PEER REVIEW DRAFT 12 JUNE 2017 Environmental Health Criteria Document Principles and Methods to Assess the Risk of Immunotoxicit

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2 PEER REVIEW DRAFT 12 JUNE 2017

4 Environmental Health Criteria Document

6 Principles and Methods to Assess the Risk of Immunotoxicity Associated

7 with Exposure to Nanomaterials

8 9 10 NOTE TO REVIEWERS 11 The target reader group for the document will be: risk assessors in a regulatory setting, researchers, 12 and industry that needs to provide the data. 13 Peer review comments should pay attention to the technical content of the draft document. The 14 document will be edited following peer review. 15 Should reviewers propose inclusion of additional material, please provide the full citation. In 16 addition, if the material cited is not available free-of-charge, for example via “open access”, please 17 provide a copy of the material. 18 19 Comments should be submitted using the template (provided on the webpage announcing the peer 20 review) to [email protected] by Friday 21 July 2017, along with a completed WHO Declaration of 21 Interests form (also available on the web page). 22 23

26 27 28

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

29 30 © World Health Organization 2017 31 32 Information on rights, cataloguing-in-publiacation and other matters will be provided in the final document. 33 This is a peer review draft, not for citing, quoting or distribution. 34

35 General disclaimers. The designations employed and the presentation of the material in this publication do not imply the 36 expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or 37 of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent 38 approximate border lines for which there may not yet be full agreement. 39 40 The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or 41 recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, 42 the names of proprietary products are distinguished by initial capital letters. 43 All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the 44 published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the 45 interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. 46 . 47 This publication contains the collective views of an international group of experts and does not necessarily represent the 48 decisions or the policies of WHO. 49 50 51 52 53 54 55

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

56 ACKNOWLEDGMENTS (to be updated upon finalization of the document) 57 This Environmental Health Criteria Document was developed by WHO with the support of the WHO 58 Collaborating Centre for Immunotoxicology and Allergic Hypersensitivity at the National Institute for 59 Public Health and the Environment (RIVM), Bilthoven, the Netherlands. 60 61 A Scoping Meeting held from 8-9 April 2015, at RIVM, Bilthoven, the Netherlands, developed an 62 outline of the document. The meeting was attended by the following: 63 Invited experts: 64 James BONNER, North Carolina State University, USA 65 Yolanda Irasema CHIRINO LOPEZ, Universidad Nacional Autónoma de México, Mexico 66 Bengt FADEEL, Karolinska Institutet, Box 210 67 Sabina HALAPPANAVAR, Health Canada, Canada 68 Jun KANNO, National Institute of Health Sciences, Japan 69 Kee Woei NG, Nanyang Technological University, Singapore 70 Kai SAVOLAINEN, Finnish Institute of Occupational Health, Finland 71 Henk van LOVEREN, RIVM, the Netherlands (Meeting Chair) 72 Il Je YU, Hoseo University, Republic of Korea 73 Representatiaves (Observers); 74 Laura ROSSI, European Chemicals Agency (ECHA), Finland 75 Jihane EL GAOUZI, OECD, France 76 The WHO Secretariat was provided by Carolyn VICKERS, WHO, Switzerland. 77 78 The invited experts contributed to the development of a first draft, which was discussed at a meeting 79 held from 30 March – 1 April 2016, at RIVM, Bilthoven, the Netherlands. The meeting was attended 80 by the following: 81 Invited experts: 82 Harri ALENIUS, Finnish Institute of Occupational Health, Finland 83 James BONNER, North Carolina State University, USA 84 Yolanda Irasema CHIRINO LOPEZ, Universidad Nacional Autónoma de México, Mexico 85 Wim H. DE JONG, RIVM, the Netherlands 86 Janine EZENDAM, RIVM, the Netherlands 87 Bengt FADEEL, Karolinska Institutet, Box 210 88 Sabina HALAPPANAVAR, Health Canada, Canada 89 Kee Woei NG, Nanyang Technological University, Singapore 90 Margriet PARK, RIVM, the Netherlands 91 Barbara ROTHEN-RUTISHAUSER, Université de Fribourg, Switzerland 92 Kai SAVOLAINEN, Finnish Institute of Occupational Health, Finland 93 Susan WIJNHOVEN, RIVM, the Netherlands 94 Henk van LOVEREN, RIVM, the Netherlands (Meeting Chair) 95 Il Je YU, Hoseo University, Republic of Korea 96 Representatiaves (Observers); 97 Laura ROSSI, European Chemicals Agency (ECHA), Finland 98 The WHO Secretariat was provided by Carolyn VICKERS, WHO, Switzerland. 99 100 Revisions to the document were made by the experts following the March 206 meeting, under the 101 Chairmanship of Henk van Loveren, in order to prepare the peer review version. 102 103 WHO intends to convene a final review group meeting to consider the peer review comments 104 received. 105 106

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

107 ABBREVIATIONS AND ACRONYMS (to be provided in the final document). 108 109 110

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

111 EXECUTIVE SUMMARY 112 113 Engeneerd nanomaterials (ENMs) or groups of ENMs may be associated with environmental and 114 human health hazards harmful effects on health [Donaldson et al., 2012]. There are different classes 115 of nanomaterials, such as carbon-based (mainly composed of carbon and examples include carbon 116 nanotubes and fullerenes); metal-based (include nano forms of titanium dioxide, gold, silver, and 117 quantum dots); dendrimers (these are polymers and are mainly used in drug delivery), and 118 composites (combination of nanomaterials; for example, carbon nanotubes coated with nano 119 titanium dioxide). These materials are being readily applied in various nanotechnology products 120 which can be categorized into four generations. 1) The first generation nanoproducts include simple 121 passive nanostructures that are used in various consumer products such as cosmetics, food, etc. The 122 properties or functions of nanomaterials are not expected to change during their use 2) The second 123 generation products include active nanostructures, the properties or function of which are expected 124 to change during their use. These changes can be unintentional in response to their local 125 environment or intentional such as in the case of drug delivery. 3) The products of the third 126 generation nanotechnology involve uses of both passive or active nanostructures to build systems of 127 nanosystems. For example, the development of synthetic organs or engineered microbes; self 128 assembling materials that assemble into new structures in the body upon their release. 4) The forth 129 generation nanostructures, which are not accomplished yet involve molecular nanosystems with a 130 specific function that include molecular devices such as in genetic therapy.. While the technology 131 itself has progressed rapidly and thousands of consumer products containing these materials are 132 beginning to appear on the market, the toxicology of these materials is less understood. Lack of a 133 clear understanding of how is general polulation exposed and the extent of exposure, and potential 134 hazard of these materials has been one of the major impediment. 135 136 Extensive research conducted in the past 20 years shows that size, shape and chemistry are among 137 the important properties that influence the interaction of nanomaterials with biological systems. 138 However, to date, no studies have conclusively shown the unique effects of nanomaterials on 139 biological systems. Considering the rapid development of technology, sheer number of 140 nanomaterials with diverse properties and the constrains associated with the conventional single 141 endpoint based toxicity assessment methods, it will be impossible for toxicologists and risk assessors 142 to keep pace with the rapidly evolving technology. Instead a new science of risk bringing together the 143 physical, chemical, and biological sciences to address the challenges presented by these novel 144 materials is needed. 145

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

146 The immune system consists of an innate part component that is able to directly respond to foreign 147 agents immediately after exposure regardless of the type of stimulus and an adaptive part 148 component that is targeted and develops over time. needs some time to develop an immune 149 response. Based on the external stimuli or toxic substance, both components may be activated, and 150 the end result could be Both of these may be affected by foreign agents. This interaction may result 151 in immunosuppression, immunostimulation, and/or hypersensitivity and autoimmunity. More 152 recently numerousRecent reports reports have identified ENMs as a potential stimulants of immune 153 response that may culminate in source for eventual immunotoxicity. While there are no validated 154 methodology available to assess immunotoxicity of nanomaterials, this document has outlined 155 several assays that are conventionally used to assess chemical-induced immunotoxicity that may be 156 compliant with nanomaterial testing. As mentioned in the earlier chapter, it is not possible to test 157 every single nanomaterial with all of the test methods mentioned; however, simple rules (outlined in 158 the paragraph below) may be followed. 159 160 Lung has been the predominantly investigated target organ and lung inflammation is the most 161 reported outcome following exposure to nanomaterials. While lung inflammation in itself may not be 162 considered as immunotoxicity, prolonged stimulation of the various components of the inflammatory 163 system including opsonization and complement activation may incite pathological conditions such as 164 asthma. Thus, nanomaterials that are immumostimulants should be investigated carefully for their 165 potential to induce immunotoxicity. Several studies have shown that nanomaterials freely 166 translocate from lungs to other immune responsive organs such as spleen, liver and lymph nodes. 167 Thus, in a scenario where lung exposure to nanomaterials results in lung inflammation and if 168 nanomaterials are found to be translocated to immune responsive organs, those nanomaterials 169 should be prioritized for a comprehensive investigation by other immunotoxicity specific assays. 170 Similarly, absorption through the layers of skin is not anticipated for nanomaterials; however, if a 171 nanomaterial is shown to induce respiratory hypersensitivity, it should be investigated for its 172 potential to induce skin irritation as well. If the bulk counterpart of a nanomaterial is a known 173 immunotoxicicant, the nanoform should be considered high risk material and investigated by the 174 appropriate methods. Although, the choice of test methods should primarily be based on their 175 potential application, in case of skin application sprays, where, by stander exposure to lung via 176 inhalation of spray mist is anticipated, in addition to skin irritation and absorption tests, lung toxicity 177 should also be investigated. Thus, the choice of methods and the extent of investigation should 178 carefully consider several factors including nanomaterials properties, their potential application, 179 route of exposure among others. Moreover, more than one test method targeting the same endpoint 180 should be included in the strategy to increase the confidence in the results derived.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

181 182 There are currently no guidelines for assessing the immuno toxicologic consequences of exposure 183 nanomaterials. For hazard identification, a variety of methods are available, that in principle have all 184 been used for classical toxicity assessment of chemicals including immunotoxicity. Given the 185 multitude of nanomaterials, and the drive to minimize the use of laboratory animals for safety 186 testing, emphasis has been on in vitro methodologies. However, many of these have not yet been 187 validated for the use of testing for nanomaterials. In general terms, risk assessment of nanomaterials 188 should follow the risk assessment paradigm for chemicals, i.e. Hazard identification, Hazard 189 characterization, Exposure assessment, and Risk characterization. Currently, the design to perform 190 risk analysis should be done on a case by case basis, flexibly including the components most 191 appropriate for the material and its proposed use. 192 193 The current EHC presents the current state of the art pertaining to principles and methods to assess 194 the risk of immunotoxicity associated with exposure to nanomaterials. 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

211 CONTENTS (to be completed following peer review) 212 213 Preambular material 214 215 1. Introduction (aim of the book) 216 217 2. Nanomaterials 218 2.1. Types of nanomaterials 219 2.2.1 Carbon-based nanomaterials 220 2.2.2. Metal-based nanomaterials 221 2.2.3 Organic-based nanomaterial 222 2.2.4 Composite nanomaterials 223 2.3. Functionalization 224 2.4 Next ENM generations 225 2.5 Life cycle of nanomaterials 226 2.6. Interaction of nanoparticles with physiological fluids 227 228 229 3. Human exposure 230 3.1 Exposure to nanomaterials 231 3.1.1 general Exposure 232 3.1.2 Occupational exposure 233 3.1.3 Consumer exposure 234 3.2 Route of exposure 235 3.2.1 Inhalation exposure 236 3.2.2 Oral exposure 237 3.2.3 Dermal exposure 238 3.2.4 Other exposure 239 3.3 Epidemiology and health surveillance 240 3.4 Toxicokinetics 241 3.4.1 Introduction to toxicokinetics 242 3.4.2 Methods to evaluate toxicokinetics of nanomaterials 243 3.4.3 Absorption, distribution, metabolism and excretion 244 3.4.3.1 Absorption 245 3.4.3.1.1 Skin

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

246 3.4.3.1.2 Gastrointestinal tract 247 3.4.3.1.3 Respiratory tract 248 3.4.4 Distribution, Metabolism and Excretion 249 3.5 Specific considerations for toxicokinetic studies on nanoparticles 250 251 4. Mechanisms of Immunotoxicity 252 4.1 General princples of toxicity of nanomaterials 253 4.1.1 Role of physico-chemical material properties 254 4.1.2 Direct cytotoxic effects of nanomaterials 255 4.1.3 Altered cellular signaling by nanomaterials 256 4.1.4 Role of oxidative stress for nanotoxicity 257 4.1.5 Impact of the bio-corona on nanoparticles 258 4.2 Nanoparticle interaction with the immune system 259 4.2.1 The innate immune system 260 4.2.1.1 Physical and biological barriers 261 4.2.2.2 Phagocytosis and clearance of nanoparticles 262 4.2.1.3 Cellular and soluble mediators of inflammation 263 4.2.1.4 Inflammasome activation by nanomaterials 264 4.3 The adaptive immune system 265 4.3.1 Nanomaterials and the adaptive immune system 266 4.4 Immunosuppression versus immune activation 267 4.5 Immunogenicity of engineered nanomaterials 268 4.6 Nanoparticles as adjuvants 269 4.7 Immunologically susceptible populations 270 4.8 The role of the biomolecule corona 271 4.9 The microbiome and the immune system 272 4.10 Routes of nanomaterial exposure 273 4.10.1 Dermal exposure to engineered nanomaterials 274 4.10.1.1 Overview of skin immunology 275 4.10.1.2 Dermal toxicity of nanomaterias 276 4.10.1.3 Results from in vivo studies 277 4.10.1.4 Results from in vitro studies 278 4.10.2 Impact of nanomaterials on the respiratory system 279 4.10.2.1 Dosimetry of nanomaterials 280 4.10.2.2 Translocation of nanomaterials

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

281 4.10.2.3 Nanomaterial interaction with lung cells 282 4.10.2.3 Alveolar Macrophages 283 4.10.2.4 Lung Epithelium 284 4.10.2.5 Dendritic Cells 285 4.10.2.6 Lymphocytes 286 4.10.2.7 Neutrophils 287 4.10.2.8 Eosinophils 288 4.10.2.9 Mast Cells 289 4.10.2.10 Animal models of immune-mediated lung disease 290 4.10.2.10.1 Pulmonary Fibrosis 291 4.10.2.10.2 Asthma 292 4.10.2.11.3 Pleural Disease 293 4.10.2.11.4 Animal models of disease susceptibility 294 4.10.2.11.5 Pre-existing Allergic Lung Disease 295 4.10.2.116 Transgenic Models 296 4.10.3 Exposure to nanomaterials via the gastrointestinal tract 297 4.10.3.1 Importance of oral exposure to nanomaterials 298 4.10.3.2 Results from in vivo studies 299 4.10.3.2.1 Carbon-based nanomaterials 300 4.10.3.2.2 Metal-based nanomaterials 301 4.10.3.3.3 Organic nanomaterials 302 4.10.3.3 Results from in vitro studies 303 4.10.3.3.1 Carbon-based nanomaterials 304 4.10.3.3.2 Metal-based nanomaterials 305 4.10.3.3.3 Organic nanomaterials 306 4.11 Human studies of nanomaterials 307 4.12 Other Exposure Routes 308 4.13 Coexposure with Allergens 309 310 5. Dosimetry 311 5.1 Introduction 312 5.2 Dispersion methods 313 5.3 Methods for administration in vivo; nose only vs whole body exposure 314 5.4 Comparison of constant vs initial bolus…exposure 315 5.5 Continuous exposure to smaller doses of NM or exposure to bolus and maintenance

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

316 thereafter 317 318 6. Hazard identification 319 6.1 Introduction 320 6.2 Test material considerations 321 6.2.1. Endotoxin contamination 322 6.2.2. Storage and handling of nanomaterials in laboratories 323 6.2.3. Endotoxin detection methods 324 6.2.4. Recommendations or suggested modifications for testing endotoxin contamination 325 6.3. Preparation of nanomaterials for exposure and characterization 326 6.4. Considerations on animal species, exposure duration and levels, routes of exposure 327 6.4.1. In vivo 328 6.4.1.1. Animal species 329 6.4.1.2. Routes of exposure 330 6.4.1.2.1 Inhalation route 331 6.4.1.2.2 Oral route 332 6.4.1.2.3 Dermal route 333 6.4.1.3. Duration and levels of exposure 334 6.4.1.3.1 Multiple exposures versus single exposure 335 6.5 Considerations on in vitro cell types, exposure methods, duration and levels 336 6.5.1 Specific cell types of the respiratory system - cells of the innate immunity 337 6.5.1.1. Epithelial cells and epithelial barrier 338 6.5.1.2. Phagocytes (neutrophils and macrophages) 339 6.5.1.3. Other innate immune cells 340 6.5.2. Specific cell types of the respiratory system - cells of the adaptive immunity 341 6.5.2.1 Dendritic cells 342 6.5.2.2 T cells 343 6.5.2. 3 B cells 344 6.5.2.4 Fibroblasts 345 6.5.3 Specific cell types of the respiratory system – co-culture systems 346 6.5.4 Ex vivo precision cut lung slice method to investigate nanomaterial induced tissue 347 6.6. Testing considerations and methods applicable to inhalation route (respiratory system) 348 6.6.1 In vivo 349 6.6.1.1. Mode of lung deposition 350 6.6.1.1.1 Inhalation:

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

351 6.6.1.1.1.1 Characterization of particles in the exposure chamber 352 6.6.1.1.2 Intratracheal instillation 353 6.6.2 In vivo testing methods applicable to inhalation route (respiratory system) 354 6.6.2.1. Body weights 355 6.6.2.2 Clinical pathology 356 6.6.2.3 Bronchoalveolar lavage fluid (refer to GD 39 for details on the test 357 6.6.2.3.1. BAL total cell counts, differential count 358 6.6.2.3.2. BAL cytotoxicity assay 359 6.6.2.3.3. BAL cytokine and chemokine profile 360 6.6.2.4. Lung tissue cytokine and chemokine profiles 361 6.6.2.5. Cellular damage 362 6.6.2.6. Cellular proliferation 363 6.6.2.7. Pleural lavage 364 6.6.2.8 Tissue Histopathology 365 6.6.2.9 Gross pathology and organ weights 366 6.6.3 In vitro testing methods applicable to inhalation route (respiratory system) 367 6.6.3.1 Immune cells damage 368 6.6.3.2 Ex vivo phagocyte function assay 369 6.6.3.3 Phagocytic activity of granulocytes, monocytes and phagocytic respiratory burst 370 6.6.3.4 Lymphocyte proliferation 371 6.6.3.5 Immune suppression 372 6.6.3.6 Respiratory sensitization 373 6.7.0. Testing considerations and methods applicable to dermal route (skin exposure) 374 6.7.1 Skin epithelium 375 6.7.2 Preparation of nanomaterials for dermal exposure 376 6.7.3 In vivo skin pre-testing considerations 377 6.7.4 In vitro/ex vivo and synthetic skin models 378 6.7.5 Skin cell cultures 379 6.7.6 Skin test methods for immunotoxicity 380 6.7.6.1 In vivo Skin sensitization 381 6.7.6.1.1 Human Repeated Insult Patch Test 382 6.7.6.1.2 Guinea Pig Maximization Test and the Buehler Test 383 6.7.6.1.3 The Local Lymph Node Assay 384 6.7.6.2 In vivo skin irritation test 385 6.7.6.2.1 The HRIPT test

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

386 6.7.6.2.2 OECD 404 (adopted in 1981, first revision in 1992, second revision in 2002) 387 6.7.6.3 In vitro skin sensitization tests 388 6.7.6.3.1 The Direct Peptide Reactivity Assay (DPRA) (OECD 442C) 389 6.7.6.3.2 ARE-Nrf2 Luciferase (OECD 442D) 390 6.7.6.3.3 In vitro skin irritation 391 6.7.6.4 Skin responses in susceptible populations 392 6.8.0. Testing considerations and methods applicable to oral route 393 394 7. Approaches to risk assessment 395 7.1. Introduction - the basic risk assessment paradigm, question/problem formulation, define 396 goals, scope, focus of the risk assessment 397 7.1.1. Hazard identification 398 7.1.2 Dose-response assessment 399 7.1.3. Exposure assessment 400 7.1.4. Risk characterization 401 7.1.5 Major issues with applying the existing HHRA paradigm to nanomaterials 402 7.1.5.1 Hazard identification process – challenges 403 7.1.5.2 Issues with dosimetry 404 7.1.5.3 Issues with exposure assessment 405 7.1.5.4 Issues with risk characterization 406 7.1.6 Conclusions – application of conventional risk assessment approach to nanomaterials 407 7.2. Alternative risk assessment approaches for nanomaterials 408 7.2.1 A tiered approach to human health risk assessment of nanomaterials 409 7.2.2 A stepwise read across approach 410 7.3. Nanomaterial immunotoxicity assessments 411 7.3.1 Summary of existing immunotoxicity risk assessment approaches for chemicals and 412 Pharmaceuticals 413 7.3.1.2 Levels of evidence for immune system toxicity 414 7.3.2 Nanomaterial immunotoxicity risk assessment 415 7.3.2.1 Animal studies 416 417 8 Future Research 418 8.1 Immunotoxicology risk assessment paradigm 419 8.2 High content omics testing 420 8.3 Adverse outcome

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

421 8.4 High throughput screening 422 423 9. Conclusions and Recommendations 424 425 10 Case studies 426 10.1 Carbon nanotubes 427 10.2 Silver nanoparticles 428 429 11 Glossary of Terms 430 431 12 References 432

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

433 1. INTRODUCTION 434 435 Scope of the document 436 437 The purpose of this document is to present an overview on the current knowledge and evidence on 438 principles and basic mechanisms of immunotoxicity caused by engineered nanomaterials (ENMs). 439 ENMs are encountered occupationally during the manufacturing process, nanomaterials present in 440 consumer products or nanomaterials that might be encountered in the environment after release 441 the manufacturing process or from consumer products. . ENMs have been shown to have the ability 442 to negatively affect the immune system. This document provides guidance on principles and 443 methods for hazard and risk assessment of different ENM and groups of ENM on the immunological 444 system in the body. Hence, the key cell types and elements and the functioning of the human 445 immunological system will be described, and key effects of various ENM on these cells and elements 446 of the immune system will be addressed. 447 The potential users of the document include the scientific community, health care professionals, 448 regulators and decision makers at the national and international levels, industry and industrial 449 associations, and important civil society organizations such as the social partners, and other civil 450 society interests groups. 451 452 Engineered nanomaterials and nanotechnologies 453 454 Nanotechnology - the ability to manipulate materials at or near-the atomic scale to produce new 455 structures, materials, and devices – holds possibilities for scientific advances in many areas including 456 consumer products, energy, as well as for revolutionizing some areas of large-scale manufacturing. 457 By arbitrary definition, nanotechnology deals with structures within the length range of 458 approximately 1 to 100 nm (ISI/TS 8004-2015). The very small size results in a relatively large surface 459 area that conveys unique specific properties to the nanoscale. Nanotechnology has the potential to 460 improve existing technologies and to enable emerging technologies to blossom [see Nickel et al., 461 2000; Savolainen et al. 2013]. Such important areas include construction industry, energy industry, 462 automobile manufacturing, pharmaceutical industry, food industry and agricultural industry [Ngô and 463 Van de Voorde, 2014; Vogel et al., 2014]. Specific applications are energy production and storage, 464 catalytic reactions, microelectronics, plastics and polymers, concrete, coatings and paints, cosmetics 465 and medical products. It is also noteworthy that ENM alone do not carry additional value in most 466 cases. They enable added value when combined with existing and emerging technologies. Indeed, 467 nanotechnology is considered as one of the ‘key enabling technologies”in the reseach and innovation

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

468 program of the European Commission (https://ec.europa.eu/programmes/horizon2020/en/what- 469 horizon-2020). 470 However, it has appeared over the years that some of the ENMs or groups of ENMs may be 471 associated with harmful effects on health [Donaldson et al., 2012]. Hence, it has become important 472 to be able to distinguish those materials which may be harmful from the innocuous ones, and to 473 develop methods for hazard identification, exposure and risk assessment for these materials which in 474 many ways differ from the traditional, soluble chemicals [Hartung and Sabbioni, 2011]. Because of 475 the diversity of potential harmful effects of ENMs it has also become necessary to identify the key 476 target organs or organ systems for potential toxic effects and risk assessment. It is therefore also 477 important to emphasize that describing ENMs as a uniform group of materials is misleading 478 (Maynard et al., 2006). There may be hundreds of thousands of ENM synthetized in the laboratory; 479 alone carbon nanotubes (CNT) alone exist in more than 50,000 different forms [Schulte et al., 2008]. 480 Hence, when discussing various ENMs, one has to be specific, and indicate the special features of the 481 nanomaterial in question, and to properly characterize the physico-chemical features in detail (e.g. 482 the size, surface area, surface to volume ratio, reactivity, chemical composition, structure, 483 crystallinity, aspect ratio, tensile strength, electrical conductivity, persistency and dissolution rate, 484 particle size distribution, and chirality) [Roco et al., 2010; NAS, 2012; Hedmer et al., 2013]. 485 The ENMs in the widest use include carbon black, amorphous silica, nano-scale metal particles such 486 as silver or gold, nano-scale metal oxides such as TiO2, iron oxide or cerium oxide, as well as carbon- 487 based materials such as fullerenes, single- (SWCNT) and multi-walled carbon nanotubes (MWCNT), 488 and carbon nanofibers. These ENMs possess specific properties that are determined by their shape, 489 size and dissolution in medium. In general three broad categories of ENMs can be distinguished: 490 carbon-based, metal and metal oxides, and organic ENMs. In view of the multitude of ENM variations 491 a proper physicochemical characterization of the ENM under investigation is essential for both 492 hazard and the risk assessment (Bhattacharya et al., 2015). In addition, such a characterization is also 493 necessary for a proper identification of the ENM (e.g. is the nanomaterial as used in the final product 494 the same as the one for which the risk assessment is done?). 495 It is not surprising that ENMs, consumer products incorporating ENMs, and nanotechnologies taking 496 advantage of ENMs in industrial materials and processes have attracted a remarkable amount of 497 attention during the past decade. The technological and beneficial properties of ENMs are 498 attributable to their unique properties described accompanying the nanoscale [Roco et al., 2010; 499 NAS, 2012]. Several of the novel applications of ENMs have now been incorporated into a variety of 500 popular consumer products, e.g. better protection against ultraviolet (UV) radiation by sun-block 501 creams containing nano-scale titanium dioxide (TiO2) as effective UV-filters. ENM are also now 502 widely used in other consumer items such as electronics, cosmetics, clothes, cleaning materials such

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

503 as bleach, sportswear and other sport products such as cross-country skis or tennis rackets. 504 Nanotechnologies and materials are also used in a number of professional protective clothing items 505 and several military applications [See: The Project of Emerging Nanotechnologies (PEN); 506 http://nanotechproject.org. 507 Industrial applications include, in addition to those mentioned above, computer hard drives with 508 much higher memory capabilities, as well as incorporation of these materials into semiconductors. 509 The unique properties of ENMs may confer remarkable economic benefits by increasing the capacity 510 of various types of information and communication technology (ICT) devices. They also may achieve 511 it possible to make savings by reducing material needs e.g. increasing the strength of concrete in the 512 construction industry means that the amount of concrete and other building materials can be 513 significantly cut back. An important nanomaterial used for this purpose is amorphous silica. The 514 economically most important applications of ENMs are likely to be found in industrial applications 515 such as coatings, optical and printed electronics, applications of nano-cellulose, and energy 516 production as well as increasingly in the future in clean-tech applications and prevention of climate 517 change by reducing energy through more efficient batteries and power sources, as well as more light 518 weight composites for cars and aircraft that will reduce fuel emissions. 519 ENM are being also used as oil (nano-diamonds) and gasoline (cerium oxide) additives where they 520 reduce friction (diamonds) or decrease the release of carbon dioxide (cerium oxide improving the 521 burning of gasoline) [Roco et al., 2010]. 522 The number of consumer products and expected size of the market products incorporating ENMs by 523 2020 has been predicted tp grow dramatically (Roco et al., 2010, The Project of Emerging 524 Technologies (PEN)). 525 Notwithstanding these beneficial properties, in some cases ENMs can also be harmful to human 526 health or the environment. In fact, many of the characteristic properties of ENMs, small size - at least 527 one dimension between 1 – 100 nm, large surface area per weight, and high surface reactivity, that 528 make them technologically so valuable, are unfortunately the very reasons for their potentially 529 harmful effects [Nel et al., 2009; Feliu and Fadeel, 2010; Savolainen et al., 2010]. In fact, the small 530 size of ENMs enable them to enter the body and penetrate biological barriers much better than their 531 chemically identical but larger counterparts (Pietrouisti et al., 2013). The small size may also have 532 unexpected effects on the kinetics, especially distribution and excretion from the body. 533 The number of consumer products and expected size of the market of products incorporating ENMs 534 by 2020 has been predicted to grow dramatically [Roco et al., 2010, The Project of Emerging 535 Nanotechnologies (PEN)]. 536

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

537 Nevertheless, inspite of the increasing knowledge of some of the possible harmful health effects of 538 ENMs, the use and the number of applications of ENM have dramatically increased during the last 539 ten years, and the growth continues at an increasing speed due to the technological and economical 540 benefits of these materials. 541 542 One major challenge in assessing the risks of safety of these materials is the lack of systematic 543 knowledge about exposure to these materials. There has been very limited systematic research into 544 the hazards of ENMs for human health or the environment with most of the studies concentrating on 545 only a few nanomaterials [see Brouwer 2010; Bouwmeester et al., 2011; Kahru and Savolainen 2010]. 546 The amount of data about exposure in workplaces and the general environment is even more 547 restricted and in fact only a few studies have been published [Brouwer 2010; Kahru and Savolainen 548 2010, NAS, 2012; Guseva Canu et al., 2015]. Without this basic information, it is not possible to 549 conduct an adequate assessment of risks of exposure to different nanomaterials. 550 Data on hazards and exposure are essential in performing a quantitative risk assessment as indicated 551 by the well-known equation: 552 553 “hazard x exposure = risk” 554 555 [NRC, 1983]. Even though some of these ENMs have been identified as being hazardous [Gulumian 556 and Savolainen, 2012; Palomäki et al., 2011; Rossi et al, 2010; Shvedova et al., 2007], it is likely, as 557 with other chemicals that many if not most of engineered nanomaterials will prove to be harmless or 558 only marginally dangerous [European Commission 2012]. However, our abilities to identify the 559 harmful materials from their harmless counterparts are very limited, and this complicates the 560 assessment of safety and risks of ENMs. The challenges to differentiate between harmful and 561 harmless ENMs which are due to limitations of the current hazard, exposure and risk assessment 562 paradigms, are a cause for concern among consumers, regulators and the industries using these 563 materials. These concerns related to the uncertainties of the health and environmental effects of 564 ENMs have been identified by the European Commission as a major obstacle preventing the efficient 565 transfer of these materials into the activities of existing and emerging technologies [European 566 Commission 2012]. 567 If one considers the specific alarms surrounding the potential health effects and safety of ENMs, the 568 concerns started to increase as the use of these materials became more widespread. The first 569 publications on potential health effects of these materials were published in the 1990’s [Linnainmaa 570 et al. 1997; Oberdorster et al., 1994]. It was only at the beginning of this century that 571 nanotoxocology became identified as a topic for ENM safety evaluation (Oberdorster et al., 2005;

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

572 2007). Similar to other particles, ENM have the potential to induce lung inflammation after inhalation 573 exposure which resulted in an interest based on occupational safety. Oberdörster et al. [Oberdörster 574 et al. 2004] and Elder et al. [Elder et al. 2006] conducted work in experimental animals and reported 575 that inhalational exposure to engineered nanomaterials could lead to the uptake of the particles by 576 the brain. In this case, axons of the olfactory nerve in the olfactory epithelium could transport the 577 particles into the olfactory bulb under the frontal cortex. Media picked this item to the headlines and 578 emphasize the possibility of brain damage to humans which has been shown largely unjustified by 579 later studies. These materials can be distributed to other parts of the brain after their uptake through 580 the nose in experimental animals, but there is no evididence on effects in humans, but some on 581 animals. Several studies have associated titanium dioxide nanoparticles with pulmonary 582 inflammation (Oberdorster et al., 1994; Grassian et al., 2007; including transient effects (Ma-Hock et 583 L,M 2009; Warheit et al., 2006). Exposure of single walled carbon nanotubes (SWCNT) resulted in 584 progressive pulmonary fibrosis [Shvedova et al., 2007] and multiwalled carbon nanotubes (MWCNT) 585 in pulmonary granulomas and inflammation [Shvedova et al. 2009] and even asbestos like changes in 586 the peritoneal cavity of experimental animals [Poland et al., 2008; Takagi et al., 2008; Takagi et al. 587 2012]. Rhyman-Rasmussen et al. (2009) have shown that carbon nanotubes can also reach the 588 subpleural tissue of the lungs and lead there to asbestos-like fibrosis and collagen accumulation. 589 In spite of the expansion of the research into the potential health effects of ENMs, none of the ENMs 590 has been systematically studied to allow the performance of a thorough evidence-based risk 591 assessment of ENMs [Kahru and Savolainen 2010; Savolainen et al., 2010; other references e.g. from 592 NIOSH by Schulte and/or Geraci ]. The lack of information about exposure to these materials is even 593 more noteworthy [Brouwer 2010]. In other words, our ignorance of both hazard and exposure data 594 on these materials means that a realistic and reliable risk assessment is difficult, if not impossible, at 595 present. 596 The predominating challenge associated with the use of these materials in various applications of 597 nanotechnologies is the uncertainty associated with their potential health effects. This uncertainty 598 represents a major obstacle to the producers of these materials as well as worrying the enterprises 599 that incorporate these materials into consumer and other products. Hence, one of the major 600 challenges facing the nanotechnologies today is to alleviate the doubts about the safety of ENMs 601 [Fadeel and Savolainen, 2013; Nature Nanotechnology Editorial, 2012; Savolainen and Alenius, 2013; 602 Manyard, 2015)]. Until the issue of safety assessment of engineered nanomaterials has been 603 resolved by research both on the potential hazard and exposure to ENMs, uncertainty of the safety 604 of these materials will remain an obstacle for the beneficial applications of engineered nanomaterials 605 in a number of technological applicationsin Europe and globally. 606

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

607 608 Immunotoxicity testing 609 610 The immune system evolved primarily to protect the host from infectious or neoplastic disease 611 (Germolec 2004). It consists of several organs and specialized cells types throughout the body 612 (Boraschi et al., 2017). As such, a normal functioning of the immune system presents a defense 613 against invading microorganisms, foreign (xenobiotic) materials and tumors, that would otherwise 614 result in disease. However, deregulation of the immune system may result in allergic and/or 615 autoimmune responses. In the modern world, humans are exposed to an ever increasing variety of 616 xenobiotics including chemicals, metals, drugs, and ENMs that disrupt immune system homeostasis 617 and normal functionality. Immunotoxicology is the study of interaction of xenobiotic agents such as 618 chemicals and drugs with the immune system. The immune system consists of an innate part that is 619 able to directly respond to foreign agents and an adaptive part that needs some time to develop an 620 immune response. Both of these may be affected by foreign agents. This interaction may result in 621 immunosuppression, immunostimulation, and/or hypersensitivity and autoimmunity. More recently 622 numerous reports have identified ENMs as a potential source for immunotoxicity Hussain et al., 623 2012; Boraschi et al., 2012; Dobrovolskaia et al., 2016).. 624 Immunotoxicity caused by chemical exposure is a topic that has been addressed in previous EHC 625 documents. For example, the International Programme on Chemical Safety (IPCS) Environmental 626 Health Criteria monograph 180: Principles and Methods for Assessing Direct Immunotoxicity 627 Associated with Exposure to Chemicals (IPCS, 1996) reviewed the causes, consequences, and 628 detection of disorders mediated by “immunotoxicity”. The IPCS Environmental Health Criteria 629 monograph 212: Principles and Methods for Assessing Allergic Hypersensitization Associated with 630 Exposure to Chemicals (IPCS, 1999) focused on mechanisms, clinical aspects, epidemiology, hazard 631 identification, and risk assessment of allergy and hypersensitivity following exposure to certain 632 chemicals. In addition adocument was published considering induction of autoimmunity due to 633 chemical exposure. (International Programme on Chemical Safety (IPCS) Environmental Health 634 Criteria monograph 236: Principles and Methods for Assessing Autoimmunity Associated with 635 Exposure to Chemicals (IPCS, 2006)). A special issue of “Methods” published in 2007 was dedicated 636 to the use of animal models for determining immunotoxicity (Methods Volume 41, January 2007). At 637 the moment it is not yet clear whether existing methods will be applicable for EMNs as well or 638 whether specific methods may be needed in view of the special characteristics of ENMs. In general 639 the OECD testing guidelines are applicable but some adaptations may be needed in the way the 640 actual testing will be performed. Immunotoxicity testing of chemicals is done in a tiered approach. In 641 the first Tier indications for effects on organs of the immune system are identified in regular toxicity

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

642 tests like a 28 or 90 day repeated dose toxicity study. In the second Tier more specifically immune 643 function assays may be applied like antibody formation and resistance against infectious agents (EHC 644 180, IPCS 1996). Although these tiered studies are performed in animals, new and improved in vitro 645 tests and in silico approaches are also being developed that will become available to study the 646 immune system. 647 This criteria document is intended to focus on ENM that would be encountered occupationally during 648 the manufacturing process, nanomaterials present in consumer products, or nanomaterials that 649 might be encountered in the environment after release the manufacturing process or from consumer 650 products. A special effort will be made to discuss only immunotoxicity studies that used well- 651 characterized ENMs. While it is acknowledged that nanomaterials used in emerging field of 652 nanomedicine (drug delivery, imaging, etc) could pose immunotoxic side-effects, this topic will not be 653 covered in this criteria document. 654 655

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

656 2. TYPES OF NANOPARTICLES

657 The emergence of nanotechnology has brought a burst in the synthesis of a broad range of 658 nanoparticles which are not only difficult to define but also to classify since the variety of them is 659 huge. However, some organizations, including including The Royal Society & the Royal Academy of 660 Engineering of the United Kingdom (2004), The European Chemicals Agency (2006), the Scientific 661 Committee on Emerging and Newly Identified Health Risks (2007) , the European Union (2011), and 662 the International Organization of Standardization (2015) have put forward a definition of 663 nanomaterial. Even if they are not exactly the same, definitions converge in structures sized between 664 1 to 100 nm in at least one of their dimensions (Table 1, Taken from Garduño-Balderas et al., 2015) 665 independently of the shape, composition, solubility or any characteristic. On the other hand, 666 according to the chemical composition, most nanomaterials can be categorized into carbon-based, 667 metal-based, organic nanomaterials and composites. In this chapter, a description of the properties 668 of these four types of nanoparticles is provided.

Table 1. Definition of nanoparticles

Organization NM definition

The International A nano-object (a discrete piece of material with one, two or three external Organization for dimensions in the nanoscale (1-100nm)). Standardization

(ISO/TS: 27687:2008)

Those which have structured components with at least one dimension less than 100 nm. Materials that have one dimension in the nanoscale and are extended in the other two dimensions, such as a thin films or surface Royal Society & the Royal coatings. Some of the features on computer chips come in this category. Academy of Engineering of Materials that are nanoscale in two dimensions (and extended in one the United Kingdom dimension) include nanowires and nanotubes. Materials that are (United Kingdom) nanoscale in three dimensions are particles, for example precipitates, colloids and quantum dots (tiny particles of semiconductor materials). Nanocrystalline materials, made up of nanometer-sized grains, also fall into this category.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Nanomaterial’ means a natural or manufactured active substance or non- Regulation (EU) No active substance containing particles, in an unbound state or as an 528/2012 of the European aggregate or as an agglomerate and where, for 50 % or more of the Parliament and of the particles in the number size distribution, one or more external dimensions council is in the size range 1-100 nm. Fullerenes, graphene flakes and single-wall

carbon nanotubes with one or more external dimensions below 1 nm shall

be considered as nanomaterials.

Scientific Committee on Emerging and Newly Identified Any form of a material that is composed of discrete functional parts, many Health Risks of which have one or more dimensions of the order of 100 nm or less. (SCENIHR; European Union) Natural, incidental or manufactured material containing particles, in an unbound state or as an aggregate or as an agglomerate and where, for European Union 50% or more of the particles in the number size distribution, one or more Recommendation external dimensions is in the size range 1 nm-100 nm. In specific cases 2011/696/EU and where warranted by concerns for the environment, health, safety or competitiveness, the number size distribution threshold of 50 % may be replaced by a threshold between 1 and 50%.

Industrial materials intentionally produced, manufactured or engineered to have unique properties or specific composition at the nanoscale, that is a NICNAS Working Definition size range typically between 1 nm and 100 nm, and is either a nano-object of Industrial Nanomaterials (i.e. that is confined in one, two, or three dimensions at the nanoscale) or (Australia) is nanostructured (i.e. having an internal or surface structure at the nanoscale.

USEPA has no formal definition. The criteria considered is as follows: A substance that is solid under 25oc under atmosphereic pressure, manufactured or processed so that its primary particles, aggregates, or Environmental Protection agglomerates are 1-100 nm in size, giving them unique and novel Agency, USA characteristics or properties. A similar particle size distribution of more than 10 % by weight in the range of 1-100 nm size is also recommended by The American Chemistry Council.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Any manufactured substance or product and any component material, ingredient, device, or structure to be nanomaterial if it is at or within the nanoscale in at least one external dimension, or has internal or surface structure at the nanoscale, or if it is smaller or larger than the nanoscale in Health Canada all dimensions and exhibits one or more nanoscale properties/phenomena”. Nanoscale properties/phenomena refer to properties that are attributable to the size of the substance and size effects.

669 670 671 2.1 Types o nanoparticles 672 Based on the Project on Emerging Nanotechnologies (PEN) (http://nanotechproject.org), the bulk of 673 nano-enabled products that are commercially available today fall within the category of consume)r 674 products. Among these, the most commonly used ENMs are metal and carbon based (Fig x). Here, we 675 have categorized ENMs into three main groups based on their chemical composition, namely carbon, 676 metal and organic based and composites. Some of their characteristics are described below. 677 678 679 680 681 682 683 684 685 686 687 688 689 690 Figure X. Composition of engineered nanomaterials (ENMs) found in nano-enabled products listed in 691 the consumer product inventory that is maintained by The Project on Emerging Nanotechnologies in 692 2005. Modified from Vance et al., 2015.] . 693 694 2.2.1. Carbon-based nanomaterials 695 a) Single walled carbon nanotubes (SWCNTs)

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

696 These structures can be viewed as cylinders of graphene rolled along a lattice vector in the graphene 697 plane and this vector determines the chirality and diameter of the nanotube; the combination of 698 these parameters, in turn, decides for example, whether the nanotubes are metallic or 699 semiconductiin (Dai, 2002). Chemical vapor deposition is one of the most common methods for 700 SWCNTs synthesis using temperatures >800 °C, which allows a highly dense and well-aligned SWCNTs 701 that enhances electrical properties, however, laser ablation and, arc-discharge are also used. The 702 presence of metals withinn SWCNTs is a disadvantage of this method which can be solved by 703 selective electrical burning of metals (Park et al., 2013). 704 b) Double-walled carbon nanotubes (DWCNTs). 705 DWCNT are coaxial nanostructures composed of SWCNTs and chemical vapor deposition method of 706 methane is used for their synthesis but the diameter cannot be entirely controlled by this method, 707 however, iron disilicid can be used for a higher nucleation efficiency to obtain a more restricted 708 diameter of DWCNTs (Qi et al., 2007). However, some other reagents used for catalysis as FeMoMgO 709 in low temperature (550°C) conditions yield high purity DWCNTs (Somanathan et al., 2014). Arc- 710 discharge and Peapod method are also used for their synthesis. SWCNTs are obtained as impurities 711 during DWCNTs synthesis and amorphous carbon, but the synthesis can be purified by scavenging 712 the compunds used as catalysts in low pH followed by high-temperature oxidation. 713 714 c) Multiwalled carbon nanotubes (MWCNTs). 715 MWCNTs are multiple rolled structures and can also be synthetized by chemical vapor deposition of 716 methane, laser ablation, arc-discharge and Peapod method. Iron used as catalyst is one of the main 717 impurities during the synthesis and it can be removed by graphitization. The inner and outer 718 diameters of the MWCNTs can be modified by catalyst and the flow rate of hydrocarbon at the 719 catalyst particle surface (Fonseca et al., 1998). 720 721 2.2.2. Metal-based nanomaterials

722 a) Titanium dioxide (TiO2) 723 TiO2 exists in three different crystalline forms: anatase, rutile and brookite. TiO2 can be engineered

724 in various shapes such as nanospheres, nanobelts. It is a highly stable semiconductor compound due 725 to its wide band gap under ultraviolet light. It has electronic properties which allow its usage in the

726 electronic field. TiO2 is often used in consumer and personal care products (e.g. sunscreens and 727 cosmetics) as a UV filter. Its ability to scatter light to give a consistent and stable whitening effect has 728 also been exploited as a pigment in paints (PW6, CI77891) and food colouring additive (E171). Being 729 photocatalytic, they produce reactive oxygen species (ROS) under light and can act as anti-microbial 730 agents. TiO2 based anti-microbial sprays have emerged as a result.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

731 b) Zinc oxide (ZnO) 732 ZnO is a semiconductor compound (band gap of 3.37 eV) that exhibits high surface area to volume 733 ratio. ZnO nanoparticles are the third most commonly produced nanomaterial worldwide (Piccino et 734 al., 2012). It can be synthetized by laser ablation, hydrothermal methods, electrochemical deposition, 735 sol-gel methods, chemical vapor deposition and combustion methods (Kumar et al., 2013). Similar to

736 TiO2, ZnO is photocatalytic and is used in sunscreens as UV filters. 737 c) Silicon dioxide (SiO2)

738 SiO2, or Silica, has different crystalline forms including α-quartz, β-quartz, α -tridymite, α and β forms 739 of cristobalite, keatite, coesite, and stishovite. Crystalline silica forms can induce “miners lungs” and 740 are generally highly toxic after inhalation. Silicon-oxygen bond lengths vary between the different 741 crystalline forms, but the Si-O-Si angle also varies between a low values α-tridymite, up to high values

742 in β-tridymite. Synthetic amorphous silica (SAS) or SiO2 nanoparticles can be produced by sol–gel or

743 pyrogenic methods to obtain homogeneous size nanoparticles. SiO2 has very wide industrial uses, 744 although they are not necessarily used in the nano-scale as SiO2 preparations mainly consist of 745 aggregates/agglomerates of the primary nanosized SiO2 particles. These applications range from 746 structural fillers in composite materials to abrasive agents in toothpastes and as anti-caking agents in 747 food (E551). 748 d) Iron based

749 Magnetite (Fe3O4), hematite (α-Fe2O3) and maghemite (γ-Fe2O3) are three of the most magnetic iron 750 based nanoparticles. Those nanoparticles have magnetic properties also called superparamagnetic 751 properties, which can be used in combination with external magnetic fields for different purposes. 752 e) Silver 753 Silver is a metal which can be found pure in the nature and as an alloy with other metals. Silver 754 nanoparticles can be obtained by laser ablation and evaporation-condensation methods, among 755 others (Iravani et al., 2014). Using reducing agents such as hydrazine, ethylene glycol, UV light and 756 some others, and silver nitrate as a precursor, silver nanoparticles can be obtained. Currently, silver 757 is the most widely used ENM among known nano-enabled products (Fig x). The most common reason 758 to use silver nanoparticles is their anti-microbial ability. Products that have exploited this property 759 include textiles (clothing), electronics, food packaging, wound dressings and more. 760 f) Gold 761 Gold has been considered as precious metal during human history and it is an inert metal that can be 762 found in the nature but also is alloyed with some other metals such as copper. Gold nanoparticles 763 synthesis can be achieved by two phase-transfer methods (based on the Brust-Schiffrin method) 764 which used a thiol group that is bound to gold in an aqueous phase which is transferred to an organic

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

765 phase. Because of their inertness, gold nanoparticles are often being explored for applications in 766 nanomedicine, such as drug carriers. 767 g) Quantum dots 768 Quantum dots (QDs) are inorganic compounds with a core of elements from groups II-VI or III-V in a 769 crystalline form made up of 100 to 100 000 atoms which has high photo stability, tunable 770 fluorescence under single wavelength excitation and longer lifetime as compared to conventional 771 fluorophores (Cinteza, 2010). QDs have a metalloid core which is formed by cadmium selenide, lead 772 selenide, or indium arsenide (CdSe, PbSe, or InAs) enclosed within a shell coating, comprised of a 773 different semiconductor material of higher bandgap followed by a polymer coating (Walling et al., 774 2009). The shell protects the core from oxidation and degradation, and the polymer coating can be 775 an amphiphilic coating such as poly (ethylene glycol)-based bidentate ligands (Mei et al., 2009). 776 777 2.2.3. Organic-based nanomaterial 778 a) Plymer-based/dendrimers. 779 Tomalia and colleagues firstly were the first to describe the synthesis of dendrimers in 1984 and they 780 defined dendrimers as structures with an initiatior core, interior layers called generations and 781 composed of repeating units, radially attached to the initiator core and an exterior attached to the 782 outermost interior generation (Tomalia et al., 1984). This first synthesis was developed using 783 polyamidoamine by a Michael addition reaction with ammonia and ethylendiamine as initiator cores. 784 b) Lipid-based nanoparticles. 785 Liposomes, micelles, niosomes and nanoclays are also included in this category and most of them 786 have silicate-related compounds. The usages of nanoclays are expanding and nowadays it has been 787 suggested that it can be used in pharmaceutical products but also for nanocomposite synthesis. 788 789 2.2.4 Composite nanomaterials 790 It is also possible for nanomaterials to be made from combinations of the different multiphase 791 materials described above, for example in the form of physical blends, co-polymers, gels or core-shell 792 structured nanoparticles and can be shaped also as flowers and diamonds among others. There is no 793 limit of components to make a nanocomposite. Some can be synthetized by few components, for 794 instance, the supercapacitors made by reduced graphene oxide and titanium dioxide nanorods; some 795 others can have a NPs core and be coated and/or functionalized by other NPs. In addition, 796 nanocomposites can be made by a multilayer system, such as those designed for solar cells, which 797 can include layers of metal-based, carbon-based and organic NPs (Figure XXX). The combination of 798 different types of NPs leads to physicochemical properties which differ completely from the isolated 799 components (Fleetham et al., 2015; Ramadoss et al., 2013).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

800 801

802 803 804 Figure XXX. This figure will be properly replaced to show the shape classification of nanoparticles. 805 806 2.2 Next ENM generations 807 The synthesis or the above mentioned ENMs as primary structures were called the first ENMs 808 generation. This generation is also called passive NPs and those NPs coated or functionalization with 809 biological effects belong to the second geneneration, which has been produced since 2000. The third 810 generation is related to the capability of assembling ENMs in a more complex systems, for instance, 811 polymer grafts used as assembly-regulating molecules able to bind to the surface of NPs and guide 812 them for a specif organization used for spectroscopy. The fourth generation, which is expected to be 813 developed in this decade, will be involved in molecular devices with active functions. 814 815 2.3 Life cycle of nanomaterials 816 The focus of this document is on nanomaterials, which are intentionally produced to satisfy a 817 particular function, or unintentionally produced because of man-made activities – engineered

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

818 nanomaterials (ENMs). Nanomaterials that occur due to natural processes, such as volcanic ash, are 819 not within the scope of this discussion. Regardless of nanomaterial type, ENMs begin their life cycle 820 at the point of synthesis, go through several stages of transformation, and end with final disposal 821 (Mitrano et al., 2015), which is usually related to waste water treatment plants or landfills deliver. 822 823 Stage one 824 The first stage can be considered the manufacturing process, in which work places represent a 825 potential hazard. The manufacturing process can include synthesis of the nanoparticles or using 826 nanoparticles acquired from manufacturers as ingredients to fabricate a nano-enabled product. 827 However, not only personnel in charge of production or direct handling of the nanomaterial could be 828 exposed, but maintenance and cleaning personnel and administrative staff can also be exposed by 829 inhalation, dermal contact or by ingestion. This first stage also includes storing nanomaterials or 830 modifying the surfaces of primary nanoparticles for further processes. Industries of textiles, paints, 831 cosmetics, food and electronics involve personnel who can be exposed in occupational settings. 832 833 Stage two 834 The second stage could be related to personnel in charge of transporting nano-enabled products and 835 in this stage nanomaterials could suffer some transformations since they can have interactions with 836 packaging materials or may be subjected to varying conditions such as temperature fluctuations or 837 UV exposure. In addition, it has been estimated that between 0.1% and 2% of ENMs could be 838 released unintentionally to the environment, which includes water, atmosphere and soil. 839 840 Stage three 841 In the third stage, nano-enabled products are used by the consumer, resulting in ENMs being 842 potentially released into the environment as a consequence of usage. As examples, anti-microbial 843 treated sports wear can release silver nanoparticles during washing process, food can be poured into 844 the raw waste water releasing food-grade titanium dioxide and silicon dioxide, sunscreens containing 845 zinc oxide or titanium dioxide may be washed of and end in swimming pools and all these ENM may 846 end up and accumulate in waste water treatment plants, among others. 847 848 Stage four 849 The fourth stage can be considered only when nano-enabled products are intentionally delivered as 850 waste (end-of-life). In most of the cases, ENMs are disposed into landfills, raw sewages or 851 incinerators. These could result in uncontrolled release of ENMs into the environment through the 852 soil, water or air, and eventually result in their presence in the food chain.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

853 854 Final disposal and recycling 855 In spite of the recycling strength in other fields, there is no information about how to dispose or 856 recycle ENMs. Indeed, technical data sheet of commercial nanomaterials for research have no 857 information about final disposal and regulation for ENM waste is not fully established even in 858 developed countries. The landfills represent the most important ending of ENMs followed by soil and 859 water while the atmosphere seems to have the less percentage of ENMs. Insoluble NPs such as 860 titanium dioxide and silicon dioxide remain as solid particles, maybe as agglomerates/aggregates, 861 while some others release some ions, such as iron and zinc NPs. This also could represent a challenge 862 in terms of recycling, since some of them could not be recovered. In addition, the concept of 863 recycling ENMs has not been fully addressed by researchers or industries since it could be hard to 864 recover ENMs from waste products or polluted water/soil and might not be cost-effective. 865 866 2.3 Functionalization 867 Surface modification is performed to confer new properties to nanomaterials, which could include a 868 decrease in toxicity (but also an inadvertent increase in toxicity), enhanced solubility, binding of 869 specific molecules, and confer stability, among others. Functionalization techniques might employ 870 hydrophilic or hydrophobic molecules, which in turn can be organic or inorganic and they can consist 871 of polymeric or non-polymeric forms (coatings). Functionalization can confer new properties to the 872 primary nanoparticle, for instance, graphene oxide can be modified with poly(ethylene glycol)- 873 poly(ethylene imine) copolymers tagged with folic acid for biological purposes, and such 874 functionalization induces enhanced cellular uptake and photoluminescence (Prabhakar et al., 2015). 875 Carbon nanotubes can be functionalized by covalently attaching carboxyl or amine groups to the 876 surface to increase solubility or by atomic layer deposition coating with metal oxide for enhancing 877 conductive properties. All of these functionalizations alter macrophage innate immune responses 878 and either increase or decrease fibroproliferative responses in the lungs of mice (Wang et al., 2011; 879 Taylor et al., 2014). Some types of functionalization have no influence on nanoparticles native 880 properties, such as functionalization of QDs with polyethylene glycol, which does not interfere with 881 its optical properties (Mei et al., 2009; Karakoti et al., 2015). However, the use of polyethylene glycol 882 as coating can change the toxicokinetic behavior of the ENM as it prolongs circulation time of ENM in 883 blood (Niidome et al., 2006; Lankveld et al., 2011). 884 Iron based nanoparticles are commonly coated with polyethylene glycol, carboxydextran, 885 polyethylene imine. 886 Recently, some old molecules such as diazodium salts, are being used for new purposes of gold 887 nanoparticles, SWCNT, nanosized TiO2 and nanodiamond functionalization (Mohamed et al., 2015).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

888 889 2.4. Interaction of NPs with physiological fluids 890 The specific primary routes by which engineered nanoparticles may interact with the human body 891 include inhalation, ingestion and application to the skin. In addition, parenteral application by 892 injection is also a primary route for medical therapeutic purposes, but will not be discussed within 893 this document. Independent of the entry route, the particles inevitably encounter a complex 894 physiological fluid populated with e.g. proteins, vitamins, lipids and salts/ions. Different 895 consequences of such may include formation of a surface-bound protein layer, particle dissolution or 896 aggregation, which are expected to have a crucial impact on cellular interaction (for reviews see 897 Monopoli et al, 2013; Urban et al., 2016; Moore et al., 2015). 898 The biomolecules bound to primary nanoparticles form a corona and the composition of those layers 899 depends on the surrounding media of the nanoparticles, which are commonly water, proteins, lipids, 900 ions, among others. The most important influence in corona formation is the physicochemical 901 characteristics of nanoparticles including size, shape, and surface charge. 902 A tightly bound immobile protein layer formed on the particle surface (the so-called hard corona) 903 and possibly a weakly associated mobile layer (the soft corona) has been described (Cedervall et al., 904 2007). In addition to protein binding, also lipids such as the surfactant lipids in the lung or lipids in 905 the blood should be considered, as well as the possible exchange of the soft corona within time. 906 The determination of the protein / lipid corona is complex and technically challenging and an 907 overview of some relevant methods will be given in the hazard identification chapter. The bio-corona 908 may be of particular importance for the interaction of ENMs with cells of the immune system 909 including phagocytic cells such as macrophages. 910 911

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

912 3. HUMAN EXPOSURE 913 914 3.1 Exposure to nanomaterials 915 With the increasing number of nanomaterial applications, the production of nanomaterials is also 916 increasing proportionally. Thus, there is an increased potential for human and environmental 917 exposure to nanomaterials, including nanomaterials released from products containing 918 nanomaterials and during the life cycle of nanomaterials, from production, use, and recycling to 919 disposal (EPA 2007). Exposure can be defined as ‘contact with a chemical, physical, or biological 920 agent by swallowing, breathing, or touching the skin or eyes” and exposure can be short-term (acute 921 exposure), intermediate duration, or long-term (chronic) (ISO TR 18637, 2015). The major exposure 922 sources for humans are occupational exposure, general population exposure, and consumer 923 exposure. 924 925 Figure 1. Life-cycle exposure to nanomaterials (EPA, 2007)

926 927 928 3.1.1 Occupational exposure

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

929 Workers involved in the life-cycle of nanomaterials, from the manufacturing, synthesis, formulation, 930 handling, and packaging disposal to the recycling of nanomaterials or products containing 931 nanomaterials, can be exposed to different amounts of nanomaterials depending on the work 932 duration, frequency, occupational setting, and mitigation strategies. The various potential sources of 933 occupational exposure are shown in the Table below and depend on the manufacturing method. 934 Workplace air is known to contain manufactured nanomaterials including nano-objects and their 935 agglomerates and aggregates (NOAA),. Along with the manufacturing process, handling, which 936 includes coating, sonicating, dispersion, bagging, and cleaning the vacuum, can also generate NOAAs 937 in workplace ambient air. 938 939 Potential sources of occupational exposure for various manufacturing methods 940 (Adapted from Aitken, 2004; 2006; EPA 2005) Manufacturing NOAA Potential exposure Route of exposure process source Gas phase In air Direct leakage from reactor Inhalation Product recovery from bag Inhalation/dermal filters in reactors. Processing and packaging Inhalation/dermal of dry powder. Equipment Dermal (and inhalation cleaning/maintenance during reactor (including reactor and evacuation) spent filters). Vapor deposition On substrate Product recovery from Inhalation reactor / dry contamination of work place

Processing and packaging Inhalation / Dermal of dry powder. Equipment Dermal (and Inhalation cleaning/maintenance during reactor (including reactor). evacuation) Colloidal Liquid suspension If liquid suspension is Inhalation / Dermal processed into powder, potential exposure during spray-drying to create powder, and processing and packaging of dry

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

powder Equipment Dermal cleaning/maintenance. Attrition Liquid suspension If liquid suspension is Inhalation / Dermal processed into powder, potential exposure during spray-drying to create powder, and processing and packaging of dry powder Equipment Dermal cleaning/maintenance

941 NOAA, nano-objects and their agglomerates and aggregates greater than 100 nm (ISO 12901-1) 942 943 3.1.2 General population exposure 944 Very little is known about general population exposure to nanomaterials, as to date there are very 945 few quantitative and specific trace analytical methods for detecting nanomaterials in the 946 environment. Notwithstanding, the general population can be exposed to nanoparticles in many 947 possible scenarios, such as the release of nanomaterials from waste control systems, leakage, 948 spillage, or other unknown or unexpected processes. 949 950 3.1.3 Consumer exposure 951 The potential use of ENMs in consumer products covers plastics, sporting equipment, electronic 952 equipment, and textiles, i.e. applications where nanomaterials are embedded into a matrix. For 953 textiles, ENM may be applied on the fibres or the surface. The main exposure from such consumer 954 products would be from abrasion, grinding, cutting, washing, and weathering of the products (Bello 955 et al., 2009; Göller et al., 2010; Froggett et al., 2014). Dermal exposure via cosmetic products or

956 sunscreens is also possible (e.g. TiO2, Zn ENMs as present in sunscreens) along with minor inhalation 957 or oral exposure. When compared with the data for occupational exposure, little is known about 958 nanoparticles released from consumer products, which may differ from the original manufactured 959 nanomaterials in terms of their physicochemical properties i.e. size, surface area characteristics 960 (chemical composition, corona), and shape. The physicochemical properties and hazard information 961 related to released ENMs compared to pristine nanomaterials have not yet been studied extensively. 962 Furthermore, exposure to ENMs fabricated for nanobio or bionano products, may represent an 963 important new safety concern, as exposure to these products and ENMs is usually via invasive

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

964 administration. However, in this case, extensive data reviews and approval by regulatory authorities 965 are normally required before use. 966 967 3.2. Route of exposure 968 Human exposure to nanomaterials can occur via various routes, although inhalation is the most 969 common route for workplace exposure, and consumer and general population exposure. Depending 970 on the use of nanomaterials, the other exposure routes include oral ingestion, dermal exposure, and 971 parenteral exposure by injection. 972 973 3.2.1. Inhalation exposure 974 Inhalation is the main route for airborne NOAA particles to enter the bodies of workers. Once 975 inhaled, nanomaterials are deposited in the various regions of the respiratory tract according to the 976 particle size. As shown in Figure 1, the ICRP predictive mathematical model describes the fractional 977 deposition of inhaled nanoparticles in the different regions of the human respiratory tract; head 978 region (nasopharyngeal), tracheobronchial region, and alveolar region. According to the ICRP model, 979 for 1 nm particles, 80 % are deposited in the head region, with 20 % in the tracheobronchial region 980 and less than 1 % in the alveolar region. However, for 20 nm particles, 50 % are deposited in the 981 alveolar region, and 25 % in the head and tracheobronchial regions. The main mechanism governing 982 nanoparticle deposition is diffusion, whereas sedimentation and impaction governs the deposition 983 mechanism for bulk particles (ref). Similarly MPPD (multiple path particle dosimetry model) is a 984 model for predicting particle dosimetry in the human lung following exposure to airborne particulate 985 matter for risk assessments (Asgharian et al., 2001). In this model, the fractional depositions of 986 particles in the human and rat lung were presented using three lung geometries: typical-path 987 symmetric, five-lobe symmetric, structurally different, and five-lobe asymmetric (stochastic) were 988 presented. Current version of MPPD model is applicable to predict deposition and clearance of 989 nanoparticle as well (Asgharian et al., 2009; Cassee et al., 2000)). Only few information are available 990 for the magnitude of exposure due to lack of personal sampling data. Most exposure studies have 991 been emission measurement from the nanomaterial manufacturing or handling or estimated 992 emission from nanocomposite producet simulation studies. Although harmonized approach to 993 esposure to nanomaterials has been suggested (Brouwer et al, 2013), standardized methods on 994 nanomaterial exposure assessment have not been finalized. 995

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

996 997 998 Figure 1. Predicted total and regional deposition of particles in human respiratory tract related to 999 particle size using ICRP 66 model. Deposition Fraction includes probability of particles being inhaled 1000 (inhalability). Subject is a nose breather, performing standard work. 1001 1002 3.2.2. Oral exposure 1003 A number of nanomaterials are already in use as food additives or in food-contact materials. These 1004 applications include the nano-encapsulation of vitamins and flavors to protect them from 1005 deterioration during storage or allow controlled release, and the creation of specific nano-sized

1006 micelles to deliver specific tastes. SAS (SiO2 in its aggregated form) is commonly used as anticlumping 1007 agent in food products manufactured as powders. Many food and beverage products also contain 1008 nano-scale particles. Moreover, some of the nanomaterials used in food packaging can migrate into 1009 the food and be ingested, while some of the nanomaterials used in biocides can be ingested as food 1010 contaminants. Once ingested, these nanoscale particles can be absorbed by the human 1011 gastrointestinal (GI) system and undergo various physical changes. Nanoscale particles may even 1012 enter the systemic circulation intact. Some nanoparticles, such as silver nanoparticles, known and 1013 used as a biocide, can affect the GI microbial milieu (Jeong et al., 2010). In contrast to inhalation 1014 exposure, there are no reliable standard methods for measuring nanomaterial characteristics and 1015 levels in complex matrices, such as food. Furthermore, the ADME (absorption, distribution,

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1016 metabolism, and excretion) of ingested nanomaterials is not well understood. Some attempts to 1017 characterize and esimate nanomaterial in the food relevant product internationally have been 1018 initiated by the ILSI Nanorelease project by organizing round robin tests. 1019 1020 3.2.3. Dermal exposure 1021 In an occupational setting, in addition to inhalation exposure, dermal exposure to nanomaterials is 1022 also a serious concern, as certain skin areas can be exposed to nanomaterials. However, unlike 1023 inhalation exposure, such dermal exposure may not result in the dermal penetration of 1024 nanomaterials into the epidermal layer and systemic uptake. The keratin layer of the skin acts as the 1025 first barrier to dermal penetration. In general, the uptake via the skin is considered negligeably. 1026 However, a number of other factors also influence the dermal absorption of nanoparticles, including 1027 the location and skin conditions at the application site, physicochemical properties of the penetrating 1028 molecules, and physicochemical properties of the vehicle dispersing the penetrating molecules 1029 (Gautam et al., 2011). The lipophilic-hydrophilic gradient, pH gradient, and isoelectric point also 1030 influence the dermal absorption of nanoparticles (Baroli et al., 2007). Furthermore, the inclusion of 1031 solvents, surfactants, enhancers, and others molecules in a nanosuspension can alter or damage the 1032 stratum corneum to induce a potential increase in the absorption of all or selected ingredients of the 1033 applied nanomaterial-containing formulation (Bangale et al. 2005). 1034 1035 3.2.4 Other routes 1036 Although it is not normal way of exposure to nanomaterials, intratracheal instillation, 1037 intrapertoneal injection, and intra venous injections were frequently performed to experimental 1038 animals to identy hazards of nanomateirals to respiratory tract, mesothlial layers and systemic 1039 circulation, respectively (Asgharian et al., 2001; 2009) 1040 1041 3.3. Epidemiology and health surveillance 1042 Under the NanOSH project, Gulumian (2015) recently conducted a systemic review of health 1043 surveillance programmes used to protect workers from the potential risks of exposure to 1044 manufactured nanomaterials. A total of 30 citations related to health surveillance and epidemiology 1045 studies were identified and an additional 33 citations were identified from the references included in 1046 the initial 30 citations. After excluding 5 duplicate references, a total of 58 references were evaluated 1047 for inclusion in the systemic review, and 5 studies were finally selected, as shown in Table 1. As the 1048 paper by Gause et al. (2011) is more conceptual and introductory with no relevant health 1049 information, it is not discussed further. 1050

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1051 Table 1. Description of epidemiology and health surveillance studies included in systemic review 1052 (kindly provided by Gulumian) Study (Gause et al., (Lee et al., 2012) (Liou et al., 2012) (Liao et al., 2014) (Lee et al., 2014) 2011) Design Descriptive Descriptive Cross-sectional Cross-sectional Cross-sectional Programme Case Programme Case exp case control exp case control exp case control Study Study study study study Country USA South Korea Taiwan Taiwan South Korea Participant s Exposure Carbon-based Silver Metal, metal- Metal, metal- Multi-walled nanomaterials nanomaterials oxide and oxide and carbon nanotubes carbon-based carbon-based NMN NMN Exposure NS 7 years NS NS 3.9 ± 3.9 years duration Occupation Commercial Commercial, Commercial , 14 Commercial , 14 Commercial, or Industry research synthesis plants handling plants handling synthesis laboratory nanomaterials nanomaterials Participant N=200, of which N = 2, both N =227 exp N =258 exp N = 9 exp N ± 20% exposed exposed N = 137 unexp N = 200 unexp N = 4 unexp Participant NS 37 and 42 years Exp < 40: 173 Exp < 40: 194 Exp: 33.8 ± 4.9 y ages (years) Exp ≥ 40: 54 Exp ≥ 40: 64 Unexp: 28.3 ± Unexp < 40: 89 Unexp < 40: 135 4.4 y Unexp ≥ 40: 48 Unexp ≥ 40: 65

(p = 0.03) (p = 0.12) Participant NS Male exp: 100% Male exp: 78% Male exp: 76% Male exp: 10% gender (% Male unexp: Male unexp: Male unexp: male) 54% 60% 75%

Medical Surveillanc e Biomarker / Clinical Biomarkers / Biomarkers / Clinical Biomarkers/Clini Clinical Clinical Clinical cal Target General Health “Routine blood” Lung Cardiovascular; Pulmonary stress inflammation respiratory; skin; biomarkers Oxidative neurological “Routine” blood,

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

damage Antioxidants Cardiovascular Neurologic Geno-toxicity Lung function Indicators “periodic health Silver (blood, Clara cell Self- Malondialdehyde

status”; lung urine) protein, heat administered , H2O2, 4-HHE, function; chest shock protein 70, symptom n-hexanal X-Ray nuclear factor kB questionnaire; Cobalt, transcription work-relatedness Molybdenum factor activation, of symptoms (blood) nitric oxide Pulmonary 8-ho-guanosine, function m-guanosine, parameters isoprostane Superoxide dismutase; glutathione peroxidase; myeloperoxidase Fibrinogen; intracellular adhesion molecule-1; IL-6 Reaction time, memory test Programme Outcomes Biomarkers N/A Blood/urine No diff in lung N/A Pulmonary stress silver not inflammation, increased, all elevated; oxidative three markers damage or No diff in genotoxicity Cobalt, Antioxidants Molybdenum decreased No changes in Cardiovascular pulmonary markers elevated function Early signs N/A 1 No diff in lung 1 out of 3 neuro- N/A Routine blood function; tests impaired not deviant

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2 Routine blood No diff in lung not deviant function Clinical N/A N/A N/A Increased N/A signs prevalence of allergic dermatitis and ‘sneezing’ Interpretati Nano-specific Surveillance Although some Allergic Oxidative stress on of surveillance not measures did not biomarkers were dermatitis has markers outcomes deemed detect any elevated, these multiple causes, increased, but necessary abnormalities, markers are not not specific for small sample and specific for nanomaterial non-specific nanomaterial exposure. exposure “Sneezing” non- specific symptom. Costs Monetary / NS NS NS NS NS Resource use Coverage NS 2 out of 5 97% of workers NS NS workers participated participated 1053 1054 Lee et al (2012) conducted a health surveillance study of a silver nanoparticle manufacturing 1055 workplace, including the assessment of personal exposure levels to silver nanoparticles, a walk- 1056 through evaluation of the manufacturing process, and the collection of blood and urine samples from 1057 the exposed workers. Two male workers, each with 7 years in silver nanoparticle manufacturing, 1058 exhibited exposures at the work environment by silver concentrations of 0.35 and 1.35 μg/m3, 1059 respectively, where the blood silver levels were 0.034 and 0.135 μg/dl, respectively, and the urine 1060 silver levels were 0.043 μg/dl and not detected, respectively. When reviewed by an occupational 1061 physician, their health status, including blood chemistry and hematology, was determined to be 1062 within a normal range. Plus, the study reported that the silver nanomaterial manufacturing workers 1063 were exposed to much lower concentrations of silver dust and soluble silver TLVs (threshold limit 1064 values), and no significant findings were recorded as regards their health status. One additional case 1065 report of work-related argyria has been documented in South Korea. A 27-year-old Korean male, 1066 employed as a technician plating mobile telephone subunits with aerosolized silver for 4 years, 1067 presented with asymptomatic blue-to-gray facial discoloration that had progressed over 4 months,

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1068 consistent with general argyria. The patient also showed silver deposits in his sclera, conjunctiva, and 1069 oral mucosa. No other adverse physical or organ effects were observed. A laboratory investigation 1070 confirmed that his complete blood count, chemistry panel, liver function test, and routine urinary 1071 analysis were all within a normal range. However, his serum silver level was elevated at 15.44 ㎍/dL 1072 (normal range, 1.1–2.5 ㎍/dL) and urinary silver was 243.2 ㎍/L (normal range, 0.4–1.4 ㎍/L). The 1073 lead, mercury, and nickel levels were all within a normal range. A histopathological examination of a 1074 3-mm punch biopsy specimen from the patient’s face revealed silver granules in the epidermal basal 1075 layer. In addition, fine, minute, round, and brown-to-black silver granules were deposited in the 1076 basement membrane zone surrounding the eccrine sweat glands. The physician reported a case of 1077 generalized argyria resulting from the use of aerosolized silver in the mobile telephone subunit 1078 industry, which is apparently the first case of its kind based on an extensive literature search of 1079 English publications (Cho et al., 2008). While unaware of the exact use of nanosilver in the 1080 workplace, the physician concluded that the patient’s 4-year work exposure to silver particles, 1081 whether nanoscale or non-nanoscale, seemed to induce the general argyria. 1082 1083 Lee et al (2014) studied a MWCNT manufacturing workplace with 4 office workers and 9 1084 manufacturing workers and used several biomarker tools, including the EBC (exhaled breath

1085 condensate) malondialdehyde, H2O2, 4-HHE, and n-hexanal, lung function parameters, blood metal 1086 (cobalt and molybdenum) concentration as a surrogate for MWCNT exposure, and routine 1087 hematology and biochemistry, along with the workplace air concentration of MWCNTs in terms of 1088 the TSP (total suspended particulate) concentration and EC (elemental carbon concentration). The 1089 workers exhibited a normal range of hematology and blood biochemistry values and normal lung 1090 function parameters. When analyzing the EBCs, the malondialdehyde (MDA), 4-hydroxy-2-hexenal (4- 1091 HHE), and n-hexanal levels in the MWCNT manufacturing workers were significantly higher than 1092 those in the office workers. The MDA and n-hexanal levels were also significantly correlated with the 1093 blood molybdenum concentration, suggesting MDA, n-hexanal, and molybdenum as useful 1094 biomarkers of MWCNT exposure (Lee et al., 2014). 1095 1096 Several epidemiological studies were recently conducted on Taiwanese workers (227 exposed and 1097 137 non-exposed workers) handling 14 different engineered nanomaterials (Liou et al., 2012; Liao et 1098 al., 2013, Liao et al 2014). The 227 exposed workers showed a depression of antioxidant enzymes, 1099 such as superoxide dismutase (SOD) and glutathione peroxidase (GPX), and increased expression of 1100 cardiovascular markers, including fibrinogen, ICAM, and Interluekin-6 (Liou et al., 2012). In a 6-month 1101 follow-up study, 38 carbon nanotube handling and manufacturing workers showed a depression of

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1102 GPX, significant changes in a comet assay (L/H ratio, tail to head ratio), and significant reduction of 1103 the lung function parameters, such as the peak expiratory flow rate and a forced expiratory flow of 1104 25% (Liao et al., 2013). However, since these studies lack of proper measurement of the ambient 1105 nanomaterial concentrations the workers were exposed to, this limits any correlation between the 1106 biomarkers and the nanomaterial exposure. Similar variables were also analyzed in 2 studies that 1107 collected samples from 14 different nanomaterial handling plants (Liao et al., 2014; Liou et al., 2012). 1108 Due to the different settings from which the workers were recruited for these studies, the exposure 1109 to NMs was heterogeneous and included exposure to metals, metal oxides, and carbon-based 1110 nanomaterials (Liao et al., 2014; Liou et al., 2012). Both studies stratified the exposed participants 1111 into two risk groups. The risk stratification was conducted following the control banding tool 1112 developed by Paik et al (2008) based on the severity score of the nanomaterial toxicity and the score 1113 of the exposure probability (Paik, Zalk, & Swuste, 2008). Risk group 1 was considered the lower-risk 1114 group and composed of 139 participants (35 female, 104 male) in Liao’s study and 128 participants 1115 (31 female, 97 male) in Liou’s study (2012). Risk group 2, the higher-risk group, had 119 participants 1116 (26 female, 93 male) in Liao’s study and 99 participants (19 female, 80 male) in Liou’s study (2012). 1117 Liao’s study also used a self-reported symptom questionnaire to assess the prevalence of 1118 cardiovascular, respiratory, skin, and neurological symptoms and diseases between the exposed and 1119 unexposed workers. The work-relatedness of the symptoms was also assessed through the 1120 questionnaire by asking whether or not the symptom was present before being exposed to NMs, plus 1121 the reported symptoms and diseases were further checked by an occupational physician (Liao et al., 1122 2014). The only symptom identified as significantly work-related was sneezing (5.88% in risk level 2 1123 and 7.91% in risk level 1 vs. 2.00% in controls, p=0.04). The prevalence of a work-related dry cough 1124 (p=0.06) and productive cough (p=0.09) among the nanomaterial handling workers was also higher 1125 than that among the controls. The only disease significantly worsened by work was allergic dermatitis 1126 (4.20% in risk level 2, 0% in risk level 1 vs. 0.50% in controls, p=0.01). The incidence of angina was 1127 also higher among the nanoworkers than among the controls (p=0.06). 1128 Recently a cross-sectional health surveillance study on large-scale manufacturing of MWCNT along 1129 with relatively high occupational exposure levels has been reported (Fatkhutdinova et al., 2016). All 1130 air samples were collected at the workplaces from both specific areas and personal breathing zones 1131 using filter-based devices to quantitate elemental carbon and perform particle analysis by TEM. 1132 Biological fluids of nasal lavage, induced sputum and blood serum were obtained from MWCNT- 1133 exposed and non-exposed workers for assessment of inflammatory and fibrotic markers. They found 1134 that exposure to MWCNTs caused significant increase in IL-1β, IL6, TNF-α, inflammatory cytokines 1135 and KL-6, a serological biomarker for interstitial lung disease in collected sputum samples. 1136 Additionally, the level of TGF-β1 was increased in serum obtained from young exposed workers,

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1137 indicating accumulation of inflammatory and fibrotic biomarkers in biofluids of workers 1138 manufacturing MWCNTs (Fatkhutdinova et al., 2016). 1139 1140 Trop et al. (2006) reported on a burn victim who had an argyria-like condition and elevated 1141 activities of liver-specific plasma enzymes when an Acticoat™ dressing (containing ionic AgNPs) 1142 was applied to his wound. No mention was made of silver being sequestered in the liver, 1143 although this was possible since levels of the metal were elevated in plasma and urine. As 1144 reported by the authors, local treatment with Acticoat™ dressings for 7 days caused the plasma 1145 activities of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) to rise 1146 incrementally to 233 and 78 units per liter (U/L), respectively (upper limits of normal: 33 U/L for 1147 ALT and 37 U/L for AST). The concentration of silver in blood plasma was 107 micrograms per 1148 kilogram (μg/kg); this value subsequently dropped toward normal levels (13.3 μg/kg after 97 1149 days) when Acticoat™ dressings were changed on day 8 to dressings containing betadine 1150 ointment. C-reactive protein level was also elevated, in parallel with plasma silver levels, 1151 reaching a maximum concentration of 128 milligrams per liter (mg/L) after 4 days. The level of 1152 this liver-synthesized marker for acute inflammation was back to normal (5 mg/L) after 8 silver 1153 treatment–free days. In a prospective study of 30 patients with graft-requiring burns, Vlachou et 1154 al. [2007] found increased concentrations of silver in serum (median, 56.8 micrograms per liter 1155 [μg/L]; range, 4.8–230 μg/L) when Acticoat™ dressings were applied to the wounds. They found 1156 no changes in hematologic or clinical chemistry parameters indicative of toxicity associated with 1157 the silver absorption, and at the 6-month follow-up, the median serum level had declined to 0.8 1158 μg/L (Vlachou et al. 2007). 1159

1160 3.4. Toxicokinetics 1161 1162 3.4.1 Introduction 1163 In the safety assessment toxicokinetics may be seen as bridging the gap between exposure and 1164 toxicology. Toxicokinetic testing gives information on the fate and behaviour of the substances in 1165 evaluation and provides insight into potential target organs and organ burden that may ultimately 1166 result in toxicity. Like for other substances it describes the study of absorption, distribution, 1167 metabolism and excretion (ADME) as potentially sequential processes. However, especially for 1168 certain metal and metaloxide nanomaterials it may be debatable whether metabolism does occur. It 1169 is very important to have a good insight in the behavior of nanomaterials in the body. The major risk

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1170 associated with nanomaterials is considered to be related to the presence or release of free nano- 1171 objects, ions, or components that comprise the individual nanomaterials e.g. when nanoparticle 1172 aggregates/agglomerates may disintegrate and release the comprising primary nanoparticles. The 1173 biodistribution over the various organs will determine the possibility for inducing harmful effects. 1174 1175 The OECD 417 test guideline describes the toxicokinetic studies performed for chemical substances 1176 and explicitly states that it is not intended for the toxicokinetic testing of nanomaterials. Analogously, 1177 both the in vivo and in vitro OECD Guidelines 427 and 428 for dermal penetration, respectively, were 1178 developed for chemicals and not proven valid for nanoparticles. Therefore, the use of such 1179 methodologies should be evaluated on a case-by-case basis. 1180 1181 The biodistribution of nanoparticles is influenced by a number of factors including size, surface 1182 charge, and surface composition like protein binding and coating (De Jong et al. 2008, Lankveld et al. 1183 2011, Landsiedel et al. 2012). When nanomaterials are used in test systems, one has to be aware that 1184 some of the properties that need to be determined can be affected and are largely dependent on the 1185 surrounding environment (e.g. tissue culture media, blood/serum, protein presence). Such 1186 interactions with the environment may result in a temporal evolution of the nanomaterials 1187 themselves e.g. by obtaining/shedding a protein coating, the formation of nano-object 1188 agglomerates/aggregates and other changes in the nanomaterials. Such changes may affect the 1189 nanomaterial characteristics, which can impact the toxicological profile of a nanomaterial. Factors 1190 such as route of administration, size of the nano-object or its aggregates/agglomerates, surface 1191 properties (chemistry and charge), animal species, dose and dosing methods have all been reported 1192 to influence the toxicokinetics in animal models (Niidome et al., 2006, De Jong et al., 2008, Lipka et 1193 al., 2010, Hirn et al., 2011, Lankveld et al., 2011). 1194 1195 3.4.2 Methods to evaluate toxicokinetics of nanomaterials 1196 The methodology to perform toxicokinetic studies for chemicals is described in OECD TG 417 which 1197 explicitly mentions that nanomaterials are not covered by the technical guideline. However, a 1198 number of basic requirements as indicated in OECD TG 417 can be applied also to 1199 nanomaterials/nanoparticles e.g. the study design, number and sex of animals to be used. In view of 1200 the possible long retention time as mentioned above, the time schedule indicated in OECD 417 of a 1201 evaluation typically at day 7 after administration is not applicable for nanomaterials. Although 1202 prolonged evaluation up to 14 days is also mentioned. A more prolonged time schedule for example 1203 of 90 days as used by Geraets et al., (2014) might be more appropriate. For certain nanomaterials

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1204 that shed ions or slowly dissolve/degrade (e.g. nanosilver, nanoZnO), also shorter time periods may 1205 need to be considered. 1206 1207 As for all (toxico)kinetic studies also for studies with nanomaterials a critical point is the availability of 1208 a measurement system for detection of the nanomaterials. However, detection of nanoparticles in 1209 tissues/organs is complex. Electron microscopy is neither applicable for quantitative measurements 1210 nor for all nanomaterials. To date, most studies on toxicokinetics of nanomaterials have used 1211 elemental analysis of the components of the nanomaterials e.g. Zn for ZnO, Ti for TiO2, Ag for Ag 1212 nanoparticles (Krystek et al., 2013). For metal and metal oxides analysis could be performed by using 1213 inductively coupled plasma mass spectroscopy (ICP-MS) or atomic absorption mass spectroscopy 1214 (AA-MS). Although this provides a good indication of the possible tissue distribution, the limitation is 1215 that the nanoparticles themselves are not detected or measured. With ICP-MS generally the tissue 1216 including the nanomaterials are completely dissolved under acid conditions, so only the elements 1217 composing the nanomaterial are detected. Whether these are present as nanoparticles or as 1218 previously dissolved ions is unknown. In combination with separation techniques like field flow 1219 fractionation (FFF) it is possible to evaluate the presence of particles using so called single particle 1220 ICP-MS (Van Der Zande et al. 2012). 1221 1222 A specific labelling of nanomaterials to follow their fate in vivo can be done by using radioactive 1223 isotopes as radiolabels or fluorescent dyes. This can be done by adding a label to the nanomaterial by 1224 specific binding, or by using a radio-isotope of the nanomaterial itself. A disadvantage of the specific 1225 labelling is that the label can detach from the nanomaterial (Geiser and Kreyling 2010). In that case 1226 the label is measured while it is detached from the nanomaterials so incorrect/false information is 1227 obtained on the distribution of the nanoparticles. So, when using any specific labeling technique a 1228 careful evaluation of the integrity of the label-nanoparticle combination is necessary. Alternatively 1229 radioactive isotopes may be used that are isotopes of a metal being part of the nanomaterial (e.g. 1230 gold or silver). With this approach, there is some certainty that the nanoparticles themselves are 1231 detected, although for silver nanoparticles there is still uncertainty regarding the release of silver 1232 ions. In addition, natural stable isotopes like may be used to demonstrate uptake from the 1233 application site (Gulson et al. 2010). By using ZnO nanoparticles enriched for 68Zn it was 1234 demonstrated that Zn present as ZnO nanoparticles in sunscreen formulations was taken up from the 1235 skin application site (Gulson et al., 2010). The 68Zn could be detected in blood and urine of the human 1236 volunteers. As the 68Zn was evaluated using ICP-MS measurements it is not known whether 68Zn has 1237 been absorbed as ZnO particles or soluble Zn or both. Also in internal organs of mice including the 1238 liver, the presence of 68Zn was observed but here also penetration of the ZnO nanoparticles

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1239 themselves could not be established (Osmond-McLeod et al. 2014). The Zn homeostasis was largely 1240 maintained and the rpresence of ZnO particles in the sunscreen did not elicit and adverse biological 1241 effect (Osmond-McLeod et al. 2014). 1242 So, depending on the methodology used for nanoparticle detection there may be uncertainty 1243 whether the nanomaterial or the released ions are detected, especially for metallic nanomaterials 1244 that can release ions (e.g. silver or zinc oxide). 1245 1246 Another method specifically suitable for metal nanoparticles maybe neutron-activation of the metal 1247 nanoparticles. Using this method with 198Au very low levels of translocation from the lung was 1248 demonstrated for 1.4 nm and 18 nm gold nanoparticles and for 192Ir aerosol of ultra fine particles 1249 (UFP) varying in size between 15 and 20 nm (Semmler et al., 2004, Semmler-Behnke et al., 2008). 1250 This neutron activation can also be done after the administration of the metal nanoparticles for 1251 detection of the presence in organs. 1252 1253 As in general uptake via the various routes of administration may be rather low (see below) the 1254 determination of such low uptake/translocation remains a challenge when studying the 1255 toxicokinetics of nanomaterials. For example the limit of detection for the used ICP-MS method to 1256 evaluate tissue distribution of metal and metaloxide nanoparticles is about 50 ng/g tissue (Krystek et 1257 al., 2014). So, instead of using one of the major exposure routes (skin, gastrointestinal tract or lung) 1258 also the intravenous route of administration is commonly used for toxicokinetic studies in order to 1259 identify potential target organs. However, one should be aware that the route of administration also 1260 may effect the systemic distribution as has been demonstrated for the intravenous versus the 1261 inhalation route (Kreyling et al., 2014, Hirn et al., 2011, Semmler-Behnke et al., 2008). This can be 1262 explained by interactions of the nanoparticles with lung fluids and/or intestinal tract fluids before 1263 reaching the absorbing epithelium. The nanoparticles can become coated with a range of different 1264 proteins when they become in contact with biological fluids; this commonly referred to as the 1265 “protein corona” (Cedervall et al., 2007, Lesniak et al., 2012). 1266 1267 3.4.3 Absorption, distribution, metabolism and excretion 1268 1269 3.4.3.1 Absorption 1270 Skin 1271 Dermal penetration can be evaluated by using various in vitro systems for which the skin of many 1272 mammalian species, including humans. OECD 428 describes the use of in vitro models for skin 1273 absorption, whereas in vivo skin absorption studies are described in OECD 427. However, none of

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1274 these tests were designed for nanoparticles. Nanoparticle quantitation remains a problem in these 1275 studies especially in view of the fact that dermal penetration of nanoparticles is generally considered 1276 to be low or absent (Butz et al. 2007, Monteiro-Riviere and Riviere 2009b, Sadrieh et al. 2010, 1277 Monteiro-Riviere and Larese Filon 2012). In general, nanoparticle penetration of the skin is limited to 1278 the first cell layers of the stratum corneum (Butz et al. 2007). However, for some nanomaterials, 1279 there seemed to be limited uptake. For example, when ZnO nanomaterial was applied on the skin in 1280 a sunscreen formulation, the presence of Zn in the blood originating from the ZnO in the sunscreen 1281 was observed (Gulson et al. 2010). 1282 Silver (Ag) nanoparticles are widely used as antimicrobial agents, for example in wound dressings (Rai 1283 et al. 2009, 2014). In an in vitro system using human skin exposed to Ag nanoparticles, a low 1284 translocation into the receptor fluid was found which was increased fivefold in damaged skin (Larese 1285 Filon et al. 2009). The presence of elemental Ag was determined with electrothermal atomic 1286 absorption spectroscopy (ETAAS) which cannot discriminate between silver ions and silver particles, 1287 so translocation of nanoparticles was not demonstrated. Some absorption of Ag does occur as after 1288 treatment of burn patients with wound dressings containing nanocrystalline silver increased blood 1289 silver serum levels were observed similar to the situation for Zn absorbance from sunscreens 1290 formulations containing ZnO nanoparticles (Vlachou et al. 2007, Gulson et al., 2010). The silver levels 1291 in blood were considered to be non-toxic to the patients (Vlachou et al., 2007). 1292 For skin penetration and absorption, the quality of the skin in terms of skin damage, like abrasions 1293 and UVB damage (sunburns), mechanical stressors (skin flexing), and the effects of solvents and 1294 vehicles used may affect the skin penetration (Monteiro-Riviere and Larese Filon 2012). 1295 In a recent review a more balanced outcome was described showing that for certain (rather) small 1296 nanoparticles skin penetration might be possible especially for metal and metal oxide nanoparticles 1297 (Filon et al., 2015). Both types of nanomaterials may consist of particles with a secondary size which 1298 occur in physiological media after interaction with media components. A size dependent skin 1299 penetration can be distinguished: NPs ≤ 4 nm can penetrate and permeate intact skin, NPs size 1300 between 4 and 20 nm can potentially permeate intact and damaged skin, NPs size between 21 and 1301 45 nm can penetrate and permeate only damaged skin, NPs size > 45 nm cannot penetrate nor 1302 permeate the skin. In addition, for metals other aspects may play an important role like dissolution 1303 which can cause local and systemic effects, the sensitizing or toxic potential and the tendency to 1304 create aggregates. A decision tree is proposed to evaluate the potential risk for consumers and 1305 workers exposed to NPs (Filon et al., 2015). 1306 1307 Gastrointestinal tract

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1308 Another route for internal; exposure is via the gastrointestinal (GI) tract. However, there is still a 1309 lack of knowledge to what extent single particles, small aggregates/agglomerates and larger 1310 aggregates/agglomerates can be translocated across the GI tract epithelium. Uptake from the GI- 1311 tract was demonstrated for several nanomaterials (Jani et al. 1990, 1994, Wang et al. 2007, Kim et al. 1312 2008, Park et al. 2010 a, b), but a lack of uptake of nanoparticles was also observed (Yang et al. 1313 2012). In general, smaller particles were found to have a higher uptake (Jani et al. 1990, Park et al. 1314 2010a). However, large titanium particles with a size of 500 nm were also absorbed via the GI-tract 1315 (Jani et al. 1994). In a human volunteer study nansilver (size around 60 nm) was orally ingested up to 1316 14 days (Munger et al., 2014). Serum silver concentration was detected in 42% of subjects receiving 1317 100 μg/day and in 92% of subjects receiving 480 μg/day but was undetectable in the urine (Munger 1318 et al., 2014). However, in the silver preparation the majority of the silver was present as silver ions, 1319 so uptake of nanoparticles could not be established. 1320 As for skin penetration studies also for the GI tract in vitro models are available to study migration 1321 over intestinal cells for which transwell systems using Caco-2 cells are commonly used (Martinez- 1322 Argudo et al., 2007, Kulkani and Feng 2013, Braakhuis et al., 2015, Lefebvre et al., 2015). A 1323 translocation for polystyrene nanopartciles was noted varying between 1.6% and 12.3% of the added 1324 dose (Martinez-Argudo et al., 2007, Walczak et al., 2015). The in vitro cellular models can indicate 1325 low, medium or high translocation, although especially low translocation can be difficult to 1326 determine in view of limitations of the sensitivity of the measurement technique. Although Caco-2 1327 cells are commonly used as model for GI tract epithelium also models with cells present in the 1328 epithelium of Peyer’s Patches (so called M cells) were used (Lefebvre et al., 2015). These M cells 1329 were found to be more efficient for particles than normal enterocytes (Des Rieux et al., 2006). So, M 1330 cells may be an important addition to a model for gastrointestinal absorption studies (Shahbazi and 1331 Santos 2013). Especially for the GI tract the potential interaction that may occur with all components 1332 and gastrointestinal fluids may warrant a more dedicated evaluation of these interactions (Lefebvre 1333 et al., 2015). For these studies various in vitro non cellular models are available as well (Oomen et al., 1334 2003, Brandon et al., 2006, MacAllister 2010, Cockburn et al., 2012, Guhmann et al., 2013). Some of 1335 these models are static without passage but also dynamic models are available each with their own 1336 limitations (Lefebvre et al., 2015). Similar to skin penetration the solubility of the nanomaterials 1337 might be an issue also in the GI tract to facilitate absorption of shed ions and/or degradation 1338 products of nanomaterials. 1339 1340 Respiratory tract 1341 In the lung the uptake of nanoparticles occurs primarily in the deep lung and depends on the 1342 deposition of the particles in the alveoli the respiratory part of the lung. This deposition is depending

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1343 on the size of the particles and can be calculated (modeled) by the ICRP and/or the MPPD model 1344 (ICRP 1994, Cassee et al., 2002). The uptake form the lung is limited (Kreyling et al., 2014). A small 1345 but significant fraction of the dose of nanoparticles may show systemic distribution, although the 1346 majority of the nanoparticles remained in the lung (Kreyling et al. 2002, Semmler-Behnke et al. 2008, 1347 Sung et al.2011, Abid et al. 2013). The elimination half-time from the lung for both fine and ultrafine 1348 (nano)particles in rats was approximately 65 days (Pauluhn 2009, 2011). Part of the inhaled particles 1349 (nano and micro) will end in the GI tract due to the mucociliary cascade that removes inhaled 1350 particles from the lung, and will be excreted in the faeces (Abid et al. 2013). The primary particle size 1351 of the nanoparticles was important as smaller (7 nm versus 20 nm) nanoparticles had a higher uptake 1352 from the lung (Balasubramanian et al. 2013). In this study, macrophage-mediated mucociliary 1353 escalation, followed by faecal excretion, was the major pathway of clearing the inhaled nanoparticles 1354 from the lungs. For lung exposure also the potential of the nanoparticles to migrate along the 1355 olfactory nerve into the brain (olfactory bulb) should be considered (Oberdörster et al. 2004, 1356 Balasubramanian et al. 2013). 1357 1358 3.4.3.2 Distribution, Metabolism and Excretion 1359 Following translocation across the portal of entry, distribution of certain ENPs to different organs and 1360 tissues may occur. Generally metabolism of ENPs is considered to be minimal or absent although 1361 especially metal and metal oxide ENPS may degrade and disappear slowly by shedding ions. After 1362 oral administration/exposure excretion of non absorbed ENPs occurs via the faeces from the GI tract. 1363 When systemically available generally a rapid sequestration of ENPs by cells of the mononuclear 1364 phagocytic system (MPS) occurs as indicated by high levels of IV administered nanoparticles into the 1365 liver and the spleen (De Jong et al., 2008, Lankveld et al., 2010, Geraets et al., 2014). The 1366 nanoparticles are actively and quasi-irreversibly removed from the blood by the phagocytizing cells 1367 of the MPS , and thus is not concentration dependent. So, the concentration in blood or plasma 1368 generally has little value for internal dose estimation, while the concentration in organs does not 1369 depend on the blood concentration. Even with low blood concentrations the uptake in organs of the 1370 MPS notably liver and spleen can occur. 1371 The ability to measure the internal dose of an ENP in tissues is essential in identifying potential target 1372 organs for toxicity testing and in the construction of an appropriate dose–response relationship. This 1373 is especially true for ENPs that can accumulate in tissues over time. For TiO2 at 90 days after 1374 administration Ti could still be detected in considerable amounts mainly in liver and spleen (Geraets 1375 et al., 2014). This presents a new toxicological paradigm, particularly for organs that are normally 1376 protected against the entry of larger particulate materials. This also makes it essential that dose–

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1377 response relationships for ENPs are determined for multiple target organs, including some that may 1378 not be a first tier consideration in the risk assessment of conventional chemicals. 1379 1380 Oral and intravenous exposure to gold nanoparticles of different sizes resulted in an increased organ 1381 distribution in mice and rats with decreasing particle size (Hillyer and Albrecht 2001, De Jong et al.,

1382 2008). For TiO2 after oral administration organ distribution was reported for 500 nm particles (Jani et

1383 al., 1994). In contrast also minimal to non existing translocation was observed for TiO2 nanoparticles 1384 Geraets et al., 2014) The size of the primary particles of the latter study was 20-25 nm while the 1385 majority of particles occurred as agglomerates/aggregates measuring between 80–150 nm. In 1386 another independent study a low but significant amount of Ti could be detected in liver and/or 1387 spleen at 24 hours after the last of five consecutively administered oral doses ( Tassinari et al., 2014). 1388 For nanosilver a dose dependent increase of Ag content was present in brain, kidneys, liver, lungs, 1389 stomach and testes after oral administration (Van Der Zande et al., 2012, Kim et al., 2008). Although 1390 the translocation of Ag from the GI tract could be mainly attributed to migration of Ag+ ions (Van Der 1391 Zande et al., 2012). The oral administration of Ag-NP resulted in a widespread presence of Ag in 1392 various organs which was mainly cleared from the organs at 8 weeks with the exception of brain and 1393 testis. The organ Ag content was highly correlated with the presence of Ag+ ions in the Ag-NP 1394 suspension indicating that Ag+ ions and to a lesser extent Ag-NP passed the intestines. Remarkably

1395 Ag-NPs could be demonstrated in animals treated with AgNO3 solutions, indicating the new 1396 formation of nanoparticles from Ag+ in vivo. It was concluded that exposure to silver nanoparticles 1397 appears to be very similar to exposure to silver salts (Van Der Zande et al., 2012). Also for two 1398 commercially available synthetic amorphous silica nanomaterials uptake from the GI tract was 1399 demonstrated with increase in Si tissue levels in brain, kidney, liver and testis (Van Der Zande et al., 1400 2014). 1401 1402 In a recent review Hougaard et al., 2015 evaluated the possibilities for migration of nanoparticles 1403 into the placenta and fetus. In addition to animal models (using intravenous administration) also in 1404 vitro and ex vivo placental models are available both from animal and human origin. The migration 1405 of nanoparticles into the placenta and pups was considered possible. For various types of 1406 nanoparticles the migration into the placenta and pups was demonstrated be it to a low extent 1407 (Hougaard et al., 2015). 1408 1409 3.4.4 Specific considerations for toxicokinetic studies on nanoparticles 1410 A major difference between the toxicokinetics of neat (soluble) chemical and nanomaterials is that 1411 for neat dissolved chemicals the tissue distribution is concentration dependent, and an equilibrium is

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1412 generally obtained between blood and organ concentration, whereas nanoparticles are rapidly 1413 removed from the circulation by cells of the mononuclear phagocytic system (MPS) as indicated by 1414 the distribution of a major fraction of an injected dose into spleen and liver (De Jong et al., 2008, 1415 Lankveld et al., 2010, Geraets et al., 2014). A toxicokinetic PBPK model indicated distribution to 1416 several compartments in which the nanomaterials are quasi-irreversibly retained in the organs using 1417 2-5 compartments for nanoparticle distribution (Lankveld et al., 2010). There is no equilibrium 1418 concentration between tissue and blood. Uptake in organs can occur independent of the blood 1419 concentration, i.e. even with a low blood concentration and high organ concentration, organ uptake 1420 can occur. These may result in the persistence of nanomaterials in organs for long periods: silver 1421 could still be detected in various organs at day 17 after intravenous administration of silver 1422 nanoparticles in rats (Fabian et al. 2008; Pauluhn 2009; Lankveld et al. 2010). Titanium nanoparticles 1423 were still detectable up to 90 days after a single and repeated intravenous administration (Geraets et 1424 al. 2014). To identify tissue distribution and the potential for tissue accumulation and persistence of 1425 a nanomaterial, it is necessary to design single and repeated kinetic studies with a representative 1426 follow-up period of time. In OECD 417 on toxicokinetic testing, the follow up period is typically up to 1427 7 days, which may be too short a period for nanomaterials in view of their potential persistence in 1428 organs. 1429 1430 The surface chemistry i.e. possible coatings and/or adhesion of biomolecules to the surface of the 1431 nanoparticles does affect the tissue distribution and toxicokinetics. For Polyethylene glycol (PEG) 1432 coated nanomaterials prolonged blood circulation was demonstrated (Niidome et al., 2006, Lankveld 1433 et al., 2011). In this respect also the passage of the nanoparticles through the respiratory and/or GI 1434 tract may influence the absorption and thus the systemic organ exposure. For several metallic 1435 nanoparticles also the possibility for dissolution and/or degradation should be considered. So, careful 1436 evaluation of the organ content is needed when performing (toxico)kinetic studies for nanomaterials. 1437 In this respect the combination of separation techniques like field flow fractionation (FFF) might be 1438 used together with so called single particle ICP-MS to identify the presence of particles in tissues. 1439 1440

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1441 4. MECHANISMS OF IMMUNOTOXICITY 1442 1443 4.1 General principles of toxicity of nanomaterials 1444 4.1.1 Role of physico-chemical material properties 1445 In order to sytematically investigate toxic effects of ENM it would be desirable to correlate their toxic 1446 effects with their physico-chemical properties. However, this approach is not straightforward, as 1447 many physico-chemical properties are strongly entangled and are difficult to control independently 1448 (Fadeel et al. 2013; Rivera-Gil et al. 2013). Nevertheless, in this section, we discuss how material 1449 properties are linked to ENM toxicity, focusing on the importance of size, shape, charge, and 1450 solubility or dissolution. 1451 1452 Size matters, in particular for cellular uptake of nanoparticles. For some nanoparticles, size has been 1453 shown to drive toxicity through direct interactions with cellular receptors or intracellular structures. 1454 For instance, Tsoli et al. (2005) found that so-called Au55 clusters display cytotoxicity which seemed 1455 to be caused by the unusually strong interaction between the 1.4 nm particles and the major grooves 1456 of DNA. Only marginally smaller or larger particles of the same chemical composition showed 1457 drastically reduced toxicity. Sargent et al. (2009) provided evidence of CNT interference with the 1458 mitotic spindle in lung epithelial cells leading to induction of aneuploidy, an early event in the 1459 progression of many cancers. The nanotube bundles are similar to the size of microtubules that form 1460 the mitotic spindle and such “bio-mimicry” may explain how these nanomaterials are incorporated 1461 into the mitotic spindle apparatus. Moreover, SWCNTs have approximately the same diameter (1-4 1462 nm) the DNA double helix and can physically intertwine with DNA (Pampaloni and Florin. 2008). It is 1463 generally believed that the greater the intracellular dose of nanoparticles, the more pronounced are 1464 the toxic effects that they generate. Chan and co-workers demonstrated, using gold nanoparticles, 1465 that there can be an optimal size for cellular uptake (Chithrani et al. 2006; Chithrani et al. 2007). 1466 However, it is important to note that cellular uptake does not necessarily lead to cell death as studies 1467 have shown mesoporous silica nanoparticles or silica-coated iron oxide nanoparticles to be relatively 1468 inert at the doses tested despite considerable cellular uptake in macrophages (Witasp et al. 2009; 1469 Kunzmann et al. 2011). Moreover, cellular uptake is not mandatory for cytotoxicity to occur: cobalt- 1470 chromium nanoparticles were shown to cause damage to fibroblasts across an intact cellular barrier 1471 without having to cross the barrier. The outcome, which included DNA damage without overt cell 1472 death, was different from that observed in cells subjected to direct exposure to nanoparticles 1473 (Bhabra et al. 2009). 1474

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1475 Nanoparticle shape is also important. In particular, phagocytosis of foreign materials such as ENMs 1476 can be influenced by their aspect ratios. Using HeLa cells as a model, particles of ellipsoid shape were 1477 found to be more readily engulfed than spherical particles (Gratton et al. 2008). In contrast, 1478 nanomaterials with dramatically high aspect ratios and rigidity (such as the “rod-like” CNTs) may 1479 resist uptake by macrophages, but could cause cellular damage through piercing of the plasma 1480 membrane (so-called frustrated phagocytosis), resulting in the leakage of cellular constitutents 1481 including cytokines and reactive oxygen species (Murphy et al. 2012). Similarly, changing a relatively

1482 inert material, such as TiO2, into a fibrous structure results in a toxic nanomaterial that provokes 1483 inflammatory responses in macrophages (Hamilton et al. 2009). In addition to the overall shape of 1484 the nanoparticles, the smoothness/roughness of the particle’s surface also affects the opsonization 1485 of the particle and its subsequent uptake by phagocytes (discussed below). Furthermore, in a study 1486 using gold nanoparticles with different aspect ratios, Schaeublin et al. (2012) found that the gold 1487 nanospheres were non-toxic, whereas the gold nanorods induced apoptosis. Notably, both 1488 nanoparticles formed agglomerates in cell culture medium, but the spherical particles had a large 1489 fractal dimension (i.e., tightly bound and densely packed) while the nanorod agglomerates had a 1490 small fractal dimension (i.e., loosely bound). Similarly, the in vivo pulmonary toxicity of SWCNTs has 1491 also been demonstrated to be highly dependent on their agglomeration; hence, ‘fiber-like’ materials 1492 may sometimes behave as agglomerated particles and not as individual fibers (Mutlu et al. 2010; 1493 Murray et al. 2012). It was suggested that the effective surface area of CNTs (based on geometrical 1494 analysis of the agglomerates) might be a more useful dose metric than specific surface area or 1495 particle number (Murray et al. 2012). 1496 1497 Surface properties, such as the charge (zeta potential) and hydrophobicity, directly affect the 1498 interaction of nanomaterials with biological surfaces, cell membranes and proteins (Stark. 2011). In 1499 general, owing to the overall net negative charge of cellular surfaces, positively charged 1500 nanoparticles are incorporated faster by cells than negatively charged ones, leading to high rates of 1501 non-specific internalization and a shorter half-life in the circulation. The higher toxicity of positively 1502 charged nanoparticles is generally correlated to their enhanced cellular uptake (Oh et al. 2010). 1503 Schaeublin et al. (2011) demonstrated the effects of surface charge on the modality of cell death 1504 induced by a particular nanomaterial. Thus, charged gold nanoparticles induced cell death through 1505 apoptosis whereas neutral gold nanoparticles triggered necrosis in a human keratinocyte cell line. To 1506 elucidate surface charge-dependent toxicity, nanoparticles with different surface charge, for which 1507 the other physico-chemical parameters remain constant are required, but this is often 1508 experimentally difficult to achieve. Overall, the physico-chemical properties of nanoparticles are 1509 strongly linked and are difficult to control independently (Fadeel et al. 2013). Moreover, size, shape

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1510 and surface charge are altered when the nanoparticles interact with the biological environment. In 1511 addition, dissolution of metal and metal oxide nanoparticles in acidic cellular compartments 1512 (lysosomes) underlies the so-called Trojan horse-type mechanism of particle toxicity (Stark. 2011). 1513 This mechanism has been attributed to oxides of zinc, iron, manganese, cobalt and copper 1514 nanoparticles which drive toxicity through releasing of toxic ions into cells. Cho et al. (2012) 1515 correlated the toxicity of 15 different metal/metal oxide nanoparticles with one of two physico- 1516 chemical parameters: zeta potential under acid conditions for low-solubility nanoparticles and 1517 solubility (degree of dissolution) for high-solubility nanoparticles. Hence, the authors found that the 1518 inflammogenic potential of high-solubility nanoparticles depends on the ions that are produced 1519 during their dissolution inside the acidic phagolysosomes of the cells (Cho et al. 2012). For some 1520 (photocatalytic) particles, the crystal structure is an important determinant of toxicity. Sayes et al.

1521 (2006) studied the effects of TiO2 nanoparticles and found that the extent to which these particles 1522 affected cellular behavior was not dependent on surface area; instead, the phase composition of the

1523 particles was important insofar as anatase TiO2 was 100 times more toxic than rutile TiO2. The most 1524 cytotoxic nanoparticles were also the most effective at generating ROS (Sayes et al. 2006). 1525 1526 4.1.2 Direct cytotoxic effects of nanomaterials 1527 With respect to immunotoxicity of ENMs, this could certainly result from a direct cytotoxic effect 1528 culminating in cell death. Understanding which cell death modality that is engaged upon nanoparticle 1529 exposure could lead to strategies by which to mitigate these effects. There are several different 1530 modes of cell death, including apoptosis (a form of programmed cell death with numerous upstream 1531 and downstream regulators), and regulated necrosis or necroptosis (cell death dependent on 1532 RIP1/RIP3 kinase activation), as well as pyroptosis (cell death mediated by caspase-1 activation), 1533 ferroptosis (cell death mediated by lipid peroxidation), autophagy (controlled destruction of survival 1534 factors or organelles within a cell) and others (Andón and Fadeel. 2013). For example, cerium 1535 dioxide, an ENM widely used as a fuel additive, induces apoptosis and autophagy in human 1536 peripheral blood monocytes (Hussain et al., 2012). Recent studies showed that ultrasmall (<10 nm in 1537 diameter) poly(ethylene glycol)-coated silica nanoparticles triggered ferroptosis in starved cancer 1538 cells, which could potentially be exploited for therapeutic purposes (Kim et al. 2016). Notably, acute 1539 and chronic exposure to the same nanomaterial could induce drastically different effects at the 1540 cellular level. Hence, exposure of cells to CNTs or CNT based scaffolds induces apoptosis (Bottini et al. 1541 2006; Dinicola et al. 2015). However, Wang et al. (2011) reported that chronic exposure to CNTs may 1542 yield an apoptosis-resistant and tumorigenic phenotype in lung epithelial cells. The same authors 1543 also demonstrated the induction of cancer stem cell-like cell upon chronic exposure of human lung 1544 epithelial cells to CNTs (Luanpitpong et al. 2014).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1545 1546 4.1.3 Altered cellular signaling by nanomaterials 1547 In addition to direct killing of immune cells, ENMs also have the potential to disrupt extracellular or 1548 intracellular signaling and thereby impair cell function or enhance pro-inflammatory activity of the 1549 cells. For example, Ni nanoparticles have been shown to prolong growth factor (PDGF)-initiated 1550 phosphorylation of MAP kinase in rat pleural mesothelial cells (Glista-Baker et al., 2012). This 1551 resulted in enhanced production of the chemokines, CCL2 and CXCL10 which play important roles in 1552 the recruitment of macrophages, monocytes, and lymphocytes during the pleural immune response 1553 to ENMs such as Ni or Ni-containing CNTs. Sustained phosphorylation of signaling intermediates such 1554 as MAP kinases or receptor tyrosine kinases is achieved by ROS generation by metal nanoparticles. 1555 ROS inactivate protein tyrosine phosphatases that are necessary to dephosphorylate activated 1556 kinases (Samet and Tal. 2010). Tian et al. (2017) reported that nano-sized graphene oxide retarded 1557 cellular migration via disruption of the actin cytoskeleton. Kodali et al. (2013) have shown that pre- 1558 treatment of macrophages with iron oxide nanoparticles at non-cytotoxic doses caused extensive 1559 transcriptional reprogramming in response to subsequent challenge with bacterial 1560 lipopolysaccharides or LPS. Macrophages exposed to nanoparticles displayed a diminished 1561 phagocytic activity toward certain bacteria. The authors concluded that the effects of nanoparticles 1562 may be indirectly manifested only after challenging normal cell function, such as phagocytosis of 1563 pathogens. 1564 1565 4.1.4 Role of oxidative stress for nanotoxicity 1566 The generation of reactive oxygen species (ROS) in cells is a common feature of nanomaterial- 1567 induced toxicity (Li et al. 2008; Nel et al. 2009). Two major mechanisms of ROS generation have been 1568 proposed. First, the direct interaction of nanomaterials with molecular oxygen generates free 1569 radicals such as superoxide ion and hydroxyl radical. The oxidative potential of metal oxide 1570 nanoparticles (e.g., Zn, Cu, Fe) is due to generation of ROS via Fenton-like reactions. Second, the 1571 activation of inflammatory cells (e.g., macrophages, neutrophils) by nanomaterials stimulates the 1572 release of ROS by these cells through activation of the NADPH oxidase. Using NADPH oxidase- 1573 deficient mice, Shvedova et al. (2008) showed that NADPH oxidase-derived ROS play a direct role in 1574 determining the course of the pulmonary response to CNTs. Nanomaterials can also disrupt 1575 mitochondrial homeostasis. For example, Ag nanoparticles inhibit the electron transport chain and 1576 adenosine triphosphate (ATP) synthesis in cells (Asharani et al. 2009). Nanomaterials also have the 1577 potential to inhibit antioxidant pathways. ROS generated by nanomaterial exposure can, in turn, act 1578 as signaling intermediates to activate intracellular signaling targets, including receptor tyrosine 1579 kinases, mitogen-activated protein (MAP) kinases, and transcription factors leading to transcriptional

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1580 activation and the expression of genes involved in pulmonary immune responses. However, while 1581 ROS generation or, more specifically, oxidative stress (i.e., the imbalance between ROS production 1582 and the cell’s inherent ability to detoxify the reactive intermediates or repair the resulting oxidative 1583 damage) is often viewed as the final common pathway of nanoparticle-induced toxicity, it is prudent 1584 to ask whether oxidative stress is a specific event that plays a role in mediating cellular damage, for 1585 instance, through the activation of apoptosis, or whether ROS generation represents a secondary 1586 event resulting inevitably from disruption of biochemical processes within the dying cell? (Shvedova 1587 et al. 2012). The answer to this question will have important ramifications for the development of 1588 strategies with which to reduce toxicity of ENMs. 1589 1590 4.2 Nanoparticle interactions with the immune system 1591 1592 The immune system represents the primary defense system against foreign intrusion, including both 1593 pathogens and particles. Immunity, in essence, is the balanced state of having adequate defenses 1594 against infection or other biological intrusion, while retaining tolerance to avoid destructive actions 1595 towards the body’s own cells. The immune system can be divided into the ‘primitive’ or innate 1596 immune system which is quick to react to foreign intrusion and the more delayed, but highly specific 1597 adaptive immune system (Boraschi et al. 2017). Importantly, there is cross-talk between the innate 1598 and the adaptive arms of the immune system, and some cell types that were previously considered 1599 as being primitive innate immune cells, such as the neutrophil, are also involved in the orchestration 1600 of adaptive responses. In order to understand the potential toxicity of engineered nanomaterials, it is 1601 necessary to understand the immune system with its multiple, specialized cell types that work in 1602 concert, and its soluble mediators, including cytokines and complement factors (Farrera and Fadeel. 1603 2015). Moreover, it is important to consider not only material-intrinsic properties of the pristine 1604 nanomaterial, but also the acquired, context-dependent ‘identity’ of a nanomaterial in a living 1605 system resulting from the adsorption of biomolecules on its surface (Fadeel et al. 2013). 1606 Furthermore, one should also consider the fact that nanomaterials may adopt a “new” identity 1607 through the adsorption of biomolecules (proteins, lipids, and so on), a phenomenon which, in turn, is 1608 linked to nanomaterial intrinsic properties eg. size and hydrophobicity. In this chapter, we discuss the 1609 structure and function of the immune system and review available literature on interactions of 1610 nanomaterials with the immune system following three main routes of exposure, namely dermal 1611 exposure, pulmonary exposure, and oral exposure. 1612 1613 4.2.1 The innate immune system

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1614 The immune system consists of an array of dedicated cell types that work in concert using complex 1615 detection, communication and execution systems to defend us from any external or internal harm. 1616 The immune system is typically divided into the innate immune system which is quick to react to 1617 foreign intrusion and the more delayed but highly specific adaptive immune system which is 1618 endowed with immunological memory after an initial response to a specific pathogen, leading to an 1619 enhanced response to subsequent encounters with that same pathogen; the latter mechanism is the 1620 basis for vaccination, one of the triumphs of modern medicine. It has been proposed (Bhattacharya 1621 et al. 2013; Farrera and Fadeel. 2015) that most if not all of the adverse effects of ENMs are exerted 1622 via direct effects on cells of the innate immune system, including macrophages, or via dendritic cells 1623 (DCs), phagocytosis-competent, antigen-presenting cells that act as a bridge between the two arms 1624 of the immune system, while the effects on the adaptive immune system (i.e., B cells and T cells) are 1625 thought to be indirect. This is not surprising as the cells of the so-called mononuclear phagocyte 1626 system (MPS) have as one of their main functions to clear the body from unwanted foreign materials 1627 (pathogens, particles). In the following sections, published data on in vitro and in vivo interactions of 1628 nanomaterials with the innate and adaptive immune system are highlighted along with a discussion 1629 of the bio-corona. 1630 1631 4.2.1.1 Physical and biological barriers 1632 In addition to specialized immune cells belonging to the innate or adaptive immune system, which 1633 are found throughout the body, the organism is also protected by anatomical barriers which may also 1634 be viewed as part of the innate immune system. These include physical, chemical and biological 1635 barriers. The skin is the main barrier that protects our body from the external environment and acts 1636 as the first line of defence against invading microorganisms. In the respiratory and gastrointestinal 1637 tract, movement due of cilia (the so-called mucociliary escalator or peristalsis in the GI tract helps 1638 remove infectious agents. On the skin and in the airways, small peptides known as defensins are 1639 secreted and contribute to the ‘primitive’ defense against infection. The microflora in the gut also 1640 protects the host organism by competing with pathogenic bacteria and studies in recent years have 1641 shown that the gut microbiome takes an active role in shaping and instructing the immune system 1642 (Schroeder and Bäckhed. 2016). This is an area of potentially great significance when it comes to 1643 exposure to ENMs (Pietroiusti et al. 2016), and more research is therefore needed on the potential 1644 interplay between ENMs, the microbiome, and the immune system. 1645 1646 Notably, the skin, gastrointestinal tract, and the lungs are in direct contact with the environment and 1647 direct exposure to nanomaterials may therefore occur. Consequently, understanding the anatomical, 1648 chemical and biological barrier function at these locations is of great importance. Furthermore, it is

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1649 also important to note that nanoparticles may translocate across various biological barriers and 1650 reach organs distal to the portal of entry. Hence, nanoparticles may avoid macrophage clearance in 1651 the lungs and cross the air-blood barrier and enter into the systemic circulation leading to their 1652 distribution to other organs (Choi et al. 2010). In the pregnant mother, the placenta protects the 1653 unborn child and studies in recent years have addressed whether nanoparticles could cross the 1654 placental barrier causing harm to the fetus (Hougard et al., 2015). 1655 1656 4.2.1.2 Phagocytosis and clearance of nanoparticles 1657 Phagocytosis, the process whereby macrophages, DCs and other myeloid cells such as neutrophils 1658 internalize various targets including microbes or cell debris, is a key mechanism of innate immunity. 1659 Numerous studies have demonstrated that phagocytosis of nanoparticles also occurs. However, 1660 when the target is physically too large for the macrophage to engulf this may trigger ‘frustrated 1661 phagocytosis’ leading to the release of pro-inflammatory mediators; this is seen, for instance, when 1662 asbestos fibers or long, rigid multi-walled CNTs are injected into mice (Bhattacharya et al. 2013). If a 1663 foreign body is too large, a granuloma is formed. In some cases, macrophages fuse with each other to 1664 form ‘giant cells’ that encapsulate the foreign body; this encapsulation of offending agents is an 1665 ancient defense strategy and may also be of relevance for immune responses to nanomaterials. 1666 Understanding how phagocytes (Figure 1) interact with nanoparticles is thus of key importance. In 1667 addition, it may be important to consider differences between the various populations of 1668 macrophages, depending on tissue of origin and activation status. Activated macrophages are 1669 commonly divided into M1, or classically activated macrophages, and M2, or alternatively activated 1670 macrophages (Mills et al. 2000). Hence, while M1 are thought to be more pro-inflammatory and 1671 oriented toward pathogen killing, M2 macrophages have been linked to anti-inflammatory, wound 1672 healing and tissue repair functions. The presence of so-called Th1 cytokines has a tendency to 1673 polarize macrophages toward the M1 phenotype while Th2 immune responses can induce 1674 macrophage polarization toward the M2 phenotype (Mills et al. 2000). It should be noted that the 1675 classical M1-M2 categorization has been viewed as an oversimplification (Murray et al. 1676 2014).Nevertheless, it has been shown that macrophage phenotype determines uptake of silica 1677 particles (Gallud et al. 2017) and gold particles (MacParland et al. 2017). Jones et al. (2013) showed 1678 that mouse strains that are prone to Th1 immune responses clear nanoparticles at a slower rate than 1679 Th2-prone mice. Interestingly, the authors could show that granulocytes – mainly neutrophils – as 1680 well as macrophages participated in the clearance observed in Th2-prone mice. The implication of 1681 this preclinical study is that the immune status of an individual may impact on nanoparticle 1682 clearance. In a subsequent study, the immune system was found to undergo changes in tumor

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1683 bearing mice leading to increased particle clearance caused by an increase in M2 macrophages (Kai 1684 et al. 2016), further supporting the view that the immune status affects nanoparticle clearance. 1685 1686

1687 1688 Figure 1. Schematic illustration showing the various organs of the immune system and the 1689 distribution of innate and adaptive immune cells, with emphasis on tissues rich in tissue resident 1690 macrophages. From: Elsabahy and Wooley. Cytokines as biomarkers of nanoparticle immunotoxicity. 1691 Chem Soc Rev. 2013;42(12):5552-76. 1692 1693 An undesired effect of nanoparticles is the impairment of phagocytic activities. For instance, SWCNTs 1694 were described to impair macrophage engulfment of apoptotic target cells (Witasp et al. 2009). 1695 Similar observations have previously been made for ultrafine carbon particles, which impair ingestion 1696 of microorganisms by human alveolar macrophages (Lundborg et al 2001). Thus, inhalation of 1697 nanoparticles could lead to increased susceptibility to pulmonary pathogens as their clearance might 1698 be impaired. Indeed, delayed bacterial clearance is seen in mice exposed to SWCNTs via pharyngeal 1699 aspiration (Shvedova et al. 2008). Furthermore, as discussed above, Kodali et al. (2013) showed that 1700 pre-treatment of macrophages with iron oxide nanoparticles caused extensive transcriptional re- 1701 programming in response to subsequent challenge with bacterial LPS. Macrophages exposed to

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1702 nanoparticles displayed a phenotype suggesting an impaired ability to transition from an M1 to M2- 1703 like activation state, associated with a diminished phagocytic activity toward certain bacteria (Kodali 1704 et al. 2013). 1705 1706 Several studies in recent years have shown that CNTs are susceptible to degradation by naturally 1707 occurring plant and human enzymes (peroxidases) (Vlasova et al. 2016). Myeloperoxidase (MPO) is a 1708 component of the microbicidal system of phagocytes, especially neutrophils, and thus an important 1709 part of the innate immune response. Notably, MPO expressed in primary human neutrophils is 1710 capable of degradation of oxidized SWCNTs (Kagan et al. 2010). Furthermore, neutrophil extracellular 1711 traps (NETs) produced by activated neutrophils can ‘capture’ and digest SWCNTs in an MPO- 1712 dependent manner (Farrera et al. 2014). Additionally, eosinophil peroxidase (EPO), the major 1713 oxidant-producing enzyme in eosinophils, degrades SWCNTs (Andón et al. 2013) and CNTs can also 1714 undergo biodegradation in activated macrophages through a superoxide/peroxynitrate-driven 1715 oxidative pathway (Kagan et al. 2014). Thus, innate immune cells are apparently capable of 1716 enzymatic degradation of bacteria and fungi as well as carbon-based nanomaterials. However, it 1717 remains unclear how important these enzymatic biodegradation mechanisms are in the lungs of 1718 experimental animals and humans. For example, while MPO-dependent degradation of SWCNTs was 1719 demonstrated in the lungs of mice (Shvedova et al. 2012), abundant intact MWCNTs remain 1720 biopersistent in the lungs and pleura of mice months after inhalation, suggesting that biodegradation 1721 may be limited (Ryman-Rasmussen et al., 2009). However, it is noted that MWCNTs are not as readily 1722 degraded as SWCNTs (Russier et al. 2011). This issue has important implications for the toxicity of 1723 ENMs as biopersistence versus biodegradability may determine whether nanomaterials exert long- 1724 term adverse effects in the body. 1725 1726 4.2.3 Cellular and soluble mediators of inflammation 1727 Inflammation (derived from Latin, ‘to set on fire’) is an adaptive response involving soluble factors 1728 and cells that is triggered in response to infection, trauma, ischemia, toxic or other injury (Nathan. 1729 2002). In inflammation, foreign agents are detected by inflammatory cells upon entry into the body, 1730 resulting in the stimulation of these cells with production of soluble factors such as cytokines and 1731 chemokines (Medzhitov. 2008). The most important inflammatory cells are the monocytes- 1732 macrophages, neutrophils, eosinophils and mast cells, all of which have been implicated in responses 1733 to nanomaterials. Importantly, both infectious and non-infectious inducers of inflammation trigger 1734 the activation of conserved signaling pathways including the activation of the inflammasome in 1735 macrophages, with release of IL-1β, one of the key mediators of inflammation (Broz and Dixit, 2016). 1736 Regardless of the trigger, the purpose of the inflammatory response is to remove or sequester the

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1737 offending agent, to allow the host to adapt, and ultimately, to restore functionality to the tissues 1738 (Medzhitov. 2008). However, if the process becomes chronic, other cell types such as lymphocytes 1739 and fibroblasts may become involved and the adaptive changes may damage the host. Thus, it is 1740 important to distinguish between transient, protective responses to an insult versus chronic, 1741 maladaptive ones. 1742

1743 Macrophages are a part of the mononuclear phagocyte system. These cells are professional 1744 phagocytic cells and are endowed with proteolytic and catabolic activities. Macrophages originate 1745 from monocytes that when they become tissuebound, convert into cells with enhanced phagocytic 1746 capacity (Underhill and Goodridge, 2012). Macrophages are usually distinguished from DCs by their 1747 expression of F4/80, CD11b and Fc surface receptors although some authors have argued that DCs 1748 are simply a heterogeneous subset of the mononuclear phagocyte system with no unique adaptation 1749 for antigen presentation (Geissmann et al. 2010; Galli et al. 2011). Macrophages detect foreign 1750 matter using receptors such as the Toll-like receptors (TLRs) (Underhill and Goodridge, 2012). TLRs 1751 belong to the pattern recognition receptors (PRRs) expressed by cells at the front line of host 1752 defense, eg. macrophages, DCs, and epithelial cells. PRRs enable these cells to detect and respond to 1753 the presence of so-called danger- and pathogen-associated molecular patterns (DAMPs and PAMPs, 1754 respectively). Neutrophils or polymorphonuclear granulocytes are granular leukocytes possessing a 1755 nucleus with three or more lobes. They are the most short-lived cells in the bloodstream. The 1756 turnover is tremendous; it has been estimated that about 60% of hematopoietic activity in the bone 1757 marrow is directed toward neutrophil production. Neutrophils are specialized in the phagocytosis of 1758 bacteria and other pathogens and are capable of phagocytosis on their own (i.e., without the 1759 participation of any other immune mechanism) or through antibody- or complement-mediated 1760 opsonization of pathogens (Nauseef and Borregaard, 2014). Neutrophils are equipped with oxidative 1761 and proteolytic enzymes for the destruction of ingested pathogens. Neutrophils also release so-called 1762 neutrophil extracellular traps or NETs that contribute to the defense against extracellular bacterial 1763 and fungal pathogens; in addition, NETs play a role in non-infectious (sterile) inflammation (Jorch and 1764 Kubes, 2017). Eosinophils are the least common white blood cell and their presence is associated 1765 with parasite infections, allergic diseases such as asthma, chronic lung inflammatory states, and 1766 hyper-eosinophilic syndrome (Fulkerson and Rothenberg, 2013). Eosinophils are found in the 1767 circulation, but reside predominantly in mucosal surfaces where they can remain for weeks. 1768 Eosinophils are particularly effective in ridding the organism of multicellular pathogens known as 1769 parasites. Following activation, eosinophils degranulate to release cytotoxic granule proteins that are 1770 capable of causing tissue damage. Mast cells play a role in innate immunity and host defense against

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1771 parasites, and in tissue repair, and angiogenesis. Mast cells are similar to basophil granulocytes in the 1772 blood and are activated by a variety of stimuli through various receptors including receptors for IgE. 1773 Once activated, the mast cell produces histamine, leukotrienes, proteases, cytokines, chemokines 1774 and other substances that cause immediate airway inflammation, leading to asthma symptoms and 1775 in some cases anaphylactic shock (Holgate. 2012)

1776 1777 The key soluble mediators of the innate immune system are the cytokines, chemokines, and 1778 complement factors. These proteins not only allow immune cells to communicate, but also facilitate 1779 the clearance of pathogens through opsonization, i.e., decoration of microbial surfaces to stimulate 1780 macrophage or neutrophil engulfment. Conceptually, this is similar to the bio-corona formation on 1781 the surface of nanoparticles (discussed below). The major role of chemokines is to act as 1782 chemoattractants to guide the migration of immune cells (Fernandez and Lolis, 2002). The release of 1783 chemokines is often stimulated by pro-inflammatory cytokines. Monocytes-macrophages, 1784 neutrophils, eosinophils, and mast cells all express receptors for different chemokines. Tumor 1785 necrosis factor (TNF)-α is a key cytokine involved in inflammation and is produced mainly by 1786 activated macrophages (Murray et al. 2014). The IL-1 family is a group of cytokines which plays a 1787 central role in the regulation of immune and inflammatory responses to infectious or non-infectious 1788 insults (Dinarello et al. 2012). Among the family members, IL-1α and IL-1β are the most studied. 1789 Finally, the complement system is an ancient and integral part of the innate immune system that 1790 helps or ‘complements’ antibodies. Traditionally, complement has been viewed as a supportive first 1791 line of defense against microbes, quickly tagging them for phagocytosis. However, it is now 1792 understood that complement factors not only complement, but even orchestrate immunological and 1793 inflammatory processes, thus contributing substantially to tissue homeostasis (Ricklin et al. 2012). 1794 The complement system is composed of thirty distinct plasma and cell-bound proteins that are 1795 activated through three different pathways, the so-called classical, alternative, and lectin pathways. 1796 It is important to note that inadvertent complement activation could lead to anaphylaxis, a serious 1797 allergic reaction that is rapid in onset and may cause death (Szebeni. 2014). Thus, it is essential to 1798 evaluate nanomaterials with respect to their propensity for complement activation. 1799 1800 4.2.4 Inflammasome activation by nanomaterials 1801 The inflammasomes are protein complexes in phagocytic cells that regulate the activation of IL-1β, a 1802 pro-inflammatory cytokine involved in infectious and non-infectious inflammation (Davis et al., 1803 2011). Inflammasomes are activated by so-called danger- and pathogen-associated molecular 1804 patterns (DAMPs and PAMPs, respectively). PAMPs include conserved molecular motifs found in

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1805 bacterial cell wall components (eg., lipopolysaccharide, LPS, and peptidoglycan, PGN) or viral 1806 DNA/RNA, as well as fungal glucans. DAMPs, on the other hand, are endogenous stress signals 1807 (sometimes referred to as ‘alarmins’) and several such factors have been identified including high- 1808 mobility group box 1 protein (HMGB1), uric acid, nucleotides, and heat shock proteins (HSPs) 1809 (Bianchi. 2007). DAMPs are released from activated or dying cells and their recognition by DCs and 1810 macrophages via pattern recognition receptors (PRRs) results in immune cell maturation and the 1811 production of pro-inflammatory cytokines (Davis et al., 2011) Notably, several studies in recent years 1812 have shown that various classes of nanomaterials also trigger inflammasome activation, specifically, 1813 the NLRP3 inflammasome (Sun et al. 2013). 1814 1815 Several members of the NOD-like receptor (NLR) family, a subset of cytoplasmic PRRs, are able to 1816 sense PAMPs and DAMPs and subsequently induce the assembly of the inflammasome, which serves 1817 as an activation platform for caspase-1, a central mediator of innate immunity (Broz and Dixit, 2016). 1818 Active caspase-1, in turn, promotes the maturation and release of interleukin-1β (IL-1β) and IL-18. It 1819 appears that DAMPs and PAMPs synergize to permit secretion of IL-1β: PAMPs (such as LPS) 1820 stimulate synthesis of pro-IL-1β, but not its secretion; while DAMPs can stimulate assembly of an 1821 inflammasome and activation of caspase-1, which cleaves pro-IL-1β into IL-1β. Dostert et al. (2008) 1822 initially demonstrated that asbestos fibers and crystalline silica trigger activation of the NLRP3 1823 [nucleotide-binding domain, leucine-rich family (NLR), pyrin-containing 3] inflammasome, and 1824 several other studies have subsequently shown that the NLRP3 inflammasome responds to a range of 1825 different ENMs including metal/metal oxides and carbon-based nanomaterials (Yazdi et al. 2010; 1826 Lunov et al. 2011; Palomäki et al. 2011; Yang et al. 2013; Baron et al. 2015; Ramadi et al. 2016; 1827 Andón et al. 2017). It is worth noting that there are several inflammasomes (Strowig et al. 2012). 1828 However, while all inflammasomes recognize certain PAMPs or DAMPs, it is a distinctive feature of 1829 the NLRP3 inflammasome that it is activated by many diverse stimuli. Overall, it seems that innate 1830 immune cells may respond in a similar manner to pathogens as well as to ENMs and other exogenous 1831 substances, eg., alum (Eisenbarth et al. 2008). 1832 1833 The mechanism of activation of the NLRP3 inflammasome has been the subject of much research, 1834 and reactive oxygen species (ROS) production, either through the activation of the NADPH oxidase 1835 and/or through perturbation of mitochondrial function as well as lysosomal damage or a drop in 1836 intracellular potassium concentrations have all been implicated in this process. Specifically, the 1837 release of the cysteine protease, cathepsin B from lysosomes has been implicated in inflammasome 1838 activation, not least by nanomaterials (Sun et al. 2013). Li et al. (2014) reported that rare earth oxide 1839 nanoparticles triggered inflammasome activation through a biotransformation process within

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1840 lysosomes resulting in lipid membrane dephosphorylation and organelle damage, with release of 1841 cathepsin B. In a subsequent study, rare earth oxide nanoparticles were found to interfere with 1842 autophagosome fusion with lysosomes, thereby disrupting the homeostatic regulation of activated 1843 NLRP3 complexes, leading to enhanced IL-1β production (Li et al. 2014a). Sun et al. (2015) provided 1844 evidence for NADPH oxidase-generated ROS in lysosomal damage and subsequent IL-1β production 1845 in macrophages exposed to MWCNTs. In a recent study, caspase-1 activation and production of IL-1β

1846 in bone marrow-derived macrophages was shown to occur upon simultaneous exposure to SiO2 and

1847 TiO2 nanoparticles at concentrations at which these nanoparticles individually did not cause

1848 macrophage activation, and marked lung inflammation was observed in mice treated with both SiO2

1849 and TiO2 (Tsugita et al. 2017). Notably, SiO2 localized in lysosomes and TiO2 nanoparticles did not,

1850 while only TiO2 nanoparticles triggered ROS production, suggesting distinct forms of cell damage 1851 leading to inflammasome activation. 1852 1853 Overall, the NLRP3 inflammasome appears to function as a sensor and integrator for pro- 1854 inflammatory stimuli. It should be noted that the in vitro assessment of inflammasome activation is 1855 usually performed using LPS-primed cells, implying that nanoparticles exposure alone might not be 1856 sufficient to induce inflammasome activation and that other concomitant signals would be required 1857 in a living organism. MWCNT exposure was recently shown to increase secretion of HMGB1 (an 1858 endogenous stress signal or DAMP) in alveolar macrophages, and neutralization of extracellular 1859 HMGB1 reduced MWCNT-induced IL-1β secretion in vivo (Jessop and Holian. 2015). These findings 1860 provide a basis for the sterile (i.e., non-infectious) inflammation triggered by MWCNTs and perhaps 1861 other nanomaterials. 1862 1863 4.3 The adaptive immune system 1864 4.3.1 Nanomaterials and the adaptive immune system 1865 The innate and adaptive immune system work together in a coordinated manner. The innate immune 1866 system is the first line of defense against foreign intruders and its activation leads to signals (eg. 1867 cytokine secretion) for adaptive immune response activation. The innate immune response is 1868 immediate and occurs, for instance, through engulfment of the offending pathogen, while the 1869 adaptive immune response takes days or weeks; however, once an immune response has been 1870 established, the immune system can respond rapidly to a new encounter with the same pathogen. 1871 Exposure to nanomaterials and other xenobiotics could in principle lead to immune suppression, in 1872 which the immune system would fail to expand in response to a pathogen, or immune stimulation, in 1873 which case the immune system would overrespond, leading potentially to autoimmune or allergic 1874 disease (Boraschi et al. 2017).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1875 1876 The adaptive immune system is composed of B cells and T cells. B cells are responsible for humoral 1877 (antibody-mediated) immunity while T cells are involved in cell-mediated responses. T helper cells 1878 (CD4+) are needed to support the production of antibodies by B cells; T helper cells, in turn, are 1879 subdivided into Th1, Th2, and Th17, depending on their specific role and cytokine profiles. Cytotoxic 1880 T cells (CD8+) are required for killing of virus-infected and malignant cells, while regulatory T cells are 1881 required for maintenance of immune tolerance, which is important for the discrimination between 1882 ‘self’ and ‘non-self’ (i.e., pathogens). DCs, in turn, constitute the bridge between the innate and the 1883 adaptive arms of the immune system. These cells are effective phagocytic cells which also exhibit a 1884 capacity for processing and presentation of antigens (i.e., a substance which provokes an adaptive 1885 immune response) (Banchereau and Steinman, 1998). Importantly, DCs migrate from peripheral 1886 tissues to lymph nodes, where they can stimulate T cells and B cells. Targeting DCs may be 1887 advantageous, for instance, in vaccination strategies (Hubbell et al. 2009). However, undesired 1888 targeting of DCs may be linked to nanomaterial toxicity. In fact, several studies have reported that 1889 ENMs may disturb the functions of DCs, which may affect B cells and/or T cells. For instance, 1890 MWCNTs were found to alter the capacity of human monocytes to differentiate into DCs (Laverny et 1891 al. 2013) and pulmonary exposure to SWCNTs has been shown to induce diminished proliferation of 1892 splenic T cells in mice through direct effects on DCs (Tkach et al. 2011). Nanoparticulate carbon black 1893 present in cigarette smoke was found to accumulate in human myeloid DCs from emphysematous 1894 lung and administration of carbon black induced inflammation and Th17-dependent emphysema in 1895 mice (You et al. 2015). 1896 1897 Importantly, surface modification may serve to mitigate or otherwise modulate ENM effects on 1898 immune cells. Zhi et al. (2013) studied the effects of GO with and without polyvinylpyrrolidone (PVP) 1899 coating on human DCs, T cells and macrophages, and found that PVP-coated GO exhibited lower 1900 immunogenicity compared with uncoated GO in terms of inducing DC differentiation and maturation. 1901 It is noteworthy that hydrophobicity has been hypothesized as a conserved damage-associated 1902 molecular pattern or DAMP that plays a role in activating the immune system (Seong and Matzinger. 1903 2004). The suggestion is that hydrophobic portions of proteins are normally hidden from the immune 1904 system, but act as danger or damage signals upon exposure or release. Recent studies using 1905 nanoparticles appear to support this view. Moyano et al. (2012) generated a panel of gold 1906 nanoparticles with varying degree of surface hydrophobicity and measured cytokine responses in 1907 splenocytes after exposure to the nanoparticles. Splenocytes are comprised mainly of B cells, but 1908 also include T cells and monocytes and thus represent both the innate and adaptive arms of the 1909 immune system. The authors could show a direct correlation between the hydrophobicity of

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1910 nanoparticles and cytokine expression, and a similar result was also noted in mice following a single 1911 intravenous injection of the nanoparticles (Moyano et al. 2012). In line with these findings, Shahbazi 1912 et al. (2014) reported that surface chemistry of porous silica particles dictated the 1913 immunostimulatory effects on human DCs. 1914 1915 4.4 Immunosuppression versus immune activation 1916 Nanomaterials can elicit either immunostimulatory or immunosuppressive effects (Dobrovolskaia 1917 and McNeil. 2007). Mitchell et al. (2009) found that inhalation of MWCNTs produced a systemic 1918 immunosuppression in mice with decreased T cell proliferation in the spleen due to a signal from the 1919 lung, most likely TGF-β secreted by alveolar macrophages. As already mentioned, nanoparticles have 1920 been shown to influence T cell proliferation, typically through an effect on antigen-presenting cells, 1921 leading to an enhanced T cell stimulatory capacity (Farrera and Fadeel. 2015). Using a fully human 1922 autologous immunological construct as a non-animal alternative to monitor nanoparticle responses, 1923 Schanen et al. (2009) showed that exposure to TiO2 nanoparticles led to elevated levels of pro- 1924 inflammatory cytokines and increased maturation of DCs. Additionally, the nanoparticles effectively

1925 primed activation and proliferation of naive CD4+ T cells in comparison to micrometer-sized TiO2. 1926 1927 Myeloid-derived suppressor cells (MDSC), a heterogenous population of myeloid progenitors, are 1928 increasingly being recognized as important players in the immunosuppression induced by tumors. 1929 This is accomplished either through direct or indirect mechanisms, i.e., by directly influencing 1930 effector T cells, or through the generation and/or expansion of other regulatory cell populations, 1931 such as regulatory T cells (reviewed in Serafini. 2013). Shvedova et al. (2013) reported that 1932 metastatic establishment and growth of Lewis lung carcinoma (LLC), a commonly used model of 1933 pulmonary metastatic disease, is promoted by SWCNTs and that this effect is likely mediated by 1934 increased accumulation of MDSC, as their depletion prevented the pro-tumor activity of the 1935 nanomaterials. Additionally, acute exposure to SWCNTs induced recruitment of MDSC in the lungs of 1936 exposed mice and MDSC-derived production of TGF-β, resulting in an increased tumor burden in the 1937 lungs (Shvedova et al. 2015). 1938 1939 Mast cells are well known to act in response to danger signals through a variety of receptors and 1940 pathways including IL-33 and the IL-1-like receptor, ST2, and as such are critical in innate and 1941 adaptive immune responses, while playing a role in allergic conditions. Katwa et al. (2012) 1942 demonstrated the crucial involvement of mast cells and the IL-33/ST2 axis in pulmonary and 1943 cardiovascular responses to MWCNTs. Thus, toxicological effects of MWCNTs were observed only in 1944 mice with a sufficient population of mast cells and were not observed when mast cells were absent

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1945 or incapable of responding to IL-33. Furthermore, in a more recent study, Rydman et al. (2014) 1946 compared two types of MWCNTs, rigid rod-like and flexible tangled MWCNT, in terms of their effects 1947 following inhalation exposure in mice. Mast cell deficient mice were used to evaluate the role of 1948 these cells in the inflammatory process. The authors found that even a short-term inhalation of the 1949 rod-like MWCNT induced allergic airway inflammation. Mast cells were found to partially regulate 1950 the inflammation caused by rod-like MWCNT, but alveaolar macrophages were also suggested to 1951 play an important role in the disease. Unlike rod-like MWCNTs, flexible or tangled MWCNT do not 1952 appear to directly cause allergic airway inflammation when administred to the lungs of mice, but 1953 tangled MWCNTs do exacerbate allergen-induced lung inflammation by enhancing levels of growth 1954 factors (PDGF and TGF-1) or mucins (MUC5AC and MUC5B) as has been demonstrated in either 1955 ovalbumin (OVA) allergen and house dust mite (HDM) allergen mouse models (Ryman-Rasmussen et 1956 al., 2009; Shipkowski et al., 2015). Li et al. (2014) reported that single walled CNTs (SWCNTs) 1957 exacerbated OVA-induced allergic asthma in rats and that this exacerbation was counteracted by 1958 concurrent administration vitamin E. This work thus points to an important role of reactive oxygen 1959 species in this model. Shurin et al. (2014) reported that graphene oxide attenuates Th2 immune 1960 response in a model of OVA-induced asthma, but leads to potentiation of airway remodeling and 1961 hyperresponsiveness. Moreover, exposure to GO increased the macrophage production of the 1962 mammalian chitinases, CHI3L1 and AMCase, whose expression is associated with asthma (Shurin et 1963 al. 2014). For a further discussion on pulmonary effects of ENMs, refer to the section below on 1964 pulmonary exposure. In general, there is considerable evidence that ENMs either cause or 1965 exacerbate allergic lung inflammation in rodents, suggesting that certain ENMs might pose a health 1966 concern for individuals with pre-existing asthma or as an enciting agent for asthma. 1967 1968 4.5 Immunogenicity of engineered nanomaterials 1969 An important issue that needs to be addressed is nanomaterial immunogenicity, i.e., the ability to 1970 elicit a specific adaptive immune response (Ilinskaya and Dobrovolskaia. 2016). To date, 1971 immunogenicity of nanomaterials has been observed only rarely. It has been shown that 1972 polyethyleneglycol (PEG)-coated nanoparticles (eg., liposomes) are able to act as T cell-independent 1973 immunogens and elicit a direct B cell response with the production of anti-PEG IgM antibodies, 1974 responsible for accelerated clearance of the nanoparticles upon subsequent administrations (Mima 1975 et al. 2015; Shiraishi et al. 2016). Recognition of PEG represents a problem in the context of 1976 nanomedicine as PEG is widely used for particle coating. However, no response against the particles 1977 themselves was detected in the aforementioned studies. Indeed, it seems unlikely that 1978 nanomaterials per se are immunogenic. However, it is well-known that certain small molecules, 1979 referred to as haptens, can elicit an immune response when coupled to a protein carrier (Boraschi et

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

1980 al. 2012). Indeed, immunization of mice with a C60 fullerene derivative conjugated to a protein 1981 (bovine thyroglobulin) yielded a population of fullerene-specific IgG antibodies (Chen et al. 1998). 1982 Similarly, induction of specific anti-particle antibody responses was reported upon in vivo inoculation 1983 of polyamidoamine (PAMAM) dendrimers conjugated with a protein (eg., the cytokine, IL-3), 1984 constructed to increase the bioavailability and half-life of the protein (Lee et al. 2004). Indeed, while 1985 the dendrimer was not immunogenic by itself, it became so when conjugated with IL-3, as well as 1986 with other proteins, such as bovine serum albumin. Furthermore, it is possible that the acquired bio- 1987 corona on the surface of nanoparticles could induce specific immune responses, if proteins bound to 1988 the nanoparticles were to undergo conformational changes, thereby revealing cryptic epitopes. 1989 Further studies are required to address this possibility. 1990 1991 4.6 Nanoparticles as adjuvants 1992 Nanosized particles have for a long time been used as adjuvants in vaccine preparations (Kuroda et 1993 al., 2013; He et al., 2015). Particulates and crystals induce inflammatory responses and thereby 1994 stimulate the immune system. One of the most commonly used adjuvants in vaccines is aluminum 1995 hydroxide (He et al., 2015). The vaccine adjuvant preparation can be prepared by adding aluminum 1996 salts to a solution of antigen (aluminum-precipitated method) or by adding antigen to a previously 1997 prepared aluminum hydroxide (aluminum-adsorbed method). The aluminum hydroxide particles may 1998 have a rather heterogeneous size distribution. It was reported that the smaller particles induced a

1999 better immune response. Antibody titers were higher in the vaccine adsorbed with Al(OH)3 with

2000 mean diameter of 200 nm compared to that containing Al(OH)3 particles of 600 nm diameter. In 2001 addition, the adjuvant with smaller particle size had better physical characteristics and absorption 2002 efficiency (Huai et al., 2011). He et al. (2015) concluded that nanoscale aluminum hydroxide induces 2003 higher antibody responses in various antigen adjuvant combinations and would be a good candidate 2004 for vaccine adjuvant. Sun et al. (2013) synthesized a library of aluminum oxyhydroxide (AlOOH) 2005 nanorods and found that the crystallinity and surface hydroxyl group display of AlOOH nanoparticles 2006 impacted the activation of the NLRP3 inflammasome in human THP-1 myeloid cells or murine bone 2007 marrow-derived DCs; thus, shape, crystallinity, and hydroxyl content play an important role in NLRP3 2008 inflammasome activation and are potentially useful for boosting of antigen-specific immune 2009 responses. PGLA (poly(lactic-co-glycolic acid) and liposomal nanoparticulates are now under 2010 investigation as adjuvants in cancer vaccines (Varypataki et al., 2016). PLGA nanoparticles induced an 2011 increase in mucosal and systemic immune responses when the particles were used as nanocarrier for 2012 hepatitis B surface antigen (Mishra et al., 2011). For thiol-organosilica particles (size range 100 to 925 2013 nm) the smallest particles induced an immune response in the Peyer’s Patches as indicated by an 2014 increased number of dendritic cells and IgA levels in intestinal secretions (Awaad et al., 2012).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2015 Enhancing pulmonary immunity may be accomplished by modification of the surface of 2016 polyanhydride nanoparticles. The surface modification with carbohydrates to target C-type lectin 2017 receptors on antigen presenting cells resulted in an increased uptake by alveolar macrophages and 2018 enhanced activation phenotype (Chavez-Santoscoy et al., 2012). In addition to the use as adjuvant 2019 itself nanomaterials are also studied for their capacity as antigen carriers. Different PLGA polymers 2020 showed a size dependent uptake in antigen presenting cells, while the uptake was independent of 2021 the polymer composition (Rietscher et al., 2016). Both in vitro and in vivo OVA directed immune 2022 responses were enhanced compared to OVA alone. Thus, important lessons may be learned from the 2023 extensive work on adjuvants in terms of how the immune system responds to nanoparticles. This has 2024 implications for innate and adaptive immune responses to coexposure to nanoparticles and 2025 allergens, for instance in the lungs or on the skin. 2026 2027 4.7 Immunologically susceptible populations 2028 Life expectancy is increasing in developed countries. Consequently, the health of the elderly is rapidly 2029 becoming a major public health concern. Specific immune responses in the elderly are less effective 2030 than in adults with decreased T and B cell activation, while innate immune reactions may be 2031 increased due to a constitutively inflammatory microenvironment (reviewed in: Boraschi et al. 2013). 2032 However, several reports show that macrophages and DC, being constitutively activated in the 2033 elderly, are less responsive to activation by external stimuli. Moreover, immunological frailty due to 2034 chronic diseases, infection and malnutrition, is often associated with a state of enhanced 2035 inflammation that induces constitutive macrophage and DC activation. Therefore, one could surmise 2036 that nanomaterials, which interact mainly with phagocytic cells, as we have discussed in preceding 2037 sections, may trigger fewer unwanted reactions in the elderly population. However, information is 2038 lacking and research on this topic is certainly needed. The situation is different in infants and very 2039 young children, in which the immune system is immature and therefore less able to cope with 2040 external challenges, but not constitutively inflamed as in the elderly. It is also important to note that 2041 barrier functions in infants may differ from those in adults (Semmler-Behnke et al. 2012) and this 2042 could certainly impact on the biodistribution of nanomaterials and, therefore, on the outcome of 2043 nanomaterial exposure in infants versus adults. In any case, focused studies using relevant models 2044 are required to assess the detrimental effects of nanomaterials in immunologically susceptible or frail 2045 populations (elderly, chronically ill) and individuals with pre-existing conditions including infections, 2046 allergy or asthma, with particular consideration of the effects of chronic exposure, in order to 2047 understand the real-life impact of ENMs. 2048 2049 4.8 The role of the biomolecule corona

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2050 The interaction of engineered nanomaterials with the surface of cells or with biomolecules has been 2051 termed the ‘nano-bio’ interface (Monopoli et al. 2012). These interactions are of critical importance 2052 for our understanding of the biological behavior of nanomaterials and not least for the effects of 2053 nanomaterials on the immune system. Indeed, the immune system has evolved to recognize and 2054 respond to foreign intruders with or without a coating of endogenous opsonins such as complement 2055 factors or immunoglobulins and it stands to reason, therefore, that immune cells would also be able 2056 to respond to nanoparticles with or without a surface corona of biomolecules (Farrera and Fadeel. 2057 2015). 2058 2059 It is important to consider that nanomaterials do not present pristine surfaces in a biological system. 2060 Instead, upon entry into a biological system, nanomaterials rapidly adsorb biomolecules forming a 2061 bio-corona on the surface (Monopoli et al. 2012). It is generally believed that the bio-corona has two 2062 components, a ‘hard’ corona that is stable and a ‘soft’ corona; the hard corona is thought to bestow 2063 a ‘biological identity’ on the nanomaterial. In other words, the hard corona is what the cells ‘see‘. 2064 Whether or not the hard corona covers the surface of the nanomaterial completely remains 2065 unresolved. Notwithstanding, the binding of proteins to nanoparticle surfaces may not only afford a 2066 new ‘identity‘ to the nanoparticle, but this interaction may also affect the protein that are adsorbed. 2067 In an illustrative example, Deng et al. (2011) showed that poly(acrylic acid)-coated gold nanoparticles 2068 bind fibrinogen, a protein involved in blood clot formation, in a charge-dependent manner inducing 2069 unfolding of the protein and that binding to integrin receptors on the surface of a monocytic cell line 2070 led to activation of the NF-кB pathway and secretion of the pro-inflammatory cytokine, TNF-α. 2071 Moreover, Yan et al. (2013) noted that albumin undergoes conformational changes upon adsorption 2072 onto nanoporous polymer particles leading to a significant decrease in internalization in 2073 undifferentiated THP1 cells, in comparison to the bare particles, while scavenger receptor-mediated 2074 uptake in differentiated, macrophage-like THP.1 cells was enhanced by the presence of the unfolded 2075 albumin. The latter study emphasizes that adsorbed proteins, such as albumin, may act as opsonins 2076 to promote uptake, or, on the contrary, as ‘dysopsonins’ depending on the cellular context. In a 2077 comprehensive study using silica and polystyrene nanoparticles of various sizes and different surface 2078 functionalizations, Tenzer et al. (2013) could show that plasma protein adsorption occurs very rapidly 2079 and that this affects haemolysis (i.e. lysis of red blood cells), thrombocyte activation, cellular uptake 2080 and endothelial cell death. However, it remains to be firmly demonstrated if and how the specific 2081 identity of the corona proteins is linked to toxicity or other cellular responses. Dobrovolskaia et al. 2082 (2014) reported that the composition of the protein corona did not correlate with compatibility of 2083 colloidal gold nanoparticles with cells of the blood. Ge et al. (2011) reported that the toxicity of 2084 SWCNTs was mitigated after the binding of serum proteins, but this does not necessarily prove that

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2085 specific proteins are involved; the serum proteins may have saturated the binding sites of the CNTs 2086 thereby preventing further interactions with cellular proteins. Furthermore, in vitro studies using 2087 various cell lines have demonstrated that the presence of a protein corona affects toxicity of silica 2088 nanoparticles, possibly through passivation of the particle surface (Dutta et al. 2007; Shi et al. 2012; 2089 Lesniak et al 2012). The relative proportion of protein versus the amount of nanoparticles also 2090 influences the extent of corona formation. An increase in serum concentration near in vivo serum 2091 levels suppressed in vitro cytotoxic activity of nanoparticles (Kim et al., 2014). Thus, while there is 2092 increasing evidence for the importance of the protein corona in regulating the cellular interactions of 2093 nanoparticles, not least cellular uptake of nanoparticles, it appears difficult at this stage to identify a 2094 role for individual protein(s) in the corona. However, some recent studies have provided evidence for 2095 a role of specific proteins in the regulation of cellular uptake. Hence, Fedeli et al. (2015) identified 2096 histidine-rich glycoprotein as the major component of the plasma derived hard corona on silica 2097 nanoparticles and showed that this protein functioned as a dysopsonin when particles were 2098 incubated with cells. Saha et al. (2016) reported that surface functionality can be used to tune the 2099 protein corona formed on surface of cationic gold nanoparticles, dictating the interaction of the 2100 nanoparticles with macrophages and evidence was provided for recognition of specific complement 2101 proteins in the bio-corona. 2102 2103 Furthermore, while the majority of bio-corona studies to date have been performed using human 2104 plasma or fetal bovine serum as a source of proteins, it is important to note that the bio-corona may 2105 consist not only of proteins, but also lipids, sugars, nucleotides and other biomolecules. Moreover, 2106 the bio-corona composition may of course vary depending on the portal of entry. Hence, 2107 nanoparticles that are inhaled into the airways are more than likely covered with a corona of lung 2108 surfactant consisting of both lipids and proteins, while nanoparticles that are taken up through the 2109 oral route would encounter an entirely different mix of biomolecules in the gastrointestinal tract. 2110 Kapralov et al. (2012) found that pharyngeal aspiration of SWCNTs in mice resulted in adsorption of 2111 lung surfactant proteins and surfactant lipids and, furthermore, that this protein-lipid bio-corona 2112 facilitated uptake of SWCNTs by murine macrophage-like cells. Gasser et al. (2012) found that pre- 2113 coating of MWCNTs with porcine pulmonary surfactant (Curosurf) affected their oxidative and pro- 2114 inflammatory potential in vitro. Raesch et al. (2015) performed proteomics and lipidomics studies of 2115 the corona on nanoparticles in contact with lung surfactant and found differences in protein, but not 2116 lipid binding for nanoparticles with different surface functionalizations. For the GI-tract also food 2117 content as a source for biomolecules in the corona might need to be considered. However, there are 2118 very few studies to date on the gastrointestinal bio-corona on nanoparticles (Ault et al. 2016). It will 2119 also be important to understand whether the bio-corona would change as nanoparticles cross

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2120 biological barriers in the body or whether a biological ‘fingerprint‘ of different anatomical 2121 compartments is retained on the surface. 2122 2123 The adsorption of complement factors may be viewed as a special case of bio-corona formation and 2124 could have pronounced effects if complement is activated (reviewed in: Moghimi et al. 2011). 2125 Polymer coating of nanomaterials may influence complement activation. Complement activation can 2126 be prevented modifying the density of the polymers on polystyrene nanoparticles (Al-Hanbali et al. 2127 2006). Moreover, Hamad et al. (2010) found that alterations of copolymer architecture on 2128 nanoparticles from “mushroom” to “brush” configuration not only switched complement activation 2129 from the C1q-dependent classical pathway to the so-called lectin pathway, but also reduced the level 2130 of generated complement activation products. These findings provide a rational basis for the 2131 intelligent design of immunologically safer nanosystems for clinical applications. In a recent study, 2132 Chen et al (2017) reported that dextran-coated superparamagnetic iron oxide core-shell 2133 nanoparticles incubated in human serum and plasma are rapidly opsonized with the third 2134 complement component (C3) via the alternative pathway. Thus, complement factors may bind to the 2135 nanoparticle surface, or to the bio-corona. 2136

2137 The conjugation of common allergens such as birch/grass pollen or house dust mite to gold 2138 nanoparticles was explored in a recent study and enhanced allergic responses were found to such 2139 conjugates (Radauer-Preiml et al. 2016). Specifically, in this in vitro study, gold nanoparticles were 2140 conjugated with the major allergens of birch pollen (Bet v 1), timothy grass pollen (Phl p 5) and house 2141 dust mite (Der p 1) and differences in the activation of human basophil cells derived from birch/grass 2142 pollen- and house dust mite-allergic patients in response to free allergen and nanoparticle-allergen 2143 conjugates were determined using the basophil activation assay. The formation of a stable corona 2144 was found for all three allergens used. The data suggested that, depending on the allergen, different 2145 effects could be observed after binding to nanoparticles, including enhanced allergic responses 2146 against Der p 1 and also, for some patients, against Bet v 1 (Radauer-Preiml et al. 2016). In summary, 2147 the data showed that conjugation of allergens to ENMs may modulate human allergic responses. 2148 Thus, the formation of an ‘allergen-corona‘ could perhaps lead to the exacerbation of symptoms in 2149 sensitized individuals.

2150 2151 4.9 The microbiome and the immune system

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2152 A wide variety of microorganisms inhabits the body surfaces of essentially all vertebrates. A 2153 microbiota is the ecological community of microorganisms that share our body space. In the 2154 lower intestine these organisms have evolved to degrade a variety of dietary substances which 2155 enhances host digestive efficiency and ensures a steady nutrient supply for the microbes. The 2156 microorganisms which are present on human skin, our largest body part, was traditionally 2157 considered as a cause of infections, but it has become clear that the specific microbes are 2158 necessary for defense and thus for the health of the skin. Invasion of host tissue by resident 2159 bacteria has serious health consequences including inflammation and even sepsis. It is evident 2160 that this host-microbiota relationship is playing an important role in the development and 2161 functioning of the immune systems and immune system is crucial to preserve the symbiotic 2162 relationship between host and microbiota (Hooper et al., 2012). 2163 2164 Much of our current understanding of microbiota-immune system interactions has been 2165 acquired from studies of germ-free animals. Such animals are housed in sterile isolators to 2166 control their exposure to microorganisms, including viruses, bacteria, and parasites (i.e., 2167 gnotobiotics animals). More recently, next-generation sequencing technologies have opened 2168 new avenues to investigate microbiota complexity. Although microbial cell cultures are still 2169 needed, the complex composition of microbiotas can be efficiently studied by 16S ribosomal 2170 RNA sequencing and by unbiased deep sequencing of microbial DNA (i.e., shotgun sequencing). 2171 These high-throughput analyses have made possible the construction of defined microbiotas in 2172 different anatomical locations of the host as well as to compare microbiotas between different 2173 animal species and between individuals. Recent technological advances have also provided tools 2174 (e.g. transcriptomics, epigenomix and metabolomix) to explore how microbiota shape many 2175 aspects of host physiology and immunity. The application of these new approaches in the 2176 context of older techonologies has revolutionized the study of interactions between the 2177 microbiota and the immune system (Hooper et al., 2012). 2178 2179 Unravelling the complex interactions between the microbial environment and host tissues in the 2180 context of exposure to various types of ENMs will provide important insights into how ENMs 2181 influence mechanisms of maintenance or disruption of tissue homeostasis. Studies in recent 2182 years have underscored that the human microbiome is critically implicated in many diseases, 2183 e.g., obesity, diabetes, atherosclerosis, cardiovascular disease, rheumatoid arthritis, COPD, 2184 asthma and skin diseases. Information about ENM-microbiota interaction will provide

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2185 information, not only about possible deleterious effects of ENMs on microbiota and associated 2186 pathophysiologies, but it also give grounds to develop effective post-exposure treatments that 2187 might attenuate or eliminate the impact of ENMs on pathological processes mediated by 2188 microbiome (Pietroiusti et al., 2016). 2189 2190 Skin provides a unique opportunity for host-microbiome interaction studies due to its 2191 continuous interactions with environment and optimal accessibility (sampling) to both 2192 microbiota and underlying tissues (Belkaid et al., 2014). As an ENM exposure route, skin is 2193 relevant target because of the extensive use of cosmetics, sun screens and contact with ENM- 2194 coated surfaces. Intact skin is an efficient penetration barrier to topical ENMs but there is a lack 2195 of knowledge with regards to the effects of ENM exposure in the case of weakened or disrupted 2196 skin, e.g. in conditions such as contact-, atopic- and irritant dermatitis. More importantly, 2197 virtually nothing is known about the possible effects of cutaneous ENM exposure on the 2198 composition of the skin microbiome and host-microbiome interactions in ENM-related tissue 2199 inflammation. 2200 2201 The airways are likely the most relevant route for ENM exposure in occupational settings. Until 2202 recently, the airways were thought to be sterile unless infected. However, 16S and shotgun 2203 sequencing approaches have revealed that the airways harbour a unique steady-state 2204 microbiota (Dy et al., 2016). Virtually nothing is known about effects of ENM exposure on airway 2205 microbiome and its role in the maintenance or disruption of airway tissue homeostasis. The 2206 gastro-intestinal tract is an important exposure route for consumers, but it is considered less 2207 relevant for workers. It should be noted, however, that a substantial percentage of inhaled 2208 nanoparticles are cleared by the muco-ciliary escalator cells into the oral cavity and thereafter 2209 into the gastrointestinal tract, and that ENMs deposited in the skin may reach the gut lumen 2210 through hand-mouth contact. Therefore, gut microbiome is probably involved even in the case 2211 of pulmonary or skin exposure (Pietroiusti et al., 2016). 2212 2213 The consequences of ENM interactions with human and environmental microbiota and the 2214 outcome of this interaction are probably among the most challenging tasks for the research in 2215 nanotoxicology in the next years. However, only a limited amount of information about the 2216 effects of ENM on microbiota exits in the literature. Most of the previous articles are focusing on 2217 ENM on in vitro setting on single microbes or using microbial culture methods for intestinal

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2218 microbial samples. Available data suggest that ENMs may cause adverse health effects through 2219 the direct killing of microorganisms, or through alterations of their function. It should be noted, 2220 however, that microbiota research in nanotoxicology is still in its infancy. To date, only a few 2221 large scale microorganism analyses using modern 16S sequencing or similar approaches have 2222 been conducted to investigate the effects of ENMs (Chen et al., 2014). It was reported recently 2223 that while silver nanoparticles did not elicit any overall toxicity in mice exposed following oral 2224 exposure to human relevant doses, the particles could induce microbial alterations in the gut (van 2225 den Brule et al. 2016). Silver nanoparticles are expected to have antibacterial effects, but data are 2226 lacking for other ENMs. 2227 2228 4.10. Immunotoxicity depending on route of exposure 2229 Inadvertent exposure to nanomaterials may occur primarily via three different routes: on the skin, or 2230 via inhalation into the lungs, or through oral exposure with transmission via the gastrointestinal 2231 tract. Therefore, in the following sections, we focus the discussion on these three routes of exposure, 2232 with a main emphasis on pulmonary exposure as this is by far the most commonly studied route. 2233 Translocation of nanoparticles to the systemic circulation may occur, but we will not discuss direct 2234 intravenous administration of nanomaterials as this is relevant only in the context of nanomedicine 2235 and nanomedicines are, in principal, covered by existing regulations for medicines or medicinal 2236 products and subjected to the same scrutiny as any other medicinal products. Neverthless, 2237 investigatingthe distribution and fate of nanomaterials in the body is of utmost importance for our 2238 understanding of nanomaterial toxicity. In particular, particles that enter into the blood stream are 2239 rapidly cleared by phagocytic cells belonging to the mononuclear phagocytic system (MPS) and 2240 subsequently primarily end up in the liver or spleen. This, again, points to the importance of the 2241 innate immune system in the surveillance and clearance of foreign objects. However, nanoparticles 2242 may also be excreted from the body. For an overview of the adsorption, distribution, metabolism, 2243 and excretion (ADME) of a nanomaterial, see the section on toxicokinetics. It is important to keep in 2244 mind that the immune system executes its surveillance role in the whole body, not least in the lungs 2245 and in the gastrointestinal tract as these organs are in direct contact with the external environment. 2246 Immune responses, on the other hand, take place largely in the secondary lymphoid organs, i.e. 2247 lymph nodes and spleen. 2248 2249 4.10.12 Dermal exposure to engineered nanomaterials 2250 4.10.2.1 Overview of skin immunology

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2251 Skin is the largest human organ by size and also by surface area in contact with the external 2252 environment (average surface area of ~2m2 in adults). This necessitates an impeccable defense 2253 system to keep foreign insults at bay. The skin does this with an intricate combination of physical and 2254 immunological barriers. The physical barrier is made up of tightly packed corneocytes and an 2255 impermeable envelope of highly crosslinked proteins and lipids. The immunological barrier provided 2256 by the human skin is expectedly complex and results from intricately orchestrated interaction 2257 cascades between multiple cell types (epithelial, stromal and immune cells) and the foreign stimulus 2258 [ Richmond et al., 2014; Shiu et al., 2001; Pasparakis et al., 2014).]. In the epidermis, keratinocytes 2259 and non-epithelial immune cells including Langerhans cells and dendritic epidermal T cells form the 2260 first line of immunological defence. In the dermis, stroma cells such as fibroblasts and endothelial 2261 cells, and a variety of immune cells including macrophages, mast cells, dermal dendritic cells, dermal 2262 T cells and innate lymphoid cells (ILCs) fortify the skin against deeper threats. 2263 2264 Keratinocytes play an integral role in the innate defence system of skin. Upon contact with a wide 2265 variety of exogenous stimuli, potentially including nanomaterials, keratinocytes serve as the first line 2266 of immunological defence by acting as cellular signal transducers. They do this by secreting a myriad 2267 of cytokines and chemotactic agents, as well as by expressing cell surface adhesion molecules (Fig X). 2268 These molecules facilitate the recruitment and retention of circulating immune cells into the 2269 epidermis. Keratinocytes then amplify this initial immune response by producing additional 2270 immunostimulatory molecules once the circulating immune cells arrive in the skin. In addition, 2271 keratinocytes themselves can act as antigen-presenting cells. Collectively, keratinocytes therefore 2272 constitute a major component that link the innate and adaptive immune system in the skin, in order 2273 to support rapid and distinct responses of human skin to exogenous stimuli [Schroder et al., 2010). . 2274 2275 Langerhans cells are the resident dendritic cells in the epidermis. Upon activation, they migrate to 2276 the lymph nodes where they interact with T cells and B cells to initiate and shape the adaptive 2277 immune response. In skin contact hypersensitivity, Langerhans cells activate regulatory T-cells while 2278 inhibiting cytotoxic T-cells [Kplan et al., 2015; Seneschal et al., 2012). Normal human skin is home to 2279 approximately 20 billion T-cells, of which 80% are α/β T-cells [Clark, 2010]. T-cells in the epidermis 2280 are dominated by CD8+ cells while CD4+ T-cells are the majority in the dermis [Foster et al., 1990]. 2281 Dermal dendritic cells are critical in cross-presenting keratinocyte-derived antigens for the activation 2282 of cytotoxic T-cells [Henri et al., 2010; Igyarto et al., 2011]. Aside from multiple immune cells, an 2283 arsenal of anti-microbial peptides (AMPs) are present in the skin, which are significant contributors in 2284 preventing microbial growth in healthy skin [Gallo and Hooper, 2012]. However it has not been 2285 established if AMPs play any role in helping skin cope with exposure to engineered nanomaterials.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2286 2287

2288 2289 2290 Figure X. Schematic figure showing immune responses in the skin. From…xxx 2291 2292 2293 Besides the respiratory tract also for skin sensitization an enhancement of reactivity to the model 2294 sensitizer DNCB was reported when nanoTiO2 was administered subcutaneously at the base of the 2295 ear in a local lymph node assay (Hussain et al., 2012). Augmentation of the Th2 response was 2296 indicated by the increased production of IL-4 and IL-10 by the lymphocyte population isolated from 2297 the draining lymph node. The effective concentration 3 (EC3) for inducing a threefold increased cell 2298 proliferation compared to controls, used as a threshold value in the local lymph node assay (LLNA),

2299 was decreased for DNCB when nanoTiO2 was used but not with pigment TiO2. When amorphous silica 2300 NPs were intradermally injected together with mite antigen of Dermatophagoides pteronyssinus an 2301 aggravation was observed in the D. pteronyssinus induced atopic dermatitis in a nanoparticle size 2302 (range investigated from 30 to 1000 nm) dependent manner, the smaller NPs inducing the most 2303 severe effects (Hirai et al., 2012). In addition, IgE and Th2 responses were enhanced as well. The 2304 dose of silica NPs was fixed at 250 µg resulting in both a higher number of NPs and higher surface 2305 area for the smaller NPs. So, it is not clear whether the effect was a matter of size only (the smaller, 2306 the more active) or depended on the dose itself as higher dosages were used when the dose was 2307 expressed as a different metric (e.g. number of particles or surface area). 2308 Both diesel exhaust particles (DEP) and carbon black (CB) had adjuvant activity in a murine OVA 2309 sensitization model when OVA and DEP or CB were co-administered. Both local lymph node 2310 responses in terms of cell proliferation and systemic IgE production were enhanced in the DEP and 2311 CB treated animals compared to controls (Løvik et al., 1997).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2312 So far, the mechanism of enhanced immune responses by (nano)particles has not yet been 2313 elucidated, even in the use of particles as vaccine adjuvants that have been applied for decades. 2314 However, it is evident that co-exposure to particulates and antigen poses a risk for sensitization, 2315 especially when the exposure occurs in the respiratory tract and the lung. 2316 2317 4.10.2.1 Dermal toxicity of nanomaterials 2318 To answer the question of whether skin exposure to nanomaterials is a threat to the human immune 2319 system, the possibility of the nanomaterial in question to penetrate the skin will need to be 2320 established first. It was widely accepted that penetration of engineered nanoparticles across the skin 2321 is highly unlikely [Dwivedi et al.,2009; Nohynek et al., 2008; Osmond et al., 2010).. Any penetration, if 2322 at all, is trivial and happens through hair follicle openings rather than across the stratum corneum 2323 [Gamer e al., 2006; Nohynek et al., 2007; Filipe et al., 2009). However, the bulk of these early studies 2324 on the topic focussed on penetration across animal or healthy adult human skin models. More 2325 recently, reports have emerged that nanomaterial penetration through skin will increase significantly 2326 when the barrier function of skin is compromised, such as in UV-damaged or tape-stripped skin 2327 [Mortensen et al., 2008; Ravichandran et al., 2011). Quantitatively, Miquel-Jeanjean et al. [Miquel- 2328 Jeanjean et al., 2012] calculated that 0.19 ± 0.15 wt% of TiO2 nanoparticles can reach viable cells in 2329 skin after 24 hr exposure on intact skin, which doubled to 0.39 ± 0.39 wt% when the skin was 2330 damaged by UV irradiation. There is also a current lack of knowledge in relation to nanomaterial 2331 penetration across the skin in situations of long-term use and across skin of varying states of health 2332 [Osmond and McCall, 2010; Zhao and Ng, 2014]. From the perspective of chronic skin exposure and 2333 potential prolonged low-level penetration of nanomaterials across the skin, there is therefore a 2334 significant need to establish immunotoxicity influences of nanomaterials in the skin. Representative 2335 reports of such studies are discussed herein. 2336 2337 4.10.2.3 Results from in vivo studies 2338 Skin irritation and sensitization tests make up the bulk of available reports about skin immunotoxicity 2339 of engineerd nanomaterials. Most of these evaluated the influence of metal-based (pure metals and 2340 metal oxides) nanoparticles due to the frequent use of these in cosmetics, consumer and personal 2341 care products primarily as UV filters and antimicrobial agents. In a recent study, TiO2 nanoparticles 2342 (anatase; < 25 nm diameter) were evaluated for local immune effects after dermal exposure on the 2343 ears of female BALB/c mice [Auttachoat et al., 2014]. After 3 days of exposure to the TiO2 2344 nanoparticles (2.5–10% w/v in 4:1 acetone:olive oil) auricular lymph node cell proliferation was not 2345 affected, based on the Local lymph node assay (LLNA). However, skin irritation, based on measuring 2346 the percentage of ear thickness change (IRR), was observed with 5% and 10% TiO2 exposure.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2347 Following these results, the group sensitized mouse ear skin with 2.5–10% TiO2 nanoparticles and 2348 subsequently challenged the skin with further exposure to 10% TiO2, with no significant change in 2349 ear thickness found. Using a similar mouse model, Smulders et al. [Smulders et al., 2015). exposed 2350 mouse ear skin with TiO2, Ag and SiO2 nanoparticles for a day before applying 2,4- 2351 dinitrochlorobenzene (DNCB) as a known dermal sensitizer for 3 days. It was found that dermal 2352 exposure to Ag or SiO2 nanoparticles prior to DNCB sensitization did not influence the stimulation 2353 index (SI) which was calculated after 6 days of exposure. However, application of 4 mg/mL TiO2 2354 nanoparticles prior to DNCB sensitization, resulted in a significant increase in SI, along with increased 2355 titanium concentration in the draining lymph node cells of this group. 2356 2357 In an attempt to understand skin sensitization of nano-sized UV absorbers, ZnO, TiO2 and SiO2 2358 nanoparticles embedded in textiles were applied onto rabbits for acute dermal irritation evaluation 2359 (OECD TG 404) and guinea pigs for sensitization tests (OECD TG 406) [Piasecka-Zelga et al., 2015]. 2360 Results suggest that none of the analyzed materials or modifiers induced major skin reactions. Only a 2361 TiO2 (TK44) and ZnO (Z11) based modifier were classified as mild sensitizers based on the calculated 2362 primary irritation index. 2363 Guinea pigs were also used for testing skin sensitization from repeated exposure to high aspect ratio 2364 hydroxyapatite nanoparticles of minor axis greater than 50 nm [Geetha et al., 2013]. Through visual 2365 inspection and analysing blood parameters, oxidative stress in liver and brain and DNA damage in 2366 liver, it was concluded that the hydroxyapatite nanoparticles did not induce skin sensitization or 2367 oxidative damage up to a delivered dose of 100μg/mL. 2368 2369 Carbon-based nanoparticles have been explored as potential antioxidants in products for topical 2370 application but little in vivo data is available, likely due to the policy of not using animal testing in the 2371 cosmetic industry and the availability of an in vitro alternative for skin irritation testing as described 2372 in guideline OECD439. In one such study, single-walled carbon nanohorns (SWCNHs) were exposed 2373 onto the back skin of rabbits by applying a cotton lint laden with 0.015g of the as-grown SWCNHs 2374 [Miyawaki et al., 2008]. The primary irritation index was found to be 0, suggesting low acute 2375 toxicities caused by dermal exposure to SWCNHs. 2376 2377 Because of the uncertainty of nanomaterial influence on skin of various conditions, there is a need to 2378 adopt in vivo models of skin diseases to study nanomaterial immunotoxicity. In this respect, atopic 2379 dermatitis models are interesting because of the hypothesis that skin with defective physical barriers 2380 would facilitate higher levels of nanomaterial penetration that could elicit an immune response. 2381 However, there are limited reports of such studies, and most focused on studying metal-based

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2382 nanoparticles. Yanagisawa et al. [Yanagisawa et al., 2009] found that repeated exposure of 20 µg of 2383 TiO2 nanoparticles (15, 50, or 100 nm) in 10 µl of saline to the ear skin of NC/Nga mice over a 17 day 2384 period resulted in significant increase of histamine levels in blood serum and IL-13 expression in the 2385 ear, independent of nanoparticle size. Moreover, the TiO2 nanoparticles aggravated epidermis 2386 thickening when exposed together with mite allergen extracts, along with elevated levels of IL-4 in 2387 the skin, total IgE and histamine in blood serum. The authors concluded that dermal exposure to 2388 TiO2 nanoparticles when skin barrier function is compromised will exacerbate atopic dermatitis 2389 symptoms through Th2-biased immune responses. The same group repeated a similar study to 2390 evaluate size effects of polystyrene nanoparticles on atopic dermatitis [Yanagisawa et al., 2010]. 2391 Similar to TiO2 nanoparticles, polystyrene nanoparticles also aggravated epidermis thickening when 2392 exposed together with mite allergen extracts. In contrast to TiO2, this effect was more significant 2393 with smaller polystyrene particles (25, 50 nm) compared to larger ones (100 nm). However, it should 2394 be noted that the nanoparticles were injected intradermally rather than applied topically in these 2395 studies. 2396 2397 In comparison, Hirai et al. [Hirai, T et al., 2015] did carry out an immunotoxicity study via repeated 2398 topical application of silica nanoparticles (30 nm) on the ears and upper backs of NC/Nga mice. When 2399 silica nanoparticles were applied together with mite allergen extracts, atopic dermatitis-like lesions 2400 were aggravated based on ear thickness and histology assessments, compared to application of 2401 either the silica nanoparticles or the mite allergen extracts alone. In addition, concurrent application 2402 of silica nanoparticles and mite allergen extracts resulted in the production of allergen-specific Th1- 2403 related IgG2a and Th2-related IgG1 which were absent when the mice were exposed to nanoparticles 2404 applied separately from the allergen or to well-dispersed nanoparticles. It was concluded that the 2405 presence of allergen adsorbed agglomerates of silica nanoparticles led to a low IgG/IgE ratio which is 2406 a key risk factor in human atopic allergies. 2407 2408 Besides a transgenic model, it is also possible to evoke local inflammation and allergy in normal skin, 2409 by sensitizing with an allergen or superantigen, to model atopic dermatitis. Using the back skin on 2410 BALB/c mice and ovalbumin/staphylococcal enterotoxin B as the allergen/superantigen cocktail, Ilves 2411 et al. [Ilves et al., 2014] showed that nano-sized ZnO (< 50 nm), as opposed to bulk (> 100 nm), is able 2412 to reach the dermis of the allergic skin. Both particle types were able to alleviate local allergic effects 2413 on the skin, although this ability was more pronounced with the nano-sized ZnO. However, ZnO 2414 nanoparticles induced systemic production of IgE more significantly than larger ZnO particles, 2415 providing evidence of the allergy promoting potential of ZnO nanoparticles. 2416

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2417 Collectively, there is a lack of in vivo studies to investigate the immunotoxicity effects of engineered 2418 nanomaterials in the skin, especially of compromised skin. This is primarily due to a lack of suitable 2419 and validated models. 2420 2421 4.10.2.3 Results from in vitro studies 2422 Nanoparticles of various configurations are being produced for a wide range of applications. These 2423 may be doped or surface treated with other materials to achieve a particular function. It is prudent 2424 that immunotoxicity of such hybrid nanoparticles be evaluated in relation to their stoichiometric 2425 composition or spacial arrangement. In this respect there is significant literature on using in vitro 2426 models, primarily monolayer skin cell cultures, for nanoimmunotoxicity studies. Similar to the in vivo 2427 studies, most in vitro studies focused on metal-based (pure metals and metal oxides) nanoparticles 2428 due to the relevance of their application in commercial products. 2429 2430 Schaeublin et al., 2012] found that nanoparticle of differing aspect ratios could elicit different cellular 2431 response in human keratinocytes. Gold nanospheres (20 nm diameter) were compared with gold 2432 nanorods (16.7 nm diameter, 43.8 nm long) for their toxic effects on immortalised human 2433 keratinocytes (HaCaT). The nanorods were found to be cytotoxic from 25 μg/mL onwards and caused 2434 more significant ROS production compared with the nanospheres. Notably, the nanorods induced 2435 higher levels of IL1α, IL1β, IL18 and TNFSF10, suggesting that gold nanoparticles of a higher aspect 2436 ratio induce greater immune response in keratinocytes. However, the nanospheres were coated with 2437 mercaptopropane sulfonate while the nanorods were coated with polyethylene glycol (PEG). 2438 Although both coating materials are known to be biocompatible, some differential effects could be 2439 present. 2440 Ryman-Rasmussen et al. [Ryman-Rasmussen et al., 2007.] determined that surface coatings can 2441 affect the irritation potential of quantum dots (QDs) in epidermal keratinocytes. QDs of a spherical 2442 core/shell structure with a diameter of 4.6 nm and QDs of an ellipsoid core/shell structure with 2443 diameters of 6 nm by 12nm were coated with PEG, PEG-amine or carboxylic acid to create neutral, 2444 cationic and anionic surfaces, respectively. Primary human keratinocytes were exposed to these QDs 2445 and were found to internalise them after 24 hour. Neutral QDs did not result in significant 2446 cytotoxicity in the keratinocytes. Conversely, only QDs with anionic surfaces significantly increased 2447 the release of pro-inflammatory cytokines IL-1β, IL-6, and IL-8. These data indicate that surface 2448 coating on nanoparticles can be an important determinant of immunotoxicity in human 2449 keratinocytes. 2450

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2451 A valid concern of in vitro nanotoxicology studies is the relevance of the dosage used, in relation to 2452 real life exposure levels. In this aspect, a recent report demonstrated that TiO2 nanoparticles can 2453 induce autophagy in primary human keratinocytes at serial diluted concentrations as low as 100 2454 fg/ml [Zhao et al., 2013], which would be a realistic level of cellular exposure even with the 2455 anticipated low levels of nanoparticle penetration across the skin. Although autophagy is inherently a 2456 survivor mechanism, its induction suggests some level of sinister influence by the nanoparticles. 2457 Autophagy was recently implicated in keratinocyte inflammation response by negatively regulating 2458 p62 expression [Lee et al., 2011], which is involved in increased inflammation and tumourigenesis 2459 [Mathew et al., 2009]. 2460 2461 Carbon-based nanoparticles are the next most common group of nanoparticles used in 2462 immunotoxicity studies. For example, fullerene-based amino acids were tested with human 2463 epidermal keratinocytes over a concentration range [Rouseet al., 2006]. At 0.04 mg/mL, the 2464 fullerene-based amino acids induced significantly higher IL-8 production after 8 to 48 h exposure, 2465 while IL-6 and IL-1β activities were greater after 24 h and 48 h exposure. Conversely, TNF-α and IL-10 2466 levels were insignificant. Various configurations of carbon nanotubes have received significant 2467 attention due to reports of cellular responses that are mechanistically similar to those seen in 2468 asbestosis [Donaldson et al., 2013; Poland et al., 2008) ]. In terms of immunotoxicity, exposing 2469 HaCaT cells to single-wall carbon nanotubes (SWCNT) at concentrations up to 0.24 mg/ml, for 18 2470 hours resulted in elevated oxidative stress levels shown by formation of free radicals, accumulation 2471 of peroxidative products and antioxidant depletion [Shvedova et al., 2003], suggesting the initiation 2472 of an immune response by keratinocytes. 2473 2474 4.10.3 Impact of nanomaterials on the respiratory system 2475 Nanomaterials pose potential risks for immune-mediated lung diseases through occupational, 2476 consumer, or environmental exposures (Card et al., 2008; Bonner, 2010). Immunotoxicity is defined 2477 as any adverse effect on the immune system following toxicant exposure that results in immune 2478 stimulation or immune suppression (Hussain et al., 2012, WHO EHC 180)). In the lung, 2479 immunostimulation increases the incidence of allergic reactions, inflammatory responses, or 2480 autoimmunity, while immunosuppression suppresses the maturation and proliferation of immune 2481 cells, resulting in increased susceptibility to infectious diseases or tumor growth. ENMs have been 2482 reported to exhibit either immunostimulatory or immunosuppressive effects in the lung, and this 2483 largely depends on the specific type of nanomaterial in question. However, the effects of 2484 nanomaterials on the immune system can also depend on the context of exposure; for example, 2485 repeated exposures versus nanomaterial exposure after the establishment of allergic inflammation.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2486

2487 2488 2489 Figure X: Immune cells in the lung under normal conditions and pre-existing asthma. Adapted from 2490 Thompson EA, Sayers BC, Glista-Baker EE, Shipkowski KA, Taylor AJ, Bonner JC. Innate Immune 2491 Responses to Nanoparticle Exposure in the Lung. J Environ Immunol Toxicol. 2014, 1, 150-156.. 2492 2493 4.103.1 Methods of administration to the lungs 2494 Inhalation is a major route of exposure for ENMs and therefore the lungs are a major target of ENM- 2495 initiated immunotoxicity. Because of their small size, inhaled nanoparticles easily reach the lower 2496 respiratory region due principally through diffusion, where they deposit on the alveolar epithelium. 2497 Larger aggregates of nanoparticles may also be subjected to the forces of sedimentation and deposit 2498 in the lower airways of the lung and at alveolar duct bifurcations. Depositions of (nano)particles in 2499 the various regions of the lung can be modelled by using the IRCP or MPPD models as described 2500 above (IRCP 1994; Cassee et al., 2002; Asgharian et al., 2009). Once deposited, ENMs trigger a 2501 complex sequelae of events that activate the immune system in the lungs, beginning with stimulation 2502 of the respiratory epithelium to release soluble chemotactic factors that attract resident alveolar 2503 macrophages as well as neutrophils, eosinophils and lymphocytes to the lungs (Fig. X). Epithelial- 2504 derived chemokines also attract dendritic cells which sample and transport inhaled foreign

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2505 substances, including nanoparticles, to draining lymph nodes to program naïve T helper cells to 2506 differentiated T helper phenotypes (Th1, Th2, Th17, T-reg). These diverse cell types participate in the 2507 complex immunotoxic response to ENMs. 2508 2509 The deposition of inhaled ENMs is determined by a number of factors including particle size, shape, 2510 electrostatic charge, and aggregation state. For example, inhalation exposure to well-dispersed 2511 carbon nanotubes (CNTs) in mice results in deposition in the distal regions; i.e., alveolar duct 2512 bifurcations and alveolar epithelial surfaces, of the lungs of mice or rats (Ma-Hock et al., 2009; 2513 Pauluhn 2010). Agglomeration of ENMs, alternatively referred to as state of dispersion, refers to 2514 nanoparticles that loosely adhere to one another through non-covalent interactions, such as 2515 electrostatic charge, whereas aggregation refers to strongly bound particles by covalent interaction 2516 between particles. Dispersion of agglomerated ENMs can be achieved through surface 2517 functionalization to reduce electrostatic charge, or by suspension of ENMs in surfactant-containing 2518 media. The relative state of dispersion often influences the type of immune or pathologic response 2519 to ENMs. For example, dispersed CNTs cause diffuse interstitial fibrosis throughout the lower lung, 2520 whereas agglomerated CNTs tend to cause focal granuloma formation. Aggregation of ENMs also 2521 depends to some extent on the method of delivery to the lungs. Intratracheal instillation or 2522 oropharyngeal aspiration techniques for delivery to the lungs of rats or mice can result in more 2523 aggregation of ENMs and generally do not faithfully reproduce deposition patterns that are achieved 2524 with inhalation exposures to dry aerosolized or nebulized suspensions of ENMs. There may also be a 2525 difference in the local dose as instillation results in the delivery of the total dose as a bolus whereas 2526 inhalation presents a more evenly distributed exposue dose. However, major advances have been 2527 made in methods for dispersing ENMs in aqueous suspension using surfactant-containing media prior 2528 to instillation or aspiration in rats or mice. Inhaled or well-dispersed instilled ENMs also reach the 2529 sub-pleural region of the lungs either via macrophage-dependent or –independent processes. 2530 Persistent ENMs such as CNTs may remain embedded within the subpleural tissue of mice for 2531 months (Ryman-Rasmussen et al., 2009a). Some CNT-bearing macrophages exit the lung via the 2532 pleural lymphatic system and enter the pleural space or can be found in lung-associated lymph 2533 nodes. This type of translocation of ENMs is discussed further below. 2534 2535 4.10.3.2 Translocation of nanomaterials 2536 The majority of the surface area of the alveolar region of the lung is lined with thin, pancake-shaped 2537 type I alveolar epithelial cells and a lesser surface area of the alveolus is occupied by type II epithelial 2538 cells, which produce surfactant that lines the alveolus and serve as progenitor cells for the type I 2539 epithelial cells. ENMs that enter the alveolus must pass through the surfactant coating covering the

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2540 type I epithelium, then pass either through or around the epithelial cells and the underlying 2541 interstitial space and finally pass through or around the capillary endothelial cells to enter the 2542 bloodstream. Following inhalation exposure, it is estimated that nanoparticles with a diameter less 2543 than ~34 nm have the ability to cross the alveolar epithelium of the lung and endothelium of the 2544 pulmonary capillaries to enter the circulation (Kreyling et al., 2010). The translocation of Au 2545 nanoparticles across the alveolar barrier (air-blood barrier) has be demonstrated to be inversely 2546 related to nanoparticle size; smaller particles with greater specific surface area crossed this barrier 2547 most effectively (Kreyling et al., 2014), although translocation ws low when xpressed as percentage 2548 of the administered dose. Moreover, translocation was greater for negatively charged nanoparticles 2549 than positively charged nanoparticles. Once in the circulation, nanoparticles can again cross the 2550 endothelial barrier of the vascular to reach organs such as the heart, liver, spleen, kidney, and brain. 2551 There is also evidence that ENMs can cross the placental barrier in pregnant mothers to gain access 2552 to the developing fetus (Hougaard et al, 2013).. 2553 2554 ENMs may also translocate from the lungs via the lymphatic system. The pulmonary lymphatic 2555 system is a vascular network that serves to remove excess fluid from the connective tissue spaces of 2556 the lung parenchyma. The lymphatic system is also known to play an important role in clearing 2557 particulate material from the lung to the lymph nodes (Bonner 2008). The lymphatic system in the 2558 lung is divided into superficial and deep portions, but these two portions are connected. The 2559 superficial portion is located in the connective tissue of the pleura lining the lung. The deep portion is 2560 in the connective tissue surrounding the bronchovascular tree. The two portions connect in the 2561 interlobular septa. The lymphatic vessels are structurally similar to thin-walled veins. The presence of 2562 valves in the lymphatic vessels and the movement of the lung during respiration promote the flow of 2563 lymph from the periphery and pleura toward the hilus. Afferent lymphatics from the lung drain into 2564 the tracheobronchial lymph nodes. Lymph from tracheobronchial and hilar nodes drain into the 2565 thoracic, right, and left lymphatic ducts and from these ducts the lymph then drains into the systemic 2566 venous system. Therefore, ENMs can reach the systemic circulation via lymphatic clearance. For 2567 fiber-like ENMs such as CNTs, lymphatic clearance could be an important conduit for translocation to 2568 other tissues. It has been reported that MWCNTs delivered to the lungs by oropharyngeal aspiration 2569 are found in extra-pulmonary organs such as liver and kidney (Mercer et al. 2013). 2570 ENMs, either by direct translocation or through the release of soluble mediators from the lungs, are 2571 capable of activating immune responses in tissues and organ systems beyond of the lung, including 2572 the spleen and heart. For example, inhaled multi-walled CNTs cause systemic immunosuppression in 2573 mice through a mechanism that involves the release of TGF-1 from the lungs, which enters the

2574 bloodstream to signal COX-2-mediated increases in PGE2 and IL-10 in the spleen, both of function to

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2575 suppress T cell proliferation (Mitchell et al., 2007; Mitchell et al., 2009). Also, SWCNTs or MWCNTs 2576 delivered to the lungs of mice have been reported to exacerbate cardiovascular dysfunction and 2577 disease (Erdely et al., 2009). These studies showed no evidence of CNT translocation from the lung. 2578 Rather, the systemic effects were likely due to the release of soluble cytokines or growth factors 2579 from the lung. However, it is possible that at least some ENMs will translocate from the lung to 2580 distant organs to modulate immune responses. 2581 2582 4.10.3.3 Nanomaterial interaction with lung cells 2583 2584 4.10.3.3.2 Lung Epithelium 2585 The lung epithelium is the primary site of deposition for ENMs. The lower respiratory tract is 2586 comprised of the conducting airways (trachea, bronchi and bronchioles) and the lung parenchyma 2587 which consists primarily of gas exchange units (alveoli) (see Fig. X ‘Immune cells in the lung’. The 2588 trachea, bronchi and bronchioles conduct air to the pulmonary parenchyma. The bronchi bifurcate to 2589 form bronchioles and continue to progressively bifurcate in a tree-like fashion to form bronchioles of 2590 decreasing diameter. The most distal conducting segment of the tracheobronchial tree is called the 2591 terminal bronchiole, which bifurcates to form respiratory bronchioles that contain some alveolar 2592 ducts and terminate in clusters of alveolar sacs. Well-dispersed ENMs reach the alveolar region 2593 where they interact with two primary alveolar epithelial cell types; type I and type II cells. Type I cells 2594 cover the majority of the alveolar surface constitute part of the air-blood barrier, which also includes 2595 the capillary endothelium and interstitial compartment sandwiched between the type I epithelium 2596 and endothelium. Type I cells comprise 8-11% of the structural cells found in the alveolar region and 2597 yet cover 90-95% of the alveolar surface (Bonner 2008). Their major function is to allow gases to 2598 equilibrate across the air-blood barrier and to prevent leakage of fluids across the alveolar wall into 2599 the lumen. The type I epithelium is particularly sensitive to damage from a variety of inhaled 2600 toxicants, including ENMs, due to their large surface area. Moreover, their repair capacity is limited 2601 because they have few organelles associated with energy production and macromolecular synthesis. 2602 Type II cells comprise 12-16% of the structural cells in the alveolar region, but cover only about 7% of 2603 the alveolar surface. They are cuboidal cells with a microvillus surface and unique organelles called 2604 lamellar bodies that store surfactant. The major function of type II cells is to secrete surfactant to 2605 lower the surface tension in the alveoli, thereby reducing the filling of the alveolar compartment with 2606 fluid and alveolar collapse. Type II cells also serve as a progenitor cell for type I cells, which cannot 2607 replicate. Therefore, type II cells are critical to alveolar epithelial repair after injury. The wall of the 2608 alveolus is composed of the alveolar epithelium, a thin layer of collagenous and elastic connective 2609 tissue interspersed with fibroblasts (termed the pulmonary interstitium), and a network of capillaries

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2610 lined by endothelial cells. This distance between the alveolar space and the capillary lumen is known 2611 as the air-blood barrier. The air-blood barrier is a multi-layered structure approximately 0.4 2612 micrometers in thickness that consists of an alveolar type I cell, alveolar basement membrane, 2613 interstitial space, endothelial basement membrane, and a capillary endothelial cell. 2614 ENMs that deposit on the type I epithelium either translocate across the air-blood barrier or remain 2615 on the surface or are taken up by the epithelium; this depends largely on the physico-chemical 2616 features of the ENM such as size, shape, and charge. ENMs are further modified by biocorona 2617 formation once they interact with the surfactant layer that covers the alveolar epithelium. ENMs are 2618 also taken up by alveolar macrophages (discussed below). As an example, inhaled MWCNTs that 2619 deposit on the alveolar surface in mice are taken up to some extent by macrophages, but some 2620 singlet MWCNTs and small agglomerates penetrate the type I epithelium. This is likely an important 2621 event in mediating alveolar cell injury and stimulating the type I cells to produce chemokines that 2622 attract circulating inflammatory cells (e.g., neutrophils) from the blood to migrate across the air- 2623 blood barrier to the alveolus. Neutrophilic influx is regulated by CXC chemokines such as CXCL1, 2624 CXCL2, and CXCL5. 2625 While well-dispersed ENMs act in the alveolar region, agglomerated ENMs deposit on the airway 2626 epithelium or in particular at regions where airways and alveolar ducts bifurcate (alveolar duct 2627 bifurcations). The airway epithelium forms a continuous lining for the conducting airways. The varied 2628 composition of the epithelium allows it to perform a variety of functions. First, the epithelium along 2629 with its apical mucus layer and its basal lamina, comprise an important barrier against inhaled ENMs. 2630 The apical surfaces of the airway epithelial cells are connected by tight junctions and effectively 2631 provide a barrier that isolates the airway lumen. Second, the various airway epithelial cells produce a 2632 mixture of secretions composed of (a) an aqueous “sol” phase containing proteins, lipids, and ions 2633 and (b) a gel phase containing mucus. Third, ciliated cells comprise the largest proportion of exposed 2634 cells in the normal airway and as discussed above, they propel the mucus within the airway lumen 2635 proximally, thereby mediating clearance of inhaled particles and debris. Fourth, the airway 2636 epithelium exhibits repair following injury, thereby establishing normal airway architecture. Fifth, the 2637 airway epithelium can produce a variety of soluble mediators (cytokines, growth factors, protease, 2638 and lipid mediators) that modulate the responses of other lung cells including airway smooth muscle 2639 cells, fibroblasts, immune cells and phagocytes. During an immune response to inhaled ENMs, the 2640 airway epithelium is stimulated to secrete chemokines that attract dendritic cells to the basal side of 2641 the epithelium, where they insert their dendritic processes between the epithelial cells to contact 2642 ENMs on the cell surface and after migration present ENMs to naïve T cells in draining lymph nodes 2643 to promote T cell differentiation (discussed further below). 2644

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2645 4.10.3.3.1 Alveolar Macrophages 2646 Alveolar macrophages serve a central immune defense role against inhaled particles and fibers. Most 2647 inhaled ENMs deposited in the distal regions of the lung are avidly taken up by alveolar 2648 macrophages. However, the majority of engulfed ENMs are in an agglomerated/aggregated form 2649 (clusters of nanoparticles). While aggregated ENMs are engulfed by macrophages, individual 2650 nanoparticles can escape immune surveillance and phagocytosis. For example, individual CNTs evade 2651 phagocytosis or uptake by macrophages and can be detected by TEM within epithelial or 2652 mesenchymal cells (Ryman-Rasmussen et al., 2009a). Agglomerated/aggregated ENMs taken up by 2653 macrophages via phagocytosis are cleared from the lungs through two primary mechanisms; 1) the 2654 mucociliary escalator and 2) the lymphatic drainage system. The mucociliary escalator is comprised 2655 of a coating of mucus on the surface of the airways that is constantly moving up the airways by the 2656 coordinated movement of cilia on the airway epithelium (Bonner 2008). Macrophages with engulfed 2657 particles or fibers migrate to the distal portion of small airways where they are transported by the 2658 escalator to larger airways and ultimately out of the trachea where they are swallowed or expelled 2659 through coughing. A secondary macrophage-mediated clearance passage for ENMs out of the lung is 2660 the lymphatic drainage system, which includes lymphatic vessels that drain into the pleural cavity. 2661 Rigid, high aspect ratio ENMs (fiber- or tube-shaped) can present a problem for macrophage- 2662 mediated clearance if the nanofiber or nanotube exceeds the width of the engulfing phagocyte. For 2663 example, migration of macrophages containing CNTs across the pleura could cause DNA damage to 2664 mesothelial cells similar to asbestos fibers (Donaldson et al., 2010). However, whether CNTs possess 2665 pleural carcinogenicity like asbestos fibers remains unknown. It seems likely that when the CNTs 2666 share similar characteristics as asbestos fibers in terms of fibre length, rigidity and biopersistence, 2667 theycan induce tumors (Takagi et al., 2008; Sakamoto et al., 2009). Whereas these non degradable 2668 persistent CNTs with a rigid structure and a certain length did induce tumprs, more tangled CNTs did 2669 not (Muller et al., 2009). Inhaled multi-walled CNTs also reach lung-associated lymph nodes in rats 2670 and macrophages likely play a role in the trafficking of ENMs to lymph nodes. 2671 Certain high aspect ratio ENMs (rigid CNTs, nanowires, nanofibers) are capable of disrupting 2672 macrophage function by causing frustrated phagocytosis, which results in the release of 2673 inflammatory mediators (ROS and cytokines) and cell death (Donaldson et al., 2010; Murphy et al., 2674 2012). The innate immune function of macrophages could also be compromised by formation of 2675 bridges composed of parallel bundles of CNTs that link two or more macrophages (Mangum et al., 2676 2006). Macrophage phagocytosis and chemotaxis in rat alveolar macrophages in vivo is impaired by 2677 exposure to TiO2 nanoparticles (Liu et al., 2010). ENMs have also been reported to impair 2678 phagocytosis of microbes. For example, mice exposed to single-walled CNTs have impaired clearance 2679 of the bacteria Listeria monocytogenase (Shvedova et al., 2008). ENMs can also cause cell death of

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2680 macrophages by inducing mitochondrial stress. For example, CeO2 nanoparticle toxicity to human 2681 peripheral blood monocytes was found to be caused by mitochondrial damage and overexpression of 2682 apoptosis-inducing factor, but was independent of ROS production, and resulted in autophagy 2683 (Hussain et al., 2012). Furthermore autophagy induced by CeO2 nanoparticles was further increased 2684 after pharmacological inhibition of tumor suppressor protein p53. Inhibition of autophagy partially 2685 reversed cell death by CeO2 nanoparticles. 2686 Macrophage activation by ENMs involves a complex network of intracellular signaling pathways, 2687 some of which are designed as protective responses to oxidative stress and cell injury. Specific types 2688 of ENMs have been reported to have either enhance or suppress ROS-mediated events in 2689 macrophages. For example, Ag nanoparticles increase the expression of NF- -2 in 2690 RAW264.7 macrophages, whereas no such pro-inflammatory effect was found the Au nanoparticles 2691 (Nishanth et al., 2011). Multi-walled CNTs also increase levels of the COX-2 enzyme and do so 2692 through a MAP kinase-dependent signaling pathway in mouse RAW264.7 cells in vitro (Lee et al., 2693 2012). Also, multi-walled CNTs activate the antioxidant mediator Nrf2 in cultured human THP-1 cells 2694 (Brown et al., 2010). Both COX-2 and Nrf2 are protective factors against lung disease that are 2695 increased to counteract ROS-induced cellular stress initiated by CNTs. In contrast to MWCNTs, 2696 platinum (Pt) nanoparticles suppress inflammatory responses of RAW264.7 cells by reducing 2697 bacterial lipopolysaccharide (LPS) induction of ROS, ERK phosphorylation and levels of COX-2 and 2698 iNOS (Rehman et al., 2012). The studies collectively suggest that the effects of ENMs on either 2699 inducing or suppressing pro-inflammatory signaling pathways in macrophages could be determined 2700 by multiple factors including shape, elemental composition, and ROS-generating potential of the 2701 ENM. 2702 2703 The uptake of ENMs could have a variety of consequences related to macrophage biology and 2704 function. The high aspect ratio ENMs (e.g., carbon nanotubes) cause inflammasome activation in 2705 macrophages which results in the processing and secretion of the pro-inflammatory cytokines IL-1β 2706 and IL-18 (Palomäki et al., 2011; Meunier et al., 2012; Sun et al., 2012). Inflammasome activation by 2707 multi-walled CNTs and other high aspect ratio materials (e.g., nanofibers, TiO2 nanobelts) is 2708 mediated by lysosomal disruption and ROS production (Sun et al., 2012). Inflammasome activation 2709 leading to the release of mature IL-1β has been proposed as a pro-fibrogenic event and downstream 2710 targets of IL-1R activation include up-regulation for growth factor receptors such as PDGF- 2711 (Hamilton et al., 2012; Wang et al., 2000). 2712 While inflammasome activation and subsequent IL-1β release has been proposed as contributing to 2713 ENM-induced immunotoxicity and lung disease pathogenesis, macrophage IL-1β release is also an 2714 important innate immune response for recruiting neutrophils to the lung to participate in microbial

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2715 killing and the resolution of inflammation (Martinon et al., 2009). Exposure to certain ENMs (TiO2, 2716 ZnO, CNTs) results in a Th1 immune cell microenvironment that promotes polarization of classically 2717 activated macrophages (M1).24 M1 macrophages are capable of inflammasome activation and are 2718 thought to play essential roles in microbial killing and innate immune responses. In allergic asthma or 2719 fibrosis, macrophages are polarized to alternatively activated macrophages (M2) in the presence of 2720 Th2 cytokines IL-4 or IL-13 and M2 macrophages are thought to play important roles in fibrosis and 2721 cancer (Gordon and Martinez, 2010). While little is known about inflammasome activation and IL-1β 2722 release by M2 macrophages, compared with M1 macrophages, a decrease in caspase-1 expression 2723 and pro-IL-1β processing has been described for human monocytes treated with IL-13 (Scotton et al., 2724 2005). IL-13 or IL-4 suppresses inflammasome activation in monocytes in vitro and mice with Th2 2725 type inflammation following house dust mite allergen sensitization display reduced inflammasome 2726 activation and IL-1β production following exposure to MWCNT (Shipkowski et al., 2015). Also, ENMs 2727 alone could modify macrophage phenotype, as discussed above. For example, alveolar macrophages 2728 from rats exposed to CeO2 nanoparticles had increased levels of arginase-1 mRNA, which is a marker 2729 of M2 macrophages (Ma et al., 2011). 2730 2731 4.10.3.3.5 Neutrophils, eosinophils and mast cells 2732 Neutrophils are a primary infiltrating cell type into the lungs following initiation of a Th1 immune 2733 response. They are an important component of the innate immune response that play roles in 2734 microbial killing via extrusion of cytoplasmic “nets” composed of nucleic acids decorated with 2735 proteinases. Neutrophilic inflammation can be rapidly resolved via recognition and phagocytosis by 2736 macrophages. A variety of ENMs stimulate neutrophilic inflammation in the lungs of mice or rats, 2737 including SWCNTs, MWCNTs, TiO2, Ag, SiO2 and Ni (Mangun et al., 2006; Bonner et al., 2013; 2738 Gustafsson et al., 2011; Silva et al., 2015; Brown et al., 2013; Glista-Baker et al., 2014). While 2739 neutrophils participate in the inflammatory response to many types of ENMs, there is little evidence 2740 that ENMs are engulfed by neutrophils. However, Ni nanoparticles are avidly engulfed by 2741 neutrophils, as well as macrophages, lavaged from the lungs of mice 1 day after oropharyngeal 2742 aspiration exposure (Glista-Baker et al., 2014). In addition, inhalation exposure to CeO2 nanoparticles 2743 results in phagocytosis by both neutrophils and macrophages (Srinivas et al., 2011). These 2744 demonstrate that some types of ENMs directly interact with neutrophils and thus could impair the 2745 normal immune functions of these inflammatory cells. 2746 2747 Eosinophils are a primary infiltrating cell type into the lungs following initiation of a Th2 immune 2748 response and are the hallmark inflammatory cell of allergic lung diseases, particularly asthma. 2749 Allergen challenge is commonly used in rodents to elicit an allergic eosinophilic lung inflammatory

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2750 response. Some ENMs have also been reported to cause eosinophilic inflammation in the lungs of 2751 mice, including certain types of MWCNT (rod-like), and certain forms of TiO2 (nanobelts). Like 2752 neutrophils, there is little information showing that ENMs are taken up by eosinophils. However, 2753 mice challenged with fungal spores and then exposed to Au nanoparticles showed eosinophilic lung 2754 inflammation and approximately 14% of eosinophils contained Au nanoparticles within intracellular 2755 vesicles (Geiser et al., 2014). 2756 2757 2758 Table 1. Type of lung inflammatory cell infiltration observed in rodents following exposure to 2759 selected ENMs. 2760 ENM Shape or Species/Strain Inflammatory Cell Reference Characteristic Type MWCNT Flexible tube Mouse, Rat Neutrophilic Rydman et al, 2014; Bonner et al., 2013 MWCNT Rod-like tube Mouse Eosinophilic Rydman et al., 2014 SWCNT flexible tube Mouse, Rat Neutrophilic Shvedova et al., 2005; Mangum et al., 2006 TiO2 Sphere Mouse, Rat Neutrophilic Gustafsson et al., 2011 Bonner et al., 20013 TiO2 Rigid nanobelt Mouse, Rat Neutrophilic/ Bonner et al., 2013 Eosinophilic Ag Sphere Mouse Neutrophilic Silva et al., 2015

SiO2 Sphere Mouse Neutrophilic Brown et al., 2013 2761 2762 Mast cells are also important in the allergic immune response (Brown et al., 2008). Upon activation, 2763 mast release histamine that alters vascular permeability. Mast cells have been shown to participate 2764 in ENM-induced immune and inflammatory responses after pulmonary exposure to influence 2765 physiological responses in other organ systems such as the heart. For example, mast cells play a 2766 central role in the activation of the IL-33/ST2 axis to mediate adverse pulmonary and cardiovascular 2767 responses to MWCNTs (Katwa et al., 2012). Mast cells also contribute to altered vascular reactivity 2768 and ischemia-reperfusion injury following CeO2 nanoparticle instillation (Wingard et al., 2011). These 2769 findings demonstrated that CeO₂ nanoparticles activate mast cells contributing to pulmonary

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2770 inflammation, impairment of vascular relaxation and exacerbation of myocardial 2771 ischemia/reperfusion injury. In addition, anatase and rutile TiO2 nanoparticles induces histamine 2772 secretion in mast cells (Chen et al., 2012). 2773 2774 2775 4.10.3.3.3 Dendritic Cells 2776 Dendritic cells (DCs) are important immune initiators that serve to capture and present allergens to 2777 naïve T cells, thereby driving T cell polarization (Gill 2010). DCs mediate the first step in the allergic 2778 immune response through uptake and presentation of allergen to Naïve T cells. Macrophages and B 2779 lymphocytes may also serve as antigen-presenting cells. DCs acquire and ‘sample’ allergens 2780 deposited on the airway epithelium. Chemokines released by the airway epithelium upon allergen 2781 stimulation attract DCs. Following recognition and uptake, dendritic cells migrate to the T-cell-rich 2782 area of draining lymph nodes, display an array of antigen-derived peptides on the surface of major 2783 histocompatibility complex (MHC) molecules, and acquire the cellular specialization to select and 2784 activate naïve antigen-specific T cells. Allergen targeting to the dendritic cells occurs via membrane- 2785 bound IgE. Dendritic cells interact with many cell types, including mast cells, epithelial cells and 2786 fibroblasts. Mediators released by these cells can activate the dendritic cells so that it is induced to 2787 mature and attract memory Th2 cells through release of Th2-selective chemokines. Mature effector 2788 Th2 cells play a central role in asthma pathogenesis by releasing cytokines (e.g., IL-13) that stimulate 2789 eosinophil recruitment, smooth muscle cell cytokine and chemokine production, and goblet cell 2790 hyperplasia. 2791 Understanding the effects of ENMs on the function and phenotype of dendritic cells (DCs) is an 2792 important area of study. The exacerbation of allergen-induced airway disease discussed above could 2793 be partly through the inappropriate activation of antigen-presenting DCs. Both ENMs and diesel 2794 pollutant nanoparticles have been shown to activate DCs (Bezemer et al., 2011). Other work has 2795 shown that ZnO nanoparticles cause DC death at relatively low concentrations, whereas these ENMs 2796 had no effect on peripheral blood mononuclear cells (Andersson-Willman et al., 2012). Interestingly, 2797 CNTs have been reported to inhibit the differentiation of peripheral blood monocytes into DCs 2798 (Laverny et al., 2012). The consequence of this would most likely be depletion of antigen-presenting 2799 DCs in the lung, which would presumably reduce immune recognition and subsequent T cell 2800 differentiation in draining lymph nodes. 2801 2802 4.10.3.3.4 Lymphocytes 2803 The type of immune and inflammatory response elicited by inhaled ENMs depends largely on T- 2804 lymphocyte programming. As discussed above, this occurs via presentation of the ENM to naïve T

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2805 cells in the draining lymph nodes and is regulated by the cytokine microenvironment. Four different 2806 T-lymphocyte populations can result as a consequence of programming; Th1, Th2, Th17, and Treg. 2807 Th1 lymphocyte differentiation is driven by IL-12. Th1 cells produce IFN- nd TNF- 2808 cell-mediated immunity and IgG2 production. IFN- 2809 phenotype, which in turn secrete IL- - 2810 orchestrates downstream CXC chemokines (in mice) or IL-8 (in humans) to promote neutrophilic 2811 inflammation. Many types of ENMs have been shown to cause a Th1 type immune response in the 2812 lungs of rodents with subsequent neutrophilic inflammation. For example, MWCNTs or TiO2 2813 nanoparticles primarily elicit neutrophilic inflammation in the lungs of mice or rats (Bonner et al., 2814 2013). However, in this study, the type of MWCNT used was tangled and flexible. As discussed below, 2815 more rigid MWCNT promote a Th2 immune response. 2816 Immunological research into the mechanisms of allergy has identified cytokine production by T- 2817 helper 2 (Th2) effector lymphocytes as being critical for orchestrating allergic inflammation rich in 2818 eosinophils. Upon recognition their cognate antigen, Th2 lymphocytes produce cytokines that 2819 regulate IgE synthesis, growth and activation of eosinophils and mast cells, and expression of 2820 endothelial cell adhesion molecules. The first step in the allergic immune response is the uptake and 2821 presentation of allergen by antigen-presenting cells called dendritic cells. Macrophages and B 2822 lymphocytes may also serve as antigen-presenting cells. Following recognition and uptake, dendritic 2823 cells migrate to the T-cell-rich area of draining lymph nodes, display an array of antigen-derived 2824 peptides on the surface of major histocompatibility complex (MHC) molecules, and acquire the 2825 cellular specialization to select and activate naïve antigen-specific T cells. Allergen targeting to the 2826 dendritic cells occurs via membrane-bound IgE. Dendritic cells interact with many cell types, 2827 including mast cells, epithelial cells and fibroblasts. Mediators released by these cells can activate the 2828 dendritic cells so that it is induced to mature and attract memory Th2 cells through release of Th2- 2829 selective chemokines. Mature effector Th2 cells play a central role in asthma pathogenesis by 2830 releasing cytokines (e.g., IL-13) that stimulate eosinophil recruitment, smooth muscle cell cytokine 2831 and chemokine production, and goblet cell hyperplasia. Certain types of ENMs stimulate a Th2 2832 immune response in the lungs. For example, while tangled MWCNT elicit a Th1 immune response 2833 with neutrophilic inflammation, rigid rod-like MWCNTs stimulate a Th2 immune response 2834 characterized eosinophilic inflammation (Rydman et al., 2014). Mice exposed to these MWCNT 2835 exhibited elevated levels of Th2 cytokines (e.g. IL-13) and mucous cell metaplasia of the airway 2836 epithelium. Interestingly, proteomic analysis of the secretome of monocyte-derived macrophages 2837 after exposure to tangled MWCNTs, rod-like MWCNTs, or crocidolite asbsestos revealed different 2838 protein profiles for tangled and rod-like MWCNTs, but similar profiles for rod-like MWCNT and 2839 asbestos (Palomaki et al., 2015).

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2840 2841 4.11.3.4 Animal models of immune-mediated lung disease 2842 4.11.3.4.1 Pulmonary Fibrosis 2843 Pulmonary fibrosis is an occupational hazard following exposure to many particles and fibers. Innate 2844 immune responses could play an important role in the progression or resolution of fibrosis. ENMs, 2845 because of their increased surface area to mass ratio and ROS-producing potential, could pose a 2846 significant risk for the development of fibrosis. Certain assumptions have been made with respect to 2847 ENM toxicity and expected disease outcome. For example, CNTs share some features with asbestos 2848 fibers, mainly with regard to their fiber-like shape and aspect (length to width) ratio. Asbestos fibers 2849 are a known cause of fibrosis and mesothelioma in humans. However, CNTs also have some uniquely 2850 different properties from asbestos, including nanoscale width and highly conformal structure. 2851 Therefore, some caution should be taken in making comparisons of fiber-like or tube-shaped ENMs 2852 with asbestos fibers with respect to fibrogenic potential. Nevertheless, pulmonary fibrosis is also a 2853 common pathologic feature observed in numerous rodent studies after exposure to CNTs (Li et al., 2854 2007; Shvedova et al., 2005; Porter et al., 2010). As mentioned above, aggregated CNTs tend to 2855 produce granulomatous lesions (Muller et al., 2005; Murray et al., 2012). In contrast, diffuse 2856 interstitial pulmonary fibrosis is associated with well-dispersed CNTs that are readily taken up by 2857 macrophages and cause greater growth factor (PDGF-AA, TGF-β1) and IL-1β production than non- 2858 dispersed CNTs (Wang et al., 2011). Other factors contribute to the fibrogenic potential of ENMs, 2859 including shape, composition, electrostatic charge, and ROS generating capacity. High aspect ratio 2860 ENMs impede clearance and structures longer than ten to fifteen micrometers (the approximate 2861 width of an alveolar macrophage) are difficult to clear from lung tissues via macrophage-mediated 2862 mechanisms. For fiber or tube-shaped ENMs, diameter is also a determinant of toxicity. For example, 2863 thinner (~10 nm diameter) multi-walled CNTs are more toxic in the lungs of mice than thicker (~70 2864 nm diameter) multi-walled CNTs (Fenoglio et al., 2012). Rigidity of ENMs is also a factor. For example, 2865 long single-walled CNT are flexible and when folded, single-walled CNTs are taken up by 2866 macrophages without causing frustrated phagocytosis or impeding macrophage clearance. The 2867 composition of ENMs must also be carefully considered. For example, metals used as catalysts in the 2868 manufacture of CNTs (e.g., nickel, cobalt, iron) are known to mediate pulmonary fibrosis in humans 2869 (Kelleher et al., 2000). For example, nickel is known to cause occupational asthma and contact 2870 dermatitis, whereas iron and cobalt cause interstitial pulmonary fibrosis in occupations related to 2871 mining and metallurgy. 2872 The emergence of the myofibroblast, a collagen synthetic mesenchymal cell, is a key step in the 2873 progression of lung fibrosis. A variety of growth factors, cytokines, and chemokines that stimulate 2874 myofibroblast differentiation, growth, migration, and extracellular matrix production are induced in

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2875 the lungs of rats or mice after exposure to ENMs (Bonner 2010b). For example, single- or multi- 2876 walled CNTs delivered to the lung by intratracheal instillation in rats or inhalation in mice increase 2877 mRNA and protein levels of platelet-derived growth factor (PDGF) (Ryman-Rasmussen et al., 2009b; 2878 Cesta et al., 2010). PDGF stimulates the replication, chemotaxis, and survival of lung mesenchymal 2879 cells (fibroblasts, myofibroblasts, and smooth muscle cells) to promote lung fibrogenesis (Bonner 2880 2004). CNTs delivered to the lungs of mice increase levels of TGF- 2881 production by fibroblasts and myofibroblasts (Shvedova et al., 2005; Wang et al., 2011). In addition 2882 to TGF- 2883 of OPN are increased in the lungs of rats exposed to SWCNTs or MWCNTs (Mangum et al., 2004; 2884 Taylor et al., 2014). Alveolar macrophages, as well as airway epithelial cells and fibroblasts, produce 2885 PDGF, TGF- y CNT exposure and drive the 2886 inflammatory response in the lung. CXCL8 (IL-8), a potent neutrophil chemoattractant, is produced by 2887 a human bronchial epithelial cell line in vitro after exposure to multi-walled CNTs (Hirano et al., 2888 2010). CCL2, also known as monocyte chemoattractant protein-1 (MCP-1), is produced by 2889 macrophages and airway epithelial cells and is increased in the bronchoalveolar lavage fluid of mice 2890 after CNT inhalation exposure. In general, a complex interaction of cytokines, chemokines and 2891 growth factors contributes to the progression of pulmonary fibrosis. 2892 2893 4.10.3.4.2 Asthma 2894 Asthma features a chronic airway remodeling response that is characterized by: (1) eosinophilic 2895 inflammation, (2) airway smooth muscle thickening, (3) mucous cell hyperplasia and mucus 2896 hypersecretion, and (4) subepithelial fibrosis. Allergic diseases, including asthma, are thought to 2897 result from a dysregulated immune response to commonly encountered antigens in genetically 2898 predisposed individuals. Immunological research into the mechanisms of allergy has identified 2899 cytokine production by T-helper 2 (Th2) effector lymphocytes as being critical for orchestrating 2900 allergic inflammation rich in eosinophils. Upon recognition their cognate antigen, Th2 lymphocytes 2901 produce cytokines that regulate IgE synthesis, growth and activation of eosinophils and mast cells, 2902 and expression of endothelial cell adhesion molecules. Common allergens that cause asthma in the 2903 human population include house dust mite and cockroach antigens. In addition to allergens that 2904 directly cause Th2 cell programming and allergic disease, a variety of environmental agents (e.g., 2905 mold, ozone, pesticides) exacerbate pre-existing asthma. There is evidence that some ENMs are 2906 capable of either directly acting as allergens or exacerbating pre-existing allergic lung disease. 2907 2908 4.10.3.4.3 Malignant Disease.

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2909 In addition to airway and interstitial lung diseases, the pleural mesothelial lining surrounding the 2910 lungs is a potentially important site of toxicity for certain ENMs. Of particular concern are high aspect 2911 ratio ENMs such as CNTs, nanofibers, and nanowires that have asbestos-like shape and therefore 2912 could be persistent in lung tissue. ENMs contained within macrophage cross the pleural lining via the 2913 lymphatic drainage and thereby interact with the mesothelial lining of the pleura. Here, the durable 2914 nature of CNTs, nanofibers, or nanowires coupled with fiber-like shape and reactivity (i.e., ROS- 2915 generating capacity) could result in immune reactions, pleural inflammation, or DNA damage to 2916 mesothelial cells. While unknown at the present time, it has been speculated that such high-aspect 2917 ratio ENMs could have asbestos-like behavior and long term effects immune or inflammatory effects 2918 that could lead to tumor formation (i.e., mesothelioma). Intraperitoneal injection of long multi- 2919 walled CNTs in mice induces inflammation and granuloma formation on the mesothelial surface of 2920 the peritoneum and a similar strategy with CNTs using mice deficient in the tumor suppressor p53 2921 showed mesothelioma formation in the abdominal cavity after injection of CNTs (Poland et al., 2009; 2922 Takagi et al., 2008). CNTs also activate p53 in mouse embryonic stem cells in vitro (Zhu et al., 2007). 2923 Multi-walled CNTs delivered to the lungs by inhalation or aspiration accumulate in subpleural tissue 2924 and some tubes penetrate the pleural lining (Ryman-Rasmussen et al., 2009a; Porter et al., 2010; 2925 Mercer et al., 2010). The inhalation of multi-walled CNTs in mice has been shown to produce pro- 2926 inflammatory lesions on the pleural surface that have been referred to as mononuclear cell 2927 aggregates (Ryman-Rasmussen et al., 2009a). These same mice had elevated levels of PDGF and CCL2 2928 in lavage fluid. Interestingly, PDGF and Ni nanoparticles synergistically increase CCL2 production by 2929 cultured rat mesothelial cells (Glista et al., 2012). The mechanism involves Ni enhancement of PDGF- 2930 induced CCL2 expression by prolonged MAP kinase activation in mesothelial cells. The accumulation 2931 of mononuclear cell aggregates at the pleural surface after exposure to CNTs, or presumably Ni 2932 nanoparticles, could be mediated by PDGF secreted by activated macrophages. PDGF, in turn, 2933 stimulates the production of CCL2 that serves to recruit mononuclear cells to the pleural surface. 2934 Moreover CCL2 is produced by pleural mesothelial cells and is a candidate chemokine that could 2935 participate in the formation of mononuclear cell aggregates observed at the pleura of mice after 2936 inhalation of multi-walled CNTs (Visser et al., 1998). The issue of whether ENMs are capable of 2937 causing immune, inflammatory, or carcinogenic effects at the pleura in humans remains a key topic 2938 of research and will have important implications for the future use and development of ENMs for a 2939 variety of applications. 2940 2941 4.10.3.4.4. Animal models of disease susceptibility 2942 It is likely that the most susceptible individuals to adverse immune effects of ENMs will be those with 2943 pre-existing respiratory disease such as asthma or chronic bronchitis. ENMs will most likely have the

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2944 most profound adverse health effects on individuals with pre-existing respiratory diseases such as 2945 asthma, bronchitis, or COPD (Bonner 2011; Inoue and Takano, 2011). For example, multi-walled CNTs 2946 exacerbate allergic airway inflammation in mice caused by ovalbumin sensitization as determined by 2947 amplified lung levels of Th2 cytokines and chemokines, as well as serum IgE levels, compared with 2948 allergen alone (Inoue et al., 2009). Single-walled CNTs have also been reported to exacerbate allergic 2949 airway inflammation in mice via enhanced activation of Th immunity and increased oxidative stress 2950 (Inoue 2010; Nygaard et al., 2009). CNTs have been shown to exacerbate airway fibrosis in mice that 2951 were first pre-challenged with ovalbumin, although in this study lung levels of allergen-induced IL-13, 2952 a principal Th2 cytokine, were reduced by multi-walled CNT exposure compared to allergen alone 2953 (Ryman-Rasmussen et al., 2009b). Repeated exposure to multi-walled CNTs has also been shown to 2954 induce Th2 allergic responses in the absence of any allergen pre-exposure (Park et al., 2009). 2955 2956 Table 2. Effect of selected ENMs on allergen-induced lung inflammation. ENM Allergen Species/Strain Effect Reference MWCNT Ovalbumin Mouse Exacerbation of Ryman-Rasmussen et airway fibrosis; al., 2009; Inoue et al., enhanced Th2 2009 cytokines and IgE MWCNT House Dust Mouse Exacerbation of Shipkowski et al., Mite airway fibrosis 2015 SWCNT Mouse Nygaard et al., 2009 TiO2 Diisocyanate Mouse Enhanced airway Hussain et al., 2011 hyperresponsiveness TiO2 Ovalbumin Mouse Suppression of Rossi et al., 2010 allergen-induced airway inflammation Au Diisocyanate Mouse Enhanced airway Hussain et al., 2011 hyperresponsiveness

SiO2 Ovalbumin Mouse Adjuvant effect; Brandenburger et al., exacerbation of 2013 airway inflammation 2957 2958 Other ENMs also exacerbate allergic inflammation in mice. TiO2 or Au nanoparticles enhance airway 2959 hyper-responsiveness in a mouse model of diisocyanate-induced airway inflammation and increase

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2960 numbers of inflammatory cells (Hussain et al., 2011). However, the majority of inflammatory cells in 2961 that study were neutrophils, which suggests a shift from the classic Th2 response to one that 2962 primarily features eosinophils. While these studies suggest that individuals with allergic asthma are 2963 susceptible to lung and airway disease caused by ENMs exposure, it remains unknown whether 2964 ENMs will cause or exacerbate asthma in humans. While there is a significant body of evidence in 2965 rodents suggesting that ENMs (e.g., CNTs) would be a hazard to individuals with asthma through 2966 exacerbation of Th2 inflammation, some studies with ENMs show suppression of allergen-induced 2967 airway disease in mice. For example, it has been shown that TiO2 nanoparticles cause neutrophilic 2968 inflammation in healthy mice, but suppresses allergic airway inflammation in mice sensitized with 2969 ovalbumin allergen (Rossi et al., 2010). Therefore, there is evidence that ENMs can cause 2970 immunostimulation or immunosuppression of local immune responses in the lung. 2971 2972 When nanoparticles (nano TiO2 or Carbon Black) were co-administered with an antigen like 2973 ovalbumin in mice an enhancement of pulmonary inflammation and allergic airway sensitization was 2974 observed (De Haar et al., 2006). Also particulate pollutants, e.g. diesel exhaust particles (DEP), have 2975 been shown to increase airway response after intratracheal instillation with an antigen and are 2976 considered to contribute to allergic diseases (Ichinose et al., 1997). Single wall and multiwall carbon 2977 nanotubes (swCNT, mwCNT, respectively) and ultrafine (nano) CB were demonstrated to promote 2978 allergic responses in the murine OVA allergic airway model. Following a subcutaneous and/or 2979 intranasal co-administration of OVA antigen with the various carbon nanomaterials, after three 2980 weeks an intranasal challenge was given to the animals. Serum levels of OVA specific IgE and cell 2981 numbers in bronchioalveolar lavage fluids (BALF) were strongly increased by the carbon 2982 nanomaterials (Nygaard et al., 2009). An IgE adjuvant effect was demonstrated for polystyrene 2983 particles (PSP) size 100 nm, in a mouse OVA IgE model (Granum et al., 2001). Irrespective of the 2984 route of administration, intranasal (i.n.) instillation, intratracheal (i.t.) instillation or intraperitoneal 2985 (i.p.) injection, and immunization protocol PSP, particles in combination with OVA elicited increased 2986 levels of both allergen specific and total IgE in mice. Promotion of allergic airway inflammation in the

2987 OVA model in mice was demonstrated for TiO2, DEP, CB, mwCNT, swCNT, SiO2, and PSP 2988 nanomaterials (Ichinose et al., 1997; Granum et al., 2001; De Haar et al., 2006; Inoue et al., 2009;

2989 Nygaard et al., 2009; Brandenberger et al., 2013; Marzaioli et al., 2014). For silica (SiO2) 2990 nanoparticles the enhanced reactions in the OVA airway allergic model could be attenuated by amino 2991 and phosphate surface modifications of the SiO2 NPs (Marzaioli et al., 2014). In an airway asthmatic 2992 hyper-reactivity model with toluene diisocyanate (TDI) immunization pulmonary pretreatment with

2993 TiO2 (15 nm) and Au (40 nm) nanoparticles one day before oropharyngeal challenge with TDI resulted 2994 in a three to fivefold increase in lung inflammatory cells (neutrophils and macrophages), and

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

2995 histologically in an increased edema, epithelial damage and inflammation (Hussain et al., 2011). So, 2996 for allergic lung diseases both particulate airborn pollutants like DEP and manufactured 2997 nanomaterials (MN) can exacerbate existing allergic disease in the lung and act as adjuvant for the 2998 induction of such allergies. Both workers and consumers with lung problems are at risk for adverse 2999 effects after airway exposure to (nano)particles. 3000 3001 The effects of ENMs in the lung could also be exacerbated by pre-existing bacterial or viral infection. 3002 Bacterial LPS is a potent pro-inflammatory agent and has been implicated in a number of 3003 occupational and environmental lung diseases in humans, including bronchitis, chronic obstructive 3004 pulmonary disease (COPD), and asthma. CNTs, either single- or multi-walled, have been reported to 3005 increase the severity of LPS-induced lung inflammation, pulmonary vascular permeability, and 3006 production of pro-inflammatory cytokines in the lungs of mice (Inoue et al., 2008). Moreover, 3007 pulmonary fibrosis induced by multi-walled CNTs is increased by LPS pre-exposure in rats and CNT- 3008 induced production of PDGF by rat alveolar macrophages and lung epithelial cells is enhanced by LPS 3009 pre-exposure (Cesta et al., 2010). These studies provide evidence that LPS-induced lung inflammation 3010 is a susceptibility factor that increases the severity of fibroproliferative lung disease caused by CNT 3011 exposure. 3012 3013 In addition to environmental susceptibility factors, a variety of genes play important roles in 3014 determining susceptibility to CNTs and other engineered nanomaterials. Transgenic mouse models 3015 are valuable towards elucidating the role of specific immune mediators (e.g., transcription factors, 3016 enzymes). For example, the STAT1 transcription factors plays a central role Th1 immunity by 3017 mediating many of the biological effects of IFNs. STAT1 knock-out mice are susceptible to 3018 exacerbation of allergen-induced airway fibrosis by MWCNTs (Thompson et al., 2015). The 3019 transcription factor T-bet maintains Th1 immunity and knock-out of T-bet in mice results in a 3020 spontaneous shift towards a Th2 phenotype. T-bet knock-out mice display enhanced airway 3021 remodeling (e.g., mucous cell metaplasia) and exaggerated levels of CCL2 after exposure to Ni 3022 nanoparticles (Glista-Baker et al., 2014). In addition to transgenic models of transcription factor 3023 deficiency, knock-out mice lacking specific enzymes reveal clues about their role in immune reactions 3024 to ENMs. Mice deficient in the COX-2 enzyme are susceptible to exacerbation of allergen-induced 3025 airway remodeling by MWCNTs (Sayers et al., 2013). While COX-2 mediates pro-inflammatory 3026 effects, it has also been implicated in the pathogenesis of asthma and pulmonary fibrosis (Bonner 3027 2010). Mice deficient in myeloperoxidase (MPO) have fewer MWCNTs in their lungs compared to 3028 wild type counterparts, suggesting that MPO mediates the degradation of some CNTs in the lungs 3029 [Shvedova et al., 2012].

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3030 3031 4.11.4 Exposure to nanomaterials via the gastrointestinal tract 3032 4.11.4.1 Overview of GIT immunology 3033 The GIT derives from the endoderm during the embryonic development and is responsible for the 3034 breakdown and absortion of food and liquids which starts by the mechanical digestion but parallely, 3035 the chemical digestion started by enzimes also takes palce in the mouth. Esofagus, stomach, small 3036 intestine, which includes duodenum, jejunum, ileum, and large intestine, rectum, and the anus can 3037 be considered as the main GIT components of accessory organs such as salivary glands, tongue, liver, 3038 gallbladder, and pancreas are also closely involved in the GIT functions. 3039 Stomach. The cardia, fundus, body, pyloric antrum, pyloric canal and pyloric sphincter are the main 3040 structures in the stomach and mucosa, submucosa, muscularis externa and serosa are the layers of 3041 this tissue. In the mucosa, the epithelium is composed by parietal cells, which produced the 3042 hydrochloric acid, chief cells, which produce the pepsinogen, and enteroendocrine G cells, which 3043 mainly produce gastrin. 3044 3045 Small and large intestine are mainly dedicated to the nutrients absortion and water through the 3046 crypts in which can be distinguished steam cells, M cells, goblets cells and epithelial cells and below 3047 the crypts, the muscularis muscularis, submucosa, muscularis externa and serosa (peritoneum )are 3048 located. In addition, the GIT contains non-lymphoid and secondary lymphoid tissues and the gut 3049 associated lymphoid tussue (GALT) is composed by immune cells which are located in the epithelium 3050 and lamina propia while mesenteric lymph nodes, Peyer´s patches are part of the seconday lymphoid 3051 tissues. One of the most important differences between small and large intestine is the villi which 3052 enhance surface for absortion. 3053 Cell types in the intestine. Stem cells of intestine are pluripotent cells located in the base of the 3054 crypts and well-kown for the capabilities of proliferation and differentiation. Once the cells have 3055 been differentiated, they migrate up to the crypt with exception of Paneth cells, which are only 3056 present in the small intestine and absent in the large. Then, some of the differentiated epithelial cells 3057 have endocrine functions through the secretion of differente molecules. For instance, Paneth and 3058 goblet cells produce antimicrobial proteins (AMPs) which are released into the intestinal lumen. 3059 Specifically, Paneth cells produce alpha-defensin, lysozyme and cathelicidins while goblet cells are 3060 well specialized in mucin 2 production as a protection barrier. Alpha-defensins are expressed 3061 prenataly and increase after birth and the fistly described functions were related to antimicrobial 3062 properties against pathogenic organisms but without effects on commensal bacteria, however, lately, 3063 it has been found that alpha-defensins also have a role in the inflammatory cytokine release. 3064 Enterocytes also produces AMPs such as C-type lectin regenerating islet-

100

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3065 also fundamental for the transcytosis of immunoglobulins from the lamina propria to the intestinal 3066 lumen. The microfold cells, also called M cells, are responsible for antigen uptake and facilitate the 3067 antigen specific immune response since they have basolateral invaginations to harbor immune cells. 3068 In the Peyer’s Patches, there are three zones that can be identified as follicular area, intrerfolliclar 3069 area and the follicle-associated epithelium. Here, the proliferative lymphocytes B, follicular dendritic 3070 cells and macrophages are located, however, B and T cells and dendritic cells are sorounding the 3071 Patches, which are connected to lymphatic and endothelial vessels (Figure XXXXX). 3072 The similitudes between small and large intestine are remarkable, however, the villi is absent in the 3073 large intestine, there is no Peyer’s Patches nor Paneth cells but solitary lymph nodes, among others. 3074 In addition, the GIT is colonized by a compley communitiy of Achaea, fungi, protozoa and viruses and 3075 the Frimicutes and Bacteroidetes have been identified as some of the main species. Collectively, this 3076 community is called microbiota and it is estimated in approximately 1x1014 organism in the gut. 3077 3078

intes.xps 3079 3080 3081 Figure XXXXX. Schematic representation of immune cells in the small intestine. 3082 3083 3084 3085 In a general view, the GIT protects itself against pathogens by the innate immune system releasing 3086 different molecules, which mainly prevent the entrance of the organism into deeper layers. However, 3087 if pathogens overcome this strategy, adaptive immune system responds through B and T cells. As it 3088 was described above, GIT posseses an immune system and specifically, if nanoparticles are consumed 3089 in the diet, they could end up internalized by the immune system. Until now, there are limited 3090 studies showing specific immunotoxicity in oral cavititiy, esophagus and stomach and more has been 3091 investigated in the small and large intestine, but not entirely related to immune system. However, 3092 the evidence of accumulation in some of those tissues opens the question of their possible 3093 immunotoxicity. For instance, titanium dioxide NPs can be accumulated in oral cavitity while silver 3094 and zinc oxide nanoparticles can be accumulated in the stomach of rats (Hendrickson et al., 2016; Ko 3095 et al., 2015), however, the amount of NPs that can be absorbed in the oral cavitity, esophagus and 3096 stomach seems to be less than the amount in the small and large intestine, and at least for titanium 3097 dioxide, the higher absortion could be estimated in 4% for the large intestine and much less for the

101

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3098 rest of the GIT. Indeed, there is evidence of accumulation of particles in Peyer´s Patches, which could 3099 affect the function of T and B cells. 3100 Once NPs have reached the intestine, there are several routes of translocation, including paracelular 3101 transport through tight juntions, transcytosis and endocytosis (includes processes mediated by 3102 clathrin, caveolae, pinocytosis and phagocytosis) by epithelial cells, however, since the normal 3103 function of M cells is the transport of molecules and microorganism by transcytosis, more attention 3104 should be addressed on this cell type as a possible route of transport of NPs and also possible lost of 3105 antigen presenting functions. 3106 Probably the most insoluble NPs could still be observed as NPs and agglomerates/aggregates, while 3107 some others, such as zinc oxide, cannot. Some soluble, organic or functionalized NPs could suffer 3108 enzymatic degradation (eg, amylases, lipases, proteases) but also the different environments could 3109 affecte them, for instance, low pH in the stomach and alkaline pH in the duodene, high mucus 3110 content in the small and large intesitine and by contrast, internalization of coated NPs (eg chitosan 3111 promotes the trascytosis of NPs in the intestine) could be favoured. 3112 3113 4.11.4.2 Importance of oral exposure to nanomaterials 3114 The rising usage of nanomaterials in foods, food packaging and pharmaceuticals has brought concern 3115 about their effects after oral exposure but the fact that part of the inhaled nanoparticles can be 3116 cleared by mucus and reach gastrointestinal tract (GIT) enhanced that concern, especially because 3117 inhalatory exposure to nanomaterials have proved toxic effects. Consequently, nanomaterials used 3118 in food and pharmaceutical-related industry (Johnson et al., 2015; Omlor et al., 2015) are highlighted 3119 the attention of toxicologist. For instance, the food industry has took advantage the migration of NPs 3120 from the package into the food, which was a problem initialy not only for safety reasons but also for 3121 modifications in the organoleptic properties of the food. Some nanocomposites are being used as 3122 monitors of the food stability and quality but also as antimicrobial agents (Honarvar et al., 2016). 3123 Nanocomposites have been also suggested as usefull tools for drug or calcium delivery and taking 3124 together all the possible oral applications, it seems that oral exposure will be increased. 3125 Nowadays some nanomaterials used as food additives are intentionally added to foods mainly to 3126 preserve aliments or to improve organoleptic properties, and letter E identifies them in the European 3127 Union. Some of the currently food additives, such as titanium dioxide (European Food Authority 3128 Safety, 2004), designated as E171, has been used since 1966. However at that time it was produced 3129 as microparticles but nowadays it is demonstrated that E171 also contains a nanofraction with a size 3130 below 100 nm and percentages of 10 – 36 have been reported (Weir et al., 2012; Peters et al, 2014). 3131 Silicon dioxide, designated as E551 and silver particles, designated as E174 are also food additives 3132 that are being used in nanometric scale (Jovanović et al., 2015; Peters et al., 2012; Gaillet et al., 2015;

102

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3133 Shi et al., 2014). However, there are some other additives including calcium carbonate (E170) and 3134 calcium silicate (E552), that are been using as nanoparticles in food additives but are less 3135 investigated. 3136 In addition, some other nanoparticles, such as zinc oxide are not considered as food additives but are 3137 intended to be present in plastics to contain foodstuff for long-term storage. The usage of 3138 nanoparticles as food additives or in the food packaging highlighted the possible adverse effects on 3139 the human health based on the research developed in the last 2 decades on inhalatory exposure to 3140 nanoparticles. 3141 One of the first studies related to the effects of particles used in food additives was performed by 3142 members of the American Cynamid Company in 1963. In this study rats received a diet containing 3143 100,000 mg per kg of diet, which is 10% of food additive, during 30 days and only a low accumulation 3144 of titanium in muscle was reported (West and Wyzan, 1963). During the following decades, in vitro 3145 and in vivo experimental models have been used to determine if oral exposure to nanoparticles 3146 might have an impact on the alimentary tract and lately, nanoparticles for medical applications 3147 related to oral exposure and inhaled nanoparticles that could reach GIT as a consequence of 3148 clearance from the respiratory system (Balasubramanian et al., 2013) have gain relevance in terms of 3149 alimentary tract effects. Taking together, the necessity of evaluation of nanoparticles effects in 3150 alimentary tract has been strengthening by the evidence that inhaled nanoparticles can reach GIT 3151 and translocate to the circulation reaching organs including liver, kidney, brain and cardiovascular 3152 system (Reviewed by Kermanizadeh et al., 2015). 3153 In spite of the different types of nanoparticles, not all of them are being under exhaustive research 3154 and those used as food additives have had special concern followed by those used for food packing 3155 and those that are potential oral drugs. 3156 In this section, effects related to exposure to carbon-based, metal-based and organic nanoparticles 3157 on alimentary tract are discussed according to in vivo, in vitro data, and human studies. 3158 There are limitations that need to be considered when assessing for toxicity of nanomaterials upon 3159 oral exposure. One of the first limitations of comparing nanomaterials effects, independently of the 3160 route of exposure, is the lack of standardized methods for characterization of nanomaterials. 3161 Secondly, preparation of nanomaterial stock and the dispersion media of nanomaterials exert a 3162 strong influence on biological effects and there are still no unified criteria to prepare a stock solution 3163 and to select a proper dispersion medium for each study. Thirdly, the dose and concentration can be 3164 higher than reached under realistic human exposure, however, the lack of quantification of 3165 nanomaterials orally ingested complicates the design of experimental studies. Fourthly, the genetic 3166 variations associated to strain of experimental animals and the age of animals used have been poorly 3167 addressed and the passages number for cell culture experiments is rarely reported. Fifthly, a few

103

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3168 studies take into consideration the interaction of nanomaterials with saliva, gastric fluids and 3169 intestinal fluids, which might modify the nanomaterials but also, a few studies addressed the effect 3170 of food/pharmaceutical process on nanomaterial properties and their impact on pharmacokinetics, 3171 which would simulate a more realistic condition to human exposure. 3172 3173 The current and potential usage of different types of nanomaterials by oral intake highlighted the 3174 attention of studies of potential adverse GIT effects and its possible immunotoxicity to humans. The 3175 lack of consensus in the experimental design for in vivo and in vitro studies complicates the 3176 interpretation of the available information, specifically by the overload used for in vivo studies and 3177 the high concentrations used for the in vitro studies. However, samples taken from human biopsies 3178 demonstrate that nanoparticles can reach important tissues involved in the immune response, which 3179 could have a detrimental impact on human health, specifically under prolonged exposure or/and 3180 when pre-existent diseases are present in adults or children. Diseases related to GIT could be 3181 worsened, since particles such as titanium dioxide and silicon dioxide can be deposited in GIT tissues. 3182 The adverse nanoparticles effects on immune system components related to GIT are not deeply 3183 evaluated. For instance, macrophages from oral administration could have an impairment in the 3184 function against further phagocytosis or nanomaterial internalization could inhibit release of cytokine 3185 to have a proper immune response. The effect of a second hit of nanomaterials or viruses/bacteria 3186 might be investigated since antigen presentation could also have deterioration, because damaged 3187 macrophages are unable to accomplish their physiological function. 3188 3189 Figure on immune cells in the GI/tract 3190 3191 Epidemiological information of consumption of nanomaterials would contribute to understand better 3192 the current situation of human alimentary tract exposure and would provide information for 3193 improved design of research, as contribution of estimation of ingested nanomaterials has done by 3194 Weir and colleagues (2012), in which document was provided a real estimation of 1 to 2 mg/kg of 3195 titanium dioxide consumption in children and less than 1 mg/kg in adults. 3196 3197 4.11.4.2 Results from in vivo studies 3198 4.11.4.2.1 Carbon-based nanomaterials 3199 Evaluation of some carbon-based nanoparticles effects on GIT is still limited and there is some data 3200 available using carbon black particles dosed at 0.064 and 0.64 mg/kg in a single and ten day 3201 intragastric administration in Zucker rats with metabolic syndrome. In this work carbon black 3202 administration of 6.4 mg/kg induced lipid load in the liver associated to hepatic steatosis (Vesterdal

104

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3203 et al., 2014). Single-walled carbon nanotubes sized in 0.9 to 1.7 nm and C60 fullerenes sized in 0.7 3204 nm intragastrically administered in rats dosed at 0.064 and 0.64 mg/kg showed DNA damage in liver 3205 and lung tissues with no alterations in colon after 24 h of administration. In this study, DNA damage 3206 was not associated to impairment of DNA repair system (Folkmann et al., 2009). A high dose of 1000 3207 mg/kg SWCNT administered orally in mice showed no behavioral changes and brain, kidney, spleen, 3208 and liver and blood analysis showed no alterations except for elevated creatinine concentration in 3209 blood (Kolosnjaj-Tabi et al., 2010). SWCNT with a higher length of 300 nm can be eliminated through 3210 the kidney and bile ducts and be detected in urine and feces after 3 days of oral administration 3211 (Kolosnjaj-Tabi et al., 2010). However, SWCNT intraperitoneal administration (50-1000 mg/kg) tested 3212 at different dose and length of nanomaterial coalesced inside the body and agglomerates higher than 3213 10 µm induced granuloma formation, which was smaller than 1 mm of diameter in mice dosed at 50 3214 mg/Kg. Accumulation of SWCNT in the peritoneal cavity was observed after 150 days of 3215 intraperitoneal exposure (300 mg/kg) with no apparent alterations (Kolosnjaj-Tabi et al., 2010). 3216 Functionalized graphene oxide and 125I labeled was dosed at 4 mg/kg orally to Balb/c mice and after 3217 4 h, high levels of radioactivity were found in the stomach and intestine but not in liver, spleen, 3218 kidney, heart, lung skin, muscle, bone, brain and thyroid. In addition, retained radioactivity was 3219 detected only between 2 to 3% after 1 day of oral administration and undetectable after 7 days (Yang 3220 et al., 2013). 3221 3222 4.11.4.2.2.Metal-based nanomaterials 3223 One of the earliest in vivo studies in metal-based particles, showed that titanium dioxide (12.5 3224 mg/kg) of 475 nm administered to rats by oral gavage for 10 days was located in the granular areas 3225 of Peyer's patches, in mesentery nodes as well as in the connective tissues of the mesentery network 3226 but also in sinusoidal liver cells. In addition, colon was the tissue with highest uptake followed by the 3227 Peyer’s patches, liver, lungs, peritoneum and small intestine (Jani et al., 1994). In the following 3228 decades, several studies were performed in order to investigate the effects of oral exposure to 3229 microparticles but with the nanotechnology development, titanium dioxide nanoparticles non-food 3230 grade have been deeply investigated, however, less have been done with E171 nanoparticles, which 3231 are the food grade particles. In this regard, E171 particles administred intragastrically in BALB/c mice 3232 enhanced tumor formation in colon in a chemically colorectal cancer model induced by 3233 azoxymethane and dextran sodium sulphate. In this model, E171 particles were not able to induce 3234 tumor formation after 10 weeks of administration, but enhanced the tumor formation in distal and 3235 medium colon and the sole E171 particles administration decreased dramatically the content of 3236 goblet cells. Isolated tissue from E171 exposed mice was cultured (ex-vivo) and the cells showed 3237 some particles which remained internalized in the cells after one week in culture (Urrutia-Ortega et

105

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3238 al., 2016). The evidence of goblet cells decrease in the colon caused by food grade titanium dioxide in 3239 mice might affect the barrier of mucus against pathogens and also might affect the enzymatic 3240 content of enzymes in the mucus layer, which are required for physiological last abstortion process in 3241 the large intestine. 3242 Biodistribution of three different sizes of titanium dioxide nanoparticles (25, 80 and 155 nm) were 3243 investigated by oral administration in mice dosed at 5 mg/kg. This is a high dose for an in vivo CD-1 3244 mice model and it showed that nanoparticles sized in 25, 80 and 155 nm dispersed in 0.5% 3245 hydroxypropylmethylcellulose were deposited in the liver by 106, 3970 and 107 ng/g of tissue, 3246 respectively and 375, 440 and 170 ng/g of tissue in the kidney and undetectable in red cells (Wang et 3247 al., 2007). Some histological alterations were found in liver such as hydropic degeneration and the 3248 spotty necrosis of hepatocytes with an increased alanine aminotransferase serum levels, swelling in 3249 the renal glomerulus, and inflammation of the stomach of mice received titanium dioxide sized at 80 3250 nm after 2 weeks. No abnormal findings in the heart, lung, testicle/ovary, and spleen tissues were 3251 reported but fatty degeneration in the hippocampus of brain tissue was found. Titanium dioxide 3252 sized at 25 nm and 150 nm showed less alterations than 80 nm (Wang et al., 2007). Titanium dioxide 3253 nanoparticles sized at 25 nm were dispersed in distilled water and dosed at 1 and 2 mg/kg in rats. 3254 Effects on spleen were analyzed and compared between males and females and accumulation of 3255 titanium dioxide nanoparticles was found in the white pulp of the spleen and ovaries of female rats. 3256 Histological alterations at both doses in thyroid gland and adrenal medulla were found with no 3257 changes in thyroid function but an increase in testosterone levels in males rats were found (Tassinari 3258 et al., 2014). The impact of spleen accumulation on immune response was not evaluated. Later, a 3259 study of titanium dioxide sized at 40, 40-50, 120 and 5000 nm was administered to Sprague-Dawley 3260 male rats at 4.6 mg/kg dispersed in deionised water or 5% ovalbumin solution. Interestingly, in this 3261 work, a background of titanium was detected in blood (4-21 ng/mL) and in urine (12-46 ng/mL) of 3262 control rats fed with normal diet. In this study, it was found that cardboard toys contained 4.5 µg/g 3263 of titanium and these toys were provided for environmental enrichers (MacNicoll et al., 2015). The 3264 dispersion media had no effect on tissue uptake and samples taken from GIT of exposed rats to 120 3265 and 5000 nm had 86.7 ng/g of titanium (MacNicoll et al., 2015). Oral administration of silicon dioxide 3266 nanoparticles dosed at 30 or 100 mg/kg during six days to mice suffering from colitis, showed that 3267 labeled silicon dioxide nanoparticles with a hydrodynamic size of 101 nm were able to accumulate in 3268 inflamed tissue while healthy tissue exhibited limited particle accumulation (Moulari et al., 2008). 3269 This study opened the potential usage of silicon dioxide nanoparticles for the treatment of 3270 inflammatory bowel diseases when nanoparticles are bounded to a therapeutic agent (Mamaeva et 3271 al., 2011; Schmidt et al., 2013) but also open the question if accumulation on inflamed tissue could 3272 worsen the inflammatory process under chronic conditions.

106

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3273 More recent in vivo studies have shown that amorphous silicon dioxide with a diameter of 70 nm, 3274 300 nm or 1,000 nm orally administered at 2.5 mg/mouse/daily for 28 days induced no significant 3275 differences in body weight, liver and kidney function markers and no abnormalities in the liver, 3276 kidney, large intestine, small intestine, brain, lung, spleen, heart and stomach (Yoshida et al., 2014). 3277 Hematological analysis of white blood cells, lymphocytes, monocytes, granulocytes, platelets, and 3278 red blood cells remained also without alterations (Yoshida et al., 2014). Oral silver nanoparticles 3279 intake sized between 53 to 71 nm, administered at 30, 300 and 1000 mg/kg/day during 28 days in 3280 rats showed a slight hepatotoxicity effect with no alterations in food consumption, in body weight or 3281 in organ weight (Kim et al., 2008). 3282 Iron oxide (Fe2O3) nanoparticles have also been investigated, however, less information is available 3283 on alimentary tract since limited iron oxide is being designed for oral consumption. Indeed, it is 3284 estimated that oral exposure to human could be attributed to accidental ingestion or ingestion 3285 through pharmaceutical product. However, rats dosed at 500, 1000, 2000 mg/kg of 30 nm iron oxide 3286 of <5 μm (bulky) showed no genotoxicity in leucocytes after 6, 24, 48 and 72 h of exposure. 3287 Micronucleus test was also negative in peripheral blood cells. The biodistribution of iron was 3288 analyzed at 6, 24, 48 and 72 h after treatment in liver, spleen, kidney, heart, brain, bone marrow, 3289 urine and feces. Internalization of iron oxide nanoparticles was size and dose dependent in all the 3290 tissues, however, spleen, kidney and brain were able to uptake nanosized iron oxide while bulk iron 3291 oxide was not no detected. Bulk iron oxide was undetectable in blood and urine samples but dose 3292 dependent distribution was detected in bone marrow and urine for nanosized iron oxide and large 3293 excretion in feces were found in both types of iron oxide (Singh et al., 2013). A chronic study of oral 3294 exposure to ɣ-iron oxide coated with amino-dextran was carried out to investigate effect of these 3295 nanoparticles sized in12 nm on liver, spleen and duodenum accumulation. Iron oxide was 3296 administered in the diet at 6 to 24 mg/kg and the variation was associated to growth of chicken since 3297 these animals started the experiment after 1 day of birth and were used for a 14-day treatment, 3298 then, the ɣ-iron oxide consumption in the food increased during the study, however, no 3299 accumulation in liver, spleen or duodenum was found and iron was excreted mainly by feces 3300 (Chamorro et al., 2015). Studies to compare different type of metal-based nanoparticles showed that 3301 14-day administration of silicon dioxide (12 nm), silver (11 nm) and iron oxide (60 nm) nanoparticles 3302 dosed at 1959, 2061 and 2000 mg/kg, respectively had no alterations in clinical observations, body 3303 weight, hematological analysis, serum parameters and histological examinations in male and female 3304 Sprague-Dawley rats. However, a chronic 13-week study with these nanoparticles showed no toxicity 3305 of silicon dioxide and iron oxide but silver nanoparticles showed systemic toxicity (Yun et al., 2015). 3306 3307 4.11.4.2.2 Organic nanomaterials

107

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3308 In spite of extensive literature on dendrimers, limited information is still available about dendrimers 3309 effects on the GIT after oral administration. One of the earliest studies showed uptake of 3310 polyamidoamine (PAMAM) dendrimers with a size of 2.5 nm (6300 KDa) dosed at 14 mg/kg 3311 accumulate on blood, liver, spleen, kidney, small intestine and large intestine in 3%, 1.5%, 0.1%, 3312 0.5%, 15% and 5%, respectively, after 6 h of oral administration in rats (Sakthivel et al., 1999). A 3313 comparison of lymphoid and non-lymphoid tissue uptake showed that dendrimers are internalized 3314 1% after 3 h in the lymphoid tissue of small intestine while non-lymphoid small intestinal internalized 3315 3.8% after 3 h. Lymphoid tissue from large intestine showed no uptake while non-lymphoid tissue 3316 increased from 1% after 3 h and from to 3.8% after 12 h of oral exposure (Sakthivel et al., 1999). This 3317 study reveals an opposite effect on dendrimers uptake in non-lymphoid tissue from small and large 3318 intestine. Regarding to dendrimers branches, the number of hyperbranched topology radiating from 3319 the core (named “generation and designated by letter G”) has an impact on absorption and cellular 3320 effects; for instance, G3 dendrimers dosed at 3.4 mg/kg had higher accumulation in kidney after 48 h 3321 of administration while G5 and G7 dendrimers were internalized in the pancreas after 24 h of 3322 administration. However G3 dendrimers also accumulated in liver, kidney and spleen. In this study, 3323 G7 dendrimers showed highly urinary excretion (74%) during the first 4 h after injection (Roberts et 3324 al., 1996). In the following decades, toxicity of dendrimers according to number of generations and 3325 modification in the surfaces, have been performed showing that both exerts distinct biological 3326 effects. In this regard, different surfaces of G3.5, G4, G6.5 and G7 dendrimers dosed between 30 and 3327 500 mg/kg orally administered were tested and only G7-NH2 dendrimers dosed at 50 mg/kg showed 3328 hemobilia, G7-OH dendrimers dosed at 300 mg/Kg showed splenomegaly and G4-OH dendrimers 3329 dosed at 300 mg/kg showed increase in blood urea nitrogen, marker of renal funtion (Thiagarajan et 3330 al., 2013a). This study showed that maximum tolerated dose was obtained by smaller dendrimers 3331 (G3.5 and G4) and signs of toxicity were observed by G7-NH2 and G7-OH dendrimers administration. 3332 In addition, oral bioavailability in mice showed that radiolabeled G6.5-COOH (125I) dendrimers orally 3333 administered at 1 mg/kg did not show signs of toxicity after 10 days of exposure. These dendrimers 3334 remained intact under simulated GIT conditions and 70% of the dose was recovered mainly from the 3335 urine. The dose remaining in the GIT was located in stomach, small intestine, cecum, post-cecum, 3336 stool and carcass (Thiagarajan et al., 2013b). However, since dendrimers are being developed for 3337 drug delivery systems and imaging, variations of central dendrimers core, including ammonia, 3338 ethylenediamine and variation of peripheral groups have been designed to test anticancer drugs and 3339 imaging molecules. Biodistribution and pharmacokinetics of these types of dendrimers, which have 3340 drugs or molecules attached have been tested, however studies on some of these promising 3341 dendrimers (for drug delivery systems or imaging) does not include biodistribution or toxicity of 3342 dendrimers without bounded drugs. Some of the studies claim that dendrimers have undetectable

108

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3343 toxicity and good transepithelial absorption occurs in the intestine after oral administration and 3344 other routes of exposure (Wijagkanalan et al., 2011; Goldberg et al., 2011; Ke et al., 2008; Sadekar et 3345 al., 2012; Thiagarajan et al., 2012b; Chauhan, 2015). 3346 3347 3348 3349 4.11.4.3. Results from in vitro studies 3350 4.11.4.3.1 Carbon-based nanomaterials 3351 Carbon black, SWCNT and C60 fullerene nanoparticles were associated to lipid accumulation in 3352 human HepG2 hepatocytes after 3 h of nanoparticles exposure (0, 0.1, 10, 50 and 100 μg/mL) 3353 followed by 18 h of incubation with oleic/palmitic acid for 18 h, which was used to mimic some 3354 features of metabolic syndrome (Vesterdal et al., 2014). This study highlighted that exposure to 3355 carbon-based nanoparticles followed by these fatty acids could worsen the accumulation of lipids in 3356 hepatocytes. Functionalization of some carbon-based nanoparticles including SWCNT and fullerenes 3357 (from to 15.6 to 1000 µg/mL) showed very low cell viability decrease of Caco-2 cells but slight 3358 increase in the cytotoxicity of functionalized SWCNT after 72 h of exposure causing also disruption of 3359 the tight junction integrity after 24 h of cell culture exposure (Coyuco et al., 2011). In addition, 3360 SWCNT functionalized with carboxylic acid induced a decrease in transepithelial electrical resistance, 3361 reflecting weakening of the tight junctions that results in paracellular transport while SWCNT 3362 functionalized with poly(ethylene glycol) had no effect on this parameter (Coyuco et al., 2011). 3363 However, both type of functionalized SWCNT inhibited the P-gp efflux pump, which is responsible for 3364 avoid xenobiotic cell accumulation. This could be interpreted as a benefit of carbon-based 3365 functionalized nanoparticles to increase oral bioavailability of P-gp drug substrates (Coyuco et al., 3366 2011) but also could be associated to unspecific xenobiotic accumulation. Human colon carcinoma 3367 cell line HT29 exposed to SWCNT (0.1 and 0.2 µg/mL) during 48 h and 72 h induced a decrease in cell 3368 growth and mitochondrial activity and increase in ROS generation but importantly, induction of DNA 3369 strand breaks can be detected after 3 h of incubation with SWCNT at concentrations ≥0.0001 µg/mL 3370 and the DNA damage started at 0.00005 µg/mL under alkaline conditions. Induction of p53 3371 phosphorylation was also observed, which is triggered by DNA damage (Pelka et al., 2013). SWCNT 3372 exposure at 100 µg/mL induced reorganization of clathrin in mast cells, which has an impact on the 3373 inflammatory response (Umemoto et al., 2014). In addition, lower exposure (10 µg/mL) to SWCNT 3374 inhibited calcium response after 30 minutes and disrupted kinase signaling pathways without 3375 alteration in proliferation even after 5 days of exposure (Umemoto et al., 2014). On the other hand, 3376 SWCNT attenuated the decrease in cell viability induced by Escherichia coli and Staphylococcus 3377 aureus infection in Caco-2 cell cultures in concentrations ranged from 1 to 10 μg/mL. In addition,

109

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3378 SWCNT reduced the NLRP3 inflammasome-mediated IL-1b secretion of Caco-2 cells induced by 3379 bacterial infection (Chen et al., 2015). This study highlights that SWCNT could interact with gut 3380 microbiota and also, the SWCNT capability to reduce components of the inflammatory response. 3381 MWCNT with length of 5-10 nm and diameter of 20-30 nm induced a slight decrease in viability of co- 3382 cultures of Caco-2 and HT29 cells exposed to 100 µg/mL but up-regulated cell proliferation, 3383 antiapoptotic and DNA repair pathways associated with cell survival (Tilton et al., 2014). In this study, 3384 a comparison with titanium dioxide nanoparticles shaped as belts showed that there are general 3385 early cellular responses to MWCNT and titanium dioxide after 1 h of exposure but unique gene and 3386 protein expression pattern was associated to each type of nanomaterial after 24 h of exposure to 10 3387 µg/mL and 100 µg/mL (Tilton et al., 2014). MWCNT exposure to Caco-2 cell cultures at 12.5, 25 and 3388 50 µg/mL showed interaction with the cell membrane without affecting cell morphology and cell 3389 membrane integrity but increased malondialdehyde concentration, a lipid peroxidation marker, in 3390 cell cultures exposed to 25 and 50 µg/mL. A downregulation of ABC transporters was found in cell 3391 exposed to 12.5, 25 and 50 µg/mL which could be useful to enhance cellular drug uptake (Wang et 3392 al., 2015). 3393 Graphene oxide flakes in vitro studies are very limited until now but no cytotoxicity has reported on 3394 Caco-2 cell culture exposed to 166 µg/mL after 24 h of exposure and no cytotoxicity on Escherichia 3395 coli exposed to 400 µg/mL of same graphene oxide, but a slight decrease of these bacteria viability 3396 after 10 h of graphene oxide flakes at 300 µg/mL (Nguyen et al., 2015). However, graphene oxide 3397 sheets can act as a biocompatible surface for Escherichia coli and promote proliferation (Akhavana 3398 and Ghaderia, 2012) highlighting the importance of shape but also, the attention on effects induced 3399 in human GIT microbiota. 3400 3401 4.11.4.3.2 Metal-based nanomaterials 3402 Several studies have conducted to evaluate up take of nanoparticles in GIT-related cell cultures and 3403 titanium dioxide nanoparticles can be internalized by Caco-2 cells (Janer et al., 2014; Faust et al., 3404 2014; Song et al., 2015), however it could be lower compared to lung epithelial cell cultures (Janer et 3405 al., 2014). Titanium dioxide nanoparticles sized in 12 nm (anatase) and rutile (22 nm) were 3406 accumulated in Caco-2 cell cultures exposed to 50 µg/mL with no alterations in cell viability or DNA 3407 damage in spite of increase in ROS generation, which was sustained for 48 h. In addition, induction of 3408 multidrug resistance protein 1, 2, 4 (MRP1, MRP2 and MRP4) and breast cancer resistance protein 3409 (BCRP) gene expression was observed in cell cultures exposed to both types of nanoparticles, but the 3410 expression of their proteins were higher in exposed cells to rutile (Dorier et al., 2015). These proteins 3411 are associated to solute-liquid transporters and efflux pumps from the ABC transporter family and 3412 their increase have been associated to xenobiotics resistance under pathological conditions.

110

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3413 The influence of digestion simulation fluids, including saliva, gastric, duodenal and bile juice, and cell 3414 differentiation has been investigated on Caco-2 cells exposed to titanium dioxide nanoparticles (50 3415 and 200 µg/mL). First, titanium dioxide nanoparticles exposed to gastric fluids induced a slight 3416 decrease in cell viability and membrane integrity and increase in ROS generation only in 3417 undifferentiated cell cultures and had no effect on differentiated ones. Undifferentiated Caco-2 cells 3418 internalized native nanoparticles easily, but not pretreated nanoparticles compared to differentiated 3419 cells (Song et al., 2015). Titanium dioxide E171 food grade induced disruption of the brush border of 3420 Caco-2 cell culture exposed to 350 ng/mL, an effect that was not associated to the nanoparticle 3421 sedimentation (Faust et al., 2014). 3422 Titanium dioxide shaped as spheres is usually evaluated but more recently, nanoparticles shaped as 3423 belts have been investigated including nanobelts with a 7 μm length, 0.2 μm width and 0.01 μm 3424 thickness in Caco-2/HT29 co-cultures showing an induction of inflammation pathways, cell cycle 3425 arrest, DNA replication stress, genomic instability and apoptosis (Tilton et al., 2014). However, until 3426 now, titanium dioxide which is in contact with GIT is more related to a sphere-shaped titanium 3427 dioxide than belts or fibers. 3428 Effects of particles such as silicon dioxide was published in 1980 by O’Neil et al., with the hypothesis 3429 that silicon dioxide particles found in Mediterranean grass Phalaris minor could be the etiology of 3430 oesophageal cancer, which had a high incidence in Iran, since this plant was present as a wheat 3431 contaminant in the Middle East and proliferation of fibroblast cell cultures was observed after 3432 incubation with fibers isolated from the Mediterranean grass Phalaris minor (O'Neill et al., 1980). In 3433 the following decades, some in vitro evidence sustained limited cytotoxicity in GIT-related cell culture 3434 including colorectal cell lines showing that silicon dioxide nanoparticles sized in 25 nm and 100 nm 3435 had low cytotoxicity in a concentration between 10 µg/mL to 150 µg/mL (Sergent et al., 2012) and 3436 Caco-2 cell lines are able to internalize 20, 60, and 90 nm nanoparticles in a time, concentration and 3437 size-dependent manner but also this type of nanoparticles could enhance the bioavailability of non- 3438 soluble hydrophilic drugs (Zhang et al., 2012) or molecules such as insulin (Andreani et al., 2014). This 3439 opens the question of the possibility of silicon dioxide nanoparticles as enhancers in the entrance of 3440 other type of molecules which could have an adverse effect on cells. 3441 The influence of digestive fluids and cell differentiation has also investigated in silicon dioxide and 3442 zinc oxide nanoparticles (0.3125, 1.25, 5, 20 and 80 μg/cm2) have also tested showing that 3443 undifferentiated cell cultures of Caco-2 cells had more sensibility to nanoparticles toxicity. However, 3444 no effects were associated to gastric fluid simulation on ROS generation (Gerloff et al., 2013). In this 3445 study, a solution of 34 mM NaCl/HCl at pH 2.7 followed by an incubation in an intestinal solution 3446 (carbonate/bicarbonate buffer 50 mM at pH 9.5 including 1.68 g of NaCO3, 7.16 g NaHCO3 and 4 g 3447 NaCl in H2O) was used as gastric fluid simulation and after incubation the samples were diluted in

111

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3448 cell culture medium. The evaluation of digestive fluids on nanoparticles-induced effects is a key 3449 element to mimic more realistic conditions of exposure to nanomaterials, however, the methodology 3450 to simulate gastric conditions has not been standardized yet. 3451 Silicon dioxide nanoparticles obtained from different companies were tested on human gastric 3452 epithelial cells and decrease in cell viability between 40 and 60% was observed after 72 h of exposure 3453 to 200 μg/mL, regardless of the company. However, same nanoparticles exposed in Caco-2 cell 3454 cultures showed less toxicity but Caco-2 cell cultures showed higher percentage of cells accumulated 3455 in G2/M cell cycle phase than gastric epithelial cells, in which this cell cycle phase remained without 3456 changes (Yang et al., 2014) which highlights the cell-type response associated to nanoparticles 3457 exposure irrespective to nanoparticle uptake, since both types of cell culture showed comparable 3458 endocytic vesicles in the cytoplasm. Other studies have shown limited toxicity associated to size of 3459 silicon dioxide nanoparticles sized in 25 and 100 nm on colorectal HT-29 cell cultures (Sergent et al., 3460 2012), however, silicon dioxide nanoparticles sized in 70, 100, 300, and 1,000 nm had an inverse 3461 correlation with the amount of albumin, transferrin, and IgG human proteins bound to the 3462 nanoparticles surface (Hata et al., 2014). 3463 Iron oxide nanoparticlessized in 6, 53, 76 and 98 nm exposed to 0.1, 0.2 and 0.3 mg/mL in Caco-2 cell 3464 cultures showed higher adsorption for 76 and 98 nm nanoparticles when rates were expressed as 3465 mg/m2/min, however, when adsorption rates were expressed as m2/min, the rates were higher for 3466 smaller 6 and 53 nm nanoparticles. After 3 h of exposure, transepithelial electrical resistance was 3467 affected followed by disturbances of microvilli (Zhang et al., 2010). The degree of substitution of 3468 carboxylic groups in iron oxide nanoparticles had a positive effect on uptake in Caco-2 cell cultures, 3469 which demonstrated the effect of nanoparticle charge on internalization (Ayala et al., 2013). 3470 3471 4.11.4.3.3 Organic nanomaterials 3472 Cytotoxicity of dendrimers have been associated to destabilization of cell membranes and hemolysis, 3473 especially cationic dendrimers compared to their anionic counterparts (Chen et al., 2004). Higher 3474 generation of cationic dendrimers can induce greater degree of cell damage in colon epithelial cell 3475 cultures (Kitchens et al., 2007) and internalization of cationic G3.5 dendrimers occurs through 3476 clathrin-endocytic pathways but also can open the tight junction of cell cultures (El-Sayed et al., 3477 2002; Goldberg et al., 2010). G3 and G5 dendrimers induced a decrease in cell viability in lung 3478 fibroblast cell cultures at 1 mM (high) and 10 μM (intermediate), respectively. However, G7 3479 dendrimers showed toxicity in 1 mM, 10 μM and 100 nM (low) concentration (Roberts et al., 1996). 3480 Further studies have shown to compare evaluate the effect of different number of generations have 3481 been done in Caco-2 cells and dendrimers from G0 to G4 but G0, G1 and G2 had similar apical to 3482 basolateral permeability that was higher than G3 dendrimers. In addition, basolateral to apical

112

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3483 permeability was higher compared to apical to basolateral dendrimers permeability except to G4 3484 dendrimers, which showed no permeability in both cases. Cell viability decreased only with G2 at 10 3485 mM after 210 minutes; G3 dendrimers at 0.1, 1 and 10 mM after 210 minutes and G4 was toxic since 3486 low time of exposure (90 minutes) with all tested concentrations (El-Sayed et al., 2002). Further 3487 studies have shown that PAMAM G4, G5 and G6 dendrimers with a diameter of 4.5, 5.4 and 6.7 nm, 3488 respectively, had an increasing generation effect on toxicity on colon SW480 epithelial cells 3489 compared to HaCaT keratinocytes (Mukherjee et al., 2010). In this study, concentrations of 0.01 to 21 3490 μM, 0.03 to 5 μM and 0.01 to 5 μM of G4, G5 and G6 dendrimers, respectively, were used for 3491 cytotoxicity in HaCaT cells while concentration of 0.01 to 7 μM, 0.03 to 5 μM and 0.01 to 5 μM were 3492 used to test cytotoxicity of G4, G5 and G6 dendrimers, respectively, for SW480 cells. Regarding to 3493 dendrimers generation, higher number was associated to an increase in ROS generation and 3494 internalization of dendrimers in the mitochondria. DNA breakage was generation-increasing 3495 dependent in HaCaT exposed cell cultures (Mukherjee et al., 2010). In addition, co-localization of 3496 PAMAM dendrimers of 45 nm with the mitochondria of human lung cells was associated to release of 3497 cytochrome C. A decrease in cell viability of 39% induced by 2 mg/mL dendrimers was associated to 3498 DNA fragmentation, activity of caspases 3 and 9, increased Bax expression and decreased Bcl-2 3499 expression, which are apoptosis markers (Lee et al., 2009). 3500 On the other hand, a three-dimensional organoid kidney proximal tubule epithelial cell culture was 3501 proposed to compare in vitro and in vivo studies on dendrimers effects. G5-OH PAMAM dendrimers 3502 effects, which suffer kidney clearance by intravenous administration (Sadekar et al., 2011), were 3503 compared to effects found in a three dimensional cell culture model. Results showed that 3504 dendrimers were accumulated in the organoids with upregulation of kidney injury molecule-1 (KIM- 3505 1) and N-acetyl-ß-D-glucosaminidase (NAG), which are used as kidney acute renal failure markers. 3506 However, G5-OH PAMAM dendrimers induced low cell damage and limited cytokine release 3507 (Astashkina et al., 2014). This three-dimensional organoid model could be useful for kidney toxicity 3508 assessment since results have correspondence to in vivo findings. 3509 Dextran nanoparticles with a mean size distribution of 5 nm can be endocyted by kidney proximal 3510 tubules of male C57BL/6 mice (8 months of age) dosed at 40 mg/Kg through the tail vein. Dextran 3511 nanoparticles were detected after 24 h of injection with no detection after 7 days and this 3512 administration. In addition, dextran nanoparticles increased the albumin uptake in proximal tubules, 3513 which was associated to an increase in megalin expression, a receptor involved in albumin uptake. In 3514 spite of changes in endocytosis, no alterations in renal function were reported (Nair et al., 2015). 3515 3516 4.11.4.3.4 Nanocomposites

113

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3517 Some nanocomposites have been tested for therapeutic purposes, which are not addressed in this 3518 document, however, some of those studies show some interactions with GIT even if are not given by 3519 oral administration. For instance, mesoporous silica NPs modified by undecylenic acid with chitosan 3520 as surface modification, was loaded with glucagon-like peptide-1 into the pores formed on the 3521 mesoporous silica NPs and then, hydroxypropylmethylcellulose acetate succinate MF was used as an 3522 enteric coating. This nanocomposite was given by gavage to diabetic animals, in which a 33% 3523 decrease in glucose concentration in blood was observed. Due to chitosan component, which is able 3524 to open the tight juntions and has mucoaddesiveness properties, these nanocomposites have good 3525 epithelial adhesion in the intestine (Shrestha et al., 2016). Fullerene functionalized with hydroxyl 3526 groups, known as fullerenol, has been suggested as possible therapy against immunological 3527 alterations derived from bone marrow transplantation, however, it seems that fullerenol NPs have 3528 some effects in the small intestine, since fullerenol administered by intraperitoneal injection 3529 prevented the disruption of the epithelium and also, the traslocation of bacteria (Bernardes et al., 3530 2015). Some other nanocomposites not related to GIT, such as nanocomposites of polycaprolactone 3531 coated with polyvinyl pyrrolidone have shown induction of cell diferentiation on goblets cells in the 3532 conjunctiva with undetectable immune cell response (Shin et al., 2007). Even if this nanocomposite is 3533 used in an artificial conjunctiva, the effects on goblets cells could be relevant in terms of GIT, in 3534 which goblet cells have a protective role through the mucin synthesis. Early cell goblet cell 3535 diferentiation and evasion of immune cell detection must be evaluated by similar nanocomposites 3536 with potential usage in the GIT. The coatings, such as those able to increase the adhesion in the 3537 epitelium have also to be evaluated, specifically if those nanocomposites are designed for longer 3538 treatments. The disruption of tight juntions caused by some coatings could have a detrimental 3539 effects and might cause translocation of microbiota. In addition, if the core or matrix of the 3540 nanocomposite is a metal or a carbon-based component, those NPs could end up in in the 3541 sourunding cells, highlighting the attention of possible effects on the epithelium or cells from the 3542 immune system. b 3543 3544 4.12 Human studies of nanomaterials 3545 Ethical considerations have limited human studies regarding to nanomaterials exposure but there is 3546 some information about human studies especially for the research on some metal-based and organic 3547 nanomaterials used as food additives or food packaging but there is still a deficiency of human 3548 studies. 3549 In this regard, one of the first reports that found the presence of dark pigment deposits in human, 3550 showed that macrophages of Peyer’s patches had dark pigments. The samples were obtained from 3551 post-mortem and intestinal resections from human samples and further analysis revealed aluminum,

114

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3552 silicon, and titanium (Shepherd et al., 1987). In addition, those elements were also found in 3553 mesenteric lymph nodes, and in some transmural inflammatory aggregates of patients with Crohn's 3554 disease (Shepherd et al., 1987). Later, titanium dioxide sized in 100 to 200 nm, aluminosilicates <100- 3555 400 nm in length and silicates ranged from 100 to 700 nm in length were present in the basal areas 3556 of ileal Peyer's patches and colonic lymphoid aggregates from 20 adult patients diagnosed with 3557 Crohn's disease, ulcerative colitis, and colonic carcinoma (Powell et al., 1996). Also Gatt (2004) 3558 demonstrated the presence of particles of various sizes in diseased colon mucosa. At that time, it was 3559 suspected that oral exposure to certain foods could include those particles and precisely, it was 3560 suspected that titanium dioxide could be contained in food as E171 additive, since it was already 3561 approved for human consumption (Powell et al., 1996). 3562 Confirmation of human absorption of titanium dioxide was reported later and particles of 160 nm of 3563 diameter showed higher absorption than particles sized in 360 nm in six male volunteers aged 3564 between 24 and 66 years-old (Böckmann et al., 2000). Then, the hypothesis that macrophages from 3565 the lung can be deposited in the bowel or in the lymphatic vassels through the bloodstream led to 3566 investigate the presence of pigments in the colon, terminal ileum, duodenum, stomach and 3567 esophagus biopsies from children suspected of having Inflammatory Bowel Diseases. Aggregates of 3568 particles in the terminal ileum and Peyer’s patches were found in 63 children and the pigment 3569 presence was higher in patients with ulcerative colitis than children with Crohn's disease. The age of 3570 patients ranged from 3.7 to 18.1 years-old (Hummel et al., 2014). Unfortunately, this study does not 3571 analyze properly the aggregates but again it was suspected that titanium dioxide and aluminosilicates 3572 could be the pigments found and they could be orally obtained though the diet and pharmaceuticals. 3573 However, a more detailed absorption study in nine human volunteers aged from 30 to 56 years 3574 received a single 5 mg/kg dose of titanium dioxide dispersed in water. Volunteers received particles 3575 sized in 15 nm (study 1), 100 nm (study 1) or 5000 nm (study 1) and majority of them participated in 3576 the three studies. No differences in gastrointestinal absorption were found between the three types 3577 of particles absorption below 0.1% (Jones et al., 2015). 3578 Even if silicon dioxide nanoparticles are being used as food additive, the effects of oral consumption 3579 are still poorly investigated in human studies. Effects of silicon dioxide particles are reported since 3580 1982, after the O'Neill and colleagues demonstrated that these particles were located in the mucosa 3581 of oesophageal tumors from Chinese patients, probably because silicon dioxide was present as 3582 impurity in some foods such as millet bran (O'Neill et al., 1982). The size of those particles were no 3583 identify but highlighted the possibility of harmful effects on human health if they are orally ingested. 3584 A sixty-patient volunteer study was performed to evaluate the effect of 14-day oral exposure to two 3585 types of commercial silver nanoparticles with an average daily ingestion of silver colloid formulation 3586 is estimated to be 100 μg/day for 10 ppm, and 480 μg/day for 32 ppm silver and silver was detected

115

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3587 in serum but no alterations in metabolic, hematologic, urine, physical parameters and sputum 3588 morphology and imaging remained without changes (Munger et al., 2014). Serum silver 3589 concentration was detected in 42% of subjects receiving 100 μg/day and in 92% of subjects receiving 3590 480 μg/day but was undetectable in the urine (Munger et al., 2014). 3591 3592 3593 4.13 Other Exposure Routes relevant for occupational setting 3594 The incidental ophthalmic exposure can be the result of manufacturing process in which workers 3595 without eye protection has eye contact with nanoparticles. The evidence of ophthalmic toxicity in

3596 environmental settings is limited but ocular exposure to TiO2 NPs in rabbits reduced goblet cell 3597 density and those cells are responsible for gel-forming mucin secreted in the conjunctival epithelium 3598 (Eom et al., 2016). Long-term ocular exposure (90 days) of ZnO NPs in rats showed not only NPs 3599 deposition in the eye ball and muscle and surrounding tissues but also retinal atrophy between the 3600 inner nuclear layer and the outer nuclear layer of exposed eyes (Kim et al., 2015). 3601 3602 The translocation of inhaled particles though olfactory bulb has arisen the concern of toxicity in this 3603 structure and there is direct evidence of olfactory bulb-brain translocation pathway in rats (Kao et al., 3604 2012). In olfactory deposition the size of nanoparticles is critical and smaller nanoparticles, sized at 3605 3nm are deposited in the anterior nose, while larger particles particles were uniformly distributed 3606 throughout the nasal structures (Garcia and Kimbell, 2009). In terms of toxicity, some metal oxide 3607 particles, such as ZnO activated the Nrf2-mediated oxidative stress response after short-period of 3608 exposure (2 h) in human olfactory neurosphere-derived cells while longer exposure (6 h) induced an 3609 activation of DNA damage repair pathways (Osmond-McLeod et al., 2013). After 24 h of exposure, 3610 cell death markers were positive and in all cases, the coating of ZnO nanoparticles decreased the 3611 toxicity (Osmond-McLeod et al., 2013). In in vivo systems, the intranasal administration of ZnO 3612 induced disruption of the olfactory epithelial accompanied by infiltration of macrophages and 3613 neutrophils in the lamina propria (Gao et al., 2013). The ZnO damage was associated to 3614 mitochondrial destruction (Gao et al., 2013). There is evidence of alteration in neurotransmitter 3615 secretion in the olfactory bulb after nanoparticle exposure, as has been demonstrated after copper 3616 nanoparticle intranasal exposure (Zhang et al., 2012). One of the main concerns of olfactory bulb- 3617 brain translocation of nanoparticles is strongly associated with apparent low bulb olfactory toxicity 3618 but some alteration in brain related to lactate dehydrogenase release and increase in glutamate 3619 brain concentrations even no evidence of pro-inflammatory cytokines were detected in brain (Liu et 3620 al., 2011). Another concern of olfactory bulb-brain translocation is the capability of deposition in 3621 deeper brain regions including cerebral cortex, hippocampus, and cerebellum. Intranasal instillation

116

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3622 of TiO2 NPs caused pathological alterations in olfactory bulb and hippocampus and induced higher 3623 TNF-alpha and IL-1 beta release (Wang et al., 2008) and as it has been demonstrated in inhalatory

3624 and oral exposure, the crystalline phase has impact on release of these cytokines since anatase TiO2 3625 NPs had higher pro-inflammatory potential that rutile (Wang et al., 2008). Similar effects were 3626 reported in olfactory bulb-brain translocation of iron NPs, in which pathological alterations in 3627 olfactory bulb, hippocampus and striatum were found and also microglial proliferation and activation 3628 in olfactory bulb was reported (Wang et al., 2011). It is important to highlight that microglia 3629 represents the immune cell response of the brain and its activation and proliferation seems to be a 3630 direct consequence of the stress caused by nanoparticle deposition. 3631 3632 4.14 Placental exposure 3633 The exposure during pregnancy in occupational settings has also been addressed more recently 3634 derived from the concern of inhalation or some other routes of exposure. While some types of 3635 nanoparticles have not been found in placenta of pregnant C57BL/6 mice or fetal tissues, such as 3636 gold nanoparticles sized in 2 nm and 40 nm injected intravenously or intraperitoneally after 24 h 3637 (Sadauskas et al., 2007) some others, have found that gold NPs modified by ferritin or polyethylene 3638 glycol showed higher internalization rate into the placenta and the fetus than gold NPs coated with 3639 citrate, which suggests that NPs with negative surface have less capability to be internalized (Yang et 3640 al., 2012). However, the surface modification, both, positively or negatively on iron oxide NPs had the 3641 same ability to cross the placenta but positively surface modification induced higher toxicity in CD-1 3642 mice (Di et al., 2014). 3643 In addition, exposure in the late phase of the pregnancy showed that placenta can uptake gold NPs 3644 coated with polyethylene glycol and citrate without reaching the fetus (Yang et al., 2012). These 3645 findings highlight the importance of the gestational age exposure and also the effect of coating. In 3646 this regard, it seems that NPs can reach the newborn after the delivery through dam´s milk as well, as 3647 it was found after fullerenes administrated by the tail vein in rats, in which fullerenes could also be 3648 distributed into the GIT of the newborn (Sumner et al., 2010). Oxidized SWCNTs administrated in 5 3649 days-pregnant mice induced fetal malformations and miscarriages (Pietroiusti et al., 2011), which 3650 contrast with the undetectable or low toxicity induced by gold NPs, however, it seems that the sole 3651 accumulation in absence of damage can cause intrauterine inflammation, specifically by gold NPs 3652 sized at 3nm and 13 nm (Tian et al., 2013). It has also demonstrated that gold NPs can be 3653 transolcated to the fetus trhough transtrophoblastic channels or via transcellular processes instead 3654 of amniotic fluids (Semmler-Behnke et al., 2014). In addition, syncytiotrophoblast could be involved 3655 in the route of internalization and at least for polyestirene NPs, the size and the coating were lees 3656 important (Grafmueller et al., 2015).

117

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3657 Functionalzation of SWCNTs with polyethylene glycol at low dose (10 μg/mouse) in pregnant CD-1 3658 mice seems to exhibit less tocxicity compared with higher dose (30 μg/mouse), which showed 3659 teratogenic effects and reduced vascularization in the placenta (Campagnolo et al., 2013). 3660 Inhaled cadmium NPs have impact on the embryo implantation, as was demonstrated in CD-1 3661 pregant mice. Cadmium was detected only in the placenta of pregnant mice which inhaled these NPs 3662 every other day, however, if inhalation was conduceted daily, placental weight decreased and the 3663 growth of the neonatal was delayed (Blum et al., 2012). Inhaled copper NPs induce also a decrease in 3664 the maternal weight of pregnant mice with similar effect on the weight of newborns, however, even 3665 if no copper was found in the placenta or the fetus, alterations in gene expression of the newborns 3666 spleen were detected, mainly related to Th1 and Th2 responses (Adamcakova-Dodd et al,. 2015). 3667 In spite of the high doses (10, 100, and 1000 mg/kg/day) used for evaluation of oral effects of silver 3668 NPs sized at 6.45 nm in pregnant Sprague-Dawley rats, no alteration in the fetus development were 3669 found and hepatotoxicity in dams was the main toxic effect (Yu et al., 2014), however, silver NPs 3670 administered intravenously induced an upregulation in the meiosis which was associated with a 3671 reduction in the DNA methylation levels of Zac1 transciption factor while Igf2r had an increase in 3672 DNA methylation of placentas derived from silver NPs exposed mice (Zhang et al., 2015). Zinc oxide 3673 NPs were dispersed in 5% glucose solution and administered by the tail vein in 9-week old Sprague- 3674 Dawley rats from day 6 to day 20 of gestation and high levels of zinc were found in liver, lung and 3675 kidney and in the fetus was high in the liver but no evidence of toxicity related to the NPs 3676 administration was found (Lee et al., 2016). 3677 Placental barrier models have compared the possible toxicity among different NPs and according to 3678 some findings, iron oxide NPs could have higher toxicity potential compared with silicon dioxide NPs 3679 (Correia Carreira et al., 2015) and maybe these could be usefult systems to predict the most toxic 3680 characteristics of NPs in terms of placental damage. Some other models, such as amnion epithelial 3681 cells or BeWo b30 cell line, both derived from human placenta, can be useful to understand some 3682 cellular mechanism of damage after NP internalization. For instance, ZnFe2O4-NPs can induce ROS 3683 generation, mitochondrial membrane disruption, DNA damage and inflammation (Saquib et al., 3684 2013) while Fe2O3 NPs can also cause disruption of tight juntions (Faust et al., 2014). 3685 3686 There are several limitations in the current studies about the placental exposure of NPs, and some of 3687 the most important for in vivo studies are related to the selected dose and the route of exposure for 3688 the pregnant animals. Those studies showing placental translocation through inhalation of NPs seem 3689 to have high relevance in terms of occupational exposure and those in which animals received the 3690 NPs by oral consumption as well, since both are realistic routes of exposure. However, tail vain 3691 injection of NPs is still helpful to understand if NPs can reach the placenta and fetus once NPs enter

118

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3692 into the blood stream independently of the exposure route. The immunotoxicity of NPs in pregnant 3693 experimental models has been less explored but at least the toxicological studies performed until 3694 now reveal that inflammation in the placenta and possible spleen immune responses in fetus could 3695 have additional impacts. Indeed, most of the studies have been focused on the evaluation of fetus or 3696 newborns and no reports until now on further effects, for instance in the growth newborns until 3697 they reach the adult size or in the following pregnancies of the dams if they are exposed again. The 3698 exposure of NPs in the second generation could also have impact on the second exposed fetus 3699 generation. Higher susceptibility for complications or alterations during the following pregnancies 3700 could also be observed. In addition, there is still limited information about the translocation of 3701 internalized NPs in the fetus to other tissues, for instance, during the clearance attempt. For 3702 instance, in the dams, if macrophages have internalized NPs, translocation to lymph nodes can occur. 3703 Another important issue is the timing of exposure during the pregnancy. The animal models showed 3704 that effect during early stages of the pregnancy could have higher impact on the development than 3705 exposure in the late stages. This might have relevance since women cannot be aware of pregnancy 3706 specifically during the firs weeks. 3707 Even if there is still unknown mechanisms of pacental NPs translocation and the evidence of real toxi 3708 city in the dams and fetus/newborns is still uncomplete, the risk for placental translocation in pregna 3709 nt woman exposed to NPs seems to deserve a warning. 3710 3711 3712 3713

3714 5. Dosimetry

3715

3716 5.1 Consideration of dosimetry in nanomaterial in vivo and in vitro study

3717

3718 5.1.1 In vivo study

3719 When designing an in vivo or in vitro toxicity study for ENM exposure, the actual workplace 3720 exposure scenario or consumer exposure should be considered. Generation of (NOAA) aerosol should 3721 attempt to simulate important characteristics of actual real life NOAA emissions and exposure. For 3722 instance, when considering workplace aerosols, concentration in terms of mass or number (if known), 3723 shape, size and size distribution of NOAA, size distribution of NOAA, frequency of exposure, and 3724 handling and manufacturing conditions are basic information for consideration of NOAA generation

119

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3725 for the in vivo study. Various methods of NOAA generation could be adopted to generate NOAA 3726 aerosols to simulate actual exposure situation. For inhalation experiments the starting NOAA could 3727 be in the form of aggregated/agglomerated dust form, well or poorly suspended in liquid media, or 3728 solid state material where the NOAA is generated by condensation/evaporation or spark generation. 3729 For the oral administration study, the actual amount or form of intake ENM in food or consumer 3730 product should be considered. The ENM suspension should reflect or simulate actual situation of 3731 intake of nanomaterial in terms of concentration, delivery or vehicle in suspension, exposure rate, 3732 and duration of exposure. 3733 Special care should be taken that exposure concentrations administered in an in vivo study are not 3734 too high. In inhalation studies, high exposure concentrations could lead to overload conditions 3735 In oral studies, effects of TiO2 were actually shown to decrease at high concentrations, presumably 3736 because particles were aggregated and no longer taken up by the gastrointestinal tract (Fabian, E., La 3737 ndsiedel, R., Ma-Hock, L., Wiench, K., Wohlleben, W., and van Ravenzwaay, B. (2008). Tissue distribut 3738 ion and toxicity of intravenously administered titanium dioxide nanoparticles in rats. Arch Toxicol 82( 3739 3), 151-7.). 3740 Dermal and oral in vivo studies will mostly be focused on nanomaterials suspended in a liquid or solid 3741 matrix. Due to the high extent of nanomaterial interaction with these matrices, it is recommended to 3742 use a matrix which resembles the exposure conditions in real life as closely as possible. 3743 3744 5.1.2 In vitro 3745 In vivo exposure studies usually expose experimental animals to chemicals or ENMs with controlled 3746 duration to simulate actual human exposure situations from short-term, long-term to life-time. 3747 Studies that are in vitro are performed with cells or biological molecules outside their normal 3748 biological context such as cells in artificial culture medium. In vitro assays are simple, fast and cost- 3749 effective as predictive screens for toxicity assessment of chemicals and nanomaterials in commerce. 3750 Moreover, in vitro models allow for extensive investigation of particle-cell interactions in human cells, 3751 which may be difficult to conduct in vivo. 3752 At the same time, there are a number of known limitations of in vitro studies for studying the effects 3753 of chemicals and ENMs, many of which are related to dosimetry. Firstly, most in vitro studies use a 3754 limited number of immortal cell types which are exposed to chemicals or ENMs under static 3755 conditions. For this reason, most in vitro studies are only capable of studying short term effects of 3756 chemicals and nanomaterials which do not involve complex interactions between different cells and 3757 organs. For the immune system, this generally means that only parts of a complete immune response 3758 can be studied in vitro, such as a specific pathway or induction of immune response mediators. A 3759 number of additional challenges occur specifically when studying ENMs in vitro.

120

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3760 Firstly, ENMs tend to interact to a great extent with components of the cell culture medium. Proteins 3761 and lipids adhere to the surface of the material, essentially changing the identity of the material 3762 interacting with the cells. Due to different components present in the cell culture medium compared 3763 to the in vivo situation, this may lead to problems extrapolating effects from in vitro to in vivo. In 3764 addition, materials will almost always aggregate or agglomerate, especially at high concentrations, 3765 leading a greater extent of cell interaction and more aggregates and agglomerates with increasing 3766 concentrations. This is an issue that needs to be taken into account when making dispersions for in 3767 vitro studies, which will be discussed further under paragraph XX 3768 Furthermore, one should be aware that administering materials under static conditions to cells 3769 cultured at the bottom of a culture plate will in most cases lead to different interaction rate of the 3770 materials with cells compared to in vivo situations, leading to different cellular concentrations. In 3771 addition, similar nominal exposure concentrations may lead to different cellular concentrations for 3772 different nanomaterials due to differences in density of the material and resulting gravitational and 3773 diffusional forces (Teeguarden et al., 20017) 3774 Moreover, when inhaling nanomaterials, nanomaterials interact with lung cells at an air liquid 3775 interface whereas for most in vitro studies, materials are suspended in a cell culture medium with all 3776 the components they interact with, again resulting in a different identity and behavior compared to 3777 materials dispersed in air. 3778 For the reasons just mentioned, effect concentrations found in vitro cannot be readily extrapolated to 3779 in vivo. Often, effects in vitro are found at concentrations that are much higher than could ever be 3780 achieved in vivo and therefore may not be relevant. Determination of cellular concentrations rather 3781 than nominal concentrations may improve the extrapolation, or at least allow for a better comparison 3782 of results between different in vitro studies. A computational In vitro Sedimentation, Diffusion and 3783 Dosimetry model was introduced by Hinderliter et al. (2010) to calculate the cellular concentrations 3784 of nanomaterials depending on particle size, density and their aggregation state, which will be 3785 updated shortly . 3786 3787 3788 5.2 Metrics of dose 3789 Traditionally, administered mass is used to describe doses of conventional chemical substances in experimental 3790 studies, both in vitro and in vivo. For deriving effect doses of ENMs, mass and chemical composition alone may 3791 not adequately describe the dose, because particles with the same chemical composition can have completely 3792 different effect mass doses depending on properties such as particle size. Other dose metrics such as particle 3793 number, volume, or surface area have been suggested, but consensus is lacking. An approach to determine the 3794 most adequate description of dose has been suggested by Delmaar et al (2015, which could in theory also

121

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3795 include particle properties such as zetapotential, surface reactivity and dispersity index. However, it should be 3796 realized that these properties are dependent on the experimental condition, and therefore the most adequate 3797 dose metrics for nanomaterials are also likely to be different for different experimental and real life situations, 3798 or even for different toxicity endpoints. 3799 Ideally, dose metrics should be derived based on local effect concentrations in different organs. However, the 3800 measurement of nanoparticles in aerosol or in inorganic pure solvents may be easier than those in wet organic 3801 samples such as lungs, liver and culture cell spin downs. Standardizing the pretreatment methods of those wet 3802 samples are sometimes very difficult to meet the detection levels required for the studies conducted at realistic 3803 ambient dose ranges. Once those problems are solved, conversion of the data between mass, number, surface 3804 area and other metrics will become possible at satisfactory accuracy for validation for the exposure 3805 assessments. 3806 3807 5.3 Dispersion of nanomaterial 3808 Dispersion, in general, is defined as an act to distribute particulate matter (dispersion phase) uniformly into 3809 another matter (continuous phase) or materials such as air, liquid (culture medium, etc) or solid via 3810 consolidation of liquefied base. In case of nanomaterials, dispersion may mean another step, i.e. uniform 3811 distribution of primary particles by breaking down secondary particles. In any case, the dispersion is flexibly 3812 defined according to the aim of dispersion. For nanomaterial safety, the aim of dispersion is to reproduce or 3813 simulate the status of exposure at the target sites of organisms in experiments conducted for the assessment of 3814 its toxicity. The needs of dispersion technique for nanomaterials is stronger than the larger particule matters. 3815 The reason is because the smaller particles have wider surface per unit weight and tends to agglomerate faster 3816 and stronger. There are three major force, electrostatic, steric hindrance and Van der Waals forces. Dispersion 3817 methods, especially in solution, are more or less designed to counteract with those forces. 3818 3819 5.3.1 Aerosol dispersion 3820 Dispersion methods of nanomaterials are largely divided into three categories; direct dispersion of dry sample 3821 into aerosol (Dry Dissemination), suspension in liquid for the use as suspension, and suspension in liquid to 3822 generate dispersed aerosol (Wet Dissemination). 3823 3824 The first category is represented by the use of dust feeder which is normally used for non-nano powder 3825 inhalation studies in the past and present. For the better dispersion, the dispersed powder is introduced to 3826 cyclonic airflow or acoustic vibration chamber for rigorous agitation, sometimes combined with a final filtration 3827 to remove large aggregates/agglomerates (ref. # Japan Bioassay Research Center for two year MWCNT rat 3828 inhalation study). 3829 3830 The second category is represented by sonication of the suspension often with an aid of dispersant including 3831 aquatic solution with detergents, with high molecular weight adsorptive polymer, and oily low-surface tension 3832 solvents. Various dispersant of high dispersion and low toxicity to the applied organs or cells in culture system

122

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3833 are reported. In intratracheal inoculation studies, artificial lung surfactant is preferentially used. 3834 3835 The third category is represented by an idea of critical point drying in order to avoid re-aggregation by surface 3836 tension during the drying process. A well-known method is to use water suspension at a controlled pressure 3837 and temperature introduced to a well-designed nozzle to continuously create the critical point drying at the tip 3838 of the nozzle for the continuous generation of dry aerosol (with some moisture inevitably). Another idea is to 3839 use tertiary butyl alcohol (TBA) as medium for dispersion and prefiltering in liquid phase, followed by freeze and 3840 sublimate to obtain the dry dispersant (ref. Taquann method) inspired by the sample preparation method for 3841 scanning electron microscopy. 3842 There are also “spray-and-dry” type system at near room temperature and atmospheric pressure. In such case, 3843 agglomeration of the sample within a droplet of the spray can be commonly seen; the fineness of the spray 3844 droplets often determines the level of dispersion. Non-volatile component of the dispersant if any will co-exist 3845 with the final aerosol sample. In some spray type methods, due to pretreatment and/or high pressure spraying, 3846 the size/shape of the sample, especially very thin fibrous particles, may be altered. 3847 3848 Table 1. Mode of Aerosol exposure

3849

Mode of Generation Advantage Disadvantage generation Techniques

Wright Dust - small amount of m - unstable concentration aterial required for - feeder also cannot be used for every kin Feeder generation d of dust - small, simple, and compact structure - manufactured nan omaterials can be dispersed Dry Brush Type - small, simple, and - possible triboelectric charging may occu Dissemination compact structure r from friction while brushing off material Aerosol - possible to use the s from a pellet Generator test material as it is manufactured - less test material is required

Small Scale - possible to use the - unstable concentration, which is affecte test material as it is d by shape or cohesiveness of the parti Powder manufactured cle when vacuuming the particle loaded

123

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Disperser - small and compact groove; structure - intertwined and tangled CNTs may not be vacuumed evenly, or particles may s tick together; - applies relatively weak forces for disper sing an agglomerate

Fluidized Bed - small, simple, and - variable aerosol concentration and alter compact structure ation of the test substance Aerosol - possible to use the - ambient humidity, shape, and/or cohesi Generator test material itself veness of the particles may cause unsta ble concentration - relatively weak mechanism for dispersin g an agglomerate - fibrous test substances may exhibit bre aking of individual fibers

Acoustic Dry - generates a stable - affected by the ambient humidity aerosol Aerosol - suitable for less co Generator hesive powder suc

h as silica (SiO2); Elutriator - possible to use the test material itself

Vilnius - possible to use the - unstable concentration test material as it is - weak mechanisms for dispersing an ag Aerosol manufactured glomerate Generator - suitable to generat - test particles may adhere to the vanes, e an aerosol for sm which will hinder the aerosol generation all volumes of pow process der - unsuitable for generating aerosols from - simple structure fibrous material - possible to generat e large amount (1 ~ 2500 mg/m3) of t est aerosol for a lo ng time (0.5 ~ 6 ho urs)

Rotating - possible to use the - concentration of the generated aerosol i test material itself s unstable and affected by shape or coh Drum - small, compact, an esiveness of the particles

124

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Generator d easy to use - relatively weak mechanisms for dispersi ng an agglomerate - unsuitable for generating aerosols from fibrous material - differences in concentration of aerosol g enerated over time

Atomizer / - particles suspende - particles may form from impurities in a s d or dispersed in li olvent such as DI water Nebulizer quid can be genera - possible to change the properties of na ted as aerosols no-objects such as CNTs by contact wit - small, compact, an h a liquid d easy to use - concentration of aerosol can also increa Wet se over time as the liquid evaporates - difficult to generate particles when the p dissemination articles are not well or uniformly dispers ed

Electro-static - effective dispersing - possibility of damage of the test substan of CNT by using ult ce by ultrasonic and introduction impurit assist axial rasonic energy ies such as biological agents from the D

atomizer I water

Evaporation / - simple and stable - difficult to generate materials with high method of generati melting temperature and low evaporatio Condensation ng metal nanoparti n rate.

Generator cles - produced nanopart icles can be compl etely contaminatio n free - can obtain high co Phase change ncentrated and no n-aggregated nano particles

Spark - can generate nano - few commercially available electrodes f particle aerosols in or aerosol generation and differences in Generator the entire range (1- the properties with the actual NOAA exp 100 nm), produced osed to workers in workplace air. nanoparticles can be completely cont amination free and

125

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

composed of one o r more materials d epending on requir ements and the sy stem used

Condensation - might be the only - limited to materials with appropriate vap way to make a con or pressure-temperature characteristics Nanoaerosols trolled source of ae and stable under the applied temperatur rosol of appropriat es e particle sizes and - coagulation with resulting particle size g concentration rowth with time may limit the ability to g enerate high concentrations

Chemical - Simple to use, effe - byproducts are generated Chemical ctive method for ge - use of inert gas may affect inhalation te Reaction reaction nerating nanomate sts, dilution and other gas conditioning rials may be required

3850

3851

3852 5.3.2 Liquid suspension 3853 Nano-objects insoluble in water need special attention to the sample preparation of in vitro toxicity 3854 assays since its behaviour is pretty different from that of materials soluble in water. Such attention is 3855 very critical in clarifying whether observed toxicity comes from tested nano-object itself or from 3856 some other sources. Liquid suspension utilizes certain kind of liquid or solution as dispersant. The 3857 main purpose of using a dispersant is the gain better and stable dispersion. The physic-chemical 3858 property of agglomeration common to all particle is known to be three fold; electrostatic, Van der 3859 Waals forces and steric hindrance. Other specific forces could be added such as magnetic force for 3860 magnetized particles. Zeta potential is known as an important measurement related to agglomeration 3861 by electrostatic mechanisms. The practical problem would be that the zeta potential is a combined 3862 function of the properties of the material surface and the solution used to disperse. The adhesive 3863 molecule in the medium, especially proteins are known to form “corona” around the particles and 3864 mask the original surface properties. The adhesiveness is again a function of original surface 3865 properties and conditions of the medium, including pH, salt concentration/polarity, status of hydrated 3866 water and others. Dispersion by adding steric mechanisms using high molecular weight dispersant 3867 such as PVC are much stable than those using low weight molecule dispersants. The major drawback 3868 is the difficulty in removing them. 3869 In any case, when the concentration of dispersant is lowered by diluted, re-agglomeration normally

126

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3870 takes place. In case of rigorous ultrasonication, at least two possible mechanisms of contamination 3871 are considered. First possibility is the mechanical abrasion of the vibrator by friction of the suspended 3872 particles. It is known that MWCNT is hard enough to abrade the nickle metal of a vibrator. From the 3873 view of sonochemistry some chemical reactions are reported to take place. When water is sonication 3874 at the strength where cavitation bubble formation is seem around to vibrator, nitrous and nitric acid 3875 can be formed from H2O and atmospheric N2 dissolved in the water. It is said that the water focally, 3876 where the cavitation bubbles collapse, can reach 5,000 K of high temperature and 1,000 atm of high 3877 pressure, breaking down H2O molecules to reactive species to react with N2. Where other chemical 3878 components are present, more complicated reaction may take place. 3879 3880 5.3.3 Identification of nanosuspension 3881 3882 In reporting the status of dispersion of a sample used for nanotoxicity studies, minimal requirement 3883 may be set and proposed. 3884 3885 For aerosol, direct filter sampling from the exposure chamber where the test system is exactly 3886 exposed should be taken and presented with both light microscopic (LM) and electron microscopic 3887 (EM) views (SEM and/or TEM) at low and high magnification. Size (at least two demensions) should 3888 be measured for sufficient number of particle in EM. For liquid suspension, the LM view of the 3889 suspension of the working concentration in a glass vial (preferentially microcuvet for spectrogram). 3890 Additionally, in case of IT studies using spray apparatus, LM views of the sprayed sample on a slide 3891 glass immediately before the exposure to the animal and immediately after the end of the exposure, 3892 and EM of the samples in suspension of the working concentration. In case of in vitro cell culture 3893 studies, LM views of the final medium smeared on a slide glass, EM of the cell surface (SEM) and 3894 cytoplasm (TEM) or image equivalent to them are requested. 3895 3896 5.4 Methods for dose administration in vivo; 3897 3898 5.4.1 Nose only vs whole body exposure and other respiratory tract delivery 3899 3900 5.4.1.1. Whole-body exposure vs nose only vs other direct dose delivery methods 3901 Whole-body exposure systems use chambers in which the animals can move around in cages. These 3902 systems allow exposure of large number of animals and are relatively non-stressful and less labor 3903 required, and good for chronic exposure studies where daily exposure durations are 6-24 hrs. On the 3904 contrary, these systems expose all part of animal surface, food and water, and require large volumes

127

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3905 of test articles. They occupy large area of building and expensive to build, install and maintain. There 3906 is a new system reported for small scale whole body inhalation studies with less expensive 3907 investments and less maintainance cost (Taquahashi et al. 2012) but still needs wider evaluation on 3908 its performance and stability. 3909 Nose-only/head-on exposure offers advantages over whole body such as less space need to install 3910 and manage, minimum skin contamination, smaller quantity of test article and better dose control 3911 comparing with the whole-body. However, higher stress to the animals and labor intensiveness to 3912 exposure large number of animals prohibit using for chronic study. Other non-physiological methods 3913 in delivering dose to respiratory system are intratracheal and pharyngeal instillation methods. They 3914 are good for screening for test article with small amount of test articles, but these methods are 3915 required anesthesia to administer dose and may induce tissue damage and occasional experimental 3916 death due to handling of animals. Appropriate risk assessment of pulmonary toxicology is not 3917 possible with these direct dosing methods (cf. section “sample modification during preparation and 3918 dispersion”). 3919 3920 Table 2. Advantages and disadvantages of pulmonary exposure methods.

3921

Advantages Disadvantages Whole-body exposure -physiological way of -large quantity material need exposure -dermal, eye, oral exposure -Long-term studies, repeated -Dose not well defined exposure possible -large space occupied -large number of animal -expensive -No anesthesia or discomfort -animal excretes for animals contamination -labor efficient Nose/head only exposure -Relatively physiological -stressful way of exposure -labor intensive -Good for repeated exposures -placement of animals in tube possible training may be needed -no anesthesia -minimum skin contamination

128

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

-better control of dose -relatively small space occupied Intratracheal instillation -deliver dose directly to -Non-physiological exposure lungs with risk assessment problem -good for screening of -Anesthesia needed respiratory toxicity -No repeated dosing -can be used of TK study -Tissue injury -Labor intensive Oropharyngeal instillation -deliver dose directly to -Non-physiological exposure lungs with risk assessment problem -good for screening of -No repeated dosing respiratory toxicity -Labor intensive -can be used of TK study Orophayngeal aspiration -deliver dose directly to -Non-physiological exposure lungs with risk assessment problem -good for screening of -Potential aspiration of oral respiratory toxicity content into lungs -can be used of TK study -No repeated dosing response -Labor intensive

3922

3923 The uses and limitations of intratracheal instillation method are well described in the paper by 3924 Driscoll et al (2000). Use of intratracheal instillation has been recommended to address certain 3925 questions important to pulmonary toxicology; 1) screening panels of test materials for their relative 3926 potential to produce toxicity in the lower respiratory tract or to screen for effects over a range of 3927 doses, 2) comparing the effects of a new material to similar materials for which inhalation data are 3928 available, 3) comparing the toxicity of fractional components of a mixture with toxicity of the whole 3929 mixture to obtain on which constituents may be most toxicologically relevant (Driscoll et al., 2000). 3930 The limitations of intratracheal instillation are 1) difficulties in extrapolating a lung dose administered 3931 by intratracheal instillation to an inhalation exposure concentration, 2) a significant differences 3932 between humans and rodents with respect to the dependence of respirability and regional deposition 3933 on the size of airborne particles. Therefore, some inhaled particles that can reach the deep lung 3934 region in humans but would not reach these regions in rodent lungs and vise versa. 3) clumping 3935 (aggregates/agglomerates) of particles and local inflammatory responses that can affect local lung

129

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3936 clearance process and biopersistence of particles. Therefore, intratracheal doses below

3937 approximately 100 ㎍/rat should be used. They also recommend intratracheal instillation should not

3938 be used 1) when determining particle deposition patterns in the lungs that would occur following 3939 inhalation, 2) when obtaining information on the upper respiratory tract toxicity of a material, 3) 3940 when evaluating short-term clearance assay 4) when materials are reacting with the vehicle or that 3941 change in the vehicle that may alter their toxicity. 3942 3943 Table 3. Recommendation for the conduct and design of studies using intratracheal instillation 3944 (summarized from Driscoll et al., 2000) Recommendation Validation -procedure be appropriately validated in the lab (whether test material can be reproducibly delivered to the lungs) -establish that the test material is not altered by the vehicle

(cf. PROPOSAL in section “sample modification during preparation and dispersion”) Study design -vehicle -endotoxin-free saline solution ?? bio-surfactant?? -dosing volume -not be so large as to damage the lung, or so small as to preclude adequate distribution of test material within the lung -dose group -control group as vehicle only, -high dose group should avoid excessive dose resulting in immediate toxic effects to the lungs due to mechanisms not expected to occur during dosing at lower rate

-dosing frequency -single exposure to characterize potential toxicity -observation -include early time after instillation (0-3 days) to evaluate immediate response -a slightly longer time after instillation (7-10 days) to determine persistence -a later observation (1-3 months) to determine long-term effects Dose metric Mass/body weight, desirably lung surface area or lung weight

130

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3945

Method Anesthesia Deposition Evenness Realistic Toxicity Use for NP Mechanistic Long-term Costs throughout Of Dose-rate Ranking Quantitative treatment Data Caveats Exposures Resp. Tract Deposition Risk artifact assessment

Intranasal Yes (No) - No Yes No Yes (Yes) - Low instillation Intratrach Yes No - No Yes No Yes (Yes) - LOw Instillation A Orophar. Yes No (-) No Yes No Yes (Yes) - Low Aspiration B Laryngeal Yes No (-) No Yes No Yes (Yes) - Low aspiration Intratrach Yes No (+) No Yes No Yes (Yes) - Low Microscopy A Intratrach Yes No (+) No Yes No Yes (Yes) - Low Insufflation A

131

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Inhalation whole body No Yes + Yes Yes Yes (No) No + High C

Nose only No Yes + Yes Yes Yes (No) No + High Intratrach Yes No + Yes Yes Yes (No) No - HIgh D

3946

3947 Table 4. Comparison of lung exposure methods (Permission from G.O.)

3948

132

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3949 5.4.2 Oral administration 3950 Nanosuspension can be administered through oral gavage or drinking water or mixed with food. The 3951 nanosuspension can be mixed with appropriate vehicles such as corn oil or carboxy methylcellulose or other 3952 vehicualr materials. In any case the dose should be monitored regularly to identify whether proper dose is 3953 delivered to experimental animals 3954 3955 5.4.3 Skin administration 3956 Dermal exposures are of particular concern in certain applications often where NP's are dispersed in liquid 3957 matrices. Methods should address how the NP's are dispersed in a way that is uniform and that can be 3958 correlated to pertinent realistic exposures. In a study on zinc oxide researchers used human skin and a 3959 commercial sunscreen was used as the source of nanoparticles [Biomedical Optics Express 2011, 2, 3321- 3960 3333ix]. This approach clearly attempted to develop a realistic exposure scenario that closely modeled human 3961 exposure and nanomaterials in commercial use. (ISO TR 16196, 2016) 3962 3963 5.4.4 Other routes administration 3964 Nanosuspension can be adminstered to intraperitoneal or intravenous to study effect of nanomaterial on 3965 mestothelial layers and systemic circulation, respectively. As with other liquid media, it should be considered 3966 how nanoparticles may transform from dry particles to particles in a liquid medium. If dispersing practices are 3967 used it should be addressed how the dispersing method is relevant in the realistic and potential biological 3968 effects from any dispersing 3969 agents should be considered and addressed. Exposure scenarios shall be described so it is clear 3970 whether a scenario is to represent potential exposures in the use of a nanomaterial or if the exposure is an 3971 overexposure intended to elicit a biological response (ISO TR 16196, 2016) 3972 3973 5.5 Methods of in vitro exposure 3974 In vitro tests are used for rapid evaluations, and screening is typically conducted using cell culture 3975 or other in vitro techniques due to fast response time, cost, infrastructure and time constraints; factors that 3976 limit most whole animal studies. The purpose of an in vitrotest is to provide an indicator of potential adverse 3977 outcomes and effects on human health or the environment. Although there are many definitions available for 3978 the term in vitro test, an in vitro test can be generally rgarded ed as a relatively simple, inexpensive test that 3979 can be administered easily and provides rapid results. ISO TR 16196 (2016, Compilation and Description of 3980 Sample Preparation & Dosing Methods for Engineered and Manufactured Nanomaterials) suggest following 3981 considerations but are not limited to physical properties that are described more fully below. 3982 3983 Size – Some nanomaterials may have the ability to reach target sites that non-nano forms cannot. An 3984 example might be inhaled particles that go deeper into lungs. 3985 3986 Size distribution – Most nanoparticles are not of a single size thus there is a distribution of sizes. For some

133

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

3987 materials there may be fractions present of nanomaterials and nonnanomaterials of the same composition. 3988 The toxicological contributions of each may need to be distinguished. 3989 3990 Solubility – Nanomaterials have a greater ratio of surface to mass so in situations where nanoparticles are in 3991 contact with liquids, sparingly soluble nanomaterials are expected to more readily dissolve than the same 3992 material in non-nano form. Examples include amorphous silicon dioxide and zinc oxides and silver. When 3993 solubility is significant the relative contributions of toxicological effects due to particulates vs. dissolved species 3994 may need to be considered. In some cases where dissolution is complete, observed effects may not be due to 3995 the presence of nanoparticles even though the initial test substance was a nanoparticle. 3996 3997 Transformations – Because nanomaterials are not a single class of materials there are an enormous number 3998 of uses. Some of these uses may involve intentional and unintentional) transformations. An example of an 3999 intentional transformation is the use of aluminum nanomaterials in energetic applications where at least some 4000 of the aluminum is converted to alumina (Darlington et al., 2009). An example of an unintentional 4001 transformation is release of CNT’s from matrices into environmental settings in which they 4002 would be subject to photochemical processes, oxidation, biotransformation, etc. CNT’s may 4003 be subject to combustion processes if they are part of matrices that are burned. With each 4004 transformation it is needed to consider if changes have occurred that require 4005 recharacterization of the test nanomaterial 4006 4007 5.6. Comparison of constant vs initial bolus…exposure (see also below) 4008 4009 5.6.1 Continuous exposure to smaller doses of NM or exposure to bolus and maintenance thereafter 4010 The major and toxicologically significant differences between continuous exposure to smallter doses (CX) and 4011 bolus exposure (BX) is the difference in the time course of lung burden. The lung burden is a function of input 4012 and output, in an inhalation study, speed of lung deposition and clearance. In CX studies, the lung burden 4013 increases with the time of exposure, in most cases linearly. 4014 (However, It is not clear whether there is an inhalation condition that does not increase the lung burden; 4015 clearance rate may not be constant at low dose levels)). 4016 4017 On the other hand, in BX studies, the lung burden decrease according to time. The decrease function seems to 4018 be a combination of a decay curves by clearance mechanisms such as mucociliary escalator and particle-laden 4019 macrophage migration towards esophagus and local lymphnodes, and a fixation curve to the lung tissue that is 4020 for particles embedded in granulomas and/or fibrotic microscars. As a whole, the shape of the curve is similar 4021 to a decay curve except the asymptote is not zero. In case of well dispersed nanoparticles without 4022 aggregates/agglomerates, the decay is fast and the level of the asymptote is relatively low, whereas in case of a 4023 sample with large aggregates/agglomerates, the decay is minimal with very high level of asymptote due to 4024 massive formation of granulomas against the aggregates/agglomerates.

134

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4025

4026

4027

4028 Long term toxicity including 2 year carcinogenesis study 4029 4030 Historically speaking, the 2 year rat carcinogenesis study was set as to expose most of the lifespan of 4031 a test animal to match with the human regulation especially the concept of ADI (Admitted daily 4032 intake dose). ADI is calculated to have no effect after life long exposure (whole life span). And this 4033 protocol was validated by testing known human carcinogens. It is concluded that 2 year daily 4034 exposure to those human carcinogens were positive for carcinogenesis in the test animal. 4035 4036 The soluble chemicals given by oral route and gaseous chemicals given by inhalation were 4037 metabolized daily and body burden or the target organ burden reach a constant value, i.e. asymptotic 4038 value, within the first few days or weeks of the study. Thus, the body burden of the test chemicals can 4039 be considered virtually constant during the 2 year study period. When the body burden of such 4040 chemicals are plotted on a graph with y-axis as body burden and x-axis as day of the inhalation, the 4041 area under the curve (AUC) is rectangular with height as asymptotic value and width as 2 year time 4042 span. 4043 4044 On the other hand, biologically persistent particulate matter including biopersistent nanomaterials 4045 constantly accumulates in the lungs, when constant concentration of aerosol is given daily to the test 4046 animal by inhalation. It is know that the lung inhales constant amount of particulate matter everyday, 4047 and smaller but constant amount is eliminated from the lung daily (to GI tract, lymphatic system, etc). 4048 It means that under such condition, the lung burden linearly increases daily up to 2 years. When 4049 plotted to the graph with y-axis as lung burden and x-axis as day of inhalation up to 2 years, in

135

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4050 contrast to the gaseous chemicals mentioned above, the graph will be a linear line from zero at day 0 4051 to a certain value at 2 years, making a triangle-shaped AUC. 4052 4053 It is well known that the chronic effect of a given test material is a function of the AUC. The 4054 particulate matter inhalation under this exposure protocol will give half of the AUC of gaseous 4055 chemcals. Moreover, the lung burden is small during the early phase of the study, where the animals 4056 are young and may be more sensitive to the carcinogenic insults, whereas in the later phase, the lung 4057 burden is high and might inpair general status of the animals also known to attenulate to growth of 4058 the induced tumor thus might lead an underestimated conclusion on carcinogenic potency. 4059 4060 If the sensitivity of the 2 year study is depending on the AUC and we would like to compare the 4061 carcinogenic potency of any test material, and, the endpoint we need is the lung burden, not the 4062 daily aerosol concentration, then we have to align the protocol so that every 2 year carcinogenesis 4063 study has constant lung burden all the time, i.e. rectangular AUC. 4064 4065 In order to make the AUC of the biopersistent particulate matter including g nanomaterials, we need 4066 to change the exposure protocol. Instead of the daily constant concentral inhalation, we need to give 4067 relatively high concentration aerosol to boost the lung burden to an aimed amount within a relatively 4068 short period. Then, a less frequent inhalation, we call maintenance inhalation, must follow in order to 4069 keep the lung burden constant. It is, again, well known that a certain amount of the particles are 4070 eliminated from the lung daily so that the lung burden gradually decreases. This maintenance 4071 inhalation keeps the lung burden constant, and thus makes the AUC rectangular. 4072 4073HAZARD ASSESSMENT 4074 40756.1. Introduction 4076As discussed in the previous sections several studies have demonstrated that nanomaterials interact with 4077 immune cells and tissues of the immune system. Such interaction is shown to trigger cytotoxicity, 4078 oxidative stress, pro-inflammatory response, complement activation related pseudoallergy (CARPA), 4079 genotoxicity, alter metabolism and proliferation of immune and inflammatory cells (Reviewed in 4080 Lapas, 2014). 4081 4082Like for chemicals, immune toxicity of nanomaterials can be associated with several adverse outcomes 4083 (reviewed in Dobrovolskaia MA and Mcneil SE). (1) Immunosuppression – decreased host resistance 4084 to infectious agents, (2) Immune activation - increased risk of developing autoimmune diseases, (3)

136

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4085 Increased risk of developing allergic diseases – atopy, food allergies and asthma, (4) Hypersensitivity 4086 to the chemical/substance – repeated exposure, (5) Abnormal inflammatory responses – unresolved 4087 inflammation - tissue or organ damage and dysfunction 4088and, (6) Abnormal inflammatory responses – altered adaptive immune response – disease. 4089 4090Nanomaterial - induced immune modulation and dysregulation can result from inhalation, dermal exposure 4091 or oral ingestion of nanomaterials in the workplace and environment or via direct injection of drugs 4092 that contain nanomaterials. While most of the effects may be anticipated to occur locally at the site 4093 of exposure, there are evidences showing low levels of translocation of nanomaterials to blood from 4094 primary organs such as the lung (Husain et al., 2015; Kreyling et al., 2013) or gastro-intestinal tract, 4095 depending on the size and properties of the nanomaterials. For example, translocation of 4096 polysterene microspheres across gastrointestinal tract or translocation of ultrafine insoluble iridium 4097 particles from lung epithelium to extrapulmonary tissues was shown to depend on their size with 4098 smaller size particles exhibiting higher translocation (Jani et al., 1990; Kreyling WG, et al., 2002). In 4099 addition, there is evidence that orally administered silver can be distributed to other organs such as 4100 the liver (reviewed in hadrup and Lam, 2014) and translocation of titanium dioxide nanoparticles 4101 (TiO2NPs) from lungs of mice exposed via intratracheal instillation to liver and heart was reported by 4102 Husain et al (2015). The translocation of TiO2NPs to heart coincided with the activation of 4103 complement proteins in heart (Husain et al 2015). Once in the biological fluid, many types of 4104 nanomaterials are shown to form protein coronas, resulting in changes to their primary physical- 4105 chemical properties and hence their interactions with cells of the immune system (please refer to the 4106 section on protein corona) and their ability to translocate. Several classes of nanomaterials with 4107 diverse properties are shown to persist in lungs, spleen, liver and other immune responsive tissues, 4108 which will result in tissue injury and toxicity (reviewed in Ptreos RA and DeSimone JM., 2010; Murphy 4109 FA et al., 2011; Husain et al., 2013; Husain et al., 2015). 4110 4111Although the extent of such translocation is generally negligible, the resulting systemic consequences may 4112 depend on the primary properties of nanomaterials such as solubility or size. In some cases, indirect 4113 systemic responses can be anticipated based on the magnitude of the target organ toxicity. A robust 4114 lung inflammation involving multi immune cell types as observed following exposures to carbon 4115 based nanoparticles including carbon black or multi-walled carbon nanotubes tend to culminate in 4116 systemic responses in liver (Poulsen et al., 2015; Bourdon et al., 2012). Whereas, such responses are 4117 not observed after lung inhalation exposure to low toxicity nanomaterials including TiO2NPs 4118 (Halappanavar et al, 2011). 4119

137

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4120The Organization for Economic Co-operation and Development (OECD) and International Organization of 4121 Standardization (ISO), and International Conference on harmonization (ICH) of technical 4122 requirements for registration of pharmaceuticals for human use have developed international 4123 guidelines for the assessment of immunotoxicity induced by chemicals and drugs, respectively. These 4124 test guidelines (OECD TG 407, 412, 413, ISO10093-4 and ICH-S8) describe various parameters to 4125 consider in determining the appropriate test methods and provide specific details of the method, 4126 animal handling and dosing considerations, as well as the reporting standards. However, no single 4127 guideline on its own is encompassing enough to cover all aspects of immunotoxicity testing. Hence, 4128 based on the amount of chemical used and the anticipated exposure route, a flexible combination of 4129 methods described in these documents may be necessary to accurately assess the chemical induced 4130 toxicity. 4131 4132Although nanomaterials cannot be considered the same as chemicals, the research conducted so far does 4133 not reveal immunotoxicity distinct to nanomaterials, suggesting possible application of the existing 4134 chemical-specific immunotoxicity testing guidelines and tools to nanomaterials. The tiered approach 4135 suggested in Environmental Health Criteria 180 (Principles and Methods for Assessing Direct 4136 Immunotoxicity Associated with Exposure to Chemicals) may be applicable to some nanomaterials 4137 that are directly acting on components of the immune system. However, interaction of 4138 nanomaterials with cells and tissues of the immune system is suggested to be distinct from that of 4139 chemicals (described in the previous sections) and moreover, due to their distinct properties, 4140 nanomaterials have been shown to interfere with some of the components of the conventional 4141 immunotoxicity assays, leading to misinterpretation of their toxic potential. In addition, the 4142 sensitivity of the conventional test methods to assess or predict nanomaterial induced 4143 immunotoxicity is not completely known. For other immunotoxic effects, such as autoimmunity and 4144 induction of CARPA, globally harmonized technical test guidelines are lacking altogether (Giannakou 4145 et al., 2016). Overall, a major challenge for nanomaterials lies in precisely understanding the 4146 interaction of nanomaterials with the immune system, determining the appropriate assays and their 4147 suitability, and in knowing the necessary modifications required to make them applicable for testing 4148 nanomaterials. 4149 4150Given the complexity associated with nanomaterial testing, the purpose of this chapter is to recommend 4151 an immune toxicity testing strategy as it relates to specific routes of exposure to nanomaterials. 4152 Before initiating the specific experiments to assess nanomaterial-induced immunotoxicity, several 4153 questions have to be addressed as shown in Figure-1. The obvious starting point would be to 4154 determine whether there is a likelyhood for exposure to the nanomaterial that is being investigated

138

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4155 in the environment. This information can be derived from information on their potential applications. 4156 If the answer to this first question is yes, then one has to find out the route of exposure and duration 4157 of exposure. Next, if the nanomaterial of interest is suspected to interact with tissues and cells of the 4158 immune system, and whether there is a chance of tissue accumulation and biopersistence of 4159 nanomaterials has to be investigated. Intravenously injected nanomedicines are anticipated to 4160 impact the immune system due to their direct interaction with blood, depending on the time in 4161 blood (for how long do nanomaterials remain in blood) and their dose. For nanomedicines, a tiered 4162 testing approach has been developed that includes screening for immunotoxicity in the first tier 4163 using immunopathology, humoral and cell-mediated immune testing (Dobrovolskaia MA et al 2009). 4164 Subsequently, for those materials screened positive in the first tier, confirmation and 4165 characterization of immunotoxicity using additional assays for assessing humoral, innate and cell- 4166 mediated effects on immunity along with effects on host resistance are conducted. Although 4167 nanomedicines are excluded from the scope of this document, the testing strategy may still be 4168 relevant for nanomaterials that become systemically available after exposure through other routes. 4169 In addition, structural similarities to known immunotoxicants such as immunotoxicity of parent bulk 4170 materials should be taken into consideration. Depending on the answers to the questions above a 4171 selection of assays can be considered for testing immunotoxicity of the nanomaterials. 4172The following sections are intended to provide some guidance on current practices of immunotoxicity 4173 testing of nanomaterials with a special emphasis on the important factors such as, handling and 4174 preparation of nanomaterials, physical-chemical characterization, and proper dose regime. 4175 4176

139

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4177Figure -1

4178 4179 41806.2. Test material considerations 41816.2.1. Endotoxin contamination 4182Endotoxins , also known as lipopolysaccharides, are large thermostable molecules that are found on the 4183 outer cell membrane walls of gram-negative bacteria such as E coli, Salmonella, Shigella, and 4184 Haemophilus. Endotoxins are made up of bioactive lipid component (Lipid A) covalently bound to a 4185 hydrophilic heteropolysaccharides of variable length (Rietschel et al., 1994). Endotoxins are 4186 ubiquitous in our environment and exposure to a small amount of endotoxins can result in systemic 4187 reactions including respiratory distress, inflammation, pyrogenic reaction, shock and coagulation 4188 related events (Hurley, 1995). In an experimental condition, 6 pg of endotoxin is the accepted safe 4189 level of endotoxin that is shown to not induce increases in cytokine levels when administered to a 20 4190 g mouse via intravenous injection over 1-hr period. 4191Many commercially available nanomaterials are not produced in a sterile environment since they may not 4192 need to be sterile for the purposes of their applications. As a result, these nanomaterials are often 4193 contaminated by endotoxin, levels of which vary with different batches even if synthesized in the 4194 same facility or laboratory due to the variations in the synthesis and handling procedures. 4195Endotoxin can bind to the surface of nanomaterials; hydrophobic nanomaterials and nanomaterials with a 4196 positive surface charge are known to adsorb biomolecules and hence may influence binding of 4197 endotoxin to surfaces. Since contamination with a very minute amount of endotoxin can trigger a

140

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4198 large immune response, it is imperative to assess endotoxin contamination for each nanomaterial 4199 before their use in the experiments (Dobrovskaia et al., 2007). For example, gold nanoparticles 4200 contaminated with endotoxin during the sample preparation process were shown to interfere with 4201 the in vitro biocompatibility tests (Vallhove H et al., 2006). In a study (Inoue et al., 2006) investigating 4202 the influence of endotoxin contamination induced by nanomaterials, mice were exposed to 4203 lipopolysaccharide (LPS) only or LPS with carbon black nanoparticles (14nm or 56 nm) and lung 4204 inflammation response was assessed. The results showed that in the presence of LPS, 14 nm carbon 4205 black aggravated the LPS-induced lung inflammation and edema with enhnaced lung expression of IL- 4206 1 beta, MIP-alpha, MCP-1, MIP-2 and keratinocyte chemoattractant, whereas 56 nm carbon black did 4207 not show apparent effects. These results indicated that nanomaterials can aggravate lung 4208 inflammation in the presence of bacterial endotoxin, which is more prominent with smaller particles 4209 (Inoue et al., 2006). A simple and straight forward method to check whether nanomaterial samples 4210 are contaminated with bacteria is to inoculate a small amount of it onto agar plates and check for 4211 growth of bacterial colonies over few days. However, the absence of bacterial growth does not 4212 guarantee the absence of endotoxins, since these molecules may still be present in the sample but in 4213 the absence of live bacteria. There are different types of endotoxin methods that can be used to test 4214 nanomaterials. The choice of method depends on the physical-chemical properties of the 4215 nanomaterials and their behaviour in aqueous media. In addition, a special consideration should be 4216 given to their optical behavior as some of them are shown to interfere with the optical reading of 4217 endotoxin assay results (Dobrovskaia et al., 2009, Giannakou 2016). Few others, owing to their 4218 hydrophobic or positively charged surfaces, more readily adsorb endotoxins. These nanomaterials 4219 should immediately be flagged for possible endotoxin contamination testing. 4220 42216.2.2. Storage and handling of nanomaterials in laboratories 4222Nanomaterials can be contaminated with endotoxin in the environment. They should be stored in 4223 commercially available endotoxin free glassware or polysterene containers. Polypropylene containers 4224 and other plastic containers should be avoided. Equipment and glassware used for the preparation of 4225 samples for endotoxin tests can be treated by heating to a temperature of greater than 250oC for at 4226 least 30 minutes prior to their use. A separate laboratory with clean water and air supply is 4227 recommended for working with nanomaterials. In addition, everything involved with the preparation 4228 of nanomaterials including pipettes, dispersion equipment, dispersion liquid, pipette tips, containers, 4229 water, etc. should all be endotoxin-free. ISO guideline (ISO 29701:2010 (E) - ISO/TC 229, 4230 Nanotechnologies) suggests that 0.05% polysorbate 20 for extraction of airborne endotoxin from 4231 glass fiber filters and 0.1% vitamin E surfactant for extraction of endotoxin from carbon-based 4232 nanomaterials were found to improve the extraction of endotoxins. More details of equipment and

141

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4233 laboratory vessels used in the preparation of test samples are provided in ISO guideline (ISO 4234 29701:2010 (E) - ISO/TC 229, Nanotechnologies). 4235 4236Endotoxin detection methods 4237The rabbit pyrogen test (RPT) and the Limulus Amoebocyte Lysate (LAL) are most commonly used tests 4238 that are internationally harmonized and validated for assessing the endotoxin contamination of 4239 chemical substances, biological products, drugs and medical devices (Hurley et al., 1995). 4240 4241The RPT test involves measuring the rise in temperature of rabbits following the intravenous 4242 administration of a test solution at a certain concentration over a set period of time. RPT method is 4243 cost and time intensive and requires animal testing. Because of these reasons, RPT is no longer 4244 preferred. However, it is recommended when the test results of more than one type of LAL method 4245 are inconsistent. 4246 4247The LAL test uses the cell lysate obtained from the circulating amoebocytes of horseshoe crab (Limulus 4248 polyphemus) as a test reagent, which when reacted with endotoxins present in the test substance, 4249 initiate a cascade of enzymatic reactions leading to a clot formation (Ketchum and Novitsky, 2000). 4250 LAL is administered in three different ways: chromogenic, kinetic turbidity and gel clot, all of which 4251 assess different aspects of this enzymatic reaction leading to the end result of clot formation. 1. 4252 Chromogenic method: This method is based on the activation of a serine protease (coagulase) by the 4253 endotoxin, critical for the development of a clot. In the test, the natural coagulate that serves as a 4254 substrate in the reaction is replaced by a chromogenic substrate. In the presence of endotoxin, this 4255 chromogenic substrate is cleaved, releasing a chromophore, which is then measured by 4256 spectrophotometry. 2. The kinetic turbimetric method: In the presence of endotoxins LAL turns 4257 turbid. The time taken to reach a level of turbidity (onset time at which turbidity appears) is directly 4258 proportional to the amount of endotoxin present in the test substance. The higher the amount of 4259 endotoxin, the shorter the turbidity onset time. 3. Gel clot method: This is based on the presence or 4260 absence of a gel clot in the test sample, the end result of the reaction between the LAL and the 4261 endotoxin. The gel forms when proteins are coagulated due to the presence of endotoxins. All three 4262 methods have been used to test endotoxin contamination in nanomaterial formulations. These 4263 studies demonstrate that nanomaterials interfere with various steps of these assays and that 4264 appropriate controls and assay modifications are required (Dobrovskaia et al., 2010, Giannalou et al., 4265 2016). The colorimetric methods in general are not suitable for assessing nanomaterial induced 4266 effects. The optical features of the nanomaterials should be taken into consideration in selecting the 4267 appropriate method.

142

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4268In an end point chromogenic LAL test, spiking in a known amount of endotoxin into citrate stabilized 4269 colloidal gold nanoparticles showed 50% lower endotoxin recovery compared to the actual endotoxin 4270 spiked-in, whereas, higher recovery rates than the original endotoxin amounts was observed for 4271 PMLA nanoparticles, suggesting both inhibitory and enhancement activity of nanoparticles on the 4272 LAL test (Dobrovskaia 2010, Eliah 2014). Similar interference was also observed in gel clot LAL assay. 4273 The same nanoparticle formulations tested by the kinetic turbidity LAL assay showed no such 4274 interference. Thus, LAL assays to detect endotoxin contamination should include appropriate 4275 controls to exclude enzyme inhibition or enzyme activity enhancement by the nanomaterials. 4276 Inclusion of spike-in control (known amount of endotoxin standard into LAL grade water and into 4277 nanomaterials suspension) should be a routine. If interference is observed, different dilutions of 4278 nanomaterials suspensions should be tried (Dobrovskaia 2010). In a chromogenic endpoint assay, 4279 chromogenic colour intensity is measured at 405 nm which overlaps with extinction peaks of some 4280 nanomaterials such as silver nanoparticles. This will mask the colorimetric signal arising from the 4281 assay. Therefore, for silver nanoparticles or for those nanomaterials that have overlapping extinction 4282 peaks with the LAL assay, chromogenic LAL test is not applicable. For such nanomaterials, the kinetic 4283 turbidity LAL assay is recommended. However, it is important to note that in the kinetic turbidity 4284 assay, OD is read at 660 nm, which can again overlap with extinction peaks of larger aggregates of 4285 nanomaterials such as silver nanoparticles. Also, depending on the concentrations, the nanomaterial 4286 suspension may cause turbidity. In this case, samples should be diluted to minimize the number of 4287 aggregates or suspensions should be centrifuged to remove particles and only the supernatants 4288 without any particles should be used. Additionally, minimizing the time duration used to collect 4289 baseline OD resulting in lesser opportunities for nanomaterials to aggregate will reduce some of the 4290 interference (Elijha et al., 2014). 4291In any case, more than one LAL format should be used; not all nanomaterials inhibit or enhance LAL 4292 activity in a similar manner. Where possible, samples should be centrifuged to remove the 4293 nanoparticles and only the supernatant should be used in the testing. In case of suspected false 4294 positive or negative results, samples can be tested for the presence of endotoxin specific 4295 lipopolysaccharides using Gas Chromatography (GC). It should be noted though that this method only 4296 identifies the presence of the lipopolysaccharides in the sample and gives no information on the 4297 activity of the endotoxin. If inhibition and enhancement controls yield inconsistent results, an in vivo 4298 RPT test should be performed. Thus, the choice of the test method will depend on the type of 4299 nanomaterial and its various properties. Details of the endotoxin assay for nanomaterials have been 4300 reviewed in a book by Marina A. Dobrovolskaia, Scott E. McNeil (2013). 4301

143

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4302Recently, TLR-4 transfected reporter cell line has been used to verify the LAL results (Smulders bet al., 4303 2012). Macrophage activation test is another test method recognized as an alternative to the LAL test 4304 by the current pyrogen and endotoxin testing guidance (Schindler et al., 2009). However, these 4305 methods are limited to the nanomaterial suspensions that do not cause cytotoxicity or 4306 nanoformulations that do not contain cytotoxic drug components. Other assays such as human 4307 PBMC activation assay and the human monocyte activation assay recommended by European Centre 4308 for the Validation of Alternative Methods (ECVAM) may also be applicable to nanomaterials, 4309 however, pyrogen induced inflammatory mediators may not be specific to endotoxins as 4310 nanomaterials themselves may also act as pyrogens. Similarly, The Human Peripheral Blood 4311 Mononuclear Cell/ Interleukin-6 In Vitro Pyrogen Test method recommended by ICCVAM can be 4312 used. It is important to note that often endotoxin is found adherent to nanomaterial surface. With 4313 the tests mentioned above, it is not known if endotoxin that is adsorbed onto the nanomaterial 4314 surface can be detected. It is also not possible to distinguish between the free endotoxin present in 4315 the suspension and endotoxin that is adherent to the surface. 4316 43176.2.3. Recommendations or suggested modifications for testing endotoxin contamination: 43181. Inoculate small amount of nanomaterial on agar plate to check for bacterial growth. 43192. Assess for inhibition or enhancement interference by spiking-in a known amount of endotoxin into 4320 LAL grade water and nanomaterial suspension. 43213. For highly aggregating nanomaterials and where aggregation is known to interfere with the OD 4322 reading, reduce the time required to collect baseline OD and read the sample immediately after the 4323 preparation. 43244. Prepare as uniform suspension as feasible – large aggregates of some nanomaterials interfere with 4325 the extinction peaks. 43265. A series of dilutions should be used to investigate the optical interference, if this does not resolve the 4327 issues, then the colorimetric assay should not be used. 43286. Use more than one type of LAL assay. If feasible, assess the contamination using all three LAL assay 4329 types. 43307. In case of suspected false positive results, check for the presence of endotoxin-specific 4331 polysaccharides using GC. 43328. If the results obtained from two LAL assay types are inconsistent, RPT should be used to confirm the 4333 results. 4334 43356.3 Preparation of nanomaterials for exposure and characterization

144

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4336For a detailed guidance on how to prepare nanomaterials for exposure and characterization methods, 4337 readers are referred to OECD TG ENV/JM/MONO(2012)40 and ENV/JM/MONO(2014)15. The 4338 biological behaviour of nanomaterials is linked to the physical and chemical properties, which include 4339 their size, shape, solubility, surface chemistry (charge, functionalised groups), and agglomeration 4340 status. These primary material properties change in the exposure medium in which they are 4341 dispersed depending on the methods used. These properties determine how nanomaterials interact 4342 with cells and tissues, how deep they travel in a tissue, their cellular or tissue localization, 4343 toxicokinetics and ultimate fate. The property dynamics make it difficult to assess the dose- 4344 dependency of toxicological effects of nanomaterials (Oberdorster, 2005; Bell et al., 2014). 4345 Nanomaterials might not disperse homogenously in the suspension medium due to their aggregating 4346 behaviour and their dispersion is influenced by the addition of factors that aid in uniform dispersion. 4347 The aggregation and agglomeration of nanomaterials, which is strongly dependent on their colloidal 4348 stability, influence the sedimentation rate of the nanomaterials and hence affect the deposited dose 4349 (Sharma et al., 2014). Additionally, the biomolecules, proteins and other constituents of the 4350 biological medium that get adsorbed to the particle surface affect their uptake by specific cells in the 4351 tissue (see also the section on the protein corona, page ). This gets even more complicated when 4352 cells or animals are exposed to aerosolized nanomaterials, where the deposition of nanomaterials is 4353 dependent on the aerosolizing method and the exposure system used (Oberdorster 2015; Wong 4354 2007; Polk et al, 2016). Dispersability of nanomaterials will depend on the surface properties, the 4355 amount of particles to be dispersed and the mechanical treatment (sonication, milling etc (Bihari, 4356 2008)) and choice of vehicle for dispersion of nanomaterial should be based on the nanomaterial 4357 properties, and route and method of exposure (Buford et al., 2007). Ideally, more than one type of 4358 dispersant should be tested before settling on one type. It is important to note the influence of the 4359 dispersant on the behaviour of particle in the exposure medium as well as on the resulting toxicity 4360 (Zhe et al., 2007; Monteiro-Riviere and Tran, 2007). Saline is used as a common vehicle to prepare 4361 nanomaterial suspension. Water or PBS should be the vehicles of choice as they may aid in eventual 4362 deagglomeration of particles. However, BALF, BSA, diluted serum from the test animals, PBS 4363 supplemented with surfactants have been routinely used as vehicles for dispersing nanomaterials 4364 (Porter et al., 2008). For in vitro studies with lung cells the coating of pulmonary surfactant has also 4365 been highly recommended since this might influence the interaction with the cells (Schleh et al., 4366 2013). However, it should be noted that the presence of BSA in suspension containing carbon-based 4367 nanomaterials may interfere with characterization of the suspension (Elgrabi et al., 2007; Buford et 4368 al., 2007). A BALF mimic consisting of Ca2+ and Mg2+ free PBS with serum albumin and the lung 4369 surfactant (1,2 dipalmitoyl-sn-glycero-3-phosphocholine (DPPC)) is recommended to circumvent the 4370 issues related to retrieval of BALF from animals (ENV/JM/MONO(2012)40). BALF mimic contains low

145

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4371 levels of protein and surfactant found in the lung alveolar lining fluid. Pluronic –F68, PEG, Triton, 4372 DMSO and Tween are the other chemical surfactants used to aid dispersion of nanomaterials in 4373 suspension. Volume of the vehicle consisting of these various dispersants should be tested for their 4374 inherent toxicity and the exposure volume should be adjusted accordingly. Freshly prepared 4375 suspensions are the choice. Since nanomaterials can agglomerate very quickly in suspension, periodic 4376 monitoring of dispersion rate is necessary, especially if suspended nanomaterials are used for 4377 aerosolisation. pH of the solution and temperature should be routinely checked 4378 (ENV/JM/MONO(2014)15). Exposure concentration should be maintained at constant, especially for 4379 inhalation chamber by checking the stability of the suspension over time. Dynamic depolarized light 4380 scattering (DDLS), DLS, UV-Vis, differential centrifugal sedimentation, modified light microscopy can 4381 be used for this purpose (Urban et al., 2015Moore et al, 2015). 4382Since the nanomaterial dose will impact the agglomeration and aggregation, independent measurements 4383 should be reported for each dose. Some dispersing methods such as sonication or milling may alter 4384 the primary particle properties, care must be taken to characterize the particles before and after the 4385 dispersion. The report should include information on number of particles, particle concentration, and 4386 for aerosol, aerosol characteristics (ENV/JM/MONO(2014)15). 4387Thus, a detailed characterization of nanomaterials as produced and in use is essential and should include 4388 characterization in dry as well as in suspended status. A list of techniques applicable to characterizing 4389 properties of nanomaterials both in dry and in the liquid suspension are discussed in the previous 4390 chapters and are provided in the OECD (ENV/JM/MONO(2012)40; ENV/JM/MONO(2014)15) 4391 guidance on sample preparation and dosimetry for the safety testing of manufactured 4392 nanomaterials. The guidance documents also provide a list of techniques that can be used for 4393 generation of exposure aerosols. 4394 43956.4. Considerations on animal species, exposure duration and levels, routes of exposure 43966.4.1. In vivo 43976.4.1.1. Animal species 4398For details readers are directed to OECD TGs 403 and 412. A special attention should be given when 4399 choosing the species and strain of animals, age of animals, duration and level of exposure as well as 4400 the route of exposure. For nanomaterials, although strain or species specific effects are not 4401 extensively characterized, it may be better to start with the species and strains 4402 included/recommended in the routine testing of chemical induced immune toxicity (OECD TG 407). 4403 Among the rodents, murine immune system is well characterized and assays for immune toxicity 4404 testing in mice and rats are well established. Thus, mouse and rats as the first choice for testing may 4405 be supported. On the other hand, for specific immunotoxicity tests such as CARPA, the minitiature

146

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4406 pig has been reported to be the preferred model (Jackman et al., 2016). Since chemical induced 4407 immune effects are pronounced during the developmental phase, if a nanomaterial is expected to 4408 induce immune effects (thymocyte proliferation and differentiation), in utero or neonatal exposure 4409 should be considered. For developmental studies, weaning animals are considered a better choice 4410 compared to the adult animals. 4411 44126.4.1.2. Routes of exposure 4413The specific routes by which nanomaterials may enter the human body, and potentially elicit adverse 4414 immune effects are the lung via inhalation, the gastrointestinal tract via digestion, the skin, and 4415 blood vessels via intravenous injection (Oberdorster et al., 2005). As nanomaterials gain contact with 4416 the respiratory tract, the gastrointestinal tract and the skin, these biological compartments are 4417 “innately designed” to act as barriers to the passage of nano-sized materials into the organism. The 4418 following sections will discuss the testing methods in the context of specific routes of entry 4419 mentioned above. 4420 44216.4.1.2.1 Inhalation route 4422Inhalation is the physiological way by which organisms are exposed to respirable substances, and 4423 therefore, inhalation is the ‘gold standard’ with respect to the choice of method for the lung 4424 exposure to nanomaterials. Although, lung deposition via pharyngeal aspiration, and intratracheal 4425 instillation are also considered relevant, the deposited dose via inhalation, instillation and aspiration 4426 vary significantly, which is an important aspect of dosimetry and consideration of the data for risk 4427 assessment. As a result, quantitative dose-response data derived from inhalation exposure form a 4428 basis for human health risk assessment and data derived from other methods are usually used as 4429 weight of evidence in support of the decisions derived. 4430The deposition of nanomaterials in the various regions of the respiratory tract after inhalation depends on 4431 1) physico-chemical properties of the inhaled nanomaterials (size), 2) the breathing frequency and 4432 tidal volume, and 3) anatomical features of the airway (Kreyling et al., 2013; Braakhuis et al., 2014). 4433 Once deposited, further interactions of nanomaterials with surrounding milieux including mucus, 4434 proteins, surfactants, biomolecules etc, determine the biokinetics of particles overtime and their 4435 effects. Inhalation of well dispersed particles results in even distribution of particles among the 4436 different lobes of lungs. In case of instillation, the particles are already suspended before their direct 4437 delivery into lungs, which is anticipated to alter the dosimetry and lung distribution patterns. For 4438 example, inhalation exposure to well dispersed carbon nanotubes results in deposition of CNTs in the 4439 alveolar duct bifurcations and alveolar epithelial surfaces in rodents. Whereas, bolus delivery of 4440 nanomaterials results in

147

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4441centra l lung deposition (Oberdurster et al., 2015). Moreover, bolus administration of very high doses 4442 (possibly crossing the maximum inhalable dose) is suggested to alter the way in which nanomaterials 4443 interact with the surrounding milieux, and in some studies, effects of particles introduced via 4444 instillation are shown to be much higher compared to the particles introduced via inhalation (Driscoll 4445 et al., 2000; Baisch et al., 2014). However, several studies have reported no such differences either in 4446 the distribution patterns or lung responses following exposure via the two methods (Jackson et al., 4447 2012). For respirable nanomaterials, aerosol characterization is important to ensure the exposure 4448 dose. 4449 44506.4.1.2.2 Oral route 4451For oral route, exposure via oral gavage or through feed is preferred over exposure via drinking water. For 4452 guidance on exposure via oral route, readers are referred to OECD TG 407. Some nanomaterials tend 4453 to precipitate in water and thus, it is difficult to control the delivered dose. A thorough 4454 characterization of nanomaterials at various stages before the exposure including in their dry form 4455 and in the exposure vehicle is a must. If nanomaterials are fed through the diet, characterization of 4456 nanomaterial transformation in the food matrix is necessary for accurate interpretation of the 4457 results. In addition, one should be aware that nanomaterials are likely to go through a number of 4458 transformations as they travel from the acidic environment of the stomach to the alkaline 4459 environment in the intestines, although the information available on this is limited (reviewed by 4460 Ballmann et al., 2015; Frohlich et al., 2011). 44616.4.1.2.3 Dermal route 4462Exposure via the dermal route is generally conducted under occlusion, to prevent oral exposure via 4463 grooming. Nanomaterials need to be dispersed in an appropriate vehicle that does not induce a 4464 dermal response itself, and at the same time allows for maximum uptake of the nanomaterials 4465 through the skin. The majority of nanoparticles will remain in the stratum corneum, and so will not 4466 become systemically available. If any uptake through the skin occurs, it may take significantly longer 4467 than it would for soluble chemicals. Therefore, before executing any dermal in vivo study, it may be 4468 wise to conduct in vitro skin uptake studies with reconstituted skin, using sufficiently long exposure 4469 and evaluation times to determine whether any systemic uptake can be expected and at what 4470 timeline (reviewed by the Danish EPA, 2013). 4471Compared to raw nanomaterials, characterisation of nanomaterials contained in the skin care products or 4472 cosmetics used in the dermal absorption studies is challenging. Thus, a separate characterisation 4473 approach may be needed that includes methods to isolate, purify and concentrate nanomaterials 4474 before their characterisation in the cosmetic products. Some of the properties that are known to 4475 influence dermal penetration of nanomaterials include particle size, chemical composition, and

148

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4476 surface chemistry. Adequate characterisation of nanomaterials for these properties before exposure 4477 in dry and in formulation should be conducted (Danish report, 2015). 4478It is suggested that nanomaterials in their dry state cannot penetrate skin. Therefore, the skin penetration 4479 studies for nanomaterials have to be conducted in their suspended forms. Adequate characterization 4480 of the solvent or formulation in which the nanomaterials are suspended is critical to the 4481 interpretation of the results. Use of standard compound (positive and negative controls) is essential 4482 for validation of results and should be routinely incorporated in the study design. 4483 44846.4.1.3. Duration and levels of exposure 4485For chemicals, 28 d repeated exposure regimen is recommended before the assessment of immune 4486 parameters (OECD guideline 408). However, since effects leading to immune dysfunction following 4487 exposures to nanomaterials may take longer than that known for chemicals, understanding 4488 toxicokinetics of the nanomaterials being investigated prior to the immune toxicity assessment may 4489 be important. 4490In general, higher toxicity is observed at the higher doses; however, nanomaterials do not show dose- 4491 dependent transition in the toxicity observed (Oberdorster, 2005). This means that the toxicity 4492 induced at a lower dose may be different from the toxicity observed at a higher dose. This is mainly 4493 due to the fact that nanomaterials aggregate (a state of dispersion, where nanomaterials are loosely 4494 clumped together through non-covalent interactions) in suspension in a dose-dependent manner. 4495 Aggregation of nanomaterials interferes with their uptake and distribution and with their ability to 4496 interact with surrounding microenvironment, influencing their toxicity potential. The relative state of 4497 dispersion influences the type of immune or pathological responses to nanomaterials. Thus, for the 4498 selection of doses, traditional methods of choosing doses such as LD50 may not be applicable 4499 (considering the fact that nanomaterials of similar composition may exhibit different LD50 4500 (Oberdorster 2005; Bell et al., 2014)). As a rule, a minimum of three doses in addition to controls is 4501 necessary. Where possible, the selection should be based on a pilot experiment. 4502 45036.4.1.3.1 Multiple exposure versus single exposure 4504In reality, exposure to low doses of respirable particles over a long period of time is the norm. However, 4505 most of the inhalation studies are limited to single exposure of a large dose of nanomaterial in a 4506 short period of time. This is especially true for intratracheal or pharyngeal aspiration methods, where 4507 a bolus amount of material is directly deposited in lungs. Although, the high bolus doses delivered by 4508 intratracheal instillation method or pharyngeal aspiration methods is suggested to be the same as 4509 deposited per unit alveolar surface area in humans exposed to occupational exposure levels over a 4510 40-yr working life, these studies do not take into account the rate at which these doses are delivered

149

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4511 and cleared (Oberdorster et al., 2015). Inhalation way of deposition offers opportunities to deliver 4512 doses at a low rate. Although this is an issue with intratracheal instillation and pharyngeal aspiration, 4513 the large bolus doses can be split into multiple smaller doses and administered over weeks. However, 4514 the latter techniques are invasive and require anesthesia, which may predispose animals to 4515 accidental injury, and repeated anesthesia can add to the stress over time. 4516 45176.5. Considerations on in vitro cell types, exposure methods, duration and levels 4518A special attention should be given when choosing the cell types and the species that they are derived 4519 from, passage numbers, duration and level of exposure as well as the way by which the cells are 4520 exposed in vitro should be noted. Guidelines for good cell culture practice are required, should be 4521 applied and documented, including the assessment of materials used in the experiments (e.g. the 4522 cultured cells, the culture medium, and the culture substratum (Gstraunthaler and Hartung, 2002)). 4523 Similar to chemicals, in vitro responses to nanomaterials will heavily depend on all of these factors 4524 (Park et al., 2009). Although, the cell types that are routinely used for in vitro testing of chemicals can 4525 be used, their sensitivity varies with each nanomaterial variant. Thus, choice of cell types should also 4526 consider the possible exposure route, the type of nanomaterial being studied and the physical- 4527 chemical properties along with the consideration of types of effects being investigated. 4528Immortalised or primary cells extracted from tissues or organs are exposed under submerged conditions 4529 directly to nanomaterials suspended in the appropriate medium (usually the serum supplemented 4530 medium in which cells are grown). Depending on cell type, a specific medium is used to culture cells 4531 in vitro. For in vivo target tissues such as lung or skin, exposure under submerged condition is not 4532 realistic, since in vivo exposures of these tissues occurs at the air-liquid interface. While the exposure 4533 in submerged condition is ideal for many chemicals that are soluble, for insoluble materials such as 4534 nanomaterials, estimation of the exact dose of exposure becomes difficult as they aggregate in liquid 4535 suspension and much of what is suspended may remain in suspension and never get in contact with 4536 the cellular surface. Nanomaterials’ movement in liquids is driven via diffusion (random motion) and 4537 they do not directly sediment. The characterization of nanomaterials in physiological fluids and the 4538 assessment of their colloidal behavior are challenging due to the complex physical and chemical 4539 forces involved, the multitude types of highly complex physiological fluids, the variety and complexity 4540 of analytical methods, and theories upon which these methods are based. The ultimate fate of 4541 nanomaterials in a fluid is then dictated by its mass density, i.e. nanomaterials will settle if their mass 4542 density is greater than that of the fluid (Balog et al., 2015). There is a steady increase in number of 4543 research papers adopting the concept of particokinetics proposed by Teeguarden et al. (2007) for 4544 interpreting dose-response curves (Hinderliter et al., 2010; Zhue et al., 2013) and it is recommended 4545 to apply such an approach for any suspension experiment applying nanomaterials in vitro.

150

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4546It is also a routine practice to conduct in vitro testing using a single cell type (mono cell culture), which 4547 does not represent the complexity of a tissue consisting of multiple cell types. Alternative methods 4548 such as, use of more than one cell types (co-culture), ex vivo tissue slice cultures (Sauer et al., 2013) 4549 or exposure at the air liquid interface are emerging (Paur et al, 2011; Frohlich et al., 2014); however 4550 these methods are not completely validated even for assessing chemical-induced effects. 4551 Internationally harmonised protocols or guidance is lacking with respect to the latter emerging 4552 techniques. Some of the cell types and methods of exposure that are currently being used or tested 4553 for nanomaterials are discussed in the following sections. 4554 45556.5 Routinely used cell types of the respiratory system 4556A detailed description of individual cell types relevant to respiratory system is provided in the Chapter on 4557 Interaction with immune system. Here, specific cell types that are routinely used for assessing in vitro 4558 immunotoxicity of nanomaterials are listed. These include in vitro cell culture models of cells of the 4559 nasal, bronchial and alveolar regions. However, adequate characterization and validation of the 4560 models is necessary before their routine integration into immunological testing of nanomaterials. 45616.5.1 Cells of the innate immunity 45626.5.1.1. Epithelial cells 4563Although primary cultures of tracheal and bronchial epithelial cells are technically feasible (Forrets et al., 4564 2005; Stokes et al., 2014) a number of studies have been conducted using immortalised cell lines. The 4565 human airway epithelial cell lines Calu-3 (human origin), the 16HBE14o- (a SV-40 large T-antigen 4566 transformed, bronchial epithelial cell line derived from a normal human airway epithelia), the BEAS- 4567 2B cells (normal human epithelial cells immortalized using the adenovirus 12-simian virus 40 hybrid 4568 virus) are the other cell types regularly used to assess cellular interactions with nanomaterials and to 4569 investigate the potential of nanomaterial-induced respiratory toxicity in vivo. (Reddel et al., Cozens 4570 et al., 1994; Gurr et al., 2005; Veranth et al., 2007; Steimer et al., 2005; Forrets et al., 2005; Stokes et 4571 al., 2014). The commercially available cell line A549, which originates from human lung carcinoma 4572 (Lieber et al., 1976), is one of the most well characterized and most widely studied in vitro models 4573 (Foster et al., 1998). It has been shown that the A549 cells have many important biological properties 4574 of alveolar epithelial type II cells. The A549 cells have been well characterised (Kemp et al., 2008; 4575 Kuehn et al., 2016) and have already been used to assess both acute and long term effects of 4576 exposures to ambient nanoparticles as well as occupational exposure to engineered nanomaterials. 4577 45786.5.1.2. Phagocytes (neutrophils and macrophages) 4579Alveolar macrophages and neutrophils are the two most commonly used cell types for assessing immune 4580 responses of the lung. Monocytes can be isolated from human buffy coat and differentiated into

151

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4581 macrophages (Monocyte Derived Macrophages: MDMs) with the addition of Monocyte Colony 4582 Stimulating Factor (M-CSF) (Lehman et al., 2010). This primary cell type can be produced in a very 4583 reproducible manner and represents a reliable in vitro system. J774.1a (mouse macrophages derived 4584 from ascites) and THP-1 (human peripheral blood, derived from monocytic leukemia patients), and 4585 RA2264.7 (mouse macrophages, derived from ascites) are the frequently used phagocytic cell types 4586 for assessing nanomaterial-induced toxicity. 4587 45886.5.1.3. Other innate immune cells 4589Other innate immune cells are not directly linked to immunity against xenobiotics, but still have an 4590 important role in lung homeostasis. Natural killer (NK) cells are innate lymphocytes that are critical in 4591 the defense against infections in the early phase. Upon ingestion of particles, phagocytes may 4592 activate NK cells in order to display effector functions (Chan et al., 2014). NK cells kill infected cells by 4593 releasing granules containing perforin and granzymes by a mechanism similar to that in CD8+ T cells 4594 (Sun et al., 2011), and by producing interferon-gamma (IFN-γ), that induces macrophages to kill 4595 phagocytosed particles (Chichocki et al., 2014). Mast cells, basophils and eosinophils have little role 4596 in particle uptake, but they play a central role as effector cells in allergies. In response to 4597 environmental allergens including nanomaterials, these cells produce TH2 cytokines, such as 4598 interleukin (IL) -4, IL-5 and IL-13, and high levels of IgE antibodies (Kudo et al., 2013; Stone et al., 4599 2010). Innate lymphoid cells (ILCs) are a newly described set of antigen nonspecific, non-T and non-B 4600 lymphocytes with conserved effector cell functions. These cells secrete a variety of cytokines, such as 4601 IFN-γ, IL-5, IL-9, IL-13, IL-17, and IL-22 and are involved in mucosal tissue homeostasis. ILCs may be 4602 involved in atopic diseases by interacting with other immune cells such as mast cells (Chang et al., 4603 2013). 4604 46056.5.2 Specific cell types of the respiratory system - cells of the adaptive immunity 46066.5.2.1 Dendritic cells 4607Dendritic cells (DCs) are the most potent antigen-presenting cells (APC) in the respiratory tract and are 4608 specialized in capturing, processing and presenting antigen. DCs also play an important role in 4609 bridging innate and adaptive immune response during infections (Wilkstrom et al, 2007). It was 4610 shown in several studies that there are different subtypes of DCs in the lungs and that their 4611 proportions and immune responsiveness depends on the respiratory compartment considered (Von 4612 Garnier et al., 2005; Satpathy et al., 2012). It is a common practice to isolate monocytes from human 4613 buffy coat and to differentiate them to dendritic cells (Monocyte Derived Dendritic Cells: MDDCs) 4614 with the addition of growth factors like Granulocyte-Monocyte Colony Stimulating Factor (GM-CSF) 4615 and interleukin 4 (IL-4) (Lehner et al., 2005). However, immportalised cell lines such as the U-937

152

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4616 (Sundstrom and Nilsen, 1976) or the MUTZ-3 (Santgoets et al., 2008) have also been used. Since 4617 distinct differences in the immunophenotypic characteristics are observed between primary DCs and 4618 the immortalized cell-lines, these factors should be considered in interpreting the results. 4619 46206.5.2.2 T cells 4621The two major subsets of T lymphocytes are CD4+ helper T cells and CD8+ cytotoxic T cells (CTLs) both 4622 expressing the αβ T cell antigen receptor (TCR). Another population of T cell, named the γδ T cells, 4623 expresses a distinct type of antigen receptor with more limited diversity. In the respiratory tract, 4624 antigen-specific T cell responses are initiated by DC following sampling of airway antigen and 4625 presentation of antigen-derived peptides to naïve T cells in peribronchial and mediastinal lymph 4626 nodes. Both in vitro and in vivo studies have shown that T cell proliferation upon nanoparticle 4627 treatment is affected (Blank et al., 2011; 2013; Hardy et al., 2012; Nembrini et al., 2011). T-cells can 4628 be generated from human buffy coat from the CD14 negative fraction. The CD14 negative fraction is 4629 further purified and the CD4 positive fraction is then used to generate human T-cells in vitro (Blank et 4630 al., 2011). The most common way to measure T-cell activation and proliferation in vitro is via 4631 radioactive labelled compounds (Thymidine), as described in Blank et al., 2011. 4632 46336.5.2.3 B cells 4634Most mature B cells belong to the follicular B cell subset and are IgD+ IgM+. They have the ability to 4635 recirculate in lymphoid organs and reside in specialized niches known as B cell follicles, where they 4636 might encounter the antigen that they are specific for (Shapiro et al., 2005). Similar to T-cells, B-cells 4637 can also be isolated from human buffy coat in vitro. Human B-cells are characterized by CD19, CD20 4638 surface marker expression 4639 46406.5.2.4 Fibroblasts 4641Excessive activation and proliferation of fibroblasts can lead to excessive deposition of extracellular matrix 4642 components including collagen and lead to fibrosis-like condition. At present it is not known if 4643 nanomaterials interact with fibroblasts directly to activate them or indirectly via inducing secretion of 4644 the extracellular matrix components. Primary mouse or human fibroblasts or cell lines such as MRC-5 4645 (human lung fibroblasts) can be used to assess the proliferative activity upon exposure to 4646 nanomaterials (Vietti et al., 2013; Hussain et al., 2014). 4647 46486.5.3 Specific cell types of the respiratory system – co-culture systems 4649The lung responses to invading particles involve the interplay of several organ systems and multiple cell 4650 types. An efficient modelling of the more complex nature of the underlying physiology contributing

153

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4651 to the response, complex culture systems that involve modelling of the interactions between 4652 different lung cell types is necessary. Co-culture models contain either primary or immortalised cell 4653 lines of epithelial origin cultured along with either primary or immortalised cells of other origin 4654 including the endothelial cells, fibroblasts, and macrophages. A third cell type, usually of immune 4655 origin, is then added on top mimicking the phagocytic environment of the lung. Although, the 4656 feasibility of culturing multiple cell types permits investigation of cell-cell communications, an 4657 important feature of lung disease responses to nanomaterials, the model is technically challenging 4658 and requires specialised expertise. The routinely employed in vitro co-cultures include epithelial 4659 (A549 cells) and endothelial cells to mimic alveolar epithelial barrier (Bermudez et al., 2002; Birkness 4660 et al., 1995). Recently, a primary co-culture system to simulate the human alveolar-capillary barrier 4661 was constructed using primary cells of human pulmonary microvascular endothelial cells and the 4662 primary human type II alveolar epithelial cells to study the impact of nanocarriers (Bermudez et al., 4663 2002REF). A triple cell culture in vitro model of the human airway wall was constructed to study the 4664 cellular interplay and intricate cellular responses of epithelial cells, human blood monocytes derived 4665 macrophages and dendritic cells exposed to nanomaterials (Rothen-Rutishauser et al., 2008). In this 4666 model, monolayers of two different epithelial cell lines, A549 and 16HBE14o- epithelia as well as 4667 primary epithelial type I cells was grown on a microporous membrane in a two chamber system 4668 (Rothen-Rutishauser et al., 2008). In addition, a quadruple-culture containing epithelial, endothelial, 4669 macrophages and mast cells has been established (Moreno et al., 2008; Klein et bal., 2013), co- 4670 cultures of epithelial cells with NK cells (Muller et val., 2013), or epithelial cells, macrophages and 4671 fibroblasts (Mangum et al., 1990). Though these models offer promising alternatives to exhaustive in 4672 vivo testing, experiments comparing the results from the two (in vivo vs in vitro) models are very 4673 scarce. Moreover, some studies suggest that the cellular responses following exposures to 4674 nanomaterials observed in the co-culture models are different from those observed following 4675 exposures of single cell types (Muller et al., 2009; Clift et al., 2011F) and that the results obtained 4676 from the single cell type exposure are more reflective of the in vivo responses than the responses 4677 observed following stimulation of co-culture systems. Regardless, the time and cost-effective in vitro 4678 systems can best be used to screen the potentially hazardous nanomaterials from the inert ones and 4679 the results derived can be used as weight of evidence. It is important to note that by merely culturing 4680 two, three or four different types in a petri, the responses occurring in humans cannot be mimicked. 4681 Thus, the architecture of the in vitro cell co-culture model in regards to the specific organ they 4682 represent is essential when nanomaterial effects are studied. 4683 4684Air -liquid cultures

154

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4685The cultivation of epithelial cells on permeable supports enables the culture medium to be separated on 4686 either side of the cultured epithelium leading to an increased differentiation of the cultured cells 4687 (Methods Enzymol 1989, 171: 736-44). Furthermore, the medium can be removed from the upper 4688 side to expose the cells to air on one side, allowing the cells to ‘feed’ from the medium in the 4689 chamber underneath (Voisin et al., 1977). The air-liquid culture technique has been described in 4690 different cell culture models (Ehrhardt et al., 2002; Ritter et al., 2003; Herzog et al., 2013; Chortarea 4691 et al., 2015). Air-exposed cell cultures allow studying the interaction of inhaled nanomaterials with 4692 cells in an environment that more closely mimics the in vivo situation. Of particular importance is 4693 that the cells are covered by a very thin liquid lining layer with a molecular surfactant film at the air- 4694 liquid interface since surfactant plays an important role in particle displacement and retention (Gehr 4695 et al., 1996). It is demonstrated that A549 cells (Blank et al., 2006), the bronchial epithelial cell line 4696 16HBE14o and Calu-3 cells can be exposed to air (Ehrhardt et al., 2002). In these studies the air- 4697 exposed cultures exhibited a clear epithelial morphology and integrity as in in situ conditions. Such in 4698 vitro cell systems combined with various air-liquid exposure systems that allow a dosimetrically 4699 accurate delivery of aerosolized nanomaterial offer a reliable method for the investigation of 4700 nanomaterial-cell interactions and possible cellular responses (for reviews see Paur et al., 2011). 4701 4702 47036.5.4 Ex vivo precision cut lung slice method to investigate nanomaterial induced tissue responses. 4704As outlined above, the establishment of in vitro systems that can accurately simulate the responses of an 4705 intact organ following exposure to toxicants continues to be a challenge. While monolayer culture or 4706 co-culture models have advanced our understanding of the cellular effects induced by nanomaterials, 4707 they lack the sohpisitcation of the intact organ system. Lung is a multicellular organ and it is unlikely 4708 that the use of few cell types alone can accurately reflect the co-ordinated responses of several 4709 different cell types in the lungs following chemical insult. The precision cut live tissue slices (PCTS) 4710 method represents miniaturized organ models from which they are prepared. In contrast to single 4711 cell type in vitro cell cultures or co-culture of select cell types, PCTS maintain original tissue 4712 architecture with relevant structural and functional features, and cell-cell interactions. PCTS can be 4713 prepared from different species. The ready application of this method to human tissues makes it 4714 specifically interesting and provides a link between animal derived data and extrapolation to human 4715 relevance. This technique is being used to investigate the lung responses to nanomaterials (Sewald 4716 and Braun, 2013; Halappanavar et al., 2014). Similar to the above in vitro techniques, this method 4717 can be used for the purposes of screening potentially harmful nanomaterials. 4718 47196.6. Pre-testing considerations and methods applicable to inhalation route (respiratory system)

155

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

47206.6.1 In vivo pre-testing considerations 4721Inhalation route of exposure applies to testing of immunostimulatory or immunosuppressive effects of 4722 nanomaterials on the respiratory tract. Inhalation is the primary route of exposure to the majority of 4723 NMs and has been extensively characterized. The entire respiratory tract should be investigated 4724 including nose, larynx, trachea, lymph nodes, lung lobes, regional lymph nodes and the pleura for 4725 different endpoints. 24 hours and 28 days post-exposure time points are recommended for sampling 4726 and standard OECD guidelines should be followed (OECD TG for NM). 4727 47286.6.1.1. Mode of lung deposition 4729Experimental approaches for assessing hazards posed by airborne particles use several methods to deposit 4730 particles in the respiratory tract; whole body or nose only inhalation, intratracheal instillation or 4731 pharyngeal aspiration. While inhalation method delivers the intended dose slowly over a period of 4732 time spanning hours, weeks or months, instillation or pharyngeal aspiration methods deliver the 4733 dose at a very high rate within half a second. This difference alters the rate at which the dose is 4734 deposited, the total deposited dose and regional distribution, all of which influence the final 4735 outcome of the exposure (Oberdorster et al., 2015; Kreyling et al., 2013). Where possible, it is 4736 recommended to obtain information on lung deposition and distribution in the body and relate it to 4737 the observed toxicity. For example, a part of the tissue exposed to NM can be set aside for ICP-MS 4738 analysis for elemental detection. This enables effective evaluation of the dose-response, taking into 4739 account how much of the exposed material actually reaches the alveoli and other organs. However, 4740 there are limitations - ICP-MS is only suitable for materials for which the elements do not have a high 4741 background in the body. In addition, it is important to be aware that ICP-MS only analyses elements 4742 and not particles. New techniques such as SP-ICP MS are being evaluated for their suitability to 4743 determine the delivered and retained dose. Nanotubes have been shown to cause a spectrum of 4744 pulmonary effects following lung deposition. However, studies have shown that these effects vary 4745 depending on the mode of exposure (Shvedova et al., 2008). For example, pulmonary inflammation, 4746 lung tissue damage and granulomas are observed in rodents exposed to CNTs administered via 4747 intratracheal administration. A direct comparison of lung effects following exposure to same dose 4748 and type of CNTs via intratracheal instillation or inhalation ways showed no lung lesions in mice 4749 exposed via inhalation (Silvia et al., 2014; Li et al., 2007). It is suggested that inhalation of uniformly 4750 dispersed CNTs causes systemic immune responses such as immune suppression observed in spleen 4751 (Mitchell et al., 2009) . 4752 47536.6.1.1.1 Inhalation:

156

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4754Whole body and nose only are the two ways by which animals are exposed to respirable substances. The 4755 former requires housing of animals individually, allowing free movement throughout the procedure. 4756 In the case of latter, animals are restrained in a narrow tube with their noses protruding into the 4757 plenum where they are exposed to aerosols. However, it should be noted that whole body exposure 4758 will lead to exposure to nanomaterials from other routes such as exposures to GI tract. On the other 4759 hand, nose-only exposures may lead to stress due to restraining the animals, which may particularly 4760 be important for immune toxicity testing as immune responses are altered during stress. Increased 4761 stress levels and loss of body weight or retarded gain in the body weight is observed in long term 4762 nose-only exposure to other substances (Phalen 2009; Coggins et al., 1993; Pauluhn and Mohr, 1999; 4763 Coggins et al., 2011; Rothenberg et al., 2000;) but could be reduced by sufficient training of the 4764 animals prior to exposure of the substance, or selecting a different strain. Since the issues related to 4765 delivery of nanomaterials via either of the methods are common to chemicals as well, general 4766 recommendations outlined in TG 403 and GD 39 should be followed. Particle size determination for 4767 all aerosols is a must. It is not possible to produce aerosols consisting of single nanoparticles. 4768 Nanoparticles will agglomerate and hence the average size in the aerosol will be much larger. For 4769 chemicals, it is suggested that mass median aerodynamic diameters (MMAD) ranging from 1-4 µm 4770 with a geometric standard deviation of 1.5 to 3.0 are recommended. However, for nanomaterials, 4771 aerosol dispersion should be as small as possible. The aerosol preparation for nanomaterials should 4772 be < 3 µm MMAD. Use of ventury for aerosolization of powder and further deagglomeration of 4773 particles in the aerosol with a jet mill are recommended. Electrospray, spark generators, atomizing a 4774 sufficiently diluted dispersion, evaporation and subsequent condensation of metal nanoparticles are 4775 some other methods that can help keep the aggregation and agglomeration of nanomaterials in the 4776 aerosol to a minimum. Because the formation of aggregates and agglomerates directly correlates 4777 with particle resident time and the number of particles in air (per cm3), aerosol should be produced 4778 as close to the breathing zone of the animal as possible. Aerodynamic particle size in the range of 0.3 4779 – 3 µm are attainable. OECD (ENV/JM/MONO(2012)40) guidance on sample preparation and 4780 dosimetry for the safety testing of manufactured nanomaterials provides a list of techniques that can 4781 be used for generation of exposure aerosols. (ENV/JM/MONO(2012)). The exposure dose should 4782 consist of maximum tolerated dose on the high end and human relevant doses on the low end of the 4783 dose spectrum. The maximum tolerated dose for nanomaterials may depend on the maximum 4784 dispersability attainable. The reporting should be based on count median aerodynamic diameter as 4785 mass-based dose metrics may not be suitable for nanomaterials. Use of SMPS and ELPI that enable 4786 single particle counting are suitable for the analysis (aacording OECD guidelines for inhalation 4787 exposures). In addition to particle counts and particle size, particle morphology in the aerosol should 4788 also be investigated. The particle collection equipment should enable collection and classification of

157

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4789 the entire range of particle sizes present in the inhalation chamber. The actual concentration of 4790 nanomaterial in the breathing zones of inhalation system has to be measured and reported. When 4791 mechanical processes are used to obtain certain particle dispersion, aerosol should be characterized 4792 to ensure that the particle has not been transformed during the process. Dry powders are preferred 4793 for aerosolization. In cases where nanomaterials are suspended in liquids prior to aerosolization, 4794 further characterization of particles in liquid suspension and aerosol is necessary to ensure that the 4795 primary particle properties are not altered. If vehicle other than water is used for suspension, gas 4796 chromatography like techniques should be used to determine the concentration of vehicle in the 4797 atmosphere of the chamber. A detailed guidance on exposure chamber conditions are provided in 4798 the inhalation toxicity guidelines including ISO10801 (2010) Nanotechnologies – Generation of metal 4799 nanoparticles for inhalation toxicity testing using the evaporation/condensation method. The 4800 exposure atmosphere should be maintained at constant as much as possible following the guidelines 4801 described in the OECD inhalation toxicity tests. Dry powders represent physiological scenario of 4802 exposure to these materials in the environment or at the occupational setting and thus, the data 4803 derived from such studies are relevant to risk assessment. The results from studies using aerosols 4804 generated from the suspended nanomaterials can inform hazard driven approaches. 4805 48066.6.1.1.2 Intratracheal instillation 4807While inhalation is the ‘gold standard’ and constitutes a physiological route of delivery of nanomaterials to 4808 the respiratory system, because of the high aggregating/agglomerating behaviour of nanomaterials, 4809 uniform aerosol preparation is very difficult, and it is not easy to characterize the aerosol in the 4810 exposure chamber and estimate the deposited dose. In addition, special expertise required to build 4811 and operate the inhalation system, amount of test material required to generate sufficient 4812 concentrations of aerosols in the exposure chamber over a long span of time or highly toxic nature of 4813 the test material preclude the routine use of this method for lung exposure (Driscoll et al., 1999). 4814 Intratracheal instillation is used as an alternative to the cumbersome inhalation procedure, which by- 4815 passes the upper respiratory system and administers a bolus amount of nanomaterials directly into 4816 the tracheobronchial and alveolar regions of the respiratory tract of animals under anesthesia. 4817 Compared to inhalation, the delivered or deposited dose by this method is more uniform, precise 4818 and reproducible. Although, it has been argued that lung responses to nanomaterials administered 4819 via inhalation or instillation may be different, several studies have shown that lung responses to 4820 particles deposited via both methods are consistent for identical nanomaterials. Thus, results 4821 obtained by these techniques can be used to rank the pulmonary hazards of nanomaterials and allow 4822 qualitative risk assessment (Oberdorster et al., 2015). These results can also be used as proof-of-

158

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4823 principle studies to generate hypothesis and to provide weight of evidence to support inhalation 4824 study results. 4825 48266.6.1.1.3. Pharyngeal or laryngeal aspirations 4827This involves deposition of suspended nanomaterials on the back of the tongue and pulling of the tongue, 4828 which results in gasping as part of reflex mechanism leading to aspiration of deposited 4829 nanomaterials. In comparison to instillation, pharyngeal aspiration is supposed to aid in uniform 4830 distribution of nanomaterials in lungs. However, comparisons of lung responses following exposure 4831 via instillation or aspiration are consistent. 4832 48336.6.2 In vivo testing methods applicable to inhalation route (respiratory system) 48346.6.2.1. Body weights 4835Body weights should be monitored weekly and if they remain unchanged in the first week, bi-weekly 4836 monitoring should be carried out. If changes in the body weights are observed, food consumption 4837 should be monitored. Specific to inhalation exposures, loss of body weight or inability to gain body 4838 weight has been observed in some cases where nose-only technique was used. The animals’ weight 4839 has been linked to immunological effects of nanomaterials (de Jong et al., 2013). The 4840 recommendations made by the OECD TG 412 are applicable for nanomaterial testing; the individual 4841 animal weight is recorded on day 0 and at least once weekly thereafter for the duration of the study 4842 and before the necropsy. 48436.6.2.2 Clinical pathology 4844For most of the respirable nanomaterials thus far investigated, the results of the standard clinical 4845 pathology parameters (OECD 412) have not revealed significant abnormalities. Thus, it needs to be 4846 discussed if clinical pathology testing should be a routine for all the nanomaterials or only for those 4847 that showed positive effects on the respiratory system. 4848 48496.6.2.3 Bronchoalveolar lavage fluid (refer to GD 39 for details on the test) 4850The airway epithelium is the target surface for any inhaled substances including nanomaterials. The fluid 4851 lining the epithelium holds a wealth of information concerning the toxic potential of inhaled 4852 substances, and thus, is routinely lavaged with isotonic buffer to examine its contents. Several 4853 biochemical and cytological constituents including types of infiltrating cells, enzymatic activity, 4854 cytokines and chemokines, as markers of lung immune responses are quantified by examining the 4855 BAL fluid. Thus, BAL fluid assessment can aid in developing dose-response of a substance, to rank a 4856 substances’ potency and to set up no effect level of exposure for regulatory purposes. Since all 4857 nanomaterials, due to their small size are expected, and in many cases, demonstrated to be

159

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4858 translocated to alveoli in the lower respiratory tract upon inhalation, bronchoalveolar lavage fluid 4859 assessment is recommended as a mandatory test for NMs (discussed ENV/JM/MONO(2012)40 and 4860 also in OECD inhalation TG for NM). Temporal changes in the BALF constituency can be prognostic of 4861 initiation, progression and/or severity of a lung immune disease, and inform the potential 4862 mechanisms involved (Wan-Seob et al., 2010). 4863 48646.6.2.3.1 BAL total cell counts, differential count 4865In normal situations, macrophages constitute 90% of the total cell population of BALF. Following exposure 4866 to lung toxicants, depending on the potency of a substance, influx of neutrophils lymphocytes and 4867 eosinophils is observed, all of which imply different degree of lung inflammation (Wan-Seob, 2010). 4868 Thus, measurement of BAL differential cell count provides information concerning the extent of 4869 pulmonary inflammation and possibly the type of inflammation induced by NMs. Ideally a separate 4870 set of animals with whole lung lavage is recommended for BAL assessment. BAL total and differential 4871 cell counts are scored using an ordinary light microscope. Macrophages, polymorphic nuclear 4872 granulocytes, lymphocytes and other cell type such as eosinophils, epithelial cells are scored. The 4873 detailed method is provided in OECD GD 39. 4874 48756.6.2.3.2. BAL cytotoxicity assay 4876Traditionally, Lactate dehydrogenase (LDH) release into the surrounding medium is measured as an 4877 indication of loss of plasma membrane integrity and consequent cellular toxicity. The amount of 4878 cellular damage is directly proportional to the amount of LDH release. This assay has been employed 4879 to investigate nanomaterial induced cytotoxicity (Gaudgagnini et al., 2015). However, several 4880 nanomaterials including titanium dioxide, silver, silica, iron oxide, cerium oxide, CdSe, zinc oxide, 4881 Qdots, PS, copper, SWCNT, CB, C60 and gold are shown to either increase the absorbance generating 4882 false positive results or quench the fluorescence, inhibit the LDH enzyme activity, or increase the 4883 enzyme adsorption thereby generating false negative results (Kroll et al., 2011). Thus, the assay must 4884 be used with appropriate controls; the incubation of nanomaterials with the different components of 4885 the assay in the absence of cells or the cells alone in the absence of any nanomaterials. Addition of 4886 BSA to the nanomaterial preparation has been suggested to minimize enzyme adsorption. 4887 Interference with the assay components can differ based on the nanomaterial doses used; a series of 4888 nanomaterial doses have to be tested with assay components without the cells to eliminate the 4889 doses at which the interference occurs. For nanomaterials such as CNTs, CB and Qdots, the 4890 interference comes from their colour, which may be unavoidable. In such cases, multiple assays that 4891 use different detection methods are useful (Xianglu et al., 2011). Low nanomaterial concentrations 4892 are recommended for methods using absorbance method of detection.

160

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4893 4894Several studies have shown that the positive cytotoxicity results for BAL cells are transient and thus, the 4895 results must accompany histopathological assessment of lung sections (Rahman et al., 2016). The 4896 measurement of total protein in BAL fluid (supernatant without cells), increases of which, is a good 4897 indicator of lung alveolar/capillary barrier permeability state, can also be used to support the LDH 4898 cytotoxicity assay. However, there is no consensus on whether this should be a routine test (OECD TG 4899 for NMs). Bradford protein quantification method is considered suitable for assessing total protein in 4900 BAL fluid. 4901The other recommended measurements for determining the cytotoxicity in BAL include Alkaline 4902 phosphatase (indicator of secretions from Type II cells), gamma-glutamyltransferase (a membrane 4903 bound surface enzyme that plays a role in the metabolism of xenobiotics and leukotrienes; is lost 4904 during lung injury) and N-acetyl-B-glucosaminidase (it is a lysosomal enzyme secreted by 4905 activated/macrophages that have phagocytosed particles, increased levels of this enzyme indicates 4906 activation or lysis of phagocytes and thus lung injury) assessment. 4907 49086.6.2.3.3. BAL cytokine and chemokine profile 4909In addition to cells and enzymes, activated epithelium and macrophages secrete additional factors such as 4910 growth factors, chemokines, and cytokines. Profiling of these factors can provide information 4911 concerning the sources of infiltrating cells and the underlying mechanisms of lung responses. Unique 4912 cytokine and chemokine patterns are associated with specific disease types and are helpful in 4913 discriminating the inflammatory, immunosuppressive and allergic responses to nanomaterials 4914 (Elsabahy and Wooley, 2013). Multiplex ELISA, single chemokine/cytokine specific ELISA, single or 4915 multiplex RT-PCR and Western Blot are the routinely used methods for assessing the cytokines and 4916 chemokines. Care must be taken during the preparation of protein or RNA preparation. Particles can 4917 interfere with the electrophoresis (Western blots) and absorbance reading. 4918In conclusion, the BAL analysis following inhalation exposure to nanomaterials is a mandatory assessment 4919 and the existing methods for chemicals are readily applicable to nanomaterials conditional to 4920 inclusion of appropriate controls. 4921 49226.6.2.4. Lung tissue cytokine and chemokine profiles 4923Several nanomaterials are shown to induce changes in expression of genes and proteins associated with 4924 inflammatory process in the lung tissue. These include several cytokines, chemokines and growth 4925 factors (Husain et al., 2013). The changes in the tissue expression of these molecules are directly 4926 correlated with the extent of lung inflammation observed. Moreover, studies have suggested that 4927 magnitude of lung inflammation following exposure via inhalation route is a good indication of

161

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4928 likelihood of occurrence of systemic effects (Poulsen et al., 2015; Saber et al., 2013). In some 4929 instances, elevated levels of cytokines and chemokines are observed in the absence of or receding 4930 infiltrating inflammatory cells, suggesting that tissue profiling may be a sensitive measure to monitor 4931 long term immune effects of nanomaterials (Husain et al., 2013). Targeted single gene and protein, 4932 custom low-throughput assays such as pathway-specific PCR arrays and multiplex ELISA or higher 4933 throughput global analysis of gene (Halappanavar et al 2011; Halappanavar et al., 2015; Nikota et al., 4934 2016; Rahman et al., 2016;) and protein expression profiles using microarray or proteomic methods 4935 are also applied to nanomaterials (Teeguarden et al., 2011;). The high throughput techniques have an 4936 advantage of revealing the unexpected or previously unknown immune effects of nanomaterials and 4937 the underlying mechanisms of such effects. 4938 49396.6.2.5. Cellular damage 4940Activated phagocytic cells release reactive oxygen species (ROS) upon internalization of pathogens or 4941 particles, resulting in respiratory burst that can be measured in macrophages, monocytes and 4942 neutrophils derived from BALF cells. A fluorimetric assay that relies on the intracellular oxidation of 4943 5- and 6-carboxy-2’,7’-dichlorodihydrofluorescein diacetate (carboxy H2DCFDA) (Molecular Probes) is 4944 currently used in some studies (Decan et al., 2016); however, it is important to note that the results 4945 are not specific and is shown to detect various types of radicals. In addition, other comprehensive 4946 methods such as in vitro and in vivo lipid peroxidation, protein oxidation, and protein carbonylation 4947 using proteomics techniques Riebeling et al., 2001) can also be used to assess oxidative stress. 4948 Intracellular glutathione levels can be measured using the ThiolTracker™ Violet assay (Decan et al., 4949 2016) and an assay that measures glutathionylation of proteins. Oxidative stress can also be 4950 measured by assessing the antioxidant pathways involving measurement of relevant genes and 4951 proteins (Hasse et al., 2016). Other biomolecule modifications such as nitrosylation, reflective of 4952 oxidative stress, can be measured by measuring nitrosylated tissue proteins, or increase in NO 4953 production, and nitrate/nitrite ratio in BAL. In addition to tissue analysis, acellular glutathione levels, 4954 antioxidants and NO production in BAL supernantant can be used to assess ROS synthesis. 4955 49566.6.2.6. Cellular proliferation 4957Labeled nucleotide precursors, such as tritiated thymidine and 5-Bromo-2deoxyuridine (BrDU) or via 4958 immunohistochemistry of lung tissue by staining for proliferating cell nuclear antigen (PCNA) can be 4959 used to assess increases in cell division of epithelial cells, an indication of excessive cell growth or 4960 hyperplasia. The methodology is the same as the ones used for chemicals, which includes pulsing of 4961 animals in the experimental groups with BrDU or tritiated thymidine. For immunohistochemistry, 4962 lung sections are reacted with antibodies against PCNA. The other cellular proliferation indicators

162

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4963 such as Ki67 have also been used to assess proliferation of cells in lungs of mice exposed to MWCNTs 4964 (Rahman et al., 2017). 4965 49666.6.2.7. Pleural lavage 4967Although not compulsory, for high aspect ratio NMs such as fibers, tubes and plates, a pleural lavage 4968 should be conducted (OECD recommended, (discussed ENV/JM/MONO(2012)40 and also in OECD 4969 inhalation TG for NM)). 49706.2.8 Tissue Histopathology 4971Cellular influx and consequent lung inflammation and immune responses induced by nanomaterials are 4972 often transient, and depend on the exposure dose and inherent properties of nanomaterials. Thus, in 4973 addition to the detailed assessment of BAL fluid, a thorough histopathology should be conducted to 4974 determine the site of nanomaterial deposition, the type of inflammatory lesions and more 4975 importantly, the sustainability of the effects. Histopathology should be conducted on a separate set 4976 of animals that are not used for BAL collection. A whole lung assessment is recommended for bio 4977 distribution assessment, and formalin fixed, hematoxylin and eosin (H&E) stained left lung lobe is 4978 used for the histopathological analyses (according to ENV/JM/MONO(2012)40). In addition to the 4979 morphological and pathological assessment, immunohistochemistry and special stains can be used to 4980 identify specific cell types, cellular apoptosis, alveolar thickening, and disease phenotypes (collagen 4981 deposition in fibrosis) (Rahman et al., 2017). 4982It is important to note that particles when stained with H&E staining appear darker under the microscope. 4983 Some particles may interfere with the fluorescence antibodies quenching or increasing the 4984 fluorescence. These factors must be kept in mind and appropriate controls (suggested under BAL 4985 cytotoxicity section) must be included in the study to avoid misinterpretation of the results. 4986Particles deposited in the lung are translocated to regional and systemic lymph nodes and extrapulmonary 4987 immune responsive tissues. Thus, a detailed analysis of particle deposition in spleen, thymus, local 4988 lung lymph nodes and of liver must be conducted and this should be mandatory (Oberdorster et al). 4989 Novel microscopic techniques such as enhanced darkfield, hyperspectral imaging techniques (Husain 4990 et al., 2015; Decan et al., 2016) are highly sensitive and enable detection of very low levels of 4991 particles lodged in the tissue. This technique does not require any pre-processing of samples and can 4992 be applied to fresh, frozen, stained and unstained particles. Moreover, this technique does not 4993 require pre-labelling of cells or particles with florescence or radiolabeling (Roth et al., 2015). 4994 Consequent to finding particles in these organs, a detailed histopathological analysis of these tissues 4995 must be performed. Thus, dark-field enhanced and hyperspectral microscopic method of detection 4996 can aid in prioritizing lung samples for a detailed histopathological investigation. In the repeat dose

163

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

4997 study, histopathology of the parietal pleura and the subpleural proliferation of lung tissue should be 4998 included for fiber-like nanomaterials. 4999 50006.6.2.9 Immunotoxicity relevant pathology and organ weights 5001Nanomaterials are suggested to translocate to various organs from the target lung tissue exposed. Since 5002 nanomaterials are mostly poorly soluble, as suggested in TG 412, draining lymph nodes from the hilar 5003 regions of the lung should be examined with regards to weight, cellularity and cell composition. In 5004 addition, nasopharyngeal tissues should be investigated. Also, a detailed investigation of other 5005 tissues for translocation and/or associated immunotoxicity should be performed qualitatively using 5006 microcoscopic methods and quantitative analytical measures and immunotoxicity endpoints should 5007 be performed. Lung, liver spleen, thymus, bone marrow, heart, kidney and central nervous system 5008 should be weighed and examined. 5009While gross morphological and pathological changes in these tissues can be assessed by the routine 5010 haemotoxylin and eosin staining (Rahman et al. 2016), immunohistochemistry against cell surface 5011 markers (macrophages, neutrophils, T-helper cells) can be used for precise detection of specific cell 5012 population that is enriched. Frozen tissue sections that are paraffin embedded are better for 5013 detecting cell surface markers, expression of which is scarce. Antigen specific antibodies can also be 5014 used to detect specific enrichment of a gene or enhanced activity of a protein (Rahman et al. 2017). 5015 Where sample is abundantly available, tissue cellularity can also be assessed using flow cytometry. 5016 For example flow cytometry with antigen specific antibodies can be used to sort T and B 5017 lymphocytes, stem cell population of the bone marrow, etc. Morphometric analysis of the 5018 histological tissue sections can also provide information on tissue distribution and localization of NM. 5019Relative effects on the T and B cell population in spleen, atrophy of the spleen and lymph nodes in case of 5020 immune suppression and lymphoid tissue hypertrophy shoud be assessed in case of immune 5021 stimulating nanomaterials. Due to the possibilities of clearance of particles from lung tissue to 5022 instestine via mucociliary transport, cellular proliferation in Peyer’s patches of small intestine should 5023 also be investigated for potential indication of elevated immune response. 5024Morphological analysis of the bone marrow containing multipotent stem cells that is capable of 5025 differentiating into B and T lymphocytes and macrophages. This can be done using bone marrow 5026 smears or by cytospin preparations of bone marrow cells. Cellular viability and bone marrow cellular 5027 differential counts are good indicators of chemical induced effects on the immune system. 5028 50296.6.3 In vitro testing methods applicable to inhalation route (respiratory system) 5030Love et al. (2012) and Fröhlich et al. (2015) summarise various in vitro toxicity testing methods available 5031 and experimental details to consider for assessing nanomaterial-induced toxic responses.

164

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5032 50336.6.3.1 Immune cells damage 50346.6.3.1.1 Cellular viability 5035Cytotoxicity to immune cells can be measured by the LDH assay as discussed above (6.6.2.3.2.). In addition, 5036 other cellular viability assays such as, MTT, WST1, neutral red, almar blue, and trypan blue exclusion 5037 methods have been routinely employed to assess this endpoint. However, as mentioned for the LDH 5038 assay, care should be taken to assess the potential interference of nanomaterials with the test 5039 reagents (Park et al., 2009; Stone et al., 2012). Appropriate assay and particle controls have to be 5040 included in the assay. 5041 5042 5043Oxidative damage is another type of cellular damage that is frequently assessed following exposures to 5044 nanomaterials. The methods and considerations are described in 6.6.2.5. Fibre-like nanomaterials 5045 have been shown to affect the phagocytic function of macrophages. Depending on shape, aspect 5046 ratio and biopersistency, these materials can induce frustrated phagocytosis (Donaldson et al., 2010; 5047 Brown et al., 2007). For such nanomaterials, ex vivo phagocytosis assay should be performed. 5048 50496.6.3.1.2 Antigen uptake, processing and presentation 5050The main function of dendritic cells is to capture antigens (i.e., the innate immune system), process them 5051 and present them to T-cells via receptor mechanisms (i.e., adaptive immunity) in order to initiate an 5052 immune response. The potential of dendritic cells to take up and process antigen following exposure 5053 to NM can be evaluated using ovalbumin (OVA) coupled to Alexa-647 and bovine serum albumin 5054 (BSA) coupled to DQ-Red to analyze antigen uptake and antigen processing, respectively (Garnier et 5055 al. 2007; Blank et al., 2011). Considering the co-stimulatory expression, the expression of surface 5056 phenotype markers (CD11c, CD11b) or activation markers (CD40, CD86 and MHCII) and the release of 5057 broad range of cytokines / chemokines can give important information in terms of measuring the 5058 activation status of DCs (Satpathy et al. 2012; Von Garnier et al. 2009). 5059 50606.6.3.1.3 Activation of T-cells 5061Following interactions with DCs, effector T-cells can be activated and differentiate towards different fates, 5062 depending on the signal from the DCs (antigenic specificity). A different cytokine secretion pattern is 5063 a common characteristic of the select differentiation pathways. Clonal expansion induced by 5064 proliferation signal then follows (Masopust et al. 2013; Holt et al. 2008) and can be determined in 5065 vitro by utilizing an autologous CD4+ T cell stimulation assay as described (Gerber et al. 2011). 5066

165

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

50676.6.3.2 Ex vivo phagocyte function assay 5068There are no guidelines or validated methods for conducting phagocyte function assay; however, various 5069 in vitro phagocytes function tests that are currently available (Ricevuti et al. 1993) and their 5070 importance to screening potentially immunotoxic nanomaterials are reviewed in Fröhlich (2015). A 5071 panel of in vitro assays including cytokine secretin, chemotaxis, phagocytosis and respiratory burst 5072 assays can be used to measure phagocyte function. Recently, a rapid screening method to evaluate 5073 the interaction of nanomaterial with phagocytic cells has been proposed (Mottas et al. 2017). The 5074 conventional ex vivo phagocytic assays require fluorescence tagging of test substance, which is not 5075 suitable for nanoparticles as it may influence cellular interaction and uptake of NMs. A luminescence 5076 based approach has been tested for NMs. Although the method has only been tested using in vitro 5077 immortalized monoculture models, it can be applied to ex vivo models as well. BALF cells from 5078 animals are retrieved and are incubated with the test particles and luminol solution. The phagocytic 5079 activity is visualized with luminol, a dye which becomes luminescent only upon exposure to the low 5080 pH of the phagolysosome. BALF cells can also be incubated with particles, sorted using flow 5081 cytometry and then scored for specific cell types that contain particles using luminol method. 5082For auto fluorescing particles such as quantum dots, BALF cells from animals exposed to vehicle alone or 5083 NMs are retrieved and are challenged with various doses of fluorescent particles and allowed to 5084 internalize particles. Uptake of particles is then evaluated by confocal microscopy. As described 5085 above, cell type specific phagocytosis can be scored by sorting the cells by flow cytometry prior to 5086 incubating with particles and visualizing by confocal microscope. For details on the various aspects of 5087 phagocyte function assay, plese refer to Fröhlich (2015). 5088 50896.6.3.3 Phagocytic activity of granulocytes, monocytes and phagocytic respiratory burst 5090Diluted human heparinized whole blood is reacted with nanomaterials for different duration. Post- 5091 exposure, samples are reacted with hydroethidine solution for the assessment of respiratory burst 5092 and fluorescein labelled S aureaus bacteria for the assessment of ability of NM exposed phagocytes 5093 to ingest bacteria. Samples are tested by flow cytometry. This is fairly a simple assay but additional 5094 steps have to be added to make sure that the nanomaterials do not interfere with the fluorescein 5095 reading as well as the instrument (Fröhlich et al. 2015). 5096 50976.6.3.4 Lymphocyte proliferation 5098Leukocyte proliferation aids in the host defense response to immunogenic substances and mitogens 5099 stimulating cell proliferation and division, respectively. Human blood cells are reacted with NMs in 5100 vitro or spleen or blood lymphocytes derived from animals exposed to NMs are examined for 5101 proliferative activity of lymphocytes (Ghoneum et al. 2014). Cells in vitro or cells derived from

166

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5102 animals exposed to NM are reacted with mitogens and cell specific antigen before pulsing with 5103 tritiated thymidine. An increase in radioactivity incorporation is measured as an indication of cell 5104 division and differentiation of lymphocytes. 5105 51066.6.3.5 Other immune response related assays 51076.6.3. 5.1 Respiratory sensitization 5108With regards to respiratory sensitization, conventional substances are classified as respiratory sensitizers 5109 ad hoc, on the basis of human data. Nano forms of the known classified substances should be 5110 regarded as respiratory sensitizers as well. At present, validated experimental models to determine 5111 the respiratory sensitizing potential of chemicals or nanomaterials do not exist. Some assumptions 5112 are excercised in evaluating respiratory sensitization potential of nanomaterials and caution should 5113 be excercised in interpretation of the results 51141) The bulk counter part of a nanomaterial that is classified as a respiratory sensitizer must be treated as a 5115 respiratory sensitizers and prioritized for further testing. 51162) Nanomaterials that are skin sensitizers should be considered respiratory sensitizers, if they are 5117 bioavailable in the lung as nanomaterials. The nanomaterials that are positive for skin sensitizing 5118 tests such as LLNA and GPMT (discussed below) should be further tested for their respiratory 5119 sensitization potential. 51203) Nanomaterials that are translocated from lungs to systemic circulation and other organs and those that 5121 exhibit reactive properties should be prioritised for sensitizer testing. As discussed above, low levels 5122 of systemic translocation has been demonstrated after inhalation or intratracheal instillation for a 5123 number of nanomaterials of various chemical compositions. In general, translocation potential seems 5124 to be higher for smaller size particles (reviewed in Braakhuis et al. ). Thus, in the absence of evidence 5125 showing their translocation, particles with primary sizes below approximately <30nm and 5126 aerodynamic sizes between 10 and 100 nm can be assumed to translocate and should be prioritised 5127 for further testing. Some studies suggest that nanomaterials may act as adjuvants, aggravating 5128 respiratory sensitization in mice models of airway allergic diseases (Brandenberger et al. 2013). 5129 However, no OECD guidelines are currently available to test this, even for conventional substances, 5130 although it could very well be a very relevant endpoint to consider in future. 51316.6.3.5.2 Immune suppression 5132Immune suppression is a result of impaired T cell development and function leading to the toxicity. The 5133 inhalation studies described above (and the other OECD standard toxicity studies via other exposure 5134 routes) focus mostly on direct immune activation. As such, nanomaterials have been extensively 5135 characterized for their immunostimulation behaviors than their immunosuppression. The 5136 immunospsuppressive and anti-inflammatory properties of nanomaterials have been reviewed in

167

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5137 (Ilinskaya and Dobrovolskaia, 2014). Spleen tissue is often investigated for markers of immune 5138 suppression following direct or systemic exposure. For systemic exposures, spleen is harvested in 5139 sterile tubes containing sterile HBSS and homogenized on ice. Splenic lymphocyte subpopulation 5140 analysis including CD4+ and CD8+ T lymphocytes, B lymphocytes and natural killer cells is conducted. 5141 In non-immunized animals this will determine if exposure leads to a shift in the helper (CD+4) or 5142 cytotoxic (CD8+) thymus derived lymphocytes or natural killer cells. 5143 5144 5145 5146 5147 5148 5149 51506.6.7. Testing considerations and methods applicable to dermal route (skin exposure) 5151 51526.6.7.1 Skin epithelium 5153The external surface of the skin consists of a keratinised squamous epithelium, known as the epidermis, 5154 which is supported and nourished by a thick underlying layer of connective tissue referred to as the 5155 dermis, which is highly vascular and contains many sensory receptors (Weather et al., 1979). A major 5156 function of the skin, especially the stratum corneum, which is the most outer layer, is to provide a 5157 protective barrier against the hazardous external environment. The skin is relatively impenetrable to 5158 lipophilic particles larger than 600 Daltons in size, whereas lipophilic particles smaller than this may 5159 passively penetrate the skin (Barry, 2001). In general, it is accepted that only compounds and drugs 5160 smaller than 500 Daltons in size can penetrate the skin readily (Bos 2000). 5161 5162Nanomaterial s have unique physical properties making them ideal for use in various skin care products 5163 currently on the market. Functionalized and/or surface modified metal oxide nanomaterials, 5164 specifically zinc oxide (nanoZnO) and titanium dioxide (nanoTiO2), are the primary nanomaterials 5165 used in sunscreen and skin care products as UV adsorber (Loden et al., 2011). Thus, the skin is 5166 exposed to nanomaterials present in cosmetic products such as moisturisers and sunscreens. The 5167 skin is also a potential target for drug delivery via nano-carriers (Nohynek et al., 2007). 5168NanoTiO2 and nanoZnO formulated in topically applied sunscreen products exist as aggregates of primary 5169 particles ranging from 30 to 150 nm in size. These aggregates are bonded in such a way that the force 5170 of sunscreen product application onto the healthy skin would have no impact on their structure or 5171 result in the release of primary particles. Many studies using skin tissue (which is easily available

168

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5172 from animal slaughterhouses) have also shown that under exaggerated test conditions neither nano- 5173 structured TiO2 nor ZnO penetrate beyond the stratum corneum of skin using the “minipig” species 5174 (Sadrieh et al., 2010). Studies of the translocation of TiO2 nanoparticles in histological skin sections 5175 suggest that these nanoparticles may only penetrate into the ‘horny’ upper layers of the stratum 5176 corneum (Lademann et al., 1999). However, other studies have shown that nano-sized particles can 5177 enter a small percentage of hair follicles and are stored in this location for a prolonged period 5178 compared to their location within the stratum corneum; a factor that may enhance drug delivery by 5179 this route, although it will also exacerbate any potential toxicity (i.e. dose kinetics) (Lademann et al., 5180 2006). While such studies suggest little, if any epidermal or dermal penetration of these 5181 nanoparticles, recent work in live mice and pigs indicates that topically applied nano-sized TiO2 5182 particles (<10 nm) may indeed pass through the stratum corneum (Wu et al., 2009). In addition, 5183 stretched porcine skin was far more susceptible to dermal translocation of a C60 fullerene 5184 substituted peptide, which could reach the intercellular spaces of the stratum granulosum in 5185 stretched skin (Yang et al., 2010). 5186 51876.6.7.2 Preparation of nanomaterials for dermal exposure 5188Engineered nanomaterials can come into contact with skin in various forms including as airborne pristine 5189 particles, aerosolized particles in a liquid vehicle or as ingredients of topical formulations. In order to 5190 clearly understand the dermal immune responses of nanomaterials, the role of the vehicle or other 5191 ingredients in the formulation in which nanomaterials are suspended must be investigated in 5192 partallel with a detailed charactersation of the specific nanomaterial. Most of the dermal studies to 5193 date have been conducted using raw nanomaterials suspended in simple solvents; however, this 5194 method does not reflect the realistic skin application scenario. Thus, investigation of dermal 5195 absorption capacity of nanomaterials contained in the actual consumer products is preferred. 5196 However, simple suspensions of nanomaterials can still be used to provide information on the 5197 potential immunogenicity of these materials on the skin. In this respect, the US FDA has recently 5198 endorsed new safety standards for testing sunscreens, where a lack of information about 5199 interactions between ingredients in sunscreen formulations was identified as a current shortfall 5200 (Sunscreen Innovation Act 5201 [http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/ucm434782.htm]). 5202 52036.6.7.3 In vivo skin pre-testing considerations 5204OECD guideline TG 427 is applicable to nanomaterial dermal testing. Several species including rodents, 5205 rabbits, monkeys and pigs are recommended although pig skin is suggested to be the choice if human 5206 studies are not possible, as it is the closest in mimicking the human dermal penetration rate (Klang et

169

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5207 al 2012). Hairless mice or immune-deficient mice (SCID) have also been tested (Wu et al 2009). 5208 Anatomical origin of skin plays an important role in substance penetration (Jepps et al 2013). Skin 5209 from forearm, leg, back, abdomen or breast are routinely used in the testing. As for chemicals, in vivo 5210 skin penetration studies for nanomaterials should consider age, sex, genetic background, skin 5211 condition of the test animal. The other important factor to consider is the presence of hair on skin 5212 since deposition and potential penetration of nanomaterials is suggested to be highest around the 5213 hair follicles (Filipe et al 2009). These parameters should be promptly recorded and reported since 5214 they are suggested to influence nanomaterial penetration into skin. Pre-treatment of skin by shaving, 5215 depilation and clipping are routinely used for in vivo studies, which are suggested to impact the 5216 absorption through skin (Ravichandran et al 2011). Tape-stripping is also used to remove layers in the 5217 stratum corneum, in cases where correlation with the physical barrier is desired. Although most in 5218 vivo skin studies are carried out in static conditions, skin massage, flexion of skin or animal 5219 movement during the exposure should be considered [Rouse et al 2007]. As suggested in TG 427, 5220 animals should be housed separately to avoid consumption of nanomaterials through oral route 5221 during grooming activities. Use of skin on the back or on the shoulders alleviates the issue with 5222 personal grooming and hence is recommended by TG 427. 5223 52246.7.4 In vitro/ex vivo and synthetic skin models 5225Ex vivo testing consists of applying a test nanomaterial onto the surface of a skin derived from animals 5226 (rodents, pig) or humans in an in vitro set up that mimics the environment (donor chamber) and the 5227 systemic circulation (receptor chamber). Following the exposure to a nanomaterial for the required 5228 amount of time at the specific doses, nanomaterial is wiped or washed off from the skin surface and 5229 the receptor fluid in the receptor chamber is tested for nanomaterial penetration. Receptor fluid can 5230 also be sampled during the experiment at different interval of time to assess penetration. This is 5231 especially important if a nanomaterial is suspected to be unstable in the experimental conditions 5232 tested. Freshly derived skin or frozen skin is used in the testing. In addition, commercially available 5233 models such as the EpiDermTM skin model (www.mattek.com) as well as the EPISKINTM 5234 (www.loreal.com) have been evaluated for corrosivity testing of chemicals and their potential in risk 5235 assessment for nanomaterials needs to be evaluated more thoroughly. 5236 52376.7.5 Skin cell cultures 5238Skin epithelial cell lines, such as the epidermal cell line A431 (Sharma et al., 2009) and the immortalized 5239 human keratinocyte cell line HaCaT (Boukamp et al., 1988) have been used to study the effects of 5240 silver (Zanette et al., 2011), TiO2 (Xue et al., 2010) or SiO2 nanoparticles (Yang et al., 2010). 5241 Alternatively, primary human keratinocytes are physiologically relevant and readily available

170

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5242 commercially (Zhao et al., 2013). However genetic variations may need to be screened depending on 5243 experimental objectives. 5244Since the epidermis is composed of many different cell layers, the optimal in vitro skin model is still lacking 5245 (Yang et al., 2010) and further research needs to be performed. Since complete skin tissue can be 5246 easily obtained from the slaughterhouse, this might be a better tool to study dermal penetration and 5247 dermal immunogenic effects of nanomaterials. 5248In vitro three-dimensional organotypic skin cultures are available and serve as potential alternatives to 5249 using real skin (Katawhala et al., 2013). These are either constituted from single cell cultures of 5250 keratinocytes and fibroblasts, or bought commercially as products of tissue engineering (e.g. 5251 EpiDerm, EpiSkin). Variations of such cultures are possible but they are essentially bilayered cultures 5252 consisting of a dermis and a stratified epidermis above. These cultures emulate the anatomy of real 5253 skin, including a proper stratum corneum, but are void of skin appendages and immune cells. 5254 5255 52566.7.6 Skin test methods for immunotoxicity 5257Depending on the invasiveness of the technique used to test dermal absorption, a special attention should 5258 be paid to potential artefacts as a result of techniques used to prepare the skin sample or the 5259 analysis of samples. For example, cross sections of the skin can be prepared by cutting from dermis 5260 to epidermis. This will prevent transfer of nanomaterials situated on the epidermis through the 5261 cutting knife. 5262 5263Cellular toxicity 5264Cellular toxicity of skin cells is measured using the same methods described under 6.6.3 52656.7.6.1 In vivo Skin sensitization 5266In order for the skin sensitization response to take place, nanomaterials will first have to be absorbed by 5267 the skin and reach the dendritic or other systemic immune cells involved in stimulating sensitization 5268 responses. Not all nanomaterials are anticipated to be absorbed through the skin; some of the 5269 nanomaterial properties may influence their skin absorption potential, therefore, a detailed 5270 characterization of nanomaterials before the exposure is a must. If a nanomaterial does not cross the 5271 skin barrier, it can be safely assumed that they are not good candidates for skin sensitization tests. 5272 52736.7.6.1.1 Human Repeated Insult Patch Test 5274Human Repeated Insult Patch Test (HRIPT) is the most relevant test for assessing skin sensitizing potential 5275 of nanomaterials. A test article is applied repeatedly to the skin of healthy human volunteers and 5276 markers of skin sensitization such as skin redness is observed. However, for ethical reasons, this

171

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5277 method cannot be used as a predictive tool to test substances with unknown dermal sensitization 5278 potential. These tests are usually only performed to confirm the safe use of potentially sensitizing 5279 substances in consumer products by confirming the dermal sensitization NOEL that has been 5280 obtained from dermal sensitization quantitative structure–activity relationships, animal pre-clinical 5281 data, and historical human data (Politana, 2008). There is no validated regulatory guideline for the 5282 HRIPT, although a review of the HRIPT design and the critical factors that can affect the induction of 5283 dermal sensitization can be found in McNamee (2008). 5284 52856.7.6.1.2 Guinea Pig Maximization Test and the Buehler Test 5286A number of OECD guidelines exist to address skin sensitization for conventional chemicals. OECD 406 5287 (adopted in 1981, revised in 1992) describes both the Guinea Pig Maximization Test (GPMT) and the 5288 Buehler Test (BT). In both tests, guinea pigs are first exposed to the test substance by intradermal 5289 injection and/or epidermal application (induction exposure). Following a rest period of 10 to 14 days 5290 (induction period) to allow for the development of possible immune responses, the animals are 5291 exposed to a challenge dose via topical application (challenge exposure). The degree of skin reaction, 5292 based on erythema and swelling, is recorded over time and compared with that in sham treated 5293 control animals. The GPMT differs from the BT in that the former uses adjuvant while the latter does 5294 not. 5295 52966.7.6.1.3 The Local Lymph Node Assay 5297The Local Lymph Node Assay (LLNA, OECD 429) is recommended to assess the potential of chemicals to 5298 induce sensitisation as a function of lymphocyte proliferative responses in regional lymph nodes in 5299 mice. It works on the basic principle that lymphocyte proliferation in the local lymph nodes will be 5300 increased proportionally to the dose and potency of the applied sensitizer. Briefly, the test substance 5301 is applied to the dorsum of each ear over the first 3 days. On day 6, the animals are injected with 20 5302 μCi of 3H-methyl thymidine or 2 μCi of 125I-iododeoxyuridine and 10-5M fluorodeoxyuridine through 5303 the tail vein, and sacrificed 5 hours post-injection. Single-cell suspensions of lymph node cells are 5304 then prepared and their proliferation quantified by 3H or 125I counting. A stimulation index, 5305 calculated as the ratio of the proliferation in treated groups to that in the concurrent vehicle control 5306 group, is used for comparing the extent of sensitization. Other observations such as ear erythema or 5307 ear thickening may be included. OECD guidelines for non-radioactive modifications of the LLNA have 5308 been adopted in 2010. The LLNA:DA test (OECD 442A) quantifies adenosine triphosphate (ATP) 5309 content via bio-luminescence as an indicator of lymphocyte proliferation, while the LLNA BrdU test 5310 (OECD 442B) utilises non-radiolabelled 5-bromo-2-deoxyuridine (BrdU) in an ELISA-based test 5311 system to measure lymphocyte proliferation. OECD recommends that LLNA (OECD 429) be used as

172

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5312 the preferred method where an in vivo test is necessary, due to animal welfare benefits and the 5313 provision of more quantitative data, compared to GPMT and BT (OECD 406). 5314Since the absorption of nanomaterials through the skin and their systemic availability is essential for a 5315 dermal sensitization reaction to occur, the standard experimental animal assays such as the OECD 5316 guidelines for the GPMT, BT and LLNA become relevant only if it has been demonstrated that the 5317 nanomaterials are absorbed through the skin. However, precaution has to be taken not to 5318 underestimate the potential since it is not entirely known if negative dermal absorption results from 5319 animal (mice, pigs) testing can be extrapolated directly to human situation. For many nanomaterials 5320 it will be difficult to assess skin absorption in these tests, for example, due to high background levels, 5321 which complicates the assessment of the sensitizing potential of NMs with animal models. 5322 53236.7.6.2 In vivo skin irritation test 53246.7.6.2.1 The HRIPT test 5325The HRIPT test discussed earlier for skin sensitization can also be used to assess skin irritation through 5326 repeated patching during the induction phase, but similar to sensitization, this test method cannot 5327 be used as a predictive tool. 5328 53296.7.6.2.2 OECD 404 (adopted in 1981, first revision in 1992, second revision in 2002) 5330In this test, young adult albino rabbits are used. The substance to be tested is applied as single dose over a 5331 shaved area of approximately 6 cm2 on the dorsal area of the trunk of the test animals. The exposure 5332 period is normally kept at 4 hours. The degree of irritation/corrosion is scored for signs of erythema 5333 and oedema, at 60 minutes and then at 24, 48 and 72 hours after removal of the substance, in order 5334 to evaluate the reversibility or irreversibility of the effects observed. This timeframe should be 5335 extended up to 14 days if there is damage to skin which cannot be identified as irritation or corrosion 5336 at 72 hours. Untreated skin areas of the test animal serve as the control. All local toxic effects (e.g. 5337 skin defatting) and any systemic adverse effects (e.g. loss of body weight) should also be recorded. 5338 Histopathological examination will be useful especially for comparing cases which are ambiguous. 5339 Effort should be made to harmonize grading of skin responses. Subjectivity of this test is a shortfall 5340 that needs to be taken into consideration. 5341 53426.7.6.3 In vitro skin sensitization tests 5343The need to conduct animal testing is being contested in many countries around the world and more 5344 emphasis is being placed on developing alternative in vitro testing methods to reduce, refine and 5345 even replace current reliance on animal testing. Although alternatives to animal testing are desired, 5346 unless they are mechanistically founded, their utility in qunatititaive rsisk assessment and decision

173

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5347 making processes is undermined. However, in the area of skin sensitization (allergic contact 5348 dermatitis in humans or contact hypersensitivity in rodents), the key events involved in the process 5349 of sensitization occurring at various levels of biological organization are identified through a well 5350 constructed skin sentisation Adverse Outcome Pathway (AOP). According to this AOP, the covalent 5351 binding of electrophilic substances to nucleophilic centres in skin proteins (molecular initiating event) 5352 initiates the sensitsation cascade leading to inflammatory responses as well as gene expression 5353 associated with specific cell signalling pathways such as the antioxidant/electrophile response 5354 element (ARE)-dependent pathways in keratinocytes resulting in the activation of dendritic cells and 5355 T cell proliferatin. Based on the key events identified in this AOP, in vitro assays addressing the 5356 specific key events were developed and were recently adopted by OECD in 2015. These in vitro 5357 methods include, Direct Peptide Reactivity Assay (DPRA), TG 442C; ARE-Nrf2 Luciferase test method, 5358 TG 442D and Human cell line activation test 9H-CLAT), TG 442E. Thus, skin sensitization is one 5359 toxicological area where validated and harmonized methods are available for in vitro assays, which 5360 may be applicable to nanomaterial testing as well with significant modifications. 5361 53626.7.6.3.1 The Direct Peptide Reactivity Assay (DPRA) (OECD 442C) 5363This test is proposed to address the molecular initiating event of the skin sensitisation AOP, namely protein 5364 reactivity, by quantifying the reactivity of test chemicals towards model synthetic peptides 5365 containing either lysine or cysteine. Cysteine and lysine percent peptide depletion values are then 5366 used to categorise a substance in one of four classes of reactivity for supporting the discrimination 5367 between skin sensitisers and non-sensitisers. The guidance documents states that the method is not 5368 applicable for testing of metal compounds since they are known to react with proteins with 5369 mechanisms other than covalent binding. As discussed before, there is a high level of both covalent 5370 and non-covalent interaction between nanomaterials and proteins. Therefore, it is questionable 5371 whether this method is applicable to nanomaterials. 5372 53736.7.6.3.2 ARE-Nrf2 Luciferase (OECD 442D) 5374The second in vitro test method, ARE-Nrf2 Luciferase (OECD 442D) is proposed to address the second key 5375 event. Skin sensitisers have been reported to induce genes that are regulated by the antioxidant 5376 response element (ARE). Small electrophilic substances such as skin sensitisers can act on the sensor 5377 protein Keap1 (Kelch-like ECH-associated protein 1), by e.g. covalent modification of its cysteine 5378 residue, resulting in its dissociation from the transcription factor Nrf2 (nuclear factor-erythroid 2- 5379 related factor 2). The dissociated Nrf2 can then activate ARE-dependent genes such as those coding 5380 for phase II detoxifying enzymes.

174

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5381The guidance states that the test method is applicable to soluble chemicals or chemcials that can form a 5382 stable dispersion (i.e. a colloid or suspension in which the test chemical does not settle or separate 5383 from the solvent into different phases) either in water or DMSO. This method may therefore be 5384 applicable to nanomaterials, provided that the stable dispersion criteria can be fulfilled. 5385 Nevertheless, the test cannot be used as a stand-alone test to predict the sensitizing potential of 5386 nanomaterials. Rather, the data should be considered in the context of integrated approaches, 5387 combining the results with other complementary information e.g., derived from in vitro assays 5388 addressing other key events of the skin sensitisation AOP as well as non-testing methods that include 5389 read-across from chemical analogues. In general, one can assume that if bulk counter part of a 5390 nanomaterial is classified as sensitizer, then the respective nanomaterials can be considered as 5391 sensitizers and should be prioritized for testing. Some nanomaterials, owing to their physical- 5392 chemical properties can be more reactive and induce localized oxidative stress, which can then lead 5393 to skin sensitization. For example photocatalytic nanomaterials may fall under this category. More 5394 research is required in this area that will result in the development of an intelligent testing strategy, 5395 similar to those recently evaluated by van der Veen et al. [16] for conventional substances. Ideally, 5396 such a testing strategy should use a combination of in silico and in vitro methods specifically 5397 designed for nanomaterials. 5398 53996.7.6.3.3 In vitro skin irritation 5400For conventional substances, the irritant and corrosive potential can be assessed using tests following 5401 OECD guidelines 439, 431, 430 and 435. Reviewing the literature, very little nano-specific data is 5402 found on irritant and corrosive properties of nanomaterials. A number of studies have addressed skin 5403 and/or eye corrosion and irritation of nanoparticles using the above mentioned OECD test guidelines 5404 or similar [1-10]. These studies reported no irritation or corrosion effects from 10 nm Ag NPs [1], 50 5405 nm polystyrene NPs, 21 nm TiO2 [2], 140 nm rutile/anatase TiO2 NPs [10], 7 and 10-20nm SiO2 NPs 5406 [6], carbon nanohorns [9] and fullerenes [3, 8]. In studies investigating CNTs, one study did not find 5407 irritation from two forms of MWCNTs [7]. In another study, two products with SWCNTs and one 5408 MWCNTs did not show any signs of skin or eye irritation, while another one with MWCNTs showed 5409 mild and reversible skin and eye irritation [4]. One study investigated unwashed, washed and carbon 5410 coated AgNPs of different sizes and did not see immediate signs of irritation, although focal points of 5411 inflammation were observed in the skin upon closer examination [5]. More recently, anatase TiO2 5412 nanoparticles less than 25 nm in diameter were applied onto the ears of female BALB/c mice for 3 5413 days [11]. Using the local lymph node assay (LLNA), it was found that auricular lymph node cell 5414 proliferation was not affected. However, skin irritation was observed with 5% and 10% TiO2 5415 exposure, based on measuring the percentage of ear thickness change (IRR).

175

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5416In summary, most reported studies using OECD guidelines to study corrosion and irritation in healthy skin 5417 do not find any effects of nanomaterials, and those that do only find very mild and reversible effects. 5418 This could either indicate that nanomaterials are generally not irritant or corrosive, or it could mean 5419 that the OECD guidelines are not appropriate to study these effects for nanomaterials, for example, 5420 due to the recommended exposure and evaluation times being too short. This introduces some 5421 uncertainty into the evaluation of this endpoint. As a rule of thumb, nanomaterials that are (partly) 5422 composed of corrosive or irritant chemicals can be assumed to cause corrosion and irritation 5423 themselves. In addition, nanomaterial suspensions with an extreme pH (<2 or > 11.5) can be assumed 5424 to cause corrosion and irritation. One could also consider that nanomaterials with reactive properties 5425 are at risk for causing irritation and/or corrosion, although this would be based mostly on speculative 5426 assumptions. In any case, the term ‘reactive properties’ would need a better definition. 5427 54286.7.6.4 Skin responses in susceptible populations 5429There is interest in extending skin exposure studies to models of skin diseases, because the defective 5430 physical barriers presented in some skin conditions could allow greater penetration of nanomaterials. 5431 There are limited reports of such studies but the transgenic NC/Nga mice to model atopic dermatitis 5432 has been used. Yanagisawa et al. [12] found that repeated exposure of TiO2 nanoparticles (15, 50, or 5433 100 nm) on the ear skin of NC/Nga mice, through intradermal injections over a 17 day period, 5434 resulted in significant increase in histamine levels in blood serum and IL-13 expression in the ear. 5435 Interestingly, co-exposure of the TiO2 nanoparticles together with mite allergen extracts aggravated 5436 epidermis thickening and elevated levels of IL-4, IgE and blood serum histamine, suggesting that 5437 atopic dermatitis symptoms were exacerbate through Th2-biased immune responses. A similar study 5438 was conducted by Hirai et al. [13] by topical application of 30 nm silica nanoparticles, together with 5439 mite allergen extracts, on the ears and upper backs of NC/Nga mice. Atopic dermatitis-like lesions in 5440 the form of increased ear thickness were recorded. Allergen-specific Th1-related IgG2a and Th2- 5441 related IgG1 levels were detected, which were absent when the mice were exposed to silica 5442 nanoparticles applied separately from the allergen or to well-dispersed silica nanoparticles. The 5443 presence of allergen adsorbed agglomerates of silica nanoparticles led to a low IgG/IgE ratio which is 5444 a key risk factor in human atopic allergies. 5445With the interest in understanding sensitization of diseased skin to nanomaterials, Ilves et al. [15] 5446 sensitized the back skin on BALB/c mice with ovalbumin/staphylococcal enterotoxin B as the 5447 allergen/superantigen cocktail to evoke local inflammation and allergy to model atopic dermatitis. 5448 Exposure of this skin to ZnO particles showed that nano-sized ZnO (< 50 nm) is able to reach the 5449 dermis layer in atopic dermatitis skin. In addition, ZnO nanoparticles induced systemic production of 5450 IgE more significantly than larger ZnO particles, suggesting the allergy promoting potential of ZnO

176

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5451 nanoparticles. Smulders et al. [14] exposed mouse ear skin with TiO2, Ag and SiO2 nanoparticles for a 5452 day before applying 2,4-dinitrochlorobenzene (DNCB) for 3 days, and found that dermal exposure to 5453 Ag or SiO2 nanoparticles prior to DNCB sensitization did not influence the stimulation index (SI) after 5454 6 days of exposure. In contrast, the SI was increased with the application of 4 mg/mL TiO2 5455 nanoparticles prior to DNCB sensitization. Other studies reported no skin sensitization from 50 nm 5456 polystyrene NPs or 21 nm TiO2 [2], 140 nm anatase/rutile TiO2 NPs [10], fullerenes [3], carbon 5457 nanohorns [9] SWCNTs and MWCNTs [4]. Thus, the responses to nanomaterial exposure in normal 5458 population and susceptible population may be entirely different and, it is important to consider these 5459 differences in the final decision making processes. 5460 54616.8. Testing considerations and methods applicable to oral route 5462The OECD has published guidelines for assessing toxicokinetics in acute, repeated (28 or 90-day) dose 5463 toxicity studies following oral exposure. According to these guidelines, administration of the test 5464 material by gavage, drinking water or via food is recommended. However, it is important to note that 5465 some nanomaterials precipitate in water and form aggregates in water; when incorporated in the 5466 food matrix, they can be transformed, which may potentially affect their uptake, bioavailability, 5467 interaction with surrounding biological milieu and toxicological activity. Characterization of the 5468 physico-chemical properties of the nanomaterial in the matrix used for the study is therefore 5469 important. Caution is needed when applying the results obtained from an oral study with a specific 5470 matrix for the safety assessment of nanomaterials contained in a different matrix. 5471During their passage through the gastrointestinal (GI) tract, nanomaterials may be exposed to a range of 5472 different physiological environments that may affect their physico-chemical properties, which will in 5473 turn affect their potential local toxicity and absorption of the particles (Liu et al. 2012). In this light, 5474 when conducting oral studies, it is important to be aware of the species-specifc differences in the 5475 physiology of the GI tract. For example pH of the GI tract in rodents and humans is different (Bergin 5476 and Witzmann, 2013). Outcomes of oral studies in rodents with particles that undergo large pH 5477 dependent changes in their physicochemical properties may not be relevant for humans. It is 5478 therefore recommended to characterize nanoparticles at various ranges of pH prior to starting oral 5479 studies. 5480The majority of oral biokinetic studies have demonstrated that most nanomaterials remain in the gut 5481 lumen and are excreted via feces. Little is known on the local effects of nanomaterials in the GI tract. 5482 Effects of nanomaterials with antibacterial activity on gut microbiota may need to be investigated, 5483 since an imbalance in gut microbiota has been linked to a number of immune related diseases such 5484 as inflammatory bowel disease, type 1 diabetes and spondyloarthropathies. Although no validated 5485 methods exist to address this issue, improvements in sequencing technologies, coupled with a

177

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5486 renaissance in 16S rRNA gene based community profiling, have enabled the characterization of 5487 microbiomes (Costello et al., 2015). 5488Since for most nanomaterials absorption through the GI tract is limited, acute systemic effects of 5489 nanomaterials are likely not encountered frequently after oral exposure. However, the clearance rate 5490 of some systemically available nanomaterials was found to be dose dependent and very low, 5491 resulting in accumulation in the body (eg Elgrabli et al., 2015, Kumari et al., 2014), which could 5492 potentially result in long term effects. Therefore, investigating the distribution and clearance of 5493 nanomaterials in the body after oral exposure is recommended prior to conducting oral toxicity 5494 studies. 5495OECD guideline 417 offers a number of considerations for conducting such toxicokinetic studies which are 5496 mostly applicable to nanomaterials as well. Where possible, nanomaterials should be labeled 5497 radioactively with C14, to allow for an adequate time course study. Alternatively, ICP-MS or other 5498 analytical techniques can be used for elemental analysis of the materials in different organs at 5499 different time points. The suitable application of ICP-MS for this purpose and other relevant 5500 analytical techniques have been discussed in detail by Krystek et al. (2014). 5501In time course studies, part of the collected organs from toxicokinetic studies could be set aside for later 5502 use to investigate systemic (immune) toxicity parameters in case the results indicate that significant 5503 absorption has taken place. However, most of the tissue material will often be needed to detect the 5504 presence of nanomaterials since their absorption is frequently low. 5505 5506Once the absorption and clearance of the nanomaterial has been determined, the appropriate duration of 5507 subsequent toxicity studies can be selected (Figure XXX). Technical guidelines and guidance 5508 documents for conducting acute and (sub)chronic repeated dose toxicity studies can be found in the 5509 library of the OECD. 5510If significant absorption of nanomaterials occurs, it will likely primarily occur via the M-cells of the Peyer’s 5511 Patches in the small intestine where they are taken up by the gut-associated lymphoid tissue 5512 (Martirosyan and Schneider, 2014). From there, systemic distribution to various organs may occur, 5513 most likely to liver, spleen and other tissues of the mononuclear phagocyte system. These tissues 5514 therefore deserve thorough histopathological investigation in subsequent studies, paying special 5515 attention to markers of inflammation responses. Apart from directly affecting immune parameters 5516 locally or systemically, it has also been suggested that nanomaterials can act as a ‘Trojan Horse’, by 5517 taking along toxins or other components that are normally not absorbed by the GI tract (Govers 5518 1994). In addition, they may act as adjuvants to cellular immune responses (Lefebvre 2014), implying 5519 that nanomaterials may exacerbate Irritated Bowel Disease and Crohn’s disease. However, studies

178

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5520 investigating this link come up with contradicting results and so far, a causal relationship has not 5521 been established. 5522 55236.9 Systemic exposure and Translocation of nanomaterials into the blood stream 5524The main administration route up to date for biomedical nanomaterials such as theranostic nanomaterials 5525 is via intravenous injection (Nichols and Bay. 2012), where they encounter a complex cellular and 5526 molecular milieu of the human blood (Clift et al., 2015). The human blood is composed of a cellular 5527 fraction (45%), i.e. red blood cells, white cells and platelets, and the blood plasma (55%) consisting of 5528 proteins, glucose, amino acids, fatty acids and others such as clotting proteins, i.e. fibrinogens. The 5529 term “blood serum” refers to plasma from which fibrinogen has been removed. Blood contains more 5530 than 1000 types of proteins and over 50 of them have been found adsorbed onto the nanomaterial 5531 surfaces (Gao and He, 2014; Nel et al., 2009). Since blood composition is highly complex, the analysis 5532 of the nanomaterial behavior in complete (human) blood is very challenging and the studies are 5533 routinely conducted ex vivo using defined suspension conditions and / or in vivo experiments (for a 5534 review see Hall et al.(2007). Indeed, most research is focused on the identification of surface 5535 adsorbed proteins on nanomaterials with respect to their physicochemical properties (Lundqvisk et 5536 al., 2008). 5537Recently the use of human blood cell models has been described to study the immunotoxic potential upon 5538 exposure to nanomaterials (Tulinska et al., 2015). This approach includes the isolation of 5539 lymphocytes, natural killer cells, granulocytes and monocytes from fresh peripheral whole-blood 5540 cultures and subsequent assessment using specific ex vivo assays such as proliferation activity of 5541 lymphocytes, killing activity of NK cells, phagocytic activity of granulocytes, etc. 5542Information on immunotoxicity of nanomaterials that become systemically available can be obtained from 5543 standard repeated dose in vivo toxicity studies, for which various technical guidelines and guidance 5544 documents are available at the library of the OECD or the ICH. Standard toxicity assays include, 5545 hematological changes such as leukocytopenia/leukocytosis, granulocytopenia/granulocytosis, or 5546 lymphopenia/lymphocytosis, alterations in immune system organ weights and/or histology (e.g., 5547 changes in thymus, spleen, lymph nodes, and/or bone marrow), changes in serum globulins not 5548 explained by other factors, changes in serum immunoglobulins, enhanced infection incidence and 5549 augmented tumor rate without a different possible explanation. The necessity for additional 5550 immunotoxicity studies can be based on the weight of evidence of the outcome of the initial 5551 immunotoxicity parameters. Additional studies that may be recommended include immune function 5552 studies in rodents or non-rodent species such as T-cell dependent antibody response (TDAR) in 5553 affected cell types. If such additional tests provide sufficient data to conclude on a risk of 5554 immunotoxicity that is considered acceptable, no extra animal testing might be called for. If it is

179

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5555 unknown what specific cell type is affected and should be used to perform a TDAR then other cell- 5556 specific assays need to be selected. An acceptable study design of the additional rodent studies 5557 represents a 28-day study with daily dosing, which includes any of the following immunotoxicity 5558 assays: immunophenotyping (flow cytometry or immunohistochemistry), natural killer cell activity 5559 assays (performed if immunophenotyping studies demonstrate a change in number or if standard 5560 toxicity studies display high viral infectious rates), host resistance studies (testing if the test 5561 compound can modulate the host resistance), macrophage/neutrophil function 5562 (macrophage/neutrophil function assessment) and assays to measure cell-mediated immunity (in 5563 vivo assays where antigens are used for sensitization). For nanomaterials demonstrating high 5564 absorption and low clearance rates, repeated dose toxicity studies should be combined with a 5565 developmental toxicity study including a wide range of immunological parameters, since the 5566 developing immune system has been shown to be very susceptible to disruption by chemicals (Hessel 5567 et al., 2015). 5568 5569Conclusions 5570The majority of assays that are mentioned in this chapter are mainly focused on the detection of direct 5571 immunotoxicty, i.e. the recommended tests are limited to evaluating the potential for inadvertent 5572 immunosuppression and immunostimulation. Both are possible scenarios for nanomaterials, but 5573 other immunotoxic effects may also be encountered, such as the hypersensitivity and autoimmunity. 5574 The ICH S8 guideline for immunotoxicity testing of pharmaceuticals states that testing for respiratory 5575 or systemic allergenicity or autoimmunity is not based on standard testing approaches, since no 5576 (validated) models are available. To adequately assess the immunosafety of nanomaterials, safety 5577 evaluation guidelines need to be expanded with a testing strategy consisting of standardized and 5578 predictive immunotoxicity tests tailored to the unique properties of nanomaterials. 5579 5580It is obvious that for safety testing, not all assays mentioned could be applied for a specific test material. 5581 Many of the tests mentioned are explorative tests. For the purpose of risk assessment, human data 5582 would be most relevant, but these are scarce, and usually lack proper exposure assessment. Animal 5583 data are helpful, but it is not always possible to extrapolate animal data to humans. In additiom, 5584 there is a strong movement to reduce animal testing. In vitro testing often lacks the complexity of 5585 the entire system or organism. Hence, a specific guidance of what could be used for hazard 5586 assessment and which risk assessment framework could be employed is not available. Thus, such 5587 decisions should be driven by the nature of the test material and the purpose of the eventual risk 5588 assessment. 5589

180

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5590 7. APPROACHES FOR RISK ASSESSMENT 5591 5592 7.1 Introduction 5593 Risk assessment is a decision-support tool that enables regulators to derive decisions on the safety of 5594 exposure to a substance, the nature of potential hazards elicited by the substance, the affected 5595 receptors/population and on how best to manage the risks. As discussed in detail in the previous 5596 chapters, nanomaterials exhibiting distinct physical-chemical properties are increasingly found in a 5597 variety of products and as a consequence, exposure to these materials in our surroundings (homes, 5598 at work and in the environment) is inevitable. It is accepted that in the nanoform, many materials 5599 behave differently than their bulk counterparts, implying that they likely have novel toxicological 5600 properties that are unique from the same chemicals in their bulk form. This also suggests that 5601 knowing the risks posed by the bulk forms is not sufficient and that separate safety testing and risk 5602 assessment frameworks are needed for nanomaterials. 5603 5604 Studies conducted so far on the human health effects of nanomaterials reveal that the potential for 5605 exposure to most of these substances in occupational settings does occur; however, exposure via 5606 consumer products or in the environment is likely but limited. Furthermore, it is noted that the 5607 mechanisms by which nanomaterials are internalized and transported in the environment or in the 5608 human body may be different (Decan et al 2015; Bellmann S 2015; Lefebvre DE 2015), but so far 5609 toxicological effects unique to nanomaterials have not been revealed. Thus, similar to bulk 5610 substances, human health and environmental safety assessment and regulation of nanomaterials 5611 may follow a conventional risk assessment framework that is established for bulk substances. 5612 However, some physical-chemical properties of nanomaterials may influence their toxic potency and 5613 thus, to be certain, a risk assessment strategy involving investigation of potential exposure and 5614 hazard should be tailored to each property and application of nanomaterials. 5615 5616 Although the steps involved in human health risk assessment (HHRA) of chemicals in general may 5617 vary from one regulatory organization to the other, as in the case of chemicals, HHRA of 5618 nanomaterials could consist of four major steps (Details of the process are available on 5619 https://www.epa.gov/risk/human-health-risk-assessment; Vermeire 2007): 1) hazard identification, 5620 2) dose response, 3) exposure assessment and 4) risk characterization. Similar to chemicals, basic 5621 questions concerning the specific nanomaterial being assessed must be addressed before the process 5622 of risk assessment is initiated, including: “Who are the receptors or who (population) is at risk?”, 5623 “What are the pathways of exposure – air, water, soil, food, consumer products, pharmaceuticals; 5624 routes of exposure – inhalation, ingestion, skin contact?”, “What are the potential health effects?”,

181

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5625 “What information is known regarding bio-distribution, absorption, metabolism and excretion?” and 5626 finally, “What is the duration of exposure required to elicit an adverse health effect?”. These are 5627 indicated below in the context of the four general steps of risk assessment (Figure 1, adapted from 5628 https://www.epa.gov/risk/human-health-risk-assessment) as applicable to nanomaterials. 5629 5630

182

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5631 Figure 1: A putative risk assessment framework for nanomaterials

5632 5633 5634 7.1.1 Hazard identification 5635 The hazard identification step determines if exposure to a substance results in an increased incidence 5636 of a specific adverse health effect and if that adverse effect is likely to occur in humans. As part of 5637 this first step, the available scientific literature concerning the substance of interest is examined and 5638 an understading of how exposure to substances leads to an adverse effect (mode of action) is 5639 established. Mode-of-action is a sequence of key biological events and processes that are initiated 5640 when a substance interacts with a cellular component leading to a cascade of changes at the 5641 biochemical, cellular and organ level evaltually resulting in an adverse effect. It was originally 5642 described for carcinogenic substances but is now commonly applied to any substance-induced 5643 adverse effect (Sonich-Mullin et al., 2001). More recently, the Adverse Outcome Pathway (AOPs) 5644 framework, another mechanism to assemble and organize biological information in a chemical 5645 agonistic manner has been proposed (Ankley et al., 2010); 5646 http://www.oecd.org/chemicalsafety/testing/adverse-outcome-pathways-molecularscreening-and- 5647 toxicogenomics.htm). The former is used to support the weight-of-evidence for the observed 5648 deleterious effects induced by the substances, which is a component of the subsequent hazard 5649 characterization process. However, the application of AOPs in the context of risk assessment is yet to 5650 be demonstrated. More often, a single substance may be responsible for multiple adverse effects. 5651 Importantly, the type of adverse endpoints that are affected (eg. cancer vs non-cancer) will

183

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5652 determine the risk assessment guidelines employed. Although human clinical studies or 5653 epidemiological studies involving evaluation of large human populations to confirm the association 5654 between exposure to a substance and a health effect are most desired, such studies are expensive 5655 and scarce. Thus, most of the human health risk assessment studies conducted on chemicals rely on 5656 data collected from animal studies, with due consideration to the uncertainties involved in 5657 extrapolating information derived from such models to humans. Understanding of how a substance 5658 is absorbed, distributed, metabolized or eliminated in/from the body (toxicokinetics), and how 5659 exposure to a substance leads to negative health effects (toxicodynamics), is very important; 5660 however, such information is not often available for consideration. 5661 5662 7.1.2 Dose-response relationship 5663 Once the potential hazard of exposure to a specific substance is identified, the amount of substance 5664 required to initiate such effects has to be determined and the specific conditions under which such 5665 exposure is likely has to be described. These data then become the basis for establishing 5666 concentration/dose-response or exposure-response relationships. It is commonly accepted that an 5667 increase in adverse effect is directly proportional to the dose of the substance administered. 5668 However, this relationship (concentration response) is influenced by various factors such as, the type 5669 of substance, the targets (humans, animals), the recipient age and sex, and the exposure routes, 5670 among other factors. For risk assessment purposes, in addition to knowing the dose at which an 5671 adverse effect is observed, determining a dose at which the substance does not induce any effects, 5672 the lowest dose at which an effect is observed and a dose beyond which the response plateaus is 5673 critical. The lowest dose at which the first observed changes in a pre-determined adverse response 5674 are observed is then used as a critical effect dose for risk assessment. Factors are applied to that 5675 value to calculate the regulatory exposure limits imposed on that substance. Although epidemiology 5676 studies involving humans are desired in obtaining the dose response relationships, for a majority of 5677 chemicals such data are non-existent. There are other challenges including 1) animal studies, which 5678 are routinely used, are very expensive and as a result many studies do not investigate the response 5679 over a range of doses; 2) many animal studies are conducted using very high concentrations of 5680 substances and non-realistic exposure methods, which may lack relevance to a human scenario and 5681 3) extrapolation from animal species to humans often involves several uncertainties. Nonetheless, 5682 using high doses can inform on the absence of adverse effects at concentrations relevant to human 5683 exposure. 5684 5685 Two types of dose-response assessments exist: non-linear dose-response assessment and linear 5686 dose-response assessment:

184

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5687 5688 Non-linear dose response assessment – this is based on the threshold hypothesis, which refers to a 5689 minimum dose of substance required to initiate a toxic response; below this concentration, a toxic 5690 effect is not observed. Based on this assessment approach, two exposure levels are determined. A 5691 No-Observed-Adverse-Effect Level (NOAEL) is the exposure level at which no statistically significant 5692 biological change or adverse outcome is observed in experimental subjects following exposure to a 5693 substance compared to the non-exposed controls. A Lowest-Observed-Adverse-Effect Level (LOAEL) 5694 is the lowest dose at which the first statistically significant biological change or the adverse outcome 5695 itself is observed and is used when a NOAEL has not been identified. An alternative to the NOAEL can 5696 also be used, such as the Benchmark Dose (BMD) or Benchmark Dose Lower Confidence Limit 5697 (BMDL). The BMD or BMDL incorporate mathematical modelling of responses over several doses to 5698 determine a critical dose at which a predetermined change in response or adverse effect occurs 5699 (referred to as benchmark response – in the range of 1- 10 % depending on the statistical power of 5700 the study and the type of data). BMDL refers to the statistical lower confidence limit on a dose that is 5701 observed to produce the predetermined response. In a non-linear approach, the LOAEL, NOAEL or 5702 BMDL are used as the point of departure. 5703 From the NOAEL, LOAEL or BMDL a safety daily intake level can be derived. These can be referred to 5704 as either a reference dose (RFD, from an oral or dermal study), a reference concentration (RFC, from 5705 an inhalation study) or a derived no-effect level (DNEL, from all studies). RFD, RFC or DNEL derivation 5706 requires application of several uncertainty factors to take into account the variability and uncertainty 5707 reflecting possible differences between the human and animal population, as well as variability 5708 within the human population. 5709 A RFD is an estimate of a daily human oral exposure (in milligrams per kilogram of bodyweight per 5710 day) that is not likely to induce significant increases in risks of developing negative health effects 5711 during a life time. A RFC is used to assess safety of exposure to substances via inhalation route. Here, 5712 concentration refers to milligrams of substance in air per cubic meter of air. Similar to RFD, a DNEL is 5713 defined as the level of exposure beyond which humans should not be exposed. It is applicable to 5714 threshold effects (not applicable for non-threshold effects). 5715 5716 Linear dose-response assessment – is used for cancer-inducing substances. When MOA information 5717 does not exist or when MOA information suggests that there is no threshold to a toxicity observed, 5718 linear dose response assessment is applied. In this type of assessment, there is no safe dose level. 5719 Extrapolation from animals to humans is based on a straight line that is drawn from the point of 5720 departure for the observed effect to the dose showing zero response. The slope of this straight line

185

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5721 referred to as the slope factor or cancer slope factor is used to estimate risk at dose levels that fall 5722 along this line. Usually BMDL values are used as points of departure. 5723 5724 7.1.3 Exposure assessment 5725 Exposure assessment determines whether people are, or will be, in contact with potentially 5726 hazardous chemicals from specific conditions of use (WHO, 2010). It measures the life cycle of a 5727 substance in the environment from its origin to the target use and the route of exposure (inhalation, 5728 skin, oral). The direct exposure data is often missing, and thus, the exposure assessment is conducted 5729 using available data on the environmental concentrations of the substance. Then the estimates of 5730 human exposure over time and mathematical models to calculate transport and fate of the chemical 5731 in the environment are applied. While it is important to know the levels of exposure to a substance in 5732 the environment, it is equally crtical to know the doses that are biologically available, which 5733 determine the bioactivity potential of a substance (Bellmann S 2015). For example, the exposure 5734 assessment must consider the amount of substance that is present in the environment, the amount 5735 that can be inhaled or ingested or may be applied to the skin, and the amount that is internalized and 5736 available for interaction with biological material. 5737 5738 7.1.4 Risk characterization 5739 The last step of the HHRA takes into account all of the information gathered from the above three 5740 steps and informs the risk management excercise on the likelihood that a substance will induce risk 5741 and the nature of those risks. It documents the various steps involved in risk assessment, the 5742 assumptions or uncertainties exercised and the existence of a policy implication. 5743 5744 Although the four steps that form the basis of HHRA of chemicals can be applied to nanomaterials, 5745 several additional factors unique to nanomaterials (described below), must be taken into 5746 consideration (OECD 2012). 5747 5748 7.1.5 Major issues with applying the existing HHRA paradigm to nanomaterials 5749 One of the major hurdles in implementing the HHRA for nanomaterials is that there is no nano- 5750 specific regulation at present. Although a handful of attempts have been made to conduct risk 5751 assessment using existing HHRA approaches for some nanomaterials, they are not complete due to 5752 the lack of quantitative data (refer to later sections in this chapter). While individual regulatory 5753 agencies have a recommendation for what constitutes a nanomaterial, an internationally harmonized 5754 definition or agreement in this respect is lacking. There are no specific ‘triggers’ (the criteria used to 5755 support a regulatory decision to review a new material or a product for human health and

186

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5756 environmental impacts before its release into commerce) assigned to nanomaterials that will initiate 5757 their risk assessment (Council of Canadian Academies, YEAR). Various limitations associated with 5758 regulating nanomaterials have been reviewed and summarized by Hansen (2010). 5759 In Europe, at least for cosmetics and food, industries are obliged to register products containing 5760 nanomaterials and are asked to add the term ‘nano’ on the label when nanomaterials are used in the 5761 products. Although this is not harmonized across Europe, some EU member states have implemented 5762 national laws that oblige the industries to register nanomaterial containing products 5763 (http://ec.europa.eu/food/food/labellingnutrition/foodlabelling/index_en.htm, and EU, 2009, 2011). 5764 However, this is not a common practice globally. In the majority of EU countries there are no 5765 mandatory reporting systems for industries when their products contain nanomaterials. As a result, 5766 there is no way of tracking these materials in products and the extent of their use. The latter also 5767 means that the workers or consumers cannot be accurately monitored for actual exposure. 5768 Moreover, there is no consensus on what type of information is required for risk assessment (e.g., 5769 what properties should be considered in asking for information from the notifier), and there is no 5770 binding legislation that obliges industry to register the use of nanomaterials in their products 5771 (Falkner and Jaspers, 2012; Aschberger et al., 2014). Recently, the European Union announced that it 5772 will not administer a mandatory EU-wide harmonized registry for nanomaterials; however, will set up 5773 a voluntary EU-Observatory for Nanomaterials (EU-ON), hosted by ECHA. The EU-ON is expected to 5774 contribute to increased transparency of uses of nanomaterials in products on the EU market. 5775 Whether the EU-ON could influence risk assessment and management strategies is debated, given 5776 the voluntariness of the Observatory and its limitation to publicly available sources. 5777 Another technological limitation that has further hampered the progress in this area is the lack of 5778 validated methodologies to effectively detect nanomaterials in products, matrices and biological 5779 materials, characterize their properties and assess their safety (Linkov et al., 2015). Although existing 5780 regulatory acts and legislations are not specific to nanomaterials, they can be used to assess and 5781 manage the potential risks they may pose (see http://www.hc-sc.gc.ca/sr-sr/pubs/nano/faq- 5782 eng.php#a3). Thus, similar to chemicals, the risk assessment approach for nanomaterials in general 5783 will depend on the type of NM and on its intended use, which will determine the legislation under 5784 which they should be assessed and regulated (Amenta et al, 2015). This in turn will depend on how a 5785 NM is defined under that legislation. 5786 Legislative provisions that explicitly address NMs would ideally refer to a definition to identify and 5787 distinguish them from other materials (Amenta et al, 2015). On 18 October 2011, the European 5788 Commission (EC) published the “Recommendation on the definition of nanomaterial” to promote 5789 consistency in the interpretation of the term “nanomaterial” for legislative and policy purposes in the 5790 EU (European Parliament and Council, 2011). This definition is not legally binding but serves as a

187

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5791 reference that is broadly applicable across different regulatory sectors and can be adapted to specific 5792 product legislation (Table 1). The impact of the EC definition on different legislative frameworks is 5793 extensively discussed by Bleeker et al (2012) and the table below summarizes which regulatory 5794 frameworks have a nano definition, whether a label (on the product) is required, whether there are 5795 specific provisions in place and any further developments that are anticipated (from Bleeker et al, 5796 2012). 5797 Table 1: Overview of the legal frameworks governing nanomaterials

Legislation Regulation Definition Label Specific Anticipated discussions/ available required provisions developments Biocides new Biocides Yes Yes Separate Guidance regulation -EU, assessment 2012a)) Plant protection ((Regulation No No None Guidance products (EC) No. 1107/2009; EU, 2009b). Cosmetics Cosmetics Yes Yes Separate Guidance (regulation (EC assessment No. 1223/2009;EU, 2009c)). Food Information to Regulation Yes Yes None None consumers (EU) No. 1169/2011 (EU, 2011d) Contact materials Regulation No No Separate Guidance (EC) No. assessment 10/2011 (EU, 2011b). Novel foods/feeds New draft Yes Yes Separate Guidance Regulation on assessment novel foods Additives Regulation No No Separate Re-evaluation of authorised food (EC) No. assessment additives; guidance 1331/2008 (EU, 2008b Medicinal products No1 No None Definition: guidance The EC new and emerging technologies working group recommended the addition of ‘‘all devices incorporating or consisting of particles, components or devices at thenanoscale’’ in the highest risk class (N&ET, 2007); currently this is being revised. Medical devices No No None Definition, risk classification, specific provisions, guidance REACH2 No No None Guidance Classification, labelling No No None Additional or adaptation of and packaging legislation and guidance

1 The legislation does not include a definition, but the European Medicines Agency provides a definition on its website. 2 Registration, Evaluation, Authorisation and Restriction of Chemicals

188

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Occupational Health and No No None Guidance and OELs Safety 5798 The European Commission definition uses size (i.e., size range 1- 100 nm) as the only defining 5799 property of the material, which is the primary criterion that is considered in the definition of 5800 nanomaterials by regulatory agencies across the world. Size in the EC definition refers to the external 5801 dimensions of the constituent particles of a material which can be unbound but may also be in the 5802 form of agglomerates and/or aggregates. The EC definition, which was reviewed in 2015 (Rauscher et 5803 al., 2014; Roebben et al., 2014) applies to all particulate NMs irrespective of their origin, i.e. natural, 5804 incidental or manufactured, whereas, some jurisdictions limit the definition to manufactured 5805 nanomaterials only. 5806 Since nanomaterials are extensively used in various products, a definition is implemented by many 5807 regulatory agencies when deciding to consider evaluating a substance or a product as a 5808 nanomaterial. Along with the size (1-100 nm) parameter, particle size distribution is also specified. 5809 The nanomaterial definitions and particle size distribution thresholds recommended by selected 5810 jurisdictions are summarized in Table – (Chapter 2 – Types of Nanomaterials). 5811 5812 However, the analytical techniques to accurately measure the particle number distribution in a 5813 material are not yet available. Thus, several deficiencies (discussed above) in the fundamental risk 5814 assessment framework have led to a current state where these materials are largely unregulated or 5815 regulated as conventional chemicals without due considerations of their nano properties. The 5816 collective technical impediments have led to insufficient hazard data to perform a detailed risk 5817 assessment in support of regulatory decision making. Specific challenges in the context of the four 5818 risk assessment steps are described below. 5819 5820 7.1.5.1 Hazard identification process - challenges 5821 Hazards posed by nanomaterials will depend on their diverse physico-chemical properties (Nel, 5822 2006). For some nanomaterials, toxicity is shown to increase with the decreasing particle size and 5823 consequent increase in the surface area. For few others, their surface properties including surface 5824 charge, chemical functionalization groups as well as their shape seems to be important (Aillon et al., 5825 2009). In addition, although nanomaterials are defined as materials in the size range of 1 -100 nm, 5826 exposure often involves aggregates or agglomerates of few nanometers to several micrometers. This 5827 may in turn influence their ability to interact with biological materials, their uptake by living cells and 5828 their biological activity or toxic effects (Oberdorster et al., 2005). Thus, in order to determine if a 5829 nanomaterial can pose risk to humans, it is important to demonstrate that they come in contact with 5830 biological material in the nano-form and that the observed toxicity following exposure to

189

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5831 nanomaterials is influenced by their nano-size enabled properties. Although at present, exposure to 5832 nanomaterials is considered highest for the workers at production sites, with several products 5833 containing nanomaterials now in use, exposure via consumer products is exponentially increasing. 5834 This suggests that the hazard identification step of the risk assessment must include a detailed 5835 characterization of exposure including the amount of materials released in the environment and 5836 their changing properties as they move through the life cycle chain (synthesis to disposal). The other 5837 factors that need to be carefully considered are dose metrics and target tissue. By virtue of their 5838 small size, it is demonstrated that many nanomaterials freely translocate to deeper regions within 5839 tissues and cells accessing vesicles or organelles that were not accessed by the bulk particles. 5840 Although translocation to other organs is expected to be low, several recent studies have shown that 5841 particles deposited in lungs or GI tract are capable of translocating to other distant organs, where 5842 they may induce biological effects (Choi et al., 2010; Yu et al., 2007; Oberdorster et al., 2004; Czarny 5843 et al., 2014; Husain et al 2015). However, limitations in the methodologies currently available to track 5844 the biodistribution and translocation make it nearly impossible to generate such information. 5845 To summarize, the available data concerning hazard potential of nanomaterials is inconclusive and 5846 often questionable, and the appropriate test methods, or benchmarks for assessing the potential 5847 risks are not established. As discussed earlier, an integral part of assessment of potential 5848 toxicological effects of nanomaterials is adequate characterisation of nanomaterials through various 5849 life cycle stages from synthesis to exposure. Significant efforts have been made internationally to 5850 identify and validate measurement and characterization tools for nanomaterials 5851 (www.oecd.org/science/nanosafety); however, at present, standardized or internationally 5852 harmonized protocols to carry out such measurements are not available, making it difficult to directly 5853 relate a specific biological response to a specific nanomaterial or its properties. 5854 5855 7.1.5.2 Issues with dose-metrics 5856 Mass per volume or per body weight is a relevant dose-metric for chemicals, however, for most 5857 nanomaterials; this seems to be a poor predictor of toxicity owing to their diverse physico-chemical 5858 properties and unique behavior (Oberdorster et al., 2005; Rushton et al., 2010). Total surface area or 5859 total particle number is suggested to be appropriate dose-metrices for most nanomaterials. For 5860 example, particle surface area and particle volume have been shown to better predict lung responses 5861 in rats or mice exposed to several nanomaterials of varying sizes. However, surface area or particle 5862 number cannot be used as universal parameters to report dose metrics since biological effects 5863 induced by some nanomaterials are not directly correlated to the total surface area or particle 5864 volume (Oberdortser et al., 2005; Rushton et al., 2010). At present, most studies report only the 5865 administered dose, ignoring the important cellular uptake processes such as diffusion or

190

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5866 sedimentation, which define the rate at which the nanomaterials become available for cellular 5867 uptake, their fate and transport within cells. To add complexity, the uptake, fate and transport of 5868 nanomaterials is heavily influenced by the particle and exposure medium characteristics, which must 5869 be accounted for in dosimetry report (Chen et al., 2013; 2014; Simko et al., 2014; De Loid et al., 5870 2013).) 5871 Traditionally, dose response assessments have been conducted in animal models (mainly rodents). 5872 The extrapolation of thus derived data to human scenarios is then carried out using several untested 5873 assumptions and applying uncertainty factors. However, the numbers of nanomaterials that require 5874 immediate assessment make it impossible to test each one of them using in vivo models. In vitro 5875 alternatives, however, are not readily applicable to nanomaterials due to issues discussed earlier 5876 including dynamic and changing particle properties in different exposure systems, dose metrics and 5877 their relationships with observed effects. Thus, a standardized approach to measuring nanomaterial 5878 properties is crucial to compare the results derived from various studies, which in turn will help 5879 identify material-specific dose metrics and allow read-across prediction of effects. Standardized 5880 parameters that can be applied to nanomaterials, a standardized classification system, and 5881 generation of toxicological data for a set of reference nanomaterials using standardized 5882 measurement methods may help tackle some of these issues. 5883 5884 7.1.5.3 Issues with exposure assessment 5885 Issues concerning the exposure assessment of nanomaterials relate back to the lack of appropriate 5886 metrics for quantifying occupational or consumer exposures, which stems from issues with finding 5887 appropriate exposure monitoring techniques (Yorkel and MacPhail, 2011; Savolainen et al., 2010). 5888 Adding to the complexity is the lack of sensitive methodologies to detect and differentiate 5889 nanomaterials from background noise and the lack of analytical characterization methodologies. A 5890 single nanomaterial may have several potential routes of exposure as it traverses through the 5891 environment and interacts with different media. Thus, a thorough investigation of exposure 5892 throughout the life cycle of a nanomaterial is needed. For consumer exposure, another problem is 5893 that there is hardly any obligation to label or register nanomaterials in consumer products. 5894 Manufacturers can label nano on the product, but this does not mean that nanomaterials are 5895 actually present in that product, but also a product without a nano label can contain nanomaterials. 5896 Information on concentrations of nanomaterials in products are also lacking. Furthermore there are 5897 few human exposure models available that are validated for nanomaterials. Some, such as the 5898 worker exposure model Stoffenmanager nano and the consumer exposure model ConsExpo nano, 5899 are currently under development. For the latter, a new version has recently been made publicly 5900 available (www.consexponano.nl). Thus, several attempts are being made to accurately measure the

191

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5901 exposure; however, in its present state, lack of accurate exposure assessment poses the largest 5902 hurdle in HHRA of nanomaterials. 5903 7.1.5.4 Issues with risk characterization 5904 Since there are very few mandatory requirements for labeling products that contain nanomaterials, it 5905 presents a significant challenge to measure realistic exposure levels and to relate actual exposure 5906 levels to adverse effects. Although comparative approaches are commonly used to characterize the 5907 risk posed by chemicals (comparison of nanomaterials properties and behavior with those known for 5908 bulk materials of similar composition), this can only provide qualitative information since derivation 5909 of quantitative risk estimates is an issue due to datagaps about the distinct size-related properties 5910 attributed to nanomaterials. 5911 While the issues related to nanomaterials are daunting, the scientific community is hopeful that the 5912 novel techniques that are being developed will permit quantitative measurement of relationships 5913 between nanomaterial properties and associated toxicity. However, in the interim, less quantitative 5914 or qualitative ranking approaches can be applied to existing nanomaterials. 5915 To this end, several strategies have been proposed, which are outlined in the following sections. 5916 5917 Before implementing any strategy, basic questions must be asked as part of the problem formulation 5918 aspect of HHRA of nanomaterials. These questions include: 5919 5920 1. What is the chemical composition of the nanomaterial? 5921 2. Are the nanomaterials sufficiently characterized; that is, are there sufficient data regarding 5922 particle size, surface coatings, and other physical-chemical properties of nanomaterials? 5923 3. Has the nanomaterial been characterized in the exposure medium (ie. 5924 aggregation/agglomeration)? 5925 4. Are these nanomaterials readily internalized by living cells? 5926 5. Is there information available regarding their Absorption, Distribution, Metabolism and 5927 Excretion (ADME)? 5928 6. Is there information regarding dose-metrics? 5929 7. Who are the populations/individuals at risk? 5930 8. What further information is needed regarding product volume, potential application and 5931 uses, behaviors through the life cycle – for example if a nanomaterial is coated masking 5932 nanomaterial contact with biological samples, how durable is the coating? 5933 9. What are the most plausible routes of exposure? 5934 10. What are the toxic effects and threshold dose? 5935

192

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5936 These and other questions, defined as the current major knowledge gaps, are depicted in Figure X 5937 below, both for human and environment (RIVM, 2014). 5938

5939 5940 5941 Figure X: A summary of the major knowledge gaps existing for risk assessment of NM, adapted from 5942 RIVM, 2014 5943 5944 Once these questions are answered satisfactorily, additional questions regarding the likelihood of 5945 exposure to these materials, their potential for adverse effects at those exposure levels, the nature 5946 or severity of risk, and the use of safety controls (personal protection equipment) to mitigate 5947 exposure and/or hazard while handling the nanomaterials, should be answered before 5948 recommending risk management measures. In the meantime, a precautionary approach should be 5949 exercised. 5950 5951 5952 7.1.6 Conclusions – application of conventional risk assessment approach to nanomaterials 5953 Attempts have been made to conduct risk assessments for NM using approaches used for chemicals 5954 (listed below), which revealed that a complete and quantitative risk assessment at present is not 5955 possible due to lack of or insufficient exposure and hazard data (Pronk et al, 2009; Aschberger et al, 5956 2010a,b; Christensen et al, 2010, 2011). Moreover, hazard information currently available is 5957 generated from methodologies that are not validated for NMs (Johnston et al, 2013). Although 5958 definitive conclusions cannot be reached, in the interim the available results are being used to 5959 support decision making. Following are some examples of authoritative nanomaterial-specific risk

193

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5960 assessments conducted, which are not complete due to lack of qualitative and quantitative data 5961 (RIVM 2014). The examples are limited to a few first generation nanomaterials. 5962 5963 • The Scientific Committee on Consumer Safety (SCCS) assessed a number of cosmetic 5964 ingredients that are in the nanoscale such as Carbon Black CI77266, Titanium Dioxide and Zinc Oxide. 5965 It was found that dermal uptake of these ingredients through dermal route is minimal and hence 5966 internal exposure is negligible. However, use of these ingredients in cosmetic products dispensed 5967 through spray bottles potentially leading to exposure via inhalation route was not recommended 5968 (SCCS, 2012a,b, 2014a,b,c,d,e). 5969 • The Scientific Committee on Existing and Newly Identified Health Risks (SCENIHR) reviewed 5970 the available information for nanosilver in the context of life cycle analysis, consumer and 5971 occupational exposure, ADME (toxicokinetics), toxicity to humans and to environmental species. The 5972 results concluded that with the available information, an adverse effect associated with nanosilver 5973 exposure could not be identified, and that a quantitiave risk assessment cannot be conducted 5974 (SCENIHR, 2014). 5975 • Nanosilver and TiO2 (2017) are also undergoing substance evaluation under REACH. Other 5976 nanomaterials such as silica (SiO2) are widely used and potential for human exposure to these 5977 materials is increasing. As a result, SiO2 has been evaluated by the SCCS (SCCS, 2015) and is 5978 undergoing a substance evaluation under REACH. 5979 • The European Safety Authority (EFSA) strategy for 2020 includes re-evaluation of safety of 5980 established food additives, which will include nanoforms of the additives. This evaluation process is 5981 scheduled to be completed in 2020. 5982 • The United States National Institute for Occupational Safety and Health (NIOSH, 2013) 5983 completed a human health risk assessment for carbon nanotubes (CNT) and carbon nanofibers (CNF)- 5984 induced lung fibrosis. The study recommended an exposure limit (REL) of 1µg/m3 elemental carbon 5985 as a respirable mass 8-hour time-weighted average concentration. The study also noted that this 5986 concentration reflected the analytical limit of quantification and thus that the exposure levels to 5987 CNTs should be kept below the levels of 1µg/m3. This study highlights additional limitations to 5988 quantititave risk assessment, which include the avaibility of sensitive analytical tools to detect 5989 nanomaterials. 5990 5991 Thus, for now, the challenges and knowledge gaps faced by risk assessors with regards to NMs have 5992 been clearly identified (Johnston et al, 2013, Thomas et al, 2009, Aschberger et al, 2011, Savolainen 5993 et al, 2010). There is also consensus among the agencies, researchers and regulators (e.g RIVM 2009,

194

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

5994 2014) that the conventional risk assessment framework currently available for chemical substances 5995 should be critically evaluated and eventually adapted for application to NMs. 5996 5997 7.2 Recent initiatives developing risk assessment approaches for nanomaterials 5998 7.2.1 Introduction 5999 As discussed in the previous chapters, nanomaterials exhibit versatile properties that influence their 6000 biological behavior and their ultimate fate. At present, the relationship between the properties and 6001 the observed toxicity is not clear. Moreover, it is not clear if some of the physico-chemical properties 6002 are more important than others. Therefore, a systematic evaluation of toxicological potential of each 6003 NM on a case-by-case basis is encouraged. However, the number of NM that require immediate risk 6004 assessment and the cost, time and experimental animals associated with extensive testing of each 6005 NM variant, make the case-by-case approach impractical. In the interim, the available information 6006 concerning physico-chemical properties, exposure, and hazard potential can be used to inform read 6007 across or grouping approaches, which will help guide the testing strategy. Alternative risk screenings 6008 tools such as the Swiss precautionary matrix, and Stoffenmanager Nano are already available; at 6009 present, these tools are designed for industrial use, and the development of similar tools to assist 6010 regulators in their decision making is on the horizon. 6011 6012 The suitability of risk assessment approaches for nanomaterials has been assessed by various 6013 international scientific organizations and committees such as the EU Scientific Committee on 6014 Emerging and Newly Identified Health Risks (SCENIHR, 2006, 2007b, 2009), Nanotechnology 6015 Industries Association (Arnold, 2007), US Environmental Protection Agency (US EPA; 2008) and Defra 6016 (Crane and Handy, 2007). A summary of these analyses has been reported by Defra (Rocks et al., 6017 2008) and an updated list of various recent risk assessment initiatives and other supporting 6018 information are included in the Table below (based on Dekkers et al, 2016). Some of the cited 6019 strategies will be explained in more detail in the paragraphs 7.2.3 and 7.2.4. 6020 6021 Table 1: Summary of recent developments in risk assessment approaches for nanomaterials

Sources Risk assessment strategy Read across and grouping Other supporting approaches information MARINA Bos et al., 2015 Oomen et al., 2015 GUIDEnano S.W.P. Wijnhoven and P. Van M. Park and G. Janer, Kesteren, personal personal communication communication ITSnano Stone et al., 2014 Stone et al., 2014 Stone et al., 2014 Intelligent testing strategy for NM NANoREG Dekkers et al, 2016 SUN Malsch et al., 2015

195

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

ECETOC Arts et al., 2014 Arts et al., 2015a Arts et al., 2015b ECHA ECHA, 2013 Regulation (EC) No 1907/2006 ECHA/RIVM/JRC, 2016 (EU 2006) Hankin et al. 2011 (RIP-oN2 on information requirements) Aitken et al., 2011 (RIP-oN3 on chemical safety assessment) EC, 2011 (updated chapters of the guidance on IR & CSA) OECD OECD, 2013 OECD, 2014 OECD, 2007-2016 Testing programme of manufactured nanomaterials SCENIHR SCENIHR, 2009 SCENIHR, 2007 Test guidelines RIVM/ARCADIS Sellers et al., 2015 Labib et al Labib et al., 2016 AOP and mechanisms driven approaches 6022 6023 6024 7.2.2 Risk Assessment approach by SCENIHR 6025 SCENIHR (2007b) was the first organisation to describe a potential risk assessment approach for 6026 nanomaterials and suggested that identification of nanomaterials of interest and in-detail 6027 characterization of a nanomaterial are the prerequisites for the success of an effective risk 6028 assessment paradigm for nanomaterials. Human and environmental risks from exposure to the 6029 specific nanomaterial is then evaluated in a 4-stage process (see also the Figure 2 and 3 below): 6030 6031 Figure 2: SCENIHR 4-stage process

6032 6033 In steps 1 and 2, the focus is on the nanomaterial exposure. If no human or environmental exposure 6034 is expected or if exposure is expected to be very low, the process stops or the nanomaterial is 6035 considered a low priority for hazard assessment. Step 3 focuses on careful selection of in silico/in 6036 vitro models for hazard testing in tier-1. If the results are positive, either limited in vivo testing (when 6037 effects are very similar to those of the bulk chemical), or in-depth specialized in vitro testing followed 6038 by in vivo testing (if effects are different from those of the bulk chemical or a different effect that 6039 was previously not known) is recommended. The data generated from the first three steps are used 6040 to derive exposure/dose-response and to assess the risk. SCENIHR, however, does not provide in this 6041 proposal details on the kind of in vitro or in vivo testing to be performed. It may be implied that the 6042 route of exposure and potential application may gude the choice of toxicity models for testing. In

196

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6043 2009, SCENIHR concluded in the opinion on “Risk Assessment of Products of Nanotechnologies” that 6044 ‘this framework remains appropriate although a few further details can be added in the light of 6045 recent publications’. 6046 6047 Figure 3:

6048 6049 7.2.3 Summary of RA approaches developed in the recent EU FP7 projects 6050 In the FP7 project MARINA, a hypothetical framework for risk assessment of NM has been developed 6051 (Bos et al, 2015). The strategy involves data generation in two phases (Figure X): Phase 1- Problem 6052 framing and Phase 2 - Risk assessment. Identification of Relevant Exposure Scenario (RES) at different

197

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6053 stages of the life cycle takes the central stage in this strategy driving the prioritization, ranking and

6054 risk characterisation decisions. 6055 The main goal of the Problem Framing phase is to set the scope of the risk assessment and to identify 6056 Relevant Exposure Scenarios (RESs) for each step in the life cycle of a NM. The Risk Assessment phase 6057 (phase 2) is a stepwise, iterative process of identification of data gaps from the perspective of risk 6058 assessment, of choosing the appropriate tools for data generation to fill in the gaps, of generation 6059 and collection of the data and of performance of a risk characterization (i.e., human health and/or 6060 ecological according to the user’s objectives) (Bos et al, 2015; Oomen et al, 2016). 6061 6062 In the Intelligent Testing Strategy (ITS nano) project, a survey was conducted to collect opinions from 6063 more than 80 stakeholders to identify research priortities that would enable effective intelligent risk 6064 evaluation strategies for NMs. Among the priorities identified, availability of standardised 6065 nanomaterials for testing, standardization of methods for characterizing nanomaterials throughout 6066 the life cycle and as pertinent to different risk assessment steps, defining important properties that 6067 are likely to change over the life cycle of a nanomaterial and those that are detrimental, 6068 characterizing the exposure in the context of dose delivered, and bioavailable dose, were the 6069 overarching and cross- cutting ones. This framework not only highlighted the important research 6070 needs but also how each of those priorities can be addressed. (Stone et al, 2014). 6071 6072 In NANoREG, the focus was to develop harmonized approaches to conduct risk assessment of 6073 nanomaterials. Taking into consideration the questions and needs of regulatory communities and the 6074 legislators, NANoREG is proposing a strategy that identifies the situations under which the 6075 nanospecific grouping, read across and (Q)SAR tools are applicable or where data gaps exist. This is 6076 the first project to recommend the types of data or endpoints to consider for any nanomaterial

198

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6077 requiring risk assessment. In the proposed flow chart for risk assessment of nanomaterials (Dekkers 6078 et al, 2016), six overlapping elements have been identified as most important nanospecific 6079 determinants: exposure potential, dissolution, nanomaterial transformation, accumulation, 6080 genotoxicity and immunotoxicity. The main objectives of the proposed strategy were (Dekkers et al, 6081 2016): -to prioritise those applications that have the highest potential to cause human health effects 6082 -to identify the most important information needed to address the nanospecific issues within 6083 the risk assessment. 6084 6085 In another European project GUIDEnano, a web-based tool, the primary focus of which is to guide 6086 nano-enabled product developers (industry) on safe by design approaches, designing product- 6087 specific risk assessment and risk mitigation approaches is being developed. This approach will also 6088 develop tools to indentify hot spots for release of nanomaterials, decision trees to provide guidance 6089 on the uses of (computational) exposure models and, when necessary, design of cost-effective 6090 strategies for experimental exposure assessment (http://www.guidenano.eu/). Specific publications 6091 on this tool are currently in preparation and will be published soon. 6092 6093 In addition to these projects, two recent reports conducted risk assessment of TiO2 and SiO2 in 6094 food, based on internal organ concentrations (instead of external concentrations) using a kinetic 6095 model taking into account accumulation of nanomaterials in the body over time (Heringa et al, 2016; 6096 Van Kesteren et al, 2015). 6097 6098 7.2.4 Read across approaches for identifying the risk of nanomaterials 6099 In general, application of grouping of substances and read-across between substances is recognized 6100 as a valuable approach in regulatory frameworks e.g. to fill potential data gaps in the hazard 6101 characterisation, based on availability of adequate data from similar substances. It is expected that 6102 for nanomaterials, grouping and read-across approaches will be an important means of addressing 6103 identified data gaps. Due to the numerous possible nanoforms with the same chemical identity (i.e. 6104 covered by the same registration dossier or CAS number) but with differences regarding other 6105 physicochemical properties (e.g. surface modification, size distribution and particle shape), there is a 6106 need for alternative approaches that would allow predicting hazardous properties by implementing 6107 readacross between nanoforms and/or from non-nanoforms to nanoforms of the same substance to 6108 minimise testing, in particular, on animals. 6109 For nanomaterials, Industry (ECETOC) (e.g. Arts et al., 2014; Arts et al., 2015a; Arts et al., 2015b) and 6110 Member States (e.g. RIVM-Arcadis project, see Sellers et al., 2015) as well as several research 6111 projects in FP7 (e.g. MARINA (Oomen et al., 2015), NanoREG and GUIDEnano) and individual

199

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6112 research groups (e.g. Gebel et al., 2014; Walser and Studer, 2015) have proposed methodologies for 6113 grouping and read-across of nanomaterials through case studies. Also, the OECD guidance on 6114 grouping of chemicals (OECD, 2014a) mentions nanomaterials; however, no concrete guidance is 6115 provided for nanomaterials in that document. 6116 6117 Based on the ideas developed in the former approaches, recently, ECHA, together with RIVM and JRC 6118 developed a new approach for read-across that aims to illustrate a number of core principles (i.e. 6119 elements to consider), which are based on the current, albeit incomplete, scientific understanding 6120 (ECHA/RIVM/JRC, 2016, summarized below in Figure X). This has led to a structured approach to 6121 guide registrants and regulators on how to apply grouping and read-across concepts to nanoforms 6122 within regulatory frameworks such as REACH. Based on this working document, ECHA is currently 6123 working on a draft guidance: Appendix R.6-1: Recommendations for nanomaterials applicable to the 6124 Guidance on QSARs and Grouping (ECHA, in preparation, 6125 https://echa.europa.eu/support/guidance/consultation-procedure/ongoing-reach). 6126 Apart from the above, approaches to classify nanomaterials to estimate their potential risks have 6127 been proposed. These include, stochastic multicriteria acceptability analysis (SMAA-TRI; Tervonen et 6128 al., 2007)), weight of evidence (Linkov et al., 2011) , grouping (Arts et al 2014, Arts et al 2016), 6129 Bayesian networks (Low-Kam et al., 2015; Marvin et al., 2017) or an approach to combine the 6130 heterogenous data available from all the above methods and various sources (Linkov et al., 6131 2015)was outlined. Thus, tremendous efforts are being made to determine the best risk assessment 6132 and management stratagies. A need for detailed physical-chemical characterization of materials in 6133 their dry and aquesous states, before and after exposure, at different stages of life cycle; analysis of 6134 bioavailable dose and material kinetics are emphasised across the various initiatives. However, there 6135 is no harmonized agreement on the types of toxicity endpoints to be included or the toxicity models 6136 used. 6137 6138 6139 6140 Figure X: Proposed strategy or stepwise procedure for using data between (nano)forms (adapted 6141 from ECHA et al, 2015).

200

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

Identification of the nanoform Substance · The potentially unassessed nanoform is characterised by 1 basic physicochemical parameters

Initial grouping of nanoforms · Find similarities on chemical identity and particle 2 characteristics · Find similarities on fundamental behaviour and reactivity

Identification of available data and data gaps Information requirements 3 · Data available on the (group of) nanomaterial(s) REACH (tonnage) · Identification of data gaps

Identification of potential source materials · Hypothesis building to justify use of source materials 4 · Consider similarities in physicochemical parameters · Consider relationships between physicochemical parameters and toxicokinetics and hazard · Identify information needed to substantiate the hypothesis

Substantiate hypothesis Consider all endpoints 5 NO required under REACH in · Information gathering · Consider data needed to improve argumentation an overall testing strategy · Build testing strategy

Assess new data 6 · Hypothesis sufficiently substantiated? YES Fill data gap 6142 6143 7.3. Nanomaterial immunotoxicity risk assessment 6144 7.3.1 Summary of existing immunotoxicity risk assessment approaches for chemicals and 6145 pharmaceuticals 6146 As described in the previous chapters, NMs that gain access through various portals of entry of the 6147 body can interact with immune cells and tissues and have the potential to induce a variety of 6148 immunotoxic effects in animal and cell culture models. Collectively, a growing body of research 6149 supports the conclusion that adverse effects on the immune system are a relevant consideration in 6150 NM risk assessment (exemplified in the previous chapters). Immunotoxicity testing guidelines and 6151 methodologies developed for pharmaceuticals, cosmetics, agrochemicals, food additives, industrial 6152 chemicals and environmental chemical contaminants can inform approaches for NM immunotoxicity 6153 risk assessment. Table 2 summarizes guidances and methodologies that are widely used for 6154 immunotoxicity testing. While some focus specifically on immunotoxicity assessment (ICH, 2005; 6155 IPCS, 2012; EPA, 1998, 2013), others incorporate selected immune parameters in protocols designed 6156 to characterize general toxicity for various routes and durations of exposure. These include OECD 6157 test guidelines under Section 4 (Health Effects) and the US EPA’s test guidelines in Series 870 (Health 6158 Effects). Basic immune parameters incorporated in these protocols usually include total and 6159 differential leucocyte counts, serum albumin and globulin levels, spleen and thymus weight, and 6160 optional immune tissue histopathological assessment that may include spleen, thymus, and selected

201

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6161 lymph nodes relevant to the route of exposure. Significant changes in such endpoints, which are 6162 largely observational, may serve as triggers for further characterization of immunotoxicity using a 6163 functional immune assay such as a T-dependent antibody response (TDAR). 6164 6165 Table 2. Immunotoxicity guidances and testing approaches for chemical and pharmaceutical risk 6166 assessment. Document Description Refer ence ICH S6 (R1) Preclinical safety Guidance for assessing immunostimulation and ICH evaluation of biotechnology- immunotoxicity assessment of biotechnology-derived 1997 derived pharmaceuticals (ICH, pharmaceuticals. Recommendations include ; ICH 1997, 2011) immunogenicity assessment (measurement of 2011 antibodies) when conducting repeated dose studies; immunotoxicity including stimulation or suppression of cellular and humoral immunity, inflammatory reactions at the site of administration, and surface antigen expression on target cells; routine tiered testing or standardized batteries not recommended. ICH S8 Immunotoxicity Studies for Guidance for assessing immunosuppression or ICH Human Pharmaceuticals (ICH, enhancement by human pharmaceuticals. 2005 2005) Recommendations on nonclinical (ie. rodent) testing to identify potential immunotoxins and guidance on a “weight-of-evidence decision making approach” for immunotoxicity testing; recommends consideration of data from standard toxicity studies; if warranted additional immunotoxicity studies may include a TDAR assay, immunophenotyping, NK cell activity, host resistance, macrophage/neutrophil function, or assays of cell mediated immunity; standardized approaches for respiratory or systemic allergenicity or drug-specific autoimmunity not recommended due to lack of availability; well-standardized methods for skin sensitization testing not included in this guidance. IPCS Environmental Health Criteria Overview of pre-1996 methods for assessing immune WHO

202

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

180. Principles and methods for status, primarily immunosuppression, inflammation and 1996 assessing direct immunotoxicity autoimmunity. The document also summarizes early associated with exposure to studies validating immunosuppression assays for chemicals (WHO, 1996) immunotoxicity risk assessment. IPCS Environmental Health Criteria Overview of allergic disease, clinical aspects, WHO 212. Principles and methods for epidemiology and experimental models. Highlights lack 1999 assessing allergic of exposure and hazard data for use in assessing allergic b hypersensitisation associated with disease risks associated with chemical exposure. exposure to chemicals (WHO, 1999b) IPCS Environmental Health Criteria Overview of autoimmune disease, clinical aspects, WHO 236. Principles and methods for epidemiology and experimental models. The document 2006 assessing autoimmunity associated also highlights exposure, mode of action, susceptibility with exposure to chemicals (WHO, and other issues, concluding that much of the 2006) information needed to address the risk of chemical-induced autoimmune diseases is not available. IPCS Harmonization Project Framework for chemical immunotoxicity risk WHO Document No. 10. Guidance for assessment, with schematics for weight-of -evidence 2012 immunotoxicity approaches, for the major categories of potential risk assessment for chemicals immune-chemical interactions: direct (WHO, 2012) immunosuppression, direct immunostimulation, immune sensitization and allergy, and autoimmunity and autoimmune disease. OECD Test Guidelines for the Test guidelines for assessing chemical toxicity using OEC Testing of Chemicals. Section 4. animal models, primarily rats and mice, including oral, D Health Effects (http://www.oecd- inhalation and dermal exposure routes. Most 2008 ilibrary.org/environment/oecd- guidelines for repeated dose studies 28d or longer ; guidelines-for-the-testing-of- incorporate descriptive endpoints relevant to OEC chemicals-section-4-health- immunotoxicity, including total and differential D effects_20745788) leucocyte counts, serum albumin and globulin levels, 1998 spleen and thymus weights and immune tissue histopathology. Test Guideline 443 (Extended One Generation Reproductive Toxicity Study) includes an

203

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

optional F1 cohort for a TDAR assay. US EPA Health Effects Test Test guideline for assessing immunosuppression due to US Guidelines OPPTS 870.7800 repeated dose chemical (pesticide) exposure using EPA, Immunotoxicity (EPA, 1998, 2013) rodents; includes descriptive parameters (hematology, 1998 lymphoid organ weights, histopathology) from standard , rodent toxicology studies; includes functional assay 2013 (TDAR), lymphocyte phenotypic analyses. The US EPA * has also developed an immunosuppression risk assessment guidance that has not been published (as of Apr 2016)a US EPA OPPTS/OCSPP Test Test guidelines for oral, dermal, inhalation, Guidelines, Series 870, Health developmental and other toxicity studies, ranging from Effects acute to chronic. Most incorporate basic immune (https://www.epa.gov/sites/produ parameters, including hematology, clinical endpoints ction/files/2015- and optional histopathology of immune tissues that 11/documents/ocspp- may provide support for more focussed immunotoxicity testguidelines_masterlist-2015-11- studies (OPPTS 870.7800). 19.pdf) US Food and Drug Administration Review of toxicological information submitted in (FDA) Guidance for Industry and support of the safety of food ingredients. Includes July Other Stakeholders: Toxicological guidelines for immunotoxicity studies. 2000 Principles for the Safety ; Assessment of Food Ingredients Upd ated Jul y 2007 US National Toxicology Program Outlines general methods for immunomodulation ? (NTP) General Methods for testing in mice, and methods for hypersensitivity testing Immunotoxicology Studies 6167 http://yosemite.epa.gov/sab/sabproduct.nsf/50E9FC926D852DA7852573FA00754AF3/$File/SAB+im 6168 muno+brief+2+29+08.pdf 6169 6170 7.3.1 Levels of evidence for concluding on immune system toxicity

204

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6171 Based on the results of individual immunotoxicity studies of chemical agents and other test articles, 6172 the US National Toxicology Program (NTP 2009) established a level of evidence system that was 6173 adopted by WHO (2012) to accurately interpret and conclude if a test article should be considered 6174 immunotoxic or not. Five categories (clear evidence, some evidence, equivocal evidence, no 6175 evidence, inadequate evidence of toxicity to the immune system) are defined by the type and 6176 magnitude of effects on immune parameters. According to these criteria, dose-related effects on 6177 functional immune parameters provide the strongest evidence of immunotoxicity and are required 6178 to demonstrate positive (clear or some) evidence of immunotoxicity. In the context of NMs, a 6179 requirement for evidence of a functional change in immunity to indicate clear evidence of 6180 immunotoxicity hazard is applicable, even if the means by which this is established diverges from 6181 those used for other test substances as methods for NM hazard characterization evolve. 6182 6183 7.3.2 Nanomaterial immunotoxicity risk assessment 6184 The adverse effects of any regulated chemical or product on the immune system are assessed as part 6185 of hazard identification according to the guidelines and regulations that govern the specific chemical 6186 or product. Immunotoxicity is one potential hazard that has been addressed, to varying degrees, by 6187 international scientific organizations and committees in their considerations of NM risk assessment 6188 approaches and methodologies (see Table X in this document for a summary of various NM risk 6189 assessment initiatives). Furthermore, immunotoxicity is identified as one of the six elements 6190 considered in the test-strategy for the identification of potential risk of a nanomaterial (Dekkers et al, 6191 2016). Although there is no formal consensus, the overall HHRA concepts for chemicals are generally 6192 considered to be applicable for nanomaterials. Likewise, within current risk assessment frameworks, 6193 hazard characterization methodologies and approaches are generally considered to be appropriate 6194 for NMs. Implicit in this statement is the recognition that issues relevant to NM risk assessment in 6195 general are also relevant to NM immunotoxicity risk assessment. These include uncertainties related 6196 to detecting NMs and characterizing their physicochemical characteristics in biological matrices, 6197 during transport and uptake into tissues, and at their active site(s) in cells, and other issues related to 6198 conducting and interpreting NM animal studies. 6199 6200 Similar to chemicals, prior to initiating risk assessment for nanomaterials induced immunotoxicity, it 6201 is important to understand whether or not imunotoxicity tests should be performed and the types of 6202 immunotoxicity (immunostimulation, immunosuppression, autoimmunity, sensitization) to be 6203 considered. The list of entry points suggested for chemicals (IPCS) are applicable to nanomaterials. 6204 However, only few effects such as, changed cytokine levels, increased incidences of inflammation or 6205 increased inflammatory markers have been extensively characterised for nanomaterials. Studies

205

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6206 reporting on other immunotoxic observations including blood counts, changes in organ weights or 6207 histopathology of immune related organs are scarce. 6208 6209 7.3.2.1 Human clinical studies on NMs 6210 Human epidemiology studies, clinical studies and/or case studies demonstrating immune 6211 suppression, immune stimulation, sensitization or autoimmunity due to NM exposure provide critical 6212 evidence for NM risk assessment; however, are not available at present. As outlined above, the 6213 limitations in the availability of these studies for NM immunotoxicity risk assessment are similar to 6214 their limitations for use in establishing toxicity to any other organ system, namely their small 6215 number, use of descriptive parameters to assess functionality, and experimental design limitations 6216 that challenge the discernment of association and causality. 6217 6218 7.3.2.2 Animal-based and alternative studies. 6219 Animal model-based regulatory guidelines for assessing immunotoxicity, summarized in Table X, have 6220 not yet been widely applied or validated for NMs, although the OECD health effects test guidelines 6221 have been assessed for applicability in nanomaterial hazard characterization (OECD, 2009). Test 6222 guidelines for dermal, oral and inhalation routes of exposure, all applicable to NMs, were considered. 6223 Some gaps and issues identified were relevant to NM immunotoxicity hazard characterization. The 6224 extent to which immune parameters are assessed in repeated dose 28- and 90-day rodent toxicity 6225 studies varies according to the route of exposure. For the oral exposure route Test Guidelines 407 6226 (28-day) and 409 (90-day) have been updated to include additional observational immune endpoints 6227 (as described in Table X above). In contrast, test guidelines for repeated dose inhalation studies of 6228 similar durations (Test Guidelines 412 and 413) have not yet been revised. Since inhalation is a 6229 potential route of human exposure to NMs, the OECD Working Party on Manufactured 6230 Nanomaterials (WPMN) separately considered inhalation toxicity testing protocols for NMs (OECD, 6231 2012). A key recommendation of the WPNM with relevance to NM immunotoxicity was mandatory 6232 bronchiolar lavage fluid (BALF) analysis to assess changes in total and percentage inflammatory cell 6233 populations (OECD 2012, please see section X under hazard identification chapter of this document). 6234 6235 An earlier review of NM risk assessment methodologies by SCENIHR (2007) also identified NM- 6236 specific concerns related to inflammation and immunotoxicity endpoints in animal studies. Although 6237 inflammation is a protective response to cellular insult and not necessarily indicative of a direct effect 6238 on adaptive immunity, inflammasome activation due to particulate exposure has been linked to 6239 adjuvanticity and exacerbation of responses to allergens in atopic individuals (refs). The potential for 6240 inappropriate immune responses due to airway, as well as dermal, NM exposure was raised by

206

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6241 SCENIHR as a question under immunotoxicity hazard assessment. Similar to the OECD Test Guideline 6242 review, bronchiolar lavage fluid assessment was recommended for airway exposure guidelines as a 6243 means of detecting inflammatory responses to NMs (SCENIHR, 2007). Among the OECD guidelines 6244 available for skin and eye, the local lymph node assay (LLNA; Test Guideline 429) was considered to 6245 be appropriate for investigating NM skin sensitization potential, as well as advantageous with respect 6246 to animal welfare considerations and quantity of well-characterized test chemical required. 6247 Few studies have examined NMs using functional assays that have been widely used, standardized or 6248 validated for immunotoxicity risk assessment. TDAR assays have been used to demonstrate 6249 significant dose-dependent adverse or beneficial effects of NMs on immune function. More often, a 6250 TDAR assay is used to demonstrate that a substance in nano form can modulate immunotoxicity due 6251 to a primary chemical treatment. For example, nanocurcumin ameliorated arsenic-induced 6252 suppression of rat IgG responses to keyhole limpet hemocyanin (KLH), and chromium(III) 6253 nanoparticles enhanced anti-Sheep red blood cell (SRBC) IgG responses in heat-stressed rats (Sankar 6254 et al 2013; Zha et al 2009). Rarely to date has a TDAR been used to demonstrate a direct adverse 6255 effect of NMs on immune responses. In one study intravenous exposure to nanosilver for 28d 6256 increased KLH-specific IgG and significantly altered a number of descriptive parameters, providing 6257 support for the plausibility of applying “standard” regulatory immunotoxicity models to NM hazard 6258 characterization (Vanderbriel et al 2014). 6259 Other functional assays used for regulatory purposes include Host-resistance assays - 6260 Sensitization assays (Ema et al, 2011. Reg Tox Pharm 61, 276-281; Park et al, 2011. Tox In Vitro 25, 6261 1863-1869) 6262 Conclusion on new risk assessment approaches for nanomaterials and links with immunotoxic 6263 substances 6264 6265 There appears to be general agreement that parameters indicative of immunotoxicity should be 6266 included in NM hazard characterization, at the very least by assessing observational parameters that 6267 are already incorporated into OECD Health Effects and other standardized test guidelines. Savolainen 6268 et al (2010) proposed a tiered risk assessment approach for well-characterized engineered NMs in 6269 which physicochemical and acellular testing would be conducted in Tier I, in vitro testing in Tier II, in 6270 vivo testing in Tier III, and carcinogenicity and reproduction studies in Tier IV, in which 6271 immunotoxicity assessment is incorporated as a Tier II and a Tier III endpoint. In the proposed 6272 MARINA risk assessment strategy for NMs, immunotoxicity data are considered part of the minimum 6273 information required for data evaluation in Phase I, which was described as the problem framing 6274 phase consisting of data evaluation and relevant exposure scenario identification (Bos et al., 2015). 6275 Furthermore, in the recently developed risk assessment strategy of the NANoREG project, which is

207

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6276 based on six different nano-specific elements, immunotox is identified as one of these six elements. 6277 These elements are the basis of the test-strategy for the identification of potential risk of a 6278 nanomaterial (Dekkers et al, 2016). 6279 Given the complexity associated with risk assessment of nanomaterials, ‘weight of evidence (WOE)’ 6280 approach may be better suited, which generally refers to a method or methods for summarizing and 6281 interpreting all available scientific evidence on health effects, including associated uncertainties, in 6282 the hazard identification step. The process is adaptable, includes expert opinion and professional 6283 judgment, and is often used when assessing complicated systems that require integrating diverse 6284 evidence (Weed et al., 2005; Zuin et al., 2011). Despite the ubiquity of the term WOE in the 6285 regulatory context, it may be used to refer to distinctly different methodological approaches, ranging 6286 from simple evidential summary to complex frameworks for quantitative weighting and scoring of 6287 the available evidence. Examples of the latter have been described for ranking and screening 6288 nanomaterial hazard (Genaidy et al., 2009; Hristozov et al., 2014; Zuin et al., 2011). In the context of 6289 chemical immunotoxicity risk assessment, WOE criteria for immunotoxicity, immunostimulation, 6290 sensitization and autoimmunity have been described (WHO, 2012). Taken together, a quantitative 6291 WOE approach for NM immunotoxicity risk assessment within the context of NM hazard assessment 6292 is feasible. 6293 6294 In conclusion, a validated risk assessment framework for immunotoxicity by nanomaterials does not 6295 exist; however, with the vast amount of data available and the ongoing risk assessment initiatives, 6296 the gaps in knowledge and limitations in the approaches have been identified. Research priorities 6297 that can generate the type of missing data are identified. Considering the potential nanomaterials 6298 variants and lack of data, categorization, grouping or binning tasks are expected to be difficult. Use of 6299 standardized well characterized materials, standardized test methods and biological endpoints, may 6300 help compare toxicity across nanomaterials. An intelligent, mechanism-based tiered testing strategy 6301 taking into account various questions related to the four fundamental steps of risk assessment will 6302 expedite the process without hampering their safe use in many applications. 6303 6304

208

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6305 8. FUTURE RESEACH 6306 There have been important developments in the field of molecular biology, bioinformatics and 6307 toxicology in general, that nanotoxicology stands to benefit from. Below, a few of the advances 6308 made including new technologies and novel toxicity testing paradigms are discussed in the 6309 context of their applicability to nanotoxicology. 6310 6311 8.1 Emerging toxicology paradigm 6312 As exemplified in the previous chapters, toxicology relies heavily on animal-based bioassays to 6313 evaluate toxicity induced by chemicals including nanomaterials. While animal testing is a gold 6314 standard for the regulatory acceptance, the approach lacks capacity to generate quantitative 6315 toxicity data required for the assessment of several classes of nanomaterials and thousands of 6316 variants of each nanomaterial class that requires immediate testing. The issues associated with 6317 the use of single-end point based assays apply to other areas of regulatory toxicology as well; for 6318 example, it is evident that animal-reliant approaches to generate data required for the risk 6319 assessment of thousands of chemicals that are already in commerce and thousands that are 6320 added to the registry every year in a timely manner, due to time and resource intensiveness of 6321 the approach, is ineffective. Thus, a call for a transformative shift in the toxicitiy testing paradigm 6322 was made in 2007 by the National Research council (National Research Council. 2007). In the 6323 paradigm shifting report ‘Toxicity Testing in the 21st Century: Vision and A Strategy’, the National 6324 Research Council called for less reliance on whole animal testing by transitioning from animal- 6325 based conventional toxicity assessment methods to mechanistically based in vitro human cell 6326 culure assays . It was hoped that these targeted in vitro assays, apart from being rapid, 6327 affordable and aligning with the 3R principles, would enable development of predictive 6328 toxicological tools leading to rapid screening of a large number of chemicals including 6329 nanomaterials for their toxic potential, enabling prioritization and categorization for further 6330 toxicological evaluation. However, it was soon realized that such a paradigm shift will require 6331 thorough knowledge of the mode of action of the chemicals being tested, which, at present, is 6332 not available for all substances including for nanomaterials. High-content technology and global 6333 scale screening of genes, proteins and metabolites using omics tools has considerably furthered 6334 the knowledge of diverse pathways or processes perturbed by the chemicals that were 6335 previously unknown. However, the application of this vast data repertoire to derive meaningful 6336 information to support targeted mechanism-based assays, data derived from which can be used 6337 to support regulatory decisions has not been realized so far. 6338 6339

209

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6340 8.2 Global molecular screening to identify perturbed toxicity pathways, biological functions and 6341 processes 6342 Regardless of the approach taken, an important question that needs to be answered for 6343 nanomaterials is ‘what toxicity endpoints should be prioritized?’ as the basis of their future 6344 hazard assessment. The development of comprehensive array of in vitro toxicity tests will require 6345 a detailed understanding of the intricate networking and interactions occurring at the molecular 6346 and cellular level, toxicity pathway responses and molecular perturbations relevant to human 6347 biology. Such understanding cannot be derived from single endpoint-based assays. The systems 6348 biology approaches that combine computational modelling, bioinformatics tools and quantitative 6349 molecular biology techniques will play an important role in revealing those complex biological 6350 interactions (Halappanavar et al, 2017 ). The ability to systematically and quantitatively measure 6351 the genome-wide changes occurring in a specific biological system (genomics, transcriptomics, 6352 proteomics, metabolomics, lipidomics, metagenomics) leading to exhaustive cataloguing of the 6353 changes in the expression of molecular entities such as genes, proteins, small biomolecules and 6354 associated pathways already exists (Costa & Fadeel, 2016). With advances in bioinformatics, an 6355 understanding of the relationships shared between the molecular perturbations and the 6356 toxicological effects are beginning to be unraveled and assembled (e.g., the Adverse Outcome 6357 Pathway of the Organisation for Economic Cooperation and Development) (Vinken, 2013; Labib 6358 et al., 2016). Although this approach is widely used in medical and pharmaceutical research, its 6359 realization in the field of toxicology and specifically in nanotoxicology, will require a large data 6360 repertoire, data visualization and analysis framework and extensive computational power to 6361 reduce and synthesize the data in a biologically meaningful manner for its application to 6362 predictive science. In silico tools for data analysis, data transformation, machine learning and 6363 computational modeling is the absolute prerequisite for the successful implementation of this 6364 new toxicology paradigm. The resulting details on the perturbed system at the different levels of 6365 biological organization can then be used to inform the specific in vitro or in chemico models, and 6366 endpoints to be used in the assessment. 6367 6368 Of the many omic systems, genome-scale gene expression analysis and mapping of biological 6369 pathways have been the mainstays and associated methodologies are well established and 6370 validated for research. Although limited in the context of nanotoxicology, the studies that have 6371 used genomics tools have shown that the gene expression data can be used to 1) qualitatively 6372 relate changes elicited by a single ENM or different classes of ENM (Nikota et al, 2016), 2) 6373 characterize the mechanistic nature of the response to individual ENM (Halappanavar et al, 6374 2016; Nikota etal, 2016), 3) identify sets of genes or gene signatures as markers of exposure to,

210

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6375 and/or adverse effects from exposure to ENM (Williams et al., 2016), 4) identify and prioritize 6376 perturbated biological pathways and processes that correlate with adverse effects , 5) identify 6377 harmful ENM from innocuous ones (Labib et al., 2016; Nikota et al., 2016; Williams et al., 2016). 6378 In addition to these advantages that have obtained regulatory acceptance for incorporation in 6379 the routine testing of chemicals as weight of evidence tools, recent progress in the field of gene 6380 expression analysis enable derivation of gene or pathway based dose-response curves, 6381 mathematical modeling (Labib et al, 2016) and points of departure calculation for the purpose of 6382 conducting human health risk assessment (Labib et al., 2016; Moffat et al., 2015). 6383 6384 While multiplex microarrays are the predominant platforms used in this type of global scale 6385 analyses, pathway-specific quantitative PCR arrays querying targeted subset of genes is also 6386 extensively used to assess chemical-induced hazards. These methods have been applied to 6387 understanding nanomaterials induced toxicity in both in vitro and in vivo models (Poulsen et al., 6388 2014). So far, there have been no reports concerning the interference of nanomaterials with the 6389 RNA isolation, cDNA preparation, dye labeling, microarray hybridization or quantification of dye 6390 signal intensity. However, care should be taken in isolating RNA from cells or tissues. In general, 6391 the established methodologies currently in use for RNA isolation are capable of efficiently 6392 removing the nanomaterials and thus do not face interference from nanomaterials at the 6393 subsequent steps leading to dye labeling and microarray hybridization. 6394 6395 There are very few studies that have used proteomics (Shedova et al., 2012; Teeguarden et al., 6396 2011), another widely used method in the field of chemicals to catalogue global expression 6397 changes at the protein level. Although the approach is used to investigate the mode of action of 6398 nanomaterials (Riebeling et al., 2016), compared to the genome, the chemical composition of 6399 proteome is highly dynamic and standardized protocols for protein separation and isolation, 6400 protein identification and quantification are still emerging. Moreover, the additional layers of 6401 regulation involving post-translation modifications of proteins make it difficult to effectively 6402 interpret the results. Other high-throughput but low-density platforms such as bead arrays that 6403 investigate several cytokines, chemokines and growth factors have been popular and have been 6404 applied to investigating the inflammatory potential of nanomaterials (Husain et al., 2015) and in 6405 determining the specific mechanisms involved in the inflammation (Husain et al., 2013). 6406 Nanomaterials have been shown to interfere with colorimetric and optometric methods involved 6407 in quantification of proteins (Bradford assay) and ELISA reading. Centrifugation of protein lysates 6408 to remove left over nanomaterials has been shown to reduce the assay interference 6409 (Halappanavar et al unpublished data). Until the proteomic methods are further validated, the

211

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6410 proteomics results can only be used to support the gene expression results. The other global 6411 analysis platforms include microRNA profiling and methylation profiling. However very few 6412 studies have applied these tools to understanding the toxicity induced by nanomaterials 6413 Halappanavar et al., 2011) and the resulting data is used to generate research hypothesis or to 6414 support the findings of the apical endpoints and to understand the gene expression results. 6415 Thus, global profiling of biomolecule expression following exposures to nanomaterials will play 6416 an important role in determining the toxicity mechanisms, pathways and processes involved, and 6417 will enable correct selection of in vitro cell systems, identification of important toxicity 6418 endpoints, and development of novel toxicity assays. 6419 6420 8.1 Mode of action and adverse outcome pathways 6421 Through the use of omics technology hundreds of toxicity pathways have been identified for a 6422 variety of nanomaterials in experimental animals or in vitro cell culture models (please see 6423 references above). However, the relationships among these various pathways or networks of 6424 pathways that eventually become toxicity pathways leading to adverse outcomes or disease 6425 progression are yet to be firmly established. In this context, the Adverse Outcome Pathway (AOP) 6426 tool was developed by the OECD to support the new paradigm (NRC 2007) that is reliant on the 6427 underlying mechanisms rather than the direct toxicity observations. AOP is a conceptual 6428 framework consisting of sequential biological events that occur with a molecular initiating event 6429 (MIE) - an initial interaction of a toxicant or a chemical substance with cellular biomolecules and 6430 progressing through several key biological events (KEs) that are interdependent and occurring at 6431 various levels of biological organisation, enventually resulting in a disease or adverse outcome 6432 (AO; Ankley et al., 2010). The MIEs and KEs describe the toxicologically relevant events that are 6433 essential for the disease progression and are measurable. Such organized representation of the 6434 perturbed biology enables development of integrated testing strategies consisting of a 6435 combination of in vitro cell culture models and in silico computational approaches to predict in 6436 vivo toxicity and support risk assessment (Horvat et al., 2016). While only a handful of AOPs have 6437 been developed so far that demonstrate their potential application in chemical risk assessment, 6438 the AOP for ‘Skin Sensitisation Initiated by Covalent Binding to Proteins’, the first AOP to be 6439 developed, reviewed and approved by OECD (McKay et al., 2013) has successfully shown that 6440 well constructed and quantitative AOPs can enable development of mechanistically driven in 6441 vitro methods that can reduce reliance on animal testing and help screen a vast number of 6442 substances for their potential to induce toxicity. As a result of this AOP, a defined approach to 6443 assessing the risk of skin sensitization without animal testing based on Integrated Assessment 6444 and Testing Approaches has been proposed.

212

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6445 6446 While putative AOPs for nanomaterial-induced toxicity are beginning to emerge (Labib et al., 6447 2016; Gerloff et al., 2016; Vietti et al., 2016), quatitative AOPs are yet to be established for 6448 nanomaterials. Hence their implementation in the mechanism-based ENM screening strategy has 6449 not yet been achieved. In the context of nanomaterial-induced immunotoxicity testing, the 6450 systematic organization of the existing scientific information concerning the known 6451 immunotoxicity potential of nanomaterials in an AOP framework will help identify the data gaps, 6452 in formulation of novel and appropriate research questions, and in design of alternative in vitro 6453 predictive assays and testing strategies. Although still in its primitive stages, incorporation of this 6454 concept in the routine investigation of nanomaterial-induced toxicity is essential for the 6455 effectiveness of the predictive strategies. 6456 6457 8.4 High throughput screening (HTS) 6458 High throughput screening refers to industrial scale screening for ‘positive hits’, the chemical 6459 compounds that display significant biological activity in a specific toxicity screen. Originally 6460 designed for drug target identification, high throughput screening has turned into a field of its 6461 own dedicated to the design of toxicological assays that are optimized for speed, efficiency, high- 6462 throughput, high sensitivity and reproducibility. In the recent years, several attempts have been 6463 made in the field of nanotoxicology to develop assays that rapidly assess the toxicity induced by 6464 nanomaterials. Nel et al (2013), Huan Meng et al. (2009), and Damoiseaux et al. (2011), have 6465 developed multi-parametric HTS assays to explore oxidative stress, inflammation and resulting 6466 cytotoxicity induced by several well characterized nanomaterials. This is in agreement with the 6467 trend setting 2007 report from the US national academy of sciences, which advocated for rapid 6468 and efficient toxicity testing by transitioning from qualitative and descriptive animal-based 6469 testing to quantitative, mechanism-based testing using in vitro human cells. However, at present, 6470 the meaning of positive hits from the High throughput oxidative stress or inflammation panel of 6471 assays to adverse outcomes observed in whole animals or in humans following nanomaterial 6472 exposure is not clear (Huang Meng et al. 2009). Although, many nanomaterials are shown to 6473 induce oxidative stress and inflammatory mediators in both in vitro and in vivo models, both of 6474 which can be mounted as organism’s defense mechanism, under what circumstances the 6475 observed oxidative stress or modulation of expression of inflammatory mediators will induce 6476 toxicity is not known. Therefore, it is imperative that detailed mechanisms underlying toxicity 6477 induced by each class of nanomaterial be first sought before establishing the HTS assays. 6478 It is also important to note, that a single HTS assay or data derived from single toxicological 6479 endpoint using a single cell type will not be sufficient to accurately reflect the affected biology,

213

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6480 rather integration of multivariate data (incorporating dose/concentration, post-exposure time 6481 points, nanomaterial properties, biological activity) derived from different types of test systems 6482 using systems biology approach as described above is required (Halappanavar et al. (2017) . 6483 6484 These newer techniques with a focus on predictive approach can help test a number of 6485 nanomaterials and aid in hazard ranking in a manner that is proportionate with the ever growing 6486 list of novel nanomaterials being commercialized. Rigourous efforts are being made in the world 6487 of chemical toxicology to deliver on the promises made by NRC’s landmark document. However, 6488 its acceptance by the regulatory community, who historically have based their regulatory 6489 decisions on observing a adverse phenotype in animals, is something that is going to be debated 6490 for a long time to come. 6491 6492 8.5 Microbiome and impacts on immunotoxicity 6493 The microbiome serves as a gatekeeper between organisms and the environment and they 6494 define the levels of organisms’ exposure to the environment. It is estimated that by cell numbers 6495 humans are ~ 90% microbial. In the recent past, increasing emphasis is laid on understading the 6496 relationhip between the altered microbiome and human health. It has now been shown that 6497 composition of microbiome impacts the toxicity of xenobiotics. Based on this increasing 6498 awareness, a proposal for a new model of health risk assessment that incorporates the link 6499 between the microbiome environmental exposures is suggested. This is especially important in 6500 the context of immunotoxicity as microbiome plays an important role in maturation and function 6501 of immune system, impacting the various immune processes. Thus, future immunotoxicity 6502 assessments must take into consideration the entire microbiome in understading the xenobiotic 6503 ior nanomaterial induced responses. 6504 6505 6506 9. Conclusions and Recommendations 6507 6508 Conclusions 6509 6510 · Numerous studies in recent years using in vitro or in vivo model systems have 6511 demonstrated that ENMs or certain classes of ENMs may be associated with harmful 6512 effects on human health. 6513 · A major challenge in assessing the health risks of ENMs is the lack of systematic 6514 knowledge about

214

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6515 · A major challenge in assessing the human health risks of ENMs to ensure safety is the 6516 lack of systematic knowledge about exposure to these materials. There is also a paucity 6517 of studies on possible long-term effects. 6518 6519 · The particle nature of ENMs predicts their uptake by phagocytic cells including 6520 professional phagocytes such as macrophages, hence the intimate interaction of ENMs 6521 with the immune system. There is considerable concern that some ENMs with fiber-like 6522 dimensions may exert asbestos-like toxicity due to the inability of the innate arm of the 6523 immune system to handle long fibers. In addition to size and shape (aspect ratio), the 6524 chemical composition, surface charge and degree of solubility have all been shown to be 6525 important determinants of ENM toxicity. 6526 6527 · The immune system consists of an innate arm that is able to directly respond to foreign 6528 agents and serves as the first line of defense and an adaptive arm that needs some time 6529 to mount an immune response and is endowed with immunological ‘memory’ (i.e. 6530 responses are specific for certain pathogens and responses are stronger upon 6531 subsequent encounter with the same pathogen). Both arms of the immune system may 6532 be affected by foreign agents. This interaction may result in immunosuppression, 6533 immunostimulation, and/or hypersensitivity and autoimmunity. Immunosuppression 6534 may lead to reduced resisyance to infectious agents and defective immune surveillance 6535 against malignant cells. Recently, numerous reports have identified ENMs as potential 6536 triggers of immunotoxicity. 6537 6538 · The immune system consists of an innate part that is able to directly respond to foreign 6539 agents and an adaptive part that needs some time to develop an immune response. Both 6540 of these may be affected by exposure foreign agents. This interaction may result in 6541 immunosuppression, immunostimulation, and/or hypersensitivity and autoimmunity. 6542 More recently numerous reports have identified ENMs as a potential trigger for 6543 immunotoxicity which has been attributed to the innate as well as to the adaptive 6544 immune system. 6545 6546 · There are several different types of ENMs including carbon-based nanomaterials (carbon 6547 nanotubes, graphene-based materials and others); metal-based nanomaterials; organic 6548 nanomaterials, and lipid-based nanomaterials. There are also composite nanomaterials. 6549 Nanomaterial-enabled products have specific life cycles. Comprehensive characterization

215

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6550 of physicochemical properties of ENMs is a prerequisite for any nanotoxicological study. 6551 Furthermore, ENMs interact with biological fluids in different compartments upon entry 6552 into the body that could have an impact on their biological effects on different organ 6553 systems, including the immune system. 6554 6555 · There are several different types of nanoparticles, such as carbon-based nanomaterials 6556 (single walled carbon nanotubes (SWCNTs, MWCNTs’graphenes and fullerenes); metal- 6557 based nanomaterials (Ag, Au, Fe, Al, etc.), oxides (SiO2, TiO2, CeO2); semiconductors; 6558 organic polymeric nanomaterials (dendrimers, nano cellulose); bio inspired 6559 nanomaterials; and lipid-lipid-based nanomaterials. There are also composite 6560 nanomaterials. All these materials have all their specific life cycles from synthesis to 6561 incorporation in products to disposal or recycling. They will interact with physiological 6562 fluids in different compartments upon entering the body and this may endow ENMs with 6563 a bio-corona of adsorbed biomolecules. They have specific biological effects on different 6564 organ systems, including the immune system. 6565 6566 · Exposure to ENMs may occur in various different settings, i.e., occupational exposure, 6567 consumer exposure, and exposure through environmental contamination. In addition, 6568 deliberate exposure to specific ENMs may occur in the clinical setting (nanomedicine). 6569 Exposure may be inhalatory, orally, or through the skin, and could also occur through 6570 direct administration into the blood stream (in the clinical setting). ENMs may also reach 6571 the blood stream following translocation across biological barriers, such as the air-blood 6572 barrier in the lungs. 6573 6574 · Organs at risk can be identified by toxicokinetic studies. There are currently no guidelines 6575 for assessing toxicokinetics of nanomaterials. It is therefore justified, whenever p[ossible, 6576 to carefully study a absorption, distribution, metabolism and exctretion of the ENMs on 6577 a case by case basis. 6578 6579 · For potential immunotoxic effects, besides composition of the nanomaterial, other 6580 physical and chemical properties of nanomaterials such as size, surface area, shape, 6581 crystallinity, charge, and aggregation are all important factors. 6582 6583 · The immunotoxic effects described for other chemicals may in principle be applicable 6584 also for ENMs. However, while ENMs may be viewed as chemical entities, they are more

216

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6585 complex than traditional chemicals due to the variations of multiple different material 6586 properties in addition to composition. In addition, the so-called bio-corona may also 6587 impact on immune response to ENMs. 6588 6589 · Challenges for toxicity studies of nanomaterials are the dispersion of nanomaterial in 6590 biological systems and the identification of appropriate dose , in terms of mass, number 6591 and surface area. 6592 6593 · For hazard identification, a variety of methods are available, that in principle have all 6594 been used for classical toxicity assessment of chemicals including immunotoxicity. Given 6595 the multitude of nanomaterials, and the drive to minimize the use of laboratory animals 6596 for safety testing, emphasis has been on in vitro methodologies as an alternative 6597 assessing immune effects in vivo . However, many of these have not yet been validated 6598 for the use of testing of ENMs. Moreover, structured (tiered) approaches for 6599 immunotoxicity testing are so far lacking. 6600 6601 · Risk assessment of ENMs should follow the risk assessment paradigm for chemicals, i.e. 6602 hazard identification, hazard characterization, exposure assessment, risk 6603 characterization. A risk assessment framework for immunotoxicity of nanomaterials has 6604 not yet been fully developed. 6605 6606 Recommendations: 6607 6608 · To define the minimal requirements for characterization of test materials before, during and 6609 after the experiments, and control the dispersion of ENMs for both in vivo and in vitro 6610 testing. 6611 6612 · To define reference materials for further testing and standardization and validation of test 6613 systems. 6614 6615 · To advance approaches to relevant exposure regimen, both in vivo and in vitro. and define 6616 the most appropriate dose metrics (i.e., particle mass, number, or surface area) 6617 6618 · To validate test methods for ENMs by ensuring that the test material itself does not interfere 6619 with the assay; and establish validated protocols for endotoxin testing of ENMs

217

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6620 6621 6622 · To explore absorption, distribution, metabolism, and excretion of ENMs on a case-by-case 6623 basis; and how interactions with the immune system affect toxicokinetics of ENMs 6624 6625 · To advance testing for effects of nanomaterials by implementing developments in the field of 6626 molecular biology, systems biology and bioinformatics, high-throughput screening, and in 6627 silico modeling. 6628 6629 · To identiy positive controls for various bioassays. 6630 6631 · To develop more advanced nanomaterial-relevant models to study immunotoxicity. These 6632 may include more advanced in vitro model systems to study the immunotoxicity of ENMs, 6633 taking into account the interplay between immune cell populations and other cells (eg., lung 6634 epithelial cells) as well as the role of the bio-corona on the surface of ENMs 6635 6636 · To investigate develop complex mixture immunotoxicology of engineered nanomaterials. 6637 6638 · To unravel mechanisms of of ENM immunotoxicity and develop adverse outcome pathways, 6639 and define patterns of responses applying to different groups of ENMs to facilitate risk 6640 assessment of ENMs. 6641 6642 • To develop an intelligent, mechanism-based, tiered testing strategy for ENMs taking into 6643 account the four fundamental steps of traditional risk assessment (i.e., hazard identification, 6644 hazard characterization, exposure assessment and risk characterization) 6645 6646 · To identify susceptible populations who are at greatest risk for potential adverse effects of 6647 nanomateriasl on immune-mediated disease (e.e. sex, age, ethnicity, pre-existing condition). 6648 6649 · To establish acceptable uncertainty associated with exposure to nanomaterials. 6650 6651 · Coordinate nanoimmunotox risk assessment vs general nanotox risk assessment. 6652 6653

218

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6654 10. CASE STUDY 1: CARBON NANOTUBES 6655 6656 10.C1.1 Introduction 6657 Immunotoxic responses to carbon nanotubes (CNTs) such as allergic sensitization and 6658 immunosuppression have been documented in experimental animals (primarily rodents) following 6659 inhalation exposure or other pulmonary delivery routes such as oropharyngeal aspiration or 6660 intranasal aspiration (Thompson et al., 2014). However, to date no epidemiological evidence is 6661 available to support immunotoxicity or any adverse outcome in humans exposed occupationally, 6662 through unanticipated environmental contact, or through exposure to consumer products containing 6663 CNTs. This is most likely due to the infancy of the nanotechnology industry, the relatively recent 6664 development of bulk manufacturing of CNTs over the past decade, and the time required for 6665 completing an epidemiological evaluation of adverse health effects in a human population. Based on 6666 data from experimental animals, there has been guidance for occupational exposure limits that 6667 consider CNTs as a single type of material (NIOSH 2013). However, CNTs represent a diverse class of 6668 nanomaterials that have variable physicochemical characteristics; e.g., single-walled (SWCNTs) vs 6669 multi-walled (MWCNTs), different aspect ratios, rigid vs flexible CNTs, or modified forms after CNT 6670 synthesis by surface functionalization. The case-study for CNTs includes a survey of the available data 6671 from rodent studies in vivo that document immunotoxicity in response to well-characterized CNTs. In 6672 addition to experimental evidence, recent risk assessment and hazard evaluation of CNTs has 6673 become available with specific emphasis on cancer (Keumpel et al., 2016; IARC 2016). Since cancer 6674 involves immune cell dysregulation, at least in part, such risk assessment has important implications 6675 for this case-study. Moreover, evidence documenting occupational levels of CNTs measured in the 6676 workplace has important implications for dose-response relationships used in studies with 6677 experimental animals (Erdely et al., 2013; Guseva et al., 2016). Finally, based on the available 6678 experimental evidence, adverse outcome pathway (AOP) information has been performed for fibrotic 6679 lung disease after CNT exposure that has relevance, since the mechanism of fibrosis involves 6680 dysregulation or changes the immune system (Vietti et al., 2015; Labib et al., 2016). 6681 6682 10.C1.2 Background on immune effects induced by CNTs 6683 Immunotoxicity is defined as any adverse effect on the immune system following toxicant exposure 6684 that results in immune stimulation or immune suppression (EHC 180, 1996; Hussain et al., 2012). 6685 Immunostimulation increases the incidence of allergic reactions, pro-inflammatory responses, or 6686 autoimmunity, while immunosuppression suppresses the maturation and proliferation of immune 6687 cells, resulting in increased susceptibility to infectious diseases or tumor growth. In addition, immune 6688 responses against the CNT itself (e.g., allergy) may occur. The available literature on CNTs indicates

219

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6689 both immunostimulatory and immunosuppressive effects in experimental animals or cultured cells. 6690 CNTs are capable of affecting both innate and acquired immune responses. 6691 6692 CNTs primarily impact innate immune function by affecting the function of macrophages, i.e., 6693 professional phagocytic cells. For example, SWCNT form ‘bridge-like’ structures between 6694 macrophages that could impair macrophage functions such as migration or particle/fiber clearance 6695 (Mangum et al., 2006). Long, rigid and biopersistent MWCNTs behave like asbestos fibers to cause 6696 frustrated phagocytosis; the rupture of lysosomal or cell membranes to cause leakage of proteases 6697 and cytokines that cause inflammation (Brown et al., 2007; Donaldson et al., 2010). Lysosomal 6698 membrane damage by MWCNTs is a key factor in releasing cathepsin B into the cytoplasm that plays 6699 a role in inflammasome activation that results in the secretion of interleukin (IL)- -18 as a 6700 primary innate immune response (Palomäki et al., 2011; Hamilton et al., 2012). IL- 6701 through its cognate receptor IL1R is necessary for acute lung inflammation caused by CNTs (Girtsman 6702 et al., 2014). In addition, there is evidence that CNTs impair macrophage clearance mechanisms. For 6703 example, SWCNT pre-exposure impairs macrophage phagocytosis and clearance of Listeria 6704 monocytogenes from the lungs of mice resulting in greater bacterial infectivity (Shvedova et al., 6705 2008). Other innate responses could involve effects of CNTs on the systemic levels of cytokines and 6706 potential effects on the coagulation cascade (Inoue et al., 2008). 6707 6708 The acquired immune response is also affected by pulmonary CNT exposure. For example, there is 6709 evidence in mouse models of allergic asthma that “challenge” exposure to either SWCNTs or 6710 MWCNTs after allergen sensitization exacerbates airway remodeling (fibrosis, mucous cell metaplasia 6711 and mucus hypersecretion, eosinophilic inflammation, increased levels of T-helper 2 (Th2) cytokines 6712 such as IL-13) and elevates levels of allergen-induced serum IgE levels (Inoue et al., 2010; Ryman- 6713 Rasmussen et al., 2009a). While exacerbation of allergen-induced airway disease has been reported 6714 with SWCNT or tangled flexible MWCNTs, most of these CNTs do not directly elicit allergic Th2 6715 responses in the lungs of mice upon a single inhalation or oropharyngeal exposure. Instead the 6716 typical pulmonary immune response to SWCNT or tangled MWCNTs is a Th1 immune response 6717 characterized by acute neutrophilic inflammation that progresses to pulmonary fibrosis or granuloma 6718 formation over the course of several weeks (Bonner et al., 2013; Porter et al., 2010). However, more 6719 recent work demonstrated that rigid. rod-like MWCNTs can directly mediate asthma-like 6720 immunologic effects (Th2 cytokines), pathologic effects (eosinophilia, mucus hypersecretion, fibrosis) 6721 and physiological effects (airway hyper-responsiveness) after inhalation exposure in the absence of 6722 any allergen co-exposure (Rydman et al., 2014). Mast cells were found to partly regulate allergic 6723 inflammation caused by rod-like MWCNTs. However, mast cells are not solely linked to allergic

220

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6724 inflammatory responses, as these inflammatory cells have been demonstrated to play a key role in 6725 pulmonary and cardiovascular response to tangled MWCNTs (Katwa et al., 2012). An emerging 6726 concept is that different types of MWCNTs with different physical and chemical characteristics can 6727 cause widely different acquired immune responses (e.g., Th1 vs Th2 lymphocyte polarization). Even 6728 when different types of CNTs produce seemingly similar inflammatory responses in the lungs of mice, 6729 they may elicit differences in transcriptional and protein biomarkers of fibrosis or innate immune 6730 responses of macrophages (Poulsen et al., 2015; Palomäki et al., 2015). 6731 6732 10.C1.3 Application of the weight of evidence approach for assessment of immunotoxicity 6733 A series of questions is presented that is intended to aid in organizing and characterizing 6734 immunotoxicity data for CNTs from the strongest and most predictive data through the least 6735 predictive evidence supporting human risk for immunostimulatory diseases (allergy, autoimmunity, 6736 hypersensitivity) or immunosuppression. The weight of evidence conclusions developed by 6737 answering these questions summarize the hazard identification for immunotoxicity and should 6738 describe the database in terms of consistency and biological plausibility, including strengths, 6739 weaknesses, uncertainties and data gaps. 6740 6741 10.C1.3.1 Are there epidemiological studies, clinical studies or case studies that provide human 6742 data on end-points relevant to immunostimulation (i.e., unintended stimulation of cellular or 6743 humoral immune function, autoimmunity or allergy?) 6744 Yes. Recently, Shvedova and colleagues conducted a workplace exposure assessment in a MWCNT 6745 manufacturing facility in Russia (Shvedova et al., 2016). They found an average inhalable element 6746 6747 clearly -weighted average of EC. 6748 The mRNA and non-coding (nc)RNA expression profiles in the blood of exposed workers (with at least 6749 six months direct contact with aerosolised MWCNTs) and non-exposed employees of the same 6750 production facility were compared. Results revealed significant changes in the expression of genes 6751 involved in lung inflammation and immunosuppression in MWCNT-exposed humans. The limitation 6752 of this study is that there was no direct evidence of exposure (e.g., CNTs in bronchoalveolar lavage 6753 fluid or nasal washes) although detection of CNTs in the workplace suggests that human exposure 6754 occurred. Other studies documenting workplace levels of CNTs have been conducted and recently 6755 reviewed (Guseva et al., 2016; Kuijpers et al., 2016). However, these studies did not address whether 6756 workers were exposed or if workers had changes in biological responses. The most abundant 6757 information on workplace concentrations is for MWCNT-composites, where the expected release 6758 scenarios may include free standing CNTs, although agglomerated CNTs, and composite fragments

221

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6759 with- and without protruding CNTs are most commonly released entities. The amounts of CNTs 6760 detected in the workplace are commonly related to the Recommended Exposure Limit (REL) (NIOSH 6761 2013). While there is emerging evidence of occupational CNT exposure in humans, it is important to 6762 note that no chronic immune diseases have yet to be linked to exposure. 6763 6764 10.C1.3.2 Is there evidence that exposure to the substance is associated with exacerbation of 6765 hypersensitivity responses, allergy or induction of autoimmune disease or alters the outcome of 6766 host resistance assays? 6767 Yes. There is evidence that CNTs exacerbate allergen-induced airway inflammation and chronic 6768 airway remodeling in mice. MWCNTs or SWCNTs delivered by oropharyngeal aspiration to ICR mice 6769 after ovalbumin allergen pre-exposure (i.e., sensitization) exacerbate allergic airway inflammation 6770 (Inoue et al., 2009; Inoue et al., 2010). Ovalbumin pre-sensitization is also exacerbated by tangled, 6771 flexible MWCNTs delivered by nose-only inhalation exposure in C57BL6 mice (Ryman-Rasmussen et 6772 al., 2009a). These same MWCNTs were shown to exacerbate house dust mite allergen-induced 6773 airway inflammation and fibrosis in C57BL6 mice (Shipkowski et al., 2015). CNTs have also been 6774 reported to directly stimulate allergic airway inflammation. For example, MWCNTs delivered by 6775 oropharyngeal aspiration to ICR mice caused increased distribution of B cells in spleen and in blood 6776 (Park et al., 2009). Whether or not CNTs directly cause allergic lung inflammation or exacerbate lung 6777 inflammation may depend in mouse strain differences. However, the ability of some CNTs to directly 6778 promote allergic lung inflammation also is influenced by physical and/or chemical characteristics. For 6779 example, a comparison of tangled, flexible MWCNTs and rigid, rod-like MWCNTs delivered by 6780 occupationally relevant doses via inhalation exposure showed that rod-like MWCNTs possessed 6781 antigenicity and caused allergic airway inflammation that was characterized by increased numbers of 6782 eosinophils and mucus hypersecretion, along with other cytokine and immunoglobulin biomarkers of 6783 allergic inflammation discussed below (Rydman et al., 2014). In contrast, tangled MWCNTs caused 6784 more conventional neutrophilic lung inflammation with no mucous cell metaplasia or mucus 6785 hypersecretion. 6786 6787 There is also evidence that specific genes play a role in susceptibility to MWCNT-induced 6788 exacerbation of allergic lung disease. For example, COX-2 KO mice have exaggerated lung 6789 inflammation when exposed sequentially to ovalbumin and then MWCNTs. These mice also exhibit 6790 elevated levels of serum IgE and increased IL-13 mRNA (Sayers et al., 2013). STAT1 KO mice are also 6791 susceptible to lung inflammation and fibrosis induced by ovalbumin pre-sensitization followed by 6792 MWCNT exposure by oropharyngeal aspiration and have increased biomarkers of Th2 inflammation 6793 (Thompson et al., 2015).

222

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6794 6795 There is thus far no direct evidence of hypersensitivity reactions or autoimmunity caused by CNTs, 6796 although studies have not yet been conducted in mouse models of autoimmunity and there is a 6797 paucity of information on dermal exposure to CNTs that might give clues to hypersensitivity 6798 reactions. 6799 6800 10.C1.3.3 Is there evidence that exposure to the substance is associated with unintended 6801 stimulation of immune function (antibody production) or alters the balance of immunoregulatory 6802 cytokines? 6803 Yes. Increased levels of IgE, associated with allergic inflammation, are increased by MWCNTs in ICR 6804 mice along with elevated levels of Th2 cytokines such as IL-4 and IL-5 (Park et al., 2009). MWCNTs 6805 delivered by oropharyngeal aspiration to ICR mice that developed allergic lung inflammation also had 6806 increased serum IgE levels (Inoue et al., 2009; 2010). Rod-like MWCNTs that directly stimulated 6807 allergic lung inflammation after a single or repeated inhalation exposure exhibited increased lung 6808 mRNA levels of the Th2 cytokines IL-5 and IL-13 (Rydman et al., 2014). In addition, rod-like MWCNTs 6809 increased the Th2-promoting cytokine IL-33, and eosinophil-attracting chemokine CCL17 was 6810 expressed in BAL cells as well as in other cell types present in lung tissue (Rydman et al., 2014). 6811 Tangled MWCNTs that exacerbated ovalbumin allergen-induced airway inflammation in C57BL6 mice 6812 had increased levels of the Th2 cytokine IL-5 (Ryman-Rasmussen et al., 2009). 6813 6814 Immune stimulation to microbial products is also increased by CNTs. For example, the mycobacterial 6815 antigen ESAT-6, a T cell activator associated with tuberculosis and sarcoidosis, causes increased CD3 6816 (+) lymphocyte infiltration in the lungs of C57BL6 mice and as discussed below also exacerbates 6817 associated lung pathology (Malur et al., 2015). MWCNTs also enhanced ESAT-6-induced production 6818 of the immunoregulatory cytokines osteopontin and chemokine (C-C motif) ligand 2 (CCL2) in mouse 6819 lung tissue. Bacterial lipopolysaccharide (LPS) pre-exposure enhances MWCNT-induced lung fibrosis 6820 in Sprague-Dawley rats and is associated with increased levels of PDGF produced by alveolar 6821 macrophages (Cesta et al., 2010). These studies highlight the overall concept that CNTs exacerbate 6822 microbial immune responses. 6823 6824 10.C1.3.4 Is there evidence that the substance causes immunosuppression and reduces immune 6825 function (antibody production, T cell proliferation, macrophage function, NK cell function, etc.)? 6826 Yes. There is some evidence that MWCNTs cause systemic immunosuppression following inhalation 6827 exposure in mice. Mice exposed to MWCNTs by whole-body inhalation exposure exhibited systemic 6828 immunosuppression (Mitchell et al., 2007). The mechanism was shown to be mediated by the release

223

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6829 of transforming growth factor beta-1 (TGF- 6830 signal cyclooxygenase-2 (COX-2)-mediated increases in PGE2 and IL-10 in the spleen, both of which 6831 function to suppress T cell proliferation (Mitchell et al., 2009). Moreover, T cell proliferation was not 6832 suppressed in COX-2 knockout mice, further confirming the importance of this enzyme in 6833 immunosuppression. 6834 6835 There is evidence that CNTs reduce or suppress macrophage function. SWCNT pre-exposure impairs 6836 macrophage phagocytosis and clearance of Listeria monocytogenes from the lungs of mice resulting 6837 in greater bacterial infectivity (Shvedova et al., 2008). Helical carbon nanotubes inhibit macrophage- 6838 mediated phagocytosis of Pseudomonas aeruginosa (Walling et al., 2013). Furthermore, SWCNTs 6839 delivered to the lungs of rats by intratracheal instillation form ‘bridge-like’ formations that link two 6840 or more macrophages that presumably would impair macrophage functions such as migration or 6841 particle/fiber clearance (Mangum et al., 2006). Long, rigid MWCNTs cause frustrated phagocytosis in 6842 macrophages resulting in leakage of proteases and cytokines that cause inflammation (Donaldson et 6843 al., 2010). 6844 6845 There is also evidence that MWCNTs reduce allergen-induced IgE production in mice when delivered 6846 by inhalation exposure prior to house dust mite allergen exposure. This contrasts to increased levels 6847 of IgE observed following ovalbumin or house dust mite allergen pre-exposure followed by MWCNT 6848 exposure. 6849 6850 10.C1.3.5 Is there histopathological evidence, hematological changes or increases in immune organ 6851 weight that suggest that the substance causes immunostimulation or modulates autoimmunity or 6852 allergy? 6853 Yes. SWCNTs or MWCNTs delivered to the lungs of rats or mice cause pulmonary inflammation and 6854 fibrotic lesions (Mangum et al., 2006; Porter et al., 2010). MWCNTs delivered to the lungs by 6855 inhalation or aspiration migrate to the subpleural tissue; more tangled flexible MWCNTs are found 6856 within subpleural macrophages while long, rigid MWCNTs penetrate the mesothelial lining of the 6857 pleura (Ryman-Rasmussen et al., 2009b; Porter et al., 2010). The inhalation of MWCNTs in mice has 6858 been shown to produce pro-inflammatory lesions on the pleural surface that have been referred to 6859 as mononuclear cell aggregates (Ryman-Rasmussen et al., 2009). Granulomas are a major 6860 pathological feature in the lungs of rodents exposed to MWCNTs. MWCNT-induced granuloma 6861 formation with infiltration of CD3 (+) lymphocytes in the lungs of mice is exacerbated by the 6862 mycobacterial antigen ESAT-6, a T cell activator associated with tuberculosis and sarcoidosis (Malur 6863 et al., 2015). Lung fibrosis is increased in the lungs of rats exposed to MWCNTs and MWCNT-induced

224

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6864 fibrosis is increased by pre-exposure to LPS (Cesta et al., 2010). Mucous cell metaplasia (also 6865 described as airway goblet cell hyperplasia) and mucus hypersecretion are pathological features in 6866 mice exposed to rod-like MWCNTs but not tangled MWCNTs (Rydman et al., 2014). Mucous cell 6867 metaplasia in mice exposed to ovalbumin allergen or house dust mite allergen is enhanced by 6868 exposure to tangled MWCNTs (Thompson et al., 2015; Shipkowksi et al., 2015). Some types of 6869 functionalized MWCNTs (e.g., zinc oxide coated MWCNTs) increase lung inflammation (Dandley et 6870 al., 2016), while other types of functionalized MWCNTs (e.g., COOH-MWCNTs, aluminum oxide 6871 coated MWCNTs) reduce inflammation or fibrosis (Bonner et al., 2013; Taylor et al., 2014). 6872 6873 10.C1.4 Conclusion 6874 The CNT case-study is an illustration of the use of the risk assessment guidance for the assessment of 6875 immunotoxicity. Carbon nanotubes were selected because of the strong database for animal studies, 6876 although gaps in our knowledge of CNTs include lack of information on the potential for autoimmune 6877 disease and systemic effects of CNTs. Also, a limitation is that there is currently no information on 6878 immunotoxic effects of CNTs in humans. The case-study also illustrates an example of the kind of 6879 variation seen with multiple types of CNTs that vary in physico-chemical properties. It should be 6880 noted that this case-study on CNTs is provided with the purpose of illustrating how the risk 6881 assessment guidance can be used for immunotoxicity, but it does not represent a comprehensive risk 6882 assessment, nor does it represent a final regulatory position. 6883 6884 101.CASE STUDY 2. Silver nanoparticles 6885 6886 10C2.1 Introduction 6887 Nanomaterials in consumer products are already widely present in our society. According to the 6888 Nanotechnology Consumer Products Inventory created in 2005 by the Woodrow Wilson 6889 International Center for Scholars and the Project on Emerging Nanotechnologies, more than 1800 6890 consumer products have a claim to contain nanomaterials (Vance et al., 2015). Silver nanoparticles 6891 (Ag-NPs) are among the most used nanomaterials in consumer products. In fact, according to the 6892 product inventory mentioned above, nano-Ag is the nanomaterial that is claimed being associated 6893 with 441 products being the most used nanomaterial (www.nanoproject.org). In addition Ag-NPs are 6894 widely used in medicine as antimicrobial agent. The antimicrobial activity includes antibacterial, 6895 antiviral and antifungal properties. Specifically, Ag-NPs are used in cosmetics, food packaging, dietary 6896 supplements, cloths, toys and especially in medical devices such as bandages, wound and burn 6897 dressing, surgical instruments and implants (Wijnhoven et al., 2009, Reidy et al., 2013). Recent 6898 development of printed electronics also uses Ag-NPs, as conductive ink to manufacture

225

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6899 semiconductors, radiofrequency identification (RFID), and flexible printed circuit boards (FPCBs) and 6900 solar cells (Weldon et al, 2016). However, because of its widespread use there are also concerns on 6901 their toxicity. Inflammatory, oxidative, genotoxic, and cytotoxic effects are observed with Ag-NP 6902 exposure and may be inherently linked (Johnston et al., 2010). For Ag-NPs a complicating factor is 6903 their capacity for (partial) dissolution. This makes it difficult to attribute the resulting positive 6904 (antimicrobial) or negative (toxic) effect to the nanoparticles as also for the silver ions such activity 6905 was demonstrated (Garcia-Reyero 2014, Rosario et al., 2016). A additional complicating factor for 6906 silver ions is that soluble silver salts can be deposit as metallic silver, silver sulfide, and silver chloride 6907 nanoparticles (Van Der Zande et al., 2012, Walczak et al., 2013, Juling et al., 2016). 6908 6909 Silver nanoparticles (Ag-NPs) can be fabricated using physical, chemical, and biological (or green) 6910 syntheses (Ge et al., 2014). For risk assessment the characterization of the Ag-NPs is essential as 6911 different effects have been observed depending on size, coating and dissolution rate of the silver 6912 nanoparticles (Lankveld et al., 2010; Park et al., 2011, Soares et al., 2016). 6913 6914 6915 6916 10.C1.2 Background on immune effects induced by silver nanoparticles 6917 Immunotoxicity is defined as any adverse effect on the immune system following toxicant exposure 6918 that results in immune stimulation or immune suppression (IPCS, 1996, Hussain et al., 2012). 6919 Immunostimulation increases the incidence of allergic reactions, inflammatory responses, or 6920 autoimmunity, while immunosuppression suppresses the maturation and proliferation of immune 6921 cells, resulting in increased susceptibility to infectious diseases or tumor growth. In addition, immune 6922 responses against nanosilver itself (e.g. allergy) may occur. Toxicity of Ag-NPs was noted for several 6923 cells of the immune system including human neutrophils (Soares et al., 2016). Cells and organs 6924 composing the mononuclear phagocytic system (MPS) are a major target of Ag-NPs as indicated by 6925 their organ localization after intravenous administration (Lankveld et al., 2010, Bachler et al., 2013). 6926 6927 Immnunotoxicity determined by cellular effects in vitro 6928 In vitro Ag-NPs of various sizes were cytotoxic for both fibroblasts and macrophage cell lines with the 6929 smaller Ag-NPs of 20 nm being more toxic (Park et al., 2011). The effective concentration inducing a 6930 20% reduced metabolic cell activity (EC20) was 2.7 and 7 µg/mL, respectively for L929 fibroblasts and 6931 RAW 264.7 macrophages. For the RAW 264.7 macrophages the EC20 value of the 20 nm Ag-NPs was 6932 similar to that of ionic silver, while for the fibroblasts the metabolic activity was affected more by the 6933 20 nm Ag-NP (EC20 = 2.8 µg/mL) than for ionic silver (EC20 = 7.1 µg/mL). Macrophages were less

226

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6934 sensitive for membrane damage as indicated by a lack of LDH release while for the fibroblasts clearly 6935 membrane damage was noted. Larger Ag-NPS of 80 and 113 nm were less effective in reducing the 6936 metabolic activity when compared to ionic silver. The 20 nm Ag-NPs induced reactive oxygen species 6937 (ROS) in the RAW 264.7 macrophages. Exposure of RAW 264.7 macrophages to Ag-NPs of all sizes 6938 tested (20 nm, 80 nm, 113 nm) resulted in the release of a variety of inflammatory markers although 6939 the reaction varied widely from a low (<5 fold) increase for IL1-β and IL-10 to a high (>500 fold) 6940 induction for granulocyte colony stimulating factor (GCSF). Comparing the effect on both cell lines, 6941 the results indicate that macrophages may not be the most sensitive cell type for Ag-NP toxicity (Park 6942 et al., 2011). In another study Pratsini et al. (2013) showed that smaller Ag-NPs exerted higher 6943 cytotoxicity in RAW 264.7 macrophages. When compared with ions released from the silver 6944 nanoparticles, the ions dominated the cytotoxicity for the smaller Ag-NPs (< 10 nm) whereas for the 6945 larger Ag-NPs cellular interactions with the nanoparticles seemed to be the dominant factor for 6946 cytotoxicity. 6947 In the study by Lankoff et al. (2012) THP-1 derived human macrophages were less sensitive than 6948 HepG2 and A549 cells for nanosilver uptake by the cells. Nanosilver decreased metabolic activation, 6949 as well as increased cell death (Lankoff et al., 2012). Also Carlson et al. (2008) showed that Ag-NPs 6950 induced ROS in a lung macrophage cell line which ultimately lead to cytotoxicity for the lung 6951 macrophages (Carlson et al., 2008). In addition, an inflammatory response was indicated by the 6952 production of tumor necrosis factor (TNF- α), macrophage inhibitory protein (MIP-2) and interleukin 6953 (IL)-1-β. Barbasz et al., (2015) showed that differentiated monocytes and macrophages are more 6954 resistant to Ag-NP cytotoxicity than their undifferentiated parent cells (Barnasz et al., 2015). For both 6955 undifferentiated and differentiated monocytes/ macrophages nitric oxide (NO) level were increased. 6956 In an in vitro study with murine peritoneal macrophages a decrease in viability and in NO production 6957 was observed at concentrations as low as 0.4 ppm (Shavandi et al., 2011). Furthermore, a significant 6958 decrease in viability was noted in another study for peritoneal macrophages at 10 µg/mL (Xu et al., 6959 2013). 6960 In conclusion, as indicated by a decreased viability and induction of ROS and cytokines Ag-NPs affect 6961 macrophage functionality. Nevertheless, macrophages may be less sensitive for Ag-NP toxic effects 6962 compared to other cell types. These effects on macrophages warrant careful evaluation of Ag-NPs 6963 when used in clinical applications. 6964 6965 Silver NPs show hemolytic activity when incubated in vitro with human blood (Choi et al., 2011). Two 6966 silver nanoparticle preparations (obtained from two different suppliers) showed a higher hemolytic 6967 activity for red blood cells (RBC) when compared to micron sized silver particles at equal mass 6968 concentrations (dose >220 µg/mL or >10 cm2/mL). The increased hemolysis was attributed to an

227

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

6969 increase in ion release for the nanoformulations when compared to the micronsized particles. For 6970 the Ag-NP preparation considerable size changes were noted when incubated with PBS and media 6971 components. The increased hemolysis was related to their greater surface area, increased silver ion 6972 release, and direct interaction with RBCs (Choi et al., 2011). When nanosilver was incorporated in a 6973 nanocomposite polymeric material (polyhedral-oligomeric-silsesquioxane-poly(carbonate- 6974 urea)urethane) intended to be used in medical devices for cardiovascular applications higher silver 6975 levels in the polymeric material (> 0.75% by weight) demonstrated a hemolytic tendency, whereas 6976 lower levels (< 0.40% by weight) did not show hemolysis (De Mel et al., 2012). The mechanisms by 6977 which Ag-NPs exert their toxic activity for erythrocytes can be due to the release of silver ions, 6978 interaction of the Ag-NPs themselves with the erythrocyte membrane or a combination thereof 6979 (Wildt et al., 2013). In contrast, in vivo exposure by inhalation or intravenous injection did not show 6980 long term effects of Ag-NPs on RBC (Ji et al., 2007, Sung et al., 2009, Tiwari et al., 2011). 6981 Nevertheless, at early time intervals after administration RBC toxicity was indicated by a reduced 6982 hemoglobin, RBC and hematocrit levels in blood (Tiwari et al., 2011). After repeated intravenous 6983 administrations for 28 days several red blood cell parameters (Hemoglobin (Hb), Haematocrit (Hct), 6984 Mean Corpuscular Volume (MCV), Mean Corpuscular Hemoglobin Concentration (MCHC)) were 6985 decreased while the RBC cell number itself was unaffected (De Jong et al., 2013). In a two weeks oral 6986 study in mice with Ag-NPs the values of RBC, Hb, and Hct did not vary significantly in the control and 6987 Ag-NP-treated (size range 35-45 nm) animals. Liver enzyme determination in blood indicated liver 6988 toxicity (Heydrnejad et al., 2015). 6989 6990 Ultra-small silver nanoclusters (size 1.8 ± 0.3 nm as determined by TEM) were cytotoxic for 6991 peripheral blood mononuclear cells (PBMC) at concentrations as low as 1 µg/mL with incubation 6992 times up to 24 hours (Orta-García et al., 2015). In addition, intracellular ROS has been induced with a 6993 similar level of ROS activation independent of the exposure dose (ranging from 0.1 – 5 µg/mL). In 6994 addition, no effect of incubation time was present over a time period ranging from 3 to 12 hours. 6995 6996 Immnunotoxicity determined by effects in vivo 6997 It is well known that particulates including nanoparticles induce lung inflammation after inhalation. 6998 For Ag-NPs a series of inhalation studies were performed by a research group from Korea (Ji et al., 6999 2007, Sung et al., 2009, Song et al., 2013). Sung et al., (2009) reported that dose dependent 7000 increases in lesions related to Ag-NP exposure were indicated by histopathological examinations 7001 including mixed inflammatory cell infiltrate, chronic alveolar inflammation, and small granulomatous 7002 lesions. In male rats the lung inflammation persisted after a 12 week recovery period post exposure 7003 while female rats showed a gradual recovery (Song et al., 2013). In a 10 day inhalation exposure

228

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7004 study in mice both macrophages and neutrophils numbers were increased in the bronchoalveolar 7005 lavage fluid (BALF) (Stebounova et al., 2011).The inflammation persisted for at least three weeks 7006 following the last exposure. 7007 One of the characteristics of particle induced lung inflammation is the attraction of neutrophils as 7008 demonstrated in the bronchioalveolar lavage fluid (BALF) (Braakhuis et al., 2016). Both similar and 7009 distinct different activities were shown in a study with human neutrophils exposed to 20 nm and 70 7010 nm Ag-NPs (Poirier et al., 2016). Both sizes of Ag-NPs acted as inhibitors of de novo protein 7011 synthesis, but showed opposing effects on neutrophil apoptosis. Opposite effects were found with 7012 regard to neutrophil apoptosis; Ag-NPs of 20 nm induced apoptosis whereas the 70 nm Ag-NPs 7013 delayed apoptosis. Furtheremore, Ag-NPs of 20 nm induced CXCL8 chemokine (IL-8) production, and 7014 albumin and metalloproteinase-9 levels in the culture supernatants. Both Ag-NPs forms did not 7015 induce ROS formation in the neutrophils. 7016 7017 Two in vivo 28 days repeated dose toxicity studies showed immunotoxicity as the most sensitive 7018 parameter (De Jong et al., 2013, Vandebriel et al., 2014). Two different sizes of Ag-NPs (20 nm and 7019 100 nm) were intravenously administered in male and female Wistar derived rats (De Jong et al., 7020 2013). The doses evaluated were ranging from 0.0082 mg/kg body weight (b.w.) per day to 6 mg/kg 7021 b.w./day for the 20 nm Ag-NPs, and 6 mg/kg b.w./day for the 100 nm Ag-NPs. General toxicity and 7022 immunotoxicity was evaluated one day after the last administration. Treatment with a maximum 7023 dose of 6 mg/kg b.w. was well tolerated by the animals. However, both for 20 nm and 100 nm Ag- 7024 NPs growth retardation was observed during the treatment. There was a decrease in body weight 7025 and thymus weight, and an increase in liver and spleen weight. Both thymus and spleen effects may 7026 be indications for possible immunotoxicity. This increase in spleen weight was due to an absolute 7027 increase in both T and B cell number whereas the relative cell numbers remained constant. Brown 7028 and black pigment indicating Ag-NP accumulation was observed by histopathology in liver, spleen, 7029 and lymph nodes. Clinical chemistry indicated liver damage (increased alkaline phosphatase, alanine 7030 transaminase, and aspartate transaminase) that could not be confirmed by histopathology. 7031 Hematology showed a decrease in several red blood cell parameters. The most striking toxic effect 7032 was the almost complete suppression of the natural killer (NK) cell activity in the spleen at high doses 7033 for both the 20 nm and 100 nm Ag-NPs. After treatment of 20 nm Ag-NPs for mitogen stimulated 7034 spleen cells a decrease in interferon-γ and interleukin (IL)-10 production after concanavalin-A 7035 stimulation was noted, while after lipopolysaccharide stimulation decreased IL-6, IL-10 levels and 7036 TNF-α production was present as well as an increased IL-1β production. After the 100 nm Ag-NP 7037 treatment only the IL-10 production was decreased. In addition, an increase in serum IgM and IgE 7038 antibodies and an increase in blood neutrophilic granulocytes was observed. For the spleen weight a

229

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7039 critical effect dose (CED) of 0.37 mg/kg b.w. could be established. The lowest critical effect dose 7040 (CED) for a 5% change compared to control animals was observed for thymus weight (CED05 0.01 7041 mg/kg b.w.) and for functional immune parameters, i.e. decrease in NK cell activity (CED05 0.06 7042 mg/kg b.w.) and LPS stimulation of spleen cells (CED05 0.04 mg/kg b.w.). These results show that for 7043 nanosilver effects on the immune system were the most sensitive parameters for potential adverse 7044 responses. In a follow-up study by Vandebriel et al., (2014) male rats were treated for 28 days with a 7045 similar dose range of 20 nm Ag-NPs to investigate functional activity of the immune system. This was 7046 performed by measuring the T-cell dependent antibody (TDAR) to keyhole limpet hemocyanin (KLH). 7047 A reduction in KLH-specific IgG was observed, with a lowest 5% lower confidence bound of the BMD 7048 (BMDL) of 0.40 mg/kg b.w./day. This suggests that Ag-NPs induce suppression of the functional 7049 immune system. Other parameters sensitive to Ag-NP exposure were in line with the previous study: 7050 a reduced thymus weight with a BMDL of 0.76 mg/kg b.w./day, and an increased spleen weight, 7051 spleen cell number, and spleen cell subsets, with BMDLs between 0.36 and 1.11 mg/kg b.w./day. 7052 Both studies show that the systemic intravenous exposure of the immune system resulted in 7053 immunotoxicity as indicated by the almost complete suppression of the NK cell activity and a reduced 7054 IgG antibody production. 7055 7056 A 28-day subacute oral toxicity of 60 nm Ag-NPs in male and female rats was performed by oral 7057 administration of doses of 30, 300, 1000 mg/kg of body weight. No significant changes in body 7058 weight observed, while significant dose-dependent changes were found in alkaline phosphatases and 7059 cholesterol values in both the male and female rats, indicating that exposure to more than 300 7060 mg/kg b.w. of Ag-NPs may result in liver damage with an increased incidence of bile-duct hyperplasia 7061 (Kim et al., 2008). An increase in some red blood cell parameters (Red Blood Cells, Hemoglobin, and 7062 Hematocrit) was observed in female rats whereas the increase present in male rats was not 7063 significantly different from controls. The NOAEL reported by this study was 30 mg/kg body weight 7064 (being the lowest tested dose). In this study, a dose-dependent increased accumulation of Ag-NP was 7065 observed in the lamina propria in the small and large intestine and in the tip of the upper villi in the 7066 ileum and protruding surface of the fold in the colon. The Ag-NP treated rats showed higher numbers 7067 of goblet cells releasing mucus granules in the crypt and ileal lumen. Lower amounts of neutral and 7068 acidic mucins were found in the goblet cells and the amount of sialomucins was increased, while the 7069 amount of sulfomucins was decreased. In the colon of the Ag-NP treated rats, sialyated mucins were 7070 detected in the lamina propria. This study suggested that Ag-NPs were a powerful intestinal 7071 secretagogue and induced an abnormal mucin composition in the intestinal mucusa (Jeong et al, 7072 2010). 7073

230

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7074 A 13 weeks subchronic oral toxicity study of Ag-NPs (56 nm) was conducted at doses of 30, 125, and 7075 500 mg/kg b.w. (Kim et al., 2010). A significant decrease in the body weight of the male rats (p<0.05) 7076 was observed after 4 weeks of exposure. Consistent with the 28-day oral toxicity study by Kim et al 7077 (2008), significant dose-dependent changes of alkaline phosphatase and cholesterol were found 7078 indicating liver damage. Histopathology of the liver indicated a higher incidence of bile-duct 7079 hyperplasia in exposed male and female rats compared with controls. There were no effects on red 7080 or white blood cells. The NOAEL reported from this study was 30 mg/kg body weight (being the 7081 lowest tested dose). 7082 7083 In a 13 week oral repeated dose toxicity study with Ag-NPs of 10 nm, 75 nm, and 110 nm from the 7084 same supplier (nanoComposix, San Diego, USA) as in the studies above no meaningful effects of Ag- 7085 NP on the absolute organ weights and relative organ weights (%; organ weight/necropsy body 7086 weight) were observed (Boudreau et al., 2016). An effect of particle size on tissue distribution was 7087 observed as the silver concentrations were higher in the blood and bone marrow of rats exposed to 7088 Ag-NP 10 nm when compared to 75 nm or 110 nm. Brown pigment was observed by histopathology 7089 in various organs which was considered a measure of silver mobility rather than toxicity. The study 7090 demonstrated for the 10 nm Ag-NPs the translocation, and thus systemic availability, from the 7091 gastro-intestinal tract to organs and tissues as mostly intact particles. Histopathology and other 7092 parameters did not indicate toxicity after the 13 week oral exposure to the Ag-NPs. 7093 7094 It has been established that percutaneous exposure to powdered silver, silver solutions, and dental 7095 amalgams can induce allergic contact dermatitis (Catsakis & Sulica, 1978; Heyl, 1979; Marks, 1966). 7096 Acute eye and dermal irritation/corrosion tests using rabbits were conducted with Ag-NPs (average 7097 10 nm) and revealed no significant clinical signs or mortality and no acute irritation or corrosion 7098 reaction for the eyes and skin (Kim et al., 2013). A skin sensitization test using guinea pigs reported 7099 Ag-NPs as a weak skin sensitizer, showing discrete or patchy erythema (Kim et al., 2013). 7100 7101 In contrast to macrophage toxicity observed in vitro (see above) Xu et al. (2013) also demonstrated 7102 that Ag-NPs can exert adjuvant activity by stimulating antigen processing cells like macrophages (Xu 7103 et al., 2013). When administered by intraperitoneal and subcutaneous immunization with ovalbumin 7104 (OVA) or bovine serum albumin (BSA) as antigen, serum antigen specific IgG and IgE levels were 7105 significantly increased. The serum levels were lower than those obtained with Freund’s Complete 7106 Adjuvant (FCA) but similar to the levels obtained with Alum adjuvant. In these studies Ag-NPs were 7107 mixed with either OVA or BSA antigen. Peritoneal macrophages were activated by the 7108 intraperitoneally administered Ag-NPs as indicated by the increase in the number harvested and

231

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7109 increase in TNF-α and IFN-γ in the peritoneal lavage fluid. Supporting in vitro studies performed with 7110 low toxicity Ag-NPs concentration (10 µg/mL) using peritoneal exudate macrophages, showed that 7111 there was no effect of Ag-NP exposure on antigen uptake by the peritoneal macrophages. An 7112 increased IgG1/IgG2a and IgE response indicated a Th2-mediated immune response. The authors 7113 attributed the adjuvant activity of the Ag-NPs to the recruitment and activation of local leukocytes 7114 (Xu et al., 2013). Also Chuang et al., (2013) demonstrated the adjuvant activity of Ag-NPs after 7115 inhalation exposure before an antigen challenge in the lung in intrperitoneally OVA sensitized mice. 7116 The Ag-NP inhalation had allergic and inflammatory effects in both healthy and allergic mice. 7117 7118 More recently for green synthesized Ag-NPs adjuvant activity was demonstrated for use in a rabies 7119 vaccine for veterinary use in vivo in mice (Asgary et al., 2016). Different amounts of Ag-NPs were 7120 mixed with inactivated rabies virus and administered intraperitoneally twice with one week interval 7121 (day 1 and day 7) in mice. Animals were intracerebrally challenged at day 14 with rabies virus and 7122 survival and neutralizing antibodies were determined at day 35. Adjuvant activity of the Ag-NPs was 7123 demonstrated as survival and induction of virus neutralizing antibodies were similar to the positive 7124 Alum control vaccine used (Asgary et al., 2016). 7125 7126 Epidemiology 7127 Colloidal silver has been used already for a long time (Nowack et al., 2011) Silver and silver 7128 compounds are considered to have low toxicity in humans (Chen and Schluesener 2008). The most 7129 well known clinical condition is argyria which is characterized by a striking blueish-gray coloring of 7130 the skin and the eye (Drake and Hazelwood 2005). The discoloration is caused by silver salt 7131 deposition in the skin and eye. 7132 7133 A surveillance case study of workers involved in the manufacture of Ag-NPs reported by Lee et al 7134 (2012) identified two male workers who had worked for 7 years manufacturing silver nanomaterials 7135 exposed to silver concentrations of 0.35 and 1.35 µg/m3 based on personal air sampling. These two 7136 workers showed silver blood levels of 0.034 and 0.0135 µg/dL and urinary silver levels of 0.043 µg/dL 7137 and not detected, respectively. No health effects were observed and blood hematology data were in 7138 the normal range. One case report identified a 27 year old man involved in plating mobile telephone 7139 subunits with aerosolized silver for 4 years. The man showed general argyria and hallmark blue-gray 7140 skin and mucosa pigmentation with elevated serum silver levels of 15.44 µg/dL (normal range 1.1-2.5 7141 µg/dL) and a urinary silver concentration of 243.2 µg/L (normal range 0.4-1.4 µg/L). No other adverse 7142 physical or organ effects were observed (Cho et al, 2008). Although the exact use of Ag-NPs in the

232

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7143 workplace was uncertain exposure to silver particles, whether nanoscale or non-nanoscale, appeared 7144 to induce the general argyria within 4 years. 7145 7146 10. C1.3 Application of the weight of evidence approach for assessment of immunotoxicity 7147 7148 A series of questions is presented that is intended to aid in organizing and characterizing 7149 immunotoxicity data for nanosilver from the strongest and most predictive data through the least 7150 predictive evidence supporting human risk for immunostimulatory diseases (allergy, autoimmunity) 7151 or immunosuppression. The weight of evidence conclusions developed by answering these questions 7152 summarize the hazard identification for immunotoxicity and should describe the database in terms of 7153 consistency and biological plausibility, including strengths, weaknesses, uncertainties and data gaps. 7154 7155 10.C1.3.1 Are there epidemiological studies, clinical studies or case studies that provide human data 7156 on end-points relevant to immunostimulation (i.e., unintended stimulation of cellular or humoral 7157 immune function, autoimmunity or allergy?) 7158 Yes. In general epidemiological studies are limited to case studies identifying an increase in silver 7159 concentration in workers without clinical signs of immune effects. However, it has been established 7160 since long that percutaneous exposure to powdered silver, silver solutions, and dental amalgams can 7161 induce allergic contact dermatitis (Catsakis & Sulica, 1978; Heyl, 1979; Marks, 1966). 7162 C1.3.2 Is there evidence that exposure to the substance is associated with exacerbation of 7163 hypersensitivity responses, allergy or induction of autoimmune disease or alters the outcome of host 7164 resistance assays? 7165 7166 Yes. Intraperitoneal and subcutaneous immunization with ovalbumin (OVA) or bovine serum albumin 7167 (BSA) as antigen, resulted in an increased serum antigen specific IgG and IgE levels when Ag-NPs 7168 were administered together with the antigen (Xu et al., 2013). Also inhalation of Ag-NPs before a 7169 nasal OVA challenge in OVA immunized mice resulted in an increased allergic response (Chuang et 7170 al., 2013). 7171 7172 10.C1.3.3 Is there evidence that exposure to the substance is associated with unintended stimulation 7173 of immune function (antibody production) or alters the balance of immunoregulatory cytokines? 7174 7175 Yes. Macrophages were activated after in vivo exposure to Ag-NPs as indicated by the increase in the 7176 number harvested and increase in TNF-α and IFN-γ in the peritoneal lavage fluid (Xu et al., 2013).

233

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7177 Intravenously administered Ag-NPs altered the cytokine pattern secreted by mitogen stimulated 7178 spleen cells (De Jong et al., 2013). 7179 7180 10.C1.3.4 Is there evidence that the substance causes immunosuppression and reduces immune 7181 function (antibody production, T cell proliferation, macrophage function, NK cell function, etc.)? 7182 7183 Yes. The study of De Jong et al., (2013) showed an almost complete disappearance of NK cell activity 7184 in spleen cells. In addition, a reduction of T and B cell mitogen responses by spleen cells was 7185 demonstrated. For T cells a decrease in IL-10 and interferon-γwas observed. For B cells a decrease in 7186 IL-6, IL-10 levels and TNF-α production was present, while IL-1β production was increased. In the in 7187 vivo study by Vandebriel et al., (2014) a reduction in KLH-specific IgG was observed after Ag-NPs 7188 exposure in KLH immunized animals. 7189 7190 10.C1.3.5 Is there histopathological evidence, hematological changes or increases in immune organ 7191 weight that suggest that the substance causes immunostimulation or modulates autoimmunity or 7192 allergy? 7193 7194 Yes. Silver NPs show hemolytic activity (Choi et al., 2011, De Mel et al., 2012). The mechanisms by 7195 which Ag-NPs exert their toxic activity for erythrocytes can be due to the release of silver ions, 7196 interaction of the Ag-NPs themselves with the erythrocyte membrane or a combination thereof 7197 (Wildt et al., 2013). Ultra-small silver nanoclusters were found to be cytotoxic for PBMC (Orta-García 7198 et al., 2015). 7199 7200 An increase spleen weight was demonstrated by the studies of De Jong et al., (2013) and Vandebriel 7201 et al., (2014) after intravenous administration of Ag-NPs. In several oral exposure studies effects on 7202 spleen weight were not observed (Kim et al., 2008, Kim et al., 2010, Boudreau et al., 2016). Changes 7203 in hematology parameters were observed in both intravenous and oral toxicity studies (Kim et al., 7204 2008, De Jong et al., 2013, Vandebriel et al., 2014). 7205 7206 10.C1.4 Conclusion 7207 7208 Silver nanoparticles (Ag-NPs) show immunotoxic properties that are depending on the size of the 7209 nanoparticles in which smaller Ag-NPs seem to have a larger effect when based on a dose expressed 7210 in mass. The evidence for immunotoxic activity of Ag-NPs is based on both in vivo animal studies and 7211 in vitro studies with various cells of the immune system. The mechanism of the toxic activity is not

234

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7212 known but is partially caused by Ag ion release. Also interactions of the Ag-NPs with the cells can 7213 lead to cytotoxicity. In humans (nano)silver toxicity is mainly observed in the form of argyria. 7214 7215 7216 7217 11. GLOSSARY OF TERMS 7218 Use existing WHO descriptions where available 7219 7220 7221 7222

235

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7223 12. REFERENCES (note to reviewers, a number of references are incomplete but will be provided 7224 in the final document). 7225 7226 Abid AD, Anderson DS, Das GK, Van Winkle LS, Kennedy IM., (2013). Novel lanthanide-labeled metal 7227 oxide nanoparticles improve the measurement of in vivo clearance and translocation. Part Fibre 7228 Toxicol 10:1 7229 7230 ADDR med Kunal 2013 7231 7232 Agban Y, Lian J, Prabakar S, Seyfoddin A, Rupenthal ID. Nanoparticle cross-linked collagen shields for 7233 sustained delivery of pilocarpine hydrochloride. Int J Pharm. 2016 Mar 30;501(1-2):96-101. 7234 7235 Aitken RJ, Creely KS, Tran CL. 2004. Nanoparticles: An occupational hygiene review, Research Report 7236 274. Prepared by the Institute of Occupational Medicine for the Health and 5 Safety Executive, North 7237 Riccarton, Edinburgh, England. 7238 7239 Aillon KL, et.al., Adv Drug Deliv Rev. 2009 Jun 21; 61(6): 457–466. 7240 7241 Akhavana O, Ghaderia E. Escherichia coli bacteria reduce graphene oxide to bactericidal graphene in 7242 a self-limiting manner. Carbon 2012; 50, 1853-1860. 7243 7244 Alfaro-Moreno E, Nawrot TS, Vanaudenaerde BM, Hoylaerts MF, Vanoirbeek JA, Nemery B et al.: Co- 7245 cultures of multiple cell types mimic pulmonary cell communication in response to urban PM10. Eur 7246 Respir J 2008, 32: 1184-1194.¨ 7247 7248 Al-Hanbali O, Rutt KJ, Sarker DK, Hunter AC, Moghimi SM. Concentration dependent structural 7249 ordering of poloxamine 908 on polystyrene nanoparticles and their modulatory role on complement 7250 consumption. J Nanosci Nanotechnol. 2006 Sep-Oct;6(9-10):3126-33. 7251 7252 Amenta V, Aschberger K, Arena M, Bouwmeester H, Moniz FB, Brandhoff P, Gottardo S, Marvin HJP, 7253 Mech A, Pesudo LQ, Rauscher H, Schoonjans R, Vettori MV, Weigel S, Peters RJ (2015) Regulatory 7254 aspects of nanotechnology in the agri-feed-food sector in EU and non-EU countries. Reg Tox Pharm 7255 DOI 10.1016/j.yrtph.2015.06.016. 7256

236

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7257 Andersson-Willman B, Gehrmann U, Cansu Z, et al. Effects of subtoxic concentrations of TiO2 and 7258 ZnO nanoparticles on human lymphocytes, dendritic cells and exosome production. Toxicol Appl 7259 Pharmacol. 2012; 264(1):94-103. 7260 7261 Andreani T, de Souza AL, Kiill CP, Lorenzón EN, Fangueiro JF, Calpena AC, Chaud MV6, Garcia ML, 7262 Gremião MP, Silva AM, Souto EB. Preparation and characterization of PEG-coated silica nanoparticles 7263 for oral insulin delivery. Int J Pharm. 2014;473(1-2):627-35. 7264 Andón FT, Fadeel B. Programmed Cell Death: Molecular Mechanisms and Implications for Safety 7265 Assessment of Nanomaterials. Acc Chem Res. 2013 Mar 19;46(3):733-42. 7266 7267 Andón FT, Mukherjee SP, Gessner I, Wortmann L, Xiao L, Hultenby K, Shvedova AA, Mathur S, Fadeel 7268 B. Hollow carbon spheres trigger inflammasome-dependent IL-1β secretion in macrophages. 7269 2017;113: 243–251. 7270 7271 Ankley,G. T., Bennett, R. S., Erickson, R. J., Hoff, D. J., Hornung, M.W., Johnson, R. D., Mount, D. R., 7272 Nichols, J. W., Russom, C. L., Schmieder, P. K., et al. (2010). Adverse outcome pathways: A conceptual 7273 framework to support ecotoxicology research and risk assessment. Environ. Toxicol. Chem. 29, 730– 7274 741. 7275 7276 OECD 2013 7277 7278 Aschberger K, Johnston HJ, Stone V, Aitken RJ, Hankin SM, Peters SA, Tran CL, Christensen FM. 7279 (2010a). Review of carbon nanotubes toxicity and exposure–appraisal of human health risk 7280 assessment based on open literature. Crit Rev Toxicol 40:759–790. 7281 7282 Aschberger K, Johnston HJ, Stone V, Aitken RJ, Tran CL, Hankin SM, Peters SA, Christensen FM. 7283 (2010b). Review of fullerene toxicity and exposure–appraisal of a human health risk assessment, 7284 based on open literature. Regul Toxicol Pharmacol 58:455–473. 7285 7286 Aschberger K, Micheletti C, Sokull-Klüttgen B, Christensen FM. (2011). Analysis of currently available 7287 data for characterising the risk of engineered nanomaterials to the environment and human health– 7288 lessons learned from four case studies. Environ Int 37:1143–1156. 7289

237

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7290 K. Aschberger, H. Rauscher, H. Crutzen, F.M. Christensen, K. Rasmussen, B. Sokull-Kluettgen, H. 7291 Stamm. Considerations on Information Needs for Nanomaterials in Consumer Products – Discussion 7292 of a Labelling and Reporting Scheme for Nanomaterials in Consumer Products in the EU. ISPRA 2014 7293 7294 Asgary V, Shoari A, Baghbani-Arani F, Sadat Shandiz SA, Khosravy MS, Janani A, Bigdeli R, Bashar R, 7295 Cohan RA. Green synthesis and evaluation of silver nanoparticles as adjuvant in rabies veterinary 7296 vaccine. Int J Nanomedicine 11, 3597-3605, 2016. 7297 7298 Asgharian B, Hofmann W, Bergmann R (2001): Particle Deposition in a Multiple-Path Model of the 7299 Human Lung, Aerosol Science and Technology, 34:4, 332-339 7300 7301 Asgharian B, Price O, Miller F, Subramaniam R, Cassee FR, Freijer J, et al. 2009. Multiple-Path Particle 7302 Dosimetry Model (MPPD v2.11): A Model for Human and Rat Airway Particle Dosimetry. In 7303 Hamner Institutes for Health Sciences Applied Research Associates (ARA), National Institute for 7304 Public Health and the Environment (RIVM), and Ministry of Housing, Spatial Planning and the 7305 Environment, editor. Raleigh, NC: Applied Research Associates (ARA). 7306 7307 AshaRani PV, Low Kah Mun G, Hande MP, Valiyaveettil S. Cytotoxicity and genotoxicity of silver 7308 nanoparticles in human cells. ACS Nano. 2009 Feb 24;3(2):279-90. 7309 7310 Astashkina A, Jones CF, Thiagarajan G, Kurtzeborn K, Ghandehari H, Brooks BD, Grainger DW. 7311 Nanoparticle toxicity assessment using an in vitro 3-D kidney organoid culture model. Biomaterials. 7312 2014;35(24):6323-31. 7313 7314 Ault AP, Stark DI, Axson JL, Keeney JN, Maynard AD, Bergin IL, Philbert MA. Protein corona-induced 7315 modification of silver nanoparticle aggregation in simulated gastric fluid. Environ Sci Nano. 2016 Dec 7316 1;3(6):1510-1520. 7317 7318 Auttachoat, W., et al., Route-dependent systemic and local immune effects following exposure to 7319 solutions prepared from titanium dioxide nanoparticles. J Immunotoxicol, 2014. 11(3): p. 273-82. 7320 7321 Awaad A, Nakamura M, Ishimura K. Histochemical and biochemical analysis of the size-dependent 7322 nanoimmunoresponse in mouse Peyer's patches using fluorescent organosilica particles. Int J 7323 Nanomedicine. 2012;7:1423-1439. 7324

238

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7325 Ayala V, Herrera AP, Latorre-Esteves M, Torres-Lugo M, Rinaldi C. Effect of surface charge on the 7326 colloidal stability and in vitro uptake of carboxymethyl dextran-coated iron oxide nanoparticles. J 7327 Nanopart Res. 2013;15(8):1874. 7328 7329 Bachler G, von Goetz N, Hungerbühler K. A physiologically based pharmacokinetic model for ionic 7330 silver and silver nanoparticles. Int J Nanomedicine. 8, 3365-3382, 2013. 7331 7332 Baig MS, Ahad A, Aslam M, Imam SS, Aqil M, Ali A. Application of Box-Behnken design for preparation 7333 of levofloxacin-loaded stearic acid solid lipid nanoparticles for ocular delivery: Optimization, in vitro 7334 release, ocular tolerance, and antibacterial activity. Int J Biol Macromol. 2016 Apr;85:258-70. 7335 7336 Baisch et al 2014, Particle and Fibre Toxicology 2014, 11:5 doi:10.1186/1743-8977-11-5; 7337 Jackson et al 2012 7338 7339 Balasubramanian SK, Poh K-W, Ong C-N, Kreyling WG, Ong W-Y, Yu LE, (2013). The effect of primary 7340 particle size on biodistribution of inhaled gold nano-agglomerates. Biomaterials 34, 5439-5452. 7341 7342 Balog S, Rodriguez-Lorenzo L, Monnier CA, Obiols-Rabasa M, Rothen-Rutishauser B, Schurtenberger 7343 P, Petri-Fink A. Characterizing nanoparticles in complex biological media and physiological fluids with 7344 depolarized dynamic light scattering. Nanoscale 7(14):5991-7 (2015). 7345 7346 J. Banchereau, R.M. Steinman. Dendritic cells and the control of immunity. Nature 392 (1998) 245- 7347 252. 7348 7349 Bangale MS, Mitkare SS*, Battani SG, Sararkar DM. 2012.Recent nanotechnological aspects in 7350 cosmetics and dermatological preparations Int J Pharm Pharm Sci, 4 (2): 88-97 7351 7352 Barbasz A, Oćwieja M, Barbasz J. Cytotoxic Activity of Highly Purified Silver Nanoparticles Sol Against 7353 Cells of Human Immune System. Appl Biochem Biotechnol. 176, 817-834, 2015. 7354 7355 Barnes PJ, Nature Reviews Immunology 8, 183-192 7356 7357 Baroli B, Ennas MG, Loffredo F, Isola M, Pinna R, López-Quintela MA. 2007. Penetration of metallic 7358 nanoparticles in human full-thickness skin. J Invest Dermatol. 2007 Jul;127(7):1701-12 7359

239

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7360 Baron L, Gombault A, Fanny M, Villeret B, Savigny F, Guillou N, Panek C, Le Bert M, Lagente V, 7361 Rassendren F, Riteau N, Couillin I. The NLRP3 inflammasome is activated by nanoparticles through 7362 ATP, ADP and adenosine. Cell Death Dis. 2015 Feb 5;6:e1629 7363 7364 Paul A. Baron et al.(2008) 7365 7366 Barry BW: Novel mechanisms and devices to enable successful transdermal drug delivery. Eur J 7367 Pharm Sci 2001, 14: 101-114. 7368 7369 Bell, Iris R.; Ives, John A.; and Jonas, Wayne B. (2014) "NONLINEAR EFFECTS OF NANOPARTICLES: 7370 BIOLOGICAL VARIABILITY FROM HORMETIC DOSES, SMALL PARTICLE SIZES, AND DYNAMIC ADAPTIVE 7371 INTERACTIONS," Dose-Response: An International Journal: Vol. 12: Iss. 2, Article 4. 7372 7373 Bello D, Wardke BL, Yamamoto H, deVilloria RG, Garcia EJ, Hart AJ, Ahn K, Ellenbecher MJ, Hallock M. 7374 2009. Exposure to nanoscale particles and fibers during machining of hybrid advanced composites 7375 containing carbon nanotubes. J. Nanopart Res 11:231-249 7376 7377 Bellmann S, Carlander D, Fasano A, Momcilovic D, Scimeca JA, Waldman WJ, Gombau L, Tsytsikova L, 7378 Canady R, Pereira DI, Lefebvre DE. Mammalian gastrointestinal tract parameters modulating the 7379 integrity, surface properties, and absorption of food-relevant nanomaterials.Wiley Interdiscip Rev 7380 Nanomed Nanobiotechnol. 2015 Sep-Oct;7(5):609-22. doi: 10.1002/wnan.1333. Epub 2015 Jan 30. 7381 7382 Belkaid Y, Segre JA. Dialogue between skin microbiota and immunity. Science 2014; 346:954-9. 7383 7384 Bergin IL and Witzmann FA. Nanoparticle toxicity by the gastrointestinal route: evidence and 7385 knowledge gaps. Int J Biomed Nanosci Nanotechnol. 2013 ; 3(1-2): doi:10.1504/IJBNN.2013.054515. 7386 7387 Bhabra G, Sood A, Fisher B, Cartwright L, Saunders M, Evans WH, Surprenant A, Lopez-Castejon G, 7388 Mann S, Davis SA, Hails LA, Ingham E, Verkade P, Lane J, Heesom K, Newson R, Case CP. 7389 Nanoparticles can cause DNA damage across a cellular barrier. Nat Nanotechnol. 2009 Dec;4(12):876- 7390 83. 7391 7392 Bhattacharya K, Andón FT, El-Sayed R, Fadeel B. Mechanisms of carbon nanotube-induced toxicity: 7393 focus on pulmonary inflammation. Adv Drug Deliv Rev. 2013 Dec;65(15):2087-97.

240

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7394 Bianchi ME. DAMPs, PAMPs and alarmins: all we need to know about danger. J Leukoc Biol. 2007 7395 Jan;81(1):1-5. 7396 7397 Boudreau MD, Imam MS, Paredes AM, Bryant MS, Cunningham CK, Felton RP, Jones MY, Davis KJ, 7398 Olson GR. Differential Effects of Silver Nanoparticles and Silver Ions on Tissue Accumulation, 7399 Distribution, and Toxicity in the Sprague Dawley Rat Following Daily Oral Gavage Administration for 7400 13 Weeks. Toxicol Sci. 150, 131-160, 2016. 7401 7402 Edilberto Bermudez et al.(2004) 7403 7404 Bermudez LE, Sangari FJ, Kolonoski P, Petrofsky M, Goodman J: The efficiency of the translocation of 7405 Mycobacterium tubercu Shapiro-Shelef M, Calame K. Regulation of plasma-cell development. Nat. 7406 Rev. Immunol. 5(3), 230–42 (2005) 7407 7408 Bermudez LE, Sangari FJ, Kolonoski P, Petrofsky M, Goodman J: The efficiency of the translocation of 7409 Mycobacterium tuberculosis across a bilayer of epithelial and endothelial cells as a model of the 7410 alveolar wall is a consequence of transport within mononuclear phagocytes and invasion of alveolar 7411 epithelial cells. Infect Immun 2002, 70: 140-6. 7412 7413 Bertrand Gamer, A.O., E. Leibold, and B. van Ravenzwaay, The in vitro absorption of microfine zinc 7414 oxide and titanium dioxide through porcine skin. Toxicology in vitro : an international journal 7415 published in association with BIBRA, 2006. 20(3): p. 301-7. 7416 7417 Braakhuis HM, Cassee FR, Fokkens PH, de la Fonteyne LJ, Oomen AG, Krystek P, de Jong WH, van 7418 Loveren H, Park MV. Identification of the appropriate dose metric for pulmonary inflammation of 7419 silver nanoparticles in an inhalation toxicity study. Nanotoxicology. 10, 63-73, 2016. 7420 7421 Bezemer GF, Bauer SM, Oberdörster G, et al. Activation of pulmonary dendritic cells and Th2-type 7422 inflammatory responses on instillation of engineered, environmental diesel emission source or 7423 ambient air pollutant particles in vivo. J Innate Immun. 2011; 3(2):150-66. 7424 7425 Peter Bihari1 et al Particle and Fibre Toxicology20085:14 7426 7427 Birkness KA, Swisher BL, White EH, Long EG, Ewing EP, Quinn FD: A tissue culture bilayer model to 7428 study the passage of Neisseria meningitidis. Infect Immun 1995, 63: 402-9

241

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7429 7430 Blank F, Gerber P, Rothen-Rutishauser B, Salaklang J, Sakulkhu U, de Peyer K, Gehr P, Gazdhar A, 7431 Nicod LP, Hofmann H, Geiser T, Fink A, von Garnier C. Polymer-coated nanoparticles alter dendritic 7432 cell-dependent T cell stimulation. Nanotoxicology 5:606-21 (2011). 7433 7434 Blank F, Rothen-Rutishauser BM, Schurch S, Gehr P: An optimized in vitro model of the respiratory 7435 tract wall to study particle cell interactions. J Aerosol Med 2006, 19: 392-405 7436 7437 Blank F, Gerber P, Rothen-Rutishauser B, et al. Biomedical nanoparticles modulate specific CD4(+) T 7438 cell stimulation by inhibition of antigen processing in dendritic cells. Nanotoxicology. 5, 606–21 (2011 7439 7440 Blank F, Stumbles PA, Seydoux E, et al. Size-dependent Uptake of Particles by Pulmonary APC 7441 Populations and Trafficking to Regional Lymph Nodes. Am J Respir Cell Mol Biol. 49, 67–77 (2013). 7442 7443 Böckmann J, Lahl H, Eckert T, Unterhalt B. Titan-blutspiegel vor und nach belastungsversuchen mit 7444 titandioxide. Pharmazie. 2000;55(2):140-3. 7445 7446 Bonner JC. Nanoparticles as a potential cause of pleural and interstitial lung disease. Proc. Am. 7447 Thorac. Soc. 2010a; 7(2): 138-141. 7448 7449 Bonner JC. Respiratory Toxicology in Molecular and Biochemical Toxicology, 4th Edition. Eds: R.C. 7450 Smart and E. Hodgson. John Wiley & Sons, Inc. New York. 2008; 639-670. 7451 7452 Bonner JC. Carbon nanotubes as delivery systems for respiratory disease: do the dangers outweigh 7453 the potential benefits? Expert Rev Respir Med. 2011; 5(6):779-787. 7454 7455 Bonner JC. Mesenchymal cell survival in airway and interstitial pulmonary fibrosis. Fibrogenesis 7456 Tissue Repair. 2010b; 3:15. 7457 7458 Bonner JC. Regulation of PDGF and its receptors in fibrotic diseases. Cytokine Growth Factor Rev. 7459 2004; 15: 255-273. 7460 7461 Bonner JC, Silva RM, Taylor AJ, Brown JM, Hilderbrand SC, Castranova V, Porter D, Elder A, 7462 Oberdörster G, Harkema JR, Bramble LA, Kavanagh TJ, Botta D, Nel A, Pinkerton KE. Interlaboratory

242

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7463 evaluation of rodent pulmonary responses to engineered nanomaterials: the NIEHS Nano GO 7464 Consortium. Environ Health Perspect. 2013;121(6):676-82. 7465 7466 Bonner JC. Nanoparticles as a potential cause of pleural and interstitial lung disease. Proc. Am. 7467 Thorac. Soc. 2010a; 7(2): 138-141. 7468 7469 Boobis AR, Doe JE, Heinrich-Hirsch B, Meek ME, Munn S, Ruchirawat M, Schlatter J, Seed J, Vickers C 7470 (2008) IPCS framework for analyzing the relevance of a noncancer mode of action for humans. Crit 7471 Rev Tox 38: 87-96. 7472 7473 Boobis AR, Cohen SM, Dellarco V, McGregor D, Meek ME, Vickers C, Willcocks D, Farland W (2006) 7474 IPCS framework for analyzing the relevance of a cancer mode of action for humans. Crit Rev Tox 36: 7475 781-792. 7476 7477 Boraschi D, Aguado MT, Dutel C, Goronzy J, Louis J, Grubeck-Loebenstein B, Rappuoli R, Del Giudice 7478 G. The gracefully aging immune system. Sci Transl Med. 2013 May 15;5(185):185ps8. 7479 7480 Boraschi D, Costantino L, Italiani P. Interaction of nanoparticles with immunocompetent cells: 7481 nanosafety considerations. Nanomedicine (Lond). 2012 Jan;7(1):121-31. 7482 7483 Boraschi D, Fadeel B, Duschl A. Immune System. In: Adverse Effects of Engineered Nanomaterials: 7484 Exposure, Toxicology, and Impact on Human Health. Second Edition. Eds: B Fadeel, A Pietroiusti, A 7485 Shvedova. Academic Pres-Elsevier. 2017. pp. 313-337. 7486 7487 Bos JD, Meinardi MM. The 500 Dalton rule for the skin penetration of chemical compounds and 7488 drugs. Exp Dermatol. 2000 Jun;9(3):165-9. 7489 7490 Bottini M, Bruckner S, Nika K, Bottini N, Bellucci S, Magrini A, Bergamaschi A, Mustelin T. Multi- 7491 walled carbon nanotubes induce T lymphocyte apoptosis. Toxicol Lett. 2006 Jan 5;160(2):121-6. 7492 7493 Bouwmeester H, Lynch I, Marvin HJ, Dawson KA, Berges M, Braguer D, Byrne HJ, Casey A, Chambers 7494 G, Clift MJ, Elia G, Fernandes TF, Fjellsbø LB, Hatto P, Juillerat L, Klein C, Kreyling WG, Nickel C, 7495 Riediker M, Stone V. 2011. Minimal analytical characterization of engineered nanomaterials needed 7496 for hazard assessment in biological matrices. Nanotoxicology 5:1-11. 7497

243

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7498 Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A, Fusenig NE: Normal 7499 keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J Cell Biol 7500 1988, 106: 761-771. 7501 7502 Bourdon JA, Halappanavar S, Saber AT, Jacobsen NR, Williams A, Wallin H, Vogel U, Yauk CL. Toxicol 7503 Sci. 2012 Jun;127(2):474-84. 7504 7505 Bourdon et al 2013 7506 7507 Boverhof DR, Ladics G, Luebke B, Botham J, Corsini E, Evans E, Germolec D, Holsapple M, Loveless SE, 7508 Lu H, van der Laan JW, White KL Jr, Yang Y (2014) Approaches and considerations for the assessment 7509 of immunotoxicity for environmental chemicals: A workshop summary. Reg Tox Pharm 68: 96-107 7510 7511 Braakhuis HM, Kloet Sk, Kezic S, Kuper F, Park MV, Bellmann S, Van Der Zande M, Le Gac S, Krystek P, 7512 Peters RJ, Rietjens IM, Bouwmeester H. (2015) Progress and future of in vitro models to study 7513 translocation of nanoparticles.Arch Toxicol. 89, 1469. 7514 7515 Braakhuis, H. M., Park, M. V., Gosens, I., De Jong, W. H., and Cassee, F. R. (2014). Physicochemical 7516 characteristics of nanomaterials that affect pulmonary inflammation. Part Fibre Toxicol 11(1), 18. 7517 7518 Brandenberger C, Rowley NL, Jackson-Humbles DN, Zhang Q, Bramble LA, Lewandowski RP, Wagner 7519 JG, Chen W, Kaplan BL, Kaminski NE, Baker GL, Worden RM, Harkema JR. Engineered silica 7520 nanoparticles act as adjuvants to enhance allergic airway disease in mice. Part Fibre Toxicol. 7521 2013;10:26. 7522 7523 Brandon EF, Oomen AG, Rompelberg CJ, Versantvoort CH, van Engelen JG, Sips AJ. (2006). Consumer 7524 product in vitro digestion model: bioaccessibility of contaminants and its application in risk 7525 assessment. Regul Toxicol Pharmacol 44:161–71. 7526 7527 Brown DM, kinloch I, Bangert U, Windle AH, Walter DM, Walker GS, Scotchford CA, Donaldson K, 7528 Stone V. An in vitro study of the potential of carbon nanotubes and nanofibres to induce 7529 inflammatory mediators and frustrated phagocytosis. Carbon 2007, 45; 1743-1756. 7530 7531 Brown JM, Wilson TM, Metcalfe DD. The mast cell and allergic diseases: role in pathogenesis and 7532 implications for therapy. Clin Exp Allergy 2008; 38(1):4-18.

244

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7533 7534 Brown DM, Donaldson K, Stone V. Nuclear translocation of Nrf2 and expression of antioxidant 7535 defence genes in THP-1 cells exposed to carbon nanotubes. J Biomed Nanotechnol. 2010; 6(3):224- 7536 33. 7537 7538 Brown DM, Kanase N, Gaiser B, Johnston H, Stone V. Inflammation and gene expression in the rat 7539 lung after instillation of silica nanoparticles: effect of size, dispersion medium and particle surface 7540 charge. Toxicol Lett. 2013; S0378-4274(13)01362-3. 7541 7542 Brown et al. Carbon 2007 7543 7544 Brouwer D. 2010. Exposure to manufactured nanoparticles in different workplaces. Toxicology. 269, 7545 120-7. 7546 7547 Broz P, Dixit VM. Inflammasomes: mechanism of assembly, regulation and signalling. Nat Rev 7548 Immunol. 2016 Jul;16(7):407-20. 7549 7550 Mary C Buford et al., Part Fibre Toxicol. 2007; 4: 6. 7551 7552 Butz T, Reinert T, Pinheiro T, Moretto P, Pallon J, Kiss AZ, Stachura J, D¸abro´s W, Stachura Z, Lekki J, 7553 Lekka M, Hunyadi J, Bıro T, Sticherling M, Van Vaeck L, Van Royen P, Surleve-Bazeille JE., (2007). 7554 NANODERM, Quality of Skin as a Barrier to ultra-fine Particles, QLK4-CT-2002-02678 Final Report. 7555 http://www.uni-leipzig.de/~nanoderm/Downloads/Nanoderm_Final_Report.pdf 7556 7557 Carbone C, Manno D, Serra A, Musumeci T, Pepe V,Tisserand C, Puglisi G. Innovative hybrid vs 7558 polymeric nanocapsules: The influence of the cationic lipid coating on the "4S".Colloids Surf B 7559 Biointerfaces. 2016 Feb 6;141:450-457. 7560 7561 Card JW, Zeldin DC, Bonner JC, et al. Pulmonary applications and toxicity of engineered 7562 nanoparticles. Am J Physiol. 2008; 295(3): L400-L411. 7563 7564 Carlson C, Hussain SM, Schrand AM, Braydich-Stolle LK. Hess KL, Jones RL, Schlager JJ.. Unique 7565 cellular interaction of silver nanoparticles: Size dependent generation of reactive oxygen species. J 7566 Phys Chem B 112:13608–13619, 2008. 7567

245

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7568 Casals E, Pfaller T, Duschl A, Oostingh GJ, Puntes VF. Hardening of the nanoparticle-protein corona in 7569 metal (Au, Ag) and oxide (Fe3O4, CoO, and CeO2) nanoparticles. Small. 2011;7(24):3479-86. 7570 7571 Cassee FR, Muijser H, Duistermaat E, Freijer JJ, Geerse KB, Marijnissen JCM, Arts JHE., (2002). Particle 7572 size-dependent total mass deposition in lungs determines inhalation toxicity of cadmium chloride 7573 aerosols in rats. Application of a multiple path dosimetry model. Arch Toxicol 76, 277-286. 7574 7575 Catsakis LH, Sulica VI. Allergy to silver amalgams. Oral Surg Oral Med Oral Pathol 46, 371–375, 1978.. 7576 Chen X, Schluesener HJ. Nanosilver: a nanoproduct in medical application. Toxicol Lett. 176, 1-12, 7577 2008. 7578 7579 Cedervall T, Lynch I, Lindman S, Berggard T, Thulin E, Nilsson H, Dawson KA, Linse S. (2007) 7580 Understanding the nanoparticle-protein corona using methods to quantify exchange rates and 7581 affinities of proteins for nanoparticles. Proc Natl Acad Sci U S A 104, 2050–2055. 7582 7583 Cesta MF, Ryman-Rasmussen JP, Wallace DG, Masinde T, Hurlburt G, Taylor AJ, Bonner JC. Bacterial 7584 lipopolysaccharide enhances PDGF signaling and pulmonary fibrosis in rats exposed to carbon 7585 nanotubes. Am J Respir Cell Mol Biol. 2010, 43(2):142-51. 7586 7587 Chamorro S, Gutiérrez L, Vaquero MP, Verdoy D, Salas G, Luengo Y, Brenes A, José Teran F. Safety 7588 assessment of chronic oral exposure to iron oxide nanoparticles. Nanotechnology. 7589 2015;26(20):205101. 7590 7591 Chan C, Smyth M, Martinet L. Molecular mechanisms of natural killer cell activation in response to 7592 cellular stress. Cell Death Differ. 21(1), 5–14 (2014) 7593 7594 Chang Y-J, DeKruyff RH, Umetsu DT. The role of type 2 innate lymphoid cells in asthma. J. Leukoc. 7595 Biol. 94(5), 933–40 (2013) 7596 7597 Chauhan AS. Dendrimer nanotechnology for enhanced formulation and controlled delivery of 7598 resveratrol. Ann N Y Acad Sci. 2015, in press. 7599 7600 Chavez-Santoscoy AV, Roychoudhury R, Pohl NL, Wannemuehler MJ, Narasimhan B, Ramer-Tait AE. 7601 Tailoring the immune response by targeting C-type lectin receptors on alveolar macrophages using 7602 "pathogen-like" amphiphilic polyanhydride nanoparticles. Biomaterials. 2012;33:4762-4772.

246

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7603 Chen F, Wang G, Griffin JI, Brenneman B, Banda NK, Holers VM, Backos DS, Wu L, Moghimi SM, 7604 Simberg D. Complement proteins bind to nanoparticle protein corona and undergo dynamic 7605 exchange in vivo. Nat Nanotechnol. 2017 May;12(4):387-393. 7606 7607 B.X. Chen, S.R. Wilson, M. Das, D.J. Coughlin, B.F. Erlanger. Antigenicity of fullerenes: antibodies 7608 specific for fullerenes and their characteristics. Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 10809-10813. 7609 7610 Chen H, Wang B, Zhao R, Gao D, Guan M, Zheng L, Zhou X, Chai Z, Zhao Y, Feng W. Coculture with 7611 Low-Dose SWCNT Attenuates Bacterial Invasion and Inflammation in Human Enterocyte-like Caco-2 7612 Cells. Small. 2015, in press. 7613 7614 Chen HT, Neerman MF, Parrish AR, Simanek EE. Cytotoxicity, hemolysis, and acute in vivo toxicity of 7615 dendrimers based on melamine, candidate vehicles for drug delivery. J Am Chem Soc. 7616 2004;126(32):10044-8. 7617 7618 Chen EY, Garnica M, Wang YC, Mintz AJ, Chen CS, Chin WC. A mixture of anatase and rutile TiO2 7619 nanoparticles induces histamine secretion in mast cells. Part Fibre Toxicol. 2012; 19: 9:2. 7620 7621 Chen JM, et al., Particle and Fibre Toxicology201411:20 7622 7623 Chen JM, et. al., Nanotoxicology. 2013 Jun; 7(4): 417–431. 7624 7625 Chen HQ, Gao D, Wang B, Zhao RF, Guan M, Zheng LN, Zhou XY, Chai ZF, Feng WY. Graphene oxide as 7626 an anaerobic membrane scaffold for the enhancement of B. adolescentis proliferation and 7627 antagonistic effects against pathogens E. coli and S. aureus. Nanotechnology 2014; 25:165101 7628 7629 Chithrani BD, Chan WCW. Elucidating the mechanism of cellular uptake and removal of protein- 7630 coated gold nanoparticles of different sizes and shapes. Nano Lett 2007, 7:1542-1550. 7631 7632 Chithrani BD, Ghazan AA, Chan CW. Determining the size and the shape dependence of gold 7633 nanoparticle uptake into mammalian cells. Nano Lett 2006, 6:662-668. 7634 7635 Cho EA, Lee WS, Kim KM, Kim SY. Occupational generalized argyria after exposure to aerosolized 7636 silver. J Dermatol 35, 759–60, 2008. 7637

247

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7638 Cho WS, Duffin R, Thielbeer F, Bradley M, Megson IL, Macnee W, Poland CA, Tran CL, Donaldson K. 7639 Zeta potential and solubility to toxic ions as mechanisms of lung inflammation caused by metal/metal 7640 oxide nanoparticles. Toxicol Sci 2012, 126:469-477. 7641 7642 Choi HS, Ashitate Y, Lee JH, Kim SH, Matsui A, Insin N, Bawendi MG, Semmler-Behnke M, Frangioni 7643 JV, Tsuda A. Rapid translocation of nanoparticles from the lung airspaces to the body. Nat Biotechnol. 7644 2010 Dec;28(12):1300-3. 7645 7646 Choi J, Reipa V, Hitchins VM, Goering PL, Malinauskas RA. Physicochemical characterization and in 7647 vitro hemolysis evaluation of silver nanoparticles. Toxicol Sci. 113, 133-143, 2011. 7648 7649 Chortarea S, Clift MJD, Vanhecke D, Endes C, Wick P, Petri-Fink A et al.: Repeated exposure to carbon 7650 nanotube-based aerosols does not affect the functional properties of a 3D human epithelial airway 7651 model. Nanotoxicology 2015, 9: 983-993 7652 7653 Christensen FM, Johnston HJ, Stone V, Aitken RJ, Hankin S, Peters S, Aschberger K. (2010). Nano- 7654 silver – feasibility and challenges for human health risk assessment based on open literature. 7655 Nanotoxicology 4:284–295. 7656 7657 Christensen FM, Johnston HJ, Stone V, Aitken RJ, Hankin S, Peters S, Aschberger K. (2011). Nano- 7658 TiO2–feasibility and challenges for human health risk assessment based on open literature. 7659 Nanotoxicology 5:110–124. 7660 7661 Chuang HC, Hsiao TC, Wu CK, Chang HH, Lee CH, Chang CC, Cheng TJ; Taiwan CardioPulmonary 7662 Research Group (T-CPR) (2013) Allergenicity and toxicology of inhaled silver nanoparticles in allergen- 7663 provocation mice models. Int J Nanomedicine. 8, 4495-4506. doi: 10.2147/IJN.S52239. 7664 7665 Cichocki F, Sitnicka E, Bryceson YT. NK cell development and function - Plasticity and redundancy 7666 unleashed. Semin. Immunol. 26(2), 114–126 (2014) 7667 7668 Cinteza LO. Quantum dots in biomedical applications: advances and challenges. J Nanophotonics, 7669 2010, 1:4. 7670 7671 Clark, R.A., Skin-resident T cells: the ups and downs of on site immunity. J Invest Dermatol, 2010. 7672 130(2): p. 362-70.

248

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7673 7674 M.J.D. Clift, J.-F. Dechézelles, B. Rothen-Rutishauser, A. Petri-Fink, Front Chem 3 (2015) 7675 7676 Clift et al. 2011 7677 7678 Cockburn A, Bradford R, Buck N, Constable A, Edwards G, Haber B, Hepburn P, Howlett J, Kampers F, 7679 Klein C, Radomski M, Stamm H, Wijnhoven S, Wildemann T. (2012). Approaches to the safety 7680 assessment of engineered nanomaterials (ENM) in food. Food Chem Toxicol 50:2224–2242. 7681 7682 Coggins, C. R. E., P. H. Ayres, A. T. Mosberg, J. W. Sagart, and A. W. Hayes. 1993. Comparative 7683 inhalation study in rats using cigarettes containing tobacco expanded with chlorofluorocarbon-11 7684 (CFC-11) or hydrochlorofluorocarbon-123 (HCFC-123). Inhalation Toxicology 5:97–115. 7685 doi:10.3109/08958379309034496. 7686 7687 Coggins, C. R. E., J. S. Edmiston, A. M., Jerome, T. B. Langston, E. J. Senaa, D. C. Smith, and M. J. 7688 Oldham. 2011. A comprehensive evaluation of the toxicology of cigarette ingredients: Essential oils 7689 and resins. Inhalation Toxicology 23(S1): 41–69. doi:10.3109/08958378.2010.543188 7690 7691 Costa PM, Fadeel B. Emerging systems biology approaches in nanotoxicology: Towards a mechanism- 7692 based understanding of nanomaterial hazard and risk. Toxicol Appl Pharmacol. 2016 May 7693 15;299:101-11 7694 7695 Costello, M. E., Robinson, P. C., Benham, H., and Brown, M. A. (2015). The intestinal microbiome in 7696 human disease and how it relates to arthritis and spondyloarthritis. Best practice & research. Clinical 7697 rheumatology 29(2), 202-12. 7698 7699 Council of Canadian Academies, Expert Panel on Nanotechnology Staff. Small is Different: A Science 7700 Perspective on the Regulatory Challenges of the nanoscale. 7701 7702 Coyuco JC, Liu Y, Tan BJ, Chiu GN. Functionalized carbon nanomaterials: exploring the interactions 7703 with Caco-2 cells for potential oral drug delivery. Int J Nanomedicine. 2011;6:2253-63. 7704 7705 Cozens AL, Yezzi MJ, Kunzelmann K, Ohrui T, Chin L, Eng K et al.: CFTR expression and chloride 7706 secretion in polarized immortal human bronchial epithelial cells. Am J Respir Cell Mol Biol 1994, 10: 7707 38-47

249

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7708 7709 Crane M., Handy R.D. (2007). An assessment of regulatory testing strategies and methods for 7710 characterizing the ecotoxicological hazards of nanomaterials. Final report of Defra project CSA 7284, 7711 Watts and Crane Associates, Faringdon, Oxfordshire, UK. Available at: 7712 randd.defra.gov.uk/Document.aspx?Document=CB01097_6262_FRP.pdf. 7713 7714 Czarny et. al., ACS Nano, 2014, 8 (6), pp 5715–5724 7715 7716 Dai H. Carbon nanotubes: synthesis, integration, and properties. Acc Chem Res. 2002;35(12):1035-44. 7717 7718 R. Damoiseaux, S. George, M. Li, S. Pokhrel, Z. Ji, B. France, T. Xia, E. Suarez, R. Rallo, L. M€adler, Y. 7719 Cohen, E. M. V. Hoek and A. Nel Nanoscale, 2011, 3, 1345 7720 7721 Dandley EC, Taylor AJ, Duke KS, Ihrie MD, Shipkowski KA, Parsons GN, Bonner JC. Atomic layer 7722 deposition coating of carbon nanotubes with zinc oxide causes acute phase immune responses in 7723 human monocytes in vitro and in mice after pulmonary exposure. Part Fibre Toxicol. 2016, 13(1):29. 7724 7725 Danish report 2015, Comparison of three different brushing techniques to isolate and culture primary 7726 nasal epithelial cells from human subjects. 7727 7728 DanishMinistry of the Environment (EPA). Dermal Absorption of Nanomaterials. Part of the ”Better 7729 control of nano” initiative 2012-2015 Environmental Project No. 1504, 2013 7730 7731 Darabi Sahneh F, Scoglio C, Riviere J. Dynamics of nanoparticle-protein corona complex formation: 7732 analytical results from population balance equations. PLoS One. 2013;8(5):e64690. 7733 7734 Darlington TK, Neigh AM, Spencer MT, Nguyen OT, Oldenburg SJ. Environ. Toxicol. Chem 2009, 28, 7735 1191-1199. 7736 7737 Da Silva SB, Ferreira D, Pintado M, Sarmento B. Chitosan-based nanoparticles for rosmarinic acid 7738 ocular delivery-In vitro tests. Int J Biol Macromol. 2016;84:112-20. 7739 7740 Davis BK, Wen H, Ting JP. The inflammasome NLRs in immunity, inflammation, and associated 7741 diseases. Annu Rev Immunol. 2011;29: 707–735 7742

250

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7743 Decan N et al., Mutat Res Genet Toxicol Environ Mutagen. 2016 Jan 15;796:8-22 7744 7745 De Haar C, Hassing I, Bol M, Bleumink R, Pieters R. Ultrafine but not fine particulate matter causes 7746 airway inflammation and allergic airway sensitization to co-administered antigen in mice. Clin Exp 7747 Allergy. 2006;11:1469-1479. 7748 7749 De Jong WH, Hagens, WI, Krystek P, Burger MC, Sips AJAM, Geertsma RE. (2008). Particle size 7750 dependent organ distribution of gold nanoparticles after intervenous administration. Biomaterials 7751 29, 1912-1919. 7752 7753 De Jong WH, Van Der Ven LT, Sleijffers A, Park MV, Jansen EH, Van Loveren H, Vandebriel RJ. 7754 Systemic and immunotoxicity of silver nanoparticles in an intravenous 28 days repeated dose toxicity 7755 study in rats. Biomaterials 34, 8333-8343, 2013. 7756 7757 Delmaar, C. J., Peijnenburg, W. J., Oomen, A. G., Chen, J., de Jong, W. H., Sips, A. J., Wang, Z., and 7758 Park, M. V. (2015). A practical approach to determine dose metrics for nanomaterials. Environ Toxicol 7759 Chem 34(5), 1015-22. 7760 7761 DeLoid G, et. al., 2013 Nature Communications 5, Article number: 3514 7762 7763 De Mel A, Chaloupka K, Malam Y, Darbyshire A, Cousins B, Seifalian AM. A silver nanocomposite 7764 biomaterial for blood-contacting implants. J Biomed Mater Res A. 100A, 2348-2357, 2012. 7765 7766 Z.J. Deng, M. Liang, M. Monteiro, I. Toth, R.F. Minchin. Nanoparticle-induced unfolding of fibrinogen 7767 promotes Mac-1 receptor activation and inflammation. Nat. Nanotechnol. 6 (2011) 39-44. 7768 7769 Des Rieux A, Fievez V, Garinot M, Schneider Y-J, Preat V. (2006) Nanoparticles as potential oral 7770 delivery systems of proteins and vaccines: a mechanistic approach.J Controlled Release, 116, 1-27. 7771 7772 Dinarello CA, Simon A, van der Meer JW. Treating inflammation by blocking interleukin-1 in a broad 7773 spectrum of diseases. Nat Rev Drug Discov. 2012 Aug;11(8):633-52. 7774 Dinicola S, Masiello MG, Proietti S, Coluccia P, Fabrizi G, Palombo A, Micciulla F, Bistarelli S, Ricci G, 7775 Catizone A, De Toma G, Bizzarri M, Bellucci S, Cucina A. Multiwalled carbon nanotube buckypaper 7776 induces cell cycle arrest and apoptosis in human leukemia cell lines through modulation of AKT and 7777 MAPK signaling pathways. Toxicol In Vitro. 2015 Oct;29(7):1298-308.

251

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7778 7779 M.A. Dobrovolskaia, B.W. Neun, S. Man, X. Ye, M. Hansen, A.K. Patri, R.M. Crist, S.E. McNeil. Protein 7780 corona composition does not accurately predict hematocompatibility of colloidal gold nanoparticles. 7781 Nanomedicine 10 (2014) 1453-1463. 7782 7783 Dobrovolskaia MA, Shurin M, Shvedova AA. Current understanding of interactions between 7784 nanoparticles and the immune system. Toxicol Appl Pharmacol. 2016 May 15;299:78-89. 7785 7786 M.A. Dobrovolskaia, S.E. McNeil. Immunological properties of engineered nanomaterials. Nat. 7787 Nanotechnol. 2 (2007) 469-478. 7788 7789 7790 7791 Marina A. Dobrovolskaia, Scott E. McNeil. Handbook of Immunological Properties of Engineered 7792 Nanomaterials 7793 7794 Marina a. dobrovolskaia, dori r. germolec and James l. Weaver. Evaluation of nanoparticle 7795 immunotoxicity. Nature nanotechnology | VOL 4 | JULY 2009 7796 7797 Dobrovolskaia MA, Neun BW, Clogston JD, Ding H, Ljubimova J, McNeil SE: Ambiguities in applying 7798 traditional Limulus amebocyte lysate tests to quantify endotoxin in nanoparticle formulations. 7799 Nanomedicine (Lond) 2010, 5: 555–562. 7800 7801 Donaldson K, Schinwald A, Murphy F, Cho WS, Duffin R, Tran L, Poland C. 2012. The Biologically 7802 Effective Dose in Inhalation Nanotoxicology. Acc Chem Res 2012 Sep 24. [Epub ahead of print] 7803 7804 Donaldson, K., et al., Pulmonary toxicity of carbon nanotubes and asbestos - Similarities and 7805 differences. Adv Drug Deliv Rev, 2013. et al. 2013. Nanosafety in Europe 2015 -2025: Towards Safe 7806 and Sustainable Nanomaterials and Nanotechnology Innovations. 7807 7808 Donaldson K, Murphy FA, Duffin R, et al. Asbestos, carbon nanotubes and the pleural mesothelium: a 7809 review of the hypothesis regarding the role of long fibre retention in the parietal pleura, 7810 inflammation and mesothelioma. Part Fibre Toxicol. 2010; 7:5. 7811

252

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7812 Dorier M, Brun E, Veronesi G, Barreau F, Pernet-Gallay K, Desvergne C, Rabilloud T, Carapito C, 7813 Herlin-Boime N, Carrière M. Impact of anatase and rutile titanium dioxide nanoparticles on uptake 7814 carriers and efflux pumps in Caco-2 gut epithelial cells. Nanoscale. 2015;7(16):7352-60. 7815 7816 C. Dostert, V. Pétrilli, R. Van Bruggen, C. Steele, B.T. Mossman, J. Tschopp, Innate immune activation 7817 through Nalp3 inflammasome sensing of asbestos and silica, Science, 320 (2008) 674-677. 7818 7819 Drake, P. L.; Hazelwood, K. J. Exposure-Related Health Effects of Silver and Silver Compounds: A 7820 Review. Ann. Occup. Hyg. 49, 575–588, 2005 7821 7822 Driscoll KE, Costa DL, Hatch G, Henderson R, Oberdorster G, Salem H, Schlesinger RB. 2000. 7823 Intratracheal instillation as an exposure technique for the evaluation of respiratory tract toxicity: 7824 uses and limitations. Toxicol Sci. 55(1):24-35. 7825 7826 Driscoll et al., Toxicological sceinces 55, 24-35 (2000); Osier M, Fundamentals and Applied Toxicology 7827 40, 220-227 (1997) 7828 7829 D. Dutta, S.K. Sundaram, J.G. Teeguarden, B.J. Riley, L.S. Fifield, J.M. Jacobs, S.R. Addleman, G.A. 7830 Kaysen, B.M. Moudgil, T.J. Weber. Adsorbed proteins influence the biological activity and molecular 7831 targeting of nanomaterials. Toxicol. Sci. 100 (2007) 303-315. 7832 7833 Dwivedi, P.D., et al., Are nanomaterials a threat to the immune system? Nanotoxicology, 2009. 3(1): 7834 p. 19-26 7835 7836 Dy R, Sethi S. The lung microbiome and exacerbations of COPD. Curr Opin Pulm Med 2016 7837 7838 EFSA [European Food and Safety Authority] (2004) Opinion of the scientific panel on food additives, 7839 flavourings, processing aids and materials in contact with food on a request from the commission 7840 related to the safety in use of rutile titanium dioxide as an alternative to the presently permitted 7841 anatase form, Question No EFSA-Q-2004-103, Adopted on 7 December 2004. EFSA J 163:1-12. 7842 7843 Ehrhardt C, Kneuer C, Fiegel J, Hanes J, Schaefer UF, Kim KJ et al.: Influence of apical fluid volume on 7844 the development of functional intercellular junctions in the human epithelial cell line 16HBE14o-: 7845 implications for the use of this cell line as an in vitro model for bronchial drug absorption studies. Cell 7846 Tissue Res 2002, 308: 391-400

253

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7847 7848 Ehrhardt C, Fiegel J, Fuchs S, Abu-Dahab R, Schaefer UF, Hanes J et al.: Drug absorption by the 7849 respiratory mucosa: cell culture models and particulate drug carriers. J Aerosol Med 2002, 15: 131- 7850 139. 7851 7852 S.C. Eisenbarth, O.R. Colegio, W. O'Connor, F.S. Sutterwala, R.A. Flavell. Crucial role for the Nalp3 7853 inflammasome in the immunostimulatory properties of aluminium adjuvants. Nature 453 (2008) 7854 1122-1126 7855 7856 Elder A, Gelein R, Silva V, Feikert T, Opanashuk L, Carter J, Potter R, Maynard A, Ito Y, Finkelstein J, 7857 Oberdörster G. 2006. Translocation of Inhaled Ultrafine Manganese Oxide Particles to the Central 7858 Nervous System. Environ Health Perspect 114:1172–1178. 7859 7860 Electro spray Generator (Figure 11) Ref ?? 7861 7862 Dan Elgrabli, Steve Abella-Gallart, Olivier Aguerre-Chariol, Franck Robidel, Francoise Rogerieux, Jorge 7863 Boczkowski & Ghislaine Lacroix (2007) Effect of BSA on carbon nanotube dispersion for in vivo and in 7864 vitro studies, Nanotoxicology, 1:4, 266-278, 7865 7866 Elgrabli, D., Beaudouin, R., Jbilou, N., Floriani, M., Pery, A., Rogerieux, F., and Lacroix, G. (2015). 7867 Biodistribution and Clearance of TiO2 Nanoparticles in Rats after Intravenous Injection. PloS one 7868 10(4), e0124490. 7869 7870 Elijah J. Petersen et al. Environ. Sci. Technol. 2014, 48, 4226−4246 7871 7872 Mahmoud Elsabahy, and Karen L. Wooley. Cytokines as biomarkers of nanoparticle immunotoxicity. 7873 Chem Soc Rev. 2013 June 21; 42(12): 5552–5576 7874 7875 El-Sayed M, Ginski M, Rhodes C, Ghandehari H. Transepithelial transport of poly(amidoamine) 7876 dendrimers across Caco-2 cell monolayers. J Control Release. 2002;81(3):355-65. 7877 7878 Ellinger-Ziegelbauer et al.(2009) 7879 7880 ENV/JM/MONO(2012)14 OECD 7881

254

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7882 ENV/JM/MONO(2012)40 OECD 7883 7884 Eom Y, Song JS, Lee DY, Kim MK, Kang BR, Heo JH, Lee HK, Kim HM. Effect of Titanium Dioxide 7885 Nanoparticle Exposure on the Ocular Surface: An Animal Study. Ocul Surf. 2016, in press. 7886 7887 Erdely A, Dahm M, Chen BT, Zeidler-Erdely PC, Fernback JE, Birch ME, Evans DE, Kashon ML, Deddens 7888 JA, Hulderman T, Bilgesu SA, Battelli L, Schwegler-Berry D, Leonard HD, McKinney W, Frazer DG, 7889 Antonini JM, Porter DW, Castranova V, Schubauer-Berigan MK. Carbon nanotube dosimetry: from 7890 workplace exposure assessment to inhalation toxicology. Part Fibre Toxicol. 2013 Oct 21;10(1):53. 7891 7892 Erdely A, Hulderman T, Salmen R, et al. Cross-talk between lung and systemic circulation during 7893 carbon nanotube respiratory exposure. Potential biomarkers. Nano Lett. 2009; 9(1), pp. 36-43. 7894 7895 European Commission COM(2012) 572 final. Second regulatory review on nanomaterials. Available at 7896 http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2012:0572:FIN:EN:PDF 7897 7898 EU. Regulation on Cosmetic Products 1223/2009. 2009 7899 7900 EU REGULATION (EU) No 1169/2011 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL. 2011 7901 7902 EU, 2011a. Commission Recommendation of 18 October 2011 on the definition of nanomaterial 7903 (2011/696/EU). O. J. L 275: 38-40. 7904 7905 Eydner et al.(2012) 7906 7907 Fabian, E., Landsiedel, R., Ma-Hock, L., Wiench, K., Wohlleben, W., Van Ravenzwaay, B. (2008). Tissue 7908 distribution and to 7909 7910 Fadeel B, Feliu N, Vogt C, Abdelmonem AM, Parak WJ. Bridge over troubled waters: understanding 7911 the synthetic and biological identities of engineered nanomaterials. Wiley Interdiscip Rev Nanomed 7912 Nanobiotechnol. 2013 Mar-Apr;5(2):111-29. 7913 7914 Fadeel B, Pietroiusti A, Shvedova AA. Adverse effects of engineered nanomaterials. Exposure, 7915 toxicology, and impact on human health. Academic Press, Elsevier, London, United Kingdom, 2012. 7916 7917 Fadeel B, Savolainen K. 2013. Broaden the discussion. Nature Nanotech 8:71.

255

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7918 7919 R. Falkner, N. Jaspers. Regulating nanotechnologies: risk, uncertainty and the global governance gap. 7920 Glob. Environ. Polit., 12 (1) (2012), pp. 30–55 7921 7922 S. Fallarini, T. Paoletti, C.O. Battaglini, P. Ronchi, L. Lay, R. Bonomi, S. Jha, F. Mancin, P. Scrimin, G. 7923 Lombardi. Factors affecting T cell responses induced by fully synthetic glyco-gold-nanoparticles. 7924 Nanoscale 5 (2013) 390-400. 7925 7926 Farrera & Fadeel 2015 Toxicity of intravenously administered titanium dioxide nanoparticles in rats. 7927 Arch. Toxicol. 82,151–157. 7928 7929 Farrera C, Fadeel B. It takes two to tango: Understanding the interactions between engineered 7930 nanomaterials and the immune system. Eur J Pharm Biopharm. 2015 Sep;95(Pt A):3-12. 7931 7932 Fatkhutdinova LM, Khaliullin TO, Vasil'yeva OL, Zalyalov RR, Mustafin IG, Kisin ER, Birch ME, Yanamala 7933 N, Shvedova AA. Fibrosis biomarkers in workers exposed to MWCNTs. Toxicol Appl Pharmacol. 2016 7934 7935 Faust JJ, Doudrick K, Yang Y, Westerhoff P, Capco DG. Food grade titanium dioxide disrupts intestinal 7936 brush border microvilli in vitro independent of sedimentation. Cell Biol Toxicol. 2014;30(3):169-88. 7937 7938 Fedeli C, Segat D, Tavano R, Bubacco L, De Franceschi G, de Laureto PP, Lubian E, Selvestrel F, Mancin 7939 F, Papini E. The functional dissection of the plasma corona of SiO₂-NPs spots histidine rich 7940 glycoprotein as a major player able to hamper nanoparticle capture by macrophages. Nanoscale. 7941 2015 Nov 14;7(42):17710-28 7942 7943 Feliu N, Fadeel B. Nanotoxicology: no small matter. Nanoscale. 2010 Dec;2(12):2514-20. 7944 7945 Fenoglio I, Aldieri E, Gazzano E, et al. Thickness of multiwalled carbon nanotubes affects their lung 7946 toxicity. Chem Res Toxicol. 2012; 25(1):74-82. 7947 7948 E.J. Fernandez, E. Lolis, Structure, function, and inhibition of chemokines, Annu Rev Pharmacol 7949 Toxicol, 42 (2002) 469-499. 7950 7951 M Filipe, P., et al., Stratum corneum is an effective barrier to TiO2 and ZnO nanoparticle 7952 percutaneous absorption. Skin Pharmacol Physiol, 2009. 22(5): p. 266-75.ay 15;299:125-31.

256

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7953 7954 Filon FL, Mauro M, Adam G, Bovenzi M, Crosera M. (2015) Nanoparticles skin absorption: New 7955 aspects for a safety profile evaluation. Regul Toxicol Pharmacol 72, 310-322. 7956 7957 Folkmann JK, Risom L, Jacobsen NR, Wallin H, Loft S, Møller P. Oxidatively damaged DNA in rats 7958 exposed by oral gavage to C60 fullerenes and single-walled carbon nanotubes. Environ Health 7959 Perspect. 2009;117(5):703-8. 7960 7961 Forrest IA, Murphy DM, Ward C, Jones D, Johnson GE, Archer L et al.: Primary airway epithelial cell 7962 culture from lung transplant recipients. Eur Respir J 2005, 26: 1080-1085. 7963 7964 Foster KA, Oster CG, Mayer MM, Avery ML, Audus KL: Characterization of the A549 cell line as a type 7965 II pulmonary epithelial cell model for drug metabolism. Exp Cell Res 1998, 243: 359-66. 7966 7967 Foster, C.A., et al., Human epidermal T cells predominantly belong to the lineage expressing 7968 alpha/beta T cell receptor. J Exp Med, 1990. 171(4): p. 997-1013. 7969 7970 Frick SU, Bacher N, Baier G, Mailänder V, Landfester K, Steinbrink K. Functionalized polystyrene 7971 nanoparticles trigger human dendritic cell maturation resulting in enhanced CD4+ T cell activation. 7972 Macromol Biosci. 12(12), 1637–47 (2012). 7973 7974 Froggett SH, Clancy SF, Boverhof DR, Canady RA, 2014. A review and perspective of existing research 7975 on the release of n 7976 7977 Fröhlich E, Salar-Behzadi S. Toxicological assessment of inhaled nanoparticles: role of in vivo, ex vivo, 7978 in vitro, and in silico studies. Int J Mol Sci. 2014 Mar 18;15(3):4795-822. doi: 10.3390/ijms15034795. 7979 Review. 7980 7981 Frohlich., Int J Nanomedicine. 2015; 10: 3761–3778. 7982 7983 Fröhlich E, Roblegg E Models for oral uptake of nanoparticles in consumer products. Toxicology. 2012 7984 Jan 27;291(1-3):10-7. doi: 10.1016/j.tox.2011.11.004. Epub 2011 Nov 18. Review. 7985 7986 P.C. Fulkerson, M.E. Rothenberg, Targeting eosinophils in allergy, inflammation and beyond, Nat Rev 7987 Drug Discov, 12 (2013) 117-129

257

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

7988 7989 Gaillet S, Rouanet JM. Silver nanoparticles: their potential toxic effects after oral exposure and 7990 underlying mechanisms--a review. Food Chem Toxicol. 2015;77:58-63. 7991 7992 S.J. Galli, N. Borregaard, T.A. Wynn, Phenotypic and functional plasticity of cells of innate immunity: 7993 macrophages, mast cells and neutrophils, Nat Immunol, 12 (2011) 1035-1044 7994 7995 Gallo, R.L. and L.V. Hooper, Epithelial antimicrobial defence of the skin and intestine. Nat Rev 7996 Immunol, 2012. 12(7): p. 503-16 7997 7998 Gallud A, Bondarenko O, Feliu N, Kupferschmidt N, Atluri R, Garcia-Bennett A, Fadeel B. Macrophage 7999 activation status determines the internalization of mesoporous silica particles of different sizes: 8000 Exploring the role of different pattern recognition receptors. Biomaterials. 2017 Mar;121:28-40. 8001 Gamer, A.O., E. Leibold, and B. van Ravenzwaay, The in vitro absorption of microfine zinc oxide and 8002 titanium dioxide through porcine skin. Toxicology in vitro : an international journal published in 8003 association with BIBRA, 2006. 20(3): p. 301-7. 8004 8005 Gao L, Yang ST, Li S, Meng Y, Wang H, Lei H. Acute toxicity of zinc oxide nanoparticles to the rat 8006 olfactory system after intranasal instillation. J Appl Toxicol. 2013;33(10):1079-88. 8007 8008 H. Gao, Q. He, Expert Opin. Drug Deliv. 11 (2014) 000 8009 8010 Garcia-Reyero N, Kennedy AJ, Escalon BL, Habib T, Laird JG, Rawat A, Wiseman S, Hecker M, Denslow 8011 N, Steevens JA, Perkins EJ. Differential effects and potential adverse outcomes of ionic silver and 8012 silver nanoparticles in vivo and in vitro. Environ Sci Technol. 48, 4546-4555, 2014. 8013 8014 Garcia GJ, Kimbell JS. Deposition of inhaled nanoparticles in the rat nasal passages: dose to the 8015 olfactory region. Inhal Toxicol. 2009;21(14):1165-75. 8016 8017 Gasser M, Wick P, Clift MJ, Blank F, Diener L, Yan B, Gehr P, Krug HF, Rothen-Rutishauser B. 8018 Pulmonary surfactant coating of multi-walled carbon nanotubes (MWCNTs) influences their oxidative 8019 and pro-inflammatory potential in vitro. Part Fibre Toxicol. 2012 May 24;9:17. 8020 8021 Gause CB, Layman RM, Small AC. (2011). A small business approach to nanomaterial environment, 8022 health, and safety. Journal of Occupational and Environmental Medicine, 53, S28–S31.

258

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8023 8024 Gatti AM. Biocompatibility of micro- and nano-particles in the colon. Part II. Biomaterials 25, 385- 8025 392, 2004. 8026 8027 Gautam A., Singh, Vijayaraghavan R, Dermal Exposure of Nanoparticles: An Understanding. Journal of 8028 Cell and Tissue Research 2011; 11(1): 2703-2708. 8029 8030 Ge L, Li Q, Wang M, Ouyang J, Li X, Xing M. Nanosilver particles in medical applications: synthesis, 8031 performance, and toxicity. Int J Nanomedicine. 4, 2399-2407, 2014. 8032 8033 C. Ge, J. Du, L. Zhao, L. Wang, Y. Liu, D. Li, Y. Yang, R. Zhou, Y. Zhao, Z. Chai, C. Chen. Binding of blood 8034 proteins to carbon nanotubes reduces cytotoxicity. Proc. Natl. Acad. Sci. U.S.A. 108 (2011) 16968- 8035 16973 8036 8037 Geetha, C.S., et al., Cells-nano interactions and molecular toxicity after delayed hypersensitivity, in 8038 guinea pigs on exposure to hydroxyapatite nanoparticles. Colloids Surf B Biointerfaces, 2013. 112: p. 8039 204-12. 8040 8041 Gehr P, Green FH, Geiser M, Im Hof V, Lee MM, Schurch S: Airway surfactant, a primary defense 8042 barrier: mechanical and immunological aspects. J Aerosol Med 1996, 9: 163-81. 8043 8044 Geiser M, Kreyling WG. (2010). Deposition and biokinetics of inhaled nanoparticles. Part Fibre 8045 Toxicol. 7:2. 8046 8047 Geiser M, Wigge C, Conrad ML, Eigeldinger-Berthou S, Künzi L, Garn H, Renz H, Mall MA. Nanoparticle 8048 uptake by airway phagocytes after fungal spore challenge in murine allergic asthma and chronic 8049 bronchitis. BMC Pulm Med. 2014; 14:116. 8050 8051 F. Geissmann, S. Gordon, D.A. Hume, A.M. Mowat, G.J. Randolph, Unravelling mononuclear 8052 phagocyte heterogeneity, Nat Rev Immunol, 10 (2010) 453-460. 8053 8054 Genicio N, Gallo Paramo J, Shortt AJ. Quantum dot labeling and tracking of cultured limbal epithelial 8055 cell transplants in vitro. Invest Ophthalmol Vis Sci. 2015;56(5):3051-9. 8056

259

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8057 Geraets L, Oomen AG, Krystek P, Jacobsen NR, Wallin H, Laurentie M, Verharen HW, Brandon EFA, De 8058 Jong WH.( 2014). Tissue distribution and elimination after oral and intravenous adminisrtation of 8059 diffenrent titanium dioxide nanoparticles in rats. Part Fibre Toxicol 11:30. 8060 8061 Gerloff K, Pereira DI, Faria N, Boots AW, Kolling J, Förster I, Albrecht C, Powell JJ, Schins RP. Influence 8062 of simulated gastrointestinal conditions on particle-induced cytotoxicity and interleukin-8 regulation 8063 in differentiated and undifferentiated Caco-2 cells. Nanotoxicology. 2013;7(4):353-66. 8064 8065 Gerloff K et al., The Adverse Outcome Pathway approach in nanotoxicology, Comput. Toxicol. (2016), 8066 http://dx.doi.org/10.1016/j.comtox.2016.07.001 8067 8068 Germolec D (2009) Explanation of levels of evidence for immune system toxicity. National Toxicology 8069 Program, U.S. Department of Health and Human Services. http://ntp.niehs.gov/go/9399 8070 8071 Ghoneum M, Pan D, Katano H (2014). Enhancement of Human T Lymphocyte Proliferation by 8072 Nanodiamond and Nanoplatinum in Liquid, DPV576. JSM Nanotechnol Nanomed 2(2): 1031 8073 8074 Giannakou, C., Geertsma, R. E., de Jong, W. H., van Loveren, H., Vandebriel, R. J., and Park, M. V. D. Z. 8075 (2016). Immunotoxicity Testing of Nanomedicinal Products: Possible Pitfalls in Endotoxin 8076 Determination. Current Bionanotechnology 2(in print). 8077 8078 Giannakou, C., Park, M. V., de Jong, W. H., van Loveren, H., Vandebriel, R. J., and Geertsma, R. E. 8079 (2016). A comparison of immunotoxic effects of nanomedicinal products with regulatory 8080 immunotoxicity testing requirements. Int J Nanomedicine 11, 2935-52. 8081 8082 Gill MA. The role of dendritic cells in asthma. J Allergy Clin Immunol. 2012; 129(4):889-901. 8083 8084 Girtsman TA, Beamer CA, Wu N, Buford M, Holian A. IL-1R signalling is critical for regulation of multi- 8085 walled carbon nanotubes-induced acute lung inflammation in C57Bl/6 mice. Nanotoxicology. 2014 8086 Feb;8(1):17-27. 8087 8088 Glista EE, Taylor AJ, Sayers BC, et al. Nickel nanoparticles enhance platelet-derived growth factor- 8089 induced chemokine expression by mesothelial cells via prolonged mitogen-activated protein kinase 8090 activation. Am J Respir Cell Mol Biol. 2012;47(4):552-61. 8091

260

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8092 Glista-Baker EE, Taylor AJ, Sayers BC, Thompson EA, Bonner JC. Nickel nanoparticles cause 8093 exaggerated lung and airway remodeling in mice lacking the T-box transcription factor, TBX21 (T-bet). 8094 Part Fibre Toxicol. 2014; 11:7. 8095 8096 Glista-Baker EE, Taylor AJ, Sayers BC, Thompson EA, Bonner JC. Nickel nanoparticles enhance 8097 platelet-derived growth factor-induced chemokine expression by mesothelial cells via prolonged 8098 mitogen-activated protein kinase activation. Am J Respir Cell Mol Biol. 2012 Oct;47(4):552-61 8099 8100 Goldberg DS, Ghandehari H, Swaan PW. Cellular entry of G3.5 poly (amido amine) dendrimers by 8101 clathrin- and dynamin-dependent endocytosis promotes tight junctional opening in intestinal 8102 epithelia. Pharm Res. 2010;27(8):1547-57. 8103 8104 Goldberg DS, Vijayalakshmi N, Swaan PW, Ghandehari H. G3.5 PAMAM dendrimers enhance 8105 transepithelial transport of SN38 while minimizing gastrointestinal toxicity. J Control Release. 8106 2011;150(3):318-25. 8107 8108 Göller D, Stintz M, Hillemann L, Vorbau M. 2010. Characterization of Nanoparticle Release from 8109 Surface Coatings by the Simulation of a Sanding Process. Ann. Occup. Hyg., 54 (6): 615–624 8110 8111 Gordon S, Martinez FO. Alternative activation of macrophages: mechanism and functions. Immunity. 8112 2010; 32(5):593-604. 8113 8114 Granum B, Gaarder PI, Løvik M. IgE adjuvant effect caused by particles — immediate and 8115 delayed effects. Toxicology 2001;156:149-159. 8116 8117 GRASSIAN VH, ADAMCAKOVA-DODD A, PETTIBONE JM, O’SHAUGHNESSY PT, THORNE PS 8118 Inflammatory response of mice to manufactured titanium dioxide nanoparticles: Comparison of size 8119 effects through different exposure routes. Nanotoxicology 1, 211-226, 2007 8120 8121 Gratton SE, Ropp PA, Pohlhaus PD, Luft JC, Madden VJ, Napier ME, DeSimone JM. The effect of 8122 particle design on cellular internalization pathways. Proc Natl Acad Sci U S A. 2008 Aug 8123 19;105(33):11613-8. 8124 8125 Gstraunthaler G, Hartung T: Good cell culture practice: good laboratory practice in the cell culture 8126 laboratory for the standardization and quality assurance of in vitro studies. In Cell culture models of

261

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8127 biological barriers. In-vitro test systems for drug absorption and delivery. Edited by Lehr C-M. 8128 London, New York: Taylor and Francis; 2002:112-120 8129 8130 Guadagnini R et al., Nanotoxicology. 2015 May;9 Suppl 1:13-24. 8131 8132 Gustafsson Å, Lindstedt E, Elfsmark LS, Bucht A. Lung exposure of titanium dioxide nanoparticles 8133 induces innate immune activation and long-lasting lymphocyte response in the dark agouti rat. J 8134 Immunotoxicol. 2011 Jun;8(2):111-21. 8135 8136 Gustafsson A, Jonasson S, Sandström T, Lorentzen JC, Bucht A. Genetic variation influences immune 8137 responses in sensitive rats following exposure to TiO2 nanoparticles. Toxicology. 2014 Dec 4;326:74- 8138 85. 8139 8140 Guhmann M, Thommes M, Gerber F, Pollinger N, Klein S, Breitkreutz J, WeitschiesW. (2013). Design 8141 of biorelevant test setups for the prediction of diclofenac in vivo features after oral administration. 8142 Pharm Res 30: 1483–501.anomaterials from solid nanocomposites, Particle and Fibre Toxicology 8143 11:17 8144 8145 Gulson B, McCall M, Korsch M, Gomez L, Casey P, Oytam Y, Taylor A, McCulloch M, Trotter J, Kinsley 8146 L, Greenoak G. (2010). Small amounts of zinc from zinc oxide particles in sunscreens applied outdoors 8147 are absorbed through human skin Toxicol Sc 118, 140-149. 8148 8149 Gulumian M, Savolainen K. 2012.Toxicological issues in developed and developing countries: the 8150 difference is in approach and not in content. Hum Exp Toxicol 31:205-6. 8151 8152 Gurr JR, Wang AS, Chen CH, Jan KY: Ultrafine titanium dioxide particles in the absence of 8153 photoactivation can induce oxidative damage to human bronchial epithelial cells. Toxicology 2005, 8154 213: 66-73 8155 8156 Gurr JR, Wang AS, Chen CH, Jan KY: Ultrafine titanium dioxide particles in the absence of 8157 photoactivation can induce oxidative damage to human bronchial epithelial cells. Toxicology 2005, 8158 213: 66-73 8159

262

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8160 Gurr JR, Wang AS, Chen CH, Jan KY: Ultrafine titanium dioxide particles in the absence of 8161 photoactivation can induce oxidative damage to human bronchial epithelial cells. Toxicology 2005, 8162 213: 66-73. 8163 8164 Guseva Canu I, Bateson TF, Bouvard V, Debia M, Dion C, Savolainen K, Yu IJ. Human exposure to 8165 carbon-based fibrous nanomaterials: A review. Int J Hyg Environ Health. 2016, 219(2):166-75. 8166 8167 Hadrup N, Lam HR. Regul Toxicol Pharmacol. 2014 Feb;68(1):1-7. 8168 8169 Sabina Halappanavar,1,* Petra Jackson,2,3 Andrew Williams,1 Keld A Jensen,2 Karin S Hougaard,2 8170 Ulla Vogel,2,3,4 Carole L Yauk,1 and Håkan Wallin. Environ Mol Mutagen. 2011 Jul; 52(6): 425–439. 8171 8172 Halappanavr et al 2017 WIREs Nanomed Nanobiotechnol 2017, e1465. doi: 10.1002/wnan.1465 8173 8174 Halappanavar S, Saber AT, Decan N, Jensen KA, Wu D, Jacobsen NR, Guo C, Rogowski J, Koponen IK, 8175 Levin M, Madsen AM, Atluri R, Snitka V, Birkedal RK, Rickerby D, Williams A, Wallin H, Yauk CL, Vogel 8176 U. Transcriptional profiling identifies physicochemical properties of nanomaterials that are 8177 determinants of the in vivo pulmonary response. Environ Mol Mutagen. 2015, 56, 245-64 8178 8179 Halappanavar S, Jackson P, Williams A, Jensen KA, Hougaard KS, Vogel U, Yauk CL, Wallin H. 8180 Pulmonary response to surface-coated nanotitanium dioxide particles includes induction of acute 8181 phase response genes, inflammatory cascades, and changes in microRNAs: a toxicogenomic study. 8182 Environ Mol Mutagen. 2011, 52, 425-439. 8183 8184 Halappanavar S, Vogel U, Wallin H, Yauk CL. Promise and peril in nanomedicine: the challenges and 8185 needs for integrated systems biology approaches to define health risk. Submitted, WIREs 2016. 8186 8187 Halappanavar et al in preparation; Sauer UG et al., Toxicology and applied pharamacology Volume 8188 276, issue 1, 1-20.; 8189 8190 Hamad I, Al-Hanbali O, Hunter AC, Rutt KJ, Andresen TL, Moghimi SM. Distinct polymer architecture 8191 mediates switching of complement activation pathways at the nanosphere-serum interface: 8192 implications for stealth nanoparticle engineering. ACS Nano. 2010 Nov 23;4(11):6629-38. 8193

263

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8194 Hamilton RF, Wu N, Porter D, Buford M, Wolfarth M, Holian A. Particle length-dependent titanium 8195 dioxide nanomaterials toxicity and bioactivity. Part Fibre Toxicol. 2009 Dec 31;6:35. 8196 8197 Hamilton RF Jr, Buford M, Xiang C, Wu N, Holian A. LRP3 inflammasome activation in murine alveolar 8198 macrophages and related lung pathology is associated with MWCNT nickel contamination. Inhal 8199 Toxicol. 2012, 24(14):995-1008. 8200 8201 Hirn S et al., Beilstein J Nanotechnology 2014, 5, 2440-2449 8202 8203 Halappanavar et al 2010 8204 8205 Halappanavar S, Jackson P, Williams A, Jensen KA, Hougaard KS, Vogel U, Yauk CL, Wallin H. Environ 8206 Mol Mutagen. 2011 Jul;52(6):425-39. 8207 8208 Halappanavar etal., Environ Mol Mutagen. 2015 Mar;56(2):245-64; 8209 8210 J.B. Hall, M.A. Dobrovolskaia, A.K. Patri, S.E. McNeil, Nanomedicine (Lond) 2 (2007) 789 8211 8212 Hamilton RF Jr, Buford M, Xiang C, et al. NLRP3 inflammasome activation in murine alveolar 8213 macrophages and related lung pathology is associated with MWCNT nickel contamination. Inhal 8214 Toxicol. 2012;24(14):995-1008. 8215 8216 Han X, Geller B, Moniz K, Das P, Chippindale AK, Walker VK. Monitoring the developmental impact of 8217 copper and silver nanoparticle exposure in Drosophila and their microbiomes. Sci Total Environ 2014; 8218 487:822-9. 8219 8220 Handler JS, Green N, Steele RE. Cultures as epithelial models: porous-bottom culture dishes for 8221 studying transport and differentiation. Methods Enzymol. 1989;171:736-44. 8222 8223 Hansen SF and Baun A. Dose Response. 2012; 10(3): 364–383.Wiley Interdiscip Rev Nanomed 8224 Nanobiotechnol. 2010 Sep-Oct;2(5):441-9. 8225 8226 Hardy CL, LeMasurier JS, Belz GT, et al. Inert 50-nm polystyrene nanoparticles that modify pulmonary 8227 dendritic cell function and inhibit allergic airway inflammation. J Immunol. 188(3), 1431–41 (2012). 8228

264

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8229 Hartung T, Sabbioni E. 2011. Alternative in vitro assays in nanomaterial toxicology. Wiley Interdiscip 8230 Rev Nanomed Nanobiotechnol. 2011 Jul 15. doi: 10.1002/wnan.153. [Epub ahead of print] 8231 8232 Hasse et al TAAP 2016 8233 8234 Hata K, Higashisaka K, Nagano K, Mukai Y, Kamada H, Tsunoda S, Yoshioka Y, Tsutsumi Y. Evaluation 8235 of silica nanoparticle binding to major human blood proteins. Nanoscale Res Lett. 2014;9:2493. 8236 8237 He P, Zou Y, Hu Z. Advances in aluminum hydroxide-based adjuvant research and its mechanism. 8238 Hum Vaccin Immunother. 2015;11:477-488. 8239 8240 Hedmer M, Kåredal M, Gustavsson P, Rissler J. 2013. The Nordic Expert Group for Criteria 8241 Documentation of Health Risks from Chemicals; Arbetsmiljöverket. 148. Carbon nanotubes. Arbete 8242 och Hälsa, 2013; 47 (5). Https://gupea.ub.gu.se/bitstream/2077/34499/1/gupea_2077_34 8243 499_1.pdf 8244 8245 Henri, S., et al., CD207+ CD103+ dermal dendritic cells cross-present keratinocyte-derived antigens 8246 irrespective of the presence of Langerhans cells. J Exp Med, 2010. 207(1): p. 189-206. 8247 8248 Herzog F, Clift MJ, Piccapietra F, Behra R, Schmid O, Petri-Fink A et al.: Exposure of silver- 8249 nanoparticles and silver-ions to lung cells in vitro at the air-liquid interface. Part Fibre Toxicol 2013, 8250 10: 11. 8251 8252 Hessel, E. V., Tonk, E. C., Bos, P. M., van Loveren, H., and Piersma, A. H. (2015). Developmental 8253 immunotoxicity of chemicals in rodents and its possible regulatory impact. Critical reviews in 8254 toxicology 45(1), 68-82. 8255 8256 Heydrnejad MS, Samani RJ, Aghaeivanda S. Toxic Effects of Silver Nanoparticles on Liver and Some 8257 Hematological Parameters in Male and Female Mice (Mus musculus). Biol Trace Elem Res. 165, 153- 8258 158, 2015 8259 8260 Heyl T. Contact dermatitis from silver coat. Contact Dermatitis 5, 197, 1979. 8261

265

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8262 Hinderliter PM, Minard KR, Orr G, Chrisler WB, Thrall BD, Pounds JG, Teeguarden JG: ISDD: A 8263 computational model of particle sedimentation, diffusion and target cell dosimetry for in vitro 8264 toxicity studies. Part Fibre Toxicol 2010, 7:36 8265 8266 Hillyer JF and Albrecht RM. (2001). Gastrointestinal persorption and tissue distribution of differently 8267 sized colloidal gold nanoparticles. J. Pharm. Sci., 90, 1927-1936. 8268 8269 Hirai T, Yoshikawa T, Nabeshi H, Yoshida T, Tochigi S, Ichihashi K, Uji M, Akase T, Nagano K, Abe Y, 8270 Kamada H, Itoh N, Tsunoda S, Yoshioka Y, Tsutsumi Y. Amorphous silica nanoparticles size- 8271 dependently aggravate atopic dermatitis-like skin lesions following an intradermal injection. Part 8272 Fibre Toxicol. 2012;9:3. 8273 8274 Hirai, T., et al., Cutaneous exposure to agglomerates of silica nanoparticles and allergen results in IgE- 8275 biased immune response and increased sensitivity to anaphylaxis in mice. Part Fibre Toxicol, 2015. 8276 12(1): p. 16 8277 8278 Hirano, S., Fujitani, Y., Furuyama, A., et al. Uptake and cytotoxic effects of multi-walled carbon 8279 nantobes in human bronchial epithelial cells. Toxicol Appl Pharmacol. 2010; 249(1): 8-15. 8280 8281 Hirn S, Semmler-Behnke M, Schleh C, Wenk A, Lipka J, Schäffler M, Takenaka S, Möller W, Schmid G, 8282 Simon U, Kreyling WG. (2011). Particle size-dependent and surface charge-dependent biodistribution 8283 of gold nanoparticles after intravenous administration. European Journal of Pharmaceutics and 8284 Biopharmaceutics 77 (2011) 407–416 8285 8286 Holgate ST. Innate and adaptive immune responses in asthma. Nat Med. 2012 May 4;18(5):673-83. 8287 8288 Holt PG, Strickland DH, Wikström ME, Jahnsen FL. Regulation of immunological homeostasis in the 8289 respiratory tract. Nat. Rev. Immunol. 8(2), 142–52 (2008). 8290 8291 Hooper LV, Littman DR, Macpherson AJ. Interactions between the microbiota and the immune 8292 system. Science 2012; 336:1268-73. 8293 8294 Horvat, T., Landesmann, B., Lostia, A. et al. From Pathways to People: Applying the Adverse Outcome 8295 Pathway (AOP) for Skin Sensitization to Risk Assessment Arch Toxicol 2016, doi:10.1007/s00204-016- 8296 1814-8

266

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8297 8298 Hougaard KS, Campagnolo L, Chavatte-Palmer P, Tarrade A, Rousseau-Ralliard D, Valentino S, Park 8299 MV, De Jong WH, Wolterink G, Piersma AH, Ross BL, Hutchison GR, Hansen JS, Vogel U, Jackson P, 8300 Slama R, Pietroiusti A, Cassee FR.(2015) A perspective on the developmental toxicity of inhaled 8301 nanoparticles. Reprod. Toxicol., 56, 118-140. 8302 8303 Huai LP, Liu H, Zhang N, Sun JF, Wang YY, Wei H. The effect of the different size of aluminum 8304 hydroxide particles on optimum immunity of diphtheria toxoid. Prog in Microbiol Immunol 2011; 8305 39:15-16. 8306 8307 Huan Meng, Tian Xia, Saji George,†,§ and Andre E. Nel 2009 ACS nano vol 3 no 7 8308 8309 Huang YJ, Boushey HA. The microbiome in asthma. J Allergy Clin Immunol 2015; 135:25-30 8310 8311 J.A. Hubbell, S.N. Thomas, M.A. Swartz. Materials engineering for immunomodulation. Nature 462 8312 (2009) 449-460 8313 8314 Hummel TZ, Kindermann A, Stokkers PC, Benninga MA, ten Kate FJ. Exogenous pigment in Peyer 8315 patches of children suspected of having IBD. J Pediatr Gastroenterol Nutr. 2014;58:477-80. 8316 8317 Hurley JC: Endotoxemia: methods of detection and clinical correlates. Clin Microbiol Rev 1995, 8: 8318 268–292. 8319 8320 Hussain S, Al-Nsour F, Rice AB, Marshburn J, Yingling B, Ji Z, Zink JI, Walker NJ, Garantziotis S. Cerium 8321 dioxide nanoparticles induce apoptosis and autophagy in human peripheral blood monocytes. ACS 8322 Nano. 2012 Jul 24;6(7):5820-9 8323 8324 Hussain S, Vanoirbeek JA, Hoet PH. Interactions of nanomaterials with the immune system. Wiley 8325 Interdiscip Rev Nanomed Nanobiotechnol. 4, 169-183, 2012. 8326 8327 Husain M, Wu D, Saber AT, Decan N, Jacobsen NR, Williams A, Yauk CL, Wallin H, Vogel U, 8328 Halappanavar S. Intratracheally instilled titanium dioxide nanoparticles translocate to heart and liver 8329 and activate complement cascade in the heart of C57BL/6 mice. Nanotoxicology. 2015; 9, 1013-22 8330

267

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8331 Husain M, Saber AT, Guo C, Jacobsen NR, Jensen KA, Yauk CL, Williams A, Vogel U, Wallin H, 8332 Halappanavar S. Pulmonary instillation of low doses of titanium dioxide nanoparticles in mice leads 8333 to particle retention and gene expression changes in the absence of inflammation. Toxicol Appl 8334 Pharmacol. 2013, 269, 250-262. 8335 8336 Husain M, Saber AT, Guo C, Jacobsen NR, Jensen KA, Yauk CL, Williams A, Vogel U, Wallin H, 8337 Halappanavar S. Toxicol Appl Pharmacol. 2013 8338 8339 Husain M, Wu D, Saber AT, Decan N, Jacobsen NR, Williams A, Yauk CL, Wallin H, Vogel U, 8340 Halappanavar S. Nanotoxicology. 2015;9(8):1013-22 8341 8342 Hussain S, Sangtian S, Anderson SM, Snyder RJ, Marshburn JD, Rice AB, Bonner JC, Garantziotis S. 8343 Inflammasome activation in airway epithelial cells after multi-walled carbon nanotube exposure 8344 mediates a profibrotic response in lung fibroblasts. Part Fibre Toxicol. 2014 Jun 10;11:28. doi: 8345 10.1186/1743-8977-11-28. 8346 8347 Husain M et al., Volume 269, Issue 3, 15 June 2013, Pages 250–262 8348 8349 Hussain SM, Warheit DB, Ng SP, Comfort KK, Grabinski CM, Braydish-Stolle LK (2015) At the 8350 crossroads of nanotoxicology in vitro: Past achievements and current challenges. Tox Sci 147: 5-16. 8351 8352 Hussain S, Vanoirbeek JA, Luyts K, et al. Lung exposure to nanoparticles modulates an asthmatic 8353 response in a mouse model. Eur Respir J. 2011;37(2):299-309. 8354 8355 Hussain S, Al-Nsour F, Rice AB, et al. Cerium dioxide nanoparticles induce apoptosis and autophagy in 8356 human peripheral blood monocytes. ACS Nano 2012;6(7):5820-9. 8357 8358 Hussain S, Smulders S, De Vooght V, Ectors B, Boland S, Marano F, Van Landuyt KL, Nemery B, Hoet 8359 PH, Vanoirbeek JA. Nano-titanium dioxide modulates the dermal sensitization potency of DNCB. 8360 Part Fibre Toxicol. 2012;9:15. 8361 8362 Hussain S, Vanoirbeek JA, Luyts K, De Vooght V, Verbeken E, Thomassen LCJ, Martens JA, Dinsdale D, 8363 Boland S, Marano F, Nemery B. Hoet PHM. Lung exposure to nanoparticles modulates an asthmatic 8364 response in a mouse model. Eur Respir J 2011; 37: 299–309. 8365

268

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8366 Hussain S, Al-Nsour F, Rice AB, Marshburn J, Yingling B, Ji Z, Zink JI, Walker NJ, Garantziotis S. Cerium 8367 dioxide nanoparticles induce apoptosis and autophagy in human peripheral blood monocytes. 8368 ACS Nano. 2012 Jul 24;6(7):5820-9. 8369 8370 IARC [in press]. IARC monographs on the evaluation of carcinogenic risks to humans: Volume 111. 8371 Some nanomaterials and some fibres. Lyon:IARC. 8372 8373 ICH (2005) S8. Immunotoxicity studies for human pharmaceuticals. 8374 http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Safety/S8/Step4/S8_Guide 8375 line.pdf 8376 8377 ICH (2011) S6(R1). Preclinical safety evaluation of biotechnology-derived pharmaceuticals. 8378 http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Safety/S6_R1/Step4/S6_R 8379 1_Guideline.pdf 8380 8381 Ichinose T, Takano H, Miyabara Y, Yanagisawa R, Sagai M. Murine strain differences in allergic airway 8382 inflammation and immunoglobulin production by a combination of antigen and diesel exhaust 8383 particles. Toxicology 1997;122:183–192. 8384 8385 ICICIS Group Investigators (1998) Report of validation study of assessment of direct immunotoxicity 8386 in the rat. Toxicology 125: 183-201. 8387 8388 ICRP, Human respiratory tract model for radiological protection. Oxford, England: Pergamon, Elsevier 8389 Science Ltd., International Commission on Radiological Protection Publication No. 66, 1994. 8390 8391 IPCS. EHC 180. Principles and Methods for Assessing Direct Immunotoxicity Associated with Exposure 8392 to Chemicals. WHO, Geneva, 1996. 8393 8394 Igyarto, B.Z., et al., Skin-resident murine dendritic cell subsets promote distinct and opposing 8395 antigen-specific T helper cell responses. Immunity, 2011. 35(2): p. 260-72. 8396 8397 Ilinskaya AN, Dobrovolskaia MA. Immunosuppressive and anti-inflammatory properties of 8398 engineered nanomaterials. Br J Pharmacol. 2014 Sep;171(17):3988-4000. 8399

269

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8400 Ilinskaya AN, Dobrovolskaia MA. Understanding the immunogenicity and antigenicity of 8401 nanomaterials: Past, present and future. Toxicol Appl Pharmacol. 2016 May 15;299:70-7. 8402 8403 Ilves, M., et al., Topically applied ZnO nanoparticles suppress allergen induced skin inflammation but 8404 induce vigorous IgE production in the atopic dermatitis mouse model. Part Fibre Toxicol, 2014. 11: p. 8405 38 8406 8407 International Organization of Satndardization. ISO/TS 80004-1, 2015. 8408 8409 Inoue K, Yanagisawa R, Koike E, Nishikawa M, Takano H. Repeated pulmonary exposure to single- 8410 walled carbon nanotubes exacerbates allergic inflammation of the airway: Possible role of oxidative 8411 stress. Free Radic Biol Med. 2010, 48(7): 924-934. 8412 8413 Inoue K, Koike E, Yanagisawa R, Hirano S, Nishikawa M, Takano H. Effects of multi-walled carbon 8414 nanotubes on a murine allergic airway inflammation model. Toxicol Appl Pharmacol. 2009, 8415 15;237(3):306-16. 8416 8417 Inoue K, Takano H, Koike E, Yanagisawa R, Sakurai M, Tasaka S, Ishizaka A, Shimada A. Effects of 8418 pulmonary exposure to carbon nanotubes on lung and systemic inflammation with coagulatory 8419 disturbance induced by lipopolysaccharide in mice. Exp Biol Med (Maywood). 2008, 233(12):1583-90. 8420 8421 Inoue K, Takano H, Yanagisawa R, Hirano S, Sakurai M, Shimada A, Yoshikawa T. 2006. Effects of 8422 airway exposure to nanoparticles on lung inflammation induced by bacterial endotoxin in mice. 8423 Environ Health Perspect. 114(9):1325-30. 8424 8425 Inoue K, Takano H, Koike E, et al. Effects of pulmonary exposure to carbon nanotubes on lung and 8426 systemic inflammation with coagulatory disturbance induced by lipopolysaccharide in mice. Exp Biol 8427 Med. 2008;233(12):1583-1590. 8428 8429 Inoue K, Takano H. (2011) Aggravating impact of nanoparticles on immune-mediated pulmonary 8430 inflammation. ScientificWorldJournal. 2011;11:382-90. 8431 8432 Inoue K, Koike E, Yanagisawa R, et al. Effects of multi-walled carbon nanotubes on a murine allergic 8433 airway inflammation model. Toxicol Appl Pharmacol. 2009;237(3): 306-316. 8434

270

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8435 Inoue K, Yanagisawa R, Koike E, et al. Repeated pulmonary exposure to single-walled carbon 8436 nanotubes exacerbates allergic inflammation of the airway: Possible role of oxidative stress. Free 8437 Radic Biol Med. 2010;48(7): 924-934. 8438 8439 Iravani S, Korbekandi H, Mirmohammadi SV, Zolfaghari B. Synthesis of silver nanoparticles: chemical, 8440 physical and biological methods. Res Pharm Sci. 2014;9(6):385-406. 8441 8442 ISO TR 18637. 2005. Nanotechnology- General Framework for the Development of Occupational 8443 Exposure Limits and Bands for Nano-Objects and their Aggregates and Agglomerates 8444 8445 Jackman JA, Mészáros T, Fülöp T, Urbanics R, Szebeni J, Cho NJ. Comparison of complement 8446 activation-related pseudoallergy in miniature and domestic pigs: foundation of a validatable immune 8447 toxicity model. Nanomedicine. 2016 May;12(4):933-43. doi: 10.1016/j.nano.2015.12.377. Epub 2016 8448 Jan 6. 8449 8450 Janer G, Mas del Molino E, Fernández-Rosas E, Fernández A, Vázquez-Campos S. Cell uptake and oral 8451 absorption of titanium dioxide nanoparticles. Toxicol Lett. 2014;228(2):103-10. 8452 8453 Jani PU, McCarthy DE, Florence AT. Titanium dioxide (rutile) particle uptake from the rat GI tract and 8454 translocation to systemic organs after oral administration. Int J Pharm. 1994, 105;157-168. 8455 8456 Jessop F, Holian A. Extracellular HMGB1 regulates multi-walled carbon nanotube-induced 8457 inflammation in vivo. Nanotoxicology. 2015 May;9(3):365-72. 8458 8459 Jepps OG, Dancik Y, Anissimov YG, Roberts MS. Modeling the human skin barrier--towards a better 8460 understanding of dermal absorption. Adv Drug Deliv Rev. 2013 Feb;65(2):152-68. 8461 8462 Jeong GN, Ryu HY, Kim YS, Song KS, Yu IJ. Histochemical study of intestinal mucins after 8463 administration of silver nanoparticles in Sprague-Dawley rats. Arch Toxicol. 84: 63-69, 2010 8464 8465 Ji JH, Jung JH, Kim SS, Yoon JU, Park JD, Choi BS, Chung YH, Kwon IH, Jeong J, Han BS, Shin JH, Sung 8466 JH, Song KS, Yu IJ. Twenty-eight-day inhalation toxicity study of silver nanoparticles in Sprague- 8467 Dawley rats. Inhal Toxicol. 19, 857-871, 2007. 8468 8469 Ji et al.(2007), Sung et al.(2008)

271

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8470 8471 Johnston HJ, Hutchison G, Christensen FM, Peters S, Hankin S, Stone V. A review of the in vivo and in 8472 vitro toxicity of silver and gold particulates: Particle attributes and biological mechanisms responsible 8473 for the observed toxicity. Crit. Rev. Toxicol. 4, 328–346, 2010. 8474 8475 Jones SW, Roberts RA, Robbins GR, Perry JL, Kai MP, Chen K, Bo T, Napier ME, Ting JP, Desimone JM, 8476 Bear JE. Nanoparticle clearance is governed by Th1/Th2 immunity and strain background. J Clin 8477 Invest. 2013 Jul;123(7):3061-73. 8478 8479 Jones K, Morton J, Smith I, Jurkschat K, Harding AH, Evans G. Human in vivo and in vitro studies on 8480 gastrointestinal absorption of titanium dioxide nanoparticles. Toxicol Lett. 2015;233:95-101. 8481 8482 Johnston et al 2012; others - US…? ) 8483 8484 Johnson DS, Duncan SE, Bianchi LM, Chang HH, Eigel WN, O'Keefe SF1. Packaging modifications for 8485 protecting flavor of extended-shelf-life milk from light. J Dairy Sci. 2015;98(4):2205-14. 8486 8487 Johnston H, Pojana G, Zuin S, Jacobsen NR, Møller P, Loft S, Semmler-Behnke M, McGuiness C, 8488 Balharry D, Marcomini A, Wallin H, Kreyling W, Donaldson K, Tran L, Stone V (2013). Engineered 8489 nanomaterial risk. Lessons learnt from completed nanotoxicology studies: potential solutions to 8490 current and future challenges. Crit Rev Tox 43: 1-20. 8491 8492 Jorch SK, Kubes P. An emerging role for neutrophil extracellular traps in noninfectious disease. Nat 8493 Med. 2017 Mar 7;23(3):279-287. 8494 8495 Jovanović B. Critical review of public health regulations of titanium dioxide, a human food additive. 8496 Integr Environ Assess Manag. 2015;11(1):10-20. 8497 8498 Juling S, Bachler G, von Götz N, Lichtenstein D, Böhmert L, Niedzwiecka A, Selve S, Braeuning A, 8499 Lampen A. In vivo distribution of nanosilver in the rat: The role of ions and de novo-formed 8500 secondary particles. Food Chem Toxicol. 97, 327-335, 2016. 8501 8502 Kahru A, Savolainen K. 2010. Potential hazard of nanoparticles: from properties to biological and 8503 environmental effects. Toxicology 269:89-91. 8504

272

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8505 Kai MP, Brighton HE, Fromen CA, Shen TW, Luft JC, Luft YE, Keeler AW, Robbins GR, Ting JP, Zamboni 8506 WC, Bear JE, DeSimone JM. Tumor presence induces global immune changes and enhances 8507 nanoparticle clearance. ACS Nano. 2016 Jan 26;10(1):861-70. 8508 8509 Kao YY, Cheng TJ, Yang DM, Wang CT, Chiung YM, Liu PS. Demonstration of an olfactory bulb-brain 8510 translocation pathway for ZnO nanoparticles in rodent cells in vitro and in vivo. J Mol Neurosci. 8511 2012;48(2):464-71. 8512 8513 Kapitel med Boraschi 2012 8514 8515 Kapitel med Fernando 8516 8517 Kaplan, D.H., et al., Epidermal langerhans cell-deficient mice develop enhanced contact 8518 hypersensitivity. Immunity, 2005. 23(6): p. 611-20. 8519 8520 Karakoti AS, Shukla R, Shanker R, Singh S. Surface functionalization of quantum dots for biological 8521 applications. Adv Colloid Interface Sci. 2015;215:28-45. 8522 8523 Kapralov AA, Feng WH, Amoscato AA, Yanamala N, Balasubramanian K, Winnica DE, Kisin ER, Kotchey 8524 GP, Gou P, Sparvero LJ, Ray P, Mallampalli RK, Klein-Seetharaman J, Fadeel B, Star A, Shvedova AA, 8525 Kagan VE. Adsorption of surfactant lipids by single-walled carbon nanotubes in mouse lung upon 8526 pharyngeal aspiration. ACS Nano. 2012 May 22;6(5):4147-56. 8527 8528 Kathawala MH, Xiong S, Richards M, Ng KW, George S, Loo SCJ. Emerging In Vitro Models for Safety 8529 Screening of High Volume Production Nanomaterials under Environmentally Relevant Exposure 8530 Conditions. SMALL 9(9-10):1504-20. 2013 8531 8532 Katwa P, Wang X, Urankar RN, Podila R, Hilderbrand SC, Fick RB, Rao AM, Ke PC, Wingard CJ, Brown 8533 JM. A carbon nanotube toxicity paradigm driven by mast cells and the IL-₃₃/ST₂ axis. Small. 2012 Sep 8534 24;8(18):2904-12. 8535 8536 Ke W, Zhao Y, Huang R, Jiang C, Pei Y. Enhanced oral bioavailability of doxorubicin in a dendrimer 8537 drug delivery system. J Pharm Sci. 2008;97:2208 – 16. 8538

273

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8539 Kelleher P, Pacheco K, Newman LS. Inorganic dust pneumonias: the metal-related parenchymal 8540 disorders. Environ Health Perspect. 2000;108 (suppl. 4): 685-696. 8541 8542 Kemp SJ, Thorley AJ, Gorelik J, Seckl MJ, O'Hare MJ, Arcaro A et al.: Immortalisation of Human 8543 Alveolar Epithelial Cells to Investigate Nanoparticle Uptake. Am J Respir Cell Mol Biol 2008 8544 8545 Kermanizadeh A, Balharry D, Wallin H, Loft S, Møller P. Nanomaterial translocation-the biokinetics, 8546 tissue accumulation, toxicity and fate of materials in secondary organs-a review. Crit Rev Toxicol. 8547 2015:1-36. 8548 8549 Ketchum P, Novitsky T: Assay of Endotoxin by Limulus Amebocyte Lysate. Methods Mol Med 2000, 8550 36: 3–1 8551 8552 Khodakovskaya MV, Kim BS, Kim JN, Alimohammadi M, Dervishi E, Mustafa T, Cernigla CE. Carbon 8553 nanotubes as plant growth regulators: effects on tomato growth, reproductive system, and soil 8554 microbial community. Small 2013; 9:115-23. 8555 8556 Kim SE, Zhang L, Ma K, Riegman M, Chen F, Ingold I, Conrad M, Turker MZ, Gao M, Jiang X, Monette 8557 S, Pauliah M, Gonen M, Zanzonico P, Quinn T, Wiesner U, Bradbury MS, Overholtzer M. Ultrasmall 8558 nanoparticles induce ferroptosis in nutrient-deprived cancer cells and suppress tumour growth. Nat 8559 Nanotechnol. 2016 Nov;11(11):977-985. 8560 8561 Kim JS, Song KS, Sung JH, Ryu HR, Choi BG, Cho HS, Lee JK, Yu IJ. Genotoxicity, acute oral and dermal 8562 toxicity, eye and dermal irritation and corrosion and skin sensitisation evaluation of silver 8563 nanoparticles. Nanotoxicology 7, 953–960, 2013. 8564 8565 Kim YS, Song MY, Park JD, Song KS, Ryu HR, Chung YH, Chang HK, Lee JH, Oh KH, Kelman BJ, Hwang K, 8566 Yu IJ. Subchronic oral toxicity of silver nanoparticles. Part Fibre Toxicol 7:20, 2010 8567 8568 Kim JA, Salvati A, Åberg C, Dawson KA. Suppression of nanoparticle cytotoxicity approaching in vivo 8569 serum concentrations: limitations of in vitro testing for nanosafety. Nanoscale. 6, 14180-14184, 8570 2014. 8571 8572 Kim YS, Kim, J.S., Cho, H.S., Rha, D.S., Kim, J.M., Park, J.D., Choi, B.S., Lim, R., Chang, H.K., Chung, Y.H., 8573 Kwon, I.H., Jeong, J., Han, B.S., and Yu, I.J. (2008). Twenty-Eight-Day Oral Toxicity, Genotoxicity, and

274

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8574 Gender-Related Tissue Distribution of Silver Nanoparticles in Sprague-Dawley Rats. Inhal Toxicol 20, 8575 575–583. 8576 8577 Kim et al.(2012) 8578 8579 Kim YH, Kwak KA, Kim TS, Seok JH, Roh HS, Lee JK, Jeong J, Meang EH, Hong JS, Lee YS, Kang JS. 8580 Retinopathy Induced by Zinc Oxide Nanoparticles in Rats Assessed by Micro-computed Tomography 8581 and Histopathology. Toxicol Res. 2015;31(2):157-63. 8582 8583 Kim et al.(2010), Sung et al.(2011) 8584 8585 Kim YS, Kim JS, Cho HS, Rha DS, Kim JM, Park JD, Choi BS, Lim R, Chang HK, Chung YH, Kwon IH, Jeong 8586 J, Han BS, Yu IJ. Twenty-eight-day oral toxicity, genotoxicity, and gender-related tissue distribution of 8587 silver nanoparticles in Sprague-Dawley rats. Inhal Toxicol. 2008;20(6):575-83. 8588 8589 Kitchens KM, Foraker AB, Kolhatkar RB, Swaan PW, Ghandehari H. Endocytosis and interaction of 8590 poly (amidoamine) dendrimers with Caco-2 cells. Pharm Res. 2007;24(11):2138-45. 8591 8592 Klang V, Schwarz JC, Lenobel B, Nadj M, Auböck J, Wolzt M, Valenta C. In vitro vs. in vivo tape 8593 stripping: validation of the porcine ear model and penetration assessment of novel sucrose stearate 8594 emulsions. Eur J Pharm Biopharm. 2012, Apr;80(3):604-14. 8595 8596 Klein SG, Serchi T, Hoffmann L, Blomeke B, Gutleb AC: An improved 3D tetraculture system mimicking 8597 the cellular organisation at the alveolar barrier to study the potential toxic effects of particles on the 8598 lung. Part Fibre Toxicol 2013, 10: 31 8599 8600 Kodali V, Littke MH, Tilton SC, Teeguarden JG, Shi L, Frevert CW, Wang W, Pounds JG, Thrall BD. 8601 Dysregulation of macrophage activation profiles by engineered nanoparticles. ACS Nano. 2013 Aug 8602 27;7(8):6997-7010. 8603 8604 Kolosnjaj-Tabi J, Hartman KB, Boudjemaa S, Ananta JS, Morgant G, Szwarc H, Wilson LJ, Moussa F. In 8605 vivo behavior of large doses of ultrashort and full-length single-walled carbon nanotubes after oral 8606 and intraperitoneal administration to Swiss mice. ACS Nano. 2010;4(3):1481-92. 8607 8608 Kroll A et al ., Part Fibre Toxicol. 2011; 8: 9.

275

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8609 8610 Kreyling WG et al J toxicol Environ Health A 65, 1513-1530, 2002 8611 8612 Kreyling et al Acc Chem Res. 2013 March 19; 46(3): 714–722. 8613 8614 Kreyling WG, Hirn S, Moller W, Schleh C, Wenk A, Celik G Lipka J, Schäffler M, Haberl N, Johnston BD, 8615 Sperling R, Schmid G, Simon U, Parak WJ, Semmler-Behnke M (2014) Air–blood barrier translocation 8616 of tracheally instilled gold nanoparticles inversely depends on particle size. ACS Nano 8, 222–233. 8617 8618 Kreyling, W.G., Semmler, M., Erbe, F., Mayer, P., Takenaka, S., Schulz, H., Oberdörster, G., Ziesenis, A. 8619 (2002). Translocation of ultrafine insoluble iridium particles from lung epithelium to extrapulmonary 8620 organs is size dependent but very low. J Toxicol Environ Health, Part A 65, 1513–1530. 8621 8622 Krystek, P., Park, M. V. D. Z., Braakhuis, H., and Jong, W. H. (2013). Applications of Inductively 8623 Coupled Plasma-Mass Spectrometry in Biodistribution Studies of (Engineered) Nanoparticles. In 8624 Encyclopedia of Analytical Chemistry (R. A. Meyers, Ed.) Eds.), Vol. Published Online : 15 MAR 2013, 8625 DOI: 10.1002/9780470027318.a9337. Wiley & Sons, Ltd. 8626 8627 Krystek P, Park M, De Jong WH, Braakhuis H. (2013). Applications of Inductively Coupled Plasma- 8628 Mass Spectrometry in Biodistribution Studies of (Engineered) Nanoparticles, Published in 8629 Encyclopedia of Analytical Chemistry, Online 2006-2013 John Wiley &Sons, Ltd. 2013, DOI: 8630 10.1002/9780470027318.a9337. 8631 8632 Krystek P, Tentschert J, Nia Y, Trouiller B, Noel L, Goetz ME, Papin A, Luch A, Guérin T, De Jong WH. 8633 (2014) Method development and inter-laboratory comparison about the determination of titanium 8634 from titanium dioxide nanoparticles in tissues by inductively coupled plasma mass spectrometry. 8635 Anal Bioanal Chem 406, 3853–3861. 8636 8637 Kudo M, Ishigatsubo Y, Aoki I. Pathology of asthma. Front. Microbiol. 4, 263 (2013). 8638 8639 Kuehn A, Kletting S, de SC-W, Repnik U, Griffiths G, Fischer U et al.: Human alveolar epithelial cells 8640 expressing tight junctions to model the air-blood barrier. ALTEX 2016, 33: 251-260. 8641 8642 Kuempel ED, Jaurand MC, Møller P, Morimoto Y, Kobayashi N, Pinkerton KE, Sargent LM, Vermeulen 8643 RC, Fubini B, Kane AB. Evaluating the mechanistic evidence and key data gaps in assessing the

276

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8644 potential carcinogenicity of carbon nanotubes and nanofibers in humans. Crit Rev Toxicol. 2016, 8645 18:1-58. 8646 8647 Kuijpers E, Bekker C, Fransman W, Brouwer D, Tromp P, Vlaanderen J, Godderis L, Hoet P, Lan Q, 8648 Silverman D, Vermeulen R, Pronk A. Occupational Exposure to Multi-Walled Carbon Nanotubes 8649 During Commercial Production Synthesis and Handling. Ann Occup Hyg. 2016, 60(3):305-17. 8650 8651 Kulkarni SA and Feng S-S. (2013) Effects of particle size and surface modification on cellular uptake 8652 and biodistribution of polymeric nanoparticles for drug delivery. Pharm. Res. 30, 2512-2522. 8653 8654 Kunzmann A, Andersson B, Vogt C, Feliu N, Ye F, Gabrielsson S, Toprak MS, Buerki-Thurnherr T, 8655 Laurent S, Vahter M, Krug H, Muhammed M, Scheynius A, Fadeel B. Efficient internalization of silica- 8656 coated iron oxide nanoparticles of different sizes by primary human macrophages and dendritic cells. 8657 Toxicol Appl Pharmacol. 2011 Jun 1;253(2):81-93. 8658 8659 Kuo TR, Lee CF, Lin SJ, Dong CY, Chen CC, Tan HY. Studies of intracorneal distribution and cytotoxicity 8660 of quantum dots: risk assessment of eye exposure. Chem Res Toxicol. 2011;24(2):253-61. 8661 8662 Kuroda E, Coban C, Ishii KJ. Particulate adjuvant and innate immunity: past achievements, present 8663 findings, and future prospects. Int Rev Immunol. 2013; 32:209-220. 8664 8665 Kumar SS, Venkateswarlu P, Ranga Rao V, Rao GN. Synthesis, characterization and optical properties 8666 of zinc oxide nanoparticles. Int Nano Lett 2013;3:30 8667 8668 Kumari, M., Kumari, S. I., Kamal, S. S., and Grover, P. (2014). Genotoxicity assessment of cerium oxide 8669 nanoparticles in female Wistar rats after acute oral exposure. Mutation research. Genetic toxicology 8670 and environmental mutagenesis 775-776, 7-19. 8671 8672 Labib S, Williams A, Yauk CL, Nikota JK, Wallin H, Vogel U, Halappanavar S. Nano-risk Science: 8673 application of toxicogenomics in an adverse outcome pathway framework for risk assessment of 8674 multi-walled carbon nanotubes. Part Fibre Toxicol. 2016, 13:15. doi: 10.1186/s12989-016-0125-9. 8675 8676 Labib S, Williams A, Kuo B, Yauk CL, White PA, Halappanavar S. A framework for the use of single- 8677 chemical transcriptomics data in predicting the hazards associated with complex mixtures of 8678 polycyclic aromatic hydrocarbons.Arch Toxicol. 2016 Nov 17. [Epub ahead of print]

277

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8679 8680 Labib S, Williams A, Yauk CL, Nikota JK, Wallin H, Vogel U, Halappanavar S. Nano-risk Science: 8681 application of toxicogenomics in an adverse outcome pathway framework for risk assessment of 8682 multi-walled carbon nanotubes. Part Fibre Toxicol. 2016, 13:15. doi: 10.1186/s12989-016-0125-9. 8683 8684 Labib S, Williams A, Yauk CL, Nikota JK, Wallin H, Vogel U, Halappanavar S. Nano-risk Science: 8685 application of toxicogenomics in an adverse outcome pathway framework for risk assessment of 8686 multi-walled carbon nanotubes. Part Fibre Toxicol. 2016, 13:15. doi: 10.1186/s12989-016-0125-9. 8687 8688 Lademann J, Richter H, Schaefer UF, Blume-Peytavi U, Teichmann A, Otberg N et al.: Hair follicles - a 8689 long-term reservoir for drug delivery. Skin Pharmacol Physiol 2006, 19: 232-236 8690 8691 Lademann J, Weigmann H, Rickmeyer C, Barthelmes H, Schaefer H, Mueller G et al.: Penetration of 8692 titanium dioxide microparticles in a sunscreen formulation into the horny layer and the follicular 8693 orifice. Skin Pharmacol Appl Skin Physiol 1999, 12: 247-256 8694 8695 Lambrecht BN, Hammad H. The role of dendritic and epithelial cells as master regulators of allergic 8696 airway inflammation. Lancet. 376(9743), 835–43 (2010) 8697 8698 Landsiedel R, Fabian E, Ma-Hock L, Wohlleben W, Wiench K, Oesch F, Van Ravenzwaay B. (2012). 8699 Toxico-/biokinetics of nanomaterials. Arch Toxicol 86, 1021–1060. 8700 8701 Lankoff A, Sandberg WJ, Wegierek-Ciuk A, Lisowska H, Refsnes M, Sartowska B, Schwarze PE, 8702 Meczynska-Wielgosz S, Wojewodzka M, Kruszewski M. The effect of agglomeration state of silver and 8703 titanium dioxide nanoparticles on cellular response of HepG2, A549 and THP-1 cells. Toxicol Lett. 8704 208, 197-213, 2012. 8705 8706 Lankveld DPK, Oomen AG, Krystek P, Troost-De Jong A, Noorlander CW, Van Eijkeren JCH, Geertsma 8707 RE, De Jong WH. (2010). The kinetics of the tissue distribution of silver nanoparticles of differents 8708 sizes. Biomaterials 31, 8350-8361. 8709 8710 Lankveld DPK, Rayavarapu RG, Krystek P, Oomen AG, Verharen HW, Van Leeuwen TG, De Jong WH, 8711 Manohar S. Blood clearance and tissue distribution of PEGylated and non-PEGylated gold nanorods 8712 after intravenous administration in rats. Nanomedicine 6, 339- 349, 2011. 8713

278

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8714 Larese Filon F, D’Agostin F, Crosera M, Adami G, Renzi N, Bovenzi M, Maina G. (2009). Human skin 8715 penetration of silver nanoparticles through intact and damaged skin. Toxicology 255, 33-37. 8716 8717 Lappas CM. Air-blood barrier translocation of tracheally instilled gold nanoparticles inversely 8718 depends on particle size. In Handbook of Nanotoxicology, Nanomedicine and Stem Cell Use in 8719 Toxicology. Edited by Saura C. Sahu and Daniel A. Casciano. © 2014 John Wiley & Sons, Ltd. ISBN 978- 8720 1-118-43926-5. 8721 8722 G. Laverny, A. Casset, A. Purohit, E. Schaeffer, C. Spiegelhalter, F. de Blay, F. Pons. 8723 Immunomodulatory properties of multi-walled carbon nanotubes in peripheral blood mononuclear 8724 cells from healthy subjects and allergic patients. Toxicol. Lett. 217 (2013) 91-101. 8725 8726 8727 Lee JH, Mun J, Duck J, Yu IJ. A health surveillance case study on workers who manufacture silver 8728 nanomaterials. Nanotoxicology 6, 667-669, 2012. 8729 8730 Lee JH, Mun J, Park JD, Yu IJ. (2012). A health surveillance case study on workers who manufacture 8731 silver nanomaterials. Nanotoxicology, 6(6), 667–669. 8732 8733 Lee JS, Choi YC, Shin JH, Lee JH, Lee Y, Park SY, Yu, IJ. (2014). Health surveillance study of workers 8734 who manufacture multi-walled carbon nanotubes. Nanotoxicology, (0), 1–10. 8735 8736 Lee, H.M., et al., Autophagy negatively regulates keratinocyte inflammatory responses via scaffolding 8737 protein p62/SQSTM1. J Immunol, 2011. 186(2): p. 1248-58. 8738 8739 Lee JK, Sayers BC, Chun KS, et al. Multi-walled carbon nanotubes induce COX-2 and iNOS expression 8740 via MAP kinase-dependent and -independent mechanisms in mouse RAW264.7 macrophages. Part 8741 Fibre Toxicol. 2012;9:14. 8742 8743 Lee JH, Cha KE, Kim MS, Hong HW, Chung DJ, Ryu G, Myung H. Nanosized polyamidoamine (PAMAM) 8744 dendrimer-induced apoptosis mediated by mitochondrial dysfunction. Toxicol Lett. 2009;190(2):202- 8745 7. 8746 8747 Lee SC, Parthasarathy R, Botwin K, Kunneman D, Rowold E, Lange G, Klover J, Abegg A, Zobel J, Beck 8748 T, Miller T, Hood W, Monahan J, McKearn JP, Jansson R, Voliva CF. Biochemical and immunological

279

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8749 properties of cytokines conjugated to dendritic polymers. Biomed Microdevices. 2004 Sep;6(3):191- 8750 202. 8751 8752 Lefebvre DE, Venema K, Gombau L, Valerio LG Jr, Raju J, Bondy GS, Bouwmeester H, Singh RP, 8753 Clippinger AJ, Collnot EM, Mehta R, Stone V. (2015) Utility of models of the gastrointestinal tract for 8754 assessment of the digestion and absorption of engineered nanomaterials released from food 8755 matrices. Nanotoxicology 9, 523-542. 8756 8757 Lehmann AD, Daum N, Bur M, Lehr CM, Gehr P, Rothen-Rutishauser BM: An in vitro triple cell co- 8758 culture model with primary cells mimicking the human alveolar epithelial barrier. Eur J Pharm 8759 8760 Lehmann A, Brandenberger Ch, Blank F, Gehr P, Rothen-Rutishauser B: A 3D model of the human 8761 epithelial airway barrer. In Methods in Bioengineering. Alternative Technologies to Animal Testing. 8762 Edited by Maguire T, Novik E. Boston, London: Artech House; 2010:239-260 8763 8764 Lehner M., Morhart P., Stilper A., Holter W. Functional characterization of monocyte-derived 8765 dendritic cells generated under serum-free culture conditions. Immunol Lett. 2005 99(2):209-16 8766 8767 Lesniak A, Fenaroli F, Monopoli MP, Aberg C, Dawson KA, Salvati A. (2012) Effects of the presence or 8768 absence of a protein corona on silica nanoparticle uptake and impact on cells. ACS Nano 6, 5845– 8769 5857. 8770 8771 Li Y, Boraschi D. (2016) Endotoxin contamination: a key element in the interpretation of nanosafety 8772 studies. Nanomedicine (Lond). 11(3):269-87. 8773 8774 R. Li, Z. Ji, C.H. Chang, D.R. Dunphy, X. Cai, H. Meng, H. Zhang, B. Sun, X. Wang, J. Dong, S. Lin, M. 8775 Wang, Y.P. Liao, C.J. Brinker, A.E. Nel, T. Xia. Surface interactions with compartmentalized cellular 8776 phosphates explains rare earth oxide nanoparticle hazard and provides opportunities for safer 8777 design. ACS Nano 8 (2014) 1771-1783. 8778 8779 R. Li, Z. Ji, H. Qin, X. Kang, B. Sun, M. Wang, C.H. Chang, X. Wang, H. Zhang, H. Zou, A.E. Nel, T. Xia. 8780 Interference in autophagosome fusion by rare earth nanoparticles disrupts autophagic flux and 8781 regulation of an interleukin-1β producing inflammasome. ACS Nano 8 (2014) 10280-10292. 8782

280

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8783 Li J, Li L, Chen H, Chang Q, Liu X, Wu Y, Wei C, Li R, Kwan JK, Yeung KL, Xi Z, Lu Z, Yang X. Application 8784 of vitamin E to antagonize SWCNTs-induced exacerbation of allergic asthma. Sci Rep. 2014 Mar 8785 4;4:4275. 8786 8787 Li JG, Li WX, Xu JY, et al. Comparative study of pathological lesions induced by multiwalled carbon 8788 nanotubes in lungs of mice by intratracheal instillation and inhalation. Environ Toxicol 2007;22: 415- 8789 421. 8790 8791 Li N, Xia T, Nel AE. The role of oxidative stress in ambient particulate matter-induced lung diseases 8792 and its implications in the toxicity of engineered nanoparticles. Free Radic Biol Med. 2008 May 8793 1;44(9):1689-99. 8794 8795 Liao HY, Chung YT, Lai CH, Wang SL, Chiang HC, Li LA, et al. 2013. Six-month follow-up study of health 8796 markers of nanomaterials among workers handling engineered nanomaterials. Nanotoxicology 8797 1:100–10. 8798 8799 Liao HY, Chung YT, Lai CH, Lin MH, Liou SH (2014). Sneezing and allergic dermatitis were increased in 8800 engineered nanomaterial handling workers. Industrial Health, 52(3), 199. 8801 8802 D Lieber M, Smith B, Szakal A, Nelson-Rees W, Todaro G: A continuous tumor-cell line from a human 8803 lung carcinoma with properties of type II alveolar epithelial cells. Int J Cancer 1976, 17: 62-70 8804 8805 Linkov I, Welle P, Loney D, Tkachuk A, Canis L, Kim JB, Bridges T.2011. Use of multicriteria decision 8806 analysis to support weight of evidence evaluation. Risk Analy 31:1211–25. 8807 8808 Linkov I, Massey O, Keisler J, Rusyn I, Hartung T. 2015. From “"weight of evidence" to 8809 quantitative data integration using multicriteria decision analysis and Bayesian methods”. Altex 32:3– 8810 8 8811 8812 Linkov P, et.al., Nanoscale, 2013, 5, 8781 8813 8814 Linkov I, Massey O, Keisler J, Rusyn I, Hartung T (2015) From “weight of evidence” to quantitative 8815 data integration using multicriteria decision analysis and Bayesian methods. ALTEX 32: 3-8. 8816

281

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8817 Linnainmaa K, Kivipensas P, Vainio H. 1997.Toxicity and cytogenetic studies of ultrafine titanium 8818 dioxide in cultured rat liver epithelial cells. Toxicol In Vitro 11:329-35. 8819 Liu R, Zhang X, Pu Y, et al. Small-sized titanium dioxide nanoparticles mediate immune toxicity in rat 8820 pulmonary alveolar macrophages in vivo. J Nanoscale Nanotechnol 2010;10: 5161-5169. 8821 8822 Liou SH, Tsou TC, Wang SL, Li LA, Chiang HC, Li WF, Lin PP, Lai CH, Lee HL, Lin MH, Hsu JH, Chen CR, 8823 Shih TS, Liao HY, Chung YT (2012) Epidemiological study of health hazards among workers handling 8824 engineered nanomaterials. J Nanopart Res 14, 878–92. 8825 8826 Lipka J, Semmler-Behnke M, Sperling RA, Wenk A, Takenaka S, Schleh C, Kissel T, Parak WJ, Kreyling 8827 WG. (2010). Biodistribution of PEG-modified gold nanoparticles following intratracheal instillation 8828 and intravenous injection. Biomaterials 31, 6574-6581. 8829 8830 Liu J, Wang Z, Liu FD, Kane AB, Hurt RH. Chemical transformations of nanosilver in biological 8831 environments. ACS Nano 2012 6(11):9887-99. 8832 8833 Liu R, Zhang X, Pu Y, et al. Small-sized titanium dioxide nanoparticles mediate immune toxicity in rat 8834 pulmonary alveolar macrophages in vivo. J Nanoscale Nanotechnol 2010;10: 5161-5169. 8835 8836 Liu Q, Shao X, Chen J, Shen Y, Feng C, Gao X, Zhao Y, Li J, Zhang Q, Jiang X. In vivo toxicity and 8837 immunogenicity of wheat germ agglutinin conjugated poly(ethylene glycol)-poly(lactic acid) 8838 nanoparticles for intranasal delivery to the brain. Toxicol Appl Pharmacol. 2011;251(1):79-84. 8839 8840 Loden M, Beitner H, Gonzalez H, Edstrom DW, Akerstrom U, Austad J et al.: Sunscreen use: 8841 controversies, challenges and regulatory aspects. Br J Dermatol 2011, 165: 255-262 8842 Nohynek GJ, Lademann J, Ribaud C, Roberts MS: Grey goo on the skin? Nanotechnology, cosmetic 8843 and sunscreen safety. Crit Rev Toxicol 2007, 37: 251-277 8844 8845 Sara A. Love, Melissa A. Maurer-Jones,* John W. Thompson,* Yu-Shen Lin,* and Christy L. Haynes. 8846 Assessing Nanoparticle Toxicity. Annual Review of Analytical Chemistry 5:181-205 8847 8848 Løvik M, Høgseth AK, Gaarder PI, Hagemann R, Eide I. Diesel exhaust particles and carbon black have 8849 adjuvant activity on the local lymph node response and systemic IgE production to ovalbumin. 8850 Toxicology 1997;121:165–178. 8851

282

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8852 Low-Kam C, Telesca D, Ji Z, Zhang H, Xia T, Zink JI, Nel AE. 2015. A Bayesian regression tree approach 8853 to identify the effect of nanoparticles' properties on toxicity profiles. Ann Appl Stat 9:383–401. 8854 8855 Luanpitpong S, Wang L, Castranova V, Rojanasakul Y. Induction of stem-like cells with malignant 8856 properties by chronic exposure of human lung epithelial cells to single-walled carbon nanotubes. Part 8857 Fibre Toxicol. 2014 May 11;11:22. 8858 8859 Lundborg M, Johard U, Låstbom L, Gerde P, Camner P. Human alveolar macrophage phagocytic 8860 function is impaired by aggregates of ultrafine carbon particles. Environ Res. 2001 Jul;86(3):244-53. 8861 8862 M. Lundqvist, J. Stigler, G. Elia, I. Lynch, T. Cedervall, K.A. Dawson, P Natl Acad Sci Usa 105 (2008) 8863 14265 8864 8865 O. Lunov, T. Syrovets, C. Loos, G.U. Nienhaus, V. Mailänder, K. Landfester, M. Rouis, T. Simmet. 8866 Amino-functionalized polystyrene nanoparticles activate the NLRP3 inflammasome in human 8867 macrophages. ACS Nano 5 (2011) 9648-9657. 8868 8869 Ma JY, Zhao H, Mercer RR, et al. Cerium oxide nanoparticle-induced pulmonary inflammation and 8870 alveolar macrophage functional change in rats. Nanotoxicology 2011;5(3):312-25. 8871 8872 MacNicoll A, Kelly M, Aksoy H, Kramer E, Bouwmeester H, Chaudhry Q. A study of the uptake and 8873 biodistribution of nano-titanium dioxide using in vitro and in vivo models of oral intake. J Nanopart 8874 Res. 2015 17:66 8875 8876 MacParland SA, Tsoi KM, Ouyang B, Ma XZ, Manuel J, Fawaz A, Ostrowski MA, Alman BA, Zilman A, 8877 Chan WC, McGilvray ID. Phenotype determines nanoparticle uptake by human macrophages from 8878 liver and blood. ACS Nano. 2017 Mar 28;11(3):2428-2443. 8879 8880 Mangum JB, Turpin EA, Antao-Menezes A, et al. Single-walled carbon nanotube (SWCNT)-induced 8881 interstitial fibrosis in the lungs of rats is associated with increased levels of PDGF mRNA and the 8882 formation of unique intercellular carbon structures that bridge alveolar macrophages in situ. Part 8883 Fibre Toxicol. 2006;3:15. 8884 8885 Ma-Hock L, Treumann S, Strauss V, et al. Inhalation toxicity of multiwall carbon nanotubes in rats 8886 exposed for 3 months.Toxicol Sci. 2009;112(2):468-81.

283

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8887 8888 Ma-Hock et al.(2009), 8889 8890 Ma-Hock L, Burkhardt S, Strauss V, Gamer AO, Wiench K, Van Ravenzwaay B, Landsiedel R. 8891 Development of a Short-Term Inhalation Test in the Rat Using Nano-Titanium Dioxide as a Model 8892 Substance. Inhalation Toxicol 21, 102-118, 2009. 8893 8894 Malur A, Barna BP, Patel J, McPeek M, Wingard CJ, Dobbs L, Thomassen MJ. Exposure to a 8895 Mycobacterial Antigen, ESAT-6, Exacerbates Granulomatous and Fibrotic Changes in a Multiwall 8896 Carbon Nanotube Model of Chronic Pulmonary Disease. J Nanomed Nanotechnol. 2015, 6(6). 8897 8898 Mamaeva V, Rosenholm JM, Bate-Eya LT, Bergman L, Peuhu E, Duchanoy A, Fortelius LE, Landor S, 8899 Toivola DM, Lindén M, Sahlgren C. Mesoporous silica nanoparticles as drug delivery systems for 8900 targeted inhibition of Notch signaling in cancer. Mol Ther. 2011;19(8):1538-46. 8901 8902 Mangum JB, Turpin EA, Antao-Menezes A, et al. Single-walled carbon nanotube (SWCNT)-induced 8903 interstitial fibrosis in the lungs of rats is associated with increased levels of PDGF mRNA and the 8904 formation of unique intercellular carbon structures that bridge alveolar macrophages in situ. Part 8905 Fibre Toxicol. 2006;3: 15. 8906 8907 Mangum JB, Everitt JI, Bonner JC, Moore LR, Brody AR: Co-culture of primary pulmonary cells to 8908 model alveolar injury and translocation of proteins. In Vitro Cell Dev Biol 1990, 26: 1135-1143 8909 8910 Marks R.. Contact dermatitis due to silver. Br J Dermatol 78, 606–607, 1966 8911 8912 Martinon F, Mayor A, Tschopp J. The inflammasomes: guardians of the body. Annu Rev Immunol. 8913 2009;27:229-265. 8914 8915 Martinez-Argudo I, Sands C, Jepson MA. (2007) Translocation of enteropathogenic Escherichia coli 8916 across an in vitro M cell model is regulated by its type III secretion system. Cellular Microbiology, 9, 8917 1538-1546. 8918 8919 Martirosyan, A., and Schneider, Y. J. (2014). Engineered nanomaterials in food: implications for food 8920 safety and consumer health. International journal of environmental research and public health 11(6), 8921 5720-50.

284

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8922 8923 Hans J. P. Marvin, Yamine Bouzembrak, Esmée M. Janssen, Meike van der Zande, Finbarr Murphy, 8924 Barry Sheehan, Martin Mullins & Hans Bouwmeester (2017) Application of Bayesian networks for 8925 hazard ranking of nanomaterials to support human health risk assessment, Nanotoxicology, 11:1, 8926 123-133, DOI: 10.1080/17435390.2016.1278481 8927 8928 8929 Marzaioli V, Aguilar-Pimentel JA, Weichenmeier I, Luxenhofer G, Wiemann M, Landsiedel R, 8930 Wohlleben W, Eiden S, Mempel M, Behrendt H, Schmidt-Weber C, Gutermuth J, Alessandrini F. 8931 Surface modifications of silica nanoparticles are crucial for their inert versus proinflammatory and 8932 immunomodulatory properties. Int J Nanomedicine. 2014, 9:2815-2832. 8933 8934 Masopust D., Schenkel J.M. The integration of T cell migration, differentiation and function. Nat Rev 8935 Immunol 2013. 13 309-320 8936 8937 Mathew, R., et al., Autophagy suppresses tumorigenesis through elimination of p62. Cell, 2009. 8938 137(6): p. 1062-75. 8939 8940 Mattei TA, Rehman AA. Technological developments and future perspectives on graphene-based 8941 metamaterials: a primer for neurosurgeons. Neurosurgery. 2014;74(5):499-516; discussion 516. 8942 8943 Maynard et al., 2006, Natl. Nanotechnol 8944 8945 Maynard AD. The (nano) enterpreneur’s dilemma. Nature Nanotechnology 10, 199-200, 2015. 8946 8947 McAllister M. (2010). Dynamic dissolution: a step closer to predictive dissolution testing? Mol Pharm 8948 7:1374–1387. 8949 8950 MacKay C, Davies M, Summerfield V, Maxwell G. Altex 30, 4/13 8951 8952 McKinney et al(2009) 8953 8954 McNamee et al., Regulatory Toxicology and Pharmacology 52 (2008) 24–34. A review of critical 8955 factors in the conduct and interpretation of the human repeat insult patch test 8956

285

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8957 R. Medzhitov, Origin and physiological roles of inflammation, Nature, 454 (2008) 428-435. 8958 8959 Mei BC, Susumu K, Medintz IL, Mattoussi H. Polyethylene glycol-based bidentate ligands to enhance 8960 quantum dot and gold nanoparticle stability in biological media. Nat Protoc. 2009;4(3):412-23. 8961 8962 Mercer RR, Hubbs AF, Scabilloni JF, et al. Distribution and persistence of pleural penetrations by 8963 multi-walled carbon nanotubes. Part Fibre Toxicol. 2010;7:28. 8964 8965 Mercer RR, Scabilloni JF, Hubbs AF, Wang L, Battelli LA, McKinney W, Castranova V, Porter DW. 8966 Extrapulmonary transport of MWCNT following inhalation exposure. Part Fibre Toxicol. 2013;10:38. 8967 8968 Merrifield DL, Shaw BJ, Harper GM, Saoud IP, Davies SJ, Handy RD, Henry TB. Ingestion of metal- 8969 nanoparticle contaminated food disrupts endogenous microbiota in zebrafish (Danio rerio). Environ 8970 Pollut 2013; 174:157-63. 8971 8972 Meunier E, Coste A, Olagnier D, et al. Double-walled carbon nanotubes trigger IL-1β release in human 8973 monocytes through Nlrp3 inflammasome activation. Nanomedicine 2012;8(6):987-95. 8974 8975 Mills CD, Kincaid K, Alt JM, Heilman MJ, Hill AM. M-1/M-2 macrophages and the Th1/Th2 paradigm. J 8976 Immunol 2000;164:6166-73. 8977 8978 Min WIREs review 8979 8980 Miquel-Jeanjean, C., et al., Penetration study of formulated nanosized titanium dioxide in models of 8981 damaged and sun-irradiated skins. Photochem Photobiol, 2012. 88(6): p. 1513-21. 8982 8983 Mima Y, Hashimoto Y, Shimizu T, Kiwada H, Ishida T. Anti-PEG IgM Is a major contributor to the 8984 accelerated blood clearance of polyethylene glycol-conjugated protein. Mol Pharm. 2015 Jul 8985 6;12(7):2429-35 8986 8987 Mishra N, Tiwari S, Vaidya B, Agrawal GP, Vyas SP. Lectin anchored PLGA nanoparticles for oral 8988 mucosal immunization against hepatitis B. J Drug Target. 2011;19:67-78. 8989 8990 Mitchell LA, Gao J, Wal RV, Gigliotti A, Burchiel SW, McDonald JD. Pulmonary and systemic immune 8991 response to inhaled multiwalled carbon nanotubes. Toxicol Sci 2007, 100: 203-214.

286

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

8992 8993 Mitchell LA, Lauer FT, Burchiel SW, McDonald JD. Mechanisms for how inhaled multiwalled carbon 8994 nanotubes suppress systemic immune function in mice. Nature Nanotech 2009, 4(7):451-456. 8995 8996 Mitrano DM, Motellier S, Clavaguera S, Nowack B. Review of nanomaterial aging and transformations 8997 through the life cycle of nano-enhanced products. Environ Int. 2015;77:132-47. 8998 8999 Miyawaki, J., et al., Toxicity of single-walled carbon nanohorns. ACS Nano, 2008. 2(2): p. 213- 9000 9001 Moffat I, Chepelev NL, Labib S, Bourdon-Lacombe J, Kuo B, Buick JK, Lemieux F, Williams A, 9002 Halappanavar S, Malik AI, Luijten M, Aubrecht J, Hyduke DR, Fornace AJ Jr, Swartz CD, Recio L, Yauk 9003 CL. Comparison of toxicogenomics and traditional approaches to inform mode of action and points of 9004 departure in human health risk assessment of benzo[a]pyrene in drinking water. Crit Rev Toxicol. 9005 2015, 45:1-43. 9006 9007 Moghimi SM, Andersen AJ, Ahmadvand D, Wibroe PP, Andresen TL, Hunter AC. Material properties in 9008 complement activation. Adv Drug Deliv Rev. 2011 Sep 16;63(12):1000-7. 9009 9010 Monteiro-Riviere, N. and Tran, L. 2007) 9011 9012 Moulari B, Pertuit D, Pellequer Y, Lamprecht A. The targeting of surface modified silica nanoparticles 9013 to inflamed tissue in experimental colitis. Biomaterials. 2008;29(34):4554-60. 9014 9015 Mohamed AA, Salmi Z, Dahoumane SA, Mekki A, Carbonnier B, Chehimi MM.Functionalization of 9016 nanomaterials with aryldiazonium salts. Adv Colloid Interface Sci. 2015, in press. pii: S0001- 9017 8686(15)00125-6. 9018 9019 Monteiro-Riviere NA, Larese Filon F. (2012). Skin. In: Adverse effects of engineered nanomaterials, 9020 1st Edition. Exposure, Toxicology, and impact on human health. Eds: Fadeel B, Pietroiusti A, Shvedova 9021 AA. Elsevier, Amsterdam, London, New York. 9022 9023 Monteiro-Riviere NA. Riviere JE. (2009b) Interaction of nanomaterials with skin: Aspects of 9024 absorption and biodistribution, Nanotoxicology, 3: 3, 188 — 193. 9025

287

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9026 Monopoli MP, Åberg C, Salvati A, Dawson KA. Biomolecular coronas provide the biological identity of 9027 nanosized materials. Nat Nanotechnol. 7, 779-786, 2012 9028 9029 Monopoli MP, Aberg C, Salvati A, Dawson KA. Biomolecular coronas provide the biological identity of 9030 nanosized materials. Nat Nanotechnol. 7, 779-786, 2012 9031 9032 Moore TL, Rodriguez-Lorenzo L, Hirsch V, Balog S, Urban D, Jud C, Rothen-Rutishauser B, Lattuada M, 9033 Petri-Fink A. Nanoparticle colloidal stability in cell culture media and impact on cellular interactions. 9034 Chem Soc Rev. 44(17):6287-305 (2015). 9035 9036 Mortensen, L.J., et al., In vivo skin penetration of quantum dot nanoparticles in the murine model: 9037 the effect of UVR. Nano Lett, 2008. 8(9): p. 2779-87. 9038 9039 Morimoto et al.(2012) 9040 9041 Inès Mottas, Ana Milosevic, Alke Petri-Fink, Barbara Rothen-Rutishauser and Carole Bourquin. A 9042 rapid screening method to evaluate the impact of nanoparticles on macrophages. Nanoscale, 2017,9, 9043 2492-2504 9044 9045 Moyano DF, Goldsmith M, Solfiell DJ, Landesman-Milo D, Miranda OR, Peer D, Rotello VM. 9046 Nanoparticle hydrophobicity dictates immune response. J Am Chem Soc. 2012 Mar 7;134(9):3965-7. 9047 9048 Mukherjee SP, Lozano N, Kucki M, Del Rio-Castillo AE, Newman L, Vázquez E, Kostarelos K, Wick P, 9049 Fadeel B. Detection of Endotoxin Contamination of graphene based materials using the TNF-α 9050 Expression Test and guidelines for endotoxin-free graphene oxide production. PLoS One. 2016 Nov 9051 23;11(11):e0166816. 9052 9053 Mukherjee SP, Lyng FM, Garcia A, Davoren M, Byrne HJ. Mechanistic studies of in vitro cytotoxicity of 9054 poly(amidoamine) dendrimers in mammalian cells. Toxicol Appl Pharmacol. 2010;248(3):259-68. 9055 9056 Muller L, Brighton LE, Jaspers I: Ozone exposed epithelial cells modify cocultured natural killer cells. 9057 Am J Physiol Lung Cell Mol Physiol 2013, 304: L332-L341. 9058 9059 Muller J, Huaux F, Moreau N, Misson P, Heilier JF, Delos M, Arras M, Fonseca A, Nagy JB, Lison D. 9060 Respiratory toxicity of multi-wall carbon nanotubes. Toxicol Appl Pharmacol 2005;207:221-231.

288

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9061 9062 Muller L, Riediker M, Wick P, Mohr M, Gehr P, Rothen-Rutishauser B: Oxidative stress and 9063 inflammation response after nanoparticle exposure: differences between human lung cell 9064 monocultures and an advanced three-dimensional model of the human epithelial airways. J R Soc 9065 Interface 2009 9066 9067 Muller J, Delos M, Panin N, Rabolli V, Huaux F, Lison D. Absence of carcinogenic response to multiwall 9068 carbon nanotubes in a 2-year bioassay in the peritoneal cavity of the rat. Toxicol Sci. 110, 442-448, 9069 2009. 9070 9071 Munger MA, Radwanski P, Hadlock GC, Stoddard G, Shaaban A, Falconer J, Grainger DW, Deering- 9072 Rice CE. In vivo human time-exposure study of orally dosed commercial silver nanoparticles. 9073 Nanomedicine. 2014;10:1-9. 9074 9075 Murray AR, Kisin ER, Tkach AV, et al. Factoring-in agglomeration of carbon nanotubes and nanofibers 9076 for better prediction of their toxicity versus asbestos. Part Fibre Toxicol. 2012;9:10. 9077 9078 Murphy FA, Poland CA, Duffin R, Al-Jamal KT, Ali-Boucetta H, Nunes A, Byrne F, Prina-Mello A, Volkov 9079 Y, Li S, Mather SJ, Bianco A, Prato M, Macnee W, Wallace WA, Kostarelos K, Donaldson K. Am J 9080 Pathol. 2011 Jun;178(6):2587-600. 9081 9082 Murphy FA, Schinwald A, Poland CA, et al. The mechanism of pleural inflammation by long carbon 9083 nanotubes: interaction of long fibres with macrophages stimulates them to amplify pro-inflammatory 9084 responses in mesothelial cells. Part Fibre Toxicol. 2012; 9:8. 9085 9086 Muir et al.(1987) Sayes et al.(2010) 9087 9088 Murphy FA, Schinwald A, Poland CA, Donaldson K. The mechanism of pleural inflammation by long 9089 carbon nanotubes: interaction of long fibres with macrophages stimulates them to amplify pro- 9090 inflammatory responses in mesothelial cells. Part Fibre Toxicol. 2012 Apr 3;9:8. 9091 9092 Murray PJ, Allen JE, Biswas SK, Fisher EA, Gilroy DW, Goerdt S, Gordon S, Hamilton JA, Ivashkiv LB, 9093 Lawrence T, Locati M, Mantovani A, Martinez FO, Mege JL, Mosser DM, Natoli G, Saeij JP, Schultze JL, 9094 Shirey KA, Sica A, Suttles J, Udalova I, van Ginderachter JA, Vogel SN, Wynn TA. Macrophage

289

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9095 activation and polarization: nomenclature and experimental guidelines. Immunity. 2014 Jul 9096 17;41(1):14-20. 9097 9098 Murray AR, Kisin ER, Tkach AV, Yanamala N, Mercer R, Young SH, Fadeel B, Kagan VE, Shvedova AA. 9099 Factoring-in agglomeration of carbon nanotubes and nanofibers for better prediction of their toxicity 9100 versus asbestos. Part Fibre Toxicol. 2012 Apr 10;9:10. 9101 9102 Mutlu GM, Budinger GR, Green AA, Urich D, Soberanes S, Chiarella SE, Alheid GF, McCrimmon DR, 9103 Szleifer I, Hersam MC. Biocompatible nanoscale dispersion of single-walled carbon nanotubes 9104 minimizes in vivo pulmonary toxicity. Nano Lett. 2010 May 12;10(5):1664-70. 9105 9106 Nair AV, Keliher EJ, Core AB, Brown D, Weissleder R. Characterizing the interactions of organic 9107 nanoparticles with renal epithelial cells in vivo. ACS Nano. 2015;9(4):3641-53. 9108 9109 C. Nathan, Points of control in inflammation, Nature, 420 (2002) 846-852. 9110 9111 National Research Council. 2007. (2007). Toxicity Testing in the 21st Century: A Vision and a 9112 Strategy. Washington, DC. 9113 9114 National Academy of Sciences (NAS). 2012. A Research Strategy for Environmental, Health, and 9115 Safety Aspects of Engineered Nanomaterials. National Research Council of the National Academies. 9116 1-211 9117 9118 National Research Council (NRC). 1983. Risk Assessment in the Federal Government: Managing the 9119 Process. Committee on the Institutional Means for Assessment of Risks to Public Health, National 9120 Research Council. Available at http://www.nap.edu/openbook.php?isbn=0309033497 9121 9122 Nature Nanotechnology Editorial. 2012. Join the dialogue. Nature Nanotech 7:545 9123 Nel AE, Mädler L, Velegol D, Xia T, Hoek EM, Somasundaran P, Klaessig F, Castranova V, Thompson M. 9124 2009. Understanding biophysicochemical interactions at the nano-bio interface. Nat Mater 7:543-57. 9125 9126 Nauseef WM, Borregaard N. Neutrophils at work. Nat Immunol. 2014 Jul;15(7):602-11. 9127 9128 ANDRE NEL, TIAN XIA, HUAN MENG, XIANG WANG, SIJIE LIN, ZHAOXIA JI, AND HAIYUAN ZHANG,Vol. 9129 46, No. 3 ’ 2013 ’ 607–621 ’ ACCOUNTS OF CHEMICAL RESEARCH

290

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9130 9131 Nel A, Xia T, Mädler L, Li N. Toxic potential of materials at the nanolevel. Science. 2006 Feb 9132 3;311(5761):622-7. 9133 9134 9135 Nel AE, Mädler L, Velegol D, Xia T, Hoek EM, Somasundaran P, Klaessig F, Castranova V, Thompson M. 9136 Understanding biophysicochemical interactions at the nano-bio interface. 9137 Nat Mater. 2009 Jul;8(7):543-57. 9138 9139 Nel AE, Mädler L, Velegol D, Xia T, Hoek EM, Somasundaran P, Klaessig F, Castranova V, Thompson M. 9140 Understanding biophysicochemical interactions at the nano-bio interface. Nat Mater. 2009 9141 Jul;8(7):543-57. 9142 9143 Nembrini C, Stano A, Dane KY, et al. Nanoparticle conjugation of antigen enhances cytotoxic T-cell 9144 responses in pulmonary vaccination. Proc. Natl. Acad. Sci. U. S. A. 108(44), E989–97 (2011). 9145 9146 Nguyen TH, Lin M, Mustapha A. Toxicity of graphene oxide on intestinal bacteria and Caco-2 cells. J 9147 Food Prot. 2015 May;78(5):996-1002. 9148 9149 J.W. Nichols, Y.H. Bae, Nano Today 7 (2012) 606. 9150 9151 Nikota J, Williams A, Yauk CL, Wallin H, Vogel U, Halappanavar S. Meta-analysis of transcriptomic 9152 responses as a means to identify pulmonary disease outcomes for engineered nanomaterials. Part 9153 Fibre Toxicol. 2016, 13, 25. doi: 10.1186/s12989-016-0137-5. 9154 9155 Nikota J, Williams A, Yauk CL, Wallin H, Vogel U, Halappanavar S . Meta-analysis of transcriptomic 9156 responses as a means to identify pulmonary disease outcomes for engineered nanomaterials. Part 9157 Fibre Toxicol. 2016, 11;13:25. doi: 10.1186/s12989-016-0137-5 9158 9159 Nikota J, Williams A, Yauk CL, Wallin H, Vogel U, Halappanavar S. Meta-analysis of transcriptomic 9160 responses as a means to identify pulmonary disease outcomes for engineered nanomaterials. Part 9161 Fibre Toxicol. 2016, 13(1):25. doi: 10.1186/s12989-016-0137-5. 9162 9163 NIOSH: Occupational exposure to carbon nanotubes and nanofibers: current intelligence bulletin 65, 9164 2013

291

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9165 9166 NIOSH Current Intelligence Bulletin 65. Occupational exposure to carbon nanotubes and nanofibers. 9167 2013. 9168 9169 Nishanth RP, Jyotsna RG, Schlager JJ, et al. Inflammatory responses of RAW 264.7 macrophages upon 9170 exposure to nanoparticles: role of ROS-NFκB signaling pathway. Nanotoxicology. 2011;5(4):502-16. 9171 9172 Nickel U, zu Castell A, Pöppl K, Schneider S. 2000. A Silver Colloid Produced by Reduction with 9173 Hydrazine as Support for Highly Sensitive Surface-Enhanced Raman Spectroscopy. Langmuir 16 9087- 9174 91 9175 9176 Niidome T, Yamagata M, Okamoto Y et al. PEG-modified gold nanorods with a stealth character for in 9177 vivo application. J. Control. Release 114, 343–347, 2006. 9178 9179 Niidome T, Yamagata M, Okamoto Y, Akiyama Y, Takahashi H, Kawano T, Katayama Y, Niidome Y. 9180 (2006). PEG-modified gold nanorods with a stealth character for in vivo application. J. Control. 9181 Release 114, 343–347. 9182 9183 No JW, Kim DH, Lee MJ, Kim DH, Kim TH, Sung AY. Preparation and Characterization of Ophthalmic 9184 Lens Materials Containing Titanium Silicon Oxide and Silver Nanoparticles. J Nanosci Nanotechnol. 9185 2015 Oct;15(10):8016-22. 9186 9187 Nohynek, G.J., E.K. Dufour, and M.S. Roberts, Nanotechnology, cosmetics and the skin: is there a 9188 health risk? Skin pharmacology and physiology, 2008. 21(3): p. 136-49 9189 9190 Nohynek, G.J., et al., Grey goo on the skin? Nanotechnology, cosmetic and sunscreen safety. Critical 9191 reviews in toxicology, 2007. 37(3): p. 251-77 9192 9193 Nowack B, Krug HF, Height M. 120 years of nanosilver history: implications for policy makers. Environ 9194 Sc Technol 45, 1177-1183, 2011 9195 9196 Nygaard UC, Hansen JS, SamuelsenM, Alberg T, Marioara CD, Løvik M. Single-walled andmulti-walled 9197 carbon nanotubes promote allergic immune responses inmice. Toxicol Sci 2009;109:113–123. 9198

292

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9199 Oberdorster G, Ferin J, Lehnert BE. 1994. Correlation between particle size, in vivo particle 9200 persistence, and lung injury. Environ Health Perspect 102 Suppl 5:173-9 9201 9202 Oberdörster G, Sharp Z, Atudorei V, Elder A, Gelein R, Kreyling W, Cox C. 2004. Translocation of 9203 inhaled ultrafine particles to the brain. Inhal. Toxicol 16:437-45 9204 9205 Oberdorster, G; Oberdorster, E; Oberdorster, J. Nanotoxicology: An emerging discipline evolving 9206 from studies of ultrafine particles. Environmental health perspectives. 113, 823 -839, 2005; 9207 9208 Oberdörster G, Stone V, Donaldson K. Toxicology of nanoparticles: A historical perspective. 9209 Nanotoxicology 1, 2-25, 2007 9210 9211 Oberdörster, et al. Translocation of Inhaled Ultrafine Particles to the Brain G. Inhalation toxicology 9212 Volume 16, Issue 6-7, 2004 9213 9214 Günter Oberdörster, Vincent Castranova, Bahman Asgharian & Phil Sayre (2015) Inhalation Exposure 9215 to Carbon Nanotubes (CNT) and Carbon Nanofibers (CNF): Methodology and Dosimetry, Journal of 9216 Toxicology and Environmental Health, Part B, 18:3-4, 121-212, DOI: 9217 10.1080/10937404.2015.1051611 9218 9219 Oberdorster, G.; Stone, V.; Donaldson, K. Toxicology of nanoparticles: A historical perspective. 9220 Nanotoxicology 2007, 1, 2-25. 9221 9222 Oberdorster G, Ferin J,Lehnert BE Correlation between Particle Size, In Vivo 9223 Particle Persistence, and Lung Injury. Environ Health Perspect 102(Suppl 5):173-179, 1994. 9224 9225 Oberdorster H, et al., Particle and Fibre Toxicology20052:8 9226 9227 OECD TG for inhalation exposures of particles 9228 9229 OECD 407, 412, ENV/JM/MONO(2012)14 9230 9231 OECD 427: Skin Absorption: In vivo Method. OECD, Paris, France (2004). 9232

293

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9233 OECD Regulatory Frameworks for Nanotechnology in Foods and Medical Products. Summary Results 9234 of a Survey Activity. Paris, 2013. From: 9235 http://search.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=DSTI/STP/NANO(2012) 9236 9237 Oh WK, Kim S, Choi M, Kim C, Jeong YS, Cho BR, Hahn JS, Jang J. Cellular uptake, cytotoxicity, and 9238 innate immune response of silica-titania hollow nanoparticles based on size and surface 9239 functionality. ACS Nano. 2010 Sep 28;4(9):5301-13. 9240 9241 Omlor AJ, Nguyen J, Bals R, Dinh QT. Nanotechnology in respiratory medicine. Respir Res. 9242 2015;16:64. 9243 9244 O'Neill C, Pan Q, Clarke G, Liu F, Hodges G, Ge M, Jordan P, Chang U, Newman R, Toulson E. Silica 9245 fragments from millet bran in mucosa surrounding oesophageal tumours in patients in northern 9246 China. Lancet. 1982;1:1202-6. 9247 9248 O'Neill CH, Hodges GM, Riddle PN, Jordan PW, Newman RH, Flood RJ, Toulson EC. A fine fibrous silica 9249 contaminant of flour in the high oesophageal cancer area of north-east Iran. Int J Cancer. 9250 1980;26(5):617-28. 9251 9252 Oomen AG, Rompelberg CJ, Bruil MA, Dobbe CJ, Pereboom DP, Sips AJ. (2003). Development of an in 9253 vitro digestion model for estimating the bioaccessibility of soil contaminants. Arch Environ Contam 9254 Toxicol. 44, 281-287. 9255 9256 Orta-García ST, Plascencia-Villa G, Ochoa-Martínez AC, Ruiz-Vera T, Pérez-Vázquez FJ, Velázquez- 9257 Salazar JJ, Yacamán MJ, Navarro-Contreras HR, Pérez-Maldonado IN. Analysis of cytotoxic effects of 9258 silver nanoclusters on human peripheral blood mononuclear cells 'in vitro'. J Appl Toxicol. 35, 1189- 9259 1199, 2015. 9260 Osmond-McLeod MJ, Osmond RI, Oytam Y, McCall MJ, Feltis B, Mackay-Sim A, Wood SA, Cook AL. 9261 Surface coatings of ZnO nanoparticles mitigate differentially a host of transcriptional, protein and 9262 signalling responses in primary human olfactory cells. Part Fibre Toxicol. 2013;10(1):54. 9263 9264 Osmond, M.J. and M.J. McCall, Zinc oxide nanoparticles in modern sunscreens: an analysis of 9265 potential exposure and hazard. Nanotoxicology, 2010. 4(1): p. 15-41. 9266

294

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9267 Osmond-McLeod MJ, Oytam Y, Kirby JK, Gomez-Fernandez L, Baxter B, McCall MJ. (2013). Dermal 9268 absorption and short term biological impact in hairless mice from sunscreens containing zinc oxide 9269 nano- or larger particles. Nanotoxicology 8 (Supplement 1) 72-84, 2014 (DOI: 9270 10.3109/17435390.2013.855832) 9271 9272 Paik SY, Zalk DM, Swuste P (2008) Application of a pilot control banding tool for risk level assessment 9273 and control of nanoparticle exposures. Ann Occup Hyg 52, 419–28. 9274 9275 Palomäki J, Sund J, Vippola M, Kinaret P, Greco D, Savolainen K, Puustinen A, Alenius H. A 9276 secretomics analysis reveals major differences in the macrophage responses towards different types 9277 of carbon nanotubes. Nanotoxicology. 2015;9(6):719-28. 9278 9279 Palomäki J, Välimäki E, Sund J, Vippola M, Clausen PA, Jensen KA, Savolainen K, Matikainen S, Alenius 9280 H. Long, needle-like carbon nanotubes and asbestos activate the NLRP3 inflammasome through a 9281 similar mechanism. ACS Nano. 2011 5:6861-70 9282 9283 Pampaloni F, Florin EL. Microtubule architecture: inspiration for novel carbon nanotube-based 9284 biomimetic materials. Trends Biotechnol. 2008 Jun;26(6):302-10. 9285 9286 Park EJ, Cho WS, Jeong J, Yi J, Choi K, Park K. Pro-inflammatory and potential allergic responses 9287 resulting from B cell activation in mice treated with multi-walled carbon nanotubes by intratracheal 9288 instillation. Toxicology. 2009, 259(3):113-121 9289 9290 Park, E.J., Bae, E., Yi, J., Kim, Y., Choi, K., Lee, S.H., Yoon, J., Lee, B.C., and Park, K. (2010a). Repeated- 9291 dose toxicity and inflammatory responses in mice by oral administration of silver nanoparticles. 9292 Environ Toxicol Pharmacol 30, 162–168. 9293 9294 Park EJ, Yi J, Kim Y, Choi K, Park K. (2010b). Silver nanoparticles induce cytotoxicity by a Trojan-horse 9295 type mechanism. Toxicol In vitro 24, 872-878. 9296 9297 Park S, Vosguerichian M, Bao Z. A review of fabrication and applications of carbon nanotube film- 9298 based flexible electronics. Nanoscale. 2013;5(5):1727-52. 9299

295

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9300 Park MV, Neigh AM, Vermeulen JP, de la Fonteyne LJ, Verharen HW, Briedé JJ, van Loveren H, de 9301 Jong WH. The effect of particle size on the cytotoxicity, inflammation, developmental toxicity and 9302 genotoxicity of silver nanoparticles. Biomaterials. 32, 9810-9817, 2011. 9303 9304 Pasparakis, M., I. Haase, and F.O. Nestle, Mechanisms regulating skin immunity and inflammation. 9305 Nat Rev Immunol, 2014. 14(5): p. 289-301. 9306 9307 Pauluhn, J., and U. Mohr. 1999. Repeated 4-week inhalation exposure of rats: Effect of low-, 9308 intermediate, and high-humidity chamber atmospheres. Experimental and Toxicologic Pathology 9309 51:178–87. 9310 9311 Pauluhn J. (2009). Retrospective analysis of 4-week inhalation studies in rats with focus on fate and 9312 pulmonary toxicity of two nanosized aluminum oxyhydroxides (boehmite) and pigment-grade iron 9313 oxide (magnetite): The key metric of dose is particle mass and not particle surface area. Toxicology 9314 259, 140-148. 9315 9316 Pauluhn, J. (2011). Poorly soluble particulates: Searching for a unifying denominator of nanoparticles 9317 and fine particles for DNEL estimation. Toxicology 279, 176-188 9318 9319 Pauluhn J. Subchronic 13-week inhalation exposure of rats to multiwalled carbon nanotubes: toxic 9320 effects are determined by density of agglomerate structures, not fibrillar structures. Toxicol Sci. 9321 2010;113(1):226-42. 9322 9323 Paur HR, Cassee FR, Teeguarden JG, Fissan H, Diabate S, Aufderheide M et al.: In-vitro cell exposure 9324 studies for the assessment of nanoparticle toxicity in the lung—A dialog between aerosol science and 9325 biology. J Aerosol Sci 2011, 42: 668-692 9326 9327 Park, M. V. D. Z., Lankveld, D. P. K., van Loveren, H., and de Jong, W. (2009). The status of in vitro 9328 toxicity studies in risk assessment of nanomaterials Nanomedicine 4(6), 669-685. 9329 9330 Pelka J, Gehrke H, Rechel A, Kappes M, Hennrich F, Hartinger CG, Marko D. DNA damaging properties 9331 of single walled carbon nanotubes in human colon carcinoma cells. Nanotoxicology. 2013;7:2-20. 9332

296

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9333 Peters R, Kramer E, Oomen AG, Rivera ZE, Oegema G, Tromp PC, Fokkink R, Rietveld A, Marvin HJ, 9334 Weigel S, Peijnenburg AA, Bouwmeester H. Presence of nano-sized silica during in vitro digestion of 9335 foods containing silica as a food additive. ACS Nano. 2012;6(3):2441-51. 9336 9337 Peters RJ, van Bemmel G, Herrera-Rivera Z, Helsper HP, Marvin HJ, Weigel S, Tromp PC, Oomen AG, 9338 Rietveld AG, Bouwmeester H. Characterization of titanium dioxide nanoparticles in food products: 9339 analytical methods to define nanoparticles. 9340 J Agric Food Chem. 62, 6285-6293, 2014. 9341 9342 Robby A. Petros* and Joseph M. DeSimone - NATURE REVIEwS VOLUmE 9 | AUGUST 2010 9343 doi:10.1038/nrd2591 9344 9345 Phalen, R. F. 2009. Inhalation studies, foundations and techniques, 2nd ed. New York, NY: Informa 9346 Healthcare. Biopharm 2010 9347 9348 Piasecka-Zelga, J., et al., Acute dermal toxicity and sensitization studies of novel nano-enhanced UV 9349 absorbers. J Occup Health, 2015. 57(3): p. 275-84 9350 9351 Piccinno F, Gottschalk F, Seeger S, Nowack B. Industrial production quantities and uses of ten 9352 engineered nanomaterials in Europe and the world. J Nanopart Res. 2012;14:1109. 9353 9354 Pietroiusti A, Campagnolo L, Fadeel B. Interactions of engineered nanoparticles with organs 9355 protected by internal biological barriers. Small. 2013 May 27;9(9-10):1557-72. 9356 9357 Pietroiusti A, Magrini A, Campagnolo L. New frontiers in nanotoxicology: Gut 9358 microbiota/microbiome-mediated effects of engineered nanomaterials. Toxicol Appl Pharmacol. 9359 2016 May 15;299:90-5. 9360 9361 Poirier M, Simard JC, Girard D. Silver nanoparticles of 70 nm and 20 nm affect differently the biology 9362 of human neutrophils. J Immunotoxicol.13, 375-385, 2016. 9363 9364 Poland CA, Duffin R, Kinloch I, et al. Carbon nanotubes introduced into the abdominal cavity of mice 9365 show asbestos-like pathogenicity in a pilot study. Nat. Nanotechnol. 2008;3(7), pp. 423-428. 9366

297

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9367 Poland CA, Duffin R, Kinloch I, Maynard A, Wallace WA, Seaton A, Stone V, Brown S, Macnee W, 9368 Donaldson K. 2008. Carbon nanotubes introduced into the abdominal cavity of mice show asbestos- 9369 like pathogenicity in a pilot study. Nature Nanotech 3:423-8 9370 9371 Politana et al. Regulatory Toxicology and Pharmacology 52 (2008) 35–38. The Research Institute for 9372 Fragrance Materials’ human repeated insult patch test protocol 9373 9374 Polk WW et al., Part Fibre Toxicol. 2016 Apr 23;13:20 9375 9376 Porter DW, Hubbs AF, Mercer RR, Wu N, Wolfarth MG, Sriram K, Leonard S, Battelli L, Schwegler- 9377 Berry D, Friend S, Andrew M, Chen BT, Tsuruoka S, Endo M, Castranova V. Mouse pulmonary dose- 9378 and time course-responses induced by exposure to multi-walled carbon nanotubes. Toxicology 2010, 9379 269:136-147. 9380 9381 Porter et al., 2008 9382 9383 Porter et al.(2012) 9384 9385 Poulsen SS, Saber AT, Mortensen A, Szarek J, Wu D, Williams A, Andersen O, Jacobsen NR, Yauk CL, 9386 Wallin H, Halappanavar S, Vogel U. Changes in cholesterol homeostasis and acute phase response 9387 link pulmonary exposure to multi-walled carbon nanotubes to risk of cardiovascular disease. Toxicol 9388 Appl Pharmacol. 2015 Mar 15;283(3):210-22. doi: 10.1016/j.taap.2015.01.011. Epub 2015 Jan 22. 9389 9390 Poulsen SS, Saber AT, Williams A, Andersen O, Købler C, Atluri R, Pozzebon ME, Mucelli SP, Simion M, 9391 Rickerby D, Mortensen A, Jackson P, Kyjovska ZO, Mølhave K, Jacobsen NR, Jensen KA, Yauk CL, 9392 Wallin H, Halappanavar S, Vogel U. MWCNTs of different physicochemical properties cause similar 9393 inflammatory responses, but differences in transcriptional and histological markers of fibrosis in 9394 mouse lungs. Toxicol Appl Pharmacol. 2015 Apr 1;284(1):16-32. doi: 10.1016/j.taap.2014.12.011. 9395 9396 Poulsen SS, Jacobsen NR, Labib S, Wu D, Husain M, Williams A, Bøgelund JP, Andersen O, Købler C, 9397 Mølhave K, Kyjovska ZO, Saber AT, Wallin H, Yauk CL, Vogel U, Halappanavar S. Transcriptomic 9398 analysis reveals novel mechanistic insight into murine biological responses to multi-walled carbon 9399 nanotubes in lungs and cultured lung epithelial cells. PLoS One. 2013, 8(11):e80452. doi: 9400 10.1371/journal.pone.0080452. 9401

298

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9402 Powell JJ, Ainley CC, Harvey RS, Mason IM, Kendall MD, Sankey EA, Dhillon AP, Thompson RP. 9403 Characterisation of inorganic microparticles in pigment cells of human gut associated lymphoid 9404 tissue. Gut. 1996;38:390-5. 9405 9406 Prabhakar N, Näreoja T, von Haartman E, Şen Karaman D, Burikov SA, Dolenko TA, Deguchi T, 9407 Mamaeva V, Hänninen PE, Vlasov II, Shenderova OA, Rosenholm JM. Functionalization of graphene 9408 oxide nanostructures improves photoluminescence and facilitates their use as optical probes in 9409 preclinical imaging. Nanoscale. 2015;7(23):10410-20. 9410 9411 Pratsinis A, Hervella P, Leroux JC, Pratsinis SE, Sotiriou GA. Toxicity of silver nanoparticles in 9412 macrophages. Small. 15, 2576-2584, 2013. 9413 9414 Priester JH, Van De Werfhorst LC, Ge Y, Adeleye AS, Tomar S, Tom LM, Piceno YM, Andersen GL, 9415 Holden PA. Effects of TiO2 and Ag nanoparticles on polyhydroxybutyrate biosynthesis by activated 9416 sludge bacteria. Environ Sci Technol 2014; 48:14712-20. 9417 9418 Project of Emerging Nanotechnologies (PEN). www.nanotechproject.org/ 9419 9420 Qu Y, Ma Q, Deng J, Shen W, Zhang X, He Z, Van Nostrand JD, Zhou J, Zhou J. Responses of microbial 9421 communities to single-walled carbon nanotubes in phenol wastewater treatment systems. Environ 9422 Sci Technol 2015; 49:4627-35. 9423 9424 Radauer-Preiml I, Andosch A, Hawranek T, Luetz-Meindl U, Wiederstein M, Horejs-Hoeck J, Himly M, 9425 Boyles M, Duschl A. Nanoparticle-allergen interactions mediate human allergic responses: protein 9426 corona characterization and cellular responses. Part Fibre Toxicol. 2016 Jan 16;13:3 9427 9428 Raesch SS, Tenzer S, Storck W, Rurainski A, Selzer D, Ruge CA, Perez-Gil J, Schaefer UF, Lehr CM. 9429 Proteomic and lipidomic analysis of nanoparticle corona upon contact with lung surfactant reveals 9430 differences in protein, but not lipid composition. ACS Nano. 2015 Dec 22;9(12):11872-85. 9431 9432 Luna Rahman, Nicklas Raun Jacobsen, Syed Abdul Aziz, Dongmei Wu, Andrew Williams, Carole L 9433 Yauk, Paul White, Hakan Wallin, Ulla Vogel and Sabina Halappanavar. Multi-walled carbon nanotube- 9434 induced genotoxic, inflammatory and pro-fibrotic responses in mice: Investigating the mechanisms of 9435 pulmonary carcinogenesis. Mutation Research – Genetic Toxicology and Environmental Mutagenesis 9436 – 2017 under peer review.

299

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9437 9438 Rai M, Kon K, Ingle A, Duran N, Galdiero S, Galdiero M. (2014). Broad-spectrum bioactivities of silver 9439 nanoparticles: the emerging trends and future prospects. Appl Microbiol Biotechnol. ;98, 1951-1961. 9440 9441 Rai M, Yadav A, Gade A. (2009). Silver nanoparticles as a new generation of antimicrobials, 9442 Biotechnol Adv. 27, 76-83. 9443 9444 Ramadi KB, Mohamed YA, Al-Sbiei A, Almarzooqi S, Bashir G, Al Dhanhani A, Sarawathiamma D, Qadri 9445 S, Yasin J, Nemmar A, Fernandez-Cabezudo MJ, Haik Y, Al-Ramadi BK. Acute systemic exposure to 9446 silver-based nanoparticles induces hepatotoxicity and NLRP3-dependent inflammation. 9447 Nanotoxicology. 2016 Oct;10(8):1061-74. 9448 9449 Ramsey & Maynard et al.(2005) 9450 9451 Rauscher, H., Roebben, G., Boix Sanfeliu, A., Emons H., Gibson N., Koeber R., Linsinger T., Rasmussen 9452 K., Riego Sintes, J., Sokull-Klüttgen, B., Stamm, H. (2015). Towards a Review of the EC 9453 Recommendation for a Definition of the Term “nanomaterial”; Part 3: Scientific-technical evaluation 9454 of options to clarify the definition and to facilitate its implementation. European Commission Joint 9455 Research Centre from: 9456 http://publications.jrc.ec.europa.eu/repository/bitstream/JRC95675/towards%20review%20ec%20re 9457 c%20def%20nanomaterial%20-%20part%203_report_online%20id.pdf 9458 9459 Rauscher, H., Roebben, G., Amenta, V., Boix Sanfeliu, A., Calzolai, L., Emons, H., Gaillard, C., Gibson, 9460 N., Linsinger, T., Mech, A., Quiros Pesudo, L., Rasmussen, K., Riego Sintes, J., Sokull-Klüttgen, B., 9461 Stamm, H. 2014. Towards a Review of the EC Recommendation for a Definition of the Term 9462 “nanomaterial”; Part 1: Compilation of Information Concerning the Experience with the Definition. 9463 European Commission Joint Research Centre from. https://ec.europa.eu/jrc/sites/default/ 9464 files/lbna26567enn.pdf. 9465 9466 Ravichandran, S., L.J. Mortensen, and L.A. Delouise, Quantification of human skin barrier function 9467 and susceptibility to quantum dot skin penetration. Nanotoxicology, 2011. 5: p. 675-86. 9468 9469 Reddel RR, Ke Y, Gerwin BI, McMenamin MG, Lechner JF, Su RT et al.: Transformation of human 9470 bronchial epithelial cells by infection with SV40 or adenovirus-12 SV40 hybrid virus, or transfection 9471 via

300

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9472 9473 Rehman MU, Yoshihisa Y, Miyamoto Y, et al. The anti-inflammatory effects of platinum nanoparticles 9474 on the lipopolysaccharide-induced inflammatory response in RAW264.7 macrophages. Inflamm Res. 9475 2012;61(11):1177-85 9476 9477 Reidy, B., Haase, A., Luch, A., Dawson, K.A., Lynch, I. Mechanisms of silver nanoparticle release, 9478 transformation and toxicity: a critical review of current knowledge and 9479 recommendations for future studies and applications. Dent. Mater. 6 , 2295–2350, 2013. 9480 9481 Ricevuti G, Mazzone A, Fossati G, Mazzucchelli I, Cavigliano PM, Pasotti D, Notario A. Assay of 9482 phagocytic cell functions. Allerg Immunol (Paris). 1993 Feb;25(2):55-66. 9483 9484 Richmond, J.M. and J.E. Harris, Immunology and skin in health and disease. Cold Spring Harb Perspect 9485 Med, 2014. 4(12): p. a015339. 9486 9487 Richter-Reichhelm H-B, Dasenbrock CA, Descotes G, Emmendörfer AC, Ernst HU, Harleman JH, 9488 Hildebrand B, Küttler K, Rühl-Fehlert CI, Schilling K, Schulte AE, Vohr H-W (1995) Validation of a 9489 modified 28-day rat study to evidence effects of test compounds on the immune system. Reg Tox 9490 Pharm 22: 54-56. 9491 9492 Ricklin D, Hajishengallis G, Yang K, Lambris JD. Complement: a key system for immune surveillance 9493 and homeostasis. Nat Immunol. 2010 Sep;11(9):785-97. 9494 9495 Riebeling C, Wiemann M, Schnekenburger J, Kuhlbusch TA, Wohlleben W, Luch A, Haase A. A redox 9496 proteomics approach to investigate the mode of action of nanomaterials. Toxicol Appl Pharmacol. 9497 2016, 299, 24-29. 9498 9499 Rietschel ET, Kirikae T, Schade FU, Mamat U, Schmidt G, Loppnow H, Ulmer AJ, Zähringer U, Seydel U, 9500 Di Padova F, Schreier M, Brade H: Bacterial endotoxin: molecular relationships of structure to activity 9501 and function. FASEB J 1994, 8: 217–225. 9502 9503 Rietscher R, Schröder M, Janke J, Czaplewska J, Gottschaldt M, Scherließ R, Hanefeld A, Schubert US, 9504 Schneider M, Knolle PA, Lehr CM.Antigen delivery via hydrophilic PEG-b-PAGE-b-PLGA nanoparticles 9505 boosts vaccination induced T cell immunity. Eur J Pharm Biopharm. 2016; 102: 20-31. 9506

301

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9507 Ritter D, Knebel JW, Aufderheide M: Exposure of human lung cells to inhalable substances: a novel 9508 test strategy involving clean air exposure periods using whole diluted cigarette mainstream smoke. 9509 Inhal Toxicol 2003, 15: 67-84 9510 9511 Rivera-Gil P, De Aberasturi DJ, Wulf V, Pelaz B, Del Pino P, Zhao Y, De La Fuente J, De Larramendi IR, 9512 Rojo T, Liang XJ, et al. The challenge to relate physico-chemical properties of colloidal nanoparticles 9513 to their cytotoxicity. Acc Chem Res. 2013;46(3):743–749. 9514 9515 RIVM (2015) Assessing health and environmental risks of nanoparticles. National Institute for Public 9516 Health and the Environment, Bilthoven, The Netherlands. 9517 9518 Roberts JC, Bhalgat MK, Zera RT. Preliminary biological evaluation of polyamidoamine (PAMAM) 9519 Starburst dendrimers. J Biomed Mater Res. 1996;30(1):53-65. 9520 9521 Rocks S., Pollard S., Dorey R., Levy L., Harrison P., Handy R. (2008). Comparison of risk assessment 9522 approaches for manufactured nanomaterials. Final report of Defra project CB403, Cranfield 9523 University, Cranfield, UK. Available at: randd.defra.gov.uk/Document.aspx?Document=CB0403_ 9524 7306_ABS.doc. 9525 9526 Roco MC, Mirkin CA, Hersam MC. 2010. Nanotechnology research directions for societal needs in 9527 2020: retrospective and outlook summary. NSF/WTEC (National Science Foundation/World 9528 Technology Evaluation Center) report [summary of the full report published by Springer]. 9529 9530 Roebben, G., Rauscher, H., Amenta, V., Aschberger, K., Boix Sanfeliu, A., Calzolai, L., Emons, H., 9531 Gaillard, C., Gibson, N., Holzwarth, U., Linsinger, T., Rasmussen, K., Riego Sintes, J., Sokull-Klüttgen, 9532 B., Stamm, H., 2014. Towards a Review of the EC Recommendation for a Definition of the Term 9533 “nanomaterial”; Part 2: Assessment of Collected Information Concerning the Experience with the 9534 Definition. European Commission Joint Research Centre from. https://ec.europa.eu/jrc/en/ 9535 publication/eur-scientific-and-technical-research-reports/towards-review-ecrecommendation- 9536 definition-term-nanomaterial-part-2-assessment-collected. 9537 9538 Rooney AA, Luebke RW, Selgrade MJK, Germolec DR (2012) Immunotoxicity and its application in risk 9539 assessment. In: Luch A ed. Clinical and Environmental Toxicology, Experientia Supplementum 101, 9540 Basel, Springer, pp 251-287. DOI 10.1007/978-3-7643-8340-4¬_9 9541

302

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9542 Rosário F, Hoet P, Santos C, Oliveira H. Death and cell cycle progression are differently conditioned by 9543 the AgNP size in osteoblast-like cells. Toxicology. 368-369, 103-115, 2016. 9544 9545 Rossi EM, Pylkkänen L, Koivisto AJ, et al. Inhalation exposure to nanosized and fine TiO2 particles 9546 inhibits features of allergic asthma in a murine model. Part Fibre Toxicol. 2010;7:35. 9547 9548 Rossi EM, Pylkkänen L, Koivisto AJ, Vippola M, Jensen KA, Miettinen M, Sirola K, Nykäsenoja H, 9549 Karisola P, Stjernvall T, Vanhala E, Kiilunen M, Pasanen P, Mäkinen M, Hämeri K, Joutsensaari J, 9550 Tuomi T, Jokiniemi J, Wolff H, Savolainen K, Matikainen S, Alenius H. Airway exposure to silica-coated 9551 TiO2 nanoparticles induces pulmonary neutrophilia in mice. Toxicol Sci. 2010 Feb;113(2):422-33. 9552 9553 Rotating Durm Generator Ref?? 9554 9555 Roth GA, et al.,, Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology 9556 Volume 7, Issue 4, pages 565–579, July/August 2015; 9557 9558 Rothen-Rutishauser B, Blank F, Mühlfeld Ch GP: In vitro models of the human epithelial airway 9559 barrier to study the toxic potential of particulate matter. Exp Opinion Drug Metabolism Toxicol 2008, 9560 4: 1075-1089 9561 9562 Rothenberg, S. J., R. M. Parker, R. G. York, G. E. Dearlove, and M. M. Martin. 2000. Lack of effects of 9563 nose-only inhalation exposure on testicular toxicity in male rats. Toxicological Sciences 53:127–34. 9564 9565 Rouse, J.G., et al., Fullerene-based amino acid nanoparticle interactions with human epidermal 9566 keratinocytes. Toxicol In Vitro, 2006. 20(8): p. 1313-20. 9567 9568 Rouse JG, Yang JZ, Ryman-Rasmussen JP, Barron AR, Monteiro-Riviere NA. Effects of mechanical 9569 flexion on the penetration of fullerene amino acid-derivatized peptide nanoparticles through skin. 9570 Nano Lett. 2007;7(1):155–160. 9571 9572 Rushton EK, et.al., J Toxicol Environ Health A. 2010; 73(5): 10.1080/15287390903489422 9573 9574 Rydman EM, Ilves M, Koivisto AJ, Kinaret PA, Fortino V, Savinko TS, Lehto MT, Pulkkinen V, Vippola 9575 M, Hämeri KJ, Matikainen S, Wolff H, Savolainen KM, Greco D, Alenius H. Inhalation of rod-like

303

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9576 carbon nanotubes causes unconventional allergic airway inflammation. Part Fibre Toxicol. 2014, 9577 16;11:48. 9578 9579 Rydman EM, Ilves M, Koivisto AJ, Kinaret PA, Fortino V, Savinko TS, Lehto MT, Pulkkinen V, Vippola 9580 M, Hämeri KJ, Matikainen S, Wolff H, Savolainen KM, Greco D, Alenius H. Inhalation of rod-like 9581 carbon nanotubes causes unconventional allergic airway inflammation. Part Fibre Toxicol. 2014; 9582 16;11:48. 9583 9584 Ryman-Rasmussen JP, Tewksbury EW, Moss OR, Cesta MF, Wong BA, Bonner JC. Inhaled multiwalled 9585 carbon nanotubes potentiate airway fibrosis in murine allergic asthma. Am J Respir Cell Mol Biol. 9586 2009 Mar;40(3):349-58. 9587 9588 Ryman-Rasmussen, J.P., J.E. Riviere, and N.A. Monteiro-Riviere, Surface coatings determine 9589 cytotoxicity and irritation potential of quantum dot nanoparticles in epidermal keratinocytes. J Invest 9590 Dermatol, 2007. 127(1): p. 143-53 9591 9592 Ryman-Rasmussen JP, Cesta MF, Brody AR, Shipley-Phillips JK, Everitt JI, Tewksbury EW, Moss OR, 9593 Wong BA, Dodd DE, Andersen ME, Bonner JC. Inhaled carbon nanotubes reach the subpleural tissue 9594 in mice. Nat Nanotechnol. 2009 Nov;4(11):747-51. 9595 9596 Ryman-Rasmussen JP, Tewksbury EW, Moss OR, et al. Inhaled multiwalled carbon nanotubes 9597 potentiate airway fibrosis in a murine model of allergic asthma. Am J Resp Cell Mol Biol. 9598 2009b;40(3):349-358. 9599 9600 Saber AT, Jacobsen NR, Jackson P, Poulsen SS, Kyjovska ZO, Halappanavar S, Yauk CL, Wallin H, Vogel 9601 U. Particle-induced pulmonary acute phase response may be the causal link between particle 9602 inhalation and cardiovascular disease. Wiley Interdiscip Rev Nanomed Nanobiotechnol. 2014, 6, 517- 9603 31 9604 9605 Sadrieh N, Wokovich AM, Gopee NV, Zheng J, Haines D, Parmiter D, Siitonen PH, Cozart CR, Patri AK, 9606 McNeil SE, Howard PC, Doub WH, Buhse LF. (2010). Lack of Significant Dermal Penetration of 9607 Titanium Dioxide from Sunscreen Formulations Containing Nano- and Submicron-Size TiO2 Particles 9608 Tox Sc 115, 156–166. 9609

304

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9610 Sadekar S, Ghandehari H. Transepithelial transport and toxicity of PAMAM dendrimers: implications 9611 for oral drug delivery. Adv Drug Deliv Rev. 2012;64(6):571-88. 9612 9613 Sadekar S, Ray A, Janàt-Amsbury M, Peterson CM, Ghandehari H. Comparative biodistribution of 9614 PAMAM dendrimers and HPMA copolymers in ovarian-tumor-bearing mice. Biomacromolecules. 9615 2011;12(1):88-96. 9616 9617 Saha K, Rahimi M, Yazdani M, Kim ST, Moyano DF, Hou S, Das R, Mout R, Rezaee F, Mahmoudi M, 9618 Rotello VM. Regulation of macrophage recognition through the interplay of nanoparticle surface 9619 functionality and protein corona. ACS Nano. 2016 Apr 26;10(4):4421-30. 9620 9621 Sahneh FD, Scoglio CM, Monteiro-Riviere NA, Riviere JE. Predicting the impact of biocorona 9622 formation kinetics on interspecies extrapolations of nanoparticle biodistribution modeling. 9623 Nanomedicine (Lond). 2015;10(1):25-33. 9624 9625 Sakamoto Y, Nakae D, Fukumori N, Tayama K, Maekawa A, Imai K, Hirose A, Nishimura, T, Ohashi N, 9626 Ogata A. Induction of mesothelioma by a single intrascrotal administration of multi-wall carbon 9627 nanotube in intact male Fischer 344 rats. J Toxicol Sci. 34, 65-76, 2009. 9628 9629 Sakthivel T, Toth I, Florence AT. Distribution of a lipidic 2.5 nm diameter dendrimer carrier after oral 9630 administration.Int J Pharm. 1999;183:51-5. 9631 9632 Samet JM, Tal TL. Toxicological disruption of signaling homeostasis: tyrosine phosphatases as targets. 9633 Annu Rev Pharmacol Toxicol. 2010;50:215-35. 9634 9635 Sankar P, Telang AG, Suresh S, Kesavan M, Kannan K, Kalaivanan R, Sarkar SN (2013) 9636 Immunomodulatory effects of nanocurcumin in arsenic-exposed rats. Int Immunopharm 17: 65-70. 9637 9638 Santegoets SJ1, van den Eertwegh AJ, van de Loosdrecht AA, Scheper RJ, de Gruijl TD. Human 9639 dendritic cell line models for DC differentiation and clinical DC vaccination studies. J Leukoc Biol. 9640 2008 Dec;84(6):1364-73. doi: 10.1189/jlb.0208092. Epub 2008 Jul 29. 9641 9642 Sargent LM, Shvedova AA, Hubbs AF, Salisbury JL, Benkovic SA, Kashon ML, Lowry DT, Murray AR, 9643 Kisin ER, Friend S, McKinstry KT, Battelli L, Reynolds SH. Induction of aneuploidy by single-walled 9644 carbon nanotubes. Environ Mol Mutagen. 2009 Oct;50(8):708-17.

305

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9645 9646 Satpathy AT, Wu X, Albring JC, Murphy KM. Re(de)fining the dendritic cell lineage. Nat. Immunol. 9647 13(12), 1145–54 (2012) 9648 9649 Sauer UG, Vogel S, Aumann A, Hess A, Kolle SN, Ma-Hock L, Wohlleben W, Dammann M, Strauss V, 9650 Treumann S, Gröters S, Wiench K, van Ravenzwaay B, Landsiedel R. Applicability of rat precision-cut 9651 lung slices in evaluating nanomaterial cytotoxicity, apoptosis, oxidative stress, and inflammation. 9652 Toxicol Appl Pharmacol. 2014 Apr 1;276(1):1-20. doi: 10.1016/j.taap.2013.12.017. Epub 2013 Dec 29 9653 9654 Savolainen K, Alenius H, Norppa H, Pylkkanen L, Tuomi T, Kasper G. 2010. Risk assessment of 9655 engineered nanomaterials and nanotechnologies-A review. Toxicology 269:92-104 9656 9657 Savolainen K, Alenius H. 2013. Disseminating widely. Nature Nanotech, 8: 72 9658 9659 Savolainen K, et al., Toxicology 269 (2010) 92–104 9660 9661 Sayes CM, Wahi R, Kurian PA, Liu Y, West JL, Ausman KD, Warheit DB, Colvin VL. Correlating 9662 nanoscale titania structure with toxicity: a cytotoxicity and inflammatory response study with human 9663 dermal fibroblasts and human lung epithelial cells. Toxicol Sci. 2006, 92:174-185 9664 9665 Sayers BC, Taylor AJ, Glista-Baker EE, Shipley-Phillips JK, Dackor RT, Edin ML, Lih FB, Tomer KB, Zeldin 9666 DC, Langenbach R, Bonner JC. Role of cyclooxygenase-2 in exacerbation of allerge-induced airway 9667 remodeling by multiwalled carbon nanotubes. Am J Respir Cell Mol Biol. 2013 Oct;49(4):525-35. 9668 9669 James F. Scabilloni et al.(2004) 9670 9671 SCCS, 2012a. Guidance on the safety assessment of nanomaterials in cosmetics. SCCS/1484/12, 9672 Scientific Committee on Consumer Safety (SCCS), Brussels, Belgium. 9673 9674 SCCS, 2012b. Opinion on Zinc oxide (nano form). SCCS/1489/12, Scientific Committee on Consumer 9675 Safety (SCCS), Brussels, Belgium. 9676 9677 SCCS, 2014a. Revision of the Addendum to the opinion SCCS/1489/12 on Zinc oxide (nano form) (S76) 9678 (SCCS/1518/13 - 23 July 2013 - revision of 22 April 2014). 9679

306

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9680 SCCS, 2014b. Revision of the memorandum on Relevance, Adequacy and Quality of Data in Safety 9681 Dossiers on Nanomaterials (SCCS/1524/13 – 12 December 2013 - revision of 27 March 2014). 9682 9683 SCCS, 2014c. Revision of the opinion on 2,2’-Methylene-bis-(6-(2H-benzotriazol-2-yl)-4-(1,1,3,3- 9684 tetramethylbutyl)phenol) nano form (S79) (SCCS/1443/11 - 18 March 2013 - revision of 23 July 2013). 9685 9686 SCCS, 2014d. Revision of the opinion on Carbon Black (nano-form) (SCCS/1515/13 – 12 December 9687 2013 - revision of 27 March 2014 ). 9688 9689 SCCS, 2014e. Revision of the opinion on Titanium Dioxide (nano form) (SCCS/1516/13 - 22 July 2013 - 9690 revision of 22 April 2014). 9691 9692 SCCS, 2015. SCCS opinion on Silica, Hydrated Silica, and Silica Surface Modified with 9693 Alkyl Silylates (nano form)(SCCS/ 1545/15- 29 September 2015- revision of 20 March 2015) 9694 9695 SCENIHR (2006). Opinion on the appropriateness of existing methodologies to assess the potential 9696 risk associated with engineered and adventitious products of nanotechnologies. EU Scientific 9697 Committee on Emerging and Newly Identified Health Risks (SCENIHR), Brussels, Belgium. 9698 9699 SCENIHR (2007a). Opinion on the scientific aspects of the existing and proposed definitions relating 9700 to products of nanoscience and nanotechnologies. EU Scientific Committee on Emerging and Newly 9701 Identified Health Risks (SCENIHR), Brussels, Belgium. 9702 9703 SCENIHR (2007b). Opinion on the appropriateness of the risk assessment methodology in accordance 9704 with the technical guidance documents for new and existing substances for assessing the risks of 9705 nanomaterials. EU Scientific Committee on Emerging and Newly Identified Health Risks (SCENIHR), 9706 Brussels, Belgium. 9707 9708 SCENIHR (2009). Risk assessment of products of nanotechnologies. EU Scientific Committee on 9709 Emerging and Newly Identified Health Risks (SCENIHR), Brussels, Belgium. 9710 9711 Schaeublin NM, Braydich-Stolle LK, Maurer EI, Park K, MacCuspie RI, Afrooz AR, Vaia RA, Saleh NB, 9712 Hussain SM. Does shape matter? Bioeffects of gold nanomaterials in a human skin cell model. 9713 Langmuir. 2012 Feb 14;28(6):3248-58. 9714

307

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9715 Schaeublin NM, Braydich-Stolle LK, Schrand AM, Miller JM, Hutchison J, Schlager JJ, Hussain SM. 9716 Surface charge of gold nanoparticles mediates mechanism of toxicity. Nanoscale. 2011 Feb;3(2):410- 9717 20. 9718 9719 Schindler, S., et al., 2009. Development, validation and application of the monocyte activation test 9720 for pyrogens based on human whole blood. ALTEX, 26 (4). 9721 9722 Schinwald et al.(2012) 9723 9724 Schleh C1, Kreyling WG, Lehr CM. Pulmonary surfactant is indispensable in order to simulate the in 9725 vivo situation. Part Fibre Toxicol. 2013 Mar 25;10:6. doi: 10.1186/1743-8977-10-6. 9726 9727 Schmidt C, Lautenschlaeger C, Collnot EM, Schumann M, Bojarski C, Schulzke JD, Lehr CM, Stallmach 9728 A. Nano- and microscaled particles for drug targeting to inflamed intestinal mucosa: a first in vivo 9729 study in human patients. J Control Release. 2013;165(2):139-45. 9730 9731 Schroder, J.M., The role of keratinocytes in defense against infection. Curr Opin Infect Dis, 2010. 9732 23(2): p. 106-10. 9733 9734 Schulte A, Althoff J, Ewe S, Richter-Reichhelm H-B, BGVV Group Investigators (2002) Two 9735 immunotoxicity ring studies according to OECD TG 407 – Comparison of data on cyclosporine A and 9736 hexachlorobenezene. Reg Tox Pharm 36: 12-21. 9737 9738 Schulte P, Geraci C, Zumwalde R, Hoover M, Castranova V, Kuempel E, Murashov V, Vainio H, 9739 Savolainen K. 2008. Sharpening the focus on occupational safety and health in nanotechnology. 9740 Scand J Work Environ Health 34, 471-8 9741 9742 B.C. Schanen, A.S. Karakoti, S. Seal, D.R. Drake, W.L. Warren, W.T. Self. Exposure to titanium dioxide 9743 nanomaterials provokes inflammation of an in vitro human immune construct. ACS Nano 3 (2009) 9744 2523-2532. 9745 9746 Schroeder BO, Bäckhed F. Signals from the gut microbiota to distant organs in physiology and 9747 disease. Nat Med. 2016 Oct 6;22(10):1079-1089. 9748 9749 Schulte A, Ruehl-Fehlert C (2006) Regulatory aspects of immunotoxicology. Exp Tox Path 57: 385-389

308

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9750 9751 Scott LT. Methods for the chemical synthesis of fullerenes. Angew Chem Int Ed Engl. 9752 2004;43(38):4994-5007. 9753 9754 Scotton CJ, Martinez FO, Smelt MJ, et al. Transcriptional profiling reveals complex regulation of the 9755 monocyte IL-1 beta system by IL-13. J Immunol. 2005;174(2):834-45. 9756 9757 Seed J, Caarney EW, Corley RA, Crofton KM, DeSesso JM, Foster PMD, Kavlock R, Kimmel G, Klaunig J, 9758 Meek ME, Preston RJ, Slikker W Jr, Tabacova S, Williams GM, Wiltse J, Zoeller RT, Fenner-Crisp P, 9759 Patton DE (2005) Overview: Usin mode of action and life stage information to evaluate the human 9760 relevance of animal toxicity data. Crit Rev Tox 35, 663-672. 9761 9762 Semmler-Behnke M, Kreyling WG, Schulz H, et al. Nanoparticle delivery in infant lungs. Proc Natl 9763 Acad Sci USA. 2012 Mar 27;109(13):5092-7. 9764 9765 Semmler-Behnke M, Kreyling WG, Lipka J, Fertsch S, Wenk A, Takenaka S, Schmid G, Brandau W. 9766 (2008). Biodistribution of 1.4- and 18-nm gold particles in rats. Small 4, 2108–2111. 9767 9768 Semmler M, Seitz J, Erbe F, Mayer P, Heyder J, Oberdörster G, Kreyling WG. (2004) Long-Term 9769 Clearance Kinetics of Inhaled Ultrafine Insoluble Iridium Particles from the Rat Lung, Including 9770 Transient Translocation into Secondary Organs, Inhalation Toxicology 16, 453-459, 2004. 9771 9772 Seneschal, J., et al., Human epidermal Langerhans cells maintain immune homeostasis in skin by 9773 activating skin resident regulatory T cells. Immunity, 2012. 36(5): p. 873-84 9774 9775 Seong SY, Matzinger P. Hydrophobicity: an ancient damage-associated molecular pattern that 9776 initiates innate immune responses. Nat Rev Immunol. 2004 Jun;4(6):469-78. 9777 9778 P. Serafini. Myeloid derived suppressor cells in physiological and pathological conditions: the good, 9779 the bad, and the ugly. Immunol. Res. 57 (2013) 172-184. 9780 Sergent JA, Paget V, Chevillard S. Toxicity and genotoxicity of nano-SiO2 on human epithelial 9781 intestinal HT-29 cell line. Ann Occup Hyg. 2012;56(5):622-30. 9782 9783 Sewald K & Braun A (2013) Xenobiotica 43(1):84-97. 9784

309

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9785 Shahbazi MA, Fernández TD, Mäkilä EM, Le Guével X, Mayorga C, Kaasalainen MH, Salonen JJ, 9786 Hirvonen JT, Santos HA. Surface chemistry dependent immunostimulative potential of porous silicon 9787 nanoplatforms. Biomaterials. 2014 Nov;35(33):9224-35. 9788 9789 Shahbazi MA, Santos HA. 2013. Improving oral absorption via drugloaded nanocarriers: absorption 9790 mechanisms, intestinal models and rational fabrication. Curr Drug Metab 14:28–56. 9791 9792 Shannahan JH, Kodavanti UP, Brown JM. Manufactured and airborne nanoparticle cardiopulmonary 9793 interactions: a review of mechanisms and the possible contribution of mast cells. Inhal Toxicol. 2012 9794 Apr;24(5):320-39. 9795 9796 Shapiro-Shelef M, Calame K. Regulation of plasma-cell development. Nat. Rev. Immunol. 5(3), 230– 9797 42 (2005) 9798 9799 Sharma V, Shukla RK, Saxena N, Parmar D, Das M, Dhawan A: DNA damaging potential of zinc oxide 9800 nanoparticles in human epidermal cells. Toxicol Lett 2009, 185: 211-218 9801 9802 Sharma G ety al., Nanotoxicology. 2014 Sep;8(6):663-75. 9803 9804 Shepherd NA, Crocker PR, Smith AP, Levison DA. Exogenous pigment in Peyer's patches. Hum Pathol. 9805 1987;18(1):50-4. 9806 9807 Shavandi Z, Ghazanfari T, Moghaddam KN. In vitro toxicity of silver nanoparticles on murine 9808 peritoneal macrophages. Immunopharmacol Immunotoxicol. 33, 135-140, 2011. 9809 9810 J. Shi, H.L. Karlsson, K. Johansson, V. Gogvadze, L. Xiao, J. Li , T. Burks, A. Garcia-Bennett, A. Uheida, 9811 M. Muhammed, S. Mathur, R. Morgenstern, V.E. Kagan, B. Fadeel. Microsomal glutathione 9812 transferase 1 protects against toxicity induced by silica nanoparticles but not by zinc oxide 9813 nanoparticles. ACS Nano 6 (2012) 1925-1938. 9814 9815 Shi LE, Li ZH, Zheng W, Zhao YF, Jin YF, Tang ZX. Synthesis, antibacterial activity, antibacterial 9816 mechanism and food applications of ZnO nanoparticles: a review. Food Addit Contam Part A Chem 9817 Anal Control Expo Risk Assess. 2014;31(2):173-86. 9818

310

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9819 Shipkowski KA, Taylor AJ, Thompson EA, Glista-Baker EE, Sayers BC, Messenger ZJ, Bauer RN, Jaspers 9820 I, Bonner JC. An allergic lung microenvironment suppresses carbon nanotube-induced inflammasome 9821 activation via STAT6-dependent inhibition of caspase-1. PLoS One. 2015 Jun 19;10(6):e0128888. 9822 9823 Shiraishi K, Kawano K, Maitani Y, Aoshi T, Ishii KJ, Sanada Y, Mochizuki S, Sakurai K, Yokoyama M. 9824 Exploring the relationship between anti-PEG IgM behaviors and PEGylated nanoparticles and its 9825 significance for accelerated blood clearance. J Control Release. 2016 Jul 28;234:59-67. 9826 9827 Shiu, J., et al., Skin: Immunological Defence Mechanisms, in eLS. 2001, John Wiley & Sons, Ltd. 9828 9829 Shurin MR, Yanamala N, Kisin ER, Tkach AV, Shurin GV, Murray AR, Leonard HD, Reynolds JS, Gutkin 9830 DW, Star A, Fadeel B, Savolainen K, Kagan VE, Shvedova AA. Graphene oxide attenuates Th2-type 9831 immune responses, but augments airway remodeling and hyperresponsiveness in a murine model of 9832 asthma. ACS Nano. 2014 Jun 24;8(6):5585-99. 9833 9834 Shvedova AA, Kisin ER, Murray AR, Kommineni C, Castranova V, Fadeel B, Kagan VE. Increased 9835 accumulation of neutrophils and decreased fibrosis in the lung of NADPH oxidase-deficient C57BL/6 9836 mice exposed to carbon nanotubes. Toxicol Appl Pharmacol. 2008 Sep 1;231(2):235-40. 9837 9838 Shvedova AA, Kisin ER, Porter D, Schulte P, Kagan VE, Fadeel B, Castranova V. 2009. Mechanisms of 9839 pulmonary toxicity and medical applications of carbon nanotubes: Two faces of Janus? Pharmacol 9840 Ther 121:192-204. 9841 9842 Shvedova AA, Kisin ER, Yanamala N, Tkach AV, Gutkin DW, Star A, Shurin GV, Kagan VE, Shurin MR. 9843 MDSC and TGFβ are required for facilitation of tumor growth in the lungs of mice exposed to carbon 9844 nanotubes. Cancer Res. 2015 Apr 15;75(8):1615-23. 9845 9846 Shvedova AA, Pietroiusti A, Fadeel B, Kagan VE. Mechanisms of carbon nanotube-induced toxicity: 9847 focus on oxidative stress. Toxicol Appl Pharmacol. 2012 Jun 1;261(2):121-33. 9848 9849 A.A. Shvedova, A.V. Tkach, E.R. Kisin, T. Khaliullin, S. Stanley, D.W. Gutkin, A. Star, Y. Chen, G.V. 9850 Shurin, V.E. Kagan, M.R. Shurin. Carbon nanotubes enhance metastatic growth of lung carcinoma via 9851 up-regulation of myeloid-derived suppressor cells. Small 9 (2013) 1691-1695. 9852

311

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9853 Shvedova AA, Fabisiak JP, Kisin ER, Murray AR, Roberts JR, Tyurina YY, Antonini JM, Feng WH, 9854 Kommineni C, Reynolds J, Barchowsky A, Castranova V, Kagan VE. Sequential exposure to carbon 9855 nanotubes and bacteria enhances pulmonary inflammation and infectivity. Am J Respir Cell Mol Biol. 9856 2008 May;38(5):579-90. 9857 9858 Shvedova AA, Yanamala N, Kisin ER, Khailullin TO, Birch ME, Fatkhutdinova LM. Integrated Analysis of 9859 Dysregulated ncRNA and mRNA Expression Profiles in Humans Exposed to Carbon Nanotubes. PLoS 9860 One. 2016 Mar 1;11(3):e0150628. 9861 9862 Shvedova AA, Kisin ER, Mercer R, et al. (2005) Unusual inflammatory and fibrogenic pulmonary 9863 responses to single-walled carbon nanotubes in mice. Am J Physiol 2005;289:L698-L708. 9864 9865 Shvedova, A.A., et al., Exposure to carbon nanotube material: assessment of nanotube cytotoxicity 9866 using human keratinocyte cells. J Toxicol Environ Health A, 2003. 66(20): p. 1909-26. 9867 9868 Shvedova AA, Kisin ER, Porter D, Schulte P, Kagan VE, Fadeel B, Castranova V. 2009. Mechanisms of 9869 pulmonary toxicity and medical applications of carbon nanotubes: Two faces of Janus? Pharmacol 9870 Ther 121:192-204 9871 9872 Shvedova AA, Sager T, Murray AR, Kisin E, Porter DW, Leonard SS, Schwegler-Berry D, Robinson VA, 9873 Castranova V. 2007. Critical Issues in the Evaluation of Possible Adverse Pulmonary Effects Resulting 9874 from Airborne Nanoparticles. In: Nanotoxicology - Characterization, Dosing and Health Effects. 9875 Monteiro-Riviere NA and Tran CL (Eds.), Informa healthcare, New York, pp. 225-36 9876 9877 Shvedova AA et al., Toxicol Appl Pharmacol. 2012 Jun 1;261(2):121-33 9878 9879 Shvedova et al 2008 Am J Physiol Lung cell mol physiol 295(4) L552-565 9880 9881 Silva RM, Anderson DS, Franzi LM, Peake JL, Edwards PC, Van Winkle LS, Pinkerton KE. Pulmonary 9882 effects of silver nanoparticle size, coating, and dose over time upon intratracheal instillation. Toxicol 9883 Sci. 2015 Mar;144(1):151-62. 9884 9885 Silva et al 2014 ACS nano vol 8 No 9 8911-8931; 9886 9887 Simko M, et al., Int. J. Environ. Res. Public Health 2014, 11, 4026-4048;

312

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9888 9889 A. Singh, H. Nie, B. Ghosn, H. Qin, L.W. Kwak, K. Roy. Efficient modulation of T-cell response by dual- 9890 mode, single-carrier delivery of cytokine-targeted siRNA and DNA vaccine to antigen-presenting cells. 9891 Mol. Ther. 16 (2008) 2011-2021 9892 9893 Singh SP, Rahman MF, Murty US, Mahboob M, Grover P. Comparative study of genotoxicity and 9894 tissue distribution of nano and micron sized iron oxide in rats after acute oral treatment. Toxicol Appl 9895 Pharmacol. 2013;266(1):56-66. 9896 9897 Smulders et al. Particle and Fibre Toxicology 2012, 9:41 9898 9899 Smulders, S., et al., Nano-TiO2 modulates the dermal sensitization potency of dinitrochlorobenzene 9900 after topical exposure. Br J Dermatol, 2015. 172(2): p. 392-9 9901 9902 Soares T1, Ribeiro D1, Proença C1, Chisté RC1, Fernandes E1, Freitas M2. 9903 Size-dependent cytotoxicity of silver nanoparticles in human neutrophils assessed by multiple 9904 analytical approaches. Life Sci. 145, 247-254, 2016. 9905 9906 Song KS, Sung JH, Ji JH, Lee JH, Lee JS, Ryu HR, Lee JK, Chung YH, Park HM, Shin BS, Chang HK, Kelman 9907 B, Yu IJ. Recovery from silver-nanoparticle-exposure-induced lung inflammation and lung function 9908 changes in Sprague Dawley rats. Nanotoxicology. 7, 169-180, 2013. 9909 9910 Song ZM, Chen N, Liu JH, Tang H, Deng X, Xi WS, Han K, Cao A, Liu Y, Wang H. Biological effect of food 9911 additive titanium dioxide nanoparticles on intestine: an in vitro study. J Appl Toxicol. 2015 Jun 23. 9912 9913 C. Sonich-Mullin, R. Fielder, J. Wiltse, K. Baetcke, J. Dempsey, P. Fenner-Crisp, D. Grant, M. Hartley, A. 9914 Knaap, D. Kroese, I. Mangelsdorf, E. Meek, J. M. Rice, M. Younes, Regulatory Toxicology and 9915 Pharmacology Vol. 34, 1, 2001. 9916 9917 Srinivas A, Rao PJ, Selvam G, Murthy PB, Reddy PN. Acute inhalation toxicity of cerium oxide 9918 nanoparticles in rats. Toxicol Lett. 2011; 205(2):105-15. 9919 9920 Stark WJ. Nanoparticles in biological systems. Angew Chem Int Ed Engl. 2011 Feb 7;50(6):1242-58. 9921

313

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9922 Stebounova LV, Adamcakova-Dodd A, Kim JS, Park H, O'Shaughnessy PT, Grassian VH, Thorne PS. 9923 Nanosilver induces minimal lung toxicity or inflammation in a subacute murine inhalation model. Part 9924 Fibre Toxicol. 8:5, 2011. 9925 9926 Steimer A, Haltner E, Lehr CM: Cell culture models of the respiratory tract relevant to pulmonary 9927 drug delivery. J Aerosol Med 2005, 18: 137-82 9928 9929 Stokes AB, Kieninger E, Schögler A, Kopf BS, Casaulta C, Geiser T, Regamey N, Alves MP.Comparison 9930 of three different brushing techniques to isolate and culture primary nasal epithelial cells from 9931 human subjects. Exp Lung Res. 2014 Sep;40(7):327-32. doi: 10.3109/01902148.2014.925987. Epub 9932 2014 Jul 24. 9933 9934 Stone; oberdorster; Holder et al chem risk tox 2012 9935 9936 Stone KD, Prussin C, Metcalfe DD. IgE, mast cells, basophils, and eosinophils. J. Allergy Clin. Immunol. 9937 125, 73–80 (2010) 9938 9939 T. Strowig, J. Henao-Mejia, E. Elinav, R. Flavell. Inflammasomes in health and disease. Nature 481 9940 (2012) 278-286. 9941 9942 Sun B, Ji Z, Liao YP, Wang M, Wang X, Dong J, Chang CH, Li R, Zhang H, Nel AE, Xia T. Engineering an 9943 effective immune adjuvant by designed control of shape and crystallinity of aluminum oxyhydroxide 9944 nanoparticles. ACS Nano. 2013 Dec 23;7(12):10834-49. 9945 9946 Sun B, Wang X, Ji Z, Wang M, Liao YP, Chang CH, Li R, Zhang H, Nel AE, Xia T. NADPH oxidase- 9947 dependent NLRP3 inflammasome activation and its important role in lung fibrosis by multiwalled 9948 carbon nanotubes. Small. 2015 May 6;11(17):2087-97. 9949 9950 Sun B, Wang X, Ji Z, Li R, Xia T. NLRP3 inflammasome activation induced by engineered 9951 nanomaterials. Small. 2013; 9(9-10):1595-607. 9952 9953 Sun JC, Lanier LL. NK cell development, homeostasis and function: parallels with CD8+ T cells. Nat. 9954 Rev. Immunol. 11(10), 645–57 (2011 9955 Sundström C, Nilsson K. Establishment and characterization of a human histiocytic lymphoma cell line 9956 (U-937). Int J Cancer. 1976 May 15;17(5):565-77.

314

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9957 9958 Sung JH, Ji JH, Park JD, Yoon JU, Kim DS, Jeon KS, Song MY, Jeong J, Han BS, Han JH, Chung YH, Chang 9959 HK, Lee JH, Cho MH, Kelman BJ, Yu IJ. Subchronic inhalation toxicity of silver nanoparticles. Toxicol 9960 Sci. 108, 452-461, 2009. 9961 9962 Sung JH, Ji, J.H., Park, J.D., Song, M.Y., Song, K.S., Ryu1, H.R., Yoon, J.U., Jeon, K.S., Jeong, J., Han, B.S., 9963 Chung, J.H., Chang, H.K., Lee, J.H., Kim, D.W., Kelman, B.J., and Yu, I.J. (2011). Subchronic inhalation 9964 toxicity of gold nanoparticles. Part Fibre Toxicol 8, 16. 9965 9966 Szebeni J. Complement activation-related pseudoallergy: a stress reaction in blood triggered by 9967 nanomedicines and biologicals. Mol Immunol. 2014 Oct;61(2):163-73 9968 9969 Takagi A, Hirose A, Nishimura T, Fukumori N, Ogata A, Ohashi N, Kitajima S, Kanno J. 2008. Induction 9970 of mesothelioma in p53+/- mouse by intraperitoneal application of multi-wall carbon nanotube. J 9971 Toxicol Sci 33:105-16 9972 9973 Takagi A, Hirose A, Futakuchi M, Tsuda H, Kanno J. 2012. Dose-dependent mesothelioma induction 9974 by intraperitoneal administration of multi-wall carbon nanotubes in p53 heterozygous mice. Cancer 9975 Sci 103: 1440–4 9976 9977 Takenaka et al.(2006), 9978 9979 Taquahashi et al.(2013) 9980 9981 Tassinari R, Cubadda F, Moracci G, Aureli F, D’Amato M, Valeri M, De Berardis B, Raggi A, Mantovani 9982 A, Passeri D, Rossi M, Maranghi F. (2014). Oral, short-term exposure to titanium dioxide 9983 nanoparticles in Sprague-Dawley rat: focus on reproductive and endocrine systems and spleen. 9984 Nanotoxicology 8, 654-662. 9985 9986 Taylor AJ, McClure CD, Shipkowski KA, Thompson EA, Hussain S, Garantziotis S, Parsons GN, Bonner 9987 JC. Atomic layer deposition coating of carbon nanotubes with aluminum oxide alters pro-fibrogenic 9988 cytokine expression by human mononuclear phagocytes in vitro and reduces lung fibrosis in mice in 9989 vivo. PLoS One. 2014 Sep 12;9(9):e106870. 9990

315

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

9991 Teeguarden, JG, Webb-Robertson BJ, Waters KM, Murray AR, . Kisin ER, Varnum SM, Jacobs JM, 9992 Pounds JG, Zanger, R. . Shvedova, AA. TOXICOLOGICAL SCIENCES 120(1), 123–135 (2011) 9993 9994 Teeguarden JG et al., Toxicol Sci. 2011 Mar;120(1):123-135 9995 Teeguarden JG, Hinderliter PM, Orr G, Thrall BD, Pounds JG: Particokinetics in vitro: Dosimetry 9996 considerations for in vitro nanoparticle toxicology assessments (vol 95, pg 300, 2007). Toxicol Sci 9997 2007, 97:614-614 9998 9999 Teeguarden JG, Hinderliter PM, Orr G, Thrall BD, Pounds JG. 2007. Particokinetics in vitro: dosimetry 10000 considerations for in vitro nanoparticle toxicity assessments. Toxicol Sci 95:300–312 10001 10002 S. Tenzer, D. Docter, J. Kuharev, A. Musyanovych, V. Fetz, R. Hecht, F. Schlenk, D. Fischer, K. Kiouptsi, 10003 C. Reinhardt, K. Landfester, H. Schild, M. Maskos, S.K. Knauer, R.H. Stauber. Rapid formation of 10004 plasma protein corona critically affects nanoparticle pathophysiology. Nat. Nanotechnol. 8 (2013) 10005 772-781. 10006 10007 Tervonen T, Lahdelma R, Almeida Dias J, Figueira J, Salminen P. 2007. SMAA-TRI. In: Linkov I, Kiker 10008 GA, Wenning RJ, eds. Environmental Security in Harbors and Coastal Areas: Management Using 10009 Comparative Risk Assessment and Multi-Criteria Decision Analysis. Dordrecht: Springer, the 10010 Netherlands,217–231 10011 10012 Thiagarajan G, Greish K, Ghandehari H. Charge affects the oral toxicity of poly(amidoamine) 10013 dendrimers. Eur J Pharm Biopharm. 2013a;84(2):330-4. 10014 10015 Thiagarajan G, Sadekar S, Greish K, Ray A, Ghandehari H. Evidence of oral translocation of anionic 10016 G6.5 dendrimers in mice. Mol Pharm. 2013b;10(3):988-98. 10017 10018 Thompson EA, Sayers BC, Glista-Baker EE, Shipkowski KA, Ihrie MD, Duke KS, Taylor AJ, Bonner JC. 10019 Role of signal transducer and activator of transcription 1 in murine allergen-induced airway 10020 remodeling and exacerbation by carbon nanotubes. Am J Respir Cell Mol Biol. 2015, 53(5):625-36. 10021 10022 Thompson EA, Sayers BC, Glista-Baker EE, Shipkowski KA, Taylor AJ, Bonner JC. Innate Immune 10023 Responses to Nanoparticle Exposure in the Lung. J Environ Immunol Toxicol. 2014, Jul-Sep;1(3):150- 10024 156. 10025

316

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10026 Thompson EA, Sayers BC, Glista-Baker EE, Shipkowski KA, Ihrie MD, Duke KS, Taylor AJ, Bonner JC. 10027 STAT1 attenuates murine allergen-induced airway remodeding and exacerbation by carbon 10028 nanotubes. Am J Respir Cell Mol Biol. 2015 Mar 25. [Epub ahead of print] 10029 10030 Anne Thoustrup Saber, Jacob Stuart Lamson, Nicklas Raun Jacobsen, Gitte Ravn-Haren, Karin Sørig 10031 Hougaard, Allen Njimeri Nyendi, Pia Wahlberg, Anne Mette Madsen, Petra Jackson, Ha°kan Wallin, 10032 Ulla Vogel. Particle-Induced Pulmonary Acute Phase Response Correlates with Neutrophil Influx 10033 Linking Inhaled Particles and Cardiovascular Risk. PLoS ONE 8(7): e69020 (2013) 10034 10035 Tian X, Yang Z, Duan G, Wu A, Gu Z, Zhang L, Chen C, Chai Z, Ge C, Zhou R. Graphene oxide 10036 nanosheets retard cellular migration via disruption of actin cytoskeleton. Small. 2017 Jan;13(3). 10037 10038 Tilton SC, Karin NJ, Tolic A, Xie Y, Lai X, Hamilton RF Jr, Waters KM, Holian A, Witzmann FA, Orr G. 10039 Three human cell types respond to multi-walled carbon nanotubes and titanium dioxide nanobelts 10040 with cell-specific transcriptomic and proteomic expression patterns. Nanotoxicology. 2014;8(5):533- 10041 48. 10042 10043 A.V. Tkach, G.V. Shurin, M.R. Shurin, E.R. Kisin, A.R. Murray, S.H. Young, A. Star, B. Fadeel, V.E. Kagan, 10044 A.A. Shvedova. Direct effects of carbon nanotubes on dendritic cells induce immune suppression 10045 upon pulmonary exposure. ACS Nano 5 (2011) 5755-5762. 10046 10047 Tomalia DA, Baker H, Dewald J, Hall M, Kallos G, Martin S, Roeck J, Ryder J, Smith P. A New Class of 10048 Polymers: Starburst-Dendritic Macromolecules. Polymer Journal 1985;17:117-132. 10049 10050 Tsoli M, Kuhn H, Brandau W, Esche H, Schmid G. Cellular uptake and toxicity of Au55 clusters. Small. 10051 2005 Aug;1(8-9):841-4. 10052 10053 Tsugita M, Morimoto N, Nakayama M. SiO2 and TiO2 nanoparticles synergistically trigger 10054 macrophage inflammatory responses. Part Fibre Toxicol. 2017 Apr 11;14(1):11. 10055 10056 Tulinska et al. zImmunotoxicity and genotoxicity testing of PLGA-PEO nanoparticles in human blood 10057 cell model. Nanotoxicology, 2015; 9(S1): 33–43 10058

317

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10059 Turabekova M, Rasulev B, Theodore M, Jackman J, Leszczynska D, Leszczynski J. Immunotoxicity of 10060 nanoparticles: a computational study suggests that CNTs an C60 fullerenes might be recognized as 10061 pathogens by toll-like receptors. Nanoscale. 2014; 6(7):3488-95 10062 10063 Turnbaugh PJ, Ley RE, Hamady M, Fraser-Liggett CM, Knight R, Gordon JI. The human microbiome 10064 project. Nature 2007; 449:804-10. 10065 10066 Tiwari DK, Jin T, Behari J. Dose-dependent in-vivo toxicity assessment of silver nanoparticle in Wistar 10067 rats. Toxicol Mech Methods. 21, 13-24, 2011. 10068 10069 Tyurina YY et al., ACS Nano. 2011 Sep 27;5(9):7342-53 10070 10071 Umemoto EY, Speck M, Shimoda LM, Kahue K, Sung C, Stokes AJ, Turner H. Single-walled carbon 10072 nanotube exposure induces membrane rearrangement and suppression of receptor-mediated 10073 signalling pathways in model mast cells. Toxicol Lett. 2014;229:198-209. 10074 10075 D.M. Underhill, H.S. Goodridge, Information processing during phagocytosis, Nat Rev Immunol, 12 10076 (2012) 492-502. 10077 10078 Urban D, Rodriguez-Lorenzo L, Balog S, Kinnear C, Rothen-Rutishauser B, Petri-Fink A. Plasmonic 10079 nanoparticles and their characterization in physiological fluids. 10080 10081 US EPA (1998) OPPTS Harmonized Test Guideline: OPPTS 870.7800. Immunotoxicity. EPA/712/C- 10082 98/351. United States Environmental Protection Agency, Office of Chemical Safety and Pollution 10083 Prevention, Washington DC. 10084 10085 US EPA, 2007. Nanotechnology white paper, EPA 100/B-07/001, US EPA, Washington DC. 10086 10087 US EPA (2008). Draft Nanomaterial Research Strategy. Report EPA/600/S-08/002, US Environmental 10088 Protection Agency (US EPA), Office of Research and Development. Available at: 10089 epa.gov/ncer/nano/publications/nano_strategy_012408.pdf. 10090 10091 Vallhov H, Qin J, Johansson SM, Ahlborg N, Muhammed MA, Scheynius A, Gabrielsson S. 2006. The 10092 importance of an endotoxin-free environment during the production of nanoparticles used in 10093 medical applications. Nano Lett. 6(8):1682-6.

318

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10094 10095 Vance ME, Kuiken T, Vejerano EP, McGinnis SP, Hochella Jr. MF, Rejeski D, Hull MS. Nanotechnology 10096 in the real world: Redeveloping the nanomaterial consumer products inventory. Beilstein J. 10097 Nanotechnol. 6, 1769–1780, 2015. 10098 10099 Vandebriel RJ, Tonk EC, De La Fonteyne-Blankestijn LJ, Gremmer ER, Verharen HW, Van Der Ven LT, 10100 Van Loveren H, De Jong WH. Immunotoxicity of silver nanoparticles in an intravenous 28-day 10101 repeated-dose toxicity study in rats. Part Fibre Toxicol. 11:21. 2014 10102 10103 van den Brule S, Ambroise J, Lecloux H, Levard C, Soulas R, De Temmerman PJ, Palmai-Pallag M, 10104 Marbaix E, Lison D. Dietary silver nanoparticles can disturb the gut microbiota in mice. Part Fibre 10105 Toxicol. 2016 Jul 8;13(1):38. 10106 10107 Van Der Zande M, Vandebriel RJ, Van Doren E, Kramer E, Herrera Rivera Z, Serrano-Rojero CS, 10108 Gremmer ER, Mast J, Peters RJ, Hollman PC, Hendriksen PJ, Marvin HJ, Peijnenburg AA, Bouwmeester 10109 H. Distribution, elimination, and toxicity of silver nanoparticles and silver ions in rats after 28-day oral 10110 exposure. ACS Nano. 6, 7427-7442, 2012. 10111 10112 Van Der Zande M, Vandebriel RJ, Groot MJ, Kramer E, Herrera Rivera ZE, Rasmussen K, 10113 Ossenkoppele JS, Tromp P, Gremmer ER, Peters RJ, Hendriksen PJ, Marvin HJ, Hoogenboom RL, 10114 Peijnenburg AA, Bouwmeester H. (2014) Sub-chronic toxicity study in rats orally exposed to 10115 nanostructured silica. Part Fibre Toxicol., 11, 8. Colloids and Surfaces B: Biointerfaces 137:39-49 10116 (2016). 10117 10118 Vance, M. E., Kuiken, T., Vejerano, E. P., McGinnis, S. P., Hochella, M. F., Jr., Rejeski, D. and Hull, M. S. 10119 (2015) Nanotechnology in the real world: Redeveloping the nanomaterial consumer products 10120 inventory. Beilstein Journal of Nanotechnology, 6, 1769-1780. 10121 10122 Vandebriel RJ, Tonk ECM, de la Fonteyne-Blankestijn LJ, Gremmer ER, Verharen HW, van der Ven LT, 10123 van Loveren H, de Jong WH (2014) Immunotoxicity of silver nanoparticles in an intravenous 28-day 10124 repeated-dose toxicity study in rats. Part Fibre Tox 11/21: 1-9. 10125 http://www.particleandfibretoxicology.com/content/11/1/21 10126

319

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10127 Varypataki EM, Silva AL, Barnier-Quer C, Collin N, Ossendorp F, Jiskoot W. Synthetic long peptide- 10128 based vaccine formulations for induction of cell mediated immunity: A comparative study of cationic 10129 liposomes and PLGA nanoparticles. J Control Release. 2016; 226:98-106. 10130 10131 Veranth JM, Kaser EG, Veranth MM, Koch M, Yost GS: Cytokine responses of human lung cells (BEAS- 10132 2B) treated with micron-sized and nanoparticles of metal oxides compared to soil dusts. Part Fibre 10133 Toxicol 2007, 4: 2 10134 10135 Verma P, Ahuja M. Cubic liquid crystalline nanoparticles: optimization and evaluation for ocular 10136 delivery of tropicamide. Drug Deliv. 2016 Feb 15:1-12. 10137 10138 Vermeire 2007 10139 10140 Vesterdal LK, Danielsen PH, Folkmann JK, Jespersen LF, Aguilar-Pelaez K, Roursgaard M, Loft S, Møller 10141 P. Accumulation of lipids and oxidatively damaged DNA in hepatocytes exposed to particles. Toxicol 10142 Appl Pharmacol. 2014;274:350-60. 10143 10144 Vickers AE & Fisher RL (2004). Chemico-biological interactions 150(1):87-96. 10145 10146 Vietti G, Ibouraadaten S, Palmai-Pallag M, Yakoub Y, Bailly C, Fenoglio I, Marbaix E, Lison D, van den 10147 Brule S. Towards predicting the lung fibrogenic activity of nanomaterials: experimental validation of 10148 an in vitro fibroblast proliferation assay. Part Fibre Toxicol. 2013 Oct 10;10:52. doi: 10.1186/1743- 10149 8977-10-52 10150 10151 Vietti G, Lison D, van den Brule S. Mechanisms of lung fibrosis induced by carbon nanotubes: towards 10152 an Adverse Outcome Pathway (AOP). Part Fibre Toxicol. 2016, 29;13:11. 10153 10154 Vilnius Aerosol Generator Ref ?? 10155 10156 Visser CE, Tekstra J, Brouwer-Steenbergen JJ, et al. (1998) Chemokines produced by mesothelial cells: 10157 huGRO-alpha, IP-10, MCP-1 and RANTES. Clin Exp Immunol. 1998;112(2):270-275. 10158 10159 Vlachou E, Chipp E, Shale E, Wilson YT, Papini R, Moiemen NS. (2007). The safety of nanocrystalline 10160 silver dressings on burns: a study of systemic silver absorption.Burns. 33, 979-985. 10161

320

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10162 Vogel U, Savolainen K, Wu Q, van Tongeren M, Brouwer D, Berges M (eds.). 2014. Handbook of 10163 Nanosafety: Measurement, Exposure and Toxicology. Academic Press. 10164 10165 Voisin C, Aerts C, Jakubczak E, Houdret JL, Tonnel TB: [Effects of nitrogen dioxide on alveolar 10166 macrophages surviving in the gas phase. A new experimental model for the study of in vitro 10167 cytotoxicity of toxic gases (author's transl)]. Bull Eur Physiopathol Respir 1977, 13: 137-144 10168 10169 Von Garnier C, Filgueira L, Wikstrom M, et al. Anatomical location determines the distribution and 10170 function of dendritic cells and other APCs in the respiratory tract. J. Immunol. 175(3), 1609–18 10171 (2005). 10172 10173 Von Garnier C, Wikstrom Me, Zosky G et al.: Allergic airways disease develops after an increase in 10174 allergen capture and processing in the airway mucosa. J Immunol 179(9), 5748-5759 (2007) 10175 10176 Von Garnier C, Nicod LP. Immunology taught by lung dendritic cells. Swiss Med Wkly. 139(13-14), 10177 186–92 (2009) 10178 10179 Walling BE, Kuang Z, Hao Y, Estrada D, Wood JD, Lian F, Miller LA, Shah AB, Jeffries JL, Haasch RT, 10180 Lyding JW, Pop E, Lau GW. Helical carbon nanotubes enhance the early immune response and inhibit 10181 macrophage-mediated phagocytosis of Pseudomonas aeruginosa. PLoS One. 2013, 18;8(11):e80283. 10182 10183 Walling MA, Novak JA, Shepard JR. Quantum dots for live cell and in vivo imaging. Int J Mol Sci. 10184 2009;10(2):441-91. 10185 10186 Walczak AP, Fokkink R, Peters R, Tromp P, Herrera Rivera ZE, Rietjens IM, Hendriksen PJ, 10187 Bouwmeester H. Behaviour of silver nanoparticles and silver ions in an in vitro human 10188 gastrointestinal digestion model. Nanotoxicology. 7, 1198-1210, 2013. 10189 10190 Walczak AP, Hendriksen PJM, Woutersen RA, Van Der Zande M, Undas AK, Helsdingen R, Van Den 10191 Berg HHJ, Rietjens IMCM, Bouwmeester H. (2015) Bioavailability and biodistribution of differently 10192 charged polystyrene nanoparticles upon oral exposure in rats. J. Nanopart. Res., 17, 231. 10193 10194 Wang L, Luanpitpong S, Castranova V, Tse W, Lu Y, Pongrakhananon V, Rojanasakul Y. Carbon 10195 nanotubes induce malignant transformation and tumorigenesis of human lung epithelial cells. Nano 10196 Lett. 2011 Jul 13;11(7):2796-803.

321

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10197 10198 Wang Y, Wang B, Zhu MT, Li M, Wang HJ, Wang M, Ouyang H, Chai ZF, Feng WY, Zhao YL. Microglial 10199 activation, recruitment and phagocytosis as linked phenomena in ferric oxide nanoparticle exposure. 10200 Toxicol Lett. 2011;205(1):26-37. 10201 10202 Wang, J., Zhoua, G., Chena, C.,Yu, H., Wang, T., Mad, Y., Jia, G., Gaoa, Y., Li, B., Suna, J., Li, Y., Jiao, F., 10203 Zhaoa, Y., Chai, Z. (2007). Acute toxicity and biodistribution of different sized titanium dioxide 10204 particles in mice after oral administration. Toxicol Lett 168, 176–185. 10205 10206 Wang X, Xia T, Ntim SA, Ji Z, Lin S, Meng H, Chung CH, George S, Zhang H, Wang M, Li N, Yang Y, 10207 Castranova V, Mitra S, Bonner JC, Nel AE. Dispersal state of multiwalled carbon nanotubes elicits 10208 profibrogenic cellular responses that correlate with fibrogenesis biomarkers and fibrosis in the 10209 murine lung.ACS Nano. 2011 Dec 27;5(12):9772-87. 10210 10211 Wang J, Liu Y, Jiao F, Lao F, Li W, Gu Y, Li Y, Ge C, Zhou G, Li B, Zhao Y, Chai Z, Chen C. Time- 10212 dependent translocation and potential impairment on central nervous system by intranasally instilled 10213 TiO(2) nanoparticles. Toxicology. 2008 Dec 5;254(1-2):82-90. 10214 10215 Wang Z, Xu Y, Meng X, Watari F, Liu H, Chen X. Suppression of c-Myc is involved in multi-walled 10216 carbon nanotubes' down-regulation of ATP-binding cassette transporters in human colon 10217 adenocarcinoma cells. Toxicol Appl Pharmacol. 2015;282(1):42-51. 10218 10219 Wang YZ, Zhang P, Rice AB, et al. Regulation of interleukin-1beta -induced platelet-derived growth 10220 factor receptor-alpha expression in rat pulmonary myofibroblasts by p38 mitogen-activated protein 10221 kinase. J Biol Chem. 2000;275(29):22550-7. 10222 10223 Wang X, Tian X, Ntim SA, et al. The dispersal state of multi-walled carbon nanotubes influences pro- 10224 fibrogenic epithelial and macrophage responses that correlate with the extent of pulmonary fibrosis. 10225 ACS Nano 2011;5(12):9772-9787. 10226 10227 Wan-Seob Cho et al ., Environmental Health Perspectives • volume 118 | number 12 | December 10228 2010 10229

322

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10230 Warheit DB, Webb TR, Sayes CM, Colvin VL, Reed KL. Pulmonary Instillation Studies with Nanoscale 10231 TiO2 Rods and Dots in Rats: Toxicity Is not Dependent upon Particle Size and Surface Area. Tox Sc 91, 10232 227-236, 2006 10233 10234 Weir A, Westerhoff P, Fabricius L, Hristovski K, von Goetz N. Titanium dioxide nanoparticles in food 10235 and personal care products. Environ Sci Technol. 2012;46(4):2242-50. 10236 10237 Weldon BA, Faustman EM, Oberdorster G. Workman T, Griffith WC, Kneuer C, Yu IJ. Occupational 10238 exposure limit for silver nanoparticles: considerations on the derivation of a general health-based 10239 value. Nanotoxicology. 10, 945-956, 2016. 10240 10241 West, B, and Wyzan, H. Investigations of the Possible Absorption of Titanium Dioxide from the 10242 Gastrointestinal Tract. American Cyanamid Company. Wayne, New Jersey. 1963. 10243 10244 Wheater PR, Burkitt HG, Daniels VG: Functional histology. Churchill Livingston; 1979. 10245 10246 White KL Jr, Gennings C, Myrray MJ, Dean JH (1994) Summary of an international methods validation 10247 study, carried out in nine laboratories, on the immunological assessment of cyclosporin A in the 10248 Fischer 344 rat. Tox In Vitro 8: 957-961. 10249 10250 WHO (1996) Principles and methods for assessing direct immunotoxicity associated with exposure to 10251 chemicals. Environmental Health Criteria 180. World Health Organization, Geneva, Switzerland. 10252 10253 WHO (1999a) Principles for the assessment of risks to human health from exposure to chemicals. 10254 Environmental Health Criteria 210. World Health Organization, Geneva, Switzerland. 10255 10256 WHO (1999b) Principles and methods for assessing allergic hypersensitization associated with 10257 exposure to chemicals. Environmental Health Criteria 212. World Health Organization, Geneva, 10258 Switzerland. 10259 10260 WHO (2006) Principles and methods for assessing autoimmunity associated with exposure to 10261 chemicals. Environmental Health Criteria 236. World Health Organization, Geneva, Switzerland. 10262 10263 WHO (2009) Principles for modelling dose-response for the risk assessment of chemicals. 10264 Environmental Health Criteria 239. World Health Organization, Geneva, Switzerland.

323

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10265 10266 WHO (2010) WHO human health risk assessment toolkit: chemical hazards. IPCS harmonization 10267 project document; no. 8. World Health Organization, Geneva, Switzerland. 10268 10269 WHO (2012) Guidance for immunotoxicity risk assessment for chemicals. IPCS harmonization project 10270 document; no. 10. World Health Organization, Geneva, Switzerland. 10271 10272 Wick P, Louw-Gaume AE, Kucki M, Krug HF, Kostarelos K, Fadeel B, Dawson KA, Salvati A, Vázquez E, 10273 Ballerini L, Tretiach M, Benfenati F, Flahaut E, Gauthier L, Prato M, Bianco A. Classification framework 10274 for graphene-based materials. Angew Chem Int Ed Engl. 2014;53(30):7714-8. 10275 10276 Wijagkanalan W, Kawakami S, Hashida M. Designing dendrimers for drug delivery and imaging: 10277 pharmacokinetic considerations. Pharm Res. 2011;28:1500-19. 10278 10279 Wijnhoven SWP, Peijnenburg WJGM, Herberts CA, Hagens WI, Oomen AG, Heugens EHW, Roszek B, 10280 Bisschops J, Gosens I, Van De Meent D, dekkers S, De Jong WH, Van Zijverden M, Sips AJAM, 10281 Geertsma RE. Nano-silver- a review of available data and knowledge gaps in human and 10282 environmental risk assessment. Nanotoxicology 3, 109-138, 2009. 10283 10284 Wildt B, Malinauskas RA, Brown RP. The effects of engineered nanomaterials on erythrocytes. In: 10285 Handbook of immunological properties of engineered nanomatyerials. EDs: Dobrovolskaia MA, 10286 McNeill SE. Frontiers in Nanobiomedical Research Vol 1. 2013. Page173-206. 10287 10288 Wilding LA, Bassis CM, Walacavage K, Hashway S, Leroueil PR, Morishita M, Maynard AD, Philbert 10289 MA, Bergin IL. Repeated dose (28-day) administration of silver nanoparticles of varied size and 10290 coating does not significantly alter the indigenous murine gut microbiome. Nanotoxicology 2015:1-8. 10291 10292 Wikstrom ME, Stumbles P a. Mouse respiratory tract dendritic cell subsets and the immunological 10293 fate of inhaled antigens. Immunol. Cell Biol. 85(3), 182–8 (2007) 10294 10295 Williams A, Halappanavar S. Application of biclustering of gene expression data and gene set 10296 enrichment analysis methods to identify potentially disease causing nanomaterials. Beilstein J 10297 Nanotechnol. 2015, 6:2438-48. 10298

324

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10299 Williams A, Halappanavar S. Application of biclustering of gene expression data and gene set 10300 enrichment analysis methods to identify potentially disease causing nanomaterials. Beilstein J 10301 Nanotechnol. 2015, 6:2438-48. 10302 10303 Wingard CJ, Walters DM, Cathey BL, Hilderbrand SC, Katwa P, Lin S, Ke PC, Podila R, Rao A, Lust RM, 10304 Brown JM. Mast cells contribute to altered vascular reactivity and ischemia-reperfusion injury 10305 following cerium oxide nanoparticle instillation. Nanotoxicology. 2011; 5(4):531-45. 10306 10307 Witasp E, Kupferschmidt N, Bengtsson L, Hultenby K, Smedman C, Paulie S, Garcia-Bennett AE, Fadeel 10308 B. Efficient internalization of mesoporous silica particles of different sizes by primary human 10309 macrophages without impairment of macrophage clearance of apoptotic or antibody-opsonized 10310 target cells. Toxicol Appl Pharmacol. 2009 Sep 15;239(3):306-19. 10311 10312 Witasp E, Shvedova AA, Kagan VE, Fadeel B. Single-walled carbon nanotubes impair human 10313 macrophage engulfment of apoptotic cell corpses. Inhal Toxicol. 2009 Jul;21 Suppl 1:131-6. 10314 10315 Brian A. Wong, Toxicol Pathol January 2007 vol. 35 no. 1 3-14 10316 10317 Wu J, Liu W, Xue C, Zhou S, Lan F, Bi L et al.: Toxicity and penetration of TiO2 nanoparticles in hairless 10318 mice and porcine skin after subchronic dermal exposure. Toxicol Lett 2009, 191: 1-8 10319 10320 Xia T, Hamilton RF, Bonner JC, Crandall ED, Elder A, Fazlollahi F, Girtsman TA, Kim K, Mitra S, Ntim SA, 10321 Orr G, Tagmount M, Taylor AJ, Telesca D, Tolic A, Vulpe CD, Walker AJ, Wang X, Witzmann FA, Wu N, 10322 10323 Xianglu Han et al., Toxicology. 2011 September 5; 287(0): 99–104. 10324 10325 Xie Y, Zink JI, Nel A, Holian A. Interlaboratory evaluation of in vitro cytotoxicity and inflammatory 10326 responses to engineered nanomaterials: the NIEHS Nano GO Consortium. Environ Health Perspect. 10327 2013 Jun;121(6):683-90. 10328 10329 Xu Y, Tang H, Liu JH, Wang H, Liu Y. Evaluation of the adjuvant effect of silver nanoparticles both in 10330 vitro and in vivo. Toxicol Lett. 219, 42-48, 2013. 10331

325

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10332 Xu X, Ray R, Gu Y, Ploehn HJ, Gearheart L, Raker K, Scrivens WA. Electrophoretic analysis and 10333 purification of fluorescent single-walled carbon nanotube fragments. J Am Chem Soc. 10334 2004;126(40):12736-7. 10335 10336 Xue C, Wu J, Lan F, Liu W, Yang X, Zeng F et al.: Nano titanium dioxide induces the generation of ROS 10337 and potential damage in HaCaT cells under UVA irradiation. J Nanosci Nanotechnol 2010, 10: 8500- 10338 8507 10339 10340 Y. Yan, K.T. Gause, M.M. Kamphuis, C.S. Ang, N.M. O'Brien-Simpson, J.C. Lenzo, E.C. Reynolds, E.C. 10341 Nice, F. Caruso. Differential roles of the protein corona in the cellular uptake of nanoporous polymer 10342 particles by monocyte and macrophage cell lines. ACS Nano 7 (2013) 10960-10970. 10343 10344 Yanagisawa, R., et al., Titanium dioxide nanoparticles aggravate atopic dermatitis-like skin lesions in 10345 NC/Nga mice. Exp Biol Med (Maywood), 2009. 234(3): p. 314-22. 10346 10347 Yanagisawa, R., et al., Size effects of polystyrene nanoparticles on atopic dermatitislike skin lesions in 10348 NC/NGA mice. Int J Immunopathol Pharmacol, 2010. 23(1): p. 131-41 10349 10350 M. Yang, K. Flavin, I. Kopf, G. Radics, C.H. Hearnden, G.J. McManus, B. Moran, A. Villalta-Cerdas, L.A. 10351 Echegoyen, S. Giordani, E.C. Lavelle. Functionalization of carbon nanoparticles modulates 10352 inflammatory cell recruitment and NLRP3 inflammasome activation. Small 9 (2013) 4194-4206. 10353 10354 Yang X, Liu J, He H, Zhou L, Gong C, Wang X et al.: SiO2 nanoparticles induce cytotoxicity and protein 10355 expression alteration in HaCaT cells. Part Fibre Toxicol 2010, 7: 1 skin. Nano Lett 2007, 7: 155-160 10356 10357 Yang Y, Gajaraj S, Wall JD, Hu Z. A comparison of nanosilver and silver ion effects on bioreactor 10358 landfill operations and methanogenic population dynamics. Water Res 2013; 47:3422-30. 10359 10360 Yang ST, Liu Y, Wang YW, Cao A. Biosafety and bioapplication of nanomaterials by designing protein- 10361 nanoparticle interactions. Small. 2013;9(9-10):1635-53. 10362 10363 Yang S-T, Wang T, Dong E, Chen X-X, Xiang K, Liu J-H, Liu Y, Wang H. (2012). Bioavailability and 10364 preliminary toxicity evaluations of alumina nanoparticles in vivo after oral exposure. Toxicology Res 10365 1, 69. 10366

326

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10367 Yang K, Gong H, Shi X, Wan J, Zhang Y, Liu Z. In vivo biodistribution and toxicology of functionalized 10368 nano-graphene oxide in mice after oral and intraperitoneal administration. Biomaterials. 10369 2013;34:2787-95. 10370 10371 Yang YX, Song ZM, Cheng B, Xiang K, Chen XX, Liu JH, Cao A, Wang Y, Liu Y, Wang H. Evaluation of the 10372 toxicity of food additive silica nanoparticles on gastrointestinal cells. J Appl Toxicol. 2014 10373 Apr;34(4):424-35. 10374 10375 A.S. Yazdi, G. Guarda, N. Riteau, S.K. Drexler, A. Tardivel, I. Couillin, J. Tschopp. Nanoparticles activate 10376 the NLR pyrin domain containing 3 (Nlrp3) inflammasome and cause pulmonary inflammation 10377 through release of IL-1α and IL-1β. Proc. Natl. Acad. Sci. U.S.A. 107 (2010) 19449-19454. 10378 10379 10380 Yoshida T, Yoshioka Y, Takahashi H, Misato K, Mori T, Hirai T, Nagano K, Abe Y, Mukai Y, Kamada H, 10381 Tsunoda S, Nabeshi H, Yoshikawa T, Higashisaka K, Tsutsumi Y. 2014. Intestinal absorption and 10382 biological effects of orally administered amorphous silica particles. Nanoscale Res Lett. 9. 10383 10384 You R, Lu W, Shan M, Berlin JM, Samuel EL, Marcano DC, Sun Z, Sikkema WK, Yuan X, Song L, Hendrix 10385 AY, Tour JM, Corry DB, Kheradmand F. Nanoparticulate carbon black in cigarette smoke induces DNA 10386 cleavage and Th17-mediated emphysema. Elife. 2015 Oct 5;4:e09623. 10387 10388 Yorkel RA and MacPhail RC, Journal of Occupational Medicine and Toxicology20116:7 10389 10390 LE. Yu, LYue, L Yung, CN Ong, YL Tan, KS Balasubramaniam, D Hartono, G Shui, MR Wenk, WY Ong. 10391 Translocation and effects of gold nanoparticles after inhalation exposure in rats,Nanotoxicology, 1:3, 10392 235-242, 2007 10393 10394 Yun JW, Kim SH, You JR, Kim WH, Jang JJ, Min SK, Kim HC, Chung DH, Jeong J, Kang BC, Che JH. 10395 Comparative toxicity of silicon dioxide, silver and iron oxide nanoparticles after repeated oral 10396 administration to rats. J Appl Toxicol. 2015;35(6):681-93. 10397 10398 Zanette C, Pelin M, Crosera M, Adami G, Bovenzi M, Larese FF et al.: Silver nanoparticles exert a long- 10399 lasting antiproliferative effect on human keratinocyte HaCaT cell line. Toxicol In Vitro 2011, 25: 1053- 10400 1060 10401

327

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10402 Zha L, Zeng J, Sun S, Deng H, Luo H, Li W (2009) Chromium (III) nanoparticles affect hormone and 10403 immune responses in heat-stressed rats. Biol Trace Elem Res 129: 157-169. 10404 10405 Zhang L, Bai R, Liu Y, Meng L, Li B, Wang L, Xu L, Le Guyader L, Chen C. The dose-dependent 10406 toxicological effects and potential perturbation on the neurotransmitter secretion in brain following 10407 intranasal instillation of copper nanoparticles. Nanotoxicology. 2012;6(5):562-75. 10408 10409 Zhang W, Kalive M, Capco DG, Chen Y. Adsorption of hematite nanoparticles onto Caco-2 cells and 10410 the cellular impairments: effect of particle size. Nanotechnology. 2010 Sep 3;21(35):355103. 10411 10412 Zhang Y, Wang J, Bai X, Jiang T, Zhang Q, Wang S. Mesoporous silica nanoparticles for increasing the 10413 oral bioavailability and permeation of poorly water soluble drugs. Mol Pharm. 2012;9(3):505-13. 10414 Zhao, Y. and K.W. Ng, Nanotoxicology in the Skin: How Deep is the Issue? Nano LIFE, 2014. 04(01): p. 10415 1440004. 10416 10417 Zhao Y, Howe JLC, Yu Z, Leong DT, Chu JJH, Loo JSC and Ng KW. Exposure to Titanium Dioxide 10418 Nanoparticles Induces Autophagy in Primary Human Keratinocytes. SMALL 9(3):387-92. 2013. 10419 10420 Zhao, Y., et al., Exposure to titanium dioxide nanoparticles induces autophagy in primary human 10421 keratinocytes. Small, 2013. 9(3): p. 387-92. 10422 10423 X. Zhi, H. Fang, C. Bao, G. Shen, J. Zhang, K. Wang, S. Guo, T. Wan, D. Cui. The immunotoxicity of 10424 graphene oxides and the effect of PVP-coating. Biomaterials 34 (2013) 5254-5261 10425 10426 Zhu L, Chang DW, Dai L, et al. DNA damage induced by multiwalled carbon nanotubes in mouse 10427 embryonic stem cells. Nano Lett. 2007;7(12): 3592-3597. 10428 10429 Zhu, S. et al., 2006 10430 10431 Zuin S, Micheletti C, Critto A, Pojana G, Johnston H, Stone V, Tran L, Marcomini A (2011) Weight of 10432 evidence approach for the relative hazard ranking of nanomaterials. Nanotox 5: 445-458. 10433 10434 Zupke O, Distler E, Baumann D, et al. Biomaterials preservation of dendritic cell function upon 10435 labeling with amino functionalized polymeric nanoparticles. Biomaterials. 31(27), 7086–7095 (2010) 10436

328

DRAFT FOR PEER REVIEW: DO NOT CITE, QUOTE OR DISTRIBUTE

10437 10438 10439 10440 10441 10442 10443 10444 10445 10446 10447 10448 10449 NOTE: an example of an earlier EHC can be found at the following link: 10450 http://www.who.int/entity/ipcs/publications/ehc/ehc236.pdf?ua=1 10451 10452 Examples of WHO risk assessment methodology documents: 10453 http://www.who.int/ipcs/publications/methods/harmonization/en/ (eg. See Document 10, on 10454 immunotoxicity)

329