IN VITRO RESILIENCE AND NANOTOXICITY IN 3D BRAIN MODELS
by Georgina Harris
A dissertation submitted to Johns Hopkins University in conformity with the requirements for the degree of Doctor of Philosophy.
Baltimore, Maryland July, 2018
© Georgina Harris 2018 All rights reserved
ABSTRACT
Current neurotoxicity testing does not meet the needs to protect human health from potential neurotoxicants. The increase in incidence of neurological disorders has shown that environmental exposures may pose a risk in conjunction with genetic factors. Pesticide exposure and aging are associated with increased Parkinson’s disease (PD) risk. To date, in vitro research focuses on apical endpoints from high-dose acute exposures. We propose to study cellular recovery and resilience in vitro, to challenge current acute toxicity testing and question (a) whether dopaminergic cells can recover from low-dose exposures and (b) how they respond to a subsequent toxicant hit. To address the current needs, we developed and characterized an in vitro human dopaminergic 3D brain model using LUHMES (Lund Human Mesencephalic cell line). Taking advantage of the fact that our model is cultured in suspension, we analyzed not only acute but also delayed response to the pesticide rotenone after compound withdrawal and 7 days recovery. Rotenone quantification demonstrated it was effectively removed from media after wash-out. We further assessed viability after second exposures to test our resilience hypothesis. Molecular and functional assays were used to assess toxicity and recover. Dopaminergic neurons were able to recover functionally (neurite outgrowth and electrical acitivty) from low-dose acute rotenone effects, however other endpoints
(complex I inhibition, gene expression) were permanently altered and pre-exposed cells were resilient to a second hit indicating long-term molecular memory after wash-out. Repeated low-dose exposures to rotenone upregulated PD-related genes. Finally, 3D LUHMES and iPSC-derived
BrainSphere model ware applied to study internalization and toxicity of nano-delivery particles
(AuSC, AuPEG and PLA). Effects on viability, mitochondrial membrane potential and oxidative response genes were observed. Our results present a different approach to studying toxicity in vitro, with the use of 3D models and compound wash-out to better understand whether acute effects are reversible (more similar to in vivo exposures). Genetic or epigenetic factors could lead to altered recovery and drive disease development. Furthermore, advances in nanotechnology require new testing strategies to assess the safety for novel drug delivery systems.
ii
THESIS COMMITTEE
Prof. Thomas Hartung Professor
Advisor Department of Environmental Health and Engineering
Johns Hopkins Bloomberg School of Public Health
Prof. Daniele Fallin Chair, Professor
Department of Mental Health
Johns Hopkins Bloomberg School of Public Health
Prof. Marsha Wills-Karp Chair, Professor
Department of Environmental Health and Engineering
Johns Hopkins Bloomberg School of Public Health
Dr. Anne Hamacher-Brady Assistant Professor
W. Harry Feinstone Department of Molecular Microbiology and Immunology
Johns Hopkins Bloomberg School of Public Health
Dr. Sin-ichi Kano Assistant Professor
Department of Psychiatry and Behavioral Sciences
School of Medicine
Dr. Wan-Yee Tang Associate Professor
Department of Environmental Health Sciences
Johns Hopkins Bloomberg School of Public Health
Dr. Jiou Wang Associate Professor
Department of Biochemistry and Molecular Biology
Johns Hopkins Bloomberg School of Public Health
iii
PREFACE
The Center for Alternatives to Animal testing (CAAT) is dedicated to the promotion, development and education of the 3Rs (reduction, refinement and replacement) in toxicology.
Since 1981, it has supported the creation, development, validation, and use of alternatives to animals in research and product safety testing by providing up to date information and acting as a hub for pharma, industry and regulatory scientists to engage in discussions about the needs in toxicology. In line with the goals of bodies such as the Organization for Economic Co-operation and Development (OECD), Environmental Protection Agency (EPA), Food and Drug administration (FDA), ICCVAM and ECVAM, the center aims to develop approaches which are more predictive of toxicity to protect human health.
Animal testing is costly, time consuming and has been shown to not be highly predictive of human adverse outcomes. Moreover, for chronic diseases, which are driven by gene-environment interactions, animal models fail to reproduce human pathology. Therefore, using novel 3D in vitro models which more closely represent human organ regions with close cell-cell interactions, we can define molecular mechanisms upon toxicant exposure and how different genetic backgrounds may impact neuronal degeneration. Once the 3D dopaminergic model was developed, we attempted to study neuronal recovery. Most in vitro research to date has been performed at high concentrations, measuring acute effects. To better recapitulate life exposures, we treated a 3D dopaminergic model acutely, but washed out the compound to determine whether these effects are relevant in long-term toxicity or whether cells can cope with low-dose effects.
We further studied the effects of low, repeated-dose in the 3D model determining that chronic exposure using a 3D model is more relevant to study neurodegenerative mechanisms in vitro.
The work in this PhD thesis was aimed at researching (i) current neurotoxicity testing and in vitro models, (ii) evidence showing how environmental exposures are a risk factor for
iv neurodegenerative diseases such as Parkinson’s and (iii) novel nanoparticle drug-delivery systems to treat neurodegenerative diseases. The scientific papers included in this thesis present Georgina
Harris’ contributions to this field; the development and use of a 3D in vitro dopaminergic model to study acute, delayed and repeated-dose effects, neuronal resilience and the use of 3D models for nanotoxicity testing.
During my PhD have been very fortunate to have had experienced mentors with a strong drive for alternative methods. My first and upmost gratitude is to my parents Daphne Lopez and Donald
Harris who provided me with all the tools and education to reach my goals. They have always been supportive of my choices and have been there to help me when it was needed. I would never achieve any of this without you. David Pamies who lept on this adventure with me in the US, he has been the best co-worker I could have asked for and has taught me patience, persistence and how we can overcome any hurdles together. Thank you for all the support these years and for taking care of me always. My sisters Carla and Vanessa, for putting up with me and always being so caring. My family in Elche, for always checking in on me, cheering me on throughout the years, visiting us in Baltimore and making me the best meals when I am in town. Thanks to my advisor, Prof. Thomas Hartung, whom I will always admire every time I attend one of his presentations. Ever since our fist meeting he has tried to open doors for my future and enabled networking and travel to further educate me. I am forever grateful for the opportunities and advice he has provided me with. I am grateful for the International Foundation for Ethical
Research who provided four years of funding to aid my PhD as well as funding for my research.
As one of the few institutions that provides fnding to non-US citizens, I am very thankful for the opportunity and support to foreign students. To my thesis committee, who met with me multiple times over the past three years, thank you for your constructive feedback and for guiding me along this path. I would also like to thank my second advisor, Dr. Lena Smirnova who taught me lab techniques and allowed me to develop scientific research independence. Thank you for always
v having time for me and helping me when I needed it. Shelly Odwin, for keeping our lab organized, supporting me, making me laugh, working with me in the summer and teaching me patience and that one can always achieve new goals (now it is my turn to support you). Dr.
Helena Hogberg, whom the lab would not exist without. Thank you for always having the time to talk and give me advice, your PhD thesis was an inspiration for me and you kept me going when I needed support. To Mounir, Andre and Alex, thank you for your help with metabolomics and transcriptomics data, your knowledge is priceless and helped me troubleshoot data throughout the years. To fellow students whom I have trained or worked closely with, Dana, Melanie, Mariana,
Johannes, Marize, Rober, Vy and Daphne, you have been my greatest inspiration and sense of achievement. Teaching you and learning from you all bring memories of the best moments of my
PhD. To Ruth, Jamie and Michelle, thank you for your patience and hard work every day, your friendliness and caring attitude provides CAAT with a lovely working environment. To my collaborators, my research would not have been possible without you and your expertise. Thank you for working hard on the vision I had and the experiments I proposed.
Throughout my PhD, some mentors went above and beyond to enquire about my research and provide constructive criticism. Dr. Milena Mennecozzi, Dr. Brigitte Landesmann and Prof. Anna
Price are my most inspiring women in science who have believed in me and helped me grow as a scientist. Dr. Maurice Whelan, Prof. Marcel Leist, Dr. Jack Fowle, Prof. Martin Wilks and Dr.
Pedro DelValle have also strongly influenced my growth and a scientist and have always had the time for me, providing me with opportunities throughout the years. I am lucky to have had such strong role models.
To my friends, thank you for always being there for the good and bad times. Moving to a different continent was worth it to meet you all and share beautiful moments. Minipandi and
Charm, you are my second family and I’m lucky to have your friendship.
vi
Relevant co-authored publications not integrated in this thesis:
Pamies D, Block K, Lau P, Gribaldo L, Pardo CA, Barreras P, Smirnova L, Wiersma D, Zhao L,
Harris G, Hartung T, Hogberg HT. Rotenone exerts developmental neurotoxicity in a human brain spheroid model. Toxicol Appl Pharmacol. 2018. pii: S0041-008X(18)30042-5. doi:
10.1016/j.taap.2018.02.003.
Pamies D, Barreras P, Block K, Makri G, Kumar A, Wiersma D, Smirnova L, Zang C, Bressler J,
Christian KM, Harris G, Ming GL, Berlinicke CJ, Kyro K, Song H, Pardo CA, Hartung T,
Hogberg HT. A human brain microphysiological system derived from induced pluripotent stem cells to study neurological diseases and toxicity. ALTEX. 2017;34(3):362-376. doi:
10.14573/altex.1609122.
Leist M, Hasiwa N, Rovida C, Daneshian M, Basketter D, Kimber I, Clewell H, Gocht T,
Goldberg A, Busquet F, Rossi AM, Schwarz M, Stephens M, Taalman R, Knudsen TB, McKim J,
Harris G, Pamies D, Hartung T. Consensus report on the future of animal-free systemic toxicity testing. ALTEX. 2014; 31(3):341-56. doi: http://dx.doi.org/10.14573/altex.1406091.
Bouhifd M, Beger R, Flynn T, Guo L, Harris G, Hogberg H, Kaddurah-Daouk R, Kamp H,
Kleensang A, Maertens A, Odwin-DaCosta S, Pamies D, Robertson D, Smirnova L, Sun J, Zhao
L, Hartung T. Quality assurance of metabolomics. ALTEX. 2015;32(4):319-26. doi:
10.14573/altex.1509161.
Hogberg HT, Bressler J, Christian KM, Harris G, Makri G, O'Driscoll C, Pamies D, Smirnova L,
Wen Z, Hartung T. Toward a 3D model of human brain development for studying gene/environment interactions. Stem Cell Res Ther. 2013;4 Suppl 1:S4. doi: 10.1186/scrt365.
vii
TABLE OF CONTENTS
ABSTRACT ...... II
PREFACE ...... IV
LIST OF TABLES, WITH TITLES AND PAGE REFERENCES ...... XIII
LIST OF FIGURES, WITH TITLES AND PAGE REFERENCES ...... XIV
1. INTRODUCTION ...... 1
1.1. NEUROTOXICITY TESTING ...... 1
1.1.1. TOWARDS 3D NEURONAL IN VITRO MODELS ...... 2
1.1.2. IN VITRO NEUROTOXICITY ENDPOINTS...... 7
1.1.3. NEUROTOXICOLOGY OF NANOPARTICLE DRUG DELIVERY SYSTEMS ... 13
1.2. PESTICIDE EXPOSURE AND PARKINSON’S DISEASE ...... 16
1.2.1. PUBLIC HEALTH SIGNIFICANCE ...... 17
1.2.2. PARKINSON’S DISEASE PATHOPHYSIOLOGY ...... 19
1.2.3. MOLECULAR MECHANISMS IN PARKINSON’S DISEASE ...... 20
1.2.4. IN VIVO AND IN VITRO PARKINSON’S DISEASE MODELS ...... 27
1.2.3.1 LUND HUMAN MESENCEPHALIC (LUHMES) CELL LINE ...... 30
1.2.3.2. INDUCED PLURIPOTENT STEM CELL (IPSC)-DERIVED NEURONS ...... 31
1.2.5. ROTENONE AS A PD MODEL COMPOUND ...... 33
1.2.6. GLYPHOSATE EXPOSURE: PUBLIC HEALTH CONTROVERSY ...... 36
1.3. 21ST CENTURY TOXICOLOGY ...... 37
1.3.1. ADVERSE OUTCOME PATHWAYS (AOPS) ...... 39
1.3.1. REPEATED LOW-DOSE EFFECTS ...... 40
1.4. CELLULAR RESILIENCE CONCEPT IN TOXICOLOGY ...... 41
viii
1.5. RESEARCH PRESENTED IN THE THESIS ...... 44
2. CELLULAR RESILIENCE ...... 48
2.1. ABSTRACT ...... 49
2.2. INTRODUCTION ...... 50
2.2.1. CONSIDERATION 1: IT IS NOT IMPORTANT WHETHER YOU FALL, BUT
WHETHER YOU GET UP AGAIN ...... 53
2.2.2. CONSIDERATION 2: ANASTASIS – AWAKEN FROM THE DEAD ...... 55
2.2.3. CONSIDERATION 3: ALL CELLS ARE EQUAL(LY VULNERABLE) ...... 56
2.2.4. CONSIDERATION 4: KINETICS CANNOT EXPLAIN ALL ORGAN-
SELECTIVITIES ...... 59
2.2.5. CONSIDERATION 5: ARE DIFFERENCES IN CELLULAR RESILIENCE
UNDERLYING ORGAN-SELECTIVITY OF TOXICANTS? ...... 60
2.2.6. CONSIDERATION 6: HOW TO CHALLENGE THE CONCEPT? ...... 64
2.2.7. CONSIDERATION 7: RESILIENCE IS NOT THE RETURN TO THE PRIOR
STATE ...... 67
2.3. CONCLUSIONS ...... 71
2.4. ACKNOWLEDGEMENTS ...... 72
2.5. REFERENCES ...... 72
3. 3D DIFFERENTIATION OF LUHMES CELL LINE TO STUDY RECOVERY AND
DELAYED NEUROTOXIC EFFECTS ...... 89
3.1. ABSTRACT ...... 91
3.2. INTRODUCTION ...... 91
3.3. MATERIALS METHODS ...... 93
ix
3.3.1. BASIC PROTOCOL 1: LUHMES DIFFERENTIATION IN 3D ...... 93
3.3.2. BASIC PROTOCOL 2: COMPOUND TREATMENT AND WASH-OUT ...... 106
3.4. TABLES AND FIGURES ...... 123
3.5. DISCUSSION ...... 129
3.6. ACKNOWLEDGEMENTS ...... 132
3.7. REFERENCES ...... 132
4. LUHMES 3D DOPAMINERGIC NEURONAL MODEL FOR NEUROTOXICITY
TESTING ALLOWING LONG-TERM EXPOSURE AND CELLULAR RESILIENCE
ANALYSIS...... 138
4.1. ABSTRACT ...... 140
4.2. INTRODUCTION ...... 141
4.3. MATERIALS AND METHODS ...... 146
4.4. RESULTS ...... 156
4.5. DISCUSSION ...... 173
4.6. ACKNOWLEDGEMENTS ...... 178
4.7. REFERENCES ...... 179
5. TOXICITY, RECOVERY AND RESILIENCE IN A 3D DOPAMINERGIC IN VITRO
MODEL EXPOSED TO ROTENONE...... 187
5.1. ABSTRACT ...... 188
5.2. INTRODUCTION ...... 189
5.3. MATERIALS AND METHODS ...... 193
5.4. RESULTS ...... 199
x
5.5. TABLES AND FIGURES ...... 206
5.6. DISCUSSION ...... 214
5.7. ACKNOWLEDGEMENTS ...... 223
5.8. REFERENCES ...... 224
6. EFFECTS OF REPEATED LOW-DOSE EXSPOSURE TO ROTENONE AND
GLYPHOSATE IN THE 3D LUHMES MODEL...... 238
6.1. MATERIALS AND METHODS ...... 239
6.2. RESULTS ...... 241
6.3. FIGURES ...... 242
7. IMPACT OF GOLD AND POLY-LACTIC ACID NANOPARTICLES ON 3D HUMAN
BRAIN SPHEROID MODELS: STUDYING BIOCOMPATIBILITY FOR BRAIN DRUG
DELIVERY...... 245
7.1. ABSTRACT ...... 247
7.2. INTRODUCTION ...... 248
7.3. MATERIALS AND METHODS ...... 250
7.4. RESULTS ...... 257
7.5. FIGURES ...... 262
7.6. DISCUSSION ...... 269
6.7. CONCLUSIONS ...... 274
6.7 ACKNOWLEDGEMENTS ...... 275
6.8. REFERENCES ...... 276
8. SUMMARIZING DISCUSSION ...... 284
xi
9. REFERENCES ...... 306
10. APPENDICES ...... 336
10.1. APPENDIX I SUPPLEMENTARY FIGURES CHAPTER 4 ...... 337
10.2. APPENDIX II. SUPPLEMENTARY FIGURES CHAPTER 5...... 340
11. CURRICULUM VITAE GEORGINA HARRIS (MSC, BSC) ...... 355
xii
LIST OF TABLES
Table 1-1. Studies which have found an association between rotenone or paraquat exposure and PD risk...... 19
Table 1-2. Cell lines used as dopaminergic cell models – table updated and modified from
(Lazaro, Pavlou, and Outeiro 2017).…………………………………………………...... 30
Table 3-1. Flask coating solution ………………………………………………………. 123
Table 3-2. LUHMES medium composition…………………………………………….. 123
Table 3-3. Solutions for immunocytochemistry and flow cytometry…………………… 124
Table 3-4Table 3-4. Antibody dilutions for immunocytochemistry…………………… .124
Table 3-5. Marker genes for 3D differentiation quality control………………………… 125
xiii
LIST OF FIGURES
Figure 1-1. Differentiation of NT2 (Eaton and Wolfe 2009), IPSC-derived cortical neurons
(www.axolbio.com), and LUHMES cells (photographed by Helena T Hogberg, Center for
Alternatives to Animal Testing) in 2D culture showing clustering and aggregate-like formations...... 4
Figure 1-2. Building 3-dimensional brain organoids by Sergiu. P. Pasca, 2017……………….….. 5
Figure 1-3. iPSC-derived 3D models developed by (A) Lancaster et al, 2013 (B) Pamies et al, 2016
(C) Qian et al, 2018………………………………………………………………………………… 6
Figure 1-4. NP features which influence systemic delivery and blood brain barrier (BBB) passage.
Published by (Saraiva et al. 2016)………………………………………………………………….15
Figure 1-5. Number of publications per year using search term ‘nanoparticles AND
Parkinson’s’...... 16
Figure 1-6. Most likely candidate genes that were significantly associated with PD. Black or grey text indicate known loci and red text indicates novel loci significantly associated with PD, published by Chang et al, 2017...... 22
Figure 1-7. The pathway of dopamine synthesis, release, metabolism and receptors published by
(Youdim, Edmondson, and Tipton 2006)…………………………………………………….….....26
Figure 1-8. Epigenetic processes in familial and sporadic PD published by (Ammal Kaidery,
Tarannum, and Thomas 2013)…………………………………………………………….………..27
Figure 1-9. NPCs generate different cell types of the nervous system (www.sigmaaldrich.com/life- science/stem-cell-biology/neural-stem-cell-biology.html)...... 33
xiv
Figure 1-10. Rotenone molecular structure………………………………………………….…… 34
Figure 1-11. Rat brain slices, showing complex I inhibition by rotenone using competitive binding with labelled dihydro-rotenone (a) and respiration rate in liver and brain mitochondria exposed to rotenone (b) published by (Betarbet et al. 2000)…………………………………………………..35
Figure 1-12. Molecular structure of glyphosate…………………………………..………………. 37
Figure 1-13. Cellular resilience concept published in (Smirnova et al. 2015)……………..………43
Figure 2-1. The Cellular Resilience Concept: Survivable toxic insults create cellular stress;
Pathways of Defense (PoD) might allow cells to return to a normal state; however, resilience programs often leave cells in an altered state, e.g. with an epigenetic scar, which might contribute to long-term manifestations of hazard but could also be a target for therapeutic strategies...... 53
Figure 3-1. Suggested treatment schedule for sample collection on day 8. In this scheme, 48 hr treatment takes place on day 6 of differentiation, while 24, 12, and 6 hr treatments take place on day 7 of differentiation. All treated and control samples are collected at the same time on day 8, at the same stage of differentiation...... 125
Figure 3-2. Pipetting schematic for serial dilution of test compound...... 126
Figure 3-3. Plate layout for concentration range finder experiments and dose-response curves.
Vehicle control (DMSO) and compound concentrations are tested across four replicates, where C1 is the lowest and C5 is the highest compound concentration...... 126
Figure 3-4. Screenshot for ImageJ software with uploaded fluorescence image of an aggregate.
Oval shape selection is used to select aggregate area for measurement of mean grey values (using the commands Analysis, Measure [Ctrl+M])...... 127
xv
Figure 3-5. High-content analysis (HCA) mask settings to identify aggregates (nuclear mask shown in blue outlining bright aggregate) and individual neurite outgrowth (neurite mask shown in purple outlining neurites) observed after 24 hr incubation in matrigel-coated wells...... 127
Figure 3-6. Time consideration scheme for Basic Protocols 1 (LUHMES Neuronal Differentiation in 3D) and 2 (Compound Treatment and Washout)...... 128
Figure 3-7. Time consideration scheme for the toxicological endpoints described in Support
Protocols 1 (immunostaining), 2 (flow cytometry), 5 (ATP assay), 6 (MitoTracker), and 7 (neurite outgrowth quantification)...... 128
Figure 4-1. Adaptation and optimization of 3D LUHMES differentiation protocol. (a) The original
2D differentiation (2D diff protocol) was adapted for 3D culture (3D diff protocol) by subjecting the single-cell suspension to continuous gyratory shaking. Protocol 3D pre-diff involves a pre-differentiation step in 2D for 2 days, trypsinization, and subsequent cultivation in 3D. Further optimization involved adding the anti-proliferation compound taxol on day 3 for 48 h (10 nM, 48 h) to reduce cell proliferation (3D + T10 protocol). Toxic compound treatment took place between days 6 and 8 for 12, 24, and 48 h, reversely.
Samples for toxicological end points were collected immediately after treatment on day 8 or on day 15 after washout of compounds and 7 days recovery. (b) Differentiation of LUHMES in monolayer. As differentiation advances, cells may detach from the surface (d9, last photograph). (c) Aggregate formation under continuous gyratory shaking (80 rpm). Note the increasing size of aggregates in course of differentiation. (d) Treatment of 3D cultures with anti-proliferation drug taxol (10 nM) for 48 h blocked the proliferation and slowed down aggregate growth. (e) Measurements of aggregate size during differentiation following different 3D protocols. At least 100 aggregates were measured per condition, per day, in at least three independent experiments, with exception of 3D pre-diff. and 3D gradient speed, where only one independent experiment was conducted. Data represent mean ± SEM, n ≥ 3 (independent experiments). Scale bars are 200 μm.……………………………………………………………………………..……..148
xvi
Figure 4-2. Quantification of apoptosis and necrosis in 3D LUHMES model. Caspase 3/7 activation
(green nuclei) as an early apoptotic marker was visualized using fluorescent microscopy in combination with Hoechst 33342 staining of nuclei (blue) in undifferentiated LUHMES monolayer cultures (d0) and 21 days after induction of differentiation in 3D. Scale bars are 50 μm. b Annexin
V/7-AAD-positive cells were quantified using flow cytometry on day 0 (as negative control) and days 12, 15, and 21 following 3D diff. protocol for differentiation. Aggregates exposed to 0.5 μM rotenone for 48 h and after 7 days recovery were used as a positive control (last panel). Annexin V- positive cells are in early apoptosis; double stained for Annexin V and 7AAD cells are in later apoptotic phases, while 7-AAD-positive cells represent a population of necrotic cells. Data are shown as mean ± SEM, n ≥ 3 (independent experiments) c Penetration assay with Hoechst 33342:
12-day-old aggregates of GFP-expressing LUHMES were differentiated according to 3D diff protocol and were stained for increasing time intervals with Hoechst 33342. Confocal optical slices through the center of the aggregates are shown to demonstrate time-dependent penetration of
Hoechst 33342 through the aggregates. Compare GFP expression (green) at all time points in the center of aggregates with the absence of Hoechst 33342 staining (blue) after 5, 15, and 30 min of incubation and penetration of Hoechst 33342 to the middle after 1 and 6 h of incubation. No apoptotic nuclei are visible in the center of aggregates (60 min and 6 h). Scale bar is 100 μm (color
Figure online)...... 157
Figure 4-3. Estimation of proliferation rate within the aggregates. a Percentage of Ki-67-positive cells on days 6 and 12 of differentiation with or without taxol in comparison with undifferentiated
LUHMES (d0). The number of Ki-67-positive cells was measured using Alexa Flour 647- conjugated anti-Ki-67 antibody by flow cytometry. Data represent mean ± SD, n ≥ 3 (independent experiments). b Ki-67 gene expression on days 0, 6, and 12 of differentiation in 3D diff and 3D +
T10 cultures. Data are normalized to Ki-67 expression on d0 and represent mean ± SEM, n ≥ 3
(independent experiments). c Immunostainings of 3D diff and 3D + T10 aggregates with antibody
xvii against KI-67 showing prolonged presence of Ki67-positive cells (red) in 3D diff cultures in comparison with 3D + T10 aggregates. The aggregates were co-stained with postmitotic neuronal marker NeuN (green). The nuclei were visualized with Hoechst 33342 staining. Scale bars are 50
μm. The aggregates were fixed on glass slides and covered with coverslips for confocal imaging, which explains the larger size of the aggregates in comparison with Fig. 1, where floating aggregates were imaged...... 161
Figure 4-4. Immunocytochemistry of neuronal differentiated LUHMES in 3D diff and 3D + T10 cultures on days 6, 12, 15, and 21 after induction of differentiation. Panel a shows the induction of expression of MAP2 (green) and NF200 (red), as well as maturation and neurite elongation in 3D cultures followed 3D +T10 protocol versus 3D diff protocol over a span of 21 days of differentiation. Panel b shows the overlay of NF200 (red) and typical punctual staining with synaptic marker, synaptophysin (Syn, green). Note higher synaptophysin expression in 3D + T10 cultures versus 3D diff cultures. The nuclei were visualized with Hoechst 33342 staining. Scale bars are 50 μm. The aggregates were fixed on glass slides and covered with coverslips for confocal imaging which explains the larger size of the aggregates in comparison with Fig. 1, where floating aggregates were imaged...... 163
Figure 4-5. Enhanced neuronal maturation in 3D + T10 cultures. a MAP2 staining of representative aggregates differentiated for 12 days following either 3D +T10 (first panel) or 3D diff. (second panel) protocols. The nuclei were visualized with Hoechst 33342 staining. b Higher magnification
(63×) of representative aggregates differentiated for 12 days under 3D + T10 conditions and stained with synaptophysin (green), NF200 (red) in the first slide and MAP2 (green), NF200 (red) in the second slide. The nuclei were visualized with Hoechst 33342 staining. Scale bars are 50 μm. Real-
Time RT-PCR of genes involved in LUHMES neuronal differentiation and maturation. LUHMES were differentiated in 3D + T10 (c) and 2D monolayer cultures (d). RNA samples were collected on days 3, 6, 9, 12, 15, and 21 of differentiation and prior induction of differentiation (day 0) as a
xviii control. Data represent mean of log2 (fold change) ± SEM normalized to d0 from at least four independent experiments. Statistical significance was calculated using one-way ANOVA test followed by Dunnett’s post hoc test. Expression of all genes was significantly (p < 0.05) different in comparison with day 0, except Nestin in 3D cultures and Ki-67 in 2D cultures (Supplementary
Table S2) (color Figure online)…………………………………………..……………………….164
Figure 4-6. Cell viability of LUHMES aggregates after exposure to rotenone and MPP+. LUHMES cells were differentiated following 3D + T10 protocol and exposed reversely to different rotenone
(a) and MPP+ (b) concentrations from day 6 to 8 (48 h) and from day 7 to 8 (24 h). c Cell viability after toxicant washout and recovery period. LUHMES were exposed reversely until day 8 for indicated period of time to 0.1 μM rotenone and 5 μM MPP+. On day 8, compounds were washed out and cells recovered for further 7 days. Chronic/repeat-dose (fresh substance was added with each medium exchange) exposure (192 h, from day 7 until day 15) was included as positive control. Cell viability was analyzed using resazurin reduction assay. Cell viability is presented in
% of solvent-treated controls in at least three independent experiments (n ≥ 3, mean ± SEM, n = 2 for MPP+ on day 15). d Mitochondrial membrane potential in individual LUHMES aggregates, exposed to rotenone for 48 h from day 6 to 8 measured by Mitotracker assay. Fluorescence intensity was measured as mean gray values using ImageJ software and normalized to the size and then to the fluorescence intensity of DMSO control aggregates (n = 3), at least 10 aggregates were assayed for each independent experiment, ***p < 0.001, Kruskal–Wallis followed by Dunn’s post hoc test)...... 166
Figure 4-7. Neurite integrity in individual LUHMES aggregates exposed to rotenone. a Exposure to
0.1 μM rotenone for 48 h from d6 to d8 perturbed neurite integrity. Confocal images showing RFP- expressing cells (red) mixed in a 1:49 ratio with wild-type cells and differentiated following the 3D
+ T10 protocol for 8 days. Note rotenone-altered neurite integrity of viable RFP-expressing cells in comparison with DMSO control samples. Nuclei are stained with Hoechst 33342. Scale bars are 50
xix
μm. b Quantification of neurite area in rotenone-treated samples versus DMSO controls, normalized to the number of RFP-positive cell bodies in three independent experiments (nine aggregates were quantified for rotenone-treated samples and 12 for DMSO control samples) (n = 3,
**p < 0.01, Mann–Whitney test)...... 169
Figure 4-8. Time-dependent perturbations of gene expression after exposure of 3D LUHMES to rotenone. a Toxicant treatment and washout scheme: LUHMES were differentiated following 3D +
T10 protocol; exposure to 0.1 μM rotenone occurred for 12 or 24 h from day 7 until day 8; Samples were collected for RT-PCR immediately after exposure on day 8 (dark arrow) or after rotenone washout and 7 days recovery on day 15 (light arrow). b Protein coding genes (ASS1, AT4, CTH,
SHMT2) and miRNA (mir-7) with counter-regulation pattern after rotenone washout in comparison with acute response. c Protein coding genes (MLF1IP, TYMS) and miRNA (mir-16) with stronger response after rotenone withdrawal in comparison with acute toxicity. Dark bars show expression of the genes on day 8, while light bars show expression of the genes after rotenone washout and 7- day recovery. The data are means of log2 (fold change) ± SEM of at least three independent experiments (9–12 technical replicates). (n ≥ 3, *p < 0.05, **p < 0.01, and ***p < 0.001, one-way-
ANOVA followed by Dunnett’s post hoc test) ...... 171
Figure 5-1. LUHMES 3D model for acute, recovery and resilience experiments. (a) LUHMES differentiated in 3D on a gyratory shaker showing (b) RFP-expressing cells (RED) and TH (green).
(c) LUHMES 3D treatment and wash-out scheme for recovery and resilience (second hit) experiments and endpoints. (d) Media rotenone quantification prior to treatment, day 8 and day 15.
From left to right bars correspond to negative control (media without rotenone), positive control
(media with rotenone prior to treatment), 24 h treatment control (media with rotenone in plates), 24 h treatment (media with rotenone in plates with aggregates), 7-day wash-out control (media with rotenone in plates on day 15 after wash-out) and 7-day wash out treated cells (media with rotenone
xx in plates on day 15 after wash-out with aggregates). (e) Amount of rotenone bound to plastic and cells after 24 h exposure (day 8)...... 206
Figure 5-2. 3D LUHMES viability after wash-out. (a) Cell viability measured over time using resazurin assay on days 8 (after 24h treatment) and 10, 12, and 15 (throughout recovery). (b)
Cytotoxicity over time during recovery measured by LDH release on days 8 (after 24h treatment) and 10, 12, and 15 (throughout recovery). (c) Protein concentration on day 15 after wash-out and 7 days recovery. (d) DNA quantification on day 15, after wash-out and 7 days recovery. All data were normalized to untreated control cells and are displayed as means ± SEM from 3 independent experiments. *p < 0.05……………………………………...... 207
Figure 5-3. Effects of rotenone on complex I activity and ATP levels. (a) Complex I activity after rotenone exposure (day 8) and after compound wash-out and recovery (day 15) in control and treated samples. (b) ATP levels after rotenone exposure (24 h, day 8) or after wash-out and 7 days recovery period (day 15) in control and treated samples. Differences in treated and control samples from at least three independent experiments were analyzed for statistical significance using unpaired Student’s t-test. A p-value < 0.05 is denoted on graphs by * and p < 0.0001 by ****, respectively…...... 208
Figure 5-4. TEM analysis of mitochondria after rotenone exposure and wash-out. Legend: M, mitochondria; G, Golgi complex; L, lipid droplets; N, nucleus; and NN, neurite. The number (a) and diameter (b) of mitochondria from random image areas were quantified on day 8 (24h) and day
15 (wash-out). Data from 20 random images from three 3 independent experiments is shown as well as means ± SD. Differences between treated and untreated samples were analyzed for statistical significance using unpaired Student’s t-test. A p Value < 0.05 is denoted on the graphs by
*. (c) Representative images are shown with arrows indicating morphological alternations to the mitochondrial membrane…………………………………………………………...……………..209
xxi
Figure 5-5. Image J Sholl analysis of neurite outgrowth after rotenone exposure (day 8) and wash- out (day 15). RFP-LUHMES aggregates were grown on Matrigel® on day 8 or day 15. (a)
Representative images for the different conditions are shown. (b) Sholl analysis (Image J) was used to calculate the number of neurites at different distances from the aggregate center on day 8 and day 15 from three independent experiments (5 individual aggregates per experiment). Curves were compared using a quadratic non-linear regression fit with confidence intervals……………………………………………………………………………………………210
Figure 5-6. 3D LUHMES electrical activity on day 15 after acute exposure on day 8 and compound wash-out. (a) Photo microscopy image of a 3D LUHMES aggregate attached to a glass pipette and a patched cell at a higher magnification. Cells on different aggregates were patched in three independent experiments. (b) Firing pattern of a representative tonic (top) and a phasic (middle) cell with voltage responses to 1 s current injections (bottom) at 4, 8, 12, 16, 20 and 24 pA, (c) Total number of tonic and phasic cells in control and treated samples on day 15 (p=0.695 two-sided
Fisher's exact test), (d) the Input resistance (Rm) of the phasic cells (p=0.963 two-sided Mann-
Whitney U Test) and (e) Minimal Spiking Latency of phasic cells (p=0.852 two-sided Mann-
Whitney U Test). Error bars represent SEM.211
Figure 5-7. Rotenone-induced transcriptome changes on day 8 (24h) vs. day 15 (wash-out). (a)
Volcano plots show the global changes in transcriptome for day 8 and day 15. (b) Venn-diagram shows the number of up- and down-regulated genes on day 8 (D8 (24h)) and on day 15 (D15
(wash-out)) (FC > 1.5, p < 0.01). 10 genes were in intersection between two conditions, which are listed in (c). For this diagram, the p-values were not adjusted for multiple testing. ACTA1 – actin alfa 1, skeletal muscles, PPP1R27 - Protein phosphatase 1, regulatory subunit 27, GDF15 - Growth differentiation factor 15, CCK – cholecystokinin, CD200 – OX-2 membrane glycoprotein, LCP1 – plastin 2 (lymphocyte cytosolic protein 1), ZFHX4 -AS1 – ZFHX4 (Zinc Finger Homeobox 4) antisense RNA 1, FRMPD2 - FERM and PDZ domain containing 2, FRMPD2 - FERM and PDZ
xxii domain containing 2, GRXCR1 - glutaredoxin and cysteine rich domain containing
1…………………………………………………………………………………………….…….212
Figure 5-8. (a) and (b) Cell viability concentration-response for aggregates on day 15, which were pre-exposed to DMSO (Control) or rotenone (pre-exposed 100 nM or 50 nM) on day 8. (c) LDH- release dose-response for aggregates on day 15 that were pre-exposed to DMSO (control) or rotenone (pre-exposed 100 nM) on day 8. Dose-response curves were generated from three independent experiments and analyzed by one-way ANOVA followed by Bonferroni’s correction.
(d) NEF2L2, ATF4, EAAC1, (e) DAT, CASP3 and (f) TYMS, MLF1IP gene expression by QT-
PCR from three independent experiments analyzed using the Student’s t-test and Bonferroni’s correction for multiple hypothesis testing. A p-value < 0.05 is denoted by *, p < 0.01 by **, and p <
0.001 by ***, respectively …………………………………………………….………………..213
Figure 6-1. Viability assays for repeated-dose toxicity of rotenone and glyphosate. Resazurin assay time-course for reonone (a) and day 15 viability for glyphosate (b). LDH assays for rotenone and glyphosate time-course experiments (c and d)...... 242
Figure 6-2. Effects of repeated-dose rotenone on ATP levels (a) and neurite outgrowth
(b)...... 243
Figure 6-3. Changes in gene expression for PD-related genes (TH, PARK2, PARK7 and ASS1) after repeated-dose exposures to rotenone...... 244
Figure 7-1. NP characterization. Representative images obtained by TEM of (A) Au-SC, (B) Au-
PEG and (C) PLA-NP. (D) Quantification of NP size by TEM. (E) Z‑average hydrodynamic diameter values of 6 µg/mL Au-SC and 20 µg/mL PLA-NP diluted in LUHMES and BrainSpheres culture media, analyzed by DLS after 0, 24 and 72 h of incubation. Results are expressed as mean
(±SD). Each experimental group corresponds to the analysis of three independent experiments with three replicates. Statistical significance was analyzed by one-way ANOVA followed by
xxiii
Bonferroni’s multiple comparisons post-test (**p < 0.01, ***p < 0.001). Scale bar (A) 100 nm, (B)
25 nm and (C) 200 nm...... 262
Figure 7-2. Confocal Images of (A) RFP-expressing LUHMES after 7 days of differentiation; and
4-week BrainSpheres expressing different neural cell type markers: (B) neuronal MAP2, (C) astrocytic GFAP and (D) oligodendrocyte-specific Olig1. Scale bars 100
µm...... 263
Figure 7-3. NP uptake by 3D LUHMES and BrainSpheres. Representative images acquired by confocal microscope after 72 h exposure to 20 µg/mL PLA-NP (green) of (A) BrainSpheres stained with MAP-2, GFAP, OLIG-1 (red) and Hoechst33342 (blue nuclei); and (B) 3D RFP-LUHMES
(red) (left panels). Right panels correspond to magnification of left panel images. Cross-hair lines in the Z-stack images evidence PLA-NP internalization. (C) Flow cytometry Scatter plots of wild type (WT) LUHMES (negative control), and RFP-LUHMES, treated with 0, 0.2, 2 and 20 µg/mL
PLA-NP. Red and green fluorescence was quantified in FL2 and FL1 channels, respectively. (D)
Intracellular levels of Au in 3D LUHMES and BrainSpheres after 24 and 72 h exposure to 6 µg/mL
Au-SC and 20 µg/mL Au-PEG quantified by ICP-MS. Results are expressed as mean (±SEM).
Each experimental group corresponds to the analysis of three independent experiments with three replicates. Student’s t-test was used to compare 24 and 72 h treatment groups (*p < 0.05, ***p <
0.001). Scale bars 100 µm (left panels) and 25 µm (right panels)...... 264
Figure 7-4. NP effect on MMP and cell viability in 3D human neural models. (A) Representative images of MitoTracker® staining in 3D LUHMES and BrainSpheres after 24 and 72 h exposure to
Au-SC (6 µg/mL), Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL). (B) MMP levels in 3D
LUHMES and BrainSpheres after 24 and 72 h exposures to Au-SC (0.06, 0.6 and 6 µg/mL), Au-
PEG (0.2, 2, 20 µg/mL) and PLA-NP (0.2, 2 and 20 µg/mL) normalized to the untreated controls.
xxiv
(C) Percent of LDH release in 3D LUHMES and BrainSpheres models after 24 and 72 h exposures to Au-SC (6 µg/mL), Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL). ‘- Control’ corresponds to
LDH released from untreated 3D human CNS models. Results are expressed as mean (±SEM).
Each experimental group corresponds to three independent experiments imaging at least 20 spheroids. One-way ANOVA with Bonferroni’s multiple comparisons post-test was used for analysis of statistical significance (*p < 0.05, **p < 0.01, ***p <
0.001)...... 265
Figure 7-5. Morphology of 3D LUHMES and BrainSpheres exposed to Au-SC (6 µg/mL), Au-PEG
(20 µg/mL) and PLA-NP (20 µg/mL) for 72 h. White arrowheads indicate cells in degeneration.
Scale bars 10 µm...... 266
Figure 7-6. Effect of NP on expression of genes related to ROS regulation in BrainSpheres. Graphs showing the relative expression of SOD1, SOD2, NF2L2, GSTO1, NFR1 and CLEC7A after exposure to Au-SC (6 µg/mL), Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL) for 72 h. Data was collected from three independent experiments with three technical replicates and represents fold changes (± SEM). One-way ANOVA with Bonferroni’s multiple comparisons post-test was used to analyze the statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001)...... 266
Figure 7-7. NP influence release of chemokines, cytokines and growth factors in 3D human neural models. Graphs showing the levels of different secreted mediators after exposure to Au-SC (6
µg/mL), Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL) for 24 and 72 h. (A) 3D LUHMES (MIP-
1β, IL-10, IL-12p70, TNFα, bFGF and VEGF) and (B) BrainSpheres (IL-1ra, IL-10, IL-12p70,
GM-CSF, bFGF and VEGF). Data were collected from three independent experiments with three technical replicates and represents mean (± SEM). One-way ANOVA with Bonferroni’s multiple comparisons post-test was used to analyze the statistical significance (*p <
0.05).…………………………………………………………………………………………..…267
xxv
Figure 7-8. Influence of NP on release of TGF-β isoforms in BrainSpheres. Graphs show the levels of secreted TGF-β1, TGF-β2 and TGF-β3 after exposure to Au-SC (6 µg/mL), Au-PEG (20
µg/mL) and PLA-NP (20 µg/mL) for 24 and 72 h. Each experimental group corresponds to the analysis of three independent experiments with three replicates and represents mean (± SEM). One- way ANOVA with Bonferroni’s multiple comparisons post-test was used to analyze the statistical significance (*p < 0.05, ***p < 0.001)...... 268
xxvi
INTENDED TO BE BLANK
xxvii
CHAPTER 1
1. INTRODUCTION
1.1. NEUROTOXICITY TESTING
The main role of the nervous system is to control all bodily processes, as well as psychological and motor functions. It is sensitive to perturbation from different sources and whether it has the ability to regenerate if at all, within the central nervous system (CNS) is still questioned (Sorrells et al. 2018; Snyder, Ferrante, and Cameron 2012). Evidence shows that very slight molecular effects to especially the central nervous system can result in adverse effects on human health.
Genetic disorders of the CNS can take a lifetime to develop and their onset is variable (Pihlstrom,
Wiethoff, and Houlden 2017). Therefore, there is a need for consistent guidance on how to evaluate data on neurotoxic substances and assess their potential to cause transient or persistent and direct or indirect effects on human health.
Regarding neurotoxicity testing and disease modelling, animal studies are the current gold standard, not only due to the complexity of organ effects, but also regulatory requirements.
OECD regulation requires chemical testing in a rodent and non-rodent species at three doses in at least 20 animals (10 female and 10 male). The oral dosing regimen can be acute 28 days, sub- chronic (90 days) or chronic (1 year or longer) (Parasuraman 2011; NRC 1993). Behavioral tests and histological assessments are performed to understand, whether the substance is neurotoxic.
Animal testing is expensive in terms of animal use, scientific resources and is not always predictive of human toxicity due to interspecies differences as well as the diversity within human responses (Olson et al. 2000; Hartung 2013, 2017b).
1
Currently, there are over 100,000 chemicals in the market (Hartung 2010a), of which relatively few have been tested for neurotoxicity. Only 1,505 substances are listed as neurotoxicants with supporting data1. This low number can be partly attributed to the time and cost of neurotoxicity testing (Schmidt et al. 2017) as well as a gap in understanding the true role of environmental exposures in long-term non-communicable neurological diseases such as Parkinson’s and
Alzheimer’s.
Neuronal regeneration remains contended and high oxygen consumption in the aging brain can make the nervous system particularly vulnerable to damage (Snyder 2018). The more we understand about neurodegenerative diseases, the less animal studies seem to fit the purpose as age and (epi)genetic factors play such an important role and are not yet factored into the gold standard animal toxicity tests (Pistollato et al. 2016). New methodologies are needed to toxicology, which is hindered by inter-species differences, time and resource consumption.
To overcome this problem, studies have been established using human-derived cell lines.
However, the difficulty to reproduce the complexity of the central nervous system represents a major challenge to in vitro models. This becomes even more complicated when brain development and neurodevelopmental toxicity (DNT) come into play. The brain develops according to specifically timed molecular signals and consists of different cell types; neurons, astrocytes, oligodendrocytes and microglia (Smirnova et al. 2014)
Cell-cell interactions and proper cell signaling play a central role in brain function as well as response to toxicants. Modelling the complexity of the brain requires an advanced in vitro system and it is therefore essential to develop 3D models (Alépée et al. 2014), which include specific targets indicative of nervous system function, for future neurotoxicity testing(Hartung et al.
2004).
1.1.1. Towards 3D Neuronal in vitro models
1 http://www.istas.net/risctox/en/dn_risctox_lista.asp? busc=1&f=neu
2
Current in vitro models (cancer cell lines, immortalized cell lines, primary cell cultures or stem cells) offer the advantage of a controlled environment to study molecular pathways involved in neurotoxicity. Different models may lack certain components of the cellular microenvironment
(cell types, signaling molecules, cellular matrices), which are critical to disease development.
Understanding the limitations of each model is important to determine, whether it can answer the question being posed. The field of in vitro disease pathology is growing exponentially with the use of human-derived induced pluripotent stem cells (iPSCs) and the ability to create microphysiological systems (MPS) with multiple cell types. MPS refer to systems, which recapitulate aspects of organ architecture and organ function. In the past, the most common in vitro models relied on cancer-derived cell lines (SH-SY5Y, SK-N-BE, BE2-M17 (human) and
PC12 (rat)) (Xicoy, Wieringa, and Martens 2017; Alberio, Lopiano, and Fasano 2012). These cell lines provide a rapid model to study very specific compound effects (Tong et al. 2017) =, however, due to their cancer origin, are unstable and likely have multiple DNA and chromosomal aberrations (leading to variability in their response across experiments), which can lead to unreliable results (Pamies and Hartung 2017). Immortalized cell lines (LUHMES (human) and
N27 (rat)) provide another option for fast differentiation in vitro (Smirnova et al. 2016;
Schildknecht et al. 2013). However, these are single cell-type models and therefore one cannot determine what role other cell types play in disease pathology. Finally, primary cells are costly to produce, human brain primary cells are rarely available, and these do not allow for long-term studies in vitro because of their short life-span in culture. Different compound-induced effects have been reported in primary different lines, therefore, caution must be taken when interpreting neurotoxicity data and testing using different models leads to a more informed approach
(Heusinkveld and Westerink 2017). Upon differentiation, neuronal models cluster and naturally form aggregate-like regions with long processes between them. This sometimes leads to detachment from plasticware which provides a disadvantage to toxicity testing (Figure 1-1).
3
Figure 1-1. Differentiation of NT2 (Eaton and Wolfe 2009), IPSC-derived cortical neurons
(www.axolbio.com), and LUHMES cells (photographed by Helena T Hogberg, Center for Alternatives to
Animal Testing) in 2D culture showing clustering and aggregate-like formations.
The major advance for in vitro toxicology began after the release of the National Academy of
Sciences Report in 2007, driving the need for alternatives to animal testing to better understand the hazards and risks posed by chemicals (NRC, 2007). This was followed shortly by funding by the NIH in 2012 to develop 3D organotypic models for pharmaceutical testing (Hartung and
Zurlo 2012). At this time, 3D skin models such as EpiSkinTM and EpiDermTM were proving to be successful in vitro alternatives (Gordon et al. 2015), but such human models had not yet been developed for the brain. Certainly, a 3D model developed from rat primary cortical neurons had been developed as an alternative for developmental neurotoxicity (Honegger et al, 1979;
Honegger and Monnet-Tschudi, 2001) and spheres from embryonic stem cells as well as immortalized cell lines using the hanging drop technique were commonly used in research. Today there are multiple methods to culture 3D brain organoids including hanging drop, centrifugation, cell aggregation by gyratory shaking, bioreactors and low-adhesion plates (Figure 1-2).
4
Figure 1-2. Building 3-dimensional brain organoids by Sergiu. P. Pasca, 2017
Using a combination of these approaches Lancaster et al., produced the first human iPSC-derived
3D brain model (Lancaster et al. 2013). This model presents regional identities organized as discrete domains capable of influencing one another. This was followed by Pasca et al., who produced similar model with different technic (Pasca et al. 2015). Using neuroprogenitor cells
(NPCs) derived from human iPSCs, our group developed BrainSpheres, containing different neuronal cell types and glia, highly reproducibly in size and shape, and described the potential utility of such a model for toxicity testing (Hogberg et al. 2013; Pamies et al. 2017), similar to the a previously published human neutrospheres DNT model (Moors et al. 2009; van Vliet et al.
2007). Qian et al developed brain-region specific organoids presenting cortical midbrain and hypothalamic organoids (Qian et al. 2018) and Choi et al differentiated human neural precursor
(ReN) cells within a 3D MatrigelTM matrix with high levels of brain extracellular matrix protein
(Choi et al. 2014) (Figure 1-3).
Although these 3D models contain different cell types and brain regions, their complexity leads to variability, which can influence reproducibility between organoids. The differentiation procedure for these models is over 10 weeks long. Prior to full differentiation; these models cannot be described as fully mature. For this reason, they may also serve to test developmental neurotoxicity (DNT). One of the main benefits of iPSC-derived models is the potential to study
5 patient-derived iPSCs (containing the mutated gene associated to a disease) as well as the gene- corrected version of these cells to understand how the specific mutations may influence neuronal differentiation, lead to disease pathology or influence the individual’s response to therapies or toxicant exposures (Martinez-Morales and Liste 2012; Lee et al. 2017; Qian et al. 2018; Qian et al. 2016; Lancaster and Knoblich 2014; Jorfi et al. 2018; Pamies et al. 2017).
Figure 1-3. iPSC-derived 3D models developed by (A) Lancaster et al, 2013 (B) Pamies et al, 2016 (C)
Qian et al, 2018
Although human-derived microphysiological systems are most representative of the human brain due to their multi-cellular composition, the complexity of these models makes it difficult to attribute mechanisms to a single cell-type. Simple, fast and highly reproducible models are needed in parallel for chemical screening in toxicity testing and to study molecular pathways of
6 toxicity (Kleensang et al. 2014). Single-cell type models, differentiated in 3D, offer a tool to study cell-specific toxicant-induced disease mechanisms. Some could argue that traditional monolayer cell culture can achieve this, however, differentiated neurons do not adhere well in 2D culture, leading to the inability to study longer exposures. A 3D model enables differentiation in suspension, which also provides the opportunity to include microfluidics, which mimic organ exposures more closely. A comparison between these models and more complex organotypic models can provide evidence towards the role of other cell types in disease progression. It has been shown that 3D cultures exhibit increased survival and enhanced differentiation compared to ones cultured in monolayer for brain and other tissue cell types (Zhang et al. 2016; Subramanian et al. 2014; Pamies and Hartung 2017; Smirnova et al. 2016). A benefit of in vitro models is the ability to identify, which cell type is affected as well as the mechanism of toxicity of a compound
(Worth and Balls, 2002). Ideally an in vitro test battery would include models ranging from single-cell type to complex multi-cellular models, which preserve tissue structure and function.
From a regulatory perspective, currently, in vitro systems are not yet adequate to fully replace in vivo testing. However, validation of these techniques can contribute to prioritizing compounds or better understanding the sites of action to complement current in vivo neurotoxicity testing strategies (Coecke et al. 2007). Any proposed in vitro assays would have to be formally validated under approval of national and international validation bodies (Bal-Price et al. 2008).
1.1.2. IN VITRO neurotoxicity endpoints
To be able to accurately test compounds for (developmental) neurotoxicity in vitro, reference compounds are needed (Aschner et al. 2017). With the limited number of compounds which have been identified as neurotoxicants, it is challenging to determine reference compounds as these present different modes of action. It has been noted that adverse outcome pathways (AOPs) do not necessarily provide enough quantitative time-concentration-effect relationships to determine
7 whether a molecular initiating event (MIE) can lead to effects at the organ level (Aschner et al.
2017). For this reason, compounds should be tested on their ability to disrupt fundamental biological processes such as proliferation, differentiation (gene expression), migration, axonal and dendritic outgrowth, synapse formation, calcium signaling, neurotransmitter release, electrical activity and apoptosis (viability) (Schmidt et al. 2017). In order to develop in vitro models, which can assess neurotoxicity, these must be carefully characterized to understand their reproducibility and utility (Schmidt et al, 2017). Different neuronal models express specific mRNA and protein markers and identifiable markers for differentiation stages. Some models include neurons and glia, others are composed of a single cell type and release specific neurotransmitters. Importantly, neurons should be functional and must show spontaneous electrical activity. Although some in vitro models contain different neuronal/glial cell types, to this date, the inclusion of microglia (resident immune system) still remains a challenge. A combination of different viability, gene expression and functional endpoints is key to study mechanisms of neurotoxicity (Harry et al. 1998; Terrasso et al. 2015).
Viability
Traditionally, in vitro assays have relied on viability tests to quantify toxicity and differentiate, target specific or unspecific effects. The Alamar Blue assay, measures resazurin reduction by mitochondrial NADH indicating mitochondrial viability and metabolic activity. The MTT assay is a colorimetric assay based on the reduction of MTT to its purple formazan, which occurs in the presence of cytosolic NAD(P)H-dependent oxidoreductase enzymes. The lactate dehydrogenase
(LDH) assay quantifies the enzyme found abundantly in cells and indicates membrane leakage due to cell death. This assay measures late-stage cell death. Although cell viability dose-response curves can indicate the concentration at which a compound is cytotoxic for the given cell type, it is not a suitable measurement to understand indirect mechanisms of toxicity, which may be triggered by a compound that could lead to toxicity in the long-term.
8
mRNA and protein expression
Cell-type specific markers are used to identify different cell types and whether they are affected by a compound. Genes that are only expressed by a given cell type are suitable to identify neurotoxicity and gene profiling can derive biomarkers of toxicity as well as molecular signatures of toxicity. Immunostaining of the proteins allows for visualization of morphological changes, cellular localization and cell loss, however, this endpoint tends to be more qualitative. The use of high-content imaging software to quantify co-localization between stainings is increasingly used as a tool to measure molecular endpoints (Wilson, Graham, and Ball 2014; Sirenko et al. 2014;
Gotte et al. 2010) (refs). Commonly, mRNA or protein expression markers for proliferating (ki-
67)), undifferentiated (SOX2, Nestin, PAX6), differentiated (B-III Tubulin, neurofilament,
MAP2, synaptophysin, NeuN) (Gordon, Amini, and White 2013), glutamatergic (vGLUT,
NMDAR), gabaergic (GAT, GABA receptors), dopaminergic (dopamine transporter (DAT),
Tyrosine Hydroxylase (TH)), cholinergic (ChAt, VAChT, Acetylcholinesterase) and serotonergic
(TPH, Serotonin transporter) neurons as well as oligodendrocytes (olig1, olig2, MBP), astrocytes
(GFAP, S100β) and microglia (CD45, Iba1) (Raff et al. 1979).
Neurotransmitter release
Neurotransmitter systems guide multiple cellular processes development, differentiation of the nervous system and synaptic plasticity. Neurotoxicants can act as agonists or antagonists at specific receptors which inhibit or increase the release of neurotransmitters, or alter the removal of neurotransmitters from the synaptic cleft (Bal-Price et al. 2008). One common example is acetylcholinesterase (AchE) inhibition by organophosphorus insecticides or flame retardants
(Hendriks and Westerink 2015). These compounds lead to irreversible AchE inhibition and acetylcholine accumulation in the synaptic cleft, which hyperstimulates nicotinic and muscarinic receptors disrupting neurotransmission (Colovic et al. 2013). Currently reversible AchE inhibitors
9 are used in Alzheimer’s disease treatment where excitotoxicity plays a role (Marambaud, Dreses-
Werringloer, and Vingtdeux 2009). Alterations in serotonin levels are linked to depression and dopamine imbalance is observed in addiction and Parkinson’s disease (Juarez Olguin et al. 2016)
Transporter activity can also be quantified by adding radiolabeled substrates in cells and supernatant (Schildknecht et al. 2009; Dingemans, van den Berg, and Westerink 2011).
Compounds which affect calcium signaling can in turn alter neurotransmitter levels.
Calcium signaling
Calcium (Ca2+) signaling is a mechanism required for neuronal processes ranging from action potentials to cell adhesion and survival. Calcium disturbances are observed in brain diseases such as Alzheimer’s, Parkinson’s and Huntington’s disease (Marambaud, Dreses-Werringloer, and
Vingtdeux 2009). Neurotransmitter release at the post-synaptic cleft is regulated by an increased influx of Ca2+ (Brini et al. 2014). Intracellular calcium is tightly regulated and maintained at 100 nM, while extracellular calcium lies in the millimolar range (Barhoumi et al. 2010). Mitochondria are calcium regulators, which can increase cytosolic Ca2+ during oxidative stress. Continuous
Ca2+ release from mitochondria can lead to cell death. To assess cellular health, cytosolic to mitochondrial Ca2+ ratio can be quantified, alternatively, short-term kinetic measurements of stimulated Ca2+ transients can be measured to test more mechanistic endpoints. Furthermore, dynamic measurements of Ca2+ waves and oscillations are measured over time and can identify how compounds affect long-term calcium homeostasis. These measurements all involve cellular imaging or biosensors.
Cellular ATP levels
The production of ATP is a direct measurement of cellular energy metabolism and functionality. Compounds, which interfere with either the mitochondrial membrane potential or
10 electron transport machinery, can affect ATP production. Several compounds are known to inhibit complex I (rotenone), III (antimycin A), IV (cyanide, carbon monoxide) and V
(oligomycin) of the electron transport chain. Furthermore, compounds which induce cellular stress through reactive oxygen species (ROS) production can lead to loss of mitochondrial membrane potential and thus decreased ATP production. Neurons require high levels of ATP because proteins are transported long distances through neurites to synapses. Improved technologies such as the Agilent Seahorse assay 2 allow real-time monitoring of cellular respiration and can quantify which stage of ATP synthesis is impaired.
Mitochondria functionality
As mentioned above, compounds can decrease mitochondrial membrane potential (MMP).
Although few studies focus on this as an endpoint, however, its close link to apoptosis means it can serve as an earlier indicator of toxicity (Sakamuru, Attene-Ramos, and Xia 2016).
Compounds can affect mitochondrial function direct or indirectly. The MMP is generated via redox reactions in the electron transport chain, which produces ATP. Over the past decade, the use of dyes to study mitochondrial functionality has increased. Examples include cell membrane permeable fluorescent dyes, such as 3, 3′-dihexyloxacarbocyanine iodide [DiOC6(3)], rhodamine-
123 (Rh123), tetramethyl rhodamine methyl and ethyl esters (TMRM and TMRE), and JC-1, which are able to cross into mitochondria, however, only remain if the mitochondrial membrane is not damaged or the mitochondrial permeability transition pore (mPTP) is not open (Rao,
Carlson, and Yan 2014). Recently, improved imaging techniques allow researchers to look at mitochondrial morphology, fission and fusion via fluorescence imaging or electron microscopy
(Wiemerslage and Lee 2016). Quantification of changes in the length, fragmentation, and location of mitochondria can be indicative of mitochondrial functionality.
2 https://www.agilent.com/en/products/cell-analysis/how-seahorse-xf-analyzers-work
11
Neurite Outgrowth and Integrity
Probably one of the better characterized organ-specific functional endpoints, neurite outgrowth assays allow to quantify the number, length and branching of neurites (Krug et al. 2013; Radio and Mundy 2008). Compounds can inhibit neurite outgrowth and integrity at concentrations, which do not affect cell viability. Although evaluated as a screen for developmental toxicants, it can also be used for neurotoxicity using terminally differentiated cells (Sherman and Bang 2018).
Automated detection and screening can provide quantitative data on how compounds increase or decrease neurite outgrowth parameters.
Reactive Oxygen species (ROS) accumulation
The production of cellular ROS can begin through various direct and indirect mechanisms. Cell stress which overwhelms cellular energy metabolism, mitochondrial dysfunction, imbalance in oxidant and antioxidant pathways, impaired cell signalling or transport mechanisms and redox cycling can increase ROS production or accumulation. ROS can act as signaling molecules, which can activate antioxidant mechanisms to stress. Although this is not an endpoint that provides target specificity for a given compound, it can provide information on whether the cell is able to cope with a toxic insult. Currently, fluorescence assays can quantify intracellular ROS (the best-known substrate is 2',7'-dichlorodihydrofluorescein diacetate (H2DCFDA), which produces a fluorescent product in the presence of ROS), while other assays can quantify oxidative damage such as DNA double strand breaks (pH2Ax staining and comet assay) or lipid peroxidation
(malondialdehyde (MDA) adducts, detected fluorometrically).
Misfolded proteins
12
The accumulation of misfolded proteins can serve as an endpoint to study neurotoxicity, since it is an important mechanism in multiple neurodegenerative disorders such as tauopathies and alpha-synucleinopathies (Takalo et al. 2013). Increased oxidative stress can lead to misfolded proteins and disruptions to protein quality control systems and the unfolded protein response, which causes accumulation of protein aggregates, which in turn are toxic to cells when not removed. Protein aggregates are measured using fluorescent ligands or Western blot after separation based on their solubility (Bondos and Bicknell 2003).
Electrical Activity
Many neurotoxicants cause changes that do not affect cell viability but affect pre- and postsynaptic mechanisms of transmission or action potential propagation. Patch clamp techniques were the method of choice to identify effects on specific ion channels. Although this method is able to take very accurate measurements, it is a specialized, low-throughput technique, which measures a single cell and does not allow for subsequent recordings. Microelectrode arrays
(MEAs) are plates containing electrodes, which measure ion currents at different sites simultaneously, without disrupting cells (Johnstone et al. 2010). They allow non-invasive high- throughput recordings, which can be performed over the cell culture or treatment period
(Novellino et al. 2011; Scelfo et al. 2012). MEAs had previously been used to study brain tissue slices or primary dissociated cultures in basic research but are currently proposed as a promising tool for high-throughput neurotoxicity testing for prioritization after evaluation by multiple independent laboratories (Hogberg et al. 2011; Johnstone et al. 2010; Novellino et al. 2011).
1.1.3. Neurotoxicology of nanoparticle drug delivery systems
Nanotechnology, alteration of material at the atomic level to develop macroscale products, was described first in the late 1950s by Nobel Prize winner Richard Feynman (Hulla, Sahu, and Hayes
13
2015). This idea was further developed with the synthesis of carbon nanotubes (Iijima, 1991). At the start of the 21st Century, nanoparticle (NP) use in technological and biomedical applications has grown exponentially. Currently, 700 types of nanoparticles are estimated to be on the market
(Wiesner et al. 2011). Nanomaterials are described as particles within 1-100 nm in size and for the past 20 years have been developed for products ranging from cosmetics to food, sensors, antimicrobials, diagnostic biosensors, imaging probes and drug delivery systems. Due to their presence in almost all consumer products today, and extensive human exposure, there is concern for the potential health risk they pose. This led to the emergence of the field of nanotoxicology, which studies the adverse human and environmental effects of nanoparticles(Oberdorster et al.
2005). Nanoparticles are also of concern due to their distinct physico-chemical properties (size, large surface area, and surface properties (e.g. formation of protein corona)), and experts in the field have concluded that a different approach to testing nanoparticles is needed instead of conventional toxicity testing 3 (Hartung and Sabbioni 2011). Studies have shown the applications of nanoparticle drug delivery to treat neurodegeneration in animals and cells (Yealland et al.
2016; Bezem et al. 2018; Khanbabaie and Jahanshahi 2012; Hawthorne et al. 2016; Rizvi and
Saleh 2018), while others have reported toxicity by different nanoparticle types to brain cells in vivo and in vitro (Yarjanli et al. 2017; Mushtaq et al. 2015; Larner et al. 2017). Testing in vitro has heavily relied on viability assays and cell imaging techniques as nanoparticles have the potential to interfere with multiple assay measurements (fluorescence recordings, RNA/DNA extraction, binding assays). Furthermore, many studies to date do not accurately characterize nanoparticles in cell culture media or blood circulation. Nanoparticle aggregation can lead to different toxicities and improved physico-chemical characterization is necessary to better understand toxicity mechanisms.
3 http://www.fda.gov/nanotechnology/taskforce/report
14
An effective nano-drug delivery system for neurodegenerative diseases would have the capacity to cross the blood brain barrier (BBB) and release the drug at targeted brain regions without exerting toxicity (Cacciatore et al. 2016). Figure 1-4 summarizes different properties which influence systemic delivery and BBB passage.
Figure 1-4. NP features which influence systemic delivery and blood brain barrier (BBB) passage.
Published by (Saraiva et al. 2016)
Specifically, for Parkinson’s disease therapy, the use of nanoparticles to target and degrade alpha- synuclein aggregates is a promising application (An et al. 2010; Helmschrodt et al. 2017). Figure
1-5 shows the rapid increase in the number of yearly publications which are found for the search term ‘nanoparticles AND Parkinson’s’. Publications almost doubled between 2014 and 2015 and the number of publications in the first quarter of 2018 is equal to the number published in 2013.
Currently there are no international nanotoxicity testing guidelines and the increasing use of nanoparticles for drug delivery, requires a faster, less costly approach. In vitro alternatives would allow for faster testing and using a range of cellular models is proposed as a beneficial strategy to nanotoxicology(Clift, Gehr, and Rothen-Rutishauser 2011).
15
Figure 1-5. Number of publications per year using search term ‘nanoparticles AND Parkinson’s’.
1.2. PESTICIDE EXPOSURE AND PARKINSON’S DISEASE
The first reports that pesticide exposure (specifically paraquat and rotenone) could lead to neurodegeneration were published in the 1990s, although these had been used in agriculture since the early 1950s (Liou et al. 1997; Betarbet et al. 2000; Dhillon et al. 2008; Hertzman et al. 1990).
Because age is a risk factor for neurodegenerative diseases, it is still challenging to associate an exposure to the development of a disease later in life. Reported symptoms after occupational exposure to pesticides include headache, dizziness, nausea, vomiting, pupillary constriction, and currently there is not enough epidemiological evidence to define, whether these events may contribute to neurodegeneration later in life (Cannon and Greenamyre 2011). The accidental exposure to MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine produced as a side product during MPPP synthesis, which was sold as illicit drug, often named “synthetic heroin”), led to the first animal models of Parkinson’s disease, linking acute neurotoxicant exposure to long-term neurodegeneration (Langston et al. 1999; Langston et al. 1983). Currently, maneb (fungicide), glyphosate (herbicide) as well as solvents and flame retardants are of concern too, although
16 evidence is not yet sufficient to proof a link between exposure and PD (Meco et al. 1994;
Gunnarsson et al. 1992).
1.2.1. Public Health Significance
PD is a complex disorder, which presents differences in pathology between patients due to heterogeneous genetic backgrounds and lifetime exposures. It is the second most common neurodegenerative disease after Alzheimer’s disease, affecting 5 million people worldwide
(Dorsey et al. 2007). This number is expected to double by 2030, with a prevalence shift towards developing Eastern nations. It the second most-common neurodegenerative disease in people over the age of 65. Aging is currently the highest risk factor for PD and although some known genetic factors play a major role in early onset of familial PD, these monogenic forms only account for 5-
10 % of sporadic PD cases (Farrer 2006; Sherer et al. 2003). Other genetic risk factors as well as environmental factors (age, stress) or exposures (pesticides, flame retardants, metals) have been associated with the remaining sporadic cases. When no genetic association to any single gene can be identified, an environmental component is likely (Meulener et al. 2005). Moreover, the increasing incidence of Parkinson’s disease is faster than genetic drift, which provides evidence towards a link between environmental exposures and PD development. Finally, multiple twin studies have not shown sufficient evidence for hereditability (Tanner 2003; Wirdefeldt et al.
2004; Tanner et al. 1999). Current research focuses on gene-environment interactions (the idea that response to an environmental exposure may differ according to an individual's genotype) and epigenetic mechanisms in neurodegeneration to better understand how acute exposures lead to chronic disease.
17
Over five billion pounds of pesticides were used worldwide in 2007 and our exposure via
different routes is increasing4. Epidemiological data have confirmed increased risk of PD after
exposure to two pesticides, rotenone and paraquat (Tanner et al. 2011; Furlong et al. 2015;
Nandipati and Litvan 2016). However, studies also indicate that not all exposed individuals
develop this neurodegenerative disease (Semchuk et al. 1992, Gorell et al. 1998, Petrovitch et al.
2002); therefore, other susceptibility factors must play a role. Although retrospective
epidemiological studies prove exposure is a risk factor for PD, they do not prove a causal link or
allow of causal relationships (Breckenridge et al. 2016; Van Maele-Fabry et al. 2012; Hernandez,
Reed, and Singleton 2016; Baltazar et al. 2014) . Novel next generation sequencing (NGS) allows
for resequencing of Parkinson’s disease data sets and it is proposed that using a ‘common
disease–multiple rare variants hypothesis’ higher effect sizes can be identified for disease-
relevant alleles (Tsuji 2010). Table 1-1 outlines studies, which have found positive associations
between rotenone or paraquat exposure and PD (Nandipati and Litvan 2016).
4 https://www.epa.gov/sites/production/files/2017-01/documents/pesticides-industry-sales-usage- 2016_0.pdf (Accessed May 21st 2018)
18
Table 1-1. Studies which have found an association between rotenone or paraquat exposure and PD risk.
Compound Association Reference Rotenone exposure associated with Tanner et al. 2011 PD, OR 2.5 (95% CI 1.3, 4.7)
Report of past rotenone use was Dhillon et al. 2008 associated with PD, OR 10.0 (95% CI 2.9, 34.3) Rotenone Protective glove use modified Furlong et al. 2015 association of paraquat and permethrin with PD, paraquat OR 3.9 (95% CI 1.3, 11.7) * & permethrin OR 4.3 (95% CI 1.2, 15.6) * but did not modify the association with rotenone
Paraquat exposure associated with Liou et al. 1997 PD, OR 3.2 (95% CI 2.41, 4.31)
Paraquat associated with higher Kamel et al. 2007 rate of prevalent PD, O.R. 1.8 (95% CI 1.0, 3.4)
Paraquat exposure associated with Costello et al. 2009 increased PD risk, OR 2.27 (95% CI 0.91, 5.70) Paraquat Paraquat exposure associated with Tanner et al. 2011 PD, OR 2.0 (95% CI 1.4, 4.7)
Paraquat exposure and history of Lee et al. 2012 traumatic brain injury associated with PD risk OR 2.77 (95% CI 1.45, 5.29)
1.2.2. Parkinson’s disease pathophysiology
Dopaminergic neurons play a pivotal rule in mechanisms such as movement, drug addiction and
reward. Although they account for less than 1% of neurons represented in the brain, degeneration
leads to Parkinson`s disease (PD) (Chinta and Andersen 2005). This is a progressive
neurodegenerative disorder, characterized by the pathological feature of excessive dopaminergic
neuron loss in the midbrain substantia nigra (SN), and in some cases, intra-cytoplasmic
19 inclusions in intact neurons (Lewy Bodies). The disease originates in the basal ganglia and affects specific neuron groups along the cortex, thalamus, brain stem, and spinal cord (Caligiore et al.
2016). Moreover, neuronal and non-neuronal interactions play a role in PD. Post mortem studies have shown that neuro-inflammatory processes are present in Parkinson’s disease, which might contribute to dopaminergic cell death (Wang, Liu, and Zhou 2015). Clinical manifestations appear only when >50% of nigral DA neurons are lost and presents motor-symptoms such as rest tremor, bradykinesia and postural disturbances. To date only symptomatic treatment for patients suffering from PD is available and there are no accepted therapies targeting neuroprotection or neuro-restoration. The most common treatment continues to be L-DOPA (dopamine precursor), due to low levels of dopamine neurotransmitter and its receptors in PD brains (Barbeau 1969). In order to make clinical progress, improved diagnosis of Parkinson’s disease in its earlier stages, and the establishment of therapies preventing progression are required.
1.2.3. Molecular mechanisms in Parkinson’s Disease
PD is a complex multifactorial disorder which involves, genetic, epigenetic and environmental factors contributing to changes at a molecular level and long-term pathology.
Genetic risk factors
The genetic factors found to be associated with PD are depicted Figure 1-6 and most significant ones are described here in more detail. The most prominent pathological characteristic of PD is accumulation of misfolded α-synuclein in the substantia nigra (Lewy Bodies) (Spillantini et al.
1997). Alpha-synuclein (α-Syn) is abundantly expressed in the nervous system, comprising 1% of total cytosolic protein (Stefanis 2012). Abnormal deposition of α-Syn and its misfolding, occurring as a downstream effect of mutations in the SNCA gene, is known to result in toxic intermediates, which affect signaling and mitochondrial functionality, and is usually observed in
PD pathology (Singleton et al. 2003). One of the genes involved in mitochondrial quality control
20
(Chartier-Harlin et al. 2004)is Parkin (PARK2). A variety of mutations in this gene (coding for a ubiquitin ligase), including point mutations, intron and exon rearrangements, lead to decreased function. Normal function of Parkin maintains proper mono- and ubiquitination to of lysine 29,
48, or 63 (Zhang et al. 2015). Dysregulation of Parkin can result in dopaminergic neurodegeneration through improper activation of pro-survival pathways like the NF-kB pathway
(Henn et al. 2007). Mutations in the PTEN-induced putative Kinase 1 (PINK1) have also been linked to hereditary forms of Parkinson’s disease. Valente and colleagues have demonstrated that wildtype forms of PINK1 may protect neurons from stress-induced mitochondrial dysfunction and apoptosis, while mutated forms do not (Valente et al. 2004). DJ-1 is an oxidative stress sensor, redox-sensitive chaperone and protease involved in protection against cell stress and cell death (da Costa 2007). It stabilizes NRF2 and PINK1, thereby enhancing cell protection mechanisms. It also plays an important role in maintaining mitochondria morphology and autophagy of disrupted mitochondria. DJ-1 has also been found to regulate tyrosine hydroxylase
(TH) expression (Ariga et al. 2013). The most common mutation in spontaneous PD is in the leucine-rich repeat kinase 2 (LRRK2) gene. As a GTPase kinase, the protein has ubiquitous functions within the cell. One of its roles is associated with vesicular transport in presynaptic vesicles (Biskup et al. 2006). It has further been found to interact with SNCA, shown by co- immunoprecipitation in PD patient samples (Yacoubian et al. 2010). GBA, the gene coding for glucocerebrosidase β acid, is absent in Gaucher’s disease. This is a lysosomal storage disorder, in which patient’s experience fatigue, anemia and enlargement of liver and spleen, and in some cases develop PD (Sidransky and Lopez 2012). As this protein plays a role in lysosomal protein degradation, loss of glucocerebrosidase enzyme (GCase) activity has been shown to cause accumulation of alpha-synuclein (Bae et al. 2015). In 2017, a meta-analysis including over
425,000 individuals identified 17 new PD risk loci (Chang et al. 2017) (Figure 1-6). These genes were mostly associated to lysosomal function and autophagy, thus more work is needed to develop novel targets for therapy. Another feature of PD pathology is deregulation of iron and
21 calcium homeostasis (Sian-Hulsmann et al. 2011). With increasing age, neurons increase reliance on L-type Calcium channels. This reliance is assumed to increase mitochondria-mediated oxidative stress, which downstream may to increased neuronal death (Branch, Sharma, and
Beckstead 2014). This provides one explanation as to why age is a known as susceptibility factor for idiopathic PD.
Figure 1-6. Most likely candidate genes that were significantly associated with PD. Black or grey text indicate known loci and red text indicates novel loci significantly associated with PD, published by Chang
et al, 2017.
Currently, diagnosis for PD relies on post-mortem brain samples looking for α-synuclein accumulation in Lewy bodies. This phenotype is present in other synucleinopathies such as PD with dementia, dementia with Lewy bodies, and multiple-system atrophy. However, in the future, measuring α-synuclein in cerebrospinal fluid may be available 5 . Diagnosis of early-stage
5 Handbook of Clinical Neurology Volume 145, 2018, Pages 309-323
22 synucleinopathies remains a challenge; therefore, more research into understanding the mechanisms in this disease is needed. A meta-analysis by Tanner et al. classified pesticides by mechanism. This demonstrated significant associations between complex I inhibition and a parkinsonian phenotype (Tanner et al, 2011). In the current approach to link changes at the molecular level with adverse outcomes (adverse outcome pathways, AOP) (Leist et al. 2017), understanding the key events in Parkinson’s disease can provide weight of evidence to describe
AOPs and apply these to integrated approach to testing and assessment (IATA) (Tollefsen et al.
2014).
Complex I inhibition
Complex I (NADH-ubiquinone oxidoreductase) of the electron transport chain is the molecular target for some compounds shown to induce PD (rotenone, 1-methyl-4-phenylpyridinium
(MPP+), an active metabolite of MPTP) (Sherer et al. 2007). In concordance, patients show complex I deficiency in the substantia nigra (Schapira et al. 1989; Parker, Parks, and Swerdlow
2008). Others have suggested that early glutathione depletions observed in the parkinsonian SN and nitric oxide (NO) signaling could be responsible for subsequent complex I inhibition, mitochondrial dysfunction, and neuronal cell loss, thus indicating that compounds acting via other mechanisms could lead to selective dopaminergic cell damage via indirect complex I inhibition
(Zhang, Dawson, and Dawson 2006; Hsu et al. 2005; Chinta and Andersen 2006).
Oxidative Stress and Mitochondrial Dysfunction
The brain consumes over 20% body oxygen while only constituting 3% of body mass (Mariani et al. 2005). Blocking of electron flow through the electron transport chain (ETC) leads to the
- generation of reactive oxygen species; superoxide (O2 ), nitric oxide (NO) and hydroxyl radicals
(OH) (Franco-Iborra, Vila, and Perier 2016; Keane et al. 2011). These molecules can act as signaling molecules or lead to lipid peroxidation or protein damage which can end in cell death.
23
Oxidative stress can further damage mitochondria and mitochondrial DNA (mtDNA). Reduced antioxidant defenses (gluthatione (GSH) and superoxide dismutase (SOD)) can also influence
ROS levels (Jenner et al. 1992). PD patients show increased levels of ROS, protein oxidation, mitochondrial damage and reduced antioxidant activity (Jenner 1993; Floor and Wetzel 1998;
Saggu et al. 1989). Increased ROS disrupts the mitochondrial membrane potential (MMP) via increases in intracellular calcium ions (Ca2+), and ATP production relies on a stable MMP (Sherer et al. 2001). Iron has also been found to play a role in oxidative stress and is related to mitochondrial viability. Increased iron has been recorded in PD brain tissue and release of reactive ferrous iron in response to ROS can lead to creation of more ROS via the Fenton reaction
(Double et al. 1998). Dopamine, released by dopaminergic neurons, can also be oxidized producing quinones, which have been reported to further inhibit the electron transport chain
(Khan et al. 2005). This evidence may explain why dopaminergic cells are more susceptible to mitochondrial dysfunction.
Decreased ATP production
As mentioned above, ETC complex inhibition or mitochondrial dysfunction lead to decreased
ATP production. Besides regular cell functions, neurons require higher levels of ATP to transport proteins from the nucleus along axons to synapses and to maintain the resting membrane potential of the cell. In the absence of ATP, cellular depolarization occurs, leading to excitotoxicity
(Sherer, Betarbet, and Greenamyre 2002). A decrease in ATP will also lead to dopaminergic susceptibility when not meeting energy demands. Neurons do not have a high glycolytic capacity
(ability to produce ATP from glycolysis), however, they can produce ATP from lactate and will do so when levels are high. Astrocytes have been found to be more efficient at glycolysis, although they have the same oxidative capacity (Belanger, Allaman, and Magistretti 2011).
Creatine, from our diet crosses the blood brain barrier. It can be converted into phospho-creatine by creatine kinase, which produces ADP. During high energy demands, creatine kinase can
24 regenerate ATP from ADP using phospho-creatine (Andres et al. 2005; Balestrino et al. 2002).
MPTP has shown to decrease ATP levels in rat brain and rotenone has similarly shown to decrease ATP production in vitro (Chan et al. 1991; Li et al. 2003; Sherer et al. 2003; Krug et al.
2014).
Accumulation of misfolded proteins and autophagy
Parkinson’s disease is also characterized by impaired proteostasis, which is the mechanism by which proteins are synthesized, modified, transported and degraded within a cell (McNaught et al.
2001) The two main systems which remove proteins from the cell are the ubiquiting proteasomal system (UPS) and the autophagy-lysosomal pathway (ALP). In a healthy cell, 30 % of all protein produced is degraded. Under stress, (impaired ATP production and increased ROS), the accumulation of misfolded proteins forms protein precipitates, which can further interfere with intracellular protein transport.
Impaired dopamine release and uptake
Dopamine is produced in peripheral tissues, however, its main role is in the SN of the ventral midbrain. Dopaminergic neurons (DN) in the substantia nigra produce dopamine and release it to postsynaptic terminals in the striatum (Figure 1-7). Here D2 receptors intake dopamine.
Remaining dopamine is removed by monoamine oxidases (MAO). Various techniques have been able to demonstrate lower dopamine content and uptake in PD patients (Leenders et al. 1990;
German et al. 1989). Dopamine is produced from the amino acid tyrosine by tyrosine hydroxylase
(TH) and when dopaminergic cells are under stress or undergoing cell death, lower levels of dopamine are produced or transported. Patients have shown that compensatory mechanisms take over, such as decreased dopamine reuptake from the synaptic cleft by the dopamine transporter
(DAT), increased dopamine metabolism, modified D2 transport and dopamine diffusion (Blesa et al. 2017).
25
Figure 1-7. The pathway of dopamine synthesis, release, metabolism and receptors published by (Youdim,
Edmondson, and Tipton 2006)
Inflammation
Although it is known that inflammation plays a role in PD and other neurodegenerative diseases, there is little quantitative data on how it contributes to the different stages of PD (McGeer and
McGeer 2008). Microglia and astrocytes have found to be activated by damaged neurons (Bartels and Leenders 2007). However, animal studies have shown cross species differences and in vitro models have not yet been able to incorporate inflammatory cells to study how neuro- inflammation contributes to neurodegeneration.
Dopaminergic neuron death
Dopaminergic cell loss is currently used for Parkinson’s disease diagnosis in post-mortem samples. PD symptoms are visible when 50-70 % of DN in the substantia nigra are lost. Although
26 some mechanisms leading to DN death are known, prevention of DN loss is still a challenge and late diagnosis means that current treatments target symptoms rather than recovery.
Epigenetic mechanisms in PD
Amongst the previously mentioned mechanisms, epigenetic changes are also known to occur.
Some miRNAs have found to play a role in PD, as well as DNA methylation changes. These are summarized in Figure 1-8.
Figure 1-8. Epigenetic processes in familial and sporadic PD published by (Ammal Kaidery, Tarannum,
and Thomas 2013)
1.2.4. In vivo and in vitro Parkinson’s disease models
Parkinson’s disease is studied in neurotoxic and genetic in vivo and in vitro models. It must be noted that current models do not fully recapitulate PD symptoms and pathology (Blesa and
27
Przedborski 2014). This could be partly due to the fact that PD is also a disease, in which aging plays an important role and animals are not kept until old age. Animals exposed to 6‐ hydroxydopamine (6‐OHDA), 1‐methyl‐1,2,3,6‐tetrahydropyridine (MPTP), paraquat or rotenone. Although monkeys, cats, dogs, and rats are all sensitive to 6-OHDA, rats are most commonly used. However, 6-OHDA does not lead to Lewy body formation. The MPTP model in primates is the most similar to human PD, and improved models with low doses leading to progressive disease development are becoming developed. Paraquat is an herbicide, which has a similar structure to MPTP. Exposure leads to the production of reactive oxygen species via redox cycling and protein aggregation in mice (Manning-Bog et al. 2002). The rotenone exposure model has mainly used rats since studies using mice or monkeys have not been successful
(Ferrante et al. 1997). The systemic rotenone model of PD has become a widely used animal model as recapitulates human PD pathology (Cannon et al. 2009; Betarbet et al. 2000).
Genetic models have also been developed with the most common mutations associated with PD
(Chesselet and Richter 2011). These include SNCA, LRKK2, DJ-1, Parkin or PINK1 mutations.
To date, these models have proven to be less pathologically similar to PD compared to chemical- induced models, as significant loss of dopaminergic neurons is not observed (Beal 2010).
Monogenic forms of PD occur earlier in life, therefore, animal models may be more representative of the development of these forms of PD.
In light of the disadvantages of existing animal models, in vitro models can provide a tool to understand the mechanisms, by which environmental exposures and genetics lead to dopaminergic neurodegeneration as well as neuroprotective pathways. Understanding these molecular pathways is critical to determine underlying mechanisms of gene-environmental interactions contributing to the progression of this neurodegenerative disease and identifying biomarkers for early diagnosis or therapy. Current dopaminergic cell models include non- neuronal models, which produce dopamine (H4 and HEK cells), cancer cell lines (SH-SY5Y), rat
28 cell lines (PC12 and primary neurons) and human immortalized cells (LUHMES) or iPSC- derived neurons (Error! Reference source not found.). Although iPSC-derived models are currently preferred as they better mimic brain physiology, identifying which cells play a role in neurotoxicity remains challenging. The 3D LUHMES cell model is currently the only 3D model from human dopaminergic cells (Smirnova et al, 2016, Harris et al, 2017). This cell line has already proven to be a useful tool in high-throughput neurotoxicity testing, specifically for dopaminergic toxicity in vitro (Lotharius et al. 2005; Tong et al. 2017; Zhang, Yin, and Zhang
2014; Krug et al. 2013; Krug et al. 2014).
29
Table 1-2. Cell lines used as dopaminergic cell models – table updated and modified from (Lazaro, Pavlou,
and Outeiro 2017).
Cell Line Advantages Disadvantages H4 -Easy culture, transfection -Not neuronal/dopaminergic cells HEK -Expres aSYN -Not neuronal/dopamiergic cells -Suitable for screening effect of compounds on aSYN (Lazaro et al, 2016) SH-5YSY -Differentiated cells express TH, dopamine D2 and D3 -Neuroblastoma origin - can lead to differences in receptors, DAT, and VMAT2 (Presgraves et al, 2004) differentiation due to genomic instability (Xicoy et al, 2017)
-Easy culture and transfection -Inconsistent response to the same differentiation treatment, depending on the cell source (Wang et al, 2007)
-Differentiated cells are more susceptible to neurtoxicants -Neuroblastoma origin - can lead to differences in than non-differentiated (Cheung et al, 2009) differentiation due to genomic instability
PC12 -Differentiate into neurons -Non-human -Secrete dopamine and levodopa (Yoshida et al, 2003) -Cultured with serum which leads to variability -Suitable model to study neurosecretion (Westerink and Ewing, 2008) LUHMES -Derived from human mesencephalic neurons - -Low transfection efficiency neurodevelopmental studies, disease modeling and neuropharmacology (Scholz et al., 2011) -Display dopaminergic features after differentiation -TH expression is heterogeneous after differentiation (Zhang (Lotharius et al, 2005; Zhang et al, 2014) et al, 2014) -Suitable for neurodevelopmental and neurotoxicity studies (Schlachetzki et al, 2013) -Can be cultured as a 3D model (Smirnova et al 2016, Harris et al, 2017) Primary dopaminergic -Suitable to study dopaminergic cell survival and neurite -Short-term studies due to difficult maintenance neurons retraction, and regeneration (Schlachetzki et al., 2013) -Not always human origin -Dissection procedure can introduce experimental variability (Xicoy et al., 2017) -Low percentage of TH cells (Falkenburger and Schulz, 2006) iPSC-derived neurons -Mimic the microenvironment of the PD cells in vitro -Differentiation protocols are long and diverse leading to (Zhang et al, 2017) variable results if not well characterized -Can be differentiated in dopaminergic neurons (Soldner -Inherently immature cells et al, 2009) -Suitable to study individual variability in response to compounds -Can be cultured as 3D models (Lancaster et al, 2013, Qian et al, 2016, Pamies et al, 2017)
1.2.3.1 Lund Human Mesencephalic (LUHMES) cell line
The LUHMES cell line was developed in 2005 (Lotharius et al. 2005). It is a subclone of the tetracycline-controlled, v-myc-overexpressing human mesencephalic-derived cell line MESC2.10
30
(Lotharius et al. 2002) Treating cells with tetracycline abolishes v-myc expression which allows cells to exit proliferation and differentiate. This overcomes the limitation of primary neuronal models which do not proliferate as large numbers of neurons can be generated prior to differentiation (Scholz et al. 2011). To maintain expression of Tyrosine Hydroxilase (TH), the continued presence of dibutyryl‐cAMP is necessary in culture media. All other differentiation expression patters are expressed in the absence of external factors (cAMP, GDNF) indicating these cells are post-mitotic and differentiate homogeneously into dopaminergic neurons (Scholz et al, 2011). In monolayer culture, LUHMES require a pre-differentiation step in culture flasks, prior to seeding on plates. In 3D culture, the anti-proliferation agent Paclitaxel (Taxol) is added
(10 nM, 48 h) and washed out to remove proliferating cells (Smirnova et al, 2016). Generation of a pure-dopaminergic, non-proliferating cell culture is necessary to assess dopaminergic neurotoxicity. Several reporter lines were generated overexpressing ASYN, DAT1, TH and carring eGFP- or turboRFP as a fluorescent tag. A GFP- or RFP-expressing LUHMES cell line can be used to image neurite outgrowth and disturbance by neurotoxicants. Mechanistically, both rotenone and MPP+ are toxic to LUHMES cells and effects on ATP synthesis, mitochondria dynamics and neurite outgrowth are observed, while inhibition of SNCA expression by siRNA in
LUHMES is protective against MPP+ induced toxicity in vitro (Schildknecht et al. 2013). 3D
LUHMES aggregates are cultured in differentiation media on a gyratory shaker. Within 3 days, cells express the same levels of differentiation makers as in 2D. Proliferation and size of aggregates is controlled and cells can be kept in culture for over 21 days. Differentiation and toxicity results have been reproduced in the newly developed 3D LUHMES model, with the added advantage that toxicant can be easily removed (Smirnova et al, 2016; Harris G et al, 2017).
1.2.3.2. Induced pluripotent stem cell (iPSC)-derived neurons
31
We live in the era of development towards personalized medicine. The generation of iPSCs by reprogramming human fibroblasts was pioneered in 2003 (Tokuzawa et al. 2003; Takahashi and
Yamanaka 2006). The work described four transcription factors (Oct4 (Pou5f1), Sox2, cMyc, and
Klf4) which when transduced into adult fibroblasts, led to the generation of iPSCs, that shared embryonic stem cell properties. With the ethical disavantages of using ESCs, this was a major breakthrough in drug development (Lorenz et al. 2017), disease modelling and cell transplantation medicine using human iPSCs (Park et al. 2008; Takahashi et al. 2007). Today, the generation of iPSC without the need for transgene integration is possible. These have been successfully generated only using purified recombinant proteins (Zhou et al. 2009), or modified
RNA molecules (Warren et al. 2010). To derive neural precursor cells (NPCs) from iPSCs, different protocols have been published, which involve growing cells in neural differentiation media containing growth factors and suppliements specific for neural development (Li et al.
2011; Gunhanlar et al. 2017). NPCs can differentiate into neuronal and glial cell populations, and this can also be directed to produce a single cellt type cultures (Error! Reference source not found.). To develop some organotypic 3D models, NPCs are differentiated into neurons and glia using different techniques allowing aggregation (Figure 1-2), with the aim to better recapitulates the cellular environment of the brain. The iPSC-derived BrainSpheres used in this thesis are differentiated for 8 weeks under gyratory shaking which forms 350 µm aggregates containing neurons, astrocytes and oligodendrocytes (Pamies et al. 2017).
As mentioned above, iPSC-derived models benefit from containing multiple cell types and better recapitulating in vivo structure. However, for toxicology, identifying which cells are affected, or contributing to a measured endpoint remains challenging. For example, rotenone, which only affects dopaminergic neurons, would not cause any effect on aggregate viability as a whole as dopaminergic cells comprise a small percentage of neurons within aggregates. Specific markers are needed to assess toxicity, but again low number of cells can lead to low expression levels to
32 start with. Differentiating iPSC into specific neuronal types and then generating 3D models is one solution. Alternatively, using markers to be able to sort cells and extract one specific cell type to measure gene expression within these cells will allow cell-specific anayalsis (Goudriaan et al.
2014).
NPCs
Figure 1-9. NPCs generate different cell types of the nervous system (www.sigmaaldrich.com/life-
science/stem-cell-biology/neural-stem-cell-biology.html)
1.2.5. Rotenone as a PD model compound
Rotenone has been widely studied as one of the best-known model PD-inducing compounds, it is extremely lipophilic, and freely crosses cellular membranes independent of any transporters
(Error! Reference source not found.).
33
Figure 1-10. Rotenone molecular structure
Rotenone animal models use chronic exposure to rotenone via intravenous infusion (2–3 mg/kg/day, 1-5 weeks), intraperitoneal injection (2-3 mg/kg/day, 1-5 weeks), intermittent low- dose i.p. dosing (1.5-2 mg/kg/day, 2 months), or intragastric administration (5 mg·kg−1 5 days a week for 2-3 months) (Cannon et al. 2009; Uversky 2004). Using quantitative autoradiography
(Higgins and Greenamyre 1996), Betarbet et al reported that the uniform, chronic exposure required in the animal model brain, to induce PD pathology is ~20-30 nM rotenone, for days to weeks (75% inhibition of specific binding to complex I) (Greenamyre, Betarbet, and Sherer
2003). Interestingly, this concentration was not sufficient to inhibit respiration in brain mitochondria, but will partially inhibit respiration in liver mitochondria (Betarbet et al. 2000).
This has been explained by the ‘threshold effect’ seen in brain mitochondria where effects on oxidative phopshorylation and ATP synthesis were not observed until 72 % complex I inhibition was reached (Davey and Clark 1996). At high concentrations, rotenone can also bind non- specifically to proteins and hydrophobic compounds (Panov et al. 2005). For this reason, specifically for in vitro experiments, rotenone binding to plastic must be taken into account.
Data published in Betarbet et al, 2000 showing (a) complex I inhibition in sections from rotenone-infused animals (bottom) vs vehicle-infused animals (top) and (b) Titration of glutamate-supported respiration in mitochondria from liver and brain exposed to rotenone. The estimated brain concentration of rotenone (30 nM) which is known to inhibit complex I, does no
34 inhibit respiration in brain mitochondria, but it does in liver mitochondria. (Error! Reference source not found.).
Figure 1-11. Rat brain slices, showing complex I inhibition by rotenone using competitive binding with labelled dihydro-rotenone (a) and respiration rate in liver and brain mitochondria exposed to rotenone (b)
published by (Betarbet et al. 2000)
Rotenone toxicity has shown conflicting results, it has been shown to bind irreversibly to complex I of the respiratory chain (Grivennikova et al. 1997), and in vitro and in vivo studies have demonstrated that binding is necessary reproduce PD mechanisms such as ROS accumulation (Sherer et al. 2003; Dhillon et al. 2008; Furlong et al. 2015). Other studies using complex I subunit knockout mice (Ndufs4-deficient), have indicated that binding is not required to induce dopaminergic cell death and unknown off target effects may lead to loss of TH positive
35 cells in vitro (Choi et al. 2008; Higgins and Greenamyre 1996). Rotenone models of PD are widely used, however, studies have not yet shown whether neurons can recover from rotenone- induced toxicity, or what the chronic effects of low-dose early life exposures are.
1.2.6. Glyphosate exposure: public health controversy
Glyphosate, the active substance in Roundup®, is the most widely used herbicide in the world
(Error! Reference source not found.). It is an organophosphorus, which interferes with a key enzyme in the shikimic acid pathway, which is present only in plants. For this reason, it is a highly effective broad-spectrum herbicide, and is microbially degraded in soil to form ammonia, inorganic phosphate, and carbon dioxide. As it has not been found to be toxic to birds, fish or insects, it is considered a safe product for use. Currently controversy has arisen based on the fact that glyphosate has shown neurotoxicity in some studies and the loose assumption that the products of glyphosate metabolism could interfere with proteins (Samsel and Seneff 2015).This is based on the fact that glyphosate derives from glycine, an inhibitory neurotransmitter and is a manganese (Mn) chelator (Mesnage and Antoniou 2017). The further associations with disease have been contested as there is not sufficient evidence to support them (Faria 2015). Animal studies have had little power to show effects, and the increasing use of this herbicide indicates that further studies are needed to better understand its potential toxicity. At millimolar concentrations (40 mM), glyphosate induced autophagy-mediated apoptosis in PC12 cells (Gui et al. 2012). Commercial glyphosate pesticides (formulated with solvents) have been reported to inhibit mitochondria (Bababunmi, Olorunsogo, and Bassir 1979; Olorunsogo, Bababunmi, and
Bassir 1980) and cause neuronal degeneration in C. elegans (Negga et al. 2011). However, this has not been shown for glyphosate alone. A single clinical case has been recorded for accidental glyphosate exposure and development of PD symptoms (Barbosa Egberto et al. 2001). In vitro
36 tests can help determine the toxic effects of acute or long-term exposure to glyphosate (Van
Bruggen et al. 2018).
Figure 1-12. Molecular structure of glyphosate
1.3. 21st CENTURY TOXICOLOGY
Toxicity testing involving animals dates back to Paracelsus in the 15th century, quoted often for
‘the dose makes the poison’. Systematic animal testing in Toxicology began in the 1920s, when the dose, which acutely leads to 50% death (LD50) was determined as a suitable toxicity endpoint. Tragically, in the 1960s thousands of children were born with developmental disorders caused by Thalidomide exposure (a drug prescribed to pregnant women for morning sickness), therefore, pharmaceuticals became subject to intensive animal testing. In the late 1980s, the
Organisation for Economic Co-operation and Development (OECD) and the International
Conference on Harmonization (ICH) published guidelines for toxicity testing of chemical and pharmaceutical substances (Parasuraman 2011). Prior, in 1959, Russel and Burch had proposed the 3Rs framework to improve scientific and animal welfare (Russel and Burch, 1959), asking for refinement, reduction or replacement of animal experimentation wherever possible. Although these are the studies we learn to accept animal tests through textbook knowledge, evidence proves that they are not sustainable, reproducible or predictive (Hartung 2017a).
The first human immortalized cell line was developed in 1951, the HeLa cell (Masters 2002). In the past decade, we have made leaps towards refining animal use for research and reducing the number of animals used for pharmaceutical and chemical toxicity testing by screening substances
37
using non-animal and computational methods. However, regulatory agencies still require animal
studies to large extent. Considering human drug toxicity, animals (rats, rabbits, dogs) have shown
to have a positive predictive value (PPV) ranging from 50-70%, indicating these tests are not
always fit-for-purpose (Bailey, Thew, and Balls 2014; Gottmann et al. 2001). If animal studies,
which are based on high-dose testing using a precautionary approach, do not accurately predict
safe substances, the identification of harmful substances cannot be precise either. Importantly,
92–94% of all drugs that pass preclinical tests fail in clinical trials due to unforeseen toxicity6.
Animals are not genetically diverse, they are exposed under a controlled environment and studies
have low power.
In 2007, the National Academy of Sciences published a report titled ‘21st Century Toxicology: A
vision a strategy’ identifying gaps in current testing strategies and a roadmap to toxicity testing,
describing the need for ‘integrating computational models and in vitro assays based on cell
culture models and endpoints that are amenable for adaptation to high throughput screening to
be able to test a large number of chemicals’ (Bal-Price et al. 2010; Bal-Price et al. 2008). This
approach can exploit the current revolution in biotechnology, with the development of better
cellular models and high-throughput screening techniques. For example, the endocrine disruptor
screening program (EDSP) is a major effort driven by the Environmental Protection Agency
(EPA) to screen pesticides. In 2006, the European legislation on the Registration, Evaluation,
Authorisation and Restriction of Chemicals (REACH) reformed the regulatory requirements for
chemicals in Europe (Hartung 2010b) as had the cosmetics regulation in 2002 (Hartung 2008),
which involved by 2013 banning any product with ingredients which have been the subject of
animal testing7 . This latter regulation transformed the market and incentivized the cosmetic
6 https://www.the-scientist.com/?articles.view/articleNo/23003/title/More-compounds-failing-Phase-I/ 7 REGULATION (EC) No 1223/2009 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 30 November 2009 on cosmetic products. Official Journal of the European Union
38
industry to look for and use alternatives. Changes in this regulation were achieved for example
via the development of alternative methods to identify and predict skin irritants and sensitizers.
Specifically, neurotoxicity cannot be quantitatively assessed for compounds which are acutely
toxic at the high doses tested orally. To overcome this issue of subchronic to chronic exposure
extrapolation in risk assessment, an uncertainty factor of 10 is added to calculate the reference
dose (RfD) 8 . This presents problems for substances which may lead to long-term
neurodegeneration at low level exposures or target organ concentrations. Screening substances
prior to animal testing could identify those, which are neurotoxic even at low concentrations.
Other areas, which in vitro methods could revolutionize, are gene-environment interactions, inter-
individual variability, biomarker development and toxicity of mixtures. Although in vitro
neurotoxicity testing is still in its early stages, the development of a battery of assays which can
quantify neuronal/glial toxicity by measuring specific neuronal endpoints and include
metabolism, would provide a feasible testing strategy for substances (Schmidt et al. 2017).
1.3.1. Adverse outcome pathways (AOPs)
Although epidemiological studies can find associations between pesticide exposure and
neurodegeneration, there is a lack of data to inform and establish causality. For this reason, the
adverse outcome pathway (AOP) framework was developed by the OECD, helping assess and
provide weight of evidence to describe a series of biological processes (molecular inhitiating
events (MIE) and key events (KE)) which lead to an adverse outcome (AO) (Villeneuve et al.
2014). It is important to emphasize that AOPs are not compound-specific, but rather pathway-
specific, with the intention to determine thresholds for one KE, which lead to the initiation of a
following KE that could result in adversity (Bal-Price and Meek 2017). Although linear, further
8 Reference dose (RfD) is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. The RfD is generally expressed in units of milligrams per kilogram of bodyweight per day (mg/kg/day).
39 work has improved AOPs to incorporate feed-forward connections, shortcuts between KEs and thresholds. Specifically, for this thesis, AOP #3 is of importance and is published on the AOP
Wiki (https://aopwiki.org/wiki/index.php/Aop:3). Terron et al., describe this AOP in detail, providing evidence for a causal role between complex I inhibition and development of parkinsonian phenotype. They do so by providing the data to support the key event relationships
(KERs) (Terron et al, 2017). Currently there is greater emphasis on determining measurable KEs, similar to the need for better biomarkers of disease, to use AOPs within the IATA framework.
Specific in vitro tests would allow for assessing whether compounds lead to specific key events.
A Similar approach, termed pathways of toxicity (PoT), focuses on molecular pathways (which are finite within a cell), and how disturbingthese networks can alter cellular fate (Hartung and
McBride 2011; Kleensang et al., 2014). Biomarkers of toxicity (BoT) are needed to better extrapolate in vitro to in vivo effects. It is required that they be quantifiable, predictive of outcome, mirror the toxic response in vivo and provide information on the rate, magnitude and reversibility of a response (Blaauboer et al. 2012).
1.3.1. Repeated low-dose effects
Although often performed in animals, repeated-dose experiments are not common in vitro. This is due to some of the disadvantages of current models used: primary cells have a short life-span in vitro and cancer cells divide over time (masking any toxicity). The repeated dose 28-day oral toxicity study in rodents (OECD Test Guideline 407; http://alttox.org/mapp/toxicity-endpoints- tests/neurotoxicity/) is commonly used to study neurotoxicity. With the current knowledge on gene-environment interactions in neurodegenerative diseases, the high doses at which compounds are tested in animals may cause overt toxicity, which masks any subtle adverse effects, which may lead to disease in the long-term, especially if the individual carries a gene variant associated
40 with the disease. Worth and Balls presented the need for mechanistically relevant alternative methods, which could be used as a test battery to assess neurotoxic endpoints in tiers. The first tier would separate general toxicity from neurotoxicity and the second tier would focus on mechanistic endpoints (Worth & Balls, 2002).
In neurodegenerative diseases, final stages are well characterized via human post-mortem samples and in vivo studies. In the case of PD, decrease in dopamine production, Lewy body formation and > 50% dopaminergic cell death are observed in the final stages. However, the mechanisms leading up to PD pathology and how low-dose, lifetime exposures impact long-term disease outcome have not been addressed in vivo or in vitro. There is a need for in vitro assays that allow for long-term, repeated, low-dose exposures (Basketter et al. 2012). With the development of 3D models, this has recently become a possibility. Neurospheres from rat primary cultures, iPSC derived 3D models and 3D LUHMES are some of the current models used to study repeated-low dose effects.
1.4. CELLULAR RESILIENCE CONCEPT IN TOXICOLOGY
Few articles can be found studying cellular resilience. It is a postulated complex cellular mechanism in toxicology (Smirnova et al., 2015), which earlier has been mostly addressed within infectious diseases. The adaptive immune response allows us to keep immunological memory after an exposure to a specific pathogen. This system is highly adaptable and leads to specific memory immune cells (memory B and T cells) which can respond to subsequent exposures and fight infection (Mueller and Mackay 2016). With advances in molecular epigenetics, more research has described ‘memory’ in other cells types, as changes to the genome are heritable and can alter the cellular response. The best-known example is the Agouti mouse model exposed to
BPA, which shows how exposures can alter gene expression (Dolinoy 2008). Interestingly, the
41 other field, which has developed interest in these mechanisms is neuroprotection and plasticity; therefore, recent studies can be found on neuronal resilience and the processes of reverting "back to normal" and reversal of apoptosis ("anastasis") (Manji et al. 2000; Tyagi et al. 2015; Smirnova et al. 2016). The foundation of the hypothesis of this thesis is that cells can overcome low-dose toxicant effects (cell death is not triggered) and then can become resilient to subsequent exposures (via activation of cell survival pathways, changes in gene expression or epigenetic modulations. The term “anastasis” (Greek for “rising to life”) was described by Tang et al. who demonstrated that cells presenting apoptotic markers could recover when an apoptosis inducer is washed away, leaving some permanent genetic changes (Tang et al. 2012). In 2015, our group published a food for thought article on cellular resilience as a concept in toxicology, describing how cells under toxicant-induced stress, may become more susceptible, more resilient, or recover.
Molecular scars or imprinting underlie these responses and can determine long-term health
(Figure 1-13). Resilience is not always just the return to the prior state, as described above.
However, a cell that is not challenged is ‘bored to death’ (Hartung 2007) and robustness of a cellular system depends on its response to environmental stressors, it would be difficult to think that our cells cannot adapt to exposures over time. Epigenetic memory in the form of DNA methylation can protect or contribute to long-term pathogenesis or vulnerability to subsequenne exposures (Fraga et al. 2005; Tyagi et al. 2015).
42
Figure 1-13. Cellular resilience concept published in (Smirnova et al. 2015)
These pathways and mechanism of resilience, recovery and delayed responses decide the final outcome, and not those which are immediately activated upon insult (pathways of toxicity).
Basically, the hypothesis is that failure in coping with the toxicant insult is what triggers toxicity and neurodegeneration specifically. One aspect that has not yet been addressed is, whether these resilience mechanisms are beneficial or detrimental to cells in the long-term as permanent activation of these pathways or the ‘molecular memory’, may contribute to disease (Daskalakis et al. 2013). In context of neurodegeneration, cellular resilience and recovery concepts with possible delayed outcomes become especially relevant, since prolonged periods of latency between exposure and onset of the symptoms has been proposed already in 1991 (Reuhl 1991).
Dopaminergic neurons show selective susceptibility to rotenone toxicity compared to other cell types in the brain, researchers have hypothesized that this can be attributed to a variety of factors specific to dopaminergic neurons; (1) there are only 500,000 dopaminergic neurons in the substantia nigra; cell loss will have a greated effect on function (2) they produce dopamine; which can contribute to ROS production (Jeon et al. 2010) (3) their energy demand is high;
43 neuronal processes for dopaminergic neurons can span 4 m in length and high levels of ATP are needed to transport proteins along these processes (Bolam and Pissadaki 2012) and (4), they have a lower number of mitochondria, or decreased biogenesis upon toxicity (Liang et al. 2007).
Differences in susceptibility between dopaminergic neurons and other cell types have been shown in vivo and in vitro (Hirsch, Graybiel, and Agid 1988; Heusinkveld and Westerink 2017; Kweon et al. 2004). However, how the susceptibility of dopaminergic neurons changes over time or with multiple exposures has not been investigated to date. More studies are needed to better determine how permanent published acute effects are, and whether subsequent exposures lead to an adverse outcome.
Currently, adult neurogenesis is controversial, with studies showing conflicting results in animals and humans (Sorrells et al. 2018; Spalding et al. 2013; Altman and Das 1965). If any, neurogenesis is slow in the brain, therefore, neurons must be able to adapt to exposures, which occur over a lifetime. An example of molecular memory was described in rats, which underwent exercise daily or intermittently, showing that a brief second re-exposure to exercise (which was not enough to induce changes in animals who had never exercised), led to fast induction of BDNF in the hippocampus. This indicated that cells were primed to induce BDNF (Berchtold et al.
2005). By studying recovery and resilience, we can better identify where the threshold of an effect lies (whether it is reversible or not), the role it plays in disease and develop therapies which target neuroprotective mechanisms. Although cellular recovery is not a novel concept, it is not commonly studied (negative results are not easily published, therefore if no effect is observed it becomes obsolete, but it may be that cells are able to overcome acute toxicity). Further research into these mechanisms are critical for understanding chronic effects, altered susceptibility and mixture toxicities.
1.5. RESEARCH PRESENTED IN THE THESIS
44
The second and third chapter of this thesis, titled ‘LUHMES 3D dopaminergic neuronal model for neurotoxicity testing allowing long-term exposure and cellular resilience analysis’, presents two manuscripts which are published in Archives of Toxicology and Current Protocols in
Toxicology. These manuscripts describe the characterization of a LUHMES 3D neuronal model, to study dopaminergic toxicity. Current monolayer LUHMES cell cultures has the limitation, as do other neuronal models, that neurons tend to dissociate from plastic and coatings in culture, which only permits keeping them in culture for up to 10-12 days. To study effects after compound removal, washing the cells is required, and to ensure no compound remains attached to plastic (as it would leach slowly during culture), cells must be transferred to a new cell culture plate. To allow this, developing and characterizing a novel model was necessary. This research shows that LUHMES differentiate into a homogeneous population of dopaminergic cells in 3D, in the same way as they do in monolayer cultures. They express the same levels of differentiation markers and can be kept in culture for three weeks and can be used to study delayed effects after acute exposures and wash-out. A detailed protocol for cell culture and endpoints for characterization and toxicity measurements was further published.
After characterization of this model, we applied it to study the acute effects of rotenone as well as how cells recover from these effects 7 days after compound wash-out. This manuscript, titled
‘Toxicity, recovery and resilience in a 3D dopaminergic neuron in vitro model exposed to rotenone’ was submitted to Archives of Toxicology and is presented in Chapter 4. We quantified rotenone in the media before and after wash-out, showing that a significant amount can bind to plastic and cells during 24 h treatment. After wash-out, there was no rotenone in the cell culture media. After confirming this, we were able to study changes in complex I inhibition, ATP synthesis, neurite outgrowth, electrical activity and gene expression after acute exposure and recovery. We show how some acute effects of PD-inducing rotenone concentrations can be overcome upon compound removal in an in vitro disease model. This is the first in vitro study to
45 look at the reversibility of acute toxicity. We further tested, whether dopaminergic cells were more or less susceptible to a second exposure. The IC50 for dopaminergic cells exposed for the second time to rotenone was higher than for cells exposed for the first time, indicating protection to second exposures and some sort of ‘memory’ within recovered cells. With the field moving rapidly in this direction, more complex 3D models will provide further insights into the mechanisms behind toxicant-induced neurodegeneration, recovery and neuroprotection. Also, complex 3D models will allow studying dopaminergic cell responses in the physiological environment of other neural cells present. This will allow assessing the mechanisms why dopaminergic neurons are more sensitive than other type of neurons to the same toxicant insult.
In the fourth manuscript presented in Chapter 5 we studied low, repeated-dose effects of rotenone and the increasingly used herbicide, glyphosate. The manuscript titled ‘In vitro repeated low-dose effects of rotenone and glyphosate in 3D LUHMES’ is in preparation. Effects on neurite outgrowth, ATP levels and PD-related gene expression were measured. Genes, which are involved in PD progression were altered after repeated-dose exposure to rotenone. Glyphosate was not cytotoxic and showed no effects on neurite outgrowth at the concentrations tested.
In the sixth chapter, the manuscript titled ‘suitability of 3D human brain spheroid models to distinguish toxic effects of gold and poly-lactic acid nanoparticles to assess biocompatibility for brain drug delivery’ demonstrates the applicability of the LUHMES 3D model and iPSC-derived
BrainSpheres to study internalization and toxicity of drug-delivery nanoparticles. The use of single-cell type 3D LUHMES together with multicellular BrainSpheres allowed assessment and comparison of different susceptibility to NP toxicity. As an area of biomedicine, which is rapidly developing, the models presented in this paper could be used in conjunction to understand how different cell types respond to different nanoparticle types or coatings. The results presented are critically discussed in the final chapter of this thesis, with the aim to identify the impact of this research, potential pitfalls and research gaps.
46
AIM 1. Characterization of a 3D dopaminergic cell model to study nerotoxicity and recovery is published in the manuscripts presented in chapters 3 and 4.
AIM 2. The use of 3D dopaminergic model to study recovery and resilience to rotenone exposure is conceptually indtroduced in chapters 2, 3 and 4 with research supporting our hypothesis in chapters 5 and 6.
AIM 3. The application of 3D brain models to study nurotoxicity of nanoparticles used for drug-delivery is presented in chapter 7.
47
CHAPTER 2
2. CELLULAR RESILIENCE
“A good half of the art of living is resilience.” Alain de Botton
*The work presented in this chapter is published in the following article:
Smirnova, L., G. Harris, M. Leist, and T. Hartung. 2015. 'Cellular resilience', ALTEX, 32: 247-
60.
48
2.1. ABSTRACT
Cellular resilience describes the ability of a cell to cope with environmental change such as the exposure to a toxicant. If the cellular metabolism does not collapse directly, programs of stress response promote adaptation to a new homeostasis under stress or trigger coordinated cell death.
The processes of reverting “back to normal” have little been studied on the cellular level, and reversal of apoptosis (“anastasis”) has been overlooked for a long time. Different cell types show astonishingly similar vulnerability to most toxicants, except for those which require a very specific target, metabolism or mechanism present only in specific cell types; i.e. the majority of chemicals trigger “general cytotoxicity” in any cell at similar concentrations. We put forward the hypothesis that cells differ less in their vulnerability to a given toxicant, but in their resilience, i.e. how they can cope with this hit. In many cases, cells do not actually return to the naïve state after a toxic insult. The phenomena of ‘pre-conditioning’ and ‘hormesis’ describe this for low-dose exposures to toxicants, which render the cell more resistant to subsequent hits. The underlying defense and resilience programs include a number of epigenetic changes, which leave a
‘memory/scar’, an alteration as a consequence of the stress the cell has experienced. These memories might have long-term consequences, positive (resistance/tolerance/robustness) and negative ones, which contribute to the chronic and delayed manifestations of hazard and ultimately disease. This article calls for more systematic analyses, of how cells cope with toxic perturbations in the long run after stressor withdrawal. A technical prerequisite for this are long- term stable (organo-typic) cultures and a characterization of molecular networks of cellular stress responses.
49
2.2. INTRODUCTION
Resilience is the ability of a system (here, a cell) to cope with negative change. The concept has been used in many areas from ecology to material sciences, engineering and disaster research.
Resilience can be seen as the opposite of vulnerability, though views differ dependent on the area
(Linkov et al., 2014). In toxicology (especially in vitro toxicology), however, the term and the concept are not well developed. Cells, organs and organisms and their vulnerability are dependenton their capacity to cope with (disastrous) changes, i.e., exposure to a toxicant. Disaster research has been moving away from preparing for each and every possible hit toward a concept of resilience, especially involving critical infrastructures1 (di Mauro et al., 2010). For example, one of the critical infrastructures of in vitro toxicology are mitochondria a, an Achilles’ heel of cells, where oxidative stress occurs in response to many hazards, triggering apoptosis by cytochrome C release. Given the endosymbiotic theory on the bacterial origin of mitochondria
(Wallin, 1923), this could be interpreted as the late manifestation of a chronic infection of the cell. It is tempting to develop testing strategies for hazardous substances based not on the apical manifestations but on the critical infrastructures that trigger the problem. This might be more efficient than identifying the many possible interactions of substances (now called molecular initiating events (MIE) in the context of Adverse Outcome Pathways (AOP)) or characterizing the entire Pathway of Toxicity (PoT, Kleensang et al., 2014). We can interpret these critical infrastructures as the nodes of the PoT networks, which would lend themselves as biomarkers of toxicity (Blaauboer et al., 2012).
50
Disaster research aims for such mapping and monitoring of critical infrastructures (providing services that are deemed vital for the functioning of society). The etymological root of ‘critical’ is linked to the term ‘crisis’, which refers to a ‘change in state of a system’, implying a time of great difficulty or danger. The logical counterpart is to characterize the vulnerability, which directly corresponds to resilience, i.e. the ability to cope with the hit. The third component is the probability of a hit, which determines the risk and is most difficult to assess both for societies and in our case toxicology. We can only say how often something has been hit in the past, i.e. the prevalence of certain modes of action of substances. But there can always be surprises, i.e. so- called “Black Swans” (Taleb, 2007): Black swan events are defined by the “triplet: rarity, extreme impact and retrospective (though not prospective) predictability”. Thalidomide was for example such a Black Swan in toxicology.
This article explores the resilience component of toxic action on a cellular level (Figure 2-1). On an organism level, this is typically covered as recovery and reversibility and plays an important role for example in the classification and labeling of substances. On a cellular level this has been not in the scope of most studies, which is likely owed to the fact that we too often study short-
51 term effects putting emphasis on cytotoxic actions of substances. However, it is of limited relevance for most hazard manifestations except for acute, high-dose intoxications.
In the second part, this article goes one step further: resilience is not just about the cell going back to “normal”, but how the insult changes the cell and imprints on its future functionality and responses. The wounds leave a systemic type of memory effect, figuratively speaking a “scar”, which is can be maintained among others by epigenetic mechanisms or mutations. A resilient cell is not necessarily a healthy cell, e.g. it could be cancerous and very resilient against chemotherapy. Some of the best examples for resilience are found in the field of chemotherapy.
Some tumor cells develop a high resilience and become resistant to drugs, although they are exposed to the same concentrations as their neighboring cells.
Such changes can be long-term, or even permanent; cellular memories can be beneficial and we will discuss cellular hormesis in this context. On the one hand, the concept of beneficial effects is more developed in biomedical research, in particular with respect to ischemia-reperfusion as a stressor to organs, and so called ‘pre-conditoning’ (= making cells more resilient to subsequent stress) is used experimentally and clinically (Wang et al., 2015; Clapp et al., 2012; Wu et al.,
2012; Yellon and Hausenloy, 2005; Dunn et al., 2012; O’Neill et al., 2012). Tolerance is a similar concept, where small doses of a toxicant (e.g. the famous arsenic eaters of Styria, Heisch, 1860) or toxin (e.g. endotoxin, Lehner and Hartung, 2002) protect against subsequent stronger hits.
These concepts from in vivo can to some extent traced back to cellular changes (Hartung and
Wendel, 1992). On the other hand, long-term effects can also be detrimental, lead to adverse outcome and will be critical for our understanding of late manifestations, changed susceptibilities and mixture toxicities, especially when the exposure is of limited duration. The resulting ‘late consequences of early life stress’, also termed ‘Barker hypothesis’ (Hales and Barker, 1992) have become a major theme in epidemiological research, public health and mechanistic research
52
(McGovan 2009; Suderman et al., 2012; Yehuda et al., 2015; Sebert et al., 2011; Lindblom et al.,
2015; Alastalo et al., 2012; Lau and Rogers, 2004).
Figure 2-1. The Cellular Resilience Concept: Survivable toxic insults create cellular stress; Pathways of Defense (PoD) might allow cells to return to a normal state; however, resilience programs often leave cells in an altered state, e.g. with an epigenetic scar, which might contribute to long-term manifestations of hazard but could also be a target
for therapeutic strategies.
2.2.1. Consideration 1: It is not important whether you fall, but whether you get up again
This is not only true for the boxer, but for each and every hit that we or a cell takes. Can we keep fighting? What is the functional impairment? Can it be restored? What is the resulting vulnerability for further hits of the same or a different type?
The in vitro toxicological literature is not rich with respect to such questions at a cellular level.
Some aspects were addressed in recent EU projects such as SEURAT-1, ESNATS, Predictomics etc., but their focus was still largely on the initial damage to the models. There are only few well- defined exceptions, mainly deriving from the fields of carcinogenesis and heat shock response, as
53 cancer cells have evolved a number of strategies to increase their resilience towards the toxic influence of chemotherapy. They involve upregulation of anti-apoptotic proteins, and drug efflux transporters (Leist and Jaattela 2001, 2002; Hansson et al., 2003; Hanahan and Weinberg, 2011).
The design of toxicological studies at the organism level, however, covers such type of questions very well. Morphological changes in the target organ as well as behavioral abnormalities are very often addressed immediately after exposure as well as after a recovery period. The same design of toxicological tests on the molecular and cellular level is of big advantage to understand molecular mechanisms of organ/organism recovery and adaption.
How long does a perturbation last? How is homeostasis reestablished? There must be elasticity, which allows returning to normal. This requires some sensing and counter-regulations. A number of cellular stress responses have been described (rearrangements in energy metabolism, oxidative stress response, activation of anti-apoptotic pathways, and DNA repair mechanisms), but their actual contribution to reestablishing homeostasis is often not clear. These stress response pathways (SRP) include hypoxia signaling via HIF-1, the heat shock response via HSH-1, the antioxidant response via NRF-2, stress kinase signaling via JNK and AP-1, DNA damage responses via e.g. p21 or BSCL2, and the unfolded protein response/amino acid starvation response via ATF-4/ATF-6 (Limonciel et al., 2015; Jennings, 2013; Wink et al., 2014; Hendriks et al., 2012) We have earlier in this series discussed homeostasis under stress (Hartung et al.,
2012), which is what we often measure when characterizing toxic signatures by omics technologies,. However, the restoration process, which occurs when removing the stressor, is much less addressed.
The hypothesis is put forward here, that these are actually the processes, which decide about the long-term manifestation of hazard or recovery. Most toxicants are encountered at doses far below cytotoxicity, but to the extent that is enough to affect biology. Thus this understanding of perturbation and restoration should drive our analysis of pathogenesis and reversibility.
54
2.2.2. Consideration 2: Anastasis – awaken from the dead
Quite surprisingly, also cellular suicide attempts can be stopped. Most recently the term
“anastasis” (Greek for “rising to life”) has been coined (Tang et al., 2012 and 2015). The group observed “unexpected reversal of late-stage apoptosis in primary liver and heart cells, macrophages, NIH 3T3 fibroblasts, cervical cancer HeLa cells, and brain cells. After exposure to an inducer of apoptosis, cells exhibited multiple morphological and biochemical hallmarks of late-stage apoptosis, including mitochondrial fragmentation, caspase-3 activation, and DNA damage. Surprisingly, the vast majority of dying cells arrested the apoptotic process and recovered when the inducer was washed away. Of importance, some cells acquired permanent genetic changes and underwent oncogenic transformation at a higher frequency than controls.
Global gene expression analysis identified a molecular signature of the reversal process”.
Transcriptional responses were found to be critical for this reversal and inhibition of classical survival genes BCL-2, XIAP, MDM2, or HSP90 significantly suppressed reversal of apoptosis.
Though this may seem an isolated finding, there are frequent reports in the literature that cells may survive apparently lethal damages, such as rupture of the plasma membrane (Roostalu und
Straehle, 2012; Jaiswal et al., 2014), release of cytochrome C to the cytoplasm (Potts et al., 2003;
Deshmukh and Johnson, 1998), membrane blebbing (Foghsgaard et al., 2001) or caspase activation (Leist and Jaattela, 2001). It needs to be clarified further whether such cell culture observations are relevant in vivo, where such cells would be removed by phagocytosis before they can recover (Leist and Jaattela 2001, Hirt et al., 2000, 2003), but at least in Drosophila, transient caspase activation has been documented in cells that were not removed (Tang et al., 2015).
So even after the most extreme impact, programmed cell death, which has initiated, is reversible to quite some extent. However, reversibility may not lead exactly to the ground state but to altered cellular states, for instance related to senescence (Jurk et al., 2012) or involving permanent DNA damage (Ono et al., 2003; Vijg et al., 1997; Tang et al., 2012).
55
2.2.3. Consideration 3: All cells are equal(ly vulnerable)
Astonishingly, cells are very similar in susceptibility to toxicants at the level of cytotoxicity as was demonstrated by several studies where different cell types have shown comparable responses to the toxicants regardless tissue of origin and significant correlation between cytotoxicity in vitro and LD50 in vivo. Probably, Willi Halle was the first to notice on a larger scale that different cells show cytotoxicity to a given chemical at very similar concentrations. He started the Halle register, a large manual collection of IC50 concentrations from published cell experiments first published in 1988 (Halle and Goeres, 1988); and later translated and published by ECVAM in 2003 (Halle,
2003). The principal idea of this work was to use the geometric mean of the collected IC50 values (in mmol/L medium) and the corresponding acute oral LD50 for rats or mice (in mmol/kg) and calculate a simple linear regression model. There was clearly a positive correlation, though this was not found good enough to predict LD50 values in later validation attempts (NIH, 2006) or even the then recommended prediction of start doses for LD50 testing was challenged (Schrage et al., 2011). It is quite remarkable, still, that this approach works to some extent, especially for the prediction of not acutely toxic substances, for which it is now recommended by ECVAM
(Prieto et al., 2013). Halle concluded (2003): “The results of linear regression analysis showed that the biostatistical parameters obtained with IC50/LD50 values for xenobiotics taken from various publications …, and from the US National Institute for Occupational Safety and Health’s
Registry of Toxic Effects of Chemicals (NIOSH RTECS; ..), are comparable within a certain range, despite the fact that the various laboratories used different cell types, Standard Operating
Procedures (SOPs) and cytotoxic endpoints”. Here, especially, the aspect that as a mean of different cytotoxicity assays a value characterizing the toxicity of a substance can be generated is of interest.
The next very similar attempt was the MEIC (Multicentre Evaluation of In vitro Cytotoxicity) program (Clemedson and Ekkwall, 1998), which showed a good correlation (around 70%)
56 between in vitro basal cytotoxicity data and human lethal blood concentrations. In MEIC, 50 reference chemicals were tested in 61 in vitro assays (Ekkwall, 1999). A principal component analysis “indicated high general similarity (around 80%) of all the results from the 61 methods.
According to the new ``random probe'' analysis, this similarity must depend on the high correlation of results from assays with different cell types (mean R2 0.81) and/ or different viability endpoints (mean R2 0.85). Main factors contributing to the 20% dissimilarity of results were different exposure times and the use of phylogenetically distant test objects in the non- analogous ecotoxicological assays.” (Clemedson and Ekwall, 1999). To study the relevance of in vitro results, IC50 values were compared with human lethal blood concentrations (LCs) by linear regression. An average IC50 for the ten 24-hour human cell line tests predicted peak LCs better
(R2 0.74) than other groups of tests (Ekwall, 1999). This claimed predictivity formed the basis for the A-cute-Tox project (Clemedson et al., 2012). In this FP6 EU project, the correlation of in vitro cytotoxicity with animal LD50 data and human lethal blood concentrations was further evaluated, and clearly lower correlations were found. However, also in this project, many different cytotoxicity assays showed a significant correlation in IC50 values, independent of the cell type used (Kinsner-Ovaskainen et al., 2013).
Recently, Lin and Will (2012) “investigated the utility of hepatic-, cardiac-, and kidney-derived cell lines to (1) accurately predict cytotoxicity and (2) to accurately predict specific organ toxicities. We tested 273 hepatotoxic, 191 cardiotoxic, and 85 nephrotoxic compounds in HepG2
(hepatocellular carcinoma), H9c2 (embryonic myocardium), and NRK-52E (kidney proximal tubule) cells for their cytotoxicity… the majority of compounds, regardless of their designated organ toxicities, had similar effects in all three cell lines. Only approximately 5% of compounds showed differential toxicity responses in the cell lines with no obvious correlation to the known in vivo organ toxicity”. Another study showed that neuronal cells do not react differently to neurotoxicants than non-neuronal cells (Stiegler et al., 2011). However, differences in sensitivity
57 to toxicants have been reported for mouse embryonic stem cells differentiated into different lineages (Visan et al., 2012; Seiler and Spielmann, 2011), suggesting, that the developing system
(differentiating cells) could be an exception and possibly linked to the fact that they are more vulnerable to the toxicants than mature or undifferentiated cells. Another exception could be higher sensitivity of the cells in S-phase of mitosis to the drugs and toxicants, which is broadly used in cancer therapy.
One reason for the non-selectivity on the level of cytotoxicity testing is that the majority of chemicals is promiscuous with respect to toxicity targets, as observed in the US EPA high- throughput screening project ToxCast : “the majority of chemicals represented in the ToxCast phase I library likely act via nonselective interactions with cellular macromolecules” (Thomas et al., 2013). “976 structurally and categorically diverse chemicals in the ToxCast library across
331 biological assays: a quarter of the 976 compounds tested showed no demonstrable activity
(AC50) in any of the assays…specific or promiscuous activities…. a chemical affected 10 assays on average, ranging from 0 (274 chemicals) to 90 (1 chemical).” (Sipes et al., 2013).
Taken together these different studies make a very strong case that different cells of the same species are similar with regard to cytotoxicity and do not explain organ-selectivity of toxicants.
Obvious exceptions are the few compounds that show differential effects in fresh primary hepatocytes due to metabolic activation or deactivation not taking place in other cells. The limited predictivity of in vitro assays for the animal toxicity in 28 day or longer-term studies (Thomas et al., 2012) means that another component is necessary to explain why a given substance targets specific organs. Perhaps measuring cytotoxicity is wrong from the beginning? Measurement of functional endpoints and activation of stress-response pathways at sub-cytotoxic concentrations may be a way forward. Unfortunately, only few studies have been comparing functional cellular endpoints at subcytotoxic concentrations for many substances . An analysis of the ToxCast dataset seems to be most promising here. ToxCast does include eight cytotoxicity tests.
58
Noteworthy, the effective concentrations of different assays for the same chemical were very close: the concentration where a substance was positive in the first assay to the concentration, where it activated 10% of the assays it was positive in, differed only by a factor below three
(Thomas et al., 2013).
2.2.4. Consideration 4: Kinetics cannot explain all organ-selectivities
Some toxicants, especially environmental chemicals, may have a promiscuous effect on many organs, but some are very target-specific and/or need to be metabolized. Thus, differences in toxicokinetics, i.e. differences in absorption, distribution, metabolism and excretion (ADME) of chemicals across different body locations create organ selectivity, e.g.
• Topical (local) toxicities of skin, eye, lung etc.
• Liver first pass effects, leading to accumulation of xenobiotics absorbed in the gut and in the liver
• Differences in metabolic activation, again especially known for the liver and kidney
• Biological barriers, such as the blood-brain barrier or the blood testes barrier
• Specific transporters into cells for e.g. microcystin (liver), paraquat (lungs), MPP+
(dopaminergic neurons)
• And others
If kinetic and ADME can be addressed in vivo, the combination of some rough pharmacokinetic modeling with in vitro cytotoxicity data, however, is challenging and does not always improve in vivo hazard prediction from high-throughput in vitro toxicity assays. In fact, (Wetmore et al.,
2013) found that: “Adjusting the in vitro assays for pharmacokinetics did not improve the ability to predict in vivo effects as either a discrete (yes or no) response or a low effect level (LEL) on a
59 continuous dose scale.“ This may again be due to the simple cytotoxicity assays being not optimal starting point altogether.
One example of organ selectivity not linked to pharmacokinetics is the selective toxicity of the neurotoxicant 1-methyl-4-phenylpyridinium (MPP+) to dopaminergic neurons of the nigrostriatal pathway (Efremova et al., 2015), while the neighboring mesolimbic pathway is hardly affected.
The different types of dopaminergic neurons seem to cope with this chemical insult in different ways.
However, as we have to assume that many substances are not eliminated very quickly from the blood stream, but actually bind to serum proteins, most cells of the body, if not protected by blood barriers such as the brain, will actually be exposed over prolonged periods of time to more or less the same concentrations
2.2.5. Consideration 5: Are differences in cellular resilience underlying organ-selectivity of toxicants?
There are two common explanations, why many chemicals show organ-selectivity in vivo as discussed above: 1. The unique presence of specific target structures leading to different susceptibilities and 2. Differences in substance kinetics reaching higher concentrations of the substance or its toxic metabolite in a certain part of the body. However, the differences in susceptibility of different cell types in vitro are often not very pronounced. However, most cells used in vitro also do not have the same phenotype as in vivo, especially concerning the specific targets of toxicity, and the required metabolism (Coecke et al., 2006). Systemic levels of the toxicant can be the same and adjustment for tissue concentrations did not dramatically improve the in vitro to in vivo extrapolations. This shall not belittle the role of kinetics for extrapolation from in vitro effective to corresponding in vivo dose (Basketter et al., 2012, Leist et al., 2014), but to point out its incomplete explanation of organ selectivity of substances. Therefore, a third
60 alternative explanation is put forward here: Perhaps it is less the susceptibility to be hit by a toxicant but the ability to recover from this hit, which makes the difference. The condensed hypothesis put forward here is: All cells are equal(ly vulnerable), but some are more resilient than others…
The concept of cellular resilience as the differing ability of cells to cope with damage includes properties such as the ability to mobilize alternative energy sources and other redirections of metabolic resources, the elasticity of the metabolic network, the synthesis of defensive molecules such as anti-oxidants and other stress response elements as well as the induction of repair.
It is often assumed that the robustness of many complex systems is rooted in their redundancy, which for networks represents the existence of many alternative paths that can preserve communication (such as metabolic flows and regulatory gene networks) among nodes, even if some nodes are absent. Reka and Barabasi (2002) review the state of the art in the field of complex communication networks and highlight that previous research attempting to address this issue in quantitative terms failed to uncover to what degree redundancy plays a role. It is quite surprising that many gene knock-outs actually have no or little phenotype without inactivation of another gene or additional environmental stress (10-15% show little or no phenotypic effects
(Melton 1994, Barbaric et al., 2007) showing a biological robustness of the system. The rate of knock-outs without phenotype is difficult to estimate in mice, because negative data are often not published; in yeast the rate is ca. 40-60%. Often stress to the system such as infection, hypoxia, temperature changes or toxicity is required to show that responses are impaired. But have some cells less redundancies than others? This is not quite clear. As redundancy and robustness refer more to the initial set-up than to the difference in coping with the hit, this does not really further the argument here. The point seems to be whether cells reach a tipping point to collapse (Scheffer et al., 2012) and whether this point is different for different cell types depending on their resilience programs?
61
Components contributing to cellular resilience are likely the stress responses of the cell, which include repair enzymes, cell membrane repair (Steinhardt, 2005), the mechanisms to remove denatured proteins and other cellular trash, heat-shock proteins (Velichko et al., 2013), anti- apoptotic mechanisms (Brink et al., 2008), released inflammatory mediators (Finch et al., 2010) and growth factors, damage limiting (e.g. anti-oxidative) components, the mobilization of additional energy etc.. So the question to be answered is: what happens in the cells after hit but before they enter into necrotic or apoptotic cell death programs. We have addressed this in our research in two studies relating to Parkinson’s disease most recently. In the first model (Krug et al. 2014), dopaminergic neurons were exposed to the Parkinson toxicant MPP+, the metabolite of the illicit drug (meperidine) contaminant 1-methyl-4-phenyl-tetrahydropyridine (MPTP). MPTP is not itself toxic, but owing to its high lipophilicity, it is able to cross the blood brain barrier, where it is metabolized in astrocytes by monoamine oxidase B (MOA-B) to MPP+, which is then transported selectively by the dopamine transporter into neurons and inhibit mitochondrial electron transport chain, that ultimately lead to oxidative stress and apoptosis.
In the (Krug et al., 2014) project, human dopaminergic neuronal cells (LUHMES) where exposed to MPP+, and then were analyzed using combined metabolomics and transcriptomics approaches to identify the earliest cellular adaptations to stress. When mitochondria parameters were at control levels, strong transcriptome and metabolome changes were seen: depletion of phosphocreatine and oxidative stress (e.g., methionine sulfoxide formation), altered glucose flux showed a complex pathway of toxicity. This included the interference of energy metabolism,
ROS formation, ER stress, gene expression that ultimately led to mitochondrial cytochrome-C release and apoptosis. A strong increase of S-adenosyl-methionine (SAM) and early activation of transsulfuration pathway was observed that increased glutathione level. Bioinformatic analysis of our data identified the transcription factor ATF-4 as an upstream regulator of early responses.
Findings on this signaling pathway and on adaptive increases of glutathione production were
62 confirmed biochemically. Metabolic and transcriptional profiling contributed complementary information on multiple primary and secondary changes that contribute to the cellular response to
MPP+. ATF4 has been identified as a key transcriptional factor in MPTP toxicity also by others
(Ye et al., 2013). This illustrates how the cells struggle to survive before apoptosis sets in, representing candidate PoD in the resilience of these cells.
In the second project (Maertens et al., 2015), we have analyzed microarray data derived from brains from MPTP treated mice (Miller et al. 2004) and carried out Weighted Gene Correlation
Network Analysis (WGCNA), supported by text-mining, and other systems-level technologies to construct a genetic regulatory network for MPTP toxicity. The paper was discussed in two guest editorials (Rahnenführer and Leist, 2015; Andersen et al., 2015). Several modules of connected genes were identified, which overrepresented annotations for neurodegenerative diseases and transcription factor analysis identified SP-1 as key regulator, which is known to regulate the dopamine transporter (Wang and Bannon, 2005), and is involved in several neurodegenerative diseases (Qiu et al., 2006, Santpere et al., 2006). Interestingly, SP-1 was not detected as an important game player using conventional statistical methods of gene expression analysis. In addition to SP-1, the network hubs consist of some candidates well known for their role in
Parkinson’s disease (STAT3, JUN). SREBF1, also identified in this study, has previously been identified as a risk locus for sporadic Parkinson’s disease (Do et al., 2011) and in a recent RNAi screening study, it was implicated in the control of the PTEN-induced kinase 1 (PINK1)/Parkin pathways that control the autophagic destruction of mitochondria (Ivatt and Whitworth, 2014).
One hub, HDAC1, has been implicated in cell-survival in neurotoxicity to dopaminergic neurons in vitro and ischemia in vivo (Kim et al., 2008), thus a candidate PoD. The WGCNA network also suggested a protein, LANCL1, that was connected to both HDAC1 and STAT3, which binds glutathione and is believed to play a role in neuronal survival following oxidative insult (Zhong et al., 2012). Notably, ATF-4, identified in the cell culture experiments above, was also present as a
63 hub in the Weighted Gene Correlation Network Analysis. This study shows, that WGCNA can help to identify not only the components of the toxic insult but also the initiation of PoD as elements of cellular resilience.
Thus, combined omics analysis is a new unbiased approach to unravel earliest metabolic changes, whose balance decides on the final cell fate. Similarly, we now hope to unravel the pathway of defense and resilience when the stressor is withdrawn. A prerequisite for this was the development of a 3D organoid culture of the LUHMES cells (Smirnova et al., submitted), which allows to culture cells longer and to transfer the organoid into uncontaminated culture dishes for toxicant withdrawal and recovery studies.
2.2.6. Consideration 6: How to challenge the concept?
The first step needs to be the characterization of cell stress and its return to normal/new homeostasis, favorably by a combination of omics technologies that includes non-coding RNAs and epigenome to generate high-content data sets. Such largely untargeted characterization comes with many challenges as experienced in the Human Toxome project (Bouhifd et al., 2015).
Central problem are the signal to noise problem and the small-n fallacy: It is very difficult to identify a few meaningful genes out of the almost 30,000 when there is a lot of biological and technical variability and only a limited number of measurements possible (Krug et al., 2013).
Other omics technologies such as metabolomics are even less standardized (Bouhifd et al., 2013,
Ramirez et al., 2013, Bouhifd et al., 2015 this issue of ALTEX). A way forward is tracing the signatures of toxicity back to mechanism (Hartung and McBride, 2011), but the incomplete mapping of pathways in the different databases is a major challenge (Kleensang et al., 2014).
Workflows as suggested earlier (Maertens et al., 2015) though can derive candidate pathways from such untargeted characterization, and from our experience weighted genome correlation network analysis represents a key tool to overcome aforementioned shortcomings. Targeted
64 follow-up measurements, transcription factor analysis and qualification of results by linguistic search engines and systematic literature reviews help further.
The next step will be the systematic intervention in these pathways with gene silencing technologies or pharmacological inhibitors, the “mechanistic validation” (Hartung et al., 2013). In case of resilience pathways, the expectation would be that these delay or hinder the restoration of homeostasis or functional capacity to levels before the hit, limit the protective effect against a second hit (see below) and will possibly result in a shift of the concentration response curve of cytotoxicity as a proxy of organ selectivity.
The ultimate step will be dynamic modeling of the perturbed cell and its resilience program.
Buchman (2002) suggested that (cellular) homeostasis arises through the combination of specific feedback mechanisms and spontaneous properties of interconnected networks, making it
“dynamically stable”. Manke et al. (2006) used dynamical systems theory for data from large- scale protein interaction screens in yeast and C. elegans to demonstrate entropy as a fundamental invariant and as a measure of structural and dynamical properties of networks. Tyson et al.,
(2003) interpreted the dynamics of regulatory and signaling pathways in the cell as “strikingly similar to the wiring diagram of a modern electronic gadget. Instead of resistors, capacitors and transistors hooked together by wires, one sees genes, proteins and metabolites hooked together by chemical reactions and intermolecular interactions.” Some reviews of methodologies are available (Koch and Ackermann, 2012; Jack et al., 2013; Hoeng et al., 2014, Sturla et al., 2014;
Sauer et al., 2015). In pharmacology, drug action is increasingly interpreted as interference with such complex networks (Hood and Perlmutter, 2004; Araujo et al., 2007, Kreeger and
Lauffenburger, 2010).
A living cell is a complex, dynamic system comprised of hundreds of thousands active genes, transcribed mRNA, proteins with all of their modifications, metabolites, structural constituents
65 from lipids and carbohydrates to mention only a few, which come to mind first. All of this is undergoing even under homeostatic conditions continuous change and exchange regulated by complex interactions in networks resulting in rhythmic and chaotic patterns. This becomes even more complex if we see a population of cells, different cell types interacting or then the organ functions they form and their systemic interaction in the organism. Even worse, life means reacting to the environment, which is constantly impacting on all levels of organization. It is illusive to fully describe such a complex system and model it. It is also naïve to take any component and expect it to reflect the whole system. The goal must be to understand enough of system to understand the major impacts, which is essentially what research into diseases or toxicology is about: Understanding the impacts which make a lasting and severe change to the system.
To use an analogy, to understand the traffic in a larger city, we need to characterize a system of hundred thousands of pedestrians, cars, bicycles etc. But we do not need and we cannot understand each and every element’s behavior to understand that something impacted. If there was a traffic accident, we see patterns of changes (traffic jam, redirection of flow, emergency forces deployed etc.). If we take a snapshot photograph from a satellite of the situation, we might already see certain clusters or the appearance of ambulance cars. Even better if we can visualize fluxes and show where flow is hindered and see the direction of movement.
Omics technologies in combination with WGCNA are these type of satellite photographs, usually just as a snapshot of the system. By comparison with the “normal” situation, we can start to identify the major cellular derangements, especially when we have time series, replicates and dose-response analysis available, best if we can determine fluxes. We do not need to monitor each and every “car”, some of them suffice to characterize what happens on the main “roads” and
“places”; some of them are more telling such as the ambulances, the police cars or the firefighters. Different types of interferences can result in similar patterns (accident, construction
66 work, a sport event) if hitting the same place/region. The stronger the disruption, the more easy to detect perturbation at places further away or whatever we measure (a traffic jam will impact little on pedestrians and bicyclists, a roadblock does).
The analogy falls short when we see that our omics snapshot photographs are selective, they see either mRNA, proteins or metabolites etc. This would be a camera seeing only cars but missing the anomalies of a marathon or a bicycle race taking place in the city. In order to understand these situations, we need to combine our monitoring.
A few lessons from our analogy:
• A dynamic system can hardly be understood from a single snapshot.
• Repeated and varied measurements especially of different components will give a more robust view on the system.
• The better we understand normal state and earlier perturbations, the better we know where and what to monitor and how to interpret it.
• Knowing the early and stress responses (the ambulance and police cars) is a good way to sense trouble even not knowing why they were deployed.
• Simulation of traffic helps planning and can be done understanding only the major principles of the system.
• The stronger the hit to the system and the longer lasting the effect, the more likely we will see it and interpret it correctly.
For toxicology, however, such systems approaches (Hartung et al., 2012) are still “pie in the sky”, but virtual experiments will at some point show how these networked systems achieve their elasticity and resilience when exposed to toxicants.
2.2.7. Consideration 7: Resilience is not the return to the prior state
67
There are three ways for the cell to deal with the hit/stress: what does not kill us, makes us either stronger or impaired. The challenge of a cell by a toxicant induces defense mechanisms
(discussed above) and this can on the long run result in protective effects. This phenomenon has been termed among others “hormesis” (Calabrese and Baldwin, 2001; Calabrese and Blain, 2005) in toxicology and radiation biology. It describes the phenomenon that cell viability or biological fitness in general increase when a system is exposed to low concentrations of a stressor. Hormesis in this sense is the result of resilience, i.e. the cell induces a stress-and-defense program.
Nicolas Taleb has addressed permutations of this concept in his book “Antifragility” (2012):
“Antifragility is beyond resilience or robustness. The resilient resists shocks and stays the same; the antifragile gets better. … Some things benefit from shocks; they thrive and grow when exposed to volatility, randomness, disorder, and stressors and love adventure, risk, and uncertainty.” Interestingly, he notes “complex systems are weakened, even killed, when deprived of stressors”, which resembles very much the note in an earlier article in this series, that cell culture bores cells to death (Hartung, 2007): We argued there, that cell mass and functionality is not maintained in cells pampered with nutrients and no demand on metabolism and cell function.
Environmental stress continuously compromises biological system (proper development, cell cycle, signaling pathways, etc.). Robustness of the biological systems against environmental stress is crucial in many aspects of the proper functionality, including development programs.
Robustness can be seen as a part of the resilience concept: there are certain regulatory molecular mechanisms, which play against the stressors to maintain the proper functioning.
Taleb addresses natural systems several times: “It is all about redundancy. Nature likes to overinsure itself. Layers of redundancy are the central risk management property of natural systems.” This is quite in line with genetics (two alleles plus many gene copies and variants) and the lack of effect of many gene knock-outs. Macia and Sole (2009) pointed out that it is not only
68 redundancy but also degeneracy, i.e. the ability of elements that are structurally different to perform the same function or yield the same output, such as alternative metabolic pathways
(Tagore and De, 2012), which results in the robustness of cellular networks. Unraveling the cellular signaling networks begins to explain how a cell can exhibit an apparent paradox of robustness to toxic perturbations while responding specifically and sensitively to relevant inputs
(Araujo and Liotta, 2006). One of these cellular signaling networks regulating robustness is posttranscriptional regulation of gene expression by microRNA through positive and negative feedback loops (Herranz and Cohen, 2010; Ebert and Sharp, 2012). Several studies have shown how microRNA may buffer the altered “noisy” gene expression and thus maintain steady-state of the system. Most important aspect in this type of regulation is genetic and functional redundancy of microRNAs that makes them stable against environmental stress. This explains little or no phenotypes in individual microRNA knockout experiments (Miska et., al 2007) and appearance of the phenotype only upon stress (summarized in Ebert and Sharp, 2012). Some microRNAs were shown to stimulate cellular resistance to environmental stress conditions, e.g., hypoxia (e.g., mir-210, mir-424, Chan et., al 2012; Loscalzo 2010) temperature changes (e.g., mir-34, mir-83,
Burke et al 2015), pathogenic stress (e.g., let-7 family, Ren and Ambros 2015), others were shown protective properties against toxicant exposure (e.g., mir-7, mir-153, Fragkouli et al 2014,
Choi et al 2014). These make microRNA a good candidate to contribute to cellular resilience.
But this setup appears to explain more why the system is robust, can take individual hits. It does not explain, how it learns and becomes better. Can other epigenetic mechanisms such as DNA methylation answer this question? The epigenome may drive response mechanisms to environmental stress on the interface between the dynamic environment and the inherited genome possibly allowing an “epigenotoxic effect” (Szyf, 2007). Alterations in DNA methylation and histone modifications have been associated with errors in autoimmune function, nervous development and diseases such as cancer and neurodegeneration. DNA methylation and histone
69 modifications itself are extensively regulated by different factors (e.g., translocation (TET) oxygenase family, DNA metyltransferases, methyl-CpG- binding proteins, histone acetylases, histone deacetylases), which, themselves, are (post)-transcriptionally regulated. Environmental exposures can lead to changes in activity of those factors and perturb cellular DNA methylation and histone modification (Smirnova et al., 2012, Szyf 2011). Epigenetic modifications coming especially into the light, when we are talking about low-dose long-term exposures. The study by
Fraga et al. (2005) on monozygotic twins revealed moderate or no differences in epigenetic profilies in 3-years old twins, while those profiles drifted apart with increasing age of probands, suggesting environmental and life stile contribution on epigenome. Environmental stressors may cause a permanent change in epigenome (so-called epigenetic memory, scar, or foot-print).
Epigenetic memory in the form of changes in the DNA methylation pattern, could protect against, or contribute to long-term pathogenesis or cellular vulnerability to subsequent hazards (Tyagi et al., 2015). Thus the epigenome serves as the adaption to stress, plasticity or resilience tool. Since it is evident that epigenetic alterations maintain a memory of the obtained signal to make the system robust and tolerant against the environment, it is possible that the epigenome may make the system “antifragile”. There are few examples of developing stress tolerance in plant biology and ecotoxicology: e.g., epigenetic silencing of flowering locus C under prolonged exposure to cold temperature that results in coordination of flowering of Arabidopsis (He et al., 2003; Kin et al., 2005). Earthworms developed a tolerance against low-dose arsenic by epeginetic adaption mechanisms (Vandegehuchte et al., 2014). It is suggested that the increased stress tolerance can be even transmitted in form of altered DNA methylation patterns to the next generation, which was not even exposed to the stress factor (reviewed in Vandegehuchte et al., 2014). For this reason, it is important to study epigenetic mechanisms in toxicology to further understand the mode of action regarding low-dose exposures (Mirbahai, and Chipman, 2014).
70
This is how the experience imprints and changes future responses. Epigenetics might serve antifragility – the system is not restored but better. Taleb again: “Antifragility has a singular property of allowing us to deal with the unknown, to do things without understanding them - and do them well.” Isn’t this exactly what a cell exposed to unknown toxicants should do? Perhaps we should not stretch the analogy of society and cells too far, but the parallels are stimulating food for thought. Friedrich Nietzsche “That which does not destroy, strengthens.” is not always correct. Sometimes the results of stressors are ‘bad memories’, such as epigenetic scars (Balmer et al., 2014 a,b), mutations or other functional impairments may predispose to the disease or lead to adverse outcome in life-time or even transgenerationally. The fine line between resilience and maladaptation may need to be defined situation by situation.
2.3. CONCLUSIONS
Nicolas Taleb was quoted already several times in this article. With his books “The Black Swan” and “Antifragility” he has popularized some ideas, which are also central to some phenomena in toxicology. We made earlier in this series of articles the notion (Bottini and Hartung, 2009) – not surprisingly referring to him in the article on economical aspects of our field, that rare events
(“Black Swans”) are typical in safety sciences. We do force our testing strategies (high-dose, oversensitive models), however, into the “Gaussian” part of probabilities, which we can handle.
His follow-up book on anti-fragility resonated well with some of the thoughts here: “Fragility is quite measurable, risk not so at all, particularly risk associated with rare events.” This is actually pretty good guidance and description what toxicology is about: We assess the fragility of our systems with high-dose experiments to be prepared for the rare event of a low-dose risk. The interesting new thought, however, is the aspect of anti-fragility. Evolution has to favor anti-fragile constructions. This elasticity affords our protection against the majority of (small) hits. We need to understand this to appreciate the limits of what we can stand and how we can reinforce this
71 defense. We need to understand, where this system fails, possibly leaving scars and maladaptations leading to hazard manifestations. It appears that we have tools in reach to address this, especially long-term cultures and high-content characterizations of responses), possibly changing the point of view on how organ-selectivity of toxic actions and chronic manifestations of toxicities come about.
2.4. ACKNOWLEDGEMENTS
The authors would like to thank Dr Imran Sha, US EPA, and Dr Igor Linkv and his team, US
Army, for valuable discussions and critically reading the manuscript.
2.5. REFERENCES
Alastalo H, Räikkönen K, Pesonen AK, Osmond C, Barker DJ, Heinonen K, Kajantie E, Eriksson
JG. Early life stress and blood pressure levels in late adulthood. J Hum Hypertens. 2013
Feb;27(2):90-4. doi: 10.1038/jhh.2012.6.
Andersen, M. E., McMullen, P. D., & Krewski, D. (2015). Developing tools for defining and
establishing pathways of toxicity. Archives of Toxicology, 89(5), 809–812.
http://doi.org/10.1007/s00204-015-1512-y
Araujo, R. P., & Liotta, L. A. (2006). A control theoretic paradigm for cell signaling networks: a
simple complexity for a sensitive robustness. Current Opinion in Chemical Biology, 10(1),
81–87. http://doi.org/10.1016/j.cbpa.2006.01.002
Araujo, R. P., Liotta, L. A., and Petricoin, E. F. (2007). Proteins, drug targets and the mechanisms
they control: the simple truth about complex networks. Nature Reviews. Drug Discovery,
6(11), 871–880. http://doi.org/10.1038/nrd2381
72
Barbaric, I., Miller, G., & Dear, T. N. (2007). Appearances can be deceiving: phenotypes of
knockout mice. Briefings in Functional Genomics & Proteomics, 6(2), 91–103.
http://doi.org/10.1093/bfgp/elm008
Basketter DA, Clewell H, Kimber I, Rossi A, Blaauboer B, Burrier R, Daneshian M, Eskes C,
Goldberg A, Hasiwa N, Hoffmann S, Jaworska J, Knudsen TB, Landsiedel R, Leist M,
Locke P, Maxwell G, McKim J, McVey EA, Ouédraogo G, Patlewicz G, Pelkonen O,
Roggen E, Rovida C, Ruhdel I, Schwarz M, Schepky A, Schoeters G, Skinner N, Trentz K,
Turner M, Vanparys P, Yager J, Zurlo J and Hartung T. A roadmap for the development of
alternative (non-animal) methods for systemic toxicity testing. ALTEX 2012, 29:3-89.
Blaauboer, B. J., Boekelheide, K., Clewell, H. J. et al. (2012). The use of biomarkers of toxicity
for integrating in vitro hazard estimates into risk assessment for humans. Altex, 29, 411–
425. http://doi.org/10.14573/altex.2012.4.411
Bottini AA and Hartung T. Food for thought… on economics of animal testing. ALTEX 2009,
26:3-16.
Bouhifd M, Beger R, Flynn T, Guo L, Harris G, Hogberg HT, Kaddurah-Daouk R, Kamp H,
Kleensang A, Maertens A, Odwin-DaCosta S, Pamies D, Robertson D, Smirnova L, Sun J,
Zhao L and Hartung T. Quality Assurance of Metabolomics. ALTEX, 2015, in press
Bouhifd M, Andersen ME, Baghdikian C, Boekelheide K, Crofton KM, Fornace AJ Jr.,
Kleensang A, Li H, Livi CB, Maertens A, McMullen PD, Rosenberg M, Thomas R,
Vantangoli M, Yager JD, Zhao L and Hartung T. The Human Toxome project. ALTEX
2015, 32:112-124.
Bouhifd M, Hartung T, Hogberg HT, Kleensang A and Zhao L. Review: Toxicometabolomics. J.
Appl. Toxicol. 2013, 33:1365-1383. DOI 10.1002/jat.2874.
Brink CB; Pretorius A; van Niekerk BP; Oliver DW; Venter DP. Studies on cellular resilience
and adaptation following acute and repetitive exposure to ozone in cultured human
epithelial (HeLa) cells. Redox Rep 2008;13(2):87-100
73
Buchman, T. G. (2002). The community of the self. Nature, 420(6912), 246–251.
http://doi.org/10.1038/nature01260
Burke, S. L., Hammell, M., & Ambros, V. (2015). Robust Distal Tip Cell Pathfinding in the Face
of Temperature Stress Is Ensured by Two Conserved microRNAS in Caenorhabditis
elegans. Genetics, 200(4), 1201–1218. http://doi.org/10.1534/genetics.115.179184
Calabrese, E. J., and Baldwin, L. A. (2001). The frequency of U-shaped dose responses in the
toxicological literature. Toxicological Sciences 62(2), 330–338.
Calabrese, E., and Blain, R. (2005). The occurrence of hormetic dose responses in the
toxicological literature, the hormesis database: an overview. Toxicology and Applied
Pharmacology, 202(3), 289–301. http://doi.org/10.1016/j.taap.2004.06.023
Chan, Y. C., Banerjee, J., Choi, S. Y., & Sen, C. K. (2012). miR-210: the master hypoxamir.
Microcirculation (New York, N.Y. : 1994), 19(3), 215–223. http://doi.org/10.1111/j.1549-
8719.2011.00154.x
Choi, D. C., Chae, Y.-J., Kabaria, S., Chaudhuri, A. D., Jain, M. R., Li, H., et al. (2014).
MicroRNA-7 protects against 1-methyl-4-phenylpyridinium-induced cell death by targeting
RelA. Journal of Neuroscience, 34(38), 12725–12737.
http://doi.org/10.1523/JNEUROSCI.0985-14.2014
Clapp C, Portt L, Khoury C, Sheibani S, Eid R, Greenwood M, Vali H, Mandato CA, Greenwood
MT. Untangling the Roles of Anti-Apoptosis in Regulating Programmed Cell Death using
Humanized Yeast Cells. Front Oncol. 2012 Jun 13;2:59. doi:10.3389/fonc.2012.00059.
Clemedson C. and Ekwall B. (1999) Overview of the final MEIC results: I. The In vitro±In vitro
Evaluation. Toxicology in vitro 13, 657-663.
Clemedson, C. (2008). The European ACuteTox Project: A Modern Integrative In vitro Approach
to Better Prediction of Acute Toxicity. Clinical Pharmacology Therapeutics 84, 200–202.
http://doi.org/10.1038/clpt.2008.135
74
Coecke S, Ahr H, Blaauboer BJ, Bremer S, Casati S, Castell J, Combes R, Corvi R, Crespi CL,
Cunningham ML, Elaut G, Eletti B, Freidig A, Gennari A, Ghersi-Egea J-F, Guillouzo A,
Hartung T, Hoet P, Ingelman-Sundberg M, Munn S, Janssens W, Ladstetter B, Leahy D,
Long A, Meneguz A, Monshouwer M, Morath S, Nagelkerke F, Pelkonen O, Ponti J, Prieto
P, Richert L,Sabbioni E, Schaack B, Steiling W, Testai E, Vericat J-A and Worth A.
Metabolism: A bottleneck in in vitro toxicological test development. ATLA – Altern. Lab.
Anim. 2006, 34:49-84
Dehay B, Bove J, Rodriguez-Muela N, Perier C, Recasens A, Boya P, Vila M (2010) Pathogenic
lysosomal depletion in Parkin- son’s disease. J Neurosci 30:12535–12544. doi:10.1523/
JNEUROSCI.1920-10.2010
Di Mauro C, Bouchon S, Logtmeijer C, Nordvik JP, Pride R and Hartung T. Structured approach
to identifying European critical infrastructures. Int. J. Critical Infrastructures, 2010, 6:277-
292.
Dunn JF, Wu Y, Zhao Z, Srinivasan S, Natah SS. Training the brain to survive stroke. PLoS One.
2012;7(9):e45108. doi:10.1371/journal.pone.0045108.
Ebert, M. S., & Sharp, P. A. (2012). Roles for MicroRNAs in Conferring Robustness to
Biological Processes. Cell, 149(3), 515–524. http://doi.org/10.1016/j.cell.2012.04.005
Ekwall, B. (1999). Overview of the Final MEIC Results: II. The In vitro--In vivo Evaluation,
Including the Selection of a Practical Battery of Cell Tests for Prediction of Acute Lethal
Blood Concentrations in Humans. Toxicology in vitro 13(4-5), 665–673.
Finch CE; Morgan TE; Longo VD; de Magalhaes JP. Cell resilience in species life spans: a link
to inflammation? Aging Cell 2010 Aug;9(4):519-26
Fraga, M. F., Ballestar, E., Paz, M. F., Ropero, S., Setien, F., Ballestar, M. L., et al. (2005).
Epigenetic differences arise during the lifetime of monozygotic twins.
75
Fragkouli, A., & Doxakis, E. (2014). miR-7 and miR-153 protect neurons against MPP(+)-
induced cell death via upregulation of mTOR pathway. Frontiers in Cellular Neuroscience,
8, 182–182. http://doi.org/10.3389/fncel.2014.00182
Hales CN, Barker DJ (July 1992). "Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty
phenotype hypothesis". Diabetologia 35 (7): 595–601. doi:10.1007/BF00400248
Halle, W. and Goeres, E. (1988). Register der Zytotoxizität (IC50) in der Zellkultur und
Möglichkeiten zur Abschätzung der akuten Toxizität (LD50). In Beiträge zur
Wirkstofforschung (ed. P. Oehme, H. Loewe and E. Goeres), 108 pp. Berlin, Germany:
Institut für Wirkstofforschung.
Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011 Mar
4;144(5):646-74.
Hansson O, Nylandsted J, Castilho RF, Leist M, Jäättelä M, Brundin P. Overexpression of heat
shock protein 70 in R6/2 Huntington's disease mice has only modest effects on disease
progression. Brain Res. 2003 Apr 25;970(1-2):47-57.
Hartung T and McBride M. Food for thought… on mapping the human toxome. ALTEX 2011,
28, 83-93. doi: 10.14573/altex.2011.2.083
Hartung T, Stephens M and Hoffmann S. Mechanistic validation. ALTEX 2013, 30:119-130.
Hartung T, van Vliet E, Jaworska J, Bonilla L, Skinner N and Thomas R. Systems toxicology.
ALTEX 2012, 29: 119-128.
Hartung T. Food for thought … on cell culture. ALTEX 2007, 24:143-147.
Hartung T and Wendel A. Endotoxin-inducible cytotoxicity in liver cell cultures - II:
Demonstration of endotoxin-tolerance. Biochem. Pharmacol. 1992, 43:191-196.
He, Y., Michaels, S. D., & Amasino, R. M. (2003). Regulation of flowering time by histone
acetylation in Arabidopsis. Science, 302(5651), 1751–1754.
http://doi.org/10.1126/science.1091109
76
Heisch C. The Arsenic Eaters of Styria. Boston Med Surg J 1860; 62:484-488July 12, 1860DOI:
10.1056/NEJM186007120622404
Hendriks G, Atallah M, Morolli B, Calléja F, Ras-Verloop N, Huijskens I, Raamsman M, van de
Water B, Vrieling H. The ToxTracker assay: novel GFP reporter systems that provide
mechanistic insight into the genotoxic properties of chemicals. Toxicol Sci. 2012
Jan;125(1):285-98.
Herranz, H., & Cohen, S. M. (2010). MicroRNAs and gene regulatory networks: managing the
impact of noise in biological systems. Genes & Development, 24(13), 1339–1344.
http://doi.org/10.1101/gad.1937010
Hoang T, Choi D-K, Nagai M, Wu D-C, Nagata T, Prou D, Wilson GL, Vila M, Jackson-Lewis
V, Dawson VL, Dawson TM, Chessel M-F, Przedborski S (2009) Neuronal NOS and
cyclooxygenase-2 contribute to DNA damage in a mouse model of Parkinson disease. Free
Radical Biol Med 47:1049–1056. doi:10.1016/j. freeradbiomed.2009.07.013
Hoeng J, Talikka M, Martin F, et al. Case study: the role of mechanistic network models in
systems toxicology. Drug Discov Today (England), Feb 2014, 19(2) p183-92
Hood, L., & Perlmutter, R. M. (2004). The impact of systems approaches on biological problems
in drug discovery. Nature Biotechnology, 22(10), 1215–1217.
http://doi.org/10.1038/nbt1004-1215
Ivatt R, Whitworth AJ (2014) SREBF1 links lipogenesis to mitophagy and sporadic Parkinson’s
disease. Autophagy 10:33–34. doi:10.4161/auto.29642
Jack, J., Wambaugh, J., & Shah, I. (2013). Systems toxicology from genes to organs. Methods in
Molecular Biology, 930, 375–397. http://doi.org/10.1007/978-1-62703-059-5_17
Jennings P. Stress response pathways, toxicity pathways and adverse outcome pathways. Arch
Toxicol. 2013 Jan;87(1):13-4. doi: 10.1007/s00204-012-0974-4.
Kim D, Frank CL, Dobbin MM, Tsunemoto RK, Wu D, Peng PL, Guan J-S, Lee B-H, Moy LY,
Giusti P, Broodie N, Mazitschek R, Delalle I, Haggarty SJ, Neve RL, Lu YM, Tsai L-H
77
(2008) Deregulation of HDAC1 by p25/Cdk5 in neurotoxicity. Neuron 60:803–817.
doi:10.1016/j.neuron.2008.10.015
Kim, S. Y., He, Y., Jacob, Y., Noh, Y.-S., Michaels, S., & Amasino, R. (2005). Establishment of
the vernalization-responsive, winter-annual habit in Arabidopsis requires a putative histone
H3 methyl transferase. The Plant Cell, 17(12), 3301–3310.
http://doi.org/10.1105/tpc.105.034645
Kleensang A, Maertens A, Rosenberg M, Fitzpatrick S, et al. Pathways of Toxicity. ALTEX
2014, 31:53-61. doi: 10.14573/altex.1309261
Kleensang A, Maertens A, Rosenberg M, Fitzpatrick S, Lamb J, Auerbach S, Brennan R, Crofton
KM, Gordon B, Fornace AJ Jr., Gaido K, Gerhold D, Haw R, Henney A, Ma'ayan A,
McBride M, Monti S, Ochs MF, Pandey A, Sharan R, Stierum R, Tugendreich S, Willett C,
Wittwehr C, Xia J, Patton GW, Arvidson K, Bouhifd M, Hogberg HT, Luechtefeld T,
Smirnova L, Zhao L, Adeleye Y, Kanehisa M, Carmichael P, Andersen E. M, Hartung T.
Pathways of Toxicity. ALTEX 2014, 31:53-61. doi: 10.14573/altex.1309261
Koch, I., and Ackermann, J. (2012). On functional module detection in metabolic networks.
Metabolites, 3(3), 673–700. http://doi.org/10.3390/metabo3030673
Kreeger, P. K., & Lauffenburger, D. A. (2010). Cancer systems biology: a network modeling
perspective. Carcinogenesis, 31(1), 2–8. http://doi.org/10.1093/carcin/bgp261
Krug AK, Gutbier S, Zhao L, Pöltl D, Kullmann C, Ivanova V, Förster S, Jagtap S, Meiser J,
Leparc G, Schildknecht S, Adam M, Hiller K, Farhan H, Brunner T, Hartung T, Sachinidis
A, and Leist M. Transcriptional and metabolic adaptation of human neurons to the
mitochondrial toxicant MPP+. Cell Death Disease, 2014, 5, e1222;
doi:10.1038/cddis.2014.166.
Lau C, Rogers JM. Embryonic and fetal programming of physiological disorders in adulthood.
Birth Defects Res C Embryo Today. 2004 Dec;72(4):300-1
78
Lehner MD and Hartung T. Endotoxin tolerance – mechanisms and beneficial effects in bacterial
infection. Rev. Physiol. Biochem. Pharmacol. 2002, 144:95-141.
Leist M, Hasiwa N, Rovida C, Daneshian M, Basketter D, Kimber I, Clewell H, Gocht T,
Goldberg A, Busquet F, Rossi A-M, Schwarz M, Stephens M, Taalman R, Knudsen TB,
McKim J, Harris G, Pamies D and Hartung T. Consensus report on the future of animal-
free systemic toxicity testing. ALTEX 2014, 31:341–356.
Leist M, Jäättelä M. Burning up TNF toxicity for cancer therapy. Nat Med. 2002 Jul;8(7):667-8.
Leist M, Jäättelä M. Four deaths and a funeral: from caspases to alternative mechanisms. Nat Rev
Mol Cell Biol. 2001 Aug;2(8):589-98. Review.
Limonciel A, Moenks K, Stanzel S, Truisi GL, Parmentier C, Aschauer L, Wilmes A, Richert L,
Hewitt P, Mueller SO, Lukas A, Kopp-Schneider A, Leonard MO, Jennings P.
Transcriptomics hit the target: Monitoring of ligand-activated and stress response pathways
for chemical testing. Toxicol In vitro. 2015 Jan 13. pii: S0887-2333(14)00251-3. doi:
10.1016/j.tiv.2014.12.011.
Lin, Z., & Will, Y. (2011). Evaluation of Drugs with Specific Organ Toxicities in Organ Specific
Cell Lines. Toxicological Sciences 126, 114–127. http://doi.org/10.1093/toxsci/kfr339
Lindblom R, Ververis K, Tortorella SM, Karagiannis TC. The early life origin theory in the
development of cardiovascular disease and type 2 diabetes. Mol Biol Rep. 2015
Apr;42(4):791-7. doi: 10.1007/s11033-014-3766-5.
Loscalzo, J. (2010). The cellular response to hypoxia: tuning the system with microRNAs. The
Journal of Clinical Investigation, 120(11), 3815–3817. http://doi.org/10.1172/JCI45105
Macia J; Sole RV. Distributed robustness in cellular networks: insights from synthetic evolved
circuits. J R Soc Interface 2009 Apr 6;6(33):393-400
Maertens A, Luechtefeld T, Kleensang A and Hartung T. MPTP’s pathway of toxicity indicates
central role of transcription factor SP1. Arch. Toxicol. 2015, 89:743-755. doi:
10.1007/s00204-015-1509-6.
79
Manke T; Demetrius L; Vingron M. An entropic characterization of protein interaction networks
and cellular robustness. J R Soc Interface 2006 Dec 22;3(11):843-50
Mattson M P (2008). Hormesis Defined. Ageing Res Rev. 2008 Jan; 7(1): 1–7. doi:
10.1016/j.arr.2007.08.007
McGowan PO, Sasaki A, D'Alessio AC, Dymov S, Labonté B, Szyf M, Turecki G, Meaney MJ.
Epigenetic regulation of the glucocorticoid receptor in human brain associates with
childhood abuse. Nat Neurosci. 2009 Mar;12(3):342-8. doi:10.1038/nn.2270.
Melton, D. W. 1994 Gene targeting in the mouse. Bioessays 16, 633–638.
(doi:10.1002/bies.950160907)
Mirbahai, L., & Chipman, J. K. (2014). Epigenetic memory of environmental organisms: a
reflection of lifetime stressor exposures. Mutation Research - Genetic Toxicology and
Environmental Mutagenesis, 764-765, 10–17.
http://doi.org/10.1016/j.mrgentox.2013.10.003
Miska, E. A., Alvarez-Saavedra, E., Abbott, A. L., Lau, N. C., Hellman, A. B., McGonagle, S.
M., et al. (2007). Most Caenorhabditis elegans microRNAs Are Individually Not Essential
for Development or Viability. PLoS Genetics, 3(12), e215.
http://doi.org/10.1371/journal.pgen.0030215
National Institutes of Health (NIH), 2006. Background Review Document (BRD): Validation of
Neutral Red Uptake Test Methods NIH /In vitro Cytotoxicity Test Methods for Estimating
Acute Oral Systemic Toxicity. Publication No. 07-4518, November 2006. Available from:
http://iccvam.niehs.nih.gov/methods/acutetox/inv_nru_brd.htm
O'Neill S, Ross JA, Wigmore SJ, Harrison EM. The role of heat shock protein 90 in modulating
ischemia-reperfusion injury in the kidney. Expert Opin Investig Drugs. 2012
Oct;21(10):1535-48.
Perier C, Vila M (2012) Mitochondrial biology and Parkinson’s disease. Cold Spring Harb
Perspect Med 2:a009332. doi:10.1101/cshperspect.a009332
80
Prieto, P., Cole, T., Curren, R., Gibson, R. M., Liebsch, M., Raabe, H., et al. (2013). Assessment
of the predictive capacity of the 3T3 Neutral Red Uptake cytotoxicity test method to
identify substances not classified for acute oral toxicity (LD50>2000mg/kg): Results of an
ECVAM validation study. Regulatory Toxicology and Pharmacology, 65(3), 344–365.
http://doi.org/10.1016/j.yrtph.2012.11.013
Qiu Z, Norflus F, Singh B, Swindell MK, Buzescu R, Bejarano M, Chopra R, Zucker B, Benn
CL, DiRocco DP, Cha JH, Ferrante RJ, Hersch SM (2006) Sp1 is up-regulated in cellular
and transgenic models of Huntington disease, and its reduction is neuroprotective. J Biol
Chem 281:16672–16680. doi:10.1074/jbc. M511648200
Rahnenführer, J., & Leist, M. (2015). From smoking guns to footprints: mining for critical events
of toxicity pathways in transcriptome data. Archives of Toxicology, 89(5), 813–817.
http://doi.org/10.1007/s00204-015-1497-6
Ramirez T, Daneshian M, Kamp H, Bois FY, Clench MR, Coen M, Donley B, Fischer SM,
Ekman DR, Fabian E, Guillou C, Heuer J, Hogberg HT, Jungnickel H, Keun HC,
Krennrich G, Krupp E, Luch A, Noor F, Peter E, Riefke B, Seymour M, Skinner N,
Smirnova L, Verheij E, Wagner S, Hartung T, van Ravenzwaay B and Leist M.
Metabolomics in Toxicology and Preclinical Research. ALTEX 2013, 30:209-225.
Reka, A. and Barabasi, A-L. (2002) ‘Statistical mechanics of complex networks’, Reviews of
Modern Physics, Vol. 74, No. 47, http://arXiv:cond-mat/0106096v1.
Ren, Z., & Ambros, V. R. (2015). Caenorhabditis elegans microRNAs of the let-7 family act in
innate immune response circuits and confer robust developmental timing against pathogen
stress. Proceedings of the National Academy of Sciences, 112(18), E2366–75.
http://doi.org/10.1073/pnas.1422858112
Santpere G, Nieto M, Puig B, Ferrer I (2006) Abnormal Sp1 transcription factor expression in
Alzheimer disease and tauopathies. Neurosci Lett 397:30–34.
doi:10.1016/j.neulet.2005.11.062
81
Sauer JM, Hartung T, Leist M, et al. Systems Toxicology: The Future of Risk Assessment. Int J
Toxicol, Jul 2015, 34(4) p346-8
Scheffer, M., Carpenter, S. R., Lenton, T. M., Bascompte, J., Brock, W., Dakos, V., et al. (2012).
Anticipating critical transitions. Science, 338(6105), 344–348.
http://doi.org/10.1126/science.1225244
Schrage, A., Hempel, K., Schulz, M., Kolle, S. N., van Ravenzwaay, B., & Landsiedel, R. (2011).
Refinement and reduction of acute oral toxicity testing: a critical review of the use of
cytotoxicity data. Alternatives to Laboratory Animals : ATLA, 39(3), 273–295.
Sebert S, Sharkey D, Budge H, Symonds ME. The early programming of metabolic health: is
epigenetic setting the missing link? Am J Clin Nutr. 2011 Dec;94(6 Suppl):1953S-1958S.
doi: 10.3945/ajcn.110.001040.
Seiler, A. E., & Spielmann, H. (2011). The validated embryonic stem cell test to predict
embryotoxicity in vitro. Nat.Protoc., 6(7), 961–978. http://doi.org/nprot.2011.348
Sipes, N. S., Martin, M. T., Kothiya, P., Reif, D. M., Judson, R. S., Richard, A. M., et al. (2013).
Profiling 976 ToxCast chemicals across 331 enzymatic and receptor signaling assays.
Chemical Research in Toxicology, 26(6), 878–895. http://doi.org/10.1021/tx400021f
Smirnova L., Harris G., Delp J., Valadares M, Pamies D., Hogberg H., Leist M., and Hartung T..
A LUHMES 3D dopaminergic neuronal model for neurotoxicity testing allowing long-term
exposure and cellular resilience analysis. Submitted.
Smirnova, L., Sittka, A., & Luch, A. (2012). On the role of low-dose effects and epigenetics in
toxicology. Exs, 101, 499–550. http://doi.org/10.1007/978-3-7643-8340-4_18
Steinhardt, R. A. (2005). The mechanisms of cell membrane repair: A tutorial guide to key
experiments. Annals of the New York Academy of Sciences, 1066, 152–165.
http://doi.org/10.1196/annals.1363.017
Sterky FH, Hoffman AF, Milenkovic D, Bao B, Paganelli A, Edgar D, Wibom R, Lupica CR,
Olson L, Larsson NG (2012) Altered dopamine metabolism and increased vulnerability to
82
MPTP in mice with partial deficiency of mitochondrial complex I in dopamine neurons.
Hum Mol Gen 21:1078–1089. doi:10.1093/hmg/ ddr537
Sturla, S. J., Boobis, A. R., FitzGerald, R. E., Hoeng, J., Kavlock, R. J., Schirmer, K., et al.
(2014). Systems toxicology: from basic research to risk assessment. Chemical Research in
Toxicology, 27(3), 314–329. http://doi.org/10.1021/tx400410s
Suderman M, McGowan PO, Sasaki A, Huang TC, Hallett MT, Meaney MJ, Turecki G, Szyf M.
Conserved epigenetic sensitivity to early life experience in the rat and human hippocampus.
Proc Natl Acad Sci U S A. 2012 Oct 16;109 Suppl 2:17266-72. doi:
10.1073/pnas.1121260109.
Szyf, M. (2007). The dynamic epigenome and its implications in toxicology. Toxicological
Sciences : an Official Journal of the Society of Toxicology, 100(1), 7–23.
http://doi.org/10.1093/toxsci/kfm177
Szyf, M. (2011). DNA methylation, the early-life social environment and behavioral disorders.
Journal of Neurodevelopmental Disorders, 3(3), 238–249. http://doi.org/10.1007/s11689-
011-9079-2
Tagore S; De RK. Detecting breakdown points in metabolic networks. Comput Biol Chem 2011
Dec 14;35(6):371-80
Tang, H. L., Tang, H. M., Hardwick, J. M., & Fung, M. C. (2015). Strategies for tracking
anastasis, a cell survival phenomenon that reverses apoptosis. Journal of Visualized
Experiments, (96), –. http://doi.org/10.3791/51964
Tang, H. L., Tang, H. M., Mak, K. H., Hu, S., Wang, S. S., Wong, K. M., et al. (2012). Cell
survival, DNA damage, and oncogenic transformation after a transient and reversible
apoptotic response. Molecular Biology of the Cell, 23(12), 2240–2252.
http://doi.org/10.1091/mbc.E11-11-0926
Thomas, R. S., Black, M. B., Li, L., Healy, E., Chu, T.-M., Bao, W., et al. (2012). A
comprehensive statistical analysis of predicting in vivo hazard using high-throughput in
83
vitro screening. Toxicological Sciences, 128(2), 398–417.
http://doi.org/10.1093/toxsci/kfs159
Thomas, R. S., Philbert, M. A., Auerbach, S. S., Wetmore, B. A., Devito, M. J., Cote, I., et al.
(2013). Incorporating New Technologies into Toxicity Testing and Risk Assessment:
Moving from 21st Century Vision to a Data-Driven Framework. Toxicological Sciences :
an Official Journal of the Society of Toxicology, 136(1), kft178–18.
http://doi.org/10.1093/toxsci/kft178
Tyagi, E., Zhuang, Y., Agrawal, R., Ying, Z., & Gomez-Pinilla, F. (2015). Interactive actions of
Bdnf methylation and cell metabolism for building neural resilience under the influence of
diet. Neurobiology of Disease, 73, 307–318. http://doi.org/10.1016/j.nbd.2014.09.014
Tyson, J. J., Chen, K. C., & Novak, B. (2003). Sniffers, buzzers, toggles and blinkers: dynamics
of regulatory and signaling pathways in the cell. Current Opinion in Cell Biology.
http://doi.org/10.1016/S0955-0674(03)00017-6
Vandegehuchte, M. B., & Janssen, C. R. (2014). Epigenetics in an ecotoxicological context.
Mutation Research - Genetic Toxicology and Environmental Mutagenesis, 764-765, 36–45.
http://doi.org/10.1016/j.mrgentox.2013.08.008
Velichko, A. K., Markova, E. N., Petrova, N. V., Razin, S. V., & Kantidze, O. L. (2013).
Mechanisms of heat shock response in mammals. Cellular and Molecular Life Sciences,
70(22), 4229–4241. http://doi.org/10.1007/s00018-013-1348-7
Visan, A., Hayess, K., Sittner, D., Pohl, E. E., Riebeling, C., Slawik, B., et al. (2012). Neural
differentiation of mouse embryonic stem cells as a tool to assess developmental
neurotoxicity in vitro. Neurotoxicology, 33(5), 1135–1146.
http://doi.org/10.1016/j.neuro.2012.06.006
Wallin IE. (1923). The mitochondria problem. Am Naturalist 57, 255-261.
Wang J, Bannon MJ (2005) Sp1 and Sp3 activate transcription of the human dopamine
transporter gene. J Neurochem 93:474–482. doi:10.1111/j.1471-4159.2005.03051.x
84
Wang Y, Reis C, Applegate R 2nd, Stier G, Martin R, Zhang JH. Ischemic conditioning-induced
endogenous brain protection: Applications pre-, per- or post-stroke. Exp Neurol. 2015 Apr
18. pii: S0014-4886(15)00123-5. doi:10.1016/j.expneurol.2015.04.009.
Wetmore, B. A., Wambaugh, J. F., Ferguson, S. S., Li, L., Clewell, H. J., Judson, R. S., et al.
(2013). Relative impact of incorporating pharmacokinetics on predicting in vivo hazard and
mode of action from high-throughput in vitro toxicity assays. Toxicological Sciences,
132(2), 327–346. http://doi.org/10.1093/toxsci/kft012
Wink S, Hiemstra S, Huppelschoten S, Danen E, Niemeijer M, Hendriks G, Vrieling H, Herpers
B, van de Water B. Quantitative high content imaging of cellular adaptive stress response
pathways in toxicity for chemical safety assessment. Chem Res Toxicol. 2014 Mar
17;27(3):338-55. doi: 10.1021/tx4004038.
Wu KH, Mo XM, Han ZC, Zhou B. Cardiac cell therapy: pre-conditioning effects in cell-delivery
strategies. Cytotherapy. 2012 Mar;14(3):260-6. doi:10.3109/14653249.2011.643780.
Ye Q, Zhang X, Huang B, Zhu Y, Chen X (2013) Astaxanthin sup- presses MPP-induced
oxidative damage in PC12 cells through a Sp1/NR1 signaling pathway. Mar Drugs
11:1019–1034. doi:10.3390/md11041019
Yehuda R, Flory JD, Bierer LM, Henn-Haase C, Lehrner A, Desarnaud F, Makotkine I,
Daskalakis NP, Marmar CR, Meaney MJ. Lower methylation of glucocorticoid receptor
gene promoter 1F in peripheral blood of veterans with posttraumatic stress disorder. Biol
Psychiatry. 2015 Feb 15;77(4):356-64. doi:10.1016/j.biopsych.2014.02.006.
Yellon DM, Hausenloy DJ. Realizing the clinical potential of ischemic preconditioning and
postconditioning. Nat Clin Pract Cardiovasc Med. 2005 Nov;2(11):568-75.
Zhong WX, Wang YB, Peng L, Ge XZ, Zhang J, Liu SS, Zhang XN, Xu ZH, Chen Z, Luo JH
(2012) Lanthionine synthetase C-like protein 1 interacts with and inhibits cystathionine
beta-syn- thase: a target for neuronal antioxidant defense. J Biol Chem 287:34189–34201.
doi:10.1074/jbc.M112.383646
85
Potts PR, Singh S, Knezek M, Thompson CB, Deshmukh M. Critical function of endogenous
XIAP in regulating caspase activation during sympathetic neuronal apoptosis. J Cell Biol.
2003 Nov 24;163(4):789-99. Epub 2003 Nov 17. PubMed PMID: 14623868; PubMed
Central PMCID: PMC2173693.
Deshmukh M, Johnson EM Jr. Evidence of a novel event during neuronal death: development of
competence-to-die in response to cytoplasmic cytochrome c. Neuron. 1998 Oct;21(4):695-
705. PubMed PMID: 9808457.
Foghsgaard L, Wissing D, Mauch D, Lademann U, Bastholm L, Boes M, Elling F, Leist M,
Jäättelä M. Cathepsin B acts as a dominant execution protease in tumor cell apoptosis
induced by tumor necrosis factor. J Cell Biol. 2001 May 28;153(5):999-1010. PubMed
PMID: 11381085;
Jaiswal JK, Lauritzen SP, Scheffer L, Sakaguchi M, Bunkenborg J, Simon SM, Kallunki T,
Jäättelä M, Nylandsted J. S100A11 is required for efficient plasma membrane repair and
survival of invasive cancer cells. Nat Commun. 2014 May 8;5:3795. doi:
10.1038/ncomms4795. PubMed
Roostalu U, Strähle U. In vivo imaging of molecular interactions at damaged sarcolemma. Dev
Cell. 2012 Mar 13;22(3):515-29. doi: 10.1016/j.devcel.2011.12.008. PubMed PMID:
22421042.
Hirt UA, Gantner F, Leist M. Phagocytosis of nonapoptotic cells dying by caspase-independent
mechanisms. J Immunol. 2000 Jun 15;164(12):6520-9. PubMed
Hirt UA, Leist M. Rapid, noninflammatory and PS-dependent phagocytic clearance of necrotic
cells. Cell Death Differ. 2003 Oct;10(10):1156-64. PubMed PMID:
Jurk D, Wang C, Miwa S, Maddick M, Korolchuk V, Tsolou A, Gonos ES, Thrasivoulou C,
Saffrey MJ, Cameron K, von Zglinicki T. Postmitotic neurons develop a p21-dependent
senescence-like phenotype driven by a DNA damage response. Aging Cell. 2012
86
Dec;11(6):996-1004. doi: 10.1111/j.1474-9726.2012.00870.x. Epub 2012 Sep 12. PubMed
PMID: 22882466; PubMed
Ono T, Ikehata H, Vishnu Priya P, Uehara Y. Molecular nature of mutations induced by
irradiation with repeated low doses of X-rays in spleen, liver, brain and testis of lacZ-
transgenic mice. Int J Radiat Biol. 2003 Aug;79(8):635-41.
Vijg J, Dollé ME, Martus HJ, Boerrigter ME. Transgenic mouse models for studying mutations in
vivo: applications in aging research. Mech Ageing Dev. 1997 Dec 30;99(3):257-71.
Review.
Kinsner-Ovaskainen A, Prieto P, Stanzel S, Kopp-Schneider A. Selection of test methods to be
included in a testing strategy to predict acute oral toxicity: an approach based on statistical
analysis of data collected in phase 1 of the ACuteTox project. Toxicol In vitro. 2013
Jun;27(4):1377-94. doi: 10.1016/j.tiv.2012.11.010. Epub 2012 Nov 21. PubMed PMID:
23178337.
Stiegler NV, Krug AK, Matt F, Leist M. Assessment of chemical-induced impairment of human
neurite outgrowth by multiparametric live cell imaging in high-density cultures. Toxicol
Sci. 2011 May;121(1):73-87. doi: 10.1093/toxsci/kfr034. Epub 2011 Feb 21. PubMed
PMID: 21342877.
Efremova L, Schildknecht S, Adam M, Pape R, Gutbier S, Hanf B, Bürkle A, Leist M. Prevention
of the degeneration of human dopaminergic neurons in an astrocyte co-culture system
allowing endogenous drug metabolism. Br J Pharmacol. 2015 Aug;172(16):4119-32. doi:
10.1111/bph.13193. Epub 2015 Jun 26. PubMed PMID:
Volbracht C, Leist M, Nicotera P. ATP controls neuronal apoptosis triggered by microtubule
breakdown or potassium deprivation. Mol Med. 1999 Jul;5(7):477-89. PubMed PMID:
10449809;
87
Pöltl D, Schildknecht S, Karreman C, Leist M. Uncoupling of ATP-depletion and cell death in
human dopaminergic neurons. Neurotoxicology. 2012 Aug;33(4):769-79. doi:
10.1016/j.neuro.2011.12.007. Epub 2011 Dec 19. PubMed PMID: 22206971.
Krug AK, Kolde R, Gaspar JA, Rempel E, Balmer NV, Meganathan K, Vojnits K, Baquié M,
Waldmann T, Ensenat-Waser R, Jagtap S, Evans RM, Julien S, Peterson H, Zagoura D,
Kadereit S, Gerhard D, Sotiriadou I, Heke M, Natarajan K, Henry M, Winkler J, Marchan
R, Stoppini L, Bosgra S, Westerhout J, Verwei M, Vilo J, Kortenkamp A, Hescheler J,
Hothorn L, Bremer S, van Thriel C, Krause KH, Hengstler JG, Rahnenführer J, Leist M,
Sachinidis A. Human embryonic stem cell-derived test systems for developmental
neurotoxicity: a transcriptomics approach. Arch Toxicol. 2013 Jan;87(1):123-43. doi:
10.1007/s00204-012-0967-3. Epub 2012 Nov 21. PubMed PMID: 23179753; PubMed
Central PMCID: PMC3535399.
Balmer NV, Leist M. Epigenetics and transcriptomics to detect adverse drug effects in model
systems of human development. Basic Clin Pharmacol Toxicol. 2014a Jul;115(1):59-68.
doi: 10.1111/bcpt.12203. Epub 2014 Mar 3. Review. PubMed PMID:
Balmer NV, Klima S, Rempel E, Ivanova VN, Kolde R, Weng MK, Meganathan K, Henry M,
Sachinidis A, Berthold MR, Hengstler JG, Rahnenführer J, Waldmann T, Leist M. From
transient transcriptome responses to disturbed neurodevelopment: role of histone
acetylation and methylation as epigenetic switch between reversible and irreversible drug
effects. Arch Toxicol. 2014b Jul;88(7):1451-68.
88
CHAPTER 3
3. 3D DIFFERENTIATION OF LUHMES CELL LINE TO STUDY
RECOVERY AND DELAYED NEUROTOXIC EFFECTS
*The work presented in this chapter is published in the following article:
Harris G, Hogberg, H.T, Hartung T. & Smirnova L. (2017). 3D differentiation of LUHMES cell line to study recovery and delayed neurotoxic effects. Current Protocols in Toxicology, 73,
11.23.1–11.23.28. doi: 10.1002/cptx.29A
Key points
• The LUHMES dopaminergic cell line can be differentiated in 3D, using gyratory shaking.
• This model can be characterized based on gene expression, immunostaining and flow
cytometry, which must be performed before implementation in a new laboratory.
• Assays to measure viability, mitochondrial membrane potential, ATP levels and neurite
outgrowth must be adapted for 3D culture.
89
• 3D LUHMES is a suitable model to study delayed neurotoxic effects after compound wash
out.
90
3.1. ABSTRACT
Current neurotoxicity testing and the study of molecular mechanisms in neurodegeneration in vitro usually focus on acute exposures to compounds. 3D Lund human mesencephalic
(LUHMES) cells allow long-term treatment or pulse exposure in combination with compound washout to study delayed neurotoxic effects as well as recovery and neurodegeneration pathways.
In this unit we describe 3D LUHMES culture and characterization. Characterization of the model involves immune cytochemistry, flow cytometry, and qPCR measurements. Studying the delayed effects of compounds is more relevant to human exposures and neurodegenerative diseases with a strong genetic or environmental component. Most assays for molecular endpoints have been developed for monolayer cell culture and therefore need to be adapted for 3D models. In this unit, we further describe toxicological assays for molecular endpoints such as ATP levels, mitochondrial viability, and neurite outgrowth, which have been adapted for use in 3D LUHMES cultures.
3.2. INTRODUCTION
Lund human mesencephalic (LUHMES) cells are a human, mesencephalon-derived cell line, immortalized by a tetracycline-regulated v-myc vector. These cells are increasingly used in in vitro research (Efremova et al., 2015; Krug et al., 2014; Noelker et al., 2015; Oliveira et al., 2015;
Poltl, Schildknecht, Karreman, & Leist, 2012; Schildknecht et al., 2013; Tong et al., 2016; Zhang,
Yin, & Zhang, 2014), owing to their fast differentiation into homogeneous dopaminergic cell populations (Scholz et al., 2011). In vitro toxicology is evolving towards the use of complex models, favoring human-derived cell lines and 3D cultures. It has been shown that 3D models show increased differentiation, signaling, and cell-cell interactions, which better recapitulate tissue structure and function when compared to traditionally used monolayer cultures (Alepee et
91 al., 2014; Pampaloni, Reynaud, & Stelzer, 2007). Fast growing technologies (bio-printing, cell culture scaffolds, and bioreactors) have made 3D cultures possible (Knight & Przyborski, 2015).
However, every model needs to be extensively characterized, and molecular and toxicological endpoints (such as those described in this unit) must be adapted to 3D cultures. For certain assays,
3D models can be treated similar to tissue samples due to their size, complexity, and cell interactions. During development of a new in vitro model, comparison to the existing model, as well as characterization, is important to understand its cellular composition and whether it can better answer a specific research question. We propose that immunostaining, flow cytometry, and qPCR methods be used to identify and quantify expression of neuronal differentiation in the 3D
LUHMES model (e.g., NeuN, TH, DAT, MAP2, VMAT2, β-III-Tubulin, SYN1) and proliferation markers (e.g., Ki67, Nestin). Controlling cellular composition and cell death (e.g.,
Annexin V, caspase 3/7 activation [for apoptosis], 7-AAD [for necrosis]) in 3D models is necessary to ensure that the results obtained are derived from differentiated cells and not undifferentiated or dead cells within aggregates.
As mentioned above, to apply 3D models for research, attention must be paid to aggregate shape and size, which must remain within 250 to 350 μm, to avoid possible necrosis in the core of an aggregate (Mehta, Hsiao, Ingram, Luker, & Takayama, 2012; Ou & Hosseinkhani, 2014; Zanoni et al., 2016). Maintaining the size below 350 μm allows for all cells throughout the aggregate to receive the nutrients and necessary media components for differentiation (Smirnova, Harris, et al.,
2015). In the case of LUHMES semi-confluent monolayer cultures, cells rapidly exit the cell cycle upon induction of differentiation (Scholz et al., 2011). In the 3D differentiation model, cell suspensions are placed on a gyratory shaker, which leads to formation of aggregates with tight cell-cell interactions. These interactions stimulate cell-to-cell signaling, and a significant
Percentage of the cells continue to proliferate. To inhibit LUHMES proliferation, treatment with the anti-proliferative drug paclitaxel (Taxol) during the early days of differentiation (days 3 to 5)
92 was introduced in the 3D protocol, showing efficient inhibition of cell proliferation and generation of pure differentiated dopaminergic neuronal aggregates, which can be kept in culture for at least 21 days (Smirnova, Harris, et al., 2015). This unit describes how to differentiate
LUHMES cells in 3D (Basic Protocol 1) and ensure quality of the model (Support Protocols 1, 2, and 3). This 3D model was developed to recapitulate tissue-like cellular interactions and prolong the shelf-life of differentiated cultures (>2 weeks). It has previously been observed that medium exchanges and certain compound exposures can lead to disruption of neurodifferentiated monolayer culture morphology (Constantinescu, Constantinescu, Reichmann, & Janetzky, 2007;
Ilieva & Dufva, 2013; Smirnova, Harris, et al., 2015). In this 3D model, aggregates are in suspension; neurite outgrowth arborization and cell connections within the floating aggregates protect these structures from the mechanical disruptions during medium changes and washing steps. Furthermore, most toxicity testing to date involves measuring acute effects of single-dose exposures; however, 3D models allow measurement of chronic effects of low-dose long-term exposures, as well as delayed toxic effects and cellular adaptation after compound removal.
The second part of this unit describes the step-by-step protocol for compound exposure and washout using this 3D model (Basic Protocol 2). Support Protocols 4, 5, 6, and 7 describe 3D protocol adaptations to measure endpoints which can be used to assess acute and chronic effects
(on viability, neurite outgrowth, mitochondrial function) of a single-dose exposure to the known dopaminergic toxicant rotenone. With the development of in vitro 3D models, testing chronic or delayed effects of compounds as well as cellular adaptation and/or recovery after stress will provide data which are more relevant to human exposures leading to disease.
3.3. MATERIALS METHODS
3.3.1. BASIC PROTOCOL 1: LUHMES DIFFERENTIATION IN 3D
93
LUHMES is a tetracycline-regulated v-myc-vector immortalized cell line. In the presence of
tetracycline, cAMP and GDNF, the v-myc promotor is switched off and LUHMES rapidly exit
the cell cycle and homogeneously differentiate to mature dopaminergic neurons (Lotharius et al.,
2005; Scholz et al., 2011). LUHMES routine culture and neuronal differentiation in monolayer
were extensively characterized (Krug et al., 2014; Scholz et al., 2011). This protocol will describe
the adaptation of LUHMES neuronal differentiation into 3D cultures. We developed a 3D
protocol in order to perform long-term exposures as well as to wash-out the compounds for
cellular recovery and adaptation studies.
Basic protocol 1 describes LUHMES cell culture and differentiation in 3D. Support protocol 1
covers the characterization of differentiation by immunohistochemistry. Support protocol 2
describes the characterization of neuronal differentiation by flow cytometry with the emphasis in
handling of spheroids for flow cytometry analysis. Support protocol 3 refers to the essential steps
of characterization by RT-PCR.
Materials
See “Reagents and Solutions” for main recipes, and stock solution preparations used in this
protocol.
- LUHMES ATCC® CRL-2927™
- Coating solution (see recipe, Table 1)
- Proliferation medium (see recipe, Table 2)
- Differentiation medium (see recipe, Table 2)
- Wash medium (see recipe, Table 2)
- Poly-L-Ornithine (Sigma-Aldrich, P-3655) (see recipe)
- Fibronectin solution (Sigma-Aldrich, F-1141) (see recipe)
- Advanced DMEM/F-12 medium (TermoFisher Scientific, cat # 12634-010)
94
- N-2 supplement (TermoFisher Scientific, cat # 17502048)
- L-Glutamine (Sigma-Aldrich, cat # G7513)
- Human basic fibroblast growth factor (bFGF, R&D Systems, cat # 4114-TC) (see recipe)
- Glia Derived Neurotrophic factor (GDNF, Gemini, cat # 300-121P) (see recipe)
- Tetracyclin (Sigma-Aldrich, cat # T-7660) (see recipe)
- cAMP (Santa Cruz, cat # 16980-89-5) (see recipe)
- TrypLETM Express (TermoFisher Scientific, cat # 12605036)
- Taxol (Sigma, cat # P4543) (see recipe)
- Dulbecco’s Phosphate buffered saline (PBS) solution without Ca2+ and Mg2+ (Quality Biological)
- TrypanBlue
- Double distilled water
- Sterile syringe filters, 0.20 µm (Corning)
- T75 and T175 Flasks (NuncEasYFlask, Nunclon Delta Surface, cat # for 75 cm2 Flask 156472)
- 6-well plates (BD)
- Cell culture centrifuge
- Phase-contrast microscope
- Cell counter or hemocytometer
- Cell culture incubator suitable to accommodate a shaker
- Cell culture gyratory shaker. Model used in this protocol: ES-X Kuhner laboratory shaker, Tray
size: 420x420 mm, orbital shaking motion with diameter 50 mm, speed 80 rpm.
Protocol Steps
All cell culture steps should be conducted in sterile laminar hood, biosafety 2. All solutions and
media should be kept sterile.
All culture reagents must be pre-warmed to 37 oC in a water bath.
95
LUHMES cells should only be used up to passage 25.
All plates and flasks for routine cultures must be NuncEasYFlask, Nunclon Delta Surface
For routine cell culture, thawing and freezing details refer to protocol described by (Krug et al.,
2014; Scholz et al., 2011).
Flask Coating
1. Prepare coating solution according to Table 1.
2. Add coating solution to flasks, (T25 Flask: 4 mL; T75 Flask: 8 mL; T175 Flask: 14 mL) and incubate overnight at 37 °C.
3. After incubation aspirate the coating-solution and wash twice with sterile distilled water.
4. Dry flasks under the laminar flow bench before use.
Routine Cell Culture and Preparation for Differentiation
Follow the cell culture protocol as described by (Krug et al., 2014; Scholz et al., 2011). To prepare cells for differentiation, expand LUHMES culture as follows:
5. Culture 2 x 106 LUHMES cells in a T75 pre-coated flask containing 12 mL proliferation medium. Incubate at 37 °C, 5% CO2, 95% humidity.
6. When 70-80% confluency is reached, passage the cells by removing the medium and
TM adding 2 mL TryplE for 3 min at 37 °C, 5% CO2, 95% humidity.
7. When the cells have detached, re-suspend the cells in 10 ml DMEM/F12 Advanced medium without supplements and pipette up and down several times to prepare a single cell suspension.
8. Transfer the cell suspension to a 50 mL falcon tube containing 10 mL DMEM/F12
Advanced medium without supplements and pipette up and down three more times.
9. Centrifuge the cells 3 min at 1000 rpm, room temperature.
96
10. Aspirate the medium, and gently re-suspend the pellet in 15 ml of proliferation medium.
11. Count the cell stock using a cell counter and trypan blue.
12. Fill the pre-coated culture Flask with proliferation medium. Final volume for T75 Flask is
12 mL, for T175 - 22 mL.
13. Passage 2 x 106 cells into T75 flask to passage cells or 4 x 106 cells into a T175 flask for differentiation.
14. Place flasks in an incubator at 37 °C, 5% CO2, 95% humidity for 48 hours.
3D culture and differentiation
Day 0: Initiation of differentiation
15. Prepare necessary volume of differentiation medium (Table 2)
Prior to use, warm DMEM/F12 Advanced medium, differentiation medium and TryplETM solution to 37 °C in a water bath.
16. Un-package and label 6-well plates and place under sterile laminar hood.
Examine cell culture from step 11 by phase-contrast microscope, which should be 70-80% confluent.
17. Remove the medium, add 2-3 mL of pre-warmed TryplETM solution and incubate 3-5 min in incubator at 37 °C, 5% CO2, 95% humidity.
18. When the cells have detached, follow the steps 6 to 8.
19. Aspirate the medium and gently re-suspend the pellet in 15 ml of differentiation medium.
20. Count the cell stock. Ideally this cell stock is between 2-3 x 106 viable cells per mL.
Note: one 80% confluent T175 flask contains sufficient cells for ~ 8 6-well plates.
21. Add 2 mL of differentiation medium to each well of desired number of 6 well plates.
22. Add the volume of cell suspension from step 18 required to reach a final concentration of
5.5 x 105 cells per well.
97
Cell density is an important parameter for differentiation. Too many cells can increase LUHMES proliferation rate during the later steps of the differentiation process. Make sure to have no more than 5.5 x 105 cell per well.
The humidified incubator containing the gyratory shaker is kept at 10 % CO2 to ensure CO2 diffusion to the inside of aggregates under constant gyratory movement.
23. Place 6-well plates on a shaker at 80 rpm (orbital diameter 50 mm) in an incubator at
37 °C, 10 % CO2, 95% humidity (Video 1).
Note: For all steps below involving medium exchange, compound treatment and washing out compounds, do not take more than two plates out from the incubator (shaker) at once and work quickly.
Make sure aggregates are not exposed to air for more than a few seconds. When fresh medium is added, gently shake the plate prior returning it to the incubator to make sure aggregates are floating as single aggregates and are not stuck together (Video 1).
Day 3: Addition of anti-proliferation drug, Taxol
Note: In monolayer cultures LUHMES rapidly exit cell cycle after induction of differentiation, however, due to increased cell-to-cell interaction within 3D aggregates, a portion of LUHMES cells continue to proliferate (observed in (Smirnova et al., 2015a)). To inhibit proliferation, ensure complete differentiation, and maintain the size of the aggregates below 350 μm; treatment with anti-proliferation drug, Taxol, was introduced into the 3D differentiation protocol.
24. Prepare fresh anti-proliferation medium (Table 2).
For anti-proliferation medium, standard differentiation medium is supplemented with 2 μM
Taxol stock to reach the final Taxol concentration of 20 nM (see recipe).
To calculate total medium amount, count 1 mL of anti-proliferation medium per well.
Pre-warm the medium to 37 °C.
98
25. Transfer plates from the shaker to the laminar hood.
Examine the culture by phase-contrast microscope: Three days after cell seeding, aggregate formation is visible.
26. Shake plate in small circle motions to allow aggregates to collect in the middle of the well (Video 1).
27. Tilt plate slightly, remove 800 μL medium and add 1mL of anti-proliferation medium to each well.
Final Taxol concentration in each well is 10 nM. Aspiration of 800 μL instead of 1 mL from each well takes in account evaporation during the incubation.
28. Return plates back to the incubator on the shaker at 80 rpm at 37 °C, 10 % CO2, 95% humidity.
Day 5: Taxol wash-out (Video 1)
29. Prepare the necessary volume of wash and differentiation media (Table 2). Pre-warm media to 37 °C.
30. Transfer plate from the shaker to the laminar hood (Video 1).
31. Shake plate in small circle motions to allow aggregates to collect in the middle of the well (Video 1).
32. Tilt plate slightly and remove the maximum possible volume of medium (leaving <200
μL in the well). Add 2 mL of wash medium (Video 1).
33. Repeat step 30 to remove wash medium and add 2 mL of differentiation medium to each well (Video 1).
34. Return plates back to the incubator on the shaker at 80 rpm at 37 °C, 10 % CO2, 95% humidity.
Day 8, 10, 12, 15, and 18: Medium exchanges
99
35. Shake plate in small circle motions to allow aggregates to collect in the middle of the
well.
36. Tilt plate slightly and remove 1 mL medium from each well.
37. Add 1.2 mL differentiation medium to each well.
38. Cultures can be kept at least up to 21 days.
SUPPORT PROTOCOL 1
CHARACTERIZATION OF NEURONAL DIFFERENTIATION BY
IMMUNOCYTOCHEMISTRY
Different antibodies can be used to visualize LUHMES differentiated cells (NeuN, MAP2, NF200
and SYN1) and proliferating cells (Ki67). Although methods such as flow cytometry are more
quantitative, immunocytochemistry can provide details on aggregate structure, as well as cell and
protein localization within aggregates. This protocol is modified from routine
immunocytochemistry for monolayer cultures to enhance antibody penetration into aggregates.
Note: Antibody dilutions suggested in Table 4 will vary if antibodies from different provider/clone
are used.
Materials
- Paraformaldehyde (PFA) 4%
- Clearing solution (see recipe, Table 3)
- Blocking solution (see recipe, Table 3)
- Wash solutions I and II (see recipe, Table 3)
- Dulbecco’s Phosphate buffered saline (PBS) solution without Ca2+ and Mg2+ (Quality Biological)
- Antibodies (see Table 4)
100
- Hoechst 33342
- Immu-Mount mounting medium (Fisher Scientific)
- 1.5 mL Eppendorf tubes (Fisher Scientific)
- Glass slides (Fisher Scientific)
- Glass cover slip (Fisher Scientific)
- Nail polish
- Eppendorf tubes
- 24-well plates
- Microplate Shaker
- Fluorescence microscope (Nikon)
- Zeiss LSM 510 Confocal III confocal microscope (Zeiss)
- ZEN imaging software (Zeiss)
- ImageJ imaging software
Protocol Steps
1. Prepare wash solutions I and II, optical clearing solution, blocking solution (see recipe).
Keep the solutions on ice.
2. Transfer plate from the shaker to the laminar hood.
3. Shake plate in small circle motions to allow aggregates to collect in the middle of the
well.
4. Collect ~ 20-40 aggregates per condition in 1.5 mL Eppendorf tubes.
The number of aggregates does not have to be exact, but similar for all conditions.
5. Allow aggregates to sink to the bottom of the tube and wash once with 500 μL cold PBS.
6. Allow aggregates to sink to the bottom of the tube (3-5 min), aspirate PBS and add 500
μL of cold wash solution I and 500 μl 4% PFA. Incubate for 45 min at 4oC.
101
Look at the tube against light every time prior to washing to determine whether aggregates have completely sunk to the bottom of the tube.
7. Wash fixed aggregates twice with 500 μL wash solution I as described in step 5. Aspirate wash solution I and add 500 μL Optical Clearing Solution. Incubate for optical clearing 48 hours at 4oC on the shaker.
Note: this and the following steps can be performed in Eppendorf tubes or 24-well plates.
Working with 24-well plates (200 μL per well) will give a more homogeneous staining throughout aggregate, however lower volumes can be used with Eppendorf tubes (50 μL per tube).
8. Remove optical clearing solution and add 500 μL blocking solution and place tubes/plate on a shaker for 1 hour at 4 oC.
9. Aspirate blocking solution.
10. Incubate aggregates in 200 μL primary antibody diluted in blocking solution for 48 h at
4oC on a shaker (see Table 4 for dilutions and list of antibodies).
11. Aspirate primary antibody and wash aggregates three times for 15 min each with wash solution II.
12. Remove wash solution II and incubate aggregates in 200 μL secondary antibody diluted in blocking solution (see Table 4) for 24 h at 4 oC on a shaker. Keep samples protected from light.
13. Aspirate secondary antibody and wash aggregates three times for 15 min each with wash solution II and incubate with Hoechst 33342 nuclear stain diluted in PBS (1:10,000) for 1 h at room temperature on a shaker.
14. Wash aggregates twice with wash solution I, and once with PBS. Keep aggregates in
PBS.
15. Label glass slides and place a drop of mounting medium to mount the aggregates.
16. Cut off the end of a 200 uL tip and transfer aggregates from the Eppendorf tube/plate to the glass slide. Ensure mounting medium is covering aggregates and place a coverslip over slide.
Note: Avoid air bubble and keep slides protected from light.
102
17. Use nail polish to seal coverslip, allow slides to dry and preserve slide at 4 oC, protected
from light.
18. Image with fluorescence or confocal microscope following the manufacturer’s
instructions.
SUPPORT PROTOCOL 2
CHARACTERIZATION OF NEURONAL DIFFERENTIATION BY FLOW
CYTOMETRY
Flow cytometry can be used to quantify percentage of proliferating cells within the aggregates
and to assess apoptotic and necrotic cells. The flow cytometry protocol is extensively described in
Current Protocols in Toxicology unit 20.9, supporting protocol 1 (Smirnova et al., 2015b). In this
protocol, only modification steps for 3D cultures will be described.
Materials
- PBS-EDTA (see recipe, Table 3)
- EDTA 0.5 M (Sigma)
- TryplETM Express (ThermoFisher Scientific, cat # 12605036)
- RNAse-free DNAse 1500 Kunitz units (Qiagen)
- 4% PFA
- Blocking solution (see recipe, Table 3)
- Wash solution I and II (see recipe, Table 3)
- Dulbecco’s Phosphate buffered saline (PBS) solution without Ca2+ and Mg2+ (Quality Biological)
- Alexa Fluor® 647 Mouse anti-Human Ki-67 antibody (BD, cat # 561126)
- Alexa Fluor® 647 Mouse IgG1 κ Isotype control (BD, cat # 557783)
- PE Annexin V Apoptosis Detection Kit I (BD, cat # 559763)
103
- 1.5 and 2 mL Eppendorf tubes
- 1 mL BD Luer-Lok™ syringe with 20 G x 1 in. BD PrecisionGlide™ needle (BD)
- Falcon® 5 mL round bottom polystyrene test tubes with cell strainer snap cap (BD Biosciences)
- Centrifuge
- Flow cytometer (BD Biosciences)
Protocol steps
1. Retrieve differentiated aggregates from the incubator.
2. Remove medium from the cultures in 6 -well plates. Rinse cells with 3 mL of PBS-
EDTA.
3. Add 1 mL dissociation solution (TryplETM Express containing 4 units/mL DNAse).
4. Return plates to the shaker in the incubator for 30 min.
TrypLETM Express, a recombinant trypsin, is very gentle and effective solution and should be used
for dissociation to achieve better cell survival.
5. Dissociate the aggregates by gently pipetting up and down using a 200 µL pipette.
If aggregates are not dissociated, return the plate to the shaker in the incubator for another 10-
15 min. Use a 1 mL syringe with 20 G x 1 in. needle to generate single cell suspension by passing
un-dissociated aggregates through the needle 2-3 times. This form of dissociation is more
aggressive and should be performed carefully.
6. Collect the entire cell suspension into a 2-mL centrifuge tube containing ice-cold
PBS/EDTA solution.
7. Count a sample of the cell suspension.
Choose a method that guarantees to calculate the amount of viable cells before fixing the cells
(e.g., trypan blue staining). Viability should be no lower than 80%
8. Centrifuge the cells 3 min at 1,500 g, 4 °C.
All subsequent steps should be performed on ice.
104
9. For Annexin staining, re-suspend the pellet in 1x Annexin V binding buffer and follow the manufacturer’s protocol for further steps
(http://www.bdbiosciences.com/ds/pm/tds/559763.pdf).
10. For intracellular antibody staining re-suspend the pellet in 500 μL wash solution I, add
500 μL 4% PFA and fix the cells 20 min at 4 °C.
11. Follow the Current Protocols in Toxicology Support Protocol 1, Unit 20.9 for washing and blocking steps (Smirnova et al., 2015b).
12. Dilute anti-Ki-67 and corresponding isotype control antibodies in blocking solution
(dilution determined by manufacturer).
13. Add 100 μL pre-diluted antibody solution to 1x106 cells and incubate 45 min on ice protected from light. Include unstained cells as additional control.
14. Wash twice with wash solution II and once with wash solution I. Re-suspend in 300 μL wash solution I.
15. Transfer the cells into the 5 mL Falcon flow cytometry test tube, passing the cell strainer, and proceed to measurement, following flow cytometer manufacturer’s instructions and Current
Protocols in Toxicology Support Protocol 1, Unit 20.9 (Smirnova et al., 2015b).
SUPPORT PROTOCOL 3
CHARACTERIZATION OF NEURONAL DIFFERENTIATION BY qRT-PCR
To ensure LUHMES differentiation in 3D aggregates, expression of specific neuronal markers should be assed as quality control of the cultures. The protocols for RNA extraction, cDNA synthesis and qRT-PCR is no different to monolayer cultures, and are described in Current
Protocols in Toxicology, Unit 20.9 (Smirnova et al., 2015b) and TaqMan gene expression assay protocol
(http://www3.appliedbiosystems.com/cms/groups/mcb_support/documents/generaldocuments/cm s_041280.pdf). To assess the quality of the cultures, differentiation markers such as tyrosine
105
hydroxylase (TH), Dopamine transporter (DAT), β-III-tubulin and Synapsin 1 (SYN1) should be
checked for significant increase in expression by day 3, compared to Day 0. Similarly, strong
down-regulation of proliferation marker Ki-67 (lower than 5-10%) and neoroprogenitor marker
Nestin (NES) should be observed by day 6 to ensure complete differentiation and verify that no
expansion of proliferating cell populations is present (refer to Figure 5 in (Smirnova et al., 2015a)
for details).
3.3.2. BASIC PROTOCOL 2: COMPOUND TREATMENT AND
WASH-OUT
The following protocol describes compound exposure scheme for 3D LUHMES in 6-well and 24-
well formats. 3D cultures are particularly useful to study delayed cellular response after
compound removal. Basic protocol 2 describes how to wash out compounds to study long-term
effects, cellular recovery and adaptation pathways. Measurement of the compound concentration
in the medium prior to and after washing steps is recommended (but not described in this
protocol), as different compounds have different degradation and diffusion rates within the
aggregates. The number of washing steps may need to be modified for different compounds.
Materials
- Differentiated 3D LUHMES (from Basic Protocol 1)
- Rotenone 100 mM (see recipe)
- Dimethyl sulfoxide (DMSO) HPLC grade, 99.9% (Sigma-Aldrich)
- Dulbecco’s Phosphate buffered saline (PBS) solution without Ca2+ and Mg2+ (Quality Biological)
- Differentiation medium (Table 2)
- Wash medium (see recipe, Table 2)
- 24-well plates (BD Biosciences)
106
- 6-well plates (BD Biosciences)
- 15 mL tubes (BD Biosciences)
Protocol Steps:
A. Treatment in 24-well plate
For concentration range finder experiments and dose-response curves, a 24-well plate format can
be used and is described below. The differentiating cultures are dynamic and changing over the
time. To minimize variability between samples due to differentiation, apply reverse treatment
scheme (Figure 1) and collect all samples at the same time/stage of differentiation.
1. On the day of treatment defrost rotenone stock (100 mM), or compound of interest, stored
at -20 °C.
2. For treatment, pre-dilute 100 mM Rotenone in differentiation medium (1:1000) to obtain
a 100 µM stock.
When adding volume of 100 mM rotenone to medium, a white precipitation is observed. Vortex
the solution for 30 seconds to ensure that the compound has fully dissolved and no white
precipitate remains.
3. Dilute DMSO (1:1000) and vortex for 30 seconds.
4. Prepare serial dilutions of stock compound (Figure 2).
5. Retrieve the plate with differentiated aggregates from the incubator. Shake plate in small
circle motions to allow aggregates to collect in the middle of the wells.
6. Collect all aggregates using a 1000 µL pipette. Pool aggregates from 6 wells together into
a 15 mL tube.
7. Allow aggregates to sink to the bottom of the tube.
8. Remove medium from 15 mL tubes and re-suspend aggregates in 6 mL differentiation
medium.
107
9. Pipette up and down carefully and transfer 250 μL of medium with aggregates into each well of 24-well plate.
Aggregates must be distributed evenly across wells. Ensure to pipette up and down carefully every time 250 μL are transferred to another well.
10. Add 250 μL of compound serial dilutions from step 4 to designated wells (see Figure 3 for plate layout).
Note: The serial dilutions are prepared 2x the final concentration desired in the well.
11. Gently tilt plate to ensure that aggregates are not collecting in the middle of the wells to avoid agglomeration.
12. Place 24-well plate in a cell culture incubator at 37 °C, 5 % CO2, 95% humidity for desired compound treatment time (up to 48 h).
Note: Handle the plate with caution, to avoid aggregates accumulation in the middle of the plate and formation of single aggregate.
B. Treatment in 6-well plate
1. Differentiate 3D LUHMES cells in 6- well plates as described in Basic Protocol 1.
2. On the day of treatment defrost rotenone stock (100 mM), or compound of interest, stored at -20 oC.
3. For treatment, pre-dilute 100 mM rotenone stock in differentiation medium (1:10,000) to obtain a 10 μM rotenone solution.
It is recommended to store rotenone at high concentrated stock solution and perform high dilution, since the compound is not soluble in water or medium.
When adding volume of 100 mM rotenone to medium, a white precipitation is observed. Vortex the solution for 30 seconds to ensure that the compound has fully dissolved and no white precipitate remains.
108
4. Dilute DMSO (1:10,000) in differentiation medium and vortex for 30 seconds
5. Add volume of 10 µM rotenone solution, required to reach desired final concentration in
6-well plate containing 2 mL medium per well. Add the same amount of 1:10,000 diluted DMSO solution to vehicle control wells.
The necessary volume to achieve the desired final concentration can be calculated as follows:
퐶1 × 푉1 푉2 = 퐶2 where,
C1 = desired final concentration in µM
C2 = 10 µM rotenone
V1 = 2000 μL volume of medium in well
V2 = required volume of 10 µM rotenone solution in μL
6. Place 6-well plate on gyratory shaker at 80 rpm, in an incubator at 37 °C, 10 % CO2, 95% humidity.
7. Incubate LUHMES aggregates for the experimental exposure time.
8. Collect the aggregates for end-point measurement (described in Support protocols 4, 5 and 6) or proceed to compound wash-out steps of this protocol.
C. Compound wash-out (Video 1)
If experimental design foresees studies of delayed compound effects after compound withdrawal, wash-out steps are introduced to the protocol. In order to avoid exposure to compound, which may be retained and released by the plastic of 6-well plates, this protocol involves washing aggregates and transferring them into new culture plates.
1. After desired exposure duration, retrieve 2 plates from the incubator at a time (Video 1).
2. Shake plate in small circle motions to allow aggregates to collect in the middle of the wells (Video 1).
109
3. Tilt plate and using a 1 mL pipette, remove medium from each well leaving 200 μL to avoid aggregate exposure to air (Video 1).
4. Add 2 mL wash medium to each well (Video 1).
These steps must be performed quickly to avoid aggregate exposure to air and clumping (sticking together) during washing.
5. Add 2 mL differentiation medium per well in new 6-well plates (Video 1).
6. Shake plate containing aggregates in wash medium in small circle motions to allow aggregates to collect in the middle of the well. Tilt plate slightly and collect aggregates and as little medium as possible using a 200 μL pipette (Video 1).
7. Transfer the aggregates into the new plate containing 2 mL differentiation medium
(prepared in step 5) (Video 1).
8. Return plates to shaker at 80 rpm, in an incubator at 37 °C, 5 % CO2, 95% humidity
(Video 1).
SUPPORT PROTOCOL 4
VIABILITY ASSAY AFTER COMPOUND TREATMENT
In this protocol two assays to study cell viability based on mitochondrial integrity and function are described. The ability of mitochondria to uptake and irreversibly reduce resazurin, informs about the aerobic respiration capacity of mitochondria (metabolic activity). The ATP assay gives total cellular ATP content and informs about cellular respiration. ATP levels will vary depending on the number of cells in the sample therefore a surrogate measure of cell number such as total protein concentration must be conducted in parallel to normalize these results.
A. Resazurin assay for 3D Culture
Materials
110
- Differentiated 3D LUHMES (after compound treatment, Basic Protocol 2)
- Resazurin solution 1 mg/mL (see recipe)
- 6-well, 24-well plates
- Black 96-well multi-well plates (BD Biosciences)
- Fluorescence Microplate Reader (PerSeptive Biosystems CytoFluor II)
Protocol Steps:
1. Differentiate cells following Basic Protocol 1.
2. Treat aggregates according to Basic Protocol 2 either in 24- or 6-well plate format.
3. Prepare 1 mg/mL resazurin solution (see recipe).
4. Retrieve the plates from the incubator and place them under laminar hood.
5. Remove half of the medium from each well.
6. Add 10 % volume 1 mg/mL resazurin solution to each well (i.e. add 100 µL 1 mg/mL
resazurin solution to 1 mL medium in each well of 6-well plate or 25 µL to 250 µL medium in
24-well plates).
7. Return the cell culture plate to the incubator at 37 °C, 5 % CO2, 95% humidity.
8. Prepare a blank solution with 90 % differentiation medium and 10 % 1 mg/mL resazurin
solution.
9. Incubate for 1 – 4 hours until a visible color change from blue to purple is observed in
control wells.
10. To measure, carefully pipette 100 µL from each well into 96-well plate in triplicates per
condition.
Note: avoid air bubble formation, since this may interfere with the measurement.
11. Read fluorescence emission at 580-610 nm.
12. Calculate % cell viability by comparing relative fluorescence units (RFU) of treated vs.
vehicle control samples using following formula:
111
푅퐹푈(푡푟푒푎푡푒푑)– 푅퐹푈(푏푙푎푛푘) 푽풊풂풃풊풍풊풕풚 (%) = × ퟏퟎퟎ (푅퐹푈(푐표푛푡푟표푙) – 푅퐹푈(푏푙푎푛푘) )
B. ATP assay for 3D Culture
Materials
- Differentiated 3D LUHMES (after compound treatment, Basic Protocol 2)
- ATP determination kit (ThermoFisher Scientific, Cat # A22066)Dulbecco’s Phosphate buffered
saline (PBS) solution without Ca2+ and Mg2+ (Quality Biological)
- Whole cell lysis buffer (see recipe)
- Pierce BCA Protein Assay Kit
- 1.5 mL Eppendorf tubes (TermoFisher Scientific)
- White flat bottom 96-well plates (BD Biosciences)
- GloMax® 96 Microplate Luminometer (Promega)
- Spectrophotometer (Molecular Devices)
Protocol steps
1. Defrost ATP assay components.
2. After compound treatment retrieve the 6-well plates with differentiated cells from the
incubator and place them under laminar hood.
3. Shake plate in small circle motions to allow aggregates to collect in the middle of the
well.
4. Tilt plate slightly and collect 1 well of aggregates into a 1.5 mL Eppendorf, using a
200 µL pipette.
5. Allow aggregates to sink, remove the medium.
6. Wash aggregates in each sample tube by adding 1 mL PBS.
112
7. Wait for aggregates to sink to the bottom of the Eppendorf tube and aspirate as much volume as possible.
Make sure aggregates have sunk to the bottom to not lose part of your sample.
8. Add 50 μL whole cell lysis buffer (see recipe) to each sample.
9. Pipette up and down carefully 5- 10 times and place for 30 min on ice until aggregates have dissolved.
Note: When pipetting aggregates up and down, avoid bubble formation.
For larger sample size, incubate for up to 1 hour on ice, and increase lysis buffer volume to 100
μL if aggregates have not dissolved.
10. Follow manufacturer’s instructions to prepare reaction buffer and reaction solution
(https://tools.thermofisher.com/content/sfs/manuals/mp22066.pdf).
Prepare fresh reaction solution immediately prior use and do not vortex once prepared. Reaction solution cannot be stored and should be discarded after use.
11. Label white flat bottom 96-well plates.
12. Add 100 μL of reaction solution to each well (each condition in triplicates + 3 blanks
(lysis buffer only)).
13. Add 10 μL of cell lysis to each well. Save the rest of the samples on ice for protein measurement. Alternatively, freeze lysates at -20 °C for later protein quantification.
14. Shake gently and incubate 15 min at room temperature protected from light.
15. If using GloMax® 96 Microplate Luminometer, click “activate PMT’ (5 minutes).
16. Measure luminescence using GloMax® 96 Microplate Luminometer immediately.
When testing different conditions/treatments which may affect cell viability within aggregates, normalization to number of cells is necessary. In this protocol, total protein concentration is used to normalize ATP measurements to cell number.
17. Prepare PierceTM BCA Protein working reagent (1:50) by following manufacturer’s instructions
113
(https://tools.thermofisher.com/content/sfs/manuals/MAN0011430_Pierce_BCA_Protein_Asy_U
G.pdf).
18. Pipette 200 µL BCA working reagent per well in 96-well plate (for each sample in duplicate, blank in duplicate and wells desired for BSA standard curve).
19. For each sample, transfer 10 µL of cell lysis (in duplicate) to wells containing 200 µL
BCA working reagent.
20. For blank measurements, transfer 10 µL lysis buffer (in duplicate) to wells containing
200 µL BCA working reagent.
21. Add 10 µL bovine serum albumin (BSA) standard protein concentrations (in duplicate) to wells containing 200 µL BCA working reagent.
Recommended concentrations are 0, 0.25, 0.5, 0.75, 1, 1.5, 1.75 and 2 mg/ml BSA.
22. Incubate plate 30 min at 37 oC.
23. Measure absorbance at 563 nm using spectrophotometer.
24. Generate a standard curve by plotting average blank-corrected absorption values of each
BSA standard vs. its concentration in mg/mL.
25. Calculate the protein concentration for each sample using the protein standard curve by following protocol described in (https://tools.thermofisher.com/content/sfs/brochures/TR0057-
Read-std-curves.pdf)
26. Normalize ATP relative luminescence values (RLU) of each sample (measured in step
16) to its protein concentration (mg/mL, calculated in step 25). Present normalized ATP luminescence for each condition as a % of vehicle control:
114
SUPPORT PROTOCOL 5
MEASURING MITOCHONDRIAL MEMBRANE POTENTIAL IN 3D CULTURES
To assess the effect of compounds on mitochondrial membrane potential, a MitoTracker probe, which enters and accumulates in active mitochondria, is used. As for most assays, manufacturer’s protocols provide experimental steps for monolayer cultures or cells in single-suspension. This support protocol provides the experimental steps to measure total mitochondrial membrane potential in aggregates after compound exposure.
Materials
- Differentiation medium (see recipe, Table 2)
- Differentiated 3D LUHMES (after compound treatment, Basic Protocol 2)
- MitoTracker ® Red CMXRos (Thermofisher Scientific)
- 4% PFA
- Dulbecco’s Phosphate buffered saline (PBS) solution without Ca2+ and Mg2+ (Quality
Biological)
- Immu-Mount mounting medium (Fisher Scientific)
- 24-well plates (BD Biosciences)
- Glass slides (Fisher Scientific)
- Glass cover slip (Fisher Scientific)
- Nail polish
- Fluorescence microscope (Nikon)
- ImageJ software (https://imagej.nih.gov/ij/)
115
Protocol steps
1. Differentiate cells following Basic Protocol 1.
Note: If using any fluorescence LUHMES reporter cell lines, make sure this wavelength does not interfere with that of the fluorescent dye used in the MitoTracker® assay.
2. Prepare 1 mM MitoTracker® stock solution by adding 94.1 μL DMSO to tube provided by manufacturer (protect from light and store at -20 oC for up to 6 months)
(https://tools.thermofisher.com/content/sfs/manuals/mp07510.pdf).
3. On the day of sample collection, add 500 μL differentiation medium to desired number of wells in a 24-well plate.
4. Retrieve 6-well plates with aggregates from the incubator.
5. Shake plate in small circle motions to allow aggregates to collect in the middle of the 6- well plate.
6. Collect ~ 20 aggregates per condition/replicate and pipette into 24-well plate (1 well per condition/replicate). Final volume 500 μL.
7. Add 0.5 μL of 1 mM MitoTracker® stock and 0.5 μL of 1:10 pre-diluted Hoechst 33342 to each well containing 500 μL medium.
o 8. Protect plate from light and incubate 45 min at 37 C, 5 % CO2, 95% humidity. Give plate a slight shake every 15 minutes to avoid aggregates clumping together.
9. After incubation, tilt the plate to allow aggregates to sink and remove medium containing
MitoTracker® and Hoechst 33342.
o 10. Add 500 μL fresh differentiation medium and incubate 1 hour at 37 C, 5 % CO2, 95% humidity. Give plate a slight shake every 15 minutes to avoid aggregates clumping together.
Note: this step is not in the manufacturer’s protocol
(https://tools.thermofisher.com/content/sfs/manuals/mp07510.pdf) and must be included to effectively wash out excess MitoTracker® from 3D cultures and obtain more accurate results.
116
Cells with disrupted mitochondria cannot retain the red fluorescent dye, however, the dye will remain in the aggregate if not washed sufficiently.
11. Tilt the plate to allow aggregates to sink and remove medium.
12. Wash aggregates twice for 3 min with pre-warmed PBS.
13. Aggregates can be imaged live (proceed to step 17), or fixed for later imaging.
14. To fix, remove 300 µL PBS and add 200 µL PFA 4 % (2 % PFA final).
15. Incubate aggregates 30 min at room temperature.
16. Wash twice with PBS.
17. Place a drop of mounting medium on glass slides.
18. Cut off the end of the yellow tip and transfer aggregates from the Eppendorf tube to the glass slide. Pipette directly in the middle of mounting medium drop. Avoid bubbles.
19. Ensure mounting medium is covering aggregates and place a coverslip over slide.
20. Allow the slides to dry at least 4 hours at 4 oC protected from light.
21. Use nail polish to seal coverslip and preserve slide at 4 oC, protected from light until imaging.
22. Take individual images of aggregates using a fluorescence microscope. Adjust exposure time to ensure that aggregates are not over-exposed. ConFigure the settings by checking negative and positive controls (or highest compound concentration).
Any over-exposed regions will interfere with accurate quantification of fluorescence intensity.
Make sure the exposure time is consistent throughout the experiments for all conditions.
23. Quantify images using ImageJ open source software (Figure 4).
24. From the dropdown menu chose “Analysis” – “Set Measurements”. Activate area, min & max gray area and mean grey area.
25. Use “oval” selection to select aggregate area. From the dropdown menu chose “Analysis”
- “Measure” or use key combination Ctrl+M (Comand+M) to quantify mean intensity (mean grey value) in each aggregates. The table with the results will be generated automatically.
117
26. Measure mean intensity for 15-20 aggregates per condition.
27. Save results table as text or excel format.
28. Normalize the intensities (mean grey values) in treated samples to the intensity of the
vehicle control. Use median with interquartile range and scatter dot plot to present the results.
SUPPORT PROTOCOL 6
NEURITE OUTGROWTH QUANTIFICATION
This protocol involves LUHMES 3D aggregates, stably expressing Red Fluorescence Protein
(RFP, (Schildknecht et al., 2013)) in order to enable easy and fast visualization of the neurites
without additional antibody staining. After differentiation and compound treatment in 3D,
LUHMES aggregates are seeded onto matrigel-coated 96-well plates. Clear radial neurite
outgrowth occurs from aggregates within 24 hours after seeding, which can be then imaged and
quantified using automated high content imaging by measuring parameters such as number and
density of neurites and neurite total length per aggregate.
Materials
- Differentiated aggregates (Basic Protocol 1) from LUHMES stably transfected with RFP reporter
(kindly provided by Prof. Dr. Leist, Konstanz, Germany).
- BD Matrigel hESC-qualified Matrix 5 mL (BD Biosciences Research #354277)
- Differentiation medium (see recipe, Table 2)
- 4% PFA
- 1.5 mL Eppendorf tubes (Fisher Scientific)
- Black 96-well plate, clear glass bottom (ThermoScientific)
- ArrayScan Confocal Module ArrayScanTM XTI (Thermo ScientificTM)
118
Protocol steps
1. Differentiate RFP-LUHMES cells following Basic Protocol 1.
2. Treat the cells following Basic Protocol 2.
3. 24 hours prior to sample collection, defrost an aliquot of matrigel on ice.
Note: Allow at least 4 hours for defrosting. Defrosting must be on ice. Keep matrigel cold, since it starts to polymerize at room temperature.
4. Dilute 125 μL matrigel in 10 mL DMEM/F12 advanced medium. Add 70 μL of diluted matrigel to each well of a black/clear bottom 96-well plate. Avoid air bubbles formation. Incubate overnight in cell culture incubator.
Note: Ensure there is sufficient matrigel and that it does not evaporate from the wells prior to seeding.
5. On the day of sample collection, aspirate matrigel from 96-well plate.
6. Add 50 μL differentiation medium per well of the 96-well plate.
7. Retrieve the 6-well plate with differentiated aggregates from the incubator and place under laminar hood.
8. Shake plate in small circle motions to allow aggregates to collect in the middle of the well. Collect aggregates using a 200 μL pipette. Collect one well of aggregates for each condition into a 1.5 mL Eppendorf tube.
9. Allow aggregates to sink to the bottom of the tube.
10. Remove medium from Eppendorf tubes and re-suspend aggregates in 200 μL of fresh differentiation medium.
11. Pipette up and down carefully and transfer 50 μL of aggregates into each well of 96-well plate from step 6. This is sufficient volume for 4 technical replicates (4 wells).
Note: Handle the plate with caution and avoid shaking. Visually check that aggregates have not clumped together in the center of the well. Aggregates should be dispersed as evenly as possible.
119
Gently pipetting 20 μL up and down can help disperse aggregates. Individual aggregates will allow for better identification, imaging and neurite quantification.
o 12. Incubate the plate 24 hours in the incubator at 37 C, 5 % CO2, 95% humidity.
13. Next day, remove 50 μL medium and add 50 μL of 4% PFA with Hoechst 33342
(1:5000) to all wells.
Plate must be handled with care to avoid disrupting neurites during fixation. When adding or removing solutions do so slowly and carefully. Don’t remove the medium or wash solution completely to avoid detachment of the aggregates from the plate.
14. Incubate the plate 1 hour at room temperature, protected from light.
15. Wash wells three times by removing 50 μL and adding 50 μL PBS to the corner of the wells. Each time incubate for 10 min.
16. Store plate at 4 oC protected from light.
17. For confocal High Content Analysis (HCA) follow steps as described by Termo
Schientific Cellomics® Neuronal Profiling V4 Bio Application Guide
(http://www.med.cam.ac.uk/wp- content/uploads/2016/02/NeuronalProfiling_V4_LC06190800.pdf).
18. Briefly, set a mask to detect nuclear stain, which will outline attached aggregates
(excluding neurites) and focus on object to obtain images (Figure 5).
19. Then, work with a second mask to identify individual neurites. Settings may need to be adapted for better neurite identification. ConFigure these settings by checking negative and positive controls (Figure 5).
Due to RFP expression in these cells, the aggregate “body” will be much brighter than individual neurites. ConFigure settings to ensure neurites are bright enough to be quantified. Overexposure of the aggregate bodies does not affect neurite measurements and is acceptable for this endpoint.
20. Once mask settings are conFigured, set machine to search and image each well upon aggregate detection.
120
Keep the settings the same for all samples within experiment.
21. Once images are obtained, apply masks to detect and quantify neurites.
22. Following manufacturer’s instructions, use data to obtain total neurite length or neurite number per aggregate.
REAGENTS AND SOLUTIONS
Use mili-Q-purified or double distilled and autoclaved water in all recipes and protocol steps.
Avoid repeated thawing/freezing for all medium supplements. Once defrosted store all medium supplements at 4 °C for no longer than one week.
Poly-L-Ornithine (PLO) (1mg/ml)
Dissolve 50 mg directly in the container in 5 mL of sterile double distilled water. Sterile filter, put this stock into a 50 ml Falcon, filled with 45 mL of sterile double distilled water. Aliquot. Store 5 mL aliquots at -20 °C.
Fibronectin (1mg/ml)
Aliquot. Store at 4°C.
N-2 supplement (100X)
Aliquot. Store at -20 °C.
L-Glutamine (200 mM)
Thaw in water bath, thoroughly mix, store aliquots of 5-7 ml in falcon tubes at -20 °C.
FGF (160 µg/ml)
Dissolve in 0.1% BSA/PBS. Sterile filter. Aliquot into 50 µL aliquots. Store at -20 °C.
GDNF (20 µg/ml)
Dissolve GDNF in 0.1% BSA/PBS and sterile filter. Aliquot into 20 µL aliquots. Store at -20 °C.
Tetracycline (2 mg/ml)
121
Weigh reasonable amount and dissolve in sterile double distilled water, sterile filter and aliquot into 500 µL aliqouots. Store in the dark at -20 °C. cAMP (100 mM)
Dissolve 500 mg in 5 mL of sterile double distilled water directly in the container; add into 50 mL Falcon filled with 5.2 mL of sterile double distilled water. Sterile filter. Aliquot 1mL in black eppendorf tubes. Store in the dark at -20 °C.
0.1% BSA/PBS
Dissolve 100mg BSA in 100 mL PBS.
Taxol (10 mM)
Weigh a reasonable amount (~15 mg) of Taxol and dissolve in the necessary volume of sterile
DMSO to reach 10 mM concentration. Place in water bath for 10 min and vortex until the powder has completely dissolved. Aliquot 5 μL of 10 mM stock into sterile 100 μL Eppendorf tubes.
Store at -20 °C for upto six months. Before use predilute the 10 mM stock 1:1000 in PBS
(working solution). Dilute the working solution 1:100 in differentiation medium to the final concentration 20 nM. Once defrosted discard the rest.
Rotenone (100 mM)
Weigh a reasonable amount (~ 80 mg) of rotenone powder and dissolve in the necessary volume of sterile DMSO to reach 100 mM concentration. Vortex until the powder has visibly dissolved.
Aliquot 5 μL of 100 mM stock into sterile 100 µL Eppendorf tubes Store at -20 °C for up to six months. Once defrosted, discard the remaining volume.
Matrigel
Defrost on ice and aliquot 125 µL matrigel in 0.5 mL Eppendorf tubes. Store at -20 °C for up to six months.
LUMES culture and differentiation media
122
See Table 2 for medium composition. Defrost components, which are stored at -20 oC. Always prepare medium fresh on the day of use. Prior to use, cell culture medium should be pre-warmed to 37 °C in a water bath.
Blocking solution, wash solution I and II, optical clearing solution
See Table 3 for clearing, blocking and wash solutions for immunocytochemistry and flow cytometry endpoints.
3.4. TABLES AND FIGURES
Table 3-1. Flask coating solution
Table 3-2. LUHMES medium composition
Components Proliferation Anti- Differentiation Wash Medium Proliferation Medium Medium (20 mL) Medium (50 mL) (100 mL) (100 mL)
Advanced 19.6 mL 49.5 mL 97 mL 98 mL DMEM/F12 Medium
N2 Supplement 200 µL 500 µL 1 mL 1 mL (100x)
Glutamine 200 µL 500 µL 1 mL 1 mL (1 mg/mL)
FGF (160 µg/mL) 5 µL cAMP (100 mM) 500 µL 1 mL
Tetracycline 50 µL 100 µL 100 µL (2 mg/mL)
123
GDNF (20 µg/mL) 5 uL 10 µL
Taxol (2 uM) 500 µL
Table 3-3. Solutions for immunocytochemistry and flow cytometry
Components PBS/EDTA Optical Blocking Wash Wash (1 mM) Clearing Solution Solution I Solution II Solution Solution (Hama et al., 2011)
PFA (4%) (R&D Systems)
Urea (8M) 10 mL (Sigma)
Triton 1 % 2 mL (Sigma)
Glycerol 2 mL (Sigma)
H2O 6 mL (Distilled)
PBS (1x) 50 mL 25 mL 100 mL 200 mL
Goat Serum 2.5 mL (Sigma)
Bovine 250 mg 1 g 2 g Serum Albumin (BSA) (Sigma)
Saponin 37.5 mg 300 mg (Sigma)
EDTA (0.5 100 uL M) (Sigma)
Table 3-4. Antibody dilutions for immunocytochemistry
Antibody Clone Catalogue # Host Dilution
124
Ki67 (H-300) Polyclonal sc15402 Rabbit 1:100 (Santa Cruz) MAP2a+b (Sigma) AP-20 M1406 Mouse 1:200 NeuN (Millipore) A60 MAB377 Mouse 1:200 NF200 (Sigma) Polyclonal N4142 Rabbit 1:200 SYP (Millipore) SY38 MAB5258 Mouse 1:200 Goat-anti-Mouse Polyclonal A-21202 Goat 1:500 Alexa 488 (ThermoFisher Scientific) Goat-anti-Rabbit Polyclonal A-11011 Goat 1:500 Alexa 568 (ThermoFisher Scientific)
Table 3-5. Marker genes for 3D differentiation quality control
Gene Symbol Gene Name TaqMan® Probe # Catalogue # TH Tyrosine Hydroxylase Hs00165941_m1 4331182 DAT Dopamine Transporter Hs00997374_m1 4331182 (SLC6A3) βIII-Tub Class III β-tubulin Hs00801390_s1 4331182 Syn1 Synapsin I Hs00199577_m1 4331182 NES Nestin Hs04187831_g1 4331182 MKI67 Antigen Ki-67 Hs01032443_m1 4331182 18S Eukaryotic 18S rRNA Hs99999901_s1 4331182
Figure 3-1. Suggested treatment schedule for sample collection on day 8. In this scheme, 48 hr treatment
takes place on day 6 of differentiation, while 24, 12, and 6 hr treatments take place on day 7 of
differentiation. All treated and control samples are collected at the same time on day 8, at the same stage
of differentiation.
125
Figure 3-2. Pipetting schematic for serial dilution of test compound.
Figure 3-3. Plate layout for concentration range finder experiments and dose-response curves. Vehicle control (DMSO) and compound concentrations are tested across four replicates, where C1 is the lowest and
C5 is the highest compound concentration.
126
Figure 3-4. Screenshot for ImageJ software with uploaded fluorescence image of an aggregate. Oval shape
selection is used to select aggregate area for measurement of mean grey values (using the commands
Analysis, Measure [Ctrl+M]).
Figure 3-5. High-content analysis (HCA) mask settings to identify aggregates (nuclear mask shown in blue
outlining bright aggregate) and individual neurite outgrowth (neurite mask shown in purple outlining
neurites) observed after 24 hr incubation in matrigel-coated wells.
127
Figure 3-6. Time consideration scheme for Basic Protocols 1 (LUHMES Neuronal Differentiation in 3D)
and 2 (Compound Treatment and Washout).
Figure 3-7. Time consideration scheme for the toxicological endpoints described in Support Protocols 1
(immunostaining), 2 (flow cytometry), 5 (ATP assay), 6 (MitoTracker), and 7 (neurite outgrowth
quantification).
128
3.5. DISCUSSION
A limited number of compounds have been tested for neurotoxicity. Current animal tests cost
$700,000 to $1 million per compound, are time consuming and have limited prediction for human health due to interspecies differences. At the same time evidence shows that exposure to certain pesticides or industrial emissions (metals/PCBs) can potentially increase the risk of neurodegenerative diseases (Chin-Chan et al., 2015; Hatcher-Martin et al., 2012; Tanner et al.,
2011; Wang et al., 2011). Chronic, non-communicable diseases such as Parkinson’s and
Alzheimer’s have been associated with early life or life-long exposures to compounds. The development of 3D organotypic models opens a new venture into studying tissue complexity in vitro (Ou and Hosseinkhani, 2014; Knight and Przyborski, 2015). Although in vitro assays are more time and cost effective than in vivo experiments, most assays to date have focused on acute effects of high-dose, exposures. Differentiation of neuronal cell types in monolayer cultures has presented challenges for repeat-dose chronic studies due to their limited shelf-life and easy destructible morphology due to experimental manipulations. In particular, monolayer cultures have limited applications for studies where delayed effects after compound withdrawal are of interest. Adherent neuronal cultures cannot be properly washed and transferred to a new culture plate which ensures complete compound removal. There is a need for models which allow for compound treatment and removal to study cellular adaptation, recovery and delayed (chronic) effects. Data from these experiments could be better correlated to animal studies and human population data. Dopaminergic cells comprise <1 % of neurons in the brain and are involved in motor function and the reward system. This neuronal type, which comprises the Substancia Nigra, is prone damage by oxidative stress. Compounds such as Paraquat, MPP+ and rotenone induce selective dopaminergic neurodegeneration (Langston et al., 1984; Gao et al., 2002; McCormack et al., 2002). The gene-environment interactions which increase susceptibility to dopaminergic neurodegeneration are obscure; therefore, a better understanding on how compounds in our
129 environment may affect this cell type is needed. Moreover, specifically low-dose, long-term effects must be studied in vitro. The 3D LUHMES model was developed to allow exposure and removal of compound to study reversible, delayed and permanent molecular changes. Although more complex models exist (Pamies et al., 2016; Lancaster et al., 2013), this model can isolate molecular and cellular changes pertinent to this neuronal cell type.
Critical Parameters and Troubleshooting: Basic Protocol 1 was developed for wild-type and
RFP-expressing 3D LUHMES. When establishing this protocol in a new laboratory, quality control of the differentiated 3D LUHMES must be carried out routinely (e.g. by RT-PCR of neural markers, flow cytometry, and immunocytochemistry). Especially, the level of proliferation after induction of differentiation should be monitored since proliferating cells may strongly affect the results of toxicological endpoints.
Compound penetration: Existing assays need to be modified for application to 3D cultures. In most cases, dye or antibody penetration will be slower; therefore incubation times need to be increased. In the same way, wash steps also need to be longer to allow for dissolution of the dye/antibody. Size and chemical properties of compounds influence the penetration rate and speed. Some compounds may lack the ability to penetrate the aggregate fully. The use of fluorescently labeled compounds can help visualize compound penetration. For example, Hoechst
33324 takes at least one hour to fully penetrate LUHMES aggregates (Smirnova et al., 2015a).
Cell number: At day 0 of induction of differentiation, careful attention must be paid to seeding: cell number per well should not exceed 6 x 105, with optimal cell count between 5-5.5 x 105 cells/well. With a higher number of cells per well, the effect of the anti-proliferative compound
Taxol is less efficient. This can lead to irregular aggregate shape and size as well as increased proliferation leading to a non-homogeneous cell population throughout differentiation. For reproducible results, cell number is a key step to ensure complete differentiation.
Wash steps: During wash steps, prolonged plate standing outside the incubator without shaking and aggregate exposure to air can lead to aggregate clumping or floating on the medium surface.
130
Enough medium must be left in the well to cover aggregates and steps must be performed quickly to reduce time plates are kept outside.
Optical clearing and antibody incubation
Due to the compactness of cells within aggregates and 3D tissue-like structure, incubation in optical clearing solution is required prior to antibody staining to optimize visualization of staining by confocal microscopy. This solution was published by (Hama et al., 2011) to clear mouse brain tissue for staining.
Assay endpoint normalization: If compounds decrease cell number due to cytotoxicity, this may be visible by phase contrast. Endpoints such as gene expression or protein expression should be normalized to a housekeeping gene or protein, respectively. Molecular endpoints such as ATP release, ROS production or GSH depletion should be normalized to cell number (can be measured by quantifying cell, protein or DNA content). Lack of normalization can lead to false positives or false negatives results.
Resazurin reduction assay, RNA isolation and flow cytometry: During 3D LUHMES differentiation, it sometimes occurs that a few larger aggregates (>300 µm diameter) form. These are visible by eye and should be removed from the plate before taking samples for RNA extraction or flow cytometry. However, these larger aggregates must remain in the plate if a cell number-dependent assay is being performed, such as resazurin reduction assay for cell viability after compound treatment.
Assay interference: When performing cell viability measurements (resazurin reduction, LDH release assays), pipette medium carefully from the surface to avoid the accidental transfer of the aggregates into measurement plate. Aggregates will strongly interfere with measurements and influence final results. When selecting other endpoints, one must take into consideration potential aggregate interference with assay and consult with the manufacturer on whether this may be an issue.
131
Anticipated Results: The expected acute, reversible and delayed effects (measured on day 8 and day 15) from 3D LUHMES exposure to rotenone following this protocol unit, are described in
(Smirnova et al., 2015a). As a known mitochondrial and dopaminergic toxicant, 24 hours exposure to rotenone decreases LUHMES cell viability (measured by resazurin reduction assay) at concentrations above 100 nM. Rotenone (100 nM, 24 h) is anticipated to decrease mitochondrial activity, ATP levels and neurite outgrowth acutely. Effects before and after compound removal (acute vs. chronic can be measured by gene expression profiling. When the compound is washed-out, some effects on gene expression are reversible while others are irreversible.
Time Considerations: Figures 6 and 7 summarize time frames needed to conduct the protocols described in this unit.
3.6. ACKNOWLEDGEMENTS
This research was supported by the International Foundation for Ethical Research (IFER) graduate fellowship and by the NIH NCATS (grant U18TR000547 "A 3D Model of Human Brain
Development for Studying Gene/Environment Interactions", PI Hartung). Authors would like to acknowledge Prof. Marcel Leist’s group (University of Konstanz, Konstanz, Germany). In particular, Prof. Leist, Dr. Krug, Dr. Schildknecht and Dr. Scholz for providing LUHMES cells and support in establishing the LUHMES culture in the laboratory at Johns Hopkins Bloomberg
School of Public Health. The video published with this paper was produced by Dr. Pamies (Johns
Hopkins University, Baltimore, USA). Authors would like to acknowledge Dr. Berlinicke for her support with HCA.
3.7. REFERENCES
Alépée, N., Bahinski, A., Daneshian, M., De Wever, B., Fritsche, E., Goldberg, A., Hansmann, J.,
132
Hartung, T., Haycock, J., Hogberg, H., et al. 2014. State-of-the-art of 3D cultures (organs-on-
a-chip) in safety testing and pathophysiology. Altex 31:441–477.
Chin-Chan, M., Navarro-Yepes, J., and Quintanilla-Vega, B. 2015. Environmental pollutants as
risk factors for neurodegenerative disorders: Alzheimer and Parkinson diseases. Frontiers in
Cellular Neuroscience 9:124.
Constantinescu, R., Constantinescu, A. T., Reichmann, H., and Janetzky, D. B. 2007. Neuronal
differentiation and long-term culture of the human neuroblastoma line SH-SY5Y. In
Neuropsychiatric Disorders An Integrative Approach Journal of Neural Transmission.
Supplementa. pp. 17–28. Springer Vienna, Vienna.
Efrémova, L., Schildknecht, S., Adam, M., Pape, R., Gutbier, S., Hanf, B., Bürkle, A., and leist,
M. 2015. Prevention of the degeneration of human dopaminergic neurons in an astrocyte co-
culture system allowing endogenous drug metabolism. British journal of pharmacology
172:n/a–n/a.
Gao, H.-M., Hong, J.-S., Zhang, W., and Liu, B. 2002. Distinct role for microglia in rotenone-
induced degeneration of dopaminergic neurons. Journal of Neuroscience 22:782–790.
Hama, H., Kurokawa, H., Kawano, H., Ando, R., Shimogori, T., Noda, H., Fukami, K., Sakaue-
Sawano, A., and Miyawaki, A. 2011. Scale: a chemical approach for fluorescence imaging
and reconstruction of transparent mouse brain. Nature neuroscience 14:1481–1488.
Hatcher-Martin, J. M., Gearing, M., Steenland, K., Levey, A. I., Miller, G. W., and Pennell, K. D.
2012. Association between polychlorinated biphenyls and Parkinson's disease
neuropathology. Neurotoxicology 33:1298–1304.
Ilieva, M., and Dufva, M. 2013. SOX2 and OCT4 mRNA-Expressing Cells, Detected by
Molecular Beacons, Localize to the Center of Neurospheres during Differentiation. PLoS
ONE 8:e73669.
Knight, E., and Przyborski, S. 2015. Advances in 3D cell culture technologies enabling tissue‐like
structures to be created in vitro. Journal of Anatomy 227:746–756.
133
Krug, A. K., Gutbier, S., Zhao, L., Pöltl, D., Kullmann, C., Ivanova, V., Förster, S., Jagtap, S.,
Meiser, J., Leparc, G., et al. 2014. Transcriptional and metabolic adaptation of human
neurons to the mitochondrial toxicant MPP(+). Cell death & disease 5:e1222–e1222.
Lancaster, M. A., Renner, M., Martin, C.-A., Wenzel, D., Bicknell, L. S., Hurles, M. E.,
Homfray, T., Penninger, J. M., Jackson, A. P., and Knoblich, J. A. 2013. Cerebral organoids
model human brain development and microcephaly. Nature 501:373–379.
Langston, J. W., Langston, E. B., and Irwin, I. 1984. MPTP-induced parkinsonism in human and
non-human primates--clinical and experimental aspects. Acta neurologica Scandinavica.
Supplementum 100:49–54.
Lotharius, J., Falsig, J., van Beek, J., Payne, S., Dringen, R., Brundin, P., and leist, M. 2005.
Progressive degeneration of human mesencephalic neuron-derived cells triggered by
dopamine-dependent oxidative stress is dependent on the mixed-lineage kinase pathway.
Journal of Neuroscience 25:6329–6342.
McCormack, A. L., Thiruchelvam, M., Manning-Bog, A. B., Thiffault, C., Langston, J. W., Cory-
Slechta, D. A., and Di Monte, D. A. 2002. Environmental risk factors and Parkinson's
disease: selective degeneration of nigral dopaminergic neurons caused by the herbicide
paraquat. Neurobiology of Disease 10:119–127.
Mehta, G., Hsiao, A. Y., Ingram, M., Luker, G. D., and Takayama, S. 2012. Opportunities and
challenges for use of tumor spheroids as models to test drug delivery and efficacy. Journal of
Controlled Release 164:192–204.
Noelker, C., Lu, L., Höllerhage, M., Vulinovic, F., Sturn, A., Roscher, R., Höglinger, G. U.,
Hirsch, E. C., Oertel, W. H., Alvarez-Fischer, D., et al. 2015. Glucocerebrosidase deficiency
and mitochondrial impairment in experimental Parkinson disease. Journal of the
Neurological Sciences 356:129–136.
Oliveira, L. M. A., Falomir-Lockhart, L. J., Botelho, M. G., Lin, K.-H., Wales, P., Koch, J. C.,
Gerhardt, E., Taschenberger, H., Outeiro, T. F., Lingor, P., et al. 2015. Elevated α-synuclein
134
caused by SNCA gene triplication impairs neuronal differentiation and maturation in
Parkinson's patient-derived induced pluripotent stem cells. Cell death & disease 6:e1994–
e1994.
Ou, K.-L., and Hosseinkhani, H. 2014. Development of 3D in vitro Technology for Medical
Applications. International Journal of Molecular Sciences 15:17938–17962.
Pamies, D., Barreras, P., Block, K., Makri, G., Kumar, A., Wiersma, D., Smirnova, L., Zhang, C.,
Bressler, J., Christian, K. M., et al. 2016. A human brain microphysiological system derived
from induced pluripotent stem cells to study neurological diseases and toxicity. Altex.
Pampaloni, F., Reynaud, E. G., and Stelzer, E. H. K. 2007. The third dimension bridges the gap
between cell culture and live tissue. Nature Reviews Molecular Cell Biology 8:839–845.
Pöltl, D., Schildknecht, S., Karreman, C., and leist, M. 2012. Uncoupling of ATP-depletion and
cell death in human dopaminergic neurons. Neurotoxicology 33:769–779.
Schildknecht, S., Karreman, C., Pöltl, D., Efrémova, L., Kullmann, C., Gutbier, S., Krug, A.,
Scholz, D., Gerding, H. R., and leist, M. 2013. Generation of genetically-modified human
differentiated cells for toxicological tests and the study of neurodegenerative diseases. Altex
30:427–444.
Scholz, D., Pöltl, D., Genewsky, A., Weng, M., Waldmann, T., Schildknecht, S., and leist, M.
2011. Rapid, complete and large-scale generation of post-mitotic neurons from the human
LUHMES cell line. J Neurochem 119:957–971.
Smirnova, L., Harris, G., Delp, J., Valadares, M., Pamies, D., Hogberg, H. T., Waldmann, T.,
Leist, M., and Hartung, T. 2015a. A LUHMES 3D dopaminergic neuronal model for
neurotoxicity testing allowing long-term exposure and cellular resilience analysis. Archives of
Toxicology:1–19.
Smirnova, L., Seiler, A. E. M., and Luch, A. 2015b. microRNA Profiling as Tool for
Developmental Neurotoxicity Testing (DNT). Current protocols in toxicology / editorial
board, Mahin D. Maines (editor-in-chief) ... [et al.] 64:20.9.1–20.9.22.
135
Tanner, C. M., Kamel, F., Ross, G. W., Hoppin, J. A., Goldman, S. M., Korell, M., Marras, C.,
Bhudhikanok, G. S., Kasten, M., Chade, A. R., et al. 2011. Rotenone, paraquat, and
Parkinson's disease. Environ Health Perspect 119:866–872.
Tong, Z. B., Hogberg, H., Kuo, D., Sakamuru, S., Xia, M., Smirnova, L., Hartung, T., and
Gerhold, D. 2016. Characterization of three human cell line models for high‐throughput
neuronal cytotoxicity screening. Journal of applied toxicology : JAT.
Wang, A., Costello, S., Cockburn, M., Zhang, X., Bronstein, J., and Ritz, B. 2011. Parkinson’s
disease risk from ambient exposure to pesticides. Eur J Epidemiol 26:547–555.
Zanoni, M., Piccinini, F., Arienti, C., Zamagni, A., Santi, S., Polico, R., Bevilacqua, A., and
Tesei, A. 2016. 3D tumor spheroid models for in vitro therapeutic screening: a systematic
approach to enhance the biological relevance of data obtained. Scientific reports 6:19103.
Zhang, X.-M., Yin, M., and Zhang, M.-H. 2014. Cell-based assays for Parkinson's disease using
differentiated human LUHMES cells. Acta Pharmacologica Sinica 35:945–956.
INTERNET RESOURCES
Technical data sheet for PE Annexin V Apoptosis Detection Kit I (BD PharmingenTM)
http://www.bdbiosciences.com/ds/pm/tds/559763.pdf
Taqman® Gene Expression Assays Protocol (Applied Biosystems)
http://www3.appliedbiosystems.com/cms/groups/mcb_support/documents/generaldocuments/
cms_041280.pdf
Product information for ATP Determination Kit (Molecular ProbesTM)
https://tools.thermofisher.com/content/sfs/manuals/mp22066.pdf
Pierce BCA Protein Assay Kit instructions (ThermoScientific)
https://tools.thermofisher.com/content/sfs/manuals/MAN0011430_Pierce_BCA_Protein_Asy
_UG.pdf
How to use a protein assay standard curve (ThermoScientific)
https://tools.thermofisher.com/content/sfs/brochures/TR0057-Read-std-curves.pdf
136
Thermo Scientific Cellomics ® Neuronal Profiling V4 BioApplication Guide
http://www.med.cam.ac.uk/wp-
content/uploads/2016/02/NeuronalProfiling_V4_LC06190800.pdf
Image Processing and Analysis in Java (ImageJ) website https://imagej.nih.gov/ij/
MitoTracker® Mitochondrion-Selective Probes (Molecular Probes®) manual
https://tools.thermofisher.com/content/sfs/manuals/mp07510.pdf
137
CHAPTER 4
4. LUHMES 3D DOPAMINERGIC NEURONAL MODEL FOR
NEUROTOXICITY TESTING ALLOWING LONG-TERM EXPOSURE
AND CELLULAR RESILIENCE ANALYSIS.
*The work presented in this chapter is published in the following article:
Smirnova L, Harris G, Delp J, Valadares M, Pamies D, Hogberg HT, Waldmann T, Leist
M, Hartung T. LUHMES 3D dopaminergic neuronal model for neurotoxicity testing allowing long-term exposure and cellular resilience analysis. Arch Toxicol. 2016.
90(11):2725-2743.
138
Key points
• 3D LUHMES express the same levels of mRNA and protein differentiation markers as
early as on day 3 of differentiation as 2D culture, demonstrated by qPCR and
immunostaining.
• LUHMES can be cultures for over 21 days in 3D, compared to less than 12 days in 2D.
• Penetration by molecules to the center of the aggregates occurs within 1 hour, shown
with Hoechst staining.
• Proliferation and apoptosis remains low in aggregates.
• Mitochondrial toxicants rotenone and MPP+ show effects on cell viability,
mitochondrial membrane potential and perturb genes involved in transsulfuration
pathways.
• Rotenone down-regulates enriched miR-7 expression (enriched in dopaminergic neurons)
after 12 hours of exposure. Mir-7 expression returns to control levels after compound
withdrawal and 7 day recovery period.
• Reversibility of gene expression alterations after rotenone wash-out indicate early responses differ from long-term mechanisms.
139
4.1. ABSTRACT
Several shortcomings of current Parkinson’s disease (PD) models limit progress in identification of environmental contributions to disease pathogenesis. The conditionally immortalized cell line
LUHMES promises to make human dopaminergic neuronal cultures more easily available, but these cells are difficult to culture for extended periods of time. We overcame this problem by culturing them in 3D with minor medium modifications. The 3D neuronal aggregates allowed penetration by small molecules and sufficient oxygen and nutrient supply for survival of the innermost cells. Using confocal microscopy, gene expression, and flow cytometry, we characterized the 3D model and observed a highly reproducible differentiation process.
Visualization and quantification of neurites in aggregates was achieved by adding 2 % red fluorescent protein-transfected LUHMES cells. The mitochondrial toxicants and established experimental PD agents, rotenone and MPP+, perturbed genes involved in one-carbon metabolism and trans-sulfuration pathways (ASS1, CTH, and SHTM2) as in 2D cultures. We showed, for the first time in LUHMES, down-regulation of mir-7, a miRNA known to target alpha-synuclein and to be involved in PD. This was observed as early as 12 h after rotenone exposure, when pro-apoptotic mir-16 and rotenone-sensitive mir- 210 were not yet significantly perturbed. Finally, washout experiments demonstrated that withdrawal of rotenone led to counter- regulation of mir-7 and ASS1, CTH, and SHTM2 genes. This suggests a possible role of these genes in direct cellular response to the toxicant, and the model appears to be suitable to address the processes of resilience and recovery in neurotoxicology and Parkinson’s disease in future studies.
140
4.2. INTRODUCTION
Progress in the study of toxicant-induced dopaminergic neurodegeneration suffers from the shortcomings of available in vivo and in vitro models, since they usually do not progress in the same way human disease does. Moreover, important aspects such as endogenous counter- regulations and recovery are particularly difficult to address in vitro. Although dopaminergic neurons correspond to <1 % of all neurons in the brain, they play a crucial role in movement, sensation of pleasure, motivation, reward, and drug addiction; are particularly sensitive to oxidative stress; and are involved in the second most common neurodegenerative disease—
Parkinson’s disease (reviewed in Chinta and Andersen 2005). This neuronal cell type, therefore, is of particular interest for understanding the molecular mechanisms of PD—which are key to the development of preventive and disease-modifying therapies. Although genetic forms of PD associated with mutations in genes for alpha-synuclein, PARKIN, PINK1, LRRK2 or DJ-1 are well established (reviewed in Henchcliffe and Beal 2008), increasing evidence suggests a role for gene–environmental interactions contributing to the sporadic form of the disease, and gene regulatory networks are being unraveled (Kumar Singh et al. 2014; Krug et al. 2014; Fujita et al.
2014; Todorovic et al. 2014; Maertens et al. 2015; Lee and Cannon 2015; Rahnenführer and Leist
2015). Exposure to pesticides such as rotenone may be associated with increased risk of PD
(Ascherio et al. 2006; Costello et al. 2009; Wang et al. 2011; Tanner et al. 2011). Mitochondrial dysfunctions (e.g., toxicant-induced mitochondrial complex I inhibition) are believed to be central in the pathophysiology of PD (reviewed in Franco-Iborra et al. 2015); however, it is not clear whether this is a primary or secondary event in PD pathogenesis. In addition, it is not clear yet why dopaminergic neurons are more vulnerable to mitochondrial complex I inhibition and degeneration. Thus, cellular responses to environmental stress and molecular perturbations upon toxicant insult leading to neurodegeneration need to be elucidated further. There are a multitude of neuronal models for studying Parkinson’s disease and neurotoxicology. These include (a) cell
141 lines, such as rat PC12 cell line (Greene and Tischler 1976; Grau and Greene 2012), SH-SY5Y neuroblastoma cell line (Constantinescu et al. 2007; Borland et al. 2008), and LUHMES cell line
(Lotharius et al. 2005; Zhang et al. 2014; Ste˛pkowski et al. 2015); (b) primary cell cultures
(Lingor et al. 1999); and (c) embryonic stem cell (ESC)- or induced pluripotent stem cell (iPSC)- derived neurons (Srikanth and Young-Pearse 2014). Different types of neuronal models have strengths and limitations (Schlachetzki et al. 2012). The PC12 cell line, for instance, is easy to handle and relatively homogeneous, but is not of human origin—making it difficult to extrapolate interspecies differences in response to toxicant treatment. SH-SY5Y is a human cell line, fast dividing but difficult to differentiate into postmitotic neurons (Constantinescu et al. 2007) and has limitations as a cancer cell line because of its relatively unstable genome. Primary rat midbrain cell cultures were established to study PD (Lingor et al. 1999), but, again, with an obstacle of interspecies differences. Primary postmortem tissues, isolated from brains of patients with PD, more closely reflect the pathogenesis of the disease, but are difficult to obtain, are already affected by the disease, and, therefore, unsuitable for studying the dynamics of pathogenesis.
Another limitation of primary cell cultures isolated from mesencephalon is low yield of biological material. In contrast to these primary or cancer tissue-originated cell models, the Lund human mesencephalic (LUHMES) cell line originates from healthy human 8-week-old embryonic mesencephalic tissue, immortalized by tetracycline regulated v-myc-vector transfection, and can be rapidly and homogeneously differentiated to mature dopaminergic neurons by cultivation in the presence of tetracycline, cAMP, and GDNF (Lotharius et al. 2005; Scholz et al. 2011).
Healthy donor origin of LUHMES makes these cells an attractive tool for studying the effects of environmental exposures leading to mitochondrial dysfunctions and/or to a PD-like phenotype.
Human ESC and iPSC applications to study neurotoxicity and neurodegeneration are emerging and are very promising (Wheeler et al. 2015; Srikanth and Young-Pearse 2014), but the differentiation protocols are demanding, cost-prohibitive, and lengthy (up to 4/8 weeks of differentiation), resulting in heterogeneous neuronal populations. Protocols for enrichment in one
142 cell type during differentiation, however, are evolving (Hu et al. 2015). Choosing one neural model over another depends on the hypotheses, specific aims, and study design. Although we believe iPSC-derived neural models will provide a broader range of options in future neurotoxicological studies, we also need models allowing easy access to homogenous cell populations for mechanistic studies. In this study, we favored the fast-differentiating and homogenous LUHMES cell line for the study of early cellular perturbations and cellular adaptation after toxicant exposure as a promising PD model. More than 50 % of known miRNAs are expressed in the brain (Li and Jin 2010) where they posttranscriptionally regulate gene expression and play important roles in cellular homeostasis, metabolism, proliferation, differentiation, and apoptosis. Elimination of all miRNAs results in loss of the dopaminergic neurons in conditional knockout animal models and stem cells (Giraldez et al. 2005; Kim et al.
2007; Huang et al. 2010). Several miRNAs were shown to regulate function of dopaminergic neurons mir-133b (Kim et al. 2007), mir-9 (Leucht et al. 2008), and mir-132 (Yang et al. 2012).
These and other dopaminergic neuronspecific miRNA were deficient in PD-affected midbrains
(Kim et al. 2002; Lau and de Strooper 2010; Mouradian 2012). mir-7, which is expressed in nigral neurons in mice and humans, was shown to target α-synuclein and is downregulated in
MPP+ PD animal models (Junn et al. 2009). The neuroprotective role of mir-7 against MPP+- induced neuronal death has been suggested (Fragkouli and Doxakis 2014; Choi et al. 2014). Thus, miRNAs may play an important role in cellular responses to toxicant exposure and PD development. An increasing number of studies are addressing whether miRNAs are involved in cellular responses to environmental stress (reviewed in Smirnova et al. 2012), and the role of miRNAs in neurotoxicity is being elucidated (Huang and Li 2009; Miranda et al. 2010; Saba et al.
2012; Tal and Tanguay 2012; Pallocca et al. 2013; Smirnova et al. 2014). Moreover, miRNAs play a significant role in mitochondrial function (Li et al. 2012), including pro-apoptotic mir-15, mir-16, and anti-apoptotic mir-21, mir-17-92 cluster. ROS-responsive and hypoxia-related mir-
210 inhibits cell proliferation and represses the mitochondrial metabolism and respiration by
143 targeting several elements of the TCA cycle (Chan et al. 2012). mir-210, together with mir- 195, was shown to be regulated by rotenone exposure (Kim et al. 2013). Most importantly, there are mitochondriaenriched miRNAs, mito-miRs (Bandiera et al. 2013), which may be crucial in managing the mitochondrial response to toxicant-induced stress. This work is based on our collaborative studies on the early responses of LUHMES cells to 1-methyl-4-phenylpyridinium
(MPP+), the metabolite of 1-methyl-4-phenyl- 1,2,3,6-tetrahydropyridine (MPTP), a common agent to experimentally induce PD (Krug et al. 2014), where perturbations and counter- regulations in the cells before mitochondrial dysfunction-initiated apoptosis were characterized.
Based on data obtained from metabolomics and transcriptomics analysis, we proposed a network of toxicant (mitochondrial complex I inhibitor)-induced adaptations in human dopaminergic neurons before a tipping point is reached that allows execution of the apoptosis program. The next question is, are these early changes permanent, or can they be reversed after compound withdrawal? To experimentally test this, we need test systems that can be maintained for longer periods of time and interrogated for cellular perturbations, their counter-regulations, and ability to return to physiological conditions. LUHMES cells, which proved so promising in our earlier studies, do not allow such analysis in standard monolayer culture because differentiated cells tend to detach from the culture dish after about 9–12 days (depending on density, medium, and surface structuring). 3D cell cultures aiming to approximate organotypic cultures are rapidly emerging
(Alépée et al. 2014; Hartung 2014), and they promise tissue-like cell density and cell/cell contacts. Therefore, for the first time, we have adapted the LUHMES cell culture to 3D using constant gyratory shaking. In the first step of the current study, we adapted the 2D protocol to generate a 3D dopaminergic neuronal model and found this prolonged survival of the differentiated cells. After the 3D protocol optimization and characterization steps, we demonstrated the suitability of the 3D model for neurotoxicity testing by using two model compounds, rotenone and MPP+. Finally, we analyzed the expression of miRNAs known to be involved in PD pathophysiology and mitochondrial function. We found that mir-7 was sensitive
144 to rotenone treatment and that its expression recovered after rotenone withdrawal, while MPP+ and rotenone-responsive genes TYMS and MLF1IP, identified previously in Krug et al. (2014), were further down-regulated with time. We demonstrated that the 3D LUHMES model is suitable for studying cellular responses after toxicant withdrawal and ultimately cellular resilience— which has previously not been possible in 2D because of short survival of these cultures.
145
4.3. MATERIALS AND METHODS
LUHMES maintenance and differentiation
Proliferation medium was prepared with Advanced DMEM/F12 (Gibco, Life Technologies) supplemented with 2 mM l-Glutamine (Sigma-Aldrich), 1× N2 (Gibco) and 0.04 μg/ml recombinant basic fibroblast growth factor (bFGF, R&D Systems). Differentiation medium was prepared with Advanced DMEM/F12 containing 2 mM l-Glutamine, 1× N2 supplement, 1 mM dibutyryl cAMP (Santa Cruz), 2 μg/ml tetracycline (Sigma-Aldrich), and 2 ng/ml recombinant human glial cell line-derived neurotrophic factor (GDNF, Gemini). For propagation and differentiation in monolayer, all flasks and plates were pre-coated with 50 μg/ml poly-l-ornithine
(PLO) and 1 μg/ml fibronectin for 12 h prior to the experiment. Wild-type LUHMES (ATCC®
CRL_2927™) human neuronal precursor cells, as well as genetically modified LUHMES, ubiquitously expressing red or green fluorescent protein (RFP/GFP) were cultured as described previously (Scholz et al. 2011). RFP- and GFP-expressing cell lines were generated earlier as described in Schildknecht et al. (2013). Briefly, the conditionally immortalized cells (v-myc transgene expressing, controlled by a tet-off system) were maintained in proliferation medium in
PLO–fibronectin pre-coated Nunclon™ (Nunc) flasks and passaged every 2–3 days. For differentiation in 2D (Fig. 1a, 2D diff protocol), cells were seeded in a pre-coated 175-cm2 flask
(Nunc) in proliferation medium. After 24 h, medium was changed to differentiation medium.
After 48 h in differentiation medium, cells were trypsinized using TripleE Express (Life
Technologies) and seeded at a density of 5 × 105 cells per 2 ml per well in pre-coated 6-well plates (Nunc™ Cell culture-treated). Medium was exchanged every second day. To induce neuronal differentiation in 3D, LUHMES progenitors were trypsinized using TryplE Express, centrifuged, and resuspended in differentiation medium. Cells were seeded in 6-well plates
(Falcon®) at 5 × 105 cells/well in 2 ml of differentiation medium and placed on a gyratory shaker
146
(ES-X, Kuhner shaker) at 80 rpm in a humidified incubator (37 °C, 10 % CO2). Medium was exchanged every third day by removing 1.2 ml and adding 1.5 ml of fresh medium (Fig. 1a, 3D diff protocol). Following protocol 3D + T10 (Fig. 1a) on day 3 of differentiation, aggregates were treated with 10 nM taxol (>97 % paclitaxel, T7191 Sigma-Aldrich), an anti-proliferation compound, for 48 h. On day 5, aggregates were washed directly in the well with washing medium
(supplemented with l-glutamine, N2, and tetracycline), and differentiation medium was added to continue cellular differentiation. For co-cultures of wild-type LUHMES with RFP- or GFP- expressing cells, 1 × 104 RFP or GFP cells were seeded with 5 × 105 cells per 2 ml per well in pre-coated 6-well plates (Nunc™ Cell culture-treated). Medium was exchanged every second day.
To induce neuronal differentiation in 3D, LUHMES progenitors were trypsinized using TryplE
Express, centrifuged, and resuspended in differentiation medium. Cells were seeded in 6-well plates (Falcon®) at 5 × 105 cells/well in 2 ml of differentiation medium and placed on a gyratory shaker (ES-X, Kuhner shaker) at 80 rpm in a humidified incubator (37 °C, 10 % CO2). Medium was exchanged every third day by removing 1.2 ml and adding 1.5 ml of fresh medium (Fig. 1a,
3D diff protocol). Following protocol 3D + T10 (Fig. 1a) on day 3 of differentiation, aggregates were treated with 10 nM taxol (>97 % paclitaxel, T7191 Sigma-Aldrich), an anti-proliferation compound, for 48 h. On day 5, aggregates were washed directly in the well with washing medium
(supplemented with l-glutamine, N2, and tetracycline), and differentiation medium was added to continue cellular differentiation. For co-cultures of wild-type LUHMES with RFP- or GFP- expressing cells, 1 × 104 RFP or GFP cells were seeded with 5 × 105 LUHMES-WT cells to comprise 2 % of the total cell population. Upon seeding, cells were differentiated as described above on a gyratory shaker at 80 rpm in a humidified incubator (37 °C, 10 % CO2).
147
Figure 4-1. Adaptation and optimization of 3D LUHMES differentiation protocol. (a) The original 2D differentiation (2D diff protocol) was adapted for 3D culture (3D diff protocol) by subjecting the single-cell suspension to continuous gyratory shaking. Protocol 3D pre-diff involves a pre-differentiation step in 2D for 2 days, trypsinization, and subsequent cultivation in 3D. Further optimization involved adding the anti- proliferation compound taxol on day 3 for 48 h (10 nM, 48 h) to reduce cell proliferation (3D + T10 protocol). Toxic compound treatment took place between days 6 and 8 for 12, 24, and 48 h, reversely.
148
Samples for toxicological end points were collected immediately after treatment on day 8 or on day 15 after washout of compounds and 7 days recovery. (b) Differentiation of LUHMES in monolayer. As differentiation advances, cells may detach from the surface (d9, last photograph). (c) Aggregate formation under continuous gyratory shaking (80 rpm). Note the increasing size of aggregates in course of differentiation. (d) Treatment of 3D cultures with anti-proliferation drug taxol (10 nM) for 48 h blocked the proliferation and slowed down aggregate growth. (e) Measurements of aggregate size during differentiation following different 3D protocols. At least 100 aggregates were measured per condition, per day, in at least three independent experiments, with exception of 3D pre-diff. and 3D gradient speed, where only one independent experiment was conducted. Data represent mean ± SEM, n ≥ 3 (independent experiments).
Scale bars are 200 μm.
Chemical preparation, storage, and treatment
A 2 mM taxol stock was prepared in 100 % DMSO (D2650, Sigma-Aldrich), aliquoted, and
stored at −20 °C. A 10 mM MPP+ iodite (>98 % HPLC powder, D048 Sigma-Aldrich) stock was
prepared in sterile H2O, aliquoted, and stored at −20 °C protected from light. A 100 mM rotenone
(>95 % powder, 84–79–4 Sigma-Aldrich) stock was prepared in 100 % DMSO, aliquoted, and
stored at −20 °C protected from light. Once defrosted, aliquots were used immediately and
discarded after use.
On day 3 of differentiation, aggregates were treated with taxol (10 nM, for 48 h) to reduce the
number of proliferating cells. Test compounds (MPP+ and rotenone) were added to cultures on
days 6, 7, or 7.5 for 48, 24, and 12 h, respectively. DMSO was used as a solvent control. End
concentration of DMSO used in the cultures was ≤0.001 %, which did not have any effects on
cell viability. Samples for toxicological end points were collected on day 8. To study cellular
recovery, the toxicants were washed out on day 8. Cells were washed once with medium directly
in the plate (supplemented with l-Glutamine, N2, and tetracycline) and plated in a new 6-well
plate with fresh differentiation medium. Medium was exchanged on days 10 and 12. On day 15,
149 aggregates were collected to measure end points relevant to cellular recovery from the effects of rotenone or MPP+ exposure.
Cell viability
Cell viability (mitochondria activity) was analyzed using the resazurin reduction assay. A 1 mg/ml resazurin sodium salt (Sigma-Aldrich) stock was prepared in PBS. 200 μl of 1 mg/ml stock was added to each well (2 ml medium), and plates were kept on a gyratory shaker at 80 rpm in a humidified incubator (37 °C, 10 % CO2) for 1.5 h. 100 μl of samples was transferred from each well in triplicates into 96-well plates, and fluorescence was measured in a fluorescence plate reader (530 nm excitation/590 nm emission). Differentiation medium was incubated with resazurin in parallel as a blank control. Cell viability (mitochondria activity) was calculated as % of fluorescence intensity relative to solvent-treated controls after subtracting blanks in three biological replicates.
Cytotoxicity (membrane integrity) was analyzed using the LDH release assay (Promega). As a positive control, aggregates were treated with 1 % Triton X100. 50 μl of medium from each well was transferred to a 96-well plate. 50 μl of LDH substrate was added to each well. The plate was incubated for 30 min at room temperature in the dark. Differentiation medium was incubated with substrate in parallel as a blank control. The reaction was stopped with 50 μl stop solution.
Absorbance was recorded at 490 nm. After subtracting blanks, cytotoxicity (%) was determined by normalization of ODs from the test sample to positive and solvent-treated controls.
Mitochondrial activity
Mitochondrial activity after rotenone treatment was measured using the red fluorescent dye
Mitotracker® Red CMXRos (Life Technologies) following the manufacturer’s instructions.
Briefly, after 48-h rotenone exposure, aggregates were transferred into 24-well plates and incubated with 1 μM Mitotracker® Red on a gyratory shaker at 80 rpm at 37 °C, 10 % CO2 for
150
30 min. Aggregates were washed twice with PBS and fixed with 4 % PFA for 20 min at 4 °C.
Fixed aggregates were washed twice in PBS and mounted on slides for fluorescence imaging
(excitation 579 nm/emission 599 nm). In every independent experiment, at least 10 single aggregates were imaged per condition, and the mean gray value (sum of the gray values of all the pixels in the area divided by the number of pixels in the area) per aggregate was quantified using the open-source ImageJ software (http://imagej.nih.gov/ij/) and normalized to solvent-treated controls. Average of normalized mean gray values ± SEM was calculated from at least three biological replicates. Differences in treated and control samples were analyzed for statistical significance using Kruskal–Wallis test followed by Dunn’s Multiple comparison test. p value is denoted on graphs by ***.
Size measurements
Aggregates were cultured as described above, and phase-contrast microscopic images were taken on days 3, 6, 9, 12, 15, and 21 of differentiation. The diameter of 20–50 aggregates was measured on each day using SPOT software 5.0 (Diagnostic Instruments, Inc.). Experiments were repeated at least three times.
Flow cytometry
2D and 3D differentiated, as well as undifferentiated, LUHMES cells were trypsinized directly in the plate on the shaker with TryplE Express containing 4 units/ml DNAse at 37 °C for 30 min.
After 30 min, aggregates were homogenized by aspiration using a 1-ml syringe with 26G3/8 needle. Single-cell suspensions in 2D were obtained by pipetting up and down several times. The previously described protocol (Smirnova et al. 2015a) was followed for further steps. Briefly, cells were fixed with 2 % PFA and stained with Alexa Flour 647 mouse antihuman Ki-67 antibody (1:20, clone B56, BD Pharmingen™) at 4 °C for 1 h in PBS/1 %BSA/0.15 % saponin/10 % goat serum. Ki-67 expression was quantified using a FACSCalibur flow cytometer
151
(BD). Unstained cells, as well as cells stained with Alexa Flour 647 mouse IgG1 κ Isotype control, were used to set the gates for Ki-67-negative cells.
Immunocytochemistry and confocal imaging
For immunostaining, aggregates were collected on days 6, 12, 15, or 21 of differentiation, and immunocytochemistry was performed as previously described (Smirnova et al. 2015a) with some adaptations for 3D cultures. Briefly, aggregates were fixed with 4 % PFA (20 min, 4 °C) and blocked with blocking solution [10 % goat serum (Sigma), 1 % BSA (Sigma), 0.15 % saponin in
PBS] for at least 1 h on a shaker at 4 °C. Subsequently, aggregates were incubated with primary antibodies diluted in blocking solution for 48 h at 4 °C on a shaker. The following antibodies were used; they are as follows: rabbit antihuman neurofilament (NF200, 1:200, Sigma); mouse antihuman MAP2 (1:100, Sigma); mouse antihuman NeuN (1:100, Millipore); mouse antihuman synaptophysin (1:200, Millipore); rabbit antihuman Ki-67 (1:100, Santa Cruz). As negative control, aggregates were incubated with blocking solution. Then, aggregates were washed twice and incubated for 24 h at 4 °C on a shaker in the dark with secondary antibodies (goat antimouse
IgG Alexa Fluor® 488 and goat antirabbit IgG Alexa Fluor® 568, 1:500, Life Technologies).
After 24 h, aggregates were washed, and nuclei were stained with Hoechst 33342 (1 μg/ml,
Invitrogen, Molecular Probes) for at least 1 h at RT. Aggregates were then mounted using mounting medium (Immu-Mount™, Thermo Scientific) on glass slides (Fisherbrand® Thermo
Scientific) and analyzed using the Zeiss LSM 510 Confocal III confocal microscope (Zeiss) and
ZEN Imaging software (Zeiss). For dye penetration assays, LUHMES ubiquitously expressing
GFP were used. GFP-LUHMES were differentiated in 3D for 12 days following 3D diff protocol
(Fig. 1a). On day 12 of differentiation, Hoechst 33342 (1 μg/ml) was added to the cultures for 5,
15, 30, 60 min, or 6 h. Aggregates were then fixed with 4 % PFA, incubated with optical clearing solution (ScaleA2: 4 M urea, 10 % wt/vol glycerol, 0.1 % wt/vol Triton-X-100, pH 7.7 (Hama et
152 al. 2011) for further 48 h at 4 °C on a shaker and mounted on the glass slides. Hoechst 33342 penetration through the aggregates was analyzed using Zeiss LSM 510 confocal III microscope.
Neurite integrity quantification
To assay neurite integrity, RFP-expressing LUHMES were mixed with wild-type LUHMES in a ratio 1:49. Cells were differentiated following 3D + T10 protocol. On day 6, cells were exposed to 0.1 μM rotenone or DMSO as control. After 48-h exposure, aggregates were collected, fixed with 4 % PFA, and nuclei were stained with 1 μg/ml Hoechst 33342. Then, aggregates were incubated with ScaleA2 solution for further 48 h at 4 °C on a shaker and mounted on glass slides for imaging by confocal microscopy. The hyperstack images were analyzed using the software
KNIME with the Image Processing plugin. The program provides algorithms and means for image (pre) processing, filtering, segmentation, and feature calculation. The provided algorithms are well known and common image analysis techniques that can be combined to fit arbitrary analysis problems. In order to quantify the neurites, edge detection was applied to the images to highlight the neurites. Then, the neurite areas as well as the cell bodies were segmented. After cell bodies were excluded from the analysis, the area of the identified neurites for every slice of the hyperstack was calculated and normalized to the number of red cells in the aggregate. Lastly, the median neurite area within the hyperstack was calculated in three biological replicates (mean
± SEM). Differences in rotenone-treated and DMSO-treated samples were analyzed for statistical significance using the Mann–Whitney test; p value <0.01 is denoted on the graphs by **.
Apoptosis assays
Two assays were used to estimate the level of apoptosis during 3D differentiation (3D diff protocol). The number of apoptotic/necrotic cells was quantified using PE Annexin V Apoptosis
Detection kit I (BD Pharmingen™) by flow cytometry and visualized using the Cell Event®
153
Caspase-3/7 Green Detection Reagent (Life technologies) by confocal fluorescent microscopy.
For Annexin V-PE/7AAD staining, aggregates were trypsinized for 30 min with TryplE Express, washed once with PBS, and resuspended in 1× Annexin V binding buffer. 4 × 105 cells were stained in 100 μl with 5 μl Annexin V-PE and 5 μl 7-AAD for 15 min at room temperature protected from light. Unstained cells and cells stained either with Annexin V or 7-AAD were used to set the gates. Cells treated with 0.5 μM rotenone were used as positive control. The percentage of Annexin V-positive cells, 7-AAD-positive cells, and double-positive cells were quantified using a FACSCalibur flow cytometer (BD). For caspase 3/7 staining, upon addition of the Cell Event reagent and 1 μg/ml Hoechst 33342, cells were incubated at 37 °C for 60 min, washed once with PBS, and fixed with 4 % PFA for 20 min at 4 °C on a shaker. Fixed aggregates were washed twice with PBS and mounted on slides for fluorescence imaging using Zeiss LSM
510 confocal III microscope. Undifferentiated LUHMES were used as negative control in both experiments.
RNA extraction, Reverse Transcription, and Real-Time PCR
Total RNA was extracted using either TRIzol® Reagent (Life Technologies) or the miRVANA miRNA isolation kit (Ambion, Life Technologies) following the manufacturer’s instructions.
RNA integrity was measured using the Nanodrop 2000 (TermoScientific) UV–Vis
Spectrophotometer (260 nm). Equal amounts of purified RNA (500 ng) were reverse transcribed to cDNA using random hexamer primers (Promega) and M-MLV reverse transcriptase Kit
(Promega) following the manufacturer’s instructions. A DNAse treatment step was included in cDNA synthesis to ensure the elimination of DNA traces. cDNA was diluted 1:5, and qRT-PCR was performed. Expression of neuronal markers during LUHMES differentiation was analyzed using TaqMan gene expression assay (Life Technologies) and TaqMan advance Master Mix (Life
Technologies) according to the manufacturer’s protocols. Expression of genes perturbed by toxicant treatment was analyzed using Fast SYBR Green master mix (Life Technologies) and
154 primers listed in Supplementary Table S1. 18S and GAPDH were used as housekeeping genes for
TaqMan gene expression and SYBR Green PCRs, respectively. All RT-PCRs were performed in duplicates on Fast Applied Biosystems 7500 System (Life Technologies) with the following thermal cycling parameters: SYBR® Green RT-PCR (95 °C for 20 s, followed by 40 cycles of 3 s at 95 °C and 30 s 60 °C); a melting curve step was included in SYBR Green reactions (95 °C for
15 s, 60 °C for 1 min, 95 °C for 15 s, and 60 °C for 15 s); TaqMan gene expression assay (95 °C for 20 s, followed by 40 cycles of 3 s at 95 °C and 30 s 60 °C). For miRNA amplification, short stem-loop cDNA libraries from individual miRNAs were generated using TaqMan® microRNA assays and TaqMan® microRNA reverse transcription kits (Life Technologies). Up to six miRNAs were multiplexed in one reaction. Quantitative real-time PCR on miRNAs was performed using the TaqMan® microRNA assay kit in combination with TaqMan® FAST advanced PCR master mix, (Life Technologies) with the following thermal cycling parameters:
95 °C for 20 s, followed by 40 cycles of 3 s at 95 °C and 30 s 60 °C. Expression of individual miRNAs was normalized to RNU44 expression and was shown relative to expression in solvent- treated LUHMES. The relative mRNA and miRNA expression was quantified using the comparative CT (2−∆∆CT) method (Schmittgen and Livak 2008). Data collected from three to four independent experiments were calculated as average log2-fold change in independent biological replicates ± SEM. Differences in treated and control samples were analyzed for statistical significance using one-way ANOVA test followed by Dunnett’s post hoc test. p value
<0.05 is denoted on graphs by *, p < 0.01 by **, and p < 0.001 by ***, respectively.
155
4.4. RESULTS
Protocol adaptation and optimization for LUHMES differentiation in 3D
Differentiation of LUHMES cells in monolayer culture (Fig. 1a, b, 2D diff protocol) was well established and extensively characterized earlier (Scholz et al. 2011; Schildknecht et al. 2013), and the described changes in morphology, proliferation (Ki-67), and differentiation (Map2 and
NF200) have been reproduced here (Supplementary Figure S1a). However, LUHMES cells differentiated in monolayer have limited life span. With increasing culture age, the interaction of the large neurite network with the extracellular matrix (plate coating) weakens, and the network either contracts into ganglion-like structures or fully detaches from the plate (Fig. 1b, last panel, d9). The brief survival of the differentiated cultures in 2D (which allows acute toxicity studies) is an obstacle for long-term, low-dose toxicological studies, as well as for cellular adaptation and resilience studies after toxicological stress. Therefore, we modified and adapted the LUHMES differentiation protocol for 3D. The 3D LUHMES model was prepared using the gyratory shaking technique as established for 3D rat primary aggregating brain cell cultures (Honegger and
Monnet-Tschudi 2001; van Vliet et al. 2008) and iPSC microphysiological systems (Hogberg et al. 2013) with few modifications (Fig. 1a, c, 3D diff protocol).
First, the size of aggregates was monitored during differentiation. By adjusting initial cell number and shaker speed, we were able to control aggregate size (Fig. 1e, 3D diff). The cultivation of the aggregates from day 0 of differentiation under a constant shaking speed of 80 rpm allowed us to keep the size of aggregates within 300–425 μm in diameter through 21 days of differentiation, while under a gradually increasing speed from 68 to 80 rpm during the first 5 days, as originally established for rat primary cultures (Honegger and Monnet-Tschudi 2001), LUHMES aggregates reached 700 μm in diameter (Fig. 1e, 3D gradient).
156
Second, in order to test sufficient oxygen and nutrient supply, early apoptosis (Annexin V and
caspase 3/7-positive cells) and necrosis (7-AAD-positive cells) were monitored by flow
cytometry and fluorescence microscopy. Although a low percentage of caspase 3/7-positive cells
were detectable on day 21 of differentiation (Fig. 2a), caspase 3/7-positive cells were distributed
equally throughout the aggregates without any visible accumulation in the middle of the
aggregates. No increase in Annexin V-positive cells was observed during the 21 days of
differentiation in 3D (Fig. 2b). The percentage of Annexin V- and 7AAD-positive cells in the 3D
cultures was comparable to those in monolayer undifferentiated LUHMES cultures, which were
subjected to the same preparation procedure—both cultures were trypsinized for 30 min prior to
Annexin V/7-AAD staining.
Figure 4-2. Quantification of apoptosis and necrosis in 3D LUHMES model. Caspase 3/7 activation (green nuclei) as an early apoptotic marker was visualized using fluorescent microscopy in combination with
157
Hoechst 33342 staining of nuclei (blue) in undifferentiated LUHMES monolayer cultures (d0) and 21 days after induction of differentiation in 3D. Scale bars are 50 μm. b Annexin V/7-AAD-positive cells were quantified using flow cytometry on day 0 (as negative control) and days 12, 15, and 21 following 3D diff. protocol for differentiation. Aggregates exposed to 0.5 μM rotenone for 48 h and after 7 days recovery were used as a positive control (last panel). Annexin V-positive cells are in early apoptosis; double stained for
Annexin V and 7AAD cells are in later apoptotic phases, while 7-AAD-positive cells represent a population of necrotic cells. Data are shown as mean ± SEM, n ≥ 3 (independent experiments) c Penetration assay with
Hoechst 33342: 12-day-old aggregates of GFP-expressing LUHMES were differentiated according to 3D diff protocol and were stained for increasing time intervals with Hoechst 33342. Confocal optical slices through the center of the aggregates are shown to demonstrate time-dependent penetration of Hoechst 33342 through the aggregates. Compare GFP expression (green) at all time points in the center of aggregates with the absence of Hoechst 33342 staining (blue) after 5, 15, and 30 min of incubation and penetration of Hoechst
33342 to the middle after 1 and 6 h of incubation. No apoptotic nuclei are visible in the center of aggregates
(60 min and 6 h). Scale bar is 100 μm (color Figure online).
Third, we investigated compound penetration by staining the live, 12-day-old aggregates with
DNA-binding blue fluorescent dye, Hoechst 33342 trihydrochloride, MW = 616 g/mol
(Invitrogen) for 5, 15, 30, 60 min, and 6 h. For this experiment, LUHMES ubiquitously
expressing GFP were used (Schildknecht et al. 2013). Hoechst 33342 dye penetration throughout
the aggregates was advancing with increasing incubation time. Hoechst 33342 reached the middle
of the aggregates after 1 h of treatment (Fig. 2c, Supplementary Figure S1b). This experiment
ensured sufficient penetration of necessary small molecule factors for differentiation and
nutrients, as well as toxicants. In addition, no visible apoptotic nuclear fragmentation
accumulated in the middle of aggregates (Fig. 2c, 1 and 6 h). Thus, 3D cultures could be kept at
least twice as long in culture than their 2D counterparts.
158
Finally, withdrawal of FGF and the addition of tetracycline, cAMP, and GDNF should rapidly induce exit from the cell cycle and differentiation to postmitotic mature neurons. We suggest that the observed continuous increase in the size of aggregates during differentiation could be due to a prolonged proliferation in 3D differentiating cultures. Higher cell density and increased cell-to- cell interactions may stimulate signaling between the cells within the aggregates that impedes exit from the cell cycle. Therefore, we quantified the expression of Ki-67, a proliferation marker, in undifferentiated cells as well as in 2D and 3D cultures. As expected, undifferentiated LUHMES were 98 ± 2 %-positive for Ki-67. Induction of differentiation in 2D reduced the expression of
Ki-67 to 16 % by day 6, while in 3D cultures 49 ± 13 % of the cells were still Ki-67-positive on day 6 and 47 ± 12 % on day 12 (Fig. 3a, Supplementary Figure S1c). Therefore, we optimized the
3D diff protocol further to accelerate the exit from cell cycle in 3D and induce homogeneous differentiation. In the first step, we evaluated whether pre-differentiation in 2D for 48 h before 3D differentiation (Fig. 1a, 3D pre-diff protocol) would decrease proliferation. No differences in the size of aggregates (Fig. 1e, 3D pre-diff), as well as no change in Ki-67 expression (data not shown), were observed, compared to the 3D diff protocol; this protocol, therefore, was not followed further. In the second step, we tested whether increasing the tetracycline concentration would reduce the proliferation rate. LUHMES were differentiated according to the 3D diff protocol in the presence of 2, 4, and 10 μg/ml tetracycline. Although the highest tetracycline concentration reduced the proportion of proliferating Ki-67 cells (Supplementary Figure S1d), it appeared to be cytotoxic for the cultures (observation based on aggregate morphology, data not shown). Next, we applied treatment with the mitotic inhibitor taxol (also known as paclitaxel).
The supplementation of neural differentiation media with anti-proliferation drugs, such as cytosine arabinofuranoside (AraC), is common and broadly used in primary neuronal cultures to block the proliferation of neuroprogenitors and astroglia without affecting postmitotic neurons
(Gerhardt et al. 2001; Volbracht et al. 2006) After optimization experiments, 10 nM taxol for 48 h, from days 3 to 5 of differentiation, was chosen as a treatment scheme (Fig. 1a, 3D + T10
159 protocol). Treatment with taxol led to a reduction in aggregate size (250–300 μm on average, Fig.
1d, e, 3D + T10) and significant decreased in Ki-67-positive cells to 6 ± 6 % on day 6 and 2 ± 2
% on day 12 of differentiation (Fig. 3a, Supplementary Figure S1c). In addition, we analyzed the expression of the Ki-67 gene prior to and six and 12 days after induction of differentiation following either 3D diff or 3D + T10 protocols by real-time RT-PCR (Fig. 3b), which confirmed our flow cytometry data. The effects of taxol on Ki-67-positive cells were confirmed morphologically by immunocytochemistry, where whole aggregates were fixed at different stages of differentiation and stained with antibodies against Ki-67 and the postmitotic neuronal marker,
NeuN (Fig. 3c). Fewer Ki-67-positive cells and higher number of NeuN-positive cells were found in 3D + T10 samples on days 6 and 12 of differentiation in comparison with 3D diff samples.
Supplementation of LUHMES differentiation medium with taxol for 48 h selectively blocked proliferation without any negative effects on neuronal cells, increased the homogeneity of the cell population, and did not interfere with further toxicological studies since taxol was washed from the cultures before toxicant exposures. Thus, we favored the 3D + T10 protocol over other differentiation conditions, and this protocol was followed as a standard differentiation protocol for further experiments.
160
Figure 4-3. Estimation of proliferation rate within the aggregates. a Percentage of Ki-67-positive cells on days 6 and 12 of differentiation with or without taxol in comparison with undifferentiated LUHMES (d0).
The number of Ki-67-positive cells was measured using Alexa Flour 647-conjugated anti-Ki-67 antibody by flow cytometry. Data represent mean ± SD, n ≥ 3 (independent experiments). b Ki-67 gene expression on days 0, 6, and 12 of differentiation in 3D diff and 3D + T10 cultures. Data are normalized to Ki-67 expression on d0 and represent mean ± SEM, n ≥ 3 (independent experiments). c Immunostainings of 3D diff and 3D + T10 aggregates with antibody against KI-67 showing prolonged presence of Ki67-positive cells (red) in 3D diff cultures in comparison with 3D + T10 aggregates. The aggregates were co-stained with postmitotic neuronal marker NeuN (green). The nuclei were visualized with Hoechst 33342 staining.
Scale bars are 50 μm. The aggregates were fixed on glass slides and covered with coverslips for confocal imaging, which explains the larger size of the aggregates in comparison with Fig. 1, where floating aggregates were imaged.
161
Characterization of LUHMES differentiation in 3D
The differentiation in 3D was characterized by immunocytochemistry. In addition to Ki-67 and
NeuN stainings (described above), aggregates were stained with further neuronal markers (MAP2 for dendrites and apical part of axons, neurofilament (NF200) for axons, and synaptophysin for synapses) at different stages of differentiation with (3D + T10) and without (3D diff) taxol treatment. Induction of differentiation in 3D induced the expression of MAP2, NeuN, and synaptophysin; reduced the expression of Ki-67; and changed the morphology of neurofilament- and MAP2-positive neurites (Figure 3). Interestingly, treatment with taxol not only inhibited proliferation, but significantly enhanced maturation, dendritic morphogenesis, and arborization as shown for MAP2 and synaptophysin stainings (Figs. 4, 5a). For more detailed visualization of long neurites protruding from the differentiated neurons, high magnification of MAP2/NF200 and synaptophysin/NF200 stainings of taxol-treated aggregates is shown (Fig. 5b). These findings are in agreement with publications showing that low taxol concentrations promote lamellipodial protrusions, stabilize microtubules, and increase spine formation (Buck and Zheng 2002; Gu et al.
2008).
Real-time PCR was performed to analyze induction of neuronal genes during differentiation of
LUHMES in 3D. Expression of general neuronal markers (β-III-tubulin, NeuN, synapsin1), marker genes specific for dopaminergic neurons [tyrosine hydroxylase (TH), dopamine transporter (DAT), and vesicular monoamine transporter member 2 (VMAT2)], as well as proliferation and neural precursor markers Ki-67 and Nestin, were analyzed in course of 3D +
T10 differentiation and normalized to the expression levels at day 0 (Fig. 5c). Ki-67 and Nestin were down-regulated during differentiation, while expression of neuronal markers was significantly induced (one-way ANOVA test followed by Dunnett’s post hoc test). The expression levels of these marker genes were similar to those in 2D differentiated LUHMES (Fig.
162
5d) with slightly higher expression of TH in 3D cultures versus 2D. Note that the expression level
of all genes plateaued on day 6 of differentiation, suggesting complete differentiation.
Figure 4-4. Immunocytochemistry of neuronal differentiated LUHMES in 3D diff and 3D + T10 cultures on days 6, 12, 15, and 21 after induction of differentiation. Panel a shows the induction of expression of MAP2
(green) and NF200 (red), as well as maturation and neurite elongation in 3D cultures followed 3D +T10 protocol versus 3D diff protocol over a span of 21 days of differentiation. Panel b shows the overlay of
NF200 (red) and typical punctual staining with synaptic marker, synaptophysin (Syn, green). Note higher synaptophysin expression in 3D + T10 cultures versus 3D diff cultures. The nuclei were visualized with
Hoechst 33342 staining. Scale bars are 50 μm. The aggregates were fixed on glass slides and covered with
163 coverslips for confocal imaging which explains the larger size of the aggregates in comparison with Fig. 1, where floating aggregates were imaged.
Figure 4-5. Enhanced neuronal maturation in 3D + T10 cultures. a MAP2 staining of representative
aggregates differentiated for 12 days following either 3D +T10 (first panel) or 3D diff. (second panel)
protocols. The nuclei were visualized with Hoechst 33342 staining. b Higher magnification (63×) of
164
representative aggregates differentiated for 12 days under 3D + T10 conditions and stained with
synaptophysin (green), NF200 (red) in the first slide and MAP2 (green), NF200 (red) in the second slide.
The nuclei were visualized with Hoechst 33342 staining. Scale bars are 50 μm. Real-Time RT-PCR of
genes involved in LUHMES neuronal differentiation and maturation. LUHMES were differentiated in 3D
+ T10 (c) and 2D monolayer cultures (d). RNA samples were collected on days 3, 6, 9, 12, 15, and 21 of
differentiation and prior induction of differentiation (day 0) as a control. Data represent mean of log2
(fold change) ± SEM normalized to d0 from at least four independent experiments. Statistical significance
was calculated using one-way ANOVA test followed by Dunnett’s post hoc test. Expression of all genes
was significantly (p < 0.05) different in comparison with day 0, except Nestin in 3D cultures and Ki-67 in
2D cultures (Supplementary Table S2) (color Figure online)
LUHMES 3D model for neurotoxicity testing
Next, we analyzed the performance of the 3D model for neurotoxicity testing by applying two well-known neurotoxicants, MPP+ and rotenone. Both chemicals are mitochondrial complex I inhibitors and cause Parkinsonism (Betarbet et al. 2000; Franco-Iborra et al. 2015). MPP+ is specific for dopaminergic neurons, because of its selective uptake by them (Langston et al. 1984), while rotenone has broader toxicity. LUHMES neuronal aggregates were treated with increasing concentrations of both compounds for 24 and 48 h. First, cell viability assay, based on mitochondria metabolic capacity, was performed to generate concentration–response curves (Fig.
6a, b). Second, a cytotoxicity assay, based on the measurement of membrane integrity, was conducted in the same samples using LDH release assay (Supplementary Figure S2). As expected, mitochondria impairment was measured at concentrations at which the cellular membrane was still intact (low LDH activity in the media), confirming the mitochondria selectivity of the test compounds by the higher sensitivity of the resazurin reduction-based assay.
The concentrations (5 μM MPP+ and 0.1 μM rotenone) with slight mitochondria impairment after
24- and 48-h exposure were used for further gene expression and washout experiments.
165
Figure 4-6. Cell viability of LUHMES aggregates after exposure to rotenone and MPP+. LUHMES cells were differentiated following 3D + T10 protocol and exposed reversely to different rotenone (a) and MPP+
(b) concentrations from day 6 to 8 (48 h) and from day 7 to 8 (24 h). c Cell viability after toxicant washout and recovery period. LUHMES were exposed reversely until day 8 for indicated period of time to 0.1 μM rotenone and 5 μM MPP+. On day 8, compounds were washed out and cells recovered for further 7 days.
Chronic/repeat-dose (fresh substance was added with each medium exchange) exposure (192 h, from day 7 until day 15) was included as positive control. Cell viability was analyzed using resazurin reduction assay.
Cell viability is presented in % of solvent-treated controls in at least three independent experiments (n ≥ 3, mean ± SEM, n = 2 for MPP+ on day 15). d Mitochondrial membrane potential in individual LUHMES aggregates, exposed to rotenone for 48 h from day 6 to 8 measured by Mitotracker assay. Fluorescence intensity was measured as mean gray values using ImageJ software and normalized to the size and then to the fluorescence intensity of DMSO control aggregates (n = 3), at least 10 aggregates were assayed for each independent experiment, ***p < 0.001, Kruskal–Wallis followed by Dunn’s post hoc test).
166
As a proof of concept, compound washout experiments were performed to address counter- regulation responses after short-term exposure in comparison with long-term chronic exposure.
LUHMES were differentiated in 3D and exposed to 0.1 μM rotenone and 5 μM MPP+ for 12, 24,
48, or 192 h. On day 8 of differentiation, after 12, 24, and 48 h of exposure, compounds were washed out, and aggregates seeded into new plates and cultivated further until day 15. In case of
192-h exposure, aggregates were exposed to the toxicants continuously from day 7 until day 15.
Cell viability (resazurin reduction assay) was assessed for all exposure conditions on day 15 of differentiation (Fig. 6c). Continuous exposure (192 h) to 5 μM MPP+ was 100 % toxic for
LUHMES, while exposure to 0.1 μM rotenone for 192 h reduced cell viability by 70 %.
Interestingly, after MPP+ wash out, around 80 % of cells were lost by day 15. This suggested either that MPP+ accumulates in the aggregates and continues to affect mitochondria after wash out, or that processes initiated by 5 μM MPP+ cannot be reversed, and cells cannot recover from the primary hit (at least at this concentration) (Fig. 6c dark gray line).
The washout effect was different for varying durations of 0.1 μM rotenone exposures (Fig. 6c light gray line). Exposure for 12 and 24 h further reduced viability by 34 % in total, while cells treated with 0.1 μM rotenone for 48 h were more strongly affected (47 % decrease in viability). However, although 5 μM MPP+ was less toxic than rotenone immediately after the hit (day 8), its withdrawal could not rescue cells from ongoing cell death, while in samples treated with 0.1 μM rotenone, cell viability continued to decline but to lesser extent than in MPP+ samples.
Since mitochondria are the primary target for rotenone, we further evaluated the effects of rotenone on mitochondrial membrane potential in individual aggregates. LUHMES were differentiated following 3D + T10 protocol, exposed to 0.05, 0.1, and 0.5, 1 and 10 μM rotenone or DMSO from d6 to d8 of differentiation. After 48-h exposure to rotenone, the aggregates were stained with the
167
MitoTracker dye, to allow its accumulation in mitochondria according to the magnitude of their membrane potential. The mean fluorescence intensity values were then estimated in individual aggregates by fluorescence microscopy and normalized to DMSO controls (Fig. (Fig.6d,6d, n ≥ 3, independent experiments with 10–20 aggregates assayed per experiment). Mitochondrial activity was significantly reduced in rotenone-treated samples. High correlation between data from resazurin and MitoTracker assays was observed for lower rotenone concentrations (0.05, 0.1, and
0.5 μM), which was not as closely related for the higher cytotoxic concentrations (1 and 10 μM), where changes in morphology and size of the aggregates prohibited precise microscopic evaluation of MitoTracker samples.
There are some limitations in imaging 3D cultures. Since 3D aggregates are ≥200 μm thick, imaging them using conventional fluorescence microscopy is very challenging due to issues with light scattering and penetration depth. Advanced confocal microscopy and/or two-photon microscopy in combination with optical clearing by treatment of the tissue with Scale clearing solution (Hama et al. 2011) prior to imaging overcome these limitations. Previously, it has been shown that rotenone perturbs neurite integrity in 2D LUHMES cultures (Schildknecht et al. 2013;
Krug et al. 2013). To confirm these findings and to optimize the imaging of neurite integrity in 3D cultures, RFP-expressing LUHMES were used. Wild-type LUHMES were mixed with RFP- expressing LUHMES in the ratio 49:1 on day 0 of differentiation and differentiated following the
3D + T10 protocol. It was shown previously that RFP is only expressed in viable cells
(Schildknecht et al. 2013). After rotenone treatment from days 6 to 8 of differentiation, RFP- expressing viable cells within the aggregates were imaged using confocal microscopy for neurite quantification. Exposure of LUHMES aggregates to 0.1 μM rotenone significantly affected the neurite integrity in comparison with DMSO controls (Fig. 7a, b). Thus, application of the fluorescent cell line mixed with wild-type cells helped to overcome the limitation of image quantification in these highly compact three-dimensional cultures.
168
Figure 4-7. Neurite integrity in individual LUHMES aggregates exposed to rotenone. a Exposure to 0.1 μM rotenone for 48 h from d6 to d8 perturbed neurite integrity. Confocal images showing RFP-expressing cells
(red) mixed in a 1:49 ratio with wild-type cells and differentiated following the 3D + T10 protocol for 8 days.
Note rotenone-altered neurite integrity of viable RFP-expressing cells in comparison with DMSO control samples. Nuclei are stained with Hoechst 33342. Scale bars are 50 μm. b Quantification of neurite area in rotenone-treated samples versus DMSO controls, normalized to the number of RFP-positive cell bodies in three independent experiments (nine aggregates were quantified for rotenone-treated samples and 12 for
DMSO control samples) (n = 3, **p < 0.01, Mann–Whitney test).
Exposure of the 3D LUHMES model to rotenone and MPP+ alters the expression of genes involved in transsulfuration and one-carbon metabolic pathways The performance of the 3D model for toxicological studies was analyzed by gene expression. We have chosen a panel of candidate genes which were shown to be involved in cellular adaptation to MPP+ exposure in 2D LUHMES cultures by regulating central carbon metabolism and amino acid turnover (ASS1, argininosuccinate synthase, SHMT2, serine hydroxymethyl transferase), transsulfuration pathway
[CTH, cystathionase (cystathionine γ-lyase)], oxidative stress and DNA replication and repair
169
[TYMS, thymidylate synthetase MLF1IP, centromere protein U (MLF interacting protein)] in earlier studies (Krug et al. 2014). LUHMES were differentiated following the 3D + T10 protocol and exposed to 0.1 μM rotenone and 5 μM MPP+ for 12 and 24 h on day 7 (Fig. 8a). In agreement with our earlier studies (Krug et al. 2014), on day 8—immediately after exposure—we observed the same regulation trends of those genes by rotenone and MPP+ [Fig. 8b, c, dark bars (rotenone),
Supplementary Figure S3 (MPP+)]. Gene expression analysis was performed in three to four independent experiments (up to 12 technical replicates) and normalized to DMSO-treated controls.
ASS1 was the most strongly up-regulated gene by MPP+ (FC = 3.3, 24 h) and rotenone (FC = 2.4,
24 h). ATF4, activating transcription factor four, was identified as upstream regulator of the cellular cascades initiated by MPP+ (Krug et al. 2014) but was less up-regulated in our 3D model by rotenone (FC = 1.5, 24 h), though 2.3 times increased by 24-h MPP+ treatment. CTH and
SHMT2 were more up-regulated by MPP+ than by rotenone. MLF1IP and TYMS were significantly down-regulated in the 3D system following MPP+ and rotenone treatment. As proof- of-concept experiments—to study cellular counter-regulation—rotenone was washed out on day 8 of differentiation, and cells were kept in culture for further 7 days. Since washout of 5 μM MPP+ did not prevent cell death, we analyzed the expression of the same panel of genes on day 15 only in rotenone-treated samples (Fig. 8b, c, light bars). Interestingly, ASS1, CTH, and SHTM2, which were up-regulated immediately after exposure, were down-regulated 7 days later after rotenone withdrawal (Fig. 8b), while down-regulated genes (MLF1IP and TYMS) were further repressed with an even stronger effect (Fig. 8c). ATF4 was only slightly up-regulated on day 8 and returned to control level 7 days after recovery (Fig. 8b). This observation suggests that certain genes and signaling pathways are counter-regulated and/or may be responsible for cellular recovery after the primary hit, while other processes cannot be restored and, thus, might be a part of the new cellular homeostasis.
170
Figure 4-8. Time-dependent perturbations of gene expression after exposure of 3D LUHMES to rotenone. a
Toxicant treatment and washout scheme: LUHMES were differentiated following 3D + T10 protocol; exposure to 0.1 μM rotenone occurred for 12 or 24 h from day 7 until day 8; Samples were collected for RT-
PCR immediately after exposure on day 8 (dark arrow) or after rotenone washout and 7 days recovery on day
15 (light arrow). b Protein coding genes (ASS1, AT4, CTH, SHMT2) and miRNA (mir-7) with counter- regulation pattern after rotenone washout in comparison with acute response. c Protein coding genes
(MLF1IP, TYMS) and miRNA (mir-16) with stronger response after rotenone withdrawal in comparison
171 with acute toxicity. Dark bars show expression of the genes on day 8, while light bars show expression of the genes after rotenone washout and 7-day recovery. The data are means of log2 (fold change) ± SEM of at least three independent experiments (9–12 technical replicates). (n ≥ 3, *p < 0.05, **p < 0.01, and ***p < 0.001, one-way-ANOVA followed by Dunnett’s post hoc test)
Altered expression of mir-7 miRNA after exposure of 3D LUHMES to rotenone
Finally, miRNAs involved in mitochondrial functions and relevant for PD were analyzed after
exposure of LUHMES aggregates to 0.1 μM rotenone. In order to test whether miRNAs are
involved in the recovery process, miRNA expression was assessed on day 8 as the reaction to the
primary toxicant hit and on day 15, 7 days after rotenone withdrawal (refer to Fig. 8a for
treatment and sampling scheme). In agreement with the literature showing down-regulation of
mir-7 in PD models (Junn et al. 2009; Fragkouli and Doxakis 2014), we observed a reduction of
miR-7 expression as early as 12 h after rotenone treatment (Fig. 8b, dark bar), while known pro-
apoptotic miR-16 remained at control level (Fig. 8c). No changes were observed in expression of
miR-210 (hypoxia-sensitive miRNA, involved in mitochondrial respiration (Chan et al. 2012,
data not shown), suggesting mir-7 as a primary rotenone miRNA target prior to mitochondria-
mediated apoptosis. On day 15 after rotenone washout, however, mir-7 expression went back to
control levels, grouping this miRNA together with other counter-regulated genes (Fig. 8b),
suggesting a possible role of this miRNA in cellular adaptation and recovery. In addition, brain-
specific miRNA, mir-124, was unchanged on day 8 of differentiation and was upregulated on day
15 after washout (data not shown).
172
4.5. DISCUSSION
The information gained from acute cytotoxicity studies appears to be of limited relevance for the understanding of chronic/slow-developing processes and low-dose chemical exposures (Hengstler et al. 2012). This situation also holds true for the study of neurodegenerative diseases such as
Parkinson’s disease. Moreover, there are indications that the use of rodent models and cells may have low predictivity for human disease states (Hartung 2008; Leist and Hartung 2013; Leist et al. 2014), and therefore, attempts are ongoing to provide toxicological/disease models on the basis of human cells (Krause et al. 2013).
At present, there are only few options to work with human dopaminergic neurons in long-term experiments. Therefore, the primary aim of this work was to adapt the LUHMES neuronal model
(Scholz et al. 2011) to 3D. Besides emerging evidence in the literature pointing out the general advantages of 3D cell systems over classical monolayer cultures (Alépée et al. 2014), the main reason for attempting the 3D LUHMES model here was the opportunity for prolonged cultivation and increased cell survival after induction of differentiation. The 2D LUHMES model has been used successfully to study neurotoxicity, especially related to Parkinson’s (Schildknecht et al.
2009; Stiegler et al. 2011; Schildknecht et al. 2013; Krug et al. 2013, 2014; Zhang et al. 2014;
Stępkowski et al. 2015). Several neurotoxicological end points were established for monolayer
LUHMES cultures, such as neurite outgrowth (Krug et al. 2013), Tau phosphorylation and associated cell death (Selenica et al. 2007), reporter cell line-based assays using high-content imaging (Schildknecht et al. 2013; van Vliet et al. 2014), omics technologies to study perturbations in cellular metabolism, and gene expression after toxicant exposure (Krug et al.
2014). However, in prolonged three-dimensional cultures, cellular junctions are more in vivo-like conditions and allow pronounced neuronal network formation. In addition, the specific composition of the extracellular matrix in the brain, which differs significantly from that of other organs (Yamaguchi 2000) and was shown to promote synaptogenesis and neuronal network
173 formation (Dityatev and Fellin 2008; Frischknecht and Gundelfinger 2012), and the lack of glia support, may make 2D neuronal cultures more sensitive and less adherent to cell culture plastic ware.
We adapted the LUHMES differentiation protocol to 3D applying a gyratory shaking technic
(Honegger and Monnet-Tschudi 2001) and showed prolonged survival in 3D in comparison with
2D cultures (Fig. 1b, c). A frequent concern about 3D cultures is the potential insufficient nutrient and oxygen supply to the center of aggregates (Minchinton and Tannock 2006; Derda et al.
2009). We demonstrated that the gyratory shaking method of LUHMES spheroid cultivation combined with taxol treatment allows controlled spheroid size and fast penetration of compounds, for the example of Hoechst 33342 dye, into the middle of aggregates. Nuclei staining with
Hoechst 33342, as well as caspase 3/7 staining, ensured the absence of cell death in the center of aggregates during differentiation (Fig. 2a–c). The absence of proliferating cells after 48-h taxol treatment (Fig. 3) also proved the efficient penetration of the drug throughout the aggregates.
Demonstration of penetration of relatively large molecules such as positive-charged Hoechst
33342 (Mw = 616 g/mol) and uncharged taxol (Mw = 854 g/mol) throughout the aggregates ensures that the model is suitable for compound testing.
LUHMES cells formed a pronounced neuronal network and continued to mature further in 3D at later days of differentiation as shown in Fig. 4a, b (note increase in synaptophysin-stained cells, as well as advanced neuronal network from day 12 to day 21 of differentiation), showing the importance of keeping cultures longer for neuronal network formation and synaptogenesis studies. Although the LUHMES 3D model consists of homotypic aggregates, we were surprised to observe cellular organization and polarity within the aggregates similar to organotypical spheroids described by (Lancaster et al. 2013) as shown in Figs. 4a, a,5a5a for MAP2-positive cells, accumulating at one side of the aggregate starting on day 12 and onwards. This observation suggests that MAP2-positive cells may migrate from the initial differentiation area through the aggregate during maturation. A similar pattern was also observed for synaptophysin staining but
174 not for NF200, which was equally distributed throughout the aggregates at all stages of differentiation (Fig. 4a, b). Further experiments using combinations of live imaging techniques and RFP/GFP expression LUHMES may clarify this observation.
Especially for long-term exposures in vivo (where organ toxicities are recorded), the first organ to decompensate leads the toxicity. Therefore, in vitro it is important to measure not only the reaction of cells to a hard, single hit, if we want to evaluate neurodegenerative processes or degenerative diseases. Instead, we should rather measure the potential to compensate/recover from multiple subtoxic hits, and over longer periods. Such differences are important to work out to follow the new toxicological strategies suggested by the national research council (US) in 2007
(NRC 2007). Thus, the long-term shelf life of our model may allow designing such experiments.
In this study, we used the well-known dopaminergic neuronal toxicants MPP+ and rotenone as model compounds. MPP+ is a toxic metabolite of 1-methyl-4-phenyl-1.2.3.6-tetrahydropyridine
(MPTP), and rotenone is a broadly used pesticide. Both MPTP and rotenone are highly lipophilic, which makes it very easy for them to cross the blood–brain barrier, and they represent prime examples of mitochondrial toxins (Miller et al. 2009). Both compounds accumulate in mitochondria and inhibit complex I of the electron transport (respiratory) chain, a major target of
ROS and the reason they are used in animal models to study Parkinsonism (Langston et al. 1984;
Betarbet et al. 2000). As a proof of concept, LUHMES were differentiated in 3D, exposed either to MPP+ or rotenone for a short period of time (12, 24, or 48 h); then, the compounds were washed out, and cells were cultivated for a further 7 days. First, we observed a similar cytotoxicity on day 8 in 3D as in our earlier studies in monolayer cultures (Fig. 6a, b; Krug et al.
2014), as measured by the resazurin reduction assay. Second, we confirmed perturbation of genes involved in one-carbon metabolism and transsulfuration pathways (ASS1, CTH, and SHTM2), on day 8 after 12 or 24 h of exposure to MPP+ and rotenone (Fig. 8b, c; Supplementary Figure S3;
Krug et al. 2014). Interestingly, MPP+ effects were stronger on certain genes than those of rotenone (much stronger induction of ASS1 and AT4, SHMT2, and CTH, for example). Third,
175 we showed a down-regulation of PD-relevant mir-7 as early as 12 h after rotenone exposure, while pro-apoptotic mir-16 and rotenone-sensitive mir-210 were not yet significantly perturbed
(Fig. 8b, c and data not shown). Finally, washout experiments demonstrated different counter- regulation after short-term exposures to sub-cytotoxic concentrations of rotenone or MPP+. The
3D cultures allow moving the aggregates into a new cell culture dish, thus avoiding remaining toxicant contamination (e.g., compound bout to plastic), which represents an enormous advantage over 2D cultures for resilience studies. Interestingly, 5 μM MPP+ was not cytotoxic 24 h after exposure, but MPP+ withdrawal was not sufficient, and 80 % of cells were dead after the 7-day recovery period, likely due to the accumulation of MPP+ within the cells. 0.1 μM rotenone decreased cell viability by 20 % after 24 h, which was further decreased by additional 16 % after rotenone withdrawal (Fig. 6c). The molecular mechanisms, which prevented 64 % of cells from death in rotenone samples after washout, shall be addressed in future experiments. We observed different patterns of gene expression after 7 days of recovery. Genes down-regulated by rotenone,
MLF1IP and TYMS, were further down-regulated, while mir-7 returned to normal level with a slight tendency for up-regulation. mir-7 is one of the few known miRNAs related to PD
(reviewed in Mouradian 2012; Kaidery et al. 2013). One of mir-7’s confirmed targets is α- synuclein, a major player in PD pathogenesis (Junn et al. 2009). Overexpression of mir-7 in murine primary cortical neurons prior to exposure to MPP+ demonstrated a neuroprotective effect against MPP+ through activation of TOR pathway (Fragkouli and Doxakis 2014). Recently, mir-
7 was shown to protect SH-SY5Y cells against MPP+-induced cell death by targeting RelA (a component of nuclear factor-κB (NF-κB) (Choi et al. 2014). These studies, together with our findings, suggest a possible role of this miRNA in counter-regulation against the stress and for cellular resilience after exposure to the mitochondria toxicants MPP+ and rotenone. Here, for the first time, we demonstrated a significant down-regulation of mir-7 expression by rotenone and its complete recovery after rotenone withdrawal in a human-relevant Parkinson’s disease in vitro model. Further functional studies are needed to confirm this observation. In contrast, rotenone-
176 induced ASS1, SHTM2, and CTH were down-regulated after rotenone withdrawal (Fig. 8b), which could also be involved in counter-regulation mechanisms.
A further advantage of the longer-lived 3D LUHMES model over monolayer cultures could be the easier co-culture and readout of LUHMES with other cell types, e.g., astrocytes or liver cells.
3D aggregates can be added to monolayer cultures of astrocytes or liver cells for the time of exposure without mixing the two cell populations as with 2D cultures (Efrémova et al. 2015).
After exposure, the response to the toxicant treatment can be assessed separately for LUHMES spheres and the other cell types. This may allow for the study of neuroprotective effects of factors released by glial cells and inclusion of metabolic competence (liver cells) in the model.
In addition, the 3D LUHMES model may allow us to study the protective effects of Parkinson’s drug therapies by first exposing the cells to the toxicant and then to the drug. Thus, the model may have the potential for restorative/disease-modifying drug screening. Using reporter cell lines, in combination with quantitative high-content imaging, may contribute significantly to
Parkinson’s drug screening (Schildknecht et al. 2013). The introduction of the fluorescent cell lines into the wild-type LUHMES population in low percentage allows clear visualization of neurites and their assignment to the corresponding cell bodies. Quantification of cellular and neurite morphology in the mixed cultures of fluorescent cell lines, together with wild-type
LUHMES as described in our proof-of-concept experiment in Fig. 7, can be used for the fast screening of potential toxicants (van Vliet et al. 2014) contributing to disease development as well as the efficiency of newly developed treatments.
In conclusion, we have established a 3D LUHMES model that will allow analysis of the long- term effects of toxicant exposure, such as delayed response to the toxicant insult, cellular resilience, and/or adaptation to a new homeostasis after toxicant withdrawal (discussed in
Smirnova et al. 2015b).
177
4.6. ACKNOWLEDGEMENTS
This project was supported by the NCATS grant “A 3D Model of Human Brain Development for
Studying Gene/Environment Interactions” (1U18TR000547) and by the US Food and Drug
Administration (FDA) Grant “DNTox-21c—Identification of Pathways of Developmental
Neurotoxicity for High-Throughput Testing by Metabolomics” (U01FD004230). For flow cytometry and confocal imaging experiments, we used equipment provided by the BD Flow
Cytometry and the Cell Sorting Facility and Integrated Imaging Centers, Johns Hopkins
University, respectively. GH was supported by an International Foundation for Ethical Research
(IFER) graduate fellowship. JD was supported by the German Research Foundation (DFG
RTG1331). We acknowledge David Kolb for his support in confocal image quantification and
Michael Hughes for proof reading.
178
4.7. REFERENCES
Alépée N, Bahinski A, Daneshian M, et al. State-of-the-art of 3D cultures (organs-on-a-chip) in safety
testing and pathophysiology. ALTEX. 2014;31:441–477.
Ascherio A, Chen H, Weisskopf MG, O’Reilly E, McCullough ML, Calle EE, Schwarzschild MA, Thun
MJ. Pesticide exposure and risk for Parkinson’s disease. Ann Neurol. 2006;60:197–203. doi:
10.1002/ana.20904
Bandiera S, Matégot R, Girard M, et al. MitomiRs delineating the intracellular localization of microRNAs
at mitochondria. Free Radic Biol Med. 2013;64:12–19. doi: 10.1016/j.freeradbiomed.2013.06.013.
Betarbet R, Sherer TB, MacKenzie G, et al. Chronic systemic pesticide exposure reproduces features of
Parkinson’s disease. Nat Neurosci. 2000;3:1301–1306. doi: 10.1038/81834.
Borland MK, Trimmer PA, Rubinstein JD, et al. Chronic, low-dose rotenone reproduces Lewy neurites
found in early stages of Parkinson’s disease, reduces mitochondrial movement and slowly kills
differentiated SH-SY5Y neural cells. Mol Neurodegener. 2008
Buck KB, Zheng JQ. Growth cone turning induced by direct local modification of microtubule dynamics. J
Neurosci. 2002;22:9358–9367.
Chan YC, Banerjee J, Choi SY, Sen CK. miR-210: the master hypoxamir. Microcirculation. 2012;19:215–
223. doi: 10.1111/j.1549-8719.2011.00154.x.
Chinta SJ, Andersen JK. Dopaminergic neurons. Int J Biochem Cell Biol. 2005;37:942–946. doi:
10.1016/j.biocel.2004.09.009.
Choi DC, Chae Y-J, Kabaria S, et al. MicroRNA-7 protects against 1-methyl-4-phenylpyridinium-induced
cell death by targeting RelA. J Neurosci. 2014;34:12725–12737. doi: 10.1523/JNEUROSCI.0985-
14.2014.
Constantinescu R, Constantinescu AT, Reichmann H, Janetzky DB. Neuropsychiatric disorders an
integrative approach. Vienna: Springer; 2007. Neuronal differentiation and long-term culture of the
human neuroblastoma line SH-SY5Y; pp. 17–28.
Costello S, Cockburn M, Bronstein J, et al. Parkinson’s disease and residential exposure to maneb and
paraquat from agricultural applications in the central valley of California. Am J Epidemiol.
2009;169:919–926. doi: 10.1093/aje/kwp006.
179
Derda R, Laromaine A, Mammoto A, et al. Paper-supported 3D cell culture for tissue-based bioassays.
Proc Natl Acad Sci USA. 2009;106:18457–18462. doi: 10.1073/pnas.0910666106.
Dityatev A, Fellin T. Extracellular matrix in plasticity and epileptogenesis. Neuron Glia Biol. 2008;4:235–
247. doi: 10.1017/S1740925X09000118.
Efrémova L, Schildknecht S, Adam M, et al. Prevention of the degeneration of human dopaminergic
neurons in an astrocyte co-culture system allowing endogenous drug metabolism. Br J Pharmacol.
2015
Fragkouli A, Doxakis E. miR-7 and miR-153 protect neurons against MPP(+)-induced cell death via
upregulation of mTOR pathway. Front Cell Neurosci. 2014;8:182. doi: 10.3389/fncel.2014.00182.
Franco-Iborra S, Vila M, Perier C. The Parkinson disease mitochondrial hypothesis: where are we at?
Neuroscientist. 2015
Frischknecht R, Gundelfinger ED. The brain’s extracellular matrix and its role in synaptic plasticity. Adv
Exp Med Biol. 2012;970:153–171. doi: 10.1007/978-3-7091-0932-8_7.
Fujita KA, Ostaszewski M, Matsuoka Y, et al. Integrating pathways of Parkinson’s disease in a molecular
interaction map. Mol Neurobiol. 2014;49:88–102. doi: 10.1007/s12035-013-8489-4.
Gerhardt E, Kügler S, Leist M, et al. Cascade of caspase activation in potassium-deprived cerebellar
granule neurons: targets for treatment with peptide and protein inhibitors of apoptosis. Mol Cell
Neurosci. 2001;17:717–731. doi: 10.1006/mcne.2001.0962.
Giraldez AJ, Cinalli RM, Glasner ME, et al. MicroRNAs regulate brain morphogenesis in zebrafish.
Science. 2005;308:833–838. doi: 10.1126/science.1109020.
Grau CM, Greene LA. Use of PC12 cells and rat superior cervical ganglion sympathetic neurons as models
for neuroprotective assays relevant to Parkinson’s disease. Methods Mol Biol. 2012;846:201–211. doi:
10.1007/978-1-61779-536-7_18.
Greene LA, Tischler AS. Establishment of a noradrenergic clonal line of rat adrenal pheochromocytoma
cells which respond to nerve growth factor. Proc Natl Acad Sci USA. 1976;73:2424–2428. doi:
10.1073/pnas.73.7.2424.
Gu J, Firestein BL, Zheng JQ. Microtubules in dendritic spine development. J Neurosci. 2008;28:12120–
12124. doi: 10.1523/JNEUROSCI.2509-08.2008.
180
Hama H, Kurokawa H, Kawano H, et al. Scale: a chemical approach for fluorescence imaging and
reconstruction of transparent mouse brain. Nat Neurosci. 2011;14:1481–1488. doi: 10.1038/nn.2928.
Hartung T and Leist M (2008) Food for thought ... on the evolution of toxicology and the phasing out of
animal testing. ALTEX 25(2):91-102
Hartung T. 3D: a new dimension of in vitro research. Adv Drug Deliv Rev. 2014
Henchcliffe C, Beal MF. Mitochondrial biology and oxidative stress in Parkinson disease pathogenesis. Nat
Clin Pract Neurol. 2008;4:600–609. doi: 10.1038/ncpneuro0924.
Hengstler JG, Marchan R, Leist M. Highlight report: towards the replacement of in vivo repeated dose
systemic toxicity testing. Arch Toxicol. 2012;86:13–15. doi: 10.1007/s00204-011-0798-7.
Hogberg, HT, Bressler, J, Christian, KM, Harris, G, Makri, G, O'Driscoll, C, et al. (2013) Toward a 3D
model of human brain development for studying gene/environment interactions. Stem Cell Research &
Therapy 4 Suppl 1, S4–S4. doi:10.1186/scrt365
Honegger P, Monnet-Tschudi F. Protocols for neural cell culture. New York: Humana Press; 2001.
Aggregating neural cell cultures; pp. 199–218.
Hu W, He Y, Xiong Y, et al. Derivation, expansion, and motor neuron differentiation of human-induced
pluripotent stem cells with non-integrating episomal vectors and a defined xenogeneic-free culture
system. Mol Neurobiol. 2015
Huang W, Li MD. Nicotine modulates expression of miR-140*, which targets the 3′-untranslated region of
dynamin 1 gene (Dnm1) Int J Neuropsychopharmacol. 2009;12:537–546. doi:
10.1017/S1461145708009528.
Huang TT, Liu YY, Huang MM, et al. Wnt1-cre-mediated conditional loss of Dicer results in malformation
of the midbrain and cerebellum and failure of neural crest and dopaminergic differentiation in mice.
Fen Zi Xi Bao Sheng Wu Xue Bao. 2010;2:152–163.
Junn E, Lee K-W, Jeong BS, Chan TW, Im JY, Mouradian MM. Repression of alpha-synuclein expression
and toxicity by microRNA-7. Proc Natl Acad Sci USA. 2009;106:13052–13057. doi:
10.1073/pnas.0906277106.
181
Kaidery NA, Tarannum S, Thomas B. Epigenetic landscape of Parkinson’s disease: emerging role in
disease mechanisms and therapeutic modalities. Neurotherapeutics. 2013;10:698–708. doi:
10.1007/s13311-013-0211-8.
Kim JH, Auerbach JM, Rodriguez-Gomez JA, et al. Dopamine neurons derived from embryonic stem cells
function in an animal model of Parkinson’s disease. Nature. 2002;418:50–56. doi:
10.1038/nature00900.
Kim J, Inoue K, Ishii J, et al. A MicroRNA feedback circuit in midbrain dopamine neurons. Science.
2007;317:1220–1224. doi: 10.1126/science.1140481.
Kim JH, Park SG, Song S-Y, et al. Reactive oxygen species-responsive miR-210 regulates proliferation and
migration of adipose-derived stem cells via PTPN2. Cell Death Dis. 2013;4:e588. doi:
10.1038/cddis.2013.117.
Krause K-H, van Thriel C, De Sousa PA, et al. Monocrotophos in Gandaman village: India school lunch
deaths and need for improved toxicity testing. Arch Toxicol. 2013;87:1877–1881. doi:
10.1007/s00204-013-1113-6.
Krug AK, Balmer NV, Matt F, et al. Evaluation of a human neurite growth assay as specific screen for
developmental neurotoxicants. Arch Toxicol. 2013;87:2215–2231. doi: 10.1007/s00204-013-1072-y.
Krug AK, Gutbier S, Zhao L, et al. Transcriptional and metabolic adaptation of human neurons to the
mitochondrial toxicant MPP(+) Cell Death Dis. 2014;5:e1222. doi: 10.1038/cddis.2014.166.
Kumar Singh N, Dev Banerjee B, Bala K, et al. Gene–gene and gene-environment interaction on the risk of
Parkinson disease. Curr Aging Sci. 2014;7(2):101–109. doi: 10.2174/1874609807666140805123621.
Lancaster MA, Renner M, Martin C-A, et al. Cerebral organoids model human brain development and
microcephaly. Nature. 2013;501:373–379. doi: 10.1038/nature12517.
Langston JW, Langston EB, Irwin I. MPTP-induced parkinsonism in human and non-human primates–
clinical and experimental aspects. Acta Neurol Scand Suppl. 1984;100:49–54.
Lau P, de Strooper B. Dysregulated microRNAs in neurodegenerative disorders. Semin Cell Dev Biol.
2010
Lee J-W, Cannon JR. LRRK2 mutations and neurotoxicant susceptibility. Exp Biol Med (Maywood) 2015
doi: 10.1007/978-1-4939-2480-6.
182
Leist M, Hartung T (2013) Inflammatory findings on species extrapolations: humans are definitely no 70-
kg mice. Arch Toxicol 87(4):563-7 doi:10.1007/s00204-013-1038-0
Leist M, Hasiwa N, Rovida C, Daneshian M, Basketter D, Kimber I, Clewell H, Gocht T, Goldberg A,
Busquet F, Rossi AM, Schwarz M, Stephens M, Taalman R, Knudsen TB, McKim J, Harris G, Pamies
D, Hartung T (2014) Consensus report on the future of animal-free systemic toxicity testing. ALTEX
31(3):341-56. doi:10.14573/altex.1406091
Leucht C, Stigloher C, Wizenmann A, et al. MicroRNA-9 directs late organizer activity of the midbrain-
hindbrain boundary. Nat Neurosci. 2008;11:641–648. doi: 10.1038/nn.2115.
Li X, Jin P. Roles of small regulatory RNAs in determining neuronal identity. Nat Publ Group.
2010;11:329–338.
Li P, Jiao J, Gao G, Prabhakar BS. Control of mitochondrial activity by miRNAs. J Cell Biochem.
2012;113:1104–1110. doi: 10.1002/jcb.24004.
Lingor P, Unsicker K, Krieglstein K. Midbrain dopaminergic neurons are protected from radical induced
damage by GDF-5 application. J Neural Transm. 1999;106:139–144. doi: 10.1007/s007020050146.
Lotharius J, Falsig J, van Beek J, et al. Progressive degeneration of human mesencephalic neuron-derived
cells triggered by dopamine-dependent oxidative stress is dependent on the mixed-lineage kinase
pathway. J Neurosci. 2005;25:6329–6342. doi: 10.1523/JNEUROSCI.1746-05.2005.
Maertens A, Luechtefeld T, Kleensang A, Hartung T. MPTP’s pathway of toxicity indicates central role of
transcription factor SP1. Arch Toxicol. 2015;89:743–755. doi: 10.1007/s00204-015-1509-6.
Miller RL, Miller RL, James-Kracke M, et al. Oxidative and inflammatory pathways in Parkinson’s
disease. Neurochem Res. 2009;34:55–65. doi: 10.1007/s11064-008-9656-2.
Minchinton AI, Tannock IF. Drug penetration in solid tumours. Nat Rev Cancer. 2006;6:583–592. doi:
10.1038/nrc1893.
Miranda RC, Pietrzykowski AZ, Tang Y, et al. MicroRNAs: master regulators of ethanol abuse and
toxicity? Alcohol Clin Exp Res. 2010;34:575–587. doi: 10.1111/j.1530-0277.2009.01126.x.
Mouradian MM. MicroRNAs in Parkinson’s disease. Neurobiol Dis. 2012;46:279–284. doi:
10.1016/j.nbd.2011.12.046.
183
NRC—National Research Council, Committee on Toxicity Testing and Assessment of Environmental
Agents . Toxicity testing in the 21st century: a vision and a strategy. Washington, DC: The National
Academies Press; 2007.
Pallocca G, Fabbri M, Sacco MG, et al. miRNA expression profiling in a human stem cell-based model as a
tool for developmental neurotoxicity testing. Cell Biol Toxicol. 2013;29:239–257. doi:
10.1007/s10565-013-9250-5.
Rahnenführer J, Leist M. From smoking guns to footprints: mining for critical events of toxicity pathways
in transcriptome data. Arch Toxicol. 2015;89:813–817. doi: 10.1007/s00204-015-1497-6.
Saba R, Störchel PH, Aksoy-Aksel A, et al. Dopamine-regulated microRNA MiR-181a controls GluA2
surface expression in hippocampal neurons. Mol Cell Biol. 2012;32:619–632. doi:
10.1128/MCB.05896-11.
Schildknecht S, Pöltl D, Nagel DM, et al. Requirement of a dopaminergic neuronal phenotype for toxicity
of low concentrations of 1-methyl-4-phenylpyridinium to human cells. Toxicol Appl Pharmacol.
2009;241:23–35. doi: 10.1016/j.taap.2009.07.027.
Schildknecht S, Karreman C, Pöltl D, et al. Generation of genetically-modified human differentiated cells
for toxicological tests and the study of neurodegenerative diseases. ALTEX. 2013;30:427–444. doi:
10.14573/altex.2013.4.427.
Schlachetzki JCM, Saliba SW, de Oliveira ACP, et al. Studying neurodegenerative diseases in culture
models. Rev Bras Psiquiatr. 2012;35:S92–S100. doi: 10.1590/1516-4446-2013-1159.
Schmittgen TDT, Livak KJK. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc.
2008;3:1101–1108. doi: 10.1038/nprot.2008.73.
Scholz D, Pöltl D, Genewsky A, et al. Rapid, complete and large-scale generation of post-mitotic neurons
from the human LUHMES cell line. J Neurochem. 2011;119:957–971. doi: 10.1111/j.1471-
4159.2011.07255.x.
Selenica M-L, Jensen HS, Larsen AK, et al. Efficacy of small-molecule glycogen synthase kinase-3
inhibitors in the postnatal rat model of tau hyperphosphorylation. Br J Pharmacol. 2007;152:959–979.
doi: 10.1038/sj.bjp.0707471.
184
Smirnova L, Sittka A, Luch A. On the role of low-dose effects and epigenetics in toxicology. EXS.
2012;101:499–550.
Smirnova L, Block K, Sittka A, et al. MicroRNA profiling as tool for in vitro developmental neurotoxicity
testing: the case of sodium valproate. PLoS ONE. 2014;9:e98892. doi: 10.1371/journal.pone.0098892.
Smirnova L, Seiler AEM, Luch A. microRNA profiling as tool for developmental neurotoxicity testing
(DNT) Curr Protoc Toxicol. 2015
Smirnova L, Harris G, Leist M, Hartung T. Cellular resilience. ALTEX. 2015;32(4):247. doi:
10.14573/altex.1509271.
Srikanth P, Young-Pearse TL. Stem cells on the brain: modeling neurodevelopmental and
neurodegenerative diseases using human induced pluripotent stem cells. J Neurogenet. 2014;28:5–29.
doi: 10.3109/01677063.2014.881358.
Stępkowski TM, Wasyk I, Grzelak A, Kruszewski M. 6-OHDA-induced changes in Parkinson’s disease-
related gene expression are not affected by the overexpression of PGAM5 in in vitro differentiated
embryonic mesencephalic cells. Cell Mol Neurobiol. 2015
Stiegler NV, Krug AK, Matt F, Leist M. Assessment of chemical-induced impairment of human neurite
outgrowth by multiparametric live cell imaging in high-density cultures. Toxicol Sci. 2011;121:73–87.
doi: 10.1093/toxsci/kfr034.
Tal TL, Tanguay RL. Non-coding RNAs: novel targets in neurotoxicity. Neurotoxicology. 2012;33:530–
544. doi: 10.1016/j.neuro.2012.02.013.
Tanner CM, Kamel F, Ross GW, et al. Rotenone, paraquat, and Parkinson’s disease. Environ Health
Perspect. 2011;119:866–872. doi: 10.1289/ehp.1002839.
Todorovic M, Newman JRB, Shan J, Bentley S, Wood SA, Silburn PA, Mellick GD. Comprehensive
assessment of genetic sequence variants in the antioxidant ‘master regulator’ nrf2 in idiopathic
Parkinson’s disease. PLoS ONE. 2014;10:e0128030. doi: 10.1371/journal.pone.0128030. van Vliet EE, Morath SS, Eskes CC, et al. A novel in vitro metabolomics approach for neurotoxicity
testing, proof of principle for methyl mercury chloride and caffeine. Neurotoxicology. 2008;29:1–12.
doi: 10.1016/j.neuro.2007.09.007.
185 van Vliet E, Daneshian M, Beilmann M, et al. Current approaches and future role of high content imaging
in safety sciences and drug discovery. ALTEX. 2014;31:479–493.
Volbracht C, van Beek J, Zhu C, et al. Neuroprotective properties of memantine in different in vitro and in
vivo models of excitotoxicity. Eur J Neurosci. 2006;23:2611–2622. doi: 10.1111/j.1460-
9568.2006.04787.x.
Wang A, Costello S, Cockburn M, et al. Parkinson’s disease risk from ambient exposure to pesticides. Eur
J Epidemiol. 2011;26:547–555. doi: 10.1007/s10654-011-9574-5. [PMC free article] [PubMed] [Cross
Ref]
Wheeler HE, Wing C, Delaney SM, et al. Modeling chemotherapeutic neurotoxicity with human induced
pluripotent stem cell-derived neuronal cells. PLoS ONE. 2015;10:e0118020. doi:
10.1371/journal.pone.0118020.
Yamaguchi Y. Lecticans: organizers of the brain extracellular matrix. Cell Mol Life Sci. 2000;57:276–289.
doi: 10.1007/PL00000690.
Yang D, Li T, Wang Y, et al. miR-132 regulates the differentiation of dopamine neurons by directly
targeting Nurr1 expression. J Cell Sci. 2012;125:1673–1682. doi: 10.1242/jcs.086421.
Zhang X-M, Yin M, Zhang M-H. Cell-based assays for Parkinson’s disease using differentiated human
LUHMES cells. Acta Pharmacol Sin. 2014;35:945–956. doi: 10.1038/aps.2014.36.
186
CHAPTER 5
5. TOXICITY, RECOVERY AND RESILIENCE IN A 3D
DOPAMINERGIC IN VITRO MODEL EXPOSED TO ROTENONE.
*The manuscript presented in this chapter has been submitted to Archives of Toxicology.
Harris G, Eschment M, Orozco S.P, McCaffery J.M, Maclennan R, Severin D, Leist M,
Keensang A, Pamies D, Maertens A, Hogberg H.T, Freeman D, Kirkwood A, Hartung T and
Smirnova L. Toxicity, recovery and resilience in a 3D dopaminergic in vitro model exposed to
Rotenone.
Key points
• LUHMES 3D model can be used to study acute effects, recovery after compound wash-
out and susceptibility to second exposures.
• Acute effects of rotenone include complex I inhibition, decreased ATP production,
inhibition of neurite outgrowth and altered mitochondrial morphology.
• After compound wash-out and 7-day recovery functianly, aggregates recover showing
ATP levels, neurite length, and mitochondrial morphology similar to vehicle controls.
However, complex I activity remains inhibited.
• Microarray analysis identified few genes which remain altered after compound removal.
• Pre-exposed aggregates were resilient to the second exposure to rotenone compared to
aggregates exposed for the first time, indicating molecular memory.
187
5.1. ABSTRACT
To date, most in vitro toxicity testing has focused on acute effects of compounds at high concentrations. This testing strategy does not reflect real-life exposures, which might contribute to long-term disease outcome. We used a 3D-human dopaminergic in vitro LUHMES cell line model to determine whether molecular effects of short-term exposure to rotenone (100 nM, 24 h), a known Parkinson’s disease inducer, are permanent or reversible. A decrease in complex I activity, ATP, mitochondrial diameter and neurite outgrowth were observed acutely. After compound removal, complex I activity was still inhibited, however, ATP levels were increased, cells were electrically active and aggregates restored neurite outgrowth function and mitochondrial morphology. We identified significant transcriptomic changes after 24 h exposure which were not present 7 days after wash-out. Our results suggested that testing short-term exposures in vitro may capture many acute effects, which cells can overcome and miss adaptive processes and long-term effects involved in disease progression. Additionally, to study cellular resilience, cells were re-exposed to rotenone after wash-out. Pre-exposed cells maintained higher metabolic activity than controls and presented a different expression pattern in genes previously shown to be altered by rotenone. NEF2L2, ATF4 and EAAC1 were downregulated upon single hit on day 15, but unchanged in pre-exposed aggregates. DAT and CASP3 were only altered after re- exposure to rotenone while TYMS and MLF1IP were downregulated in both single-exposed and pre-exposed aggregates. In summary, our study shows that a human cell-based 3D model can be used to assess cellular adaptation, resilience and long-term changes pertaining to neurodegeneration.
188
5.2. INTRODUCTION
Parkinson’s disease (PD) is a consequence of the interplay between genetic and environmental factors affecting dopaminergic neurons (Kalia et al. 2015). Although they account for less than
1% of neurons in the brain, their degeneration and loss leads to neurodegenerative diseases such as PD (Chinta and Andersen 2005). This progressive neurodegenerative disorder is characterized by the pathological feature of excessive dopaminergic neuron loss in the midbrain (substantia nigra pars compacta), and in most cases, intra-cytoplasmic inclusions in intact neurons called
Lewy Bodies (Jankovic et al. 2008). It is estimated that by 2030 9 million individuals will be affected worldwide (Dorsey et al. 2007). While some known genetic factors play a role in early onset of familial PD, monogenic forms only account for ~10% of patients. Around 90% of PD cases are sporadic and represent the interplay of genetic risk and environmental factors (age, stress) or exposures (suspected are pesticides, flame retardants, metals etc.) contributing to PD risk (Beclin and Westerlund 2008). To date, only symptomatic treatment for patients suffering from PD is available, and there are no accepted neuroprotective or neuro-restorative therapies
(Ortel et al. 2017).
Several compounds are known to interfere with normal function of dopaminergic neurons; examples include pesticides (e.g. rotenone, paraquat and maneb) (Gorell et al. 1998) and flame retardants (Ahmed et al. 2017, Claudle et al. 2012). Epidemiological studies have demonstrated that rotenone use can increase the risk of developing PD by >2.5 fold (Tanner et al. 2011).
Complex I (NADH-ubiquinone oxidoreductase) of the electron transport chain is the molecular target for some compounds shown to induce PD-like symptoms (rotenone and MPP+, the active metabolite of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)) in animals and humans
(Sherer et al. 2007) with patients showing complex I deficiency in the substantia nigra (Shapira et al. 1989, Parker et al. 2008). Other reported cellular events include mitochondrial dysfunction
(decreased ATP production, decreased membrane potential), oxidative stress, impaired
189 proteostasis and accumulation of misfolded proteins, dopaminergic cell degeneration, reduced dopamine release and neuro-inflammation (https://aopwiki.org/aops/3) (Bal-Price et al. 2017,
Keane et al. 2011, Terron et al. 2018).
Rotenone has been widely studied as one of the best-known PD-inducing model compounds; it is extremely lipophilic, and freely crosses cellular membranes independently of any transporters.
Rotenone toxicity exhibits disparate symptoms. It has been shown to bind irreversibly to complex
I of the respiratory chain (Öberg et al. 1961, Grivennikova et al. 1997), and in vitro and in vivo studies have demonstrated that binding is necessary to reproduce PD mechanisms such as ROS accumulation (Sherer et al. 2003, Dhillon et al. 2008, Furlong et al. 2015). Systemic rotenone exposure has become a widely used animal model of PD (Cannon et al. 2009), with the number of
PD animal studies on a continuous rise (Daneshian et al. 2015). It is calculated, that the concentration required in the animal brain to induce PD pathology is ~20-30 nM ‘free’ rotenone
(Greenamyre et al. 2003). It has also been shown that rotenone can bind non-specifically to proteins other than complex I; therefore, it is considered that higher concentrations (> 30 nM) could have off-target effects (Higgins et al. 1996, Grefte et al. 2015).
In vitro models to study PD are needed as animal testing is demanding in terms of animal use, resources and time; they also suffer from inter-species differences making them not always predictive for human health (Olson et al. 2000, Hartung 2013). To overcome these problems, the use of human cell lines can avoid some of these issues, however, the complexity of the central nervous system represents a major challenge for in vitro models. Current in vitro models (cancer cell lines, immortalized cell lines, primary cell cultures or stem cells) offer the advantage of a controlled environment to study molecular pathways involved in PD development (Hogberg et al.
2013, Falkenburger et al. 2016). Understanding the limitations of each model is important to determine whether it can answer the question being posed (Schmidt et al. 2017). Although iPSC- derived 3D models would be the most representative of the human brain due to their multi- cellular composition (Pamies et al. 2017b), their complexity makes it difficult to attribute
190 mechanisms to the respective cell-type. Single-cell type models, differentiated in 3D, can therefore provide a tool to study cell-specific toxicant-induced disease mechanisms. It has been shown that many 3D cultures exhibit increased survival and enhanced neuronal differentiation compared to ones cultured in monolayer (Pamies et al., 2017a, Smirnova et al. 2016, Alepee et al.
2014). The use of in vitro models provides a tool to understand the mechanisms, by which environmental exposures lead to PD as well as to study neuroprotective pathways, which is critical also to identify biomarkers for early diagnosis and therapy.
LUHMES (Lund Human Mesencephalic) is a conditionally-immortalized cell line, which overexpresses tetracycline-controlled v-myc (Lotharius et al. 2005, Scholz et al. 2011). These cells are suitable as a dopaminergic-cell model as they homogeneously differentiate, are electrically active and express functional dopamine transporter (DAT), vesicular monoamine transporter (VMAT-2) and the PD-related protein α-Synuclein (ASYN) (Schildknecht et al.
2009). Furthermore, the LUHMES 3D model that we have developed can be kept in culture for up to 21 days and is suitable for long-term and wash-out studies. These 3D aggregates are cultured in suspension, therefore can be washed and transferred to a new culture plate after exposure. Compounds that easily bind to plastic such as rotenone can be washed out more effectively than in monolayer cell models (Smirnova et al. 2016, Harris et al., 2017).
A question that has yet to be addressed, to a greater extent, in in vitro toxicology is cellular recovery and resilience (Smirnova et al. 2015). Cellular resilience is a complex cellular mechanism, which to date has been mostly studied within infectious diseases (Richardson 2016) as well as neuroprotection after trauma and plasticity. Some recent studies address neuronal processes of reverting "back to normal" and reversal of apoptosis ("anastasis") (Manji et al. 2000,
Tyagi et al. 2015, Pfau et al. 2015). Our hypothesis is that cells can overcome low-dose toxicant effects (in which cell death is not triggered), and then, become either resilient or more susceptible to subsequent exposures (via activation of cell survival/death pathways, changes in gene expression or epigenetic modulations) (Smirnova et al. 2015). One hypothesis that has yet not
191 been tested is whether resilience mechanisms are beneficial or detrimental to cells in the long- term perspective as permanent activation of these pathways may contribute to disease pathology
(Daskalakis et al. 2013, Pfau et al. 2015). In neurodegenerative diseases, the final steps of disease manifestation have been well characterized in human post-mortem samples and in vivo studies.
However, early mechanisms linking environmental exposures to disease are still unknown.
Although some in vitro studies have focused on low-dose, chronic exposures to toxicants showing long-term lesions (Sherer et al. 2002, Drolet et al. 2009), recovery and resilience to neurotoxicity has not been addressed in depth.
192
5.3. MATERIALS AND METHODS
A detailed description of materials and methods can be found in Supplementary Methods.
LUHMES 3D culture and differentiation
LUHMES (ATCC® CRL_2927™) 3D cell culture and differentiation protocol was followed as described (Harris et al. 2017). Briefly, cells were used between passages 15 and 25. 4x106 cells were placed in a 175 cm² flask for 48 hours to expand cells. On day 0, 3D-differentiation was initiated: 5.5x105 cells were seeded into each well of a 6-well plate and placed on a gyratory shaker at 80 rpm (50 mm orbit) in an incubator at 37°C, 10% CO2 and 95% humidity.
Toxicant treatment and wash-out
To study delayed effects of toxicant treatment and resilience, aggregates were exposed to rotenone or DMSO (vehicle control) for 24 h with subsequent wash-out of these compounds. A concentration of 100 nM rotenone was chosen for treatment as this concentration represented the lowest observed adverse effect level (LOAEL) for viability in LUHMES (Krug et al. 2014,
Smirnova et al. 2016). Rotenone was dissolved in 100% DMSO at a stock concentration of 100 mM (aliquoted and stored at -20°C). Experiments were performed in 6-well plates. Aggregates were treated with 100 nM on day 7 of differentiation. 24 hours later, on day 8, half of the samples were collected and, in the remaining cultures, rotenone was washed out as described before
(Harris et al. 2017). Importantly, in the wash-out experiments, the aggregates were transferred to new cell culture plates to end also exposure to rotenone, which might stick to the plastic. Media was changed every other day up to day 15.
Viability assays
193
Resazurin assay was performed as described in Harris et al., 2017. Experiments were performed in three independent experiments with technical triplicates. LDH was measured in the media in control and treated samples following manufacturer’s instructions (CytoTox 96® NonRadioactive
Cytotoxicity Assay, Promega).
DNA quantification
Aggregates were lysed and DNA extracted using Phenol:Chloroform:Isoamyl (24:25:1, Sigma) extraction. DNA quantification was performed using the Qubit dsDNA Broad Range Assay Kit
(Invitrogen) and Qubit 2.0 Fluorometer (Invitrogen) according to manufacturer’s instructions.
RNA extraction, reverse transcription, and real-time PCR
Total RNA was extracted using either TRIzol® Reagent (Life Technologies) followed by RNA
Clean & ConcentratorTM-Kit (Zymo Research®) or mirVana microRNA isolation kit (for microarray analysis) following the manufacturer’s instructions. Detailed description of cDNA synthesis and PCR is described in Supplementary Methods. Primers used for PCR are listed in
Supplementary Table S1.
Microarray analysis
Microarray analysis was conducted at The Johns Hopkins Bloomberg School of Public Health
Genomic Analysis and Sequencing Core Facility. RNA was extracted from three samples per condition of LUHMES aggregates on day 8 and 15 using the mirVana miRNA Isolation kit
(Ambion/Thermo Fisher Scientific) according to the manufacturer’s protocol. Following elution of purified RNA from the mirVana miRNA columns with Nuclease-free water with RNasin, quantitation was performed using a NanoDrop spectrophotometer and quality assessment determined by RNA LabChip analysis on an Agilent BioAnalyzer 2100 or RNA Screen tape on
194 an Agilent TapeStation 2200. One hundred nanograms of total RNA was processed for hybridization to Agilent SurePrint G3 Human Gene Expression v2 8x60K Arrays according to
Agilent’s One Color Microarray-Based Analysis (Low Input QuickAmp Labeling) protocol, including cDNA synthesis, cRNA synthesis with Cy3-labeling and purification, fragmentation, hybridization, and washing. Spike-in controls were utilized and processed according to Agilent’s
One-Color RNA Spike-In kit protocol.
The arrays were scanned in the Agilent G2600D SureScan Microarray Scanner using scan protocol AgilentG3_GX_1color for gene expression arrays. Agilent’s Feature Extraction
Software Version 11.5.1.1 was used to assign grids, provide raw image files per array, and generate QC metric reports from the microarray scan data. The QC metric reports were used for quality assessment of all hybridizations and scans.
Txt-files from Feature Extraction Software were exported for further analysis with R version
3.4.2 (https://www.R-project.org/) and Bioconductor version 3.6 (Huber et al. 2015, Gentleman et al. 2004). In a first step, the arrays gMedianSignal were imported, normexp background corrected
(Ritchie et al. 2007, Silver et al. 2009)and quantile normalized between arrays (Bolstad et al.
2003). Primary QC by principal components analysis revealed a batch effect on a subset of arrays, which were processed at a different time point with a different washing procedure, which was corrected by parametric empirical Bayes frameworks for adjusting data for batch effects as implemented in ComBat (Johnson et al. 2007).
Probes were than filtered out if they are not least 10% brighter than the 95% percentile of the negative control probes on each array on at least three arrays (original 62,976 probes, after 54,135 probes). In a next step control probes were filtered out (after 51,849 probes) and duplicate probes were summarized (after 44,414 probes). Individual probes which were either labeled by the
Agilent Feature Extractor Software as not to be used, Non-uniform outlier or Population outlier were removed as well (174 probes over all arrays).
195
In a last step, probes, which did not map to Entrez-ID, were removed (32,123 probes left) and probes were averaged per Entrez-ID (22,150 unique Entrez-IDs left).
Differential expression was estimated by empirical Bayes moderation of the standard errors towards a common value (Empirical Bayes moderated t-test) (Smyth et al. 2004). The transcriptomics microarray datasets have been deposited in the Gene Expression Omnibus (GEO).
GeneOntology and Pathways Over-Representation Analysis were performed using clusterProfiler
(Yu et al. 2012) based on FC > 1.5 and p(FDR) < 0.05 for the day 8 data set and on FC > 1.5 and p (not adjusted) < 0.01 for the day 15 data set.
Complex I activity assay
Mitochondria Isolation was performed on ice using the reagent-based method (Mitochondria
Isolation Kit for Tissue and Cultured Cells, BioVision); Complex I activity using mitochondrial
Complex I Activity Colorimetric Assay Kit (BioVision) following manufacturer’s instructions.
See Supplementary Materials for details. Activity was measured in three independent experiments in technical duplicates.
ATP assay
The bioluminescence ATP Assay Kit (Thermo Fisher Scientific, A22066) was used to determine the amount of intracellular ATP in aggregates according to manufacturer’s instructions. See
Supplementary Materials for details. Average luminescence values ± SEM were calculated from at least four independent experiments with technical triplicates.
Electron microscopy and mitochondria quantification
Reagents were bought from Electron Microscopy Sciences (Fort Washington). 3D LUHMES aggregates were fixed using a solution of 3.0% formaldehyde, 1.5% glutaraldehyde contained in
196
100 mM sodium cacodylate, 5 mM Ca2+ and 2.5% sucrose at pH 7.4 for one hour at room temperature. Subsequently, samples were washed three times for 15 minutes using a solution of
100 mM cacodylate containing 2.5% sucrose at pH 7.4; post-fixed with Palade’s OsO4 at 4°C; and rinsed 1X using Kellenberger UA (uranyl acetate) and left in UA at RT overnight in the dark.
Samples were then dehydrated through a graded series of ethanol (50%, 70%, 95%, and 100%) at
4°C; followed by three 15 min washes in fresh 100% ethanol at RT. Following, two 5 min exchanges with propylene oxide (PO), samples were placed in a mixture of 50:50 Epon/propylene oxide and left overnight, uncovered, under vacuum. The resin mixture was replaced with fresh
100% Epon and left under vacuum an additional 4 - 6 h; and subsequently polymerized in an oven at 60°C for 24 to 48 hours. 80 nm sections were then cut on a Leica UCT ultramicrotome and placed on 400 mesh copper grids. Samples were imaged using a Philips EM 420 transmission electron microscope. Images were collected with a Megaview III side-entry camera from
Olympus Soft Imaging Systems (OSIS); and mitochondrial area and diameter assayed using iTEM software (also available from OSIS). Quantification was performed by selecting 20 random images from either vehicle control or treated, ranging from low (3,300x) to high (31,000x) magnification in three independent experiments. The square area was measured for the entire image excluding the grid bars, when present. The mitochondria in each image were counted; and discrimination of healthy vs unhealthy was assessed based on the appearance of the mitochondrial matrix density. Using a straight-line measurement tool, the length from the narrowest part of the mitochondria was measured and used as the diameter.
Neurite outgrowth imaging and analysis
Red fluorescent protein (RFP) expressing-LUHMES (Schildknecht et al. 2013) were differentiated and treated as described above. On day 8 or day 15, aggregates were seeded on
MatrigelTM (BD Biosciences) pre-coated, flat-bottom, black 24 or 96-well plates (Thermo Fisher
Scientific). After 24 hours, wells were fixed in 4% PFA and imaged using a confocal microscope
197
(with open pinhole) and analyzed using Sholl Image J Software
(https://imagej.net/Sholl_Analysis). To analyze this data, the ratio was calculated for each shell
(number of intersections/distance from aggregate) and the mean plotted. Curves were compared using a quadratic non-linear regression fit with confidence intervals.
Electrical activity
Whole cell recordings were performed under a DIC microscope (eclipse E600FN, Nikon). 3D
LUHMES were transferred to the recording chamber with culture media at 37oC. Every 30 min, the media were replaced. To target whole cell recordings, 3D LUHMES aggregates were attached to a glass pipette by means of a gentle negative pressure, which was released once the aggregate was attached. Cells were visualized at high magnification (40X objective, water immersion) and chosen with respect to their morphological phenotype (small, round, phase-bright cell bodies).
Patch pipettes (4 – 5 MΩ resistance) made of borosilicate glass were filled with an internal solution containing 130 mM K-gluconate, 10 mM KCl, 0.2 mM EGTA, 10 mM HEPES, 0.5 mM
Na3GTP, 4mM MgATP and 10 mM Na-Phosphocreatine (pH adjusted to 7.3 with KOH, 285–
295 mOsm). Once stable, whole-cell recordings were performed and basic electrophysiological properties were examined through depolarizing current injections. Electrophysiological data were acquired with a Multiclamp 700A amplifier (Molecular Devices), data acquisition board (model
PCI MIO 16E-4, National Instruments), and Igor Pro (Wavemetrics). Data were filtered at 4 kHz and digitized at 10 kHz. Minimal spike latency was measured using a single exponential fit for the spike latency versus the current injection strengths. Differences in treated and control samples were analyzed for statistical significance using Mann-Whitney U-test.
198
5.4. RESULTS
The LUHMES 3D cell model is suitable for recovery and resilience experiments
In this study, we used the previously developed and characterized 3D LUHMES in vitro model
(Smirnova et al. 2016) (Fig. 1). Cells (5.5x105) were seeded into each well of a 6-well plate, forming ~ 200 aggregates/well (250-300 µm in diameter), over 15 days of differentiation. By quantifying DNA, each aggregate was calculated to be composed of 3,000-5,000 cells
(Supplementary Figure S1). To determine whether LUHMES 3D in vitro cultures are capable of recovering from low-dose rotenone effects, we followed the same treatment protocol as previously reported (Harris et al., 2017, Smirnova et al., 2016) focusing on 24 h exposure to 100 nM rotenone. In addition to (i) acute cellular response on day 8 of differentiation (D8 24h); we studied (ii) delayed response and recovery on day 15 after compound wash-out and additional 7 days in culture (D15 wash-out); and (iii) cellular response to a second rotenone exposure after recovery on day 15 (D15 R100-R100). The treatment scheme for these conditions is depicted in
Fig. 1. A key advantage of the 3D culture is that in the wash-out experiments the aggregates can be transferred to clean cell culture plastic ware to ensure that the highly lipophilic rotenone is removed by washing. Using mass spectroscopy we quantified free rotenone in the media before treatment, 24h after treatment and 7 days after wash-out. These results demonstrate a decrease in free rotenone when 100 nM is added to cell culture plates in media, therefore 23.5 ng binds to plastic (Fig. 1d (bar I minus bar II), Fig. 1e). Furthermore, in the presence of cells, a further decrease in free rotenone was recorded; demonstrating 21.1 ng binds to cells within aggregates
(Fig. 1d (bar II minus bar III), Fig. 1e). No free rotenone was detected in media 7 days after wash-out.
In previous work, we have shown that 100 nM rotenone produces ~15 % decrease in viability (as measured by resazurin reduction assay) after 24 h exposure (day 8) and further 10~15 %
199 reduction until day 15 after compound wash-out (Smirnova et al., 2016 and Fig. 2a). The acute effect on cell viability has been shown also in other in vitro cultures (Sherer et al. 2003, Krug et al. 2014). To define when cytotoxicity occurs between day 8 and 15, we measured viability and
LDH release every 48 hours. The viability level on days 10 was similar to day 8. We identified that the further 10 % decrease in viability occurs between days 12 and 15 (Fig. 2a). Additionally, we observed a small but significant increase in LDH release on day 8 (Fig. 2b). However, from days 10 to 15 no significant increase in LDH was observed compared to controls. This indicated that the remaining cells could recover from the short-term rotenone exposure.
Since the resazurin assay specifically detects cellular metabolic activity, we studied whether the observed reduction was due to a decrease in metabolic activity or due to a lower cell number in rotenone-treated samples. For this reason, we measured DNA content (surrogate for cell number) and protein concentration, as a true measure of sphere viability, after rotenone wash-out and 7 days recovery. Both, DNA and protein levels were lower indicating some cells were lost after wash-out and recovery period (Fig. 2c and d).
ATP levels are increased and mitochondria recover after rotenone wash-out
Rotenone is a known complex I inhibitor which decreases ATP production (Sherer et al. 2003).
To better understand what changes are occurring at a molecular level in the remaining cells, we assessed complex I activity and ATP levels after 24 h exposure to 100 nM rotenone (D8 (24h)) and 7-days after wash-out (D15 (wash-out)). Mitochondria were extracted from control and treated samples, Complex I activity was measured in samples using a colorimetric assay and normalized to protein content. As expected, complex I activity was reduced after 24 h rotenone treatment; and remained inhibited after wash-out (Fig. 3a). ATP was measured in cell lysates using a luminescence assay (normalized to protein levels). In concordance with earlier acute in vitro reports (Sherer et al. 2003), a significant decrease in ATP levels (~30 %) was observed after
24 h exposure (Fig. 3b). However, on day 15 (after wash-out and recovery period), ATP levels
200 were significantly increased in rotenone-treated samples compared to controls, despite the decreased cell number and ~20% inhibition of complex I activity (Fig. 3b). Importantly, ATP levels and complex I activity were normalized to protein content. Thus, although there was a decrease in number of viable cells on day 15 after wash-out (Fig. 2), the remaining cells had increased ATP levels compared to vehicle-treated control samples.
As the observed effects on ATP levels could not be explained by changes in cell number, we examined mitochondria physiology by electron microscopy. There was no change in the number of mitochondria (Fig. 4a). Mitochondrial diameter was recorded to investigate whether shrinking or swelling was taking place after treatment and wash-out. After 24 h rotenone exposure, mitochondrial diameter was increased by ~27% on average (Fig. 4b and c, day 8 control avg =
0.33 µm; day 8 treated avg = 0.42 µm, *p=0.0378). After compound wash-out and 7-day recovery, mitochondria diameter was comparable to controls.
In summary, these data suggest that the cells remaining in the aggregates after acute exposure and the recovery period were able to compensate for complex I inhibition, recover ATP production and restore mitochondria morphology.
Neurons recover neurite outgrowth and electrical activity after rotenone wash-out
Next, we analyzed whether the remaining viable neurons are fully functional after the recovery period. We tested whether cells are still able to prolong their neurites when given space and an appropriate stimulus. For this reason, the aggregates were plated on Matrigel®, a condition favoring neurite outgrowth from aggregates. Aggregates were plated either immediately after rotenone treatment (D8 (24h)) or after a 7-day recovery period (D15 (wash-out)) and outgrowth was quantified using the Image J Sholl image analysis (Fig. 5, Figure S2). Our results show that acute exposure (100 nM, 24 h) decreased the number and length of neurites (R100 D8 slope = -
0.1226 ± 0.003 vs. DMSO D8 slope = -0.1822 ± 0.006). Aggregates plated after the 7-day
201 recovery period, showed no differences in number or length when compared to control samples
(R100 D15 slope = -0.245 ± 0.015 vs. DMSO D15 slope = -0.279 ± 0.016) (Fig. 5b).
LUHMES monolayer cultures have been shown to be electrically active (Scholz et al. 2011). As a further functional endpoint, we used whole-cell patch clamp recording to evaluate whether (i)
LUHMES 3D culture contains electrically active cells, (ii) electrical activity is affected by treatment and (iii) it can recover after compound wash-out. Both tonic and phasic modes of activity were identified in LUHMES aggregate cells (Fig. 6b). On day 15, no differences were observed in the number of tonic vs. phasic cell types (Fig. 6c). We focused on the physiological properties of phasic cells. No changes in the input resistance (Fig. 6d) or spike latency (Fig. 6e) were detected in phasic neurons after the treatment, confirming that there were no delayed effects of rotenone on electrical activity in measured cells.
In summary, these data suggest that the functionality of the cells was restored after the acute rotenone effects, wash-out and the recovery period.
Acute transcriptomic changes are overcome after recovery period
We further analyzed the effects of 100 nM rotenone exposure on the LUHMES aggregates transcriptome 24 hours after treatment (D8 (24h)) and after compound wash-out and 7-day recovery period (D15 (wash-out)). We found 708 genes significantly changed on day 8 (FC >
1.5 & p (adjusted) < 0.05). On day 15, after multiple hypothesis testing correction, no significantly changed genes remained. Since we performed low-dose short-term exposure and compound wash-out, we did not expect dramatic changes in gene expression on day 15, especially taking into account that the functional endpoints, described above, indicated recovery.
However, previously we could observe some slight changes in gene expression on day 15 by qPCR (Smirnova et al. 2016). Also, because qPCR is more sensitive than microarray method and the FDR correction of a big data set (over 20,000 genes) with a small sample size (three replicates per condition) may hide slight but still significant changes, we used unadjusted p-values for
202 further analysis. To be more stringent, we decreased the p-value cut-off for all microarray analysis (p < 0.01 vs. classically used p < 0.05). On day 8, 809 genes were significantly changed, with 343 upregulated and 466 downregulated genes (Supplemental Table S2). On day 15, a significantly lower number of genes were perturbed (107, FC > 1.5, p < 0.01) with 52 up- and
55 downregulated genes (Fig. 7a and b, Supplemental Table S3), There were 10 genes in the intersection of day 8 and day 15 (Fig. 7b and c). The same analysis was performed for samples exposed to 50 nM for 12 or 24 hours showing less of an effect with the lower concentration and shorter exposure time, as expected (Supplemental Tables S4-S7 and Figure S3).
The genes from day 8 (R100, 24 h) with an FDR corrected p-value of less than 0.05 (1516 genes total) as well as the genes from day 15 with an uncorrected p-value of 0.05 (1092) genes) were examined for enrichment analysis; samples were highly enriched for genes related to neurogenesis, oligodendrocyte differentiation as well as genes associate with Alzheimer’s on day
8 (Supplemental Table S8). After wash-out and recovery on day 15, samples were enriched for neurogenesis as well as plasma membrane components, and CNS development (Supplemental
Table S9). Additionally, both gene sets from day 8 and day 15 were explored for potential interactions via the STRING database; both were significantly enriched for known protein interactions. The resulting network from day 8 had TOP2B as the main “hub” (most highly connected protein). The switch from TOP2A to TOP2B is known to be critical for neuronal differentiation in vitro and in vivo; additionally TOP2B appears to selectively occupy regulatory regions in the genome where it modulates the transcription of genes involved in neuronal survival
(https://doi.org/10.1073/pnas.1119798109). (TOP2B subnetwork, Supplemental Figure S4a.)
Of the genes in common between day 8 and day 15, CCK was consistently highly connected (i.e. a “hub”) in both interaction networks. The role of CCK in the brain is poorly understood, however, in this dataset, the subnetwork of CCK for day 8 and day 15 were both highly enriched
203 for genes on the pathway for non-odorant GPCR (GPCR, class A, Rhodopsin-like), (corrected p- value of 2.072 *10-15 Day 8; .000008884 Day 15) (day 8 Supplemental Figure S4b, day 15
Supplemental Figure S4c). Given the relatively weak signal from the transcriptomics data from day 15, it is difficult to draw firm conclusions, but the data suggest there is a persistent alteration in non-odorant g-protein coupled receptors mediated in part by CCK signaling.
Metabolic resilience is observed with second exposure to rotenone after recovery
After observing that aggregates compensated for the inhibition of complex I and functionally recovered after the first insult, we tested our resilience hypothesis (Smirnova et al. 2015), by measuring susceptibility of pre-exposed 3D LUHMES to a second exposure to rotenone. After the
7-day recovery from 100 nM rotenone treatment, aggregates were re-exposed to increasing rotenone concentrations (0 – 10 µM). Control LUHMES aggregates, exposed to rotenone for the first time on day 15 (Control), showed a similar dose-response (Fig. 8a) to that previously observed on day 8 (published data Smirnova et al. 2016) indicating that the dopaminergic cell response to rotenone in this model does not change between days 8 and 15 of differentiation. In contrast, aggregates pre-exposed to 100 nM rotenone on day 8 and re-exposed on day 15 (pre- exposed), showed a significant increase in viability/metabolic activity at concentrations between
100 nM and 1 µM (Fig. 8a) compared to controls. To determine, whether observed effects were concentration-dependent, we pre-treated the aggregates on day 8 for 24 h with 25 and 50 nM rotenone. Aggregates pre-exposed to 50 nM also showed increased viability compared to DMSO controls at 316 nM (Figure 8b), while those pre-exposed to 25 nM were more similar to controls in response to the second hit (Supplementary Figure S5). Thus, we conclude that the observed effect was also concentration-dependent and exposure to 50 nM and 100 nM, but not 25 nM rotenone, led to increased viability (metabolic activity) upon a second hit. Pre-exposed samples
(100 nM) also showed lower level of released LDH than controls at higher concentrations (46 to
1000 nM) (Fig. 8c), confirming resilience of pre-exposed aggregates to the second hit.
204
A second exposure elicits a different gene expression pattern
To further understand the changes taking place after a second exposure to rotenone, we measured changes in expression of genes previously shown to be altered by rotenone (Krug et al. 2014,
Smirnova et al. 2016) as well as genes specific for dopaminergic neurons and PD. Gene expression was analyzed in three independent experiments for the following conditions: control cultures, never exposed to rotenone (DMSO-DMSO); cultures exposed to 100 nM rotenone on day 14 for the first time (DMSO-R100); cultures pre-exposed to 100 nM rotenone on day 7 and re-exposed on day 14 (R100-R100). Samples for qPCR were collected on day 15. NEF2L2, ATF4 and EAAC1 were significantly downregulated when aggregates were exposed to rotenone for the first time on day 14 (Fig. 8d DMSO-R100). However, no perturbation in expression of those genes was observed if aggregates were pre-exposed to rotenone on day 8 (Fig. 8d R100-R100).
On the contrarily, expression of DAT and CASP3 was not perturbed in aggregates exposed acutely to rotenone only on day 14 (Fig. 8e DMSO-R100), but was significantly changed in pre- exposed samples (Fig. 8e, R100-R100). Similar downregulation of TYMS and MLF1IP
(previously shown to be downregulated by both rotenone and MPP+) was observed in both conditions whether aggregates had been pre-exposed or not (Fig. 8f). These data showed that the genetic response varied depending on whether aggregates were exposed for the first or second time.
205
5.5. TABLES AND FIGURES
Figure 5-1. LUHMES 3D model for acute, recovery and resilience experiments. (a) LUHMES differentiated in 3D on a gyratory shaker showing (b) RFP-expressing cells (RED) and TH (green). (c) LUHMES 3D treatment and wash-out scheme for recovery and resilience (second hit) experiments and endpoints. (d)
206
Media rotenone quantification prior to treatment, day 8 and day 15. From left to right bars correspond to negative control (media without rotenone), positive control (media with rotenone prior to treatment), 24 h treatment control (media with rotenone in plates), 24 h treatment (media with rotenone in plates with aggregates), 7-day wash-out control (media with rotenone in plates on day 15 after wash-out) and 7-day wash out treated cells (media with rotenone in plates on day 15 after wash-out with aggregates). (e) Amount of rotenone bound to plastic and cells after 24 h exposure (day 8).
Figure 5-2. 3D LUHMES viability after wash-out. (a) Cell viability measured over time using resazurin assay on days 8 (after 24h treatment) and 10, 12, and 15 (throughout recovery). (b) Cytotoxicity over time during recovery measured by LDH release on days 8 (after 24h treatment) and 10, 12, and 15 (throughout recovery). (c) Protein concentration on day 15 after wash-out and 7 days recovery. (d) DNA quantification on day 15, after wash-out and 7 days recovery. All data were normalized to untreated control cells and are displayed as means ± SEM from 3 independent experiments. *p < 0.05
207
Figure 5-3. Effects of rotenone on complex I activity and ATP levels. (a) Complex I activity after rotenone exposure (day 8) and after compound wash-out and recovery (day 15) in control and treated samples. (b)
ATP levels after rotenone exposure (24 h, day 8) or after wash-out and 7 days recovery period (day 15) in control and treated samples. Differences in treated and control samples from at least three independent experiments were analyzed for statistical significance using unpaired Student’s t-test. A p-value < 0.05 is denoted on graphs by * and p < 0.0001 by ****, respectively
208
Figure 5-4. TEM analysis of mitochondria after rotenone exposure and wash-out. Legend: M, mitochondria;
G, Golgi complex; L, lipid droplets; N, nucleus; and NN, neurite. The number (a) and diameter (b) of mitochondria from random image areas were quantified on day 8 (24h) and day 15 (wash-out). Data from 20 random images from three 3 independent experiments is shown as well as means ± SD. Differences between treated and untreated samples were analyzed for statistical significance using unpaired Student’s t-test. A p
Value < 0.05 is denoted on the graphs by *. (c) Representative images are shown with arrows indicating morphological alternations to the mitochondrial membrane
209
Figure 5-5. Image J Sholl analysis of neurite outgrowth after rotenone exposure (day 8) and wash-out (day
15). RFP-LUHMES aggregates were grown on Matrigel® on day 8 or day 15. (a) Representative images for the different conditions are shown. (b) Sholl analysis (Image J) was used to calculate the number of neurites at different distances from the aggregate center on day 8 and day 15 from three independent experiments (5 individual aggregates per experiment). Curves were compared using a quadratic non-linear regression fit with confidence intervals
210
Figure 5-6. 3D LUHMES electrical activity on day 15 after acute exposure on day 8 and compound wash- out. (a) Photo microscopy image of a 3D LUHMES aggregate attached to a glass pipette and a patched cell at a higher magnification. Cells on different aggregates were patched in three independent experiments. (b)
Firing pattern of a representative tonic (top) and a phasic (middle) cell with voltage responses to 1 s current injections (bottom) at 4, 8, 12, 16, 20 and 24 pA, (c) Total number of tonic and phasic cells in control and treated samples on day 15 (p=0.695 two-sided Fisher's exact test), (d) the Input resistance (Rm) of the phasic cells (p=0.963 two-sided Mann-Whitney U Test) and (e) Minimal Spiking Latency of phasic cells (p=0.852 two-sided Mann-Whitney U Test). Error bars represent SEM.
211
Figure 5-7. Rotenone-induced transcriptome changes on day 8 (24h) vs. day 15 (wash-out). (a) Volcano plots show the global changes in transcriptome for day 8 and day 15. (b) Venn-diagram shows the number of up- and down-regulated genes on day 8 (D8 (24h)) and on day 15 (D15 (wash-out)) (FC > 1.5, p < 0.01). 10 genes were in intersection between two conditions, which are listed in (c). For this diagram, the p-values were not adjusted for multiple testing. ACTA1 – actin alfa 1, skeletal muscles, PPP1R27 - Protein phosphatase 1, regulatory subunit 27, GDF15 - Growth differentiation factor 15, CCK – cholecystokinin,
CD200 – OX-2 membrane glycoprotein, LCP1 – plastin 2 (lymphocyte cytosolic protein 1), ZFHX4 -AS1 –
ZFHX4 (Zinc Finger Homeobox 4) antisense RNA 1, FRMPD2 - FERM and PDZ domain containing 2,
FRMPD2 - FERM and PDZ domain containing 2, GRXCR1 - glutaredoxin and cysteine rich domain containing 1
212
Figure 5-8. (a) and (b) Cell viability concentration-response for aggregates on day 15, which were pre- exposed to DMSO (Control) or rotenone (pre-exposed 100 nM or 50 nM) on day 8. (c) LDH-release dose- response for aggregates on day 15 that were pre-exposed to DMSO (control) or rotenone (pre-exposed 100 nM) on day 8. Dose-response curves were generated from three independent experiments and analyzed by one-way ANOVA followed by Bonferroni’s correction. (d) NEF2L2, ATF4, EAAC1, (e) DAT, CASP3 and
(f) TYMS, MLF1IP gene expression by QT-PCR from three independent experiments analyzed using the
Student’s t-test and Bonferroni’s correction for multiple hypothesis testing. A p-value < 0.05 is denoted by *, p < 0.01 by **, and p < 0.001 by ***, respectively
213
5.6. DISCUSSION
Previously, we hypothesized that low-dose short-term stress can lead to several cellular outcomes
(death, recovery, resilience, or increased susceptibility) (Smirnova et al. 2015). If cell death is not
induced, cells that recover may have molecular signatures, which lead to resilience or
susceptibility to a second hit. With the experiments described here, we have demonstrated that 3D
LUHMES can restore functionality after the low-dose short-term (100 nM, 24 h) exposure to
rotenone. First, we measured free rotenone concentration in medium prior to treatment, after 24 h
treatment and 7 day post wash-out. These data showed that from 78.8 ng (100 nM), 23.5 ng
bound to plastic, and 21.1 ng bound to cells. We further confirmed that by transferring aggregates
to a new plate and performing a wash-out step on day 8, no further rotenone was present in the
culture medium on day 15 (Fig. 1d, e). Although rotenone could remain bound to cells, and this
would have to be measured to determine whether rotenone remains available to cells after wash-
out, there was no further exposure in to rotenone from the medium during recovery.
We observed an acute and delayed decrease in cell viability after compound wash-out (D8 82.6%,
D10 89.4%, D12 69.0%, D15 72.3%) (Fig. 2a). A 25% cell loss on day 15 was confirmed by
DNA and protein quantification (Fig. 2c, d). The focus of this study was to determine what occurs
to cells that survive the first exposure to rotenone. To assess whether surviving cells were
metabolically affected after the cell viability stabilizes, we assessed complex I activity, ATP
production, and mitochondria morphology after acute exposure and wash-out. Then, we assessed
whether functionality of the neurons could be restored after compound wash-out by measuring
outgrowth of neurites out from the aggregates. Delayed effects of rotenone on electrical activity
were assessed after wash-out and recovery. Acute and delayed effects of rotenone on gene
expression were analyzed by microarray. Finally, we measured the cellular response to a second-
rotenone exposure after the recovery period.
214
To study recovery, we selected the LOAEL for rotenone as reported in 2D and 3D LUHMES cultures (Krug et al. 2014; Smirnova et al. 2016). Although it is suggested that culturing in 3D increase survival vs. 2D cultures (Alépée et al. 2014), culturing LUHMES in 3D did not increase cell survival to rotenone exposure, likely because rotenone is lipophilic and can easily diffuse to the center of aggregates. A lower concentration (50 nM) was found to alter gene expression but had no effect on other molecular and functional endpoints (data not shown), therefore, was only included in microarray analysis. Our previous work also found no difference in day 15 viability after 12 or 24 h exposure on day 7 and subsequent wash-out, but further toxicity after 48 h exposure and subsequent wash-out; therefore, 24 h was selected as the most relevant time-point for this study (Smirnova et al. 2016). Rotenone was shown to inhibit mitochondria complex I and studies have demonstrated that this inhibition is necessary for dopaminergic toxicity (Sherer et al.
2003); while others have shown off-target effects prior to complex I inhibition (Choi et al. 2008).
Recovery and resilience have not yet been studied in a 3D in vitro human dopaminergic model, which is suitable for wash-out experiments (Smirnova et al. 2016; Alépée et al. 2014). Acutely, we observed a decrease in complex I activity (24 h) as observed in the previous studies (Choi et al. 2011; Richardson et al. 2005). Activity remained inhibited on day 15 (Fig. 3a), indicating that this effect was permanent after wash-out. We must note that this could be due to rotenone remaining bound within aggregates on day 15. An alternative explanation is damage to complex I or decreased expression of complex I subunits. Irreversible complex I inhibition has been shown previously (Lindahl and Öberg 1961). Acute reduction in ATP production has been well documented (Li et al. 2003; Sherer et al. 2003; Krug et al. 2014) and was confirmed in the present study, but in vitro models have not been able to demonstrate whether this is reversible. After rotenone wash-out, we observed an increase in ATP production on day 15, indicating that cells overcame complex I inhibition and have increased energy metabolism 7 days after compound removal (Fig. 3b). It is known that cells can shift from aerobic to anaerobic respiration to compensate for a decrease in ATP production in response to environmental stress (Zeiger et al.
215
2010). Zeiger et al. demonstrated that neurons will enhance ATP production following mild stress to survive. As multiple neuronal processes require ATP, an increase in ATP production may be necessary to recover cellular homeostasis in surviving cells. Dopaminergic neurons also have shown to have a large glycolytic spare capacity which could help to overcome lower ATP levels
(Delp et al. 2017).
Upon studying mitochondrial morphology, our experiments showed a reversible increase in mitochondria diameter (Fig. 4). We have previously reported acute loss of mitochondrial membrane potential (quantified using MitoTracker®; Smirnova et al. 2016) which could lead to fission defects. Studies have shown that inhibition of mitochondrial fission or promotion of mitochondrial fusion has protective effects in rotenone-induced neurotoxicity (Peng et al. 2017), and studies have documented rotenone-induced effects on mitochondrial trafficking and movement (Fang et al. 2016; Haddad and Nakamura 2015; Borland et al. 2008). Mitochondria undergo dynamic changes using fusion and fission to maintain function and morphology during stress (Knott et al. 2008). Imbalances in these mechanisms and fragmented mitochondria have been found in PD patients and only recently in vitro (Reddy 2008; Peng et al. 2017). The analysis of mitochondria number also allowed us to confirm that the differences observed in complex I activity and ATP were not due to changes in the number of mitochondria.
Measuring neurite outgrowth is a common functional endpoint to test adverse effects of compounds on neuronal cells (Stiegler et al. 2011; Scholz et al. 2011; Sun et al. 2016). Neurite outgrowth requires ATP; therefore, the decrease in ATP production could be the reason for impaired outgrowth observed on day 8 (Fig. 5). In addition, the production of reactive oxygen species, which has previously been reported at this concentration, likely also plays a role (Li et al.
2003; Han et al. 2014). The inhibition of complex I leads to electron leaking and a higher number of free electrons are, therefore, available to react with molecular oxygen to produce O2−. It has also been shown that oxidative stress, induced by rotenone (Sherer et al. 2003), increases microtubule disruption (Ren et al. 2005; Feng 2006; Choi et al 2011). As observed with ATP
216 production and mitochondria diameter, after the 7 day recovery period, neurite outgrowth was restored suggesting functional recovery, even while complex I remained inhibited.
LUHMES monolayer cultures have shown to be electrically active (Scholz et al. 2011), but this had not yet been studied in 3D LUHMES cultures. Patch clamp on day 15 revealed that 3D
LUHMES aggregates are made up of both phasic and tonic (Fig. 6) dopaminergic neuronal cell types. It has been shown that dopaminergic neurons can be either of these types resulting in different amounts of dopamine release in the striatum (Vandecasteele et al. 2005). There were a higher number of phasic cells, which were further analyzed, and no difference in input resistance
(Rm) or minimal spike latency (Fig. 6d, e) was observed between treated and control samples on day 15. Together, these functional endpoints confirm that acute effects on metabolic activity,
ATP production, mitochondria, and neurite outgrowth are reversible, and no delayed effects on electrical activity are observed after rotenone wash-out.
Functional endpoints were further confirmed with whole-genome microarray analysis (Fig. 7). As expected, the transcriptome was significantly perturbed (708 genes) on day 8, immediately after rotenone exposure, but after compound wash-out global gene expression was close to control (no significantly changed genes after FDR correction, 107 prior to FDR correction). Ten genes were altered on both day 8 and day 15, suggesting that the pathways that they are involved in are permanently perturbed. Most of those ten genes are enriched in the brain. At least two of the downregulated genes (CD200 and CCK) are strongly associated with PD pathology with a significant literature support (Wang et al. 2007; Nilsson et al. 2009). CD200 was shown to be downregulated in the substantia nigra of aging rats and blocking of the CD200 receptor significantly increased susceptibility of dopaminergic neurons to rotenone (Wang et al. 2011).
CD200 downregulation is associated with induced inflammation in PD, since this gene has anti- inflammatory and neuroprotective properties in dopaminergic neurons by inhibiting microglia activation and release of ATP and inflammatory factors (Ren et al. 2016). CCK is enriched in the brain (FANTOM5 atlas, http://fantom.gsc.riken.jp), which regulates release of dopamine that
217 affect dopamine-related behavior. Its polymorphism is associated with PD symptoms (Lenka et al. 2016; Fujii et al. 1999). ACTA1—actin alfa 1 skeletal muscles—was strongly upregulated on day 8 and still elevated on day 15. Although highly enriched in muscle tissue, ACTA1 is expressed in developing brain, especially in mesencephalon in various vertebrate species (Bertola et al. 2008). It is suggested to regulate axonal guidance, cellular motility, and cytoskeleton, and is a hub in the regulatory network of LRRK2, a high-risk PD gene (Dusonchet et al. 2014). Since we observed recovery in neurite outgrowth, overexpression of this gene may support this result.
GDF15—growth differentiation factor 15—is a secreted ligand of the TGF-beta (transforming growth factor-beta) superfamily of proteins. It is involved in the stress response after cellular injury. Elevation of GDF15 is associated with tissue hypoxia, inflammation, acute injury, and oxidative stress (Wiklund et al. 2010). It is precarious to over-interpret single gene changes, but the fact that the ultimately identified a few genes are consistently involved in PD and neuronal processes, stresses that there may be causal involvement.
From our results, we can conclude that although LUHMES 3D cultures were able to recover from acute rotenone exposure at molecular and functional levels, there was permanent complex I inhibition which cells need to adapt to. Several questions remain to be answered: what is threshold of complex I inhibition for which dopaminergic neurons, can compensate for? How long can cells overcompensate for the loss in aerobic respiration and maintain “normal” functionality? How detrimental could it be for the cells to maintain this response in the long term?
How do cells react to repeated exposures?
Yet, it is not clear why dopaminergic neurons are more susceptible to toxicity by compounds such as rotenone than other cell types (Haddad and Nakamura 2015; Schildknecht et al. 2017). Some hypotheses refer to the low number of dopaminergic neurons in the brain (~ 500,000 in healthy subjects) (Pakkenberg et al. 1991), axonal length (Surmeier et al. 2010), increased ATP demand
(Haddad and Nakamura 2015), and increased susceptibility to ROS and role of dopamine in ROS production (Gaki and Papavassiliou 2014). Based on our previous hypotheses (Smirnova et al.
218
2015), the next question that we posed was whether pre-exposed aggregates respond differently to a second exposure compared to controls, which have not previously been treated with rotenone.
From this experiment, two outcomes were possible: (1) the cells could become robust/resilient or
(2) more sensitive. To test our hypothesis, aggregates were washed and allowed to recover for 6 days, and then exposed a second time to increasing concentrations of rotenone for 24 h on day 14
(Fig. 8a). Viable mitochondria have a reducing environment due to NADPH or NADH being present (O’Brien et al. 2000). NADPH dehydrogenase or NADH dehydrogenase enzymes reduce resazurin into the fluorescent product resorufin (Riss et al. 2013). For this reason, this assay is used to measure mitochondrial metabolic activity/cell viability. Our results described in Fig. 8a showed that mitochondrial metabolic activity in the aggregates pre-exposed to rotenone at 50 or
100 nM was higher than controls showing resilience to a second exposure to rotenone. Pre- exposure to 25 nM, however, did not lead to resilience, which means that the response to a second exposure is likely dependent on the concentration of the first exposure. Sherer et al. (2003) reported that Ndufs4−/− (complex I accessory subunit) primary cells had increased NADH content but were more susceptible to rotenone toxicity. It will have to be further determined, which molecular signatures or possible epigenetic mechanisms lead to resilience and whether this is a short-term or long-term phenomenon. Further research is also needed to better understand whether the altered metabolic response is an adaptive response, making the cells robust; or rather detrimental, leading to a disease pathway in a long-term perspective. This experimental approach could provide quantitative data for different key events in adverse outcome pathways (AOPs).
To identify changes in gene expression after a second exposure compared to alterations observed after a single exposure, we assessed genes which had previously found to be altered by rotenone in LUHMES. Three genes were less sensitive to rotenone exposure on day 14 after being pre- exposed to rotenone on day 7, suggesting their role in resilience. NEF2L2, the gene coding for
Nrf2, a protein involved in the oxidative stress response (Shih et al. 2005); ATF4, previously found altered by rotenone and involved in cell stress and proteasome inhibition (Krug et al. 2014;
219
Smirnova et al. 2016); and EAAC1, responsible for glutamate uptake and found to be downregulated in PD models (Kinoshita et al. 2014; Zhang et al. 2016a, b). This could indicate that pre-exposed cells do not activate these response mechanisms upon a second exposure, and may be more resilient to the activation of specific pathways. Conversely these could be protective pathways, which the cell cannot activate upon a second exposure (point of no return; Krug et al.
2014), or have reached a tipping point (Schildknecht et al. 2017; Jennings et al 2004;
Koppelstaetter et al. 2004). Our results suggest new questions as to where the threshold of an effect lies (Bal-Price et al. 2015, 2017a, b; Terron et al. 2018).
Conversely, we found genes, which were altered to a greater extent upon a second exposure compared with a single exposure on day 14. This was observed for the dopamine transporter DAT and calcium-mediated apoptosis protein CASP3 (Fig. 8e). Furthermore, MLF1IP, the gene coding for a centromere protein involved in mitotic progression and transcriptional regulation; and
TYMS, an enzyme involved in the synthesis of thymidine nucleotides for DNA repair and mitochondrial thymidylate biosynthesis were downregulated to the same level in single-exposed and pre-exposed aggregates (Fig. 8f). We confirmed the downregulation of MLF1IP and TYMS by 100 nM rotenone that has previously been reported and was permanent after wash-out (Krug et al. 2014; Smirnova et al. 2016). For these genes, a second hit did not lead to further downregulation, indicating that there is likely a threshold for their permanent downregulation.
TYMS downregulation has been found to increase oxidative stress production as well as activate protective pathways in multiple cancer lines (Ozer et al. 2015; Xu et al. 2017) but has not been extensively studied in neurons.
These results show that second exposures lead to activation of different expression patterns and, therefore, wash-out and repeated exposures could provide more insight for adverse outcome pathways (Leist et al. 2017) and potential therapeutic targets. Further experiments are needed to study whether, in the long term, these pathways are detrimental or cells continue to confer resilience (Karatsoreos and McEwen 2013; Delp et al. 2017). The present work does not inform
220 us on the specificity of rotenone in inducing resilience, since no other cell types or chemicals were tested yet, but as, with an AOP approach, demonstrates how events which can lead to adversity and their reversibility can be addressed. Some have studied repeated-dose chronic effects in vitro (Borland et al. 2008; Shaikh and Nicholson 2009; Gourov and Currran 2014), but not with a focus on recovery, adaptation, and resilience in dopaminergic neurons. Gene– environment interactions play an important role in neurodegeneration, e.g., in PD, and an altered genetic/epigenetic response to toxicants is thought to primarily drive sporadic PD (Miranda-
Morales et al. 2017). In the context of resilience, epigenetic mechanisms may play a more crucial role, supported by the abolished transcriptional changes after compound wash-out and fact that epigenetics lay in the interplay between genetic and environmental interactions. Post-translational regulation may also be important as many PD-related genes are tightly regulated via phosphorylation and ubiquitination (Oueslati 2016; Nakazawa et al. 2016; Xu et al. 2015; Wani et al. 2015).
Although studies have focused on neuroprotective mechanisms in animal and in vitro models via silencing of pathways involved in degeneration or overexpression of neuroprotective pathways
(Yacoubian et al. 2010; Zharikov et al. 2015; Zhang et al. 2016a, b; Basil et al. 2017; Lee et al.
2017; to name a few), the reversibility of morphological and functional endpoints has not been shown in cultured cells. Understanding changes, which occur after compound removal is not only a new approach as to how in vitro toxicity testing should be addressed but also is crucial to understand long-term toxicity. In the field of neurodegenerative diseases, there is a need to better understand the interplay between degenerative, adaptive, and protective pathways to identify complex gene–environment interactions and therapeutic targets. 3D in vitro models, which allow for repeated exposures and recovery periods, will better help to understand how low-dose exposures may lead to long-term disease. Furthermore, more complex multicellular test systems would help to identify the role of support cells such as astrocytes and microglia in recovery as
221 well as how the differentiation stage or ‘age’ affects dopaminergic toxicity (Pamies et al. 2017,
2018).
Taken together, this study shows that cells which seem to be functionally ‘recovered’ from a toxicant hit retain some form of memory and are not the same anymore. This is largely neglected in the many acute high dose in vitro experiments reported. Cellular resilience and/or the
‘molecular scar’ concept in neurotoxicology and neurodegeneration can be compared to our immune system response, which develops memory and prior stimulation can lead to a different response to subsequent stimuli (Henn et al. 2011). The imprint from earlier exposures, which can manifest as either a molecular scar (rendering cells more sensitive), or resilience (more tolerant) needs to be considered to understand real-life exposures and measure risk. The demonstration of resilience here as a type of chemical tolerance would suggest that we might be overestimating toxic effects from commonly performed acute toxicity studies. It will be most interesting to see whether these phenomena are toxicant-selective, i.e., whether tolerance is observed only for the same toxicant or a class of toxicants or whether the cells are more robust in general. The model system presented here will allow the characterization of such mechanisms in the future.
222
5.7. ACKNOWLEDGEMENTS
We would like to acknowledge the International Foundation for Ethical Research Graduate
Fellowship Funding provided to Georgina Harris, and help from Erin Pryce from the Integrated
Imaging Core Facility at Johns Hopkins University. This project has received funding from the
European Union’s Horizon 2020 research and innovation programme under Grant Agreement No.
681002.
223
5.8. REFERENCES
Ahmed H, Abushouk AI, Gabr M, Negida A, Abdel-Daim MM (2017) Parkinson's disease and
pesticides: A meta-analysis of disease connection and genetic alterations. Biomed
Pharmacother 90:638-649. doi: 10.1016/j.biopha.2017.03.100.
Alberio T, Lopiano L, Fasano M (2012) Cellular models to investigate biochemical pathways in
Parkinson's disease. FEBS J 279(7):1146-55. doi: 10.1111/j.1742-4658.2012.08516.x.
Alépée N, Bahinski A, Daneshian M, et al (2014) State-of-the-art of 3D cultures (organs-on-a-
chip) in safety testing and pathophysiology. ALTEX 31:441–477.
Ariga H, Takahashi-Niki K Kato I, Maita H, Niki T, Iguchi-Ariga SMM (2013) Neuroprotective
Function of DJ-1 in Parkinson's Disease. Oxid Med Cell Longev. 683920. doi:
10.1155/2013/683920
Bal-Price A, Crofton KM, Sachana M, Shafer TJ, Behl M, Forsby A, Hargreaves A, Landesmann
B, Lein PJ, Louisse J, Monnet-Tschudi F, Paini A, Rolaki A, Schrattenholz A, Suñol C, van
Thriel C, Whelan M, Fritsche E (2015) Putative adverse outcome pathways relevant to
neurotoxicity. Crit Rev Toxicol 45(1):83-91. doi: 10.3109/10408444.2014.981331.
Bal-Price A, Lein PJ, Keil KP, Sethi S, Shafer T, Barenys M, Fritsche E, Sachana M, Meek ME
(2017) Developing and applying the adverse outcome pathway concept for understanding and
predicting neurotoxicity. Neurotoxicology 59:240-255. doi: 10.1016/j.neuro.2016.05.010.
Bal-Price A, Leist M, Schildknecht S, Tschudi-Monnet F, Paini A, Terron A. (2017) Inhibition of
the mitochondrial complex I of nigro-striatal neurons leads to parkinsonian motor deficits.
https://aopwiki.org/aops/3
Bal-Price A, Meek MEB (2017) Adverse outcome pathways: Application to enhance mechanistic
understanding of neurotoxicity. Pharmacol Ther. 179:84-95. doi:
10.1016/j.pharmthera.2017.05.006.
224
Basil AH, Sim JPL, Lim GGY, Lin S, Chan HY, Engelender S, Lim KL (2017) AF-6 Protects
Against Dopaminergic Dysfunction and Mitochondrial Abnormalities in Drosophila Models
of Parkinson's Disease. Front Cell Neurosci 11:241. doi: 10.3389/fncel.2017.00241.
Belin AC, Westerlund M (2008) Parkinson's disease: a genetic perspective. FEBS J 275(7):1377-
83. doi: 10.1111/j.1742-4658.2008.06301.x.
Bertola LD, Ott EB, Griepsma S, Vonk FJ, Bagowski CP (2008) Developmental expression of
the alpha-skeletal actin gene. BMC Evolutionary Biology 8:166. doi:10.1186/1471-2148-8-
166.
Bolstad BM, Irizarry RA, Astrand M, Speed TP (2003) A comparison of normalization methods
for high density oligonucleotide array data based on variance and bias. Bioinformatics
19(2):185-93.
Borland MK, Trimmer PA, Rubinstein JD, Keeney PM, Mohanakumar K, Liu L, Bennett JP Jr
(2008) Chronic, low-dose rotenone reproduces Lewy neurites found in early stages of
Parkinson's disease, reduces mitochondrial movement and slowly kills differentiated SH-S
Y5Y neural cells. Mol Neurodegener 3:21. doi: 10.1186/1750-1326-3-21.
Cannon JR, Tapias VM, Na HM, Honick AS, Drolet RE,Greenamyre JT (2009) A highly
reproducible rotenone model of Parkinson's disease. Neurobiol Dis 34(2): 279–290. doi:
10.1016/j.nbd.2009.01.016
Caudle WM, Guillot TS, Lazo CR, and Miller GW (2012) Industrial toxicants and Parkinson’s
disease. Neurotoxicology 33(2): 178–188. doi:10.1016/j.neuro.2012.01.010
Choi WS, Kruse SE, Palmiter RD, Xia Z (2008) Mitochondrial complex I inhibition is not
required for dopaminergic neuron death induced by rotenone, MPP+, or
paraquat. Proceedings of the National Academy of Sciences of the United States of America
105(39):15136-15141. doi:10.1073/pnas.0807581105.
225
Choi W-S, Palmiter RD, Xia Z (2011) Loss of mitochondrial complex I activity potentiates
dopamine neuron death induced by microtubule dysfunction in a Parkinson’s disease model.
The Journal of Cell Biology 192(5):873-882. doi:10.1083/jcb.201009132.
Collier TJ, Kanaan NM, Kordower JH (2017) Aging and Parkinson's disease: Different sides of
the same coin? Mov Disord 32(7):983-990. doi: 10.1002/mds.27037.
Delp J, Gutbier S, Cerff M, Zasada C, Niedenführ S, Zhao L, Smirnova L, Hartung T,
Borlinghaus H, Schreiber F, Bergemann J, Gätgens J, Beyss M, Azzouzi S, Waldmann T,
Kempa S, Nöh K, Leist M (2017) Stage-specific metabolic features of differentiating
neurons: Implications for toxicant sensitivity. Toxicol Appl Pharmacol 17:30494-5. doi:
10.1016/j.taap.2017.12.013.
Daneshian M, Busquet F, Hartung T, Leist M (2015) Animal use for science in Europe. ALTEX
32(4):261-74. doi: 10.14573/altex.1509081.
Daskalakis NP, Bagot RC, Parker KJ, Vinkers CH, de Kloet ER (2013) The three-hit concept of
vulnerability and resilience: toward understanding adaptation to early-life adversity outcome.
Psychoneuroendocrinology 38(9):1858-73. doi: 10.1016/j.psyneuen.2013.06.008.
Dhillon AS, Tarbutton GL, Levin JL, Plotkin GM, Lowry LK, Nalbone JT, Shepherd S (2008)
Pesticide/environmental exposures and Parkinson's disease in East Texas. J Agromedicine
13(1):37-48. doi: 10.1080/10599240801986215.
Dorsey ER, Constantinescu R, Thompson JP, Biglan KM, Holloway RG, Kieburtz K, Marshall
FJ, Ravina BM, Schifitto G, Siderowf A, Tanner CM (2007) Projected number of people with
Parkinson disease in the most populous nations, 2005 through 2030. Neurology 68(5):384-6.
doi: 10.1212/01.wnl.0000247740.47667.03
Drolet RE, Cannon JR, Montero L, Greenamyre JT (2009) Chronic rotenone exposure reproduces
Parkinson's disease gastrointestinal neuropathology. Neurobiol Dis 36(1):96-102. doi:
10.1016/j.nbd.2009.06.017.
226
Dusonchet J, Li H, Guillily M, Liu M, Stafa K, Derada Troletti C, Boon JY, Saha S, Glauser L,
Mamais A, Citro A, Youmans KL, Liu L, Schneider BL, Aebischer P, Yue Z, Bandopadhyay
R, Glicksman MA, Moore DJ, Collins JJ, Wolozin B (2014) A Parkinson's disease gene
regulatory network identifies the signaling protein RGS as a modulator of LRRK2 activity
and neuronal toxicity. Hum Mol Genet 23(18):4887-905. doi: 10.1093/hmg/ddu202.
Falkenburger BH, Saridaki T, Dinter E (2016) Cellular models for Parkinson's disease. J
Neurochem 139. Suppl 1:121-130. doi: 10.1111/jnc.13618
Fang D, Qing Y, Yan S, Chen D, Yan SS (2016) Development and Dynamic Regulation of
Mitochondrial Network in Human Midbrain Dopaminergic Neurons Differentiated from
iPSCs. Stem Cell Reports 7(4):678-692. doi:10.1016/j.stemcr.2016.08.014.
Fujii C, Harada S, Ohkoshi N, Hayashi A, Yoshizawa K, Ishizuka C, Nakamura T (1999)
Association between polymorphism of the cholecystokinin gene and idiopathic Parkinson's
disease. Clin Genet 56(5):394-9.
Furlong M, Tanner CM, Goldman SM, Bhudhikanok GS, Blair A, Chade A, Comyns K, Hoppin
JA, Kasten M, Korell M, Langston JW, Marras C, Meng C, Richards M, Ross GW, Umbach
DM, Sandler DP, Kamel F (2015) Protective glove use and hygiene habits modify the
associations of specific pesticides with Parkinson's disease. Environ Int 75:144-50. doi:
10.1016/j.envint.2014.11.002.
Gaki GS, Papavassiliou AG (2014) Oxidative stress-induced signaling pathways implicated in the
pathogenesis of Parkinson's disease. Neuromolecular Med 16(2):217-30. doi:
10.1007/s12017-014-8294-x.
Gorell JM, Johnson CC, Rybicki BA, Peterson EL, Richardson RJ (1998) The risk of Parkinson's
disease with exposure to pesticides, farming, well water, and rural living. Neurology
50(5):1346-50.
Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, Ellis B, Gautier L, Ge Y,
Gentry J, Hornik K, Hothorn T, Huber W, Iacus S, Irizarry R, Leisch F, Li C, Maechler M,
227
Rossini AJ, Sawitzki G, Smith C, Smyth G, Tierney L, Yang LYH and Zhang (2004)
Bioconductor: open software development for computational biology and bioinformatics.
Genome Biology 5:R80 doi: https://doi.org/10.1186/gb-2004-5-10-r80
Gourov AV, Curran B (2014) An in vitro model to study brain tissue recovery. Med Hypotheses
82(6):674-7. doi: 10.1016/j.mehy.2014.03.001.
Greenamyre JT, Betarbet R, Sherer TB (2003) The rotenone model of Parkinson's disease: genes,
environment and mitochondria. Parkinsonism Relat Disord 9 Suppl 2:S59-64.
Grefte S, Wagenaars JA, Jansen R, Willems PH, Koopman WJ (2015) Rotenone inhibits primary
murine myotube formation via Raf-1 and ROCK2. Biochim Biophys Acta 1853(7):1606-14.
doi: 10.1016/j.bbamcr.2015.03.010.
Grivennikova VG, Maklashina EO, Gavrikova EV, Vinogradov AD (1997) Interaction of the
mitochondrial NADH-ubiquinone reductase with rotenone as related to the enzyme
active/inactive transition. Biochim Biophys Acta 1319(2-3):223-32.
Haddad D, Nakamura K (2015) Understanding the susceptibility of dopamine neurons to
mitochondrial stressors in Parkinson's disease. FEBS Lett 589(24 Pt A):3702-13. doi:
10.1016/j.febslet.2015.10.021.
Harris G, Hogberg H, Hartung T, Smirnova L (2017) 3D Differentiation of LUHMES Cell Line
to Study Recovery and Delayed Neurotoxic Effects. Curr Protoc Toxicol. 73:11.23.1-
11.23.28. doi: 10.1002/cptx.29.
Hartung, T (2011) Food for Thought Look Back in Anger – What Clinical Studies Tell Us About
Preclinical Work J Immunol 186(5):3237-47. doi: 10.4049/jimmunol.1002787. Epub 2011
Jan 31.
Henn A, Kirner S, Leist M (2013) TLR2 hypersensitivity of astrocytes as functional consequence
of previous inflammatory episodes. ALTEX. 30(3): 275-291.
228
Higgins DS Jr, Greenamyre JT (1996) [3H]dihydrorotenone binding to NADH: ubiquinone
reductase (complex I) of the electron transport chain: an autoradiographic study. J Neurosci
16(12):3807-16.
Hogberg HT, Bressler J, Christian KM, Harris G, Makri G, O'Driscoll C, Pamies D, Smirnova L,
Wen Z, Hartung T (2013) Toward a 3D model of human brain development for studying
gene/environment interactions. Stem Cell Res Ther 4 Suppl 1:S4. doi: 10.1186/scrt365.
Huber W, Carey VJ, Gentleman R, et al (2015) Orchestrating high-throughput genomic analysis
with Bioconductor. Nature methods 12(2):115-121. doi:10.1038/nmeth.3252.
Jankovic J (2008) Parkinson’s disease: clinical features and diagnosis. Journal of Neurology,
Neurosurgery and Psychiatry 79(4):368–376. doi: 10.1136/jnnp.2007.131045
Jennings P, Koppelstaetter C, Pfaller W, Morin JP, Hartung T, Ryan MP (2004) Assessment of a
new cell culture perfusion apparatus for in vitro chronic toxicity testing. Part 2: toxicological
evaluation. ALTEX 21(2):61-6.
Johnson WE, Li C, Rabinovic A (2007) Adjusting batch effects in microarray expression data
using empirical Bayes methods. Biostatistics 8(1):118-27. Epub 2006 Apr 21.
Kalia LV, Lang AE (2015) Parkinson's disease. Lancet 386(9996):896-912. doi: 10.1016/S0140-
6736(14)61393-3.
Karatsoreos IN, McEwen BS (2013) Annual Research Review: The neurobiology and physiology
of resilience and adaptation across the life course. J Child Psychol Psychiatry 54(4):337-47.
doi: 10.1111/jcpp.12054.
Keane PC, Kurzawa M, Blain PG, Morris CM (2011) Mitochondrial dysfunction in Parkinson's
disease. Parkinsons Dis 2011:716871. doi: 10.4061/2011/716871.
Kinoshita C, Aoyama K, Matsumura N, Kikuchi-Utsumi K, Watabe M, Nakaki T (2014)
Rhythmic oscillations of the microRNA miR-96-5p play a neuroprotective role by indirectly
regulating glutathione levels. Nat Commun 5:3823. doi: 10.1038/ncomms4823.
229
Kleensang A, Vantangoli MM, Odwin-DaCosta S, et al (2016) Genetic variability in a frozen
batch of MCF-7 cells invisible in routine authentication affecting cell function. Scientific
Reports 6:28994. doi:10.1038/srep28994.
Kolodziejczyk AA, Lönnberg T (2017) Global and targeted approaches to single-cell
transcriptome characterization. Brief Funct Genomics [Epub ahead of print] doi:
10.1093/bfgp/elx025.
Koppelstaetter C, Jennings P, Ryan MP, Morin JP, Hartung T, Pfaller W (2004) Assessment of a
new cell culture perfusion apparatus for in vitro chronic toxicity testing. Part 1: technical
description. ALTEX 21(2):51-60.
Krug AK, Gutbier S, Zhao L, et al (2014) Transcriptional and metabolic adaptation of human
neurons to the mitochondrial toxicant MPP+. Cell Death & Disease 5(5):e1222-.
doi:10.1038/cddis.2014.166.
Lee Y, Kim MS, Lee J (2017) Neuroprotective strategies to prevent and treat Parkinson's disease
based on its pathophysiological mechanism. Arch Pharm Res 40(10):1117-1128. doi:
10.1007/s12272-017-0960-8.
Lees AJ, Hardy J, Revesz T (2009) Parkinson’s disease. The Lancet 373(9680):2055–2066. doi:
10.1016/S0140-6736(09)60492-X.
Leist M, Ghallab A, Graepel R, Marchan R, Hassan R, Bennekou SH, Limonciel A, Vinken M,
Schildknecht S, Waldmann T, et al (2017) Adverse outcome pathways: opportunities,
limitations and open questions. Arch Toxicol 91(11):3477-3505. doi: 10.1007/s00204-017-
2045-3.
Lenka A, Arumugham SS, Christopher R, Pal PK (2016) Genetic substrates of psychosis in
patients with Parkinson's disease: A critical review. J Neurol Sci 364:33-41. doi:
10.1016/j.jns.2016.03.005.
230
Li N, Ragheb K, Lawler G, Sturgis J, Rajwa B, Melendez JA, Robinson JP (2003) Mitochondrial
complex I inhibitor rotenone induces apoptosis through enhancing mitochondrial reactive
oxygen species production. J Biol Chem 278(10):8516-25.
Lindahl PE, Öberg KE (1961) The effect of rotenone on respiration and its point of attack
Experimental Cell Research 23(2):228-237.
Lotharius J, Falsig J, van Beek J, Payne S, Dringen R, Brundin P, Leist M (2005) Progressive
degeneration of human mesencephalic neuron-derived cells triggered by dopamine-dependent
oxidative stress is dependent on the mixed-lineage kinase pathway. J Neurosci 25(27): 6329-
6342.
Manji HK, Moore GJ, Rajkowska G, Chen G (2000) Neuroplasticity and cellular resilience in
mood disorders. Mol Psychiatry 5(6):578-93.
Miranda-Morales E, Meier K, Sandoval -Carrillo A, Salas-Pacheco J, Vázquez-Cárdenas P,
Arias-Carrión O (2017) Implications of DNA Methylation in Parkinson's Disease. Front Mol
Neurosci 10:225. doi: 10.3389/fnmol.2017.00225.
Nakazawa S, Oikawa D, Ishii R, et al (2016) Linear ubiquitination is involved in the pathogenesis
of optineurin-associated amyotrophic lateral sclerosis. Nature Communications 7:12547.
doi:10.1038/ncomms12547.
National Research Council (2007) Toxicity Testing in the 21st Century: A Vision and a Strategy.
Washington, DC: The National Academies Press. https://doi.org/10.17226/11970.
Nilsson A, Fälth M, Zhang X, Kultima K, Sköld K, Svenningsson P, Andrén PE (2009) Striatal
alterations of secretogranin-1, somatostatin, prodynorphin, and cholecystokinin peptides in an
experimental mouse model of Parkinson disease. Mol Cell Proteomics 8(5):1094-104. doi:
10.1074/mcp.M800454-MCP200.
O'Brien J, Wilson I, Orton T, Pognan F (2000) Investigation of the Alamar Blue (resazurin)
fluorescent dye for the assessment of mammalian cell cytotoxicity. Eur J Biochem
267(17):5421-6.
231
Oertel WH (2017) Recent advances in treating Parkinson’s disease. F1000Research 6:260.
doi:10.12688/f1000research.10100.1.
Olson H, Betton G, Robinson D, Thomas K, Monro A, Kolaja G, Lilly P, Sanders J, Sipes G,
Bracken W, Dorato M, Van Deun K, Smith P, Berger B, Heller A (2000) Concordance of the
toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharmacol 32(1):56-67.
Oueslati A (2016) Implication of Alpha-Synuclein Phosphorylation at S129 in Synucleinopathies:
What Have We Learned in the Last Decade? Journal of Parkinson’s Disease 6(1):39-51.
doi:10.3233/JPD-160779.
Ozer U, Barbour KW, Clinton SA, Berger FG (2015) Oxidative Stress and Response to
Thymidylate Synthase-Targeted Antimetabolites. Mol Pharmacol 88(6):970-81. doi:
10.1124/mol.115.099614.
Pakkenberg B, Møller A, Gundersen HJ, Mouritzen Dam A, Pakkenberg H (1991) The absolute
number of nerve cells in substantia nigra in normal subjects and in patients with Parkinson’s
disease estimated with an unbiased stereological method. Journal of Neurology,
Neurosurgery, and Psychiatry 54(1):30-33.
Pamies D, Hartung T (2017a) 21st Century Cell Culture for 21st Century Toxicology. Chem Res
Toxicol 30(1):43-52. doi: 10.1021/acs.chemrestox.6b00269.
Pamies D, Barreras P, Block K, Makri G, Kumar A, Wiersma D, Smirnova L, Zang C, Bressler J,
Christian KM, Harris G, Ming GL, Berlinicke CJ, Kyro K, Song H, Pardo CA, Hartung T,
Hogberg HT. (2017b) A human brain microphysiological system derived from induced
pluripotent stem cells to study neurological diseases and toxicity. ALTEX. 34(3):362-376.
doi: 10.14573/altex.1609122. Epub 2016 Nov 24.
Pamies D, Block K, Lau P, Gribaldo L, Pardo C, Barreras P, Smirnova L, Wiersma D, Zhao L,
Harris G, Hartung T, Hogberg HT. (2018) Rotenone exerts developmental neurotoxicity in a
human brain spheroid model. Toxicol Appl Pharmacol. pii: S0041-008X(18)30042-5. doi:
10.1016/j.taap.2018.02.003. [Epub ahead of print]
232
Parker WD, Parks JK, Swerdlow RH (2008) Complex I Deficiency in Parkinson’s Disease
Frontal Cortex. Brain Res 1189: 215–218. doi:10.1016/j.brainres.2007.10.061
Pfau ML, Russo SJ (2015) Peripheral and central mechanisms of stress resilience. Neurobiology
of Stress 1:66-79. doi:10.1016/j.ynstr.2014.09.004.
Ren Y, Ye M, Chen S, Ding J (2016) CD200 Inhibits Inflammatory Response by Promoting
KATP Channel Opening in Microglia Cells in Parkinson's Disease. Med Sci Monit 22:1733-
41.
Richardson JR, Quan Y, Sherer TB, Greenamyre JT, Miller GW (2005) Paraquat neurotoxicity is
distinct from that of MPTP and rotenone. Toxicol Sci 88(1):193-201. Epub 2005 Sep 1.
Richardson LA (2016) Understanding Disease Tolerance and Resilience. PLoS Biol 14(7):
e1002513. doi: 10.1371/journal.pbio.1002513
Riss TL, Moravec RA, Niles AL, Duellman S, Benink HA, Worzella TJ, Minor L (2013) Cell
Viability Assays. Assay Guidance Manual. (Book chapter) Eli Lilly & Company and the
National Center for Advancing Translational Sciences
Ritchie ME, Silver J, Oshlack A, Holmes M, Diyagama D, Holloway A, Smyth GK (2007) A
comparison of background correction methods for two-colour microarrays. Bioinformatics
23(20):2700-7.
Schapira AH, Cooper JM, Dexter D, Jenner P, Clark JB, Marsden CD (1989) Mitochondrial
complex I deficiency in Parkinson's disease. Lancet 1(8649):1269.
Schildknecht S, Di Monte DA Pape R, Tieu K, Leist M (2017) Tipping Points and Endogenous
Determinants of Nigrostriatal Degeneration by MPTP. Trends Pharmacol Sci 38(6):541-555.
doi: 10.1016/j.tips.2017.03.010.
Schildknecht S, Karreman C, Pöltl D, Efrémova L, Kullmann C, Gutbier S, Krug A, Scholz D,
Gerding HR, Leist M (2013) Generation of genetically-modified human differentiated cells
for toxicological tests and the study of neurodegenerative diseases. ALTEX 30(4):427-44.
233
Schildknecht S, Pöltl D, Nagel DM, Matt F, Scholz D, Lotharius J, Schmieg N, Salvo-Vargas A,
Leist M (2009) Requirement of a dopaminergic neuronal phenotype for toxicity of low
concentrations of 1-methyl-4-phenylpyridinium to human cells. Toxicol Appl Pharmacol
241(1):23-35. doi: 10.1016/j.taap.2009.07.027.
Schmidt BZ, Lehmann M, Gutbier S, Nembo E, Noel S, Smirnova L, Forsby A, Hescheler J,
Avci HX, Hartung T, Leist M, Kobolák J and Dinnyés A (2017) In vitro neurotoxicity
screening: an overview of cellular platforms and high-throughput technical possibilities.
Arch. Toxicol in press. doi: 10.1007/s00204-016-1805-9.
Scholz D, Pöltl D, Genewsky A, Weng M, Waldmann T, Schildknecht S, Leist M (2011) Rapid,
complete and large-scale generation of post-mitotic neurons from the human LUHMES cell
line. J Neurochem 119(5):957-71. doi: 10.1111/j.1471-4159.2011.07255.x.
Sherer TB, Betarbet R, Stout AK, Lund S, Baptista M, Panov AV, Cookson MR, Greenamyre JT
(2002) An in vitro model of Parkinson's disease: linking mitochondrial impairment to altered
alpha-synuclein metabolism and oxidative damage. J Neurosci 22(16):7006-15.
Sherer TB, Betarbet R, Testa CM, Seo BB, Richardson JR, Kim JH, Miller GW, Yagi T,
Matsuno-Yagi A, Greenamyre JT (2003) Mechanism of toxicity in rotenone models of
Parkinson's disease. J Neurosci 23(34):10756-64.
Sherer TB, Richardson JR, Testa CM, Seo BB, Panov AV, Yagi T, Matsuno-Yagi A, Miller GW,
Greenamyre JT (2007) Mechanism of toxicity of pesticides acting at complex I: relevance to
environmental etiologies of Parkinson's disease. J Neurochem 100(6):1469-79. doi:
10.1111/j.1471-4159.2006.04333.x
Silver JD, Ritchie ME, Smyth GK (2009) Microarray background correction: maximum
likelihood estimation for the normal–exponential convolution. Biostatistics (Oxford,
England) 10(2):352-363. doi:10.1093/biostatistics/kxn042.
Smirnova L, Harris G, Delp J, Valadares M, Pamies D, Hogberg HT, Waldmann T, Leist M,
Hartung T (2016) A LUHMES 3D dopaminergic neuronal model for neurotoxicity testing
234
allowing long-term exposure and cellular resilience analysis. Arch Toxicol 90(11):2725-
2743.
Smirnova L, Harris G, Leist M, Hartung T (2015) Cellular resilience. ALTEX 32(4):247-60. doi:
10.14573/altex.1509271.
Smyth GK (2004) Linear models and empirical bayes methods for assessing differential
expression in microarray experiments. Stat Appl Genet Mol Biol 3:Article3.
Surmeier J, Guzman JN, Sanchez-Padilla J, Goldberg JA (2010) Chapter 4 - What causes the
death of dopaminergic neurons in Parkinson’s disease? rogress in Brain Research 183:59-77
https://doi.org/10.1016/S0079-6123(10)83004-3
Tabrez S, Jabir NR, Shakil S,Greig NH, Alam Q, Abuzenadah AM, Damanhouri GA, Kamal MA
(2012) A Synopsis on the Role of Tyrosine Hydroxylase in Parkinson's Disease. CNS Neurol
Disord Drug Targets 11(4): 395–409.
Tanner CM, Kamel F, Ross GW, Hoppin JA, Goldman SM, Korell M, Marras C, Bhudhikanok
GS, Kasten M, Chade AR, Comyns K, Richards MB, Meng C, Priestley B, Fernandez HH,
Cambi F, Umbach DM, Blair A, Sandler DP, Langston JW (2011) Rotenone, paraquat, and
Parkinson's disease. Environ Health Perspect 119(6):866-72. doi: 10.1289/ehp.1002839.
Terron A, Bal-Price A, Paini A, Monnet-Tschudi F, Bennekou SH, EFSA WG EPI1 Members,
Leist M, Schildknecht S (2018) An adverse outcome pathway for parkinsonian motor deficits
associated with mitochondrial complex I inhibition. Arch Toxicol 92(1):41-82. doi:
10.1007/s00204-017-2133-4.
Tiwari BS, Belenghi B, Levine A (2002) Oxidative stress increased respiration and generation of
reactive oxygen species, resulting in ATP depletion, opening of mitochondrial permeability
transition, and programmed cell death. Plant Physiol 128(4):1271-81.
Tyagi E, Zhuang Y, Agrawal R, Ying Z, Gomez-Pinilla F (2015) Interactive actions of Bdnf
methylation and cell metabolism for building neural resilience under the influence of diet.
Neurobiol Dis 73:307-18. doi: 10.1016/j.nbd.2014.09.014.
235
Wang XJ, Ye M, Zhang YH, Chen SD (2007) CD200-CD200R regulation of microglia activation
in the pathogenesis of Parkinson's disease. J Neuroimmune Pharmacol 2(3):259-64. doi:
10.1007/s11481-007-9075-1
Wang XJ, Zhang S, Yan ZQ, Zhao YX, Zhou HY, Wang Y, Lu GQ, Zhang JD (2011) Impaired
CD200-CD200R-mediated microglia silencing enhances midbrain dopaminergic
neurodegeneration: roles of aging, superoxide, NADPH oxidase, and p38 MAPK. Free Radic
Biol Med 50(9):1094-106. doi: 10.1016/j.freeradbiomed.2011.01.032.
Wani W, Boyer-Guittaut M, Dodson M, Chatham J, Darley-Usmar V, Zhang J (2015) Regulation
of autophagy by protein post-translational modification. Laboratory investigation; a journal of
technical methods and pathology 95(1):14-25. doi:10.1038/labinvest.2014.131.
Wiklund FE, Bennet AM, Magnusson PK, Eriksson UK, Lindmark F, Wu L, Yaghoutyfam N,
Marquis CP, Stattin P, Pedersen NL, Adami HO, Grönberg H, Breit SN, Brown DA (2010)
Macrophage inhibitory cytokine-1 (MIC-1/GDF15): a new marker of all-cause mortality.
Aging Cell 9(6):1057-64. doi: 10.1111/j.1474-9726.2010.00629.x.
Xicoy H, Wieringa B, Marten GJM (2017) The SH-SY5Y cell line in Parkinson’s disease
research: a systematic review. Mol Neurodegener 12: 10. doi: 10.1186/s13024-017-0149-0
Xu W, Jiang H, Zhang F, Gao J, Hou J (2017) MicroRNA-330 inhibited cell proliferation and
enhanced chemosensitivity to 5-fluorouracil in colorectal cancer by directly targeting
thymidylate synthase. Oncol Lett 13(5):3387-3394. doi: 10.3892/ol.2017.5895.
Xu Y, Deng Y, Qing H (2015) The phosphorylation of α-synuclein: development and implication
for the mechanism and therapy of the Parkinson's disease. J Neurochem 135(1):4-18. doi:
10.1111/jnc.13234.
Yacoubian TA, Slone SR, Harrington AJ, Hamamichi S, Schieltz JM, Caldwell KA, Caldwell
GA, Standaert DG (2010) Differential neuroprotective effects of 14-3-3 proteins in models of
Parkinson's disease. Cell Death Dis 1:e2. doi: 10.1038/cddis.2009.4.
236
Yu G, Wang L, Han Y and He Q (2012) ClusterProfiler: an R package for comparing biological
themes among gene clusters. OMICS: A Journal of Integrative Biology 16(5):284-287.
doi:10.1089/omi.2011.0118
Zhang JY, Deng YN, Zhang M, Su H, Qu QM (2016) SIRT3 Acts as a Neuroprotective Agent in
Rotenone-Induced Parkinson Cell Model. Neurochem Res 41(7):1761-73. doi:
10.1007/s11064-016-1892-2.
Zhang Y, Tan F, Xu P, Qu S (2016) Recent Advance in the Relationship between Excitatory
Amino Acid Transporters and Parkinson’s Disease. Neural Plasticity 2016:8941327.
doi:10.1155/2016/8941327.
Zeiger SLH, McKenzie JR, Stankowski JN, Martin JA, Cliffel DE, McLaughlin B. (2010)
Neuron Specific Metabolic Adaptations Following Multi-Day Exposures to Oxygen Glucose
Deprivation. Biochimica et biophysica acta. 1802(11):1095-1104.
doi:10.1016/j.bbadis.2010.07.013.
Zharikov AD, Cannon JR, Tapias V, et al (2015) shRNA targeting α-synuclein prevents
neurodegeneration in a Parkinson’s disease model. The Journal of Clinical Investigation
125(7):2721-2735. doi:10.1172/JCI64502.
237
CHAPTER 6
6. EFFECTS OF REPEATED LOW-DOSE EXSPOSURE TO
ROTENONE IN THE 3D LUHMES MODEL.
* Data presented in this chapter will be included in the following manuscript under preparation.
Harris G, Hartung T, Smirnova L. Repeated-low dose effects of rotenone and glyphosate in 3D
LUHMES model.
Key points
• Regulatory animal testing involves repeated-dosing for neurotoxicity testing, however,
limited research has been performed on repeated-dose effects in vitro.
• Two European projects; DETECTIVE project (detection of endpoints and biomarkers of
repeated dose toxicity using in vitro systems, within SEURAT-1) and EUToxRisk are
aimed at replacing animal repeated-dose testing.
• Neuronal models cultured in monolayer, have limitations and relative short life-span due
to low adherence. As 3D LUHMES are cultured in suspension, we can study repeated-
dose effects after differentiation.
• Repeated doses of 30 nM rotenone did not affect viability but were found to increase
ATP levels, and decrease neurite outgrowth.
• Increased expression of PD and autophagy-related genes (TH, PARK2, PARK7 and
ASS1) was observed after repeated low-dose exposures to rotenone.
238
6.1. MATERIALS AND METHODS
LUHMES 3D culture and compound exposure
LUHMES (ATCC® CRL_2927™) 3D cell culture and differentiation protocol was followed as described (Harris et al. 2017). Briefly, cells were used between passages 15 and 25. 4x106 cells were placed in a 175 cm² flask for 48 hours to expand cells. On day 0, 3D-differentiation was initiated: 5.5x105 cells were seeded into each well of a 6-well plate and placed on a gyratory shaker at 80 rpm (50 mm orbit) in an incubator at 37°C, 10% CO2 and 95% humidity.
Toxicant treatment and wash-out
To study repeated-dose effects, non-cytotoxic concentrations of rotenone and glyphosate were selected based on previously performed viability assays. Aggregates were exposed to rotenone, glyphosate or vehicle controls for 24 h on days 7, 9, 11 and 14. After 24h exposure, cells were washed to remove the exposure.
Viability assays
Resazurin assay was performed as described in Harris et al., 2017. Experiments were performed in three independent experiments with technical triplicates. LDH was measured in the media in control and treated samples following manufacturer’s instructions (CytoTox 96® NonRadioactive
Cytotoxicity Assay, Promega).
ATP assay
The bioluminescence ATP Assay Kit (Thermo Fisher Scientific, A22066) was used to determine the amount of intracellular ATP in aggregates according to manufacturer’s instructions. See
Supplementary Materials for details. Average luminescence values ± SEM were calculated from at least four independent experiments with technical triplicates.
239
Neurite outgrowth imaging and analysis
Red fluorescent protein (RFP) expressing-LUHMES (Schildknecht et al. 2013) were differentiated and treated as described above. On day 8 or day 15, aggregates were seeded on
MatrigelTM (BD Biosciences) pre-coated, flat-bottom, black 24 or 96-well plates (Thermo Fisher
Scientific). After 24 hours, wells were fixed in 4% PFA and imaged using a confocal microscope
(with open pinhole) and analyzed using Sholl Image J Software
(https://imagej.net/Sholl_Analysis). To analyze this data, the ratio was calculated for each shell
(number of intersections/distance from aggregate) and the mean plotted. Curves were compared using a quadratic non-linear regression fit with confidence intervals.
RNA extraction, reverse transcription, and real-time PCR
Total RNA was extracted using either TRIzol® Reagent (Life Technologies) followed by RNA
Clean & ConcentratorTM-Kit (Zymo Research®) or mirVana microRNA isolation kit (for microarray analysis) following the manufacturer’s instructions. Detailed description of cDNA synthesis and PCR is described in the Appendix, Supplementary Methods. Primers used for PCR are listed in Supplementary Table S1.
240
6.2. RESULTS
Repeated-dose effects of rotenone and glyphosate on 3D LUHMES viability
We used the resazurin assay to test cell viability after exposure to rotenone (30 nM) and glyphosate (5.18 µg/mL). Rotenone (30 nM), which has previously shown to be the free concentration of rotenone which inhibits 72 % complex I acitivity, had no effects on cell viability after first two exposures. After the third exposure, a 20 % decrease in viability was observed.
After repeated exposures, glyphosate (5.18 µg/mL) also led to a 20 % decrease in viability
(Figure 1a and 1b). The LDH assay indicated that rotenone led to low levels of cytotoxicity over time, which were only significant on day 12, while glyphosate showed no cytotoxicity (Figure 1c and 1d).
Rotenone repated-dose increases ATP levels and decreases neurite outgrowth.
After repeated low-dose rotenone exposures, ATP levels doubled in treated aggregates vs controls
(Figure 2a). As a functional endpoint, we measured neurite outgrowth. This was slightly impaired showing a decrease in the number of neurites, and length (Figure 2b). These effects were similar to those seen after recovery to 100 nM rotenone (Chapter 5).
Gene expression changes indicates PD-related genes play a role after in repeated low-dose response
Geneswhich have been found to play a role in PD were up-regulated after repeated-dose exposures to 30 nM rotenone, but not to 100 nM exposures on day 8 or day 15 (R100-DMSO and
DMSO-R100). Even a second exposure to 100 nM rotenone (R100-R100) did not alter the expression of these genes to the same extent indicating these may be a specific response to low- dose exposures (Figure 3).
241
6.3. FIGURES
Figure 6-1. Viability assays for repeated-dose toxicity of rotenone and glyphosate. Resazurin assay time- course for reonone (a) and day 15 viability for glyphosate (b). LDH assays for rotenone and glyphosate time- course experiments (c and d).
242
Figure 6-2. Effects of repeated-dose rotenone on ATP levels (a) and neurite outgrowth (b).
243
Figure 6-3. Changes in gene expression for PD-related genes (TH, PARK2, PARK7 and ASS1) after repeated-
dose exposures to rotenone.
244
CHAPTER 7
7. IMPACT OF GOLD AND POLY-LACTIC ACID
NANOPARTICLES ON 3D HUMAN BRAIN SPHEROID MODELS:
STUDYING BIOCOMPATIBILITY FOR BRAIN DRUG DELIVERY.
*The manuscript presented in this chapter has been submitted to ACS NANO.
Corrêa Leite P.R#, Pereira M.R#, Harris G#, Pamies D, Gobbo dos Santos L.M , Granjeiro J.M,
Hogberg H.T, Hartung T and Smirnova L. Suitability of 3D human brain spheroid models to distinguish toxic effects of gold and poly-lactic acid nanoparticles to assess biocompatibility for brain drug delivery
# shared first authorship, contributed equally
245
Graphical abstract
246
7.1. ABSTRACT
The brain blood barrier (BBB) is the bottleneck of brain-targeted drug development. Due to their physico-chemical properties, nanoparticles (NP) can cross the BBB and accumulate in different areas of the central nervous system (CNS), thus are potential tools to carry drugs and treat brain disorders. In vitro systems and animal models have demonstrated that some NP types promote neurotoxic effects such as neuroinflammation and neurodegeneration in the CNS. Thus, risk assessment of the NP is required, but current 2D cell cultures fail to mimic complex in vivo cellular interactions, while animal models do not necessarily reflect human effects due to physiological and species differences. We evaluated the suitability of in vitro models that mimic the human CNS physiology, studying the effects of metallic gold NP (AuNP) functionalized with sodium citrate (Au-SC), or polyethylene glycol (Au-PEG), and polymeric polylactic acid NP
(PLA-NP). Two different 3D neural models were used (i) human dopaminergic neurons differentiated from the LUHMES cell line (3D LUHMES) and (ii) human iPSC-derived brain spheroids (BrainSpheres). We evaluated NP uptake, mitochondrial membrane potential, viability, morphology, secretion of cytokines, chemokines and growth factors, and expression of genes related to ROS regulation after 24 and 72 h exposure. NP were efficiently taken up by spheroids, especially when PEGylated and in presence of glia. AuNP, especially PEGlated AuNP, effected mitochondria and anti-oxidative defense. PLA-NP were slightly cytotoxic to 3D LUHMES with no effects to BrainSpheres, demonstrating that different in vitro models can help identify how specific cell types are affected by NP.
247
7.2. INTRODUCTION
Nanoparticles (NP) can be synthesized from different materials and have generated increasing interest in nanomedicine, mostly as drug delivery systems 1, 2. NP capable of delivering drugs exhibit advantages such as drug stability, increased bioavailability and reduced drug concentrations required to reach the target, thus decreasing side effects 3. Some NP have the ability to cross the blood brain barrier (BBB) and reach the central nervous system (CNS), providing promising drug delivery systems (especially for the treatment of CNS diseases 4). In this sense, NP may serve as an important alternative for invasive CNS procedures such as implantation of catheters, imaging and therapy of brain tumors, and improvement of drug delivery 3, 5-7.
However, toxic effects of nanomaterials have been described 1, including induction of oxidative stress, inflammation 8, DNA damage, and alterations in gene expression 9. We found earlier, for example, that cobalt nanoparticles were able to induce cell transformation, a hallmark of cancer, in contrast to cobalt ions 10. NP can also promote neurotoxic effects such as neuroinflammation and neurodegeneration 11. Despite gold NP (AuNP) being one of the mostly commonly used nanomaterials (due to their versatility in particle size and surface modification) and their biocompatibility for applications such as drug and gene delivery 12, they can induce toxicity in different cellular models 13. In addition, studies have shown that AuNP induce apoptosis via caspase-dependent mechanisms as well as increase cell susceptibility to apoptosis induced by other agents 14, 15.
Recently, polylactic acid NP (PLA-NP) emerged as an alternative for drug delivery. PLA-NP allow sustained therapeutic drug levels for longer periods due to their polymeric matrix that prevents drug degradation, allowing better release kinetics. In addition, these NP exhibit biodegradable characteristics (unlike the metallic ones) 16. A detailed evaluation of neurotoxicity
248 mediated by different NP that are designed to be drug carriers in the CNS, is necessary and will contribute to the development of safer nanocarriers.
The current bioengineering revolution of cell culture makes organo-typic cell models, which overcome many limitations of traditional cell culture, increasingly available 17. Three- dimensional (3D) in vitro models are the most novel approach in this development, as they present closer cell-to-cell interactions, often include different cell types, and can better reproduce in vivo physiology 18-21. 3D CNS cultures have demonstrated advantages compared to two- dimensional (2D) cultures, such as increased cell survival and differentiation and better reproduction of electrical activity 18, 22. In addition, the use of 3D CNS models derived from human cells, especially human stem cells, provide more reliable data compared to animal models, due to species differences 23. However, such human 3D models have so far been rarely considered in the emerging test strategies for NP 24, 25.
Therefore, in this study we evaluated cellular effects of AuNP functionalized with sodium citrate
(Au-SC), polyethylene glycol (Au-PEG), and PLA-NP due to their potential use as drug delivery systems, using two human 3D CNS in vitro models: (i) 3D LUHMES (Lund human mesencephalic) spheroids and (ii) human iPSC-derived brain spheroids (BrainSpheres). Both models have shown to be reproducible in content, size and shape; forming spheroids of approximately 250 and 350 µm, respectively 26. LUHMES is an immortalized dopaminergic precursor cell line, derived from healthy human 8-week-old embryonic mesencephalic tissue, which rapidly differentiates into pure dopaminergic neurons 27, 28. BrainSpheres derived from human iPSCs is a multicellular 3D brain model that contains various types of neurons, astrocytes, and oligodendrocytes, and show spontaneous electrical activity and myelination 26, 29. By using these two models in parallel, we can compare a single-cell type 3D model with a more complex
3D system. After NP characterization, we evaluated NP uptake, morphological and molecular alterations such as viability and mitochondrial membrane potential, genes related to cytotoxicity
249 and oxidative stress and secretion of cytokines, chemokines, and growth factors in both 3D models.
7.3. MATERIALS AND METHODS
Nanoparticles and chemicals
Spherical monodisperse NP diluted in ultrapure water were used in this study. Au-SC were produced with 15 nm nominal diameter through reduction of 1% tetrachloroauric acid (Sigma
Chem. Co.) by 1% sodium citrate aqueous solution (Merck KGaA), based on the Turkevich method revised by Kimling 52. An Au-SC suspension was produced at 58 µg/mL and 1.7x1012 particles/mL concentration. Au-PEG 5-kDa at 1 mg/mL Au mass concentration (1015 particles/mL) were purchased from Nanocomposix (batch JMW1410). Green-fluorescent
Coumarin-6 PLA-NP at 6.3 mg/mL polymer concentration (1.05x1013 particles/mL) were acquired from IBCP (Lyon, France). All NP were stored and protected from light. Dimethyl sulfoxide (DMSO), paraformaldehyde (PFA) and cell lysis buffer (CelLytic M) were from Sigma
Aldrich.
Nanoparticle characterization
NP were sonicated for 5 min and particle suspensions (10 µL of 1µg/mL) were deposited on copper grids, air-dried and imaged in Tecnai G2 Spirit BioTwin 12 (Au-SC and Au-PEG) or LEO
912 Omega (PLA-NP) (FEI) transmission electron microscopes (TEM), both operated at 120 kV.
NP diameter size were determined by ImageJ software (https://imagej.nih.gov/ij/index.html NIH).
For analysis of NP hydrodynamic diameter, Au-SC (6 µg/mL), Au-PEG and PLA-NP (20 µg/mL each NP) were diluted in 3D LUHMES and BrainSpheres media and incubated in cell culture flasks without cells following the same conditions as for cell cultures. Then, 1 mL of sample was
250 transferred to an appropriate cuvette for subsequent analysis of dynamic light scattering (DLS) in the Malvern Zetasizer Nano ZS apparatus (Malvern Instruments Ltd).
3D LUHMES cell culture
Wild-type and red fluorescent protein (RFP) genetically modified LUHMES human neuronal precursor cells 53 were kindly provided by Prof. Marcel Leist (University of Konstanz) and maintained and cultured as previously described 27, 28, 54. Briefly, flasks were pre-coated with
50 μg/mL poly-L-ornithine and 1 μg/mL fibronectin (both from Sigma Aldrich) for 12h. Cells were maintained in Advanced DMEM/F12 (ThermoFisher) supplemented with 2 mM L- glutamine (Sigma Aldrich), 1x N2 (ThermoFisher) and 40 ng/mL recombinant basic Fibroblast
Growth Factor (bFGF, R&D Systems) (LUHMES proliferation medium), and passaged every 2-3 days. For 3D neuronal differentiation, cells were seeded in 6-well plates at 5 x 105 cells/well in 2 mL Advanced DMEM/F12 supplemented with 2 mM L-glutamine, 1x N2, 1 mM dibutyryl cAMP
(Santa Cruz), 2 μg/mL tetracycline (Sigma Aldrich) and 2 ng/mL recombinant human Glial cell line-Derived Neurotrophic Factor (GDNF, R&D Systems) (LUHMES differentiation medium).
The spheroids were placed on an orbital shaker (ES-X, Kuhner shaker) with 50 mm orbit diameter at 80 rpm in a humidified incubator at 37oC and 10% CO2. As per the differentiation protocol 54, paclitaxel (Sigma Aldrich) was added on day 3 to block proliferation and washed-out on day 5. On day 7, spheroid size was quantified using SPOT software 5.0 (Diagnostic
Instruments Inc). 3D LUHMES were differentiated up to 10 days.
BrainSpheres cell culture
Neural progenitor cells (NPC) were differentiated from iPSC 55 and kindly provided by Professor
Hongjun Song's lab within our joint project 26. iPSC were derived from C1 (CRL-2097) fibroblasts purchased from ATCC 55. NPC were maintained in KO DMEM/F12 medium
251 supplemented with 1x StemPro supplement (ThermoFisher), 20 ng/mL human bFGF
(ThermoFisher), 20 ng/mL Epidermal Growth Factor (EGF, ThermoFisher), 4 mM L-Glutamine
(ThermoFisher), 500 Units Penicillin and 500 g Streptomycin (ThermoFisher). Half of the medium was replaced every 24h. For BrainSpheres differentiation (previously described 26), cells were mechanically detached when reached 100% confluence and seeded in 6-well plates at 2x106 cells in 2 mL Neurobasal Electro medium (ThermoFisher) supplemented with B-27-electro
(ThermoFisher), 10 ng/mL Brain-Derived Neurotrophic Factor (BDNF) and 10 ng/mL GDNF
(Gemini), 4 mM L-glutamine (ThermoFisher), 500 Units Penicillin, and 500 g Streptomycin
(ThermoFisher). Cells were placed on an orbital shaker with 19 mm orbit diameter at 88 rpm into humidified incubator at 37oC and 5% CO2. Medium was replaced every 48h. After 4 weeks of differentiation, the BrainSpheres were used for the experiments. Spheroid size was quantified using SPOT software 5.0 (Diagnostic Instruments, Inc.).
NP treatment
NP stock suspensions were diluted in differentiation media on the day of treatment to prepare following final concentrations: 0.06, 0.6 and 6 µg/mL of Au-SC; 0.2, 2 and 20 µg/mL of Au-PEG and PLA nanoparticles. The used Au mass concentrations were in the range of previous in vitro studies 33-35. 3D LUHMES were treated with NP on day 7 of differentiation for 24 or 72 hours.
BrainSpheres were treated with NP after 4 weeks of differentiation for 24 or 72 hours. Then, spheroid and supernatant samples were collected for endpoint measurements.
Immunocytochemistry and confocal microscopy
3D cultures were fixed with 4% paraformaldehyde (PFA) for 1 h, washed 3 times with PBS and incubated for 2 h with blocking buffer (1% BSA, 5% goat serum, 0.15% saponin (Sigma
Aldrich)). Samples were incubated 48 h with primary antibodies (1:200 mouse anti-MAP2; 1:200
252 rabbit anti-GFAP and 1:200 mouse anti-Olig1, diluted in blocking buffer, from Chemicon, Dako and Millipore, respectively) at 4ºC, followed by three washing steps and incubation with secondary antibodies (1:500 goat anti-mouse or 1:500 goat anti-rabbit Alexa fluor 568, diluted in blocking buffer, Molecular Probes) for overnight. Then, samples were washed and incubated with
Hoechst 33342 (1:10000, 1 μg/mL, Molecular Probes) for at least 1 h at room temperature (RT).
After three washing steps, the samples were mounted on glass slides with Prolong Gold-antifade reagent (Molecular Probes) for confocal microscopy. Z-stacks started at the top of the sample with 100-200 nm intervals were taking. Images were obtained with Zeiss LSM-510 (Zeiss) and
Leica SP5 (Leica) confocal microscopes with identical time exposure and image settings.
Flow cytometry
3D RFP-expressing LUHMES treated with different concentrations of PLA-NP for 72 h were trypsinized with TryplE Express containing 4 units/mL DNase at 37 oC for 30 min on the shaker.
Then, samples were homogenized using a 1 mL syringe with a 26G3/8 needle. Cells were washed with PBS twice, fixed with 4% PFA for 30 min and the co-localization between 3D RPF- expressing LUHMES cells and PLA-NP green fluorescence was quantified using a FACSCalibur flow cytometer (BD). The instrument was calibrated using fluorescent beads and wild-type
LUHMES cells were used as negative control to set the gates.
Inductively coupled plasma mass spectrometry (ICP-MS)
AuNP stock solution was used to prepare the calibration solutions through serial aqueous dilutions for ICP-MS determination (NexION 300D, PerkinElmer). 1% (v/v) nitric acid (Merck) and 10 µg/L rhodium from Perkin-Elmer was added to the calibration solution, blank, and samples to improve analytical performance. The Au197 isotope was measured. For quantification of intracellular Au mass, spheroids of both models were treated for 24 or 72 h with Au-SC or Au-
253
PEG. Then, spheroids were collected and lysed (CelLytic M lysis buffer, Molecular Probes) for analysis.
Mitochondrial membrane potential assay
After NP treatment, mitochondrial membrane potential (ΔΨm, MMP) was analyzed using
MitoTracker® Red CMXRos (ThermoFisher), according to manufacturer’s recommendations.
After 45 min treatment with MitoTracker Red CMXRos reagent, spheroids were fixed with 4%
PFA for 1h at RT, washed with PBS and mounted on glass slides. Images were acquired with fluorescence microscope (Olympus BX60) and red fluorescence intensity was quantified by
ImageJ software (https://imagej.nih.gov/ij/index.html, NIH).
Lactate dehydrogenase (LDH) release assay
LDH release was determined by colorimetric CytoTox 96 Cytotoxicity Assay kit (Promega). As positive control, spheroids were treated with 1% TX-100 for 30 min. After NP treatments, 20 µL of supernatants were transferred to 96-well plates followed by the addition of 20 µL of substrate solution. After 30 min of incubation in the dark at RT, 20 µL of stop solution was added to each sample. Color development was proportional to the number of cells with disruption of plasma membrane. Absorbance was measured at 490 nm. For evaluation of eventual colorimetric interference, NP diluted in culture medium were incubated with LDH positive control and substrate according to manufacturer’s instructions.
Scanning electron microscopy (SEM)
Spheroids were fixed with 2.5% glutaraldehyde for 1h at RT. After three washing steps with PBS, samples were postfixed with 1% osmium tetroxide for 90 min in the dark, followed by three washing steps with 0.1M cacodilate buffer (pH 7.4) and distilled water. Samples were gradually
254 dehydrated, dried at the critical-point and finally vaporized with platinum. Photomicrographs were obtained at 3 kV in a SEM Helios Nanolab 650 (FEI, ThermoFisher).
RNA extraction and quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA was extracted from 3D cultures after NP treatments using Tripure isolation reagent
(Roche) according to Chomczynski and Sacchi 56. RNA quantity and purity was determined using NanoDrop 2000c (ThermoFisher). One microgram of RNA was reverse-transcribed using the M-MLV Promega Reverse Transcriptase (Promega) according to the manufacturer’s recommendations. The expression of genes was evaluated using specific TaqMan gene expression assays (ThermoFisher). qRT-PCRs were performed using a 7500 Fast Real Time system (Applied
Biosystems). Relative mRNA expression was calculated using the 2−ΔΔCt method 57. The analyzed genes were: SOD1 (Hs00533490_m1), NF2L2 (Hs00975961_g1), NFR1
(Hs00602161_m1), GSTO1 (Hs02383465_s1), CLEC7A (Hs01902549) and SOD2
(Hs00167309_m1). β-actin was used as housekeeping gene.
Analysis of multiple secreted mediators
Determination of cytokines, chemokines and growth factors secreted by 3D LUHMES and
BrainSpheres cultures upon NP exposure was carried through Luminex (Austin TX, USA) xMAP magnetic technology for the following analytes: IL-1β, IL-1ra, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8,
IL-9, IL-10, IL-12 (p70), IL-13, IL-15, IL-17, eotaxin, bFGF, GCSF, GM-CSF, IFN-γ, IP-10,
MCP-1 (MCAF), MIP-1α, MIP-1β, PDGF-BB, RANTES, TNFα and VEGF, and TGF-β1, TGF-
β2 and TGF-β3. Analysis was performed following the manufacturer’s recommendations. Briefly, after calibration and validation of Bio-Plex Magpix (Bio-Rad), reagent reconstitution and standard curve preparation, magnetic beads were added to each well of the assay plate. Each step was preceded by washing steps using an automated Bio-Plex Pro wash station (Bio-Rad). Then,
255 samples, standard and controls were added, followed by detection antibodies and streptavidin-PE.
Finally, magnetic beads were re-suspended and read. The number of analytes detected in culture medium without spheres (background) was subtracted from the samples, allowing to access the protein levels secreted by cultures.
256
7.4. RESULTS
NP characterization
In order to assess NP diameter, micrographs of Au-SC (Fig. 1A), Au-PEG (Fig. 1B) and PLA-NP
(Fig. 1C) were acquired by TEM. Diameter of Au-SC, Au-PEG and PLA-NP was 17.5 ± 1.4, 5.4
± 0.7 and 67 ± 3 nm, respectively (Fig. 1D). Hydrodynamic diameter of 6 µg/mL Au-SC and 20
µg/mL PLA-NP was assessed by DLS in medium without cells just after dilution and was 27 ± 1 and 116 ± 1.5 nm (0 h, 3D LUHMES medium) and 21 ± 2 and 62 ± 2 nm (0 h, BrainSpheres medium), respectively. Although, Au-SC and PLA-NP were larger in 3D LUHMES medium at 0 h, the diameters of these NP remained stable after 24 and 72 h. In contrast, Au-SC and PLA-NP diameters increased over time (Au-SC: 29 nm at 24 h to 38 nm at 72 h; PLA-NP: 99 nm at 24 h to
137 nm at 72 h) in BrainSpheres differentiation medium (Fig 1E). Due to the small diameter of
Au-PEG (~5 nm), they fall below the accurate range for DLS and therefore this measurement was not performed for this NP type.
3D LUHMES and BrainSpheres production
3D LUHMES and BrainSpheres with controlled size (up to 350 µm) and without necrotic center were produced by constant gyratory shaking 26. At day 7 of differentiation, 3D LUHMES expressing RFP (red fluorescent protein) exhibited neuronal morphology with several processes
(Fig 2A). At 4 weeks of differentiation, BrainSpheres expressed different neural markers consistent with phenotypes of neurons (MAP2, Fig 2B), astrocytes (GFAP, Fig 2C) and oligodendrocytes (Olig1, Fig 2D).
NP uptake in 3D LUHMES and BrainSpheres
Since PLA-NP have green fluorescence property, we analyzed their internalization by confocal imaging and flow cytometry. After 72 h exposure, uptake of 20 µg/mL PLA-NP by BrainSpheres
257
(Fig 3A) and 3D LUHMES (Fig 3B) was visualized by confocal imaging. For better visualization,
BrainSpheres were stained for neuronal (MAP-2), astrocyte (GFAP) and oligodendrocyte (OLIG-
1) markers.
In BrainSpheres, PLA-NP were distributed more evenly throughout spheroids reaching the core
(Fig 3A). Despite most NP were detected close to the border of 3D LUHMES, we also observed
PLA-NP in the spheroid core (Fig 3B). PLA-NP were co-localized with the antibody staining or
RFP in both models not only on the outside layer of spheroids, but also in the cells within the core of spheroids (Fig 3A, B).
As 3D LUHMES contain a single RFP-labelled cell type, we further quantified PLA-NP internalization in RFP-LUHMES by flow cytometry. 3D LUHMES were exposed to 0, 0.2, 2, and
20 µg/mL PLA-NP for 72 h, then spheroids were dissociated and fixed for subsequent analysis of co-localization between RFP (red) and PLA-NP (green) using flow cytometry. Results indicated increased co-localization between PLA-NP and RFP-expressing LUHMES cells in a concentration-dependent manner. Only 7.2 % cells showed co-localization after exposure to 0.2
µg/mL while 96 % cells co-localized with PLA-NP at 2 µg/mL, reaching 100 % at 20 µg/mL (Fig
3C). These data showed that PLA-NP were internalized by cells throughout the spheroids at 2 and
20 µg/mL. Due to the complex composition of BrainSpheres containing multiple unlabeled cell types, we did not assess PLA-NP uptake by flow cytometry in this model.
Internalization of Au-SC and Au-PEG in both 3D models was assessed by ICP-MS. Our results showed Au-NP uptake in both models after 24 and 72 h exposure. Uptake of both Au-NP types was delayed in 3D LUHMES (increased after 72 h vs. 24 h exposure). In BrainSpheres, Au-SC uptake increased after 72 h compared to 24 h, while levels of internalization of Au-PEG at 24 and
72 h exposure time points were similar (Fig. 3D).
258
Effects of NP on mitochondrial membrane potential (MMP) and cell viability in both 3D human neural models
Both 3D LUHMES and BrainSpheres were exposed to PLA, Au-SC, and Au-PEG at different concentrations for 24 or 72 h, followed by analysis of MMP using MitoTracker® Red CMXRos
(Fig. 4A, B). After 24 h exposure to 0.6 and 6 µg/mL Au-SC, MMP was reduced in both models in a concentration-dependent manner (Fig. 4B). After 72 h, a further decrease in MMP was observed in 3D LUHMES while BrainSpheres showed reduction only at 6 µg/mL. After 24 h exposure, only the highest concentration of Au-PEG (20 µg/mL) led to reduced MMP levels in
3D LUHMES, with no significant reduction at 72 h exposure. On the contrary, BrainSpheres showed a strong concentration-dependent reduction of MMP after 24 and 72 h exposure to Au-
PEG. PLA-NP reduced MMP in 3D LUHMES at all concentrations after 24 h exposure, which reversed at 0.2 µg/mL and further decreased at 2 and 20 µg/mL after 72 h exposure. MMP of
BrainSpheres was not affected by any of the PLA-NP tested concentrations. Taken together, we observed stronger acute than prolonged effects of tested NP on mitochondria functionality in
BrainSpheres, while 3D LUHMES sensitivity to these NP was generally amplified with increased exposure concentrations and time. These data suggest different susceptibility of these 3D models to the studied nanomaterials, most probably due to the cellular complexity of the models (Fig. 4A,
B).
To determine whether the observed effect on mitochondria was linked to cytotoxicity, LDH release was measured at the highest concentrations of NP. The viability of 3D LUHMES was significantly affected (~25 % cell death) by Au-PEG and PLA-NP after 72 h exposure. The studied NP did not affect BrainSpheres viability at the tested concentrations (Fig. 4C).
Morphology changes in 3D LUHMES and BrainSpheres after NP treatment
259
The morphology of both spheroid models was analyzed by scanning electron microscopy (SEM) after 72 h treatment with the highest NP concentrations (Au-SC (6 µg/mL), Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL)). Morphology of 3D LUHMES was affected by Au-PEG and PLA-NP, but not by Au-SC. Significant cell debris attached to the spheroids and less neuronal projections were observed (Fig. 5, depicted with arrowheads). BrainSpheres did not show any alterations under the investigated conditions, which correlates with above cytotoxicity data (Fig 5).
NP effect on expression of oxidative stress related genes in BrainSpheres
Since NP exposure affected mitochondria functionality in BrainSpheres without affecting viability, we further investigated changes in gene expression related to ROS regulation as a possible mechanism of toxicity: SOD1, SOD2, NF2L2, GSTO1, and NFR1 are involved in antioxidant responses and CLEC7A is associated with ROS production and inflammation.
BrainSpheres were exposed to Au-SC (6 µg/mL), Au-PEG (20 µg/mL), and PLA-NP (20 µg/mL) for 72 h, then gene expression was analyzed by Real-Time qPCR. SOD1 expression was increased by Au-SC and Au-PEG, but not PLA, whereas NF2L2 was increased only by Au-PEG.
NFR1 expression was up-regulated by Au-PEG and PLA-NP, while expressions of GSTO1 and
CLEC7A were up-regulated by all NP. SOD2 expression was not altered upon NP challenge (Fig.
6). These data showed that the exposure to NP increased expression of genes related to oxidative stress protection in BrainSpheres, and the strongest effect was observed with Au-PEG, in line with the higher uptake of this NP.
NP influence on release of chemokines, cytokines and growth factors in 3D human neural models
Analysis of multiple secreted products from both 3D models exposed to 6 µg/mL Au-SC, 20
µg/mL Au-PEG or 20 µg/mL PLA-NP for 24 and 72 h showed alterations in the levels of some
260 mediators (Fig. 7). In general, neural cells produce lower levels of chemokines and cytokines compared to cells of the immune system, but such levels are critical to maintaining their homeostasis and, consequently, the microenvironment. Thus, any imbalance may affect their physiological behavior 30.
From all conditions tested, Au-PEG had the strongest effect on cytokines release in 3D LUHMES cultures. It significantly downregulated the levels of all cytokines tested. No significant changes in cytokine release were observed in cultures treated with Au-SC and PLA-NP (Fig. 7A).
BrainSpheres were not as sensitive to Au-PEG with only two cytokines (IL-10, 12p7) and two growth factors (bFGF, VEGF) downregulated in these cultures, while IL-1ra was upregulated by
Au-PEG and PLA-NP. In addition, Au-SC reduced the levels of IL12p70 and VEGF in
BrainSpheres (Fig. 7B).
TGF-β1, TGF-β2, and TGF-β3 levels were not significantly altered in 3D LUHMES
(Supplementary Figure 1). In BrainSpheres, TGF-β1 levels were reduced by Au-SC and Au-PEG after 72 h. These NP also eliminated TGF-β3 in the first 24 h, but not after 72 h. PLA-NP did not affect the levels of these soluble mediators (Fig. 8). Altogether, these data suggested different responses in the profile of chemokines, cytokines, and growth factors to the studied nanomaterials in both models, which are both devoid of immune cells. The very low levels and frequent reductions in cytokine levels compared to control are difficult to interpret. However, there seems to some correspondence of cytokine patterns in response to the NP between the two models, which corroborates the findings and could support mechanistic reasoning. At this stage, these findings mainly illustrate that cytokines could be highly sensitive quantitative biomarkers for perturbation of cells by NP.
261
7.5. FIGURES
Figure 7-1. NP characterization. Representative images obtained by TEM of (A) Au-SC, (B) Au-PEG and
(C) PLA-NP. (D) Quantification of NP size by TEM. (E) Z‑average hydrodynamic diameter values of 6
µg/mL Au-SC and 20 µg/mL PLA-NP diluted in LUHMES and BrainSpheres culture media, analyzed by
DLS after 0, 24 and 72 h of incubation. Results are expressed as mean (±SD). Each experimental group corresponds to the analysis of three independent experiments with three replicates. Statistical significance was analyzed by one-way ANOVA followed by Bonferroni’s multiple comparisons post-test (**p < 0.01,
***p < 0.001). Scale bar (A) 100 nm, (B) 25 nm and (C) 200 nm.
262
Figure 7-2. Confocal Images of (A) RFP-expressing LUHMES after 7 days of differentiation; and 4-week
BrainSpheres expressing different neural cell type markers: (B) neuronal MAP2, (C) astrocytic GFAP and
(D) oligodendrocyte-specific Olig1. Scale bars 100 µm.
263
Figure 7-3. NP uptake by 3D LUHMES and BrainSpheres. Representative images acquired by confocal microscope after 72 h exposure to 20 µg/mL PLA-NP (green) of (A) BrainSpheres stained with MAP-2,
GFAP, OLIG-1 (red) and Hoechst33342 (blue nuclei); and (B) 3D RFP-LUHMES (red) (left panels). Right panels correspond to magnification of left panel images. Cross-hair lines in the Z-stack images evidence
PLA-NP internalization. (C) Flow cytometry Scatter plots of wild type (WT) LUHMES (negative control), and RFP-LUHMES, treated with 0, 0.2, 2 and 20 µg/mL PLA-NP. Red and green fluorescence was quantified in FL2 and FL1 channels, respectively. (D) Intracellular levels of Au in 3D LUHMES and
264
BrainSpheres after 24 and 72 h exposure to 6 µg/mL Au-SC and 20 µg/mL Au-PEG quantified by ICP-MS.
Results are expressed as mean (±SEM). Each experimental group corresponds to the analysis of three independent experiments with three replicates. Student’s t-test was used to compare 24 and 72 h treatment groups (*p < 0.05, ***p < 0.001). Scale bars 100 µm (left panels) and 25 µm (right panels).
Figure 7-4. NP effect on MMP and cell viability in 3D human neural models. (A) Representative images of
MitoTracker® staining in 3D LUHMES and BrainSpheres after 24 and 72 h exposure to Au-SC (6 µg/mL),
Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL). (B) MMP levels in 3D LUHMES and BrainSpheres after 24 and 72 h exposures to Au-SC (0.06, 0.6 and 6 µg/mL), Au-PEG (0.2, 2, 20 µg/mL) and PLA-NP (0.2, 2 and
20 µg/mL) normalized to the untreated controls. (C) Percent of LDH release in 3D LUHMES and
BrainSpheres models after 24 and 72 h exposures to Au-SC (6 µg/mL), Au-PEG (20 µg/mL) and PLA-NP
(20 µg/mL). ‘- Control’ corresponds to LDH released from untreated 3D human CNS models. Results are
265 expressed as mean (±SEM). Each experimental group corresponds to three independent experiments imaging at least 20 spheroids. One-way ANOVA with Bonferroni’s multiple comparisons post-test was used for analysis of statistical significance (*p < 0.05, **p < 0.01, ***p < 0.001).
Figure 7-5. Morphology of 3D LUHMES and BrainSpheres exposed to Au-SC (6 µg/mL), Au-PEG (20
µg/mL) and PLA-NP (20 µg/mL) for 72 h. White arrowheads indicate cells in degeneration. Scale bars 10
µm.
Figure 7-6. Effect of NP on expression of genes related to ROS regulation in BrainSpheres. Graphs showing the relative expression of SOD1, SOD2, NF2L2, GSTO1, NFR1 and CLEC7A after exposure to Au-SC (6
266
µg/mL), Au-PEG (20 µg/mL) and PLA-NP (20 µg/mL) for 72 h. Data was collected from three independent experiments with three technical replicates and represents fold changes (± SEM). One-way ANOVA with
Bonferroni’s multiple comparisons post-test was used to analyze the statistical significance (*p < 0.05, **p <
0.01, ***p < 0.001).
Figure 7-7. NP influence release of chemokines, cytokines and growth factors in 3D human neural models.
Graphs showing the levels of different secreted mediators after exposure to Au-SC (6 µg/mL), Au-PEG (20
µg/mL) and PLA-NP (20 µg/mL) for 24 and 72 h. (A) 3D LUHMES (MIP-1β, IL-10, IL-12p70, TNFα,
267 bFGF and VEGF) and (B) BrainSpheres (IL-1ra, IL-10, IL-12p70, GM-CSF, bFGF and VEGF). Data were collected from three independent experiments with three technical replicates and represents mean (± SEM).
One-way ANOVA with Bonferroni’s multiple comparisons post-test was used to analyze the statistical significance (*p < 0.05).
Figure 7-8. Influence of NP on release of TGF-β isoforms in BrainSpheres. Graphs show the levels of
secreted TGF-β1, TGF-β2 and TGF-β3 after exposure to Au-SC (6 µg/mL), Au-PEG (20 µg/mL) and PLA-
NP (20 µg/mL) for 24 and 72 h. Each experimental group corresponds to the analysis of three independent
experiments with three replicates and represents mean (± SEM). One-way ANOVA with Bonferroni’s
multiple comparisons post-test was used to analyze the statistical significance (*p < 0.05, ***p < 0.001).
Supplementary Figure 7-9. Influence of NP on release of TGF-β isoforms in 3D LUHMES. Graphs show the levels of secreted TGF-β1, TGF-β2 and TGF-β3 after exposure to Au-SC (6 µg/mL), Au-PEG (20
µg/mL) and PLA-NP (20 µg/mL) for 24 and 72 h. Each experimental group corresponds to the analysis of three independent experiments with three replicates and represents mean (± SEM).
268
7.6. DISCUSSION
The use of NP as carriers promises advantages such as better drug stability, bioavailability, improved dosing and reduced side effects. Many NP drug delivery systems are being developed, but nanotoxicity must be controlled before administrating them 3, 31. Despite AuNP being widely used for drug delivery, recent studies have shown their potential to induce (neuronal) cytotoxicity and neuroinflammation 13, 14. PLA-NP emerged as an alternative due to their biocompatibility, biodegradability, and drug release kinetics 32.
NP derived from different materials are able to cross the BBB, reaching the CNS and becoming excellent potential drug carriers for reaching the cerebral parenchyma. However, reported AuNP toxicity to the CNS includes apoptosis and alterations in retinal layers structure, microglia activation, increase of neuronal excitability in CA1 region of hippocampus and the potential to aggravate seizure activity 33-35. The use of PLA-NP for drug delivery is a relatively recent approach, showing positive results in treating hearing loss and Alzheimer’s disease in animal models, without inducing toxicity 36, 37.
Human-derived 3D neural in vitro models, which mimic complex human CNS interactions are a promising tool 26, 38 that can be applied to study the potential cytotoxic effects of nanocarriers.
To characterize the interaction between NP, media and cells in vitro, we assessed NP size and internalization. Our NP characterization showed that Au-SC and PLA-NP did not change in size over time in LUHMES differentiation medium. Although, immediately after resuspension of NP in LUHMES differentiation medium, NP size was higher than previously measured by TEM in water, they were then stable in size for 24 and 72 h (Fig. 1C). However, in BrainSpheres differentiation medium, we observed a gradual increase in NP size for both Au-SC and PLA-NP.
In both media, the immediate (3D LUHMES) or delayed (BrainSpheres) increase in NP size can be attributed to aggregation or the formation of a protein corona due to NP interaction with proteins in cell culture medium. It is possible that proteins from the medium supplements attach
269 to the NP, affecting their hydrodynamic analysis. More work is needed within the characterization field to better predict how these media alter NP physico-chemical properties.
Due to the proprietary composition of cell culture supplements, further analysis to identify these proteins was not possible. However, a protein corona could facilitate cellular NP uptake, as is known to occur in blood circulation 39. Studying the composition of the protein corona would be necessary to assess and improve drug delivery for CNS therapies. A recent study showed that artificial apolipoprotein E4 adsorption to NP forming corona increases NP translocation through
BBB, improving the brain parenchyma accumulation threefold when compared to undecorated particles 40.
A major question for the use of 3D organotypic cultures to assess NP toxicity is, whether the NP can actually penetrate the spheroid and reach the inner cell mass. The 3D models, studied here, were, indeed, able to take up the three NP types. As quantified by flow cytometry, PLA-NP were internalized by LUHMES in a dose-dependent manner, with 2 µg/mL being sufficient to penetrate 96% of cells. These results support the confocal images that showed PLA-NP in the core of both 3D models. Au-SC uptake was similar in both 3D models with increased levels after
72 h. Au-PEG showed the same result as Au-SC in 3D LUHMES, but the uptake in BrainSpheres was 5-fold higher after 24 h and stabilized after 72 h. The differences in Au-PEG internalization between the 3D models could be attributed to the presence of glial cells in BrainSpheres that may take up these NP more efficiently than neurons and/or regulate neuronal uptake as previously shown 41. In fact, PEG functionalization contributes to a better NP uptake by neural cells, and their smaller diameter may be the reason for higher intracellular levels compared to Au-SC in
BrainSpheres. This is in line with studies showing that PEGylation increases NP accumulation in the brain compared to non-PEG-coated NP 3, 42. Therefore, PEG-coating and the presence of glial cells facilitating uptake could explain the high Au-PEG accumulation in BrainSpheres in the first 24 h and their maintenance until 72 h.
270
Due to the efficient NP internalization by neural cells in the 3D models, their potential to induce nanotoxicity needs to be considered. In 3D LUHMES, Au-SC but not Au-PEG significantly reduced mitochondria function (MMP). Au-PEG had only slight acute effects at the highest concentration tested, which was no longer significant after 72 h exposure. However, Au-PEG increased LDH release, possibly due to the high contact surface area of this NP that may contribute to neuronal lysis. In BrainSpheres, MMP was significantly reduced by both AuNP but no increase in LDH release was observed. In this model, glial cells may be involved in the AuNP clearance, reducing their availability and damage to neuronal cells.
PLA-NP reduced MMP and increased LDH release in 3D LUHMES, but showed no effect in
BrainSpheres. The 3D LUHMES model is a monoculture model, consisting only of dopaminergic neurons, and may display higher susceptibility to harmful agents that induce mitochondrial damage, oxidative stress, and neuroinflammation 43. In fact, NP cytotoxicity has shown to be lower when neurons and astrocytes are co-cultured, compared to monocultures 44. The absence of neuroprotective glial cells, together with the known susceptibility of dopaminergic neurons to oxidative stress, could contribute to the toxicity observed in this model from all studied NP. On the other hand, astrocytes in BrainSpheres as well as polymeric composition of PLA-NP may be critical to increase NP biocompatibility and cell tolerance.
Mitochondrial dysfunctions are associated with ROS production, promoting cell stress and death
45. NP can mediate this deregulation in neural cells 46. Since AuNP decreased MMP without cell viability loss in BrainSpheres, we investigated the expression of genes involved in antioxidant responses and ROS production. Although all studied NP led to increased CLEC7A expression, related to ROS production and inflammation, SOD1, NF2L2, GSTO1, and NFR1, related to antioxidant responses in general, were increased by Au-PEG and PLA-NP but not by Au-SC.
This activation of antioxidant genes may indicate that the cell activated antioxidant response as a reaction to the NP challenge.
271
These results showed that the studied NP do not necessarily induce toxicity directly but are able to unbalance cell physiology. Amongst these NP, Au-PEG had the strongest effect, in line with their increased uptake.
In BrainSpheres, exposure to NP did not decrease viability or lead to any apparent morphological alterations. However, AuNP affected mitochondrial activity and increased antioxidant genes with a possible activation of cell survival programs, which may be sufficient to maintain cell viability and morphology. In contrast, 3D LUHMES exposed to Au-PEG and PLA-NP showed morphological alterations on the surface of spheroids, which was confirmed by cell death measured by LDH release. These results reinforce the impact of glial cells in NP tolerance to toxic insults.
Glial and neuronal cells produce and secrete factors to maintain survival. Some of them are considered pro-inflammatory, but also important at physiological levels for neuron-glia communication 47. The level of these factors can be altered in response to harmful agents. Au-
PEG reduced most detected cytokines, chemokines, and growth factors in both 3D models. In 3D
LUHMES, a reduction in MIP-1β, IL-10, IL-12p70, TNFα, bFGF and VEGF levels are in line with the observed effects on viability. In BrainSpheres, Au-PEG reduced the levels of IL-10, IL-
12p70, bFGF, VEGF, and TGF-β1, as well as increased the IL-1ra levels. Au-SC affected fewer secreted mediators, probably due to their smaller contact area and low cell penetration rate compared to PEGylated AuNP. In CNS, IL-12 is manly related to pathogenesis of autoimmune diseases; VEGF to neurogenesis, neuronal migration, neuroprotection, and blood vessel growth; and TGF-β1 displays neuroprotective role and promotes glial scar and fibrosis, induced by acute and chronic brain injury 48-50. IL-10 is an important anti-inflammatory mediator. The reduced levels of such mediators suggest that AuNP in a first interaction with neural cells may turn them susceptible to other subsequent harmful agents in the CNS. However, such alterations together with reduced MMP were not sufficient to induce cell death in BrainSpheres, possibly due to the
272 increased expression of antioxidant genes and the neuroprotective role of astrocytes that may maintain cell survival.
PLA-NP had minimal effects on secreted cytokines, chemokines, and growth factors, restricted to
IL-1ra increase in BrainSpheres. IL-1ra binds to IL-1 receptor in cell membranes, preventing IL-1 downstream signaling and activation of inflammation 51. This change, however, was small and may not be relevant for the cell functionality.
Altogether, we consider PLA-NP a stealth nanomaterial in multicellular 3D human neural models, containing neuron and glial cells, without altering cell physiology and functionality and suitable for carrying drugs of interest in the CNS.
273
6.7. CONCLUSIONS
This work showed that 3D brain spheroid models are well suited to comparatively characterize
NP neurotoxicity. The simpler, single-cell model was more sensitive to the toxic effects, in line with the lack of glia support to neurons. The use of multiple models, which encompass simplicity and physiological relevance, serves as tool for more NP drug-delivery focused research.
In conclusion, we have demonstrated that AuNP, the most used nanocarrier, as well as PLA-NP, may be harmful to pure dopaminergic neurons at the highest concentrations tested, which might represent a risk to contribute to Parkinson’s disease development. In particular, in midbrain substantia nigra, where there is a high number of dopaminergic neurons, the use of nanocarriers should be carefully evaluated. However, this risk might be overestimated as the model lacks neuroprotection provided by astrocytes. Thus, BrainSpheres, are more appropriate model to study general neurotoxicity because they contain the glia, which can provide such neuronal support.
In a mixed population of neural cells within the BrainSpheres, the studied nanocarriers did not affect viability, likely due to the presence of glial cells and their participation in brain clearance.
Our findings showed that AuNP promoted alterations in cell physiology that may contribute to increased susceptibility to other subsequent harmful agents in the CNS. Therefore, the use of
AuNP as drug carrier in the CNS must be further evaluated. PLA-NP induced minor alterations in
BrainSphere neural cell physiology, emerging as a safer alternative for brain drug delivery.
274
6.7 ACKNOWLEDGEMENTS
We are very grateful for MSc. Vania da Silva Vieira for the essential assistance in scanning electron microscope, and Dr. Emile Santos Barrias, for the valuable assistance in fluorescence confocal microscope for image acquisition of Figure 3, both from INMETRO. This study was supported by grants from FAPERJ (Fundação de Amparo à Pesquisa do Rio de Janeiro), CNPQ
(Conselho Nacional de Desenvolvimento Científico e Tecnológico), and CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior). Funding bodies were not involved in study design, data collection, analysis, interpretation, or writing of the manuscript.
275
6.8. REFERENCES
1. Wu, T.; Tang, M. Review of the effects of manufactured nanoparticles on mammalian target
organs. Journal of applied toxicology : JAT 2017.
2. Zhang, X. D.; Wu, D.; Shen, X.; Liu, P. X.; Yang, N.; Zhao, B.; Zhang, H.; Sun, Y. M.;
Zhang, L. A.; Fan, F. Y. Size-dependent in vivo toxicity of PEG-coated gold nanoparticles.
International journal of nanomedicine 2011, 6, 2071-81.
3. Leite, P. E.; Pereira, M. R.; Granjeiro, J. M. Hazard effects of nanoparticles in central
nervous system: Searching for biocompatible nanomaterials for drug delivery. Toxicology in
vitro : an international journal published in association with BIBRA 2015, 29, 1653-60.
4. Bhaskar, S.; Tian, F.; Stoeger, T.; Kreyling, W.; de la Fuente, J. M.; Grazu, V.; Borm, P.;
Estrada, G.; Ntziachristos, V.; Razansky, D. Multifunctional Nanocarriers for diagnostics,
drug delivery and targeted treatment across blood-brain barrier: perspectives on tracking and
neuroimaging. Particle and fibre toxicology 2010, 7, 3.
5. Dinda, S. C.; Pattnaik, G. Nanobiotechnology-based drug delivery in brain targeting. Current
pharmaceutical biotechnology 2013, 14, 1264-74.
6. Cheng, Y.; Morshed, R. A.; Auffinger, B.; Tobias, A. L.; Lesniak, M. S. Multifunctional
nanoparticles for brain tumor imaging and therapy. Advanced drug delivery reviews 2014,
66, 42-57.
7. Tam, V. H.; Sosa, C.; Liu, R.; Yao, N.; Priestley, R. D. Nanomedicine as a non-invasive
strategy for drug delivery across the blood brain barrier. International journal of
pharmaceutics 2016, 515, 331-342.
8. Delgado-Buenrostro, N. L.; Medina-Reyes, E. I.; Lastres-Becker, I.; Freyre-Fonseca, V.; Ji,
Z.; Hernandez-Pando, R.; Marquina, B.; Pedraza-Chaverri, J.; Espada, S.; Cuadrado, A.;
Chirino, Y. I. Nrf2 protects the lung against inflammation induced by titanium dioxide
276
nanoparticles: A positive regulator role of Nrf2 on cytokine release. Environmental
toxicology 2015, 30, 782-92.
9. Li, J. J.; Hartono, D.; Ong, C. N.; Bay, B. H.; Yung, L. Y. Autophagy and oxidative stress
associated with gold nanoparticles. Biomaterials 2010, 31, 5996-6003.
10. Hartung, T.; Sabbioni, E. Alternative in vitro assays in nanomaterial toxicology. Wiley
interdisciplinary reviews. Nanomedicine and nanobiotechnology 2011, 3, 545-73.
11. Mushtaq, G.; Khan, J. A.; Joseph, E.; Kamal, M. A. Nanoparticles, Neurotoxicity and
Neurodegenerative Diseases. Current drug metabolism 2015, 16, 676-84.
12. Nicolardi, S.; van der Burgt, Y. E. M.; Codee, J. D. C.; Wuhrer, M.; Hokke, C. H.; Chiodo, F.
Structural Characterization of Biofunctionalized Gold Nanoparticles by Ultrahigh-Resolution
Mass Spectrometry. ACS nano 2017, 11, 8257-8264.
13. Chueh, P. J.; Liang, R. Y.; Lee, Y. H.; Zeng, Z. M.; Chuang, S. M. Differential cytotoxic
effects of gold nanoparticles in different mammalian cell lines. Journal of hazardous materials
2014, 264, 303-12.
14. Leite, P. E.; Pereira, M. R.; do Nascimento Santos, C. A.; Campos, A. P.; Esteves, T. M.;
Granjeiro, J. M. Gold nanoparticles do not induce myotube cytotoxicity but increase the
susceptibility to cell death. Toxicology in vitro : an international journal published in
association with BIBRA 2015, 29, 819-27.
15. Engin, A. B.; Nikitovic, D.; Neagu, M.; Henrich-Noack, P.; Docea, A. O.; Shtilman, M. I.;
Golokhvast, K.; Tsatsakis, A. M. Mechanistic understanding of nanoparticles' interactions
with extracellular matrix: the cell and immune system. Particle and fibre toxicology 2017, 14,
22.
16. Vasir, J. K.; Labhasetwar, V. Biodegradable nanoparticles for cytosolic delivery of
therapeutics. Advanced drug delivery reviews 2007, 59, 718-28.
17. Pamies, D.; Hartung, T. 21st Century Cell Culture for 21st Century Toxicology. Chemical
research in toxicology 2017, 30, 43-52.
277
18. Pamies, D.; Hartung, T.; Hogberg, H. T. Biological and medical applications of a brain-on-a-
chip. Experimental biology and medicine 2014, 239, 1096-1107.
19. Alepee, N.; Bahinski, A.; Daneshian, M.; De Wever, B.; Fritsche, E.; Goldberg, A.;
Hansmann, J.; Hartung, T.; Haycock, J.; Hogberg, H.; Hoelting, L.; Kelm, J. M.; Kadereit, S.;
McVey, E.; Landsiedel, R.; Leist, M.; Lubberstedt, M.; Noor, F.; Pellevoisin, C.; Petersohn,
D.; Pfannenbecker, U.; Reisinger, K.; Ramirez, T.; Rothen-Rutishauser, B.; Schafer-Korting,
M.; Zeilinger, K.; Zurich, M. G. State-of-the-art of 3D cultures (organs-on-a-chip) in safety
testing and pathophysiology. Altex 2014, 31, 441-77.
20. Hartung, T. 3D - a new dimension of in vitro research. Advanced drug delivery reviews 2014,
69-70, vi.
21. Marx, U.; Andersson, T. B.; Bahinski, A.; Beilmann, M.; Beken, S.; Cassee, F. R.; Cirit, M.;
Daneshian, M.; Fitzpatrick, S.; Frey, O.; Gaertner, C.; Giese, C.; Griffith, L.; Hartung, T.;
Heringa, M. B.; Hoeng, J.; de Jong, W. H.; Kojima, H.; Kuehnl, J.; Leist, M.; Luch, A.;
Maschmeyer, I.; Sakharov, D.; Sips, A. J.; Steger-Hartmann, T.; Tagle, D. A.; Tonevitsky,
A.; Tralau, T.; Tsyb, S.; van de Stolpe, A.; Vandebriel, R.; Vulto, P.; Wang, J.; Wiest, J.;
Rodenburg, M.; Roth, A. Biology-inspired microphysiological system approaches to solve
the prediction dilemma of substance testing. Altex 2016, 33, 272-321.
2. Jurga, M.; Lipkowski, A. W.; Lukomska, B.; Buzanska, L.; Kurzepa, K.; Sobanski, T.;
Habich, A.; Coecke, S.; Gajkowska, B.; Domanska-Janik, K. Generation of functional neural
artificial tissue from human umbilical cord blood stem cells. Tissue engineering. Part C,
Methods 2009, 15, 365-72.
3. Gibb, S. Toxicity testing in the 21st century: a vision and a strategy. Reproductive toxicology
2008, 25, 136-8.
4. Hartung, T. Food for thought ... on alternative methods for nanoparticle safety testing. Altex
2010, 27, 87-95.
278
5. Silbergeld, E. K.; Contreras, E. Q.; Hartung, T.; Hirsch, C.; Hogberg, H.; Jachak, A. C.;
Jordan, W.; Landsiedel, R.; Morris, J.; Patri, A.; Pounds, J. G.; de Vizcaya Ruiz, A.;
Shvedova, A.; Tanguay, R.; Tatarazako, N.; van Vliet, E.; Walker, N. J.; Wiesner, M.;
Wilcox, N.; Zurlo, J. t(4) workshop report. Nanotoxicology: "the end of the beginning" -
signs on the roadmap to a strategy for assuring the safe application and use of nanomaterials.
Altex 2011, 28, 236-41.
6. Pamies, D.; Barreras, P.; Block, K.; Makri, G.; Kumar, A.; Wiersma, D.; Smirnova, L.; Zang,
C.; Bressler, J.; Christian, K. M.; Harris, G.; Ming, G. L.; Berlinicke, C. J.; Kyro, K.; Song,
H.; Pardo, C. A.; Hartung, T.; Hogberg, H. T. A human brain microphysiological system
derived from induced pluripotent stem cells to study neurological diseases and toxicity. Altex
2017, 34, 362-376.
7. Scholz, D.; Poltl, D.; Genewsky, A.; Weng, M.; Waldmann, T.; Schildknecht, S.; Leist, M.
Rapid, complete and large-scale generation of post-mitotic neurons from the human
LUHMES cell line. Journal of neurochemistry 2011, 119, 957-71.
8. Smirnova, L.; Harris, G.; Delp, J.; Valadares, M.; Pamies, D.; Hogberg, H. T.; Waldmann, T.;
Leist, M.; Hartung, T. A LUHMES 3D dopaminergic neuronal model for neurotoxicity
testing allowing long-term exposure and cellular resilience analysis. Archives of toxicology
2016, 90, 2725-2743.
9. Hogberg, H. T.; Bressler, J.; Christian, K. M.; Harris, G.; Makri, G.; O'Driscoll, C.; Pamies,
D.; Smirnova, L.; Wen, Z.; Hartung, T. Toward a 3D model of human brain development for
studying gene/environment interactions. Stem cell research & therapy 2013, 4 Suppl 1, S4.
10. Galic, M. A.; Riazi, K.; Pittman, Q. J. Cytokines and brain excitability. Frontiers in
neuroendocrinology 2012, 33, 116-25.
11. Tsou, Y. H.; Zhang, X. Q.; Zhu, H.; Syed, S.; Xu, X. Drug Delivery to the Brain across the
Blood-Brain Barrier Using Nanomaterials. Small 2017, 13.
279
12. da Luz, C. M.; Boyles, M. S.; Falagan-Lotsch, P.; Pereira, M. R.; Tutumi, H. R.; de Oliveira
Santos, E.; Martins, N. B.; Himly, M.; Sommer, A.; Foissner, I.; Duschl, A.; Granjeiro, J. M.;
Leite, P. E. Poly-lactic acid nanoparticles (PLA-NP) promote physiological modifications in
lung epithelial cells and are internalized by clathrin-coated pits and lipid rafts. Journal of
nanobiotechnology 2017, 15, 11.
13. Hutter, E.; Boridy, S.; Labrecque, S.; Lalancette-Hebert, M.; Kriz, J.; Winnik, F. M.;
Maysinger, D. Microglial response to gold nanoparticles. ACS nano 2010, 4, 2595-606.
14. Soderstjerna, E.; Bauer, P.; Cedervall, T.; Abdshill, H.; Johansson, F.; Johansson, U. E. Silver
and gold nanoparticles exposure to in vitro cultured retina--studies on nanoparticle
internalization, apoptosis, oxidative stress, glial- and microglial activity. PloS one 2014, 9,
e105359.
15. Jung, S.; Bang, M.; Kim, B. S.; Lee, S.; Kotov, N. A.; Kim, B.; Jeon, D. Intracellular gold
nanoparticles increase neuronal excitability and aggravate seizure activity in the mouse brain.
PloS one 2014, 9, e91360.
16. Horie, R. T.; Sakamoto, T.; Nakagawa, T.; Ishihara, T.; Higaki, M.; Ito, J. Stealth-
nanoparticle strategy for enhancing the efficacy of steroids in mice with noise-induced
hearing loss. Nanomedicine (Lond) 2010, 5, 1331-40.
17. Liu, Z.; Gao, X.; Kang, T.; Jiang, M.; Miao, D.; Gu, G.; Hu, Q.; Song, Q.; Yao, L.; Tu, Y.;
Chen, H.; Jiang, X.; Chen, J. B6 peptide-modified PEG-PLA nanoparticles for enhanced
brain delivery of neuroprotective peptide. Bioconjugate chemistry 2013, 24, 997-1007.
18. Pamies, D.; Block, K.; Lau, P.; Gribaldo, L.; Pardo, C. A.; Barreras, P.; Smirnova, L.;
Wiersma, D.; Zhao, L.; Harris, G.; Hartung, T.; Hogberg, H. T. Rotenone exerts
developmental neurotoxicity in a human brain spheroid model. Toxicology and applied
pharmacology 2018.
280
19. Lundqvist, M.; Augustsson, C.; Lilja, M.; Lundkvist, K.; Dahlback, B.; Linse, S.; Cedervall,
T. The nanoparticle protein corona formed in human blood or human blood fractions. PloS
one 2017, 12, e0175871.
20. Dal Magro, R.; Albertini, B.; Beretta, S.; Rigolio, R.; Donzelli, E.; Chiorazzi, A.; Ricci, M.;
Blasi, P.; Sancini, G. Artificial apolipoprotein corona enables nanoparticle brain targeting.
Nanomedicine : nanotechnology, biology, and medicine 2017, 14, 429-438.
21. Jenkins, S. I.; Weinberg, D.; Al-Shakli, A. F.; Fernandes, A. R.; Yiu, H. H. P.; Telling, N. D.;
Roach, P.; Chari, D. M. 'Stealth' nanoparticles evade neural immune cells but also evade
major brain cell populations: Implications for PEG-based neurotherapeutics. Journal of
controlled release : official journal of the Controlled Release Society 2016, 224, 136-145.
22. Nance, E. A.; Woodworth, G. F.; Sailor, K. A.; Shih, T. Y.; Xu, Q.; Swaminathan, G.; Xiang,
D.; Eberhart, C.; Hanes, J. A dense poly(ethylene glycol) coating improves penetration of
large polymeric nanoparticles within brain tissue. Science translational medicine 2012, 4,
149ra119.
23. Haddad, D.; Nakamura, K. Understanding the susceptibility of dopamine neurons to
mitochondrial stressors in Parkinson's disease. FEBS letters 2015, 589, 3702-13.
24. De Simone, U.; Caloni, F.; Gribaldo, L.; Coccini, T. Human Co-culture Model of Neurons
and Astrocytes to Test Acute Cytotoxicity of Neurotoxic Compounds. International journal of
toxicology 2017, 36, 463-477.
25. Cui, H.; Kong, Y.; Zhang, H. Oxidative stress, mitochondrial dysfunction, and aging. Journal
of signal transduction 2012, 2012, 646354.
26. Haase, A.; Rott, S.; Mantion, A.; Graf, P.; Plendl, J.; Thunemann, A. F.; Meier, W. P.;
Taubert, A.; Luch, A.; Reiser, G. Effects of silver nanoparticles on primary mixed neural cell
cultures: uptake, oxidative stress and acute calcium responses. Toxicological sciences : an
official journal of the Society of Toxicology 2012, 126, 457-68.
281
27. Choi, S. S.; Lee, H. J.; Lim, I.; Satoh, J.; Kim, S. U. Human astrocytes: secretome profiles of
cytokines and chemokines. PloS one 2014, 9, e92325.
28. Sun, L.; He, C.; Nair, L.; Yeung, J.; Egwuagu, C. E. Interleukin 12 (IL-12) family cytokines:
Role in immune pathogenesis and treatment of CNS autoimmune disease. Cytokine 2015, 75,
249-55.
29. Rosenstein, J. M.; Krum, J. M.; Ruhrberg, C. VEGF in the nervous system. Organogenesis
2010, 6, 107-14.
30. Doyle, K. P.; Cekanaviciute, E.; Mamer, L. E.; Buckwalter, M. S. TGFbeta signaling in the
brain increases with aging and signals to astrocytes and innate immune cells in the weeks
after stroke. Journal of neuroinflammation 2010, 7, 62.
31. Akdis, M.; Aab, A.; Altunbulakli, C.; Azkur, K.; Costa, R. A.; Crameri, R.; Duan, S.;
Eiwegger, T.; Eljaszewicz, A.; Ferstl, R.; Frei, R.; Garbani, M.; Globinska, A.; Hess, L.;
Huitema, C.; Kubo, T.; Komlosi, Z.; Konieczna, P.; Kovacs, N.; Kucuksezer, U. C.; Meyer,
N.; Morita, H.; Olzhausen, J.; O'Mahony, L.; Pezer, M.; Prati, M.; Rebane, A.; Rhyner, C.;
Rinaldi, A.; Sokolowska, M.; Stanic, B.; Sugita, K.; Treis, A.; van de Veen, W.; Wanke, K.;
Wawrzyniak, M.; Wawrzyniak, P.; Wirz, O. F.; Zakzuk, J. S.; Akdis, C. A. Interleukins (from
IL-1 to IL-38), interferons, transforming growth factor beta, and TNF-alpha: Receptors,
functions, and roles in diseases. The Journal of allergy and clinical immunology 2016, 138,
984-1010.
32. Kimling, J.; Maier, M.; Okenve, B.; Kotaidis, V.; Ballot, H.; Plech, A. Turkevich method for
gold nanoparticle synthesis revisited. The journal of physical chemistry. B 2006, 110, 15700-
7.
33. Schildknecht, S.; Karreman, C.; Poltl, D.; Efremova, L.; Kullmann, C.; Gutbier, S.; Krug, A.;
Scholz, D.; Gerding, H. R.; Leist, M. Generation of genetically-modified human
differentiated cells for toxicological tests and the study of neurodegenerative diseases. Altex
2013, 30, 427-44.
282
34. Harris, G.; Hogberg, H.; Hartung, T.; Smirnova, L. 3D Differentiation of LUHMES Cell Line
to Study Recovery and Delayed Neurotoxic Effects. Current protocols in toxicology 2017, 73,
11 23 1-11 23 28.
35. Wen, Z.; Nguyen, H. N.; Guo, Z.; Lalli, M. A.; Wang, X.; Su, Y.; Kim, N. S.; Yoon, K. J.;
Shin, J.; Zhang, C.; Makri, G.; Nauen, D.; Yu, H.; Guzman, E.; Chiang, C. H.; Yoritomo, N.;
Kaibuchi, K.; Zou, J.; Christian, K. M.; Cheng, L.; Ross, C. A.; Margolis, R. L.; Chen, G.;
Kosik, K. S.; Song, H.; Ming, G. L. Synaptic dysregulation in a human iPS cell model of
mental disorders. Nature 2014, 515, 414-8.
36. Chomczynski, P.; Sacchi, N. Single-step method of RNA isolation by acid guanidinium
thiocyanate-phenol-chloroform extraction. Analytical biochemistry 1987, 162, 156-9.
37. Livak, K. J.; Schmittgen, T. D. Analysis of relative gene expression data using real-time
quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001, 25, 402-8.
283
CHAPTER 8
8. SUMMARIZING DISCUSSION
The work presented in this thesis encompasses (1) the development and characterization of a 3D
LUHMES dopaminergic cell model, (2) the application of this 3D model to study acute toxicity of the dopaminergic toxicants rotenone and MPP+, as well as recovery after wash-out, (3) the first in vitro study to look at resilience after second exposures to rotenone, (4) repeated dose-toxicity of rotenone and glyphosate to 3D LUHMES and (5) the assessment of drug-delivery nanoparticle toxicity using 3D brain in vitro models. The overarching concept for this work is to explore acute toxicity vs. recovery and resilience to challenge current acute toxicity approaches in vitro and study key events, which can lead to delayed or permanent adverse outcomes.
LUHMES was selected as the cell line of choice because of its human and non-cancerous origin, fast and homogeneous differentiation, dopaminergic markers (TH, DAT, and dopamine production), sensitivity to dopaminergic toxicants and applicability for neurotoxicity (NT) and developmental neurotoxicity (DNT) studies. It is also being used increasingly to study
Parkinson’s disease (PD) by altering gene expression or using a chemical exposure to identify novel drug targets. The generation of floating aggregate cultures developed by Honegger et al.
(1979) was modified in this project to differentiate LUHMES in 3D. Some baseline standards were kept in mind; (1) cell differentiation must not be compromised in 3D, (2) aggregate size must be controlled by gyratory shaking speed, cell concentration and media volume to avoid necrosis in the core of the aggregates, (3) proliferation within aggregates must be controlled to maintain a homogeneously differentiated cell population and (4) compounds must be able to penetrate throughout the aggregate.
284
Development of a 3D dopaminergic cell model
Although animal testing is required to test compounds for neurotoxicity, these are time- consuming, costly and have recently demonstrated to not be highly predictive of human health.
There is an increasing need and push for better predictive models (in vitro and in silico). Diseases of aging are also currently rising as the life expectancy of the westernized population increases and Parkinson’s disease is the second most common neurodegenerative disorder in people over the age of 65 with currently no preventive therapies. A 3D in vitro model to study dopaminergic toxicity and recovery is an important development in the field, which can help better understand mechanisms of dopaminergic neurodegeneration, survival and therapies. We first tested different protocols for differentiation in 2D and 3D and determined how long cultures could be kept under both conditions. It is common knowledge in the field that without a cell matrix-like coating (for example MatrigelTM, a gelatinous protein mixture produced from mouse sarcoma cells), but not physiologically present in the brain, differentiated neurons will detach from plastic-ware. This is due to the formation of long processes (neurites), which do not adhere easily, except where they find other cell clusters to form synapses (Chapter 1, Figure 1-1-1). This can present major challenges for neurotoxicity testing as ideally cells need to be fully differentiated and medium changes would not disrupt cells. Furthermore, toxicity to neurite integrity (which some compounds can cause at sub-toxic concentrations) would also cause detachment and this can be misidentified as effects on cell viability (number of cells). Imaging of dense neural networks has proven to be challenging in 2D and is overcome by either decreasing cell density or using mixed cell populations with a small portion expressing a fluorescent signal (Schildknecht et al., 2013).
Based on the results of this thesis, we conclude that a 3D model, in suspension, overcomes the pitfalls of 2D culture described above, as well as provides cells a less static environment, increased cell-cell interactions and cell culture shelf-life.
285
Differentiation of LUHMES in 3D was achieved by placing cells in differentiation media on a shaker, whereby aggregation occurred spontaneously due to their nature to form clusters (Chapter
1, Figure 1-2). Our research showed that within aggregates, due to close cell-cell interaction between cells, proliferation continued to take place unless the anti-proliferation drug Paclitaxel
(Taxol) was added (10 nM, 48 h). This modification to the protocol for 3D culture was influenced by other differentiation protocols, which use AraC to block the proliferation of neuroprogenitors and astroglia without affecting post-mitotic neurons ((Gerhardt et al. 2001) Volbracht et al. 2006).
This was a critical step to be added to the protocol. Presence of proliferating cells would make this model not any longer homogeneous in terms of cell composition and, more importantly, mask toxic effects of test compounds on mature neurons. As described above, a homogeneous cell population is needed for identify molecular signatures of neurotoxicity or to recapitulate PD pathology – degeneration of post-mitotic dopaminergic neurons. Although anti-proliferation agents are commonly used for differentiation, this could be a disadvantage of the model, as Taxol may be affecting differentiated cells and altering their response to subsequent toxicity tests. Taxol is used as a cancer treatment drug, which can be neurotoxic, leading to peripheral neuropathy due to its effects on microtubule assembly. Dose and duration of treatment is important, only patients with higher doses of Taxol have developed peripheral neuropathy (Scripture, Figg, and
Sparreboom 2006). Therefore, we titrated the concentration to lowest possible as well as limited the Taxol treatment time to 48 hours. Although the controls were differentiated the same way as compound-treated samples, Taxol pre-treatment must be taken into account when formulating a research question. Our model characterization also showed the only difference between cultures differentiated with or without Taxol was the percentage of proliferating cells and, hence, increased differentiation marker expression.
Once Taxol is removed, no further proliferation is observed and complete differentiation takes place within 6 days in culture. Our model reproduces the differentiation expression pattern
286 reported for 2D cultures (Chapter 4, Figures 4-1, 4-4 and 4-5). We hypothesized that 3D cell culture would increase differentiation, however, these cells are post-mitotic and differentiate rapidly upon myc silencing by tetracycline, reaching maximum expression levels for differentiation markers within 3 days in vitro (which remain at the same level throughout differentiation). We further characterized the model showing low levels of necrosis or apoptosis throughout aggregates during differentiation (Chapter 4, Figure 4-2), low Ki-67 (proliferation marker) mRNA and protein expression after 6 days of differentiation (Chapter 4, Figure 4-3) and penetration of the compound Hoechst 33342 (Mw = 452.5 g/mol) through aggregates within 1 hour (Chapter 4, Figure 4-2). New in vitro models must be thoroughly characterized; determining the level of apoptosis within aggregates allows us to prove that this model is suitable for toxicity testing as cells are viable throughout 21 days in culture, low Ki-67 expression shows that cells are homogeneously differentiated and the response to toxicants would not arise from a mixed population and, finally, compound penetration shows that when testing compounds of similar size, these are able to reach cells within the center of the aggregate. Although it is likely that cells on the outer part of aggregates are exposed to higher concentrations of compounds initially than the center, this better mimics human exposure, whereby affected cells would signal nearby cells to respond to a toxicant hit.
In order to study toxicity in 3D models, certain challenges must be overcome: most assays and endpoints have been developed for monolayer cultures, therefore must be adapted to measure outcomes from 3D culture. Some examples include the need to dissociate aggregates for endpoints, which require single cells suspensions, slow penetration of large molecules or antibodies, required confocal imaging for single-cell-imaging within aggregates, the need for clearing solutions to improve imaging depth throughout aggregates and 3D culture is not adapted
96-well plate formats. In Chapter 2, the protocol describes an assay for neurite outgrowth, which can be scaled up to a 96-well format by seeding differentiated aggregates on Matrigel® to induce
287 outgrowth and use high-content imaging to quantify neurite number and length (Chapter 3, Figure
3-5).
Within aggregates, LUHMES are able to grow long neurites, which wrap around the sphere. After characterization, we tested the toxicity of two well-known dopaminergic toxicants and PD inducers, rotenone and MPP+, to compare the cellular response in 3D vs. 2D. In our model, we recorded the same dose-response curves (24 h exposure, day 8) as those reported in 2D (Krug et al. 2013). This is an indication that the compounds are able to affect all the cells within aggregates equally and 3D culture does not increase survival to toxicity by these compounds or mask the effect.
Animal models (chemically induced or genetically modified) for PD are not fully representative of human pathophysiology (Beal 2010). Moreover, studying long-term effects of low-dose environmental exposures in animals is costly and difficult to perform for behavioral assessments.
Screening the high number of potential neurotoxicants within the chemical universe (100,000 chemicals) in animals also does not seem feasible. On the other hand, the development of neuroprotective therapies is slow, and is also driven by mechanistic research. There is a need for fast, reproducible in vitro models to study molecular pathways and screen drugs for efficacy or compounds for toxicity.
Rotenone is highly lipophilic and therefore binds to plastic. The only way to study cell recovery is to remove it completely from the cultures by replating the cells into the new dish. In 2D culture, cells would have to be trypsinized and re-plated (which is almost impossible for neuronal cultures, since the long processes will be disrupted by the procedure and neurons would need to form new processes once replated). 3D culture provides the enormous advantage to investigate toxicity after compound wash-out, because of their easy handling to move into new cell culture
288 plates and the ability to maintain in culture for longer (at least 21 days vs. 12 days in 2D)
(Chapter 4, Figure 4-1). Despite many 3D models being available today, recovery, adaptation and delayed response to the toxicant insult has not been studied largely in vitro.
Microphysiological systems which include microfluidics, somewhat mimic this process as media is continuously flowing and exposures can be added or removed. However, in these systems, binding to plastic must also be assessed and reported to better understand exposures over time.
Studying recovery and resilience using the 3D LUHMES model
We developed a protocol for compound wash-out which is outlined extensively in Chapter 3 to test our resilience hypothesis (Chapter 2). As a proof of concept, we applied out 3D model to study toxicity of rotenone and MPP+ after wash-out. Rotenone-treated cultures were able to maintain viability after wash-out, while MPP+-treated cultures declined drastically (Chapter 3,
Figure 3-2). We must note that the non-cytotoxic concentration of MPP+ used (5 µM), was higher than that of rotenone (100 nM) (Chapter 4, Figure 4-6). Although rotenone and MPP+ have presumably the same molecular initiating event (MIE), their long-term toxicities differed. In fact, different effects from these compounds on cellular bioenergetics have been reported (Giordano et al, 2012 (Giordano et al. 2012). It could be that at µM concentrations, MPP+ wash-out was less successful, although rotenone’s lipophilicity would in theory also make this compound stick to cells. From this data, we could observe that cells exposed to 100 nM rotenone were fighting to survive, while 5 µM MPP+ triggered cell death pathways, which cells were not able to cope with
(Krug et al, 2013). This is a novel finding as it is rarely studied in toxicology. From dose- response curves, which are currently performed in most high-throughput screening assays (HTS) in toxicology, these two compounds would have very distinct IC50s (concentration, which leads to
50 % loss in viability), that of MPP+ being at least one order of magnitude higher (chapter 4,
Figure 4-6). To be able to support this conclusion, intracellular MPP+ concentrations would need to be measured to determine whether it is effectively washed out, as well as other mechanistic
289 endpoints such as complex I inhibition, ATP production and apoptotic markers to understand why cells may not be able to cope. Other possible explanations include differences in agonistic potency on the pharmacological receptor (complex I), intracellular kinetics to reach complex I or off-target effects. An advance in 21st century toxicology is the development of AOPs (adverse outcome pathways) (Leist et al. 2017). These intend to provide a framework that encompasses available information to link MIEs with sequential mechanisms (KE) which lead to an adverse outcome (AO). An ‘AOP for Parkinsonian motor deficits associated with mitochondrial complex
I inhibition’ has been developed recently (Terron et al., 2017). These AOPs are compound agnostic, and do not factor in toxicokinetics, however, if the MIE is known, or a compound leads to any given KE, the AOP is plausible mechanistically. Experimental (in vivo and in vitro) studies support the causative link between one KE and another. This is termed the key event relationship
(KER). To improve AOPs, in vitro models and assays, which can test for one or more KE, are needed. The 3D LUHMES model is suitable to study multiple key events in vitro.
Our aim to study cellular recovery and resilience led us to select rotenone at 100 nM for 24 h.
One of the strengths of this work is the quantification of free rotenone in media prior to, during and after wash-out. The human exposome and the kinetics of substances in the organism are largely understudied but so are relative in vitro free concentrations, which need to be extrapolated to potential human in vivo exposures at a target organ (bioavailability). Most research uses the nominal concentration to study in vitro concentration-effect relationships, and does not consider free concentration or the concentration, which is available for binding. PBPK modelling should be applied for in vitro studies as multiple factors (non-specific binding to plastics, media components or cells, degradation, evaporation or metabolism) alter the bioavailability of the compound (Groothuis et al. 2015; Seibert, Morchel, and Gulden 2002). We demonstrate that in our model, rotenone can be washed-out effectively and no rotenone remains in the media on day
15, when we assess recovery and resilience to a second exposure. This measurement also allowed
290 us to quantify how much rotenone remains bound to plastic (~ 30 %) or cells (~ 30 %) after wash-out (Chapter 5, Figure 5-1d-e). To improve this analysis, we would further perform a time- course free concentration quantification to determine the rate of binding to plastic and cells.
Furthermore, measuring cellular rotenone (within aggregates) would be a better direct measurement of bound rotenone). Thus, our data demonstrated the importance of estimation of free concentration for better description of the observed effects and extrapolation to human exposures as the observed acute effects occur at lower than nominal concentrations.
Firstly we ensured that our model was able to reproduce the rotenone effects reported in vivo and in vitro (Shapire et al., 1989; Sherer et al., 2007; Bertarbet et al., 2011). In fact, the model reproduced findings very closely (Chapter 4 Figure 4-6, Chapter 5, Figures 5-2a, 5-3a-b, 5-5a-b).
Additionally, for the first time in LUHMES, we demonstrated rotenone-induced downregulation of miR-7 and recovery of its expression after compound wash-out. MiR-7 is known to target alpha-synuclein (α-Syn) and protect SH-5YSY cells against MPP+ toxicity when overexpressed
(Choi et al, 2014) (Chapter 4, Figure 4-8b). Future research would involve studying the role of this and other microRNAs in dopaminergic toxicity and recovery via miRNA silencing or target gene overexpression.
We also were able to image mitochondria by electron microscopy. Methods for mitochondrial imaging and tracking have improved and mitochondrial biogenesis, mitophagy and motility play a role in neurodegeneration. We observed an increase in diameter but not in number of mitochondria after acute rotenone exposure (Chapter 5, Figure 5-4). Although not confirmed here, an increase in mitochondrial diameter can indicate mitochondrial swelling, which can occur prior to bursting due to oxidative stress and opening of the mitochondrial permeability transition pore (mPTP) (Peng and Jou 2004). This was not studied here, and is currently only a hypothesis, and should be included in future studies. Interestingly, although the majority of these mechanisms
291 are all well studied for rotenone-induced PD pathology, research gaps remain as to how these mechanisms contribute to long-term neurodegeneration. One of the main problems is that most studies focus on the loss of dopaminergic neurons rather than their recovery. A model to study recovery or long-term toxicity could help answer questions such as whether there is a threshold past which cells can no longer recover from a toxic insult, or do cells retain ‘molecular memory’ from past exposures leading to a different ‘cellular homeostasis’ or response to subsequent exposures (Smirnova et al., 2015). Moreover, as AOPs, need evidence to demonstrate causality between endpoints, acute toxicity endpoints may not be the most adequate to provide sufficient causal evidence, because adaptation and recovery mechanisms are not assessed (Leist et al., 2017;
Bal-Price et al., 2017).
To better understand whether the observed effects would worsen over time or cells were able to recover, we measured molecular and functional neurotoxicity endpoints, which are relevant to PD mechanisms, after wash-out. In our earlier work, we had observed that after wash-out, the gene expression pattern of LUHMES was distinct to that prior to wash-out as well as to control non- treated cultures. This indicated that surviving cells have a different transcriptomic profile to untreated controls; and some early response genes are no longer involved after compound removal (Chapter 4, Figure 4-8). This had not been reported previously as time-points are usually studied in isolation and especially in toxicology, if no effect is observed, (negative) results are rarely reported (Matosin et al. 2014). Hypothetically, a gene, which was unchanged acutely but was altered after wash-out, could potentially play a more important role in disease progression or prevention. This yet has to be proven.
By assessing viability, DNA and protein levels, we concluded that 25-30% of cell loss occurs over time, after wash-out. Our interest was not in the cells, which are on the ‘road to death’, but
292 rather those, which are fighting to survive, similar to (evolutionary) biology’s “survival of the fittest”.
As complex I is the primary molecular target for rotenone (although off target binding can also occur) (Panov et al, 2005), we observed permanent inhibition of complex I activity after rotenone wash-out (Chapter 5, Figure 5-3a). With our free rotenone quantification, these results can be explained through availability kinetics. As rotenone is lipophilic, it can bind transiently to proteins and slowly become available for binding to its target. We were able to quantify that around 30 % of total rotenone remains bound or trapped within aggregates (Chapter 4, Figure x).
Although it is generally thought that complex I inhibition leads to a decrease in ATP production, a study using isolated mitochondria from rat brain showed that > 70 % complex I inhibition can be reached before decreases in ATP levels are recorded (Davey and Clark 1996). Although ATP levels decreased acutely, after wash-out, ATP levels were above controls, indicating the compensatory mechanisms in place lead to increased ATP production despite inhibited complex I
(Chapter 5, Figure 5-3b). Indeed, ATP would be required for antioxidant defense, de novo protein synthesis and DNA repair mechanisms. Future research would delve into the metabolic characterization of the cells to determine whether generated ATP is derived from glycolysis or oxidative phosphorylation. This can be achieved using isotopic flux studies of metabolism by mass spectroscopy or using metabolic assays. We generated some preliminary data using Agilent
Seahorse Technologies, but the assay needs further development before we can draw conclusions.
Although a promising, relatively simple and fast assay, Seahorse needs to be adapted for high- throughput metabolic measurements in 3D. The main limitation of Seahorse for our model was the small size of LUHMES aggregates (<300 m). The Seahorse plates are designed for much larger aggregates. Using Seahorse, Delp et al. have shown that LUHMES cells have high glycolytic capacity in 2D and further experiments as to how 3D LUHMES respond to first and second exposures to rotenone using another energy source would be a next step (Delp et al.,
293
2017). Others have described how a glycolytic cell state leads to less deleterious effects of
mitochondrial complex-I inhibitors (Lassus et al. 2016).
A flaw at this point in our assessment of recovery is the lack of oxidative stress quantification.
This is one of the most commonly reported effects of rotenone, not only in neurons. However, we
were unable to detect any levels of reactive oxygen species (ROS) in cells using fluorescent
staining techniques (penetration of fluorophores into aggregates could be an issue) or cell lysates
(small sample size could mean ROS levels were below the detection limit for the assay developed
for human samples). Other measurements such as S-nitrosylation (Nakamura et al. 2013) or
glutathione levels (Nakamura et al. 2000) could serve as surrogates for future studies to quantify
and trace oxidative stress.
An increase in ATP after compound removal indicated recovery at the energy level, this led us to
ask the question, where is ATP coming from? Our DNA and protein measurements could confirm
that cells were not dividing and the observed increase was not due to an increase in cell mass. We
further confirmed this by qPCR of the proliferation marker Ki-67, which did not increase after
wash-out. Thus, even though the cell number had decreased, those remaining cells were more
metabolic active synthesizing more ATP than controls. The next possible explanation was that
changes were taking place at the mitochondrial level. Mitochondria, which had shown an acute
increase in diameter, were similar to controls after wash-out. Therefore, the increase in diameter
was reversible, or mitochondrial biogenesis over time was able to generate healthy mitochondria
(Chapter 5 Figure 5-4b). Mitochondria are actively changing and moving throughout cells,
biogenesis is an important response to energy demand and oxidative stress regulation (Touyz,
2012)9. The next step being performed is mtDNA (mitochondrial DNA) qPCR as a relative
9 http://www.academia.edu/9801501/Primer_on_the_Autonomic_Nervous_System
294 quantification of mitochondria, which could provide an insight into whether mitochondrial biogenesis takes place during recovery.
In addition to the assays described above, mass-spectrometry-based metabolomics offers an opportunity to study metabolic capacities of the cells at the given time point. Targeted metabolomics of energy metabolism pathways, transsulfuration and oxidative stress as well as flux analysis can provide more insights into the metabolic adaptation after recovery period.
As recovery was observed at the molecular level, we tested functional endpoints. Aggregates were seeded on Matrigel®, which induces neurite outgrowth from the aggregates. Although generally considered a DNT endpoint, this allowed us to see whether these differentiated neurons could form new processes if given the medium to do so. This is an endpoint, which could be performed in high-throughput manner and our results demonstrate the importance of wash-out studies. After 24 h exposure, aggregates were unable to produce as many neurites as controls, and the neurites, which did form, were overall shorter in length. The observed effect was completely reversed on day 15 after wash-out and treated cells were able to form neurites, which were comparable in number and length to controls (Chapter 5, Figure 5-5a-b). In a neurotoxicity screen based on this as an acute endpoint, one could infer that rotenone inhibits neurite outgrowth. We demonstrate that this effect is transient and not observed long-term in this model. We further assessed functionality through electrical activity, and no differences between control and treated were recorded for % tonic vs. phasic cells, input resistance or minimal spike latency (Chapter 5.
Figure 6). This model can also be applied for assessment using high-throughput microelectrode arrays (MEAs) recordings, and preliminary results from our group (data not shown) have demonstrated LUHMES can produce spontaneous electrical activity that can be measured in this way. A time-course experiment, measuring spontaneous activity after exposure, wash-out and subsequent exposures would provide a better picture of how toxicity affects neuronal function.
295
We conclude that cells are able to react to a toxicant, but then return to normal function, and therefore acute endpoints may not be as suitable to study long-term neurotoxicity.
With the aim to identify the gene expression patterns for acute toxicity vs. recovery, we performed whole genome microarray analysis (Chapter 5. Figure 5-7). The expression profiles corroborated our molecular and functional data, indicating that 708 genes were regulated after exposure, and none were regulated after wash-out. We further looked at our wash-out samples on day 15, and analyzed data without FDR correction, to increase the discovery of genes, which were possibly altered but not significantly due to multiple testing correction. The low sensitivity of microarrays (due to FDR correction for over 20,000 genes and small samples size) could be a disadvantage in studying the effects of low dose exposure or recover periods where changes are subtle. With this further analysis, we identified 107 genes, which were perturbed on day 15. Only
10 of these genes overlapped with day 8. Pathway enrichment analysis of genes on day 8 and day
15 showed enrichment for neurogenesis and Alzheimer’s disease on day 8. On day 15, genes were enriched for neurogenesis, plasma membrane components and CNS development. This is interesting as the gene enrichment on day 15 indicates recovery (Appendix I, Chapter 5, supplementary material). Most of the 10 genes, which were altered both on day 8 and day 15, are enriched in the brain, indicating they may play crucial roles in function differentiation and homeostasis. Although functional roles and relationships to inflammation, PD, oxidative stress, and rotenone were supported in the literature, it is precarious to over-interpret these as our analysis was not confirmed yet by qPCR or subsequent experiments. This should be performed in future work, if possible compared to another compound with a different mode of action, to define whether regulation of these genes is specific to rotenone toxicity and recovery.
Furthermore, comparison between our model and a more complex co-culture model could identify, which pathways evoke signaling to other cell types that could contribute to toxicity. In
296 heart tissue, Murphy et al., recently described that endothelial cells could send nitric oxide signals to cardiomyocytes in response to estrogen and protect them from ischemic reperfusion injury
(Menazza et al. 2017; Murphy et al. 2014). Studies have also shown that in vitro, astrocytes protect against dopaminergic toxicity (Efremova et al. 2015; Du et al. 2018; De Simone et al.
2017). For this reason, comparisons between single-cell models and co-cultures can provide an insight into which pathways could be targeted in different cell types to protect against toxicity, and how different cell types support each other. This is the can be addressed in the future by (i) introducing RFP or GFP-expressing LUHMES to iPSC-derived BrainSpheres or (ii) switching to iPSC and using bioinformatic cell separation methods that allow molecular analysis of different cell types (Goudriaan et al. 2014).
Currently, the ideal model to study neurotoxicity of a compound and its relevance to long-term neurodegenerative diseases would be one that can represent the three main risk factors; (1) a genetic risk variant, (2) environmental exposure to a compound and (3) an aging phenotype. Liu et al. published a model that encompasses these elements, and provides an example of how complex chronic diseases are and how we should best approach the environmental impact on their development (Liu et al. 2017). They showed how primary cortical neuronal cultures from
LRRK2R1441G knock-in mice had decreased ATP levels and dopamine uptake upon rotenone treatment compared to WT. Furthermore, age was then introduced as a third factor in the in vivo experiments and showed that when treated, LRRK2R1441G aged mice had greater locomotor deficits than aged controls. The field of in vitro neurodegenerative disease research will need better characterization of aging to improve current models, similar to what has been done for early development.
Although our model only represents a response to environmental exposure from healthy cells.
The contribution of a neurotoxicant towards neurodegeneration can be answered partly by how cells recover from toxicity caused by an environmental exposure. If not fully functional (due to a
297 mutation or aging), cells would not be able to recover from toxicity, leading to a long-term adverse outcome. As we wrote in our food for thought article, ‘It is not important whether you fall, but whether you get up again’ and this may well be one of the factors which leads to heterogeneity in the development of sporadic PD (Smirnova et al, 2015).
To challenge ‘recovered cells’ and determine whether dopaminergic cells were more susceptible or resilient to subsequent exposures to rotenone, we treated control cells and pre-exposed cells with rotenone at different doses. Our dose response curves showed decreased sensitivity of pre- exposed cells to rotenone. This was observed for both aggregates pre-exposed to 100 nM and 50 nM rotenone. The IC20 for aggregates, which were pre-exposed to 100 nM rotenone, was >1 µM, compared to controls’ IC20 of ~100 nM. Thus, protection was observed at concentrations which upon first exposure were cytotoxic (Chapter 5 Figure 5-8a-c). A similar protective effect was observed with the LDH assay. From a philosophical perspective, one could hypothesize that the brain, which is capable of creating memories through molecular signals, may be also able to
‘remember’ past exposures in the same way as it remembers past events. To test effects on gene expression, we measured genes, which had previously shown to be altered by rotenone (Krug et al., 2013; Smirnova et al., 2016; Shih et al., 2015; Zhang et al., 2016). (Chapter 5, Figure 5-8d-f).
Although these changes differ upon second exposures and resilience, it is unknown as to whether the pathways involved are detrimental in the long-term or continue to drive resilience (Delp et al.,
2017). Waldmann et al. (2017) performed an elegant study in which they looked at microarray data for samples treated with a range of non-cytotoxic to cytotoxic concentrations of VPA
(valproic acid) and MeHg (methylmercury). They were able to identify the transition from adaptive responses to cytotoxicity, and found potential changes in expression, which could serve as biomarkers of toxicity (BoT) (Waldmann et al. 2017). To quote Paracelsus, “the dose makes the poison’, as we can adapt to low-dose effects. However, the time of exposure and genetic differences could also make a relatively inert substance poisonous. This brings us back to the
298 downregulation of mir-7, which we reported after acute exposure to rotenone (Chapter 4, Figure
4-8); this epigenetic change can have major implications in PD development. Mir-7 targets α-Syn, which is involved in PD development and pathology forming Lewy bodies, our experiments also demonstrated acute α-Syn up-regulation. Down-regulation of mir-7 would lead to increased levels of SNCA mRNA available for translation, which could lead to increased protein levels. Although transient miR-7 downregulation (observed after 24 h rotenone exposure, but not wash-out), did not alter SNCA protein levels after wash-out, longer exposures, leading to sustained downregulation could have marked effects (Kaidery, Tarannum, and Thomas 2013). Next, it would be interesting to measure microRNA expression after second exposures.
Repeated-dose toxicity of rotenone and glyphosate
Diseases for which an environmental component is likely are thought to develop years after exposure or later in life. For example, low-dose, repeated exposures to pesticides of individuals are reported to increase risk of PD development (Liou et al. 1997; Betarbet et al. 2000; Dhillon et al. 2008; Hertzman et al. 1990). Taking advantage of the ability to culture these cells for longer time-periods in suspension, we further studied repeated-dose effects of rotenone and the currently most-widely used herbicide, glyphosate. After repeated-doses, rotenone and glyphosate decrease viability by ~ 25 %. If cells could recover from 100 nM, we hypothesized, that challenging them multiple times at a lower-dose, may trigger protective pathways to better study resilience.
Astonishingly, ATP levels doubled in cells treated repeatedly with 30 nM rotenone, similar to the effect observed after wash-out of 100 nM rotenone (Chapter 6, Figure 6-1; Chapter 5, Figure 5-
3b). We observed less of an effect on neurite outgrowth than with acute 100 nM exposure, which leads to questions as to where the threshold for irreparable toxicity lies and what level of stress cells can cope with (Chapter 6, Figure 6-2). Few repeated low-dose studies have been performed in vitro, but the importance of time-dependent effects may be undervalued. Two European projects, DETECTIVE (http://www.seurat-1.eu/) and EU-ToxRisk (http://www.eu-toxrisk.eu)
299 include projects aimed towards the replacement of repeated-dose testing. In these studies, quantification of the compound, which is added to cell culture media, is important to be able to study dose-response kinetics and in vitro to in vivo extrapolations. We have already mentioned the challenges met with 2D cultures, but our 3D model could prove to be well-suited. Currently we are quantifying the free rotenone concentration in media to determine whether these effects are additive due to cumulative toxicity. The accumulation of low-dose effects and exposures to mixtures (multiple compounds at low-doses), is more relevant to human exposures and organ responses at the cellular level. In our experiments, we identified that genes involved in PD mechanisms or that are known risk factors for PD, were upregulated after repeated-dose exposures (Chapter 6 Figure 6-3). These changes were not significant for acute treatment (100 nM) or wash-out, therefore may be more specific to cellular protection against low-dose, long- term exposures. Further confirmation of up-regulation at the protein level is required to better investigate the role of these proteins in survival. As this is a ‘healthy’ cell line, the next step would be to use CRISPR-Cas9 to insert a disease-causing mutation into these cells, to understand whether this leads to an adverse outcome such as protein aggregation (used as a marker for PD), or renders cells more susceptible to repeated exposures. Although cell lines (PC12, SH-5YSY) are used to study chemical-induced PD mechanisms, to date, recovery, resilience and repeated low-dose effects have not been studied in a chemically induced human-3D model.
Despite use as animal models of PD, rotenone is unlikely to contribute significantly to the occurrence of PD in the general population. Rotenone has limited uses, poor oral bioavailability, and a short half-life in the environment. It is however, a good model compound to develop assays and identify mechanisms of dopaminergic degeneration, which can serve as endpoints to study other chemicals such as current pesticides (glyphosate, paraquat, maneb) and environmental toxicants thought to play a role in neurodegenerative diseases.
300
Assessing nanoparticle toxicity using 3D human brain spheroid models
Novel therapies to target and treat PD involve improved drug delivery systems such as the use of
nanoparticle (NP) carriers. To treat diseases of the brain, these nanoparticles (NPs) need to be
able to cross the blood brain barrier (BBB). Gold and PLA (poly-lactic acid) NP have the
potential to cross the BBB (Saraiva et al. 2016). Moreover, PEGylation increases circulation time,
which leads to NP accumulating more efficiently in the brain. Nanoparticle surface properties can
be altered biochemically. Currently it is estimated that >700 NP types exist on the market, with
many more, which have likely not yet been characterized. Within the context of PD, NP could be
used to deliver drug/gene therapies, siRNAs to target SNCA mRNA for degradation or NP that
target miRNAs and stabilize or increase their expression as an epigenetic mechanism. The
possibility to deliver these approaches is increasingly promising for future therapies and NP are a
key enabling technology in many other areas. NP, however, are not only offering treatment
options, but might also represent a health risk, in our case of neurotoxicity as they cross the BBB.
The health effects of NP exposure are emerging after years of exposure and use is only increasing
with very limited regulation10 (Hansen and Baun 2012; Busquet and Hartung 2017). Some of the
best-known examples are the effects of carbon nanotubes in the lung where deposition can occur
and similar mechanisms to asbestos toxicity are observed (Donaldson et al. 2013).
In this study, we compared 3D LUHMES to a multi-cellular iPSC-derived 3D brain model. Our
iPSC-derived BrainSphere model was developed and characterized previously (Pamies et al,
2017) (Chapter 7, Figure 7-2b-d). The model shows advantages to other currently-developed
iPSC-derived models as aggregates are homogeneous in size, containing different neuronal and
glial cells as well as spontaneous electrical activity and myelination (which has been difficult to
achieve it vitro) within 8 weeks of differentiation. The disadvantage of our models for NP drug-
10 https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1541-4337.2009.00088.x
301 delivery research is the absence of the blood-brain-barrier and immune function (microglia). In our study, three nanoparticle types were chosen, gold NP (functionalized with either sodium citrate (Au-SC) or polyethylene glycol (PEG-SC) and polymeric polylactic acid (PLA-NP)).
Nanoparticles have a large overall surface area due to their small size and can exhibit unique surface properties, which can change when they come into contact with different media proteins and compounds and this can lead to aggregation. Characterization of nanoparticle physicochemical properties (shape, size, aggregation, surface area, charge, pH, electric) in the test medium is becoming increasingly needed to make use of in vitro toxicity data (Sayes and Warheit
2009). Furthermore, these descriptors could help in silico approaches to model toxicity (Burello and Worth 2011). Currently, there is a need for better collaboration between bio- and chemical engineers, who develop NP, and toxicologists to improve toxicity assessment. Different NP characterization techniques are commonly used; these include scanning electron microscopy
(SEM), transmission electron microscopy (TEM), dynamic light scattering (DLS), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), infrared (IR), high-resolution technique of differential centrifugal sedimentation (DCS), polarized optical microscopy (POM) and Brunauer–Emmett–Teller (BET) (Khan, Saeed, and Khan 2017). In the presented work, we characterized nanoparticles using SEM and DLS to determine whether they changed in size in cell culture media (Chapter 7, Figure 7-1). PLA-NP almost doubled in size in both LUHMES and
BrainSphere differentiation media. This could be explained by the formation of a protein corona as these medias contain protein and differentiation supplements, however, better characterization is needed to demonstrate how aggregation is taking place.
We further characterized nanoparticle uptake into the brain 3D models using confocal microscopy and flow cytometry for PLA-NP and inductively coupled plasma mass spectrometry (ICP-MS) for gold NPs. As the PLA-NPs tested were tagged with a green fluorescent probe, we were able to image NPs within aggregates and co-localized within cells (Chapter 7, Figure 7-2). Using RFP-
302 expressing LUHMES cells, we were further able to confirm cellular localization by flow cytometry. Further high resolution imaging, e.g. electron microscopy, would allow better localization of these NPs to see whether they accumulate within cell mitochondria, nuclei, lysosomes or simply bind to the cellular membrane. ICP-MS results show that gold nanoparticles were able to accumulate within cells after 72 hours.
To assess toxicity of these NP, we performed the LDH assay in parallel with mitochondrial membrane potential (MMP) measurements and imaging. Our cytotoxicity data and imaging show that PLA-NPs and Au-PEG were cytotoxic to 3D LUHMES but inert to BrainSpheres (Chapter 7,
Figures 7-4 and 7-5). The MMP provides more of a mechanistic endpoint as effects can be observed prior to cell death. We observed that Au-SC decreased the MMP in both models, while
Au-PEG altered the MMP in BrainSpheres and PLA-NPs decreased the MMP in 3D LUHMES
(Chapter 7, Figure 7-4). These results indicate that although cytotoxicity is not observed in
BrainSpheres, MMP in different cell types may be affected differently, and this can be captured with MMP assay but not with general cytotoxicity assay. Moreover, Au-PEG increased the expression of specific antioxidant response genes (Chapter 7, Figure7-6). Although NPs altered the release of chemokines and growth factors in our 3D models, levels were low, and no firm conclusions can be made due to the absence of positive or negative controls (Chapter 7, Figure 7-
7).
These nanoparticles have been well characterized, therefore our research deferentially demonstrates that our 3D brain models can be applied to study neuronal toxicity of NPs in vitro.
Especially, to study how dopaminergic cells respond to NPs without support cells, and NP effects in a multi-cellular model. Further research is needed to define the mechanism of toxicity and which cell types are being affected. Intracellular NP localization assays would have provided a better understanding of how these NPs cause toxicity. The 3D LUHMES model consists of a
303 single cell-type, without support cells such as astrocytes, therefore this could explain why NP show greater cytotoxicity. Thus, using an approach with a single-cell model and a multi-cellular model provides an insight as to how different cells play a role in the response to NP toxicity. , our models proved to be applicable for this area of research.
Next steps
Overall, the work presented in this thesis describes a 3D model which can be used to assess dopaminergic toxicity in vitro. For the first time we have shown that a 3D in vitro model can be used as a tool to study recovery, resilience and repeated dose-effects as well as NP toxicity. The aspect of resilience is probably the most innovative aspect. The future directions for this research, therefore, should primarily be to better characterize the resilience. Is this an event, which determines cell fate in the long-term, or have cells fully recovered? Mechanistic evaluation over time before and after second exposures using high-content omics technologies could better characterize the cellular response. Then, loss and gain of function experiments can validate the pathways involved in resilience and recovery. In terms of assays, metabolic assays or targeted metabolomics would be used to determine the new cellular homeostasis after recovery. Is resilience maintained? How do cells respond if allowed to recover again? Using other cell types to study this could answer whether this is cell-specific or a similar response is observed in other neuronal cell types, as well as, studying other compounds to determine whether it is compound specific. Looking at epigenetic mechanisms in resilience would be a future question, a-syn is known to interact with DNA Methyltransferase 1 (DNMT1, involved DNA methylation maintenance) and therefore methylation patterns could be studied in this model after acute exposure, recovery and second exposures. Developing a co-culture using LUHMES cells would help determine the role of other support cells such as astrocytes and provide a comparison to the
304 single cell-type response. As a step further, investigating recovery in multi-cellular models using cell-sorting techniques would also closer mimic in vivo neurodegeneration. The big question, however, is whether in vitro toxicology based on recovery and resilience becomes more predictive of human health effects than currently used acute toxicity measurements?
This model is suitable for high throughput screening (HTS) approaches, therefore, for toxicity testing, the next steps would be to further develop assays such as neurite outgrowth, calcium signaling, ROS generation, dopamine production or SNCA accumulation using fluorescence- based imaging. The epitome would be to develop a combination of reporter approaches with the expression of disease risk genes in different LUHMES cell lines using CRISRCas-9 technologies to identify how genetic variants alter the response to compounds as well as therapeutic targets.
Altogether, 3D brain models have presented themselves as enabling technologies to bring neurotoxicology and research into neurodegenerative diseases to a different level. They are opening up to a mechanistic analysis as exemplified here for recovery and resilience hardly possible in traditional animal models. These humble first steps demonstrate the potential of 21st century cell culture and measurement technologies to advance our understanding and design testing strategies for old and new challenges such as nanoparticles to improve public health.
305
9. REFERENCES
Alberio, T., L. Lopiano, and M. Fasano. 2012. 'Cellular models to investigate biochemical
pathways in Parkinson's disease', FEBS J, 279: 1146-55.
Alépée, N, Bahinski T, Daneshian M, De Wever B, Fritsche E, Goldberg A, Hansmann J, Hartung
T, Haycock J, Hogberg H, Hoelting L, Kelm JM, Kadereit S, McVey E, Landsiedel R,
Leist M, Lübberstedt M, Noor F, ellevoisin C, Petersohn D, Pfannenbecker U, Reisinger K,
Ramirez T, Rothen-Rutishauser B, Schäfer-Korting M, Zeilinger K, and and Zurich M-G.
2014. ' State-of-the-art of 3D cultures (organs-on-a-chip) in safety testing and
pathophysiology - a t4 report. ', ALTEX, 31:: 441-77.
Altman, J., and G. D. Das. 1965. 'Autoradiographic and histological evidence of postnatal
hippocampal neurogenesis in rats', J Comp Neurol, 124: 319-35.
Ammal Kaidery, N., S. Tarannum, and B. Thomas. 2013. 'Epigenetic landscape of Parkinson's
disease: emerging role in disease mechanisms and therapeutic modalities',
Neurotherapeutics, 10: 698-708.
An, Y., L. Tang, X. Jiang, H. Chen, M. Yang, L. Jin, S. Zhang, C. Wang, and W. Zhang. 2010. 'A
photoelectrochemical immunosensor based on Au-doped TiO2 nanotube arrays for the
detection of alpha-synuclein', Chemistry, 16: 14439-46.
Andres, R. H., A. W. Huber, U. Schlattner, A. Perez-Bouza, S. H. Krebs, R. W. Seiler, T.
Wallimann, and H. R. Widmer. 2005. 'Effects of creatine treatment on the survival of
dopaminergic neurons in cultured fetal ventral mesencephalic tissue', Neuroscience, 133:
701-13.
Ariga, H., K. Takahashi-Niki, I. Kato, H. Maita, T. Niki, and S. M. Iguchi-Ariga. 2013.
'Neuroprotective function of DJ-1 in Parkinson's disease', Oxid Med Cell Longev, 2013:
683920.
306
Aschner, M., S. Ceccatelli, M. Daneshian, E. Fritsche, N. Hasiwa, T. Hartung, H. T. Hogberg, M.
Leist, A. Li, W. R. Mundi, S. Padilla, A. H. Piersma, A. Bal-Price, A. Seiler, R. H.
Westerink, B. Zimmer, and P. J. Lein. 2017. 'Reference compounds for alternative test
methods to indicate developmental neurotoxicity (DNT) potential of chemicals: example
lists and criteria for their selection and use', ALTEX, 34: 49-74.
Bababunmi, E. A., O. O. Olorunsogo, and O. Bassir. 1979. 'The uncoupling effect of N-
(phosphonomethyl)glycine on isolated rat liver mitochondria', Biochem Pharmacol, 28:
925-7.
Bae, E. J., N. Y. Yang, C. Lee, H. J. Lee, S. Kim, S. P. Sardi, and S. J. Lee. 2015. 'Loss of
glucocerebrosidase 1 activity causes lysosomal dysfunction and alpha-synuclein
aggregation', Exp Mol Med, 47: e153.
Bailey, J., M. Thew, and M. Balls. 2014. 'An analysis of the use of animal models in predicting
human toxicology and drug safety', Altern Lab Anim, 42: 181-99.
Bal-Price, A. K., H. T. Hogberg, L. Buzanska, and S. Coecke. 2010. 'Relevance of in vitro
neurotoxicity testing for regulatory requirements: challenges to be considered',
Neurotoxicol Teratol, 32: 36-41.
Bal-Price, A. K., C. Sunol, D. G. Weiss, E. van Vliet, R. H. Westerink, and L. G. Costa. 2008.
'Application of in vitro neurotoxicity testing for regulatory purposes: Symposium III
summary and research needs', Neurotoxicology, 29: 520-31.
Bal-Price, A., and M. E. B. Meek. 2017. 'Adverse outcome pathways: Application to enhance
mechanistic understanding of neurotoxicity', Pharmacol Ther, 179: 84-95.
Balestrino, M., M. Lensman, M. Parodi, L. Perasso, R. Rebaudo, R. Melani, S. Polenov, and A.
Cupello. 2002. 'Role of creatine and phosphocreatine in neuronal protection from anoxic
and ischemic damage', Amino Acids, 23: 221-9.
307
Baltazar, M. T., R. J. Dinis-Oliveira, M. de Lourdes Bastos, A. M. Tsatsakis, J. A. Duarte, and F.
Carvalho. 2014. 'Pesticides exposure as etiological factors of Parkinson's disease and other
neurodegenerative diseases--a mechanistic approach', Toxicol Lett, 230: 85-103.
Barbeau, A. 1969. 'L-dopa therapy in Parkinson's disease: a critical review of nine years'
experience', Can Med Assoc J, 101: 59-68.
Barbosa Egberto, R., D. Leiros da Costa Maria, A. Bacheschi Luiz, Milberto Scaff, and C. Leite
Claudia. 2001. 'Parkinsonism after glycine‐derivate exposure', Movement Disorders, 16:
565-68.
Barhoumi, R., Y. Qian, R. C. Burghardt, and E. Tiffany-Castiglioni. 2010. 'Image analysis of Ca2+
signals as a basis for neurotoxicity assays: promises and challenges', Neurotoxicol Teratol,
32: 16-24.
Bartels, A. L., and K. L. Leenders. 2007. 'Neuroinflammation in the pathophysiology of Parkinson's
disease: evidence from animal models to human in vivo studies with [11C]-PK11195 PET',
Mov Disord, 22: 1852-6.
Basketter, D. A., H. Clewell, I. Kimber, A. Rossi, B. Blaauboer, R. Burrier, M. Daneshian, C.
Eskes, A. Goldberg, N. Hasiwa, S. Hoffmann, J. Jaworska, T. B. Knudsen, R. Landsiedel,
M. Leist, P. Locke, G. Maxwell, J. McKim, E. A. McVey, G. Ouedraogo, G. Patlewicz, O.
Pelkonen, E. Roggen, C. Rovida, I. Ruhdel, M. Schwarz, A. Schepky, G. Schoeters, N.
Skinner, K. Trentz, M. Turner, P. Vanparys, J. Yager, J. Zurlo, and T. Hartung. 2012. 'A
roadmap for the development of alternative (non-animal) methods for systemic toxicity
testing - t4 report*', ALTEX, 29: 3-91.
Beal, M. F. 2010. 'Parkinson's disease: a model dilemma', Nature, 466: S8-10.
Belanger, M., I. Allaman, and P. J. Magistretti. 2011. 'Brain energy metabolism: focus on astrocyte-
neuron metabolic cooperation', Cell Metab, 14: 724-38.
308
Berchtold, N. C., G. Chinn, M. Chou, J. P. Kesslak, and C. W. Cotman. 2005. 'Exercise primes a
molecular memory for brain-derived neurotrophic factor protein induction in the rat
hippocampus', Neuroscience, 133: 853-61.
Betarbet, R., T. B. Sherer, G. MacKenzie, M. Garcia-Osuna, A. V. Panov, and J. T. Greenamyre.
2000. 'Chronic systemic pesticide exposure reproduces features of Parkinson's disease', Nat
Neurosci, 3: 1301-6.
Bezem, M. T., F. G. Johannessen, K. Jung-Kc, E. T. Gundersen, A. Jorge-Finnigan, M. Ying, D.
Betbeder, L. Herfindal, and A. Martinez. 2018. 'Stabilization of Human Tyrosine
Hydroxylase in Maltodextrin Nanoparticles for Delivery to Neuronal Cells and Tissue',
Bioconjug Chem, 29: 493-502.
Biskup, S., D. J. Moore, F. Celsi, S. Higashi, A. B. West, S. A. Andrabi, K. Kurkinen, S. W. Yu, J.
M. Savitt, H. J. Waldvogel, R. L. Faull, P. C. Emson, R. Torp, O. P. Ottersen, T. M.
Dawson, and V. L. Dawson. 2006. 'Localization of LRRK2 to membranous and vesicular
structures in mammalian brain', Ann Neurol, 60: 557-69.
Blaauboer, B. J., K. Boekelheide, H. J. Clewell, M. Daneshian, M. M. Dingemans, A. M. Goldberg,
M. Heneweer, J. Jaworska, N. I. Kramer, M. Leist, H. Seibert, E. Testai, R. J. Vandebriel,
J. D. Yager, and J. Zurlo. 2012. 'The use of biomarkers of toxicity for integrating in vitro
hazard estimates into risk assessment for humans', ALTEX, 29: 411-25.
Blesa, J., and S. Przedborski. 2014. 'Parkinson's disease: animal models and dopaminergic cell
vulnerability', Front Neuroanat, 8: 155.
Blesa, J., I. Trigo-Damas, M. Dileone, N. L. Del Rey, L. F. Hernandez, and J. A. Obeso. 2017.
'Compensatory mechanisms in Parkinson's disease: Circuits adaptations and role in disease
modification', Exp Neurol, 298: 148-61.
Bolam, J. P., and E. K. Pissadaki. 2012. 'Living on the edge with too many mouths to feed: why
dopamine neurons die', Mov Disord, 27: 1478-83.
309
Bondos, S. E., and A. Bicknell. 2003. 'Detection and prevention of protein aggregation before,
during, and after purification', Anal Biochem, 316: 223-31.
Branch, S. Y., R. Sharma, and M. J. Beckstead. 2014. 'Aging decreases L-type calcium channel
currents and pacemaker firing fidelity in substantia nigra dopamine neurons', J Neurosci,
34: 9310-8.
Breckenridge, C. B., C. Berry, E. T. Chang, R. L. Sielken, Jr., and J. S. Mandel. 2016. 'Association
between Parkinson's Disease and Cigarette Smoking, Rural Living, Well-Water
Consumption, Farming and Pesticide Use: Systematic Review and Meta-Analysis', PLoS
One, 11: e0151841.
Brini, M., T. Cali, D. Ottolini, and E. Carafoli. 2014. 'Neuronal calcium signaling: function and
dysfunction', Cell Mol Life Sci, 71: 2787-814.
Burello, E., and A. Worth. 2011. 'Computational nanotoxicology: Predicting toxicity of
nanoparticles', Nat Nanotechnol, 6: 138-9.
Busquet, F., and T. Hartung. 2017. 'The need for strategic development of safety sciences', ALTEX,
34: 3-21.
Cacciatore, I., M. Ciulla, E. Fornasari, L. Marinelli, and A. Di Stefano. 2016. 'Solid lipid
nanoparticles as a drug delivery system for the treatment of neurodegenerative diseases',
Expert Opin Drug Deliv, 13: 1121-31.
Caligiore, Daniele, Rick C. Helmich, Mark Hallett, Ahmed A. Moustafa, Lars Timmermann, Ivan
Toni, and Gianluca Baldassarre. 2016. 'Parkinson’s disease as a system-level disorder', Npj
Parkinson'S Disease, 2: 16025.
Cannon, J. R., and J. T. Greenamyre. 2011. 'The role of environmental exposures in
neurodegeneration and neurodegenerative diseases', Toxicol Sci, 124: 225-50.
Cannon, J. R., V. Tapias, H. M. Na, A. S. Honick, R. E. Drolet, and J. T. Greenamyre. 2009. 'A
highly reproducible rotenone model of Parkinson's disease', Neurobiol Dis, 34: 279-90.
310
Chan, P., L. E. DeLanney, I. Irwin, J. W. Langston, and D. Di Monte. 1991. 'Rapid ATP loss
caused by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine in mouse brain', J Neurochem, 57:
348-51.
Chang, D., M. A. Nalls, I. B. Hallgrimsdottir, J. Hunkapiller, M. van der Brug, F. Cai, Consortium
International Parkinson's Disease Genomics, Team andMe Research, G. A. Kerchner, G.
Ayalon, B. Bingol, M. Sheng, D. Hinds, T. W. Behrens, A. B. Singleton, T. R. Bhangale,
and R. R. Graham. 2017. 'A meta-analysis of genome-wide association studies identifies 17
new Parkinson's disease risk loci', Nat Genet, 49: 1511-16.
Chartier-Harlin, M. C., J. Kachergus, C. Roumier, V. Mouroux, X. Douay, S. Lincoln, C.
Levecque, L. Larvor, J. Andrieux, M. Hulihan, N. Waucquier, L. Defebvre, P. Amouyel,
M. Farrer, and A. Destee. 2004. 'Alpha-synuclein locus duplication as a cause of familial
Parkinson's disease', Lancet, 364: 1167-9.
Chesselet, M. F., and F. Richter. 2011. 'Modelling of Parkinson's disease in mice', Lancet Neurol,
10: 1108-18.
Chinta, S. J., and J. K. Andersen. 2005. 'Dopaminergic neurons', Int J Biochem Cell Biol, 37: 942-6.
Chinta S.J. 2006. 'Reversible inhibition of mitochondrial complex I activity following chronic
dopaminergic glutathione depletion in vitro: implications for Parkinson's disease', Free
Radic Biol Med, 41: 1442-8.
Choi, S. H., Y. H. Kim, M. Hebisch, C. Sliwinski, S. Lee, C. D'Avanzo, H. Chen, B. Hooli, C.
Asselin, J. Muffat, J. B. Klee, C. Zhang, B. J. Wainger, M. Peitz, D. M. Kovacs, C. J.
Woolf, S. L. Wagner, R. E. Tanzi, and D. Y. Kim. 2014. 'A three-dimensional human
neural cell culture model of Alzheimer's disease', Nature, 515: 274-8.
Choi, W. S., S. E. Kruse, R. D. Palmiter, and Z. Xia. 2008. 'Mitochondrial complex I inhibition is
not required for dopaminergic neuron death induced by rotenone, MPP+, or paraquat', Proc
Natl Acad Sci U S A, 105: 15136-41.
311
Clift, M. J., P. Gehr, and B. Rothen-Rutishauser. 2011. 'Nanotoxicology: a perspective and
discussion of whether or not in vitro testing is a valid alternative', Arch Toxicol, 85: 723-31.
Coecke, S., A. M. Goldberg, S. Allen, L. Buzanska, G. Calamandrei, K. Crofton, L. Hareng, T.
Hartung, H. Knaut, P. Honegger, M. Jacobs, P. Lein, A. Li, W. Mundy, D. Owen, S.
Schneider, E. Silbergeld, T. Reum, T. Trnovec, F. Monnet-Tschudi, and A. Bal-Price.
2007. 'Workgroup report: incorporating in vitro alternative methods for developmental
neurotoxicity into international hazard and risk assessment strategies', Environ Health
Perspect, 115: 924-31.
Colovic, M. B., D. Z. Krstic, T. D. Lazarevic-Pasti, A. M. Bondzic, and V. M. Vasic. 2013.
'Acetylcholinesterase inhibitors: pharmacology and toxicology', Curr Neuropharmacol, 11:
315-35. da Costa, C. A. 2007. 'DJ-1: a newcomer in Parkinson's disease pathology', Curr Mol Med, 7: 650-
7.
Daskalakis, N. P., R. C. Bagot, K. J. Parker, C. H. Vinkers, and E. R. de Kloet. 2013. 'The three-hit
concept of vulnerability and resilience: toward understanding adaptation to early-life
adversity outcome', Psychoneuroendocrinology, 38: 1858-73.
Davey, G. P., and J. B. Clark. 1996. 'Threshold effects and control of oxidative phosphorylation in
nonsynaptic rat brain mitochondria', J Neurochem, 66: 1617-24.
De Simone, U., F. Caloni, L. Gribaldo, and T. Coccini. 2017. 'Human Co-culture Model of Neurons
and Astrocytes to Test Acute Cytotoxicity of Neurotoxic Compounds', Int J Toxicol, 36:
463-77.
Dhillon, A. S., G. L. Tarbutton, J. L. Levin, G. M. Plotkin, L. K. Lowry, J. T. Nalbone, and S.
Shepherd. 2008. 'Pesticide/environmental exposures and Parkinson's disease in East Texas',
J Agromedicine, 13: 37-48.
312
Dingemans, M. M., M. van den Berg, and R. H. Westerink. 2011. 'Neurotoxicity of brominated
flame retardants: (in)direct effects of parent and hydroxylated polybrominated diphenyl
ethers on the (developing) nervous system', Environ Health Perspect, 119: 900-7.
Dolinoy, D. C. 2008. 'The agouti mouse model: an epigenetic biosensor for nutritional and
environmental alterations on the fetal epigenome', Nutr Rev, 66 Suppl 1: S7-11.
Donaldson, K., C. A. Poland, F. A. Murphy, M. MacFarlane, T. Chernova, and A. Schinwald.
2013. 'Pulmonary toxicity of carbon nanotubes and asbestos - similarities and differences',
Adv Drug Deliv Rev, 65: 2078-86.
Dorsey, E. R., R. Constantinescu, J. P. Thompson, K. M. Biglan, R. G. Holloway, K. Kieburtz, F. J.
Marshall, B. M. Ravina, G. Schifitto, A. Siderowf, and C. M. Tanner. 2007. 'Projected
number of people with Parkinson disease in the most populous nations, 2005 through
2030', Neurology, 68: 384-6.
Double, K. L., M. Maywald, M. Schmittel, P. Riederer, and M. Gerlach. 1998. 'In vitro studies of
ferritin iron release and neurotoxicity', J Neurochem, 70: 2492-9.
Du, F., Q. Yu, A. Chen, D. Chen, and S. S. Yan. 2018. 'Astrocytes Attenuate Mitochondrial
Dysfunctions in Human Dopaminergic Neurons Derived from iPSC', Stem Cell Reports,
10: 366-74.
Efremova, L., S. Schildknecht, M. Adam, R. Pape, S. Gutbier, B. Hanf, A. Burkle, and M. Leist.
2015. 'Prevention of the degeneration of human dopaminergic neurons in an astrocyte co-
culture system allowing endogenous drug metabolism', Br J Pharmacol, 172: 4119-32.
Faria, M. A. 2015. 'Glyphosate, neurological diseases - and the scientific method', Surg Neurol Int,
6: 132.
Farrer, M. J. 2006. 'Genetics of Parkinson disease: paradigm shifts and future prospects', Nat Rev
Genet, 7: 306-18.
313
Ferrante, R. J., J. B. Schulz, N. W. Kowall, and M. F. Beal. 1997. 'Systemic administration of
rotenone produces selective damage in the striatum and globus pallidus, but not in the
substantia nigra', Brain Res, 753: 157-62.
Floor, E., and M. G. Wetzel. 1998. 'Increased protein oxidation in human substantia nigra pars
compacta in comparison with basal ganglia and prefrontal cortex measured with an
improved dinitrophenylhydrazine assay', J Neurochem, 70: 268-75.
Fraga, M. F., E. Ballestar, M. F. Paz, S. Ropero, F. Setien, M. L. Ballestar, D. Heine-Suner, J. C.
Cigudosa, M. Urioste, J. Benitez, M. Boix-Chornet, A. Sanchez-Aguilera, C. Ling, E.
Carlsson, P. Poulsen, A. Vaag, Z. Stephan, T. D. Spector, Y. Z. Wu, C. Plass, and M.
Esteller. 2005. 'Epigenetic differences arise during the lifetime of monozygotic twins', Proc
Natl Acad Sci U S A, 102: 10604-9.
Franco-Iborra, S., M. Vila, and C. Perier. 2016. 'The Parkinson Disease Mitochondrial Hypothesis:
Where Are We at?', Neuroscientist, 22: 266-77.
Furlong, M., C. M. Tanner, S. M. Goldman, G. S. Bhudhikanok, A. Blair, A. Chade, K. Comyns, J.
A. Hoppin, M. Kasten, M. Korell, J. W. Langston, C. Marras, C. Meng, M. Richards, G.
W. Ross, D. M. Umbach, D. P. Sandler, and F. Kamel. 2015. 'Protective glove use and
hygiene habits modify the associations of specific pesticides with Parkinson's disease',
Environ Int, 75: 144-50.
Gerhardt, Ellen, Sebastian Kügler, Marcel Leist, Christoph Beier, Laura Berliocchi, Christiane
Volbracht, Michael Weller, Mathias Bähr, Pierluigi Nicotera, and Jörg B. Schulz. 2001.
'Cascade of Caspase Activation in Potassium-Deprived Cerebellar Granule Neurons:
Targets for Treatment with Peptide and Protein Inhibitors of Apoptosis', Molecular and
Cellular Neuroscience, 17: 717-31.
German, D. C., K. Manaye, W. K. Smith, D. J. Woodward, and C. B. Saper. 1989. 'Midbrain
dopaminergic cell loss in Parkinson's disease: computer visualization', Ann Neurol, 26:
507-14.
314
Giordano, S., J. Lee, V. M. Darley-Usmar, and J. Zhang. 2012. 'Distinct effects of rotenone, 1-
methyl-4-phenylpyridinium and 6-hydroxydopamine on cellular bioenergetics and cell
death', PLoS One, 7: e44610.
Gordon, J., S. Amini, and M. K. White. 2013. 'General overview of neuronal cell culture', Methods
Mol Biol, 1078: 1-8.
Gordon, S., M. Daneshian, J. Bouwstra, F. Caloni, S. Constant, D. E. Davies, G. Dandekar, C. A.
Guzman, E. Fabian, E. Haltner, T. Hartung, N. Hasiwa, P. Hayden, H. Kandarova, S.
Khare, H. F. Krug, C. Kneuer, M. Leist, G. Lian, U. Marx, M. Metzger, K. Ott, P. Prieto,
M. S. Roberts, E. L. Roggen, T. Tralau, C. van den Braak, H. Walles, and C. M. Lehr.
2015. 'Non-animal models of epithelial barriers (skin, intestine and lung) in research,
industrial applications and regulatory toxicology', ALTEX, 32: 327-78.
Gotte, M., G. Hofmann, A. I. Michou-Gallani, J. F. Glickman, W. Wishart, and D. Gabriel. 2010.
'An imaging assay to analyze primary neurons for cellular neurotoxicity', J Neurosci
Methods, 192: 7-16.
Gottmann, E., S. Kramer, B. Pfahringer, and C. Helma. 2001. 'Data quality in predictive
toxicology: reproducibility of rodent carcinogenicity experiments', Environ Health
Perspect, 109: 509-14.
Goudriaan, A., N. Camargo, K. E. Carney, S. H. Oliet, A. B. Smit, and M. H. Verheijen. 2014.
'Novel cell separation method for molecular analysis of neuron-astrocyte co-cultures',
Front Cell Neurosci, 8: 12.
Greenamyre, J. T., R. Betarbet, and T. B. Sherer. 2003. 'The rotenone model of Parkinson's disease:
genes, environment and mitochondria', Parkinsonism Relat Disord, 9 Suppl 2: S59-64.
Grivennikova, V. G., E. O. Maklashina, E. V. Gavrikova, and A. D. Vinogradov. 1997. 'Interaction
of the mitochondrial NADH-ubiquinone reductase with rotenone as related to the enzyme
active/inactive transition', Biochim Biophys Acta, 1319: 223-32.
315
Groothuis, F. A., M. B. Heringa, B. Nicol, J. L. Hermens, B. J. Blaauboer, and N. I. Kramer. 2015.
'Dose metric considerations in in vitro assays to improve quantitative in vitro-in vivo dose
extrapolations', Toxicology, 332: 30-40.
Gui, Y. X., X. N. Fan, H. M. Wang, G. Wang, and S. D. Chen. 2012. 'Glyphosate induced cell
death through apoptotic and autophagic mechanisms', Neurotoxicol Teratol, 34: 344-9.
Gunhanlar, N., G. Shpak, M. van der Kroeg, L. A. Gouty-Colomer, S. T. Munshi, B. Lendemeijer,
M. Ghazvini, C. Dupont, W. J. G. Hoogendijk, J. Gribnau, F. M. S. de Vrij, and S. A.
Kushner. 2017. 'A simplified protocol for differentiation of electrophysiologically mature
neuronal networks from human induced pluripotent stem cells', Mol Psychiatry.
Gunnarsson, L. G., L. Bodin, B. Soderfeldt, and O. Axelson. 1992. 'A case-control study of motor
neurone disease: its relation to heritability, and occupational exposures, particularly to
solvents', Br J Ind Med, 49: 791-8.
Hansen, S. F., and A. Baun. 2012. 'European regulation affecting nanomaterials - review of
limitations and future recommendations', Dose Response, 10: 364-83.
Harry, G. J., M. Billingsley, A. Bruinink, I. L. Campbell, W. Classen, D. C. Dorman, C. Galli, D.
Ray, R. A. Smith, and H. A. Tilson. 1998. 'In vitro techniques for the assessment of
neurotoxicity', Environ Health Perspect, 106 Suppl 1: 131-58.
Hartung, T. 2007. 'Food for thought... on cell culture', ALTEX, 24: 143-52.
Hartung T. 2008. 'Food for thought ... on alternative methods for cosmetics safety testing', ALTEX,
25: 147-62.
Hartung T . 2010a. 'Food for thought ... on alternative methods for nanoparticle safety testing',
ALTEX, 27: 87-95.
Hartung T. 2010b. 'Food for thought...on alternative methods for chemical safety testing', ALTEX,
27: 3-14.
Hartung T. 2013. 'Look back in anger - what clinical studies tell us about preclinical work', ALTEX,
30: 275-91.
316
Hartung T. 2017a. 'Opinion versus evidence for the need to move away from animal testing',
ALTEX, 34: 193-200.
Hartung T. 2017b. 'Utility of the adverse outcome pathway concept in drug development', Expert
Opin Drug Metab Toxicol, 13: 1-3.
Hartung, T., S. Bremer, S. Casati, S. Coecke, R. Corvi, S. Fortaner, L. Gribaldo, M. Halder, S.
Hoffmann, A. J. Roi, P. Prieto, E. Sabbioni, L. Scott, A. Worth, and V. Zuang. 2004. 'A
modular approach to the ECVAM principles on test validity', Altern Lab Anim, 32: 467-72.
Hartung, T., and E. Sabbioni. 2011. 'Alternative in vitro assays in nanomaterial toxicology', Wiley
Interdiscip Rev Nanomed Nanobiotechnol, 3: 545-73.
Hartung, T., and J. Zurlo. 2012. 'Alternative approaches for medical countermeasures to biological
and chemical terrorism and warfare', ALTEX, 29: 251-60.
Hawthorne, G. H., M. P. Bernuci, M. Bortolanza, V. Tumas, A. C. Issy, and E. Del-Bel. 2016.
'Nanomedicine to Overcome Current Parkinson's Treatment Liabilities: A Systematic
Review', Neurotox Res, 30: 715-29.
Helmschrodt, C., S. Hobel, S. Schoniger, A. Bauer, J. Bonicelli, M. Gringmuth, S. A. Fietz, A.
Aigner, A. Richter, and F. Richter. 2017. 'Polyethylenimine Nanoparticle-Mediated siRNA
Delivery to Reduce alpha-Synuclein Expression in a Model of Parkinson's Disease', Mol
Ther Nucleic Acids, 9: 57-68.
Hendriks, H. S., and R. H. Westerink. 2015. 'Neurotoxicity and risk assessment of brominated and
alternative flame retardants', Neurotoxicol Teratol, 52: 248-69.
Henn, Iris H., Lena Bouman, Julia S. Schlehe, Anita Schlierf, Julia E. Schramm, Elmar Wegener,
Kazuhiro Nakaso, Carsten Culmsee, Benedikt Berninger, Daniel Krappmann, Jörg Tatzelt,
and Konstanze F. Winklhofer. 2007. 'Parkin mediates neuroprotection through activation of
IkappaB kinase/nuclear factor-kappaB signaling', The Journal of neuroscience : the official
journal of the Society for Neuroscience, 27: 1868-78.
317
Hernandez, D. G., X. Reed, and A. B. Singleton. 2016. 'Genetics in Parkinson disease: Mendelian
versus non-Mendelian inheritance', J Neurochem, 139 Suppl 1: 59-74.
Hertzman, C., M. Wiens, D. Bowering, B. Snow, and D. Calne. 1990. 'Parkinson's disease: a case-
control study of occupational and environmental risk factors', Am J Ind Med, 17: 349-55.
Heusinkveld, H. J., and R. H. S. Westerink. 2017. 'Comparison of different in vitro cell models for
the assessment of pesticide-induced dopaminergic neurotoxicity', Toxicol In vitro, 45: 81-
88.
Higgins, D. S., Jr., and J. T. Greenamyre. 1996. '[3H]dihydrorotenone binding to NADH:
ubiquinone reductase (complex I) of the electron transport chain: an autoradiographic
study', J Neurosci, 16: 3807-16.
Hirsch, Etienne, Ann M. Graybiel, and Yves A. Agid. 1988. 'Melanized dopaminergic neurons are
differentially susceptible to degeneration in Parkinson's disease', Nature, 334: 345.
Hogberg, H. T., J. Bressler, K. M. Christian, G. Harris, G. Makri, C. O'Driscoll, D. Pamies, L.
Smirnova, Z. Wen, and T. Hartung. 2013. 'Toward a 3D model of human brain
development for studying gene/environment interactions', Stem Cell Res Ther, 4 Suppl 1:
S4.
Hogberg, H. T., T. Sobanski, A. Novellino, M. Whelan, D. G. Weiss, and A. K. Bal-Price. 2011.
'Application of micro-electrode arrays (MEAs) as an emerging technology for
developmental neurotoxicity: evaluation of domoic acid-induced effects in primary cultures
of rat cortical neurons', Neurotoxicology, 32: 158-68.
Honegger, P., Lenoir, D., & Favrod, P. (1979). Growth and differentiation of aggregating fetal
brain cells in a serum-free defined medium. Nature, 282(5736), 305–308.
Honegger P, Monnet-Tschudi F. 2001. Aggregating neural cell cultures. In: Protocols for Neural
Cell Culture. (Fedoroff S, Richardson A, eds). 3rd ed. Ottawa:Humana Press, 199–218.
Hsu, M., B. Srinivas, J. Kumar, R. Subramanian, and J. Andersen. 2005. 'Glutathione depletion
resulting in selective mitochondrial complex I inhibition in dopaminergic cells is via an
318
NO-mediated pathway not involving peroxynitrite: implications for Parkinson's disease', J
Neurochem, 92: 1091-103.
Hulla, J. E., S. C. Sahu, and A. W. Hayes. 2015. 'Nanotechnology: History and future', Hum Exp
Toxicol, 34: 1318-21.
Jenner, P. 1993. 'Altered mitochondrial function, iron metabolism and glutathione levels in
Parkinson's disease', Acta Neurol Scand Suppl, 146: 6-13.
Jenner, P., D. T. Dexter, J. Sian, A. H. Schapira, and C. D. Marsden. 1992. 'Oxidative stress as a
cause of nigral cell death in Parkinson's disease and incidental Lewy body disease. The
Royal Kings and Queens Parkinson's Disease Research Group', Ann Neurol, 32 Suppl: S82-
7.
Jeon, S. M., S. M. Cheon, H. R. Bae, J. W. Kim, and S. U. Kim. 2010. 'Selective susceptibility of
human dopaminergic neural stem cells to dopamine-induced apoptosis', Exp Neurobiol, 19:
155-64.
Johnstone, A. F., G. W. Gross, D. G. Weiss, O. H. Schroeder, A. Gramowski, and T. J. Shafer.
2010. 'Microelectrode arrays: a physiologically based neurotoxicity testing platform for the
21st century', Neurotoxicology, 31: 331-50.
Jorfi, M., C. D'Avanzo, D. Y. Kim, and D. Irimia. 2018. 'Three-Dimensional Models of the Human
Brain Development and Diseases', Adv Healthc Mater, 7.
Juarez Olguin, H., D. Calderon Guzman, E. Hernandez Garcia, and G. Barragan Mejia. 2016. 'The
Role of Dopamine and Its Dysfunction as a Consequence of Oxidative Stress', Oxid Med
Cell Longev, 2016: 9730467.
Kaidery, A.N., S. Tarannum, and B. Thomas. 2013. 'Epigenetic landscape of Parkinson's disease:
emerging role in disease mechanisms and therapeutic modalities', Neurotherapeutics, 10:
698-708.
Keane, P. C., M. Kurzawa, P. G. Blain, and C. M. Morris. 2011. 'Mitochondrial dysfunction in
Parkinson's disease', Parkinsons Dis, 2011: 716871.
319
Khan, F. H., T. Sen, A. K. Maiti, S. Jana, U. Chatterjee, and S. Chakrabarti. 2005. 'Inhibition of rat
brain mitochondrial electron transport chain activity by dopamine oxidation products
during extended in vitro incubation: implications for Parkinson's disease', Biochim Biophys
Acta, 1741: 65-74.
Khan, Ibrahim, Khalid Saeed, and Idrees Khan. 2017. 'Nanoparticles: Properties, applications and
toxicities', Arabian Journal of Chemistry.
Khanbabaie, R., and M. Jahanshahi. 2012. 'Revolutionary impact of nanodrug delivery on
neuroscience', Curr Neuropharmacol, 10: 370-92.
Kleensang, A., A. Maertens, M. Rosenberg, S. Fitzpatrick, J. Lamb, S. Auerbach, R. Brennan, K.
M. Crofton, B. Gordon, A. J. Fornace, Jr., K. Gaido, D. Gerhold, R. Haw, A. Henney, A.
Ma'ayan, M. McBride, S. Monti, M. F. Ochs, A. Pandey, R. Sharan, R. Stierum, S.
Tugendreich, C. Willett, C. Wittwehr, J. Xia, G. W. Patton, K. Arvidson, M. Bouhifd, H. T.
Hogberg, T. Luechtefeld, L. Smirnova, L. Zhao, Y. Adeleye, M. Kanehisa, P. Carmichael,
M. E. Andersen, and T. Hartung. 2014. 't4 workshop report: Pathways of Toxicity', ALTEX,
31: 53-61.
Krug, A. K., N. V. Balmer, F. Matt, F. Schonenberger, D. Merhof, and M. Leist. 2013. 'Evaluation
of a human neurite growth assay as specific screen for developmental neurotoxicants', Arch
Toxicol, 87: 2215-31.
Krug, A. K., S. Gutbier, L. Zhao, D. Poltl, C. Kullmann, V. Ivanova, S. Forster, S. Jagtap, J.
Meiser, G. Leparc, S. Schildknecht, M. Adam, K. Hiller, H. Farhan, T. Brunner, T.
Hartung, A. Sachinidis, and M. Leist. 2014. 'Transcriptional and metabolic adaptation of
human neurons to the mitochondrial toxicant MPP(+)', Cell Death Dis, 5: e1222.
Kweon, G. R., J. D. Marks, R. Krencik, E. H. Leung, P. T. Schumacker, K. Hyland, and U. J. Kang.
2004. 'Distinct mechanisms of neurodegeneration induced by chronic complex I inhibition
in dopaminergic and non-dopaminergic cells', J Biol Chem, 279: 51783-92.
320
Lancaster, M. A., and J. A. Knoblich. 2014. 'Organogenesis in a dish: modeling development and
disease using organoid technologies', Science, 345: 1247125.
Lancaster, M. A., M. Renner, C. A. Martin, D. Wenzel, L. S. Bicknell, M. E. Hurles, T. Homfray, J.
M. Penninger, A. P. Jackson, and J. A. Knoblich. 2013. 'Cerebral organoids model human
brain development and microcephaly', Nature, 501: 373-9.
Langston, J. W., P. Ballard, J. W. Tetrud, and I. Irwin. 1983. 'Chronic Parkinsonism in humans due
to a product of meperidine-analog synthesis', Science, 219: 979-80.
Langston, J. W., L. S. Forno, J. Tetrud, A. G. Reeves, J. A. Kaplan, and D. Karluk. 1999. 'Evidence
of active nerve cell degeneration in the substantia nigra of humans years after 1-methyl-4-
phenyl-1,2,3,6-tetrahydropyridine exposure', Ann Neurol, 46: 598-605.
Larner, S. F., J. Wang, J. Goodman, M. B. O. Altman, M. Xin, and K. K. W. Wang. 2017. 'In vitro
Neurotoxicity Resulting from Exposure of Cultured Neural Cells to Several Types of
Nanoparticles', J Cell Death, 10: 1179670717694523.
Lassus, B., S. Magnifico, S. Pignon, P. Belenguer, M. C. Miquel, and J. M. Peyrin. 2016.
'Alterations of mitochondrial dynamics allow retrograde propagation of locally initiated
axonal insults', Sci Rep, 6: 32777.
Lazaro, D. F., M. A. S. Pavlou, and T. F. Outeiro. 2017. 'Cellular models as tools for the study of
the role of alpha-synuclein in Parkinson's disease', Exp Neurol, 298: 162-71.
Lee, C. T., R. M. Bendriem, W. W. Wu, and R. F. Shen. 2017. '3D brain Organoids derived from
pluripotent stem cells: promising experimental models for brain development and
neurodegenerative disorders', J Biomed Sci, 24: 59.
Leenders, K. L., E. P. Salmon, P. Tyrrell, D. Perani, D. J. Brooks, H. Sager, T. Jones, C. D.
Marsden, and R. S. Frackowiak. 1990. 'The nigrostriatal dopaminergic system assessed in
vivo by positron emission tomography in healthy volunteer subjects and patients with
Parkinson's disease', Arch Neurol, 47: 1290-8.
321
Leist, M., A. Ghallab, R. Graepel, R. Marchan, R. Hassan, S. H. Bennekou, A. Limonciel, M.
Vinken, S. Schildknecht, T. Waldmann, E. Danen, B. van Ravenzwaay, H. Kamp, I.
Gardner, P. Godoy, F. Y. Bois, A. Braeuning, R. Reif, F. Oesch, D. Drasdo, S. Hohme, M.
Schwarz, T. Hartung, T. Braunbeck, J. Beltman, H. Vrieling, F. Sanz, A. Forsby, D.
Gadaleta, C. Fisher, J. Kelm, D. Fluri, G. Ecker, B. Zdrazil, A. Terron, P. Jennings, B. van
der Burg, S. Dooley, A. H. Meijer, E. Willighagen, M. Martens, C. Evelo, E. Mombelli, O.
Taboureau, A. Mantovani, B. Hardy, B. Koch, S. Escher, C. van Thriel, C. Cadenas, D.
Kroese, B. van de Water, and J. G. Hengstler. 2017. 'Adverse outcome pathways:
opportunities, limitations and open questions', Arch Toxicol, 91: 3477-505.
Li, N., K. Ragheb, G. Lawler, J. Sturgis, B. Rajwa, J. A. Melendez, and J. P. Robinson. 2003.
'Mitochondrial complex I inhibitor rotenone induces apoptosis through enhancing
mitochondrial reactive oxygen species production', J Biol Chem, 278: 8516-25.
Li, W., W. Sun, Y. Zhang, W. Wei, R. Ambasudhan, P. Xia, M. Talantova, T. Lin, J. Kim, X.
Wang, W. R. Kim, S. A. Lipton, K. Zhang, and S. Ding. 2011. 'Rapid induction and long-
term self-renewal of primitive neural precursors from human embryonic stem cells by
small molecule inhibitors', Proc Natl Acad Sci U S A, 108: 8299-304.
Liang, C. L., T. T. Wang, K. Luby-Phelps, and D. C. German. 2007. 'Mitochondria mass is low in
mouse substantia nigra dopamine neurons: implications for Parkinson's disease', Exp
Neurol, 203: 370-80.
Liou, H. H., M. C. Tsai, C. J. Chen, J. S. Jeng, Y. C. Chang, S. Y. Chen, and R. C. Chen. 1997.
'Environmental risk factors and Parkinson's disease: a case-control study in Taiwan',
Neurology, 48: 1583-8.
Liu, H. F., P. W. Ho, G. C. Leung, C. S. Lam, S. Y. Pang, L. Li, M. H. Kung, D. B. Ramsden, and
S. L. Ho. 2017. 'Combined LRRK2 mutation, aging and chronic low dose oral rotenone as
a model of Parkinson's disease', Sci Rep, 7: 40887.
322
Lorenz, C., P. Lesimple, R. Bukowiecki, A. Zink, G. Inak, B. Mlody, M. Singh, M. Semtner, N.
Mah, K. Aure, M. Leong, O. Zabiegalov, E. M. Lyras, V. Pfiffer, B. Fauler, J. Eichhorst, B.
Wiesner, N. Huebner, J. Priller, T. Mielke, D. Meierhofer, Z. Izsvak, J. C. Meier, F.
Bouillaud, J. Adjaye, M. Schuelke, E. E. Wanker, A. Lombes, and A. Prigione. 2017.
'Human iPSC-Derived Neural Progenitors Are an Effective Drug Discovery Model for
Neurological mtDNA Disorders', Cell Stem Cell, 20: 659-74 e9.
Lotharius, J., S. Barg, P. Wiekop, C. Lundberg, H. K. Raymon, and P. Brundin. 2002. 'Effect of
mutant alpha-synuclein on dopamine homeostasis in a new human mesencephalic cell line',
J Biol Chem, 277: 38884-94.
Lotharius, J., J. Falsig, J. van Beek, S. Payne, R. Dringen, P. Brundin, and M. Leist. 2005.
'Progressive degeneration of human mesencephalic neuron-derived cells triggered by
dopamine-dependent oxidative stress is dependent on the mixed-lineage kinase pathway', J
Neurosci, 25: 6329-42.
Manji, H. K., G. J. Moore, G. Rajkowska, and G. Chen. 2000. 'Neuroplasticity and cellular
resilience in mood disorders', Mol Psychiatry, 5: 578-93.
Manning-Bog, A. B., A. L. McCormack, J. Li, V. N. Uversky, A. L. Fink, and D. A. Di Monte.
2002. 'The herbicide paraquat causes up-regulation and aggregation of alpha-synuclein in
mice: paraquat and alpha-synuclein', J Biol Chem, 277: 1641-4.
Marambaud, P., U. Dreses-Werringloer, and V. Vingtdeux. 2009. 'Calcium signaling in
neurodegeneration', Mol Neurodegener, 4: 20.
Mariani, E., M. C. Polidori, A. Cherubini, and P. Mecocci. 2005. 'Oxidative stress in brain aging,
neurodegenerative and vascular diseases: an overview', J Chromatogr B Analyt Technol
Biomed Life Sci, 827: 65-75.
Martinez-Morales, P. L., and I. Liste. 2012. 'Stem cells as in vitro model of Parkinson's disease',
Stem Cells Int, 2012: 980941.
323
Masters, J. R. 2002. 'HeLa cells 50 years on: the good, the bad and the ugly', Nat Rev Cancer, 2:
315-9.
Matosin, N., E. Frank, M. Engel, J. S. Lum, and K. A. Newell. 2014. 'Negativity towards negative
results: a discussion of the disconnect between scientific worth and scientific culture', Dis
Model Mech, 7: 171-3.
McGeer, P. L., and E. G. McGeer. 2008. 'Glial reactions in Parkinson's disease', Mov Disord, 23:
474-83.
McNaught, K. S., C. W. Olanow, B. Halliwell, O. Isacson, and P. Jenner. 2001. 'Failure of the
ubiquitin-proteasome system in Parkinson's disease', Nat Rev Neurosci, 2: 589-94.
Meco, G., V. Bonifati, N. Vanacore, and E. Fabrizio. 1994. 'Parkinsonism after chronic exposure to
the fungicide maneb (manganese ethylene-bis-dithiocarbamate)', Scand J Work Environ
Health, 20: 301-5.
Menazza, S., J. Sun, S. Appachi, K. L. Chambliss, S. H. Kim, A. Aponte, S. Khan, J. A.
Katzenellenbogen, B. S. Katzenellenbogen, P. W. Shaul, and E. Murphy. 2017. 'Non-
nuclear estrogen receptor alpha activation in endothelium reduces cardiac ischemia-
reperfusion injury in mice', J Mol Cell Cardiol, 107: 41-51.
Mesnage, R., and M. N. Antoniou. 2017. 'Facts and Fallacies in the Debate on Glyphosate
Toxicity', Front Public Health, 5: 316.
Meulener, M., A. J. Whitworth, C. E. Armstrong-Gold, P. Rizzu, P. Heutink, P. D. Wes, L. J.
Pallanck, and N. M. Bonini. 2005. 'Drosophila DJ-1 mutants are selectively sensitive to
environmental toxins associated with Parkinson's disease', Curr Biol, 15: 1572-7.
Moors, M., T. D. Rockel, J. Abel, J. E. Cline, K. Gassmann, T. Schreiber, J. Schuwald, N.
Weinmann, and E. Fritsche. 2009. 'Human neurospheres as three-dimensional cellular
systems for developmental neurotoxicity testing', Environ Health Perspect, 117: 1131-8.
Mueller, S. N., and L. K. Mackay. 2016. 'Tissue-resident memory T cells: local specialists in
immune defence', Nat Rev Immunol, 16: 79-89.
324
Murphy, E., M. Kohr, S. Menazza, T. Nguyen, A. Evangelista, J. Sun, and C. Steenbergen. 2014.
'Signaling by S-nitrosylation in the heart', J Mol Cell Cardiol, 73: 18-25.
Mushtaq, G., J. A. Khan, E. Joseph, and M. A. Kamal. 2015. 'Nanoparticles, Neurotoxicity and
Neurodegenerative Diseases', Curr Drug Metab, 16: 676-84.
Nakamura, K., V. P. Bindokas, J. D. Marks, D. A. Wright, D. M. Frim, R. J. Miller, and U. J. Kang.
2000. 'The selective toxicity of 1-methyl-4-phenylpyridinium to dopaminergic neurons: the
role of mitochondrial complex I and reactive oxygen species revisited', Mol Pharmacol, 58:
271-8.
Nakamura, T., S. Tu, M. W. Akhtar, C. R. Sunico, S. Okamoto, and S. A. Lipton. 2013. 'Aberrant
protein s-nitrosylation in neurodegenerative diseases', Neuron, 78: 596-614.
Nandipati, S., and I. Litvan. 2016. 'Environmental Exposures and Parkinson's Disease', Int J
Environ Res Public Health, 13.
Negga, R., D. A. Rudd, N. S. Davis, A. N. Justice, H. E. Hatfield, A. L. Valente, A. S. Fields, and
V. A. Fitsanakis. 2011. 'Exposure to Mn/Zn ethylene-bis-dithiocarbamate and glyphosate
pesticides leads to neurodegeneration in Caenorhabditis elegans', Neurotoxicology, 32:
331-41.
Novellino, A., B. Scelfo, T. Palosaari, A. Price, T. Sobanski, T. J. Shafer, A. F. Johnstone, G. W.
Gross, A. Gramowski, O. Schroeder, K. Jugelt, M. Chiappalone, F. Benfenati, S. Martinoia,
M. T. Tedesco, E. Defranchi, P. D'Angelo, and M. Whelan. 2011. 'Development of micro-
electrode array based tests for neurotoxicity: assessment of interlaboratory reproducibility
with neuroactive chemicals', Front Neuroeng, 4: 4.
NRC. 1993. Pesticides in the Diets of Infants and Children (Washington (DC)).
Oberdorster, G., A. Maynard, K. Donaldson, V. Castranova, J. Fitzpatrick, K. Ausman, J. Carter, B.
Karn, W. Kreyling, D. Lai, S. Olin, N. Monteiro-Riviere, D. Warheit, H. Yang, and Ilsi
Research Foundation/Risk Science Institute Nanomaterial Toxicity Screening Working
325
Group. 2005. 'Principles for characterizing the potential human health effects from
exposure to nanomaterials: elements of a screening strategy', Part Fibre Toxicol, 2: 8.
Olorunsogo, O. O., E. A. Bababunmi, and O. Bassir. 1980. 'Interaction of N-
(phosphonomethyl)glycine with some respiratory chain enzymes of isolated corn-shoot
mitochondria', Arch Environ Contam Toxicol, 9: 109-14.
Olson, H., G. Betton, D. Robinson, K. Thomas, A. Monro, G. Kolaja, P. Lilly, J. Sanders, G. Sipes,
W. Bracken, M. Dorato, K. Van Deun, P. Smith, B. Berger, and A. Heller. 2000.
'Concordance of the toxicity of pharmaceuticals in humans and in animals', Regul Toxicol
Pharmacol, 32: 56-67.
Pamies, D., P. Barreras, K. Block, G. Makri, A. Kumar, D. Wiersma, L. Smirnova, C. Zang, J.
Bressler, K. M. Christian, G. Harris, G. L. Ming, C. J. Berlinicke, K. Kyro, H. Song, C. A.
Pardo, T. Hartung, and H. T. Hogberg. 2017. 'A human brain microphysiological system
derived from induced pluripotent stem cells to study neurological diseases and toxicity',
ALTEX, 34: 362-76.
Pamies, D., and T. Hartung. 2017. '21st Century Cell Culture for 21st Century Toxicology', Chem
Res Toxicol, 30: 43-52.
Panov, A., S. Dikalov, N. Shalbuyeva, G. Taylor, T. Sherer, and J. T. Greenamyre. 2005. 'Rotenone
model of Parkinson disease: multiple brain mitochondria dysfunctions after short term
systemic rotenone intoxication', J Biol Chem, 280: 42026-35.
Parasuraman, S. 2011. 'Toxicological screening', J Pharmacol Pharmacother, 2: 74-9.
Park, I. H., P. H. Lerou, R. Zhao, H. Huo, and G. Q. Daley. 2008. 'Generation of human-induced
pluripotent stem cells', Nat Protoc, 3: 1180-6.
Parker, W. D., Jr., J. K. Parks, and R. H. Swerdlow. 2008. 'Complex I deficiency in Parkinson's
disease frontal cortex', Brain Res, 1189: 215-8.
Pasca, A. M., S. A. Sloan, L. E. Clarke, Y. Tian, C. D. Makinson, N. Huber, C. H. Kim, J. Y. Park,
N. A. O'Rourke, K. D. Nguyen, S. J. Smith, J. R. Huguenard, D. H. Geschwind, B. A.
326
Barres, and S. P. Pasca. 2015. 'Functional cortical neurons and astrocytes from human
pluripotent stem cells in 3D culture', Nat Methods, 12: 671-8.
Peng, T. I., and M. J. Jou. 2004. 'Mitochondrial swelling and generation of reactive oxygen species
induced by photoirradiation are heterogeneously distributed', Ann N Y Acad Sci, 1011: 112-
22.
Pihlstrom, L., S. Wiethoff, and H. Houlden. 2017. 'Genetics of neurodegenerative diseases: an
overview', Handb Clin Neurol, 145: 309-23.
Pistollato, F., E. L. Ohayon, A. Lam, G. R. Langley, T. J. Novak, D. Pamies, G. Perry, E. Trushina,
R. S. Williams, A. E. Roher, T. Hartung, S. Harnad, N. Barnard, M. C. Morris, M. C. Lai,
R. Merkley, and P. C. Chandrasekera. 2016. 'Alzheimer disease research in the 21st
century: past and current failures, new perspectives and funding priorities', Oncotarget, 7:
38999-9016.
Qian, X., F. Jacob, M. M. Song, H. N. Nguyen, H. Song, and G. L. Ming. 2018. 'Generation of
human brain region-specific organoids using a miniaturized spinning bioreactor', Nat
Protoc, 13: 565-80.
Qian, X., H. N. Nguyen, M. M. Song, C. Hadiono, S. C. Ogden, C. Hammack, B. Yao, G. R.
Hamersky, F. Jacob, C. Zhong, K. J. Yoon, W. Jeang, L. Lin, Y. Li, J. Thakor, D. A. Berg,
C. Zhang, E. Kang, M. Chickering, D. Nauen, C. Y. Ho, Z. Wen, K. M. Christian, P. Y.
Shi, B. J. Maher, H. Wu, P. Jin, H. Tang, H. Song, and G. L. Ming. 2016. 'Brain-Region-
Specific Organoids Using Mini-bioreactors for Modeling ZIKV Exposure', Cell, 165: 1238-
54.
Radio, N. M., and W. R. Mundy. 2008. 'Developmental neurotoxicity testing in vitro: models for
assessing chemical effects on neurite outgrowth', Neurotoxicology, 29: 361-76.
Raff, M. C., K. L. Fields, S. I. Hakomori, R. Mirsky, R. M. Pruss, and J. Winter. 1979. 'Cell-type-
specific markers for distinguishing and studying neurons and the major classes of glial cells
in culture', Brain Res, 174: 283-308.
327
Rao, V. K., E. A. Carlson, and S. S. Yan. 2014. 'Mitochondrial permeability transition pore is a
potential drug target for neurodegeneration', Biochim Biophys Acta, 1842: 1267-72.
Reuhl, K. R. 1991. 'Delayed expression of neurotoxicity: the problem of silent damage',
Neurotoxicology, 12: 341-6.
Rizvi, Syed A. A., and Ayman M. Saleh. 2018. 'Applications of nanoparticle systems in drug
delivery technology', Saudi Pharmaceutical Journal, 26: 64-70.
Saggu, H., J. Cooksey, D. Dexter, F. R. Wells, A. Lees, P. Jenner, and C. D. Marsden. 1989. 'A
selective increase in particulate superoxide dismutase activity in parkinsonian substantia
nigra', J Neurochem, 53: 692-7.
Sakamuru, S., M. S. Attene-Ramos, and M. Xia. 2016. 'Mitochondrial Membrane Potential Assay',
Methods Mol Biol, 1473: 17-22.
Samsel, A., and S. Seneff. 2015. 'Glyphosate, pathways to modern diseases III: Manganese,
neurological diseases, and associated pathologies', Surg Neurol Int, 6: 45.
Saraiva, C., C. Praca, R. Ferreira, T. Santos, L. Ferreira, and L. Bernardino. 2016. 'Nanoparticle-
mediated brain drug delivery: Overcoming blood-brain barrier to treat neurodegenerative
diseases', J Control Release, 235: 34-47.
Sayes, C. M., and D. B. Warheit. 2009. 'Characterization of nanomaterials for toxicity assessment',
Wiley Interdiscip Rev Nanomed Nanobiotechnol, 1: 660-70.
Scelfo, B., M. Politi, F. Reniero, T. Palosaari, M. Whelan, and J. M. Zaldivar. 2012. 'Application of
multielectrode array (MEA) chips for the evaluation of mixtures neurotoxicity', Toxicology,
299: 172-83.
Schapira, A. H., J. M. Cooper, D. Dexter, P. Jenner, J. B. Clark, and C. D. Marsden. 1989.
'Mitochondrial complex I deficiency in Parkinson's disease', Lancet, 1: 1269.
Schildknecht, S., C. Karreman, D. Poltl, L. Efremova, C. Kullmann, S. Gutbier, A. Krug, D.
Scholz, H. R. Gerding, and M. Leist. 2013. 'Generation of genetically-modified human
328
differentiated cells for toxicological tests and the study of neurodegenerative diseases',
ALTEX, 30: 427-44.
Schildknecht, S., D. Poltl, D. M. Nagel, F. Matt, D. Scholz, J. Lotharius, N. Schmieg, A. Salvo-
Vargas, and M. Leist. 2009. 'Requirement of a dopaminergic neuronal phenotype for
toxicity of low concentrations of 1-methyl-4-phenylpyridinium to human cells', Toxicol
Appl Pharmacol, 241: 23-35.
Schmidt, B. Z., M. Lehmann, S. Gutbier, E. Nembo, S. Noel, L. Smirnova, A. Forsby, J. Hescheler,
H. X. Avci, T. Hartung, M. Leist, J. Kobolak, and A. Dinnyes. 2017. 'In vitro acute and
developmental neurotoxicity screening: an overview of cellular platforms and high-
throughput technical possibilities', Arch Toxicol, 91: 1-33.
Scholz, D., D. Poltl, A. Genewsky, M. Weng, T. Waldmann, S. Schildknecht, and M. Leist. 2011.
'Rapid, complete and large-scale generation of post-mitotic neurons from the human
LUHMES cell line', J Neurochem, 119: 957-71.
Scripture, C. D., W. D. Figg, and A. Sparreboom. 2006. 'Peripheral neuropathy induced by
paclitaxel: recent insights and future perspectives', Curr Neuropharmacol, 4: 165-72.
Seibert, H., S. Morchel, and M. Gulden. 2002. 'Factors influencing nominal effective
concentrations of chemical compounds in vitro: medium protein concentration', Toxicol In
vitro, 16: 289-97.
Sherer, T. B., R. Betarbet, and J. T. Greenamyre. 2002. 'Environment, mitochondria, and
Parkinson's disease', Neuroscientist, 8: 192-7.
Sherer, T. B., R. Betarbet, C. M. Testa, B. B. Seo, J. R. Richardson, J. H. Kim, G. W. Miller, T.
Yagi, A. Matsuno-Yagi, and J. T. Greenamyre. 2003. 'Mechanism of toxicity in rotenone
models of Parkinson's disease', J Neurosci, 23: 10756-64.
Sherer, T. B., J. R. Richardson, C. M. Testa, B. B. Seo, A. V. Panov, T. Yagi, A. Matsuno-Yagi, G.
W. Miller, and J. T. Greenamyre. 2007. 'Mechanism of toxicity of pesticides acting at
329
complex I: relevance to environmental etiologies of Parkinson's disease', J Neurochem,
100: 1469-79.
Sherer, T. B., P. A. Trimmer, K. Borland, J. K. Parks, J. P. Bennett, Jr., and J. B. Tuttle. 2001.
'Chronic reduction in complex I function alters calcium signaling in SH-SY5Y
neuroblastoma cells', Brain Res, 891: 94-105.
Sherman, S. P., and A. G. Bang. 2018. 'High-throughput screen for compounds that modulate
neurite growth of human induced pluripotent stem cell-derived neurons', Dis Model Mech,
11.
Sian-Hulsmann, J., S. Mandel, M. B. Youdim, and P. Riederer. 2011. 'The relevance of iron in the
pathogenesis of Parkinson's disease', J Neurochem, 118: 939-57.
Sidransky, E., and G. Lopez. 2012. 'The link between the GBA gene and parkinsonism', Lancet
Neurol, 11: 986-98.
Singleton, A. B., M. Farrer, J. Johnson, A. Singleton, S. Hague, J. Kachergus, M. Hulihan, T.
Peuralinna, A. Dutra, R. Nussbaum, S. Lincoln, A. Crawley, M. Hanson, D. Maraganore,
C. Adler, M. R. Cookson, M. Muenter, M. Baptista, D. Miller, J. Blancato, J. Hardy, and K.
Gwinn-Hardy. 2003. 'alpha-Synuclein locus triplication causes Parkinson's disease',
Science, 302: 841.
Sirenko, O., J. Hesley, I. Rusyn, and E. F. Cromwell. 2014. 'High-content high-throughput assays
for characterizing the viability and morphology of human iPSC-derived neuronal cultures',
Assay Drug Dev Technol, 12: 536-47.
Smirnova, L., G. Harris, J. Delp, M. Valadares, D. Pamies, H. T. Hogberg, T. Waldmann, M. Leist,
and T. Hartung. 2016. 'A LUHMES 3D dopaminergic neuronal model for neurotoxicity
testing allowing long-term exposure and cellular resilience analysis', Arch Toxicol, 90:
2725-43.
Smirnova, L., G. Harris, M. Leist, and T. Hartung. 2015. 'Cellular resilience', ALTEX, 32: 247-60.
330
Smirnova, L., Hogberg H.T., Leist M., and and Hartung T. 2014. 'Developmental neurotoxicity –
challenges in the 21st century and in vitro opportunities. ', ALTEX: 129-56.
Snyder, J. S. 2018. 'Questioning human neurogenesis', Nature, 555: 315-16.
Snyder, J. S., S. C. Ferrante, and H. A. Cameron. 2012. 'Late maturation of adult-born neurons in
the temporal dentate gyrus', PLoS One, 7: e48757.
Sorrells, Shawn F., Mercedes F. Paredes, Arantxa Cebrian-Silla, Kadellyn Sandoval, Dashi Qi,
Kevin W. Kelley, David James, Simone Mayer, Julia Chang, Kurtis I. Auguste, Edward F.
Chang, Antonio J. Gutierrez, Arnold R. Kriegstein, Gary W. Mathern, Michael C. Oldham,
Eric J. Huang, Jose Manuel Garcia-Verdugo, Zhengang Yang, and Arturo Alvarez-Buylla.
2018. 'Human hippocampal neurogenesis drops sharply in children to undetectable levels in
adults', Nature, 555: 377.
Spalding, K. L., O. Bergmann, K. Alkass, S. Bernard, M. Salehpour, H. B. Huttner, E. Bostrom, I.
Westerlund, C. Vial, B. A. Buchholz, G. Possnert, D. C. Mash, H. Druid, and J. Frisen.
2013. 'Dynamics of hippocampal neurogenesis in adult humans', Cell, 153: 1219-27.
Spillantini, M. G., M. L. Schmidt, V. M. Lee, J. Q. Trojanowski, R. Jakes, and M. Goedert. 1997.
'Alpha-synuclein in Lewy bodies', Nature, 388: 839-40.
Stefanis, L. 2012. 'alpha-Synuclein in Parkinson's disease', Cold Spring Harb Perspect Med, 2:
a009399.
Subramanian, K., D. J. Owens, R. Raju, M. Firpo, T. D. O'Brien, C. M. Verfaillie, and W. S. Hu.
2014. 'Spheroid culture for enhanced differentiation of human embryonic stem cells to
hepatocyte-like cells', Stem Cells Dev, 23: 124-31.
Takahashi, K., K. Tanabe, M. Ohnuki, M. Narita, T. Ichisaka, K. Tomoda, and S. Yamanaka. 2007.
'Induction of pluripotent stem cells from adult human fibroblasts by defined factors', Cell,
131: 861-72.
Takahashi, K., and S. Yamanaka. 2006. 'Induction of pluripotent stem cells from mouse embryonic
and adult fibroblast cultures by defined factors', Cell, 126: 663-76.
331
Takalo, M., A. Salminen, H. Soininen, M. Hiltunen, and A. Haapasalo. 2013. 'Protein aggregation
and degradation mechanisms in neurodegenerative diseases', Am J Neurodegener Dis, 2: 1-
14.
Tang, H. L., H. M. Tang, K. H. Mak, S. Hu, S. S. Wang, K. M. Wong, C. S. Wong, H. Y. Wu, H.
T. Law, K. Liu, C. C. Talbot, Jr., W. K. Lau, D. J. Montell, and M. C. Fung. 2012. 'Cell
survival, DNA damage, and oncogenic transformation after a transient and reversible
apoptotic response', Mol Biol Cell, 23: 2240-52.
Tanner, C. M. 2003. 'Is the cause of Parkinson's disease environmental or hereditary? Evidence
from twin studies', Adv Neurol, 91: 133-42.
Tanner, C. M., F. Kamel, G. W. Ross, J. A. Hoppin, S. M. Goldman, M. Korell, C. Marras, G. S.
Bhudhikanok, M. Kasten, A. R. Chade, K. Comyns, M. B. Richards, C. Meng, B. Priestley,
H. H. Fernandez, F. Cambi, D. M. Umbach, A. Blair, D. P. Sandler, and J. W. Langston.
2011. 'Rotenone, paraquat, and Parkinson's disease', Environ Health Perspect, 119: 866-72.
Tanner, C. M., R. Ottman, S. M. Goldman, J. Ellenberg, P. Chan, R. Mayeux, and J. W. Langston.
1999. 'Parkinson disease in twins: an etiologic study', JAMA, 281: 341-6.
Terrasso, A. P., C. Pinto, M. Serra, A. Filipe, S. Almeida, A. L. Ferreira, P. Pedroso, C. Brito, and
P. M. Alves. 2015. 'Novel scalable 3D cell based model for in vitro neurotoxicity testing:
Combining human differentiated neurospheres with gene expression and functional
endpoints', J Biotechnol, 205: 82-92.
Tokuzawa, Y., E. Kaiho, M. Maruyama, K. Takahashi, K. Mitsui, M. Maeda, H. Niwa, and S.
Yamanaka. 2003. 'Fbx15 is a novel target of Oct3/4 but is dispensable for embryonic stem
cell self-renewal and mouse development', Mol Cell Biol, 23: 2699-708.
Tollefsen, K. E., S. Scholz, M. T. Cronin, S. W. Edwards, J. de Knecht, K. Crofton, N. Garcia-
Reyero, T. Hartung, A. Worth, and G. Patlewicz. 2014. 'Applying Adverse Outcome
Pathways (AOPs) to support Integrated Approaches to Testing and Assessment (IATA)',
Regul Toxicol Pharmacol, 70: 629-40.
332
Tong, Z. B., H. Hogberg, D. Kuo, S. Sakamuru, M. Xia, L. Smirnova, T. Hartung, and D. Gerhold.
2017. 'Characterization of three human cell line models for high-throughput neuronal
cytotoxicity screening', J Appl Toxicol, 37: 167-80.
Tsuji, S. 2010. 'Genetics of neurodegenerative diseases: insights from high-throughput
resequencing', Hum Mol Genet, 19: R65-70.
Tyagi, E., Y. Zhuang, R. Agrawal, Z. Ying, and F. Gomez-Pinilla. 2015. 'Interactive actions of
Bdnf methylation and cell metabolism for building neural resilience under the influence of
diet', Neurobiol Dis, 73: 307-18.
Uversky, V. N. 2004. 'Neurotoxicant-induced animal models of Parkinson's disease: understanding
the role of rotenone, maneb and paraquat in neurodegeneration', Cell Tissue Res, 318: 225-
41.
Valente, E. M., P. M. Abou-Sleiman, V. Caputo, M. M. Muqit, K. Harvey, S. Gispert, Z. Ali, D.
Del Turco, A. R. Bentivoglio, D. G. Healy, A. Albanese, R. Nussbaum, R. Gonzalez-
Maldonado, T. Deller, S. Salvi, P. Cortelli, W. P. Gilks, D. S. Latchman, R. J. Harvey, B.
Dallapiccola, G. Auburger, and N. W. Wood. 2004. 'Hereditary early-onset Parkinson's
disease caused by mutations in PINK1', Science, 304: 1158-60.
Van Bruggen, A. H. C., M. M. He, K. Shin, V. Mai, K. C. Jeong, M. R. Finckh, and J. G. Morris,
Jr. 2018. 'Environmental and health effects of the herbicide glyphosate', Sci Total Environ,
616-617: 255-68.
Van Maele-Fabry, G., P. Hoet, F. Vilain, and D. Lison. 2012. 'Occupational exposure to pesticides
and Parkinson's disease: a systematic review and meta-analysis of cohort studies', Environ
Int, 46: 30-43. van Vliet, E., L. Stoppini, M. Balestrino, C. Eskes, C. Griesinger, T. Sobanski, M. Whelan, T.
Hartung, and S. Coecke. 2007. 'Electrophysiological recording of re-aggregating brain cell
cultures on multi-electrode arrays to detect acute neurotoxic effects', Neurotoxicology, 28:
1136-46.
333
Villeneuve, D. L., D. Crump, N. Garcia-Reyero, M. Hecker, T. H. Hutchinson, C. A. LaLone, B.
Landesmann, T. Lettieri, S. Munn, M. Nepelska, M. A. Ottinger, L. Vergauwen, and M.
Whelan. 2014. 'Adverse outcome pathway (AOP) development I: strategies and principles',
Toxicol Sci, 142: 312-20.
Volbracht C, van Beek J, Zhu C. 2006. Neuroprotective properties of memantine in different in
vitro and in vivo models of excitotoxicity. Eur J Neurosci. 23:2611–2622. doi:
10.1111/j.1460-9568.2006.04787.x.
Waldmann, Tanja, Marianna Grinberg, André König, Eugen Rempel, Stefan Schildknecht, Margit
Henry, Anna-Katharina Holzer, Nadine Dreser, Vaibhav Shinde, Agapios Sachinidis, Jörg
Rahnenführer, Jan G. Hengstler, and Marcel Leist. 2017. 'Stem Cell Transcriptome
Responses and Corresponding Biomarkers That Indicate the Transition from Adaptive
Responses to Cytotoxicity', Chemical Research in Toxicology, 30: 905-22.
Wang, Q., Y. Liu, and J. Zhou. 2015. 'Neuroinflammation in Parkinson's disease and its potential as
therapeutic target', Transl Neurodegener, 4: 19.
Warren, L., P. D. Manos, T. Ahfeldt, Y. H. Loh, H. Li, F. Lau, W. Ebina, P. K. Mandal, Z. D.
Smith, A. Meissner, G. Q. Daley, A. S. Brack, J. J. Collins, C. Cowan, T. M. Schlaeger,
and D. J. Rossi. 2010. 'Highly efficient reprogramming to pluripotency and directed
differentiation of human cells with synthetic modified mRNA', Cell Stem Cell, 7: 618-30.
Wiemerslage, L., and D. Lee. 2016. 'Quantification of mitochondrial morphology in neurites of
dopaminergic neurons using multiple parameters', J Neurosci Methods, 262: 56-65.
Wiesner, M. R., G. V. Lowry, E. Casman, P. M. Bertsch, C. W. Matson, R. T. Di Giulio, J. Liu, and
M. F. Hochella, Jr. 2011. 'Meditations on the ubiquity and mutability of nano-sized
materials in the environment', ACS Nano, 5: 8466-70.
Wilson, M. S., J. R. Graham, and A. J. Ball. 2014. 'Multiparametric High Content Analysis for
assessment of neurotoxicity in differentiated neuronal cell lines and human embryonic stem
cell-derived neurons', Neurotoxicology, 42: 33-48.
334
Wirdefeldt, K., M. Gatz, M. Schalling, and N. L. Pedersen. 2004. 'No evidence for heritability of
Parkinson disease in Swedish twins', Neurology, 63: 305-11.
Xicoy, H., B. Wieringa, and G. J. Martens. 2017. 'The SH-SY5Y cell line in Parkinson's disease
research: a systematic review', Mol Neurodegener, 12: 10.
Yacoubian, T. A., S. R. Slone, A. J. Harrington, S. Hamamichi, J. M. Schieltz, K. A. Caldwell, G.
A. Caldwell, and D. G. Standaert. 2010. 'Differential neuroprotective effects of 14-3-3
proteins in models of Parkinson's disease', Cell Death Dis, 1: e2.
Yarjanli, Z., K. Ghaedi, A. Esmaeili, S. Rahgozar, and A. Zarrabi. 2017. 'Iron oxide nanoparticles
may damage to the neural tissue through iron accumulation, oxidative stress, and protein
aggregation', BMC Neurosci, 18: 51.
Yealland, G., G. Battaglia, O. Bandmann, and H. Mortiboys. 2016. 'Rescue of mitochondrial
function in parkin-mutant fibroblasts using drug loaded PMPC-PDPA polymersomes and
tubular polymersomes', Neurosci Lett, 630: 23-29.
Youdim, M. B., D. Edmondson, and K. F. Tipton. 2006. 'The therapeutic potential of monoamine
oxidase inhibitors', Nat Rev Neurosci, 7: 295-309.
Zhang, C. W., L. Hang, T. P. Yao, and K. L. Lim. 2015. 'Parkin Regulation and Neurodegenerative
Disorders', Front Aging Neurosci, 7: 248.
Zhang, L., V. L. Dawson, and T. M. Dawson. 2006. 'Role of nitric oxide in Parkinson's disease',
Pharmacol Ther, 109: 33-41.
Zhang, X., M. G. Hu, K. Pan, C. H. Li, and R. Liu. 2016. '3D Spheroid Culture Enhances the
Expression of Antifibrotic Factors in Human Adipose-Derived MSCs and Improves Their
Therapeutic Effects on Hepatic Fibrosis', Stem Cells Int, 2016: 4626073.
Zhang, X. M., M. Yin, and M. H. Zhang. 2014. 'Cell-based assays for Parkinson's disease using
differentiated human LUHMES cells', Acta Pharmacol Sin, 35: 945-56.
335
Zhou, H., S. Wu, J. Y. Joo, S. Zhu, D. W. Han, T. Lin, S. Trauger, G. Bien, S. Yao, Y. Zhu, G.
Siuzdak, H. R. Scholer, L. Duan, and S. Ding. 2009. 'Generation of induced pluripotent
stem cells using recombinant proteins', Cell Stem Cell, 4: 381-4.
336
10. APPENDICES
10.1. Appendix I Supplementary Figures Chapter 4
337
338
339
10.2. Appendix II. Supplementary Figures Chapter 5.
Supplementary Methods
3D LUHMES Differentiation
Cell Culture and Differentiation: The cell culture protocol was followed as described by (Krug et al. 2014 and Scholz et al., 2011). Flasks were coated with a coating solution. Coating solution was incubated over night at 37°C. Before flasks were used for cell culture, they were washed twice with sterile distilled water. LUHMES were cultured in a 75 cm² pre-coated flask, containing 12 mL proliferation medium. The Flask was incubated at 37 °C, 5% C02 and 95% humidity. When cells reached 70-80% confluency after 2-3 days, they were passaged. Medium was aspirated and 1 mL TryplETM was added for 3 min at 37 °C. After detachment LUHMES were re-suspended in DMEM/F12 Medium (15 mL). After transferring the cell suspension to a falcon tube, they were centrifuged for 3 min at 1000 rpm. The trypsin containing medium was aspirated and the pellet was suspend in 10 mL of proliferation medium. Cells were counted using a cell counter and trypan blue. 2x106 cells were transferred into a 75 cm² flask, or 4x106 in a 175 cm² flask. At Day 0, 3D-differentiation was initiated as described by Harris et al., 2017 and
Smirnova et al., 2016. Cells were detached and counted as described above. 5.5x105 – 6x105 cells were seeded to each well of a 6-well plate. Each well contained 2 mL of differentiation medium. It was necessary to stay in the given range of cell density for differentiation to prevent increased proliferation of LUHMES during later steps of differentiation. The 6-well plates were kept on a shaker at 80 rpm in an incubator at 37 °C, 10% CO2 and 95% humidity. Due to increased cell—cell interaction in 3D aggregates compared to LUHMES cultured in monolayers, proliferation is partly increased (this observation has also been made in confluent 2D cultures).
Therefore, after 3 Days of differentiation Taxol (the anti-proliferative compound Paclitaxel) was
340 added to the 6-well plates to prevent proliferation within aggregates (Smirnova et al. 2016). A low-dose of this compound was added in this protocol to inhibit further proliferation and ensure differentiation. Anti-Proliferation Media is composed of differentiation medium with a Taxol concentration of 20 nM. To change media/ add anti-proliferation media, the plates were shaken in slow circle motions to allow aggregates accumulate in the middle. The plate was tilted, 800 µL of the medium was removed and 1 mL of anti-proliferation medium was added. The volume of removed medium varied slightly to account for evaporation. Finally the 6-well plates were placed again in the incubator at 37 °C, 10% CO2 and 95% humidity. On Day 5 of differentiation, Taxol was washed out. The maximum of volume (leaving ca. 200 μL) was removed from each well and
2 mL of fresh, pre-warmed Wash Media was added to the wells. This step was repeated one more time adding Differentiation Media. For medium exchange (on days 8, 10 and 12) medium was exchanged by removing 1 mL and adding of 1.2 mL differentiation medium. All media preparations can be found in Harris et al. 2017.
Treatment and wash-out
Treatment in 6-well plate (Day 7 treatment): Rotenone stock (100 mM in DMSO) was defrosted on the day of treatment (day 7). Prior to treatment, the 100 mM rotenone stock was diluted
1:10,000 in differentiation media and vortexed for 15 seconds. DMSO (100 %) was prepared in the same way, diluted 1:10,000 in differentiation media (0.1 % DMSO) and vortexed for 15 seconds. To treat 3D LUHMES cells, 20 µL of the 10 μM rotenone solution was added in the 6- well plate containing 2 mL medium per well, to reach the desired final concentration (100 nM).
The same volume of the 0.01 % DMSO solution was added for controls (non-cytotoxic). 6-well pates were placed on the gyratory shaker in an incubator with same settings as previously described. LUHMES were incubated with the rotenone or DMSO for 24h.
341
Treatment in 24-well plate (Day 15 dose-response): Rotenone stock (100 mM in DMSO) was defrosted on the day of treatment (day 14). The stock was diluted to 100 µM (1:1000). Then, a 2x serial dilution of the rotenone stock was prepared (63.2 nM, 200 nM, 632 nM, 2µM, 20 µM) in differentiation medium. A DMSO control was prepared in the same way as the highest rotenone concentration (final 0.002 % DMSO). Differentiated aggregates were retrieved from the incubator, shaken in circle motions and collected into a 15 mL tube. After allowing the aggregates to sink to the bottom of the tube, medium was aspirated and aggregates were suspend in 6 mL fresh differentiation medium. 250 µL of the aggregate-suspension was added to each well of the 24-well plate. Then, 250 mL of the 2x serial dilution was added to wells in triplicates.
The 24-well plate was placed in an incubator at 37 °C, 10% CO2 and 95% humidity for 24 h. The final dose-response concentration was 0, 31.6 nM, 100 nM, 316 nM, 1 µM and 10 µM.
Compound wash-out: After 24h of exposure in 6-well plates (day 8), the compound was washed out. Plates were taken from the incubator and shaken in small circle motions to allow aggregates to collect in the center of the plate. The plate was tilted and almost all medium was aspirated
(leaving 200 μL). 2 mL wash medium (Supplementary Table S2) was added to each well. Then 2 mL differentiation medium were added to each well of a newly labelled 6-well plate (transfer to a new plate is required as rotenone can adhere to plastic and could slowly release into new media).
Aggregates in wash medium were then transferred to the new 6-well plates containing 2mL differentiation medium using a 200 µL pipette (as little media as possible was transferred). These steps were performed quickly to avoid aggregate clumping during washing. New plates containing washed aggregates were then returned to the incubator on the gyratory shaker.
Viability
342
Resazurin Assay: viability was measured as described in Harris et al., 2017. Fluorescence emission at 580-610 nm was measured and %-cell viability was calculated by comparing relative fluorescence units (RFU) of treated vs. control samples after subtracting blank values.
LDH assay: LDH was measured in the media in control and treated samples following manufacturer’s instructions (CytoTox 96® NonRadioactive Cytotoxicity Assay, Promega).
Briefly, 1 well was lysed as a positive control and 50 uL media from each well were transferred into a 96-well plate. 50 uL substrate were added and the plate was incubated at RT for 30 min.
Then, 50 uL stop solution were added to stop the reaction and the absorbance was measured at
490 nm in a spectrophotometer. After blank subtraction, cytotoxicity was calculated by the following formula:
Percent cytotoxicity = 100*(Experimental LDH (OD490) /Maximum LDH (OD490))
DNA extraction and quantification
Collected cells were resuspended in 1x Tris-EDTA pH 8.0 (Quality Biological), 0.3 % (v/v) SDS
(10%, Quality Biological), and Proteinase K (1mg/ml, Invitrogen). Cells were lysed overnight, rotating at 65 °C. Lysed cells were treated with 10 µg of RNAse A (Thermo Scientific) for 30 min at 37 °C. Following digestion, cells were transferred to heavy phase lock gel (Quantabio) and an equal volume of Phenol: Chloroform: Isoamyl (24:25:1, Sigma) was added to the gel and mixed with inversion. Cells were centrifuged at 12,000 x g for 15 min and the aqueous phase separated. DNA was precipitated using Ethanol (100%, Pharamco-AAPER), 2 % (v/v) Sodium
Acetate (3M, Mediatech), and 0.3 % (v/v) Linear Acrylamide (Ambion). DNA was pelleted by centrifugation at 4,000 x g for 15 min, washed with 70% Ethanol, and resuspended in 10 mM
Tris-Cl pH 8.0 (Qiagen). Eluted DNA was incubated at 37 °C for at least one hour prior to quantification. DNA quantification was performed using the Qubit dsDNA Broad Range Assay
Kit (Invitrogen) and Qubit 2.0 Fluorometer (Invitrogen) according to manufacturer’s instructions.
343
RNA extraction, reverse transcription, and real-time PCR
Total RNA was extracted using either TRIzol® Reagent (Life Technologies) and RNA Clean &
ConcentratorTM- Kit (Zymo Research) or Mirvana microRNA isolation kit (for microarray analysis) following the manufacturer’s instructions. RNA integrity was measured using the
Nanodrop 2000 (ThermoScientific) UV–Vis Spectrophotometer (260 nm). Equal amounts of purified RNA (500 ng) were reverse transcribed to cDNA using random hexamer primers
(Promega) and M-MLV reverse transcriptase Kit (Promega) following the manufacturer’s instructions. A DNAse treatment step was included in cDNA synthesis to ensure the elimination of DNA traces. The cDNA was diluted 1:5, and qRT-PCR was performed. Expression of genes after LUHMES exposure to compounds was analysed using TaqMan gene expression assay (Life
Technologies) and TaqMan FAST advance Master Mix (Life Technologies) according to the manufacturer’s protocols. Expression of genes perturbed by toxicant treatment was analysed using Fast SYBR Green master mix (Life Technologies) and primers listed in Supplementary
Table S3. 18S and GAPDH were used as housekeeping genes for TaqMan gene expression and
SYBR Green PCRs, respectively. All RT-PCRs were performed in duplicates on Fast Applied
Biosystems 7500 System (Life Technologies) with the following thermal cycling parameters:
SYBR® Green RT-PCR (95 °C for 20 s, followed by 40 cycles of 3 s at 95 °C and 30 s 60 °C); a melting curve step was included in SYBR Green reactions (95 °C for 15 s, 60 °C for 1 min, 95 °C for 15 s, and 60 °C for 15 s); TaqMan gene expression assay (95 °C for 20 s, followed by 40 cycles of 3 s at 95 °C and 30 s 60 °C). 2-∆∆Ct method was used to calculate the fold changes
(REF: Schmittgen et al 2008, doi:10.1038/nprot.2008.73) Data collected from three independent experiments were calculated as average log2-fold change in independent biological replicates ±
SEM. Differences in treated and control samples were analyzsed for statistical significance using an unpaired t-test and Bonferroni correction. A p value <0.05 is denoted in graph by *, p < 0.01 by **, and p < 0.001 by ***, respectively.
344
Complex I Activity Assay
Mitochondria Isolation was performed on ice using the reagent-based method (Mitochondria
Isolation Kit for Tissue and Cultured Cells, BioVision); Complex I activity using mitochondrial
Complex I Activity Colorimetric Assay Kit, BioVision following manufacturer’s instructions.
Briefly, cells were collected by pelleting 3 wells from a 6-well plate, centrifuging at 600 x g for
10 min at 4°C and discarding the supernatant. 500 µL of mitochondria isolation buffer (with protease inhibitor cocktail (Sigma-Aldrich)) was then added to pellet, vortexed and placed on ice for 2 min. Then, 5 µL of Reagent A was added, vortexed for 5 sec and incubated on ice for 5 min, vortexing every min for 5 sec. The dissociated cells were then centrifuged at 7000 x g for 10 min at 4°C. The supernatant was discarded and the pellet was washed with mitochondria isolation buffer. The supernatant was carefully removed and mitochondria were re-suspended in storage buffer. The protein concentration was measured using BCA protein assay (1:20 sample to working reagent ratio) using a Nanodrop2000 spectrophotometer. Sample concentration was adjusted to 1 µg/µL using storage buffer. Sample protein concentration > 0.5 µg/µL is needed for complex I activity measurements. Complex I activity was measured following manufacturer’s instructions (mitochondrial Complex I Activity Colorimetric Assay Kit, BioVision). For Complex
I Activity assayBriefly, the standard curve was measured by diluting the complex I dye stock solution in complex I assay buffer. Absorbance at 600 nm was measured. The reaction mix was them prepared and added to each well as described by the manufacturer. 2 µL (4 µg) mitochondrial samples were added to wells containing “sample mix” and “sample + inhibitor mix”. NADH 1x working solution was prepared and 30 µL were added to each well using a multichannel pipette to avoid any differences in starting time for the reactions in the wells.
Immediately, the plate was read at 600 nm on kinetic mode at 30 second intervals for 5 min at
RT. The kinetic data collected was then analyzed using the following equation.
345
Reduced Complex Dye = Total Complex Dye (9 nmol/well) - Oxidized complex I Dye (as read from standard curve)
Sample Complex I Activity = ∆[reduced complex Dye concentration] / (∆t x p) (mUnits / µg)
∆[reduced complex Dye concentration] = change in reduced Complex I Due concentration during
∆t
∆t = t2 – t1 (min) p = mitochondrial protein (µg)
Cellular ATP assay
ATP quantification was performed on Day 8 (after toxicant treatment) and Day 15 (after wash- out) of differentiation. The bioluminescence ATP Assay Kit (Thermofisher, A22066) was used to determine the amount of intracellular ATP in aggregates according to manufacturer’s instructions.
The ATP assay gives total cellular ATP content and informs about cellular respiration. Firstly, one well of aggregates for each condition was collected into an Eppendorf tube and washed once with cold PBS. All steps were performed on ice. Aggregates were allowed to sink to the bottom of the tube and all PBS was removed. 50 µL of whole cell lysis buffer (0.3 g NaCl, 1 mL Tris
(1M), 1 mL 10% NP-40, 0.2 mL EDTA (0.5M), 17.8 mL dH2O) was added to each tube and aggregates were pipetted up and down. Tubes were placed on ice for 20-30 min to allow for complete lysis. Samples were mixed to ensure lysis and centrifuged at 10,000 g for 5 min. The assay buffer was prepared following the manufacturer’s instructions. Assay buffer (100 µL) was added to each well of a 96-well plate and 10 µL was added for each sample as well as positive (1,
5 and 10 µM ATP) and negative (lysis buffer) controls, in triplicate. The plate was kept in the dark at room temperature for 15 min. Luminescence was read using GloMax® 96 Microplate
Luminometer (Promega). To normalize ATP measurements to cell number, total protein
346 concentration was used. Therefore Pierce BCA Protein Assay Kit was used, as described in manufacturer’s instructions (https://tools.thermofisher.com/content/sfs/manuals/
MAN0011430_Pierce_BCA_Protein_Asy_UG.pdf). BCA Protein working reagent (1:50) was prepared. 200 µL were transferred into a 96-well plate (samples and blank in duplicates). For each sample 10 µL of cell lysis were added in duplicate as well as blank wells containing whole cell lysis buffer. 10 µL bovine serum albumin (BSA) standard protein concentrations were added to wells containing 200 µL BCA working reagent. Concentrations for the standard curve were 0,
0.25, 0.5, 0.75, 1, 1.5, 1.75 and 2 mg/ml BSA. The plate was incubated for 30 min at 37 °C.
Absorbance was measured at 570 nm using a spectrophotometer. Average luminescence values ±
SEM was calculated from at least four biological replicates and technical duplicates. Differences in treated and control samples were analyzed for statistical significance using unpaired Student’s
T test. A p value<0.05 is denoted on graphs by *, p<0.0001 by ***, respectively.
Neurite outgrowth imaging and Sholl analysis
RFP-LUHMES (Schildknecht et al. 2013) were differentiated and treated as described previously.
Aggregates were collected into an Eppendorf tube and washed once using Wash Medium
(Supplementary Table 1) on day 8 or day 15. Wash Medium was removed and replaced with
Differentiation Medium (Supplementary Table 1). Differentiation Medium (100 µL) was added to each well of MatrigelTM (BD Biosciences) pre-coated, flat-bottom, black 24 or 96-well plates
(ThermoScientific). 5-10 aggregates were seeded into each well in triplicates (ThermoScientific) ensuring they were well spread out within the well for better image quantification. Plates were incubated for 24h at 37 °C, 5% CO2. After 24 h, media was removed and aggregates were washed twice with warm PBS. Aggregates were then fixed with 24% PFA, 1:10,000 Hoechst
33342 (Invitrogen, Molecular Probes) for 1 hour and washed three times with PBS. Confocal images of 5 aggregates per sample in three independent experiments were obtained for neurite
347 outgrowth analysis using Sholl Image J Software. Images were obtained using the LSM Zeiss
Confocor2000. 16-bit Images were obtained with 20x objective for blue (Hoechst 42333) and red
(RFP-LUHMES) channels. Images were opened using Image J (Fiji Image J Open Source
Software), converted to B&W and the threshold adjusted to allow software to identify individual neurites (the same threshold was kept for all images in each individual experiment). A straight line was drawn from the center of the aggregate to the most distal point of outgrowth. Then neurites were counted in shells starting at the surface of the aggregate. allowed the software to determine the start and end for multiple shells (radii). The software then automatically detected the number of neurite intersections per radius (set at 10 pixel widths) and the number of intersections (Y) vs distance from aggregate (X) was recorded, plotted and analyzed.
348
Supplementary Table S1 PCR primer sequences used for SYBRGreen RT-PCR
GENE Forward sequence, 5’-3’ Reverse sequence, 3’-5’
ATF4 GGCTGGCTGTGGATGGGTTG CTCCTGGACTAGGGGGGCAA
GAPDH CACCATCTTCCAGGAGCGAGATC GCAGGAGGCATTGCTGATGATC
CASP3 TGGTTTTCGGTGGGTGTG CCACTGAGTTTTCAGTGTTCTC
TYMS CAGCTTCAGCGAGAACCCAG ACCTCGGCATCCAGCCCAAC
MLF1IP TTTGTAAGGCAGCCATCGCC CTGTGGCTCTAACCGAAGCA
349
Supplemental Figures
Suplimental Figure S1. DNA quantification and estimation of number of cells in one aggregate on day 15 of differentiation (Total DNA in sample ÷ 6 pg DNA per cell)
350
Supplemental Figure S2. Neurite Outgrowth analysis using Image J Sholl Analysis Software. (a)
Sample images taken using a confocal microscope. (b) Analysis performed counting the number of intersections at each ring every 10 µm from the aggregate center. The number of intersections is marked from high (red) to low (blue). The data obtained gave number of intersections over distance. Red and blue indicate high and low number of intersections, respectively
351
Supplemental Figure S3. 50 nM rotenone-induced transcriptome changes on day 8 (12 or 24h) vs. day 15 (wash-out). Volcano plots show significantly lower number of perturbed genes after compound wash-out and 7-day recovery period on day 15 vs. effects on day 8, after 12 h (a) or 24 h (b) exposure to 50 nM rotenone. (c) Venn-diagrams show the number of up- and down- regulated genes on day 8 (D8 (12 or 24h)) and on day 15 (D15 (wash-out)) (FC > 1.5, p < 0.01).
One gene (FGFR3) was in intersection between two conditions in samples, treated with 50 nM for
24h
352
Supplemental Figure S4. All genes were visualized in the STRING interaction database for connections; all genes that are a “first neighbor” of (a) DNA topoisomerase II beta (TOP2B, the gene with the highest connection on day 8), (b) Cholecystokinin (CCK) on day 8 and (c) CCK and Guanine nucleotide-binding protein 1 (GNB1) subnetworks on day15 are shown. Both
353 subnetwork (b) and (c) were highly enriched for G-protein coupled receptor (GPCR, yellow).
Genes with a degree higher than 20 (i.e. hubs) are indicated in magenta; genes in common between day 8 and day 15 are indicated with circles
Supplemental Figure S5. Cell viability concentration-response for aggregates on day 15, pre- exposed to DMSO (Control) or rotenone (pre-exposed 25 nM) on day 8
354
11. CURRICULUM VITAE Georgina Harris (MSc, BSc)
Current Address: 615 N Wolfe St. W7032, Johns Hopkins School of Public Health, Baltimore,
MD, USA
ResearchGate Profile: https://www.researchgate.net/profile/Georgina_Harris2
Personal E-mail: [email protected] Contact number: +1
410-340-6538
CURRENT PhD Candidate in Environmental Health Sciences - Molecular and
PROFILE Translational Toxicology, Bloomberg School of Public Health, Johns Hopkins
University. Having gained much experience through my MSc and working with
various research groups prior, I have defined my career path in molecular biology
and applied toxicological research. My interests are currently focused on
developing and optimising endpoints using human-relevant 3D in vitro models for
toxicity testing as well as studying key molecular events which determine toxicity,
recovery and resilience in an in vitro Parkinson’s disease model.
EDUCATION
2018 PhD Student, Environmental Health Sciences, Johns Hopkins Bloomberg School
of Public Health, Baltimore, USA.
2010 MSc Toxicology, The University of Birmingham, Birmingham, United Kingdom
2009 BSc Biochemistry with molecular cell biology (Bachelor of Science with
Honours). The University of Birmingham, United Kingdom.
MSc Toxicology Class Representative, 2010
355
PROFESIONAL Home Office Accredited Training Course Modules 1-3 Animal Handling
DEVELOPMENT Certificate, 2010
GLP and Laboratory Safety Training, Institute for Health and Consumer
Protection, 2011
Advanced Toxicology Course (EUROTOX Program), 2011
Lush Science Prize for co-authoring work on toxicity pathways in
hepatotoxicology and developmental toxicology, 2012
Oral presentation Award World Congress on Alternatives to Animal
Experimentation, 2014
NCAC Society of Toxicology Graduate Student Representative, 2015-2016
In vitro Specialty Section Officer: Graduate Student Representative, 2016-
2017
Teaching Assistant PhD Molecular Toxicology Course 2016
IFER graduate fellowship for alternatives to the use of animals in science,
2014-2017
English /Spanish – Bilingual, Proficient
LANGUAGES Italian – Level B1
German – level A2
WORK PhD Candidate in Environmental Health Sciences - Molecular and
EXPERIENCE Translational Toxicology, Bloomberg School of Public Health, Johns Hopkins
2014-Current University, Baltimore, USA: This program is composed of two years of
coursework and three years of thesis research. During this period I obtained an
unconditional pass on my thesis proposal, have authored two publications and co-
authored six publications.
2013 Research technologist, Centre for Alternatives to Animal Testing, Johns
356
Hopkins Bloomberg School of Public health, Baltimore, USA: My work
focused on projects concerning the use of metobolomics to predict adverse
outcome pathways, development of 3D model in vitro systems to assess neuronal
toxicity LUHMES and hiPSC cell lines. Techniques: 3D cell culture,
metabolomics, transcriptomics, RNA, DNA, protein extraction, PCRs, western
blots, siRNA transfection, protocol adaptation for 3D cultures, DNA and protein
dot blots, electron microscopy, immunocytochemistry, confocal imaging.
2012 Junior Fellow, Centre for Biodiversiy, Functional and Integrative Genomics
(BioFIG), Faculty of Sciences, University of Lisbon, Portugal: My research
was aimed at optimising small-molecule screens to test their effect upon mutant
CFTR (cystic fibrosis transmembrane conductance regulator) trafficking to the cell
membrane. Production and management of cell lines (quality control, mycoplasma
tests, proliferation and freezing for stock storage) with different mutations or
fluorescently labelled protein. Fluorescence and confocal imaging of samples to
select small-molecules which are able to increase CFTR localisation to the
membrane in transduced A549 and CFBE cells. Techniques: Cell differentiation
protocols, cell transduction, siRNA gene silencing via lipotransfection and
electroporation, protein extraction, western blots, high throughput fluorescence
imaging, confocal imaging, image quantification.
2011 Traineeship at the Institute for Health and Consumer Protection (IHCP),
Ispra, Italy: My research entailed optimization of high throughput screening (96-
well format) of substances and the implementation of high content fluorescence
imaging so to identify the molecular mechanisms involved in hepatotoxicity for
various projects being carried out at an international level (working with the
HepaRG cell model). Furthermore, I was responsible for testing nanomaterial
toxicity (within the NanoTEST FP7 European project and OECD working party)
357
which focused on investigating the role of ROS in determining potential genotoxic
nanomaterials for which I presented my results at the EUROTOX conference in
June 2012. Our main aim was to develop in vitro methodologies which can
eventually lead to the assessment of the in vivo toxic potential of these chemicals
and nanomaterials. I also worked extensively with HepRG cell line using
hepatotoxicity in vitro assays. Techniques: High throughput screening, High
Content Imaging, 96-well plate reading using fluorescence microscopy) use of
staining kits, immunostaining, PCR and western blots. Genotoxicity measurements
using the MN assay. Multi-electrode array (MEA) Chip recordings and data
analysis. Development of standard operating procedures (SOPs) for the
methodologies used. Further experience: SOP writing, GLP implementation,
experimental planning, protocol optimisation, experimental validity testing,
presentations, meetings and reports for project deliverables.
2010 Laboratory Asistant, Physiology Department, Universidad de La Laguna,
Tenerife, Spain: ‘investigating the effects of different treatments with chaperones
on AGT (Alanine-glyoxylate aminotransferase) localisation in primary
hepatocytes from the transgenic mouse model for Primary Hyperoxaluria’. My
research was focused on investigating the effects of gene therapy on the non-
functional enzyme in Pompe´s disease (acid maltase) as well as the molecular
basis of the disease in transgenic mice. Techniques: primary cell cultures,
transfections, cell fixing, immunohistochemistry, TIRFM and confocal
microscopy, protein extraction, purification and quantification, western blots.
Mitochondria and peroxisome extraction from cell suspensions.
Biomedical Research Centre of Tenerife (CIBITEN), University Hospital,
2010 Tenerife, Spain: My position involved correcting, translating and keeping the
web pages for two biomedical research centres; CIBITEN (Centre for Biomedical
358
Research of Tenerife) and ITB (Institute for Biomedical Technologies), up to date.
Furtheremore, I produced a report published towards development of the centre
towards a European level which the Centre was awarded.
2009 MSc Toxicology Lab Research project, University of Birmingham, School of
Biosciences: ‘Cellular entry, trafficking, localisation and effect of nanoparticles’ I
chose this project because of my interest in the emerging field of nanotoxicology.
My thesis focused on assay development to identify the molecular mechanisms by
which nanoparticles enter and leave various cell lines (HepG2, MDCK, COS-7) as
well as their fate within the cell. Work presented in my thesis was published in Int
J Nanomedicine. 2012; 7: 2045–2055 DOI: 10.2147/IJN.S29334. Techniques:
Cell culture, cell transfection, Real-time Total Internal Reflection Fluorescence
(TIR-FM) and Confocal microscopy.
2008 (Summer Research Lab assistant, University of Birmingham, School of Biosciences: internship) Under the supervision of Dr. Janet Smith, working on ‘Biochemical mechanisms
underlying Muscular Dystrophy and the possible therapeutic role of stem cells’.
As an early opportunity I applied for, self-motivation, organisational skills,
dedication and fast learning played an important role. Training: Animal handling,
i.p injections, i.v. injections, Necropsy, dissections, tissue sample preparation,
microtome sectioning, cell culture, immunohistochemistry
PUBLICATIONS • Repeated-low dose effects in a 3D dopaminergic in vitro model. Harris G, Vy
Tran, Erin Pryce, Melanie Eschment, McCaffery M, Hartung T, Smirnova L.
(Manuscript under preparation)
• Toxicity, recovery and resilience in a 3D dopaminergic in vitro model exposed to
359
Rotenone. Harris G, Eschment M, Perez S, McCaffery M, Severin D, Freeman D,
Pamies D, Delp J, Hogberg H, Hartung T, Smirnova L. (Submitted Arch Toxicol)
• 3D Differentiation of LUHMES Cell Line to Study Recovery and Delayed
Neurotoxic Effects. Harris G, Hogberg H, Hartung T, Smirnova L. Curr Protoc
Toxicol. 2017 Aug 4;73:11.23.1-11.23.28. doi: 10.1002/cptx.29.
• A human brain microphysiological system derived from induced pluripotent
stem cells to study neurological diseases and toxicity. Pamies D, Barreras P,
Block K, Makri G, Kumar A, Wiersma D, Smirnova L, Zang C, Bressler J,
Christian KM, Harris G, Ming GL, Berlinicke CJ, Kyro K, Song H, Pardo CA,
Hartung T, Hogberg HT. ALTEX. 2017;34(3):362-376. doi:
10.14573/altex.1609122.
• A LUHMES 3D dopaminergic neuronal model for neurotoxicity testing allowing
long-term exposure and cellular resilience analysis. Smirnova L, Harris G, Delp
J, Valadares M, Pamies D, Hogberg HT, Waldmann T, Leist M, Hartung T. Arch
Toxicol. 2016 Nov; 90(11):2725-2743.
• Cellular resilience. Smirnova L, Harris G, Leist M, Hartung T. ALTEX. 2015;
32(4):247-60. doi: 10.14573/altex.1509271.
• Quality assurance of metabolomics. Bouhifd M, Beger R, Flynn T, Guo L, Harris
G, Hogberg H, Kaddurah-Daouk R, Kamp H, Kleensang A, Maertens A, Odwin-
DaCosta S, Pamies D, Robertson D, Smirnova L, Sun J, Zhao L, Hartung T.
ALTEX. 2015;32(4):319-26. doi: 10.14573/altex.1509161.
• Sex differences in liver toxicity-do female and male human primary hepatocytes
360
react differently to toxicants in vitro? Mennecozzi M, Landesmann B, Palosaari
T, Harris G, Whelan M. PLoS One. 2015 Apr 7;10(4):e0122786. doi:
10.1371/journal.pone.0122786.
• Iron oxide nanoparticle toxicity testing using high-throughput analysis and
high-content imaging. Harris G, Palosaari T, Magdolenova Z, Mennecozzi M,
Gineste JM, Saavedra L, Milcamps A, Huk A, Collins AR, Dusinska M, Whelan
M. Nanotoxicology. 2015 May; 9 Suppl 1:87-94. doi:
10.3109/17435390.2013.816797.
• Building shared experience to advance practical application of pathway-based
toxicology: liver toxicity mode-of-action. Willett C, Caverly Rae J, Goyak KO,
Minsavage G, Westmoreland C, Andersen M, Avigan M, Duché D, Harris G,
Hartung T, Jaeschke H, Kleensang A, Landesmann B, Martos S, Matevia M,
Toole C, Rowan A, Schultz T, Seed J, Senior J, Shah I, Subramanian K, Vinken
M, Watkins P. ALTEX. 2014;31 (4):500-19. doi:
http://dx.doi.org/10.14573/altex.1401281.
• Consensus report on the future of animal-free systemic toxicity testing. Leist M,
Hasiwa N, Rovida C, Daneshian M, Basketter D, Kimber I, Clewell H, Gocht T,
Goldberg A, Busquet F, Rossi AM, Schwarz M, Stephens M, Taalman R,
Knudsen TB, McKim J, Harris G, Pamies D, Hartung T. ALTEX. 2014; 31(3):341-
56. doi: http://dx.doi.org/10.14573/altex.1406091.
• High throughput screening and high content imaging techniques to test
nanomaterial toxicity to 3T3 fibroblasts in vitro. Harris G, Palosaari T,
Milcamps A, Whelan M (JRC Publications, Scientific and Technical Report,
2014).
361
• Genomic and phenotypic alterations of the neuronal-like cells derived from
human embryonal carcinoma stem cells (NT2) caused by exposure to
organophosphorus compounds paraoxon and mipafox. Pamies D, Sogorb MA,
Fabbri M, Gribaldo L, Collotta A, Scelfo B, Vilanova E, Harris G, Bal-Price A. Int
J Mol Sci. 2014 Jan 9;15(1):905-26. doi: 10.3390/ijms15010905.
• Silencing of PNPLA6, the neuropathy target esterase (NTE) codifying gene,
alters neurodifferentiation of human embryonal carcinoma stem cells (NT2).
Pamies D, Bal-Price A, Fabbri M, Gribaldo L, Scelfo B, Harris G, Collotta A,
Vilanova E, Sogorb MA. Neuroscience. 2014 Dec 5; 281:54-67. doi:
10.1016/j.neuroscience.2014.08.031.
• Toward a 3D model of human brain development for studying
gene/environment interactions. Hogberg HT, Bressler J, Christian KM, Harris
G, Makri G, O'Driscoll C, Pamies D, Smirnova L, Wen Z, Hartung T. Stem Cell
Res Ther. 2013;4 Suppl 1:S4. doi: 10.1186/scrt365..
• A holistic assessment of realistic mixtures for human health: An approach to
tap water. Zaldívar JM, Loos R, Hoekstra EJ, Mennecozzi M, Tavazzi S,
Paracchini B, Saavedra L, Harris G, Hannaert P. and Dachs J. (JRC Publications,
Scientific and Technical Report, 2013)
• Hepatotoxicity Screening Taking a Mode-Of-Action Approach Using HepaRG
Cells and HCA. Mennecozzi M, Landesmann B, Harris G, Liska R, and Whelan
M, ALTEX Proceedings, 1/12, Proceedings of WC8. 2012. doi:
10.1016/j.toxlet.2012.03.687
362
• Dopaminergic cell recovery in an in vitro 3D model to study Parkinson's disease.
SELECTED Harris G., Eschment M, McCaffery M, Severin D, Freeman D, Pamies D,
ORAL Delp J, Hogberg H, Hartung T, Smirnova L. (World Congress on Alternatives
PRESENTA- Conference Oral Presentation, 2017)
TIONS • Human iPSC-derived 3D model to study neurotoxicity in vitro Harris G., Pamies
D, Barreras P, Block K, Makri G, Kumar A, Wiersma D, Smirnova L, Zang C,
Bressler J, Christian KM, Ming GL, Berlinicke CJ, Kyro K, Song H, Pardo CA,
Hartung T, Hogberg HT. (World Congress on Alternatives Conference Best Oral
Presentation Award, 2015)
• Automated high content fluorescence imaging for in vitro assessment of
nanomaterial toxicity. Harris G, Palosaari T, Magdolenova Z, Mennecozzi M,
Gineste JM, Saavedra L, Milcamps A, Huk A, Collins AR, Dusinska M, Whelan
M. (Eurotox 2012 Conference Oral Presentation, 2012)
REFERENCES Prof Thomas Hartung: Director for the Centre for Alternatives to Animal
Testing (CAAT), Bloomberg School of Public Health, Johns Hopkins University.
Dr. Helena Hogberg: Co-director for the Center for Alternatives to Animal
Testing (CAAT), Bloomberg School of Public Health, Johns Hopkins University.
Dr. Milena Mennecozzi: Principal Scientist, Plasticell Limited, Stevenage
Bioscience Catalyst. [email protected]
363