1 2 3 4 5 6 7 8 9 10 11 Draft NORDIC WORKING PAPERS

The Role of Retinoids in Female and Male Reproduction Including a Scoping Effort to Identify Potential Methods for Regulatory Use

This working paper has been published with financial support from the Nordic Council of Ministers. However, the contents of this working paper do not necessarily reflect the views, policies or recommendations of the Nordic Council of Ministers.

Nordisk Council of Ministers – Ved Stranden 18 – 1061 Copenhagen K – www.norden.org 12

1

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 The Role of Retinoids in Female and Male 37 Reproduction Including a Scoping Effort to Identify 38 Potential Methods for Regulatory Use

1 39 Table of contents 40 41 Table of contents ...... 2 42 1. Preface ...... 4 43 Financing and workforce ...... 5 44 2. Abbreviation list ...... 6 45 3. Summary ...... 7 46 4. Introduction to the retinoid system ...... 8 47 History and general background ...... 8 48 Retinoid Storage, Transport and Metabolism...... 9 49 Retinoid Receptors and gene regulation ...... 11 50 Crosstalk between the retinoid signaling pathway and other endocrine signaling pathways ...... 13 51 Epigenetics and its role in the retinoid system ...... 14 52 5. Formation of female and male reproductive organs, and the role of retinoids...... 15 53 6. Germ cells, meiosis and the role of retinoids ...... 16 54 7. Female reproductive organ development, function and health and the role of retinoids ...... 19 55 Oogenesis ...... 19 56 Ovarian somatic cells and folliculogenesis ...... 19 57 Ovarian steroidogenesis ...... 21 58 Other female reproductive organs and pregnancy ...... 23 59 Female reproductive pathologies and retinoid involvement ...... 24 60 Retinoid deficiency/excess in female reproduction ...... 25 61 Crosstalk with other signaling pathways ...... 25 62 8. Male reproductive organ development, function and health, and the role of retinoids ...... 26 63 Testicular somatic cells and steroidogenesis ...... 26 64 Spermatogenesis ...... 27 65 Secondary male reproductive organs ...... 31 66 Male reproductive pathologies ...... 31 67 Retinoid deficiency in male reproduction ...... 32 68 9. Impact on female and male reproduction by chemicals acting via the retinoid system ...... 33 69 Chemicals affecting female reproduction...... 35 70 Chemicals affecting male reproduction ...... 36 71 10. Potential adverse outcome pathways in female and male reproduction ...... 39 72 Proposed AOP for in utero CYP26B1 inhibition effects on male fertility ...... 40 73 Proposed AOP for effects of RDH10/RALDH inhibition in the adult on male fertility ...... 41 74 Proposed AOP for meiosis initiation, oogenesis, folliculogenesis and female fertility, focusing on Stra8 75 ...... 42 76 Proposed AOP for endometriosis in the adult female, focusing on CYP26A1 ...... 43 77 11. Initial scoping effort: assays and endpoints for effects of chemicals on female and male

2 78 reproduction via the retinoid system ...... 44 79 In silico methods ...... 44 80 In vitro assays and ex vivo methods ...... 44 81 In vivo endpoints ...... 45 82 12. Concluding remarks ...... 50 83 13. References ...... 51 84 14. Appendix ...... 63 85 Assay methods and availability: ...... 63 86 PubMed search strategy – retinoid parameters, chemical names, female/male reproduction 87 parameters...... 64 88 89

3 90 1. Preface 91 92 Terms of reference and Scope of report 93 This Nordic report represents part of the outcome of a project within the Organisation for Economic 94 Co-operation and Development (OECD) test-guideline (TG) program. This project was initiated by 95 Sweden (Helen Håkansson, Institute of Environmental Medicine, Karolinska Institutet, Sweden) and 96 coordinated by the Swedish Chemicals Agency under supervision of both a Steering committee from 97 the funding body (Nordic Co-ordination for the Development of Test Methods in Toxicology and 98 Ecotoxicology, Nord-UTTE) and by the OECD Endocrine Disrupters Testing and Assessment Advisory 99 Group. 100 101 This project builds on the ideas brought forward in the OECD Detailed Review Paper (DRP) 178 on 102 endocrine disrupter testing1 which addressed multiple aspects of the endocrine system, beyond 103 estrogens, androgens, thyroid disruption and steroidogenesis (EATS). The OECD DRP 178 suggests 104 projects in several areas, and particularly mentions assays relevant to the retinoid system, such as 105 Retinoid X Receptor (RXR) and Retinoic Acid Receptor (RAR)-reporter assays, Aryl hydrocarbon 106 Receptor (AhR) reporter assays, adipocyte differentiation, and retinoid serum levels. As identification 107 of retinoid system modulation is not included in any OECD test guideline it is urgent to cover this 108 identified knowledge gap. 109 110 The long-term aim was to develop methods that facilitate early screening, to enhance existing in vivo 111 test guidelines and to identify markers of biological effects for use in population studies, based on 112 information from retinoid biology and disturbed retinoid signalling in several organ systems. The EU 113 Commission funded development of a draft DRP on the retinoid system, that covered e.g. overall 114 biology of the retinoid system in human health and the environment, including a section on retinoic 115 acid and the reproductive system as well as a substantial annex on the retinoid system in male 116 reproduction2. However, it became apparent that the scope of the aim was too broad, and that for 117 certain aspects science was not mature enough for regulatory purposes. Consequently, the scope was 118 limited, and prioritisation was given to reproduction, as female reproduction had been identified as a 119 highly prioritized area in the recommendation from the EU Commission prioritisation workshop 20173. 120 In addition, it was decided that the annex on male reproduction should be further developed. Sweden, 121 as lead country, with support from all parties involved, decided to, as a first step of the DRP Retinoid 122 Process, publish this Nordic Working Paper. The present report will be followed up in the upcoming 123 Detailed Review Paper on retinoids, due to be published by the OECD in 2020. 124 125 Disposition 126 The report covers an overview of the retinoid system, the role of retinoids in the reproductive organs, 127 the impact of chemicals on reproduction via the retinoid system, potential adverse outcome pathways, 128 and an initial scoping effort for the possible role of chemical-induced retinoid disruption of the male 129 and female reproductive systems. In this report, the term reproduction refers to the formation and

1 OECD (2012), Detailed Review Paper on the State of the Science on Novel In Vitro and In Vivo Screening and Testing Methods and Endpoints for Evaluating Endocrine Disrupters, OECD Series on Testing and Assessment, No. 178, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264221352-en. 2 OECD (2017), “Detailed review paper on the retinoid system (draft)”, OECD, Paris. 3 Setting priorities for further development and validation of test methods and testing approaches for evaluating endocrine disruptors, in Brussels 2017 (Final Report European Commission Directorate- General for Environment Directorate B. (ISBN 978-92-79-83076-1), Available at https://publications.europa.eu/en/publication-detail/-/publication/6b464845-4833-11e8-be1d- 01aa75ed71a1/language-en

4 130 development of reproduction organs and their normal function. 131 132 Financing and workforce 133 The Nordic Working Group for Chemicals, Environment and Health (NKE) (in NKE Contracts 2017-003, 134 2018-026, 2019-003) and The Swedish Chemicals Agency provided the main funding for this work. 135 Furthermore, the European Commission (Framework Contract ENV.A.3/FRA/2014/0029) provided 136 funding for the draft DRP reported in 2017 (mainly written by Alice Baynes, Edwin Routledge, Sofie 137 Christiansen and Ulla Hass), parts of which has been incorporated in this report. 138 139 The main author of this Nordic report has been Dr. Charlotte Nilsson, RISE Research Institutes of 140 Sweden. The in-kind contributions from the scientific experts involved in writing and reviewing this 141 document is highly acknowledged (see list of contributors). 142 143 This draft report describes the status as of September 2019. The final report will be publicly available 144 via the NKE website4, and will describe the status as of December 2019. 145 146 147 List of contributors 148 149 EU Framework contract (2016-2017): 150 Alice Baynes & Edwin Routledge (Brunel University London); Sofie Christiansen & Ulla Hass (Technical 151 University of Denmark). 152 153 Nordic Working Paper (2018-2019): Miriam Jacobs (Public Health England); Norbert Ghyselinck & 154 Manuel Mark (Institute of Genetics and Molecular and Cellular Biology, France); Stefano Lorenzetti & 155 Alberto Mantovani (Instituto Superiore di Sanità, Italy); Taisen Iguchi (National Institute for Basic 156 Biology, Japan); Pauliina Damdimopoulou & Astrud Tuck (CLINTEC, Karolinska Institutet, Sweden); 157 Helen Håkansson (Institute of Environmental Medicine, Karolinska Institutet, Sweden); Anne-Lee 158 Gustafson & Margareta Halin Lejonklou (Swedish Chemicals Agency); Sofie Christiansen (Technical 159 University of Denmark). Nancy Baker (United States Environmental Protection Agency); Jodi Flaws 160 (University of Illinois, USA); Elise Grignard & Sharon Munn (Joint Research Center, DG Environment 161 European Commission); Peter Korytar (DG Environment European Commission); Anne Gourmelon & 162 Patience Browne (OECD Test Guidelines Programme, Environment Directorate, OECD). Contributions 163 from the OECD Retinoid EG will be added to this list. 164 165 166 Nord-UTTE steering group, subgroup NKG (Nordic Chemicals Group) (2017-2019): 167 Sjur Andersen (The Norwegian Environment Agency), Knud Ladegaard Pedersen, Henrik Tyle (until 168 ultimo 2018) & Marie-Louise Holmer (until primo 2018 (The Danish Environmental Protection Agency); 169 Sofie Christiansen (Technical University of Denmark); Matti Leppänen (The Finnish Environment 170 Institute); Petteri Talasniemi (The Finnish Safety and Chemicals Agency); Anne-Lee Gustafson (The 171 Swedish Chemicals Agency).

4 https://www.norden.org/en/organisation/nordic-working-group-chemicals-environment-and-health-nke 5 172 2. Abbreviation list 173

174

6 175 3. Summary

7 176 4. Introduction to the retinoid system 177 178 179 History and general background 180 Retinoids5, a chemically related group of compounds which includes vitamin A, can exist in several 181 different forms (see Figure 1). As the name “vitamin” implies, they are essential micronutrients that 182 must be supplied from the diet, either in the form of carotenoids from vegetable sources, or retinol 183 and retinyl esters from animal sources (Al Tanoury et al. 2013). Carotenoids are responsible for the 184 bright red, yellow and orange colors of many vegetables. The liver, an organ found in all vertebrate 185 organisms, can be a rich source of retinoids and may have been used by the ancient Egyptians as a 186 cure for night blindness, which is a typical symptom of retinoid deficiency (Wolf 1996). Conversely, 187 Arctic cultures have long known to avoid eating polar bear liver, which can contain extremely high 188 levels of retinoids, to avoid adverse effects such as blurred vision, nausea, skin loss, coma and even 189 death (Rodahl and Moore 1943). Thus, history has shown that both hypovitaminosis A and 190 hypervitaminosis A can be detrimental. 191 192 The requirement for retinoids in normal physiological functions has been studied for more than a 193 century, by examining the effects of retinoid deficiency or excess in different species, and, more 194 recently, in genetically modified mice (reviewed in Clagett-Dame and Knutson 2011). Retinoids are 195 essential for vision, reproduction, embryonic development, adult growth and development, and 196 maintenance of immunity and epithelial barriers (O’Byrne and Blaner 2013). Vitamin A deficiency 197 (VAD) is the main cause of preventable blindness in the world (Bastos Maia et al. 2019), and it has 198 been hypothesized that hearing loss among humans in the developing world is due to gestational 199 VAD (Emmet and West 2014). Recently, new metabolic functions for retinoids have been discovered, 200 in e.g. lipid metabolism and insulin response (Napoli 2017, Cione et al. 2016). The importance of 201 retinoids and the retinoid signaling pathways is reflected in both the ancestry and the conservation 202 of genes and pathways among both vertebrates and invertebrates (André et al. 2014). 203 204 The most fat-soluble forms of retinoids (retinyl esters) can be stored within the body in relatively 205 high levels, which can counteract periods of low dietary retinoid dietary intake (O’Byrne and Blaner 206 2013). Despite this, retinoid deficiency is one of the most prevalent nutrient deficiencies worldwide 207 (Napoli 2017). 208

5 The term “retinoids” covers all analogues of vitamin A, naturally occurring as well as synthetic. 8 209 210 211 212 213 214 13-cis-retinoic 9, 13-di-cis- 215 acid retinoic acid 216 217 218 Figure 1. Chemical structures of various forms of retinoid precursors (β-carotene), intermediate 219 transformation, transport form (all-trans-retinol; ROH), storage form (all-trans-retinyl ester), 220 intermediate forms also important in vision (11-cis-retinaldehyde, all-trans-retinaldehyde), and the 221 nuclear receptor-binding forms (all-trans-retinoic acid, 9-cis-retinoic acid, 13-cis-retinoic acid, 9,13-di- 222 cis-retinoic acid). 223 224 Retinoid Storage, Transport and Metabolism. 225 226 Storage 227 The vast majority (approximately 80-85%) of retinoids found in the body is stored in the liver (O’Byrne 228 et al. 2005). Two main cell types have been identified as being central to metabolism and storage of 229 retinoids, namely hepatocytes and hepatic stellate cells (HSC), respectively. Hepatocytes contain 230 around 10-20% of hepatic Vitamin A and are not only involved in the initial uptake of Vitamin A, but 231 are also the site of RBP synthesis and secretion (Sauvant et al. 2001). Therefore, hepatocytes are 232 important in both the uptake and the mobilization of vitamin A into, and out of, the liver. HSC have 233 been found to contain the remainder (80-90%) of hepatic vitamin A (Blaner and Li 2015), and are 234 therefore considered the main vitamin A storage cells. Studies in mice, using radiolabeled retinyl 235 esters, have shown dietary vitamin A to be initially taken up by the hepatocytes and then rapidly 236 transferred to HSC (Blomhoff et al. 1982). 237 238 Another well-known retinoid storage organ is the eye. Retinoids has long been known to play a vital 239 role in vision and eye function. The retina and retinal pigmented epithelium (RPE) contain 11-cis and 240 all-trans-isomers of retinaldehyde, ROH, and RE (Palczewski 2012), with the RPE being the main site 241 of RE storage. In contrast, the photoreceptors of the retina contain large amounts of retinaldehyde.

9 242 RE stores are also found in a number of other organs/tissues, but at lower levels than the liver, 243 including the lung, brain, skin, muscle, kidney, spleen, white and brown adipose tissue, and testis 244 (e.g. Sertoli cells) (summarized in Blaner and Li, 2015). 245 246 Within the (mammalian) body, the most abundant forms of retinoids are ROH and RE (see figure 1). 247 RE are often found in the circulation after a meal, and are the main storage forms of retinoids in the 248 liver but can also be found in other tissues. ROH can also be found in the blood and is the main 249 circulating form of retinoids during fasting. ROH is generally bound to a number of chaperone 250 proteins, known collectively as retinol-binding proteins. Of these, retinol-binding protein (RBP) 251 transports ROH in the blood, and cellular retinol-binding protein-1 (CRBP1), CRBP2, and inter- 252 photoreceptor retinoid-binding protein (IRBP) occur intracellularly (Blaner and Li 2015). Cellular 253 retinoic acid binding proteins (CRABPs) are involved in RA transport from the cytoplasm to the 254 nuclear retinoic acid receptor (RAR) and also play a role in determining intracellular RA levels by 255 controlling the amount of RA that is available for degradation (Napoli 2017). 256 257 A number of enzymes are required for the balance of storing and re-mobilizing of retinoids during 258 times of abundance and insufficiency, respectively. The most important enzymes are those 259 synthesizing RE (such as LRAT; lecithin:retinol acyltransferase) and those used to hydrolyze RE to ROH 260 (such as REH; retinyl ester hydrolase) (Figure 2). These enzymes, along with (i) the capacity to 261 accumulate RE in lipid droplets known as chylomicrons within cells and tissues and (ii) the ability to 262 re-mobilize these stores when needed, compensate for fluctuations in retinoid intake via the diet so 263 as not to negatively impact the individual. The process of remobilization of stored RE in to circulating 264 ROH seems reliant on interactions with RBP, as evidenced by studies using RBP-deficient mice 265 (Quadro et al. 1999). 266 267 Transport 268 A range of retinoid forms are found in fasting and/or non-fasting states. These include RE in 269 chylomicrons and chylomicron remnants (fat droplets exported to blood and lymph from epithelial 270 cells of the small intestine); RE in very low density lipoproteins (VLDL), low density lipoprotein (LDL) 271 and high density lipoprotein (HDL); ROH bound to RBP4 (serum retinol-binding protein type 4); RA 272 bound to albumin; and water-soluble beta-glucuronide conjugates of ROH and RA. The delivery of 273 retinoids to tissues involves many different vitamin A forms and carriers. However, ROH bound to 274 RBP4 and chylomicron RE are considered the most important (largest volume) transport pathways 275 (Laudet et al. 2015). 276 277 It is worth noting that retinoid delivery to tissues works in a similar way to vitamin D and thyroid 278 hormone (Blaner and Li 2015), by transport of large quantities of the biologically-inactive form (i.e. 279 ROH, 25-hydroxy-vitamin D, or thyroid hormone T4), and relatively low quantities of the biologically- 280 active form (RA, 1,25- dihydroxy-vitamin D, or thyroid hormone T3) in the blood (Blaner and Li 2015). 281 ROH, bound to RBP4, forms a larger complex the T4 transporting protein transthyretin (TTR) in the 282 blood (Figure 2). Both TTR and RBP4 are mainly synthesized in the liver and choroid plexus, and 283 secreted into the plasma by the liver and cerebrospinal fluid, respectively (Shirakami et al. 2012). In 284 contrast, there is no specific binding protein for the transport of RA in the plasma, as it is synthesized 285 locally in target cells by SDR, ADH, RALDH enzymes (Figure 2). 286 287 Metabolism 288 RA is considered to be the most biologically important member of the retinoid family for nuclear 289 signaling (see retinoid receptor section below). However, as stated above, the most abundant forms 290 of internal circulating vitamin A are ROH and RE. Therefore, within target cells, a number of 291 conversion steps are required to transform relatively inactive forms of retinoids into the biologically 292 active RA form (reviewed in Shannon et al. 2017). Once ROH-RBP4 has been bound to STRA6 and 293 taken up into a target cell, ROH is converted to RA via a two-step oxidation process (Figure 2). The 294 first step, in which ROH is converted into retinaldehyde, is reversible and is carried out either by a 10 295 short-chain dehydrogenase/reductase (SDR; a complex consisting of retinol dehydrogenase 10 296 [RDH10], and retinaldehyde reductase 3 [DHRS3]) or by cytosolic alcohol dehydrogenases (ADH) 297 (Shannon et al 2017). The majority of ROH is, in normal physiological conditions, bound to CRBP and 298 directed towards SDR rather than the more unspecific ADH (Napoli 2017). 299 300 The retinaldehyde molecule can undergo the second, irreversible, step and be further oxidized into 301 RA by retinaldehyde dehydrogenases RALDH 1, 2 or 3; (also known as aldehyde dehydrogenases 302 ALDH1A1 - ALDH1A3 (Shannon et al 2017). However, retinaldehyde can also be reduced back to ROH 303 by SDR, or, under VAD conditions, by retinol dehydrogenase 11 (RDH11) (Belyaeva et al. 2018). The 304 ratio of CRBP-ROH/retinaldehyde (holo-CRBP) to CRBP without ROH/retinaldehyde (apo-CRBP) 305 appears to signal whether retinaldehyde should be reduced back to ROH or further oxidized to RA 306 (Napoli et al. 2017). The fate of the retinaldehyde also depends on a negative feedback response to 307 available RA (Shannon et al. 2017). 308 309 RA is eventually catabolized, mainly by cytochromes P450 hydroxylases 26A1, B1 and C1 (Topletz et 310 al. 2015). A large number of other cytochrome p450 enzymes have also been shown to degrade RA in 311 vitro, although the relevance of these different CYP enzymes to normal retinoid homeostasis in vivo 312 is unclear (Laudet et al. 2015). In the adult human body, CYP26A1 is mainly expressed in the liver and 313 is responsible for more than 90% of the hepatic clearance of RA (Thatcher et al. 2010). CYP26A1 is 314 also expressed in e.g. testis, epididymis, uterus, endometrium and placenta, while CYP26B1- 315 expression is more dispersed and found e.g. in the brain, testes, placenta, ovaries, endometrium. 316 (Human Protein Atlas6). In the human fetus, CYP26A1 is the form predominantly expressed in the 317 brain, whereas CYP26B1 is expressed in all tissues except the brain (Kedishvili 2013). In humans, 318 CYP26C1 is mainly expressed during embryo development, but can be expressed at low levels in 319 adult tissues, e.g. in testis (Ross and Zolfaghari 2011). 320 321 While levels of RA in tissues are governed mainly by a balance of production and metabolic 322 breakdown with RA feedback loops (Kedishvili 2013, Teletin et al. 2017). Some tissue-specific uptake 323 of RA from serum appears to take place, via a, to our knowledge, still not established mechanism 324 (Kurlandsky et al. 1995, O’Bryan and Blaner 2013). In the Kurlandsky study, performed only with male 325 rats, the contribution of serum RA to the tissue pool of RA was 0.7% for the testis, 10% for the 326 epididymis and 23% for the seminal vesicles. 327 328 The physiological role of RA in cell differentiation has led to its use in cancer treatment, including 329 reproductive organ malignancies (Siddikuzzaman et al. 2011). The treatment increases the normal 330 serum nM levels of RA up to µM levels, which over time leads to autoinduction of RA clearance via 331 upregulation of CYP26 enzymes (Jing et al. 2017). Attempts have been made to counteract therapy 332 resistance by developing e.g. CYP26 inhibitors (Nelson et al. 2013). However, side effects and low 333 potency has limited the use of these inhibitors (Jing et al. 2017). Similar efforts to maintain RA 334 homeostasis has also been observed in models of teratogenicity, and it has been speculated that the 335 teratogenic effects of RA are actually due to a prolonged local RA deficiency, caused by differential 336 induction of genes coding for enzymes producing (Raldh1-3) or breaking down (Cyp26a1 and 337 Cyp26b1) RA (Lee et al. 2012). Interestingly, maternal supplementation with low doses of RA 338 following a teratogenic insult has been shown to rescue or abrogate teratogenic effects (Lee et al. 339 2012). 340 341 Retinoid Receptors and gene regulation 342 RA acts mainly transcriptionally via specific nuclear retinoic acid receptors, RARs, which can form 343 homodimers with other RAR isomers, or heterodimers with retinoid X receptors, RXRs to influence a 344 variety of cellular processes (Germain et al. 2006). The receptor dimers bind to, and activate, the 345 retinoic acid response element (RARE), inducing transcription of the target gene in the presence of

6 https://www.proteinatlas.org/ accessed in May 2019 11 346 the ligand. RARs can be activated both by RA and 9-cis-RA (Allenby et al. 1993). The endogenous. The 347 endogenous ligand for RXRs appears to be either 9,13-di-cis-RA or 9-cis-RA, formed via isomerization 348 from RA (de Lera et al. 2016). In addition, RARs have a vital role without activation by retinoids. The 349 RAR_RXR heterodimer can bind to DNA in the absence of a ligand, leading to the recruitment of a 350 corepressor complex that suppress transcription (Vilhais-Neto and Pourquié 2008). 351 352 RARs and RXRs belong to the same nuclear hormone receptor family as steroid hormones, thyroid 353 hormone and vitamin D receptors, as well as various orphan receptors and receptors activated by 354 intermediary metabolites (e.g. PPAR; peroxisome proliferator-activated receptor, LXR; liver X 355 receptor, FXR; farnesoid X receptor and PXR; pregnane X receptor) (Szanto et al. 2004). Some of 356 these nuclear receptors may also be involved in retinoid signaling responses. For example, PPARs, α, 357 β/δ and γ heterodimerize with RXR and function as transcription factors (Mangelsdorf and Evans 358 1995). RA can also serve as a ligand for PPAR β/δ, but at a much lower affinity than RAR (Al Tanoury 359 et al. 2013). 360 361 Three of retinoid receptors (RARα, RXRα and RXRβ) have widespread expression patterns, whereas 362 the others (RARβ, RARγ and RXRγ) show a more complex, tissue-specific expression. Therefore, 363 most tissues are potential targets of retinoid signalling, although different heterodimeric 364 complexes can transduce the RA signal (Rhinn and Dollé 2012). 365 In addition to the classical genomic effects, retinoic acid has been found to have a number of non- 366 genomic mechanisms such as kinase activation (Rochette-Egly 2015). 367

12 368 369 370 Figure 2: Retinoid metabolism, transport and nuclear signaling. Abbreviations: alcohol 371 dehydrogenase (ADH), β-carotene (βcar), cellular retinoic acid-binding protein (CRABP), cellular 372 retinol-binding protein (CRBP), cytochrome P450 subunit 26 protein (CYP26), lecithin:retinol 373 acyltransferase (LRAT), retinoic acid (RA), retinaldehyde dehydrogenase (RALDH), retinoic acid 374 receptor (RAR), retinoic acid response element (RARE), retinol binding protein (RBP), retinyl esters 375 (RE), retinyl ester hydrolase (REH), retinol (ROH), retinoid X receptor (RXR), short-chain 376 dehydrogenase/reductase (SDR; an enzymatic complex consisting of retinol dehydrogenase RDH10 377 and retinaldehyde reductase DHRS3), stimulated by retinoic acid protein 6 (STRA6), transthyretin 378 (TTR). 379 380 381 Crosstalk between the retinoid signaling pathway and other endocrine signaling pathways 382 RXR can heterodimerize with other steroid and nuclear receptors, and functions essentially as a 383 master switch between different signaling pathways. Consequently, the retinoid signaling pathway 384 can crosstalk with other signaling pathways: in addition to the RARs, the heterodimerization partners 385 of RXRs include PPARs, TR, and VDR; the vitamin D receptor (Chambon 1996). In addition, crosstalk 386 with androgenic signaling (Long et al. 2019), as well as with xenobiotica-related receptor pathways, 387 via e.g. CAR; the constitutive androstane receptor, PXR, and AhR; the aryl hydrocarbon receptor has 388 been demonstrated. Possible responses to toxic insult include induction of xenobiotica-metabolizing 389 enzymes (Murphy et al. 2007, Shmarakov et al. 2019). 390 391 Crosstalk has also been shown on the level of regulating the expression of e.g. RA-catabolizing 392 enzymes in the CYP26 family, it has been demonstrated that in human liver cells, PPARγ agonists 393 rosiglitazone and pioglitazone induce CYP26A1 as well as the normally less abundant CYP26B1 (Tay 394 et al. 2010). The (non-endocrine) sonic hedgehog (SHH) pathway is reported to regulate expression 395 of the Cyp26a1 and Cyp26b1 genes, thereby preventing excessive RA levels during mouse embryo 396 development (El Shahawy et al. 2019). In the testicular Leydig cells, crosstalk between the retinoid

13 397 signaling system and testosterone signaling has been shown to be crucial for steroidogenic cell 398 function (Jauregui et al. 2018). 399 400 Epigenetics and its role in the retinoid system 401 Epigenetic changes are changes in gene expression that a) do not involve gene sequence alterations 402 and b) may persist after the initial trigger is long gone (reviewed in e.g Villota-Salazar et al. 2016, and 403 Jacobs et al. 2017). Examples of epigenetic changes are histone modification and DNA methylation. 404 Histone deacetylation/methylation and DNA methylation are usually associated with repression of 405 gene transcription, while histone acetylation/demethylation and lack of DNA methylation instead 406 leads to transcriptional activation. 407 408 RA is known to mediate cell differentiation also via epigenetic mechanisms (reviewed in Urvalek et al. 409 2014). Epigenetic changes are also involved in, e.g., the process of transient induction of the Cyp26a1 410 gene by RA (Yuan et al. 2012). In some cases, epigenetic actions of RARs appear to be independent of 411 the ligand, i.e., RA (Laursen et al. 2012). 412 413

14 414 5. Formation of female and male reproductive organs, and the role of 415 retinoids 416 417 The formation of the female and male reproductive organs is initiated very early in embryogenesis. 418 The genital ridges (the primordial gonads) are formed approximately halfway through gestation in 419 mice (Spiller et al. 2017) and during the first trimester in humans (Johansson et al. 2017). In mice, the 420 genital ridges are formed at 10.5 days post coitum (dpc), and soon begin to differentiate as testes, in 421 the presence of the Y-chromosomal gene Sry, or as ovaries, in the absence of Sry. The transcription 422 factor SRY upregulates the expression of the gene Sox9, which in turn governs the expression of 423 several male-specific genes, leading to differentiation of the bipotential somatic cell precursors into 424 Sertoli cells (Kashimada et al. 2011, Rossitto et al. 2015). In the absence of Sry, as in females, a 425 different set of genes are expressed by default, such as Wnt4 and Foxl2, leading to differentiation of 426 somatic precursor cells into female granulosa cells and subsequently to ovarian development 427 (Kashimada et al. 2011). 428 429 Degradation of endogenous RA appears to be critical for normal testis development, based on 430 observations of ovotestes and a feminized reproductive tract in male Cyp26b1-null mouse embryos, 431 which died at embryonic day (ED) 14.5-15.5 (Bowles et al. 2018). However, in an earlier study using 432 Cyp26b1-null mouse on a different genetic background, testis formation and somatic cell 433 differentiation in neonatal pups appeared essentially normal, although the testis were smaller 434 (MacLean et al. 2007). In an ex vivo model using rat fetal testis, exogenous RA could disrupt the 435 proper formation of seminiferous cords, as well as the maintenance of the testicular cell fate of the 436 somatic cells; following RA exposure, expression of ovarian-specific marker FOXL2 was observed 437 (Spade et al. 2018a). 438 439 Even in adulthood, the sexual fate of the male gonads must be maintained, probably via the active 440 presence of the transcription factor DMRT17 (Minkina et al. 2014). The role of DMRT1 in the testes 441 appears to be to prevent RA-initiated activation of specific potential feminizing genes (Foxl2 and 442 Esr2) in the Sertoli cells; without DMRT1 present, male-to-female trans-differentiation of somatic 443 cells was observed in the pre- and postnatal mouse testes (Minkina et al. 2014). 444 445 In contrast, in the female mouse embryo it seems like RA signaling is not required for correct ovarian 446 development, based on observations in mice with deletions/disruptions of either RARs or RA- 447 synthesizing enzymes (Minkina et al. 2017), although earlier studies have shown that RA was 448 required in order to maintain ovarian differentiation and development in the developing fetal female 449 murine gonads (Minkina et al. 2014, Suzuki et al. 2015). Additional studies suggest that abnormal 450 endogenous RA levels may be able influence somatic cell differentiation and/or function (Bowles et 451 al. 2018). 452

7 DMRT1: Doublesex and mab-3 related transcription factor 1 15 453 6. Germ cells, meiosis and the role of retinoids 454 455 In both male and female mouse embryos, a cluster of pluripotent primordial germ cell (PGC) 456 precursors arise starting 6.25 dpc at the posterior primitive streak of the embryo. Subsequently, they 457 migrate to the genital ridges (formed at 10.5 dpc), which will differentiate into testis or ovaries (see 458 section 5). 459 460 In both males and females, the haploid gametes are produced from germ cells via meiosis in the 461 gonads, following a similar process of events and requiring expression of many of the same genes. 462 However, the timing of these events is completely different. Around the same time as PGCs in fetal 463 ovaries enter prophase of the first meiotic division (ca 13 dpc in mice), PGCs in a testicular 464 environment enter quiescence; they slow down their proliferation towards mitotic arrest as GO/G1- 465 arrested gonocytes (Kashimada et al. 2011). Thus, in the fetal gonad the commitment of germ cells 466 towards oogenesis involves entry into meiosis, whereas commitment to spermatogenesis involves 467 the inhibition of meiotic initiation, suppression of pluripotency and mitotic arrest. This mitotic arrest 468 is maintained until after birth (Spiller and Bowles 2015, Spiller et al. 2017) and consequently meiosis 469 initiation occurs postnatally in males. Thus, all oocytes are produced before birth, while 470 sperm(atocytes) are produced continuously during post-pubertal life in males. 471 472 Entrance into meiosis seems to be regulated by two master genes: Stra8 (“Stimulated by retinoic 473 acid, gene 8”) required for the initiation of meiosis in female fetal germ cells and Nanos2 (expressed 474 in male fetal germ cells), required to prevent Stra8 expression and meiosis initiation in male fetal 475 germ cells (Rossitto et al. 2015). Other genes, such as Fgf9, have also been implicated as being of 476 importance for proper control of meiotic entry; figure 3 is an attempt to summarize this information. 477 The Stra8 gene contains putative retinoid acid response elements (RAREs) and its expression is 478 activated by RA (Spiller et al. 2017). The Stra8 locus in non-germ cells appears to be epigenetically 479 silenced, thus, only pre-meiotic germ cells respond to RA by Stra8 induction (Spiller and Bowles 480 2015). The critical role of Stra8 is evident in Stra8-/- mice, as both females and males are infertile, 481 while heterozygotes of both sexes are fertile (Baltus et al. 2006). Stra8-/- females display smaller 482 ovaries lacking oocytes and follicles, and in the corresponding males, testes are smaller and testicular 483 germ cells severely reduced in numbers; in addition, spermatocytes undergo apoptosis before 484 reaching the leptotene and zygotene stages of meiotic prophase (Baltus et al. 2006). 485 486 Another meiosis-essential gene, the meiotic recombination component gene Rec8, is also an RA 487 target and is activated independently of Stra8: Rec8 encodes a meiosis-specific component of the 488 cohesion complex and is required for several steps in meiotic chromosomal activities (e.g. chromatid 489 cohesion and chiasmata formation) (Kuobova et al. 2014). Neither female nor male Rec8-/- mice that 490 survive to reach sexual maturity are fertile (Xu et al. 2005). 491

16 492 493 494 495 Figure 3: Comparison of male and female germ cell development and meiotic progression in mice (in 496 the draft DRP, re-drawn from Dollé and Niederrether 20158). Blue indicates male-specific events, 497 red/pink indicated female-specific events and grey, events shared by both sexes. Numbers indicate 498 days post coitum (dpc). Arrows and bars indicate rough timing of gene expression, solid arrows 499 indicate direct effect, and broken arrows indicate an incompletely characterized or likely indirect 500 effect. Plus (+) or minus (-) indicate supportive/additive or inhibitory action on other genes. GC; germ 501 cell, PGC; Primordial Germ Cells, RA; retinoic acid, mitotic pro; mitotic proliferation. Summary figure 502 in the draft DRP, adapted from Clagett-Dame and Knutson 2011, Feng et al. 2014, Rossitto et al. 2015 503 and Spiller and Bowles 2015. 504 505 Retinoic acid has been implicated in controlling the onset of meiosis. The RA necessary for meiosis 506 initiation is available locally in both the fetal testis and fetal ovaries. Several sources for RA have 507 been suggested. It can diffuse into the primordial gonads from the nearby mesonephros, where, in 508 mice, retinol dehydrogenase 10 (Rdh10) and retinaldehydrogenase 2 (Raldh2) as well as Raldh3, at a 509 lower level, are expressed (Niederreither, Vermot et al. 2002, Niederreither, Fraulob et al. 2002, 510 Bowles et al. 2006, Spiller and Bowles 2015, Bowles et al. 2018). An alternative hypothesis is that RA 511 is produced by the coelomic epithelium, which covers the embryonic gonad and where RALDH2 is 512 also expressed (Niederreither et al. 1997, Teletin et al. 2017). RA-producing (retin)aldehyde 513 dehydrogenases are also present in both female and male fetal gonads in humans (Childs et al. 2011, 514 Le Bouffant et al. 2010). 515 516 The crucial difference between female and males is that the RA-catabolizing enzyme CYP26B1, which 517 is expressed in mouse fetal gonads of both sexes at 11.5 dpc, is no longer expressed in the female 518 gonad from 12.5 dpc (Bowles et al. 2006). RA would thus be degraded in the fetal testis, and thereby 519 meiotic entry is blocked (MacLean et al. 2007). In contrast, in the fetal ovary RA levels would be

8 Dollé P, Neiderreither K. The Retinoids: Biology, Biochemistry, and Disease. Wiley-Blackwell; 2015. 17 520 maintained, and Stra8 and Rec8 expressed, and subsequently female gonocytes would enter into 521 meiosis I (Rossitto et al. 2015). 522 523 The importance of sufficient RA levels in the fetal ovary is implicated in a retinoid-deficient rat 524 model, where meiosis initiation was prevented (Li and Clagett-Dame 2009). As discussed in the 525 previous paragraph, RA levels are maintained in both male and female gonads alike in the absence of 526 CYP26B1. In a Cyp26b1-/- mouse model, germ cell development was studied in ovaries and testes 527 from embryos at 13.5 dpc and from neonatal pups (MacLean et al. 2007). In that study, ovarian germ 528 cells at either stage appeared to be unaffected by the lack of CYP26B1; in the embryonic Cyp26b1-/- 529 testes, some germ cells had prematurely entered meiosis while others appeared apoptotic. Increased 530 levels of RA were demonstrated in the embryonic Cyp26b1-/- testes. Germ cells were essentially 531 absent in testes from neonatal Cyp26b1-/- pups. Premature germ cell meiosis could also be observed 532 in male wildtype genital ridges cultured in the presence of a synthetic retinoid (Am580, resistant to 533 CYP26B1 metabolism), suggesting that excess RA was responsible for the effect (MacLean et al. 534 2007). It has also been shown in vitro that embryonic testes grown in the presence of RA express 535 Stra8 (Koubova et al. 2006). 536 537 The role of RA in the induction of meiosis may be facilitating (Kumar et al. 2011, Teletin et al. 2019) 538 rather than critical (Bowles et al. 2016, Spiller et al. 2017, Teletin et al. 2017). Kumar and coworkers 539 (Kumar et al. 2011) reported that meiosis occurred normally in mouse fetal ovary lacking RA due to 540 RALDH2 and RALDH3 ablations. In the post-natal male mouse, Teletin and coworkers (Teletin et al. 541 2019) showed that meiosis is not blocked, only slightly delayed, in mice lacking all RALDH isomers in 542 the seminiferous epitelium (SE) of the testis. 543

18 544 7. Female reproductive organ development, function and health and the 545 role of retinoids 546 547 The ovary is the site for differentiation and release of mature oocytes for fertilization, but it is also 548 where sex hormones (necessary for follicle development, estrous cyclicity, maintenance of the 549 reproductive tract and its function) are synthesized and then released (Barnett et al. 2006). 550 551 Retinoids are now known to play an important role in the reproductive organs of females, especially 552 in sex determination (described in the previous section) and oogenesis/folliculogenesis during 553 embryogenesis but probably also in adult pathologies such as endometriosis. 554 555 Oogenesis 556 Oogenesis is the formation of haploid mature gametes from diploid PGCs, and the entire process 557 begins in utero with primary oocytes eventually arresting at prophase I (the first phase in meiosis) in 558 primordial follicles, which then remain inactive during childhood (Barnett et al. 2006, Teletin et al. 559 2017). In more detail, PGCs differentiate into oogonia upon arrival to the developing female gonad 560 and then undergo mitotic proliferation to create a stock of millions of oogonia that form germ cell 561 nests/cysts, i.e., clusters of cells connected by intercellular bridges (reviewed in Johansson et al. 562 2017). These nests break down when some of the oogonia within the nest undergo apoptosis, thus 563 breaking the connections between the cells and thereby allowing somatic cells to enter and surround 564 the germ cells, thus creating primordial follicles each containing a primary oocyte arrested in 565 prophase I. The breakdown of germ cell nests, constituting the beginning of primordial follicle 566 formation (initiating folliculogenesis, see next section), results in an oocyte number decrease. The 567 oogonia will enter meiosis I around E13.5 in mice and between 10 and 12 weeks of gestation in 568 humans (Grive and Freiman 2015). By entering meiosis I, there is a loss of pluripotency and the cells 569 can no longer divide mitotically. Thus, when all oogonia have entered the first meiotic division, no 570 further female germ cells can be produced, and the size of the ovarian reserve is fixed. 571 572 As already mentioned in the earlier section on sex determination, RA is widely accepted to play an 573 important (Bowles et al. 2016, Spiller et al. 2017, Teletin et al. 2017), although not critical (Kumar et 574 al. 2011, Teletin et al. 2019) role in the induction of meiosis. RA is believed to exerts its role by 575 activating expression of Stra8 (Spiller et al. 2017, Baltus et al. 2006) and possibly also Rec8 (Xu et al. 576 2005, Kuobova et al. 2014). However, meiosis-initiation via Stra8 induction reportedly occurs also in 577 a fetal ovary devoid of RA, as observed in Raldh2-/- mice (Kumar et al. 2011). 578 579 Ovarian somatic cells and folliculogenesis 580 As the genital ridges begin to differentiate as ovaries in female fetuses (at approx. 11 dpc in mice), 581 the gonadal somatic cells will differentiate into granulosa cells, after which meiosis will commence in 582 the oogonia (Minkina et al. 2017). The germ cell nests with the proliferating oogonia will break down 583 and somatic cells will enter and surround the germ cells, thereby forming the primordial follicles. The 584 processes of nest breakdown, germ cell death and formation of primordial follicles are similar in mice 585 and humans, but more sequentially organized in mice (Johansson et al. 2017). Nest breakdown takes 586 place around the time of birth in mice and begins during mid-gestation (around 16 weeks) in humans 587 (Grive and Freiman 2015). In addition, follicle assembly takes place prenatally in humans and 588 postnatally in rodents (Johansson et al. 2017). These differences may suggest different initiating 589 mechanisms. The primordial follicles, each of which contains an oocyte surrounded by a single layer 590 of somatic pre-granulosa cells, represent the ovarian reserve that must last for the entire 591 reproductive life in both mice and women (Grive and Freiman 2015). 592 593 Folliculogenesis, the process of follicle maturation, primarily takes place from the commencement of 594 puberty (Johansson et al. 2017). The first (primordial) follicles each consist of a single oocyte 595 surrounded by a layer of flat granulosa cells (GC). GCs progressively form several layers around the

19 596 oocyte with an outer layer of androgen-producing theca cells. The GCs subsequently convert 597 androgens to estrogens. As GCs continue to proliferate, the antrum is formed. At this stage, selection 598 occurs between growing follicles, so that only one or a limited number of follicles continue growing 599 to the preovulatory stage whereas others undergo atresia. After ovulation, luteinized theca cells and 600 mural GCs produce progesterone. 601 602

603 604 605 Figure 4: Main steps of folliculogenesis (adapted from Georges et al. 2014). Folliculogenesis starts 606 with primordial follicles that are subsequently recruited to become primary follicles. Proliferating 607 granulosa cells (GC) progressively form several layers around the oocyte (secondary to late preantral 608 follicles). From this stage, a layer of theca cells surrounds the follicle, and the follicles start to 609 produce estrogens. Theca cells produce androgens, which are converted into estrogens in granulosa 610 cells. Whereas follicular growth to the secondary/preantral stage is independent of gonadotropins, 611 progression beyond this stage strictly depends on follicle stimulating hormone (FSH) stimulation. As 612 GC continue to proliferate, the antrum is formed and separates granulosa cells into two general 613 populations: cumulus and mural GCs. At this stage, selection occurs between growing follicles, and 614 only one or a limited number of follicles continue growing to the preovulatory stage whereas others 615 undergo atresia. Two layers of theca (theca interna and externa) can be distinguished around 616 selected follicles. After ovulation, which is triggered by a peak of FSH and luteinizing hormone (LH), 617 theca cells and mural granulosa cells luteinize to produce progesterone. 618 619 For several years, RA was believed to be required in order to maintain ovarian differentiation and 620 development in the developing fetal female gonads (Minkina et al. 2014, Suzuki et al. 2015). RA is 621 also believed to regulate mouse ovarian follicle development in the adult, and it has, e.g., been 622 studied studied in vitro by stimulating granulosa cell proliferation (Demczuk et al. 2016). Ex vivo, 623 using cat ovarian cortices, RA activates the development of primordial follicles into primary and 624 secondary follicles, possibly via differential MMP regulation (Fujihara et al. 2018). It has been shown 625 that murine theca cells and granulosa cells both express the enzymes necessary for conversion of

20 626 retinol to retinoic acid, which consist of different forms of ADH and ALDH, and that cyp26B1 is 627 expressed in both theca and granulosa cells, but is differentially regulated after injection of equine 628 chorionic gonadotropin (eCG), an LH/FSH analogue (Kawai et al. 2016, Kawai et al. 2018). Also 629 demonstrated in vitro, ovarian de novo synthesis of RA is required for the follicular expression of the 630 LH receptor in granulosa cells of mouse ovaries and their ability to respond to the ovulatory LH surge; 631 oocytes appear to negatively regulate RA synthesis in pre-ovulatory follicles, impacting the LH 632 receptor expression in follicular somatic cells possibly via an epigenetic mechanism (Kawai et al. 633 2016, Kawai et al. 2018). Cultured human ovarian cumulus granulosa cells (obtained during oocyte 634 retrieval during the course of in vitro fertilization) produce RA from retinol in the media. RA causes 635 de-phosphorylation of connexin 43, involved in the gap junction cellular communication (GJIC) 636 between the granulosa cells in the cumulus-oocyte complex. De-phosphorylation of connexin 43 637 increases GJIC, which plays an important role in oogenesis and successful fertilization (Best et al. 638 2015). 639 640 However, at present, RA appears not to be required for ovarian granulosa cell specification, 641 differentiation or function (Minkina et al. 2017). In genetic mouse models, where all three RARs were 642 deleted in the XX somatic gonad at the time of sex determination, no significant effects were found 643 on ovarian differentiation, follicle development or female fertility. Nor did the knockout of all three 644 RA-producing aldehyde dehydrogenases (Aldh1a1-3) at the same timepoint appear to masculinize 645 the mouse ovaries. Additionally, an RAR antagonist (BMS-189453) was capable of blocking meiotic 646 initiation in the germ cells of 10.5 dpc XX gonads (from wild-type mice) in vitro but had little effect on 647 the expression of markers for either granulosa or Sertoli cells; thus, disruption of RA signaling does 648 not appear to disrupt early somatic differentiation (Minkina et al. 2017). 649 650 In murine ovaries, RBP4 is expressed before puberty but increases significantly in the peripubertal 651 period. In adult mice, RBP4 expression increased at pro-estrous and peaked at estrous and was 652 localized mainly in the granulosa and theca cells of follicles. Expression was also induced by FSH, 653 alone or in combination with LH; LH alone had no effect (Jiang et al. 2018b). 654 655 Ovarian steroidogenesis 656 Ovarian steroidogenesis, or the production of sex steroid hormones, is regulated by the pituitary 657 hormones LH and FSH; LH stimulates the ovarian thecal cells to produce androgens and FSH 658 stimulates the granulosa cells to convert these androgens to estrogens (Hannon and Flaws 2015). 659 The actual synthetic process (starting with cholesterol in ovarian follicles) is complex and requires the 660 action of several enzymes (see figure 5). 661 662

21 663 664 Figure 5: Ovarian steroidogenesis is primarily conducted by the mature antral follicle and the corpus 665 luteum following ovulation. This process requires both the theca cells and granulosa cells and 666 involves the enzymatic conversion of cholesterol to 17β-estradiol and other necessary sex steroid 667 hormones. Hormones: listed in the white text boxes; steroidogenic enzymes: listed adjacent to the 668 arrows (adapted/copied from Hannon and Flaws 2015). 669 670 Ovarian steroidogenesis is reported to be regulated by RA in several systems (recently reviewed in 671 Damdimopoulou et al. 2019). For example, in rats, low levels of RA and retinol stimulate the 672 formation of progesterone (Bagavandoss and Midgley 1987). In a human ovarian surface epithelium 673 cell line, RA significantly induces production of progesterone (Papacleovoulou et al. 2009). In women, 674 retinol in plasma were associated with higher estradiol and testosterone levels (Mumford et al. 675 2016). It is likely that RA regulates either the expression or the activity (or both) of the steroidogenic 676 enzymes (Wickenheisser et al. 2005), a hypothesis that is supported by data from retinoid-deficient 677 rats (Jayaram et al. 1973). In addition, RA may modulate the pituitary gonadotropins, that in turn 678 regulate steroidogenesis (Minegishi et al. 2000) . 679 680 RA also appears to be one of the critical paracrine factors that stimulate the production of 17β- 681 hydroxysteroid dehydrogenase type 2 (HSD17B2; which catalyzes the conversion of estradiol to 682 estrone) in neighboring epithelial cells, an important mechanism for local estradiol inactivation in the 683 endometrium (Taylor et al. 2015, Jiang et al. 2018a). 684 685 A connection between retinol/RA and ovarian function has been hypothesized based on the 686 observation that serum retinol levels in women vary with the stages of the estrous cycle (highest 687 levels in the proestrus and estrus phases), and that RBP4 serum levels are positively correlated with 688 that of gonadotropin (such as FSH and LH; see figure 6 for the roles of these gonadotropins in the 689 ovarian-estrous cycle) (Jiang et al. 2017). Similarly, ovarian expression of both mRNA and protein 690 levels of RBP4 in adult female mice increased at proestrus and peaked at estrus (Jiang et al. 2018b). 691 In ovaries from mice treated with FSH, increased levels of RA (and other retinoid forms) and RBP4, as 692 well as increased expression levels of RA-producing enzymes (Adh1, Aldh1a1), were observed (Jiang 693 et al. 2018b, Liu et al. 2018). Similar effects were observed in FSH-treated mouse granulosa cells, in 694 addition to increases in e.g. STRA6 (the specific receptor for RBP4) and cellular retinol binding 695 protein, suggesting that the gonadotropin FSH modulates the pathways for retinol uptake and its 22 696 metabolism to RA in the mouse ovary (Liu et al. 2018). 697 698 While Stra6 knockout mice are fertile and overall healthy, this is not the case in humans with 699 mutations in the Stra6 gene (leading to the so-called Matthew-Woods syndrome9), pointing to a clear 700 species difference in functional importance of Stra6 (Szwarc et al. 2018). 701 702 Other female reproductive organs and pregnancy 703 The uterus is the key organ for reproductive processes such as blastocyst implantation and 704 placentation. In the female fetus, the Müllerian ducts differentiate into the oviduct, uterus and 705 vagina, with the resulting epithelia having distinct and separate organ-specific morphology and 706 function (Nakajima et al. 2016). At least in the mouse, RA signaling via RAR during embryo 707 development may determine the fate of Müllerian duct stroma into either uterus or vagina, as seen 708 in vitro with Müllerian ducts treated with either RA (leading to uterine epithelial differentiation) or 709 RAR inhibitors (leading to vaginal epithelial differentiation) (Nakajima et al. 2016). 710 711 The inner lining of the uterus is the endometrium, and its structure and function are regulated by the 712 ovarian sex steroid hormones estradiol and progesterone. In humans, the endometrium undergoes 713 major changes during the menstrual cycle (see figure 6). During the proliferative phase, ovarian 714 estrogens stimulate proliferation and growth, and during the secretory phase, progesterone prepares 715 the endometrium for potential implantation. Following fertilization, the implanting blastocyst first 716 attaches to the epithelial cells and then invades the endometrium by displacing the epithelial cells. It 717 will finally be embedded in the endometrial stroma where the formation of the placenta is initiated 718 (Su and Fazleabas 2015). The details of these processes can be species-specific. 719 720 RA regulates expression of matrix metalloproteinases (MMPs), which are produced by endometrial 721 stromal cells during decidualization (the process of endometrial changes that prepare for pregnancy). 722 RARs are expressed in the uterine stroma of mice and uterine epithelium of rats. In rats, RAR protein 723 expression is influenced by ovarian steroids; RAR expression increases under the influence of 724 estradiol, suggesting involvement of retinoids in growth and proliferation of endometrial epithelia. In 725 postmenopausal women taking estrogen, uterine RAR expression is reported to increase. Increased 726 RAR levels have also been observed in premenopausal women during the proliferative phase; a 727 phase which is associated with elevated estradiol levels (Sayem et al. 2018). 728 729 In the murine uterine epithelium, both mRNA and protein Cyp26a1 are expressed during the 730 blastocyst implantation period. Removing/inactivating Cyp26a1 in mice led to decreased pregnancy 731 rate, and reduced numbers of implantation sites (Han et al. 2010). It appears like the cyp26a1 732 enzyme might block the otherwise inhibitory effects of RA on implantation-related genes in the 733 endometrial epithelium during the murine embryo implantation period (Ma et al. 2012). RA 734 treatment of mouse blastocysts in vitro caused inhibition of cell proliferation and growth retardation, 735 and following implantation, the RA-treated blastocysts were also resorbed to a greater degree than 736 untreated blastocysts (Huang et al. 2005). 737 738 Via the heterodimer peroxisome proliferator-activated receptor (PPAR)/retinoid X receptor (RXR), 739 the retinoid signaling pathway is involved in human placental processes such as the invasion of the 740 uterine epithelium by the extraembryonic trophoblast, or fetal placental, cells. Both trophoblasts and 741 decidual cells appear to be capable of synthesizing RA and (the RXR agonist ligand) 9-cis-RA (Tarrade 742 et al. 2001). At least in vitro, trophoblast invasion is inhibited by PPAR and RXR agonists, while 743 antagonists increased invasion. Functional retinoid signaling pathways have also been demonstrated 744 in human amniotic membranes (Marceau et al. 2006). 745

9 Matthew-Woods syndrome is characterized by anophthalmia or severe microphthalmia, and pulmonary hypoplasia or aplasia (https://www.orpha.net/consor/cgi-bin/OC_Exp.php?Expert=2470&lng=EN). 23 746 747 RA concentration in follicular fluid (obtained during oocyte retrieval during the course of in vitro 748 fertilization) were found to be positively correlated with embryo quality (scored on day 3 after 749 fertilization). RA levels in follicular fluid also correlated moderately with plasma RA. Women with 750 endometriosis had significantly lower concentration of RA in follicular fluid and in plasma than 751 women with no endometriosis (Pauli et al. 2013). 752 753 The genital tubercle is a tissue present during embryonic development of the reproductive system; it 754 later develops into either the glans clitoridis or the glans penis in humans; RA appears to be involved 755 in its development, along with Rarb, Raldh2 and Cyp26b1 (Liu, Suzuki et al . 2012). 756 757 During preimplantation development of the blastocyst, RA signaling appear not to be involved (Rout 758 and Armant 2002). Functional retinoid signaling pathways have been demonstrated in human 759 amniotic membranes (Marceau et al. 2006). 760 761 Female reproductive pathologies and retinoid involvement 762 Endometriosis 763 The endometrium is the mucosal lining of the uterus, which changes in thickness during the 764 menstrual cycle. The endometrium consists of a single layer of glandular epithelial cells resting on a 765 stroma of connective tissue (Jiang et al. 2017). A number of retinoid-related genes (ALDH1A1, 766 ALDH1A2, CYP26A1, CRABP2, RARα, RARγ, and RXRα) is expressed in the endometrium of both 767 rodents and humans, and their expression is regulated by estrogen and/or progesterone (Jiang et al. 768 2017). As reviewed by Jiang et al. 2017, expression of these and other retinoid-related genes have 769 been noted in either stromal or epithelial cells. 770 771 Endometriosis is defined as the presence of endometrial tissue outside of the uterus, usually on the 772 ovaries, fallopian tubes, and peritoneum; a state which can cause severe pain. In addition, 773 fertilization problems have been associated with endometriosis (Taylor et al. 2015, Jiang et al. 774 2018a). 775 776 Several studies have demonstrated altered retinoid pathway signaling associated with endometriosis. 777 In human normal premenopausal endometrial tissue, mRNA expression of the RA-catabolizing 778 enzyme cyp26a1 increases substantially during the progesterone-dominated secretory phase 779 (leading to a resistance to RA signaling) when compared to the estrogen-dominated proliferative 780 phase, during which RA-synthesizing enzymes such as RALDHs were increased (Deng et al. 2003). At 781 least in vitro, RA has been shown to inhibit the decidualization of stromal cells; thus, reduced 782 concentration of RA in the endometrial tissue during the secretory phase could be necessary for a 783 successful implantation (Deng et al. 2003). In the mouse uterus, endogenous RA levels appear to be 784 controlled both by estrogen-dependent expression of RALDH enzymes and by progesterone- 785 dependent expression of CYP26A1 (Fritzsche et al. 2007). Expression of CYP26A1 is down-regulated in 786 both the secretory and proliferative phases in endometrial biopsies from women with moderate or 787 severe endometriosis, when compared to healthy women (Burney et al. 2007). The availability of RA 788 may therefore be increased in the endometrial tissue of women with endometriosis. In 789 endometriosis, there is evidence for progesterone resistance, and cultured stromal fibroblasts 790 originating from endometriotic lesions have decreased ability to decidualize (which is necessary for a 791 successful blastocyst implantation) (Burney et al. 2007). 792 793 Reduced STRA6, CRBP1 and ALDH1A2 expression have been demonstrated to reduce RA in 794 endometrial stromal cells (Jiang et al. 2018a). Transcriptional activation via RA-CRABP2-RAR pathway 795 have been reported to trigger cell cycle arrest and apoptosis; thus, reduced signaling can cause 796 endometrial cells to escape apoptosis and contribute to survival of ectopic cells (Jiang et al. 2018a). 797 Since RA appears to stimulate the production of HSD17B2, which inactivates estradiol in the 798 endometrium, an abnormal RA pathway in endometriosis may explain the aberrant HSD17B2 24 799 expression and the high local estradiol concentrations in endometriosis. In addition, altered retinoid 800 action may cause decreased CX43 expression and GJIC, reducing the decidualization capacity of the 801 stromal cells in endometriosis, which could contribute to progression of endometriosis lesions and 802 the associated subfertile uterine phenotype (Taylor et al. 2015, Jiang et al. 2018a). 803 804 Polycystic ovarian syndrome 805 Polycystic ovarian syndrome (PCOS) is a heterogenous disorder, characterized both by signs of 806 androgen excess and ovarian dysfunction (such as irregular ovulation and/or polycystic ovarian 807 morphology) (Escobar-Morreale 2018). Increased ovarian androgen production from theca cells, 808 elevated LH:FSH hormone ratio, and enlarged ovaries containing many antral follicles is often 809 observed; the associated arrest in follicular growth and anovulation can cause subfertility or 810 infertility (Jiang et al. 2017). 811 812 When comparing the response to RA, 9-cis-RA and retinol in cultured theca cells isolated from 813 normal-cycling women and women with PCOS, only RA led to increased testosterone production 814 (possibly via increased Cyp17 expression) in normal theca cells, while in the theca cells of PCOS 815 patients, all three (RA, retinol, and 9-cis-RA) variants have the same effect (Wickenheisser et al. 816 2005). Thus, in PCOS, theca cells may be sensitized to retinoid signaling stimulation. Moreover, 817 mRNA expression of RA-synthesizing enzymes RoDH2 and ALDH6 was increased in PCOS theca cells 818 (Wood et al. 2003), and PCOS ovaries show enhanced expression of RoDH (Marti et al. 2017). These 819 data suggest an increased rate of RA synthesis in the thecal cells of PCOS women (Jiang et al. 2017). 820 In women of reproductive age with acne and PCOS, treatment with oral isotretinoin decreased 821 ovarian volume (Acmaz et al. 2019). 822 823 Retinoid deficiency/excess in female reproduction 824 As reviewed in Clagett-Dame and Knutson 2011 (citing publications going back to the 1920’s), 825 retinoids are required for successful fertilization, implantation, placentation, embryogenesis and full- 826 term pregnancy: in severely retinoid-deficient female animals, reproduction fails prior to 827 implantation, while in less severe deficiency, fertilization and implantation occur, but embryonic 828 death at mid-gestation is often observed. Retinoid deficiency has adverse effects on placenta 829 morphology (Noback and Takahashi 1978). RA can be supplemented in order to maintain normal 830 fertilization, implantation and early embryogenesis (White et al. 1998), but unless supplementation is 831 sufficiently high by 8.5 dpc, fetuses will be reabsorbed (White et al. 2000). Retinoids are also 832 required for the normal onset of meiosis in the developing embryo (discussed in detail in an earlier 833 section) and germ cells in rat embryos with severe retinoid deficiency fail to enter meiosis (an 834 accompanying decrease in ovarian Stra8 expression was observed); supplementation of small 835 amounts of RA to dams was sufficient to initiate meiosis (Li and Clagett-Dame 2009). Retinoid- 836 deficient mice have a prolonged estrous cycle and also display a decreased rate of oocyte maturation 837 and number of ovulated oocytes after gonadotropin treatment (Kawai et al. 2016). 838 839 Conversely, a state of retinoid excess during pregnancy is known to cause teratogenic effects in 840 humans and animals. The pattern of birth effects, sometimes called retinoic acid syndrome, can be 841 observed in several organ systems, however does not appear to specifically target the female 842 reproductive organs (Azais-Braesco and Pascal 2000, Pennimpede et al. 2010). Retinoid excess may 843 be caused by diet or supplements, but also by administration of synthetic retinoids used for e.g. 844 treatment of severe acne. 845 846 Crosstalk with other signaling pathways 847 Via the heterodimer PPAR/RXR, the retinoid signaling pathway is involved in human placental 848 processes such as the invasion of the uterine epithelium by the extraembryonic trophoblast, or fetal 849 placental cells. Both trophoblasts and decidual cells appear to be capable of synthesizing RA and the 850 RXR agonist ligand 9-cis-RA (Tarrade et al. 2001). At least in vitro, trophoblast invasion is inhibited by 851 PPAR and RXR agonists, while antagonists increased invasion. 25 852 8. Male reproductive organ development, function and health, and the 853 role of retinoids 854 855 The testis produces large numbers of gametes (sperm) throughout the reproductive age of the male 856 and is also the primary source of androgens, required for spermatogenesis and the development and 857 maintenance of male secondary sex characteristics throughout the body (Bittman 2015). 858 859 RA plays several critical roles in spermatogenesis including spermatogonia differentiation and 860 spermiation (i.e. release of spermatids into the lumen of the seminiferous tubuli) (Mark et al. 2015, 861 Teletin et al. 2017), and a facilitating role in spermatocyte meiosis (Teletin et al. 2019). In addition, 862 RA signaling appears to be necessary for proper development of the testis itself (Spade et al. 2018a). 863

864 865 866 Figure 6: The seminiferous tubule and spermatogenesis (adapted from Rato et al. 2012). The 867 seminiferous epithelium is composed of Sertoli cells and developing germ cells at different stages. 868 Leydig cells and blood vessels are located in the interstitium. The supporting Sertoli cells adhere to 869 the basement membrane where spermatogonia are also adherent. Spermatogonia type A divide and 870 develop into spermatogonia type B, which enter meiotic prophase and differentiate into primary 871 spermatocytes that undergo meiosis I to separate the homologous pairs of chromosomes and form 872 the haploid secondary spermatocytes. Meiosis II yields four equalized spermatids that migrate 873 toward the lumen where fully formed spermatozoa are finally released. Abbreviation: BTB, blood– 874 testis barrier. 875 876 Testicular somatic cells and steroidogenesis 877 The bipotential somatic precursor cells in the embryonic testis develop into Sertoli cells, due to the 878 expression of the Y chromosome gene Sry (Minkina et al. 2014). During the embryonic and early 879 postnatal period, Sertoli cells divide mitotically, when the cessation of proliferation is followed by the 880 formation of tight junctions (creating the blood-testis barrier) between adjacent Sertoli cells (Nicholls

26 881 et al. 2013). Sertoli cells together with the germ cells at different stages make up the seminiferous 882 tubule, which is surrounded by peritubular myoid cells and an interstitium containing the androgen- 883 producing Leydig cells (Figure 6; Rato et al. 2012). The production of testosterone by Leydig cells, 884 stimulated by LH, depends on several steroidogenic enzymes, such as CYP11A1, HSD3B1 and 885 CYP17A1 (Jauregui et al. 2018). Testosterone secreted from the Leydig cells regulates the end of 886 meiosis, the establishment and maintenance of the BTB, and spermiation (Jauregui et al. 2018). 887 888 RA signaling, via RAR/RXR, appears to be necessary for development and function of the Sertoli cells 889 (Lucas et al. 2014). In primary rat Sertoli cells isolated on PND 10 and 20, RA suppressed proliferation 890 and initiated tight junction formation (Nicholls et al. 2013). Leydig cells express proteins for RA 891 synthesis, breakdown and signaling (Griswold and Hogarth 2018). Based on impairments observed in 892 VAD mice, the proper differentiation of Leydig cells appears to depend on sufficient retinoid levels 893 (Yang et al. 2018). 894 895 Fertility studies of conditional transgenic adult mice lacking functional RARα in the Leydig cells show 896 that RA signaling via RAR/RXR is required in Leydig cells for normal Leydig cell function and 897 subsequent male fertility; the knockout male mice were unable to produce pups (Jauregui et al. 898 2018). The same conditional knockouts also had changed expressions levels of steroidogenic 899 enzymes in the Leydig cells, increased BTB permeability as well as apoptotic pachytene 900 spermatocytes (Jauregui et al. 2018). The observed phenotype is similar to mice with low or no levels 901 of testosterone (Jauregui et al. 2018). Increased BTB permeability has also been shown in neonatal 902 mice after treatment with the RALDH inhibitor WIN 18,466 (Kent et al. 2016). 903 904 In cultured human fetal testes (ex vivo), addition of RA increased testosterone production and 905 expression of steroidogenic enzymes such as P450 cholesterol side-chain cleavage (P450SCC), 17α- 906 hydroxylase/C17-20 lyase (P450C17), and steroidogenic acute regulatory protein (StAR) (Lambrot et 907 al. 2006). 908 909 Spermatogenesis 910 Spermatogenesis (figure 7) is the formation of haploid gametes (spermatozoa) from the diploid stem 911 spermatogonia, which includes the process of spermatogonia differentiation, meiosis, differentiation 912 of spermatids (spermiogenesis) and spermatid release (spermiation). In the mouse, after birth, the 913 mitotically quiescent gonocyte (also termed pro-spermatogonia) re-enter the cell cycle at 914 approximately PND 1-2, turn into spermatogonia at PND 3-6 as they migrate to the periphery of the 915 testis cords. Within the testis cord, the spermatogonia are flanked by Sertoli cells, and the 916 peritubular myoid cells, surrounding the outside of the cord (Teletin et al. 2017, Teletin et al. 2019). 917 During this first post-natal week, a sub-population of gonocytes differentiates directly (without 918 passing through the undifferentiated spermatogonia stages) into differentiating spermatogonia that 919 support the so-called first wave of pubertal spermatogenesis. After the first wave of spermatogenesis 920 (completed by postnatal day 35 in the mouse), subsequent waves derive from the undifferentiated 921 spermatogonia that has self renewal capacity (Yoshida et al 2006).

27 922 923 Figure 7 (adapted from Mark et al. 2015): Schematic representation of adult spermatogenesis in the 924 mouse. In mouse spermatogonia in the single-cell state (AS) are considered as the true stem cells of 925 the spermatogenic lineage. In the first (proliferative) phase of spermatogenesis, As spermatogonia 926 divide to maintain a stem cell population (As) and expand the population of cells (Apr and Aal4–16; ) 927 which enter the differentiation pathway (Stages A1–A4, Int and B, collectively referred to as 928 differentiating spermatogonia (the previous stages, As – Aal16, are collectively also referred to as 929 undifferentiated spermatogonia). In the second, meiotic, phase , spermatocytes undergo 930 recombination and segregation of homologous chromosomes during the first two meiotic divisions 931 (M1 and M2) to generate step 1 spermatids (St1). The last phase, spermiogenesis, is subdivided into 932 16 steps based on morphological criteria (round spermatids, steps 1 to 8; elongating spermatids, 933 steps 9 to 16). The 12 first steps span the entire cycle of the seminiferous epithelium. Step 16 934 spermatids are released into the lumen of the seminiferous tubules as spermatozoa, during a process 935 called spermiation. 936 Abbreviations: PR, preleptotene; L, leptotene; Z, zygotene; P, pachytene; D, diplotene [the stages of 937 the first meiotic prophase]; 2S, secondary spermatocytes. 938 939 The timing of the progenitor cell commitment and subsequent differentiation and maturation of 940 spermatogonia along the seminiferous tubule is staggered, forming a spermatogenic wave (Griswold 941 2016). The cycle of the seminiferous epithelium (SE) is initiated by the precisely timed transition of A 942 (undifferentiated) spermatogonia into A1 spermatogonia, which subsequently will divide to generate 943 successively A2, A3, and A4 (differentiating) spermatogonia (Teletin et al. 2019). 944 945 The length of this wave is species-specific, e.g., 8.6 days in mice, 13 days in rats and 16 days in 946 humans (Bittman 2016). The spermatogonia continue to divide mitotically to produce cells that 947 replenish the stem cell pool and cells that undergo a series of mitotic divisions to increase the 948 number of spermatocytes, that then progress through meiotic sub-stages. In addition, it takes 8.6

28 949 days for the A1 spermatogonia to become preleptotene spermatocytes (green) and enter meiosis 950 and an additional 8.6 days ×3 to form elongated spermatids ready for spermiation. The net result is 951 that once fully established, the same cell associations or the same group of cell types appear every 952 8.6 days. In the 8.6-day period between the transition of A spermatogonia to A1 spermatogonia, 953 there is a continuum of development of each cell type. 954 955 RA is considered to be critically involved in several steps of spermatogenesis: Spermatogonia 956 differentiation specifically during Aal-A1 transition, for spermiation and in regulation of the SE cycle 957 (Teletin et al. 2019). RA is indispensable in vivo to trigger the Aal – A1 transition (Teletin et al. 2019) 958 and is also required for the survival of some Aundiff spermatogonia. Just as for oocytes, RA plays a role 959 in the entry of the spermatocytes into meiosis (Raverdeau et al. 2012, Teletin et al. 2019), possibly 960 via control of replication-dependent core histone gene expression necessary for entry into S phase 961 (Chen et al. 2016). In the developing and adult testis, spermatogonia that are destined for meiosis 962 express Stra8 about a week before meiotic entry. Synthesis of both mRNA and protein Stra8 occurs 963 exclusively in two distinct phases of spermatogonia differentiation; in differentiating type A 964 spermatogonia (A1-A4), as well as in spermatocytes in the preleptotene (PR) and leptotene (L) phase 965 at meiosis entry.. At least in the mouse, RA is also required to disengage spermatozoa from the 966 Sertoli cell cytoplasm during spermiation (Spiller and Bowles 2015, Teletin et al. 2017). 967 968 During the onset of puberty in mice, sperm production is initiated by RA produced in Sertoli cells, 969 since these are the only cells in the seminiferous epithelium with RALDH activity; spermatogonia do 970 not possess the enzymatic machinery to produce RA (Ref. to be inserted). The RA produced in the 971 Sertoli cells acts in a paracrine manner on the spermatogonia, to regulate the Aal – A1 transition. 972 Spermatogenesis is blocked in juvenile male mice deficient in RDH10 in Sertoli cells and germ cells, 973 leading to a local lack of RA (Tong et al. 2013). Once the first wave of spermatogenesis has 974 progressed beyond a certain point, it appears (based on data from mice in which all three Aldh1a1 975 genes were deleted in Sertoli cells) that RALDH activity in Sertoli cells is no longer indispensable for 976 the spermatogonia differentiation to proceed. Thus, RA must be produced elsewhere, most likely by 977 RALDH2 in spermatocytes (preleptotene spermatocytes appear at every stage VIII of the cycle) and 978 spermatids. RA acts in a paracrine manner, inducing spermatogonia differentiation from the second 979 round of spermatogonia differentiation onward. The endogenous RA levels in the seminiferous 980 epithelium are tightly regulated. Pulses of increased levels of RA move along the tubule correlated 981 with the appearance of stages VIII-IX of the seminiferous cycle. In addition, the expression of the RA 982 target genes Stra8 (in spermatocytes) and Stra6 (in Sertoli cells) peak at stages VII-VIII of the SE cycle 983 (Teletin et al. 2017, Griswold and Hogarth 2018). However, at least in adult mice, ALDH enzymes 984 (ALDH1A1-3 and ALDH8A1) were not expressed in a stage-specific manner (Kent et al. 2016). In the 985 postnatal mouse testis, there appears to be other enzymes than ALDH1A isozymes that can oxidize 986 retinal to RA, since severe deficiency (obtained via genetic approaches) of the Aldh1a2 gene, 987 considered to be the major isomer involved, had no adverse effects on male fertility or health 988 (Beedle et al. 2019). In contrast, studies using human testis biopsies suggest that ALDH1A1 is the 989 predominant form that produces RA in the Sertoli cells, while ALDH1A2 has the corresponding role in 990 the developing sperm (Arnold et al. 2015). 991 992 For the regulation of spermatogonial exposure to RA, two hypotheses has been presented: 1) all 993 spermatogonia are primed to respond to RA, but the exposure to RA is periodic and tightly 994 controlled, and 2) all spermatogonia are exposed to RA but only some can respond (Busada and 995 Geyer 2016). The differential expression of RARγ between spermatogonia subpopulations could 996 explain how the As spermatogonia (with no RARγ expressed) remain undifferentiated and maintain 997 self-renewal capabilities even in the presence of high RA concentrations, which induce differentiation 998 in the (RARγ-expressing) Aal progeny populations (Teletin et al. 2017). Observations from cell-specific 999 conditional knockout studies of RA receptor isomers suggest that the RA signal in spermatogonia is 1000 transduced via RAR/RXR heterodimers (Gely-Pernot et al. 2015). Another suggestion is that RA 1001 signaling also operates via Sertoli cells, as germ cell differentiation can proceed in the absence of 29 1002 functional RAR and RXR isotypes in spermatogonia (Teletin et al. 2017). RA may also act via non 1003 receptor-mediated pathways, such as via kinase singaling (Busada and Geyer 2016). 1004 1005 Spermatogonia exposure to RA may also be regulated by the presence of the RA-catabolizing CYP26 1006 enzymes. All three isoforms of CYP26 family are present in the murine postnatal testis, although 1007 observations after conditional knocking-out of the A1 and/or B1 genes in germ and/or Sertoli cells 1008 reveal that CYP26B1 is the critical isoform (Hogarth et al. 2015). In the fetal testis, elimination of RA 1009 by CYP26B1 is necessary not only to avoid meiosis, but also for normal mitotic arrest of male PGCs as 1010 well as for avoiding germ cell apoptosis (Rossitto et al. 2015). CYP26B1 is initially produced by the 1011 Sertoli and Leydig cells and/or interstitial somatic cells, although after birth, Cyp26b1 transcripts are 1012 confined to the peritubular myoid cells (Rossitto et al. 2015). CYP26A1 and CYP26B1 form a catabolic 1013 barrier against any RA present in the immediate environment of the seminiferous tubules. No 1014 variations in the expression of CYP26 enzymes exist across the SE cycle, unlike several proteins 1015 involved in retinoid storage (e.g. LRAT) and RA production (RALDH1 in Sertoli cells and RALDH2 in 1016 germ cells) which are regulated in a periodic manner along the seminiferous tubule (Teletin et al. 1017 2017). 1018 1019 Differential expression of CYP26B1/A1 in their microenvironment, rather than differential expression 1020 of RARs or RXRs within stem and non-stem spermatogonia, has been suggested as the mechanism for 1021 protecting the stem spermatogonia from differentiation in response to RA signaling (Lord et al. 1022 2018). In neonatal mice, heterogeneity in terms of responsiveness to RA in the gonocytes (pro- 1023 spermatogonia) was observed: 1) undifferentiated spermatogonia capable of responding to RA (due 1024 to RARγ being expressed) can either respond to low levels of RA or can only respond when higher 1025 levels of RA are added (the lack of response to low levels was hypothesized to be due to CYP26 1026 catabolism, and 2) undifferentiated spermatogonia that cannot respond to RA (due to lack of RARγ 1027 expression) and are destined to become foundational spermatogonial stem cells (Velte et al. 2019). It 1028 is also of note that RA deficiency may cause undifferentiated spermatogonia to retain their stem cell 1029 state (Agrimson et al. 2017). Thus, differential expression of RARγ and CYP26 are important for 1030 proper maintenance of spermatogonia differentiation. 1031 1032 Knockout studies in mice show that despite the critical role of RA in spermatogonia differentiation 1033 and in male germ cell meiosis, both differentiation and meiosis can occur in germ cells where RAR or 1034 RXR are absent, although a fraction of the A1 spermatogonia are adversely affected (Gely-Pernot et 1035 al. 2015). Initiation and progression of meiosis also proceeded in these mice. The conclusion was 1036 that the RA signaling pathway was not autocrine, but rather operated in Sertoli cells (Gely-Pernot et 1037 al. 2015). 1038 1039 In male mice with arrested spermatogonia differentiation, due to either vitamin A deficiency or 1040 genetic knockout of Aldh1a1 in their Sertoli cells), a single injection of RA leads to resumed Aal – A1 1041 transition, with preleptotene spermatocytes expressing Stra8 and Rec8 appearing synchronously, 8 1042 days post-RA injection, in all seminiferous tubules. The same effect is observed following treatment 1043 with the RALDH inhibitor WIN 18,446, with a subsequent RA injection (Hogarth et al. 2013). RA 1044 administration in vitamin A-deficient mice or rats results in truncation of the normal 12 (mice) or 14 1045 (rat) stages of the seminiferous epithelium cycle to only three or four stages (still enabling normal 1046 sperm production). Additionally, the normally asynchronous wave of spermatogenesis is 1047 synchronized, with near-identical timelines of the differentiation progression in all seminiferous 1048 tubules, leading to a pulsatile rather than continuous sperm production (Teletin et al. 2017). 1049 1050 Recent findings, reviewed in Teletin et al. 2019, suggest that RA is not decisive to meiotic initiation in 1051 male PGCs, but merely facilitates this process. Data indicates that meiosis is not blocked, but slightly 1052 delayed, in mice lacking all RALDH in the SE. Leptotene spermatocytes in Aldh1a1-3Ser−/− and Aldh1a1- 1053 3Germ−/−;Ser−/− mutants appear 24 h later than their counterparts in control mice, but express STRA8 1054 and subsequently progress to the pachytene stage. The strong effect of BMS-204,493 on STRA8 30 1055 expression and meiotic initiation is probably due to BMS-204,493 stabilizing the binding of RAR to co- 1056 repressor complexes that would cause transcriptional silencing at the Stra8 locus through histone 1057 deacetylation and chromatin condensation (Teletin et al. 2019). 1058 1059 The role of Stra8 gene expression in normal spermatogenesis is supported by an observation in 1060 humans: an association was found in men with a specific single nucleotide polymorphism in their 1061 Stra8 gene with spermatogenic impairment evident as azoospermia or oligozoospermia (Lu et al. 1062 2013). 1063 1064 Secondary male reproductive organs 1065 The prostate produces slightly alkaline prostate fluid that makes up approximately 30% of the 1066 volume of semen in humans and is critical for male reproductive health (Verze et al. 2016). The 1067 retinoid signaling pathway has several functions in differentiation and maintenance of secondary 1068 male reproductive organs. Presence of RA is necessary for prostate formation from the urogenital 1069 sinus during sexual differentiation (Vezina et al. 2008, Bryant et al. 2014). Genetic deletion studies in 1070 mice have also demonstrated critical roles for RA via Rarγ in prostate and in seminal vesicles (Lohnes 1071 et al. 1993). 1072 1073 RA signaling appears to be necessary for the function of epididymis, into which sperm are released 1074 after spermiation, undergo maturation, and are stored until ejaculation (Jauregui et al. 2018). In a 1075 conditional transgenic model with a dominant negative form of RARα expressed in Leydig cells and in 1076 the epididymis, the resulting abnormal epididymis phenotype may contribute to the infertility 1077 observed in these mice (Jauregui et al. 2018). These observations support previous findings that lack 1078 of RA signaling in the epididymis results in squamous metaplastic epididymal epithelium; 1079 alternatively, the abnormal epididymal phenotype may have been due to lack of testosterone 1080 (Jauregui et al. 2018). 1081 1082 Male reproductive pathologies 1083 Impaired RA signaling may play a role in several male reproductive pathologies. The transition from 1084 PGCs to differentiating spermatogonia is impaired in the absence of RA in the testis cords (Teletin et 1085 al. 2017).PGCs that fail to differentiate into spermatogonia may be the source of carcinoma in situ, 1086 which can develop into testicular germ cell cancer in humans (Busada and Geyer 2016, Teletin et al. 1087 2017). 1088 1089 In prostate tumor tissue, RA concentrations are lower than in normal prostate tissue, and since it has 1090 been hypothesized that this is due to increased RA catabolism, inhibitors of CYP26 enzymes, called 1091 RA metabolism blocking agents (RAMBAs) have been used in the treatment of prostate cancer 1092 (Nelson et al. 2013). RAMBA therapy promotes differentiation and inhibits proliferation by increasing 1093 endogenous RA in tumors (Denis et al. 1998). Some RAMBAs also inhibit estrogen synthesis via 1094 inhibition of aromatase (CYP19) and testicular androgen synthesis via inhibition of 17,20-lyase 1095 (CYP17) (Bryson and Wagstaff 1996). The ALDH1A2 enzyme expression is also altered in prostate 1096 cancer. The expression of the Aldh1a1 gene is lower and its promoter region is hypermethylated in 1097 epithelia from malignant prostate tumors (Kim et al. 2005). 1098 1099 Cryptorchidism, or non-descended testis, is a common congenital malformation that also leads to 1100 increased risk of testicular cancer and infertility in adult life (Bay et al. 2011). Complete testicular 1101 descent is necessary for normal testicular function in adult males, and the process of testis descent is 1102 regulated by the Leydig cell hormones insulin-like peptide 3 (INSL3) and testosterone (Bay et al. 1103 2011). The spectrum of reproductive system malformations caused by VAD in rats includes 1104 cryptorchidism (See et al. 2008). Both RA levels and Stra8 expression were significantly lower in the 1105 rat cryptorchid testis (induced by in utero exposure to the anti-androgen flutamide) compared to the 1106 normal testis (Peng et al. 2016). In an in vitro system, RA upregulated the expression of the gene 1107 LGR8, which encodes for the receptor for INSL3 (Klonish et al. 2005). 31 1108 1109 Retinoid deficiency in male reproduction 1110 The importance of narrowly regulated retinoid levels on male reproduction can be illustrated in 1111 animal studies measuring effects of experimentally increased or decreased levels of retinoids. 1112 Adverse effects on the male reproductive system following experimental limitations of RA have been 1113 demonstrated in multiple in vivo studies (discussed in previous sections of this report and reviewed 1114 in Clagett-Dame and Knutson 2011). In rodents maintained on a vitamin A-deficient diet, testicular 1115 degeneration with impaired spermatogenesis and a complete disappearance of all meitotic and 1116 postmeiotic cells was observed (Coward et al. 1969, Morales and Griswold 1987, van Pelt and de 1117 Rooij 1990). Spermatogenesis is arrested at the preleptotene spermatocyte stage in VAD rats and at 1118 the spermatogonia stage in VAD mice; however, spermiation failure is observed in both species 1119 under VAD conditions (Ghyselinck et al. 2006). Vitamin A deficiency also leads to replacement of 1120 normal glandular epithelium in the epididymis, prostate and seminal vesicles, by a stratified 1121 squamous keratinizing epithelium, resulting in inhibited seminal fluid production (Wiseman et al. 1122 2017). The RBP4-/- mouse model phenotype is similar to VAD, including the inability to mobilize 1123 hepatic retinoid stores into circulation, progressive testicular degeneration together with 1124 spermatogonia differentiation arrest and failure of spermiation (Ghyselinck et al. 2006). 1125 1126 In human populations with insufficient circulating retinoids, symptoms such as blindness is well- 1127 known, but there appears to be no information on reproductive parameters such as sperm 1128 production or fertility (Hogarth and Griswold 2010). However, observations made on testicular tissue 1129 samples obtained in different clinical situations, suggest a correlation between adverse reproductive 1130 parameters and an affected retinoid system. A decrease in RA and 13-cis-RA was measured in testis 1131 tissue from men with abnormal sperm production undergoing scrotal surgery (Nya-Ngatchou et al. 1132 2013). In another study, lower levels of the RA-producing RALDH1A2 enzyme in testicular tissue were 1133 associated with male infertility (Amory et al. 2017). 1134 1135

32 1136 9. Impact on female and male reproduction by chemicals acting via the 1137 retinoid system 1138 1139 Exposure to chemicals early in life may affect both female and male reproductive organs and their 1140 functions, most likely via a large number of mechanisms, with effects via estrogen and/or androgen 1141 pathways being most frequently discussed. A testicular dysgenesis syndrome (Skakkebaek et al. 1142 2001) as well as an ovarian dysgenesis syndrome (Johansson et al. 2017) have been described. Both 1143 dysgenesis syndromes are defined as early (fetal) alterations in testicular or ovarian structure or 1144 function that cause an impairment of reproductive parameters in adulthood. One of the mentioned 1145 fetal alterations in the ovarian dysgenesis syndrome was disruption of the RA-dependent meiosis 1146 initiation (Johansson et al. 2017). Also for embryonic males, it has been suggested that testicular 1147 toxicity could result from disruption of local RA homeostasis or signaling (Spade et al. 2018b). In the 1148 adult organism, chemicals may interfere with e.g. normal sperm production (Sharpe 2010) and 1149 normal function and morphology of female reproductive tissue. 1150 1151 Some observed effects and mechanisms by which chemicals could interfere with the retinoid 1152 pathway are summarized in table 1 (and also in e.g. Nilsson and Håkansson 2002, Novak et al 2008, 1153 and Shmarakov 2015). In brief, chemicals have been shown to deplete tissue retinoid levels by 1154 affecting retinoid metabolism, also after in utero exposure. Chemicals can also cause activation or 1155 inactivation of retinoid receptors. As described in this report, the retinoid pathway is important, or 1156 even critical, in several aspects of both female and male reproduction and during fetal development. 1157 It is thus conceivable that chemicals capable of interfering with the retinoid pathway may, as a result, 1158 cause adverse effects on reproductive parameters, assuming that such chemicals reach target cells 1159 during critical windows. 1160 1161 Under the course of this project, very few articles have been retrieved that focus on effects of 1162 contaminants on the female reproductive system while simultaneously looking at retinoid 1163 parameters such as e.g. RA-synthesizing/metabolizing enzymes. For similar studies on the male 1164 reproductive system, there are a slightly higher number of available references. There are also 1165 publications were chemicals have had adverse effects on reproductive parameters while 1166 simultaneously causing decreased liver retinoid stores; however, no retinoid-related parameters 1167 were measured in the reproductive organs (see e.g. van der Ven et al. 2008 and 2009 in table 1). 1168 While no conclusions on any causal role of retinoid perturbation in the adverse reproductive effects 1169 can be made from such studies, causality cannot be excluded. It should be noted that chemical- 1170 induced decreases of hepatic retinoid levels appear to be accompanied by decreased testicular levels 1171 (Håkansson et al. 1991). 1172 1173 While most studies use animals, there is human data as well. E.g., in the 1960’s, it was demonstrated 1174 that the compound WIN 18,446 could reversibly inhibit spermatogenesis in men (Heller et al. 1961). 1175 More recently, the same compound was shown to inhibit the conversion of retinal to RA, most likely 1176 by inhibiting ALDH1A2 (Paik et al. 2014), and this finding has led researchers to suggest inhibition of 1177 retinoid metabolism as an approach to male contraception (Hogarth et al. 2011). In males, treatment 1178 with 13-cis-RA (used in the treatment of e.g. acne) can improve sperm production (Cinar et al. 2016, 1179 Amory et al. 2017). 1180 1181

33 1182 Table 1: Examples of chemicals interfering with the retinoid system in different models. Chemical(s) Model system Endpoint Observed effect Reference 2,3,7,8- Male rats, single Retinoid levels in several ↓Retinyl esters in liver, testes, Håkansson et al. tetrachlorodibenzo-p- dose organs, including testes epididymis 1991 dioxin (TCDD) and epididymis ↑Retinyl esters in kidney 2,3,7,8- Pregnant rats and Retinoid levels in liver, ↓Retinyl esters in maternal and Kransler et al. 2007 tetrachlorodibenzo-p- PND 7 pups, lung and kidney perinatal liver and lung dioxin (TCDD) single gestational ↑Retinyl esters in maternal and dose perinatal kidney 2,3,7,8- Male rats, single RA and retinyl ester ↑ RA in liver and kidney Hoegberg et al. tetrachlorodibenzo-p- dose levels and LRAT ↓Retinyl esters in liver 2003 dioxin (TCDD) expression in liver and ↑Retinyl esters in kidney kidney ↑ LRAT in kidney 2,3,7,8- Male rats, single RA metabolism in ↑ RA metabolism in all Fiorella et al. 1995 tetrachlorodibenzo-p- dose microsomes from liver, investigated tissues dioxin (TCDD) lungs, kidneys, testes Technical Female and male Retinoid levels in liver ↓Retinyl esters in liver van der Ven et al. pentabromodiphenyl rats (enhanced 2008a ether mixture 28d TG407 study) Hexabromocyclodecane Female and male Retinoid levels in liver ↓Retinyl esters in liver van der Ven et al. rats (enhanced 1- 2009b gen TG415 study) Fluconazole Mice (exposed mRNa induction of Upregulation of CYP26A1 and Tiboni et al. 2009 during pregnancy) CYP26A1 and CYP26B1 CYP26B1 Organ development (not Abnormal branchial arch repro) development. Organochlorine Reporter cell lines Activation of RARs Activation was observed. Lemaire et al. 2004 pesticides (chlordane, (RARE & RARα, β, dieldring, aldrin, γ) endrin, endosulfan) CYP26A1 Induction of CYP26A1 CYP26A1 induction was observed. 543 environmental Yeast cells RARγ agonistic activity 85 of the 543 chemicals had RARγ Kamata et al. 2008 chemicals transfected with agonistic effects (especially the human RARγ monoalkylphenols and styrene dimers). 309 environmental HepG2 cells Transcription factor Max responders were lindane, Martin et al. 2010 chemicals transfected with activity oxadiazon and imazalil cis/trans-reporter transcription units for RARα, β, γ 28 environmental and Human uterus Chemical displacement; A number of the chemicals could Paganetto et al. other compounds and prostate inhibition of 3H-RA displace 3H-RA. MEHP most 2000 cytosol binding to cytosol potent. PPARγ agonists HepG2 cells mRNA induction of mRNA induction of CYP26B1 and Tay et al. 2010 rosiglitazone and CYP26B1 and CYP26A1 (to a lesser extent) CYP26A1 pioglitazone Phenobarbital Human mRNA induction of Weak induction of CYP26B1 Finkelstein et al. hepatocytes CYP26B1 2006 Tributyl tin (TBT) Human and mice Cholesterol homeostasis Impaired cholesterol homeostasis Pu et al. 2019 ovarian theca via RXR pathway cells 1183 a Also observed in males: decreased weight of epididymis, seminal vesicles and prostate, sperm head deformities; in 1184 females: induced adrenal activity of the steroidogenic CYP17 enzyme 1185 b Also observed in males: decreased weight of testis 1186 1187 It should be pointed out that effects of chemicals on retinoid homeostasis can be indirect, since some 1188 metabolic enzymes are used both for detoxification and for retinoid homeostasis, such as 1189 RALDH/ALDH class 1 (Alnouti and Klaassen 2008). Induction of Aldh isomers (as studied in mouse 1190 livers) after in vivo administration of several activators of different nuclear receptors (such as the 1191 constitutive androstane receptor CAR, pregnane X receptor PXR, PPAR, AhR) was isomer- and 1192 activator-specific (Alnouti and Klaassen 2008). The well-known induction of cytochrome P450 1193 enzymes via AhR after exposure to dioxins and dioxin-like polychlorinated biphenyls (PCBs) may be 1194 relevant since some of these P450 enzymes are believed to be involved in either synthesis or 1195 oxidation of RA (Murphy et al. 2007). If the end result is altered RA concentrations in fetal or adult 1196 reproductive organs, correct development and/or function of these organs could be compromised. 1197

34 1198 Since RXR can heterodimerize with several other nuclear receptors, it is also conceivable (and has 1199 indeed been shown; see e.g. Tarrade et al. 2001) that any interference with the retinoid system can 1200 also have effects on other signaling pathways, via e.g. PPAR, pregnane X receptor (PXR), constitutive 1201 androstane receptor (CAR) and vitamin D receptor (VDR). Also, interactions with AhR signaling 1202 pathways on several levels have been shown (Murphy et al. 2007, Vezina et al. 2009). From a 1203 functional/endpoint perspective, crosstalk with the other endocrine systems (estrogen, androgen, 1204 thyroid) is important to consider. 1205 1206 1207 Chemicals affecting female reproduction 1208 1209 Analgesics: 1210 Exposure to the analgesic chemicals acetaminophen (a.k.a. paracetamol) and indomethacin 1211 appeared to give rise to delayed entry into meiosis in female rats after in utero exposure, with 1212 changes in ovarian Stra8 levels reflecting this delay (Dean et al. 2016). Other effects, not currently 1213 linked to Stra8 or RA, included decreased fetal ovarian germ cell numbers and (in adult females 1214 exposed in utero) reduced ovarian size and female fertility (measured as number of pups per litter) 1215 were observed (Dean et al. 2016). 1216 However, in a mouse study with in utero exposure to paracetamol, ovarian Stra8 levels were not 1217 affected, while in the same females, germ cell numbers decreased (not a Stra8-related effect) (Holm 1218 et al. 2016). The diverging results may be due to e.g. differences in species and/or dose or time of 1219 exposure (350 mg/kg on days 13.5-21 in the Dean study (Dean et al. 2016); 50 or 150 mg/kg on days 1220 7 until birth in the Holm study (Holm et al. 2016)). 1221 1222 Bisphenol A (BPA): 1223 In utero exposure to BPA in mice has been associated with a delay in meiotic prophase I, 1224 hypothesized to be due to decreased expression of Stra8 in the fetal oocytes (Zhang et al. 2012). In 1225 the same mice at postnatal day three, increased BPA exposure levels were associated with more 1226 oocytes in germ cell cysts and fewer oocytes in primordial follicles (Zhang et al. 2012). One suggested 1227 mechanism for how BPA affects Stra8 expression is increased DNA methylation, meaning that the 1228 effect of BPA appears to be independent of any direct effect on retinoid metabolism or homeostasis, 1229 although it exerts (one of) its effects on (the retinoid-regulated gene) Stra8 (Zhang et al. 2012). 1230 Not all data support an effect of BPA on Stra8. In human fetal oocytes cultured in BPA-containing 1231 media, the expression pattern of Stra8 was similar to that in control cultures (Brieno-Enriquez et al. 1232 2012. In fetal ovaries originating from pregnant mice exposed to BPA, the observed increase in Stra8 1233 expression did not differ from that in unexposed mice (Lawson et al. 2011). 1234 1235 Phthalates: 1236 Only a few publications were retrieved in which retinoid-related endpoints were investigated 1237 following DEHP exposure. DEHP caused a delay in meiosis in mouse fetal germ cells following in utero 1238 exposure; the concomitant decrease in mRNA and protein expression of Stra8 was suggested to be 1239 related (Zhang et al. 2015). The DNA methylation level of Stra8 in oocytes of the F1 generation was 1240 increased as a result of DEHP exposure, and these changes were inherited by the F2 generation 1241 (Zhang et al. 2015). 1242 DEHP may also exert effects via PPAR-RXR. In vitro, the DEHP metabolite monoethylhexyl phthalate 1243 (MEHP) suppressed aromatase, the rate-limiting enzyme for conversion of testosterone to estradiol, 1244 expression via activation of PPAR-RXR heterodimers in rat ovarian granulosa cells (Lovekamp-Swan et 1245 al. 2003). 1246 Phthalates can target many aspects of ovarian development and normal function (reviewed by 1247 Hannon and Flaws 2015); the exact mechanisms are unknown. 1248 1249

35 1250 Organotins: 1251 In marine gastropods, tributyltin, used in antifouling paints for ships and fishing nets, induces 1252 imposex (the development of male genitals in females) via binding RXR and thereby activating the 1253 RXR-RAR heterodimer (Nishikawa et al. 2004). 1254 In human placental choriocarcinoma cells, trialkyltins stimulate human chorionic gonadotropin 1255 production and aromatase activity by acting as RXR agonists (Nakanishi et al. 2005). 1256 The RXR signaling pathway may be involved in the observed increase in progesterone production in 1257 human placental cells (in vitro) following organotin exposure (Macejova et al. 2016). 1258 No RXR activation was observed in rodent or human placenta tissue in which ng-mg/kg levels of 1259 organotin levels were present (de Araújo et al. 2018). 1260 In ovarian theca cells from humans, mice and other mammalian species, TBT stimulated cholesterol 1261 extracellular efflux via the RXR pathway (Pu et al. 2019). 1262 1263 R115866: 1264 This triazole-containing molecule is a CYP26 inhibitor (more potent than liarozole and largely without 1265 the inhibitory ability of liarozole on the CYP-dependent formation of estradiol and testosterone) 1266 (Stoppie et al. 2000). R115866 administration leads to increased endogenous levels of RA, with 1267 subsequent RA-like effects such as inhibition of vaginal keratinization in estrogen-stimulated rats 1268 (Stoppie et al. 2000). 1269 1270 Isotretinoin/13-cis-RA: 1271 This pharmaceutical retinoid, used for e.g. treatment of acne has been found to lead to reduced 1272 antral follicle count, ovarian volume and levels of anti-Müllerian hormone (AMH; a marker of ovarian 1273 follicle number) in humans, suggesting negative effects on the ovarian reserve (Aksoy et al. 2015). 1274 Similar effects have been observed in rats (Abali et al. 2013). The effects appear to be transient both 1275 in humans (Cinar et al. 2016) and in rats (Korkmaz et al. 2017) once treatment ceases. Isotretinoin 1276 administered to rats at doses up to five times higher than clinical doses used for acne treatment 1277 reportedly had no adverse effects on fertility, conception rate, gestation or parturition, as 1278 summarized in an US FDA Pharmacology Review of isotretinoin10. The ICH reproductive toxicology 1279 guideline for registrations of pharmaceuticals for human use does not require histopathological 1280 examination of ovaries in reproductive toxicity studies and therefore it is not possible to draw any 1281 firm conclusions on possible effects of isotretinoin on ovarian volume or follicle count from the US 1282 FDA summary of this study. 1283 1284 Chemicals affecting male reproduction 1285 1286 Thiocarbamate herbicides: 1287 Molinate, a known testicular toxicant in the rat, is also an inhibitor of RALDH and has been shown to 1288 inhibit the conversion of retinal to RA; decreased testicular levels of RA were observed in rats dosed 1289 with molinate (Zuno-Floriano et al. 2012). 1290 1291 Conazoles: 1292 Conazole fungicides are triazole-based compounds used both in agriculture and pharmacologically, 1293 and which exert their fungicidal effects via broad inhibition of CYP enzymes. In vertebrates, effects 1294 include inhibition of CYPs involved in vertebrate sex steroid synthesis as well as hepatic xenobiotic 1295 and hepatic metabolism (Chen et al. 2009, Tonk et al. 2015). Reprotoxic effects have been observed 1296 following in vivo exposure (Vickery et al. 1985, Taxvig et al. 2007). 1297 In the pharmaceutical industry, structurally related compounds such as liarozole and talarozole have 1298 been considered for treatment of some cancers and dermatological diseases (Stevison et al. 2017). 1299 One mechanism of action of these compounds is the blocking of RA catabolism via inhibition of

10 https://www.accessdata.fda.gov/drugsatfda_docs/nda/2012/021951Orig1s000PharmR.pdf, downloaded Sep 4, 2018. 36 1300 CYP26. Transient increased testicular RA levels have been observed in mice given talarozole (Stevison 1301 et al. 2017). In VAD mice administered liarozole after a dose of RA, the RA-induced proliferative 1302 effects on A spermatogonia was less than in VAD mice given RA but not liarozole (Gaemers et al. 1303 1997). However, to date, no data are available indicating causal relationships between conazole 1304 exposure, altered retinoid synthesis, and impaired male reproduction 1305 1306 Bisphenol A: 1307 In mice, neonatal exposure to bisphenol A has been shown to cause decrease in the number of 1308 sperm and damage to the motility and morphology of sperm in adult mice, and the effect was, to 1309 some extent, ameliorated by vitamin A supplementation and aggravated under vitamin A-deficient 1310 conditions (Nakahashi et al. 2001, Aikawa et al. 2004). The authors suggested that the development 1311 and functional differentiation of the reproductive tracts and the gonads may be controlled by a 1312 balance between levels of estrogen and retinoids. 1313 In mouse embryonic stem cells, bisphenol A was found to upregulate the expression of Stra8 1314 (amongst other genes) in a manner that appeared to be consistent with a feminizing effect, but 1315 bisphenol A seemed to act via a non-RA and non-RAR mechanism (Aoki et al. 2012). 1316 1317 Triorganotin compounds: 1318 Triorganotins are considered endocrine-disrupting compounds and have been shown to bind RXR 1319 (Ref to be inserted). Macejova et al. (2016) reviewed effects of triorganotin compounds on male 1320 reproductive organs in rats, but it was unclear if these effects were the result of a disrupted retinoid 1321 pathway. 1322 1323 Phthalates: 1324 Some phthalate esters are anti-androgenic and thus considered potential causing agents in the 1325 “testicular dysgenesis syndrome” (Skakkebæck et al. 2001). These compounds have now been shown 1326 to also interfere with RA synthesis, both in vitro (Chen and Reese 2016) and ex vivo (Spade et al. 1327 2018b). It was suggested (in Chen and Reese 2016) that a dual mechanism of both anti-androgen and 1328 retinoid disruption may lead to developmental effects in humans and rodents. 1329 1330 Hexachlorocyclohexane: (the gamma isomer is known under the name of lindane) 1331 Lindane administration caused atrophy of the epididymidis and seminal vesicles, along with 1332 decreased sperm count in the epididymis and reduced activities of steroidogenic enzymes, in VAD 1333 rats but not in rats given sufficient amounts of retinoids in their diet (Pius et al. 1990). 1334 1335 BMS-189453: 1336 BMS-189453 is a synthetic retinoid RARα, β and γ antagonist. Following oral administration to rats 1337 and rabbits for up to one month, BMS-189453 caused testicular degeneration and atrophy (Schulze 1338 et al. 2001). In later studies using lower doses, this compound was shown to reversibly inhibit 1339 spermatogenesis in rats, without other adverse testicular effects (Chung et al. 2016). 1340 1341 WIN 18,446: 1342 While the BMS synthetic retinoid above inhibits RA signaling, WIN 18,466 (a bisdichloroacetyl 1343 diamine) acts by lowering local RA concentration via inhibition of RALDH2 (Kogan et al. 2014). WIN 1344 18,466 administration to neonatal mice caused meiotic defects in spermatocytes (Kent et al. 2016). 1345 In WIN 18,466-treated rodents, the progression of progenitor cells from A spermatogonia into A1 1346 spermatogonia is blocked (Griswold and Hogarth 2018). 1347 1348 R115866: 1349 This triazole CYP26 inhibitor was capable of mimicking the effects of RA in an in vitro dog testis 1350 model, such as upregulation of e.g. Stra8, an effect that was also observed at protein level 1351 (Kasimanickam and Kasimanickam 2014). 1352 37 1353 Ro 23-2895: 1354 High doses of the synthetic retinoid Ro 23-2895 administered to rats caused testicular degeneration, 1355 decreased testicular weight, delayed sperm release and retention, and 1356 disorganization/desquamination of the tubular epithelium accompanied by reduced numbers of 1357 mature elongated spermatids (Bosakowski et al. 1991). These effects resemble those caused by 1358 vitamin A deficiency. In a parallel study plasma and testis retinol levels were lower compared to 1359 controls, suggesting that Ro 23-2895 caused testicular degeneration by interfering with normal 1360 retinoid homeostasis (Bosakowski et al. 1991). 1361 1362 Isotetinoin/13-cis-RA: 1363 Results from a pilot study suggest that treatment with 13-cis-RA (used for e.g. acne treatment) can 1364 improve sperm production (Amory et al. 2017). 1365 1366

38 1367 10. Potential adverse outcome pathways in female and male reproduction 1368 1369 Adverse Outcome Pathways (AOPs) form a framework for organizing data on the relationships 1370 between a molecular initiating event (MIE) induced by the interaction of a stressor (i.e. a chemical) 1371 with a molecular target and the resulting sequence of events (KE), leading to an adverse outcome 1372 (AO). AOPs can establish the rationale for the use of particular assays or in vivo endpoints, and/or 1373 highlight the need for the development of assays or exploring existing test guidelines to cover one or 1374 more components of an AOP. 1375 1376 Currently, there are no AOPs published focusing on the disruption of the retinoid pathway in the AOP 1377 Wiki database11, with the exception of one AOP under development, linking RALDH inhibition with 1378 visual impairment in fish (not open for citing). In the scientific literature, attempts to build AOPs and 1379 other frameworks to understand RA-dependent effects on embryogenesis have been published 1380 (Tonk et al. 2015, Baker et al. 2018, Battistoni et al. 2019, Di Renzo et al. 2019, Piersma et al. 201912), 1381 with disruption of CYP26 enzymes and RALDH2 suggested to be important KE. Previous efforts have 1382 not focused specifically on the reproduction system per se, even though embryo development of 1383 several other organ systems is discussed. Below, drafts of such potential AOPs are described. 1384 1385 Since there is substantial crosstalk between the retinoid system and other nuclear hormone systems 1386 (see section 4), there are additional AOPs that may be involved with retinoid pathway effects on 1387 reproduction., For example, activation of PPARα may impair steroidogenesis, and could lead to 1388 impaired fertility in males (AOP1813). RXR is a PPAR heterodimer partner, and it is possible that RXR 1389 may be involved in this AOP. 1390 1391

11 https://aopwiki.org/aops; accessed in September 2019. 12 https://cfpub.epa.gov/si/si_public_record_Report.cfm?dirEntryId=345956&Lab=NCCT 13 https://aopwiki.org/aops/18 39 1392 Proposed AOP for in utero CYP26B1 inhibition effects on male fertility 1393 As described in sections 5 and 6, temporally and spatially regulated RA concentrations in the fetal 1394 gonads appear to play a role both in the proper development of the testis and in the differentiation 1395 of PGCs into either oocytes or spermatogonia. In the male fetal testis, perturbed RA catabolism 1396 caused by inhibition of CYP26B1 (putative MIE) could disrupt the normal RA signaling pathway, which 1397 could lead to the AO impaired male fertility (Figure 10). 1398 1399 Following inhibition of the CYP26B1 enzyme or complete knockout of the cyp26b1 gene in the male 1400 mouse fetus, aberrant meiotic and apoptotic GC are observed in the testis (Bowles et al. 2006, 1401 MacLean et al. 2007, Teletin et al. 2017), and virtually no germ cells are present in testes from 1402 neonatal pups. In vitro, CYP26 inhibition prevents the normal mitotic arrest of the male GC and 1403 induces their death by apoptosis (Teletin et al. 2017). Male cyp26b1-/- homozygote 13.5 dpc mouse 1404 embryos display a mild ovotestis phenotype, with an “ovarian component” at the anterior end of the 1405 gonad which is where RA levels are expected to be higher due to the connection to the mesonephric 1406 tubules at this end (Bowles et al. 2018). In addition, abnormal development of the Leydig cells and of 1407 the Müllerian and Wolffian ducts was observed in the same male mouse embryos (Bowles et al. 1408 2018). Thus, it is possible that the (observed) effects on the germ cells and on testis development 1409 may lead to (possible) adverse effects on spermatogenesis and malfunctioning testes, which could 1410 eventually lead to (possible) impaired male fertility (see figure 10 below). 1411 1412 1413

1414 1415 1416 Figure 8: Proposed AOP for how CYP26B1 inhibition in the fetal male testis may lead to impaired 1417 male infertility. Studies looking at CYP26B1 inhibition were using deletion of the gene rather than 1418 chemical inhibition of the enzyme. 1419 1420

40 1421 Proposed AOP for effects of RDH10/RALDH inhibition in the adult on male fertility 1422 Based on the critical roles of RA in spermatogenesis in the adult male described in section 8, it seems 1423 possible that chemicals that inhibits the enzymes that produce RA (putative MIE), via decreased RA 1424 levels, may cause adverse effects on the spermatogenic process leading to the AO impaired male 1425 fertility (see figure 11 below). 1426 1427 In juvenile male mice deficient in RDH10 in Sertoli cells and germ cells (leading to a local lack of RA, 1428 observed via effects Stra8 expression levels and by using a RA-responsive reporter cell system), 1429 spermatogenesis is blocked (Tong et al. 2013). Also in juvenile mice, the differentiation of A aligned 1430 to A1 spermatogonia has been shown to be dependent on RA synthesis by RALDH in Sertoli cells 1431 (Raverdeau et al. 2012). At least in the mouse, RA is also required to disengage spermatozoa from 1432 the Sertoli cell cytoplasm during spermiation (Spiller and Bowles 2015, Teletin et al. 2017). 1433 1434 1435

1436 1437 1438 1439 Figure 9: Proposed AOP for how RDH10 or RALDH inhibition in the adult male testis may lead to 1440 impaired male infertility. Studies looking at RDH10 or RALDH inhibition were using deletion of the 1441 genes rather than chemical inhibition of the enzymes. 1442 1443 1444 Although no such data have been retrieved, it also seems possible that decreased RA testis levels via 1445 chemically-induced CYP26B1 inhibition (putative MIE) could lead to similar adverse effects on 1446 spermatogenesis. 1447

41 1448 Proposed AOP for meiosis initiation, oogenesis, folliculogenesis and female fertility, 1449 focusing on Stra8 1450 The role for RA for meiosis initiation, oogenesis and folliculogenesis in the female fetus is discussed 1451 in section 7. In female fetal offspring of VAD rats, a delay or failure of meiosis initiation was observed 1452 in oogonia (Li and Clagett-Dame 2009). Although RA levels were not measured, the low Stra8 1453 expression suggest lower than normal RA levels in the ovaries. The critical role of Stra8 in meiosis, 1454 oogenesis and follicular development is also evident in Stra8-/- female mice, which are infertile and 1455 have smaller ovaries with no oocytes or follicles (heterozygotes are fertile) (Baltus et al. 2006). 1456 Animal experiments with BPA and DEHP have connected lack of increased Stra8 expression with 1457 impaired female fertility; however, the effects of BPA (Zhang et al. 2012) and DEHP (Zhang et al. 1458 2015) appear to by-pass the retinoid system and instead affect Stra8 expression via epigenetic 1459 mechanisms. 1460 1461 1462

1463 1464 1465 1466 Figure 10: Proposed AOP for absence of ovarian Stra8 induction in utero possibly leading to impaired 1467 female fertility. 1468 1469

42 1470 Proposed AOP for endometriosis in the adult female, focusing on CYP26A1 1471 In endometrial tissue, RA appears to inhibit the decidualization of stromal cells, which is a 1472 prerequisite for blastocyst implantation after fertilization; thus, reduced concentration of RA in the 1473 endometrial tissue could be necessary for successful implantation (Deng et al. 2013). As discussed in 1474 section 7, the RA-catabolizing enzyme in the endometrium is CYP26A1, and its expression increases 1475 during the endometrial secretory phase (lowering the RA levels) when compared to the proliferative 1476 phase, during which RA-synthesizing enzymes such as RALDHs were increased (Deng et al. 2003). 1477 Expression of CYP26A1 is down-regulated in both the secretory and proliferative phases in 1478 endometrial biopsies from women with moderate or severe endometriosis, when compared to 1479 healthy women (Burney et al. 2007). The availability of RA may therefore be increased in the 1480 endometrial tissue of women with endometriosis. 1481 1482 1483

1484 1485 1486 Figure 11: Proposed AOP for lack of CYP26A1 increase in the adult female endometrium possibly 1487 leading to impaired female fertility. 1488

43 1489 11. Initial scoping effort: assays and endpoints for effects of chemicals on 1490 female and male reproduction via the retinoid system 1491 1492 Several existing OECD test guidelines measure endpoints relevant to the estrogen, androgen, 1493 steroidogenesis, and thyroid hormone pathways, but currently, no OECD test guidelines include 1494 endpoints specifically indicative of retinoid system modulation. Some reproductive parameters 1495 already included in existing test guidelines, would also provide information on adverse effects 1496 stemming from retinoid disruption. However, due to extensive cross-talk with other pathways, there 1497 is no endpoint identified that will give information specifically on disruption of retinoid signaling. The 1498 information assembled in this report regarding the role of retinoids in female and male reproduction 1499 indicates possible in vitro/ex vivo assays that could associate adverse reproductive outcomes to 1500 disruption of retinoid signaling. 1501 1502 Some of the proposed assays are suitable for screening larger numbers of chemicals (e.g. in silico/in 1503 vitro, OECD Conceptual Framework [CF] levels 1, 2), while in vivo test methods address more 1504 complex mechanistic and organ system effects (CF level 3, 4, 5). One other option could be to include 1505 ex vivo models. In order to correctly interpret positive results in lower-tier models, the ability of a 1506 substance to reach reproductive target cells/tissues during the relevant developmental window must 1507 be evaluated. 1508 1509 In silico methods 1510 In silico methods, such as molecular docking models and quantitative structure-activity relationship 1511 (QSAR) models for binding to e.g. CYP26, RALDH, Stra8, RAR and RXR, could, if available, be 1512 integrated into CF Level 1. A quantitative AOP has been developed for craniofacial malformations 1513 caused by azole fungicides, acting as CYP26 inhibitors (Battistoni et al. 2019). Also, homology 1514 modeling and computational docking simulations for CYP26A1 and CYP26B1 have been performed to 1515 identify inhibitors for these enzymes (Foti et al. 2016). Several studies describe QSAR models for 1516 binding to and activating e.g. Pregnane X receptor (PXR; an RXRα-heterodimerizing partner) and 1517 CYP3A4 (Rosenberg et al. 2017b), thyroperoxidase (TPO; Rosenberg et al. 2017a) and AhR (Klimenko 1518 et al. 2019). Similar models for e.g. RARs and RXRs would be valuable, and should be developed. 1519 1520 In vitro assays and ex vivo methods 1521 Temporal and spatial regulation of local RA is critical for normal reproductive development. Thus, in 1522 addition to nuclear receptor activation assays, assays measuring expression and/or activity of the 1523 enzymes involved in the metabolism of RA can be considered as candidate in vitro assays (Piersma et 1524 al. 2017; see also section 10). 1525 1526 Considering the complex RA modulation of both female and male reproduction, a panel of in vitro 1527 assays with the following endpoints is suggested: 1528 1529  CYP26A1, CYP26B1 (induction, inhibition) 1530  RALDH (induction, inhibition) 1531  STRA8 (induction) 1532  RAR/RXR transactivation 1533 1534 The proposed top candidate enzyme for fetal exposure to chemicals is the RA-catalyzing CYP26B1 1535 enzyme. Inhibition of CYP26B1 in the fetal testis (leading to increased RA levels) would cause serious 1536 adverse effects on normal development of testicular somatic and germ cells, while induction of the 1537 CYP26B1 enzyme in the female fetus could potentially lead to decreased RA tissue levels and 1538 subsequent adverse effects on initiation of germ cell meiosis. For evaluating adult females, CYP26A1 1539 is the more relevant enzyme, considering its role in the endometrium. 1540 44 1541 Inhibition of the RA-synthesizing RALDH enzymes could decrease RA levels, and assays measuring 1542 RALDH inhibition may therefore be considered for screening purposes. In the postnatal male, RALDH 1543 inhibition and the subsequent decrease in RA levels would interfere with spermatogenesis. 1544 1545 Stra8 is regulated by RA, and is expressed only in germ cells (Oulad-Abdelghani et al. 1996, Mark et 1546 al. 2008). Stra8 is required for germ cells to enter meiosis (Baltus et al. 2006). Therefore, an in vitro 1547 screening assay for Stra8 expression could be used to identify substances that, through disturbed RA 1548 signaling, affect meiosis initiation. 1549 1550 The different RAR and RXR transcription factors mediate the retinoid signal when activated by RA or 1551 9-cis-RA (RXRs can also act as a silent partner, i.e. no RXR ligand is necessary for RXR to bind and 1552 active genes). Activation or inactivation of these transcription factors by chemicals could lead to 1553 different effects on the cellular level, which may be translated to effects on tissue/organ level. 1554 1555 For any in vitro assay, the choice of cellular system is important in terms of both relevance and 1556 practicality. Ovarian stem cell test systems have been described, in which pre-meiotic oogonia in the 1557 post-natal ovary are used (Bhartiya and Patel 2018). This type of cells could theoretically be used to 1558 evaluate meiotic initiation in vitro. For screening large numbers of chemicals, gonadal cell or tissue 1559 models (preferably of human origin) would probably be necessary. Given the sex-specific effects of 1560 RA pathways, using two model systems in parallel (ovarian/follicular and testicular) is recommended. 1561 Any in vitro model should be characterized in terms of expression and function of e.g. CYP26, RoDH, 1562 and RALDH). 1563 1564 Ex vivo model systems consisting of rodent follicular cells or whole rodent neonatal ovaries (Hannon 1565 and Flaws 2015) could generate valuable information on chemical disruption of the possible 1566 association between RA and the formation and maturation of follicles (Minkina et al. 2017, 1567 Damdimopoulou et al. 2019). Only pre-meoitic germ cells, which can be obtained from fetal rodent 1568 ovaries and cultured in vitro (Liu, Paczkowski et al. 2012) can provide information on chemical effects 1569 on RA-mediated meiotic induction. Both human and rat fetal testis cultures have been used to study 1570 the effects of e.g. hormone disrupting chemicals on the development of human fetal testis (Lambrot 1571 et al. 2006, Spade et al. 2018b). 1572 1573 In vivo endpoints 1574 Expression measurements of e.g. CYP26, RoDH, RALDH, and Stra8 using immunohistochemistry or in 1575 situ hybridization could be performed on tissues from in vivo studies, for obtaining information on 1576 the mechanistic level. 1577 1578 The detailed histopathological (morphology) examination of reproductive organs/organ structures 1579 that are part of animal studies performed according to OECD reproductive development guideline 1580 studies could add valuable information when evaluating a possible retinoid-disrupting impact of the 1581 test chemical on these organs (Table 2 and 3; CF Level 3 or 4). Serum levels of AMH, as a proxy for 1582 follicle counting, could potentially be evaluated. 1583

45 1584 Table 2. Current OECD TG endpoints which might capture possible retinoid effects on mammalian female reproductive health, at CF level 4 and 5 (adapted 1585 from the draft DRP). Also indicated are limitations and data gaps. 1586 Retinoid effects OECD Test Guidelines (TG) Comments Reproductive development (female rodents) Oogenesis, Extended one-generation reproductive TG 443: Quantitative (most sensitive). Histopathological examination should be aimed at detecting a follicular count toxicity study (TG 443); quantitative evaluation of primordial and small growing follicles, as well as corpora lutea, in F1 females.

2-Generation reproduction toxicity study TG 416: only qualitative (limited sensitivity). A quantitative evaluation of primordial follicles should be (TG 416). conducted for F1 females. Oestrus cycles Extended one-generation reproductive TG 443: Vaginal smears should be examined daily for all F1 females in cohort 1A, after the onset of toxicity study (TG 443); vaginal patency, until the first cornified smear is recorded, in order to determine the time interval between these two events. Oestrous cycles for all F1 females in cohort 1A should also be monitored for a Reproductive screening test (TG 421); period of two weeks, commencing around PND 75. Combined 28- day/reproductive screening assay (TG 422); TG 421/422: Not included in the offspring and also not possible due to, termination of the offspring on PND 14, i.e. before puberty. 2-Generation reproduction toxicity study (TG 416 most recent update). TG 416: Oestrous cycle length and normality are evaluated in F1 females by vaginal smears prior to mating, and optionally during mating, until evidence of mating is found. Vagina, uterus Extended one-generation reproductive TG 443: Uterus (with oviducts and cervix), ovaries will be weighed (F1). Full histopathology is performed with cervix, and toxicity study (TG 443); for all high-dose and control F1 animals. All litters should be represented by at least 1 pup per sex. ovaries Organs and tissues demonstrating treatment-related changes and all gross lesions should also be Reproductive screening test (TG 421); examined in all animals in the lower dose groups to aid in determining a NOAEL. Combined 28- day/reproductive screening assay (TG 422); TG 421/422: Not included and not possible due to termination on PND 14 (F1).

2-Generation reproduction toxicity study TG 416: Vagina, uterus with cervix, and ovaries (preserved for histopathology (parental F1 animals) (TG 416 most recent update). determining a NOAEL.

TG 421/422: Not included and not possible due to termination on PND 14, although follicular counts can be made at this age (F1).

TG 416: Vagina, uterus with cervix, and ovaries (preserved for histopathology (parental F1 animals).

Adult exposure (female rodents) Vagina, uterus Extended one-generation reproductive TG 443: Uterus (with oviducts and cervix), ovaries will be weighed. Full histopathology is performed for all with cervix and toxicity study (TG 443); high- dose and control P animals. Organs demonstrating treatment-related changes should also be ovaries examined in all animals at the lower dose groups to aid in determining a NOAEL. Additionally, (histopathology) Reproductive screening test (TG 421); reproductive organs of all animals suspected of reduced fertility should be subjected to histopathological Combined 28- day/reproductive screening evaluation. assay (TG 422); TG 421/422: optional: paired ovaries (wet weight) and uterus (including cervix) in females (P). 2-Generation reproduction toxicity study (TG 416 most recent update). TG 416: Examination of the ovaries of the P animals is optional. Oestrus cycles Extended one-generation reproductive TG 443: Normally the assessment of oestrous cyclicity (by vaginal cytology) will start at the beginning of toxicity study (TG 443); the treatment period and continue until confirmation of mating or the end of the 2-week mating period.

Reproductive screening test (TG 421); TGs 421/422: Oestrous cycles should be monitored before treatment to select study females with Combined 28- day/reproductive screening regular cyclicity. Vaginal smears should also be monitored daily from the beginning of the treatment assay (TG 422); period until evidence of mating.

2-Generation reproduction toxicity study TG 416: Oestrous cycle length and normality are evaluated in P females by vaginal smears prior to mating, (TG 416 most recent update). and optionally during mating, until evidence of mating is found.

1587 1588

1589 Table 3. Current OECD TG endpoints which might capture possible retinoid effects on mammalian male reproductive health, at CF level 4 and 5. Also indicated 1590 are limitations and data gaps (adapted from the draft DRP). 1591 Retinoid effects OECD Test Guidelines (TG) Comments Reproductive development (male rodents) Hypospadias, Extended one-generation reproductive TG443: Most sensitive design due to 20 litters per group, 2 or more males per litter and termination dysgenesis of toxicity study (TG 443); after puberty. external reproductive Reproductive screening test (TG 421); TG 421/422: Limited sensitivity due to limited group size (n~8-10) and termination before puberty organs, Combined 28- day/reproductive screening (hypospadia should be detectable at birth). cryptorchidism assay (TG 422); TG416: Sensitive design due to 20 litters per group and termination after puberty, but only 1 male 2-Generation reproduction toxicity study per litter decreases the sensitivity compared to TG443. (TG 416 most recent update); TG 414: Sensitive design due to 20 litters per group, but limited/unknown sensitivity as the Prenatal Developmental Toxicity Study (TG offspring is terminated before birth (although e.g. hypospadia should still be detectable). 414). Testes development Extended one-generation reproductive TG443: Most sensitive design due to 20 litters per group and 2 adult male offspring per litter. (weight and toxicity study (TG 443); histopathology) TG 421/422: Not included but could be assessed at termination on PND 14. Limited sensitivity due to Reproductive screening test (TG 421); limited group size (n~8-10) and termination well before puberty. Combined 28- day/reproductive screening assay (TG 422); TG416: Sensitive design due to 20 litters per group and termination after puberty, but only 1 male per litter decreases the sensitivity compared to TG443. 2-Generation reproduction toxicity study (TG 416 most recent update).

Spermatogenesis Extended one-generation reproductive TG443: sensitive design due to 20 litters per group, 1 adult male offspring per litter examined. (sperm quality toxicity study (TG 443); and testicular TG 421/422: Not included and not possible due to termination on PND 14. histology Reproductive screening test (TG 421); TG416: Sensitive design due to 20 litters per group, 1 adult male offspring per litter examined Combined 28- day/reproductive screening (only included in the most recent update in 2001). assay (TG 422);

2-Generation reproduction toxicity study (TG 416 most recent update).

Adult exposure (male rodents) Spermatogenesis Extended one- generation reproductive TG 443: Sperm parameters assessed in around 20 parental males unless there is existing data to (sperm quality, toxicity study (TG 443); show that sperm parameters are unaffected in a 90-day study. testis weight and testicular Reproductive screening test (TG 421); TG 421/422: Limited sensitivity due to limited group size (n~8-10) and assessment of sperm quality is histopathology) Combined 28- day/reproductive screening not required; TG416: Sperm parameters assessed in around 20 parental males (sperm quality only assay (TG 422); included in the most recent update in 2001).

2-Generation reproduction toxicity study TG 408, 90-day study: Limited sensitivity due to limited group size (n~10) and assessment of sperm (TG 416 most recent update);, quality is not required.

Repeated Dose 90-day Oral Toxicity Study in TG407, 28 day study: Very limited sensitivity due to limited group size (n~5), assessment of Rodents (TG 408); epididymal sperm parameters are optional.

Repeated Dose 28-Day Oral Toxicity Study in Rodents (TG 407).

1592 1593

1594 12. Concluding remarks 1595 1596 Potential endpoints for studying effects on female and male reproduction caused by chemical 1597 exposure, possibly affecting the retinoid pathway, have been identified and is described in this 1598 report. 1599 1600 In silico methods, such as QSAR and molecular docking models for e.g. CYP26, RALDH, Stra8, RAR, 1601 RXR could be integrated into OECD CF Level 1. A small number of in silico tools for retinoid signaling 1602 pathway (all focusing on CYP26) have been identified (Battistoni et al. 2019, Foti et al. 2016). 1603 Multiple such tools exist for other signaling pathways (Rosenberg et al. 2017b, Rosenberg et al. 1604 2017a, Klimenko et al. 2019), which suggests that they could be developed also for the retinoid 1605 pathway. 1606 1607 For in vitro assays, there are commercial as well as governmental/academic assays that measure 1608 retinoid-relevant endpoints such as RAR/RXR activation (see section 14). For others (e.g. CYP26, 1609 Stra8, RALDH), assays suitable for integration into OECD CF Level 2 remain to be developed (see 1610 section 14 for more information). The CYP26 and RALDH isomers appear to be the best candidates at 1611 present, since they play a critical role in regulating RA concentrations in several reproductive tissues 1612 in both males and females. Also, CYP26 isomer expression and activity appears to be well studied in 1613 many labs, suggesting that these assays might be mature enough for further development. However, 1614 a more comprehensive analysis has to be performed in order to identify the most suitable candidate 1615 assay. 1616 1617 Regarding in vivo models (current mammalian OECD TG studies), no retinoid-relevant endpoints that 1618 could be added to already existing test methods has been identified, especially for TG443 which has 1619 the most sensitive design. One exception could be to evaluate the value of addition of ovary 1620 histopathology of pups in TGs 421/422. Certain already existing endpoints could be informative 1621 regarding retinoid disruption. For females ovaries and follicles are of particular interest, and for 1622 males sperm parameters. However, we have not identified any endpoints that would provide 1623 information specifically on retinoid disruption, since the control of male and female reproduction 1624 involves many other pathways. 1625 1626 Several RA-targeted drugs have been demonstrated to affect reproductive function while also 1627 interfering with retinoid homeostasis. However, no single chemical compound has been identified 1628 that is known to affect female or male reproduction specifically via perturbation of the retinoid 1629 pathway(s). This means that validation of the potential endpoints or assays described in this report, 1630 with regard specifically to retinoid disruption, will be challenging. 1631 1632

50 1633 13. References 1634 1635 Abali, R., Yuksel, M. A., Aktas, C., Celik, C., Guzel, S., Erfan, G., & Sahin, O. (2013). Decreased ovarian reserve in 1636 female Sprague–Dawley rats induced by isotretinoin (retinoic acid) exposure. Reproductive 1637 BioMedicine Online, 27(2), 184-191. doi:10.1016/j.rbmo.2013.04.010 1638 Acmaz, G., Cmar, L., Acmaz, B., Aksoy, H., Kafadar, Y. T., Madendag, Y., . . . Muderris, I. (2019). The Effects of Oral 1639 Isotretinoin in Women with Acne and Polycystic Ovary Syndrome. BioMed Research International, 2019, 5. 1640 doi:10.1155/2019/2513067 1641 Agrimson, K. S., Oatley, M. J., Mitchell, D., Oatley, J. M., Griswold, M. D., & Hogarth, C. A. (2017). Retinoic acid 1642 deficiency leads to an increase in spermatogonial stem number in the neonatal mouse testis, but excess 1643 retinoic acid results in no change. Developmental Biology, 432(2), 229-236. 1644 doi:https://doi.org/10.1016/j.ydbio.2017.10.002 1645 Aikawa, H., Koyama, S., Matsuda, M., Nakahashi, K., Akazome, Y., & Mori, T. (2004). Relief effect of vitamin A on the 1646 decreased motility of sperm and the increased incidence of malformed sperm in mice exposed neonatally 1647 to bisphenol A. Cell and Tissue Research, 315(1), 119-124. doi:10.1007/s00441-003-0806-1 1648 Aksoy, H., Cinar, L., Acmaz, G., Aksoy, U., Aydin, T., Vurdem, U. E., . . . Kartal, D. (2015). The Effect of Isotretinoin on 1649 Ovarian Reserve Based on Hormonal Parameters, Ovarian Volume, and Antral Follicle Count in Women 1650 with Acne. Gynecologic and Obstetric Investigation, 79(2), 78-82. doi:10.1159/000371551 1651 Al Tanoury, Z., Piskunov, A., & Rochette-Egly, C. (2013). Vitamin A and retinoid signaling: genomic and 1652 nongenomic effects. Journal of Lipid Research, 54(7), 1761-1775. doi:10.1194/jlr.R030833 1653 Allenby, G., Bocquel, M. T., Saunders, M., Kazmer, S., Speck, J., Rosenberger, M., . . . Chambon, P. (1993). Retinoic 1654 acid receptors and retinoid X receptors: interactions with endogenous retinoic acids. Proceedings of the 1655 National Academy of Sciences of the United States of America, 90(1), 30. doi:10.1073/pnas.90.1.30 1656 Alnouti, Y., & Klaassen, C. D. (2008). Tissue Distribution, Ontogeny, and Regulation of Aldehyde Dehydrogenase 1657 (Aldh) Enzymes mRNA by Prototypical Microsomal Enzyme Inducers in Mice. Toxicological Sciences, 1658 101(1), 51-64. doi:10.1093/toxsci/kfm280 1659 Amory, J. K., Ostrowski, K. A., Gannon, J. R., Berkseth, K., Stevison, F., Isoherranen, N., . . . Walsh, T. (2017). 1660 Isotretinoin administration improves sperm production in men with infertility from 1661 oligoasthenozoospermia: a pilot study. Andrology, 5(6), 1115-1123. doi:10.1111/andr.12420 1662 André, A., Ruivo, R., Gesto, M., Castro, L. F. C., & Santos, M. M. (2014). Retinoid metabolism in invertebrates: When 1663 evolution meets endocrine disruption. General and Comparative Endocrinology, 208, 134-145. 1664 doi:https://doi.org/10.1016/j.ygcen.2014.08.005 1665 Aoki, T., & Takada, T. (2012). Bisphenol A modulates germ cell differentiation and retinoic acid signaling in mouse 1666 ES cells. Reproductive Toxicology, 34(3), 463-470. doi:https://doi.org/10.1016/j.reprotox.2012.06.001 1667 Arnold, S. L., Kent, T., Hogarth, C. A., Schlatt, S., Prasad, B., Haenisch, M., . . . Isoherranen, N. (2015). Importance of 1668 ALDH1A enzymes in determining human testicular retinoic acid concentrations. Journal of Lipid Research, 1669 56(2), 342-357. doi:10.1194/jlr.M054718 1670 Azaïs-Braesco, V., & Pascal, G. (2000). Vitamin A in pregnancy: requirements and safety limits. The American 1671 Journal of Clinical Nutrition, 71(5), 1325S-1333S. doi:10.1093/ajcn/71.5.1325s 1672 Bagavandoss, P., & Midgley, A. R. J. (1987). Lack of Difference Between Retinoic Acid and Retinol in Stimulating 1673 Progesterone Production by Luteinizing Granulosa Cells in Vitro. Endocrinology, 121(1), 420-428. 1674 doi:10.1210/endo-121-1-420 1675 Baker, N., Boobis, A., Burgoon, L., Carney, E., Currie, R., Fritsche, E., . . . Daston, G. (2018). Building a developmental 1676 toxicity ontology. Birth Defects Res. doi:10.1002/bdr2.1189 1677 Balmer, J. E., & Blomhoff, R. (2002). Gene expression regulation by retinoic acid. Journal of Lipid Research, 43(11), 1678 1773-1808. doi:10.1194/jlr.R100015-JLR200 1679 Baltus, A. E., Menke, D. B., Hu, Y.-C., Goodheart, M. L., Carpenter, A. E., de Rooij, D. G., & Page, D. C. (2006). In germ 1680 cells of mouse embryonic ovaries, the decision to enter meiosis precedes premeiotic DNA replication. 1681 Nature Genetics, 38, 1430. doi:10.1038/ng1919 1682 Barnett, K. R., Schilling, C., Greenfeld, C. R., Tomic, D., & Flaws, J. A. (2006). Ovarian follicle development and 1683 transgenic mouse models. Human Reproduction Update, 12(5), 537-555. doi:10.1093/humupd/dml022 1684 Bastos Maia, S., Rolland Souza, A. S., Costa Caminha, M. d. F., Lins da Silva, S., Callou Cruz, R. d. S. B. L., Carvalho 1685 Dos Santos, C., & Batista Filho, M. (2019). Vitamin A and Pregnancy: A Narrative Review. Nutrients, 11(3), 1686 681. doi:10.3390/nu11030681 1687 Battistoni, M., Di Renzo, F., Menegola, E., & Bois, F. Y. (2019). Quantitative AOP based teratogenicity prediction for 1688 mixtures of azole fungicides. Computational Toxicology, 11, 72-81. 1689 doi:https://doi.org/10.1016/j.comtox.2019.03.004 1690 Bay, K., Main, K. M., Toppari, J., & Skakkebæk, N. E. (2011). Testicular descent: INSL3, testosterone, genes and the

51 1691 intrauterine milieu. Nature Reviews Urology, 8, 187. doi:10.1038/nrurol.2011.23 1692 Beedle, M.-T., Stevison, F., Zhong, G., Topping, T., Hogarth, C., Isoherranen, N., & Griswold, M. D. (2018). Sources of 1693 all-trans retinal oxidation independent of the aldehyde dehydrogenase 1A isozymes exist in the postnatal 1694 testis†. Biology of Reproduction, 100(2), 547-560. doi:10.1093/biolre/ioy200 1695 Belyaeva, O. V., Wu, L., Shmarakov, I., Nelson, P. S., & Kedishvili, N. Y. (2018). Retinol dehydrogenase 11 is essential 1696 for the maintenance of retinol homeostasis in liver and testis in mice. The Journal of biological chemistry, 1697 293(18), 6996-7007. doi:10.1074/jbc.RA117.001646 1698 Best, M. W., Wu, J., Pauli, S. A., Kane, M. A., Pierzchalski, K., Session, D. R., . . . Sidell, N. (2015). A role for retinoids in 1699 human oocyte fertilization: regulation of connexin 43 by retinoic acid in cumulus granulosa cells. 1700 Molecular Human Reproduction, 21(6), 527-534. doi:10.1093/molehr/gav017 1701 Bhartiya, D., & Patel, H. (2018). Ovarian stem cells—resolving controversies. Journal of Assisted Reproduction and 1702 Genetics, 35(3), 393-398. doi:10.1007/s10815-017-1080-6 1703 Bittman, E. L. (2015). Timing in the Testis. Journal of Biological Rhythms, 31(1), 12-36. 1704 doi:10.1177/0748730415618297 1705 Blaner, W. S., & Li, Y. (2015). Vitamin A Metabolism, Storage and Tissue Delivery Mechanisms. In P. Dollé & K. 1706 Neiderreither (Eds.), The Retinoids: Biology, Biochemistry and Disease (pp. 1-34). 1707 Blomhoff, R., Helgerud, P., Rasmussen, M., Berg, T., & Norum, K. R. (1982). In vivo uptake of chylomicron 1708 [3H]retinyl ester by rat liver: evidence for retinol transfer from parenchymal to nonparenchymal cells. 1709 Proceedings of the National Academy of Sciences of the United States of America, 79(23), 7326. 1710 doi:10.1073/pnas.79.23.7326 1711 Bosakowski, T., Levin, A. A., & Durham, S. K. (1991). Time course of testicular degeneration in rats induced by a 1712 synthetic retinoid (Ro 23-2895) and evidence for induction of hypovitaminosis A in the testes. Toxicology, 1713 66(1), 105-118. doi:10.1016/0300-483X(91)90182-Z 1714 Bowles, J., Feng, C.-W., Ineson, J., Miles, K., Spiller, C. M., Harley, V. R., . . . Koopman, P. (2018). Retinoic Acid 1715 Antagonizes Testis Development in Mice. Cell Reports, 24(5), 1330-1341. 1716 doi:https://doi.org/10.1016/j.celrep.2018.06.111 1717 Bowles, J., Feng, C.-W., Miles, K., Ineson, J., Spiller, C., & Koopman, P. (2016). ALDH1A1 provides a source of 1718 meiosis-inducing retinoic acid in mouse fetal ovaries. Nature Communications, 7, 10845. 1719 doi:10.1038/ncomms10845 1720 Bowles, J., Knight, D., Smith, C., Wilhelm, D., Richman, J., Mamiya, S., . . . Koopman, P. (2006). Retinoid Signaling 1721 Determines Germ Cell Fate in Mice. Science, 312(5773), 596. 1722 Brieño-Enríquez, M. A., Reig-Viader, R., Cabero, L., Toran, N., Martínez, F., Roig, I., & Garcia Caldés, M. (2012). Gene 1723 expression is altered after bisphenol A exposure in human fetal oocytes in vitro. MHR: Basic science of 1724 reproductive medicine, 18(4), 171-183. doi:10.1093/molehr/gar074 1725 Bruner-Tran, K. L., & Osteen, K. G. (2011). Developmental Exposure to TCDD Reduces Fertility and Negatively 1726 Affects Pregnancy Outcomes across Multiple Generations. Reproductive toxicology (Elmsford, N.Y.), 31(3), 1727 344-350. doi:10.1016/j.reprotox.2010.10.003 1728 Bryant, S. L., Francis, J. C., Lokody, I. B., Wang, H., Risbridger, G. P., Loveland, K. L., & Swain, A. (2014). Sex specific 1729 retinoic acid signaling is required for the initiation of urogenital sinus bud development. Developmental 1730 Biology, 395(2), 209-217. doi:10.1016/j.ydbio.2014.09.016 1731 Bryson, H. M., & Wagstaff, A. J. (1996). Liarozole. Drugs & Aging, 9(6), 478-484. doi:10.2165/00002512-199609060- 1732 00010 1733 Burney, R. O., Talbi, S., Hamilton, A. E., Vo, K. C., Nyegaard, M., Nezhat, C. R., . . . Giudice, L. C. (2007). Gene 1734 Expression Analysis of Endometrium Reveals Progesterone Resistance and Candidate Susceptibility Genes 1735 in Women with Endometriosis. Endocrinology, 148(8), 3814-3826. doi:10.1210/en.2006-1692 1736 Busada, J. T., & Geyer, C. B. (2016). The Role of Retinoic Acid (RA) in Spermatogonial Differentiation1. Biology of 1737 Reproduction, 94(1), 10, 11-10-10, 11-10. doi:10.1095/biolreprod.115.135145 1738 Chambon, P. (1996). A decade of molecular biology of retinoic acid receptors. FASEB journal : official publication of 1739 the Federation of American Societies for Experimental Biology, 10(9), 940. doi:10.1096/fasebj.10.9.8801176 1740 Chen, P.-J., Padgett, W. T., Moore, T., Winnik, W., Lambert, G. R., Thai, S.-F., . . . Nesnow, S. (2009). Three conazoles 1741 increase hepatic microsomal retinoic acid metabolism and decrease mouse hepatic retinoic acid levels in 1742 vivo. Toxicology and Applied Pharmacology, 234(2), 143-155. 1743 doi:https://doi.org/10.1016/j.taap.2008.10.004 1744 Chen, Y., Ma, L., Hogarth, C., Wei, G., Griswold, M. D., & Tong, M.-H. (2016). Retinoid signaling controls 1745 spermatogonial differentiation by regulating expression of replication-dependent core histone genes. 1746 Development, 143(9), 1502. doi:10.1242/dev.135939 1747 Chen, Y., & Reese, D. H. (2013). A screen for disruptors of the retinol (vitamin A) signaling pathway. Birth Defects 1748 Res B Dev Reprod Toxicol, 98(3), 276-282. doi:10.1002/bdrb.21062 1749 Childs, A. J., Cowan, G., Kinnell, H. L., Anderson, R. A., & Saunders, P. T. K. (2011). Retinoic Acid Signalling and the

52 1750 Control of Meiotic Entry in the Human Fetal Gonad. PLOS ONE, 6(6), e20249. 1751 doi:10.1371/journal.pone.0020249 1752 Chlapek, P., Slavikova, V., Mazanek, P., Sterba, J., & Veselska, R. (2018). Why Differentiation Therapy Sometimes 1753 Fails: Molecular Mechanisms of Resistance to Retinoids. International Journal of Molecular Sciences, 19(1), 1754 132. doi:10.3390/ijms19010132 1755 Chung, S. S. W., Wang, X., & Wolgemuth, D. J. (2016). Prolonged Oral Administration of a Pan-Retinoic Acid 1756 Receptor Antagonist Inhibits Spermatogenesis in Mice With a Rapid Recovery and Changes in the 1757 Expression of Influx and Efflux Transporters. Endocrinology, 157(4), 1601-1612. doi:10.1210/en.2015-1675 1758 Çinar, L., Kartal, D., Ergin, C., Aksoy, H., Karadag, M. A., Aydin, T., . . . Borlu, M. (2016). The effect of systemic 1759 isotretinoin on male fertility. Cutaneous and Ocular Toxicology, 35(4), 296-299. 1760 doi:10.3109/15569527.2015.1119839 1761 Cione, E., Maria Cristina, C., Roberto, C., Mariarita, P., Attilio, P., & Giuseppe, G. (2016). Vitamin A and Diabesity: 1762 New Insight for Drug Discovery. Mini-Reviews in Medicinal Chemistry, 16(9), 738-742. 1763 doi:http://dx.doi.org/10.2174/1389557515666150709112822 1764 Clagett-Dame, M., & Knutson, D. (2011). Vitamin A in reproduction and development. Nutrients, 3(4), 385-428. 1765 doi:10.3390/nu3040385 1766 Clermont, Y. (1963). The cycle of the seminiferous epithelium in man. The American journal of anatomy, 112, 35. 1767 Coward, W. A., Howell, J. M., Thompson, J. N., & Pitt, G. A. J. (1969). The retinol requirements of rats for 1768 spermatogenesis and vision. British Journal of Nutrition, 23(3), 619-626. doi:10.1079/BJN19690069 1769 de Araújo, J. F. P., Podratz, P. L., Merlo, E., Sarmento, I. V., da Costa, C. S., Niño, O. M. S., . . . Graceli, J. B. (2018). 1770 Organotin Exposure and Vertebrate Reproduction: A Review. Frontiers in Endocrinology, 9(64). 1771 doi:10.3389/fendo.2018.00064 1772 de Lera, Á. R., Krezel, W., & Rühl, R. (2016). An Endogenous Mammalian Retinoid X Receptor Ligand, At Last! 1773 ChemMedChem, 11(10), 1027-1037. doi:10.1002/cmdc.201600105 1774 Dean, A., van den Driesche, S., Wang, Y., McKinnell, C., Macpherson, S., Eddie, S. L., . . . Sharpe, R. M. (2016). 1775 Analgesic exposure in pregnant rats affects fetal germ cell development with inter-generational 1776 reproductive consequences. Scientific Reports, 6, 19789. doi:10.1038/srep19789 1777 Demczuk, M., Huang, H., White, C., & Kipp, J. L. (2016). Retinoic Acid Regulates Calcium Signaling to Promote 1778 Mouse Ovarian Granulosa Cell Proliferation1. Biology of Reproduction, 95(3), 70, 71-10-70, 71-10. 1779 doi:10.1095/biolreprod.115.136986 1780 Deng, L., Shipley, G. L., Loose-Mitchell, D. S., Stancel, G. M., Broaddus, R., Pickar, J. H., & Davies, P. J. A. (2003). 1781 Coordinate Regulation of the Production and Signaling of Retinoic Acid by Estrogen in the Human 1782 Endometrium. The Journal of Clinical Endocrinology & Metabolism, 88(5), 2157-2163. doi:10.1210/jc.2002- 1783 021844 1784 Denis, L., Debruyne, F., De Porre, P., & Bruynseels, J. (1998). Early clinical experience with liarozole 1785 (Liazal^™) in patients with progressive prostate cancer. European Journal of Cancer, 1786 34(4), 469-475. doi:10.1016/S0959-8049(97)10120-4 1787 Di Renzo, F., Metruccio, F., Battistoni, M., Moretto, A., & Menegola, E. (2019). Relative potency ranking of azoles 1788 altering craniofacial morphogenesis in rats: An in vitro data modelling approach. Food and Chemical 1789 Toxicology, 123, 553-560. doi:https://doi.org/10.1016/j.fct.2018.12.004 1790 El Shahawy, M., Reibring, C.-G., Hallberg, K., Neben, C. L., Marangoni, P., Harfe, B. D., . . . Gritli-Linde, A. (2019). 1791 Sonic Hedgehog Signaling Is Required for Cyp26 Expression during Embryonic Development. 1792 International Journal of Molecular Sciences, 20(9), 2275. 1793 Elabbas, L. E., Esteban, J., Barber, X., Hamscher, G., Nau, H., Bowers, W. J., . . . Håkansson, H. (2014). In Utero and 1794 Lactational Exposure to a Mixture of Environmental Contaminants Detected in Canadian Arctic Human 1795 Populations Alters Retinoid Levels in Rat Offspring with Low Margins of Exposure. Journal of Toxicology 1796 and Environmental Health, Part A, 77(5), 223-245. doi:10.1080/15287394.2013.861776 1797 Emmett, S. D., & West, K. P., Jr. (2014). Gestational vitamin A deficiency: a novel cause of sensorineural hearing loss 1798 in the developing world? Medical hypotheses, 82(1), 6-10. doi:10.1016/j.mehy.2013.09.028 1799 Escobar-Morreale, H. (2018). Polycystic ovary syndrome: definition, aetiology, diagnosis and treatment. Nature 1800 Reviews Endocrinology, 14(5), 270-284. doi:10.1038/nrendo.2018.24 1801 Feng, Y.-M., Liang, G.-J., Pan, B., Qin, X.-S., Zhang, X.-F., Chen, C.-L., . . . Shen, W. (2014). Notch pathway regulates 1802 female germ cell meiosis progression and early oogenesis events in fetal mouse. Cell cycle (Georgetown, 1803 Tex.), 13(5), 782-791. doi:10.4161/cc.27708 1804 Finkelstein, D., Lamba, V., Assem, M., Rengelshausen, J., Yasuda, K., Strom, S., & Schuetz, E. (2006). ADME 1805 transcriptome in Hispanic versus White donor livers: Evidence of a globally enhanced NR1I3 (CAR, 1806 constitutive androstane receptor) gene signature in Hispanics. Xenobiotica, 36(10-11), 989-1012. 1807 doi:10.1080/00498250600861769 1808 Fiorella, P. D., Olson, J. R., & Napoli, J. L. (1995). 2,3,7,8-Tetrachlorodibenzo-p-dioxin Induces Diverse Retinoic Acid

53 1809 Metabolites in Multiple Tissues of the Sprague-Dawley Rat. Toxicology and Applied Pharmacology, 134(2), 1810 222-228. doi:https://doi.org/10.1006/taap.1995.1187 1811 Foti, R. S., Diaz, P., & Douguet, D. (2016). Comparison of the ligand binding site of CYP2C8 with CYP26A1 and 1812 CYP26B1: a structural basis for the identification of new inhibitors of the retinoic acid hydroxylases. 1813 Journal of Enzyme Inhibition and Medicinal Chemistry, 31(sup2), 148-161. 1814 doi:10.1080/14756366.2016.1193734 1815 Fritzsche, B., Vermot, J., Neumann, U., Schmidt, A., Schweigert, F. J., Dollé, P., & Rühl, R. (2007). Regulation of 1816 expression of the retinoic acid metabolizing enzyme CYP26A1 in uteri of ovariectomized mice after 1817 treatment with ovarian steroid hormones. Molecular Reproduction and Development, 74(2), 258-264. 1818 doi:10.1002/mrd.20526 1819 Fujihara, M., Yamamizu, K., Comizzoli, P., Wildt, D. E., & Songsasen, N. (2018). Retinoic acid promotes in vitro 1820 follicle activation in the cat ovary by regulating expression of matrix metalloproteinase 9. PLOS ONE, 1821 13(8), e0202759. doi:10.1371/journal.pone.0202759 1822 Gaemers, I. C., van Pelt, A. M. M., van der Saag, P. T., Hoogerbrugge, J. W., Themmen, A. P. N., & de Rooij, D. G. 1823 (1997). Effect of Retinoid Status on the Messenger Ribonucleic Acid Expression of Nuclear Retinoid 1824 Receptors α, β, and γ, and Retinoid X Receptors α, β, and γ in the Mouse Testis*. Endocrinology, 138(4), 1825 1544-1551. doi:10.1210/endo.138.4.5051 1826 Gely-Pernot, A., Raverdeau, M., Teletin, M., Vernet, N., Féret, B., Klopfenstein, M., . . . Ghyselinck, N. B. (2015). 1827 Retinoic Acid Receptors Control Spermatogonia Cell-Fate and Induce Expression of the SALL4A 1828 Transcription Factor. PLOS Genetics, 11(10), e1005501. doi:10.1371/journal.pgen.1005501 1829 Georges, A., Auguste, A., Bessière, L., Vanet, A., Todeschini, A.-L., & Veitia, R. A. (2014). FOXL2: a central 1830 transcription factor of the ovary. 52(1), R17. doi:10.1530/jme-13-0159 1831 Germain, P., Chambon, P., Eichele, G., Evans, R. M., Lazar, M. A., Leid, M., . . . Gronemeyer, H. (2006). International 1832 Union of Pharmacology. LX. Retinoic Acid Receptors. Pharmacological Reviews, 58(4), 712. 1833 doi:10.1124/pr.58.4.4 1834 Ghyselinck, N. B., Vernet, N., Dennefeld, C., Giese, N., Nau, H., Chambon, P., . . . Mark, M. (2006). Retinoids and 1835 spermatogenesis: Lessons from mutant mice lacking the plasma retinol binding protein. Developmental 1836 Dynamics, 235(6), 1608-1622. doi:10.1002/dvdy.20795 1837 Griswold, M. D. (2016). Spermatogenesis: The Commitment to Meiosis. Physiological reviews, 96(1), 1-17. 1838 doi:10.1152/physrev.00013.2015 1839 Griswold, M. D., & Hogarth, C. (2018). Beyond stem cells: Commitment of progenitor cells to meiosis. Stem Cell 1840 Research, 27, 169-171. doi:https://doi.org/10.1016/j.scr.2018.01.032 1841 Grive, K. J., & Freiman, R. N. (2015). The developmental origins of the mammalian ovarian reserve. Development, 1842 142(15), 2554. 1843 Han, B. C., Xia, H. F., Sun, J., Yang, Y., & Peng, J. P. (2010). Retinoic acid‐metabolizing enzyme cytochrome P450 1844 26a1 (cyp26a1) is essential for implantation: Functional study of its role in early pregnancy. Journal of 1845 Cellular Physiology, 223(2), 471-479. doi:10.1002/jcp.22056 1846 Hannon, P. R., & Flaws, J. A. (2015). The Effects of Phthalates on the Ovary. Frontiers in Endocrinology, 6, 8. 1847 doi:10.3389/fendo.2015.00008 1848 Heller, C. G., Moore, D. J., & Paulsen, C. A. (1961). Suppression of spermatogenesis and chronic toxicity in men by a 1849 new series of bis(dichloroacetyl) diamines. Toxicology and Applied Pharmacology, 3(1), 1-11. 1850 doi:https://doi.org/10.1016/0041-008X(61)90002-3 1851 Hoegberg, P., Schmidt, C. K., Nau, H., Catharine Ross, A., Zolfaghari, R., Fletcher, N., . . . Håkansson, H. (2003). 1852 2,3,7,8-Tetrachlorodibenzo-p-dioxin induces lecithin: retinol acyltransferase transcription in the rat 1853 kidney. Chemico-Biological Interactions, 145(1), 1-16. doi:https://doi.org/10.1016/S0009-2797(02)00157-6 1854 Hogarth, C. A., Amory, J. K., & Griswold, M. D. (2011). Inhibiting vitamin A metabolism as an approach to male 1855 contraception. Trends Endocrinol Metab, 22(4), 136-144. doi:10.1016/j.tem.2011.01.001 1856 Hogarth, C. A., Evanoff, R., Mitchell, D., Kent, T., Small, C., Amory, J. K., & Griswold, M. D. (2013). Turning a 1857 spermatogenic wave into a tsunami: synchronizing murine spermatogenesis using WIN 18,446. Biology of 1858 Reproduction, 88(2), 40-40. doi:10.1095/biolreprod.112.105346 1859 Hogarth, C. A., Evans, E., Onken, J., Kent, T., Mitchell, D., Petkovich, M., & Griswold, M. D. (2015). CYP26 Enzymes 1860 Are Necessary Within the Postnatal Seminiferous Epithelium for Normal Murine Spermatogenesis1. 1861 Biology of Reproduction, 93(1). doi:10.1095/biolreprod.115.129718 1862 Hogarth, C. A., & Griswold, M. D. (2010). The key role of vitamin A in spermatogenesis. J Clin Invest, 120(4), 956- 1863 962. doi:10.1172/JCI41303 1864 Holm, J. B., Mazaud-Guittot, S., Danneskiold-Samsøe, N. B., Chalmey, C., Jensen, B., Nørregård, M. M., . . . 1865 Kristensen, D. M. (2016). Intrauterine Exposure to Paracetamol and Aniline Impairs Female Reproductive 1866 Development by Reducing Follicle Reserves and Fertility. Toxicological Sciences, 150(1), 178-189. 1867 doi:10.1093/toxsci/kfv332

54 1868 Huang, F.-J., Hsuuw, Y.-D., Lan, K.-C., Kang, H.-Y., Chang, S.-Y., Hsu, Y.-C., & Huang, K.-E. (2005). Adverse effects of 1869 retinoic acid on embryo development and the selective expression of retinoic acid receptors in mouse 1870 blastocysts. Human Reproduction, 21(1), 202-209. doi:10.1093/humrep/dei286 1871 Håkansson, H., Manzoor, E., & Ahlborg, U. G. (1991). Interaction between dietary vitamin A and single oral doses of 1872 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on the TCDD-induced toxicity and on the vitamin A status in 1873 the rat. Journal of nutritional science and vitaminology, 37(3), 239. doi:10.3177/jnsv.37.239 1874 Jacobs, M. N., Marczylo, E. L., Guerrero-Bosagna, C., & Rüegg, J. (2017). Marked for Life: Epigenetic Effects of 1875 Endocrine Disrupting Chemicals. Annual Review of Environment and Resources, 42(1), 105-160. 1876 doi:10.1146/annurev-environ-102016-061111 1877 Jana, K., Douglas, B. M., Qing, Z., Blanche, C., Michael, D. G., & David, C. P. (2006). Retinoic acid regulates sex- 1878 specific timing of meiotic initiation in mice. Proceedings of the National Academy of Sciences of the United 1879 States of America, 103(8), 2474. doi:10.1073/pnas.0510813103 1880 Jauregui, E. J., Mitchell, D., Topping, T., Hogarth, C. A., & Griswold, M. D. (2018). Retinoic acid receptor signaling is 1881 necessary in steroidogenic cells for normal spermatogenesis and epididymal function. Development. 1882 Jayaram, M., Murthy, S. K., & Ganguly, J. (1973). Effect of vitamin A deprivation on the cholesterol side-chain 1883 cleavage enzyme activity of testes and ovaries of rats (Short Communication). Biochemical Journal, 1884 136(1), 221-223. 1885 Jiang, Y., Chen, L., Taylor, R. N., Li, C., & Zhou, X. (2018a). Physiological and pathological implications of retinoid 1886 action in the endometrium. J Endocrinol, 236(3), R169-R188. doi:10.1530/JOE-17-0544 1887 Jiang, Y., Li, C., Chen, L., Wang, F., & Zhou, X. (2017). Potential role of retinoids in ovarian physiology and 1888 pathogenesis of polycystic ovary syndrome. Clin Chim Acta, 469, 87-93. doi:10.1016/j.cca.2017.03.025 1889 Jiang, Y., Zhao, Y., Chen, S., Chen, L., Li, C., & Zhou, X. (2018b). Regulation by FSH of the dynamic expression of 1890 retinol-binding protein 4 in the mouse ovary. Reproductive Biology and Endocrinology, 16(1), 25. 1891 doi:10.1186/s12958-018-0348-8 1892 Jing, J., Nelson, C., Paik, J., Shirasaka, Y., Amory, J. K., & Isoherranen, N. (2017). Physiologically Based 1893 Pharmacokinetic Model of All-trans-Retinoic Acid with Application to Cancer Populations and Drug 1894 Interactions. The Journal of pharmacology and experimental therapeutics, 361(2), 246-258. 1895 doi:10.1124/jpet.117.240523 1896 Johansson, H. K. L., Svingen, T., Fowler, P. A., Vinggaard, A. M., & Boberg, J. (2017). Environmental influences on 1897 ovarian dysgenesis - developmental windows sensitive to chemical exposures. Nat Rev Endocrinol, 13(7), 1898 400-414. doi:10.1038/nrendo.2017.36 1899 Jørgensen, A., Nielsen, J. E., Perlman, S., Lundvall, L., Mitchell, R. T., Juul, A., & Rajpert-De Meyts, E. (2015). Ex vivo 1900 culture of human fetal gonads: manipulation of meiosis signalling by retinoic acid treatment disrupts 1901 testis development. Human Reproduction, 30(10), 2351-2363. doi:10.1093/humrep/dev194 1902 Kamata, R., Shiraishi, F., Nishikawa, J.-i., Yonemoto, J., & Shiraishi, H. (2008). Screening and detection of the in vitro 1903 agonistic activity of xenobiotics on the retinoic acid receptor. Toxicology in Vitro, 22(4), 1050-1061. 1904 doi:https://doi.org/10.1016/j.tiv.2008.01.002 1905 Kane, M., Folias, A., Wang, C., & Napoli, J. (2008). Quantitative Profiling of Endogenous Retinoic Acid in Vivo and in 1906 Vitro by Tandem Mass Spectrometry. Analytical Chemistry, 80(5), 1702-1708. doi:10.1021/ac702030f 1907 Kashimada, K., Svingen, T., Feng, C.-W., Pelosi, E., Bagheri-Fam, S., Harley, V. R., . . . Koopman, P. (2011). 1908 Antagonistic regulation of Cyp26b1 by transcription factors SOX9/SF1 and FOXL2 during gonadal 1909 development in mice. FASEB journal : official publication of the Federation of American Societies for 1910 Experimental Biology, 25(10), 3561-3569. doi:10.1096/fj.11-184333 1911 Kasimanickam, V., & Kasimanickam, R. (2014). Exogenous Retinoic Acid and Cytochrome P450 26B1 Inhibitor 1912 Modulate Meiosis‐Associated Genes Expression in Canine Testis, an In Vitro Model. Reproduction in 1913 Domestic Animals, 49(2), 315-323. doi:10.1111/rda.12276 1914 Kawai, T., Richards, J. S., & Shimada, M. (2018). The Cell Type–Specific Expression of Lhcgr in Mouse Ovarian Cells: 1915 Evidence for a DNA-Demethylation–Dependent Mechanism. Endocrinology, 159(5), 2062-2074. 1916 doi:10.1210/en.2018-00117 1917 Kawai, T., Yanaka, N., Richards, J. S., & Shimada, M. (2016). De Novo-Synthesized Retinoic Acid in Ovarian Antral 1918 Follicles Enhances FSH-Mediated Ovarian Follicular Cell Differentiation and Female Fertility. 1919 Endocrinology, 157(5), 2160-2172. doi:10.1210/en.2015-2064 1920 Kedishvili, N. Y. (2013). Enzymology of retinoic acid biosynthesis and degradation. Journal of Lipid Research, 54(7), 1921 1744-1760. doi:10.1194/jlr.R037028 1922 Kent, T., Arnold, S. L., Fasnacht, R., Rowsey, R., Mitchell, D., Hogarth, C. A., . . . Griswold, M. D. (2016). ALDH Enzyme 1923 Expression Is Independent of the Spermatogenic Cycle, and Their Inhibition Causes Misregulation of 1924 Murine Spermatogenic Processes1. Biology of Reproduction, 94(1). doi:10.1095/biolreprod.115.131458 1925 Kim, H., Lapointe, J., Kaygusuz, G., Ong, D. E., Li, C., van de Rijn, M., . . . Pollack, J. R. (2005). The retinoic acid 1926 synthesis gene ALDH1a2 is a candidate tumor suppressor in prostate cancer. Cancer research, 65(18),

55 1927 8118. doi:10.1158/0008-5472.CAN-04-4562 1928 Klimenko, K., Rosenberg, S. A., Dybdahl, M., Wedebye, E. B., & Nikolov, N. G. (2019). QSAR modelling of a large 1929 imbalanced aryl hydrocarbon activation dataset by rational and random sampling and screening of 1930 80,086 REACH pre-registered and/or registered substances.(Research Article)(quantitative structure- 1931 activity relationship)(Report). PLOS ONE, 14(3), e0213848. doi:10.1371/journal.pone.0213848 1932 Klonisch, T., Muller-Huesmann, H., Riedel, M., Kehlen, A., Bialek, J., Radestock, Y., . . . Hombach-Klonisch, S. (2005). 1933 INSL3 in the benign hyperplastic and neoplastic human prostate gland. Int J Oncol, 27(2), 307-315. 1934 Kogan, P., & Wald, M. (2014). Male Contraception: History and Development. Urologic Clinics of North America, 1935 41(1), 145-161. doi:https://doi.org/10.1016/j.ucl.2013.08.012 1936 Korkmaz, E., Cetinkaya, N., Oz, M., Ozgu, E., Bas, S., Isikalan, M., . . . Gungor, T. (2017). The Possibly Reversible 1937 Isotretinoin Effect of Decreased Ovarian Reserve in Sprague-Dawley Albinos: Part I, Biochemical Analyses. 1938 Gynecologic and Obstetric Investigation, 82(1), 72-77. 1939 Koubova, J., Hu, Y.-C., Bhattacharyya, T., Soh, Y. Q. S., Gill, M. E., Goodheart, M. L., . . . Page, D. C. (2014). Retinoic 1940 acid activates two pathways required for meiosis in mice. PLOS Genetics, 10(8), e1004541-e1004541. 1941 doi:10.1371/journal.pgen.1004541 1942 Kransler, K. M., Tonucci, D. A., McGarrigle, B. P., Napoli, J. L., & Olson, J. R. (2007). Gestational exposure to 2,3,7,8- 1943 tetrachlorodibenzo-p-dioxin alters retinoid homeostasis in maternal and perinatal tissues of the 1944 Holtzman rat. Toxicology and Applied Pharmacology, 224(1), 29-38. 1945 doi:https://doi.org/10.1016/j.taap.2007.06.006 1946 Kumar, S., Chatzi, C., Brade, T., Cunningham, T. J., Zhao, X., & Duester, G. (2011). Sex-specific timing of meiotic 1947 initiation is regulated by Cyp26b1 independent of retinoic acid signalling. Nature Communications, 2, 1948 151-151. doi:10.1038/ncomms1136 1949 Kurlandsky, S. B., Gamble, M. V., Ramakrishnan, R., & Blaner, W. S. (1995). Plasma Delivery of Retinoic Acid to 1950 Tissues in the Rat. Journal of Biological Chemistry, 270(30), 17850-17857. doi:10.1074/jbc.270.30.17850 1951 Lambrot, R., Coffigny, H., Pairault, C., Donnadieu, A.-C., Frydman, R., Habert, R., & Rouiller-Fabre, V. (2006). Use of 1952 Organ Culture to Study the Human Fetal Testis Development: Effect of Retinoic Acid. The Journal of 1953 Clinical Endocrinology & Metabolism, 91(7), 2696-2703. doi:10.1210/jc.2005-2113 1954 Laudet, V., Zieger, E., & Schubert, M. (2015). Evolution of the Retinoic Acid Signaling Pathway. In P. Dollé & K. 1955 Neiderreither (Eds.), The Retinoids: Biology, Biochemistry and Disease (pp. 75-90). 1956 Laursen, K. B., Wong, P.-M., & Gudas, L. J. (2012). Epigenetic regulation by RARα maintains ligand-independent 1957 transcriptional activity. Nucleic acids research, 40(1), 102-115. doi:10.1093/nar/gkr637 1958 Lawson, C., Gieske, M., Murdoch, B., Ye, P., Li, Y., Hassold, T., & Hunt, P. A. (2011). Gene Expression in the Fetal 1959 Mouse Ovary Is Altered by Exposure to Low Doses of Bisphenol A1. Biology of Reproduction, 84(1), 79-86. 1960 doi:10.1095/biolreprod.110.084814 1961 Le Bouffant, R., Guerquin, M. J., Duquenne, C., Frydman, N., Coffigny, H., Rouiller-Fabre, V., . . . Livera, G. (2010). 1962 Meiosis initiation in the human ovary requires intrinsic retinoic acid synthesis. Human Reproduction, 1963 25(10), 2579-2590. doi:10.1093/humrep/deq195 1964 Lee, L. M. Y., Leung, C.-Y., Tang, W. W. C., Choi, H.-L., Leung, Y.-C., McCaffery, P. J., . . . Shum, A. S. W. (2012). A 1965 paradoxical teratogenic mechanism for retinoic acid. Proceedings of the National Academy of Sciences of 1966 the United States of America, 109(34), 13668-13673. doi:10.1073/pnas.1200872109 1967 Lemaire, G., Balaguer, P., Michel, S., & Rahmani, R. (2005). Activation of retinoic acid receptor-dependent 1968 transcription by organochlorine pesticides. Toxicology and Applied Pharmacology, 202(1), 38-49. 1969 doi:10.1016/j.taap.2004.06.004 1970 Li, H., & Clagett-Dame, M. (2009). Vitamin A Deficiency Blocks the Initiation of Meiosis of Germ Cells in the 1971 Developing Rat Ovary In Vivo1. Biology of Reproduction, 81(5), 996-1001. 1972 doi:10.1095/biolreprod.109.078808 1973 Liu, C., Paczkowski, M., Othman, M., & Yao, H. H.-C. (2012). Investigating the Origins of Somatic Cell Populations in 1974 the Perinatal Mouse Ovaries Using Genetic Lineage Tracing and Immunohistochemistry. In W.-Y. Chan & 1975 L. A. Blomberg (Eds.), Germline Development: Methods and Protocols (pp. 211-221). New York, NY: 1976 Springer New York. 1977 Liu, L., Suzuki, K., Nakagata, N., Mihara, K., Matsumaru, D., Ogino, Y., . . . Yamada, G. (2012). Retinoic acid signaling 1978 regulates sonic hedgehog and bone morphogenetic protein signalings during genital tubercle 1979 development. Birth defects research. Part B, Developmental and reproductive toxicology, 95(1), 79-88. 1980 doi:10.1002/bdrb.20344 1981 Liu, Z., Sun, Y., Jiang, Y., Qian, Y., Chen, S., Gao, S., . . . Zhou, X. (2018). Follicle-stimulating hormone (FSH) promotes 1982 retinol uptake and metabolism in the mouse ovary. Reproductive biology and endocrinology : RB&E, 16(1), 1983 52-52. doi:10.1186/s12958-018-0371-9 1984 Lohnes, D., Kastner, P., Dierich, A., Mark, M., Lemeur, M., & Chambon, P. (1993). Function of retinoic acid receptor γ 1985 in the mouse. Cell, 73(4), 643-658. doi:10.1016/0092-8674(93)90246-M

56 1986 Long, M. D., Singh, P. K., Russell, J. R., Llimos, G., Rosario, S., Rizvi, A., . . . Campbell, M. J. (2019). The miR-96 and 1987 RARγ signaling axis governs androgen signaling and prostate cancer progression. Oncogene, 38(3), 421- 1988 444. doi:10.1038/s41388-018-0450-6 1989 Lord, T., Oatley, M. J., & Oatley, J. M. (2018). Testicular Architecture Is Critical for Mediation of Retinoic Acid 1990 Responsiveness by Undifferentiated Spermatogonial Subtypes in the Mouse. Stem cell reports, 10(2), 538- 1991 552. doi:10.1016/j.stemcr.2018.01.003 1992 Lovekamp-Swan, T., Jetten, A. M., & Davis, B. J. (2003). Dual activation of PPARα and PPARγ by mono-(2- 1993 ethylhexyl) phthalate in rat ovarian granulosa cells. Molecular and Cellular Endocrinology, 201(1), 133-141. 1994 doi:https://doi.org/10.1016/S0303-7207(02)00423-9 1995 Lu, C., Xu, M., Wang, Y., Qin, Y., Du, G., Wu, W., . . . Wang, X. (2013). Genetic variants in meiotic program initiation 1996 pathway genes are associated with spermatogenic impairment in a Han Chinese population. PLOS ONE, 1997 8(1), e53443-e53443. doi:10.1371/journal.pone.0053443 1998 Lucas, T. F., Nascimento, A. R., Pisolato, R., Pimenta, M. T., Lazari, M. F. M., & Porto, C. S. (2014). Receptors and 1999 signaling pathways involved in proliferation and differentiation of Sertoli cells. Spermatogenesis, 4, 2000 e28138-e28138. doi:10.4161/spmg.28138 2001 Ma, J.-j., Han, B.-c., Yang, Y., & Peng, J.-p. (2012). Retinoic acid synthesis and metabolism are concurrent in the 2002 mouse uterus during peri-implantation. Cell and Tissue Research, 350(3), 525-537. doi:10.1007/s00441- 2003 012-1507-4 2004 Macejova, D., Toporova, L., & Brtko, J. (2016). The role of retinoic acid receptors and their cognate ligands in 2005 reproduction in a context of triorganotin based endocrine disrupting chemicals. Endocr Regul, 50(3), 154- 2006 164. doi:10.1515/enr-2016-0018 2007 Maclean, G., Li, H., Metzger, D., Chambon, P., & Petkovich, M. (2007). Apoptotic Extinction of Germ Cells in Testes 2008 of Cyp26b1 Knockout Mice. Endocrinology, 148(10), 4560-4567. doi:10.1210/en.2007-0492 2009 Mangelsdorf, D. J., & Evans, R. M. (1995). The RXR heterodimers and orphan receptors (Vol. 83, pp. 841-850). 2010 Marceau, G., Gallot, D., Borel, V., Lémery, D., Dastugue, B., Dechelotte, P., & Sapin, V. (2006). Molecular and 2011 metabolic retinoid pathways in human amniotic membranes. Biochemical and Biophysical Research 2012 Communications, 346(4), 1207-1216. doi:https://doi.org/10.1016/j.bbrc.2006.06.024 2013 Mark, M., Jacobs, H., Oulad-Abdelghani, M., Dennefeld, C., Féret, B., Vernet, N., . . . Ghyselinck, N. B. (2008). STRA8- 2014 deficient spermatocytes initiate, but fail to complete, meiosis and undergo premature chromosome 2015 condensation. Journal of Cell Science, 121(19), 3233. 2016 Mark, M., Teletin, M., Vernet, N., & Ghyselinck, N. B. (2015). Role of retinoic acid receptor (RAR) signaling in post- 2017 natal male germ cell differentiation. Biochimica et Biophysica Acta (BBA) - Gene Regulatory Mechanisms, 2018 1849(2), 84-93. doi:https://doi.org/10.1016/j.bbagrm.2014.05.019 2019 Marti, N., Galván, J. A., Pandey, A. V., Trippel, M., Tapia, C., Müller, M., . . . Flück, C. E. (2017). Genes and proteins of 2020 the alternative steroid backdoor pathway for dihydrotestosterone synthesis are expressed in the human 2021 ovary and seem enhanced in the polycystic ovary syndrome. Molecular and Cellular Endocrinology, 441, 2022 116-123. doi:https://doi.org/10.1016/j.mce.2016.07.029 2023 Martin, M. T., Dix, D. J., Judson, R. S., Kavlock, R. J., Reif, D. M., Richard, A. M., . . . Houck, K. A. (2010). Impact of 2024 Environmental Chemicals on Key Transcription Regulators and Correlation to Toxicity End Points within 2025 EPA’s ToxCast Program. Chemical Research in Toxicology, 23(3), 578-590. doi:10.1021/tx900325g 2026 Minegishi, T., Hirakawa, T., Kishi, H., Abe, K., Tano, M., Abe, Y., & Miyamoto, K. (2000). The mechanisms of retinoic 2027 acid-induced regulation on the follicle-stimulating hormone receptor in rat granulosa cells. Biochimica et 2028 Biophysica Acta (BBA) - Molecular Cell Research, 1495(3), 203-211. doi:https://doi.org/10.1016/S0167- 2029 4889(00)00003-3 2030 Minkina, A., Lindeman, R. E., Gearhart, M. D., Chassot, A. A., Chaboissier, M. C., Ghyselinck, N. B., . . . Zarkower, D. 2031 (2017). Retinoic acid signaling is dispensable for somatic development and function in the mammalian 2032 ovary. Dev Biol, 424(2), 208-220. doi:10.1016/j.ydbio.2017.02.015 2033 Minkina, A., Matson, Clinton K., Lindeman, Robin E., Ghyselinck, Norbert B., Bardwell, Vivian J., & Zarkower, D. 2034 (2014). DMRT1 Protects Male Gonadal Cells from Retinoid-Dependent Sexual Transdifferentiation. 2035 Developmental Cell, 29(5), 511-520. doi:10.1016/j.devcel.2014.04.017 2036 Morales, C. R., & Griswold, M. D. (1987). Retinol Induces Stage Synchronization in Seminiferous Tubules of Vitamin 2037 A Deficient Ratsa. Annals of the New York Academy of Sciences, 513(1), 292-293. doi:10.1111/j.1749- 2038 6632.1987.tb25018.x 2039 Mumford, S. L., Browne, R. W., Schliep, K. C., Schmelzer, J., Plowden, T. C., Michels, K. A., . . . Schisterman, E. F. 2040 (2016). Serum Antioxidants Are Associated with Serum Reproductive Hormones and Ovulation among 2041 Healthy Women. The Journal of Nutrition, 146(1), 98-106. doi:10.3945/jn.115.217620 2042 Murphy, K. A., Quadro, L., & White, L. A. (2007). The Intersection Between the Aryl Hydrocarbon Receptor (AhR)‐ 2043 and Retinoic Acid‐Signaling Pathways Vitamins & Hormones (Vol. 75, pp. 33-67): Academic Press. 2044 Nakahashi, K., Matsuda, M., & Mori, T. (2001). Vitamin A Insufficiency Accelerates the Decrease in the Number of

57 2045 Sperm Induced by an Environmental Disruptor, Bisphenol A, in Neonatal Mice. Zoological Science, 18(6), 2046 819-821. doi:10.2108/zsj.18.819 2047 Nakajima, T., Iguchi, T., & Sato, T. (2016). Retinoic acid signaling determines the fate of uterine stroma in the 2048 mouse Müllerian duct. Proceedings of the National Academy of Sciences, 113(50), 14354-14359. 2049 doi:10.1073/pnas.1608808113 2050 Nakanishi, T., Nishikawa, J.-i., Hiromori, Y., Yokoyama, H., Koyanagi, M., Takasuga, S., . . . Tanaka, K. (2005). 2051 Trialkyltin Compounds Bind Retinoid X Receptor to Alter Human Placental Endocrine Functions. 2052 Molecular Endocrinology, 19(10), 2502-2516. doi:10.1210/me.2004-0397 2053 Napoli, J. L. (2012). Physiological insights into all-trans-retinoic acid biosynthesis. BBA - Molecular and Cell Biology 2054 of Lipids, 1821(1), 152-167. doi:10.1016/j.bbalip.2011.05.004 2055 Napoli, J. L. (2017). Cellular retinoid binding-proteins, CRBP, CRABP, FABP5: Effects on retinoid metabolism, 2056 function and related diseases. Pharmacology & therapeutics, 173, 19-33. 2057 doi:10.1016/j.pharmthera.2017.01.004 2058 Nelson, C. H., Buttrick, B. R., & Isoherranen, N. (2013). Therapeutic potential of the inhibition of the retinoic acid 2059 hydroxylases CYP26A1 and CYP26B1 by xenobiotics. Current topics in medicinal chemistry, 13(12), 1402- 2060 1428. 2061 Nicholls, P. K., Harrison, C. A., Rainczuk, K. E., Wayne Vogl, A., & Stanton, P. G. (2013). Retinoic acid promotes 2062 Sertoli cell differentiation and antagonises activin-induced proliferation. Molecular and Cellular 2063 Endocrinology, 377(1-2), 33-43. doi:10.1016/j.mce.2013.06.034 2064 Niederreither, K., Fraulob, V., Garnier, J.-M., Chambon, P., & Dollé, P. (2002). Differential expression of retinoic acid- 2065 synthesizing (RALDH) enzymes during fetal development and organ differentiation in the mouse. 2066 Mechanisms of Development, 110(1), 165-171. doi:https://doi.org/10.1016/S0925-4773(01)00561-5 2067 Niederreither, K., McCaffery, P., Dräger, U. C., Chambon, P., & Dollé, P. (1997). Restricted expression and retinoic 2068 acid-induced downregulation of the retinaldehyde dehydrogenase type 2 (RALDH-2) gene during mouse 2069 development. Mechanisms of Development, 62(1), 67-78. doi:https://doi.org/10.1016/S0925- 2070 4773(96)00653-3 2071 Niederreither, K., Vermot, J., Fraulob, V., Chambon, P., & Dolle, P. (2002). Retinaldehyde dehydrogenase 2 2072 (RALDH2)- independent patterns of retinoic acid synthesis in the mouse embryo. Proceedings of the 2073 National Academy of Sciences of the United States of America, 99(25), 16111-16116. 2074 doi:10.1073/pnas.252626599 2075 Nilsson, C. B., & Håkansson, H. (2002). The Retinoid Signaling System — A Target in Dioxin Toxicity. Critical 2076 Reviews in Toxicology, 32(3), 211-232. doi:10.1080/20024091064228 2077 Nishikawa, J.-i., Mamiya, S., Kanayama, T., Nishikawa, T., Shiraishi, F., & Horiguchi, T. (2004). Involvement of the 2078 Retinoid X Receptor in the Development of Imposex Caused by Organotins in Gastropods. Environmental 2079 Science & Technology, 38(23), 6271-6276. doi:10.1021/es049593u 2080 Noback, C. R., & Takahashi, Y. I. (1978). Micromorphology of the placenta of rats reared on marginal vitamin-A- 2081 deficient diet. Cells Tissues Organs, 102(2), 195-202. doi:10.1159/000145637 2082 Novak, J., Benisek, M., & Hilscherova, K. (2008). Disruption of retinoid transport, metabolism and signaling by 2083 environmental pollutants. Environ Int, 34(6), 898-913. doi:10.1016/j.envint.2007.12.024 2084 Nya-Ngatchou, J. J., Arnold, S. L. M., Walsh, T. J., Muller, C. H., Page, S. T., Isoherranen, N., & Amory, J. K. (2013). 2085 Intratesticular 13-cis retinoic acid is lower in men with abnormal semen analyses: a pilot study. Andrology, 2086 1(2), 325-331. doi:10.1111/j.2047-2927.2012.00033.x 2087 O'Byrne, S. M., & Blaner, W. S. (2013). Retinol and retinyl esters: biochemistry and physiology. Journal of Lipid 2088 Research, 54(7), 1731-1743. doi:10.1194/jlr.R037648 2089 O'Byrne, S. M., Wongsiriroj, N., Libien, J., Vogel, S., Goldberg, I. J., Baehr, W., . . . Blaner, W. S. (2005). Retinoid 2090 absorption and storage is impaired in mice lacking lecithin:retinol acyltransferase (LRAT). The Journal of 2091 biological chemistry, 280(42), 35647-35657. doi:10.1074/jbc.M507924200 2092 Oulad-Abdelghani, M., Bouillet, P., Décimo, D., Gansmuller, A., Heyberger, S., Dollé, P., . . . Chambon, P. (1996). 2093 Characterization of a premeiotic germ cell-specific cytoplasmic protein encoded by Stra8, a novel retinoic 2094 acid-responsive gene. The Journal of Cell Biology, 135(2), 469. 2095 Paganetto, G., Campi, F., Varani, K., Piffanelli, A., Giovannini, G., & Borea, P. A. (2000). Endocrine-Disrupting Agents 2096 on Healthy Human Tissues. Pharmacology & Toxicology, 86(1), 24-29. doi:10.1034/j.1600- 2097 0773.2000.pto860105.x 2098 Paik, J., Haenisch, M., Muller, C. H., Goldstein, A. S., Arnold, S., Isoherranen, N., . . . Amory, J. K. (2014). Inhibition of 2099 Retinoic Acid Biosynthesis by the Bisdichloroacetyldiamine WIN 18,446 Markedly Suppresses 2100 Spermatogenesis and Alters Retinoid Metabolism in Mice. Journal of Biological Chemistry, 289(21), 2101 15104-15117. doi:10.1074/jbc.M113.540211 2102 Palczewski, K. (2012). Chemistry and biology of vision. The Journal of biological chemistry, 287(3), 1612. 2103 doi:10.1074/jbc.R111.301150

58 2104 Papacleovoulou, G., Edmondson, R. J., Critchley, H. O. D., Hillier, S. G., & Mason, J. I. (2009). 3β-Hydroxysteroid 2105 dehydrogenases and pre-receptor steroid metabolism in the human ovarian surface epithelium. 2106 Molecular and Cellular Endocrinology, 301(1), 65-73. doi:https://doi.org/10.1016/j.mce.2008.08.010 2107 Pauli, S. A., Session, D. R., Shang, W., Easley, K., Wieser, F., Taylor, R. N., . . . Sidell, N. (2013). Analysis of follicular 2108 fluid retinoids in women undergoing in vitro fertilization: retinoic acid influences embryo quality and is 2109 reduced in women with endometriosis. Reproductive sciences (Thousand Oaks, Calif.), 20(9), 1116-1124. 2110 doi:10.1177/1933719113477487 2111 Peng, J., Shen, L., Chen, J., Cao, X., Zhou, Y., Weng, H., . . . Wei, G. (2016). New discovery of cryptorchidism: 2112 Decreased retinoic acid in testicle. Saudi pharmaceutical journal : SPJ : the official publication of the Saudi 2113 Pharmaceutical Society, 24(3), 279-285. doi:10.1016/j.jsps.2016.04.020 2114 Pennimpede, T., Cameron Don, A., MacLean Glenn, A., Li, H., Abu-Abed, S., & Petkovich, M. (2010). The role of 2115 CYP26 enzymes in defining appropriate retinoic acid exposure during embryogenesis. Birth Defects 2116 Research Part A: Clinical and Molecular Teratology, 88(10), 883-894. doi:10.1002/bdra.20709 2117 Piersma, A. H., Hessel, E. V., & Staal, Y. C. (2017). Retinoic acid in developmental toxicology: Teratogen, morphogen 2118 and biomarker. Reprod Toxicol, 72, 53-61. doi:10.1016/j.reprotox.2017.05.014 2119 Pius, J., Shivanandappa, T., & Krishnakumari, M. K. (1990). Protective role of vitamin A in the male reproductive 2120 toxicity of Hexachlorocyclohexane (HCH) in the rat. Reproductive Toxicology, 4(4), 325-330. 2121 doi:10.1016/0890-6238(90)90045-W 2122 Pu, Y., Pearl, S., Gingrich, J., Jing, J., Martin, D., Murga-Zamalloa, C., & Veiga-Lopez, A. (2019). Multispecies study: 2123 low-dose tributyltin impairs ovarian theca cell cholesterol homeostasis through the RXR pathway in five 2124 mammalian species including humans. Archives of Toxicology, 93(6), 1665-1677. doi:10.1007/s00204-019- 2125 02449-y 2126 Quadro, L., Blaner, W. S., Salchow, D. J., Vogel, S., Piantedosi, R., Gouras, P., . . . Gottesman, M. E. (1999). Impaired 2127 retinal function and vitamin A availability in mice lacking retinol‐binding protein. EMBO Journal, 18(17), 2128 4633-4644. doi:10.1093/emboj/18.17.4633 2129 Rato, L., Alves, M. G., Socorro, S., Duarte, A. I., Cavaco, J. E., & Oliveira, P. F. (2012). Metabolic regulation is 2130 important for spermatogenesis. Nature Reviews Urology, 9, 330. doi:10.1038/nrurol.2012.77 2131 Raverdeau, M., Gely-Pernot, A., Féret, B., Dennefeld, C., Benoit, G., Davidson, I., . . . Ghyselinck, N. B. (2012). Retinoic 2132 acid induces Sertoli cell paracrine signals for spermatogonia differentiation but cell autonomously drives 2133 spermatocyte meiosis. Proceedings of the National Academy of Sciences, 109(41), 16582. 2134 doi:10.1073/pnas.1214936109 2135 Rhinn, M., & Dollé, P. (2012). Retinoic acid signalling during development. Development (Cambridge, England), 2136 139(5), 843. doi:10.1242/dev.065938 2137 Rochette-Egly, C. (2015). Retinoic acid signaling and mouse embryonic stem cell differentiation: Cross talk 2138 between genomic and non-genomic effects of RA. Biochimica et Biophysica Acta (BBA) - Molecular and 2139 Cell Biology of Lipids, 1851(1), 66-75. doi:https://doi.org/10.1016/j.bbalip.2014.04.003 2140 Rodahl, K., & Moore, T. (1943). The vitamin A content and toxicity of bear and seal liver. The Biochemical journal, 2141 37(2), 166-168. doi:10.1042/bj0370166 2142 Rosenberg, S. A., Watt, E. D., Judson, R. S., Simmons, S. O., Paul Friedman, K., Dybdahl, M., . . . Wedebye, E. B. 2143 (2017a). QSAR models for thyroperoxidase inhibition and screening of U.S. and EU chemical inventories. 2144 Computational Toxicology, 4, 11-21. doi:https://doi.org/10.1016/j.comtox.2017.07.006 2145 Rosenberg, S. A., Xia, M., Huang, R., Nikolov, N. G., Wedebye, E. B., & Dybdahl, M. (2017b). QSAR development and 2146 profiling of 72,524 REACH substances for PXR activation and CYP3A4 induction. Computational 2147 Toxicology, 1, 39-48. doi:https://doi.org/10.1016/j.comtox.2017.01.001 2148 Ross, A. C., & Zolfaghari, R. (2011). Cytochrome P450s in the regulation of cellular retinoic acid metabolism. 2149 Annual review of nutrition, 31, 65-87. doi:10.1146/annurev-nutr-072610-145127 2150 Rossitto, M., Philibert, P., Poulat, F., & Boizet-Bonhoure, B. (2015). Molecular events and signalling pathways of 2151 male germ cell differentiation in mouse. Seminars in Cell & Developmental Biology, 45, 84-93. 2152 doi:https://doi.org/10.1016/j.semcdb.2015.09.014 2153 Rout, U. K., & Armant, D. R. (2002). Expression of genes for alcohol and aldehyde metabolizing enzymes in mouse 2154 oocytes and preimplantation embryos. Reproductive Toxicology, 16(3), 253-258. 2155 doi:https://doi.org/10.1016/S0890-6238(02)00022-9 2156 Sauvant, P., Sapin, V., Alexandre-Gouabau, M.-C., Dodeman, I., Delpal, S., Quadro, L., . . . S-Braesco, V. (2001). 2157 Retinol mobilization from cultured rat hepatic stellate cells does not require retinol binding protein 2158 synthesis and secretion. International Journal of Biochemistry and Cell Biology, 33(10), 1000-1012. 2159 doi:10.1016/S1357-2725(01)00066-8 2160 Sayem, A. S. M., Giribabu, N., Karim, K., Si, L. K., Muniandy, S., & Salleh, N. (2018). Differential expression of the 2161 receptors for thyroid hormone, thyroid stimulating hormone, vitamin D and retinoic acid and extracellular 2162 signal-regulated kinase in uterus of rats under influence of sex-steroids. Biomedicine & Pharmacotherapy,

59 2163 100, 132-141. doi:https://doi.org/10.1016/j.biopha.2018.02.008 2164 Schmidt, C. K., Hoegberg, P., Fletcher, N., Nilsson, C. B., Trossvik, C., Håkansson, H., & Nau, H. (2003). 2,3,7,8- 2165 Tetrachlorodibenzo-p-dioxin (TCDD) alters the endogenous metabolism of all-trans-retinoic acid in the 2166 rat. Archives of Toxicology, 77(7), 371-383. doi:10.1007/s00204-003-0457-8 2167 Schulze, G. E., Clay, R. J., Mezza, L. E., Bregman, C. L., Buroker, R. A., & Frantz, J. D. (2001). BMS-189453, a Novel 2168 Retinoid Receptor Antagonist, Is a Potent Testicular Toxin. Toxicological Sciences, 59(2), 297-308. 2169 doi:10.1093/toxsci/59.2.297 2170 See, A. W.-M., Kaiser, M. E., White, J. C., & Clagett-Dame, M. (2008). A nutritional model of late embryonic vitamin 2171 A deficiency produces defects in organogenesis at a high penetrance and reveals new roles for the 2172 vitamin in skeletal development. Developmental Biology, 316(2), 171-190. 2173 doi:https://doi.org/10.1016/j.ydbio.2007.10.018 2174 Shannon, S. R., Moise, A. R., & Trainor, P. A. (2017). New insights and changing paradigms in the regulation of 2175 vitamin A metabolism in development. Wiley interdisciplinary reviews. Developmental biology, 6(3), 2176 10.1002/wdev.1264. doi:10.1002/wdev.264 2177 Sharpe, R. M. (2010). Environmental/lifestyle effects on spermatogenesis. Philosophical transactions of the Royal 2178 Society of London. Series B, Biological sciences, 365(1546), 1697-1712. doi:10.1098/rstb.2009.0206 2179 Shirakami, Y., Lee, S.-A., Clugston, R. D., & Blaner, W. S. (2012). Hepatic metabolism of retinoids and disease 2180 associations. BBA - Molecular and Cell Biology of Lipids, 1821(1), 124-136. 2181 doi:10.1016/j.bbalip.2011.06.023 2182 Shmarakov, I. O. (2015). Retinoid-xenobiotic interactions: the Ying and the Yang. Hepatobiliary Surg Nutr, 4(4), 2183 243-267. doi:10.3978/j.issn.2304-3881.2015.05.05 2184 Shmarakov, I. O., Lee, Y. J., Jiang, H., & Blaner, W. S. (2019). Constitutive androstane receptor mediates PCB- 2185 induced disruption of retinoid homeostasis. Toxicology and Applied Pharmacology, 381, 114731. 2186 doi:https://doi.org/10.1016/j.taap.2019.114731 2187 Siddikuzzaman, Guruvayoorappan, C., & Berlin Grace, V. M. (2011). All Trans Retinoic Acid and Cancer. 2188 Immunopharmacology and Immunotoxicology, 33(2), 241-249. doi:10.3109/08923973.2010.521507 2189 Skakkebæk, N. E., Rajpert-De Meyts, E., & Main, K. M. (2001). Testicular dysgenesis syndrome: an increasingly 2190 common developmental disorder with environmental aspects: Opinion. Human Reproduction, 16(5), 972- 2191 978. doi:10.1093/humrep/16.5.972 2192 Spade, D. J., Dere, E., Hall, S. J., Schorl, C., Freiman, R. N., & Boekelheide, K. (2018a). All-Trans Retinoic Acid Disrupts 2193 Development in Ex Vivo Cultured Fetal Rat Testes. I: Altered Seminiferous Cord Maturation and Testicular 2194 Cell Fate. Toxicological Sciences, 167(2), 546-558. doi:10.1093/toxsci/kfy260 2195 Spade, D. J., Hall, S. J., Wortzel, J. D., Reyes, G., & Boekelheide, K. (2018b). All-trans Retinoic Acid Disrupts 2196 Development in Ex Vivo Cultured Fetal Rat Testes. II: Modulation of Mono-(2-ethylhexyl) Phthalate 2197 Toxicity. Toxicological Sciences, 168(1), 149-159. doi:10.1093/toxsci/kfy283 2198 Spiller, C., Koopman, P., & Bowles, J. (2017). Sex Determination in the Mammalian Germline. Annu Rev Genet, 51, 2199 265-285. doi:10.1146/annurev-genet-120215-035449 2200 Spiller, C. M., & Bowles, J. (2015). Sex determination in mammalian germ cells. Asian J Androl, 17(3), 427-432. 2201 doi:10.4103/1008-682X.150037 2202 Stevison, F., Hogarth, C., Tripathy, S., Kent, T., & Isoherranen, N. (2017). Inhibition of the all trans-retinoic acid 2203 hydroxylases CYP26A1 and CYP26B1 results in dynamic, tissue-specific changes in endogenous atRA 2204 signaling. Drug Metabolism and Disposition. 2205 Stoppie, P., Borgers, M., Borghgraef, P., Dillen, L., Goossens, J., Sanz, G., . . . Van Wauwe, J. (2000). R115866 inhibits 2206 all-trans-retinoic acid metabolism and exerts retinoidal effects in rodents. The Journal of pharmacology 2207 and experimental therapeutics, 293(1), 304. 2208 Su, R.-W., & Fazleabas, A. T. (2015). Implantation and Establishment of Pregnancy in Human and Nonhuman 2209 Primates. Advances in anatomy, embryology, and cell biology, 216, 189-213. doi:10.1007/978-3-319- 2210 15856-3_10 2211 Suzuki, H., Kanai-Azuma, M., & Kanai, Y. (2015). From Sex Determination to Initial Folliculogenesis in Mammalian 2212 Ovaries: Morphogenetic Waves along the Anteroposterior and Dorsoventral Axes. Sex Dev, 9(4), 190-204. 2213 doi:10.1159/000440689 2214 Szanto, A., Narkar, V., Shen, Q., Uray, I. P., Davies, P. J. A., & Nagy, L. (2004). Retinoid X receptors: X-ploring their 2215 (patho)physiological functions. Cell death and differentiation, 11 Suppl 2, S126. 2216 Szwarc, M., Lan, H., William, E. G., Lisa, D. W., Qianxing, M., Ramakrishna, K., . . . John, P. L. (2018). Retinoid signaling 2217 controlled by SRC-2 in decidualization revealed by transcriptomics. REPRODUCTION, 156(5), 387-395. 2218 doi:10.1530/REP-18-0282 2219 Tarrade, A., Schoonjans, K., Pavan, L., Auwerx, J., Rochette-Egly, C. c., Evain-Brion, D. l., & Fournier, T. (2001). 2220 PPARγ/RXRα Heterodimers Control Human Trophoblast Invasion. The Journal of Clinical Endocrinology & 2221 Metabolism, 86(10), 5017-5024. doi:10.1210/jcem.86.10.7924

60 2222 Taxvig, C., Hass, U., Axelstad, M., Dalgaard, M., Boberg, J., Andeasen, H. R., & Vinggaard, A. M. (2007). Endocrine- 2223 Disrupting Activities In Vivo of the Fungicides Tebuconazole and Epoxiconazole. Toxicological Sciences, 2224 100(2), 464-473. doi:10.1093/toxsci/kfm227 2225 Tay, S., Dickmann, L., Dixit, V., & Isoherranen, N. (2010). A comparison of the roles of peroxisome proliferator- 2226 activated receptor and retinoic acid receptor on CYP26 regulation. Molecular pharmacology, 77(2), 218- 2227 227. doi:10.1124/mol.109.059071 2228 Taylor, R. N., Kane, M. A., & Sidell, N. (2015). Pathogenesis of Endometriosis: Roles of Retinoids and Inflammatory 2229 Pathways. Semin Reprod Med, 33(4), 246-256. doi:10.1055/s-0035-1554920 2230 Teletin, M., Vernet, N., Ghyselinck, N. B., & Mark, M. (2017). Roles of Retinoic Acid in Germ Cell Differentiation. Curr 2231 Top Dev Biol, 125, 191-225. doi:10.1016/bs.ctdb.2016.11.013 2232 Teletin, M., Vernet, N., Yu, J., Klopfenstein, M., Jones, J. W., Féret, B., . . . Mark, M. (2019). Two functionally 2233 redundant sources of retinoic acid secure spermatogonia differentiation in the seminiferous epithelium. 2234 Development, 146(1), dev170225. doi:10.1242/dev.170225 2235 Thatcher, J. E., Zelter, A., & Isoherranen, N. (2010). The relative importance of CYP26A1 in hepatic clearance of all- 2236 trans retinoic acid. Biochemical pharmacology, 80(6), 903-912. doi:10.1016/j.bcp.2010.05.023 2237 Theodosiou, M., Laudet, V., & Schubert, M. (2010). From carrot to clinic: an overview of the retinoic acid signaling 2238 pathway. Cellular and Molecular Life Sciences, 67(9), 1423-1445. doi:10.1007/s00018-010-0268-z 2239 Tiboni, G. M., Marotta, F., & Carletti, E. (2009). Fluconazole alters CYP26 gene expression in mouse embryos. 2240 Reproductive Toxicology, 27(2), 199-202. doi:https://doi.org/10.1016/j.reprotox.2009.01.001 2241 Tong, M.-H., Yang, Q.-E., Davis, J. C., & Griswold, M. D. (2013). Retinol dehydrogenase 10 is indispensible for 2242 spermatogenesis in juvenile males. Proceedings of the National Academy of Sciences of the United States 2243 of America, 110(2), 543-548. doi:10.1073/pnas.1214883110 2244 Tonk, E. C. M., Pennings, J. L. A., & Piersma, A. H. (2015). An adverse outcome pathway framework for neural tube 2245 and axial defects mediated by modulation of retinoic acid homeostasis. Reproductive Toxicology, 55, 104- 2246 113. doi:https://doi.org/10.1016/j.reprotox.2014.10.008 2247 Topletz, A. R., Tripathy, S., Foti, R. S., Shimshoni, J. A., Nelson, W. L., & Isoherranen, N. (2015). Induction of CYP26A1 2248 by metabolites of retinoic acid: evidence that CYP26A1 is an important enzyme in the elimination of 2249 active retinoids. Molecular pharmacology, 87(3), 430-441. doi:10.1124/mol.114.096784 2250 Urvalek, A., Laursen, K. B., & Gudas, L. J. (2014). The roles of retinoic acid and retinoic acid receptors in inducing 2251 epigenetic changes. Sub-cellular biochemistry, 70, 129-149. doi:10.1007/978-94-017-9050-5_7 2252 van Der Ven, L. T. M., van de Kuil, T., Leonards, P. E. G., Slob, W., Lilienthal, H., Litens, S., . . . Piersma, A. H. (2009). 2253 Endocrine effects of hexabromocyclododecane (HBCD) in a one-generation reproduction study in Wistar 2254 rats. Toxicology Letters, 185(1), 51-62. doi:10.1016/j.toxlet.2008.12.003 2255 van Der Ven, L. T. M., van de Kuil, T., Verhoef, A., Leonards, P. E. G., Slob, W., Cantón, R. F., . . . Vos, J. G. (2008). A 2256 28-day oral dose toxicity study enhanced to detect endocrine effects of a purified technical 2257 pentabromodiphenyl ether (pentaBDE) mixture in Wistar rats. Toxicology, 245(1), 109-122. 2258 doi:10.1016/j.tox.2007.12.016 2259 Van Pelt, A. M. M., & De Rooij, D. G. (1990). The Origin of the Synchronization of the Seminiferous Epithelium in 2260 Vitamin A-Deficient Rats after Vitamin A Replacement. Biology of Reproduction, 42(4), 677-682. 2261 doi:10.1095/biolreprod42.4.677 2262 Velte, E. K., Niedenberger, B. A., Serra, N. D., Singh, A., Roa-Delacruz, L., Hermann, B. P., & Geyer, C. B. (2019). 2263 Differential RA responsiveness directs formation of functionally-distinct spermatogonial populations at 2264 the initiation of spermatogenesis in the mouse. Development (Cambridge, England), 146(12). 2265 doi:10.1242/dev.173088 2266 Verze, P., Cai, T., & Lorenzetti, S. (2016). The role of the prostate in male fertility, health and disease. Nature 2267 Reviews Urology, 13, 379. doi:10.1038/nrurol.2016.89 2268 Vezina Chad, M., Allgeier Sarah, H., Fritz Wayne, A., Moore Robert, W., Strerath, M., Bushman, W., & Peterson 2269 Richard, E. (2008). Retinoic acid induces prostatic bud formation. Developmental Dynamics, 237(5), 1321- 2270 1333. doi:10.1002/dvdy.21526 2271 White, J. C., Highland, M., & Clagett-Dame, M. (2000). Abnormal development of the sinuatrial venous valve and 2272 posterior hindbrain may contribute to late fetal resorption of vitamin A-deficient rat embryos. Teratology, 2273 62(6), 374-384. doi:10.1002/1096-9926(200012)62:6<374::AID-TERA4>3.0.CO;2-5 2274 White, J. C., Shankar, V. N., Highland, M., Epstein, M. L., DeLuca, H. F., & Clagett-Dame, M. (1998). Defects in 2275 embryonic hindbrain development and fetal resorption resulting from vitamin A deficiency in the rat are 2276 prevented by feeding pharmacological levels of all-<em>trans-</em>retinoic acid. 2277 Proceedings of the National Academy of Sciences, 95(23), 13459. 2278 Wickenheisser, J. K., Nelson-DeGrave, V. L., Hendricks, K. L., Legro, R. S., Strauss, J. F., & McAllister, J. M. (2005). 2279 Retinoids and retinol differentially regulate steroid biosynthesis in ovarian theca cells isolated from 2280 normal cycling women and women with polycystic ovary syndrome. J Clin Endocrinol Metab, 90.

61 2281 doi:10.1210/jc.2005-0330 2282 Vickery, B., Burns, J., Zaneveld, L., Goodpasture, J., & Bergström, K. (1985). Orally administered ketoconazole 2283 rapidly appears in seminal plasma and suppresses sperm motility. Advances in Contraception, 1(4), 341- 2284 353. doi:10.1007/BF01849310 2285 Vilhais-Neto, G. C., & Pourquié, O. (2008). Retinoic acid. Current Biology, 18(7), 550-552. 2286 doi:10.1016/j.cub.2008.03.032 2287 Villota-Salazar, N. A., Mendoza-Mendoza, A., & González-Prieto, J. M. (2016). Epigenetics: from the past to the 2288 present. Frontiers in Life Science, 9(4), 347-370. doi:10.1080/21553769.2016.1249033 2289 Wiseman, E. M., Bar-El Dadon, S., & Reifen, R. (2017). The vicious cycle of vitamin a deficiency: A review. Crit Rev 2290 Food Sci Nutr, 57(17), 3703-3714. doi:10.1080/10408398.2016.1160362 2291 Wolf, G. (1996). A history of vitamin A and retinoids. The FASEB Journal, 10(9), 1102-1107. 2292 doi:10.1096/fasebj.10.9.8801174 2293 Wood, J. R., Nelson, V. L., Ho, C., Jansen, E., Wang, C. Y., Urbanek, M., . . . Strauss, J. F. (2003). The Molecular 2294 Phenotype of Polycystic Ovary Syndrome (PCOS) Theca Cells and New Candidate PCOS Genes Defined by 2295 Microarray Analysis. Journal of Biological Chemistry, 278(29), 26380-26390. doi:10.1074/jbc.M300688200 2296 Xu, H., Beasley, M. D., Warren, W. D., van der Horst, G. T. J., & McKay, M. J. (2005). Absence of Mouse REC8 2297 Cohesin Promotes Synapsis of Sister Chromatids in Meiosis. Developmental Cell, 8(6), 949-961. 2298 doi:10.1016/j.devcel.2005.03.018 2299 Yang, Y., Luo, J., Yu, D., Zhang, T., Lin, Q., Li, Q., . . . Huang, Y. (2018). Vitamin A Promotes Leydig Cell Differentiation 2300 via Alcohol Dehydrogenase 1. Frontiers in Endocrinology, 9, 644-644. doi:10.3389/fendo.2018.00644 2301 Yashiro, K., Zhao, X., Uehara, M., Yamashita, K., Nishijima, M., Nishino, J., . . . Hamada, H. (2004). Regulation of 2302 Retinoic Acid Distribution Is Required for Proximodistal Patterning and Outgrowth of the Developing 2303 Mouse Limb. Developmental Cell, 6(3), 411-422. doi:https://doi.org/10.1016/S1534-5807(04)00062-0 2304 Yoshida, S., Sukeno, M., Nakagawa, T., Ohbo, K., Nagamatsu, G., Suda, T., & Nabeshima, Y.-i. (2006). The first round 2305 of mouse spermatogenesis is a distinctive program that lacks the self-renewing spermatogonia stage. 2306 Development, 133(8), 1495. doi:10.1242/dev.02316 2307 Yuan, W., Wu, T., Fu, H., Dai, C., Wu, H., Liu, N., . . . Zhu, B. (2012). Dense Chromatin Activates Polycomb Repressive 2308 Complex 2 to Regulate H3 Lysine 27 Methylation. Science, 337(6097), 971. doi:10.1126/science.1225237 2309 Zhang, H.-Q., Zhang, X.-F., Zhang, L.-J., Chao, H.-H., Pan, B., Feng, Y.-M., . . . Shen, W. (2012). Fetal exposure to 2310 bisphenol A affects the primordial follicle formation by inhibiting the meiotic progression of oocytes. 2311 Molecular Biology Reports, 39(5), 5651-5657. doi:10.1007/s11033-011-1372-3 2312 Zhang, X.-F., Zhang, T., Han, Z., Liu, J.-C., Liu, Y.-P., Ma, J.-Y., . . . Shen, W. (2015). Transgenerational inheritance of 2313 ovarian development deficiency induced by maternal diethylhexyl phthalate exposure. Reproduction, 2314 Fertility and Development, 27(8), 1213-undefined. doi:10.1071/RD14113 2315 Zolfaghari, R., Mattie, F. J., Wei, C.-H., Chisholm, D. R., Whiting, A., & Ross, A. C. (2019). CYP26A1 gene promoter is 2316 a useful tool for reporting RAR-mediated retinoid activity. Analytical Biochemistry, 577, 98-109. 2317 doi:https://doi.org/10.1016/j.ab.2019.04.022 2318 Zuno-Floriano, F. G., Holstege, D., Hengel, M. J., Gaikwad, N. W., Aldana-Madrid, M. L., & Miller, M. G. (2012). 2319 Determination of Vitamin A and its Metabolites in Rat Testis: Possible Involvement of Vitamin A in 2320 Testicular Toxicity Caused by Molinate. Bulletin of Environmental Contamination and Toxicology, 88(6), 2321 1038-1042. doi:10.1007/s00128-012-0612-0 2322

62 2323 14. Appendix 2324 2325 Assay methods and availability: 2326 2327 RA levels: 2328 Can be measured in tissues (organs, blood) or cells with e.g. liquid chromatography methods coupled 2329 to ultraviolet or mass spectrometric detection (a multitude of publications reviewed in the D-DRP). 2330 Can also be measured with ELISA kits (available for several species) 2331 https://www.mybiosource.com 2332 CYP 26: 2333 The different isomers can be studied on mRNA and protein level using commercially available probes 2334 and antibodies (see e.g. Topletz et al. 2012). 2335 ELISA kits for CYP26A1, B1 and C1 for several species are available. 2336 https://www.mybiosource.com 2337 Enzyme activity can be measured in microsomes as conversion of RA to its metabolites (Thatcher et 2338 al. 2010 analyzed human liver microsomes). 2339 RoDH 10: 2340 ELISA kits for several species are commercially available. 2341 https://www.mybiosource.com 2342 Stra8: 2343 mRNA and protein levels can be measured (Mark et al. 2008). 2344 cDNA and antibodies for multiple species are commercially available, as are ELISA kits 2345 https://www.mybiosource.com, https://www.labome.com, https://www.biocompare.com 2346 RARs, RXRs: 2347 Several assays are commercially available, many use HepG2 cells: 2348 http://alttox.org/application-of-medium-and-high-throughput-screening-for-in-vitro-toxicology-in- 2349 the-pharmaceutical-industry/ 2350 http://www.attagene.com (used in ToxCast™) 2351 https://ncats.nih.gov/tox21/projects/assays 2352 AhR: 2353 AhR mRNA induction or activation can be measured in e.g. human HepG2 cells; commercially 2354 available. 2355 http://www.attagene.com (Attagene used in ToxCast™) 2356 https://ncats.nih.gov/tox21/projects/assays (Tox21 used in ToxCast™) 2357 CYP1A1: 2358 CYP1A1 activity can be measured in microsomes 2359 A multitude of assays (mRNA, protein, activity) are commercially available. 2360 2361 A description of a cell-based method (using mouse pluripotent P19 cells) for identifying chemicals 2362 that disrupt RA signaling pathways was published by Chen and Reese in 2013. It has been used to 2363 test phthalate esters (Chen and Reese 2016). 2364 2365

63 2366 PubMed search strategy – retinoid parameters, chemical names, female/male 2367 reproduction parameters 2368 2369 Retinoid parameters were combined with chemical names (in Search table 1 below) and with either 2370 female or male reproduction parameters (in Search tables 2 and 3 below). No cutoffs (for e.g. year of 2371 publication) were used. The searches yielded several hundred of hits. However, only a small 2372 percentage was deemed relevant. In fact, a larger number of relevant articles was found via non- 2373 structured searches (using e.g. references in articles, citing articles, etc). 2374 2375 Retinoid parameters: (retinoic OR retinoid OR retinol OR retinyl OR retinal). Inclusion of the term 2376 “vitamin A” rendered the exact same number of hits as when this term was excluded. 2377 2378 The selected chemicals are a combination of a) chemicals handpicked by Nancy Baker and coworkers 2379 at EPA, from ToxCast™ and other sources; b) ToxCast™ chemicals with activity on RARs/RXRs and 2380 CYP1A1 (information found in tables in the D-DRP); c) CYP26 inhibitors found via text mining 2381 (information found in table in the D-DRP); and finally, d) chemicals classified as Persistent Organic 2382 Pollutants by the Stockholm Convention. 2383 2384 Search table 1 (chemicals) "ser 2–7" (methyl 3-(1H-imidazol-1-yl)-2,2-dimethyl-3-(4-(naphthalen-2-ylamino)phenyl)propanoate 1-(6-tert-Butyl-1,1-dimethyl-2,3-dihydro-1H-inden-4-yl)ethanone 1,2,3-Trichlorobenzene 1,3,5-Triisopropylbenzene 2,2-dimethyl-3-(4-(naphthalen-2- ylamino)phenyl)-3-(1,2,4)triazol- 1-yl-propionic acid methyl ester 2,4,5-Trichlorophenol 2,4,6-Tris(tert-butyl)phenol 2,4-Bis(1-methyl-1-phenylethyl)phenol 2,6-Di-tert-butyl-4-ethylphenol 2,6-Di-tert-butyl-4-methoxyphenol 2-Mercaptobenzothiazole 2-Naphthylamine 3-(6-(2-dimethylamino-1- imidazol-1-yl-butyl)naphthalen-2- yloxy)-2,2-dimethyl-propionic acid 4,4'-Sulfonylbis[2-(prop-2-en-1-yl)phenol] 4-Nonylphenol 4-Nonylphenol, branched 4'-octyl-4-biphenylcarboxylic acid AC-41848 hydrate Acitretin Adapalene Aldrin all/trans-Retinoic acid Alpha-hexachlorocyclohexane AM580 AR7 Azinphos-methyl Bensulide Benzyl alcohol Beta-hexachlorocyclohexane Bifenazate

64 BMS 493 BMS-189453 BMS-195614 Butylated hydroxytoluene CD437 Chlordane Chlordecone Chlorthal-dimethyl Citric acid Clorophene Coumaphos CP-532623 DDT OR Dichlorodiphenyltrichloroethane Dieldrin diethylaminobenzaldehyde Difenoconazole Diniconazole Dioxin OR dibenzofuran Dodecyl gallate Dysprosium(III) chloride Econazole nitrate Endosulfan Endrin EPN OR ethyl phenylphosphonothioate Esfenvalerate Fludioxonil Gentian Violet Heptachlor Hexabromobiphenyl Hexabromocyclododecane Hexabromodiphenyl ether and heptabromodiphenyl ether Hexabromodiphenyl ether OR heptabromodiphenyl ether Hexachlorbenzene Hexachlorobutadiene Imazalil Imidazole Isoniazid Isopentyl benzoate LE 135 Liarozole Lindane OR gamma-hexachlorocyclohexane Mepanipyrim Methoprene acid methyl 3-(4-(6-bromopyridin-3- ylamino)phenyl)-3-(1H-imidazol- 1-yl)-2,2-dimethylpropanoate Mirex N,N-Dimethylformamide Nicotine

65 N-Phenyl-1,4-benzenediamine Organotin Pentachlorobenzene Pentachlorphenol Perfluorooctanesulfonic acid OR perfluorooctanesulfonyl fluoride Phosalone polychlorinated AND (naphtalene OR naphtalenes OR napthalene OR napthalenes) Polychlorinated biphenyls OR PCBs Polychlorinated AND (dibenzo-p-dioxins OR dibenzofurans) Prochloraz Pyraclostrobin R115866 OR rambazole R116010 SR271425 Symclosene Talarozole Tetrabromodiphenyl ether and pentabromodiphenyl ether Tetrabromodiphenyl ether OR pentabromodiphenyl ether Tetrabutyltin Toxaphene Tributyltin benzoate Tributyltin chloride Tributyltin methacrylate Triflumizole Triphenyltin hydroxide Triticonazole TTNPB

2385 2386

66 2387 2388 Search table 2 (female reproduction parameters) “female genitalia” “female gonad” “Female reproduction” “Granulosa cells” “Polycystic ovarian syndrome” “Theca cells” Abortion Cervix OR cervical Corpus Luteum Decidua OR decidualization Eclampsia Embryo OR embryonic Endometriosis Endometrium OR endometrial Estrous Fallopian Fertility OR infertility Fetus OR fetal Folliculogenesis Gestational Lactation Litter Luteal Maternal Menopause Menstrual OR menstruation Oocyte Oogenesis Oogonia Ovary OR Ovarian Ovulation Placenta OR placental Pregnancy OR pregnant Prenatal OR postnatal Puberty Trimester Uterus OR uterine Vagina OR vaginal 2389 2390

67 2391 Search table 3 (male reproduction parameters) Aspermia Asthenozoospermia Azoospermia Blood-Testis Barrier Cryptorchidism Ejaculation epididymis Epididymitis erectile fertilization Hemospermia Hypospadia impotence Leydig male Infertility Oligospermia Orchitis penile prostate prostatic Prostatitis Rete Testis scrotum Seminiferous Sertoli Sperm Spermatid Spermatocele Spermatocyte Spermatogenesis Spermatogonia Spermatozoa Testes Testicular Testis 2392

68