Genomic and Hormonal Biomarkers of Phthalate-Induced Male Rat Reproductive Developmental Toxicity Part II: A Targeted RT-qPCR Array Approach That Defines a Unique Adverse Outcome Pathway

Abstract

Previously, we demonstrated that exposure to some diortho-phthalate esters during sexual differentiation disrupts male reproductive development by reducing fetal rat testis testosterone production (T Prod) and gene expression in a dose-related manner. The objectives of the current project were to expand the number of test compounds that might reduce fetal T Prod, including phthalates, phthalate alternatives, pesticides, and drugs, and to compare reductions in T Prod with altered testis mRNA expression. We found that PEs that disrupt T Prod also reduced expression of a unique “cluster” of mRNAs for about 35 genes related to sterol transport, testosterone and insulin-like hormone 3 hormone syntheses, and lipoprotein signaling and cholesterol synthesis. However, phthalates had little or no effect on mRNA expression of genes in peroxisome proliferator-activated receptor (PPAR) pathways in the fetal liver, whereas the 3 PPAR agonists induced the expression of mRNA for multiple fetal liver PPAR pathway genes without reducing testis T Prod. In summary, phthalates that disrupt T Prod act via a novel adverse outcome pathway including down regulation of mRNA for genes involved in fetal endocrine function and cholesterol synthesis and metabolism. This profile was not displayed by PEs that did not reduce T Prod, PPAR agonists or the other chemicals. Reductions in fetal testis gene expression and T Prod in utero can be used to establish relative potency factors that can be used quantitatively to predict the doses of individual PEs and mixtures of phthalates that produce adverse reproductive tract effects in male offspring.

Administration of di-ortho phthalate esters with straight chain lengths of 3–7 carbons during male rat sexual differentiation induces “Phthalate Syndrome” (Foster, 2005; Gray and Foster, 2003), which is a unique phenotype seen in exposed male offspring that includes reduced anogenital distance (AGD), retained female-like areolae/nipples in infant and adult males, agenesis of the epididymis, vas deferens, and sex accessory glands, testicular abnormalities and agenesis or elongation of the gubernaculum resulting in testis nondescent. These abnormalities result directly from reductions in testosterone production (T Prod) and insulin-like hormone 3 (Insl3) production by the fetal testis, which in turn are caused by reductions in the expression of mRNA for key testis genes including those involved in hormone synthesis, sterol transport, and lipid metabolism and cholesterol synthesis (Gray et al., 2016; Hannas et al., 2012; Lahousse et al., 2006).

This study is a continuation and expansion of 2 previous series of experiments. One of these measured the effects of chemical exposures administered during the masculinizing window, including several phthalates, on fetal testis T Prod (Furr et al., 2014). Furr et al. (2014) demonstrated that the phthalates that reduced T Prod were restricted to phthalates that have straight chain lengths of 3–7 carbons in the para position and diisononyl phthalate (DINP), a mixture of isononyl esters. The second study examined the effects of 6 phthalates on gestational day (GD)18 T Prod and testis mRNA expression using a custom 96-well reverse transcriptase quantitative PCR (qRT-PCR) array containing key target genes representing sexual determination and differentiation, steroidogenesis, gubernaculum development, and androgen signaling pathways (Hannas et al., 2012). Diisobutyl (DIBP), dipentyl (DPeP), dihexyl (DHP), diheptyl (DHeP), and DINP all reduced T Prod and these 5 phthalates also down-regulated a cluster of mRNA transcripts measured with the custom arrays including Cyp11b1, Scarb1, Star, Cyp17a1, Cyp11a1, Hsd3b, Insl3, Cyp11b2, Lhcgr, Dhcr7, and Inha, whereas diisodecyl phthalate (DIDP) did not reduce T Prod or mRNA expression levels (Hannas et al., 2012).

Although scientists have been studying the in utero effects of phthalate esters on sexual differentiation in the rat for over 20 years (Gray et al., 2000; Mylchreest et al., 1998) and several key events in the adverse outcome pathway (AOP) are known, the molecular initiating event (MIE; or proximate receptor) is unknown. Over the past several decades, there has been an ongoing debate with regards to the role of the peroxisome proliferator-activated receptors (PPARs) in phthalate-induced reproductive toxicity. The OECD AOP Wiki website includes an AOP that reports that the phthalates induced fetal testis toxicity is regulated by PPAR activation (AOP 18; Nepelska et al., 2015). PPARs are nuclear receptors that play roles in regulating cellular differentiation, development, and metabolism, including cholesterol uptake and transport. Several scientists have proposed that phthalates interfere with testicular development and function or T Prod by interacting with PPARs (Bhattacharya et al., 2005; Boberg et al., 2008; Corton and Lapinskas, 2005; Gazouli et al., 2002; Plummer et al., 2007) and interfering with cholesterol transport (Borch et al., 2006). In contrast, we demonstrated that administration of the potent PPAR agonist WY 14643 during sexual differentiation of the rat did not reduce fetal T Prod (Furr et al., 2014). Similarly, administration of the PPARγ agonist rosiglitazone from GD7 to 19 and 21 did not induce phthalate -like effects on the fetal testis or the male offspring (Boberg et al., 2008). In fact, no data support the hypothesis that PPAR agonists produce effects that resemble the effects of phthalates on the expression of the PPAR pathway or reduce gene expression in the fetal testis during the sexual differentiation period or induce reproductive abnormalities in male offspring. In this study, we expanded the dataset for phthalate induced alterations of T Prod and gene expression and directly tested the hypothesis that administration of 3 potent PPAR agonists, WY-14643, HFPO-DA, and clofibrate, during sexual differentiation GD14–18 would reduce fetal testicular T Prod, as is seen with phthalate exposure. We also examined if these PPAR agonists altered gene expression in the fetal testis and fetal liver.

MATERIALS AND METHODS

Chemicals

Chemicals studied in the current project include phthalates, phthalate alternatives, drugs, pesticides, and toxic chemicals that act as androgen receptor (AR) antagonists or inhibit testosterone synthesis, paracetamol, and PPAR agonists WY 14643 (pirinixic acid), clofibrate, perfluorooctane sulfonate (PFOS), and hexafluoropropylene oxide dimer acid (HFPO-DA) . The chemicals included in this study are listed in Figure 2. The chemicals, the source of the chemicals and purity for blocks 1–66 are presented in a Supplementary Data file associated with Furr et al. (2014) and in a Supplementary Table 1 herein for subsequent blocks.

Experiments

The overall design of this project and sequence of experiments is shown in Figure 1. T Prod was measured at necropsy and mRNA was extracted from fetal tissues for RT-qPCR analysis using our custom arrays (Qiagen). Initially chemicals (Figure 2A ) were administered at a single, high dosage level that did not produce overt maternal toxicity to screen for those that did or did not reduce fetal T Prod (Furr et al., 2014). Those found to reduce T Prod (Figure 2B) were subsequently evaluated in dose response studies designed to determine the full range from inactive to a maximum reduction so that relative potencies based upon ED₅₀s could be determined.

Following this, twenty additional RT² Profiler PCR Arrays (Qiagen) were screened for potential phthalate-induced alterations and additional experiments were conducted with 2 of these pathway arrays to verify alterations in mRNA expression noted in the “screening” study. The effects of 9 phthalates (Figure 4), 8 active (ie, reduced T Prod) and 1 inactive, on mRNA levels on the Lipoprotein Signaling & Cholesterol Metabolism Array (Qiagen, PARN-080Z) were assessed, after which the dose-related effects of dibutyl phthalate (DBP) were examined so that median effect dose (ED₅₀) values of the alterations on this array could be compared with the effects of DBP on T Prod in the GD18 Charles River Sprague Dawley (CRSD) fetal rat.

In addition, we evaluated the effects of 7 active phthalates on mRNA expression on the Drug Metabolism: Phase 1 Enzymes Array (Qiagen, PARN-068Z). Since 3 of 5 genes affected on this array recapitulated effects on the custom arrays, we did not pursue any additional experiments with PARN 068Z. In addition, we also evaluated the effects of 3 PPAR agonists and 1 active phthalate on maternal liver, fetal liver, and fetal testis mRNA levels in the PPAR Targets Array (Qiagen, PARN-149Z) to determine if the PPAR agonists effects on gene expression in the fetus and T Prod were similar to an active phthalate (Figure 7). These results also enabled us to determine that all 3 PPAR agonists cross the placenta and affect fetal liver gene expression.

Animals—Rats

This project was conducted over several years in multiple blocks (Supplementary Table 1). Each block consisted of about 15 pregnant rats that were typically divided into 4–5 different treatment groups with 3–4 dams/group.

For the first set of blocks adult, 70- to 90-day, old virgin female Harlan Sprague Dawley (SD) rats (Harlan Laboratories, Inc., Indianapolis, Indiana) were mated by the supplier and shipped on GD2. Mating was confirmed by sperm presence in vaginal smears by the supplier (GD0 = bred date, GD1 = plug positive date). Following this, the blocks were conducted with adult 70- to 90-day old virgin female SD (Crl:(CD)SD) rat dams from Charles River Laboratories. Dams were housed individually in clear polycarbonate cages (20 × 25 × 47 cm) with laboratory-grade heat-treated pine shavings (Northeastern Products, Warrensburg, New York) as bedding. Pregnant dams were fed NIH07 Rat Chow and filtered (5 µm filter) municipal drinking water (Durham, North Carolina) ad libitum.

Pregnant rats were maintained on a 12:12 h. photo period (light/dark cycle, lights off at 18:00) and 20°C–22°C temperature with a 45–55% relative humidity. Water was tested monthly for Pseudomonas and every 4 months for a suite of chemicals, including pesticides and heavy metals. This study was conducted under protocols approved by laboratory Institutional Animal Care and Use Committee and the Association for Assessment and Accreditation of Laboratory Animal Care.

Dosing and Administration of Chemicals

Pregnant rat dams were randomly assigned to treatment groups on GD14 in a manner that provided each group with similar means and variances in body weight. Dams were weighed and dosed daily by oral gavage at approximately 07:30 h. (EST) from GD14 to 18 with the vehicle (laboratory-grade corn oil [CAS no. 8001-30-7] or deionized distilled water) or the chemical plus vehicle at 2.5 ml/kg. Corn oil was the vehicle of choice for most chemicals except HFPO-DA (GenX) which was administered in high-performance liquid chromatography (HPLC) grade water, and PFOS which was administered in HPLC grade water with 0.5% Tween-80. Paracetamol and DiBP were administered to dams at 0 or 400 mg/kg/day (3 litters/group) on GD17–21 in block 129.

Fetal necropsies

Dams were rapidly euthanized by decapitation on GD18 or 21 and fetuses were removed, decapitated, and dissected under a Leica MZ6 dissecting microscope (Wetzlar, Germany). For all experiments, a single testis from 3 males in a litter were removed and used for evaluating ex vivo T Prod. All remaining testes in each litter were pooled, immediately homogenized in TRI reagent (Sigma, Saint Louis, Missouri) on ice using a Kontes pestle homogenizer, and stored at −80°C until used to extract RNA. All necropsies began 1 h following administration of the final maternal dose and were conducted within a 2-h time frame between 08:00 and 10:00 h (EST) to avoid any potential confounding effects of fetal growth or time of day on the fetal endpoints.

Ex vivo fetal testicular T Prod

Following removal, fetal testes were immediately transferred individually into a single well on a 24-well plate containing 500 µl of M-199 media without phenol red (Gibco Life Technologies, Product no. A31224DK), supplemented with 10% dextran-coated charcoal stripped fetal bovine serum (Hyclone Laboratories, Logan, Utah). Testes were incubated for 3 h at 37°C on a rotating platform. Following incubation, media was removed and stored at −80°C until used for testosterone measurement. The level of testosterone in the media samples was measured by radioimmunoassay according to the manufacturer’s instructions (Diagnostic Products Corporation Coat-A-Count kits, Siemens Corp, Los Angeles, California) for blocks 1–84 after which time the kit was no longer was available. The intraassay coefficient of variation for this kit was 3.1% (based on variability of the standard curve) and the interassay coefficient of variation was 13.7%. Cross-reactivity of the T antibody with dihydrotestosterone (DHT) was 3.2%. The limit of detection was 0.2 ng testosterone/ml. T Prod after block 84 was measured using ALPCO kits (Salem NH, Cat no. 72-TESTO-CT2). The intraassay coefficient of variation for this kit was 5.6% (based on variability of the standard curve) and the inter-assay coefficient of variation was 8%. Cross-reactivity of the T antibody with DHT was 2.6%. The limit of detection was about 0.08 ng testosterone/ml. T Prod for the blocks 1–66 are presented in Furr et al. (2014) and the new data are presented here.

Fetal testis gene expression analysis

In addition to measuring fetal testis T Prod ex vivo, in this study we measured messenger ribonucleic acid (mRNA) expression levels in the fetal rat testis using targeted, custom-designed qRT-PCR 96-well gene arrays (arrays described in detail by Conley et al., 2019; Hannas et al., 2012). Each array includes wells to quantify mRNA for 89 genes associated with sex determination, steroid and peptide hormone synthesis and transport, and PPAR activation.

RNA was extracted from the fetal testes homogenate as previously described ( Hannas et al., 2012) and cleaned to eliminate any potential genomic DNA contamination using Qiagen RNeasy Mini Kit (Valencia, California) with the on-column DNase treatment step according to the manufacturer’s instructions. RNA concentration and purity were determined with a NanoDrop 2000 Spectrophotometer (Thermo Scientific, Wilmington, Delaware). An additional genomic DNA elimination reaction and complementary DNA (cDNA) synthesis was performed on the RNA samples using the Q RT² First Strand Kit according to the kit instructions.

For each individual sample, 300 ng of RNA was added to a single reaction, to be used across a 96-well array plate. The template cDNA was then added to RT² SYBR Green qPCR Master Mix (Qiagen) and 25 μl was added to each well of the plate. The 96-well gene array plate (purchased from Qiagen was custom designed to contain 89 individual target genes, 4 housekeeping genes (beta-actin, beta-2 microglobulin, beta-glucuronidase, and lactate dehydrogenase A), a genomic DNA control, a reverse transcription control and a positive PCR control (Hannas et al., 2012).

The PCR reaction was run on an iCycler iQ Real-Time Detection System (Bio-Rad, Hercules, California) using the following cycling parameters: 1 cycle of 95°C for 10 min, followed by 40 cycles of 95°C for 15 s, and 60°C for 1 min. Product purity was verified by melting curve analysis. The ΔC_T value for each gene was determined by dividing the gene C_T value by the mean C_T value generally using of the housekeeping genes. The selection of housekeeping genes was based upon the consistency of each transcript across all the arrays in a block. The $2−ΔΔCT$ method was used to analyze data and change in gene expression levels were reported as fold change (Tusher et al., 2001) or the ratio of the phthalate-treated sample group to the respective control group. The sensitivity, specificity and reproducibility of the gene array system are discussed in detail by Arikawa et al. (2011).

In addition to the custom arrays, we screened active phthalates for activity using 20 additional RT-qPCR pathway arrays to determine if the phthalates that reduced T Prod were also impacting other pathways in addition to those assessed on the custom array. Initially for screening all the arrays, except the PPAR Targets array, the mRNA from 3 control litters and 3 litters from a block exposed to a high dose of an active phthalate were pooled into a single sample and run on the rat variant of the following RT² Profiler Arrays: The 20 commercial pathways included:

1. Cell death pathway finder (PARN-212Z),

2. Cell lineage identification (PARN-508Z),

3. Lipoprotein Signaling and Cholesterol Metabolism (PARN-080Z),

4. Drug Metabolism Phase I enzymes (PARN-068Z)

5. Drug Metabolism Phase II enzymes (PARN-069Z),

6. Embryonic stem cell (PARN-081Y),

7. Growth factors (PARN-041Z),

8. Hedgehog signaling pathway (PARN-078Z),

9. Homeobox HOX genes (PARN-083Z),

10. Insulin signaling pathway (PARN-030Z),

11. JAK STAT signaling pathway (PARN-039Y),

12. Neurotrophins and receptors (PARN-031Z),

13. Nitric Oxide signaling pathway (PARN-062Z),

14. Nuclear receptors and coregulators (PARN-056Y),

15. Osteogenesis (PARN-026Z),

16. PPAR targets (PARN-149Z),

17. Signal transduction Pathway Finder (PARN-014Z),

18. VEGF signaling (PARN-091Z),

19. WNT signaling pathway (PARN-043Z) and

20. WNT signaling targets (PARN-243Z).

The fold values for the control and treated samples from the above arrays (not including the PPAR Targets array) were compared with identify pathway arrays where several genes were decreased by >50% or increased by >2-fold (Supplementary File). As a result, we conducted 2 experiments with the Lipoprotein Signaling and Cholesterol Metabolism (PARN-080Z) array and 1 with the Drug Metabolism Phase I Enzymes (PARN-068Z).

In the first of 2 experiments with Lipoprotein Signaling and Cholesterol Metabolism arrays we measured the fold values for a single high dose of 9 individual phthalates including diisohexyl (DiHP; 750 mg/kg), DHP (750 mg/kg), DPeP (100 mg/kg), DiBP (900 mg/kg), DiNP (1500 mg/kg), dicyclohexyl (DCHP; 900 mg/kg), diethylhexyl (DEHP) (750 mg/kg), DBP (300 mg/kg/day), and diethyl (DEP) (900 mg/kg/day). All these phthalates reduce T Prod except for DEP (n = 3 litters/phthalate). Individual phthalate fold values were statistically compared with control mRNA fold values and we also ran an orthogonal contrast comparing the fold values of the 8 positive phthalates versus the 1 negative phthalate. Following this experiment, a dose response study with the Lipoprotein Signaling and Cholesterol Metabolism array was conducted with DBP (samples pooled from 5 blocks, at 0 [n = 10 litters], 33 [n = 8], 50 [n = 3], 100 [n = 6], 300 [n = 10 ], 600 [n = 7], and 900 mg/kg/day [n = 9]) to determine ED₅₀ values for the affected genes.

In the screening level assessment with the pathway arrays (listed above several gene expression levels were altered on the Drug Metabolism Phase 1 Enzymes array [PARN-068Z] ). As a follow-up experiment 7 phthalates that were shown to reduce T Prod and gene expression on our custom array after treatment from GD14 to 18 were evaluated in this gene array at a high dosage level including DiHP (750 mg/kg), DHP (750 mg/kg), DPeP (100 mg/kg), DiBP (900 mg/kg), DiNP (1500 mg/kg), DCHP (900 mg/kg), DEHP (750 mg/kg; n = 3 litters/phthalate). This array contains 84 genes involved in reactions including oxidation, reduction, hydrolysis, cyclization, and decyclization.

In the next experiment, 3 PPAR agonists were compared with the effects of the potent phthalate DPeP on the dam and fetus in a series of blocks to test the hypothesis posed in OECD AOP 18 that claims the phthalates reduce fetal T Prod and disrupt male reproductive tract differentiation by acting as PPAR agonists. The effects of DPeP (300 mg/kg/day) on T Prod and testis, maternal and fetal liver gene expression were compared with the potent PPAR agonists WY 14643 (300 mg/kg/day; blocks 93 and 99; Table 1), clofibrate (100–400 mg/kg/day; blocks 103 and 106; Supplementary Table 1), and HFPO-DA (in blocks 124 and 127; Supplementary Table 1). The effects of HFPO-DA on fetal, neonatal, and maternal PPAR pathway were recently described by (Conley et al., 2019, 2021). Test chemicals were administered orally on GD14–18, and rats necropsied on GD18 to determine if phthalates like DPeP act via the PPAR pathway similar to WY 14643, clofibrate and HFPO-DA (GenX) or if the 3 PPAR agonists reduced T Prod as suggested (Martin et al. 2011) (OECD AOP WIKI [https://aopwiki.org/aops/18] last accessed July 7, 2020). In this experiment, fetal T Prod, maternal liver weight and fetal and maternal liver mRNA expression levels on genes represented on the PPAR pathway array (PARN 149) were measured. In addition, the effect of WY 14643 (n = 5) and DPeP (n = 5) at 300 mg/kg/day on the PPAR Targets array also were measured in the fetal testis to determine if these treatments activated this pathway in this tissue.

Data analysis and statistics

Litter mean values were used to determine if T Prod was affected by the test chemical administration using data from individual testis incubations (1 testis/male, from 3 males/litter). T Prod, maternal weight gain and liver weight (with body weight at GD18 as a covariate) and litter size were log10-transformed to correct for heterogeneity of variance and then analyzed with analysis of variance (ANOVA) model to detect significant differences using PROC GLM in SAS, version 9.4 (SAS Institute, Cary, North Carolina). Post hoc t tests were performed with the LSMEANS option to compare individual doses when a significant (p < .05) overall dose effect was determined by the ANOVA model.

Gene expression levels were analyzed using pooled testes samples from each litter (1 sample/litter). Fold changes in mRNA expression were considered to be statistically significant if the F-value for the overall ANOVA model was p < .01, followed by t test (p < .01). Expression levels with a mean C_T value >33 were considered as “undetected.”

To determine ED₅₀s, dose-response T Prod data were converted to percent of control and analyzed using a nonlinear 4 parameter regression analysis (sigmoidal fit with variable slope using Prism GraphPad 7.01 software, GraphPad Software, Inc., La Jolla, California). The top and bottom parameters were constrained to 100% and 0% of control, respectively. Dose-response of gene expression fold reductions were determined using a sigmoidal fit with variable slope model with the top constrained to 1 and the bottom constrained to 0. Potency comparisons between treatments were made using the ED₅₀ values derived from dose response curves.

Proc Factor on SAS 9.4 (method = principal components) was used to generate factor scores data for the different groups to statistically describe the degree of separation between them based upon all the affected variables on the custom arrays with and without T Prod in the model. Factor analysis is a method of data reduction and it does this by seeking underlying unobservable (latent) variables that are reflected in the observed variables (manifest variables).

Analysis of the fetal testis custom array data included fold expression values for 14 mRNAs and T Prod (as shown in Figure 1; n = 40 chemicals/treatments). The number of factors retained in the determination of the scores included those with eigenvalues > 1 and the total variance of all the standardized values for the 15 variables were calculated. A single factor score for each litter was determined using SAS 9.4 (as above) and we then grouped the chemicals into 4 clusters (phthalates that reduce T Prod, phthalates that do not reduce T Prod, nonphthalates that reduce T Prod, and PPAR agonists) and compared the mean (±SE) factor scores for each cluster using the LSMEANS option of PROC GLM.

Analysis of the Lipoprotein Signaling and Cholesterol Metabolism array (PARN-080Z) and the Drug Metabolism Phase 1 Enzymes array (PARN-068Z) fold values were analyzed on using PROC GLM and phthalate groups compared with control using LSMEANS. In addition, for the mRNA fold expression levels on Lipoprotein Signaling and Cholesterol Metabolisms array the 8 phthalates that reduced T Prod were pooled and compared with the 1 phthalate that does not reduce T Prod and the expression levels that were consistently and significantly affected (p < .001) were ranked in order of the magnitude of the changes. The fold values for the control and treatment groups for the 84 mRNA levels on the PPAR Targets arrays were all compared with one-another using the LSMEANS option in PROC GLM.

In addition, fetal liver factor scores derived from mRNA expression levels on the PPAR Targets array was conducted, as above, with data the from the different blocks pooled together (control [n = 13 litters], DPeP [n = 5], HFPO-DA [n = 9], WY 14643 [n = 5], and clofibrate [n = 10]) using 27 (out of the 84 mRNA) fold values that were significantly altered by treatments (vs control fold = 1) in order to display the statistical separation among the 4 chemicals.

RESULTS

T Prod and Custom Array

The results of this study include T Prod data, maternal weight gain, and fetal viability on several chemicals (Supplementary File), measured since the publication by Furr et al. (2014) and data on the effects of phthalates, phthalate alternatives, pesticides, drug, and toxic substances (Figs. 2A–D) on fetal testis mRNA expression of genes associated with sex determination, steroid and peptide hormone synthesis and transport, and PPAR activation on the custom array. The data also include the effects of several phthalates that reduce T Prod and one that does not on testis mRNA expression on an qRT-PCR array for Lipoprotein Signaling and Cholesterol Metabolism (PARN-080Z) genes. Phthalates that reduced T Prod also reduced mRNA expression of about 35 genes in the fetal testis on GD18 on these 2 array platforms. This cluster includes gene involved in synthesis of key proteins essential for steroid hormone transport into the cell (Scarb1) and mitochondria (Star), rate limiting enzymes for steroid hormone synthesis (Cyp11a1 and Cyp17a1), peptide hormone (Insl3), gonadotrophin receptors (Lhcgr), testis development (Inha), and several genes involved with cholesterol metabolism (below). In contrast, phthalates that did not reduce testis T Prod did not alter mRNA expression on these arrays. Maternal weight gain was reduced by a few chemicals but only DEX and the high dose of Clofibrate reduced total maternal weight (Supplementary File). Among the chemicals, only the highest dose of DPeP (600 mg/kg/d) increased fetal mortality from about 2% on controls to 13% in the treated group (Supplementary File).

Figure 2A shows the mRNA expression levels (fold vs control) among the chemicals that were at a single high dose level and chemicals were classified as a “positive” or “negative” based upon ta reduction in T Prod. Figure 2B presents the results of the dose response studies with active (reduce T Prod) and Figure 2C presents the results of inactive (did not reduce T Prod) phthalate (DIDP) and the phthalate alternatives 1,2-cyclohexane dicarboxylic acid diisononyl ester (DINCH) and that failed to reduce T Prod and did not reduce mRNA expression.

Figure 2C also presents the results of 2 pesticides (vinclozolin and prochloraz), 2 industrial chemicals (bisphenol C [BPC] and 4-methoxyimadizole [a by-product of caramel coloring manufacturing]) and 2 drugs (dexamethasone and paracetamol) on mRNA expression on the custom arrays.

Dexamethasone reduced T Prod and 9 out of 14 of the mRNA levels from the “phthalate-cluster” on the custom arrays; however, the effects of dexamethasone occurred at dosage levels that are known to toxic to the dam and fetus (5 and 50 mg/kg/day; Figure 2C, Supplementary File) and for this reason it is unclear if these endocrine alterations are specific or related to the overt toxicity. Nr5a1 which codes for the transcription factor SF1 also was reduced in the 2 highest dose groups by 22% and 41%, respectively, and 6 other transcripts were significantly affected on the custom arrays in this block at 50 mg/kg/day (data not shown).

BPC reduced T Prod by about 42% and mRNA expression of 6 of the 14 genes in this cluster from the custom array at 300 mg/kg/day. BPC also reduced maternal weight gain at this dose, but it was not as overtly toxic as dexamethasone which reduced maternal weight, not just weight gain. Prochloraz and 4-methylimidazole (4-MI) both reduced T Prod in a dose-related manner but had little or no effect on custom array mRNA expression. Vinclozolin and paracetamol had no effect on T Prod or mRNA expression at any dose level.

Data for the 3 PPAR agonists found that WY 14643 reduced T Prod by 28% at 300 mg/kg/day but had no effect on mRNA in the “phthalate cluster.” Clofibrate had no effect on T Prod at dosage levels up to 400 mg/kg/day but reduced expression of several mRNAs in the “phthalate-cluster,” and HFPO-DA (Conley et al., 2019) had no effect on T Prod or mRNA expression on the custom arrays at any dose level (Figure 2D).

Factor Analysis of Custom Array Data

The factor analysis of fold expression values of 14 mRNAs on the custom arrays and T Prod (n = 40 treatments) (Figure 3) retained 2 factors with eigenvalues > 1 (Factor 1 eigenvalue = 9.96, 66.4% of the standardized total variance of 15; Factor 2 eigenvalue = 1.89, 12.6% of the standardized total variance, cumulative variance described = 79% of 15). The standardized scores for 15 variables contribution each factor was multiplied by each groups value for each variable (a 1 × 15 by 15 × 1 matrix) yielding a single value for each treatment on both factors (Figure 3). Statistical analysis of the group means on each factor demonstrates that only the phthalate-negative and PPAR agonist clusters did not differ from one another on the custom arrays. All the other contrasts differed significantly (p < .01). Excluding T Prod from the variable list and using only the 14 mRNA expression values did not alter the statistical comparisons among the clusters on Factor 1. With factor 2 (data not shown), the only group that differed significantly from the others was the nonphthalate chemicals that reduced T Prod.

Assessment of mRNA Alterations in the Fetal Testis Using Multiple Pathway Arrays

Data from 19 pathway arrays examined to screen for potential alterations of mRNA expression are presented in Supplementar File 2 (excluding no. 16. PPAR Targets pathway [PARN-149Z]). The data are presented by C_T cycle and fold of the mRNA from phthalate-treated testes versus control litters (sorted by fold change, by array and alphabetically on separate excel sheets within S2). The arrays (no. 3 on the list) for Lipoprotein Signaling and Cholesterol Metabolism (PARN-080Z) and (no. 4) Drug metabolism Phase I Enzymes (PARN-068Z) were selected for follow-up studies because several mRNAs were affected.

Lipoprotein Signaling and Cholesterol Metabolism Array

Analysis of the Lipoprotein Signaling and Cholesterol Metabolism array (PARN-080Z) revealed that more than a dozen fetal testis mRNA expression levels were decreased by exposure to the 8 phthalates that reduced T Prod but not by DEP, a phthalate that does not reduce T Prod (Figure 4A). All the affected genes were involved in cholesterol metabolism. In addition to the mRNA fold expression levels on the Lipoprotein Signaling and Cholesterol Metabolism array, when mRNA data for the 8 phthalates that reduced T Prod were grouped and compared with the one phthalate (DEP) that does not reduce T Prod the expression levels were consistently affected (15 down- and 1 upregulated, p < .001) by the active phthalates but not by DEP. The data in Figure 4A were ranked by fold values for the 8 active phthalates with the most affected (lowest fold vs control) at the top (Figure 4A). In the DBP dose response study, the effects on mRNA expression in this pathway in Figure 4B were ranked by the ANOVA F-value with the largest value at the top.

In the second experiment with the Lipoprotein Signaling and Cholesterol Metabolism array, DBP was administered at several dosage levels to CRSD pregnant rats to identify the sensitivity of the transcript levels to phthalate administration (4b) in the fetal testis. In the DBP dose response experiment, T Prod, and mRNA for 18 genes were significantly reduced in a dose-related manner (p < .0003). Nine genes were reduced in a dose related manner with ED₅₀ values <1000 mg/kg/day and another 6 had ED₅₀s between 1000 and 2000 mg/kg/day (Figure 4B). Again, none of these declined to 0% of control and plateaued at levels ranging from 15 to 80% of control. The LOEL was 100 mg DBP/kg/day and the NOEL was 50 mg/kg/day.

The transcripts on this array and their role are as follows: (yellow highlight p < .01, blue p < .05)

Cholesterol Metabolism

Cholesterol Absorption Cel and Ldlr.

Cholesterol Catabolism Akr1d1, Apoe, Cel, Cyp39a1, Cyp46a1, Cyp7a1, Scarf1, Snx17, and Trerf1.

Cholesterol Homeostasis Abca1, Abcg1, Angptl3, Apoa1, Apoa2, Apoa4, Apoe, Lcat, Ldlr, Ldlrap1, Pcsk9.

Cholesterol Biosynthesis Acaa2, Cnbp, Cyb5r3, Cyp51, Dhcr24, Dhcr7, Ebp, Fdft1, Fdps, Hmgcr, Hmgcs1, Hmgcs2, Idi1, Mvd, Mvk, Nsdhl, Pmvk, Prkaa1 (Ampk), Prkaa2, Prkag2, RGD1564999, and Tm7sf2.

Other Cholesterol Metabolism Genes Abca2, Apob, Apoc1 (QuantiNova Symbol: NEWGENE_2134), Apoc3, Apof, Apol2, Cela3b, Cyp11a1, Cyp7b1, Hdlbp, Il4, Insig1, Insig2, Lep (Leptin), Lipe, Mbtps1, Nr0b2, Nr1h2, Nr1h3, Nr1h4, Osbpl1a, Osbpl5, Ppard, Scap, Soat1, Soat2, Sorl1, Srebf1, Srebf2, Stard3, and Vldlr.

ED50 < 1000 mg/kg/day

Eight genes in cholesterol biosynthesis, 1 with metabolism

Drug Metabolism Phase I Enzymes (PARN-068Z) Array Results

Seven phthalates also were examined using Rat Drug Metabolism: Phase 1 Enzymes arrays including DiHP (750 mg/kg), DHP (750 mg/kg), DPeP (100 mg/kg), DiBP (900 mg/kg), DiNP (1500 mg/kg), DCHP (900 mg/kg), DEHP (750 mg/kg), and control. Of the 84 genes interrogated, the 7 phthalates all significantly reduced mRNA expression for 4 genes on the Drug Metabolism Phase I Enzymes (Figure 5). Three of which recapitulated effects seen with the custom arrays including Cyp11b1, Cyp11a1, Cyp17a1, and the fourth and novel observation was ALDH₂, which is a member of the aldehyde dehydrogenase family and has a role in the metabolism of alcohol.

Effects of 3 PPAR Agonists Versus a Phthalate

The experiment that examined the effects DPeP and DiBP, phthalates that reduce T Prod, versus 3 PPAR agonists on the PPAR targets arrays found that only the PPAR agonists had major effects on the PPAR pathway in the fetal and maternal livers. Furthermore, among these 5 chemicals, DiBP (mRNA from block 19) and DPeP (blocks 93 and 99) failed to increase maternal liver weight whereas all the PPAR agonists (clofibrate [blocks 103 and 106], HFPO-DA [Conley et al., 2019]) and WY 14643 (blocks 93 and 99) increased maternal liver weight (P < .0001). In addition, PPAR agonists had little or no effect on testis T Prod. Two of 3 PPAR agonists clofibrate (100–400 mg/kg/day), and HFPO-DA (some of the data were previously published by Conley et al., 2019) did not reduce fetal T Prod (Figure 6, Table 1) and WY 14643 at 300 mg/kg/day (Table 1, blocks 93 and 99) only reduced T Prod by 28%; whereas, DiBP and DPeP produce dose related reductions in T Prod up to about 90% (Figure 6, Table 1).

The analysis of fetal liver PPAR pathway mRNA expression levels among the treatments indicated that several mRNA transcripts were consistently affected by the PPAR agonists on the PPAR targets arrays (Figure 7, columns 4, 6 to 17). DPeP affected fewer genes in the fetal liver and to a lesser degree than did any of the 3 PPAR agonists (Figure 7, column 2).

In addition to the fetal liver, our analyses also compared the effects of DPeP and WY 14643 on mRNA expression in the fetal testis using the PPAR targets arrays (Figure 7). DPeP treatment had no effect on any of the mRNA expression values whereas WY 14643 increased mRNA expression levels for only 3 genes (Fabp1, Angptl4, and Hmgcs2) in the fetal testis (Figure 7, columns 3 vs 5). However, the fold inductions with WY 14643 in the fetal testis were much smaller than those in the fetal liver (Figure 7, contrast of columns 5 vs 4).

Figure 8 presents a heat map of the contrast among the 4 treatments with control fold = 1 and also compares the 3 PPAR agonists with fold values recalculated using DPeP as the “control” with fold values for DPeP = 1.

When the fold values for the mRNA transcripts affected by the PPAR agonists were analyzed using factor analyses, 3 factors were retained in the analyses with eigenvalues greater than 1 (factor 1 eigenvalue = 19.8, factor 2 = 2.79 and factor 3 = 1.43; Figure 9). The total variance of all the standardized values explained by the 3 factors was 89% of the total variance of 27 mRNA values showing statistically significant changes. Since factor 1 was the most significant factor explaining 73% of total variance the following analyses focus on factor 1. Control and DPeP levels had significantly lower mean scores on factor 1 than did the 3 PPAR agonists (p < .001; Figure 9) and HFPO-DA and WY 14643 had significantly higher factor 1 scores than did clofibrate (p < .02). It was evident from the dramatic increases in fetal PPAR pathway mRNA transcript expression that the fetus was exposed to each of the PPAR agonists.

In addition, it is important to note that the custom arrays assess mRNA transcript levels for key genes in the PPAR pathway including Acox1, Cyp4a1, Fabp1, Apoa1, and 3 PPAR receptors α, β, and γ. Among the several hundred custom arrays run on fetal testes with phthalates that reduced T Prod the expression levels of these transcripts were never significantly affected.

DISCUSSION

In the rat, the masculinization programming window corresponds with the initiation of fetal testicular T Prod and dramatic changes in fetal testis gene expression (Carruthers and Foster, 2005; Gray et al., 1999; Scott et al., 2008; Wolf et al., 2000). The consistent down-regulation of mRNA for several key regulatory genes involved in hormone action by phthalates that reduce T Prod supports the causative link between these genomic endpoints, the fetal T Prod, and the expression of postnatal male reproductive tract abnormalities. Figure 2A shows the differences in mRNA expression on the custom arrays among the chemicals that were run at a single high dose level to identify positives and negatives. The phthalate (DIDP) and the phthalate alternative (DINCH) that failed to reduce T Prod did not reduce mRNA expression. Several nonphthalate chemicals that reduced T Prod (linuron, prochloraz, and 4-methoxyimadizole) did not affect these mRNA levels either. Dexamethasone reduced T Prod and several of these mRNA levels, but only at dosage levels that reduced maternal body weight and weight gain over the dosing period (Supplementary Table 1). Similarly, BPC reduced T Prod by about 40% in the highest dose group (300 mg/kg/day) and expression of 6 of the 14 “phthalate cluster” mRNA levels on the custom arrays, but only at dose levels that reduced maternal weight gain (Table 1). Figures 2B1 and 2B2 present the results of dose response studies with active (reduce T Prod) and inactive phthalates (did not reduce T Prod).

The cluster of transcripts displaying reduced expression on our custom arrays due to in utero phthalate exposure includes those involved in synthesis of key proteins essential for steroid hormone transport into the cell (Scarb1) and mitochondria (StAR), rate limiting enzymes for steroid hormone synthesis (Cyp11a1 and Cyp17a1), synthesis of the peptide hormone Insl3, gonadotrophin receptors (Lhcgr), and testis development (Inha), among others. Scarb1 and Star are transcripts for SR-B1 and StAR proteins, which are involved in cholesterol transport in the cell and across mitochondrial membranes. Following delivery of this cholesterol precursor into the Leydig cell mitochondria, Cyp11A1, Hsd3b, and Cyp17A1 encode enzymes sequentially involved in the transformation pathway leading to androgen synthesis. The luteinizing hormone/choriogonadotropin receptor (Lhcgr) gene codes for a G protein-coupled receptor involved in regulating testosterone and INSL3 hormone secretion from the testes, critical for fetal testis development and descent (Klonisch et al., 2004). Similar to Lhcgr; Inha, and Dhcr7 were down-regulated by most, but not all, of phthalates that reduced T Prod. 7-dehydrocholesterol reductase (Dhcr7) is the protein product of the Dhcr7 gene, which converts 7-dehydrocholesterol to cholesterol in the sterol biosynthesis pathway. Reduction of Dhcr7 in the fetal testis has been previously documented following in utero phthalate exposure (Johnson et al., 2011; Lahousse et al., 2006). Likewise, phthalate-induced down-regulation of Inha in the fetal testis has been reported by several groups (Liu et al., 2005). This gene encodes the alpha subunit of Inhibin B, which is produced by the fetal Leydig cells and involved in testis development (Bardin et al., 1989). After birth Inhibin B production occurs in both the Leydig and the Sertoli cells (Majdic et al., 1997) and contributes to Sertoli cell population expansion. Therefore, a decrease in inhibin levels during gestation may have an effect on Sertoli cell differentiation, as has been reported following in utero (Liu et al., 2005; Plummer et al., 2007) or ex vivo testis (Li and Kim, 2003) phthalate exposure. Indeed, Liu et al. (2005) proposed that phthalate inhibition of Inhibin secretion by Leydig cells may be the cause for modified Sertoli cell development.

Down-regulation of Cyp11b1 mRNA in the fetal testes was consistently the most sensitive effect following in utero exposure to phthalates that reduced T Prod. Cyp11b2 mRNA also was reduced. These genes code for enzymes that are involved in adrenal, but not testis, hormone synthesis (Hu et al., 2007; Val et al., 2006). CYP11B1 or 11 b-hydroxylase is an enzyme expressed in the adrenal cortex that converts 11-deoxycortisol into corticosterone and Cyp11b2 codes for an adrenal enzyme active in aldosterone synthesis. These genes are not found in adult Leydig cells. Steroidogenic cells of the fetal adrenal and testis are thought to be derived from a common primordium that divides into separate tissues during embryogenesis and mixed adrenal and Leydig cell properties were found dispersed in the embryonic mouse testis (Val et al., 2006). In addition to the expression of Cyp11b1, the GD17.5 testis produces corticosterone upon ACTH stimulation (O’Shaughnessy et al., 2002). However, the levels of corticosterone are low and unlikely relevant to phthalate-induced abnormal reproductive differentiation.

In 2 separate experiments, we also evaluated the effects of phthalates on fetal testis mRNA expression levels using the Lipoprotein Signaling and Cholesterol Metabolism array (PARN-080Z). In the first experiment with this array platform, we found that 8 phthalates that reduced T Prod and mRNA expression on the custom arrays also altered fetal testis mRNA expression in genes that were involved in cholesterol metabolism. When the data from this array from the 8 phthalates that reduced T Prod were pooled and compared with the one phthalate (DEP) that did not reduce T Prod, the expression levels of 16 genes were consistently affected (15 down- and 1 upregulated, p < .001; Figs. 4A and 4B).

Effects of Pesticides, Drugs, and Toxic Substances on T Prod and Testis Gene Expression

Prochloraz

In contrast to the active phthalates the fungicide prochloraz reduced T Prod by about 40% but had no effect on custom array mRNA expression levels (Figure 2C). The reduction in T Prod is similar to those previously reported by (Blystone et al., 2007 ) who reported that prochloraz/kg/day reduced ex vivo fetal T Prod (NOEL 62.5 mg) and androstenedione production (NOEL 32 mg) by about 30% and 40%, respectively. In contrast, progesterone and 17α hydroxyprogesterone production were increased by almost 10-fold (NOEL 15 mg) due to direct inhibition of CYP17 hydroxylase activity. Blystone et al. (2007) also reported that prochloraz failed to reduce Cyp17, Star, and Cyp11a mRNA expression. In addition, prochloraz is an AR antagonist ( Blystone et al., 2009) and male rats, exposed in utero, were demasculinized and displayed reduced AGD, retained female-like areolae/nipples, reduced androgen-dependent organ weights, and hypospadias (Noriega et al., 2005) due to disruption of the androgen signaling pathway by these 2 MIEs.

4-Methylimidazole

The food coloring agent 4-MI reduced T Prod by about 25–40%, at dose levels that reduced maternal weight gain, but failed to alter custom array mRNA expression levels (Figure 2C). In a multigenerational study (Behl et al., 2020), dose levels of 4-MI up to 400 mg/kg/day did not reduce AGD and had little effect on female-like nipple retention and the males did not display reproductive tract malformations indicating that 4-MI had little or no effect on male rat sexual differentiation in utero. However, continuous 4-MI exposure delayed preputial separation and vaginal opening, reduced androgen-dependent organ weights and sperm counts, and induced histopathological lesions in the prostate, testis, and epididymis (LOEL of 750 ppm [50–60 mg/kg/day]) by inhibiting steroidogenesis in both males and females (Behl et al., 2020).

Dexamethasone

The drug dexamethasone (DEX) reduced T Prod and the expression of several of the genes on the custom arrays at high maternally toxic dosage levels (5 and 50 mg/kg/day) (Figure 2C, Table 1) and the fetuses were noticeably smaller than controls in the highest dose group. Lower dosage levels (0.05, 0.1, and 0.5 mg/kg/day) also reduced maternal weight but had no effect on T Prod on GD18. DEX is teratogenic to the rodent fetus with a no observed adverse effect level (NOAEL) of 0.010 mg/kg/day and higher dose levels reduce fetal weight and viability (WHO Food Additive Series 33, 2015). Administration of DEX by oral gavage from GD14 to 20 at 0.65, 6.5, and 13 mg/kg/day reduced maternal weight gain, placental and fetal weight and fetal viability were reduced in a dose-related manner (Wangui et al., 2019). In addition, 0.2 mg/kg/day in the drinking water from GD14 to 19 led to intrauterine growth retardation and reduced pup survival (Motta et al., 2018). Taken together, it is unclear if the testicular endocrine alterations seen in this study are related to direct effects on the testis of if they are a consequence of DEX-induced overt maternal and fetal toxicity.

Bisphenol C

BPC which is as potent as flutamide as an AR antagonist in vitro, reduced T Prod by about 42% and mRNA expression of 6 of genes on the custom array at 300 mg/kg/day (Figure 2C). BPC also reduced maternal weight gain at this dose (Table 1). When BPC was administered orally to pregnant rats on GD14 to 18 at 100 and 200 mg/kg/day the reproductive tract of the male offspring was minimally affected (Gray et al., 2019). For comparison, in utero administration of flutamide has been shown to induce malformations in 100% of males at 6 mg/kg/day (McIntyre et al. 2001).

Vinclozolin and flutamide

The fungicide vinclozolin and the drug flutamide failed to alter T Prod on GD18. In addition, vinclozolin did not reduce custom array mRNA levels (flutamide was not evaluated on arrays; Figure 2C). If these AR antagonists had any effect at all on T Prod, we hypothesized that vinclozolin and flutamide would increase, rather than decrease T Prod, similar to their effects in the pubertal rat (Monosson et al., 1999) and adult male rat (Matsuura et al., 2005). Testis T Prod is increased during pubertal and adulthood by potent AR antagonists because they block the action of androgens on the hypothalamic-pituitary axis resulting in an increase in serum LH and a subsequent stimulation of testis Leydig cell T Prod. However, in contrast to postweaning life, fetal T Prod is constitutive during sexual differentiation and not regulated by LH and the fetal rat pituitary does not begin to secret LH until late in fetal life and fetal testis Leydig cells do not express LH receptors until the very end of sexual differentiation as T Prod declines (Scott et al., 2009).

Paracetamol

In this study, administration of high dosage levels of paracetamol on GD14 to 18 and GD17 to 21 had no effect of fetal testis T Prod. Although several authors have proposed that paracetamol reduces fetal rat T Prod in vivo (Kristensen et al., 2012; Sharpe, 2020; Fisher et al., 2016), the literature provides conflicting support for this hypothesis. For example, Kristensen et al. (2011) reported that GD13.5 to 21.5 paracetamol exposure did not significantly reduce testis T Prod and intratesticular testosterone levels increased significantly rather than decreased as a result of treatment (p < .02; linear regression model of data from Supplementary Table 5 of Kristensen et al., 2011). Similarly, Mazaud-Guittot et al. (2013) reported that paracetamol or its metabolite at concentration of 10⁻⁴ to 10⁻⁷ M did not have any effect on human T Prod using a culture of human fetal testes exposed to paracetamol in vitro.

The effects of in utero paracetamol on AGD in male rats also are inconsistent. For example, paracetamol reduced AGD on GD21 at 250 or 350 mg/kg/day in 1 experiment but not the other experiment in 1 study (overall reduction about 5%, Supplementary Table 4; Kristensen et al., 2011) and other papers also have reported that in utero paracetamol did not reduce neonatal male rat AGD (Pereira et al., 2020; Axelstad et al., 2014) or induce permanent alterations in any androgen-dependent tissues later in life in rats (Pereira et al., 2020) or mice (Reel et al., 1992). In addition, few if any permanent effects on androgen dependent endpoints have been seen in long-term studies of the male rat F1 offspring (Dean et al., 2016; Pereira et al., 2020, Figure 2B).

In contrast to the apparent lack of effect of in utero paracetamol of androgen-dependent endpoints in the F₁ male rat, there are several published studies indicating effects on the ovary and F₁ female fertility that suggest that this drug may have the potential to act as a fetal germ cell toxicant (Dean et al. 2016; Holm et al. 2016). There are several MIEs that could account for such effects in F₁ females, but none of these necessarily include alterations of fetal androgen levels.

PPAR agonists: HFPO-DA (GenX), WY 14643, PFOS, and clofibrate

The 4 PPAR agonists all failed to reproduce the fetal testis endocrine profile in a phthalate-like manner (Figure 2D). Examination of the multivariate factors scores based upon fetal T Prod and testis gene expression from the custom arrays (Figure 3) demonstrates that the “positive” phthalates clearly separate in multidimensional space (15 variables) from the “negative” phthalates, the 3 of the PPAR agonists, and the 5 nonphthalate chemicals (PCZ, LIN, 4MI, DEX, and BPC) that reduced T Prod but had minimal effect on testis gene expression on the custom arrays. Zhao et al. (2014) reported that PFOS reduced testis T content and AGD on GD20 but these effects were concurrent with overt maternal toxicity. In that study, the control females gained 75 g during dosing on GD11–20; whereas, reported effects on T content and AGD were only seen in dams that weighed 35% less than control dams after dosing and fetal weights were reduced by 25% versus controls (Table 1 in Zhao et al., 2014).

In addition, the PPAR agonists dramatically induced genes in the PPAR pathway and they significantly increased maternal liver weight (Figure 8). In contrast, while DPeP dramatically reduced T Prod (Figure 4) it did not increase maternal liver weight (Figure 8) and had minimal effects on PPAR mRNA expression as compared with the 3 PPAR agonists. It also is noteworthy that the PPAR agonist WY 14643 had minimal effects on the PPAR pathway in the fetal testis as compared with the fetal liver (Figure 7, columns 5 vs 4) suggesting that this pathway is of minimal biological relevance in the fetal testis during sexual differentiation.

In addition to mRNA for genes involved in sex differentiation and sex determination, every custom array contains mRNA for key genes in the PPAR pathway including Acox1, Cyp4a1, Fabp1, Apoa1, and 3 PPAR receptors. Among the several hundred arrays run on fetal testes with PHTHALATEs that reduce T Prod and affect the mRNA levels described earlier, the expression levels of mRNA for these PPAR pathway genes were never significantly affected.

Taken together, our data do not support the hypothesis that phthalates disrupt reproductive development via the PPAR pathway. Overall, our results demonstrate that the MIE for the in utero effects of phthalate esters does not involve PPAR activation or the PPAR pathway and the key events in the AOP for disrupted fetal testis development and function extend well beyond effects on StAR and TSPO mRNA and T Prod levels, as described in OECD AOP-18 (Nepelska et al., 2015) (https://aopwiki.org/aops/18; accessed February 9, 2020, S2) (Baken et al., 2019).

Phthalates disrupt a common AOP, MIE, and key events: basis for cumulative toxicity

Taken together, these results demonstrate that the phthalates that disrupt male rat sexual differentiation act via a common mode of action and induce a unique cascade of hormonal and mRNA expression alterations in the fetal testis which ultimately result in the Phthalate Syndrome in male offspring (Foster, 2005). Given that active chemicals in this class all act via a common AOP, and very likely a common MIE, is it not surprising that combinations of phthalates administered in utero as binary mixtures (Gray et al., 2001; Howdeshell et al., 2007), a mixture of 5 phthalates (Howdeshell et al., 2008) produce dose-additive cumulative effects on male rat reproductive development with life-long consequences. Based upon the above research, in part, regulatory agencies like the Consumer Product Safety Commission (CPSC, 2017) have conducted cumulative risk assessments on phthalates assuming that they act in a dose-additive manner (CPSC, 2014) In addition, numerous studies have shown that mixtures of phthalates with AR antagonists produce cumulative, dose additive adverse effects on male rat reproductive tract development (Conley et al., 2018; Gray et al., 2006b; Howdeshell et al., 2005, 2015, 2017; Rider et al., 2008, 2009,). Narrowly limiting common mechanism groups in cumulative assessments to toxicants that share identical mechanisms of toxicity is not fully protective of the potential adverse in utero effects of mixtures of toxicants with diverse mechanisms of toxicity (CPSC, 2017; NRC, 2008; Schettler, 2006).

Phthalate-Induced Reproductive Toxicity in Females and Males at Different Life-Stages

Although the MIE(s) for phthalate-induced reproductive toxicity is still unknown after 40 years of investigation, it is informative to consider all the reproductive effects across life stages, genders, and species to determine if a parsimonious hypothesis might provide a biologically plausible explanation of phthalate-induced mammalian reproductive toxicity and facilitate identification of a common MIE for multiple AOPs. This approach seems appropriate because the relative potencies of the phthalates that induce reproductive alterations during different life stages and genders appear to be consistent among many of phthalates. The following section briefly reviews some the effects of phthalates on the reproductive system of male and female rats and other species during several life stages that provide support for this hypothesis.

For example, the rank order of potencies for phthalates that disrupt fetal T Prod (Furr et al., 2014; Hannas et al., 2012) is similar to that seen with effects on the testes of immature male rats (Foster et al., 1980, 1981; Gray et al., 1982; Noriega et al., 2009). Furthermore, the phthalates that have been shown to induce reproductive tract agenesis in female rats (Gray et al., 1999; Hannas et al., 2013), induce mid-pregnancy abortions in rats (Gray et al., 2006a), or induce multinucleated gonocytes (MNGs) in the testis of treated fetuses(Spade et al., 2018) also reduce fetal T Prod. When Spade et al. (2018) exposed rats daily by oral gavage from GD17 to 21 with 1 of 8 phthalate compounds those that reduced T Prod like DBP, BBP, DPeP, and DEHP induced MNGs whereas DMP, DEP, DoTP and the brominated phthalate di-(2-ethylhexyl) tetrabromophthalate did not reduce T Prod or induce MNGs.

Prior to the 1990s, research on the reproductive effects of phthalates focused on effects on the testis and induction of testis atrophy (Foster et al., 1980; Harris et al., 1956; Shaffer et al., 1945; Gray and Butterworth 1980), tubular atrophy and hypospermatogenesis via disruption of Sertoli cell function in vivo and in vitro (Lloyd and Foster, 1988; Richburg and Boekelheide, 1996; Sekaran et al., 2015), effects seen in several mammalian species including a variety of taxonomic groups: murid rodents from the Order Rodentia (suborder Myomorpha-the rat [Rattus norvegicus], mouse [Mus musculus domesticus], and Syrian hamster [Mesocricetus auratus]), the Guinea pig (Cavia porcellus) from Order Rodentia (suborder Hystricomorpha) and a nonrodent species the ferret (Mustela putorius furo) a member of the weasel family (Order Carnivora, family Mustela).

It is noteworthy that in both the fetal and peripubertal (Sjoberg et al., 1986) stages of development the maturing testis with differentiating testicular Leydig cells or Sertoli cells, respectively, DPeP is the most potent PE, DEHP and DBP are active and DEP, DMP, DOP, and DPP are inactive (Foster et al., 1980; Furr et al., 2014; Gangolli, 1982) suggesting that these diverse effects may result from a common MIE disrupting different testicular pathways at different life stages.

These phthalates also induce adverse effects in fetal and adult female rats and mice (Gray et al. 2006a; Hannon and Flaws, 2015). DBP treatment, initiated at weaning, induced mid-pregnancy abortions in female rats and treated females displayed corpora hemorrhagic and reduced ex vivo ovarian hormone production and serum progesterone levels on GD13. Ema et al. reported similar effects in pregnant rats dosed with the DBP metabolite monobutyl phthalate (Ema and Miyawaki, 2001) and -induced pre- and postimplantation embryo/fetal loss (Ema et al., 1993, 1994).

Few studies have examined the effects of in utero phthalate exposure on females. However, studies from our laboratory and a few others have shown that phthalates like DEHP and DBP can induce malformations including vaginal agenesis and uterus unicornis (Gray et al., 1990, 1992), effects that resemble some of the abnormalities in seen in the Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome Type 1, a congenital disorder that affects about 1/5000 women. Girls with MRKH-1 have normal ovaries and fallopian tubes, an absent or incomplete vagina, no cervix and either partial or complete uterine agenesis (Herlin et al., 2020). Similar reproductive tract malformations were seen in phthalate studies in which a mixture of 5 phthalates administered to pregnant rats from GD8 to postnatal day 3 (Howdeshell et al., 2015), or from 8 to GD19 (Hannas et al., 2013). Phthalate-treatment on GD8–19 or GD8–13 induced uterine and vaginal agenesis in 88% and 22%, respectively, whereas GD14–19 had no effect on these tissues. In contrast, in these studies F1 males were demasculinized only when treatment included GD14–19 whereas GD8–13 had no effect.

Taken together, these studies indicate that in utero exposure to phthalates that produce adverse effects in F₁ male rats also induce reproductive abnormalities in F1 females at similar dosage levels. However, the critical in utero period in the female rat for the induction of these malformations includes appearance and cranial to caudal extension of the mesonephric duct (Wolffian duct—WD) and initiation of development of the Mullerian duct (MD) during phases 1 (initiation) and 2 (invagination) of MD development (Goldman and Cooper, 2010; Mullen and Behringer, 2014) which is prior to the sexual differentiation of the tract on GD14–19. The development of the WD is critical for normal development of the MD as MD cranial to caudal extension proceeds along the WD and in the absence of the WD the MD fails to develop (Mullen and Behringer, 2014). This indicates that phthalate-induced alterations of fetal testis T, AMH or Insl3 hormone levels are not MIEs relevant to the induction of these uterine and vaginal malformations.

As discussed earlier, because all the reproductive effects of phthalates on the reproductive system among the genders, life stages and species are the result of exposure to the same phthalates with about the same relative potencies, it seems reasonable to consider the hypothesis that these effects arise from a common MIE among many AOPs in different life-stages and genders.

CONCLUSIONS

Data from this study demonstrate that phthalates that disrupt T Prod also disrupted testis expression of a unique “cluster” of mRNAs for several genes related to sterol transport, testosterone, and insl3 hormone syntheses, and lipoprotein signaling and cholesterol metabolism. phthalates had little or no effect however on mRNA expression for genes in PPAR pathways in the fetal liver whereas the 3 PPAR agonists induced the expression of mRNA for multiple PPAR pathway genes without reducing T Prod. Several nonphthalates tested including linuron, prochloraz, and 4-MI partially reduced T Prod without affecting this mRNA cluster. Dexamethasone and BPC reduced T Prod and reduced some but not all the mRNAs in the cluster. Finally, several chemicals had no effect on testis T Prod including paracetamol, vinclozolin, flutamide, DINCH, dipropylheptyl phthalate, and hexaconazole. It is evident that the hormonal and genomic alterations induced by phthalates that induce the Phthalate Syndrome are not displayed by phthalates that did not reduce T Prod, PPAR agonists or the other tested chemicals.

In summary, phthalate esters that disrupt T Prod act via a novel AOP that includes down regulation of mRNA transcripts for genes involved in fetal endocrine function and cholesterol synthesis and metabolism resulting in a unique postnatal phenotype resulting from reduced T Prod and Insl3 hormone synthesis, identified as the Phthalate Syndrome, a Syndrome not mediated via the PPAR pathway. The fact that some phthalates cause the Phthalate Syndrome via a common AOP provides biological plausibility for their inclusion in a common mechanism group for cumulative risk assessments and establishing relative potency factors for T Prod and gene expression. Additional research also shows that the reductions in fetal testis gene expression and T Prod in utero reported herein can be used to establish relative potency factors to quantitatively predict the doses of individual phthalates and mixtures of phthalates that produce adverse reproductive tract effects in male offspring (Gray et al. 2016).

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.

DECLARATION OF CONFLICTING INTERESTS

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Disclaimer: The research described in this article has been reviewed by U.S. Environmental Protection Agency and approved for publication. Approval does not signify that the contents necessarily reflect the views and policies of the Agency nor does mention of trade names or commercial products constitute endorsement or recommendation for use. This article is the work product of an employee or group of employees of the National Institute of Environmental Health Sciences (NIEHS), National Institutes of Health (NIH); however, the statements, opinions or conclusions contained therein do not necessarily represent the statements, opinions, or conclusions of NIEHS, NIH, or the United States government.

ACKNOWLEDGMENTS

We thank Dr P. Hartig and Dr E.M. Kakaley and Mary Cardon for their technical assistance

FUNDING

National Toxicology Program at the National Institute of Environmental Health Sciences Interagency Agreement with the U.S. Environmental Protection Agency (Cooperative Agreement no. RW-75-92285501-1).

REFERENCES