N-Acetyl-l-cysteine and aminooxyacetic acid differentially modulate trichloroethylene reproductive toxicity via metabolism in Wistar rats

Abstract

Exposure to the industrial solvent trichloroethylene (TCE) has been associated with adverse pregnancy outcomes in humans and decreased fetal weight in rats. TCE kidney toxicity can occur through formation of reactive metabolites via its glutathione (GSH) conjugation metabolic pathway, largely unstudied in the context of pregnancy. To investigate the contribution of the GSH conjugation pathway and oxidative stress to TCE toxicity during pregnancy, we exposed rats orally to 480 mg TCE/kg/day from gestational day (GD) 6 to GD 16 with and without N-acetyl-l-cysteine (NAC) at 200 mg/kg/day or aminooxyacetic acid (AOAA) at 20 mg/kg/day as pre/co-treatments from GD 5–16. NAC is a reactive oxygen species scavenger that modifies the GSH conjugation pathway, and AOAA is an inhibitor of cysteine conjugate β-lyase (CCBL) in the GSH conjugation pathway. TCE decreased fetal weight, and this was prevented by AOAA but not NAC pre/co-treatment to TCE. Although AOAA inhibited CCBL activity in maternal kidney, it did not inhibit CCBL activity in maternal liver and placenta, suggesting that AOAA prevention of TCE-induced decreased fetal weight was due to CCBL activity inhibition in the kidneys but not liver or placenta. Unexpectedly, NAC pre/co-treatment with TCE, relative to TCE treatment alone, altered placental morphology consistent with delayed developmental phenotype. Immunohistochemical staining revealed that the decidua basale, relative to basal and labyrinth zones, expressed the highest abundance of CCBL1, flavin-containing monooxygenase 3, and cleaved caspase-3. Together, the findings show the differential effects of NAC and AOAA on TCE-induced pregnancy outcomes are likely attributable to TCE metabolism modulation.

Introduction

Trichloroethylene (TCE) is a chlorinated environmental contaminant commonly used as a metal degreaser and in the synthesis of various chemicals (Agency for Toxic Substances and Disease Registry 2019). TCE is a suggested toxicant to multiple organs, including liver (Bull, 2000), kidney (Green et al. 1997a), lungs (Forkert et al. 1985; Green et al. 1997b), and the immune (Cooper et al. 2009), nervous (Bale et al. 2011), reproductive (Healy et al. 1982; Manson et al. 1984; Lamb and Hentz 2006; Loch-Caruso et al. 2019), and developmental systems (Saillenfait et al. 1995). Of specific relevance to the present study, epidemiological studies have identified adverse effects of TCE on pregnancy outcomes, including low birth weight and small for gestational age (Rodenbeck et al. 2000; Forand et al. 2012; Ruckart et al. 2014).

The metabolism of TCE is thought to be responsible for the toxicity of TCE. TCE is metabolized through a cytochrome P450 (CYP)-dependent oxidative pathway and a glutathione (GSH) conjugation metabolic pathway (Lash et al. 2014). The CYP-dependent oxidative pathway includes formation of unstable epoxide intermediates that precedes formation of stable metabolites such as dichloroacetic acid and trichloroacetic acid (Lash et al. 2014). In contrast, the GSH conjugation pathway involves formation of S-(1,2-dichlorovinyl)-l-cysteine (DCVC) (Lash et al. 2014). DCVC metabolism by cysteine conjugate β-lyase (CCBL) and flavin-containing monooxygenase 3 (FMO3) yield the reactive metabolites of 1,2-dichlorovinylthiol (DCVT) and DCVC sulfoxide (Lash et al. 2014), respectively. DCVC metabolism by N-acetyl-transferases and CYP3As can form the toxic N-acetyl DCVC sulfoxide (Werner et al. 1996; Lash et al. 2014).

Toxicological studies have contributed to the understanding of how TCE metabolism contributes to its toxicity. Inhibiting CCBL activity in kidney cells via aminooxyacetic acid (AOAA) treatment reduces DCVC-stimulated toxicity (Elfarra and Anders, 1984; Lash et al. 1986, 1994). DCVC stimulates reactive oxygen species (ROS) generation, pro-inflammatory response, and cell death in the HTR-8/SVneo human placental cytotrophoblast cell line (Hassan et al. 2016; Elkin et al. 2018), suggesting that the GSH conjugation pathway has a role in placental toxicity. Additionally, rat placenta contains upstream GSH conjugation enzymes GSH S-transferase and γ-glutamyltransferase (Loch-Caruso et al. 2019), supporting DCVC relevance to placenta.

Because of the requirement of metabolism in TCE toxicity, modulation of TCE metabolism has the potential to alter TCE toxicity. One approach can be through use of N-acetyl-l-cysteine (NAC). NAC can augment cellular production of GSH as a scavenger of ROS and free radicals (Aldini et al. 2018). NAC can also release an acetyl group via aminoacylase I metabolism (Uttamsingh and Anders, 1999; Uttamsingh et al. 2000) to contribute to formation of N-acetyl DCVC and the downstream toxic N-acetyl DCVC sulfoxide metabolite (Werner et al. 1996; Lash et al. 2014). Another approach is use of AOAA, a CCBL inhibitor in kidney (Elfarra and Anders, 1984; Lash et al. 1986, 1994). Although CCBL isoform expression varies among tissues (Lash, 2009) and AOAA inhibition of CCBL in placenta has not been characterized, because AOAA decreases DCVC-stimulated release of interleukin-6 in HTR-8/SVneo cells (Hassan et al. 2016), AOAA may have a mechanism of action relevant to DCVC toxicity in placental cells.

Investigation of mechanisms to explain TCE-induced decreased fetal weight (Loch-Caruso et al. 2019) is a goal of the present study. Apoptosis in placenta or increased pro-inflammatory response in maternal serum are both mechanisms associated with fetal weight decrease. As an example, preeclampsia is associated with decreased birth weight but increased ratio of pro-apoptotic Bax to anti-apoptotic Bcl-2 in placenta (Afroze et al. 2016). Additionally, intrauterine growth retardation is associated with placental syncytiotrophoblast apoptosis (Ishihara et al. 2002). Regarding markers for inflammation, decreased fetal weight is associated with increased maternal serum levels of high sensitive-C-reactive protein, interleukin-6, and tumor necrosis factor-α (Guven et al. 2009).

Because TCE toxicity can be manifest via the GSH conjugation metabolic pathway and involve ROS generation, pro-inflammatory response, and cell death, the current study examined these mechanistic pathways in conjunction with TCE-stimulated reproductive toxicity using NAC and AOAA as potential modifiers of TCE toxicity. Using a timed-pregnant Wistar rat model of fetal growth restriction (Loch-Caruso et al. 2019), the present study tests the hypothesis that TCE-induced fetal weight decrease is due to apoptosis, oxidative stress, and pro-inflammatory responses that require metabolic activation through the TCE GSH conjugation pathway, and thus are modifiable by NAC and AOAA.

Materials and methods

Chemicals and reagents

Trichloroethylene (TCE), potassium phosphate dibasic (K₂HPO₄), potassium phosphate monobasic (KH₂PO₄, anhydrous), sucrose, potassium hydroxide (KOH), boric acid, trichloroacetic acid, and aminooxyacetic acid (as O-(carboxymethyl)hydroxylamine hemihydrochloride) (AOAA) were purchased from Sigma-Aldrich (St. Louis, MO). N-acetyl-l-cysteine (NAC) was obtained from the University of Michigan Hospital pharmacy (pharmaceutical grade). Mini vanilla wafers (Nabisco) were purchased locally. S-(2-Benzothiazolyl)-l-cysteine (BTC) was purchased from Toronto Research Chemicals (Toronto, ON, Canada).

Handling and treatment of rats

Timed-pregnant Wistar rats (Wistar IGS Rat, Strain Code 003) between 60 and 90 days of age were purchased from Charles River Laboratories (Portage, MI). Rats arrived at the University of Michigan School of Public Health Animal Facility on gestational day (GD) 2, with GD 0 designated as day of copulation. Wistar rats were used because prior studies reported decreased fetal weight after maternal exposure to TCE in Wistar rats (Healy et al. 1982; Loch-Caruso et al. 2019), consistent with epidemiology reports that TCE exposure during pregnancy was associated with small for gestational age and low birth weight (Forand et al. 2012; Ruckart et al. 2014). Rats were individually housed in a controlled environment with a 12-h light/dark cycle and provided with standard rat chow (Purina 5001) and water ad libitum. On GD 3, the rat weights ranged from 121 to 224 g. All procedures with the rats were approved by the University of Michigan Institutional Animal Care and Use Committee (IACUC) (Approval #PRO00006981).

Rats were orally exposed to chemical treatments using a mini vanilla wafer treat as vehicle (Seegal et al. 1997). Rats were placed individually in exposure cages without food for one hour prior to presentation with a wafer. Rats were presented with a wafer on GD 3 and GD 4 to train for recognition of the wafers as food. Rats were weighed daily, and treatment doses were adjusted to rat weight. Chemicals were added directly to the wafers and immediately offered to the rats for consumption. NAC and AOAA were dissolved in double-distilled and filtered water prior to pipetting onto the wafer. TCE was pipetted undiluted onto the wafer. The rats typically finished eating the wafer within 10 minutes of wafer presentation.

The treatment regimen is depicted in Fig. 1. Pregnant rats were exposed from GD 6–16 to 480 mg TCE/kg/day, a dose that is within one order of magnitude of the U.S. Occupational Safety and Health Administration Permissible Exposure Level (Agency for Toxic Substances and Disease Registry 2007) and similar to effective TCE dosages used in prior studies with rats (Toraason et al. 1999; Liu et al. 2010; Loch-Caruso et al. 2019). Some rats also received NAC or AOAA as modifiers of the GSH conjugation metabolic pathway, delivered daily on GD 5–16 as pre/co-exposures with TCE. The NAC dosage of 200 mg/kg/day (Fukami et al. 2004; Chang et al. 2005; Naik et al. 2006) and AOAA dosage of 20 mg/kg/day (Perry and Hansen, 1978; Donoso and Banzan, 1984) are within the range of effective dosages used in rats in past reports. The time points of treatment encompass the gestational period when fetuses and placentae begin rapid growth (Furukawa et al. 2011) and are similar to TCE exposure regimens used in prior studies with rats (Healy et al. 1982; Fisher et al. 1989; Loch-Caruso et al. 2019). In total, 49 timed-pregnant Wistar rats were acquired in eight batch deliveries. Within each batch, at least one rat was assigned randomly to each treatment group. Details of the assignments of rats to treatment group and batch are included in Supplemental Table 1. Three of the 49 rats were not pregnant.

Fig. 1.

Treatment schedule of the timed-pregnant Wistar rats. Gestation day (GD) 0 was designated as the day of copulation. Rats arrived on GD 2 were first trained to eat the vanilla wafer on GD 3 and were killed on GD 16. The following dosages of each treatment were used: 200 mg NAC/kg/day, 20 mg AOAA/kg/day, and 480 mg TCE/kg/day. This figure was modified from Su and Loch-Caruso (2020)

Rat tissue collection

Rats were killed on GD 16 by carbon dioxide asphyxiation. Euthanasia was performed directly after completion of vanilla wafer treatment on GD 16, and dams were euthanized in random order. Blood was collected from the rats directly after euthanasia by cardiac puncture. Typically, 6–8 mL of blood was collected from each rat. Blood was allowed to coagulate by incubating at 37 °C in 15-mL conical tubes. Subsequently, the blood was centrifuged at 2000×g for 30 min at 4 °C to separate clotted blood from serum, yielding 3–4 mL of serum per dam. Serum samples were aliquoted into 1.5-mL microcentrifuge tubes and stored at − 80 °C. Maternal kidney, maternal liver, placenta, and fetuses were removed by dissection and weighed individually. Tissues for histology were stored in 1:10 dilution-buffered formalin (Fisherbrand, Waltham, MA; product number 245–684) for a minimum of 24 h. Placenta, maternal kidney, and maternal liver tissues were snap frozen in liquid nitrogen and stored at − 80 °C, with the exception of tissue for RNA analysis, which was stored overnight in RNAlater (Qiagen, Germantown, MD) prior to RNAlater removal and sample transfer to − 80 °C. A subset of placentae were sampled for some endpoint analyses. Sampling strategy in those cases, within a given fetal sex, consisted of taking placentae from distinct litters as opposed to placentae from the same litter. To the extent possible, whereas placentae used for histology came from the center of the uterine horn, placentae used for all other assays came from the left uterine horn.

Measurement of cysteine conjugate β-lyase (CCBL) activity

CCBL activity was measured according to a method developed by Dohn and Anders using S-(2-benzothiazolyl)-l-cysteine (BTC) as the substrate and 2-mercaptobenzothiazole as the product produced (Dohn and Anders 1982). Briefly, placenta, maternal kidney, or maternal liver were thawed on ice, and then added at a 25% (w/v) concentration to a solution of 0.1 M potassium phosphate (pH 7.4) containing 0.25 M sucrose. Tissues were sonicated twice at 3.5 Hz for five seconds (MiSonix Sonicator 3000, Misonix Inc., Farmingdale, NY) to dissociate the tissue for enzyme analysis. The potassium phosphate buffer (0.1 M, pH 7.4) was made from a mixture of potassium phosphate dibasic (K₂HPO₄) at 69.6 mM and potassium phosphate monobasic (KH₂PO₄, anhydrous) at 30.4 mM prior to sucrose addition.

The substrate solution was made by suspending 10.15 mg of BTC in 4 mL ddH₂O representing 40% of total final desired volume. The mass of BTC added corresponds to a final BTC concentration in substrate solution of 4 mM. Potassium hydroxide (KOH) (1 M in ddH₂O) was added at 0.25 mL to dissolve BTC. After the BTC dissolved, 5 mL of boric acid (0.2 M in ddH₂O) was added. The solution pH was adjusted to 8.6 using 1 M KOH, and ddH₂O was then added to achieve 10 mL desired total volume. In aliquots, this substrate solution was added to 0.1 M potassium borate buffer (made from KOH and boric acid) in a 1.5:1.3 (v/v) ratio and incubated for 4 min at 37 °C. Subsequently, enzyme sample was added to this reaction mixture in a 1:14 (v/v) ratio. Incubation time with the enzyme sample was 30 min at 37 °C.

After the incubation, trichloroacetic acid at a 10% (w/v) concentration in ddH₂O was added to the reaction mixture in a 1:5 (v/v) ratio to terminate the reaction and precipitate protein. A negative control for each sample was made by adding 10% (w/v) trichloroacetic acid into the reaction mix before addition of the enzyme source. All reactions were centrifuged at 11,000×g for 1.25 min to remove precipitated protein. Absorbance of each supernatant was measured at 321 nm wavelength, with a molar extinction coefficient (ε) of 21,600 M⁻¹ cm⁻¹ for 321 nm, using a SpectraMax M2e microplate reader (Molecular Devices, San Jose, CA). The absorbance corresponding to the negative control was subtracted from the absorbance corresponding to the appropriate reaction to generate a net absorbance used to convert absorbance into nmol of product produced/min/mL, which was converted into nmol/min/mg tissue, nmol/min/mg protein, or nmol/min/total tissue. The bicinchoninic acid (BCA) assay commercial kit (ThermoFisher Scientific, Waltham, MA) was used to determine protein concentration.

Hematoxylin and eosin (H&E) staining

Placentae stored in 1:10 dilution-buffered formalin (Fisherbrand, Waltham, MA) were cut in half transversely to expose the maternal and fetal layers. Each half was placed in a biopsy cassette (Fisherbrand, Waltham, MA) submerged in the 1:10 dilution formalin and transferred to the University of Michigan In Vivo Animal Core. Placentae were embedded in paraffin using an automated tissue processor (TissueTek, Sakura) and typical histological methods at the core. Sections were cut to 4 μm thickness using a rotary microtome and mounted on glass slides. All placental sections mounted on slides were female in sex because not enough male placentae were collected in formalin at the time of dissection (when fetal sex was yet to be determined). Tissues were stained with hematoxylin and eosin using a standard hematoxylin and eosin staining protocol on an automated histostainer (Leica Autostainer XL, Leica Biosystems Inc., Buffalo Grove, IL). The slides were digitally scanned using an Aperio AT2® scanner (Leica Biosystems Inc.), allowing imaging of two vertical transverse sections of the whole placenta. Quantitative morphometric measurements with drawn annotations of placental zones were taken from these images using Aperio ImageScope software (Leica Biosystems Inc.). The placental zones quantified included the labyrinth zone area, basal layer area, total placenta area, labyrinth zone width, basal zone width, and decidua basale width. Decidua basale area was calculated as total placenta area minus labyrinth zone area minus basal layer area. Total placenta width was calculated as labyrinth zone width + basal area width + decidua basale width. Three different measures of width for each of the zones of interest were made at the center region of each zone. Because two sections were taken from each placenta (one from each half) for a slide, a slide for one given placenta had two different area measurements for each of the area metrics and six different width measurements for each of the width metrics. The morphometric measurements were averaged for each placenta prior to further analysis. In the case of the labyrinth zone, basal zone, and decidua basale, absolute dimensions (length and area) and relative dimensions (normalized to the total placenta and reported as percentages) were reported.

Immunohistochemistry (IHC) staining of cleaved caspase-3, CCBL1, FMO3, 3-nitrotyrosine, and CYP3A1

Placental tissue sections were subjected to immunohistochemistry (IHC) staining to detect protein expression of cysteine conjugate β-lyase (CCBL1), flavin-containing monooxygenase 3 (FMO3), and CYP3A1. These enzymes were selected for IHC analysis because of their importance for TCE metabolic activation (Lash et al. 2014). In addition, 3-nitrotyosine and cleaved caspase-3 were assessed by IHC as markers of protein oxidative stress and apoptosis, respectively (Ahsan 2013; Crowley and Waterhouse 2016).

Unstained sections were deparaffinized and rehydrated, then subject to heat-induced antigen retrieval with a commercial antigen retrieval solution (DIVA, Biocare Medical, Pacheco, CA) within a pressurized retrieval chamber (Decloaker, Biocare Medical). Staining was performed using an automated immunohistochemical stainer (Biocare Intellipath, Biocare Medical) with inclusion of endogenous peroxidase and non-specific blocking steps. Negative control slides were run concurrently using naïve mouse/rabbit serum (Polymer Universal negative control serum, Biocare Medical) in place of the primary antibody. Positive detection was performed with commercial dextran polymer-based, biotin-free reagents for mouse (MACH4 HRP, Biocare Medical) or rabbit (Biocare Rabbit on Rodent Poly HRP, Biocare Medical) with the chromogen diaminobenzidine. Hematoxylin was utilized as the nuclear counterstain. Slides were dehydrated and mounted with glass cover slips. Details of the primary antibodies, dilutions, positive controls, and negative control images are provided in Supplemental Table 2. Staining was performed at the University of Michigan In Vivo Animal Core with the exception of CYP3A1 staining, which was performed at Histoserv, Inc. (Germantown, MD).

After IHC staining, tissue sections were digitally scanned using an Aperio AT2® scanner (Leica Biosystems Inc.) or Histoserv Digital in the case of CYP3A1 (Germantown, MD), allowing imaging of two vertical whole sections of the placenta. For all IHC stains, stain quantification was performed using the open source image analysis software platform QuPath (Bankhead et al. 2017) version 0.2.0, with the exception of cleaved caspase-3, for which version 0.1.2 was used. Each zone of the placenta was traced to allow for quantification specific to each region. From QuPath, we identified a setting to detect all staining, regardless of localization to cytoplasm or nucleus, and a setting to detect only nuclear staining. All detections were based on achieving a specific color threshold. The nucleus-only stain detection values were subtracted from the total stain detection values to derive quantification of cytoplasmic stain. For a given placenta, total positive count was normalized to total area for each placental zone. These normalized values were used in statistical analyses. In the case of CYP3A1, which had a darker background compared to other stains, QuPath quantification used two different thresholds: a normal one used for all other stains and a higher threshold to account for the darker background to ensure that the background did not make a difference.

Detection of fragmented DNA via labeling with terminal deoxynucleotidyl transferase (Tdt)

Unstained placental tissues were embedded in paraffin and mounted on glass slides as described for the H&E staining. Sections were stained for expression of the Tdt enzyme to detect fragmented DNA. Slides were subject to labeling with the Tdt enzyme using the FragEL^™ DNA Fragmentation Detection Kit, Fluorescent—TdT Enzyme kit (Millipore Sigma, Burlington, MA) per the manufacturer protocols. Positive and negative control slides with HL-60 cells (provided by manufacturer) were stained according the manufacturer’s protocol. The fragmented DNA was detected as a fluorescent green stain. Nuclei were co-stained with 4′,6-diamidino-2-phenylindole (DAPI) in the color blue and used to normalize green staining to total nuclear stain.

An EVOS M7000 Microscope (ThermoFisher Scientific, Waltham, MA) with fluorescence and image stitching capabilities was used to capture images of the placenta stained with Tdt. Images of whole placenta were captured at 100× (10 × 10 ×) magnification under the DAPI, GFP, and DAPI + GFP overlay channels. QuPath v0.2.0 was used to quantify the staining on the DAPI + GFP overlay images, with quantification performed specific to the labyrinth zone, basal zone, and decidua basale of the placenta as manually annotated using QuPath v0.2.0. A ratio of green to blue stain was obtained for each of the two sections for the placenta for each zone, and the average of these was used for statistical analysis for each zone of a given placenta.

Alizarin red staining to detect placental calcification

Paraffin-embedded unstained slides of placental tissue were prepared as described for H&E staining. Sections were subject to alizarin red staining for mineral detection, using a commercial reagent kit according to the manufacturer’s recommendations (Alizarin Red C-206, Rowley Biochemical, Inc, Danvers, MA). As with H&E and IHC, slides were digitally scanned using an Aperio AT2® scanner (Leica Biosystems Inc.), allowing imaging of two vertical whole sections of the placenta. The labyrinth zone, basal zone, and decidua basale were annotated using Aperio ImageScope software (Leica Biosystems Inc.). Digital quantitation of staining for each zone was performed using a commercial algorithm (Leica Aperio Color Deconvolution, v.9.1, Leica Biosystems Inc.). Settings were calibrated for detection of alizarin red slide using a test area on the slide per manufacturer instructions. Calibrated parameters were set as R = 0.223, G = 0.579, B = 0.785 and default parameters for intensity were used (range 0–255; weak 220, moderate 175, strong 100). For each zone of the placenta, the average from the two placental sections for a given placenta was used for statistical analysis. The percentage of strong positive and total positive stains, as defined in the above calibrated parameters with total including all positive detections, were used for statistical analysis.

RNA extraction

Placental samples were thawed and homogenized using a FastPrep-24 tissue and cell lyser (MP Biomedicals, Solon, OH) with the placenta submerged in RLT Buffer PLUS (Qiagen, Germantown, MD) containing 1% (v/v) 2-mercaptoethanol (Sigma-Aldrich, St. Louis, MO). Following homogenization, RNA was extracted using an RNeasy Plus Mini kit (Qiagen, Germantown, MD) according to the manufacturer’s instructions. RNA concentration and purity were verified using a NanoDrop 2000 UV–Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). RNA was stored at − 80 °C until further analysis.

Placental sex determination

The sex of the rat placenta and associated fetus was primarily determined by the presence or absence of the Sry genomic DNA (gDNA). The gDNA was isolated from rat placenta using a NucleoSpin^™ Tissue kit (Machery-Nagel, Bethlehem, PA) according to the manufacturer’s instructions. The concentration and purity of the gDNA were verified using a Nanodrop 2000 UV–Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA) and stored at − 20 °C until further analysis using qRT-PCR. A subset of these placentae were also evaluated for Sry messenger RNA (mRNA) expression as described previously (Su and Loch-Caruso 2020). In the case of one discrepancy between the mRNA and gDNA Sry reading (male classification from gDNA but female classification from mRNA), sex was assigned based on the gDNA result. Because the highly concordant nature of the mRNA and gDNA classifications (Su and Loch-Caruso 2020), a few placentae were evaluated for Sry mRNA but not gDNA and included in subsequent analyses that involved consideration of sex.

Quantitative real-time polymerase chain reaction (qRT-PCR)

Detection of Sry gDNA and mRNA expression of Nfkb1, Bcl2, Bax, Prdx1, Lgals3, Kyat1, and Fmo3 was performed with qRT-PCR. Briefly, cDNA was synthesized using a Bio-Rad iScript^™ cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA) according to the manufacturer’s instructions. A Bio-Rad CFX Connect^™ Real-Time System (Bio-Rad Laboratories, Hercules, CA) synthesized cDNA under the following protocol: (1) 25 °C for 5 min, (2) 42 °C for 30 min, (3) 85 °C for 5 min, and then (4) cool down to 4 °C. The cDNA samples were stored at − 20 °C until further use.

Reaction mixtures of 25 μL total were prepared in SsoAdvanced^™ Universal SYBR® Green Supermix (60% v/v; Bio-Rad Laboratories, Hercules, CA) containing 0.32 μM of each (forward and reverse) primer, and 40 ng of either cDNA template or 50 ng of gDNA. Primer sequences and sources are listed in Supplemental Table 3.

Samples were run and analyzed in Hard-Shell® 96-well plates (Bio-Rad Laboratories, Hercules, CA) using a Bio-Rad CFX Connect^™ Real-Time System using the following protocol: (1) 95 °C for 10 min, (2) 95 °C for 15 s, (3) 60 °C for 1 min, (4) repeat 39 times steps 2 and 3, (5) 95 °C for 1 min, (6) 65 °C for 2 min, (7) 65 °C to gradual increase to 95 °C, stopping at every 0.5 °C interval for 5 s each. Analysis was performed using the ΔΔCt method (Yuan et al. 2006). All samples were run and analyzed in duplicate. B2m served as the reference gene.

Detection of IL-1β, TNF-α, and IL-6 in maternal serum

Maternal serum samples stored at − 80 °C were thawed and pipetted into round-bottom 96-well plates. Samples were then delivered to the University of Michigan Immunology Core for IL-1β, TNF-α, and IL-6 for enzyme-linked immunosorbent assay (ELISA) using Duoset ELISA kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s recommended protocol.

Statistical analysis

Statistical analyses included mixed model analysis of variance (ANOVA), one-way ANOVA, or two-way ANOVA, each followed by Tukey’s post-hoc comparison of means. In the case where only two groups were compared, an unpaired two-tailed t-test was used. Mixed model ANOVA was performed using SPSS software (Chicago, IL), whereas one-way ANOVA, two-way ANOVA, and t tests were performed using GraphPad Prism (San Diego, CA). Graphs were generated using GraphPad Prism (San Diego, CA). A p < 0.05 was considered statistically significant. Percentage data were converted to proportion values and arcsine transformed prior to statistical analysis. The ratio data from the Tdt staining analysis were also arcsine transformed prior to statistical analysis.

Results

Fetal weight

Fetal weight was analyzed for all fetuses irrespective of sex, as well as separately by sex in male and female sub-sets as determined primarily by placental gDNA. In the analysis of all fetuses, exposure to TCE during pregnancy decreased average fetal weight per litter on GD 16 by 12.8% (p < 0.0001) (Fig. 2a1). TCE suppressed average fetal weight in males by 21.7% relative to control (p < 0.0001) but no significant effect was observed in female fetuses (Fig. 2a2 and a3, respectively). Whereas NAC alone had no significant effect on average fetal weight relative to control, AOAA alone decreased average fetal weight relative to control by 15.1% (p < 0.0001) in all fetuses (Fig. 2a1), and the same pattern was observed in separate analyses of male and female fetuses (p = 0.018 and 0.001, respectively) (Fig. 2a2 and a3, respectively). The combined treatment of AOAA with TCE prevented the depressed average fetal weight elicited by treatment with TCE alone and AOAA alone in all fetuses and in the male subset (Fig. 2a1 and a2, respectively). In females, the combined AOAA and TCE treatment prevented depressed average fetal weight elicited by treatment with AOAA alone (Fig. 2a3). The combined treatment of NAC with TCE decreased average fetal weight relative to control in all fetuses, male fetuses, and female fetuses by 13.3%, 16.7%, and 15.6%, respectively (p < 0.0001, p = 0.003, and p = 0.002, respectively) (Fig. 2a1, a2, and a3, respectively) but did not differ from TCE alone.

Fig. 2.

Effects of trichloroethylene (TCE), in the presence and absence of N-acetyl-L-cysteine (NAC) and aminooxyacetic acid (AOAA), on a fetal weight, b placental weight, and c placental efficiency. The effects for all fetuses/placentae, a male subset and female subset are presented. We were able to determine the sex for only a subset of placentae. Placental efficiency is defined as the ratio of fetal to placental weight. Individual data points correspond to average fetal weight, average placental weight, or average placental efficiency for a given dam. Data were analyzed by mixed model ANOVA with treatment as the fixed variable and litter as the random variable. Statistically significant differences are indicated by non-overlapping letters (p ≤ 0.018). Error bars represent mean ± SEM. For all columns (1–3), N = 8 litters for each treatment group with the exception of the AOAA-only and TCE-only treatment groups, which had N = 7 each

Placental weight and efficiency

Investigation of placenta weight and placental efficiency revealed treatment-related effects. Placental efficiency is the ratio of fetal weight to placenta weight and is used as an index of placental function (Wilson and Ford, 2001). Although TCE alone had no significant effect on average placental weight (Fig. 2b1, b2 and b3), the combination treatment of TCE with AOAA increased average placental weight compared with the control group by 8.09% (p = 0.008) and compared with the TCE-only treatment group by 15.2% (p = 0.001) (Fig. 2b1). Moreover, the combination treatment of TCE + AOAA increased average placenta weight relative to TCE-only treatment by 20.7% in males (p = 0.003) (Fig. 2b2), whereas there was no statistically significant effect on female placentae (Fig. 2b3). There were no remarkable treatment-related effects on average placental efficiency (Fig. 2c1, c2 and c3), with the sole statistically significant difference being between AOAA treatment alone and NAC treatment alone in all placenta (Fig. 2c1).

Maternal body weight, tissue weights, and litter size

Treatments with TCE, NAC, or AOAA—alone or in combination—had no significant effects on maternal kidney weight, maternal liver weight, litter size, maternal weight gain (GD3 to GD16), or fetal sex ratio (Supplemental Fig. 1), indicating lack of overt treatment-related maternal toxicity.

CCBL activity in maternal tissue and placenta

AOAA effectively inhibited CCBL activity in maternal kidney by at least 73.0%, whether administered alone or as pre/co-treatment with TCE, compared with controls or compared with the TCE-only treatment group when normalized to specific activity by mg tissue (p ≤ 0.0239) (Fig. 3a1). A similar pattern was observed when data were normalized to mg protein or expressed as total organ activity (Fig. 3a2 and a3). In contrast, AOAA had no significant effect on CCBL activity in maternal liver or placenta regardless of normalization procedure (Fig. 3b–d, respectively). Neither NAC nor TCE, alone or in combination with each other, had a significant effect on CCBL specific or total activity in maternal kidney, maternal liver, or placenta (Fig. 3a–d, respectively).

Fig. 3.

Effects of AOAA, TCE, and NAC on CCBL activity in a maternal kidney, b maternal liver, c male placenta, and d female placenta. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison of means. Statistically significant differences are indicated by non-overlapping letters (p ≤ 0.0297).. Error bars represent mean ± SEM. For a–d, N = 4 specimens for each experimental group, with each N coming from distinct dams

Placental zone areas and widths

Because pregnancy progression is associated with changes in size to each zone of the rat placenta (Furukawa et al. 2011), we assessed treatment effects on placental morphology. A representative image for a placental section illustrates our method for quantification of area and width in hematoxylin and eosin-stained placental sections (Fig. 4). TCE treatment alone did not alter absolute or relative width and area dimensions in any of the zones or in the absolute width or area for total placenta (Fig. 5). Unexpectedly, NAC pre/co-treatment with TCE decreased relative area and width of the labyrinth zone by 16.9% and 17.1%, respectively (p = 0.0483 and 0.0250, respectively) (Fig. 5a1 and a3, respectively) and increased relative area and width of the basal zone by 38.1% and 47.2%, respectively (p = 0.0167 and 0.0395, respectively) (Fig. 5b1 and b3, respectively) relative to TCE treatment alone. No other changes in absolute or relative dimension were observed with treatments (Fig. 5).

Fig. 4.

Representative image of hematoxylin and eosin (H&E) stained placental section indicating quantification method for placenta morphometric analysis of different placental zones

Fig. 5.

Effects of TCE, in the presence and absence of NAC and AOAA, on placental zone morphology. Effects on a labyrinth zone dimensions, b basal zone dimensions, c decidua basale dimensions, and d total placenta dimensions are presented. Relative values indicate normalization to the total placenta value. Data were analyzed by one-way ANOVA followed by Tukey’s post-hoc comparison of means. Data of relative dimensions were converted from percent to fraction, then arcsine transformed prior to statistical analysis. N = 6 placentae for each experimental group, with each placenta from a different litter. All placenta used for this figure are female

Markers of oxidative stress, apoptosis, and pro-inflammatory response

No treatment-related changes were detected with IHC staining of placenta for 3-nitrotyrosine and cleaved caspase-3, markers of oxidative stress and apoptosis, respectively (Supplemental Table 4). Staining intensity varied but was detectable for all IHC stains for any given placental zone (representative images shown in Fig. 6). No treatment-related changes were observed for Tdt enzyme staining in placenta (quantification method shown in Supplemental Fig. 2). Likewise, no treatment-related changes were observed for mRNA expression of Prdx1, Nfkb1, Bcl2, Bax, and Lgals3, genes relevant to oxidative stress, inflammation, and apoptosis (Supplemental Table 5). Moreover, we failed to detect treatment-dependent changes of the pro-inflammatory cytokine IL-1β in maternal serum (Supplemental Fig. 3). We also attempted to detect TNF-α and IL-6 in maternal serum, but stopped at N = 2 (12 total samples) because one-third (4/12) of the samples for each endpoint had zero values (data not shown).

Fig. 6.

Representative images for immunohistochemistry (IHC) staining of control placentae for a 3-nitrotyrosine, b cleaved caspase-3, c cysteine conjugate β-lyase 1, d flavin-containing monooxygenase 3, and e CYP3A1 stains. Quantification method for the f total detection reading and the g nuclear detection reading. In f and g, the image on the left an unannotated image, whereas the image on the right is the equivalent image containing annotations from the detections. Red markings symbolize a positive detection, whereas blue markings symbolize a negative detection. Red arrows in a–e denote regions with considerable positive stain. The images from f and g are all from the labyrinth zone of placenta stained with 3-nitrotyrosine and have scale bars representing 100 μm

In placenta of control rats (analysis only reported for this group to understand zonal differences in normal conditions), differences in 3-nitrotyrosine and cleaved caspase-3 staining by placental zone were considerable (Fig. 7). Expression of 3-nitrotyrosine was lower in the basal zone compared with the labyrinth zone and dedicua basale for nuclei detection by 42.8% and 43.9%, respectively (p = 0.0123 and 0.0093, respectively), but similar differences were not observed for the total or cytoplasmic metrics (Fig. 7a). For the nuclear metric, cleaved caspase-3 expression was greater in the decidua basale compared to the labyrinth and basal zones by 1113% and 1003%, respectively (p = 0.0110 and 0.0117, respectively) (Fig. 7b).

Fig. 7.

Effects of IHC staining of a 3-nitrotyrosine, b cleaved caspase-3, c cysteine conjugate β-lyase 1, d flavin-containing monooxygenase 3, and e and f CYP3A1 as a function of placental zone in control placenta. Data were analyzed by one-way ANOVA followed by Tukey’s post-hoc comparison of means. For each zone, N = 6 female placentae, each from distinct litters

Placental enzymes involved in TCE metabolism

Treatment-related changes were not observed by IHC detection of CCBL1, FMO3, and CYP3A1, enzymes that play important roles in TCE metabolism (Supplemental Table 4). Likewise, no treatment-related changes were observed for mRNA expression of Kyat1 and Fmo3, genes relevant to TCE metabolism (Supplemental Table 5). However, significant differences in staining occurred among the placental zones within the control group (Fig. 7). Pronounced differences occurred for CCBL1, with increased staining in the decidua basale compared with the labyrinth zone for total, nuclear, and cytoplasmic detections (increased 87.0%, 303%, and 56.1%, respectively; p = 0.0072, 0.0005, and 0.0422, respectively) (Fig. 7c1, c2, and c3, respectively). Some differences in staining between the different zones is specific to particular metrics (e.g., nuclear). Detection of FMO3 was less in the labyrinth zone compared to the basal zone and decidua basale for the total and cytoplasmic metrics (for total metric, decreased 99.9% and 99.9%, respectively; p = 0.0028 and 0.0027, respectively) (for cytoplasmic metric, decreased 99.9% and 99.9%, respectively; p = 0.0016 and 0.0016, respectively) (Fig. 7d). Although the yolk sac expression of FMO3 was not quantified because the yolk sac was not present in all placenta sections, qualitative assessment of appearance suggests that the yolk sac expresses FMO3 at a very high level (Fig. 6). For the normal threshold and nuclear metric, CYP3A1 staining was lower in the basal zone compared to the labyrinth zone and decidua basale by 47.9% and 51.5%, respectively (p < 0.0001 and p < 0.0001, respectively) (Fig. 7e). A similar pattern was observed for the normal threshold for the total metric but not the cytoplasmic metric (Fig. 7e).

Alizarin red staining

Intensity of alizarin red staining for calcium deposition did not differ by treatment group (Supplemental Table 4). Representative images with and without quantification marking are displayed in Fig. 8a and b. Nonetheless, alizarin red staining intensity differed by placental zone within control rats (analysis only reported for this group to understand zonal differences under normal conditions). Specifically, the percentage of strong positive stain was 52.9% higher in the decidua basale zone than the labyrinth zone of control rats (p = 0.0274) (Fig. 8c), but no differences were observed for total positive stain among the placental zones (Fig. 8d).

Fig. 8.

Alizarin red staining of control placentae. a Representative image. b Quantification method for analysis of alizarin red staining showing alizarin red stain in the basal zone without (left) and with (right) marking from the total positive stain algorithm. The green lines are derived from tracing the basal zone and decidua basale to mark their borders. c Strong positive stain and d total positive stain of alizarin red stain intensity as a function of placental zone. Percent strong positive stain or total positive stain were not altered by treatment group for any zone of the placenta (Supplemental Table 4). Percentage data were converted to fraction values and arcsine transformed prior to statistical analysis by one-way ANOVA followed by Tukey’s post-hoc comparison of means. For each zone, N = 6 female placentae, each from distinct litters. The images from b come from the same placenta as (a)

Discussion

This study confirms our prior finding that TCE exposure during mid-pregnancy decreases fetal weight, using the same exposure regimen (Loch-Caruso et al. 2019). New to the present study, we report that AOAA pre/co-treatment prevented the TCE-induced fetal weight decrease. AOAA is an inhibitor of CCBL (Elfarra and Anders 1984; Lash et al. 1986, 1994), which metabolizes the TCE glutathione pathway metabolite DCVC into DCVT, a toxic and reactive metabolite important for kidney toxicity (Lash et al. 2014). We suggest that the effect on fetal weight may be due to AOAA inhibition of CCBL in maternal kidney because we observed that AOAA strongly inhibited CCBL activity in maternal kidney but not in placenta or maternal liver. Moreover, although NAC pre/co-treatment did not significantly affect fetal weight relative to TCE treatment alone, the study suggested that NAC pre/co-exposure with TCE could have the potential to compromise placental function because NAC pre/co-treatment altered in placental zone dimensions compared with TCE treatment alone.

Unexpectedly, AOAA treatment alone decreased fetal weight. We suggest that AOAA inhibition of CCBL may have disrupted essential physiological processes for fetal growth and development. Examples of potentially vulnerable CCBLs include alanine aminotransferases and mitochondrial aspartate aminotransferases (Cooper and Pinto, 2006). Alanine aminotransferase above 50 U/L are used in clinical diagnosis of hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome, which includes preeclampsia (Vigil-De Gracia et al. 2003). Similarly, mitochondrial aspartate aminotransferase is part of the malate-aspartate shuttle that is crucial to the metabolism of lactate in the embryo (Lane and Gardner 2005). Together, we suggest that inhibition of aminotransferase-derived CCBL activity critical for fetal development could at least partially explain a detrimental effect of AOAA treatment alone on fetal weight.

Although AOAA treatment alone depressed fetal weight, the mean fetal weight of the TCE + AOAA group increased relative to the AOAA alone group. While speculative, we suggest that the AOAA inhibition of CCBL in maternal kidney in the TCE + AOAA group might have allowed DCVC build-up that subsequently depressed activity of upstream enzymes such as γ-glutamyltransferase and cysteinyl-glycine dipeptidase. Inhibition of γ-glutamyltransferase reduces circulating levels of cysteine (Hanigan 2014). Potential benefits of reducing systemic cysteine extends to pregnancy because total plasma cysteine is positively associated with preeclampsia, very low birth weight, and premature delivery (El-Khairy et al. 2003). Likewise, because the reaction catalyzed by cysteinyl-glycine dipeptidase forms cysteine and cysteine conjugates as products (Knapen et al. 1999; Hanigan 2014), inhibition of cysteinyl-glycine dipeptidase metabolism could be beneficial for pregnancy by reducing systemic cysteine.

Our findings provide new insight into AOAA inhibition of CCBL activity, showing that AOAA had no significant effect on placental or maternal liver CCBL activity. Additionally, we report that AOAA strongly inhibited CCBL activity in maternal kidney, consistent with previous studies showing that AOAA is an effective CCBL inhibitor in kidney cells of nonpregnant animals and humans (Elfarra and Anders 1984; Lash et al. 1986, 1994). It is known that kidney and liver express different forms of CCBL (Lash 2009), but the exact forms of CCBL in placenta are presently unknown. Our data suggest that the form(s) of CCBL in the placenta are more similar to those of liver as opposed to kidney based on responses to AOAA.

Mechanistic and physiological explanations exist for how AOAA could inhibit CCBL activity in kidney to prevent TCE-induced decreased fetal weight. DCVG and DCVC are TCE metabolites generated through the GSH conjugation metabolic pathway that circulate systemically (Lash et al. 2014). Thus, it is plausible that AOAA pre/co-treatment with TCE could thwart the ability of the kidneys to convert DCVC into reactive and toxic downstream metabolites (Lash et al. 2014). Although untested, this mechanism to reduce toxicity may also be manifest in fetal kidney. The existence of peptide transporters in placenta (Leazer and Klaassen 2003) that could transport DCVC from mother to fetus supports the relevance and plausibility that AOAA pre/co-treatment could exert an effect on fetal weight via fetal kidney.

Because the magnitude of decreased fetal weight relative to control was similar for TCE treatment alone and AOAA treatment alone, an alternative possibility exists that TCE treatment prevented AOAA-stimulated fetal weight decrease. However, literature supporting a mechanism for TCE to counter an AOAA-stimulated effect, possibly through impact on alanine aminotransferase and mitochondrial aspartate aminotransferase, is not as clear to date. In light of our results, investigation into such phenomenon would be critical and would be best examined using a modified exposure regimen in which TCE is administered as a pre/co-treatment to AOAA treatment in experimental animals.

Because the labyrinth zone of the placenta increases as pregnancy progresses in rat and nearly the opposite is true of the basal zone, especially around GD 16 (Furukawa et al. 2011), the changes observed in the NAC + TCE treatment group are consistent with a delayed developmental phenotype compared to the TCE treatment alone group. These findings are important because changes in relative areas and widths of multiple placental zones were discovered. Follow-up investigation found a lack of treatment effects on alizarin red staining of the rat placentae, indicating that enlargement of the basal zone for the NAC pre/co-treatment group was not attributable to placental calcification. Rather, alteration of placental morphology may be related to other mechanisms, such as disruption of growth factors and cytokines critical for placental growth (Furukawa et al. 2011).

The suggestion that NAC pre/co-treatment increased TCE toxicity to placenta is consistent with our findings in a BeWo human trophoblast model in which NAC exacerbated some responses stimulated by DCVC (Su et al. in preparation). In BeWo cell studies, NAC enhancement of DCVC toxicity was associated with an NAC effect of increased CYP3A4 mRNA expression (Su et al. in preparation). A possible mechanism could be that NAC provides an acetyl group via its breakdown by aminoacylase I (Uttamsingh and Anders 1999; Uttamsingh et al. 2000) to generate N-acetyl DCVC from DCVC (Lash et al. 2014) and further promote formation of the toxic N-acetyl DCVC sulfoxide via CYP3As (Werner et al. 1996; Lash et al. 2014). However, treatments did not affect IHC staining of CYP3A1, a prominent CYP3A isoform in rat placenta (Ejiri et al. 2001, 2003). It may be that effects are more subtle than can be detected by IHC, and further studies utilizing placental explants may detect more subtle differences. Another potential explanation for further investigation can be that other CYP3A isoforms could be more relevant in the context of TCE metabolism in rats.

The only statistically significant treatment effects in CCBL activity compared with controls were for maternal kidney from rats treated with AOAA alone and rats treated with AOAA + TCE. Although the mean values for CCBL activity were higher compared with controls in some tissues from rats receiving other treatments, these differences in CCBL activity were not statistically significant. Even so, the trend observed for maternal kidney CCBL activity for NAC treatment alone and NAC + TCE treatment relative to control is not the same as the respective trends for fetal weight, in contrast to the situation with AOAA treatment alone and AOAA + TCE treatment. To explain these observations, we suggest that the effects of NAC and AOAA could manifest through different organs because AOAA inhibited maternal kidney but not placental CCBL activity whereas NAC + TCE treatment had an effect on placental morphology relative to TCE treatment alone, a phenomenon our data would suggest to be independent of fetal weight, which contrasts from the case with AOAA-inhibited maternal kidney CCBL activity. Additional experiments that include a larger sample size, inclusion of all groups in H&E and IHC analyses, and improved methodology to isolate specific placental zones prior to assays could clarify the explanation.

Although DCVC stimulates ROS generation, pro-inflammatory cytokine release, and apoptosis in HTR-8/SVneo human placental cells in vitro (Hassan et al. 2016; Elkin et al. 2018), exposure of pregnant rats to TCE failed to stimulate changes in markers of oxidative stress (3-nitrotyrosine) and apoptosis (cleaved caspase-3 and Tdt enzyme staining) in placental tissue. Additionally, there were no significant changes in placental mRNA expression of genes relevant to oxidative stress and apoptosis pathways (Prdx1, Bcl2, Bax, Nfkb1 or Lgals3). These findings suggest that TCE, NAC, and AOAA may have acted by mechanisms involving tissues other than placenta in vivo, or else their actions on the placenta may have involved mechanisms other than oxidative stress and apoptosis. Alternate explanations include differences in species and concentrations used in the studies. Rats exhibit higher GGT and renal CCBL activity compared to humans (Hinchman and Ballatori 1990; Lash et al. 1990) as examples of species differences that can affect detected concentrations of TCE metabolites. Additionally, systemic or urinary concentrations of TCE CYP-oxidation metabolites trichloroacetic acid and dichloroacetic acid (Lash et al. 2014) were unknown in the present study. However, trichloroacetic acid and dichloroacetic acid exhibited nominal toxicity in HTR-8/SVneo and BeWo placental cells in our preliminary experiments (unpublished data).

IHC staining indicated that differences in protein expression were more notable across different zones in placenta, rather than between treatment groups. The differences in zone were highlighted using placenta from the control group of rats, in which we detected greater CCBL1 expression in the decidua basale compared to the labyrinth zone. Because CCBL1 can potentially bioactivate DCVC to DCVT (Lash et al. 2014), this finding suggests that the decidua basale may be more sensitive to DCVC toxicity. However, the decidua basale also contains the least cytoplasmic CYP3A1 staining compared to the two other zones under the normal threshold. Because CYP3A1 bioactivates N-acetyl DCVC to N-acetyl DCVC sulfoxide (Werner et al. 1996; Lash et al. 2014), this finding may at least partially explain why TCE did not increase 3-nitrotyrosine or cleaved caspase-3 in the decidua basale or other placental zones compared with control.

Notable limitations exist in our study. Although our TCE dosage of 480 mg/kg/day is within one order of magnitude of the 100 parts per million U.S. Occupational Safety and Health Administration Permissible Exposure Level (Agency for Toxic Substances and Disease Registry 2007), our TCE dosage is several orders of magnitude higher than environmental TCE exposures for human exposures. As an example, our 480 mg TCE/kg/day dosage is approximately 24,000 times higher than the highest detected TCE concentration of 1400 parts per billion in contaminated water at Camp LeJeune (Ruckart et al. 2014). Future studies are needed to investigate TCE reproductive toxicity across a larger range of dosages, including those more relevant to human environmental exposures. Relatedly, although differences between inhalation and ingestion exposure routes, and subsequent distribution and metabolism, could influence TCE toxicity, our study design choice of ingestion represents a common route of TCE exposure, and many similarities between inhalation- and ingestion-derived metabolism of TCE exist (Lash et al. 2014; Agency for Toxic Substances and Disease Registry 2019). The sample size in a few of our analyses is another shortcoming of the current study. For example, the placental CCBL analyses only used a sample size of four placentae, with each placentae derived from a distinct litter, for each fetal sex. It is currently unknown if increasing the sample size in these experiments or using multiple placentae per litter for a given sex would lead to detection of additional statistical significance.

Another limitation relates to exclusion of the NAC-only and AOAA-only treatment groups from the H&E and IHC experiments. Although our intent was to preserve power to test for statistically significant differences for TCE effects, inclusion of these groups in the H&E and IHC analyses could have provided important insight. For example, inclusion of the NAC-only treatment group would have clarified if NAC alone is sufficient to alter placental dimensions or if the alterations are truly dependent on the presence of TCE. Additionally, information on effects of AOAA treatment alone on placental dimensions could have informed conclusions about the role of maternal kidney CCBL activity in the mechanism of action of AOAA. Ultimately, inclusion of H&E and IHC data on placental effects of NAC-only and AOAA-only treatments may have provided additional evidence to support a chain of events for the mechanisms of action for differential treatment effects on fetal weight.

In summary, this is the first report that AOAA prevented TCE-induced fetal weight reduction, using a rat model. Our data suggest that the mechanism was related to renal CCBL activity because AOAA effectively suppressed CCBL in maternal kidney but not in placenta or maternal liver. This is also the first report that NAC pre/co-treatment was more detrimental to pregnancy than TCE treatment by itself. Moreover, this study failed to find evidence of oxidative stress in the placenta under conditions of TCE exposure that reduced fetal weight. Figure 9 depicts our proposed mechanism of action of TCE on timed-pregnant Wistar rats, with modulation of TCE toxicity by NAC and AOAA, to explain the impact of the treatments on fetal weight. The present study highlights that future investigations of modulation of the TCE GSH conjugation pathway, including quantification of metabolite concentrations, will advance understanding of TCE toxicity during pregnancy.

Fig. 9.

Proposed mechanism of action of trichloroethylene (TCE) in the timed-pregnant Wistar rat with modulation by N-acetyl-l-cysteine (NAC) and aminooxyacetic acid (AOAA). Interventions to modulate TCE toxicity are listed in blue. TCE and its metabolites are in rectangular boxes, and an enzyme, cysteine conjugate β-lyase (CCBL), involved in TCE metabolism is in a circular box. DCVC S-(1,2-dichlorovinyl)-l-cysteine, DCVT, 1,2-dichlorovinylthiol