Iodoacetic Acid Exposure Alters the Transcriptome in Mouse Ovarian Antral Follicles

Abstract

Iodoacetic acid (IAA) is an unregulated water disinfection byproduct that is an ovarian toxicant. However, the mechanisms of action underlying IAA toxicity in ovarian follicles remain unclear. Thus, we determined whether IAA alters gene expression in ovarian follicles in mice. Adult female mice were dosed with water or IAA (10 or 500 mg/L) in the water for 35–40 days. Antral follicles were collected for RNA-sequencing analysis and sera were collected to measure estradiol. RNA-sequencing analysis identified 1063 differentially expressed genes (DEGs) in the 10 and 500 mg/L IAA groups (false discovery rate FDR < 0.1), respectively, compared to controls. Gene Ontology Enrichment analysis showed that DEGs were involved with RNA processing and regulation of angiogenesis (10 mg/L) and the cell cycle and cell division (500 mg/L). Pathway Enrichment analysis showed that DEGs were involved in the phosphatidylinositol 3-kinase and protein kinase B, gonadotropin-releasing hormone (GnRH), estrogen, and insulin signaling pathways (10 mg/L). Pathway Enrichment analysis showed that DEGs were involved in the oocyte meiosis, GnRH, and oxytocin signaling pathways (500 mg/L). RNA-sequencing analysis identified 809 DEGs when comparing the 500 and 10 mg/L IAA groups (FDR < 0.1). DEGs were related to ribosome, translation, mRNA processing, oxidative phosphorylation, chromosome, cell cycle, cell division, protein folding, and the oxytocin signaling pathway. Moreover, IAA exposure significantly decreased estradiol levels (500 mg/L) compared to control. This study identified key candidate genes and pathways involved in IAA toxicity and can help to further understand the molecular mechanisms of IAA toxicity in ovarian follicles.

Introduction

Water disinfection is a public health triumph that eliminates water pathogens and provides safe drinking water (Calderon 2000). Disinfectants, such as chlorine, are widely used in water disinfection facilities for this purpose. However, disinfectants can react with organic or inorganic matter in source water to generate compounds called disinfection byproducts (DBPs) (Richardson, Plewa et al. 2007). More than 700 DBPs have been identified in drinking water, and many studies have shown that these compounds are associated with increased risk of cancer and adverse pregnancy outcomes in humans (Villanueva, Cantor et al. 2004, Agency 2010, Richardson S.D. 2011, Wright, Evans et al. 2017, Rivera-Nunez, Wright et al. 2018, Jones, DellaValle et al. 2019). For instance, DBP exposure is associated with increased risk of bladder cancer, colorectal cancer, stillbirth, small gestational age, and birth defects (Villanueva, Cantor et al. 2004, Bove, Rogerson et al. 2007, Hwang, Jaakkola et al. 2008, Wright, Evans et al. 2017, Rivera-Nunez, Wright et al. 2018, Diana, Felipe-Sotelo et al. 2019, Jones, DellaValle et al. 2019, Summerhayes, Rahman et al. 2021). In addition, DBPs have been shown to impair reproductive outcomes in animals (Bodensteiner, Sawyer et al. 2004, Melnick, Nyska et al. 2007, Narotsky MG 2011, Narotsky, Klinefelter et al. 2013, Gonsioroski, Meling et al. 2019, Jiao, Gonsioroski et al. 2021). Specifically, exposure to DBPs can impair ovarian function, spermatogenesis, and fertility outcomes in animals (Bodensteiner, Sawyer et al. 2004, Melnick, Nyska et al. 2007, Narotsky MG 2011, Narotsky, Klinefelter et al. 2013, Gonsioroski, Meling et al. 2019, Jiao, Gonsioroski et al. 2021). Thus, DBPs are considered environmental contaminants and a concern for public health (Wagner 2015).

To date, only eleven DBPs are regulated by the EPA (EPA 2006). Iodoacetic acid (IAA) is one unregulated DBP that has received attention due to its high cytotoxicity and genotoxicity compared to other DBPs (Plewa, Wagner et al. 2004, Plewa, Simmons et al. 2010, Jeong, Gao et al. 2016, Jeong, Machek et al. 2017). People are exposed to IAA on a daily basis mainly via drinking water or by consuming food and beverages that were prepared with disinfected water (Richardson S.D. 2011, Li, Aziz et al. 2021). IAA levels in drinking water are reported to be up to 2.18 μg/L (Richardson, Fasano et al. 2008, Wei, Wang et al. 2013). However, the levels of IAA in drinking water could be underestimated because several factors can influence IAA formation such as the proximity of water facilities to coastal regions and hydraulic fracturing activities (Ersan, Liu et al. 2019, Liu, Ersan et al. 2019). Moreover, IAA formation can be elevated due to source water contamination with iodinated X-ray contrast media and iodine-based sanitizers (Jeong, Machek et al. 2017, Dong and Qiang 2019).

Studies demonstrated that IAA is a toxicant in vitro. Specifically, exposure to IAA inhibited glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in Chinese hamster ovary K1 cells, reducing pyruvate and adenosine triphosphate (ATP) levels, and leading to increased oxidative stress in the cells (Pals, Ang et al. 2011, Dad, Jeong et al. 2013). In addition, IAA exposure altered gene expression in DNA repair and oxidative stress pathways in non-transformed human FHs 74 intestinal cells (Attene-Ramos, Wagner et al. 2010, Pals, Attene-Ramos et al. 2013). Moreover, IAA interacts with catalase and IAA exposure inhibits catalase activity inducing cytotoxicity in mouse primary hepatocytes (Cemeli, Wagner et al. 2006, Wang, Jia et al. 2018).

IAA was shown to be a reproductive toxicant in vitro and in vivo. Specifically, IAA exposure decreased mouse antral follicle growth in vitro and impaired the ability of ovarian follicles to produce sex steroid hormones (Jeong, Gao et al. 2016, Gonsioroski, Meling et al. 2019). IAA also altered expression of important steroidogenic factors, apoptotic genes, and cell cycle regulator genes in mouse antral follicles in vitro (Gonsioroski, Meling et al. 2019). IAA exposure also interfered with mouse oocyte maturation by elevating reactive oxygen species (ROS) levels, disrupting spindle assembly, inducing DNA damage, and causing metaphase I arrest in vitro (Jiao, Gonsioroski et al. 2021). Additionally, IAA exposure in vivo reduced ovarian weight in rats (Xia, Mo et al. 2018). IAA exposure increased mRNA levels of kisspeptin in the arcuate nucleus in the mouse hypothalamus, and reduced follicle stimulating hormone (FSHβ)-positive cell number and FSHβ mRNA expression in vivo and in vitro (Gonzalez, Weis et al. 2021). IAA exposure also affected the mouse pituitary, inducing DNA damage and P21/Cdkn1a expression in vitro and DNA damage and Cdkn1a expression in vivo (Gonzalez, Weis et al. 2021). In our previous study, IAA exposure in vivo led to mice spending less time in the proestrus stage of the estrous cycle and decreased androstenedione and estradiol serum levels compared to control. Additionally, we observed that IAA exposure altered ovarian gene expression of cell cycle regulators, apoptotic factors, the steroidogenic factor Cyp19a1, and estrogen receptor 1 in mice. We did not observe alterations in the appearance or the number of ovarian follicles in mice exposed to IAA (Gonsioroski, Meling et al. 2021), suggesting that IAA exposure may affect the function, but not the structure of the ovary. Although these studies show that IAA is a reproductive toxicant, the underlying mechanisms of IAA-induced toxicity in the ovarian follicles are unclear.

With advancements in genetic sequencing technology, we can observe an extensive range of mRNA transcripts and investigate their levels of transcription. RNA-sequencing transcriptomic analysis could address fundamental questions surrounding IAA-induced reproductive toxicity. Thus, the objective of this study was to investigate differential gene expression and biomolecular mechanisms in ovarian follicles of mice exposed to IAA via drinking water. We dosed female mice with IAA through the drinking water for 35 days and evaluated the ovarian follicles using RNA-sequencing analyses. IAA exposure altered gene expression levels compared to the control groups. Functional annotation clustering analysis described biological processes and molecular functions of IAA toxicity in ovarian antral follicles.

1. Materials and Methods

1.1. Chemical

IAA was purchased from Sigma-Aldrich (St. Louis, MO). We selected IAA because previous studies showed that IAA is cytotoxic and genotoxic to Chinese hamster ovary (CHO) cells (Plewa, Simmons et al. 2010, Wagner and Plewa 2017). In addition, IAA has been shown to be an ovarian toxicant in vitro and in vivo in rodents (Jeong, Gao et al. 2016, Xia, Mo et al. 2018, Gonsioroski, Meling et al. 2019, Gonsioroski, Meling et al. 2021).

1.2. Animals

Female CD-1 mice (26 days old) were purchased from Charles River Laboratories (Charles River, CA). Animals were housed at the University of Illinois at Urbana-Champaign, College of Veterinary Medicine Animal Facility. Mice were kept under 12-h light-dark cycles at 22 ± 1°C and were provided food and water ad libitum. Mice were acclimated to the facility for two weeks before use. The Institutional Animal Use and Care Committee at the University of Illinois at Urbana-Champaign approved all procedures involving animal care, euthanasia, and tissue collection.

1.3. Study Design

Adult female mice (40 days old) were dosed with vehicle (deionized, reverse osmosis filtered water) or IAA (10 or 500 mg/L) dissolved in reverse osmosis water for 35–40 days (n=12 per treatment group) via the drinking water. Two experiments were performed a month apart. In experiment 1, animals were exposed to only water (control) or 500 mg/L of IAA in drinking water. In experiment 2, animals were exposed to only water (control) or 10 mg/L of IAA in drinking water. Bottles were rinsed three times with reverse osmosis water prior to use to ensure no contamination with disinfection byproducts from tap water and detergents that are normally used to clean the bottles in the animal facility. Reverse osmosis water samples from rinsed bottles were analyzed via gas chromatography/mass spectrometry assays for the presence of disinfection byproducts and no contamination was found after the rinsing procedure. The dosing range was based on our previous in vivo study with IAA (Gonsioroski, Meling et al. 2021). Specifically, in our previous study, we observed that 10 and 500 mg/L of IAA altered estrous cyclicity and ovarian gene expression in mice (Gonsioroski, Meling et al. 2021). The doses 10 and 500 mg/L correspond to approximately 70 and 1700 μg of IAA ingested per mouse per day. In detail, we dosed each cage of mice with one bottle of IAA solution weekly. Per week, the average of the solution consumed per cage was 95.8 mL (for the 500 mg/L of IAA group). Since we placed 4 mice per cage, we estimated that each animal drank 23.9 mL of IAA solution per week. Thus, on average, each mouse consumed 3.42 mL of IAA solution per day. We estimated that each mouse would be ingesting approximately 70 and 1700 μg per day of IAA (10 and 500 mg/L of IAA). On average, an adult woman consumes more than 3 liters of water per day, including plain water, beverages, and water in food (Drewnowski, Rehm et al. 2013). Because IAA can be found in drinking water in levels up to 2.18 μg/L (Richardson, Fasano et al. 2008, Wei, Wang et al. 2013) and the concentration of haloacetic acids can reach up to 136 μg/L in drinking water (Srivastav, Patel et al. 2020, Li, Song et al. 2021), this dose range could be in the human exposure range when taking in account all possible routes of exposure to DBPs (3 liters of water consumed per day multiplied by 136 μg of HAA = 408 μg of HAA exposure daily plus dermal and inhalation absorption). Moreover, Teuschler et al. estimated that adult women absorb 3.42 mg/day of total DBPs and about 3 μg/L/hr reach the ovaries (Teuschler, Rice et al. 2004). Oral dosing via drinking water was used to mimic the most common route of exposure to DBPs for humans (Richardson S.D. 2011). Solutions were prepared weekly and immediately placed in water bottles in cages.

1.4. Ovarian follicle dissection

After 35–40 days of dosing, mice were euthanized and antral follicles were dissected from the ovaries. Specifically, we examined estrous cyclicity of the animals every day after the 35 days exposure. Animals that were in diestrus in the day of collections (from 35^th to 40^th days) were euthanized and follicles were dissected from the ovaries. We focused on antral follicles because they are the functional units in the ovary and are required for fertility and sex steroid hormone production. In addition, our previous studies indicate that IAA exposure causes antral follicle toxicity in vitro and in vivo (Jeong, Gao et al. 2016, Gonsioroski, Meling et al. 2019, Gonsioroski, Meling et al. 2021). On the morning of collection, estrous cyclicity was determined and ovaries were collected from mice in diestrus. Immediately after euthanasia, antral follicles were meticulously dissected from the ovaries based on size (220–400 μm) using watchmaker’s forceps as described previously (Jeong, Gao et al. 2016, Zhou and Flaws 2017). Antral follicles (30–40 follicles per mouse, 12 mice per treatment) were dissected, snap frozen in liquid nitrogen separately per mouse, and then stored at −80 °C for RNA sequencing as described below. Blood was collected and sera were used for hormone assays.

1.5. RNA sequencing

Each sample was comprised of 30–40 antral follicles from a single mouse and 5 mice per group in each experiment were selected for RNA-Sequencing. Total RNA was isolated from each sample using a RNeasy Micro kit (Qiagen, Inc., Valencia, California) following the manufacturer’s instructions. RNA was eluted in 14 μl of RNase-free water and the concentration was determined using a NanoDrop (λ = 260 nm; ND 1000; Nanodrop Technologies Inc., Wilmington, Delaware).

The RNAseq libraries were prepared with Illumina’s ‘TruSeq Stranded mRNAseq Sample Prep kit’ (Illumina). Fastq files were generated and demultiplexed with with the bcl2fastq v2.20 Conversion Software (Illumina). The libraries were pooled, quantitated by qPCR, and sequenced on one SP lane for 101 cycles from one end of the fragments on a NovaSeq 6000. Raw reads were checked for quality using FastQC (version 0.11.8) (Andrews 2010) on individual samples then were summarized into a single html report by using MultiQC version 1.9 (Ewels, Magnusson et al. 2016). Average per-base read quality scores were over 30 in all samples and no adapter sequences were found, indicating these reads were high in quality. Thus, the trimming step was skipped and directly proceeded to transcripts mapping and quantification. Salmon version 1.2.0 (Patro, Duggal et al. 2017) was used to quasi-map reads to NCBI’s GRCm39 transcriptome (Annotation Release 109); the transcriptome was first indexed using the decoy-aware method in Salmon with the entire GRCm39 genome as the decoy sequence. Then, quasi-mapping was performed to map reads to transcriptome with additional arguments seqBias and gcBias to correct sequence-specific and guanine-cytosine (GC) content biases, numBootstraps=30 to compute bootstrap transcript abundance estimates and validateMappings and recoverOrphans to help improve the accuracy of mappings. Gene-level counts were then estimated based on transcript-level counts using the “bias corrected counts without an offset” method from the tximport package.

1.6. RNA sequencing analysis

The 20 total samples (five of 10 mg/L, five of 500 mg/L and ten controls (5 from each experiment) were analyzed together. RNA sequencing data from both control groups were combined to facilitate analysis. When comparing expression levels, the numbers of reads per gene need to be normalized because of the differences in total number of reads and because of potential differences in RNA composition such that the total number of reads would not be expected to be the same. The TMM (trimmed mean of M values) normalization (Robinson and Oshlack 2010) in the edgeR (McCarthy, Chen et al. 2012) package uses the assumption of most genes do not change to calculate a normalization factor for each sample to adjust for such biases in RNA composition. In this dataset, TMM normalization factors fluctuate between 0.96 and 1.03, but no consistent differences in RNA composition exist between treatments or experiments. Peixoto et al. (2015) (Peixoto, Risso et al. 2015) show that additional normalization is often necessary in RNA-Seq experiments using a method called “Remove Unwanted Variation” (RUV) (Peixoto, Risso et al. 2015), which estimates extra factors to add to the statistical model. Differential gene expression (DE) analysis was performed using the limma-trend modeling (Chen, Lun et al. 2016) of treatment + 5 RUV factors, of which only the 4^th factor weakly correlated with control experimental batch (Pearson correlation coefficient = 0.67) indicating that any experimental batch effects were minor. All three pairwise comparisons between the treatments were pulled from the model. P-values were adjusted for multiple testing correction by applying a “global” False Discovery Rate (FDR) correction across p-values for all three comparisons together (Benjamini and Hochberg 1995).

Data obtained from RNA sequencing were functionally analyzed using The Database of Annotation, Visualization, and Integrated Discovery Bioinformatics (DAVID) 6.8 according to a previously published protocol (Huang da, Sherman et al. 2009, Huang da, Sherman et al. 2009). A total of 410 DEGs were entered into DAVID (FDR < 0.1) for the 10 mg/L of IAA group and 653 DEGs were entered (FDR < 0.1) for the 500 mg/L of IAA group. “Gene_Ontology” and “Pathways” denoted DAVID defined defaults were selected for functional annotation clustering. To determine if functional gene groups were valuable, annotation clusters with a significant enrichment score > 0.5 were further explored (Huang da, Sherman et al. 2009). In addition, RNA sequencing data were inserted into the web-based portal Metascape (Metascape) to produce functional annotation network cluster graphs and ontology enrichment bar plots. The “Express Analysis” in Metascape was used because it automatically removes ontology terms that did not satisfy the minimal statistical criteria (Zhou, Zhou et al. 2019).

1.7. Analysis of estradiol levels

Blood was obtained from mice of all treatment groups during the collection of the ovaries. Sera were subjected to enzyme-linked immunosorbent assays (ELISAs, DRG International Inc., Springfield, New Jersey) for measurement of estradiol (analytical sensitivity was 10.6 pg/mL and both intra- and inter-assay coefficients of variation were below 15%). Estradiol was selected because it is the primary ovarian estrogen and previous studies showed that IAA treatment decreases estradiol levels (Jeong, Gao et al. 2016, Gonsioroski, Meling et al. 2019, Gonsioroski, Meling et al. 2021). Assays were run according to the manufacturer’s instructions. Lypocheck from Bio-Rad Laboratories was used as a control with known values in all ELISAs in this study.

1.8. Statistical analysis

Statistical analysis for RNA sequencing results were completed using R version 4.0.3. (Team 2018). Hormone level data were expressed as the mean ± standard error of the mean (SEM). Data were analyzed by comparing treatment groups to controls using IBM SPSS version 27 software (SPSS Inc., Chicago, IL, USA). Outliers were removed by the Grubb’s test using GraphPad outlier calculator software (GraphPad Software Inc., La Jolla, CA, USA). Continuous data were assessed for normal distribution by Shapiro-Wilk analysis. If data met assumptions of normal distribution and homogeneity of variance, data were analyzed by one-way analysis of variance (ANOVA) followed by Tukey HSD or Dunnett 2-sided post-hoc comparisons. However, if data met assumptions of normal distributions, but not homogeneity of variance, data were analyzed by ANOVA followed by Games-Howell or Dunnett’s T3 post-hoc comparisons. For all comparisons, statistical significance was determined by p-value ≤ 0.05. When p-values were > 0.05, but < 0.10, data were considered to exhibit borderline significance.

2. Results

2.1. RNA sequencing analysis

Guanine-cytosine (GC) content average ranged from 50–52%, percentage of reads mapped to the transcriptome ranged from 73.5 to 88.4%, and the number of reads ranged from 39.9 to 67.5 million per sample. Following RUV normalization, 172 genes were up-regulated and 238 were down-regulated at global FDR = 0.1 in the 10 mg/L IAA group compared to controls (Table 1). Further, 340 genes were up-regulated and 313 were down-regulated at global FDR = 0.1 in the 500 mg/L IAA group compared to controls (Table 1). Finally, 474 were upregulated and 335 genes were downregulated in the 500 mg/L IAA group compared with the 10 mg/L IAA group (Table 1).

Table 1.

Number of differentially expressed genes (global p < 0.1) from model with 5 RUV factors.

	10mg/L vs. controls	500mg/L vs. controls	500mg/L vs. 10mg/L
Downregulated	238	313	335
Not significant	17759	17516	17360
Upregulated	172	340	474

All genes that had at least one pairwise comparison with global FDR p < 0.1 were selected, resulting in 1443 genes visualized in a heatmap (Figure 1). Heatmaps of these genes show clear overall gene expression level differences across all 3 treatments simultaneously (Figure 1). Clusters with distinct expression patterns across controls and IAA treated groups were separated by color as follows: blue (decreased expression in 500 mg/L only), red (decreased expression in both 10 and 500 mg/L), green (increased expression in 10 mg/L only), turquoise (decreased expression in 10, but increased expression in 500 mg/L with controls intermediate), brown (increased expression in 500 mg/L only), and yellow (decreased expression in 10 mg/L only).

Fig. 1.

Heat map of differentially expressed genes with FDR p-value < 0.1 from limma-trend model of treatment + 5 RUV. Blue cluster (decreased expression in 500 mg/L only), red cluster (decreased expression in both 10 and 500 mg/L), green cluster (increased expression in 10 mg/L only), turquoise cluster (decreased expression in 10, but increased expression in 500 mg/L with controls intermediate), brown cluster (increased expression in 500 mg/L only), and yellow (decreased expression in 10 mg/L only).

2.2. Functional annotation clustering analysis

RNA sequencing data of ovarian follicles of mice treated with IAA were entered into DAVID for Gene Ontology and Pathway enrichment analysis. A total of 410 genes were entered into DAVID (FDR < 0.1) for the 10 mg/L of IAA group compared to controls and 653 genes (FDR < 0.1) were entered for the 500 mg/L of IAA group compared to controls. Gene Ontology results yielded 4 annotation clusters with an enrichment score > 1 (10 mg/L IAA group) and 7 annotation clusters with an enrichment score > 1 (500 mg/L IAA group) (Appendix A Table S2 and S4). Gene Ontology annotation clusters containing the highest enrichment scores were RNA splicing, RNA processing, RNA binding, and regulation of angiogenesis for the 10 mg/L IAA group (Appendix A Table S2). Cell cycle, cell division, and mitotic nuclear division were in the Gene Ontology annotation clusters containing the highest enrichment scores for 500 mg/L group (Appendix A Table S4). Pathway results yielded a total of 2 annotation clusters (10 mg/L IAA group) and 3 annotation clusters with an enrichment score > 0.7 in the 500 mg/L IAA group (Appendix A Table S3 and S5). Within the Pathway annotation clusters containing the highest enrichment scores for 10 mg/L IAA group were the phosphatidylinositol 3-kinase and protein kinase B (PI3K-Akt) signaling pathway (Appendix A Table S3 and Figure S1), gonadotropin releasing-hormone (GnRH) signaling pathway (Appendix A Table S3 and Figure S2), estrogen signaling pathway (Appendix A Table S3 and Figure S3), and insulin signaling pathway (Appendix A Table S3 and Figure S4). For the 500 mg/L IAA group, within the Pathway annotation clusters containing the highest enrichment scores were oocyte meiosis (Appendix A Table S5 and Figure S5), the oxytocin signaling pathway, and the GnRH signaling pathway (Appendix A Table S5 and Figure S6).

RNA-sequencing analysis identified 809 DEGs when comparing the 500 and 10 mg/L IAA groups (FDR < 0.1) (Table 1). Clusters with distinct expression patterns were separated by color (Figure 1 and Appendix A Table S1). Patterns of gene expression between the 10 and 500 mg/L groups were analyzed by functional annotation analyses. Red cluster (Appendix A Table S1) presented downregulated genes related to ribosome, translation, mRNA processing (Gene Ontology) and oxidative phosphorylation (Pathways) in both 10 and 500 mg/L. Turquoise cluster included downregulation of genes involved with chromosome, cell cycle, cell division, protein folding, platelet activation, and the oxytocin signaling pathway in 10 mg/L, but upregulation of these genes in the 500 mg/L group.

RNA sequencing data of ovarian follicles treated with IAA were entered in the Metascape (Metascape) bioinformatics platform. A total of 410 genes were entered into Metascape (FDR < 0.1) in the 10 mg/L IAA group and 653 genes were entered (FDR < 0.1) for the 500 mg/L of IAA group. Metascape generated functional annotation network clusters and bar plots of enriched ontology terms for visualization (Figures 2 and 3). Functional enrichment analyses of differentially expressed genes (DEGs) in the 10 mg/L of IAA group showed that some of these clusters of genes are involved in the response to steroid hormones and insulin secretion (Figure 2). Functional enrichment analyses in the 500 mg/L IAA group showed clusters of genes involved in the RNA metabolic process, the cell cycle, and chromatin organization (Figure 3).

Fig. 2.

Functional enrichment analyses of differentially expressed genes of ovarian follicles of mice treated with 10 mg/L IAA. (a) The enrichment ontology cluster graph represents each term as a circle. The size of the circle is proportional to the number of differentially expressed genes associated with that term. Each cluster is colored uniquely meaning that circles of the same color are associated with the same cluster. (b) Bar chart of Gene Ontology (GO) and Kegg Orthology (KO) terms colored by p-values. Metascape (http://metascape.org) was utilized for visualization.

Fig. 3.

Functional enrichment analyses of differentially expressed genes of ovarian follicles of mice treated with 500 mg/L IAA. (a) The enrichment ontology cluster graph represents each term as a circle. The size of the circle is proportional to the number of differentially expressed genes associated with that term. Each cluster is colored uniquely meaning that circles of the same color are associated with the same cluster. (b) Bar chart of Gene Ontology (GO) and Kegg Orthology (KO) terms colored by p-values. Metascape (http://metascape.org) was utilized for visualization.

2.3. Effects of IAA exposure on estradiol hormone levels

IAA exposure did not alter levels of estradiol in the 10 mg/L group compared to control. However, IAA exposure significantly decreased the levels of estradiol in the 500 mg/L group compared to control (Fig. 4, n = 10–12 per group, p ≤ 0.05).

Fig. 4.

Effects of IAA exposure on estradiol hormone levels. Sera were subjected to enzyme-linked immunosorbent assay for measurement of estradiol. Graphs represent means ± standard error of the mean from 10–12 females per treatment group. *Significant differences compared to control (p ≤ 0.05).

3. Discussion

To our knowledge, this is the first study using RNA sequencing to determine that IAA exposure dysregulates mRNA in mouse ovarian follicles. Our data indicate that exposure to IAA through drinking water for 35 days altered gene expression levels of mouse ovarian follicles compared to the control groups. Functional annotation clustering analysis described alterations in genes involved with RNA processing, regulation of angiogenesis, cell cycle, mitotic cellular division, cell division, the PI3K-Akt signaling pathway, the estrogen signaling pathway, the GnRH signaling pathway, the insulin signaling pathway, and the oxytocin signaling pathway. In addition, exposure to IAA disrupted estradiol levels in mice. This study provided a transcriptome analysis that explores how ovarian antral follicles respond to IAA exposure and presented new targets of interest to be investigated in IAA-induced ovarian toxicity.

Gene Ontology enrichment analysis of antral follicles from mice treated with 10 mg/L of IAA revealed alterations in genes involved with RNA processing and binding compared to controls. Specifically, RNA-binding motif (RBM) genes were downregulated (Rbm4, Rbm14, Rbm8a) among other DEGs by IAA treatment compared to controls groups. The RBM protein family is involved in RNA metabolism including mRNA splicing, RNA stability, and mRNA translation (Li, Guo et al. 2021). A recent study showed that RBM14 is an essential modulator of oocyte meiotic maturation by regulating α-tubulin acetylation that affects spindle morphology and chromosome alignment (Qin, Qu et al. 2021). In addition, RBM proteins are known to be involved in physiological and pathological processes including development, metabolism, proliferation, pluripotency, tumors, and immunity (Fu, Yuan et al. 2021). Thus, alterations in the expression of RBM genes could lead to decreased oocyte quality, oocyte maturation failure, and increased risk of cancers. Moreover, Gene Ontology enrichment analysis of antral follicles from mice treated with 10 mg/L of IAA revealed alterations in the expression of genes involved with regulation of angiogenesis. Precisely, FMS-like tyrosine kinase 1 (Flt1, a receptor for vascular endothelial growth factor), endoglin (Eng), and endothelial specific receptor (Tek) among other DEGs were upregulated with IAA treatment compared to controls. Ovarian function is dependent on the angiogenic vessel network, which enables the follicle to receive oxygen, nutrients, and hormonal support. Abnormal angiogenesis could impair folliculogenesis and increase the risk of polycystic ovary syndrome and ovarian cancer (Xie, Cheng et al. 2017).

Pathway results yielded a total of six annotation clusters for the 10 mg/L IAA group. The cluster with the highest enrichment score contained genes in the PI3K-Akt signaling pathway (Appendix A Figure S1). The PI3K/Akt pathway plays an important role in ovarian cancer pathogenesis and it is involved in folliculogenesis in cows, humans, sheep, and pigs (Makker, Goel et al. 2014, Li, Mo et al. 2017). Specifically, the PI3K-Akt signaling pathway is associated with the recruitment of primordial follicles, granulosa cell proliferation, corpus luteum survival, and oocyte maturation (Adhikari and Liu 2009, Makker, Goel et al. 2014). Mitogen-activated protein kinase kinase 1 (Map2k1), toll-like receptor 4 (Tlr4), integrin beta 5(Itgb5), and phosphoinositide-3-kinase adaptor protein 1 (Pik3ap1) were among the DEGs revealed in the PI3K-Akt signaling pathway. Alterations in these genes could lead to impaired folliculogenesis and subsequent infertility. In addition, within Pathway annotation clusters containing the highest enrichment scores for 10 mg/L IAA group were the GnRH signaling pathway and estrogen signaling pathway (Appendix A Figure S2 and S3). Ovarian follicular development and ovulation are regulated by hypothalamic (GnRH) and pituitary (LH, FSH) hormones. The participation of two cell types, thecal and granulosa cells, and of the two gonadotropins, FSH and LH, underlies the concept of the two cell/two gonadotropin theory, an integrative process required for ovarian estrogen biosynthesis (Luderer 2014). Alterations in the GnRH and estrogen signaling pathway genes could lead to disrupted folliculogenesis and impaired fertility. Further, they could lead to alterations in estradiol levels such as the IAA-induced decrease in estradiol observed in the current study. Interestingly, IAA has been shown to be a hypothalamic-pituitary-gonadal axis toxicant (Gonzalez, Weis et al. 2021). Specifically, IAA exposure increased mRNA levels of kisspeptin in the arcuate nucleus in the mouse hypothalamus, and reduced FSHβ-positive cell number and FSHβ mRNA expression in vivo and in vitro (Gonzalez, Weis et al. 2021). IAA exposure also affected the mouse pituitary, inducing DNA damage and P21/Cdkn1a expression in vitro and DNA damage and Cdkn1a expression in vivo (Gonzalez, Weis et al. 2021). Moreover, IAA exposure affected estradiol levels and estrous cyclicity in mice (Gonsioroski, Meling et al. 2021). Further analysis of the affected genes in the PI3K-Akt, GnRH, and estrogen signaling pathways could provide a greater understanding of how IAA impacts ovarian follicles.

Gene Ontology enrichment analysis of antral follicles of mice treated with 500 mg/L of IAA revealed alterations in genes involved with mitotic nuclear division, cell cycle, and cell division, among other biological processes compared to controls. These biological processes are of great importance for ovarian follicle formation and growth. Folliculogenesis is the process describing the growth and development or atresia of follicles through a series of morphological and functional stages from primordial to ovulatory follicles (Hannon and Curry 2018). Circumstances that disrupt the cell cycle and cell division during this process could affect the growth of follicles and cause infertility. Previous studies have shown that environmental contaminants can disrupt the expression of proteins involved in cell cycle regulation and proliferation, affecting ovarian cell growth (Gupta, Singh et al. 2010, Hannon, Brannick et al. 2015, Zhou and Flaws 2017, Gonsioroski, Meling et al. 2021). In an ovarian follicle culture system, IAA exposure decreased expression of the proliferation marker Ki67, an important factor that regulates the cell cycle progression and cell proliferation (Sun and Kaufman 2018). Our previous study showed that IAA exposure through the drinking water increased expression of cell cycle promoters (Ccna2, Ccnb1, Ccne1, and Cdk4) and decreased expression of the cell cycle promoter Ccnd2 (Gonsioroski, Meling et al. 2021). Future studies on IAA toxicity should investigate the effects of this compound on genes involved in the cell cycle.

Pathway annotation clusters containing the highest enrichment scores showed that the insulin signaling pathway (10 mg/L group) and oocyte meiosis (500 mg/L) (Appendix A Figure S4 and S5, respectively), among other molecular pathways, were altered in the follicles of mice treated with IAA compared to controls. Glucose is an important energy substrate for the generation of the adenosine triphosphate (ATP) required for the metabolic and physiological functions of the ovary (Dupont and Scaramuzzi 2016). It nourishes the development of follicles, the maturation and ovulation of the selected follicles, and formation and maintenance of a functional corpus luteum (Dupont and Scaramuzzi 2016). Insulin allows glucose available in the blood to enter cells and disruptions in the uptake of glucose involving insulin by ovarian cells could lead to impaired folliculogenesis (Dupont and Scaramuzzi 2016). Alterations in the insulin signaling pathway may be explained in part by the fact that IAA, and other regulated monohaloacetic acid DBPs, inhibit GAPDH and block glycolysis, a process that breaks down glucose into two three-carbon compounds and generates energy (Pals, Ang et al. 2011, Pals, Attene-Ramos et al. 2013, Kumari 2018). Facing an energy deficiency, ovarian follicle cells could be trying to overcome this deficit by altering expression of genes involved in the uptake of glucose, tentatively elevating glucose uptake and breakdown. Other genes involved in insulin signaling pathway should be investigated to improve the understanding of ovarian follicle toxicity induced by IAA.

Additionally, genes in the oocyte meiosis pathway (Appendix A Figure S5) were affected in the 500 mg/L IAA group compared to controls. This pattern is consistent with a recent study that showed that IAA exposure interfered with mouse oocyte maturation by elevating reactive oxygen species (ROS) levels, disrupting spindle assembly, inducing DNA damage, and causing metaphase I arrest in vitro (Jiao, Gonsioroski et al. 2021). Further investigation should be done on the other targets disrupted by IAA exposure on the oocyte.

Pathway enrichment analysis of antral follicles of mice treated with 500 mg/L of IAA also revealed alterations in genes involved in the oxytocin signaling pathway (Appendix A Figure S6) compared to controls. Specifically, IAA exposure altered expression of Harvey rat sarcoma virus oncogene (Hras), mitogen activated protein kinase 3 (Mapk3), and natriuretic peptide receptor (Npr2) among other genes compared to controls. Studies have demonstrated that preovulatory follicles in cattle and primates produce oxytocin close to the time of ovulation, suggesting an important role of oxytocin in ovulation and luteinization of follicular cells (Fortune, Komar et al. 2000). In addition, MAPK3 has been shown to be essential for female fertility, being necessary for LH-induced oocyte resumption of meiosis, ovulation, and luteinization (Fan, Liu et al. 2009). Further, NPR2 is known to have a critical role in the ovaries. Mouse females lacking Npr2 function are infertile and present abnormal estrous cyclicity (Geister, Brinkmeier et al. 2011). Thus, alterations in the oxytocin signaling process could cause impaired ovulation and infertility. Further analysis of the altered genes in the oxytocin signaling pathway are necessary for better understanding IAA toxicity on ovarian follicles.

In addition, patterns of gene expression between the 500 and 10 mg/L groups were analyzed. Clusters with distinct expression patterns were separated by color and functional annotation analyses were performed (Gene Ontology and Pathways). Red cluster presented downregulation of genes involved in ribosome, translation, mRNA processing, and oxidative phosphorylation in both 10 and 500 mg/L of IAA (Figure 1 and Appendix A Table S1). As mentioned above, alterations in the expression of genes involved with RNA metabolism could lead to decreased oocyte quality, oocyte maturation failure, and increased risk of cancers. Genes involved in oxidative phosphorylation were also downregulated in both IAA treatment groups. The energy obtained from the breakdown of carbohydrates or fats is derived by oxidative phosphorylation, which is critical for cell function and survival (Wilson 2017). Alterations in this process could cause a depletion of usable energy, leading to increased oxidative stress and eventually leading ovarian follicle cells to apoptosis (Pals, Ang et al. 2011, Dad, Jeong et al. 2013).

Moreover, the turquoise cluster included downregulated genes involved with chromosome, cell cycle, cell division, protein folding, platelet activation, and the oxytocin signaling pathway in 10 mg/L of IAA group. Interestingly, this same group of genes was upregulated in the 500 mg/L of IAA group (Appendix A Table S1). This opposite pattern in gene expression between 10 and 500 mg/L is consistent with the known effects of endocrine disrupting chemicals (EDCs). EDCs are compounds capable of interfering with the endocrine system in any way (NIEHS 2021). EDCs present non-monotonic responses, in which different concentrations of an EDC show opposite effects (Vandenberg, Colborn et al. 2012, Vandenberg 2014). Several hypotheses have been proposed to explain nonmonotonic profiles of EDCs. Some hypotheses include actions at several molecular targets with differing affinities, antagonistic effects, negative feedback regulation to reduce responses, or receptor desensitization (Vandenberg, Colborn et al. 2012). IAA has been shown to affect hormone levels by altering their gene expression (Jeong, Gao et al. 2016, Xia, Mo et al. 2018, Gonsioroski, Meling et al. 2019, Gonsioroski, Meling et al. 2021); therefore, IAA can be considered an EDC. Further research should focus on investigating the effects of IAA exposure on the endocrine system.

4. Conclusion

In conclusion, this study demonstrates that IAA exposure through drinking water for 35 days can have direct effects on the transcriptome of ovarian antral follicles of mice. These discoveries, in conjunction with other studies on IAA toxicity, will serve as the basis for future investigations on ovarian molecular targets for IAA-induced toxicity. These investigations are important for understanding the underlying mechanisms through which IAA causes adverse effects on the female reproductive system.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author. In addition, the data were uploaded in Gene Expression Omnibus (GEO) database at https://www.ncbi.nlm.nih.gov/geo/.