Ex vivo exposures to arsenite and its methylated trivalent metabolites alter gene transcription in mouse sperm cells

Abstract

We have previously reported that preconception exposure to iAs may contribute to the development of diabetes in mouse offspring by altering gene expressions in paternal sperm. However, the individual contributions of iAs and its methylated metabolites, monomethylated arsenic (MAs) and dimethylated arsenic (DMAs), to changes in the sperm transcriptome could not be determined because all three As species are present in sperm after in vivo iAs exposure. The goal of the present study was to assess As species-specific effects using an ex vivo model. We exposed freshly isolated mouse sperm to either 0.1 or 1 μM arsenite (iAs^III) or the methylated trivalent arsenicals, MAs^III and DMAs^III, and used RNA-sequencing to identify differentially expressed genes, enriched pathways, and associated protein networks. For all arsenicals tested, the exposures to 0.1 μM concentrations had greater effects on gene expression than 1 μM exposures. Transcription factor AP-1 and B cell receptor complexes were the most significantly enriched pathways in sperm exposed to 0.1 μM iAs^III. The Mre11 complex and Antigen processing were top pathways targeted by exposure to 0.1 μM MAs^III and DMAs^III, respectively. While there was no overlap between gene transcripts altered by ex vivo exposures in the present study and those altered by in vivo exposure in our prior work, several pathways were shared, including PI3K-Akt signaling, Focal adhesion, and Extracellular matrix receptor interaction pathways. Notably, the protein networks associated with these pathways included those with known roles in diabetes. This study is the first to assess the As species-specific effects on sperm transcriptome, linking these effects to the diabetogenic effects of iAs exposure.

1. Introduction

Inorganic arsenic (iAs) is a collective name for several naturally occurring inorganic forms of the toxic metalloid arsenic (As). Drinking water is the major source of exposure to iAs for humans. The most common forms of iAs present in natural water are oxyanions of tri- and pentavalent As, arsenite (iAs^III) and arsenate (iAs^V), both of which exhibit high solubility over a wide range of pH and Eh conditions (Duker et al., 2005). Globally, >200 million people from at least 40 countries are at risk of adverse health effects associated with iAs exposure via drinking water at levels that exceed 10 μg As/L, the WHO guideline value for drinking water As concentration (WHO, 2017). Food is a significant contributor to iAs exposure, especially in regions with low levels of iAs in drinking water (Ravenscroft and Brammer, 2009; FAO/WHO, 2011).

Humans and many other mammalian species metabolize iAs in a pathway that methylates iAs to yield trivalent and pentavalent mono- and dimethylated metabolites (MAs and DMAs, respectively) (Cullen, 2014). The methylation of iAs is a detoxification mechanism, which is essential for efficient clearance of As from the body. However, overwhelming evidence suggests that the trivalent MAs^III and DMAs^III, which are more toxic than iAs species or the pentavalent methylated arsenicals, significantly contribute to adverse health effects associated with iAs exposure (Stýblo et al., 2021). Epidemiological studies have linked chronic iAs exposure to skin, lung, liver and bladder cancers, as well as cardiovascular diseases, type II diabetes, and neurological and reproductive disorders (Abdul et al., 2015). Most of these studies were carried out in populations with a long history of iAs exposure, often affecting multiple generations. Yet, little is known about how exposures to iAs at different stages of life or transgenerational exposures contribute to the adverse health outcomes.

Several population studies have shown that prenatal exposure to iAs could be a significant risk factor for development of diseases during adolescence or adulthood, pointing to altered epigenetic mechanisms as drivers of this risk (Reichard and Puga, 2010). However, these studies did not differentiate between the effects of in utero exposure of the fetus and effects of preconception exposure of mothers or fathers, specifically effects of iAs or its toxic methylated metabolites on germ cells, which may affect development of the fetus and/or the phenotype of the offspring. Our laboratory has recently addressed this knowledge gap using laboratory mouse models. We have shown that a combined preconception and in utero exposure to iAs resulted in insulin resistance in a sex-specific manner in mice from one of the collaborative cross strains (CC004) (Huang et al., 2018a; Fry et al., 2019). In a follow up study, iAs exposure of C57BL/6 sires and dams before mating was found to alter transcriptomic profiles in paternal sperm and lead to development of a sex-specific diabetic phenotype in offspring (Venkatratnam et al., 2021). Notably, some of the genes altered by iAs exposure in paternal sperm were associated with pathways that were also altered in livers of the offspring, including pathways linked to diabetes.

Laboratory mice, like humans, methylate iAs to MAs and DMAs metabolites (Vahter, 2002). These metabolites are found in tissues of mice exposed to iAs along with the parent compound (Currier et al., 2016). Therefore, it is practically impossible to determine the extent by which iAs or its toxic methylated metabolites contribute to the adverse effects of iAs exposure in specific tissues. This question can be addressed using in vitro or ex vivo models, in which cells from target tissues are exposed to individual arsenicals, the known metabolites of iAs. The goal of the present study was to compare transcriptomic profiles in freshly isolated mouse sperm cells exposed ex vivo to trivalent iAs^III, MAs^III or DMAs^III and in control, unexposed sperm cells. We hypothesized that ex vivo exposures to the trivalent arsenicals would lead to differential, As species-specific expression of genes in pathways that were identified as targets of the preconception iAs exposure in mouse sperm in our published in vivo study (Venkatratnam et al., 2021).

2. Material and methods

2.1. Chemicals and reagents

Bovine serum albumin (BSA) (heat shock fraction, pH = 7, ≥ 98%), calcium chloride dihydrate $\left({\text{CaCl}}_2^{.}2{\text{H}}_2\text{O}\right)$, magnesium sulfate heptahydrate $\left({\text{MgSO}}_4^{.}7{\text{H}}_2\text{O}\right)$, monopotassium phosphate (KH₂PO₄), potassium chloride (KCl), sodium pyruvate (Na-pyruvate), and sodium lactate (Na-lactate) 60% syrup were purchased from Sigma-Aldrich (St. Louis, MO). Sodium chloride (NaCl) and anhydrous granular glucose were purchased from Mallinckrodt Chemicals (St. Louis, MO). Sodium bicarbonate (NaHCO₃) was from VWR Life Science Amresco (Cleveland, OH). Gentamicin 50 mg/mL and 1 M 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer solutions were purchased from Thermo Fisher Scientific/Gibco (Hampton, NH). Dimethyl sulfoxide (DMSO) and 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT) were purchased from EMD Millipore Corporation (Billerica, MA). iAs^III (sodium arsenite, >99% pure) was from Sigma-Aldrich. The methylated trivalent arsenicals, MAs^III (methylarsine oxide, >98% pure) and DMAs^III (iododimethylarsine, >98% pure), were obtained from Dr. William Cullen (university of British Columbia, Canada).

2.2. Mice

Eleven-week-old male C57BL/6 J mice were purchased from Jackson Laboratory (Bar Harbor, ME), and housed five mice per cage at a 12-h light–dark cycle in the UNC animal facilities. All mice were fed a pelleted AIN-93G Purified Rodent Diet from Dyets Inc. (Bethlehem, PA) and drank deionized water for 2 weeks prior to experiments. This 2-week conditioning allows for clearance of As to which the mice were exposed in Jackson facilities while fed a regular laboratory chow, which can contain substantial amounts of iAs or organic As species (Murko et al., 2018). All procedures involving mice were approved by the University of North Carolina (UNC) Institutional Animal Care and Use Committee.

2.3. Sperm collection and culture

Modified Human Tubal Fluid (HTF) -HEPES culture medium without phenol red was prepared freshly before each experiment (Cold Spring Harbor Protocols, 2017; The Jackson Laboratory, 2020). Briefly, 2.969 g NaCl, 0.175 g KCl, 0.025 g ${\text{MgSO}}_4^{.}7{\text{H}}_2\text{O}$, 0.025 g KH₂PO₄, 0.15 g ${\text{CaCl}}_2^{.}2{\text{H}}_2\text{O}$, 1.05 g NaHCO₃, 0.25 g glucose, 0.018 g Na-pyruvate, 1.71 mL Na-lactate 60% syrup, 0.5 mL Gentamycin, 52.5 mL 1 M HEPES solution, and 2.0 g fetal bovine serum albumin were mixed in 500 ml deionized water. The pH of the medium was adjusted to 7.4 with HCl. The medium was then sterile filtered with 0.22 μm filter purchased from Corning Life Sciences (Corning, NY), and stored at 4 °C overnight. Prior to sperm isolation, the medium was pre-warmed in a 37 °C water bath for at least 15 min. Sperm was collected following the previously published swim out method (The Jackson Laboratory, 2020). Briefly, mice were euthanized by cervical dislocation. Cauda epididymis and vas deferens were collected and placed in an untreated cell culture dish with 2 ml pre-warmed culture medium. The cauda epididymis was then punctured with syringes under microscope, and placed in 37 °C water bath for 10 min to release sperm cells. Only freshly collected sperm was used in this study.

2.4. Ex vivo exposures to trivalent arsenical

Sperm suspension was prepared by pooling sperm cells from 20 mice. Cell count in the suspension was determined using a 5 × 5 grid-bordered hemocytometer. Cells were then aliquoted into 6-well non-treated culture plates purchased from Corning Life Sciences (Corning, NY), 3 million cells in 3 ml culture medium per well. The cells were incubated for 4 h with iAs^III, MAs^III, or DMAs^III at final concentration of 0.1, 1 or 10 μM in five technical replicates. The incubation was carried out in open air at room temperature. Unexposed sperm cells incubated in parallel with the exposed cells were used as controls.

2.5. Cell viability assay

Cell viability was measured in sperm exposed to the trivalent arsenicals and in control cells using MTT assay (Riss et al., 2004; Buranaamnuay, 2021). This assay measured conversion of 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT) to purple formazan by viable cells. After 2-h incubation with MTT, cells were dissolved in DMSO, and optical density (OD) was measured at 570 and 620 nm wavelengths, corresponding to the maximum absorbance of formazan and MTT, respectively. Delta OD (OD₅₇₀ - OD₆₂₀) was used as the quantifying measurement of sperm cell viability. One-way ANOVA and Bonferroni post-hoc tests were used to evaluate effects of exposures on cell viability. Differences between exposed cells and controls with p < 0.05 were considered statistically significant.

2.6. Gene expression analysis

2.6.1. RNA-seq analysis

RNA extraction, library preparation, RNA-seq, and reads alignment were performed at the UNC Center for Gastrointestinal Biology and Disease Advanced Analytical Core. RNA was extracted using Quick-RNA Microprep (Zymo Research, Irvine, CA), with the on-column DNaseI step included. Qubit RNA Broad-Range Assay Kit (Invitrogen, Carlsbad, CA) was used for RNA quantification. RNA quality was determined using RNA 6000 Pico kit (Agilent Technologies, Santa Clara, CA). mRNA was selected with NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs, Ipswich, MA), and cDNA library was prepared using NEBNext Ultra II Directional RNA Library Prep Kit and NEB Next Multiplex Oligos for Illumina (96 Unique Dual Index Primer Pairs) from the same company. The sequencing libraries were amplified using 14 PCR cycles. RNA single-end sequencing (read 1 = 76 bp) was conducted by Illumina Next-Seq 500 High output kit v2.5 (75 cycles) from Illumina (San Diego, CA), with 8 bp dual index, 1% Phix control, and 1.8 pM loading concentration. Reads were aligned to the mm9 genome using Spliced Transcripts Alignment to a Reference (STAR) (Dobin et al., 2013).

Differences in gene expression between the control group and each of the treatment groups were determined using the R package DESeq2 (Love et al., 2014). Before running the comparison test, lowly expressed genes with the total sum of raw read counts no >20 across sperm samples used in that set of comparison were pre-filtered. DESeq2 first converted raw counts into normalized counts by dividing the raw counts by a sample-size specific size factor. The analysis then applied default independent filtering of DESeq2 for removing genes with low mean normalized counts, which is independent of the manual pre-filtering, and a Cook’s cutoff of 0.4 that identifies outliers in normalized counts. Those outliers are then replaced with a 20% trimmed mean of their genes’ normalized counts. Genes with significantly altered normalized expression in each comparison were identified using Benjamini–Hochberg false discovery rate (FDR) correction (FDR < 0.10).

2.6.2. Quantitative RT-PCR analysis

To validate RNA-seq results, the most down- or upregulated genes with the highest absolute counts were selected from each of the 0.1 μM exposure groups for quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis. RNA was extracted from aliquots of the sperm samples analyzed by RNA-seq using the RNEasy Mini Kit from Qiagen (Germantown, MD). For Dnm3os, there were no commercially available qRT-PCR primers so the following forward and reverse PCR primers were ordered from IDT DNA (Research Triangle Park, NC) 5′-gggcagggactttcttcctc-3′ and 5′-tacgcggtctgtttctgctt-3′, respectively. The qRT-PCR analysis was performed by a StrataGene Mx3005P qPCR analyzer (Agilent Technologies, Santa Clara, CA) using QuantiTect Reverse Transcription Kit and QuantiTect Primer Assays purchased from Qiagen (Germantown, MD) corresponding to the selected genes. Beta-actin (Actb) was used as the housekeeping gene for all qRT-PCR tests. Each sample was tested in 5 technical replicates. Outlier cycle threshold (CT) values (identified by the Grubbs’ Test) were excluded from further calculations. Fold change values was calculated by average 2^(−ΔΔCT). The ΔCT value was calculated as difference between target gene CT value and housekeeping gene CT value in the same sample replicate (Xu et al., 2013). The ΔΔCT values were calculated by subtracting the average ΔCT value in the control (unexposed) sperm samples from ΔCT values of the target genes in each exposure groups. Student’s t-test was used to evaluate differences between the relative mean expression of a gene in the exposed and control groups; p value<0.05 was considered significant.

2.6.3. Pathway and protein network analyses

The KEGG and Gene Ontology (GO) database were used to identify biological pathways enriched for the differentially expressed genes, annotating genes in the following functional categories: GO term biological process (BP), GO term cellular component (CC), GO term molecular function (MF), and KEGG pathways. The related protein networks were identified using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) (Szklarczyk et al., 2019). Pathways and functional protein networks with an FDR < 0.05 were considered statistically significant. Enrichment strength of the protein clusters and pathways was calculated by log₁₀ of the ratio between the number of proteins related with the differentially expressed genes that were annotated with a term and the number of proteins that were expected to be annotated with this term in a random network of the same size (Szklarczyk et al., 2019).

2.7. Speciation analysis of As

Quantitative speciation analysis of As was carried out by hydride generation-cryotrapping- inductively coupled-plasma-mass spectrometry (HG-CT-ICP-MS), using Agilent 7900 mass spectrometer as a detector (Matoušek et al., 2017). Before analysis, lysed sperm cells were treated with 2% cysteine which converts the pentavalent arsenicals to trivalency (Matoušek et al., 2008). This method determined concentrations of iAs (iAs^III + iAs^V), MAs (MAs^III + MAs^V) and DMAs (DMAs^III + DMAs^V). Total As concentration was calculated as sum of the concentrations of each of the As species.

3. Results

In the first approach, we determined viability of sperm cells exposed to a wide range of iAs^III, MAs^III or DMAs^III concentrations. The concentrations of arsenicals that were not cytotoxic or caused only a minor decrease in cell viability were then used in the second step to characterize effects of the arsenicals on gene expression profiles, and to identify pathways and protein cluster associated with the differentially expressed genes. The concentrations and chemical forms of As retained in the sperm after the exposures were also determined.

3.1. Viability of sperm exposed to trivalent arsenicals

Freshly collected sperm pooled from 20 mice was used in this experiment. Cell viability was measured in sperm cells after 4-h exposure to 0.1, 1 or 10 μM iAs^III, MAs^III or DMAs^III and in control (unexposed) sperm cells using MTT assay (Fig. 1). Exposure to iAs^III did not affect cell viability at any of the exposure levels tested. No significant differences were found in viability of control sperm cells and sperm cells exposed to 0.1 μM MAs^III and 0.1 or 1 μM DMAs^III. However, exposures to 1 and 10 μM MAs^III decreased cell viability by 19.1% (p = 0.010) and 41.7% (p = 0.021) respectively. Similarly, viability of sperm cells exposed to 10 μM DMAs^III was lower than that of control sperm by 38.6%, (p = 0.007).

Fig. 1.

Viability of sperm cells exposed to 0.1, 1 or 10 μM iAs^III, MAs^III or DMAs^III for 4 h and the control, unexposed sperm determined by MTT assay (Delta OD = OD570 - OD620); mean ± SD, N=3.* Viability of sperm exposed to arsenicals is significantly different from that of control sperm (p < 0.05).

3.2. Effects of trivalent arsenicals on gene expression in sperm cells

Sperm pooled from additional 20 mice was exposed for 4 h to 0.1 or 1 μM iAs^III, MAs^III or DMAs^III, the concentrations that had no or minor effects on cell viability. Unexposed cells incubated for 4 h in parallel with the exposed cells were used as controls. After incubation, sperm cells from each treatment group were separated into 3 aliquots. Each aliquot was centrifuged at 15,000 rpm for 5 min. Cell pellets were stored at −80 °C. Cells from these aliquots were used for RNA-seq and qRT-PCR analyses, and for analysis of As species. The results of these analyses are as follows:

3.2.1. Transcriptomic effects of trivalent arsenicals determined by RNA-seq

Analysis of RNA-seq data identified a total of 1177 and 106 genes which transcripts were significantly altered by exposures to trivalent arsenicals (FDR < 0.10) at 0.1 μM and 1 μM level, respectively (Suppl. Table 1). Thus, the 0.1 μM exposures had more widespread effects on transcript levels than the 1 μM exposures. Exposures to 1 μM iAs^III, MAs^III and DMAs^III altered transcript levels of 15, 72, and 27 genes, respectively. In comparison, 851, 437 and 309 gene transcripts were altered by exposures to 0.1 μM iAs^III, MAs^III and DMAs^III, respectively. The top 10 upregulated and downregulated genes in sperm exposed to each of the arsenicals are shown in Table 1. Among the top altered genes in sperm exposed to 1 μM arsenicals were Jun, Fos, Junb and Egr1 (MAs^III), Fn1 (iAs^III and DMAs^III), and Spns2 (DMAs^III). There were no genes differentially expressed by exposures to all three arsenicals at 1 μM level. In contrast, 101 identical genes were differentially expressed in sperm exposed to either of the three arsenicals at 0.1 μM level (Fig. 2, Table 1G), (highlighted in Suppl. Table 1A–C). Among the genes that were dysregulated by all three arsenicals at 0.1 μM concentration were Rilp, Tnfaip3, Col4a4, Itga4, and Kcnq1ot1, with Rilp being one of the most upregulated genes. Other top dysregulated genes identified in sperm exposed to 0.1 μM arsenicals included Vgf (iAs^III) and Fabp2 (DMAs^III). The directions of change for all the overlapping genes in sperm exposed to 0.1 μM arsenicals were consistent: 61 genes were upregulated and 40 downregulated (Fig. 2).

Table 1.

Top 10 genes that were differentially expressed in sperm cells exposed to (A) 1 μM iAs^III, (B) 1 μM MAs^III, (C) 1 μM DMAs^III, (D) 0.1 μM iAs^III, (E) 0.1 μM MAs^III and (F) 0.1 μM DMAs^III. Only 2 down-regulated genes were present in sperm exposed to 1 μM iAs^III.

Gene	log2 Fold Change	p value	FDR
(A)
Protein Disulfide Isomerase Family A Member 6 (Pdia6)	−0.24	1.27E-04	2.36E-02
45S Pre-ribosomal RNA (Rn45s)	−0.22	1.30E-05	3.82E-03
THO Complex 2 (Thoc2)	0.37	4.76E-05	1.22E-02
Arginine and Glutamate Rich 1 (Arglu1)	0.40	1.03E-08	2.11E-05
Fibronectin 1 (Fn1)	0.57	5.42E-04	8.54E-02
Arachidonate 15-Lipoxygenase (Alox15)	0.61	1.01E-06	5.33E-04
Metastasis Associated in Lung Adenocarcinoma Transcript 1 (Malat1)	0.82	1.04E-06	5.33E-04
(B)
Fos Proto-Oncogene, AP-1 Transcription Factor Subunit (Fos)	−1.15	2.31E-34	2.55E-31
Jun Proto-Oncogene, AP-1 Transcription Factor Subunit (Jun)	−1.08	3.07E-30	1.70E-27
Early Growth Response 1 (Egr1)	−0.77	1.84E-25	6.80E-23
Tristetraprolin (Zfp36)	−0.62	3.73E-18	1.03E-15
Transcription Factor JunB (Junb)	−0.56	7.42E-17	1.64E-14
Guanylate Kinase (Guk1)	0.25	4.27E-03	8.02E-02
Transgelin 2 (Tagln2)	0.26	4.89E-05	3.61E-03
ATP Synthase F1 Subunit Epsilon (Atp5e)	0.27	1.01E-03	3.12E-02
Chloride Intracellular Channel Protein 1 (Clic1)	0.28	5.90E-04	1.98E-02
Dipeptidyl Peptidase Like 6 (Dpp6)	0.28	2.25E-03	5.02E-02
(C)
Tripartite Motif Containing 66 (Trim66)	−4.51	8.77E-05	6.84E-02
UDP Glycosyltransferase Family 3 Member A2 (Ugt3a2)	−3.14	7.92E-05	6.52E-02
Double PHD Fingers 1 (Dpf1)	−2.93	1.03E-04	7.26E-02
2810471M01Rik	−1.89	5.91E-05	5.16E-02
Sodium Voltage-Gated Channel Alpha Subunit 7 (Scn7a)	−1.50	1.45E-05	3.58E-02
Proteoglycan 4 (Prg4)	0.73	4.56E-11	6.75E-07
Selectin P Ligand (Selplg)	0.74	4.58E-05	4.68E-02
Fibronectin 1 (Fn1)	0.86	9.65E-08	7.15E-04
Serpin Family B Member 2 (Serpinb2)	1.01	1.74E-05	3.69E-02
Sphingolipid Transporter 2 (Spns2)	1.19	9.92E-05	7.26E-02
(D)
Interleukin-1 Family Member 8 (Il1f8)	−6.07	1.47E-06	6.00E-04
2610316D01Rik	−5.78	2.06E-06	7.85E-04
A330076H08Rik	−5.42	1.40E-04	1.27E-02
Solute Carrier Family 16 Member 4 (Slc16a4)	−5.37
		2.52E-05	4.08E-03
Claudin 19 (Cldn19)	−5.29	9.33E-05	9.78E-03
B830017H08Rik	4.00	4.64E-03	9.68E-02
Calcium Voltage-Gated Channel Auxiliary Subunit Gamma 1 (Cacng1)	4.35	9.94E-04	4.24E-02
VGF Nerve Factor Inducible (Vgf)	4.40	2.96E-03	7.68E-02
Rab-interacting lysosomal protein (Rilp)	4.60	1.92E-03	6.07E-02
Small Nucleolar RNA, H/ACA Box 70 (Snora70)	4.81	8.01E-05	9.00E-03
(E)
Leucine Rich Repeat Containing 69 (Lrrc69)	−5.39	7.26E-05	1.43E-02
Gm15881	−4.80	2.62E-05	6.98E-03
Gm12228	−4.43	2.39E-03	9.27E-02
Cytochrome P450 2D11 (Cyp2d11)	−4.42	1.32E-03	6.87E-02
Cytochrome P450 2E1 (Cyp2e1)	−4.26	1.98E-03	8.39E-02
Small Nucleolar RNA, H/ACA Box 70 (Snora70)	4.02	1.24E-03	6.51E-02
Ras-related Protein Rab39 (Rab39)	4.08	1.04E-03	5.87E-02
Cytochrome B5 Reductase 2 (Cyb5r2)	4.15	1.46E-03	7.27E-02
C920025E04Rik	4.56	9.08E-04	5.55E-02
Rab-interacting lysosomal protein (Rilp)	4.66	8.90E-04	5.51E-02
(F)
Leucine Rich Repeat Containing 69 (Lrrc69)	−5.34	1.14E-04	2.19E-02
Gm15881	−5.33	5.90E-06	3.56E-03
Potassium Voltage-Gated Channel Subfamily H Member 6 (Kcnh6)	− 5.11	1.04E-04	2.08E-02
Transmembrane Protein 246 (Tmem246)	−5.07	1.91E-04	3.05E-02
Tetratricopeptide Repeat Domain 30A2 (Ttc30a2)	−4.95	3.17E-04	4.22E-02
4930548G14Rik	4.61	8.18E-04	6.95E-02
9930012K11Rik	4.67	1.79E-04	2.88E-02
Rab-interacting lysosomal protein (Rilp)	4.71	2.41E-04	3.44E-02
Fatty Acid Binding Protein 2 (Fabp2)	4.74	4.66E-04	5.21E-02
C130050O18Rik	4.76	8.64E-05	1.92E-02

Fig. 2.

Numbers of differentially expressed genes in sperm exposed to trivalent arsenicals (FDR < 0.10). Venn diagrams show overlapping and exposure-specific genes differentially expressed in sperm exposed to 1 μM (Panel A) and 0.1 μM (Panel B) of iAs^III, MAs^III and DMAs^III. The arrows indicate genes that were significantly up- or downregulated. The directions of change of 2 of the 3 significantly altered genes differ in cells exposed to 1 μM iAs ^III vs cells exposed to 1 μM MAs ^III.

3.2.2. Validation of RNA-seq results by qRT-PCR

To validate results of RNA-seq analysis, transcripts of thirteen genes among the most up- and downregulated genes in the sperm exposed to 0.1 μM iAs^III, MAs^III or DMAs^III were analyzed by qRT-PCR, namely Abca13, Arpp21, Cacna1e, Dnm3os, H2-DMb2, Hapln1, Hfm1, Mmp12, Ms4a6d, Myrf, Sema3a, Slc22a3, and Spns2. These genes were selected because they had the highest transcript copy numbers among the top altered genes, thus ensuring successful amplification during qRT-PCR. Specificity of the QuantiTect and IDT DNA primers for the selected genes was first examined using melting curve plots (Suppl. Fig. 1). Arpp21 and Hapln1 failed to amplify while melting curves for the other 11 genes indicated successful amplification yielding a single product. Results of the qRT-PCR analysis of nine of these genes in the exposed vs. control sperm cells were consistent with results of the RNA-seq analysis, showing the same direction of change and, for four genes (H2-Dmb3, Ms4a6d, Mmp12, and Spns2), also similar degrees (folds) of change (Fig. 3; Suppl. Table 2A). The agreement is statistically significant (p = 0.033). The transcription of five genes (Myrf, Slc22a3, Abca13, Cacnale, and Sema3a) changed in the same direction, but the fold change determined by qRT-PCR for each of these genes was smaller than that determined by RNA-seq. The qRT-PCR analysis of two of the 11 genes, Hfm1 and Dnm3os, showed opposite directions of change compared to the RNA-seq analysis.

Fig. 3.

Validation of RNA-seq data by qRT-PCR analysis of eleven most up- and downregulated genes from sperm exposed to 0.1 μM iAs^III, MAs^III or DMAs^III (Mean + SE, n = 3–5).

Differences in the transcript levels were determined by qRT-PCR in the control and the exposed sperm cells. The differences in H2-Dmb2, Mmp12, and Slc22a3 transcripts between control cells and cells exposed to 0.1 μM MAs^III were found to be statistically significant (p < 0.05); the differences in other gene transcripts did not reach statistical significance (Suppl. Table 2B).

3.2.3. Pathways enriched for the differentially expressed genes and the associated protein clusters

We used the functional annotation tool for GO Term and KEGG pathway analysis in the STRING database to identify biological pathways enriched for the genes differentially expressed by exposures to trivalent arsenicals and the associated protein networks (FDR < 0.05) (Suppl. Table 3). We found that more pathways and clusters were associated with exposures to arsenicals at 0.1 μM as compared to 1 μM level, including 610, 223, and 70 functional clusters and pathways in the sperm exposed to iAs^III, MAs^III and DMAs^III, respectively. In comparison, one and 107 significantly enriched biological pathways were associated with exposure to 1 μM iAs^III and MAs^III, respectively. Transcription factor AP-1 complex and B cell receptor complex were the most significantly enriched pathways in sperm exposed to 0.1 μM iAs^III, Mre11 complex was the most significantly enriched pathway in sperm exposed to 0.1 μM MAs^III, and pathways associated with antigen processing were most enriched in sperm exposed to 0.1 μM DMAs^III.

Using the STRING local network clustering analysis, we identify specific protein networks associated with the genes differentially expressed by exposures to trivalent arsenicals (FDR 0.05) (Suppl. Table 4). Four local networks of functional protein clusters were identified in sperm exposed to 0.1 μM MAs^III. Among these networks was the Mixed, incl. ECM-receptor interaction and matrix metalloproteinases network, which included the significantly enriched ECM-receptor interaction KEGG pathway. Another four local protein networks were significantly enriched in sperm exposed to 1 μM MAs^III, including two networks surrounding the protein cluster transcription factor AP-1 complex.

Protein clusters and biological pathways with shared proteins were mapped together in interaction networks by STRING’s network tool. For example, a protein cluster within the focal adhesion pathway was identified as significantly enriched in sperm exposed to 0.1 μM iAs^III (enrichment strength = 0.44, FDR = 0.0072). Other protein clusters and pathways associated with this exposure were also identified, specifically, PI3K-Akt signaling pathway (enrichment strength = 0.28, FDR = 0.0476), NF-Kappa B signaling pathway (enrichment strength = 0.54, FDR = 0.0123), and transcription factor AP-1 complex (enrichment strength = 1.33, FDR = 0.0358) (Fig. 4A). The focal adhesion pathway was also enriched by exposure to 0.1 μM MAs^III (enrichment strength = 0.51, FDR = 0.0299) and DMAs^III (enrichment strength = 0.64, FDR = 0.0151) (Fig. 5A, Fig. 6). Some of the most enriched protein networks in sperm exposed to 0.1 μM iAs^III were linked to adaptive immunity, including B-cell receptor complex and MHC class ii complex (Fig. 4B). The protein clusters associated with ECM-receptor interaction pathway, which shared multiple proteins with focal adhesion pathway, was significantly enriched in sperm exposed to 0.1 μM MAs^III (Fig. 5A). Additionally, pathways related with Wnt signaling and Type I Diabetes Mellitus were significantly enriched for exposure to 0.1 μM MAs^III, while no protein was shared with previously annotated pathways and protein clusters (Fig. 5B, C).

Fig. 4.

Protein network clusters and biological pathways corresponding to differentially expressed genes in sperm exposed to 0.1 μM iAs^III. Line thickness represents the strength of the interaction between nodes, with thicker lines representing stronger interactions. Color of a protein node symbolizes its role in specific pathways or protein networks. (A) Red: focal adhesion pathway; Blue: NF-Kappa B signaling pathway; Green: PI3K-Akt signaling pathway; Yellow: Transcription factor AP-1 complex. (B) Red: MHC Class II protein complex.

Fig. 5.

Protein network clusters and biological pathways corresponding to differentially expressed genes in sperm exposed to 0.1 μM MAs^III. Line thickness represents the strength of the interaction between nodes, with thicker lines representing stronger interactions. Color of a protein node symbolizes its role in specific pathways or protein networks. (A) Red: focal adhesion pathway (KEGG pathways); Blue: focal adhesion pathway (Gene Ontology Term, Cellular Component); Green: ECM-receptor interaction pathway; Yellow: Mixed, incl. ECM-receptor interaction and matrix metalloproteinases protein cluster. (B) Purple: regulation of Wnt signaling pathway; Blue: Wnt signaling pathway; Red: positive regulation of Wnt signaling pathway; Green: negative regulation of Wnt signaling pathway. (C) Red: Type 1 Diabetes Mellitus pathway.

Fig. 6.

Biological pathway corresponding to differentially expressed genes in sperm exposed to 0.1 μM DMAs^III. Line thickness represents the strength of the interaction between nodes, with thicker lines representing stronger interactions. Color of a protein node symbolizes its role in specific pathways or protein networks. Red: focal adhesion pathway.

3.2.4. Arsenic species in sperm cells exposed to trivalent arsenicals

Speciation analysis of As in sperm cells exposed to iAs^III, MAs^III and DMAs^III and control sperm cells was carried out by HG-CT-ICP-MS. Results of this analysis are shown in Fig. 7 and Suppl. Table 5. Control sperm cells contained on average 658.5 picogram (pg) of total As per million cells, with DMAs, MAs and iAs representing 407.1 pg (61.8%), 161.7 pg (24.5%), and 89.8 pg (13.6%), respectively. The sperm cells exposed to trivalent arsenicals contained almost exclusively the As species to which they were exposed, suggesting that a little or no conversion of As species took place during the 4-h exposures. As expected, more As was retained in sperm exposed to 1 μM than 0.1 μM concentrations, although the levels of total As in the cells were not in proportion with the exposure level. At both exposure levels, MAs^III was retained to a greater extent than DMAs^III and iAs^III.

Fig. 7.

Arsenic species in sperm cells exposed to 0.1 or 1 μM iAs^III, MAs^III or DMAs^III, and the control sperm cells (pg As per 10⁶ cells). (Mean + SD, n = 5).

4. Discussion

Results of previously published population and laboratory studies have linked iAs exposure to diabetes and identified insulin resistance and/or beta cell dysfunction as the underlying mechanisms (Walton et al., 2004; Navas-Acien et al., 2006; Huang et al., 2011). Exposures to iAs^III or its metabolites, MAs^III and DMAs^III, have been shown to impair glucose stimulated insulin secretion and beta-cell function through multiple pathways, including ROS induction and ROS/NRF2 mediated apoptosis of pancreatic beta cells, impairment of mitochondrial metabolism, and inhibition of calcium influx in beta cells (Douillet et al., 2013; Pachauri et al., 2013; Duan et al., 2015; Dover et al., 2018; Huang et al., 2019). Inorganic As and the methylated trivalent arsenicals have also been shown to inhibit insulin-dependent glucose uptake and glucose metabolism by inhibiting the insulin activated PI3K-Akt pathway in cultured cells (Paul et al., 2007; Zhang et al., 2017). Other mechanisms by which iAs exposure may interfere with processes regulating glucose and insulin homeostasis have also been described. For example, exposure to iAs have been shown to alter DNA methylation and transcription of diabetes-associated genes in human cohorts exposed to iAs in drinking water (Smeester et al., 2011; Bailey et al., 2013). Notably, in some of these cohorts, iAs exposure has been linked to an increased risk of diabetes.

Growing evidence from human and mouse studies suggests that prenatal iAs exposure may contribute to the development of diabetic symptoms in adult offspring (Rager et al., 2014; Rojas et al., 2015; Huang et al., 2018a; Fry et al., 2019). However, the role of parental exposure to iAs prior to conception has not been systematically examined. We have recently shown that an isolated preconception exposure to iAs (200 ppb As in drinking water for 10 week prior to mating) resulted in a diabetic phenotype in adult C57BL/6 mice and that the diabetogenic effects of this exposure were sex-specific (Venkatratnam et al., 2021). We linked these effects to altered gene expression profiles in paternal sperm. Six diabetes-associated pathways that were enriched for differentially expressed genes were identified in the sperm of sires. Three of these pathways were also altered in livers of the offspring, namely focal adhesion, phosphoinositide 3 kinase/protein kinase B (PI3K-Akt) signaling, and extracellular matrix (ECM) receptor interaction. Because mice methylate iAs, producing MAs and DMAs metabolites that are distributed in tissues including testes (Currier et al., 2014), we were not able to determine which of the metabolites of iAs were responsible for the transcriptomic effects in the sires’ sperm. To address this research gap, the present study used ex vivo exposure model to characterize the As species-specific transcriptomic effects. We focused on the trivalent arsenicals because pentavalent arsenicals are much less toxic (Stýblo et al., 2000, 2021), and we used low concentrations that are more likely to be compatible with concentrations of As species after in vivo exposures. Indeed, the concentrations of total As in the sperm cells exposed to arsenicals at 0.1 μM level were only several times higher (for MAs^III and DMAs^III exposures) or even lower (for iAs^III exposure) than in the control sperm, which was isolated from mice drinking DIW and fed semi-purified diet that contains relatively low levels of As (Murko et al., 2018). Thus, the exposure levels used in this study are relevant to the environmental exposures.

Previous studies have shown that trivalent MAs^III and DMAs^III are more potent than trivalent iAs^III as cytotoxins, genotoxins, and inhibitors of enzymes and signaling pathways (Stýblo et al., 2021). MAs^III and DMAs^III were also more cytotoxic than iAs^III in the mouse sperm in the present study. However, the most robust changes in the sperm transcriptome were associated with the exposure to 0.1 μM iAs^III, which did not affect sperm viability. Somewhat weaker effects were observed in sperm cells exposed to 0.1 μM MAs^III or DMAs^III. Still, the exposures to 0.1 μM concentrations of the trivalent arsenicals resulted in more widespread changes in sperm transcriptome than 1 μM exposures. Thus, the extent of the effects did not correlate with the exposure level or the cytotoxicity associated with 1 μM MAs^III exposure. These results are consistent with published data. Lower, more biologically relevant in vitro As exposures have led to more robust transcriptomic dysregulations than higher, potentially cytotoxic exposures, also in other types of cells (Xu et al., 2013).

Our data suggest that the effects of ex vivo exposure on sperm transcriptome were mostly As species-specific. Multiple components of the transcription factor activator protein 1 (AP-1) complex, including Jun, Junb, Fos, and Egr1, were among the top downregulated genes in sperm exposed to 0.1 μM MAs^III, yet these genes were all moderately upregulated by exposures to the other arsenicals. The AP-1-regulated gene transcription plays important roles in cell proliferation, differentiation, and apoptosis (Shaulian and Karin, 2002). Previous studies have shown that iAs can directly bind to the AP-1 complex, altering the related gene expression at non-cytotoxic level.

Many of the other differentially expressed genes are linked to diabetes pathogenesis. Here are several examples:

1. Vgf, which was significantly upregulated in sperm exposed to 0.1 μM MAs^III, encodes for the prohormone VGF that facilitates granule formation (Fargali et al., 2014). In pancreatic beta cells, Vgf is involved in biogenesis of insulin secretory granule and proinsulin processing, indicating a potential role in insulin secretion (Stephens et al., 2017). Additionally, we have identified Vgf in several networks and pathways related to cellular response to stimulus and reproductive structure development, including response to cAMP, gonad formation and sex differentiation. Fabp2, which was upregulated in sperm exposed to 0.1 μM DMAs^III, encodes for intestinal fatty acid binding protein 2. This protein is involved in glucose metabolism and insulin secretion (Albala et al., 2004). Fabp2 polymorphisms, which affect dietary fat absorption, have been linked to susceptibility to diabetes, while transcriptional overexpression may contribute to insulin resistance (Weiss et al., 2002).

2. Rilp encodes the Rab7a interacting lysosomal protein, a downstream effector of protein Rab7 that facilitates endocytosis, retrograde trafficking and autophagy. A previously published study has suggested that overexpression of Rilp could result in restricted insulin secretion via clustering of insulin secretory granules, inducing lysosomal degradation of proinsulin in immature insulin granules in pancreatic beta cells (Zhou et al., 2019). In the present study, Rilp was among the most upregulated genes in sperm exposed to all three arsenicals at 0.1 μM exposure level.

3. Tnfaip3 encodes the zinc finger A20 protein induced by the tumor necrosis factor, and plays a critical role in the downstream termination of NF-κB responses through tumor necrosis factor receptor (TNFR) (Rimoin et al., 2013). Overexpression of this gene has been shown to suppress apoptosis of beta cells, suggesting a potential a role in both Type 1 and 2 diabetes (Cheng et al., 2016).

4. Kcnq1ot1, a long non-coding (Lnc) chromatin regulatory gene, is located on the opposite strand of Kcnq1, which encodes the voltage gated potassium channels that are involved in regulation of insulin secretion from pancreatic beta cells (Yamagata et al., 2011; Rojas et al., 2015). The transcription of Kcnq1ot1 greatly overlaps with Kcnq1 transcription unit on the anti-strand; thus, the transcription of Kcnq1 could be affected through transcriptional interference (Pandey et al., 2008). However, Kcnq1 transcription in the sperm was not dysregulated by ex vivo exposures to arsenicals in the present study or by in vivo iAs exposure in our published study (Venkatratnam et al., 2021). In addition, neither Kcnq1 nor Kcnq1ot1 transcription was altered in livers of offspring from parents exposed to iAs prior to conception (Venkatratnam et al., 2021). Thus, the association between Kcnq1ot1 dysregulation in the sperm and the diabetogenic effects of preconception exposure to iAs remains unclear.

One of the goals of this study was to compare the transcriptomic effects in the sperm after ex vivo exposures to individual arsenicals with those of in vivo iAs exposure. None of the genes that were differentially expressed in sperm of male mice exposed to 200 ppb iAs in drinking water, Ptpn22, CD248, Ccl8, Mvp, or Rab3d (Venkatratnam et al., 2021), was significantly altered by ex vivo exposure in this study. However, four of the six pathways that were enriched for differentially expressed gene by in vivo exposure were also identified as targets of the ex vivo exposures: (1) the Focal adhesion pathway was significantly enriched in sperm exposed to all three trivalent arsenicals at 0.1 μM level; (2) the ECM receptor interaction pathway was significantly enriched by exposure to 0.1 μM MAs^III; (3) the PI3K-Akt signaling pathway was enriched by exposure to 0.1 μM iAs^III, and (4) the Regulation of non-canonical Wingless-related integration site (Wnt) signaling pathway was significantly enriched in sperm exposed to 0.1 μM MAs^III.

Focal adhesions are structures formed at the cell-ECM contact points, consisting of actin filaments bundles anchored to trans-membrane receptors of the integrin family through a complex of proteins (Petit and Thiery, 2000). Mediated by transmembrane molecules like integrins, the ECM-receptor interaction pathway plays a significant role in modulating adhesion and migration activities, and multiple members of this pathway regulates PI3K-Akt-mTOR signaling cascade in its upstream (Lee and Juliano, 2004). Several significantly altered genes in sperm exposed to all three arsenicals at 0.1 μM level, including Fn1, Itga4, and Col4a4, encode for crucial components of this pathway.

Insulin is the main activator of the PI3K-Akt signaling pathway, regulating glucose metabolism in multiple organs including liver, brain and adipose tissue (White and Kahn, 1994; Huang et al., 2018b). We found multiple downstream targets of the PI3K-Akt signaling pathway to be significantly upregulated, including Cdkn1a (whose expression should be repressed by activation of Akt), Pkn1 and Kit. We found multiple members of growth factor ligands upstream of PI3K activation, namely Fgf8, Fgf10, and Efna5, to be significantly downregulated in sperm exposed to 0.1 μM iAs^III, yet Vegfa was moderately upregulated by 1.6-fold change. We have also found Lpar5 and Il2rg, one G-protein coupled receptor and one cytokine receptor that facilitates PI3K activation by cytokines, chemokines and other hormonal ligands, to be among the top upregulated genes. Overall, our data provides evidence about dysregulation of the PI3K-Akt pathway as a result of arsenic exposure, bur cannot address specific mechanism underlying this event.

The Wntsignaling pathway is involved in enhancement of insulin action and susceptibility to type 2 diabetes (Chen et al., 2021), With KEGG analysis, we were able to identify 6 proteins that negatively regulate the Wnt signaling pathway, and 8 proteins that positively regulate the pathway, including Nfkb1 and Tnfaip3. However, the directions of changes within either of these functional clusters were not uniform, making prediction of the general direction of change in the Wnt signaling pathways difficult.

Among the top protein-protein interaction clusters that were significantly enriched by ex vivo exposures were clusters centered around two transcription complexes, AP-1 and NF-kappaB. These transcription factors have been previously identified as targets of As exposure (Simeonova et al., 2001; Hu et al., 2002). Exposures to iAs^III and/or the trivalent methylated arsenicals, MAs^III and DMAs^III, have been shown to alter the phosphorylation, composition and/or DNA binding of these transcription complexes, leading to activation of gene transcription associated with oxidative stress and antioxidant defense, inflammation, or pathobiology of type 2 diabetes (Barchowsky et al., 1996; Kapahi et al., 2000; Roussel and Barchowsky, 2000; Drobná et al., 2003; Patel and Santani, 2009). In the present study, we observed a moderate transcriptional upregulation of Jun, Junb and Fos, the key components of the AP-1 complex, in sperm exposed to 0.1 μM iAs^III. This observation is consistent with published data linking ex vivo acute iAs exposure to elevated c-Jun and c-Fos protein levels in lung epithelial cells (Li et al., 2002). Unfortunately, levels or DNA binding activity of AP-1 or NF-kappaB could not be assessed due to the limited amount of sperm.

While the results from this study are supportive of effects on sperm cells, the study is not without limitations. These are as follows:

1. Because of a relatively short life span of freshly isolated sperm cells, we were able to employ only a short, 4-h exposure. Transcriptomic effects of longer exposures, which would be more consistent with the in vivo exposure paradigm, may result in different transcriptomic profiles, perhaps profiles that would better compare with those described in our in vivo study.

2. We chose to work with technical replicates prepared from sperm pooled from 20 mice, rather than biological replicates. This is because the amount of sperm we could collect from a single mouse (10–20 million cells) was not sufficient to cover all experimental conditions (i.e., 3 arsenic species at 2 concentrations, plus control), given that the minimal requirement for RNA-seq analysis is 3 million cells per single sample (i.e., 21 million for all conditions). Using biological replicates would better reflect variability among the mice.

3. This study used enrichment analysis that was able to provide information about pathways enriched for differentially expressed genes. Yet this method does not take into account directions of changes in gene transcripts, or provide mechanism-based insights for the cascade of signaling in most pathways. Additionally, we used GO Term analysis together with KEGG analysis, while the former might incorporate unconfirmed pathways and protein clusters based merely on regions of sequence similarity with annotated genes in model species (Stanford et al., 2020).

4. The protein-protein interaction functional clustering analysis was performed under the premise that differences in gene expression will be reflected in the abundance of their corresponding proteins, thus ignoring the role of transcript or protein stability and effects of posttranslational modification on the actual protein levels in the exposed sperm. In spite of these limitations, our study provides unique insight into the transcriptomic effects of As exposure in mouse sperm, while linking these effects to specific, biological relevant As species.

Taken together, we identified several diabetes-associated genes and pathways that were altered in sperm exposed ex vivo to trivalent arsenicals. The dysregulation of these genes/pathways may explain the diabetogenic effects of preconception iAs exposure described in our published study (Venkatratnam et al., 2021). Other genes, pathways and protein clusters with no obvious links to diabetes were also identified. Data on these genes/pathways may inform studies that examine adverse effects of As exposure on spermatogenesis, sperm morphology, viability, life span and storage, or sperm motility. Several studies focusing on these or similar endpoint have been previously published (Danielsson et al., 1984; Uckun et al., 2002; Das et al., 2009; Zubair et al., 2017; Anwar and Qureshi, 2019). Future studies will focus on potential epigenetic effects of As exposure in germ cells and on inheritance of these effects by offspring, thus, exploring a plausible mechanism for the adverse effects of preconception exposure to iAs, including the diabetogenic effects.

Data availability

We have shared all essential data presented in the manuscript in the supplemental material attached to this submission. Codes described in methodology and other data are available on request.