Hepatic inflammation after exposure to a mixture of low-dose arsenic and cadmium in a murine model of fatty liver disease

Abstract

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a growing global health burden and a major contributor to chronic liver disease. Evidence suggests arsenic (As) and/or cadmium (Cd) exposure may influence MASLD development, yet the effects of chronic low-dose metal mixtures on hepatic inflammation and immune responses remain unclear. Using the apolipoprotein E-knockout mouse model, we examined low-dose As and Cd exposure in male and female mice. We focused on hepatic steatosis and inflammation. Steatosis-related changes were assessed via lipid metabolism gene expression and PLIN2 levels. No significant changes were observed in males; however, females exposed to the metal mixture showed increased PLIN2 expression. In contrast, males exhibited an inflammatory phenotype following combined exposure. High-plex single-cell imaging (PhenoCycler) in male livers revealed increased Ki67+ hepatocytes, enhanced β-catenin signal, and elevated CD8+ T-cell infiltration, indicating enhanced proliferation and immune activation. These findings suggest sex-dependent responses to low-dose As and Cd, with females showing subtle steatotic changes and males a pronounced inflammatory signature. Collectively, combined metal exposure induces hepatic priming in both sexes; males show an increased inflammatory response with cellular proliferation in the absence of overt steatosis or fibrosis, whereas females demonstrate increased steatosis without inflammation or fibrosis.

The global prevalence of metabolic dysfunction-associated steatotic liver disease (MASLD), previously known as non-alcoholic fatty liver disease (NAFLD), continues to rise, making it the most common chronic liver disease worldwide (Schaffner and Thaler 1986; Le et al. 2022; Younossi et al. 2025). Currently affecting approximately 38% of the population, its prevalence is projected to increase to 55.4% by 2040 (Le et al. 2022; Paik et al. 2022; Younossi et al. 2023). MASLD is initiated by the accumulation of lipids within hepatocytes, resulting in steatosis and progressive liver dysfunction. Moreover, MASLD can progress to metabolic dysfunction-associated steatohepatitis (MASH), which is characterized by liver inflammation, oxidative stress, cellular damage, and fibrotic remodeling (Li et al. 2024a). It is now widely accepted that MASLD development and progression follow the “multiple hit hypothesis,” which proposes that a combination of metabolic, genetic, and environmental factors, including exposure to environmental metals, contributes to disease progression (Buzzetti et al. 2016; Li et al. 2023b).

Among the various factors implicated in MASLD progression, environmental exposure to metals, such as arsenic (As) and cadmium (Cd), has gained increasing attention. Both metals are commonly found in contaminated drinking water, as well as in soil, food, and air (Ryan et al. 2000). The World Health Organization (WHO) has set maximum containment levels in drinking water at 10 µg/l(10 ppb) for As and 3 µg/l (3 ppb) for Cd (Edition 2011). Importantly, previous studies in rodents have demonstrated that both As and Cd, when administered individually, can promote hepatic lipid accumulation (Tan et al. 2011; Zhu et al. 2022). However, these effects were observed at exposure levels significantly higher than the WHO guidelines. In addition to mechanistic studies, epidemiological evidence has linked elevated blood (Xu et al. 2025) or urinary levels (Xie et al. 2023) of these metals with increased risk of MASLD. Moreover, emerging evidence suggests that combined exposure to As and Cd is linked to increased risk of fatty liver disease (Xie et al. 2023) and liver dysfunction (Huang et al. 2021), underscoring the need to investigate the effect of metal mixtures in addition to individual exposures alone.

Although As and Cd share overlapping hepatic targets, their mechanisms of action are distinct. Arsenic primarily contributes to hepatic lipid dysregulation by promoting fatty acid synthesis, whereas Cd shows dual functionality, activating hepatic de novo lipogenesis (DNL) and impairing the mitochondrial fatty acid β-oxidation (FAO) pathway (Zhang et al. 2018; Wan et al. 2021; Fan et al. 2024). DNL is a key metabolic process involving the synthesis, elongation, and desaturation of fatty acids, followed by the formation of triglycerides (Ameer et al. 2014). The FAO pathway serves as a metabolic counterbalance to DNL. In the liver, sterol-regulatory element binding protein 1 (SREBP-1) is the main transcription factor regulating genes involved in lipogenesis, with the Srebp-1c isoform specifically driving fatty acid synthesis (Li et al. 2023a). Fatty acid binding protein-1 (FABP1) also plays a critical role in lipid uptake, transport, metabolism, and storage of lipids within hepatocytes (Wang et al. 2015). Lipid accumulation is largely determined by the abundance and size of intracellular lipid droplets, with perilipin-2 (PLIN2) serving as a major lipid droplet-associated protein and marker of hepatic lipid storage (Kimmel et al. 2010; Kimmel and Sztalryd 2016; Itabe et al. 2017). Moreover, hepatic lipid homeostasis is further regulated by sirtuin 1 (SIRT1), which suppresses DNL and enhances FAO (Zuo et al. 2025). Peroxisome proliferator-activated receptor alpha (PPARα) acts as a transcriptional regulator of FAO (Kersten 2014). Key enzymes involved in this pathway include carnitine palmitoyl transferase 1A (CPT1a) and medium-chain acyl-coenzyme A dehydrogenase (MCAD), both essential for efficient fatty acid catabolism (Houten and Wanders 2010) .

Beyond hepatic steatosis, oxidative stress is a critical driver of MASLD/MASH progression, with both As (Sumedha and Miltonprabu 2015; Fatemi et al. 2021; Adetutu et al. 2024) and Cd (Sanjeev et al. 2019; Gelen et al. 2023) known to induce hepatic oxidative damage (Martín-Fernández et al. 2022). The nuclear factor erythroid 2-related factor 2 (Nrf2)/heme-oxygenase-1 (HO-1) signaling pathway plays a pivotal role in modulating the cellular response to oxidative stress. HO-1, a downstream effector of Nrf2, contributes to hepatoprotection through its anti-inflammatory, antioxidant, and anti-apoptotic activities (Biswas et al. 2014).

In addition to lipid accumulation and oxidative stress, hepatic inflammation drives MASLD progression to MASH (Kim et al. 2025). As a key immunological organ, the liver rapidly recruits immune cells following injury, a process that supports tissue repair but can also perpetuate inflammation, leading to hepatic injury and fibrosis (Robinson et al. 2016; Huby and Gautier 2022). Recent single-cell transcriptomic studies highlight immune landscape remodeling in MASH, marked by increased CD4+ and CD8+ T cells during MASLD progression (Xiong et al. 2019; Remmerie et al. 2020; Li et al. 2024b; Ge et al. 2025). Among CD4+ T cell subsets, Th1 (Kremer et al. 2006) and Th17 cells drive inflammation (Rau et al. 2016; Moreno-Fernandez et al. 2021), whereas Th2 cells are linked to profibrogenic responses (Gieseck et al. 2018). CD8+ T cells are involved in multiple aspects of MASLD pathogenesis. Activated hepatic CD8+ T cells are elevated in patients with MASH (Wolf et al. 2014). Moreover, inhibition of CD8+ T cells attenuates hepatic inflammation in a murine model (Bhattacharjee et al. 2017). These cells contribute to the pro-inflammatory environment through enhanced production of cytokines and can exhibit cytotoxicity by releasing perforin and/or granzymes, leading to hepatocyte death (Dudek et al. 2021). However, CD8+ T cells may also exert context-dependent regulatory or protective effects (Koda et al. 2021).

Exposure to As or Cd activates hepatic inflammatory signaling pathways and disrupts immune balance. Both As (Yang et al. 2021; Mirzaei et al. 2023; Molavinia et al. 2023) and Cd (Liu et al. 2019; Hayat et al. 2024) exposure activate nuclear factor kappa B (NF-kB), a key transcriptional regulator of immune and inflammatory responses. NF-kB activation promotes the transcription of pro-inflammatory cytokines, chemokines, and adhesion molecules, facilitating leukocyte recruitment and sustaining hepatic inflammation (Pahl 1999; Luedde and Schwabe 2011). Consistent with this, elevated levels of pro-inflammatory cytokines and increased infiltration of inflammatory cells follow exposure to As or Cd (Dkhil et al. 2020; Li et al. 2021a; Zhang et al. 2021; Mondal et al. 2022; Wang et al. 2022; Daryagasht et al. 2023; Fang et al. 2023; Molavinia et al. 2023; Nikravesh et al. 2023). However, these studies used only hematoxylin and eosin (H&E) staining and did not define specific immune cell populations. Notably, one study found that chronic exposure to 250 ppb As significantly increased the number of CD45+ immune cells in the liver (Straub et al. 2007b). Similarly, Cd exposure in pubertal mice led to an increase in CD45+ immune cells, including both M1 pro-inflammatory and M2 anti-inflammatory macrophage populations (Li et al. 2021b). These findings suggest that exposure to As and Cd can alter the hepatic immune microenvironment, potentially contributing to the development and progression of MASLD.

Emerging evidence suggests that dysregulation of the Wnt/β-catenin signaling pathway contributes to MASLD/MASH progression (Shree Harini and Ezhilarasan 2023). β-catenin plays key roles in liver physiology and injury response (Monga 2015). It maintains tissue integrity via adherens junctions and acts as a central mediator of the Wnt signaling pathway, which regulates cell proliferation, differentiation, and development (Bienz 2005). The role of β-catenin/Wnt signaling in MASLD/MASH is complex and context-dependent, with studies highlighting its involvement in lipid metabolism, hepatocyte proliferation, inflammatory signaling, and fibrotic progression (Perugorria et al. 2019; Shree Harini and Ezhilarasan 2023).

Although evidence supports a role for As- or Cd-induced liver injury, 2 key knowledge gaps remain. First, most studies have focused on high-dose exposures, leaving a limited understanding of how chronic, low-dose levels of these metals, both individually and in combination, contribute to the development and progression of fatty liver disease. Second, although inflammation is a hallmark of MASH, the specific immune cell populations involved in metal-induced liver inflammation remain poorly characterized, particularly in the context of co-exposure. To address these gaps, we investigated the effects of low-dose exposure to As and Cd in the apolipoprotein E-knockout (ApoE^−/−) mouse model, with a particular focus on hepatic steatosis and inflammation. We specifically characterized immune cell populations using a multiplexed spatial imaging technique. We found that the As/Cd mixture induced hepatotoxicity specifically in male mice. Although no changes were observed in lipid regulation, there was a significant increase in hepatocyte proliferation, β-catenin signal and CD8+ T cell infiltration.

Materials and methods

In vivo arsenic and cadmium exposures

All mouse procedures were approved by the McGill Animal Care and Use Committee. Samples analyzed in the present study were obtained from a previously described cohort (Subramaniam et al. 2023), and an additional cohort of mice was used to validate the initial findings obtained with the PhenoCycler. Briefly, male and female ApoE^−/− mice (B6.129P2-apoE^tm1Unc/J, Cat No. 002052, Jackson Laboratory, California) were bred in-house and, starting at 5 week of age, were fed a purified diet with low arsenic/cadmium (AIN-76A, Cat No. D10001, Research Diets Inc). Mice received drinking water containing sodium arsenite (5 or 50 ppb arsenic), cadmium chloride (1.5 or 5 ppb cadmium), or a combination of both for 13 weeks (n = 8). Water was changed 3 times per week. At the endpoint, mice were euthanized with isoflurane, and livers were harvested for analyses.

Plasma analyses

Blood (0.6 ml) was collected by cardiac puncture, and plasma was obtained using collection tubes (BD Vacutainer SST, Cat No. 367986). Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were assessed by the pathology services at The Centre of Phenogenomics in Toronto, Ontario.

ICP-MS metal analysis

Metal analyses of hepatic As and Cd were performed at the Mass Spectrometry Core at Université de Montréal. Liver tissue samples (0.1 to 0.5 g) were digested in 50 ml trace metal–grade polyethylene tubes with 5 ml concentrated ultra-trace nitric acid (67% to 70%, PlasmaPURE Plus, SCP Science) and 0.2 ml hydrogen peroxide (30%, VWR BDH Chemicals). Samples were predigested at room temperature for ≥48 h and subsequently heated at 85°C for 4 h (DigiPrep; SCP Science). Digests were diluted to 20 ml with ultrapure water (Milli-Q; resistivity >18.2 MΩ·cm; total organic carbon < 2 µg/l) and further diluted 10-fold prior to analysis. Dogfish liver certified reference material (DOLT-5; National Research Council Canada) was processed alongside samples for quality assurance.

As and Cd were quantified using an inductively coupled plasma mass spectrometer (NexION 5000, PerkinElmer). Isotopes ⁷⁵As and ¹¹¹Cd were monitored. To minimize polyatomic interference (⁴⁰Ar³⁵Cl at m/z 75), oxygen was used as a reaction gas (0.7 ml/min), and As was measured in mass-shift mode (75 As→91AsO). Instrument conditions were as follows: RF power 1,600 W; plasma gas flow 16 l/min; auxiliary gas flow 1.2 l/min; nebulizer gas flow 0.96 l/min; integration time 1 s; 3 replicates per sample.

External calibration was performed using EPA 200.7 Calibration Standard 6 (20 mg/l; High Purity Standards), with working standards prepared in 2% (v/v) ultra-trace nitric acid. Quality control standards (QCS-27, 100 mg/l; High Purity Standards; and QCP-QC3, 10 mg/l; Inorganic Ventures) were analyzed periodically and were traceable to NIST. Yttrium and indium (20 µg/l each) were introduced online as internal standards to correct for instrumental drift and matrix effects. The limit of detection for analytes ranged from 0.0001 to 0.0002 $\mathrm{\mu }$g/g.

Hematoxylin and eosin staining

Liver tissues were fixed with 10% neutral buffered formalin for 24 to 48 h and embedded in paraffin. Sections were collected at 4 µm thickness. Briefly, sections were deparaffinized in xylene and rehydrated in ethanol (100%, 95%, 70%) and distilled water. They were stained with Hematoxylin (Ricca Chemical, Cat No. 3530-32) for 1 min and 30 s, rinsed in running tap water, and counterstained with Eosin Y solution (Abcam, Cat No. Ab246824) for 20 s. Slides were then dehydrated in ethanol, cleared in xylene, and mounted with permount mounting media (Fisher Chemical, Cat No. SP15-500).

Immunohistochemistry

FFPE liver sections (4 μm) were deparaffinized, rehydrated, and underwent heat-induced antigen retrieval using TRIS/EDTA buffer (pH 9.0). Following blocking, sections were incubated with a primary antibody against perilipin-2 (Novus Biologicals, Cat No. NB110-40877, polyclonal antibody, 1:600). Detection was performed using an HRP-conjugated secondary antibody (Agilent Technologies, DAKO EnVision+ System HRP labeled polymer anti-rabbit, Cat No. K4003) and visualized with 3,3′-diaminobenzidine (DAB). Slides were counterstained with hematoxylin, dehydrated, and mounted with Permount media. Images were acquired using a Zeiss Axioscan slide scanner. Quantification was performed in Qupath by manually annotating perilipin-2-positive regions and calculating the positive area fraction.

Masson trichrome

Liver sections were deparaffinized, rehydrated, and stained with Trichrome Stain kit (Abcam, Cat No. ab150686). Briefly, nuclei were stained with Weigert’s iron hematoxylin, followed by staining with Biebrich scarlet–acid fuchsin to label cytoplasm, and differentiation in phosphomolybdic/phosphotungstic acid solution prior to collagen staining with aniline blue. Sections were then dehydrated, cleared, and mounted. Collagen fibers were stained blue, cytoplasm and muscle fibers red, and nuclei black.

Gene expression

Total RNA was isolated from whole liver tissue using a FastRNA pro green kit (MP Biomedicals, Cat No. 6045050). RNA concentration and purity were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific). Complementary DNA (cDNA) was synthesized from 1 µg total RNA using the iScript cDNA synthesis kit (Bio-Rad, Cat No. 1708891). Gene expression was analyzed by real-time polymerase chain reaction (qPCR). The Applied Biosystems 7500 Fast RT-PCR system (Life Technologies) and GoTaqPCR Master Mix (Promega, Cat No. A6001) were used. QPCR primers were designed and purchased from Integrated DNA Technologies (Table S1) (He et al. 2019). Each reaction contained a final primer concentration of 4 µM. The thermal cycling conditions consisted of initial denaturation at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 15 s, annealing/extension at 60 °C for 20 s. Relative gene expression was calculated using the double delta Ct (ΔΔCt) method with 36B4 (housekeeping gene) as an internal control. All samples were run in technical triplicate.

In situ immunofluorescence

Liver sections were processed as described for immunohistochemistry. Following antigen retrieval, sections were blocked with normal 2% horse serum and incubated overnight at 4 °C with primary antibodies against CD4 (Thermo Fisher Scientific, Cat No. 14-9766-82, monoclonal antibody, 1:100, clone 4SM95), CD8 (Thermo Fisher Scientific, Cat No. 14-0808-82, monoclonal antibody, 1:100, clone 4SM15), CD45 (R&D Systems, Cat No. AF114, polyclonal antibody, 1:100) or ß-catenin (BioLegend, Cat No. 844602, monoclonal antibody, 1:50, clone CTNNB1). A separate set of sections was co-stained at room temperature for 1 h with primary antibodies against carbamoyl-phosphate synthetase 1 (CPS-1) (Abcam, Cat No. ab12907, 1:100, clone EPR7493-3) and Ki67 (Thermo Fisher Scientific, Cat No. 14-5698-82, 1:100, clone SolA15) following the same protocol. Following washes, tissue sections were incubated for 1 h at room temperature in the dark with appropriate fluorophore-conjugated secondary antibodies: Alexa Fluor 647 goat anti-rat IgG, donkey anti-goat IgG, Alexa Fluor 750 goat anti-rabbit, and/or Alex Fluor 647 goat anti-mouse (all from Invitrogen, Cat No. A-21247, A-21447, A-21039, and A21237, 1:500). Nuclei were counterstained with DAPI, and slides were mounted with Fluoromount-G (Invitrogen). Images were acquired using a Zeiss Axioscan slide scanner. CD4 and CD8 T cell quantification was performed using the threshold function, whereas $\beta$-catenin expression was analyzed through pixel-based quantification in QuPath. Additionally, co-localization of CPS-1 and Ki-67 was done using the classifier tool within QuPath.

PhenoCycler

FFPE liver sections (4 µm) were processed and stained using the PhenoCycler multiplexed imaging platform (Akoya Biosciences). The detailed staining protocol is described here (Abraham et al. 2024). Briefly, tissue sections were deparaffinized, rehydrated, and underwent antigen retrieval. After blocking, the samples were incubated with a cocktail of DNA-barcoded primary antibodies at 4 °C overnight (Table S2). The tissues were then loaded onto the PhenoCycler for iterative fluorescent imaging. Complementary fluorescent reporters (ATTO550, AF647, AF750) were sequentially hybridized to the DNA-barcoded antibodies, imaged, and then removed. This process was repeated in cycles.

Images were acquired using Akoya’s PhenoCycler software and analyzed using QuPath. For single-cell analysis, nuclear segmentation was carried out using the StarDist algorithm based on DAPI staining, enabling precise identification and delineation of individual cells (Bankhead et al. 2017). Marker expression was quantified on a per-cell basis. Cells were then classified into distinct phenotypic populations based on combinatorial marker expression profiles. The Classifier tool was used to define positivity thresholds for each marker. Raw mean fluorescence intensity values for each marker were exported and used to generate heatmaps using CytoMAP for visualization of expression patterns across single cells.

Statistical analysis

Statistical analyses were performed using GraphPad software using 1-way ANOVA, Kruskal-Wallis, or Brown-Forsythe and Welch ANOVA with a Dunnett’s, Dunn’s, or Dunnett’s T3 post hoc multiple comparison test. A 2-sided P < 0.05 was considered statistically significant. Statistical test is included in the figure legends where applicable.

Results

Characterization of hepatic tissue exposed to low concentrations of Cd and As

We previously profiled the atherosclerotic lesions from ApoE^−/− mice given low concentrations of Cd, As, or the combinations in their drinking water for 13 weeks (Subramaniam et al. 2023). In addition to cardiovascular disease, the ApoE^−/− mouse model is a suitable model for studying MASLD due to its hyperlipidemic profile (Lu et al. 2020). Therefore, we utilized these tissues to profile hepatic toxicity associated with low-dose exposures. The exposure doses used in this study were based on human data from the Danish Diet, Cancer and Health Cohort (DCH), one of the largest cohorts in the world with baseline urine and dietary assessment data (Tjønneland et al. 2007). Specifically, exposure levels were defined using the 25th and 95th percentiles of Cd and As concentrations, respectively, estimated from dietary intake in the DCH cohort. The 25th percentile corresponded to 1.5 ppb for Cd and 5 ppb for As, whereas the 95th percentile corresponded to 5 ppb for Cd and 50 ppb for As. This cohort provides comprehensive information on metal exposure, and importantly, the Cd distribution reflects non-smoking individuals. We focused on the lowest concentration of Cd (1.5 ppb), but the higher 50 ppb concentration of As, because As is metabolized significantly faster in mice than humans (Stýblo et al. 2019). Both male and female mice were included in the study to evaluate potential sex-specific effects.

To determine whether low-dose exposure to As and/or Cd induced gross pathological changes in ApoE^−/− mice, hematoxylin H&E staining was first performed. H&E analysis revealed no overt histopathological alterations in liver architecture across treatment groups in either sex (Fig. 1A). However, clear cytoplasmic vacuoles consistent with lipid droplet accumulation were evident, in line with the hyperlipidemic phenotype of ApoE^−/− mice. Notably, females appeared to have more lipid droplet accumulation compared with males. We next assessed hepatic connective tissue deposition by Masson’s trichrome staining, as deposition of extracellular matrix proteins such as collagens is important to the fibrotic process of MASH. Overall, very little collagen was detected in livers from these mice, regardless of treatment group. Collagen content was significantly decreased in females following exposure to 50 ppb As (P < 0.01) and the combined low-dose exposure (P < 0.05), whereas no changes were observed in males (Fig. 1B and C).

Fig. 1.

Hepatic histology, serum ALT, and metal accumulation in male and female ApoE^−/− mice following low-dose arsenic and cadmium exposure. (A) Representative hematoxylin and eosin (H&E)-stained liver sections showing preserved liver architecture across all groups. Cytoplasmic vacuolation consistent with lipid droplets is present in females. (B and C) Masson’s trichrome staining indicates minimal collagen deposition in all groups, with reduced staining observed in females following 50 ppb As and combined exposure. (D and E) Serum alanine aminotransferase (ALT) levels in males (D) and females (E). (F to I) Hepatic arsenic and cadmium concentrations measured by ICP-MS in males (F, H) and females (G, I). Hepatic Cd levels increased in males following combined exposure, and hepatic As levels increased following 50 ppb As exposure alone; no significant changes were detected in females. The limit of detection (LOD) ranged from 0.0001 to 0.0002 μg/g. Statistics were performed using Kruskal-Wallis test followed by Dunn’s post-hoc test (n = 7 to 8). Statistical significance is indicated as follows: *P < 0.05 and **P < 0.01. Scale bars represent 100 $\mathrm{\mu }$m (A) and 20 $\mathrm{\mu }$m (A, C).

We also measured serum levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), which, when elevated, are widely recognized as biomarkers of hepatocellular injury, with ALT being more liver-specific (Lala et al. 2025). Consistent with the lack of fibrotic changes, exposure to As or Cd alone, or in combination, did not significantly alter serum ALT (Fig. 1D and E) or AST levels in either sex (Fig. S1). Of note, several mice exposed to the combination of Cd and As exhibited increased serum ALT levels (Fig. 1D), suggesting that there is variability in the hepatic response. Importantly, we analyzed an extended set of concentrations utilized in our previous paper and again, found no significant changes in ALT or AST (Table S2).

Finally, we measured As and Cd concentrations in liver tissue by inductively coupled plasma mass spectrometry (ICP-MS), and these levels were found to be low. To our knowledge, data on combined As and Cd exposure at low levels and resulting hepatic metal burden remain limited. Interestingly, Cd levels were significantly increased in males following the combined exposure only (P < 0.05) (Fig. 1F). A previous study showed that exposure to As at ppb levels resulted in minimal hepatic accumulation in mice (Straub et al. 2007). However, in our study, hepatic As levels were increased in males exposed to 50 ppb As alone (P < 0.01) (Fig. 1H). In contrast, no changes in hepatic As or Cd levels were observed in females across any groups (Fig. 1G and I). Together, these data suggest that despite the hyperlipidemic phenotype and the presence of low levels of metal in the liver, the level of hepatic injury was not severe.

Analysis of hepatic lipid droplet and gene expression is largely unchanged by Cd, As, or the combination

Based on our observations from the H&E staining (Fig. 1A), we wanted to confirm whether there was an increase in lipid droplets following exposure to the combination of Cd and As, particularly in females. To further quantify hepatic steatosis, we stained liver sections for perilipin-2 (PLIN2), a protein that resides constitutively in the membranes of lipid droplets and is required for diet-induced hepatic steatosis, inflammation, and fibrosis in mice (Najt et al. 2016). Consistent with H&E observations, an increased PLIN2-positive fraction was observed in hepatic tissue from females following exposure to 50 ppb As (P < 0.05) and the low-dose combination (P < 0.01), whereas no changes were detected in males (Fig. 2A and B). These findings were consistent with the Plin2 mRNA expression changes. Although no significant changes in Plin2 mRNA expression were observed in males, the combination of As and Cd significantly increased Plin2 mRNA expression in females (P < 0.05) (Fig. 2C).

Fig. 2.

Females demonstrate a steatotic hepatic phenotype following low-dose arsenic (As) and/or cadmium (Cd) exposure in ApoE^−/− mice. (A and B) Quantification and representative images of perilipin-2 (PLIN2) immunohistochemistry (IHC) staining in liver sections. PLIN2-positive lipid droplets increased in females following exposure to 50 ppb As and combined exposure, with no changes detected in males. Scale bars represent 20 μm. (C) Hepatic Plin2 gene expression, showing increased expression in females following combined exposure, with no significant changes in males. (D) Hepatic expression of genes involved in de novo lipogenesis (Srebp1, Fabp1) and fatty acid β-oxidation (Sirt1, Pparα, Cpt1a, Mcad). No significant differences were observed across treatment groups or between sexes. Statistical analysis was performed using 1-way ANOVA with Dunnett’s multiple comparisons test; *P < 0.05 (n = 6 to 8).

To further characterize molecular changes associated with hepatic lipid accumulation, expression of genes involved in hepatic DNL or mitochondrial FAO pathway was measured. In contrast to Plin2 mRNA levels, no significant changes were observed in Srebp-1c, Fabp1, Sirt1, Pparα, Cpt1a, or Mcad mRNA (Fig. 2D) with any exposure or by sex. To further investigate the role of HO-1 in the liver after low-dose exposure to As and/or Cd, we measured Hmox1 gene expression. However, no statistically significant changes were observed in either sex (Fig. S2), suggesting minimal oxidative stress is induced by the exposures. Thus, although we observed a sex-specific increase in PLIN2, indicative of steatotic phenotype, no other changes were detected in genes involved in lipid metabolism or catabolism.

Co-exposure to As and Cd increased hepatic CD45+ immune cell population in males

Hepatic inflammation plays a central role in the progression from MASLD to MASH, and ultimately, to advanced liver diseases such as fibrosis, cirrhosis, and hepatocellular carcinoma. Previously, studies have shown increased hepatic CD45+ leukocytes after exposure to As (Straub et al. 2007b) or Cd (Li et al. 2021b). Here, we observed a significant increase in hepatic CD45 expression in male mice, specifically in the 1.5 ppb Cd and 50 ppb As exposure group (P < 0.001; Fig. 3A and B). No significant changes in CD45 expression were observed in female mice. Although this suggests a sex-specific inflammatory response, it does not give us an indication of what cell types are important.

Fig. 3.

Low-dose cadmium (1.5 ppb Cd) and moderate dose arsenic (50 ppb As) increased hepatic immune cell population in males. Representative immunofluorescent staining (A) and quantification (B) of CD45+ cells (magenta) and nuclei (DAPI, blue) in liver sections from male and female ApoE^−/− mice. Statistics were performed using a 1-way ANOVA with a Dunnett’s multiple comparisons test (n = 8). Statistical significance as follows: ****P < 0.001. Scale bars represent 100 μm (low magnification) and 20 μm (higher magnification).

To characterize immune cell populations within liver tissue in the ApoE^−/− mouse model, we performed a hypothesis-generating PhenoCycler (formerly known as Codex) experiment using liver tissue from male mice in either the control or the As/Cd co-exposure group (Abraham et al. 2024). We utilized this high-plex, single-cell spatial biology imaging platform using an antibody panel that included markers for cell lineage, major immune cell subsets, and signaling or functional proteins to assess cellular proliferation and tissue architecture (Fig. 4A; Table S3). A representative image from control male liver tissue with selected markers demonstrates successful multiplex staining and spatial resolution of key immune and structural markers (Fig. 4B).

Fig. 4.

Multiplex imaging and phenotypic analysis of mouse liver tissues exposed to Cd and As (n = 1). (A) Marker panel used for 16-plex PhenoCycler analysis of mouse liver, categorized into 4 groups: cell lineage, immune cell, structural, and other markers. (B) Representative image of control liver tissue stained with selected markers, including DAPI (nuclear stain, blue), Na⁺/K⁺ATPase (membrane marker, red), α-SMA (activated stellate cells or smooth muscle cells, green), CD45 (pan-leukocyte marker, yellow), and MPO (myeloid cells, white). (C) Heatmap showing relative mean fluorescence intensity (MFI) of individual markers across 8 identified cell phenotypes. (D) Pie charts depicting the relative abundance of cell phenotypes in control and 1.5 ppb Cd 50 ppb As-exposed liver tissues. (E) Composition of less abundant phenotypes across experimental groups. (F and G) PhenoCycler imaging analysis of control and co-exposed (1.5 ppb Cd 50 ppb As) liver tissue from male ApoE^−/− mice, showing increased Ki67⁺ proliferative cells (cyan) and CD8⁺ T cell infiltration (green). β-catenin (pink), Ki67 (cyan), and Na⁺/K⁺ ATPase (red) are also shown.

We quantified marker expression on a per-cell basis and classified cells into 8 distinct phenotypic populations based on combinatorial marker profiles. These phenotypes included CD3+ T cells (defined as CD45+ CD3+), CD4+ T cells (CD45+ CD3+ CD4+), CD8+ T cells (CD45+ CD3+ CD8+), CD31+ endothelial cells, α-SMA+ cells, Ki67+ proliferative cells (CD45-), other immune cells (CD45+), and an unclassified “other” category (Fig. 4C). We then quantified the relative proportions of these phenotypes in both control and As/Cdco-exposed tissues. Although this analysis was limited to a single sample per group, we confirmed our CD45+ staining (Fig. 3) by showing an increase in T cells as well as other CD45+ immune cells (Fig. 4D and E). Interestingly, in these 2 samples, there were increased numbers of CD4+, CD8+, and double-negative CD3+ T cells. To validate and extend our PhenoCycler findings, hepatic T cell populations were assessed by immunostaining for CD4 and CD8 in an expanded cohort of liver sections from mice exposed to Cd, As, or the combination (n = 8). No significant changes were observed in CD4+ T cell populations in either sex, although perhaps there is a trend towards increased numbers in the males (Fig. 5A). In contrast, a significant increase in CD8+ T cells were detected in males following exposure to 1.5 ppb Cd and 50 ppb As (Fig. 5B). Moreover, in the As/Cd-exposed liver, we also observed increased αSMA+ cells (Fig. 4D and E), which can be a marker for activated hepatic stellate cells, an early response to tissue injury, although additional data will be needed to support this interpretation (Kamm and McCommis 2022).

Fig. 5.

Increased CD8+ T cells in male ApoE^−/− mice after exposure to 1.5 ppb Cd 50 ppb As. (A) No changes in the CD4+ T cell population in both sexes. (B) Low-dose cadmium (1.5 ppb Cd) and moderate dose arsenic (50 ppb As) increased CD8+ T cells in males only. (C) Representative immunofluorescent staining shows CD4+ T cells (yellow), CD8+ T cells (green), and nuclei (DAPI, blue). Scale bar represents 50 μm. Statistical analyses were performed using a Brown-Forsythe and Welch ANOVA with a Dunnett’s T3 multiple comparisons test (n = 8). Statistical significance as follows: *P < 0.05.

Increased β-catenin within livers of mice co-exposed to Cd and As

Amongst the antigens detected in the PhenoCycler experiment were a number of structural and signaling molecules. We also observed abnormal localization and increased signal intensity of β-catenin in the As/Cd-exposed liver (Fig. 4F and G). Quantification of mean fluorescence intensity (MFI) per cell showed elevated β-catenin levels in the As/Cd group compared the control (Fig. 6A). In addition to β-catenin, we also quantified the mean MFI per cell for β-actin, cytokeratin, and vimentin (Fig. S3). Although there appeared to be modest changes in β-actin expression in the As/Cd-exposed tissue, further validation will be necessary to determine the significance of these observations. We then stained the extended cohort of liver tissues (n = 7 to 8) by immunofluorescence for β-catenin. β-catenin signal intensity was independently assessed by pixel quantification. Increased β-catenin signal was observed in males following exposure to 50 ppb As (P < 0.001) and in the combined exposure group (P < 0.01) (Fig. 6B and C). In females, no significant differences were detected, although the data is variable (Fig. 6B and C).

Fig. 6.

Increased hepatic β-catenin signal in male ApoE^−/− mice after exposure to 50 ppb arsenic (As) and a combination of As and cadmium (Cd). (A) PhenoCycler analysis showed a non-significant increase in β-catenin signal following exposure to 1.5 ppb Cd 50 ppb As; this analysis was based on mean fluorescence intensity (MFI) across all cells within the tissue from a single sample (n = 1). (B) In the validation cohort, β-catenin significantly increased in male ApoE^−/− mice exposed to 50 ppb As and the combination of Cd and As, whereas no significant changes were observed in females. (C) Representative immunofluorescence images of liver tissue are shown (β-catenin, magenta; DAPI, blue). Scale bars represent 20 μm. Statistical analysis was performed using 1-way ANOVA with Dunnett’s post hoc multiple comparisons test (n = 7 to 8). Statistical significance was defined as **P < 0.01 and ***P < 0.001.

Co-exposure of low concentrations of Cd and As generates a proliferative response in hepatocytes

Finally, we observed an increase in Ki67+ cells, a marker for proliferation, in our Phenocycler experiment (Fig. 4D and E). Although a hepatocyte-specific marker was not included in the antibody panel, the Ki67+ cells were largely CD45- and based on size, were likely hepatocytes (Fig. 7A). To confirm this, the extended cohort of liver tissues was co-stained with Ki67 and CPS-1, a hepatocyte-specific marker (n = 8). A significant increase in proliferative hepatocytes was observed in the combined exposure group, only in males (Fig. 7B and C). These data indicate that low dose Cd/As combination exposure generates a proliferative phenotype in hepatocytes, which may be a response increased hepatic injury.

Fig. 7.

Increased hepatocyte proliferation in male ApoE^−/− mice following exposure to low-dose Cd (1.5 ppb) and moderate dose As (50 ppb). (A) PhenoCycler analysis showed an increase in Ki67+ cells (cyan), consistent with hepatocyte origin based on increased cell size (n = 1). Representative immunofluorescence images (B) and quantification (C) of hepatocytes (CPS1, green), proliferating cells (Ki67+, red), and nuclei (DAPI, blue) in liver sections from male and female ApoE^−/− mice in the validation cohort. Scale bars represent 50 μm (low magnification) and 10 and 20 μm (higher magnification). Statistics were performed using a Kruskal-Wallis test with a Dunn’s post hoc multiple comparisons test (n = 8). Statistical significance as follows: *P < 0.05.

Discussion

We investigated the hepatic effects of chronic low-dose exposure to a combination of 1.5 ppb Cd and 50 ppb As in the ApoE^−/− mouse model, focusing on sex-specific responses. Chronic low-dose co-exposure to Cd (1.5 ppb) and As (50 ppb) in ApoE^−/− mice did not induce a MASLD-like phenotype as initially hypothesized. No significant changes in circulating liver enzymes were observed in either sex. In males, co-exposure increased hepatic Cd accumulation and elicited a proliferative and inflammatory response, characterized by elevated Ki67+ hepatocytes, β-catenin signal, and CD8+ T cell infiltration, without changes in lipid metabolism gene expression or fibrosis. In contrast, females showed increased PLIN2 expression at both the gene and protein levels, suggestive of early lipid accumulation, along with decreased collagen levels and no evidence of inflammation. These findings demonstrate a sex-specific divergence in hepatic responses, with a primed, inflammation-associated phenotype in males and a lipid-associated, non-fibrotic response in females.

Our findings provide important mechanistic insight and align with emerging evidence from human and mouse model studies. Transcriptomic analyses of liver biopsies from patients with MASLD and MASH have identified increased hepatic immune cell infiltration, with T cells representing the most abundant immune population (Bai et al. 2023). Similarly, we observed increased CD8+ T cell infiltration in male mice following metal exposure, suggesting a shared immune-mediated mechanism of liver injury, even in the absence of overt steatosis. It is important to note, however, that these mice exhibit an inherently hyperlipidemic phenotype due to their ApoE^−/− background. Although we have not validated this in our extended cohort, it would be interesting to further investigate the increase in DN T cells observed in the Phenocycler data. DN T cells are increased in mice with diet-induced MAFLD (Li et al. 2022). However, depending upon TCR expression (αβ vs γδ), the effects may be to enhance liver inflammation or inhibit immunosuppression, both of which would be predicted to enhance disease.

The increase in Ki67+ hepatocytes may reflect compensatory proliferation following CD8+ T cell-mediated hepatocyte injury triggered by metal-induced stress, as both increased Ki67+ expression and CD8+ infiltration were male-specific and observed only in the combination exposure. Under homeostatic conditions, hepatocyte turnover is minimal, with an average hepatocyte half-life of approximately 100 d and a daily renewal rate of about 0.5% to 1% in mice (Magami et al. 2002). In contrast, hepatocyte proliferation rises significantly in response to liver injury. For example, a recent study using Alb-CreERT2 Rosa26^rbw mice reported a substantial rise in Ki67+ hepatocytes, up to 7.5% to 8%, following chronic carbon tetrachloride (CCl₄) exposure (0.25 µl/kg every 3 d for 4 wk), with levels returning to baseline during the 2- and 4-wk recovery periods (Ruz-Maldonado et al. 2024). In the same study, partial hepatectomy (Phx) induced a rapid regenerative response, peaking at about 15% Ki67+ hepatocytes within 24 to 48 h, and declining to about 4% by 72 h. Notably, a comparable staining method was used, allowing direct comparison with our results. In our model, after 13 weeks of chronic As and Cd exposure, we observed an average of approximately 2% Ki67+ hepatocytes in males. Although this proliferation rate is lower than that induced by Phx (acute injury) or CCl₄ (potent chronic injury), it is markedly above basal levels, indicating a proliferative response. Taken together, these data suggest that prolonged low-dose As and Cd exposure imposes a chronic proliferative burden on male hepatic tissue.

β-catenin plays key roles in liver physiology and injury response (Monga 2015). It maintains tissue integrity via adherens junctions and acts as a central mediator of the Wnt signaling pathway, which regulates cell proliferation, differentiation, and development (Bienz 2005). In our study, β-catenin signal intensity increased following co-exposure, suggesting dysregulation and potential hepatic injury. However, the absence of clear nuclear localization in histological analyses suggests that this may reflect altered membrane-associated β-catenin dynamics, potentially indicative of disruption of the cadherin–catenin complex or loss of junctional integrity in response to cellular stress (Krutsenko et al. 2021).

Our results indicate that the combination of low doses of As and Cd promotes inflammation, Ki67+ expression, and β-catenin increases more than either metal alone. Although several studies have reported exposure to As or Cd-induced hepatic inflammation, to our knowledge, few have investigated the effects of metal mixture-induced hepatic toxicity. One study in male and female Swiss albino mice reported that exposure to varying concentrations of a metal mixture, including As, Cd, lead, chromium, mercury, iron, manganese, and nickel, via drinking water for 8-wk led to elevated hepatic expression of inflammatory markers, such as tumor necrosis factor-α and interleukin-6 (Singh et al. 2025). In this study, the lowest concentrations of As and Cd were 380 and 98 ppb, respectively, thus considerably higher than reported herein. Notably, significant (P < 0.05) inflammatory responses were only observed at 10× and higher concentrations, suggesting that hepatic inflammation occurred only at substantially elevated exposure levels. Additionally, a separate study in Sprague-Dawley rats demonstrated that intraperitoneal administration of As (5.55 mg/kg) and Cd (0.44 mg/kg) for 14 d resulted in increased hepatic expression of pro-inflammatory cytokines, interleukin-1β and NF-kΒ (Cengiz, Gür et al. 2023) Similarly, our findings indicate that low-dose co-exposure to As and Cd can elicit an enhanced inflammatory response in the liver, even at concentrations significantly lower than those used in prior studies. Importantly, although previous work primarily assessed hepatic inflammation through cytokine and/or protein expression analyses, we characterized immune cell populations, providing deeper insight into the cellular mechanisms underlying metal-induced liver inflammation.

Previous studies in both human populations and animal models have demonstrated sex-specific susceptibility to As or Cd-induced fatty liver disease. Data from the 2016-2017 Korean National Health and Nutrition Examination Survey (NHANES) revealed positive associations with blood Cd levels and liver steatosis and fibrosis, with stronger associations observed in women compared with men (Chung et al. 2020). In contrast, a separate analysis of U.S. NHANES data found a slightly stronger association between blood Cd levels and MASLD/MASH endpoints in men (Hyder et al. 2013). These discrepant findings may be explained by population-level differences in age distributions between the cohorts, as blood Cd levels are known to increase with age (Park et al. 2021). Additionally, menopausal status may contribute to sex-specific differences, as both progesterone and estrogen can modulate Cd-mediated toxicity through regulation of metallothionein expression and metal-protein complex formation (Sogawa et al. 2001; Shimada et al. 2012). In support of sex-specific hepatic responses to metal exposure, a zebrafish model was used to examine liver protein expression following a 7-d exposure to 50 ppb As in both sexes (Carlson et al. 2013). Proteomic analyses identified 25 hepatic proteins differentially expressed, revealing sex-dependent pathway activation: males exhibited enrichment of fibrosis-related pathways, whereas females showed activation of pathways associated with hepatic lipid accumulation.

Sex-specific hepatic outcomes following metal exposure may, in part, be attributed to differences in metal retention and metabolism between males and females. In rats, males are more susceptible to As-induced liver damage compared with females, a difference linked to sex-dependent As biotransformation (Muhetaer et al. 2022). Males tend to have higher hepatic levels of monomethylarsonic acid (MMA), a more toxic intermediate, whereas females tend to have lower MMA levels and higher levels of dimethylarsinic acid (DMA), a less toxic metabolite. Consistent with this, female C57BL/6J mice exposed to 100 ppb As in drinking water show higher urinary total As and a greater proportion of DMA compared with males under varying folate diets (Huang et al. 2018). Here, we reported increased hepatic As concentration in males after exposure to 50 ppb As. Population studies similarly indicate more efficient As methylation in women, reflected by increased urinary DMA proportions (Lindberg et al. 2007, 2008a, 2008b). Sex differences have also been reported for Cd. Chronic exposure to 32 ppm Cd in drinking water for 11 wk led to significantly higher hepatic Cd accumulation in female mice compared with males (Yamanobe et al. 2015). In humans, women generally exhibit higher blood and urinary Cd levels than men, potentially due to greater gastrointestinal absorption or lower iron stores (Gade et al. 2021).

Together, these findings suggest that sex-dependent differences in As metabolism and Cd burden may contribute to divergent hepatic responses following exposure to individual metals. However, metal accumulation and toxicokinetics may also be modified in the context of co-exposure, where interactions between elements can alter absorption, distribution, or retention profiles. In line with this, whereas prior studies report greater Cd accumulation in females, we observed increased hepatic Cd levels following combined Cd and As exposure only in males, suggesting that co-exposure may shift metal retention dynamics in a sex-specific manner and contribute to the divergent hepatic phenotypes observed.

Despite the valuable insights gained, several limitations should be acknowledged. First, although the ApoE^−/− mouse model is well-established for investigating lipid metabolism and vascular inflammation, its inherent dyslipidemia may influence hepatic responses, potentially limiting the generalizability of our findings to other models or human populations. The dyslipidemia may also mask potentially subtle metal-induced changes to lipid-associated gene expression. Additionally, mice metabolize inorganic As more rapidly than humans, primarily due to species-specific differences in methylation capacity (Vahter 1999; Koller et al. 2020). Therefore, future studies utilizing the humanized arsenic-3-methyltransferase (AS3MT) mouse model would better mimic human As metabolism and provide a more accurate assessment of As-induced hepatotoxicity (Koller et al. 2020). Future work should include extended exposure periods at low doses, as well as expanded immune and tissue profiling using comprehensive liver PhenoCycler panels, which can provide spatially resolved data on cell populations and tissue architecture. This spatial information will be critical for understanding how immune and parenchymal cells interact within the hepatic microenvironment in response to chronic low-dose metal exposure.

Our objective was to investigate the effects of chronic low-dose exposure to a mixture of As and Cd on fatty liver disease in both male and female mice. In a previous study, we examined the pro-atherogenic effects of the same metal doses-derived from the Danish Diet Cancer and Health (DCH) Cohort-using the ApoE^−/− mouse model (Subramaniam et al. 2023). Consistent with the human epidemiological data, that study revealed minimal significant changes, suggesting limited impact on atherosclerotic-mediated cardiovascular risk at these exposure levels (Poulsen et al. 2021; Sears et al. 2021). In contrast, our current findings indicate that the same low-dose exposure may promote hepatic inflammation, highlighting a potentially tissue-specific response that requires further investigation of metal mixtures and MASLD in prospective longitudinal cohorts such as the DCH.

Supplementary material

Supplementary material is available at Toxicological Sciences online.

Funding

This work was supported by NIEHS VICTER grant (R01ES030938) and NIEHS funded Columbia University Northern Plains Superfund program (P42ES033719).

Conflicts of interest: None declared.