Arsenic and diabetes mellitus: a putative role for the immune system

Abstract

Diabetes mellitus (DM) is an enormous public health issue worldwide. Recent data suggest that chronic arsenic exposure is linked to the risk of developing type 1 and type 2 DM, albeit the underlying mechanisms are unclear. This review discusses the role of the immune system as a link to possibly explain some of the mechanisms of developing T1DM or T2DM associated with arsenic exposure in humans, animal models, and in vitro studies. The rationale for the hypothesis includes: (1) Arsenic is a well-recognized modulator of the immune system; (2) arsenic exposures are associated with increased risk of DM; and (3) dysregulation of the immune system is one of the hallmarks in the pathogenesis of both T1DM and T2DM. A better understanding of DM in association with immune dysregulation and arsenic exposures may help to understand how environmental exposures modulate the immune system and how these effects may impact the manifestation of disease.

KEYWORDS:

• Arsenic
• diabetes mellitus
• immune system
• oxidative stress
• cytokines

Introduction

Diabetes mellitus (DM) is a chronic, metabolic disease characterized by elevated levels of blood glucose, which can lead to serious damage to the heart, blood vessels, eyes (vessels of retina), kidneys and peripheral nerves over time. DM is a public health burden affecting 422 million people worldwide, and 1.5 million deaths are directly attributed to DM each year (WHO 2022). In all ages of the US population, there is an upward trend in the diagnosis of DM that has increased from 10.3% in 2001 to 13.2% in 2020, with an additional 38.0% of US adults being prediabetic (CDC 2022).

There are two major forms of DM, type 1 (T1DM) and type 2 (T2DM). T1DM, also referred to as insulin-dependent diabetes or juvenile diabetes, is characterized by the absence of or diminished production and release of insulin, resulting from autoimmune destruction of the islet beta-cells by autoantibodies and dysregulation of immune cells (Yoon and Jun 2005). So T1DM can be defined as a hypoinsulinemic and hyperglycemic state (Figure 1). T2DM, also known as non-insulin-dependent diabetes, is a hyperinsulinemic state that occurs initially to compensate for insulin resistance in muscle, liver, and adipose tissues due to obesity, inflammatory responses and endoplasmic reticulum stress that alter insulin sensitivity (da Silva Rosa et al. 2020); but over time insulin levels are reduced due to exhaustion of islet beta-cells (Muoio and Newgard 2008), while the hyperglycemic state maintains (Figure 1).

Figure 1. The process of development of DM. T1DM is characterized as a hypoinsulinemic and hyperglycemic state resulting from the autoimmune destruction of the islet beta-cells. T2DM is a hyperinsulinemic state initially to compensate insulin resistance in muscle, liver, and adipose tissues due to inflammatory responses or altered insulin receptor signaling; but over time insulin levels are reduced due to dysfunction of islet beta-cells, i.e. hypoinsulinemia. Hyperglycemic state remains in T2DM.

Despite significant advances in blood glucose management by insulin therapy, patients with DM remain at risk for developing lifetime complications. DM is considered to be the fifth-leading cause of death worldwide due to associated comorbidities, and it is one of the most costly diseases requiring life-long chronic treatment (Grandl and Wolfrum 2017). DM may be also associated with the COVID-19 pandemic (also known as severe acute respiratory syndrome (SARS)-CoV-2), which began at the end of 2019 and spread extremely quickly around the world. Recent data have shown that DM, cardiovascular diseases and chronic obstructive pulmonary diseases are the most prevalent comorbidities in COVID-19 patients (Sanyaolu et al. 2020). Individuals with DM, hypertension, and severe obesity are more likely to be infected and at a higher risk for progression, unfavorable prognosis, and death from COVID-19 (Chang et al. 2020; Guan, Ni, et al. 2020; Guan, Liang, et al. 2020; Muniyappa and Gubbi 2020; Onder et al. 2020; Yang et al. 2020). Additionally, there is a similar increased risk for severe acute respiratory syndrome (SARS) or Middle East Respiratory Syndrome (MERS) in individuals with DM (Muniyappa and Gubbi 2020).

The specific etiology of DM is unclear, but important and emerging risk factors include genetics, age, obesity, environment, and lifestyle. Recently, growing evidence supports the role of environmental contaminants such as pesticides, heavy metals and endocrine disruptors as possible risk factors in the development of T2DM (Chen et al. 2009; Alonso-Magdalena et al. 2011; Evangelou et al. 2016). Additionally, many environmental risk factors might also trigger autoimmune mechanisms possibly critical in the etiology of T1DM (Butalia et al. 2016; Rewers and Ludvigsson 2016). As such, it is particularly important to identify modifiable environmental risk factors for DM and to develop primary prevention strategies.

Arsenic is an environmental heavy metal with ubiquitous human exposure that is present in air, water and soil. Arsenic is biphasic in terms of its role both as a carcinogen and as a therapeutic agent depending upon concentration in a cellular milieu (Prajapati et al. 2011). Arsenic exposure has been linked to various pathological conditions including DM, many types of cancer (lung, kidney, liver, and bladder cancers), cardiovascular disease (hypertension and atherosclerosis), neurological disorders, gastrointestinal disturbances, liver disease, renal disease, reproductive health effects, and dermal changes (Jomova et al. 2011). In this review, we separately summarize the association of arsenic exposure and the risk of T1DM and T2DM, as well as mechanisms involving the immune system.

Inflammation and DM

T1DM is characterized by the chronic inflammation state of the pancreatic islets, in which the insulin-producing beta-cells are destroyed by auto reactive T-cells, inflammatory cytokines and monocytes (Grunnet and Mandrup-Poulsen 2011). A number of proinflammatory cytokines, such as interleukin (IL)-1β, tumor necrosis factor (TNF)-α, IL17, IL21 and IL23, have all been demonstrated to be important for T1DM development (Clark et al. 2017), and are upregulated significantly in T1DM patients compared to age-matched healthy control groups (Fatima et al. 2016).

T2DM is a chronic, low-grade inflammatory disease characterized by a long-standing imbalance of the immune system, metabolic syndrome, or excess nutrients associated with obesity (Dasu et al. 2007; Dasu et al. 2008; Shu et al. 2012). Cytokines produced by immunocompetent cells influence local or systemic inflammation, which are critical contributors to the development of insulin resistance and T2DM (Ballak et al. 2015). The IL1 family can affect the initiation and progression of insulin resistance by the upregulation of proinflammatory cytokines, such as IL1β and TNF-α, which can directly interfere with insulin signaling in adipocytes, hepatocytes, fibroblasts and myocytes (Tseng 2004; Koenen et al. 2011; Ballak et al. 2015). Besides, IL1β can inhibit beta-cell function and destroy beta-cell mass (Koenen et al. 2011). Insulin signaling is initiated by binding of insulin to the insulin receptors, then the receptor activates insulin receptor substrates (IRS-1 and IRS-2) through tyrosine phosphorylation. IRS are adaptor molecules in the insulin signaling pathway and dysregulation of IRS proteins may lead to glucose intolerance and insulin resistance (Kaleem et al. 2021). It is proposed that IL6 can also induce insulin resistance in adipocytes (Lagathu et al. 2003; Rotter et al. 2003) by decreasing expression of the beta-subunit of the insulin receptor, insulin receptor substrate 1 (IRS1), and the glucose transporter GLUT4, thereby reducing glucose transport and interfering with insulin receptor signaling (Muoio and Newgard 2008; Grunnet and Mandrup-Poulsen 2011). TNF-α has been found to decrease insulin sensitivity by increasing serine phosphorylation of IRS1and IRS2, reducing the docking of IRS to the insulin receptor and interfering with downstream signaling pathways (Le Roith and Zick 2001). Reactive oxygen species (ROS) resulting from hyperglycemia can induce insulin resistance, impair insulin synthesis and secretion, through its interactions with cytokines and other mediators (Tseng 2004; Yan 2014). A free fatty acid-bound glycoprotein, fetuin-A, has been shown to be an endogenous ligand for toll-like receptor 4 (TLR4) and can promote lipid-induced insulin resistance through TLR4–NF-κB pathway to produce proinflammatory cytokines (Pal et al. 2012).

These findings reveal that several pathogenic cascades of both T1DM and T2DM are mediated by inflammation and proinflammatory cytokines. In addition, hyperglycemia can cause dysfunction of the immune response and vice versa, which increases the incidence of infectious diseases and related comorbidities in patients with DM (Berbudi et al. 2020). It has been found that DM-associated inflammation can enhance the progression and be detrimental to the prognosis of patients with COVID-19 (Guo et al. 2020). In diabetic patients with COVID-19, serum levels of inflammation-related biomarkers such as IL6, C-reactive protein (CRP), serum ferritin and coagulation index, D-dimer, were all significantly higher than those in non-diabetic patients with COVID-19, indicating that patients with DM are more predisposed to form a hyperinflammatory state ultimately leading to rapid deterioration of COVID-19-infected patients (Guo et al. 2020). Also, in patients with DM, compromised innate immunity, pro-inflammatory cytokine milieu, reduced expression of coronavirus surface receptor angiotensin-converting enzyme 2 (ACE2) and use of renin-angiotensin-aldosterone system antagonists exacerbate the severity of COVID-19 (Pal and Bhadada 2020). All of the above reveal the close relationship between inflammatory responses and DM.

Arsenic exposure and development of DM

Arsenic characteristics

Arsenic is a ubiquitous environmental contaminant naturally occurring in organic and inorganic forms. Inorganic arsenic (iAs) is the predominant form in surface and groundwater reservoirs and much more toxic than organic arsenic (Sung et al. 2015). The most common trivalent iAs compounds are sodium arsenite, arsenic trioxide, and arsenic trichloride; whereas arsenic pentoxide, arsenic acid and arsenates are the most common pentavalent compounds. Arsenic can also occur in organic forms, such as arsanilic acid, dimethylarsinic acid, arsenobetaine, and methylarsonic acid (Straif et al. 2009).

Arsenic exposure in humans

People are widely exposed to arsenic compounds through contaminated drinking-water and food, industrial processes (alloying agent, glass, pigments, textiles, paper, metal adhesives, wood preservatives and ammunition, to a limited extent in pesticides and pharmaceuticals) and smoking (Figure 2). The range of total arsenic concentrations in natural water is very large, from less than 0.5 µg/L to even more than 5000 µg/L, in which iAs (III) and iAs (V) are generally the dominant species, while organic forms of arsenic are minor (Smedley and Kinniburgh 2002). The WHO recommended limit of arsenic is 10 µg/L; however, high groundwater levels of arsenic (100 µg/L, predominantly iAs) affects over 200 million people worldwide, including areas of Bangladesh, India, Pakistan, China, the USA, Canada, Brazil, Vietnam, Indonesia, Hungary, and Mexico (Shahid et al. 2018). Small amounts of iAs are also consumed in the diet. Dietary intake of iAs in the US has been estimated to range from 1 to 20 µg/day (ATSDR 2007). Normal levels of arsenic in unexposed humans are: < 1µg/L in blood, ≤ 1 mg/L in nails and ≤ 1 mg/L in hair (ATSDR 2007). The Agency for Toxic Substances and Disease Registry (ATSDR) in the US has estimated minimal risk levels (MRLs) of iAs 5 µg/kg/day and 0.3 µg/kg/day, respectively for acute-duration (14 days or less) and chronic-duration (365 days or more) oral exposure to iAs (ATSDR 2007).

Figure 2. Main sources of exposure to arsenic for people. People can be exposed to arsenic through contaminated drinking-water and food, occupational exposures, and smoking tobacco.

‘Dual-sided personality’ of arsenic in health and disease

Although the ratio of arsenic species exposure varies greatly in different places, globally, chronic exposure to iAs (especially sodium arsenite and arsenic trioxide) from drinking water has been reported to be associated with most health risks, such as cancers, developmental effects, diabetes, pulmonary and cardiovascular disease (Goldman 2020).

In addition to its harmful effects, an arsenic derivative, arsenic trioxide (As₂O₃) has also been used in the treatment of cancers, such as acute promyelocytic leukemia, usually at the dose of 10 mg per day (0.06–0.2 mg per kg body weight/day), which contributes significantly to improving survival rates exceeding 90–95% (Bachleitner-Hofmann et al. 2002; Rojewski et al. 2004; Fortier et al. 2020; Ferrara et al. 2021). As₂O₃ has also shown activity in solid tumors, multiple myeloma, myelodysplasia syndrome, high-grade neuroepithelial tumor of the central nervous system, breast cancer, pancreatic ductal adenocarcinoma and some autoimmune diseases, similar to other chemotherapeutic agents (Baj et al. 2002; Goel et al. 2012; Kritharis et al. 2013; Lang et al. 2016; Paret et al. 2017). In addition, sodium meta-arsenite has been reported to have anti-tumor effects in human prostate cancer and is a candidate for the treatment of refractory non-Hodgkin’s lymphoma, as well as salvage chemotherapy in refractory biliary tract cancers (Jo et al. 2019). The anticancer activity of arsenic is partly dependent on the production of ROS to inhibit cancer cell migration, shorten telomeres, and enhance cell death (Li et al. 2006; Liu et al. 2011). Besides, based on its immunosuppressive effects, As₂O₃ might be used as an anti-rejection agent in allotransplant for cardiac disease or T1DM. In a mouse heart transplantation model and islet xenotransplantation, arsenic trioxide was reported to alleviate allograft rejection and prolong xenograft survival in mice (Yan et al. 2009; Zhao et al. 2018). However, as As₂O₃ is also a well-known toxicant and carcinogen because of its pro-inflammatory effects, the key point lies in balancing the pros and cons of accurate As₂O₃ applications (Zhang et al. 2018). Research on the ‘dual-sided personality’ of As₂O₃ can help further elucidate its therapeutic and toxic effects for appropriate use in curing diseases.

Arsenic exposure and the risk of DM

There is existing epidemiological evidence that chronic arsenic exposure, especially iAs and methylated metabolites, monomethylated arsenic (MMA) and dimethylated arsenic (DMA), is associated with DM, primarily T2DM (Navas-Acien et al. 2008; Maull et al. 2012; Bahadar et al. 2014; Currier et al. 2014; Huang et al. 2014; Brown 2015; Martin et al. 2015). Numerous epidemiological studies show a strong diabetogenic effect of arsenic in human populations globally, such as Pakistan (Bahadar et al. 2014), Mexico (Currier et al. 2014; Martin et al. 2015), Cambodia (Huang et al. 2014), Chile (Castriota et al. 2018), Bangladesh (Paul et al. 2019) and Canada (Brown 2015). People in these regions, regardless of if there is low or high iAs water concentrations, have a higher risk of developing T2DM compared to unexposed populations, but the risk exhibits a dose- and time-response pattern associated with the increase of iAs concentrations in the water and long-term exposure times (Martinez-Castillo et al. 2021). A National Toxicology Program (NTP) workshop concluded that arsenic is linked to an increased risk of T2DM in populations exposed to high levels of arsenic (≥ 150 µg/L arsenic in drinking water), including iAs, MMA and DMA (Maull et al. 2012). Moreover, investigators have also reported that long-term exposure to low-level arsenic (< 10 µg/L arsenic in drinking water) also contributes to the development of T2DM (Bahadar et al. 2014). A study from rural Bangladesh explored the dose-dependent association of drinking water containing low-to-moderate levels of arsenic (50.01–150 µg/L) confers a greater risk of hyperglycemia than safe drinking water (As ≤ 10 µg/L) (Paul et al. 2019). In the 2003–2004 National Health and Nutrition Examination Survey (NHANES), 788 adults were studied and found that participants with T2DM had a 26% higher level of total arsenic in their urine than those without T2DM (Navas-Acien et al. 2008). Similarly, researchers pointed out that workers from high arsenic exposure occupations, such as copper smelters and the art glass industry, had a higher prevalence of DM in long-term follow-up studies (Rahman and Axelson 1995; Rahman et al. 1996). A meta-analysis of dose–response indicated that T2DM risk increased by 13% for every 100 µg/L increment of iAs in drinking water (Wang et al. 2014).

In addition, particular subpopulations especially children may be at increased risk for health effects resulting from arsenic exposure, due to altered metabolism or underlying genetic predispositions, life stages, and cultural dietary factors (Lai et al. 2015). The fetus and young children may be at particular risk as a result of developmental windows of unique sensitivity to arsenic toxicity (Miller et al. 2002). In a SEARCH for Diabetes in a Youth Case–Control (SEARCH-CC) study, arsenic metabolism, characterized by the relative proportions of iAs% versus MMA% and DMA% in plasma, was associated with a higher odds of T1DM in children and adolescents (Grau-Perez et al. 2017). A growing number of epidemiology studies also support a potential role of maternal arsenic exposure in the development of gestational diabetes mellitus (GDM) (Salmeri et al. 2020). GDM is a common metabolic disease that develops during pregnancy and usually resolves after delivery. However, GDM is detrimental for both mothers and their children, with a higher possibility for mothers to develop T2DM later in life (Bellamy et al. 2009), and for children to exhibit impaired fetal endothelial function and have greater risks of developing T2DM (Dabelea et al. 2008; Kelstrup et al. 2013; Sultan et al. 2015). Farzan found that for every 5 µg/L increase in total arsenic in a woman’s household well water, there was a nearly 10% greater risk of GDM (Farzan, Gossai, et al. 2016). Another study of maternal-infant and birth cohort found that in utero arsenic exposure was associated with cord blood DNA methylation of genes involved in signaling pathways important in T1DM and T2DM development, as several DNA methylation patterns positively correlated with elevated low-density lipoprotein (LDL) levels later in life, one characteristic of diabetic dyslipidemia (Kaushal et al. 2017). Epidemiological evidence from different countries/regions has suggested the positive correlation between maternal urinary arsenic concentrations and GDM (Ashley-Martin et al. 2018; Wang et al. 2019; Wang et al. 2020). The detrimental effects of arsenic on glucose homeostasis have also been found in animal studies. In pregnant rats, sodium arsenite exposure from drinking water alters glucose homeostasis by altering beta-cell function, and increasing the risk of GDM development; in pups, it induces low body weight from birth to 8 weeks of age, and glucose intolerance in females (Bonaventura et al. 2017).

Humans are susceptible to extended exposures to multiple toxic metals daily, so sometimes it is rather difficult to find the exact mechanism through which each individual heavy metal causes DM (Javaid et al. 2021). Many epidemiological studies have investigated the associations between multiple heavy metal exposures in combination with arsenic and the impact on DM risk. A retrospective nested case–control study of 1359 pregnant women in China measured concentration of heavy metals (including arsenic, mercury, lead, cadmium, and chromium) in their newborns’ meconium samples and found that arsenic, cadmium, and chromium were positively associated with GDM prevalence in a dose-dependent manner (Peng et al. 2015). In occupational worker populations, a large prospective cohort study demonstrated that multiple urinary metals, particularly arsenic, nickel, zinc and cobalt, were associated with elevated glucose and dysglycemia among occupational workers in the metal industry (Yang et al. 2017).

Arsenic and immune responses

Dual adverse effects of arsenic on immune responses

The immune system is a protective complex of cells, special organs, and chemical mediators that prevent cellular or tissue damage in response to various pathogens or toxic insults. Increasing lines of evidence indicate that arsenic may adversely affect the immune system, by inducing proinflammatory or immunosuppressive effects, which we will discuss separately below.

In populations chronically exposed to low levels of arsenic (in the range above the WHO-recommended limit of 10 µg/L but ≤ 50 µg/L) in drinking water, arsenic has been found to stimulate the immune response (Sinha et al. 2014; Dutta et al. 2015). Prenatal and childhood exposure to arsenic (iAs + MMA + DMA) were found to be significantly associated with the occurrence of asthma and allergic rhinitis in children (Tsai et al. 2021). iAs, mainly sodium arsenite, has been well-documented to be associated with inflammatory responses in bronchial epithelial cells (Luo et al. 2013), urothelial cells (Liu, Sun, et al. 2014), skin (Li et al. 2011), atherosclerotic lesions (Srivastava et al. 2009), vascular (Kesavan et al. 2014), keratinocytes (Lu et al. 2015), and pulmonary tissue and also systemically in populations with histories of low level arsenic exposures (Sinha et al. 2014; Dutta et al. 2015). In mice prenatally exposed to sodium arsenite, the concentrations of arsenic in liver reached 1.2 mg/kg, and the mice showed altered developmental programming in fetal liver and a proinflammatory state postnatally (States et al. 2012). Also, it has been shown that chronic low-level arsenic-exposure in women can evoke systemic inflammation by upregulating proinflammatory mediators such as TNF-α, IL6, IL8, IL12, and CRP (Prasad and Sinha 2017), many of which have been proposed as important regulators of insulin resistance in T2DM. The inflammatory state, in the long run, may contribute towards the onset of chronic diseases including atherosclerosis, rheumatoid arthritis, cancers and DM.

Alternatively, increasing literature from epidemiological and experimental studies have reported that arsenic exposure is immunosuppressive and enhances susceptibility to infections or induces immune tolerance (Rahman et al. 2011; Smith et al. 2011; Ramsey, Foong, et al. 2013; Farzan, Li, et al. 2016; Li, Guo, et al. 2021). In some prospective study cohorts, prenatal exposure to iAs through maternal drinking water was associated with a higher morbidity in respiratory infections, diarrhea during infancy or hepatitis B seroprevalence (Rahman et al. 2011; Farzan, Li, et al. 2016; Cardenas et al. 2018). Also, arsenic exposure increased susceptibility to opportunistic infections like tuberculosis (Smith et al. 2011), or other diseases associated with immunosuppression, such as skin cancers in HIV-positive or organ transplantation populations (Soto-Pena et al. 2006). Mechanisms proposed for arsenic-induced immunosuppression has been changing CD4/CD8 T cell ratio (Soto-Pena et al. 2006) and interference with local immunity (Ramsey, Bosco, et al. 2013), altered microRNA patterns involved in innate and adaptive immune pathways (Rager et al. 2014), as well as decreased numbers of antigen presenting dendritic cells (DCs) in lymph nodes (Kozul et al. 2009) or induced expression of Heme oxygenase-1 (HO-1) protein to assist arsenic-suppressed production of pro-inflammatory factors and Th1/Th17-inducible cytokines in DCs (Li, Guo, et al. 2021) and inhibited proteolysis in antigen processing (Harrison and McCoy 2001).

Taken together, these findings suggest that arsenic exposure can result in immune stimulation or suppression. It is unclear if these immunotoxicological effects of arsenic are time-dependent or dose-dependent. An in vivo study of chronic sodium arsenite exposure in mice via drinking water showed that chronic arsenic exposure induces a time-dependent modulation of the inflammatory response at the initial stage of exposure and immunosuppression at a later stage of treatment in the spleen (Yan et al. 2020). There are in vitro studies showing that arsenic can either induce suppression or proliferation of immune cells in a dose-dependent manner (Meng and Meng 2000), but not all reflected a clear dose-dependent effect (Soto-Pena and Vega 2008). For example, the immunosuppressive effects were pronounced in male cotton rats exposed to 5 mg/L sodium arsenite in drinking water for 6 weeks (Savabieasfahani et al. 1998); whereas in another study, 0.5 mg/L sodium arsenite exposure in drinking water for 5 weeks in C57BL/6 mice enhanced inflammatory-induced angiogenesis (Straub et al. 2007). The reasons for these differences may be related to gender, species, strain, cell types, arsenic concentration or metabolic capacities of systems evaluated.

Mechanisms of arsenic-induced dual effects on innate and adaptive immunity

Dual effects of arsenic on innate immunity

Innate immunity serves as the first line of defense to protect against deleterious microorganisms or foreign materials or substances. Arsenic has been shown to have dual immunosuppressive and proinflammatory effects on innate immunity. On one hand, arsenic is increasingly being recognized as a suppressor of innate immunity (Kozul et al. 2009; Xu et al. 2018; Medina et al. 2020). Arsenic (mainly iAs) exposures from drinking water have been implicated in persistent opportunistic bacterial and viral infections in adults and in children exposed early in life (Mazumder et al. 2000; Smith et al. 2006; Raqib et al. 2009). One of the major components of innate immunity is the phagocytic leukocytes, mainly neutrophils and macrophages. Human macrophages are the major targets of arsenic-mediated immunotoxicity (Lemarie et al. 2006). Several studies as shown below have indicated that iAs can induce cell apoptosis and functional suppression of macrophages. Macrophages from arsenic-exposed people with skin lesions and whose urine arsenic levels averaged 642 µg/L, show cell rounding accompanied by significantly reduced adhesion, decrease in nitric oxide production, and impaired phagocytosis (Banerjee et al. 2009). Functional suppression of macrophages induced by environmentally relevant concentrations of iAs (50-500 nM sodium arsenite) in THP-1 derived macrophages is manifested by increased cell apoptosis, suppression of two pro-inflammatory cytokines, IL-1β and TNF-α, as well as inhibition of nitric oxide production (Xu et al. 2018). Epidemiologic evidence also shows the altered suppressive effects on innate immune function in children exposed to low-dose arsenic early in life (Parvez et al. 2019). In an experimental mouse model of bacteremic Staphylococcus aureus (S. aureus) infection, sodium arsenite (0.5 mg/kg/d) exposure via i.p. injection for 15 days significantly increased bacterial load in the blood, decreased adhesion properties and phagocytic activity of splenic macrophages, and delayed bacterial clearance by the spleen (Bishayi and Sengupta 2003). Another mouse study found iAs exposure was immunotoxic to intraepithelial lymphocytes (IELs) and innate immune cells in the small intestine of male and female mice following chronic drinking water exposures to sodium arsenite (Medina et al. 2020). Exposure to iAs resulted in a reduction of small intestinal CD⁴⁺ IELs (TCRαβ+, CD8αα+) and a decrease in mature macrophages of innate immune cell subsets (Medina et al. 2020).

Another major component of innate immunity is the inflammatory response which involves the release of various cytokines. Arsenic can impair the ability of macrophages to produce cytokines. Expression of macrophagic surface markers is also affected altering endocytosis and phagocytosis (Abdul et al. 2015). Microarray-based genome-wide expression studies of peripheral blood mononuclear cells (PBMC) from populations have shown the association between arsenic exposure and expression of inflammatory genes, such as IL2RA, IL1B, IL6 and TNF (Dangleben et al. 2013).

Alternatively, there are reports that arsenic can promote innate immune responses. The microglial cell is the key cell type involved in innate immune responses in the CNS, and its activation is linked to inflammation and neurotoxicity (Chen et al. 2016). In cultured microglial cells treated with arsenic trioxide from 0.5 to 20 nm, the expression and secretion of IL6 was increased in a dose- and time-dependent manner (Chen et al. 2016). In another in vitro study, treatment with sodium arsenate (0.5–100 µm) in a murine keratinocyte cell line, arsenic induced overproduction of cytokine IL1α which promotes skin hyperkeratosis and cancer (Corsini et al. 1999). In pregnant mice exposed to 100 µg/L sodium arsenite via drinking water from gestation day 8 until birth, genes related to mucus production and innate immunity were increased in the lungs of their offspring (Ramsey, Bosco, et al. 2013).

Dendritic cells play a critical role that link innate and adaptive immunity. They are central in the initiation of immune responses that include processes such as phagocytosis, antigen presentation and processing, in addition to T-cell activation and cytokine secretion. Innate and adaptive responses are orchestrated by dendritic cells, and minimal alterations in their function and stability and can majorly impact the immune system (Bahari and Salmani 2017). Many responses of dendritic cells are decreased in the presence of very low and non-cytotoxic concentrations of arsenic, through decreased secretions of the key cytokines like IL2, that reduce dendritic cell maturation and cellular immune responses (Macoch et al. 2013; Macoch et al. 2015).

Studies summarizing the effects of arsenic on the innate immune response are shown in Table 1. So far, the present studies may not draw a conclusion to elucidate whether there is commonality in the immunosuppressive or proinflammatory effects of arsenic on innate immunity. Further multidisciplinary studies including molecular epidemiologic investigations are needed.

Table 1. Effects of arsenic on the immune system.

Model	Arsenic exposure	Immune response to arsenic exposure	Major findings induced by arsenic exposure	References
Innate immune system
In humans:
A large population in West Bengal, India	Arsenic exposure through drinking water	Immunosuppressive	Increased the prevalence of cough, shortness of breath, and chest sounds (crepitations and/or rhonchi) in the lungs	Mazumder et al. (2000)
People in West Bengal, India	Arsenic exposure via ground water: urine arsenic averaged 642 µg/L	Immunosuppressive	Monocyte-derived macrophages cell rounding and reduced adhesion, nitric oxide anion production, and phagocytic capacity	Banerjee et al. (2009)
Pregnant women and their children in rural Bangladesh	Arsenic exposure via tubewell water	Immunosuppressive	Arsenic exposure was negatively correlated with child immunity (IL-7, lactoferrin and child thymic index), and positively associated with acute respiratory infections	Raqib et al. (2009)
Health Effects of Arsenic Longitudinal Study cohort in rural Bangladesh	Arsenic exposure via household water (0.6–326 µg/L)	Immunosuppressive	Early-life arsenic exposure disrupted the innate host defense pathway in children (plasma LL-37 levels, decreased MDM-mediated Spn killing and SBA responses, and PBMC formation of cytokines)	Parvez et al. (2019)
In rodent models:
Mouse model of bacteremic S. aureus infection	sodium arsenite (i.p. injection) at 0.5 mg/kg/d for 15 days	Immunosuppressive	Increased bacterial load in blood, decreased adhesion properties and phagocytic activity of splenic macrophages and delayed bacterial clearance by spleen	Bishayi and Sengupta (2003)
Pregnant mice	100 µg/L sodium arsenic via drinking water from gestation day 8 until birth.	Proinflammatory	Increased genes expression in the lung related to mucus production and innate immunity in their offspring	Ramsey, Bosco, et al. (2013)
Mouse model	drinking water exposures to sodium arsenite (100 µg/L or 500 µg/L) for 45–60 days	Immunosuppressive	Reduction of small intestinal CD4+ IELs (TCR$\text{α}\text{β}2$+, CD8$\text{α}\text{α}2$+) and mature macrophages of innate immune cell subsets	Medina et al. (2020)
In vitro models:
Murine keratinocyte cell line (HEL30)	sodium arsenate (0.5–100 µm)	Proinflammatory	cytokine (IL1$\text{α}2$) overproduction which promotes skin hyperkeratosis and cancer	Corsini et al. (1999)
HAPI microglia cells	arsenic trioxide (0.5–20 nm)	Proinflammatory	Upregulation of IL-6 in a dose- and time-dependent manner	Chen et al. (2016)
Macrophages derived from THP-1 cells	sodium arsenite (50 nM–500 nM)	Immunosuppressive	Suppression of pro-inflammatory cytokines (IL-1$\text{β}2$ and TNF-$\text{α}2$), nitric oxide, and phagocytosis	Xu et al. (2018)
Adaptive immune system
In humans:
Humans in Taiwan, China	Long-term exposure to arsenic through arsenic-tainted well water	Proinflammatory	Increased gene expression of inflammatory molecules (IL1B, IL6, CCL2/MCP1, MMP1 etc.) in circulating lymphocytes of people with intermediate (4.64–9.00 µg/L), and high (9.60–46.5 µg/L) levels of blood arsenic	Wu et al. (2003)
Cross-sectional study of children in a high arsenic polluted region in Mexico	Mean concentration in urine arsenic of 186.7 µg/L	Immunosuppressive	Reduced CD4 subpopulation proportion, CD4/CD8 ratio and IL-2 secretion	Soto-Pena et al. (2006)
Cross-sectional study in a high arsenic polluted region in India	Arsenic exposure via drinking water at the average level of 189.65 µg/L	Immunosuppressive	Decreased T-cell proliferation and secretion of cytokines (TNF-$\text{α}2$, IFN$\text{γ}2$, IL2, IL10, IL5 and IL4)	Biswas et al. (2008)
Copper smelter workers in Chile	Chronic arsenic exposure	Immunosuppressive	Lymphocyte oxidative damage with a positive correlation between urinary arsenic levels and oxidative stress markers	Escobar et al. (2010)
Male workers from nonferrous plants in the northeastern part of China	The subjects are separated into two groups: group 1 with urine total arsenic < 50 µg/g; group 2 with urine total arsenic ≥ 50 µg/g	Proinflammatory	Elevated concentrations of IL-8, TNF-$\text{α}2$, and TGF-$\text{α}2$ in urine in the high urinary arsenic workers (≥ 50 µg/g) compared with the low urinary arsenic workers (< 50 µg/g)	Liu, Sun, et al. (2014)
In rodent models:
Mice	Following a two-week inhalation exposure to arsenic trioxide at concentrations of 50 µg/m^3 and 1 mg/m^3	Immunosuppressive	T-dependent humoral immune response is highly susceptible to arsenic trioxide suppression	Burchiel et al. (2009)
Male ApoE-/- mice	Early post-natal exposure (3-week-old mice exposed to 49 mg/L arsenic as NaAsO2 in drinking water for 7–33 weeks) or adult exposure to 49 mg/L arsenic for 24 weeks	Proinflammatory	Increased oxidative stress, inflammation, and atherosclerotic lesion formation	Srivastava et al. (2009)
SKH-1 mice	via drinking water at 0, 50, 100 and 200 mg/L for 1 month	Proinflammatory	Enhanced ROS production, cutaneous inflammation-related p38/MAPKAPK-2 MAPK signaling, alterations in Th-1/Th-2/Th-17 cytokines/chemokines and their receptors	Li et al. (2011)
Rats	sodium arsenite (100 mg/L) through drinking water for 90 days	Proinflammatory	Increased production of pre-inflammatory mediators (IL-1$\text{β}2$, IL-6, MCP-1, VCAM, sICAM) and serum C-alterations	Kesavan et al. (2014)
In vitro models:
Isolated human peripheral blood mononuclear cells (PBMC)	sodium arsenite at 1 µM	Immunosuppressive	Reduced IL2 secretion, T cell proliferation and activity as well as immune and stress genes expression	Martin-Chouly et al. (2011)
Human bronchial epithelial (HBE) cells	sodium arsenite at 1.0 µM for 0, 10, 20, or 30 cell generations	Proinflammatory	Increased IL-6 secretion, activated STAT3 and miR-21 expression	Luo et al. (2013)
Human keratinocytes (HaCaT) cell line	Sodium arsenite at 1.0 µM for 30 cell passages	Proinflammatory	Increased IL-6 secretion, activated STAT3 and miR-21 expression	Lu et al. (2015)

Dual effects of arsenic on adaptive immunity

Adaptive immunity involves specific T-lymphocytes and B-lymphocytes, which respond to different antigens during our lifetime. Arsenic has been shown to have dual immunosuppressive and proinflammatory effects on adaptive immunity. One of the major ways arsenic interferes with the adaptive immune system is through its immunosuppressive effects by impairing the development, activation, proliferation and function of lymphocytes and B- and T-cell subpopulations (Abdul et al. 2015). In a cross-sectional study from a high arsenic polluted region in India, arsenic exposure through drinking water at the average level of 189.65 µg/L suppressed T-cell proliferation and decreased the levels of cytokines (TNF-$\text{α}2$, IFN$\text{γ}2$, IL2, IL10, IL5 and IL4) secreted by the T cells (Biswas et al. 2008). In children with a mean arsenic concentration (186.7 µg/L in urine) of iAs and metabolites MMA and DMA, CD4 subpopulation proportions, CD4/CD8 ratio, and IL2 secretion levels were all reduced, as well as a tendency of increased incidences of immune-related conditions (asthma, allergies, and parasitic infections) were noted among individuals with arsenic values higher than 50 µg/L in their urine (Soto-Pena et al. 2006). Chronic arsenic exposure in copper smelter workers also induces lymphocyte oxidative damage showing a positive correlation between urinary arsenic levels and oxidative stress markers (Escobar et al. 2010). Additionally, impaired T cell functions have also been reported in experimental animals exposed to arsenic. In mice following a two-week inhalation exposure to arsenic trioxide at concentrations of 50 µg/m³ and 1 mg/m³, the T-dependent humoral immune response was highly susceptible to arsenic suppression (Burchiel et al. 2009). In vitro experiments also show that sodium arsenite reduces T cell proliferation induced by phytohemaglutinin (PHA), and interferes with expression of several immune and stress genes (Martin-Chouly et al. 2011).

Additionally, it has been found that arsenic can stimulate the adaptive immune system into a proinflammatory state. Long-term exposure to arsenic increased gene expression of inflammatory molecules (IL1B, IL6, CCL2/MCP1, MMP1, etc.) in circulating lymphocytes of people with intermediate (4.64–9.00 µg/L), and high (9.60–46.5 µg/L) levels of blood arsenic (Wu et al. 2003). In male workers from nonferrous plants located in the northeastern part of China with exposure to arsenic, urinary IL8, TNF-$\text{α}2$ and TGF-$\text{α}2$ levels are also positively associated with urinary iAs concentrations (Liu, Sun, et al. 2014). Arsenic can promote inflammation in vivo (Srivastava et al. 2009; Li et al. 2011; Kesavan et al. 2014) and in vitro (Luo et al. 2013; Lu et al. 2015), all of which provide evidence of a proinflammatory response induced by arsenic. Studies characterizing the effects of arsenic on the adaptive immune system are summarized in Table 1.

Mechanisms of arsenic-induced DM through immunity

Based on the above associations between arsenic, the immune system and DM, which includes: (1) Arsenic is a well-recognized modulator of the immune system; (2) dysregulation of the immune system is one of the hallmarks in the pathogenesis of DM; and (3) iAs exposure increases the risk of development of DM, we suggest that iAs exposure predisposes to dysfunctional immune responses, thereby alters the expression and secretion of inflammatory cytokines and mediators that could possibly initiate the development of T1DM or T2DM (as illustrated in Figures 3 and 4, respectively). Therefore, we review the role of immune system as a link to explain potential mechanisms of iAs-induced DM in human populations, animal models, and in vitro studies.

Figure 3. Mechanisms of arsenic-induced effects on immunity in T1DM. Arsenic can promote the development of T1DM through direct pancreatic beta-cell destruction, gut microbiome alterations inducing abnormal autoimmune or inflammatory responses.

Figure 4. Mechanisms of arsenic interference of immunity in T2DM. Arsenic-induced T2DM is mainly through insulin resistance due to impairing the insulin signaling cascade, affecting oxidative stress and antioxidant pathways and inducing chronic inflammation; pancreatic beta-cell damage and dysfunction due to ROS production and enzymatic impairment; stimulation of liver gluconeogenesis due to Pck1 expression, among which the immune system is implicated to be involved in these processes.

Mechanisms of arsenic-induced effects on immunity in development of T1DM

T1DM is primarily characterized by impaired insulin secretion as a result of an autoimmune assault on pancreatic islet beta-cells. There are few experimental animal studies or epidemiological studies that have directly evaluated the role of arsenic in T1DM. An epidemiological study supports the role of arsenic metabolism in T1DM, showing that low iAs% versus a high MMA% and DMA% in plasma of youth was associated with a higher odds of T1DM (Grau-Perez et al. 2017).

Several points of evidence of mechanistic data that supports an association between arsenic and T1DM have been proposed and includes the following.

Direct beta-cells damage. Pancreatic beta-cells are recognized as a target of iAs and methylated metabolites (MMA and DMA) in humans, animals, and in in vitro models (Fu et al. 2010; Douillet et al. 2013; Wu et al. 2018), by inducing the release of TNF-α and ILs, activating oxidative stress and transcription factor NFE2L2 (Nuclear factor, erythroid 2 like 2; also known as NRF2)-mediated antioxidant response, and eliciting a pancreatic islet-autonomous effect by an endoplasmic reticulum stress-autophagy pathway, which all play an important role in islet beta-cell destruction in T1DM (Tseng 2004; Chen et al. 2009; Khurana et al. 2018; Wu et al. 2018). Low-level sodium arsenite can also induce apoptosis by inhibiting TrxR activity in pancreatic beta-cells (Yao et al. 2015). These direct undesirable effects on beta-cell function and enhanced apoptosis may be possible mechanisms of T1DM development due to arsenic exposure (Tseng 2004).

Altered gut microbiome. The gut and intestinal microbiome composition and metabolic profile can be altered by chronic arsenic exposure during prenatal development and early life in children (Dong et al. 2017) and by acute/chronic sodium arsenite exposure in adult mice (Lu et al. 2014; Li, Yang, et al. 2021). In a recent study of mice exposed to sodium arsenite (50 mg/L) for 6 months, long-term exposure to iAs changed the structure of intestinal tissues, affected the Beta diversity of intestinal flora, altered the composition of the intestinal microbiota from phylum to species, and changed the composition of metabolites (Li, Yang, et al. 2021). The changes in the gut/intestinal microbiome and metabolome induced by arsenic may potentially influence susceptibility to metabolic diseases such as T1DM. Some human population studies have observed an altered microbiota in T1DM patients (de Goffau et al. 2014; Pellegrini et al. 2017; Vatanen et al. 2018). The association between T1DM and the gut microbiome is that microbiome dysbiosis in individuals may lead to abnormal immunoregulation that enables autoimmune destruction of islet beta-cell. Also, the loss of epithelial barrier integrity augments gut/intestinal permeability, and may ultimately trigger beta-cell autoimmunity by activating islet-reactive T cells, release of cytokines, beta-cell destruction with induction of T1DM (Butalia et al. 2016; Zhang et al. 2021).

Abnormal autoimmune or inflammatory responses. iAs may impair the immune system in humans and animals, especially interfering with the adaptive immune system by altering the development, activation, proliferation and function of subpopulations of lymphocytes (Dangleben et al. 2013; Abdul et al. 2015). Chronic iAs exposure from drinking water increases serum levels of antinuclear antibodies (ANA) and anti-doubled stranded DNA (dsDNA) antibodies, thereby possibly increasing the incidence of autoimmune diseases like T1DM (Das et al. 2012). Improper activation of the immune system may lead to allergic responses or trigger autoimmunity, and exposure to arsenic has been also reported to be associated with asthma and/or allergies in an epidemiological study (Das et al. 2014). Most likely, chronic long-term arsenic exposure enhances immune responses against self-antigens within the islet beta-cells and disrupts beta-cell function, thereby exacerbating the severity of the autoimmune diabetogenic process.

The mechanisms of arsenic interference on immunity in T1DM are summarized in Figure 3 and associated studies are shown in Table 2.

Table 2. Molecular mechanisms of arsenic-induced immune dysfunction in T1DM.

Model	Arsenic exposure	Major findings	References
In humans:
Humans in a highly arsenic-contaminated district in West Bengal, India	Arsenic content in water at 202.68 ± 188.12 µg/L; in urine at 612.90 ± 336.91 µg/L	Secretion of inflammatory cytokines IL-6 and IL-8, and serum ANA levels (an autoimmunity marker)	Das et al. (2012)
Humans in five sites of US (South Carolina, Ohio, Colorado, California, and Washington)	Mean total Arsenic level at 83.1 ng/L in plasma	Low iAs% versus high monomethylated arsenic [MMA]% and dimethylated arsenic [DMA]% is associated with a higher odds of T1DM	Grau-Perez et al. (2017)
Children in Bangladeshi aged 4–6 years old	Mean drinking water arsenic concentration for the high and low exposed groups at 218.8 and 1.7 µg/L respectively	Alteration of the gut microbiome composition in early childhood	Dong et al. (2017)
In vivo rodent models:
Mice	Oral exposure to 10 mg/L arsenite for 4 weeks	Changes in composition of gut microbiota	Lu et al. (2014)
C57/BL6J mice	As2O3 at 1, 2 and 4 mg/L via drinking water for 12 weeks	Islet $\text{β}2$-cell destruction: impaired $\text{β}2$-cell function through endoplasmic reticulum stress-autophagy pathway	Wu et al. (2018)
C57BL/6 mice	sodium arsenite at 50 mg/L in drinking water for 6 months	Disruption in structure of intestinal tissues, the intestinal microbiome and metabolome	Li, Yang, et al. (2021)
In vitro models:
Murine pancreatic islets	48-hour exposures to low subtoxic concentrations of iAsIII, MAsIII or DMAsIII.	Islet $\text{β}2$-cell destruction: Inhibition of glucose stimulated insulin secretion	Douillet et al. (2013)
INS-1 cells	sodium arsenite at 0.25-1 µM for 96 h	Apoptosis of pancreatic beta-cells through inhibiting TrxR activity	Yao et al. (2015)

Molecular mechanisms of arsenic interference of immunity in development of T2DM

Many epidemiological studies have reported an association between arsenic exposure and T2DM (Navas-Acien et al. 2008; Maull et al. 2012; Brauner et al. 2014), with the major mechanisms including: (1) inhibition of insulin-dependent glucose uptake (insulin resistance) due to inhibition of the insulin-activated signaling cascade and glucose transport system; (2) pancreatic beta-cell damage and dysfunction due to oxidative stress and enzymatic impairment; (3) stimulation of liver gluconeogenesis due to expression of phosphoenolpyruvate carboxykinase (PCK1, also known as PEPCK), among which the immune system is involved in the above processes (Martin et al. 2017). The mechanisms of arsenic interference in immunity and development of T2DM are summarized in Figure 4 and associated studies are shown in Table 3.

Table 3. Molecular mechanisms of arsenic-induced immune dysfunction in T2DM.

Model	Arsenic exposure	Major findings	References
Arsenic-induced insulin resistance and inhibition of glucose uptake:
3T3-L1 adipocytes	arsenite (up to 50 µM) and methylarsonous acid (up to 2 µM) for 4h	Inhibited ISGU by suppressing PKB/Akt phosphorylation	Walton et al. (2004)
3T3-L1 adipocytes	arsenite (up to 100 mM) and methylarsonous acid (up to 300 µM) for 4h	inhibited ISGU by inhibition of PDK-1/PKB/Akt-mediated transduction	Paul et al. (2007)
Mice	arsenite (up to 50 mg/L) or methylarsonous acid (up to 5 mg/L) in drinking water for 8 weeks	Impaired glucose tolerance by iAs^III	Paul et al. (2008)
3T3-L1 adipocytes	iAs^3+ up to 2 µM for 7 days	Suppressed glucose uptake by stimulating ROS accumulation and activation of NFE2L2 pathway, increased inflammatory response genes and decreased GLUT4 expression	Xue et al. (2011)
Mice	NaAsO2 orally 3 mg/kg body weight for 12 weeks	Cyto-degenerative effect and ROS accumulation in pancreatic beta-cells and hepatocytes; over-expression of TNF-$\text{α}2$ and IL6; impaired insulin signaling by reducing GLUT4, IRS1, IRS2, PI3 K, AKT and PPAR$\text{γ}2$	Chakraborty et al. (2012)
Mice	sodium arsenite at 5, 10 and 20 mg/kg, intra-gastrically	Activation of NFE2L2-mediated cellular antioxidant system in spleen, thymus, and PBMCs	Duan et al. (2015)
Mice 3T3L1 preadipocytes and C2C12 myotubes	arsenic at 0.5-2 µM for 8 weeks	Decreased insulin-stimulated glucose uptake and GLUT4 expression via oxidative stress-regulated mitochondrial Sirt3-FOXO3a signaling	Padmaja Divya et al. (2015)
Arsenic-induced beta-cell dysfunction:
Pancreatic beta-cells obtained from young adult male Wistar rats	sodium arsenite at 0.5, 1, 2, 5 and 10 µM for 72 and 144 h	Impaired pancreatic beta-cell insulin synthesis and secretion	Diaz-Villasenor et al. (2006)
Male Wistar rats	sodium arsenite orally at 1.7 mg/kg (12 h) for 90 days	Induced oxidative stress, pancreas damage, and insulin resistance	Izquierdo-Vega et al. (2006)
Hamster pancreatic HIT-T15 beta-cell; ICR mice	In HIT-T15 beta-cells, at 1–10 µM for 24 h; in mice, As2O3 at 10 mg/L in drinking water for 12 weeks	In HIT-T15 cells, induced oxidative stress. In mice, induced sub-acute peri-pancreatitis and islet damage	Yen et al. (2007)
RINm5F rat pancreatic beta-cell	Sodium arsenite (up to 5 µM) for 72 h	Impaired insulin secretion by decreasing the oscillations of [Ca^2±]i and calpain-10 partial proteolysis of SNAP-25	Diaz-Villasenor, Burns, et al. (2008)
INS-1(832/13) cells	Arsenite at 0.05–0.5 µM for 96 h	Activation of NFE2L2 and oxidative stress response, dampened ROS signaling, decreased GSIS and disturbed beta-cell function	Fu et al. (2010)
Pancreatic beta-cell-derived RIN-m5F cells and male ICR mice	As2O3 at 10 µm in RIN-m5F cells; at 5 mg/L in drinking water in mice for 6 weeks	Decreased insulin secretion and cell viability, increased beta-cell apoptosis	Lu et al. (2011)
Normal and male C57BKS/Leprdb (db/db) mice	sodium arsenite at 3 mg/L for 16 weeks	Increased gluconeogenesis, pancreatic beta-cell dysfunction, and oxidative damages in pancreas in both normal and diabetic mice	Liu, Guo, et al. (2014)
INS-1 cells	sodium arsenite at 0.25–1µM for 96 h	Apoptosis through inhibiting TrxR activity in pancreatic beta-cells	Yao et al. (2015)
INS-1 cells	sodium arsenite at 10 µm for 72 h	Increased ROS, reduced mitochondrial membrane potential, disrupted lysosomal membrane integrity, and induced pancreatic beta-cell apoptosis	Pan et al. (2016)
Arsenic-associated stimulation of gluconeogenesis:
Chick embryo in vivo and rat hepatoma H411E cells in culture	sodium arsenite at 100 µM/kg in vivo and 0.33 µM in cells	Increased both basal and hormone-induced Pck1 gene expression in chick embryo and H411E cells	Hamilton et al. (1998)
Mice	As2O3 at 0.05 or 0.5 mg/L via drinking water for 2–6 weeks	Increased hepatic gluconeogenesis by up-regulating Pck1 gene expression	Huang et al. (2015)
Arsenic induced inflammatory cytokines:
Humans in Taiwan, China	Subjects with low (0–4.32 µg/L), intermediate (4.64–9.00 µg/L), and high (9.60–46.5 µg/L) levels of blood arsenic	Elevated gene expressions of IL1B, IL6 chemokine C-C motif ligand 2/monocyte chemotactic protein-1 (CCL2/MCP1), chemokine C-X-C motif ligand 1/growth-related oncogene alpha, chemokine C-X-C motif ligand 2/growth-related oncongene beta, CD14, and matrix MMP1	Wu et al. (2003)
Pregnant women from a longitudinal, mother–child cohort in Matlab, Bangladesh	The median Urine arsenic concentrations at GW8 and GW30 were at 66 and 60 µg/L respectively	Enhanced placental inflammatory responses (IL1$\text{β}2$, TNF-$\text{α}2$, and IFN$\text{γ}2$), reduced placental T cells, and altered cord blood cytokines (IL1$\text{β}2$, IL8, IFN$\text{γ}2$, TNF-$\text{α}2$)	Ahmed et al. (2011)
Pregnant women in the New Hampshire area in US	High level (> 5 µg/L in water, and > 5 µg/L in urine) and low level (< 0.5 µg/L in water, and < 5 µg/L in urine)	Maternal urinary arsenic concentrations are related to impaired function of T-cell in cord blood; in placental samples, higher in utero urinary arsenic concentrations are positively associated with IL1$\text{β}2$ expression	Nadeau et al. (2014)

Arsenic-induced insulin resistance and inhibition of insulin-stimulated glucose uptake (ISGU)

A major mechanism involved in the development of insulin resistance and pancreatic beta-cell dysfunction is oxidative stress, and this process is thought to be critical in the pathogenesis of T2DM (Pi et al. 2010). One of mechanisms of arsenic-associated T2DM is inhibition of insulin-dependent glucose uptake through generation of ROS and upregulation of antioxidant pathways including NFE2L2, leading to insulin resistance through decreasing ISGU in insulin sensitive peripheral tissues such as adipose and muscle (Xue et al. 2011; Padmaja Divya et al. 2015). In adipocytes and myotubes, sodium arsenite induces insulin resistance via oxidative stress-regulated mitochondrial Sirt3-FOXO3a signaling pathway (Padmaja Divya et al. 2015). In primary mouse pancreatic beta-cells and hepatocytes, sodium arsenite produced cyto-degenerative effects and ROS accumulation, and in vivo, sodium arsenite impaired insulin signaling by reducing expression levels of GLUT4, IRS1, IRS2, PI3 K, AKT, and PPAR signaling molecules (Chakraborty et al. 2012). Moreover, excessive levels of ROS result in upregulation of antioxidant pathways. In mice, iAs exposure stimulates generation of ROS and activates the NFE2L2-mediated cellular antioxidant pathway in the immune system (Duan et al. 2015). In 3T3-L1 adipocytes, sodium arsenite exposure resulted in upregulation of NFE2L2 and the subsequent increase in antioxidant enzyme expression, which blunted insulin-stimulated ROS signaling and thus impairing ISGU (Xue et al. 2011). Also, sodium arseniteinduced significant induction of inflammatory response genes and decreased expression of GLUT4, which meant decreased transport and utilization of the glucose, suggesting the involvement of chronic inflammation and reduction of GLUT4 in iAs-induced insulin resistance (Xue et al. 2011).

Interestingly, the exact mechanisms underlying NFE2L2’s role in insulin resistance have not been fully clarified. Most studies found that increased NFE2L2 inhibited insulin signaling and aggravated insulin resistance, while NFE2L2 deletion ameliorated insulin resistance; however, some conflicting results have been reported based on the different approaches and animal models (Li et al. 2020). Therefore, more research is needed to elucidate the role of NFE2L2 in arsenic-induced insulin resistance.

Another possible mechanism for arsenic-induced T2DM is the inhibition of ISGU by disrupting insulin-stimulated signal transduction pathway PKB/Akt-dependent mobilization of GLUT4 transporters in adipocytes. A series of studies from Dr Miroslav Styblo’s group have shown that trivalent arsenicals, arsenite and methylarsonous acid decreased ISGU in 3T3-L1 adipocytes by inhibiting PKB/Akt phosphorylation (Walton et al. 2004), and they further identified the key mechanism responsible for this effect was by the inhibition of the PDK-1/PKB/Akt-mediated transduction step (Paul et al. 2007). Follow-up studies in male C57BL/6 mice chronically exposed to arsenite or methylarsonous acid in drinking water confirmed that arsenite could induce impaired glucose tolerance in mice, but only at a higher concentration (50 mg/L) than that in humans (Paul et al. 2008). This may because of the higher rate of iAs metabolism and clearance in mice compared to humans.

In addition, arsenic may induce a status of insulin resistance through its molecular effects on gene expression of certain cytokines, which may contribute to arsenic-induced T2DM. Arsenic has been demonstrated to increase the production of cytokines involved in inflammation. Expression of inflammatory genes IL1B and IL6 were elevated in association with arsenic exposure in human subjects (Wu et al. 2003). In a US pregnancy cohort in New Hampshire, where arsenic levels exceeded the current EPA maximum contaminant level of 10 µg/L, in utero arsenic exposure has shown greater IL1β expression in the placenta and changes of T cell phenotypes and T cell function in cord blood, which may alter fetal immune system development and lead to immune dysregulation (Nadeau et al. 2014). Another study in Bangladesh showed that maternal exposure to iAs during pregnancy increased oxidative stress and proinflammatory cytokines (IL1β, TNF-α, and IFNγ) in the placenta and altered cord blood cytokines (IL1β, IL8, IFNγ, TNF-α) (Ahmed et al. 2011). Abundant evidence has confirmed TNF-α, IL6 and IL1β are mediators of insulin resistance due to their interference with insulin receptor signaling in adipocytes, hepatocytes, fibroblasts and myocytes (Le Roith and Zick 2001; Lagathu et al. 2003; Rotter et al. 2003; Koenen et al. 2011; Ballak et al. 2015), supporting that chronic arsenic exposure can lead to insulin resistance by inducing cytokines and regulating related genes, thereby promoting the development of T2DM.

Arsenic-induced beta-cell dysfunction and impaired insulin secretion

Pancreatic beta-cells can be targeted and damaged by oxidative stress induced by arsenic in the development of T2DM. Studies report that arsenic can increase ROS generation, which impairs insulin expression or glucose-stimulated insulin secretion (GSIS), and induces beta-cell dysfunction and apoptosis (Fu et al. 2010; Liu, Guo, et al. Diaz-Villasenor et al. 2006; Izquierdo-Vega et al. 2006; Lu et al. 2011; 2014; Yao et al. 2015; Pan et al. 2016). In primary rat pancreatic beta-cells, 5 µM sodium arsenite exposure contributed to the development of DM by impairing beta-cell function, mainly in insulin synthesis and secretion (Diaz-Villasenor et al. 2006). Subchronic oral exposure of rats to sodium arsenite (1.7 mg/kg) for 90 days produced oxidative stress-induced pancreatic damage, which could activate stress-sensitive signaling pathways to induce insulin resistance and also impair insulin secretion due to cellular damage (Izquierdo-Vega et al. 2006). Another in vivo study showed that iAs exposure could cause pre-diabetic effects in normal mice and worsen the diabetic effects in diabetic mice, by altering lipid metabolism, gluconeogenesis and insulin secretion, and causing oxidative damage and inflammation in the pancreas (Liu, Guo, et al. 2014).

Numerous in vitro studies also consistently demonstrated that iAs can increase ROS generation to reduce insulin synthesis and/or secretion, and induce beta-cell apoptosis (Fu et al. 2010; Lu et al. 2011; Yao et al. 2015; Pan et al. 2016). In rat INS-1 cells exposed to 10 µm of sodium arsenite for 72 h, iAs induced pancreatic beta-cell apoptosis through activation of a lysosome-mitochondrial pathway by increasing ROS level (Pan et al. 2016). Additionally, exposure to low levels of sodium arsenite in INS-1 cells also enhanced NFE2L2 and other antioxidant enzymes levels, which are inclined to dampen ROS signaling in GSIS, thus disturbing beta-cell function (Fu et al. 2010). Interestingly, similar to NFE2L2’s role in insulin resistance, some studies have reported contradictory results on the role of NFE2L2 in arsenic-induced beta-cell damage. In mouse MIN6 cells or islets with stable knockdown of Nfe2l2, deficiency of NFE2L2 rendered beta-cells vulnerable to iAs-induced cell damage, and pretreatment of MIN6 cells with NFE2L2 activator protected the cells from iAs-induced cell damage in a NFE2L2-dependent fashion (Yang et al. 2012). It’s not clear if NFE2L2 plays paradoxical roles in beta-cell dysfunction induced by environmental arsenic exposures. Further studies are needed to investigate the precise role of NFE2L2 in arsenic-induced beta-cell dysfunction. In addition to causing cell death, the accumulation of ROS has been shown to participate in chronic inflammation which results in pancreatitis in vivo, leading to the development of DM (Yen et al. 2007).

There is another mechanism reported by A. Díaz-Villaseñor et al, by which arsenic impairs insulin secretion without altering its synthesis. They found that arsenite-impaired insulin secretion in RINm5F rat pancreatic beta-cells was due to the reduced levels of free intracellular calcium, thus decreasing calpain-10-mediated proteolysis of SNAP-25, a member in the insulin secretory machinery, and limiting insulin exocytosis (Diaz-Villasenor, Burns, et al. 2008). Since calpains play an important role in insulin secretion and action as nonlysosomal calcium-dependent cysteine proteases, they compared the activity and expression in lymphocytes of individuals with or without T2DM, and showed that the activity of calpain-10 in lymphocytes was glucose-dependent and decreased in diabetic patients (Diaz-Villasenor, Hiriart, et al. 2008). This indicates the important role of enzymatic impairment in the development of T2DM, which not only occurs in tissues involved in glucose-uptake and metabolism such as the pancreas, adipose, liver, and skeletal muscle which directly regulate glucose homeostasis; but also can arise in white blood cells such as lymphocytes, which contain insulin receptors involved in insulin clearance and metabolism.

Arsenic-associated stimulation of gluconeogenesis

In addition to insulin resistance and insulin secretion alteration, arsenic can induce hepatic gluconeogenesis, which is the pathway for synthesis of glucose from noncarbohydrate sources during fasting. PCK (also known as PEPCK) is a key rate-limiting enzyme in hepatic gluconeogenesis, resulting in fasting hyperglycemia (Friedman et al. 1993). Sodium arsenite has been shown to increase both basal and hormone-induced Pck1 gene expression in chick embryos in vivo and in cultured rat hepatoma H411E cells at non-overtly toxic doses (Hamilton et al. 1998). Increased expression of Pck1 is also found in mice exposed to arsenic trioxide (Huang et al. 2015). Sodium arsenite can also upregulate Pck1 gene expression, whose overexpression can increase gluconeogenesis in liver and blood glucose levels in mice (Liu, Guo, et al. 2014). These data collectively, suggest arsenic-induced hepatic gluconeogenesis which may result in hyperglycemia and DM.

Mitigation strategies for arsenic exposure

A major threat to public health posed by arsenic is from drinking water and dietary intake of food crops irrigated by contaminated water worldwide (WHO 2006). Therefore, mitigation strategies focus primarily on reducing consumption of arsenic-rich drinking-water and arsenic-contaminated water sources. Various solutions for reduction in sources of arsenic contamination fall into two broad categories: alternative arsenic-safe water sources and treatment of arsenic-contaminated water supplies (Sanyal et al. 2020). Alternative water sources include groundwater (deep tube wells, shallow shrouded tube wells, or infiltration gallies), cleaning up of surface water (community filtration systems, household filters), and rainwater harvesting (Krupoff et al. 2020; Sanyal et al. 2020). Drinking water highly contaminated with arsenic can be mitigated in several different ways, such as sedimentation, oxidation/filtration to remove iron and manganese, enhanced coagulation/filtration and lime softening, as well as membrane processes including reverse osmosis, electrodialysis, ultrafiltration, and nanofiltration (WHO 2006; Sanyal et al. 2020).

In regards to dietary absorption of arsenic, crops irrigated with arsenic-contaminated water pose another major source of human exposure. Recent studies have reported that the adaptation of different agronomic techniques can effectively reduce bioaccumulation of arsenic in rice grains, such as rice grown aerobically, sprinkler irrigation, and genetically modified rice (Xu et al. 2008; Moreno-Jimenez et al. 2014; Chen et al. 2017).

In addition, several microorganisms have been isolated as they are arsenate-resistant or use arsenic to ‘breathe’, which may be used as a way of arsenic removal from water (Oremland and Stolz 2005; Dey et al. 2016). Some experimental studies have found that dietary supplements like vitamin C, different spectra of medicinal plants, and drugs can reduce arsenic toxicity through antioxidant effects, although these haven’t been assessed in humans (Bhattacharya 2017; Ling et al. 2017).

Conclusions

In summary, both experimental and epidemiological studies have demonstrated that arsenic exposure may potentiate the development of DM through dysfunctional immune responses. Arsenic exposure can adversely affect the immune system by inducing dual proinflammatory or immunosuppressive effects on both the innate and adaptive immune systems. Arsenic exposure predisposes to dysfunctional immune responses, which alters the expression and secretion of inflammatory cytokines and mediators that could possibly initiate the development of T1DM or T2DM.

The mechanisms of T1DM development due to arsenic exposure may include direct beta-cells damage, altered gut microbiome and abnormal autoimmune or inflammatory responses induced by arsenic. However, the role of arsenic in the development of T1DM needs further exploration due to lack of enough direct supportive evidence from epidemiologic, in vitro, or animal studies. For example, if arsenic exposure truly predisposes individuals to enhanced autoimmunity, associations with other autoimmune diseases, such as rheumatoid arthritis, lupus etc., might be detected in arsenic exposed populations and should be explored. Clearly, additional research needs to be done to address these concerns.

In the development of T2DM, arsenic exposure can induce the major mechanisms involved in T2DM initiation including: Inhibition of insulin-dependent glucose uptake (insulin resistance); induction of pancreatic beta-cell damage and dysfunction; stimulation of liver gluconeogenesis, among all of which the immune system is involved in the above processes. In regards to the association of multiple heavy metal exposures and DM, especially T2DM and GDM, some results are conflicting. Large prospective cohort studies of well-characterized populations, standardization across study designs, and consideration of molecular mechanisms informed by model systems studies are needed for further investigation of combined exposures of arsenic and other metals as a risk factor for DM.

The role of the immune system in DM and the association between arsenic exposure and DM have important clinical implications. First, specific molecules involved in arsenic-activated immune responses may hold the promise of potential biomarkers, and even preventive targets for arsenic-associated DM. Second, great efforts are underway to prospectively evaluate the relationship between arsenic exposure and other chronic low-grade inflammatory diseases, which are commonly associated with DM, such as atherosclerosis and stroke. Nevertheless, it is crucial to further understand the potential mechanisms of how arsenic regulates immune responses in the setting of DM. This could undoubtedly contribute to preventive approaches in the management of diabetes and define environmental risk factors for the disease. Therefore, it is imperative that further in vitro and in vivo studies are needed, in addition to epidemiologic studies to further define underserved populations that may be overly burdened with contaminants, to determine how pollutants like arsenic can induce DM or other diseases through perturbation of the immune system.

Acknowledgements

The authors greatly appreciate the assistance of Mr Paul Cacioppo in the compilation of the figures. The authors sincerely thank Dr Jun Wang for his critical review and comments on the manuscript. JL, TH and XG wrote the paper. DD and HX revised and approved the version to be published. All authors agree to be accountable for all aspects of the work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

No data were generated for this review manuscript.

Additional information

Funding

This work was funded by the Intramural Research Program of the NIH, NIEHS and DTT [ZIA ES021196-27].