Arsenic immunotoxicity: a review

Exposure to arsenic (As) is a global public health problem because of its association with various cancers and numerous other pathological effects, and millions of people worldwide are exposed to As on a regular basis. Increasing lines of evidence indicate that As may adversely affect the immune system, but its specific effects on immune function are poorly understood. Therefore, we conducted a literature search of non-cancer immune-related effects associated with As exposure and summarized the known immunotoxicological effects of As in humans, animals and in vitro models. Overall, the data show that chronic exposure to As has the potential to impair vital immune responses which could lead to increased risk of infections and chronic diseases, including various cancers. Although animal and in vitro models provide some insight into potential mechanisms of the As-related immunotoxicity observed in human populations, further investigation, particularly in humans, is needed to better understand the relationship between As exposure and the development of disease.


Background
Exposure to arsenic (As) is a global public health concern because As is widely distributed and associated with numerous adverse effects. As is a well-established cause of skin, lung and bladder cancers in humans [1], and is associated with skin lesions, diabetes, cardiovascular disease and other disorders [1][2][3]. Well over 100 million people worldwide are exposed to As, particularly through ingestion of contaminated food and water in countries such as India, Bangladesh, Taiwan, Chile, and the United States [1,4]. Exposures also occur through inhalation, especially in agricultural and industrial settings [4].
Inorganic As exists in the environment as arsenite (As III ) or arsenate (As V ) and is metabolized in humans via conversion of As V to As III with subsequent methylation to mono-and di-methylated arsenicals (MMA and DMA, respectively) [5]. MMA III is considered the most toxic arsenical in vitro [6][7][8][9] and individuals who excrete a higher proportion of ingested As as urinary MMA have increased risks of As-associated cancers [10,11], suggesting a key role for MMA in As toxicity. Proposed mechanisms of toxicity include oxidative stress, inhibition of DNA repair, chromosomal aberrations, micronuclei formation, induction of apoptosis, modification of cellular signaling via altered activation, expression and DNA binding activity of transcription factors, epigenetic modifications resulting in aberrant gene expression, and altered phenotype of stem cell populations [12][13][14][15][16]. As toxicity is thus complex and multifaceted, but is not yet well understood.
Although extensive research has focused on investigating As carcinogenicity, growing evidence indicates that As also has deleterious effects on the immune system [17,18]. This may potentially play a role in As carcinogenesis of various tissues through reduced immune surveillance. However, the specific effects of As on immune function remain poorly understood. Therefore, we considered that further investigation of As immunotoxicity is warranted and conducted a PubMed search of As exposure and non-cancer immune-related effects through October 2012. Here, we summarize the known toxicological effects of As on immune function in humans, laboratory animals and in vitro models, and identify possible future research directions to help close the gaps in knowledge.

Effects in adults Gene expression
Microarray-based assays are widely used for identifying differentially expressed genes in investigations of As carcinogenicity. However, a limited number of reported epidemiological studies have employed this powerful method to investigate As toxicity in immune cells from otherwise healthy persons. A microarray-based genomewide expression study of peripheral blood mononuclear cells (PBMC) from 21 subjects in New Hampshire whose drinking-water As averaged 0.7 μg/L (range 0.007-5.3 μg/L, n = 10) and 32 μg/L (range 10.4-74.7 μg/L, n = 11) showed significant differences between exposure groups in transcripts with functions in T-cell receptor signaling, cell cycle regulation and apoptosis, and most strikingly defense and immune response [19]. Notably, higher As exposure was associated with increased expression of killer cell immunoglobulin-like receptors that inhibit natural killer cell-mediated cytotoxicity, as well as decreased expression of MHC class II molecules, HLA-DQB1, HLA-DPA1, and HLA-DRB1; defense response genes, CD69, HSPA9B and MALT1; and inflammatory genes, IL2RA and IL1B [19]. Exposure was determined by levels of drinking-water As combined with urinary or toenail As as internal markers of exposure, and control and exposed subjects were matched for age, sex and smoking status.
Down-regulated IL1B was also identified in a microarray study of PBMC from an As-exposed Bangladeshi population with (n = 11) and without skin lesions (n = 5) [20]. An overall suppression of 467/468 differentially expressed genes was observed. These findings contradict those from a microarray analysis of PBMC from 24 individuals in Taiwan with low (0-4.32 μg/L), intermediate (4.64-9 μg/L), and high (9.6-46.5 μg/L) blood As levels [21]. Among 62/708 significantly altered genes were several inflammatory molecules that were up-regulated, including IL1B, IL6, CCL2 and CD14, indicating that prolonged exposure may induce ongoing inflammation that could contribute to As-associated disease [21].
More recently, a cDNA microarray study of PBMC from 10 individuals in Mexico having urinary As levels between 117.23 and 435.12 mg/g creatinine (n = 5) revealed significant differences in expression of apoptosisand inflammation-related genes compared to unexposed subjects (n = 5) [22]. Exposure was associated with down-regulated inflammatory genes, including TNF, IL11, IL10RB, CCR1, and CXCL2 [22], which is in stark contrast with up-regulated inflammatory genes reported in chronically-exposed persons in Taiwan [21]. However, the finding of decreased TNF concurs with data from the Bangladeshi study [20]. Some apoptosis-related genes were significantly up-regulated, including BCL2L1 and CASP2, whereas others, namely TRAIL and FASLG, were suppressed [22]. Such contradictory results could be influenced by differences in exposures, sampling, methodology, population genetics and/or environmental factors. For instance, the Mexican study [22] analyzed individual RNA samples from each subject, whereas the Taiwanese study [21] used pooled samples. Additionally, the small number of participants may not be large enough to draw definitive conclusions.
One caveat is that changes in expression may not necessarily be viewed as toxic responses; some changes may occur following chemical exposure as adaptive responses, e.g. acquired resistance to acute toxicity, over time [23]. However, changes reported at the gene and/or protein level are provided to illustrate possible effects resulting from As exposure that may alter cellular function and ultimately the immune system's ability to defend the host.

Lymphocyte activation
Impaired T-cell activation and functional responses have been observed in As-exposed persons. Analysis of 11 exposed and 13 control individuals in Mexico indicated that chronic exposure significantly decreased proliferation of mitogen-stimulated lymphocytes [24], which is supported by a later study identifying delayed cell cycle progression from S-to M-phase in chronically-exposed persons whose drinking-water levels averaged 412 μg As/L compared with persons consuming water averaging 37.2 μg As/L [25]. Similarly, a cross-sectional study in West Bengal, India of 18 controls and 20 As-exposed individuals with skin lesions found significant reductions in lymphoproliferation and Th1/Th2 secretion of IL-2, IL-4, IL-5, IL-10, IFN-γ and TNF-α in As-exposed compared with unexposed individuals [26]. In contrast, lymphocytes from Chilean copper smelter workers exposed to Ascontaminated air (n = 40) displayed higher proliferation rates than those of As-unexposed individuals from the same region (n = 32) [27]. While reasons for the observed differential effects on T-cell proliferation remain unclear, it is plausible that differences in population genetics, metabolism, and/or exposure could be involved. The study also found that urinary As levels positively correlated with oxidative stress markers serum superoxide dismutase (SOD) and lymphocyte malondialdehyde (MDA), and negatively correlated with serum vitamin E levels, suggesting that chronic exposure induces lymphocyte oxidative damage [27].

Lymphocyte subpopulations
Increasing evidence indicates that As affects numerous immune cell subpopulations. Subjects exhibiting dermatological symptoms (n = 30) from exposure to > 100 μg As/L for > 10 years in Eastern India demonstrated significantly increased eosinophil numbers and decreased monocyte counts compared to unexposed persons (n = 25) [28]. As also disrupts macrophage function: monocyte-derived macrophages from As-exposed individuals with skin lesions (n = 70) demonstrated cell rounding and significantly reduced adhesion, nitric oxide anion (NO -) production and phagocytic capacity compared to macrophages from non-exposed persons (n = 64) [29]. Down-regulated F-actin and CD54 adhesion molecule, and altered Rho A-ROCK signaling likely contributed to impaired macrophage function.
Investigation of As influence on immune regulation revealed that in chronically-exposed but otherwise healthy individuals (n = 47), urinary As levels (range, 8.1 -448 μg/g creatinine) significantly inversely correlated with the number and function of natural T regulatory (nTreg) lymphocytes but not other regulatory T-cells [30]. nTreg lymphocytes are CD25 + Foxp3 + T-cells that constitute approximately 10% of circulating CD4 + T-cells and play a critical role in immune homeostasis by suppressing immune response [31]. Although increased apoptotic PBMC were evident in exposed subjects, no significant correlation was found with urinary As levels, suggesting that the effect on nTreg cells was not mediated by apoptosis induction [30]. Alternatively, this could be due to individual differences in As susceptibility. Exposure positively correlated with monocyte innate immune receptor complex TLR4/CD14 and TNF-α secretion [30], which may be causally associated with As effects on nTreg lymphocytes given their inhibitory effect on pro-inflammatory TNF-α release [32]. These results concur with previous findings that chronic human As exposure induces inflammation, including CD14 [21].

Humoral immunity
Studies evaluating antibody levels in As-exposed individuals yield conflicting results. One study observed no changes in serum IgM, IgA or IgG in 47 adult male workers exposed to As in a coal-burning power plant compared to 27 workers from another plant in the same district whose As coal content was > 10 times lower [33]. It should be noted that exposure duration and internal As doses are unknown; thus, negative results could be due to acute or low-dose internal As levels. In contrast, Bangladeshi subjects (n = 125) chronically exposed to drinking-water As demonstrated significantly elevated serum IgA, IgG and IgE compared to unexposed persons [34]. IgG and IgE levels were significantly higher during initial stages of skin manifestations, and IgE continued to increase with prolonged exposure. Moreover, increased prevalence of respiratory complications including cough, chest sound, bronchitis and asthma were evident in exposed individuals, and mean serum IgE was higher in subjects with respiratory symptoms relative to exposed subjects without [34]. No effect on eosinophils was observed [34], in contrast with a reported As-associated increase in eosinophil numbers [28], suggesting that increased serum IgE may be due to direct inflammatory effects of As rather than allergic disease [34].

Pulmonary effects
Health outcomes of As immunosuppression are evidenced by increased prevalence of opportunistic infections such as tuberculosis and fungal and respiratory tract infections (RTI) [26,29,35]. A recent report from Chile revealed increased mortality from As-associated pulmonary tuberculosis [35]. Reports of As-associated pulmonary effects [26,29,[34][35][36] support growing evidence indicating that long-term exposure increases risk of reduced lung function and non-malignant lung disease [36][37][38][39][40][41]. Moreover, epidemiological investigations provide compelling evidence that As increases the incidence of bronchiectasis [42,43], a pulmonary disease characterized by chronic infection, inflammation, irreversible bronchial damage, and respiratory failure [44,45]. Chronically-exposed subjects from West Bengal, India with As-related skin lesions (n = 108) demonstrated a 10-fold higher prevalence of bronchiectasis compared with subjects without lesions (n = 150) [42]. A later report from Chile indicated elevated mortality rates for bronchiectasis in adults aged 30-49 resulting from early-life As exposure; compared with controls, mortality rates for those with childhood and in utero exposure were 12-and 46-fold higher, respectively [43]. Also observed were 6-to 7-fold increases in lung cancer mortality rates resulting from early-life exposures. Studies on this As-exposed Chilean population indicate long latency patterns of increased lung, kidney and bladder cancer mortality continuing for > 25 years after exposures ended [46,47]. Overall, these reports indicate that As not only exerts severe respiratory effects, but that early-life exposures have pronounced long-term consequences that may include higher prevalence of and mortality from cancers of different tissues. Intriguingly, women appear to be somewhat protected from skin and respiratory manifestations [36,48], possibly due to sex hormonerelated increased methylation capacity of As in women than in men [49].

HBD1 involvement
Interestingly, we previously reported in two As-exposed populations from Nevada and Chile a significant inverse correlation in men between urinary levels of As and antimicrobial peptide human β-defensin-1 (HBD1) [50]. Studies suggest a primary role for HBD1 against pulmonary pathogens relevant to bronchiectasis [44,45] and an association between HBD1 antimicrobial inactivation and recurrent airway infections in cystic fibrosis patients [51,52]. Further, observations from transgenic mice deficient in the mouse ortholog of HBD1 indicate that β-defensin-1 serves as an initial barrier to pulmonary bacterial colonization [53]. Given growing evidence that DEFB1, the gene encoding HBD1, is a putative tumor suppressor whose down-regulation may be involved in tumorigenesis of multiple tissues [54][55][56][57][58][59][60][61][62], it is tempting to speculate that HBD1 suppression may contribute to As-induced carcinogenesis or bronchiectasis. Although our ongoing studies demonstrate As-induced reductions in DEFB1 mRNA and protein in human cell lines (unpublished data), confirmatory evidence of HBD1 inhibition is needed from other As-exposed populations. Thus, it remains to be determined whether HBD1 is suppressed in lungs of As-exposed individuals, and further investigations are needed to elucidate the role of down-regulated HBD1 in As immunotoxicity and carcinogenicity.

Effects in children and infants
The fetus, infant and young child, each at critical stages in development, are particularly sensitive to stressors that could have short-and long-term effects. Yet, few epidemiological studies have investigated the influence of early-life As exposure on immunological outcomes in children and even fewer in newborns and infants. Evidence indicates that early-life As exposure may have consequences that manifest much later in adulthood [18,63], as evidenced by increased prevalence of and mortality from bronchiectasis and lung cancer in young adults [43]. Therefore, biomarkers indicative of future disease following early-life exposure could be evident in young subjects.

Induction of apoptosis
Indeed, studies of early-life As exposure have detected markers of immune dysfunction in infants and children. Studies of Mexican children aged 4-13 have reported higher incidences of apoptotic PBMC in As-exposed children relative to controls [64,65]. Although apoptosis is important in immune homeostasis, abnormal immune cell apoptosis can contribute to dysregulated immune function, which may result in immunodeficiency, autoimmune disease or malignant transformation [66]; thus, induced apoptosis may be important in As-mediated immunosuppression. The larger study of 40 children (high and low mean urinary As levels = 46.3 and 14.2 μg/g creatinine, respectively) found a significant positive association between As exposure and apoptotic PBMC [65]. However, despite elevated apoptotic PBMC in chronicallyexposed children from the smaller study of 7 highlyexposed and 5 non-exposed children (mean urinary As levels = 143.9 and 24.8 μg/g creatinine, respectively), no significant correlation was observed between exposure and apoptotic cells [64], in agreement with a study on adults [30], possibly due to small sample size or individual differences in As susceptibility [64].

Lymphocyte activation
Consistent with findings from adults [24][25][26], significant reductions in PBMC IL-2 secretion and proliferation were observed in As-exposed children aged 6-10 (n = 90, mean urinary As levels of high-and low-exposure group = 194.9 and 29.3 μg/L, respectively) [67]. Also noteworthy were increased granulocyte-macrophage colony stimulating factor (GM-CSF) secretion and reduced CD4 + cell count and CD4/CD8 ratio without altered CD8 + cell proportion [67]. Because low CD4/CD8 is considered a surrogate marker of immunosuppression [68][69][70], the observed decrease in CD4/CD8 may be an early indicator of As-mediated immunosuppression. Furthermore, the increased GM-CSF secretion may indicate chronic inflammation given growing evidence of elevated GM-CSF levels in initiating/mediating chronic inflammation [71], and is consistent with a previous study of As-exposed adults demonstrating up-regulated inflammatory molecules [21].

ROS production
Production of the reactive oxygen species (ROS) NOand superoxide anion (O 2 -) by activated PBMC is an important innate immune response to destroy invading microbes. Cross-sectional studies assessing As influence on ROS production in children have yielded conflicting results. Analysis of 87 children in Mexico ingesting Ascontaminated water showed that exposure positively associated with O 2 production by mitogen-stimulated monocytes and basal NOand O 2 levels in PBMC and monocytes [72]. This is inconsistent with an earlier study of 65 children living near a primary smelter in Mexico in which As exposure was negatively associated with NOand O 2 production by stimulated monocytes [73]. The discrepancy may be due to differences in exposure; children in the earlier study had lower urinary As levels (range 16.7-465.7 μg/g creatinine) [73] than those in the more recent study (range 12.3-1411 μg/g creatinine) [72]. Regardless of the source of variation in results, these studies suggest that As could alter circulating cells' ability to respond to immunological challenge. For example, elevated ROS levels in un-stimulated PBMC indicate Asinduced oxidative stress, concurrent with findings from copper smelter workers [27]; ROS overproduction by activated PBMC could cause oxidative damage to surrounding tissues, whereas diminished ROS production could weaken PBMC defense against pathogens.

Prenatal exposure
Because As readily crosses the placenta [74], it could potentially alter prenatal development. Indeed, gestational As exposure is linked to increased fetal loss and infant mortality [75,76]. However, reports on immune-related effects of prenatal exposure in newborns and infants are scarce. In a mother-child cohort study in Bangladesh (n = 140), maternal urinary As levels were significantly negatively correlated with child thymic index and breast milk trophic factors IL-7 (needed for thymic and T-cell development) and lactoferrin (an antioxidant and factor in innate immunity), and positively correlated with maternal morbidity and male infant RTI [77]. These findings are supported by a more recent prospective population-based cohort study of 1,552 infants born in Bangladesh, which revealed dose-dependent increases of 69% and 20% in infant lower RTI and diarrhea, respectively, related to exposure during pregnancy (maternal urinary As levels, lowest quintile < 39 μg/L; highest quintile = 262-977 μg/L) [78]. The observed increased prevalence of infant respiratory illness is consistent with As-associated adult non-malignant lung disease [35,42] and marked increased risk of such disease following early-life exposures [43]. Moreover, enhanced male infant susceptibility to RTI is consistent with increased As-related pulmonary effects in men and not in women [36].
Another study of women delivering babies in Bangladesh (n = 130) found that gestational As exposure induced placental inflammation (IL-1β, TNF-α and IFN-γ) via oxidative stress (8-oxoguanine), reduced placental CD3 + T-cell numbers, and increased umbilical cord blood IL-8, IL-1β, TNF-α and IFN-γ [79]. These findings concur with reports of elevated oxidative stress [27] and inflammation [21] in chronically-exposed adults. In a follow-up study (n = 44), As levels in maternal urine and placental and cord blood positively associated with cord blood 8-hydroxy-2'-deoxyguanosine and inversely associated with infant thymic function at birth, as measured by signal-joint Tcell receptor-rearrangement excision circles in cord blood mononuclear cells (CBMC) [80]. Further, prenatal As exposure was associated with down-regulated oxidativestress defense genes, including SOD3, and up-regulated apoptosis-related genes in CBMC, including CASP2 [80], the latter consistent with results from adults [22].
Overall, these data indicate that in utero As exposure reduces infant thymic size and function, likely through inhibiting breast milk trophic factors and/or inducing apoptosis and oxidative stress. These effects may contribute to infant immune deficiency evidenced by increased RTI prevalence. Lack of data supporting a relationship between early-life As exposure and nonpulmonary infections suggests that the developing lung is specifically targeted by As. Furthermore, given increasing evidence of As-associated adverse immunerelated outcomes, it is likely that immune disruption resulting from early-life As exposure will have long-term detrimental consequences well into adulthood, as seen in increased prevalence of bronchiectasis and lung, kidney and bladder cancers.

Gene expression
In various animal models, As exposure is associated with altered expression of genes involved in immune response. In lungs of mice exposed to As III (< 100 ppb) for 5-6 weeks, significant changes were identified in transcripts encoding humoral immune response, antigen binding, TLRs, cytokines, cytokine receptors and genes involved in cell adhesion and migration [81,82]. Specifically, down-regulated expression of genes encoding TLR/IL1R signaling pathway, including Il1b, was identified [82]. In zebrafish embryos, As significantly inhibited induction of genes involved in regulating innate immune responses against viral and bacterial infection, including il1b [83], tnfa, ifnphi1 (type1 interferon) and mx (interferon-inducible Mx) [84,85]. As also disrupted JAK/ STAT pathway, which is critical in cytokine regulation [84]. These effects concur with epidemiological findings of As-associated decreased expression of IL1B [19,20] and TNF [22].

Lymphocyte activation
Consistent with epidemiological observations [24][25][26]67], chronic As exposure inhibits mitogen-stimulated proliferation of PBMC and SMC in broiler chickens [91] and SMC in mice [92], and T-cell and B-cell proliferation in catfish spleen and HK [90,93]. Consequently, decreases have been observed in secretion of IFN-γ, IL-2, IL-6 and IL-12 in mice [92], and "IL-4-like factors" from HK T-cells in catfish [93]. An important consideration regarding animal studies is that As concentrations administered typically far exceed human exposures, which may account for differential effects observed.

Humoral and hypersensitivity responses
As can inhibit humoral immunity, as evidenced by suppressed in vitro primary and/or secondary antibodyforming cell (AFC) responses of rodent splenocytes [87,89,[94][95][96][97]. IL-2 is a primary target of this inhibition in mice [98]. Further, As suppressed delayed-type hypersensitivity reaction, a response to cutaneous sensitization, in mice [94,99], rats [88,100] and chickens [91]. Compared to controls, As III -exposed sensitized mice demonstrated reduced lymph node cell proliferation, ear swelling, activated Langerhans cells (LC) in cervical lymph nodes, peritoneal macrophages and circulating neutrophils [99], suggesting that As inhibits LC migration to lymph nodes and subsequent T-cell activation.

Macrophages
Similar to humans [29], As exposure in animals suppresses macrophage production of NOand/or O 2 -[91,101-103], release of TNF-α [104], and phagocytosis [90,102,103]. In chronically-exposed animals, such effects may be longterm. Exposure of freshwater bivalve L. marginalis to As III (1-5 ppm) for < 30 days resulted in time-and dosedependent decreases in phagocytic efficiency and NOproduction in haemocytes, the major phagocytes and immune-effector cells in bivalves [102]. In a recovery assay, animals were maintained in As-free water for the same duration as exposure to evaluate immune efficiency. They demonstrated partial recovery of phagocytic potential, but inhibitory effects were still apparent; whereas NOproduction was restored to control levels in animals exposed to 1 ppm As III for < 4 days, NOgeneration remained suppressed in high-dose-and long-term-exposed animals [102]. As can also induce apoptosis in macrophages, as seen in 3-fold increased DNA fragmentation in splenic macrophages from As III -exposed mice [103]. Further, splenic macrophages from As-treated mice demonstrated reduced adhesion and chemotactic index [105], surface I-A k (MHC) class II molecule expression and antigen presentation to T-cells [106]. Similar observations have been made in humans [19,29], thus giving these results biological plausibility.

Immune surveillance
Altogether, these data indicate that in vivo As exposure can disrupt innate and humoral immunity. Studies of As influence on allogeneic immune response, i.e. rejection of MHC-mismatched allografts, suggest As disrupts the immune system's ability to distinguish "self" from "non-self" [107,108]. In a mouse heart transplantation model, arsenic trioxide (As 2 O 3 ) significantly reduced allograft rejection relative to control [107]. Similarly, As 2 O 3 radically reduced severe symptoms of graft-versus-host disease in mice following allogeneic hematopoietic stem cell transplantation [108]. Together with reported decreased macrophage I-A k class II expression and antigen presentation [106], these findings provide potential mechanisms whereby As can suppress the immune system's ability to discriminate self from non-self antigens.
Investigation of As influence on immune system regulation revealed redistributed nTreg lymphocytes following 3-week As III exposure in a rat model of multiple sclerosis, an autoimmune disease characterized by decreased nTreg cell number and function [30]. Whereas low As III doses increased nTreg cell number in spleen and alleviated severity of the autoimmune condition, concentrations > 100 μg/L reduced cell numbers in blood and spleen, consistent with epidemiological findings from that study [30]. Evidently, lowdose As-mediated increased number of splenic nTreg lymphocytes inhibited generation of (auto-) immune responses, hence the beneficial effect of immunosuppression by low-dose As. Thus, two possible scenarios whereby As can interfere with self/non-self recognition exist: by i) preventing immune surveillance from recognizing "non-self" from "self", leading to increased non-self antigen survival, as in allograft transplantation; or ii) inhibiting recognition of self antigens as "self", which could arise from As-induced reduction of nTreg cell inhibitory activity, leading to "anti-self" antibody production indicative of autoimmune disease. Such effects would likely render the host immunocompromised and could have detrimental health consequences.

Lymphocyte activation
In vitro As exposure suppressed IL-2 secretion and proliferation of mitogen-stimulated lymphocytes from humans and various animal species [113][114][115][116][117][118]. A biphasic dose-dependent response was observed following As III or As V exposure of mitogen-stimulated human and bovine PBMC [119], demonstrating As immunosuppressive effects depend on the dose. As markedly suppressed lymphocyte secretion and/or mRNA levels of IFN-γ, IL-4 and IL-10 in different in vitro models [114,116,117]. As III also significantly impaired differentiation of human Th17 cells by repressing their expression and release of IL-17 and decreasing expression of RORγt, which regulates IL-17, through inactivation of JNK/c-Jun pathway [120]. As III further impaired Th17 cells by disrupting functions of DC, which regulate Th17 cell differentiation, via i) blocking DC differentiation through induced necrosis; ii) decreasing DC endocytotic activity; iii) repressing secretion of IL-12p70 and IL-23, two major regulators of Th17 activities, by activated DC; and iv) reducing ability of activated DC to stimulate IFN-γ and IL-17 release from Th17 cells [121]. ↑ Pathogen colonization & ulcers/septicemia following bacterial infection in catfish [90,93] ↓ Splenic clearance of S. aureus in mice [105] ↑ Morbidity & respiratory viral titers following H1N1 viral infection in mice [110] Pulmonary health

Induction of apoptosis and humoral immunity
As immunosuppressive effects are further demonstrated in vitro by increased apoptotic rates in B-cells, T-cells, macrophages and neutrophils [64,[123][124][125][126][127]. Prior to inducing apoptosis in TA3 mouse antigen-presenting B-cells, As III inhibited activity of lysosomal protease cathepsin L, which is important in antigen processing/presentation to elicit T-cell responses [125]. This could possibly potentiate As-induced suppression of humoral immunity, for in vitro As exposure inhibits AFC responses of mouse splenocytes [128,129], consistent with animal studies.

Pulmonary effects
Proposed mechanisms for As-associated compromised respiratory immunity include impaired pulmonary alveolar macrophage (PAM) function, demonstrated by markedly reduced LPS-stimulated TNF-α and O 2 production [104], and decreased airway epithelial (AE) chloride secretion cystic fibrosis transmembrane conductance regulator (CFTR), an essential chloride channel for mucociliary clearance of pathogens, as seen in human CFBE41o-AE cells [135] and killifish gills [136]. In human 16HBE14o-AE cells, As restricted wound response, i.e. monolayer reformation following scraping of cultures, through MMP9 up-regulation [137] and inhibition of paracrine Ca 2+ signaling [138], consistent with downregulated adhesion-and migration-related genes in lungs of mice [81,82] and altered airway proteins in mice and humans [139][140][141]. Collectively, these data suggest As disrupts pulmonary defense through mechanisms involving i) altered PAM function, ii) decreased AE chloride secretion resulting in depressed clearance of pathogens, and iii) AE remodeling due to impaired wound response, ultimately promoting chronic lung diseases such as bronchiectasis.

Discussion
Studies reviewed here show that As significantly impacts both innate and adaptive immune defenses. Likely mechanisms involve altered expression of key immune regulators, induced apoptosis, oxidative stress and inflammation in circulating PBMC, impaired lymphocyte activation and macrophage function, and altered cellular and humoral immunity (Table 1). Specific examples of concordance between epidemiological and experimental data are i) reduced expression of MHC class II molecules, CD69, IL-1β and TNF-α; ii) altered expression of airway adhesion-and migration-related genes/proteins; iii) decreased stimulated lymphocyte proliferation and IL-2 secretion; iv) impaired macrophage adhesion, phagocytosis and stimulated ROS production involving altered Rho A-ROCK signaling; v) induced apoptosis of PBMC; and vi) decreased stimulated ROS production by PBMC (Table 1). These effects can result in immunosuppression, as evidenced by reduced microbial clearance in animals and increased prevalence of opportunistic infections in humans, particularly RTI. Furthermore, epidemiological data suggest marked susceptibility of the lung to perturbation by As, especially during prenatal and childhood development, which results in unprecedented rates of chronic lung diseases, notably lung cancer and bronchiectasis. Experimental data suggest that such pulmonary effects could involve disrupted PAM function and airway remodeling resulting in impaired clearance of pulmonary pathogens. The pleiotropic effects of As on the immune system, including specific examples of compromised immune surveillance such as decreased rejection of MHC mismatched allografts and reduced migration of PBMC, neutrophils and DC to sites of infection in various animal models (Table 2), lend biological plausibility to increased rates of infection, cancers and other immune-related illnesses observed in As-exposed human populations, and are illustrated in Figure 1.

As As
Infection Cancer Lung Disease Reduced Immune Surveillance Oxidative stress Inflammation Apoptosis As exposure Figure 1 Simplified scheme of select As-induced alterations of immune responses. Also illustrated is how such effects might contribute to development of As-associated illnesses. Antigen (Ag) can be foreign or tumor cell.

Conclusions
In vivo and in vitro studies depict As as an immunomodulator that could render the host immunocompromised. Such immune alterations could help explain increased risk of infections and several cancers observed in chronically-exposed human populations. As-mediated alterations of cellular and humoral immunity reported in animal and in vitro models generally agree with immunological outcomes in humans. However, more work is needed to close the gap between experimental data and risk of human immunotoxicity. Moreover, inconsistencies in epidemiological findings, possibly due to differences in dose, sampling, genetic background, and environmental/ nutritional factors, indicate need for larger participant numbers and diverse ethnic populations. Due to differential effects of exposures, populations having low, intermediate and high exposure should be evaluated to better understand dose-dependent relationships. Furthermore, strong evidence for an association between developmental As exposure and elevated risk of human disease necessitates more investigations of early-life exposure outcomes. Finally, comprehensive genomic, proteomic and metabolomic profiling will be critical for identifying and validating potential molecular targets of As to monitor progression of As-associated diseases and elucidate mechanisms of As immunotoxicity.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions NLD, CFS and MTS conceived of the study; NLD reviewed the literature and wrote and edited the manuscript; CFS and MTS critically reviewed the manuscript. All authors read and approved the final manuscript.