Arsenic immunotoxicity: a review

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 genome-wide 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 apoptosis- and 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 As-contaminated 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, 34–36] support growing evidence indicating that long-term exposure increases risk of reduced lung function and non-malignant lung disease [36–41]. 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 hormone-related 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 [54–62], 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.

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 chronically-exposed children from the smaller study of 7 highly-exposed 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 [24–26], 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 [68–70], 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) NO^- and superoxide anion (O₂ ^-) 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 As-contaminated water showed that exposure positively associated with O₂ ^- production by mitogen-stimulated monocytes and basal NO^- and O₂ ^- 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 NO^- and O₂ ^- 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 As-induced 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 T-cell receptor-rearrangement excision circles in cord blood mononuclear cells (CBMC) [80]. Further, prenatal As exposure was associated with down-regulated oxidative-stress 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 non-pulmonary infections suggests that the developing lung is specifically targeted by As. Furthermore, given increasing evidence of As-associated adverse immune-related 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

In vitro As exposure suppressed IL-2 secretion and proliferation of mitogen-stimulated lymphocytes from humans and various animal species [113–118]. 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].

In contrast with reduced CD4⁺ and unaltered CD8⁺ T-cell populations in children [67] and mice [92], in vitro As^III treatment of mitogen-stimulated mouse T-lymphocytes decreased CD8⁺ counts without affecting CD4⁺ counts [115]. Additionally, As^III inhibited early activation of mouse CD4⁺ and CD8⁺ cells, as evidenced by reduced surface CD69 expression [115], an effect that was reported at the protein level in in vitro As-exposed human CD4⁺ and CD8⁺ lymphocytes [122] and also at the mRNA level in lymphocytes of As-exposed humans [19].