A historical perspective of immunotoxicology

Abstract

An historical perspective of immunotoxicology is presented beginning from early observations in which exposure to workplace environments led to unexpected immune-mediated lung diseases to its eventual evolution as a sub-discipline in toxicology. As with most toxicology disciplines, immunotoxicology originated from concerns by scientists within industry and government as well as medical professionals to limit human exposure to agents that can potentially effect human health. The basis for these concerns originated from laboratory studies in experimental models and clinical observations that suggested certain industrial and agrochemicals, pharmaceuticals and consumer products were capable of inadvertently interacting with the immune system and cause adverse health effects. The types of immunopathologies observed and mechanisms responsible were found to be broad, being dependent upon the physiochemical properties of an agent, exposure route, and target organ/tissue, and included allergic/hypersensitivity responses, immune dysfunction, manifested by suppression or, in rare instances, stimulation, autoimmune or autoimmune-like diseases, and chronic inflammatory disorders.

• Environmental immunology
• historical perspective
• immunotoxicology

Introduction

Although the term “immunotoxicology” was first coined in the 1970s, it has been known for centuries within the medical community that certain xenobiotics can adversely affect the immune system. The very earliest observations focused on pulmonary immune diseases and, although the mechanisms were unknown, affected individuals often developed severe fibrotic lung disease. It was not until the latter half of the 20th Century, when the chemical properties of these xenobiotics were determined and the immune system was better characterized, that it was recognized that these materials represented a broad spectrum including environmental chemicals, air particulates, and pharmaceutical drugs and biologics, and that various immune-mediated diseases could result including chronic inflammation, allergic or hypersensitivity reactions, immune dysfunction (most notably suppression), or autoimmune disease.

Occupational/environmental immune-mediated lung diseases

Occupational/environmental immune-mediated lung diseases represent a dynamic and heterogeneous group of diseases induced by a wide variety of causative agents. Although asbestosis was observed in miners in ancient Egypt, it was not until 1713 that Bernardo Ramazzini described the health hazards associated specifically with 52 occupations including, among others, bakers and handlers of flax, grain silk, and hemp (Ramazzini, 2001). It was not until the 20th Century, however, that the causative agents were first identified and lung immunity actually implicated. These immune-mediated environmental lung diseases can be grouped, according to Coombs and Gell (1968) classifications, as those that evoke Type 1 reactions (i.e. IgE), often referred to as occupation asthma, those that evoke Types 3 or 4 reactions, such as chronic beryllium disease and hypersensitivity pneumonitis (HP), and those mediated primarily by innate, rather than adaptive, immunity and that represent chronic inflammatory lung disease.

Occupational asthma

Following an analysis of 150 unpublished cases, Henry Slater first described many of the characteristic features of occupational asthma, describing it as “hyper-responsiveness provoked by exposure to chemical and mechanical irritants, as well as to particular atmospheres” (Salter, 1866). However, it was not until later in that century that Ehrlich identified the presence of eosinophils in the sputum of workers, a hallmark of immune-mediated asthma (reviewed by Hirsch et al., 1980). Subsequently, numerous agents, either proteins or reactive small molecular weight chemicals, have been shown to act as allergens and elicit IgE-mediated asthma. The foundations for our understanding of allergic responses to low molecular weight chemicals originated from observations by Landsteiner and Jacobs (1935) who showed that some highly reactive chemicals are capable of binding to host proteins (i.e., haptens) acting as antigens and evoking an immune response. Equally important was the work of Coombs et al. (1968) that showed IgE was associated with allergy from castor beans. Mapp (2005) estimated that up to 20% of adult onset asthma is caused by occupational factors and that roughly 90% of these cases involve immunological mechanisms. The isocyanates and anhydrides chemical classes are common causes of occupational asthma, with over 200 000 workers in the US exposed every year and 5–10% of them developing asthma. Some of the more notable work-related protein allergens include alanase, an enzyme found in soap detergent, latex proteins, and flour, the cause of baker’s asthma (e.g., Baur & Posch, 1998; Dry et al., 1989; Schweigert et al., 2000).

Animal and human research into various aspects of occupational asthma continues to be an active pursuit in immunotoxicology, with current focus on further clarification of cellular and molecular mechanisms that participate in the response, the role of gene–environmental interactions, understanding changes responsible for shifting demographics in population frequency, and development of a simple predictive screening test to identify potential respiratory sensitizers. More recently, based on epidemiological observations of increasing asthma rates in industrialized cities, it was found that many common air pollutants exacerbate existing asthma by acting as adjuvants and enhance asthmatic responses to common allergens, such as pollen and house dust mites. These seminal observations were followed by another major observation that suggested co-exposure to endotoxin early in life regulates the response to other allergens later in life (reviewed by Selgrade et al., 2008). The latter observations were first suggested by epidemiological observations that children in frequent contact with farm animals and livestock feed have a marked reduction in asthma frequency and has led to the “hygiene hypothesis” which states that exposure to endotoxin-containing microorganisms early in life allow for the generation of an appropriate immune balance (Strickland & Holt, 2011).

Hypersensitivity pneumonitis (HP)

Campbell (1932) described what is now referred to as hypersensitivity pneumonitis in a group of workers from a Michigan plant that made railroad ties. It was later determined that this lung disease was caused by the fungus Cryptostroma corticale. The agents most often responsible for HP are diverse and include microbes, animal and plant proteins, and low-molecular weight chemicals that act as haptens. HP is relatively common among workers that are exposed to significant concentrations and/or for prolonged periods of time to an allergen with estimated frequencies as high as 5–15% of individuals with such exposure (Hirschmann et al., 2009). Of the many forms of HP, Farmer’s Lung and Pigeon-Breeder’s Lung are the most common. Acute forms of the disease are represented by Type 3 and 4 reactions and usually involve immune complexes and complement activation (Fink, 1976; Lopez & Salvaggio, 1976). Chronic HP is marked by interstitial T-cells and macrophages, as well as non-caseating granulomas that can lead to significant alveolar destruction and parenchymal fibrosis. In rats, the disease can be transferred with T-cells, suggesting an important role for cell-mediated immunity (Richerson et al., 2005).

Chronic beryllium disease (CBD)

CBD is a Type 4 reaction to beryllium. The key pathological manifestations are similar to sarcoidosis and include granulomas composed of epithelioid cells along with macrophages and T-cells. The pathogenesis was first described in 1951 (Sterner & Eisenbud, 1951), but was met by skepticism by the medical and scientific communities for over 15 years. Hanifin et al. (1970) showed the beryllium oxide stimulated lymphocytes from sensitized, but not sensitized patients, which supported a role for adaptive immunity and also led to the beryllium lymphocyte transformation test (BeLPT) that is still used to help establish a clinical diagnosis. The findings that the primary HLA Class II molecules involved in antigen presentation are HLA DPB1 with glutamic acid at position 69 (Glu69) and that the majority of patients with beryllium sensitization (BeS) or CBD possess this allele has led to one of the first genetic screening tests used in industry to help identify those at high risk for disease (Richeldi et al., 1993).

Chronic inflammatory lung diseases

There are innumerable examples of agents that activate innate immunity without significant involvement of adaptive immunity. When the material persists or causes excessive tissue damage, chronic, in contrast to acute inflammation, can result which often leads to significant pathology. Agents that produce chronic inflammation often target the lung, but the liver and skin may also be affected and it is likely that chronic inflammation can occur in any organ where macrophages reside or infiltrate. One early example is byssinosis. Byssinosis (aka brown lung disease) is caused by exposure to endotoxin in cotton dust and often occurs in inadequately ventilated working environments (Hollander, 1953). The earliest observations of environmental agents that induce chronic inflammatory conditions were in miners exposed to natural fibers, including many of the polymorphs of asbestos and silica, and which often lead to fibrotic lung diseases. Many of these agents interact with a family of innate immune receptors, known as nucleotide-binding domain leucine-rich repeat containing (NLR) inflammosomes, and interfere with their normal function (Dostert et al., 2008; Wilson & Cassel, 2010).

Although an association between asbestos and lung disease was identified over 2000 years ago in asbestos miners from ancient Egypt, it was not until the early 1900s that a physician diagnosed asbestosis in a 23-year-old woman who had been working with asbestos since she was 13 (Cooke, 1927). Subsequently, a study conducted on asbestos workers in England showed that nearly a quarter of them had asbestos-related lung disease (Selikoff et al., 1967). It was not until the mid-20^th Century that animal models were developed and researchers started to unravel to complex events responsible for asbestosis. The general consensus is that the pathogenesis of asbestosis is derived from the long-term interplay between persistent free radical production and the expression of cytokines, growth factors, and other inflammatory cell products that ultimately leads to excessive deposition of collagen and other extracellular matrix components by mesenchymal cells resulting in fibrotic lung disease (Kamp & Weitzman, 1999). Cumulative observations from animal studies beginning in the 1970s suggested that asbestos could also interfere with adaptive immunity (Maeda et al., 2008; Miller et al., 1981). Although evidence for causality remains lacking, these data have led to suggestions that immunosuppression may contribute to the occurrence and progression of asbestos-related malignant diseases (Maeda et al., 2008).

Silicosis, caused by inhalation of silica dust, like asbestosis is a fibrotic lung disease triggering an inflammatory cascade culminating in fibrosis and progressive impairment of lung function (Dostert et al., 2008; Leung et al., 2012). However, they differ in their threshold dose for diseases as well as histological appearance with silicosis manifested as simple (nodular) silicosis, progressive massive fibrosis, silicoproteinosis or diffuse interstitial fibrosis (IARC, 1997). As with asbestos, early studies focused on the role pulmonary macrophages and their scavenger receptors play in the disease process (Hamilton et al., 2008). Notable highlights of silicosis immunotoxicology research have included elucidating the role of lipopolysaccharide (LPS) priming in the cytokine signaling pathway and the role of caspase-1 in mediating secretion of fibrotic proteins (Leung et al., 2012).

Allergic contact dermatitis (ACD)

Interest in ACD originated from observations that many industrial, therapeutic, and consumer products that came into contact with the skin frequently were responsible for dermatitis. This stimulated government agencies, industry, and medical professionals to try to limit their exposure through animal testing. The first definition of a real predictive test was provided by Draize et al. (1944). Since that time numerous protocols have been described whose aims have been, in one way or another, to make improvements to the sensitivity and predictivity using guinea pigs as a surrogate for man. All of these test protocols followed similar principles; a combination of intra-dermal and/or epicutaneous treatments is administered to guinea pigs over a several week period in an attempt to induce skin sensitization, then a 1–2 week rest period to allow for an immune response to mature, followed finally by a topical challenge to assess the extent to which skin sensitization might have been induced. Sham-treated controls are also challenged for comparison. Evaluation of skin reactions was usually by subjective visual assessment 24–48 h after challenge application, the main reaction element being erythema. The protocols of Magnusson and Kligman (1969) and Buehler (1965) have been the two most studied and accepted guinea pig methods used for regulatory purposes worldwide (OECD, 2002).

Over the last several decades, the Local Lymph Node Assay (LLNA) has replaced traditional guinea pig models and is routinely used as a validated alternative approach for skin sensitization testing. The method evokes lymph node cell proliferative responses induced in mice following repeated topical exposure to a test material as a relative measure of sensitizing potential (Kimber et al., 1986). In part due to the improved animal welfare benefits, the LLNA has become the preferred method for assessing skin sensitization hazard by various regulatory authorities in most developed countries. An important point was that the LLNA was subjected to rigorous independent scrutiny and validated by the International Coordinating Committee on the Validation of Alternative Methods (ICCVAM) (Dean et al., 2001) and a similar endorsement by the European Centre for the Validation of Alternative Methods (ECVAM) (Balls & Hellsten, 2000). With the recent adoption in Europe of limitations on animal testing, alternative in vitro methods are increasingly being explored (Mehling et al., 2012).

Immundysfunction (suppression and enhancement)

In contrast to hypersensitivity/allergic reactions, it was not until the early 1970s that evidence began accumulating demonstrating that exposure to certain agents produced immuno-dysfunction. In almost all cases, dysfunction was associated with suppression, although in rare instances, such as in the case of hexachlorobenzene, immunoenhancement or stimulation was observed The earliest classes of immunosuppressive chemicals to be studied included, among others, heavy metals, halogenated aromatic hydrocarbons, drugs of abuse, and air pollutants (Gardner et al., 1977; Koller, 1973; Munson et al., 1976; Vos et al., 1973). A number of these chemicals, such as TCDD, are still being studied today to establish their specific mechanisms of action (Kerkvliet, 2012). Of particular interest was that in some of these studies that were conducted in perinatatally-exposed animals, effects were found to be more severe or persistent than those arising from an exposure of adult hosts. This has led to suggestions that the developing immune system is particularly sensitive to chemical-induced immunosuppression (Burns-Naas et al., 2008; Luebke et al., 2006).

While initially limited to experimental animal models, subsequent epidemiological studies, although often cross-sectional in nature and of limited statistical power, supported many of these experimental observations. The health outcomes most commonly observed were increased incidences of certain cancers, such as non-Hodgkin’s lymphoma and common infections such as respiratory infection or otitis media, the latter in children (e.g. Kramer et al., 2012; Lee & Chang, 1985; Luebke, 2002). Establishing a direct link between exposure and disease manifestations for immunosuppression in humans, however, remains difficult because of the inherent limitations of epidemiological studies to draw causal conclusions, particularly for common diseases like respiratory infections.

Although the experimental methods initially adopted by immunotoxicologists to assess immunosuppression in animals were those common to most immunology laboratories, the tests that were commonly performed and the experimental design by which they were conducted were performed ad hoc. The lack of standardized testing made it difficult to compare chemical-specific effects and led Dean et al. (1982), after sponsoring several workshops, to suggest a “Tier” approach with the idea that each subsequent tier provided an opportunity to better define a specific target within the immune system. Two major points were agreed upon from these workshops. First, since the immune system is not fully operational until it was challenged, the most appropriate tests would be those that incorporate an antigen challenge and, second, since it may be construed that an inadequate response to antigenic challenge does not represent an “adverse effect”, tests should also be added which could better relate to human disease. The former recommendation highlighted several common assays, including measurement of an antibody response following antigen challenge as a measure of humoral immunity and quantification of delayed hypersensitive response (DHR) or cytotoxic T-lymphocyte response (CTL) as a measure for cell-mediated immunity. To address the need to identify a clear adverse effect, tests, referred to as “host resistance assays”, were suggested that involved challenging groups of control and treated experimental animals with low levels of infectious agents or transplantable tumor cells. These assays helped predict the potential for environmental agents to alter host susceptibility in the human population as well as to validate the relevance of the immune tests. An inter-laboratory validation effort was sponsored by the NTP using this tiered testing panel (Luster et al., 1988). This effort was followed several years later by an analysis using a relatively large database to determine the concordance between the various histological, hematological, immune function, and host susceptibility assays (Luster et al., 1992, 1993). These studies were important, not only as a validation exercise for tier testing, but for providing a basis for risk assessment using immunotoxicology data. Subsequent validation studies were conducted in both mice and rats and, for the most part, results from the two species were found to be comparable (Ladics et al., 1998; van Loveren & Vos, 1989; White et al., 1994).

Tiered screening panels have been the basis for several risk assessment guidelines and most regulatory agencies in the US, European Union, and Japan have established or are developing requirements or guidelines (House & Luebke, 2007). However, the configurations of these testing panels vary depending upon the agency/organization/program under which they are conducted. The most notable difference being whether a functional immune test (i.e. incorporates antigen challenge) is included in Tier 1 or Tier 2.

Less effort has been devoted to establishing a predictive test panel to assess the immune system of humans. The United States National Academy of Sciences (NAS) and the International Program on Chemical Safety of the World Health Organization (IPCS/WHO) have proposed a three-tier testing scheme to be used for epidemiologic studies of known or suspected immunotoxic agents (NRC, 1992; WHO, 1996). Neither the sensitivity nor the predictivity of these test panels have been established, although foremost amongst these tests include quantitative measure of the immune response to a vaccine.

Autoimmunity

In most instances, the events that initiate an autoimmune response to self are unknown, although intrinsic factors (e.g. specific gene polymorphisms, sex-related hormones, age) and extrinsic factors (e.g. lifestyle, infectious agents) are associated with the induction, development, and exacerbation of autoimmunity. Over the past 40 years, epidemiological studies have provided evidence that exposure to certain drugs or environmental agents represent a significant contributing factor for the onset, remission, or severity of autoimmune diseases (see reviews by Bigazzi, 1997; Miller, 2006; IPCS, 2006, 2012). Although some of these claims, like silicone breast implants, turned out to be false, they brought attention to the need to have better predictive models to evaluate the potential for induction or exacerbation of autoimmune diseases. Detailed discussions of endpoints and methods that may be useful in characterizing autoimmunity were provided in EHC 236 (IPCS, 2006). The lack of standardized methods lie in direct contrast to many excellent studies in which autoimmune-prone animal models, such as the MRL^+/+ mouse and Lewis rat, have served to help characterize the cellular and molecular events responsible for chemical- or drug-induced autoimmune diseases. Central to many of these studies has been establishing the influence of circulating immunocompetent T-cells, particularly the activation of CD4⁺25⁺FoxP3⁺ regulatory (T_reg) and responder (T_resp) T-cells (e.g. Gilbert et al., 1999; Hayashi et al., 2011).

Environmental-induced autoimmunity

Since the observation by Pernis et al. (1965) of an increased prevalence of Rheumatoid Factor in asbestos-exposed individuals, there has been a growing body of epidemiological evidence that occupational exposure to fibrogenic fibers, as well as some heavy metals and organic solvents, are associated with systemic autoimmune diseases. Epidemiological studies have established that exposure to silica-containing mineral dusts are associated with elevated risk for rheumatoid arthritis (known as Caplan’s syndrome), systemic sclerosis, systemic lupus erythematosus, and anti-neutrophil cytoplasmic antibody (ANCA)-related vasculitis/nephritis (IPCS, 2012). Although some investigators have considered these as adjuvant effects, silica can influence circulating immunocompetent cells, particularly T_reg and T_resp cells. Correlations between asbestos exposure and autoimmune disease also suggest a higher-than-expected risk of systemic autoimmune disease among asbestos-exposed populations (Bunderson-Schelvan et al., 2011), although in general this association is less compelling than that found with silica. An animal model has been developed for asbestos-induced autoimmune disease, which is significant since the method employs normal rather than genetically prone mice (Pfau et al., 2008).

There are a large number of studies that have suggested an association between solvent exposure, predominantly work-related, and various autoimmune or autoimmune-like diseases. For example, a meta-analysis, looking at 10 epidemiological surveys, showed a fairly consistent but modest association (RR = 3.14; 95% CI = 1.56–6.33) between solvent exposure and systemic sclerosis or connective tissue disorders that are autoimmune in nature (Aryal et al., 2001). Epidemiological studies, case reports, and experimental studies have also suggested that exposure to occupational/environmental levels of mercury may contribute to idiosyncratic autoimmune disease in humans, although data from autoimmune-prone rodents are more convincing (IPCS, 2006).

Drug-induced autoimmunity

Since the observation by Hoffman in 1945 (reviewed by Sarzi-Puttini et al., 2005) that administration of sulfadiazine often coincided with the development of SLE, over 35 drugs have been implicated in the onset of autoimmune responses and autoimmune-like disease (Bigazzi, 1997; Lee & Chase, 1975; Tan, 1974). The disease is different than its classical spontaneous counterpart. For example, in drug-induced autoimmune disease, the disease is usually milder, there is minimal organ involvement, autoantibodies to native DNA are seldom observed in the circulation, and disease remission occurs following cessation of drug treatment. The latter phenomenon may be unique to compounds that induce autoimmunity via the haptenization of native proteins. Nonetheless, the health significance of this adverse effect can be significant.

Immunotoxic drug reactions

It has been suggested that a large percentage of idiosyncratic drug reactions may be immune-mediated (Uetrecht, 2009; Waldhauser & Uetrecht, 1991). If this is the case, then most of these reactions are one of two types of immunopathologies: (a) drugs that produce a state of autoallergy in which the specificity of the immune response is directed to self-tissues or (b) drugs that function as antigens or haptens and stimulate specific immune responses. Much of the early understanding of human drug allergy originated from clinical and laboratory studies of allergy to β-lactam antibiotics, particularly penicillin (Parker et al., 1962), which represented a Type I hypersensitivity reaction. For most pharmaceuticals, immunotoxic drug reactions are rare but, for others this is not so. For example, up to 1% of patients treated with β-lactam antibiotics develop allergy, with anaphylaxis occurring in ∼0.01%.

These early studies were important as they provided suggestive evidence for a multi-genetic basis for the individual variability observed in immuntoxic drug reactions; that now being recognized as genetic variation in drug metabolism and epitope recognition associated, the latter associated with HLA. The former is highlighted by the findings of Perry et al. (1970), who found N-acetyltransferase activity was highly related to the development of drug-induced lupus erythematosus. The latter was exemplified by the importance of HLA in abacavir-induced hyper-sensitivity which is so strongly associated with HLA-B*5701 (Martin et al., 2004), that excluding those individuals with the particular allele prior to abacavir treatment is effective in preventing drug reactions.

A more common occurrence in patients receiving drug therapy, particularly in those receiving proteins, is the possibility that neutralizing anti-drug antibodies (ADA) develop that can significantly alter efficacy and sometimes safety. The presence of ADA was first noted in the 1980s in the treatment of hairy-cell leukemia with interferon-α which resulted in drug resistance (Steis et al., 1988). Over the years ADA has been an area of concern within the pharmaceutical industry and screening tests for predicting ADA have been discussed (Gupta et al., 2011).

Conclusions and current challenges

A brief historical perspective is presented describing the evolution of immunotoxicology from its conception to its current state. Even with our increasing ability to better identify potential immunotoxic agents before they are introduced into humans or the environment, it is unlikely that the discipline will remain without new challenges. For example, two classes of agents that currently offer unique challenges to the immunotoxicologist are biotherapeutic monoclonal antibodies (mAb), particularly those that are immunomodulatory, and various types of nanomaterials. While generally proven to be safe, on occasion biotherapeutics present exaggerated pharmacology that has not been predicted based on an understanding of the intended function or in non-clinical studies (Stebbings et al., 2009). This became evident following the well-publicized adverse event that occurred with an immunomodulatory anti-CD28 super-agonist mAb (TGN-1412) in a clinical trial in the UK (Suntharalingam et al., 2006). Regarding nanomaterials, in vitro and experimental animal studies have provided evidence that, depending on the material (e.g. carbon nanotubes, titanium), they can potentially produce a wide range of immuntoxic effects including sensitization, immunosuppression, and chronic inflammation (Di Gioacchino et al., 2011). While it was initially thought that these materials would have similar modes of action as other immunotoxic chemicals, increasing evidence demonstrates they intercalate into DNA or incorporate directly into cell organelles (Jang et al., 2010), leaving one to question the suitability of some of the current test methods.

Declaration of interest

The author reports no conflicts of interest. The author alone is responsible for the content and writing of the paper.

Acknowledgement

Parts of this work are reproduced from Molecular Immunotoxicology (E. Corsini and H. van Loveren, Eds.) by permission of the publisher (John Wiley & Sons, Publishers: Weinheim, Germany).