Toxicopathology of the Developing Immune System

Abstract

Developmental immunotoxicity (DIT) has gained attention with the recognition that environmental chemicals can potentially affect the developing immune system and the incidence of childhood allergic diseases. Preclinical safety assessment of pharmaceuticals for men and women of childbearing potential as well as for pediatric and juvenile indications may require DIT assessments. Draft documents from environmental and chemical regulatory agencies propose strategies that use the rat as a test species and incorporate histopathology and functional testing as endpoints. While there are no guidelines for DIT assessment of pharmaceuticals, current discussions suggest that combining immunotoxicity and developmental and reproductive toxicology studies may serve this purpose. Knowledge of the principles and applications of DIT will facilitate participation in strategy development and effective conduct of relevant studies.

Introduction

There is increasing international awareness of the potential effects of xenobiotics on the developing human immune system. This awareness has in part been triggered by the increasing incidence of childhood allergic disease including asthma (Dietert and Zelikoff 2008) and the documented effects of environmental chemicals on the developing endocrine system (Colborn, vom Saal, and Soto 1993), nervous system (Bellinger 2004), and immune system (Luebke et al. 2006). Efforts are under way worldwide to develop guidelines to define risk and limit exposure of children to environmental chemicals. The World Health Organization–sponsored International Program on Chemical Safety (IPCS) Harmonization Project (IPCS Harmonization Project 2009) is developing guidance on immunotoxicity risk assessment that will be released for public and peer review in 2010. The current Office for Economic Co-operation and Development (OECD) Draft Proposal for an extended one-generation reproductive toxicity study (OECD 2009) proposes including one cohort of rats to assess the potential impact of chemical exposure on the developing immune system. Endpoints include T-dependent antigen response, organ weights, immunophenotyping, and histology. Data for multigeneration studies have been requested.

Likewise, there is increasing awareness and interest in the pharmaceutical arena to define and manage risk for immunotoxicity of pharmaceuticals with indications for use by men and women of childbearing potential, pediatric populations, and juvenile populations. Currently, there are no guidelines that specifically address the need for developmental immunotoxicology (DIT) evaluations. Guidelines for developmental and reproductive toxicology (DART) testing (ICH 2005) and immunotoxicity testing in adult animals (ICH 2006) have been established. Proposals for combined DART and DIT assessment are currently under discussion. Issues associated with the inclusion of immunotoxicology assessments in reproductive studies were the topic of a session at the 2005 annual meeting of the Society of Toxicology (Ladics et al. 2005). A proposed testing framework for DIT was published with authors from the U.S. Food and Drug Administration (FDA), U.S. Environmental Protection Agency, National Institute for Occupational Safety and Health, International Life Sciences Institute Health Science and Environmental Sciences Institute (ILSI/HESI), a major pharmaceutical company, and a major chemical company (Holsapple 2005). DIT assessment remains an area of ongoing active discussion. An ILSI/HESI-sponsored workshop on DIT testing of pharmaceuticals was held in May 2010.

Mice have been used extensively as experimental models for study of the immune system but are less widely used than rats in the regulatory toxicology environment. Although significant similarities of immune system function are shared among mammals, in-depth study of the mouse and human allows unique understanding of species-specific differences. While these differences cannot be directly extrapolated to the lesser studied immune systems in the rat, dog, rabbit, nonhuman primate, and pig, these differences highlight the importance of understanding the limitations of the specific model being used and the need for further immunological research to appropriately apply these models to risk assessment in humans.

While mice are more extensively studied for research in immunology, rats are the preferred model for preclinical safety assessment and are being proposed as a potential model for combined DART and DIT studies. Histopathology of the developing rat immune system is poorly described in the literature. Understanding of the development of this system in the normal rat is required for the proposed use of histologic endpoints in future DIT studies.

Immunomodulatory compounds in and of themselves raise specific concerns for DIT assessment based on their mechanism of action. Monoclonal antibodies are a rapidly expanding segment of the modern pharmacopoeia. In combination, immunomodulatory monoclonal antibodies pose unique challenges for preclinical DIT assessment. Successful strategies for assessment of immunomodulatory monoclonal antibodies highlight some of the potential challenges and solutions in the field of DIT.

Current preclinical DIT assessment typically involves the use of either rat or nonhuman primate test systems. An understanding of the overall approach as well as specific aspects of the proposed tests can yield insight into the development of an appropriate integrated strategy. Understanding of needs, strengths, weaknesses, and knowledge gaps for assessment of the developing immune system prior to initiation of a preclinical testing program will assist in appropriate use of animals and increase the likelihood of a useful and relevant outcome.

Ontogeny of the Immune Response in Humans and Mice: Implications for Developmental Immunotoxicology

David B. Lewis, MD, Stanford University, School of Medicine, Stanford, California, USA

Based on transgenic technology and the availability of a large number of specific antibody reagents, mice have provided invaluable insights into the general features of innate and adaptive immunity. However, there are important differences in the immune systems of rodents and humans that need to be considered when applying rodent models to more specialized questions, such as DIT. Differences in the immune systems of mice and humans that may have important biological consequences for the fetal and early postnatal development of the immune system are reviewed. The potential value of high-throughput analysis of the immune system including deep sequencing of the transcriptome and phosphoflow cytometric analysis for DIT investigation is highlighted.

Innate Immunity

Starting with the 1997 report of mammalian Toll-like receptors (TLRs; Medzhitov, Preston-Hurlburt, and Janeway 1997), there has been continuing intense interest in defining innate immune mechanisms. These analyses have revealed some striking differences between mice and human TLR expression. For example, mice but not humans express a functional TLR11 gene, which may be involved in the protection of mice from uropathogenic Escherichia coli (Zhang et al. 2004) and the unicellular protozoan, Toxoplasma gondii (Lauw, Caffrey, and Golenbock 2005). Rodents and humans also have major differences in the cell types that express particular TLRs, most notably TLR9. While TLR9 is expressed by both humans and mice in plasmacytoid dendritic cells and B lymphocytes (Kadowaki et al. 2001), it is expressed only in mice by conventional (myeloid) dendritic cells (cDCs), mononuclear phagocytes, and activated T lymphocytes. These species differences in TLR9 expression may have important biological implications in immunotoxicology because they likely account for why TLR9 stimulatory ligands (e.g., unmethylated CpG-containing oligonucleotides) elicit tumor necrosis factor–α–dependent toxicity in mice but not in humans (Campbell et al. 2009). The expression of TLR9 by cDCs in mice but not humans also suggests that there should be caution in assuming that the ability of TLR9 ligands to serve as adjuvants for vaccine responses in neonatal mice (Brazolot-Millan 1998) applies to the vaccination of human newborns because the activation of cDCs is typically critical for robust induction of adaptive immunity.

Adaptive Immunity

A comparison of the immune phenotype of mouse gene knockouts with naturally occurring human defects of the homologue has demonstrated important differences in the role of particular gene products for T- and B-lymphocyte development. For example, human genetic deficiency of proteins involved in IL-7 receptor signaling, including the IL-7 receptor alpha chain (CD127), the common gamma chain cytokine receptor (CD132), or the JAK3 tyrosine kinase that is associated with the cytoplasmic domain of CD132, result in a form of severe combined immunodeficiency (SCID) in which T-lymphocyte development is markedly impaired but B lymphocytes are present in normal numbers (T-B+ SCID; Buckley 2004). In contrast, mouse knockouts of these gene homologues have markedly impaired B-lymphocyte development. Thus, IL-7 plays a nonredundant role in B-lymphocyte development in mice but not humans (Mestas and Hughes 2004). There are also species differences in the impact of impaired signal transduction on intrathymic T-lymphocyte development. For example, ZAP-70 is a cytoplasmic tyrosine kinase that is important for signaling via the T-cell receptor (TCR) for antigen during intrathymic positive selection of T lymphocytes. However, mouse ZAP-70 deficiency impairs both CD4 and CD8 T-lymphocyte intrathymic development, whereas human ZAP-70 deficiency permits CD4 T-lymphocyte development (Mestas and Hughes 2004). For some proteins, such as the cytotoxin granulysin used by cytotoxic T lymphocytes and NK cells in humans, the mouse lacks the gene entirely (Krensky and Clayberger 2009). Finally, there can be species differences in the impact of genetic deficiencies that are sex specific. For example, in genetic deficiency of signal transducer and activator of transcription 5b (STAT5b), which impairs the generation of regulatory T lymphocytes in mice and humans (Cohen et al. 2006), there is impaired growth of male but not female mice. In contrast, both females and males are affected in STAT5b-mediated Laron dwarfism in humans (Rosenfeld 2004).

Systems Biology Approaches for Developmental Immunology

When considering using mice for fetal DIT, it is also important to consider that this model may not be informative for the impact on T-lymphocyte development because almost all CD4 and CD8 T-lymphocyte development in mice occurs postnatally. In contrast, human peripheral T lymphocytes first appear in the fetus at approximately 85 days' gestation and by birth are found at higher concentrations in the blood than postnatally (Lewis and Wilson 2010). Interestingly, humans have evolved active mechanisms to maintain tolerance of the fetus toward maternal antigens encoded on hematopoietic cells by the induction of fetal regulatory T lymphocytes (Mold et al. 2008), whereas prenatal regulatory T lymphocytes are not a feature of mice. Despite all of the potential caveats implied by the species differences described above, there is a great value to performing parallel human and mouse developmental immunology studies. To avoid missing important markers and biological processes, one can use systems biology approaches, for example transcriptome analysis, to find new markers for immune processes and then perform species comparison studies to determine the extent that the gene or protein signature are shared. For example, gene expression profiling by microarrays in human T-lineage cells was used to identify protein tyrosine kinase 7 (PTK7), a member of the receptor tyrosine kinase superfamily without a known ligand or demonstrable catalytic activity, as a novel marker for human CD4 T lymphocytes that have recently been produced by the thymus (Haines et al. 2009). Here again, major species differences were found in gene expression: although PTK7 is highly expressed by the mouse and human thymus, PTK7 expression by peripheral mouse T lymphocytes is completely absent, precluding its use as a mouse recent thymic emigrant (RTE) marker. Nevertheless, mouse RTEs marked by transgenic recombinant activating gene-driven expression of green fluorescent protein (Boursalian et al. 2004) can now be compared with bona fide human RTEs using system biology approaches.

These human versus mouse RTE studies, which are in progress, include a quantitative comparison of mRNA gene expression profiles using deep sequencing. This deep-sequencing approach has important advantages over hybridization-based arrays, in that quantitative measurement of 10 to 15 million transcripts can be obtained in 2 days using relatively small amounts of starting total RNA and without quantitative distortion over 10,000-fold range of levels of mRNA expression (Wang, Gerstein, and Snyder 2009). Using deep sequencing, additional markers for RTEs have been identified in humans, and at least one of these may be useful in mouse studies. Use of these additional markers will allow rapid assessment of mouse versus human species differences in RTE gene expression and, by gene ontology analysis, likely differences in biological pathways, which can be confirmed by more targeted experiments. A comparison of the T-lymphocyte transcriptomes of neonatal mice, of human neonates, and of human adult RTEs will be of particular interest, as this may help identify features that are unique to human neonatal T lymphocytes that may be involved in maternal-fetal tolerance, as opposed to being part of general RTE signature at all ages.

Another systems biology approach that is being pursued is phosphoflow cytometric analysis of signal transduction events following engagement of cytokine receptors or the TCR/CD3 complex. This analysis combines surface staining with fluorochrome-conjugated antibodies for markers of interest, such as PTK7, with internal staining for phospho-specific forms of proteins, such as STAT proteins that have undergone tyrosine phosphorylation. A potential advantage of this analysis is that relatively large numbers of samples, such as might be pursued in developmental immunotoxicology studies, can be analyzed. For example, cells that have been stained from six individual animals can be “barcoded” (i.e., labeled with different fluorescent dyes, such as Pacific orange or CFSE) and then pooled together for flow cytometric analysis to improve throughput (Krutzik and Nolan 2006). The growing availability of phospho-specific antibodies should allow an interrogation of most signaling pathways of interest in T lymphocytes and other cell types for a particular set of antibody stains.

Histology of the Developing Rat Immune System

George A. Parker, DVM, PhD, WIL Research Labs, Ashland, Ohio, USA

Proposals for incorporation of DIT studies into existing DART protocols include provisions for histopathologic examination as a major endpoint. Histopathologic evaluations performed during the conduct of routine toxicology studies have resulted in a substantial background database on the histologic features of adult and aged rats, but less is known about the histologic features of perinatal rats. The study reported in this presentation was designed to fill information gaps regarding the histologic features of immune system organs in late gestational, neonatal, and adolescent rats (data and figures are not shown in this publication but were presented at symposium).

Light microscopic examination of routine hematoxylin and eosin–stained histologic sections was performed on thymus, spleen, gut-associated lymphoid tissue (GALT) of small intestine, bone marrow, mandibular lymph node, mesenteric lymph node, and bronchus-associated lymphoid tissue (BALT) of lung from rats at gestational day (GD) 15, GD 20, postnatal day (PDN) 0, PND 1, PND 2, PND 3, PND 10, PND 21, and PND 42. Immunohistochemical (IHC) staining for CD3, CD45RA, and Ki67 was performed on selected specimens to define T-lymphocyte, B-lymphocyte, and proliferative cell populations, respectively.

At GD 15, only the thymus had histologic features that were somewhat similar to those of the adult organ. Other lymphoid organs had only vague features of the adult organs, to the degree that many would be difficult to identify without the context of adjacent identifiable tissues. A substantial degree of differentiation occurred by GD 20, but most lymphoid organs remained markedly immature in histologic appearance. At PND 0, only the thymus and bone marrow had histologic features that were generally similar to the adult organs, and in each case, the juvenile tissue was less cellular than the adult tissue. Lymph nodes, spleen, and GALT appeared markedly immature at PND 0, and BALT was nonexistent, thus reflecting the potential influence of environmental factors on the “normal” morphologic features of the lymphoid organs of adults. During the period PND 1 to PND 3, there was histologic evidence of responses to environmental stimulation of lymphoid organs, particularly mandibular lymph nodes, mesenteric lymph nodes, and GALT, which are known to develop de novo in response to the influx of antigenic stimuli from the gastrointestinal tract. During the period between PND 10 and PND 42, there was a gradual expansion and increase in cellularity of lymphoid organs to a point where the organs had the expected adult morphology at PND 42. GALT, diffuse intestinal lymphocytic populations, and mesenteric and mandibular lymph nodes reached normal adult morphology before the formation of BALT and splenic B-lymphocyte follicles, which were the last structures to appear in the development process. BALT is known to develop in response to antigen influx into the respiratory tract; thus, its development in the present study may have been retarded by the hygienic environment of a typical modern rodent facility. IHC staining for B lymphocytes was particularly helpful in evaluation of spleen sections, and IHC staining for Ki67 was helpful in defining focal cellular expansions such as follicle development in the spleen; therefore, optimal fixation for IHC staining is indicated for DIT studies. Thin (2- to 3-µm) rather than standard 5-µm sections allowed better visualization of fine structural details, particularly in gestational and neonatal rats.

Developmental Immunotoxicology Case Studies: Efalizumab and Rituximab

Joseph Beyer, DVM, PhD, Genentech, South San Francisco, California, USA

DIT programs designed to assess potential exposure effects of efalizumab and rituximab (therapeutic monoclonal antibodies) on the developing embryo and fetus were conducted. Efalizumab is a monoclonal antibody that binds CD11a and blocks LFA-1-ICAM interactions. Efalizumab binds to human and chimpanzee CD11a. Because of its narrow species specificity, a mouse surrogate antibody, muM17, was used for DIT studies in mice. Rituximab binds human and macaque CD20, which results in depletion of CD20+ B lymphocytes by a number of mechanisms, including antibody-dependent cell-mediated cytotoxicity, complement-mediated cytotoxicity, apoptosis, and possibly others. All general safety and developmental immunotoxicology studies for rituximab were performed in cynomolgus monkeys.

Efalizumab

The DIT program for efalizumab was composed of immune system evaluations integrated into a standard prenatal and postnatal development safety study. The study was performed in CD-1 mice using the mouse anti-CD11a surrogate antibody, muM17. A prior immunotoxicology study in adult female mice demonstrated four weekly subcutaneous doses of 3, 10, or 30 mg/kg, which resulted in markedly decreased IgM and IgG responses to injected sheep red blood cells (SRBCs), a T-lymphocyte–dependent antigen, and reduced NK cell responses. The immunological endpoints in the F₁ mice were initially evaluated at 11 weeks of age corresponding to 8 weeks after the last potential exposure through breast milk. Antibody responses to SRBCs were decreased at 11 weeks of age compared with age-matched controls, despite the absence of measureable plasma muM17. Immune parameters were evaluated again in a second cohort of F₁ mice at 25 weeks of age corresponding to 22 weeks postexposure. Anti-SRBC IgM and IgG responses continued to be mildly decreased relative to age-matched controls. In addition, CD11a expression on peripheral blood mononuclear cells was decreased statistically relative to age-matched controls despite the absence of drug. In summary, the administration of the mouse surrogate anti-CD11a antibody, muM17, to CD-1 mice had no effects on pregnancy or survival but resulted in prolonged pharmacological effects on F₁ immune systems, even in the absence of drug.

Rituximab

The DIT program for rituximab similarly integrated immune system assessments into standard embryo-fetal, prenatal, and postnatal developmental safety studies conducted in cynomolgus macaques. In the embryo-fetal development study using doses of 20, 50, and 100 mg/kg given weekly, a dose-dependent depletion of splenic B lymphocytes was demonstrated by IHC in fetuses collected at Day 100 by cesarean section. Fetal serum drug levels collected at Day 100 cesarean sections were 35%, 74%, and 74% of maternal levels. The prenatal and postnatal development study incorporated immunology endpoints and immune organ evaluations to assess the effect of rituximab on the developing immune system of cynomolgus macaques. The F₁ animals had dose-related B-lymphocyte depletion at PND 28 and 90; however, B-lymphocyte counts normalized by PND 180. Functional immune system evaluations indicated no pronounced or lasting effects due to the B-lymphocyte reductions. In summary, exposure to rituximab in utero or during lactation resulted in no effects on embryo-fetal survival or prenatal or postnatal development and was associated with no prolonged effects on immune function in F₁ animals.

Integrated Approaches to Developmental Immunotoxicology Assessments

Mark Collinge, PhD, Pfizer Global Research and Development, Groton, Connecticut, USA

DIT testing of pharmaceuticals in preclinical species is becoming increasingly important to support clinical studies for pediatric and juvenile indications. There is evidence that the developing immune system exhibits greater susceptibility to immunotoxic insult than the immune system of adults. This may be manifested in the immature immune system being more sensitive (effects seen at lower doses) than the adult immune system, or the immunotoxic effects may be more persistent. However, to date, there are no examples of compounds that have been demonstrated to be immunotoxic to the developing immune system but not to the adult at any dose.

There are no regulatory guidelines specifically focused in the area of DIT and no required routine assessments of DIT with pharmaceuticals. However, there are a number of regulatory guidelines from which some guidance can be derived (EMEA 2000, 2008; FDA 2002, 2006; ICH 2006). Similar to immunotoxicity testing in general, decisions about DIT testing are based on weight-of-evidence review, which includes (1) data from standard toxicity studies in adult animals, (2) the intended patient population, (3) clinical findings in adults, (4) expression of the drug target in the immune system, (5) structural similarities to known immunosuppressive compounds, and (6) disposition of the drug. In most cases, the primary driver for conducting DIT testing is whether the compound has been shown to be immunotoxic in adults. Current recommendations indicate that if a drug is expected to be used in pregnant women and has been shown to induce immunosuppression in adults, incorporation of immunotoxicology into a reproductive toxicology study should be considered (FDA 2002). The expectation is that these studies would aid in understanding the potential for permanent/irreversible changes and increased sensitivity in young populations.

Recommendations from past workshops focusing on DIT (Holsapple et al. 2005; Luster, Dean, and Germolec 2003) indicate the rat is the preferred model for DIT testing. This is due largely to the fact that the rat is the recommended species for DART studies, and consequently there already exists an extensive knowledge base for this species. It is also recommended that DIT studies be added to the required developmental/reproductive toxicity studies wherever possible. Animals should be exposed throughout the treatment protocol, including gestation, lactation, and postnatal development, such that all critical windows of vulnerability (reviewed in Burns-Naas et al. 2008; Dietert et al. 2000; Holsapple et al. 2004; Landreth 2002; Luster, Dean, and Germolec 2003; West 2002) can be assessed at once. If an immunotoxic effect is observed, then each specific window could be further investigated as appropriate. These recommendations were made based on consideration of both environmental chemicals and low-molecular-weight pharmaceuticals. However, in some cases, exposure paradigms may be different for pharmaceuticals compared to environmental chemicals, and testing may be restricted to the period of development with similar immunologic development to the human population (e.g., juvenile). Proposed DIT protocols in rats (reviewed in Burns-Naas et al. 2008), in which exposure occurs throughout all stages of development, indicate that immunopathology can be performed on PND 4 and/or PND 21 (weaning) but that functional assessments are generally performed in young adults at about PND 44 to 49. This is due largely to the difficulty in performing functional tests in rodents during the juvenile stage because development of the immune system in rodents is delayed compared with humans. While humans and nonhuman primates (NHPs) are born with a functional immune system, much of the rodent immune system development occurs postpartum (Holsapple, West, and Landreth 2003; Burns-Naas et al. 2008). This postpartum development of the immune system in rats is of particular concern for immune assessments in neonatal and juvenile animals (investigating transient effects during early windows of development) to support pediatric investigational plans for pediatric and juvenile indications.

While rats are the preferred species for DIT testing of pharmaceuticals in most cases, there are a number of reasons that NHPs may be preferred when investigating biologics. These include (1) cross-reactivity of biologics; (2) similarity to humans, including a functional immune system in NHP at birth; (3) expected similar transplacental transfer of antibody therapeutics; (4) immunogenicity in NHPs is closer to humans than rats; and (5) adequate blood volume can be obtained from NHPs for sequential assessments of immune endpoints, whereas blood volume may be limiting in rats. One difficulty in using NHPs is that compared to rodent immune system development, relatively little is known about the developing immune system in NHPs. Determining the corresponding immunologic age of an NHP to that of a human can be difficult. Parameters that can be used include the relative ratios of CD4:CD8 cells (Hendrickx, Peterson, and Makori 2005; Weinbauer et al. 2008) and also serum immunoglobulin levels (Buse 2005; Hendrickx, Peterson, and Makori 2005; Isaacs et al. 1983) because both change after birth over time.

Whether rats or NHPs are chosen, the selection of endpoints used should be based on the mechanism by which immunotoxicity is expected and the changes observed in adult studies. A combination of both structural (histopathology, immune organ weights, lymphocyte subset analysis) and functional (antibody T-lymphocyte–dependent antigen response, cytotoxic T-lymphocyte activity, NK cell activity, delayed type hypersensitivity, etc.) may be the best approach and may be more predictive of immunotoxicity in humans.

There are a number of key gaps that need to be addressed as the field of DIT progresses. Knowledge of immune system development in preclinical species is incomplete, and further work is needed in this area. The normal ranges of some immune parameters (e.g., cytokines, lymphocyte subsets, immunoglobulins) that may change over time during immune system development have not been adequately defined, and this may limit their value as DIT endpoints. There is incomplete knowledge of placental transfer of pharmaceuticals, an important consideration for in utero exposure to biologic therapeutics. Most methods for immunotoxicology assessments have been validated in adult animals but not in animals with immature immune systems, so such methods need further evaluation. These gaps in the state of the science affect our ability to select the most relevant species for DIT assessments, develop appropriate exposure paradigms, and conduct appropriate assays to evaluate immune function. In addition, new assays are needed that focus on immune enhancement and hypersensitivity, since immunosuppression has been the focus of most DIT studies thus far.

Summary

Developmental immunotoxicology is a rapidly developing field, especially with regard to formulation of regulatory guidances for pharmaceuticals. Guidelines for chemicals and environmental toxicants are more advanced, and the imminent publication of proposed guidelines for public and peer review will speed this advancement. Knowledge of the principles of DIT and DART methodologies can result in efficient and effective assessment of the potential developmental immunotoxicity of compounds. Although current approaches continue to be successfully applied, key gaps in knowledge of species-specific differences of the developing immune system, exposure paradigms, and validated assays affect our ability to apply a fully integrated strategy that can appropriately address the full range of potential effects on the developing immune system.