Abstract
The number of anti-inflammatory and immunomodulatory drugs being developed in the pharmaceutical industry has increased considerably in the past decade. This increase in research and development has been paralleled by questions from both regulatory agencies and industry on how best to assess decreased host resistance to infections or adverse immunostimulation caused by immunomodulatory agents such as anti-cytokine antibodies (e.g., the tumor necrosis factor-α inhibitors), anti-adhesion molecule antibodies (e.g., anti-α-4 integrin inhibitors) and immunostimulatory molecules (e.g., anti-CD28 antibodies). Although several methods have been developed for nonclinical assessment of immunotoxicity, highly publicized adverse events have brought to light significant gaps in the application of nonclinical immunotoxicity testing in assessing potential risk in humans. Confounding this problem is inconsistent application of immunotoxicology methods for risk assessment within the scientific community, limited understanding of appropriate immunotoxicity testing strategy for immunomodulators and inconsistent testing requests by regulatory agencies. To address these concerns, The Immunotoxicology Technical Committee (ITC) of the International Life Science Institute (ILSI) Health and Environmental Sciences Institute (HESI) organized a workshop on Immunomodulators and Clinical Immunotoxicology in May 2007. The Workshop was convened to identify key gaps in nonclinical and clinical immunotoxicity testing of anti-inflammatory and immunomodulatory agents and to begin to develop consistent approaches for immunotoxicity testing and risk assessment. This paper summarizes the outcome of the HESI ITC Immunomodulators and Clinical Immunotoxicology Workshop. Topics not discussed at the Workshop were outside the scope of this report. Although more work is needed to develop consistent approaches for immunotoxicity assessment of immunomodulators, this Workshop provided the foundation for future discussion.
Introduction
The International Conference on Harmonization (ICH) S8 immunotoxicity testing guidance was approved in 2006 (ICH, Citation2006). This document focused on immunotoxicity testing of drugs not intended to alter immune function. The focus on unintended immunotoxicity was decided upon by the ICH Expert Working Group early in the writing process because of difficulties providing standard testing paradigms for immunomodulatory and anti-inflammatory agents. It was believed that the strategy for these types of compounds should be developed on a case-by-case basis (Weaver et al., Citation2005; ICH, Citation2006).
Strategies used to assess potential immunotoxicity for compounds intended to modulate the immune system vary significantly between pharmaceutical companies and regulatory agencies. Some companies routinely conduct nonclinical immunotoxicity studies beyond standard toxicity studies (STS) (ICH, Citation2000) to better identify potential hazards. On the opposite end of the spectrum, some believe that immunotoxicity testing should be conducted primarily in the clinic given the difficulty in translating nonclinical data to the clinical setting. Most pharmaceutical companies fall in the middle of these two paradigms and find it difficult to determine when nonclinical immunotoxicity testing is needed and which assays should be used. Furthermore, these companies question how nonclinical data can be used to help assess potential risk, make drug development decisions and design clinical studies. This confusion has led to situations in which the expectations for immunotoxicity testing of one regulatory agency or division may differ significantly from those of the sponsor or even other regulatory authorities.
To address these concerns, the Immunotoxicology Technical Committee (ITC) of the International Life Science Institute (ILSI) Health and Environmental Sciences Institute (HESI) convened the Immunomodulators and Clinical Immunotoxicology Workshop in May 2007. The Workshop consisted of approximately 45 participants representing industry (67%), government (29%), and academia (4%). Approximately 22% of the Workshop participants were clinicians. The format of the Workshop was comprised of plenary lectures followed by breakout sessions that addressed a series of predetermined discussion topics and questions. This Workshop was a continuation of the HESI ITC Clinical Immunotoxicology Workshop that convened in February 2005. The ITC Clinical Immunotoxicology Workshop focused specifically on clinical immunotoxicity assessment and aimed to foster a dialogue with regulatory agencies (HESI ITC, Citation2005). Some gaps identified at the Clinical Immunotoxicology Workshop included difficulties in extrapolating immunotoxicity findings between species, unknown relevance of nonclinical testing to humans, lack of validated methods for monitoring immunotoxicity in the clinic and the need for more communication with regulatory authorities. These limitations remained salient as meeting participants gathered for the Immunomodulators and Clinical Immunotoxicology Workshop more than two years later and are still among the challenges faced today.
Plenary lectures
The plenary presentations () were intended to set the stage for the breakout sessions by providing: (1) background on the current strategies and perceived gaps for evaluating immune system status in nonclinical studies and in clinical trials, (2) an overview of immunodeficiencies seen in humans and the available clinical assays to assess immunity and (3) a summary of a HESI ITC survey on current practices in industry. The presentations provided relevant background information and posed several challenging questions that helped foster discussions in the breakout sessions.
Table 1. Immunomodulators and clinical immunotoxicology Workshop plenary presentations.
Current strategies and perceived gaps
Dr. Hastings raised concerns about the adequacy of the ICH S8 immunotoxicity testing guidance for evaluating the immunotoxicity potential of immunomodulatory compounds. For example, would the T-cell-dependent antibody response (TDAR) assay have predicted the liability of tuberculosis attributed to the tumor necrosis-α (TNFα) inhibitors (O’Brien et al., Citation1999; Treacy, Citation2000; Mohan et al., Citation2001)? Would host resistance assays have predicted activation of latent JC virus by natalizumab (anti-α-4 integrin antibody), which led to progressive multi-focal leukoencephalopathy (PML) in humans (Yousry et al., Citation2006; Burleson and Burleson, Citation2008)? Are there models for predicting adverse immunostimulation (Chapman et al., Citation2007)? Dr. Hastings commented that the nonclinical assays for assessing immunosuppression described in the ICH S8 guidance (e.g., TDAR, immunophenotyping, cell-mediated immunity assays, host resistance assays) are designed primarily for hazard identification of immunotoxicity, not for risk assessment. In addition, there are no recommendations in the ICH S8 guidance for bridging nonclinical immunotoxicity studies to humans, which should be the ultimate goal of immunotoxicity assessment of drug candidates. In writing ICH S8, a further complication was the explicit exclusion of biologic therapeutics from the scope of the document. Thus, many of the issues that have arisen with immunomodulatory biologics would not necessarily have been addressed even if the assays recommended in ICH S8 were appropriate and predictive.
Dr. Gourley presented case studies on immune cell activation and immunosuppression to illustrate the gaps in clinical assessment of immunomodulatory drugs. In light of the recent adverse events seen with the anti-CD28 super-agonist, TGN1412 (Suntharalingram et al., Citation2006), the risk of cytokine storm is an important consideration when evaluating new biologic therapies that target the immune system. Dr. Gourley reiterated the concern raised by Dr. Hastings on the ability of nonclinical assays to assess human risk for adverse immunostimulation, stating that the cytokine storm seen in humans with TGN1412 was not seen in cynomolgus monkeys during nonclinical testing (ESG on Phase One Clinical Trials: Final report, Citation2006). Nevertheless, there may now be an expectation that appropriate nonclinical assessments will be made prior to testing in humans in cases where there is a higher perceived risk of immune activation (e.g., the target is a receptor on an immune cell rather than a soluble factor). In addition, a number of other approaches to ensure subject safety and investigate the potential for cytokine release were discussed, drawing from a recent example of a first-in-man study of an anti-cytokine monoclonal antibody. These safeguards included: (1) requests from regulatory authorities for a more conservative dose escalation at the lower end of the dose range; (2) confirmation that the minimal anticipated biological effect level (MABEL) approach had been used to set the starting dose; and, (3) a request from the site institutional review board (IRB) to stagger doses within each cohort. In this study, a measurement of serum cytokines was made predose and at intervals in the first 24 hr in order to identify possible ‘sub-clinical’ cytokine release.
Dr. Gourley discussed issues regarding this approach, namely that the utility of such measurements to predict clinically significant cytokine release at higher doses has not been shown. The paucity of historical data from such assessments in a 24-hr period also led to difficulty in interpreting small variations in cytokine concentrations observed in some subjects. With respect to the potential for immunosuppression with biologics developed against immune targets, Dr. Gourley discussed points, echoed many times in the Workshop, that the predictivity of nonclinical testing in animals to humans remains to be determined. In addition, there is a lack of knowledge and experience concerning the best assays to use in humans in a given setting. The choice of assays used in clinical trials (e.g., functional assays like the TDAR or static endpoints such as measurement of immunoglobulins and immunophenotyping) may be decided by the types of tests already performed in animals. Also discussed was the problem of how results from human studies should be interpreted in terms of clinically significant change (e.g., predictive of increased risk of infection) in an assay such as the TDAR/vaccine challenge study.
Human immunodeficiencies and clinical assays
Dr. Harbeck presented a series of flow diagrams that provided guidance on evaluating the underlying causes of increased susceptibility to infections in humans. The presentation focused on deficiencies in humoral immunity, cellular immunity, complement and phagocyte activity. summarizes several clinical endpoints that may be used to investigate effects on immune function in humans. These assays are consistent with the list proposed by the individual breakout groups. Dr. Harbeck also provided a summary of potential pathogens associated with specific immunodeficiencies, which was taken from Bonilla et al. (Citation2006). Although there are numerous assays available to measure adverse effects on the immune system in humans, the challenge remains to determine which assays should be employed, when testing is needed and how the data would be used in the risk assessment process.
Table 2. Clinical endpoints for assessing immunodeficiencies.
HESI ITC immunomodulator survey
As a starting point for the Workshop, the ITC conducted a survey of member companies; 12 companies (Amgen, AstraZeneca, Bayer AG, Boehringer Ingelheim, Bristol-Myers Squibb, Eli Lilly & Company, F. Hoffmann-LaRoche Ltd., GlaxoSmithKline, Pfizer, Sanofi-Aventis, Schering-Plough, and Wyeth Research) responded. The purpose of the survey was to determine how various companies approach safety assessment of immunomodulatory agents and to identify shared issues, level of consistency and the rationale for varied approaches in different organizations. The survey was to serve as a tool to direct and help focus the Workshop discussions. For the purposes of this overview, “immunotoxicity” assessment refers to the conduct of specialized studies or the collection of immunotoxicity endpoints in addition to those routinely evaluated in STS.
The majority of respondents replied that they currently assess immunotoxicity (unintended effects or exaggerated pharmacologic effects) for drugs intended to be immunomodulatory. One rationale provided by those who do not routinely do so was that there are no thresholds that define the point when pharmacology becomes toxicity. The types of studies done and their timing within a program are generally determined on a case-by-case basis, depending on the molecule, target, information from literature, intended patient population or findings in nonclinical safety studies or clinical trials. Almost all respondents indicated that they have used each of the following criteria: class of drug and known liability, mechanism of action, target population, indication, review of literature and concern from clinicians conducting clinical trials. All respondents indicated that regulatory requirements or requests were frequent drivers to conduct immunotoxicity studies and the types of assays chosen; however, there was no consensus about whether some regulatory authorities (e.g., Europe, Japan, USA) were more or less likely to make specific requests.
The most commonly used methods for the evaluation of immunotoxicity are the TDAR assay and peripheral blood lymphocyte immunophenotyping, particularly when looking for broad unintended effects, but a variety of functional assays are employed. Many of the assays incorporated into nonclinical safety programs are conducted in a non-GLP (Good Laboratory Practice) fashion in a discovery setting to determine efficacy, aid in compound selection and/or to demonstrate competitive advantage. While these studies are typically conducted prior to compound recommendation for development as a clinical candidate, the results may be used to inform the conduct of nonclinical toxicity studies or clinical trials. Most companies have translated specific nonclinical assessments to clinical trials. Examples include immunophenotyping, delayed-type hypersensitivity (DTH) assessment, and TDAR evaluation by means of vaccine studies. In general, immunotoxicity studies are being conducted either exclusively in nonclinical models or in both nonclinical models and clinical studies at a given company. None of the companies indicated a preference for immunological assessments in the clinic in the absence of immunotoxicity assessments in nonclinical models.
Some pharmaceutical companies reported different approaches for immunomodulatory biologics than for immunomodulatory small molecules. Compared to small molecules, some companies were less likely to perform a TDAR assay for biologics and more likely to assess additional endpoints/models such as cytokine release (in vivo or in vitro), histamine release, complement activation, knockout (KO) mice and host resistance assays. Although the reasons for this approach were unclear from the responses, it likely reflects biology-driven decisions and a presumed lack of relevance of the TDAR for the particular biologic mechanism of the molecule. Approximately half of the respondents indicated there had been a change in their approach since the experiences with Tysabri (natalizumab) and TGN1412. For example, nonclinical evaluation of latent viruses is receiving more attention in programs targeting lymphocytes due to the PML cases seen with Tysabri. Also, as a direct result of the TGN1412 experience, increased attention has been focused on cytokine release and cell activation assays. There is now more confusion regarding regulatory agency expectations; clinical trial approaches for certain targets have changed in accordance with new recommendations such as the MABEL approach (CHMP, Citation2007). The MABEL is the anticipated dose level leading to a minimal biological effect level in humans. Safety factors that take into account criteria such as the novelty of the active substance, its potency and mode of action, and the degree of species specificity are used along with the MABEL to calculate the first dose in humans (CHMP, Citation2007).
A considerable number of questions in the survey focused on host resistance assays. There was very little consensus on the triggers for conducting such assays for small molecules and biologics, which may reflect both general skepticism about the usefulness of animal host resistance assays in predicting human risk as well as challenges with assay development and data interpretation. Some respondents felt that similarly useful but more relevant data could be obtained by conducting other tests of immune function. To the question “When will your organization choose to do a host resistance assay?” the following options were listed in the survey: (a) never; (b) only when specifically requested by regulatory/health authority; (c) class of drug and known liability; (d) mechanism of action; e) target population; (f) indication; (g) literature on rodent gene knockout models; (h) concern of clinicians running clinical trials (internal due diligence); and, (i) findings in nonclinical studies. Each potential choice received only 1–5 “yes” responses out of 12 responding companies and the three most commonly cited choices were “only when specifically requested by regulatory/health authority”, “concern of clinicians running clinical trials (internal due diligence)” and “findings in nonclinical studies”. Other reasons listed included understanding significance of findings from other immunotoxicity studies, assessing effects of a drug candidate in comparison with current standard of care therapies, determining competitive safety advantage, confirming pharmacologic activity/mechanism of action and screening candidates for development.
When host resistance assays are conducted, species, assays and challenge organisms are generally selected case-by-case on a scientifically driven, weight-of-evidence (WoE) basis, with preference given to well-established models and those with relevance to the molecule or human disease. Sometimes assay selection is driven by a specific request from a regulatory authority. The results from host resistance assays have occasionally (1–2 responses each) affected a decision to conduct or continue clinical trials, progress development or select clinical monitoring strategies.
In summary, the survey uncovered a wide range of routine approaches, experience with regulatory authorities and internal risk assessment philosophies; however, these observations may simply reflect a wide range of targets and patient populations. The most consistent response was that programs are evaluated on a case-by-case basis and that immunotoxicity assessments are conducted based on cause-for-concern, which is essentially determined by a holistic evaluation (i.e., WoE approach). Importantly, there was no standard paradigm considered appropriate by survey respondents for all immunomodulatory agents.
Breakout sessions
Following the plenary lectures, the Workshop transitioned to the breakout sessions. There were four breakout groups that were asked to address the predetermined discussion topics and questions listed in . Each breakout group focused on their assigned mechanism or target, which included: (a) migration/adhesion and neutrophil trafficking; (b) macrophage and antigen-presenting cell (APC) function; (c) T- and B-cell function; and, (d) cytokines. Several topics of consensus were reached by Workshop participants including causes-for-concern that would warrant additional specialized immunotoxicity assessment and identified gaps in testing strategy, which are discussed next. In general, the causes-for-concern and gaps identified were relevant across all breakout groups, although some differences were identified based on mechanism of action or target. Topics such as the timing of nonclinical and clinical immunotoxicity testing, how regulatory agencies view nonclinical immunotoxicity data for risk assessment in humans and examples of nonclinical safety endpoints translated to clinical trials were not adequately discussed at the Workshop.
Table 3. Breakout session discussion topics and questions.
Causes-for-concern that would warrant additional immunotoxicity studies
Characterization of off-target effects
As described in the ICH S8 guidance document for all new chemical entities, various factors are considered in a WoE review to determine if there is a significant cause-for-concern that warrants the need for additional immunotoxicity studies (ICH, Citation2006). Data from in vivo pharmacology, exploratory toxicity and standard toxicity studies may reveal unexpected changes in hematology, organ weight or histopathology data, either qualitatively or quantitatively (i.e., in severity or duration), possibly suggesting an increased risk of infection. These findings may raise concerns and additional immunotoxicity testing may be needed. However, there was consensus among the breakout groups that immunomodulatory drugs should be considered to bear immunotoxicity risks de facto and therefore may be evaluated on a case-by-case basis for potential off-target or unexpected effects on the immune system in nonclinical and clinical studies. Off-target in this context implies effects on an unintended cell type or function.
For example, a drug targeted for a specific signal transduction molecule could, in theory, inhibit not only macrophage activation (target), but B-cell function (off-target) as well. As another theoretical example, compound X is an antagonist for cytokine receptor alpha (target) on macrophages, but at higher concentrations it also binds to cytokine receptor beta (off-target) on T-cells. Thus, at higher exposure levels, T-cell activity may also be decreased in addition to the intended reduction in macrophage function. The selection of studies to identify such off-target effects should be based on a WoE review of information already available (e.g., literature on knockout mice or molecules directed at the same target). If the test compound has an unprecedented mechanism of action or target, very little may be known about the possible off-target effects and, therefore, more studies may be warranted.
There was significant discussion on what types of data on off-targets effects are needed versus those that are informative but not critical to assess safety and whether evaluation of off-target toxicity should be conducted in the clinical rather than nonclinical setting. The difficulty in deciding if additional immunotoxicity testing should be done is illustrated in the following example. Although the TDAR assay is the model in this example, this scenario may be instructive for other immune function tests. A discussion by one of the breakout groups focused on a hypothetical drug that targets monocyte migration. If a nonclinical TDAR assay was conducted and a dose-dependent decrease in antibody response was observed with an adequate safety margin, the group suggested that more aggressive monitoring of infection rates and exclusion criteria should be utilized in the clinical studies. However, it was pointed out that these types of precautions already would have been planned in clinical studies based on the target and that a TDAR study, in this case, may be of little additional value. Other participants pointed out that in nonclinical immunotoxicity studies, such a TDAR assay is used primarily to assess potential hazard but are not useful for detecting subtle effects.
In determining off-target effects, it was pointed out that one must be cognizant of species differences that could mislead investigators. Specifically, issues when testing cytokines or other soluble factors may be a concern. For example, it was noted that interleukin-(IL)-2 induces autoimmune disease in mice, but rats are resistant to the effects of IL-2. The effects of IL-10 on mice and humans also are different. Cytokine signal transduction pathways and redundancies between pathways may differ between rodents, canine, non-human primates (NHPs), and humans and thus, alterations observed in, animals may not predict changes in humans.
Other causes-for-concern
Additional causes-for-concern were emphasized by the different groups. Non-reversibility of toxicity was noted as being a significant issue that would increase the level of concern. It was emphasized that reversibility from immunosuppression should be demonstrated in appropriate nonclinical studies. There also were concerns raised with regard to patient populations that either may already be immunosuppressed (e.g., HIV+ or elderly) or populations for whom there would be a significant clinical impact if they were immunosuppressed (e.g., those with latent infections such as Mycobacterium tuberculosis or diseases such as Type-2 diabetes that predispose to infection). Further, immunotoxicity does not appear to differ significantly from other toxicities in that the safety margin between the efficacious exposure and the plasma concentration that results in changes in immunotoxicity parameters was felt to be important in determining if the off-target effects would be a significant safety concern in clinical studies. Last, immunostimulatory compounds were recognized as requiring a higher degree of immunotoxicity characterization. For example, agents that induce T-cells to release pro-inflammatory cytokines (e.g., OKT3 monoclonal antibody) or activate innate immunity via toll-like receptors should be evaluated in pharmacologically-relevant species to determine human risk potential.
Nonclinical gaps in immunotoxicity assessment of immunomodulators
Translation to humans
summarizes some of the gaps in nonclinical assessment of immunomodulators identified at the Workshop. Most of the participants attending the Workshop believed that nonclinical immunotoxicity testing is useful for hazard identification. However, the translation of the hazard to actual risk assessment in humans is unclear and was identified as one of the major gaps. Take, for example, the case where an anti-inflammatory drug that targets macrophage function is being developed and the project team wants to learn if there would be a significant effect on innate and adaptive immune responses. To address this, monocytes from rats treated with test compound were evaluated ex vivo and it was determined that phagocytosis and killing of bacteria was slightly decreased. A rat TDAR assay also was conducted and a 50% decrease in the antibody response was observed at the high dose. Based on this information, one can conclude that there is a potential to decrease monocyte/macrophage function and antibody responses. The data could be used to determine a safety factor, assist in clinical study design, and to determine potential immunotoxicity endpoints to include in clinical studies. For this example, however, the primary challenge is how to use hazard identification data to understand potential risk of decreased host resistance in humans. It is unclear how a 50% decrease in the antibody response in a rat translates into an antibody response in humans.
Table 4. Gaps and areas of controversy in nonclinical assessment of immunomodulators.
To address this issue, one of the groups recommended that nonclinical and clinical immunotoxicity data be compared to determine applicability of animal data to humans. Given the increased number of immunomodulatory drugs that have been marketed over the past 10 years (e.g., transplantation drugs, anti-inflammatory agents); it may be possible to compare nonclinical and clinical data at pharmacologically effective human exposure to help bridge this gap in knowledge. Further, the translation of animal data to human risk assessment can be greatly facilitated by understanding target differences between species. Some targets may function similarly between species, while others may not. For example, if the pro-inflammatory effects of a specific cytokine in rats differ from humans, inhibition of this cytokine may produce changes observed in the rat but not in NHPs or humans. Thus, a good understanding of the pharmacology and pathways affected in both the toxicology species and in humans will enhance the translation of nonclinical findings to humans.
Host resistance assays
The topic that elicited the greatest debate among participants was nonclinical host resistance assays. However, due to time limitations, the use and relevance of host resistance assays were not thoroughly explored. In general, Workshop participants agreed that alternatives to host resistance models (in relevant species) are needed to provide more meaningful data to clinical investigators, help translate findings from animals to humans and help better determine the no observed adverse effect level (NOAEL). The concerns expressed by Workshop participants about host resistance assays reflected the lack of consensus in responses to survey questions on the use of these assays in immunotoxicity assessment (see above). The primary concern expressed by most participants, including clinicians, was the translation of the findings to man and the interspecies relevance of organisms used in the assay. Thus, several participants felt that data from host resistance assays have limited value in risk assessment. Others expressed concern that a safety margin could not be determined from host resistance studies due to use of multiple pathogen challenge levels and lack of agreement on the most clinically relevant challenge level.
Additional opinions expressed by participants on the limitations of host resistance assays included the uncertainty of using survival versus tissue burden as the endpoint, the difficulty of obtaining approval by institutional animal care and use committees to conduct studies with a survival endpoint, and the lack of infection models in primates. Some of the participants, however, supported using well-characterized host resistance models. For example, the group that focused on drugs that target macrophages indicated that the mouse Listeria monocytogenes model could be used as a comprehensive test for innate immunity by testing for different functions of this arm of the immune response in one assay. With that said, several participants expressed a need for in vivo models, other than host resistance assays, to assess innate immune function. In vivo assays that measure migration of immune cells and are translatable to humans would be helpful for certain drugs.
A few participants suggested a host resistance assay using a standard battery of organisms would be useful for hazard identification. However, more participants expressed a preference for the use of specific, target- or target population-based organisms and use of these assays only on a scientifically justified, case-by-case basis. Although there was significant disagreement on the broad applicability of host resistance assays, there was general consensus that these assays may be helpful, given the appropriate situation. Some of the participants supported the use of host resistance assays in certain scenarios: (1) for hazard identification and not risk assessment; (2) to identify potential risk of infection with pathogens that may be of particular concern in humans; (3) to investigate whether the alteration of a specific arm of the immune response would translate to decreased host resistance given the redundancy of the immune system; and, (4) to understand mechanisms underlying unexpected infections in nonclinical or clinical studies. To help address the latter, several groups suggested that animal health documentation and diagnosis of infections need to be improved to better understand the significance of unexpected infections occurring in the context of STS and to determine whether or not the changes observed were related to drug treatment.
T-cell-dependent antibody response
The nonclinical use of a TDAR assay to investigate unintended immunosuppression is well accepted; however, the utility of this assay to assess the immunotoxicity potential of immunomodulators remains to be determined. Some participants questioned the usefulness of the TDAR assay or other immune function assays in the development of known immunosuppressive agents. For immunomodulatory agents, some meeting participants believed the TDAR assay or other nonclinical assays should only be used when the assay will influence clinical monitoring (e.g., types of infections, biomarkers), selection of patient populations and/or indication, dose selection and escalation strategies, and stopping criteria of clinical studies.
In addition, TDAR data could guide business decisions by determining what safety liabilities are either unacceptable or perhaps less significant in comparison to competitor compounds. The utility of the TDAR can also depend on the mechanism of action or target of the immunomodulatory agent. For compounds that inhibit macrophage adhesion, the TDAR may be useful to determine potential off-target effects. For drug candidates that inhibit T- or B-cell function, the effects on a TDAR at efficacious doses may be investigated in discovery studies. In these instances, the TDAR could be tested in toxicology studies at higher doses to investigate the potential impact of exaggerated pharmacology. When effects on the TDAR assay are anticipated, assessing reversibility is important.
The participants emphasized that toxicologists need to work early and closely with clinicians to develop the appropriate type of tests and testing strategies. This type of collaboration is critical so that information gained from the TDAR assay and other nonclinical studies can be used in a WoE review of all the toxicity data to determine the appropriate clinical testing strategy. Although both the TDAR assay and host resistance models have limitations, the majority of participants favored a TDAR assay for routine investigations, consistent with survey results. One of the breakout groups felt the TDAR and immunophenotyping used as standard assays is a reasonable approach to search for unexpected liabilities or to gauge the acceptable degree of immunomodulation relative to desired pharmacology, margin of safety and the intended disease indication. In general, however, participants did not favor the conduct of nonclinical studies that would not be informative in making clinical decisions.
Clinical gaps in methods and knowledge
Clinicians must be prepared to address real or perceived concerns about immunotoxicity. Immunotoxicity studies, if conducted early in development, provide useful data to address these concerns and to implement safety monitoring and vigilance measures. The information from immunotoxicity studies also may provide a rationale and criteria for patient inclusion/exclusion in clinical trials. For this to occur, an understanding of the nature of the dose-response relationship observed in immunotoxicity studies is especially important. In some cases, nonclinical immunotoxicity testing can identify biomarkers for clinical monitoring of immunopharmacology or immunotoxicity. Important nonclinical findings related to adverse immune effects also may affect drug labeling.
summarizes the gaps in assessing immunotoxicity potential of immunomodulators in humans identified by Workshop participants. There was consensus that we need more and better clinical assays to monitor different components of the immune response. A key gap identified by the breakout groups was that a commonly-accepted approach for the human TDAR assay or vaccine study has not been developed. Recall responses to antigens, such as tetanus and primary immunization with bacteriophage or KLH, were discussed as potential options. Some participants pointed out the hypothetical risk of cross-sensitization to KLH with shellfish antigens using this approach. The triggers for conducting a human TDAR assay and the usefulness of this approach in risk assessment was not adequately addressed. Several groups expressed the need for a better in vivo assay to evaluate cell-mediated immune responses in humans. Improving delayed-type hypersensitivity (DTH) responses through reduction of inter-subject variability is one possible approach to achieving this. Other participants suggested the assessment of contact hypersensitivity as a potential approach to evaluate cell-mediated immune responses. There was consensus that clinical trials should conduct more focused monitoring for specific infections in patients based on known mechanisms of action (e.g., herpes or cytomegalovirus (CMV) outbreaks) and not just general infection rates. To accomplish this, better approaches to describe the infection types in clinical events are needed. Another important point was made that for many of these assays, especially those that require live cells, more cross-lab validation and improvements in sample collection, cell isolation, freezing methods and sample stability are needed.
Table 5. Research gaps in clinical assessment of.
During discussions on the use of cytokines (e.g., interleukins) as clinical markers of immunomodulation, participants acknowledged that although serum cytokine levels may have diagnostic value for certain autoimmune or inflammatory conditions such as cytokine storm, there are currently too many assay limitations, e.g., high individual variability and lack of clinical validation, for cytokines to be presently adopted as a routine biomarker of immune status.
Specialized nonclinical immunotoxicity tests
Each breakout group, based on their assigned mechanism of action or target, provided a summary of specialized nonclinical immunotoxicity endpoints to assess potential immunotoxicity. As anticipated, there were consistencies across the breakout groups in the endpoints used to assess immunomodulation. Although questions remain on the utility of the TDAR assay to detect the immunotoxicity potential of known immunomodulators and anti-inflammatory agents (see above), all breakout groups recommended using this assay to assess either on- or off-target effects when appropriate. The groups also recommended immunopheno-typing of lymphocyte subpopulations by flow cytometry and/or immunohistochemistry, including assessing activation markers on a case-by-case basis.
Assays recommended to test compounds affecting migration/adhesion and macrophage/APC function included chemotaxis assays, phagocytic and cytolytic assays and tests for respiratory oxidative burst. Host resistance assays and cytokine evaluation also were listed as possible endpoints. The breakout group discussed clinical tests that could be used to assess immunomodulation. The endpoints proposed were consistent with the clinical tests described during the plenary lecture ().
The Workshop participants briefly discussed the extent of assay validation needed for immunotoxicity tests. Although it was generally accepted that an assay used to make a regulatory decision should be fully validated, it was recognized that few immunotoxicity assays rise to this level of scrutiny. If the assay is used along with other endpoints (i.e., in a WoE approach), which is more likely the case, then extensive assay validation may not be needed. Importantly, it was noted that if an assay is used enough and shown to be useful, it may be considered informally validated. It was acknowledged that there are significant differences in opinion on the definition of assay validation (e.g., predictivity, performance). In addition, more work is needed to develop appropriate reagents for rats, dogs and NHPs (e.g., for cytokine analysis and immunopheno-typing). To date, most of the focus in the area of immunology and, hence, the development of reagents and tools, has been on the mouse and human.
Summary
When assessing the immunotoxicity potential of immunomodulators and anti-inflammatory agents, one size clearly does not fit all. There are significant differences among companies in approaches used to assess the immunotoxicity potential of these compounds. Most companies support a WoE approach based on factors such as STS findings, pharmacology, patient population and indication. There was consensus at the Workshop that assessment of immunomodulatory agents should be conducted on a case-by-case basis to help evaluate potential off-target effects and exaggerated pharmacology. The TDAR and immunophenotyping assays are commonly utilized endpoints for nonclinical studies, whereas there was significant disagreement over the usefulness of host resistance assays.
Questions remain regarding the value of nonclinical immunotoxicity studies for immunomodulatory compounds. Given issues such as translation of nonclinical findings to humans, some believe that potential adverse effects should be evaluated primarily in clinical studies. Although a focus on clinical evaluation of immunotoxicity was favored by several Workshop participants, it was acknowledged that these evaluations are limited by the lack of clinical assays to assess immune function. Additional meetings with key stakeholders such as clinicians, regulatory agencies and industry and academic toxicologists are needed to address the gaps identified at this meeting and to develop consistent approaches for nonclinical and clinical immunotoxicity assessment of immunomodulators and anti-inflammatory drugs.
Acknowledgments
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Related Research Data
References
- Bonilla, F. A. and Geha, R. S. 2006. Update on primary immunodeficiency diseases. J. Allergy Clin. Immunol. 117:S435–441.
- Burleson, G. R. and Burleson, F. G. 2008. Testing human biologicals in animal host resistance models. J. Immunotoxicol. 5:23–31.
- Chapman, K., Pullen, N., Graham, M., andRagan, I. 2007. Nonclinical safety testing of monoclonal antibodies: the significance of species relevance. Nature Rev./Drug Discov. 6:120–126.
- CHMP (Committee for Medicinal Products for Human Use). 2007. Guideline on strategies to identify and mitigate risks for first-in-man clinical trials with investigational medicinal products. EMEA/CHMP/SWP/28367.
- ESG (Expert Scientific Group) on Phase One Clinical Trials: Final report (Norwich, US: Stationery Office, 2006. http://www.dh.gov. uk/en/Publicationsandstatistics/Publications/ PublicationsPolicyAnd Guidance/DH_063117.
- HESI Immunotoxicology Technical Committee, Clinical Immuno toxicology Workshop, Meeting Minutes, February, 2005. http://www.hesiglobal.org/NR/rdonlyres/18E51EC7-BFEC-4799-AA30-87FBE57D8D23/0/AttachmentG.doc.
- ICH (International Conference on Harmonization). 2000. Guidance for Industry. M3 Nonclinical Safety Studies for the Conduct of Human Clinical Trials for Pharmaceuticals.
- ICH (International Conference on Harmonization). 2006. Guidance for Industry. S8 Immunotoxicity Studies for Human Pharmaceuticals.
- Mohan, V. P., Scanga, C. A., Keming, Y., Scott, H. M., Tanaka, K. E., Tsang, E., Tsai, M. C., Flynn, J. L., and Chan, J. 2001. Effects of tumor necrosis factor alpha on host immune response in chronic persistent tuberculosis: Possible role of limiting pathology. Infect. Immun. 69:1847–1855.
- O’Brien, D. P., Briles, D. E., Szalai, A. J., Tu, A. H., Sanz, I., and Nahm, M. H. 1999. Tumor necrosis factor-α receptor I is important for survival from Streptococcus pneumoniae infections. Infect. Immun. 67:595–601.
- Suntharalingam, G., Perry, M. R., Ward, S., Brett, S. J., Castello-Cortes, A., Brunner, M. D., and Panoskaltsis, N. 2006. Cytokine storm in a Phase 1 trial of anti-CD28 monoclonal antibody TGN1412. New Engl. J. Med. 355:1018–1028.
- Treacy, G. 2000. Using an analogous monoclonal antibody to evaluate the reproductive and chronic toxicity potential for a humanized anti-TNFα monoclonal antibody. Hum. Exp. Toxicol. 19: 226–228.
- Weaver, J. L., Tsutsui, N., Hisada, S., Vidal, J.-M., Spanhaak, S., Sawada, J., Hastings, K. L., van der Laan J.W., van Loveren H., Kawabata, T.T., Sims, J., Durham, S. K., Fueki, O., Matula, T. I., Kusunoki, H., Ulrich, P., and Nakamura, K. 2005. Development of the ICH guidelines for immunotoxicology evaluation of pharmaceuticals using a survey of industry practices. J. Immunotoxicol. 2:171–180.
- Yousry, T. A., Major, E. O., Ryschkewitsch, C., Fahle, G., Fischer, S., Hou, J., Curfman, B., Miszkiel, K., Mueller-Lenke, N., Sanchez, E., Barkhof, F., Radue, E. W., Jäger, H. R., and Clifford, D. B. 2006. Evaluation of patients treated with natalizumab for progressive multifocal leukoencephalopathy. New Engl. J. Med. 354:924–933.