Abstract
With the advances in cell culture methodologies and molecular biology that have occurred over the past several decades, biologics have become as common as small molecules within the portfolios of the pharmaceutical industry. Toxicologic pathologists should be aware of some of the fundamental differences between small molecules and biologics. Effects are not always observed in studies following administration of biologics. When findings are observed, the toxicologic pathologist should initially determine whether the effect(s) are mediated (directly or indirectly) via the intended pharmacology, exaggerated pharmacology, an immune response, and/or off target effects. Following this determination, the toxicologic pathologist should provide an assessment regarding the relevance of the findings to the intended clinical population, usually humans. The toxicologic pathologist may also be asked to assess unusual species and models. Given their broad background in physiology and immunology, toxicologic pathologists are uniquely positioned to provide this input to drug development teams.
Keywords
Introduction
Toxicologic pathologists work with a variety of classes of molecules and devices as they relate to drug discovery and drug development. In terms of molecule classes evaluated by toxicologic pathologists, two main categories are chemical entities (more commonly referred to as small molecules or compounds) and biological products (more commonly referred to as biologics or biotherapeutics). Historically, small molecules have formed the basis of the pharmaceutical industry, are produced through a synthetic chemical process (which often involves joining molecules together, hence the use of the term compound to describe them), and typically have a molecular weight of less than 500 daltons (hence the term small molecules).
With the advances in cell culture methodologies and molecular biology that have occurred over the past several decades, biologics have become as common as small molecules within the portfolios of the pharmaceutical industry. I believe the broad background of toxicologic pathologists, with our extensive knowledge of physiology and the immune system, allows us to be in a key position to support biologics discovery and development. However, to optimally provide this support, toxicologic pathologists should be aware of some of the fundamental differences between small molecules and biologics. This commentary will discuss some of the differences between biologics and small molecules that are of particular relevance to toxicologic pathologists, and highlight emerging issues of relevance to biologics.
What Are Biologics?
What are biologics? They are an amazingly diverse group of products whose common feature is that they are made using biological methods (which may or may not utilize recombinant DNA technology). They are typically derived from well-characterized cells through the use of a variety of expression systems (including bacteria, yeast, insect, plant, and mammalian cells) and can be produced using cell cultures, animals, or plants (ICH S6 1997). Biologics can be composed of sugars, proteins, or nucleic acids or complex combinations of these substances (FDA 2012a). Biologics are typically larger than the chemicals that make up small molecules; for example, an IgG monoclonal antibody is approximately 150,000 daltons. Examples of biologic products include proteins, vaccines, whole cells, or even tissues. While biologic molecules may be used in their native structure, they are often modified, sometimes significantly, to alter function and/or pharmacokinetics. Examples of structurally modified biologic molecules include small amino acid changes in antibodies (e.g., to alter complement dependent cytotoxicity), fusion proteins such as etanercept (a fusion of the Fc portion of an antibody to a tumor necrosis factor receptor to improve half-life), and bispecific antibodies (that contain portions of two different monoclonal antibodies and bind two different antigens, thereby potentially bringing two cells together). While these are examples, the variety of modifications currently being evaluated is enormous. With these alterations come the potential that the modified structures will be seen as foreign, resulting in an immune response (discussed in more detail below).
In addition to molecules that are only small molecules or only biologics, combination or hybrid molecules exist, where a small molecule is coupled with a biologic. In some cases, this is done to modify the pharmacokinetics (usually to increase exposure, e.g., with polyethylene glycol [PEG]). In other cases, a toxin is attached to the biologic (i.e., antibody-drug conjugates [ADCs]), as a mechanism to specifically target a toxin to a particular cell type and reduce nonspecific toxicity.
Because of the immense diversity of products within biologics, with their associated different drug discovery and development paradigms, the remainder of this commentary will focus on a discussion of protein-based biologics.
Mechanism of Toxicity and No Observed Adverse Effect Levels (NOAELs)
One of the fundamental differences between small molecules and protein-based biologics relates to the underlying mechanism of toxicity. With small molecules the mechanism/mechanisms of toxicity often involve direct pharmacologic effects and/or effects related to activity outside of the intended pharmacology (off-target effects). Furthermore, metabolism can play a major role in small molecule toxicity, resulting in creation or elimination of one or more toxic moieties. Species differences in metabolism can significantly affect the development or abrogation of toxicity. It is typically expected that doses high enough to result in toxicity will be tested in small molecule development programs.
In contrast, toxicity is not always observed following administration of protein-based biologics. In fact, in some cases (in particular biologics inhibiting soluble cytokines), no effects may be seen
Whether such effects should be labeled as toxicity, or should factor into the calculation of a NOAEL, is the matter of some philosophical debate. This has been driven in part by occasional interactions with global regulatory agencies where it has sometimes been difficult to convey the difference between an anticipated exaggerated pharmacologic effect and an unexpected toxicity. As a case example based on actual data, an immunosuppressive resulted in a systemic infection. However, this only occurred in one monkey in the low-dose group; there were no other findings other than the expected pharmacologic effect on a subpopulation of leukocytes at two higher dose levels (the highest dose level being 10× the lowest). If one considered the findings in this low-dose animal as adverse and used them to set the NOAEL, then one would conclude a NOAEL had not been determined, as adverse effects were seen at the lowest dose. From experience, some regulatory agencies may not allow clinical trials until a NOAEL is established and may require an additional toxicity study. An alternative strategy, and one that I have used successfully, is to state that there were no primary toxicologic effects that were adverse, and (in this case) state that the NOAEL for primary toxicity was the highest dose tested. At the same time, one should also state that there was one animal with mortality considered secondary to exaggerated pharmacology, so that all information is transparently presented. Such differences in presentation of the data can have significant impacts on regulatory acceptance, and toxicologic pathologists should be aware of the importance of how they write their reports.
Dose Levels
The toxicologic pathologist may be asked to comment on dose-level selection in toxicity studies. In general, for protein-based biologics that do not have substantial toxicity, a low dose is selected that matches human exposure, a high dose is selected that provides an exposure multiple of at least 10× [ICH S6(R1)], and a mid-dose is selected that is in between (often the geometric mean). This paradigm differs from typical small molecule programs, where one is expected to demonstrate toxicity, even if very high doses are needed. It should also be noted that, if the protein-based biologic is very potent or is excessively potent in the normal animals used in toxicity studies (such as might occur with replacement of clotting factors to support hemostasis), lower doses may be needed. In fact, there may be cases where the normal animal cannot tolerate the exposures that are expected in human patients, and high dose in animals provides an exposure multiple of <1. The toxicologic pathologist, with his or her expansive knowledge of disease mechanisms and physiology, can play a key role in the appropriate dose selection before the study and in interpretation of findings and assessment of risk to humans at the end of the study.
Dose Responses
With small molecules, the toxicologic pathologist is often presented with a typical dose response relationship for findings, where higher doses result in greater effects. In contrast, the relationship of dose to response can be quite variable in toxicity studies with protein-based biologics. In studies without significant immune responses to the test article, one may see a typical dose response relationship based on pharmacologic effects. However, in many cases, a full pharmacologic effect may be present at more than the highest dose level, or at all dose levels, and in these groups any observed effects may be similar. It is also possible that immunogenicity may impact the pharmacologic effects, as discussed in more detail below. Thus, the dose response pattern observed with small molecules may not be observed with protein-based biologics.
Immunogenicity
One issue that commonly occurs with protein-based biologics, but less often with small molecules, is the development of immune responses to the drug. Immune responses are typically thought of as relating to an adaptive immune response and development of antidrug antibodies (ADAs; immunogenicity), and in that sense the potential effects related to the immune response are not typically seen until after 10 to 14 days of dosing (or longer). However, effects related to activation of the innate immune system can occur after the first dose (or subsequent doses), and for example, may include complement activation or development of an acute phase response (Clarke 2010). In rare cases, preexisting antibodies to the test article can be present and may result in effects after the first dose. As mentioned above, the alterations in structure of protein-based biologics may result in the molecule being detected as foreign, with the development of an immune response, as can the presence of protein (drug) aggregates in the dosing solution (Rosenberg 2006). However, alterations from the native structure do not necessarily result in an immune response, and the development of an immune response is dependent on many factors, both innate and external (Schellekens 2002). At the present time, prediction of immune responses is an area of intense interest. While progress has been made, I do not believe there are consistently accurate, validated methods of predicting immune responses in animals or humans based on the structure of the molecule.
When an immune response occurs, in my experience it often is seen to a greater extent at lower doses, resulting in an inverse dose response relationship for immunopathology, but with retention of pharmacodynamic effects at higher doses. Thus, it is possible that the lack of effects observed at lower doses may be related to a lack of exposure, and interpretation of findings in light of the toxicokinetics and ADA data can be critical. Reasons for greater immunogenicity at lower doses can be variable and may include pharmacologic immune suppression that reduces or prevents immunogenicity at higher doses, large amounts of test article that overwhelm the immune response at higher doses, and/or induction of tolerance at higher doses. An important point related to detection of an immune response is that, in most ADA assays, the presence of the test article will interfere with detection of ADAs (Ponce et al. 2009). Thus, an animal may test negative for ADAs, even if ADAs are present, because of assay interference related to ADA–test article interactions. Study interpretations by toxicologic pathologists should consider this possibility and evaluate the exposure data in individual animals when necessary.
If an immune response to a test article occurs, the effects related to immune responses on the intended pharmacology can be quite varied (Ponce et al. 2009). If the immune response abrogates the activity of the test article, for example through the development of antibodies that neutralize and/or clear the biologic, then one can observe a reduction or complete loss of pharmacologic effects over time (Hu et al. 2011). Depending on the particular test article, target, and immune response, one may still observe a full pharmacologic effect immediately after dosing, but without as sustained an effect as would otherwise be expected. Alternatively, although rare, it is also possible that the immune response might result in enhanced or prolonged activity of the test article. Enhancement or prolongation of activity could occur if an antibody response was not clearing or neutralizing, but instead increased the time in circulation of the active molecule without blocking its activity. Another rare event can be the neutralization of the endogenous molecule following development of ADAs to a replacement biologic. This can have devastating consequences to the animal or human patient, as occurred in rare instances with immune responses to erythropoietin therapy resulting in aplastic anemia (Randolph et al. 1999; Verhelst et al. 2004). The toxicologic pathologist must take into account all of the information available when making study interpretations, and evaluation of the ADA and exposure data should be part of the study assessment.
In addition to affecting the pharmacologic activity to an administered biologic, an immune response can result in hypersensitivity reactions (Kumar et al. 2010), typically in the form of either immediate (type 1) or immune complex–mediated (type III) reactions. With immediate reactions, these may take the form of classical anaphylaxis, with effects occurring within minutes after dosing. However, experience suggests effects may not be observed for up to an hour or more after dose administration. Whether such reactions are related to classic IgE-mediated anaphylaxis, non-IgE-mediated mechanisms, other immune reactions, or combinations of these is not usually clear. In rodents, these reactions may simply result in mortality with limited or no clinical signs. In monkeys or dogs, affected animals typically will display clinical signs, and if they are severe enough, elective euthanasia may be appropriate. It may also be possible to premedicate animals, for example with an antihistamine, to reduce or prevent reactions. Interpretation of the data should take into account any potential drugs given to prevent such immune responses, and for this reason more powerful premedications such as corticosteroids are generally not used. Alternative dosing regimens, such as more frequent dosing, may also be used to reduce these reactions (Rohde et al. 2012), and do not potentially confound interpretation by concurrent medications. The other hypersensitivity reaction most commonly observed in toxicity studies is immune complex–mediated injury related to deposition of immune complexes in tissues. Such immune complexes are typically thought of as forming from ADAs binding to the administered biologic. However, it is also possible that aggregated monoclonal antibodies with intact effector functions in the dosing solution could fix complement upon administration and result in circulating immune complexes that could deposit in tissues. Alternatively, the aggregates could lodge in tissues and activate the immune system
Regardless of the mechanism/mechanisms of immune complex formation, in my experience glomerular injury and vasculitis following deposition of immune complexes in these tissues are the most common expressions of immune complex disease following administration of protein-based biologics in toxicity studies. Demonstration of immune complexes in affected glomeruli or blood vessels can be useful to attempt to identify the cause. However, it should be noted that immune complexes can be detected in normal animals, and a comparison to concurrent controls (usually limited in number) and knowledge of historical controls (to provide more information on background) is often required to make an appropriate interpretation. In some cases, the deposition of immune complexes in glomeruli can result in findings such as proteinuria and hypoalbuminemia. However, in other cases the glomerular injury does not result in clinicopathologic alterations despite the presence of morphologic changes. Sometimes the test article itself can be demonstrated in the findings using antibodies that detect the test article specifically, for example, using antibodies that detect the human biologic that do not cross react with the animal molecules (i.e., anti-complementarity determining region [CDR] antibodies, or specific anti-human Fc antibodies). While immediate or immune complex–mediated reactions are the most common hypersensitivity reactions seen in toxicity studies, it is possible that other types of hypersensitivity reactions could occur.
Additional evaluations can be used to help determine whether an immune response to the drug is occurring. I have used pharmacodynamic activity assessments, evaluation of exposure and toxicokinetics, determination of ADAs, assessment of complement activation (CH50, C3b, etc.), determination of circulating immune complexes, and evidence of inflammation from hematology, clinical chemistry, or cytokine analyses to determine whether an immune response is occurring. Reduced levels of complement receptor 1 (CD35) on erythrocytes, responsible for clearing immune complexes in the blood, may also be seen. More sophisticated methods are also possible (although they do not always work), including immunologic or proteomic analyses of circulating immune complexes or tissue immune complexes to ascertain whether the test article is present. In the end, a determination of whether an immune response is affecting individual animals or a group is often based on a weight of evidence approach.
While classic hypersensitivity reactions can been seen, in many cases, I have experienced that animals do “not read the book,” and the interpretation of findings as being related to a hypersensitivity reaction may not be as clear as hoped. It must be emphasized that immune responses are variable, and not all findings on an individual animal or group basis may have a classic set of findings. Effects may not consistently occur between doses, such that an animal might display moderate signs, minimal signs, and then severe signs on consecutive doses. In the end, the study scientists must use a weight of evidence approach to determine the cause of the findings. Such determinations can be critical to the continued development or termination of a program. For example, a glomerular change based on the pharmacological mechanism may not be acceptable, while a glomerular change based on an immune response to the test article with deposition of immune complexes might not be considered predictive of what would happen in humans and might allow the program to progress.
In addition to variable reactions between animals and groups, the presence of new infections, or recrudescence of latent infections that may occur secondary to pharmacologic immunosuppression, or “normal” background findings, can potentially confound the interpretation of hypersensitivity reactions. In some cases, the etiologic agent is readily apparent, for example bacteria infiltrating the intestinal tract, an overwhelming
When interpreting suspected immune-mediated effects, scientists must consider the potential synergistic or antagonistic effects of pharmacology coupled with an immune response. When the test article is an immunomodulatory drug, there may be variable pharmacologic impact on the ability of the animal to develop an immune response to the test article at different doses, with potentially greater levels of immune suppression or enhancement at higher doses. Layered on top of this variability is the potential for an immune response to the test article to alter the pharmacology (usually reducing the pharmacologic effect). In my experience, such interactions can vary significantly between individual animals and groups and can make interpretation difficult. For example, within a single group one might have some animals with a pharmacologic response, other animals have lesser or no pharmacologic effect because they developed clearing or neutralizing ADAs, and other animals develop immunopathology secondary to immune complex deposition. Any range of these scenarios is possible. When such a scenario develops in a monkey study in which there are only a small number of animals, and there are background changes present that are often seen in monkeys, the interpretation can be quite challenging. The reasons for such individual animal variability are likely complex, but are probably related to a combination of genotype and environmental factors. It should be noted that, while immune responses can occur in toxicity studies, those responses in animals are not considered predictive of what will occur in humans (although this does not mean that they will not occur; Bugelski and Treacy 2004). In that sense, in most cases the findings related to immune responses in animals are not considered to be “show stoppers” and a reason to halt drug development. Therefore, it is critical for toxicologic pathologists to differentiate a finding related to an immune response that is not predictive for humans from a finding related to pharmacology (possibly exaggerated) that is more likely to be predictive. That said, immune responses can also occur in humans, and when they do occur, they can sometimes result in findings that are similar to what is seen in animals.
While determination of whether effects are related to an immune response can be challenging, I use some general guidelines related to immunogenic responses in animals that may be useful when interpreting data and determining the potential cause for findings:
Immunogenicity related to an adaptive immune response usually occurs somewhere after 10 to 14 days. However, findings may not be observed until much later (possibly not for many doses or many months after the first dose).
Immunogenicity may occur using homologous proteins.
Immunogenicity may not occur using nonhomologous proteins. It may be possible to dose rodents with a human protein in chronic studies without development of a detectable immune response.
Immunogenicity (based on clinical signs and pathology) is often greater at lower doses.
Immunogenicity can occur in a bell-shaped dose response curve.
Morphologic tissue injury is much less common than the development of immunogenicity.
Effects can vary significantly between animals and groups within and between studies.
Responses in an individual animal can vary throughout a study, with the severity of reactions waxing and waning despite administration of the same dose level.
Administration of human protein-based biologics in the context of toxicity studies (high doses, frequent doses) often does not result in immunogenicity.
More frequent dosing (and/or higher doses) can sometimes overcome immunogenicity in the context of toxicity studies (Ponce et al. 2009; Rohde et al. 2012).
A weight of evidence approach is usually necessary when determining whether findings in animals and groups are related to an immune response.
If rodents, and in particular mice, are a relevant pharmacologic species, it may be possible to do some interesting studies to attempt to dissect out whether an effect is related to primary pharmacology or an immune response. For example, if a suspected immune-mediated effect occurs in normal mice which develop ADAs, but not in nude mice which do not develop ADAs, it can provide evidence that the immune system is involved in the finding (Leach 2010).
One additional thought related to the immune system is that a pharmacologic effect on the immune system, an immune response to the test article, or a combination of these may result in some form of immune dysregulation. As different populations of immune cells are variably deleted, blocked, stimulated, and so on, it is easy to conceive that some form of immune dysregulation might occur. For example, the anti-CD52 monoclonal antibody alemtuzumab has been associated with autoimmunity, and it has been proposed that residual lymphocytes may expand into the depleted lymphocyte compartment left by therapy and enhance the risk of autoimmunity (Bell, Reynolds, and Isaacs 2011). This is an area that has received limited attention to date.
Combination Molecules
As noted above, some protein-based biologics contain small molecule elements, and both small molecule and biologic strategies must be employed in the development of these agents, including considerations for genotoxicity and carcinogenicity when appropriate for the indication. Whenever possible, a determination on whether observed effects are related to the biologic or the small molecule component/components should be provided. This usually should be based on studies that evaluate the individual components of the agent, either concurrently or based on historical data. For example, PEGylation of protein-based biologics can result in vacuolation in cells in a variety of tissues that is not related to the mechanism of the drug (EMEA 2009), and this is often known based on historical information. However, in my experience, novel PEGs may have different effects from those that have been used in the past, such as less reversibility of vacuolation. In the case of ADCs, which is an area of great interest at the present time, the test article consists of an antibody (or antibody-like molecule), a toxic payload, and a cleavable or noncleavable linker that combines them. The goal of the ADC is to precisely target a toxic payload using the antigen binding (Fab) domain of an antibody to specific cells. The mechanism of action of the payload, antibody binding characteristics, amount of drug loaded onto the antibody, location of linking, and linker characteristics are some of the variables that can be modulated in an attempt to maximize the effects on the target cells and minimize the effects on nontarget cells. Given the number of variables, complex interactions that are not yet well understood can occur. While the antigen-binding portion of the antibody is designed to provide specificity, the Fc portion of the ADC may also lead to uptake by nontarget cells expressing Fc receptors and may lead to unwanted effects in these cells. In practice, despite the goal of specificity, the toxic payload often results in tissue damage unrelated to the intended specific target of the antibody (FDA 2011). Administration of ADCs to species which lack pharmacologic activity associated with the antigen-binding portion of the test article can highlight what effects are unrelated to binding to the intended target, although effects related to Fc binding may still be seen. Administration of the toxic payload or conjugated linker–payload alone can highlight the toxicity of these moieties, unrelated to any Fab or Fc targeting by the antibody portion of the ADC. Companies may develop linker–payload platforms, thus developing a database on the selected linker–payloads that can be used for new ADCs to minimize additional testing and provide greater certainty that the linker–payload will be acceptable. However, even when the toxicity of a given payload or linker–payload has been well characterized, experience has shown they may behave differently when conjugated to different antibodies, likely related to the variables noted above. In addition, one must consider potential effects of impurities or reactive metabolites related to either the small molecule linker and/or payload, and these may require an assessment consistent with that of typical small molecules.
Tissue Cross-reactivity Studies
The tissue cross-reactivity (TCR) study is unique to antibody and antibody-like molecules and is not done for small molecules. The role of TCR studies in biologics drug development has been a topic of significant debate in the past several years. Opinions ranged from believing these studies provide no value, to believing they are an essential part of drug development. In the past decade, regulatory agencies were requesting TCR studies be conducted across a wide range of species, including those in which there was no demonstrated pharmacology, and then requesting
Reproductive Evaluations
The toxicologic pathologist may become involved in aspects of drug development that have not traditionally been part of standard toxicity evaluations. A current topic of great interest in the development of biopharmaceuticals is the appropriate evaluation of developmental and reproductive toxicity, particularly in cases where the test article only has activity in monkeys. Over the past several years, the study design for testing developmental toxicity in monkeys has become relatively standardized in what is termed the enhanced prenatal and postnatal toxicity study (Weinbauer et al. 2011). In these studies, dams are dosed from early gestation (GD20 in cynomolgus monkeys) until birth, and sometimes into the first month after birth. Infants are usually necropsied at up to 6 months after birth, and the assessments of the infants often include some pathology evaluations that involve the toxicologic pathologist. For example, if the test article is an immunomodulator, a detailed evaluation of the immune system organs may be included, including immunohistochemistry. If dams, fetuses, or infants experience mortality, the toxicologic pathologist will likely be asked to evaluate the animals. In this regard, an understanding of normal fetal and infant development is very helpful, as internal controls may not be available from unscheduled time points. However, this study design does not address potential effects on fertility. To address fertility (for programs where fertility is relevant and rodents are not a pharmacologically-relevant species), it is now recommended that a toxicity study of at least 3 months duration be conducted using sexually mature animals (ICH S6(R1) 2011). Industry is currently developing standards for assessing maturity pretest and effects on reproduction during the in life phase of the studies. At the end of such studies, toxicologic pathologists will usually be asked to confirm that the animals were mature. This is in contrast to what happened historically, where immaturity was only sometimes noted, and usually only in males. An assessment of maturity can be based on a combination of gross examination, organ weights, and microscopic examination. For males, the presence of larger reproductive organs, higher organ weights, and microscopic evidence of mature testes as well as sperm in the epididymides are good markers. For females, one can look for evidence of ovulation in the ovaries or evidence of maturity in the endometrium. These assessments may require the collection of additional endpoints in chronic studies that may not have been routinely collected in the past (e.g., not all companies routinely collect weights from ovaries, uterus, epididymides, seminal vesicles, etc.). Toxicologic pathologists should be prepared to make this assessment routinely in studies in the future and may be asked to make these assessments retrospectively if chronic studies have already been completed.
Cytokine Release Assays
Another relatively new assay in which toxicologic pathologists developing biologics may become involved is the cytokine release assay (or cytokine storm assay). The addition of this assay to nonclinical testing of biologics is the result of a clinical trial in which an agonistic anti-CD28 antibody (TGN1412) resulted in severe clinical reactions in all six volunteers that were dosed (Suntharalingam et al. 2006). Because the nonclinical testing of TGN1412 had not identified this risk, a cytokine release assay is now considered a required component of nonclinical development programs for molecules that are high risk (i.e., those that are agonists, bind to immune cells, etc.). Some companies conduct this assay on all new antibody programs regardless of their target, although in my opinion this is not necessary. The recommended assays are somewhat complex (Stebbings et al. 2007), and revisions to these assays have also been proposed (beyond the scope of this commentary). Industry as a whole has not yet developed a standard set of assays, or determined exactly how to use them. The toxicologic pathologist may be asked to comment on the risk to humans based on the results of cytokine release assays. In addition, cytokine concentrations are being measured in the blood in toxicity studies with greater frequency, and these results require interpretation. Unfortunately, there is limited information about how differences
Alternative Species and Models
Because of the specificity of protein-based biologics, toxicity studies may be run in species that are outside of those typically used. Examples include using the common marmoset, guinea pig, minipig, hamster, and rabbit in general and/or reproductive toxicity evaluations. Transgenic rodents can be used as well. They may be evaluated in a phenotyping manner, for example evaluating a “knockout” to determine the potential worst-case effect of antagonizing a cytokine or a “knockin” to determine the potential worst-case effect of a replacement therapy (Cavagnaro 2010). Human receptors, immune cells, cytokines, and so on, can also be inserted into mice, and then those animals can be dosed with the test article. Furthermore, in some cases the target may only be expressed in a disease model, or the toxicity may occur to a greater extent in the disease model compared with normal animals; in these cases, toxicity studies may be run in animal models. In rare cases, a surrogate (homologous species) molecule may be used in toxicity studies. Thus, the toxicologic pathologist may be asked to evaluate studies in unusual species with limited or no historical controls, in disease models, in transgenic animals, in studies using surrogate molecules, in studies using combinations of these options, and in the end determine the relevance (or lack thereof) of any findings to humans. In these cases, it is very important to have good study designs to develop strong concurrent control data.
Central Nervous System Effects
Although it is tempting to believe that protein-based biologics will not enter the CNS and cause effects because of the blood brain barrier, in my experience this is not true. Some percentage of the protein-based biologic in the peripheral blood will enter the CNS, even with larger proteins such as monoclonal antibodies. How much depends on a variety of factors, including the molecular size, presence of receptors that might allow entry into the CNS, and how close targets in the brain might be to areas that are more “leaky.” Whether the amount of the biologic that enters the CNS is enough to cause an effect will depend on factors such as the potency of the antibody on the target and the number of target epitopes that need to be engaged to cause an effect on the cell. Thus, small amounts of protein-based biologics in the CNS could cause CNS effects. As an example, dose- and time-related tremors and convulsions have been described for a monoclonal antibody for a non-CNS indication, but which also had the target in the CNS (Cannon 2012).
Biosimilars
Biosimilars are currently an area of great interest in the pharmaceutical industry worldwide. In contrast to small molecules, which can be identically copied, biologics are large and complex (especially monoclonal antibodies) and cannot be copied exactly. While not an exact copy, a biosimilar molecule is a very close physicochemical copy with highly similar biological effects. The availability of regulatory guidance on biosimilars varies significantly between countries/regions, with some having finalized documents (e.g., European Medicines Agency (EMA), World Health Organization (WHO)), others having draft documents (e.g., United States), and others having no guidance (e.g., China, Russia). Most guidance suggest that, when
ICH S6(R1)
As discussed above in several places, the recent addendum to ICH S6 [ICH S6(R1)] has clarified a number of topics regarding the development of biologics. In particular, the addendum discusses rational species selection, general toxicity and developmental and reproductive toxicity study design, carcinogenicity assessment, and TCR study design and interpretation. One item to highlight that differs from small molecules is the opportunity in some cases to conduct general toxicity studies of 6 months duration in only one species, even if there is pharmacologic activity in both a rodent and a nonrodent. Furthermore, that study can be in a rodent species if it is pharmacologically relevant and demonstrated effects in shorter duration studies that were similar to those seen in the nonrodent species. Another interesting comment is that “recovery from pharmacological and toxicological effects with potential adverse clinical impact should be understood when they occur at clinically relevant exposures.” This statement appears to suggest that evaluating recovery at exposures that are higher than those seen in the clinic may not be warranted, which is somewhat counter to the frequent strategy of including recovery at the highest dose level only. Another option would be to include recovery groups at the high dose and at a lower dose level that is clinically relevant. If findings did not recover at the high dose, but did at the clinically relevant dose, this might provide an easier path forward in development. It remains to be seen how this will be applied within the pharmaceutical industry. Overall, I believe the addendum provides a very useful update to ICH S6, it should responsibly minimize animal use, and toxicologic pathologists evaluating biologics should be familiar with the document, as well as the original ICH S6.
Summary
In summary, when findings are observed in studies following administration of protein-based biologics, the toxicologic pathologist should initially determine whether the effect/effects are mediated (directly or indirectly) via the intended pharmacology, exaggerated pharmacology, an immune response, and/or off target effects. Following this determination, the toxicologic pathologist should provide an assessment regarding the relevance of the findings to the intended clinical population, usually humans. The toxicologic pathologist may also be asked to assess unusual species and models. Given their broad background in physiology and immunology, toxicologic pathologists are uniquely positioned to provide this input to drug development teams.
Footnotes
*This is an opinion article submitted to the Regulatory Forum and does not constitute an official position of the Society of Toxicologic Pathology or the journal
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author received no financial support for the research, authorship, and/or publication of this article.
Abbreviations
Acknowledgments
The author would like to thank the many colleagues who provided helpful suggestions and input, in particular Natasha Neef, Ken Frazier, and the other Regulatory Forum Committee members; and Robert Veneziale.