Considerations Regarding Nonhuman Primate Use in Safety Assessment of Biopharmaceuticals

Abstract

Selection of a pharmacologically responsive species can represent a major challenge in designing nonclinical safety assessment programs for many biopharmaceuticals (eg, monoclonal antibodies (mAbs)). Frequently, the only relevant species for nonclinical testing of mAbs is the non-human primate (NHP). This situation, coupled with a rapidly increasing number of mAb drugs in development, has resulted in a significant increase in the number of NHPs used in nonclinical safety assessment. Apart from ethical considerations related to responsible animal use, there is a clear need for more efficient and innovative approaches to drug discovery and development; these factors drive the need to investigate alternative approaches and strategies for the safety assessment. This review summarizes important scientific and regulatory perspectives derived from presentations and audience discussions in an educational forum at the 2010 annual American College of Toxicology meeting regarding opportunities for employing alternative approaches to minimize NHP use in mAb drug development.

A Collaborative Approach to the Use of NHPs in mAb Development

Drug development is increasingly expensive and currently estimated to cost about $1 billion for a drug to reach market approval. ¹ This high cost and the lack of productivity in the pharmaceutical pipeline, particularly for novel targets, create significant challenges.^2–4 The pharmaceutical industry has invested heavily in biopharmaceuticals, such as monoclonal antibodies (mAbs), to fill this gap and to provide therapies in areas of unmet medical need. The number of mAbs in development is rising rapidly^5–7 and due to the high specificity and limited species cross-reactivity of these drugs, the nonhuman primate (NHP) is frequently the only relevant species for safety assessment. The majority of NHPs used in research are for pharmaceutical research and development^8,9 and the growing pipeline of biopharmaceuticals, such as mAbs, is driving an increase in NHP use worldwide. Both cost and ethical considerations require that industry and regulatory scientists ensure the responsible use of NHPs in biopharmaceutical development.

Many companies are working together to share data on mAb development programs in order to understand ways to minimize the use of NHPs. By translating this experience and knowledge into practical advice and recommendations, the efficiency and predictivity of drug development can be improved. The most practical published recommendations recognize the unique challenges of mAb development compared to those for small molecules and were developed to help identify the pertinent scientific questions for any specific mAb. For example, when is the use of the NHP necessary, and which study designs are the most appropriate? Ultimately, the aim of nonclinical studies is to provide the most relevant information on safety and for predicting human risk. The most efficient way to achieve this goal is to use as few animals as possible without compromising the scientific objective of the study, human safety, or regulatory acceptance.

Novel approaches that minimize the use of NHPs in preclinical development of mAB have been identified. When based on solid scientific rationale, these can include (i) incorporating safety pharmacology end points into repeated-dose toxicity studies, (ii) minimizing the number of dose groups on chronic toxicity studies, (iii) conducting a maximum of 2 general toxicology studies (1 month or longer)—one to support first-in-human (FIH) trials and one to support registration, (iv) avoiding the conduct of multiple chronic toxicity studies, (v) the inclusion of recovery animals in fewer dose groups, (vi) utilization of study designs that minimize the number of animals used in developmental and reproductive toxicity (DART) studies, and (vii) the use of only the rodent for chronic toxicity studies, when appropriate.^10–17

A number of these strategies are also reflected in recent regulatory guidelines that cover mAb drug development. For instance, International Conference on Harmonization ¹⁸ ([ICH S9] nonclinical evaluation for anticancer pharmaceuticals) eliminates the need for stand-alone safety pharmacology studies, allows registration using studies of 3 months in duration (rather than 9-12 months), and does not require specific combination safety studies to allow combination dosing in the clinic. This guideline also allows for developmental toxicity in a single relevant species in certain circumstances: if the clinical molecule shows appropriate pharmacological activity in the rat, it may be used rather than the NHP and does not require fertility studies to be conducted. The ICH M3 (R2) guidance ¹⁹ enables developmental studies to be conducted during phase III for certain molecules; thus, studies are conducted later in development for molecules with a higher probability of success. There is also a specific comment allowing for conduct of embryo-fetal studies during phase III where exposure during organogenesis is understood to be low.

Similarly, the draft addendum of ICH S6 (S6[R1] ²⁰ ) represents translation of current scientific knowledge in mAb development into regulatory guidance and reflects the principles of reduction of animal use and more responsible nonclinical testing programs. The ICH S6(R1) changes that impact NHP utilization include acceptance of in vitro alternatives for some types of molecules/therapeutic indications; elimination of the need for studies with nonrelevant species simply to examine off-target toxicity; elimination of stand-alone safety pharmacology studies unless there is specific cause for concern; the potential ability to use a single species for longer-term toxicology studies if the toxicity profile in the short-term studies is the same in a second relevant species; the expanded pre- and postnatal study design to combine reproductive and developmental toxicity studies; and elimination of the need for recovery animals in all dose groups, as well as the notation that recovery simply to assess for the presence of antidrug antibodies is not necessary. Many of the approaches discussed in this article are consistent with the current ICH S6(R1) guidance.

The following sections describe the presentations and resulting discussion from an American College of Toxicology (ACT) workshop ²¹ in which the speakers and audience explored novel approaches and case studies to ensure quality drug development while minimizing NHP use. These debates and whether the novel approaches may contradict regulatory expectations have been gathered into general recommendations to aid study directors and other scientists involved in mAb development.

Considerations for the Reduction in NHP Use in the Drug Development Paradigm: Can 1 + 1 Really Equal 4? (L. A. Burns-Naas, M. D. Todd)

The economics of the pharmaceutical industry requires efficiency without compromising quality. Accomplishing this goal often involves “out-of-the-box” innovative thinking to identify opportunities for improving efficiency and effectiveness. Frequently, this requires cross-functional collaboration, trust, and a willingness to take a little risk for the greater good of all those involved. Science—in particular, regulatory science—is no different. Protection of patient safety is paramount, but this does not mean that we cannot be more judicious in our animal use, particularly for large animals such as dogs and NHPs. This section provides suggestions for improving the responsible use of animals in research while delivering new drugs to patients sooner and at lower cost. Some of these opportunities are already being implemented at the regulatory level (see previous section).

Beyond regulatory guidance-driven opportunities, there are other considerations that sponsors can employ to reduce animal use. One obvious option already in use is to move away from terminal single-dose studies to evaluate toxicokinetics (TK). Consequently, animals are more often returned to a stock colony for reuse. Animals used for mAb TK are typically used subsequently only for evaluation of nonprotein therapeutics, though data are being developed that might provide insight into the potential for reuse of NHPs for multiple protein therapeutics. Another obvious option is to reduce the overall number of repeated-dose toxicity studies performed. Often sponsors run 1-, 3- and then 6-, or 9-/12-month studies over the course of the drug development paradigm. With very careful planning, data from 1-month studies can be used to support dose selection for a 6-month study. For NHPs, this could reduce the numbers for a given program by 30 to 60 animals, depending on the final study designs. There are clear risks associated with this approach, and a good working knowledge of the TK as well as the dose-response nature of the observed toxicities is required. Inadequate dose selection for the chronic study could necessitate repeating the study, although this risk could be reduced by employing 4 (rather than the usual 3) dose levels on the 1-month study. While the additional dose group increases animal usage on that study, there is still a net reduction when considering that the need for an intermediate-duration study could be eliminated. Sponsors would also need to work closely with their colleagues engaged in manufacture of the active pharmaceutical ingredient (API). At least 1 definitive toxicity study should be performed with the API with the final commercial route of manufacturing, and some companies wait until later in development to finalize the process or route. If a sponsor elected to go from 1- to 6-month studies for the toxicology program, it is likely that the 6-month study would be needed proximal to the start of phase II, much earlier than some companies might normally have the commercial route selected. However, in some circumstances, a “bridging strategy” might be employed by conducting definitive reproductive studies (inclusive of primary target organs and key nontarget organs) with material from the commercial route and demonstrating that effects (or lack thereof) are similar to those observed in the repeated-dose study with an earlier lot of material.

It is not uncommon for drug candidates to progress through the drug discovery phase only to encounter unacceptable toxicity or unmanageable margins of exposure when the dose-ranging or FIH-enabling studies are performed. There are several ways that sponsors can help facilitate better survival of the molecule and thus, reduced animal usage. The proverbial “bottom line” is, get the most out of the work you do. When feasible, safety end points can be included in early pharmacokinetic/pharmacodynamic (PK/PD) or other efficacy studies. Also, target knowledge reviews can be conducted at the target selection stage and, where appropriate and scientifically justified, an early, but limited, in vivo target de-risking safety evaluation can be conducted. Two examples were provided which addressed these points.

The first example involved an agonist mAb that was cross-reactive in humans and NHPs only. The target was expressed at very low levels on normal cells but was upregulated in the disease state. A cause for concern existed around the potential for cytokine release and the duration of immune activation (or the inability to downregulate the immune response). A murine homologue was assessed and differences were observed as compared to the clinical candidate in human cells. A biomarker strategy was also being developed in NHPs to monitor the PD effect. To help address the safety concerns, the sponsor elected to add specific safety end points to the nonterminal biomarker study including clinical signs, serum chemistry, hematology, immunophenotyping, specific cytokines, complement, and immune response assessment. In addition to identifying a biomarker for clinical use, the sponsor gained confidence in the safety of the target modulation and moved into FIH-enabling studies in the NHP. One consideration for these types of studies with a mAb is that repeated dosing can sometimes influence the safety profile. When feasible, it is recommended to include at least a second dose on a PK/PD study and evaluate the safety end points 1 to 5 days after the second dose.

The overarching principle of early target de-risking is that the earlier safety issues are identified, the greater the potential reduction in the overall animal use for the individual project. An obvious mechanism to improve target selection and differentiation for both efficacy and safety is through developing a deep knowledge of the target and pathway by conducting literature and competitive intelligence reviews, gathering or generating knowledge around target expression in humans and animals, and evaluating the target using a variety of available tools (RNA interference [RNAi], genetically engineered mouse models, pathway profiling, etc). The goal is to identify potential risks that should be monitored in toxicity and efficacy studies and which may require an early target de-risking evaluation to differentiate chemotype effects (eg, off-target effects resulting from the toxicity of the chemical itself) from on-target-related liabilities. Such early de-risking studies are typically conducted in vitro/ex vivo or in small rodents (if pharmacologically responsive/relevant). However, when scientifically justified, a small pilot NHP study can provide key safety data to redirect or terminate candidates or targets. In the second example, an early safety evaluation for a mAb which did not cross-react in rodents was performed in NHPs due to significant cardiovascular safety concerns for the target. The study utilized 6 NHPs and resulted in a “No Go” decision for the program. Had the sponsor elected to take this candidate into an FIH-enabling study, 48 NHPs would have been utilized to achieve the same objective.

Ideally, toxicology studies should be performed with the clinical candidate in the most relevant species to maximize the potential to identify adverse effects in humans. When this is not possible (eg, lack of cross-reactivity) or when scientifically justified, one can consider the use of alternative models to evaluate safety (eg, homologous proteins, knock-in/knock-out models, transgenic animals, and animal models of disease). Homologous proteins—molecules that recognize the equivalent target in another species—may be the most commonly used alternative approach to date. The advantages and disadvantages of each of these alternative approaches are the subject of a recent review article. ¹⁵ The clear challenges with these alternative models are (i) timing—the decision to use them should occur early in development and in consultation with global regulatory authorities, if possible; (ii) resources—material and/or animal needs can be significant, assay development is required, and there is an ultimate need to demonstrate comparable biology with the clinical candidate; (iii) data interpretation—for example, are the toxicology findings relevant? Is antidrug antibody formation an issue?; and (iv) regulatory expectations—there is a lack of any established regulatory criteria or guidance around the acceptability and use of these alternative models. Finally and significantly, the surrogate is not the clinical candidate.

So, when considering the question “Can 1 + 1 = 4?” (“Can we do more with less?”), the answer is clearly, “Yes!” As scientists, we always need to be thinking innovatively. The responsible use of animals in research is an important ethical consideration in drug development, and there is a global commitment in both industry and regulatory agencies to reduce the numbers of NHPs (and animals, in general) used in drug development while avoiding any compromise in patient safety. While full replacement is not yet feasible, we should strive to reduce animal use whenever possible. This can be done without compromising effective and efficient science.

Need for and Timing of DART Testing in NHPs (P. L. Martin)

The NHPs are used for the evaluation of DART when they are the only pharmacologically or physiologically relevant species. Pharmacological relevance can be established using in vitro cell-based systems or in vivo if there is a measurable pharmacological response. ²² If rats and rabbits are pharmacologically relevant species, then these are the species of choice for DART. ²² However, some biopharmaceuticals, especially mAbs, are pharmacologically active only in humans and NHPs. For these molecules, an assessment of DART in NHPs may be required.^14,23,24 The NHPs may also need to be used if the biopharmaceutical is too immunogenic in a nonprimate species to conduct a meaningful toxicological evaluation. Finally, the target biology may be different in primates versus nonprimates, in which case the NHP may be the only biologically or physiologically relevant species.

The primary regulatory guidance that provides a recommended basic framework for the nonclinical safety evaluation of biotechnology-derived pharmaceuticals (biopharmaceuticals) is ICH S6. With regard to DART, ICH S6 states that “the need for reproductive/developmental toxicity studies is dependent upon the product, clinical indication and intended patient population.” However, the original ICH S6 ²⁵ did not provide any additional guidance on study design. Although not stated in ICH S6, it was presumed that ICH S5 would be referred to for additional information regarding DART study designs. However, as ICH S5 ²² was not written with a focus on study designs using NHPs, some additional text regarding such studies has been proposed to be included in an addendum to ICH S6 (ICH S6 [R1]). ²⁰ A major focus of the ICH S6 (R1) addendum was to support the reduction, refinement, and replacement of animals in research (3Rs) especially with regard to the use of NHPs. The recommendations include (1) evaluation of reproductive potential by histopathological evaluation of reproductive tissues in sexually mature NHPs as part of a repeated-dose toxicology study of at least 3 months in duration in place of separate fertility studies and (2) the conduct of a single developmental toxicology study (expanded peripostnatal development [ePPND] study) that evaluates all stages of development rather than conducting separate embryo/fetal development and pre- and postnatal development studies. The ICH S6 (R1) addendum further proposes that a single species may be adequate for the evaluation of effect on embryo/fetal development. If rats and rabbits are pharmacologically relevant, then DART studies should be conducted in those species per ICH S5 (R2). However, if the rats and NHP are pharmacologically relevant but not the rabbit then the DART program can be conducted in rats only, thereby eliminating the need to conduct developmental studies in NHPs. However, it is important to understand the physiological and pharmacological species differences and to provide a scientific justification for why the studies in rodents, for example, adequately address human safety. Further, it may be necessary to adapt study designs in order to cover periods in development relevant to potential human exposure.

The primary regulatory guidance document that describes a framework for the nonclinical detection of toxicity to reproduction for medicinal products and toxicity to male fertility is ICH S5 (R2). This guidance provides a “most probable option” for study designs when rats and or rabbits are pharmacologically relevant species. However, the guidance states that the testing strategy should be determined by the anticipated use (especially in relation to reproduction), the form of the substance, and route of administration intended for humans and should leverage any existing data on toxicity, PDs, kinetics, and similarity to other compounds in structure/activity. It further states that “to employ this concept successfully, flexibility is needed and that all persons involved should be willing to discuss and consider variations in test strategy according to the state-of-the art and ethical standards in human and animal experimentation.” The flexible testing approach emphasized in ICH S5 (R2) is consistent with the case-by-case approach emphasized in ICH S6. ²⁵ This is particularly important to consider for NHP DART studies.

The ICH M3 (R2) guidance on nonclinical safety studies for the conduct of human clinical trials and marketing authorization for pharmaceuticals describes the types of nonclinical studies that are required for development of pharmaceutical drugs. For biotechnology-derived products (biopharmaceuticals), ICH M3 (R2) provides guidance only with regard to the timing of the nonclinical studies relative to clinical development. For products that are intended to be administered to men or to women of reproductive age, an assessment of male and female fertility/reproductive potential is expected to be completed before the initiation of large-scale or long-duration clinical trials (eg, phase III trials). For most products a repeated-dose toxicology study of at least 3 months in duration would be completed to support phase III clinical trials. If the NHP is the only pharmacologically relevant species and if the toxicology study included sexually mature animals with histopathology of the reproductive organs, then no additional NHP studies would be required to meet this requirement (ICH S6[R1] ²⁰ ). Therefore, inclusion of sexually mature NHPs in repeated-dose studies conducted prior to phase III can eliminate the need to evaluate reproductive potential in separate studies and thus reduce the use of NHPs.

With regard to the inclusion of women of childbearing potential (WOCBP), for molecules that show target cross-reactivity in rats and rabbits, preliminary reproduction toxicity data in these species together with precautions to prevent pregnancy in clinical trials are generally sufficient to support the inclusion of WOCBP (up to 150) in clinical trials of relatively short duration (up to 3 months). This is based on the very low rate of pregnancy in controlled clinical trials of this size and duration. The agreement between the United States, European Union, and Japanese regulatory agencies regarding the sufficiency of preliminary embryo-fetal development studies reduces the need for a large number of animals to be tested prior to understanding whether a molecule has any potential for clinical safety or efficacy. Definitive studies should be conducted prior to phase III.

Studies can be conducted in WOCBP without the completion of nonclinical developmental toxicity testing if there is extensive knowledge of the mechanism of action of the agent or if there is difficulty in conducting developmental toxicity studies in an appropriate animal model. This statement would apply to biopharmaceuticals that are pharmacologically active only in NHPs. Because of their large molecular size, nonantibody biopharmaceuticals are not expected to cross the human placenta to any great extent, ¹⁴ and any potential effects on the embryo/fetus are likely to be secondary effects of maternal or placental toxicity. For mAbs, although placental transfer is expected during the fetal period, transfer during the period of major organogenesis is likely to be low. ²⁴ Therefore, for mAbs, the type of embryo-fetal study described in ICH S5 (R2) that limits dosing only to the period of major organogenesis would provide only limited information. For antibodies, the greatest degree of exposure occurs during the fetal and the postnatal periods which are generally addressed in pre- and postnatal development studies. ²⁴ Pre- and postnatal studies are only required to be completed prior to registration (ICH M3[R2]). Therefore, delaying the developmental toxicity assessment in NHPs until phase III and dosing animals throughout pregnancy (eg, ePPND study design) is an acceptable approach so long as precautions are taken to avoid pregnancy in the clinical trials and the patients are informed of any potential risk based upon the mechanism of action of the molecule. Delaying the developmental evaluation until phase III reduces testing in NHPs for those molecules that do not proceed to phase III. Also, conducting a single study that covers all stages of development rather than a segmented approach can reduce the number of NHPs needed for the developmental toxicity evaluation.

Two case studies were provided to demonstrate a reduction in NHP use in DART studies in practice. In the first case, a human αv-integrin mAb was being developed for the treatment of solid tumors. As the clinical indication was oncology, the only DART study required for a biopharmaceutical according to ICH S9 is an embryo-fetal development study in a single species. This mAb showed similar binding affinity to human and cynomolgus monkey αv-integrins but 20-fold lower affinity to rabbit integrins and 40-fold lower to rat integrins. Despite the lower affinity in rabbits, PK and PD modeling showed that rabbits could be dosed using a regimen sufficient to produce sustained high blood levels of the antibody that were in excess of 10-fold greater than the concentrations required to saturate the rabbit αv-integrins. ²⁶ The affinity for the rat integrins was considered too low to be practical for toxicology testing. Therefore, the rabbit could be used as a suitable pharmacologically relevant model for evaluation of potential effects of the human antibody on embryo-fetal development, thereby eliminating the need to conduct an embryo-fetal development study in the cynomolgus monkey.

Second, a human anti-Type 2 helper T cells (Th2) cytokine mAb was being developed for the treatment of asthma. As asthma is a chronic indication that includes patients of reproductive age, a full DART program is required to support this indication (ICH S5 [R2] and ICH S6). In vitro characterization studies showed that this antibody bound to and neutralized the pharmacological actions of both cynomolgus monkey and rat cytokine but did not neutralize the rabbit cytokine. Furthermore, this mAb was shown to be efficacious in a rat model of asthma. Because the rat was a pharmacologically relevant species, the entire DART program was conducted in rats, thereby eliminating the need for DART assessments in cynomolgus monkeys.

A Regulatory Perspective on the Use of the NHP Model in Safety Assessment of Biologics (J. A. Lansita)

Toxicity studies should be designed to provide an adequate safety assessment for clinical development while minimizing the numbers of animals utilized. Toxicology testing in NHPs should be conducted only when absolutely necessary. The ICH S6(R1) ²⁰ clarifies specific aspects of the nonclinical development of biologics including species selection, study design and duration, immunogenicity, reproductive and developmental toxicity, and carcinogenicity. This section briefly summarizes the key points in ICH S6(R1) that may be used to streamline the conduct of nonclinical studies and potentially reduce the numbers of animals including NHPs used for the development of biologics. At the time these presentations were given, ICH S6(R1) was at step 3 of the ICH process; step 4 has now been finalized, and there are no differences from step 3 to step 4 regarding adherence to principles of reduction in animal use. The nonclinical development of biologics should be evaluated using a case-by-case approach, and sponsors are encouraged to consult with the appropriate Food and Drug Administration (FDA) reviewing division for specific advice.

A thorough evaluation of target expression and activity in rodent and nonrodent species is critical to ensure that meaningful nonclinical studies are performed in pharmacologically responsive species to assess the safety of biologics. In the case of mAbs with activity to a foreign target, the toxicology package can be limited to a safety evaluation in a disease model or, when no disease model is available, to a short-term safety study in a single species. In one case study provided, the sponsor conducted exploratory toxicology studies in cynomolgus monkeys and rabbits with longer duration studies in the rabbit. The sponsor proposed to conduct a pivotal toxicology study in the monkey and reproductive toxicology in the rabbit. Both the rabbit and monkey showed similar toxicology profiles. The FDA reviewing division questioned the following: (1) Which nonrodent species, the monkey or rabbit, should be used for pivotal nonclinical studies for a mAb to a foreign target? (2) Is reproductive toxicity testing for a mAb to a foreign target needed? The FDA reviewing division decided that although a study on monkeys would be acceptable, the use of rabbits in the pivotal toxicology study would be preferable, due to the ability to use a higher number of animals per group and compare the results to previously conducted rabbit studies, as well as an ethical concern of using monkeys when a valid alternative model exists. The FDA further determined that reproductive toxicity studies were not necessary in this case since the target is foreign/exogenous, and off-target reproductive organ toxicity was not observed in general toxicology studies.

In the case where both the rodent and nonrodent are pharmacologically responsive and findings in the 2 species are comparable in short-duration (eg, 1 month) studies, the use of rodents alone is preferred for longer-term studies (unless there is appropriate rationale for using the nonrodent species). If general toxicology studies fail to define a no observed adverse effect level (NOAEL) due to observed toxicity, alternative studies to define the mechanism of toxicity and its relevance to the clinical population may provide more useful information than conducting additional NHP general toxicology studies and thereby reduce the number of animals used.

The reproductive toxicity assessment of biologics should be conducted in sexually mature pharmacologically responsive species. If the NHP is the only relevant species, a study in NHPs is generally preferred over alternative models. No further reproductive toxicity testing may be required in cases where the potential for an adverse pregnancy outcome is suggested using a weight of evidence assessment. In the absence of any relevant species, the use of transgenic mice or a homologous protein can be considered. A thorough evaluation of the pharmacological and toxicological relevance of the homologous protein to the clinical candidate should be conducted prior to using a homologous protein in pivotal toxicology studies. For example, in the case of 1 biologic, a homologous protein in rodents showed a different toxicity profile than the clinical candidate in rodents. The clinical candidate was pharmacologically active in monkeys and rats but had lower potency in rats compared with the clinical candidate in monkeys. The sponsor decided to conduct reproductive toxicity studies with both the clinical candidate and a more potent homologue in rats. The use of the more potent homologue identified a potential new safety hazard. Since the sponsor did not provide data to further understand the clinical relevance of the new toxicity with the more potent homologue, the FDA-reviewing division assumed the toxicity was clinically relevant and represented a “worst-case” outcome.

The FDA accepts that stand-alone fertility studies are not recommended when the NHP is the only relevant species; instead, fertility end points can be included in the general toxicology studies. As described in S6(R1), the assessment of embryo-fetal and pre-/postnatal development for products active only in NHPs can be conducted in one study that includes dosing from gestation day 20 through birth rather than conducting separate studies. The number of dose groups in reproductive toxicity studies may be limited to a control group and a high-dose group (with adequate scientific justification), since the main purpose of such studies is to identify a hazard. If effects on conception and implantation are a concern and the NHP is the only relevant species, studies in transgenic animals or the use of a homologous protein may be warranted.

Examples of the use of surrogate or homologous antibodies for assessing reproductive toxicity are described in the US package inserts (USPIs ²⁷ ) for Arcalyst (Regeneron Pharmaceuticals, Inc.), Actemra (Genentech, Inc.), Cimzia (UCB, Inc.), Ilaris (Novartis Pharmaceuticals Corporation), Remicade (Janssen Biotech, Inc.), and Soliris (Alexion Pharmaceuticals, Inc.). For Ilaris (canakinumab ²⁸ ), an antihuman interleukin 1β (IL-1β) mAb, a surrogate antimurine mAb was used for reproductive toxicity studies in mice. Interestingly, delays in fetal skeletal development were observed in mice with the surrogate and were also observed in marmosets with the clinical candidate; these data indicate that a study with the antimurine mAb in mice may have been sufficient to predict the pregnancy risk in patients. Prolia (denosumab ²⁹ ), a human immunoglobulin (Ig) G2 mAb to human receptor activator of nuclear factor kappa-B ligand (RANKL), utilized data in knock-out mice to inform the pregnancy risk to patients. Alternative models can be acceptable to regulatory authorities with adequate scientific justification. Sponsors are encouraged to consult with the appropriate regulatory authority to ensure acceptability of an alternative model prior to conducting pivotal studies.

The Center for Drug Evaluation and Research (CDER)/FDA is a signatory to the approach described in S6(R1) that if a carcinogenicity assessment of a biologic is needed, the assessment often can be based on published data (such as those from human genetic diseases, transgenic animals, knock-out animals, or animal disease models, information on class effects, target biology, in vitro data, chronic toxicity data, and clinical data), thereby reducing the need for additional animal studies. If the weight-of-evidence suggests a concern for carcinogenic potential, the labeling can reflect this concern or additional nonclinical studies may be performed to mitigate the concern. Stelara, Janssen Biotech, Inc. (ustekinumab ³⁰ ), a human mAb to p40 protein subunit that disrupts IL-12 and IL-23 signaling, utilized literature reports on mouse models and a transgenic model to inform carcinogenic potential in labeling. From the USPI, “Published literature showed that administration of murine IL-12 caused an anti-tumor effect in mice that contained transplanted tumors and that IL-12/IL-23p40 knockout mice or mice treated with anti-IL-12/IL-23p40 antibody had decreased host defense to tumors. Mice genetically manipulated to be deficient in both IL-12 and IL-23 or IL-12 alone developed UV-induced skin cancers earlier and more frequently compared to wild-type mice.”

Generally, human biologic products show cross-reactivity to targets in NHPs but not always to targets in lower-order species such as rabbits, rats, and mice. However, in the case of Orencia, Bristol-Myers Squibb Company (abatacept ³¹ ), a fusion protein comprised of the cytotoxic T-lymphocyte antigen 4 (CTLA-4) extracellular-binding domain linked to the modified Fc domain of human IgG1, pharmacologic activity was seen in both rodent and nonrodent species. The clinical candidate was evaluated in both rodent and nonrodent species to support the chronic use of Orencia for adult rheumatoid arthritis and juvenile idiopathic arthritis. The complete toxicology package (submitted to the FDA prior to ICH S6[R1]) included reproductive toxicology in rats and rabbits, juvenile toxicology in rats, general toxicology in rats and NHPs, chronic toxicology in NHPs, and carcinogenicity in mice. In hindsight, based on the experience and lessons learned from the conduct of previous toxicology studies with biologics, it is possible that fewer nonclinical studies could have been conducted to fully assess the nonclinical safety of Orencia today compared with the time at which it was reviewed and approved.

Summary

Often, approaches to nonclinical safety assessment of drugs can appear formulaic in nature, for example, “always use 2 species”; “the NHP is the most relevant to humans”; “use 4 dose groups: control, low, mid, and high”; “use 4/sex/group for NHPs”; “include recovery animals in each study”; and so on. While recent surveys on industry practice regarding NHP use in the development of mAbs showed a range of program and study designs, many appeared to employ more animals than necessary or defaulted to “standard company practices”; there were clear areas of opportunity for optimizing nonclinical programs and individual studies.^16,17

“Case-by-case” is a valued mantra for the design of nonclinical safety assessment packages supporting the development of biopharmaceuticals. In fact, ICH guidance ^18–20 offers many opportunities for case-by-case approaches with possibilities for reductions in the use of NHPs. These and additional recommendations for nonclinical biopharmaceutical development were presented and further illustrated during discussions in the workshop.

Know Your Target

Leverage existing data/literature with a thorough understanding of the target, other molecules of the class, and knock-out/in transgenic models.

Species Selection

Do not assume the NHP is the most relevant or the only relevant species without assessing pharmacologic activity across species. Some differences in the magnitude of target binding can be overcome, as can some immunogenicity responses, with higher doses or modified dosing schedules.

Study Design

Leverage early pharmacology and PK studies; for example, incorporate PK/PD or safety end points in the nonclinical pharmacology (proof-of-concept) studies.

Use of Homologous Molecules

Homologous molecule strategies (eg, employing a homologous protein instead of the clinical candidate in animal models) could be appropriate following careful characterization of the homologue with regard to comparative target affinity, biologic effect, and (for DART) placental transfer properties. Homologues are most commonly employed when there is no relevant nonclinical species or as an alternative to NHP use in, for example, DART assessments. However, such an approach needs to be balanced with the burden of developing an alternative molecule of proven relevance with the more direct approach of using the clinical candidate in the NHP. ²⁰

Design of DART Studies

Rodents and rabbits are the preferred models for DART studies as long as they are pharmacologically relevant and exposure can be demonstrated. Developmental and reproductive toxicity studies could be performed in NHPs; however, issues of low statistical power (resulting from one offspring/mating) and less historical control data need to be carefully considered to ensure that the use of these species is scientifically justified and that the study design will deliver the stated objectives.

Regulatory Recommendations

Regulatory agencies are open to alternative—acceptance of such models and approaches requires adequate scientific justification/rationale. Early engagement with regulatory agencies can reduce the overall time and cost of nonclinical development.

In summary, there is a global commitment in both industry and regulatory agencies to reduce the numbers of NHPs (and animals in general) used in drug development while avoiding any compromise to patient safety. The soon-to-be-finalized ICH S6(R1) addendum will provide further guidance on options for alternative nonclinical safety strategies which may result in a reduction in NHP use. As is demonstrated by the principles and examples discussed herein, nonclinical safety assessment of biopharmaceuticals must be approached largely on a case-by-case basis and driven by an understanding of the science and considerations of human safety with regard to the biology of the molecule and the intended clinical use.