Abstract
The determination of “safe use conditions” for the use of new therapeutic agents in clinical trials is a complex undertaking that involves collaboration among many development scientists who are trained in various scientific disciplines (e.g., biopharmaceutics, biochemistry, immunology, metabolism, molecular biology, pharmacology, pharmacokinetics, pathology, toxicology, and medicine). How different companies and regulatory agencies integrate and review the massive amounts of information produced across these disciplines varies widely; however, the end objective is the same, which is to conclude whether, on the basis of data produced in laboratory and animal settings, a new pharmaceutical agent can be safely administered to human volunteers for the purpose of assessing human safety and to enable further clinical development of new medicines.
Two articles appearing in this month’s journal critically review the TeGenero incident (Expert Scientific Group on Phase One Clinical Trials 2006; Suntharalingam et al. 2006), which brought into question current practices for determining and predicting human safety. The first article (Horvath and Milton 2009) raised the issue of whether the TeGenero incident was avoidable or just the consequence of a series of unfortunate events. The second article (Milton and Horvath 2009) reviews the associated European Union regulatory response to this event, which culminated in the recent European Medicines Agency (EMEA) guideline (EMEA 2007) that addresses first in human dosing considerations. In addition, this article assesses the impact of this event and new regulatory guidance on the various scientific disciplines involved in preclinical safety assessment programs. Both articles are balanced in their review of the circumstances around the TeGenero event and the subsequent actions taken by regulatory authorities and pharmaceutical developers and make many useful points; I would recommend that they be carefully read and considered by the readership. As a scientist who has spent his career working in the area of preclinical safety assessment and having responsibility for determining safe use conditions for clinical trials, I would like to offer the following additional perspective.
In 1997, the ICHS6 guidance document, which deals with the design of preclinical safety evaluation programs for biologics, was officially released as a “Step 5” document (EMEA, 1998). I was a member of the expert working group that authored this guidance document, which has been very important in the history of biologic drug development because it stated for the first time the internationally accepted scientific principles that needed to be considered by scientists during the design of preclinical safety assessment programs of biopharmaceuticals. ICHS6 offers specific program design recommendations in areas of pharmacology, toxicology, and pharmacokinetics and describes certain unique considerations that must be dealt with for the vast majority of biologics. A central component of ICHS6 advocates a “case-by-case” approach, which acknowledges that one program’s preclinical study design may not look like another due to product-specific biology, unique issues, or concerns. Today, we have over a decade of experience across a wide range of product classes in which the principles described in ICHS6 have been successfully used. For those who are new to the field or for those who are very experienced, these principles remain valid. Safety assessment programs designed by experienced scientists and based on this guidance have successfully supported the safe development of biopharmaceuticals across the globe.
Additional scrutiny on ICHS6 and other guidance documents occurred in the wake of the recent TeGenero incident, in which a cohort of Phase I normal volunteers treated with a single dose of TGN1412 (an anti-CD28 humanized antibody) developed severe adverse drug reactions. Some have considered this unfortunate outcome as an additional indication that existing considerations used to design preclinical programs for the support of the safe conduct of clinical trials were outdated and in need of revision. This conclusion is wrong, as I believe that the TeGenero incident was avoidable. To this point, a careful review of the published literature and nonredacted data that existed at the time and were released for public review shortly after the incident supported the following relatively simple conclusions: (1) the agonist anti-CD28 action of a homologue antibody (JJ316) in rats led to a quick, dramatic polyclonal stimulation manifested as splenomegaly and splenic lymphocytosis by day 3; and (2) the sponsor (TeGenero) in fact reported a delay in the kinetics of the minimal T cell lymphocytosis induced by TGN1412 in the nonhuman primate compared to that observed with the homologue in the rodent. This data set alone portrays a major pharmacodynamic (PD) difference and should have prompted a detailed discussion around the “relevance” of the species selected for safety assessment testing and an explanation of any inconsistencies between observations across species.
In many experienced companies, including my own, this data set would have prompted additional work and/or caution. At a minimum, the basic safety implications of large PD discrepancies suggest that the biology of the primate and the rodent is different, or there is something technical that needs to be better understood in order to explain the PD difference. Both of these points would offer major safety “red flags” that should have warranted a high degree of caution in defining the development path forward. Unfortunately, the human starting dose was based on doses used in the primate model, which in this case was not pharmacologically sensitive to the antibody. As described in much greater detail by Horvath and Milton, the “totality” of the data needed to be considered at this point, and for reasons that are uncertain, it was not. This apparent omission underscores the importance of the scientific rationale for species selection as noted by the authors. This point cannot be understated as to its central importance to the adequacy of preclinical programs designed to support human safety.
TeGenero is an unfortunate outlier from the broad experience of biologic drug development and should not, in itself, warrant a complete redesign of nonclinical development program requirements for biologics or small molecules. Going forward, we must always keep in mind that biologics have unique safety issues that differ from small molecules drugs. If preclinical programs are well designed and all data carefully considered by experienced preclinical scientists, the vast history of experience to date supports the conclusions that laboratory and robust toxicology assessments can provide predictive information regarding human toxicities. Not unexpectedly, the TeGenero incident prompted the expeditious development and release of a new guidance document in the European Union that specifically addresses in a detailed manner preclinical program design considerations for first-in-human trials for new chemical and biologic products (EMEA 2007). This recent guidance document is comprehensive and well written, and it should be consulted in conjunction with ICHS6 by those developing new pharmaceuticals. The additional points raised by Milton and Horvath (2009) are a useful supplement to this guidance and should be seriously considered by those working in the field of safety assessment.