Abstract
Pathology is the science of using changes in tissues, cells, or body fluids to understand organic responses to injury, and in the case of toxicologic pathology injury by xenobiotics, in which animal models are used to understand the effects of these drugs and chemical compounds for their potential impact on humans and other species. As such, toxicologic pathologists are tasked with the interpretation of their data relative to comparative risk assessment, in addition to the description and diagnosis of the changes seen in animal studies.
Recently in our journal, the “safety function” of the non-clinical safety disciplines has been put into perspective, in an article discussing the recent EMEA guidelines for first in human (FIH) clinical trials (Milton and Horvath 2009). In their article, the authors emphasize that the need to focus on risk identification and risk mitigation, before performing a FIH study, is a somewhat new approach. Previously, the risk characterization approach had been focused primarily on hazard identification (a description of the danger, i.e., what are the target organ[s] and adverse effects caused by the agent); dose-response assessment (the dose at which a toxic effect is likely [LOAEL] or unlikely [NOAEL] to occur); exposure assessment; and risk characterization (the probability of a potential adverse effect when exposure is defined). Typically, once the potential risks have been defined and semi-quantified, risk management and risk mitigation follow. Today’s toxicologic pathologist needs to be acutely aware of the distinction between hazard identification and semi-quantitative risk characterization.
To ensure consistency in defining the hazard and in characterizing the risk, the toxicologic pathology community has dedicated innumerable hours attempting to standardize their nomenclature. These days there is every opportunity for the toxicologic pathologist to be involved in management decisions concerning risk characterization and mitigation. As the data that we generate are used in many ways for decision making, we need to be intimately and directly involved in developing the guidelines for their use, beyond simply using the results to stop compound development.
The Therapeutic Index (TI), the ratio of the effective dose to the dose at which adverse effects are noted, provides a rough estimate of the potential risk and is an essential calculation for risk mitigation strategies. Increasingly, management and investors make multimillion-dollar decisions on the progression of compounds, or lose millions more on the elimination of compounds based on the TI; hence, it behooves our discipline to be aware of, and participate in, the risk management process.
The terms used in defining a cutoff threshold include the lowest observable effect level (LOEL) and the no observable effect level (NOEL). These informational terms indicate a threshold at which any event occurs: background, confounding, or biologically significant. However, the uncommonly used term—lowest toxic effect level (LTEL)—allows for exaggerated pharmacologic effects, which may not assist in an adequate characterization of the potential health risks. This term introduces the fact that at very high doses, as required by toxicology studies, activation of pharmacology targets may reach a point of being truly “adverse, ” even though the affected targets may not be relevant to the discovery of off-target effects at the physiologically/pharmacologically relevant therapeutic dose levels. Many of the pharmacologic effects could be primary, but there are many secondary effects that damage the host and are logical indirect consequences of an exaggerated pharmacologic effect sustained for longer periods of time.
The NOAEL and the LOAEL are calculations that are germane to our discipline. One of the responsibilities of the senior toxicologic pathologist is to define what effect is adverse, in addition to fully reporting all lesions. Such a task is commonly difficult and may lead to issues arising from the background, indirect, or incidental lesions present in a study. The challenge we all face is defining what constitutes an adverse effect, be it a clinical, a clinical pathologic, an anatomical finding, or a process that involves some or all of these. A review of these concepts was published in Toxicologic Pathology in 2002 (Lewis 2002). Since toxicologic evaluation is not a stochastic, deterministic process, the subjective nature of such an assessment means that professional opinion has to be used, and that this opinion is based on professional beliefs acquired during their prescribed training period and subsequent on-the-job experience, but acknowledgement needs to be made for the considerable differences that exist in an individual’s training and job experience. Nevertheless, the toxicologic pathologist is the person most likely to be relied upon to provide an informed opinion of what constitutes adverse for the purpose of risk assessment and management.
It is of interest to note that the regulatory agencies also have difficulty in defining adverse! For example, the lesions seen in toxicology studies following administration of a biological drug to a species other than the intended recipient may be owing to the foreign host’s reaction to the biological material, rather than a true toxicity per se. It is for this reason that the risk assessment of adverse effects performed in many biological products is approached on a case-by-case basis using a semi-quantitative, and highly subjective, “weight-of-evidence” approach. This approach underscores the importance of toxicologic pathologists expressing their subjective experience to assist teams of scientists when making decisions concerning adverse effects associated with the compound under review.
Regulatory authorities use the NOAEL in determining thresholds for toxicity. The United States Food and Drug Administration defines the NOAEL as the highest tested dose, or concentration, of a substance that results in no observed adverse effect in exposed test organisms where higher doses or concentrations would result in an adverse effect. Similarly, the NOAEL is used for dose selection for FIH phase I studies. The use of the NOAEL, following the application of acceptable margins, to calculate a starting dose is assumed to give a reasonable expectation of safety in single-dose and multiple-ascending-dose FIH trials. Unfortunately, using the NOAEL is not always appropriate. The possibility of errors of omission became patently clear in 2006, when eight healthy volunteers were dosed in an FIH study of the agonistic anti-CD28 mono-clonal antibody [mAb], TGN1412 (TeGenero) (Horvath and Milton 2009). For a review of this event, we recommend the article referenced above in the March issue of Toxicologic Pathology (Lin et al. 2004).
Errors of commission are commonly incurred in the use of NOEL and NOAEL, in that the extrapolation of the effects is often performed without a full understanding of species differences in susceptibility to the compounds, the species specificity of the response, the higher or lower susceptibility of the different species to the effects of the compound, and even sometimes the influence of procedures, such as dosing route or frequency, on the outcome of the study. For example, pulmonary histiocytosis is often observed in gavage studies in different species, owing to the lung irritancy of the compound unintentionally introduced into the airways by the prosector or the animal itself. Although these compounds are intended for use by oral administration in humans by different formulations, these findings have been considered adverse, as they may not be observed in the controls (personal experience). As in this example, there are many other such examples documenting that “adverse” can be interpreted in very dissimilar ways by different scientists on the basis of their differing personal training and/or work experience.
In the TeGenero incident, a conventional NOAEL of 50 mg/kg body weight was derived from studies on nonprimates, and the starting dose for the FIH was set at 0.1 mg/kg, using a margin of 500 times lower. It has been suggested that the starting dose in FIH trials should be 1/100th the pharmacological dose, based on appropriate animal data, or a dose in or below the low microgram range but not exceeding 100 μg (Cohen 2007), which at face value suggests a very cautious approach. This suggestion was unfortunately not followed in this instance, with tragic consequences.
To avoid similar incidents in the future, the European Medicines Agency in 2007 adopted the use of the “minimum anticipated biological effect level” (MABEL), designed to improve the selection of a safe FIH dose (EMEA 2007). The MABEL also considers NOAEL derived from nonclinical studies, the dose response for the adverse effect[s], and pharmacodynamic and pharmacokinetic relationships defined using in vitro studies in human cells or tissues in vitro, and in vivo studies in animals.
The MABEL is calculated using the effect of the compound rather than the tolerability. As the primary pharmacological effect of the drug is not often monitored in animal toxicology studies, serious toxicity may occur in humans who respond to the biological stimulus at doses far lower than any NOAEL established from animal studies (Goodyear 2006). The starting human dose at MABEL is a dose that does not cause any clinically measurable effects, any nonclinical animal effects, or any effects in in vitro studies. Therefore, dosing should come close to the pharmacological activity, but below the minimum effective dose or pharmacologically active dose; however, establishing this dose must take into consideration all nonclinical data including the in vitro efficacy studies. The next higher dose given is expected to cause the first pharmacological effects in healthy volunteers without a toxic effect, and this is considered to be the MABEL. The use of a MABEL approach is, as a consequence, a very conservative one in projecting an FIH starting dose, and although avoiding the toxicity inherent in a molecule will not predict an idiosyncratic reaction, the probability of finding idiosyncratic reactions in a small FIH study is remote.
So what have we learned? The NOAEL will continue to be the main method of determining the starting dose in FIH trials for small molecules in Europe. The role of the toxicologic pathologist is central to determining what constitutes adverse, and the incorporation of species differences into the risk equation is an improvement over the methods that were used previously. So far neither the International Committee on Harmonization nor the national regulatory agencies have officially adopted the MABEL approach. However, in many cases, because of the difficulties of determining compound-related adverse effects in the nonclinical assessment of highly active biologic compounds, the NOAEL method may not be the most appropriate approach for every compound. In such cases, the MABEL approach should seriously be considered. Therefore, professional organizations such as our own have a golden opportunity to participate with regulatory agencies in addressing this issue and ensuring an outcome that satisfies both the science and safety considerations of FIH studies.
The Society of Toxicologic Pathology can play an important role by educating its membership on the best way to implement these changes in regulation. The long-awaited discussion on what constitutes an adverse effect is a treasure trove of learning and can produce significant improvement in the way we view study findings. As we are uniquely positioned to understand and communicate the issues of comparative pathophysiology, these data will become increasingly important to help in overall risk management.