Introduction

In contrast to small molecular weight drugs (SMD), the development of in vitro-in vivo correlations (IVIVCs) for monoclonal antibody (mAb) therapeutics is still in its infancy. High-throughput in vitro assays with established IVIVC that currently exist for SMD have not been so strongly pursued for antibody (Ab) therapeutics, reflecting to a certain degree the substantially lower attrition rate and the associated lower numbers of potential clinical mAb candidates. This chapter represents one of the first attempts at summarizing the available data correlating preclinical in vitro and in vivo data of mAbs.

In recent years, the correlation between in vitro and in vivo properties of chemically derived SMD has become an established methodology to predict single or multiple components of in vivo pharmacokinetics (PK), pharmacodynamics (PD), safety/toxicity, and efficacy (1-3). Accessibility to physiologically relevant in vitro assays has fostered the development of IVIVC involving critical PK, PD, safety, and efficacy data, allowing prediction of in vivo behavior prior to in vivo studies. In fact, because of the often high-throughput capabilities of such assays, thousands of new compounds can be screened for their potential to become viable clinical drug candidates in a highly efficient manner. Given the relatively high attrition rate of SMD, the pressure to combine in vitro screening tools with established IVIVC has increased the demand for tools that facilitate a more successful and efficient selection process of drug-able clinical candidate molecules. Besides the impact on the overall drug discovery and development productivity, in vitro assays provide mechanistic insight into subprocesses that are often not easily assessable under in vivo conditions. Complex mechanistic processes can be isolated into mechanistic subprocesses at organ, cellular, and subcellular levels. One example of a mechanistic dissection of complex in vivo processes is the development of in vitro methodologies to characterize the essential components of PK, namely absorption, distribution, metabolism, and excretion (ADME). Predictive in vitro and even in silico tools have been recently developed to obtain estimates for individual ADME parameters that describe the manner in which drug candidates are being absorbed, distributed, and eliminated under in vivo conditions. In a second step, the predicted ADME parameters can be combined to estimate the overall in vivo PK as a composite of the individual ADME processes. Physiologically based pharmacokinetic (PBPK)

16 Boswell et al.

modeling tools have been developed to predict the in vivo PK solely based on in vitro and in silico ADME data together with established physiological information required to describe the mammalian body (4,5). This concept of predicting human PK on the basis of in vitro and in silico data using mechanistic PBPK models has been successfully implemented in the realm of drug discovery and development, thereby partly replacing empirical prediction methods. The establishment of IVIVC derived from mechanistic models has resulted in a somewhat paradigm shift in that the human PK of SMD is now predicted primarily on the basis of in vitro and in silico data, whereas in vivo data from preclinical animal species are rather used to verify prediction success as surrogate for human prior to the actual prediction of human PK. In addition to IVIVC established for PK parameters, very successful applications of similar types of methods are known in the area of PD, safety, and efficacy of SMD (see section "Modeling of Antibody In Vitro-In Vivo Correlations").

In contrast, IVIVC for mAb therapeutics are not as well established as that for SMD. Still, the concept of developing and using IVIVC involving mAbs is not a new one. Seminal work reported by Day et al. in 1974 probed the relationship between in vitro binding and in vivo localization of 125I-labeled Abs; the results of this work helped elucidate the blood-brain barrier (BBB) concept for these molecules (6). In present day, the selection of appropriate in vitro and in vivo biological assays in evaluation of biotechnological products remains a critical challenge (7-9). In theory, relationships between in vitro and in vivo data may encompass a broad spectrum of situations ranging from qualitative explanations to true statistical evaluations. In practice, however, the vast majority of cases are best described as qualitative correlations or trends, as demonstrated by the diverse examples discussed herein.

There are more than 20 therapeutic Abs, fragments, and immunoconju-gates approved by the Food and Drug Administration (FDA) (Table 1), and hundreds more are in development (10-14). For a mAb to enter clinical trials, it must be first tested preclinically using a battery of methods; several of these are listed in Table 2 (15,16). Because of the interdependencies that exist between these and other in vitro and in vivo phenomena, it is possible to establish correlations between various combinations of these measurements. For example, the in vitro information on binding characteristics is of particular interest considering that the Ab PK and PD can be influenced by crystallizable fragment (Fc) receptor binding and several other factors (11,17). In addition, the binding affinity to cognate target antigens and the level of target expression are known to influence PK/PD behavior.

Within this chapter, we will initially present a brief overview of mAb structures and functions, followed by an examination of the specific approaches used to correlate in vitro-in vivo data related to PK, PD, and metabolism of mAb therapeutics. Because of the growing number of mAb derivatives that are being generated, the quantity of correlative data is very large, making it difficult to discuss all study cases in this work. As such, representative examples involving mostly preclinical studies throughout the literature will be described in this chapter, but an emphasis on more recent work will be apparent in most cases. In this context, we will examine three major in vitro properties: (i) binding,

(ii) potency, and (iii) cellular metabolism. In parallel, we will consider four major in vivo phenomena: (i) binding and/or targeting, (ii) PD and/or potency,

(iii) PK, and (iv) in vivo metabolism. Because IVIVC can be assessed using

TABLE 1 Current Approved Antibodies for the Parenteral Use in Treatment of Disease

Generic name

Trade name

Binding target/antibody class and isotype

Indication

Approval

Naked Antibodies

0 0

Post a comment