INTRODUCTION
With three approvals and another ~100 in clinical
development, bispecific antibodies (BsAbs) represent an important class
of therapeutic modalities.1, 2 The intent of BsAb
therapy is for a single molecule to interfere with multiple disease
pathways by recognizing two different epitopes or antigens. These
interactions can expand and prolong the efficacy of these modalities in
complex disease indications. Another attractive quality of BsAbs is
their potential to provide novel functionalities that do not exist in
mixtures of the parental antibodies leading to synergistic biological
effects. Given it has become clear that many disorders including cancer,
metabolic diseases (including diabetes and cardiovascular illnesses),
and autoimmune diseases display multiple and/or redundant mechanisms
that fuel their progression, BsAbs have the potential to provide
increasingly effective therapeutic options to patients compared with
antibodies and other therapeutic entities that interact or modulate the
activity of a single target.3-5 Innovation in the
field of protein engineering and advancements in technology have led to
the design enablement of over 100 BsAb formats.6 While
some BsAbs are simply smaller proteins comprised of two linked
antigen-binding fragments, a number of other BsAbs formats leverage the
basic modular nature of the IgG structure. The IgG-like BsAb molecules
consist of subunits on individual antibodies attached to an
agonistic/antagonistic mAb that impart the ability to bind dual soluble
or membrane bound ligands or a combination of both. These formats
include DVD-Ig, cross-mAbs, IgG-extracellular domain (ECD) and IgG-scFv
constructs.
Despite their exceptional therapeutic promise and structural
tractability, the translation of BsAbs as medicines has been relatively
slow compared to mAbs.2 For example, the dual activity
of T cell redirection and engagement was described approximately
>30 years prior to the 2009 launch of catumaxomab
(withdrawn in 2017 for commercial reasons) and more recently
blinatumomab (approved 2017) and amivantamab (approved 2021) both for
treatment of cancer. The first BsAb approved outside of oncology is
emicizumab for the treatment of haemophilia which also occurred more
recently in 2017. Similar to most antibody therapeutics, the causalities
of the slow clinical success for BsAbs can be generally related to
several factors, including an incomplete understanding of the biological
mechanism of action, poorly defined exposure-response profiles,
insufficient safety margins, strategic industry decisions and
immunogenicity. The increased inherent structural diversity and
tractability BsAbs afford relative to mAbs also leads to potentially
greater uncertainty in their pharmacokinetic and disposition profiles.
Thus, in addition to the aforementioned challenges, unpredicted aberrant
pharmacokinetic profiles requiring increased empirical protein
engineering can also limit the potential advantages BsAbs offer
pharmacologically relative to classical monospecific mAbs.
As a means to mitigate poor pharmacokinetics for mAbs, several studies
have reported leveraging preclinical in vivo and in vitrophysiochemical characterization-based PK developability strategies
during the discovery process.7-9 These approaches have
been used to improve the probability of success by selecting or
engineering mAbs with increased stability (physical, chemical and
thermal stabilities) and lower non-specific or unintended
interactions.10, 11 Improving the stability and
lowering the risk of unintended interactions, in turn provides enhanced
human exposure profiles to support the intended dose and frequency of
administration. Indeed, we and other groups have reported connecting
preclinical pharmacokinetics with various physiochemical
characterization along with FcRn interaction analyses in an integrated
manner to inform the selection and engineering of mAbs with optimized
pharmacokinetic profiles.10-16 Our laboratories have
extended these approaches to some BsAbs that utilize IgG-ECD and
IgG-scFv formats.10, 13 These studies revealed that
poor physiochemical properties in some BsAb formats contributed to
increased clearance rate, driven by endothelial cell-based
association/clearance mechanisms in the liver; moreover, the studies
showed that engineering the structural configuration of the ECD
mitigated aberrant pharmacokinetic behavior of the BsAbs. While these
initial studies lay an important foundation for understanding non-target
related factors influencing the disposition and pharmacokinetics of
BsAbs, there remains a paucity of data and an incomplete understanding
of the balance between the in vitro physiochemical factors andin vivo physiological mechanisms that influence the peripheral
clearance and disposition of BsAbs. Moreover, while previous studies
were able to connect physiochemical properties to IgG-ECD BsAb
pharmacokinetics in a post-hoc analysis, there remains considerable
opportunity to define the relative contribution of the various
non-target related factors influencing the non-specific clearance of
another BsAb format in an a priori manner. With these points in
mind, we designed the present study to evaluate the physiochemical
properties and connectivity of these with in vivo mechanism(s)
involved in the clearance of two IgG-scFv constructs (deemed BsAb-1 and
BsAb-2; Figure 1A) using preclinical models.
The IgG-scFv constructs were made with scFv units and mAbs targeting two
distinct soluble ligands having minimal peripheral concentrations in
normal animals, so that the in vivo kinetics and disposition
could be evaluated in the absence of target mediated drug disposition
(TMDD). The Fab (Fab-1) region of BsAb-1 binds to the same ligand as the
scFv (scFv-1) component of BsAb-2 and relatedly, the Fab-2 region of
BsAb-2 binds to the same target as scFv-2 in BsAb-1 (Figure 1A). While
both BsAbs orientations were imparted with the same antigen binding
properties, we observed rapid clearance (~2 mL/hr/kg) of
BsAb-1 and acceptable clearance (~0.2 mL/hr/kg) of
BsAb-2 in cynomolgus monkeys. Characterization of the two BsAbs revealed
differences in physical and thermal stability profiles, as well as, in
FcRn based interactions. The evaluation of the biodistribution of the
two molecules in cynomolgus monkeys indicated distribution to the same
organs to the same quantitative extent, but BsAb-1 was more rapidly
cleared from tissue. Taken together, the in vitro and in
vivo data indicate that the inferior physical stability properties and
the poor release of BsAb-1 from FcRn at neutral pH are likely linked to
its aberrant clearance in cynomolgus monkeys. The observation is
mechanistically distinct than the proposed increased hydrophobic
interaction findings alone that led to aberrant kinetics observed for
other BsAb formats in the earlier studies, highlighting the complexity
of the issue.10, 13 The findings in this report
confirm the need for continued evaluation and delineation of the balance
between factors influencing the disposition and pharmacokinetics of
various BsAbs and the interplay of the BsAb format on these parameters.