DISCUSSION
During our pharmacokinetic characterization of two related scFv-based
BsAb constructs, we observed an unusually rapid ~10-fold
faster clearance of BsAb-1 relative to BsAb-2. Both BsAb-1 and BsAb-2
constructs were made with scFv domains fused to the HC C-terminal of
IgG4 mAbs. The two BsAbs displayed comparable binding
affinity to the same soluble ligands that had no/minimal peripheral
concentrations in normal animals; moreover, the BsAbs had no specific
interaction with cell surface receptors thus, eliminating both
circulating ligand-mediated and cell surface target-mediated drug
disposition as potential mechanisms for the observed clearance. In
addition, since both BsAbs examined use an IgG4 parental
Fc that has been engineered to eliminate interactions with Fcγ
receptors, direct binding with blood cells is not expected to be a
viable clearance mechanism either. Given the disparity in clearance was
not related to target binding or Fcγ receptor interactions, the focus of
the present effort became delineating the non-target related
physiological mechanism(s) affecting the BsAb in vivo behavior.
Non-target related physiological factors influencing the disposition and
pharmacokinetics of BsAbs are poorly understood. In previous reports we
demonstrated that association with liver sinusoidal endothelial cells
(LSECs) led to the accumulation of BsAbs in liver and was responsible
for the unusually rapid clearance of several BsAbs including some
IgG-scFv constructs.13 Given these findings, LSEC
clearance became an initial plausible mechanism to explore for BsAb-1
and BsAb-2 that could readily explain the atypical elements of the BsAb
clearance. Interestingly, IHC analyses of liver from cynomolgus monkey
PK studies conducted with the BsAbs showed that LSECs were not the root
cause of the rapid clearance of BsAb-1 relative to BsAb-2 (Figure 4).
These negative IHC data were also consistent with radiolabel
biodistribution studies in cynomolgus monkeys, which showed that BsAb-1
and BsAb-2 had a similar rate and extent of distribution to tissues
(Figure 2). The similar rate and extent of accumulation in the tissues
observed for BsAb-1 and BsAb-2 indicate comparable tissue disposition
for the molecules and by extension, that the PK differences were not due
to clearance via a specific organ. The key difference noted in the
biodistribution studies of BsAb-1 and BsAb-2 was the rate of clearance
of the molecules from the major organs of clearance. The increased
clearance of BsAb-1 relative to BsAb-2 from all the major tissues of
elimination along with the greater amount of catabolite recovery of
BsAb-1 in urine, strongly suggests that following uptake into tissues
BsAb-1 was not efficiently recycled back to the blood circulation and
instead degraded. Given these data, we speculated other plausible
mechanisms, such as physiochemical properties leading to differential
molecular stability, non-specific binding profiles and/or FcRn
interactions, may be causative and account for the rapid tissue
elimination of BsAb-1.
Physiochemical properties including positive charge, poor thermal
stability and hydrophobic-based interaction potential that can
facilitate non-specific interactions have been linked to the
pharmacokinetic developability of mAbs.7-9 While there
are a paucity of such studies for BsAbs, in previous reports we found
that increased global structural stabilization and reduced
hydrophobicity were connnected with in vivo kinetics for BsAbs
with an ECD format.10 In the case of BsAb-1 and BsAb-2
in this report, the molecules have largely comparable physiochemical
profiles, with BsAb-1 exhibiting lower first melting temperature than
BsAb-2 (Table 1). Taken together, the physiochemical differences are
relatively small and we speculate that these factors alone are not the
major contributors to the differential clearance of BsAb-1 and BsAb-2.
Another plausible mechanism that we and others have suggested which can
negatively affect the pharmacokinetics of mAbs (and by extension BsAbs)
is the FcRn interaction profile.16, 24, 25 The FcRn
interaction profile includes the direct binding interactions of
molecules to FcRn at both acidic (pH ~6) and neutral pH
(pH ~7.4), as well as, characterization of the rate of
the dissociation of the IgG:FcRn complex as the pH increases. The later
parameter contextualizes interactions with FcRn which would impede mAb
recycling and release within the endosomal compartment and into the
peripheral circulation. 20, 24, 26 In vitroanalyses of BsAb-1 and BsAb-2 using previously published approaches
showed that BsAb-1 and BsAb-2 bound to FcRn similarly at pH 6 and showed
no binding to FcRn at pH 7.4 (Table 2). These data indicated that direct
FcRn binding was unlikely related to the differential clearance
observations between BsAb-1 and BsAb-2. However, additional
characterization of the FcRn release profile for each BsAb did show
striking differences. In the FcRn release assessment, complexes of each
of the BsAb with FcRn at acidic pH (pH ~6) were formed
and the amount of BsAb which remained bound to FcRn once the complex was
exposed to neutral pH (pH ~7.4) was measured as a
surrogate of FcRn intracellular binding and extracellular release
activities. There is ~8 times larger amount of BsAb-1
that remained bound to FcRn once the complex was exposed to neutral pH
compared to BsAb-2, indicating that BsAb-1 is less efficiently released
from FcRn upon the pH change (Table 2). Taken together, the poor FcRn
release profile may also contribute to the more rapid in vivoclearance of BsAb-1 and provide an additional and perhaps a major
culprit mechanism for the PK difference observed relative to BsAb-2.
Studies with mAbs have shown that altered release from FcRn can affect
both the distribution phase (α phase) and the elimination phase (β
phase) of the kinetic time course.27 The PK profile of
BsAb-1 showed a rapid distribution and short half-life consistent with
the hallmarks of altered FcRn release at neutral pH, further supporting
that poor release from FcRn is a likely the perpetrator mechanism for
the poor PK behavior of BsAb-1. Interestingly, analyses of the FcRn
release profile of the parental mAb (mAb-1) used to construct BsAb-1 did
not show any evidence of poor dissociation from the receptor at neutral
pH (refer to Results section). This suggests that the fusion of scFv-2
to mAb-1 altered the FcRn interactions of BsAb-1, which was not the case
when scFv-1 was fused to mAb-2 to construct BsAb-2.
While dysfunctional FcRn interactions have been noted to negatively
impact mAb PK, previous studies of other BsAbs showed no connectivity to
FcRn as causative in rapid clearance observations.10,
13, 15, 28-30 To the best of our knowledge, the data presented herein
is the first report connecting altered FcRn release to BsAb PK. We
speculate that there are likely differences in the intracellular
trafficking of BsAb-1 and BsAb-2 linked mechanistically with
FcRn-mediated recycling that connect to the in vivo catabolism,
elimination and PK observations for the two molecules (Figure 5). We
postulate that at the cellular level the molecules are largely
comparably internalized via fluid phase endocytosis into endosomes which
facilitates binding FcRn within the acidic environment of this
compartment. The relative similarity in the BsAb-1 and BsAb-2
physiochemical and direct FcRn binding properties (at acidic and neutral
pH) are consistent with this proposed non-specific intracellular
internalization and endosomal FcRn binding mechanisms. Next, in the case
of BsAb-2, the in vitro and in vivo data indicate the
molecule likely undergoes ‘productive recycling’ which is connected to
antibody-based biologics with acceptable PK developability. In this
situation, intracellular BsAb-2 is mostly salvaged from lysosomal
degradation by FcRn-mediated recycling. Indeed, the efficient release
from FcRn at neutral pH in vitro supports the molecule’s PK
profile. Additionally, the slowed blood clearance of BsAb-2 supports the
molecule being productively recycled back into the blood circulation
once the receptor:BsAb complex is exposed to the neutral pH outside
cells. In contrast, for BsAb-1 which displayed poor PK, we hypothesize
that the molecule undergoes ‘non-productive recycling’ whereby when the
FcRn:BsAb complex is exposed to neutral pH there is inefficient release
of BsAb-1 from FcRn. The inefficient release may shift the trafficking
equilibrium such that BsAb-1 is not released from cells and eventually
degraded. The ~54% of BsAb-1 that remained bound to
FcRn at neutral pH in vitro is consistent with this speculation
(Table 2). Along those lines, the greater rate and extent of BsAb-1
catabolites found in urine (relative to BsAb-2) is also supportive of
the proposed mechanism (Figure 3). Additional interrogation of our BsAb
constructs using other approaches including cell-based trafficking and
imaging studies may provide further insight in future studies.
In summary, the findings in this report are an important demonstration
that BsAb PK can be impacted by a variety of physiological and
biochemical factors. There are multiple PK developability considerations
including the nature of the BsAb targets (target/turnover/tissue
distribution), the physiochemical properties of the BsAbs, and the BsAb
structural configuration that can influence disposition and elimination
differentially. Careful delineation of preponderance of these factors on
a molecule-to-molecule basis can ultimately lead to the selection and
design of BsAbs with increased therapeutic value for patients.