Biology Reference
In-Depth Information
more than 50% for the dimer as closing UCR2 across one active site also reduced
the
k
cat
of the other active site (Burgin et al.
2010
). We thought such partial
inhibitors might provide a “safety valve” preventing complete enzyme inhibition
and thereby improve tolerability.
Compounds able to bind UCR2 when located in the active site share a common
pharmacophore and binding pose. The pharmacophore consists of four elements; a
planar scaffold providing a hydrogen bond to Gln535, a linker, and two aromatic
substituents which create a clamp that holds UCR2 in the closed conformation. Key
to the design of partial inhibitors was the deliberate weakening of the P-clamp
interaction with the hydrophobic walls of the active site. For example, we obtained
potent compounds based on pyridine and quinoline scaffolds, but compounds based
on bicyclic ring systems fully inhibited PDE4 activity. In contrast, compounds
designed with catechol and biaryl scaffolds having less aromaticity afforded com-
pounds with both full and partial inhibition behavior. With respect to the aromatic
arms, subtle changes to Ar1 substituents affected full or partial inhibition behavior.
Flexibility in Ar1 may allow full inhibitors to access additional interactions with
active site residues. Ar2 substituents affected potency but not full or partial kinetic
behavior. Full and partial inhibitors close UCR2 across the active site identically, so
differences in the positioning of UCR2 do not explain the differences in kinetic
behavior (Burgin et al.
2010
).
We found that examples of PDE4 allosteric modulators were potent in cellular
assays. Studies of eosinophil production of LTE4 in human whole blood demon-
strated that PDE4 allosteric modulators could provide complete inhibition of a
biological response, even though they have partial enzyme inhibition kinetics in
biochemical assays (Burgin et al.
2010
). For reference, simple competitive PDE4
inhibitors, such as roflumilast, exhibit similar potency in biochemical assays, by
assay of inhibition of cAMP hydrolysis in forskolin-stimulated HEK293 cells, and
in an eosinophil assay of LTE4 production (IC
50
values were 5.8 nM, 4 nM and
5.6 nM, respectively, across the three assays). PDE4D is the dominant PDE4
subtype in human embryonic kidney HEK293 cells with respect to cAMP hydroly-
sis (Lynch et al.
2005
). In contrast to roflumilast, PDE4 allosteric modulators
showed similar potency with respect to inhibition of PDE4D in biochemical assays
and in the eosinophil assay, but were 10- to15-fold less potent with respect to
inhibition of cAMP hydrolysis in HEK293 cells. For example, the clinical lead
compound D159687 had IC
50
of 28 nM and 44 nM, respectively, for biochemical
inhibition of PDE4D and inhibition of eosinophil LTE4 production, but an IC
50
of
253 nM with respect to inhibition of cAMP hydrolysis in HEK293 cells (Burgin
et al.
2010
). Rolipram similarly is potent in biochemical and eosinophil assays but
tenfold less potent versus inhibition of cAMP hydrolysis. Why should this be so?
Differences between membrane-bound and cytosolic PDE4 with respect to
inhibition by rolipram have been noted previously (Huston et al.
1996
; Souness
and Rao
1997
). Huston et al. report that rolipram inhibits membrane-bound PDE4
with high affinity, suggesting that the compound is inhibiting PDE4 through closing
UCR2. Rolipram was found to be much less potent against cytosolic PDE4,
which behaved kinetically as if rolipram only competed with cAMP for binding
