Discussion
The most important facet of JV-GL1 pharmacology is arguably its
ultra-long duration of action (Woodward et al ., 2019a). The
present studies have demonstrated that this prolonged effect on
intraocular pressure is entirely mediated by a single pharmacological
entity, the EP2 receptor. Class selective long-acting
β2-agonists (LABAs) and muscarinic antagonists (LAMAs)
have previously been discovered (Wold et al., 2019; Wendell et al.,
2020) but their clinical dosing regimen is once daily, even when used in
combination (Maqsood et al., 2019). For JV-GL1, a once weekly dosing
regimen is feasible. Further research on JV-GL1 and the
EP2 receptor may lead to a wide range of small molecules
dosed once-weekly that offer improved disease control and lower drug
costs. At this juncture, the existing data suggests JV-GL1 may be an
important new drug for women’s health and inflammation (Coleman et
al ., 2019). For glaucoma, the data indicate that JV-GL1 is a realistic
proposition for a breakthrough drug in treating glaucoma (Colemanet al ., 2018; Woodward et al., 2019a).
The long-acting effects of JV-GL1 on IOP in monkeys extend far beyond
time points when detectable levels of the drug can be measured in
anterior segment tissues of the eye (Woodward, et al. , 2019a).
JV-GL1 was designed as an EP2 receptor agonist and the
long acting effects of a week or more were quite unexpected, given that
the ocular hypotensive duration of action of the numerous
EP2 agonist family is essentially one day (Woodward,et al. , 2019b). This raised the possibility that the effects of
JV-GL1 on IOP occurred as a result of two separate pharmacological
activities; an initial EP2 receptor mediated effect
followed by a secondary long-acting phase that was independent of
EP2 receptors. In order to address this question, we
studied the effects of JV-GL1 in gene deleted mice (Saeki et al. ,
2009). Our studies demonstrate that the entire duration of ocular
hypotensive action of JV-GL1 is EP2 receptor mediated.
A single topical application of 0.01% JV-GL1 significantly reduced IOP
in both ocular normotensive and steroid-induced ocular hypertensive mice
after 3 hours. However, only steroid treated mice exhibited a long-term
pressure reduction, persisting for 4 to 6 days. JV-GL1 was incapable of
reducing IOP in mice lacking the EP2 receptor under both
ocular normotensive and steroid ocular hypertensive conditions.
Conventional interpretation of G-protein coupled receptor signalling
would suggest that upon ligand binding to the EP2receptor Gαs coupling activates adenylate cyclase which
converts adenosine triphosphate to the second messenger, cyclic
adenosine monophosphate (cAMP) (Regan et al. , 1994; Pierceet al. , 1995). The cAMP-dependent pathway promotes matrix
metalloproteinase (MMP’s) secretion (Shim, Kim and Ju, 2017). MMP’s
break down extracellular matrix material in the outflow tissues of the
eye, reducing resistance and lowering IOP (Nilsson et al. , 2006).
In this study, topical JV-GL1 did not significantly affect outflow
facility in the steroid ocular hypertensive mouse eye, nor in ocular
normotensive mouse eyes perfused with de-esterified JV-GL1 over an acute
time-scale. A study using cellular dielectric spectroscopy found JV-GL1
to have high activity in human ciliary muscle cells but little activity
in human trabecular meshwork cells (3.9 nM vs >10 µM,
EC50 values respectively) (Woodward, et al. ,
2019a). This suggests a uveoscleral mechanism of action for JV-GL1,
which correlates with the minimal effects on outflow facility observed
in this study and with the mechanism of action proposed for the standard
and most studied EP2 agonist, butaprost (Nilssonet al. , 2006). Our perfusion results largely corroborate those
obtained from perfusion studies in monkeys, where JV-GL1 did not
significantly affect outflow facility. However, in monkeys JV-GL1
significantly increases uveoscleral outflow, as measured by
fluorophotometry, and significantly reduces inflow by 20% over an acute
timescale (Woodward, et al. , 2019a). A limitation of measuring
outflow facility in enucleated mouse eyes is that inflow is essentially
terminated once the eye is removed and the pressure independent nature
of the uveoscleral pathway makes it challenging to measure using a
pressure controlled perfusion system. Thus, any effect JV-GL1 may have
on inflow or uveoscleral outflow cannot be measured. Unfortunately in
monkey studies, inflow measurements were not continued over the
time-course of IOP measurements, and the relationship between a
reduction in inflow and IOP suppression remains unanswered.
In monkeys, JV-GL1 exhibited long-term IOP effects in both ocular
normotensive and laser-induced ocular hypertensive animals, and
demonstrated considerably greater pressure reduction (Woodward et
al. , 2019a) compared to the observations made in this study with mice.
These inconsistencies may simply be due to species differences. The
relative size of the ciliary muscle is much smaller in mice than in
primates (Ko and Tan, 2013) as mice do not accommodate for vision (Tamm
and Lutjen-Drecoll, 1996). If JV-GL1 is thought to work primarily via
the uveoscleral route, then the more developed ciliary muscle of the
monkey, which is in frequent use and therefore containing greater
amounts of extracellular matrix material, would be expected to respond
better to remodelling effects of MMP’s than the more vestigial ciliary
muscle of the mouse.
Alternatively, it may be the different methods employed to induce ocular
hypertension. Unlike laser-induced ocular hypertension, the steroid
induced ocular hypertensive model relies on the persistent presence of
dexamethasone in the eye, eluting from periocular depots of
dexamethasone loaded nanoparticles. Steroid ocular hypertension is
brought about by dexamethasone induced changes in extracellular matrix
deposition and stiffness in the trabecular meshwork (Overby et
al. , 2014; Raghunathan et al. , 2015). Increased extracellular
matrix deposition is likely due to dexamethasone inhibiting MMP’s,
either through inhibition of transcription factors (Forster et
al. , 2007) or by inducing tissue inhibitors of matrix metalloproteases
(MMP’s) (Xu et al. , 2001). Dexamethasone may have similar effects
on the uveoscleral pathway. MMP’s 2, 3 & 9 were found to be
significantly reduced in ciliary body explants treated with
dexamethasone, with MMP-3 eliminated after 72 hours of dexamethasone
treatment (el-Shabrawi et al. , 2000). MMP-3 can break down a
diverse range of extracellular matrix components (Matrisian, 1990). What
is more, dexamethasone has been shown to inhibit the production of
endogenous prostaglandins in human trabecular meshwork cells and scleral
fibroblasts (Gerritsen et al. , 1986).
Prostaglandins have been shown to induce MMP production in the ciliary
body (el-Shabrawi et al. , 2000). These MMP’s are thought to be
released into the aqueous and diffuse into the ciliary muscle where they
break down extracellular matrix components between the longitudinal
fibres of the ciliary muscle, allowing less restricted flow into the
choroid (Ocklind, 1998). Consequently, we can speculate that JV-GL1 in
the steroid ocular hypertensive mouse model would have to compete
against the pro-fibrotic effects of constant dexamethasone delivery to
the eye, limiting its ability to lower IOP to the same degree observed
in the laser-induced ocular hypertensive monkey model, where no
dexamethasone is present.
It is of interest that JV-GL1 has an effect on inflow in monkeys, as
this has not been demonstrated with any other EP2agonists to date (Woodward, et al. , 2019b). The most studied
EP2 agonist, butaprost, is efficacious at lowering IOP
but has a short duration of action, with no effect on inflow and no
effect on conventional outflow facility. The entirety of butaprost
effects appear to be mediated by changes in uveoscleral outflow,
specifically ciliary muscle resistance (Nilsson et al. , 2006).
Therefore, JV-GL1 behaves somewhat differently when compared to other
EP2 agonists. Aqueous flow can be reduced by direct
activators of adenylate cyclase such as forskolin (Caprioli et
al. , 1984) and cholera toxin (Gregory et al. , 1981), via a
cAMP-dependent process. In the latter case, reduction in IOP persisted
for 6 days after a single intravitreal injection.
It is, of course, highly unlikely that JV-GL1 is directly and/or
irreversibly activating adenylate cyclase. However, recent research has
disrupted the conventional view of G-protein coupled receptor
signalling, whereby active receptor conformations engage and activate
only one of four classes of cytoplasmic heterotrimeric G-proteins
(Gαi/o, Gαs, Gαq/11 and
Gα12) and a more promiscuous model, whereby receptors
can bind to more than one G-protein subtype, eliciting multiple
G-protein dependent signalling events has emerged (Stallaert, et al.,
2011; Hermans, 2003). Recent work investigating EP2receptor activation with butaprost, demonstrated not only the
conventional Gαs – cAMP pathway activation but also
Gαq/11 – calcium pathway activation upon ligand binding
(Kandola et al. , 2014). Therefore, like forskolin and cholera
toxin, which directly activate adenylate cyclase and thereby stimulate
only the cAMP pathway, JV-GL1 may induce signal bias (Wold et al., 2019)
and signal primarily via the Gαs-cAMP pathway by virtue
of a novel interaction with the EP2 receptor. A study
using cellular dielectric spectroscopy demonstrated that the
EP2 antagonist AH6809 could not completely block JV-GL1
at the EP2 receptor, in human ciliary muscle cells
(Woodward, et al. , 2019a). This finding suggests that JV-GL1 may
be binding to a possible allosteric site on the EP2receptor where endogenous PGE2 does not bind, or bind to
the EP2 receptor in a different way. Allosteric binding
and a potential Gαs-cAMP pathway bias may be responsible
for JV-GL1 long duration effects. This would set it apart from unbiased
EP2 agonists such as butaprost and merit further
research into its mechanism of action.
The long duration of action of JV-GL1 may result from allosteric
binding, thereby implying that the binding of other anti-glaucoma
EP2 agonists (Woodward et al., 2019b) to their target
receptor is orthosteric. The long duration of action on IOP and the
inherent activity on human ciliary smooth muscle cells could be
indicative of JV-GL1 as a positive allosteric modulator. One, and only
one, possible published explanation (Jiang et al., 2010) directly
relevant to the ultra-long action of JV-GL1 on IOP involves allosteric
modulation of the EP2 receptor. More pertinent to the
extended ocular hypotensive activity is that allosteric potentiation of
EP2 receptor activity is favoured by an ester moiety
compared to a carboxylate22. The unexpected activity
of JV-GL1 on ciliary smooth muscle cells correlates with manifestation
of activity in intact cells rather than cell-free assays for
EP2 allosteric potentiators22. JV-GL1
would be an ideal compound for elucidating the crystal structure of the
EP2 and thereby gain valuable insights into orthosteric
and allosteric binding sites and drug design.
In summary, the long-acting effects of JV-GL1 on IOP are
EP2 receptor mediated and this provides an intriguing
avenue of future investigation with respect to mechanisms that impart
such extended biological activities. Beyond the significant potential
benefits of JV-GL1 for improved glaucoma treatment, there are also
implications for pre-term labour, asthma, and the anti-inflammatory
effects of EP2 agonists. Future research may reveal
similar advantages for compounds designed to be ultra-long acting at
other GPCRs.