(Investigative Ophthalmology and Visual Science. 2001;42:1126-1133.)
© 2001
by The Association for Research in Vision and Ophthalmology, Inc.
A New Approach to the Medical Management of Glaucoma, from the Bench to the Clinic, and Beyond: The Proctor Lecture
Laszlo Z. Bito
 |
Introduction
|
|---|
Although prostaglandins (PGs) were said to have detrimental
ocular effects, they and their derivatives were found later to
represent a new approach to the medical management of glaucoma, being
local hormones rather than the heretofore used agonists of the
autonomic system or inhibitors that are totally foreign to the body. An
esterified PG prodrug, PGF2
-isopropyl ester
(PGF2-IE), was found to be a particularly
effective and potent ocular hypotensive agent, but PGs, in general,
offer great therapeutic advantage in terms of their local and systemic
pharmacokinetics and mechanism of effect. By increasing uveoscleral
outflow, the hypotensive effect of PGs is not limited by episcleral
venous pressure that may be particularly elevated in sleep.
Furthermore, this mechanism, in contrast to the aqueous humor
production-reducing effect of ß-blockers, provides around-the-clock
pressure reduction. This was actually shown to be the case with
latanoprost, the active ingredient of Xalatan (Pharmacia,
Uppsala, Sweden), the first PG-based glaucoma drug released in the
United States and Europe.
It is argued that, in the future, not only the extent but also the
mechanism of pressure reduction should be considered in the selection
of therapy for a given patient. The major goal of this article is,
however, to provide an account of the many hurdles and frustrating
obstacles that had to be overcome in the development of Xalatan, from
the original misconceptions, through the development of new concepts,
to its therapeutic realization.
During much of the 20th century, glaucoma therapy was dominated by
drugs relating to the mediators or inhibitors of the sympathetic and
parasympathetic systems. In the absence of prior evidence of an
autonomic role in the control of intraocular pressure (IOP), this was
apparently because these mediators were discovered in the 19th century
and their study dominated pharmacology during the first decades of the
20th century. But then, a new class of autacoids was discovered in the
1930s, named prostaglandins (PGs) by von Euler1
in 1934,
and found in the 1960s and 1970s to be produced by, and to affect
virtually all, tissues of the body.2
Consequently, much effort was devoted to the synthesis of large numbers
of PG derivatives aimed at the development of PG-based therapeutic
agents for cardiovascular, gastrointestinal, and other major areas of
pharmacology, but these efforts eventually faded out. In my judgment,
both the initiation and cessation of these attempts were premature,
because both preceded the understanding of the pharmacodynamics of PGs.
The plethora of studies on the ocular effects of PGs, including the
numerous clinical studies with PGF2-IE, which
became the first PG derivative used to demonstrate the feasibility of
this new approach to glaucoma management as well as the mechanism of
PG-mediated lowering of the IOP, have been repeatedly
reviewed.3
4
5
6
7
An account of the development of
latanoprost, with its superior selectivity for the F type of PG
receptor, making it therefore essentially free of the conjunctival
hyperemic and corneal sensory side effects of naturally occurring PGs,
is described in this issue of IOVS by Johan Stjernschantz in
his Proctor Medal award lecture.8
It is for this reason that I write here primarily about my own
involvement and the work of my own laboratory, although I do not mean
to imply that all the relevant ideas were generated by my group.
|
The First Hurdle: The Concept that PGs Increase IOP and Have Other
Deleterious Ocular Effects
|
|---|
My exposure to, and interest in, PGs dates back to 1964, when I
started a postdoctoral fellowship with Hugh Davson. I learned then that
he had given up research on aqueous humor dynamics, because the
required cannulation of the eye invariably led to some breakdown of the
blood–aqueous barrier (BAB), preventing evaluation of normal
physiologic or pharmacologic influences. A search for the mediator of
this annoying effect led Ambache et al.,9
10
11
in the
1950s, to discover, in crude iridial extracts, lipidic mediators that
reproduced some of the signs of ocular irritation and that were
subsequently shown to include PGs. This discovery, however, did not
help to protect the BAB, because there was no known inhibitor of the PG
system. Thus, we focused in the 1960s on experiments that did not
require cannulation of the eye, including the elucidation of transport
mechanisms that control the unique and remarkably stable
microenvironment of the retina and the brain.12
13
14
Interestingly, these studies led me back to PGs a few years later,
after I established my own laboratory at Columbia University.
Meanwhile, in the late 1960s, Upjohn (Kalamazoo, MI) had begun
to provide free PG samples, and several researchers began to administer
PGs to the eyes of experimental animals (primarily rabbits), mostly
through cannulas inserted into the anterior chamber also for measuring
IOP, in amounts sufficient to produce signs of ocular irritation and
inflammation. These included BAB breakdown and large
increases in IOP, which these researchers must have assumed
to be the primary ocular effects of PGs, based on the findings of
Ambache and Brummer.10
Being aware of the unphysiological
state of the cannulated eye, I viewed these results with much
skepticism, which I did not hesitate to air at scientific meetings.
Yet, the resultant concept that PGs are the mediators of
ocular inflammation became more and more generally accepted during the
1970s.15
This preoccupation with the role of PGs in ocular
inflammation did, however, lead to the demonstration that PGs are
synthesized by ocular tissues, most notably by the anterior uvea, but
are not effectively metabolized within the eye.16
17
This
led me to the hypothesis that there must be an active transport
mechanism for the removal of PGs from intraocular fluids.
|
The PG Transport System and Its Implications
|
|---|
My already-mentioned interest in the homeostatic transport
processes of the blood–ocular and blood–brain barriers extended over
the years to several classes of substances of known biological
importance.14
Only the so-called organic acid transport
system presented an enigma, because its known substrates, such as
iodipamide, were not produced within the body and had no known
physiological roles. It was one of the great challenges of this field
to find the natural substrate of this transport system. However, our
first reports18
19
demonstrating that PGs are transported
by this system were received with great skepticism, because PGs, being
fatty acid derivatives, were assumed to freely cross cellular
membranes. Starting in 1970, we conducted in my laboratory—with the
able cooperation of Erica Salvador, Roger Baroody, and Martin
Wallenstein, among others—a series of experiments demonstrating that
not only the normal functioning of the eye and the brain, but also the
pulmonary metabolism and renal excretion of PGs and PG derivatives
depend on this active PG transport system.18
19
20
21
Our
studies of these transport processes gave us a unique insight into the
ocular and systemic pharmacokinetics of PGs and also led me to the
conclusion that PGs, functioning in the eye like local hormones, are
uniquely well suited to achieve localized effects, after topical
application, without systemic side effects.5
This became a
particularly important consideration after the recognition of the
systemic and central side effects of the eye drop formulation of a
cardiac drug, timolol, because it called attention to the fact that
topical ocular instillation may result in a more effective systemic
delivery of a drug than oral administration, by avoiding the
"first-pass" effect of hepatic metabolism.22
Our studies of the kinetics of the ocular PG transport system also
indicated that PG concentrations in the intraocular fluids
(IOFs) are normally maintained at very low levels—orders of
magnitude lower than those resulting from the doses of PGs that were
introduced in the early 1970s into cannulated eyes, yielding very large
IOP increases and other ill effects. The resultant concept
that PGs are the mediators of the ocular irritative response
and inflammation was also supported at that time by several other
observations, most notably in the studies of Kenneth Eakins et
al.16
showing very large increases in
PGE2-like activity in the aqueous humor in
uveitis.
|
The First Breakthrough: A Tonometer Probe that Allowed Us to Study
the Whole Time Course of IOP Changes in Uveitis
|
|---|
By that time we had already shown in our PG transport studies that
experimental uveitis eliminates the ability of the anterior uvea to
accumulate PGs.23
This suggested that, during uveitis, the
PG accumulation in the aqueous humor may be due not only to increased
PG synthesis, but also to impaired PG removal across the BAB. This
realization gave me the impetus to undertake a long-term study to
define the development and resolution of the various signs of uveitis,
including the whole time course of its effect on IOP over several
weeks. This was made possible by a breakthrough in 1972, when, with the
help of Maurice Langham, we were able to obtain a floating-tip IOP
sensor that allowed us to repeatedly measure the IOP of unanesthetized
rabbits, without any sign of BAB breakdown.
Intravitreal injection of bovine serum albumin caused a biphasic
inflammatory response, with respect to iritis, BAB breakdown, and
cellular invasion of the anterior chamber, and a profound
decrease in IOP.24
The first phase peaked
within the first 2 days, followed by a temporary recovery to normal
values. The lowest IOP was observed on day 2, and it was preceded by,
or paralleled with, a loss of anterior uveal PG transport capacity.
This capacity did not show substantial recovery, before a second phase
of inflammation, associated with a second cellular invasion of the
anterior chamber (peaking at around 12 days), became evident. Only the
first phase of this uveitic response was observed when bacterial
endotoxin was injected. This led us to the finding that all commercial
protein preparations we tested contained such
endotoxins.25
|
The Ocular Effects of Exogenous PGs on Eyes Not Traumatized by
Anterior Chamber Cannulation
|
|---|
We found no ill effects when, in our studies on the ocular PG
transport system,13
26
we injected intravitreally amounts
of PGs in excess of those that had been shown previously to cause BAB
breakdown when introduced into the anterior chamber. These and many
other observations eventually confirmed my conviction that the IOP
increases observed in earlier experiments were due to the combined
effects of excessively high PG doses and of endogenous mediators
released by the trauma of anterior chamber cannulation. However, the
implications of these conclusions were generally ignored throughout the
1970s.
It was a very refreshing experience, therefore, when in the fall of
1975, while I was instructing in the histology laboratory, a first-year
medical student introduced himself to me and demonstrated not only
great familiarity with the PG field, including my own prior
publications, but also expressed skepticism regarding the prevailing
views on the ocular effects of PGs. This is how I met Carl Camras, who
had worked in eye research at Yale as a college student. From some
experiments he had done there, Carl had independently concluded that
the role of PGs in ocular inflammation is primarily a hypotensive one.
Carl demonstrated a keen ability to analyze and critically evaluate
relevant publications, as well as to defend his viewpoints vigorously
and effectively—virtues he never compromised over the years. Carl
joined my laboratory in the summer of 1976, and, using the
aforementioned tonometer probe, produced a seminal
paper,27
demonstrating that low doses of PGs can reduce
IOP in uncannulated rabbit eyes.
Unfortunately, further studies in rabbits revealed that the
demonstrated IOP reduction, lasting for a few hours with a single
topically applied PG dose, could not be maintained with repeated
twice-daily topical applications, because of rapid development of
tachyphylaxis, which could not be overcome by increasing the PG dose.
In fact, doubling the PG dose from 5 to 10 µg on the fifth day of
treatment increased, rather than decreased, the IOP.
Tachyphylaxis to the IOP effects of PGs had also been reported earlier,
in short-term studies on cannulated eyes,28
29
but there
was no guidance in the literature as to the possibility of overcoming
it.
We postponed the publication of this most disappointing finding for
several years,30
hoping to first establish its mechanism.
Assuming that it had to do with BAB breakdown, we redoubled our ongoing
efforts to elucidate the true role of PGs in the mediation of the
ocular irritative and inflammatory response and the role of our PG
transport system in its prevention. Carl had also worked on one of
these projects showing that the large initial IOP increase after
nitrogen mustard (NM) application could be blocked by prior alcohol
denervation or by capsaicin-induced neuropeptide depletion, but not by
indomethacin.31
32
This indicated that the initial IOP
response to ocular irritation is mediated by neuropeptides rather than
PGs, whereas a second phase of IOP increase, occurring at 3 to 7 hours
after NM application, was not similarly blocked but could be reduced by
indomethacin. These experiments clearly demonstrated that PGs are just
one of the mediators of the complex phenomenon of the ocular irritative
response, even in rabbits.
Because one of my coworkers, Susan Merritt, had already trained six owl
monkeys to accept handling and tonometry for our IOP studies on the
autoregulation of cholinergic mechanisms, Carl also tested the effects
of PGs on their eyes. The IOP reduction obtained was very impressive,
yet the findings were rather disappointing: even a 40-fold higher dose
of PGF2 than was observed to effectively
reduce IOP in rabbits was found ineffective in this species.
Significant IOP reduction was obtained with 1000 µg per eye, but this
dose also produced some initial pressure increase and miosis, and in
some animals even a few cells were detected in the aqueous humor of the
treated eyes. There was some contralateral IOP effect that appeared to
be increased when alternate eyes were treated with the same single dose
of PG at several-week intervals. These and other findings suggested
that, in these very small-bodied animals (only 0.8–1.0 kg body
weight), topically applied drugs can have a systemic effect, thus
complicating the interpretation of results.33
These
findings also indicated that there are astonishing species differences
with respect to both the nature of ocular response to PGs and the doses
required to obtain an effect. The differences in these two divergent
species made it impossible to predict the possible responses of the
human eye to similar treatment. The eyes of these nocturnal monkeys, in
particular, could not be assumed to be an adequate model for the
diurnal type of eyes of humans.
By the beginning of the 1980s, our studies, combined with available
information on morphologic and behavioral differences in mammals,
convinced me that the data obtained from the rabbit, which has a
monitoring type of visual system, are irrelevant with respect to the
searching-type human eye, which, as we later concluded, has a very
different protective mechanism, in which PGs also play an important
role, but a role very different from their observed effects on rabbit
eyes.
|
The Unique Sensitivity of the Rabbit Eye to BAB Breakdown and
Its Apparent Evolutionary Significance
|
|---|
Using many different experimental models, ranging from
paracentesis to x-ray or NM-induced ocular irritation and intravitreal
endotoxin or antigen-induced uveitis,31
34
35
I became
more and more convinced that the great vulnerability of the rabbit eye
to BAB breakdown is not a freakish evolutionary oversight. Rather, it
represents a uniquely well-developed prophylactic mechanism to allow
the entry of the blood-borne components of the clotting mechanism,
required for the sealing of corneal wounds, into the anterior chamber.
The importance of this mechanism becomes evident if we consider that
the avascular cornea overlying the essentially protein-free aqueous
humor is the only part of the body surface that has no ready access to
such a clotting mechanism. Because the lateral placed, protruding,
unprotected eyes of a rabbit, which allows the virtually
360o field of vision required to monitor its
environment for predators, are very vulnerable to injury, there must
have been a particularly high selective pressure in this species for a
mechanism to allow plasma proteins to enter the anterior chamber after
penetrating ocular injury. This, in fact, is a truly prophylactic
mechanism in the rabbit, because its aqueous humor can become plasmoid
even before corneal penetration occurs, when the corneal surface or the
surrounding skin is irritated.
In contrast, for species with great dependence on high visual acuity
(such as primates, particularly arboreal primates), such prophylactic
breakdown of the BAB would be detrimental because of the
interference—caused by the resultant light-scattering of the increased
protein content in the anterior chamber—with the very high visual
acuity required by such species. However, the frontally oriented,
deeply seated eyes in these species are much better protected, thus
minimizing corneal injuries, although greatly limiting the visual
field. Although the BAB breakdown mechanism in rabbits is mediated by
axon reflexes releasing substance P and other neuropeptides, it became
clear that PGs play at least a modulatory role in BAB breakdown in this
species.32
35
We argued, therefore, that tachyphylaxis to
PGs may also be unique to the rabbit eye, rendering it particularly
unsuitable for predicting the effects of PGs on the human eye.
|
Selecting More Appropriate Animal Models
|
|---|
For these reasons, we decided to use rhesus monkeys (Macaca
mulatta) instead, because this diurnal species has much more
humanlike eyes, and macaques have been shown, particularly by the
elegant studies of Ernst Bárány, Anders Bill, and
Paul Kaufman, to exhibit many physiological and pharmacological
responses very similar to those exhibited by the human
eye.36
However, to develop optimal glaucoma drugs, we expected to assay many
dozens of PGs in different concentrations and formulations. Because of
their high cost of maintenance and difficulty of handling, we judged
monkeys alone insufficient for our purposes. Because of their ease of
handling and relatively well protected, searching-type eyes, we chose
cats as a second model. And, indeed, we found cats to be much less
sensitive to BAB breakdown than rabbits.
|
Maintained IOP Reduction in Cats and Rhesus Monkeys, with Daily
Topical PG Application
|
|---|
Because any drug advocated for the treatment of chronic simple
glaucoma must maintain effective IOP reduction indefinitely, we
initiated a long-term study in cats that continued for 9 months.
Effective IOP reduction was maintained so long as a dose of 100 µg of
PGE2 was applied topically to the eye once or
twice daily.30
These studies, together with shorter term
studies in rhesus monkeys, clearly demonstrated that in contrast to
rabbits, continuous PG treatment in cats maintains IOP reduction
without tachyphylaxis or breakdown of the BAB. Virtually all ocular
parameters that could be assessed noninvasively were evaluated in at
least some of the six cats studied. All these parameters, which
included ERG, the integrity of intraocular muscle function as
determined with dynamic pupillography, maintenance of endothelial
cellularity as assessed by specular microscopy, and complete
biomicroscopic evaluations, were found to remain normal.37
These were the key experiments that finally allowed me to sufficiently
convince myself, and thereby to be able to begin convincing others,
that PGs could be used for long-term reduction of IOP for the medical
management of glaucoma—i.e., that our ideas, dating back to the early
1970s, could indeed be reduced to practice.
Coincidentally, it was at this time, in 1980, when I first heard about
the Bayh-Dole Act, which encouraged, and in essence required, the
patent protection of findings arising from National Institutes of
Health (NIH)–sponsored research that have therapeutic potential, as
well as their licensing for commercial development. Accordingly, with
the help of university attorneys, I applied for a patent for the use of
PGs to reduce IOP in glaucoma.38
|
The Side Effects of PGs on the Ocular Surface and the Need to
Reduce the Topically Applied Dose
|
|---|
The anticipation of using PGs to reduce IOP in patients led me to
put the same dose of 100 µg PGE2 that we used
on the cats into my own eye, but I quickly washed it out, because I
experienced a very irritating foreign-body sensation. Actually, I
should have been prepared for this effect, because we had noted that
after topical PG application, cats would hold their treated eyes closed
for varying periods, indicating some irritation of the ocular surface.
Luckily, we soon realized that the duration of eye closure in these
cats was dependent on the dose and type of PG applied. I found that
these animals’ reactions correlated with the extent of the
foreign-body sensation that I experienced, after putting various PGs in
different doses into my own eyes.
Such comparison also made me realize that some of these side effects on
the ocular surface must be due to the very high topical PG
concentrations required to deliver effective amounts to intraocular
target sites, because of the low permeability of the cornea to PGs.
This concept was also contrary to the accepted view on PGs before our
work, which was that these lipidic mediators could readily dissolve in
and pass through membrane lipids. This, of course, implied that they
could be expected to freely penetrate even tight-junctional membranes,
such as the corneal epithelium. Contrary to this a priori view, our
studies demonstrated that the basic cell membrane, thus also the
tight-junctional corneal epithelium, represents an effective
permeability barrier to PGs.39
|
The Search for a Commercial Partner
|
|---|
By this time, officers of the newly established Office of Science
and Technology Development at Columbia University had approached the
relevant pharmaceutical companies in the United States, but none was
inclined to license my patent, presumably because of the still
prevalent prejudices against the ocular use of PGs. It was then that,
through the mediation of my colleague and friend, Endre Balazs of
Healon fame (Pharmacia & Upjohn), we established contact with Pharmacia
in Uppsala, Sweden. Much credit is due to Bengt Agerup of Pharmacia,
who could see the significance of our findings, and who became a great
supporter of this project within the company.
|
The Pharmacokinetic and Pharmacologic Advantages of Local Hormones
Such as PGs
|
|---|
Our studies on the control of cholinergic sensitivity showed, for
example, that from a physiological point of view, pilocarpine is
particularly ill-suited for use as a topical ocular hypotensive agent
because of its antagonistic effects within the eye, decreasing
uveoscleral outflow while facilitating conventional outflow. We could
even demonstrate that under some conditions, it increases rather than
decreases IOP.33
These and other considerations led me to
develop the concept that the use of local hormones has great
therapeutic advantages over the use of neurotransmitters and their
analogues. In contrast to these, local hormones must have evolved to
have a consonant effect in the diffusional domain of their site of
release.4
5
They also offer great advantage when applied
topically, because of the very specific distinguishing characteristic
of local hormones—i.e., their rapid inactivation in, and removal from,
the circulation, virtually eliminating the possibility of systemic side
effects.5
These apparent advantages of this new approach
convinced me to stay with this project and to fight for its success,
overcoming all obstacles.
|
Esterified PGs, as the First Approach for Overcoming Local Side
Effects
|
|---|
In the ensuing collaboration with Pharmacia, we initially hoped to
minimize the known effects of sensory irritation and conjunctival
hyperemia by finding a PG ester (or diester) that was stable enough not
to be rapidly de-esterified on the ocular surface, hence remaining
inactive, yet hydrolyzed by tissue esterases rapidly enough after it
enters the cornea to have a full intraocular hypotensive effect. By
that time, we were collaborating with Bahram Resul, a medicinal chemist
who was the first person hired by Pharmacia for this project, in 1983.
Bahram synthesized many different esters and sent them to us to test in
our cat and monkey IOP models. In addition, by measuring the isotonic
contraction of bovine iris sphincter preparations in vitro, we could
estimate the rate of hydrolysis of various PG esters within this
tissue, from the time elapsed between their addition to the bathing
fluid and the onset (or half-time) of contraction. We could also
estimate the efficacy and potency of the free PG released, by
determining the maximum sphincter muscle contraction achieved. As we
could create beautiful structure–activity relationships that could
also be confirmed in vivo by measuring miosis in cats,40
Bahram and I were very optimistic that we would find a PG derivative
with reduced local side effects, but support for this project within
Pharmacia started to dwindle.
|
The Savior Arrives and Sets His Sights Very High
|
|---|
The arrival of Johan Stjernschantz, who joined the project in
1986, was a godsend. Johan immediately saw the great value of this new
approach to glaucoma management and gave up the higher management
position he was originally hired for by Pharmacia to become the project
leader. Johan set his sights very high: to go beyond esterification
alone and to develop a PGF2 derivative that
would retain the potency and efficacy of
PGF2-esters but would not have the known local
side effects of foreign-body sensation and conjunctival hyperemia,
which would have limited its use to cases of glaucoma that could not be
controlled by milder drugs.
At first, I was disappointed by the obvious delay that ensued as a
result of this high standard, because in my case the foreign body
sensation caused by PGF2-IE and some of the
higher esters was quite tolerable, at least compared with some of the
PG-free acids I had tried. I thought that we should at least make
PGF2-IE available to patients with glaucoma
that was not controlled by other drugs. But finally, I had to accept as
an unfortunate fact of life that a company cannot go to the expense of
long-term toxicology and phase 3 studies unless the drug is acceptable
for a large enough segment of patients, making it commercially viable.
Carl Camras, who had returned to New York by then and was working at
Mt. Sinai with Steve Podos, continued to advocate making
PGF2-IE available for clinical use. When in
the summer of 1988, Carl, Johan, and I went to Umeå, Sweden, to visit
Albert Alm (who by then had conducted very important phase 2 studies
with PGF2-IE in human volunteers), Carl kept
insisting that patients are used to such side effects, but then
admitted that he had never put any PGs into his own eye. That was
easily remedied. We did put a drop of the
PGF2-IE into one of Carl’s eyes, and he
quickly admitted that even though the foreign-body sensation was
tolerable, many patients would not comply with its daily use.
Johan Stjernschantz and Bahram Resul, with their teams, worked many
long days producing new PG derivatives that they tested in Uppsala and
we in New York. They came up with latanoprost, the isopropyl ester form
of a 17-phenyl PGF2 derivative (the active
ingredient of Xalatan) in record time. For this, they deserve great
credit. The series of steps that led to this derivative, as well as its
pharmacology, are described in the presentation of Johan
Stjernschantz,8
the corecipient of the 2000 Proctor Medal.
|
The Next Crisis
|
|---|
I thought my job would end with the completion of the laboratory
research, but biology is not a science of the readily predictable. I
was ready to settle back, primarily to our fascinating studies on
accommodation and presbyopia with Paul Kaufman and
others,41
42
43
44
knowing that the PG project was in good
hands, when I received a crisis call from Johan. In the toxicology
study in France, they noted that the irides of the latanoprost-treated
eyes in some monkeys became noticeably darker. I was stunned. "And in
the rabbits?" I asked after I recovered my wits. "None of the
rabbits," was his response. Species differences had struck again!
This time, however, closer to home, because the unwanted effect
occurred in a primate. On the other hand, I had convinced myself so
thoroughly by then that local hormones can do no wrong that I was able
to regain my composure. My first hope was that the increased
pigmentation might reflect some peculiarity related to the
characteristic coloration of the rhesus monkeys’ eyes.
At that point, however, the initiation of the phase 3 studies was at
stake, and Johan therefore asked me to try to figure out what was going
on. My job was easy: I simply had to work day and night to educate
myself in melanin chemistry and the fascinating world of melanosomes
which, as I expected, had many aficionados in dermatology. It did not
take me long to learn about the fundamental difference between the
melanin system of the skin and the eye—namely, that dermal melanocytes
continually produce melanin and transfer it to keratocytes, whereas
iridial melanocytes are continent and presumably have no mechanism to
lose melanin, which itself has no intracellular
turnover.45
Thus, if PGs stimulate melanogenesis even to a
minimal extent, it would obviously have to result eventually in an
observable darkening of a light-colored iris. Thus, there was no need
to assume that the iridial pigmentation we observed was due to
proliferation of melanocytes.
After many long transatlantic calls and emergency meetings in Uppsala,
we were able to persuade management to continue with the project, thus
gaining some time to learn more about the melanin system and to
initiate some studies.
Luckily, there were some very good electron microscopic studies
available in rhesus monkeys showing that their iridial melanosomes are
morphologically different from those of the human eye and that
sulfur-containing pheomelanin is a major form of pigment in rhesus
monkeys, rather than eumelanin, which is the major form in most humans
(or in pigmented rabbits, for that matter). In short, there was reason
to believe that the increased iridial pigmentation that occurred in
monkeys but not in pigmented rabbits, might not necessarily occur in
humans. Thus, Johan’s decision to perform one of the toxicology
studies in a primate, resulting in the early discovery of this side
effect, turned out to be a godsend. It led to the inclusion of
photographic documentation of the eye color of all patients when they
were enrolled in the phase 3 studies and periodically during the whole
study. This allowed the documentation of the extent, incidence, and
time course of this fascinating effect.
|
The Third (Not Entirely Unexpected) Crisis
|
|---|
In the meantime, our efforts to gain a better understanding of the
melanin system, the peculiarities of the iridial melanocytes, and the
possible effects of PGs thereon, continued full steam. Thus, by the
time the first human case of increased pigmentation was reported in the
United Kingdom, we had reason to assume what later would be
experimentally supported46
—that PGs produced within the
globe may be involved with the normal maintenance of iridial
pigmentation and that this phenomenon may simply represent just one
more normal physiological mechanism mediated or modulated by endogenous
PGs.
The future of the whole project depended on the answer to the question
of what happens to the pigmentation after treatment is stopped, which
the protocol of the phase 3 study was designed to answer. If the
pigmentation continues to increase after treatment is stopped, as soon
as increased pigmentation is noted in a given patient, the project must
be terminated, at least temporarily, because this would imply that the
exogenous PGs induce a self-propagating, and thus possibly malignant,
process. If the effect is rapidly reversed, it would imply that PGs
stimulate very excessive melanin production that is partially
compensated for by loss of melanosomes, which is likely to occur also
during treatment. In this case, further studies would be required to
study the fate of these melanosomes, including the possibility that
they accumulate in the trabecular meshwork.
|
Anxious Waiting
|
|---|
The ensuing months were nerve wracking, until the third
possibility was shown to be the case. After the pigmentation increase
was noted and the treatment stopped, the pigmentation remained constant
in all cases, showing neither a continuing increase nor a rapid
decrease and implying that this side effect on eye color is primarily
of cosmetic concern. This, however, does not mean that I did not
consider the effect to be a potentially very important finding,
particularly because my extensive reading and the studies we initiated
revealed how little is known about iridial melanocytes and their
physiological functions. My great interest in this phenomenon took me
back to my long-time interest in the homeostasis of IOF composition and
the microenvironment of intraocular tissues. Given the capacity of the
melanin system to scavenge free radicals, I began to consider the
possible role of iridial melanocytes in free radical and hydrogen
peroxide management of the aqueous humor.
|
Beneficial Potentials and Missed Opportunities
|
|---|
The possible beneficial effects of PG-induced stimulation of the
melanin system of the iris began to intrigue me, particularly when I
found two articles in journals not usually monitored by researchers in
our field. These provided evidence that iridial pigmentation is not
stable in all individuals throughout their lifetimes. One study
provided cross-sectional data strongly suggesting that iris color can
change in adults.47
The other report, based on the
Louisville Twin Study, showed that eye color changes occur well past
infancy—i.e., up to the age of six, the age to which these twins had
been observed until then.48
I had good luck when I contacted Dr. Matheny, the head of the
Louisville Twin Study. It turned out that even though his team had lost
interest in eye color, because much better genetic markers had become
available over the years, they had kept on measuring it, and for many
twins they had eye color records, made at regular intervals, from
infancy up to 20 to 24 years of age. Not only that, but they also
recorded the eye color of the parents at the time the twins were first
observed. Because the parents were mostly 20 to 30 years of age at that
time (up to 20–30 years before) reexamining them would give us data on
possible eye color changes covering most of adulthood. I went to Johan,
"begging" him to finance the data analysis to be done at the
University of Wisconsin, Madison, through the mediation of Paul
Kaufman, who has made many important contributions to this field and
with whom we had also collaborated on unrelated studies. This data
analysis showed that eye color changes occurred both in the twins up to
24 years of age and in the parents during adulthood. Interestingly,
approximately 15% of both the young-adult and the adult population
exhibited such changes in the sample of white twins.49
This was about the same percentage as the overall incidence of
PG-induced eye color change in the phase 3 Xalatan
study.8
5
This led us to the working hypothesis that
approximately 15% of white persons retain a mechanism that influences
the melanin synthesis of iridial melanocytes past infancy. Considering
the typically uneven initial coloration in the eyes that are most
likely to show this PG-response,8
50
our working
hypothesis further assumed that this type of iridial coloration is due
to age-dependent focal or regional loss of pigmentation that can be
reversed by PGs.
If you will excuse me for a nonscientific observation: Ever since we
observed this phenomenon and I read the literature on eye color
changes, I was looking in a conscious way at the irides of just about
everybody I met. And I became convinced that the multicolored irides
with irregular areas of hypo- or de-pigmentation and that exhibit the
greatest tendency toward PG-induced increased pigmentation do not occur
in children or young adults. Thus, it must represent an age-dependent
focal loss of pigmentation. It would also be of interest to compare the
appearance of the iris in persons with ocular hypertension and in
normal individuals and also in those with ocular hypertension and
normal persons who do and do not have a tendency toward development of
glaucomatous changes. I must emphasize that I do not believe in
"iridology," but it is indeed likely that age-dependent loss of
melanin from the iridial melanocytes reflects some local insufficiency,
possibly insufficient endogenous PG production that may also contribute
to the pathophysiology of the outflow routes and may reflect an
underlying condition that affects the whole eye, contributing even to
age-related macular degeneration.51
|
The Possible Role of the Iridial Melanin System in Intraocular
Homeostasis
|
|---|
Top
Introduction
The First Hurdle: The...
