The formation of NO (and
l-Cit) from
l-Arg
plays a major role in the regulation of cardiovascular functions. Under
physiological conditions, the formation of NO involves a constitutively
active calcium/calmodulin-dependent isoform of the enzyme NOS, eNOS,
that is predominantly localized to the endothe-lium.
4 Moreover, the production of NO can be evoked by inflammatory mediators
and cytokines because of the induction of a calcium-insensitive type of
NOS, iNOS,
4 mainly in the vascular smooth muscle. In the
course of NO formation, small amounts (2%–5% of the metabolized
l-Arg) of a stable intermediate, HOArg, can be released
from the active site of the enzyme in addition to NO and
l-Cit.
8 Because NOS is the only known source
of HOArg in the human body, HOArg constitutes a highly specific
indicator for NOS activity and NO production in vivo. In our study,
significantly elevated levels of HOArg were observed in the aqueous
humor of diabetic patients. The release of substantial amounts of HOArg
into the extracellular space has been shown in vivo and in vitro for
various cell types expressing iNOS.
8 Once released from
the NO-producing cell, HOArg competes with
l-Arg for
re-entry, most likely through the y
+ amino acid
carrier, and usually accumulates in the extracellular space. HOArg
uptake and metabolism may only be facilitated when the extracellular
concentration of
l-Arg decreases below a critical
threshold.
8 Because HOArg production reflects only a small
proportion of total NO activity, a significant increase in HOArg in the
aqueous humor of diabetic patients suggests a considerable increase in
NOS activity.
Interestingly, diabetic eyes with and without retinopathy contained
comparable levels of HOArg, indicating that a maximally increased
generation of NO occurs early in the course of diabetic retinal
disease. However, there was no correlation of HOArg concentrations with
duration of diabetes, although such an association is to be expected.
From our data, it is not possible to state whether there is a
relationship between the duration of diabetes mellitus and aqueous
humor levels of HOArg, because all diabetic subjects had a history of
ongoing diabetes mellitus for at least 6 years.
In the present study, we observed lower concentrations of
l-Arg in the aqueous humor of diabetic patients than in
nondiabetic patients, although this difference was not statistically
significant. This is consistent with previous reports on NOS activity
in vivo and may be in part the result of an increased uptake: Cellular
sources of
l-Arg include uptake of extracellular
l-Arg, intracellular synthesis from
l-Cit
(recycling) as well as, to a lesser extent, intracellular
proteolysis.
4 10 Moreover,
l-Cit levels in
diabetic and nondiabetic patients were not significantly different. The
most likely explanation for this observation is that many NO-producing
cells such as macrophages or endothelial cells can recycle
l-Cit to
l-Arg.
10 Because the
availability of
l-Arg constitutes a rate-limiting factor
for NO production in vivo, the generation of
l-Arg from
l-Cit appears to be an important mechanism in cells with
high NO activity, in particular in those cells expressing iNOS.
Experimental studies have demonstrated that retinal vascular
endothelial cells and pericytes synthesize iNOS under stimulated
conditions, a mechanism that may be expected in diabetic retinal
vascular disease.
11
It is noteworthy that, under normal conditions, the vascular
endothelium of the choroidal and anterior uveal circulations is likely
to be the main source of NO in aqueous humor.
12 Because NO
production has been shown to be lower in the diabetic vascular
endothelium, it may be speculated that aqueous humor levels of NO
metabolites are decreased in diabetic patients. However, it is not yet
clear whether the bioavailability of NO (due to the increased formation
of reactive oxygen species) rather than endothelial NOS activity or
expression is responsible for the endothelial dysfunction associated
with diabetes. Findings in studies of ocular hemodynamic reactivity to
NO inhibition suggest that either NOS activity is increased or
sensitivity to NO is decreased in patients with diabetes
mellitus.
13
The observation of enhanced NO production in diabetic eyes in vivo is
consistent with reports on NO as a mediator of physiological and
pathologic processes in the retina. It has been speculated that iNOS
activity may be involved in diabetic retinal vascular damage through
NO-induced increases in blood flow and vascular permeability. In
diabetic rats, [
125I]-labeled albumin entry across the
blood–retinal barrier, an indicator of altered vascular permeability,
has been shown to be decreased by continuous administration of
guanidine compounds.
7 Because guanidines inhibit iNOS, it
may be hypothesized that iNOS activity is increased in the diabetic
retina. However, because effects other than NOS inhibition may be
mediated by these compounds, there is no definitive proof that NO
produces diabetic retinal damage in vivo.
An increasing body of evidence that NO is an antiproliferative and
cytotoxic agent for retinal cells has been provided in several
experimental studies. Recent findings indicate that NO mediates
neurotoxic actions of glutamate that are responsible for ischemic
retinal injury.
14 In addition, glutamate neurotoxicity has
been shown to be attenuated by the inhibition of NOS
activity.
15 Other reports have demonstrated that bovine
retinal pigmented epithelial cells express iNOS after activation with
interferon-γ and lipopolysaccharide. It also has been shown that NO
exhibits an antiproliferative action on RPE cells, whereas fibroblast
growth factor has a protective effect by inhibiting the induction of
NOS.
3
From the current observations, it is not possible to implicate NO
in the initiation of retinal damage in diabetic retinopathy, because
factors promoting the expression of iNOS (and thus, enhanced NO
formation) may precede its release into the aqueous humor. Because
cytokines responsible for cellular proliferation and angiogenesis, such
as vascular endothelial growth factor, basic fibroblast growth factor,
or transforming growth factor-β
2 are found in vitreous
and aqueous humor or in neovascular membranes from eyes with
proliferative vitreoretinal disorders,
5 16 it is possible
to suggest that persistently raised intraocular levels of NO may depend
on a continuous operation of cytokine-mediated cellular reactions.
Moreover, because NO has been shown to be a potent inhibitor of
cytokine-induced proliferation of endothelial cells,
6 enhanced levels of NO-derived metabolites in aqueous humor may reflect
a regulatory mechanism that is potentially relevant to ocular
angiogenesis.
5 Therefore, we view the presence of high NOS
activity in diabetic eyes as a reactive process.
Our observation that NO production appears to be increased in eyes of
diabetic subjects without retinopathy supports the hypothesis that
cytokine-mediated effects occur before the onset of morphologic changes
in the vasculature. There is a reasonable likelihood that in pathologic
situations such as diabetic mellitus in which the expression of iNOS is
turned on, the generation of high concentrations of NO may be a
significant factor in the regulation of retinal vascular functions and
intraocular proliferative changes. The current findings merit further
investigations to elucidate whether there is a direct relationship
between the action of various cytokines and NO in vivo. An
understanding of the endogenous inhibition of ocular cellular
proliferation has a high degree of relevance, given the spectrum of
potentially blinding complications in diabetic retinopathy.
The authors thank Felicia Grimm for expert technical assistance.