An eye on insulin
Diabetic retinopathy, the most frequent complication of diabetes and leading cause of vision loss, involves vascular and neural damage in the retina. Insulin and IGF-1 signaling are now shown to contribute to retinal neovascularization, in part, by modulating the expression of various vascular mediators.
causative role to arginase will de-
pend on the results of the genetic
studies that Zimmermann and col-
leagues’ work warrants. Do single
nucleotide polymorphisms in hu-
man arginase dysregulate expression
and/or function so as to contribute
to asthma pathogenesis? The role of
arginase in the realm of asthma will
ultimately be dictated by the answer
to this question.
1. 2003. Asthma Prevalence, Health Care Use and Mor-
tality, 2000-2001. National Center for Health Sta-
tistics, Center for Disease Control.
2. Zimmermann, N., et al. 2003. Dissection of
experimental asthma with DNA microarray
analysis identifies arginase in asthma pathogen-
esis. J. Clin. Invest. 111:1863–1874.
3. Morris, S.M., Jr. 2002. Regulation of enzymes of
the urea cycle and arginine metabolism. Annu.
Rev. Nutr. 22:87–105.
4. Fischer, A., Folkerts, G., Geppetti, P., and
Groneberg, D.A. 2002. Mediators of asthma:
nitric oxide. Pulm. Pharmacol. Ther. 15:73–81.
5. Wechsler, M.E., et al. 2000. Exhaled nitric oxide in
patients with asthma: association with NOS1 geno-
type. Am. J. Respir. Crit. Care Med. 162:2043–2047.
6. Endo, M., et al. 2003. Induction of arginase I and
II in bleomycin induced fibrosis of mouse lung.
Am. J. Physiol. Lung Cell. Mol. Physiol. doi:10.1152/
7. Meurs, H., et al. 2002. Increased arginase activity
underlies allergen-induced deficiency of cNOS-
derived nitric oxide and airway hyperresponsive-
ness. Br. J. Pharmacol. 136:391–398.
8. Wei, L.H., Jacobs, A.T., Morris, S.M.J., and Ignar-
ro, L.J. 2000. IL-4 and IL-13 upregulate arginase I
expression by cAMP and JAK/STAT6 pathways in
vascular smooth muscle cells. Am. J. Physiol. Cell.
9. Chang, C., Zoghi, B., Liao, J.C., and Kuo, L. 2000.
The involvement of tyrosine kinases, cyclic
AMP/protein kinase A, and p38 mitogen-activat-
ed protein kinase in IL-13-mediated arginase I
induction in macrophages: its implications in
IL-13-inhibited nitric oxide production. J. Immunol.
10. Mowen, K.A., et al. 2001. Arginine methylation of
STAT1 modulates IFNalpha/beta-induced tran-
scription. Cell. 104:731–741.
11.Liu, B., Gross, M., ten Hoeve, J., and Shuai, K.
2001. A transcriptional corepressor of Stat1 with
an essential LXXLL signature motif. Proc. Natl.
Acad. Sci. U. S. A. 98:3203–3207.
12. Shi, O., Morris, S.M.J., Zoghbi, H., Porter, C.W., and
O’Brien, W.E. 2001. Generation of a mouse model
for arginase II deficiency by targeted disruption of
the arginase II gene. Mol. Cell. Biol. 21:811–813.
13. Iyer, R.K., et al. 2002. Mouse model for human
arginase deficiency. Mol. Cell Biol. 22:4491–4498.
The Journal of Clinical Investigation | June 2003 | Volume 111 | Number 12 1817
An eye on insulin
Sarah K. Bronson,
Chad E.N. Reiter,
and Thomas W. Gardner
Department of Cellular and Molecular Physiology, and
Department of Ophthalmology, Penn State College of Medicine, Hershey,
Diabetic retinopathy, the most frequent complication of diabetes and
leading cause of vision loss, involves vascular and neural damage in the
retina. Insulin and IGF-1 signaling are now shown (see the related arti-
cle beginning on page 1835) to contribute to retinal neovascularization,
in part, by modulating the expression of various vascular mediators.
J. Clin. Invest. 111:1817–1819 (2003). doi:10.1172/JCI200318927.
Physiologic and pathologic blood ves-
sel growth patterns are stimulated by
local and systemic influences. The
hope of angiogenesis research is to
understand these complex interac-
tions in order to provide better means
to control pathologic vessel forma-
tion, or perhaps stimulate appropri-
ate vessel growth, to reduce maladap-
tive consequences. In the retina, nor-
mal vessel growth occurs in the plane
of the retina from the optic nerve
toward the periphery in a radial pat-
tern and is guided by cues from astro-
cytes in the inner retina (1). This
growth is mediated by VEGF and
other ligands (2), while angioblasts
from the circulation can provide
endothelial progenitors (3).
The problem of retinal
Pathologic neovascularization of the
retina is a common and serious com-
plication of retinopathy of prematu-
rity (ROP) and diabetic retinopathy
(DR). The treatment of DR, ablation
of the diseased retina with laser pho-
tocoagulation or cryotherapy to cause
involution of the new vessels, has
remained fundamentally unchanged
for almost 50 years. The nature of
the growth promoting stimuli is not
well understood, but in ROP the
stimulus is assumed to be in part due
to perinatal retinal hyperoxia fol-
lowed by hypoxia. The classic re-
sponse to hypoxia includes hypoxia-
inducing factor-1 (HIF-1) translocation
to the nucleus and subsequent
downstream events such as the
upregulation of VEGF, eNOS, and
endothelin-1 (ET-1). The fact that
VEGF is increased in the vitreous of
diabetic patients makes it tempting
to speculate that diabetes induces a
hypoxic, or HIF-1–driven response
that is similar to that observed in
ROP. Current animal models of dia-
betes do not develop proliferative
retinopathy and the only model that
simulates the neovascularization
seen clinically is that induced by rel-
ative hypoxia in developing retinas.
Postnatal mice are placed in hyper-
oxic conditions for several days, at a
time when their vessels have not yet
reached the peripheral retina, which
causes vasoconstriction; when they
are returned to room air, the vaso-
constriction is relieved and neovas-
cularization develops when the reti-
na perceives relative hypoxia (4). This
model provides a rodent model of
neovascularization in the absence of
systemic metabolic defects due to
insulin depletion or resistance.
Smith et al. (5) previously showed
that an IGF-1 inhibitor blocked neo-
vascularization in this model but no
studies have examined the effect of
the insulin receptor.
Address correspondence to: Thomas W.
Gardner, Department of Cellular and
Molecular Physiology, and Department of
Ophthalmology, Penn State College of
Medicine, 500 University Drive, Hershey,
Pennsylvania 17033, USA.
Phone: (717) 531-6711;
Fax: (717) 531-7667;
Conflict of interest: The authors have
declared that no conflict of interest exists.
Nonstandard abbreviations used:
retinopathy of prematurity (ROP); diabetic
retinopathy (DR); hypoxia-inducible factor-1
(HIF-1); endothelin-1 (ET-1); insulin receptor
(IR); insulin-like growth factor-1 receptor
(IGF-1R), vascular endothelial insulin
receptor knockout (VENIRKO); vascular
endothelial insulin-like growth factor
receptor knockout (VENIFARKO).
1818 The Journal of Clinical Investigation | June 2003 | Volume 111 | Number 12
A role for insulin and IGF-1
in retinal neovascularization
In this issue of the JCI, Kondo and
colleagues (6) have utilized a clever
genetic system in mice to probe the
role of insulin and IGF signaling in
this experimental model of relative
hypoxia. They use two different
genetically altered lines of mice, one
of which contains an insulin receptor
(IR) gene that can be deleted in the
presence of Cre recombinase (7),
resulting in an absence of the IR in
all cells that have expressed the
recombinase, and a similar line that
contains an IGF-1 receptor (IGF-1R)
gene that can be deleted in the pres-
ence of the recombinase (8). Experi-
mental animals are bred to be
homozygous for either of the modi-
fied receptor genes and to carry a
transgene that expresses Cre recom-
binase under the control of the Tie-2
promoter/enhancer (9). The endothe-
lium in these animals is systemically
devoid of the IR or IGF-IR, and
referred to as vascular endothelial
insulin receptor knockout (VENIRKO)
and vascular endothelial insulin-like
growth factor receptor knockout
(VENIFARKO), respectively. While
neither Kondo et al. (6) nor Vicent et
al. (10) confirm the loss of IR or IGF-1R
proteins, it is presumed that the loss
is nearly complete given the severe
reduction in message detected by
quantitative PCR amplification of
endothelial cell RNA from these mice
and the careful description of the
Tie2-Cre transgenic line (9). The
results are quite interesting.
First, Kondo (6) and Vicent (10)
found no developmental or physiolog-
ical consequences of IR or IGF-1R loss
in endothelium; i.e., no gross or histo-
logical changes in the vasculature, and
no metabolic changes such as those
seen in subsets of the tissue-specific
knockouts for each of these receptors
(reviewed in ref. 11). However, when
young mice are put through the relative
hypoxia protocol there are distinct dif-
ferences between control mice and those
with an absence of the IR or IGF-1R in
the endothelium. They find reduced neo-
vascularization, and concomitantly
less immunoreactivity for markers of
neovascularization including, VEGF,
eNOS, and ET-1 in VENIRKO and to
a lesser extent in VENIFARKO mice,
compared to wild-type mice. One can-
not exclude the possibility that the
reduction of VEGF, eNOS, and ET-1
is simply because there are fewer
endothelial cells to express these pro-
teins. Nonetheless, this result is inter-
esting because deficiencies in the IR
and the IGF-1R signaling pathways
are critically linked to metabolic
defects in Type I and Type II diabetes
and insulin resistance — disease states
where neovascularization is a fre-
New findings beget new questions
The question is, what are the normal
functions of the IR and IGF-1R in the
retina? They probably do not regu-
late acute glucose utilization as they
do in skeletal muscle, liver, and fat
because their activity is constant
under physiological conditions (12).
In the CNS, the insulin and IGF sig-
naling pathways stimulate embryon-
ic growth, and loss of function muta-
tions cause developmental eye defects
in Drosophila (13). In retinal endothe-
lial cells, insulin stimulates mitogen-
esis and insulin transport (14, 15).
Hence, it is possible that the IR and
IGF-1R in the retina function largely
to provide trophic stimuli for main-
tenance of cell numbers. Therefore,
signaling via factors upregulated by
HIFs may intersect or augment the
IR and IGF-1R pathways to promote
neovascularization. The articles by
Kondo et al. (6) and Vicent et al. (10)
demonstrate that, from a develop-
mental standpoint, endothelial cell
insulin and IGF signaling are unnec-
essary, and consequences due to loss
of signaling through these receptors
may only be observed in response to
some perturbation or injury. The neo-
vascularization described by Kondo et
al. presumably requires increased pro-
liferation and/or survival of endothe-
lial cell progenitors, so it is not sur-
prising that disruption of the insulin
and IGF signaling pathways results
in less neovascularization. It would
be helpful to know if the reduced vas-
cularization in the knockout mice
resulted from less proliferation or
Retinal neovascularization is the result of hypoxia-induced damage to the neural retina and
its capillaries (a). Following investigation of the effects of hypoxia in the presence and absence
of IR/IGF-1R on retinal vascularization, Kondo et al. (6) found that under normoxic condi-
tion (b), the retinas of mice develop normally in the absence of endothelial IR/IGF-1R. Pre-
sumably, sufficient growth factors (e.g., VEGF) are present to facilitate normal development.
Under conditions of relative hypoxia and in the presence of endothelial IR/IGF-1R (c), VEGF,
eNOS, and ET-1 are increased, leading to extra-retinal neovascularization. Under conditions
of relative hypoxia and in the absence of endothelial IR/IGF-1R (d), VEGF, eNOS, and ET-1
are reduced, possibly due to impaired HIF-1 activation or reduced PI3K activity related to
IG/IGF-1R. Reduced neovascularization results from less IR/IGF-IR input.
more apoptosis. The deciphering of
downstream signals such as the activ-
ity of PI3K in the endothelium in
response to relative hypoxia might
also shed some light on this provoca-
Looking forward therapeutically
The observations that loss of the IR
and IGF-1R signaling pathways appear
to counter the response to relative
hypoxia and do not alter vasculogene-
sis in the absence of relative hypoxia
suggest that it is not insulin and IGF-1
signaling in endothelial cells alone that
promotes neovascularization in dia-
betes. It also prompts us to look to
other cell types in the retina for a
response to diminished insulin action.
The experiments of Kondo et al. (6) do
not address the role of retinal glia or
neurons in vascular regulation. An-
swers to these questions will directly
impact therapeutic choices for diabet-
ic patients. Kondo et al. suggest that, in
diabetes, inhibition of retinal insulin or
IGF-1 signaling in the eye might be
beneficial, however, in the context of
what is understood to be systemically
reduced insulin signaling in these dis-
ease states, therapeutic blocking of
insulin signaling appears counterintu-
itive. Identification of specific mole-
cules at the intersection of the HIF-1
and insulin and IGF-1 signals, as well
as a thorough understanding of how
the varied cell types in the retina
respond to the diabetic state, will nec-
essarily precede therapeutic trials to
prevent loss of vision.
1. Provis, J.M., et al. 1997. Development of the
human retinal vasculature: cellular relations and
VEGF expression. Exp. Eye. Res. 65:555–568.
2. Neely, K.A., and Gardner, T.W. 1998. Ocular neo-
vascularization: clarifying complex interactions.
Am. J. Pathol. 153:665–670.
3. Grant, M.B., et al. 2002. Adult hematopoietic
stem cells provide functional hemangioblast
activity during retinal neovascularization. Nat.
4. Smith, L.E., et al. 1994. Oxygen-induced retinopa-
thy in the mouse. Invest. Ophthalmol. Vis. Sci.
5. Smith, L.E., et al. 1999. Regulation of vascular
endothelial growth factor-dependent retinal neo-
vascularization by insulin-like growth factor-1
receptor. Nat. Med. 5:1390–1395.
6. Kondo, T., et al. 2003. Knockout of insulin and
IGF-1 receptors on vascular endothelial cells pro-
tects against retinal neovascularization. J. Clin.
Invest. 111:1835–1842. doi:10.1172/JCI200317455.
7. Kulkarni, R.N., et al. 1999. Tissue-specific
knockout of the insulin receptor in pancreatic
beta cells creates an insulin secretory defect
similar to that in type 2 diabetes. Cell.
8. Holzenberger, M., et al. 2003. IGF-1 receptor reg-
ulates lifespan and resistance to oxidative stress
in mice. Nature. 421:182–187.
9. Kisanuki, Y.Y., et al. 2001. Tie2-Cre transgenic
mice: a new model for endothelial cell-lineage
analysis in vivo. Dev. Biol. 230:230–242.
10. Vicent, D., et al. 2003. The role of endothelial
insulin signaling in the regulation of vascular
tone and insulin resistance. J. Clin. Invest.
11.Kahn, C.R., Bruning, J.C., Michael, M.D., and
Kulkarni, R.N. 2000. Knockout mice challenge
our concepts of glucose homeostasis and the
pathogenesis of diabetes mellitus. J. Pediatr.
Endocrinol. Metab. 13:1377–1384.
12. Reiter, C.E.N., and Gardner, T.W. 2003. Retinal
insulin and insulin signaling: implications for
diabetic retinopathy. Prog. Ret. Eye Res.
13. Broglio, W., et al. 2001. An evolutionarily con-
served function of the Drosophila insulin receptor
and insulin-like peptides in growth control. Curr.
14. King, G.L., Goodman, A.D., Buzney, S., Moses, A.,
and Kahn, C.R. 1985. Receptors and growth-pro-
moting effects of insulin and insulinlike growth
factors on cells from bovine retinal capillaries and
aorta. J. Clin. Invest. 75:1028–1036.
15. Stitt, A.W., Anderson, H.R., Gardiner, T.A., Bailie,
J.R., and Archer, D.B. 1994. Receptor-mediated
endocytosis and intracellular trafficking of
insulin and low-density lipoprotein by retinal vas-
cular endothelial cells. Invest. Ophthalmol. Vis. Sci.
The Journal of Clinical Investigation | June 2003 | Volume 111 | Number 12 1819
Tolerance: Of mice and men
David H. Sachs
Transplantation Biology Research Center, Massachusetts General Hospital, Boston,
Little is known about the effect of an individual’s immune history on his
or her response to an allogeneic tissue transplant. An important study
(see the related article beginning on page 1887) now reveals that individ-
uals harboring virally-induced memory T cells that are cross reactive with
donor alloantigen are resistant to conventional strategies designed to
induce transplant tolerance.
J. Clin. Invest. 111:1819–1821 (2003). doi:10.1172/JCI200318926.
Enormous progress has been made in
the field of transplantation during
the past three decades, due in large
part to the availability of effective
immunosuppressive drugs. Although
all of these agents suppress the
immune response nonspecifically
with respect to antigen, the most ef-
fective ones exhibit sufficient selec-
tivity so that rejection can be avoided
without undue compromise of the
host’s ability to respond to microbial
pathogens. Nevertheless, patients on
immunosuppressive medications are
constantly walking a tightrope be-
tween the consequences of too little
suppression (i.e., rejection) and of
too much suppression (infections or
cancer) of their immune system. In
addition, even in patients without
complications due to their immuno-
suppression, there is an inexorable
loss of transplanted organs due to
chronic rejection at a rate of approx-
imately 5% per year (1).
For these reasons, ever since the
description of acquired tolerance to
allografts in mice by Medawar and
colleagues appeared in 1953 (2), a
major goal of both clinicians and
immunologists in the field of trans-
plantation has been the induction of
tolerance in transplant recipients.
What has been most frustrating
about this quest has been the fact
that a very large number of successful
approaches to the induction of toler-
ance have been reported in rodent
models, but have failed when
attempted in large animals, especial-
ly in nonhuman primates and in
humans (Table 1). Indeed, as clinical
results of organ transplants using
standard immunosuppression are so
good, at least in the short term, many
clinicians are no longer interested in
Address correspondence to: David H. Sachs,
Transplantation Biology Research Center,
Massachusetts General Hospital, MGH East,
Building 149-9019, 149 13th Street,
Boston, Massachusetts 02129, USA.
Phone: (617) 726-4065; Fax: (617) 726-4067;
Conflict of interest: The author has declared
that no conflict of interest exists.