© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 655–660

655

Introduction

Ocular neovascularization, which is strongly associated withretinal ischaemia, is the hallmark of proliferative diabetic retino-pathy (PDR). Nitric oxide (NO) is a free radical, produced

Correspondence to

: Dr Rafael Simó, Diabetes Unit, Endocrinology Division, Hospital General Universitari Vall d’Hebron, Pg. Vall d’Hebron, 119–129, 08035 Barcelona, Spain. E-mail: rsimo@hg.vhebron.es

Abstract

Aims

Several reports have implicated nitric oxide (NO) in the angiogenicprocess. The assessment of NO stable end products, nitrite and nitrate (NOx),is commonly used as a measure of NO production in biological fluids. The aimsof the study were to investigate NOx concentrations in the vitreous fluidof patients with proliferative diabetic retinopathy (PDR) and to evaluate therelationship between NOx and vascular endothelial growth factor (VEGF).

Patients and methods

Serum and vitreous fluid samples were obtained simul-taneously at the time of vitreoretinal surgery from 23 patients with PDR, and 17control non-diabetic patients with non-proliferative ocular disease. NOxwas determined by using the Griess reaction and VEGF levels were assessedby ELISA.

Results

The intravitreous concentration of NOx was significantly elevated inpatients with PDR in comparison with the control group (31.6

±

2.96

µ

mol/ lvs. 18

±

2.46

µ

mol/ l;

P

= 0.01). However, we did not detect any differencesbetween NOx serum concentrations. We observed a correlation between serumand vitreous levels of NOx in diabetic patients (

r

= 0.79;

P

< 0.001), but not inthe control group. Intravitreous levels of VEGF in patients with PDR werehigher than those obtained in serum (1.42 ng/ml (0.12–7.62) vs. 0.12 ng/ml(0.03–0.42);

P

< 0.01). Vitreal levels of VEGF were strikingly higher in patientswith PDR than in the control subjects (1.42 ng/ml (0.12–7.62) vs. 0.009 ng/ml(0.009–0.04);

P

< 0.001). No correlation between vitreal concentrations ofNOx and VEGF was observed, either in diabetic patients or in the controlgroup.

Conclusions

NOx and VEGF are increased but not related in the vitreous fluidof diabetic patients with PDR. Our results suggest that serum diffusion couldplay a significant role in explaining the increase of NOx. By contrast, intraocularproduction seems to be the main factor responsible for the intravitreousenhancement of VEGF.

Diabet. Med. 19, 655–660 (2002)

Keywords

nitric oxide, vascular endothelial growth factor, vitreous fluid,diabetic retinopathy

Blackwell Science, LtdOxford, UKDMEDiabetic Medicine0742-3071Blackwell Science Ltd, 200219Original ArticleOriginal articleNO and VEGF in diabetic retinopathy

C. Hernández et al.

Nitric oxide and vascular endothelial growth factor concentrations are increased but not related in vitreous fluid of patients with proliferative diabetic retinopathy

C. Hernández, A. Lecube, R. M. Segura*, L. Sararols† and R. Simó

Endocrinology Division, *Biochemistry Department and †Department of Ophthalmology, Diabetes Unit, Hospital General Vall d’Hebron, Barcelona, Spain

Accepted 8 February 2002

DME_768.fm Page 655 Tuesday, July 23, 2002 1:53 PM

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 655–660

656

NO and VEGF in diabetic retinopathy •

C. Hernández et al

.

from L-arginine by NO synthases (NOS) that promotes angio-genesis in several tissues [1–3]. It has been demonstrated

in vitro

that NO plays a role in different aspects of the angiogenicresponse to vascular endothelial growth factor (VEGF) [4],a recognized growth factor involved in the development ofPDR [5,6]. Furthermore, VEGF-induced angiogenesis,

in vivo

,is significantly attenuated by NOS inhibition [7].

It is rather difficult to measure NO since it is a moleculewith a short half-life and reacts rapidly with free oxygen,oxygen radicals, redox metals, sulphhydrils, disulphides andoxygenated haemoglobin [8]. However, determination ofthe stable end products of NO radical, nitrite and nitrate(NOx), by the Griess procedure can be successfully used as ameasure of the production of NO radical in human bodyfluids [9,10].

Because it is not possible to explore the human retina

in vivo

, the vitreous fluid obtained from diabetic patients sub-mitted to vitreoretinal surgery could be a useful material withwhich to investigate the mechanisms involved in the develop-ment of PDR. However, there are two main problems thatcould lead to a misinterpretation of the results. First, vitreoushaemorrhage that often occurs in advanced PDR precludes theusefulness of vitreous fluid to explore indirectly the intraocularproduction of NO. Second, the serum levels of NO could influ-ence their vitreal concentrations, and, in consequence, thispossibility should be considered both in the design of the studyand the analysis of the results.

In the present study we have determined the intravitreousconcentrations of NOx in diabetic patients with PDR whiletaking into account the confounding factors mentioned above.In addition, we have investigated the relationship between theintravitreous levels of NOx and VEGF in these patients.

Patients and methods

Subjects

The study included 23 patients with PDR (nine with Type 1diabetes and 14 with Type 2; mean age 50

±

15 years) in whoma classic three port pars plana vitrectomy was performed.The main clinical features of the diabetic patients are shownin Table 1. Seventeen non-diabetic patients (mean age62

±

14 years) with other conditions requiring vitrectomy, butin whom the retina was not directly affected by neovasculariza-tion, served as a control group. In the control group thediagnosis included macular oedema (

n

= 1), epiretinal mem-brane (

n

= 6), rhegmatogenous retinal detachment (

n

= 3),and macular hole (

n

= 7).Retinopathy was graded intraoperatively in all eyes by the

same ophthalmologist and classified as active or quiescentaccording to the characteristics previously reported by Aiello

et al

. [5].Patients with a history of previous vitreoretinal surgery

and vitreous haemorrhage within the previous 2 months wereexcluded. Moreover, we excluded all cases in which intravitreoushaemoglobin was detectable.

Sample collection

Undiluted vitreous samples (0.5–1 ml) were obtained at theonset of vitrectomy by aspiration into a 1-ml syringe attachedto the vitreous cutter (Alcon Model, Ten-Thousand Ocutome;Irvine, CA, USA) before starting intravitreal infusion of abalanced salt solution. The vitreous samples were transferredto a tube, placed immediately on ice and centrifuged at 16 000

g

for 5 min at 4

°

C. Supernatants were frozen at

−

80

°

C until assayed.For serum determinations, blood samples were collected

simultaneously with the vitrectomy, then centrifuged at3000

g

for 10 min at 4

°

C to obtain serum, then aliquotedand stored at

−

80

°

C until assayed.The protocol for sample collection was approved by the

hospital ethics committee and all patients were fully informedbefore they gave their consent.

Laboratory assays

NO assay

Serum and vitreous levels of the NO stable metabolites, nitrite+ nitrate (NOx), were measured as nitrite using the Griess reac-tion [11]. Nitrate was enzymatically reduced to nitrite withnitrate reductase [12]. Samples were diluted four-fold withphosphate buffer pH 7.5. Nitrate reductase, NADPH, andFAD (Boehringer Mannheim, Mannheim, Germany) were addedto duplicates of potassium nitrate standards, reagent blankscontaining phosphate buffer, and diluted samples to yield finalconcentrations of 234 U/l, 155

µ

mol/l, and 35

µ

mol/l, respect-ively. The reaction mixture was incubated for 40 min at roomtemperature. To oxidize the unreacted NADPH and avoidinterference with the Griess reaction, pyruvic acid (Sigma-Aldrich,St Louis, MO, USA) and LDH from hog muscle (BoehringerMannheim) were then added to give final concentrations of970

µ

mol/ l and 7460 U/l, respectively [9]. The mixture wasfurther incubated for 10 min at room temperature. At the endof this incubation step, proteins were precipitated by adding1/17th volume of ZnSO

4

300 g/ l and centrifuging at 12 000 gfor 10 min [9]. We applied 150

µ

l of the corresponding super-natants to microtitre plate wells, followed by 75

µ

l of eachsulphanilamide (1 g/ l) and N-1-naphthylethylenediamine

Table 1 Main clinical features of diabetic patients included in the study

Type 1 n = 9

Type 2 n = 14

Age (years) 36 (24–43) 60 (51–72)Gender (M/F) 5/4 4/10BMI (kg/m2) 22.3 ± 2.2 26.6 ± 2.7Duration of diabetes (years) 22.5 (12–32) 22 (6–38)HbA1c (%) 9.1 ± 1.5 8.2 ± 4.4AER > 20 µg/min (%) 70 53Treatment (n)

Oral anti-diabetic agents 0 1Insulin 9 10Oral agents + insulin 0 3

Prior laser treatment (n) 8 9

Data are mean (range) or mean ± SD.

DME_768.fm Page 656 Tuesday, July 23, 2002 1:53 PM

Original article

657

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 655–660

(0.1 g/l), both in 25 g/l phosphoric acid (Boehringer Mann-heim). After 10 min of colour development, absorbanceswere measured on a microplate reader (BioWhittaker Inc.,Walkersville, MD, USA) at 540 nm as the main wavelength and630 nm as the reference wavelength. Results are given as NOx(mmol/l) calculated from a nitrate standard curve with sixconcentrations ranging from 2.5 to 79

µ

mol/l (linear regressionadjust), after correction for the dilution factor.

VEGF assay

VEGF was determined from undiluted vitreous and serumsamples by ELISA (R&D Systems, Abingdon, UK), which uses ananti-human monoclonal antibody specific to VEGF precoatedonto a microtitre plate, and an enzyme-linked polyclonalantibody specific to VEGF as a second antibody. The coeffi-cients of variation intra-assay and interassay were 3.8% and5.1%, respectively. For data processing, we allocated theminimum value detected by ELISA (0.009 ng/ml) to all sampleswith concentrations below the detection threshold.

Intravitreous haemoglobin and protein measurements

Vitreous haemoglobin levels were determined by specto-photometry (Uvikon 860; Kontron Instruments, Zurich, Swit-zerland) using the classic method of Harboe for measuringplasma haemoglobin in micromolar concentration [13]. Thismethod has been recently revalidated [14], and in our handsthe lowest limit of detection was 0.03 mg/ml. Vitreal proteinswere measured by a previously validated microturbidimetricmethod with an autoanalyser (Hitachi 917; BoehringerMannheim). This method, based in the benzetonium chloridereaction, is a highly specific method for detection of proteinsand has a higher sensibility and reproducibility than the classicLowry method. The lowest level of proteins detected is0.02 ng/ml. The coefficients of variation intra-assay and inter-assay were 2.9% and 3.7%, respectively.

Statistical analysis

The Kolmogorov–Smirnov test was employed to confirm theassumption of the normality of the variables. Student’s

t

-testwas used to compare NOx concentrations between diabeticand control patients and the results were expressed as mean

±

SE

. Intravitreous VEGF and protein concentrations weredisplayed as median and range, in view of their skewed dis-tribution, and for comparisons the Mann–Whitney

U

-test wasused. Correlations were examined by Spearman’s rank correla-tion. Levels of statistical significance were set at

P

< 0.05.

Results

The results of all the measurements in the serum and the vitreousfluid of diabetic patients and control subjects are summarizedin Table 2.

The intravitreous concentration of NOx was significantlyelevated in patients with PDR in comparison with the controlgroup (31.6

±

2.96

µ

mol/ l vs. 18

±

2.46

µ

mol/ l;

P

= 0.01).However, we did not detect significant differences in NOxserum concentrations between diabetic patients and control

subjects (48.1

±

9

µ

mol/ l vs. 39.5

±

3.6

µ

mol/ l;

P

= NS). Astrong correlation between the serum and vitreous levels ofNOx was observed in diabetic patients (Fig. 1), but not in thecontrol group. We did not detect differences between Type 1and Type 2 diabetic patients in either the serum or the vitreouslevels of NOx. In addition, age and gender appeared to have noinfluence on serum and vitreous NOx levels, either in diabeticpatients or control subjects.

Vitreal NOx concentrations were not related to the severityof diabetic retinopathy (Table 3). In addition, no correlationbetween glucose and NOx levels in the vitreous fluid wasdetected either in patients with PDR or in the control group.

Vitreous VEGF concentrations were higher in patients withPDR than in the control group (1.42 ng/ml (0.12–7.62) vs.0.009 ng/ml (0.009–0.04);

P

< 0.001). Intravitreous proteinlevels were also higher in diabetic patients with PDR than incontrol subjects (Table 2). However, the increase of intra-vitreal proteins did not explain the enhancement of VEGFdetected in diabetic patients because the ratio of VEGF/vitrealproteins remained statistically different between diabeticpatients and the control group (0.46 ng/mg (0.017–4.72) vs.0.01 ng/mg (0.003–0.03);

P

< 0.0001). In addition, a relation-ship between diabetic retinopathy severity and VEGF levelswas observed (Table 3). Intravitreous levels of VEGF inpatients with PDR were higher than those obtained in serum(1.42 ng/ml (0.12–7.62) vs. 0.12 ng/ml (0.03–0.42);

P

< 0.01).

Table 2 Laboratory data in patients with proliferative diabetic retinopathy (PDR) and control subjects

PDR

n = 23Control n = 17 P

VitreousGlucose (mg/dl) 125.6 ± 18.9 71.5 ± 5.1 < 0.05Proteins (mg/ml) 2.78 (1.4–13.8) 0.76 (0.2–2.6) < 0.0001NOx (µmol/ l) 31.6 ± 2.96 18 ± 2.46 < 0.01VEGF (ng/ml) 1.42 (0.12–7.62) 0.009 (0.009–0.04) < 0.001SerumNOx (µmol/ l) 48.1 ± 9 39.5 ± 3.6 NSVEGF (ng/ml) 0.12 (0.03–0.43) 0.15 (0.03–0.54) NS

Data are expressed as mean ± SE and median (range).

Figure 1 Correlation between serum and vitreous levels of nitrite and nitrate (NOx) in patients with proliferative diabetic retinopathy.

DME_768.fm Page 657 Tuesday, July 23, 2002 1:53 PM

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 655–660

658

NO and VEGF in diabetic retinopathy •

C. Hernández et al

.

However, in the control group, serum levels of VEGF werehigher than obtained in vitreous fluid. No correlation betweenvitreal and serum levels of VEGF was detected in eitherdiabetic patients or the control group.

Finally, in both the diabetic patients with PDR and controlsubjects, there was no correlation between VEGF and NOx.This was true for both the serum and the vitreous fluid (Fig. 2).

Discussion

NO has been implicated in all steps of angiogenesis includingdissolution of matrix, endothelial cell migration, proliferationand organization into a network structure, followed by lumenformation [1–4]. An increased NO concentration in theaqueous humour of diabetic patients with retinopathy hasbeen reported, especially in patients with neovascular glaucoma[15]. However, to the best of our knowledge, there are no studiesevaluating the levels of NOx in the vitreous fluid.

In the present study we observed a higher level of NOxin the vitreous fluid of patients with PDR in comparisonwith the control group. This finding could not be attributedto the higher NOx detected in the systemic circulation becausewe did not detect any significant difference between theserum levels of NOx in diabetic patients and the controlgroup. Moreover, the influence of vitreous haemorrhage onintravitreous NOx levels can be ruled out because thosepatients in whom intravitreous haemoglobin was detectablewere excluded. We detected a strong direct correlationbetween the serum and vitreous levels of NOx in diabeticpatients, but not in the control group. Certainly, the dis-ruption of the blood retinal barrier that exists in diabeticretinopathy could favour the passage of NO from systemiccirculation into the vitreous body. This concept is supportedby the increase of intravitreal protein concentration observedin diabetic patients in comparison with non-diabetic controls.Therefore, serum diffusion could well be a significant cause ofthe enhancement of NOx observed in diabetic patients withPDR. However, it has been demonstrated that interleukin-1, tumour necrosis factor, and interferon-gamma, which areincreased in the vitreous fluid of diabetic patients with PDR,could induce expression of the inducible NOS (iNOS)isoform in retinal pigment epithelium and Müller cells [16–18].In addition, it has been recently reported that iNOS mediatesthe change from retinal to vitreal neovascularization in ischaemicretinopathy [19]. Therefore, although serum diffusion seemsto play a significant role in the increase of NOx observed inpatients with PDR, our study does not allow us to measure thecontribution of iNOS to its enhancement.

It could be argued that the increase of serum-derivedNOx observed in the vitreous of diabetic patients mightcontribute to the angiogenic process, but we did not find arelationship between NOx intravitreous levels and PDRactivity. Furthermore, as mentioned above, there is increasingevidence that several retina cell types express not onlyconstitutive NOS with proangiogenic proprieties [4,7], but arealso capable of generating large amounts of NO through theiNOS pathway after cytokine stimulation or hypoxia [16–20],which has an anti-angiogemic effect [19,21,22]. Therefore,future studies of NOS isoforms in order to evaluate moreaccurately the balance between proangiogenic and anti-angiogenic signals are needed.

A dose-dependent reduction of NO synthesis from retinalmicrovascular endothelial cells after exposure to supraphys-iological glucose concentrations (15 and 25 mmol/l) has beendemonstrated [23]. However, we have not detected a relation-ship between NOx and intravitreous glucose concentration.One possible explanation for this finding is that the vitreousglucose was not high enough to influence the synthesis of NO.In fact, in only one case did the intravitreous glucose achievethe levels which have been reported as necessary to suppressNO synthesis. Alternatively, the lack of a relationship betweenintravitreous glucose and NOx could be explained by theintravitreous NOx being mainly serum-derived.

Table 3 Intravitreous nitrite and nitrate (NOx) and vascular endothelial growth factor (VEGF) levels according to the activity of proliferative diabetic retinopathy.

Quiescent n = 8

Active n = 15 P

NOx (µmol/l) 32.5 ± 9.4 35.4 ± 3.5 NSVEGF (ng/ml) 0.24 (0.12–1.26) 2.13 (0.37–6.66) < 0.01

Figure 2 Lack of relationship between vascular endothelial growth factor (VEGF) and nitrite and nitrate (NOx) in either the serum or vitreous fluid in patients with proliferative diabetic retinopathy.

DME_768.fm Page 658 Tuesday, July 23, 2002 1:53 PM

Original article

659

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 655–660

We have confirmed previous reports published by otherauthors [5] and ourselves [6], that intravitreal levels of VEGFwere significantly higher in patients with PDR in comparisonwith the control group. This difference remained significantafter adjusting for vitreal proteins. Furthermore, intravitreouslevels of VEGF were 12-fold higher than those obtained inserum in diabetic patients with PDR, although intravitreousprotein concentration is at least 20-fold less than serum [6,24].Thus, the higher intravitreal concentration of VEGF in relationto its serum levels observed in patients with PDR stronglysupports its intraocular production. Finally, we did not findany correlation between serum and vitreous VEGF concentra-tions. Taken together, these data suggest that exaggeratedintraocular synthesis of VEGF but not serum diffusion is themain factor responsible for their intravitreous enhancement indiabetic patients with PDR. Another possible mechanism thatcould contribute to the elevated concentrations of VEGF invitreous fluid is the ability of heparan sulphate proteoglycans,abundant molecules in the vitreous body, to sequestrate VEGFby means of recognized binding sites [25].

VEGF produces a dose-dependent up-regulation of NOgeneration in human endothelial cells [26] and it has beenshown that NO production by NOS is involved in signallingVEGF’s permeability-enhancing effects [27,28]. Thus, NOSinhibition blocks VEGF-induced vascular hyperpermeabilityin all ocular tissues [29]. NO production also contributes tothe angiogenic properties of VEGF in human endothelial cells[4]. On the other hand, NO has been shown to down-regulateVEGF and their receptor (VEGFR2)

in vitro

[30,31] and

in vivo

[32,33]. However, in the present study we did not detecta relationship between VEGF and NOx concentrations in thevitreous fluid of patients with PDR. This unexpected resultmight be because the enhancement of VEGF in the vitreousfluid of patients with PDR is mainly due to intraocularsynthesis. In contrast, the intravitreous increase of NOx is anunreliable indicator of its retinal synthesis since it is signi-ficantly affected by serum diffusion.

We conclude that NOx and VEGF are increased but notrelated in the vitreous fluid of diabetic patients with PDR.In addition, our results suggest that the lack of relationshipbetween intravitreous concentrations of NOx and VEGFmight be partly explained by the different origins of theirintravitreal enhancement. However, further studies arerequired to clarify this issue.

Acknowledgements

This work was supported by grants from the Ministerio deSanidad y Consumo (FIS 98/1270), the Ministerio de Cienciay Tecnología (PM 99-0136), the Associació Catalana deDiabetis (Ajut per la Recerca 1999) and Novo Nordisk S.A.(01/0066). We thank Dr F. Campos for the analysis of vitreoushaemoglobin and Dr José García-Arumí for his contribution tothe vitreous collection, and Michael Willy for his assistancewith manuscript preparation.

References

1 Ziche M, Morbidelli L, Masini E, Amerini S, Granger HJ, Maggi CA

et al.

Nitric oxide mediates angiogenesis in vivo and endothelial cellgrowth and migration in vitro promoted by substance P.

J Clin Invest

1994;

94

: 2036–2044.2 Leibovich SJ, Polverini PJ, Fong TW, Harlow LA, Koch AE.

Production of angiogenic activity by human monocytes requires anL-arginine/nitric oxide-synthase-dependent effector mechanism.

Proc Natl Acad Sci USA

1994;

91

: 4190–4194.3 Andrade SP, Hart IR, Piper PJ. Inhibitors of nitric oxide synthase

selectively reduce flow in tumor-associated neovasculature.

Br JPharmacol

1992;

107

: 1092–1095.4 Papapetropoulos A, Garcia-Cardena G, Madri JA, Sessa WC. Nitric

oxide production contributes to the angiogenic properties of vascularendothelial growth factor in human endothelial cells.

J Clin Invest

1997;

100

: 3131–3139.5 Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD, Shah ST

et al.

Vascular endothelial growth factor in ocular fluid of patients withdiabetic retinopathy and other retinal disorders.

N Engl J Med

1994;

331

: 1480–1487.6 Burgos R, Simó R, Audí L, Mateo C, Mesa J, García-Ramirez M

et al.

Vitreous levels of vascular endothelial growth factor are notinfluenced by its serum concentrations in diabetic retinopathy.

Diabetologia

1997;

40

: 1107–1109.7 Ziche M, Morbidelli L, Choudhuri R, Zhang HT, Donnini S,

Granger HJ

et al.

Nitric oxide synthase lies downstream from vascularendothelial growth factor-induced but not basic fibroblast growthfactor-induced angiogenesis.

J Clin Invest

1997;

99

: 2625–2634.8 Stamler JS, Singel DJ, Loscalzo J. Biochemistry of nitric oxide and its

redox-activated forms.

Science

1992;

258

: 1898–1902.9 Moshage H, Kok B, Huizenga JR, Jansen PLM. Nitrite and nitrate

determinations in plasma: a critical evaluation.

Clin Chem

1995;

41

:892–896.

10 Guevara I, Iwanejko J, Dembinska-Kiec A, Pankiewicz J, Wanat A,Anna P

et al.

Determination of nitrite /nitrate in human biologicalmaterial by the simple Griess reaction.

Clin Chim Acta

1998;

274

: 177–188.11 Green LC, Wagner DA, Glogowski J, Skipper PL, Wishnok JS,

Tannenbaum SR. Analysis of nitrate, nitrite, and [15N]nitrate inbiological fluids.

Anal Biochem

1982;

126

: 131–138.12 Gilliam MB, Sherman MP, Griscavage JM, Ignarro LJ. A spectropho-

tometric assay for nitrate using NADPH oxidation by Aspergillusnitrate reductase.

Anal Biochem

1993;

212

: 359–365.13 Harboe M. A method for determination of hemoglobin in plasma by

near-ultraviolet spectrophotometry.

Scand J Clin Lan Invest

1959;

11

: 66–70.14 Fairbanks VF, Ziesmer SC, O’Brien PC. Methods for measuring

plasma hemoglobin in micromolar conentration compared.

ClinChem

1992;

38

: 132–140.15 Chiou S-H, Chang C-J, Chou C-K, Hsu W-M, Liu J-H, Chiang C-H.

Increased nitric oxide levels in aqueous humor of diabetic patientswith neovascular glaucoma.

Diabetes Care

1999;

22

: 861–862.16 Goreau O, Lepoivre M, Becquet F, Courtois Y. Lipopolysaccharide

and cytokines induce a macrophage-type of nitric oxide synthase inbovine retina pigmented epithelial cells.

Biochem Biophys ResCommun

1992; 186: 854–859.17 Goreau O, Lepoivre M, Becquet F, Courtois Y. Differential regula-

tion of inducible nitric oxide synthase by fibroblast growth factorsand transforming growth factor beta. Proc Natl Acad Sci USA 1993;90: 4276–4280.

18 Goreau O, Hicks D, Courtois Y. Human retina pigment epithelialcells produce nitric oxide in response to cytokines. Biochem BiophysRes Commun 1994; 198: 120–126.

19 Sennlaub F, Courtois Y, Goureau O. Inducible nitric oxide synthasemediates the change from retinal to vitreal neovascularization inischemic retinopathy. J Clin Invest 2001; 107: 717–725.

DME_768.fm Page 659 Tuesday, July 23, 2002 1:53 PM

© 2002 Diabetes UK. Diabetic Medicine, 19, 655–660

660 NO and VEGF in diabetic retinopathy • C. Hernández et al.

20 Nathan C. Inducible nitric oxide synthase: what difference does itmake? J Clin Invest 1997; 100: 2417–2423.

21 Sennlaub F, Courtois Y, Goureau O. Nitric oxide synthase-II isexpressed in severe corneal alkali burns and inhibits neovasculariza-tion. Invest Ophthalmol Vis Sci 1999; 40: 2773–2779.

22 Pipili-Synetos E, Kritikou S, Papadimitriou E, Athanassiadou A,Flordellis C, Maragoudakis ME. Nitric oxide synthase expression,enzyme activity and NO production during angiogenesis in the chickchorioallantoic membrane. Br J Pharmacol 2000; 129: 207–213.

23 Chakravarthy U, Hayes RG, Stitt A, McAuley E, Archer DM.Constitutive nitric oxide synthase expression in retinal vascularendothelial cells is suppressed by high glucose and advanced glyca-tion end products. Diabetes 1998; 47: 945–952.

24 Hernández C, Burgos R, Cantón A, García-Arumí J, Segura RM,Simó R. Vitreous levels of vascular cell adhesion molecule and vascularendothelial growth factor in patients with proliferative diabetic retino-pathy. Diabetes Care 2001; 24: 516–521.

25 Robinson CJ, Stringer SE. The splice variants of vascular endothelialgrowth factor (VEGF) and their receptors. J Cell Sci 2001; 114: 853–865.

26 Hood JD, Meininger CJ, Ziche M, Granger HJ. VEGF upregulatesecNOS message, protein, and NO production in human endothelialcells. Am J Physiol 1998; 274: H1054–H1058.

27 Feng Y, Venema VJ, Venema RC, Tsai N, Behzadian MA, Caldwell RB.

VEGF-induced permeability increase is mediated by caveolae. InvestOphthalmol Vis Sci 1999; 40: 157–167.

28 Lakshminarayanan S, Antonetti DA, Gardner TW, Tarbell JM.Effect of VEGF on retinal microvascular endothelial hydraulicconductivity: the role of NO. Invest Ophthalmol Vis Sci 2000; 41:4256–4261.

29 Tilton RG, Chang KC, LeJeune WS, Stephan CC, Brock TA,Williamson JR. Role for nitric oxide in the hyperpermeabilityand hemodynamic changes induced by intravenous VEGF. InvestOphthalmol Vis Sci 1999; 40: 689–696.

30 Ghiso N, Rohan RM, Amano S, Garland R, Adamis AP. Suppressionof hypoxia-associated vascular endothelial growth factor geneexpression by nitric oxide via cGMP. Invest Ophthalmol Vis Sci1999; 40: 1033–1039.

31 Shen BQ, Lee DY, Zioncheck TF. Vascular endothelial growth factorgoverns endothelial nitric-oxide synthase expression via a KDR/Flk-1 receptor and a protein kinase C signaling pathway. J Biol Chem1999; 274: 33057–33063.

32 Folkman J. Angiogenesis in cancer, vascular, rheumatoid and otherdisease. Nat Med 1995; 1: 27–31.

33 Tuder RM, Flook BE, Voelkel NF. Increased gene expression forVEGF and the VEGF receptors KDR/Flk and Flt in lungs exposed toacute or to chronic hypoxia. Modulation of gene expression by nitricoxide. J Clin Invest 1995; 95: 1798–1807.

DME_768.fm Page 660 Tuesday, July 23, 2002 1:53 PM