Deficit of Somatostatin-Like Immunoreactivity in the Vitreous Fluid of Diabetic Patients: Possible role in the development of proliferative diabetic retinopathy

The study included 14 consecutive diabetic patients with PDR (6 type 1 and 8 type 2 diabetic patients) in whom a vitrectomy was performed. Sixteen nondiabetic patients with other conditions requiring vitrectomy, but in whom the retina was not directly affected by neovascularization, served as a control group. Both groups were matched by age and sex. In the control group, the diagnoses included macular hole (n = 8), rhegmatogenous retinal detachment (n = 4), and epiretinal membrane (n = 4). Both venous blood and vitreous samples were collected at the time of vitreoretinal surgery. Patients with renal failure (creatinine ≥120 μmol/l), who had undergone previous vitreoretinal surgery, had recently had a vitreous hemorrhage (<2 months), or had received photocoagulation in the previous 3 months were all excluded.

In all cases, a classic three-port pars plana vitrectomy was performed. For visualization of the vitreous cavity, we used a wide-field system with a precorneal Volk lens of 130° and inversion image system Moeller-Wedel (Hamburg, Germany). During vitrectomy, partial posterior vitreous detachment (PVD) was observed in 11 of 14 diabetic patients and in almost all nondiabetic subjects of the control group (6 of 8 macular holes and in all cases with either rhegmatogenous retinal detachment or epiretinal membranes).

Undiluted vitreous samples (0.5–1 ml) were obtained at the onset of vitrectomy by aspiration into a 1-ml syringe attached to the vitreous cutter (Ten-Thousand Outcome; Alcon, Irvine, CA) before starting intravitreal infusion of balanced salt solution. The vitreous samples were transferred to a sterile tube containing EDTA (0.054 ml, 0.34 mol/l) and aprotinin (500 KIU/ml), placed immediately on ice, and centrifuged at 16,000g for 5 min at 4°C. Supernatants were frozen at −80°C until assayed.

During vitrectomy, blood samples were collected in chilled tubes containing EDTA (0.054 ml, 0.34 mol/l) and aprotinin (500 KIU/ml) and then centrifuged immediately at 3,000g for 10 min at 4°C. The plasma obtained was aliquoted and stored at −80°C until assayed. The protocol for sample collection was approved by the hospital ethics committee, and informed consent was obtained from patients.

SLI was measured by radioimmunoassay (Euro-Diagnostica, Malmö, Sweden). Before assay, somatostatin was extracted using Vycor, a leached silica glass, as previously described (26). The extraction recovery was similar for plasma and vitreous samples (75–85%). ¹²⁵I-thyrosine somatostatin was used as a tracer together with a highly specific rabbit antisomatostatin serum (initial dilution 1:30,000). The antigenic site is directed toward the central part of the molecule containing the tryptophan residue. The antiserum shows no cross-reaction with a wide range of hypothalamic, pancreatic, gastrointestinal, or pituitary hormones. This antibody cross-reacts with cyclic somatostatin (100%), Tyr1-somatostatin (100%), linear somatostatin (50%), and des-ala-gly-somatostatin (25%). The lower detection limit was 10 pg/ml. The intra-assay coefficient of variation was 8.3% for values ∼27 pg/ml and 2.8% for concentrations ∼94 pg/ml. All samples were diluted in parallel to the cyclic somatostatin standard. Validation of the assay has previously been reported in detail (26).

Vitreal proteins were measured by a previously validated microturbidimetric method with an autoanalyzer (Hitachi 917; Boehringer, Mannheim, Germany). This method, based on the benzetonium chloride reaction, is a highly specific method for the detection of proteins and has a higher sensibility and reproducibility than the classic method of Lowry. The lowest level of proteins detected was 0.02 mg/ml. Intra- and interassay coefficients of variation were 2.9 and 3.7%, respectively.

Apart from excluding patients with recent vitreous hemorrhage, we also excluded those in which intravitreous hemoglobin was detected. For this purpose, vitreous hemoglobin levels were measured by spectrophotometry (Uvikon 860; Kontron Instruments, Zürich, Switzerland) using the classic method of Harboe for measuring plasma hemoglobin in micromolar concentration (27). This method has been further validated (28), and in our studies, the lowest limit of detection was 0.03 mg/ml.

Values of SLI were compared using Student’s t test. However, because of their skewed distribution, the statistical comparisons of the ratios (vitreous SLI/plasma SLI and vitreous SLI/intravitreal proteins) were performed by means of a nonparametric test (Mann-Whitney U test). The Spearman rank correlation coefficient was used to examine correlations. Levels of statistical significance were set at P < 0.05. All statistical analyses were performed with the Statistical Package for the Social Sciences (SPSS-PC). The results are expressed either as the means ± SE or median and range.

The main clinical characteristics of patients with PDR and nondiabetic control subjects are summarized in Table 1. SLI in the vitreous fluid was significantly lower in diabetic patients than in the control group (68 ± 18.7 vs. 193.6 ± 30.8 pg/ml, P < 0.01). By contrast, plasma SLI was higher in diabetic patients than in nondiabetic control subjects (88.2 ± 11.5 vs. 43.5 ± 10.7 pg/ml, P < 0.01) (Fig. 1.).

We did not observe significant differences in SLI levels between type 1 and type 2 diabetic patients in either plasma (75.7 ± 18 vs. 100.6 ± 17 pg/ml, P = 0.34) or the vitreous fluid (96.5 ± 24 vs. 52.6 ± 28 pg/ml, P = 0.25). In addition, diabetic patients who received major laser treatment (n = 9) presented similar SLI intravitreous concentration to those in whom retinal photocoagulation was not performed (n = 5; 62.55 ± 22.50 vs. 77.90 ± 36.51 pg/ml, P = 0.71).

Because vitreous SLI levels may in part reflect those in the plasma, we calculated the ratio of vitreous to plasma SLI concentrations in each patient. The vitreous SLI-to-plasma SLI ratio was strikingly higher in nondiabetic subjects than in diabetic patients with PDR (P < 0.01) (Fig. 2). This result was similar when the control group was compared separately with either type 1 (5.3 [1.2–71.1] vs. 1.0 [0.5–4.1], P < 0.01) or type 2 diabetic patients (5.3 [1.2–71.1] vs. 0.14 [0.03–1.94], P < 0.001).

We detected higher intravitreous protein concentrations in diabetic patients with PDR than in the control group (2.6 mg/ml [1.1–14.5] vs. 0.9 mg/ml [0.4–2.3], P < 0.0001). Therefore, after correcting for total vitreous protein concentration (vitreal SLI [pg/ml]-to-vitreal proteins [mg/ml] ratio), SLI (pg/mg of proteins) remained significantly higher in nondiabetic control subjects than in diabetic patients with PDR (186 [51–463] vs. 7.5 [0.8–82], P < 0.0001).

Remarkably, intravitreous levels of SST were significantly higher than those obtained in plasma in nondiabetic control subjects (193.6 ± 30.8 vs. 43.5 ± 10.7 pg/ml, P < 0.0001). By contrast, in diabetic patients, we observed higher levels of SLI in plasma than in the vitreous fluid, but in this case, the differences did not reach statistical significance (88.2 ± 11.5 vs. 68 ± 18.7 pg/ml, P = 0.59).

Finally, a lack of relationship between plasma and vitreous levels of SST was observed in both diabetic patients with PDR (r = 0.18, P = 0.55) and nondiabetic control subjects (r = −0.09, P = 0.72).

Vitreous fluid obtained from diabetic patients undergoing vitreoretinal surgery is currently used for indirectly exploring the synthesis of several peptides by the retina. However, we are unaware of previous reports in which SLI was determined in the vitreous fluid. In the present study, after considering the main confounding factors that could lead to misinterpretation of the results (vitreous hemorrhage, intravitreal protein concentration, and plasma SLI levels), we have provided evidence that there is a deficit of somatostatin in the vitreous fluid of diabetic patients with PDR. This finding was observed in absolute terms and was even overstated when either plasma levels of SLI or intravitreal protein concentrations were considered. It could be speculated that the high prevalence of PVD among diabetic patients with PDR could prevent the passage of somatostatin within the vitreous. However, the similar prevalence of PVD detected in both diabetic patients and nondiabetic control subjects makes its potential influence on the results highly unlikely. Another potential cause of the deficit of SLI found in the vitreous fluid of diabetic patients could have been previous photocoagulation. In this regard, it should be emphasized that we excluded all the patients who had received laser treatment in the preceding 3 months. In addition, we did not find any significant difference in vitreous SLI concentration between diabetic patients receiving laser treatment and those in whom retinal photocoagulation was not performed. Therefore, the influence of previous photocoagulation, at least in the group of patients included in the present study, would appear to be negligible.

Remarkably, higher SLI concentrations in the vitreous fluid than in plasma were detected in nondiabetic control subjects. It must be emphasized that the intravitreous protein concentration is at least 20-fold less than in serum (29,30). Thus, the higher intravitreal concentration of a particular protein in relation to its plasma levels strongly supports its intraocular production. Moreover, SLI levels found in the vitreous fluid of nondiabetic subjects included in our study were even higher than those observed by several authors in the cerebrospinal fluid of normal healthy control subjects (31–33). These findings suggest that somatostatin exerts an important role in retinal homeostasis. In this regard, there is growing evidence that in the retina, somatostatin acts as a neuromodulator through multiple pathways, including intracellular Ca²⁺ signaling (34), nitric oxide function (35), and glutamate release from photoreceptors (36). In addition, the loss of somatostatin immunoreacitivy was found after degeneration of the ganglion cells (37). Therefore, the neuroretinal damage that occurs in diabetic retinopathy might be the reason for the decreased SLI detected in the vitreous fluid of these patients. Similarly, levels of SLI have been consistently decreased in the cerebrospinal fluid of patients with various neurodegenerative diseases (33,38–41).

The observation that GH secretion is reduced by somatostatin analogs has been the basis for the clinical trials in severe PDR (5–7,9,10). Animal models also support the role of somatostatin analogs in the suppression of GH and IGF-I production and the subsequent inhibition of retinal neovascularization (25,42). However, somatostatin may reduce endothelial cell proliferation and neovascularization by multiple mechanisms, including inhibition of postreceptor signaling events of peptide growth factors such as IGF-I, vascular endothelial growth factor, epidermal growth factor, and platelet-derived growth factor (11). In addition, both SSTR2- and SSTR3-selective analogs directly inhibit retinal endothelial cell growth in vitro (12,43). Therefore, it appears that the inhibitory effect of somatostatin on retinal endothelial cell proliferation can be achieved independently of the modulation of systemic GH and IGF-I levels. In recent years, several natural inhibitors of angiogenesis that counterbalance the inducers of angiogenesis have been identified. The major natural inhibitor in the vitreous and cornea of the eye is pigment epithelium-derived factor (PEDF), a protein produced by retinal pigment epithelial cells and found in high concentrations within the retina as well as in the vitreous, where it is responsible for the antiangiogenic activity of this fluid (44). Furthermore, a deficit of PEDF has been found in the vitreous fluid of patients with PDR (45,46), thus suggesting that the loss of angigogenic inhibitors has a central role in mediating the angiogenic response of retinal ischemia, such as that seen in PDR. The strikingly higher SLI concentration observed in nondiabetic control subjects lead us to propose somatostatin as a good candidate to be added to these natural inhibitors of angiogenesis. In addition, it is possible that the decreased concentration of SLI observed in the vitreous fluid of diabetic patients might contribute to the process of retinal neovascularization. However, because SLI vitreous concentration does not necessary imply somatostatin activity, future studies exploring the binding to SSTRs are needed.

In summary, we have detected significantly higher SLI in the vitreous fluid than in the plasma in nondiabetic control subjects. This raises the possibility that SLI could be relevant for retinal homeostasis. In addition, we have observed a significant deficit of SLI in the vitreous fluid of diabetic patients. Obviously, further studies are needed to establish the cause of this finding as well as the potential role in the etiopathogenesis of diabetic retinopathy.

This work was supported by grants from the Ministerio de Sanidad y Consumo (FIS 98/1270), the Ministerio de Ciencia y Tecnología (PM 99-0136), and Novo Nordisk Pharma SA (01/0066).

We thank Dr. F. Campos for the analysis of vitreous hemoglobin, Dr. Magela Garat and Dr. Ana Cantón for their contribution to the vitreous collection and processing of samples, and Michael Willy for his assistance with manuscript preparation.