Vascular Adhesion Molecules in Vitreous from Eyes with Proliferative Diabetic Retinopathy

Abstract

purpose. To investigate whether proliferative vitreoretinopathy (PDR) is associated with a selective increase in vitreous levels of soluble vascular cell adhesion molecules that mediate leukocyte extravasation and interaction with endothelium during processes of inflammation and neovascularization.

methods. Vitreous from 55 patients undergoing vitrectomy for treatment of PDR complicated by vitreous hemorrhage and/or traction retinal detachment was assayed for the presence of the soluble vascular cell adhesion molecules sICAM-1, sVCAM-1, and sE-selectin using a standard enzyme-linked immunosorbent assays (ELISA). Vitreous from 12 cadaveric eyes matching age and sex of the patients were used as control samples.

results. Vitreous levels of sICAM-1, sVCAM-1, and sE-selectin were significantly higher in eyes with PDR than in control cadaveric vitreous, and levels of all three molecules did not relate to the type or duration of diabetes mellitus. However, eyes with either traction retinal detachment alone or both traction retinal detachment and vitreous hemorrhage exhibited significantly higher levels of sICAM-1 and sE-selectin than eyes with vitreous hemorrhage alone. Vitreous levels of sVCAM-1 were similar in eyes with either vitreous hemorrhage or traction retinal detachment alone.

conclusions. The present observations suggest that molecular inflammatory mechanisms may contribute to processes of neovascularization and fibrosis observed in PDR, possibly not as the causative event, but as a result of endothelial, Müller, and retinal pigment epithelial cell activation. The results also indicate that retinal detachment amplifies the existing inflammation within the diabetic retina. Identification of any abnormalities in the production and control of specific adhesion molecules could have important implications in the design of new therapeutic regimens to treat and prevent this sight-threatening complication of diabetes mellitus.

Proliferative diabetic retinopathy (PDR) is a common complication of diabetes mellitus, characterized by active neovascularization of the optic disc and the retina, with formation of fibrovascular tissue at the vitreoretinal interface. ¹ Neovascularization is thought to be induced by retinal ischemia caused by capillary occlusion, in which platelets and leukocytes may play an important role. Although the idea that inflammation may be involved in the development of insulin-dependent diabetes mellitus (IDDM) is controversial, ² to our knowledge the concept that inflammation may play a role in the pathogenesis of PDR has not been addressed in the literature. However, all the molecular mechanisms implicated in new vessel formation and fibrosis in PDR ^{3 4} are those that characterize inflammation in general. ⁵ Therefore, leukocytes often found within PDR membranes ⁶ may well have migrated into the retina by the same mechanisms by which they would normally migrate into sites of inflammation. 

Inflammation is initiated by activation of endothelial cells by cytokines, which results in their enhanced expression and shedding of vascular cell adhesion molecules. ⁷ Upregulation of vascular cell adhesion molecules is critical for leukocyte migration through endothelial cell junctions into the abluminal surface of the vessels. ^{7 8} Rolling and adhesion of leukocytes to vascular endothelium, which are the initial steps for leukocyte extravasation, are mediated by the adhesion molecules P-selectin, E-selectin, and L-selectin, ^{7 8} whereas more firm leukocyte–endothelial interactions and cell migration are governed by intercellular adhesion molecule (ICAM)-1 and the vascular cell adhesion molecule (VCAM)-1. ^{7 8}  

Extracellular domains of vascular cell adhesion molecules are often found as soluble (s) forms in serum and body fluids, after cleavage by metalloproteinases, ⁹ and both E-selectin and VCAM-1 are recognized to be angiogenic in vivo and in vitro. ¹⁰ It is possible that these molecules play a role in the pathogenesis of PDR. Supranormal serum levels of sICAM-1, sVCAM-1, and sE-selectin may be found in patients with chronic inflammatory ¹¹ and ocular disease ¹² and in patients with IDDM, particularly those with retinopathy. ^{13 14} In addition, high vitreous levels of sICAM-1 are observed in eyes with anterior uveitis ¹⁵ and in eyes with proliferative vitreoretinopathy, ¹⁶ where they constitute a marker of inflammation severity and a risk factor for development of this complication of retinal detachment. 

In view of the above evidence and of the potential angiogenic role of vascular cell adhesion molecules, we investigated whether high vitreous levels of sICAM-1, sVCAM-1, and sE-selectin may be found in vitreous from eyes with PDR and whether they relate to retinal complications of this condition, such as vitreous hemorrhage and traction retinal detachment. 

Methods

Vitreous samples were obtained from 55 patients undergoing vitreoretinal surgery for the treatment of vitreous hemorrhage alone (23 patients), traction retinal detachment alone (21 patients), or both vitreous hemorrhage and traction retinal detachment (11 patients) complicating PDR. Of the patients investigated, 31 had IDDM and 17 had non-IDDM. Forty-one patients had diabetes of more than 20 years’ duration (mean, 22 ± 7.3 years), and 16 patients had diabetes of less than 10 years’ duration (mean, 6 ± 2.6 years). All patients had undergone laser photocoagulation for treatment of PDR. Cadaveric vitreous samples obtained within 6 to 18 hours after death from donors with no known ocular or systemic inflammatory disease, and matching age and sex of the patients, were used as normal control specimens. Undiluted vitreous samples (approximately 0.75 ml) were centrifuged for 5 minutes at 600g to remove contaminating cells and then transferred to cryotubes for storing at −70°C until use. The study was approved by the ethics committee of St. Thomas’ Hospital, and it was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. 

Determination of sICAM-1, sVCAM-1, and sE-Selectin

Levels of vitreous sICAM-1, sVCAM-1, and sE-selectin were determined using commercial enzyme-link immunosorbent assay (ELISA) kits (R&D Systems, Oxon, UK) as follows: Microtiter well plates coated with specific antibodies to individual adhesion molecules were incubated with 100 μl of a 1:10 dilution of vitreous, together with 100 μl of the respective anti-adhesion molecule antibody, after which antibodies and test samples were removed and the plate washed six times with phosphate-buffered saline (PBS) containing 0.05% Tween-20. The amount of conjugated antibodies was detected by addition of 100 μl tetramethylbenzidine (substrate) and incubation for 30 minutes at room temperature. The enzymatic reaction was stopped by addition of 100 μl of 1 M H₂SO₄ and the absorbance read at 450 nm, with a correction wavelength of 620 nm (model MR 5000; Dynatech, Cambridge, MA). Levels of specific adhesion molecules present in vitreous samples were interpolated from specific calibration curves prepared with standard reagents. 

Statistical Analysis of the Results

The significance of difference between corresponding groups of observations was evaluated by the nonparametric Mann–Whitney test. Acceptable significance was achieved at P < 0.05. 

Results

Vitreous Levels of sICAM-1, sVCAM-1, and sE-Selectin in Eyes with PDR

Figure 1 shows that vitreous from eyes with PDR contained significantly higher levels of sICAM-1 (range, 6.0–65.0 ng/ml) than vitreous from cadaveric control eyes (range, 4.17–9.81 ng/ml; P = 0.00008). Vitreous sVCAM-1 levels were also higher in eyes with PDR (range, 5.2–81.6 ng/ml) than in control cadaveric eyes (range, 2.8–14 ng/ml; P = 0.00001). In addition, vitreous levels of sE-selectin were significantly higher (range, 0.3–3.93 ng/ml) in eyes with PDR than in cadaveric control eyes (range, 0.1–0.59 ng/ml; P = 0.000005). There were no differences in the levels of these molecules between vitreous from patients with IDDM and vitreous from patients with non-IDDM, nor was there any relationship between vitreous levels of these molecules and duration of diabetes mellitus (Table 1) . 

Vitreous Levels of sICAM-1 in Relation to Retinal Complications of PDR

As observed in Figure 2 , breakdown of the diabetic group into the various retinal complications of PDR showed that all groups contained significantly higher vitreous levels of sICAM-1 (P < 0.00001) when compared with cadaveric control samples. In addition, vitreous from eyes with PDR complicated by traction retinal detachment contained significantly higher levels of sICAM-1 (range, 6.04–35.6 ng/ml) than vitreous from eyes with vitreous hemorrhage alone (range, 8.7–22 ng/ml; P = 0.017). Likewise, vitreous levels of sICAM-1 in PDR eyes with both vitreous hemorrhage and traction retinal detachment exhibited higher levels of sICAM-1 (range, 15.7–65 ng/ml) than eyes with vitreous hemorrhage alone (P = 0.0021). There were no differences between the vitreous levels of this adhesion molecule in eyes with traction retinal detachment alone and eyes with both traction retinal detachment and vitreous hemorrhage (P = 0.15). 

Vitreous Levels of sVCAM-1 in Relation to Retinal Complications of PDR

Figure 3 shows that vitreous levels of sVCAM-1 in eyes with PDR complicated by vitreous hemorrhage, traction retinal detachment, or both were significantly higher (P < 0.001) than in cadaveric eyes. Vitreous levels of sVCAM-1 in eyes with vitreous hemorrhage alone (range, 5.2–32.6 ng/ml) did not differ from those observed in eyes with PDR complicated by traction retinal detachment alone (range, 5.2–38 ng/ml; P = 0.94). In contrast, vitreous levels of this molecule were significantly higher in eyes with both vitreous hemorrhage and traction retinal detachment (range, 10.8–71.4 ng/ml) than in eyes with either vitreous hemorrhage (P = 0.012) or traction retinal detachment alone (P = 0.007). 

Vitreous Levels of sE-Selectin in Relation to Retinal Complications of PDR

Figure 4 shows that similar to that seen with sICAM-1 and sVCAM-1, breakdown of the diabetic group into the various retinal complications of PDR showed that all groups contained significantly higher vitreous levels of sE-selectin than cadaveric control samples (P < 0.0008). Vitreous levels of sE-selectin in eyes with PDR complicated by traction retinal detachment (range, 0.2–2.9 ng/ml) were significantly higher than in eyes with vitreous hemorrhage alone (range, 0.22–1.19 ng/ml; P = 0.021). Similarly, vitreous from eyes with PDR complicated by both traction retinal detachment and vitreous hemorrhage exhibited higher levels of this molecule (range, 0.5–2.71 ng/ml) than eyes with vitreous hemorrhage alone (P = 0.021). There were no differences between the vitreous levels of this molecule in eyes with traction retinal detachment alone or traction retinal detachment together with vitreous hemorrhage (P = 0.7). 

Discussion

In this study, we investigated whether PDR as a complication of IDDM or non-IDDM was associated with raised vitreous levels of soluble vascular cell adhesion molecules that mediate leukocyte extravasation and interaction with endothelium during the inflammatory process. We observed that vitreous levels of sICAM-1, sVCAM-1, and sE-selectin were significantly higher in eyes with PDR than in control cadaveric vitreous, and that levels of these molecules did not relate to the type or duration of diabetes mellitus. Assessment of clinical complications of PDR leading to vitrectomy showed that eyes with either traction retinal detachment alone or with both traction retinal detachment and vitreous hemorrhage exhibited significantly higher levels of sICAM-1 and sE-selectin than those with vitreous hemorrhage alone. These findings suggest that inflammation caused by retinal detachment further amplifies the existing inflammatory process that leads to neovascularization in the diabetic eye. This is supported by previous observations that retinal detachment also amplifies inflammation in eyes with anterior uveitis. ¹⁵ The observations that vitreous levels of sICAM-1 and sE-selectin were higher in eyes with traction retinal detachment than in those with vitreous hemorrhage alone and that sVCAM-1 levels in eyes with vitreous hemorrhage were similar to those in eyes with traction retinal detachment suggest that these molecules may be locally produced within the retinal environment and that their levels in the vitreous may not depend on extravasation from the circulation. 

Expression of ICAM-1 has been demonstrated in vivo and in vitro in various cells of the retina and choroid and in leukocytes. ^{8 17 18} Presence of both molecules has been observed in proliferating vascular endothelium of PDR membranes, ¹⁹ and ICAM-1, normally expressed by RPE cells, ¹⁸ may also be found in the extracellular matrix of these membranes. ¹⁹ These observations support the view that sICAM-1 found in vitreous from eyes with PDR may derive from local retinal cells, including retinal vascular endothelium and RPE cells, as well as from leukocytes migrating into the retina. Expression of the vascular adhesion molecules ICAM-1, VCAM-1, and E-selectin on endothelium is crucial for leukocyte recruitment into the inflammatory site, ^{7 8} and cytokines such as tumor necrosis factor-α, which is found in vitreous ²⁰ and in the extracellular matrix and luminal and abluminal surfaces of vessels in PDR membranes, ¹⁹ promote the upregulation and release of these molecules during the inflammatory process. ²¹ It is therefore possible that release of soluble adhesion molecules into the vitreous from eyes with PDR may be promoted by this cytokine, which is also known to play an important role in the pathogenesis of diabetes mellitus and in the development of proliferative retinopathy complicating this disease. ^{22 23}  

General features that characterize the inflammatory process in general ⁵ are those that promote neovascularization and fibrosis within the diabetic retina, ^{3 4} suggesting that cellular and molecular mechanisms of inflammation may operate during the development of PDR. Although there is no evidence to suggest that inflammation may be the main trigger of fibrovascular proliferation in PDR, there is evidence that cell adhesion and angiogenesis are strongly linked. ¹⁰ Evidence for the angiogenic properties of VCAM-1 and E-selectin derives from observations that both molecules induce angiogenesis of rat cornea and chemotaxis of human endothelial cells. ^{10 14} On this basis, it is reasonable to suggest that these molecules may contribute to the angiogenic process observed in PDR, and that the severity of this condition may well be related to levels of production and release of vascular cell adhesion molecules. In turn, adhesion molecule production within the diabetic retina may depend on the profile of cytokine production induced by either retinal hypoxia, as suggested by experimental findings in vitro, ²⁴ or by abnormal glucose metabolites, as indicated by observations that methyl glyoxal–modified proteins, which are highly increased in poorly controlled diabetes, ²⁵ induce expression of mRNA coding for tumor necrosis factor-α. ²⁶  

Further investigation of the mechanisms that promote and control the production and activity of vascular cell adhesion molecules within the eye may aid in a better understanding of the microvascular process leading to PDR and in the development of new therapeutic approaches to treat and prevent this sight-threatening complication of diabetes mellitus. 

 Supported by Gift of Thomas Pocklington Trust and the Guide Dogs for the Blind Association.

 Submitted for publication January 21, 1999; revised March 30, 1999; accepted April 20, 1999.

 Proprietary interest category: N.

 Corresponding author: G. Astrid Limb, Department of Pathology, Institute of Ophthalmology, Bath Street, London EC1V 9EL UK.

 

Figure 1.

 

Histogram showing the mean levels ± SEM of soluble vascular cell adhesion molecules in vitreous from eyes with proliferative diabetic retinopathy. Comparison with vitreous levels of vascular cell adhesion molecules in normal cadaveric vitreous. Mann–Whitney test:* P = 0.00008 (versus control cadaveric vitreous),** P = 0.00005 (versus control cadaveric vitreous),*** P = 0.000005 (versus control cadaveric vitreous).

Table 1.

 

Vitreous Levels of Vascular Adhesion Molecules in Relation to Type and Duration of Diabetes Mellitus

	sICAM-1 (ng/ml)	sVCAM-1 (ng/ml)	sE-Selectin (ng/ml)
Type of diabetes mellitus
IDDM	22.4 (6.0–65.0) [37]	21.4 (5.2–71.3) [30]	0.76 (0.36–4.3) [39]
Non-IDDM	18.0 (8.7–34.2) [11]	19.0 (5.2–93.6) [12]	0.79 (0.2–3.9) [11]
Duration of diabetes
More than 10 years	20.3 (6.0–64.9) [34]	21.7 (0.22–81.6) [33]	0.72 (0.2–3.9) [38]
Less than 10 years	21.3 (9.7–34.2) [9]	16.4 (5.2–93.6) [10]	0.8 (0.35–4.3) [13]

 The figures indicate the median values. Numbers in parentheses indicate the range of values within each group; numbers in brackets indicate the number of samples examined for each adhesion molecule.

Figure 2.

 

Scatter diagram showing the vitreous levels of sICAM-1 in eyes with PDR complicated by vitreous hemorrhage (VH) alone, traction retinal detachment (TRD) alone, or both. Mann–Whitney test:* P < 0.00001 (versus control cadaveric vitreous);** P < 0.00001 (versus control cadaveric vitreous), P = 0.017 (versus VH alone);*** P < 0.00001 (versus control cadaveric vitreous), P = 0.0021 (versus VH alone). The lines represent the median values.

Figure 3.

 

Scatter diagram showing the vitreous levels of sVCAM-1 in eyes with PDR complicated by vitreous hemorrhage (VH) alone, traction retinal detachment (TRD) alone, or both. Mann–Whitney test:* P < 0.001 (versus control cadaveric vitreous);** P < 0.001 (versus control cadaveric vitreous), P = 0.94 (versus VH alone);*** P < 0.001 (versus control cadaveric vitreous), P = 0.012 (versus VH alone), P = 0.007 (versus TRD alone). The lines represent the median values.

Figure 4.

 

Scatter diagram showing the vitreous levels of sE-selectin in eyes with PDR complicated by vitreous hemorrhage (VH) alone, traction retinal detachment (TRD) alone, or both. Mann–Whitney test:* P < 0.0008 (versus control cadaveric vitreous);** P < 0.0008 (versus control cadaveric vitreous), P = 0.021 (versus VH alone);*** P < 0.0008 (versus control cadaveric vitreous), P = 0.021 (versus VH alone), P = 0.81 (versus TRD alone). The lines represent the median values.