Transforming Growth Factor β2 in the Vitreous in Proliferative Diabetic Retinopathy FREE

PROLIFERATIVE diabetic retinopathy (PDR) is a representative proliferative disorder of the eye that causes vitreous hemorrhage (VH) and tractional retinal detachment (TRD) due to fibrovascular proliferation. The involvement of various cytokines has been suggested in the pathogenesis of this condition,^{1 - 3} and growth factors such as basic fibroblast growth factor,⁴ insulinlike growth factor,⁵ and vascular endothelial growth factor^{6 - 7} have been found at high concentrations in the vitreous of patients with PDR.

Transforming growth factor β (TGF- β), a polypeptide originally named on the basis of its ability to transform fibroblasts, is now well known as a multifunctional growth factor that either stimulates or inhibits cell proliferation and differentiation, depending on the cell type or environment.^{8 - 11} This factor can also cause an accumulation of extracellular matrix. Transforming growth factor β is secreted as an inactive latent complex that must be transformed into a mature form to become active. In vitro, the latent complex can be activated by exposure to acidic or alkaline environments or by heat.¹² At present, however, the mechanism for activation in vivo is not fully understood.

In humans, there are 3 known TGF- β isoforms (TGF- β₁, - β₂, and - β₃) that localize in the anteroposterior segment of the human eye.^{13 - 16} Of the 3, TGF- β₂ is considered to be the predominant form in the posterior segment of the eye.^{17 - 18} There have been several reports of the presence of TGF- β₂ in the human vitreous.^{18 - 19} Connor et al¹⁸ found that TGF- β₂ levels were increased in proliferative vitreoretinopathy. The amount of increase correlated with the severity of proliferative vitreoretinopathy, suggesting that TGF- β₂ has a role in the formation of the proliferative membrane in this disorder. Based on the finding that TGF- β₂ levels were reduced in the ocular fluid of patients with uveitis, de Bore et al¹⁹ suggested a relation between TGF- β₂ and immunosuppressive effects.

Panretinal photocoagulation is effective in the treatment of retinal neovascularization. The mechanism of its efficacy, however, has not been completely elucidated. Recently, there have been several reports^{20 - 22} on the relation between TGF- β₂ production and laser photocoagulation. These studies reported increased levels of TGF- β₂ both in vitro²¹ and in vivo (in the rabbit vitreous)²⁰ after photocoagulation, suggesting that TGF- β₂ may play a part in the efficacy of retinal photocoagulation for the treatment of retinal neovascularization.

In the present study, we quantified TGF- β₂ levels in vitreous fluid from patients with PDR. Furthermore, we investigated whether TGF- β₂ levels correlated with the extent of preoperative laser retinal photocoagulation or the severity of fibrous proliferation.

The study was approved by the Human Research Committee of the Kyoto Prefectural University of Medicine, Kyoto, Japan. Vitreous fluid was obtained from 49 eyes of 48 patients with PDR (49 specimens) at the time of therapeutic vitrectomy, and 19 patients who underwent nondiabetic vitrectomies (19 specimens) served as controls. The specimens were collected with proper informed consent. The diabetic subjects comprised 20 men and 28 women whose mean (±SD) age was 54.3 (±11.9) years. The duration from the initial diagnosis of diabetes mellitus to surgical procedure varied from 0.3 to 30 years, with a mean of 12.4 years. The glycosylated hemoglobin value, measured within 1 month of the operation, was 0.08±0.02% (7.7±1.9%). Of these 48 patients, 27 (28 specimens) had been receiving insulin treatment as their glycemic management, and 21 (21 specimens) had not. Of the 49 specimens, 21 had TRD combined with VH, 19 had TRD without VH, 6 had VH without TRD, and 3 had neither VH nor TRD. Of these PDR specimens, 5 patients had rubeotic glaucoma, and 9 had "burned-out" retinopathy, whose neovascular activity is inactive. The 19 control subjects comprised 8 men and 11 women, whose mean (±SD) age was 60.2 (±12.3) years; 9 had an idiopathic macular hole, 6 had retinal detachment due to a macular hole, and 4 had rhegmatogenous retinal detachment. No proliferative changes were found in the retinal detachments. Table 1 details the patients in the group with PDR and the control group.

Before intraocular infusion, the vitreous cutter was used to remove much of the vitreous, which was collected undiluted and stored in polypropylene tubes at − 80°C until assay.

Each PDR specimen was classified into 2 groups by the same surgeon (T.I.) at the time of vitrectomy. One classification was based on the extent of retinal photocoagulation before vitrectomy. These groups included: no retinal photocoagulation (NRP) (10 specimens), focal retinal photocoagulation (FRP) (14 specimens), panretinal photocoagulation (PRP) (17 specimens), and supplemental photocoagulation after the initial PRP (sPRP) (8 specimens). Another classification was made, depending on the degree of fibrous proliferation, as follows: no apparent fibrous proliferation or fibrous proliferation around the disc only (FPD) (25 specimens), fibrous proliferation around the vascular arcade (FPA) (18 specimens), and fibrous proliferation throughout the posterior pole (FPP) (6 specimens).

Transforming growth factor β₂ levels were assessed with a specific-capture enzyme-linked immunosorbent assay kit (Amersham Life Science, Buckinghamshire, England). This immunoassay, using the quantitative "sandwich" enzyme immunoassay technique, detects only mature TGF- β₂; latent TGF- β₂ is not recognized by the antibodies in this kit. The 96-well plates provided were coated with murine antibody to human TGF- β₂. After 50 µL of assay diluent was pipetted into each well, 200 µL of specimen dilutions or standards were incubated for 2 hours at room temperature. After washing, enzyme-linked polyclonal anti–TGF- β₂ was added for detection and incubated for 2 hours at room temperature. After further washing, the substrated solution was added, and the reaction was stopped with sulfuric acid, 1 mol/L. The optical density was read at the 450-nm wavelength. Recombinant human TGF- β₂, provided in the kit, was used as a standard. The standard curve showed a linear response in the range of 31.3 to 2000.0 pg/mL, using logarithm/logarithm paper. The TGF- β₂ concentration was calculated by comparing specimen absorption with that of serial dilutions of human recombinant TGF- β₂.

Before the assay, vitreous specimens were diluted 1:2 with the standard diluent provided in the kit. The assay detection limit was defined as 2 pg/mL. Given the specimen dilution, the lowest detectable level was 4 pg/mL. Most TGF- β₂ found in intraocular fluid has been identified as the latent form.¹⁸ To assay total TGF- β₂ (mature form plus the latent form), specimens had to be activated before the assay. Thus, to assay total TGF- β₂, half the volume of each vitreous specimen was activated by acidification with a solution of hydrochloride, 150 mmol/L, for 30 minutes at room temperature, followed by sodium hydroxide neutralization before the assay.¹⁸ In our study, each specimen, therefore, was assayed both with and without activation.

Data were statistically analyzed with the Mann-Whitney U test for comparison of the 2 groups and with the Kruskal-Wallis test for multiple groups. For statistical calculations, specimens with concentrations below the assay detection limit were entered as 4.0 pg/mL. The results were declared as significant at P <.05. Data are given as mean (±SD).

Mature TGF- β₂ was detected in 46 of the 49 PDR specimens and in 16 of the 19 control specimens. After acid activation, TGF- β₂ was detected in all specimens. The total TGF- β₂ level in controls was 1305 (±972) pg/mL. The levels of TGF- β₂ in controls were similar to those previously reported in vitreous specimens from eye bank eyes.¹⁹ In patients with PDR, the total level of TGF- β₂ was 2634 (±1652) pg/mL. This was significantly greater than in the controls (P<.001) (Figure 1, left panel). The level of mature TGF- β₂ was 79(±81) pg/mL in controls and 244 (±316) pg/mL in patients with PDR, a significant difference (P<.01) (Figure 1, right panel). The total and mature TGF- β₂ levels in the PDR group were analyzed with multiple regression by age, sex, duration of diabetes, extent of photocoagulation, and the glycosylated hemoglobin value. No significant correlations were found between the TGF- β₂ levels and these factors. There was no correlation between TGF- β₂ levels and the extent of VH or TRD. No difference was found in glycemic management. The mature-total ratio was 0.15 in the PDR group and 0.14 in the control group; not a significant difference. The levels of total and mature TGF- β₂ of the 5 specimens that had rubeotic glaucoma were 2538 (±1719) and 533 (±451) pg/mL, respectively, and the levels in the 9 specimens that had burned-out retinopathy were 1428 (±984) and 179 (±108) pg/mL, respectively. Mature TGF- β₂ levels tended to be higher in the group that had rubeotic glaucoma than in the group that had burned-out retinopathy; however, the difference was not significant (P=.26 for both).

According to the extent of retinal photocoagulation, the total TGF- β₂ levels were 2276 (±1353) pg/mL in the NRP group, 2638 (±1632) pg/mL in the FRP group, 2704 (±1905) pg/mL in the PRP group, and 2818 (±1776) pg/mL in the sPRP group (Figure 2, left panel). The respective mature TGF- β₂ levels were 485 (±524), 278 (±289), 124 (±112), and 139 (±125) pg/mL (Figure 2, right panel). There was no statistical difference among the 4 groups (mature TGF- β₂, P=.06; total TGF- β₂, P=.42), although there was a tendency toward a high level of mature TGF- β₂ in the NRP group preoperatively. The mature-total TGF- β₂ ratios were 0.24 (±0.27) in the NRP group, 0.19 (±0.26) in the FRP group, 0.09 (±0.14) in the PRP group, and 0.11 (±0.17) in the sPRP group. No statistical difference was found among the 4 groups, but there was a tendency toward a higher ratio in the NRP group (P=.07).

Grouping by the degree of fibrous proliferation revealed the following: the total TGF- β₂ levels were 2440 (±1426) pg/mL in the FPD group, 3084 (±1919) pg/mL in the FPA group, and 2091 (±1626) pg/mL in the FPP group (Figure 3, left panel). No significant difference among the 3 groups was detected (P=.41). The respective mature TGF- β₂ levels were 133 (±106), 202 (±202), and 832 (±533) pg/mL (Figure 3, right panel). These 3 groups were statistically different in the mature TGF- β₂ levels (P<.01). When the mature TGF- β₂ levels of the 4 groups were analyzed independently with each of the other groups, the levels in the FPP group were greater than in the other groups (FPD group and FPA group, P<.01 for both).

In the present study, we measured TGF- β₂ levels in the vitreous fluid collected at the time of vitreous surgery from patients with PDR and other disorders (the control group). The total TGF- β₂ level in patients with PDR was approximately 2 times greater and the mature TGF- β₂ level was 3 times greater than in the controls. These findings are of interest in the study of PDR because TGF- β₂ is known to be a growth factor with diverse functions, including those related to the formation of fibrous proliferative membranes and neovascularization.

The total TGF- β₂ levels in the vitreous from patients with PDR was quantified by Boulton et al.²³ In their study, total TGF- β₂ levels in diabetic vitreous was not different from those in nondiabetic controls, but vitreous from the patients receiving insulin treatment contained TGF- β₂ at relatively high concentrations when the diabetic group was subdivided according to glycemic management. This result did not agree with our findings. With regard to glycemic management, the dose of insulin administered to each patient may be related.

The total TGF- β₂ levels in the vitreous of patients with proliferative vitreoretinopathy has been reported to correlate with the severity of fibrosis.¹⁸ Similarly, in our study of patients with PDR, greater intraocular fibrosis was associated with increased levels of mature TGF- β₂. It seems that the process of TGF- β₂ synthesis may be enhanced in patients with PDR, although a reduction in the degradation of this cytokine cannot be excluded. The mechanism by which TGF- β₂ may enhance fibrous proliferation is unclear, but it is likely to involve its chemotactic effects on monocytes and fibroblasts and its effects on the production of the extracellular matrix.⁸

There have been several reports on the relation between TGF- β₂ production and photocoagulation. Previous studies have reported that TGF- β₂ levels were increased in rabbit vitreous and in the culture medium of retinal pigment epithelial cells after photocoagulation.^{20 - 21} Yoshimura et al²² reported that photocoagulated retinal pigment epithelial cells in vitro produced an inhibitor of endothelial cell proliferation and that the character of the inhibitor corresponded to that of TGF- β₂. These studies suggest that TGF- β₂ is involved in the suppression of neovascularization after retinal photocoagulation. According to these studies, the increase in TGF- β₂ production after photocoagulation was considered to reach a fairly high level. The present study, however, showed no clear correlation between the extent of preoperative retinal photocoagulation and TGF- β₂ concentration. The TGF- β₂ level was measured early after photocoagulation in other studies, but in our clinical specimens, the time after photocoagulation varied widely, possibly resulting in inconsistencies in the findings. The increase in TGF- β₂ concentration in the vitreous in patients with PDR in the present study may be a result of, not only the photocoagulation undergone preoperatively, but other factors also.

In this study, TGF- β₂ levels were shown to increase in the vitreous of patients with PDR. This cytokine is thought to promote the pathogenic process of PDR by inducing the formation of proliferative membranes. The specific actions of TGF- β₂ on neovascularization remain largely unknown; however, they are likely to be complex because the effects of this growth factor vary depending on its concentration, its interaction with the extracellular matrix, and the presence of other cytokines. Further investigations of these and other aspects of the mechanisms of action of TGF- β₂ are needed to better understand the role of this molecule in PDR.

Accepted for publication February 19, 1998.

This study was supported in part by a grant-in-aid for scientific research 08672031 from the Ministry of Education, Culture, and Science of Japan, Tokyo, a research grant from Kyoto Foundation for the Promotion of Medical Science, and an intramural research fund of Kyoto Prefectural University of Medicine, Kyoto, Japan.

Reprints: Tsunehiko Ikeda, MD, PhD, Department of Ophthalmology, Kyoto Prefectural University of Medicine, Hirokoji Kawaramachi Kamigyo-ku, Kyoto 602-0841, Japan (e-mail: tikeda@eyeeye.ophth.kpu-m.ac.jp).