Cytokines in the Vitreous of Patients With Proliferative Diabetic Retinopathy Ahmed M. Abu El Asrar, M.D., Davide Maimone, M.D., Peter H. Morse, M.D., Susan Gregory, Ph.D., and Anthony T. Reder, M.D.
Sixteen vitreous and paired serum samples from 13 patients with proliferative diabetic retinopathy, vitreous samples from seven ca daveric control subjects, and aqueous humor samples from 15 normal control subjects were assayed for the cytokines interleukin-1, tumor necrosis factor-a, interleukin-6, and interfer o n e · Interleukin-6 was detected in 15 of 16 vitreous samples (94%) from diabetic patients, but it was not detected in any of the aqueous humor samples. Vitreous interleukin-6 levels positively correlated with ocular disease ac tivity. Interleukin-1 was detected in seven of 16 vitreous samples (44%) and in four of ten aqueous humor samples (40%), whereas tumor necrosis factor-a and interferon-γ were never detected in vitreous or aqueous fluid. Serum samples from diabetic patients and control subjects contained comparable low levels of interleukin-6. Interleukin-1, tumor necrosis factor-a, and interferon-γ were not found in any of the sera. Because interleukin-6 can function as B-cell differentiation factor, this cytokine may have a role in immunoglobulin deposition in the ocular tissues and in the immunopathologic characteristics of prolifer ative retinopathy.
Accepted for publication Sept. 14, 1992. From the Department of Ophthalmology, Mansoura University Hospital, Mansoura, Egypt (Dr. Abu El Asrar); and the Departments of Neurology and Brain Research Institute (Drs. Maimone and Reder), Ophthal mology and the Visual Sciences Center (Dr. Morse), and Medicine (Dr. Gregory), The University of Chicago, Chicago, Illinois. This study was supported in part by the Peace Fellowship Programme for Egypt, sponsored by the Agency for International Development, Washing ton, DC; a Harry Weaver Neuroscience Award from the National Multiple Sclerosis Society, New York, New York; and grant 1 K08 NS 01068-01 from the National Institutes of Health-Public Health Service (Dr. Reder). Reprint requests to Peter H. Morse, M.D., Visual Sciences Center, 939 E. 57th St., Chicago, IL 606371454.
©AMERICAN JOURNAL OF OPHTHALMOLOGY
114:731-736,
1 HE PATHOGENESis of proliferative retinopathy is incompletely understood, but immunologie mechanisms may have a role in the develop ment of the characteristic microangiopathy. 1 Ocular deposits of immunoglobulin and com plement have been detected in patients with diabetic retinopathy. 24 The preretinal mem branes in individuals with diabetes contain activated B lymphocytes, T cells, and monocytes, and class II major histocompatibility complex antigen expression is found at the level of the pigmented epithelium. 3 ' 4 Addition ally, it is our clinical impression that some eyes with proliferative diabetic retinopathy, espe cially in type I diabetics, show clinical signs of inflammation such as injected, irritated eyes, and cells and flare in the anterior chamber. Cytokines are polypeptides produced by im mune cells that regulate immune responses and inflammatory reactions. Some cytokines could trigger or maintain proliferative retinopathy. Among relevant cytokines, interleukin-lß and tumor necrosis factor-a have angiogenic ac tivity, 5,6 stimulate glial cell proliferation, 7,8 and induce fibroblast proliferation and colla gen synthesis. 9 Interleukin-6 induces the final differentiation of B cells to antibody-produc ing cells.10 Interferon-^ is a potent inducer of class II major histocompatibility complex anti gens. 1113 Several studies in mice suggest that the dysregulation of cytokine production or the production of abnormal cytokines could be re sponsible for the hyperactivation of autoantibody-producing B cells. 1416 Intraocular secre tion of cytokines could explain ocular deposits of immunoglobulin in patients with prolifera tive diabetic retinopathy. To explore immunologie mechanisms in the pathogenesis of proliferative retinopathy, we examined the vitreous and serum from patients with proliferative diabetic retinopathy for the presence of interleukin-lß, tumor necrosis factor-ct, interleukin-6, and interferon-ã.
DECEMBER,
1992
731
732
AMERICAN JOURNAL OF OPHTHALMOLOGY
Subjects and Methods Subjects—Vitreous and paired serum samples were obtained from 13 patients with proliferative diabetic retinopathy who underwent vitrectomy for nonclearing vitreous hemorrhage (13 eyes) or traction macular detachment (three eyes). In three patients, a second vitreous and serum sample was obtained intraoperatively on the fellow eye. The patients included eight men and five women whose ages ranged from 24 to 88 years, with an average of 51 years. The patients had a duration of diabetes from nine to 30 years, with an average of 20 years, and all were insulin dependent. The preoperative and operative findings were recorded and the clinical disease severity was graded from 0 to 4 according to the presence and extent of active fibrovascular tissue, vitre ous hemorrhage, iris neovascularization (rubeosis iridis), and hyphema. Grade 0 denoted inactive residual fibrous tissue from proliferative diabetic retinopathy; grade 1+ denoted vitreous hemorrhage without active neovascu larization; grade 2+ denoted vitreous hemor rhage with actively growing fibrovascular tis sue; grade 3+ denoted iris neovascularization in addition to the findings in grade 2 + ; and grade 4+ denoted hyphema in addition to the findings in grade 3 + . Two patients had renal transplants and were taking immunosuppressive drugs. The first pa tient was being treated with prednisone, cyclosporine, and azathioprine (1+ disease activity; the interleukin-6 level was 1.84 pg/ml). The second patient was being treated with predni sone and cyclosporine (3+ disease activity; the interleukin-6 level was 15 pg/ml). Vitreous samples were also obtained after death in seven patients with no history of intraocular disease, ocular trauma, or sepsis. Vitreous was harvested four to seven hours post mortem. One additional control donor under went a vitrectomy during repair of a macular hole. Aqueous samples from 15 nondiabetic sub jects who had no history of ocular inflammation were obtained at the time of cataract extraction. These samples were used as a control for the normal intraocular cytokine levels. The aque ous fluid was used as a control sample because removal of vitreous from normal eyes is contraindicated. Furthermore, there is communica tion and exchange between the fluid content in the anterior chamber and the posterior segment
December, 1992
of the eye, especially in elderly patients with liquified vitreous. In preliminary experiments, interleukin-6 was found at marked levels in the vitreous of seven nondiabetic cadaveric eyes within three to eight hours of death. Because of the variability of fresh donor vitreous, aqueous was used as a control. The samples were centrifuged, and the supernatants were frozen at —70 C until they were assayed. Interleukin-6 bioassay—An interleukin-6-dependent mouse hybridoma cell line, B9, was used (L. A. Aarden, M.D., Central Laboratory of Netherlands Red Cross Blood Transfusion Service, University of Amsterdam, The Nether lands). 16 After one wash, cells were resuspended in Iscove's modified Dulbecco's medi um supplemented with 5% fetal bovine serum, 5 x 10_B mol/1 mercaptoethanol, penicillin, and streptomycin, and seeded in 96-well flatbottomed microwell plates at a density of 5,000 cells/200 ìÀ/well. Vitreous samples were as sayed without further treatment. All sera were heat-inactivated at 56 C for 30 minutes to pre vent a nonspecific toxic effect to the B9 cells. Serial dilutions of each sample were added in triplicate to the wells. Proliferation was evalu ated after 72 hours by measuring the amount of radioactivity incorporated after a four-hour pulse with 1 ì è of [ 3H] thymidine. Sample values were always compared to a standard curve generated with recombinant human in terleukin-6 (S. Gillis, M.D., Immunex, Inc., Seattle, Washington) or natural human inter leukin-6 (Endogen, Inc., Boston, Massachu setts). No inhibitory activity on B9-cell prolifer ation was observed in the vitreous or serum after the addition of various amounts of recom binant human interleukin-6. The detection lim it in vitreous and serum was 1.0 p g / m l . Neutralization with anti-interleukin-6 antiserum—To assess the specificity of the interleu kin-6 activity in our vitreous samples, we per formed neutralization tests with a commercially available rabbit antiserum specific for human interleukin-6 (Endogen, Inc.). Aliquots of vitre ous were incubated at 37 C for three hours with various concentrations of rabbit anti-interleu kin-6 or equivalent doses of preimmmune rab bit serum (Endogen, Inc.). Duplicate aliquots (100 ìÀ each) of neutralized specimens were assayed for their ability to induce proliferation of the B9 cells as described. Tumor necrosis factor-a enzyme-linked immunosorbent assay (ELISA)—Tumor necrosis fac tor-a measurements were performed with an
Proliferative Diabetic Retinopathy
Vol. 114, No. 6
ELISA kit (Cistron Biotechnology, Pine Brook, New Jersey). Undiluted vitreous and sera were tested in duplicate by adding 100 ìÀ/well. A standard curve was run in each assay with predetermined amounts of recombinant human tumor necrosis factor-a (range, 20 to 1,000 pg/ml). This method showed a detection limit of 20 p g / m l of tumor necrosis factor-a. The standard curve was not altered by dilution of recombinant human tumor necrosis factor-a in pooled normal human sera. Interleukin-lß ELISA—Interleukin-lß was quantitated with an ELISA kit (Cistron Biotech nology). The assay was used according to the manufacturer's instructions. Duplicate aliquots of 100 ìÀ of undiluted vitreous and serum were tested. The standard curve obtained with recombinant human interleukin-lß ranged from 20 to 1,000 p g / m l . The detection limit was 20 pg/ml. The standard curve was not altered by dilution of recombinant interleukin-lß in pooled normal human sera. Interferon-^ ELISA—Interferon-^ was quanti tated with an ELISA kit (Amgen Biological, Thousand Oaks, California). The assay was used according to the manufacturer's instruc tions. Duplicate aliquots of 100 ìÀ of undiluted vitreous and serum were tested. The detection limit was 6.25 U / m l . The standard curve ranged from 6.25 to 50 U / m l . Statistical analysis—Data were analyzed by using one-tailed paired or unpaired f-tests, Mann-Whitney U tests, and linear regression analysis as appropriate. A P value less than .05 was considered significant. Values were ex pressed as mean ± standard error of the mean.
Results Cytokines in ocular fluids—Interleukin-6 was detected in 15 of 16 vitreous samples (94%) from patients with proliferative diabetic reti nopathy. The levels of detectable interleukin-6 in the vitreous samples ranged from 1.14 to 190 pg/ml ± (43.3 ± 15.7 p g / m l ; mean ± standard error of the mean) (Fig. 1). Interleukin-6 was not detected in any of the aqueous samples from 15 normal control subjects (vitreous sam ples vs aqueous samples, P < .005; MannWhitney U test). In vitreous fluid from control subjects (seven post mortem), interleukin-6 levels ranged from 0 to 49.5 pg/ml (9.80 ± 6.05 pg/ml). Interleu kin-6 values in these control subjects were
733
r
Ö 100
=
so
•i **«»-
•JMW^
Vitreous Serum Fig. 1 (Abu El Asrar and associates). Detectable interleukin-6 (IL-6) levels in vitreous and paired serum samples from 16 eyes. One vitreous sample and eight serum samples had interleukin-6 levels less than 1 pg/ml. lower than in patients with proliferative diabet ic retinopathy (P < .03, unpaired f-test). Interleukin-6 levels in the vitreous samples of patients with proliferative diabetic retinopa thy were positively correlated with the grade of clinical ocular disease severity (Fig. 2) (R = 0.677, P < .004; linear regression analysis). The two diabetic patients with renal transplants had interleukin-6 levels of 1.84 p g / m l (disease activity, 1+) and 15 p g / m l (disease activity, 3 + ). The presence of specific interleukin-6 activity in vitreous samples was confirmed by using a rabbit anti-human interleukin-6 in neutraliza tion tests. The proliferative response of B9 cells to interleukin-6 from vitreous samples was completely blocked by the rabbit antiserum. Interleukin-lß protein was detected in seven of 16 vitreous samples (44%), with levels rang ing from 21 to 68 p g / m l . Interleukin-lß was present in four of ten aqueous samples, with levels ranging from 36 to 59 p g / m l . Tumor necrosis factor-a and interferon--y were not detected in any of the vitreous sam ples from patients with proliferative diabetic retinopathy or aqueous samples from control subjects. Cytokines in the serum—Interleukin-6 was detected in eight of the 16 serum samples (50%). The serum interleukin-6 level ranged from 1.1 to 21.4 p g / m l (Fig. 1). These levels were not different from those found in the serum of normal control samples assayed in our laboratory. Interleukin-6 levels in the vitreous samples were higher than those in paired sera (P < .05; paired f-test). Interleukin-lß, tumor necrosis factor-á, and
734
AMERICAN JOURNAL OF OPHTHALMOLOGY
Disease Activity
Fig. 2 (Abu El Asrar and associates). Correlation of interleukin-6 (IL-6) levels in vitreous samples with clinical disease activity in 16 eyes from patients with proliferative diabetic retinopathy. interferon-ã were not detected in any of the sera.
Discussion In our study, interleukin-6 was found in 15 of 16 vitreous samples (94%) from patients with proliferative diabetic retinopathy. Interleu kin-6 was not present in aqueous samples from normal control subjects and was also signifi cantly lower in postmortem vitreous samples from subjects without intraocular disease. In postmortem eyes, the delay between death and harvesting of vitreous would be expected to increase vitreous protein, 17 including interleu kin-6, which implies that the interleukin-6 found in fresh vitreous in proliferative diabetic retinopathy is biologically important. It is pos sible that compartments such as the vitreous act as a reservoir for cytokines, thus amplifying local effects. The levels of interleukin-6 correlate with clinical disease severity in eyes as graded by our criteria (Fig. 2). An increase of interleu kin-6 was not detected in paired serum sam ples, suggesting that interleukin-6 in the vitre ous is intraocular in origin and designates proliferative retinopathy. The inference that interleukin-6 synthesis is local would theoreti cally be confirmed by Northern blot analysis of retinal tissue, but retinal or iris biopsy is not performed in patients undergoing vitrectomy for proliferative diabetic retinopathy. Interleukin-6 can be produced by various cell types, including monocytes, activated T cells,18
December, 1992
vascular endothelial cells, fibroblasts, 19 and glial cells.20 These cells could be the source of intraocular interleukin-6 in patients with pro liferative retinopathy caused by diabetes and possibly by sickle cell disease, retinal detach ment, and retinal vein occlusion. Interleukin-6, also called B-cell stimulating factor-2, may have an important role in anti body-mediated diseases, including those in re stricted compartments of the body. Interleu kin-6 is present in the aqueous humor of 63% of patients with uveitis. 21 This cytokine is also detected in the synovial fluid of patients with active rheumatoid arthritis,22"24 in which inter leukin-6 levels correlate with disease activity. 24 Several tumors, such as cardiac myxomas and cervical carcinomas, produce interleukin-6. Pa tients with such tumors have hypergammaglobulinemia and multiple autoantibodies that dis appear from the serum after surgical removal of the tumor. 25 Unregulated production of inter leukin-6 might be responsible for autoantibody production in these patients. The interleukin-6 we detected in the vitreous from patients with proliferative diabetic reti nopathy could be involved in deposits of immunoglobulins in the walls of newly formed preretinal vessels. 2 Biopsies of the pars plana in proliferative diabetic retinopathy demonstrate immunoglobulin deposits at the level of the pigmented epithelium 3 and the preretinal mem branes show extensive immunoglobulin depos its in the connective stroma and within the vascular walls. 4 In addition to increased intra ocular interleukin-6 and ocular deposits of im munoglobulin, patients with type I diabetes mellitus have autoantibodies that are directed against a variety of pancreatic and nonpancreatic antigens. Moreover, the frequency of other autoimmune diseases is higher in diabetic pa tients than in the normal population. 26 We have also found interleukin-6 in the vitre ous from four patients with proliferative vitreoretinopathy (13.7 ± 4.7 pg/ml) (Reder, A. T., Abu El Asrar, A. M., and Maimone, D., unpub lished data, 1991). In this disease also, immu noglobulin and complement deposits are seen at the level of the pigmented epithelium. 27 How ever, the disease mechanism differs from pro liferative diabetic retinopathy, for there is metaplasia of the retinal pigment epithelium without vascular proliferation. A pathogenetic role could be envisaged in proliferative retinopathy for tumor necrosis fac tor-a, possibly with a synergistic effect from interleukin-lß. 2 8 Tumor necrosis factor-a and interieukin-Iß induce angiogenesis in ocular
Vol. 114, No. 6
Proliferative Diabetic Retinopathy
interleukin-lß induce angiogenesis in ocular tissues5·6 and proliferation of glial cells7,8 and fibroblasts. We did not detect tumor necrosis factor-a in any of the vitreous samples we examined. Tumor necrosis factor-a clears rapid ly from the circulation. 29 Similar rapid clearance or degradation of proteins in the eye could have hampered the detection of this cytokine in our study. 30 Interleukin-lß was detected in seven of 16 vitreous samples from patients with prolifera tive diabetic retinopathy. Interleukin-lß, how ever, was also present in four of ten aqueous humor samples from normal control subjects. Normal aqueous humor contains a remarkably consistent spectrum of proteins, including immunoglobulin, albumin, and lower molecular weight proteins. 17 The pattern does not reflect serum proteins. The interleukin-1 in the aque ous and vitreous may have been synthesized within these compartments, for serum interleu kin-1 was undetectable. Finally, surgical trau ma to the iris markedly increases protein levels in the aqueous. 17 Our samples were obtained with exquisite care and trauma was unlikely. It is also possible that interleukin-6 down-regu lates production of tumor necrosis factor-a and to some extent interleukin-lß in the eyes of patients with proliferative diabetic retinopa thy.31 In previous immunopathologic studies, class II major histocompatibility complex antigen expression was noted in the pigmented epithe lium, 3 as well as on many cells in the preretinal membranes 4 from patients with proliferative di abetic retinopathy. Interferon-ã is a potent inducer of class II antigen expression on a variety of epithelial cells in vitro, 1113 and activates retinal pigmented epithelial cells, which allows them to present antigens. 32 Interferon--y is de tected in the vitreous of patients with acquired immunodeficiency syndrome retinitis 33 and sympathetic ophthalmia 84 and could amplify proliferative retinopathy. In our study, howev er, we failed to detect interferon-ã in the vitre ous samples. We cannot exclude the possibility that small amounts of interferon-ã, not detect able with our ELISA system at the time of sampling, were present in the vitreous from patients with proliferative diabetic retinopa thy. In this regard, the synovial fluid of patients with rheumatoid arthritis contains little or no interferon-ã, despite class II major histocom patibility complex antigen expression by the synovial cells.35 This suggests that factors other than interferon--y may be responsible for class II
735
major histocompatibility complex antigen ex pression in localized compartments. We found that detectable levels of interleu kin-6 in vitreous samples from patients with proliferative diabetic retinopathy, and interleu kin-6 levels correlate with clinically detectable ocular disease activity. The presence of inter leukin-6 may account for the well-established finding of immunoglobulin deposition in ocular tissues. Our data suggest that an inflammatory reaction may further exacerbate disease severi ty in eyes of patients with proliferative retinop athy. Therefore, the use of anti-inflammatory drugs or cytokine antagonists before, during, and after surgical intervention warrants further evaluation.
References 1. Rahi, A. H. S., and Addison, D. J.: Autoimmunity and the outer retina. Trans. Ophthalmol. Soc. U.K. 103:428, 1983. 2. Melato, M., Antonutto, G., Manconi, R., and Ponte, E.: Ocular deposits of immunoglobulin in diabetic retinopathy. Can. J. Ophthalmol. 17:45, 1982. 3. Baudouin, C, Fredj-Reygrobellet, D., Lapalus, P., and Gastaud, P.: Immunohistopathologic findings in proliferative diabetic retinopathy. Am. J. Ophthal mol. 105:383, 1988. 4. Baudouin, C, Gordon, W. C , Fredj-Reygrobel let, D., Baudouin, F., Peyman, G. ]., Gastaud, P., and Bazan, N. G.: Class II antigen expression in diabetic preretinal membranes. Am. J. Ophthalmol. 109:70, 1990. 5. Ben Ezra, D., Hemo, I., and Maftzir, G.: In vivo angiogenic activity of interleukins. Arch. Ophthal mol. 108:573, 1990. 6. Frater-Schroder, M., Risau, W., Hallmann, R., Gautschi, P., and Bohlen, P.: Tumor necrosis factor type a, a potent inhibitor of endothelial cell growth in vitro, is angiogenic in vivo. Proc. Nati. Acad. Sci. U.S.A. 84:5722, 1987. 7. Guilan, D., and Lachman, L. B.: Interleukin-1 stimulation of astroglial proliferation after brain in jury. Science 220:497, 1985. 8. Selmaj, K. W., Farooq, M., Morton, W. T., Raine, C. S., and Brown, C. F.: Proliferation of astrocytes in vitro in response to cytokines. A primary role for tumor necrosis factor. J. Immunol. 144:129, 1990. 9. Kohase, M., May, L. T., Tamm, I., Vilcek, J., and Sehgal, P. B.: A cytokine network in human diploid fibroblasts. Interactions of b-interferons, tumor ne crosis factor, platelet-derived growth factor, and in terleukin-1. Mol. Cell. Biol. 7:273, 1987. 10. Hirano, T., Taga, T., Nakano, N., Yasukawa, K., Kashiwamura, S., Shimizu, K., Nakajima, K., Pyun,
736
AMERICAN JOURNAL OF OPHTHALMOLOGY
K. H., and Kishimoto, T.: Purification to homogenei ty and characterization of human B-cell differentia tion factor (BCDF or BSF p-2). Proc. Nati. Acad. Sci. U.S.A. 82:5490, 1985. 11. Basham, T. Y., Nickoloff, B. J., Merigan, T. C , and Morhenn, V. B.: Recombinant y interferon differ entially regulates class II antigen expression and biosynthesis on cultured normal human keratinocytes. J. Interferon Res. 5:23, 1985. 12. El-Asrar, A. M., Van Den Gord, J. J., Billiau, A., Desmet, V., Emarah, M. H., and Missotten, L.: Recombinant interferon-7 induces HLA-DR expression on human corneal epithelial and endothelial cells in vitro. A preliminary report. Br. J. Ophthalmol. 73:587, 1987. 13. Tabibzadeh, S. S., Gerber, M. A., and Satyaswaroop, P. G.: Induction of HLA-DR antigen expres sion in human endometrial epithelial cells in vitro by recombinant-interferon. Am. J. Pathol. 125:90, 1986. 14. Preud'homme, G. J., Fieser, T. M., Dixon, F. ]., and Theofilopoulos, A. N.: B cell-tropic interleukins in murine systemic lupus erythematosus (SLE) 1. Immunol. Rev. 78:159, 1984. 15. Sidman, C. L., Marshall, J. D., Masiello, N. C., Roths, ]. B., and Shultz, L. D.: Novel B-cell matura tion factor from spontaneously autoimmune viable motheaten mice. Proc. Nati. Acad. Sci. U.S.A. 81:7199, 1984. 16. Aarden, L., DeGroot, E. R., Schaap, O. L., and Lansdorp, P. M.: Production of hybridoma growth factor by human monocytes. Eur. J. Immunol. 17:1411, 1987. 17. Tripathi, R. C , Millard, C. B., and Tripathi, B. J.: Protein composition of human aqueous humor. SDS-PAGE analysis of surgical and post-mortem samples. Exp. Eye Res. 48:117, 1989. 18. Houssiau, F. A., Coulie, P. G., and Van Snick, J.: Distinct roles of IL-I and IL-6 in human T-cell activation. J. Immunol. 143:2520, 1989. 19. Mizel, S. B.: The interleukins. FASEB J. 3:2379, 1989. 20. Frei, K., Malipiero, U. V., Leist, T. P., Zinkernagel, R. M., Schwab, M. E., and Fontana, A.: On the cellular source and function of interleukin 6 pro duced in the central nervous system in viral diseases. Eur. J. Immunol. 19:689, 1989. 21. Murray, P. I., Hoekzema, R., van Haren, M. A. C , de Hon, F. D., and Kijlstra, A.: Aqueous humor interleukin-6 levels in uveitis. Invest. Oph thalmol. Vis. Sci. 31:917, 1990. 22. Hirano, T., Matsuda, T., Turner, M., Miyasaka, N., Buchan, G., Tang, B., Sato, K., Shimizu, M., Maini, R., Feldmann, M., and Kishimoto, T.: Exces sive production of interleukin 6/B cell stimulating factor-2 in rheumatoid arthritis. Eur. J. Immunol. 18:1797, 1988. 23. Bhardwaj, N., Santhanam, U., Lau, L. L., Tat ter, S. B., Ghrayeb, J., Rivelis, M., Steinman, R. M.,
December, 1992
Sehgal, P. B., and May, L. T.: IL-6/IFN-ß 2 in synovial effusions of patients with rheumatoid arthritis and other arthritides. Identification of several isoforms and studies of cellular sources. J. Immunol. 143:2153, 1989. 24. Waage, A., Kaufmann, C , Espevik, T., and Husby, G.: Interleukin-6 in synovial fluid from pa tients with arthritis. Clin. Immunol. Immunopathol. 50:394, 1989. 25. Hirano, T., Taga, T., Yasukawa, K., Nakajima, K., Nakano, N., Takatsuki, F., Shimizu, M., Murashima, A., Tsunasawa, S., Sakiyama, F., and Kishimoto, T.: Human B-cell differentiation factor defined by an antipeptide antibody and its possible role in autoantibody production. Proc. Nati. Acad. Sci. U.S.A. 84:228, 1987. 26. Drell, D. W., and Notkins, A. L.: Multiple immunological abnormalities in patients with type I (insulin-dependent) diabetes mellitus. Diabetologia 30:132, 1987. 27. Baudouin, C , Fredj-Reygrobellet, D., Bau douin, F., Lapalus, P., and Gastaud, P.: Immunohistologic study of proliferative vitreoretinopathy. Am. J. Ophthalmol. 108:387, 1989. 28. Dinarello, C. A.: The biology of interleukin 1 and comparison to tumor necrosis factor (correspon dence). Immunology 16:227, 1987. 29. Michie, H. R., Manague, K. R., Spriggs, D. R., Revhaug, A., O'Dwyer, A., Dinarello, C. A., Cerami, A., Wolff, S. M., and Wilmore, D. W.: Detection of circulating tumor necrosis factor after endotoxin ad ministration. N. Engl. J. Med. 318:1481, 1988. 30. Tripathi, B. J., Geanon, J. D., and Tripathi, R. C : Distribution of tissue plasminogen activator in human and monkey eyes. Ophthalmology 94:1434, 1987. 31. Schindler, R., Mandila, J., Endres, S., Ghorbani, R., Clark, S. C , and Dinarello, C. A.: Correla tion and interactions in the production of interleu kin-6 (IL-6), IL-1, and tumor necrosis factor (TNF) in human blood mononuclear cells. IL-6 suppresses IL-1 and TNF. Blood 75:40, 1990. 32. Percopo, C. M., Hooks, J. ]., Shinohara, T., Caspi, R., and Detrick, B.: Cytokine-mediated activa tion of a neuronal retinal resident cell provokes antigen presentation. J. Immunol. 145:4101, 1990. 33. Mondino, B. J., Sidikaro, Y., Mayer, F. J., and Sumner, H. L.: Inflammatory mediators in the vitre ous humor of AIDS patients with retinitis. Invest. Ophthalmol. Vis. Sci. 31:798, 1990. 34. Hooks, J. J., Chan, C. C , and Derrick, B.: Iden tification of the lymphokines, interferon-ã and interIeukin-2, in inflammatory eye diseases. Invest. Oph thalmol. Vis. Sci. 29:144, 1988. 35. Firestein, G. S., and Zavifler, N. J.: Peripheral blood and synovial fluid monocyte activation in in flammatory arthritis. Arthritis Rheum. 30:864, 1987.
Sixteen vitreous and paired serum samples from 13 patients with proliferative diabetic retinopathy, vitreous samples from seven ca daveric control subjects, and aqueous humor samples from 15 normal control subjects were assayed for the cytokines interleukin-1, tumor necrosis factor-a, interleukin-6, and interfer o n e · Interleukin-6 was detected in 15 of 16 vitreous samples (94%) from diabetic patients, but it was not detected in any of the aqueous humor samples. Vitreous interleukin-6 levels positively correlated with ocular disease ac tivity. Interleukin-1 was detected in seven of 16 vitreous samples (44%) and in four of ten aqueous humor samples (40%), whereas tumor necrosis factor-a and interferon-γ were never detected in vitreous or aqueous fluid. Serum samples from diabetic patients and control subjects contained comparable low levels of interleukin-6. Interleukin-1, tumor necrosis factor-a, and interferon-γ were not found in any of the sera. Because interleukin-6 can function as B-cell differentiation factor, this cytokine may have a role in immunoglobulin deposition in the ocular tissues and in the immunopathologic characteristics of prolifer ative retinopathy.
Accepted for publication Sept. 14, 1992. From the Department of Ophthalmology, Mansoura University Hospital, Mansoura, Egypt (Dr. Abu El Asrar); and the Departments of Neurology and Brain Research Institute (Drs. Maimone and Reder), Ophthal mology and the Visual Sciences Center (Dr. Morse), and Medicine (Dr. Gregory), The University of Chicago, Chicago, Illinois. This study was supported in part by the Peace Fellowship Programme for Egypt, sponsored by the Agency for International Development, Washing ton, DC; a Harry Weaver Neuroscience Award from the National Multiple Sclerosis Society, New York, New York; and grant 1 K08 NS 01068-01 from the National Institutes of Health-Public Health Service (Dr. Reder). Reprint requests to Peter H. Morse, M.D., Visual Sciences Center, 939 E. 57th St., Chicago, IL 606371454.
©AMERICAN JOURNAL OF OPHTHALMOLOGY
114:731-736,
1 HE PATHOGENESis of proliferative retinopathy is incompletely understood, but immunologie mechanisms may have a role in the develop ment of the characteristic microangiopathy. 1 Ocular deposits of immunoglobulin and com plement have been detected in patients with diabetic retinopathy. 24 The preretinal mem branes in individuals with diabetes contain activated B lymphocytes, T cells, and monocytes, and class II major histocompatibility complex antigen expression is found at the level of the pigmented epithelium. 3 ' 4 Addition ally, it is our clinical impression that some eyes with proliferative diabetic retinopathy, espe cially in type I diabetics, show clinical signs of inflammation such as injected, irritated eyes, and cells and flare in the anterior chamber. Cytokines are polypeptides produced by im mune cells that regulate immune responses and inflammatory reactions. Some cytokines could trigger or maintain proliferative retinopathy. Among relevant cytokines, interleukin-lß and tumor necrosis factor-a have angiogenic ac tivity, 5,6 stimulate glial cell proliferation, 7,8 and induce fibroblast proliferation and colla gen synthesis. 9 Interleukin-6 induces the final differentiation of B cells to antibody-produc ing cells.10 Interferon-^ is a potent inducer of class II major histocompatibility complex anti gens. 1113 Several studies in mice suggest that the dysregulation of cytokine production or the production of abnormal cytokines could be re sponsible for the hyperactivation of autoantibody-producing B cells. 1416 Intraocular secre tion of cytokines could explain ocular deposits of immunoglobulin in patients with prolifera tive diabetic retinopathy. To explore immunologie mechanisms in the pathogenesis of proliferative retinopathy, we examined the vitreous and serum from patients with proliferative diabetic retinopathy for the presence of interleukin-lß, tumor necrosis factor-ct, interleukin-6, and interferon-ã.
DECEMBER,
1992
731
732
AMERICAN JOURNAL OF OPHTHALMOLOGY
Subjects and Methods Subjects—Vitreous and paired serum samples were obtained from 13 patients with proliferative diabetic retinopathy who underwent vitrectomy for nonclearing vitreous hemorrhage (13 eyes) or traction macular detachment (three eyes). In three patients, a second vitreous and serum sample was obtained intraoperatively on the fellow eye. The patients included eight men and five women whose ages ranged from 24 to 88 years, with an average of 51 years. The patients had a duration of diabetes from nine to 30 years, with an average of 20 years, and all were insulin dependent. The preoperative and operative findings were recorded and the clinical disease severity was graded from 0 to 4 according to the presence and extent of active fibrovascular tissue, vitre ous hemorrhage, iris neovascularization (rubeosis iridis), and hyphema. Grade 0 denoted inactive residual fibrous tissue from proliferative diabetic retinopathy; grade 1+ denoted vitreous hemorrhage without active neovascu larization; grade 2+ denoted vitreous hemor rhage with actively growing fibrovascular tis sue; grade 3+ denoted iris neovascularization in addition to the findings in grade 2 + ; and grade 4+ denoted hyphema in addition to the findings in grade 3 + . Two patients had renal transplants and were taking immunosuppressive drugs. The first pa tient was being treated with prednisone, cyclosporine, and azathioprine (1+ disease activity; the interleukin-6 level was 1.84 pg/ml). The second patient was being treated with predni sone and cyclosporine (3+ disease activity; the interleukin-6 level was 15 pg/ml). Vitreous samples were also obtained after death in seven patients with no history of intraocular disease, ocular trauma, or sepsis. Vitreous was harvested four to seven hours post mortem. One additional control donor under went a vitrectomy during repair of a macular hole. Aqueous samples from 15 nondiabetic sub jects who had no history of ocular inflammation were obtained at the time of cataract extraction. These samples were used as a control for the normal intraocular cytokine levels. The aque ous fluid was used as a control sample because removal of vitreous from normal eyes is contraindicated. Furthermore, there is communica tion and exchange between the fluid content in the anterior chamber and the posterior segment
December, 1992
of the eye, especially in elderly patients with liquified vitreous. In preliminary experiments, interleukin-6 was found at marked levels in the vitreous of seven nondiabetic cadaveric eyes within three to eight hours of death. Because of the variability of fresh donor vitreous, aqueous was used as a control. The samples were centrifuged, and the supernatants were frozen at —70 C until they were assayed. Interleukin-6 bioassay—An interleukin-6-dependent mouse hybridoma cell line, B9, was used (L. A. Aarden, M.D., Central Laboratory of Netherlands Red Cross Blood Transfusion Service, University of Amsterdam, The Nether lands). 16 After one wash, cells were resuspended in Iscove's modified Dulbecco's medi um supplemented with 5% fetal bovine serum, 5 x 10_B mol/1 mercaptoethanol, penicillin, and streptomycin, and seeded in 96-well flatbottomed microwell plates at a density of 5,000 cells/200 ìÀ/well. Vitreous samples were as sayed without further treatment. All sera were heat-inactivated at 56 C for 30 minutes to pre vent a nonspecific toxic effect to the B9 cells. Serial dilutions of each sample were added in triplicate to the wells. Proliferation was evalu ated after 72 hours by measuring the amount of radioactivity incorporated after a four-hour pulse with 1 ì è of [ 3H] thymidine. Sample values were always compared to a standard curve generated with recombinant human in terleukin-6 (S. Gillis, M.D., Immunex, Inc., Seattle, Washington) or natural human inter leukin-6 (Endogen, Inc., Boston, Massachu setts). No inhibitory activity on B9-cell prolifer ation was observed in the vitreous or serum after the addition of various amounts of recom binant human interleukin-6. The detection lim it in vitreous and serum was 1.0 p g / m l . Neutralization with anti-interleukin-6 antiserum—To assess the specificity of the interleu kin-6 activity in our vitreous samples, we per formed neutralization tests with a commercially available rabbit antiserum specific for human interleukin-6 (Endogen, Inc.). Aliquots of vitre ous were incubated at 37 C for three hours with various concentrations of rabbit anti-interleu kin-6 or equivalent doses of preimmmune rab bit serum (Endogen, Inc.). Duplicate aliquots (100 ìÀ each) of neutralized specimens were assayed for their ability to induce proliferation of the B9 cells as described. Tumor necrosis factor-a enzyme-linked immunosorbent assay (ELISA)—Tumor necrosis fac tor-a measurements were performed with an
Proliferative Diabetic Retinopathy
Vol. 114, No. 6
ELISA kit (Cistron Biotechnology, Pine Brook, New Jersey). Undiluted vitreous and sera were tested in duplicate by adding 100 ìÀ/well. A standard curve was run in each assay with predetermined amounts of recombinant human tumor necrosis factor-a (range, 20 to 1,000 pg/ml). This method showed a detection limit of 20 p g / m l of tumor necrosis factor-a. The standard curve was not altered by dilution of recombinant human tumor necrosis factor-a in pooled normal human sera. Interleukin-lß ELISA—Interleukin-lß was quantitated with an ELISA kit (Cistron Biotech nology). The assay was used according to the manufacturer's instructions. Duplicate aliquots of 100 ìÀ of undiluted vitreous and serum were tested. The standard curve obtained with recombinant human interleukin-lß ranged from 20 to 1,000 p g / m l . The detection limit was 20 pg/ml. The standard curve was not altered by dilution of recombinant interleukin-lß in pooled normal human sera. Interferon-^ ELISA—Interferon-^ was quanti tated with an ELISA kit (Amgen Biological, Thousand Oaks, California). The assay was used according to the manufacturer's instruc tions. Duplicate aliquots of 100 ìÀ of undiluted vitreous and serum were tested. The detection limit was 6.25 U / m l . The standard curve ranged from 6.25 to 50 U / m l . Statistical analysis—Data were analyzed by using one-tailed paired or unpaired f-tests, Mann-Whitney U tests, and linear regression analysis as appropriate. A P value less than .05 was considered significant. Values were ex pressed as mean ± standard error of the mean.
Results Cytokines in ocular fluids—Interleukin-6 was detected in 15 of 16 vitreous samples (94%) from patients with proliferative diabetic reti nopathy. The levels of detectable interleukin-6 in the vitreous samples ranged from 1.14 to 190 pg/ml ± (43.3 ± 15.7 p g / m l ; mean ± standard error of the mean) (Fig. 1). Interleukin-6 was not detected in any of the aqueous samples from 15 normal control subjects (vitreous sam ples vs aqueous samples, P < .005; MannWhitney U test). In vitreous fluid from control subjects (seven post mortem), interleukin-6 levels ranged from 0 to 49.5 pg/ml (9.80 ± 6.05 pg/ml). Interleu kin-6 values in these control subjects were
733
r
Ö 100
=
so
•i **«»-
•JMW^
Vitreous Serum Fig. 1 (Abu El Asrar and associates). Detectable interleukin-6 (IL-6) levels in vitreous and paired serum samples from 16 eyes. One vitreous sample and eight serum samples had interleukin-6 levels less than 1 pg/ml. lower than in patients with proliferative diabet ic retinopathy (P < .03, unpaired f-test). Interleukin-6 levels in the vitreous samples of patients with proliferative diabetic retinopa thy were positively correlated with the grade of clinical ocular disease severity (Fig. 2) (R = 0.677, P < .004; linear regression analysis). The two diabetic patients with renal transplants had interleukin-6 levels of 1.84 p g / m l (disease activity, 1+) and 15 p g / m l (disease activity, 3 + ). The presence of specific interleukin-6 activity in vitreous samples was confirmed by using a rabbit anti-human interleukin-6 in neutraliza tion tests. The proliferative response of B9 cells to interleukin-6 from vitreous samples was completely blocked by the rabbit antiserum. Interleukin-lß protein was detected in seven of 16 vitreous samples (44%), with levels rang ing from 21 to 68 p g / m l . Interleukin-lß was present in four of ten aqueous samples, with levels ranging from 36 to 59 p g / m l . Tumor necrosis factor-a and interferon--y were not detected in any of the vitreous sam ples from patients with proliferative diabetic retinopathy or aqueous samples from control subjects. Cytokines in the serum—Interleukin-6 was detected in eight of the 16 serum samples (50%). The serum interleukin-6 level ranged from 1.1 to 21.4 p g / m l (Fig. 1). These levels were not different from those found in the serum of normal control samples assayed in our laboratory. Interleukin-6 levels in the vitreous samples were higher than those in paired sera (P < .05; paired f-test). Interleukin-lß, tumor necrosis factor-á, and
734
AMERICAN JOURNAL OF OPHTHALMOLOGY
Disease Activity
Fig. 2 (Abu El Asrar and associates). Correlation of interleukin-6 (IL-6) levels in vitreous samples with clinical disease activity in 16 eyes from patients with proliferative diabetic retinopathy. interferon-ã were not detected in any of the sera.
Discussion In our study, interleukin-6 was found in 15 of 16 vitreous samples (94%) from patients with proliferative diabetic retinopathy. Interleu kin-6 was not present in aqueous samples from normal control subjects and was also signifi cantly lower in postmortem vitreous samples from subjects without intraocular disease. In postmortem eyes, the delay between death and harvesting of vitreous would be expected to increase vitreous protein, 17 including interleu kin-6, which implies that the interleukin-6 found in fresh vitreous in proliferative diabetic retinopathy is biologically important. It is pos sible that compartments such as the vitreous act as a reservoir for cytokines, thus amplifying local effects. The levels of interleukin-6 correlate with clinical disease severity in eyes as graded by our criteria (Fig. 2). An increase of interleu kin-6 was not detected in paired serum sam ples, suggesting that interleukin-6 in the vitre ous is intraocular in origin and designates proliferative retinopathy. The inference that interleukin-6 synthesis is local would theoreti cally be confirmed by Northern blot analysis of retinal tissue, but retinal or iris biopsy is not performed in patients undergoing vitrectomy for proliferative diabetic retinopathy. Interleukin-6 can be produced by various cell types, including monocytes, activated T cells,18
December, 1992
vascular endothelial cells, fibroblasts, 19 and glial cells.20 These cells could be the source of intraocular interleukin-6 in patients with pro liferative retinopathy caused by diabetes and possibly by sickle cell disease, retinal detach ment, and retinal vein occlusion. Interleukin-6, also called B-cell stimulating factor-2, may have an important role in anti body-mediated diseases, including those in re stricted compartments of the body. Interleu kin-6 is present in the aqueous humor of 63% of patients with uveitis. 21 This cytokine is also detected in the synovial fluid of patients with active rheumatoid arthritis,22"24 in which inter leukin-6 levels correlate with disease activity. 24 Several tumors, such as cardiac myxomas and cervical carcinomas, produce interleukin-6. Pa tients with such tumors have hypergammaglobulinemia and multiple autoantibodies that dis appear from the serum after surgical removal of the tumor. 25 Unregulated production of inter leukin-6 might be responsible for autoantibody production in these patients. The interleukin-6 we detected in the vitreous from patients with proliferative diabetic reti nopathy could be involved in deposits of immunoglobulins in the walls of newly formed preretinal vessels. 2 Biopsies of the pars plana in proliferative diabetic retinopathy demonstrate immunoglobulin deposits at the level of the pigmented epithelium 3 and the preretinal mem branes show extensive immunoglobulin depos its in the connective stroma and within the vascular walls. 4 In addition to increased intra ocular interleukin-6 and ocular deposits of im munoglobulin, patients with type I diabetes mellitus have autoantibodies that are directed against a variety of pancreatic and nonpancreatic antigens. Moreover, the frequency of other autoimmune diseases is higher in diabetic pa tients than in the normal population. 26 We have also found interleukin-6 in the vitre ous from four patients with proliferative vitreoretinopathy (13.7 ± 4.7 pg/ml) (Reder, A. T., Abu El Asrar, A. M., and Maimone, D., unpub lished data, 1991). In this disease also, immu noglobulin and complement deposits are seen at the level of the pigmented epithelium. 27 How ever, the disease mechanism differs from pro liferative diabetic retinopathy, for there is metaplasia of the retinal pigment epithelium without vascular proliferation. A pathogenetic role could be envisaged in proliferative retinopathy for tumor necrosis fac tor-a, possibly with a synergistic effect from interleukin-lß. 2 8 Tumor necrosis factor-a and interieukin-Iß induce angiogenesis in ocular
Vol. 114, No. 6
Proliferative Diabetic Retinopathy
interleukin-lß induce angiogenesis in ocular tissues5·6 and proliferation of glial cells7,8 and fibroblasts. We did not detect tumor necrosis factor-a in any of the vitreous samples we examined. Tumor necrosis factor-a clears rapid ly from the circulation. 29 Similar rapid clearance or degradation of proteins in the eye could have hampered the detection of this cytokine in our study. 30 Interleukin-lß was detected in seven of 16 vitreous samples from patients with prolifera tive diabetic retinopathy. Interleukin-lß, how ever, was also present in four of ten aqueous humor samples from normal control subjects. Normal aqueous humor contains a remarkably consistent spectrum of proteins, including immunoglobulin, albumin, and lower molecular weight proteins. 17 The pattern does not reflect serum proteins. The interleukin-1 in the aque ous and vitreous may have been synthesized within these compartments, for serum interleu kin-1 was undetectable. Finally, surgical trau ma to the iris markedly increases protein levels in the aqueous. 17 Our samples were obtained with exquisite care and trauma was unlikely. It is also possible that interleukin-6 down-regu lates production of tumor necrosis factor-a and to some extent interleukin-lß in the eyes of patients with proliferative diabetic retinopa thy.31 In previous immunopathologic studies, class II major histocompatibility complex antigen expression was noted in the pigmented epithe lium, 3 as well as on many cells in the preretinal membranes 4 from patients with proliferative di abetic retinopathy. Interferon-ã is a potent inducer of class II antigen expression on a variety of epithelial cells in vitro, 1113 and activates retinal pigmented epithelial cells, which allows them to present antigens. 32 Interferon--y is de tected in the vitreous of patients with acquired immunodeficiency syndrome retinitis 33 and sympathetic ophthalmia 84 and could amplify proliferative retinopathy. In our study, howev er, we failed to detect interferon-ã in the vitre ous samples. We cannot exclude the possibility that small amounts of interferon-ã, not detect able with our ELISA system at the time of sampling, were present in the vitreous from patients with proliferative diabetic retinopa thy. In this regard, the synovial fluid of patients with rheumatoid arthritis contains little or no interferon-ã, despite class II major histocom patibility complex antigen expression by the synovial cells.35 This suggests that factors other than interferon--y may be responsible for class II
735
major histocompatibility complex antigen ex pression in localized compartments. We found that detectable levels of interleu kin-6 in vitreous samples from patients with proliferative diabetic retinopathy, and interleu kin-6 levels correlate with clinically detectable ocular disease activity. The presence of inter leukin-6 may account for the well-established finding of immunoglobulin deposition in ocular tissues. Our data suggest that an inflammatory reaction may further exacerbate disease severi ty in eyes of patients with proliferative retinop athy. Therefore, the use of anti-inflammatory drugs or cytokine antagonists before, during, and after surgical intervention warrants further evaluation.
References 1. Rahi, A. H. S., and Addison, D. J.: Autoimmunity and the outer retina. Trans. Ophthalmol. Soc. U.K. 103:428, 1983. 2. Melato, M., Antonutto, G., Manconi, R., and Ponte, E.: Ocular deposits of immunoglobulin in diabetic retinopathy. Can. J. Ophthalmol. 17:45, 1982. 3. Baudouin, C, Fredj-Reygrobellet, D., Lapalus, P., and Gastaud, P.: Immunohistopathologic findings in proliferative diabetic retinopathy. Am. J. Ophthal mol. 105:383, 1988. 4. Baudouin, C, Gordon, W. C , Fredj-Reygrobel let, D., Baudouin, F., Peyman, G. ]., Gastaud, P., and Bazan, N. G.: Class II antigen expression in diabetic preretinal membranes. Am. J. Ophthalmol. 109:70, 1990. 5. Ben Ezra, D., Hemo, I., and Maftzir, G.: In vivo angiogenic activity of interleukins. Arch. Ophthal mol. 108:573, 1990. 6. Frater-Schroder, M., Risau, W., Hallmann, R., Gautschi, P., and Bohlen, P.: Tumor necrosis factor type a, a potent inhibitor of endothelial cell growth in vitro, is angiogenic in vivo. Proc. Nati. Acad. Sci. U.S.A. 84:5722, 1987. 7. Guilan, D., and Lachman, L. B.: Interleukin-1 stimulation of astroglial proliferation after brain in jury. Science 220:497, 1985. 8. Selmaj, K. W., Farooq, M., Morton, W. T., Raine, C. S., and Brown, C. F.: Proliferation of astrocytes in vitro in response to cytokines. A primary role for tumor necrosis factor. J. Immunol. 144:129, 1990. 9. Kohase, M., May, L. T., Tamm, I., Vilcek, J., and Sehgal, P. B.: A cytokine network in human diploid fibroblasts. Interactions of b-interferons, tumor ne crosis factor, platelet-derived growth factor, and in terleukin-1. Mol. Cell. Biol. 7:273, 1987. 10. Hirano, T., Taga, T., Nakano, N., Yasukawa, K., Kashiwamura, S., Shimizu, K., Nakajima, K., Pyun,
736
AMERICAN JOURNAL OF OPHTHALMOLOGY
K. H., and Kishimoto, T.: Purification to homogenei ty and characterization of human B-cell differentia tion factor (BCDF or BSF p-2). Proc. Nati. Acad. Sci. U.S.A. 82:5490, 1985. 11. Basham, T. Y., Nickoloff, B. J., Merigan, T. C , and Morhenn, V. B.: Recombinant y interferon differ entially regulates class II antigen expression and biosynthesis on cultured normal human keratinocytes. J. Interferon Res. 5:23, 1985. 12. El-Asrar, A. M., Van Den Gord, J. J., Billiau, A., Desmet, V., Emarah, M. H., and Missotten, L.: Recombinant interferon-7 induces HLA-DR expression on human corneal epithelial and endothelial cells in vitro. A preliminary report. Br. J. Ophthalmol. 73:587, 1987. 13. Tabibzadeh, S. S., Gerber, M. A., and Satyaswaroop, P. G.: Induction of HLA-DR antigen expres sion in human endometrial epithelial cells in vitro by recombinant-interferon. Am. J. Pathol. 125:90, 1986. 14. Preud'homme, G. J., Fieser, T. M., Dixon, F. ]., and Theofilopoulos, A. N.: B cell-tropic interleukins in murine systemic lupus erythematosus (SLE) 1. Immunol. Rev. 78:159, 1984. 15. Sidman, C. L., Marshall, J. D., Masiello, N. C., Roths, ]. B., and Shultz, L. D.: Novel B-cell matura tion factor from spontaneously autoimmune viable motheaten mice. Proc. Nati. Acad. Sci. U.S.A. 81:7199, 1984. 16. Aarden, L., DeGroot, E. R., Schaap, O. L., and Lansdorp, P. M.: Production of hybridoma growth factor by human monocytes. Eur. J. Immunol. 17:1411, 1987. 17. Tripathi, R. C , Millard, C. B., and Tripathi, B. J.: Protein composition of human aqueous humor. SDS-PAGE analysis of surgical and post-mortem samples. Exp. Eye Res. 48:117, 1989. 18. Houssiau, F. A., Coulie, P. G., and Van Snick, J.: Distinct roles of IL-I and IL-6 in human T-cell activation. J. Immunol. 143:2520, 1989. 19. Mizel, S. B.: The interleukins. FASEB J. 3:2379, 1989. 20. Frei, K., Malipiero, U. V., Leist, T. P., Zinkernagel, R. M., Schwab, M. E., and Fontana, A.: On the cellular source and function of interleukin 6 pro duced in the central nervous system in viral diseases. Eur. J. Immunol. 19:689, 1989. 21. Murray, P. I., Hoekzema, R., van Haren, M. A. C , de Hon, F. D., and Kijlstra, A.: Aqueous humor interleukin-6 levels in uveitis. Invest. Oph thalmol. Vis. Sci. 31:917, 1990. 22. Hirano, T., Matsuda, T., Turner, M., Miyasaka, N., Buchan, G., Tang, B., Sato, K., Shimizu, M., Maini, R., Feldmann, M., and Kishimoto, T.: Exces sive production of interleukin 6/B cell stimulating factor-2 in rheumatoid arthritis. Eur. J. Immunol. 18:1797, 1988. 23. Bhardwaj, N., Santhanam, U., Lau, L. L., Tat ter, S. B., Ghrayeb, J., Rivelis, M., Steinman, R. M.,
December, 1992
Sehgal, P. B., and May, L. T.: IL-6/IFN-ß 2 in synovial effusions of patients with rheumatoid arthritis and other arthritides. Identification of several isoforms and studies of cellular sources. J. Immunol. 143:2153, 1989. 24. Waage, A., Kaufmann, C , Espevik, T., and Husby, G.: Interleukin-6 in synovial fluid from pa tients with arthritis. Clin. Immunol. Immunopathol. 50:394, 1989. 25. Hirano, T., Taga, T., Yasukawa, K., Nakajima, K., Nakano, N., Takatsuki, F., Shimizu, M., Murashima, A., Tsunasawa, S., Sakiyama, F., and Kishimoto, T.: Human B-cell differentiation factor defined by an antipeptide antibody and its possible role in autoantibody production. Proc. Nati. Acad. Sci. U.S.A. 84:228, 1987. 26. Drell, D. W., and Notkins, A. L.: Multiple immunological abnormalities in patients with type I (insulin-dependent) diabetes mellitus. Diabetologia 30:132, 1987. 27. Baudouin, C , Fredj-Reygrobellet, D., Bau douin, F., Lapalus, P., and Gastaud, P.: Immunohistologic study of proliferative vitreoretinopathy. Am. J. Ophthalmol. 108:387, 1989. 28. Dinarello, C. A.: The biology of interleukin 1 and comparison to tumor necrosis factor (correspon dence). Immunology 16:227, 1987. 29. Michie, H. R., Manague, K. R., Spriggs, D. R., Revhaug, A., O'Dwyer, A., Dinarello, C. A., Cerami, A., Wolff, S. M., and Wilmore, D. W.: Detection of circulating tumor necrosis factor after endotoxin ad ministration. N. Engl. J. Med. 318:1481, 1988. 30. Tripathi, B. J., Geanon, J. D., and Tripathi, R. C : Distribution of tissue plasminogen activator in human and monkey eyes. Ophthalmology 94:1434, 1987. 31. Schindler, R., Mandila, J., Endres, S., Ghorbani, R., Clark, S. C , and Dinarello, C. A.: Correla tion and interactions in the production of interleu kin-6 (IL-6), IL-1, and tumor necrosis factor (TNF) in human blood mononuclear cells. IL-6 suppresses IL-1 and TNF. Blood 75:40, 1990. 32. Percopo, C. M., Hooks, J. ]., Shinohara, T., Caspi, R., and Detrick, B.: Cytokine-mediated activa tion of a neuronal retinal resident cell provokes antigen presentation. J. Immunol. 145:4101, 1990. 33. Mondino, B. J., Sidikaro, Y., Mayer, F. J., and Sumner, H. L.: Inflammatory mediators in the vitre ous humor of AIDS patients with retinitis. Invest. Ophthalmol. Vis. Sci. 31:798, 1990. 34. Hooks, J. J., Chan, C. C , and Derrick, B.: Iden tification of the lymphokines, interferon-ã and interIeukin-2, in inflammatory eye diseases. Invest. Oph thalmol. Vis. Sci. 29:144, 1988. 35. Firestein, G. S., and Zavifler, N. J.: Peripheral blood and synovial fluid monocyte activation in in flammatory arthritis. Arthritis Rheum. 30:864, 1987.
