Association of Interferon-γ, Interleukin-10, and Tumor Necrosis Factor-α Gene Polymorphisms with Occurrence and Severity of Eales' Disease

Abstract

Purpose.: Eales' disease (ED) is an idiopathic retinal vasculitis characterized by capillary nonperfusion and neovascularization. Previous reports on ED demonstrated that T-cell-mediated immunoresponse and differential cytokine production in inflammatory and angiogenic stage seem to influence the extent and severity of this disease. Therefore, the purpose of this study is to investigate the influence of cytokine gene polymorphisms on occurrence and severity of ED.

Methods.: One hundred twenty-one patients with ED were recruited from an Eastern Indian population and compared with 223 matched healthy control subjects. Genotyping of IFN-γ, IL-10, and TNF-α were performed by amplification refractory mutation system polymerase chain reaction (ARMS-PCR).

Results.: A statistically significant association was found between the IL-10 −1082AA (P = 0.002), TNF-α −308AA (P = 0.0017) genotypes and the IL-10 ATA haplotype (P = 0.0123) and the occurrence of ED. In addition IL-10 −1082GG (P = 0.0005), TNF-α −308GG (P < 0.0001) genotype were found to be protective against disease occurrence. A synergistically low IL-10/high TNF-α genotype increased the risk of development (P < 0.0001) and the severity (P = 0.019) of ED.

Conclusions.: These data suggest that a low IL-10–expressing and high TNF-α–expressing genotype of the host can influence the occurrence and severity of outcome of ED.

Eales' disease (ED), first described by the British ophthalmologist Henry Eales in 1882, is an idiopathic inflammatory vasoproliferative retinal disease primarily affecting the peripheral retina of young and apparently healthy adult men between 20 and 40 years of age. ^1,2 Although present in the Western world, it is predominantly found in the Indian subcontinent. ³ It is a clinical condition characterized by the sequential cascade of peripheral retinal inflammation (that is, perivasculitis; the peripheral sheathing of veins leading to sclerosis of the retinal vessels; ischemia of the retina, hence, retinal nonperfusion culminating in retinal neovascularization; and recurrent vitreous hemorrhage, with or without retinal detachment. ^3,4  

Despite a quite well-defined clinical profile, the etiopathogenesis of this disease is still debatable. Among the most favored theories are mycobacterial infection and hypersensitivity to tuberculoprotein. ^{5 –8} Association of human leukocyte antigen (HLA), predominantly HLAB5 (B51), DR1, and DR4 ⁹ , and T-cell involvement in the lymphocytic infiltration of epiretinal membrane (ERM) and subretinal membrane (SRM) in patients with ED (Badrinath SS, et al. IOVS 1992;33:ARVO Abstract 857) indicate that a T-cell-mediated immune mechanism may play a key role in retinal vasculitis (i.e., the inflammatory stage) in this disease. 

T-cell immunity is mediated by the dynamic balance between pro- and anti- inflammatory cytokines. Interferon gamma (IFN)-γ is a key T-helper type 1 (Th 1) cytokine produced by T-cells and natural killer (NK) cells and acts as an activator of macrophages, thus playing a central role in influencing the host defense. ¹⁰ It also upregulates several proinflammatory cytokines, including tumor necrosis factor (TNF)-α. ¹¹ TNF-α, also known as an angiogenic cytokine, is involved in inflammation-mediated angiogenesis. ¹² TNF-α stimulates a dose-dependent response of matrix metalloproteinase (MMP) ¹³ and vasoformative substance (growth factor) ¹⁴ activation and promotes peripheral neovascularization of the retina, which leads to severe outcome of this disease. ¹⁵ On the contrary, anti-inflammatory effectors play an important role in countering the effects of proinflammatory cytokines. Interleukin (IL)-10 is classically described as an anti-inflammatory cytokine with pleiotropic effects in immunoregulation and inflammation by inhibiting the antigen-specific T-cell proliferation, nitric oxide secretion, and major histocompatibility complex (MHC) class 2 expression. ¹⁶ In animal models, IL-10 inhibits TNF-α production and neutrophil activation, thus diminishing lung tissue injury. ¹⁷ It also inhibits tumor growth in human melanoma cells by inhibition of angiogenesis. ¹⁸  

The cascade of cytokines that emerge during the inflammatory stage of ED ¹⁹ clearly indicates their contribution to the immunomodulation of the defense against antigens, and the variation in cytokine production affects the extent and severity of disease in all probability. As the magnitude of cytokine production depends not only on antigenic challenge but also on genetic factors in the host, such as single nucleotide polymorphisms (SNP) located in the coding and regulatory regions of cytokine-expressing genes, ²⁰ the search for SNPs has now become a potential tool for the genetic prediction of disease susceptibility and rate of development of severity. 

The purpose of this study was to determine whether there is any association between the SNPs of IFN-γ +874 A/T; IL-10 −1082G/A, −819C/T, and −592C/A; and TNF-α −308G/A and −238G/A, and the occurrence and severity of ED, taking into account the role of these cytokines in the inflammatory and angiogenic stages of this disease. 

Methods

Study Subjects

One hundred twenty-one patients (97 males, 24 females) with ocular findings suggestive of ED were recruited from the Retina Research Clinic at the Regional Institute of Ophthalmology, Kolkata, India. Forty-one patients presented with active vasculitis in the retinal periphery, 50 had neovascularization in the periphery, and 30 had advanced neovascularization in the disc and periphery with vitreous hemorrhage and tractional retinal detachment. 

Diagnosis was made with the help of direct and indirect ophthalmoscopy, slit lamp biomicroscopy with +90 D and three-mirror lenses, and fluorescein angiography. Ultrasonography was performed to detect tractional retinal detachment in eyes affected by vitreous hemorrhage. Other ophthalmic examinations included best corrected visual acuity determined by ETDRS chart, intraocular pressure measurement by applanation tonometry, and anterior segment evaluation by slit lamp examination. The location and extent of retinal involvement due to vasculitis, neovascularization, and fibrovascular traction were documented in all patients by fundus drawing and digital fundus color photography. 

Our control group was composed of 223 age-, sex-, and ethnically matched, healthy adults, without any history suggestive of ED, attending the outpatient department of the same institute for treatment of visual difficulties due to refractive errors, and without any other systemic or ocular disease. 

Subjects with a history of diabetes mellitus, hypertension, collagen vascular disease, sarcoidosis, Behçet's disease, systemic lupus erythematosus, Coats' disease, or syphilis were excluded from the study. The study protocol complied with the Declaration of Helsinki and was approved by the institution's ethics committee. Informed consent was obtained from each subject. 

DNA Isolation and Genotyping of Cytokine Gene Polymorphism

Venous blood samples were collected by venipuncture from the study subjects and genomic DNA was isolated from peripheral blood mononuclear cells (PBMCs) by proteinase-K digestion and a standard high salt-extraction method. ²¹ All polymorphisms were typed by using amplification refractory mutation system polymerase chain reaction (ARMS-PCR). ²² Briefly, 100 ng genomic DNA was amplified with DNA polymerase (Ampli Taq Gold; Applied Biosystems, Inc. [ABI], Foster City, CA) in two separate PCR reaction mixtures, containing 1× PCR buffer (Applied Biosystems), 2 mM MgCl₂, 0.25 mM each dNTP, 20 picomoles generic antisense primer, and 20 picomoles of one of the two allele-specific primers. To assess the success of PCR, an internal control 426-bp nucleotide sequence of human growth hormone or 266 bp of the β₂-microglobulin gene was amplified in both reactions with specific primers (20 picomoles/reaction). The primer sequences, primer annealing temperature (Ta°), and PCR product sizes are given in Table 1. The PCR reactions were performed in the following conditions: 95°C 4 minutes, followed by 10 cycles of 95°C for 15 seconds, Ta° for the internal control (Table 1) for 50 seconds, 72°C for 40 seconds, and then 25 cycles of 95°C for 20 seconds, Ta° for each SNP (Table 1) for 50 seconds, 72°C for 50 seconds, and 72°C for 10 minutes for the final extension. The amplified products were separated by electrophoresis on 2% agarose gel stained with 0.5 mg/mL ethidium bromide and visualized and photographed under a UV transilluminator (Fig. 1). 

Table 1.

 

Primer Sequence and Condition Used for ARMS-PCR Method

Primer Name	Primer Sequence (5′–3′)	Product Size (bp)	Annealing Temperature (°C)
IFN-γ +874(A/T)	TCAACAAAGCTGATACTCCA	261	56
IFN-γ +874A	TTCTTACAACACAAAATCAAATCA
IFN-γ +874T	TTCTTACAACACAAAATCAAATCT
IL-10 −1082 (G/A)	CAGCCCTTCCATTTTACTTTC	550	56
IL-10 −1082 G	TACTAAGGCTTCTTTGGGAG
IL-10 −1082 A	CTACTAAGGCTTCTTTGGGAA
IL-10 −819 (C/T)	GCTGTCCCCCACCCCAACTGTG	166	58
IL-10 −819 T	ACCCTTGTACAGGTGATGTAAT
IL-10 −819 C	ACCCTTGTACAGGTGATGTAA
IL-10 −592 (C/A)	TAACTTAGGCAGTCACCTTAGG	151	58
IL-10 −592 A	ACATCCTGTGACCCCGCCTGTA
IL-10 −592 C	ACATCCTGTGACCCCGCCTGTC
TNF-α −308 (G/A)	TCTCGGTTTCTTCTCCATCG	184	58
TNF-α −308 G	ATAGGTTTTGAGGGGCATGG
TNF-α −308 A	AATAGGTTTTGAGGGGCATGA
TNF-α −238 (G/A)	CCGGATCATGCTTTCAGTGC	590	58
TNF-α −238 G	AGACCCCCCTCGGAATCG
TNF-α −238 A	AAGACCCCCCTCGGAATCA
β2-Microglobulin FP	CCAAAGATTCAGGTTTACTCACG	266	66
β2-Microglobulin RP	ACTTAACTATCTTGGGCTGTGAC
Growth factor FP	CCTTCCAACCATTCCCTTA	426	62
Growth factor RP	TCACGGATTTCTGTTGTGTTTC

 FP, forward primer; RP, reverse primer.

Figure 1.

 

Representative genotypes of five patients according to their (A) IFN-γ +874A/T, (B) IL-10 −1082G/A, (C) IL10 −819C/T, (D) IL10 −592C/A, (E) TNF-α −308G/A, and (F) TNF-α −238G/A polymorphisms. The two alleles of the polymorphic region of each SNP were analyzed in each patient in two different assays by ARMS-PCR. To assess the success of PCR, internal control region of either growth hormone (426 bp; A–E) or the β2-microglobulin (266 bp; F) gene were amplified along with the each SNP. Products amplified were run in a 2% agarose gel. M1, DNA/BsuRI (HaeIII) marker; M2, 100-bp DNA ladder.

Statistical Analysis

Age and sex differences between the patients and control subjects were investigated by Student's t-test and χ² test, respectively (GraphPad, San Jose, CA). The study groups were tested for Hardy-Weinberg equilibrium (HWE) and the expected and observed frequencies were compared by χ² analysis (by an online calculator: provided by Tufts University, Boston, MA: http://www.tufts.edu/∼mcourt01/Documents/Court%20lab%20-%20HW%20calculator.xls). Haplotype frequencies were estimated by SNPStats software (http://bioinfo.iconcologia.net/snpstats/start.htm provided in the public domain by the Biostatistics and Bioinformatics Unit, Catalan Institute of Oncology, Barcelona, Spain). The allele, genotype, haplotype, and combined genotype frequencies of each SNP were compared between ED patients and control subjects by using the χ² analysis of 2 × 2 contingency tables and calculation of the odds ratio (OR) with 95% confidence interval (95% CI). Fisher's exact P-values were calculated for analysis in which one or more variables within 2 × 2 tables were less than 5 (by online calculator http://faculty.vassar.edu/lowry/odds2x2.html/ provided in the public domain by Vassar College, Poughkeepsie, NY). The level of statistical significance was set at P < 0.05, except for test results to which the Bonferroni adjustment was applied. 

Results

Demographic and clinical characterizations of the study participants are presented in Table 2. There was no significant difference between the groups in age, sex, ethnicity, and geographic origin. 

Table 2.

 

Demographic and Clinical Characteristics of the Study Participants

Characteristics	ED Patients (n = 121)	Controls (n = 223)	P
Ratio of men to women	97:24	183:40	0.662
Mean age in years ± SD	30.6 ± 10.7	32.26 ± 9.6	0.1424
Bilateral involvement	83	—	—
Type of lesion
Active vasculitis in peripheral retina	41	—	—
Neovascularization in the peripheral discs	50	—	—
Vitreous hemorrhage with tractional retinal detachment	30	—	—
Visual acuity
20/20–20/40	38	212	—
20/50–20/100	53	11	—
20/200–20/400	30	—	—

Genotyping was performed for the entire group of 121 ED patients and 223 healthy control subjects. No significant departure from HWE was observed for the IFN-γ +874A/T; IL-10 −1082G/A and −592C/A; TNF-α −308 G/A; and TNF-α −238 G/A variant in the ED patients (χ² = 0.15, P = 0.69; χ² = 2.59, P = 0.107; χ² = 1.95, P = 0.16; χ² = 1.06, P = 0.3; and χ² = 1.8, P = 0.17, respectively) or the control subjects (χ² =1.79, P = 0.18; χ² = 0.39, P = 0.52; χ² = 0.02, P = 0.88; χ² =2.20, P = 0.13; and χ² = 3.12, P = 0.07, respectively). However, there was a significant departure from HWE for the IL-10 −819C/T variant in the ED patients (χ² = 7.26; P = 0.007) and the control subjects (χ² = 7.30; P = 0.0068), where the observed heterozygote frequency was greater than expected. 

A Trend toward a High IFN-γ–Expressing Genotype in ED Occurrence

In our study, the minor allele (T) frequency was higher in the ED patients (n = 121) than in the control subjects (n = 223; 28.92% vs. 25.78%), but was statistically nonsignificant. During comparison of genotype frequencies, a statistically nonsignificant trend has also been observed that shows a higher frequency of the +874TT genotype in the ED patients than in the control subjects (9.09% vs. 4.93%; P = 0.13; Table 3). 

Table 3.

 

Distribution of Allele and Genotype Frequencies of IFN-γ Gene Polymorphisms

IFN-γ Gene Variants	ED Patients n (%)	Controls n (%)	P	OR (95% CI)
+874A/T alleles
A	172 (71.07)	331 (74.21)		0.8537 (0.6018–1.211)
T	70 (28.92)	115 (25.78)	0.37	1.1714 (0.8258–1.6616)
Genotypes
AA	62 (51.23)	119 (53.36)	0.7	0.9184 (0.5897–1.4303)
AT	48 (39.66)	93 (41.70)	0.71	0.9191 (0.5854–1.4431)
TT	11 (9.09)	11 (4.93)	0.13	1.9273 (0.8099–4.586)

Association between Low IL-10–Expressing Promoter Polymorphisms and ED Occurrence

Allele and genotype frequencies of all three IL-10 promoter polymorphisms (−1082G/A, −819C/T, and −592C/A) were compared between the ED patients (n = 121) and the control subjects (n = 223) and are presented in Table 4. The IL-10 −1082AA genotype frequency was significantly higher among the ED patients than in the control subjects (42.14% vs. 26.0%; P = 0.002), with a concomitant significant decrease in −1082GG genotype frequency among the ED patients (7.43% vs. 21.97%; P = 0.0005), but no significant difference in heterozygote frequency. The IL-10 −1082A allele frequency was also significantly higher in the ED patients (67.35% vs. 52.01%; P = <0.0001). No other IL-10 SNPs showed significant association with the occurrence of ED. It may be noted that all the P-values are significant at the 5% level even after Bonferroni correction for multiple tests (12 tests = 3 SNP loci × 4 tests per locus) since all P-values are lower than 0.004. 

Table 4.

 

Distribution of Allele and Genotype Frequencies of IL-10 Promoter Polymorphisms

IL-10 Gene Variants	ED Patients n (%)	Controls n (%)	P	OR (95% CI)
−1082G/A alleles
G	79 (32.64)	214 (47.98)		0.5254 (0.379–0.7284)
A	163 (67.35)	232 (52.01)	<.0001*	1.9032 (1.3728–2.6385)
Genotypes
AA	51 (42.14)	58 (26.0)	0.002*	2.0727 (1.297–3.3121)
AG	61 (50.41)	116 (52.01)	0.77	0.9378 (0.6023–1.4601)
GG	9 (7.43)	49 (21.97)	0.0005*	0.2853 (0.1349–0.6037)
−819C/T alleles
C	113 (46.69)	239 (53.58)	0.08	0.7587 (0.5544–1.0382)
T	129 (53.30)	207 (46.41)		1.3181 (0.9632–1.8038)
Genotype
TT	27 (22.31)	38 (17.04)	0.23	1.3984 (0.805–2.4293)
CT	75 (61.98)	131 (58.74)	0.55	1.145 (0.7274–1.8024)
CC	19 (15.7)	54 (24.21)	0.065	0.583 (0.3272–1.0388)
−592C/A alleles
A	106 (43.80)	190 (42.60)	0.8	1.050 (0.7658–1.4401)
C	136 (56.19)	256 (57.39)		0.9522 (0.6944–1.3058)
Genotype
AA	27 (22.6)	41 (18.2)	0.39	1.275 (0.7387–2.2008)
CA	52 (42.98)	108 (48.32)	0.36	0.802 (0.514–1.253)
CC	42 (34.42)	74 (33.48)	0.81	1.070 (0.671–1.707)

*  Significant after Bonferroni correction for multiple testing (i.e., the adjusted P-value at the 0.05 significance level is 0.004).

Further stratification of subjects into haplotype groups based on the three promoter polymorphisms yielded the following distribution: 41.32% (50/121) GCC; 75.20% (91/121) ATA; 42.14% (51/121) ACC; 17.35% (21/121) ATC in the ED patients and 53.36% (119/223) GCC; 61.88% (138/223) ATA; 43.04% (96/223) ACC; 16.59% (37/223) ATC in the control subjects (Table 5). Data are not presented for the other haplotypes (GTA, ACA, GTC, GCA), because their overall frequencies were <0.05% in this population. These analyses reveal that the ATA haplotype frequency was significantly higher in the ED patients than in the control subjects (75.20% vs. 61.88% P = 0.0123). These associations remained significant after Bonferroni correction for testing of the four major haplotypes. The GCC haplotype frequency was significantly decreased in the ED patients compared with that in the control subjects (41.32% vs. 53.36% P = 0.032); however, these associations are no longer significant after Bonferroni correction (the adjusted P-value at the 0.05 significance level for four major haplotypes is 0.0125). No other haplotypes differed significantly among the ED patients and control subjects. 

Table 5.

 

Distribution of IL-10 Promoter Haplotype Frequency

Haplotype (−1082/−819/−592)	ED Patients n (%)*	Controls n (%)*	P	OR (95% CI)
GCC	50 (41.32)	119 (53.36)	0.032	0.6155 (0.3935–0.9627)
ATA	91 (75.20)	138 (61.88)	0.0123†	1.8684 (1.1409–3.0597)
ACC	51 (42.14)	96 (43.04)	0.86	0.9638 (0.616–1.5082)
ATC	21 (17.35)	37 (16.59)	0.86	1.0557 (0.5863–1.9008)

*  Haplotype present in the number of individuals shown in each group.

†  Significant after Bonferroni correction for four major haplotypes (i.e., the adjusted P-value at the 0.05 significance level is 0.0125).

Significant Association between the TNF-α −308G/A Polymorphism but Not the −238G/A Polymorphism and ED Occurrence

Table 6 represents the allelic and genotypic distributions of the TNF-α −308G/A and −238G/A polymorphisms in the patients and control subjects. There was a significant difference in the allelic distribution of TNF-α −308G/A polymorphism between the groups (P = <0.0001), indicating that the −308A allele (34.71% ED patients vs. 19.50% control subjects) may be related to ED occurrence. Individuals with the −308AA and −308GA genotype were also overrepresented among the patients with ED compared with the control subjects (9.91% vs. 2.24%, P = 0.0017 and 49.58% vs. 34.52%; P = 0.00645, respectively). In addition, the −308G allele (65.28% ED patients vs. 80.49% control subjects) and the −308GG genotype (40.49% ED patients vs. 63.22% control subjects, P = <0.0001) was found to be protective against the disease. 

Table 6.

 

Distribution of Allele and Genotype Frequencies of TNF-α Promoter Polymorphisms

TNF-α Gene Variants	ED Patients n (%)	Controls n (%)	P	OR (95% CI)
−308G/A alleles
G	158 (65.28)	359 (80.49)		0.4558 (0.3201–0.6491)
A	84 (34.71)	87 (19.50)	<0.0001*	2.1938 (1.5407–3.1238)
Genotypes
GG	49 (40.49)	141 (63.22)	<0.0001*	0.3958 (0.2514–0.623)
GA	60 (49.58)	77 (34.52)	0.00645	1.865 (1.1883–2.9272)
AA	12 (9.91)	5 (2.24)	0.0017*	4.8 (1.6492–13.9704)
−238G/A alleles
G	204 (84.29)	376 (84.30)		0.9994 (0.6501–1.5366)
A	38 (15.70)	70 (15.69)		1.0006 (0.6508–1.5383)
Genotype
GG	84 (69.42)	155 (69.50)		0.996 (0.616–1.6103)
GA	36 (29.75)	66 (29.59)		1.0075 (0.6207–1.6352)
AA	1 (0.82)	2 (0.896)		0.9208 (0.0826–10.2604)

*  Significant after Bonferroni correction for multiple testing (i.e., the adjusted P-value at the 0.05 significance level is 0.00625).

The distribution of the TNF-α promoter polymorphism at position −238 was verified in all the subjects who also provided samples for −308 typing. No significant difference, however, was observed in allelic or genotypic frequencies between the groups. 

We note that all the P-values reported are significant at the 5% level, even after Bonferroni correction for multiple testing (8 tests = 2 SNP loci × 4 tests per locus), except for the TNF-α −308GA genotype which is marginally nonsignificant (0.00625; significance threshold after correction). However, our inference that the TNF-α −308AA genotype is significantly associated with ED remains unaltered, even though the heterozygotes may show only marginal overrepresentation among patients than in control subjects. 

Combined IL-10 −1082G/A and TNF-α −308G/A Genotypes and ED Occurrence

On the basis of the findings, the patients and control subjects were classified into high and low functional genotypes of IL-10 −1082G/A and TNF-α −308G/A SNP, those that are significantly associated with the occurrence of ED, and also analyzed for the synergistic effect of these two polymorphisms (Table 7). According to previous reports, the IL-10 −1082A and TNF-α −308A alleles are associated with low IL-10 and high TNF-α production, respectively, ^23,24 and the single-locus model study showed that low IL-10 producer and high TNF-α producer allele carriers were significantly overrepresented in the patient group than in the control group (92.56% vs. 78.02% P = 0.00058 and 59.5% vs. 36.77% P = <0.0001, respectively). 

Table 7.

 

Distribution of IL-10 −1082G/A and TNF-α −308G/A SNP Combined Genotype

Genotype	ED Patients n (%)	Controls n (%)	P	OR (95% CI)
IL-10 −1082
Low (AA/AG)	112 (92.56)	174 (78.02)	0.00058*	3.5045 (1.6564–7.4146)
High (GG)	9 (7.43)	49 (21.97)		0.2853 (0.1349–0.6037)
TNF-α −308
Low (GG)	49 (40.49)	141 (63.22)		0.3958 (0.2514–0.623)
High (AA/AG)	72 (59.5)	82 (36.77)	<0.0001*	2.5266 (1.6051–3.9773)
Combined IL-10/TNF-α
Low/low	44 (36.36)	102 (45.73)	0.09	0.6779 (0.4303–1.068)
Low/high	68 (56.19)	72 (32.28)	<0.0001*	2.6908 (1.7058–4.2446)
High/low	5 (4.13)	39 (17.48)	0.0003*	0.2034 (0.0779–0.5309)
High/high	4 (3.30)	10 (4.48)	0.7	0.7282 (0.2235–2.3729)

*  Significant after Bonferroni correction for multiple testing (i.e., the adjusted P-value at the 0.05 significance level is 0.0083).

Analysis of the combined IL-10 −1082G/A and TNF-α −308G/A genotype yielded a stronger association of the low IL-10/high TNF-α genotype with ED occurrence (P = <0.0001). Whereas, the high IL-10/low TNF-α genotype was represented as a protective genotype (P = 0.0003) of this disease. 

Discussion

The pathogenesis of ED remains largely unknown. There is a great controversy regarding the underlying mechanism responsible for the initiation and occurrence of ED. But the disease ultimately progresses to a florid presentation mimicking any other vasoproliferative disorder where endothelial dysfunction plays a key role. Endothelial dysfunction, a concomitant of endothelial cell activation, is characterized by the expression of leukocyte adhesion molecules, vascular permeability, thrombogenic propensity, dysregulation of vasoreactivity, remodeling of abnormal blood vasculature, and, most important, angiogenesis. ²⁵ Angiogenesis is a complex process characterized by the sprouting of new blood vessels from preexisting ones. Apart from its important role in embryonic growth and wound healing, excess angiogenesis is also a determinant of many pathologic conditions, including retinal neovascularization. ^15,26 The finding that inflammation is often associated with increased angiogenesis can be explained by inflammation-induced production of angiogenic factors, either from vascular or other tissue cells or from infiltrating leukocytes. ²⁵ TNF-α, a major inflammatory cytokine acts as an angiogenic stimulator in inflammation-mediated pathologic angiogenesis. ¹² Although there are some variations of thought regarding the angiogenic properties of TNF-α, ^27,28 it has been proposed that it may induce angiogenesis through the various secondary angiogenic factors such as platelet-derived growth factor (PDGF), vascular endothelial growth factor (VEGF), IL-8, and basic fibroblast growth factor (bFGF). ^14,29 The finding that the elevated level of TNF-α in the inflammatory stage of ED increased considerably in the proliferative stage, on a backdrop where inflammation has subsided clinically but retinal neovascularization and vitreous hemorrhage have already developed as a consequence of retinal hypoxia and ischemia, ¹⁹ providing the indication of TNF-α-mediated angiogenesis in this disease. However, angiogenesis is also accompanied by extracellular remodeling and controlled degradation of specific components of basement membrane and extracellular matrix involving different proteolytic components, among which MMPs play an essential role. ³⁰ TNF-α also increases the production of several members of the MMP family, ¹³ that are secreted as proenzymes activated by proteolytic cleavage and regulated by the family of inhibitors called tissue inhibitors of metalloproteinase (TIMP), ³¹ and promote the neovascularization of the retina. These matrix-degrading enzymes are also reported to be involved in the process of new vessel formation in retina affected by diabetic retinopathy. ³² Although, no previous reports have confirmed the involvement of these types of proteinases in the pathogenesis of ED, because of their important role in neovascularization, participation of these proteinases in the immunopathogenesis of ED cannot be ruled out. 

Pro- and anti-inflammatory cytokines, mainly IFN-γ and IL-10, respectively, regulate production of angiogenic cytokines which determine severity of inflammation thus influencing susceptibility of the individual to the disease. The proposed involvement of cytokine gene polymorphism in in vivo production of cytokines encouraged us to analyze the different polymorphisms of cytokine genes and to provide some insight into genotypes that are protective or predisposing for this disease. 

The functional study of IFN-γ +874A/T polymorphism, lies within a putative nuclear factor κβ (NF-κβ) binding site, have found that T allele and the TT genotype are linked to the high production of IFN-γ. ³³ Several reports have shown that the high IFN-γ production gives protection against the clinical manifestations of different diseases including tuberculosis, ³⁴ outlining its function in host defense against antigenic exposure. On the contrary, other studies have demonstrated that the high-expressing TT genotype is associated with disease occurrence. ³⁵ High IFN-γ production is associated with ocular disease outcome, as reported by others. ³⁶ In this study we observed a statistically nonsignificant trend of the IFN-γ +874T allele, which reflected a constitutively higher production of IFN-γ toward the occurence of ED. IFN-γ production is the hallmark of a Th1 response, and the polymorphism of its gene resulting in high IFN-γ production may be expected to produce a severe inflammatory response by increasing the transcription of the TNF-α gene, increasing the stability of TNF-α mRNA, with both as a whole increasing TNF-α production. ¹¹  

IL-10 is a multifunctional immunoregulatory cytokine produced by macrophages, T-cells, and retinal cells, ³⁷ usually considered to mediate potent downregulation of inflammatory responses. ¹⁶ Among the identified polymorphisms in the 5′ flanking promoter region of the IL-10 gene, the best described SNPs in this region −1082G/A, −819C/T, and −592C/A combine with microsatellite alleles to form haplotypes related to differential IL-10 gene expression as well as IL-10 production. ³⁸ It has been established that the IL-10 −1082/−819/−592 combinations GCC, ACC, and ATA are associated with high, intermediate and low IL-10 production, respectively. ³⁹ However, in the proximal region of the promoter, the homozygous genotype for the −1082G allele seems to be associated with the higher IL-10 production with respect to the GA heterozygous and AA homozygous genotypes. ²³ In this case-control study, we observed that the presence of the ATA haplotype was significantly associated with an increased risk of ED. In a single-locus study, −1082A allele carrier individuals were more susceptible to ED. In addition, the −1082G allele and −1082GG genotypes were found to be protective against the ED occurrence. Similar to our findings, the association between the low-IL-10-production genotype with occurrence and severity of the different inflammatory mediated ocular diseases has also been investigated by different groups. ^40,41 IL-10, besides its anti-inflammatory role, also plays a protective role in inflammation-mediated tissue destruction by downregulating the synthesis of VEGF along with angiogenic cytokines—namely, IL-1β, TNF-α, IL-6, and MMPs—and by inducing the upregulation of TIMP production. ^42,43 The lo- expressing genotype of the IL-10 gene can therefore be predicted to create a severe outcome of inflammatory reaction in the eye and drive the disease course toward blindness. 

The important immunomodulatory cytokine TNF-α plays a complex role in the pathogenesis of ED. Among the different polymorphisms, two common functional polymorphisms in the promoter region of TNF-α, −308G/A and −238G/A, have been well studied. In the present study, we found a significant association of the −308A allele as well as the −308AA genotype with the occurrence of ED. The TNF-α −308A allele affects the binding of transcription factor, either by increasing promoter activity or by inhibiting the repressor of transcription and thereby increasing the production of TNF-α. ²⁴ It can be hypothesized that the high expression of TNF-α acts as a central factor responsible for the inflammation-associated angiogenesis in the proliferative stage of ED by inducing several angiogenesis-associated factors. At the same time, it can be said that the TNF-α −308G allele and the −308GG genotype is linked with downregulation of TNF-α production, ²⁴ which tends to afford better protection against ED because of low TNF-α production during the advanced phase of the disease. 

The synergistic effect of cytokine gene polymorphisms, particularly in this case IL-10 and TNF-α, has complex and predominantly opposing roles in ED, may also regulate the phenotypic outcome of this disease. Because of that, we have also evaluated the interaction between IL-10 and TNF-α in promoting the occurrence of ED. Our results showed a strong association between susceptibility to development of ED and the low-IL-10/high-TNF-α genotype. These results suggest that the angiogenic properties of an elevated TNF-α level are not counterbalanced by the low amounts of IL-10 in patients with the low-IL-10/high-TNF-α genotype, thus increasing the susceptibility and triggering a severe disease course and outcome. However, most of TNF-α activity is mediated through TNF-α receptors; the future study of receptor gene polymorphisms may further elucidate the immunopathogenesis of this disease. 

Recently, anti-TNF-α therapy served to facilitate intraretinal angiogenesis and recovery of diabetic maculopathy, associated with lessening of intravitreal neovascularization also become a prospective clinical tool for treating retinal neovascularization in conjunction with laser photocoagulation. ⁴⁴ Anti-TNF-α therapy with infliximab, a humanized mouse monoclonal antibody to TNF-α, is also effective in the treatment of different diseases by minimizing the severe inflammatory response. ^{45 –47} Because of the pathologic role of TNF-α in the inflammatory and proliferative stages of ED, infliximab may seem to be beneficial. ¹⁹ However, TNF-α has a vital role in host defense against Mycobacterium tuberculosis and, in conjunction with TNF-α-dependent chemokines, helps in the development and maintenance of granulomas, which compartmentalize tubercle bacilli during infection. ^48,49 So TNF-α inhibition by infliximab may increase the risk of M. tuberculosis infection. The idea of analyzing IL-10 and TNF-α genotype at an early stage of ED affected patients to decide whether to apply immunosuppressive therapy has important clinical implications, because genetically predisposed low-IL-10/high-TNF-α–producing patients can only benefit from this type of treatment. 

To conclude, in the present study we have identified that common genetic variants of two key genes of the T-cell-mediated immune pathway—IL-10 and TNF- α—were associated with risk and severity of ED in our population. To the best of our knowledge, this is the first effort to investigate the cytokine gene polymorphism in patients with ED. This study also has pharmacogenomic value and may be clinically useful in imparting the treatment and the development of new immunotherapies in the future. 

Footnotes

 Supported by a grant from the Indian Council of Medical Research (ICMR).

Footnotes

 Disclosure: A. Sen, None; S.K. Paine, None; I.H. Chowdhury, None; L.K. Mondal, None; A. Mukherjee, None; A. Biswas, None; S. Chowdhury, None; S. Bhattacharya, None; B. Bhattacharya, None

The authors thank the medical institute for their cooperation in allowing us to contact the patients and to obtain information from their clinical records and Partha P. Majumder (Director, National Institute of Biomedical Genomics) for helpful suggestions and support.