Genetics of Uveitis

Uveitis refers to a diverse group of several dozen disorders characterized and connected by the shared presence of intraocular inflammation.^1,2 While the etiology of infectious uveitis is often clear and well-defined, the pathogenesis of most forms of noninfectious uveitis is believed be both complex and multifactorial, with important contributions coming from the environment,³ the microbiome,^4–6 and the patient’s underlying genetics.^7–10 In many instances, these seemingly unrelated factors may act in concert to increase disease susceptibility, as has been proposed with HLA-B27-associated uveitis¹¹ and HLA-A29-positive birdshot chorioretinopathy (BSCR).¹² Many additional HLA and non-HLA polymorphisms have been found to either increase or decrease a given patient’s or population’s risk of developing uveitis.^7–10,13,14 Rarely, specific gene mutations may produce a distinctive uveitis syndrome, as with alterations in the calcium-dependent cysteine protease Calpain-5 (CAPN5) gene, hemizygous mutations in which can confer a neovascular inflammatory vitreoretinopathy (NIV) characterized by the progressive development of uveitis, retinal degeneration, neovascularization, and intraocular fibrosis.^15,16 Uveitis can also occur with inherited retinal dystrophies (IRD), a large and diverse group of retinal degenerations with autosomal dominant, autosomal recessive, X-linked, or mitochondrial inheritance.¹⁷ Last, both paradoxical autoinflammatory uveitis, typically with granulomatous features,¹⁸ and opportunistic infectious uveitis, most commonly cytomegalovirus (CMV) retinitis¹⁹ or bacterial or fungal endophthalmitis,²⁰ can occur in patients with inherited or primary immunodeficiency diseases (PIDs), such as common variable immune disorder, chronic granulomatous disease (CGD), severe combined immunodeficiency, or Job and Blau syndromes.^21,22 In this issue of Ocular Immunology & Inflammation (OII), seven original articles^23–29 and three letters^30–32 touch upon aspects of the development, characteristics, or management of several forms of ocular inflammation that have strong genetic determinants.

Kumar et al²³ prospectively evaluated the interleukin (IL)-6 and IL-10 levels in tear samples from 20 consecutive HLA-B27-associated and 10 consecutive non-HLA-B27 associated uveitis control subjects seen at a tertiary referral center in Chandigarh, India. All patients had unilateral, active anterior uveitis, which was recurrent in six of the HLA-B27 positive cases. Tears were collected from both the actively inflamed and quiescent fellow eye. Cytokine levels were measured before initiation of therapy using bead-based flow cytometric assays. Both mean IL-6 level (751.4 ± 934.7 vs 261.9 ± 274.3; p = 1 × 10⁻⁶) and mean IL-6/IL-10 ratio (4.85 ± 6.42 vs 2.16 ± 3.28; p = .001) were significantly higher in HLA-B27-positive as compared to HLA-B27-negative subjects, whereas IL-10 levels trended higher in HLA-B27 cases, but the difference failed to achieve statistical significance (220.8 ± 208.1 vs 142 ± 23.36; p = .035). In addition, both mean IL-6 level and mean IL-6/IL-10 ratio were higher in tears from actively inflamed HLA-B27-positive eyes as compared to measurements obtained from the quiescent fellow eyes, and IL-6 levels tended to be higher in the subgroup of actively inflamed HLA-B27-positive eyes with past recurrences. The authors acknowledged the prior, corroborative work of El-Asrar et al,³³ who showed that aqueous humor IL-6 levels were significantly higher in uveitis patients who were HLA-B27-positive or who had Behçet disease (BD) compared to patients with Vogt-Koyanagi-Harada (VKH) disease, ocular sarcoidosis (OS), or idiopathic granulomatous uveitis (P < .0001), and suggested that measurement of tear IL-6 and IL-10 levels might offer a noninvasive way to predict disease recurrences over time in HLA-B27-positive patients with uveitis.

Lakra et al²⁴ retrospectively evaluated the efficacy of the subcutaneously administered tumor necrosis factor (TNF) inhibitors adalimumab (ADL) and golimumab (GLM) in 14 patients with HLA-B27 associated uveitis seen at a tertiary referral center in Chennai, India, between 2014 and 2018. Twelve patients (85.7%) were male. Ages in the cohort ranged from 17 to 53 years, with a median 31.5 years. Thirteen of the 14 (92.9%) had ankylosing spondylitis and 10 (71.4%) had isolated anterior uveitis, whereas four (28.6%) had both anterior and intermediate uveitis. The uveitis was bilateral in six subjects (42.9%). All patients were treated prior to the introduction of a TNF inhibitor topically with 1% prednisolone acetate and 2% homatropine, and with oral corticosteroids. In addition, eleven subjects were on methotrexate and two were prescribed oral sulfasalazine. One or more periocular corticosteroid injection was given to eleven patients. Twelve of the subjects (85.7%) were administered ADL, whereas two (14.3%) received GLM. Follow up ranged from seven to 36 months (mean 23 months). Vision improved in one or both eyes in 10 subjects (71.4%), with an improvement in median best-corrected visual acuity (BCVA) from logMAR 0.30 (Snellen 20/40) at baseline to logMAR 0.097 (~Snellen 20/20) at last visit. Cystoid macular edema, present in two study subjects (14.3%) at baseline, resolved following initiation of ADL. The authors concluded that their experience with the use of TNF inhibitors in India supported previous studies³⁴ demonstrating a role for such therapy in the management of both ocular and systemic inflammation in patients with HLA-B27-associated uveitis.

Anukul et al²⁵ compared the frequency of HLA-DRB1 and DQB1 class II allele genotypes in 23 patients with VKH disease to 20 patients with other forms of uveitis seen in a referral clinic in Northern Thailand. Forty healthy participants served as an additional control group. Both HLA–DRB1*04:05 (frequency 35%; OR 42.1; 95% CI 5.4-331.8; p < .001) and HLA-DQB1*04:01 (frequency 71.4% with none in either control group) were more prevalent in patients with VKH disease. The authors concluded that their findings supported studies from other populations demonstrating the association of by DRB1 and DQB1 HLA alleles with VKH disease.

Tsui et al²⁶ described the use of high-resolution DNA typing of HLA-A29 in two patients with family history of BSCR seen at a tertiary referral clinic in Northern California, USA. The first family consisted of a mother, her daughter, and a material aunt with BSCR. The second family consisted of two affected sisters. All five affected individuals carried a single HLA-A29*02 allele. The authors cited previous reports demonstrating the high rate of HLA-A29*02 carrier status among patients with BSCR,^35,36 and three previous studies suggesting a positive family history in 3% to 4% of affected patients.^37–39

Garman et al.²⁷ used Genome-Wide Association Study (GWAS) to compare OS in European- (EA) and African-American (AA) cohorts to healthy controls so as to identify HLA and non-HLA variants significantly associated with this condition. Both cases and controls were identified in one of three existing sarcoidosis cohorts: A Case Control Etiologic Study of Sarcoidosis (ACCESS) Group, the Sarcoidosis Genetic Analysis (SAGA) study, and the Henry Ford Family Study (HFFS). The GWAS analysis was run on 1271 AA sarcoidosis cases, 1551 AA controls, 332 EA sarcoidosis cases, and 2046 EA controls. Among the sarcoidosis cases, 260 AA (20.5%) and 35 EA (10.6%) subjects had ocular involvement. Statistically significant associations on GWAS were defined as those that achieved p < 5 × 10⁻⁸. The authors identified seven novel OS associations in the Membrane Associated Guanylate kinase with Inverted structure 1 (MAGI1) gene, which encodes for a scaffolding protein that appears to be integral to tight junction and epithelial barrier integrity, and which regulates both type 1 interferon and several immune regulating transcriptions factors, including BLIMP-1, STAT1, and NFkB. The authors also confirmed an association between HLA-DRB1*04:01 and OS in EA, and identified an association between HLA-DRB1*03:01, *11:01, and *12:01 and non-ocular sarcoidosis in AA. They suggested that further studies into the role of MAGI1 and epithelial barrier function in OS may be warranted. Of note, there is precedent to support the concept that in a multisystem disease, genes might affect the likelihood that a specific organ is affected. For example, Robinson and colleagues⁴⁰ found genes specifically associated with uveitis, but not joint disease in HLA B27-associated anterior uveitis. Although not discussed directly in this study or the previously cited report by Anukul et al,²⁵ it is worth noting that broad HLA testing is not generally recommended to aid in syndromic diagnosis of uveitis patients. Only HLA-B27⁴¹ and HLA-A29⁴² have sufficient positive predictive value to be used to support the diagnosis of HLA-B27-associated uveitis and birdshot chorioretinopathy (BSCR), respectively, and even then, only when the clinical presentation and findings suggest the diagnosis.

Babu et al²⁸ described seven subjects with Blau syndrome, a rare systemic autoinflammatory disorder characterized by arthritis, dermatitis and uveitis, all of whom were seen at a tertiary referral center in Bangalore, India. All cases were genetically confirmed to be associated with mutations in the gene encoding Nucleotide-binding Oligomerization Domain containing 2 (NOD2), including two de novo mutations and five inherited in an autosomal dominant fashion. Ages ranged from three to 25 years, with a median of 10 years. Four of the subjects were female. All patients had a history of early onset skin rashes, arthritis and uveitis, and were antinuclear antibody (ANA) negative. Ocular manifestations included keratoconjunctivitis sicca (6/6; 100.0%), conjunctival granulomas (3/6; 50.0%), and either granulomatous anterior (3/6; 50.0%) or panuveitis (3/6; 50.0%). Common longer-term ocular complications included band keratopathy (5/6; 83,3%) and cataract (3/6; 50.0%). All patients were treated successfully with a combination of oral and topical corticosteroids together with a corticosteroid-sparing immunosuppressive agent, typically methotrexate. The authors encouraged consideration of Blau syndrome in young patients who present with granulomatous uveitis, particularly when accompanied by rash and arthritis.

Li et al²⁹ examined six known single nucleotide polymorphisms (SNPs) in the Caspase Recruitment Domain Family Member 9 (CARD9) gene from 480 subjects with BD, 1151 with acute anterior uveitis (AAU), and 1440 healthy controls in China. The authors chose CARD9 as it is known to play a key role in the regulation of both innate and adaptive immunity, and alterations in the CARD9 gene have been linked to both primary immunodeficiencies and inflammatory disorders in humans.⁴³ While none of the six individual SNPs showed an association with either BD or AAU, a haplotype analysis did reveal a significant decrease in a specific 5-SNP CARD9 gene haplotype in BD compared to healthy controls (p = .012; OR = 0.59, 95% CI = 0.41 to 0.84), leading the authors to suggest that this particular haplotype might be protective for BD in their Han population. No haplotype combinations were associated with AAU.

Nieves-Moreno et al³⁰ described recurrent, giant chalazion formation as a complication in two separate children with unrelated PIDs: X-linked CGD (Bridges-Good Syndrome) associated with a hemizygous mutation in CYBB gene in a four-year-old boy, and hyperimmunoglobulin E syndrome (Job syndrome) due to a de novo mutation in the STAT3 gene in a five-year-old girl. Both patients had multiple, recurrent childhood infections. Bargir et al^31,32 commented on the report to remind us that children with giant chalazion formation, recurrent or unusual infections, and/or unusual granuloma formation should be evaluated for PIDs.

Together, these studies highlight the growing importance genetics plays in the development, characteristics, and management of uveitis. While testing for most identified HLA and non-HLA risk polymorphisms plays little role in the day-to-day diagnosis or management of uveitis, HLA-B27 and HLA-A29 tests are recommended and used routinely in patients suspected, based on demographic and clinical findings, to have HLA-B27-associated uveitis and BSCR, respectively. An IRD-associated uveitis should be suspected when there is a relevant family history or either symptomatic or anatomical evidence of widespread photoreceptor degeneration, such as nyctalopia or diffuse outer retinal atrophy on optical coherence tomography. Electrophysiological and genetic testing can be useful in such settings to confirm the diagnosis.⁴⁴ One of several genetically defined PIDs²¹ should be considered in those who develop uveitis with unexplained granulomatous features, or due to otherwise opportunistic CMV retinitis or bacterial or fungal endophthalmitis, particularly in young patients and when there is a history of multiple prior cutaneous, oral, or pulmonary infections or systemic granuloma formation. Here too, genetic testing can be fruitful.^45,46

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Financial conflicts

ETC, FP, and VBM have no relevant financial conflicts. JTR is a consultant to AbbVie, Gilead, Roche, Novartis, Horizon, Revolo, Corvus, Kyverna, Santen, and Affibody. He receives research support from Horizon and Pfizer, and receives royalties from UpToDate. MZ is a consultant to AbbVie, Alimera, Santen and Gilead.