Prevalence of anti-retinal autoantibodies in different stages of Age-related macular degeneration

Age-related macular degeneration (AMD) is the leading cause of central vision loss in older adults [1, 2]. The etiology of AMD appears to be diverse, including age, genetic predisposition, diet, smoking, and other environmental risk factors [3, 4]. Considerable evidence supports a strong role for local inflammation including the accumulation of macrophages, lymphocytes, and mast cells found in association with both forms of late AMD: geographic atrophy (GA) and neovascular AMD [5‐9]. Large drusen deposits in the macula (intermediate AMD) generally precede development of the late stages of AMD. The composition of these immunologically active drusen deposits includes proteins, lipids, complement, and other substances that may act as triggers for immune responses in the eye [7, 10]. In addition, the activation of complement factors and secondary mediators of inflammation such as cytokines and chemokines has been demonstrated in serum from AMD patients [3, 11‐13]. These findings suggest the possible involvement of IL-22 and IL-17 in the inflammation that contributes to pathogenicity of AMD, and C5a may be one of the factors contributing to the elevated serum levels in AMD patients [14, 15].

There is increasing evidence for the presence of anti-retinal autoantibodies (AAbs) in association with AMD, although it is unclear whether such AAbs play an active role in the etiology of disease or if they are generated in a response to retinal injury from the underlying disease processes [16‐21]. It is possible that both the generation of AAbs as well as the activation of complement could be responses to retinal damage/degradation [21‐23]. Nevertheless, recent studies have demonstrated the presence of various AAbs in individuals with AMD, including anti-aldolase C, anti-pyruvate kinase isoform M2, anti-retinaldehyde binding protein 1, and anti-retinol binding protein 3 [16, 17, 20, 24]. The detection of AAbs could conceivably be important in subtyping the disease with specific antibody signatures (multiple antibody arrays), eventually helping define pathogenesis and optimum therapy. The goal of our study was to determine the AAb specificity in different stages of AMD, and prevalence of AAb signatures in relation to AMD severity.

The etiology of AMD is complex and includes genetic risk factors, environmental factors, age, and immune and autoimmune causes [30]. Several potential immune mediators are known to play a role in the pathogenic process of AMD, such as infiltration of macrophages, presence of cytokines/chemokines, T-lymphocytes, as well as formation of autoantibodies [1, 5]. Our studies demonstrated a complex pattern of AAbs against several retinal proteins in individuals with AMD with fairly constant presence over the severity spectrum of the disease. We also identified AAbs that could be markers of disease activity (Figures 3 and 4). We believe that early generation of anti-retinal AAbs may create an environment that promotes cell loss, angiogenesis, and progression to late stages of AMD [31]. Since in autoimmune diseases, the immune response is itself part of the disease process, it is possible that the consistent presence of AAbs suggests their contribution to pathogenic processes through the availability of the antigens from degenerating retina.

Previously published studies showed the association of AAbs with AMD, suggesting a contribution of the immune system to pathogenicity of the disease [7, 17, 19, 32, 33]. Several autoantibodies to retinal antigens have been identified, including anti-glial fibrillary acidic protein (GFAP), anti-α-enolase, and anti-carboxyethylpyrrole (CEP), an oxidized component of drusen [17, 32, 34, 35]. GFAP is a 52-kDa antigen, the main intermediate filament protein in mature astrocytes, that was targeted in 44% of the AMD patients’ population studied [17]; however, we have not found AAbs against 52-kDa antigen in a measurable level in our cohort. The study examined approximately the same number of patients as our study reported anti-enolase AAbs in 67% of the AMD patients, which are similar to our findings [17]. Other studies showed that AAbs in individuals with neovascular AMD recognized retinol-binding protein 3 (RBP3, 120-kDa), retinol-binding protein 1 (RLBP1, 36-kDa) and aldolase C (39-kDa) [20]. RBP3 is an essential protein for the exchange of retinoid between the RPE and photoreceptors and their survival because it prevents the potentially cytotoxic effects of retinoids. Autoantibodies against RBP3 were detected in 33% (6 of 18) of patients with AMD and also in 24% (11 of 45) of patients with Macular Telangiectasia Type 2 [36]. The authors suggested that the existence of mutual AAbs in MacTel-2 and AMD shared some common etiologic or pathogenic mechanisms for both conditions. RBP3, also known as interphotoreceptor retinoid binding protein (IRBP), is a unique protein to the photoreceptor cells in the retina, which has been found to be highly pathogenic in animals. Immunization of animals with IRBP induced an intraocular inflammatory disease that is primarily mediated by T cells but anti-IRBP antibodies are also present, suggesting their role in inflammation. Moreover, anti-IRBP autoantibodies and T cell have been also found in patients with uveitis, retinitis pigmentosa, and progressive rod-cone degeneration [37‐41]. Cellular retinaldehyde-binding protein (CRALBP), transcribed from the RLBP1 gene, is a 36-kDa protein found in the RPE and in retinal Müller cells [42]. Both elevated seroreactivity to RBP3 and RLBP1 in AMD patients suggests that the inflammation, in particular, autoimmunity, is strongly associated with the pathogenesis of the disease [11]. In our cohort, a 35-kDa antigen is likely to be CRALBP although its identity wasn’t confirmed (due to insufficient quantity of samples) if confirmed, these AAbs would be found to be 2 times more frequent in individuals with AMD than in controls. Nonetheless, anti-CRALBP AAbs have been found in uveitis, showing 54% seropositivity in patients with uveitis compared to 17% in normal subjects [43], they were also detected in patients with cancer-associated retinopathy, CAR (Adamus, unpublished information). Altogether, these findings indicated that anti-IRBP and anti-CRALBP AAbs are not highly specific for AMD individuals.

It is not surprising to find autoantibodies against α-enolase (46-kDa), aldolase C (40-kDa), and pyruvate kinase M2 (60-62-kDa) that previously were found in both neovascular and geographic atrophy AMD [17, 20]. These key enzymes of the glycolytic pathway may promote autoimmunity by acting as autoantigens. α-Enolase, also called non-neuronal enolase, belongs to a family of glycolytic enzymes but also has other cellular functions related to its subcellular localizations that are distinct from its well-established activity in glycolysis [44‐46]. Furthermore, differential expression of α-enolase and presence of specific AAbs have been related to several pathologies, such as cancer, Alzheimer’s disease, autoimmune diseases, and rheumatoid arthritis, among others. Antibodies against α-enolase have been strongly associated with CAR [25, 47‐49]. Recently, we showed that AAbs against anti-glycolytic enzymes were highly associated with CAR and gynecological cancers [50]. In vitro and in vivo studies showed that anti-enolase AAbs have pathogenic potential in killing retinal cells [51‐53]. Generation of AAbs against enolase and other anti-glycolytic enzymes may be a normal process since they can be found in healthy individuals. However, it is possible that excessive production of such autoantibodies can be generated as a consequence of enolase uptake by antigen-presenting cells and subsequent B cell activation, can potentially initiate a tissue injury as a result of immune complex deposition, or an induction of apoptosis leading to the death of retinal cells.

Autoantibodies against CEP, an adduct that develops from an oxidation fragment of docosahexaenoic acid (DHA), are present in plasma and are more abundant in AMD than in controls, suggesting their potential as a biomarker for AMD [32]. Animal studies showed that mice immunized with CEP-modified mouse serum albumin generated anti-CEP antibodies that consequently induced AMD-like lesions in the outer retina, suggesting that autoimmunity was associated with the initiation or progression of AMD [34]. CEP is typically present in photoreceptor rod outer segments, and RPE in the mouse retina, and its reactivity is more intense in photoreceptors of human AMD retina than healthy retina [32].

In AREDS, persons with either intermediate drusen, extensive small drusen, or the pigment abnormalities associated with AMD were more likely to be female, and more likely to have a history of arthritis [4]. This might suggest an association with chronic inflammatory disease. We found that women with arthritis and intermediate AMD had significantly greater incidence of AAbs than men, suggesting possible sex hormone effects. Changes in sex hormones mainly influence onset of rheumatoid arthritis in older individuals of both sexes [54]. Lower levels of testosterone in men, and early menopause in women, may be signs of premature aging, putting them at greater risk of developing diseases and thus antibodies.

Epidemiologic evidence indicates that smoking cigarettes results in an increased risk of AMD [27, 55]. Smokers of both sexes have an increased risk of developing seropositive rheumatoid arthritis (RA) [56]. For example, in RA, the presence of anti-citrullinated protein antibodies has been described to be specifically associated with smoking. Also, smoking has also been shown to be associated with AAbs in various other autoimmune diseases, such as anti-dsDNA in systemic lupus erythematosus, and anti-Jo1 in idiopathic inflammatory myopathy. There is evidence that citrullinated α-enolase is present in the lungs of smokers [57]. However, the precise mechanisms whereby smoking may trigger immunity to citrullinated α-enolase remain to be elucidated. Gender differences in our cohort were evident in AAb frequencies in never, current, and ever smokers with somewhat surprising results. Women who had never smoked presented with a higher prevalence of anti-retinal AAbs than women who were former smokers, which suggests a possible influence of sex hormones (e.g., estrogens). Estrogen strongly stimulate the immune response [58] and also can induce AAb levels in persons after quitting smoking [59]. Never-smoking men with AMD were protected from developing anti-retinal AAbs, in contrast to ex-smokers who had higher prevalence of AAbs. Testosterone plays an important role in the modulation of inflammatory processes and may neutralize the adverse changes of immune response [60]. Testosterone is known to promote apoptosis of Th2 cells and development of the Th1 phenotype of cytokine production, which results in suppression of humoral immune response. A fall of testosterone concentration in aging men can influence oxidative modification and the immune response, which is important in the pathogenesis of atherosclerosis.

The role of AAbs in the induction or acceleration of retinal deterioration is uncertain. Our studies showed that AAbs against retinal proteins are apparently different in various AMD severity subgroups. However, the differences in frequency of specific AAbs between AMD subgroups may suggest that they participate in pathogenicity of AMD. Also, different AAbs co-exist in individual patients, possibly creating antibody signatures specific to each stage of AMD. If confirmed by other studies, these elevated levels of specific antibodies may be a useful predictor or biomarker of AMD progression from intermediate to late AMD. On the other hand, it is reasonable to suggest that the high frequency of AAbs in persons with AMD as well as healthy individuals is a consequence of this progressive ‘aging’ of the immune system. However, pathogenic AAbs that can be detected in peripheral blood years before the destruction of retinal cells may lead to obvious clinical symptoms. Regardless of their role, AAbs may be useful as biomarkers and we presented a few possible candidates for each stage of AMD progression even though we could not provide the identity of the targeted antigens (except enolase). This was a limitation of this research because insufficient serum amounts did not allow for full identification of these retinal antigens. Further studies are necessary to confirm our results in the larger cohort of patients with AMD, and also to examine individual patients over the progression of their macular disease, and to identify the targeted antigens.