Anti-Retinal Antibodies in Patients With Macular Telangiectasia Type 2

Abstract

Purpose.: Macular telangiectasia type 2 (MacTel-2) is a retinal disease that can cause loss of central vision. To gain better understanding of the etiology and pathogenesis of MacTel-2, we investigated antigens that prompt the generation of retinal autoantibodies in the serum of patients with MacTel-2.

Methods.: We screened for the presence of retinal autoantibodies in 45 serum samples collected from patients with MacTel-2 and 58 serum samples from healthy control subjects by Western blot. We then isolated and identified three retinal proteins that are putative targets of three of the most frequently detected autoantibodies in the serum of patients with MacTel-2 using chromatographic fractionation and liquid chromatography coupled to tandem mass spectrometry. We also validated the retinal location of the three antigens by immunohistochemisty using MacTel-2 sera as primary antibodies and commercial antibodies.

Results.: Retinal autoantibodies were detected in a significantly higher proportion of patients with MacTel-2 than in controls (31 of 45 [69%] vs. 9 of 58 [16%], P < 0.0001). The three antigens that were targeted by the most frequently detected MacTel-2 autoantibodies were identified as glycogen debranching enzyme (hereafter AGL, named for the gene symbol AGL), retinol-binding protein 3 (RBP3), and creatine kinase type B (CK-B); autoantibodies against these antigens were found in four, eleven, and nine MacTel-2 serum samples, respectively.

Conclusions.: We found that most patients with MacTel-2 possess retinal autoantibodies, the most prevalent of which were directed against AGL, RBP3, and CK-B. The localization of retinal proteins bound by AGL, RBP3, and CK-B autoantibodies is consistent with their putative physiological functions. These findings provide potentially novel mechanisms for the etiology and pathogenesis of MacTel-2.

Introduction

The retina is thought to be an “immune privileged” site, which allows it to tolerate the introduction of antigens without eliciting an inflammatory immune response. This may protect the retina from inflammatory damage that could have potentially disastrous consequences for vision because the retina has very limited regenerative capacity. ¹  

The blood–retina barrier (BRB) is the interface between systemic circulation and the retina. It consists of nonfenestrated capillaries of the retinal circulation with tight junctions between retinal vascular endothelial and retinal pigment epithelium (RPE) cells. The ability of the BRB to prevent free entry and exit of cells and larger molecules into and out of the retina is critical for the maintenance of immune privilege. Breakdown of the BRB is an important feature of many retinal diseases, including macular telangiectasia type 2 (MacTel-2), age-related macular degeneration (AMD), and diabetic retinopathy (DR). ^2–4 Breakdown of the BRB exposes retinal antigens to the immune system, eliciting an inflammatory response that can cause tissue damage and vision loss. ⁵ Autoantibodies against retinal antigens have been detected in a number of ocular disorders, including autoimmune retinopathy ^6,7 and AMD. ^8–10  

MacTel-2 is a potentially blinding condition of the retina, yet its etiology and pathological mechanisms remain poorly understood. ⁴ Although it has been reported that MacTel-2 is present in 0.1% of the population older than 43 years, the actual prevalence may be higher. ^4,11 MacTel-2 affects both eyes, with a predilection for the temporal parafovea. ¹² Leak is seen on fluorescein angiography, but cystic edema is usually absent. Biomicroscopic features of MacTel-2 include right angle vessels and deep retinal neovascularization, loss of macular transparency, superficial white crystals, inner retinal cavitation without retinal thickening, foveal thinning, and photoreceptor atrophy. 

To further explore this retinal disease, retinal autoantibodies were profiled in serum of patients with MacTel-2, and the results were compared with those obtained from healthy control subjects. We first screened for the presence of autoantibodies in serum of patients with MacTel-2 and then isolated the target antigens through chromatography, which were subsequently subjected to liquid chromatography coupled to tandem mass spectrometry (LC–MS/MS) for protein identification. We then performed immunohistochemical studies on human retinas using MacTel-2 sera as primary antibodies to identify the retinal cell types that were bound by the target autoantibodies. These results were further validated using commercially available antibodies. 

Materials and Methods

Patient Information and Serum Sample Collection

The study was approved by the Human Research Ethics Committee of the South Eastern Sydney Local Health District (Northern Sector). Written informed consent was obtained from all participants who donated serum. This research adhered to the tenets of the Declaration of Helsinki. 

Peripheral blood samples were collected from participants in the Natural History Observation Study of MacTel-2 (NHO MacTel-2). ¹³ The inclusion and exclusion criteria for enrollment in the NHO MacTel-2 were a clinical diagnosis of bilateral MacTel-2, no DR or only background DR (<10 microaneurysms without retinal hemorrhage), and no other confounding ocular conditions that might have prevented the evaluation of the condition (Table 1). 

Table 1

 

Patient Information and Autoantibodies Present

Group	n	Mean Age, y	Age Range, y	Ratio of Men to Women	Ratio of Diabetes to Nondiabetes	Autoantibiodies Present, n (%)
Control	58	45	21–78	28:30	13:45	9 (16)
MacTel-2	45	59	37–79	18:27	11:34	31 (69)

The final diagnosis of MacTel-2 was confirmed by the Moorfields Reading Centre at Moorfields Eye Hospital, London, UK, on the basis of color fundus photographs, optical coherence tomography (OCT) scans, fundus fluorescein angiography, fundus autofluorescence, blue light reflectance, and microperimetry. The diagnosis was confirmed in all cases by attenuation of central masking of background autofluorescence by luteal pigment with other signs, including temporal parafoveal telangiectasis with late staining seen on fluorescein angiography, superficial white crystals, foveal thinning, inner retinal cavitation, and photoreceptor disruption seen on OCT. Breakdown of the BRB was graded by an experienced retinal clinician (MCG) on the extent of leakage during angiography 5 minutes after fluorescein injection as absent, mild, moderate, or severe. 

For healthy controls, blood samples were collected from a group of persons who were enrolled in the NHO MacTel-2 study as blood relatives or non–blood-related controls of patients with MacTel-2. The Moorfields Reading Centre confirmed that the controls had neither MacTel-2 nor other retinal diseases based on their clinical signs and the ocular images. 

Retina Tissue Processing

Five postmortem human retinas (age range, 62–75 years) were obtained from the Lions NSW Eye Bank in accordance with the Australian Therapeutic Goods Administration guidelines for the use of human tissues for research. The ethics approval for the use of postmortem human tissues for this project was obtained from The University of Sydney Ethics Committee. 

Human retinas were homogenized in ristocetin-induced platelet agglutination (RIPA) buffer with protease inhibitor complex (Roche, Dee Why, Australia). The homogenates were constantly shaken for 30 minutes (60 rpm at 4°C) and then centrifuged at 12,000g for 30 minutes at 4°C. Supernatants were collected, pooled, and stored at −80°C until use. Protein concentration was determined using BCA protein assay (Sigma, Castle Hill, Australia). Total retinal protein homogenates were mixed with NuPAGE loading dye (4×) and reducing buffer (10×) (Invitrogen, Mount Waverley, Australia) and then heated for 10 minutes at 70°C. The supernatant was collected after centrifugation at 12,000g for 5 minutes. 

Autoantibody Screening

Approximately 1 μg of protein was loaded per 1 mm of the NuPAGE Tris-Bis gel (Invitrogen). The precast gels were manually modified to include a single small well for loading the protein standard and a large “prep” well (∼6.5 mm in length), allowing the total retinal protein sample to be applied across the whole gel. Proteins were separated by electrophoresis at constant 180 V at 4°C for 1 hour. The gels were transferred onto polyvinylidene difluoride (PVDF) membrane (Millipore, North Ryde, Australia) at constant 100 V at 4°C for 1 hour. 

The PVDF membranes were incubated in 5% BSA, 0.1% Tween, and Tris-buffered saline and Tween 20 (TBST) for 1 hour at room temperature to block nonspecific binding. Serum from each participant was diluted (1:200) and reacted with the membrane overnight at room temperature. A multiscreen apparatus (Bio-Rad, Gladesville, Australia) allowed the large area of blotted protein to be divided into individual chambers; therefore, a single blot could be used to test sera from up to 18 individuals. Each chamber had a total volume of 600 μL. 

After overnight incubation in sera, membranes were washed in 0.1% TBST three times and then incubated with goat anti-human IgGAM (1:10,000; Invitrogen) at room temperature for 1 hour. This polyclonal antibody recognizes both heavy and light chains of human IgG, IgA, and IgM. After three washes in 0.1% TBST, the blot was incubated in electrochemiluminescence Western blot (WB) detection reagent (Millipore) for 5 minutes and photographed using the G-Box imaging system (In Vitro Technologies, Lane Cove, Australia). 

Chromatography Fraction

Total human retinal protein was extracted by RIPA buffer as described above. The dialysis was performed in buffer A (50 mM Tris, pH 8.0) and then loaded into an ion exchange column (HiTrap Q HP, 1 mL; GE Healthcare, Rydalmere, Australia). Proteins were eluted out based on their different surface charges. Different elution fractions of 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 100% of buffer B (1 M NaCl, 50 mM Tris buffer, pH 8.0) were collected and subjected to WB for tracking the antigen, in which sera from patients with MacTel-2 were used as primary antibodies. 

The antigen-containing fraction was further dialyzed into 1 M ammonium sulfate and 50 mM potassium phosphate buffer (pH 7.5) and then loaded onto a hydrophobic interaction chromatography column (HiTrap Butyl HP; GE Healthcare). Proteins were eluted out based on their different hydrophobicity. Different elution fractions were collected and subjected to WB to track the antigens as described above. 

The antigen-containing fractions were dialyzed to exclude the salt and then concentrated. Fractions were run in parallel on two identical 4% to 12% NuPAGE Tris-Bis gels (Invitrogen) at a constant 180 W for 1 hour. The first gel was processed for WB using MacTel-2 sera to identify the exact position of the target antigen. Protein bands were stained in the second gel with SimplyBlue SafeStain (Invitrogen). The desired antigen-containing bands in the second gel were identified using the corresponding migration distance of the target proteins in the membrane subjected to WB. The antigen-containing bands in the second gel were excised, destained, and dried before being digested overnight using trypsin (Promega, Alexandria, Australia). Digests were extracted using 60% acetonitrile (ACN)–0.1% trifluoroacetic acid. The extracts were dried with a speed vacuum and reconstituted in 5% ACN–0.1% formic acid and applied to reverse-phase NanoLC mass spectrometry (Australian Proteome Analysis Facility, Sydney, Australia). 

Tandem Mass Spectrometry Data Acquisition

The digests were dissolved in 40 μL of electrospray ionization loading buffer, injected onto a peptide trap (CapTrap; Michrom, Auburn, CA) for preconcentration, and desalted with 0.1% formic acid–2% ACN at 8 μL/min. The peptide trap was switched in line with the analytical column. Peptides were eluted from the column using a linear solvent gradient stepwise from H₂O:CH₃CN (100:0 + 0.1% formic acid) to H₂O:CH₃CN (10:90 + 0.1% formic acid) at 500 nL/min over 80 min. The LC eluent was analyzed with positive ion nanoflow electrospray MS on QSTAR (AB SCIEX, Mt. Waverley, Australia), which was operated in an information-dependent acquisition mode (IDA). In IDA mode, a time-of-flight MS survey scan was acquired (ratio of mass to charge [m/z] 400–1600, 0.5 seconds) with the three largest multiply charged ions (counts >25) sequentially subjected to tandem mass spectrometry analysis. The MS/MS spectra were accumulated for 2 seconds (m/z 100–1600). 

Data Processing

The data were exported and submitted to the database search program Mascot (Matrix Science Ltd., London, UK). Peak lists were searched against Homo sapiens in the Swiss-Prot database (the European Molecular Biology Laboratory, Heidelberg, Germany, and the Swiss Institute of Bioinformatics, Lausanne, Switzerland). High scores in the database search indicate a likely match, which is confirmed or qualified by operator inspection. 

Immunohistochemistry

Two postmortem human eyes from healthy donors were used for immunohistochemical studies. The approximate time from the death of the patient to storage of the globe was 12 hours. After removing corneas, eyecups were fixed in 4% paraformaldehyde for 4 hours, followed by equilibration in 20% sucrose overnight. After dissecting eyecups into smaller pieces, tissues were embedded in optical cutting temperature compound (ProSciTech, Thuringowa, Australia) for cryosectioning. 

To identify the types of retinal cells targeted by autoantibodies detected in MacTel-2, cryosections (10 μm) were incubated with 1:50 dilutions of MacTel-2 sera overnight at 4°C. Antibody binding was detected by incubation of sections with goat anti-human IgG (H+L) conjugated with AlexaFluor 488 (1:1000; Invitrogen) at 37°C for 2 hours. The nuclei were counterstained with 10 μg/mL of propidium iodide in PBS for 5 minutes. The stained sections were examined with a confocal microscope (LSM700; Carl Zeiss, North Ryde, Australia). 

To validate the results of immunohistochemistry (IHC) in which MacTel-2 sera were used as primary antibodies, cryosections were also incubated with commercial antibodies against proteins revealed by tandem mass spectrometry data acquisition. The commercial antibodies used for further validation included anti–creatine kinase type B (anti–CK-B) (sc-1517; Santa Cruz, Dallas, TX), anti–retinol-binding protein 3 (anti-RBP3) (ab101456; Abcam, Cambridge, MA), and anti–glycogen debranching enzyme (hereafter anti-AGL, named for the gene symbol AGL) (C-term, AP2402b; Abgent, San Diego, CA). Immunohistochemistry was performed as described above. 

Results

Retinal Autoantibody Detection

Forty-five blood samples were collected from patients with confirmed MacTel-2 and 58 blood samples from controls. To detect autoantibodies present in MacTel-2 sera, we performed WB analysis using sera as primary antibodies to probe the total proteins extracted from human retinas. Various band patterns were detected, especially in membranes blotted with MacTel-2 sera (Fig. 1A). Overall, we detected autoantibodies against human retinal proteins in 69% (31 of 45) of MacTel-2 serum samples, while only 16% (9 of 58) of controls showed positive bands (P < 0.0001) (Table 1). Fisher exact test gave an odds ratio of 12.1, indicating that the likelihood of patients with MacTel-2 having retinal autoantibodies is significantly greater than that among controls. 

Figure 1

 

Retinal autoantibody detection by WB. (A) Total human retinal protein blotted with serum from an individual patient with MacTel-2 or control serum. Bands indicate different retinal autoantibodies in MacTel-2 sera. (B) Three different sizes of bands were detected by WB. Antigen x (∼150 kDa), antigen y (∼120 kDa), and antigen z (∼40 kDa) were frequently observed in MacTel-2 sera (asterisks) but were not present in controls.

Three bands were frequently detected in MacTel-2 sera that did not appear in sera from controls. We named these three unknown proteins antigen x (∼150 kDa), antigen y (∼120 kDa), and antigen z (∼40 kDa) (Fig. 1B). Autoantibodies against antigens x, y, and z were found in four, eleven, and nine MacTel-2 serum samples, respectively. In addition, we detected low-molecular-weight (low MW) (<40 kDa) autoantibody bands in seven MacTel-2 serum samples, but they were not detected in multiple patients. Not surprisingly, 16% (9 of 58) of control serum samples also contained low MW autoantibody bands when probing against human total retinal proteins. No samples shared a MW similar to that of antigens x, y, or z. This is consistent with a study ¹⁴ identifying a certain percentage (19% [15 of 79]) of low MW autoantibodies detected in control serum samples. 

Reviewing the clinical data of all these patients having Mactel-2 with antigens x, y, or z revealed that most patients had advanced MacTel-2. Two of eleven patients with antigen y were observed to have retinal hyperautofluorescence, which occurs in only approximately 5% of the general population with MacTel. ¹⁵ Patients with antigen y exhibited more pronounced BRB breakdown than other cases (data not shown). 

Antigen Fractionation and Identification

To identify the three retinal proteins targeted by the three autoantibodies that were detected in MacTel-2 sera, chromatography was used to fractionate the human retinal proteins. As described in the Materials and Methods section, we first separated human retinal protein lysates by surface charge using an ion exchange column (Q column) and then chose the fractions containing our target antigens to be further isolated by hydrophobicity using a hydrophobic column (butyl column). Each isolation step was validated by WB using MacTel-2 sera containing the autoantibody against antigen x, y, or z. (Figs. 2A–C, asterisks). The antigen-containing protein fraction was then run in parallel on two identical SDS-PAGE gels. The first gel was used to identify the exact position of the target antigen by WB using MacTel-2 serum containing target autoantibody. The desired antigen-containing band stained by SimplyBlue in the second gel was identified using the corresponding migration distance of the target protein in the membrane subjected to WB. The antigen-containing band in the second gel was excised (Figs. 2D–F, asterisks), purified, and analyzed with LC–MS/MS. 

Figure 2

 

Fractionation and separation of target antigen. (A–C) Chromatographic separation of antigens x, y, and z from total human retinal proteins identified by autoantibody detection through WB. Asterisks indicate fractions containing target antigens. (D–F) The SDS-PAGE gels further separated the target antigens. Asterisks indicate the target bands cut out for MS.

As summarized in Table 2, the LC–MS/MS analysis revealed antigen x as AGL (a protein related to energy metabolism ¹⁶ ), antigen y as RBP3 (which is known to have important roles in the visual light cycle ¹⁷ ), and antigen z as CK-B (which is a factor involved in energy metabolism in the visual cycle ¹⁸ ). The LC–MS/MS analysis detected multiple matching sequences consistent with our predicted target proteins, which confirmed our earlier analysis (see Supplementary Fig. S1 for details of the LC–MS/MS results for identification of the three antigens). 

Table 2

 

Detailed Information About the MS/MS Identified Target Antigens

Target Peptide	Protein Name	Organism Name	Mass, Da	Protein Score	Matches	Sequences	Exponentially Modified Protein Abundance Index
Antigen x	AGL	H. sapiens	174,652	301	17 (8)	17 (8)	0.28
Antigen y	RBP3	H. sapiens	135,278	2746	83 (61)	45 (43)	2.84
Antigen z	CK-B	H. sapiens	42,617	8314	164 (153)	27 (27)	25.72

Antigen Localization

Immunohistochemistry was conducted using MacTel-2 sera and commercial antibodies against AGL, RBP3, and CK-B to localize the target cell types in human retinas (Figs. 3, 4, Table 3). Immunohistochemistry using MacTel-2 serum containing anti-AGL as a primary antibody produced strong immunostaining around retinal blood vessels in the inner retina, along with weak staining in astrocytes and cells in both the inner nuclear layer (INL) and outer nuclear layer (ONL) (Figs. 3A–C). Immunohistochemistry using a commercial anti-AGL antibody produced a similar immunoreactivity pattern (Figs. 4A–C). Immunohistochemistry with MacTel-2 serum containing anti-RBP3 showed that RBP3, a molecule responsible for the transfer of 11-cis retinol and all-trans retinol between the photoreceptors and RPE, ¹⁹ was abundantly expressed in the photoreceptor outer segments (Figs. 3D–F). This finding was further confirmed by immunostaining using a commercial anti-RBP3 antibody (Figs. 4D–F). Immunohistochemistry with MacTel-2 serum containing anti–CK-B showed that CK-B was globally expressed throughout the entire retina and that high CK-B expression was observed in the photoreceptor outer segments (Figs. 3G–I). Immunostaining using a commercial anti–CK-B antibody revealed staining results similar to those observed with CK-B serum (Figs. 4G–I). To further confirm the identity of antigen z as CK-B, we blotted patient serum against a recombinant CK-B protein. Sera from all nine patients were positive for recombinant CK-B protein compared with control sera (Supplementary Fig. S1). There was no obvious immunoreactivity in the human retina when control sera were used as primary antibodies for IHC (Figs. 3J–L). Thus, these IHC results indicate that patients with MacTel-2 can produce autoantibodies to target specific retinal proteins. 

Figure 3

 

Immunohistochemistry in human retinas using MacTel-2 sera containing different autoantibodies identified by chromatographic fractionation and tandem mass spectrometry. Healthy subject serum was used as a control. Sections were also stained with propidium iodide (PI) for nuclear counterstaining. (A–C) MacTel-2 serum containing anti-AGL autoantibody. (D–F) MacTel-2 serum containing anti-RBP3 autoantibody. (G–I) MacTel-2 serum containing anti–CK-B autoantibody. (J–L) Control serum without autoantibody. GCL, ganglion cell layer, Scale bar: 60 μm.

Figure 4

 

Immunohistochemistry using commercial antibodies against AGL (A–C), RBP3 (D–F), and CK-B (G–I). Sections were also stained with propidium iodide (PI) for nuclear counterstaining. GCL, ganglion cell layer, Scale bar: 60 μm.

Table 3

 

Immunohistochemistry in Human Retinas Using MacTel-2 Sera as Primary Antibodies and Confirmation of the Results Using Commercial Antibodies

Protein Name	Cellular Localization in the Human Retina	Function
AGL	Strong staining around inner retinal vessels, weak staining in astrocytes and cells in the INL and ONL	Glycogen metabolism, facilitating glycogen breakdown
RBP3	Photoreceptor outer segments	Visual cycle trafficking, responsible for transfer of 11-cis retinol and all-trans retinol between photoreceptors and RPE
CK-B	Globally expressed throughout the entire retina, high expression was observed in the photoreceptor outer segments	Energy metabolism, providing energy for the visual cycle

Discussion

This study is the first to date to report the presence of autoantibodies against retinal proteins in sera from persons with MacTel-2. We identified three retinal proteins, AGL, RBP3, and CK-B, that appear to be the targets for many of these autoantibodies. These data may serve as a foundation for further studies aimed at elucidating the pathogenic mechanisms that cause MacTel-2. 

AGL is involved in glycogen metabolism through facilitating glycogen breakdown. ²⁰ Immunohistochemistry revealed that AGL was found mainly around the inner retinal vessels, with weak expression in astrocytes and cells of both the INL and ONL. It has been reported that glycogen may be used by Müller cells to generate lactate for neighboring cells. ^21,22 Our previous proteomics study ²³ demonstrated downregulation of glycolytic pathway proteins in the retina of a patient with MacTel-2. The glycolytic pathway is dominant in glial cells, including astrocytes in the superficial retina and Müller cells that span the full thickness of the neural retina. Compromise of this pathway may affect the energy supply of photoreceptors. This might be exacerbated by the presence of AGL autoantibodies in patients with MacTel-2. Autoantibodies against AGL have not been reported to date for any other retinal disease. Whether anti-AGL autoantibody could be used as a biomarker unique to MacTel-2 warrants further investigation. 

RBP3 is essential for the exchange of retinoid between the RPE and photoreceptors. The presence of sufficient quantities of functional RBP3 is also critical to photoreceptor survival because it prevents the potentially cytotoxic effects of retinoids. We detected RBP3 autoantibody in approximately 24% (11 of 45) of patients with MacTel-2, and our IHC studies found that these autoantibodies bound photoreceptor outer segments strongly. Intriguingly, we also found RBP3 autoantibody in 33% (6 of 18) of serum samples from patients with AMD (Zhu L, unpublished data, 2013). Detection of a common autoantibody might indicate that both MacTel-2 and AMD share some common etiologic or pathogenic mechanisms. 

It is possible that RBP3 autoantibodies could contribute to outer retinal degeneration. Reduced levels of functional RBP3 protein in the interphotoreceptor matrix might interfere with the visual cycle that represents the natural exchange of proteins between the photoreceptors and RPE. This disruption could result in the accumulation of all-trans retinol, which is readily transformed through condensation to N-retinylidene-N-retinylethanolamine, ²⁴ a prominent component of the retinal “waste product” lipofuscin. This component is also known to be responsible for retinal autofluorescence, a phenomenon that is a standard AMD diagnostic marker. ^25,26 In a small proportion of patients with MacTel-2, increased autofluorescence due to accumulation of subretinal debris has also been observed. ^15,27  

CK-B is an important enzyme for vertebrate energy metabolism. ¹⁸ It catalyzes the conversion of creatine, consuming adenosine triphosphate, phosphocreatine, and adenosine diphosphate in the visual cycle. ¹⁸ CK-B is critical for providing energy for the visual cycle in photoreceptors. CK-B autoantibody has been detected in 25% of persons with Vogt-Koyanagi-Harada disease and in 38% of those with sarcoidosis, ²⁸ prevalences that are similar to the 20% (9 of 45) that we found in patients with MacTel-2. Notably, sera from patients with DR, paraneoplastic sensory-dominant neuropathy, and nonparaneoplastic autoimmune retinopathy have also been reported to contain autoantibodies binding to CK-B. ^29–31 We found CK-B antibodies globally expressed throughout the retina, with a particularly high expression in photoreceptor outer segments. Because CK-B is abundantly expressed in the retina, ¹⁸ the presence of CK-B antibody in patients with MacTel-2 may adversely affect their visual function. After reviewing the clinical data of all nine patients whose sera contained CK-B antibody, we found that all these patients exhibited more pronounced BRB breakdown, determined angiographically, than individuals from the set of persons with MacTel-2. This suggests that the presence of CK-B autoantibody in these patients may contribute to retinal vasculopathy and may have clinical significance for the early diagnosis of retinal vascular diseases such as MacTel-2 and DR. 

A potential mechanism underlying the generation of retinal autoantibodies is the activation of microglia by the stressed retina. Activated microglia may phagocytose target retinal antigens and then enter the systemic circulation, where they may become antigen-presenting cells. ³² This could result in a vicious circle, whereby more autoantibodies pass the compromised BRB, leading to an increase in phagocytosed antigens. This would not only impair the function of these antigens but also damage surrounding areas because of inflammation. 

MacTel-2 remains poorly understood both etiologically and pathologically. The presence of different autoantibodies in persons with MacTel-2 suggests that the clinical picture might be produced by different pathogenic processes. For instance, autoantibodies against AGL may contribute to impairment of the glycolytic pathway. RBP3 autoantibodies might contribute to photoreceptor damage; most of the patients having MacTel-2 with this autoantibody were at a more advanced pathogenic stage. CK-B autoantibody was detected in patients having MacTel-2 with more pronounced vasculopathy. Clinically, there is great variation in the degree of neuronal or vascular damage from one patient to another. Further studies are warranted on the relationships among different autoantibodies, clinical phenotypes, and progression rates. 

In summary, we found that most patients with MacTel-2 possess retinal autoantibodies, the most prevalent of which were directed against AGL, RBP3, and CK-B. The localization of retinal proteins bound by AGL, RBP3, and CK-B autoantibodies was consistent with their putative physiological functions. These findings provide potentially novel mechanisms for the etiology and pathogenesis of MacTel-2. Further studies will be required to determine whether these autoantibodies have a causative role in photoreceptor damage and BRB breakdown or whether they are simply a secondary phenomenon that is a nonspecific reaction to retinal disease. 

Supplementary Materials

pdf S1 

pdf S2 

pdf S3 

pdf S4 

Acknowledgments

Supported by a grant from Lowy Medical Research Institute, Sydney, Australia. Mark C. Gillies is a fellow of Sydney Medical School Foundation and is supported by a National Health and Medical Research Council (Australia) Practitioner Fellowship. The MS/MS analysis was undertaken at the infrastructure of Australian Proteome Analysis Facility (APAF) which was sponsored by the Australian Government through the National Collaborative Research Infrastructure Strategy (NCRIS). 

Disclosure: L. Zhu, None; W. Shen, None; M. Zhu, None; N.J. Coorey, None; A.P. Nguyen, None; D. Barthelmes, None; M.C. Gillies, None