Photographs of the right (A) and left (B) optic nerve heads of the patient showing crisp margins and a completely normal salmon color.
Montages of retinal photographs taken of the right (A) and left (B) eyes of the patient. The optic nerve heads appear much more pale in these images than in Figure 1 because of the exposure settings required to show retinal details in this darkly pigmented fundus. There is arteriolar narrowing in both eyes, but it is much more pronounced in the right eye than in the left eye. There is an oval area of yellowish discoloration centered on the macula of the right eye and a much smaller lesion in the left eye just inferior to the fovea. There is a granular appearance to the entire fundus in the right eye, which is much less noticeable in the left eye. There is some perivenous hypopigmentation in both eyes, most noticeable in the left eye.
Higher-magnification color photographs of the superotemporal (A) and inferotemporal (B) aspects of the left fundus showing marked perivenous hypopigmentation.
Goldmann perimetry visual field of the left eye showing a near normal V4e isopter (magenta) and I4e isopter (blue) and moderate symmetrical constriction of the I2e isopter (red).
Aqueous soluble (S) and detergent-soluble (ie, insoluble [I]) fractions of human retina probed with the patient's serum or with secondary antibody only. Note the prominent γ-enolase band at approximately 47 kDa.
Probing of a polyvinylidene fluoride membrane containing retinal proteins separated by 2-dimensional gel electrophoresis with the patient's serum. Note the major reactive spot (arrow). The corresponding spot on a silver-stained gel was identified as γ-enolase. The approximate positions of the molecular weight markers based on the corresponding silver-stained gel are indicated to the right, and the approximate pH gradient is indicated at the top of the blot. pI indicates isoelectric point.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Ko AC, Hernández J, Brinton JP, et al. Anti–γ-Enolase Autoimmune Retinopathy Manifesting in Early Childhood. Arch Ophthalmol. 2010;128(12):1590–1595. doi:10.1001/archophthalmol.2010.295
JANEY L.WIGGSMD, PhD
To describe the clinical, molecular, and serologic findings of a case in which autoimmune retinopathy and early-onset heritable retinal degeneration were both considered in the differential diagnosis.
A 3-year-old girl had clinical findings suggestive of a childhood-onset retinal degeneration. Samples of DNA and serum were collected. The coding regions of 11 genes associated with Leber congenital amaurosis were sequenced. The patient's serum reactivity to soluble and insoluble fractions of human retinal protein was compared with that of healthy control subjects (n = 32), patients with inflammatory eye disease (n = 80), and patients with molecularly confirmed retinal degenerations (n = 11). Two-dimensional gel electrophoresis and mass spectrometry were used to identify a protein that corresponded to a reactive band on Western blot.
No plausible disease-causing mutations were identified in any of the retinal disease genes tested. However, the patient's serum showed reactivity to a single retinal antigen of approximately 47 kDa. Two-dimensional gel electrophoresis and mass spectrometry revealed the major reactive species to be neuron-specific enolase (NSE). Autoantibodies targeting NSE were not observed in any healthy control subjects or patients with inflammatory eye disease. However, anti-NSE activity was found in 1 child with molecularly confirmed Leber congenital amaurosis.
This patient's clinical and laboratory findings coupled with the recently discovered role of anti-NSE antibodies in canine autoimmune retinopathy suggest that autoantibodies targeting NSE are involved in the pathogenesis of her disease.
Infection or inflammation within the retina early in life may lead to an autoimmune phenocopy of early-onset inherited retinal degeneration.
Autoimmune retinopathy (AIR) is a pathogenic immunologic process in which circulating antibodies recognize normal retinal proteins and cause retinal degeneration. The retina is a relatively immune-privileged tissue that is somewhat isolated from the immune system by the blood-retina barrier and local inhibition of adaptive and innate immune cells.1,2 However, several mechanisms can lead to the development of autoantibodies that recognize retinal proteins. Exposure to certain microorganisms can result in the development of antibodies that recognize normal proteins in the retina.3 This mechanism is sometimes known as molecular mimicry. In addition, ineffective peripheral tolerance can result in inadequate suppression of autoreactive lymphocytes that recognize retinal proteins.4 These antiretinal antibodies and self-reactive lymphocytes can gain access to ocular tissues if the blood-retina barrier is disrupted through inflammation or trauma.5
Patients with AIR usually have symptoms of sudden visual field loss with a previously normal visual history.6 Although the fundus may initially look normal, visual deficits are often accompanied by severe electroretinographic abnormalities. A diagnosis of AIR is based on clinical evidence coupled with laboratory findings, particularly the finding of antiretinal antibodies.6 Autoantibodies to several retinal proteins have been associated with AIR, most commonly recoverin and α-enolase.6-9
The fundus appearance of a patient who has had AIR in the past can be similar to that seen in heritable retinal degenerations such as Leber congenital amaurosis (LCA) and retinitis pigmentosa. It seems likely that some patients, especially young children who are unlikely to report a sudden loss of peripheral vision, are misdiagnosed with an inherited photoreceptor degeneration when in fact their disease is of autoimmune origin. We report a case in which AIR and early-onset heritable retinal degeneration were both strongly considered in the differential diagnosis. On the basis of negative genetic testing results and positive serologic findings, we now believe this young patient to have an autoimmune phenocopy of an inherited retinal disease.
The patient was birthed by cesarean delivery after an uncomplicated pregnancy. There was no family history of impaired vision of any kind. Her growth, physical development, and cognition are completely normal. In the first few months of life, she was hospitalized for a severe febrile illness. No specific cause for this illness was found and she made a complete recovery. At age 3 years, she told her mother that she could not see when her left eye was covered. This prompted urgent visits to 2 of us (M.B.M. and E.M.S.). Her visual acuity was found to be hand motions OD and 20/20 OS. There was no strabismus or nystagmus, and no relative afferent defect was detected at that time. However, the patient's irides were so dark that the pupil responses were difficult to evaluate with penlight examination alone. Fundus examination revealed normal optic nerves (Figure 1) but significant retinal abnormalities in both eyes consisting of arteriolar narrowing, yellowish macular changes that were more prominent in the right eye than in the left eye, and some perivascular hypopigmentation (Figure 2 and Figure 3). There was no anterior segment or vitreous inflammation. A nonsedated electroretinogram was performed and was found to be nonrecordable in both eyes under all stimulus conditions. The patient's age precluded reliable assessment of her peripheral vision. Although the good central vision in the left eye and the absence of nystagmus were atypical for a congenital photoreceptor abnormality, the abnormal fundus examination results and extinguished electroretinogram in both eyes resulted in the provisional diagnosis of LCA. Blood samples were obtained for routine serologic studies and molecular analysis of all known LCA genes (described later). The serologic studies revealed only mild IgG reactivity to Toxocara, herpes simplex virus type 1, and rubella. These 3 serologic results were of sufficiently low titer that one could interpret them to reflect incidental contact with pets (Toxocara), community exposure (herpes simplex virus type 1), and immunization (rubella) rather than a cause of the patient's retinal disease.
The patient was reexamined at age 4 years; on that visit, infrared pupillometry revealed a 2.7–log unit relative afferent pupillary defect. Magnetic resonance imaging of the brain and orbits ruled out an intracranial cause. Dichromatic pupillometry10 revealed profound retinal dysfunction in the right eye but detectable retinal function in the left eye. The visual acuity and fundus findings on this visit were unchanged from those at her initial examination. Goldmann perimetry revealed no detectable peripheral vision in the right eye but quite well-preserved peripheral vision in the left eye (Figure 4). At this point, the asymmetry of the pupil responses, visual acuity, visual field, and fundus examination results caused us to consider the possibility of an inflammatory or autoimmune retinal insult as an explanation for her retinal findings. It was suspected that the symmetrical nonrecordability of the unsedated electroretinogram obtained at age 3 years was artifactual. However, given the clear asymmetry of her pupil responses and the overall stability of the clinical findings on this visit, we chose to defer an electroretinogram under anesthesia. Instead, an additional blood sample was obtained for antiretinal antibody studies (described later).
The patient was most recently seen at age 6 years, at which time her visual acuity, visual fields, and fundus examination results were all unchanged from those of her previous visits.
This study was approved by the Human Subjects Committee of the University of Iowa. Written informed consent was obtained from all subjects or their parents or guardians. All patients were examined at the University of Iowa by 1 or more of us, and these examinations included at a minimum visual acuity and intraocular pressure measurements, slitlamp biomicroscopy, and indirect ophthalmoscopy.
The coding sequences of 11 genes previously associated with LCA (CRB1, RDH12, GUCY2D, AIPL1, RPE65, CRX, RPGRIP1, LRAT, CEP290, RD3, TULP1) were screened for disease-causing mutations using bidirectional automated DNA sequencing as described previously.11 In addition, the gene encoding the retinal protein against which the patient serum showed reactivity (ENO2) was screened.
In addition to the samples described earlier in the case report, serum samples were also obtained from 11 patients with molecularly confirmed photoreceptor degeneration (9 with retinitis pigmentosa and 2 with LCA) and 80 patients with ocular inflammatory disease (10 with autosomal dominant neovascular inflammatory vitreoretinopathy, 7 with acute zonal occult outer retinopathy, 3 with birdshot chorioretinopathy, 14 with multifocal choroiditis, 2 with pars planitis, 29 with ocular histoplasmosis, and 15 with unspecified posterior uveitis). Blood was also collected from 32 control individuals (21 women and 11 men) who had normal, healthy retinas and no significant medical history of disease (no visual disturbances, diabetes, cancer, systemic disease, or inflammatory disease). Ten of the control individuals were between ages 18 and 30 years, 16 were between ages 31 and 60 years, and 6 were older than 60 years. Blood samples were collected in glass tubes without anticoagulants, allowed to clot, and centrifuged at 1000 g for 10 minutes. All sera were stored at −80°C until tested.
Human donor eyes were obtained from the Iowa Lions Eye Bank, Iowa City. Human retina was collected from the entire posterior pole from 3 donors without known ocular pathology within 6.5 to 17 hours post mortem. Each whole retina was homogenized in phosphate-buffered saline containing protease inhibitors (Roche complete kit; Roche Diagnostics Corp, Indianapolis, Indiana). The solution was centrifuged at 16 300 g for 15 minutes, and the aqueous soluble supernatant from this preparation was pooled from all 3 donors. The remaining pellets were then resuspended in 500 μL of phosphate-buffered saline with protease inhibitors and centrifuged for 15 minutes. The supernatants were discarded after each wash, and after 3 washes the pellets were resuspended in 175 μL of phosphate-buffered saline with protease inhibitors and with 1% Triton X-100. Pellets were then homogenized and centrifuged at 3000 g for 2 minutes. The resulting supernatants containing detergent-soluble retinal proteins were pooled from the 3 donors. Aqueous soluble and insoluble (ie, detergent-soluble) fractions were analyzed.
Soluble and insoluble retinal protein fractions were suspended in a solution containing 1X NuPAGE LDS sample buffer and 1X NuPAGE Reducing Agent (Invitrogen Corp, Carlsbad, California). Proteins were separated with a 1.5-mm NuPAGE 4% to 12% BIS-TRIS gel at 200 V. After electrophoresis, the proteins were transferred from the gel to a polyvinylidene fluoride membrane. The membrane was dried overnight and cut into 5-mm-wide strips that were wetted in methanol and blocked for 1 hour in 5% nonfat dry milk in phosphate-buffered saline. The strips were then each incubated with human serum at a dilution of 1:200 to 1:500 in 2% nonfat dry milk for 1 hour, rinsed with 1X TRIS-buffered saline with 0.1% Tween 20, and washed twice for 10 minutes in 1X TRIS-buffered saline with 0.1% Tween 20. For detection of autoantibody binding, strips were then incubated in a 1:30 000 dilution of horseradish peroxidase–conjugated goat-antihuman IgG/IgA/IgM antibody (Pierce, Rockford, Illinois) and washed 3 times for 10 minutes each in 1X TRIS-buffered saline with 0.1% Tween 20. The membranes were then developed using the ECL Plus Western Blotting Detection System (GE Healthcare, Milwaukee, Wisconsin). For some experiments, blots were probed with rabbit antienolase antibodies (H-300; Santa Cruz Biotechnology, Inc, Santa Cruz, California).
Whole soluble retinal protein (prepared with protease inhibitors and Triton X-100) from 1 donor without known ocular pathology was sent for 2-dimensional electrophoresis analysis at Kendrick Laboratories, Madison, Wisconsin. Ampholytes ranging from a pH of 3.5 to 10 were separated for 13.75 hours at 700 V for the first dimension, and the second dimension used 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Duplicate gels were run; one gel was silver stained, and proteins from the other gel were transferred to a polyvinylidene fluoride membrane. The membrane was probed with serum from the patient to identify the retinal protein bound by antibodies in the serum sample. The resulting film image revealed a single major reactive spot on the membrane. The corresponding silver-stained spot from the duplicate gel was subsequently excised and sent to Columbia University Protein Chemistry Core Facility, New York, New York, for identification via matrix-assisted laser desorption/ionization mass spectrometry.
Sera that showed reactivity to retinal proteins within 5 kDa of the molecular weight of neuron-specific enolase (NSE) were used to probe a Western blot of 117 ng of commercial purified human NSE (Lee Biosolutions, Inc, St Louis, Missouri).
Screening of the coding sequences of 11 genes known to cause early-onset retinal degeneration did not reveal any disease-causing mutations in the patient. However, the patient's serum intensely labeled a single major band with a molecular weight of approximately 47 kDa within the soluble fraction of retinal protein (Figure 5). The antibodies that reacted to this band included antibodies from IgG, IgA, and IgM isotypes. This band was not detected in the retinal aqueous insoluble fraction (Figure 5) or in protein samples of retinal pigment epithelium–choroid (data not shown).
To determine the identity of the reactive band, the patient's serum was used to probe a Western blot of proteins separated by 2-dimensional gel electrophoresis. The major reactive spot was a 47-kDa protein with an acidic isoelectric point (pI) (Figure 6). Stripping and reprobing of the blot with anti–pan enolase antibodies revealed a smear of reactivity at the same molecular weight as the spot widely distributed across the first dimension, corresponding to all 3 enolase proteins and their isoforms (data not shown). The retinal protein present in this spot was identified conclusively as NSE (SwissProt P09104) by matrix-assisted laser desorption/ionization mass spectrometry. Screening of the coding sequence of the patient's NSE gene (ENO2) showed only normal sequence.
We then sought to determine whether NSE reactivity is a common finding in our clinic's patients with retinal disease. The sera of 26 individuals including healthy control subjects, patients with ocular inflammatory diseases, and patients with molecularly confirmed hereditary retinal degeneration showed reactivity to proteins that had molecular weights similar to that of NSE. These sera were applied to a Western blot of purified NSE protein, and anti-NSE activity was found in one 7-year-old girl who had molecularly confirmed LCA. However, antibodies directed against NSE were not observed in any control subjects or patients with ocular inflammatory disease.
In the premolecular era, the diagnosis of LCA would not have been entertained for a child with visual acuity better than 20/200. However, as the genes responsible for LCA were discovered, a number of examples were found of children with 20/50 or better visual acuity with mutations in the same genes that cause profound visual loss in other individuals.12 In our patient, the ophthalmoscopic findings of narrowed arterioles and retinal thinning in both eyes, pink optic discs, and a nonrecordable electroretinogram suggested the diagnosis of LCA even in the face of 20/20 visual acuity in the better eye. The profound visual acuity difference was felt to be secondary to the greater involvement of the macula in the right eye coupled with some amount of superimposed amblyopia. The patient's very dark irides obscured the large relative afferent papillary defect on her initial evaluation. The perivascular hypopigmentation was somewhat suggestive of periarteriolar preservation of the retinal pigment epithelium seen in some patients with CRB1 - or RDH12- associated LCA,13,14 but careful scrutiny revealed that these perivascular changes were in fact associated with venules rather than arterioles, more consistent with a previous inflammatory insult with periphlebitis. Infrared pupillometry allowed the afferent defect to be convincingly demonstrated, and dichromatic pupillometry and Goldmann perimetry (performed a year later) revealed a degree of functional asymmetry that was more compatible with inflammatory or autoimmune disease than with primary photoreceptor degeneration. It seems likely that the electroretinographic findings in our patient were an artifact of the unsedated testing conditions. Had we performed the electroretinogram under anesthesia15 or examined the pupils with infrared videography at the time of her initial visit, the significant asymmetry of our patient's retinal disease would likely have been recognized at that time.
Plausible disease-causing mutations can be identified in approximately 65% of patients with the clinical diagnosis of LCA, and it is currently unknown what fraction of patients with LCA without molecular findings have a nongenetic inflammatory or autoimmune disease. That is, because LCA is usually inherited in an autosomal recessive fashion, most families have a single affected child and no family history of a similar disease. If our patient had had an inflammatory insult in the first few weeks of life that affected both eyes to the same degree as her right eye, she undoubtedly would have developed sensory nystagmus and exhibited clinical findings that would be very difficult to distinguish from LCA.
Once a nongenetic cause was suspected, antiretinal antibodies were sought using Western blotting and a single strong band was observed. It has been known for some time that low-titer antibodies to many antigens can develop following compromise of the blood-retina barrier from a variety of insults.5,16 Our current thinking is that an autoantibody to a single antigen is more likely to be involved in the pathogenesis of a patient's disease than antibodies directed against multiple retinal antigens. Isolated anti-NSE antibodies were not observed in any of our 32 healthy control subjects or our 80 patients with various forms of uveitis. However, we did observe an isolated anti-NSE antibody in a single patient with a molecularly confirmed photoreceptor degeneration—a 7-year-old girl with CRB1 -associated LCA (Cys195Phe and Gly750Asp). Although it is impossible to determine a pathogenic mechanism from 2 patients, it is interesting that the patient with the anti-NSE autoantibodies had much more severe disease than her older brother, who has the same CRB1 genotype. This suggests that the anti-NSE antibody may be augmenting the photoreceptor degeneration in the younger sibling.
Antibodies to another isoform of enolase, α-enolase, have also been associated with AIR in humans.17 In vivo and in vitro studies have shown that anti–α-enolase antibodies cause the death of retinal cells.18,19 Anti–α-enolase antibodies inhibit the normal function of the targeted enzyme and cause a depleted adenosine triphosphate state within the cell, an increase in intracellular calcium, increased Bax translocation in the mitochondria, induction of cytochrome- c release, and ultimately apoptosis.8 When cultured ex vivo or injected in vivo, anti–α-enolase antibodies showed the ability to penetrate the layers of the retina and induce the apoptotic death of cells in the inner nuclear layer and the ganglion cell layer.20 The latter finding is consistent with the fairly frequent observation of optic nerve involvement in AIR.21-23
Although anti–α-enolase antibodies have been described in autoimmune retinal disease, descriptions of anti-NSE autoantibodies in humans are relatively rare. The major spot bound by the patient's serum exhibited an acidic pI consistent with NSE (estimated pI 5.07) but not α-enolase (estimated pI 7.38) or β-enolase (estimated pI 7.70). The identity of the reactive spot was subsequently confirmed to be NSE by matrix-assisted laser desorption/ionization mass spectrometry as well as by the reactivity of the patient's serum with purified NSE.
Additional evidence for the pathogenicity of the anti-NSE antibodies in our patient is the recent identification of similar antibodies in a significant fraction of dogs with a canine form of AIR known as sudden acquired retinal degeneration syndrome.24,25 Braus et al26 studied the antibody profile of a cohort of dogs diagnosed with sudden acquired retinal degeneration syndrome and found that 25% had strong binding of IgG to NSE. None of the 13 control animals showed any serum reactivity to NSE. This is also consistent with a study by Maruyama et al,27 who injected anti-NSE serum into the vitreous of Lewis rats and found a lowering of the electroretinographic b-wave amplitude in the treated eyes.
In some human patients with AIR, intravenous immunoglobulin has been shown to arrest vision loss and improve their visual fields.28-30 Similarly, intravenous immunoglobulin treatment of dogs with sudden acquired retinal degeneration syndrome has also shown promising results.25 We have not used intravenous immunoglobulin to treat our patient because the visual function in her better eye is so stable.
In summary, it is important for clinicians and scientists who study and care for patients with inherited photoreceptor degenerations to consider the possibility of an autoimmune phenocopy of these diseases, especially for disorders affecting children who are too young to report sudden changes in their vision. Serologic studies of larger numbers of patients with inherited retinal disease as well as additional animal studies will be needed to further test the possibility that anti-NSE antibodies can cause or augment photoreceptor degeneration.
Correspondence: Edwin M. Stone, MD, PhD, Department of Ophthalmology and Visual Sciences, University of Iowa Carver College of Medicine, 4111 MERF, Iowa City, IA 52242 (firstname.lastname@example.org).
Submitted for Publication: September 23, 2009; final revision received April 2, 2010; accepted April 6, 2010.
Financial Disclosure: None reported.
Funding/Support: This work was supported in part by grants EY-017451 (Dr Mullins) and EY-016822 (Dr Stone) from the National Institutes of Health, the Doris Duke Charitable Foundation (Ms Ko), the Foundation Fighting Blindness (Dr Stone), and the Grousbeck Family Foundation (Dr Stone).
Additional Contributions: We thank the patients and control subjects for their participation in this research study and gratefully acknowledge the Iowa Lions Eye Bank for their support of vision research.