Article Text

Download PDFPDF

A bird in the hand . . .
  1. A D Dick
  1. Division of Ophthalmology, University of Bristol, UK; a.dick{at}

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Pathological correlate confirms clinical and experimental observations in posterior uveitis

    Continuing our understanding of the basic pathobiology of non-infectious posterior segment intraocular inflammatory conditions has relied largely on experimental models of uveitis, such as experimental autoimmune uveoretininitis (EAU). Arguably such models are not well supported by human data in that there is still no definitive evidence of a role for retinal autoantigens in posterior uveitic conditions,1,2 despite us moving on to clinical trials of tolerance induction.3 On the other hand the models are useful. The pathological changes observed appear to explain what we may see clinically—namely, vitritis, retinal vasculitis, chorioretinal leucocytic infiltrates, and optic nerve head and macula oedema.4,5 As such we are able to discern common immune mediated processes that lead to inflammation, in particular T cell and macrophage behaviour, cytokine mediation of inflammatory response, and other immune regulatory mechanisms in play in the eye.6 To this end, preclinical studies of novel immune modulatory agents (a classic example being cyclosporin7) have now been successfully translated into clinical practice.8 There are caveats in our interpretations. Although in animal models the photoreceptors are the target tissue for retinal antigen specific autoreactive T cells precipitating the non-specific inflammatory process that follows,9,10 in humans there is little evidence that photoreceptors are primarily destroyed during posterior uveitis. Definitively, therefore, we require support of human pathological studies to confirm pathological processes we observe in experimental models. To date such an aim has been restricted because specimens have frequently been of end stage disease, which does not allow us to understand the evolution of ocular inflammation. Moreover, in humans non-infectious ocular inflammation represents a spectrum of conditions, the study of which may not permit an iteric approach to investigating pathology. We thus resort to modulating experimental models by varying antigen, antigen dose, species, and strain of animals used so as to mimic the spectrum of clinical non-infectious ocular inflammatory conditions, in particular Vogt-Koyanaga-Harada disease, sympathetic ophthalmia, and birdshot retinochoroidopathy. These conditions demonstrate a significant HLA association with both MHC class II molecules in the former two11 and MHC class I molecules in birdshot retinochoroidopathy,12 13 providing some relevance to the notion of an autoimmune pathogenesis. If we were to consider animal models as a true representation of posterior uveitis in humans and not just providing excellent experimental conditions to investigate general immunobiological processes and cell behaviour in the eye then we need more pathological information on specific uveitic conditions in order to translate our experimental findings more purposefully.

    We require support of human pathological studies to confirm pathological processes we observe in experimental models

    In this issue of the BJO (1439) Gaudio and co-workers have enlightened our understanding of an archetypical uveitis, birdshot retinochoroidopathy. Their clinicopathological description of “non-end stage” disease in an HLA A29 patient is timely. Birdshot retinochoroidopathy was initially described in patients with profuse retinal vascular leakage with resultant retinal, macular, and disc oedema14,15 with a poor prognosis over many years.16 The recently described animal model of spontaneous retinopathy in HLA-A29 transgenic mice17 supports the high specificity and sensitivity of HLA-A29 positivity in the diagnosis of birdshot retinochoroidopathy,12,13 and the increased relative risk of developing the disease when this allele is present. HLA-A29 Tg mice are characterised histologically by changes that are almost identical to those described in the report by Gaudio et al, including perivasculitis, vitritis, leucocytic infiltration of the optic nerve head, and choroiditis. More immunohistochemical information of, for example, cell phenotype, microglial activation, GFAP reactivity, and cytokine profile would have assisted in comparing findings with those of experimental models.

    Why does HLA-A29 lead to increased susceptibility to develop birdshot retinochoroidopathy? Are we assisted by isolated clinicopathological reports in our understanding of the pathogenesis of uveitis? Reverting back to the animal model EAU, the most potent autoantigens that induce disease are soluble S-antigen (S-Ag) and interphotoreceptor retinoid binding protein (IRBP). Consequently, T cell responses in humans to these antigens have been studied extensively in a variety of uveitic conditions, and have been found particularly prevalent in birdshot retinochoroidopathy.18 There is now conceivably more relevance of S-Ag reactivity in uveitis. Firstly, there is biochemical evidence that two peptides from the carboxy terminus of S-Ag bind efficiently with the peptide binding motif of HLA-A29.19 Secondly, the previous clinicopathological report of birdshot chorioretinopathy exhibited strong T cell proliferative responses to S-Ag18 despite being no pathological features that truly paralleled the classic clinical features of birdshot during its evolution or paralleled the pathological observations in the HLA-A29 Tg mouse (the eye, however, was phthisical and frustratingly the patient was also HLA-A29 negative). Thirdly, although both reports show no conclusive evidence of photoreceptor involvement as seen during the development and end stage of inflammation in animal models, there was a significant granulomatous response in the outer retina in late stages of disease, which could be interpreted as targeting the photoreceptors. Clinically, end stage disease is highlighted by chorioretinal atrophy and reduced ERG responses but we still lack the definitive evidence to support the primary targeting of photoreceptors, as the clinical findings may be explained by vasculitis and ischaemia.

    Finally, if we were to conclude that clinicopathological studies in uveitis or more pertinently birdshot retinochoroidopathy are represented well by animal models how, in the case of an MHC class I restricted disease, do we generate S-Ag specific CD4+ T cell activation and retinal inflammation? Historically the evidence for molecular mimicry, where infectious agents cross react with self determinants on, for example, HLA A29 is poor, despite isolated clinical reports.20 The combination, however, of this current report and the generation of HLA-A29 Tg mice that demonstrate pathological features which strikingly resemble this present case offer an exciting future to develop our understanding of the pathogenesis of birdshot retinochoroidopathy and the role of retinal antigens, the results of which may then be translated to other uveitic conditions.

    Note in Proof

    Pathological correlate confirms clinical and experimental observations in posterior uveitis


    Linked Articles