A non membrane-targeted human soluble CD59 attenuates choroidal neovascularization in a model of age related macular degeneration

Abstract

Age related macular degeneration (AMD) is the most common cause of blindness amongst the elderly. Approximately 10% of AMD patients suffer from an advanced form of AMD characterized by choroidal neovascularization (CNV). Recent evidence implicates a significant role for complement in the pathogenesis of AMD. Activation of complement terminates in the incorporation of the membrane attack complex (MAC) in biological membranes and subsequent cell lysis. Elevated levels of MAC have been documented on choroidal blood vessels and retinal pigment epithelium (RPE) of AMD patients. CD59 is a naturally occurring membrane bound inhibitor of MAC formation. Previously we have shown that membrane bound human CD59 delivered to the RPE cells of mice via an adenovirus vector can protect those cells from human complement mediated lysis ex vivo. However, application of those observations to choroidal blood vessels are limited because protection from MAC- mediated lysis was restricted only to the cells originally transduced by the vector. Here we demonstrate that subretinal delivery of an adenovirus vector expressing a transgene for a soluble non-membrane binding form of human CD59 can attenuate the formation of laser-induced choroidal neovascularization and murine MAC formation in mice even when the region of vector delivery is distal to the site of laser induced CNV. Furthermore, this same recombinant transgene delivered to the intravitreal space of mice by an adeno-associated virus vector (AAV) can also attenuate laser-induced CNV. To our knowledge, this is the first demonstration of a non-membrane targeting CD59 having biological potency in any animal model of disease in vivo. We propose that the above approaches warrant further exploration as potential approaches for alleviating complement mediated damage to ocular tissues in AMD.

Figure 1. sCD59 is expressed and efficiently secreted from adenovirus in vitro.

(A) Expression cassettes for both sCD59 and GFP or without a transgene were cloned into the deleted E1 region of an E1/E3-deleted adenovirus. (B) ARPE-19 cells infected with AdCAGsCD59 express sCD59 and exhibit robust secretion into the media. CAG, chicken b-actin promoter; pA, polyA; LITR/RITR, left/right inverted terminal repeat; ¥, packaging signal; ΔE1/E3, deleted early regions 1/3; E2/E4, early regions 2/4.

Figure 2. AdCAGsCD59-conditioned media confers significant protection from human serum-mediated cell lysis.

(A) A representative FACS histogram for hepa-1c1c7 cells treated with either NHS or HI-NHS in media conditioned with either AdCAGGFP or AdCAGsCD59. (B) While no significant difference in cell lysis was observed in hepa-1c1c7 cells treated with HI-NHS in media conditioned with either AdCAGGFP or AdCAGsCD59, a significant (34.08±6.40%, **p<0.01) reduction was observed in cell lysis in hepa-1c1c7 cells treated with NHS in media conditioned with AdCAGsCD59 relative to those cells treated with NHS in media conditioned with the control virus, AdCAGGFP. HI-NHS, heat-inactivated normal human serum; NHS, normal human serum. n = 4.

Figure 3. Delivery of adenovirus to mouse subretinal space does not affect size of laser-induced CNV.

Representative micrographs of FITC-GSL I stained laser-induced CNV spots from each of AdCAGpA-injected and uninjected eyecups. No significant difference was observed in the size of spots induced by laser in AdCAGpA-injected relative to uninjected eyes (p>0.5). n = 12 spots/5 eyes (uninjected), n = 26 spots/9 eyes (AdCAGpA); ns, not significant. Scale bar = 50 µm.

Figure 4. AdCAGpA and AdCAGsCD59 coinjected with AdCAGGFP result in equivalent levels of transduction and at sites distal to the sites of laser burn.

Representative micrographs showing RPE/choroid flatmounts transduced with AdCAGGFP:AdCAGpA (1∶10) and AdCAGGFP:AdCAGsCD59 (1∶10). The region of transduction is demarcated and each of the three laser spots labeled 1–3. No significant difference in area of transduction was observed between AdCAGpA-injected and AdCAGsCD59-injected eyecups. n = 10 eyes (AdCAGsCD59), n = 8 eyes (AdCAGpA). ns, not significant. Scale bar = 200 µm.

Figure 5. AdCAGsCD59 delivered to murine RPE results in a significant reduction in laser-induced CNV.

Representative micrographs showing FITC-GSL I stained laser-induced CNV spots from eyes injected with either AdCAGpA or AdCAGsCD59. AdCAGsCD59 transduction of murine RPE results in a 61.7±19.9% reduction in size of CNV spot relative to eyes injected with AdCAGpA (**p<0.01). n = 33 spots/12 eyes (AdCAGsCD59), n = 26 spots/9 eyes (AdCAGpA). Scale bar = 50 µm.

Figure 6. AdCAGsCD59 results in a significant reduction in MAC deposition at the site of laser-induced CNV.

Representative micrographs showing laser-induced CNV spots from eyes injected with AdCAGpA and AdCAGsCD59 and co-stained with GSL I and anti-C9 antibody (MAC). For more accurate representation of MAC staining, CNV spots of equivalent size are shown for each of AdCAGpA and AdCAGsCD59. The overlay illustrates deposition of MAC extending beyond the region staining positive for GSL I, particularly in the AdCAGpA-injected eyecup. A significant (52.5±13.4%, ***p<0.001) reduction in MAC deposition was observed in regions of laser burn in AdCAGsCD59-injected eyecups relative to those injected with AdCAGpA. n = 11 spots/4eyes (AdCAGpA), n = 20 spots/7 eyes (AdCAGsCD59). Scale bar = 50 µm.

Figure 7. Intravitreal delivery of an AAV expressing sCD59 results in a significant protection against laser-induced CNV.

(A) The expression cassette containing sCD59 under the control of the chicken β-actin promoter was cloned into an AAV serotype 2 vector. A control virus expressing GFP was also generated. (B) Western blot analysis confirms secretion of sCD59 by human embryonic retinoblasts infected with AAVCAGsCD59. (C) Representative micrographs of laser-induced CNV spots observed in AAVCAGGFP-injected and AAVCAGsCD59-injected eyecups. (D) A significant 62.6±16.9% (***p<0.001) reduction in size of CNV was observed in AAVCAGsCD59-injected eyecups relative to those injected with AAVCAGGFP. n = 42 spots/11 eyes (AAVCAGGFP), n = 53 spots/14 eyes (AAVCAGsCD59). Scale bar = 50 µm.