A human retinal microvascular endothelial-pericyte co-culture model to study diabetic retinopathy in vitro

This human primary co-culture model using human retinal microvascular endothelial cells (hREC) and human retinal pericyte cells (hRP) aims to improve current understanding of the cellular changes occurring in the retinal microvasculature during diabetic retinopathy (DR). Currently, patients often present in clinic with late-stage DR, only when vision becomes impaired. Therefore, new strategies for earlier detection in clinic, combined with novel pharmaceutical and cellular interventions are essential in order to slow or halt the progression of DR from background to sight-threatening stage. This co-culture model can be used as a simple, replicable in vitro tool to discover and assess novel drug therapies and improve fundamental understanding of alterations to cell behaviour in the human retinal microvasculature during DR. hRP and hREC were cultured for up to 21 days in normoxic (20%) or hypoxic (2%) oxygen levels and physiological (5.5 mM) or very high (33 mM) glucose, to maintain a healthy, or induce a diabetic-like phenotype in vitro. Mono- or co-cultured hREC and hRP were seeded 1:1 in healthy (20% oxygen and 5.5 mM glucose) or diabetic-like (2% oxygen and 33 mM glucose) conditions, on either side of untreated polyethylene terephthalate (PET) transwell inserts, and cultured for 21 days. Mono- and co-cultures were analysed for changes in metabolic activity, angiogenic response and junctional protein expression, using immunofluorescence antibody labelling, flow cytometry and multiplex ELISA technology. hRP and hREC were successfully co-cultured, and the glucose and oxygen concentrations selected for the in vitro healthy and diabetic-like conditions were sufficient for cell viability and EC monolayer integrity, with evidence of an angiogenic response in diabetic-like conditions within the 21 day timeframe. Angiopoietin-2 (Ang-2), vascular endothelial growth factor (VEGF), and platelet-derived growth factor (PDGF) secretion were all increased, whilst hepatocyte growth factor (hHGF), tissue inhibitor for metalloproteinase-2 (TIMP-2) and interleukin-8 (IL-8) secretion were all reduced in the in vitro diabetic-like conditions. The secretion profile of co-cultures was different to mono-cultures, highlighting the importance of using co-culture models to collect data more reflective of the close relationship between hRP-hREC in vivo. Previous groups have developed useful co-culture models utilising non-human, immortalised or large vessel-sourced cells to explore changes to the vasculature during hypoxia and/or high glucose insult. In this study the use of human primary, retina-specific microvascular cells, mono- and co-cultured, collected over a longer culture period, has enabled detection of changes that may have been missed in previous models.