#4455 DISCOVERY AND VALIDATION OF NOVEL RENAL EPITHELIAL SENESCENCE ASSOCIATED BIOMARKERS

Abstract

Abstract Background and Aims Cellular senescence is characterised by irreversible cell cycle arrest and marked changes in transcriptional and secretory activity. It can occur as a physiological component of healthy organismal development or in response to a range of cellular insults including DNA damage, oncogenic mutations and metabolic stress. Renal tubular senescence in response to ageing and injury is proposed as a driver of kidney fibrosis. Senescent cell depletion in mice improves outcomes in multiple organs including the kidney. There are currently no non-invasive biomarkers for quantifying renal senescence available. We are using a multi-omics approach and utilising human renal proximal tubular epithelial cells (hRPTECs) in culture, a murine model of renal senescence and human samples, to identify urinary biomarkers of renal tubular senescence and to determine if they can predict decline in kidney function. Method In vitro: We optimised a model of induced senescence in hRPTECs in vitro using 10Gy irradiation and treatment with MDM2-antagonist Nutlin 3A. Transcriptomic studies using bulk RNAseq were performed comparing senescent cells with proliferating controls (each group n = 5). Genes with >x2 fold-change between groups and adjusted p values <0.01 were regarded as differentially expressed. In vivo: Senescence was induced using 7Gy total body irradiation (TBI) in mice whilst the Bcl2/w/xL inhibitor ABT-263, was used to deplete senescent cells. We performed LC-MS proteomic studies on urine from mice comparing healthy controls with mice exposed to TBI with or without subsequent senolytic therapy. In human kidney biopsy samples from 55 patients with CKD, we performed immunofluorescence staining for senescence marker p21CIP1, proliferation marker Ki67 (absent in senescent cells) and tubular markers CD10 and CKPAN and quantified tubular senescence using machine learning algorithms in the software QuPath (see figure). Tubular cells were classified as senescent when p21CIP1 positive and Ki67 negative and expressed as a percentage of all tubular cells. Results Irradiated and Nutlin 3A treated cells had increased mRNA levels of CDKN1A and reduced LMNB1 and MKI67 in keeping with senescence induction. Other transcripts including CXCL8 and IL6 rose in irradiated cells compared to controls but fell in Nutlin 3A treated cells. 1272 genes where differentially expressed in irradiated cells compared to controls; 760 of these genes were differentially expressed in the same direction in Nutlin 3A treated samples. Over representation analysis highlighted pathways relating to the cell cycle, consistent with senescence induction. LC-MS studies on murine urine samples identified 15 proteins that fell in mice exposed to TBI compared to healthy controls but reverted towards baseline with senolytic treatment. By combining my datasets with publicly available data, we identified several candidate biomarkers of senescence. This includes urokinase plasminogen activator surface receptor, a protein linked with senescence and ageing as well as other novel senescence markers (not named due to pending patent applications). In the patients where senescence was quantified, the mean age at the time of biopsy was 54 years (range 19-81 years) and the mean eGFR was 56 ml/min/1.73 m2 [s.d. 34]). Renal tubular senescence correlated with age (rho = 0.64, p<0.001) and inversely with baseline eGFR (rho = -0.51, p<0.001). LC-MS and ELISA studies in matched urine samples are ongoing to determine which molecules most closely correlate with senescence histologically in the human kidney; these will be presented at the meeting. Conclusion Through a combination of transcriptomic studies and LC-MS proteomics, we have identified several candidate urinary biomarkers of senescence; ongoing studies will determine which molecules correlate with renal senescence histologically. Further studies in a cohort of >380 patients with >4 years follow-up will determine if the most promising biomarkers predict renal outcomes and will also be presented at the meeting.