#3789 SERUM AND URINARY SOLUBLE PD-1, PD-L1 AND PD-L2 AS BIOMARKERS FOR CHECKPOINT INHIBITOR RELATED ACUTE INTERSTITIAL NEPHRITIS

Abstract Background and Aims Acute interstitial nephritis (AIN) related to Immune Checkpoint Inhibitors (ICI) is a frequently described adverse effect, although the mechanisms are not fully understood. The gold standard for the diagnosis is kidney biopsy. Although it is invasive, it is necessary as AIN can be confounded with other causes of acute kidney injury (AKI) in oncologic patients. The diagnosis of AIN has therapeutical consequences such as the need to withdraw the drug if confirmed, possibly diminishing patients’ life expectancy. Few biomarkers have been proposed for non-invasive diagnosis or the follow-up of these patients, which could be useful when the biopsy is contraindicated or not available. Our aim was to study the Immune Checkpoint pathway during the event, as well as analyse potential biomarkers directly related to the pathway. Method We performed an observational study. We recruited patients with incident diagnosis of AIN related to ICI in our institution from 2018 to 2022. We obtained serum and urine at the time of diagnosis after signing informed consent. We recruited n = 18 patients with the diagnosis of AIN related to ICI. For comparison, we recruited patients with non ICI-related AIN (n = 18), patients with active ANCA-vasculitis (n = 35), patients with acute tubular necrosis (ATN) (n = 22), and controls (n = 36) (patients with diagnosis of malignancy without AKI (n = 22), and healthy controls (n = 14)). We determined soluble PD-1 (sPD-1), sPD-L1 and sPD-L2 in serum and urine using a multiplex bead-based Luminex assay. Besides, we performed PD-L1 immunohistochemistry and PD-L2 immunofluorescence staining of kidney biopsies from patients with ICI-related AIN (n = 14) and compared to patients with non ICI-related AIN (n = 10). Results Serum sPD-1 was higher in patients with AIN compared to controls (p = 0.0004) and patients with ATN (p = 0.021). There were no differences between controls and ATN (p = 0.76). Urinary sPD-1 was lower in subjects with AKI compared to controls (control vs AIN (p = 0.019), control vs active vasculitis (p<0.0001) and control vs ATN (p = 0.0304)). Patients with ATN exhibited the lowest urinary sPD-1 levels, and there were statistically significant differences when compared to AIN (p = 0.0304). Focusing on oncologic patients, urinary sPD-1 concentration <129.3 pg/ml had a 71.43% sensitivity and 94.12% specificity to differentiate ATN from AIN, with a likelihood ratio of 12.14. Serum and urinary sPD-L1 and sPD-L2 did not show statistically significant differences between AIN and ATN. In kidney biopsies, patients with ICI related AIN showed higher density of PD-L1 positive tubules than patients with non-ICI related AIN (p = 0.037). Proportion of patients having more than 2.64/mm2 PD-L2 positive tubules was higher among patients with ICI related AIN compared to non ICI-related AIN (p = 0.034). There was a positive correlation (p = 0.009, r = 0.72) between urinary sPD-1 and the density of PD-L1 positive tubules (tubules/mm2). As for PD-L2, serum PD-L2 was higher in patients with positive tubular PD-L2 expression (p = 0.009). Conclusion Serum sPD-1 increases in AIN, contrary to patients with ATN and controls. Urinary sPD-1 decreases in patients with AKI. This decrease is more intense in the case of patients with ATN, thus being useful in the differentiation between ATN and AIN, especially in patients with neoplasms. Patients with ICI-related AIN exhibit higher positivity for PD-L1 and PD-L2 staining compared to other causes of AIN. Altogether, our findings suggest a role for these soluble molecules in the ICI-related AIN pathophysiology.