Executing precision longitudinal flow to differentiate natural variation of immune populations from treatment response

Abstract

Abstract One challenge in assessing modulation of the immune response following immune therapy is understanding natural variation in the same immune populations of specific patient cohorts. Confident interpretation of data requires standardization of procedures and proper use of appropriate controls. Here, we describe the steps taken to address potential sources of technical variability in a 31-parameter flow cytometry panel built for a Type 1 Diabetes study with over 450 samples. This includes titering and bridging antibodies, selection and integration of multiple compensation bead control options, optimizing machine performance through voltage scans and 8-peak bead matching, and post-acquisition FlowJo gating for data cleanup. When applying this process to our large study, we were able to utilize the coefficient of variation (CV) in combination with our internal control sample to show low overall technical variation. Of the 64 frequency and intensity outputs we monitored, 75% had CVs less than 25%. Higher CVs were seen exclusively in low abundance or intensity populations more subject to measurement error. When looking at the natural variation of the populations we monitored, none of the technical variation we noted exceeded the biological variation present in the trial samples. Furthermore, we calculated CVs for each population over time in each individual and averaged those CVs within their cohorts to quickly identify populations of interest with treatment—in this case, modulation of follicular helper cells. Employing this streamlined process produces clean data that can be analyzed confidently and allows for quick discernment of biological changes in subjects with or without immune perturbation across time. ITN (NIAID) and TrialNet (NIDDK) funding along with BRI internal HIP Core funds