| Summary: | To understand how immune balance is affected in immunoscnescencc and discover opportunities for correction with immunotherapy, a cross-sectional study of pro- and anti-inflammatory CD4+ T cell subsets in the peripheral blood of donors of different, ages was conducted. A whole blood assay using extra- and intracellular flow cytometry was designed to enumerate anti-inflammatory natural CD127lowCD25highFoxp3+ T regulatory cells (nTreg) and inducible IL-10+IFNy- T regulatory cells (iTreg) versus proinflammatory T helper 17 cells (Thl7) and T helper 1 cells (Thl), both IL-10+ and IL-10- . The frequency of these cell types was also linked to functional measurements of IL-17, IL-10 and IFN7 in whole blood supernatants after an overnight stimulation. We tested the hypothesis that the balance between these subsets changes with older age. This work shows for the first time that iTreg increase relative to nTreg and Thl7, but not Thl, with healthy ageing in man. We also identify that the previously reported increases in IL-10 levels relate to changes in the iTreg population. However, they did not increase relative to the Thl subset. During this project, scope for improvement in multicolour flow cytometry data analysis was identified, in order to minimise subjectivity and maximise efficiency of nTreg enumeration. The hypothesis that automating nTreg analysis with the SamSPECTRAL algorithm is superior to baseline C means clustering and traditional manual gating was investigated. SamSPECTRAL was qualitatively better than /c-means in clustering nTreg from flow cytometry data containing overlapping, non-spherical clusters with different densities, and was more objective than traditional manual gating. However, across 90 data files, an optimal solution was not always achieved, and statistical measures of cluster validity did not support the visual evidence that SamSPECTRAL better captured the natural structure of the data. A novel extension of SamSPECTRAL to include an automated elliptical gating step allowed for comparison of test and control datasets to correct nTreg frequency measurements for false positive events. As manual inspection of each solution was required, however, the ability to entirely automate nTreg analysis was prevented. We hope that this work will encourage further collaboration between the disciplines of immunology and computer science to advance the study of cancer and ageing.
|
|---|
