Mass Cytometry for Comprehensive Profiling of Systemic Immune Composition in Head and Neck Squamous Cell Carcinoma Patients Reveals Dynamic Changes Early in Therapy

Abstract

We hypothesized that peripheral blood from patients undergoing RT would demonstrate quantifiable changes in immune cell composition early during therapy, and that some of these features would ultimately correlate with clinical outcomes. To this end, we leveraged mass cytometry (CyTOF) to comprehensively profile all major immune subpopulations in peripheral blood of head and neck cancer (HNC) patients before and up to two weeks after initiation of RT. As a pilot, we obtained peripheral blood samples from 5 patients with HNC and 5 healthy control patients. For cancer patients, samples were acquired prior to initiation of any therapy, prior to fraction 4 and prior to fraction 9 of RT. Samples were barcoded and stained using a panel of 40 metal-conjugated antibodies prior to data acquisition on Helios mass cytometer (Fluidigm). Debarcoded and normalized data were analyzed by UMAP dimensionality reduction and unsupervised segmentation using CLARA (Clustering Large Applications). Cancer patients had a median age of 60 (range 49-67), HNC squamous cell carcinoma (2 tonsil, 1 base of tongue, 1 glottic larynx, 1 unknown primary), locoregionally advanced disease, and median RT fraction size of 212 cGy. The majority of patients had p16+ HNC (3) and received concurrent platinum-based chemotherapy (4). All patients had primary and nodal disease in situ at time of radiation. Neutrophils predominated in whole blood as expected (71.9% controls, 80.6% patients, p = 0.056). When analysis was limited to non-neutrophils, 16 immune subpopulations were additionally identified via unsupervised clustering. At baseline, there were no differences in proportions of most subsets except for an increase in regulatory T cells (Tregs, p = 0.016) and decreases in naïve CD8 T cells (p = 0.032) and gdT cells (p = 0.016) in cancer patients as compared to healthy controls. However, Post-RT samples displayed dynamic and divergent changes in cell frequencies. 3 patients demonstrated up to a 5.6-fold increase in CD14+ HLA-DR-/lo monocytes (often referred to as monocytic myeloid-derived suppressor cells or M-MDSCs) typically associated with immunosuppression. Type 2 conventional dendritic cells, critical mediators of antigen cross-presentation, were increased in 3 patients, stable in 1 patient and decreased in 1 patient. Interestingly, Treg frequency was dramatically reduced at timepoint 2 before returning to near baseline levels at timepoint 3, a pattern seen in 4 patients. Major immune subpopulations can be robustly quantified in pre- and post-RT whole blood samples using CyTOF and patterns of change were evident even in a small number of patients. Dynamic changes in immune cell composition are notable even within 1-2 weeks of RT initiation. More detailed correlation with treatment regimens and therapeutic outcomes in a larger patient cohort may allow identification of prognostic immunologic biomarkers.