Multiplex Analysis of Peripheral Blood in Cancer Patients Predicts Humoral and Cellular Responses to Sars-Cov-2 Vaccines

Patients with cancer on active treatments have an immunosuppressed immune system both as the direct effect of their condition and indirect effects of receipt of anti-cancer therapy. Previous studies suggest that patients with cancer may have poor responses to SARS-CoV-2 vaccines and there are currently no clinically useful predictive biomarkers to determine if immunosuppressed cancer patients will respond to vaccines. This study aims to characterize which specific immune populations are associated with that poor vaccine response, so we can begin to gain a mechanistic understanding of how specific cancer therapies lead to immune dysfunction.

Patients with hematological malignancies and solid tumors were enrolled from December 2020 to January 2022 into a prospective cohort study, U.S. National Cancer Institute-funded Serological Sciences Network (SeroNet)-Coronavirus Risk Associations and Longitudinal Evaluation (CORALE). This study has collected blood from over 1,100 patients with cancer. In this cohort, we oversampled patients with altered immunity including patients with B-cell/plasma cell malignancies, stem cell transplant, and patients receiving different immunotherapies e.g. daratumumab, rituximab, and immune checkpoint inhibitors (Figueiredo et al. 2021, Figueiredo et al. 2024). In this study, we focused on a subset of patients (B-cell/plasma cell malignancies, n = 44; severe aplastic anemia, n = 3; cancer, n = 19) who had no history of SARS-CoV-2 infection and an available blood sample prior to vaccination. Immunophenotyping of patients was performed by CyTOF on PBMCs before vaccination. SARS-CoV-2 IgG antibody titers were determined with serological assays on follow-up samples (before booster). TCR repertoire analysis was performed by ImmunoSEQ Assay. The TCR breadth (number of spike-specific clones) and depth (number of cells with a specific TCR sequence), were calculated as previously described (Xu et al. 2022).

To obtain the immunological features, CyTOF data were analyzed and combined with complete blood count lab data. This yielded a feature set of 680 variables that included phenotypic frequencies, concentrations, and functional state of detected immune subsets. Humoral response scores were estimated as z-scores against time-dependent distribution of healthcare worker data (healthy controls) (Figueiredo et al. 2021). Pearson correlation and linear model regression using ElasticNet modeling against anti-Spike IgG (IgG-S) antibody titer and TCR breadth and depth were performed to determine predictive immunological features of response. For ElasticNet regression analysis, only the feature set with significant (p < 0.01) correlations were included.

Multivariate modeling showed several determinants of antibody and cellular response to vaccine. For IgG-S antibody response, we observed presence of type 2 innate lymphoid cells (ILC2), TCF1+ T-reg cells, and higher frequency of CD27- IgD+ B cells (naïve antibody-secreting B cells) as positive predictors of response. In contrast, expression of inhibitory receptor CD161 on CD127+ CD8 TCM cells (long-lived memory T cells) and CTLA4+ CD4 T-reg cells negatively correlated with antibody response. For spike-specific TCR breadth and depth, PDL1 expression on CD25+ CD4 T helper cells, expression of CD28 and CXCR3, the marker for T-cell trafficking and function, on double negative T cells, concentration of CD4 naïve T cells and frequency of CD8 naïve T cells showed a positive correlation, whereas T-bet and CTLA4 expression on CD4 T-reg cells and presence of senescent CD57+ CD127+ CD8+ memory T cells negatively predicted TCR responses. We next integrated the humoral and cellular responses and identified 5 clusters of patients with distinct patterns of response. Of note, CXCR3 and CD28 expressing double negative T cells and ILC2 cells were among the top immunological features in a cluster of patients with strong antibody and TCR responses.

We identified distinct immune features predictive of humoral and cellular response to SARS-CoV-2 vaccination. These biomarkers could guide recommendations for vaccine timing after immuno-altering cancer therapy. Moreover, SARS-CoV-2 vaccine response data give an insight into the overall immune fitness that is highly relevant for patients to go on with the subsequent immune-based therapies. We will validate these candidate biomarkers in an external cohort of patients.