Dexamethasone and brain metastasis-infiltrating CD8+ T cells.

Abstract

e15075 Background: Nearly all patients who develop large and/or symptomatic brain metastases receive corticosteroids such as dexamethasone (Dex) to treat or prevent symptoms like nausea, headaches, or focal neurologic deficits. Dex has immunosuppressive effects such as induction of T cell apoptosis, so it’s possible that this treatment may impair the anti-tumor immune response. To address this question, we performed a secondary analysis of our in-press data describing the phenotype of brain metastasis-infiltrating CD8+ T cells to determine whether the duration of Dex therapy prior to surgical resection correlated with phenotypes we described our cohort. Methods: Fresh brain metastases specimens were weighed, and immune cells were isolated. CD45+ and CD8+ lymphocytes were counted via flow cytometry from 35 and 34 samples, respectively. A subset of samples were analyzed by high-parameter spectral flow cytometry to gain phenotypic information. Dex treatment was later extracted from the electronic medical record, and Spearman’s correlation coefficients (CC) were calculated to evaluate the association between duration of Dex therapy and brain metastasis-infiltrating lymphocyte density or phenotype. Results: Patients were on Dex for a median of 3 days (4mg q6 hours) prior to surgery. On univariate analysis, the CC between CD45+ lymphocytes or CD8+ T cells per gram of tumor and duration of Dex treatment was -0.185 (P=0.288) and -0.122 (P=0.49), respectively. The CC between frequency of PD-1+ CD8+ T cells and duration of Dex was -0.064 (p=0.784, 21 patients). In our previous work, we used the FLOWSOM algorithm to identify distinct populations of CD8+ T cells that were enriched in brain metastases compared to blood. We described two such populations, one with a tissue residence phenotype, and the other with terminally-differentiated phenotype. The CC between frequency of these populations and duration of Dex was 0.043 (p=0.889) and 0.238 (p=0.433), respectively (13 patients). We also tested for correlation between frequency of individual PD-1+ CD8+ T cell populations and duration of Dex treatment (13-18 patients). The CC for Tim3+, CD39+, CTLA4+, CD127+, Tcf-1+, CD28+, and GZMB+ populations of PD-1+ CD8+ T cells were 0.037 (p=0.891), 0.096 (p=0.705), -0.094 (p=0.738), 0.011 (p=0.971), -0.026 (p=0.918), 0.384 (p=0.128), and 0.236 (p=0.358), respectively. Conclusions: Within this cohort of patients who were on Dex for a median of 3 days prior to surgery, duration of pre-surgery Dex did not significantly affect density of brain metastasis-infiltrating CD45+ lymphocytes or CD8+ T cells or frequency of PD-1+ CD8+ T cells.