T cell profiling reveals high CD4+CTLA-4 + T cell frequency as dominant predictor for survival after prostate GVAX/ipilimumab treatment

Abstract

Immune checkpoint blockade enhances antitumor responses, but can also lead to severe immune-related adverse events (IRAE). To avoid unnecessary exposure to these potentially hazardous agents, it is important to identify biomarkers that correlate with clinical activity and can be used to select patients that will benefit from immune checkpoint blockade. To understand the consequences of CTLA-4 blockade and identify biomarkers for clinical efficacy and/or survival, an exploratory T cell monitoring study was performed in a phase I/II dose escalation/expansion trial (n = 28) of combined Prostate GVAX/ipilimumab immunotherapy. Phenotypic T cell monitoring in peripheral blood before and after Prostate GVAX/ipilimumab treatment revealed striking differences between patients who benefited from therapy and patients that did not. Treatment-induced rises in absolute lymphocyte counts, CD4(+) T cell differentiation, and CD4(+) and CD8(+) T cell activation were all associated with clinical benefit. Moreover, significantly prolonged overall survival (OS) was observed for patients with high pre-treatment frequencies of CD4(+)CTLA-4(+), CD4(+)PD-1(+), or differentiated (i.e., non-naive) CD8(+) T cells or low pre-treatment frequencies of differentiated CD4(+) or regulatory T cells. Unsupervised clustering of these immune biomarkers revealed cancer-related expression of CTLA-4(+) in CD4(+) T cells to be a dominant predictor for survival after Prostate GVAX/ipilimumab therapy and to thus provide a putative and much-needed biomarker for patient selection prior to therapeutic CTLA4 blockade.

Fig. 1

Increased absolute lymphocyte counts (ALC) following Prostate GVAX/ipilimumab therapy are associated with prolonged survival. Absolute numbers of lymphocytes were determined before (w0), during (w4, w8, w16), and after (fu) Prostate GVAX/ipilimumab treatment. a The number of lymphocytes (×1,000) per μl blood is shown for 28 patients. b Kaplan–Meier curve for increases in ALC. The number of patients and corresponding median survival for each group are given. c Levels of circulating CD3⁺, CD4⁺, and CD8⁺ T lymphocytes and CD3⁻CD56⁺ NK cells were determined before (w0), during (w4, w8, w12, w16, w20, and w24), and after (follow-up (fu)) Prostate GVAX/ipilimumab treatment by whole-blood flow cytometric analysis. Mean percentage ± SEM of CD3⁺, CD4⁺, CD8⁺, and CD3-CD56⁺ cells is shown for 28 patients. Differences between pre- and on- or post-treatment were analyzed with the repeated measures ANOVA with a post hoc Dunnett’s multiple comparisons test. Differences were considered significant when p < 0.05, as indicated with an asterisk (*p < 0.05, **p < 0.01)

Fig. 2

Increased Treg frequencies following Prostate GVAX/ipilimumab therapy are associated with reduced survival. The number of circulating CD4⁺CD25^hiFoxP3⁺ regulatory T cells was determined before (w0), during (w4, w8, w12, w16, w20, and w24), and after (fu) Prostate GVAX/ipilimumab treatment by flow cytometric analysis on isolated PBMC. a Individual percentages (black lines) and means with range (open box plot) of circulating CD4⁺CD25^hiFoxP3⁺ cells are given for patients that experienced partial PSA response (PR; left panel), disease stabilization (SD; middle panel), or disease progression (PD; right panel) during/after Prostate GVAX/ipilimumab treatment. b Kaplan–Meier curve for increases in CD4⁺CD25^hiFoxP3⁺ Tregs. The number of patients and the corresponding median survival for each group are given. c Kaplan–Meier curve for increases in Teff/Treg ratio. Teff/Treg ratio was determined by dividing frequency of CD4⁺CD45RO⁺ T cells by the CD4⁺CD25highFoxP3⁺ Treg frequency. The number of patients and the corresponding median survival for each group are given. Differences between pre- and on- or post-treatment were analyzed with the repeated measures ANOVA with a post hoc Dunnett’s multiple comparisons test. Differences were considered significant when p < 0.05, as indicated with an asterisk (*p < 0.05, **p < 0.01)

Fig. 3

High pre-treatment frequencies of CD8⁺ Teff/Tcm/Tem, CD4⁺PD-1⁺, or CD4⁺CTLA4⁺ T cells, and low pre-treatment frequencies of CD4⁺ Teff/Tem and Tregs are predictive parameters for survival benefit after Prostate GVAX/ipilimumab therapy. Pre-treatment frequencies of circulating CD4⁺ and CD8⁺ effector/memory, CD4⁺PD-1⁺, CD4⁺CD25^hiFoxP3⁺, or CD4⁺CTLA-4⁺ T cells were determined by flow cytometry. Kaplan–Meier curves for pre-treatment frequencies of a CD8⁺ effector/memory T cells, b CD4⁺PD-1⁺, c CD4⁺CD25^hiFoxP3⁺, and d CD4⁺ effector/memory cells. The number of patients and the corresponding median survival for each group are given. e Percentage of CTLA-4-expressing cells within total and conventional CD4⁺ T cells (i.e., minus Tregs) in patients (black bars) and age- and sex-matched healthy volunteers (open bars) before treatment. f Kaplan–Meier curve for pre-treatment frequencies of CD4⁺CTLA-4⁺ conventional T cells. The number of patients and the corresponding median survival for each group are given

Fig. 4

Cancer-related expression of CTLA-4⁺ in CD4⁺ T cells is the dominant predictor for survival after Prostate GVAX/ipilimumab therapy. a Unsupervised cluster analysis of the expression of the treatment-induced and pre-treatment T cell markers. To identify clusters of correlated markers, hierarchical cluster analysis using TIGR software was performed, and average linkage analysis was done by Pearson correlation analysis. Values of the treatment-induced and pre-treatment parameters are given relative to the cut-off value (determined by Cox regression model as described in “Materials and methods”): below cut-off in green and above cut-off in red. Kaplan–Meier curve for b group 1 versus group 2 and c group 3 (with high cancer-related CD4⁺CTLA-4⁺ T cell rates: highlighted by white box in the heat plot) versus group 2 (with generally low CD4⁺CTLA-4⁺ T cell rates). The number of patients and the corresponding median survival for each group are given