Systemic Immune Activity Predicts Overall Survival in Treatment-Naïve Patients with Metastatic Pancreatic Cancer
Matthew R Farren, Thomas A Mace, Susan Geyer, Sameh Mikhail, Christina Wu, Kristen Ciombor, Sanaa Tahiri, Daniel Ahn, Anne M Noonan, Miguel Villalona-Calero, Tanios Bekaii-Saab, Gregory B Lesinski, Matthew R Farren, Thomas A Mace, Susan Geyer, Sameh Mikhail, Christina Wu, Kristen Ciombor, Sanaa Tahiri, Daniel Ahn, Anne M Noonan, Miguel Villalona-Calero, Tanios Bekaii-Saab, Gregory B Lesinski
Abstract
Purpose: Pancreatic ductal adenocarcinoma (PDAC) is an aggressive cancer with a 5-year survival rate <7% and is ultimately refractory to most treatments. To date, an assessment of immunologic factors relevant to disease has not been comprehensively performed for treatment-naïve patients. We hypothesized that systemic immunologic biomarkers could predict overall survival (OS) in treatment-naïve PDAC patients.
Experimental design: Peripheral blood was collected from 73 patients presenting with previously untreated metastatic PDAC. Extensive immunologic profiling was conducted to assess relationships between OS and the level of soluble plasma biomarkers or detailed immune cell phenotypes as measured by flow cytometry.
Results: Higher baseline levels of the immunosuppressive cytokines IL6 and IL10 were strongly associated with poorer OS (P = 0.008 and 0.026, respectively; HR = 1.16 and 1.28, respectively), whereas higher levels of the monocyte chemoattractant MCP-1 were associated with significantly longer OS (P = 0.045; HR = 0.69). Patients with a greater proportion of antigen-experienced T cells (CD45RO(+)) had longer OS (CD4 P = 0.032; CD8 P = 0.036; HR = 0.36 and 0.61, respectively). Although greater expression of the T-cell checkpoint molecule CTLA-4 on CD8(+) T cells was associated with significantly shorter OS (P = 0.020; HR = 1.53), the TIM3 molecule had a positive association with survival when expressed on CD4(+) T cells (P = 0.046; HR = 0.62).
Conclusions: These data support the hypothesis that baseline immune status predicts PDAC disease course and overall patient survival. To our knowledge, this work represents the largest cohort and most comprehensive immune profiling of treatment-naïve metastatic PDAC patients to date. Clin Cancer Res; 22(10); 2565-74. ©2015 AACR.
Trial registration: ClinicalTrials.gov NCT01280058.
Conflict of interest statement
Conflicts of Interest: GBL and TBS received independent grant support from Oncolytics Inc.
©2015 American Association for Cancer Research.
Figures
(a–b) IL-6. (a) Patients were stratified based on IL-6 levels. (b) IL-6 levels in patients stratified around the median OS. (c–d) IL-10. (c) Patients were stratified based on IL-10 levels. (d) IL-10 levels in patients stratified around the median OS. (e) Patients were stratified based on IL-6 and IL-10 levels (cutpoints as in (a) and (c)) (f) Patients were stratified based on MCP-1 levels. P-values based on dichotomized analysis.
(a–b) IL-6. (a) Patients were stratified based on IL-6 levels. (b) IL-6 levels in patients stratified around the median OS. (c–d) IL-10. (c) Patients were stratified based on IL-10 levels. (d) IL-10 levels in patients stratified around the median OS. (e) Patients were stratified based on IL-6 and IL-10 levels (cutpoints as in (a) and (c)) (f) Patients were stratified based on MCP-1 levels. P-values based on dichotomized analysis.
(a–b) IL-6. (a) Patients were stratified based on IL-6 levels. (b) IL-6 levels in patients stratified around the median OS. (c–d) IL-10. (c) Patients were stratified based on IL-10 levels. (d) IL-10 levels in patients stratified around the median OS. (e) Patients were stratified based on IL-6 and IL-10 levels (cutpoints as in (a) and (c)) (f) Patients were stratified based on MCP-1 levels. P-values based on dichotomized analysis.
(a–b) IL-6. (a) Patients were stratified based on IL-6 levels. (b) IL-6 levels in patients stratified around the median OS. (c–d) IL-10. (c) Patients were stratified based on IL-10 levels. (d) IL-10 levels in patients stratified around the median OS. (e) Patients were stratified based on IL-6 and IL-10 levels (cutpoints as in (a) and (c)) (f) Patients were stratified based on MCP-1 levels. P-values based on dichotomized analysis.
(a–b) IL-6. (a) Patients were stratified based on IL-6 levels. (b) IL-6 levels in patients stratified around the median OS. (c–d) IL-10. (c) Patients were stratified based on IL-10 levels. (d) IL-10 levels in patients stratified around the median OS. (e) Patients were stratified based on IL-6 and IL-10 levels (cutpoints as in (a) and (c)) (f) Patients were stratified based on MCP-1 levels. P-values based on dichotomized analysis.
(a–b) IL-6. (a) Patients were stratified based on IL-6 levels. (b) IL-6 levels in patients stratified around the median OS. (c–d) IL-10. (c) Patients were stratified based on IL-10 levels. (d) IL-10 levels in patients stratified around the median OS. (e) Patients were stratified based on IL-6 and IL-10 levels (cutpoints as in (a) and (c)) (f) Patients were stratified based on MCP-1 levels. P-values based on dichotomized analysis.
(a) Analysis of T cell CD45RO expression. The proportion of CD4+ (top) and CD8+ T cells (bottom) that express CD45RO was determined by gating on cells staining positive for CD4+ or CD8+ (left) and then gating on the population staining positive for CD45RO (right). (b–c) Association between CD45RO+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells (b) or CD8 T cells (c) expressing CD45RO. (d) Association between CD45RO on both CD4+ and CD8+ T cells is a predictor of survival. Patients were stratified based on the proportion of both CD4+ and CD8+ T cells that expressed CD45RO (cutpoints as used in (b) and (c)). P-values based on dichotomized analysis.
(a) Analysis of T cell CD45RO expression. The proportion of CD4+ (top) and CD8+ T cells (bottom) that express CD45RO was determined by gating on cells staining positive for CD4+ or CD8+ (left) and then gating on the population staining positive for CD45RO (right). (b–c) Association between CD45RO+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells (b) or CD8 T cells (c) expressing CD45RO. (d) Association between CD45RO on both CD4+ and CD8+ T cells is a predictor of survival. Patients were stratified based on the proportion of both CD4+ and CD8+ T cells that expressed CD45RO (cutpoints as used in (b) and (c)). P-values based on dichotomized analysis.
(a) Analysis of T cell CD45RO expression. The proportion of CD4+ (top) and CD8+ T cells (bottom) that express CD45RO was determined by gating on cells staining positive for CD4+ or CD8+ (left) and then gating on the population staining positive for CD45RO (right). (b–c) Association between CD45RO+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells (b) or CD8 T cells (c) expressing CD45RO. (d) Association between CD45RO on both CD4+ and CD8+ T cells is a predictor of survival. Patients were stratified based on the proportion of both CD4+ and CD8+ T cells that expressed CD45RO (cutpoints as used in (b) and (c)). P-values based on dichotomized analysis.
(a) Analysis of T cell CD45RO expression. The proportion of CD4+ (top) and CD8+ T cells (bottom) that express CD45RO was determined by gating on cells staining positive for CD4+ or CD8+ (left) and then gating on the population staining positive for CD45RO (right). (b–c) Association between CD45RO+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells (b) or CD8 T cells (c) expressing CD45RO. (d) Association between CD45RO on both CD4+ and CD8+ T cells is a predictor of survival. Patients were stratified based on the proportion of both CD4+ and CD8+ T cells that expressed CD45RO (cutpoints as used in (b) and (c)). P-values based on dichotomized analysis.
(a) Association between T cell expression of alternative co-stimulatory molecules, checkpoint molecules, and OS. The proportion of CD4+ (left) or CD8+ (right) T cells that express the indicated co-stimulatory or checkpoint molecule was analyzed versus OS. p-values based on continuous measure distributions (b) Analysis of T cell CTLA-4 expression. The proportion of CD4+ (top) and CD8+ (middle) T cells that express CTLA-4 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for CTLA-4, based on the fluorescence of isotype control antibody stained cells. (bottom) Typical expression of CTLA-4 on CD4+ (left) and CD8+ (right) T cells. Grey histogram – isotype control; black line – CTLA-4 staining. (c) Negative association between CD8+ CTLA-4+ T cells and OS. Patients were dichotomized based on the proportion of CD8+ T cells expressing CTLA-4. (d) Analysis of T cell TIM3 expression. The proportion of CD4+ (left) and CD8+ (right) T cells that express TIM3 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for TIM3 based on the fluorescence of isotype control antibody stained cells. (e) Positive association between TIM3 expression on CD4+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells expressing TIM3. P-values based on continuous measure distributions (a) or dichotomized analysis (c, e).
(a) Association between T cell expression of alternative co-stimulatory molecules, checkpoint molecules, and OS. The proportion of CD4+ (left) or CD8+ (right) T cells that express the indicated co-stimulatory or checkpoint molecule was analyzed versus OS. p-values based on continuous measure distributions (b) Analysis of T cell CTLA-4 expression. The proportion of CD4+ (top) and CD8+ (middle) T cells that express CTLA-4 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for CTLA-4, based on the fluorescence of isotype control antibody stained cells. (bottom) Typical expression of CTLA-4 on CD4+ (left) and CD8+ (right) T cells. Grey histogram – isotype control; black line – CTLA-4 staining. (c) Negative association between CD8+ CTLA-4+ T cells and OS. Patients were dichotomized based on the proportion of CD8+ T cells expressing CTLA-4. (d) Analysis of T cell TIM3 expression. The proportion of CD4+ (left) and CD8+ (right) T cells that express TIM3 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for TIM3 based on the fluorescence of isotype control antibody stained cells. (e) Positive association between TIM3 expression on CD4+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells expressing TIM3. P-values based on continuous measure distributions (a) or dichotomized analysis (c, e).
(a) Association between T cell expression of alternative co-stimulatory molecules, checkpoint molecules, and OS. The proportion of CD4+ (left) or CD8+ (right) T cells that express the indicated co-stimulatory or checkpoint molecule was analyzed versus OS. p-values based on continuous measure distributions (b) Analysis of T cell CTLA-4 expression. The proportion of CD4+ (top) and CD8+ (middle) T cells that express CTLA-4 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for CTLA-4, based on the fluorescence of isotype control antibody stained cells. (bottom) Typical expression of CTLA-4 on CD4+ (left) and CD8+ (right) T cells. Grey histogram – isotype control; black line – CTLA-4 staining. (c) Negative association between CD8+ CTLA-4+ T cells and OS. Patients were dichotomized based on the proportion of CD8+ T cells expressing CTLA-4. (d) Analysis of T cell TIM3 expression. The proportion of CD4+ (left) and CD8+ (right) T cells that express TIM3 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for TIM3 based on the fluorescence of isotype control antibody stained cells. (e) Positive association between TIM3 expression on CD4+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells expressing TIM3. P-values based on continuous measure distributions (a) or dichotomized analysis (c, e).
(a) Association between T cell expression of alternative co-stimulatory molecules, checkpoint molecules, and OS. The proportion of CD4+ (left) or CD8+ (right) T cells that express the indicated co-stimulatory or checkpoint molecule was analyzed versus OS. p-values based on continuous measure distributions (b) Analysis of T cell CTLA-4 expression. The proportion of CD4+ (top) and CD8+ (middle) T cells that express CTLA-4 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for CTLA-4, based on the fluorescence of isotype control antibody stained cells. (bottom) Typical expression of CTLA-4 on CD4+ (left) and CD8+ (right) T cells. Grey histogram – isotype control; black line – CTLA-4 staining. (c) Negative association between CD8+ CTLA-4+ T cells and OS. Patients were dichotomized based on the proportion of CD8+ T cells expressing CTLA-4. (d) Analysis of T cell TIM3 expression. The proportion of CD4+ (left) and CD8+ (right) T cells that express TIM3 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for TIM3 based on the fluorescence of isotype control antibody stained cells. (e) Positive association between TIM3 expression on CD4+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells expressing TIM3. P-values based on continuous measure distributions (a) or dichotomized analysis (c, e).
(a) Association between T cell expression of alternative co-stimulatory molecules, checkpoint molecules, and OS. The proportion of CD4+ (left) or CD8+ (right) T cells that express the indicated co-stimulatory or checkpoint molecule was analyzed versus OS. p-values based on continuous measure distributions (b) Analysis of T cell CTLA-4 expression. The proportion of CD4+ (top) and CD8+ (middle) T cells that express CTLA-4 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for CTLA-4, based on the fluorescence of isotype control antibody stained cells. (bottom) Typical expression of CTLA-4 on CD4+ (left) and CD8+ (right) T cells. Grey histogram – isotype control; black line – CTLA-4 staining. (c) Negative association between CD8+ CTLA-4+ T cells and OS. Patients were dichotomized based on the proportion of CD8+ T cells expressing CTLA-4. (d) Analysis of T cell TIM3 expression. The proportion of CD4+ (left) and CD8+ (right) T cells that express TIM3 was determined by gating on CD4 or CD8 expressing cells and then gating on the population staining positive for TIM3 based on the fluorescence of isotype control antibody stained cells. (e) Positive association between TIM3 expression on CD4+ T cells and OS. Patients were dichotomized based on the proportion of CD4+ T cells expressing TIM3. P-values based on continuous measure distributions (a) or dichotomized analysis (c, e).
Source: PubMed