Bcl-2 and β1-integrin predict survival in a tissue microarray of small cell lung cancer

Abstract

Introduction: Survival in small cell lung cancer (SCLC) is limited by the development of chemoresistance. Factors associated with chemoresistance in vitro have been difficult to validate in vivo. Both Bcl-2 and β(1)-integrin have been identified as in vitro chemoresistance factors in SCLC but their importance in patients remains uncertain. Tissue microarrays (TMAs) are useful to validate biomarkers but no large TMA exists for SCLC. We designed an SCLC TMA to study potential biomarkers of prognosis and then used it to clarify the role of both Bcl-2 and β(1)-integrin in SCLC.

Methods: A TMA was constructed consisting of 184 cases of SCLC and stained for expression of Bcl-2 and β(1)-integrin. The slides were scored and the role of the proteins in survival was determined using Cox regression analysis. A meta-analysis of the role of Bcl-2 expression in SCLC prognosis was performed based on published results.

Results: Both proteins were expressed at high levels in the SCLC cases. For Bcl-2 (n=140), the hazard ratio for death if the staining was weak in intensity was 0.55 (0.33-0.94, P=0.03) and for β(1)-integrin (n=151) was 0.60 (0.39-0.92, P=0.02). The meta-analysis showed an overall hazard ratio for low expression of Bcl-2 of 0.91(0.74-1.09).

Conclusions: Both Bcl-2 and β(1)-integrin are independent prognostic factors in SCLC in this cohort although further validation is required to confirm their importance. A TMA of SCLC cases is feasible but challenging and an important tool for biomarker validation.

Figure 1

Staining pattern of SCLC cores for β₁-integrin and Bcl-2. The TMAs were stained and scanned onto the Ariol system. Consensus scoring was performed by two blinded observers. Scores were assigned based on staining intensity. The images shown are from the Ariol system and were cropped to size using Photoshop.

Figure 2

Analysis of the role of Bcl-2 staining in survival in SCLC. (A) Overall survival analysis was performed based on the patients who had Bcl-2 staining scored on the TMA (n=140). Variables were assessed by Kaplan–Meier analysis and survival curves were plotted. The variables that were significant in the whole patient cohort remained significant in this cohort. The Kaplan–Meier plot for Bcl-2 intensity (n=140) is shown, P=0.004 by log rank test. The numbers of patients at risk in each group at each time point are shown below the plot. (B) Variables that were positive in univariable analysis were used for Cox regression analysis. A multivariable model was constructed using a forward stepwise method based on likelihood ratios. Four variables were independent predictors of outcome; stage, sex, first-line treatment and intensity of staining for Bcl-2 (P=0.03, n=117). Age at diagnosis and PS were rejected. XRT=radiotherapy; BST=best supportive care; L=limited; E=extensive; HR=hazard ratio; CI=confidence interval.

Figure 3

Analysis of the role of β₁-integrin staining in survival in SCLC. (A) Overall survival analysis was performed based on the patients who had CD29 (β₁-integrin) staining scored on the TMA (n=151). Variables were assessed by Kaplan–Meier analysis and survival curves were plotted. The variables that were significant in the whole patient cohort remained significant in this cohort. The Kaplan–Meier plot for β₁-integrin intensity (n=151) is shown, P=0.05 by log rank test. The numbers of patients at risk in each group at each time point are shown below the plot. (B) Variables that were positive in univariable analysis were used for Cox regression analysis. A multivariable model was constructed using a forward stepwise method based on likelihood ratios. Three variables were independent predictors of outcome; stage, first-line treatment and intensity of staining for β₁-integrin (P=0.02, n=121). Age at diagnosis, sex and PS were rejected. XRT=radiotherapy; BST=best supportive care; L=limited; E=extensive; HR=hazard ratio; CI=confidence interval.

Figure 4

Combining results of β₁-integrin staining and Bcl-2 staining identifies three groups with differing prognoses. Cases were grouped based on expression of β₁-integrin and Bcl-2 into cases where (1) both proteins had high expression (n=74; predicted to have poor prognosis), (2) both had low expression (n=9; predicted to have good prognosis) and (3) those with discordant expression levels (n=47; predicted to have intermediate prognosis). Survival was plotted by the Kaplan–Meier method showing significant differences between the three groups (log rank=8.98; P=0.01). The numbers of patients at risk in each group at each time point are shown below the plot.

Figure 5

Forest plot of role of Bcl-2 in SCLC survival from the literature. Previous studies were combined in a meta-analysis with the results of the current study and a forest plot generated. The summary hazard ratio (HR) suggests better survival in Bcl-2-negative cases but does not reach significance. Box size is proportional to accuracy of estimate. CI=confidence interval.