Identification of a novel subpopulation of Caspase-4 positive non-small cell lung Cancer patients

Abstract

Background: Therapy/prognosis of Non-Small Cell Lung Cancer (NSCLC) patients are strongly related to gene alteration/s or protein expression. However, more than 50% of NSCLC patients are negative to key drugable biomarkers.

Methods: We used human samples of NSCLC and mouse models of lung adenocarcinoma.

Results: We showed that caspase-4 was highly present in the tumor mass compared to non-cancerous human tissues. Interestingly, the orthologue murine caspase-11 promoted lung carcinogenesis in mice. Carcinogen-exposed caspase-11 knockout mice had lower tumor lesions than wild type mice, due to the relevance of caspase-11 in the structural lung cell as demonstrated by bone marrow transplantation and adoptive transfer experiments. Similarly to what observed in mice, caspase-4 was correlated to the stage of lung cancer in humans in that it induced cell proliferation in a K-Ras, c-MyC and IL-1α dependent manner. Caspase-4 positive adenocarcinoma (79.3%) and squamous carcinoma (88.2%) patients had lower median survival than patients who had lower levels of caspase-4. Moreover, PD-L1 expression and gene mutation (i.e. EGFR) were not correlated to caspase-4 expression. Instead, NSCLC patients who had K-Ras or c-MyC gene alteration were positively correlated to higher levels of caspase-4 and lower survival rate.

Conclusions: We identified a subgroup of NSCLC patients as caspase-4 positive among which double and triple positive caspase-4, K-Ras and/or c-MyC patients which prognosis was poor. Because K-Ras and c-MyC are still undrugable, the identification of caspase-4 as a novel oncoprotein could introduce novelty in the clinical yet unmet needs for NSCLC patients.

Keywords: Caspase-4; Cell proliferation; K-Ras; Lung cancer; Oncoprotein; Survival rate; Tumor progression; cMyc.

Fig. 1

Caspase-4 positive NSCLC patients show poor survival rate. Higher levels of tumor caspase-4 were detected histologically (n = 89) (a), by means of RT-PCR (n = 8) (b) and western blotting (n = 9) (c). Representative blots are shown for immunohistochemistry (a). Healthy tissues represent the non-cancerous tissues obtained after surgical resection of tumor areas. d Quantitative analysis of western blotting data. The levels of caspase-4 were related to the TNM stage (e) of NSCLC patients. f Survival rate of stage I NSCLC patients who presented > or < levels of caspase-4 according to the cut-off = 377 pg/ml; g Survival rate of stage II NSCLC patients who presented > or < levels of caspase-4 according to the cut-off = 0.377 ng/ml; h Survival rate of stage III NSCLC patients who presented > or < levels of caspase-4 according to the cut-off = 377 pg/ml. Levels of IL-1β (i) and IL-1α (j) in lung homogenates deriving by non-cancerous (healthy) and cancerous (LC) tissues. k Survival rate of NSCLC patients according to IL-1α and caspase-4 levels (n = 61). Data are expressed as median ± quartile range and represented as violin plots. One Way ANOVA followed by Bonferroni’s post-test was applied to Fig. B. Mann Whitney test was applied to Fig. D, I and J; Two-Way ANOVA was applied to Fig. E. Log-rank Mantel-Cox and Wilcoxon test were performed to statistically analyze the survival rate between the four groups

Fig. 2

Poor survival rate of adenocarcinoma and squamous lung cancer patients according to the levels of tumor-tissue caspase-4. a and b Venn’s diagrams obtained by the analysis of genomic alteration on CASP4 (mutation or amplification or higher gene copy number) correlated to other genes involved cell death, inflammation, gene expression and cell proliferation in adenocarcinoma (a) and squamous (b) lung cancer patients by using a public database (www.cbioportal.org). Common genes were related to KRAS, CMYC and CASP5. ROC analyses for caspase-4 and caspase-5 analyzed in samples obtained by adenocarcinoma (n = 66) (c) and squamous (n = 13) (d) patients. Tumor levels of caspase-4 and caspase-5 analysed by means of ELISA were correlated to the available survival rate of (e) adenocarcinoma (n = 53) and (d) squamous (n = 17) NSCLC patients. Median survival of adenocarcinoma patients with high of caspase-4 expression (e, red line) was 0.925 years vs median survival of patients with lower expression (blue line, 3.028 years). Median survival of squamous carcinoma patients with high expression of caspase-4 (f, red line) was 0.864 years vs median survival of patients with lower expression (blue line, not defined). Log-rank Mantel-Cox and Wilcoxon test were performed to statistically analyse the survival rate between the groups

Fig. 3

Triple positive Caspase-4, K-Ras and c-Myc NSCLC patients have poor survival rate. NSCLC patients were stratified according to the levels of caspase-4 related to PD-L1 expression (a), gene mutation (i.e. EGFR mutation, ALK or MET or ROS1 translocation) (b), or c-Myc overexpression (g, m) and K-Ras mutations (j, m). In addition, fraction of total was analysed (b, e, h, k, n) to better highlight the fraction of analysed patients according to the levels of tumor tissue caspase-4 based on a cut-off (377 pg/ml) obtained by ROC analysis. Furthermore, the percentage of caspase-4 positive patients were related to PD-L1 expression (c), gene mutation (i.e. EGFR mutation, ALK or MET or ROS1 translocation) (f), or c-Myc overexpression (i, o) and K-Ras mutations (l, o). Log-rank Mantel-Cox and Wilcoxon test were performed to statistically analyse the survival rate between the groups

Fig. 4

Caspase-11 is involved in tumor progression in mice. a Mouse model of NMU-induced carcinogenesis. b Wild type mice (n = 23) showed higher tumor lesions than caspase-11 ko (n = 12) mice in a time-dependent manner. c C3H mice (n = 14) did not show any difference in terms of tumor areas than wild type mice. d Lung tissues of K-Ras ^LA1 (n = 10) and K-Ras ^LA1p53^R172HB (n = 10) had similar levels of cleaved caspase-11 as wild type NMU-treated C57Bl/6 mice. e BAL levels of IL-1β in caspase-11 ko mice (n = 12) were solely reduced at longer time points compared to wild type mice (n = 23); instead, f IL-1α levels in lung homogenates were significantly reduced at all time points in caspase-11 ko mice (n = 12) compared to wild type mice (n = 23) subjected to NMU. g The neutralization of IL-1α by means of a monoclonal antibody significantly reduced tumor hyperplastic cells in the lung of NMU-treated C57BL/6 mice. Lung immune infiltration was characterized by MDSC (h), but not by Treg (i) in NMU-treated wild type but not in caspase-11 ko mice. j No differences in tumor area was observed in NMU-treated mice or in mice who were subjected to the adoptive transfer (n = 10) of CD4+ T cells obtained by C57Bl/6 or caspase-11 ko mice. k Bone marrow transplantation experiments showed that wt into ko mice had similar tumor area as caspase-11 ko mice treated with NMU (n = 12). Data are expressed as mean ± SEM or as median ± quartile range (violin plots). Two-Way ANOVA followed by Sidak’s multiple comparison test was applied to Fig. B, C, E, F, H, I. The statistical analysis performed for Fig. B and C considered the comparison among and within the groups according to two variables, tumor lesion vs time; instead the statistical analysis for Fig. E and F was performed according to the variation of the amount of cytokine released (expressed as pg/ml) vs time, as well as in Fig. H and I the variables were percentage of cells vs time. In contrast, One Way ANOVA followed by Bonferroni’s post test was applied to Fig. G, J, K, as the statistical test was related to the variation within and among the groups according the only one variable

Fig. 5

Large subunit of caspase-4 facilitates tumor cell proliferation. A549 cells were transfected with PC4–1, PC4–2, PC4–3 and PC4–4 as explained in Materials and Method section. a Cells transfected with PC4–4 (the sole large subunit of caspase-4) showed higher CFSE positive cells than the other groups (n = 15). b Co-culture of A549 cells and PBMCs (ratio 1:5) showed similar proliferation expressed as %CFSE+ cells than A549 alone treated with the recombinant large subunit of capase-4. The neutralization of EGFR by means of cetuximab (c), inhibition of HDAC by means of SAHA (d), inhibition of methyltransferase (e) and mTOR (g) did not alter cell proliferation as instead observed when the inhibitor of k-RAS (FTI-276) was added (f). Experiments were performed n = 15 in duplicate. Data are expressed as median ± quartile range and represented as violin plots. One-Way ANOVA followed by Bonferroni’s post test was applied