Caspase activation in tumour-infiltrating lymphocytes is associated with lymph node metastasis in oral squamous cell carcinoma

Abstract

Oral squamous cell carcinomas (OSCCs) are genetically heterogeneous and exhibit diverse stromal and immune microenvironments. Acquired resistance to standard chemo-, radio-, and targeted therapies remains a major hurdle in planning effective treatment modalities for OSCC patients. Since Caspase 8 (CASP8) is frequently mutated in OSCCs, we were interested to explore a potential interaction between tumour-infiltrating lymphocytes (TILs) and CASP8 activation using high-content image analysis of human tumour (n = 32) sections. Despite the lymphocyte-rich tumour microenvironment, we observed lower activation of CASP8 (0-10% of tumour area) and its downstream effector CASP3 (0-6%) in tumours than in normal oral epithelium. Conversely, we found apoptosis was high for all the lymphocyte subtypes examined (38-52% of lymphocytes within tumour islands). Tumours with higher Fas ligand (FasL) expression had a significantly higher proportion of cleaved CASP3/8 positive cytotoxic T cells within the tumour islands (p = 0.05), and this was associated with the presence of lymph node metastatic disease [odds ratio: 1.046, 95% confidence interval (1.002-1.091), p = 0.039]. Our finding of extensive activation of the extrinsic pathway of apoptosis in TILs, together with evidence of higher FasL in CASP8 mutated tumours, may be useful in predicting the course of disease in individual patients. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Keywords: CASPASE8; OSCC; T cell exhaustion; apoptosis; apoptotic T lymphocytes; cytotoxic T lymphocytes; immune evasion; lymph node metastasis; tumour-infiltrating lymphocytes (TILs).

Figure 1

Activation of CASP8 and CASP3 expression in a small subset of human OSCC. (A) cCASP8 or cCASP3 (green) and keratin 14 (red) expression in OSCC. (B) cCASP8 (green) expression (marked with asterisk) in keratinised, well‐differentiated areas. (C and D) Box plots showing CASP3 and CASP8 activation based on tumour differentiation status (well, moderate, and poor differentiation) and in keratinised areas. (E) Kaplan–Meier plots of DSS (log‐rank test) in patients with high or low cCASP8 or cCASP3 expression in total study cohort (n = 32) or cases with lymph node metastasis (N+, n = 18). Data shown as mean ± SD; one‐way ANOVA with Šidák's multiple comparisons test. ns, not significant. Scale bars, 100 μm; n = 32.

Figure 2

Correlation between apoptosis in tumour cells and TIL density. (A) Immunofluorescence staining for TILs (green) in OSCC sections labelled with anti‐keratin 14 (red) (representative images). ^#CD57 labelling detects NK cells and a subset of lymphocytes. Note that the same tumour sections are shown in Figure 3A. (B) Dot plots representing total lymphocyte density enumerated by high content imaging analysis (HCA) pipeline. (C) Correlation between individual lymphocyte densities in tumour (T) or stroma (S) with cCASP8 expression. (D and E) Representative images of PD‐L1 expression in OSCC and its quantification. Note patchy membranous PD‐L1 staining in tumour islands (T) and lymphocytes (zoomed image, z). (F) Correlation between tumour PD‐L1 and cCASP8 expression. Data shown as mean ± SD. *p < 0.05; one‐way ANOVA with Šidák's multiple comparisons test. ns, not significant. Scale bars, 100 μm; n = 32.

Figure 3

Apoptosis of TILs. Representative images of cCASP3 (orange) expression in lymphocytes (green) within tumour islands (T) and stroma (S) of OSCC stained with keratin 14 (red) (A) and in benign inflamed oral surface squamous epithelium control samples (IE and S regions indicated) (B) ^#CD57 labelling detects NK cells and a subset of lymphocytes. Note that the same sections are shown in Figures 2A and 3A. (C and D) Dot plots representing percentage of lymphocytes with activated CASP3 in OSCC and controls. Tumour islands (T) and stroma (S) are shown separately in C (n = 32); IE and S are shown separately in D (n = 5). (E) Positive correlations observed between cCASP3 and cCASP8 activation in Tc in tumour epithelium (correlation coefficient r = 0.53) and stroma (r = 0.56). Data shown as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001, ****p < 0.0001; one‐way ANOVA with Šidák's multiple comparisons test. ns, not significant (C and D); Spearman correlation (E). Scale bars, 100 μm.

Figure 4

FasL‐mediated apoptosis of Tc in OSCC. (A) Representative IHC images showing high FasL expression in tumour islands (T) (i), loss of FasL expression in keratinised tumour islands (marked with *) (ii), loss of diffuse expression of FasL in dysplastic surface epithelium (yellow dotted line) compared to tumour (T) (iii), and a basal/lower prickle cell pattern of FasL expression in non‐dysplastic surface epithelium (marked by askterisk) (iv). Scale bars, 200 μm (n = 32). (B) Dot plots comparing FasL expression represented as H‐score in tumour islands and stroma. (C) Box plots representing correlation between FasL expression and high or low levels (based on median values) of cCASP3 and cCASP8 in Tc. (D and E) Caspase 3/7 and Caspase 8 activity in CASP8 mutant (SJG6, SJG13, SJG18) versus CASP8 wild type lines (OK, SJG17, SJG33) upon FasL treatment. (F) TCGA data showing higher mean FASLG expression in samples harbouring CASP8 mutations. Data shown as mean ± SD (B, C, and F); *p < 0.05, **p < 0.01, and ***p < 0.001, ****p < 0.0001; Mann–Whitney test (B); multiple comparisons using ordinary one‐way ANOVA with the two‐stage linear step‐up procedure of Benjamini, Krieger, and Yekutieli (C); one‐way ANOVA with Tukey's multiple comparisons test, assays performed in triplicate and represented as mean ± SEM (D and E); unpaired t‐test (F). ns, not significant.

Figure 5

Apoptosis of Tc in OSCC patient samples. (A and B) Kaplan–Meier plots of DSS (log‐rank test), total Tc density, and cCASP3‐activated Tc in tumour epithelium and stroma (n = 32). (C) Dot plots representing percentage cCASP3+Tc in OSCC tumour islands (T) and stroma (S) (n = 32), dysplasias (IE and S)) (n = 13) and control oral mucosa (IE and S)) (n = 5). (D–F) Quantification of total Tc density and percentage of cCASP3+ Tc in tumour (T) and stroma (S) of lymph node metastatic OSCC patients. (G) Schematic describing a possible mechanism of tumour immune evasion in OSCC. Data shown as mean ± SD. *p < 0.05 and **p < 0.01, one‐sample t‐test and Wilcoxon signed‐rank test (C); one‐way ANOVA with Dunn's multiple comparisons test (D–F). ns, not significant.