A novel ΔNp63-dependent immune mechanism improves prognosis of HPV-related head and neck cancer

Background Deconvoluting the heterogenous prognosis of Human Papillomavirus (HPV)-related oropharyngeal squamous cell carcinoma (OSCC) is crucial for enhancing patient care, given its rapidly increasing incidence in western countries and the adverse side effects of OSCC treatments. Methods Transcriptomic data from HPV-positive OSCC samples were analyzed using unsupervised hierarchical clustering, and clinical relevance was evaluated using Kaplan-Meier analysis. HPV-positive OSCC cell line models were used in functional analyses and phenotypic assays to assess cell migration and invasion, response to cisplatin, and phagocytosis by macrophages in vitro. Results We found, by transcriptomic analysis of HPV-positive OSCC samples, a ΔNp63 dependent molecular signature that is associated with patient prognosis. ΔNp63 was found to act as a tumor suppressor in HPV-positive OSCC at multiple levels. It inhibits cell migration and invasion, and favors response to chemotherapy. RNA-Seq analysis uncovered an unexpected regulation of genes, such as DKK3, which are involved in immune response-signalling pathways. In agreement with these observations, we found that ΔNp63 expression levels correlate with an enhanced anti-tumor immune environment in OSCC, and ΔNp63 promotes cancer cell phagocytosis by macrophages through a DKK3/NF-κB-dependent pathway. Conclusion Our findings are the first comprehensive identification of molecular mechanisms involved in the heterogeneous prognosis of HPV-positive OSCC, paving the way for much-needed biomarkers and targeted treatment.


1
Supplementary Tables Supplementary Table 1: Patients' demographics.Two independent cohorts of patients with HPVpositive OSCC from the tumor banks of Strasbourg, France (N=34) and Liège, Belgium (N=43) were used for the validation of the transcriptomic data.77 tumors specimens were used for gene expression assays by RT-qPCR.71 FFPE samples were available for immunohistochemistry analyses.Patients' gender and age, history of tobacco smoking, pathological tumor size staging (pT), pathological lymph node invasion staging (pN), tumor stage, treatment, occurrence of metastasis within three years after treatment and 5-year overall survival are shown.Number and percentage (in brackets) for each cohort and combined cohorts are shown.NA: non available.Supplementary Table 2: List of oligonucleotide primer pairs.Gene of interest expression analysis was carried out by using a RT-qPCR approach.The names of analyzed gene are shown, as well as the 5'-to-3' sequence of forward and reverse oligonucleotide primers used in this study.

Gene name
Forward primer (5'-3') Reverse primer (5'-  S4).The median expression (Med) in metastatic (Meta) and non-metastatic (Non-meta) lesions is shown.Median expressions were compared using a two-sample Wilcoxon rank-sum test, and differences were considered statistically significant when p<0.05 (shown in bold).A ROC-curve analysis of the relationship between expression and the occurrence of metastatic relapse within 3 years was carried out, and optimal cut-off values were determined.The predictive power of these values was assessed by determining their sensibility, specificity and area under the curve (AUC).Immunocytofluorescence staining of p65 expression in THP-1 macrophages transfected with either scrambled (siCtrl) or anti-CKAP4 (siCKAP4) siRNAs, and incubated with DMEM (negative control) or 0.5 µg of hrDKK4 for 6 h prior to staining.DAPI, p65 staining and merge are shown.Magnification: X200.A magnification (right panels) of the inset in the merge is shown.White and yellow arrowheads highlight p65 staining in the cytoplasm and the nuclei, respectively.Supplemental Video S1.In vitro time-lapse analysis of the phagocytosis of SCC90 cells by THP-1 macrophages Green-labelled SCC90 cells were co-cultured with red-labelled THP-1 macrophages, cultures were analyzed using a time-lapse video-microscopy approach.Images were acquired every 10min for 22h and 20min using the 20X objective of a IncuCyte® S3 Live-Cell Analysis Instrument.

Table 3 :
List of antibodies and experimental conditions.Protein of interest expression analysis was carried out by using Western Blot, immunohistochemistry or immunocytofluorescence approaches.Protein names are shown, as well as the references of used antibodies (including clone reference and provider) and the dilution at which they were used is shown.

Table 4 :
(1)t of 148 genes found to be differentially expressed in Cluster 1 and Cluster 2. Gene symbols, Gene names, gene id, log2 fold-change (logFC), average expression (AveExpr) and adjusted p-value (Adj.p.Val) are shown.Please note that only genes found to be deregulated with a logFC>1 and an Adj.p.Val<0.05 in our unsupervised hierarchical clustering analysis (Fig.1A) are shown.Gene names highlighted in bold were previously shown to be transcriptionally regulated by Np63 in Barbieri et al.(1).

Table 5 :
Meta-analysis and two-two comparison of our tumor transcriptomic data with independent publicly available data sets.Our transcriptomic data set was compared to data sets from Slebos et al. (2), Mirghani et al. (3) and Pyeon et al. (4), which allowed the recovery of gene modules named with colors.The Z-score, correlation coefficient (Corr.)and p-value of the two-two comparisons are shown.The modules showing the most genes in common with Cluster 2 are shown in italic.

Table 6 :
Analysis of the correlation of gene expression with metastasis occurrence.