FAK Drives Resistance to Therapy in HPV-Negative Head and Neck Cancer in a p53-Dependent Manner

Abstract

Purpose: Radiation and platinum-based chemotherapy form the backbone of therapy in human papillomavirus (HPV)-negative head and neck squamous cell carcinoma (HNSCC). We have correlated focal adhesion kinase (FAK/PTK2) expression with radioresistance and worse outcomes in these patients. However, the importance of FAK in driving radioresistance and its effects on chemoresistance in these patients remains unclear.

Experimental design: We performed an in vivo shRNA screen using targetable libraries to identify novel therapeutic sensitizers for radiation and chemotherapy.

Results: We identified FAK as an excellent target for both radio- and chemosensitization. Because TP53 is mutated in over 80% of HPV-negative HNSCC, we hypothesized that mutant TP53 may facilitate FAK-mediated therapy resistance. FAK inhibitor increased sensitivity to radiation, increased DNA damage, and repressed homologous recombination and nonhomologous end joining repair in mutant, but not wild-type, TP53 HPV-negative HNSCC cell lines. The mutant TP53 cisplatin-resistant cell line had increased FAK phosphorylation compared with wild-type, and FAK inhibition partially reversed cisplatin resistance. To validate these findings, we utilized an HNSCC cohort to show that FAK copy number and gene expression were associated with worse disease-free survival in mutant TP53, but not wild-type TP53, HPV-negative HNSCC tumors.

Conclusions: FAK may represent a targetable therapeutic sensitizer linked to a known genomic marker of resistance.

Figure 1.

In vivo shRNA screen revealed FAK is a target for radiosensitization and chemosensitization in mutant TP53 HPV-negative HNSCC. A and B,In vivo shRNA screen following treatment with radiation (A) or carboplatin (B) in 3 mutant TP53 HPV-negative tumor models (UM-SCC-22a, HN31, and Cal27). Average RSA log P values versus reference for each screened gene shown, with values ≤ −1.3 considered significant. Orange points represent genes in the focal adhesion kinase pathway. C, Venn diagram of genes from the in vivo screen significantly reduced in tumors treated with XRT (n = 92), carboplatin (n = 92), or both (n = 72). D, Pathway analysis of genes associated with response to both radiation and carboplatin in the in vivo screen (FDR = false discovery rate). E and F, In vivo shRNA screen identified FAK (PTK2) as a radiosensitizing and antitumor target in terms of both magnitude of effects (≤ −1.3 log P; E) and as a chemosensitizing and antitumor target in terms of magnitude of effects (≤ −1.3 log P). All genes shown are statistically significant compared with FDR.

Figure 2.

FAK inhibition leads to radiosensitization in mutant, but not wild-type TP53, HPV-negative HNSCC. A, Tumor-growth curves for mutant TP53 HPV-negative HN31 xenograft model expressing either control or FAK shRNA in the absence or presence of radiation at 2 Gy for 4 days (control vs. shFAK + RT; *, P < 0.01; shFAK vs. shFAK + RT; #, P< 0.01; control + RT vs. shFAK + RT; ⁺, P < 0.01). B, Average area under the curve for HN31 shRNA in vivo experiment from A (*, P < 0.05). C, Immunoblots of the indicated proteins in mutant TP53 HPV-negative HN31 cells or wild-type TP53 HPV-negative HN30 with pretreatment of 0.5 μmol/L defactinib for 24 hours, irradiation with 2 Gy, and collected at 1-hour time point. D and E, Mutant TP53 HN31 (D) and wild-type TP53 HN30 (E) cell lines were treated with defactinib and RT at 2 Gy. Clonogenic survival was determined and normalized to the vehicle-treated cells (control vs. RT; *, P < 0.05). F, A similar clonogenic survival experiment performed in TP53-null UM-SCC-1 cells expressing empty vector, wild-type TP53, or two missense TP53 (G245C and R282W) mutations (defactinib vs. 2 Gy + defactinib; *, P < 0.05; 2 Gy vs. 2 Gy + defactinib; #, P < 0.05). Individual points within experiments were compared using ANOVA with correction for multiple comparisons. All P values are two-sided.

Figure 3.

FAK inhibition leads to chemosensitization in mutant TP53 HPV-negative HNSCC. A and B, Mutant TP53 HPV-negative HN31 cell lines were treated with carboplatin combined with either FAK shRNA (A) or defactinib (B) and carboplatin. Clonogenic survival was determined and normalized to the vehicle-treated cells. C, Western blots of the indicated proteins from mutant TP53 HN31 cells or wild-type TP53 HN30 cells pretreated with defactinib for 24 hours, followed by the addition of carboplatin for 24 hours. D, The indicated cell lines were pretreated with 0.5 μmol/L defactinib for 2 hours and then treated with 250 nmol/L cisplatin (125 nmol/L for Detroit562) for 24 hours, at which point TUNEL staining was performed. Quantification of TUNEL-positive cells per 40× field. E, Representative images of TUNEL staining in cells from D. F, MTT assay of mutant TP53 HN31 and wild-type TP53 HN30 parental and CR cell lines treated with indicated doses of cisplatin and defactinib normalized to 0 μmol/L cisplatin and 0 μmol/L defactinib. G, MTT assay of mutant TP53 HN31 and wild-type TP53 HN30 CR cell lines treated with indicated doses of cisplatin and defactinib normalized to cisplatin dose. H, Western blots of the indicated proteins in CR cells. ANOVA with correction for multiple comparisons was used for the clonogenic assay and paired Student t test was used for the TUNEL assay. *, two-sided P < 0.01.

Figure 4.

FAK phosphorylation, ROS, and DDR are differentially affected by TP53 status in HPV-negative HNSCC. A, Western blots of the indicated proteins from UM-SCC-1 cells forced to express wild-type or mutant (G245D) p53 treated with 1.0 μmol/L defactinib and ionizing radiation (2 Gy). B, ROS production in mutant TP53 HN31 and wild-type TP53 HN30 cells following treatment with defactinib and ionizing radiation (2 Gy) measured via DCF-ROS flow cytometry assay. C and D, I-Scel assay for HR (C) and NHEJ (D) in mutant TP53 HN31 and wild-type TP53 HN30 cells following treatment of defactinib. HR, homologous recombination; NHEJ, nonhomologous end joining. ANOVA with correction for multiple comparisons was used. *, two-sided P < 0.01.

Figure 5.

FAK is associated with worse DFS following radiation in mutant TP53 HPV-negative HNSCC. A and B, HPV-negative tumors treated with surgery and postoperative radiation (n = 94) for the institutional cohort were separated into either TP53 mutant (A, n = 49) or wild-type (B, n = 31). Each cohort was then further stratified by PTK2 copy number as performed previously. C, Patients were stratified by TP53 status (mutant vs. wild-type) and PTK2 mRNA expression (upper tertile vs. others) as described in the Materials and Methods. Statistical analysis via Kaplan–Meier with log-rank analysis. Two-sided P values are shown in the figure.