Phosphorylation of TRIP13 at Y56 induces radiation resistance but sensitizes head and neck cancer to cetuximab

Abstract

Radiation therapy, a mainstay of treatment for head and neck cancer, is not always curative due to the development of treatment resistance; additionally, multi-institutional trials have questioned the efficacy of concurrent radiation with cetuximab, the epidermal growth factor receptor (EGFR) inhibitor. We unraveled a mechanism for radiation resistance; that is, radiation induces EGFR, which phosphorylates TRIP13 (thyroid hormone receptor interactor 13) on tyrosine 56. Phosphorylated (phospho-)TRIP13 promotes non-homologous end joining (NHEJ) repair to induce radiation resistance. NHEJ is the main repair pathway for radiation-induced DNA damage. Tumors expressing high TRIP13 do not respond to radiation but are sensitive to cetuximab or cetuximab combined with radiation. Suppression of phosphorylation of TRIP13 at Y56 abrogates these effects. These findings show that EGFR-mediated phosphorylation of TRIP13 at Y56 is a vital mechanism of radiation resistance. Notably, TRIP13-pY56 could be used to predict the response to radiation or cetuximab and could be explored as an actionable target.

Keywords: DSB repair; EGFR; NHEJ; TRIP13; cetuximab; radiation resistance; squamous cell carcinoma.

Graphical abstract

Figure 1

TRIP13 promotes radiation resistance in vivo (A) UM-SCC-1 and UM-SCC-11A were stably transfected with FLAG-tagged TRIP13 or empty vector (pCMV6) and overexpression was verified by immunoblot analysis (left and right panels, respectively). GAPDH was used as a loading control. (B and C) Clonogenic survival assays were performed with UM-SCC-1-TRIP13 (B) or UM-SCC-11A-TRIP13 (C) and control cells were irradiated at different doses as indicated. The inhibitory radiation dose for 50% survival (ID₅₀) was calculated. Data are presented from three independent experiments with three replicates in each (p < 0.001, two-way ANOVA). (D) Schematic design for the mouse experiment. (E) UM-SCC-1-pCMV6 or UM-SCC-1-TRIP13-FLAG (1 × 10⁶) cells were injected subcutaneously, bilaterally in male athymic nude mice (n = 5 for pCMV6; n = 4 for TRIP13). At 8 days, when most tumors reached ∼150 mm³ in size, tumors on the right in each group were irradiated (2 Gy/day, 5 days). ∗∗∗p < 0.0001 for pCMV6-IR versus TRIP13-IR; ∗∗∗p < 0.0001 for pCMV6-no IR versus pCMV6-IR; ∗∗∗p < 0.0001 for pCMV6-no IR versus TRIP13-no IR; linear mixed model of tumor volume with random effect for mouse and fixed effects for time. (F) UM-SCC-1-pCMV6 or UM-SCC-1-TRIP13-FLAG (0.5 × 10⁶) cells were injected subcutaneously, bilaterally in female athymic nude mice (n = 4 for pCMV6; n = 5 for TRIP13). At 11 days, when most tumors reached ∼150 mm³, tumors on the right in each group were irradiated (2 Gy/day, 5 days). ∗∗∗p < 0.0001 for pCMV6-IR versus TRIP13-IR; ∗∗p < 0.001 for pCMV6-no IR versus pCMV6-IR; ∗∗∗p < 0.0001 for pCMV6-no IR versus TRIP13-no IR; linear mixed model. (G and H) Kaplan-Meier curve of tumor doubling time from initial day of radiation. ∗∗p ≤ 0.001, ∗∗∗p ≤ 0.0001 for pCMV6-no IR group versus TRIP13-no IR or IR and pCMV6-IR versus TRIP13-no IR or IR for male (G) and female (H) mice; log-rank test. (I) Mitoses (white ars) in representative fields of H&E-stained tissue sections. Scale bars, 300 μm (upper) and 25 μm (lower). (J) Mitoses in 10 high-power fields were averaged for each tumor from the four groups. (K) Representative tumor images.

Figure 2

Overexpression of TRIP13 promotes HNSCC cell survival and DNA repair (A) A TUNEL assay was performed on irradiated or non-irradiated UM-SCC-1-(pCMV6 and TRIP13-FLAG) after 24 h. Nicked DNA of apoptotic cells was fluorescently labeled (green) and nuclei were stained with propidium iodide (red). Representative fields are shown (scale bar, 50 μm). (B) Average apoptosis from 12 randomly selected fields. ∗∗p ≤ 0.001, ∗∗∗p ≤ 0.0001, one-way ANOVA. (C) Comet assay with irradiated (2.5 Gy) or non-irradiated UM-SCC-1-(pCMV6 and TRIP13-FLAG). Representative fields are shown. (D) Tail length, moment, and percent DNA in the tail were calculated from randomly selected cells. ∗∗∗p ≤ 0.0001, t test. (E) Irradiated and non-irradiated UM-SCC-1-(pCMV6 and TRIP13-FLAG) were lysed at indicated time points and immunoblotted with anti-TRIP13, anti-cleaved PARP1, and anti-γH2AX. Actin was used as loading control. Band intensities were quantified and normalized to corresponding actin, then to UM-SCC-1-pCMV6, and expressed as arbitrary densitometry units (DU) normalized to percent of control. Data are representative of two independent experiments. (F) Time-response curve for γH2AX intensity from two independent experiments. ∗∗p ≤ 0.001, ∗∗∗p ≤ 0.0001, t test. (G) Irradiated UM-SCC-1-pCMV6 and UM-SCC-1-TRIP13-FLAG were immunofluorescently labeled with anti-γH2AX and DAPI nuclear stain. Non-irradiated cells at 0 h served as control (scale bar, 10 μm). (H) γH2AX foci per nucleus were calculated from three replicate slides. The average foci/nucleus in four fields from each slide are shown. ∗∗∗p ≤ 0.0001, t test.

Figure 3

Phosphorylation of TRIP13 is associated with radiation resistance (A) Lysates from UM-SCC-1, UM-SCC-22B, and UM-SCC-29 were immunoblotted with anti-TRIP13 and anti-actin, as a loading control. Band intensities were quantified and normalized to corresponding actin and expressed as arbitrary DU. (B) UM-SCC-1, UM-SCC-22B, and UM-SCC-29 were irradiated with different doses (IR) as indicated and the surviving fractions and ID₅₀ were calculated (two-way ANOVA). (C) Overall survival of HNSCC patients treated with radiation with low or high TRIP13 expression (from TCGA). (D) Representative image of TRIP13-stained human HNSCC TMA (scale bar, 50 μm). (E) Recurrence-free survival in HNSCC TMA with TRIP13^positive and TRIP13^negative expression. (F) Schematic diagram showing how radiation-resistant cells were established. (G) Clonogenic assay with parent and radiation-resistant cells, UM-SCC-1, UM-SCC-1-IRR (top panel), and UM-SCC-29, UM-SCC-29-IRR (lower panel). The ID₅₀ was calculated (two-way ANOVA). (H) Lysates from UM-SCC-1 and UM-SCC-29 parent and radiation-resistant cells were immunoblotted with anti-TRIP13 and anti-GAPDH as a loading control. TRIP13 was immunoprecipitated and blotted with anti-phospho-tyrosine. 5% of eluant of immunoprecipitated TRIP13 served as a loading control. Band intensities were quantified and normalized to corresponding total TRIP13 (bottom) or GAPDH (top) and expressed as arbitrary DU. (I) Clonogenic sensitivity assay with UM-SCC1-TRIP13-FLAG and radiation-resistant UM-SCC-1-TRIP13-FLAG-IRR. ∗∗∗p ≤ 0.001, one-way ANOVA. Each color represents an independent experiment with three replicates in each experiment. (J) Lysates from UM-SCC1-(TRIP13-FLAG and TRIP13-FLAG-IRR) were immunoprecipitated and blotted with anti-phospho-tyrosine. Total TRIP13 was assessed as loading control in 5% of input. Band intensities were quantified and normalized to corresponding FLAG-tagged TRIP13 (52 kDa) and expressed as arbitrary DU.

Figure 4

EGFR phosphorylates TRIP13 (A) Mass spectrometric analysis of purified TRIP13 protein with its binding partner EGFR as a co-complex and its peptide coverage. (B and C) TRIP13 and EGFR were immunoprecipitated from UM-SCC-1 (B) and UM-SCC-11A (C) and blotted with anti-TRIP13 and anti-EGFR. Immunoglobulin G (IgG) was used as an immunoprecipitation control. (D) UM-SCC-1 and UM-SCC-11A were irradiated for indicated time periods, lysed, and immunoblotted with anti-phospho-EGFR(Y1068) and anti-EGFR. Band intensities were quantified and normalized to corresponding total EGFR and expressed as arbitrary DU. (E) TRIP13-FLAG from UM-SCC-1-TRIP13 irradiated as indicated was immunoprecipitated with anti-mouse M2 FLAG beads and blotted with anti-rabbit phospho-tyrosine and anti-TRIP13. Band intensities were quantified and normalized to corresponding total TRIP13 and expressed as arbitrary DU. (F) UM-SCC-1-TRIP13-FLAG was irradiated (2.5 Gy) in the presence of either 2 nM cetuximab or 5 μM erlotinib. After 24 h, lysates were prepared and immunoblotted with anti-pEGFR(Y1068), anti-EGFR, and anti-TRIP13. TRIP13 was immunoprecipitated with M2-FLAG beads and blotted with anti-phospho-tyrosine and anti-TRIP13. Band intensities were quantified and normalized to corresponding EGFR (top) or total TRIP13 (bottom) and expressed as arbitrary DU. (G) UM-SCC-1-TRIP13-FLAG was stimulated with 5 nM EGF for 30 min in the presence of either cetuximab or erlotinib. Lysates were immunoblotted with anti-pEGFR(Y1068), anti-EGFR, and anti-TRIP13. TRIP13 was immunoprecipitated with M2-FLAG beads and blotted with anti-phospho-tyrosine and anti-TRIP13. Band intensities were quantified and normalized to corresponding EGFR (top) or total TRIP13 (bottom) and expressed as arbitrary DU. (H) Representative image of EGFR-stained human HNSCC TMA (scale bars, 50 μm). (I) Recurrence-free survival in HNSCC-TMA with EGFR^positive and EGFR^negative expression.

Figure 5

EGFR phosphorylates TRIP13 at Y56 (A) Multiple sequence alignment from known database showing conservation of tyrosine residues across species. (B) TRIP13 was mutated at tyrosine residues and stably expressed in UM-SCC-1. Wild-type and mutant TRIP13 were immunoprecipitated from irradiated and non-irradiated cells and blotted with anti-phospho-tyrosine and anti-TRIP13. Band intensities were quantified and normalized to corresponding total TRIP13 and expressed as arbitrary DU. (C) Clonogenic sensitivity assay. pCMV6 vs TRIP13 ∗∗∗p<0.0001, pCMV6 vs Y56F p=0.9039 (NS), pCMV6 vs Y152F ∗∗p=0.0011, pCMV6 vs Y206F p=0.2174 (NS), TRIP13 vs Y56F ∗∗∗p=0.0004, TRIP13 vs Y152F p=0.6974 (NS), TRIP13 vs Y206F ∗p=0.0105, Y56F vs Y152F ∗p=0.01, Y56F vs Y206F p=0.6854 (NS), Y152F vs Y206F p=0.1705 (NS); one-way ANOVA. (D) Wild-type, mutated (Y56F), and pre-phosphorylated (pY56) TRIP13 peptides were generated across the EGFR kinase motif as indicated (top) and were used for the EGFR kinase assay. The slot blot was blotted with anti-phospho-tyrosine. Blots were also stained with Ponceau S as a loading control. Control samples were treated identically except for omission of EGFR kinase. (E) EGF-treated wild-type (EGFR-WT) and kinase dead (EGFR-KD) EGFR were stably expressed in Cho-K1 cells and validated for EGFR activation. (F) EGFR-WT and EGFR-KD proteins were immunoprecipitated with anti-EGFR-bound magnetic beads and used for a kinase assay with wild-type and scrambled TRIP13 peptides. Ponceau S confirmed equal loading of peptides. (G) UM-SCC-1-TRIP13-FLAG cells were transfected with peptides as indicated and irradiated. TRIP13 from irradiated and non-irradiated cells was immunoprecipitated and immunoblotted with anti-phospho-tyrosine as well as anti-TRIP13. Band intensities were quantified and normalized to corresponding total TRIP13 and expressed as arbitrary DU. (H) Clonogenic assay with UM-SCC-1-TRIP13-FLAG transfected with peptides as indicated and normalized to corresponding non-irradiated cells. ∗p ≤ 0.01, ∗∗∗p ≤ 0.0001, one-way ANOVA. (I) A phospho-tag gel was blotted with anti-phospho-tyrosine antibody (left) and then with anti-TRIP13 (right); arrows indicate pTRIP13 (left) and total TRIP13 (right).

Figure 6

EGFR-phosphorylated TRIP13 promotes radiation resistance via NHEJ (A) Mini-NHEJ assay: extracts from UM-SCC-1-(pCMV6, TRIP13, and TRIP13^Y56F) were used to multimerize a linearized pcDNA vector; un-ligated linearized vector (red ar) and multimerization (black ars) are shown. T4 ligase and water were positive and negative controls, respectively. (B) A plasmid recircularization assay was performed by transforming bacterial cells with re-ligated plasmid from the NHEJ assay. Colonies were counted and graphed (data from two independent experiments with three replicates in each). ∗∗∗p ≤ 0.0001, TRIP13 versus pCMV6 and Y56F for no IR; ∗∗∗p ≤ 0.0001 for TRIP13 versus pCMV6 and Y56F; ∗∗p ≤ 0.001 for pCMV6 and Y56F for IR (one-way ANOVA). (C) A mini-NHEJ assay with the indicated conditions was performed in the presence of cetuximab. T4 ligase and water were positive and negative controls, respectively. Un-ligated linearized vector (red ar) and multimerization (black ars) are shown. (D) Clonogenic survival assay with irradiated or non-irradiated UM-SCC-1-(pCMV6, TRIP13, and TRIP13-Y56F) in the presence or absence of cetuximab. Each color represents an independent experiment with three replicates in each experiment. A two-way ANOVA was performed and results are presented in Table 1. (E) Schematic diagram shows cetuximab and radiation schedule for in vivo experiments presented in (F) and (G). (F and G) UM-SCC-1-pCMV6 (F) or UM-SCC-1-TRIP13 (G) cells were injected bilaterally in each mouse and tumors were grown until ∼150 mm³. Either cetuximab or saline (five mice/group for UM-SCC-1-pCMV6; six mice/group for UM-SCC-1-TRIP13) was injected as shown in (E) and tumors were irradiated. Tumors were measured for indicated periods and tumor volume was calculated. Significance of data was determined as follows: pCMV6-Cont-no IR versus pCMV6-Cont IR (∗∗p ≤ 0.001); pCMV6-Cont-no IR versus pCMV6-Ctx-no IR (∗∗p ≤ 0.001); pCMV6-Cont-no IR versus pCMV6-Ctx IR (∗∗p ≤ 0.001); TRIP13-Cont-no IR versus TRIP13-Ctx-no IR (∗∗p ≤ 0.001); TRIP13-Cont-no IR versus TRIP13-Ctx-IR (∗∗p ≤ 0.001); TRIP13-Cont-IR versus TRIP13-Ctx-no IR (∗∗p ≤ 0.001); TRIP13-Cont-IR versus TRIP13-Ctx-IR (∗∗p ≤ 0.001).

Figure 7

Summary: radiation induces EGFR, which promotes NHEJ repair, radiation resistance, and poor outcome via phosphorylation of TRIP13 at Y56