Role of IQGAP1 in Papillomavirus-Associated Head and Neck Tumorigenesis

Abstract

Approximately 25% of head and neck squamous cell carcinomas (HNSCC) are associated with human papillomavirus (HPV) infection. In these cancers as well as in HPV-associated anogenital cancers, PI3K signaling is highly activated. We previously showed that IQ motif-containing GTPase activating protein 1 (IQGAP1), a PI3K pathway scaffolding protein, is overexpressed in and contributes to HNSCC and that blocking IQGAP1-mediated PI3K signaling reduces HPV-positive HNSCC cell survival and migration. In this study, we tested whether IQGAP1 promotes papillomavirus (PV)-associated HNSCCs. IQGAP1 was necessary for optimal PI3K signaling induced by HPV16 oncoproteins in transgenic mice and MmuPV1 infection, a mouse papillomavirus that causes HNSCC in mice. Furthermore, we found that, at 6 months post-infection, MmuPV1-infected Iqgap1^-/- mice developed significantly less severe tumor phenotypes than MmuPV1-infected Iqgap1^+/+ mice, indicating a role of IQGAP1 in MmuPV1-associated HNSCC. The tumors resulting from MmuPV1 infection showed features consistent with HPV infection and HPV-associated cancer. However, such IQGAP1-dependent effects on disease severity were not observed in an HPV16 transgenic mouse model for HNC. This may reflect that IQGAP1 plays a role in earlier stages of viral pathogenesis, or other activities of HPV16 oncogenes are more dominant in driving carcinogenesis than their influence on PI3K signaling.

Keywords: HPV; IQGAP1; MmuPV1; PI3K signaling; head and neck cancer; infection model; mouse model; papillomavirus.

Figure 1

IQGAP1 is necessary for efficient HPV oncoprotein-induced PI3K signaling. (A) Immunoblot to confirm the expression of IQGAP1, HPV16 E6, and HPV16 E7 in the constructed IQGAP1^KO NOKs-HPV cell lines. Relative band intensity (RBI) was calculated by normalizing to the band intensity of NOKs lane. (B) Immunoblot shows that stable expression of HPV16 E6 and E7 regulates PI3K signaling in NOKs cell, in part depending on IQGAP1. RBI was calculated by normalizing to NOKs lane. (C) Quantification of immunoblot in (B). AKT activation was determined by calculating the ratio of p^S473AKT over the levels of total AKT. The intensity of the immunoblots was analyzed by ImageJ, and the graph shows mean ± standard deviation of three independent experiments. Statistics: NOKs vs. NOKs IQGAP1^KO, p = 0.12; NOKs vs. NOKs E6, p = 0.001; NOKs E6 vs. NOKs IQGAP1^KO E6, p = 0.05; NOKs IQGAP1^KO vs. NOKs IQGAP1^KO E6, p = 0.005; NOKs vs. NOKs E7, p = 0.34; NOKs E7 vs. NOKs IQGAP1^KO E7, p = 0.03; NOKs IQGAP1^KO vs. NOKs IQGAP1^KO E7, p = 0.19. All statistical analysis in this figure was conducted with two-sided t-test. Asterisks represent statistical significance. Note: all uncropped western blot images for this study are summarized in Figure S8.

Figure 2

MmuPV1 infection induces PI3K signaling in keratinocytes. (A) MmuPV1 infection induces PI3K signaling in primary mouse keratinocytes. Left: immunoblotting detection of S6 activation levels in mock-infected and MmuPV1-infected primary mouse keratinocytes. Relative band intensity (RBI) was calculated by normalizing to the first mock-infected lane from the left. Right: quantification of the band intensity. PI3K signaling level was represented by the ratio of phosphorylated S6 (activated form of S6 protein) over total S6. Statistics: mock vs. MmuPV1, p = 0.069, two-sided t-test. (B) MmuPV1 infection induces PI3K signaling in NOKs via an IQGAP1-dependent mechanism. Left: immunoblot detection of AKT activation levels in mock-infected and MmuPV1-infected NOKs and NOKs IQGAP1^KO cells. RBI was calculated by normalizing to the first mock-infected NOKs lane from the left. Right: quantification of band intensity. The PI3K signaling level was represented by the ratio of phosphorylated AKT (activated form of AKT protein) over total AKT. Statistics: NOKs mock vs. NOKs MmuPV1, p = 0.053; NOKs MmuPV1 vs. NOKs IQGAP1^KO MmuPV1, p = 0.06; NOKs IQGAP1KO Mock vs. NOKs IQGAP1KO MmuPV1, p = 0.25, two-sided t-test. Note: all uncropped western blot images for this study are summarized in Figure S8.

Figure 3

IQGAP1 contributes to MmuPV1-induced head and neck tumorigenesis. (A) Overview of experimental protocol for MmuPV1 infection of tongue tissue in Iqgap1^+/+ and Iqgap1^−/− mice. (B) Tumor incidence in each experimental group at 6 months post-infection. Statistical analysis was performed with the two-sided Fisher’s exact test: MmuPV1-infected Iqgap1^+/+ vs. mock-infected Iqgap1^+/+, p = 0.14; MmuPV1-infected Iqgap1^+/+ vs. MmuPV1-infected Iqgap1^−/−, p = 0.069; mock-infected Iqgap1^+/+ vs. mock-infected Iqgap1^−/−, p = 1; mock-infected Iqgap1^−/− vs. MmuPV1-infected Iqgap1^−/−, p = 1, two-sided Fisher’s exact test. (C) Tumor multiplicity at 6 months post-MmuPV1 infection. Statistical analysis was performed with the two-sided Wilcoxon rank-sum test: MmuPV1-infected Iqgap1^+/+ vs. mock-infected Iqgap1^+/+= 2.1 vs. 0.5, p = 0.02; MmuPV1-infected Iqgap1^+/+ vs. MmuPV1-infected Iqgap1^−/−= 2.1 vs. 0.3, p = 0.006. Mock-infected Iqgap1^+/+ vs. mock-infected Iqgap1^−/−, p = 1; mock-infected Iqgap1^−/− vs. MmuPV1-infected Iqgap1^−/−, p = 0.89. (D) Disease severity at 6 months post-infection. Statistical analysis was performed with the two-sided Wilcoxon rank-sum test. MmuPV1-infected Iqgap1^+/+ vs. MmuPV1-infected Iqgap1^−/−= 1.8 vs. 0.56, p = 0.04; mock-infected Iqgap1^+/+ vs. mock-infected Iqgap1^−/−= 0.6 vs. 0.3, p = 0.82; MmuPV1-infected Iqgap1^+/+ vs. mock-infected Iqgap1^+/+=1.8 vs. 0.6, p = 0.06; mock-infected Iqgap1^−/− vs. MmuPV1-infected Iqgap1^−/− = 0.3 vs. 0.56, p = 0.84. Asterisks represent statistical significance.

Figure 4

Biomarker analysis of MmuPV1-associated cytoplastic mild dysplastic lesions arising in Iqgap1^+/+ and Iqgap1^−/− mice. The presence of MmuPV1 was detected by in situ hybridization against the MmuPV1 E6/E7 transcripts. MCM7 and BrdU were detected by immunohistochemistry. pERK and pS6 were detected by TSA-enhanced immunofluorescence (red: pERK or pS6; blue: DAPI).