Fig 2: EGFR (ERBB1) exon 20 mutation induces inverted papilloma formation. (A) The effect of a specific EGFR exon 20 mutation on sinonasal inverted papilloma formation was investigated by organotypic raft culture. (B) H&E staining of a primary tumor where iIP4 cell are derived. Black arrows indicated an inverted growth pattern observed in the raft culture.

Fig 3: ErbB ligands equally increased amiloride-insensitive ISC in both sexes. FDLE cells were cultured in SF-Med and stimulated with TGF-α, HB-EGF or NRG1 for 24 h. Data are displayed as box and whiskers with the 10–90 percentile, mean (+) and median (horizontal line). TGF-α increased Iamil-insens and decreased ΔIamil in male (a) and female (b) FDLE cells (n = 23–38; ***p < 0.001 by t-test). HB-EGF increased Iamil-insens in male cells (n = 15–23; *p < 0.05). (c) TGF-α increased Rte in male and female cells, and HB-EGF increased Rte in male cells, with the same trend observed in females (*p < 0.05). (d) NRG1 increased Iamil-insens in male and female cells, while NRG1 decreased ΔIamil in female FDLE cells (n = 25–27; *p < 0.05; ***p < 0.001). (e) Rte was not affected by NRG1 in male and female cells. (f) NRG1 increased Iamil-insens, ΔIforsk and ΔICFTR in mixed FDLE cells (n = 39–43; **p < 0.01; ***p < 0.001). (Grey boxes) male; (yellow boxes) female; (white square) mix.

Fig 4: A PI3K inhibitor and MEK inhibitor inhibited migration and invasion abilities of SIP cell, iNP cell and iNP cell carrying EGFR (ERBB1) mutation. (A) The cells were cultured in 6-well plate, starved overnight in serum- and EGF-free FYAD medium, then incubated in the presence or absence of 20 µM LY294002 or 10µM U0126 for 48 h prior to harvesting cell lysates and analyzed by western blotting. The total and the phosphorylated levels of EGFR, AKT, and MAPK were detected by the antibodies described in “Methods” (left panel), and the intensities of individual bands were indicated by bar graphs (right panels). To expose one blot to several antibodies at the same time, the membrane was cut into several pieces at desired positions prior to exposure to specific antibodies. The grouping of blots in (A) was cropped from the same samples loaded in 2 gels with 7 blots. The original blots are presented in Supplementary Fig. S3. (B) Cell migration was investigated by wound healing assay as described in “Method” (left panel), and the wound area at 24 h after 20 µM LY294002 or 10 µM U0126 treatment were indicated by bar graphs (right panels). (C) Cell invasion was investigated by using transwell invasion assay as described in “Method” (left panel), and the number of invaded cells at 24 h after 20 µM LY294002 or 10 µM U0126 treatment were indicated by bar graphs (right panels). Note that treatment with 20 µM LY294002 and 10 µM U0126 significantly reduced migration and invasion abilities of the tested cells. “a, b, c” indicate a statistically significant difference (a: p < 0.05; b: p < 0.01; c: p < 0.001), and “ns” indicates not significant.

Fig 5: Major signaling pathways downstream of ErbB receptors remain intact in EGFR KO mice. (A) Upstream regulator analysis using IPA showing major downstream signaling mediators of ErbB signaling (PI3K/AKT, ERK, p70 S6K, Jnk, RAS), and growth factor signaling via NRG1 (ligand for ErbB3) and HGF predicted to be activated in EGFR-KO mice at 5 months. IPA analysis of RNA sequencing data showing predicted activation of (B) PI3K and (C) NRG1 downstream network in EGFR KO vs WT mice at 5 months. Positive z-scores represents predicted activation of upstream regulator (absolute z-score > 2 considered as significant) based on expression profile of downstream genes. (D) Western blot and (E) densitometric analysis of total p-38, AKT, ERK, and JNK along with their phosphorylation forms in WT and EGFR KO mice at 5 months. (F) Western blot along with densitometric analysis showing protein expression of total ErbB2 and ErbB3 in WT and EGFR-KO mice at 5 months. (G) Western blot and densitometric analysis of phospho-MET and MET in EGFR-KO mice at 5 months. n = 3–5 mice/group.