Fig 2: Increased expression of HES5 in human IMA Representative photographs of H&E and HEY1, HES1, HES5 IHC staining in human ADC and IMA. Scale bar, 50µm.Statistic analysis of low, medium, high expression ratio of IHC staining for HES5 in human ADC and IMA. Significance was calculated by ?2 test for trend.Expression of HES5, NANOG, LKB1, NKX2-1 in matching IMA tissues (1 row/patient).

Fig 3: Nanog deletion triggers gastric differentiation in KL model A, BHuman (A) and KLN murine (B) lung sections were stained with H&E or with Periodic acid–Schiff (PAS) and Alcian blue (AB). Scale bar, 50 µm.CStatistical analysis of mucinous tumor numbers (no.) in KL (n = 7) or KLN (n = 7) mice. Significance was calculated by two-tailed unpaired Student’s t-test with Welch’s correction. Results were shown as mean. **P = 0.0066.DRepresentative pictures of lung tumors at 4 weeks and 8 weeks post-Ad-Cre infection in KLN mice. Scale bar, 50 µm.EStatistical analysis of IMA ratio at 4 and 8 weeks post-Ad-Cre treatment (n = 5 per group). Results were shown as mean ± SEM.FStatistical analysis of ratio of mucinous phenotype in AAH, AD, and ADC (n = 5 per group). Results were shown as mean ± SEM. Significance was calculated by one-way ANOVA with Dunnett’s multiple comparisons test. *P = 0.0215.GqPCR analysis of stomach-restricted gene set between KL ADC and KLN mucinous tumor sections (n = 3 or 4 per group). Results were shown in a box and whisker plot with the minimum value, first quartile, median, third quartile, and maximum value. The box represented the 50% of the central data, with a line inside represented the median. The significance was calculated by two-tailed unpaired Student’s t-test with Welch’s correction. *P = 0.0260 (Hnf4a), *P = 0.0275 (Vsig1), *P = 0.0122 (Gkn3), *P = 0.0324 (Ctse), **P = 0.0084 (Onecut2).HRepresentative histological and IHC staining for HNF4a, GKN1, NKX2-1, MUC5AC, CK20 in lung tumors derived from ADC and IMA. Scale bar, 50 µm.

Fig 4: (A) Positive thyroglobulin in papillary carcinoma. (B) Positive thyroglobulin in follicular carcinoma. (C) Positive TTF1 in follicular carcinoma showing vascular invasions. (D) Positive TTF1 in papillary carcinoma.

Fig 5: Frequently concurrent loss of NANOG, LKB1, and NKX2-1 expression in human IMA Representative images of low, medium, high protein expression of NANOG in human lung adenocarcinomas. Scale bar, 150 µm (top), 50 µm (bottom).Representative H&E and IHC staining for NANOG in human ADC and IMA. Scale bar, 50 µm.Statistic analysis of NANOG expression in ADC and IMA. The tumor tissues were analyzed by IHC staining and graded as low, medium, high expression. Significance was calculated by ?2 test for trend.Western blot for NANOG and TUBULIN in 7 paired human normal lungs (N) and ADC samples (T) and human ESCs. Mr, relative molecular mass. kDa, killoDalton.Western blot for NANOG and TUBULIN in human adenocarcinoma cell lines transfected with empty vector or human NANOG overexpression vector. The arrow indicated the band of exogenous human NANOG. The asterisk indicated the band of endogenous human NANOG. Mr, relative molecular mass. kDa, killoDalton.Representative images of low, medium, high protein expression of LKB1 in human lung adenocarcinomas. Scale bar, 150 µm (top), 50 µm (bottom).Representative H&E and IHC staining for LKB1 in human ADC and IMA. Scale bar, 50 µm.Statistic analysis of LKB1 expression in ADC and IMA. The tumor tissues were analyzed by IHC staining and graded as low, medium, high expression. Significance was calculated by ?2 test for trend.The expression grade of IHC staining of NANOG, NKX2-1, LKB1 in 76 human IMA (1 sample/row). Source data are available online for this figure.