Fig 2: Model of the effect of RIP4 in tumor differentiation. RIP4 inhibits STAT3 activity, and maintains lung cancer cells in a well-differentiated epithelial state marked by high levels of TTF1 and no HMGA2. However, upon RIP4 knockdown, sustained signaling through STAT3 promotes the expression of ECM remodeling genes. This results in cancer cell dedifferentiation marked by expression of HMGA2 and low levels of TTF1

Fig 3: Knockdown of H19 attenuates high-mobility group AT-hook 2 (HMGA2)-mediated epithelial–mesenchymal transition (EMT). The mRNA (A) and protein levels (B) of HMGA2, vimentin, N-cadherin, and E-cadherin were confirmed by RT-qPCR and Western blot in H19-depleted CAL27 cells. GAPDH was probed as the loading control. The mRNA (C) and protein levels (D) of EMT-related transcription factors twist, zeb1, and snail1 were confirmed by RT-qPCR and Western blot. GAPDH was probed as loading control. Numbers are mean ± SD (n = 3, *p < 0.05 , **p < 0.01, and ***p < 0.001).

Fig 4: H19 knockdown decreases the expression of HMGA2 in a let-7-dependent manner. (A) The let-7a binding sites within the full-length transcripts of H19 were validated by the bioinformatics analysis Miranda. (B) let-7 sensor (psiCHECK2-let-7 4×) was transfected into CAL27 cells, together with 0, 20, 40, or 80 ng of sponge plasmid wild type (WT) H19 or mutant (Mut) H19. (C) The diagram represents the psiCHECK2-HMGA2 sensor plasmid. The let-7 binding sites of HMGA2 were inserted to the cloning site. (D) Using psiCHECK2-HMGA2, interaction of HMGA2 and miRNA let-7a was measured by luciferase assay in CAL27. (E) CAL27 cells were transfected with 48 nM control miRNA (NC), let-7 mimic (mlet-7), or let-7 inhibitor (ilet-7). RNAs were extracted 48 h later, and RT-qPCR analysis was performed. (F) let-7 sensor (psiCHECK-HMGA2) was transfected into CAL27 cells, together with 0, 20, 40, or 80 ng of sponge plasmid WT H19 or Mut H19. The expression of HMGA2 in (G) TSCC tissues and (H) cell lines was detected by RT-qPCR. GAPDH was probed as the loading control. Numbers are mean ± SD (n = 3, *p < 0.05, **p < 0.01, and ***p < 0.001).

Fig 5: Let-7 miRNAs are highly expressed during myelination and inversely correlated to Lin28B expression.(a) Forty most abundant miRNAs in the SN of wild-type mice at PN4 corrected for values in SN of PN4 mice with SC-specific deletion of Dicer (Dicer KO; see the Methods section). (b) Mean levels of abundant let-7 isoforms and the Lin28-dependent miR-98-5p at embryonic day (ED) 17.5, PN1, PN4, PN10, PN30 and PN60 in SN of wild-type mice in reads per million (r.p.m.). (c) Levels of Lin28B mRNA and of mature let-7f-5p and let-7i-5p during SN development in wild-type mice, normalized to GAPDH mRNA for Lin28B and to snoRNA-202 for let-7 miRNAs (n=3 mice per time point). (d) Cumulative distribution of differential expression of all expressed mRNAs and let-7 targets, predicted with TargetScan (TS) or miRWalk (MW), in Dicer KO versus control SN at PN1. The rightward shifts in the curves for targets indicate enrichment compared with all expressed mRNAs. (e) Hmga2 mRNA levels during wild-type mouse SN development (n=3 mice per time point). Error bars: s.e.m. (c,e). One-way ANOVA with Dunnett's multiple comparison test **P<0.01 (e).