Fig 2: HOXA9 is downregulated in cSCC cell lines and primary tumors and acts as a direct target of miR-365. The expression levels of HOXA9 and miR-365 were detected by qPCR and western blot in a HaCaT keratinocytes, primary keratinocytes and cSCC cell lines (A431 and HSC-1) or b normal skin tissues and cSCC tumors. c IHC detection of HOXA9 on paraffin sections of cSCC tumors and normal skin specimens. Representative images with various levels of staining (Strong from normal tissues, negative or weak from tumor tissues) are shown. Scale bar: 100 µm. (200×) d Association of HOXA9 IHC-staining scores with tumor grades (I, II, and III–IV). The numbers of cases are shown below. Data are plotted as the means of 95% confidence interval ± s.d. e Schematic illustration of base pairing between miR-365 and its putative binding sequence in the 3′ UTR of HOXA9. The mutated sequences are underlined. f Wild-type (WT) or mutant reporter constructs were transfected into A431 cells with NC or miR-365 mimics. A dual luciferase assay was performed 48 h post transfection and normalized to Renilla luciferase activities. Data represent the average of three independent experiments ± s.d. g HOXA9 mRNA expression was measured in NC mimic, miR-365 mimic, antagomiR NC, or antagomiR-365 transfected A431 cells by qRT-PCR and normalized with GAPDH. Expression folds are shown with respect to NC mimic or antagomiR NC cells where normalized copy numbers were set to 1. h HOXA9 protein expression was measured in the above transfected A431 cells by western blot using GAPDH as a loading control. Each experiment was performed at least in triplicate and results are presented as mean ± s.d except in (d). One-Way ANOVA and Dunnett’s multiple comparison test were used to analyze the data. (*P < 0.05, **P < 0.01, ***P < 0.001)

Fig 3: HOXA9 represses glycolysis and promotes oxidative phosphorylation. a Metabolic phenotypes of keratinocyte and cSCC cell lines were assessed by the ratio of OCR/ECAR. b OCR and ECAR levels of the above cell lines were plotted as fold change relative to HaCaT. c, d Left: ECAR and OCR analysis of A431 cells, in which HOXA9 was depleted using two siRNAs, followed by treatment with the indicated compounds; Right: glycolytic variations (glycolysis, glycolytic capacity, and glycolytic reserve) or OXPHOS variations (basal, maximal respiration, ATP production, and spare respiratory capacity) were summarized from raw data. e, f Left: ECAR and OCR analysis of A431 cells, in which HOXA9 was overexpressed, followed by treatment with the indicated compounds; right panels: glycolytic variations (glycolysis, glycolytic capacity, and glycolytic reserve) or OXPHOS variations (basal, maximal respiration, ATP production, and spare respiratory capacity) were summarized from raw data. Each experiment was performed in triplicate and data are presented as mean ± s.d. One-Way ANOVA and Dunnett’s multiple comparison test were used to analyze the data (*P < 0.05, **P < 0.01, ***P < 0.001)

Fig 4: Genome-wide analysis of HOXA9-regulated transcriptomic changes by RNA-Seq in cSCC cells. a The expression of HOXA9 mRNA and protein was verified in A431 cells treated with siRNA targeting HOXA9 by qRT-PCR and western blot. Each experiment was performed in triplicate and data are presented as mean ± s.d. One-Way ANOVA and Dunnett’s multiple comparison test were used to analyze the data (*P < 0.05, **P < 0.01, ***P < 0.001). b Total RNAs were isolated from A431 cells treated with siNC or siHOXA9 oligos and subjected to high throughput mRNA sequencing. Differentially expressed genes between siNC-treated and siHOXA9-treated A431 cells were determined by RNA-Seq and shown by volcano plot. c Over-represented Gene Ontology categories by Gene Ontology analysis of differently-expressed genes. BP biological process, MF molecular function, CC cellular component, KEGG Kyoto Encyclopedia of Genes and Genomes. The HIF-1 signaling pathway is highlighted. d Differentially expressed genes were clustered and shown in a heat map. Color bars at the right represent gene clusters established through k-means clustering. e Validation of identified glycolytic genes in the HIF-1 pathway including HIF1A, HK2, GLUT1, PDK1, PFKL, PGK1, PFK2, PDHB, and ENO2

Fig 5: The glycolysis-inhibitory role of HOXA9 is dependent on CRIP2. a Western blot detection showed that CRIP2 knockdown led to significant upregulation of HIF-1α, HK2, GLUT1, and PDK1, whereas HOXA9 expression was not affected. b–d Depletion of CRIP2 enhanced cell proliferation, migration, and invasiveness of A431 cells by CCK-8 assay, transwell migration assay, and Matrigel invasiveness measurement. e, f ECAR and OCR analysis of A431 cells following depletion of CRIP2 by siRNA as summarized from raw data. g Western blot detection showing the variations of protein expression levels of HOXA9, CRIP2, HIF-1α, HK2, GLUT1, and PDK1 in response to CRIP2 knockdown and/or HOXA9 overexpression. h, i ECAR and OCR analysis of A431 cells depleted of CRIP2 by siRNA and/or overexpressing HOXA9 as summarized from raw data. Each experiment was performed in triplicate and data are presented as mean ± s.d. One-Way ANOVA and Dunnett’s multiple comparison test were used to analyze the data (*P < 0.05, **P < 0.01, ***P < 0.001)