br Fig CPSF facilitated stemness

Fig. 4. CPSF4 facilitated stemness maintenance in colon cancer cells.
(A) The representative images of sphere formation after CPSF4 knockdown or overexpression. (B) The expression of CSC-related markers in colon cancer USP7/USP47 inhibitor following CPSF4 knockdown or overexpression tested by Western blot. (C) The analysis of CD44+ cells by flow cytometry in colon cancer cells following CPSF4 silencing. (D) Quantitative analysis of CD44+ cells. (E) The sensitivity of colon cancer cells to oxaliplatin treatment at different concentrations detected by MTT assay when CPSF4 was silenced or overexpressed and its IC50 values were calculated accordingly. The data were shown as the mean ± s.d. (n = 3) of three independent experiments, *P < 0.05, **P < 0.01, ***P < 0.001, Student's t-test.
(−321 to −234) by PROMO (a transcriptional factor prediction web-site). NF-kB1 was one of them and has been demonstrated to be a common transcriptional factor involved in the regulation of hTERT expression [38]. We therefore examined the possible synergy of NF-kB1 in CPSF4-mediated hTERT expression regulation in colorectal cancer as proof-of-concept. The interaction between CPSF4 and NF-kB1 was confirmed by co-immunoprecipitation assay firstly (Supplementary Fig. 3A). Chromatin immunoprecipitation assay was then performed and more chromatin hTERT promoter DNA was amplified with NF-kB1 antibody than IgG itself (Supplementary Fig. 3B). Furthermore, CPSF4 knockdown decreased the binding of NF-kB1 at hTERT promoter (Supplementary Fig. 3B). Next, LOVO cells stably overexpressing CPSF4 were treated with NF-κB Activation Inhibitor (QNZ) and then were transfected with hTERT promoter-driven luciferase plasmids (−321 to +40). The result revealed that QNZ treatment inhibited the up-reg-ulation of hTERT transcription mediated by CPSF4 overexpression (Supplementary Fig. 3C). Taken together, we concluded that CPSF4 might work as a transcriptional co-activator by recruiting NF-kB1 at hTERT promoter to co-regulate hTERT transcription in colon cancer.
As noted above, the change of CPSF4 expression influenced hTERT expression levels. We further investigated CPSF4's dependence on hTERT in mediating colon cancer cell proliferation and stemness maintenance. We first constructed LOVO cells stably overexpressing CPSF4 control vector, CPSF4 overexpression plasmids, CPSF4 over-expression plasmid plus hTERT shCtrl plasmids, and CPSF4 over-expression plasmids plus hTERT shRNA plasmids. The expressing level of CPSF4 and hTERT was detected by Western blot (Fig. 5E). Next, MTT assay, colony formation assay and sphere formation analysis were re-spectively performed in these cells and the results revealed that knockdown of hTERT significantly attenuated the proliferative capacity and stemness-promoting potential caused by CPSF4 overexpression 
(Fig. 5F–H). Furthermore, we found that the expression elevations of the proliferation-related and CSC-related markers mediated by CPSF4 overexpression were also rescued by hTERT knockdown (Fig. 5I). To-gether, these results suggest that hTERT was essential for the tumor-promoting role of CPSF4 in colon cancer.
3.6. CPSF4 promoted tumorigenicity in mouse models with colon cancer xenografts by upregulating hTERT
Next, we validated the effect of CPSF4 on tumorigenesis via hTERT signaling in vivo using CRC cell-derived xenografts. The LOVO cells with different expression of CPSF4 and hTERT were subcutaneously injected into BALB/c nude mice, which were divided into 4 groups (n = 5 for each group). One month later, tumor volume and weight were respec-tively monitored and protein expressions in tumor tissues were de-termined after the mice were sacrificed. As shown in Fig. 6A–E, over-expression of CPSF4 significantly promoted tumor growth. However, inhibition of hTERT expression effectively reversed the in vivo tumor progression enhanced by CPSF4 overexpression. Moreover, both im-munohistochemical analysis and immunoblotting analysis of the tumor tissues exhibited that CPSF4 overexpression promoted the expression of Ki67, p-AKT, CD133 and MMP9 in vivo, and such promotion was si-milarly reversed by hTERT knockdown (Fig. 6F, G). All these results confirmed that CPSF4-mediated upregulation of tumor growth was at least partially realized through targeting hTERT in colon cancer.
3.7. CPSF4 positively correlates with hTERT in human colon cancer tissue samples
Finally, we evaluated the clinical significance of CPSF4 and its correlation with hTERT in a cohort of 81 patients with colorectal
Fig. 5. The transcriptional activation of hTERT by CPSF4 and its key role in CPSF4-mediated proliferation and stemness promotion in colon cancer cells.
(A) The protein levels of hTERT in colon cancer cells detected by immunoblotting following CPSF4 silencing. (B) The analysis of different hTERT promoter fragments-driven luciferase expression in colon cancer cells following CPSF4 knockdown. (C) Pulldown assay to verify the binding of CPSF4 at hTERT promoter in different colon cancer cells or in SW620 cells following CPSF4 silencing. (D) ChIP assay to verify the binding of CPSF4 at the −321 to −234 sites of hTERT promoter in different colon cancer cells or in SW620 cells upon CPSF4 knockdown. LOVO cells were respectively transected with CPSF4 control vector, CPSF4 overexpression plasmids, CPSF4 overexpression plasmids plus hTERT shCtrl plasmids, CPSF4 overexpression plasmids plus hTERT shRNA plasmids, and then (E) the expression of CPSF4 and hTERT was detected by Western blot; (F) cell viability was tested by MTT assay; (G) colony formation assay was performed and colony number was counted; (H) the representative images of tumor sphere formation; (I) the expression of PI3K/AKT signaling pathway and CSC-related markers analyzed by im-munoblotting. Data in panel B, F and G were shown as the mean ± s.d. from three separate experiments (ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, Student's t-test).