Histone methyltransferase SETDB1 regulates liver cancer cell growth through methylation of p53

Abstract

SETDB1 is a histone H3K9 methyltransferase that has a critical role in early development. It is located within a melanoma susceptibility locus and facilitates melanoma formation. However, the mechanism by which SETDB1 regulates tumorigenesis remains unknown. Here we report the molecular interplay between SETDB1 and the well-known hotspot gain-of-function (GOF) TP53 R249S mutation. We show that in hepatocellular carcinoma (HCC) SETDB1 is overexpressed with moderate copy number gain, and GOF TP53 mutations including R249S associate with this overexpression. Inactivation of SETDB1 in HCC cell lines bearing the R249S mutation suppresses cell growth. The TP53 mutation status renders cancer cells dependent on SETDB1. Moreover, SETDB1 forms a complex with p53 and catalyses p53K370 di-methylation. SETDB1 attenuation reduces the p53K370me2 level, which subsequently leads to increased recognition and degradation of p53 by MDM2. Together, we provide both genetic and biochemical evidence for a mechanism by which SETDB1 regulates cancer cell growth via methylation of p53.

Figure 1. SETDB1 is overexpressed in liver cancers.

(a) Publically available clinical cancer gene expression data (GSE6764) were downloaded from GEO to assess SETDB1 expression in HCC tumour samples. The expression of SETDB1 was extracted and box-plotted using Prism. (b) Total RNA was isolated from six pairs of human primary HCC tissues. The expression of SETDB1 in these samples was analysed using RT–qPCR. Data are presented as mean±s.d. (c) Representative immunohistochemical staining of the SETDB1 protein using HCC TMA. The sections were counterstained with haematoxylin and eosin (H&E) staining. Scale bar, 100 μM. (d) Percentages of strong positive, positive, weak positive and negative samples for SETDB1 in tumours (n=59) and normal adjacent controls (n=35).

Figure 2. Association of TP53 mutations with SETDB1 copy number gain or overexpression in HCC tumours.

Waterfall plots were generated to illustrate the association of TP53 mutations with SETDB1 copy number gain (a) or overexpression (b). Eighty-four HCC tumours were either ranked by SETDB1 copy number or expression level. Each bar represents one tumour sample, which is coloured by the TP53 mutation status. Tumours with TP53 mutation are marked as red.

Figure 3. Inhibition of SETDB1 suppresses HCCLM3 cell growth in vitro.

(a) HCCLM3 cells were transfected with two independent siRNAs against SETDB1. Cells were grown for 5 days and cell proliferation was measured with CellTiter-Glo (n=6). The knockdown efficiency of these two siRNAs was confirmed by using RT–qPCR (n=3). Data are presented as mean±s.d. *P<0.001, t-test. (b) Cells were infected with inducible lentiviral-based shRNA against SETDB1. Cells were treated with 1 μg ml⁻¹ doxycycline to induce SETDB1 knockdown and cell growth was measured on days 3, 5 and 7, respectively. Data are presented as mean±s.d. *P<0.001, t-test, n=6. Target knockdown was confirmed using RT–qPCR. (c) PLC cells were transfected with FLAG-tagged SETDB1 overexpression construct or pcDNA3 vector control or a set domain-deleted SETDB1 mutant. Cell growth was measured 5 days after transfection. Data are presented as mean±s.d. *P<0.05, t-test, n=6. Expression level of the wild-type or mutant SETDB1 was confirmed with western blot analysis using the antibody against FLAG. FL, full length; Mu, mutated SETDB1. (d) HCCLM cells were treated with 1 μg ml⁻¹ doxycycline for 3 days, and 10 μM EdU was added to the culture for the last 6 h of culture. EdU incorporation was assessed by immunofluorescence using the antibody against EdU. Two representative images of the control and SETDB1 knockdown cells are shown here. Inhibition of SETDB1 significantly reduced the number of EdU-positive cells. The relative ratios of the EdU-positive cells are also plotted. SETDB1 knockdown significantly reduced the number of EdU-positive cells. (e) HCCLM3 cells were grown in three-dimension in soft agar at the density of 32 or 64K cells ml⁻¹. Cells were treated with doxycycline for 3 or 4 days, and the cultures were stained for p-iodonitrotetrazolium violet overnight at 37 °C.

Figure 4. R249S mutation renders cancer cell sensitive to SETDB1 knockdown.

(a) Parental or wild-type p53 or p53R249S-expressing Hep3B cells were infected with an inducible lentiviral-based shRNA against SETDB1. Cells were treated with 1 μg ml⁻¹ doxycycline to induce SETDB1 knockdown, and cell growth was measured on days 2, 5 and 7. Data are presented as mean±s.d. n=6. (b) Assessing growth phenotypes on SETDB1 knockdown in isogenic pair of somatic p53 knockout HCT116 cells, or R249S restored HCT116 cells. Cells were infected with SETDB1 shRNA, and growth was measured on days 3 and 5, the same as above. Data are presented as mean±s.d. n=6. (c) Growth inhibition on SETDB1 knockdown was measured in HCCLM3 with stably infected SETDB1-inducible shRNA. The cells were treated with a control scramble shRNA or shRNA against p53. Cell growth assessment was performed the same as above. Data are presented as mean±s.d; *P<0.05, t-test, n=6. (d) Hep3B cells with stably infected SETDB1 shRNA were transfected with vector control, wild-type p53 or p53R249S mutants. Growth inhibition on SETDB1 knockdown was measured 2 days after transfection. Data are presented as mean±s.d.; *P<0.05, t-test, n=6. (e) Cell growth analysis was performed the same as above, except that the cells were treated with 0.05 μg ml⁻¹ Doxorubicin for 2 or 5 days. Data are presented as mean±s.d; *P<0.05, t-test, n=6.

Figure 5. LC-MS/MS analysis of the p53 methylation by SETDB1.

Unmethylated or chemically methylated p53 peptides were incubated with the endogenous SETDB1 complex. The end products were analysed using LC-MS/MS to confirm the p53 methylation. Corresponding peptide sequences and methylation status were marked by the identified peaks. (a) p53K370me0 peptide was used as the substrate. (b) K370me1 peptide was used as the substrate. (c) SMYD2 was used to induce K370me1. SETDB1 mainly catalyses p53K370me1 to K370me2. (d) SMYD2 alone with the p53K370me0 peptide was used as a control.

Figure 6. Regulation of p53K370 methylation by SETDB1.

(a) To confirm that endogenous SETDB1 methylates endogenous p53, HCCLM3 cells stably infected with inducible SETDB1 shRNA were treated with doxycycline for 3 days, and then exposed to Doxorubicin for 24 h. Methylation of p53 was assessed using the antibody against p53K370me2. (b) SNU182 cells stably infected with inducible SETDB1 shRNA were treated with or without 1 μg ml⁻¹ doxycycline for 3 days after transfection and then exposed to 0.05 μg ml⁻¹ Doxorubicin for 24 h. The cells were harvested for western blot analysis using p53K370me2 antibody (1.2 μg ml⁻¹) alone or competed with 5 μM p53K370 unmethylated or monomethylated peptides or di-methylated peptide at the dose of 1, 5 and 10 μM. All competing peptides were of 31-mer long in length. (c) p53 null HCT116 cells were transfected with wild-type p53 or a mutant p53 with K370 replaced with A (p53K370A). Cells were also treated with 0.05 μg ml⁻¹ Doxorubicin for 24 h and harvested for western blot analysis using p53K370me2 antibody. GAPDH was analysed as the control. (d) p53 null HCT116 cells were transfected with wild type or p53R249S mutant together with SETDB1 or the set domain-deleted SETDB1 control. The methylation of p53 at K370 was measured by western blot analysis. (e) In vitro p53 methylation assays were performed using the endogenous SETDB1 complex pulled down from HCCLM3 cells that were treated with 0.5 μg ml⁻¹ of Doxorubicin for 6 h. The immunoprecipitated complex was used as the enzyme in the assay with full-length p53 protein as the substrate and S-adenosyl methionine (SAM) as the methyl donor. Methylation was assayed using western blot analysis. The reaction was carried out at room temperature or at 37 °C for the duration as indicated.

Figure 7. Regulation of p53 stability by p53K370 di-methylation.

Wild-type p53 (a) or p53R249S (b) were transfected to p53 null HCT116 cells with stably infected inducible SETDB1 shRNA. The cells were treated with or without doxycycline for 3 days after transfection. Then, the cells were treated with 50 μg ml⁻¹ cycloheximide and the turnover of p53 was measured at the time points indicated by western blot analysis using the antibody against total p53. (c) Similar experiments were performed to analyse the endogenous p53 turnover in HCCLM3 cells without p53 overexpression on SETDB1 knockdown. (d) SETDB1 knockdown increases p53 ubiquitination. HCCLM3 cells were treated with doxycycline as described above, and the cells were harvested for p53 immunoprecipitation and western blot analysis for p53 (left) or ubiquitin (right). The loading was normalized to total p53. (e) HCCLM3 cells infected with inducible SETDB1 shRNA were induced SETDB1 knockdown first and then treated with Doxorubicin before being harvested for immunoprecipitation. The cell lysates were immunoprecipitated with antibodies against total MDM2 or IgG control. The samples were then analysed using western blot analysis. Increased p53/MDM2 association was observed on SETDB1 knockdown. (f) SETDB1 knockdown reduces p53S15 phosphorylation. HCCLM3 cells were treated with doxycycline as described above, and the cells were harvested for western blot analysis on p53 phosphorylation (S15).

Figure 8. Inhibition of SETDB1 suppresses HCC cancer cell growth in vivo.

(a) HCCLM3 was stably infected with lentiviral SETDB1 shRNA. Overall, 3 × 10⁶ microplasma-free cells were subcutaneously injected to the flank region of the female athymic nude mice. When tumour size reached 100 mm³, mice were randomized and fed with doxycycline-containing drinking water (5% sucrose with 0.5 mg ml⁻¹ doxycycline, n=12) or sucrose only as the control (n=12). Doxycycline was changed twice every week. Tumour growth was measured using caliper. Body weight was monitored simultaneously. Data are presented as mean±s.e.m. P<0.001, t-test. (b) Total RNA was isolated from the xenograft tumour samples. Target knockdown in doxycycline-treated tumours was confirmed using RT–qPCR analysis as shown in the scatter plot. The blue line indicates the mean of each group. P<0.001, t-test, n=8. (c) Each of the four doxycycline-treated tumours or the control tumours were taken out for western blot analysis using antibody against SETDB1. Inhibition of SETDB1 protein expression was confirmed in doxycycline-treated tumours. (d) Immunohistochemical staining of SETDB1 in paraffin section of the xenograft tumour samples confirmed the target knockdown in doxycycline-treated tumours. Scale bar, 100 μM. The SETDB1 IHC signal was quantified using the Aperio algorithm. (e) Xenograft paraffin blocks were prepared and an H&E staining was applied to the sections for the pathological analysis. SETDB1 knocked down tumours appeared to be more differentiated than the controls. Shown here are representative fields. Scale bar, 100 μM.

Figure 9. Working model of how SETDB1 regulates cancer cell growth through methylation of p53.

GOF TP53 mutations, such as R249S, are relatively stable and often oncogenic. The stability of these GOF p53 can be enhanced through interaction with and being di-methylated by SETDB1 at K370, which results in less MDM2 association. In cancer cells bearing GOF p53 mutation and SETDB1 overexpression, attenuation of SETDB1 leads to less p53K370me2, enhanced p53 turnover and growth inhibition. Although SETDB1 can also interact with and methylate wild-type p53 at K370, it has little effect on cell growth as the wild type p53 is inherently very unstable and acts as a tumour suppressor. Therefore, SETDB1 can regulate cancer cell growth, at least in part, through methylation of GOF p53 at K370.