HCT116 colon cancer cell line was purchased from Horizon Discovery (Cambridge, UK). LoVo colon and SNU-1 gastric cancer cell lines were purchased from the American Type Culture Collection. HCT116 and SNU-1 cell lines were cultured in the RPMI-1640 medium (Sigma-Aldrich; Merck KGaA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Sigma-Aldrich; Merck KGaA). LoVo cell line was cultured in Ham's F-12 nutrient mixture medium (Sigma-Aldrich; Merck KGaA) containing 10% FBS. Trametinib was purchased from Cayman Chemical (Ann Arbor). Nutlin-3 was purchased from Calbiochem.

Control siRNA and MDM2- and MDM4-targeting DNA-modified siRNA sequences used in this study were adopted from a previous report (10). Reverse siRNA transfection was performed using lipofectamine RNAiMAX (Invitrogen; Thermo Fisher Scientific, Inc.), in accordance with the manufacturer's instructions. Transfection of the SNU-1 cell line was performed as described previously (9). Transfection effects of chiMDM2 and chiMDM4 are shown in Fig. S1.

Cell viability was determined using the WST-8 colorimetric assay with the Cell Count Reagent SF (Nacalai Tesque) as previously described (9). The combination index (CI) was determined by the Chou-Talalay method using the CalcuSyn software (Biosoft) (15). The CI values of <0.9, ≥0.9 and <1.1, and ≥1.1 were defined as the synergistic effect, additive effect, and antagonistic effect, respectively.

Protein extraction, sodium dodecyl sulfate polyacrylamide gel electrophoresis, and immunoblot analyses were performed as previously described (9). Twenty micrograms of protein samples were applied to gels. The primary antibodies used in this study are shown in Table SI. Cell cycle assay was performed using a Cycletest Plus DNA Reagent Kit (BD Biosciences) as previously described (16).

Human BCL2 cDNA was isolated from the pSVBT plasmid (a kind gift from Dr Tsujimoto) (17) by XhoI digestion and was cloned into pENTR1A (Invitrogen; Thermo Fisher Scientific, Inc.) at the SalI-XhoI sites. Then, BCL2 cDNA was subcloned into a lentivirus expression plasmid (pLenti6.3/V5-DEST, Invitrogen) with LR Clonase II (Invitrogen; Thermo Fisher Scientific, Inc.), and the resultant lentivirus plasmids were designated as BCL2/pLenti6.3. Lentiviruses were produced by the transfection of 293FT cells (Invitrogen; Thermo Fisher Scientific, Inc.) with BCL2/pLenti6.3 or EGFP/pLenti6.3 along with pLP1, pLP2, and pLP/VSVG (Invitrogen; Thermo Fisher Scientific, Inc.) mixtures using the Lipofectamine 2000 (Invitrogen) reagent according to the manufacturer's instructions. Cells were infected with lentivirus at a multiplicity of infection of five in the presence of 10 µg/ml of polybrene (Sigma-Aldrich; Merck KGaA) overnight and replaced with fresh media. After 48 h, the infected cells were selected using 10 µg/ml of blasticidin (Invitrogen; Thermo Fisher Scientific, Inc.).

Each experiment was performed in triplicate, and all data were shown as mean ± standard deviation. The Shapiro-Wilk test was used to evaluate whether data were normal distribution or not. The significance among three different groups was evaluated by one-way ANOVA. Then, for post hoc pairwise multiple comparisons, Tukey's test was used if Levene's test showed homogeneity of variance, and Games-Howell's test was used if not. P-value of <0.05 was considered a statistically significant difference. All statistical analyses were performed using the SPSS version 27.0 (SPSS, Inc.).

We examined whether chiMDM4 could enhance the antitumor effects of chiMDM2 and trametinib in colon (HCT116 and LoVo) and gastric (SNU-1) cancer cells harboring wt TP53 and mt KRAS (wt/G13D). As shown in Fig. 1, trametinib alone and chiMDM2 alone decreased cell viability in a dose-dependent manner. Further, the combination of chiMDM2 and trametinib induced synergistic antitumor effects. The CIs of chiMDM4, chiMDM2, and trametinib were calculated and summarized (Table I). The addition of chiMDM4 augmented the antitumor effects of the combination of chiMDM2 and trametinib, which suggested that the simultaneous inhibition of MDM2 and MDM4 might have an advantage over the inhibition of MDM2 alone.

Similarly, we examined the cell viability after the treatment of cell lines with chiMDM4, nutlin-3, and trametinib (Fig. 2). The CI values of chiMDM4, nutlin-3, and trametinib are listed in Table II. The addition of chiMDM4 synergistically enhanced nutlin-3 and trametinib mediated growth suppression in the tumor cell lines.

To explore the mechanism by which trametinib treatment and dual inhibition of MDM4/MDM2 exerted an enhanced antitumor activity, their effects on cell cycle distribution and apoptosis were analyzed using flow cytometry (Fig. S2). As shown in Fig. 3, trametinib alone and chiMDM4/chiMDM2 alone increased the cell fractions in the G1 phase, whereas it reduced those in the S phases in all cell lines in a similar manner, which indicated an induced G1 arrest. Simultaneous exposure to trametinib and chiMDM4/chiMDM2 induced a profound G1 arrest. Trametinib alone did not or only slightly increased the sub-G1 population, which is representative of cells undergoing apoptotic cell death, whereas chiMDM4/chiMDM2 increased the sub-G1 population moderately in HCT116 and SNU-1 cells and slightly in LoVo cells. The combination of trametinib and chiMDM4/chiMDM2 further increased the sub-G1 population in all three cell lines. These results suggested that trametinib enhanced chiMDM4/chiMDM2-induced G1 arrest and apoptosis.

The combination of trametinib and chiMDM4/chiMDM2 treatment induced the synergistic antitumor activity by enhancing G1 arrest and apoptosis in all cells with mt KRAS. To explore the mechanisms, proteins regulating cell cycle and apoptosis were examined in HCT116 cells using immunoblot analysis. Results were summarized using a heatmap (Fig. 4). If the upregulation or downregulation of the proteins by the combination of trametinib and chiMDM4/chiMDM2 was two-fold or more than the control and each agent alone, then it was considered as an important synergistic antitumor effect of the combination treatment. The proteins such as p53, p15, p27, and p53-activated proteins [p21, Fas, p53 upregulated modulator of apoptosis (PUMA), and 14-3-3 σ] were upregulated. On the other hand, the downregulated proteins included MDM2, p-ERK2, MYC, E2F1, and E2F1-activated proteins (cyclin A, cyclin B1, DNA polymerase δ, TYMS, CDC2, and CDC25A).

We also analyzed the expression of those proteins in LoVo and SNU-1 cell lines that were greatly upregulated or downregulated in HCT116 cells (Fig. 5). Fold change values of the proteins in all three cell lines are summarized in Table III. Notably, p-ERK2, which functions as a positive regulator of retinoblastoma (RB) phosphorylation and an inhibitor of cleavage of both caspase-8 and caspase-9, was suppressed by the combination of trametinib and chiMDM4/chiMDM2 treatment in all three cell lines (4.0–5.3-fold) compared to controls. With respect to the changes in protein expressions related to G1 arrest, the combination treatment markedly reduced phosphorylated RB (a master regulator of E2F-mediated transcription), cyclin A (one of the most efficiently activated proteins by E2F1), and MYC (a direct target of ERK1/2) in all three cell lines. ChiMDM4/chiMDM2 induced p53 and p21 in all three cell lines (Table III). In SNU-1 cells, as was observed in HCT116 cell, chiMDM4/chiMDM2 induced p53 and p21, which was enhanced by addition of trametinib. In LoVo cells, chiMDM4/chiMDM2 similarly induced p53 and p21. However, trametinib did not enhanced chiMDM4/chiMDM2-mediated p53 accumulation, or even decreased p21 accumulation. With regard to the changes in protein expressions related to apoptosis, chiMDM4/chiMDM2 induced moderate to high levels of Fas in HCT116 and SNU-1 cells, and a low level of Fas in LoVo cells (Table III). The addition of trametinib increased Fas levels in HCT116, LoVo, and SNU-1 cells and markedly increased cleaved caspase-8 in HCT116 cells, and slightly increased it in LoVo cells; however, no increase of cleaved caspase-8 was detected in SNU-1 cells until twice the amount of protein samples was used in immunoblotting. Even so, trametinib slightly increased cleaved caspase-8 in chiMDM4/chiMDM2-treated SNU-1 cells (Fig. S3). ChiMDM4/chiMDM2 induced PUMA expression in HCT116 and LoVo cells, but not in SNU-1 cells. Trametinib further enhanced the expression of PUMA that was induced by chiMDM4/chiMDM2 in HCT116 and LoVo cells (Table III). This combination treatment caused caspase-9 cleavage in HCT116 and LoVo cells, but not in SNU-1 cells.

Bcl2 is an antiapoptotic protein of Bcl2 family, regulating the mitochondria-mediated apoptosis in cells. Overexpression of Bcl2 antagonizes proapoptotic Bcl2 family members, such as PUMA, and represses caspase-9 activation by reducing the translocation of proapoptotic Bax and cytochrome C release, then inhibits the mitochondria-mediated apoptosis (18,19). To investigate whether activation of caspase-8 and caspase-9 was involved in the apoptosis caused by combination treatment, the effect of Bcl2 expression on apoptosis was analyzed in HCT116 and LoVo cells. BCL2 and a control EGFP were stably transduced in HCT116 and LoVo cells using lentiviruses (Fig. 6A). As shown in Figs. 6B and S4, the combined treatment using chiMDM4/chiMDM2 and trametinib increased the sub-G1 population from 5 to 14% in the control HCT116 cells, whereas this combination increased the sub-G1 fraction from 3 to 7% in BCL2-transduced HCT116 cells. This suggests that Bcl2 overexpression partially blocks apoptosis induction in HCT116 cells via the combination treatment strategy. The combined treatment increased the sub-G1 fraction from 5 to 10% in the control LoVo cells and from 4 to 6% in BCL2-transduced LoVo cells, suggesting that Bcl2 overexpression strongly blocked apoptosis in LoVo cells via the combination treatment strategy.