Abstract
Mcl-1, a prosurvival Bcl-2 family protein, is frequently overexpressed in cancer cells and plays a critical role in therapeutic resistance. It is well known that anticancer agents induce phosphorylation of Mcl-1, which promotes its binding to E3 ubiquitin ligases and subsequent proteasomal degradation and apoptosis. However, other functions of Mcl-1 phosphorylation in cancer cell death have not been well characterized. In this study, we show in colon cancer cells that histone deacetylase inhibitors (HDACi) induce GSK3β-dependent Mcl-1 phosphorylation, but not degradation or downregulation. The in vitro and in vivo anticancer effects of HDACi were dependent on Mcl-1 phosphorylation and were blocked by genetic knock-in of a Mcl-1 phosphorylation site mutant. Phosphorylation-dead Mcl-1 maintained cell survival by binding and sequestering BH3-only Bcl-2 family proteins PUMA, Bim, and Noxa, which were upregulated and necessary for apoptosis induction by HDACi. Resistance to HDACi mediated by phosphorylation-dead Mcl-1 was reversed by small-molecule Mcl-1 inhibitors that liberated BH3-only proteins. These results demonstrate a critical role of Mcl-1 phosphorylation in mediating HDACi sensitivity through a novel and degradation-independent mechanism. These results provide new mechanistic insights on how Mcl-1 maintains cancer cell survival and suggest that Mcl-1–targeting agents are broadly useful for overcoming therapeutic resistance in cancer cells.
Significance: These findings present a novel degradation–independent function of Mcl-1 phosphorylation in anticancer therapy that could be useful for developing new Mcl-1–targeting agents to overcome therapeutic resistance. Cancer Res; 78(16); 4704–15. ©2018 AACR.
Introduction
Stress-induced apoptosis in mammalian cells is regulated by the Bcl-2 family proteins. Proapoptotic BH3-only members such as PUMA, Bim, and Noxa sense and become activated by different stress signals (1, 2). They indirectly or directly activate the multi-BH domain containing proapoptotic members Bax/Bak, which triggers downstream events including mitochondrial outer membrane permeabilization (MOMP), cytosolic release of cytochrome c, and activation of caspases (1, 2). Myeloid cell leukemia 1 (Mcl-1) is a prosurvival Bcl-2 family member that is frequently overexpressed or amplified in various human tumors (3). Mcl-1 suppresses apoptosis by binding to the BH3 domains of proapoptotic proteins through a hydrophobic surface groove to inhibit MOMP and caspase activation (4, 5). A number of recent studies indicate a critical role of Mcl-1 in tumor cell survival and therapeutic resistance (5).
Mcl-1 protein stability and activity are regulated by posttranslational modifications including phosphorylation (6). Mcl-1 protein contains a proline/glutamic acid/serine/threonine (PEST) region subjected to phosphorylation (3). Phosphorylation of Mcl-1 by glycogen synthase kinase 3β (GSK3β) or other kinases promotes the binding of Mcl-1 to E3 ubiquitin ligases including F-box and WD repeat domain-containing 7 (FBW7), Mule, and β-TrCP, leading to Mcl-1 ubiquitination and subsequent proteasomal degradation (7–9). Several studies using overexpressed Mcl-1 mutants suggest that Mcl-1 phosphorylation also affects its antiapoptotic activity and interactions with other Bcl-2 proteins (7, 10, 11). However, it remains unclear if and how Mcl-1 phosphorylation affects its endogenous function and interactions with other Bcl-2 family members and apoptosis.
Histone deacetylases (HDAC) are epigenetic enzymes that regulate gene expression by removing the acetyl group from histones. They are well-established targets in anticancer therapy (12, 13). Several HDAC inhibitors (HDACi), such as suberoylanilide hydroxamic acid (SAHA; Vorinostat), have been approved for treating hematopoietic malignancies (12,13). However, HDACi as single agents are generally ineffective for solid tumors, and resistance to HDACi represents a major obstacle in their clinical applications (12). Induction of apoptosis has been implicated as a key effect of HDACi (14). HADCi-induced apoptosis was shown to require the BH3-only proteins Bim and Noxa in leukemia cells (15, 16). Recent studies suggest that Mcl-1 is a major determinant of HADCi sensitivity in some leukemia and solid tumor cells (17–19). Nonetheless, how Mcl-1 mediates sensitivity and resistance to HDACi has remained unclear. The functional relationship between Mcl-1 and other Bcl-2 family proteins in therapeutic response to HDACi is obscure.
In this study, we used a genetic knock-in (KI) approach to identify a critical role of GSK3β-mediated Mcl-1 phosphorylation in apoptosis induced by HDACi in colon cancer cells. Our results demonstrate that Mcl-1 phosphorylation determines HDACi sensitivity by suppressing its bindings to PUMA, Bim, and Noxa, independent of ubiquitination and proteasomal degradation. They provide novel mechanistic insight on how Mcl-1 maintains cancer cell survival, and new clues for targeting Mcl-1 to overcome therapeutic resistance.
Materials and Methods
Cell culture
Colon cancer cell lines, including HCT116, DLD1, and RKO, were obtained from the ATCC. HCT116 cells with knock-in of the Mcl-1 phosphorylation site mutant S121A/E125A/S159A/T163A (Mcl-1-KI) were generated by homologous recombination as described previously (20). PUMA-knockout (KO) and FBW7-KO HCT116 and DLD1 cells were described previously (20, 21). Cells were authenticated in 2017 by genotyping and analysis of protein expression by Western blotting, and routinely checked for Mycoplasma contamination by PCR. All cell lines were maintained at 37°C in 5% CO2 and cultured in McCoy 5A modified media (Invitrogen) supplemented with 10% defined FBS (HyClone), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen). For drug treatment, cells were plated in 12-well plates at 20%–30% density 24 hours before treatment. DMSO (Sigma) stocks of SAHA, MS-275, SB216763, ABT-737, ABT-263, sunitinib, sorafenib, regorafenib, TW-37 (Selleck Chemicals), and Mcl-1 inhibitors including UMI-77 (22), UMI-212 (compound 21), and UMI-36 (compound 36; ref. 23), were diluted to appropriate concentrations with cell culture medium.
Targeting Bim and Noxa in HCT116 cells
Gene targeting vectors were constructed by using the recombinant adeno-associated virus (rAAV) system (24). Briefly, two homologous arms flanking exon 2 of Bim or Noxa, along with a neomycin-resistant gene cassette (Neo), were inserted between two Not I sites in the AAV shuttle vector pAAV-MCS (Agilent Technologies). Packaging of rAAV was performed by using the AAV Helper-Free System (Agilent) according to the manufacturer's instructions. HCT116 cells containing two copies of Bim and Noxa were infected with the rAAV and selected by G418 (0.5 mg/mL, Mediatech) for 3 weeks. Drug-resistant clones were pooled and screened by PCR for targeting events with the following primers for Bim: P1, 5′-TGATTGGATGTATTCAGAGG/P2, 5′-CATTACTACAGCACTCTCCTC, or for Noxa: P1, 5′- GCGAAGGAAGTGGTGCATTG/P2, 5′- CTGAAGAAAACAGATGTGAGG, in combination with a primer for Neo 5′-TCTTGACGAGTTCTTCTGAG/5′-TTGTGCCCAGTCATAGCCG. Prior to targeting the second allele, the Neo cassette, which is flanked by Lox P sites, was excised from a heterozygous clone by infection with an adenovirus expressing Cre recombinase (Ad-Cre). Single clones were screened by PCR for Neo excision using P1/P2, and two independent positive clones were infected again with the Bim- or Noxa-targeting construct. After the second round, Neo was excised by Ad-Cre infection, and gene targeting was verified by PCR and Western blotting.
MTS assay
Cells seeded in 96-well plates at a density of 1 × 104 cells/well were treated with SAHA or MS-275 for 72 hours. 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay was performed using the MTS Assay Kit (Promega) according to the manufacturer's instructions. Chemiluminescence was measured by using a Wallac Victor 1420 Multilabel Counter (Perkin Elmer). Each assay was conducted in triplicate and repeated three times.
Western blotting
Western blotting was performed as described previously (25). Antibodies include those for PUMA, phospho-Mcl-1 (S159/T163), cleaved caspases 3, 8, and 9, ERK, phospho-ERK (T202/Y204), AKT, phospho-AKT (S473), GSK3β, phospho-GSK3β (S9), Bim, Noxa, Bad (Cell Signaling Technology), V5, cytochrome oxidase subunit IV (Invitrogen), Bax, HA (Santa Cruz Biotechnology), Bcl-XL, Mcl-1 (BD Biosciences), Bid (EMD Biosciences), cytochrome c, β-actin (Sigma), Bak (Millipore), Bcl-2 (Dako), and FBW7 (Bethyl Laboratories).
Transfection and siRNA knockdown
Transfection was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Expression constructs of HA-tagged PUMA, Bim, and Noxa were described previously (25). Mcl-1 expression construct was generated by cloning a PCR-amplified full-length human Mcl-1 cDNA fragment into pcDNA3.1/V5-His vector (Invitrogen; ref. 20). Mutations were introduced into Mcl-1 using QuickChange XL Site-Directed Mutagenesis Kit (Agilent Technologies). siRNA transfection was done 24 hours before drug treatment using 200 pmol of control scrambled siRNA or human Mcl-1 siRNA (CGCCGAATTCATTAATTTATT-dTdT; Dharmacon).
Immunoprecipitation
After treatment, cells were harvested and resuspended in 1 mL of EBC buffer (50 mmol/L Tris-HCl, pH 7.5, 100 mmol/L NaCl, 0.5% Nonidet P-40) supplemented with a protease inhibitor cocktail (Roche). Cell suspensions were sonicated and spun at 10,000 × g for 10 minutes to prepare cell lysates. For immunoprecipitation (IP), 1–2 μg of IP antibodies were mixed with protein G/A-agarose beads (Invitrogen) for 20 minutes at room temperature. The beads were washed twice with PBS containing 0.02% Tween 20 (pH 7.4), incubated with cell lysates on a rocker for 6 hours at room temperature, and then washed three times with PBS (pH 7.4). Beads were then boiled in 2× Laemmli sample buffer and subjected to SDS-PAGE and Western blot analysis.
Reverse transcription PCR
Total RNA was isolated using the Mini RNA Isolation II kit (ZYMO Research) according to the manufacturer's protocol. One-μg of total RNA was used to generate cDNA by the SuperScript II reverse transcriptase (Invitrogen). PCR was performed using previously described conditions (26). PCR primers include those for PUMA: 5′-CGACCTCAACGCACAGTACGA-3′/5′-AGGCACCTAATTGGGCTCCAT-3′; Bim: 5′-GGAGACGAGTTTAACGCTTAC-3′/5′-AAGCAAAATGTCTGCATGG-3′; Noxa: 5′-TTCAGCTCGCGTCCTGCAG-3′/5′-GTTCCTGAGCAGAAGAGTTTGG-3′; Mcl-1: 5′-ATGCTTCGGAAACTGGACAT-3′/5′-TGGAAGAACTCCACAAACCCA-3′; FBW7: 5′-GTGATAGAACCCCAGTTTCA-3′/5′-CCTCAGCCAAAATTCTCCAG-3′; Mule: 5′-CAGTTCCATAGAGCATTTGA-3′/5′-GCCATTCTCCTTTCCACCCC-3′; and β-actin: 5′-GACCTGACAGACTACCTCAT-3′/5′-AGACAGCACTGTGTTGGCTA-3′.
Analysis of apoptosis
Apoptosis was measured by counting condensed and fragmented nuclei after nuclear staining with Hoechst 33258 (Invitrogen) as previously described (25). At least 300 cells were analyzed for each sample. Colony formation assays were performed by plating treated cells in 12-well plates at appropriate dilutions, followed by crystal violet staining 14 days after plating as described previously (27). Each experiment was performed in triplicate and repeated at least twice. Annexin V/propidium iodide (PI) staining was performed using Annexin-Alexa Fluor 488 (Invitrogen) and PI as described previously (27). For analysis of cytochrome c release, cytoplasmic and mitochondrial fractions were separated by Mitochondrial Fractionation Kit (Active Motif) according to the manufacturer's instructions, followed by Western blotting of cytochrome c in the cytoplasmic and mitochondrial fractions.
Xenograft tumor experiments
All animal experiments were approved by the University of Pittsburgh Institutional Animal Care and Use Committee. Female 5- to 6-week-old Nu/Nu mice (Charles River) were housed in micro isolator cages in a sterile environment, and allowed access to water and chow ad libitum. Xenograft tumors were established by subcutaneously injecting mice with 4 × 106 of WT or Mcl-1-KI HCT116 cells. After tumor growth for 7 days, mice were treated daily for 10 consecutive days with SAHA (50 mg/kg/day) or the control vehicle. Tumor volumes were measured by calipers and calculated according to the formula 1/2 × length × width2. After tumors reached 1.0 cm3 in size, mice were euthanized and tumors were dissected and fixed in 10% formalin and embedded in paraffin. Terminal deoxynucleotidyl transferase mediated dUTP Nick End Labeling (TUNEL; Millipore) and active caspase-3 (Cell Signaling Technology) immunostaining was performed on 5-mm paraffin-embedded tumor sections. Signals were detected by Alexa Fluor 488–conjugated secondary antibody (Invitrogen) with nuclear counter staining by 4′ 6-Diamidino-2-phenylindole (DAPI).
Interactions of PUMA and Mcl-1 were detected by in situ proximity ligation assay (PLA) in paraffin-embedded sections of WT or Mcl-1-KI HCT116 tumor tissues from nude mice treated with SAHA for 1 day. PLA was performed using the Duolink In Situ Red Starter Kit (Sigma) according to the manufacturers’ instructions. The sections were incubated with PUMA (rabbit, 1:50; Abcam) and Mcl-1 (mouse, 1:100; BD Biosciences) antibodies for at 4°C overnight, and then mounted with Vectashield Mounting Medium (Vector Laboratories) with DAPI (Sigma) for nuclear counter staining. PLA signals were visualized by fluorescence microscopy (Olympus) and quantified by counting.
Statistical analysis
Statistical analysis was carried out using GraphPad Prism VI software. P values were calculated by the Student t test and were considered significant if P < 0.05. Means ± SD were displayed in the figures.
Results
HDACi induce GSK3β-dependent Mcl-1 phosphorylation and apoptosis, but not Mcl-1 degradation or downregulation, in colon cancer cells
We recently showed that treating HCT116 colon cancer cells with a multi-kinase inhibitor, such as regorafenib, sorafenib, UCN-01, or sunitinib, induced S159/T163 phosphorylation of Mcl-1, which is essential for Mcl-1 ubiquitination, proteasomal degradation, and subsequent apoptosis induction (20). In addition to the kinase inhibitors, the HDACi SAHA and MS-275 were also found to induce Mcl-1 phosphorylation (Fig. 1A), which correlated with apoptosis induction (Fig. 1B, right). However, unlike the kinase inhibitors, HDACi treatment did not promote Mcl-1 degradation as shown by unchanged Mcl-1 protein expression (Fig. 1A and C) and half-life (Supplementary Fig. S1A). Mcl-1 mRNA expression was slightly increased in response to SAHA, but unchanged after MS-275 treatment (Supplementary Fig. S1B). HDACi treatment also induced Mcl-1 phosphorylation without degradation in DLD1 and RKO colon cancer cells, and even increased total Mcl-1 level in DLD1 cells after 16 hours (Supplementary Fig. S1C and S1D). These intriguing observations suggest that the level of total Mcl-1 may not be the key determinant of its activity, and prompted us to further investigate the mechanism and functional role of Mcl-1 phosphorylation in HDACi-induced apoptosis.
HDACi-induced Mcl-1 phosphorylation requires GSK3β, and could be blocked by the GSK3 inhibitor SB216763 or GSK3β knockdown by siRNA of (Fig. 1B, left; Supplementary Fig. S1E, left). SAHA or MS-275 treatment reduced ERK phosphorylation (T202/Y204), and relived ERK-mediated inhibitory phosphorylation (S9) of GSK3β (Fig. 1A; ref. 28). Mcl-1 phosphorylation by GSK3β or other kinases was shown to involve 3 sites including S121, S159, and T163 (8, 9). Phosphorylation-mediated binding of Mcl-1 to FBW7 could be blocked by a quadruple mutant of these 3 sites and the phospho-mimic site E125 (4A mutant: S121A/E125A/S159A/T163A; ref. 9). Transfection of the 4A mutant (20), not wild-type (WT) Mcl-1, completely blocked SAHA- and MS-275–induced Mcl-1 phosphorylation (Fig. 1D). Furthermore, GSK3β inhibition or knockdown suppressed apoptosis induced by SAHA or MS-275 (Fig. 1B, right; and Supplementary Fig. S1E, right). Compared with WT Mcl-1, the 4A mutant Mcl-1 expressed at a similar level had the stronger effect of suppressing HDACi-induced apoptosis in HCT116 cells (Fig. 1E), which was confirmed in DLD1 and RKO cells (Supplementary Fig. S1F and S1G). These results suggest a critical role GSK3β-mediated Mcl-1 phosphorylation in apoptotic response to HDACi.
Knock-in of Mcl-1 phosphorylation site mutant suppresses HDACi-induced apoptosis
We used a genetic knock-in (KI) approach to further delineate the functional role of Mcl-1 phosphorylation independent of degradation (Supplementary Fig. S2A). HCT116 cells with KI of Mcl-1 phosphorylation site mutant (Mcl-1-KI; ref. 20) were completely resistant to SAHA- and MS-275–induced phosphorylation (Fig. 1F), and had suppressed nuclear fragmentation (Fig. 2A), and increased viability (Fig. 2B) and clonogenic survival (Fig. 2C). Blocked apoptosis was further confirmed by reduced Annexin V staining (Supplementary Fig. S2B), caspase-3, 8, and 9 activation (Fig. 2D) and cytochrome c release (Fig. 2E). Consistent with Mcl-1 dependency, knockdown of Mcl-1 restored SAHA and MS-275 sensitivity in Mcl-1-KI cells (Fig. 2A and B; Supplementary Fig. S2B). Mcl-1-KI cells were also resistant to the kinase inhibitors that induce Mcl-1 degradation, including regorafenib, sorafenib, UCN-01, and sunitinib (Supplementary Fig. S2C; ref. 20), but remained sensitive to other anticancer agents such as TRAIL, etoposide, and sulindac (20). These results demonstrate that Mcl-1 phosphorylation, but not its degradation, is critical for HDACi-induced apoptosis.
To better understand unaltered Mcl-1 stability in response to HDACi, we examined the E3 ubiquitin ligases of Mcl-1, including FBW7 and Mule. SAHA or MS-275 at the apoptosis-inducing concentration did not affect the expression of FBW7 and Mule, in contrast to their marked induction by regorafenib and sorafenib (Supplementary Fig. S3A–S3C). SAHA- and MS-275-induced apoptosis was not suppressed, but slightly enhanced in isogenic FBW7-knockout (KO) HCT116 and DLD1 cells, which had reduced apoptotic response to regorafenib and other kinase inhibitors (Supplementary Fig. S3D and S3E; ref. 20). Furthermore, transfection of WT FBW7, but not the tumor-derived inactivating mutants including R465C and R505C (29), was sufficient to promote Mcl-1 degradation in HDACi-treated HCT116 cells (Supplementary Fig. S3F), indicating that insufficient expression of FBW7 or other Mcl-1 E3 ligases is responsible for the lack of Mcl-1 degradation in response to HDACi. These results suggest that depending on specific stimuli and whether an E3 ligase of Mcl-1 is induced, Mcl-1 phosphorylation can promote apoptosis through a degradation-dependent or -independent mechanism.
PUMA, Bim, and Noxa are induced and required for HDACi-induced apoptosis
To determine how Mcl-1 phosphorylation mediates HDACi-induced apoptosis, we analyzed other Bcl-2 family members. Interestingly, we found that three BH3-only proteins, including PUMA, Bim, and Noxa, were all induced in response to HDACi treatment (Fig. 3A; Supplementary Fig. S1C). PUMA, Bim, and Noxa mRNA expression was also markedly elevated following SAHA or MS-275 treatment (Supplementary Fig. S4A–C). In contrast, other proapoptotic and antiapoptotic members, including Bad, Bid, Bax, Bak, Bcl-2, and Bcl-XL, were not substantially altered in response to SAHA and MS-275 treatment (Fig. 3A).
We then analyzed isogenic HCT116 cells with KO of PUMA (21), Bim (Supplementary Fig. S4D), or Noxa (Supplementary Fig. S4E) to determine whether the induced BH3-only proteins are necessary for HDACi-induced apoptosis. Indeed, deletion of PUMA, Bim, or Noxa led to substantially increased cell viability (Fig. 3B), decreased nuclear fragmentation (Fig. 3C) and Annexin V staining (Supplementary Fig. S5A), blocked activation of caspases 3 and 9 (Fig. 3D), and markedly improved clonogenic survival (Fig. 3E), in response to SAHA or MS-275. Decreased sensitivity to HDACi was also observed in PUMA-KO DLD1 cells (Supplementary Fig. S5B). Consistent with the functions of BH3-only proteins in activating Bax and/or Bak, HDACi-induced apoptosis and caspase activation were suppressed in BAX-KO HCT116 cells (Supplementary Fig. S5C and S5D). Therefore, HDACi induce apoptosis in colon cancer cells by upregulating multiple BH3-only proteins, which cooperatively activate Bax to cause caspase activation.
Phosphorylation-dead Mcl-1 binds to PUMA, Bim, and Noxa to inhibit cell death, which is abrogated by small-molecule Mcl-1 inhibitors
We further investigated whether Mcl-1 phosphorylation regulates the activity of BH3-only proteins in HDACi-induced apoptosis through protein–protein interactions. Following SAHA treatment, phosphorylation-dead Mcl-1 showed markedly enhanced binding to PUMA, Bim, and Noxa (Fig. 4A), which led to reduced interaction of Bcl-XL with PUMA, Bim, and Noxa, but increased interaction of Bcl-XL with Bax (Fig. 4B). These differences in binding were confirmed by reciprocal IP using anti-Bim or anti-Noxa antibody (Supplementary Fig. S6A and S6B). Similar changes were observed in MS-275–treated cells (Supplementary Fig. S6C). To further determine how Mcl-1 phosphorylation affects the binding to BH3-only proteins, WT or the 4A mutant Mcl-1, along with PUMA, Bim, or Noxa, was transfected into HCT116 cells. IP analysis revealed that the mutant Mcl-1 bound to substantially more PUMA, Bim, and Noxa compared with WT Mcl-1 (Fig. 4C), indicating an intrinsic difference in binding capacity to BH3-only proteins between WT and the mutant Mcl-1. Furthermore, inhibiting Mcl-1 phosphorylation by the GSK3 inhibitor SB216763 enhanced the binding of endogenous Mcl-1 to PUMA, Bim, and Noxa (Fig. 4D). Because E125 is a phospho-mimic but not a phosphorylation site, we also generated and analyzed the E125A mutant, and found this mutant did not affect the binding of Mcl-1 to PUMA and Bim (Supplementary Fig. S6D). Together, our data suggest that Mcl-1 phosphorylation liberates PUMA, Bim, and Noxa upon their induction, and allows them to bind to Bcl-XL or other prosurvival factors to promote apoptosis.
Mcl-1 phosphorylation is likely compromised in drug-resistant cancer cells due to insufficient GSK3β activity caused by aberrant ERK activation. We therefore tested whether HDACi resistance mediated by the lack of Mcl-1 phosphorylation could be overcome by direct inhibition of Mcl-1. Indeed, treating cells with an Mcl-1-selective small-molecule inhibitor, including UMI-77 (22), UMI-212, or UMI-36 (23), or the pan-Bcl-2 inhibitor TW-37 with a strong effect on Mcl-1 (30), restored SAHA- and MS-275–induced apoptosis and caspase activation in Mcl-1-KI cells (Fig. 5A and B; Supplementary Fig. S7). IP analysis showed TW-37 treatment restored the binding of endogenous PUMA, Bim, and Noxa to Bcl-XL (Fig. 5C). These results suggest that Mcl-1 inhibitors could be useful for enhancing the sensitivity to anticancer agents such as HDACi, which induce Mcl-1 phosphorylation but not degradation.
Mcl-1 phosphorylation without degradation mediates the antitumor effects of SAHA in vivo
We then used xenograft models to analyze the in vivo effects of Mcl-1 phosphorylation and HDACi sensitivity. Compared with WT HCT116 tumors, Mcl-1-KI tumors were significantly more resistant to SAHA treatment (Fig. 6A and B). Apoptosis was markedly reduced in Mcl-1-KI tumors compared with WT tumors as indicated by TUNEL and active caspase-3 staining (Fig. 6C and D). Consistent with the findings from cultured cells, SAHA treatment did not cause Mcl-1 degradation, but induced the expression of PUMA, Bim and Noxa in WT and Mcl-1-KI xenograft tumors (Fig. 6E). Probing in situ protein–protein interactions by PLA (31) revealed a significantly enhanced interaction between the mutant Mcl-1 and PUMA in Mcl-1-KI tumors (Fig. 6F). These results confirmed the pivotal role of Mcl-1 phosphorylation in mediating the in vivo antitumor effects of HDACi by liberating PUMA and other BH3-only proteins, independent of Mcl-1 degradation.
Discussion
A hallmark of cancer is resistance to apoptosis, which maintains survival of cells en route to oncogenic transformation (32). Overexpression or amplification of Mcl-1 is one of the most frequent alterations in human cancers and underlies development of therapeutic resistance by evading cell death (33). Among the antiapoptotic Bcl-2 family proteins, Mcl-1 regulation is uniquely dynamic and complex. Mcl-1 harbors a long unstructured N-terminus containing the PEST region, which includes many phosphorylation sites. Phosphorylation of Mcl-1 at these sites by GSK3β, p38, JNK, CDK1, casein kinase II, or other kinases may have a variety of functional roles in apoptosis induced by different stress conditions including therapeutic treatment (7–9).
Our study provides convincing evidence that drug-induced Mcl-1 phosphorylation directly affects the interactions of Mcl-1 with multiple BH3-only proteins to inhibit other antiapoptotic members and activate Bax/Bak. Our biochemical and genetic data demonstrate that HDACi require Mcl-1 phosphorylation by GSK3β on at least 3 sites, including S121, S159, and T163, to induce apoptosis in colon cancer cells. The E125 site analyzed in the 4A mutant and Mcl-1-KI cells had no effect on these interactions. Mcl-1 phosphorylation unleashes the full proapoptotic activity of PUMA, Bim, and Noxa, and allows them to bind to and neutralize other prosurvival factors such as Bcl-XL, which plays a key role in colon cancer cell survival (34). This is in contrast to apoptosis induced by the kinase inhibitors, which largely relies on a single BH3-only protein PUMA (20), but is enhanced by FBW7-mediated Mcl-1 degradation to achieve a threshold level for cell death. Therefore, whether Mcl-1 degradation is necessary for apoptosis induction may be determined by the strength of BH3 signals and the induction of FBW7 or other E3 ligases (Fig. 7). If the BH3 signal is strong enough, such as in cases where multiple BH3-only proteins are induced, Mcl-1 only needs to be phosphorylated to release sufficient BH3 signals to inactivate key prosurvival factors to induce cell death. If the BH3 signal is not very strong, such as when a single BH3-only protein is activated, an E3 ligase needs to be activated to promote Mcl-1 degradation for unleashing sufficient BH3 signal for cell killing (Fig. 7).
There are a number of published three-dimensional structures of the complexes formed between Mcl-1 and BH3 domains of BH3-only proteins, including PUMA, Bim, and Noxa (35–37). All of these structures were obtained from recombinant Mcl-1 with deletions of N-terminal 153 residues including the PEST region, as well as the C-terminal 11 or 23 residues including the transmembrane anchor. The Mcl-1 protein used in these studies retains the BH3-binding groove and thus the binding affinity and selectivity for the BH3 peptides derived from BH3-only proteins (36, 38). While these complex structures provided essential knowledge about the interactions in the BH3 domain, they are not informative for the interactions between the PEST region and BH3-only proteins, which can be used as a template to build a model to provide insights on the interactions between the mutant Mcl-1 and BH3-only proteins. However, it is well established that posttranslational modifications such as phosphorylation are specifically recognized by interacting proteins, and can modulate the strength of the interactions, conformational changes, steric constraints, and thus potentially affect the protein–protein interactions (39–41). Indeed, in Mcl-1-KI cells, phosphorylation-deficient Mcl-1 has increased binding to PUMA, Bim and Noxa, which frees Bcl-XL and other antiapoptotic proteins to more efficiently inhibit Bax/Bak. In line with our findings, several previous studies using transfected mutant Mcl-1 suggest that Mcl-1 phosphorylation has other effects, in addition to promoting the binding to E3 ligases and degradation. For example, cophosphorylation of S121 and T163 impaired its antiapoptotic function in response to H2O2 treatment (11). GSK3-mediated phosphorylation at S159 inhibits the interaction of Mcl-1 with Bim (7). Cophosphorylation of Mcl-1 at S159 and T163 was shown to reduce the antiapoptotic function of Mcl-1 (10). Our results reveal the relevant context for further investigating the degradation-independent function of Mcl-1 phosphorylation that might be amenable for drug development and optimization. However, it should be noted that the observed effects of HDACi on Mcl-1 may not be generalizable to all cell types, as the mechanisms of apoptosis are dependent on cell types and cellular contexts. For example, Mcl-1 accumulation caused by FBW7 deficiency had opposite effects on sorafenib-induced apoptosis in colon cancer cells and T-cell acute lymphoblastic leukemia cells (8, 29).
SAHA and other HDACi are approved by the FDA for use in cutaneous T-cell leukemia (12, 13). However, HDACi alone are not effective for treating solid tumors. Overexpression of Bcl-2 has previously been shown to cause HADCi resistance (12). The dependence on Mcl-1 phosphorylation suggests Mcl-1 as a critical mediator of HDACi sensitivity and resistance. Similar to our findings, previous studies showed that SAHA induces PUMA, Bim, and Noxa, and that Mcl-1 is a dominant prosurvival factor in head and neck squamous cell carcinoma (HNSCC) cells (19). In this case, FBW7 mutations also did not protect, but sensitize HNSCC cells to SAHA, which may be due to the effects of FBW7 on other substrates such as Jun, Myc, cyclin E, and Notch 1 (42). SAHA has synergy with BH3 mimetics in killing HNSCC cells (19), also suggesting a threshold level of BH3 signal as the key determinant of SAHA sensitivity. In leukemia cells, PUMA, Bim, and Noxa could also be induced by SAHA (16). However, only Bim, but not PUMA and Noxa, was required for SAHA-induced apoptosis (16), reflecting the cell type–specific function of these proteins. Furthermore, the E3 ubiquitin ligase Mule was shown to mediate HDACi sensitivity in MEF cells by promoting the ubiquitination and degradation of HDAC2 (43). These findings further support the critical role of Bcl-2 family in HDACi-induced therapeutic response and cancer type–specific resistance mechanisms.
Inhibiting prosurvival Bcl-2 family proteins in tumor cells has emerged as an attractive therapeutic strategy, leading to recent approval of the Bcl-2-selective inhibitor ABT-199 (Venetoclax) for the treatment of chronic lymphocytic leukemia (44). However, ABT-199 and related agents do not bind to Mcl-1, and resistance to these agents quickly emerges due to overexpression of Mcl-1 (45–47). There is an urgent need for developing small-molecule Mcl-1 inhibitors. Several different chemical classes of Mcl-1 inhibitors have been described (22, 48, 49). Our data suggest that Mcl-1-binding BH3 mimetics can be used to overcome therapeutic resistance caused by compromised Mcl-1 phosphorylation. Mcl-1 inhibitors may be broadly useful for potentiating many anticancer agents that promote Mcl-1 phosphorylation, even without a change in Mcl-1 stability or expression. The isogenic Mcl-1-KI cells lacking drug-induced phosphorylation can help identify new inhibitors or drug combinations with improved potency and on-target activity.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: J. Yu, L. Zhang
Development of methodology: X. Zheng, J. Yu, L. Zhang
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Tong, R. Fletcher
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Tong, X. Tan, J. Yu, L. Zhang
Writing, review, and/or revision of the manuscript: J. Tong, X. Tan, Z. Nikolovska-Coleska, J. Yu, L. Zhang
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Tan, J. Yu, L. Zhang
Study supervision: L. Zhang
Other (provided key reagents): Z. Nikolovska-Coleska
Acknowledgments
We thank our lab members for critical reading and discussion. This work is supported by NIH grants (R01CA172136 and R01CA203028 to L. Zhang; U19AI068021 and R01CA215481 to J. Yu; R01CA149442 to Z. Nikolovska-Coleska; R01CA217141 to Z. Nikolovska-Coleska and L. Zhang). This project used the UPMC Hillman Cancer Center shared facilities that were supported in part by award P30CA047904.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.