Discovery of a novel, highly potent EZH2 PROTAC degrader for targeting non-canonical oncogenic functions of EZH2

Abstract

Aberrant expression of EZH2, the main catalytic subunit of PRC2, has been implicated in numerous cancers, including leukemia, breast, and prostate. Recent studies have highlighted non-catalytic oncogenic functions of EZH2, which EZH2 catalytic inhibitors cannot attenuate. Therefore, proteolysis-targeting chimera (PROTAC) degraders have been explored as an alternative therapeutic approach to suppress both canonical and non-canonical oncogenic activity. Here we present MS8847, a novel, highly potent EZH2 PROTAC degrader that recruits the E3 ligase von Hippel-Lindau (VHL). MS8847 degrades EZH2 in a concentration-, time-, and ubiquitin–proteasome system (UPS)-dependent manner. Notably, MS8847 induces superior EZH2 degradation and anti-proliferative effects in MLL-rearranged (MLL-r) acute myeloid leukemia (AML) cells compared to previously published EZH2 PROTAC degraders. Moreover, MS8847 degrades EZH2 and inhibits cell growth in triple-negative breast cancer (TNBC) cell lines, displays efficacy in a 3D TNBC in vitro model, and has a pharmacokinetic (PK) profile suitable for in vivo efficacy studies. Overall, MS8847 is a valuable chemical tool for the biomedical community to investigate canonical and non-canonical oncogenic functions of EZH2.

1. Introduction

Epigenetic regulation plays a critical role in cancer initiation, progression, and metastasis [1]. Therefore, targeting aberrant epigenetic factors using small molecules is an attractive therapeutic strategy. Protein lysine methyltransferases (PKMTs) catalyze lysine methylation by transferring the methyl group from the co-factor S-5′-adenosyl-l-methionine (SAM), leading to various histone methylation marks corresponding to transcriptional activation or repression [2]. As such, dysregulated or mutated PKMTs have been heavily implicated in the onset and progression of cancer [1,2].

EZH2 (Enhancer of Zeste Homolog 2), EED, and SUZ12 are core components of the polycomb repressive complex 2 (PRC2), which regulates chromatin compaction through tri-methylation of histone H3 lysine 27 (H3K27me3), a transcriptionally silent mark [3]. Mutated or overexpressed EZH2, the main catalytic subunit of PRC2, has been implicated in numerous cancers, including breast [4], leukemia [5], gastric [6], ovarian [7], lung [8], lymphoma [9] and prostate [10]. Thus, EZH2 is a therapeutic target of high interest [11,12].

Over the past decade, many catalytic inhibitors of EZH2 have been developed [13,14], including the Food and Drug Administration (FDA)-approved Tazemetostat (EPZ-6438) [15]. While some of these inhibitors have achieved anti-cancer activity, they face challenges including EZH2 mutations and chemoresistance [16–19]. Moreover, recent studies have identified non-catalytic functions of EZH2 in cancers such as MLL1-rearranged acute myeloid leukemia (MLL-r AML) [20], breast [21], castration-resistant prostate [22–24], SWI/SNF-mutant cancer [25], multiple myeloma [26], and lung [27]. Due to these non-catalytic oncogenic activities, treatment of these cancers with EZH2 enzymatic inhibitors does not produce robust anti-cancer effects. To target the non-catalytic oncogenic activities of EZH2, a new therapeutic approach is needed.

To effectively target all oncogenic functions of EZH2, we enlisted the Proteolysis Targeting Chimera (PROTAC) approach. PROTACs have become a promising therapeutic strategy for targeting oncoproteins with non-catalytic activity [28–31]. These heterobifunctional small molecules are comprised of a protein of interest (POI) binder linked to an E3 ligase ligand. The PROTAC-induced proximity between the POI and E3 ligase leads to polyubiquitination of the POI, triggering its degradation via the 26S proteasome [28–30]. Our group has developed multiple degraders targeting EZH2 [20,24,26,32,33] and several additional PRC2-targeting PROTAC degraders have since been developed by others (Fig. 1) [34], including the first-in-class EZH2 selective degrader MS1943, which is a hydrophobic tag-based EZH2 degrader [33], von Hippel-Lindau (VHL)-recruiting EZH2 PROTACs YM281 [35] and MS8815 [32], and cereblon (CRBN)-recruiting EZH2 PROTACs E7 [36], MS177 [20], and U3i [37]. In addition, several EED-binding PRC2 PROTAC degraders such as PROTAC-1 [38] and UNC6852 [39] have been reported (Fig. 1). Furthermore, we recently reported MS147 [40], an EED-binding PROTAC that preferentially degrades polycomb repressive complex 1 (PRC1) over PRC2 (Fig. 1).

Fig. 1.

Chemical structures of EZH2- and EED-binding degraders. MS1943 is a hydrophobic tag-based EZH2 degrader. YM281 and MS8815 are VHL-recruiting EZH2 PROTAC degraders. E7, MS177, and U3i are CRBN-recruiting EZH2 PROTAC degraders. PROTAC-1 and UNC6852 are EED-binding PRC2 PROTAC degraders. MS147 is an EED-binding PRC1 PROTAC degrader. Parent inhibitors are colored in pink, linkers in black and the E3 ligase ligands in blue.

Here, we report the discovery and characterization of a novel, highly potent EZH2 PROTAC degrader, MS8847, which is based on EPZ-6438 and recruits the VHL E3 ligase with an optimized linker composition. We characterized MS8847 in MLL-r AML and TNBC cell lines because we and others uncovered the non-canonical oncogenic functions of EZH2 in MLL-r AML [20] and TNBC [21]. Importantly, we show that MS8847 induces superior EZH2 degradation and anti-proliferative effects in multiple MLL-r AML cell lines compared to previously reported EZH2 PROTACs such as MS8815, YM281, U3i and E7. MS8847 also degrades EZH2 and inhibits cell growth in TNBC cell lines and has a mouse PK profile suitable for in vivo efficacy studies.

2. Results

2.1. Design and synthesis of VHL-recruiting EZH2 PROTAC degraders

Previously, we discovered a novel, potent CRBN-recruiting EZH2 PROTAC degrader MS177 [20] that effectively targets both canonical and non-canonical functions of EZH2 in MLL-r AML (Fig. 1). However, this degrader also effectively induced degradation of c-MYC and CRBN neo-substrates IKZF1/3 [20]. To further dissect the effects of EZH2 degradation in MLL-r AML without affecting other targets, we sought to explore novel, potent EZH2 PROTAC degraders by hijacking the VHL E3 ligase complex. We utilized EPZ-6438 (Fig. 2), a potent and selective EZH2 inhibitor (with a half-maximal inhibitory concentration (IC₅₀) = 2.5 nM), which has been approved by the U.S. FDA for the treatment of epithelioid sarcoma and follicular lymphoma [15], as an EZH2 binder. By exploring several linkers, we previously discovered MS8815 (Fig. 1), a VHL-recruiting EZH2 PROTAC degrader, which contains seven methylene units in the linker portion. MS8815 effectively induced EZH2 degradation and inhibited the growth in TNBC cells [32]. To further optimize the linker region, we designed a novel precursor, compound 1, which contains the propionic acid group attached to the piperazine moiety for linker connection (Fig. 2). Based on this precursor, we designed and synthesized a series of novel EZH2 putative degraders 2–8 by conjugating this precursor to VHL-1, a classic ligand of the E3 ligase VHL, through a set of alkylene linkers in different lengths (Fig. 2).

Fig. 2.

Schematic design and chemical structures of precursor 1 and putative EZH2 PROTAC degraders 2–8, based on EPZ-6438.

The synthetic route for compounds 2–8 is depicted in Scheme 1. Following the previously reported procedures [32], the intermediate 9 was prepared and converted to intermediate 10 by the removal of the Boc group under acidic conditions. Amide coupling between 10 and ethyl 3-bromopropanoate (11), and subsequent hydrolysis of the ester group yielded the key precursor 1. Compounds 2–8 were subsequently generated by conjugating intermediates 12–18, which connect VHL-1 to linkers with various carbon chain length, to the carboxylic acid of precursor 1.

Scheme 1.

Synthetic route for the synthesis of compounds 2–8. Reagents and conditions: (a) TFA/DCM, rt, 30 min, yield: 98 %; (b) 11, Cs₂CO₃, DMF, 80 °C, 4 h, yield: 51 %; (c) LiOH, THF/H₂O, rt, 12 h, yield: 97 %; (d) 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI), 1-Hydroxy-7-azabenzo-triazole (HOAt), N-Methylmorpholine (NMM), DMSO, rt, 12 h, yields: 19–60 %.

2.2. Evaluation of EZH2 PROTAC degraders in MLL-r AML cells

We first evaluated the effect of putative degraders 2–8 on reducing the EZH2 protein level in the MLL-r AML cell line EOL-1 (Fig. 3). Compounds 2–4 with shorter linkers (4–6 methylene units) did not induce EZH2 degradation, while compounds 5–6 bearing longer linkers (7 and 8 methylene units, respectively) induced partial degradation. Interestingly, compound 7 (9 methylene units) reduced the EZH2 protein level completely at 0.1 μM, but less profoundly at 1 μM. This may be due to the hook effect commonly observed with some PROTAC degraders, where oversaturation of the PROTAC decreases degradation efficiency [30]. We also evaluated the effect of compounds 2–7 on reducing the H3K27me3 mark and found that they in general had little to mild effect on reducing the mark (Fig. S1). Notably, compound 8 (10 methylene units), named as MS8847, induced complete EZH2 degradation at both 0.1 and 1 μM. We, therefore, selected MS8847 as our lead degrader to move forward.

Fig. 3.

Compound 8 (MS8847) induces the most potent EZH2 degradation in a structure-activity relationship (SAR) study. The effect of designed EZH2 putative degraders 2–8 on reducing the EZH2 protein level in EOL-1 cells following 24 h treatment at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the EZH2 protein level with β-Actin as the loading control. Results shown are representative of two independent biological experiments.

2.3. MS8847 potently degrades EZH2 in a concentration- and time-dependent manner

Next, we characterized the EZH2 degradation activity of MS8847 in EOL-1 cells. MS8847 effectively induced EZH2 degradation following 24 h treatment while its parent inhibitor, EPZ-6438, did not (Fig. 4A). MS8847 also reduced the H3K27me3 mark, though not as effectively as EPZ-6348 (Fig. 4A). We determined that MS8847’s half-maximal degradation concentration (DC₅₀) for EZH2 was 34.4 ± 10.7 nM in EOL-1 cells (24 h treatment) (Fig. 4B). MS8847 also potently degraded PRC2 core components EED, and SUZ12 to a lesser extent (Fig. 4B). With 300 nM of MS8847, EZH2 degradation started at 8 h and reached maximum degradation by 24 h (Fig. 4C). Overall, MS8847 potently degrades EZH2 in a concentration- and time-dependent manner.

Fig. 4.

MS8847 induces EZH2 degradation in a concentration- and time-dependent manner. A) Western blot analysis of EZH2 and H3K27me3 protein levels in EOL-1 cells following 24 h treatment with MS8847 and EPZ-6438 at the indicated concentration. B) Western blot analysis of PRC2 core components, EZH2, EED and SUZ12, in EOL-1 cells after 24 h treatment with MS8847 at the indicated concentration. C) Western blot analysis of EZH2 and H3K27me3 protein levels after treatment with 300 nM of MS8847 at the indicated time point. The cell lysates were analyzed by Western blotting to examine the protein levels with β-Actin as the loading control. WB results shown in panels A–C are representative of two independent biological experiments. Error bars in panels B–C represent ± SD.

2.4. Mechanism of action studies and selectivity assessment for MS8847

To verify the mechanism of action (MOA) of MS8847-mediated EZH2 degradation, we first developed two structurally similar analogs, 19 (MS8847N1) and 20 (MS8847N2), as negative controls of MS8847 (Fig. 5). MS8847N1 was designed to block EZH2 binding without affecting VHL recruitment by adding methyl groups to the two amide groups in the EPZ-6438 portion of MS8847 [20,26]. MS8847N2 was designed to abolish binding to the VHL E3 ligase but maintain binding to EZH2 by incorporating a diastereomer of VHL-1 [41].

Fig. 5.

Chemical structures of the negative controls MS8847N1 and MS8847N2. Red marks indicate alterations in chemical structures. 19 (MS8847N1) was designed to abolish binding to EZH2. 20 (MS8847N2) was designed to abolish binding to VHL.

The synthesis of these two negative controls is outlined in Scheme 2. Compound 19 was afforded by the amide coupling of intermediate 22 and intermediate 18 that was prepared according to the previously reported procedures [42]. First, intermediate 21 was synthesized by introduction of two methyl groups to the amide moieties under basic conditions and followed by Boc deprotection. Following the same procedures for preparation of 1, intermediate 22 was synthesized from intermediate 21 and ethyl 3-bromopropanoate (11). In a similar manner, compound 20 was prepared by amide coupling between intermediates 1 and 23 that was prepared following the reported procedures [43].

Scheme 2.

Synthetic route for the synthesis of compounds 19 and 20. Reagents and conditions: (a) Na₂CO₃, CH₃I, DMF, rt, 3 h; (b) TFA/DCM, rt, 30 min, yield: 89 % for two steps; (c) 11, Cs₂CO₃, DMF, 80 °C, 3 h, yield: 65 %; (d) LiOH, THF/H₂O, rt, 2 h, yield: 80 %; (e) 18, EDCI, HOAt, NMM, DMSO, rt, 12 h, yield: 55 %; (f) EDCI, HOAt, NMM, DMSO, rt, 12 h, yield: 36 %.

We next compared the effects of MS8847 and its negative controls, MS8847N1 and MS8847N2, on reducing EZH2 protein levels in EOL-1 cells (Fig. 6A). While MS8847 effectively induced EZH2 degradation, MS8847N1 and MS8847N2 did not, indicating that MS8847-mediated EZH2 degradation is dependent on both EZH2 and VHL binding. To further validate MOA of MS8847, we conducted rescue experiments in EOL-1 cells that were pre-treated with a neddylation inhibitor (MLN4942, 0.05 μM), the parent inhibitor (EPZ-6438, 5 μM), or a VHL ligand (VHL-1, 5 μM) for 2 h. The cells were then treated with 0.5 μM of MS8847 for 24 h. MS8847 treatment induced EZH2 degradation that was rescued upon pre-treatment with MLN4942, EPZ-6438, or VHL-1 (Fig. 6B). These results confirm that MS8847-mediated EZH2 degradation requires EZH2 and VHL binding, and neddylation, a key component of the ubiquitin proteasome system (UPS). Overall, MS8847’s MOA is consistent with the degradation mechanism expected of PROTAC degraders.

Fig. 6.

MS8847 degrades EZH2 in a VHL- and UPS-dependent manner and is selective for EZH2 over other protein methyltransferases. A) Western blot analysis of EZH2 and H3K27me3 protein levels in EOL-1 cells following 24 h treatment with MS8847 or its negative control MS8847N1 or MS8847N2 at the indicated concentration. B) Western blot analysis of the EZH2 protein level following 2 h pre-treatment with a neddylation inhibitor (MLN4924, 0.05 μM), parent inhibitor (EPZ-6438, 5 μM), or VHL ligand (VHL-1, 5 μM), and subsequent 24 h treatment with 0.5 μM MS8847. For (A) & (B) The cell lysates were analyzed by Western blotting to examine the protein levels with β-Actin as the loading control. WB results shown in panels A–B are representative of two independent biological experiments. C) Selectivity of MS8847 at 10 μM against a panel of 20 protein methyltransferases measured using radioactive methyltransferase assays. Results shown are the mean values ± SD from two replicates.

We next evaluated selectivity of MS8847 against a panel of 20 protein methyltransferases. MS8847 did not exhibit appreciable inhibition (>50 % at 10 μM) for any of the 20 methyltransferases (Fig. 6C), suggesting that MS8847 is selective for EZH2 over other methyltransferases. Additionally, while MS8847 inhibits the enzymatic activity of both EZH2 (Fig. S2A) and EZH1 (Fig. S2B), it does not degrade EZH1 (Fig. S3), indicating that MS8847 is a selective degrader of EZH2 over the structurally similar EZH1. Together, these results demonstrate that MS8847 is a bone fide EZH2 PROTAC degrader and is selective for EZH2 over other protein methyltransferases.

2.5. MS8847 induces superior EZH2 degradation and anti-proliferative activity compared to previously reported EZH2 degraders in MLL-r AML cells

After assessing MOA and selectivity of MS8847, we compared EZH2 degradation and anti-proliferative effects of MS8847 to previously published EZH2 PROTAC degraders (i.e., MS8815, YM281, U3i and E7) in multiple MLL-r AML cell lines (EOL-1, MV4; 11, and RS4; 11). We did not include MS177 in this study because MS177 degrades not only EZH2, but also c-MYC and IKZF1/3. As shown in Fig. 7A, MS8847 is the most potent EZH2 degrader among these five EZH2 degraders in EOL-1 cells, with complete degradation at 0.1 and 0.3 μM. Furthermore, MS8847 also displayed the most potent anti-proliferative activity among the five EZH2 degraders tested in EOL-1 cells. The IC₅₀ value of MS8847 in EOL-1 cells was 0.11 μM, more than 3-fold more potent than MS8815 (IC₅₀ = 0.42 μM), the second most potent compound (Fig. 7B). Similarly, MS8847 also displayed superior EZH2 degradation and anti-proliferative effects in MV4; 11 and RS4; 11 cells compared to the other four EZH2 PROTAC degraders. In MV4; 11 cells, MS8847 induced near-complete degradation at all tested concentrations (Fig. 8A) and suppressed cell growth with an IC₅₀ of 0.17 μM, over 2-fold more potent than U3i (0.46 μM), the second most potent compound in this cell line (Fig. 8B). We also evaluated the EZH2 degradation and cell growth inhibition effects of EPZ-6438 and the two negative controls MS8847N1 and MS8847N2 in MV4; 11 cells. As expected, EPZ-6438, MS8847N1 and MS8847N2 did not induce EZH2 degradation in MV4; 11 cells (Fig. 8C). Moreover, MS8847 (IC₅₀ = 0.19 μM), was the only compound that effectively inhibited the growth in MV4; 11 cells (Fig. 8D), providing support that the cell growth inhibition effect of MS8847 in MV4; 11 cells is likely due to its EZH2 degradation activity, not its EZH2 inhibition activity. This result also provides additional support that PROTAC degraders can be advantageous over catalytic inhibitors for cancers that are dependent on oncoproteins with non-enzymatic activity. Furthermore, in RS4; 11 cells, MS8847 induced EZH2 degradation at both 0.1 and 0.3 μM (Fig. 9A) and inhibited cell growth with an IC₅₀ of 0.41 μM (Fig. 9B). MS8847 is again the most potent among the five EZH2 degraders tested. Moreover, MS8847 did not inhibit cell growth in K562 cells (Fig. S4), which are insensitive to EZH2 KO [33], suggesting that MS8847 selectively inhibits EZH2-dependent cell growth. Overall, MS8847 displayed more potent EZH2 degradation and anti-proliferative activities than previously published EZH2 PROTAC degraders in multiple MLL-r AML cell lines.

Fig. 7.

MS8847 induces superior EZH2 degradation and anti-proliferative activity in EOL-1 cells compared with previously reported EZH2 degraders. A) Western blot analysis of the EZH2 protein level in EOL-1 cells following 24 h treatment with MS8847, MS8815, YM281, U3i or E7 at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the EZH2 protein level with β-Actin as the loading control. Results shown are representative of two independent biological experiments. B) Cell viability of EOL-1 cells following 5 d treatment with MS8847, MS8815, YM281, U3i or E7 at the indicated concentration in a WST-8 assay (CCK-8). Results shown are the mean values ± SD from two independent biological experiments.

Fig. 8.

MS8847 is more potent than previously reported EZH2 degraders in MV4; 11 cells. A) Western blot analysis of the EZH2 protein level in MV4; 11 cells following 24 h treatment with MS8847, MS8815, YM281, U3i or E7 at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the EZH2 protein level with β-Actin as the loading control. Results shown are representative of two independent biological experiments. B) Cell viability of MV4; 11 cells following 5 d treatment with MS8847, MS8815, YM281, U3i or E7 at the indicated concentration in a WST-8 assay (CCK-8). Results shown are the mean values ± SD from two independent biological experiments. C) Western blot analysis of EZH2 and H3K27me3 protein levels in MV4; 11 cells following 24 h treatment with MS8847, its parent inhibitor EPZ-6438, or its negative control MS8847N1 or MS8847N2 at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the protein levels with β-Actin as the loading control. Results shown are representative of two independent biological experiments. D) Cell viability of MV4; 11 cells following 5 d treatment with MS8847, EPZ-6438, MS8847N1 or MS8847N2 at the indicated concentration in a WST-8 assay (CCK-8). Results shown are the mean values ± SD from two independent biological experiments.

Fig. 9.

MS8847 is more potent than previously reported EZH2 degraders in RS4; 11 cells. A) Western blot analysis of the EZH2 protein level in RS4; 11 cells following 24 h treatment with MS8847, MS8815, YM281, U3i or E7 at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the EZH2 protein level with β-Actin as the loading control. Results shown are representative of two independent biological experiments. B) Cell viability of RS4; 11 cells following 5 d treatment with MS8847, MS8815, YM281, U3i or E7 at the indicated concentration in a WST-8 assay (CCK-8). Results shown are the mean values ± SD from two independent biological experiments.

2.6. MS8847 potently degrades EZH2 and inhibits cell growth in TNBC cells

In addition to evaluating MS8847 in MLL-r AML cells, we assessed its effect in a solid tumor type: TNBC. We selected TNBC because it has been shown that EZH2’s non-catalytic activity is a significant contributor to TNBC progression and metastasis [21]. We utilized two TNBC cell lines, BT549 and MDA-MB-468, which we showed previously are dependent on EZH2 [33]. As shown in Fig. 10, MS8847 induced potent EZH2 degradation in BT549 and MDA-MB-468 cells following 48 h treatment. MS8847 treatment also led to anti-proliferative activity in BT549 and MDA-MB-468 cells with IC₅₀ values of 1.45 μM and 0.45 μM, respectively (Fig. 10). While MS8847 effectively suppressed cell growth, the parent EZH2 inhibitor EPZ-6438 had no effect, indicating that the growth of BT549 and MDA-MB-468 cells is dependent on EZH2, but not its methyltransferase activity. Overall, MS8847’s potent EZH2 degradation and cell growth inhibition activities in both TNBC and MLL-r AML cells suggest that EZH2 degraders could be effective therapeutics for both hematological and certain solid tumor types.

Fig. 10.

MS8847 potently degrades EZH2 and inhibits proliferation in TNBC cells. A) Western blot analysis of the EZH2 protein level in BT549 cells following 48 h treatment with MS8847 at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the EZH2 protein level with β-Actin as the loading control. Results shown are representative of three independent biological experiments. B) Cell viability of BT549 cells following 5 d treatment with MS8847 at the indicated concentration in a WST-8 assay (CCK-8). Results shown are the mean values ± SD from three independent biological experiments. C) Western blot analysis of the EZH2 protein level in MDA-MB-468 cells following 48 h treatment with MS8847 at the indicated concentration. The cell lysates were analyzed by Western blotting to examine the EZH2 protein level with β-Actin as the loading control. Results shown are representative of three independent biological experiments. D) Cell viability of MDA-MB-468 cells following 5 d treatment with MS8847 at the indicated concentration in a WST-8 assay (CCK-8). Results shown are the mean values ± SD from three independent biological experiments.

2.7. MS8847 inhibits cell growth in an in vitro TNBC 3D tumor model

MS8847 was also evaluated in a 3D cell culture assay that recapitulates an in vitro TNBC tumor model of BT549 cells. 3D cell culture methods have enhanced the scientific community’s ability to evaluate tool compounds and assess complex biological questions of cellular penetrance through 3D cellular architecture [44,45]. Utilizing magnetically levitated TNBC cells and the Bio-Assembler^™ (Nano3D Biosciences, Inc.), in vitro BT549 tumor spheroids were generated following a previously described methodology [46]. MS8847 treatment significantly reduced the size of BT549 in vitro spheroids (Fig. 11A and B) compared to DMSO-treated control tumors. Furthermore, the cell viability of BT549 cells was significantly reduced in MS8847-treated tumors (Fig. 11C). In the 3D BT549 model, higher concentrations (e.g., 5 μM) of MS8847 were needed to induce the growth inhibition phenotype compared to the 2D BT549 model. This is consistent with other reports of 3D cell culture [47]. Overall, these results highlight the efficacy of MS8847 in a 3D in vitro TNBC model and further support that MS8847 is a promising potential anti-cancer therapeutic.

Fig. 11.

MS8847 displays cell growth inhibition in an in vitro 3D spheroid TNBC model. A) Representative display of BT549 in vitro spheroids after 5 d treatment of MS8847 at the indicated concentration. Spheroids were imaged, measured using ImageJ analysis, and plotted in Prism-GraphPad. Results shown are representative of three independent biological experiments. B) The average relative area values of each spheroid taken from three independent biological experiments. Error bars represent ± SD. C) Each spheroid was analyzed for cell viability using Cell Titer Glo 3D assay and results shown are the mean values ± SD from three independent biological experiments. One-way ANOVA. ** = P ≤ 0.01, *** = P ≤ 0.001, **** = P ≤ 0.0001.

2.8. MS8847 is bioavailable in mice

Finally, pharmacokinetic (PK) properties of MS8847 were evaluated in Swiss albino mice following a single-dose IP injection of 50 mg/kg (Fig. 12). The maximum plasma concentration (C_max) of 3.9 μM was observed at 4 h. Over 1 μM of MS8847 remained in the plasma around 9 h. Moreover, it should be noted that the mice treated with MS8847 in the PK study did not exhibit obvious clinical signs of toxicity, suggesting that MS8847 was well tolerated at the tested dose. These results indicate that MS8847 has sufficient PK properties that enable in vivo efficacy studies.

Fig. 12.

MS8847 is bioavailable in mice. Plasma concentrations of MS8847 in male Swiss Albino mice over 12 h following a single IP injection of 50 mg/kg. Experiments were performed in biological triplicate per time point, with the values representing the mean concentrations ± SEM.

3. Conclusion

We discovered MS8847, a novel, highly potent EZH2 PROTAC degrader, which effectively degrades EZH2 in a concentration-, time-, and UPS-dependent manner. Importantly, MS8847 displays superior EZH2 degradation and anti-proliferative activity in multiple MLL-r AML cell lines compared to the previously published EZH2 PROTAC degraders, MS8815, YM281, U3i, and E7. MS8847 also potently degrades EZH2 and suppresses cell growth in TNBC cells. Moreover, MS8847 significantly inhibits cell growth in a 3D in vitro TNBC model. MS8847 also has sufficient mouse PK properties, making it suitable for in vivo efficacy studies. Overall, MS8847 is an enhanced EZH2 PROTAC degrader that outperforms previously reported EZH2 PROTAC degraders, making it not only a useful chemical tool for the scientific community to further investigate EZH2/PRC2 pathophysiology, but also a potential therapeutic for treating EZH2-dependent cancers.

4. Experimental section

4.1. Chemistry

4.1.1. General

All commercially available solvents and reagents were used without further purification. Reverse phase flash chromatography was performed using a HP C18 RediSep Rf reverse phase silica columns. Final compounds for biological evaluation were purified with preparative high-performance liquid chromatography (HPLC) on an Agilent Prep 1200 series with the UV detector set to 254 nm with solvent A (0.1 % of TFA in water) and solvent B (acetonitrile) as eluents with a flow rate of 40 mL/min at room temperature (rt). Purities of the final compounds were assessed >95 % by HPLC under following conditions: Agilent 1200 series system, 2.1 mm × 150 mm Zorbax 300SB-C18 5 μm column, 1–99 % gradient of 0.1 % TFA in water, and 0.1 % TFA in acetonitrile; flow rate, 0.4 mL/min; High-resolution mass spectra (HRMS) data was acquired on an Agilent G1969A API-TOF with an electrospray ionization (ESI) source. ¹H NMR and ¹³C NMR spectra were recorded on either AVANCE NEO 800 MHz, or Bruker 600 MHz 400 NMR spectrometer, and were reported in parts per million (ppm) on the δ scale in the following format: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quarter, m = multiplet), coupling constant (J, Hz), and integration. Intermediates 10 [32], 12–18 [43] and 21 [42] were prepared according to the published procedures.

4.1.2. General procedures for synthesis of intermediate 1

3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl) methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanoic acid (1).

To a solution of intermediate 10 [32] (800 mg, 1.4 mmol) in DMF (15 mL) were added ethyl 3-bromopropanoate (11, 380 mg, 2.1 mmol, 1.5 equiv) and cesium carbonate (910 mg, 2.8 mmol, 2 equiv), and the reaction was stirred at 80 °C for 4 h until the complete consumption of 22. Water was added, extracted with ethyl acetate (3 × 30 mL), washed by wanter and brine, dried over Na₂SO_4. The solvent was removed, and the resulting mixture was purified by reverse phase ISCO (10 %–100 % methanol/0.1 % TFA in H₂O) to afford ester intermediate as yellow solid (477 mg, yield 51 %). ESI (m/z) [M+H]⁺: 672.8; To a solution of this obtained intermediate (477 mg, 0.71 mmol) in THF (5 mL) and H₂O (3 mL) was added lithium hydroxide (33 mg, 1.42 mmol, 2.0 equiv). The reaction was stirred at rt overnight, and the mixture was purified by reverse phase ISCO (10 %–100 % methanol/0.1 % TFA in H₂O) to afford intermediate 1 as white solid (470.5 mg, yield 97 %).¹H NMR (600 MHz, CD₃OD) δ 7.86–7.71 (m, 3H), 7.70–7.59 (m, 1H), 7.57 (d, J = 10.1 Hz, 2H), 6.13 (s, 1H), 4.50 (s, 2H), 4.20–3.91 (m, 3H), 3.79–3.52 (m, 2H), 3.49–3.33 (m, 6H), 3.26–3.03 (m, 8H), 2.42 (s, 3H), 2.39 (s, 3H), 2.25 (s, 3H), 2.03–1.71 (m, 4H), 1.01 (t, J = 7.2 Hz, 3H). ESI (m/z) [M + H]⁺: 643.9.

4.1.3. General procedures for synthesis of compounds 2–8

(2S,4R)-1-((S)-2-(5-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)pentanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (2).

To the solution of intermediate 1 (12.8 mg, 0.02 mmol) in DMSO (1 mL) were added linker 12 (14.8 mg, 0.02 mmol, 1.0 equiv), EDCI (5.8 mg, 0.03 mmol, 1.5 equiv), HOAt (4.1 mg, 0.03 mmol, 1.5 equiv), and NMM (6.1 mg, 0.06 mmol, 3.0 equiv). After stirring overnight at rt, the resulting mixture was purified by preparative HPLC (10 %–100 % methanol/0.1 % TFA in H₂O) to afford compound 2 as white solid (13.7 mg, yield 60 %). ¹H NMR (800 MHz, CD₃OD) δ 8.95 (s, 1H), 7.85 (s, 1H), 7.75 (d, J = 8 Hz, 2H), 7.69 (s, 1H), 7.55 (d, J = 8 Hz, 2H), 7.46 (d, J = 7.8 Hz, 2H), 7.41 (d, J = 7.8 Hz, 2H), 6.13 (s, 1H), 4.63 (s, 1H), 4.58–4.52 (m, 2H), 4.51 (s, 4H), 4.37 (d, J = 15.5 Hz, 1H), 4.07 (s, 2H), 3.99 (s, 3H), 3.89 (d, J = 10.9 Hz, 1H), 3.81 (dd, J = 10.9, 4.0 Hz, 2H), 3.66–3.60 (m, 2H), 3.42–3.34 (m, 3H), 3.27–3.23 (m, 4H), 3.20–3.19 (m, 3H), 3.12 3.08 (m, 3H), 2.64–2.59 (m, 2H), 2.50–2.45 (m, 3H), 2.44–2.42 (m, 3H), 2.41–2.39 (m, 3H), 2.34–2.27 (m, 2H), 2.25 (s, 3H), 2.23–2.20 (m, 2H), 2.16 (s, 1H), 2.11–2.06 (m, 1H), 1.83–1.75 (m, 2H), 1.66–1.61 (m, 2H), 1.52 (s, 2H), 1.45–1.13 (m, 2H), 1.03–0.99 (m, 9H). HRMS (m/z) for C₆₄H₈₇N₁₀O₈S⁺ [M + H]⁺: calculated 1155.6424, found 1155.6434.

(2S,4R)-1-((S)-2-(6-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)hexanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (3).

Compound 3 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (12.8 mg, 0.02 mmol) and 13 (11.3 mg, 0.02 mmol, 1.0 equiv). Compound 3 was obtained as white solid (7.4 mg, yield 32 %). ¹H NMR (800 MHz, CD₃OD) δ 8.95 (s, 1H), 7.83 (s, 1H), 7.74 (d, J = 8 Hz, 2H), 7.68 (s, 1H), 7.54 (d, J = 8 Hz, 2H), 7.47 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz, 2H), 6.13 (s, 1H), 4.63 (s, 1H), 4.53–4.59 (m, 2H), 4.51 (s, 4H), 4.37 (d, J = 15.4 Hz, 1H), 4.02 (s, 2H), 4.01–3.95 (m, 3H), 3.90 (d, J = 10.9 Hz, 1H), 3.81 (dd, J = 10.8, 3.7 Hz, 2H), 3.71–3.61 (m, 2H), 3.41–3.35 (m, 3H), 3.27–3.24 (m, 4H), 3.21–3.17 (m, 3H), 3.11–3.04 (m, 3H), 2.64–2.61 (m, 2H), 2.48–2.47 (m, 3H), 2.43 (s, 3H), 2.40 (s, 3H), 2.33–2.27 (m, 2H), 2.25 (s, 3H), 2.23–2.21 (m, 2H), 2.16 (s, 1H), 2.11–2.07 (m, 2H), 1.82–1.75 (m, 2H), 1.63–1.60 (m, 2H), 1.53–1.50 (m, 2H), 1.41–1.27 (m, 4H), 1.03–0.99 (m, 9H). HRMS (m/z) for C₆₅H₈₉N₁₀O₈S⁺ [M + H]⁺: calculated 1169.6580, found 1169.6582.

(2S,4R)-1-((S)-2-(7-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)heptanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (4).

Compound 4 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (12.8 mg, 0.02 mmol) and 14 (11.6 mg, 0.02 mmol, 1.0 equiv). Compound 4 was obtained as white solid (9.6 mg, yield 41 %). ¹H NMR (800 MHz, CD₃OD) δ 8.94 (s, 1H), 7.81 (s, 1H), 7.73 (d, J = 8 Hz, 2H), 7.66 (s, 1H), 7.53 (d, J = 8 Hz, 2H), 7.47 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz, 2H), 6.13 (s, 1H), 4.64 (s, 1H), 4.59–4.53 (m, 2H), 4.51 (s, 4H), 4.37 (d, J = 15.4 Hz, 1H), 3.99 (s, 2H), 3.99–3.94 (m, 3H), 3.90 (d, J = 10.9 Hz, 1H), 3.81 (dd, J = 11.0, 3.9 Hz, 2H), 3.71–3.59 (m, 2H), 3.42–3.36 (m, 3H), 3.26–3.23 (m, 4H), 3.19–3.16 (m, 3H), 3.08–3.02 (m, 3H), 2.62–2.59 (m, 2H), 2.48–2.46 (m, 3H), 2.42 (s, 3H), 2.40 (s, 3H), 2.32–2.27 (m, 2H), 2.25 (s, 3H), 2.24–2.21 (m, 2H), 2.16 (s, 1H), 2.11–2.07 (m, 1H), 1.82–1.75 (m, 2H), 1.62–1.58 (m, 2H), 1.50 (s, 2H), 1.36–1.29 (m, 6H), 1.03–0.97 (m, 9H). HRMS (m/z) for C₆₆H₉₁N₁₀O₈S⁺ [M + H]⁺: calculated 1183.6737, found 1183.6731.

(2S,4R)-1-((S)-2-(8-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)octanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (5).

Compound 5 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (12.8 mg, 0.02 mmol) and 15 (11.9 mg, 0.02 mmol, 1.0 equiv). Compound 5 was obtained as white solid (4.5 mg, yield 19 %). ¹H NMR (800 MHz, CD₃OD) δ 8.95 (s, 1H), 7.84 (s, 1H), 7.74 (d, J = 8 Hz, 2H), 7.68 (s, 1H), 7.54 (d, J = 8 Hz, 2H), 7.47 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz, 2H), 6.13 (s, 1H), 4.64 (s, 1H), 4.58–4.53 (m, 2H), 4.51 (s, 4H), 4.36 (d, J = 15.5 Hz, 1H), 4.02 (s, 2H), 4.02–3.96 (m, 3H), 3.90 (d, J = 10.9 Hz, 1H), 3.81 (dd, J = 10.9, 3.9 Hz, 2H), 3.72–3.60 (m, 2H), 3.40–3.38 (m, 3H), 3.28–3.16 (m, 4H), 3.19 3.15 (m, 3H), 3.08–3.03 (m, 3H), 2.64–2.61 (m, 2H), 2.49–2.47 (m, 3H), 2.43 (s, 3H), 2.40 (s, 3H), 2.31–2.27 (m, 2H), 2.25 (s, 3H), 2.23–2.21 (m, 2H), 2.16 (s, 1H), 2.12–2.07 (m, 1H), 1.83–1.76 (m, 2H), 1.62–1.58 (m, 2H), 1.50 (s, 2H), 1.33 (s, 8H), 1.03–1.00 (m, 9H). HRMS (m/z) for C₆₇H₉₃N₁₀O₈S⁺ [M + H]⁺: calculated 1197.6893, found 1197.6899.

(2S,4R)-1-((S)-2-(9-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)nonanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (6).

Compound 6 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (12.8 mg, 0.02 mmol) and 16 (16 mg, 0.02 mmol, 1.0 equiv). Compound 6 was obtained as white solid (6.6 mg, yield 28 %). ¹H NMR (800 MHz, CD₃OD) δ 8.93 (s, 1H), 7.81 (s, 1H), 7.73 (d, J = 8 Hz, 2H), 7.65 (s, 1H), 7.53 (d, J = 8 Hz, 2H), 7.47 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz, 2H), 6.13 (s, 1H), 4.64 (s, 1H), 4.59–4.52 (m, 2H), 4.51 (s, 4H), 4.36 (d, J = 15.4 Hz, 1H), 4.00 (s, 2H), 4.00–3.94 (m, 3H), 3.90 (d, J = 10.9 Hz, 1H), 3.81 (dd, J = 10.7, 3.4 Hz, 2H), 3.68–3.60 (m, 2H), 3.38–3.40 (m, 3H), 3.26–3.23 (m, 4H), 3.17–3.18 (m, 3H), 3.08–3.03 (m, 3H), 2.62–2.60 (m, 2H), 2.48–2.47 (m, 3H), 2.42 (s, 3H), 2.40 (s, 3H), 2.32–2.27 (m, 2H), 2.25 (s, 3H), 2.23–2.20 (m, 2H), 2.16 (s, 1H), 2.11–2.05 (m, 1H), 1.81–1.75 (m, 2H), 1.63–1.59 (m, 2H), 1.49 (s, 2H), 1.32 (s, 10H), 1.03–0.99 (m, 9H). HRMS (m/z) for C₆₈H₉₅N₁₀O₈S⁺[M + H]⁺: calculated 1211.7050, found 1211.7053.

(2S,4R)-1-((S)-2-(10-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)decanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (7).

Compound 7 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (12.8 mg, 0.02 mmol) and 17 (12.4 mg, 0.02 mmol, 1.0 equiv). Compound 7 was obtained as white solid (8.1 mg, yield 33 %). ¹H NMR (800 MHz, CD₃OD) δ 8.96 (s, 1H), 7.85 (s, 1H), 7.75 (d, J = 8 Hz, 2H), 7.69 (s, 1H), 7.55 (d, J = 8 Hz, 2H), 7.47 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz, 2H), 6.14 (s, 1H), 4.64 (s, 1H), 4.59–4.52 (m, 2H), 4.51–4.50 (m, 4H), 4.36 (d, J = 15.4 Hz, 1H), 4.02 (s, 2H), 4.02–3.95 (m, 3H), 3.90 (d, J = 10.9 Hz, 1H), 3.81 (dd, J = 10.8, 3.6 Hz, 2H), 3.73–3.64 (m, 2H), 3.39–3.35 (m, 3H), 3.29–3.26 (m, 4H), 3.18–3.15 (m, 3H), 3.11–3.03 (m, 3H), 2.63 (t, J = 6.6 Hz, 2H), 2.48 (s, 3H), 2.43 (s, 3H), 2.40 (s, 3H), 2.34–2.26 (m, 2H), 2.25 (s, 3H), 2.24–2.22 (m, 2H), 2.18–2.13 (m, 1H), 2.10–2.07 (m, 1H), 1.83–1.77 (m, 2H), 1.63–1.58 (m, 2H), 1.50–1.49 (m, 2H), 1.36–1.31 (m, 12H), 1.02–1.00 (m, 9H). HRMS (m/z) for C₆₉H₉₇N₁₀O₈S⁺ [M + H]⁺: calculated 1225.7206, found 1225.7213.

(2S,4R)-1-((S)-2-(11-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)undecanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (8).

Compound 8 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (12.8 mg, 0.02 mmol) and 18 (16.5 mg, 0.02 mmol, 1.0 equiv). Compound 8 was obtained as white solid (14 mg, yield 57 %). ¹H NMR (800 MHz, CD₃OD) δ 8.95 (s, 1H), 7.83 (s, 1H), 7.74 (d, J = 8 Hz, 2H), 7.67 (s, 1H), 7.54 (d, J = 8 Hz, 2H), 7.47 (d, J = 7.8 Hz, 2H), 7.42 (d, J = 7.8 Hz, 2H), 6.13 (s, 1H), 4.64 (s, 1H), 4.59–4.53 (m, 2H), 4.52 (s, 4H), 4.36 (d, J = 15.4 Hz, 1H), 4.03 (s, 2H), 4.00–3.99 (m, 3H), 3.90 (d, J = 10.9 Hz, 1H), 3.82–3.80 (m, 2H), 3.68–3.62 (m, 2H), 3.38 (t, J = 11.2 Hz, 3H), 3.29–3.26 (m, 4H), 3.18 3.15 (m, 3H), 3.09 (s, 3H), 2.63 (t, J = 6.7 Hz, 2H), 2.48 (s, 3H), 2.43 (s, 3H), 2.40 (s, 3H), 2.30–2.28 (m, 2H), 2.25 (s, 3H), 2.24–2.21 (m, 2H), 2.16 (s, 1H), 2.10–2.07 (m, 1H), 1.79 (s, 2H), 1.65–1.58 (m, 2H), 1.52–1.47 (m, 2H), 1.31–1.30 (m, 14H), 1.02–1.01 (m, 9H). ¹³C NMR (101 MHz, CD₃OD) δ 174.63, 173.06, 170.96, 170.31, 152.46, 151.87, 146.86, 143.76, 139.15, 132.50, 131.10, 129.71, 128.97, 127.65, 127.36, 120.94, 109.80, 69.70, 67.68, 65.93, 59.96, 59.46, 57.57, 56.62, 54.76, 52.76, 49.96, 49.08, 42.91, 42.29, 39.91, 39.17, 37.55, 36.11, 35.43, 35.29, 35.20, 33.56, 31.67, 29.34, 29.18, 29.10, 29.00, 28.94, 28.89, 26.58, 25.69, 25.66, 25.61, 24.70, 22.33, 18.34, 17.24, 14.19, 13.03. t_R = 4.40 min, HRMS (m/z) for C₇₀H₉₉N₁₀O₈S⁺ [M + H]⁺: calculated 1239.7363, found 1239.7371.

4.1.4. General procedures for synthesis of compounds 19 and 20

(2S,4R)-1-((S)-2-(11-(3-(4-((3’-(ethyl(tetrahydro-2H-pyran-4-yl) amino)-4′-methyl-5’-(methyl((1,4,6-trimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-[1,1′-biphenyl]-4-yl)methyl) piperazin-1-yl)propanamido)undecanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (19).

Compound 19 was synthesized following the standard procedures for preparing compound 2 from intermediate 22 (64.6 mg, 0.096 mmol) and 18 (58.8 mg, 0.096 mmol, 1.0 equiv). Compound 19 was obtained as white solid (67.4 mg, yield 55 %). ¹H NMR (400 MHz, CD₃OD) δ 9.05 (s, 1H), 7.90 (d, J = 9.6 Hz, 1H), 7.83–7.71 (m, 2H), 7.62–7.51 (m, 3H), 7.46–7.28 (m, 4H), 6.17 (s, 1H), 4.82–4.63 (m, 2H), 4.56 (s, 1H), 4.53–4.15 (m, 7H), 4.02–3.89 (m, 4H), 3.88–3.69 (m, 4H), 3.49 (d, J = 1.8 Hz, 3H), 3.47–3.38 (m, 6H), 3.34 (s, 3H), 3.24 (s, 3H), 3.10 (t, J = 7.1 Hz, 2H), 2.91 (d, J = 1.8 Hz, 1H), 2.73 (d, J = 1.8 Hz, 3H), 2.62 (t, J = 6.9 Hz, 2H), 2.43 (d, J = 2.0 Hz, 4H), 2.36–2.29 (m, 6H), 2.28–1.97 (m, 6H), 1.94–1.69 (m, 1H), 1.60–1.48 (m, 3H), 1.46–1.34 (m, 2H), 1.24 (s, 14H), 0.96 (s, 9H). ¹³C NMR (101 MHz, CD₃OD) δ 174.66, 173.10, 170.96, 170.10, 169.86, 164.00, 163.15, 159.90, 159.53, 152.27, 151.02, 150.05, 146.06, 139.42, 131.29, 131.26, 129.34, 129.13, 129.00, 128.97, 127.69, 127.66, 127.51, 127.46, 119.09, 118.44, 110.58, 109.88, 65.86, 59.73, 59.42, 57.58, 56.62, 52.71, 49.57, 48.63, 42.47, 42.28, 39.18, 37.55, 35.28, 35.18, 35.11, 30.79, 30.42, 29.31, 29.18, 29.12, 29.01, 28.95, 28.92, 28.86, 26.61, 25.67, 25.64, 25.61, 19.25, 18.33, 13.79, 13.37, 9.52. t_R = 5.27 min, HRMS (m/z) for C₇₂H₁₀₃N₁₀O₈S⁺ [M + H]⁺: calculated 1267.7676, found 1267.7696.

(2R,4S)-1-((S)-2-(11-(3-(4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanamido)undecanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-(4-(4-methylthiazol-5-yl)benzyl)pyrrolidine-2-carboxamide (20).

Compound 20 was synthesized following the standard procedures for preparing compound 2 from intermediate 1 (40.3 mg, 0.06 mmol) and 23 (38.4 mg, 0.06 mmol, 1.0 equiv). Compound 20 was obtained as white solid (26.5 mg, yield 36 %).¹H NMR (800 MHz, CD₃OD) δ 9.01 (s, 1H), 7.92 (s, 1H), 7.78 (d, J = 8.0 Hz, 2H), 7.75 (s, 1H), 7.58 (d, J = 8.0 Hz, 2H), 7.45 (d, J = 8.1 Hz, 2H), 7.40 (d, J = 8.1 Hz, 2H), 6.15 (s, 1H), 4.61–4.57 (m, 1H), 4.56–4.45 (m, 5H), 4.44 (s, 1H), 4.36 (d, J = 15.7 Hz, 1H), 4.16 (s, 2H), 4.01–3.99 (m, 4H), 3.78 (s, 2H), 3.74–3.72 (m, 1H), 3.44–3.34 (m, 8H), 3.25 (s, 4H), 3.16 (t, J = 7.1 Hz, 2H), 2.67 (t, J = 6.8 Hz, 2H), 2.50 (s, 3H), 2.45 (s, 3H), 2.40 (s, 3H), 2.28–2.24 (m, 4H), 2.23–2.10 (m, 3H), 2.06–2.01 (m, 1H), 1.83 (s, 2H), 1.64–1.34 (m, 5H), 1.33–1.13 (m, 14H), 1.08–0.99 (m, 9H). ¹³C NMR (151 MHz, CD₃OD) δ 175.55, 173.10, 171.17, 170.31, 169.64, 164.03, 152.60, 151.87, 146.84, 143.83, 140.38, 139.12, 132.56, 130.96, 129.64, 129.00, 128.31, 127.34, 127.31, 120.90, 109.92, 69.12, 65.88, 59.93, 59.57, 58.65, 55.43, 52.74, 50.03, 49.05, 42.09, 39.17, 37.71, 35.43, 34.91, 33.97, 29.47, 29.19, 29.12, 28.99, 28.96, 28.88, 26.60, 25.65, 25.61, 25.40, 18.29, 17.18, 14.23, 13.64, 9.53. t_R = 5.80 min, HRMS (m/z) for C₇₀H₉₉N₁₀O₈S⁺[M + H]⁺: calculated 1239.7363, found 1239.7376.

4.1.5. General procedures for synthesis of intermediate 22

3-(4-((3’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-5’-(methyl((1,4,6-trimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl) carbamoyl)-[1,1′-biphenyl]-4-yl)methyl)piperazin-1-yl)propanoic acid (22).

To the solution of tert-butyl 4-((3’-(((4,6-dimethyl-2-oxo-1,2-dihydropyridin-3-yl)methyl)carbamoyl)-5’-(ethyl(tetrahydro-2H-pyran-4-yl)amino)-4′-methyl-[1,1′-biphenyl]-4-yl)methyl)piperazine-1-carboxylate (9, 335.2 mg, 0.5 mmol) in DMF (5 mL) was added sodium carbonate (212 mg, 2.0 mmol, 4.0 equiv) at rt. Iodomethane (141.7 mg, 1.0 mmol, 2.0 equiv) was added slowly. After the reaction was stirred at rt for 3 h, water was added, extracted with ethyl acetate (3 × 20 mL), dried over Na₂SO_4. The solvent was removed, and the resulting mixture was dissolved in TFA (3 mL) and DCM (3 mL). The reaction was stirred at rt for 30 min. Solvent was evaporated and the mixture was purified by reverse phase ISCO (10 %–100 % methanol/0.1 % TFA in H₂O) to afford intermediate 21 as white solid (266.5 mg, 89 % yield for two steps). ESI (m/z) [M+H]⁺: 600.6; To a solution of intermediate 21 (235 mg, 0.4 mmol) in DMF (5 mL) were added ethyl 3-bromopropanoate (11, 108 mg, 0.6 mmol, 1.5 equiv) and cesium carbonate (260 mg, 0.8 mmol, 2 equiv), and the reaction was stirred at 80 °C for 4 h until the complete consumption of 21. Water was added, extracted with ethyl acetate (3 × 30 mL), washed by water and brine, dried over Na₂SO_4. The solvent was removed, and the resulting mixture was purified by reverse phase ISCO (10 %–100 % methanol/0.1 % TFA in H₂O) to afford ester intermediate as white solid (182 mg, yield 65 %). ESI (m/z) [M+H]⁺: 700.8; To a solution of this intermediate (83 mg, 0.12 mmol) in THF (3 mL) and H₂O (3 mL) was added lithium hydroxide (6 mg, 0.24 mmol, 2.0 equiv). The reaction was stirred at rt for 2 h, and the mixture was purified by preparative HPLC (10 %–100 % methanol/0.1 % TFA in H₂O) to afford intermediate 22 as white solid (64.6 mg, yield 80 %).¹H NMR (600 MHz, CD₃OD) δ 7.75–7.51 (m, 3H), 7.51–7.35 (m, 3H), 6.14 (s, 1H), 4.37–4.19 (m, 1H), 3.96–3.84 (m, 6H), 3.63–3.42 (m, 4H), 3.38 (s, 1H), 3.34–3.24 (m, 2H), 3.18–3.02 (m, 6H), 3.02–2.86 (m, 6H), 2.69 (s, 3H), 2.67–2.58 (m, 3H), 2.31 (s, 4H), 2.28 (s, 3H), 2.23 (d, J = 9.1 Hz, 4H). ESI (m/z) [M + H]⁺: 672.6.

4.2. Cell culture

EOL-1 cells were received from our collaborator Gang Greg Wang’s lab at UNC Chapel Hill. MV4; 11, RS4; 11, BT549, MDA-MB-468, and K562 cells were purchased from ATCC. EOL-1, MV4; 11, and RS4; 11, and BT549 cells were grown in RPMI-1640 medium containing 10 % FBS and 1 % penicillin-streptomycin. K562 cells were cultured in IMDM containing 10 % FBS and 1 % penicillin-streptomycin. MDA-MB-468 cells were cultured in DMEM containing 10 % FBS and 1 % penicillin-streptomycin. All cells were cultured at 37 °C with 5 % CO₂.

4.3. Western blot assay

EOL-1, MV4; 11, RS4; 11, BT549, and MDA-MB-468 cells were seeded in 6 well plates at 0.5–2 million cells per well. Following compound treatment, cells were lysed with RIPA buffer containing 1x protease-proteasome inhibitor cocktail. Following 30 min of incubation on ice with periodical mixing, samples were centrifuged at 4 °C for 10 min at 15000 RPM, and the pellets discarded. Protein concentration was quantified using the Pierce Gold Standard BCA kit (Thermofisher). 1x Laemmli buffer was added to cell lysates and samples were heated at 100 °C for 10 min. Samples (~10 μg) were run on 4–15 % or 4–20 % precast SDS-PAGE gels (BioRad) then transferred to PVDF membranes using the Transblot rapid transfer machine (BioRad). Membranes were blocked for 1 h shaking at room temperature in LI-COR TBS or PBS blocking buffer, then incubated with primary antibody overnight at 4 °C. Membranes were washed with TBST or PBST (0.1 % Tween 20) before and after incubation with secondary antibody (LI-COR) for 1 h at room temperature. Blots were imaged on the Odyssey CLx Imagining system (LI-COR) and analyzed using the LI-COR ImageStudio software. The following antibodies were purchased from Cell Signaling Technology: EZH2 (D2C9), SUZ12 (D39F6), EED (E4L6E), EZH1 (D7D5D), H3K27me3 (C36B11), β-Actin (13E5).

4.4. Cell viability assay

EOL-1, MV4; 11, RS4; 11, K562, BT549, and MDA-MB-468 cells were seeded in 96 well plates at 3,000–10,000 cells per well. Cells were treated with compound at indicated serial dilutions and incubated at 37 °C for 5 days. Cell viability was determined using CCK-8 (Cell Counting Kit-8, Dojindo, CK04), with 1x solution added to each well and re-incubated at 37 °C for 2–4 h. Absorbance was measured at 450 nm and reference was measured at 690 nm using the Infinite F PLEX plate reader (TECAN). IC₅₀ values were determined using GraphPad Prism 6.

4.5. Griener Bio-Assembler 3D in vitro assay

In vitro BT549 spheroids were designed using the magnetic levitation Bio-Assembler^™ (Griener Bio One #655846) kit. Method described in detail by Haisler et al. [46]. Briefly, wells were incubated with Nanoshuttle^™ (proprietary nanoparticle assembly consisting of gold, iron oxide, and poly-l-lysine) for 24 h at a ratio of 1:2000 (1 μL of nanoshuttle^™:2000 cells). The nanoshuttle-coated cells were then trypsinized, washed and plated into low-adhesion 6 or 24 well Bio-Assembler plates at the following densities: 6 well plate 5.0–7.5 × 10⁵ CFU and 24 well plate 3.0–5.0 × 10⁵ CFU. Bio-Assembler magnetic cores were then applied externally to the plates, and cells were then placed in an incubator for 7 days at 37 °C in a humidified environment with medium changed every 3–4 days, until tumor formed into singular solid sphere (approximately 30 mm²). MS8847 and/or DMSO was applied to the cells for 5 d treatment. Spheroids were washed twice with sterile PBS, imaged on Epsilon Perfection V3 scanner and image analysis was conducted in ImageJ to measure the area of the spheroid. Cell viability of spheroids were then assessed utilizing Promega Cell Titer Glo 3D Cell Viability assay (Promega G9681). Area and luminescence values were analyzed using GraphPad Prism 6.

4.6. EZH2/EZH1 methyltransferase inhibition and selectivity assays

EZH2 and EZH1 inhibition assays were performed in duplicate at indicated concentrations by Reaction Biology Corp. using the radioisotope-based MT HotSpot^™ system. The selectivity assay was conducted by Reaction Biology Corps., with 10 μM of compound tested against a panel of 20 methyltransferases using the same platform, performed in duplicate.

4.7. Mouse pharmacokinetic study

The six male Swiss Albino mice were intraperitoneally administered a single dose of MS8847 in HCl form at 50 mg/kg in a solution formulation of 5 % NMP, 5 % solutol HS-15, and 90 % normal saline. After administration, blood samples were collected at 0.5, 1, 2, 4, 8 and 12 h, and centrifuged to obtain the plasma. The resulting plasma samples were stored at −70 ± 10 °C and the plasma concentration-time data of MS8847 was used for the pharmacokinetic analysis. Pharmacokinetic analysis was performed using NCA module of Phoenix WinNonlin (Version 8.0). Plasma concentrations were quantified by fit-for-purpose LC-MS/MS method (LLOQ: 2.02 ng/mL). Compound concentrations in plasma at each time point are the average values from 3 test mice.

Data availability

Data will be made available on request.