Macrocyclic Carbon-Linked Pyrazoles As Novel Inhibitors of MCL-1

Abstract

Myeloid cell leukemia-1 (MCL-1) is a member of the antiapoptotic BCL-2 proteins family and a key regulator of mitochondrial homeostasis. Overexpression of MCL-1 is found in many cancer cells and contributes to tumor progression, which makes it an attractive therapeutic target. Pursuing our previous study of macrocyclic indoles for the inhibition of MCL-1, we report herein the impact of both pyrazole and indole isomerism on the potency and overall properties of this family of compounds. We demonstrated that the incorporation of a fluorine atom on the naphthalene moiety was a necessary step to improve cellular potency and that, combined with the introduction of various side chains on the pyrazole, it enhanced solubility significantly. This exploration culminated in the discovery of compounds (R_a)-10 and (R_a)-15, possessing remarkable cellular potency and properties

The BCL-2 family of proteins tightly control cell death by regulating the intrinsic apoptosis pathway.^1,2 Cancer cells highly express members of the BCL-2 family, and their expression has been shown to be essential, making them promising targets for cancer treatment. Besides BCL-2, BCL-XL, BCL-W, and BFL-1, MCL-1 is among the five key antiapoptotic proteins in the BCL-2 protein family responsible for controlling cell death.³ MCL-1 directly inhibits the activity of pro-apoptotic BCL-2 family proteins, namely, BAK and BAX, while indirectly preventing apoptosis by capturing apoptotic sensitizer proteins, such as BIM and NOXA, which contain BH3 domains.⁴⁻⁸ Several types of cancers, among which are chronic myeloid leukemia (CML),⁹ acute myeloid leukemia (AML),¹⁰ B-cell acute lymphoblastic leukemia¹¹ and multiple myeloma (MM),¹² are characterized by MCL-1 gene amplification. Overexpression of the human MCL-1 gene elevates MCL-1 protein levels, which contribute to tumorigenesis and resistance, not only to conventional chemotherapies but also to targeted therapies, including the BCL-2 selective inhibitor Venetoclax (ABT-199).^13,14 Researchers from both the pharmaceutical industry and academia have been pursuing MCL-1 inhibitors and outstanding progress has been made.¹⁵⁻¹⁷ A few MCL-1 selective inhibitors have advanced to clinical trials such as AZD5991 (1),¹⁸ AMG176 (2,)¹⁹ and S64315 (3),^20,21 which show picomolar biochemical potency and on-target cell activity (Figure 1).

Figure 1.

Examples of MCL-1 inhibitors tested in clinical trials.

We recently disclosed JNJ-4355 (4, Figure 2) as a highly potent MCL-1 inhibitor with properties optimized to maximize its therapeutic index.²²⁻²⁴ In the present work, we wanted to explore the impact of both pyrazole and indole isomerism on the overall properties and potency of the molecule.

Figure 2.

Novel macrocyclic C4-linked isomeric pyrazoles.

We hypothesized that analogues of 4, with isomeric pyrazoles on the 3,7-indolyl core, would offer the opportunity to modulate both potency and properties, such as solubility. Indeed, the unfavorable physicochemical properties of these rigid high MW protein–protein inhibitors have been challenging to marry with high solubility, a requirement for IV dosing which most players have selected as the administration route in the clinic to maximize therapeutic index. The novel chemotype differs from the previously reported analogues by two structural aspects: on the one hand, a 3,7-indolyl core which is attached to a pyrazole ring through its C7; and on the other hand, an isomeric pyrazole is attached to C7 of the indolyl core through its C5. Our objectives were to explore two vectors on this novel chemotype to modulate potency and solubility: first, by introduction of various side chains at C3 (R¹) on the pyrazole, then by incorporating diverse side chains on the 3,7-indolyl core nitrogen (R²). The comparison of the two sets of compounds would then allow us to evaluate the best vector for fine-tuning potency and properties. Herein, we describe the synthesis and biological results of those two sets of analogues.

We first installed various side chains (R¹) on C3 of the pyrazole. Compound (rac)-5 (R¹ = H, Table 1) was the initial starting point to benchmark the novel C4-linked isomeric pyrazole. Separation by means of SFC gave both atropisomers (R_a)-5 and (S_a)-5. The absolute configuration was not determined, but (R_a)-5 atropisomer was assumed to be the active compound based on docking and analogy with other cocrystallized structures such as compound (R_a)-10 (see later). The stereo configuration assignment of atropisomers was based on the rules described by Eliel et al.²⁵ Compound (R_a)-5 demonstrated very promising potency in our biochemical assay (MCL-1 K_i = 0.069 nM) and cellular assay (MOLP-8 AC₅₀ = 51 nM) while its atropisomer (S_a)-5 showed poor potencies in the same assays (MCL-1 K_i = 11 nM and MOLP-8 AC₅₀ > 10 000 nM, respectively; see Table 1). Only the active (R_a) atropisomers will be further discussed in this article. We further explored the effect of a methyl substituent at C3 on the isomeric pyrazole ((R_a)-6), which resulted in similarly attractive potency but with a higher cell shift (MOLP-8/MCL-1), suggesting a lower cellular permeability. However, the solubility was poor for both compounds. We then replaced the thioether linker by its methylene isostere to (R_a)-7 in the hope to improve permeability by reducing the TPSA. Since the solubility of the corresponding carboxylic acid was often poor in DMSO we opted to isolate some compounds as sodium salt such as analogues (R_a)-7 and (R_a)-9, which were well soluble in DMSO. Compound (R_a)-7 had somewhat better activity in the biochemical and cellular assays but also increased in solubility at physiological pH (Table 1). To optimize solubility further, we explored structural modifications which had been successful during the optimization of compound 4 (Figure 2): introduction of a fluorine atom on the naphthalene in combination with a solubilizing side chain on the pyrazole ring.²³ Keeping this in mind, we first synthesized analogue (R_a)-8 with a solubilizing side chain at the pyrazole 3-position. Analogue (R_a)-8 indeed demonstrated improved solubility at pH 7 (590 μM) while maintaining good potency. In addition, only low acute cardiac hazard was detected in our calcium transient cardiomyocyte (CTCM) assay²⁶ at 10 μM (data not shown). We next explored the effect of 6-fluorine on the naphthalene ring, which previously proved to boost the potency in 4. To confirm the effect of the 6-fluoronaphthyl, compound (R_a)-9, a matched-pair analogue of (R_a)-7, was synthesized. Compound (R_a)-9 was isolated as a sodium salt but expected to behave like its analogues under buffered conditions. To our delight, we observed a 4-fold increase in cellular potency and an improved solubility compared to (R_a)-7 at physiological pH (Table 1).

Table 1. Effects of Side Chains R¹ and Fluoronaphthyl Group on Potency and Solubility.

Compd	MCL-1 Ki (nM)a	MOLP-8 AC50 (nM)b	Cell shift MOLP-8/MCL-1c	Solubility (μM) pH 2/pH 7d
(Ra)-5	0.069	51	738	<0.55/<0.49
(Sa)-5	11	>10 000	>900	NA
(Ra)-6	0.040	54	1350	<0.27/9.7
(Ra)-7	0.032	45	1406	<0.04/210
(Ra)-8	0.023	43	1870	<0.06/590
(Ra)-9	0.041	10	243	<0.03/550
(Ra)-10	0.018	14	777	<0.02/580
(Ra)-11	0.013	16	1230	<0.29/460

^aK_i represents the MCL-1 binding affinity of a compound that disrupts the MCL-1:Bim construct interaction, as determined in Tb-MCL-1 HTRF assay. The HTRF K_i values are a mean of at least 4 determinations.

^bMOLP-8 AC₅₀ represents the cellular activity of compounds inducing apoptosis in MCL-1 dependent MOLP-8 cells in the CaspaseGlo assay.²⁷ MOLP-8 AC₅₀ values are a mean of at least 2 determinations.

^cCell shift represents the ratio between cellular and biochemical potency.

^dThe compound solubility at pH 2.0 and pH 7.0 in phosphate buffer. NA = not available.

Combining a solubilizing side chain R¹ and the 6-fluoronaphthalene led to compound (R_a)-10. This resulted in attractive potencies in both biochemical (MCL-1 K_i = 0.018 nM) and cellular assay (MOLP-8 AC₅₀ = 14 nM) with enhanced solubility at pH = 7. Also, low acute cardiac hazard in CTCM was found at 10 μM and acceptable moderate clearance was observed (Table 3). Extending the solubilizing side chain by one carbon yielded compound (R_a)-11, which gave similar biological results but higher CYP (2C8) inhibition (data not shown) and lower solubility. In short, introduction of solubilizing side chains on the pyrazole (R¹) and addition of a 6-fluorine atom on the naphthalene enhanced both the potency and solubility. Combining both modifications led to (R_a)-10 that emerged as a promising compound. The absolute configuration of (R_a)-10 was obtained from a cocrystal structure with MBP-MCL-1 (Figure 3). The crystal structure of (R_a)-10 was determined at 1.47 Å resolution and confirmed the (R_a) absolute configuration, with the carboxylic acid moiety forming two hydrogen bonds and a salt bridge with Arg263. The pyrazole nitrogen establishes a water-mediated interaction with His224 and two π–π stacking interactions were observed between Phe270 and both the naphthyl part and the indole scaffold. Compound (R_a)-10 was the only analogue where the (R_a) configuration has been proven. The (R_a) configurations of all other compounds described in this article are based on the similarity of biological results and docking analogy with other cocrystallized structures.

Table 3. Comparison of Biological Activities and Properties of (R_a)-10 and (R_a)-15.

	(Ra)-10	(Ra)-15
MCL-1 Ki (nM)a	0.018	0.020
MOLP-8 AC50 (nM)b	14	76
Cell shift MOLP-8/MCL-1c	777	3800
KMS-12-PE AC50 (nM)d	111	407
KMS-12-PE-MCL-1 KO (μM)e	>30	>30
CHI logD (pH 7.4)f	2.49	2.53
Microsomal CLint h/m (μL/(min mg))	48/157	43/24
CYP IC50 (μM) 1A2/2C19/2C8/2C9/2D6/3A4	>20/>20/9.6/19/>20 > 20	>20/>20/5.6/>20/10.5/9.5
Solubility (μM) pH 2/pH 7g	<0.020/580	580/550
CTCM risk scoreh0.1/0.3/1/3/10 (μM)	No/No/No/No/Low	No/No/No/No/Low
hERG IC50 (μM)i	>30	>30

^aK_i represents the MCL-1 binding affinity of a compound which disrupts the MCL-1:Bim interaction, as determined in Tb-MCL-1 HTRF assay. The HTRF K_i values are a mean of at least 4 determinations.

^bMOLP-8 AC₅₀ represents the cellular activity of compounds inducing apoptosis in MCL-1 dependent MOLP-8 cells in the CaspaseGlo assay.²⁷ MOLP-8 AC₅₀ values are a mean of at least 2 determinations.

^cCell shift represents the ratio between cellular and biochemical potency.

^dHuman multiple myeloma cell line KMS-12-PE.

^eKMS-12-PE MCL-1 gene knockout BFL-1 overexpressing cell line.

^fChromatographic hydrophobicity index.

^gThe compound solubility at pH 2.0 and pH 7.0 in phosphate buffer.

^hCalcium transient cardiomyocyte assay.²⁶

ⁱhERG gene product inhibition measured in the patchXpress assay.

Figure 3.

Co-crystal structure of (R_a)-10 (green) with MCL-1 structure (purple) (PDB 8SVY) at 1.47 Å. Main binding interactions for (R_a)-10 with MCL-1: carboxylic acid forming hydrogen bonds and a salt bridge with Arg263, pyrazole nitrogen establishes a water-mediated hydrogen bond with His224, and two π–π stacking interactions between Phe270 and both naphthyl part and indole scaffold. *The methoxyethoxymethyl side chain adopts two conformations in the crystal lattice.

In the second part of this work, we used indole core nitrogen as a handle to install diverse side chains (R², Figure 2). This could be conveniently done in a rapid late-stage functionalization of the indole nitrogen. Docking studies using our previously cocrystallized structures (data not shown) highlighted the possibility to reach the external solvent region and supported the synthesis of compounds (R_a)-12 to 16 (Table 2). We first profiled (R_a)-12 as a reference compound. Modest cellular potency was observed relative to its potent binding affinity. Incorporation of a dimethylamino ethyl chain in compound (R_a)-13 slightly improved cellular potency, but still a similar high cell shift probably due to protonation and solvation of the tertiary amine side chain was observed. Modification to a methyl ethyl ether derivative (R_a)-14 gave a remarkable improvement in cellular potency, but still low solubility was observed. Insertion of a 2-morpholinoethyl side chain in compound (R_a)-15 allowed to combine high cellular potency with dramatically improved solubility especially at acidic pH. In addition, the intrinsic clearance in human and rat liver microsomes values were relatively low for (R_a)-15 (see Table 3). Further elongation of the ether chain yielded compound (R_a)-16 with good solubility at pH 7, but low solubility at acidic pH, while good cellular potency was retained. However, a higher intrinsic clearance in human liver microsomes was observed compared with previous compounds in this series (data not shown). From our exploration of 3,7-substituted indoles carrying a C4-linked isomeric pyrazole, compounds (R_a)-10 and (R_a)-15 emerged as having the most promising properties warranting further profiling.

Table 2. R² Side-Chain Variations.

^aK_i represents the MCL-1 binding affinity of a compound which disrupts the MCL-1:Bim construct interaction, as determined in Tb-MCL-1 HTRF assay. The HTRF K_i values are a mean of at least 4 determinations.

^bMOLP-8 AC₅₀ represents the cellular activity of compounds inducing apoptosis in MCL-1 dependent MOLP-8 cells in the CaspaseGlo assay.²⁷ MOLP-8 AC₅₀ values are a mean of at least 2 determinations.

^cCell shift represents the ratio between cellular and biochemical potency.

^dThe compound solubility at pH 2.0 and pH 7.0 in phosphate buffer. NA = not available.

An overview of the respective biological activities and properties of both compounds is shown in Table 3. Compound (R_a)-10 exhibited a similar binding affinity to compound (R_a)-15, suggesting that introduction of solubilizing side chains R¹ or R² respectively, on the pyrazole and the 3,7-indolyl core moiety did not affect the binding affinity. However, improved cellular potency was observed for (R_a)-10, in both the MOLP-8 and KMS-12-PE cell lines, which resulted in a lower cell shift for compound (R_a)-10 relative to compound (R_a)-15. In addition, (R_a)-10 and (R_a)-15 were inactive in an in-house generated KMS-12-PE-MCL-1 knockout (KO) and BFL-1 overexpressing cell line. This confirmed cell killing was mediated via inhibition of the MCL-1 protein. Finally, the CHI logD (pH 7.4) values for compounds (R_a)-15 and (R_a)-10 were in an optimal range of around 2.5.

Like with clinical compounds 1, 2, and 3, we aimed for IV dosing to control exposure well and maximize therapeutic index. This required us to develop potent moderate to high clearance compounds that would provide sufficient exposure to exert cell killing but a short plasma half-life to avoid organ toxicity. The moderate to high intrinsic clearance observed for (R_a)-10 and (R_a)-15 in human liver microsomes (hLM CLint) and in mouse liver microsomes (mLM CLint) fell within our projected range and warranted further profiling. The compounds exhibited a weak inhibition of some CYP450 enzymes (Table 3). The strongest inhibition was observed for compound (R_a)-15 on the 2C8 isoform, with an IC₅₀ of 5.6 μM, which is well above its cell killing activity (MOLP-8 AC₅₀ = 76 nM), and hence, it was judged to pose a minimal DDI risk. Thermodynamic solubility measurements in buffered aqueous media confirmed good solubility for both compounds at pH 7. However, poor solubility was observed for compound (R_a)-10 at pH 2 (Table 3). Finally, no significant drug-induced inhibition of the hERG channel was observed for both compounds in an automated patch-clamp assay. Acute cardiac risk hazard was assessed using a calcium transient assay in human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs).²⁶ A low pro-arrhythmic risk score was identified for both compounds only at the highest concentration (10 μM).

The synthetic routes to compounds (R_a)-10 and (R_a)-15 are outlined below. The detailed synthesis protocols and analytical data for all compounds can be found in the Supporting Information. Lead compound (R_a)-10 (Scheme 1) could be accessed through a multistep synthesis from commercially available indole 17. Selective ester reduction of 17 with borane at 50 °C, followed by alcohol protection, gave 18 in good yields. Next, Suzuki cross-coupling with boronate ester 19, obtained in five steps from commercial reagents (Supporting Information), delivered 20 in quantitative yield. Alkylation of crude 20 with methyl iodide followed by PMB deprotection with DDQ delivered 21 in good yields. Alcohol 21 was converted to the reactive mesylate intermediate with mesyl chloride in DCM at 0 °C and this intermediate was then converted in situ to thioacetate 22 by addition of potassium thioacetate in DMF in good yield. Next, coupling of 22 with intermediate 23 (obtained in seven steps; see the Supporting Information) via nucleophilic substitution under basic conditions, followed by deprotection with p-toluenesulfonic acid, delivered 24 in 69% yield in two steps. Macrocyclization by an intramolecular Mitsunobu reaction of diol 24 afforded (rac)-25 as a white solid. Chiral SFC separation of (rac)-25 delivered pure atropisomers, (R_a)-25 and (S_a)-25. Finally, the hydrolysis of (R_a)-25 with LiOH afforded compound (R_a)-10. Compound (R_a)-15 was also accessed from intermediate 18 (Scheme 2). Stille coupling of 18, preferred over its Suzuki alternative due to better yield, with stannane 26 (prepared in 2 steps; see the Supporting Information) delivered indole 27. Deprotection of the PMB group with DDQ afforded 28. Mesylation of alcohol 28 followed by in situ treatment with potassium thioacetate in DMF afforded 29. Coupling with iodide 30 (which could be made in seven steps, see Supporting Information), followed by deprotection of the alcohols, delivered diol 31. Next, macrocyclization by the intramolecular Mitsunobu reaction gave (rac)-32, which was further purified by chiral SFC to afford atropisomer (R_a)-32 as a suitable intermediate to install diverse side chains. Reaction of (R_a)-32 with 4-(2-bromoethyl)morpholine hydrobromide under basic conditions followed by hydrolysis gave compound (R_a)-15.

Scheme 1. Synthesis of Compound (R_a)-10.

Reagents and conditions: (a) BH₃.THF, THF, 50 °C, 16 h, 86%; (b) TBDMSCl, imidazole, DMAP, DCM, rt, 16 h, 59%; (c) CPhos-Pd-G3, K₂CO₃, 1,4-dioxane/water (10:1), 40 °C, 3 h; (d) MeI, Cs₂CO₃, DMF, rt, 2 h, 39%; (e) DDQ, DCM/water (10:1), 2 h, rt, 98%; (f) DIPEA, MsCl, DCM, 0 °C, 30 min; then KSAc, DMF, rt, 16 h, 82%; (g) MeOH/THF (3:1), K₂CO₃, rt, 48 h; (h) p-TsOH, MeOH, rt, 1 h, 69% (2 steps); (i) DTBAD, PPh₃, toluene/THF (5:1), 70 °C, 4 h, 60%; (j) Chiral SFC separation, 25%; (k) LiOH, MeOH/THF/water (1:1:1), 60 °C, 3 h, 88%.

Scheme 2. Synthesis of Compound (R_a)-15.

Reagents and conditions: (a) PdCl₂(dtbpf), K₃PO₄, toluene, 125 °C, 16 h, 78%; (b) DDQ, DCM/water (9:1), rt, 30 min, 58%; (c) DIPEA, MsCl, DCM, 0 °C, 10 min.; then KSAc, DMF, rt, 15 min, 64%; (d) MeOH/THF (2:1), K₂CO₃, rt, 1 h, 39%; (e) p-TsOH, MeOH, rt, 30 min, 98%; (f) DTBAD, PPh₃, toluene/THF (20:1), 70 °C, 10 min.; (g) Chiral SFC separation, 14% (from 31); (h) 4-(2-bromoethyl)morpholine hydrobromide, Cs₂CO₃, DMF, 60 °C, 16 h; (i) LiOH, water/MeOH/THF (1:2:2), 60 °C, 3 h, 37% (2 steps).

In summary, we disclose here a novel series of MCL-1 inhibitors with a unique isomeric pyrazole core structure. In the optimization of these leads, we first demonstrated that incorporation of the fluorine atom on the naphthalene moiety was necessary to improve cellular potency and that combined with the introduction of various side chains on the isomeric pyrazole (position R¹) enhanced solubility significantly without compromising potency. This exploration culminated in the discovery of compound (R_a)-10. We also introduced solubilizing side chains on the indole core at position R² while maintaining the fluorine atom on the naphthalene ring, using a late-stage functionalization approach with the hope of modulating properties even further. This led to the discovery of compound (R_a)-15, which also possesses remarkable cellular potency and properties. Both (R_a)-10 and (R_a)-15 fulfilled our target lead profile of combining high cell killing potency with good solubility, clean in vitro ADME-tox profile, and moderate-to-high clearance, making them suitable drugs for IV dosing with well-controlled exposure to maximize therapeutic index.

Glossary

Abbreviations

AML

acute myeloid leukemia

BAD

BCL-XL/BCL-2 associated death protein

BAK

BCL-2 homologous antagonist killer protein

BAX

BCL-2-associated X protein

BCL-2

B-cell lymphoma 2

BCL-W

BCL-2-like-2

BCL-XL

B-cell lymphoma extra-large isoform

BFL-1/A1

BCL-2-related protein A1

BH3

BCL-2 homology domain 3

BID

BH3 interacting domain death agonist protein

BIK

BCL-2 interacting killer

BIM

BCL-2 interacting mediator of cell death

BOK

BCL-2-related ovarian killer

CL_int

intrinsic clearance

CML

chronic myeloid leukemia

CTCM

calcium transient cardiomyocytes assay

CYP

cytochromes P450

hERG

human Ether-à-go-go related gene

KMS-12-PE

human multiple myeloma cell line KMS-12-PE

KO

gene knockout

MCL-1

myeloid cell leukemia-1

MM

multiple myeloma

MOLP-8

human multiple myeloma cell line MOLP-8

NOXA

NADPH oxidase activator

SFC

supercritical fluid chromatography.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.3c00141.

• Pharmacological analysis, general methods, analytical analysis, experimental procedures, synthesis schemes, compound analysis, and crystallography data (PDF)

The authors declare no competing financial interest.