The succinct synthesis of AT13387, a clinically relevant Hsp90 inhibitor

Abstract

AT13387 is an orally bioavailable clinical candidate developed to inhibit theheat shock protein 90 (Hsp90). This article describes a modified synthetic route for the multi-gram production of AT13387 in 46% overall yield. The modified synthetic route is short, avoids stringent reaction conditions and difficult purifications, which led to increase in an overall yield.

Introduction

Eukaryotic cells have evolved to express the molecular chaperone family of heat shock proteins (Hsp) to ensure the conformational function of client proteins as well as their degradation via the proteasome. In fact, molecular chaperones are responsible for thematuration, maintenance andproper folding of nascent polypeptides into their biologically active three-dimensional structures.^[1–4]However, Hsp’scan also servea detrimentalrole in cancer, neurodegeneration, and/or autoimmune/inflammatory diseases.^[5–11]In fact, cancer cells utilize the chaperone machinery to increasetheir survival and proliferation by stabilizing mutant and overexpressed oncoproteins, such as protein kinase B (Akt), cyclin-dependent kinases 4 (Cdk4), epidermal growth factor receptor (EGFR), Her-2 and c-Met among others. Since cancer cells are highly dependent upon the chaperonemachinery, inhibition of the heat shock proteins is widely sought as a promising strategy for the treatment of cancer.^[12–14]The 90 kDa heat shock protein, Hsp90 is an extensively studied chaperone that modulates proteostasis activity^[15–18]and prevents apoptosis via the stabilization of oncogenic signaling proteins.^[13,19,20]As a consequence of these roles, Hsp90 inhibitors have been developed and have advanced to clinical evaluation.^[22–26]One such compound, AT13387 is orally bioavailable and currently undergoingphase II clinical trials. AT13387 was developed by researchers at Astex Pharmaceuticals^{[27, 28]}usinga fragment-based drug design approach and the utilization of extensive SAR studies (Hsp90 K_d = 0.71 nM, LE = 0.42).^[27,29] The original synthesis of AT13387 was achieved in thirteen steps and in an overall yield of 2.6% (Scheme 1).^[27]As illustrated in scheme 1, resorcylic acid 6and the isoindolineintermediates (5 and 10) were synthesized, and then coupled to afford 11.The 1-methylpiperazine moiety was introduced through a lengthy sequence of transformations.Although this route was usefulfor the preparation ofAT13387, it required several laborious steps and long reaction times (>3 days)in addition to the chromatographic separation of highly polar intermediates (Scheme 1).^[27]After the identification of AT13387 as a clinical candidate,attempts to reduce the complexity of its preparation and to improve the overall yieldwere pursued (Scheme 1 and 2).^{[30, 31]}Patel and coworkers reduced the number of steps and improved the overall yield to 13.3%via an alternative synthetic procedure (Scheme 2, route B), that involvedthe preparation of dioxinone-resorcylate 21utilizing a base-mediated cyclization/aromatization process, as well as isoindoline27via a Suzuki-Miyaura cross-coupling reaction. Both units (21 and 27) were combined through a Grignard-mediated amine dioxinonetransacylation. Despite an improvement in the overall yield, the reported procedure required the use of hazardous reagents and stringent reaction conditions, which is challenging to implement on a large-scale.The Liang team reported anotherprocedure, wherein the piperazino-isoindoline (27) was prepared via a [2+2+2] cycloaddition of dipropargylamine derivative 29 and propargyl alcohol, while resorcylate 6 was prepared following the synthetic route developed by Astex researchers (Scheme 2, route C). Although, the Liang team was able to produce an alternative synthetic route to prepare the piperazino-isoindoline scaffold, the synthesis of AT13387 was lengthy, and required challenging purifications with a low overall yield of 9.3%.

Scheme 1.

Reagents and reaction conditions: a) Ac₂O, BF₃.OEt₂, 75 °C, 3h; b) benzyl bromide, K₂CO₃, acetone, reflux, 8h; c) MePPh₃Br, KOC(CH₃)₃, THF, rt, 1h; d) NaOH, MeOH/H₂O, 70 °C, 3h; e) MeOH, SO₂Cl; f) NBS, benzoyl peroxide; g) acetone, Na₂CO₃; h) anisole, TFA, microwave; i) EDC, HOBt; j) NaOH, 48 h; k) N,O-dimethylhydroxylamine hydrochloride, EDC, HOBt, NEt₃, DMF; l) LiAlH₄, THF; m) NaBH(OAc)₃, DCE, AcOH; n) 10% Pd/C, EtOH, H₂.

Scheme 2.

Retrosynthetic routes developed for AT13387.

Therefore, to address these synthetic concerns and to improve the overall yield,a synthetic procedure for the multigram synthesis of AT13387 was pursued, which ultimately led to a 46% overall yield. As compared to the other procedures, thismethod producedAT13387 in seven steps, many of which did not require a protection/deprotection protocol nor column purification ofthe highly polar and weakly UV active,t-butyl 5-((4-methylpiperazin-1-yl)methyl)isoindoline-2-carboxylate (26).

Results and Discussion

2,4-Bis(benzyloxy)-5-(prop-1-en-2-yl) benzoic acid (6)was prepared by modification ofthe previously reported route^[27]as noted in Scheme 3. The phenols of 4-dihydroxybenzoic acid (32)were masked as the benzyl ethers using benzyl bromide in acetonitrile,which was then treatedwith N-bromosuccinimide (NBS)to givebenzyl 2,4-bis(benzyloxy)-5-bromobenzoate (34) as a colorless solid in high yield. A palladium-mediated crosscoupling reaction was then utilized along with potassium trifluoro(prop-1-en-2-yl)borate to produce 35 as an amorphous solid (88% yield), which was then hydrolyzed with lithium hydroxideto give the corresponding benzoic acid6 in excellentyield. The synthesis of 6did not require column purification throughout its preparation, except for intermediate35.

Scheme 3.

Reagents and reaction conditions: a) Benzyl bromide, K₂CO₃, CH₃CN, 70 °C, 5h; b) NBS,CH₃CN, rt; c) potassium trifluoro(prop-1-en-2-yl)borate, Pd(OAc)₂, XPhos, Cs₂CO₃, THF:H₂O (9:1), 90 °C, 8h; d) LiOH, THF:MeOH:H₂O (2:1:1), rt, 3h; e) 5-bromoisoindoline, EDC.HCl, HOBt, DIPEA, DCM, 0 °C to rt, 13h; f) potassium 4-methylpiperazinomethyl-trifluoroborate, Pd(OAc)₂, t-Bu₃P-HBF₄, Cs₂CO₃, Toluene:H₂O (10:3), 100 °C, 14h; g) 10% Pd/C, CH₃OH, H₂, rt, 12h.

Benzoic acid 6was then subjected to an 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride (EDC) mediated amide coupling reaction with 5-bromoisoindoline to afford (2,4-bis(benzyloxy)-5-(prop-1-en-2-yl)phenyl)(5-bromoisoindolin-2-yl)methanone(36). Optimization of these coupling conditions led to the use of EDC withhydroxybenzotriazole (HOBt) in anhydrous N,N-diisopropylethylamine (DIPEA) and dichloromethane (DCM),which gave36 in 95% yield (Table 1, entry 3).

Table 1:

Optimization of amide coupling reaction conditions.

Entry	Coupling reagent	Basea	Solventb	Temperature	Time (hrs)	Isolated Yield
1	EDC/HOBt	Et3N	DCM	25 °C	20	81%
2	EDC/HOBt	Et3N	DMF	25 °C	18	76%
3	EDC/HOBt	DIPEA	DCM	25 °C	13	95%
4	EDC/HOBt	DIPEA	DMF	25 °C	12	83%
5	EDC/HOAt	Et3N	DCM	25 °C	30	75%
6	EDC/HOAt	Et3N	DMF	25 °C	18	70%
7	EDC/HOAt	DIPEA	DCM	25 °C	18	76%
8	EDC/HOAt	DIPEA	DMF	25 °C	16	78%

^aAll reactions involve the use of anhydrous base.

^bAll reactions performed in anhydrous solvent.

^cIsolated yield, average of three independent reactions (n = 3).

The next key step that required optimization was the palladium-catalyzed cross coupling reactionbetween (2,4-bis(benzyloxy)-5-(prop-1-en-2-yl)phenyl)(5-bromoisoindolin-2-yl)methanone (36) and potassium 4-methylpiperazinomethyl-trifluoroborate (25). After extensive screening of reagent combinations (Table 2), (2,4-bis(benzyloxy)-5-(prop-1-en-2-yl)phenyl)(5-((4-methylpiperazin-1-yl)methyl)isoindolin-2-yl)methanone (17) was producedingood yield (85%) viapalladium acetate (5 mol%), tri-tert-butylphosphoniumtetrafluoroborate (10 mol%) and cesium carbonate (3 equivalents)in a heterogeneous mixture of toluene and water (10:3). As shown in Table 2, the reaction time was minimized via the use of a sealed tube and under microwave irradiation, unfortunately, the yields were modest at best (38%−67%, Table 2). It is noteworthy to mention that in comparison to previously reported methods, the late stage installation of 1-methylpiperazine avoided the need to prepare theweakly UV active intermediate26, which also contributed to an increase in the overall yield.For the final step, we observed that hydrogenolysiswascleanerwhen methanol was used as a solvent in lieu of ethanolunder the presence of a hydrogen atmosphere and 10% palladium on carbon.However, silica chromatographywasrequired to achieve>98% purity of AT13387. Overall, in comparison to route A (13 steps, overall yield = 2.6 %), route B (8 steps,overall yield = 13.3 %), or route C (>12 steps, overall yield = 9.3%), the syntheticroute described herein requires 7 steps to produce AT13387 in anoverall yield of 46%.

Table 2:

Optimization of the palladium-catalyzed cross coupling reaction for the preparation of 17.

Entry	Catalysta	Solvents (ratio)	Temp.	Time (hrs)	Isolated yieldb
1	Pd(OAc)2, XPhos	THF: H2O (7:3)	110 °C	18	52%
2	Pd(OAc)2, XPhos	Toluene: H2O (10:3)	110 °C	18	55%
3	Pd(OAc)2, t-Bu3P-HBF4	THF: H2O (7:3)	100 °C	14	63%
4	Pd(OAc)2, t-Bu3P-HBF4	Toluene: H2O (10:3)	100 °C	14	85%
5	Pd(PPh3)4	THF: H2O (7:3)	100 °C	18	47%
6	Pd(PPh3)4	Toluene: H2O (10:3)	100 °C	18	35%
8	Pd(OAc)2, XPhos	THF: H2O (7:3)	110 °C#	0.5	46%
9	Pd(OAc)2, XPhos	Toluene: H2O (10:3)	110 °C#	0.5	51%
10	Pd(OAc)2, t-Bu3P-HBF4	THF: H2O (7:3)	100 °C#	0.5	40%
11	Pd(OAc)2, t-Bu3P-HBF4	Toluene: H2O (10:3)	100 °C#	0.5	67%
12	Pd(PPh3)4	THF: H2O (7:3)	100 °C#	0.5	41%
13	Pd(PPh3)4	Toluene: H2O (10:3)	100 °C#	0.5	38%

^aAll reactions performed using, Pd-catalyst (5 mol%), ligand (10 mol%) and Cs₂CO₃(3 equivalents).

^bIsolated yield, average of three independent reactions (n = 3).

^#Reactions performed under microwave irradiations in a sealed tube.

Conclusions

In conclusion, a synthetic method for the multigram synthesis of the Hsp90 inhibitor AT13387 was achieved in high yield. In fact, the route is very succinct,producing AT13387 in high yieldswhich may be useful for assessing the clinical utility of AT13387 or the generation of analogs. Since, many of the intermediates do not require chromatographic purification, which further enhances the usefulness of this method for the large scale synthesis of AT13387.

Experimental Section

¹H NMR and ¹³C NMR spectra were recorded on Bruker AVIIIHD 400 MHz NMR with a broadband X-channel detect gradient probe using CDCl₃ as a solvent. Chemical shifts wererecorded in δ (ppm) with tetramethylsilane (TMS) as an internal reference. J (coupling constant) values were estimated in Hertz (Hz). The following notation is used: br – broad, s – singlet, d – doublet, t – triplet, q – quartet, m – multiplet, dd – doublet of doublets. Mass spectra (MS) were recorded on a Waters 2695 mass spectrometer using ESI ionization mode. Thin layer chromatography was performed using TLC silica gel 60 F₂₅₄. Normal phase column chromatography was performed on a flash chromatography system (AI-580S, Yamazen) using 30 μm silica gel column.All solvents were reagent grade and used without further purification.Compound 1, 6, 17 and 33-36 are already reported. ^{[27, 32, 33]}

Procedure for the synthesis of (2,4-Dihydroxy-5-isopropylphenyl)(5-((4-methylpiperazin-1-yl)methyl)isoindolin-2-yl)methanone (1)^[27]

Palladium on carbon (10%) (214.6 mg, 2.01 mmol, 0.3 eq.) was added to a solution of (2,4-Bis(benzyloxy)-5-(prop-1-en-2-yl)phenyl)(5-((4-methylpiperazin-1-yl)methyl)isoindolin-2-yl)methanone (3.95 g, 6.72 mmol, 1 eq.) in methanol (30 mL). The reaction mixture was then stirred under a hydrogen atmosphere for 12 h. The mixture was filtered through a pad of celite, eluted with ethyl acetate and the eluent was concentrated. The resulting residue was purified by silica chromatography (0–10% MeOH: DCM) to afford (2,4-dihydroxy-5-isopropylphenyl)(5-((4-methylpiperazin-1-yl)methyl)isoindolin-2-yl)methanone as a colorless solid (2.2 g, 81%): ¹H NMR (400 MHz, CDCl₃) δ 7.38 (s, 1H), 7.23–7.18 (br m, 3H), 6.33 (s, 1H), 5.01 (s, 2H), 4.98 (s, 2H), 3.55 (s, 2H), 3.26 – 3.19 (m, 1H), 2.89–2.68 (br m, 8H), 2.57 (s, 3H), 1.24 (d, J = 6.8 Hz, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 171.22, 159.93, 158.76, 136.71, 136.46, 135.72, 128.97, 126.20, 126.07, 123.47, 122.53, 108.48, 103.62, 62.10, 54.12 (2C), 53.69 (2C), 50.92 (2C), 44.43, 26.58, 22.87 (2C); HRMS (ESI+): m/z [M+H⁺]calcd. for C₂₄H₃₁N₃O₃, 410.2438; found: 410.2438.