Discovery of a Potent and Isoform-Selective ... - ACS Publications

Jan 29, 2013 - Celgene Avilomics Research, 45 Wiggins Avenue, Bedford, Massachusetts 01730, United States. •S Supporting Information. ABSTRACT: ...
4 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/jmc

Discovery of a Potent and Isoform-Selective Targeted Covalent Inhibitor of the Lipid Kinase PI3Kα Mariana Nacht,* Lixin Qiao, Michael P. Sheets, Thia St. Martin, Matthew Labenski, Hormoz Mazdiyasni, Russell Karp, Zhendong Zhu, Prasoon Chaturvedi, Deepa Bhavsar, Deqiang Niu, William Westlin, Russell C. Petter, Aravind Prasad Medikonda, and Juswinder Singh Celgene Avilomics Research, 45 Wiggins Avenue, Bedford, Massachusetts 01730, United States S Supporting Information *

ABSTRACT: PI3Kα has been identified as an oncogene in human tumors. By use of rational drug design, a targeted covalent inhibitor 3 (CNX-1351) was created that potently and specifically inhibits PI3Kα. We demonstrate, using mass spectrometry and X-ray crystallography, that the selective inhibitor covalently modifies PI3Kα on cysteine 862 (C862), an amino acid unique to the α isoform, and that PI3Kβ, -γ, and -δ are not covalently modified. 3 is able to potently (EC50 < 100 nM) and specifically inhibit signaling in PI3Kα-dependent cancer cell lines, and this leads to a potent antiproliferative effect (GI50 < 100 nM). A covalent probe, 8 (CNX-1220), which selectively bonds to PI3Kα, was used to investigate the duration of occupancy of 3 with PI3Kα in vivo. This is the first report of a PI3Kα-selective inhibitor, and these data demonstrate the biological impact of selectively targeting PI3Kα.

■

inhibitor that does not bind to β, δ, and γ may spare toxicities caused by prevention of platelet aggregation and activation,13 inhibition of immune responses mediated by T, B, and mast cells,14 and blocking anti-inflammatory signals,15 respectively. Recently, there has been a resurgence of interest in the use of covalent inhibitors to address challenges in potency and selectivity to drug targets.16 There are many examples of covalent drugs that have proven to be safe and successful therapies for a wide variety of indications, including blockbuster drugs such as esomeprazole (AstraZeneca) and clopidogrel (Sanofi-Aventis/Bristol-Myers Squibb).16 The majority of these covalent drugs have been discovered serendipitously, but recently there has been progress in the rational design of covalent drugs, termed targeted covalent inhibitors, which has been described mainly in the area of kinases17 and more recently in the protease field.18 Here we describe the first example of a targeted covalent inhibitor of the lipid kinase family that is an isoform-selective inhibitor of PI3Kα.

INTRODUCTION The family of lipid phosphoinositide 3-kinases (PI3Ks) plays a key regulatory role in many cellular processes including proliferation, cell survival, differentiation, and metabolism (reviewed in refs 1 and 2). There are three classes of PI3Ks. The class I family is the best studied and the family most implicated in cancer cell growth. The class I PI3Ks are heterodimers consisting of a p110 catalytic subunit and a p85 regulatory subunit. The PI3K class I family consists of four isoforms, α, β, γ, and δ. PI3Ks α and β are ubiquitously expressed, whereas expression of γ and δ is more restricted, mostly to leukocytes. The lipid kinases convert phosphatidylinositol 4,5-bisphosphate to phosphatidylinositol 3,4,5-trisphosphate, and the process is reversed by the phosphatase and tumor suppressor protein PTEN.3,4 In recent years, activated class IA PI3K signaling has been increasingly linked to cancer. PI3K can be activated by receptor tyrosine kinase (RTK) signaling1,5−7 and inactivation of the negative regulator PTEN,8 and through direct somatic mutations in PIK3CA, the gene that encodes the p110α subunit of PI3K.1,9 PIK3CA is mutated in up to 30% of human cancers, including breast, colon, prostate, and endometrial cancers.1,10 Importantly, it has been demonstrated that these mutations are in fact oncogenic.11 Several PI3K inhibitors are currently in clinical development, but most are pan-PI3K inhibitors that bind reversibly to the ATP site.12 The human tumor genetic data suggest that mutations in PI3Kα drive tumor growth and therefore support the need for a PI3Kα selective inhibitor. Moreover, a selective © 2013 American Chemical Society

■

DESIGN OF A TARGETED COVALENT INHIBITOR OF PI3Kα A major challenge in the design of isoform-selective PI3K inhibitors has been the high similarity of the lipid kinase ATP binding sites. We examined the ATP binding site and nearby residues of PI3Kα to identify opportunities for selective Received: June 21, 2012 Published: January 29, 2013 712

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

Figure 1. Binding site sequence and structure of PI3Kα complexed with compound 3. (a) Binding site composition of class 1 PI3K isoforms α, β, γ, and δ. (b) X-ray complex of 3 to PI3Kα. The protein backbone is shown as a cartoon. The structure of 3 is shown as a stick representation, and the binding site residues are shown as lines. Hydrogen bonds between 3 and the protein are shown as dashed lines. C862 is highlighted, as well as D810, Y836, and V851 (hinge region) which are involved in hydrogen bonds with the inhibitor. (c) Overlay of 3−PI3Kα (yellow) X-ray complex with Xray complex of GDC-0941 bound to PI3Kγ (3DBS, Ggreen). (d) Electron density of 3 bound to PI3Kα. The ligand and neighboring protein side chains are shown as a stick model, colored according to the chemical atom type (3 in cyan, human PI3K in gray, N in blue, O in red, and S in yellow). The ligand molecule is shown superimposed with the refined 2Fo − Fc electron density map contoured at 1.0.

covalent modification. Our design strategy was to use the binding mode of the reversible pan-PI3K inhibitor GDC094119 to create a targeted covalent inhibitor. The binding mode of GDC-0941 to PI3Kγ had been determined by X-ray crystallography, and we used molecular modeling to model the binding of GDC-0941 to the X-ray structure of PI3Kα.20 By aligning the PI3Kα model with the X-ray complex of PI3Kγ, we identified C862 as a promising amino acid to target, since it is near the small molecule binding site (∼10 Å), provides a relatively strong nucleophilic center, and is unique to PI3Kα (it is not present in β, γ, and δ; Figure 1a). Through a series of design cycles exploring both linker spacing and electrophilic functional groups, we identified 3 (CNX-1351) as a targeted covalent inhibitor of PI3Kα (Figure 2a). In addition, we developed an irreversible biotinylated covalent probe 8 (CNX1220; see Figure S1 in Supporting Information and Scheme 2) that allows for the measurement of free protein versus bound target protein.

■

Figure 2. 2D chemical structure of (a) compound 3 and (b) GDC0941.

coupling with (1H-indazol-4-yl)boronic acid to give 4indazolylthienopyrimidine 2. After Boc deprotection, the resulting piperazine was acylated with carbonyldiimidazoleactivated 6-methyl-4-oxohept-5-enoic acid, which delivered the desired gem-dimethylenone compound 3. The biotinylated covalent probe 8, was designed to measure the occupancy of PI3Kα by covalent inhibitors (i.e., 3). The synthesis of 8 is described in Scheme 2. Compound 1 was treated with 3-hydroxyphenylboronic acid under Suzuki conditions to give phenol 5. Upon deprotection, the resulting piperazine was treated with 5-(diethoxyphosphoryl)-4-oxopen-

CHEMISTRY

The synthetic approach for the covalent inhibitor 3 (Figure 2a) is outlined in Scheme 1. Thienopyrimidyl chloride 1 was prepared using methods similar to those reported for GDC0941 (Figure 2b).21,22 Compound 1 then underwent Suzuki 713

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

Scheme 1a

Reagent and conditions: (a) (1H-indazol-4-yl)boronic acid, Na2CO3, H2O, Pd(PPh3)2Cl2, DMA, microwave, 120 °C, 1 h, 73%; (b) HCl in dioxane, DCM−MeOH, 100%; (c) 6-methyl-4-oxohept-5-enoic acid, carbonyl diimidazole, DMA, rt, 45 min; then de-Boc intermediate, DIPEA, DMA, rt, overnight, 62%.

a

Scheme 2a

Reagent and conditions: (a) 3-hydroxyphenylboronic acid, Pd(PPh3)4, Na2CO3, toluene/ethanol/H2O, 120 °C, 1 h, 69%; (b) HCl in dioxane, DCM−MeOH, 100%; (c) compound 3, carbonyldiimidazole, DMA, rt, 45 min; then de-Boc intermediate, DIPEA, DMA, rt, overnight, 47%; (d) K2CO3, DMA−H2O, 70 °C, 4 h; (e) CF3CO2H, dichloromethane; (f) N-biotinyl-NH-(PEG)2-COOH-DIPEA (20 atoms), HATU, DIPEA, DMA. a

■

RESULTS AND DISCUSSION Mass spectrometry was used to confirm that 3 covalently bonds to PI3Kα but not to the other class I isoforms (see Figure S2a in Supporting Information). To confirm that C862 was the amino acid on which the bond was formed, PI3Kα protein was incubated with 3 and then subjected to tryptic digestion. Subsequent peptide analysis of the digested, modified protein confirmed modification of C862 (see Figure S2b in Supporting Information).

tanoic acid 4, which was prepared by condensing succinic anhydride with deprotonated diethyl methylphosphonate. The resulting phosphonate 6 subsequently underwent Wittig− Horner−Emmons reaction with tert-butyl 4-formylbenzylcarbamate, giving trans-enone 7 only. Final assembly of covalent probe 8 was achieved via de-Boc of compound 7 followed by amide formation with commercially available polyethylene glycol linked biotin glutaric acid. 714

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

dimethylenone of 3 would react with the thiol present in GSH. After 2 h at 37 °C, >80% of free 3 was recovered, consistent with the no-GSH control samples (see Figure S3a in Supporting Information). This indicates that 3 is not indiscriminately reactive toward GSH. 3 was incubated in albumin-depleted human plasma at 37 °C for 1 or 3 h to determine whether it would covalently bond with any highly abundant human plasma proteins. None of the observed human plasma proteins showed mass shifts consistent with the mass of 3, suggesting an inability to covalently modify these off-target plasma proteins at detectable levels (see Figure S3b in Supporting Information). Compound 3 Inhibits PI3Kα Signaling in Cells and Shows Prolonged Inhibition Consistent with a Covalent Mechanism of Action. SKOV3 ovarian cancer cells have a mutation in PIK3CA that results in a constitutively activated PI3Kα signaling pathway.23 These cells were used to examine if 3 could inhibit P-AktSer473 as a measure of inhibition of PI3K signaling. 3 (1−1000 nM) was added to SKOV3 cells for 1 h. Lysates of the cells were interrogated by immunoblot analysis for P-AktSer473. A noncovalent pan-PI3K inhibitor, GDC-0941, was used as a control and demonstrates that 3 inhibits PI3K signaling in SKOV3 cells, with potency (EC50 of 10−100 nM; see Figure S4 in Supporting Information) similar to that of the pan-PI3K inhibitor. To demonstrate the covalent mechanism of action in cells, SKOV3 cells were transiently treated with the covalent inhibitor, and prolonged inhibition of P-AktSer473 was examined. SKOV3 cells were incubated with 3 or the reversible inhibitor GDC-0941 (500 nM each) for 1 h. Cells were washed, and medium was replaced with fresh compound-free medium at regular intervals (approximately once every hour for 3 h). Cell lysates were prepared at 1, 6, and 24 h after compound removal. Cell lysates were interrogated by immunoblot analysis for PAktSer473 (Figure 3). As expected, signaling returned immediately after removal of the reversible inhibitor, GDC-0941, but was inhibited >6 h after removal of 3, consistent with the activity of an irreversible compound. Compound 3 Inhibits Growth of Cells Dependent on PI3Kα. To investigate the functional consequence of inhibiting PI3Kα in cells, two cell lines with different PIK3CA activating mutations, SKOV3 ovarian cancer cells (H1047R) and MCF-7 breast cancer cells (E545K), were treated with 3 and growth was monitored. Both PIK3CA-driven cell lines were growth inhibited by exposure to 3 for 96 h (GI50 of 78 and 55 nM, respectively) (Table 2). These data suggest that selective inhibition of PI3Kα will lead to an antiproliferative effect on PIK3CA-driven tumors. Compound 3 Inhibits PI3Kα Signaling in Vivo and Bonds to p110α. To examine the activity of 3 on PI3Kα signaling in vivo, 3 was delivered to nude mice (100 mg/kg ip) once a day for 5 consecutive days. Because PI3Kα is ubiquitously expressed, spleens were harvested from the mice at 1, 4, and 24 h after the last administration of compound to measure the impact of 3 on PI3K signaling. Lysates from the spleens were immunoblotted for P-AktSer473 (Figure 4a). Inhibition of PI3K signaling was detected as a decrease in PAktSer473 at 1 and 4 h after the last dose. An irreversible covalent inhibitor allows for the measurement of free protein versus bound target protein using a covalent probe that bonds only to free protein. A biotinylated covalent probe that is selective for p110α (8; see Figure S5a and Figure S5b in Supporting Information) was used to detect free,

Using X-ray crystallography, we confirmed our predicted binding mode of 3 to PI3Kα (Figure 1b). The X-ray complex of GDC-0941 has been described previously, binding to PI3Kβ, -γ, and δ, and has been shown to adopt a three-point binding mode to the PI3K ATP site.19 Comparison of the published binding mode of GDC-0941 (PDB code 3DBS) with our X-ray complex of 3 to PI3Kα shows similar binding modes in the PI3K ATP binding site but also important differences (Figure 1c). The two compounds adopt a similar packing arrangement of the thienopyrimidine core. Both employ the morpholine oxygen to form a hydrogen bond with the hinge region and place the indazole into the affinity pocket. There are differences in the binding mode of the substituted piperazine functionality between the two compounds. Unlike the binding mode of GDC-0941, in which the piperazine is kinked out of the plane relative to the thienopyrimidine core and packs against Arg770 and Ser773, the substituted piperazine of 3 is directed toward the other side of the binding site toward the side chain of C862. The dimethylenone is clearly seen in the electron density to form a covalent bond with C862 (Figure 1d). Compound 3 Is a Selective Covalent Inhibitor of PI3Kα. The biochemical potency and selectivity of 3 were examined. 3 was tested against all four of the class I PI3K enzymes α, β, γ, and δ. In an end point assay, 3 potently inhibited PI3Kα and was 20−400 times less potent against β, γ, and δ (Table 1). Because of the irreversible mechanism of Table 1. Biochemical Selectivity of Compound 3a kinase

IC50_App b

PI3Kα PI3Kβ PI3Kγ PI3Kδ PI3KC2A PI3KC3 PI4Kα PI4Kβ SPHK1 SPHK2

6.8 nM 166.0 nM 240.3 nM 3020.0 nM >1 μM >1 μM >1 μM >1 μM >1 μM >1 μM

a

Compound 3 was tested in a panel of 10 lipid kinases, and an IC50 apparent was determined (IC50_App). bIrreversible inhibitors show time-dependent inhibition. Thus, an IC50 apparent is used to determine the inhibitory activity of irreversible molecules at a given time point.

action, 3 shows time-dependent inhibition. Thus, IC 50 apparents are used to report the inhibitory activity of 3 at a given time point. 3 showed even greater selectivity against the class II lipid kinases, none of which were appreciably inhibited at the highest concentration of drug tested (1 μM) (Table 1). Moreover, 3 was tested against a panel of 60 kinases, representing all branches of the kinome tree, and showed excellent selectivity; no kinases were inhibited by greater than 40% using 1 μM compound (see Table S1 in Supporting Information), suggesting that 3 is a very selective PI3Kα inhibitor. Compound 3 Shows No Reactivity toward Glutathione or Abundant Human Plasma Proteins. The offtarget reactivity of 3 was assessed using mass spectrometry methods. 3 was reacted with a 500-fold molar excess of glutathione (GSH), a tripeptide and abundant intracellular thiol, in a metabolism-free setting to determine whether the 715

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

PI3Kα but not in the other isoforms has enabled us to achieve selective covalent inhibition of PI3Kα. This is in contrast to the natural product wortmannin, which is also a covalent inhibitor of the PI3K family but which targets a conserved Lys and therefore shows no isoform selectivity. This is also the first reported X-ray structure of a targeted covalent inhibitor to the lipid kinase family. Our predicted binding model of 3 was in good agreement with the PI3Kα Xray crystal structure that we determined. We have recently compared the reactivity and geometry of the β,β-dimethylenone that is present in 3 with other five α,β-unsaturated ketones using high level quantum mechanical calculations.24 The β,β-dimethylenone showed the highest activation barrier in aqueous solution, indicating that it is a highly stable electrophile compared to α- or β-methylenone and α,β-dimethylenone. The covalent mechanism of action enables complete and durable inhibition of PI3Kα, which leads to silencing of the target protein and its downstream oncogenic signaling. Our data show that the inhibition of PI3Kα persists even after washout of drug, which could enable less frequent dosing and minimize off-target activity. Less frequent dosing coupled with specificity toward the α isoform may enable a higher therapeutic index by sparing potential toxicities from pan inhibition of this central lipid kinase family with pleiotropic functions. Another important benefit of the covalent strategy is the companion covalent probe that was used preclinically to assess occupancy of PI3Kα in vivo and could be a powerful translational tool in assessing PK/PD in the clinic. 3 is a powerful tool compound that will allow several important questions about PI3K biology to be addressed. The selectivity for PI3Kα and the prolonged duration of inhibition will allow the investigation of the role of PI3Kα in different tumor models with different genetic contexts. Moreover, there are reports regarding the contribution of α and β to glucose homeostasis offering diverging conclusions; a selective α inhibitor may resolve these open questions. Although 3 has useful properties as a tool compound, it also has deficiencies as a lead compound. The pharmacokinetic properties of the compound are suboptimal and are the focus of a medicinal chemistry effort to improve its oral bioavailability while maintaining its potency and selectivity. This will be the focus of a future publication. Finally, emerging clinical data support the notion that combinations of targeted therapies may be more effective than monotherapy and may help reduce the emergence of drug resistance. Thus, a selective and durable inhibitor of PI3Kα may be a very important component of future combination approaches.

Figure 3. Inhibition of p-AktSer473 continues after removal of compound 3. SKOV-3 cells were incubated with 3 or GDC-0941 (500 nM each) for 1 h. Medium was replaced with fresh compoundfree medium at regular intervals, and cell lysates were prepared at the indicated times after compound removal. Cells were processed for incell Western analysis (see Experimental Section) to determine expression of P-AktSer473. Signaling returned immediately after removal of the reversible inhibitor, GDC-0941, but was inhibited >6 h after removal of 3. Data shown are from duplicates. Experiment was performed at least 3 times: (∗∗) p < 0.001; (∗) p < 0.05.

Table 2. Compound 3 Inhibits Growth of Cells Dependent on PI3Kαa GI50 ± SEM (nM) GDC-0941 3

SKOV3 (PIK3CAH1047R)

MCF-7 (PIK3CAE545K)

147.6 ± 36.6 77.6 ± 16.8

70.2 ± 13.5 54.7 ± 14.3

a

Cells with activating mutations in PI3KCA (H1047R or E545K) were growth inhibited by the PI3Kα selective compound 3 or the pan-PI3K inhibitor GDC-0941. Tumor cells (3000−7000 cells/well) were placed in a 96-well plate and allowed to settle for 5 h. Compounds were serially diluted (1 μM to 4 nM) and added to the cells. After a 96 h incubation, cell viability was measured using Cell Titer-Glo (Promega) to determine the growth inhibition. Data shown are an average of four experiments, each performed in duplicate.

■

uninhibited p110α in spleen cells. There was little free p110α at 1 and 4 h after the last dose, indicating that 3 irreversibly bonded to p110α in vivo (Figure 4b). At 24 h after the final dose, free p110α was detectable because of protein resynthesis, concomitant with the return of P-AktSer473 expression.

EXPERIMENTAL SECTION

Chemistry. General Experimental Details. All commercial reagents and anhydrous solvents were obtained from commercial sources and were used without further purification unless otherwise specified. Microwave chemistry was done using CEM Explorer. LC−MS was used to analyze the purity of all compounds on an Agilent LC-1200 series coupled with Agilent 6120 quadruple mass spectrometer. The mobile phase for all mass analysis consisted of acetonitrile−water mixtures with 5 mM ammonium acetate and 5 mM acetic acid. All final compounds were greater than 95% pure. Preparative reverse-phase chromatography was carried out using using a Prep Varian ProStar model 210 on a Microsorb-MV 100-8 C18 Dynamax 250 mm × 21.4 mm column with a linear gradient from 10% to 90% CH3CN in H2O over 20 min (0.1% trifluoroacetic acid) and a flow rate of 20 mL/min. 1 H and 13C NMR spectra (δ, ppm) were recorded on a JEOL AMX-

■

CONCLUSION This is the first detailed report of a selective covalent inhibitor of PI3Kα. Recently, Novartis and Intellikine launched phase I clinical trials for α-selective inhibitors (BYL719 and INK1117, respectively, AACR 2012), but detailed reports of these drugs have not been published. Our data demonstrate that the application of a targeted covalent inhibitor strategy to PI3Kα has a number of unique features related to the covalent mechanism of action. The targeting of a Cys which is present in 716

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

Figure 4. Compound 3 inhibited p-AktSer473 in mouse spleens and bonds to PI3Kα in vivo. Compound 3 was delivered into the intraperitoneal cavity of nude mice at 100 mg/kg once a day for 5 consecutive days (n = 3 mice per group). Spleens were harvested from the mice at the indicated times after the last dose (1−24 h) and interrogated by immunoblot for P-AktSer473 (a) or for PI3Kα occupancy (b) as described in Experimental Section. Inhibition of PI3K signaling was detected as a decrease in P-AktSer473 at 1 and 4 h after last dose. Similarly, there was little/no free PI3Kα at 1 and 4 h after the last dose ((∗∗) p < 0.001). Protein resynthesis accounts for the return of free protein and signaling activity at 24 h ((∗∗) p < 0.001). 400 (400 MHz) or Bruker AVANCE-3 (400 MHz) instrument. Highresolution mass spectra were measured using a Waters Q-Tof Premier operating in +ESI W-mode. tert-Butyl 4-((2-Chloro-4-morpholinothieno[3,2-d]pyrimidin6-yl)methyl)piperazine-1-carboxylate (1). 2-Chloro-4morpholinothieno[3,2-d]pyrimidine-6-carbaldehyde (0.40 g, 1.5 mmol), tert-butylpiperazine 1-carboxylate, and 0.2 mL of acetic acid were dissolved in 12 mL of dichloroethane. The mixture was stirred at room temperature for 2 h. NaBH(OAc)3 (0.54g, 2.5 mmol) was added to the reaction mixture, and the resulting mixture was stirred at room temperature for 10 h. Then 20 mL of NaHCO3 aqueous solution and 10 mL of DCM were added. The organic layer was separated and dried over Na2SO4. After removal of solvent, the crude product was subjected to chromatography on silica gel (eluent, EtOAc/hexane 3:7). A total of 0.30 g of the title compound was obtained. MS: m/z 454.2 (M + H). 1H NMR (CDCl3, 400 MHz): δ 7.16 (s, 1H), 4.37 (q, J = 4.4 Hz, 4H), 3.90 − 3.78 (m, 6H), 3.46 (t, J = 4.8 Hz, 4H), 2.49 (t, J = 4.8 Hz, 4H), 1.45 (s, 9H) ppm. tert-Butyl 4-((2-(1H-Indazol-4-yl)-4-morpholinothieno[3,2d]pyrimidin-6-yl)methyl)piperazine-1-carboxylate (2). To a 10 mL microwave tube were added Boc-piperazine 1 (91 mg, 0.20 mmol), (1H-indazol-4-yl)boronic acid (40 mg, 0.25 mmol), Na2CO3 (84 mg, 0.79 mmol) with 1.6 mL of degassed N,N-dimethylacetamide, and 0.4 mL of degassed water, followed by Pd(PPh3)2Cl2 (8.2 mg, 0.01 mmol). The whole system was repurged with nitrogen and then heated in a microwave reactor at 120 °C for 1 h. The reaction mixture was diluted with water and extracted with dichloromethane (DCM). The combined organic layers were dried over anhydrous Na2SO4. After the removal of solvent under reduced pressure, the crude material was purified by silica gel column chromatography (DCM/ MeOH v/v 30/1) to afford compound 2 (78 mg, 73%) as slight yellow solid. 1H NMR (CDCl3, 400 MHz): δ 9.0 (s, 1H), 8.26 (d, J = 6.8 Hz, 1H), 7.6−7.5 (m, 3H), 7.38 (s, 1H), 4.15−4.06 (m, 4H), 3.98−3.75 (m, 6H), 3.48 (t, J = 4.6 Hz, 4H), 2.60−2.43 (m, 4H), 1.46 (s, 9H) ppm. MS: m/z 536.1 [M + H]. 1-(4-((2-(1H-Indazol-4-yl)-4-morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-6-methylhept-5-ene1,4-dione (3). To a stirred solution of compound 2 (54 mg, 0.1 mmol) in 1 mL of DCM and 1 mL of MeOH was added 4 M HCl in dioxane (2 mL), and the resulting reaction mixture was stirred at room temperature for 1 h. The solvent was then removed under reduced pressure, and the residue was dried in vacuum and used directly for the following step. MS: m/z 436.1 (M + H). 6-Methyl-4-oxohept-5-enoic acid (17 mg, 0.11 mmol) and carbonyldiimidazole (17 mg, 0.11 mmol) were mixed in 0.5 mL of N,N-

dimethylacetamide for 30 min before transferring into the de-Boc intermediate above in 1 mL of N,N-dimethylacetamide. The reaction mixture was stirred overnight, then subjected to normal EtOAc/H2O workup. After solvent removal, the resulting solid was filtered and washed with acetonitrile, giving 36 mg of the desired product as a pale white solid (62%). 1H NMR (400 MHz, CDCl3): δ 8.97 (s, 1H), 8.22 (d, 1H, J = 6.4 Hz), 7.52 (d, 1H, J = 6.4 Hz), 7.44 (t, 1H, J = 7.2 Hz), 7.36 (s, 1H), 6.11 (d, 1H, J = 1.2 Hz), 4.06 (t, 4H, J = 4.4 Hz), 3.89 (t, 4H, J = 5.2 Hz), 3.79 (s, 2H), 3.63 (br t, J = 4.4 Hz), 3.53 (br t, J = 4.4 Hz), 2.76 (t, 2H, J = 6.4 Hz), 2.59 (t, 2H, J = 6.4 Hz), 2.48 (m, 4H), 2.11 (d, 3H, J = 1.2 (Hz), 1.84 (d, 3H, J = 1.2 Hz) ppm. 13C NMR (100 Hz, CDCl3): δ 199.5, 170.34, 162.5, 160.4, 158.1, 155.6, 149.1, 140.9, 136.5, 132.0, 126.4, 123.8, 123.7, 121.9, 121.6, 113.1, 111.5, 66.8, 57.4, 53.4, 52.9, 52.7, 46.4, 45.1, 41.6, 38.6, 30.9, 27.7, 26.9, 20.8 ppm. MS: m/z 574.2 (M + H). HRMS calcd for C30H37N7O3S 574.2600, found 574.2634. 5-(Diethoxyphosphoryl)-4-oxopentanoic Acid (4). To a solution of diethyl methylphosphonate (0.76g, 5.0 mmol) in 20 mL of THF at −78 °C was added n-BuLi (2.5 N in hexane, 5.0 mmol) slowly. The reaction mixture was stirred at −78 °C for 1 h. Succinic anhydride (0.50 g 5.0 mmol) in 5.0 mL of anhydrous THF was introduced slowly into the mixture at −78 °C. The reaction mixture was stirred for 1 h at −78 °C. Then 1 N HCl (5.0 mL) aqueous solution was added and the mixture was warmed to room temperature. The THF was then removed under vacuum, and the remaining aqueous residue was extracted with dichloromethane (3 × 10 mL). The organic layer was dried over Na2SO4, filtered, and the solvent was removed. The residue was purified by chromatography on silica gel (eluent, EtOAc/MeOH 20:1) to provide the acid 4. MS: m/z 253.1 (M + H); 1H NMR (400 MHz, CDCl3): δ 4.15 (4H, m), 3.18 (1H, s), 3.13 (1H, s), 2.95 (2H, t, J = 6.4 Hz), 2.63 (2H, t, J = 6.4 Hz), 1.33 (6H, m) ppm. tert-Butyl 4-((2-(3-Hydroxyphenyl)-4-morpholinothieno[3,2d]pyrimidin-6-yl)methyl)piperazine-1-carboxylate (5). To a mixture of chloropyrimidine 1 (305 mg, 0.67 mmol), 3-hydroxyphenylboronic acid (139 mg, 1.0 mmol), and sodium carbonate (214 mg, 2 mmoL) in degassed toluene/ethanol/water (6 mL/3.6 mL/1.8 mL) was added tetrakis(triphenylphosphine)palladium (51 mg, 0.067 mmol). The reaction mixture was heated to 120 °C for 1 h under nitrogen, and the solvent was then removed under reduced pressure. The residue was subject to a quick aqueous EtOAc/water workup and purified by chromatography on silica gel (EtOAc/heptane v/v 1/1), giving 238 mg of yellow foamy solid as the title compound (69%). MS: m/z 512.3 (M + H). 1H NMR (400 MHz, DMSO-d6): δ 9.49 (s, 1H), 7.87−7.81 (m, 2H), 7.38 (s, 1H), 7.26 (t, J = 8.11 Hz, 1H), 6.87−6.83 717

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

26.9, 23.2 ppm. MS: m/z 1169.4 (M + H); 1167.6 (M-H). HRMS Calcd for C59H81N10O11S2 1169.5528, found 1169.5529. Assays. Mass Spectrometry. Mass spectrometry was used to confirm that 3 covalently bonds to PI3Kα but not to the other class I isoforms of PI3K. 3 was incubated with PI3Kα (Millipore, 14-602), PI3Kβ (Millipore, 14-603), PI3Kδ (Millipore, 14-604), PI3Kγ (BPS bioscience, BPS-40625) at a 10:1 compound-to-enzyme molar ratio at room temperature for 1 h. After the incubation, 10 mL of 0.2% TFA in water was added to each sample. Protein enrichment was achieved with a C4-packed ZipTip (Millipore), direct spotting onto a MALDI target plate, and combination with equal volume of sinapinic acid (10 mg/mL in 50% acetonitrile) as the desorption matrix. Once dry, the samples were analyzed on an AB Sciex 4800 MALDI TOF/TOF instrument. Trypsin Digest. After the reaction of PI3Kα with 3 as described above, the protein was separated by electrophoresis on a 4−12% BisTris gel and stained with a Coomassie blue dye. The PI3Kα band was then excised from the gel and subjected to a standard in-gel trypsin digest. Peptide extract was then spotted directly on a MALDI target plate and combined with an equal volume of α-cyano-4-hydroxycinnamic acid (5 mg/mL in 80% acetonitrile) as the desorption matrix. Once dried, the spots were analyzed on an AB Sciex 4800 MALDI TOF/TOF instrument in both MS and MS/MS modes to confirm modification of the target amino acid for 3. The targeted Cys862 in peptide NSHTIMQIQCK was clearly modified. GSH Reactivity. GSH reactivity was assessed by incubating 10 μM 3 at 37 °C in PBS at pH 7.4 with or without GSH (5 mM). At various designated time points, 50 μL of the reaction mixture was diluted with 50 μL of internal standard (1 μg/mL carbutamide in acetonitrile), injected onto a C8 column (Agilent Technologies, Zorbax, 3.5 μm, SBC18, 2.1 mm × 30 mm), and analyzed with an AB Sciex QTrap 4000 mass spectrometer. Parent recovery was assessed and calculated as percent of the time zero value. Plasma Protein Bonding. Human plasma, obtained from Bioreclamation (Hicksville, NY), was subjected to albumin depletion using the Pierce Blue albumin removal kit (catalog no. 89845, Thermo Scientific, Rockford, IL) according to product procedures. The albumin-depleted plasma was then diluted to a final concentration of 1 mg/mL. It is important to note that albumin was depleted but not completely removed. A 10 mM DMSO stock of 3 was diluted 1:10 into 10 mg of the albumin-depleted human plasma. The reaction was carried out in a 96-well plate format in a 37 °C incubator. At the end of the incubation, 10 μL aliquots of the samples were diluted with 10 μL of 0.2% TFA and then pipetted with C4 ZipTips (Millipore) directly onto a MALDI target plate, using sinapinic acid as the desorption matrix (10 mg/mL in 0.1%TFA in acetonitrile/water 50:50 v/v) and ovalbumin as the internal standard. Samples were analyzed on an ABSciex 4800 MALDI TOF/TOF mass spectrometer. pAktS473 Immunoblot. SKOV3 cells were plated in SKOV3 growth medium (DMEM supplemented with 10% FBS and penicillin/ streptomycin) at a density of 4 × 105 cells per well of 12-well plates. Twenty-four hours later the medium was removed and replaced with 1 mL of medium containing test compound in 0.1% DMSO and cells were returned to the incubator for 1 h. At the end of the hour, the medium was removed and the cells were washed with PBS, then lysed and scraped into 30 μL of cell extraction buffer (Biosource, Camarillo, CA) plus Complete protease inhibitor and PhosStop phosphatase inhibitor (Roche, Indianapolis, IN). Cell debris was spun down at 13000g for 1 min, and the supernatant was taken as the cell lysate. Protein concentration of the lysate was determined by BCA Assay (Pierce Biotechnology, Rockford, IL), and 50 μg of protein was loaded per well onto a NuPAGE Novex 4−12% Bis-Tris gel (Invitrogen, Carlsbad, CA) and then transferred to Immobilon PVDF-FL (Millipore, Billerica, MA). The blot was blocked in Odyssey blocking buffer (Li-Cor Biosciences, Lincoln, NE) for 1 h and then incubated overnight at 4 °C with mouse anti-Akt (no. 2920) and rabbit anti-phospho-Akt (Ser473) (no. 9271); Cell Signaling Technology, Beverley, MA) antibodies, both diluted 1:1000 in PBS/Odyssey Buffer (1:1) + 0.1% Tween-20. The blots were washed 3 times for 5 min each in PBS +

(m, 1H), 4.00−3.93 (m, 4H), 3.87 (s, 2H), 3.84−3.77 (m, 4H), 3.39− 3.33 (m, 4H), 2.44 (s, 4H), 1.39 (s, 9H) ppm. 13C NMR (100 Hz, CDCl3): δ 159.8, 158.9, 157.6, 156.5, 154.7, 149.7, 139.2, 129.5, 123.0, 120.1, 117.6, 115.0, 113.0, 79.9, 66.7, 57.5, 52.8, 46.3, 28.4 ppm. HRMS calcd for C26H34N5O4S 512.2331, found 512.2332 (M + H). Diethyl (5-(4-((2-(3-Hydroxyphenyl)-4-morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1-yl)-2,5-dioxopentyl)phosphonate (6). Boc-piperazine 5 (360 mg, 0.7 mmol) was dissolved in 3 mL of 4 N HCl dioxane solution and dichloromethane (5 mL), and the reaction was continued for 3 h at room temperature. After removal of solvents, a white solid (350 mg) was obtained and used directly for the next step. MS: m/z 412.1 (M + H). Phosphonate 4 (750 μL, 0.25 M acetonitrile solution, 187 μmol) was treated with carbonyldiimidazole (30 mg, 187 μmol) in 1 mL of N,N-dimethylacetamide for 30 min, followed by addition of the de-Boc intermediate obtained above (27.5 mg, 50 μmol). After being stirred for an additional 30 min, the reaction mixture was concentrated and the residue was purified by preparative HPLC, giving 15 mg of white powder after lyophilization (47%). 1H NMR (400 MHz, CD3OD): δ 7.71 (t, J = 5.3 Hz, 2H), 7.23 (s, 1H), 7.18 (t, J = 7.9 Hz, 1H), 6.79 (ddd, J = 8.1, 2.5, 1.0 Hz, 1H), 4.09−4.00 (m, 4H), 3.98 (dd, J = 8.9, 4.4 Hz, 4H), 3.82 (s, 2H), 3.80−3.74 (m, 4H), 3.56−3.48 (m, 4H), 2.81 (t, J = 6.3 Hz, 2H), 2.60−2.49 (m, 4H), 2.49−2.42 (m, 2H), 1.23 (t, J = 7.1 Hz, 6H) ppm. 13C NMR (100 Hz, CD3OD): δ 202.8 (d, J = 6.2 Hz), 172.2, 163.0, 161.4, 159.0, 158.7, 151.9, 140.8, 130.4, 123.6, 120.5, 118.3, 116.1, 114.2, 67.8, 64.1 (d, J = 6.4 Hz), 58.0, 53.9 (d, J = 25.6 Hz), 47.7, 46.6, 43.0, 39.8 (d, J = 2.4 Hz), 28.0, 16.6 (d, J = 6.2 Hz) ppm. MS: m/z 646.5 (M + H). HRMS calcd for C30H41N5O7S 646.2464, found 646.2466. (E)-tert-Butyl 4-(6-(4-((2-(3-Hydroxyphenyl)-4morpholinothieno[3,2-d]pyrimidin-6-yl)methyl)piperazin-1yl)-3,6-dioxohex-1-enyl)benzylcarbamate (7). A mixture of the phosphonate 5 (13 mg, 20 μmol), tert-butyl 4-formylbenzylcarbamate (10 mg, 40 μmol), potassium carbonate (40 mg) in 1 mL of N,Ndimethylacetamide, and 100 μL of water was heated at 70 °C for 4 h. After filtration, the reaction mixture was purified by preparative HPLC, giving 10 mg of desired enone 7 as a white solid. 1H NMR (400 MHz, DMSO-d6): δ 9.49 (s, 1H), 7.87−7.82 (m, 2H), 7.67 (d, J = 8.0 Hz, 2H), 7.58 (d, J = 16.3 Hz, 1H), 7.39 (s, 2H), 7.28 (dd, J = 8.3, 2.7 Hz, 3H), 6.89 (d, J = 16.2 Hz, 2H), 4.15 (d, J = 6.1 Hz, 2H), 3.96 (dd, J = 19.8, 14.9 Hz, 6H), 3.86−3.78 (m, 4H), 3.50 (d, J = 22.2 Hz, 4H), 3.32 (d, J = 9.5 Hz, 2H), 2.90 (d, J = 6.4 Hz, 2H), 2.62 (s, 2H), 2.44 (s, 2H), 1.37 (d, J = 24.0 Hz, 9H) ppm. 13C NMR (100 Hz, MeOD): δ 199.3, 170.3, 162.2, 160.0, 157.8, 156.6, 156.0, 155.9, 149.2, 142.6, 139.7, 133.5, 129.5, 128.6, 127.9, 126.0, 123.4, 120.0, 117.4, 115.1, 113.1, 79.8, 66.8, 57.4, 53.0, 46.3, 45.3, 41.7, 35.4, 28.4, 27.1 ppm. MS: m/z 727.3 (M + H). HRMS calcd for C39H47N6O6S 727.3278, found 727.3276. Biotinylated Covalent Probe (8). To 7.5 mg of enone 7 in 1 mL of dichloromethane was added 1 mL of trifluoroacetic acid. After the mixture was stirred for 30 min, the solvent was removed under reduced pressure, giving de-Boc intermediate with m/z 627.3 (M + H). Then 1 mL of N,N-dimethylacetamide, 60 μL of DIPEA, and 9.0 mg of commercially available N-biotinyl-NH(PEG)2-COOH-DIPEA (20 atoms) (Novabiochem catalog no. 851029) were added, followed by 9 mg of HATU. After being stirred for an additional 30 min, the crude mixture was purified by preparative HPLC, giving 9 mg of desired compounds as a white powder after lyophilization. 1H NMR (400 MHz, CD3OD): δ 7.76 (s, 1H), 7.67 (br t, 2H, J = 8.4 Hz), 7.60 (m, 2H), 7.44 (t, 1H, J = 8.0 Hz), 7.32 (d, 2H, J = 8.4 Hz), 7.13 (dd, 1H, J = 2.8, 16.4 Hz), 6.83 (d, 1H, J = 16.4 Hz), 4.66 (s, 2H), 4.48 (m, 2H), 4.38 (s, 2H), 4.30 (m, 4H), 3.92 (m, 4H), 3.83 (m, 2H), 3.62 (m, 4H), 3.57 (m, 4H), 3.50 (m, 4H), 3.15−3.30 (m, 8H), 3.70 (br t, 2H, J = 6.4 Hz), 2.90 (m, 2H), 2.67−2.76 (m, 4H), 2.15−2.30 (m, 8 H), 1.90 (m, 4H), 1.70−1.80 (m, 8H), 1.60 (m. 6H,), 1.40 (m, 4H) ppm. 13 C NMR (100 Hz, CD3OD): δ 201.3, 175.9, 175.3, 173.0, 166.0, 160.7, 160.3, 159.6, 158.1, 157.4, 152.2, 150.3, 144.0, 143.1, 134.8, 133.6, 131.6, 130.0, 129.7, 129.1, 126.8, 122.6, 122.2, 121.5, 120.3, 118.3, 116.4, 115.6, 71.5, 71.2, 69.9, 67.4, 63.4, 61.6, 57.0, 55.1, 53.2, 52.9, 43.8, 41.0, 40.8, 37.8, 36.8, 36.3, 36.2, 36.1, 30.4, 29.8, 29.5, 27.5, 718

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

0.2% Tween-20 and then incubated for 1 h at room temperature with fluorescently labeled secondary antibodies (Li-Cor) diluted 1:10000 in PBS/Odyssey buffer (1:1) + 0.1% Tween-20. The blots were washed 2 times for 5 min each in PBS + 0.2% Tween-20, once in distilled water, then scanned on an Odyssey machine (Li-Cor). Band intensity was determined using the Odyssey software, and the phopho-Akt signal was normalized to total Akt within samples, then expressed as a percentage of the untreated phospho-Akt signal. Demonstration of Prolonged Duration of Inhibition. SKOV3 cells were plated in SKOV3 growth medium (DMEM supplemented with 10% FBS and pen/strep) at a density of 2.5 × 104 cells per well of Costar no. 3603 black 96-well clear flat-bottom plates. Plates were set up in quadruplicate with one plate each for the 0, 1, 6, and 24 h time points. Twenty-four hours later the medium was removed and replaced with 100 μL of medium containing test compound or DMSO as a control and cells were returned to the incubator for 1 h. At the end of the hour, the medium was removed and the cells were washed 2 times with PBS. The PBS was removed from three of the plates, replaced with 100 μL of growth medium, and the plates were returned to the incubator. The fourth plate was taken as the “0”-hour time point and developed as described for in-cell Western or Western blot. Thirty minutes after the first wash, the medium was removed from the remaining plates, replaced with 100 μL of fresh growth medium, and then the plates were returned to the incubator. At 1 h after the first wash, one plate was taken as the 1 h time point and developed as an in-cell Western. The remaining two plates were washed two more times at 1 h intervals and developed as in-cell Western or Western blot at 6 and 24 h after the first wash. In-Cell Western. Treated SKOV3 cells were washed once with PBS, then fixed for 20 min at room temperature in 4% formaldehyde in PBS. The formaldehyde was removed, and cells were washed 5 times for 5 min each with 100 μL of permeabilization buffer (PBS + 0.1% Triton X-100) at room temperature with gentle shaking. The last wash was removed and replaced with 150 μL of Odyssey blocking buffer (Li-Cor, Lincoln, NE) and incubated for 90 min at room temperature with gentle shaking. The blocking buffer was then replaced with 50 μL of primary antibody mix (rabbit anti-phospho-Akt(Ser473) at 1:100 (Cell Signaling Technology, Boston, MA) and mouse anti-tubulin at 1:5000 (Sigma Aldrich, St. Louis, MO) diluted in Odyssey blocking buffer) and incubated overnight at room temperature with gentle shaking. The next morning, the antibody mix was removed and the wells were washed 5 times for 5 min each with PBS + 0.1% Tween-20. The last wash was replaced with 50 μL of secondary antibody mix (goat anti-rabbit-IRDye-680 and goat anti-mouse-IRDye-800 (Li-Cor), both diluted 1:1000 in Odyssey blocking buffer + 0.2% Tween-20) and incubated for 1 h at room temperature with gentle shaking. The antibody mix was removed, and the wells were washed 5 times for 5 min each in PBS + 0.1% Tween-20, then 1 time with doubly distilled H2O. The plates were scanned on an Odyssey machine (Li-Cor) with a 3 mm focus offset at an intensity of 8 in both channels, and the data were analyzed using the Odyssey software. Cell Proliferation Assays and GI50 Determinations in SKOV3 and MCF-7 Cells. SKOV3 cells or MCF-7 cells were plated in SKOV3 proliferation assay medium (DMEM supplemented with 5− 10% FBS and pen/strep) at a density of 5000 cells in 180 μL volume per well in Costar no. 3610 white 96-well clear flat-bottom plates and incubated overnight in a humidified 37 °C incubator. A standard curve ranging from 10 000 to 50 000 cells was set up in a separate plate and allowed to adhere to the plate for 4−6 h, at which time the plate was developed using Cell Titer-Glo (Promega, Madison, WI) according to the manufacturer’s instructions. The next morning, 3-fold compound dilutions ranging from 10 000 to 40 nM were prepared in proliferation medium containing 1% DMSO. Then 20 μL of each dilution was added to the SKOV3 or MCF-7 cells plated the previous day, resulting in a dose−response

curve from 1000 to 4 nM. The cells were incubated for 96 h and then developed with Cell Titer Glo. The cell numbers at the end of the assay were determined using the standard curve generated at the start of the assay. Growth inhibition was calculated using the following formulas, and GI50 values were determined by plotting the percent growth inhibition vs log of the compound concentration in GraphPad:

% growth = 100 ×

T − T0 C − T0

where T is the cell number at end of assay, T0 is the cell number at start of assay (5000), C is the mumber of cells in DMSO controls at end of assay, and

% growth inhibition = 100 − % growth Pharmacokinetic Study of 3. Fasted male C57BL/6 mice were dosed intraperitoneally (3 animals/time point) at 10 mg/kg using a dosing formulation consisting of 15% Solutol HS 15/5% DMSO/80% PBS. Blood samples (300 μL) were collected via terminal cardiac puncture and placed into chilled tubes containing EDTA as the anticoagulant. The samples were centrifuged at 4 °C at 13 000 rpm for 5 min, and plasma was collected and stored frozen until analysis. 3 was extracted from the plasma samples by protein precipitation, and the plasma concentration of 3 was assessed by LC−MS/MS using an API 4000 QTrap mass spectrometer. The bioanalytical method details are shown in Table S2, and the pharmacokinetic data are shown in Table S3 in Supporting Information. The calibration curve of 3 was constructed using standards ranging from 1 to 5000 ng/mL. The regression analysis of 3 was performed by plotting the peak area ratio of 3 over internal standard (y) against the 3 concentration (x) in ng/mL. Pharmacokinetic analysis was conducted using Analyst software (version 1.4.2; Life Technologies; Carlsbad, CA) for LC−MS/MS data management and WinNonlin (version 5.2; Pharsight Corporation; Mountain View, CA) for PK analysis. Mean plasma concentration values for each time point were used to generate time to achieve peak plasma concentration [Tmax (h)], plasma terminal half-life [t1/2 (h)], peak plasma concentration [Cmax (ng/mL)], and the area under the plasma concentration−time curve [AUC ((h·ng/mL)]. In Vivo Pharmacodynamic Evaluation of the PI3Kα Covalent Inhibitor. Female nu/nu (n = 3/group) were administered compound (GDC-0941 or 3) delivered ip at 100 mg/kg once daily for 5 consecutive days. After delivery of the last dose, spleens from treated animals were harvested at 1, 4, and 24 h time points. Spleens were immediately frozen in liquid nitrogen. Samples were stored at −80 °C until processing for homogenates. Homogenates were made by adding approximately 100 μL of spleen sample to a Precellys homogenizing tube (Precellys, Montigny, France) containing 300 μL of cell extraction buffer (Biosource, Camarillo, CA) plus Complete protease inhibitor and PhosStop phosphatase inhibitor (Roche, Indianapolis, IN) and kept on ice. The sample was homogenized in a Precellys 24 homogenizer for 15 s followed by centrifugation at 16000g for 20 min at 4 °C. The supernatant was moved to a new tube, and the protein concentration was determined by BCA Assay (Pierce Biotechnology, Rockford, IL). An amount of 150 μg of protein sample was added to a 0.2 mL tube and the volume brought up to 100 μL with IP buffer from the protein A/G plate IP kit (Pierce Biotechnology, Rockford, IL). The covalent probe compound, 8, was added at 1 μM, and the tube was incubated at room temperature with rocking for 1 h. Protein A/G coated wells from the protein A/G plate IP kit were washed 3× with 200 μL of IP buffer before coating with 4 μL of rabbit anti-p110α antibody (no. 4249, Cell Signaling Technology, Danvers, MA) plus 36 μL of IP buffer per well. After incubating at room temperature with shaking for 1 h, the wells were washed 5× with 200 μL of IP buffer and the protein samples, preincubated with 8, were added to the wells. The wells were incubated overnight at 4 °C with shaking. The wells were then washed 5× with 200 μL of IP buffer. The immuoprecipitate was eluted from the plate with 40 μL of Pierce elution buffer (Pierce Biotechnology, Rockford, IL), and the eluate was moved to a 1.5 mL tube containing 4 719

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

■

μL of Pierce neutralization buffer (Pierce Biotechnology, Rockford, IL). The proteins were separated by electrophoresis and immunoblotted as described above. p110α was detected using an anti-p110α antibody (Epitomics, Burlingame, CA) diluted 1:2500 in PBS/ Odyssey buffer (1:1) + 0.1% Tween-20. The blot was subsequently incubated with a fluorescently labeled secondary antibody, goat antirabbit-IRDye800 (Li-Cor) diluted 1:10000 in PBS/Odyssey buffer (1:1) + 0.1% Tween-20, and streptavidin-AlexaFluor-680 (Invitrogen) diluted 1:1000 to detect the biotinylated probe, The streptavidin signal, detecting covalent probe binding, was normalized to total p110α signal within samples, then expressed as a percentage of the vehicle treated samples. Kinase Selectivity Panel. 3 was tested in a panel of 10 lipid kinases at the Reaction Biology Corporation (Malvern, PA). Briefly, compound was tested in a 10-concentration IC50 curve with 3-fold serial dilution starting at 1 μM. Reactions were carried out at 10 μM ATP. An HTRF assay format was used for PI3Kα, PI3Kβ, PI3Kγ, and PI3Kδ; ADP-GLO assay format was used for other kinases. The following substrates were used: for HTRF, phosphatidylinositol 4,5bisphosphate; for SPHK1 and SPHK2, sphingosine; for other ADPGLO enzymes, phosphatidylinositol. For general kinase selectivity, 3 was run in a kinase selectivity panel at Reaction Biology Corporation (Malvern, PA) using HotSpot technology and radioisotope-based P81 filtration. 3 was dissolved in pure DMSO to the final 1 μM test concentration. Substrates for the various kinases tested against 3 were prepared fresh daily in reaction buffer. Any required cofactors were then added to the substrate solution followed by kinase addition and preincubated for 30 min at room temperature. 33P-ATP (10 μM) was delivered into the reaction mixture to initiate the reaction, and reaction continued for 2 h at room temperature. The reaction was terminated, and any unreacted phosphate was washed away using 0.1% phosphoric acid prior to detection utilizing a proprietary technology (Reaction Biology Corp.; Malvern, PA, U.S.). The study was performed in duplicate, and 10 μM staurosporine, a nonselective, ATP-competitive kinase inhibitor, was used as the positive control. Molecular Modeling of 3. The binding mode of 3 to PI3Kα was obtained through molecular modeling to the X-ray structure of PI3Kα (p110α/p85α complex, PDB code 2RD0). We initially superimposed the complex structures of PI3Kγ (GDC-0941, PDB code 3DBS) and PI3Kα, then docked GDC-0941 into the PI3Kα protein. The key hinge binding interactions of the thienopyrimidine scaffold between PI3Kα and PI3Kγ are almost identical. We identified Cys862 as being unique to PI3Kα and determined it to be approximately 10 Å from the thienopyrimidine of GDC-0941. Through an iterative structure-based drug design process where we explored variations in linker length and electrophilic functionality, we identified 3. The covalent bond between the thiol (Cys862) and 3 was formed in silico through a local minimization using CHARMM force field in Discovery Studio (Discovery Studio Modeling Environment, release 2.1; Accelrys Software Inc.: San Diego, CA, 2008).

■

Article

AUTHOR INFORMATION

Corresponding Author

*Phone: 781-541-3700. Fax: 781-541-5101. E-mail: mnacht@ celgene.com. Author Contributions

All authors contributed equally. Notes

The authors declare the following competing financial interest(s): The authors are employees of Celgene (except D. Bhavsar, M. P. Sheets, W. Westlin). All authors except for D. Bhavsar, M. P. Sheets, and W. Westlin are shareholders in Celgene.

■

ACKNOWLEDGMENTS We thank Proteros Biostructures for assistance in solving the complex of PI3Kα with compound 3.

■

ABBREVIATIONS USED PI3K, phosphoinositol 3-kinase; PTEN, phosphate and tensin homologue; GSH, glutathione; ip, intraperitoneally

■

REFERENCES

(1) Engelman, J. A. Targeting PI3K signalling in cancer: opportunities, challenges and limitations. Nat. Rev. Cancer 2009, 9, 550−562. (2) Liu, P.; Cheng, H.; Roberts, T. M.; Zhao, J. J. Targeting the phosphoinositide 3-kinase pathway in cancer. Nat. Rev. Drug Discovery 2009, 8, 627−644. (3) Stambolic, V.; Suzuki, A.; de la Pompa, J. L.; Brothers, G. M.; Mirtsos, C.; Sasaki, T.; Ruland, J.; Penninger, J. M.; Siderovski, D. P.; Mak, T. W. Negative regulation of PKB/Akt-dependent cell survival by the tumor suppressor PTEN. Cell 1998, 95, 29−39. (4) Vanhaesebroeck, B.; Waterfield, M. D. Signaling by distinct classes of phosphoinositide 3-kinases. Exp. Cell Res. 1999, 253, 239− 254. (5) Engelman, J. A.; Janne, P. A.; Mermel, C.; Pearlberg, J.; Mukohara, T.; Fleet, C.; Cichowski, K.; Johnson, B. E.; Cantley, L. C. ErbB-3 mediates phosphoinositide 3-kinase activity in gefitinibsensitive non-small cell lung cancer cell lines. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 3788−3793. (6) Stommel, J. M.; Kimmelman, A. C.; Ying, H.; Nabioullin, R.; Ponugoti, A. H.; Wiedemeyer, R.; Stegh, A. H.; Bradner, J. E.; Ligon, K. L.; Brennan, C.; Chin, L.; DePinho, R. A. Coactivation of receptor tyrosine kinases affects the response of tumor cells to targeted therapies. Science 2007, 318, 287−290. (7) Mellinghoff, I. K.; Wang, M. Y.; Vivanco, I.; Haas-Kogan, D. A.; Zhu, S.; Dia, E. Q.; Lu, K. V.; Yoshimoto, K.; Huang, J. H.; Chute, D. J.; Riggs, B. L.; Horvath, S.; Liau, L. M.; Cavenee, W. K.; Rao, P. N.; Beroukhim, R.; Peck, T. C.; Lee, J. C.; Sellers, W. R.; Stokoe, D.; Prados, M.; Cloughesy, T. F.; Sawyers, C. L.; Mischel, P. S. Molecular determinants of the response of glioblastomas to EGFR kinase inhibitors. N. Engl. J. Med. 2005, 353, 2012−2024. (8) Cully, M.; You, H.; Levine, A. J.; Mak, T. W. Beyond PTEN mutations: the PI3K pathway as an integrator of multiple inputs during tumorigenesis. Nat. Rev. Cancer 2006, 6, 184−192. (9) Zhao, L.; Vogt, P. K. Class I PI3K in oncogenic cellular transformation. Oncogene 2008, 27, 5486−5496. (10) Karakas, B.; Bachman, K. E.; Park, B. H. Mutation of the PIK3CA oncogene in human cancers. Br. J. Cancer 2006, 94, 455−459. (11) Kang, S.; Bader, A. G.; Vogt, P. K. Phosphatidylinositol 3-kinase mutations identified in human cancer are oncogenic. Proc. Natl. Acad. Sci. U.S.A 2005, 102, 802−807. (12) Shuttleworth, S. J.; Silva, F. A.; Cecil, A. R.; Tomassi, C. D.; Hill, T. J.; Raynaud, F. I.; Clarke, P. A.; Workman, P. Progress in the preclinical discovery and clinical development of class I and dual class

ASSOCIATED CONTENT

S Supporting Information *

Chemical structure of the covalent probe 8; mass spectrometry on 3 with PI3K family members; specificity of 3 toward PI3Kα vs other kinases; lack of reactivity of 3 to nonspecific thiols including glutathione and plasma proteins; inhibition of PI3Kα signaling in SKOV3 cells by 3; specificity of the covalent probe 8 for PI3Kα vs PI3Kβ in vitro; X-ray crystallography parameters of 3 complexed with PI3Kα. This material is available free of charge via the Internet at http://pubs.acs.org. Accession Codes

The structure of 3 complexed with PI3Kα has been deposited in the Protein Data Bank, and the PDB code is 3zim. 720

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721

Journal of Medicinal Chemistry

Article

I/IV phosphoinositide 3-kinase (PI3K) inhibitors. Curr. Med. Chem. 2011, 18, 2686−2714. (13) Jackson, S. P.; Schoenwaelder, S. M.; Goncalves, I.; Nesbitt, W. S.; Yap, C. L.; Wright, C. E.; Kenche, V.; Anderson, K. E.; Dopheide, S. M.; Yuan, Y.; Sturgeon, S. A.; Prabaharan, H.; Thompson, P. E.; Smith, G. D.; Shepherd, P. R.; Daniele, N.; Kulkarni, S.; Abbott, B.; Saylik, D.; Jones, C.; Lu, L.; Giuliano, S.; Hughan, S. C.; Angus, J. A.; Robertson, A. D.; Salem, H. H. PI 3-kinase p110beta: a new target for antithrombotic therapy. Nat. Med. 2005, 11, 507−514. (14) Okkenhaug, K.; Vanhaesebroeck, B. PI3K in lymphocyte development, differentiation and activation. Nat. Rev. Immunol. 2003, 3, 317−330. (15) Ruckle, T.; Schwarz, M. K.; Rommel, C. PI3Kgamma inhibition: towards an “aspirin of the 21st century”? Nat. Rev. Drug Discovery 2006, 5, 903−918. (16) Singh, J.; Petter, R. C.; Baillie, T. A.; Whitty, A. The resurgence of covalent drugs. Nat. Rev. Drug Discovery 2011, 10, 307−317. (17) Singh, J.; Petter, R. C.; Kluge, A. F. Targeted covalent drugs of the kinase family. Curr. Opin. Chem. Biol. 2010, 14, 475−480. (18) Hagel, M.; Niu, D.; St, M. T.; Sheets, M. P.; Qiao, L.; Bernard, H.; Karp, R. M.; Zhu, Z.; Labenski, M. T.; Chaturvedi, P.; Nacht, M.; Westlin, W. F.; Petter, R. C.; Singh, J. Selective irreversible inhibition of a protease by targeting a noncatalytic cysteine. Nat. Chem. Biol. 2011, 7, 22−24. (19) Shuttleworth, S. J. Progress in the preclinical discovery and clinical development of class I and dual class I/IV phosphoinositide 3kinase (PI3K) inhibitors. Curr. Med. Chem. 2011, 18, 2686−2714. (20) Huang, C. H.; Mandelker, D.; Schmidt-Kittler, O.; Samuels, Y.; Velculescu, V. E.; Kinzler, K. W.; Vogelstein, B.; Gabelli, S. B.; Amzel, L. M. The structure of a human p110alpha/p85alpha complex elucidates the effects of oncogenic PI3Kalpha mutations. Science 2007, 318, 1744−1748. (21) Sutherlin, D. P.; Sampath, D.; Berry, M.; Castanedo, G.; Chang, Z.; Chuckowree, I.; Dotson, J.; Folkes, A.; Friedman, L.; Goldsmith, R.; Heffron, T.; Lee, L.; Lesnick, J.; Lewis, C.; Mathieu, S.; Nonomiya, J.; Olivero, A.; Pang, J.; Prior, W. W.; Salphati, L.; Sideris, S.; Tian, Q.; Tsui, V.; Wan, N. C.; Wang, S.; Wiesmann, C.; Wong, S.; Zhu, B. Y. Discovery of (thienopyrimidin-2-yl)aminopyrimidines as potent, selective, and orally available pan-PI3-kinase and dual pan-PI3kinase/mTOR inhibitors for the treatment of cancer. J. Med. Chem. 2010, 53, 1086−1097. (22) Folkes, A. J.; Ahmadi, K.; Alderton, W. K.; Alix, S.; Baker, S. J.; Box, G.; Chuckowree, I. S.; Clarke, P. A.; Depledge, P.; Eccles, S. A.; Friedman, L. S.; Hayes, A.; Hancox, T. C.; Kugendradas, A.; Lensun, L.; Moore, P.; Olivero, A. G.; Pang, J.; Patel, S.; Pergl-Wilson, G. H.; Raynaud, F. I.; Robson, A.; Saghir, N.; Salphati, L.; Sohal, S.; Ultsch, M. H.; Valenti, M.; Wallweber, H. J.; Wan, N. C.; Wiesmann, C.; Workman, P.; Zhyvoloup, A.; Zvelebil, M. J.; Shuttleworth, S. J. The identification of 2-(1H-indazol-4-yl)-6-(4-methanesulfonyl-piperazin1-ylmethyl)-4-morpholin-4-yl-t hieno[3,2-d]pyrimidine (GDC-0941) as a potent, selective, orally bioavailable inhibitor of class I PI3 kinase for the treatment of cancer. J. Med. Chem. 2008, 51, 5522−5532. (23) Ikediobi, O. N.; Davies, H.; Bignell, G.; Edkins, S.; Stevens, C.; O’Meara, S.; Santarius, T.; Avis, T.; Barthorpe, S.; Brackenbury, L.; Buck, G.; Butler, A.; Clements, J.; Cole, J.; Dicks, E.; Forbes, S.; Gray, K.; Halliday, K.; Harrison, R.; Hills, K.; Hinton, J.; Hunter, C.; Jenkinson, A.; Jones, D.; Kosmidou, V.; Lugg, R.; Menzies, A.; Mironenko, T.; Parker, A.; Perry, J.; Raine, K.; Richardson, D.; Shepherd, R.; Small, A.; Smith, R.; Solomon, H.; Stephens, P.; Teague, J.; Tofts, C.; Varian, J.; Webb, T.; West, S.; Widaa, S.; Yates, A.; Reinhold, W.; Weinstein, J. N.; Stratton, M. R.; Futreal, P. A.; Wooster, R. Mutation analysis of 24 known cancer genes in the NCI60 cell line set. Mol. Cancer Ther. 2006, 5, 2606−2612. (24) Krenske, E. H.; Petter, R. C.; Zhu, Z.; Houk, K. N. Transition states and energetics of nucleophilic additions of thiols to substituted alpha,beta-unsaturated ketones: substituent effects involve enone stabilization, product branching, and solvation. J. Org. Chem. 2011, 76, 5074−5081.

721

dx.doi.org/10.1021/jm3008745 | J. Med. Chem. 2013, 56, 712−721