Design, synthesis, and structure-activity relationships of 3-ethynyl-1H-indazoles as inhibitors of Phosphatidylinositol 3-kinase signaling pathway

Abstract

A new series of 3-ethynyl-1H–indazoles has been synthesized and evaluated in both biochemical and cell-based assays as potential kinase inhibitors. Interestingly, a selected group of compounds identified from this series exhibited low micromolar inhibition against critical components of the PI3K pathway, targeting PI3K, PDK1 and mTOR kinases. Combination of computational modeling and structure-activity relationships studies reveal a possible novel mode for PI3K inhibition, resulting in a PI3Kα isoform specific compound. Hence, by targeting the most oncogenic mutant isoform of PI3K, the compound displays anti-proliferative activity both in monolayer human cancer cell cultures and in three-dimensional tumor models. Because of its favorable physicochemical, in vitro ADME and drug-like properties, we propose that this novel ATP mimetic scaffold could result useful in deriving novel selecting and multi-kinase inhibitors for clinical use.

Introduction

The PI3K/AKT/mTOR cascade is an important cellular signaling pathway that regulates intersecting biological processes such as cell growth and proliferation, cell survival, protein synthesis, and glycolysis metabolism.^1–3 Physiological stimulation of this cascade occurs through the binding of growth factors (e.g., insulin or insulin-like growth factor; IGF-1) to receptor tyrosine kinases (RTK) on the cell surface. Subsequent activation of the lipid phosphatidylinositol 3-kinase (PI3K) leads to the phosphorylation of phosphatidylinositol-4,5-bisphosphate (PIP₂) to produce phosphatidylinositol-3,4,5-triphosphate (PIP₃), which in turn interacts with the pleckstrin homology (PH) domain of AKT and recruits the kinase to the plasma membrane. Full activation of AKT requires phosphorylation of two distinct residues: Thr308 by the upstream kinase PDK1 and Ser473 by the mammalian target of rapamycin complex 2 (mTORC2).^1–3 Activated AKT, among its wide array of downstream effects, increases protein synthesis rate by phosphorylation at Thr246 of the proline-rich substrate of 40 kDa (PRAS40). This substrate is further regulated through phosphorylation at Ser183 by the mTOR complex 1 (mTORC1). Downstream of mTORC1, signaling events can go through p70 S6 kinase (S6K1), which in turn phosphorylates the programmed cell death protein (PDCD4) at residue Ser457 (Figure 1).

Figure 1.

Schematic representation of the PI3K/mTOR signaling pathway highlighting the phosphorylation sites tested in our cellular and biochemical assays.

Because of its involvement in a variety of cellular events, such as survival, proliferation, cell motility and invasion,⁴ hyperactivation of the PI3K signaling pathway is one of the most common molecular events in nearly every form of human cancer.^5–8 Indeed, aberrant activation of the PI3K cascade, including PTEN inactivation and PI3K activating mutations, are present in about 32 % of colon cancers, 30 % of breast cancers, 30% of melanomas,⁹ 27 % of brain cancers, and 25 % of stomach cancers.^{10, 11} It is for these reasons that considerable drug discovery efforts are ongoing targeting several components of this signaling cascade, using either single and multiple target strategies.^12–16 Interestingly, clinical data suggest that multikinase inhibitors (MKIs) produce greater benefit over single kinase inhibition, especially in solid tumors where different kinases and different pathways synergistically contribute to tumor proliferation.^17–19 Indeed, emerging Phase II clinical trials data reveal that dual PI3K/mTOR inhibitors may be more effective having the advantage of being less susceptible to PI3K drug resistance owing their preserved activity against mTOR.¹⁶

In our drug discovery program we interrogated the PI3K signaling pathway and we report on the identification of a novel scaffold of the 3-ethynylindazole family as multiple PI3K/PDK1/mTOR inhibitor. Such scaffold represents a novel lead structure for the PI3K pathway, suitable for further functionalization and drug development.

Results and Discussion

Compounds 6–19 (Table 1) were prepared as shown in Scheme 1. Compound 1 was iodinated and Boc protection was performed according to the previously reported procedures.²⁰ Compound 3 (Scheme 1) was coupled with the appropriate alkyne using the Sonogashira reaction conditions to give 4 and its analogs in moderate yield (45–65%). Compound 4 was treated with TFA to afford the final compound 5 and its analogs (6–19, Table 1) in good yields (89–95%). Similarly, compounds 20–21 (Table 1) were synthesized starting from 7-aza-indazole instead of indazole and following the same synthetic procedure reported in Scheme 1 for indazole derivatives.

Table 1.

Kinase and cell-based assay results for compounds 6–21^a

compd ID	R	X	Kinase assay IC50 (µM)			Cell-based Assayd IC50 (µM)		Cell-based Assay GI50 (µM)
			PI3Kαb	mTORb	PDK1c	AKT [pT308]	PRAS40 [pS183]	PC3	MDA- MB-231	MCF7	Hela
6		CH	1.05	4.97	39.20	17.09	10.80	2.51	8.28	3.23	3.60
7		CH	>100	>100	>50	>100	>50	ND	ND	ND	ND
8		CH	>100	>100	>100	>100	>200	ND	ND	ND	ND
9		CH	1.85	>100	20.80	23.5	>50	6.81	10.48	3.66	5.21
10		CH	0.36	10.70	3.87	12.85	13.60	2.43	7.84	2.58	2.53
11		CH	>50	>100	>100	>100	>50	ND	ND	ND	ND
12		CH	>100	>100	>25	>100	>50	ND	ND	ND	ND
13		CH	5.12	>100	12.60	12.90	5.10	2.84	8.29	10.39	5.21
14		CH	>100	>100	>100	>50	>50	ND	ND	ND	ND
15		CH	>100	>100	>100	>50	>50	ND	ND	ND	ND
16		CH	>100	>100	>100	>100	>50	ND	ND	ND	ND
17		CH	>100	>100	>100	>100	>100	ND	ND	ND	ND
18		CH	>100	>100	>100	>100	>100	ND	ND	ND	ND
19		CH	>100	>100	>100	>50	>50	ND	ND	ND	ND
20		N	3.05	>10	>50	ND	19.1	ND	ND	ND	ND
21		N	9.0	5.0	>50	ND	>10	ND	ND	ND	ND

^aValues are means of at least three or more experiments with a typical standard deviation of less than ± 20%.

^bLanthaScreen Eu kinase binding assay.

^cZ’-LYTE kinase activity assay.

^dLanthaScreen cellular assays in HEK293E cell line; AKT [pThr308] for PI3K/PDK1 readout and PRAS40 [pSer183] for mTORC1 readout. ND: not determined (compound was not tested).

Scheme 1. Synthetic scheme for compounds 6–21.

R = aryl or alkyl, see Table 1

X= CH (1-19) or N (20-21)

Reagents and conditions. (a) Iodine, KOH, DMF, rt; (b) di-tert-butyl dicarbonate, Et₃N, DMAP, CH₃CN, rt; (c) aryl or aliphatic alkyne, Pd(Ph₃P)₂Cl₂, CuI, Et₃N, CH₃CN, rt; (d) TFA, CH₂Cl₂, rt.

We chose to initially screen the synthesized compounds (shown in Table 1) in a cell-based format using a LanthaScreen cellular assay technology, recently reported and successfully applied to the analysis of the PI3K/AKT/mTOR pathway.^21–23 This assay platform uses individual cell lines that stably express key kinase substrate within the cascade as GFP-fusion proteins. Cells are stimulated, in the presence or absence of an inhibitor, to trigger endogenous kinase activity and phosphorylation of the substrate. The phosphorylation status of the GFP-substrate is detected after cell lysis by addition of a terbium (Tb)-labeled phosphospecific antibody, via time-resolved Förster resonance energy transfer (TR-FRET) signal between Tb and GFP.²¹ We used three phosphorylation readouts to interrogate pathway signaling: AKT [pThr308] for PI3K/PDK1, PRAS40 [pThr246] for AKT, and PRAS40 [pSer183] for mTORC1 activity (Figure 1).

Compounds were incubated with the different cell lines in 10-point dose-response format for 60 minutes to generate IC₅₀ curves. Among all newly synthesized series of 3-ethynyl-1H–indazoles (6 to 19), we found that compounds 6, 10, 13 inhibited both AKT and PRAS40 phosphorylation, while compound 9 inhibited only AKT phosphorylation, with potencies in the low micromolar range (Table 1).

To understand if the inhibition of AKT phosphorylation at Thr308 was a result of a direct inhibition of PDK1 or PI3K, we performed a cell-free kinase assay against both kinases. All the compounds which displayed inhibition in cell based assays (6, 9, 10, 13) resulted to be active against both kinases, with the highest potency for 10 (IC₅₀ = 361 nM) against PI3Kα. Similarly, we tested the compounds against mTOR through a biochemical assay. Compounds 6 and 10 are multiple PI3Kα/PDK1/mTOR inhibitors, while compounds 9 and 13 are dual PI3Kα /PDK1 inhibitors (Table 1). All compounds tested were inactive against AKT as measured by a biochemical kinase assay for AKT1 (data not shown). Further in vitro displacement assays against the PI3K isoforms revealed that compounds 6, 10 and 13 are α-isoform specific and ATP competitive with about 100-fold selectivity over the β and γ isoforms (Table 2).

Table 2.

Biochemical Selectivity Profile against PI-3K isoforms, IC₅₀ (µM)

Cmpound ID	PI-3Kαa	PI-3Kγa	PI-3Kδa
6	1.05	>100	>100
10	0.323	40	39
13	5.12	>100	>100

^aThese assays were conducted by Invitrogen’s SelectScreen Biochemical Kinase (SSBK) Profiling Service. ^a Values are means of at least three or more experiments with a typical standard deviation of less than ± 20%.

In order to rationalize the affinity observed for the lead structure 10, we studied its binding mode within the ATP binding site of PI3Kα by using computational docking studies. A target binding pocket was derived from the X-ray crystal structure of the ternary complex involving the most common mutant of PI3K catalytic domain p110α (H1047R), its regulatory subunit (p85α) and the drug wortmannin (PDB id: 3hhm). From the docked binding pose, compound 10 appears to be deeply inserted in the ATP-binding site (Figure 2A): the N-2 atom of compound 10 seems to be involved in hydrogen bonding interactions with the OH of Tyr836 and the indazole moiety further stabilized via hydrogen bond contacts between the NH-1 and the N-2 of the heterocycle with the backbone carbonyl of Asp810 and the backbone NH of Asp933, respectively (Figure 2B). Furthermore, the compound binds to the hinge region of the kinase via a hydrogen bond between the protonated NH of the pyridine ring and the backbone carbonyl of Val851 (Figure 2). From these studies it became apparent that compounds of this series capable of interacting with the hinge region are most effective against PI3Kα. Indeed, the isomer 9, whose protonated nitrogen could still constitutes a suitable hydrogen bonding donor group interacting with the hinge residue Val851, showed good potency (IC₅₀ = 1.85 µM); in compounds 6 and 13, where the ethynyl-indazole moiety is linked to an aniline, is the amino group at the 3- or 4-position on the phenyl ring to be involved in the hydrogen bond with the backbone carbonyl of Val851. Indeed, they both showed a low micromolar inhibition (IC₅₀ values of 1.05 µM and 5.12 µM respectively).

Figure 2.

Docking studies of compound 10 within the ATP-binding site of PI3Kα (H1047R) mutant (PDB id: 3hhm).²⁴ (A) Surface representation of the active site of p110α with the ATP mimic compound 10. Surface generated with MOLCAD³⁷ and color coded according to cavity depth (blue, shallow; yellow, deep). Protein residues involved in hydrogen bonds are highlighted. (B) Ribbon representation of the PI3Kα (H1047R) mutant in complex with compound 10 in the same binding pose of A). Protein residues involved in hydrogen bonds showed as capped sticks.

The shift of the amino group at the 2-position on the phenyl ring (16) doesn’t allow the compound to form the critical interaction with the hinge region, thus explaining its loss of activity (IC₅₀ > 100 µM). As expected, the simple phenyl (7), fluorophenyl (8), trifluoromethylphenyl (17), thiophenyl (18), 4-cyanophenyl (19) substitutions on 3-ethynylindazole showed poor activity. Accordingly, aliphatic alkynes such as cyclopentyl (14), cyclopropyl (15), and N,N-dimethyl (11) were inactive both in the biochemical and cell-based assays (Table 1).

Keeping the pyridine group as the most efficient substituent, we explored possible modifications on the indazole ring (20–21). Isosteric replacement of the indazole with the pyrazolopyridine ring revealed to be detrimental for PI3Kα inhibition. Indeed, by comparing the most active compound of the 3-ethynylindazole series (10, IC₅₀ = 361 nM) with the corresponding pyrazolopyridine analog (20, IC₅₀ = 3.05 µM) a 10-fold loss of activity was observed and the same trend was followed by compound 9 (IC₅₀ = 1.85 µM) versus compound 21 (IC₅₀ = 9.0 µM). Such decrease in activity is expected given that the additional nitrogen atom is not engaged in hydrogen bonding interactions in the docked structures of compounds 20 and 21 in the binding pocket of PI3Kα (Supporting Information).

Since the ATP binding pocket of mTOR is highly homologous to that of PI3K, it is not surprising that some of the compounds inhibit both kinases (Table 1). Indeed, even previously reported selective PI3K inhibitors show some degree of activity against mTOR.¹⁶ However, noteworthy compound 10 in our cellular and biochemical assays inhibits also the less structurally related PDK-1. Inhibition of multiple kinases on the PI3K pathway may be highly desirable to suppress oncogenic transformations that result from the activation of the pathway in cancer cells.

To evaluate the drug-likeness of the proposed series on empirical ground^{25, 26}, we gathered information about physicochemical properties of the compounds, such as solubility, cell permeability, ligand efficiency indices,^27–29 as well as plasma and microsomal stability (Table 3 and Supporting Information). Hence, investigation of these parameters suggests that the compounds of this novel scaffold series may be deemed suitable as lead-candidates for further drug optimizations.³⁰

Table 3.

Ligand efficiency indices and in vitro ADME properties of 10 and 13

Parameter		10	13
Ligand Efficiency	LE (Kcal mol−1)	0.5	0.4
Binding Efficiency Index	BEI	29.4	22.7
Surface Binding Efficiency Index	SEI	17.5	10.5
Lipophilic Ligand Efficiency	LLE	4.2	2.5
Microsomal stability	t ½ (min)	35	>60
Plasma stability	% remaining after 60 min	97	100

LE: Ligand efficiency,²⁸ BEI: binding efficiency index,²⁷ SEI: surface efficiency index, ²⁷ LLE: lipophilic ligand efficiency indices.²⁷–²⁹

On the basis of their cellular efficacy and favorable biopharmaceutical properties, we further investigated the anti-proliferative activity of the most promising compounds (6, 9, 10, and 13) against human cancer cells lines including prostatic (PC3), breast (MD-MA-231 and MCF7), and cervical (HeLa) cancers. The results are summarized in Table 1. The anti-tumor activity of compound 10, which displayed the highest cellular potency inhibiting PI3Kα, mTOR and PDK1, has been further evaluated against growth/proliferation of the glioblastoma U87 cell line when tested in three-dimensional cultures. The three-dimensional cell culture system has been chosen because it better recapitulates real human tissue with respect to oxygen and nutrient levels, cell-cell contacts and cellular architecture compared with conventional two-dimensional cell culture systems.³¹ It has also been observed that spheroid cultures exhibit differential sensitivity to known chemotherapeutic agents, as reported for the wortmannin derivative PX-866 (22)³² which showed higher potency in multicellular 3D system then in monolayer.³³ We chose U87 glioblastoma cells because they exhibit deregulated PI3K signaling due to the functional loss of the tumor suppressor PTEN (Figure 1). U87 spheroids, generated through the previously reported hanging drop method,³⁴ were treated daily, for a total of 6 days, with compound 10 and 22 as control at different concentrations (Figure 3). Treatment with 10 µM of compound 10 every day strongly suppressed cell growth/proliferation, as proved by detected spheroid volumes, comparable to those produced by the potent PI3K inhibitor 22 (Figure 3).

Figure 3.

U87 cells grown as spheroids generated by the hanging drop method³⁴ and treated with the indicated concentration of 22 or compound 10 every day, for a total of 6 days. Left panel: Phase-contrast photos representative of four independent experiments. Right panel: spheroid length and width (measured with an optical micrometer) were used to calculate spheroid volumes (µm³).

In summary, we successfully developed a useful synthetic route to obtain 3-ethynylindazole derivatives which lead to the identification of a multiple PI3K/PDK1/mTOR inhibitor, with a nanomolar inhibition against PI3Kα. Other indazoles were recently reported as possible inhibitors of other protein kinases,³⁵ however the scaffolds reported here, in particular compound 10, represent a novel and valuable starting point for the development of inhibitors of the PI3K pathway and potentially other kinases (see below). Moreover, due to the observed drug-likeness and in vitro ADME properties of this novel scaffold (Table 3), we anticipate that further selective and/or multi-kinase inhibitors could arise from further derivatization of this novel ATP mimetics. In fact, preliminary selectivity panels with compound 10 against 314 protein kinases (Supporting Information) reveals that less than 10% of these proteins showed ≥ 50% inhibition at 10 µM (Table 4). Hence, based on these data, we envision that further elaborations of compound 10 could lead to further selective or multi-kinase inhibitors against a variety of drug targets.

Table 4.

Kinases which showed an inhibition higher than 50 % at 10µM concentration of compound 10.

KINASES TESTED
ABL1 T3151I	MARK4
CDK5/p25	MELK
CHUK (IKK alpha)	MINK1
CLK2	MUSK
DNA-PK	MYLK2 (skMLCK)
FLT3	NTRK1 (TRKA)
FLT3 D835Y	NTRK2 (TRKB)
FLT4 (VEGFR3)	NTRK3 (TRKC)
GSG2 (Haspin)	NUAK1 (ARK5)
IRAK1	PDGFRA T674I
JNK1 (MAPK8)	PDGFRA V561D
JNK2 (MAPK9)	RET V804L
KDR (VEGFR2)	RET Y791F
LRKK2	SGK (SGK1)
LRRK2 G2019S	SYK
MAP4K4 (HGK)

Experimental Section

Chemistry

Unless otherwise indicated, all anhydrous solvents were commercially obtained and stored in Sure-seal bottles under nitrogen. All other reagents and solvents were purchased as the highest grade available and used without further purification. Thin-layer chromatography (TLC) analysis of reaction mixtures was performed using Merck silica gel 60 F254 TLC plates, and visualized using ultraviolet light. NMR spectra were recorded on Varian 300 or 500 MHz instruments. Chemical shifts (δ) are reported in parts per million (ppm) referenced to ¹H (Me₄Si at 0.00). Coupling constants (J) are reported in Hz throughout. Mass spectral data were acquired on Shimadzu LCMS-2010EV for low resolution, and on an Agilent ESI-TOF for either high or low resolution. Purity of all compounds was obtained in a HPLC Breeze from Waters Co. using an Atlantis T3 3µm 4.6×150 mm reverse phase column. The eluant was a linear gradient with a flow rate of 1 ml/min from 95% A and 5% B to 5% A and 95% B in 15 min followed by 5 min at 100% B (Solvent A: H₂O with 0.1% TFA; Solvent B: acetonitrile with 0.1% TFA). The compounds were detected at λ=254 nm. Purity of key compounds was established by elemental analysis as performed on a Perkin Elmer series II-2400 and HPLC analysis and determined to be > 95%. Combustion analysis was performed by NuMega Resonance Labs, San Diego, CA, USA.

Synthesis of tert-butyl 3-(phenylethynyl)-1H-indazole-1-carboxylate (4, R = Ph)

To a solution of compound 3 (343 mg, 1 mmol) in CH₃CN (5 mL) were added phenyl acetylene (0.13 mL, 1.2 mmol), CuI (20 mg, 0.10 mmol), Pd(Ph₃P)₂Cl₂ (70 mg, 0.10 mmol), and Et₃N (0.42 mL, 3 mmol) respectively at room temperature under nitrogen atmosphere. The reaction mixture was stirred at room temperature for 16 h then solvent was removed in vacuo. The residue was chromatographed over silica gel (2% ethyl acetate in hexane) to give a Boc-protected product 4 (270 mg, 85%).

Synthesis of 3-(phenylethynyl)-1H-indazole (7)

To a solution of 4 (150 mg, 0.47 mmol) in CH₂Cl₂ (3 mL) was added TFA (1 mL) at room temperature. The resulting mixture was stirred at room temperature for 3 h then solvent and TFA were removed in vacuo. The residue was extracted with CH₂Cl₂ (50 mL), washed with saturated NaHCO₃ aqueous solution (3 × 30 mL), brine (30 mL), dried (MgSO₄) and concentrated in vacuo. The residue was chromatographed over silica gel (40% ethyl acetate in hexane) to afford a white solid compound 7 (94 mg, 92%). ¹H NMR (300 MHz, CDCl₃) δ 7.26–7.34 (m, 1H), 7.40–7.50 (m, 5H), 7.66–7.76 (m, 2H), 7.95 (d, J = 7.8 Hz, 1H), 11.68 (br s, 1 H, NH); MS m/z 219 (M+H)⁺, 166, 155, 121, 88; HRMS calcd for C₁₅H₁₁N₂ 219.0917 (M+H), found 219.0918.

Following the above mentioned procedure and the use of the appropriate starting materials and reagents, compounds 6 to 16 were prepared. Yields refer to the final deprotection step.

Synthesis of 4-((1H-indazol-3-yl)ethynyl)aniline (6)

Yield: 65%; ¹H NMR (300 MHz, CD₃OD) δ 5.24 (br s, 2 H, NH₂), 6.57–6.60 (m, 2 H), 7.12–7.15 (m, 1H), 7.24 (d, J = 7.5 Hz, 2 H), 7.28–7.46 (m, 2 H), 7.70 (d, J = 6.9 Hz, 1 H), 11.82 (br s, 1 H, NH); MS m/z 234 (M+H)⁺, 158, 149, 121, 102, 65; HRMS calcd for C₁₅H₁₂N₃ 234.1031 (M+H), found 234.1021.

Synthesis of 3-((3,5-difluorophenyl)ethynyl)-1H-indazole (8)

Yield: 89%; ¹H NMR (300 MHz, DMSO) δ 7.27 (t, J = 7.5 Hz, 1H), 7.36–7.51 (m, 4H), 7.63 (d, J = 8.4 Hz, 1H), 7.92 (d, J = 8.1 Hz, 1H), 13.32 (br s, 1 H); MS m/z 255 (M+H)⁺, 149, 121; HRMS calcd for C₁₅H₉F₂N₂ 255.0728 (M+H), found 255.0726.

Synthesis of 3-(pyridin-3-ylethynyl)-1H-indazole (9)

Yield: 92%; ¹H NMR (300 MHz, CDCl₃) δ 7.24–7.41 (m, 3H), 7.46 (m, 1H), 7.59 (d, J = 7.8 Hz, 1H), 7.87–7.99 (m, 2H), 8.63 (s, 1H), 11.32 (br s, 1 H, NH); MS m/z 220 (M+H)⁺, 192, 121, 102; HRMS calcd for C₁₄H₁₀N₃ 220.0875 (M+H), found 220.0877.

Synthesis of 3-(pyridin-4-ylethynyl)-1H-indazole (10)

Yield: 91%; ¹H NMR (400 MHz, DMSO) δ 7.29 (t, J = 7.3 Hz, 1 H), 7.45 (t, J = 7.3 Hz, 1 H), 7.62–7.67 (m, 3 H), 7.89 (d, J = 7.9 Hz, 1 H), 8.65 (d, J = 6.1 Hz, 2 H), 13.68 (br s, 1 H, NH); HRMS calcd for C₁₄H₁₀N₃ 220.0875 (M+H), found 220.0878. Anal. calcd for C₁₄H₉N₃: C, 76.70; H, 4.14; N, 19.17. Found: C, 76.58; H, 4.29; N, 19.01.

Synthesis of 3-(1H-indazol-3-yl)-N,N-dimethylprop-2-yn-1-amine (11)

Yield: 90%; ¹H NMR (300 MHz, CDCl₃) δ 2.47 (s, 6H), 3.67 (s, 2H), 7.18–7.27 (m, 1H), 7.40 (t, J = 7.6 Hz, 1H), 7.56 (d, J = 8.4 Hz, 1H), 7.65 (d, J = 7.8 Hz, 1H), 11.65 (br s, 1 H, NH); MS m/z 200 (M+H)⁺, 155, 127, 83; HRMS calcd for C₁₂H₁₄N₃ 200.1182 (M+H), found 200.1190.

Synthesis of 3-(pyridin-2-ylethynyl)-1H-indazole (12)

Yield: 91%; ¹H NMR (300 MHz, CD₃OD) δ 7.29 (t, J = 6.9 Hz, 1H), 7.40–7.55 (m, 2H), 7.60 (d, J = 8.1 Hz, 1H), 7.77 (d, J = 6.6 Hz, 2H), 7.93 (d, J = 6.9 Hz, 2H), 11.31 (br s, 1 H, NH); HRMS calcd for C₁₄H₁₀N₃ 220.0875 (M+H), found 220.0878.

Synthesis of 3-((1H-indazol-3-yl)ethynyl)aniline (13)

Yield: 91%; ¹H NMR (300 MHz, CD₃OD) δ 5.26 (br s, 2 H, NH₂), 6.76 (d, J = 7.8 Hz, 1 H), 6.94 (d, J = 7.5 Hz, 1 H), 6.97 (s, 1 H), 7.13 (t, J = 8.1 Hz, 1H), 7.24 (t, J = 8.1 Hz, 1 H), 7.44 (t, J = 8.1 Hz, 1 H), 7.56 (d, J = 7.8 Hz, 1 H), 7.84 (d, J = 8.7 Hz, 1 H), 11.32 (br s,1 H, NH); HRMS calcd for C₁₅H₁₂N₃ 234.1031 (M+H), found 234.1025. Anal. calcd for C₁₅H₁₁N₃: C, 77.23; H, 4.75; N, 18.01. Found: C, 77.01; H, 4.89; N, 17.89.

Synthesis of 3-(cyclopentylethynyl)-1H-indazol (14)

Yield: 92%; ¹H NMR (300 MHz, CDCl₃) δ 1.66–1.69 (m, 2H), 1.80–189 (m, 4H), 2.02–2.19 (m, 2H), 2.94–3.08 (m, 1H), 7.22 (t, J = 7.4 Hz, 1H), 7.41 (t, J = 7.4 Hz, 1H), 7.70 (d, J = 7.8 Hz, 1H), 7.82 (d, J = 7.8 Hz, 1H), 11.84 (br s, 1 H, NH); HRMS calcd for C₁₄H₁₅N₂ 211.1235 (M+H), found 211.1233.

Synthesis of 3-(cyclopropylethynyl)-1H-indazole (15)

Yield: 90%; ¹H NMR (300 MHz, CDCl₃) δ 0.92–1.00 (m, 4H), 1.56–1.67 (m, 1H), 7.17–7.27 (m, 1H), 7.36–7.46 (m, 1H), 7.66 (d, J = 8.4 Hz, 1H), 7.82 (d, J = 8.1 Hz, 1H), 9.86 (br s, 1 H, NH); HRMS calcd for C₁₂H₁₁N₂ 183.0922 (M+H), found 183.0925.

Synthesis of 2-((1H-indazol-3-yl)ethynyl)aniline (16)

Yield: 81%; ¹H NMR (300 MHz, CD₃OD) δ 5.26 (br s, 2 H, NH₂), 6.58–6.60 (m, 2 H), 7.12–7.15 (m, 1H), 7.24 (d, J = 7.5 Hz, 2 H), 7.28–7.46 (m, 2 H), 7.70 (d, J = 6.9 Hz, 1 H), 11.80 (br s, 1 H, NH); MS m/z 234 (M+H)⁺, 158, 149, 121, 102, 65; HRMS calcd for C₁₅H₁₂N₃ 234.1031 (M+H), found 234.1021.

Synthesis of 3-((3,5-bis(trifluoromethyl)phenyl)ethynyl)-1H-indazole (17)

Yield: 90%; ¹H NMR (300 MHz, CDCl₃) δ 7.32 (t, J = 8.5 Hz, 1H), 7.49 (t, J = 6.6 Hz, 1H), 7.85–7.94 (m, 2H), 8.07 (s, 2H), 9.56 (br s, 1 H, NH); MS m/z 355 (M+H)⁺, 275, 244, 149, 130, 127, 121, 118, 88; HRMS calcd for C₁₇H₉F₆N₂ 355.0664 (M+H), found 355.0664.

Synthesis of 3-(thiophen-3-ylethynyl)-1H-indazole (18)

Yield: 92%; ¹H NMR (300 MHz, CDCl₃) δ 7.22–7.46 (m, 4H), 7.63–7.70 (m, 2H), 7.90 (d, J = 8.1 Hz, 1H), 9.52 (br s , 1 H, NH); MS m/z 225 (M+H)⁺, 130, 121, 102; HRMS calcd for C₁₃H₉N₂S 225.0481 (M+H), found 225.0484.

Synthesis of 4-((1H-indazol-3-yl)ethynyl)benzonitrile (19)

Yield: 91%; ¹H NMR (300 MHz, CD₃OD) δ 7.28 (t, J = 7.5 Hz, 1H), 7.47 (t, J = 7.8 Hz, 1H), 7.59 (d, J = 8.1 Hz, 1H), 7.74–7.76 (m, 4H), 7.85 (d, J = 8.1 Hz, 1H), 10.89 (br s, 1 H, NH); MS m/z 244 (M+H)⁺, 217, 149, 141, 130, 127, 121, 102; HRMS calcd for C₁₆H₁₀N₃ 244.0869 (M+H), found 244.0875.

Synthesis of 3-(pyridin-4-ylethynyl)-1H-pyrazolo[3,4-b]pyridine (20)

Yield: 86%; ¹H NMR (300 MHz, DMSO-d6) δ 7.37 (dd, J = 8.1 and 4.5 Hz, 1 H), 7.65–7.69 (m, 2 H), 8.42 (dd, J = 8.1 and 1.2 Hz, 1 H), 8.65 (dd, J = 4.5 and 1.2 Hz, 1 H), 8.67–8.71 (m, 2 H), 13.89 (br s, 1 H, NH); MS m/z 221 (M+H)⁺, 159, 130, 121, 102, 88, 85, 56; HRMS calcd for C₁₃H₉N₄ (M+H) 221.0827, found 221.0822.

Synthesis of 3-(pyridin-2-ylethynyl)-1H-pyrazolo[3,4-b]pyridine (21)

Yield: 83%; ¹H NMR (300 MHz, DMSO-d₆) δ 7.35 (dd, J = 8.1 and 4.2 Hz, 1 H), 7.42 (t, J = 5.7 Hz, 1 H), 7.79 (d, J = 7.5 Hz, 1 H), 7.90 (t, J = 7.7 Hz, 1 H), 8.35 (d, J = 8.1 Hz, 1 H), 8.60–8.67 (m, 2 H), 13.87 (br s, 1 H, NH); HRMS calcd for C₁₃H₉N₄ (M+H) 221.0827, found 221.0824.

Assays

LanthaScreen Cellular assay

Cell-based assays for both AKT and PRAS40 phosphorylation were carried out using LanthaScreen cellular assay technology from Invitrogen (Carlsbad, CA).²¹ Briefly, the assay protocol for compound screening is as follows. Cells were plated in white tissue culture-treated 384-well assay plates at a density of 20,000 cells per well in 32 µL of assay medium (low glucose DMEM + 0.1% bovine serum albumin, BSA). After overnight serum starvation, cells were pretreated with 4 µL of compound at the indicated concentrations (10-point dose-response in duplicate) for 60 minutes. The cells were then stimulated with insulin (EC₈₀ concentration of ~5 ng/mL) for 30 minutes to activate PI3K/AKT/mTOR signaling. The assay medium was subsequently removed via aspiration and cells were lysed by the addition of 20 µL of LanthaScreen cellular assay lysis buffer supplemented with protease and phosphatase inhibitor cocktails (Sigma P8340 and P2850, respectively) as well as 2 nM of the Tb-labeled detection antibody, also from Invitrogen. Following 2 hour assay equilibration at room temperature, the TR-FRET emission ratios were acquired on a PerkinElmer EnVision fluorescence plate reader (Waltham, MA) with TRF laser excitation and emission wavelengths of 520 nm and 495 nm. Data analysis and curve-fitting was per formed using XLfit4 and GraphPad Prism4 software.

PI3Kα in vitro assay

The lipid kinase PI3K was assayed by using a LanthaScreen Eu Kinase Binding^® Assay technology from Invitrogen. To a solution of the compounds diluted in assay buffer (25 mM TrisHCl, 5 mM MgCl2, 0.02% chaps) the following components were added: the kinase tracer (PV6088, 20 nM final concentration), the kinase PIK3CA/PIK3R1 (PV4788, 5 nM final concentration) and Eu-labeled anti-His tag antibody (PV6089 2 nM final concentration).

The assay was performed in Corning 3673 white 384-well assay plates. Following 1 hour assay equilibration at room temperature, the TR-FRET emission ratios were acquired on a PerkinElmer EnVision fluorescence plate reader (Waltham, MA) with TRF laser excitation and emission wavelengths of 665 nm and 615 nm. Data analysis and curve-fitting was per formed using XLfit4 and GraphPad Prism4 software.

mTOR in vitro assay

Mammalian target of rapamycin (mTOR) was assayed by using a LanthaScreen Eu Kinase Binding^® Assay technology from Invitrogen. To a solution of the compound diluted in assay buffer (25 mM TrisHCl, 5 mM MgCl2, 0.02% chaps) the following components were added: the kinase tracer (PV6087, 20 nM final concentration), the kinase FRAP1 (PV4753, mTOR, 5nM final concentration) and Eu-labeled anti-GST tag antibody (PV5594, 2nM final concentration).

The assay was performed in Corning 3673 white 384-well assay plates. Following 1 hour assay equilibration at room temperature, the TR-FRET emission ratios were acquired on a PerkinElmer EnVision fluorescence plate reader (Waltham, MA) with TRF laser excitation and emission wavelengths of 665 nm and 615 nm. Data analysis and curve-fitting was per formed using XLfit4 and GraphPad Prism4 software.

PDK1 in vitro assay

The PDK1 kinase was tested by using the Z´-LYTE^® biochemical assay technology from Invitrogen (PV 3180). The assay was performed in Corning 3676 black 384-well assay plate. The final 10 µL Kinase Reaction consisted of 9.75 – 49.4 ng PDK1 and 2 µM of the substarte Ser/Thr 07 in 50 mM of HEPES buffer at pH 8.0, containing 0.01% BRIJ-35, 10 mM MgCl₂, 1 mM EGTA, 0.01% NaN₃. ATP concentration at the Km value was used. After the 1 hour Kinase Reaction incubation, 5 µL of a 1:32768 dilution of Development Reagent A was added. After the development reaction, where a site-specific protease recognizes and cleaves the non-phosphorylated peptide, a ratiometric read-out of the donor emission (Coumarin, 445 nm) over the acceptor (Fluorescein, 520 nm) was detected by a PerkinElmer EnVision fluorescence plate reader (Waltham, MA). Data analysis and curve-fitting was per formed using XLfit4 and GraphPad Prism4 software.

Anti-proliferative Assay

All human cancer cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA) and were maintained in 5% CO₂ at 37°C. Human cervical carcinoma (Hela) and U87 glioblastoma cells were grown in Dulbecco’s modified Eagle’s medium (DMEM, Cellgro) supplemented with 10% fetal bovine serum (FBS, Omega Scientific) and 1% penicillin/streptomycin (Omega Scientific). MDA-MB-231 and MCF7 breast cancer cells and PC3 prostate cancer cells were cultured in RPMI + GlutaMAX Medium (Gibco) plus 10% FBS and 1% penicillin/streptomycin.

Approximately 3000 cells were seeded into individual wells of a 96-well tissue culture plate and incubated for 24hours. Cells were replenished with fresh medium (0.1ml/well) and exposed to triplicates of different concentration solutions (from 0.5 to 100µM) of test compounds. The analyzed inhibitors were dissolved in DMSO reaching a final DMSO concentration of 0.5%. After incubation for 72 hours at 37 °C and 5% CO₂, cell viability was assessed using ATPlite assay from Perkin Elmer (Waltham, MA). Viability was normalized to control cells which were treated with the vehicle, DMSO. The reported IG₅₀ values were calculated by PRISM 5 (Graphpad).

3D culture assay

U87 cells were induced to form spheroids via a hanging drop method.³⁴ Cells were plated at 200 cells per well (20ul) of a Nunc-60 well microwell MiniTray (Polystyrene). The trays were covered, inverted and placed in a humidity chamber for 5 days till one spheroid formed in each well. The spheroids were then transferred into a 48 well plate coated in 1% low melting point agarose to prevent them from adhering. The spheroids were measured and DMSO or drug was added every 24 hours for 6 days. Spheroid length and width (measured with an optical micrometer) were used to calculate spheroid volumes (µm³).

Microsomal Stability Assay (Rat Liver Microsome assay)

Test compound solutions were incubated with rat liver microsomes (RLM) for 60 minutes at 37.5 °C. The final incubation solutions contained 4µM test compound, 2mM NADPH, 1 mg/ml (total protein) microsomes, and 50mM phosphate (pH 7.2). Compound solutions, protein, and phosphate were pre-incubated at 37.5 °C for 5 min and the reactions were initiated by the addition of NADPH and incubated for 1 hour at 37.5 °C. Aliquots were taken at 15 minute time-points and quenched with the addition of methanol containing internal standard. Following protein precipitation and centrifugation, the samples were analyzed by LC-MS. Test compounds were run in duplicate with 2 control compounds of known half life.

Plasma Stability Assay

Test compound solution was incubated (1 µM, 2.5% final DMSO concentration) with fresh rat plasma at 37 °C. The reactions were terminated at 0, 30, and 60 min by the addition of two volumes of methanol containing internal standard. Following protein precipitation and centrifugation, the samples were analyzed by LC-MS. The percentage of parent compound remaining at each time point relative to the 0 min sample is calculated from peak area ratios in relation to the internal standard. Compounds were run in duplicate with a positive control known to be degraded in plasma.

Molecular Modeling

Molecular modeling studies were conducted on a Linux workstation and a 64 3.2-GHz CPUs Linux cluster. Docking studies were performed using the X-ray coordinates of p110alpha (H1037R) mutant in complex with niSH2 of p85alpha and the drug wortmannin (PDB code: 3hhm).²⁴ The PI3Kα crystal structure was extracted from the protein data bank and the complexed ligand was used to define the binding site for docking of small molecules. The genetic algorithm (GA) procedure in the GOLD docking software performed flexible docking of small molecules whereas the protein structure was static.^36–38 For each compound, 20 solutions were generated and subsequently ranked according to GoldScore. Molecular surfaces were generated with MOLCAD ³⁸ and docked structures analyzed with Sybyl (Tripos Inc., ST. Louis).

Abbreviations list

PI3K

Phosphatidylinositol 3-kinase

PDK1

3-phosphoinositide dependent protein kinase-1

RTK

receptor tyrosine kinases

AKT

Protein Kinase B

mTOR

mammalian target of rapamycin

PTEN

phosphatase and tensin homolog protein

PIP₂

phosphatidylinositol-4,5-bisphosphate

PIP₃

phosphatidylinositol-3,4,5-triphosphate

PDCD4

programmed cell death protein

S6K1

S6 kinase beta-1

GFP

Green fluorescent protein