Novel pyrazolo[4,3-d]pyrimidine microtubule targeting agents (MTAs): Synthesis, structure–activity relationship, in vitro and in vivo evaluation as antitumor agents

Abstract

The design, synthesis, and biological evaluation of a series novel N1‑methyl pyrazolo[4,3-d]pyrimidines as inhibitors of tubulin polymerization and colchicine binding were described here. Synthesis of target compounds involved alkylation of the pyrazolo scaffold, which afforded two regioisomers. These were separated, characterized and identified with ¹H NMR and NOESY spectroscopy. All compounds, except 10, inhibited [³H] colchicine binding to tubulin, and the potent inhibition was similar to that obtained with CA-4. Compounds 9 and 11–13 strongly inhibited the polymerization of tubulin, with IC₅₀ values of 0.45, 0.42, 0.49 and 0.42 μM, respectively. Compounds 14–16 inhibited the polymerization of tubulin with IC_50s near ~1 μM. Compounds 9 12, 13 and 16 inhibited MCF-7 breast cancer cell lines and circumvented βIII-tubulin mediated cancer cell resistance to taxanes and other MTAs, and compounds 9–17 circumvented Pgp-mediated drug resistance. In the standard NCI testing protocol, compound 9 exhibited excellent potency with low to sub nanomolar GI₅₀ values (≤10 nM) against most tumor cell lines, including several multidrug resistant phenotypes. Compound 9 was significantly (P <0.0001) better than paclitaxel at reducing MCF-7 TUBB3 (βIII-tubulin overexpressing) tumors in a mouse xenograft model. Collectively, these studies support the further preclinical development of the pyrazolo[4,3-d]pyrimidine scaffold as a new generation of tubulin inhibitors and 9 as an anticancer agent with advantages over paclitaxel.

Cancer is the second-largest cause of mortality in the world and accounted for an estimated 9.6 million deaths in 2018.¹ An effective strategy for cancer therapy is to target the mitotic pathways of rapidly proliferating tumor cells.² Microtubules are hollow tube-like assemblies consisting of heterodimers of two globular protein subunits, α- and β-tubulin (Fig. 1). Tubulin dimers are longitudinally arrayed to form protofilaments, and the protofilaments in turn interact laterally to form the microtubule³ (Fig. 1). Microtubules play essential roles in multiple eukaryotic cellular processes, such as cell growth and division, motility, intracellular trafficking, and the ability to adapt cellular shape in response to the environment.⁴ Molecules that bind to tubulin and interrupt the dynamics of microtubules can be classified as microtubule targeting agents (MTAs),⁴ which possess highly diverse chemical structures, a few of which are shown in Figure 2.^5,6

Fig. 1.

Formation of microtubules: polymerization of dimers of α- and β-tubulin, formation of protofilaments and finally cyclization into microtubules. (Modified from Kaul et al.²⁶).

Fig. 2.

Representative microtubule targeting agents.

Chemotherapy with MTAs has led to improvement in the duration and quality of life for many patients with cancer.⁷ Despite their success, there are several shortcomings associated with MTAs. Treatment with MTAs can result in drug resistance that often is a major obstacle to treatment success.⁷ The clinical utility of the taxanes and the vinca alkaloids is severely limited by two major mechanisms of drug resistance: the expression of the drug efflux pump P-glycoprotein (Pgp) and overexpression of the βIII-isotype of tubulin.^8,9 Pgp expression is very common, particularly in patients who have received prior chemotherapy.¹⁰ The use of Pgp inhibitors to overcome this resistance has been explored but is unsuccessful due to side effects, which occur with high doses.¹¹ The overexpression of βIII-tubulin is involved in clinical resistance to taxanes and vinca alkaloids in nonsmall cell lung,^12,13 breast,¹⁴ ovarian,¹⁵ and gastric¹⁶ cancers. In addition to the resistance issues, the complex natural product MTAs have neurotoxicity, low oral bioavailability and narrow therapeutic indexes as major impediments to clinical success.^17–19

Many reports have shown that CS agents frequently do not display either Pgp or βIII-isotype dependent resistance.²⁰ Thus, development of MTAs targeting the CS can circumvent the resistance limitations associated with taxanes and vinca alkaloids, and their use should improve clinical outcomes.^21,22 Colchicine itself is an approved drug for gout but is not employed as an anticancer agent due to its toxic side effects at higher doses. These include neutropenia, gastrointestinal upset, bone marrow damage, and anemia.²³ Other colchicine binding inhibitors have demonstrated promise and some are currently in clinical trials as anticancer candidates.²⁴

One of the most potent anti-mitotic CS MTAs that gained the attention of medicinal chemists and cancer biologists is combretastatin A-4 (CA-4, Fig. 2), a trimethoxyphenyl (TMP) containing stilbenoid.²¹ CS inhibitors have been extensively studied, and several have entered clinical trials, including 2-methoxyestradiol, combretastatin A-4 phosphate (CA-4P) (fosbretabulin), the combretastain CA-1P prodrug OXi4503 (3-methoxy-2-phosphonooxy-6-[(Z)-2-(3,4,5-trimethoxyphenyl)ethenyl]phenyl] dihydrogen phosphate), AVE8062 ((2S)-2-Amino-3-hydroxy-N-[2-methoxy-5-[(Z)-2-(3,4,5-trimethoxyphenyl)ethenyl]phenyl]propenamide), CKD-516 ((S)-N-(4-(3-(1H-1,2,4-triazol-1-yl)-4-(3,4,5-trimethox-ybenzoyl)phenyl)thiazol-2-yl)-2-amino-3-methylbutanamide hydrochloride), BNC105P (sodium 6-methoxy-2-methyl-3-(3,4,5-trime-thoxybenzoyl)benzofuran-7-yl phosphate), ABT-751 (N-[2-(4-hydrox-yanilino)pyridin-3-yl]-4-methoxybenzenesulfonamide), CYT-997 (1-ethyl-3-[2-methoxy-4-[5-methyl-4-[[(1S)-1-pyridin-3-ylbutyl]amino] pyrimidin-2-yl]phenyl]urea), ZD6126 (phosphoric acid mono-(5-acety-lamino-9,10,11-trimethoxy-6,7-dihydro-5H-dibenzo[a,c]cyclohepten-3-yl) ester), plinabulin (NPI-2358); ((3Z,6Z)-3-benzylidene-6-[(5-tert-butyl-1H-imidazol-4-yl)methylidene]piperazine-2,5-dione), MN-029 (methyl N-[6-[4-[[(2S)-2-aminopropanoyl]amino]phenyl]sulfanyl-1H-benzimidazol-2-yl]carbamate).^13,14,25 Currently there are no CS agents approved by the FDA for cancer, and this demonstrates the need to develop additional CS agents for potential clinical use⁷ because overcoming Pgp and βIII-tubulin mediated resistance is required to improve the overall survival rates observed with MTAs currently in clinical use.

We previously reported^27,28 compounds 2–7 (Table 1 and Fig. 3) as potent CS antiproliferative agents. In the structure–activity relationship (SAR) studies of these pyrrolo[3,2-d]pyrimidine based antitubulin agents, we demonstrated the importance of the 5‑methyl moiety on compounds 3, 5 and 7 for significant improvement of biological activities over the 5‑desmethyl compounds 2, 4 and 6. Compound 3 with a 5‑methyl is 160- and 22-fold more potent in cellular microtubule depolymerization activity and antiproliferative effects, respectively than the corresponding 5‑desmethyl 2. Moreover, compounds 5 and 7 are 43- and 4-fold more potent in inhibiting tubulin assembly respectively, than compounds 4 and 6. On the basis of our previous studies it was of interest to design a novel series of N1-methylated pyrazolo[4,3-d]pyrimidines (Fig. 4) as potential MTAs and as anticancer agents. As the first step of our efforts, we utilized scaffold hopping on the pyrrolo[3,2-d] pyrimidines as an approach to design novel pyrazolo[4,3-d]pyrimidines. Scaffold hopping is a drug design strategy to identify structurally diverse compounds that are similar in activity/property space.²⁹ An additional aspect incorporated into the pyrazolo[4,3-d]pyrimidine structure was a 2‑chloro substituent. It was observed that a 2‑chloro substituted purine compound 8 displayed IC₅₀ values of 39.6 ± 2.9, 41.2 ± 3.2 and 58.9 ± 3.1 nM in three human malignant melanoma cell lines A375, M14 and RPMI7951, respectively.³⁰ Thus, it was of interest to explore the antitumor activity of these compounds with the novel N1-methyl-pyrazolo[4,3-d]pyrimidine scaffold with a chlorine substitution at the 5-position as MTAs. We used various additional structural modifications for the design of compounds 9–17 (Fig. 4) that include d i) rigidification or conformational restrictions, ii) homologation and branching on the alkyl chain and iii) variation in electronic effects on the aromatic ring.

Table 1.

Microtubule depolymerization activity, antiproliferative effects, inhibition of colchicine binding and tubulin assembly.

The increase of activity of lead compounds with a 5-methyl group
No	Cellular microtubule depolymerization EC50 (nM)	Antiproliferative effects MDA-MB-435 IC50 (nM) ± SD	Inhibition of colchicine binding		Inhibition of tubulin assembly IC50 (μM ± SD)
			5 μM inhibitor % inhibition ± SD	0.5 μM inhibitor% inhibition ± SD
1	>10000	ND	0	ND	ND
2	1200	96 ± 5	60 ± 1	ND	10 ± 0.6
3	7.4	4.3 ± 0.3	99 ± 2	79 ± 0.8	ND
4	-	-	ND	ND	21 ± 1
5	-	-	92 ± 0.2	73 ± 3	0.48 ± 0.008
6	-	-	43 ± 3	ND	3.3 ± 0.3
7	-	-	72 ± 2	ND	0.91 ± 0.03
CA-4	9.8	4.4 ± 0.5	99 ± 0.8	84 ± 3	0.54 ± 0.06

Fig. 3.

Lead compounds.

Fig. 4.

Compounds 9–17 and the rationale for drug design.

Our optimization effort started by replacing the pyrrole ring in 3 with a pyrazolo ring and a 5‑chloro group to afford compound 9. Using molecular modeling (Maestro 2020–2),³¹ we docked several conformationally restricted analogs in the CS of tubulin. As shown in Fig. 4, the bioactive conformations of 9 and 10 are determined by three rotatable single bonds: 4-position C–N bond (bond a), 1^′-position C–N bond (bond b) and 4^′-position C–O bond (bond c). Literature findings³² suggest that conformational preference via molecular modeling and ¹H NMR studies indicate that a methyl group on the aniline nitrogen in 1 and 2 restricts the free rotation of both bond a and bond b (Fig. 3) and thus restricts the conformation of the anilino ring. Conformational constraint of a small-molecule inhibitor can enhance the binding affinity to its intended target protein by reducing conformational entropic costs upon binding and can also improve the binding selectivity by reducing accessible, low-energy conformational space.^33,34 We, therefore, modeled the conformationally restricted analogs 9, 11, and 12. In compound 9, the restriction of bond a was accomplished by incorporating a N7‑methyl group in 10. Compound 11, with the 1,2,3,4-tetrahydroquinoline at the N7-position was designed from 9 by further restricting bond b. Compound 12 restricts both bonds a and b by incorporating a fused bicyclic 6ʹ-methoxynaphthyl-1ʹ-amino ring at the N7-position as a replacement of the monocyclic 4^′-methoxyphenyl group of 9. In compounds 9, 11 and 12, bond rotation restrictions were anticipated to reduce the energy to adopt the bound conformation, thus allowing improved inhibitory activity.

Next, we focused on the homologation and branching of the 4^′-OMe (9) to 4^′-OEt (13) and 4^′-O-iPr (14) to allow for better penetration into the hydrophobic pocket of the CS. We also explored the impact of electron density and hydrogen bonding ability in compounds 15–17. The distance and the nature of heteroatom substitution affects hydrogen bond (HB) strength.³⁵ Thus, it was of interest to isosterically replace the oxygen atom of the 4^′-OCH₃ of 9 with a sulfur and synthesize 15. Compounds 16 and 17 were designed and synthesized to probe the electron withdrawing effect of fluorine on the inhibition of colchicine binding and microtubule assembly and on anticancer activity.

Compounds 1–17 were docked in the X-ray crystal structure of the CS (PDB: 6BS2, 2.65 Å)³⁰ using Maestro, Schrödinger 2020–2.³¹ Multiple low energy conformations were obtained from docking. Fig. 5 shows the docked conformation of lead 3 (pink) and compound 9 (tan) superposed with CA-4 (cyan). The pyrazolo[4,3-d]pyrimidine scaffold of 9–17 form hydrophobic interactions at the α_β tubulin interface with Alaβ314, Alaβ352 and Ileuβ316 similar to the lead pyrrolo[3,2-d]pyrimidine scaffold. The pyrazole ring is stacked amid the Metβ257 and Cysβ239. The N7-CH₃ moiety of 9 and 11–17 forms hydrophobic interactions with Alaβ314 and Leuβ253. The 4^′-oxygen of compounds 9–14, and 16–17 and the 4^′-sulfur of compound 15 are oriented towards the pocket formed by Thrα179, Lysβ350, Asnβ256 and Asnβ348. The 5-Cl of 9–17 makes hydrophobic interactions with Leuβ253 and Leuβ240 (only 9 is shown in Fig. 5B). The N4 of the pyrazolo[4,3-d]pyrimidine ring of all the compounds makes water mediated hydrogen bond interactions with the backbone of Cysβ239 similar to the original X-ray crystallized ligand.³⁰ The docked score of compound 9 is − 10.71 kcal/mol and is better than the lead compound 3 (− 9.76 kcal/mol). The docked scores for 10 is − 9.32, and for 11–17 are in a range of − 9.95 to − 10.65 kcal/ mol. The N7‑desmethyl compound 10 has the lowest docked score of the designed analogs and was predicted to be the least active.Fig. 6

Fig. 5.

A) Docked pose of lead 3 (pink) with CA-4 (cyan) in the CS. B) Docked pose of 9 (tan) with CA-4 (cyan) in the CS. (Maestro 2020–2. Docked scores of 3 and 9 are − 9.76 and − 10.71 kcal/mol, respectively).

Fig. 6.

Treatment with compound 9 decreased primary tumor growth in MCF-7 TUBB3 (βIII-tubulin overexpressing) breast cancer cell model. * = P <0.05; *** P <0.001 (0.0007 for control vs 9, 0.0003 for PTX vs 9. A) 10⁷ cells/100 uL Matrigel as implanted into fat pad #4 of 8 wk old athymic female mice. The mice were treated with compound 9 at the maximally tolerated dose (MTD) of 30 mg/kg two times weekly or with paclitaxel at its MTD of 10 mg/kg/week and tumor volumes were determined. Statistical analysis was performed with two-way ANOVA repeated measures post test. Compound 9 was significantly (P <0.0001) better than paclitaxel at reducing MCF-7 TUBB3 (βIII-tubulin overexpressing) tumor volume. B) Animal weights were graphed as percent weight change at day 41 over the starting weight. Statistical analysis was performed with one-way ANOVA. Only the control and paclitaxel mice lost weight at the end of the study. Sample sizes are 7 for control (an animal was lost toward the end of the experiment) and 8 for PTX and 9.

Boc-protection of 18 afforded 19 in 72% yield. Methylation of 19 to 20 and subsequent deprotection³⁶ provided the aniline intermediate 21 (Scheme 1). Treatment of 22a-27a with paraformaldehyde, followed by reduction with NaBH₄ gave the anilino intermediates 22b-27b (Scheme 2). In Scheme 3, pyrazole 28 was alkylated with methyl iodide in the presence of base K₂CO₃ in DMF, to provide after separation two regioisomers, 29a and 29b, in 35 and 55% yields, respectively. Intermediates 29a and 29b were characterized by ¹H NMR and NOESY (Nuclear Overhauser Enhancement Exchange Spectroscopy). Varying the base has been used in the literature to alkylate the N1 position.^37,38 Our use of K₂CO₃ is the first reported to yield 29a and 29b in 35 and 55% yields, respectively. The separation method is discussed in the Experimental Section. The ¹H NMRs of 29a and 29b agree with those in the literature.³⁹ Reduction of the nitro group in 29a, yielded 30, which was cyclized to 31 with urea. Chlorination of 31 with POCl₃ and pyridine in toluene provided 32. Nucleophilic aromatic substitution of 32 using anilines 4^′-methoxyaniline, 4^′-methoxy-N-methylaniline, 21, 22b–27b afforded final compounds 9–17 in 68–84% yields.

Scheme 1.

a) Di-tert-butyl dicarbonate, DCM, rt, 4 h, 72%; b) CH₃I, DMF, rt, 12 h, 55%; c) CF₃COOH, THF, rt, 6 h, 69%.

Scheme 2.

i) Paraformaldehyde, DCM, 12 h; ii) NaBH₄, DCM, rt, 6–12 h, 56–72% over two steps.

Scheme 3.

CH₃I, K₂CO₃, DMF, rt, 12 h (29a: 35%, 29b: 55%); b) 10% Pd/C, H₂, MeOH, 40 psi, 1 h, 72%; c) urea, 180 ^◦C, 2 h; d) POCl₃, pyridine, toluene, 4 h, reflux, 42%; e) anilines, acetonitrile, reflux, 4–12 h, 35–49%.

Compounds 9–17 and the reference compound CA-4 were evaluated for inhibitory effects on tubulin polymerization and on the binding of [³H]colchicine to tubulin as an indication of whether their antiproliferative effects are the result of an interaction with tubulin (Table 2). CA-4 is a drug candidate in clinical trials,^40,41 and it is a highly potent, competitive inhibitor of the binding of colchicine to tubulin.⁴² Except for 10, all the compounds at, 5 μM, inhibited [³H]colchicine binding to the protein, and the extent of inhibition was similar to that obtained with CA-4. With equal or more than 94% and 68% inhibition at 5 and 0.5 μM, respectively, compounds 9, 11, 13 and 16 showed similar potency as the lead compounds and CA-4 (Tables 1 and 2). Isosteric replacement of 4^′-OMe with 4^′-SMe generated compound 15, which was equipotent to 9 in the inhibition of colchicine binding assay. Compounds 11 and 12, the conformationally restricted analogs of 9, in the colchicine binding assay (Table 2), displayed comparable potency to 9 (94 and 91% at 5 μM and 68 and 66% at 0.5 μM, respectively). Next, we assessed the effects of homologation and branching at the 4^′-O-position of 9. Homologation with a single methylene at the 4^′-position of 9 (4^′-OMe) afforded 13 (4^′-OEt), which was more potent (99% at 5 μM and 88% at 0.5 μM) than both 9 and CA-4 in the colchicine inhibition assay. Branching with an isopropyl moiety afforded compound 14, which was slightly less potent than 9, with 85% inhibition at 5 μM. Introduction of a fluorine atom at the 3^′-position of the phenyl ring of 9 yielded compound 16, which had 97 and 77% inhibition at 5 and 0.5 μM, respectively, in the inhibition of colchicine binding assay. Compounds 11–16 at 5 μM, inhibited the binding of [³H]colchicine by 85–99%, whereas compound 17 showed only a 64% inhibition at 5 μM of [³H]colchicine binding. Compound 10 was completely inactive in inhibiting colchicine binding. The lower activity of 17 could be attributed to the strong electron withdrawing effect of the CF₃ moiety on the oxygen of the 4^′- position. This oxygen (or sulfur in 16) does interact with the pocket formed by Thrα179, Lysβ350, Asnβ256 and Asnβ348 at the colchicine site, and the CF₃ reduces this interaction. Compound 10, lacking the N7‑methyl moiety, did not inhibit tubulin assembly, nor did it have a significant effect on colchicine binding, emphasizing the critical role of this methyl moiety at the N7-position in retaining the bioactive conformation for anti-tubulin activity (described below). The lack of activity of 10 corroborates the lowest docked score of 10 in the CS obtained from our molecular modeling studies.

Table 2.

Inhibition of colchicine binding and tubulin assembly.

Compound No	Inhibition of colchicine binding		Inhibition of tubulin assembly IC50 (μM ± SD)
	5 μM inhibitor% inhibition ± SD	0.5 μM inhibitor% inhibition ± SD
2	60 ± 1	ND	10 ± 0.6
3	99 ± 2	79 ± 0.8	ND
4	ND	ND	21 ± 1
5	92 ± 0.2	73 ± 3	0.48 ±0.008
6	43 ± 3	ND	3.3 ±0.3
7	72 ± 2.0	ND	0.91 ± 0.03
9	94 ± 3	68 ± 3	0.45 ±0.09
10	7.1 ± 3	-	>20
11	94 ± 0.4	68 ± 2	0.42 ± 0.07
12	91 ± 1	66 ± 0.7	0.49 ±0.01
13	99 ± 0.4	88 ± 0.007	0.42 ± 0.07
14	85 ± 1	-	1.1 ± 0.08
15	93 ± 0.5	64 ± 2	0.82 ± 0.02
16	97 ± 0.3	77 ± 1	0.74 ±0.04
17	64 ± 5	-	2.9 ±0.2
CA-4	99 ± 0.8	84 ± 3	0.54 ±0.06

All of the compounds (except 10) were highly active in the tubulin assembly assay with low to sub micromolar inhibitory IC₅₀ values (0.42–1.1 μM), comparable with lead compounds 5 and 7 (0.48 and 0.91 μM, respectively) and CA-4 (0.54 μM) (Table 2). Compounds 9 and 11–13 strongly inhibited the polymerization of tubulin, with IC₅₀ values 0.45, 0.42, 0.49 and 0.42 μM, respectively. Compound 10 had no effect on tubulin polymerization. Compounds 14–16 inhibited the polymerization of tubulin with an IC_50s near ~1 μM, 2-fold less potent than CA-4.

The potent MTAs 9, 12, 13 and 16 were selected for evaluation of their inhibitory activities towards the growth of MCF-7 wild-type (WT) human breast cancer cells and and a MCF-7 cell line overexpressing βIII-tubulin (data shown in Table 3). Comparing the EC₅₀ values in these cell lines, compounds 9 and 16 were the most potent of the series (EC₅₀s 2.0 and 3.3 nM for 9 and 1.7 and 2.6 nM for 16 in MCF-7 WT and MCF7-βIII-tubulin overexpressing cell lines, respectively). Compounds 9 and 16 were 100- and ~600-fold, respectively more active than paclitaxel in the MCF-7 WT and MCF-7-βIII-tubulin cell lines. On the other hand, compounds 12 and 13 were 5- and 25-fold more potent, respectively, in the MCF-7 WT cells; and 120- and 220-fold more potent, respectively, in the MCF-7-βIII-tubulin cell line compared to paclitaxel. Comparison of the EC₅₀ values in the parental MCF-7 WT and genetically manipulated MCF-7 βIII cell line allows for the calculation of a relative resistance value, designated as Rr. This value is calculated by dividing the EC₅₀ value obtained in the βIII-overexpressing MCF-7 cells by the EC₅₀ obtained in the parental MCF-7 WT cells. The isoform βIII is an important determinant in cellular resistance towards paclitaxel, which is a known Pgp substrate.^11,16 The Rr value of paclitaxel was 10.6 (resistance ratio: βIII-tubulin/WT). Compound 12 had an Rr of 0.5, indicating that it is able to overcome drug resistance mediated by βIII. Compounds 9, 13, and 16 also had Rr values ≤1.6, suggesting that they are all poor substrates for βIII-tubulin.

Table 3.

Activities of compounds 9, 12, 13 and 16 in a βIII-tubulin overexpressing cell line.

Agent	MCF-7 WT EC50 (nM)	MCF-7 TUBB3 βIII-tubuliu EC50 (nM)	Resistance Ratio MCF-7 TUBB3/WT
9	2.0 ± 0.3	3.3 ± 0.4	1.1
12	35.2 ± 4.7	16.8 ±1.5	0.5
13	7.7 ±0.7	9.7 ±1.0	1.3
16	1.7 ±0.2	2.6 ±0.4	1.6
Paclitaxel	194.2	2065.8	10.6

The ability of 9–17 to circumvent Pgp-mediated drug resistance was evaluated using an ovarian cancer cell line pair (Table 4). The Rr value is calculated by dividing the IC₅₀ value obtained in the Pgp overexpressing NCI/ADR-RES cells by the IC₅₀ obtained in the parental OVCAR-8 cells. The Rr value for paclitaxel was 150 whereas most of the synthesized compounds had Rr values less than 1. The <1 Rr values indicated that these compounds were poor substrates for Pgp and hence have minimal Pgp-mediated transport and consequently should have advantages over MTAs that are good Pgp substrates.

Table 4.

Compound activity in a Pgp overexpressing cell line.

Compound No	Compound activity in a Pgp overexpressing cell linea		Rr
	Parental OVCAR-8 IC50 (nM) ± SD	Pgp overexpressing NCI/ADR-RES IC50 (nM) ± SD
9	9.0 ± 0.7	5.0 ± 0	0.55
10	>5000	>5000	-
11	17 ± 2	8.0 ± 0.7	0.47
12	14 ± 3	11 ± 0.7	0.78
13	2.0 ± 1	4.0 ± 1	2.0
14	49.0 ± 10	44 ± 8.0	0.89
15	31 ± 2	17 ± 3	0.54
16	9.0 ± 1	7.0 ± 3	0.77
17	880 ± 40	730 ±100	0.83
Paclitaxel	5.3 ± 2	800 ± 200	150
CA-4	1.8 ± 0.4	1.8 ± 0.4	1.0

To determine its antitumor spectrum, compound 9 was tested in the National Cancer Institute’s panel of 60 human tumor cell lines (NCI-60). As shown in Table 5, compound 9 exhibited excellent potency with low GI₅₀ values ≤10 nM against most tumor cell lines, including many multidrug resistant phenotypes, in the standard NCI testing protocol. The GI₅₀ values of 9 was single digit nanomolar in 19 tumor cell lines.

Table 5.

Human cancer cell growth inhibitory activity GI₅₀ (nM) of 9 in NCI 60 cell line panel.

Panel/Cell line	GI50 (nM)	Panel/Cell line	GI50 (nM)	Panel/Cell line	GI50 (nM)	Panel/Cell line	GI50 (nM)
Leukemia		Colon Cancer		Melanoma		Renal Cancer
CCRF-CEM	9.96	COLO 205	9.78	LOX IMVI	8.53	786-0	13.23
HL-60 (TB)	11.23	HCC-2998	23.56	MALME-3M	12.07	A498	13.42
K-562	6.46	HCT-116	9.07	M14	10.56	ACHN	8.40
MOLT-4	19.85	HCT-15	5.14	MDA-MB-435	3.90	CAKI-1	15.72
RPMI-8226	15.29	HT29	7.60	SK-MEL-2	20.97	RXF 393	8.75
SR	9.64	KM12	10.76	SK-MEL-28	17.04	SN12C	15.76
NSCLC		SW-620	6.67	SK-MEL-5	14.58	TK-10	16.31
A549/ATCC	12.67	CNS Cancer		UACC-257	22.69	UO-31	10.01
EKVX	15.47	SF-268	12.17	UACC-62	15.90
HOP-62	15.41	SF-295	11.43	Ovarian cancer		Breast Cancer
HOP-92	16.03	SF-539	8.76	IGROVI	13.30	MCF7	10.81
NCI-H226	9.49	SNB-19	13.88	OVCAR-3	6.37	MDA-MB-231/ATCC	14.29
NCI-H23	11.87	SNB-75	10.81	OVCAR-4	12.93	HS 578T	11.19
NCI-H322M	14.09	U251	7.97	OVCAR-5	11.76	BT-549	24.12
NCI-H460	4.52	Prostate Cancer		OVCAR-8	14.69	T-47D	16.44
NCI-H522	14.43	PC-3	10.73	NCI/ADR-RES	8.24	MDA-MB-468	9.09
		DU-145	8.89	SK-OV-3	13.93

Compound 9 was selected for a tubulin III overexpressing, antimicrotubule drug resistant in vivo xenograft mouse study on the basis of its nanomolar potency in vitro in the NCI cancer cell line panel and its potent activities in inhibiting microtubule polymerization and colchicine binding assay. Compound 9 significantly reduced primary tumor growth vs paclitaxel in the tubulin III overexpressing MCF-7 TUBB3 orthotopic xenograft, with no significant weight loss in the study as opposed to paclitaxel, which resulted in significant weight loss. Thus compound 9 is effective in vivo against antimicrotubule drug resistant breast cancer.

N-Alkylation of the pyrazole scaffold 28 afforded two regioisomers, 29a and 29b. Methylation of unsymmetrically substituted pyrazole derivatives usually affords mixtures of both possible alkylated products.³⁹ The regiochemistry of both alkylated intermediates 29a and 29b was identified by ¹H NMR³⁹ and NOESY spectroscopy (Figs. 7 and 8). In ¹H NMR (only δ values are included, spectra not shown). For the N1-CH₃ protons the chemical shift for 29a and 29b occurred as singlets at 4.02 and 3.95 ppm, respectively. It is noteworthy that the presence of the adjacent C5-carboxylate to the N1-CH₃ in 29a caused the N1-CH₃ protons of 29a to shift ~0.1 ppm deshielded compared to the N1-CH₃ of 29b. The initial studies involved NMR utilizing NOESY spectroscopy, which is one of the most direct ways to determine spatial proton-proton/proton-heteronucleus/heteronucleus-heteronucleus correlations within a molecule in the range up to 5 Å or less.^39,43 This is also an excellent method to determine solution conformation. Intermediate 29a was identified as the methyl 1-methyl-4-nitro-1H-pyrazole-5-carboxylate regioisomer, as no strong NOE correlation was observed between the H3 proton and the N1-CH₃ protons (Fig. 7). Intermediate 29b was determined to be the methyl 1-methyl-4-nitro-1H-pyrazole-3-carboxylate regioisomer, as a characteristic NOE correlation was observed between the H5 proton and the N1-CH₃ protons (Fig. 8). The observed H5 proton and N1-CH₃ cross peak (δ 8.95, 3.95) was dominated by scalar coupling effects, with evidence of only a NOE proton contributing to it in 29b. For 29a, the H3 proton occurs at δ 8.36, and, for 29b, the H5 proton is at δ 8.95. The probable deshielding effects of the H5 proton in 29b is due to the direct electron withdrawing effect of C-5. ¹H NMR and crystal structure from a previous report, along with the 2D NMR from our studies, confirm the structure of regioisomers 29a and 29b as designated.

Fig. 7.

NOESY NMR of Intermediate 29a.

Fig. 8.

NOESY NMR of Intermediate 29b.

The ¹H NMR spectra of the N7-H analog 10 (Fig. 9) and the N7-CH₃ analog 9 (Fig. 10) in DMSO‑d6 afforded additional information related to conformational restriction in 9 as compared to 10. For compound 9, the “N1-CH₃” protons appeared at δ 3.04, whereas for 10 they were significantly deshielded at δ 4.35 ppm. This shielding of the “N1-CH₃” protons in 9 was attributed to a diamagnetic anisotropic effect in 9 arising from the proximity of the phenyl ring as shown in Fig. 11 (more favored anti-conformation for 9 on the basis of the N7-CH₃ and the N1- CH₃ groups). The steric bulk of the N7-CH₃ resulting in a steric clash of the N7-CH₃ and the N1-CH₃ groups in 9 restricts the conformation and positions the phenyl ring on top of the N1-CH₃ moiety (Fig. 11), resulting in the observed shielding effect, in the ¹H NMR, on the N1-CH₃ group in 9 as compared to that in 10. This shielding effect of the phenyl group on the N1-CH₃ group (δ _≈ 1.1) was also observed for the N7-CH₃ analogs 11–17. This provides an estimation of the solution conformation of the active analog 9 compared to the inactive analog 10. In addition, the docked conformation of 9 in Fig. 5B is similar to the ¹H NMR solution conformation of 9. This indicates a low energy barrier for 9 to adopt the bound conformation and results in potent activity.

Fig. 9.

¹H NMR of 10.

Fig. 10.

¹H NMR of 9.

Fig. 11.

Possible explanation of orientation of compound 9.

In this study, a series of nine novel N1-methyl-pyrazolo[4,3-d]pyrimidines were designed and synthesized as anticancer agents that function as CS binding agents and as inhibitors of tubulin polymerization. Synthesis of the target compounds involved alkylation of the pyrazolo scaffold to provide two regioisomers that were separated, characterized and structurally identified via ¹H NMR and NOESY spectroscopy. Compounds 9 and 11–13 strongly inhibited the polymerization of tubulin with IC₅₀s near ~0.5 μM., whereas compounds 14–16 inhibited the polymerization of tubulin with IC_50s near ~1 μM. Compounds 9, 12, 13 and 16 circumvented βIII-tubulin mediated cancer cell resistance and compounds 9–17 circumvented Pgp mediated drug resistance. Compound 9 showed GI₅₀ values ≤10 nM against many tumor cell lines, including several multidrug resistant phenotypes, in the standard NCI testing protocol. SAR studies indicated that the N7-CH₃ moiety of 9 was crucial for activity and that the 4^′-OMeC₆H₄ was the best substitution for biological activity. In addition, the lower activity of 11 and 12 compared to 9 and 13 suggested that some flexibility of bond ‘a’ and ‘b’ were necessary for potent activity in this series. Compound 9 was significantly (P <0.0001) better than paclitaxel at reducing MCF-7 TUBB3 (βIII-tubulin overexpressing) tumor growth in an in vivo study and is poised for further preclinical development.

Abbreviations:

CS

colchicine site

CA-1

combretastatin A-1

CA-4

combretastatin A-4

CA-4P

combretastatin A-4 phosphate

DMSO

dimethyl sulfoxide

DIPEA

N,N-Diisopropylethylamine

MTAs

microtubule targeting agents

NCI

National Cancer Institute

NOESY

Nuclear Overhauser Effect Spectroscopy

Pgp

P-glycoprotein

PDB

Protein Data Bank

SRB

sulforhodamine B

TFA

trifluoracetic acid