Discovery of N,4-Di(1H-pyrazol-4-yl)pyrimidin-2-amine-Derived CDK2 Inhibitors as Potential Anticancer Agents: Design, Synthesis, and Evaluation

Cyclin-dependent kinase 2 (CDK2) has been garnering considerable interest as a target to develop new cancer treatments and to ameliorate resistance to CDK4/6 inhibitors. However, a selective CDK2 inhibitor has yet to be clinically approved. With the desire to discover novel, potent, and selective CDK2 inhibitors, the phenylsulfonamide moiety of our previous lead compound 1 was bioisosterically replaced with pyrazole derivatives, affording a novel series of N,4-di(1H-pyrazol-4-yl)pyrimidin-2-amines that exhibited potent CDK2 inhibitory activity. Among them, 15 was the most potent CDK2 inhibitor (Ki = 0.005 µM) with a degree of selectivity over other CDKs tested. Meanwhile, this compound displayed sub-micromolar antiproliferative activity against a panel of 13 cancer cell lines (GI50 = 0.127–0.560 μM). Mechanistic studies in ovarian cancer cells revealed that 15 reduced the phosphorylation of retinoblastoma at Thr821, arrested cells at the S and G2/M phases, and induced apoptosis. These results accentuate the potential of the N,4-di(1H-pyrazol-4-yl)pyrimidin-2-amine scaffold to be developed into potent and selective CDK2 inhibitors for the treatment of cancer.


Introduction
Cyclin-dependent kinase 2 (CDK2) is a serine/threonine protein kinase that is activated by binding to cyclin A or E and phosphorylation at its Thr160 residue by the CDK-activating kinase (CAK, i.e., CDK7-cyclin H-MAT1). The activity of CDK2 is also negatively regulated by phosphorylation at its Thr14 and Tyr15 by Wee1/Myt1 as well as by binding to CDK inhibitory proteins such as the CDK-interacting protein (Cip)/kinase inhibitory protein (Kip) family members (i.e., p21 Cip1 , p27 Kip1 , and p57 Kip2 ) [1]. CDK2 plays important roles in overseeing various facets of the cell division cycle, including regulation of the G1-to-S transition, centrosome duplication, DNA replication and repair, and activation of CDK1/cyclin B for the G2-to-M transition [2,3]. Aberrant activity of CDK2 is linked with a myriad of human cancer types, including those originating from the ovary [4], breast [5], lung [6], and brain [1]. Moreover, this kinase is associated with resistance to CDK4/6 inhibitors [7,8]. Therefore, CDK2 has been an attractive target for the development of new anticancer treatments and the amelioration of resistance to CDK4/6 inhibitors [9]. Indeed, several studies have demonstrated the potential of CDK2 inhibition in cancer treatments both as a monotherapy and as part of combination therapies [9][10][11][12].
Structurally-diverse small-molecule CDK2 inhibitors have been developed, including AT7519, CYC065, dinaciclib, PF-06873600, and PF-07104091, all of which are currently at different stages of clinical trials, however, there is not yet any clinically-approved selective CDK2 inhibitor [1,13]. Consequently, it is imperative to develop novel, potent, and selective CDK2 inhibitors.
Structurally-diverse small-molecule CDK2 inhibitors have been developed, including AT7519, CYC065, dinaciclib, PF-06873600, and PF-07104091, all of which are currently at different stages of clinical trials, however, there is not yet any clinically-approved selective CDK2 inhibitor [1,13]. Consequently, it is imperative to develop novel, potent, and selective CDK2 inhibitors.

Inhibitor Design
With the desire to discover potent and selective CDK2 inhibitors with drug-like properties, we used a bioisosteric replacement strategy, i.e., substitution of the phenylsulfonamide moiety of 1, our recently-reported lead compound [20], with a wide range of pyrazole-derived groups to conceive a series of N,4-di(1H-pyrazol-4-yl)pyrimidin-2-amine analogues ( Figure 1). We envisioned that this replacement might improve the potency, selectivity, and physiochemical properties. Moreover, the N-NH motif of the newly introduced pyrazole ring could serve as both a hydrogen-bond acceptor and donor to interact with CDK2, and such an interaction might contribute positively to the potency of the entire molecule [27,28].
To understand the potential of these newly designed compounds to inhibit CDK2, a molecular modeling study on 14, an exemplar of the proposed series (Scheme 1), or 1 in

Inhibitor Design
With the desire to discover potent and selective CDK2 inhibitors with drug-like properties, we used a bioisosteric replacement strategy, i.e., substitution of the phenylsulfonamide moiety of 1, our recently-reported lead compound [20], with a wide range of pyrazolederived groups to conceive a series of N,4-di(1H-pyrazol-4-yl)pyrimidin-2-amine analogues ( Figure 1). We envisioned that this replacement might improve the potency, selectivity, and physiochemical properties. Moreover, the N-NH motif of the newly introduced pyrazole ring could serve as both a hydrogen-bond acceptor and donor to interact with CDK2, and such an interaction might contribute positively to the potency of the entire molecule [27,28].
To understand the potential of these newly designed compounds to inhibit CDK2, a molecular modeling study on 14, an exemplar of the proposed series (Scheme 1), or 1 in complex with CDK2/cyclin E was performed, and their respective plausible binding modes are shown in Figure 2. Both compounds are predicted to engage the hinge region, forming hydrogen bonds through their common pyrimidinyl-N1-C2-NH motif with Leu83 of CDK2. These highly similar CDK2-binding features between 14 and 1 encouraged us to synthesize the N-pyrazole analogues. complex with CDK2/cyclin E was performed, and their respective plausible binding modes are shown in Figure 2. Both compounds are predicted to engage the hinge region, forming hydrogen bonds through their common pyrimidinyl-N1-C2-NH motif with Leu83 of CDK2. These highly similar CDK2-binding features between 14 and 1 encouraged us to synthesize the N-pyrazole analogues.

Structure-Activity Relationships
With the desire to improve the potency and selectivity of our previously-identified lead compound 1 [20], its phenylsulfonamide moiety was replaced with pyrazole-derived groups, affording a series of N,4-di(1H-pyrazol-4-yl)pyrimidin-2-amine derivatives. These new analogues were then evaluated for their activities against CDK2E; in order to under-

Structure-Activity Relationships
With the desire to improve the potency and selectivity of our previously-identified lead compound 1 [20], its phenylsulfonamide moiety was replaced with pyrazole-derived groups, affording a series of N,4-di(1H-pyrazol-4-yl)pyrimidin-2-amine derivatives. These new analogues were then evaluated for their activities against CDK2E; in order to understand their selectivity profiles, they were also tested against other CDKs, namely CDK1B, CDK5p25, and CDK9T1, all of which share high amino acid sequence similarity with CDK2E. Additionally, the antiproliferative activities of these compounds were screened against A2780 ovarian cancer cells.
As shown in Table 1, substitution of the phenylsulfonamide moiety of 1 with an unsubstituted pyrazol-4-yl ring gave 14, which displayed more potent CDK2 and CDK5 inhibition with K i values of 0.007 and 0.003 µM, respectively. Compound 14 also improved the selectivity of 1 for CDK2 over CDKs 1 and 9. However, 14 showed about 28-fold lower antiproliferative activity against A2780 ovarian cancer cells when compared to 1. Replacement of the fluorine atom in 14 (R 2 = F) with chlorine afforded 15 (R 2 = Cl), which maintained the inhibitory potencies towards CDK2 (K i = 0.005 µM) and CDK5 (K i = 0.003 µM) and slightly enhanced the selectivity for CDK2 over CDK1 and CDK9. Furthermore, in comparison with 1, 15 showed improved potency against and selectivity for CDK2 but reduced antiproliferative activity against A2780 cells (15: GI 50 = 0.158 µM versus 1: GI 50 = 0.018 µM). Encouraged by these results (i.e., enhanced CDK2 inhibitory activity and selectivity), further SAR studies were carried out by varying R 1 and R 3 substituents while retaining R 2 as chlorine or fluorine.  018 The enzymatic inhibition percentage was determined using a luminescent ADP-Glo™ assay. Apparent inhibition constant (Ki) values were calculated using their corresponding half maximal inhibition (IC50) values and the appropriate Km(ATP) of each kinase as described in the experimental section. The antiproliferative activity was determined by 72 h 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays using A2780 cells. The asterisk (*) indicates % for either biochemical or cell-based data. Previously reported data for 1 [20] were included in the table for comparison purposes. CDK1B, CDK2E, CDK5p25, and CDK9T1 denote CDK1/cyclin B, CDK2/cyclin E, CDK5/p25, and CDK9/cyclin T1, respectively. Me: methyl.
We next investigated the SAR around the other pyrazole ring occupying the pyrimidinyl-C4 position (i.e., R and R 1 ). As shown in Table 2 Table 1)), thus signifying that diversifica- The enzymatic inhibition percentage was determined using a luminescent ADP-Glo™ assay. Apparent inhibition constant (K i ) values were calculated using their corresponding half maximal inhibition (IC 50 ) values and the appropriate K m (ATP) of each kinase as described in the experimental section. The antiproliferative activity was determined by 72 h 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays using A2780 cells.
Substitution of the N-pyrazol-4-yl-NH in 14 and 15 (R 3 = H) by a methyl, affording 16 and 17 (R 3 = Me), respectively, led to a reduction in the inhibitory potencies across all the CDKs tested and, consequently, resulted in diminished antiproliferative activities against A2780 cells. An acetamide group was then introduced at the same position to test whether the terminal amide could interact with Asp86 and/or Lys89 residues of CDK2. However, 18 and 19 (R 3 = CH 2 CONH 2 ) thus obtained had detrimental effects on the inhibition of all four CDKs and A2780 cells. Taken together, these results suggest the importance of the unmasked pyrazole ring (R 3 = H) at the pyrimidinyl-C2-NH position for inhibitory activities towards both CDK2 and A2780 cells.
We next investigated the SAR around the other pyrazole ring occupying the pyrimidinyl-C4 position (i.e., R and R 1 ). As shown in Table 2, methylation of the pyrazolyl-C3 position as in 20 and 21 (R 1 = Me) led to significant decreases in inhibitory activities against all the CDKs tested with concomitant reductions in the cellular activities. Similarly, substitution of the pyrazolyl-N1-methyl group by a larger 2-morpholinoethyl as in 31 and 32 (R = 2-morpholinoethyl) had detrimental effects on both kinase and cellular inhibition (e.g., 32: K i (CDK2) = 0.252 µM and %GI(A2780) = 3 at 1 µM versus 17: K i (CDK2) = 0.011 µM and %GI(A2780) = 87 at 1 µM (not shown in Table 1)), thus signifying that diversification of the pyrazole ring by introduction of a larger substituent at the N1 or C3 position may not be tolerated. To explore the effect of regioisomerism regarding the pyrazole ring at the pyrimidinyl-C4 position on the CDK2 inhibition, 35 with a 1-methyl-1H-pyrazol-5-yl moiety was synthesized and found to be less active towards CDK2 than its 1-methyl-1H-pyrazol-4-yl counterpart 16 (% CDK2 inhibition = 64 and 97 (not shown in Table 1), respectively), suggesting the dependency of the kinase inhibition on the topology of the pyrazole ring at the pyrimidinyl-C4 position. The enzymatic inhibition percentage was determined using a luminescent ADP-Glo™ assay. Apparent inhibition constant (Ki) values were calculated using their corresponding half maximal inhibition (IC50) values and the appropriate Km(ATP) of each kinase as described in the experimental section. The antiproliferative activity was determined by 72 h 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays using A2780 cells. The asterisk (*) indicates % for either biochemical or cell-based data. Previously reported data for 1 [20] were included in the table for comparison purposes. CDK1B, CDK2E, CDK5p25, and CDK9T1 denote CDK1/cyclin B, CDK2/cyclin E, CDK5/p25, and CDK9/cyclin T1, respectively. Me: methyl. We next investigated the SAR around the other pyrazole ring occupying the pyrimidinyl-C4 position (i.e., R and R 1 ). As shown in Table 2, methylation of the pyrazolyl-C3 position as in 20 and 21 (R 1 = Me) led to significant decreases in inhibitory activities against all the CDKs tested with concomitant reductions in the cellular activities. Similarly, substitution of the pyrazolyl-N1-methyl group by a larger 2-morpholinoethyl as in 31 and 32 (R = 2-morpholinoethyl) had detrimental effects on both kinase and cellular inhibition (e.g., 32: Ki(CDK2) = 0.252 µM and %GI(A2780) = 3 at 1 µM versus 17: Ki(CDK2) = 0.011 µM and %GI(A2780) = 87 at 1 µM (not shown in Table 1)), thus signifying that diversification of the pyrazole ring by introduction of a larger substituent at the N1 or C3 position may not be tolerated. To explore the effect of regioisomerism regarding the pyrazole ring at the pyrimidinyl-C4 position on the CDK2 inhibition, 35 with a 1-methyl-1H-pyrazol-5yl moiety was synthesized and found to be less active towards CDK2 than its 1-methyl-1H-pyrazol-4-yl counterpart 16 (% CDK2 inhibition = 64 and 97 (not shown in Table 1), respectively), suggesting the dependency of the kinase inhibition on the topology of the pyrazole ring at the pyrimidinyl-C4 position. Having identified the dependence of the CDK2 inhibitory activity on the topology of the 1-methyl-1H-pyrazol-4-yl group at the pyrimidinyl-C4 position, it seemed logical to investigate if a similar dependence existed on the other pyrazole ring at the pyrimidinyl-C2-NH position. As a result, 5-chloro-4-(1-methyl-1H-pyrazol-4-yl)-N-(1Hpyrazol-5-yl)pyrimidin-2-amine 23 was synthesized. As Table 3 shows, the compound displayed greatly diminished CDK inhibitory activities, particularly with an 18-fold reduced CDK2 inhibition when compared to its 1H-pyrazol-4-yl counterpart 15 (K i = 0.090 and 0.005 µM, respectively). Consequently, 23 showed 47-fold less potent antiproliferative activity against A2780 cells than did 15 (GI 50 = 7.350 and 0.158 µM, respectively), indicating the dependency of the CDK2 inhibitory activity on the orientation of the 1H-pyrazolyl ring at the pyrimidinyl-C2-NH position.
Next, we investigated the effect of a related heteroaromatic system-thiazol-2-yl-at the pyrimidinyl-C2-NH position (i.e., 25) on the CDK and A2780 inhibitory activities. As shown in Table 3, the introduction of such a ring led to a significant reduction of the inhibitory activities toward all the CDKs tested as well as A2780 cells.
Taken together, compounds with chlorine at the pyrimidinyl-C5 position (R 2 = Cl) were generally more potent CDK2 inhibitors than their fluorine counterparts (R 2 = F), which may be attributed to a stronger hydrophobic interaction between the chlorine and Phe80. against A2780 cells than did 15 (GI50 = 7.350 and 0.158 µM, respectively), indicating the dependency of the CDK2 inhibitory activity on the orientation of the 1H-pyrazolyl ring at the pyrimidinyl-C2-NH position.
Next, we investigated the effect of a related heteroaromatic system-thiazol-2-yl-at the pyrimidinyl-C2-NH position (i.e., 25) on the CDK and A2780 inhibitory activities. As shown in Table 3, the introduction of such a ring led to a significant reduction of the inhibitory activities toward all the CDKs tested as well as A2780 cells.
Taken together, compounds with chlorine at the pyrimidinyl-C5 position (R 2 = Cl) were generally more potent CDK2 inhibitors than their fluorine counterparts (R 2 = F), which may be attributed to a stronger hydrophobic interaction between the chlorine and Phe80.

Docking Study on 15
To understand the likely binding mode of 15, the most potent CDK2 inhibitor in the series, with CDK2 and to provide a rationale for the observed potency, the compound was docked into the crystal structure of CDK2/cyclin E (PDB ID: 7KJS) using OEDocking 4.1.2.1 from OpenEye Scientific Software. The result revealed that 15 may form two hydrogen bonds with Leu83 in the ATP binding pocket of CDK2 ( Figure 3). Specifically, while the 2-amino group of the pyrimidine ring could develop a hydrogen bond with the carbonyl of Leu83, the pyrimidinyl-N1 is likely to form a second hydrogen bond with the NH of the same amino acid residue. This binding pose is similar to those of 1 or 14 in complex with CDK2/cyclin E ( Figure 2) and those of other inhibitors co-crystallized with CDK2 [15,27,29].

Docking Study on 15
To understand the likely binding mode of 15, the most potent CDK2 inhibitor in the series, with CDK2 and to provide a rationale for the observed potency, the compound was docked into the crystal structure of CDK2/cyclin E (PDB ID: 7KJS) using OEDocking 4.1.2.1 from OpenEye Scientific Software. The result revealed that 15 may form two hydrogen bonds with Leu83 in the ATP binding pocket of CDK2 ( Figure 3). Specifically, while the 2amino group of the pyrimidine ring could develop a hydrogen bond with the carbonyl of Leu83, the pyrimidinyl-N1 is likely to form a second hydrogen bond with the NH of the same amino acid residue. This binding pose is similar to those of 1 or 14 in complex with CDK2/cyclin E ( Figure 2) and those of other inhibitors co-crystallized with CDK2 [15,27,29].

Cellular Mechanistic Studies of 15
Given that A2780 and OVCAR5 were two of the most sensitive cancer cell lines to 15 in the above cancer cell panel screen, both ovarian cell lines were selected for cellular mechanistic studies. The effects of 15 on cell cycle progression and the level of phosphorylated Rb were determined by flow cytometry and Western blotting, respectively. In addition, the impacts of 15 on colony formation and apoptosis were analyzed by staining with crystal violet and annexin V/propidium iodide (PI), respectively. CYC065, a CDK2/9 inhibitor and a clinical candidate [30], was used as a control in all mechanistic studies conducted. The flow cytometric analysis (Figure 4; top) showed that the compound at 0.5 µM a the cells at the G2/M phase (about 52%) when compared to the untreated cells (23 increased subpopulation of S-phase cells was also observed with the treatment of 2 15 (26% versus 19% in untreated cells). At 0.5 µM, CYC065 caused the cell to arrest G1 (27% versus 0.3% in untreated cells) and S-phase (24% versus 19% in untreated  Similar cell-cycle effects on OVCAR5 cells were also observed (Figure 4, bottom). At the concentration of 0.5 µM, 15 resulted in a 39% increment in the subpopulation of G2/M cells when compared to the vehicle control. At 2 µM, the compound also showed a similar effect (53% G2/M-phase cells versus 22% in untreated cells) with an increased S-phase subpopulation (26% versus 19% in untreated cells). CYC065 caused a similar cell-cycle profile with an increased percentage of G2/M cells, but to a lesser extent (a 12% increase when compared to the vehicle control).

Effect of 15 on Apoptosis in Cancer Cells
To probe whether the antiproliferative activities of 15 were due to the induction of apoptosis, A2780 and OVCAR5 cells were incubated with 15 at concentrations of 0.5 and 2 µM for 48 h, and apoptosis was determined by annexin V and PI double staining. As Figure 5 shows, in A2780 cells, 15 induced 47% and 49% apoptosis at 0.5 and 2 µM, respectively, while only 1% was found in untreated cells. In contrast, CYC065 at 0.5 µM showed 73% apoptosis. In OVCAR5 cells, 15 caused 30% and 38% apoptosis at 0.5 and 2 µM, respectively, whereas 2% was observed in untreated cells. On the other hand, CYC065 induced 16% apoptosis. CYC065 induced more apoptosis in A2780 cells than did 15, and the opposite held true in OVCAR5 cells.
spectively, while only 1% was found in untreated cells. In contrast, CYC065 at showed 73% apoptosis. In OVCAR5 cells, 15 caused 30% and 38% apoptosis at 0. µM, respectively, whereas 2% was observed in untreated cells. On the other hand, C induced 16% apoptosis. CYC065 induced more apoptosis in A2780 cells than did the opposite held true in OVCAR5 cells.

Effect of 15 on Colony Formation
As Figure 6 illustrates, 15 inhibited the ability of three ovarian cancer ce (A2780, OVCAR5, and OV90) to form viable colonies at 0.5 and 2 µM. However, inhibition of colony formation was observed at 0.5 µM in OV90 cells. These result agreement with the observed antiproliferative effects of 15 (Tables 1 and 4). Sim CYC065 reduced the clonogenic growth of all three cell lines, with the lowest eff OVCAR5.

Effect of 15 on Colony Formation
As Figure 6 illustrates, 15 inhibited the ability of three ovarian cancer cell lines (A2780, OVCAR5, and OV90) to form viable colonies at 0.5 and 2 µM. However, partial inhibition of colony formation was observed at 0.5 µM in OV90 cells. These results are in agreement with the observed antiproliferative effects of 15 (Tables 1 and 4). Similarly, CYC065 reduced the clonogenic growth of all three cell lines, with the lowest effect on OVCAR5.

Western Blot Analysis
To explore whether 15 could inhibit its target kinases within cancer cells and to understand the molecular mechanism underlying its anticancer effects, OVCAR5 cells were incubated with 0.5 or 2 µM of 15 or 0.5 µM of CYC065 for 48 h, and the phosphorylation levels of the physiological substrates of CDK2, CDK5, and CDK9 (i.e., Rb, FAK, and RNAPol-II, respectively) were determined by Western blotting. As shown in Figure 7, 15 completely abolished the kinase activity of CDK2, CDK5, and CDK9 at 2 µM as demonstrated by the disappearance of p-Rb(Thr821), p-FAK(Ser732), and p-RNAPol-II(Ser2), respectively. Additionally, 15 also caused double-strand DNA breaks, showcased by the appearance of phosphorylated histone H2AX (p-H2AX), which is an independent marker of DNA damage. The DNA damage resulted in apoptotic cell death which was evidenced by the cleavage of caspase-3 and the downregulation of the anti-apoptotic protein MCL1.

Western Blot Analysis
To explore whether 15 could inhibit its target kinases within cancer cells and to understand the molecular mechanism underlying its anticancer effects, OVCAR5 cells were incubated with 0.5 or 2 µM of 15 or 0.5 µM of CYC065 for 48 h, and the phosphorylation levels of the physiological substrates of CDK2, CDK5, and CDK9 (i.e., Rb, FAK, and RNAPol-II, respectively) were determined by Western blotting. As shown in Figure 7, 15 completely abolished the kinase activity of CDK2, CDK5, and CDK9 at 2 µM as demonstrated by the disappearance of p-Rb(Thr821), p-FAK(Ser732), and p-RNAPol-II(Ser2), respectively. Additionally, 15 also caused double-strand DNA breaks, showcased by the appearance of phosphorylated histone H2AX (p-H2AX), which is an independent marker of DNA damage. The DNA damage resulted in apoptotic cell death which was evidenced by the cleavage of caspase-3 and the downregulation of the anti-apoptotic protein MCL1.
Molecules 2023, 28, x FOR PEER REVIEW 12 of 22 Figure 7. Western blot analysis of OVCAR5 cells treated with 15. Cells were exposed to 15 at 0.5 or 2 µM or to CYC065 at 0.5 µM for 48 h. DMSO diluent was employed as a control. Experiments were performed twice, and representative data are presented.

Conclusion
The bioisosteric replacement of the phenylsulfonamide moiety of our previous lead compound 1 with pyrazole-derived groups led to the discovery of a new chemotype of CDK2 inhibitors, namely 4-(1-methyl-1H-pyrazol-4-yl)-N-(1H-pyrazol-4-yl)pyrimidin-2amines. Among them, 14 and 15 inhibited CDK2 at single-digit nanomolar concentrations (Ki = 0.007 and 0.005 µM, respectively). The SAR analysis revealed that N-alkylation or topological change of the pyrazol-4-yl ring at the pyrimidinyl-C2-NH position had a detrimental effect on both the CDK2 inhibition and the antiproliferative activity. Similar ef- Figure 7. Western blot analysis of OVCAR5 cells treated with 15. Cells were exposed to 15 at 0.5 or 2 µM or to CYC065 at 0.5 µM for 48 h. DMSO diluent was employed as a control. Experiments were performed twice, and representative data are presented.

Conclusions
The bioisosteric replacement of the phenylsulfonamide moiety of our previous lead compound 1 with pyrazole-derived groups led to the discovery of a new chemotype of CDK2 inhibitors, namely 4-(1-methyl-1H-pyrazol-4-yl)-N-(1H-pyrazol-4-yl)pyrimidin-2amines. Among them, 14 and 15 inhibited CDK2 at single-digit nanomolar concentrations (K i = 0.007 and 0.005 µM, respectively). The SAR analysis revealed that N-alkylation or topological change of the pyrazol-4-yl ring at the pyrimidinyl-C2-NH position had a detrimental effect on both the CDK2 inhibition and the antiproliferative activity. Similar effects were also observed upon derivatization of the other pyrazol-4-yl moiety at the pyrimidinyl-C4 position. Mechanistically, the antiproliferative activity of 15 was shown to be associated with inhibition of the phosphorylation of Rb, FAK, and RNAPol-II, cell cycle arrest at the S and G2/M phases, and induction of apoptosis. Thus, 15 can be considered a lead compound that could be further optimized into a potent and selective CDK2 inhibitor with potential anticancer activity.

Chemistry
Chemicals and solvents were obtained from commercial sources and were used without further purification. A CEM Discover SP and Explorer 48/72/96 microwave system (CEM corporation, Matthews, NC, USA) controlled by the Synergy TM software was used for microwave-assisted synthesis. The progression of a reaction was monitored by thin-layer chromatography using Merck 60 F 254 silica gel-precoated aluminum plates and visualized by UV irradiation (254 nm). Flash column chromatography was carried out using a glass column packed with silica gel (230-400 mesh). An AB SCIEX TripleTOF ® 5600 mass spectrometer (Concord, Ontario, Canada) with ESI ionization mode was used to determine the high-resolution mass spectra. 1 H and 13 C NMR spectra were determined by a Bruker Avance III HD spectrometer (Faellanden, Switzerland) at 500.16 and 125.76 MHz, respectively, and were analyzed using the Bruker Topspin 4.1 program. Chemical shifts are reported in parts per million (ppm) and referenced to 1 H signals of residual nondeuterated solvents and 13 C signals of the deuterated solvents. The multiplicity of 1 H NMR signals was reported as: s = singlet, d = doublet, and t = triplet. Coupling constant (J) values are reported in Hz. A Shimadzu Prominence UltraFast Liquid Chromatograph System (Kyoto, Japan) equipped with a CBM-20A communications bus module, a DGU-20A5R degassing unit, an LC-20AD liquid chromatograph pump, an SIL-20AHT auto-sampler, an SPD-M20A photodiode array detector, a CTO-20A column oven, and a Phenomenex Kinetex 5u C18 100A 250 mm × 4.60 mm column was used to determine the purity of the final compounds. All the final compounds used for biological assays had a purity greater than 95%. Method A (gradient 5%-95% methanol containing 0.1% formic acid (FA) over 20 min at a flow rate of 1 mL/min, followed by 95% methanol containing 0.1% FA over 5 min) and method B (gradient 5%-95% MeCN containing 0.1% FA over 20 min at a flow rate of 1 mL/min, followed by 95% MeCN containing 0.1% FA over 5 min) were used for analytic HPLC. The melting points (m.p.) of final compounds were determined using a Stuart SMP10 melting point apparatus and are uncorrected.

General Synthetic Procedure B: Buchwald-Hartwig Amination [33]
A mixture of a 2-chloropyrimidine (1.20 eqv.), an arylamine (1.00 eqv.), Cs 2 CO 3 (2.50 eqv.), Pd 2 (dba) 3 (0.05 eqv.), and Xantphos (0.10 eqv.) in 1,4-dioxane was degassed, purged with nitrogen (3×), and subjected to microwave irradiation at 140 • C for 1 h. The reaction mixture was concentrated under reduced pressure, and the residue was partitioned between water and ethyl acetate (1:2, v/v). The aqueous layer was extracted with ethyl acetate (3×), and the organic layer and extracts were combined, washed with brine, dried over Na 2 SO 4 , and filtered. The filtrate was concentrated under reduced pressure, and the residue was purified by flash column chromatography (silica gel) to give the desired product. 1  2-(4-Nitro-1H-pyrazol-1-yl)acetamide (12): To a solution of 4-nitro-1H-pyrazole (10) (1.70 g, 15.0 mmol) in acetonitrile (150 mL) were added K 2 CO 3 (4.10 g, 30.0 mmol) and 2-bromoacetamide (11) (2.10 g, 15.0 mmol), and the mixture was heated at 60 • C overnight, cooled down, filtered, and washed with acetonitrile (50 mL). The filtrate and washing were concentrated under reduced pressure, and the residue was triturated with diethyl ether:petroleum ether (2:1) to yield 12 as a white solid (2.40 g, 95%). 1 H NMR (DMSO-d 6 ):  (2.40 g, 14.1 mmol) in methanol (100 mL) was added 10% palladium on carbon (150 mg, 141 µmol), and the reaction mixture was stirred at room temperature under hydrogen overnight. The reaction product was filtered through a pad of Celite, and the residue was washed with methanol (100 mL). The filtrate and washing were combined and concentrated under reduced pressure to afford 13c as a dark brown solid, which was used in the next step without further purification (1.80 g, 90%). 1 13    The luminescent ADP-Glo™ assay, which measures the amount of ADP formed in a kinase reaction as described previously [34], was used to determine the kinase inhibition profile of the compounds. Initially, the kinase inhibitory activity of the compounds was evaluated at 1 µM, and the compounds that displayed ≥ 80% kinase inhibition were selected for the determination of their apparent inhibition constant (K i ) values. K i values were calculated from their corresponding IC 50 values (the concentrations of inhibitors that inhibit the kinase activity by 50%) and the Michaelis-Menten constant (K m ) of ATP for each kinase using the Cheng-Prusoff equation as previously stated [34]. In brief, 1 µL of each concentration of the test compounds was added to the respective wells, and to it were added 1 µL of the kinases and their respective substrates, diluted to the desired concentration in kinase reaction buffer (40 nM tris base pH 7.5, 20 mM MgCl 2 , 0.4 µM dithiothreitol, and 0.1 mg/mL bovine serum albumin), and the pH was adjusted by adding 1 µL of a two-fold diluted kinase reaction buffer. To initiate the kinase activity, 1 µL of ATP solution (the concentration varies depending on the type of kinase under investigation) was added to each well, and the mixture was incubated at room temperature (20 to 22 • C) for 30-60 min depending on the kinase being tested. The reaction was then halted, 5 µL of the ADP-Glo reagent was added to each well (removing any unreacted ATP), and it was incubated at room temperature for 40 min. Then, to each well, 10 µL of the kinase detection reagent containing luciferin was added, which was incubated once more at room temperature in the dark for 30 min and the luminescence was measured by an EnVision ® multi-label plate reader with an integration time of 1.5 s per well. Positive controls (the wells without the inhibitor) and negative or blank controls (the wells without the inhibitor and the enzyme, but with 1 × kinase reaction buffer) were also determined in 2.5% DMSO.
The percentage of kinase inhibition was determined using the following equation: % Residual kinase activity = Luminescence in the presence of inhibitor − Luminescence of blank control Luminescence in the absence of inhibitor − Luminescence of blank control × 100% %kinase inhbition = 100% − %Residual kinase activity To determine the IC 50 values, each compound was serially diluted in 2.5% DMSO, and the kinase inhibitory activities at eight different concentrations were determined and analyzed by a four-parameter logistic non-linear regression model using GraphPad Prism version 7.03 (San Diego, CA, USA). K m is defined as the concentration of ATP or the substrate required to produce 50% of the maximal reaction rate in the kinase reaction. Therefore, since we used the concentration of ATP and the substrate at their K m values, K i simply equals IC 50 /2 assuming these inhibitors are ATP-competitive based on the Cheng-Prusoff equation: K i = IC 50 /(1 + ([ATP]/K m (ATP))) where [ATP] is the ATP concentration used for the IC 50 determination and K m (ATP) for each kinase was determined experimentally.

Cell Viability Assays
The MTT (Life Technologies, Mulgrave, VIC, Australia) and resazurin (Sigma-Aldrich) assays were used to determine the antiproliferative activities of the compounds initially at 1 µM against the cancer cell lines stated above. Those compounds that inhibited the growth by ≥80% were subjected to the determination of GI 50 using nonlinear regression analysis.

Cell Cycle Analysis
The effect of the compounds on cell cycle distribution was evaluated following a previously reported method [34]. Briefly, six-well plates (Sigma-Aldrich, NSW, Australia) were seeded with about 1-2 × 10 5 cells/well in 2 mL growth medium, incubated at 37 • C and 5% CO 2 overnight, and treated with various concentrations of the test compounds and incubated for 48 h. The wells were collected by trypsinization, transferred into fluorescenceactivating cell sorting (FACS) tubes, washed with 1 × phosphate-buffered saline (1 × PBS) twice, fixed in 70% (v/v) ice-cold ethanol for 15 min, and centrifuged at 300× g for 5 min. The cell pellets were then collected, stained with 200 µL of the cell cycle solution (comprising 50.0 µg/mL PI, 0.1 mg/mL RNase A, and 0.05% Triton X-100), and incubated for 1.5 h at room temperature in the dark. Finally, the DNA content was determined by a flow cytometer (CytoFLEX) and analyzed by CytExpert 2.1 software (Beckman Coulter, Brea, CA, USA).

Apoptosis Assay
Annexin V-fluorescein isothiocyanate (FITC) apoptosis detection kit I (consisting of a ten-time stock solution of binding buffer, FITC-annexin V, and PI staining solution) was used to determine the effect of the compounds on cancer cell survival as previously reported [34]. Briefly, six-well plates (Sigma-Aldrich, NSW, Australia) were seeded with about 1-2 × 10 5 cells/well in 2 mL growth medium, incubated at 37 • C and 5% CO 2 overnight, treated with various concentrations of the test compounds, and incubated for 48 h. The wells were collected by trypsinization, transferred into FACS tubes, washed twice with 1 × PBS, and centrifuged at 300× g for 5 min. The pellets were then suspended in 100 µL of 1 × binding buffer and incubated in the dark with 3 µL of PI and 3 µL of FITC-conjugated annexin V solutions for 15 min. Finally, 200 µL of binding buffer was added to each sample and analyzed by the CytoFLEX cytometer. The extent of apoptotic cell death was quantified as the sum of the percentages of early apoptotic cells (annexin V-positive and PI-negative) and late apoptotic cells (both annexin-V-and PI-positive) and presented graphically as a contour diagram relative to the untreated control.

Colony Formation
The effect of compound 15 on colony formation was investigated as previously reported [34]. Briefly, ovarian cancer cells (i.e., A2780, OVCAR5, and OV90) seeded in six-well plates were adhered for 6 h and treated with 15. The medium that contained the compound was then changed every 72 h. Ten days after treatment, the cells were washed with PBS, fixed, and stained with crystal violet staining solution (0.05% crystal violet, 1% formaldehyde, 1% PBS, and 1% ethanol), and colonies were quantified using ImageJ software (National Institute of Health, Maryland, USA), with data presented as percentages of the control cells.

Western Blot Analysis
Western blot analysis of the compounds was carried out as previously reported [35]. The antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA).