Synthesis and Biological Activity of 3-(Heteroaryl)quinolin-2(1H)-ones Bis-Heterocycles as Potential Inhibitors of the Protein Folding Machinery Hsp90

In the context of our SAR study concerning 6BrCaQ analogues as C-terminal Hsp90 inhibitors, we designed and synthesized a novel series of 3-(heteroaryl)quinolin-2(1H), of types 3, 4, and 5, as a novel class of analogues. A Pd-catalyzed Liebeskind–Srogl cross-coupling was developed as a convenient approach for easy access to complex purine architectures. This series of analogues showed a promising biological effect against MDA-MB231 and PC-3 cancer cell lines. This study led to the identification of the best compounds, 3b (IC50 = 28 µM) and 4e, which induce a significant decrease of CDK-1 client protein and stabilize the levels of Hsp90 and Hsp70 without triggering the HSR response.


Introduction
The 90-kDa heat shock protein (Hsp90) has emerged recently as a promising therapeutic target for the treatment of cancer [1][2][3][4][5] and other diseases [6,7]. As a chaperone protein, Hsp90 is evolved in the conformational maturation, folding, stabilization, activation, and degradation of over 400 client proteins in healthy cells as well as in cancerous cells which are directly associated with all hallmarks of cancer [8][9][10][11]. This Hsp90 chaperone cycle depends on the ATPase activity. ATP binding to the N-terminal domain (NTD) and hydrolysis by Hsp90 drive a conformational cycle necessary for chaperone function [12][13][14]. The binding of ATP to each monomer shifts Hsp90 to a "closed" formation that can bind, fold, and activate client proteins [15,16]. Thus, inhibition of Hsp90 function results in the simultaneous interruption of many signal transduction pathways which are pivotal to tumor progression and survival.
Several structurally distinct Hsp90 inhibitors that target the ATP binding pocket are currently being evaluated for anticancer activity in numerous Phase II and several Phase III clinical trials. However, they are ineffective over time due to the compensatory mechanism involving the induction of a heat shock response. The expression of chaperones Hsp27, Hsp70, Hsp40, and Hsp90 increases [17][18][19], leading to undesirable chemoprotective effects [20][21][22]. Clinical resistance has been attributed to this chemoprotective effect, and dosage increases to overcome resistance are not a viable option due to toxicity. These results continue to motivate the pursuit of alternative strategies for modulating heat shock protein complexes [23][24][25].
An alternative molecular mechanism of inhibition is through binding to the C-terminal domain of Hsp90. The CTD has been implicated biochemically as the site of a possible  In this context, we previously reported a novel series of simplified 3 amido-quinolin-2-one analogues related to Nvb as a class of highly potent hsp90 inhibitors [32][33][34][35][36]. From the structure-activity relationship (SAR) studies, 6BrCaQ ( Figure 1) [37,38] was identified as a very promising C-terminal Hsp90 inhibitor displaying an antiproliferative activity LC 50 of 5-50 µM [39,40] against various cancer cell lines (MCF7, MDA MB231, Caco2, IGROV, ISHIKAWA, PC3, and HT29 cells). Further studies on its mode of action revealed that 6BrCaQ manifests downregulation of several Hsp90 client proteins (HER2, Raf-1 and cdk-4), induces a high apoptosis level in MCF-7 breast cancer cell line and PC3. In addition, encapsulated in liposomes, 6BrCaQ exerted an improved in vitro activity on breast cancer cells (MDA-MB-231) and displays an in vivo anti-tumor activity on an orthotopic breast cancer model in nude mice [40].
More recently, we demonstrated that conjugation of 6BrCaQ with the cationic head triphenylphosphonium (TPP) leads to the conjugate 6BrCaQ-C10-TPP for the targeting of the mitochondrial heat shock protein TRAP1. Hence, 6BrCaQ-C10-TPPdisplays an antiproliferative activity with mean GI 50 values at a nanomolar level in a diverse set of human cancer cells (GI 50 = 0.008-0.30 µM) including MDA-MB-231, HT-29, HCT116, K562 and PC-3 cancer cell lines. This study showed that this compound 6BrCaQ-C 10 -TPP induces a significant mitochondrial membrane disruption and interferes with TRAP1 function in colon carcinoma cells without inducing the heat-shock response HSF1 [41].
On the other hand, Blagg and co-workers reported during their various SAR studies that the coumarin analogue (I) (Figure 1), which possesses a benzothiophen heterocycle at the C3 position, is able to induce the cell death with an IC 50 of 0.98 µM against SkBr3 cell lines [42]. Inspired by this study and the promising activity displayed by 6BrCaQ, we proposed to design a news series of quinolinone based heterocycle analogues ( Figure 1) in which the amide function of 6BrCaQ will be replaced by various heterocycles, including benzoxazoles, benzothiazoles, indoles, benzimidazoles, and purines, in the aim to better understand the SAR in this novel series. In this article, the synthesis and biological evaluation of analogues of type 3, 4, and 5 are described.
In the pursuit of our SAR-study, we considered the possibility of functionalization of the thiomethyl group attached to 6′-position of the purine motif in the derivative 3e. This motif is found in a series of hsp90 inhibitors, such as PU-H71 disclosed in Figure 1. If succeeded, this approach would provide a fast and easy access to a small library of more sophisticated purine-quinolinone analogues. We have rationalized that an additional aromatic group in molecule 1 would modify the intercalation ability due to changes in the planarity and in the extension of conjugation.
After a detailed survey about this topic, only few examples of this derivatization of thiopurines were found in the literature. One of the methodologies available to introduce diversity in this particular position is the scarcely explored Liebeskind-Srogl coupling [44]. This approach exploits the pseudo-halogen character of CH3Sgroup (thio-organyl in general) as partner in a Suzuki-like coupling reaction, involving an arylboronic acid under palladium catalysis in the presence of a copper salt [45,46]. We decided to start our investigations with the adoption of two approaches. The first one involved the coupling of thiopurinoquinolone 3e with p-methoxyphenyl-boronic acid under PdDppfCl2·CH2Cl2 catalysis and conventional heating [47], while the second method chosen to promote the desired transformation employed Pd(OAc)2 and 1,10phenanthroline under microwave irradiation [48]. To our delight, both approaches were capable to afford the expected product 4a, in 61% and 49% yield, respectively (Scheme 2). In order to explore the scope of this synthetic transformation, we decided to use the first method with a sort of arylboronic acid. This selection was based on the rationale of the electronic and steric effects exerted by the chosen substituents. The reactions proceeded Scheme 1. Synthetic strategy to target 3-(heteroaryl)quinolin-2(1H)-ones 3.
In the pursuit of our SAR-study, we considered the possibility of functionalization of the thiomethyl group attached to 6 -position of the purine motif in the derivative 3e. This motif is found in a series of hsp90 inhibitors, such as PU-H71 disclosed in Figure 1. If succeeded, this approach would provide a fast and easy access to a small library of more sophisticated purine-quinolinone analogues. We have rationalized that an additional aromatic group in molecule 1 would modify the intercalation ability due to changes in the planarity and in the extension of conjugation.
After a detailed survey about this topic, only few examples of this derivatization of thiopurines were found in the literature. One of the methodologies available to introduce diversity in this particular position is the scarcely explored Liebeskind-Srogl coupling [44]. This approach exploits the pseudo-halogen character of CH 3 S-group (thio-organyl in general) as partner in a Suzuki-like coupling reaction, involving an arylboronic acid under palladium catalysis in the presence of a copper salt [45,46]. We decided to start our investigations with the adoption of two approaches. The first one involved the coupling of thiopurinoquinolone 3e with p-methoxyphenyl-boronic acid under PdDppfCl 2 ·CH 2 Cl 2 catalysis and conventional heating [47], while the second method chosen to promote the desired transformation employed Pd(OAc) 2 and 1,10-phenanthroline under microwave irradiation [48]. To our delight, both approaches were capable to afford the expected product 4a, in 61% and 49% yield, respectively (Scheme 2). In order to explore the scope of this synthetic transformation, we decided to use the first method with a sort of arylboronic acid. This selection was based on the rationale of the electronic and steric effects exerted by the chosen substituents. The reactions proceeded smoothly, affording the expected products 4a-h with yields ranging from 39% to 89% (Scheme 2). smoothly, affording the expected products 4a-h with yields ranging from 39% to 89% (Scheme 2).
Taking advantage of this synthetic procedure, we decided to examine the scope of Liebeskind-Srogl reaction with anilines (Scheme 3). The introduction of a nitrogen atom at the 6′-position of the purine ring would imply its overall transformation into a nucleic acid analogue, i.e., an [(N-phenyl)adenine] motif. The possibility to have an adenine ring attached to a quinolone nucleus would increase its biological resemblance [49,50].
It is interesting to denote that this reaction can also be reached by an alternative twostep procedure involving the initial thioether-sulfone oxidation, followed by a nucleophilic heteroaromatic substitution with the appropriate amine. This protocol has been previously employed by Piguel and coworkers during the synthesis of 6,8,9-purinederivatives [47]. To the best of our knowledge, the introduction of an amine to the purine ring in 6′-position through a Liebeskind-Srogl reaction has never been reported in the literature.
We started our investigations by adapting the arylsulfide amination protocol described by the group of Murakami [51]. Unfortunately, only the degradation of starting material 3e was detected. Several conditions were assayed, including the palladium source, ligand, base, and microwave heating [52][53][54]. To our surprise, with a slight modification of the previously used PdDppf·Cl2/CuTC protocol, the reaction proceeded until completeness, giving the desired product 5a in 67% yield (Scheme 3). It is important to observe that the presence of CuTC and the Cs2CO3 base were mandatory to accomplish the expected transformation.
Under this condition, we succeed to generate the coupling product from 3,4,5trimethoxyaniline (5b, 51%). Unfortunately, however, the reaction of 3e with aniline, benzyamine, butylamine, and pyrrolidine could not be driven to completeness and the expected product could not be separated from the starting material by current chromatographic purification conditions (CC and preparative TLC). Scheme 2. Synthetic strategy to target 3-(purino)-quinolin-2(1H)-ones 4a-h.
Taking advantage of this synthetic procedure, we decided to examine the scope of Liebeskind-Srogl reaction with anilines (Scheme 3). The introduction of a nitrogen atom at the 6 -position of the purine ring would imply its overall transformation into a nucleic acid analogue, i.e., an [(N-phenyl)adenine] motif. The possibility to have an adenine ring attached to a quinolone nucleus would increase its biological resemblance [49,50].

Biological Evaluation of Quinolones Analogues
Antiproliferative Activity Upon completion of their syntheses, the in vitro activity of quinolone derivatives 3af, 4a-h, and 5a,b was evaluated by their growth-inhibitory potency in three cancer cell lines. At first, the viability of the synthesized compounds was examined with the MDA-MB-231 MCF-7 breast cancer cell line at concentrations of 10 µM, 15 µM, and 25 µM. Prostate cancer PC-3 cells and human fetal lung fibroblast MRC-5 cell lines were also subjected to this series of compounds at a unique concentration of 15 µM. The It is interesting to denote that this reaction can also be reached by an alternative twostep procedure involving the initial thioether-sulfone oxidation, followed by a nucleophilic heteroaromatic substitution with the appropriate amine. This protocol has been previously employed by Piguel and coworkers during the synthesis of 6,8,9-purine-derivatives [47].
To the best of our knowledge, the introduction of an amine to the purine ring in 6 -position through a Liebeskind-Srogl reaction has never been reported in the literature.
We started our investigations by adapting the arylsulfide amination protocol described by the group of Murakami [51]. Unfortunately, only the degradation of starting material 3e was detected. Several conditions were assayed, including the palladium source, ligand, base, and microwave heating [52][53][54]. To our surprise, with a slight modification of the previously used PdDppf·Cl 2 /CuTC protocol, the reaction proceeded until completeness, giving the desired product 5a in 67% yield (Scheme 3). It is important to observe that the presence of CuTC and the Cs 2 CO 3 base were mandatory to accomplish the expected transformation.
Under this condition, we succeed to generate the coupling product from 3,4,5-trimethoxyaniline (5b, 51%). Unfortunately, however, the reaction of 3e with aniline, benzyamine, butylamine, and pyrrolidine could not be driven to completeness and the expected product could not be separated from the starting material by current chromatographic purification conditions (CC and preparative TLC).

Biological Evaluation of Quinolones Analogues
Antiproliferative Activity Upon completion of their syntheses, the in vitro activity of quinolone derivatives 3a-f, 4a-h, and 5a,b was evaluated by their growth-inhibitory potency in three cancer cell lines. At first, the viability of the synthesized compounds was examined with the MDA-MB-231 MCF-7 breast cancer cell line at concentrations of 10 µM, 15 µM, and 25 µM. Prostate cancer PC-3 cells and human fetal lung fibroblast MRC-5 cell lines were also subjected to this series of compounds at a unique concentration of 15 µM. The quantification of cell survival in these cell lines was established using MTS assays after 72 h exposure (Table 1), and GI 50 values were estimated at the concentration required to produce 50% inhibition ( Table 2).
As shown in Table 1, all these series of analogues induced a significant decrease of the cell viability in MDA-MB-231 cells in a concentration depend manner. At 10 µM concentration the viability percentage of MDA-MB-231 cells decreased until less than 47% under 3a and 5a exposure (Table 1). In addition, increasing the concentration at 25 µM, analogues 3a, 3b, 4g, and 4h importantly affect the growth of MDA-MB-231 cells (~30% survival), clearly demonstrating the bioactivity potential of these compounds.
Then, the cytotoxicity activity was examined with two other cancer cell lines: PC-3 cells and human fetal lung fibroblast MRC-5. As shown in Table 1, almost all the reported compounds do not present any effect against MRC-5 cell lines (>82% survival) at 15 µM concentration, except compounds 4d, 4e, and 4g, which induce a slight effect on the growth of MRC-5 cells (71% to 79% survival). In contrast, PC-3 cells seem to be more sensitive to these derivatives than MRC-5 cells, as we can see in Table 1. Upon exposure of these cell lines at 15 µM concentration, compounds 3b, 3h, 3e, and 4e were able to decrease the cell viability in PC-3 cells until 56%.
Then, the growth inhibitory activities against PC-3 prostate cancer cell line were measured for the selected 3-heteroaryl-quinolin-2(1H)-one derivatives 3a-e. All the compounds shown in Table 2 display an estimated GI 50 ranging between 28 and 48 µM. Of the selected derivatives, 3b showed a significant ability to inhibit cell growth and was the most cytotoxic (GI 50 = 28 µM) against the PC-3 prostate cancer cell lines.
To provide additional evidence of the growth inhibitory activity manifested by the derivatives, the most active compounds 3a-e, 4e, and 5b were evaluated for their ability to induce the degradation of Hsp90-dependent client protein Cdk4, the most widely studied molecular signature indicative of Hsp90 blockade.                 Cell viability effect of 3a-f, 4a-h and 5a,b derivatives against MDA-MB-231, PC-3 and MRC-5 cell lines measured through cell metabolic activity (MTS-based assay).

Cell viability [%] [a]
MDA Then, the growth inhibitory activities against PC-3 prostate cancer cell line were measured for the selected 3-heteroaryl-quinolin-2(1H)-one derivatives 3a-e. All the compounds shown in Table 2 display an estimated GI50 ranging between 28 and 48 µM. Of the selected derivatives, 3b showed a significant ability to inhibit cell growth and was   Cell viability effect of 3a-f, 4a-h and 5a,b derivatives against MDA-MB-231, PC-3 and MRC-5 cell lines measured through cell metabolic activity (MTS-based assay).

Cell viability [%] [a]
MDA Then, the growth inhibitory activities against PC-3 prostate cancer cell line were measured for the selected 3-heteroaryl-quinolin-2(1H)-one derivatives 3a-e. All the compounds shown in Table 2 display an estimated GI50 ranging between 28 and 48 µM. Of the selected derivatives, 3b showed a significant ability to inhibit cell growth and was Then, the growth inhibitory activities against PC-3 prostate cancer cell line were measured for the selected 3-heteroaryl-quinolin-2(1H)-one derivatives 3a-e. All the compounds shown in Table 2 display an estimated GI50 ranging between 28 and 48 µM. Of the selected derivatives, 3b showed a significant ability to inhibit cell growth and was Then, the growth inhibitory activities against PC-3 prostate cancer cell line were measured for the selected 3-heteroaryl-quinolin-2(1H)-one derivatives 3a-e. All the compounds shown in Table 2 display an estimated GI50 ranging between 28 and 48 µM. Of the selected derivatives, 3b showed a significant ability to inhibit cell growth and was   Cell viability effect of 3a-f, 4a-h and 5a,b derivatives against MDA-MB-231, PC-3 and MRC-5 cell lines measured through cell metabolic activity (MTS-based assay).

Cell viability [%] [a]
MDA Then, the growth inhibitory activities against PC-3 prostate cancer cell line were measured for the selected 3-heteroaryl-quinolin-2(1H)-one derivatives 3a-e. All the compounds shown in Table 2 display an estimated GI50 ranging between 28 and 48 µM. Of the selected derivatives, 3b showed a significant ability to inhibit cell growth and was  As depicted in Figure 2, the cyclin-dependant kinase CdK4 was degraded following treatment with 3a-e, 4e, and 5b. The GAPDH protein was not affected by the tested compounds, indicating the selective degradation of hsp90-dependent clients. CDK-4 level was more decreased by compounds 3b and 4e at a concentration of 15 µM. One can note that the anti-proliferative activity of 3b (IC 50 = 28 µM, Table 2) and 4e correlate well with the concentration needed to induce Hsp90/CDK-4 client protein degradation. 3c 37 3e 38 [a] GI50 is the concentration of compound needed to reduce cell growth by 50% following 72 h cell treatment with the tested drug.
To provide additional evidence of the growth inhibitory activity manifested by the derivatives, the most active compounds 3a-e, 4e, and 5b were evaluated for their ability to induce the degradation of Hsp90-dependent client protein Cdk4, the most widely studied molecular signature indicative of Hsp90 blockade.
As depicted in Figure 2, the cyclin-dependant kinase CdK4 was degraded following treatment with 3a-e, 4e, and 5b. The GAPDH protein was not affected by the tested compounds, indicating the selective degradation of hsp90-dependent clients. CDK-4 level was more decreased by compounds 3b and 4e at a concentration of 15 µM. One can note that the anti-proliferative activity of 3b (IC50 = 28 µM, Table 2) and 4e correlate well with the concentration needed to induce Hsp90/CDK-4 client protein degradation.
Hsp90 N-terminal inhibitors induce a Heat-shock-response by releasing a transcription factor (HSF1) of the genes of Hsp27, Hsp70, and Hsp90. This increase in transcription leads to opposition, to apoptosis, and thus resistance to treatment.
It is important to check that the levels of these proteins are not increasing with our compounds. We showed by Western blot that 3a-e, 4e, and 5b stabilize the levels of Hsp90 and Hsp70 without triggering the HSR. This result was already observed with 6-BrCaQ in PC-3 cell lines, as we reported previously: liposomal 6-BrCaQ stabilized levels of Hsp70 and decreased the level of Hsp90 [40]. Effects of quinolone analogues 3a-e, 4e and 5b on HSP90 machinery protein levels and on CDK-4 stability. PC-3 cells were grown and exposed to Hsp90 inhibitors (3a-e, 4e and 5b, 15 µM) as described in Experimental section for 72 h and cell lysates were analyzed by Western blotting with regard to the levels of CDK-4, Hsp90α/β and Hsp70. NT corresponds to untreated cells; D, DMSO-treated cells were used as controls, GADPH level is used for control in protein loading on gels.

Conclusion
In summary, we have designed and synthesized a new series of 3-heteroarylquinolin-2(1H)-one derivatives as potential Hsp90 inhibitors. During this study, we developed a Pd-catalyzed Liebeskind-Srogl cross-coupling reaction between an SMecontaining quinolinyl-purine derivative and various aryl boronic acids. We reported also, for the first time, that anilines may be used as nucleophilic partners during this coupling. From these SAR studies, 3a-e, 4e, and 5b were found to display the strongest cell viability effect against MDA-MB 231 and PC-3 cancer cell lines. In addition, compounds 3b and 4e were found to be able to induce a significant decrease of CDK-1 client protein and stabilize the levels of Hsp90 and Hsp70 without triggering the HSR response.

General Experimental Methods
NT D 3a 3b 3c 3e 4e 5b HSP90 α/β CDK-4 Hsp70 GAPDH Figure 2. Effects of quinolone analogues 3a-e, 4e and 5b on HSP90 machinery protein levels and on CDK-4 stability. PC-3 cells were grown and exposed to Hsp90 inhibitors (3a-e, 4e and 5b, 15 µM) as described in Experimental section for 72 h and cell lysates were analyzed by Western blotting with regard to the levels of CDK-4, Hsp90α/β and Hsp70. NT corresponds to untreated cells; D, DMSO-treated cells were used as controls, GADPH level is used for control in protein loading on gels.
Hsp90 N-terminal inhibitors induce a Heat-shock-response by releasing a transcription factor (HSF1) of the genes of Hsp27, Hsp70, and Hsp90. This increase in transcription leads to opposition, to apoptosis, and thus resistance to treatment.
It is important to check that the levels of these proteins are not increasing with our compounds. We showed by Western blot that 3a-e, 4e, and 5b stabilize the levels of Hsp90 and Hsp70 without triggering the HSR. This result was already observed with 6-BrCaQ in PC-3 cell lines, as we reported previously: liposomal 6-BrCaQ stabilized levels of Hsp70 and decreased the level of Hsp90 [40].

Conclusions
In summary, we have designed and synthesized a new series of 3-heteroaryl-quinolin-2(1H)-one derivatives as potential Hsp90 inhibitors. During this study, we developed a Pdcatalyzed Liebeskind-Srogl cross-coupling reaction between an SMe-containing quinolinylpurine derivative and various aryl boronic acids. We reported also, for the first time, that anilines may be used as nucleophilic partners during this coupling. From these SAR studies, 3a-e, 4e, and 5b were found to display the strongest cell viability effect against MDA-MB 231 and PC-3 cancer cell lines. In addition, compounds 3b and 4e were found to be able to induce a significant decrease of CDK-1 client protein and stabilize the levels of Hsp90 and Hsp70 without triggering the HSR response.