Phenothiazine- and Carbazole-Cyanochalcones as Dual Inhibitors of Tubulin Polymerization and Human Farnesyltransferase

In the search for innovative approaches to cancer chemotherapy, a chemical library of 49 cyanochalcones, 1a-r, 2a-o, and 3a-p, was designed as dual inhibitors of human farnesyltransferase (FTIs) and tubulin polymerization (MTIs) (FTIs/MTIs), two important biological targets in oncology. This approach is innovative since the same molecule would be able to interfere with two different mitotic events of the cancer cells and prevent these cells from developing an emergency route and becoming resistant to anticancer agents. Compounds were synthesized by the Claisen–Schmidt condensation of aldehydes with N-3-oxo-propanenitriles under classical magnetic stirring and under sonication. Newly synthesized compounds were screened for their potential to inhibit human farnesyltransferase, tubulin polymerization, and cancer cell growth in vitro. This study allowed for the identification of 22 FTIs and 8 dual FTIs/MTIs inhibitors. The most effective molecule was carbazole-cyanochalcone 3a, bearing a 4-dimethylaminophenyl group (IC50 (h-FTase) = 0.12 µM; IC50 (tubulin) = 0.24 µM) with better antitubulin activity than the known inhibitors that were previously reported, phenstatin and (-)-desoxypodophyllotoxin. The docking of the dual inhibitors was realized in both the active site of FTase and in the colchicine binding site of tubulin. Such compounds with a dual inhibitory profile are excellent clinical candidates for the treatment of human cancers and offer new research perspectives in the search for new anti-cancer drugs.


Introduction
The majority of new cancer cases detected worldwide each year are treated with anti-cancer drugs. Unfortunately, many cancer mutations become resistant to these drugs. An alternative consists of using cocktails of drugs acting on various biological targets of interest in oncology to overcome resistant cells. This type of approach can give good results, but it often leads to a large increase in side effects. Advances in the field of cell biology have allowed for the identification of new targets for the treatment of cancer, opening up new therapeutic perspectives. Another alternative to avoid the use of multiple anticancer drugs to stop cancer cell growth proliferation is to use a single compound acting on two different biological targets [1,2]. Our approach concerns the inhibition of two aspects of cell division occurring at two very different times in the life of a cancer cell: one involving farnesyltransferase and the other involving tubulin. Designed compounds target farnesyltransferase, a zinc metalloenzyme, and also inhibit tubulin polymerization. Tubulin is involved in cell proliferation due to its ability to polymerize and form microtubules, key components of the cytoskeleton. This protein is the target of a large panel of small molecules that interfere with the dynamics of its polymerization or depolymerization. Most of them bind to the laulimalide, maytansine, taxane/epothilone, vinca alkaloid, and colchicine sites [3]. By interacting with tubulin and microtubules, the tubulin polymerization inhibitors block cells in mitosis; this results in their accumulation in the G2/M phase of the cancer cell cycle. For the design of our potential dual inhibitors, two known strong inhibitors of tubulin polymerization, combretastatin A-4 (CA-4) (I, Figure 1) and phenstatin (II, Figure 1), were considered as reference molecules. Most of the modifications previously described in the structure of CA-4 and phenstatin involved either the ethylenic or carbonyl bridge or the methoxyphenol B ring. However, the 3,4,5-trimethoxyphenyl group (ring A) has long been kept intact, as it was considered essential for cytotoxic activity as well as the inhibition of tubulin polymerization. Our group previously described that a completely different A ring consisting of a phenothiazine unit can successfully replace the 3,4,5-trimethoxyphenyl of phenstatin and provide effective tubulin polymerization (e.g., compounds III and IV, Figure 1) [4,5].
The other target of molecules from this study was human protein farnesyltransferase (FTase). FTase is a heterodimeric metalloenzyme that belongs to the protein prenyl transferase family and is composed of two subunits: α (48 kDa) and β (45 kDa). Farnesylation is a post-translational modification occurring in several cell signaling proteins such as small GTPases, including the oncogenic Ras proteins that play a fundamental role in cancer cell growth and division [6]. FTase catalyzes the transfer of a farnesyl group (C 15 ) from farnesyl pyrophosphate or farnesyl diphosphate (FPP) to the free thiol group of a cysteine residue embedded in the C-terminal CaaX motif of proteins where C is a cysteine, a is an aliphatic amino acid, and X is a serine, a methionine, an alanine, or a glutamine [7]. Preventing the farnesylation process may constitute an approach in the treatment of cancers, and, therefore, farnesyltransferase inhibitors (FTIs) were developed for anticancer therapy, and diverse compounds with druglike properties are available [7][8][9][10][11]. The use of farnesyltransferase inhibitors was disappointing in clinical trials for cancer treatment. Indeed, even if FTase is completely inhibited, a bypass is always possible for the cancerous cell. This alternative path involves a protein very similar to FTase, which is geranylgeranyltransferase I (GGTase-I) [12]. However, proving the effectiveness of dual compounds FTIs/MTIs, which are inhibitors of FTase (FTIs) and of tubulin polymerization (MTIs), may lead to an innovative approach for the design of new anti-cancer compounds.
Several associations between FTIs and MTIs were described in the literature. The association of lonafarnib (SCH66336, compound V, Figure 1) with paclitaxel resulted in an enhanced cytotoxic effect in ovarian cancer cells in vitro and in vivo [13]. The same association of lonafarnib/paclitaxel (Taxol) or lonafarnib/docetaxel (Taxotere) is synergistic in vivo in NCI-460 lung cancer cells, and lonafarnib could also be used by patients who develop resistance to taxanes. Another FTI (FTI-277, compound VI, Figure 1) displayed synergistic effect with paclitaxel or docetaxel in cells resistant to paclitaxel [14].
Based on all these previous findings presented above, a new series of anticancer agents, dual inhibitors of farnesyltransferase and tubulin polymerization FTIs/MTIs, were developed in this study (compounds 1a-r, 2a-o, and 3a-p, Figure 2). These compounds are not prodrugs or a combination of two known specific inhibitors, rather they are original structures rationally designed from previous studies and the literature data. These compounds, 1a-r, 2a-o, and 3a-p (Figure 2), share a common bridge between the A and B rings, which is a cyanochalcone group. Another particularity of these target molecules is phenothiazine (compounds 1a-r), 2-methylthiophenothiazine (compounds 2a-o), and carbazole (compounds 3a-p) as the A ring. The literature analysis allowed for the identification of some similar phenothiazine-cyanochalcones that are nitric oxide (NO) inhibitors, preventing diseases mediated by lipid peroxidation (compound VII, Figure 2) [15], or display antibacterial activity against the Gram-positive bacteria Bacillus subtilis (compounds VIII and  [4,5] and of human farnesyltransferase (Lonafarnib (SCH66336) (V) [13] and FTI-277 (VI) [14]) investigated as anticancer compounds.
Based on all these previous findings presented above, a new series of anticancer agents, dual inhibitors of farnesyltransferase and tubulin polymerization FTIs/MTIs, were developed in this study (compounds 1a-r, 2a-o, and 3a-p, Figure 2). These compounds are not prodrugs or a combination of two known specific inhibitors, rather they are original structures rationally designed from previous studies and the literature data. These compounds, 1a-r, 2a-o, and 3a-p (Figure 2), share a common bridge between the A and B rings, which is a cyanochalcone group. Another particularity of these target molecules is phenothiazine (compounds 1a-r), 2-methylthiophenothiazine (compounds 2a-o), and carbazole (compounds 3a-p) as the A ring. The literature analysis allowed for the identification of some similar phenothiazine-cyanochalcones that are nitric oxide (NO) inhibitors, preventing diseases mediated by lipid peroxidation (compound VII, Figure 2) [15], or display antibacterial activity against the Gram-positive bacteria Bacillus subtilis (compounds VIII and IX, Figure 2) or the Gram-negative bacteria Escherichia coli (compound VIII, Figure 2)

Synthetic Strategy
The cyanochalcones (1a-r, 2a-o, and 3a-p) of this study were prepared by the Claisen-Schmidt condensation of the corresponding N-3-oxo-propanenitriles 4a, 4b, and 5 and (hetero)aryl aldehydes 8-36 (Scheme 1). N-3-oxo-propanenitriles 4a, 4b, and 5 were obtained by treating phenothiazine 6a, 2-methylthiophenothiazine 6b, and carbazole 7 with the mixed anhydride of acetic acid and cyanoacetic acid obtained in situ from an equimolar mixture of cyanoacetic acid and acetic anhydride (Scheme 1). The resulting N-3-oxopropanenitriles 4a and 5 were previously described, and their physicochemical characterization corresponded to that reported in the literature [24,25]. Next, the key condensation reaction was conducted under classical magnetic stirring, and, based on the effective results obtained previously for the Claisen-Schmidt condensation of (hetero)aryl ketones with (hetero)aryl aldehydes [11], also under sonication of the mixture instead of classical magnetic stirring. In the classical magnetic stirring procedure (procedure A, Scheme 1), piperidine and glacial acetic acid were used as catalysts and ethanol or acetonitrile as solvents. The reaction media were stirred under reflux for the phenothiazine derivatives and at rt for the carbazole derivatives. The sonication method (procedure B, Scheme 1) used LiOH as a base and ethanol as a solvent. Both procedures, applied to the corresponding N-3-oxo-propanenitriles 4a, 4b, and 5 and aldehydes 8-36, allowed for the obtainment of a large panel of 49 cyanochalcones in medium to high yields (40-87%) (see Table 1 and Charts 1 and 2). All compounds were obtained as E-isomers. No trace of Z-isomers was detected in this series. In order to conduct the greenest and least energy-consuming synthetic method possible for obtaining cyanochalcones, the synthesis of the same compound was carried out following the two procedures. This resulted in comparable yields, but the reaction time was significantly reduced from hours to minutes by sonication. As an example, the phenothiazine derivative 1m was obtained in 74% yield under sonication and in 67% yield under magnetic stirring, but after only 2 min (procedure B) against 24 h (procedure A) (Chart 1). Consequently, a major part of 2-methylthiophenothiazine-cyanochalcones 2a, 2c-f, 2h-j, and 2o was further synthesized under sonication in less than 90 s (Table 1). Under the standard conditions of procedure A, the synthesis of these latter

Synthetic Strategy
The cyanochalcones (1a-r, 2a-o, and 3a-p) of this study were prepared by the Claisen-Schmidt condensation of the corresponding N-3-oxo-propanenitriles 4a, 4b, and 5 and (hetero)aryl aldehydes 8-36 (Scheme 1). N-3-oxo-propanenitriles 4a, 4b, and 5 were obtained by treating phenothiazine 6a, 2-methylthiophenothiazine 6b, and carbazole 7 with the mixed anhydride of acetic acid and cyanoacetic acid obtained in situ from an equimolar mixture of cyanoacetic acid and acetic anhydride (Scheme 1). The resulting N-3-oxo-propanenitriles 4a and 5 were previously described, and their physicochemical characterization corresponded to that reported in the literature [24,25]. Next, the key condensation reaction was conducted under classical magnetic stirring, and, based on the effective results obtained previously for the Claisen-Schmidt condensation of (hetero)aryl ketones with (hetero)aryl aldehydes [11], also under sonication of the mixture instead of classical magnetic stirring. In the classical magnetic stirring procedure (procedure A, Scheme 1), piperidine and glacial acetic acid were used as catalysts and ethanol or acetonitrile as solvents. The reaction media were stirred under reflux for the phenothiazine derivatives and at rt for the carbazole derivatives. The sonication method (procedure B, Scheme 1) used LiOH as a base and ethanol as a solvent. Both procedures, applied to the corresponding N-3-oxo-propanenitriles 4a, 4b, and 5 and aldehydes 8-36, allowed for the obtainment of a large panel of 49 cyanochalcones in medium to high yields (40-87%) (see Table 1 and Charts 1 and 2). All compounds were obtained as E-isomers. No trace of Z-isomers was detected in this series. In order to conduct the greenest and least energy-consuming synthetic method possible for obtaining cyanochalcones, the synthesis of the same compound was carried out following the two procedures. This resulted in comparable yields, but the reaction time was significantly reduced from hours to minutes by sonication. As an example, the phenothiazine derivative 1m was obtained in 74% yield under sonication and in 67% yield under magnetic stirring, but after only 2 min (procedure B) against 24 h (procedure A) (Chart 1). Consequently, a major part of 2-methylthiophenothiazine-cyanochalcones 2a, 2c-f, 2h-j, and 2o was further synthesized under sonication in less than 90 s (Table 1). Under the standard conditions of procedure A, the synthesis of these latter compounds would have required refluxing

Biological Evaluation
Synthesized compounds were further evaluated in vitro on the two biological targets: farnesyltransferase and tubulin. The results of these biological evaluations are presented in Table 2. Potential inhibitors were first screened at a high concentration (100 µM), and only compounds that generally inhibited more than 60% of the proteins were selected for IC50 calculation. Dimethylsulfoxide (DMSO) was used as a negative reference, while phenstatin (II) and (-)-desoxypodophyllotoxin were positive references for the tubulin polymerization assay, and FTI-276 was the positive reference for the evaluation on human FTase (Table 2). FTI-277 (VI, Figure 1) is a prodrug of FTI-276, the latter being more affine to FTase than parent FTI-277. Interestingly, a large part of the tested cyano-chalcones inhibited FTase and presented a moderate to potent effect (IC50 values ranging from tens of micromoles (e.g., phenothiazine 1q: IC50 (h-FTase) = 44.85 µM) on submicromolar (e.g., carbazole 3a: IC50 (h-FTase) = 0.12 µM) concentrations (Table 2). In the tubulin polymerization assay, fewer compounds were found to be tubulin polymerization compounds, but two of these inhibitors significantly outperformed the potencies of the positive references phenstatin (II) and (-)-desoxypodophyllotoxin (e.g., compare phenothiazine 1l: IC50 (tubulin) = 0.71 µM or carbazole 3a: IC50 (tubulin) = 0.24 µM with the positive reference phenstatin (II), IC50 (tubulin) = 3.43 µM, and with (-)-desoxypodophyllotoxin: IC50 (tubulin) = 1.76 µM, Table 2). Moreover, four carbazole-cyanochalcones 3b, 3i, 3j, and 3l displayed similar inhibitory activity to that of the reference inhibitors (Table 2). Considering both biological evaluations, it can be concluded that several compounds inhibited the two biological targets of interest and may be considered as dual FTIs/MTIs. This is the case for one phenothiazine-cyanochalcone, 1l, and seven carbazole-cyanochalcones, 3a, 3b, 3d, 3e, 3i, 3j, and 3l. The carbazole-cyanochalcone 3a was the most potent inhibitor discovered in this study, inhibiting both targets and presenting submicromolar IC50 (0.12 µM for h-FTase and 0.24 µM for tubulin polymerization, respectively). The corresponding phenothiazine 1c was not active (Table 2). Now, looking at the chemical structures of phenothiazine 1l and carbazole 3j, they have the same classical B-ring as CA-4 (I, Figure 1) and phenstatin (II, Figure 1). This ring seems important to the biological activity against tubulin, especially in the phenothiazine series. Its replacement by other substituents in the phenothiazine cyanochalcones (1a-k, 1m-r) abolished the inhibitory effect (Table 2). On the contrary, in the series of carbazoles, the replacement of the classical 3′-hydroxy-4′-methoxyphenyl B ring was tolerated. The 4-dimethylaminophenyl group in compound 3a had the best modulation in the current study. Moreover, the reverse substitution of the classical B ring (3′-methoxy-4′-hydroxyphenyl in compound 3l instead of 3′-hydroxy-4′-methoxyphenyl Chart 2. Compounds with carbazole unit 3a-p synthesized in this study (procedure A).

Biological Evaluation
Synthesized compounds were further evaluated in vitro on the two biological targets: farnesyltransferase and tubulin. The results of these biological evaluations are presented in Table 2. Potential inhibitors were first screened at a high concentration (100 µM), and only compounds that generally inhibited more than 60% of the proteins were selected for IC 50 calculation. Dimethylsulfoxide (DMSO) was used as a negative reference, while phenstatin (II) and (-)-desoxypodophyllotoxin were positive references for the tubulin polymerization assay, and FTI-276 was the positive reference for the evaluation on human FTase (Table 2). FTI-277 (VI, Figure 1) is a prodrug of FTI-276, the latter being more affine to FTase than parent FTI-277. Interestingly, a large part of the tested cyano-chalcones inhibited FTase and presented a moderate to potent effect (IC 50 values ranging from tens of micromoles (e.g., phenothiazine 1q: IC 50 (h-FTase) = 44.85 µM) on submicromolar (e.g., carbazole 3a: IC 50 (h-FTase) = 0.12 µM) concentrations (Table 2). In the tubulin polymerization assay, fewer compounds were found to be tubulin polymerization compounds, but two of these inhibitors significantly outperformed the potencies of the positive references phenstatin (II) and (-)-desoxypodophyllotoxin (e.g., compare phenothiazine 1l: IC 50 (tubulin) = 0.71 µM or carbazole 3a: IC 50 (tubulin) = 0.24 µM with the positive reference phenstatin (II), IC 50 (tubulin) = 3.43 µM, and with (-)-desoxypodophyllotoxin: IC 50 (tubulin) = 1.76 µM, Table 2). Moreover, four carbazole-cyanochalcones 3b, 3i, 3j, and 3l displayed similar inhibitory activity to that of the reference inhibitors (Table 2). Considering both biological evaluations, it can be concluded that several compounds inhibited the two biological targets of interest and may be considered as dual FTIs/MTIs. This is the case for one phenothiazine-cyanochalcone, 1l, and seven carbazole-cyanochalcones, 3a, 3b, 3d, 3e, 3i, 3j, and 3l. The carbazole-cyanochalcone 3a was the most potent inhibitor discovered in this study, inhibiting both targets and presenting submicromolar IC 50 (0.12 µM for h-FTase and 0.24 µM for tubulin polymerization, respectively). The corresponding phenothiazine 1c was not active (Table 2). Now, looking at the chemical structures of phenothiazine 1l and carbazole 3j, they have the same classical B-ring as CA-4 (I, Figure 1) and phenstatin (II, Figure 1). This ring seems important to the biological activity against tubulin, especially in the phenothiazine series. Its replacement by other substituents in the phenothiazine cyanochalcones (1a-k, 1m-r) abolished the inhibitory effect (Table 2). On the contrary, in the series of carbazoles, the replacement of the classical 3 -hydroxy-4 -methoxyphenyl B ring was tolerated. The 4-dimethylaminophenyl group in compound 3a had the best modulation in the current study. Moreover, the reverse substitution of the classical B ring (3 -methoxy-4 -hydroxyphenyl in compound 3l instead of 3 -hydroxy-4 -methoxyphenyl in compound 3j) conserved antitubulin activity. The 3 -fluoro-4 -methoxyphenyl substitution in compound 3i was also tolerated, while the substitution of the 3 -fluoro by a 3 -chloro in compound 3h resulted in the loss of the biological activity ( Table 2). The suppression of the 4 -methoxy group in carbazole 3d dramatically decreased the antitubulin potential (compare carbazole 3d (IC 50 (tubulin) = 69.8 µM) with 3j (IC 50 (tubulin) = 2.92 µM), Table 2). The 4 -nitro substitution in carbazole 3e conserved an inhibitory potential (IC 50 (tubulin) = 10.45 µM), Table 2) but was significantly reduced compared to that in carbazole 3a. Table 2. Inhibitory activities of studied molecules on human farnesyltransferase and tubulin polymerization in vitro.

Entry
Compound To better visualize the distribution of the cyanochalcones from this study into selective human FTase inhibitors or dual inhibitors, their clustering was realized using POWER BI Desktop software version 2.117.984.0. (Figure 3). Eight dual FTIs/MTIs (cluster in pink,  Figure 3) and twenty-two inhibitors of human farnesyltransferase (cluster in green, Figure 3) were found. The dual inhibitors were generally carbazole cyanochalcones (3a, 3b, 3d, 3e,  3i, 3j, and 3l), except for phenothiazine 1l, which also displayed dual inhibitory potential. The phenothiazine cyanochalcones displayed an FTIs profile.

Molecular Docking of Dual FTIs/MTIs Inhibitors
The docking study was next realized on the dual FTIs/MTIs inhibitors identified in this study in the active site of FTase and in the colchicine binding site of tubulin to understand their binding mode. All the data obtained for the docking of the eight dual inhibitors are available in the Supplementary Materials section ( Figure S2). The structure of the human FTase was obtained from its complexed X-ray crystal structure in the RCSB Protein Data Bank (1LD7) with FPP and the inhibitor molecule described by Bell et al. [26]. The flexible docking of FTIs into the enzyme active site was performed using GOLD 5.1 [27]. The binding site was defined by a 10 Å sphere around the cocrystallized ligand of 1LD7, and 30 poses were generated for each compound using GoldScore as the scoring function. The solutions were selected by checking the superimposition of the poses, keeping the most representative of the largest clusters. The protocol used for the docking of the selected molecules in the tubulin binding site (colchicine site) was realized as previously reported [4].

Docking on FTase
All the investigated compounds fit well in the pocket (Figures 4 and S2). The largest number, consisting of compounds 1l, 3b, 3e, 3i, 3j, and 3l, has their tricyclic group toward the entry of the pocket, and all form interactions with the zinc ion. Tyr 601 is involved in a  hydrogen bond with 1l, 3j, and 3l (Figures 4a,d and S2).  The tricyclic part of compounds 3a (Figure 4b) and 3d (Figure 4c) is superimposed, though more toward Trp 407 than the other molecules. Moreover, only 3d can interact with zinc, as the fluorobenzene moiety of 3a is also oriented toward Trp 407, stabilizing the compound by a distant and not optimal double stacking with it.

Docking in the Tubulin Binding Site (Colchicine Site)
The reference, phenstatin II, binds to the backbone of Ala 732, while mostly being in a hydrophobic region (Figure 5f). Compounds 3b and 3a (Figure 5c) and the 40% highest score of compound 1l (Figure 5a) all form a cluster closer to the entry of the binding site, where they can form a hydrogen bond with Asn 682. Compounds 3d, 3e (Figure 5d), and 3i all superimpose well in a conformation deeper in the pocket, with the cyano able to form a hydrogen bond with the Ser 168 of α tubuline. The tricyclic part of compounds 3a (Figure 4b) and 3d (Figure 4c) is superimposed, though more toward Trp 407 than the other molecules. Moreover, only 3d can interact with zinc, as the fluorobenzene moiety of 3a is also oriented toward Trp 407, stabilizing the compound by a distant and not optimal double stacking with it.

Docking in the Tubulin Binding Site (Colchicine Site)
The reference, phenstatin II, binds to the backbone of Ala 732, while mostly being in a hydrophobic region (Figure 5f). Compounds 3b and 3a (Figure 5c) and the 40% highest score of compound 1l (Figure 5a   A third cluster is formed by compounds 3j and 3l (Figure 5e) and the 60% lowest score of compound 1l (Figure 5b), with a better occupation of the deepest part of the pocket than phenstatin but fully lacking any hydrogen bond and counting on their hydrophobic fitting to stay in the binding site.

Conclusions
In this study, a chemical collection of 49 cyanochalcones, 1a-r, 2a-o, and 3a-p, decorated with phenothiazine, 2-methylthiophenothiazine, and carbazole rings was designed and synthesized by the Claisen-Schmidt condensation of the corresponding N-3-oxopropanenitriles and aldehydes. The synthetic procedure was realized either under classical magnetic stirring and heating or under sonication of the medium. The ultrasound-assisted condensation allowed for a reduction in the reaction time from hours to minutes, especially for the synthesis of phenothiazine cyanochalcones. The therapeutic strategy described in this report was used to obtain dual FTIs/MTIs inhibitors. This approach is innovative, since the same molecule would be able to interfere with two different mitotic events of cancer cells and prevent these cells from developing an emergency route and becoming resistant to anticancer agents. Synthesized compounds were evaluated in vitro on human farnesyltransferase, on tubulin polymerization, and on the NCI-60 cancer cell lines panel. Phenothiazine derivatives proved to be inhibitors of human FTase, while carbazole derivatives displayed dual inhibition of FTase and tubulin polymerization. Of interest, phenothiazine cyanochalcone 1l and carbazole cyanochalcone 3a displayed better antitubulin activity than that of the known inhibitors previously reported: phenstatin II and (-)-desoxypodophyllotoxin. Carbazole derivatives were more active than the phenothiazine analogues. This study allowed for the identification of 22 FTIs and 8 dual FTIs/MTIs inhibitors. The most effective molecule was carbazole-cyanochalcone 3a bearing a 4-dimethylaminophenyl group (IC 50 (h-FTase) = 0.12 µM; IC 50 (tubulin) = 0.24 µM). The docking of the dual inhibitors was realized both in the active site of FTase and in the colchicine binding site of tubulin and allowed for the visualization of their binding modes. The biological evaluation of these promising dual inhibitors in several cancer cell lines and the evaluation of their pharmacokinetic parameters will be realized in due course. Such compounds with a dual inhibitory profile are excellent clinical candidates for the treatment of human cancers and offer new research perspectives in the search for new anti-cancer drugs.

Materials and Methods for Synthesis and Characterizations
Starting materials were commercially available and were used without further purification (suppliers: Carlo Erba Reagents S.A.S., Val de Reuil, France, Thermo Fisher Scientific Inc., Illkirch-Graffenstaden, France, and Sigma-Aldrich Co., Saint-Quentin-Fallavier, France). Ultrasound-mediated reactions were realized using Q700S apparatus (QSonica, LLC, Newton, MA, USA) and CL-334 model probe. Melting points were measured on a MPA 100 OptiMelt ® apparatus (Stanford Research Systems, Sunnyvale, CA, USA) and a KRÜSS Optronic KSP1N apparatus (A.KRÜSS Optronic GmbH, Hamburg, Germany) and were uncorrected. Nuclear magnetic resonance (NMR) spectra were acquired at 500 MHz for 1 H NMR and at 125 MHz for 13 C NMR on a Bruker Avance III spectrometer (Bruker, Mannheim, Germany) and at 400 MHz for 1 H NMR and at 100 MHz for 13 C NMR on a Varian 400-MR spectrometer (Varian, Les Ulis, France) with tetramethylsilane (TMS) as internal standard, at room temperature (RT). All spectra were realized using deuterated solvents (CDCl 3 99.8%D + 0.03% TMS V/V or DMSO-d 6 99.8%D + 0.03% TMS V/V), purchased from Eurisotop, Saint-Aubin, France. The calibration was realized using TMS pic as the 0.00 ppm value in the registered spectra. Chemical shifts (δ) were expressed in ppm relative to TMS. Splitting patterns were designed: s, singlet; d, doublet; dd, doublet of doublets; t, triplet; td, triplet of doublets; q, quadruplet; quint, quintuplet; m, multiplet; sym m, symmetric multiplet; br s, broaden singlet; br t, broaden triplet. Coupling constants (J) were reported in hertz (Hz). Thin-layer chromatography (TLC) was realized on Macherey Nagel silica gel plates with fluorescent indicator and were visualized under a UV lamp A mixture of cyanoacetic acid (2 equiv.) and acetic anhydride (2 equiv.) was stirred at 50-80 • C. After complete solubilization, phenothiazine derivative 6a or 6b or carbazole 7 (1 equiv.) was added, and the mixture was stirred at 100 • C for 1 h. The precipitate formed was filtered and purified by recrystallization from ethanol to provide pure product 4a, 4b, or 5.
The physicochemical characterization of compounds 4a and 5 corresponded to that previously described in the literature [24,25].