Biological Evaluation of Arylsemicarbazone Derivatives as Potential Anticancer Agents

Fourteen arylsemicarbazone derivatives were synthesized and evaluated in order to find agents with potential anticancer activity. Cytotoxic screening was performed against K562, HL-60, MOLT-4, HEp-2, NCI-H292, HT-29 and MCF-7 tumor cell lines. Compounds 3c and 4a were active against the tested cancer cell lines, being more cytotoxic for the HL-60 cell line with IC50 values of 13.08 μM and 11.38 μM, respectively. Regarding the protein kinase inhibition assay, 3c inhibited seven different kinases and 4a strongly inhibited the CK1δ/ε kinase. The studied kinases are involved in several cellular functions such as proliferation, migration, cell death and cell cycle progression. Additional analysis by flow cytometry revealed that 3c and 4a caused depolarization of the mitochondrial membrane, suggesting apoptosis mediated by the intrinsic pathway. Compound 3c induced arrest in G1 phase of the cell cycle on HL-60 cells, and in the annexin V assay approximately 50% of cells were in apoptosis at the highest concentration tested (26 μM). Compound 4a inhibited cell cycle by accumulation of abnormal postmitotic cells at G1 phase and induced DNA fragmentation at the highest concentration (22 μM).


Chemistry
The aryl semicarbazone derivatives 3a-3m were obtained by the classical condensation reaction between suitable aryl and (hetero)aryl aldehydes and semicarbazide in the presence of catalytic hydrochloric acid in ethanol (Scheme 1). Selective one-pot N 2 -alkylation of 5-nitro-2-furfuraldehyde semicarbazone 3m was carried out by nucleophilic substitution on butyl bromide using basic conditions (K 2 CO 3 ) in DMF as previously described in the literature by Brondani et al. [22].
Pharmaceuticals 2018, 11, x FOR PEER REVIEW 3 of 23 panel of 10 protein kinases. Compounds 3c and 4a showed cytotoxicity for all tested cell lines, particularly on HL-60, so both were chosen for further studies of possible mechanisms of action.

Chemistry
The aryl semicarbazone derivatives 3a-3m were obtained by the classical condensation reaction between suitable aryl and (hetero)aryl aldehydes and semicarbazide in the presence of catalytic hydrochloric acid in ethanol (Scheme 1). Selective one-pot N 2 -alkylation of 5-nitro-2-furfuraldehyde semicarbazone 3m was carried out by nucleophilic substitution on butyl bromide using basic conditions (K2CO3) in DMF as previously described in the literature by Brondani et al. [22].
The introduction of such a butyl chain would allow us to check the influence of such a lipophilic appendage on biological activities examined within the framework of the study. Indeed, the other analogues of the series remained unsubstituted at the level of this nitrogen.

Cytotoxicity Assays
The cytotoxicity assessment was performed against three human cancer cell lines in suspension (K562, HL-60 and MOLT-4) and four adhered (HEp-2, NCI-H292, HT-29 and MCF-7). Initially, all compounds 3a-3m and 4a) were tested in a single concentration of 25 μg/mL. Compounds that exhibited a percentage of inhibition greater than 75% in at least three cell lines were considered active. Among the 14 evaluated derivatives, seven (3c-3h, 3m and 4a) met this criterion and had their IC50 values determined after 72 h of incubation. As shown in Table 1, the derivatives showed different cytotoxicity values.
Only compounds 3c (2-hydroxynaphthyl) and 4a (N 2 -butyl-5-nitro-2-furfuraldehyde semicarbazone) showed antiproliferative activity against all the tested cell lines, whereas 3a-b, bearing activating groups (electron-donating groups) and heterocyclic derivatives 3i-l were inactive. Compound 3m presented high toxicity on MCF-7 (IC50 = 8.56 ± 0.01) and low cytotoxic activity on HL-60, with IC50 =24.33 ± 0.16. In addition, cytotoxic activity against some cancer cell lines was observed for halogenated and alkylated compounds 3d-h. The introduction of such a butyl chain would allow us to check the influence of such a lipophilic appendage on biological activities examined within the framework of the study. Indeed, the other analogues of the series remained unsubstituted at the level of this nitrogen.

Cytotoxicity Assays
The cytotoxicity assessment was performed against three human cancer cell lines in suspension (K562, HL-60 and MOLT-4) and four adhered (HEp-2, NCI-H292, HT-29 and MCF-7). Initially, all compounds 3a-3m and 4a) were tested in a single concentration of 25 µg/mL. Compounds that exhibited a percentage of inhibition greater than 75% in at least three cell lines were considered active. Among the 14 evaluated derivatives, seven (3c-3h, 3m and 4a) met this criterion and had their IC 50 values determined after 72 h of incubation. As shown in Table 1, the derivatives showed different cytotoxicity values.
Compound 3l, which bears a simple furan, was inactive, while its nitro derivative showed some activity. One possible mechanism of the cytotoxic action of 3m could be the bioreduction of the nitro group to highly reactive radical intermediates and reactive oxygen species (ROS) generation. This is a mechanism already well described in the anticancer activity of other nitro compounds [34]. However, more specific experiments should be conducted to prove this hypothesis. Interestingly, the counterpart 4a remained active after the alkylation of 3m (Table 1). We next decided to determine the lipophilicity of each tested compound. Indeed, lipophilicity is one of the fundamental factors for the biological activity of drugs because it is directly linked to the degree of permeability of the cell membranes and how much the molecule will be available to interact with the intended biological target(s) [35]. Moreover, solubility was calculated to verify whether compounds were sufficiently water soluble to allow a relevant interpretation for cytotoxicity assays. ClogD 7.4 and solubility values are shown in Table 2.       Overall, the predicted solubility obtained for the synthesized compounds were moderate for 3c, 3f, 3h and 4a (S values between 0.01 and 0.06 mg/mL) and high for all other molecules (S values higher than 0.06 mg/mL).
Analyzing the aromatic core of the arylsemicarbazones, it was observed that ClogD 7.4 values span values ranging from −0.07 to 2.87. Notably lipophilicity for compounds 3c-h and 4a, higher than 1.5 (i.e. 1.54-2.87), could be considered as a suitable parameter for the active compounds to display enhanced pharmacokinetic properties.
After discussing the cellular activity against cancer cell lines and the importance of some physicochemical parameters, the data allowed us to identify compounds 3c and 4a as the most promising derivatives. These derivatives displayed cytotoxic activity against all the tested cell lines and the best results were against the HL-60 cell line. Compound 3c showed IC 50 values ranging from 13.08 to 56.27 µM, while 4a had IC 50s between 11.38 to 37.89 µM, being more cytotoxic for all the panel. In order to evaluate the cytotoxic potential in normal cells, all the derivatives were tested against Peripheral Blood Mononuclear Cells (PBMCs) and the most promising compounds were again compounds 3c and 4a, which offered more selectivity against the cancer cells. The selectivity indexes (SI = IC 50 PBMC/ IC 50 cancer cells) of derivatives 3c and 4a for the cell line with the lowest IC 50 (HL-60) were 6.2 and 4.5, respectively.

Protein Kinases
Data expressed as mean ± SD of at least three assays. NT: no tested.
Compound 3c was the most active kinase inhibitor, with IC 50 values ranging from 0.073-5.1 µM, except towards HASPIN, CK1 and GSK3 kinases. Compound 3c showed inhibitory activity against seven protein kinases and the highest activity was against DYRK1A. The other counterparts remained inactive in this assay. Compound 4a exhibited a selective submicromolar IC 50 of 0.76 µM against the CK1δ/ε isoforms. More specifically, in comparison with other compounds, the presence of a H-donor group in position 2 of the naphthalene ring (compound 3c) could display efficient interactions with the tested kinases.

Molecular Modeling
For the docking study, three kinases were chosen according to the results of the in vitro assays: tyrosine phosphorylation regulated kinase 1A (DYRK1A), in which compound 3c was active in the nanomolar range, casein kinase 1 (CK1δ/ε), agianst which compound 4a was active, and CDK2 for its importance in the cell cycle. Several CDK2 inhibitors have been developed and some of them (including roscovitine, CYC065, dinaciclib, AT7519, milciclib) have undergone clinical evaluation [36]. On the other hand, with dual specificity, DYRK1A plays a key role in Down syndrome and Alzheimer's disease. In addition, DYRK1A has been shown to be expressed ubiquitously, and its functional role in cancer is still largely obscure, however several compounds caused inhibition of this kinase in HL-60 cells, decreasing its proliferation in vitro [37].
In order to elucidate how the 3c molecule interacts with CDK2, a detailed docking study was conducted, also including compound 3g for comparison reasons. Figure 1A illustrates the docking pose for 3c and Figure 1B illustrates the docking pose for 3g. As can be seen in Figure 1, the molecular reasons for the greater stability (greater docking score) or affinity of 3c ligand in the active site of CDK2 can be explained mainly by the presence of important intermolecular interactions, in particular one hydrogen bond that the hydroxyl group in position 2 of the naphthalene ring of compound 3c additionally establishes with the residue ASP-145 (3.0 Å), when compared with molecule 3g ( Figure 1B). A detailed comparison between the docking results obtained for these two ligands can be found in Table 4. The fact that the molecule 3c has a greater docking score in comparison with the molecule 3g, i.e., a greater in silico affinity for the target, correlates with the higher in vitro activity of the molecule 3c (see Table 3), or the higher inhibition capacity of the molecule 3c on this target (CDK2). importance in the cell cycle. Several CDK2 inhibitors have been developed and some of them (including roscovitine, CYC065, dinaciclib, AT7519, milciclib) have undergone clinical evaluation [36]. On the other hand, with dual specificity, DYRK1A plays a key role in Down syndrome and Alzheimer's disease. In addition, DYRK1A has been shown to be expressed ubiquitously, and its functional role in cancer is still largely obscure, however several compounds caused inhibition of this kinase in HL-60 cells, decreasing its proliferation in vitro [37].
In order to elucidate how the 3c molecule interacts with CDK2, a detailed docking study was conducted, also including compound 3g for comparison reasons. Figure 1A illustrates the docking pose for 3c and Figure 1B illustrates the docking pose for 3g. As can be seen in Figure 1, the molecular reasons for the greater stability (greater docking score) or affinity of 3c ligand in the active site of CDK2 can be explained mainly by the presence of important intermolecular interactions, in particular one hydrogen bond that the hydroxyl group in position 2 of the naphthalene ring of compound 3c additionally establishes with the residue ASP-145 (3.0 Å), when compared with molecule 3g ( Figure  1B). A detailed comparison between the docking results obtained for these two ligands can be found in Table 4. The fact that the molecule 3c has a greater docking score in comparison with the molecule 3g, i.e., a greater in silico affinity for the target, correlates with the higher in vitro activity of the molecule 3c (see Table 3), or the higher inhibition capacity of the molecule 3c on this target (CDK2). Additionally, docking of compound 3c was made on two of the six kinases, and a comparison with compound 3g was done. This study supported the fact that 3c inhibits DYRK1A well and better than 3g (see Table 3). Figure 2A illustrates the docking pose for 3c and Figure 2B illustrates the docking pose for 3g. The molecular reasons for the greater stability of the 3c ligand in the active site of DYRK1A can be explained mainly by the presence of two additional hydrogen bonds that the hydroxyl group in position 2 of the naphthalene ring of compound 3c establishes with the residues GLU-239 (3.0 Å) and LEU-241 (2.9 Å), when compared with molecule 3g ( Figure 2B). A detailed comparison between the docking results obtained for these two ligands, regarding the DYRK1A target, can also be found in Table 4.  Additionally, docking of compound 3c was made on two of the six kinases, and a comparison with compound 3g was done. This study supported the fact that 3c inhibits DYRK1A well and better than 3g (see Table 3). Figure 2A illustrates the docking pose for 3c and Figure 2B illustrates the docking pose for 3g. The molecular reasons for the greater stability of the 3c ligand in the active site of DYRK1A can be explained mainly by the presence of two additional hydrogen bonds that the hydroxyl group in position 2 of the naphthalene ring of compound 3c establishes with the residues GLU-239 (3.0 Å) and LEU-241 (2.9 Å), when compared with molecule 3g ( Figure 2B). A detailed comparison between the docking results obtained for these two ligands, regarding the DYRK1A target, can also be found in Table 4. The fact that DYRK1A has a greater docking score in comparison with CDK2 target, regarding compound 3c, agrees with the in vitro assay (see Table 3) results and can be explained by the shorter hydrogen bonds and the greater amount of hydrophobic contacts observed in the complex formed by compound 3c and the DYRK1A enzyme. It is also important to highlight that the hydroxyl group in position 2 of the naphthalene ring of compound 3c establishes at least one hydrogen bond with the CDK2 and DYRK1A kinases. According to in vitro assay, compound 4a is active only at the CK1δ/ε target, therefore docking analysis was performed in order to elucidate the molecular reasons of his potency (0.76 μM). It can be explained mainly by the presence of three hydrogen bonds and by five residues interacting through hydrophobic contacts, as can be seen at Figure 3 and Table 4. The fact that DYRK1A has a greater docking score in comparison with CDK2 target, regarding compound 3c, agrees with the in vitro assay (see Table 3) results and can be explained by the shorter hydrogen bonds and the greater amount of hydrophobic contacts observed in the complex formed by compound 3c and the DYRK1A enzyme. It is also important to highlight that the hydroxyl group in position 2 of the naphthalene ring of compound 3c establishes at least one hydrogen bond with the CDK2 and DYRK1A kinases.
According to in vitro assay, compound 4a is active only at the CK1δ/ε target, therefore docking analysis was performed in order to elucidate the molecular reasons of his potency (0.76 µM). It can be explained mainly by the presence of three hydrogen bonds and by five residues interacting through hydrophobic contacts, as can be seen at Figure 3 and Table 4.
Protein kinases are involved in several cellular functions, such as gene transcription, cell cycle progression, cytoskeletal rearrangement, proliferation, differentiation, migration, and cell death [38,39]. Deregulation of the signaling pathways of these proteins is associated with several diseases, including cancer [40,41]. Consequently, this class of enzymes represents important therapeutic targets for treatment of diseases [42].
In the literature some semicarbazone derivatives are reported as protein kinase inhibitors [19,20]. Thus, this information motivated the accomplishment of the kinase testing that indicated 3c as a promising compound for further studies to determine the possible mechanism of action. Compound 4a was also included, because it was active for all the tested cancer cell lines, and presented CK1δ/ε kinase inhibitory activity, of interest because casein kinase 1δ/ε has been identified as a promising therapeutic target for oncology application. including cancer [40,41]. Consequently, this class of enzymes represents important therapeutic targets for treatment of diseases [42].
In the literature some semicarbazone derivatives are reported as protein kinase inhibitors [19,20]. Thus, this information motivated the accomplishment of the kinase testing that indicated 3c as a promising compound for further studies to determine the possible mechanism of action. Compound 4a was also included, because it was active for all the tested cancer cell lines, and presented CK1δ/ε kinase inhibitory activity, of interest because casein kinase 1δ/ε has been identified as a promising therapeutic target for oncology application.

Detection of Cell Viability Using the Guava ViaCount Kit
Initially, the percentage of viable cells in apoptosis or necrosis was determined by the ViaCount assay at the Guava EasyCyte HT flow cytometer (Merck-Millipore, Hayward, CA, USA). The ViaCount kit (Merck-Millipore, Hayward, CA, USA) determines cell viability based on membrane permeability. The concentrations used in the cytometric tests were values corresponding to the IC50 and 2x IC50 of 72 h and the chosen cell line was HL-60 (promyelocytic leukemia). The results showed that the two compounds significantly decreased cell viability (Figure 4).
At the two concentrations tested, 3c increased the number of cells in apoptosis. Compound 4a increased the number of cells in early apoptosis, mainly at the concentration of 22 μM, where 65.3% of the cells were in early apoptosis. Doxorubicin increased the number of cells in apoptosis and that died.

Evaluation of Mitochondrial Transmembrane Potential (ΔΨm)
Mitochondria play an essential role in the life and death of cells, as they are responsible for the energy production necessary for cell survival and also regulate apoptosis. The good performance in energy production and the integrity of the mitochondria are guaranteed by the maintenance of mitochondrial electrical potential [43]. Some drugs act by inducing the loss of mitochondrial transmembrane potential leading to a process called mitochondrial depolarization, which is one of the early events in the process of cell death by apoptosis triggered by the intrinsic (mitochondrial)

Evaluation of Mitochondrial Transmembrane Potential (∆Ψm)
Mitochondria play an essential role in the life and death of cells, as they are responsible for the energy production necessary for cell survival and also regulate apoptosis. The good performance in energy production and the integrity of the mitochondria are guaranteed by the maintenance of mitochondrial electrical potential [43]. Some drugs act by inducing the loss of mitochondrial transmembrane potential leading to a process called mitochondrial depolarization, which is one of the early events in the process of cell death by apoptosis triggered by the intrinsic (mitochondrial) pathway [44].
In this sense, in order to evaluate if compounds 3c and 4a induce apoptosis by alterating the mitochondria transmembrane potential, this assay was performed by flow cytometry using the fluorochrome rhodamine 123, because this is able to accumulate in cells with unchanged mitochondrial transmembrane potential. After 72 h of incubation, 3c and 4a were found to be able of inducing mitochondrial depolarization in HL-60 cells ( Figure 5).
Cells treated with 3c at concentration values of 13 and 26 µM produced 30.9% and 34.2% of depolarized cells, respectively. These data suggest the cell death caused by 3c involves other death pathways beyond the intrinsic mitochondrial pathway of apoptosis. On the other hand, compound 4a was more active and induced depolarization in 59.3% of the cells at a concentration of 22 µM. The positive control, doxorubicin, led to 41.8% of depolarized cells.
Apoptosis is a key programmed cell-death pathway involved in numerous processes. One of them is the balance between cell proliferation and death, essential for the maintenance of tissue homeostasis. In general, two major signaling pathways control apoptosis: (i) mitochondria-mediated or intrinsic pathway and (ii) death receptor-mediated or extrinsic pathway [45]. When the cell undergoes pro-apoptotic stimuli, such as deprivation of growth factors, DNA damage, hypoxia, activation of oncogenes, among others, the signals that are translated converge mainly to mitochondria causing the collapse of the potential of the internal mitochondrial membrane (∆ψm) that trigger death by apoptosis [46]. The results obtained in this test demonstrated that the compounds elicited pro-apoptotic effects that induced mitochondrial depolarization in HL-60 cells.

Cell Cycle Assay
In order to improve the study of the mechanism of death induction by compounds 3c and 4a in HL-60 cells, a test was performed on the flow cytometer after staining with propidium iodide to

Cell Cycle Assay
In order to improve the study of the mechanism of death induction by compounds 3c and 4a in HL-60 cells, a test was performed on the flow cytometer after staining with propidium iodide to evaluate the effect of the compounds on cell cycle progression ( Figure 6).
After the treatment period with 3c (13 µM), 63.3% of the cells were in the G1 phase. At the concentration of 26 µM, the population in the G1 phase increased to 66.2%. These results indicate that 3c causes cell cycle arrest in the G1 phase.
Treatment with 4a increased the cell population in the G1 phase by 62.7% (11 µM) and 81.9% (22 µM), respectively, indicating cell cycle arrest in the G1 phase. In the histogram (4a; 22 µM) it is possible to visualize large amounts of cell debris. The chemotherapeutic doxorubicin caused arrest at G2/M. The cell cycle is the process by which cells multiply. This is divided into four phases: G1, S, G2 and M. To ensure correct progression through the cell cycle, cells have developed checkpoints that are mechanisms that evaluate cell conditions before beginning the next phase, thereby ensuring cell division fidelity [47]. At these checkpoints, if necessary, cell cycle arrest occurs and DNA repairs are performed, however, depending on the efficacy of the process, the cycle is not completed and the cell is referred for apoptosis. The regulation of the cell cycle is a balance between the positive regulators which induce cell replication and negative regulators that prevent replication [48].

Cell Cycle Assay
In order to improve the study of the mechanism of death induction by compounds 3c and 4a in HL-60 cells, a test was performed on the flow cytometer after staining with propidium iodide to evaluate the effect of the compounds on cell cycle progression ( Figure 6).  Positive regulators are formed by protein kinases that act together stimulating the continuity of the cycle [49]. Previously, CDK5 was thought to function in a cell cycle independent manner, however, recently the retinoblastoma protein (Rb) was discovered as a downstream target of CDK5. Expression of CDK5 leads to the phosphorylation of Rb, ultimately leading to the expression of cyclins and others cdks [50]. In the protein kinase inhibition test, it was found that 3c inhibited the CDK5/p25 complex, probably preventing the phosphorylation of pRb protein, preventing the cell leaving the G1 stage and entering the S phase. The stop of the cell cycle in G1 caused by 3c corroborates with this hypothesis.
According to Huart et al. [51], more than four hundred proteins interact with p53 and many of them act as inhibitors. Studies have shown that the casein kinase 1 (CK1) family phosphorylates the MDM2 protein and modulates the interaction with p53 under different conditions. Inuzuka et al. [52] observed that under normal conditions, the CK1δ/ isoforms are able to phosphorylate several serine residues in the acid domain of MDM2 that are involved in the degradation of p53, acting as a negative regulator of this protein.
Studies suggest that dysregulation of these cellular processes contributes to oncogenesis, and that pharmacological inhibition of CK1 can increase p53 protein levels in cells and induce cell death. Thus, the inhibition of these kinases represents an important approach in anticancer therapy [53,54].
Compound 4a inhibited CK1δ/ε isoform activity. That result may be associated with cell cycle arrest at G1 checkpoint caused by 4a in HL-60 cells, since CK1δ/ε inhibition prevents the transition of abnormal postmitotic cells from G1 to S phase [55].

DNA Fragmentation Assay
Cell death by apoptosis is characterized by marked morphological changes such as chromatin condensation, DNA fragmentation, cell membrane blebbing and formation of apoptotic bodies [56].
DNA fragmentation is initiated after chromatin condensation and depends on several variables, such as the time required for the depolarization of the mitochondria and activation of the caspase cascade. In response to DNA damage the p53 protein is activated causing cell cycle arrest in G1 [57]. At flow cytometry the apoptotic cells with fragmented DNA can be identified as the sub-G1 population, seen to the left of the G1 peak [58]. Thus, using the fluorochrome propidium iodide and the same principle of the cell cycle test, an assay was performed to identify HL-60 cells with fragmented DNA (Figure 7).

DNA Fragmentation Assay
Cell death by apoptosis is characterized by marked morphological changes such as chromatin condensation, DNA fragmentation, cell membrane blebbing and formation of apoptotic bodies [56].
DNA fragmentation is initiated after chromatin condensation and depends on several variables, such as the time required for the depolarization of the mitochondria and activation of the caspase cascade. In response to DNA damage the p53 protein is activated causing cell cycle arrest in G1 [57]. At flow cytometry the apoptotic cells with fragmented DNA can be identified as the sub-G1 population, seen to the left of the G1 peak [58]. Thus, using the fluorochrome propidium iodide and the same principle of the cell cycle test, an assay was performed to identify HL-60 cells with fragmented DNA (Figure 7).
Compound 3c induced fragmentation in 3.2% and 7.5% of cells at concentrations of 13 μM and 26 μM, respectively. Compound 4a induced fragmentation in 7.0% and 40.3% of the cells at concentrations of 11 and 22 μM, respectively. This result corroborates the inhibition of CK1δ/ε and the accumulation of abnormal postmitotic cells at G1 phase [55].

Study of the Induction of Cell Death by the Annexin V test
Aiming to better study the mechanism of induction of death involved in HL-60 cells after treatment with compounds 3c and 4a, a flow cytometric test was performed after double staining with annexin V-FITC/7-AAD. This assay is based on the ability of annexin V to bind with high affinity to phosphatidylserine which is translocated to the outer membrane of apoptotic cells and the 7-

Study of the Induction of Cell Death by the Annexin V Test
Aiming to better study the mechanism of induction of death involved in HL-60 cells after treatment with compounds 3c and 4a, a flow cytometric test was performed after double staining with annexin V-FITC/7-AAD. This assay is based on the ability of annexin V to bind with high affinity to phosphatidylserine which is translocated to the outer membrane of apoptotic cells and the 7-AAD dye binds to the DNA of cells with loss of membrane integrity. Thus, in this test it is possible to identify four distinct populations: viable cells [Annexin V(−) and 7-AAD(−)], initial apoptotic cells [Annexin V(+) and 7-AAD(−)], late apoptotic cells [Annexin V(+) and 7-AAD(+)] and dead cells [Annexin V(−) and 7-AAD (+)] [59]. The results are shown in Figure 8.
After treatment with 3c (13 and 26 µM), it was observed that 16.7 and 27.2% of the cells were in initial apoptosis and 7.4 and 18.1% in late apoptosis phase, respectively. In the treatment with 4a (11 µM) 5.4% of the cells were in initial apoptosis, 8.3% in late apoptosis and 9.1% were dead. Cells treated with 4a (22 µM) showed 4.7% of the cells in initial apoptosis, 12.5% in late apoptosis and almost 50% of the cells were dead. Regarding the treatment with doxorubicin, 23.7% of the cells were in initial apoptosis and 19% in late apoptosis.
In cell death by apoptosis, the cell undergoes morphological changes striking and coordinated. Initially, the cellular plasma membrane remains intact but undergoes structural changes, such as the distribution of phosphatidylserine in the outer layer of the membrane, which is a signal for recognition by phagocytes. Thus, the dead cell is rapidly phagocytosed by macrophages and removed without extravasation of the cytoplasmic content [60]. After treatment with 3c (13 and 26 μM), it was observed that 16.7 and 27.2% of the cells were in initial apoptosis and 7.4 and 18.1% in late apoptosis phase, respectively. In the treatment with 4a (11 μM) 5.4% of the cells were in initial apoptosis, 8.3% in late apoptosis and 9.1% were dead. Cells treated with 4a (22 μM) showed 4.7% of the cells in initial apoptosis, 12.5% in late apoptosis and almost 50% of the cells were dead. Regarding the treatment with doxorubicin, 23.7% of the cells were in initial apoptosis and 19% in late apoptosis.
In cell death by apoptosis, the cell undergoes morphological changes striking and coordinated. Initially, the cellular plasma membrane remains intact but undergoes structural changes, such as the distribution of phosphatidylserine in the outer layer of the membrane, which is a signal for recognition by phagocytes. Thus, the dead cell is rapidly phagocytosed by macrophages and removed without extravasation of the cytoplasmic content [60]. The main death mode caused by 3c was apoptosis, confirmed in all the tests. The action of 3c on cell death by apoptosis is probably related to strong inhibition of protein kinases, which causes cell cycle arrest and triggers pro-apoptotic signals that lead to the collapse of mitochondrial membrane potential, culminating in apoptosis.
Derivative 4a, at the highest concentration (22 µM), showed strong cytotoxicity and caused DNA fragmentation in HL-60 cells. The compound 4a is a nitro derivative and the effect can be associated with the mechanism of action of this class of substances that is related to the reduction of the nitro group, resulting in reactive species that interact with the plasma membrane of the cells and with essential biomolecules, causing the cytotoxic effect. However, further studies need to be done to elucidate the mechanism of action of this compound.

General Experimental Procedures
All melting points were determined using a Fisatom 430D apparatus (Fisatom, São Paulo, Brazil) and are uncorrected. FTIR spectra were obtained on a model IFS66 spectrophotometer (Bruker, Billerica, MA, USA) using KBr pellets. NMR spectra were recorded on Bruker Avance 400 spectrometer ( 1 H-NMR at 400 MHz and 13 C-NMR at 100 MHz), the residual signal of the solvent was used as internal reference. Chemical shift (δ) values are expressed in parts per million (ppm) and coupling constants (J) values are given in Hz. Abbreviations: s = singlet; d = doublet; t = triplet; q = quadruplet; m = multiplet. Elemental analyses were performed on a PE-2400 instrument (PerkinElmer Inc., Waltham, MA, USA) and the results were within acceptable range. Thin layer chromatography (TLC) was carried out on silica gel plates with a fluorescence indicator of F254 (0.2 mm, E. Merck, Darmstadt, Germany). All reagents used in the present study were of analytical grade and were purchased from Aldrich or Fluka (both São Paulo, Brazil) and were used without additional purification.

General Procedure for the Preparation of Arylsemicarbazones 3a-3m
Compounds 3a-3i and 3k-3m were obtained by reaction of 2.5 mmol of semicarbazide hydrochloride and an equimolar amount of the appropriate aryl aldehyde in 15 mL of absolute ethanol as solvent. Three drops of hydrochloric acid were added to the medium as catalyst. The reaction was kept at room temperature and under constant magnetic stirring for a period of 3-5 h. To obtain derivative 3j, 5 mmol of semicarbazide hydrochloride (two equivalents) were added to 2.5 mmol of the respective aldehyde. The reactions were monitored by means of thin layer chromatography and, upon completion, the arylsemicarbazone precipitated in the reaction medium was filtered in sintered funnel and washed with water, followed by washing in hexane. Purification of the synthesized compounds was performed by recrystallization from distilled water.

Molecular Modeling
To elucidate the binding mode of the compound 3c that showed the highest activity among the compounds, docking calculations were done against CDK2, CDK5, CDK9, CLK1, DYRK1A, PIM1 and CK1δ. For comparison purposes, the compound 3g, with IC 50 > 10 µM, were also included, as a negative result. The compound 4a, that showed activity against CK1δ target was also evaluated. The optimized structure for ligands 3c, 3g and 4a were obtained by application of the RM1 method [61], available as part of SPARTAN program [62], using internal default settings for convergence criteria. The targets structures used in the docking studies were taken from the RCSB Protein Data Bank, available online: http://www.rcsb.org (accessed on: 17 July 2019), under the following codes: CDK2 (6GUE), CDK5 (1UNL), CDK9 (3MY1), CLK1 (6FT8), DYRK1A (4YLK), PIM1 (3VBV) and CK1δ (3UYT). GOLD software [63] was used for these calculations. The binding mode for the ligands was determined as the highest (most positive) score among the possible solutions for each ligand, generated according to the Chemscore fitness function. The active site was defined after performing the alignment of all six targets, as all atoms within a radius of 7.0 Å from the co-crystallized ligand of CDK2 "FB8". Ten aminoacids residues in each target were treated as flexible during the calculations, in order to take into account the effects of the induced fit, see Table S1 at Supplementary Materials. Then, BINANA program [64] was used to analyze the molecular interactions found in the docking results, using default settings, with the exception of the hydrogen bond distance that was changed to 3.5 Å.

Ethical Aspects of the Peripheral Blood Mononuclear Cell Assay
The project was submitted and approved by the Human Research Ethics Committee of the Federal University of Pernambuco, which is in accordance with Resolution 466/12 of the National Health Council and approved with CAAE opinion 67004017.3.0000.5208.

Isolation of Peripheral Blood Mononuclear Cells
Mononuclear cells were obtained from the peripheral blood of healthy volunteers by heparin tube collection and isolated by Ficoll-Histopaque gradient centrifugation. After centrifugation, the cells were harvested, washed and resuspended in RPMI 1640 medium supplemented with 20% fetal bovine serum, 2 mM glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin (Gibco™ Life Technologies, Waltham, MA, USA). The phytohemagglutinin mitogen 2% was also added into the culture medium. After that, using the neubauer chamber, the cells were counted and plated in a 96-well plate at a concentration of 10 6 cells/mL and then were maintained for overnight under incubation at 37 • C and 5% CO 2 [65].

Culture of Tumor Cells
The cells were obtained from the Cell Bank of Rio de Janeiro and maintained in the Laboratory of Cell Culture of the Department of Antibiotics of the Federal University of Pernambuco. The lines K562 (myeloblastic leukaemia), HL-60 (promyelocytic leukemia) and MOLT-4 (lymphoblastic leukemia) were cultured in RPMI 1640 medium and the lines HEp-2 (larynx carcinoma), NCI-H292 (lung carcinoma), HT-29 (colon adenocarcinoma) and MCF-7 (breast adenocarcinoma) were cultured in DMEM medium, supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin (Gibco™ Life Technologies), at 37 • C and 5% CO 2 .

Cytotoxicity Assessment by MTT Assay
The evaluation of the cytotoxicity of the compounds was performed by the method of 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) (Sigma Aldrich Co., St. Louis, MO, USA). To perform the assay the cells were seeded in 96-well plates at the following concentrations: HL-60 and K562 (0.3 × 10 6 cells/mL); MOLT-4 and PBMC (10 6 cells/mL); NCI-H292, HT-29, HEp-2 e MCF-7 (10 5 cells/mL). Briefly, the compounds previously dissolved in DMSO (dimethylsulfoxide, Vetec, São Paulo, Brazil), were diluted in medium and added to each well at the desired concentrations. For the calculation of the concentration that inhibits 50% of cell growth (IC 50 ) the compounds were serially diluted to obtain the final concentrations between 0.19-25 µg/mL. Doxorubicin was used as the standard in the concentration range between (0.078-10 µg/mL). After 72 h of contact of the cells with the compounds, it was added to each well 25 µL of MTT solution (5 mg/mL). Plates were left for another 3 h in incubator at 37 • C and at the end of that period the supernatant was aspirated. To perform the reading, 100 µL of DMSO was added in each well for the dissolution of the formazan crystals. The amount of formazan was measured by reading the plate at 560 nm absorbance. The IC 50 was calculated from the nonlinear regression in the Prisma 7.0 software (GraphPad, San Diego, CA, USA). Each sample was tested in triplicate in three independent experiments.

Kinase Inhibition Assays
Kinase activities were assayed in appropriate kinase buffer, with either protein or peptide as substrate in the presence of 10 µM ATP in a final volume of 6 µL using the ADP-Glo TM assay kit (Promega, Madison, WI, USA, as described in Zegzouti et al. [66]). Controls were performed with appropriate dilutions of dimethylsulfoxide (solvent of the tested compounds). Peptide substrates were obtained from Proteogenix (Schiltigheim, France).