Highly Antiproliferative Latonduine and Indolo[2,3-c]quinoline Derivatives: Complex Formation with Copper(II) Markedly Changes the Kinase Inhibitory Profile

A series of latonduine and indoloquinoline derivatives HL1–HL8 and their copper(II) complexes (1–8) were synthesized and comprehensively characterized. The structures of five compounds (HL6, [CuCl(L1)(DMF)]·DMF, [CuCl(L2)(CH3OH)], [CuCl(L3)]·0.5H2O, and [CuCl2(H2L5)]Cl·2DMF) were elucidated by single crystal X-ray diffraction. The copper(II) complexes revealed low micro- to sub-micromolar IC50 values with promising selectivity toward human colon adenocarcinoma multidrug-resistant Colo320 cancer cells as compared to the doxorubicin-sensitive Colo205 cell line. The lead compounds HL4 and 4 as well as HL8 and 8 induced apoptosis efficiently in Colo320 cells. In addition, the copper(II) complexes had higher affinity to DNA than their metal-free ligands. HL8 showed selective inhibition for the PIM-1 enzyme, while 8 revealed strong inhibition of five other enzymes, i.e., SGK-1, PKA, CaMK-1, GSK3β, and MSK1, from a panel of 50 kinases. Furthermore, molecular modeling of the ligands and complexes showed a good fit to the binding pockets of these targets.

Latonduines (backbones C and D in Chart 1) were first extracted from the Indonesian sponge Stylissa carteri and are not cytotoxic to cancer cells. 28,29 However, substitution of their pyrrole ring by an indole unit made them cytotoxic. 30,31 The resulting indolo [2,3-d]benzazepine (backbone E in Chart 1) is a microtubule destabilizing agent (MDA) targeting the colchicine binding site. 30 The two isomeric backbones indolo [3,2d]benzazepine A and indolo [2,3-d]benzazepine E are related structurally as shown in Chart 1. Nevertheless, by flipping the indole moiety and shifting the lactam unit in paullone A, one obtains not only increased cytotoxicity but also a different mode of action. 30 Being intrigued by the activity of indolo [3,2-d]benzazepineand indolo [3,2-c]quinoline-based molecules as potential anticancer drugs, we decided to extend our chemistry to other related isomeric systems, namely, indolo [2,3-d]benzazepineand indolo [2,3-c]quinoline-derived species, with unexplored chemistry and biological effects. We envisioned exciting new results in the field of metal-based anticancer drugs and, in particular, new structure−activity relationships.
One of the major drawbacks of these isomeric scaffolds is their limited aqueous solubility and bioavailability. This issue was successfully addressed for many indolo [3,2-d]benzazepine and indolo [3,2-c]quinoline derivatives and several latonduines by creating metal binding sites at their backbones and metal complex formation. Werner-type coordination complexes of copper(II), ruthenium(II), osmium(II), gallium(III), and organometallic compounds were synthesized and investigated as potential anticancer drugs. 32−40 The reported results revealed that the metal complexes did not only enhance the aqueous solubility and bioavailability but also augmented their antiproliferative activity both in vitro and in vivo. Nevertheless, bioavailability and aqueous solubility need further improvement, requiring other approaches to enhance their pharmacological profile.
Morpholine, as a known biologically active moiety, has been attached to the main scaffolds since it is considered to improve the necessary pharmacological parameters of drug candidates. 41,42 In our recent paper, 43 we reported that the Schiff base resulted from condensation of the 11-bromo-7-hydrazin-yl derivative of E (Chart 1) with 2-acetylpyridine and its copper(II) complex showed the highest cytotoxicity among the compounds tested. This prompted us to further develop this backbone and prepare 2-acetylpyridine with a morpholine unit. As a starting material, the respective aldehyde was used, which was recently reported by us. 44 Protein kinases represent an excellent target for cancer therapy. 45−49 It should be also stressed that the multitargeted kinase inhibitors have become a "hot topic", accounting for about 25% of drug discovery research. 48,49 Initial attempts to create highly selective mono-kinase inhibitors to avoid unexpected toxic effects have been steadily displaced by two anticancer therapies that target several kinases and block distinct kinase signaling pathways as they showed therapeutic benefits in the treatment of complex cancer diseases. 46 The first therapy is based on using several selective mono-kinase inhibitors simultaneously, while the second is based on using a single drug as a multikinase inhibitor. Advantages and hurdles of both therapies have been discussed in the literature. 46,48,49 The second therapy, which implies the use of a single drug as a multikinase inhibitor that revealed higher potency, permits avoiding the consequences of drug−drug interactions, which can affect absorption, metabolism, excretion, plasma level, and, finally, activities, as well as reducing side effects and is much easier to apply. 46 Herein, we report the synthesis and characterization of new chelating systems derived from indolo [2,3-d]benzazepine E and indolo [2,3-c]quinoline F and of their copper(II) complexes (Chart 2), speciation in aqueous solution, and antiproliferative activity. The inhibition ability of the lead compounds in a panel of 50 kinases was investigated in vitro and by molecular modeling, providing insights into the mode of action of the most potent copper(II) complex and its metal-free ligand. These modified molecules offer a broad spectrum of various effects on malign cells, while some of them show a marked increase in aqueous solubility, thus increasing the bioavailability and improving the pharmacological profile. Lead compound 8 and its proligand HL 8 do not only target cancer specific kinases but also offer an excellent pharmacological profile.

RESULTS AND DISCUSSION
2.1. Synthesis and Characterization of Starting Building Blocks and Ligands. The aldehyde G prepared as reported elsewhere 44 was converted into secondary alcohol I by reaction with the Grignard reagent (CH 3 MgBr) and workup in 92% yield. Swern oxidation of I resulted in ketone J (Scheme 1), which was purified chromatographically to give an easily crystallizable product in 69% yield.
The 1 H NMR spectrum of J agreed with the expected structure, which, in addition, has been confirmed by SC-XRD (see Chart S1 for atom numbering scheme and Figure S1 in the Supporting Information).
The derivative IVb (Scheme 2) has not been reported previously. Its synthesis has been performed by following the procedures described in the literature for unsubstituted indolo [2,3-c]quinoline IVa 50 as shown in Scheme 2. In the first step, ethyl 5-bromo-1-ethoxymethyl-1H-indol-2-carboxylate was allowed to react with 2-iodoaniline in the presence of AlMe 3 in dichloromethane (DCM) to give Ib in 84% yield. Protection of carboxamide nitrogen atom and isolation of IIb were realized in 99% yield by treatment of Ib with di-tert-butyl dicarbonate Boc 2 O in dry acetonitrile in the presence of catalytic amount of N,N-dimethyl-4-aminopyridine (DMAP). The intramolecular Heck cyclization reaction of IIb in the presence of Pd(OAc) 2 , PPh 3 , and Ag 2 CO 3 followed by workup afforded IIIb in 19% yield. Full deprotection of IIIb and formation of IVb were accomplished in 79% yield by refluxing IIIb in EtOH:12 M HCl 4:1.
Then, compounds IVa and IVb were chlorinated with excess POCl 3 at 120°C to give rise to Va/Vb in >90% yield. Finally, the treatment of Va/Vb with excess hydrazine hydrate at reflux delivered the desired species M and N in >95% yield. This pathway to create chelating molecules with some modifications was also successful with core structures A and B (Chart 1). 51,52 The potential ligands HL 1 −HL 8 were synthesized by Schiff base condensation reactions of hydrazin-yl derivatives K−N with aldehyde G or ketone J in anoxic ethanol (Scheme 3) in 57−98% yields by adapting literature protocols. 43,51,53 1 H NMR spectra of the potential ligands HL 1 −HL 8 show the typical peak pattern of the morpholine unit at around 2.35 and 3.55 ppm, as well as proton resonances of the linking methylene group between the morpholine unit and the pyridine ring at around 3.55 ppm sometimes overlapping with H 25 for indolo [2,3-d]benzazepines or H 24 for indolo [2,3-c]quinolines, respectively (for atom numbering scheme, see Chart S2). The additional methyl group as R 2 in Scheme 3 is seen as a singlet at around 2.49 ppm. The 2D NMR spectra provided evidence that indolo [2,3-d]benzazepines HL 1 −HL 4 are solely present as tautomers with an exocyclic double bond between C 7 and N 13 as evidenced by weak coupling between H 6 and C 5 and triplet resonance in the 1 H NMR spectra, suggesting the presence of two protons in the closest vicinity of H 6 . 43 There were no other tautomeric forms identified. In contrast, NMR spectra of HL 5 − HL 8 indicate that these indolo [2,3-c]quinolines exist in two tautomeric forms in the solution. The major species possesses an exocyclic double bond with a hydrogen atom at N 5 , while the minor species contains an endocyclic double bond with a hydrogen atom at N 12 . The ratio between these two species is solvent-and concentration-dependent. At a concentration of about 10 mg/mL in DMSO-d 6 , the ratios between the major vs minor species are 1:0.02, 1:0.01, 1:0.80 and 1:0.15 for HL 5 − HL 8 , respectively. In most cases, a complete assignment of all 1 H and 13 C resonances was impeded by low signal intensity for minor species and signal overlapping. However, the high signal intensity of the minor species in the case of HL 7 made a complete assignment of all signals in both species possible. The chemical shift of NH 5 for the species with an exocyclic double bond is 11.97 ppm, while that of NH 12 for the species with an endocyclic double bond is 14.54 ppm. This was confirmed by a long-range 1 H− 13 C HMBC 2D NMR experiment, where the proton N 5 H showed 3 J couplings to quaternary carbons C 11c and C 6a , while N 12 H revealed such couplings to quaternary carbon C 6a and ternary carbon C 14 , which is possible, if the hydrogen is bound to a hydrazinic nitrogen. The structural change in the molecule from an exocyclic to an endocyclic double bond leads to a shift of all 1 H and 13 C signals. While the 1 H and 13 C resonances of the morpholine moiety are only marginally affected, those near the hydrazinic moiety show major changes. In particular, the signals for the hydrogen C 14 H and imine carbon are upfield shifted from 8.52 to 7.57 ppm and from 152.19 to 131.14 ppm, respectively, when going from major to minor species (see Chart S2 for the NMR atom numbering scheme in the Supporting Information). At close inspection of the 1 H NMR spectrum of HL 7 , two more sets of NMR signals with low intensities become apparent, which are presumably attributed to E and Z isomers of the previously described tautomers, leading to a total of four signal sets. However, low signal intensity, signal overlapping, and the absence of non-cross Scheme 1. Synthesis of 2-Acetyl-5-(morpholinomethyl)pyridine J Starting from 2-Formyl-5-(morpholinomethyl)pyridine G 44a peaks in a two-dimensional 1 Figures S3 and S4, respectively, with pertinent bond distances (Å), bond angles, and torsion angles (deg) quoted in the legends. Details of data collection and refinement are given in Table S1. The complexes crystallized in the monoclinic space groups C2/c, P2 1 /c, P2 1 /c, and non-centrosymmetric triclinic P1, respectively.
The coordination geometry of Cu(II) in [CuCl(L 1 )(DMF)] ( Figure 1a) is four-coordinate square-planar, even though very weak coordination of DMF molecule can be considered, taking into account the apical position of oxygen atom O1 with respect to the basal plane determined by the metal ion, the coordinated three nitrogen donor atoms, and the chlorido co-ligand. In this latter case, the coordination geometry can be interpreted as 4 + 1 binding. Comparison with coordination geometry in [CuCl-(L 2 )(CH 3 OH)] (Figure 1b), which is best described as fivecoordinate square-pyramidal, and bond lengths around copper-(II), which are significantly expanded when compared to those in [CuCl(L 1 )(DMF)], provides further evidence for a more appropriate description of coordination geometry in [CuCl-(L 1 )(DMF)] as four-coordinate. The increase in coordination number in [CuCl(L 2 )(CH 3 OH)] to five leads to a significant expanding of interatomic distances between Cu(II) and donor atoms due to increase in interatomic repulsions. The latonduine backbone in both complexes has almost identical folding due to the presence of an sp 3 -hybridized carbon atom in the sevenmembered azepine ring. The torsion angle Θ C4a−C5−N6−C7 is almost the same in both complexes (see the values quoted in the legends to Figures 1a,b).
Complex [CuCl(L 3 )] forms a weak dimeric associate ( Figure  S2), in which the coordination environment of Cu1 can be described as slightly distorted square-planar (see also Figure 1c). The atom Cl1 acts as a bridging μ-chlorido co-ligand to Cu2 of the second half of the dimeric associate with formation of a long contact of 2.8970(8) Å. Therefore, the coordination geometry of Cu2 can be described as 4 + 1 as was the case for complex [CuCl(L 1 ) (DMF)]. The latonduine derivatives adopt the same binding mode to both Cu(II) atoms Cu1 and Cu2, and each acts as a monoanionic tridentate ligand. The bond lengths in each chromophore of the two Cu(II) ions are very similar to those in [CuCl(L 1 ) (DMF)], in accordance with small structural difference between the two coordinated ligands (L 1 ) − and (L 3 ) − .
In contrast to Cu(II) compounds with strongly folded indolo [2,3-d] The coordination geometry in [CuCl 2 (H 2 L 5 )] + is five-coordinate and can be described as intermediate (τ 5 = 0.40) between square-pyramidal (τ 5 = 0) and trigonal bipyramidal (τ 5 = 1). 54 The indolo [2,3-d] 33 The presence of a six-membered pyridine-like ring in [CuCl 2 (H 2 L 5 )] + and HL 6 instead of a seven-membered azepine ring makes these indoloquinoline systems flat, a premise to intercalate into DNA. The SC-XRD studies revealed that HL 1 −HL 3 and HL 5 act as tridentate ligands but adopt different protonation states depending on conditions. Therefore, it was of interest to study the solubility of metal-free ligands and Cu(II) complexes as well as their protonation state at physiological pH.
2.4. Solubility Studies. The thermodynamic solubility of selected metal-free ligands and Cu(II) complexes was characterized in water at pH 7.4 and 5.0 using UV−vis spectrophotometry for the determination of the concentration of the saturated solutions. The determined solubility (S) values are shown in Figure 2.
The obtained solubility values indicate that the copper(II) complexes are commonly more soluble in water than their corresponding metal-free ligands. The positive effect of the morpholine moiety on the solubility is measurable but minor. Comparison of the S values for the HL 5 and HL 5 nm , 5 and 5 nm , and 6 and 6 nm pairs under the conditions used indicates that all compounds have significantly better solubility at pH 5 than at pH 7.4. For HL 5 and HL 5 nm , this can be explained by the partial protonation of the morpholine nitrogen and pyridine nitrogen with the decreasing pH. The morpholine-containing complexes (3 and 5−8) also can get partially protonated at the noncoordinating morpholine nitrogen when the pH is lowered. However, formation of the aqua complex from the mixed hydroxido species (vide infra) can also contribute to the increased solubility even for the non-morpholine complexes at pH 5. The presence of the bromo-substituent (5 vs 6, 5 nm vs 6 nm , and 7 vs 8) results in a considerable decrease of the aqueous solubility at both tested pH values, while the effect of the methyl group at the Schiff base ketimine bond is not significant.
2.5. Solution Equilibrium Studies. The protonation processes of HL 1 nm and its morpholine counterpart HL 1 (Chart 3) and the solution stability of their copper(II) complexes (1 nm and 1) were characterized by UV−vis spectrophotometry. Since the organic compounds and their copper(II) complexes possess limited aqueous solubility, the solution equilibrium studies were performed in 30% (w/w) DMSO at low concentrations (12.5 μM (HL 1 nm ) or 50 μM (HL 1 )). Based on the characteristic changes in the UV−vis spectra for HL 1 nm in the pH range 2−6 ( Figure 3a), two  relatively well-separated proton dissociation processes were observed and their pK a values were determined (Table 1). Notably, upon increasing the pH to >∼6.6 precipitation occurred in the solution, leading to the elevation of the base line most probably due to the formation of the neutral metal-free ligand species. The first proton dissociation step was accompanied by a blueshift (λ max : 352 nm → 342 nm) in the pH range between 2 and 3.45, while the λ max is redshifted (342 nm → 377 nm) upon the second step. These spectral changes and the spectra of the individual ligand species (Figure 3b) calculated by deconvolution of the recorded UV−vis spectra are fairly similar to those found for the ketimine derivative of HL 1 nm in our recent work. 43 Thus, a similar deprotonation pattern is feasible for the non-substituted and the methyl-substituted ligands. The neutral species HL 1 nm can be present in two tautomeric forms (due to the rearrangement of the NC− NH−N and NH−CN−N bonds) and can be protonated at two sites (Scheme 4). Therefore, the first proton dissociation step (pK a1 ) is attributed to deprotonation of the pyridinium nitrogen, while the second process (pK a2 ) is attributed to the deprotonation of the benzazepinium nitrogen. The ligand HL 1 nm possesses somewhat lower pK a values than its methyl derivative as a result of the electron-donating property of the methyl group.
Two pK a values were determined for HL 1 as well (Table 1) by the deconvolution of the recorded spectra (Figure 3b), although the spectral changes were different. This ligand contains also the morpholinium group (Scheme 4). By the careful analysis of the spectral changes, it is suggested that the deprotonation of the pyridinium nitrogen takes place at fairly acidic pH, and its pK a value is lower compared to that of HL 1 nm as a result of the electron-withdrawing effect of the protonated morpholinium moiety. Even though the deprotonation of the benzazepinium and morpholinium nitrogens is overlapping, the spectra of the individual species (Figure 3d) suggest that pK a2 mostly belongs to the benzazepinium moiety, and the deprotonation of the non-Chart 3. Metal-Free Ligands HL 1 and HL 1 nm Used for the Solution Equilibrium Studies  chromophoric morpholinium nitrogen is accompanied by a minor spectral change, as expected.
The pK a values collected in Table 1 indicate that the neutral form (HL) predominates at physiological pH in solution in the case of both ligands, contributing to their strong lipophilic nature (log D 7.4 values +4.75 (HL 1 nm ) and +4.30 (HL 1 ) estimated by the MarvinSketch program. 55 The UV−vis spectra recorded for 1 nm in the pH range 2−11 ( Figure 4a) showed strong similarities to those of the methylated complex reported recently. 43 (Figure 4b), the same coordination modes are likely in these species in couples. Therefore, the deprotonation of [CuH(HL 1 )] 3+ takes place most probably at the hydrazinic nitrogen (characterized by a pK a of 3.81) and the morpholinium nitrogen remains protonated. Increasing the pH (pH > 5), the absorbance is decreased in the whole wavelength range, no isosbestic points are found (indicating the formation of some precipitate), while the λ max is increased. Formation of a mixed hydroxido species [CuH(L 1 )(OH)] + is also possible. However, the deprotonation of the morpholinium group can take place as well (formation of [Cu(L 1 )] + ) in this pH range. As these two processes are overlapping, the two species could not be well distinguished, and the obtained stability constant is quite uncertain.
Using the determined stability constants, concentration distribution curves were computed ( Figure 5) for both copper(II)−ligand systems. The results imply that the complexes do not dissociate at the used 12.5 μM concertation at neutral pH (the fraction of the free metal ion is negligible). However, the concentration of the free metal ion is significantly higher in the acidic pH range than it was found for the methylated complex, suggesting the somewhat lower stability of  1 nm and 1. Notably, while the ketimine derivative of 1 nm (or 3 nm in terms of nomenclature used herein) was identified as the predominant species at neutral pH, 43 for 1 nm and 1, formation of considerable amounts of mixed hydroxido species is also suggested ( Figure 5).

Stability of Selected Compounds in a Buffered
Medium and Blood Serum. Prior to the biological assays, the aqueous stability of selected compounds (HL 4 , 4, HL 8 , and 8) was measured as a function of time by UV−vis spectrophotometry at pH = 7.40 in 10 mM HEPES in PBS and in blood serum diluted by factor 3 (dilutions were made in HEPES and in PBS as well). In the buffer solutions, slow precipitation of the compounds was observed as shown for 4 in Figure S5 in the Supporting Information. This process took several hours, and it was in all cases less pronounced in the PBS buffer than in the HEPES medium.
Measurements with diluted serum show a more elaborate picture. As can be seen in Figure S6, complex 8 appears to react with serum components in a fast process and then slow precipitation and a second type of interaction become dominant. The first process (0.1−13 min) has only moderate effect on the spectral properties of the complex, and the N, N, N coordination sphere is not altered, while slow development of a new band at 400 nm indicates partial decomposition of the complex. The new band cannot be undoubtedly attributed to the liberation of ligand HL 8 . For complex 4, a similar behavior was observed, even though changes were smaller. Interestingly, HL 4 itself interacts with serum components ( Figure S7), and complex formation with metal ions (i.e., Zn(II), Cu(II), and Fe(III)) can be supposed as well.

Lead Morpholine-Indolo[2,3-c]quinolone and Latonduine Derivatives as well as Their Cu(II) Complexes
Exhibit Antiproliferative Activity in a Sub-micromolar Concentration Range and Trigger Apoptosis. The in vitro antiproliferative activity of the compounds was tested in doxorubicin-sensitive Colo205 and multidrug-resistant Colo320 human colon adenocarcinoma cell lines as well as in normal human embryonal lung fibroblast cells (MRC-5) by  MTT assay. As seen in Table 2, the morpholine-hybrid ligands nm )] 2+ = 6.39) 43 indicates that methylation at the Schiff base azomethine bond increases markedly the stability constants and is in agreement with the enhancement of antiproliferative activity.
The bromo-substituent brings much smaller changes in the cytotoxicity but enhances selectivity for cancer cells in the case of methylated (ketimine) Schiff bases.
An additional trend can be seen when comparing the IC 50 values of metal-free ligands HL 1 −HL 8 with their respective copper(II) complexes 1−8. Upon complex formation, the IC 50 values generally decrease, showing a positive effect on the cytotoxic behavior of the organic molecules in both cancer cell lines except for the HL 5 /5 pair. The nickel(II) complex [Ni(HL 7 ) 2 ]Cl 2 showed inferior cytotoxicity when compared to compound 7 and other copper(II) complexes tested. Even though the stoichiometry of copper(II) complexes and nickel-(II) complexes is different, we assume that not only the stoichiometry has an effect on cytotoxicity but also the metal ion identity. Most of the proligands as well as their copper(II) complexes (except HL 5 , HL 6 , and 6) show selectivity toward the doxorubicin-sensitive Colo205 cells over the normal cells (selectivity factor (SF) > 1), while the SF values were smaller with regard to the resistant Colo320 cell line. Notably, the HL 7 / 7 pair in both cell lines and the HL 8 /8 pair in the case of Colo205 display very good selectivity. An additional positive effect of complex formation with copper(II), besides the increase in cytotoxicity, is the generally increased selectivity of the complexes toward the cancer cells. Therefore, upon complex formation with copper(II), a marked enhancement of the pharmacological profile is noticed. In all, complexes 4 and 7 are not only characterized by lower IC 50 values on both malign cell lines than their corresponding metal-free ligands and the reference compound doxorubicin but are also selective. Additionally, complex 4 is found to be somewhat more cytotoxic against the resistant Colo320 cells in comparison to the sensitive cells, which is another noteworthy feature.
The influence of the morpholine moiety on both cytotoxicity and selectivity was assessed by comparison of the cytotoxicity for the metal-free ligands HL 4 and HL 4 nm and complexes 4 and 4 nm (Chart 4).
While the cytotoxicity of HL 4 nm and 4 nm in Colo205 cells exceeds that of the respective morpholine-bearing molecules HL 4 and 4 by a factor of ca. 3, this trend is inverted for Colo320 cells. Interestingly, the increase in the cytotoxicity in the Colo205 cells upon the complex formation is much smaller for both ligands in comparison to the Colo320 cells. The morpholine-bearing complex 4 is more active than 4 nm in Colo320 cells.
The lead compound 8 and its metal-free ligand HL 8 , as well as their indolo[2,3-d]benzazepine analogues (4 and HL 4 ), were further investigated to elucidate the cytotoxic mechanism. An apoptosis assay was performed by flow cytometry via the analysis of multidrug-resistant Colo320 cells stained with Annexin-V-FITC and propidium iodide (PI). The compounds were tested at two concentrations in the range of their IC 50 values, and 12Hbenzophenothiazine (M627) and cisplatin were used as positive controls. Apoptosis is a form of programmed cell death, which is the preferred mode of action for an anticancer drug. 57 The fluorescence of PI (FL3) was plotted versus Annexin-V fluorescence (FL1) as shown in Figure 6 for the positive controls, for DMSO, and for the tested compounds. The percentage of the gated events regarding the early apoptosis, the late apoptosis and necrosis, and cell death is quoted in Table S2. These data revealed that all four tested compounds (HL 4 , 4, HL 8 , and 8) could trigger apoptosis in Colo320 cells more efficiently than cisplatin. Of note is the high percentage of early apoptotic (10.7%) and late apoptotic and necrotic cells (20.1%) for HL 4 at 2 μM, which further increases for HL 8 to 15.8 and 44.5%, respectively. A high population of apoptotic and necrotic cells is also observed for complex 4 (12.7%) at 0.25 μM and for complex 8 (12.1%) at 0.5 μM.
The interaction of lead drug candidates with DNA was further studied to reveal peculiarities in their behavior.
2.8. Lead Cu(II) Complex 8 Interacts with Calf Thymus (ct)-DNA More Effectively than Complex 4. The interaction of 4 and 8 and HL 4 and HL 8 with ct-DNA was investigated by spectrofluorometry in ethidium bromide (EtBr) displacement studies. EtBr is a fluorescent probe, and its fluorescence intensity increases upon intercalation into the DNA helix. The ligands did not affect the fluorescence of the EtBr−ct-DNA system ( Figure  S8). Addition of complexes, however, decreased the emission intensity. Interestingly, the solubility of 8 (which was low) increased significantly in the presence of ct-DNA, and this complex reduced most significantly the fluorescence of EtBr. Decrease in the fluorescence may indicate (i) the displacement Journal of Medicinal Chemistry pubs.acs.org/jmc Article of EtBr or (ii) (partial) quenching of the fluorescence of the bound probe. In order to separate these processes, fluorescence lifetime measurements were carried out. These experiments indicated that the decrease in intensity is due to both EtBr displacement and alterations in the close environment of the intercalated EtBr (see Figure S9 in the Supporting Information). In addition, they provided evidence that the two copper(II) complexes bind to ct-DNA. However, complex 8 replaced EtBr more effectively than 4. To get further insight into the mechanism of action of lead compounds, their antiproliferative activity in wild-type cells HCT116 and HCT116 cell subline with knocked out p53 gene was investigated.
2.9. Is DNA a Crucial Target for Lead Drug Candidates HL 4 , HL 8 , 4, and 8? The oncosuppressor protein p53 controls Journal of Medicinal Chemistry pubs.acs.org/jmc Article the cellular response to DNA strand breaks induced by cytotoxic drugs or by radiation. 58 The p53 protein may enhance cell chemosensitivity by promoting apoptosis via different mechanisms including activation of proapoptotic genes such as bax and repression of antiapoptotic genes such as bcl-2 or in contrast increase chemoresistance by promoting p21-mediated and p21independent growth arrest and DNA repair and by activation of antiapoptotic genes such as bcl-x. There is strong evidence that the modulation of drug sensitivity by p53 may be both drug-and cell type-specific. 59 Targeted p53 inactivation in human cancer cells was shown to enhance their chemosensitivity to the drugs able to induce DNA strand breaks such as doxorubicin but at the same time make them quite resistant to drugs with other nucleicacid-related mechanisms of action, e.g., 5-fluorouracil (5-FU). Accordingly, essential differences in chemosensitivity were observed between cells with wild-type p53 gene and cells with knocked out p53 gene by homologous recombination. 60 Based on this knowledge, we used two isogenic cell lines, namely the wild-type HCT116 and HCT116 cell line with knocked out p53 gene, and treated them with our lead compounds as well as by cisplatin used as a DNA-damaging (positive control) drug. The results of these MTT assays summarized in Table 3 clearly show that the sensitivity of a p53-deficient HCT116 subline toward compounds HL 4 , HL 8 , 4, and 8 remains intact when compared to wild-type cells with proficient p53 gene. The sensitivity data are in strong contrast with the response of the cells to DNA cross-linking drug cisplatin, which showed lowered cytotoxicity in the subline with knocked out p53 gene. The data obtained strongly suggest that DNA is not a crucial target for the lead drug candidates evaluated in this study. Among other possible mechanisms underlying the antiproliferative activity of the lead drug candidates, kinase inhibition was further considered.  Table  2), were submitted to the International Centre for Kinase Profiling at Dundee University and screened against 50 enzymes using an inhibitor concentration of 10 μM. Figure 7 (see also  Table S4) summarizes the results of this assay as a histogram plotting the percentage of the remaining enzyme activity (x-axis) as a function of added lead compound (HL 8 : blue trace, 8: red trace) for each of the 50 enzymes assayed (y-axis). These data revealed fully distinct enzyme inhibitory patterns and selectivity for HL 8 and its complex 8. HL 8 showed good selectivity and notable potency for one of the 50 kinases assayed, namely for the serine and threonine protein kinase PIM-1, while 8 significantly inhibited (below 68% of original activity) the activity of five enzymes, namely of serum and glucocorticoid-regulated kinase SGK-1, cAMP-dependent protein kinase PKA, calcium/calmodulin-dependent protein kinase CaMK-1, mitogen stressactivated kinase MSK1, and glycogen synthase kinase GSK3β. Thus, the coordination to copper(II) completely changes the kinase inhibition profile of HL 8 .
A closer look at these proteins reveals common features. All of these are serine/threonine-protein kinases, which have at least one ATP binding site, indicating a competitive behavior with this molecule. They use ATP to phosphorylate protein residues, e.g., L-serine one. 61−66 SGK-1 is closely related with cancer growth, survival, and metastasis in a variety of tumors; in these malign tissues, SGK-1 is upregulated. 67−71 By downregulating this kinase, tumor growth and metastasis can be slowed down or even stopped. PKA is an important kinase often found in mitochondria that is able to modulate the energy household. 72,73 Since cancer cells have a very high demand for energy, suppression of that enzyme can assist in starving them. MSK1 is related to cancer growth, metastasis, and increased aggressiveness with an overall poor survival for certain types of cancers. 74,75 By downregulating this protein, the favorable effect of cancer therapy might be enhanced. CaMK-1 is expressed in all tissues but overexpressed in cancers. 76,77 Furthermore, there is evidence that CaMK-1 impacts chemoresistance in ovarian cancer. 78 This overexpressed kinase that affects cancer survival and growth by controlling the cell cycle 77 offers a potent target for anticancer therapy. Upregulation of PIM-1 is directly connected with tumor progression, survival, and even transformation 79−81 and, therefore, is a good target for chemotherapy. Due to its ability to initiate the transformation of healthy cells to malign cells, it is considered a proto-oncogene. Resveratrol inhibits PIM-1 activity via binding to the ATP pocket, reducing cancer cell proliferation and survival. 82 Both compounds were screened against the respective enzymes, and IC 50 values were determined. The IC 50 Table S5 and Figures S10−S15). GSK3β was chosen for determining IC 50 values since the previously reported paullones showed both in silico and in vitro inhibition of this enzyme. 22,83,84 Hence, we are prone to assume that mono-kinase and multikinase inhibition is a more plausible underlying mechanism of cytotoxicity for organic lead drug candidate HL 8 , and the copper(II) complex 8. Even though the multikinase inhibitory activity of HL 8 is likely, it still has to be confirmed by increasing the panel of available kinases. At the same time inhibition of enzymes might not be the only and also not the main mode of action contributing to cell death for the selected molecules, and other underlying mechanisms might be responsible for the apoptosis. Among other possible targets, tubulin is worthy to be mentioned as some of the indolo[2,3-d]benzazepine derivatives were reported to effectively inhibit tubulin polymerization. 30 To further provide evidence for the potential binding of copper(II) complexes and the respective ligands to PIM-1, CaMK-1, SGK-1, PKA, and GSK3β kinases and, in particular, for the lead drug candidates HL 8 and 8, molecular docking calculations were conducted.  The available X-ray diffraction structures in Protein Data Bank contain co-crystalized ligands, which were removed and redocked into the binding sites of the kinases to test the robustness of the scoring functions used, namely, GoldScore (GS), 90 ChemScore (CS), 91,92 ChemPLP (Piecewise Linear Potential), 93 and ASP (Astex Statistical Potential) 94 embedded in the GOLD (v2020.2.0) docking algorithm. The predicted poses were overlaid with the co-crystalized ligands, and the rootmean-square deviation (RMSD) was calculated for the heavy atoms. The results are shown in Tables S6−S11 in the Supporting Information. In general, good results were obtained. Molecular docking with the ligands HL 1 −HL 8 and copper(II) complexes 1−8 showed reasonable scores for the five kinases (see Tables S6−S11 in the Supporting Information), implying potential binding to the respective kinase pockets. In the case of copper(II) complexes, the GS scoring function was used. The PIM-1 scores (Table S7) indicate that the metal complexes and their ligands bind with greater affinity than the co-crystalized ligand LI7. It should be also stressed that HL 1 −HL 8 showed better scores than the complexes 1−8, i.e., for 8 and HL 8 , the metal-free ligand has a considerable better score. This is in line with the experimental IC 50 value where the ligand HL 8 gave  18 μM), whereas the Cu(II) complex did not register binding. Table S8 contains the results for the CaMK-1 kinase, and the scores for both the copper(II) complexes and ligands are comparable to those of the co-crystalized ligand J60. Complex 8 has the best score of the Cu(II) complexes, which fits the experimental results in being the most active at least in the Colo205 cell-based assay (see Table 2). The scores for the GSK3β kinase (Table S9) have lower values than the cocrystallized ligand Z48, and similar scores were obtained for both the ligands and copper(II) complexes, but 8 has the best score of the complexes and a considerably better score than its HL 8 counterpart. For the PKA enzyme (Table S10), the ligands have better scores than the CMP co-crystallized ligand only for CS, similar to ChemPLP but worse scores for the other functions as well as for the complexes, which fits the modest IC 50 value of 6.69 μM for 8. Finally, the results for the SGK-1 kinase (Table  S11), the ligands, and the complexes have better scores than the co-crystallized MMG ligand; the complex 8 has a better GS than HL 8 . The modeling for PIM-1 revealed that the ligands can adopt two plausible conformations in the binding pocket, whereas the copper(II) complexes are predicted to bind in an unfavorable pose, e.g., complex 8 has its morpholine ring buried within the binding pocket, while its bromine-containing moiety is pointing into the aqueous phase, which is in line with only the HL 8 ligand giving good binding results. It is important to note that the metal complexes were docked in their forms shown in Chart 2 (CuCl 2 (HL)); however, they are present in their [Cu(L)] + forms in aqueous solution at pH 7.4 based on the solution equilibrium studies, and the displacement of the coordinated chlorido ligands by the side chain donor atoms of the enzymes is also possible. Thus, the coordinative binding of the complexes to the proteins cannot be excluded. The predicted poses for HL 8 are shown in Figure 8, while those for complex 8 are shown in Figure S16. The tetracyclic motif of HL 8 for both predicted poses overlap with the co-crystalized ligand LI7 but with the morpholine-containing fragments occupying different clefts on the enzyme's surface.
On the one hand, the modeling for CaMK-1 using the ligands resulted in several different predicted conformations, indicating poor binding reflected in the experimental binding results. On the other hand, the complexes had consistent pose prediction neatly overlapping the J60 co-crystalized ligand as shown in Figure 9A. The general binding of J60 and 8 is similar in that the chlorine atom of J60 and the bromine atom in 8 are in a similar position, as well as the solubilizing groups of both, i.e., they are pointing into the water environment. This is a strong indication that 8 is tightly bound to CaMK-1 as seen in the IC 50 value of 0.75 μM. Furthermore, the copper(II) ion may potentially bind to the oxygen atoms in the carboxylic side chain of Glu105 with a 4.7 Å distance between copper(II) and the proximal oxygen atom. The binding is plausible since the side chain of Glu105 is quite flexible with three aliphatic carbon atoms ( Figure 9B).
The modeling for GSK3β showed that the ligands have two plausible binding modes, and the copper(II) complexes have a good fit in the pocket completely overlapping the co-crystallized Z48 ligand as shown in Figure 10A, indicating good binding. Cys199 is relatively close to the copper atom at 4.5 Å, but this ; the co-crystalized ligand LI7 is shown in ball-and-stick format, and its hydrogens are not shown for clarity (see circled area). The configuration of the ASP prediction is shown in green, and the ChemPLP pose is blue, both are in the stick format. The protein surface is rendered; blue depicts regions with a partial positive charge on the surface, red depicts regions with a partial negative charge, and gray shows neutral areas.  Journal of Medicinal Chemistry pubs.acs.org/jmc Article amino acid residue is embedded deep within the protein, and it is therefore questionable how far the mercaptan moiety can reach ( Figure 10B). Docking studies of 8 in SGK-1 and PKA both show good overlap with the co-crystallized ligands. However, the binding can be characterized as weak. For both enzymes, the hydrophilic morpholine moiety is buried in the hydrophobic binding pocket and the bromine substituent pointing into the hydrophilic area. The results as well as the docking parameters are shown in Figures S17 and S18 and Tables S10 and S11 in the Supporting Information.
Thus, the molecular docking calculations are in agreement with the results of the kinase inhibition assays; good binding is seen for HL 8 to PIM-1 but not its copper(II) complex 8. CaMK-1 and GSK3β both are predicted to have good binding to 8 in agreement with the kinase binding data.
2.12. Molecular Descriptors Indicate Drug-like Chemical Space for Lead Drug Candidates. The calculated molecular descriptors MW (molecular weight), log P (octanol− water partition coefficient), HD (hydrogen bond donors), HA (hydrogen bond acceptors), PSA (polar surface area), and RB (rotatable bonds) are given in Table S13 derived using the QikProp software. 95 QikProp is not parameterized for copper-(II) complexes. Therefore, Scigress 96 was used instead with the available MW, log P, HD, and HA descriptors (Table S14). The values for the proligands' descriptors lie mostly within drug-like chemical space with some exceptions. HD is in lead-like chemical space and for some proligands, MW and log P reach into the known drug space (KDS) (for the definition of lead-like, drug-like, and KDS regions see ref 97 and Table S12). The complexes obviously have higher MW and are all in the KDS, but the HD and HA remained intact (Table S14).
The known drug indexes (KDIs) for the proligands were calculated to gauge the balance of the molecular descriptors (MW, log P, HD, HA, PSA, and RB). This method is based on the analysis of drugs in clinical use, i.e., the statistical distribution of each descriptor is fitted to a Gaussian function and normalized to 1, resulting in a weighted index. Both the summation of the indexes (KDI 2A ) and multiplication (KDI 2B ) methods were used 98 as shown for KDI 2A in eq 1 and for KDI 2B in eq 2; the numerical results are given in Table S13 The KDI 2A values for the proligands range from 4.68 to 5.46 with a theoretical maximum of 6 and the average of 4.08 (±1.27) for known drugs. The KDI 2B range is from 0.14 to 0.55, with a theoretical maximum of 1 and with a KDS average of 0. 18 (±0.20). This means that the molecular descriptors' balance is reasonable with good biocompatibility. The low values can be explained by the high MW and log P values of some the compounds.
The proligands contain two imine groups linked by a N−N single bond. These groups are quite electron-rich, rendering them susceptible to an electrophilic attack. To test this, the ionization potential (one-electron oxidation) and electron affinity (one-electron reduction) were derived for HL 1 using DFT and compared to the statistical distribution of known drugs. 99 The ionization potential is 6.6 eV, and 95% of drugs lie in the 6.0−9.0 eV range. The electron affinity is −1.3 eV with drugs in the −1.5 to 2.0 eV range. 99 Thus, HL 1 is within the ranges of known drugs albeit with relatively low values. The bond dissociation energy (BDE) for the N−N single bond was also derived using DFT, resulting in 60.9 kcal/mol, which is substantially higher than the average for drugs (53.9 kcal/mol, n = 23). 100  The determined pK a values for HL 1 indicate that this morpholine-indolo[2,3-d]benzazepine hybrid is present in its neutral form at pH 7.4, and a similar behavior is suggested for the other ligands of the series. These are predicted to possess high lipophilic character, especially in the case of the bromosubstituted compounds (HL 2 and HL 6 ). The thermodynamic solubility of copper(II) complexes was higher than that of the corresponding organic hybrids. The bromo-substituted derivatives showed lower aqueous solubility, unlike at the Schiff-base bond-methylated compound, where the effect was negligible. The determined stability constants of copper(II) complexes indicate high thermodynamic stability, and a low extent of dissociation is suggested at low micromolar concentrations at neutral pH. At this pH, the [Cu(L)] + species predominates, in which the non-coordinating hydrazonic nitrogen is most likely deprotonated.

CONCLUSIONS
Overall, the compounds studied in this work are highly antiproliferative in cancer cells and deserve further development as potential anticancer drugs. The indolo[2,3-d]benzazepine proligands HL 1 −HL 4 were found to be generally less cytotoxic than the analogous indolo[2,3-c]quinoline compounds HL 5 − HL 8 against human colon adenocarcinoma cell lines, and the presence of the bromo-substituent enhanced the selectivity for cancer cells. The copper(II) complexes are more cytotoxic than the corresponding metal-free hybrids. Some of the compounds (HL 7 , 7, HL 8 , and 8) displayed very good selectivity against cancer cells over the normal ones. Complex 4 was more cytotoxic against the resistant Colo320 cells in comparison to the sensitive cells, and it was more active than the 4 nm , which does not contain the morpholine moiety. Compounds HL 4 , 4, HL 8 , and 8 could trigger apoptosis in the multidrug-resistant Colo320 cells more efficiently than cisplatin. Unlike HL 4 and HL 8 , complexes 4 and 8 were also able to replace the intercalative EtBr from ct-DNA. However, MTT assays with wild-type HCT116 cells with intact p53 gene and its p53deficient subline strongly suggest that DNA is not an effective target responsible for the inhibition of cell proliferation by lead drug candidates HL 4 , 4, HL 8 , and 8.

Journal of Medicinal Chemistry
pubs.acs.org/jmc Article Enzyme inhibition assays against a panel of 50 related kinases revealed fully distinct inhibitory profiles for HL 8 and 8, stressing again a special role of the metal on both the antiproliferative activity and the underlying mechanisms of cytotoxicity of this class of compounds at the molecular level. HL 8 proved to be selective and a potent inhibitor of PIM-1, while 8 strongly inhibited the activity of five other enzymes, namely, SGK-1, PKA, CaMK-1, MSK1, and GSK3β. The disclosed kinase inhibition was further supported by molecular docking calculations.
To further increase the affinity and selectivity of the lead drug candidates for particular enzymes, modification of lead structures by introduction of substituents and functional groups, which will allow for adjusting their electronic and steric properties, is needed. This work is going on in our laboratory and will be reported in due course.

EXPERIMENTAL SECTION
4.1. Starting Materials. 2-Formyl-5-(morpholinomethyl)pyridine (G in Scheme 1) was synthesized in several steps as described elsewhere. 44 The starting backbones K and L (see Scheme 3) were prepared as reported recently. 30,43 4.2. Synthesis of Precursors and Metal-Free Ligands. The isolated yield and analytical data for HL 1 −HL 8 are summarized in Tables S15 and S16. The experimental CHN contents were within ±0.4% with those calculated, providing evidence for >95% purity. 4.2.2. 2-Acetyl-5-(morpholinomethyl)pyridine (J in Scheme 1). Oxalylchloride (517 μL, 5.93 mmol) was diluted with dry DCM (9 mL) and cooled to −80°C. Dry DMSO (890 μL, 12.5 mmol) was added dropwise. Five minutes later, to this solution, a solution of I (1.29 g, 5.80 mmol) in DCM (10 mL) was added. The reaction mixture was stirred at −80°C for 15 min. Then, triethylamine (3.64 mL) was added dropwise (the solution turned turbid) and stirring was continued overnight while the reaction mixture slowly reached room temperature. The reaction was quenched with water (20 mL). The organic phase was separated, and the aqueous phase was extracted with DCM (3 × 20 mL). The combined organic phases were dried over MgSO 4 . The solvent was removed under reduced pressure, and the crude product was purified on silica by using DCM/MeOH 98:2 as an eluent to give the product as a pale-yellow solid. Yield: 878 mg, 69%. Anal. calcd for C 12

5-Bromo-1-(ethoxymethyl)-N-(2-iodophenyl)-1H-indole-2carboxamide (I).
Under an argon atmosphere, 2-iodoaniline (3.54 g, 16.16 mmol) was dissolved in dry DCM (10 mL). The solution was cooled to −20°C, and AlMe 3 (10.1 mL, 20.2 mmol) was added dropwise. The mixture was stirred at −20°C for 45 min before slowly reaching 0°C. Then, ethyl 5-bromo-1-ethoxymethyl-1H-indol-2carboxylate 43 (1.32 g, 4.04 mmol) in dry DCM (8 mL) was added dropwise. The mixture was stirred at room temperature for 18 h before being cooled to 0°C again. Then, 1 M HCl (90 mL, 90 mmol) was added slowly. The solution was stirred for 10 min at room temperature until phase separation was clearly seen. The aqueous phase was extracted with DCM (3 × 100 mL). The combined organic phases were dried over MgSO 4 and concentrated in vacuo. The crude product was refluxed in methanol (30 mL) and cooled to 4°C. The product was collected by filtration as a white, cloudy solid. Yield: 1.74 g, 84%. 1 30 (9.36 g, 40.0 mmol) and phosphorus oxychloride (140 mL) were mixed in a 250 mL nitrogen flask under an argon atmosphere. The mixture was refluxed overnight. The next day, it was cooled to room temperature and poured on ice water (800 mL), which was cooled additionally by an ice bath (strongly exothermic reaction). Solid sodium hydroxide was added in portions under cooling (strongly exothermic reaction) until a pH of about 12 was reached. Then, the reaction mixture was extracted with dichloromethane (3 × 500 mL). The organic phase was washed with water (3 × 200 mL) and brine (1 × 200 mL) and dried over magnesium sulfate. The solvent was removed under reduced pressure. Yield: 9.47 g, 94%. overnight. The product was precipitated with water. Ethanol (10 mL) was added to the suspension to dissolve the precipitate. The solution was slowly concentrated by 1/2 under reduced pressure. The product was isolated by filtration, washed with cold ethanol (1 mL), and dried in vacuo overnight to give a yellow powder. Yield: 398 mg. 1 Table S1. The structures were solved by direct methods and refined by full-matrix least-squares techniques. Non-H atoms were refined with anisotropic displacement parameters.  102 The average water ionization constant (pK w ) is 14.52 ± 0.05, which corresponds well to the literature data. 103 Argon was passed over the solutions during the titrations. Proton dissociation constants (pK a ) of the ligand, overall stability constants (log β) of the copper(II) complexes, and the individual spectra of the various species present in the solution were calculated by the computer program PSEQUAD. 104  Thermodynamic solubility of proligands (HL 5 and HL 5 nm ) and copper(II) complexes (3, 5, 5 nm , 6, 6 nm , 7, and 8) was measured for the saturated solutions in water at pH 5 (20 mM 2-(N-morpholino)ethanesulfonic acid (MES, Sigma Aldrich) buffer) and 7.4 (20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Sigma Aldrich) buffer) at 25.0 ± 0.1°C. The concentration of the compounds was determined by UV−vis spectrophotometry using stock solutions of the compounds with a known concentration dissolved in pure DMSO and 50% (v/v) DMSO/buffered aqueous solution for the calibration.
Aqueous stability of 4, 8, and HL 4 was investigated in phosphatebuffered saline (PBS) and 10 mM HEPES buffers and in three-times diluted blood serum (Sigma Aldrich, from human male AB plasma) at pH = 7.40. The concentration of the compounds was between 5 and 10 μM. Blood serum was filtered on a 1.25 μm polyethersulfone membrane syringe filter and diluted with 10 mM HEPES or PBS buffer.
4.8. DNA Binding Studies. Fluorescence measurements were carried out on a Fluoromax (Horiba Jobin Yvon) fluorometer. A stock solution of calf thymus DNA (ct-DNA, Sigma Aldrich) was prepared as described in our former work. 105 Samples contained 10 μM ct-DNA expressed in base pairs, 5 μM ethidium bromide (EtBr, Sigma Aldrich), and different concentrations of complexes 4 or 8 or proligands HL 4 or HL 8 in 10 mM HEPES buffer (pH = 7.40). The excitation wavelengths were 510 or 455 nm, and the fluorescence emission was measured in the range 530−750 nm. Corrections for self-absorbance and inner filter effect were done according to our former work. 106 Fluorescence lifetime was measured on the same fluorometer equipped with a DeltaHub time-correlated single photon counting (TCSPC) controller using a NanoLED light source N-455 (Horiba Jobin Yvon). Details on the instrument parameters are found in Table S3. The fluorescence intensity decay over time is described by a sum of exponentials, where α i and τ i are the contribution to the total intensity (I) at time point 0 and lifetime of component i, respectively. 107 The quality of the fit was judged from a χ 2 R value close to 1.0 and a random distribution of weighted residuals. Fractional intensities were calculated as described in our former work. 105 4.9. Cell Lines. Human colonic adenocarcinoma cell lines Colo 205 doxorubicin-sensitive (ATCC-CCL-222) and Colo 320/MDR-LRP multidrug resistant over-expressing ABCB1 (MDR1)-LRP (ATCC-CCL-220.1) were purchased from LGC Promochem, Teddington, UK. The cells were cultured in an RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, 1 mM Napyruvate, and 10 mM HEPES. The semi-adherent human colon cancer cells were detached with Trypsin-Versene (EDTA) solution for 5 min at 37°C. An MRC-5 human embryonal lung fibroblast cell line (ATCC CCL-171) was purchased from LGC Promochem, Teddington, UK. The cells were cultured in Eagle's minimal essential medium (EMEM, containing 4.5 g L −1 glucose) supplemented with a non-essential amino acid mixture, a selection of vitamins, and 10% heat-inactivated fetal bovine serum. The HCT-116 colon carcinoma cells and isogenic p53knock-out subline were a gift of Prof. Bert Vogelstein from The Johns Hopkins Oncology Center. Cells were cultured in McCoy's 5A medium supplemented with 10% heat-inactivated fetal bovine serum and 4 mM L-glutamine. All cell lines were incubated at 37°C, in a 5% CO 2 , 95% air atmosphere.
4.10. Assay for Cytotoxic Effect. MRC-5 non-cancerous human embryonic lung fibroblast and human colonic adeno-carcinoma cell lines (doxorubicin-sensitive Colo 205 and multidrug resistant Colo 320 colonic adenocarcinoma cells) were used to determine the effect of compounds on cell growth. The effects of increasing concentrations of compounds on cell growth were tested in 96-well flat-bottomed microtiter plates. The compounds were dissolved in DMSO and stock solutions of 10 mM were prepared. These were further diluted in the appropriate cell culture medium by twofold serial dilution starting from 100 or 10 μM for the compounds. The adherent human embryonal lung fibroblast cells were cultured in 96-well flat-bottomed microtiter plates using EMEM supplemented with 10% heat-inactivated fetal bovine serum. The density of the cells was adjusted to 1 × 10 4 cells in 100 μL per well, the cells were seeded for 24 h at 37°C, 5% CO 2 , then the medium was removed from the plates containing the cells, and the dilutions of compounds previously made in a separate plate were added to the cells in 200 μL. In the case of the colonic adenocarcinoma cells, the twofold serial dilutions of compounds were prepared in 100 μL of RPMI 1640, horizontally. The semi-adherent colonic adenocarcinoma cells were treated with Trypsin-Versene (EDTA) solution. They were adjusted to a density of 1 × 10 4 cells in 100 μL of RPMI 1640 medium and were added to each well, with the exception of the medium control wells. The final volume of the wells containing compounds and cells was 200 μL. The culture plates were incubated at 37°C for 72 h; at the end of the incubation period, 20 μL of MTT (thiazolyl blue tetrazolium bromide, Sigma) solution (from a stock solution of 5 mg mL −1 ) was added to each well. After incubation at 37°C for 4 h, 100 μL of sodium dodecyl sulfate (SDS) (Sigma) solution (10% in 0.01 M HCI) was added to each well and the plates were further incubated at 37°C overnight. Cell growth was determined by measuring the optical density (OD) at 540/630 nm with a Multiscan EX ELISA reader (Thermo Labsystems, Cheshire, WA, USA). Inhibition of the cell growth (expressed as IC 50 : inhibitory concentration that reduces by 50% the growth of the cells exposed to the tested compounds) was determined from the sigmoid curve where 100 − ((OD sample − OD medium control )/ (OD cell control − OD medium control )) × 100 values were plotted against the logarithm of compound concentrations. Curves were fitted by GraphPad Prism software 57 using the sigmoidal dose−response model (comparing variable and fixed slopes). The IC 50 values were obtained from at least three independent experiments. Tests in adherent HCT-116 colon carcinoma cells and an isogenic p53-knockout subline thereof were performed in a similar manner, with 1.5 × 10 3 cells seeded in 100 μL of McCoy's 5A medium (Sigma-Aldrich) per well. From fresh stock solutions in DMSO (10 mM HL 4 and HL 8 and 2 mM 4 and 8), test compounds were diluted in a medium in adjusted concentration ranges and applied for 96 h. Upon removal of the medium containing the test compounds and 4 h of staining with 100 μL of MTT/medium mixture, the latter was exchanged for 150 μL of DMSO per well and photometric measurement performed immediately thereafter.
4.11. Mechanisms of Cell Death: Assay for Apoptosis Induction. The assay was carried out for selected compounds using an Annexin V-FITC Apoptosis Detection Kit (cat. no. APOAF-50TST) from Sigma according to the manufacturer's instructions. The concentration of the Colo320 cell suspension was adjusted to approximately 0.5 × 10 6 cells/mL in an RPMI 1640 medium, and the cell suspension was distributed in 1 mL of aliquots into a 24-well plate and then incubated overnight at 37°C and 5% CO 2 . On the following day, the medium was removed and replaced by 1 mL of RPMI medium containing the compounds except the control samples. Colo320 cells were incubated in the presence of the compounds at 2 and 4 μM (HL 8 and HL 4 ), or 0.5 and 2 μM (8), or 0.25 and 0.5 μM (4), in the 24-well plate at 37°C for 3 h, and 12H-benzo[α]phenothiazine (M627, 20 μM) 108 and cisplatin (Teva, 15 and 30 μM) were used as positive controls. After the incubation period, the samples were washed with PBS and fresh RPMI 1640 medium was added to the samples. The cells were incubated overnight at 37°C and 5% CO 2 . The next day, 200 μL of 0.25% Trypsin (Trypsin-Versen) was added to the samples until cells appeared detached followed by the addition of 400 μL of RPMI 1640 medium supplemented with 10% bovine serum. The cells were collected in Eppendorf tubes and centrifuged at 2000g for 2 min. The harvested cells were resuspended in fresh serum-free RPMI 1640 culture medium. After this step, the apoptosis assay was carried out according to the instructions of the manufacturer. The fluorescence was analyzed immediately using a ParTec CyFlow flow cytometer (Partec GmbH, Munster, Germany).
4.12. Enzyme Inhibition Tests. All kinase assays were carried out using Multidrop 384's at room temperature in a total assay volume of 25.5 μL. To plates containing 0.5 μL of HL 8 and 8, DMSO control or acid blank (0.51 mM), 15 μL of an enzyme mix containing enzyme (0.51 mM), and excess peptide/protein substrate in buffer was added. Compounds were pre-incubated in the presence of the enzyme and peptide/protein substrate for 5 min before initiation of the reaction by addition of 10 μL of ATP (final concentration selected for each kinase at 5, 20, or 50 μM). Assays were carried out for 30 min at room temperature before termination by the addition of 5 μL of orthophosphoric acid. The assay plates were then harvested onto P81 Unifilter Plates by a PerkinElmer Harvester and air dried. The dry Unifilter plates were then sealed on the addition of MicroScint O and were counted in PerkinElmer Topcount scintillation counters.
4.13. Determination of IC 50 Values against Protein Kinases. IC 50 values were determined by the International Centre for Kinase Profiling at the University of Dundee according to standard protocols published elsewhere. 109,110 4.14. Molecular Docking. The ligands and complexes were docked against the crystal structures of PIM-1 (PDB ID: 1YXX, resolution: 2.00 Å), 85 calmodulin-dependent protein kinase I G (CaMK-1) (PDB ID: 2JAM, resolution: 1.70 Å), GSK3β (PDB ID: 3I4B, resolution: 2.30 Å), 87 PKA (PDB ID: 3OF1, resolution: 2.21 Å), 88 and SGK-1 (PDB ID: 3HDM, resolution: 2.60 Å) 89 kinases, which were obtained from the Protein Data Bank (PDB). 111,112 The GOLD (v2020.2.0) software suite was used to prepare the crystal structures for docking, i.e., the hydrogen atoms were added, water molecules deleted, and the co-crystallized ligands identified: PIM-1: . The docking center for the binding pockets was defined as the position of the co-crystallized ligands with a 10 Å radius. The GoldScore (GS) 90 and ChemScore (CS) 91,92 ChemPLP (Piecewise Linear Potential) 93 and ASP (Astex Statistical Potential) 94 scoring functions were implemented to predict the binding modes and relative energies of the ligands using the GOLD (v2020.2.0) software suite. The GOLD docking algorithm is reported to be an excellent modeling tool. 113,114 The Scigress version FJ 2.6 program 96 was used to build the ligands and complexes; the MM3 115−117 force field was applied to identify the global minimum using the CONFLEX method 118  The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.1c01740.