Synthesis, Characterization and Evaluation of the Antibacterial and Antitumor Activity of Halogenated Salen Copper(II) Complexes derived from Camphoric Acid

Platinum metal complexes are the most common chemotherapeutics currently used in cancer treatment. However, the frequent adverse effects, as well as acquired resistance by tumor cells, urge the development of effective alternatives. In the recent past, copper complexes with Schiff base ligands have emerged as good alternatives, showing interesting results. Accordingly, and in continuation of previous studies in this area, three new camphoric acid-derived halogenated salen ligands and their corresponding Cu(II) complexes were synthesized and their antitumor activity was evaluated in order to determine the influence of the type and number of halogens present (Br, Cl). The in vitro cytotoxic activity was screened against colorectal WiDr and LS1034 and against breast MCF-7 and HCC1806 cancer cell lines. The results proved the halogenated complexes to be very efficient, the tetrachlorinated Cu(II) complex being the most promising, presenting IC50 of 0.63-1.09 μM for the cell lines studied. The complex also shows selectivity to colorectal cancer cells compared to non-tumor colon cells. It is worth highlighting that the tetrachlorinated Cu(II) complex, our most efficient complex, shows a significantly more powerful antitumor effect than the reference drugs currently used in conventional chemotherapy. The halogenated salen and corresponding complexes were also screened for their antimicrobial activity against four bacterial species-Staphylococcus aureus, Enterococcus faecalis, Escherichia coli and Pseudomonas aeruginosa-and four fungal species-Candida albicans, Candida glabrata, Aspergillus fumigatus and Alternaria alternata. The compounds were found to exhibit moderate to strong antibacterial activity against the bacterial strains studied. NMR studies and theoretical calculations provided some insight into the structure of the ligands and copper complexes. Considering the results presented herein, our work validates the potential use of copper-based chemotherapeutics as alternatives for cancer treatment.


Introduction
Cancer is one of the leading worldwide causes of death. Colorectal and breast cancer are two of the most incident and mortal types of cancer. [1] The discovery of platinum-based drugs in the twentieth century embodied an enormous breakthrough for cancer chemotherapy. Cisplatin, oxaliplatin and carboplatin are platinum metal complexes currently used in the treatment of cancer. However, these chemotherapeutic compounds, and others, namely, fluoropirimidines, irinotecan, taxanes and anthracyclines, besides having many adverse side effects, have led to the development of increased resistance of cancer cells to the pharmaceuticals used. [2] A similar problem exists concerning antibiotic resistance. Although this is an ancient phenomenon that occurs naturally in the environment, [3] it has now reached alarming proportions, with the emergence and spreading of multidrug-resistant bacteria, making the treatment of infectious diseases increasingly difficult. Several approaches to circumvent this problem are currently being investigated and one of these is the development of new metal-based drugs with different modes of action that are not affected by the current resistance mechanisms, since their targets are expected to be different.
Hence, in order to solve these problems, it is of fundamental importance to search for alternative therapeutic agents, with improved pharmacological properties and minimal negative characteristics.
In the recent past considerable attention has been given to the use of metal complexes in medicine for various types of therapy and diagnosis. In cancer treatment, over the last decades, many non-Pt metal complexes have emerged and have been investigated as potential cancer treatments. Among these, copper complexes seem to be good alternatives to Pt complexes and those with Schiff base ligands have shown interesting results. [4]- [6] The biological activity of many Schiff base and salen metal complexes, Figure 1, has been recognized, specifically with respect to antimicrobial, antifungal, antiviral, antioxidant, anti-inflammatory, insecticidal and anticancer activities. [4][6]- [11]  Among these complexes, the salen metal complexes are emerging as potential therapeutic agents. The excellent chelating ability of these compounds with many metals is responsible for their improved biological activity. [8] These complexes are very attractive alternatives due to the versatility of the salen ligands, which result from the condensation of amines with salicylaldehydes or aromatic hydroxyketones, and thus can be easily fine-tuned both sterically and electronically by varying one or both of the reagents used. The structural diversity and reactivity of these complexes also makes them very useful in organic synthesis, particularly in catalytic processes, namely, alkene epoxidations, hydroxylations, cyanations and alkylations, among others.
Increasing attention is being given to the study of the therapeutic activity of salen metal complexes, with greater focus on copper, iron and manganese complexes. [8] [9][12] [13] Our previous studies on this topic involved the assessment of the in vitro cytotoxic activity of Cu(II), Fe(III) and Mn(III) metal complexes of (1R,3S)-N,N'-bis(salicylidene)-1,3-diamino-1,2,2trimethylcyclopentane and of the corresponding dibrominated derivative, (1R,3S)-N,N'-bis(5bromosalicylidene)-1,3-diamino-1,2,2-trimethylcyclopentane, against human melanoma, colorectal and breast cancer cell lines. Our studies showed that Cu(II) complexes presented higher antitumour activity, possibly due to the square planar geometry of the complexes which can favor their interaction with biological targets. [14] Our results further indicated that the copper (II) complex of the dibrominated derivative showed the highest cytotoxic activity towards all cell lines studied, with IC50 values of 0.95-2.32 M, a 3-5 fold increase when compared to the non-halogenated counterpart, [10] a marked improvement relatively to current conventional chemotherapy. These results are in agreement with published studies, which refer that many pharmaceuticals that incorporate halogen atoms in their structures show enhanced biological activity concerning cytotoxic effects, as well as improved pharmacodynamic and pharmacokinetic properties. [15]- [17] These results led us to question whether increasing the number of bromine atoms in the structure or exchanging bromine for chlorine would have any effect on the cytotoxicity of the complexes. With this in mind, we synthesized the analogous tetrabrominated, tetrachlorinated and dichlorinated salen

Nuclear Magnetic Resonance Studies
The 1 H and 13 C NMR spectra were obtained on a Bruker Avance III HD 500 MHz NMR spectrometer.
The residual signal of the solvent, CDCl3, was used as the internal reference for 1 H (δ 7.27) and 13 C (δ 77.23) shifts. 2D NMR spectra, COSY, NOESY, HSQC and HMBC were recorded on the same spectrometer. The 13 C spectra were recorded using proton decoupling techniques, taking advantage of the nuclear Overhauser effect.

Theoretical Studies
All calculations were carried out with the ORCA electronic structure package version 4.0.0.2. [19] Structures were optimized at the density functional theory (DFT) level using the meta-GGA exchange-correlation functional of Tao et al. (TPSS) with the flexible multiply polarized triple-Def2-TZVP basis set. [20]- [22] Dispersion corrections to the TPSS functional were added using Grimme's D3 method. [23] [24] Solvation effects (water) were treated with the SMD continuum solvent model. [25] [26] The same level of theory as was in the geometry optimizations was also employed in a subsequent ab-initio molecular dynamics (AIMD) simulation to ensure compatibility with the minimized structures. The AIMD simulation used a time-step of 0.5 fs and was run for 1000 steps with thermostatic coupling to a 335K heat-bath. The Berendsen thermostat was employed to regulate the temperature with the coupling parameter  set to 10 fs.
In order to evaluate the effects of halogenation on the global characteristics of the complexes the total electronic dipole moment, µ, and the isotropically averaged polarizability, α, were calculated at the TPSS-D3/Def2-TZVP level of theory with SMD aqueous solvation. The polarizability α was calculated using the coupled-perturbed SCF approach. The local effects of halogenation were evaluated by fitting condensed atomic charges to the molecular electrostatic potential using the CHELPG method. [19]

Biological Activity Studies
Cell Culture . Cell lines were propagated as previously described. [10] Metabolic Activity Cancer cells were seeded in 48-well plates in a concentration of 50-70x10 5 cells/mL. After 24 hours, cells were treated with the synthesized complexes (0.5 to 5 µM) for 48 hours. Cell proliferation was then assessed by the colorimetric assay MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), which allows to measure metabolically active cells, through the quantification of formazan crystals produced, due to the relationship of direct proportionality between the quantity of crystals and the cells' metabolic activity. After the dissolution of the formazan crystals, absorbances were measured in a spectrophotometer at 570 and 620 nm, and the data processed using Origin Pro 8.5, making it possible to establish dose-response curves and to determine the respective IC50 (halfmaximal inhibitory concentration) values.

Flow Cytometry
The effect of the synthesized compounds on cell viability and cell cycle was evaluated through flow cytometry, using the cytometer FACSCalibur. 1x10 6 cells were used per experience, for three conditions: the control, the concentration corresponding to the IC50 of each cell line and 5 µM. All cell lines were labelled with annexin V bound to the fluorochrome fluorescein isothiocyanate (AV-FITC, Immunostep, ANXVF-200T) and propidium iodide (Immunostep, PI), as previously described. [27] In order to evaluate alterations on the cell cycle, tumor cell lines were fixed and labelled with a solution of PI and RNAse (Immunostep, PI/Rnase), as previously described. [28] The number of events obtained through the CellQuest TM program, corresponding to the number of cells, was 1x10 4 for the annexin and propidium iodide (AV/IP) double labelling and 2.5x10 4 for cell cycle study. The analysis and quantification of the information was made using a specific software that processes in a dedicated computer (Paint-a-Gate 3.02, Macintosh Software).

Antimicrobial Activity
Fresh stock solutions of salen ligands 2a-c and copper(II) complexes 3a-c were prepared in dimethyl sulfoxide (DMSO) at a concentration of 10 mg/mL (eq. to 15 mM for 2a; 14 mM for 3a; 20 mM for 2c; 18 mM for 3c; 24 mM for 2b; 21 mM for 3b).
The bacteria and fungi used in this study belong to the collection of the Institute of Microbiology, FMUC and were obtained from clinical samples with the exception of Enterococcus faecalis, that was purchased from American type culture collection (ATCC® 29212™).
The antibacterial activity of all compounds was tested against two Gram-positive bacteria -S. aureus and E. faecalis -and two Gram-negative bacteria -E. coli and P. aeruginosa -according to standard M07-A10 of the Clinical Laboratory Standards Institute. [29] Briefly, overnight cultures of each strain were transferred into a tube containing 2 mL of normal saline (NaCl 0.9%), until the turbidity standard 0.5 McFarland.
A 96-well microtiter plate was inoculated with 100 µl of tested compounds ten-fold serial diluted in Muller-Hinton broth medium (over the range 1 ng/mL to 100 µg/mL) and 10 µL of the bacterial suspensions at a final concentration of 5x10 4 CFU/well. To ensure that DMSO had no effect on The decimal reduction ratio (RD) of the microbial population was determined by aspirating 0.1 mL of suspensions from the wells which did not show any growth after incubation during MIC assays and plated on Columbia agar medium followed by counting the CFU after 18-24 h of incubation at 37 ºC. RD was calculated according the formula: RD = log N0/N [30] wherein N0 is the CFU of untreated bacteria and N the CFU after antimicrobial treatment. A condition that causes one log10 reduction of the initial population corresponds to 90% kill of the initial population. [31] The antifungal activity was evaluated against two yeasts, C. albicans and C. glabrata, by M44.P standard of the National Committee for Clinical Laboratory Standards [32] using the paper disk diffusion method in Mueller-Hinton agar supplemented with 2% glucose and methylene blue and against two filamentous fungi, Asp. fumigatus and Alt. alternata, according to the method described by EUCAST standard 9.3.1 (2017) using the microdilution assay in RPMI medium.

Statistical Analysis
Statistical analysis was performed using the IBM™ SPSS™ 24.0 software (IBM Corporation, USA).
Normality of the quantitative variable distribution and variance homogeneity was assessed using Shapiro-Wilk and Levene tests, respectively. Student T-test or one-factor ANOVA were used in parametric analysis and Mann-Whitney or Kruskal-Wallis, in non-parametric analysis. For more than 2 groups, post-hoc analysis was also made using the Games-Howel test (in case of variances homogeneity) and using the Bonferroni correction (without variances homogeneity). A P value of less than 0.05 was considered statistically significant. (1) was obtained from (1R,3S)-camphoric acid in a simple one-step procedure. [18] Subsequent condensation of (1) with 2,5-dibromosalicylaldehyde, 5chlorosalicylaldehyde or 2,5-dichlorosalicylaldehyde in ethanol, in the presence of activated silica gel, under ultrasound irradiation for 30 minutes at room temperature, gave the halogenated salen ligands 2a-2c, Scheme 1.
Mass spectrometry and elemental analysis are in agreement with a 1:1 metal:ligand stoichiometry for all three complexes synthesized.

Nuclear Magnetic Resonance Studies
A study of the structure of the salen ligands and complexes was carried out by NMR using the tetrachlorinated derivatives 2c/3c as models, Scheme 2.

Scheme 2.
Assignments were made using homonuclear and heteronuclear 2D correlations ( Figures S1-S4, supplementary material). The 1 H and 13 C NMR spectra of 2c/3c were obtained in CDCl3-d1 solution, between various metal ions and relevant ligands, [33]- [41] including the paramagnetic copper(II) [38] and chromium(III) [37] ions.   The complex is paramagnetic and, because of the rapid nuclear relaxation induced by the paramagnetic properties of the Cu(II) metal ion, 1 H NMR signals of bound ligands (inner sphere) are broadened and/or shifted to lower and higher frequencies ( Figure 2 and Table 1 These findings are in agreement with similar effects observed in previous studies on the paramagnetic 8-hydroxyquinoline complex of Cr(III) [37] and other mononuclear complexes of the Cu(II) metal ion. [38] Accordingly, the 13 C NMR signals of the ligand in the complex are broadened (Figure 3 In contrast, binuclear copper(II) complexes show relatively sharp NMR line widths which are two orders of magnitude narrower than the mononuclear analog. [42] The observed broad signals support the hypothesis that the copper(II)/ligand complex is a mononuclear CuL species.

Geometries
To provide something to gauge the properties of the halogenated complexes against, it was decided where it can be seen that the lower aquo ligand has been ejected from the metal center and relocated to a position hydrogen-bonded to the two ligand oxygen atoms.
This water molecule was removed and the system re-optimized to give the structure shown in Figure   4 (right). Loss of the lower water had essentially no effect on the structure of the complex and in particular the metal environment suggesting that in this complex the preferred metal environment is square-pyramidal. In order to investigate the stability of the square pyramidal aquo complex an ab-initio molecular dynamics (AIMD) simulation was performed. Figure 5 shows two snapshots from the AIMD simulation. At 100 fs the axial water ligand has already dissociated from the metal center and is moving away from the complex. 200 fs into the simulation the water has moved further away and the remaining complex has lost its symmetry and the metal center is transitioning towards a distorted square planar geometry. Subsequent removal of the dissociated upper water molecule and reoptimization of the remaining complex resulted in only minor changes to the geometry

Electronic Properties
The electronic properties of the Cu(II) complexes were investigated using the previously optimized geometries for 3a-c and the non-halogenated complex. CHELPG charges for the Cu(II) center and the four phenyl ring substituent atoms at the halogen substitution positions are shown in Table 3  This reverses to being negative in all four positions in complexes 3a and 3c and due to the much greater size of the halogens compared with adjacent hydrogen atoms this suggests a uniform negatively charged surface on the edges of the phenyl rings which would lead to very different behavior in intermolecular interactions than would be expected for the non-halogenated complex.
Complex 3b, having only two chlorine substituents, displays a more oscillatory charge distribution amongst the four phenyl positions considered with larger positive and negative charges than seen in any of the other complexes. Table 3.
Several global electronic properties for the complexes are shown in Table 4 The total electric dipole moment µ undergoes a small increase on going from the non-halogenated complex to the dichlorinated 3b but can only be expected to produce a small change in the polarity of the complex. However, in both of the tetrahalogenated complexes 3a and 3c a much larger increase of approximately 50% is seen indicating significantly more polar character for these complexes.
The orientationally-averaged polarisability, α also increases with the degree and type of halogen substitution. Here, an increase of 6% and 12% are seen for the dichloro-and tetrachloro-complexes, respectively indicating a stepwise augmentation of α with the number of Cl atoms. As might be expected, the tetrabrominated complex 3a displays the largest α value with a 20% increase over that of that of the non-halogenated complex.

In Vitro Cytotoxic Activity
In order to determine the cytotoxicity of complexes 3a-c, their effects on the proliferation of four human cancer cell lines were first investigated. Cell lines were incubated with increasing concentrations of the complexes (0.5 to 5 μM), and subsequently, their metabolic activity was evaluated using MTT assay for 48 h. Cell proliferation was determined as a measure of metabolic activity relative to the control experiments. A sigmoid fitting was performed to determine the IC50 values. Results are presented in Figure 7 as dose-response curves.
In all human cancer cell lines, the increase in concentration of the Cu(II) complexes led to cell proliferation inhibition. Also, cell proliferation inhibition is dependent on the Cu(II) complex tested.
Tetrachlorinated Cu(II) complex 3c has the strongest anti-proliferative activity against HCC1806, MCF-7 and WiDr cells. Cell proliferation inhibition in LS1034 cell line was more evident when cells were treated with the dichlorinated Cu(II) complex 3b. Figure 7. Table 5 shows IC50 values for the three complexes studied. Included in the table, for comparative purposes, are the results of the analogous non-halogenated and dibrominated complexes previously studied by us. [10] Complex 3c is the most cytotoxic against three of the four cell lines studied, with On the other hand, the influence of the number of bromine atoms is not as obvious: the tetrabrominated complex 3a only presented higher antiproliferative effect than the dibrominated Cu(II) complex in LS1034 and HCC1806 cell lines. Table 5.
Comparison of IC50 values of reference drugs currently used in conventional chemotherapeutic protocols, with our most efficient complex, 3c, Table 6, prove this complex to be very promising. It shows a significantly more powerful cytotoxic effect than 5-FU, oxaliplatin and irinotecan against WiDr cells. A more effective cytotoxicity is also observed with 3c than with 5-FU and irinotecan against LS1034 cells and the complex is more potent than cisplatin against MCF-7 cells.
Being 3c the most promising complex, the antiproliferative activity of the corresponding free ligand 2c was evaluated, in order to ascertain the influence of metal on the antitumor effect. Figure 8 shows the results of this study. The importance of the metal, Cu(II), is obvious since the free ligand shows little or no effect on tumor cell proliferation (p<0.001 for all conditions above 1 µM). This data is in agreement with our previous studies concerning the dibrominated ligand and its corresponding Cu(II) complex. [10] It also coincides with reports referring that Cu(II) plays a crucial role in antitumor activity, inducing tumor cell death through various mechanisms, among which are ROS production, DNA damage and angiogenesis inhibition. [45]    The WiDr cell line was found to be the less sensitive to the complex. The reduction in cell viability was accompanied by an increase in cell death, normally by late apoptosis/necrosis and necrosis, except for LS1034 cells whose predominant type of cell death was early apoptosis. WiDr cells and HCC1806 cells die predominantly by necrosis, as indicated from the largest population of dead cells observed when they are exposed to 5 μM of 3c.
When analyzing data from non-tumor cells, the results showed that HFF1 had no response to 3c in Cell cycle studies were assessed using incubation of cell lines with IC50 and 5 μM of 3c through staining with IP/RNAase. Figure 10 shows that 3c, in a concentration of 5 μM, induced a significant increase of the apoptotic peak in MCF-7 (p<0.05) and HCC1806 (p<0.05) cell lines.

Figure 10.
In addition to the increase of the apoptotic peak, results also show a cell cycle blockage in the S phase regarding MCF-7 (p<0.01) cell line, also observed with the IC50 concentration in HCC1806 cells. No significant and biologically relevant differences were obtained after incubation of both colorectal tumor cell lines with 3c.
Metal complexes are described as being capable of inhibiting enzymes, interacting with intracellular biomolecules, enhancing lipophilicity, altering the cell membrane functions, and blocking the cell cycle and other functions. [8] The ability of copper complexes to interact with DNA is well documented. However, recent studies indicate other cell constituents as possible targets, namely topoisomerases and the proteasome multiprotein complex. [13] In Another advantage of 3c is its selectivity towards the colorectal cancer chemoresistant cell line (LS1034) when compared to the results obtained with the non-tumor cell line (CCD-841 CoN).

In Vitro Antimicrobial Activity
Over the years, it has been shown that copper has the ability to kill bacteria, yeast and viruses, with the assumption that copper ions are released, leading to a disruption of the bacterial cell membrane and causing cell death. The entry of these ions into the bacteria leads to the production of reactive oxygen species (ROS) and the degradation of plasmid and genomic DNA. [48] Complexes 3a-c as well as the corresponding free salen ligands 2a-c were screened for their antibacterial activity against S. aureus, E. faecalis, E. coli and P. aeruginosa and for their antifungal effect against C. albicans, C. glabrata, Asp. fumigatus and Alt. alternata. The tests were carried out at different concentrations, between 1 ng/mL to 100 µg/mL.
In the presence of free ligands 2a-c and copper complexes 3a-c no significant antifungal activity was observed against the fungal species tested.
Antibacterial activity was observed with 2a-c and 3a-c, as indicated by the minimal inhibitory concentration (MIC) and decimal reduction ratio (RD) values obtained,    Table 5. IC50 values obtained for the four tumor cell lines after incubation with Cu(II) complexes 3a-c and the corresponding r 2 (coefficient of determination of the fitted curves).