Metal Coordination and Biological Screening of a Schiff Base Derived from 8-Hydroxyquinoline and Benzothiazole

Designing new metallodrugs for anticancer therapy is a driving force in the scientific community. Aiming to contribute to this field, we hereby report the development of a Schiff base (H2L) derived from the condensation of 2-carbaldehyde-8-hydroxyquinoline with 2-hydrazinobenzothiazole and its complexation with transition metal ions. All compounds were characterised by analytical and spectroscopic techniques, which disclosed their structure: [Cu(HL)Cl], [Cu(HL)2], [Ni(HL)(acetate)], [Ni(HL)2], [Ru(HL)Cl(DMSO)], [VO(HL)2] and [Fe(HL)2Cl(H2O)]. Different binding modes were proposed, showing the ligand’s coordination versatility. The ligand proton dissociation constants were determined, and the tested compounds showed high lipophilicity and light sensitivity. The stability of all complexes in aqueous media and their ability to bind to albumin were screened. Based on an antiproliferative in vitro screening, [Ni(HL)(acetate)] and [Ru(HL)Cl(DMSO)] were selected for further studies aiming to investigate their mechanisms of action and therapeutic potential towards colon cancer. The complexes displayed IC50 < 21 μM towards murine (CT-26) and human (HCT-116) colon cancer cell lines. Importantly, both complexes exhibited superior antiproliferative properties compared to the clinically approved 5-fluorouracil. [Ni(HL)(acetate)] induced cell cycle arrest in S phase in CT-26 cells. For [Ru(HL)Cl(DMSO)] this effect was observed in both colon cancer cell lines. Additionally, both compounds significantly inhibited cell migration particularly in the human colon cancer cell line, HCT-116. Overall, the therapeutic potential of both metal complexes was demonstrated.


Introduction
Cancer is one of the most challenging diseases and related deaths are estimated to rise as life expectancy increases [1]. Although relevant progresses in therapy have been attained over the past decades, the incidence and mortality continue to increase. Particularly, colorectal arises as the fourth most prevalent form of cancer both in the EU and the USA [2]. In Portugal, this malignancy has the highest incidence and represents the second most lethal type of cancer [3]. As such, knowledge about its pathophysiology and aetiology is crucial to design effective treatments. Colorectal cancer develops when the healthy colorectal epithelium suffers changes (due to genetic and environmental factors), which can lead to the formation of malignant adenomatous polyps that proliferate in an uncontrolled manner, increasing in size and invading nearby tissues. In the worst cases, they can reach ide) tetrachlororuthenate(III), [(DMSO) 2 H][Ru(DMSO) 2 Cl 4 ] was synthetised according to a literature procedure [24]. All solvents were of p.a. grade and used as received. Phosphate buffered saline (PBS) was obtained from Sigma as readily soluble tablets and HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) was purchased from Sigma. The water used in all studies with biomolecules was doubly deionised in a MilliQ system. The spectroscopic data were visualised and treated using Spectragryph software v. 1.2.15 [25]. Elemental analysis for C, H, N and S were carried on a FISONS EA 1108 CHNS-O apparatus at Laboratório de Análises of Instituto Superior Técnico. A 500-MS Varian Ion Trap Mass Spectrometer was used to measure ESI-MS spectra of methanolic solutions of the compounds in both positive and negative ion modes. A Bruker Avance II + 400 (UltraShieldTM Magnet) spectrometer operating at 400. 13 MHz for proton and at 81 MHz for carbon was used to obtain the 1 H and 13 C NMR spectra. The chemical shifts are reported in ppm using tetramethylsilane as internal reference. A Perkin Elmer Lambda 35 spectrophotometer was used to measure the UV-vis electronic absorption spectra, while in the infrared region a JASCO FT/IR 4100 spectrophotometer was used. The first derivative X-band EPR spectra of the frozen solutions at ca. 100 K were recorded on a Bruker ESP 300E operated at ≈ 9.51 GHz with a frequency modulation of 100 KHz.

Synthesis of the Copper(II) Complexes
In a round bottom flask, 0.25 mmol of H 2 L was suspended in methanol (MeOH), deprotonated with 0.25 mmol of CH 3 ONa and the yellow mixture was stirred at room temperature (r. t.) for ca. 20 min. In a small vial, 1 eq. (1:1) or 0.5 eq. (2:1) of CuCl 2 ·2H 2 O was dissolved in MeOH and added to the previous mixture, which immediately became red. The reaction mixture was left under stirring at r. t. overnight. The resulting solids were collected by filtration, washed with ice-cold MeOH and dried under vacuum in a desiccator over silica-gel.

Synthesis of the Nickel(II) Complexes
In a two-entry reaction vessel 0.25 mmol of H 2 L was suspended in ethanol (EtOH) and deprotonated with KOH in MeOH. In a vial, 1 eq. (1:1) or 0.5 eq. (2:1) of Ni(CH 3 COO) 2 ·3H 2 O was dissolved in EtOH and added dropwise to the ligand solution and allowed to reflux for 5 h. The reaction mixture was left overnight in the freezer. The resulting solids were The procedure was similar to the one used for the Ni-complexes, but 1 eq. of triethylamine (Et 3 N) was used to deprotonate the ligand and the metal precursor used was [(DMSO) 2 H][trans-Ru(DMSO) 2 Cl 4 ] [24]. [

Synthesis of the Oxidovanadium(IV) Complex
The procedure was similar to the one used for the Ni(II) complexes, although VO(acac) 2 was used as the metal precursor. The solutions were prepared with degassed solvents and the reaction was performed under nitrogen atmosphere.

UV-Visible Spectrophotometry
The pH-potentiometric titrations for the calibration of the electrode system were performed in 30% (v/v) and 60% (v/v) DMSO/H 2 O solvent mixtures using 0.10 M KCl background electrolyte at 25.0 ± 0.1 • C as described in our previous works [26,27].
A Thermo Evolution 220 spectrophotometer was used to record the UV-visible (UV-vis) spectra in the 260 to 750 nm window. The path length was varied between 2 and 50 mm. As a result of the limited solubility of the compounds in water, the spectrophotometric titrations were performed to determine the proton dissociation constants (pK a ) of the ligand and its complex at 5 µM concentration using 5 cm path length. Due to the light sensitivity of the studied compounds, the samples were covered and were kept in the dark during the titrations. The starting sample volume was 50 mL. These titrations were performed in the pH range 2.0-12.5 (30% (v/v) DMSO/H 2 O) and 2.0-14.0 (60% (v/v) DMSO/H 2 O). Samples were deoxygenated by bubbling purified argon through them for approximately 10 min prior to the measurements and argon was also passed over the solutions during the spectrophotometric titrations. Calculation of the proton dissociation constants was performed with the computer program PSEQUAD [28].
For the light sensitivity studies, DMSO solutions were prepared. The starting sample was separated equally for more portions, one was irradiated with diffuse light, while other portions were stored in the dark and measured only once.

Lipophilicity Measurements
The distribution coefficient values (D 7.4 ) of the compounds were determined by the traditional shake-flask method in n-octanol/buffered aqueous solution at pH 7.40 (10 mM HEPES buffer, in the absence or in the presence of 0.10 M KCl) at 25.0 ± 0.2 • C. The samples and stock solutions were protected from light until the spectrophotometric concentration determination. The ligand and complex were dissolved at 150 and 100 µM concentrations in n-octanol, respectively. n-Octanol was presaturated with the HEPES solution with the given KCl concentration. The aqueous solution and n-octanol ratio was 9:1 (instead of the usual 1:1 ratio due to the high lipophilicity of the compounds). The samples were mixed with vertical rotation (~20 rpm) for 4 h, then were centrifuged at 2415 g for 8 min. After separation, UV-vis spectra of the compounds in the n-octanol phase were recorded and compared to the spectra of the n-octanol stock solution. D 7.4 values were calculated as in the following Equation (1): where A oct,0 and A oct,1 are the absorbance values of the n-octanol phase before and after separation, respectively, V water phase and V octanol phase are the volumes of the two phases.
2.5. 1 H NMR Spectroscopy 1 H NMR spectroscopic studies for the ligands and the ruthenium complex were carried out on a Bruker Avance III HD Ascend 500 Plus instrument. DMSO_d 6 and 60% (v/v) DMSO_d 6 /HEPES mixture were used as solvent. DMSO_d 6 was used as reference (δ = 2.50 ppm) and WATERGATE method was applied to suppress the solvent resonance in the aqueous samples. The samples contained the ligand or the complex in 200 µM-2 mM concentration.

Stability Assays under Aqueous Conditions
The complexes were monitored for their stability in PBS, pH 7.4 at 25 • C in the absence and presence of equimolar amounts of bovine serum albumin (BSA). Stock solutions of each complex were prepared in DMSO and then diluted in PBS in order to obtain final concentrations of 20 to 30 µM and ensuring that the organic solvent was less than 1% (v/v). For the assays with BSA, stock solutions of the protein were prepared in PBS using 20 mg of lyophilised powder in 5 mL solvent, which were allowed to hydrate for 24 h in the refrigerator. Samples were monitored by UV-vis absorption spectroscopy for 6 h and a final measurement was also taken after 24 h.

Fluorescence Assays
Based on 3D scans, it was possible to conclude that the ligand, the Ru-complex and their HSA adducts are not fluorescent in 5% (v/v) DMSO/HEPES mixture. Prior to all spectrofluorometric measurements, samples were protected from light.
The concentration of BSA in the experiments was ca. 1.5 µM (in aqueous PBS). The fluorescence emission spectra were recorded between 310 and 575 nm, with λ ex = 295 nm. Suc-cessive additions of the complexes were done directly to the cuvette (optical path 1 cm) and the fluorescence emission spectra were recorded for each one under the same conditions. Blank assays for each sample, consisting in solutions with the same concentration of complex and no BSA, were recorded and subtracted to the corresponding emission spectra containing the fluorophore. The UV-vis spectra of the sample solutions were used to correct the emission spectra, minimizing inner filter and reabsorption phenomena interference [29][30][31].

Circular Dichroism Spectroscopy with Bovine Serum Albumin
The concentration of BSA in the experiments was ca. 20-25 µM (in aqueous PBS). The circular dichroism (CD) spectra are an average of three scan accumulations, recorded between 300 and 500 nm, with bandwidth of 1 nm and response of 2 s, using a scanning speed of 100 nm/min. Successive additions of the complexes stock solutions in DMSO were done directly to the cuvette (1 cm) and the CD spectra were recorded under the same conditions after 10 min equilibration. Induced CD spectra were obtained as the CD spectrum of compound-BSA mixture minus the CD spectrum of BSA alone and are expressed as ellipticity in millidegrees.

In Vitro Antiproliferative Activity and Viability
The MTT [32] and Guava ViaCount assays [33,34] were used to assess cell viability in the absence (negative control) or presence of increasing concentrations of [Ni(HL)(acetate)], [Ru(HL)Cl(DMSO)] and 5-FU (positive control). Briefly, cells were plated at a concentration of 5 × 10 4 cells/mL in 96-well plates (200 µL/well) and cultured for 24 h in the culture conditions described above. Afterwards, supernatant was removed, and cells were incubated with [Ni(HL)(acetate)] and [Ru(HL)Cl(DMSO)] at concentrations ranging from 0.5 to 40 µM or from 0.5 to 100 µM for the 5-FU (positive control). Negative controls were cells maintained in complete culture medium. After 48 h, culture medium was discarded, cells were washed twice with PBS and 50 µL of MTT at 0.5 mg/mL in incomplete medium was added to all wells. After a 2 to 3 h incubation period at 37 • C and 5% CO 2 environment, 100 µL DMSO was added to each well to solubilise the formazan crystals, and absorbance was measured at 570 nm using a microplate reader Model 680 (Bio-Rad, CA, USA). Between three to four independent experiments were carried out, with six replicates per condition. GraphPad Prism ® 8 (GraphPad Software, San Diego, CA, USA) was used to analyse cell proliferation and values were plotted and fit to a standard log dose-response curve. The IC 50 values were calculated using Equation (2) [35]: where, Y 0 /2 is the half-cell density of the negative control; Y 1 is the cell density above Y 0 /2; X 1 is the concentration corresponding to Y 1 ; Y 2 is the cell density below Y 0 /2; X 2 is the concentration corresponding to Y 2 . The IC 50 was determined by linear interpolation between X 1 and X 2. For Guava ViaCount assay, cells were plated at a concentration of 1 × 10 5 cells/mL in 24-well plates (1 mL/well) and cultured for 24 h in the culture conditions described above [33]. CT-26 cells were incubated with [Ni(HL)(acetate)] or [Ru(HL)Cl(DMSO)] at concentrations of 5 and 10 µM or 10 and 20 µM, respectively. For the HCT-116 cell line, the tested concentrations were 5 and 7.5 µM, and 7.5 and 12.5 µM, respectively. The positive control group, 5-FU, was tested at 0.5 and 2 µM, and at 20 and 40 µM for CT-26 and HCT-116, respectively. Negative control group was constituted by cells in the presence of complete culture medium. After a 48 h incubation period, cell culture supernatants were collected, and adherent cells were detached using TrypLE™ Express (Invitrogen). Detached cells were then added to supernatants and the wells were washed using PBS with 2% FBS to recover any remaining cells. The recovered suspension was added to the previously collected cells and supernatants and centrifuged for 5 min at 650× g. The supernatants were discarded, and cells were suspended in PBS with 2% FBS. Then, on a 96-well plate, 20 µL of cell suspension was incubated with 180 µL of Guava ViaCount ® reagent (Merck Millipore, Darmstadt, Germany) for 5 min at r. t. Using a Guava EasyCyte 5HT Flow cytometer (Guava Technologies, Inc., Hayward, CA, USA), data were collected and analysed with the ViaCount software module with the acquisition of 10,000 events per sample. GraphPad Prism ® 8 was used to analyse cell proliferation (GraphPad Software, San Diego, CA, USA). Two independent experiments were carried out, with three replicates per condition.

Cell Cycle Analysis
To understand the potential effects of [Ni(HL)(acetate)] and [Ru(HL)Cl(DMSO)] on cell cycle progression, CT-26 and HCT-116 cells were used [36]. Briefly, cells were plated at a density of 5 × 10 4 cells/mL in 6-well plates (2 mL) and allowed to grow for 24 h at 37 Prior to the analysis, samples were kept at 4 • C for at least 16 h. On the day of data collection, cells were centrifuged at 850× g and suspended in PBS with 25 µg/mL PI (Fluka, Sigma-Aldrich) and 50 µg/mL RNase A (Sigma-Aldrich), then incubated at 37 • C for 30 min. The Guava easyCyte™ Flow Cytometer (Merck Millipore, Darmstadt, Germany) was used to collect the data corresponding to 10,000 events per sample. The ModFit LT™ 5.0 software was used to analyse the data and the percentages of cells in each cell cycle phase was determined (Verity Software House, Topsham, ME, USA).

Migration Assay
To assess cellular migration in the presence or absence of tested compounds, HCT-116 and CT-26 cells were plated at 1.5 × 10 5 and 3 × 10 5 cells/mL in a 12-well plate (1 mL/well) and were allowed to adhere and grow until reaching around 80-90% confluence [36,37]. A sterile 10 µL pipette tip was used to wound the cell monolayer, followed by a washing step with PBS to remove debris. Then, cells were incubated with the tested compounds suspended in RPMI-1640 (CT-26) or DMEM media (HCT-116) supplemented with 2% FBS and 100 IU/mL of penicillin and 100 µg/mL streptomycin (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), and hereafter designated as RPMI 2% and DMEM 2%, respectively. Under the above mentioned culture conditions, CT-26 cells were incubated with . ImageJ (https://imagej.net (accessed on 15 September 2022)) was used to analyse the images, and wound closure was normalised to the original wound area at time 0 h using Equation (2), where W 0 represents the wound width at time 0 h and W t represents the wound width at the studied timepoints. Three independent experiments were performed, with three replicates for each tested condition.

Statistical Analysis
Results are expressed as mean ± standard deviation (SD). Statistical analysis was performed with two-way ANOVA followed by Dunnett post-hoc test using GraphPad Prism ® 8 for Windows (GraphPad Software, San Diego, CA, USA). p < 0.05 was considered statistically significant.

Computational Methodologies
All theoretical calculations were of the density functional theory/ time-dependent density functional theory (DFT/TDDFT) type, carried out using GAMESS-US version R3 [38]. A range corrected LC-BPBE (µ = 0.20) functional as implemented in GAMESS-US [38] was used in both ground-and excited-state calculations. The solvent was included using the polarisable continuum model with the solvation model density to add corrections for cavitation, dispersion, and solvent structure. In TDDFT calculation of FC (Franck-Condon) excitations the dielectric constant of the solvent was split into a "bulk" component and a fast component, which is essentially the square of the refractive index. In "adiabatic" conditions only the static dielectric constant is used. A SBKJC effective core potential and corresponding-31G basis set was used in either DFT or TDDFT calculations.

The Ligand Precursor
The condensation reaction of 2-carbaldehyde-8-hydroxyquinoline with 2-hydrazinoben zothiazole provided the ligand precursor, H 2 L, as shown in Figure 1. Since its characterisation has been already published by us [23], details are included in Supplementary Material and here we report only the new solution studies.

The Ligand Precursor
The condensation reaction of 2-carbaldehyde-8-hydroxyquinoline with 2hydrazinobenzothiazole provided the ligand precursor, H2L, as shown in Figure 1. Since its characterisation has been already published by us [23], details are included in Supplementary Material (SM, S1) and here we report only the new solution studies. Spectroscopic and elemental studies confirmed the structure and purity of the prepared Schiff base. In the electronic absorption spectrum measured in DMSO ( Figure  S1a), it is possible to identify three maxima at λ = 265 nm (ε = 2.60 × 10 4 M −1 cm −1 ), 368 nm (ε = 2.87 × 10 4 M −1 cm −1 ) and 476 nm (ε = 9.78 × 10 3 M −1 cm −1 ). According to the definition presented by Kasha and Rawls for the electronic transitions [39], the two bands in the UV correspond to π→π* transitions within the quinoline and benzothiazole chromophores. The band at 476 nm is more difficult to assign, and a study carried in the absence of light, and described below, indicates that H2L is light sensitive.
Nevertheless, UV-vis spectroscopic studies revealed an interesting behaviour for H2L. The lower energy band is only present in solvents of higher donor number (DMSO and DMF). Its extinction coefficient decreases as the concentration increases, while the band at 372 nm increases, this being more pronounced in DMSO. Concentration Spectroscopic and elemental studies confirmed the structure and purity of the prepared Schiff base. In the electronic absorption spectrum measured in DMSO ( Figure S1a), it is possible to identify three maxima at λ = 265 nm (ε = 2.60 × 10 4 M −1 cm −1 ), 368 nm (ε = 2.87 × 10 4 M −1 cm −1 ) and 476 nm (ε = 9.78 × 10 3 M −1 cm −1 ). According to the definition presented by Kasha and Rawls for the electronic transitions [39], the two bands in the UV correspond to π→π* transitions within the quinoline and benzothiazole chromophores. The band at 476 nm is more difficult to assign, and a study carried in the absence of light, and described below, indicates that H 2 L is light sensitive.
Nevertheless, UV-vis spectroscopic studies revealed an interesting behaviour for H 2 L. The lower energy band is only present in solvents of higher donor number (DMSO and DMF). Its extinction coefficient decreases as the concentration increases, while the band at 372 nm increases, this being more pronounced in DMSO. Concentration dependence effects have been described for donor-acceptor intramolecular interactions [40,41]. Considering the Gutmann basicity scale [42], only DMSO and DMF possess donor numbers (DN) which seem to be sufficiently high to deprotonate the hydrazone moiety of HL. These experimental results are well described by DFT calculations, which showed that only the deprotonated species possesses the band in the visible range (notice that the calculations indicate that this deprotonation refers to the NH of the hydrazone moiety and not the hydroxyl group in the 8-hydroxyquinoline).
DFT calculations were used to access the structural conformations of H 2 L in both gas phase and DMSO. Figure S2 shows the tested conformers, isomer D, presented in Figure 1, being the most stable by about 9 kJ/mol (DMSO), which corresponds to being almost (97.5% Boltzmann distribution) the only conformer present in solution. The DMSO TDDFT simulated spectrum, Figure 2A, is identical to the experimental one with absorption bands at 250 and 360 nm plus a shoulder at 340 nm and a weak band at 305 nm. Probing the S1 potential energy surface (PES) showed an internal excited state proton transfer (ESPT) of the hydroxyl proton to the neighbour nitrogen creating a zwitterionic structure, which has an emission at 580 nm close to the experimentally observed. However, the absorption band at 476 nm is absent in simulations, which prompted us to find other possible species in solution. The deprotonation of H 2 L leads to an anionic base with an extra absorption at 480 nm, Figure 2B, also observed in the experimental spectrum. The combined simulated data suggest that in high donor number, high dielectric constant solvents, the separation of charges of the zwitterionic form are stabilised leading to continuous irradiation to the formation of this species, without a band at 480 nm. Putting the sample in darkness re-establishes the pH equilibrium and the band of anionic base resulting from proton transfer to the solvent. Simulations also predict a 15 nm red shift of the 360 nm band upon formation of the zwitterionic form, in good agreement with the shift observed under continuous irradiation. Since emission is from the zwitterionic form, H 2 L only has emission in solvents, such as DMSO, that can stabilise the charges in this species.
Fluorescence characterisation showed that H 2 L has emissive properties in DMSO, showing a band centred at λ = 600 nm when excited at λ = 475 nm ( Figure S1b). This property vanishes in 60% (v/v) DMSO/H 2 O solvent mixtures.

Metal Coordination
Following our previous studies with Zn(II) [23], we decided to evaluate the ability of the ligand to coordinate other transition metal ions with different properties, such as oxidation state and preferential coordination geometry. Cu(II), Ni(II), Ru(III), V(IV)O and Fe(III) ions were selected, and synthetic procedures were performed to obtain complexes with metal-to-ligand (M:L) stoichiometries of 1:1 and 1:2. Nevertheless, for some metal ions (V, Ru and Fe), only one type of stoichiometry was obtained based on the characterisation data. Interestingly, from a Ru(III) precursor we obtained a Ru(II) complex. Overall, the elemental analysis and mass spectrometry performed for the complexes are in good agreement with the following structural formulae: [ the zwitterionic form, H2L only has emission in solvents, such as DMSO, that can stabilise the charges in this species.

Metal Coordination
Following our previous studies with Zn(II) [23], we decided to evaluate the ability of the ligand to coordinate other transition metal ions with different properties, such as oxidation state and preferential coordination geometry. Cu(II), Ni(II), Ru(III), V(IV)O and Fe(III) ions were selected, and synthetic procedures were performed to obtain complexes with metal-to-ligand (M:L) stoichiometries of 1:1 and 1:2. Nevertheless, for some metal ions (V, Ru and Fe), only one type of stoichiometry was obtained based on the characterisation data. Interestingly, from a Ru(III) precursor we obtained a Ru(II) complex. Overall, the elemental analysis and mass spectrometry performed for the complexes are in good agreement with the following structural formulae: [Cu II Figure S3). The FTIR of [VO(HL)2] (in KBr pellet, Figure S4) presents a medium sharp band  Figure S3). The FTIR of [VO(HL) 2 ] (in KBr pellet, Figure S4) presents a medium sharp band at 983 cm −1 corresponding to the characteristic V = O stretching vibration, confirming the presence of an oxidovanadium centre [43]. The V(IV)O and Fe(III) complexes' FTIR spectra ( Figure S4) show broad bands at 3400 cm −1 , in a clear indication of the deprotonation of the phenol group upon coordination, while little changes are observed in the region 1620-1440 cm −1 , dismissing the coordination of the imine moiety in these complexes.
A broad band centred at 3400 cm −1 is also present in [Cu II (HL)Cl] and [Ni II (HL)(acetate)] spectra, supporting the coordination through the O-phenolate of the hydroxyquinoline moiety (see Figures S5 and S6). The absence of NMR spectra for the d 8 Ni(II) 1:1 complex reveals an octahedral geometry around this metal centre, with an electronic distribution that yields two unpaired electrons. In contrast, the FTIR spectra of [Cu(HL) 2 ], [Ni(HL) 2 ] and [Ru(HL)Cl(DMSO)] (see also Figure S7) display sharp bands at 3411 cm −1 , indicating that in these complexes the phenol group is protonated and not involved in the metal coordination. Substantial changes in the 1620-1144 cm −1 region point to the coordination of imine and benzothiazole nitrogen atoms in the first two, while the disappearance of the signal at 520 cm −1 in the Ru complex points to 8-hydroxyquinolinate coordination. This coordination sphere is further corroborated by the 1 H NMR spectrum of the [Ni(HL) 2 ] complex in DMSO_d 6 ( Figure S8), in which the presence of the OH signal can be observed at the same chemical shift as found for the ligand precursor, and as in [Ru(HL)Cl(DMSO)] ( Figure S9). The FTIR spectrum of [Ru(HL)Cl(DMSO)] ( Figure S7) does not show the NH band, indicating the deprotonation of the ligand at this moiety. A strong band at 1019 cm −1 , corresponding to the rocking (ρ) of the CH 3 groups, and another at 1108 cm −1 [ν(S = O)] confirm the presence of a dimethyl sulfoxide molecule in this complex [24]. The deprotonated form of the ligand in [Ru(HL)Cl(DMSO)] is further confirmed in the 1 H NMR spectrum ( Figure S9) with the absence of the NH signal, while a broad peak at 10 ppm shows the presence of OH. The observation of a diamagnetic character for the Ru-complex indicates that reduction took place during the synthetic procedure and from a Ru(III) precursor, a Ru(II) species was obtained. This is not uncommon since Ru has a low reduction potential and triethylamine, the base used to undertake ligand deprotonation, might have been oxidised [44]. This has been previously observed with other Ru(III) complexes [45].
The Cu(II) and V(IV)O complexes, being paramagnetic, were also analysed by electron paramagnetic resonance (EPR, see Figures S10 and S11) spectroscopy and the spin Hamiltonian parameters were obtained by spectral simulation [46], Table 1. In both Cu(II) complexes, the frozen solution X-band EPR spectra in DMF are axial, presenting g z > g x,y > 2.0, which suggests the presence of a d x Table 1. Spin Hamiltonian parameters determined by simulation of spectra measured at ca. 100 K [46]. The copper complexes were dissolved in DMF (≈3 mM), while for vanadium the solvent was methanol (≈1 mM).  [48]. This suggests that in DMF solution the chloride ion is not coordinated to Cu in the 1:1 complex and acts as a counter ion (binding set N 3 O). For [Cu(HL) 2 ], both N 2 O 2 and N 4 are possible donor sets. Bearing in mind the FTIR data, N 4 is proposed. Moreover, the values obtained for the g z /|A z | ratio are higher than 135 cm, indicating the presence of tetrahedral distortions from square planar-based coordination geometries [49], due to steric hindrance, imposed by the ligand planarity.  [51], which involves the coordination of two 2-aldehyde-8hydroxyquinoline molecules, the precursor used for the synthesis of H 2 L [52].

Complex
The UV-vis absorption spectra of the complexes in DMSO (see Figures S12-S14) show the characteristic intraligand π→π* transitions, in roughly the same wavelengths, with new bands appearing at lower energy, corresponding to charge transfer mechanisms between the ligand and the metal centres-see Table 2. Due to low solubility in common solvents, no d-d bands could be observed and assigned. All complexes show low solubility and stability in PBS buffered solutions, and the Fe(III) complex is almost insoluble. Spectral changes suggest precipitation, except for the V(IV)O complex for which oxidation of the metal centre likely takes place. To corroborate this, a small amount of [VO(HL) 2 ] was mixed with 2 equivalents of KOH and left stirring under air in methanol at r. t. over two days. An orange solid was collected and analysed by 51 V NMR spectroscopy showing the presence of a resonance at δ= −527 ppm (standard VOCl 3 at δ = 0 ppm), in agreement with the formation of a [V V O 2 L] type complex. Its chemical shift is in the same range as found for similar complexes [53]. In general, the spectral intensity in buffered aqueous solutions strongly decreases within 24 h ( Figure 3A,B and Figure S15); however, the presence of BSA clearly avoids this problem and allows the recording of almost unchanged UV-vis spectra for at least 24 h- Figure 3C,D and Figure S15. This is important since mammal cells incubation media contains ca. 40 µM of BSA added in FBS and will help to protect the complexes from precipitation. Overall, the analytic and spectroscopic data suggest that the ligand adopts different coordination modes with different metal ions and stoichiometries. In the 1:1 stoichiometry, the ligand is probably tridentate and acts as monoanionic through the hydroxyl group (Cu II and Ni II ) or through the hydrazinic moiety (Ru II ). For the 1:2 stoichiometry, the ligand is bidentate and monoanionic, but coordination involves either the hydroxyquinolinate moieties (V IV O and Fe III ) or the imine and benzothiazolate nitrogen atoms (Cu II and Ni II ), while keeping the phenol protonated. Considering all the Overall, the analytic and spectroscopic data suggest that the ligand adopts different coordination modes with different metal ions and stoichiometries. In the 1:1 stoichiometry, the ligand is probably tridentate and acts as monoanionic through the hydroxyl group (Cu II and Ni II ) or through the hydrazinic moiety (Ru II ). For the 1:2 stoichiometry, the ligand is bidentate and monoanionic, but coordination involves either the hydroxyquinolinate moieties (V IV O and Fe III ) or the imine and benzothiazolate nitrogen atoms (Cu II and Ni II ), while keeping the phenol protonated. Considering all the experimental evidence, structural formulae for the complexes are proposed in Scheme 1.

Light Sensitivity of H2L and Its Ru-Complex
The light sensitivity of H2L and its Ru-complex (as model metal compound) were tested by UV-vis spectrophotometry to further understand their behaviour. Samples were dissolved in DMSO and stored in the dark, while one portion was irradiated continuously by diffuse light. Both the ligand and the complex showed strong light sensitivity, which resulted in substantial changes in the UV-vis spectra. A relatively fast change is observed Scheme 1. Schematic representation of 1:1 copper(II), nickel(II) and ruthenium(II) and 2:1 copper, nickel, vanadium and iron complexes.

Light Sensitivity of H 2 L and Its Ru-Complex
The light sensitivity of H 2 L and its Ru-complex (as model metal compound) were tested by UV-vis spectrophotometry to further understand their behaviour. Samples were dissolved in DMSO and stored in the dark, while one portion was irradiated continuously by diffuse light. Both the ligand and the complex showed strong light sensitivity, which resulted in substantial changes in the UV-vis spectra. A relatively fast change is observed for the ligand in the first 1.5 h, and after that the spectrum has only minor changes ( Figure 4A). In dark, only a small change is visible, accompanied by the very slow development of a new band with λ max = 474 nm ( Figure 4B). Figure S16 depicts the spectral changes visible in Figure 4B. The inset shows the change of the absorbance values at λ max = 474 nm as a function of time. The changes follow a saturation curve, however, this reaction did not reach the final stage during the monitored period (4 days). In the ligand, this may be due to a photoisomerisation process of the imine bond (Z/E isomers), which is quite common in hydrazones [54]), but not in the Ru-complex, unless what is being observed in the complex solution is due to unbound ligand molecules. The Ru-complex showed identical changes during irradiation ( Figure S17), while the samples in dark showed no changes during the 4 days of monitoring.
not reach the final stage during the monitored period (4 days). In the ligand, this may be due to a photoisomerisation process of the imine bond (Z/E isomers), which is quite common in hydrazones [54]), but not in the Ru-complex, unless what is being observed in the complex solution is due to unbound ligand molecules. The Ru-complex showed identical changes during irradiation ( Figure S17), while the samples in dark showed no changes during the 4 days of monitoring. 1 H NMR spectra were also recorded for samples protected from light and irradiated. Figure S18 shows that in the samples protected from light a slight change is visible, namely a singlet peak that goes from 8.34 ppm to 5.77 ppm. In the chemical shift range 3.2−4.9 ppm, a new peak set is developed (not shown) which can be due to complexation with metal ion traces. A parallel sample was irradiated with diffuse light, and after 2.5 days it showed a clear difference from the light protected sample, since a new peak set is observed, which is not similar to the spectrum of 2-hydrazinobenzothiazole, a possible decomposition product.
DFT calculations in the Ru complex predict a tetra-bonded ligand with the hydroxyl hydrogen making an intramolecular H-bond to the chloride atom, Figure 5A. The absorption spectrum depicts a low energy band at 760 nm, which is observed in the measured spectrum at 770 nm ( Figure S14). This is the HOMO-LUMO transition of MLCT character and is antibonding towards the hydroxyquinoline moiety, Figure 5B, between metal and ligand leading to the HLun-coordination upon irradiation explaining the behaviour described above. 1 H NMR spectra were also recorded for samples protected from light and irradiated. Figure S18 shows that in the samples protected from light a slight change is visible, namely a singlet peak that goes from 8.34 ppm to 5.77 ppm. In the chemical shift range 3.2−4.9 ppm, a new peak set is developed (not shown) which can be due to complexation with metal ion traces. A parallel sample was irradiated with diffuse light, and after 2.5 days it showed a clear difference from the light protected sample, since a new peak set is observed, which is not similar to the spectrum of 2-hydrazinobenzothiazole, a possible decomposition product.
DFT calculations in the Ru complex predict a tetra-bonded ligand with the hydroxyl hydrogen making an intramolecular H-bond to the chloride atom, Figure 5A. The absorption spectrum depicts a low energy band at 760 nm, which is observed in the measured spectrum at 770 nm ( Figure S14). This is the HOMO-LUMO transition of MLCT character and is antibonding towards the hydroxyquinoline moiety, Figure 5B, between metal and ligand leading to the HLun-coordination upon irradiation explaining the behaviour described above.

Lipophilicity of the Title Ligand and Its Ru Complex
The distribution coefficients (D7.4) for H2L and the Ru-complex (as model complex) were determined using HEPES buffer (pH = 7.4) for the aqueous phase. Distribution was measured in the absence and presence of 100 mM KCl. The latter condition mimics the chloride ion concentration of the blood plasma. Due to the predicted high lipophilicity of these compounds at neutral pH, an aqueous phase:n-octanol phase = 9:1 volume ratio was used. Samples were dissolved in n-octanol, and after partitioning and separation the UVvis spectra for both phases were recorded. Both the ligand and the Ru-complex showed high lipophilicity and the presence of chloride ion had only a slight effect on the coefficients, as shown in Table 3. While the UV-vis spectrum of ligand remained unchanged after partitioning, the complex suffered changes in the wavelength range λ > 400 nm ( Figure S19), which indicates changes in the coordination sphere. This may be the result of the possible coordination of water or additional chloride ions.

Determination of Proton Dissociation Constants
The pKa values of bioactive compounds are crucial factors influencing the protonation state under physiological conditions and the solubility at a given pH, having therefore a great impact on the pharmacokinetic properties. H2L has four dissociable protons, namely the protonated benzothiazole nitrogen (NH + btia), the quinolinium

Lipophilicity of the Title Ligand and Its Ru Complex
The distribution coefficients (D 7.4 ) for H 2 L and the Ru-complex (as model complex) were determined using HEPES buffer (pH = 7.4) for the aqueous phase. Distribution was measured in the absence and presence of 100 mM KCl. The latter condition mimics the chloride ion concentration of the blood plasma. Due to the predicted high lipophilicity of these compounds at neutral pH, an aqueous phase:n-octanol phase = 9:1 volume ratio was used. Samples were dissolved in n-octanol, and after partitioning and separation the UV-vis spectra for both phases were recorded. Both the ligand and the Ru-complex showed high lipophilicity and the presence of chloride ion had only a slight effect on the coefficients, as shown in Table 3. While the UV-vis spectrum of ligand remained unchanged after partitioning, the complex suffered changes in the wavelength range λ > 400 nm ( Figure S19), which indicates changes in the coordination sphere. This may be the result of the possible coordination of water or additional chloride ions.

Determination of Proton Dissociation Constants
The pK a values of bioactive compounds are crucial factors influencing the protonation state under physiological conditions and the solubility at a given pH, having therefore a great impact on the pharmacokinetic properties. H 2 L has four dissociable protons, namely the protonated benzothiazole nitrogen (NH + btia ), the quinolinium nitrogen (NH + quin ), the hydroxyl group (OH), and the hydrazinic nitrogen (NH + hydr ) ( Figure 6). The pK a values predicted by the Marvin software [55] suggest two overlapping deprotonation processes in the acidic pH range and two overlapping processes in the basic pH range (Table 4). nitrogen (NH + quin), the hydroxyl group (OH), and the hydrazinic nitrogen (NH + hydr) ( Figure 6). The pKa values predicted by the Marvin software [55] suggest two overlapping deprotonation processes in the acidic pH range and two overlapping processes in the basic pH range (Table 4).  Titrations were performed first in 30% (v/v) DMSO/H2O mixture due to the limited water solubility of the ligand. The sample was protected from light and, for a given portion of the sample, the UV-vis spectrum was recorded only once. Despite the high DMSO content and low concentration (5 μM), the ligand precipitated in a wide pH range (3.8−11.0, Figure S20), indicating that a neutral species is dominating under these conditions. Following this experiment, the titration was repeated in 60% (v/v) DMSO/H2O mixture (Figure 7), and no precipitation was found in the whole pH range (1.73−13.67) under this condition. However, only three deprotonation steps were observed, most probably because one of the deprotonation steps was already completed at the starting pH. The determined pKa values are listed in Table 4.  Titrations were performed first in 30% (v/v) DMSO/H 2 O mixture due to the limited water solubility of the ligand. The sample was protected from light and, for a given portion of the sample, the UV-vis spectrum was recorded only once. Despite the high DMSO content and low concentration (5 µM), the ligand precipitated in a wide pH range (3.8−11.0, Figure S20), indicating that a neutral species is dominating under these conditions. Following this experiment, the titration was repeated in 60% (v/v) DMSO/H 2 O mixture (Figure 7), and no precipitation was found in the whole pH range (1.73−13.67) under this condition. However, only three deprotonation steps were observed, most probably because one of the deprotonation steps was already completed at the starting pH. The determined pK a values are listed in Table 4.
From these results, it is still unclear to which functional group belongs the first deprotonation process. To elucidate the pK a assignment, 2-hydrazinobenzothiazole (considered as a fragment of the title ligand) was titrated under similar conditions to monitor the spectral changes during the deprotonation of the benzothiazole moiety. All deprotonation steps were accompanied by a bathochromic shift for this fragment, while the first deprotonation resulted in a hypochromic shift in the UV-vis spectrum of the title ligand. Based on this observation, the benzothiazole moiety was already deprotonated at the starting pH and the first process is considered the deprotonation of NH + quin (pK a = 1.99 ± 0.03). The spectrophotometric titration of the ruthenium complex was performed in the same way with samples carefully protected from light. Surprisingly, all spectral changes are identical with those of the ligand. Most probably the complex dissociates in the 60% (v/v) DMSO/H 2 O mixture, under the conditions used. To confirm this, 1 H NMR spectra were recorded for the ligand and the complex in DMSO_d 6 and in 60% (v/v) DMSO_d 6 /aqueous buffer (pH = 7.4, 10 mM HEPES) mixture. While the complex has a different spectrum from the ligand in DMSO_d 6 ( Figure S21a), in buffered aqueous solution as co-solvent very similar spectra are recorded ( Figure S21b). Most of the complex is dissociated in the solvent mixture, which can be responsible for the similarity in the UV-vis titration. From these results, it is still unclear to which functional group belongs the first deprotonation process. To elucidate the pKa assignment, 2-hydrazinobenzothiazole (considered as a fragment of the title ligand) was titrated under similar conditions to monitor the spectral changes during the deprotonation of the benzothiazole moiety. All deprotonation steps were accompanied by a bathochromic shift for this fragment, while the first deprotonation resulted in a hypochromic shift in the UV-vis spectrum of the title ligand. Based on this observation, the benzothiazole moiety was already deprotonated at the starting pH and the first process is considered the deprotonation of NH + quin (pKa = 1.99 ± 0.03).
The spectrophotometric titration of the ruthenium complex was performed in the same way with samples carefully protected from light. Surprisingly, all spectral changes are identical with those of the ligand. Most probably the complex dissociates in the 60% (v/v) DMSO/H2O mixture, under the conditions used. To confirm this, 1 H NMR spectra were recorded for the ligand and the complex in DMSO_d6 and in 60% (v/v) DMSO_d6/aqueous buffer (pH = 7.4, 10 mM HEPES) mixture. While the complex has a different spectrum from the ligand in DMSO_d6 (Figure S21a), in buffered aqueous solution as co-solvent very similar spectra are recorded ( Figure S21b). Most of the complex is dissociated in the solvent mixture, which can be responsible for the similarity in the UVvis titration.

Albumin Binding
Spectrofluorimetric studies show that mixtures of all compounds and BSA display significant fluorescence quenching in a concentration-dependent way. However, due to the light sensitivity of the samples, these experiments were not quantitatively analysed. The intense light of the fluorimeter can cause decomposition of the compounds, even in experiments performed in the dark. Thus, we cannot exclude that the quenching is due to binding of side-products to albumin, instead of the original ligand or complex. Moreover, circular dichroism (CD) titrations showed the development of high intensity induced CD bands centred at ca. 360 nm, particularly for H 2 L, [Ni(HL)(acetate)], [Ru(HL)Cl(DMSO)] and [VO(HL) 2 ]. Since under the conditions used [c compoud between 5 and 70 µM)] no metal centred bands are observed, we cannot exclude that these are due to the ligand or its side-products binding the protein. Nonetheless, spectroscopic titrations seem to suggest that the compounds can bind albumin, as already indicated by the increased stability of the compounds in the presence of this protein.

MTT Assay
Several metal-based complexes were synthesised using the Schiff base derived from 8-hydroxyquinoline and benzothiazole, aiming to obtain new metallodrugs for anticancer therapy. Thus, to evaluate the antiproliferative properties of the newly synthetised metal complexes, a preliminary screening was performed using human (HCT-116 and A375) and murine (CT-26 and B16F10) colon cancer and melanoma cell lines, respectively. The human keratinocyte cell line, HaCaT, was also included in this in vitro study aiming to evaluate the specificity of the new compounds against tumour cells. Complexes were evaluated at concentrations of 20 and 40 µM and the obtained results are depicted in Figure S22. The ligand, H 2 L, as well as the molecules used to synthetise it was also tested. While the benzothiazole precursor (HzBzTz) presented no effect, the aldehyde (HQ2CHO) showed cytotoxicity against the human melanoma and colon cancer cell lines. Regarding the ligand, a high cytotoxicity was observed for all tested cell lines ( Figure S23 and Table S1). Similar results have been also described for other different Schiff bases [56,57].
MTT assay was performed for the selected Ni and Ru complexes using a wide range of concentrations from 0.5 to 40 µM. The respective dose-response curves are shown in Figure 8A,B and the calculated IC 50 values are presented in Table 5.
therapy. Thus, to evaluate the antiproliferative properties of the newly synthetised m complexes, a preliminary screening was performed using human (HCT-116 and A and murine (CT-26 and B16F10) colon cancer and melanoma cell lines, respectively human keratinocyte cell line, HaCaT, was also included in this in vitro study aimi evaluate the specificity of the new compounds against tumour cells. Complexes evaluated at concentrations of 20 and 40 μM and the obtained results are depict Figure S22. The ligand, H2L, as well as the molecules used to synthetise it was also te While the benzothiazole precursor (HzBzTz) presented no effect, the alde (HQ2CHO) showed cytotoxicity against the human melanoma and colon cancer cell Regarding the ligand, a high cytotoxicity was observed for all tested cell lines (Figur  and Table S1). Similar results have been also described for other different Schiff [56,57].
Overall, most complexes exhibited antiproliferative activity towards all cell with the exception of [Fe(HL)2Cl(H2O)], which showed no cytotoxic effect in the t concentrations. Among all screened compounds, [Ru(HL)Cl(DMSO)] [Ni(HL)(acetate)] were selected for further in vitro studies based on their higher selec towards colon cancer cell lines in comparison to HaCaT ( Figure S22).
MTT assay was performed for the selected Ni and Ru complexes using a wide r of concentrations from 0.5 to 40 μM. The respective dose-response curves are show Figure 8A,B and the calculated IC50 values are presented in Table 5.    Both complexes demonstrated notable antiproliferative activity in murine and human colon cancer cell lines, with [Ni(HL)(acetate)] proving to be more effective than [Ru(HL)Cl(DMSO)]. Importantly, in HCT-116 cells, both complexes displayed higher cytotoxicity than the clinically approved 5-FU.

IC50 (µM) ± SD
In the literature, several other nickel complexes have also exhibited antiproliferative activity against colon cancer cell lines. For example, two Ni-Schiff base complexes exhibited IC 50 values of 7.9 [58] and 26.5 µM [59] towards HCT-116 cells, after 48 h incubation. Additionally, three other Ni complexes containing thiosemicarbazones were incubated with HCT-116 cells for 72 h, displaying IC 50 values between 11.7 and 18 µM [56], the same range as our nickel complex after a shorter (48 h) incubation period. In fact, some of the nickel complexes, reported by Heng and co-workers, were equally or less cytotoxic when compared with their respective ligand, suggesting that the observed antitumour activity may not derive from complexation with nickel [56].
Likewise, many ruthenium complexes have been described in the literature as potential compounds against colon cancer. When tested against CT-26 cells, two ruthenium compounds demonstrated IC 50 values of 50.9 and 148 µM, after a 72 h incubation time, a much lower cytotoxic effect compared to our ruthenium complex [60]. Many other complexes of this metal have been tested against HCT-116 cells. For example, Kubanik et. al. evaluated a variety of Ru complexes with the same 8-hydroxyquinoline scaffold but with different substituents. Almost all showed high cytotoxic activity with IC 50 values ranging from 1.98 to 14.6 µM [61]. Furthermore, in a study conducted by Xu and co-workers [62], several Ru complexes have also exhibited IC 50 values in the low micromolar concentration range against colorectal, liver, lung and cervical carcinoma cell lines, being more potent than the clinically approved cisplatin [62].
On the other hand, several other studies reported lower potency of ruthenium compounds against HCT-116 cells (IC 50 values from 22 to > 100 µM) [58,62] compared to our Ru complex, [Ru(HL)Cl(DMSO)]. Moreover, two 8-hydroxyquinoline Ru complexes with a similar structure to [Ru(HL)Cl(DMSO)] exhibited up to 18-fold lower antiproliferative activity [63]. In another study, a ruthenium-based 5-FU complex was synthetised and tested against a large panel of different cell lines. In most cases, this hybrid complex demonstrated superior cytotoxicity compared to 5-FU alone and, a 3-fold increase in potency was observed towards HCT-116 cells [64].
Noteworthy, NAMI-A was the first Ru complex to enter clinical trials in patients with non-small cell lung cancer [65] and solid tumours [66]. Curiously, previous in vitro studies showed minimal cytotoxicity against cancer cells [67] exhibiting a 1000-fold lower activity than cisplatin [68]. Nevertheless, in vivo experiments demonstrated NAMI-A antimetastatic potential and safety [69].

Guava ViaCount Assay
To corroborate data from MTT assays, the analysis of cell viability by Guava Via-Count flow cytometry was also performed for the murine colon cancer cell line, CT-26 ( Figure 9A,B). A concentration-dependent decrease in cell viability was observed for the tested compounds, being in-line with the data described in Section 3.7. A higher percentage of viable cells was observed for the tested compounds in comparison to data from MTT assay.
Importantly, when compared to the negative control, a much lower cell concentration (three-fold) was observed following incubation with the metal-based complexes under study ( Figure 9B). While MTT assay evaluates the mitochondrial activity of adherent cells, in Guava Via count all cells are analysed using two DNA binding dyes [33,34]. Nevertheless, these results confirm the inhibition of cancer cell proliferation and induction of cell death by the Ni and Ru complexes. For 5-FU, a higher percentage of dead and apoptotic cells was observed. These results might suggest an advantage of our complexes in terms of safety as their action was predominantly to halt cell proliferation, unlike 5-FU, which considerably induced cell death.

Cell Cycle Analysis
One of the hallmarks of cancer is the rapid overgrowth of malignant cells, which mainly happens due to a dysfunction in their cell cycle regulation [56,70,71]. The induction of cell cycle arrest is, therefore, an important anticancer mechanism of action which helps reduce the uncontrolled cell proliferation. This mechanism is exploited by many anticancer drugs, including the positive control used in this study, 5-FU [56,58,70]. In this sense, to better understand how [Ni(HL)(acetate)] and [Ru(HL)Cl(DMSO)] exert their antitumour effect, cell cycle progression was evaluated in CT-26 and HCT-116 cell lines ( Figure 10). To accomplish this, cells were incubated for 24 h with [Ni(HL)(acetate)], [Ru(HL)Cl(DMSO)] and 5-FU at sub-toxic concentrations (below the IC 50 ) to avoid excessive cytotoxicity, thus ensuring the quality of obtained results [71,72].
[Ni(HL)(acetate)] arrested cell cycle at S phase in murine CT-26 cells at a concentration close to its IC 50 value ( Figure 10A,C); however, no such effect was observed in the human HCT-116 cells at the tested concentrations ( Figure 10B,D). In the literature, different nickel complexes also showed the capacity to arrest the cell cycle of leukocytes [73] and human gastric carcinoma cells [74] at S phase. Furthermore, in MCF7 human breast cancer cells, another nickel complex triggered cell cycle arrest at S phase and enhanced DNA fragmentation, ultimately inducing cell apoptosis [59]. On the other hand, two other Schiff base ligands, along with their respective nickel complexes, arrested cell cycle at G0/G1 phase, in HCT-116 cells [56].
For [Ru(HL)Cl(DMSO)] a cell cycle arrest at S phase in both murine and human colon cancer cell lines was achieved. The action of other ruthenium complexes on cell cycle progression has also been widely described in the literature. A Ru triphenylphosphine complex [59] and a Ru terpyridine complex [60] demonstrated the ability to arrest cell cycle at G2/M phase in HCT-116 cells. Additionally, a bis-pyrazolylpyridine Ru(III) complex was shown to arrest CT-26 cell cycle at G0/G1 phase. This mechanism of action is useful, since it stops cancer cell proliferation; however, acting at early phases of the cell cycle may allow cells to repair the damage caused by the anticancer agent [68]. Human HT-29 colon cancer cell lines incubated with a Ru quercetin complex [75] and a Ru phloretin complex [76], arrested cell cycle at G0/G1 and G2/M phases, respectively. The ruthenium 5-FU complex, previously mentioned in Section 3.7, did not arrest the cell cycle of HCT-116 cells at any phase, opposed to 5-FU alone that increased the number of cells in S phase. These results demonstrate the impact that both metals and ligands may exert in the biological activity of compounds [64].
In the present work, 5-FU was used as a positive control group. This clinically approved molecule is an antimetabolite anticancer drug that mainly impairs DNA replication during S phase [64,77,78], inhibiting cell cycle progression and inducing cell apoptosis [77,78]. In vitro studies performed by Afrin et al. and Yang et al. on HCT-116 cells have showed that 5-FU, at concentrations around 20 µM, arrested cell cycle at S phase, [79,80] thus validating our results.
Overall, it is well-known the variety of alternative mechanisms of action by which metal complexes may exert their biological or pharmacological activity, besides impacting cell cycle [4,8,67]. Various nickel complexes have shown to act through the formation of ROS species and inducing mitochondrial-mediated apoptosis [56,74,81,82] or by activating the tyrosine kinase proapoptotic pathway [58]. In addition, other ruthenium complexes have also demonstrated to exert activity through the apoptotic pathway via cleaved PARP, caspase-3 and caspase-9 [64,83]; lysosome-mediated apoptosis [62]; necrosis [60]; and possibly inhibiting DNA replication by binding to DNA strands [59].

Migration Assay
Cancer metastasis occurs in latter stages of the disease, involving the spreading of cancer cells from the primary tumour site to secondary sites, and represents a major factor in cancer-related deaths [84,85]. Cell migration plays a pivotal role in cancer metastasization and, therefore, it is important to investigate the potential effect of novel compounds in this cellular process [36,62,67,81]. In this regard, the antimetastatic potential of [Ni(HL)(acetate)] and [Ru(HL)Cl(DMSO)] was evaluated through a wound-healing assay in CT-26 and HCT-116 cell lines. To accomplish this, cells were incubated with [Ni(HL)(acetate)], [Ru(HL)Cl(DMSO)] and 5-FU at concentrations close to the IC 50 value of each compound [81]. Cell migration was monitored over time after wound scratch, as depicted in Figure 11A

Conclusions
In this study, novel metal complexes were synthetised using a Schiff base derived from 8-hydroxyquinoline and benzothiazole. Seven new complexes were obtained and characterised by analytical and spectroscopic techniques, with the following stoichiometries being assigned: Some nickel complexes have shown high suppression of cell migration in lung carcinoma cell lines, possibly due to the silencing of genes related to epithelial-mesenchymal transition [86]. In a study conducted by Qi and co-workers, a nickel complex also inhibited cell migration of hepatoma carcinoma cells [87]. Similarly, in the literature, various examples of ruthenium complexes demonstrating antimigratory effects can be found. A study performed by Xu et al. showed that CT-26 cells treated with a ruthenium complex displayed a pronounced inhibitory effect likely due to downregulation of the cell migration-related proteins metalloproteinase-2 and metalloproteinase-9 [62]. Additionally, as aforementioned, the ruthenium complex NAMI-A has entered clinical trials due to its specific antimetastatic activity [67]. This compound has also demonstrated an inhibitory effect in reducing cell adhesion and migration in HCT-116 cell line. This complex impaired the interaction between the cells and fibronectin and collagen I, two extracellular proteins closely related to cellular migration in the liver, an organ where colon cancer often metastasizes [67].

Conclusions
In this study, novel metal complexes were synthetised using a Schiff base derived from 8-hydroxyquinoline and benzothiazole. Seven new complexes were obtained and characterised by analytical and spectroscopic techniques, with the following stoichiometries being assigned: . This mechanism of action is in accordance with the positive control used in this study, 5-FU, which interferes with DNA replication. In addition, cell migration of HCT-116 was significantly impaired by both complexes.
Overall, these promising results emphasise the therapeutic potential of these nickel and ruthenium complexes as alternative and effective therapies against colorectal cancer.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pharmaceutics14122583/s1. Reference [88] is cited in the supplementary materials. Figure Figure S22: Antiproliferative effect of a panel of metal-based complexes towards murine (B16F10) and human (A375) melanoma cell lines, human epidermal keratinocytes (HaCat) and murine (CT-26) and human (HCT-116) colon cancer cell lines. All compounds were tested at 20 and 40 µM. Cell viability was determined by MTT assay after a 48 h incubation period. Figure  S23: Antiproliferative effect of the ligand H 2 L towards murine (CT-26) and human (HCT-116) colon cancer cell lines. Cell viability was determined by MTT assay after a 48 h incubation period. Table S1: Half-inhibitory concentration (IC 50 ) of H 2 L towards human and murine colon cancer cell lines, after a 48 h incubation period.