Purine and Purine Isostere Derivatives of Ferrocene: An Evaluation of ADME, Antitumor and Electrochemical Properties.

Novel purine and purine isosteres containing a ferrocene motif and 4,1-disubstituted (11a−11c, 12a−12c, 13a−13c, 14a−14c, 15a−15c, 16a, 23a−23c, 24a−24c, 25a−25c) and 1,4-disubstituted (34a−34c and 35a−35c) 1,2,3-triazole rings were synthesized. The most potent cytotoxic effect on colorectal adenocarcinoma (SW620) was exerted by the 6-chloro-7-deazapurine 11c (IC50 = 9.07 µM), 6-chloropurine 13a (IC50 = 14.38 µM) and 15b (IC50 = 15.50 µM) ferrocenylalkyl derivatives. The N-9 isomer of 6-chloropurine 13a containing ferrocenylmethylene unit showed a favourable in vitro physicochemical and ADME properties including high solubility, moderate permeability and good metabolic stability in human liver microsomes.


Introduction
Nucleoside analogues were among the first chemotherapeutic agents that were introduced for the medical treatment of cancer [1,2]. Cytotoxic nucleobase-derived compounds have gained a lot of importance in recent years for combating cancer of different types, usually in combination with other agents [3][4][5]. They were shown to interact with various biological targets, such as cellular kinases, ribonucleotide reductase, and cellular polymerases in order to produce cytotoxic effects [6,7]. The interest in metal complexes of ferrocene-based ligands is due to their favourable physicochemical properties and reversible redox properties that enable ferrocene derivatives to be used as excellent candidates for drug development [8,9]. The antitumor activity of ferrocene-containing compounds is shown to be related to cytotoxic ferrocenium cation and generation of reactive oxygen species (ROS), especially hydroxyl radicals that can be produced under oxidizing conditions through a Fenton-type reaction, whose presence may lead to the damage of the cells [10,11]. To enhance the cytotoxic activity, the ferrocenyl moiety has been successfully incorporated into molecules that have biological activity [12]. In this regard, ferrocifen, obtained by replacing the phenyl group Molecules 2020, 25 in tamoxifen with ferrocene, exhibits a strong effect against both hormone-dependent (ERα + ) and hormone-independent (ERα − ) breast cancer cell types in contrast to tamoxifen, which is only active against ERα + cells [13,14]. Furthermore, the antimalarial drug ferroquine was recently found to enhance anticancer activity of several chemotherapeutics suggesting its potential application as an adjuvant to existing anticancer therapy [15]. It negatively regulates hypoxia-inducible factor-1α (HIF-1α) and showed to be effective under starved and hypoxic conditions frequently observed in advanced solid cancers. The anticancer potential of organometallic nucleoside analogues was explored [16][17][18] and showed that nucleoside analogues of ferrocene exhibited potent antineoplastic activity [19][20][21][22]. Combined application of ferrocenylalkyl nucleobase with anticancer drug cyclophosphamide demonstrated therapeutic synergism of antitumor activity against Lewis lung carcinoma (LLC) [23]. Besides, ferrocene derivatives of thieno [2,3-d]pyrimidine, as purine isosteres, exhibited potent anticancer effects in several breast cancer and AML (acute myeloid leukemia) cell lines, despite a loss of mitogen-activated protein kinase-interacting kinase (MNK) potency [24]. On the other hand, ferrocenes incorporating cinchona or carbohydrate moiety through 1,2,3-triazole linker showed significant cytostatic activity [25,26]. Some ferrocene-cinchona hybrids increased reactive oxygen species (ROS) production and induced mitochondrial damage in human multidrug resistant (MDR) cancer cells [27]. Taking into consideration the biological relevance of nucleoside analogues of ferrocene, here we have described the synthesis of novel ferrocene-tagged purine and purine isosteres with the aim to evaluate their antiproliferative activity. Preliminary ADME properties and electrochemical oxidation potential of compounds that exhibited best growth-inhibitory activity were also assessed.

Evaluation of Permeability and Metabolic Stability
Compounds 11a−11c, 13a, 15a and 15b with highest cytostatic activities were further screened for permeability and P-glycoprotein (P-gp) substrate assessment, as well as metabolic stability in liver microsomes (human and mouse). These properties have been shown to significantly affect the bioavailability of drugs and to be important in the early drug discovery phases and the lead optimization process [39].
The Madin-Darby canine kidney cells with overexpressed human multidrug-resistant gene (MDCKII-hMDR1) have been frequently used to study bidirectional transport of compounds and assess P-gp substrate potential [40]. Apparent permeability (P app ) values were determined from the amount permeated through the MDCKII-hMDR1 cell monolayer in both the apical-basolateral (AB) and basolateral-apical (BA) direction. P app values and the efflux ratios with and without P-gp inhibition with Elacridar are listed in Table 2.
Compound 11b is characterized by a high permeability, with an efflux ratio of 1.5 suggesting no influence of P-pg on permeability. All the remaining compounds exhibited a moderate permeability (P app A/B), with efflux ratios above 2, classifying them as P-gp substrates, which was further confirmed when they were incubated with a P-gp inhibitor, elacridar. As previously reported in the literature [41], a comparison of efflux ratio generated in the presence and absence of a P-gp inhibitor can be used for the identification of P-gp substrates. A compound is typically considered to be a P-gp substrate when the efflux ratio in the absence of inhibitor is greater than 2 and/or is at least 50% drop is observed in the presence of inhibitor [42]. Among compounds with moderate permeability the 6-chloro-7-deazapurine derivative 11c with a directly connected ferrocene moiety and N-7 regioisomers of 6-chloropurine 15a and 15b showed moderate AB permeability, while compounds 11a and 13a had moderate to high AB permeability in the range from 8.3 to 8.48 × 10 −6 cm/s.  11a−11c, 13a, 15a, 15b in MDCKII-hMDR1 cell assay from apical-to-basolateral (AB) and basolateral-to-apical (BA) side without and with inhibition of P-glycoprotein (P-gp) efflux pump with Elacridar. bioavailability of drugs and to be important in the early drug discovery phases and the lead optimization process [39]. The Madin-Darby canine kidney cells with overexpressed human multidrug-resistant gene (MDCKII-hMDR1) have been frequently used to study bidirectional transport of compounds and assess P-gp substrate potential [40]. Apparent permeability (Papp) values were determined from the amount permeated through the MDCKII-hMDR1 cell monolayer in both the apical-basolateral (AB) and basolateral-apical (BA) direction. Papp values and the efflux ratios with and without P-gp inhibition with Elacridar are listed in Table 2.  11a−11c, 13a, 15a, 15b in MDCKII-hMDR1 cell assay from apical-to-basolateral (AB) and basolateral-to-apical (BA) side without and with inhibition of Pglycoprotein (P-gp) efflux pump with Elacridar. Compound 11b is characterized by a high permeability, with an efflux ratio of 1.5 suggesting no influence of P-pg on permeability. All the remaining compounds exhibited a moderate permeability (PappA/B), with efflux ratios above 2, classifying them as P-gp substrates, which was further confirmed when they were incubated with a P-gp inhibitor, elacridar. As previously reported in the literature [41], a comparison of efflux ratio generated in the presence and absence of a P-gp inhibitor can be used for the identification of P-gp substrates. A compound is typically considered to be a P-gp substrate when the efflux ratio in the absence of inhibitor is greater than 2 and/or is at least 50% drop is observed in the presence of inhibitor [42]. Among compounds with moderate permeability the 6chloro-7-deazapurine derivative 11c with a directly connected ferrocene moiety and N-7 regioisomers of 6-chloropurine 15a and 15b showed moderate AB permeability, while compounds 11a and 13a had moderate to high AB permeability in the range from 8.3 to 8.48 × 10 −6 cm/s. bioavailability of drugs and to be important in the early drug discovery phases and the lead optimization process [39].

Compd Heterocycle R
The Madin-Darby canine kidney cells with overexpressed human multidrug-resistant gene (MDCKII-hMDR1) have been frequently used to study bidirectional transport of compounds and assess P-gp substrate potential [40]. Apparent permeability (Papp) values were determined from the amount permeated through the MDCKII-hMDR1 cell monolayer in both the apical-basolateral (AB) and basolateral-apical (BA) direction. Papp values and the efflux ratios with and without P-gp inhibition with Elacridar are listed in Table 2.  11a−11c, 13a, 15a, 15b in MDCKII-hMDR1 cell assay from apical-to-basolateral (AB) and basolateral-to-apical (BA) side without and with inhibition of Pglycoprotein (P-gp) efflux pump with Elacridar. Compound 11b is characterized by a high permeability, with an efflux ratio of 1.5 suggesting no influence of P-pg on permeability. All the remaining compounds exhibited a moderate permeability (PappA/B), with efflux ratios above 2, classifying them as P-gp substrates, which was further confirmed when they were incubated with a P-gp inhibitor, elacridar. As previously reported in the literature [41], a comparison of efflux ratio generated in the presence and absence of a P-gp inhibitor can be used for the identification of P-gp substrates. A compound is typically considered to be a P-gp substrate when the efflux ratio in the absence of inhibitor is greater than 2 and/or is at least 50% drop is observed in the presence of inhibitor [42]. Among compounds with moderate permeability the 6chloro-7-deazapurine derivative 11c with a directly connected ferrocene moiety and N-7 regioisomers of 6-chloropurine 15a and 15b showed moderate AB permeability, while compounds 11a and 13a had moderate to high AB permeability in the range from 8.3 to 8.48 × 10 −6 cm/s. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and square- Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments. Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and square- Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and square- Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and square- Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and squarewave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and square- Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds.

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of Incubation with mouse liver microsomes suggested that that all tested compounds are metabolically unstable with a predicted in vivo clearance above 97% of liver blood flow as summarized in Table 3. No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of No substantial differences in clearance values for these compounds were observed in mouse microsomes. On the other hand, incubation with human liver microsomes showed the significant impact of type of nitrogen heterocycle on stability. Namely, 6-chloro-7-deazapurine derivatives 11a−11c exhibiting high clearance values (>90% LBF) can be considered as highly labile compounds. On the contrary, 6-chloropurine derivative 13a, as structural analogue of 6-chloro-7-deazapurine 11a, with clearance value of 65% LBF showed to be moderately stable. Additionally, N-7 regioisomer 15a had somewhat increased microsomal stability (58% LBF) than its N-9 analogue 13a (65% LBF).

Electrochemical Properties of Selected Compounds by Voltammetric Assays
Compounds 11c, 13a and 15b with strong antiproliferative activity and good physicochemical properties were further evaluated for their electrochemical properties under a wide range of solution conditions. Although the electrochemical behaviour of ferrocene, triazole and their derivatives has been studied in both protic and aprotic solvents [43][44][45][46][47], we have been interested to examine the redox mechanism of the above-mentioned compounds, which may be influenced by the type of nitrogen heterocycle, ferrocene, and alkyl spacer between ferrocene and triazole unit.
The electrochemical properties of 6-chloro-7-deazapurine and 6-chloropurine derivatives of ferrocene 11c, 13a and 15b on a glassy carbon electrode were studied using cyclic (CV) and square-wave voltammetry (SWV) in aqueous electrolyte solutions over a wide pH range. The pH was varied from 2 to 10 in order to determine the effect of pH on the voltammograms (net peak currents and potentials). Figure 1 shows square-wave and cyclic voltammograms for the oxidation of 0.1 mM of 13a in aqueous buffer electrolytes 3 ≤ pH ≤ 10. The pH of the solution affects the voltammetric response of this compound, i.e., the oxidation of evaluated compounds 11c, 13a and 15b is strongly dependent on it. Furthermore, as can be seen, the best results were obtained at pH 9, so this value was selected for the following experiments.   Figure 2). Two peaks at ~0.4 V (P1) and ~1.1 V (P2) can be observed, indicating that these molecules have two redox active centres. The peak P1 occurs in all supporting electrolytes (i.e., at 2 ≤ pH ≤ 10), while the peak P2 is only observed at pH ≥ 9 (see Figure 1). The potential of peak P1 is pH-independent. Comparing to the literature data, peak P1 can be ascribed to the following charge transfer reaction [Fe II (C5H5)2] ⇌ [Fe III (C5H5)2] + + e − [44,48]. The forward and backward components of peak P1 are oxidative and reductive currents, respectively (see inset in Figure 2), which indicates that the oxidation of ferrocene is reversible electrode reaction.  Figure 2 shows square-wave voltammograms for the oxidation of 11c, 13a and 15b on glassy carbon electrode, in 0.5 mol/L NaClO 4 at pH 9. The forward (i f ) and backward (i b ) components of the net response (∆i = i f − i b ) are shown as well (inset in Figure 2). Two peaks at~0.4 V (P1) and~1.1 V (P2) can be observed, indicating that these molecules have two redox active centres. The peak P1 occurs in all supporting electrolytes (i.e., at 2 ≤ pH ≤ 10), while the peak P2 is only observed at pH ≥ 9 (see Figure 1). The potential of peak P1 is pH-independent. Comparing to the literature data, peak P1 can be ascribed to the following charge transfer reaction [Fe II (C 5 [Fe III (C 5 H 5 ) 2 ] + + e − [44,48]. The forward and backward components of peak P1 are oxidative and reductive currents, respectively (see inset in Figure 2), which indicates that the oxidation of ferrocene is reversible electrode reaction.
(P2) can be observed, indicating that these molecules have two redox active centres. The peak P1 occurs in all supporting electrolytes (i.e., at 2 ≤ pH ≤ 10), while the peak P2 is only observed at pH ≥ 9 (see Figure 1). The potential of peak P1 is pH-independent. Comparing to the literature data, peak P1 can be ascribed to the following charge transfer reaction [Fe II (C5H5)2] ⇌ [Fe III (C5H5)2] + + e − [44,48]. The forward and backward components of peak P1 are oxidative and reductive currents, respectively (see inset in Figure 2), which indicates that the oxidation of ferrocene is reversible electrode reaction. Furthermore, in order to clarify the origin of peak P2, ethynyl ferrocene (i) and 6-chloropurine (ii) (bearing no triazole moiety) were analysed as well (under otherwise identical conditions, Figure  3). No anodic peak P2 was detected on SWV responses of these compounds within the investigated potential window. Comparing with the SWV response of 13a (black curve in Figure 3), it can be Furthermore, in order to clarify the origin of peak P2, ethynyl ferrocene (i) and 6-chloropurine (ii) (bearing no triazole moiety) were analysed as well (under otherwise identical conditions, Figure 3). No anodic peak P2 was detected on SWV responses of these compounds within the investigated potential window. Comparing with the SWV response of 13a (black curve in Figure 3), it can be concluded that the peak P2 corresponds to the electrode reaction of triazole ring. Both components (i f and i b ) of the peak P2 are oxidative currents (see inset in Figure 2), which indicate that the electro-oxidation of 1,2,3-triazole in 11c, 13a and 15b at pH 9 is totally irreversible electrode reaction. The same results were obtained by cyclic voltammetry (under otherwise identical conditions). These observations are in agreement with existing literature on the electrochemical oxidation of triazole-acridine conjugates [49].
Molecules 2020, 25, x FOR PEER REVIEW 10 of 24 concluded that the peak P2 corresponds to the electrode reaction of triazole ring. Both components (if and ib) of the peak P2 are oxidative currents (see inset in Figure 2), which indicate that the electrooxidation of 1,2,3-triazole in 11c, 13a and 15b at pH 9 is totally irreversible electrode reaction. The same results were obtained by cyclic voltammetry (under otherwise identical conditions). These observations are in agreement with existing literature on the electrochemical oxidation of triazoleacridine conjugates [49]. In addition, it can be seen from Figure 2 that all investigated compounds provided similar electrochemical responses. More precisely, the electro-oxidation of the 1,2,3-triazole moiety (i.e., peak P2) in 11c, 13a and 15b takes place at almost the same potential. Taking into account the structural difference between 13a and 15b in type of the alkyl bridge between the two redox-active centres triazole and ferrocene, it is reasonable to conclude that its effect on the oxidation of the triazole in In addition, it can be seen from Figure 2 that all investigated compounds provided similar electrochemical responses. More precisely, the electro-oxidation of the 1,2,3-triazole moiety (i.e., peak P2) in 11c, 13a and 15b takes place at almost the same potential. Taking into account the structural difference between 13a and 15b in type of the alkyl bridge between the two redox-active centres triazole and ferrocene, it is reasonable to conclude that its effect on the oxidation of the triazole in studied compounds is negligible. However, in the case of 11c, where the triazole ring is directly attached to the ferrocene, the electrochemical oxidation of ferrocene (i.e., peak P1) occurs at more positive potentials. This result indicates that iron(II) in 11c is more difficult to oxidize due to the stronger electron-withdrawing effect of the 1,2,3-triazoles ring directly attached to the ferrocene nucleus, or due to its steric hindrance effect on the Fe(II) ion in 11c.

General Procedure for the Synthesis of Purine and Purinomimetics with Ferrocene at N-1 of 1,2,3-Triazole
The corresponding N-propargylated heterocyclic base 5−10 (1 eq.) was dissolved in methanol, and the corresponding terminal azide (1.2 eq.) and Cu(OAc) 2 (0.05 eq.) were added. The reaction mixture was stirred at room temperature overnight. The solvent was removed under reduced pressure and the residue was purified by column chromatography (CH 2 Cl 2 :CH 3 OH = 60:1).

Permeability and P-glicoprotein Substrate Assessment
MDCKII-hMDR1 cell monolayer was used for determination of permeability and P-gp substrate assessment. MDCKII-hMDR1 cells (Solvo Biotechnology, Szeged, Hungary) were grown in the controlled atmosphere (37 • C, 95% humidity, 5% CO 2 ) in MDCK cell culture medium containing Dulbecco's Modified Eagle Medium (DMEM) + 10% Fetal bovine serum (FBS, inactivated) + 1% glutamax-100 + 1% antibiotic/antimycotic + 1% of Non-essential amino acids Minimum Essential Medium (MEM NEAA). Cultures were split every 3-4 days using 0.05% Trypsin-EDTA. Cells were seeded 4 days prior to the experiment in concentration of 0.3 × 10 6 cells mL −1 and cultured in CO 2 incubator. On the experiment day, cell monolayers were washed with Dulbecco's phosphate buffer saline (D-PBS) and equilibrated for 45 min in CO 2 incubator in D-PBS containing 1% DMSO (1% v/v) with or without Elacridar (2 µM). Compounds, in final concentration of 10 µM, were prepared in transport medium containing Lucifer yellow (100 µM), with DMSO contentment of 1% (v/v) and in the presence and without Elacridar. Lucifer yellow calibration curve was prepared by serial dilution (12 points, with 100 µM as the highest point). Monolayer integrity was examined by fluorescent measurement on a Microplate reader Infinite F500 (Tecan) by using an excitation of 485 nm and emission of 530 nm. Experiment was started by applying the solutions containing test compounds to apical and basolateral side of the cell monolayer. Starting compound concentration (C 0 ) was sampled at the start of the experiment and apical and basolateral compartments were sampled after incubation at 37 • C with gentle shaking for 60 min. Compound were tested in duplicate. Amprenavir was used as a control with low permeability and as a P-gp substrate without inhibitor presence, turning into a highly permeable compound with the inhibitor present. Diclofenac was used as a control with high permeability and no interaction with P-gp.
Samples were analyzed on an ABSciex API 4000 Triple Quadrupole Mass Spectrometer (Sciex, Division of MDS Inc., Toronto, ON, Canada) coupled to a UHPLC Nexera X2 (Shimadzu, Kyoto, Japan). Samples (1 µL) were injected onto a reversed phase HPLC column Luna Omega Polar C18, 30 × 2.1 mm i.d., 1.6 µm particle size (Phenomenex) which was kept at 50 • C. Aqueous solution was 0.1% formic acid in deionized water and the organic mobile phase was water/acetonitrile/formic acid mixture (90/10/0.1, v/v/v). Flow rate was kept at 0.7 mL min −1 for all measurements. Fast-gradient elution was as follows: 2% of organic mobile phase through 0.15 min, from 2 to 95% of organic mobile phase (0.15−0.7 min), 95% of organic mobile phase (0.7−1.1 min), from 95 to 2% of organic mobile phase (1.1−1.11 min), and re-equilibration with 98% of aqueous mobile phase (1.11−1.5 min). Positive ion mode with turbo spray, an ion source temperature of 550 • C and a dwell time of 75 ms were utilized for mass spectrometric detection. Quantitation was performed using multiple reaction monitoring (MRM) at the specific transition corresponding to the compound of interest. Warfarin was used as an internal standard. The ratios between compound and internal standard peak areas were used instead of real compound concentration. Apparent permeability coefficient, P app (nm s −1 ) values were calculated by using the following Equation (1): where: dQ/dT = permeability rate; C 0 = initial concentration in donor compartment; A = surface area of the cell monolayer (0.7 cm 2 ). The efflux ratio in the presence or absence of the P-gp inhibitor was calculated from P app values, using following Equation (2): Efflux ratio = P app (BA)/P app (AB)

Metabolic Stability in Liver Microsomes
Metabolic stability of compounds was assessed in human and mouse liver microsomes (Corning, Tewksbury, MA, USA). Compounds, in final concentration of 1 µM, were incubated in phosphate buffer (50 mM, pH 7.4) for 60 min at 37 • C together with liver microsomes and NADPH generating system (nicotinamide adenine dinucleotide phosphate (NADP, 0.5 mM), glucose-6-phosphate (G6P, 5 mM), glucose-6-phosphate dehydrogenase (1.5 U mL −1 ) and magnesium chloride (0.5 mM)). Also, compounds were incubated without the presence of NADPH cofactor, as a buffer stability control. Metabolic activity of liver microsomes was verified by including testosterone and propranolol as positive controls, as well as caffeine as a negative control. Sampling was performed at six time points (0, 10, 20, 30, 45 and 60 min), followed by reaction termination by addition of acetonitrile/methanol mixture (2:1, v/v) containing dicofenac as internal standard. Samples were analyzed as previously described for samples from permeability assay. The in vitro half-life (t 1/2 ) was calculated from the slope of the linear regression by plotting ln % remaining of parent compound against incubation time. In vitro intrinsic clearance, CL int was calculated from half-life using following Equation (3)

Voltammetric Measurements
All chemicals used in the experiments were of the best grade commercially available (Sigma-Aldrich) and were used without further purification. Stock standard solutions of compounds 11c, 13a, 15b (c = 2.0 × 10 −3 mol/L) and 6-chloropurine (c = 1.2 × 10 −2 mol/L) were prepared from dry pure substances in dimethyl sulfoxide (DMSO, p.a.), purchased from Kemika (Zagreb, Croatia). Stock standard solution of ethynyl ferrocene (c = 1.1 × 10 −2 mol/L) was prepared from dry pure substance in ethanol (Kemika). For the supporting electrolyte, analytical grade NaClO 4 (Kemika) was used. Water was deionized by the Millipore Milli-Q system to the resistivity ≥ 18 MΩcm. Voltammetric measurements were carried out using the computer-controlled electrochemical system "PGSTAT 101" (Eco-Chemie, Utrecht, The Netherlands), controlled by the electrochemical software "NOVA 1.5". A three-electrode system (BioLogic, Claix, France) with glassy carbon electrode (GCE) of 3 mm in diameter as a working electrode, Ag/AgCl (3 mol/L NaCl) as a reference electrode and a platinum wire as a counter electrode were used. All potentials were expressed versus Ag/AgCl (3 mol/L NaCl) reference electrode. The supporting electrolyte (0.5 mol/L NaClO 4 ), adjusted to the desired pH value, was placed in the electrochemical cell and the required aliquot of the standard analyte solution was added. The solution in the electrochemical cell was degassed with high purity nitrogen for 10 min before measurement, and the nitrogen blanket was maintained thereafter. Before each run, the glassy carbon working electrode was polished with diamond suspension in spray (grain size 6 µm) and rinsed with ethanol and deionized water. All experiments were performed at room temperature. The presented results are reported as the mean value of three independent measurements.
Taken together, we may conclude that compound 13a is highlighted for further structural optimization to obtain more effective and selective ferrocene-tagged purine and/or purine isostere derivatives against SW620 cell lines.