TSPO Ligand-Methotrexate Prodrug Conjugates: Design, Synthesis, and Biological Evaluation

The 18-kDa translocator protein (TSPO) is a potential mitochondrial target for drug delivery to tumors overexpressing TSPO, including brain cancers, and selective TSPO ligands have been successfully used to selectively deliver drugs into the target. Methotrexate (MTX) is an anticancer drug of choice for the treatment of several cancers, but its permeability through the blood brain barrier (BBB) is poor, making it unsuitable for the treatment of brain tumors. Therefore, in this study, MTX was selected to achieve two TSPO ligand-MTX conjugates (TSPO ligand α-MTX and TSPO ligand γ-MTX), potentially useful for the treatment of TSPO-rich cancers, including brain tumors. In this work, we have presented the synthesis, the physicochemical characterizations, as well as the in vitro stabilities of the new TSPO ligand-MTX conjugates. The binding affinity for TSPO and the selectivity versus central-type benzodiazepine receptor (CBR) was also investigated. The cytotoxicity of prepared conjugates was evaluated on MTX-sensitive human and rat glioma cell lines overexpressing TSPO. The estimated coefficients of lipophilicity and the stability studies of the conjugates confirm that the synthesized molecules are stable enough in buffer solution at pH 7.4, as well in physiological medium, and show an increased lipophilicity compared to the MTX, compatible with a likely ability to cross the blood brain barrier. The latter feature of two TSPO ligand-MTX conjugates was also confirmed by in vitro permeability studies conducted on Madin-Darby canine kidney cells transfected with the human MDR1 gene (MDCK-MDR1) monolayers. TSPO ligand-MTX conjugates have shown to possess a high binding affinity for TSPO, with IC50 values ranging from 7.2 to 40.3 nM, and exhibited marked toxicity against glioma cells overexpressing TSPO, in comparison with the parent drug MTX.


Introduction
The subcellular 18-kDa translocator protein (TSPO) [1], is an attractive biomarker for molecular imaging and a potential therapeutic target for drug delivery to tumors overexpressing TSPO [2,3]. TSPO is mostly located in the outer mitochondria membrane of steroid-synthesizing cells in peripheral organs system. In contrast, its presence in the central nervous system is delimited to ependymal cells including the glia. TSPO is involved in several pathophysiological processes, such as steroidogenesis, immunomodulation, apoptosis, brain injury, neurodegeneration, and cancer [4][5][6][7]. TSPO is upregulated in neuroinflammation and its overexpression has been proved in several types of cancers, including as steroidogenesis, immunomodulation, apoptosis, brain injury, neurodegeneration, and cancer [4][5][6][7]. TSPO is upregulated in neuroinflammation and its overexpression has been proved in several types of cancers, including gliomas, whereas expression in the healthy brain is low [8,9]. Thus, TSPO could serve as potential therapeutic tool to target brain tumors using TSPO-ligands conjugates with anti-cancer drugs. Classical synthetic TSPO ligands include phenylisoquinoline carboxamides (e.g., PK 11195), benzodiazepines (e.g., RO5-4864), phenoxyphenyl acetamide (e.g., DAA1106), pyrazolo[1,5-a]pyrimidine acetamides (e.g., DPA-713), indoleacetamides (FGIN-1-27), and imidazo[1,2-a]pyridines (e.g., alpidem) [5,10] (Figure 1a). Recently, we have synthesized new 2-phenylimidazo[1,2-a]pyridine acetamides, designed from alpidem by introducing hydrophilic substituents on the imidazopyridine nucleus, that bind the mitochondrial protein with high affinity and selectivity (Figure 1b) [10]. The structure-activity analysis has shown that the substitution at the 8-position of the imidazopyridine nucleus with appropriate lipophilic or hydrophilic groups, combined with a chlorine atom added at the para-position of the phenyl ring are key factors to increase the affinity and the selectivity toward the TSPO binding sites [10]. Additionally, the incorporation of amino-, hydroxyl-, or carboxylic groups allows for convenient bioconjugation with anticancer drugs.  Methotrexate (MTX; 2,4-diamino-N 10 -methylpteroylglutamic acid; 1) is an anticancer drug of choice for the treatment of several cancers such as acute lymphocytic leukemia, choriocarcinoma, non-Hodgkin's lymphoma, gastric, breast, head, and neck cancers [11]. MTX and its active metabolites (polyglutamates) are competitive inhibitors of the enzyme dihydrofolate reductase (DHFR) that lead to blockage of tetrahydrofolate synthesis and the depletion of nucleotide precursors. MTX is a class IV drug in the FDA's Biopharmaceutical Classification System (BCS, Amidon, cFederal Drug and Food Administration, Rockville, MD, USA) with a low permeability (see Table 1) and poor aqueous solubility [12]. It has a short plasma half-life and poor permeability across blood-brain barrier (BBB), when used in normal dosages of the protocols, making it unsuitable to brain tumors [13]. The use of high-dose MTX alone was proposed as the first-line treatment for primary central nervous system (CNS) lymphoma, but the side effects were more obvious. In fact, the myelosuppression, a common side effect and meningitis that often appears in the intrathecal treatment are dose-dependent and are more severe in patients receiving MTX in high dose. Moreover, CNS toxicities, including acute encephalopathy, also occur when the drug is used in high doses [13]. For these reasons, MTX is a suitable candidate for this study and therefore it was selected as a model drug to achieve TSPO ligand-MTX conjugates, potentially useful for the treatment of TSPO-rich cancers, including brain tumors overexpressing the TSPO. In this paper, we describe the synthesis, the physicochemical characterization, as well as the in vitro stabilities of the new TSPO ligand-MTX conjugates. The binding affinity for TSPO and selectivity versus the central-type benzodiazepine receptors (CBR) of conjugates were also evaluated. Furthermore, the cytotoxicity of prepared compounds was investigated on human SF126 and SF188, and rat RG2 and C6 glioma cell lines, together with their ability to permeate MDCK-MDR1 cell monolayers.

Results and Discussion
A drug delivery system, such as a bio-conjugate that carries an anticancer drug to brain tumors overexpressing the mitochondrial protein, should have favorable biopharmaceutical properties that include membrane permeability, high receptor binding affinity and selectivity, cytotoxicity or ability to convert itself into a cytotoxic moiety. The conjugate strategy is widely implemented to optimize biopharmaceutical and pharmacokinetic characteristics of drugs, including the transport across biological barriers and the reduction of adverse side effects [14]. Usually, the design of a conjugate involves the formation of a covalent chemical bond between the drug and a pharmacologically-inactive portion (e.g., a backbone polymer), whilst activation occurs after in vivo administration (conjugate prodrugs) or after reaching the target site upon cellular internalization. The further development of the conjugate is the strategy of the bioconjugate, which contributes in the direct linkage of the drug to a pharmacologically-active portion (e.g., a selective ligand or a peptide), or by the intermediacy of a spacer. In this regard, a number of bioconjugates with selective TSPO ligands have been developed for molecular imaging or for the delivery of hydrophilic anticancer drugs into brain tumors across the BBB. Bioconjugation of nanodevices with TSPO ligands (bio-conjugates with low molecular weight, TSPO decorated nanoparticles, and TSPO-targeted dendrimers) have also been described [15][16][17][18]. Moreover, in our previous studies we pointed out that some selective TSPO-ligands showed apoptotic effects and that the simultaneous transport of a TSPO-ligand with an anticancer drug may result in synergistic effects, precisely the synergism of the antitumor activity of the anticancer drug and of the TSPO ligand [16]. Thus, the aim of this study was to synthesize two new bio-conjugates of the anticancer drug MTX with the potent and highly selective TSPO ligand 2 (Scheme 1).
The TSPO-ligand MTX conjugates were completely characterized by spectroscopic techniques and mass spectrometry. The ESI-HRMS spectra showed a peak at m/z = 876.3459 (TSPO-ligand
The TSPO-ligand MTX conjugates were completely characterized by spectroscopic techniques and mass spectrometry. The ESI-HRMS spectra showed a peak at m/z = 876.3459 (TSPO-ligand MTX conjugate 3, [M´H]´) or at 876.3430 (TSPO-ligand MTX conjugate 4, [M´H]´), both in agreement with the expected chemical formula, C 43 H 48 ClN 13 O 6 . Additionally, the one-dimensional (1D-) and two-dimensional (2D-) nuclear magnetic resonance (NMR) characterization ( 1 H, correlation spectroscopy (COSY), and nuclear overhauser effect spectroscopy (NOESY)) provided spectra in full agreement with the structures assigned to 3 and 4. The interpretation of combined 2D spectra can prove extremely useful in discriminating the structure of regioisomers [19]. In the case at hand, the NOESY spectra of the two regioisomers 3 and 4 show major differences in the intensity of cross-peaks occurring between the Gly NH of the TSPO moiety and the protons of Glu side-chain. Figure 2 summarizes these relevant NOESY correlations. In one case, the strong NH/α-CH correlation and the weak NH/β-CH 2 are consistent with conjugation of TSPO ligand to the α-COOH of MTX. In the other case, the absence of the NH/α-CH correlation and the strong NH/γ-CH 2 correlation are distinctive features of the conjugation to the γ-COOH of MTX. prove extremely useful in discriminating the structure of regioisomers [19]. In the case at hand, the NOESY spectra of the two regioisomers 3 and 4 show major differences in the intensity of cross-peaks occurring between the Gly NH of the TSPO moiety and the protons of Glu side-chain. Figure 2 summarizes these relevant NOESY correlations. In one case, the strong NH/α-CH correlation and the weak NH/β-CH2 are consistent with conjugation of TSPO ligand to the α-COOH of MTX. In the other case, the absence of the NH/α-CH correlation and the strong NH/γ-CH2 correlation are distinctive features of the conjugation to the γ-COOH of MTX.

Lipophilicity
The lipophilicity of a molecule, quantitatively expressed as LogP can be useful to predict its permeability through biological barriers. The lipophilicity of TSPO-ligand 2, MTX and of the TSPO-ligand MTX conjugates 3 and 4 was estimated by calculating their 1-octanol/water partition coefficients, using CLOGP software (Toronto, ON, Canada), based on the fragmental method of Hansch and Leo [20]. Compounds that possess a value of LogP ≥2.5 are able to cross the BBB. As it can be seen from data in Table 1, both TSPO-ligand MTX conjugates have logP values higher than +2.5 (i.e., +5.21 for 3 and +4.40 for 4), and appeared to be suitably lipophilic to cross the BBB. Conversely, the parent drug has a negative value of CLogP (−0.24), in line with the experimental evidence (as mentioned above) that MTX is not able to cross the BBB effectively.

Stability Studies
The chemical stability of the TSPO-ligand MTX conjugates 3 and 4 was evaluated in 50 mM phosphate buffer solution at pH 7.4 and 37 °C; the physiological stability was also determined using a dilute rat serum solution (50% v/v) at 37 °C. The half-lives were calculated by the disappearance of the starting conjugate and are reported in Table 1. The aim of this work is the synthesis of TSPO ligand-MTX conjugates in order to improve the plasma stability and having less side effects than MTX. In addition TSPO ligand-MTX conjugates could enhance the transport across BBB and due a tumor targeting effect for the TSPO moiety, might selectively deliver the antineoplastic agent to brain tumors and prolong its efficacy.  (Figure 1), most likely originating from the degradation of ligand 2, as previously shown [18], could be detected. Other products arising from the further degradation of the anticancer drug were found.

Radioligand Binding Assays
Binding affinities of the TSPO-ligand MTX conjugates 3 and 4 for TSPO and central-type benzodiazepine receptor (CBR) were assessed on membrane preparations from rat cerebral cortex. The IC50 values were determined from the displacement curves using the reference compound

Lipophilicity
The lipophilicity of a molecule, quantitatively expressed as LogP can be useful to predict its permeability through biological barriers. The lipophilicity of TSPO-ligand 2, MTX and of the TSPO-ligand MTX conjugates 3 and 4 was estimated by calculating their 1-octanol/water partition coefficients, using CLOGP software (Toronto, ON, Canada), based on the fragmental method of Hansch and Leo [20]. Compounds that possess a value of LogP ě2.5 are able to cross the BBB. As it can be seen from data in Table 1, both TSPO-ligand MTX conjugates have logP values higher than +2.5 (i.e., +5.21 for 3 and +4.40 for 4), and appeared to be suitably lipophilic to cross the BBB. Conversely, the parent drug has a negative value of CLogP (´0.24), in line with the experimental evidence (as mentioned above) that MTX is not able to cross the BBB effectively.

Stability Studies
The chemical stability of the TSPO-ligand MTX conjugates 3 and 4 was evaluated in 50 mM phosphate buffer solution at pH 7.4 and 37˝C; the physiological stability was also determined using a dilute rat serum solution (50% v/v) at 37˝C. The half-lives were calculated by the disappearance of the starting conjugate and are reported in Table 1. The aim of this work is the synthesis of TSPO ligand-MTX conjugates in order to improve the plasma stability and having less side effects than MTX. In addition TSPO ligand-MTX conjugates could enhance the transport across BBB and due a tumor targeting effect for the TSPO moiety, might selectively deliver the antineoplastic agent to brain tumors and prolong its efficacy.  (Figure 1), most likely originating from the degradation of ligand 2, as previously shown [18], could be detected. Other products arising from the further degradation of the anticancer drug were found.

Radioligand Binding Assays
Binding affinities of the TSPO-ligand MTX conjugates 3 and 4 for TSPO and central-type benzodiazepine receptor (CBR) were assessed on membrane preparations from rat cerebral cortex. The IC 50 values were determined from the displacement curves using the reference compound [ 3 H]-PK 11195, for TSPO, or [ 3 H]flunitrazepam for CBR. The data obtained, expressed as IC 50 values, as well as the ratios between CBR and TSPO affinity, respectively (Selectivity Index, SI), as a measure of the in vitro selectivity, are shown in Table 2 and were compared with unlabeled TSPO selective ligand PK 11195, and with CBR selective ligand flunitrazepam. The results of the binding affinities indicate that TSPO-ligand MTX conjugates 3 and 4 show high affinity for TSPO, with IC 50 values ranging from 7.2 to 40.3 nM. Moreover, the SI of TSPO-ligand MTX conjugate 3 is similar to the reference compound PK 11195. Radioligand binding assays on human brain tissue showed a genetic polymorphism and the variation of binding affinity of selective ligands for the TSPO, due to presence of differing binders in the general population (i.e., high-affinity binders, HABs, low-affinity binders, LABs and mixed-affinity binders, MABs) [21]. Thus, the binding quantification of TSPO ligands is confounded by the expression of different forms of TSPO and to assess the potentially therapeutic use of TSPO-ligands would be appropriate to genetically test the patients.

Cytotoxicity Studies
Cytotoxicity assays of the TSPO-ligand MTX conjugates 3 and 4 were conducted against SF126 and SF188 human glioma cells and, RG2 and C6 rat glioma cells and results are also shown in Table 2. These cellular lines were selected for their high expression level of TSPO and, therefore, have been used extensively for in vitro experiments on the brain tumor models and for cytotoxicity studies relating to anticancer drugs, as well as TSPO-ligands and its nanodevices. The tested compounds show high cytotoxicity on all cell lines, especially for C6 rat glioma cells. In particular, the TSPO-ligand MTX α-conjugates 3 appears to be the most cytotoxic displaying a lower IC 50 value compared to MTX on C6 cells.

Transport Studies
Transport studies of the TSPO-ligand MTX conjugates 3 and 4 were carried out on a MDCKII-MDR1 cell monolayer and results are shown in Table 2. This approach is an established and versatile in vitro method to evaluate the drug permeability through the BBB, as well MDCKII-MDR1 was identified as the most promising cell line for qualitative predictions of brain distribution [22]. However He and co-worker have reported the brain microvascular endothelial cells (BMECs) as a more established approach to address the permeability through the BBB [23]. This method was previously described by Audus and has been also used in our recent work [24,25]. The apparent permeability coefficient (P app ) of each compound was calculated on the basis of the amounts permeated through MDCKII-MDR1 monolayer by the equation: where V A (dC/dt) is the linear appearance rate of mass in the receiver solution; A is the filter/cell surface area and [D] in is the initial tested compound concentration in the apical (AP) chamber. The P app values were evaluated in apical to basolateral direction. Moreover, the flux of fluorescein isothiocyanate-dextran (FD4, Sigma, Milano, Italy) (200 µg/mL) and diazepam (75 µM) as paracellular and transcellular markers, respectively, was measured as an internal control to verify the cell monolayer's integrity and tight junction integrity during the assay. On an experimental day, the integrity of the cell monolayer was assessed by measuring the transendothelial electrical resistance (TEER) value. The TEER of the BBB was assessed close to 200 Ω/cm 2 , precluding the access to a wide number of common anticancer drugs to brain tissue, including MTX. The results of the transport studies demonstrate that both TSPO-ligand MTX conjugates 3 and 4 were characterized by higher P app values than MTX and paracellular marker FD4 (3.54˘(0.39ˆ10´1 0 )), with a P app values being about. 30 times greater than the parent drug. Moreover, the permeability of MTX-conjugates was found to be lower than that of the transcellular marker diazepam (6.40˘(0.31ˆ10´5)). Finally, these results demonstrate that although the MTX-conjugates possessed an adequate calculated lipophilic character, which result in a better permeation capacity, compared to the MTX, the P app values were predictive of a modest ability to overcome the BBB by the transcellular pathway. However, due to the development of brain tumors, the structure of tumor vasculature changes and the BBB is gradually replaced by the more permeable blood-brain tumor barrier (BBTB) [26].

High-Performance Liquid Chromatography (HPLC) Analyses
HPLC analyses were performed with an Agilent 1260 Infinity Quaternary LC System equipped with an Agilent variable wavelength UV detector, a Rheodyne injector (Rheodyne, Model 7725i, Agilent) equipped with a 20 µL loop and a OpenLAB CDS ChemStation software (Agilent). A reversed phase Symmetry C18 column (25 cmˆ3.9 mm; 5 µm particles) as the stationary phase and a mixture of methanol and deionized water 80:20 (v/v) as the mobile phase with the flow rate of 1.0 mL/min were utilized and the column effluent was monitored continuously at 254 nm. The compounds were quantified by measuring the areas of the peaks, and using, as references, suitable standard solutions, chromatographed under the same conditions. The data were processed using Microsoft Excel 2010 or GraphPad Software (La Jolla, CA, USA).

Stability Studies in Rat Serum Solution
The TSPO ligand-MTX conjugates 3 and 4 stabilities in physiological medium were studied at 37˝C in 0.05 M phosphate buffer and 0.14 M NaCl at pH 7.4, containing 50% v/v of rat serum. The in vitro stabilities were assayed by adding 100 µL of the stock solution of conjugates in DMSO (1 mg/mL) to 1.1 mL of preheated serum solution, and the mixture was maintained in water bath at 37˝C (˘0.2). Thus, the final concentration of the compounds in the tested solutions was 1 M. At different time points over a period of 0-120 min, aliquots of 100 µL of seach sample were withdrawn and added to 500 µL of cold acetonitrile for deproteinize the serum. Samples were centrifuged at 3500 rpm for 10 min, and the supernatant was removed, filtered through cellulose acetate membranes (0.22 µm, Advantec MFS) and then 20 µL of filtrates were immediately analyzed by HPLC. The studies were done in triplicate and the pseudo-first-order rate constants were obtained from the slopes of the linear plot of the logarithms of the residual concentrations of conjugates against time.

Synthesis of TSPO-Ligand-MTX Conjugates Prodrugs
To a solution of MTX (0.25 g, 0.55 mmol) in anhydrous DMF (2.5 mL) cooled at 0˝C in ice bath was added CDI (0.10 g, 0.68 mmol), and the reaction mixture stirred for 30 min. After this time, a solution of compound 2 (0.306 g, 0.55 mmol) in anhydrous DMF (2.5 mL) was added dropwise, and stirring prolonged for 12 h at room temperature in the dark. Solvent was evaporated under reduced pressure, and the residue dissolved in CH 2 Cl 2 (20 mL). The resulting solution was washed with 5% NaHCO 3 , dried over Na 2 SO 4 , and rotovapored to dryness. Purification of the residue by silica-gel column chromatography using 15% MeOH/CH 2 Cl 2 as the eluting solvent, afforded regioisomers 3 (0.34 g, 70% yield) and 4 (0.1 g,~25% yield) as orange powders. Cell monolayers with TEER 120-140 Ohm/cm 2 were used. In each experiment the cells were equilibrated for 30 min at 37˝C in transport medium which include 0.4 mM K 2 HPO 4 , 25 mM NaHCO 3 , 3 mM KCl, 122 mM NaCl, 10 mM glucose, and the pH was 7.4, with osmolarity of 300 mOsm as determined by a freeze point-based osmometer. For apical-to-basal permeability (AP-BL), test and control compounds solutions were prepared in transport medium at concentration of 75 µM (or 200 µg/mL for FD4), and added to the apical side of the cell monolayer (0.5 mL). Fresh assay medium was placed in the receiver compartment. The transport experiments were carried out in incubator at 37˝C, 5% CO 2 , and 95% humidity. After incubation time of 2 h, samples were removed from the apical side of the monolayer and then analyzed. MTX and TSPO ligand-MTX conjugates 3 and 4 were analyzed by HPLC as described above in the stability studies, while diazepam was analyzed with a PerkinElmer double-beam UV-visible spectromphotometer Lamba Bio 20 (Milan, Italy), equipped with 10 mm path-length-matched quartz cells. The FD4 samples were analyzed with a Victor3 fluorometer (Wallac Victor3, 1420 Multilabel Counter, Perkin-Elmer) at excitation and emission wavelengths of 485 and 535 nm, respectively. All experiments were repeated three times and each compound was tested in triplicate.

Statistical Analysis
The statistical analysis was accomplished using one-way analysis of variance (ANOVA) followed by the Tukey post hoc tests (GraphPad Prism version 5.04 for Windows, GraphPad Software, San Diego, CA, USA). Differences were considered statistically significant at p < 0.05.

Conclusions
TSPO ligand-MTX conjugates have shown to possess a high binding affinity and selectivity for TSPO, and exhibited marked toxicity against glioma cells, in comparison with the parent drug MTX. These results also highlight the ability of the TSPO-ligand to transport the hydrophilic drug through biological membranes and determine its accumulation in target cells overexpressing the TSPO. The present work describes the proof-of-concept demonstration of the strength of the bio-conjugate strategy that simultaneously carries inside of cancer cells two agents with distinct modes of action, in the treatment of brain tumors. For this reason, the TSPO ligand-MTX conjugates could be potential tools to increase the effectiveness of the drug in the treatment of brain tumors overexpressing the mitochondrial target TSPO.