Benzothiazole Derivatives Endowed with Antiproliferative Activity in Paraganglioma and Pancreatic Cancer Cells: Structure–Activity Relationship Studies and Target Prediction Analysis

The antiproliferative effects played by benzothiazoles in different cancers have aroused the interest for these molecules as promising antitumor agents. In this work, a library of phenylacetamide derivatives containing the benzothiazole nucleus was synthesized and compounds were tested for their antiproliferative activity in paraganglioma and pancreatic cancer cell lines. The novel synthesized compounds induced a marked viability reduction at low micromolar concentrations both in paraganglioma and pancreatic cancer cells. Derivative 4l showed a greater antiproliferative effect and higher selectivity index against cancer cells, as compared to other compounds. Notably, combinations of derivative 4l with gemcitabine at low concentrations induced enhanced and synergistic effects on pancreatic cancer cell viability, thus supporting the relevance of compound 4l in the perspective of clinical translation. A target prediction analysis was also carried out on 4l by using multiple computational tools, identifying cannabinoid receptors and sentrin-specific proteases as putative targets contributing to the observed antiproliferative activity.


Introduction
Heterocycles represent precious scaffolds in natural and synthetic molecules, endowed with a great variety of biological activities. Benzothiazoles are members of the bicyclic heteroaromatic family, and they are widely used in medicinal chemistry as the scaffolds of several drugs, including antimicrobial, anti-inflammatory, anticonvulsant, neuroprotective and many others [1,2]. The easy functionalization of the aromatic ring of benzothiazole and the 2-amino or 2-mercapto substituents, which are frequently used as building blocks, makes them attractive and reactive components that are useful in organic and medicinal chemistry programs [3,4]. The development of benzothiazole-based drugs has led to a number of derivatives currently marketed to treat different pathologies, such as the neuroprotective agent riluzole, the diuretic ethoxzolamide, the antidiabetic zopolrestat, the immunosuppressant frentizole, or the diagnostic tool thioflavin T ( Figure 1A). zopolrestat, the immunosuppressant frentizole, or the diagnostic tool thioflavin T ( Figure  1A). The potent antiproliferative effects exerted by benzothiazoles in different cancer models has elicited interest in these molecules as promising antitumor agents ( Figure 1B) [5,6]. Considering that 2-phenylbenzothiazoles have been recognized as highly potent cytotoxic compounds, several derivatives were identified and tested in different cancer cell lines. In particular, the chemical manipulation of the lead compound DF203 [7], a 2phenylbenzothiazole derivative, led to the development of the clinic candidate prodrug Phortress [8]. More recently, the derivative BLZ945, acting as a CSF-1R kinase inhibitor [9,10], has been identified, and it is currently under evaluation as single agent and in combination with spartalizumab for the treatment of advanced solid tumors in adults (ClinicalTrials.gov, NCT02829723) [11]. The 2-ureidobenzothiazole derivative KST016366 has been reported as a potent multikinase inhibitor, displaying a broad-spectrum antiproliferative activity against a wide panel of cancer cell lines [12].
In previous studies, our research group synthesized 2-mercaptobenzothiazole derivatives as Peroxisome Proliferator-Activated Receptor Alpha (PPARα) antagonists [13] (Figure 2). These compounds were able to antagonize PPARα at low micromolar concentrations, and displayed also interesting antiproliferative effects when tested in paraganglioma, glioblastoma, colorectal and pancreatic cancer cell models [14,15]. Interestingly, the benzothiazole derivative 2b showed a marked cytotoxic effect, mainly in paraganglioma cells, with a dose-dependent inhibition profile and a potency The potent antiproliferative effects exerted by benzothiazoles in different cancer models has elicited interest in these molecules as promising antitumor agents ( Figure 1B) [5,6]. Considering that 2-phenylbenzothiazoles have been recognized as highly potent cytotoxic compounds, several derivatives were identified and tested in different cancer cell lines. In particular, the chemical manipulation of the lead compound DF203 [7], a 2-phenylbenzothiazole derivative, led to the development of the clinic candidate prodrug Phortress [8]. More recently, the derivative BLZ945, acting as a CSF-1R kinase inhibitor [9,10], has been identified, and it is currently under evaluation as single agent and in combination with spartalizumab for the treatment of advanced solid tumors in adults (ClinicalTrials.gov, NCT02829723) [11]. The 2-ureidobenzothiazole derivative KST016366 has been reported as a potent multikinase inhibitor, displaying a broad-spectrum antiproliferative activity against a wide panel of cancer cell lines [12].
In previous studies, our research group synthesized 2-mercaptobenzothiazole derivatives as Peroxisome Proliferator-Activated Receptor Alpha (PPARα) antagonists [13] (Figure 2). These compounds were able to antagonize PPARα at low micromolar concentrations, and displayed also interesting antiproliferative effects when tested in paraganglioma, glioblastoma, colorectal and pancreatic cancer cell models [14,15]. Interestingly, the benzothiazole derivative 2b showed a marked cytotoxic effect, mainly in paraganglioma cells, with a dose-dependent inhibition profile and a potency comparable to that of the commercially available PPARα antagonist GW6471 [16,17]. However, it cannot be excluded that other mechanisms of action, or molecular targets, in addition to PPARα inhibition, may emerge for explaining the antiproliferative activity of this compound. comparable to that of the commercially available PPARα antagonist GW6471 [16,17]. However, it cannot be excluded that other mechanisms of action, or molecular targets, in addition to PPARα inhibition, may emerge for explaining the antiproliferative activity of this compound.  To further explore the potential of this class of molecules as antiproliferative agents, we synthetized novel derivatives of the lead compound 2b by keeping unaltered the benzothiazole scaffold and the amide functional group, and by introducing substituents on the distal aromatic ring. In particular, substituents with different electronic properties and steric hindrance were selected to obtain a series of para-substituted (4a-f), metasubstituted (4g-l) and disubstituted analogs that bear electron-withdrawing substituents (4m-p). These modifications were performed to test how this molecular portion could modulate the cytotoxic activity in different cancer cell lines.
In this regard, the synthesized compounds 4a-p were tested for their antiproliferative activity in three distinct pancreatic cancer cell lines (AsPC-1, Capan-2, BxPC-3) and two paraganglioma cell lines (PTJ64i, PTJ86i) that have been established by Prof. Cama's research group at the University of Chieti [17]. Compounds of the series that consistently affected cell viability across the tested cancer cell lines more potently than the lead compound 2b were further analyzed against normal HFF-1 fibroblast cells, to evaluate their toxicity. In the perspective of clinical translation, we also tested whether the most potent and less toxic derivatives could be usefully combined with already approved drugs by analyzing the effects of combined treatments on cancer and normal cell viability.
A target prediction study was also performed to shed light on the putative mechanisms of action contributing to the antiproliferative effects of novel compounds.

Chemistry
The synthesis of the planned compounds 4a-p was carried out as depicted in Scheme 1. Briefly, the commercially available ethyl mandelate was treated with mesyl chloride and triethylamine in THF at 0 °C. The resulting mesylate 1 was reacted with 5-chloro-2mercaptobenzothiazole and triethylamine in THF at 50 °C for 24 h; the ester 2 was hydrolyzed in a basic medium to afford acid 3. The direct coupling of 3 with the proper amines HOBt, DCC, and N-methylmorpholine in DMF led to the amides 4a-p. Crude products were purified by column chromatography or crystallization, obtaining desired amides in good purity and discrete yields. To further explore the potential of this class of molecules as antiproliferative agents, we synthetized novel derivatives of the lead compound 2b by keeping unaltered the benzothiazole scaffold and the amide functional group, and by introducing substituents on the distal aromatic ring. In particular, substituents with different electronic properties and steric hindrance were selected to obtain a series of para-substituted (4a-f), meta-substituted (4g-l) and disubstituted analogs that bear electron-withdrawing substituents (4m-p). These modifications were performed to test how this molecular portion could modulate the cytotoxic activity in different cancer cell lines.
In this regard, the synthesized compounds 4a-p were tested for their antiproliferative activity in three distinct pancreatic cancer cell lines (AsPC-1, Capan-2, BxPC-3) and two paraganglioma cell lines (PTJ64i, PTJ86i) that have been established by Prof. Cama's research group at the University of Chieti [17]. Compounds of the series that consistently affected cell viability across the tested cancer cell lines more potently than the lead compound 2b were further analyzed against normal HFF-1 fibroblast cells, to evaluate their toxicity. In the perspective of clinical translation, we also tested whether the most potent and less toxic derivatives could be usefully combined with already approved drugs by analyzing the effects of combined treatments on cancer and normal cell viability.
A target prediction study was also performed to shed light on the putative mechanisms of action contributing to the antiproliferative effects of novel compounds.

Chemistry
The synthesis of the planned compounds 4a-p was carried out as depicted in Scheme 1. Briefly, the commercially available ethyl mandelate was treated with mesyl chloride and triethylamine in THF at 0 • C. The resulting mesylate 1 was reacted with 5-chloro-2-mercaptobenzothiazole and triethylamine in THF at 50 • C for 24 h; the ester 2 was hydrolyzed in a basic medium to afford acid 3. The direct coupling of 3 with the proper amines HOBt, DCC, and N-methylmorpholine in DMF led to the amides 4a-p. Crude products were purified by column chromatography or crystallization, obtaining desired amides in good purity and discrete yields. Final compounds, including para-substituted 4a-f, meta-substituted 4g-l and a group of m-p disubstituted analogues 4m-p bearing electron-withdrawing substituents, are reported in Table 1. For each compound, the method of purification, the yield, and the melting point are also specified.

Antiproliferative Activity
We analyzed by MTT assays the effects of synthesized compounds 4a-p on the viability of three pancreatic cancer cell lines (AsPC-1, Capan-2, BxPC-3) and two paraganglioma cell models (PTJ64i, PTJ86i), based on relevant antiproliferative effects previously shown by the lead compound 2b in the same cells [16].
Compounds 4a-p were submitted to a preliminary MTT assay at a one-point screening concentration of 75 μM for 72 h, including 2b as a reference compound ( Figure  3). Overall, the majority of the tested compounds affected cancer cell viability, with a potency comparable, or superior to, 2b with the exception of 4g (m-methoxy) and the 2bromo-substituted derivatives 4n (2-Br, 5-NO2), 4o (2-Br, 4-CF3), and 4p (2-Br, 5-CF3). The presence of the electron-donating methoxy group decreased the activity of 2b, in both the Final compounds, including para-substituted 4a-f, meta-substituted 4g-l and a group of m-p disubstituted analogues 4m-p bearing electron-withdrawing substituents, are reported in Table 1. For each compound, the method of purification, the yield, and the melting point are also specified.

Antiproliferative Activity
We analyzed by MTT assays the effects of synthesized compounds 4a-p on the viability of three pancreatic cancer cell lines (AsPC-1, Capan-2, BxPC-3) and two paraganglioma cell models (PTJ64i, PTJ86i), based on relevant antiproliferative effects previously shown by the lead compound 2b in the same cells [16].
Compounds 4a-p were submitted to a preliminary MTT assay at a one-point screening concentration of 75 µM for 72 h, including 2b as a reference compound ( Figure 3). Overall, the majority of the tested compounds affected cancer cell viability, with a potency compara- ble, or superior to, 2b with the exception of 4g (m-methoxy) and the 2-bromo-substituted derivatives 4n (2-Br, 5-NO 2 ), 4o (2-Br, 4-CF 3 ), and 4p (2-Br, 5-CF 3 ). The presence of the electron-donating methoxy group decreased the activity of 2b, in both the para and meta positions. Conversely, the introduction of electron-withdrawing substituents, such as halogens, nitro, trifluoromethyl and acetylamino groups, improved the antiproliferative activity of the lead compound 2b, and this effect was observed for meta and para positions, including the m,p-dichloro derivative 4m. This general trend was observed for 4a-p in all the selected cancer cell lines. Starting from these data, we selected nine compounds displaying a greater antiproliferative activity than 2b to perform concentration-response curves. Specifically, these compounds, namely 4d, 4e, 4f, 4h, 4i, 4j, 4k, 4l, and 4m, were tested at concentrations of 3, 6, 12 and 24 µM for 72 h. In addition, the nine compounds, together with 2b, were tested against human fibroblasts HFF-1, in order to evaluate their selectivity against tumor cells, as compared to normal cells. Effects on cancer or normal cell viability were extrapolated from concentration-response curves and are expressed as IC 50 values in Table 2. para and meta positions. Conversely, the introduction of electron-withdrawing substituents, such as halogens, nitro, trifluoromethyl and acetylamino groups, improved the antiproliferative activity of the lead compound 2b, and this effect was observed for meta and para positions, including the m,p-dichloro derivative 4m. This general trend was observed for 4a-p in all the selected cancer cell lines. Starting from these data, we selected nine compounds displaying a greater antiproliferative activity than 2b to perform concentration-response curves. Specifically, these compounds, namely 4d, 4e, 4f, 4h, 4i, 4j, 4k, 4l, and 4m, were tested at concentrations of 3, 6, 12 and 24 μM for 72 h. In addition, the nine compounds, together with 2b, were tested against human fibroblasts HFF-1, in order to evaluate their selectivity against tumor cells, as compared to normal cells. Effects on cancer or normal cell viability were extrapolated from concentration-response curves and are expressed as IC50 values in Table 2.   Notably, the compounds 4k and 4l displayed the greatest and most consistent selectivity index (SI) values across the tested cancer cell lines (Table 3 and Table S1), which supported their potential as effective and safe anticancer agents in pancreatic cancer and paraganglioma treatment.
Considering that gemcitabine is one of the first-line therapies in pancreatic cancer, to deepen the potential clinical relevance of compounds 4k and 4l we further analyzed the effects of the combinations between each of them and gemcitabine on pancreatic cancer cell line viability (Figure 4 and Figure S1, and Table 4).  Notably, the compounds 4k and 4l displayed the greatest and most consistent selectivity index (SI) values across the tested cancer cell lines (Tables 3 and S1), which supported their potential as effective and safe anticancer agents in pancreatic cancer and paraganglioma treatment. Considering that gemcitabine is one of the first-line therapies in pancreatic cancer, to deepen the potential clinical relevance of compounds 4k and 4l we further analyzed the effects of the combinations between each of them and gemcitabine on pancreatic cancer cell line viability (Figures 4 and S1, and Table 4).   Remarkably, the combination between the lowest concentrations of gemcitabine and 4l (0.1 and 0.5 µM, respectively) decreased pancreatic cancer cell viability in a more marked and synergistic manner (CIs < 1) across the three pancreatic cancer cell lines, as compared to combinations with higher concentrations of the two agents ( Figure 4 and Table 4). Notably, the combination with the lowest concentrations of the two compounds appeared safer in normal fibroblast HFF-1 cells, as compared with the pancreatic cancer cell lines (inhibition rate 37% in HFF-1, as compared with 63% in AsPC-1, 66% in BxPC-3 and 48% in Capan-2), supporting the relevance of compound 4l in the perspective of clinical translation (Table 4).

Target Prediction Studies
In principle, there could be multiple putative molecular targets for explaining the cytotoxic activity of derivatives 4a-p, and their identification may contribute to a more comprehensive understanding of their bioactivity. Since the high-throughput in vivo target profiling of compounds to identify a potential binding protein for a specific molecule could be expensive and time-consuming, an in silico target fishing study was applied to identify novel proteins possibly involved in the network of molecular events underlying the cytotoxic activity against the cancer cell lines. In silico target fishing (also known as target prediction or target identification) is emerging as an efficient alternative to predict the macromolecular target speedily [18]. A primary distinction divides these methods into ligand-based, i.e., comparing the characteristics (fingerprints) of the query molecule with those present in the databases, and structure-based, i.e., comparing the putative ligand with the features of the target protein's active site [19]. Various tools, freely accessible or not, have been designed to this aim.
In the present work, all query molecules were analyzed using different web-based tools including ligand-based methods such as SEA SEARCH, PLATO, PPB2, and SuperPred, and structure-based tools such as PharmMapper ( Table 5). The resulting targets were filtered, selecting only those expressed in the human organism. Furthermore, as most tools retrieve many possible targets, only the highest-ranked (more reliable) were selected and analyzed. Results for the selected derivative 4l are shown in Table 6, and the prediction results for other analogue compounds are very similar to those presented.
By analyzing the obtained results, to reduce the risk of false-positive prediction [30] we focused on the target protein classes more frequently retrieved by multiple tools, namely cannabinoid receptors (CBR1, CBR2, and G-protein coupled receptor 55-GPCR55-an orphan GPCR binding cannabinoid), which were predicted by four over six programs, as well as sentrin-specific proteases (SENP6, SENP7, SENP8), which were suggested by three over six programs (Table 6). To further assess the significance of our prediction, the potential role of these two targets in pancreatic cancer was explored by a literature research.  [31]. It is worth noting that the cannabinoid system is a well-known player in cancer biology [32]. The overexpression of both CB1 and CB2 receptors on pancreatic cancer cells, as compared to a very limited expression in healthy pancreatic cells, was revealed by a study in patients [33]. In addition, antagonists to these receptors seem to be extremely promising as antitumor agents, since they are selective in interacting with pancreatic tumor cells, as compared to healthy cells [34].
Sentrin-specific proteases, also known as SUMO-specific proteases, are cysteine proteases responsible for the DeSUMOylation of target proteins, an essential post-translational modification process [35]. To date, seven SENP isoforms have been identified, showing different localization and substrate preference [36]. Considering that SENPs regulate proteins involved in DNA repair, cell cycle, and neovascularization, the deregulation (overexpression or downregulation) of SENP activity results in cellular dysfunction associated with the development of different diseases, including prostate, thyroid, colon, lung, and pancreatic cancer [37]. For instance, the upregulation of SENP3 is a prognostic marker of pancreatic cancer according to the Cancer Genome Atlas (TCGA) [38]. For these reasons, in recent years SENP proteases have emerged as potential targets for cancer therapy [39]; the high sequence homology of members of human SENPs, along with the differences in substrate specificity and subcellular localization might be an interesting way to discover selective SENP inhibitors [40].

Docking Studies
To further validate our predictions, docking calculations were carried out on two representative target proteins. Ligands were flexibly docked in the binding site of the Cannabinoid CB1 receptor (CBR1) using the crystallographic structure of CBR1 retrieved by the Protein Data Bank (PDB: 6KPG) [41].
Docked poses in the CBR1 binding site are pretty well conserved among studied compounds: Figure 5a reports the binding mode of compound 4l highlighting the optimal fitting of the ligand in the CB1 binding pocket. The reported ligand establishes interactions with Phe108, Val110, Leu111, Asn112, Phe170, Phe174, His178, Phe189, Val196, Thr197, Ile267, Phe268, Leu359, Met363, Lys376, Phe379, Ala380, and Ser383 (Figure 5b). These contacts are common to the other ligands as demonstrated by the interaction diagram reported in the Supplementary Materials (Supplementary Figure S2).
identified, showing different localization and substrate preference [36]. Considering that SENPs regulate proteins involved in DNA repair, cell cycle, and neovascularization, the deregulation (overexpression or downregulation) of SENP activity results in cellular dysfunction associated with the development of different diseases, including prostate, thyroid, colon, lung, and pancreatic cancer [37]. For instance, the upregulation of SENP3 is a prognostic marker of pancreatic cancer according to the Cancer Genome Atlas (TCGA) [38]. For these reasons, in recent years SENP proteases have emerged as potential targets for cancer therapy [39]; the high sequence homology of members of human SENPs, along with the differences in substrate specificity and subcellular localization might be an interesting way to discover selective SENP inhibitors [40].

Docking Studies
To further validate our predictions, docking calculations were carried out on two representative target proteins. Ligands were flexibly docked in the binding site of the Cannabinoid CB1 receptor (CBR1) using the crystallographic structure of CBR1 retrieved by the Protein Data Bank (PDB: 6KPG) [41].
Docked poses in the CBR1 binding site are pretty well conserved among studied compounds: Figure 5a reports the binding mode of compound 4l highlighting the optimal fitting of the ligand in the CB1 binding pocket. The reported ligand establishes interactions with Phe108, Val110, Leu111, Asn112, Phe170, Phe174, His178, Phe189, Val196, Thr197, Ile267, Phe268, Leu359, Met363, Lys376, Phe379, Ala380, and Ser383 (Figure 5b). These contacts are common to the other ligands as demonstrated by the interaction diagram reported in the Supplementary Materials (Supplementary Figure S2). Since the crystallographic structure of SENP6 and SENP7 is not available, the chimeric form of SENP2 containing the SENP6 characteristic loop 1 (similar to SENP7) was used for the structure-based studies (PDB: 3ZO5) [42]. To explore the possible binding sites of the protein, the Sitemap tool was used. Among the identified sites, the one that was superimposable to the most studied SENPs was chosen for the grid generation. This binding site possesses a characteristic form that is able to host all studied compounds. The tridentate structure of the compounds fits well with the structure of the binding site. In Figure 6a, the binding mode of 4l is shown by the surface representation of the protein, while in Figure 6b, the residues that make contacts with the ligand are reported. In detail, these residues are Leu401, Asn 423, Asn427, Val430, Lys434, Leu441, His442, Val443, Phe444, Ser445, Thr464, Val467, and Gln472. The interaction diagram of all compounds and residues is reported in Supplementary Figure S3.
Although the identified protein classes seem to be promising targets of our benzothiazole derivatives, further in vitro and in vivo studies will be necessary to validate our prediction. Since the crystallographic structure of SENP6 and SENP7 is not available, the chimeric form of SENP2 containing the SENP6 characteristic loop 1 (similar to SENP7) was used for the structure-based studies (PDB: 3ZO5) [42]. To explore the possible binding sites of the protein, the Sitemap tool was used. Among the identified sites, the one that was superimposable to the most studied SENPs was chosen for the grid generation. This binding site possesses a characteristic form that is able to host all studied compounds. The tridentate structure of the compounds fits well with the structure of the binding site. In Figure 6a, the binding mode of 4l is shown by the surface representation of the protein, while in Figure 6b, the residues that make contacts with the ligand are reported. In detail, these residues are Leu401, Asn 423, Asn427, Val430, Lys434, Leu441, His442, Val443, Phe444, Ser445, Thr464, Val467, and Gln472. The interaction diagram of all compounds and residues is reported in Supplementary Figure S3.
Although the identified protein classes seem to be promising targets of our benzothiazole derivatives, further in vitro and in vivo studies will be necessary to validate our prediction.

Physicochemical and Pharmacokinetic Properties Calculation
QikProp [43] calculations were carried out to predict the parameters affecting the drug-likeness and bioavailability of the studied benzothiazoles ( Table 7). The calculated physicochemical and pharmacokinetic properties denote that the studied compounds have a drug-like profile, presenting at most one violation of the rule of five due to their lipophilicity (logPoct/water > 5). The predicted low solubility and possible HERG

Physicochemical and Pharmacokinetic Properties Calculation
QikProp [43] calculations were carried out to predict the parameters affecting the drug-likeness and bioavailability of the studied benzothiazoles ( Table 7). The calculated physicochemical and pharmacokinetic properties denote that the studied compounds have a drug-like profile, presenting at most one violation of the rule of five due to their lipophilicity (logPoct/water > 5). The predicted low solubility and possible HERG inhibition represent limiting aspects that need further optimization. On the other hand, our compounds present excellent oral absorption and remarkable cell permeability. Moreover, some of the analyzed compounds show good CNS activity and a promising brain/blood partition coefficient. QPlogPo/w: Predicted octanol/water partition coefficient (−2.0-6.5); RuleOfFive: Number of violations of Lipinski's rule of five; PSA: Van der Waals surface area of polar nitrogen and oxygen atoms and carbonyl carbon atoms ; #rotor: Number of non-trivial (not CX3), non-hindered (not alkene, amide, small ring) rotatable bonds. (0-15); CIQPlogS: Conformation Independent predicted aqueous solubility, log S. S in mol dm −3 is the concentration of the solute in a saturated solution (−6.5-0.5); PercentHumanOralAbsorption: Predicted human oral absorption on a 0 to 100% scale; QPPCaco: Predicted apparent Caco-2 cell permeability in nm/sec. Caco-2 cells are a model for the gut-blood barrier (<25 poor, >500 great); QPPMDCK: Predicted apparent MDCK cell permeability in nm/sec. MDCK cells are considered to be a good mimic for the blood-brain barrier (<25 poor, >500 great); QPlogBB: Predicted brain/blood partition coefficient (−3.0-1.2); QPlogHERG: Predicted IC50 value for the blockage of HERG K+ channels (concern below −5); QPlogKhsa: Prediction of binding to human serum albumin (−1.5-1.5); CNS: Predicted central nervous system activity on a −2 (inactive) to +2 (active) scale; #metab: Number of likely metabolic reactions.

Chemistry
Melting points were determined with a Buchi Melting Point B-450 and were uncorrected. NMR spectra were recorded on a Varian Mercury 300 spectrometer with 1 H at 300.060 MHz and 13 C at 75.475 MHz. Proton chemical shifts were referenced to the TMS internal standard. Chemical shifts are reported in parts per million (ppm, δ units). Coupling constants are reported in units of Hertz (Hz). Splitting patterns are designed as s, singlet; d, doublet; t, triplet; q, quartet; dd, double doublet; m, multiplet; b, broad. Elemental analyses were carried out on a PerkinElmer 240B micro-analyzer, obtaining results within ± 0.4 % of the theoretical values. The purity of all compounds was over 98%. All commercial chemicals and solvents were reagent grade and were obtained from Merck; they were used without further purification, unless otherwise specified. Chemical reactions were monitored by thin layer chromatography on silica gel plates (60F-254, Sigma Aldrich, Italy) and the analysis of the plates was carried out using a UV lamp 254/365 nm. Flash chromatography was performed on silica gel 60 (Merck).

Synthesis of Ethyl 2-[(Methylsulfonyl)oxy]-2-phenylacetate 1
Triethylamine (834 µL, 6 mmol) was added to a solution of ethyl mandelate (162 mg, 1 mmol) in THF (8 mL) at 0 • C under stirring. Then methanesulfonyl chloride (232 µL, 3 mmol) was added dropwise, and the reaction was allowed to stir for two hours at room temperature. A saturated solution of ammonium chloride was added to the mixture, that was extracted with dichloromethane. Combined organic phases were washed with a saturated sodium chloride solution, dried on sodium sulfate, and the solvent was evaporated under reduced pressure. The crude material was purified by column chromatography on silica gel, with petroleum ether/ethyl acetate 8:2 as eluent. White solid, 93% yield, m.p. 60-61 • C; 1  Triethylamine (278 µL, 2 mmol) was added to a solution of 5-cloro-2-mercaptobenzothiazole (202 mg, 1 mmol) in THF (10 mL) at 0 • C under stirring. After 30 min mesylate 1 (258 mg, 1 mmol) was added dropwise, and the temperature was allowed to reach room temperature. The mixture was then heated at 50 • C and stirred for 48 h. After the removal of THF, the resulting oil was dissolved in distilled water and extracted with dichloromethane. The organic phase was washed with a saturated solution of sodium chloride, dried on sodium sulfate, and evaporated under reduced pressure. The crude product was purified by column chromatography on silica gel, with petroleum ether/ethyl acetate 8:1 as eluent. Pale yellow solid, 87% yield; 1  NaOH 2N (4 mL, 8 mmol) was added to ester 2 (363 mg, 1 mmol) in THF (5 mL) and the solution was stirred at room temperature overnight. THF was removed under reduced pressure, and the aqueous phase was acidified by HCl 2N, obtaining a precipitate that was collected by filtration under vacuum and recrystallized from cyclohexane. White crystals, 99% yield, m.p. 186-188 • C; 1

General Procedure for the Synthesis of Amides 4a-p
To a stirred solution of acid 3 (168 mg, 0.5 mmol) in DMF (5 mL) at 0 • C, N,N'dicyclohexylcarbodiimide (DCC, 103 mg, 0.5 mmol) and hydroxybenzotriazole (HOBt, 77 mg, 0.5 mmol) were added. After 15 min, N-methylmorpholine (NMM, 55 µL, 0.5 mmol) and the selected amine (0.5 mmol) were added in sequence to the reaction mixture. The resulting solution was then stirred at room temperature for 24 h, concentrated and the residue was dissolved in dichloromethane, and washed with a saturated solution of sodium bicarbonate and brine. The combined organic layers were dried on sodium sulfate and concentrated under vacuum to provide the crude products, which were purified by column chromatography or crystallization.

Cell Lines, Treatments, and Cell Viability Assay
Pancreatic cancer (AsPC-1, BxPC-3, and Capan-2), paraganglioma (PTJ86i and PTJ64i) and normal fibroblast (HFF-1) cell lines were cultured as previously described [17]. All compounds were dissolved in DMSO (stock solutions) and then diluted in culture media to the final working concentrations. In this way, the solutions were completely clear and devoid of any undissolved material by microscopic inspection. The final concentration of DMSO in the experiments was at most 0.18% and showed no cell toxicity. The effects of compounds 4a-p and 2b on cancer cell viability were tested by an MTT assay (Sigma-Aldrich, St. Louis, MO, USA), as described by Florio et al. [17]. In the initial screening all compounds were tested at a one-point screening concentration of 75 µM for 72 h (five replica wells per each condition). Then, concentration-response curves were generated by incubating cancer or normal fibroblast cell lines for 72 h with compounds showing a greater antiproliferative activity than 2b, at concentrations ranging from 0 µ M to 24 µM (five replica wells per each condition). For combined treatments, the experimental design was made according to the Chou-Talalay method for drug combination studies, as previously described [44].

Calculation of Half Maximal Inhibitory Concentration (IC 50 ), Selectivity Index (SI), Combination Index (CI) Values and Statistical Analysis
IC 50 values were extrapolated from concentration-response curves and calculated using the CompuSyn software [45]. The interactions between compound 4l and gemcitabine were assessed by calculating the CI values using the CompuSyn software. Based on this analysis, a CI < 1 indicates synergism, a CI = 1 indicates additive effects and a CI > 1 indicates antagonism. SI values were calculated as previously described [46]. Comparisons of mean values were performed using an unpaired Student's t-test. For multiple compar-isons, a one-way ANOVA followed by Dunnett's test were employed. A p-value < 0.05 was estimated as statistically significant.
The structure-based analysis was carried out using the Schrödinger Life-Sciences Suite 2021-4 [43]. The 2D sketcher in Maestro was used to construct ligand structures that were submitted to LigPrep to obtain ligand 3D geometry, identify all potential tautomers, and protonation states at pH 7.0 ± 0.4. The resulting structures were minimized utilizing Macro-Model, the OPLS4 force field, and 5000 steps of the PRCG algorithm with a convergence criterion of 0.05 KJ/mol. Docking calculations were carried out on the 3D coordinates of CBR1 with PDB ID 6KPG and the chimeric form of SENP2 containing the SENP7 loop, PDB ID 3ZO5. Before docking calculations, protein structures were prepared using the Protein Preparation routine in Maestro [47] that fixes the protein structure and relaxes it through a constrained minimization. For the CBR1 the docking protocol was validated by reobtaining the X-ray geometry of the crystallographic ligand (RMSD 0.5135Å). To define the more proper binding site in SENP6, a SiteMap [48,49] calculation was carried out and the subsequent Glide Grid was generated on the predicted site. Glide [50,51] SP flexible docking calculations were carried out on both proteins

Conclusions
In conclusion, a library of phenylacetamide derivatives containing the benzothiazole nucleus was synthesized and tested for its antiproliferative activity in three pancreatic and two paraganglioma cancer cell lines. Most of the derivatives showed an improved antiproliferative activity, as compared to the lead compound 2b, allowing us to trace some preliminary structure-activity relationships. Considering both the antiproliferative activities and selectivity profiles, 4l was further analyzed in combination with gemcitabine on the three pancreatic cancer cell lines. Notably, the combination of the two compounds when used at the lowest concentrations (0.1 µM gemcitabine and 0.5 µM 4l, respectively) affected pancreatic cancer cell viability in a potent and synergistic manner. This synergistic effect supports the interest in this class of benzothiazoles in view of a possible translation into cancer treatment. A computational study was also conducted to predict the putative targets involved in the observed antiproliferative activity in pancreatic cancer cells: the different tools used to this aim identified the cannabinoid receptors and the sentrin-specific proteases as potential targets contributing to the bioactivity of this family of compounds. Docking studies on representative targets (CBR1 and SENP6) show that all compounds fit well in their binding site. Future in vitro and in vivo studies will be necessary to validate the predicted molecular targets of our compounds and gain novel insights into the underexplored potential of these molecules as antitumor agents in paraganglioma and pancreatic cancer treatment.