Haloperidol Metabolite II Valproate Ester (S)-(−)-MRJF22: Preliminary Studies as a Potential Multifunctional Agent Against Uveal Melanoma

Increased angiogenesis and vascular endothelial growth factor (VEGF) levels contribute to higher metastasis and mortality in uveal melanoma (UM), an aggressive malignancy of the eye in adults. (±)-MRJF22, a prodrug of the sigma (σ) ligand haloperidol metabolite II conjugated with the histone deacetylase (HDAC) inhibitor valproic acid, has previously demonstrated a promising antiangiogenic activity. Herein, the asymmetric synthesis of (R)-(+)-MRJF22 and (S)-(−)-MRJF22 was performed to investigate their contribution to (±)-MRJF22 antiangiogenic effects in human retinal endothelial cells (HREC) and to assess their therapeutic potential in primary human uveal melanoma (UM) 92-1 cell line. While both enantiomers displayed almost identical capabilities to reduce cell viability than the racemic mixture, (S)-(−)-MRJF22 exhibited the highest antimigratory effects in endothelial and tumor cells. Given the fundamental contribution of cell motility to cancer progression, (S)-(−)-MRJF22 may represent a promising candidate for novel antimetastatic therapy in patients with UM.


■ INTRODUCTION
Uveal melanoma (UM) is a rare and aggressive intraocular tumor, which arises from melanocytes such as cutaneous melanoma, but presents unique biology and genetic traits. 1 Approximately 90% of all UMs involve the choroid, while the rest involve the ciliary body (6%) or iris (4%). 2 Despite early diagnosis and treatment with conventional chemotherapy and surgery, nearly 50% of all UM patients develop hepatic metastases, which are usually fatal within 1 year from diagnosis. 3−5 Lacking an intraocular lymphatic system, UM tends to spread via a hematogenous route, and the presence of microvascular loops and networks is clinically related to UM progression and a worse prognosis. 6 Recent studies have shown that different UM cell lines produce a copious amount of vascular endothelial growth factor (VEGF), the primary activator of tumor angiogenesis in mammals. 9 Treatment of inoperable UM patients with the VEGF-trap aflibercept (AFL), a popular drug in ophthalmology, resulted in 50% progressionfree survival at 4 months. 7−9 Intravitreal injection of bevacizumab, an antiangiogenic monoclonal antibody targeting all isoforms of vascular endothelial growth factor A (VEGF-A), is currently under evaluation through a phase II trial for the treatment of UM metastatic disease. 10−12 Sigma (σ) receptors are involved in different biological functions, including cell proliferation and survival, and are overexpressed in several tumor cell lines. 13 This unique class of receptors consists of two subtypes, sigma-1 (σ 1 ) and sigma-2 (σ 2 ). The σ 1 receptor is a chaperon protein at the mitochondria-associated membrane (MAM) involved in apoptosis, and its in vivo silencing modulates endothelial cell proliferation and inhibits angiogenesis. 14 The presence of the σ 1 receptor in cancer cells increases VEGF secretion and stimulates motility, in part through the regulation of the human voltage-dependent K + channel (hERG) membrane expression. 15 Recently identified as the ER-resident transmembrane protein 97 (TMEM97), the σ 2 receptor is poorly understood. 16 Recognized as a biomarker of cell proliferation, the mechanism by which the σ 2 receptor promotes apoptosis and autophagy remains unclear. 17,18 Data suggest that σ receptors are able to induce apoptosis and autophagy in UM, whereas they mediate opposite biological effects on cell proliferation. 19,20 Notably, compounds endowed with a σ 1 receptor antagonists/σ 2 receptor agonist functional profile such as haloperidol (HP) and haloperidol metabolite II (HP-mII) have been shown to reduce human UM cell proliferation. Different from HP-mII, which displays a preferential activity for σ receptors compared to other receptor systems, antiproliferation by HP on UM 92-1 cells may be due to additional nonspecific effects. 20,21 Histone deacetylases (HDACs) are enzymes involved in specific epigenetic changes associated with cancer and other diseases. 22 Inhibition of HDACs induces hyperacetylation of histones, which affects gene expression. 23 HDAC inhibitors (HDACis) induce the inhibition of angiogenesis through various mechanisms, such as activation of cell-cycle arrest or induction of apoptosis and autophagy. 22,24,25 More than 20 HDACis have entered clinical studies, with vorinostat and romidepsin approved for the treatment of cutaneous T-cell lymphoma. 26,27 HDACi may play a role in the adjuvant therapy of patients with UM by inducing differentiation and prolonged dormancy of micrometastases. 28 In this context, HDACi valproic acid (VPA) induces G1 cell-cycle arrest of UM cells and reduces UM progression in vivo and has recently undergone a phase II clinical trial for high-risk UM patients. 29−31 To improve antiangiogenic and anticancer capabilities, dual ligands targeting σ receptors and HDAC were previously developed employing a prodrug approach. 32,33 (±)-MRJF22 [(±)-1], a prodrug of (±)-HP-mII with VPA ( Figure 1), significantly reduced cell migration and proliferation (20 and 120 times more than (±)-HP-mII and VPA, respectively) in VEGF-A-stimulated human retinal endothelial cells (HRECs). 34 Angiogenesis within UM progression is the result of a complex interplay between endothelial and tumor cells, in which VEGF-A may play a significant role. 35,36 The involvement of σ receptors and HDAC in antiangiogenic and antiproliferative activities makes the prodrug (±)-1 a potential pharmacological tool exploitable for the treatment of UM metastatic disease. To gain more insights and define contribution to antiangiogenic effects by (±)-1, the asymmetric synthesis of (R)-(+)-MRJF22 [(+)-1] and (S)-(−)-MRJF22 [(−)-1] and their evaluation on VEGF-Astimulated HRECs is described here. Moreover, present studies investigate the effects of (±)-1 and its enantiomers on human UM 92-1 cell proliferation and migration, which together with angiogenesis represents the principal determinants of metastasis development in UM.
Enzymatic stability studies in rat and human plasma were performed for both the racemate and its enantiomers ( Table  1). As we expected, the compounds were rapidly hydrolyzed by carboxylesterases present in rat plasma (t 1/2 , 0.2−0.4 h) compared to human plasma (t 1/2 , >58 h) due to the huge disparity between the esterase content in both plasma. 38 Considering the high stability of (±)-1 and its enantiomers in simulated physiological fluids and human plasma, these compounds could reach the target site without undergoing esterase metabolism.
At 24 h, the enantiomers showed similar extent and efficacy in reducing HREC cell viability compared to the racemic mixture (±)-1 but, contrary to this, reduced the cell viability in a time-dependent manner ( Figure S2). Indeed, the IC 50 ranged from 9.8 to 4.4 μM for (+)-1 and from 10.1 to 6.8 μM for (−)-1, not significantly different from those of (±)-1 (from 10.5 to 10.1 μM). Moreover, confirming previous results with (±)-1, 34 both enantiomers exhibited higher antiproliferative potencies in HREC with respect to precursors VPA and (±)-HP-mII (Table 3). In contrast, HP remains the most potent compound in reducing HREC viability, compared to (±)-1 and its enantiomers (Table 3). 34 The results suggest Table 1. Chemical and Enzymatic Stabilities of (±)-1 and Its Two Enantiomers   that the enantiomers exert a toxicity comparable to that of the racemic mixture on HREC and, in any case, more tolerated by the cells than their precursor HP.
Confirming previous results, 34 PTZ partially counteracted the anti-VEGF-A activity of the (±)-1 racemic mixture ( Figure  2), suggesting that (±)-1 may function, at least in part, as a σ 1 antagonist in the inhibition of VEGF-A-stimulated HREC proliferation. Surprisingly, coincubations with PTZ or AC927 did not alter the anti-VEGF-A effects on proliferation by (+)-1 and (−)-1 (Figure 2), ruling out a relevant role of σ receptor in antiangiogenetic effects by the prodrugs with respect to VEGF-A-mediated HREC proliferation. Presumably, a minimal contribution of σ 1 receptors to racemic mixture effects ( Figure   2) could reflect differences in overall affinities toward σ 1 and σ 2 receptors compared to the single enantiomer components.
Unregulated stimulation of endothelial cell motility by VEGF-A underlies pathological angiogenesis. 41 Therefore, the effects of (±)-1, (+)-1, and (−)-1 on HREC motility stimulated by VEGF-A (80 ng/mL) were evaluated with the wound healing assays (Figures 3 and S3−S5). VEGF-A promoted HREC migration compared to untreated controls, inducing complete wound closures at 24 h ( Figures S3−S5). In contrast, HREC monolayers did not close their wounds over 48 h incubation in the presence of 5 μM (±)-1 or its enantiomers ( Figure 3A−C). Of note, while HREC monolayers did not significantly alter wound changes in unstimulated cells, the three prodrugs completely abrogated VEGF-Amediated induction of HREC migration ( Figure 3A−C). Cotreatments with both AC927 and PTZ reduced (±)-1 and (+)-1 abilities to inhibit VEGF-A-mediated wound healing ( Figure 3A,B). However, while PTZ or AC927 exhibited similar actions with (±)-1 and significantly, but partially, opposed the inhibition of VEGF-A-mediated cell motility by the racemic mixture ( Figure 3A), (+)-1 effects were significantly and completely blocked only by (+)-PTZ coincubation ( Figure 3B). Distinctively, the inhibition of VEGF-A-mediated HREC motility by (−)-1 was prevented, in part, only by cotreatment with the σ 2 receptor antagonist AC927 ( Figure 3C). These data suggest a plausible involvement of both σ 1 and σ 2 receptors in the regulation of VEGF-A-mediated HREC motility by the racemic mixture, reflecting the contribution of σ 1 and σ 2 receptor regulation by enantiomer (+)-1 and enantiomer (−)-1, respectively. Moreover, compared to the other prodrugs, enantiomer (−)-1 produced the highest reduction of HREC motility stimulated by VEGF-A ( Figure 3A−C), below the basal levels of the untreated control ( Figure 3C). Since (−)-1 has the greatest σ 2 binding affinity of the dual-ligand prodrugs (Table 2), these results suggest that σ receptor-mediated regulation of HREC migration by VEGF-A is principally sensitive to σ 2 agonist activity. The representative images of HREC treated with VEGF-A and three prodrug derivatives after the scratch wound are shown in the Supporting Information ( Figures S2−S4A).
To further explore effects on angiogenesis, tube formation assays with HREC were evaluated. Following seeding onto basement membranelike Matrigel and stimulation for 24 h with VEGF-A (80 ng/mL), HREC organized into networks of tubular structures mimicking capillary formation in neoangiogenesis ( Figure 4A). As expected, VEGF-A significantly increased total tube length by HREC, an effect prevented by cotreatment with VEGF-trap AFL ( Figure 4B). Treatments with (±)-1, (+)-1, or (−)-1 also blocked VEGF-A-stimulated tube formation in HREC, with effects similar to those of AFL ( Figure 4B). In close agreement with findings in HREC motility (Figure 3), while (±)-1 actions on VEGF-Astimulated tube formation were partially and equally blocked by both σ 1 receptor agonist PTZ and σ 2 receptor antagonist AC927, (+)-1 and (−)-1 effects were prevented selectively by AC927 and PTZ, respectively ( Figure 4B). Again, these results suggest the simultaneous involvement of σ 1 and σ 2 receptors in the regulation of VEGF-A-induced HREC tube formation by the racemic mixture, reflecting the selective contribution of the σ 1 receptor antagonism by enantiomer (+)-1 and the σ 2 receptor agonism by enantiomer (−)-1.
Overall, (±)-1 and its enantiomers exhibit significant but distinct antiangiogenic effects as assessed with the HREC model in vitro. On the one hand, all three dual-ligand prodrugs induce HREC cytotoxicity and inhibit VEGF-A-mediated HREC proliferation with greater potency than precursor compounds, mostly by inducing molecular mechanisms independent of σ receptor signaling (Table 3 and Figure 2). In contrast, (±)-1, (+)-1, and (−)-1 inhibit VEGF-A-stimulated HREC motility and tubelike structure formation stimulated by VEGF-A through the regulation of σ receptor pathways (Figures 3 and 4). In particular, the actions of enantiomers (+)-1 and (−)-1 appear to be selectively mediated by σ 1 and σ 2 receptor signaling, respectively (Figures 3 and 4). Accordingly, the enantiomer mixture (±)-1 affects VEGF-Amediated HREC migration and tube formation by regulating simultaneously both σ 1 and σ 2 receptors (Figures 3 and 4). Our data highlighted a similar aptitude of (±)-1 and (+)-1 in counteracting the VEGF-A proangiogenic effect through a σ 1 receptor antagonist profile, as indicated by the PTZ capability to mitigate or completely block their effects. However, the (−)-1 enantiomer displays the strongest antiangiogenic activity, probably as a consequence of the superior binding activity as a σ 2 receptor agonist compared to the other dualligand prodrugs ( Table 2). The superior inhibitory ability of (−)-1 against VEGF-A-stimulated HREC motility is of great significance, given the established role of cell migration and invasion in metastatic UM progression. 42 Antiproliferative Activity on Human Uveal Melanoma 92-1 Cells. To evaluate signaling directly on UM cancer cells, the pharmacological effects of (±)-1, (+)-1, and (−)-1 were examined in human UM 92-1 cells. First, the presence of  Quantification of tube length was carried out using the Angiogenesis Analyzer tool for ImageJ software. HREC were treated with 80 ng/mL VEGF-A in the presence or absence of 5 μM (±)-1, (+)-1, and (−)-1 or further supplemented with selective σ 1 receptor agonist PTZ (2 μM) and σ 2 receptor antagonist AC927 (2 μM). We included HREC treated with 40 μg/mL of AFL (B). Values are expressed as mean ± SEM of three independent experiments, each conducted in triplicate. Statistical analysis was performed using one-way ANOVA, followed by Tukey's test. *p < 0.05 vs Ctrl; # p < 0.05 vs VEGF-A; § p < 0.05 vs the same condition without agonist or antagonist. prodrug target molecules in the 92-1 tumor cell line was checked. HDACs are ubiquitous proteins expressed in all cell types, including UM 92-1 cells. 43 Gene expressions of σ 1 and σ 2 receptors, in turn, were confirmed by reverse transcription polymerase chain reaction (RT-PCR). Similar to MCF-7 human breast cancer cells (used as positive controls), UM 92-1 cells possess transcripts for both the SIGMAR1 gene (σ 1 receptor) and the TMEM97 gene (σ 2 receptor), a novel finding in the field ( Figure 5).
In fact, PTZ and AC927 alone promote, as expected, the proliferation of human UM 92-1 cells, while they are not able to restore the loss of cell viability induced by (±)-1 ( Figure  6B). These results are in agreement with antiproliferative effects in HREC (Table 3 and Figure 2) and suggest that the inhibition of proliferation by prodrugs (±)-1, (+)-1, and (−)-1 may reflect the HDACi activity by the VPA molecular component but not the σ receptor signaling by the HP-mII. 29,30 Indeed, the almost identical antiproliferative potencies exhibited by the three prodrugs would be consistent with the invariable chemistry of VPA constituents but not with the diverse stereochemistry of HP-mII enantiomers, which substantially influences their different binding affinities for σ receptors ( Table 2). While excluding σ receptor involvement in the regulation of proliferation by the three prodrug derivatives, present results do not rule out the possible offtarget contributions to VPA-dependent inhibition of proliferation, nor elucidate the exact molecular configuration of VPA interaction with the target (as a prodrug or separate, released metabolite).

■ CONCLUSIONS
In this report, the synthesis and pharmacological characterization of the two enantiomers of (±)-HP-mII valproate ester, (±)-1, is presented. The results indicate that the three prodrugs exhibit antiangiogenic activity in vitro, comparable to the VEGF-trap AFL but with distinct pharmacological profiles. Indeed, while all compounds induce σ receptor-independent HREC cytotoxicity to a similar extent, HREC migration and tube formation stimulated by VEGF-A are inhibited differently by the three prodrugs through the activation of selective σ receptor signaling. In this context, the enantiomer (−)-1 displays the highest antimigratory effect on VEGF-Astimulated HREC by acting as a σ 2 receptor agonist. Investigations in human UM 92-1 cells demonstrated the potential of the dual-ligand prodrugs as novel anti-UM agents and confirmed molecular pharmacology findings obtained in HREC. In particular, (±)-1 and its enantiomers induce pharmacologically equivalent cytotoxic effects on 92-1 cells, which are significantly greater than precursor compounds, probably dependent on the HDACi activity by the VPA molecular component but not related to σ receptor signaling. In close agreement with the findings in HREC, all compound prodrugs also significantly inhibit UM 92-1 cell motility with different potencies. Again, the (−)-1 enantiomer displays the highest antimigratory activity on 92-1 cells, followed by the (+)-1 stereoisomer and the racemic mixture, pointing toward a presumable involvement of σ receptor pathways and a biological dominance of σ 2 over σ 1 signaling in antimigratory effects by the dual-target/dual-function prodrugs. However, the exact molecular mechanisms underlying anti-UM effects by (−)-1 and its related prodrugs remain unclear and will be the focus of future studies.
On the basis of present findings, (−)-1 might represent a promising candidate for the development of pharmacological strategies to treat UM patients. In this context, beyond prodrugs with esterification, multiligand drug designs with simultaneous regulation of multiple targets could represent an alternative synthetic approach currently explored in this laboratory.

■ EXPERIMENTAL SECTION
Chemistry. Reagent grade chemicals were purchased from Merck KGaA (Darmstadt, Germany) and TCI Europe−Tokyo chemical industry (Tokyo, Japan) and were used without further purification. All reactions involving air-sensitive reagents were performed under N 2 in oven-dried glassware using the syringe-septum cap technique. The reactions were monitored by thin-layer chromatography (TLC) performed on silica gel Merck 60 F 254 -coated aluminum plates; the spots were visualized by UV light (λ = 254 nm) or iodine chamber. Melting points were determined on a Buchi B-450 apparatus in glass capillary tubes and are uncorrected. Optical rotations were taken at 20°C with a PerkinElmer 241 polarimeter (Llantrisant, U.K.). Flash chromatography purification was performed on a Merck silica gel 60 (40−63 μm; 230−400 mesh) stationary phase. Nuclear magnetic resonance spectra ( 1 H NMR and 13 C NMR recorded at 500 MHz) were obtained on Varian INOVA spectrometers using CDCl 3 , D 2 O, and DMSO-d 6 with 0.03% tetramethylsilane (TMS) as the internal standard. Chemical shifts (δ) are given in parts per million (ppm) and coupling constants (J) in Hertz (Hz). Signal multiplicities are characterized as s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, and br = broad. High-resolution mass spectrometry (HR-MS) was performed using an Orbitrap Fusion Tribrid mass spectrometer (Thermo Scientific). Compounds at the concentration of 10 μg/mL were dissolved in ACN/H 2 O (80/20) + 0.1% formic acid. Parameters: polarity (positive). Full scan mass was analyzed at a resolution of 120 000. Purities of all compounds reached at least 95% as determined by microanalysis (C, H, N)   Infinity II Multicolumn Thermostat (model G7116A), and a 1260 Infinity II Diode Array Detector (model G7115A). Data were acquired and integrated using the software Agilent OpenLab CDS LC ChemStation. The separation was performed using a Poroshell 120 EC-C18 (150 × 4.6 mm i.d., particle size 4 μm; Agilent, Santa Clara), maintained at 20°C. The samples were run using a mixture of water (A) and acetonitrile (B) enriched with trifluoroacetic acid (0.1% v/v). The gradient used was from 80% A to 100% B over 10 min. The flow rate was 0.8 mL/min. The UV detector was set at a length of 254 nm. Enantioselective HPLC analyses were performed using the same above-described conditions except for the stationary phase, which was a Chiralcel OJ[-RH] column (150 × 4.6 mm, 5 μm).
Kinetics of Chemical Hydrolysis. Pseudo-first-order rate constants (k obs ) for the hydrolysis of the compounds were then calculated considering the slopes of the linear plots of log (% residual compound) against time. The analyses were run in triplicate, and the mean values of the rate constants were calculated.
Kinetics of Enzymatic Hydrolysis. Human plasma was purchased from 3H Biomedical (Uppsala, Sweden, Europe). Plasma aliquots (4 mL) were diluted with 0.02 M PBS (pH 7.4) to obtain a final volume of 5 mL containing 80% plasma. Studies were performed at 37 ± 0.5°C using a shaking bath. Each experiment was started by adding a 10 −4 M drug stock solution (200 μL) to the preheated plasma. Hundred microliters of the thermostated medium was taken at various times, treated with cold methanol (500 μL) to precipitate plasma proteins, and centrifugated (5 min at 5000g). The supernatant was analyzed by HPLC to quantify the amount of the residual intact compound.
Receptor Binding Studies. The σ 1 and σ 2 receptor binding studies were performed according to the literature. 46,47 Briefly, for the σ 1 receptor binding assay, guinea pig brain membranes (400 μL, 500 μg protein) were incubated for 150 min at 37°C with 3 nM of the radiolabeled ligand [ 3 H]-(+)-pentazocine (45 Ci/mmol) and increasing concentrations of tested compounds in 50 mM Tris-HCl (pH 7.4) to a total volume of 1 mL. Nonspecific binding was assessed in the presence of 10 μM unlabeled haloperidol. Moreover, σ 2 receptor binding assays were made according to the following protocol: the guinea pig brain membranes (300 μL, 360 μg protein) were incubated for 120 min at room temperature with 3 nM [ 3 H]-DTG (31 Ci/mmol) in the presence of 0.4 mM radiolabeled ligand (+)-SKF10,047 to block the σ 1 sites. The incubation was performed in 50 mM Tris-HCl (pH 8.0) to a total volume of 0.5 mL with increasing concentrations of each test compound. Nonspecific binding was evaluated in the presence of 5 μM DTG.
Cell Proliferation. HREC and human 92-1 cell (seeded 4 × 10 3 and grown for 24 h at optimum culture conditions) proliferation was measured using crystal violet staining after treatment in 96-well plates with the indicated concentrations of chemicals for an additional 48 h. Control cells received an equal volume of vehicle (DMSO). At the end of the treatment, each well was washed with phosphate-buffered saline (PBS). Following this, the cells were fixed (in 4% paraformaldehyde) and stained with crystal violet solution (0.5% in 20% methanol for HREC or 1% aqueous solution for 92-1 cells). Subsequently, the plate was washed with water and left to dry. Crystal violet staining was evaluated by measuring the absorbance at 570 nm or 590 nm, after crystal violet extraction with 10% acetic acid (at room temperature for 10 min), with a microplate reader with the plate reader (Synergy 2-BioTek). For quantification, each assay was carried out in triplicate.
Wound Healing Assay. HREC and 92-1 cells (2.5 × 10 5 cells/ well) were seeded into 24 well plates and grown to confluency. Confluent cell monolayers were scratched with a p200 pipet tip to create 1 mm wide wounds. After washing (three times) with PBS to remove cell debris, wounded monolayers were incubated for 48 h (37°C ) in the complete medium in the presence of the indicated treatments. HREC were treated for 48 h in the presence of (±)-1, (+)-1, or (−)-1 with or without VEGF-A (80 ng/mL), AC927, and PTZ. 92-1 cells were treated for 48 h in the presence of (±)-1, (+)-1, or (−)-1. Wound closure was monitored by photographs at 40X using a phase-contrast microscope in each culture condition and at each time point (0, 24, and 48 h for HREC and 0, 6, 24, 30, and 48 h for 92-1 cells). The numbers of cells toward the wounds were counted using ImageJ software (ImageJ 1.50e, National Institutes of Health, NIH, Bethesda, MD) and were quantified by measuring the distance traveled over time by both cell fronts into the wound area.
Tube Formation Assay. The tube formation assay was analyzed in vitro in BD Matrigel, following the manufacturer's instructions (BD, Bedford). In brief, 96-well plates were coated with 50 μL of Matrigel and allowed to solidify at 37°C for 2 h. HREC were seeded at 1.5 × 10 4 cells/well in 100 μL of medium with (±)-1, (+)-1, or (−)-1 in the presence or absence of VEGF-A (80 ng/mL), AC927, and PTZ. After 8 h of incubation, tubelike structures were photographed at 10× magnification using an inverted microscope (Leica DM IRB) equipped with a CCD camera. Tube formations were quantified with ImageJ software (NIH, Bethesda, MD).
Statistical Analysis. All results are reported as mean ± SEM from at least two or three independent experiments (n = 2 or 3) performed at least in duplicate or triplicate. The results were analyzed using oneway ANOVA followed by Tukey−Kramer multiple comparisons test; differences between groups were considered significant for p-value < 0.05.