Synthesis and Biological Evaluation of (E)-4-Hydroxy-3-methylbut-2-enyl Phosphate (HMBP) Aryloxy Triester Phosphoramidate Prodrugs as Activators of Vγ9/Vδ2 T-Cell Immune Responses

The aryloxy triester phosphoramidate prodrug approach has been used with success in drug discovery. Herein, we describe the first application of this prodrug technology to the monophosphate derivative of the phosphoantigen HMBPP and one of its analogues. Some of these prodrugs exhibited specific and potent activation of Vγ9/Vδ2 T-cells, which were then able to lyse bladder cancer cells in vitro. This work highlights the promise of this prodrug technology in the discovery of novel immunotherapeutics.


■ INTRODUCTION
Present since birth, 1 Vγ9/Vδ2 T-cells represent the dominant subtype of human γδ T-cells in adult peripheral blood. 2 They expand in response to various infections, including tuberculosis, leprosy, typhoid, malaria, and toxoplasmosis, and studies in primate models have suggested a role in immunity to Mycobacterium tuberculosis. 3 Interestingly, they also exhibited an ability to target and lyse a diverse range of cancer cells in vitro. 2 Such properties have made the Vγ9/Vδ2 subset a major focus in the therapeutic exploitation of γδ T-cells. 4 Interestingly, Vγ9/Vδ2 T-cells have been shown to be activated by small molecule phosphoantigens (PAg) such as (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP) and isopentenyl pyrophosphate (IPP) ( Figure 1). 5,6 Beyond these natural ligands, two synthetic molecules, risedronate and zoledronate, activate Vγ9/Vδ2 T-cells through accumulation of IPP and are currently used in the clinic to treat osteoporosis and some types of cancer ( Figure 1). 7−9 The mechanism by which these small molecule phosphoantigens activate Vγ9/Vδ2 T-cells is understood to be mediated by the type-1 transmembrane protein butyrophilin 3A1. 10,11 Although conflicting reports exist as to whether PAgs bind to the extracellular or intracellular domains of this transmembrane protein, there is an increasing body of evidence that supports the notion that these PAgs bind the intracellular B30.2 domain of butyrophilin 3A1. 10,12−15 Encouraged by Vγ9/Vδ2 T-cells' ability to mount immune responses toward pathogens, lyse tumor cells, as well as their amenability to be targeted and modulated by small molecules (PAgs and their synthetic mimics), we embarked on the discovery of small molecules that have the potential to activate Vγ9/Vδ2 T-cells. Given our interest in the discovery and development of phosphorylated molecules and their prodrugs as therapeutics, 16−18 we focused our efforts on the natural phosphoantigen HMBPP as it is the most potent activator of Vγ9/Vδ2 T-cells reported to date ( Figure 1). 19 HMBPP has a pyrophosphate group, which accounts for poor drug-like properties, namely, poor cell membrane permeability, due its charged nature under physiological conditions (pH ≤ 7.4) and limited in vivo stability. These properties have hindered the development of many drugs with unmasked phosphate or pyrophosphate groups. To overcome these drawbacks, numerous phosphate prodrug strategies have been developed and used with success in the discovery of mostly nucleotide monophosphates and monophosphonates. 20−22 There have been reports in the literature of the application of the bis-pivaloyloxymethyl (bisPOM) phosphate prodrug technology and its derivatives to HMBP and its diphosphonates, which resulted in potent phosphoantigens though these were less potent than the parent phosphoantigen HMBPP. 19,23−25 Aiming to discover phosphoantigen prodrugs that are as potent as HMBPP, we focused our work on the aryloxy triester phosphoramidate 26 prodrug approach, in which the monophosphate or monophosphonate groups are masked by an aryl motif and an amino acid ester moiety. This prodrug technology is known to be more efficient in delivering monophosphorylated molecules than the bisPOM approach. 27 Over the past decade or so, it has led to at least 10 clinical candidates with two being eventually approved for clinical use. 28 Notably, this prodrug approach has mostly been used on nucleotide monophosphates and monophosphonates. In this work, we applied this powerful phosphate prodrug technology to the monophosphate derivative of the phosphoantigen HMBPP, (E)-4-hydroxy-3-methylbut-2-enyl phosphate (HMBP). The aryloxy triester phosphoramidate prodrugs of the phosphoantigen HMBP will be referred to as HMBP ProPAgens in this work.

■ RESULTS AND DISCUSSION
The synthesis of HMBP ProPAgens was similar to that reported by Reichenberg et al. 29 Briefly, the hydroxyl group of 1-hydroxypropan-2-one (1) was protected with TBS in the presence of imidazole (Scheme 1a). The ether product 2 was then treated with a Horner−Wadsworth−Emmons reagent, specifically triethyl phosphonoacetate, to create the double bond with E selectivity and yield the ester product 3. This compound, 3, was produced in a mixture of E/Z isomers with a 5:1 ratio similar to what is reported in the literature. 5 The E/Z isomers were separated by column chromatography using solvent mixture of 5−7% diethyl ether in hexane. The two products were then evaluated with 1 H and 13 C NMR utilizing previous work by Hintz et al. 5 who performed NOESY experiments to determine the identity of each isomer. The desired E-isomer, 3, was subsequently reduced using lithium aluminum hydride in THF to afford the key intermediate 4.
The coupling of compound 4 to phosphorochloridates to afford the desired prodrugs was achieved using standard procedures. 17,18 Phosphorochloridates, 9a−d, bearing methyl, isopropyl, tert-butyl, or benzyl esters, were synthesized following the procedure reported by Mehellou et al. 17 and as shown in Scheme 1b. These were coupled to 4 in the presence of triethylamine or N-methylimidazole (NMI) to afford the desired HMBP ProPAgens, 5a−d, in modest yields. Notably, NMI was only used in cases when TEA gave very low yields. The removal of the TBS protecting group was initially pursued using TBAF in THF. 30 However, this strategy did not work in our hands as the reaction yielded so many products and no traces of the desired HMBP ProPAgens were detected by mass spectroscopy. As an alternative, we used mild acidic conditions, 0.1 equiv of 1.25 M HCl in methanol, and this achieved the TBS deprotection without degrading the phosphate masking moieties to afford the desired HMBP ProPAgens 6a−d in good yields.
The unprotected HMBP ProPAgens 6a−d, however, exhibited low stability and underwent rapid degradation, which prevented their biological testing (purity <95%). In order to get an insight into the metabolism of ProPAgens 6a− d, compound 6d was incubated with the carboxypeptidase cathepsin A in vitro and the reaction was monitored with 31 P NMR. The data show that after 72 h, the major metabolite had a phosphorus peak at ∼1.95 ppm (Supporting Information Figure S1). Mass spectroscopy analysis of these samples showed that the degradation product, which had a 31 P NMR peak of ∼1.95 ppm, was the phosphate group masked with the aryl motif and the amino acid ester moiety (Supporting Information Figure S2). This indicated that the P−O bond in HMBP ProPAgens 6a−d was labile and was cleaved off to release the unphopshorylated PAg backbone rather than HMBP. This may explain literature reports pursuing the phosphonates of HMBP where the −O-Pbond was replaced by a −CH 2 -P-one and lack of reports on the native HMBP phosphate prodrugs. 19 Interestingly, HMBP ProPAgens 5a−d, which had the side chain hydroxyl group protected with a TBS moiety, exhibited better stability than 6a−d, and hence we were able to characterize them fully and obtain a measure of their purity (see Supporting Information). Prodrugs 5a−d were then investigated for their ability to activate Vγ9/Vδ2 T-cells. For this, peripheral blood mononuclear cells (PBMCs) containing Vγ9/Vδ2 T-cells derived from healthy donors were incubated with increasing concentrations of HMBPP, zoledronate, or HMBP ProPAgens 5a−d (up to 100 μM) (Figure 2a and Figure 2b). Peripheral blood γδ T-cells lack appreciable levels of surface CD69 or CD25 under steady state conditions, but Tcell receptor (TCR) stimulation upregulates both T-cell activation markers. 31 PAg responsive Vγ9/Vδ2 T-cells were then distinguished by TCR Vγ9 and Vδ2 expression and assessed for the upregulation of CD69 and CD25.
As shown in Figure 2c and Figure 2d, the natural phosphoantigen HMBPP showed significant activation of Vγ9/Vδ2 T-cells, EC 50 = 0.06 nM, comparable to its reported potency. 19 Additionally, zoledronate showed a moderate activation, EC 50 ≈ 500 nM as expected (
Next, we tested the potential for ProPAgens to sensitize the urinary bladder carcinoma cell line T24 for targeted killing by in vitro expanded Vγ9/Vδ2 T cells. Without sensitization, medium pulsed T24 cells were poorly targeted by increasing ratios of Vγ9/Vδ2 T cells, but upon pulsing for 4 h with zoledronate or ProPAgens 5a−d, T24 cells were specifically lysed (Figure 3a). The short pulsing period, poor lipophilicity, and requirement to target a metabolic enzyme resulted in only a marginal increase in specific T24 cell killing by zoledronate (Figure 3b). In contrast, a 1000-fold lower concentration of each ProPAgen mediated a 2-to 4-fold increase in specific lysis of T24 over the same period (Figure 3b).
The activity across the different HMBP ProPAgens in both biological assays was in agreement with what is typically observed with aryloxy triester phosphoramidate (ProTide) prodrugs of nucleoside analogues as it correlated with the lipophilicity and rate of degradation. The HMBP ProPAgen with benzyl esters, e.g., 5b, has higher lipophilicity and thus better (passive) cellular uptake. Also, the benzyl group is a better leaving group than the other aliphatic esters, and thus the metabolism of the benzyl ester of 5b proceeds faster than the other ProPAgens, 5a, 5c, and 5d. Together, these explain the superior activity of ProPAgen 5b. HMBP ProPAgen 5a also exhibited relatively potent activation of Vγ9/Vδ2 T-cells with EC 50 = 1.38 nM. The fact that ProPAgen 5c, which has an isopropyl ester, exhibited lower potency in activating Vγ9/Vδ2

Journal of Medicinal Chemistry
Brief Article T-cells as compared to 5a and 5d was surprising as often phosphoramidate prodrugs with this ester exhibit similar activity to those with a methyl ester. However, the less potent activity observed with the HMBP ProPAgen 5d was expected and is in agreement with the literature where phosphoramidate prodrugs with t Bu esters show lower biological activity than those with other ester motifs. This is because the hydrolysis of the t Bu ester group of the phosphoramidates by esterase enzymes proceeds much slower than those with Me, iPr, and Bn esters.

■ CONCLUSION
We reported the application of the aryloxy triester phosphoramidate technology to the phosphoantigen HMBP and one of its synthetic analogues to generate HMBP ProPAgens. Although the ProPAgens of the native HMBP compound (6a−d) were of very low stability, those of the synthetic analogue (5a−d) were more stable. These later ones exhibited potent activation of Vγ9/Vδ2 T-cells. Notably, HMBP ProPAgen 5b was the most potent activator of Vγ9/Vδ2 Tcells, EC 50 = 0.45 nM with the other three ProPAgens, 5a, 5c, and 5d, also showing low nanomolar activation. Impressively, Vγ9/Vδ2 T-cells activated by HMBP ProPAgens exhibited potent lysis of urinary bladder carcinoma cancer cells (T24) in vitro. In terms of specificity, HMBP ProPAgens, 5a−d, showed excellent specificity toward the activation of Vγ9/Vδ2 T-cells as they had no effects of other T-cells such as CD8+ αβ T-cells. Efforts aimed at improving the stability of the native HMBP ProPAgens are currently underway and will be reported in the future. In conclusion, the application of the aryloxy triester phosphoramidate prodrug technology to HMBP to generate HMBP ProPAgens yielded potent and specific activators of Vγ9/Vδ2 T-cells. Together, this showcases the promise of this prodrug technology in the discovery of novel immunotherapeutic agents.
■ EXPERIMENTAL SECTION General Information. All of the reactions were carried out under argon atmosphere and were monitored with analytical thin layer chromatography (TLC). NMR data were recorded on a Bruker AV300, AVIII300, AV400, AVIII400, or DRX500 spectrometer in the deuterated solvents indicated, and the spectra were calibrated on residual solvent peaks. Chemical shifts (δ) are quoted in ppm, and J values are quoted in Hz. In reporting spectral data, the following abbreviations were used: s (singlet), d (doublet), t (triplet), q (quartet), dd (doublet of doublets), td (triplet of doublets), and m (multiplet). HPLC was carried out on a DIONEX summit P580 quaternary low-pressure gradient pump with a built-in vacuum degasser using a Summit UVD 170s UV/vis multichannel detector. HPLC grade solvents were used. Chromeleon software was used to visualize and process the obtained chromatograms. Analytical separations used a flow rate of 1 mL/min and preparative used a flow rate of 20 mL/min. The purity of the tested compounds was determined by high-performance liquid chromatography (HPLC) where all of the tested compounds had ≥95% purity.
1-((tert-Butyldimethylsilyl)oxy)propan-2-one, 2. Hydroxyacetone 1 (1.00 g, 13.50 mmol, 1 equiv) and TBSCl (3.05 g, 20.25 mmol, 1.5 equiv) were combined in dry CH 2 Cl 2 (50 mL) under an argon. Imidazole (2.02 g, 29.70, 2.2 equiv) was then added portionwise at 0°C . The mixture was allowed to warm to room temperature and left to stir at that temperature for 2.5 h. The reaction mixture was then washed with brine (50 mL), extracted with Et 2 O (50 mL × 3), dried (MgSO 4 ) and the solvents were removed under reduced pressure to give a crude oil, which was purified via column chromatography (10:1 hexane/EtOAc). 2 was obtained as a colorless oil (2.45 g, 95%). (E)-4-((tert-Butyldimethylsilyl)oxy)-3-methylbut-2-en-1-ol, 4. In a flame-dried Schlenk flask were placed ester 3 (0.5 g, 1.93 mmol, 1 equiv) and 20 mL of dry THF, and the resulting mixture was cooled to 0°C with stirring. Then LiAlH 4 (1 M solution in THF) (1.93 mL, 1.93 mmol, 1 equiv) was added dropwise, and the solution was allowed to warm to room temperature and stirred for 2 h. MeOH (1 mL) was then added cautiously, after which a saturated solution of sodium potassium tartrate (30 mL) was added which formed a white gel. This was stirred overnight at room temperature, and then the phases were allowed to separate. The organic layer was collected and the aqueous layer was washed with Et 2 O (20 mL × 3), dried (MgSO 4 ) and the solvents were removed under reduced pressure to leave a crude oil. This was then purified by column chromatography using hexane/ethyl acetate (4:1) as an eluant to give product 4 (0.1677 g, 40%) as a colorless oil.

Journal of Medicinal Chemistry
Brief Article