Solid-Phase Synthesis of New Trp(Nps)-Containing Dipeptide Derivatives as TRPV1 Channel Blockers

Trp(Nps)-Lys-NH2 derivatives, bearing alkyl or guanidine groups either at the N–terminus or on the Lys side-chain or at both positions were conveniently prepared on solid-phase and evaluated as TRPV1 channel antagonists.


Introduction
TRPV1 (Transient Receptor Potential Vanilloid 1) is a non-selective, Ca 2+ preferring, ion channel, activated by temperatures higher than 42 ºC, acidic pH, vanilloids such as capsaicin, and the endogenous cannabinoid receptor ligand anandamide [1]. TRPV1 expression is up-regulated in a number of painful disorders, including chronic, neuropathic and acute inflammatory pain, and OPEN ACCESS consequently, it is a promising therapeutic target for pain relief [1,2]. In this respect, numerous research programs have been dedicated to the search of TRPV1 modulators to be used as pharmacological tools for better understanding the pharmacology of this cation channel. In addition, few modulators that emerged from these programs have reached clinical trials for multiple therapeutic indications [3,4]. However, due to possible side effect issues, such as diminished response to damaging heat stimuli, altered body temperature and reduction in the perception of taste, the development of TRPV1 antagonists as analgesic drugs still needs compounds with greater efficacy and fewer adverse effects [3,4].
Like ruthenium red [5], Arg-rich peptides have been described as non-competitive antagonists of TRPV1 receptors, through binding to a site located near the entryway of the aqueous pore [6]. A few years ago, we found that dipeptide derivatives Xaa-Trp (Nps) and Trp(Nps)-Xaa (Xaa=Lys, Arg, Nps = 2-o-nitrophenylsulfenyl), first described as analgesic dipeptides of unknown mechanism of action [7,8], inhibited the activation of TRPV1 in the micromolar range, and also acted as NMDA channel blockers, although with lower potency [9]. In view of their structural analogy with the Argrich peptides, it was hypothesized that these Nps-containing dipeptides could interact with the vanilloid receptor through the pore entrance. Now, in an attempt to fine-tune the potency/selectivity balance within this family of channel blockers, we have prepared and evaluated a new series of H-Trp(Nps)-Lys-NH 2 dipeptide derivatives, incorporating alkyl and guanidino moieties at the N-terminal group and on the Lys side-chain. This paper deals with the solid-phase synthesis and the biological activity displayed by this new series of Trp(Nps)-containing dipeptides, compared to model compounds H-Trp(Nps)-Lys-NH 2 (1) and H-Trp(Nps)-Arg-NH 2 (2).

Synthesis
The synthesis of the Trp(Nps)-Lys-NH 2 dipeptide derivatives was performed by parallel solidphase synthesis on a Rink-amide MBHA polystyrene resin (NH 2 -PS), following a Fmoc/ t Bu strategy (Scheme 1). Fmoc-Trp-Lys(Boc)-NH-PS (3) and Fmoc-Trp-Lys(Alloc)-NH-PS (4) were prepared from the corresponding amino acid derivatives as appropriate intermediates for modifications at the Nterminus and the Lys side-chain, respectively. Then the most demanding step, the incorporation of the Nps moiety, was performed by reaction with 2-nitrobenzenesulphenyl chloride (Nps-Cl) in an 8:2 AcOH-DMF mixture, to produce dipeptidyl resins 5 and 6. After the orthogonal removal of the Fmoc group of 5, two portions of the resin were subjected to reductive amination by condensation with benzaldehyde and isobutylaldehyde, followed by treatment with NaBH 3 CN [10], to give compounds 7 and 8. Additional aliquots of the Fmoc intermediate 5 were deprotected and then treated with N,Ndi(Boc)-S-methylisothiourea, HBTU and HBPyU in the presence of a base to provide the corresponding N-guanylated derivatives 9-11 [11,12]. Finally, resin-bound derivatives 7-11 were cleaved from the resin with TFA:H 2 O (19:1, v/v) to render the corresponding N-alkyl (17, 18) and Nguanidyl derivatives 19-21 in low-moderate yield and high purity after extraction in reverse-phase cartridges. When the same reactions were applied to dipeptidyl resin 6, after Alloc removal, the expected N ε -benzyl and N ε -isobutyl derivatives 22a, 23a were obtained after cleavage. However, guanidylation reactions mainly resulted in the corresponding diguanidylated compounds 24b-26b. A plausible explanation could be that the free amino group of the Lys side-chain, generated after removal of the Alloc group, could be basic enough under the guanidylation reaction conditions to promote the Fmoc removal in intermediate 6. As reported by Farrera-Sinfreu et al. [13], this undesired removal of the Fmoc group from Lys peptides seems to be time-and base-dependent, and could be prevented by quick chemical capture of the ε-NH 2 group, not applicable in our case. Accordingly, the treatment with N,N-di(Boc)-Smethylisothiourea, requiring the presence of mercury (II) chloride as catalyst and TEA and repeated twice for reaction completion [11], the only isolated product was the disubstituted compound 24b. However, guanidylation using HBTU or HBPyU in the presence of DIEA led to inseparable mixtures of mono-and di-substituted derivatives, 25a + 25b and 26a + 26b in a ~1:3 ratio (Scheme 1, Table 1). All compounds were obtained as single diastereoisomers, as deduced by HPLC and 1 H-NMR, indicating the stereochemical integrity of all the steps in the process.

Biological activity
Compounds 17-26 were evaluated for their ability to inhibit Ca 2+ influx through TRPV1 channels, activated by capsaicin, and to block the glutamate-evoked activity of NMDA receptors (Figure 1). Model compounds 1 and 2 were included for comparative purposes. As shown in Figure 1, the incorporation of alkyl groups at the N-terminus of Trp residue decreases ( i Bu, 17) or maintains (Bzl, 18) the TRPV1 blockage with respect to the parent compound H-Trp(Nps)-Lys-NH 2 (1). A free guanidinium moiety at this position also reduces the TRPV1 antagonist properties, while the incorporation of tetramethyl-and pyrrolidyl-substituted guanidinos improves the percentage of TRPV1 blockage (20 and 21). It is noteworthy that, in general, substitution at the N-terminus leads to better TRPV1/NMDA selectivity than the model dipeptide derivative.
N ε -Alkyl Lys derivatives 22a and 23a show comparable and slightly better activity and enhanced selectivity with respect to the Lys-model dipeptide 1. Compared to the H-Trp(Nps)-Arg-NH 2 (2) prototype, diguanidino derivatives 24b-25b do not result in significant improvement of the blocking effects. While they either retain or slightly improve TRPV1 blockade, selectivity towards NMDA diminished, especially for compound 24b, with two free guanidino groups.

General
All reagents were of commercial quality. Solvents were dried and purified by standard methods. 1 H-NMR spectra were recorded at 300 MHz on a Varian XL-300 instrument (samples at 5−10 mM concentrations). Electrospray mass spectra (ES-MS) were performed, in positive ion mode, on a Hewlett-Packard 1100SD or an HPLC-MS WATERS equipment, the last integrated by a 2695 separation module, a 2996 photodiode array detector and a Micromass ZQ 2000 spectrometer. Analytical HPLC chromatograms were done on a C 18 reverse-phase column (SunFire C 18 , 4.6 × 50 mm; Flow rate: 1 mL/min; Gradient: 10% A to 50% A in 8 min, were A: CH 3 CN + 0.08% formic acid and B: H 2 O + 0.1% formic acid). The solid-phase syntheses were carried out manually, on plastic syringes with a porous filter, and attached to a vacuum manifold for fast removal of excess of reagents and solvent. Resins were swollen in DCM/DMF/DCM/DMF (4 × 0.5 min).

Typical procedure for Alloc removal
To the pre-swollen resin, under Ar atmosphere, dried DCM (1 mL) and phenylsilane (25 equiv.) were added. After 5 min, a solution of Pd[PPh 3 ] 4 (0.1 equiv.) in dried DCM (1 mL) was added and the mixture was allowed to stand for 1 h 30 min. After this time the resin was drained off, washed with DCM (5 x 1 min) and the deprotection procedure was repeated once more, allowing the mixture to stand for 30 min. Finally, the resin was successively washed with DCM/dioxane-H 2 O (9:1)/DMF/DCM (5 × 1 min).

Cleavage reactions
For resins 12-16, the removal of the Fmoc protecting group before cleavage was required. Resins 7-11 and 12-16 were treated with a mixture of TFA-H 2 O, 19:1 (1 mL for 100 mg) at r.t. for 2 h. After washings with DCM (5 × 1mL), the filtrates were evaporated and the residue solved in water and washed with ethyl acetate. The aqueous phase was lyophilized to give derivatives 17-26, isolated as trifluoroacetate salts. When required, the products were purified through reverse-phase cartridges using mixtures of H 2 O/ACN as solvent.  ) with a two-microelectrode voltage-clamp amplifier at 20 ºC. TRPV1 channels were activated by application of 10 µM capsaicin in the absence or presence of compounds (10 µM) at a holding potential (Vh) of −60 mV. Whole-cell currents from oocytes injected with NR1/NR2A (1:3, w/w) subunits were recorded in standard Ringer solution upon activation with L-glutamate (100 µM) supplemented with glycine (10 µM) at Vh = −60 mV. Responses were normalized with respect to that evoked without channel blockers. Experiments were performed in triplicate; each bar in Figure 1 represents the mean ± SEM, n (number of oocytes) = 12.

Conclusions
Convenient solid-phase procedures have been developed for the modification of Trp(Nps)-Lys dipeptides at the N-terminus or on the Lys side-chain. Main steps of these procedures are the AcOHcatalyzed incorporation of the Nps moiety at the indole ring, compatible with TFA labile resins, the reductive amination to N-alkyl derivatives, and the guanidylation reactions with thioureas or uronium salts. Diguanidylated by-products were obtained during the incorporation of guanidino groups on the Lys side-chain, probably due to Fmoc-removal by the free Lys ε-NH 2 group under the basic media. From the biological point of view, the presence of substituted guanidino groups at N-terminus led to the best TRPV1 activity/NMDA selectivity ratio. Based on the significant activation of NMDA receptor by compound 20 (∼30%), derivative 21 could be considered as a suitable molecule for further modifications in the search for new TRPV1 channel blockers.