Photoswitchable Inhibitor of the Calcium Channel TRPV6

Abstract

Herein we report the first photoswitchable inhibitor of Transient Receptor Potential Vanilloid 6 (TRPV6), a selective calcium channel involved in a number of diseases and in cancer progression. By surveying analogs of a previously reported TRPV6 inhibitor appended with a phenyl-diazo group, we identified a compound switching between a weak TRPV6 inhibitor in its dark, E-diazo stereoisomer (Z/E = 3:97, IC₅₀ ≫ 10 μM) and a potent inhibitor as the Z-diazo stereoisomer accessible reversibly by UV irradiation at λ = 365 nm (Z/E = 3:1, IC₅₀ = 1.7 ± 0.4 μM), thereby allowing precise spatiotemporal control of inhibition. This new tool compound should be useful to deepen our understanding of TRPV6.

TRPV6 is a member of the Transient Receptor Potential Vanilloid (TRPV) family¹⁻⁵ with affinity for divalent ions, critically involved in Ca²⁺ uptake in intestine, placenta, and exocrine tissues.^6,7 TRPV6 expression is altered in a diversity of diseases, including Crohn’s disease and tumors.⁸⁻¹⁰ Nevertheless, the exact role of TRPV6 in these pathologies remains to be clarified.¹¹⁻¹⁴ Recently, we reported the first selective and potent TRPV6 inhibitors, e.g. 1 and 2, based on a (cis-4-phenyl-cyclohexyl)piperazine scaffold (Figure 1). These inhibitors were used to probe the pharmacological effect of TRPV6 inhibition on cancer cell growth.¹⁵ Here, we set out to develop a photoswitchable version of our inhibitors to facilitate further studies on the mechanism of TRPV6.

Figure 1.

Chemical structures of TRPV6 inhibitors 1–2 and investigated analog photoswitches 9a–e.

Photoswitchable inhibitors can be obtained by incorporating a diazo group into their structure, which isomerizes between the more stable E- and the less stable Z-isomer upon UV irradiation at 365 nm (E → Z) and 470 nm (Z → E).¹⁶ The diazo moiety may substitute a two-atom linker separating two aromatic groups in an existing inhibitor to form an azoster.^17,18 Alternatively, a phenyldiazo group is appended at a position such that E/Z photoisomerization modulates inhibition.¹⁹ The latter design has been used in all photoswitchable ion channel inhibitors reported to date,²⁰ including the four cases of TRP ion channel modulators, which were used for optical control of nociceptor TRPV1²¹ and sensory neuron receptors TRPC2/3/6.^22,23 In all of these cases the phenyldiazo group replaced or extended a lipophilic group present in the parent inhibitor, and the Z-diazo isomer was more inhibitory than the E-diazo isomer (Figure S1).

To design our TRPV6 photoswitches, we followed the phenyldiazo appendage approach starting from inhibitors 1 and 2. Structure–activity relationship studies had shown that only minimal perturbations were allowed on the cis-4-phenyl-cyclohexyl group, while a range of different aromatic groups were allowed at the second piperazine substituent, the most potent inhibitors containing a meta-substituted pyridine.¹⁵ We therefore extended the pyridine ring in 1 or 2 to form a diazobenzene group, hoping to observe differential TRPV6 inhibition between the Z-diazo and E-diazo isomers, introducing a pyridine nitrogen atom where allowed by synthesis (Figure 1).

Phenyldiazo analogs 9a–b were prepared by first reacting N-Boc-protected 4-amino or 4-nitro-aryl-piperazines (3a–b) with nitrosobenzene to obtain the corresponding diazo intermediates (4a–b), followed by Boc removal to the free piperazines 5a–b and reductive alkylation of the piperazine with 4-phenyl-cyclohexanone 6. Analog 9c was similarly obtained from a diazo intermediate 5c obtained by converting 3-aminopyridine (3c) to the corresponding phenyldiazonium (4c) and coupling with 1-phenylpiperazine (7). However, we obtained 9d–e by alkylation of (4-phenylcyclohexyl)piperazine 8(24,25) with 4-hydroxymethyl or 4-bromomethyl-aryldiazoaryl intermediates 5d–e obtained separately (Scheme 1).²⁶ In all cases the TRPV6 inhibitory cis-1,4-cyclohexyl stereoisomer with the diazo group in the dark, E-configuration, was isolated by column chromatography. All diazo compounds were finally precipitated as dihydrochloride salts. The structure of our photoswitches was confirmed by X-ray crystallography in the case of E-9a, E-9c, and E-9e, which showed that the phenyl substituent was axial and the piperazine equatorial relative to the cyclohexane ring.

Scheme 1. Synthesis of TRPV6 Photoswitches 9a–e, and X-ray Crystal Structure of E-9a, E-9c, and E-9e.

Reagent and conditions: (a) nitrosobenzene, NaOH, benzene, H₂O, reflux, 10 min (45%); (b) HCl, H₂O, reflux, o.n. (84%); (c) (i) 4-phenyl-cyclohexanone (6), Et₃N, NaBH(OAc)₃, DCE, r.t., 48 h; (ii) HCl, MeOH (34–40%, over 2-steps); (d) (i) H₂, Pd/C, AcOEt, r.t., 48 h; (ii) nitrosobenzene, AcOH, r.t., 24 h (61%, over 2-steps); (e) NaNO₂, HCl, H₂O, 0 °C to r.t., 17 h (65%); (f) LiAlH₄, THF, Ar, 0 °C to r.t., 24 h (76%); (g) TBDMSCl, DIPEA, DMAP, CH₂Cl₂, DMF, 0 °C to r.t., 8 h (53%); (h) nitrosobenzene, NaOH, toluene, H₂O, 50 °C, 17 h (55%); (i) TBAF, THF, r.t., 2 h (90%); (j) (i) MsCl, Et₃N, THF, 0 °C to r.t., 2 h; (ii) (4-phenylcyclohexyl)piperazine (8), K₂CO₃, NaI, DMF, 60 °C, 2 h; (iii); HCl, MeOH (20%, over 3-steps); (k) nitrosobenzene, AcOH, r.t., 24 h (93%); (l) NBS, BPO, CCl_4, reflux, 24 h (99%); (m) (i) (4-phenylcyclohexyl)piperazine (8), K₂CO₃, DMF, r.t., 24 h; (ii) HCl, MeOH (46%, over 2-steps).

We analyzed the photoswitchable behavior of 9a–e using RP-UHPLC and UV–vis spectrometry. The 500 μM stock solutions in DMSO maintained in the dark for at least 48 h contained almost pure E-isomers (Table 1). These E-isomers displayed a characteristic maximum absorbance (λ_max(E)) between 300 and 400 nm (Figure 2). In the case of 9d and 9e, UV irradiation at λ = 365 nm resulted in a decrease of this absorbance concomitantly with the appearance of a new peak detected by RP-UPLC (Figures S4–S6) and a new maximum absorbance at UV–vis (λ_max(Z)) between 400 and 500 nm corresponding to the Z-isomer.^27,28 In the case of 9a, irradiation at λ = 365 nm did not result in a new λ_max(Z), probably due to overlap of both E- and Z-λ_max. The photostationary state of 9a, 9d, and 9e was reached within 5 min and featured the less stable Z-isomer (Z/E ratios of 1:7 to 3:1), which returned to E-isomer in the dark within minutes to days upon thermal isomerization (Figure 2B,C). We did not observe any switching effect for 9b–c (Figure S3), probably due to fast thermal isomerization.^29,30

Table 1. Photoisomerization Properties of Compounds 9a–e.

entry	dark-adapteda	UV-adaptedb	T1/2c
9a	100:0	88:12	8.2 min
9b	93:7	93:7	-
9c	100:0	100:0	-
9d	100:0	50:50	3.7 h
9e	97:3	25:75	3.9 d

^aE/Z ratio for the dark state.

^bE/Z ratio after photoisomerization at λ = 365 nm for 5 min. Ratios determined by RP-UHPLC.

^cHalf-life for thermal Z → E isomerization in the dark, determined by UV–vis spectroscopy.

Figure 2.

(A) Individual UV–vis spectra of compounds 9a, 9d, and 9e (DMSO, 500 μM), in the dark-adapted (black), UV-adapted (magenta, 365 nm), blue-adapted (blue, 470 nm) states. The DMSO solutions were irradiated for 5 min in a self-built chamber equipped with LED lamps³¹ (90 mW). (B) Light-magenta lines represent the UV–vis spectra of the thermal relaxation of the UV-adapted state (dark-magenta line) in direction to the dark-adapted state (black line) over time. (C) Absorption at λ_max(E) was plotted as a function of time (blue cross), with exponential fit (black line).

To evaluate TRPV6 inhibition by our diazo analogs, we measured their effect on channel activity by means of cadmium (Cd²⁺) uptake into HEK293 cells stably overexpressing human TRPV6 (HEK-hTRPV6) as described previously.¹⁵ In this assay, a solution of Cd²⁺ is applied extracellularly, and its transport through TRPV6 leads to complexation and fluorescence of an intracellular dye. Screening of the dark state revealed partial inhibition of Cd²⁺ transport compared to vehicle control for E-9a (30%), E-9d (70%), and E-9e (13%) (Figure 3, black bars).

Figure 3.

Inhibition of Cd²⁺ (50 μM) influx into HEK293-hTRPV6 cells. Data was normalized to the maximum entry in the vehicle group (buffer/DMSO). Compounds 1–2 (positive controls) 9a, and 9d–e were tested at 10 μM. Diazo derivatives 9a–e showed limited solubility and could not be measured above 10 μM. **P < 0.01; ***P < 0.001. Data shown are mean + SEM (n = 3 or 6).

To access the Z-isomers, we irradiated the 10 mM DMSO stock solutions of each individual compound at λ = 365 nm for 5 min. Then, the solutions containing Z-isomers were immediately diluted with assay buffer to the same concentration of screening used before (10 μM). Switching did not affect channel blockage of 9a but increased the inhibitory activity of compounds 9d and 9e (Figure 3, magenta bars).

Unfortunately, dose–response measurements for the dark and photoisomerized states of 9d had comparable IC₅₀ values (IC₅₀ ≈ 0.6 μM, Figure S8). However, we observed a striking switching effect with 9e, which upon UV irradiation turned from a weak E-isomer (<15% inhibition at 10 μM) to a potent Z-isomer (IC₅₀ = 1.7 ± 0.4 μM, Figure 4A) matching the inhibitory potency of its template 2 (IC₅₀ = 1.7 μM).¹⁵ Furthermore, irradiation of 9e with UV light at alternate 365 and 470 nm led to alterations of the UV–vis spectrum indicating fully reversible isomerization between E-9e and Z-9e, suggesting that this inhibitor would be well suited for reversible photoswitching of TRPV6 (Figure 4B).

Figure 4.

TRPV6 inhibition and photoswitching of 9e. (A) Inhibition curve of Cd²⁺ influx into HEK293-hTRPV6 cells by Z-9e. Data shown are mean ± SEM (n = 6/concentration) from four independent experiments. (B) UV absorbance at 327 nm indicating photoswitching between E-9e and Z-9e upon irradiation alternatively at 365 and 470 nm, measured at 500 μM in DMSO.

To evaluate photoswitching of the TRPV6 channel with inhibitor 9e we conducted electrophysiological experiments in transiently transfected HEK293-hTRPV6 cells (Figure 5). The experiment started in 10 mM Ca²⁺ solution. After 120 s the constitutively active TRPV6 currents reached a plateau, and a subsequent 10 mM Ca²⁺ solution containing 10 μM E-9e or an equivalent amount of DMSO as control was applied (Figure 5A, time point I), followed by applying continuous irradiation at λ = 365 and 470 nm. Finally, 10 μM La³⁺ was added for full inhibition at the end of the experiment. While currents were stable over the course of the experiment in the control sample, addition of E-9e led to a significant decrease of the TRPV6 current densities. The partial inhibition by E-9e in this experiment contrasts with the very weak inhibition observed in the FLIPR assay (Figure 3) and probably reflects inherent differences in the two assay formats. Upon irradiation of E-9e at 365 nm, the currents were further reduced, which was reversible upon excitation with 470 nm. Subsequent irradiation via alternating wavelength (365/470 nm) showed the repeatability and photostability of this process (Figure 5B). The current–voltage relationship of TRPV6 currents before and after the application of 9e revealed positive reversal potentials, indicating unchanged Ca²⁺ selectivity (Figure 5C).

Figure 5.

Electrophysiological characterization of the photoswitchable TRPV6 inhibitor 9e. (A) Averaged time course of whole-cell currents of YFP-TRPV6 transfected HEK293 cells. The diagram illustrates the whole-cell inward current density at −74 mV during a 200 ms voltage ramp from −90 to +90 mV (applied every 5 s, HP = 50 mV), recorded by patch clamp technique, across time. Black arrows indicate time points of 9e photoswitching by continuous irradiation at λ = 365 nm (polychrome IV monochromator, 150 W Xe lamp, first arrow: turn-on) and 470 nm. Data shown are mean – SEM (n = 4, DMSO or n = 10, 9e) from at least two independent experiments. I–VI illustrate the data points used for graphs (B) and (C). (B) Bar chart (mean + SEM) of current density of YFP-TRPV6 expressing HEK293 cells before and after (I, white; II, gray) addition of 10 μM 9e and after the exposure with 365 nm (III, V, magenta) and 470 nm (IV, VI, blue). **P < 0.0005. (C) Representative current–voltage characteristic before the addition of 9e (I, black), as well as after the exposure with 365 nm (III, magenta) and 470 nm (IV, blue).

In summary, by surveying phenyldiazo analogs of the previously reported TRPV6 inhibitors 1 and 2, we identified 9e as the first photoswitchable inhibitor of TRPV6. This inhibitor switches reversibly between a weak E-isomer in its dark state and an inhibitory Z-isomer upon UV irradiation at λ = 365 nm, which matches the activity of its template inhibitor 2. Rapid photoswitching of TRPV6 was illustrated by monitoring time-dependent inhibition of TRPV6 by 9e using electrophysiology. Further studies with 9e should provide new insights into the physiological and pathological roles of TRPV6.

Glossary

ABBREVIATIONS

TRPV

transient receptor potential vanilloid

TRPA

transient receptor potential ankyrin

TRPC

transient receptor potential channel

SAR

structure–activity relationships

RP-UHPLC

reversed phase-ultra high performance liquid chromatography

NCF

nominal calcium free

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.9b00298.

• Figures S1–S8, Table S1, experimental details for chemistry, photochemical characterization, biological assays, spectra copies of ¹H/¹³C NMR, crystal structure reports, and purity of compounds 9a–e (PDF)

Author Contributions

M.R.C. designed, synthesized, and purified the compounds. R.B. and M.R.C. performed FLIPR assays. S.L. and C.B. designed and performed electrophysiology experiments. M.R.C. and J.L.R. wrote the manuscript. M.A.H., C.R., and J.L.R. supervised the study. All the authors discussed the results and commented on the manuscript.

This work was supported financially by the Swiss National Science Foundation, NCCR TransCure.

The authors declare no competing financial interest.