In this work, 53 α-phenylglycine amides were analyzed by a diverse range of in silico approaches like activity cliff, molecular docking, molecular dynamics and 3D-QSAR model construction. The analysis assisted the design new compounds that could potentially inhibit the Transient Receptor Potential Melastatin 8 (TRPM8). This non-selective cation channel has a link with some diseases such as migraine, overactive bladder, and prostate cancer. A hybrid QSAR model, with acceptable figures of merit (R2adj = 0.87, Q2LOO = 0.86, Q2ext = 0.75), was used to predict the pIC50 for various designed structures. The synthetic routes employed in previous works was used to guide structure planning ensuring synthetic accessibility. Druglikeness properties were analyzed by the SwissADME website to filter out non-suitable compounds. It was possible to create four prototypes with higher pIC50. All designed compounds can be readily synthesized and tested for TRPM8 inhibition.
Research Article
Employing QSAR to design synthetic accessible TRPM8 Inhibitors
https://doi.org/10.21203/rs.3.rs-2371227/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Transient Receptor Potential Melastatin 8
TRPM8
Antagonist
Rational planning
QSAR
TRPM8 is a non-selective polymodal ion channel receptor activated by pressure, refrigeration compounds such as menthol and allicin[1], and low temperatures (15–25 ºC) [2]. TRPM8 having altered expression or function has been linked to different stages of different diseases. As it has a relatively restricted expression in normal tissues, an abnormal expression offers the possibility of its use as a marker for diagnosis and therapeutic intervention [3].
Inhibitors of this receptor began to be sought, developed, and studied but the rational planning was hindered until the elucidation by cryo-electron microscopy in 2018 was unveiled [4]. Rational drug design enabled the design of selective TRPM8 inhibitors, dodging the Transient Receptor Potential V1 (TRPV1) and Transient Receptor Potential Ankyrin 1 (TRPA1) [5]. Such selective antagonists are being studied for migraine [6, 7], prostate cancer [8], breast cancer [9, 10], overactive bladder [11], eye pain [12], allodynia [13], neuropathic pain [14], among other disorders.
A series of selective TRPM8 α-phenylglycinamides antagonists were synthetized giving a very promising compound, (1) [15]. The objective was the discovery of new drugs for overactive bladder. Efficacy in vivo is proven by test wet-dog shake (WSD) (Fig. 1). Despite good bioavailability, KPR-2579 turned out to be hepatotoxic able to cause seizures at doses above 300 mg kg− 1. Compound (2) [16] was able to mitigate such effects, despite being much less potent against human TRPM8 (hTRPM8).
To better understand the mechanism of interaction of these series of α-phenylglycinamides compounds with TRPM8, an extensive in silico study was carried out. Drug-target interactions were estimated by molecular docking and molecular dynamics simulations for a few representative compounds. Activity cliffs [17] were investigated to discover important SAR features and QSAR models were built to predict pIC50 for newly designed compounds. ADMET predictions were extensively applied to all compounds to design safe TRPM8 inhibitors.
A total of 53 hTRPM8 α-phenylglycinamides inhibitors molecular structures were retrieved from the literature [15, 16] to create a 3D structure library. Each compound had its 3D structure, and protonation state built using the MarvinSketch software [18]. All 3D coordinates were optimized using the PM7 semiempirical method using Mopac software [19].
The data obtained were used to calculate the Structure Activity Landscape Index (SALI) equation [20] (Eq. 1).
$$SALI=\frac{\left|{A}_{i}-{A}_{j}\right|}{\left(1-{SIM}_{ij}\right)} \left(1\right)$$
where, Ai and Aj are the biological activity values of two different compounds i and j and SIMij is the Tanimoto similarity between them. Higher values of SALI aided the identification of Activity Cliffs [20, 21].
To make possible the estimation of the ligands’ biological conformation, the hTRMP8 structure PDB ID:6O72 [22] was used to perform molecular dynamics and molecular docking simulations. Short, restrained and simplified molecular dynamics simulations were used better accommodate α-phenylglycinamides to a binding site co-crystallized with a unrelated ligand (TC-I 2014 [22]). AutoDock VINA software [23] was used to estimate the biological posed and score for selected compounds to the relaxed binding sites. The simulations were performed with GROMACS 5 [24] using CHARMM force field [25] for the ligand and protein topologies. The TRPM8 is naturally imbedded in a bilaminar membrane, not simulated by this study. To minimize any artifacts generated by this approximation, the system was simulated in a non-water solvent with dielectric constant of approximately 5, i.e., chloroform, 4.81. TRPM8 coordinated were constrained allowing only the binding side aminoacid side chains and the ligands to move about.
QSAR models were built using 41 compounds (training set) and 12 used for external validation (test set). Molecular interaction field descriptors computed with aligned 3D structures using 3D-QSAR-R [26] to compute van der Waals, electrostatic, hydrophobic and hydrogen bonding descriptors. The descriptors were these potentials interacting with a carbon sp3 probe with point charge + 1 vising lattice points around all compounds. Grid spacing was set to 1 Å. Descriptors with variance less than 0.02 were removed, as well as intercorrelated data (c = 0.9) [27]. 2D descriptors were obtained with PaDEL-descriptor software [28]. Variable selection was used to reduce the complexity of the models as described in the literature [29]. The internal validation of the models was performed using the y-randomization and Leave-N-out tests [30]. The QSAR model applicability domain was assess by the AD-MDI tool [31].
To design new inhibitors, the method of synthesis of the compounds was analyzed [15, 16]. Of the reagents used, various aldehyde containing building blocks were researched, available in the market through the Acros Organics platform (acros.com). Pharmacokinetic parameters and ADME-Tox data were obtained from the SwissADME [32] and ICdrugs [33].
To better understand the Structure-activity relationship, SALI was used in the chosen cluster and 4 activity cliffs were detected (Fig. 2). Cliff C1 suggests that the phenyl group must be in a specific configuration to be more active (3, pIC50 = 5.7). Such configuration is also important for the methyl substituent (7, pIC50 = 6.6). Methyl substituted esters (9, pIC50 = 6.6) and amides (10, pIC50 = 5.0) are more potent when compared carboxyl (8, pIC50 = 4.2) or N,N-dimethyl groups (11, pIC50 = 4.0). The pointed out that ester analogues are readily converted to its less potent derivative carboxylic acid [15], so primary amides are preferable due to the higher stability.
Molecular dynamics and docking simulations were used to estimate how α-phenylglycinamides inhibitors may bind to hTRMP8. We found two uncertain binding poses for ligand (2) used as template. One of the binding poses (Pose 2, Fig. 3) according to the interaction energy with the binding pocket possibly is more favorable when compared to (Pose 1, Fig. 3). A net 28.3 kJ mol− 1 (~ 6.8 kcal mol− 1) is gained upon simultaneous cation-π interaction between ARG832 and ligand (2) aromatic moieties. Furthermore, the primary amine can make two hydrogen bond interaction with ASN732 and ASP772. Pose 1 hydrogen bond interaction are prone to occur with TYR995 and ASP793. Less effective aromatic interaction can be formed with TYR736 (Pose 1) and with PHE1003 (Pose 2).
Once the reasonable binding mode was estimated, all molecules were aligned according to compound (2). 3D and 2D descriptors, once filtered, were concatenated in order to build models with S-MLR with high leave-one-out cross-validation correlation coefficient (Q2LOO > 0.5) [30]. The prediction power was tested against an external data set (Q2EXT, n = 14) to attest its usefulness as hTRMP8 pIC50 activity predictor. The best showed Q2LOO = 0.86 and Q2EXT = 0.75 (Fig. 4).
On Fig. 4 3D descriptors LJ1 demonstrated that the presence of a methyl group in the region is better than the absence of groups. An example of this is the better performance of the (1) molecule in this descriptor. Another information acquired is that the presence of an ethyl group provides worse performance than the absence of groups. This factor shows that small-volume groups are best in this region. The next descriptor, HF1, indicates that the presence of indane brings better activity than benzene alone. This can be observed by the better activity in the region of the (2) molecule when compared to (1). The addition of this group generates a repositioning of the ring that favors its activity, indicating the already expected hydrophobicity of the region. HF2 correlates to the presence of an ester group as observed in the (9), giving the same information as in the cliff C3.
The 2D descriptors part of the model are also list in Fig. 4. The GATS4s is a local spatial correlation index in a molecule whose resulting high values are related to the presence of electronegative atoms. It is also related to the probability of intermolecular interaction in the form of hydrogen bonds. When its value is > 1, electronegative atoms are distributed throughout the molecule and there is no local correlation[34]. The AATSC3s and ATSC8i are weighted average centered Broto-Moreau Autocorrelation, and VR1_Dzv (Randic-like eigenvector-based index) are harder to interpret [35]. The other descriptors such as nsssCH refers to the number of -CH moieties present in each molecule, related possibly to the same interpretation as in HF1.
About 61 (supporting information) new prototypes were designed. All planned compounds were design as readily accessible Uri reaction reactants [16]. Among them, 5 obtained more expressive values (pIC50 > 7.0) (Table 1). Regarding the pharmacokinetic evaluations, all prototypes have the characteristic of high gastrointestinal absorption and two are estimated to have high blood-brain barrier permeation.
In this work, computational methods were applied to design new TRPM8 inhibitors. For this, a literature search was carried out to build the database, perform activity cliff studies, molecular docking and molecular dynamics in the selected molecules. Additionally, a hybrid QSAR model was constructed and computationally validated. This collection of information made it possible to plan and predict the biological activity of a series of prototypes with synthetic accessibility for receptor inhibition.
Acknowledgements The authors thank the Federal University of Rio Grande do Norte (UFRN) for providing the infrastructure and support for this project. The authors are indebted to the High-Performance Computing Center (NPAD) at UFRN for the availability of computational resources and the Instituto Metrópole Digital (IMD) for the support in the realization of this work.
Author contributions CCRM and EGB contributed to the conceptualization, methodology and analysed the data. All authors wrote, review and editing this paper. All the authors have read and approved the manuscript for submission.
Conflict of interest The authors declare that they have no competing financial interests or personal relationships that could to infuence the work reported in this paper.
Ethical approval None of the authors conducted any human or animal experiments for this publication.
Supplementary Information The online version contains supplementary material available at [url].
- Bandell M, Story GM, Hwang SW, et al (2004) Noxious cold ion channel TRPA1 is activated by pungent compounds and bradykinin. Neuron 41:849–857. https://doi.org/10.1016/s0896-6273(04)00150-3
- Bautista DM, Siemens J, Glazer JM, et al (2007) The menthol receptor TRPM8 is the principal detector of environmental cold. Nature 448:204–208. https://doi.org/10.1038/nature05910
- Liu Y, Mikrani R, He Y, et al (2020) TRPM8 channels: A review of distribution and clinical role. Eur J Pharmacol 882:173312. https://doi.org/10.1016/j.ejphar.2020.173312
- Yin Y, Wu M, Zubcevic L, et al (2018) Structure of the cold- and menthol-sensing ion channel TRPM8. Science 359:237–241. https://doi.org/10.1126/science.aan4325
- Beccari AR, Gemei M, Monte M Lo, et al (2017) Novel selective, potent naphthyl TRPM8 antagonists identified through a combined ligand-and structure-based virtual screening approach. Sci Rep 7:1–15. https://doi.org/10.1038/s41598-017-11194-0
- Horne DB, Biswas K, Brown J, et al (2018) Discovery of TRPM8 Antagonist (S)-6-(((3-Fluoro-4-(trifluoromethoxy)phenyl)(3-fluoropyridin-2-yl)methyl)carbamoyl)nicotinic Acid (AMG 333), a Clinical Candidate for the Treatment of Migraine. J Med Chem 61:8186–8201. https://doi.org/10.1021/acs.jmedchem.8b00518
- Jahanfar F, Sadofsky L, Morice A, D’Amico M (2022) Nebivolol as a Potent TRPM8 Channel Blocker: A Drug-Screening Approach through Automated Patch Clamping and Ligand-Based Virtual Screening. Membranes (Basel) 12:. https://doi.org/10.3390/membranes12100954
- Di Donato M, Ostacolo C, Giovannelli P, et al (2021) Therapeutic potential of TRPM8 antagonists in prostate cancer. Sci Rep 11:23232. https://doi.org/10.1038/s41598-021-02675-4
- Huang Y, Li S, Jia Z, et al (2020) Transient Receptor Potential Melastatin 8 (TRPM8) Channel Regulates Proliferation and Migration of Breast Cancer Cells by Activating the AMPK-ULK1 Pathway to Enhance Basal Autophagy. Front Oncol 10:
- Yapa KTDS, Deuis J, Peters AA, et al (2018) Assessment of the TRPM8 inhibitor AMTB in breast cancer cells and its identification as an inhibitor of voltage gated sodium channels. Life Sci 198:128–135. https://doi.org/10.1016/j.lfs.2018.02.030
- Nakanishi O, Fujimori Y, Aizawa N, et al (2020) KPR-5714, a Novel Transient Receptor Potential Melastatin 8 Antagonist, Improves Overactive Bladder via Inhibition of Bladder Afferent Hyperactivity in Rats. J Pharmacol Exp Ther 373:239–247. https://doi.org/10.1124/jpet.119.263616
- Yoon HJ, Kim J, Yang JM, et al (2021) Topical TRPM8 Agonist for Relieving Neuropathic Ocular Pain in Patients with Dry Eye: A Pilot Study. J Clin Med 10:250. https://doi.org/10.3390/jcm10020250
- Aierken A, Xie Y-K, Dong W, et al (2021) Rational Design of a Modality-Specific Inhibitor of TRPM8 Channel against Oxaliplatin-Induced Cold Allodynia. Adv Sci 8:2101717. https://doi.org/10.1002/advs.202101717
- Cao S, Li Q, Hou J, et al (2019) Intrathecal TRPM8 blocking attenuates cold hyperalgesia via PKC and NF-κB signaling in the dorsal root ganglion of rats with neuropathic pain. J Pain Res 12:1287–1296. https://doi.org/10.2147/JPR.S197168
- Kobayashi J ichi, Hirasawa H, Ozawa T, et al (2016) Synthesis and optimization of novel α-phenylglycinamides as selective TRPM8 antagonists. Bioorganic Med Chem 25:727–742. https://doi.org/10.1016/j.bmc.2016.11.049
- Kobayashi J ichi, Hirasawa H, Fujimori Y, et al (2021) Identification of N-acyl-N-indanyl-α-phenylglycinamides as selective TRPM8 antagonists designed to mitigate the risk of adverse effects. Bioorganic Med Chem 30:115903. https://doi.org/10.1016/j.bmc.2020.115903
- Cruz-Monteagudo M, Medina-Franco JL, Pérez-Castillo Y, et al (2014) Activity cliffs in drug discovery: Dr Jekyll or Mr Hyde? Drug Discov Today 19:1069–1080. https://doi.org/10.1016/j.drudis.2014.02.003
- Marvin Sketch Group R (2013) Marvin Sketch. In: https://www.chemaxon.com/products/marvin/
- James J. P. Stewart, Stewart Computational Chemistry, Colorado Springs, CO U MOPAC2016
- Guha R (2012) Exploring Structure-Activity Data Using the Landscape Paradigm. Wiley Interdiscip Rev Comput Mol Sci 2:10.1002/wcms.1087. https://doi.org/10.1002/wcms.1087
- Bajorath J (2017) Representation and identification of activity cliffs. Expert Opin Drug Discov 12:879–883. https://doi.org/10.1080/17460441.2017.1353494
- Diver MM, Cheng Y, Julius D (2019) Structural insights into TRPM8 inhibition and desensitization. Science (80- ) 365:1434–1440. https://doi.org/10.1126/science.aax6672
- Trott O, Olson AJ (2010) AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem 31:455–461. https://doi.org/10.1002/jcc.21334
- David Van Der S, Erik L, Berk H, et al (2005) GROMACS: Fast, flexible, and free. J Comput Chem 26:1701–1718
- K. V, E. H, C. A, et al (2009) CHARMM general force field: A force field for drug‐like molecules compatible with the CHARMM all‐atom additive biological force fields. J Comput Chem 31:671–690. https://doi.org/10.1002/jcc.21367
- Patil RB, Barbosa EG, Sangshetti JN, et al (2018) LQTA-R: A new 3D-QSAR methodology applied to a set of DGAT1 inhibitors. Comput Biol Chem 74:. https://doi.org/10.1016/j.compbiolchem.2018.02.021
- Barbosa EG, Ferreira MMC (2012) Digital Filters for Molecular Interaction Field Descriptors. Mol Inform 31:75–84. https://doi.org/10.1002/minf.201000181
- Yap CW (2011) PaDEL-descriptor: An open source software to calculate molecular descriptors and fingerprints. J Comput Chem 32:1466–1474. https://doi.org/10.1002/jcc.21707
- de Araújo Santos R, Braz C, Ghasemi J, et al (2015) Mixed 2D–3D-LQTA-QSAR study of a series of Plasmodium falciparum dUTPase inhibitors. Med Chem Res 24:1098–1111. https://doi.org/10.1007/s00044-014-1189-4
- Kiralj R, Ferreira M (2009) Basic validation procedures for regression models in QSAR and QSPR studies: theory and application. J Braz Chem Soc 20:770–787. https://doi.org/http://dx.doi.org/10.1590
- Yan J, Zhu WW, Kong B, et al (2014) A combinational strategy of model disturbance and outlier comparison to define applicability domain in quantitative structural activity relationship. Mol Inform 33:503–513. https://doi.org/10.1002/minf.201300161
- Daina A, Michielin O, Zoete V (2017) SwissADME: A free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci Rep 7:1–13. https://doi.org/10.1038/srep42717
- Wei M, Zhang X, Pan X, et al (2022) HobPre: accurate prediction of human oral bioavailability for small molecules. J Cheminform 14:1. https://doi.org/10.1186/s13321-021-00580-6
- Fagerberg JH, Bergström CA (2015) Intestinal solubility and absorption of poorly water soluble compounds: predictions, challenges and solutions. Ther Deliv 6:935–959. https://doi.org/10.4155/tde.15.45
- Roy K, Das R (2014) A Review on Principles, Theory and Practices of 2D-QSAR. Curr Drug Metab 15:. https://doi.org/10.2174/1389200215666140908102230
No competing interests reported.