International Journal of Molecular Sciences Article Understanding the Molecular Basis of 5-HT4 Receptor Partial Agonists through 3D-QSAR Studies Alejandro Castro-Alvarez 1, * , Emigdio Chávez-Ángel 2 and Ronald Nelson 3, * 1 Laboratorio de Bioproductos Farmacéuticos y Cosméticos, Centro de Excelencia en Medicina Traslacional, Facultad de Medicina, Universidad de La Frontera, Av. Francisco Salazar 01145, Temuco 4780000, Chile 2 Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193 Barcelona, Spain; emigdio.chavez@icn2.cat 3 Departamento de Química, Facultad de Ciencias, Universidad Católica del Norte, Av. Angamos 0610, Antofagasta 1270709, Chile * Correspondence: alejandro.castro.a@ufrontera.cl (A.C.-A.); rnelson@ucn.cl (R.N.) Abstract: Alzheimer’s disease (AD) is a neurodegenerative disorder whose prevalence has an incidence in senior citizens. Unfortunately, current pharmacotherapy only offers symptom relief for patients with side effects such as bradycardia, nausea, and vomiting. Therefore, there is a present need to provide other therapeutic alternatives for treatments for these disorders. The 5-HT4 receptor is an attractive therapeutic target since it has a potential role in central and peripheral nervous system disorders such as AD, irritable bowel syndrome, and gastroparesis. Quantitative structure-activity relationship analysis of a series of 62 active compounds in the 5-HT4 receptor was carried out in the present work. The structure-activity relationship was estimated using three-dimensional quantitative structure-activity relationship (3D-QSAR) techniques based on these structures’ field molecular   (force and Gaussian field). The best force-field QSAR models achieve a value for the coefficient of determination of the training set of R2 training = 0.821, and for the test set R2 test = 0.667, while for Citation: Castro-Alvarez, A.; Chávez-Ángel, E.; Nelson, R. Gaussian-field QSAR the training and the test were R2 training = 0.898 and R2 test = 0.695, respectively. Understanding the Molecular Basis of The obtained results were validated using a coefficient of correlation of the leave-one-out cross- 5-HT4 Receptor Partial Agonists validation of Q2 LOO = 0.804 and Q2 LOO = 0.886 for force- and Gaussian-field QSAR, respectively. through 3D-QSAR Studies. Int. J. Mol. Based on these results, novel 5-HT4 partial agonists with potential biological activity (pEC50 8.209– Sci. 2021, 22, 3602. https://doi.org/ 9.417 for force-field QSAR and 9.111–9.856 for Gaussian-field QSAR) were designed. In addition, for 10.3390/ijms22073602 the new analogues, their absorption, distribution, metabolism, excretion, and toxicity properties were also analyzed. The results show that these new derivatives also have reasonable pharmacokinetics Academic Editor: Ashu Johri and drug-like properties. Our findings suggest novel routes for the design and development of new 5-HT4 partial agonists. Received: 1 March 2021 Accepted: 26 March 2021 Keywords: Alzheimer’s disease; 5-HT4 ; partial agonist; 3D-QSAR; force and gaussian fields Published: 30 March 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in 1. Introduction published maps and institutional affil- iations. Alzheimer’s disease (AD) is a neurodegenerative disorder that mainly affects people over 60 years old. The current pharmacotherapy only provides palliative treatments, reducing the associated symptoms through the increase of cholinergic function. This pharmacotherapy can produce unwanted side effects such as abdominal pain, muscle Copyright: © 2021 by the authors. cramps, tremors, and fatigue, among others [1]. In this sense, there is a need for new Licensee MDPI, Basel, Switzerland. therapeutic targets for the treatment of this disorder. This article is an open access article The 5-HT4 receptor (5-HT4 R) belongs to a superfamily of G-protein coupled receptors distributed under the terms and (GPCRs) [2–4]. This receptor is highly expressed in the brain regions of the hippocampus, conditions of the Creative Commons amygdala, and cerebral cortex, areas of the brain related to short- and long-term memory Attribution (CC BY) license (https:// and cognitive processing, so that deterioration of this region would be associated with creativecommons.org/licenses/by/ neurological diseases such as Alzheimer’s disease [5,6]. The 5-HT4 R has been reported to 4.0/). play an essential role in disorders of the central nervous system (CNS) such as AD [7,8], Int. J. Mol. Sci. 2021, 22, 3602. https://doi.org/10.3390/ijms22073602 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2021, 22, 3602 2 of 18 peripheral nervous system (PNS) disorders [9], irritable bowel syndrome [10–12], and gastroparesis [13–15]. Moreover, 5-HT4 R agonists modulate peptides derived from the soluble amyloid precursor protein-α (a non-amyloidogenic protein) that plays a role in neuroprotection against the neurotoxic effects of β-amyloid [16]. Therefore, 5-HT4 R partial agonists show very promising activity for symptomatic treatments of cognitive disorders in AD [17]. Its dual mechanism of action in treating AD and other cognition-related diseases makes 5-HT4 R a very attractive target for new drug discovery. Consequently, several structurally diverse heteroaromatic compounds [18–21] have been explored as 5-HT4 R total or partial agonists for both CNS and PNS. Nirogi et al. reported a series of 5-HT4 R compounds with 3-isopropylimidazo [1,5-a]-pyridine-carboxamide scaffold, most of which showed cognition-enhancing properties in animal models [22]. However, their absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties were not satisfactory due to their low ability to penetrate the blood-brain barrier. Their results revealed that these molecules are composed of an aromatic fragment, a coplanar functional group, and a bulky substituent. Recently, Nirogi reported new 5-HT4 R partial agonists with good ADMET properties and potential drug candidates [23]. To design new 5-HT4 R agonists, theoretical studies are substantially essential to expedite and save resources. Several computational methods simplify the drug discovery process. Quantitative structure-activity relationship (QSAR) is a ligand-based drug design method, which relates to the biological activity of compounds with several physicochemical properties [24]. However, QSAR techniques have limited efficacy for designing new functional molecules due to the lack of three-dimensional (3D) molecules’ structures. Consequently, 3D-QSAR averts this problem by using the 3D-attributes of ligands and chemometric tools. That significantly improves the predictability of the biological activity of the model [25–28]. In this work, we present a computational study of a three-dimensional quantitative structure-activity relationship (3D-QSAR) of a set of molecules with agonist activity on 5-HT4 receptors. The calculations were carried out by using force- and Gaussian-field based QSAR models. Our 3D-QSAR study aims to obtain helpful information to guide future 5-HT4 R agonists’ design with promising therapeutic activity and that these new analogues have good ADMET properties as prospective drug candidates. 2. Results and Discussion 2.1. Studied Compounds The studied dataset was based on Brodney et al. [18] and Nirogi et al. [22,23]. They reported different compounds with biological activity (5-HT4 receptor partial agonist.) expressed in EC50 in nanomolar concentration (see Table S1 of the Supplementary Mate- rial). In total, 62 compounds were divided in training (43 compounds) and test dataset (19 compounds), as is shown in Figure 1. The biological activity was expressed in terms of pEC50 for this study. Finally, the ADMET properties were estimated by using the pkCSM [29,30] and SwissADME [31,32] web services. On one hand, pkSCM calculates the pharmacokinetic properties using structural similarity of the new molecules with molecules with known pharmacokinetic properties and, from this comparison, the pkSCM program returns es- timated values for the new molecules. On the other hand, SwissADME calculates the different drug-likeness parameters by calculating physicochemical descriptors for each of the designed molecules. These descriptors are lipophilic, water-solubility, surface volume, among others (an extended description can be found in Section S3 of the Supplementary Material). From these descriptors, the program generates an estimate of drug-likeness based on parameters from Lipinski [33], Ghose [34], Veber [35], and Egan [36]. Int. J. Mol. Sci. 2021, 22, 3602 3 of 18 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 3 of 19 O O O N N N OH OH O N N O O O O O O NH NH NH NH NH F N F N O O N N N N N N N 1* 2* 3* 4 5 O N O N N HO N N N N O O N O O O O OH NH Cl NH N N N N N H2N O N N N N 6* 7 8* 9 10 O OH N N N N N N O O N O O O NH O NH NH NH N N N N N N N N N N 11 12 13 14 15 O O N N O N N N OH O NH O F N N O O NH NH N O N O NH N N N O H N N N N N N 16* 17* 18 19 20 O N N N N O OH O O OH O F O NH NH NH NH N N N H N N N N N N S N O N 21* 22 23* 24 25 OH N N N O N N O O O N NH NH NH O N N N N N N N N 26 27 28 29* O N H + N O O O NH NH O N N Cl NH2 30 31 Figure 1. Cont. Int. J. Mol. Sci. 2021, 22, 3602 4 of 18 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 4 of 19 O HO O N N N O N HO N N O O O N N NH NH O N O O NH N N N N N N N N N N 32 33* 34 35 36 O O O N N O OH H N + N N N O O N O O NH O NH O NH O NH NH N N N N N N N N N N N 37* 38 39 40 41 O O N N O N O N O N N N O O NH O NH O NH O NH N N N N N N N N N N N N 42 43 44* 45* 46 N O N N N N OH N O N N O NH N O O NH N O NH O O N N N N N N N N N N 47 48 49 50 51 O O N N N O N N N O O NH O NH N O NH O NH O N N N N N N N N N N 52* 53 54* 55 56* HO O OH N N O N N O O O O NH O NH O NH O O N N N N N O N N 57 58 59* Cl 60 O N O N N N O O NH N N N N 61* 62 Figure 1. Figure 1. Dataset Dataset of 62 compounds. of 62 compounds. Structures with an Structures with an asterisk asterisk (*) (*) were were used used as as aa test test dataset. dataset. 2.2. Statistical Results The statistical results for force- and Gaussian-field QSAR (FFQSAR and GFQSAR, respectively) are presented in Tables 1 and and 2. 2. All possible field combinations were tested for both FFQSAR and GFQSAR. In the case of FFQSAR, the combination of the steric and Int. J. Int. J. Mol. Mol. Sci. Sci. 2021, 22, 3602 2021, 22, x FOR PEER REVIEW 55 of 19 of 18 electrostatic fields was statistically significant (see Table 1) with2R2training of 0.821 and R 2test electrostatic fields was statistically significant (see Table 1) with R training of 0.821 and R2 test 0.667. The GFQSAR models with the highest R2 test values were those that considered the 2 0.667. The GFQSAR models with the highest R test values were those that considered the field combinations (see Figure 2 and Table 2). The best model with the highest R2training 2 and field combinations (see Figure 2 and Table 2). The best model with the highest R training R2test was chosen for the study. This model presented steric (0.420), electronic (0.125), ac- and R2 test was chosen for the study. This model presented steric (0.420), electronic (0.125), ceptor (0.304), and donor hydrogen-bond (0.151) contribution, with a correlation between acceptor (0.304), and donor hydrogen-bond (0.151) contribution, with a correlation between experimental and predicted data showing R22training 0.898 and with an external validation experimental and predicted data showing R 0.898 and with an external validation 0.695 (R22test) (see Table S2). The experimentaltraining activities, the predicted values, and the re- 0.695 (R sidual test ) for values (seethis Table S2).are model The experimental shown in Table 3.activities, the predicted All the compounds showedvalues, and the low residual residual values for this model are shown in Table 3. All the compounds showed low values with a range from −1.1 to 1.5 for FFQSAR and −1.2 to 1.2 for GFQSAR. residual values with a range from −1.1 to 1.5 for FFQSAR and −1.2 to 1.2 for GFQSAR. Table 1. Summary of statistical results from force field quantitative structure-activity relationship (FFQSAR) and field Table 1. Summary of statistical results from force field quantitative structure-activity relationship (FFQSAR) and field contributions. contributions. Fraction of Fields Included in the Model Fields SD R2training R2Scramble R2test Stability Fraction Steric of Fields Included in the Model Electrostatic Fields SD 2 2 R2 test0.229 Stability S 0.654 R R 0.719 training 0.554 Scramble 0.329 1 Steric Electrostatic ES 0.633 0.654 0.7370.719 0.230 0.554 0.314 0.3290.614 0.229 1 1 E All 0.633 0.522 0.8210.737 0.188 0.230 0.667 0.3140.120 0.614 0.574 1 0.426 All SD: Standard 0.522 deviation of0.821 the regression. 0.188 0.667 for the 0.120 R2training is the value 0.574 regression (the coefficient of determination)0.426 of the 2 training SD: set. R Standard 2 scramble isof deviation the theaverage value regression. of R from 2 R training is theavalue seriesforofthe models built(the regression using scrambled coefficient activities; thisofvalue of determination) measures the training set. R 2 is the averagethe value of R2 from a series the degree scramble to which molecular fields canoffitmodels built using meaningless scrambled data. The valueactivities; this of R2test value for the measures predictedthe degree toon activities which the the test molecular set. Stabilityfieldsofcanthefitmodel meaningless data. The predictions tovalue changesof R2in for the test the predicted training set activities on the The composition. test set. Stability steric (S) andof the model predictions electrostatic to (E) field changes in the training set composition. The steric (S) and electrostatic (E) field contributions in each model. contributions in each model. Figure 2. Figure 2. The The results results of of the the distribution distribution of of R R22test test values were obtained from 31 combinations of Gaussian fields. S, steric; E, values were obtained from 31 combinations of Gaussian fields. S, steric; E, electrostatic; H, hydrophobic; A, hydrogen-bond electrostatic; H, hydrophobic; A, hydrogen-bond acceptor;acceptor; D: D: hydrogen-bond hydrogen-bond donor. donor. The The best best Gaussian Gaussian fields fields combination combination is highlighted with an asterisk on the bars with R22training 0.898 and R2test 2 0.695. is highlighted with an asterisk on the bars with R training 0.898 and R test 0.695. A A detailed detailed discussion discussionabout aboutthe thebest bestFFQSAR FFQSAR andandGFQSAR GFQSAR models’ internal models’ andand internal ex- ternal validation parameters is presented in Section 2.4 of Supplementary Material. external validation parameters is presented in Section S2.4 of Supplementary Material. Int. J. Mol. Sci. 2021, 22, 3602 6 of 18 Table 2. Summary of statistical results from Gaussian field quantitative structure-activity relationship (GFQSAR) and field contributions. Field Contributions Fields SD R2 training R2 Scramble R2 test Stability S E H A D S 0.742 0.639 0.681 0.229 0.507 1.000 E 0.362 0.914 0.769 0.573 0.688 1.000 H 0.498 0.837 0.804 0.460 0.341 1.000 A 0.567 0.789 0.645 0.548 0.587 1.000 D 0.810 0.570 0.382 0.178 0.362 1.000 SE 0.563 0.792 0.765 0.557 0.017 0.708 0.293 SH 0.565 0.791 0.784 0.405 0.019 0.509 0.491 SA 0.507 0.831 0.790 0.319 0.253 0.534 0.466 SD 0.572 0.785 0.763 0.657 0.226 0.675 0.325 EH 0.343 0.923 0.862 0.577 0.431 0.304 0.696 EA 0.479 0.849 0.743 0.656 0.593 0.345 0.655 ED 0.431 0.878 0.702 0.548 0.475 0.609 0.391 HA 0.387 0.902 0.835 0.615 0.453 0.561 0.440 HD 0.473 0.853 0.824 0.405 0.217 0.734 0.266 AD 0.504 0.833 0.681 0.526 0.328 0.710 0.290 SEH 0.483 0.847 0.817 0.560 0.141 0.417 0.170 0.413 SHA 0.443 0.871 0.824 0.628 0.225 0.356 0.326 0.318 SHD 0.498 0.838 0.817 0.422 0.017 0.416 0.375 0.210 SEA 0.467 0.857 0.807 0.681 0.250 0.466 0.147 0.387 SED 0.498 0.837 0.799 0.492 0.027 0.555 0.194 0.251 SAD 0.486 0.845 0.812 0.617 0.148 0.470 0.356 0.175 EHA 0.332 0.928 0.846 0.604 0.454 0.171 0.481 0.348 EHD 0.368 0.911 0.846 0.627 0.344 0.230 0.576 0.194 EAD 0.455 0.864 0.739 0.611 0.447 0.275 0.504 0.221 HAD 0.377 0.907 0.844 0.601 0.346 0.490 0.366 0.145 SEHA 0.410 0.890 0.834 0.670 0.257 0.321 0.109 0.295 0.275 SEHD 0.443 0.871 0.836 0.556 0.099 0.365 0.131 0.331 0.172 SEAD 0.442 0.898 0.826 0.695 0.172 0.420 0.125 0.304 0.151 SHAD 0.430 0.879 0.840 0.644 0.186 0.328 0.281 0.260 0.132 EHAD 0.356 0.917 0.850 0.639 0.396 0.145 0.426 0.305 0.125 SEHAD 0.395 0.898 0.847 0.674 0.213 0.302 0.095 0.258 0.228 0.117 R2 SD is the standard deviation of the regression. training is the value for the regression (the coefficient of determination) of the training set. R2 scramble is the average value of R2 from a series of models built using scrambled activities; this value measures the degree to which the molecular fields can fit meaningless data. The value of R2 test for the predicted activities on the test set. Stability of the model predictions to changes in the training set composition. The steric (S), electronic (E), hydrophobic (H), hydrogen-bond donor (D), and hydrogen-bond acceptor (A) field contributions in each model. Table 3. Experimental and calculated pEC50 and residual values for the analyzed compounds obtained with the force-field QSAR (FFQSAR) and Gaussian-field QSAR (GFQSAR) model. The highlighted rows show the test set compounds. FFQSAR GFQSAR FFQSAR GFQSAR Comp. pEC50_exp pEC50_calc Res. pEC50_calc Res. Comp. pEC50_exp pEC50_calc Res. pEC50_calc Res. 1 8.921 8.996 −0.075 8.404 0.518 32 7.246 6.825 0.421 6.993 0.253 2 8.585 8.836 −0.251 9.127 −0.542 33 6.842 7.797 −0.955 8.065 −1.223 3 7.398 8.416 −1.018 7.211 0.187 34 6.315 6.676 −0.361 6.240 0.076 4 7.509 7.745 −0.236 7.716 −0.207 35 7.268 6.893 0.375 7.441 −0.173 5 5.648 6.033 −0.385 5.894 −0.246 36 6.942 6.924 0.018 7.180 −0.238 6 6.284 6.418 −0.134 6.683 −0.399 37 8.310 8.321 −0.011 8.586 −0.276 7 6.331 5.984 0.347 6.177 0.154 38 9.523 9.551 −0.028 9.631 −0.108 8 6.133 6.173 −0.040 6.098 0.035 39 8.060 7.735 0.325 7.305 0.755 9 8.699 8.736 −0.037 8.733 −0.034 40 8.076 8.292 −0.216 7.911 0.165 10 7.141 8.076 −0.935 7.222 −0.081 41 7.102 8.080 −0.978 8.083 −0.981 11 7.703 7.675 0.028 7.967 −0.264 42 6.223 5.805 0.418 5.591 0.632 12 6.067 6.801 −0.734 6.559 −0.492 43 5.712 6.893 −1.181 6.000 −0.288 13 8.244 7.459 0.785 8.088 0.156 44 7.983 7.459 0.524 7.512 0.471 Int. J. Mol. Sci. 2021, 22, 3602 7 of 18 Table 3. Cont. FFQSAR GFQSAR FFQSAR GFQSAR Comp. pEC50_exp pEC50_calc Res. pEC50_calc Res. Comp. pEC50_exp pEC50_calc Res. pEC50_calc Res. 14 8.000 7.932 0.068 8.196 −0.196 45 5.712 5.226 0.486 6.499 −0.787 15 8.009 7.568 0.441 7.775 0.234 46 8.824 9.078 −0.254 9.024 −0.200 16 8.824 8.294 0.530 8.722 0.102 47 9.155 8.822 0.333 8.597 0.558 17 8.469 7.331 1.138 7.541 0.928 48 7.208 7.169 0.039 6.673 0.535 18 9.301 8.681 0.620 9.184 0.117 49 7.009 6.466 0.543 7.435 −0.426 19 6.301 6.646 −0.345 5.982 0.319 50 6.120 6.400 −0.280 6.303 −0.183 20 5.867 6.333 −0.466 6.292 −0.425 51 9.222 9.270 −0.048 9.368 −0.146 21 7.658 7.808 −0.150 7.641 0.017 52 7.866 7.670 0.196 7.525 0.341 22 7.237 7.235 0.002 7.383 −0.146 53 10.000 8.778 1.222 8.970 1.030 23 7.469 8.233 −0.764 7.944 −0.475 54 8.046 8.168 −0.122 7.983 0.063 24 7.745 7.490 0.255 7.766 −0.021 55 6.099 6.030 0.069 5.722 0.377 25 7.738 7.225 0.513 7.848 −0.110 56 8.056 7.701 0.355 8.100 −0.044 26 7.409 7.349 0.060 7.366 0.043 57 8.886 8.488 0.398 9.006 −0.120 27 7.301 7.367 −0.066 7.527 −0.226 58 9.523 8.838 0.685 8.928 0.595 28 7.959 8.241 −0.282 7.843 0.116 59 9.398 9.487 −0.089 9.815 −0.417 29 6.076 6.727 −0.651 6.162 −0.086 60 5.963 6.700 −0.737 6.536 −0.573 30 8.319 7.986 0.333 7.491 0.829 61 7.678 6.177 1.501 6.403 1.275 31 8.284 8.662 −0.378 8.659 −0.375 62 7.377 7.497 −0.120 8.061 −0.684 Res.: Residual value. Comp.: Compound number 2.3. Analysis of the 3D-QSAR Models The dataset of 62 compounds was randomly separated into a training set (43 compos- ites) and a test set (19 composites). The training set was used to run different 3D-QSAR models with FFQSAR and GFQSAR (different field combinations). The best models were evaluated with the test set. The visualization of the best 3D-QSAR models were analyzed by recognizing the colored regions highlighting the favored and disfavored areas that explain the 5-HT4 R partial agonist activity of the different compounds studied. For FFQSAR, the best model was achieved with the combination of steric and electrostatic fields with R2 0.821 for the training set and R2 0.667 for the test set, respectively. 2.3.1. Force-Field Based 3D-QSAR Model—Steric and Electrostatic Contour Map Green and yellow colors represent the force field-based steric interactions. The green and the yellow regions represent zones where bulky substituents’ addition can increase or decrease activity, respectively (see Figure 3A). In the most active molecule (compound 53) the green contours surround the tetrahydropyran and the heterocyclic aromatic ring. In contrast, the less active compounds have an aromatic ring, whose orientation is in the yellow regions, resulting in decreased biological activity. Figure 3B shows field-based electrostatic interactions which are represented by red (electronegative) and blue (electropositive) contours. The regions coloured in blue and red represent the most influential electropositive and electronegative zones in biological activity. One of the largest regions for electronegative interactions is around the tertiary amine of the piperidine and the amide’s carbonyl. Moreover, electropositive regions corresponding to the nitrogen atoms of the oxadiazole ring are present in some compounds. The less active molecules of the dataset have a disfavored conformation due to the imidazo[1,5-a]pyridine heterocyclic ring. We observed that all compounds with low partial 5-HT4 R agonist activity had a chain length from amide to morpholine N of three carbon atoms. It suggests that this distance is responsible for the decrease in biological activity. Mol. Sci. 2021, 22, x FOR PEER REVIEW 8 of 19 Int. J. Mol. Sci. 2021, 22, HT 3602 4R agonist activity had a chain length from amide to morpholine N of three carbon 8 of 18 atoms. It suggests that this distance is responsible for the decrease in biological activity. Figurecontour Figure 3. (A) Steric 3. (A) Steric contour and (B) and (B) electrostatic electrostatic contour contour maps for maps the best for the based force-field best force-field 3D-QSARbasedmodel.3D-The active molecules areQSAR shownmodel. The in sticks foractive molecules (A) and are shownThe (B), respectively. in sticks for (A) favorable andand (B), respectively. disfavored regions ofThe the favorable steric field shown and disfavored in (A) are highlighted regions in green of the steric and yellow, field shown respectively, whilein the (A) are highlighted electrostatic in green regions are and shownyellow, respec- in (B), the favorable tively, while the electrostatic regions are shown in (B), the favorable electropositive and unfavorable electropositive and unfavorable electronegative regions are highlighted in blue and red, respectively. electronegative regions are highlighted in blue and red, respectively. 2.3.2. Gaussian-Field 2.3.2. Gaussian-Field Based 3D-QSAR Based 3D-QSAR Model—Steric Model—Steric Contour MapContour Map The GFQSAR was generated using five-factors partial least squares (PLS) and correlat- The GFQSAR was generated using five-factors partial least squares (PLS) and corre- ing four fields: steric (S), electrostatic (E), hydrogen bond donors (HBD), and hydrogen lating four fields: steric (S), electrostatic2 (E), hydrogen bond donors (HBD), and hydrogen bond acceptor (HBA). A Q value of 0.886 was derived from the leave-one-out (LOO) cross- bond acceptor (HBA). A Q2 value of 0.886 was derived from the leave-one-out (LOO) validation method. A non-cross-validation analysis yielded an R2 = 0.898 with a standard cross-validation method. A non-cross-validation analysis yielded an R2 = 0.898 with a error of estimate SD = 0.377 and F ratio of 360.13 (see Table S2). The steric, electrostatic, standard error of estimate SD = 0.377 and F ratio of 360.13 (see Table S2). The steric, elec- HBA, and HBD fields contributions (ranging from 0 to 1) were 0.420, 0.125, 0.304, and trostatic, HBA, and HBD fields contributions (ranging from 0 to 1) were 0.420, 0.125, 0.304, 0.151, respectively (see Table 2). The field contributions of steric (0.420) and HBA (0.304) and 0.151, respectively (see Table 2). The field contributions of steric (0.420) and HBA intensities were higher than the electrostatic (0.125) and HBD (0.151), indicating a larger (0.304) intensities were higher than the electrostatic (0.125) and HBD (0.151), indicating a requirement of steric and hydrogen-bond acceptor for protein-ligand interactions. larger requirement of steric and hydrogen-bond acceptor for protein-ligand interactions. The steric interactions represented in green and yellow contour are shown in Figure 4. The steric interactions represented in green and yellow contour are shown in Figure The green regions of the molecules represent the favorable effect of the bulky substituents, 4. The green i.e., regions of the at these molecules positions, therepresent the favorable bulky groups will haveeffect of the higher bulkyConversely, activity. substitu- the yellow ents, i.e., at these positions, outlines representthethe bulky groups regions will where have the bulkyhigher groupsactivity. Conversely, will reduce activity.the yellow outlines represent the regions where the bulky + groups will reduce activity. For case 53 (Figure 4A,B), the S 2 contour (green contour) shows a favorable activity For casecloser 53 (Figure to the4A,B), the Sof isopropyl +2 contour (green contour) shows a favorable activity the aromatic ring. This trend was also observed for the highly closer to the active isopropyl of the 38, molecules aromatic 58, andring. This 59. On thetrend otherwas also hand, anobserved opposite for the highly behavior was observed for active molecules 38, 58, and 59. On the other hand, an opposite behavior was observed moderately and less active molecules with substitutions with different orientations for on the moderately and less active aromatic ring, molecules as is compoundwith substitutions 5 (Figure 4C,D),with different that a molecule orientations shows a on the biological deficient aromatic ring, as is compound activity, 5 (Figure whose isopropyl 4C,D), a molecule orientation is opposite that shows to that a deficient biologi- of compound 53. Similar behavior cal activity, whose isopropyl orientation is opposite to that of compound was also observed for compounds 45, 43, and 20 with imidazo[1,5-a]pyridine 53. Similar be- ring. havior was also observed for compounds 45, 43, and 20 with imidazo[1,5-a]pyridine ring. Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW Int. J. Mol. Sci. 2021, 22, 3602 9 of 18 Figure4. Figure Contour maps 4. Contour maps obtained obtainedfor the forbest theGaussian-based 3D-QSAR 3D-QSAR best Gaussian-based model steric model (hydrogensteric (hyd bond donors, HD), green and yellow regions indicate a favorable and unfavorable steric bond donors, HD), green and yellow regions indicate a favorable and unfavorable steric inter interaction, respectively. An respectively. Anactive molecule active molecule(53) (53) is represented in (A) (sticks is represented in (A)representation) and (B) (as draw (sticks representation) and (B) (a representation). Less active molecule (5) is shown in (C) (sticks representation) and (D) (as draw representation). Less active molecule (5) is shown in (C) (sticks representation) and (D) (as representation). representation). 2.3.3. Gaussian-Field Based 3D-QSAR Model—Electrostatic Contour Map 2.3.3. Gaussian-Field Based As for the electrostatic 3D-QSAR contour Model—Electrostatic maps, the blue (E+ 1 , E+ 2 , E+ 3 andContour Map E+ 4 ) and red (E− 1 , E− 2 , E− 3 and E− 4 ) contours represent the favorable and unfavorable As for the electrostatic contour maps, the blue (E+1, E+2, components E+3 and E+4of the ) and red (E electrostatic field (Figure 5). The bulky E+ 1 contour surrounding the tetrahydropyran ring E−(Figure 3 and E−4) contours represent the favorable and unfavorable components of the el 5A,B) provides information on analogues that have electron-donating substituents static that field favor (Figure biological5). The bulky activity, such asE+compounds 1 contour surrounding the 50 18, 38, and 59 (pEC tetrahydropyran 9.301, 9.523, and ring (F 9.398, respectively), compounds that have a tertiary hydroxyl 5A,B) provides information on analogues that have electron-donating group on the tetrahydropyran substituent ring (compounds 18 and 58). The electropositive E + contour highlights the importance of favor biological activity, such as compounds 18, 38, and 59 (pEC50 9.301, 9.523, 3 and amide hydrogen for biological activity (contour that reappears in the study of hydrogen- respectively), compounds that have a tertiary hydroxyl group on the tetrahydropyra bonding donor groups, Section 2.3.5). (compounds 18 and 58). The electronegative E−The electropositive − E+3 contour highlights the importance o 3 and E 4 contours highlight the importance of the nitrogen ide hydrogen specificity of thefor biological aromatic activity ring that (contour would provide that density electron reappears in thebiological and answer study of hydr activity; however, the E − contour stands out (Figure 5A,B), in proportion, more than the bonding donor groups,3 Section 2.3.5). E− 4 contour, therefore, compounds that have a nitrogen in E− 3 are more active than those The electronegative − E−3 and E−4 contours highlight the importance of the nit that have a nitrogen in E 4 (Figure 5C,D). Finally, the electropositive contours E+ 2 and specificity E+ 4 wouldof the aromatic respond to compoundsring with that 1,3,4-oxadiazole would provide electron groups; density however, theirand answer biol biological activity; however, activity is moderate the E 3(61 to low − contour with pEC stands out 50 7.678, 48 (Figure with pEC5A,B), 50 7.208 in andproportion, 35 with pEC50 more th −7.268, see Table 3). E 4 contour, therefore, compounds that have a nitrogen in E 3 are more active than − that have a nitrogen in E−4 (Figure 5C,D). Finally, the electropositive contours E+2 a would respond to compounds with 1,3,4-oxadiazole groups; however, their biologi tivity is moderate to low (61 with pEC50 7.678, 48 with pEC50 7.208 and 35 with pEC50 see Table 3). Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 1 Int. J. Mol. Sci. 2021, 22, 3602 10 of 18 Figure 5. Contour maps obtained for the best Gaussian-based 3D-QSAR model with electrostatic Figure 5. Contour maps obtained for the best Gaussian-based 3D-QSAR model with elect interaction (HD), blue and red regions indicate favorable electropositive and electronegative inter- interaction (HD), blue and red regions indicate favorable electropositive and electronegativ actions, respectively. Active molecule (53) is represented in (A) (sticks representation) and (B) (as actions, respectively. Active molecule (53) is represented in (A) (sticks representation) and draw representation). Less active molecule (5) is represented in (C) (sticks representation) and (D) draw representation). (as draw representation). Less active molecule (5) is represented in (C) (sticks representation) a (as draw representation). 2.3.4. Gaussian-Field Based 3D-QSAR Model—Hydrogen Bond Acceptor Contour Maps 2.3.4. Hydrogen bond acceptor Gaussian-Field Based functional 3D-QSARgroups provide properties Model—Hydrogen that determine Bond Acceptorthe Contour biological activity of a drug candidate compound. Therefore, the contour map obtained withHydrogen the GFQSAR bond model acceptor functional establishes which HBA groups regionsprovide would helpproperties that determi in the biological biological activity of activity of a Figure the molecule. drug candidate 6 shows the compound. impact of the Therefore, HBA groups the contour on the 5-HT4 Rmap ob partial agonist activity of the molecules. The regions of highest with the GFQSAR model establishes which HBA regions would help in the and lowest affinity are biologi shown with red and magenta contours, respectively. In general, in all the studied molecules, tivity of the molecule. Figure 6 shows the impact of the HBA groups on the 5-HT4R the tetrahydropyran group is responsible for establishing hydrogen bonds (HA+ 3 contour). agonist activity of the molecules. + The regions of highest and lowest The large affinity region (HA 1 contour) near the amide carbonyl suggests that it favours affinity are with 5-HTred and magenta 4 R partial contours, agonist activity. respectively. In contrast, In general, fewer affinity in all the regions indicate thatstudied functionalmolecul groups attached directly tetrahydropyran group to the is amide, such as for responsible hydroxyl on the piperidine establishing ring or hydrogen replacing bonds (HA+3 con the piperidine with morpholine will reduce biological activity such as compounds 5, 10, The large affinity region (HA+1 contour) near the amide carbonyl suggests that it fa and 32 with EC50 of 5.648, 7.141, and 7.246 (see Table 3), respectively. That finding is 5-HT 4R partial agonist activity. In contrast,− fewer affinity regions indicate that func represented by a large magenta colored area (HA 1 ). groups attached directly to the amide, such as hydroxyl on the piperidine ring or rep 2.3.5. the Gaussian-Field piperidine withBased 3D-QSARwill morpholine Model—Hydrogen reduce biologicalBond Donor Contour activity suchMaps as compounds The contour of hydrogen bond donor (HBD) maps provides and 32 with EC50 of 5.648, 7.141, and 7.246 (see Table 3), respectively.significant information That finding about the functional groups involved in the biological activity in these compounds. Figure 7 resented by a large magenta colored area (HA 1). − shows the favorable (purple) and unfavorable (cyan) HBD regions. The favorable contour (HD+ 1 ) highlights the hydrogen atom of the amide, suggesting that HBDs are favored at that position and explaining why molecules with 1,3,4-oxadiazole have less 5-HT4 R partial agonist activity (compounds 8, 12, 19, 34, 35, 42, 48, 50, 55, and 61; EC50 values between 6.067 to 7.678), due to the change of the hydrogen bond donor capacity of the amide by the hydrogen bond acceptor group such as the oxadiazole ring. Int. J.J. Mol. Int. Mol. Sci. Sci.2021, 22, x3602 2021, 22, FOR PEER REVIEW 1111of of 19 18 Figure 6. Contour maps obtained for the best Gaussian-based 3D-QSAR model hydrogen bond ac- ceptors (HBA), red and magenta regions indicate a favourable and unfavourable hydrogen bond donor interaction, respectively. The active molecule (53) is represented in (A) (sticks representation) and (B) (as draw representation). The less active molecule (5) is represented in (C) (sticks represen- tation) and (D) (as draw representation). 2.3.5. Gaussian-Field Based 3D-QSAR Model—Hydrogen Bond Donor Contour Maps The contour of hydrogen bond donor (HBD) maps provides significant information about the functional groups involved in the biological activity in these compounds. Figure 7 shows the favorable (purple) and unfavorable (cyan) HBD regions. The favorable con- Figure Figure tour (HD 6. 6. Contour Contour maps obtained +1) highlights obtained forthe for the hydrogen thebest bestGaussian-based atom Gaussian-based 3D-QSAR of the amide,3D-QSAR suggesting model model hydrogen thathydrogen HBDs bond bond are ac- accep- favored ceptors tors at that (HBA), (HBA), red red and position and magenta andmagenta regions regions explaining why indicate indicate a favourable a favourable molecules and unfavourable and unfavourablehave with 1,3,4-oxadiazole hydrogen hydrogen bond bond4Rdonor less 5-HT par- donor interaction, respectively. interaction, The molecule active molecule (53) is represented in (A) representation) (sticks representation) tial agonistrespectively. The active activity (compounds 8, 12, 19,(53) 34,is35, represented 42, 48, 50,in55, (A)and (sticks and (B) 61; EC50 values between and (B) (as draw representation). The less active molecule (5) is represented in (C) (sticks represen- (as draw 6.067 to representation). 7.678), due to Thethe less activeofmolecule change (5) is represented the hydrogen bond donorin (C)capacity (sticks representation) of the amideand by tation) and (D) (as draw representation). (D) (as draw representation). the hydrogen bond acceptor group such as the oxadiazole ring. 2.3.5. Gaussian-Field Based 3D-QSAR Model—Hydrogen Bond Donor Contour Maps The contour of hydrogen bond donor (HBD) maps provides significant information about the functional groups involved in the biological activity in these compounds. Figure 7 shows the favorable (purple) and unfavorable (cyan) HBD regions. The favorable con- tour (HD+1) highlights the hydrogen atom of the amide, suggesting that HBDs are favored at that position and explaining why molecules with 1,3,4-oxadiazole have less 5-HT4R par- tial agonist activity (compounds 8, 12, 19, 34, 35, 42, 48, 50, 55, and 61; EC50 values between 6.067 to 7.678), due to the change of the hydrogen bond donor capacity of the amide by the hydrogen bond acceptor group such as the oxadiazole ring. Figure 7. Contour maps obtained for the best Gaussian-based 3D-QSAR model hydrogen bond Figure 7. Contour maps obtained for the best Gaussian-based 3D-QSAR model hydrogen bond do- donors nors (HD), (HD), violet violet regionsindicate regions indicatefavorable favorablehydrogen hydrogenbond bonddonor donor interactions interactions andand cyan cyan regions regions indicate indicate unfavorable unfavorablehydrogen hydrogenbondbonddonor donorinteractions. The interactions. active The molecule active molecule(53)(53) is represented in (A) is represented in (sticks representation) (A) (sticks and (B) representation) and(as(B) draw (as representation). The less draw representation). Theactive lessmolecule (5) is represented active molecule in (5) is repre- sented (C) in (C) (sticks (sticks representation) representation) and (D) (as and (D)representation). draw (as draw representation). 2.4. Design of New Derivatives Based on the results of the 3D-QSAR studies, thirty-nine new compounds (see Table S3, Supplementary Material) were designed and evaluated using molecule 53 as a template. Ten new 5-HT4 R partial agonists with bioactivity greater than that predicted from the tem- plate molecule Figure 7. Contourweremapsselected obtainedusing for thethe GFQSAR best model from Gaussian-based thesemodel 3D-QSAR thirty-nine hydrogencompounds. bond do- The (HD), nors structures violetof the ten regions new designed indicate favorable compounds hydrogen bond anddonor theirinteractions pEC50 values andpredicted by cyan regions the constructed indicate unfavorable FFQSAR and hydrogen GFQSAR bond models are The donor interactions. shown in molecule active Table 4. The design (53) is of these represented in (A) new(sticks representation) derivatives seeks toand (B) (as the enhance drawsteric (S+ 1 ), electrostatic representation). The less (E + ), and active 1 molecule (5) is repre- hydrogen bond sented in (C) acceptor (HA + and (sticks HA+ 2 ) regions representation) andby(D) (as draw representation). modifying fragment 3 of the template molecule (see 3 Figure 8). All the proposed molecules have a predicted activity better than the template Int. J. Mol. Sci. 2021, 22, 3602 12 of 18 Int. Int. J. Int. J.J. Mol. Mol. Sci. Mol. Sci. 2021, 2021, Sci. 22, 2021, 22, x FOR xx FOR 22, PEER PEER REVIEW REVIEW 13 of 19 13 of 19 Int. Int. J. Mol. Int. J.J. Int. Int. Mol. J.J. Mol. Sci. Mol. Sci. Mol. Sci. 2021, Sci. 2021, Sci. 22,22, 2021, 22, 2021, 2021, 22,22, xx xxx FOR FOR FOR FOR FOR FOR PEER PEER PEER REVIEW REVIEW molecule PEER REVIEW PEER PEER (53, REVIEW REVIEW REVIEW 13 13 13 1313 pEC50 = 8.970) according to the GFQSAR model. However, only compound13 of of of of of of 19 19 19 1919 19 var8 showed higher activity than 53 using both models. Table Table 4. The Table 4. Table Table 4. The 4. 4. 4. The TheThe proposed proposed proposed proposed proposed structures structures of new structures of structures structures of of ofof new new new derivates derivates new andand derivates derivates derivates andand their and their their predicted predicted their pEC predicted their predicted 50pEC pEC predicted pECpEC 50 50 values values values 50 valuesusingusing using values using using the thethe the FFQSAR FFQSAR FFQSAR the FFQSAR FFQSAR and andand GFQSAR GFQSAR GFQSAR and and GFQSAR GFQSAR Table Table Table 4. 4. The The The proposed proposed proposed structures structures structures of of newnew new derivates derivates derivates andand and their their their predicted predicted predicted pEC pEC pEC 50 50 50 50 values 50 values values using using using the the the FFQSAR FFQSAR FFQSAR and and and GFQSAR GFQSAR GFQSAR models. models. models. models. models. models. models. models. ID ID IDID ID Structures Structures Structures Structures FFQSAR FFQSAR GFQSAR GFQSAR FFQSAR GFQSAR FFQSAR GFQSAR ID ID IDID ID Structures Structures Structures Structures FFQSAR FFQSAR FFQSAR FFQSAR GFQSAR GFQSAR GFQSAR GFQSAR IDID Structures Structures Structures FFQSAR FFQSAR FFQSAR GFQSAR GFQSAR GFQSAR IDID Structures Structures Structures FFQSAR FFQSAR FFQSAR GFQSAR GFQSAR GFQSAR N N N N N O N N N N O O O O var1 var1 var1 var1 O O O O HNO O 8.283 8.283 8.283 9.221 9.221 9.221 9.221 var6 var6 var6 var6 8.698 8.698 8.698 9.111 9.111 9.111 9.111 var1 var1 var1 var1 HN HN HNO HN HNO HN HN HN O O O O O 8.283 8.283 8.283 8.283 9.221 9.221 9.221 9.221 var6 var6 var6 var6 8.698 8.698 8.698 8.698 9.111 9.111 9.111 9.111 N N N NN N N NNN N NNN N N N N var2 var2 var2 8.209 8.209 8.209 9.375 9.375 9.375 var7 var7 var7 8.690 8.690 8.690 9.547 9.547 9.547 var2 var2 var2 var2 8.209 8.209 8.209 8.209 9.375 9.375 9.375 9.375 var7 var7 var7 var7 8.690 8.690 8.690 8.690 9.547 9.547 9.547 9.547 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 12 of 19 var3 var3 var3 8.285 8.285 8.285 9.432 9.432 9.432 var8var8 var8 9.417 9.417 9.417 9.259 9.259 9.259 var3 var3 var3 8.285 8.285 8.285 9.432 9.432 9.432 var8 var8 var8 9.417 9.417 9.417 9.259 9.259 9.259 2.4. Design of New Derivatives Based on the results of the 3D-QSAR studies, thirty-nine new compounds (see Table S3, Supplementary Material) were designed and evaluated using molecule 53 as a tem- plate. Ten new 5-HT4R partial agonists with bioactivity greater than that predicted from var4 var4 var4var4 the template 8.719 8.719 8.719 8.719 molecule 9.114 9.114 9.114 were selected 9.114 var9 var9 var9 var9using the GFQSAR model 8.818 8.818 8.818 from8.818these thirty-nine 8.818 9.7009.700 9.7009.700 var4 var4 var4 8.719 8.719 8.719 9.114 9.114 9.114 var9var9 var9 8.818 8.818 9.700 9.700 9.700 compounds. The structures of the ten new designed compounds and their pEC50 values predicted by the constructed FFQSAR and GFQSAR models are shown in Table 4. The design of these new derivatives seeks to enhance the steric (S+1), electrostatic (E+1), and hydrogen bond acceptor (HA+3 and HA+2) regions by modifying fragment 3 of the tem- var5 var5 var5var5 8.775 8.775 8.775 8.775 (see Figure 9.513 9.513 9.513 9.513 var10var10 var10 var10 8.641 8.6418.641 8.641 9.8499.849 9.8499.849 var5 var5 var5 var5 var5 8.775 plate molecule 8.775 9.513 9.513 8). Allvar10 var10 the proposed molecules have a predicted 8.641 8.641 activity 9.849 9.849 better than the template molecule (53, pEC50 = 8.970) according to the GFQSAR model. However, only compound var8 showed higher activity than 53 using both models. 2.5. 2.5. ADMET ADMET 2.5. ADMET Predictions Predictions Predictions 2.5. 2.5. 2.5. ADMET ADMET ADMET Predictions Predictions Predictions MostMost Most drugdrug drug candidates candidates candidates fail fail toto fail to make to make make it through it through through it through through clinical clinical clinical trials trials trialsinininthe in the thedrugdrug drug discovery discovery discovery pro- pro- pro- MostMost Most drug drug drug candidates candidates candidates failfail fail to to makemake make it it it through clinical clinical clinical trials trials trials in in the the thedrugdrug drug discovery discovery discovery pro- pro- pro- cess cess cess because because because of of of their their theirpoor poor poor pharmacokinetics. pharmacokinetics. pharmacokinetics. To To To assess assess assess whether whether whether all all newly newly all newly designed designed designed com- com- com- cess cessbecause because ofof their their poorpoor pharmacokinetics. pharmacokinetics. ToTo assess assess whether whether allallnewly newly designed designed com-com- pounds pounds pounds could could could become become become potential potential potential drugs, drugs, drugs, we we we perform perform perform ADMET ADMET ADMET predictions. predictions. predictions. pounds pounds pounds could could could become become become potential potential potential drugs, drugs, drugs, we we we perform perform perform ADMET ADMET ADMET predictions. predictions. predictions. ADMET ADMET ADMET properties properties properties areare are shown shown shown in in Table in Table Table 5, 5, and 5, and and drugdrug drug similarity similarity similarity predictions predictions predictions are are are shown shownshown ADMET ADMET ADMET properties properties properties are are are shown shown shown ininin Table Table Table 5, 5, 5, and and and drugdrug drug similarity similarity similarity predictions predictions predictions are are are shown shown shown in in in Table Table Table 6. 6. The 6. The The intestine intestine intestine is is the is the themainmain main sitesite site of of absorption absorption of absorption of ofan of an an orally orally orally administered administered administered drug.drug. drug. AA AA in inin Table Table Table 6. 6. 6.The The The intestine intestine intestine is is is the the the main main main sitesite site ofofof absorption absorption absorption ofofof anan an orally orally orally administered administered administered drug.drug. drug. A A molecule molecule molecule withwith with an an an absorbance absorbance absorbance of ofmore of more more thanthan than 30%30% 30% is is considered considered is considered considered well wellwell absorbed. absorbed. absorbed. As As Asshownshown shown in in in molecule molecule molecule with with with ananan absorbance absorbance absorbance ofof of more more more than than than 30% 30% 30% is is is considered considered well well well absorbed. absorbed. absorbed. AsAsAsshownshown shown ininin Table Table Table 5, 5, 5, the 5, the the intestinal intestinal intestinal absorbance absorbance absorbance ofofof ten of ten ten molecules molecules molecules is is is between between is between 92.7% 92.7%92.7% andand and 96.3%, 96.3%, 96.3%, which which which Table Table 5, the the intestinal intestinal absorbance absorbance of ten ten molecules molecules is between between 92.7% 92.7% andand 96.3%, 96.3%, which which reveals reveals a very good absorption in the human gut. Significantly, the intestinal absorbance reveals reveals reveals reveals aa very very aaa very very very good good good good good absorption absorption absorption absorption absorption inin in the in in the the the the human humanhuman human human gut. gut.gut. gut. gut. Significantly, Significantly, Significantly, Significantly, Significantly, the the the the theintestinal intestinal intestinal intestinal intestinal absorbance absorbance absorbance absorbance absorbance of of of compound compound compound var6var6 var6 is is is higher higherhigher thanthan than 96%.96%. 96%. A A A volume volume volume of of distribution distribution of distribution (VDss) (VDss) (VDss) greater greater greater than than than 0.45 0.45 0.45 of ofof compound compound compound var6 var6 var6 is is is higherhigher higher thanthan than 96%.96%. 96%. AA A volume volume volume ofofof distribution distribution distribution (VDss) (VDss) (VDss) greater greater greater than than than 0.45 0.45 0.45 is is is considered considered considered high. high. high. High High High VDss VDssVDss indicates indicates indicates thatthat that more more more drugs drugs drugs are are are distributed distributed distributed in in tissues tissues in tissues than thanthan in in in is considered is is considered considered high. high. high. HighHigh High VDss VDss VDss indicates indicates indicates thatthat that more more more drugs drugs drugs are are are distributed distributed distributed ininin tissues tissues tissues than than than ininin plasma. plasma. plasma. TheThe The VDssVDss VDss of of the of the the ten ten tennew new new compounds compounds compounds werewere were higher higher higher thanthan than 0.45. 0.45. 0.45. plasma. plasma. plasma. The The The VDssVDss VDss ofofof the the the ten ten tennew new new compounds compounds compounds were were were higher higher higher than than than 0.45. 0.45. 0.45. Metabolism Metabolism Metabolism plays plays plays anan an an essential essential essential rolerole role ininin converting converting in converting pharmacological pharmacological pharmacological compounds. compounds. compounds. Cyto- Cyto- Cyto- Metabolism Metabolism plays plays an essential essential rolerole in converting converting pharmacological pharmacological compounds. compounds. Cyto- Cyto- chromes chromes chromes CYP2D6 CYP2D6 CYP2D6 and and and CYP3A4 CYP3A4CYP3A4 are are arethe the thetwo two two main main main P450P450 P450 isoforms isoforms isoforms responsible responsible responsible forfordrug for drug drug me-me- me- chromes chromes chromes CYP2D6 CYP2D6 CYP2D6 and and and CYP3A4CYP3A4 CYP3A4 are are arethe the the two two two main main main P450 P450 P450 isoforms isoforms isoforms responsible responsible responsible for forfor drug drug drug me- me- me- tabolism. tabolism. tabolism. As As As indicated indicated indicated in in in Table TableTable 5, 5, all 5, all designed designed all designed small small small molecules molecules molecules werewere were neither neither neither substrate substrate substrate nor nor nor tabolism. tabolism. tabolism. Figure Figure AsAs 8. Template 8. TemplateAs indicated indicated indicated molecule molecule in in in(53) Table Table Table (53) used 5, 5, used 5, all all allfor for designed thedesigned designed thederivation small small small derivation molecules ofmolecules molecules of new werewere were molecules/compounds new neither neither neither molecules/compounds substrate substrate substrate with en- with nor nor nor inhibitor inhibitor inhibitor of of of CYP2D6, CYP2D6, CYP2D6, a aa feature feature feature thatthat that may may may be be bean an an advantage advantage advantage since sincesince compound compound compound 53 53is 53 is inhibitory inhibitory is inhibitory inhibitor hanced bioactivity. inhibitor inhibitor enhanced ofof of CYP2D6, CYP2D6, CYP2D6, bioactivity. aa feature a featurefeature that that that may may may bebe be ananan advantage advantage advantage since since since compound compound compound 535353 is inhibitory is is inhibitory inhibitory to toto to CYP2D6. CYP2D6. CYP2D6. Furthermore, Furthermore, Furthermore, the the the compounds compounds compounds were werewere substrates substrates substrates of ofof of CYP3A4 CYP3A4 CYP3A4 except except except for forforfor com- com- com- to to CYP2D6. CYP2D6. CYP2D6. Furthermore, Furthermore, Furthermore, the the the compounds compounds compounds were were were substrates substrates substrates of of CYP3A4 CYP3A4 CYP3A4 except except except for forcom-com- com- pounds pounds pounds In general, In general,var8, var8, var8, var9, the var9, var9, andand and presence presence var10, var10, var10, of ofthe the indicating indicating indicating thatthat that pyrazolo[1,5-a]pyridine pyrazolo[1,5-a]pyridine CYP3A4 CYP3A4 CYP3A4 cancan can ring metabolize metabolize ringmetabolize provides provides compounds compounds compounds moremore active active an-var1- var1- var1- pounds pounds var8, var8, var9,var9, andand var10, var10, indicating indicating that thatCYP3A4 CYP3A4 cancan metabolize metabolize compounds compounds var1- var1- var7. var7. var7. alogues. analogues. Compounds Compounds Compounds This Thisaromatic aromatic var2, var2, var2, var3, var3, var3, heterocyclic heterocyclic var4, var4,var4, var5, var5, var5, system system var6,var6, var6, provides providesvar7, var7, var7, andand and var8 lipophilic lipophilic var8 var8 areare are and not not not CYP3A4 CYP3A4 CYP3A4 electronic electronic inhibitors, inhibitors, inhibitors, features. features. var7. var7. var7. Compounds Compounds Compounds var2, var2, var2, var3, var3, var3, var4,var4, var4, var5,var5, var5, var6,var6, var6, var7, var7, var7, and and and var8 var8 var8 are are are not not not CYP3A4 CYP3A4 CYP3A4 inhibitors, inhibitors, inhibitors, implying implying implying Furthermore, Furthermore, thatthat that the the theythey they amide amide will willwill not attached attached not not affect affect affect totothe thenormal normal normal drug heterocyclic heterocyclic drug drug metabolism. metabolism. metabolism. ringatatposition ring position77has hashigher higherbiological biological implying implying implying that that that theythey they willwill will not not not affect affect affect normal normal normal drug drug drug metabolism. metabolism. metabolism. activity. activity. DrugDrug Drug Onthe On the clearance clearance clearance otherhand, other hand,related related related to allto all to the the bioavailability bioavailability to bioavailability analogueshave analogues have is is is essential essential is essential aaessential fragment fragment ininin in determining determining determining 33whose whoseextension extension dosing dosing dosing isis rates rates rates four four to toto to DrugDrug Drug clearance clearance clearance related related related to to bioavailability bioavailability bioavailability is is essential essential in in determining determining determining dosing dosing dosing rates rates rates to to carbonachieve achieve achieve carbon atoms atoms steady-state steady-state steady-state from fromthe concentrations concentrations concentrations amide’s theconcentrations amide’s nitrogen nitrogen in in inthe to the the tothe the body. body. body. FromFrom From heterocycle’s heterocycle’s the the the predicted predicted predicted nitrogen nitrogen except except total total total clearance, clearance, clearance, compound compound allall all achieve achieve achieve steady-state steady-state steady-state concentrations concentrations ininin the the body. body. body. From From From the the the predicted predicted predicted total total total clearance, clearance, clearance, all allall var8,compounds compounds compounds which has cancan can a be be bebe be secondaryexcreted excreted excreted without without without aliphatic problems problems problems amine. This at at the atextension the the renal renal renal level. level. level. oflevel. fragment 3 provides good compounds compounds can can excreted excreted without without problems problems at the at the renal renal level. biological Furthermore, Furthermore, Furthermore, activity, where drug drugdrug var1,toxicity toxicity toxicity var2, is is isis var3,another another another var6, important important important var7, andindex index index index var9 have for forfor for andrug drug drug screening. screening. screening. aliphatic ringDrugs Drugs Drugs Drugs that Furthermore, Furthermore, Furthermore, drug drug drug toxicity toxicity toxicity is is another another another important important important index index for fordrugdrug drug screening. screening. screening. Drugs Drugs should should should favors bebe beas thebesteric as as non-toxic non-toxic non-toxic region to (green to to human human human color,health health health health see as Figure as as possible possible possible 3) of the or or or have havehave 3D-QSAR a a wide a wide wide therapeutic therapeutic models.therapeutic margin. margin. Thesemargin. margin. rings should should should be be asasas non-toxic non-toxic non-toxic tototo human human human health health asas as possible possible possible or oror have have have aa wide a wide wide therapeutic therapeutic therapeutic margin. margin. have at least one functional group with hydrogen bond acceptor properties favoring the Int. J. Mol. Sci. 2021, 22, 3602 13 of 18 var8, which has a secondary aliphatic amine. This extension of fragment 3 provides good biological activity, where var1, var2, var3, var6, var7, and var9 have an aliphatic ring that favors the steric region (green color, see Figure 3) of the 3D-QSAR models. These rings have at least one functional group with hydrogen bond acceptor properties favoring the HA+ 3 regions’ interaction (see Figure 6). In contrast, var4, var5, var8, and var10 compounds have an aliphatic chain with an oxygenated functional group interacting in the HA+ 2 hydrogen bond acceptor region. 2.5. ADMET Predictions Most drug candidates fail to make it through clinical trials in the drug discovery process because of their poor pharmacokinetics. To assess whether all newly designed compounds could become potential drugs, we perform ADMET predictions. ADMET properties are shown in Table 5, and drug similarity predictions are shown in Table 6. The intestine is the main site of absorption of an orally administered drug. A molecule with an absorbance of more than 30% is considered well absorbed. As shown in Table 5, the intestinal absorbance of ten molecules is between 92.7% and 96.3%, which reveals a very good absorption in the human gut. Significantly, the intestinal absorbance of compound var6 is higher than 96%. A volume of distribution (VDss) greater than 0.45 is considered high. High VDss indicates that more drugs are distributed in tissues than in plasma. The VDss of the ten new compounds were higher than 0.45. %clearpage Table 5. Absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of new designed molecules. Absorption Distribution Metabolism Excretion Toxicity Substrate Inhibitor Skin ID IA 1 VDss 2 2D6 3A4 1A2 2C19 2C9 2D6 3A4 TC 3 AMES Hepatotoxicity Sensitization 53 92.738 1.473 Yes Yes No No No Yes Yes 0.825 No Yes No var1 95.344 1.158 No Yes No No No No Yes 0.626 No Yes No var2 95.054 1.152 No Yes No No No No No 0.776 No Yes No var3 94.483 1.084 No Yes No No No No No 0.789 No Yes No var4 95.045 0.941 No Yes No No No No No 0.837 No Yes No var5 95.572 0.826 No Yes No No No No No 1.258 No Yes No var6 96.306 0.84 No Yes No No No No No 0.936 No Yes No var7 95.844 0.952 No Yes No No No No No 1.06 No Yes No var8 94.815 0.584 No No No No No No No 1.171 No Yes No var9 92.905 0.961 No No No No No No Yes 0.719 No Yes No var10 95.300 0.671 No No No No No No Yes 0.881 No Yes No 1 IA is intestinal absorption, values expressed in % absorption. 2 VDss is volume of distribution, values expressed in log L kg−1 . 3 TC is total clearance, values expressed in log mL min−1 kg−1 . Table 6. Drug likeness of novel designed molecules based on Lipinski, Ghose, Veber, and Egan rules, and their synthetic accessibility. ID Lipinski Ghose Veber Egan Synthetic Accessibility 53 Yes Yes Yes Yes 3.26 var1 Yes No Yes Yes 4.77 var2 Yes Yes Yes Yes 4.55 var3 Yes Yes Yes Yes 4.4 var4 Yes Yes Yes Yes 3.52 var5 Yes Yes No Yes 3.68 var6 Yes Yes Yes Yes 4.07 var7 Yes Yes Yes Yes 3.97 var8 Yes Yes Yes Yes 2.71 var9 Yes No Yes Yes 4.89 var10 Yes Yes Yes Yes 3.48 Int. J. Mol. Sci. 2021, 22, 3602 14 of 18 Metabolism plays an essential role in converting pharmacological compounds. Cy- tochromes CYP2D6 and CYP3A4 are the two main P450 isoforms responsible for drug metabolism. As indicated in Table 5, all designed small molecules were neither substrate nor inhibitor of CYP2D6, a feature that may be an advantage since compound 53 is in- hibitory to CYP2D6. Furthermore, the compounds were substrates of CYP3A4 except for compounds var8, var9, and var10, indicating that CYP3A4 can metabolize compounds var1- var7. Compounds var2, var3, var4, var5, var6, var7, and var8 are not CYP3A4 inhibitors, implying that they will not affect normal drug metabolism. Drug clearance related to bioavailability is essential in determining dosing rates to achieve steady-state concentrations in the body. From the predicted total clearance, all compounds can be excreted without problems at the renal level. Furthermore, drug toxicity is another important index for drug screening. Drugs should be as non-toxic to human health as possible or have a wide therapeutic margin. All new derivatives are non-toxic to AMES (estimation of the mutagenic potential of chemical compounds) and do not cause skin sensitization. However, all new compounds have potential hepatotoxicity, which could possibly alter normal liver function. To further understand this unfavorable side effect, the synthesis of the proposed compounds must be tested in a living organism. However, such studies go beyond the current stage of this study. Finally, the online tool SwissADME (http://www.swissadme.ch/, accessed on 19 February 2021), which provides access to several different rule-based filters, was used to predict drug similarity. As shown in Table 6, all new compounds meet the Lipinski and Egan drug similarity rules; only compounds var1 and var9 fail the Ghose filter, and compound var5 fails the Veber filter. According to Lipinski and Egan’s drug similarity rules, the results of multiple evaluations indicate that these computationally designed compounds can be converted into oral drugs. The synthetic accessibility values of all designed molecules are approximately 4, meaning that they are synthesizable compounds (synthetic accessibility ranges from 1–10). 3. Materials and Methods 3.1. Dataset Collection A total of 62 partial agonists of the 5-HT4 receptor (Figure 1), which showed promis- sory potency, were collected from the literature [18,22,23]. All the compounds with pEC50 values ranging from 5.64 to 10.0 were used in this study. The geometry for all these molecules was converted into a 3D structure using OCHEM. The 3D structure of the molecules was processed with OMEGA [37] module using the following parameters: (i) AM1_BCC Force field, (ii) FixpKa from the QUAPAC package for all possible ionization states at a given biological pH, (iii) one low energy conformation per ligand. Force- and Gaussian-field 3D-QSAR calculations were performed for all the molecules. All the training and test set molecules with experimental and predicted EC50 values were listed in Table S1 of the Supplementary Material. 3.2. Alignment The alignment of molecules is the most crucial input for the generation of 3D-QSAR models. The compound with the highest activity (53) was used as the template molecule. A shape-based alignment was used for all conformers of each ligand. These alignments were carried out with ROCS suite [38]. Finally, each ligand’s best conformer was filtered considering electrostatic field compound 53, as is shown in Figure S1 of the Supplementary Material. 3.3. Field-Based QSAR Model The 3D-QSAR analysis using field-based methods was performed with the QSAR tool of the Schrodinger Suite. The 3D-QSAR method builds the model by relating the known activities and molecular elements of the training set using the OPLS_2005 force Int. J. Mol. Sci. 2021, 22, 3602 15 of 18 field. The steric and electrostatic field around the ligand on a 3D grid was calculated using the field-based 3D-QSAR. The force-field-based QSAR model is an alignment-dependent method in which the interaction energy terms of the molecular field are correlated with biological activities using multivariate statistical analysis. In the 3D-QSAR model based on Gaussian force fields, interaction energy calculations were performed using steric, electrostatic, hydrogen bond donor (HBD), and hydrogen bond acceptor (HBA) potential fields using Gaussian equations for the field calculations. The fields were calculated on an orthohedral grid enclosing the training set molecules, with a spacing of 1 Å and extending 3 Å beyond the boundaries of this set. The threshold for van der Waals and electrostatic interactions was set at 30 kcal/mol, eliminating points closer than 2 Å from any of the atoms in the training set. During the PLS procedure, all variables (grid points) with a standard deviation less than 0.05 were removed. The lattice and probe step sizes were adjusted automatically. The partial least squares (PLS) analysis is applied to construct the best model through the linear correlation of FFQSAR and GFQSAR concerning pEC50 [18,22,23]. The maximum number of PLS factors was set to 5. A cross-validation analysis was performed using the leave-one-out method. The external predictive ability of each model constructed was assessed by calculating the predictive correlation coefficient (R2 test ). In addition, the models were also subjected to external validation criteria according to the test proposed by Golbraikh and Tropsha (see Supplementary Material in Section S2.4) [39,40]. All these calculations were carried out with the DTC Lab software tools (https://dtclab.webs.com/software-tools, accessed on 25 March 2021). 3.4. Prediction ADMET Properties Drug candidates need to have good ADMET (absorption, distribution, metabolism, excretion, and toxicity) and drug-likeness profiles to initially estimate pharmacokinetic and drug-likeness parameters in the drug discovery process [41]. In this work, new candidates with ADMET properties include human intestinal absorption, steady-state volume of distribution (VDss), hepatic metabolism, total clearance, AMES toxicity, hepatotoxicity, and skin sensitization properties. ADMET can be predicted using pkCSM [29]. The prediction of drug similarity of new molecules is estimated using parameters based on Lipinski, Ghose, Veber, and Egan rules and their synthetic accessibility by ap- plying the SwissADME web tool [31] (http://www.swissadme.ch, accessed on 25 March 2021). The SwissADME synthetic accessibility score is mainly based on the assumption of the molecular fragments in the “actually” obtainable molecules, which correlates with the ease of synthesis. The score is normalized to range from 1 (very easy) to 10 (very difficult to synthesise). 4. Conclusions The structures included in this study have a reasonable structure–activity relationship and good correlation. The force and Gaussian-field models were generated and showed good R2 and Q2 LOO values for the models. The field-based model has R2 = 0.821 and Q2 = 0.804 based on the steric and electrostatic fields. The Gaussian model has R2 = 0.898 and Q2 = 0.886 based on the four field intensities of steric, electrostatic, hydrogen-bond acceptor (HBA), and hydrogen-bond donor (HBD). The analysis of both 3D-QSAR models indicates that the largest contributions are provided by steric and hydrogen bond acceptors properties (0.420 and 0.304, respectively). The models developed herein can be further applied to design new compounds with potent 5-HT4 receptor partial agonists. Finally, we found three factors that could effectively enhance the activity of 5-HT4 R partial agonists: (1) The four-carbon atom distance between the amide nitrogen and the aliphatic amine corresponding to fragment 3. (2) Structural variability in fragment 3 considering aliphatic rings that provide a favourable hydrophobic source for activity Int. J. Mol. Sci. 2021, 22, 3602 16 of 18 (3) The hydrogen bond acceptor groups in fragment 3 can enhance the activity of com- pounds. Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW The 17 of 19 structural elements related to the biological activity of these compounds studied are shown in Figure 9. Figure9.9.Main Figure Mainstructure-activity structure-activityrelationships relationshipsderived derivedfrom fromthis thisstudy. study. Supplementary Finally, theMaterials: The following combination are available of the three online at better factors showed www.mdpi.com/xxx/s1. predicted biological ac- tivity than the single or the two factors (var 8). According to Author Contributions: Conceptualization, methodology, and investigation,these rules, thirty-nine A.C.-A. new and R.N.; su- molecules were designed, and the constructed 3D-QSAR models were used to predict pervision, E.C.-Á. and R.N.; writing—original draft, A.C.-A.; writing—review and editing, A.C.-A., the pEC R.N., and value of 50 E.C.-Á. Allthe newly authors designed have read andmolecules. Tenpublished agreed to the new 5-HT 4 R partial version of theagonists were manuscript. selected as having promised biological activity compared to the studied compounds. Fur- Funding: This research was funded by FONDECYT, grant number 11200620 and 3190118. thermore, the results of in silico studies suggested that these new 5-HT4 R partial agonists have reasonable Institutional ADMET Review Boardproperties Statement:and Notdrug-likeness. applicable. These results establish a theoretical basis for further study of these compounds. Informed Consent Statement: Not applicable. A deeper study focused on synthesizing of these compounds and the experimental study of their biological activity will pursue in Data Availability future research. Statement: The data presented in this study are available in Supplementary Ma- terial or on request from the corresponding author. Supplementary Acknowledgments: Materials: A.C.-A.The following and are available R.N. would online at https://www.mdpi.com/article/10 like to acknowledge the FONDECYT program. Ronald .3390/ijms22073602/s1. Nelson would like to acknowledge the Postdoctoral FONDECYT Grant. ICN2 is supported by the Severo Ochoa program, the Spanish Research Agency (AEI, grant no. SEV-2017-0706), and the Author Contributions: Conceptualization, methodology, and investigation, A.C.-A. and R.N.; super- CERCA Programme/Generalitat de Catalunya. E.C.-Á. acknowledges support from Spanish vision, E.C.-Á. and R.N.; writing—original draft, A.C.-A.; writing—review and editing, A.C.-A., R.N. MICINN project SIP (PGC2018-101743-B-I00). and E.C.-Á. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest: The authors declare no conflict of interest. Funding: This research was funded by FONDECYT, grant number 11200620 and 3190118. Institutional AbbreviationsReview Board Statement: Not applicable. 5-HT4RConsent Informed Receptor 5-HT Statement: Not 4 applicable. AD Alzheimer’s disease Data Availability Statement: The data presented in this study are available in Supplementary S Material Stericfrom the corresponding author. or on request E Electrostatic Acknowledgments: H A.C.-A. and R.N. would like to acknowledge the FONDECYT program. Ronald Hydrophobic Nelson would like to acknowledge the Postdoctoral FONDECYT Grant. ICN2 is supported by the HBA Hydrogen bond acceptor Severo Ochoa program, the Spanish Research Agency (AEI, grant no. SEV-2017-0706), and the CERCA HBD Hydrogen bond donor Programme/Generalitat de Catalunya. E.C.-Á. acknowledges support from Spanish MICINN project SIPADMET Absorption, distribution, metabolism, excretion, and toxicity (PGC2018-101743-B-I00). FFQSAR Force-field based QSAR Conflicts GFQSAR of Interest: The authors declare Gaussian-field no conflict of interest. based QSAR CCC Correlation coefficient of concordance 3D-QSAR Three-dimensional quantitative structure–activity relationship References 1. McDowell, S.E. Adverse reactions to drugs used in the treatment of Alzheimer’s disease. Adverse Drug React. Bull. 2011, 1031– 1034, doi:10.1097/FAD.0b013e328348c804. 2. Gerald, C.; Adham, N.; Kao, H.T.; Olsen, M.A.; Laz, T.M.; Schechter, L.E.; Bard, J.A.; Vaysse, P.J.; Hartig, P.R.; Branchek, T.A.; et al. The 5-HT4 receptor: Molecular cloning and pharmacological characterization of two splice variants. EMBO J. 1995, 14, 2806–2815. Int. J. Mol. Sci. 2021, 22, 3602 17 of 18 Abbreviations 5-HT4 R Receptor 5-HT4 AD Alzheimer’s disease S Steric E Electrostatic H Hydrophobic HBA Hydrogen bond acceptor HBD Hydrogen bond donor ADMET Absorption, distribution, metabolism, excretion, and toxicity FFQSAR Force-field based QSAR GFQSAR Gaussian-field based QSAR CCC Correlation coefficient of concordance 3D-QSAR Three-dimensional quantitative structure–activity relationship References 1. McDowell, S.E. Adverse reactions to drugs used in the treatment of Alzheimer’s disease. Adverse Drug React. Bull. 2011, 1031–1034. [CrossRef] 2. Gerald, C.; Adham, N.; Kao, H.T.; Olsen, M.A.; Laz, T.M.; Schechter, L.E.; Bard, J.A.; Vaysse, P.J.; Hartig, P.R.; Branchek, T.A.; et al. The 5-HT4 receptor: Molecular cloning and pharmacological characterization of two splice variants. EMBO J. 1995, 14, 2806–2815. [CrossRef] [PubMed] 3. Eglen, R.M.; Wong, E.H.F.; Dumuis, A.; Bockaert, J. Central 5-HT4 receptors. Trends Pharmacol. Sci. 1995, 16, 391–398. [CrossRef] 4. Bockaert, J.; Claeysen, S.; Compan, V.; Dumuis, A. 5-HT4 receptors. Curr. Drug Targets CNS Neurol. Disord. 2004, 3, 39–51. [CrossRef] 5. Bockaert, J.; Claeysen, S.; Compan, V.; Dumuis, A. 5-HT4 receptors, a place in the sun: Act two. Curr. Opin. Pharmacol. 2011, 11, 87–93. [CrossRef] [PubMed] 6. Spencer, J.P.; Brown, J.T.; Richardson, J.C.; Medhurst, A.D.; Sehmi, S.S.; Calver, A.R.; Randall, A.D. Modulation of hippocampal excitability by 5-HT 4 receptor agonists persists in a transgenic model of Alzheimer’s disease. Neuroscience 2004, 129, 49–54. [CrossRef] 7. Maillet, M.; Robert, S.; Lezoualc’h, F. New Insights into Serotonin 5-HT4 Receptors: A Novel Therapeutic Target for Alzheimers Disease? Curr. Alzheimer Res. 2004, 1, 79–85. [CrossRef] 8. Rebholz, H.; Friedman, E.; Castello, J. Alterations of Expression of the Serotonin 5-HT4 Receptor in Brain Disorders. Int. J. Mol. Sci. 2018, 19, 3581. [CrossRef] [PubMed] 9. Liu, Q.Q.; Yao, X.X.; Gao, S.H.; Li, R.; Li, B.J.; Yang, W.; Cui, R.J. Role of 5-HT receptors in neuropathic pain: Potential therapeutic implications. Pharmacol. Res. 2020, 159, 104949. [CrossRef] [PubMed] 10. Tonini, M.; Pace, F. Drugs Acting on Serotonin Receptors for the Treatment of Functional GI Disorders. Dig. Dis. 2006, 24, 59–69. [CrossRef] 11. Gershon, M.D. Review article: Serotonin receptors and transporters—Roles in normal and abnormal gastrointestinal motility. In Alimentary Pharmacology and Therapeutics; Supplement; Blackwell Publishing Ltd.: Hoboken, NJ, USA, 2004; Volume 20, pp. 3–14. 12. Nam, Y.; Min, Y.S.; Sohn, U.D. Recent advances in pharmacological research on the management of irritable bowel syndrome. Arch. Pharm. Res. 2018, 41, 955–966. [CrossRef] 13. Bouras, E.P.; Camilleri, M.; Burton, D.D.; McKinzie, S. Selective stimulation of colonic transit by the benzofuran 5HT4 agonist, prucalopride, in healthy humans. Gut 1999, 44, 682–686. [CrossRef] 14. Konen, J.R.; Haag, M.M.; Guseva, D.; Hurd, M.; Linton, A.A.; Lavoie, B.; Kerrigan, C.B.; Joyce, E.; Bischoff, S.C.; Swann, S.; et al. Prokinetic actions of luminally acting 5-HT4 receptor agonists. Neurogastroenterol. Motil. 2020, 33, e14026. [CrossRef] 15. Gwynne, R.M.; Bornstein, J.C. Luminal 5-HT4 receptors—A successful target for prokinetic actions. Neurogastroenterol. Motil. 2019, 31, e13708. [CrossRef] [PubMed] 16. Lezoualc’h, F. The serotonin 5-HT4 receptor and the amyloid precursor protein processing. Exp. Gerontol. 2003, 38, 159–166. [CrossRef] 17. Lanthier, C.; Dallemagne, P.; Lecoutey, C.; Claeysen, S.; Rochais, C. Therapeutic modulators of the serotonin 5-HT4 receptor: A patent review (2014-present). Expert Opin. Ther. Pat. 2020, 30, 495–508. [CrossRef] 18. Brodney, M.A.; Johnson, D.E.; Sawant-Basak, A.; Coffman, K.J.; Drummond, E.M.; Hudson, E.L.; Fisher, K.E.; Noguchi, H.; Waizumi, N.; McDowell, L.L.; et al. Identification of Multiple 5-HT 4 Partial Agonist Clinical Candidates for the Treatment of Alzheimer’s Disease. J. Med. Chem. 2012, 55, 9240–9254. [CrossRef] 19. Ahmad, I.; Nirogi, R. 5-HT4 Receptor Agonists for the Treatment of Alzheimer’s Dsease. Neurosci. Med. 2011, 02, 87–92. [CrossRef] 20. Modica, M.; Pittala, V.; Romeo, G.; Salerno, L.; Siracusa, M. Serotonin 5-HT3 and 5-HT4 Ligands: An Update of Medicinal Chemistry Research in the Last Few Years. Curr. Med. Chem. 2010, 17, 334–362. [CrossRef] Int. J. Mol. Sci. 2021, 22, 3602 18 of 18 21. Castriconi, F.; Paolino, M.; Giuliani, G.; Anzini, M.; Campiani, G.; Mennuni, L.; Sabatini, C.; Lanza, M.; Caselli, G.; De Rienzo, F.; et al. Synthesis and structure–activity relationship studies in serotonin 5-HT4 receptor ligands based on a benzo[de][2,6]naphthridine scaffold. Eur. J. Med. Chem. 2014, 82, 36–46. [CrossRef] 22. Nirogi, R.; Mohammed, A.R.; Shinde, A.K.; Bogaraju, N.; Gagginapalli, S.R.; Ravella, S.R.; Kota, L.; Bhyrapuneni, G.; Muddana, N.R.; Benade, V.; et al. Synthesis and SAR of Imidazo[1,5-a]pyridine derivatives as 5-HT4 receptor partial agonists for the treatment of cognitive disorders associated with Alzheimer’s disease. Eur. J. Med. Chem. 2015, 103, 289–301. [CrossRef] 23. Nirogi, R.; Mohammed, A.R.; Shinde, A.K.; Gagginapally, S.R.; Kancharla, D.M.; Middekadi, V.R.; Bogaraju, N.; Ravella, S.R.; Singh, P.; Birangal, S.R.; et al. Synthesis, Structure–Activity Relationships, and Preclinical Evaluation of Heteroaromatic Amides and 1,3,4-Oxadiazole Derivatives as 5-HT 4 Receptor Partial Agonists. J. Med. Chem. 2018, 61, 4993–5008. [CrossRef] [PubMed] 24. Abbasi, M.; Sadeghi-Aliabadi, H.; Hassanzadeh, F.; Amanlou, M. Prediction of dual agents as an activator of mutant p53 and inhibitor of Hsp90 by docking, molecular dynamic simulation and virtual screening. J. Mol. Graph. Model. 2015, 61, 186–195. [CrossRef] 25. Athar, M.; Lone, M.Y.; Khedkar, V.M.; Jha, P.C. Pharmacophore model prediction, 3D-QSAR and molecular docking studies on vinyl sulfones targeting Nrf2-mediated gene transcription intended for anti-Parkinson drug design. J. Biomol. Struct. Dyn. 2016, 34, 1282–1297. [CrossRef] [PubMed] 26. Roy, K.; Kar, S.; Ambure, P. On a simple approach for determining applicability domain of QSAR models. Chemom. Intell. Lab. Syst. 2015, 145, 22–29. [CrossRef] 27. Tripuraneni, N.S.; Azam, M.A. A combination of pharmacophore modeling, atom-based 3D-QSAR, molecular docking and molecular dynamics simulation studies on PDE4 enzyme inhibitors. J. Biomol. Struct. Dyn. 2016, 34, 2481–2492. [CrossRef] [PubMed] 28. Zhu, X.; Zhou, L.; Zhong, L.; Dai, D.; Hong, M.; You, R.; Wang, T. Exploration of potential RSK2 inhibitors by pharmacophore modelling, structure-based 3D-QSAR, molecular docking study and molecular dynamics simulation. Mol. Simul. 2017, 43, 534–547. [CrossRef] 29. Pires, D.E.V.; Blundell, T.L.; Ascher, D.B. pkCSM: Predicting small-molecule pharmacokinetic and toxicity properties using graph-based signatures. J. Med. Chem. 2015, 58, 4066–4072. [CrossRef] [PubMed] 30. pkCSM. Available online: http://biosig.unimelb.edu.au/pkcsm/ (accessed on 19 February 2021). 31. Daina, A.; Michielin, O.; Zoete, V. SwissADME: A free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci. Rep. 2017, 7, 42717. [CrossRef] 32. Molecular Modeling Group. Swiss Institute of Bioinformatics SwissADME. Available online: http://www.swissadme.ch/ (accessed on 19 February 2021). 33. Lipinski, C.A.; Lombardo, F.; Dominy, B.W.; Feeney, P.J. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 1997, 23, 3–25. [CrossRef] 34. Ghose, A.K.; Viswanadhan, V.N.; Wendoloski, J.J. A knowledge-based approach in designing combinatorial or medicinal chemistry libraries for drug discovery. 1. A qualitative and quantitative characterization of known drug databases. J. Comb. Chem. 1999, 1, 55–68. [CrossRef] 35. Veber, D.F.; Johnson, S.R.; Cheng, H.-Y.; Smith, B.R.; Ward, K.W.; Kopple, K.D. Molecular Properties That Influence the Oral Bioavailability of Drug Candidates. J. Med. Chem. 2002, 45, 2615–2623. [CrossRef] 36. Egan, W.J.; Merz, K.M.; Baldwin, J.J. Prediction of drug absorption using multivariate statistics. J. Med. Chem. 2000, 43, 3867–3877. [CrossRef] [PubMed] 37. Hawkins, P.C.D.; Skillman, A.G.; Warren, G.L.; Ellingson, B.A.; Stahl, M.T. Conformer Generation with OMEGA: Algorithm and Validation Using High Quality Structures from the Protein Databank and Cambridge Structural Database. J. Chem. Inf. Model. 2010, 50, 572–584. [CrossRef] 38. Paul, C.D.; Hawkins, A.; Geoffrey, S.; Nicholls, A. Comparison of Shape-Matching and Docking as Virtual Screening Tools. J. Med. Chem. 2006, 50, 74–82. [CrossRef] 39. Golbraikh, A.; Tropsha, A. Beware of q2! J. Mol. Graph. Model. 2002, 20, 269–276. [CrossRef] 40. Tropsha, A. Best Practices for QSAR Model Development, Validation, and Exploitation. Mol. Inform. 2010, 29, 476–488. [CrossRef] [PubMed] 41. Ferreira, L.L.G.; Andricopulo, A.D. ADMET modeling approaches in drug discovery. Drug Discov. Today 2019, 24, 1157–1165. [CrossRef]