Phytosterols and triterpenes from Morinda lucida Benth. exhibit binding tendency against class I HDAC and HDAC7 isoforms
Abstract
The integral and multifaceted role of histone deacetylases, commonly referred to as HDACs, in the complex molecular mechanisms underlying the initiation, progression, and maintenance of various forms of cancer has been extensively and unequivocally demonstrated by a multitude of comprehensive studies. These enzymes are crucial epigenetic regulators that influence gene expression by modifying the acetylation status of histones and other non-histone proteins. Their aberrant activity in cancer can lead to dysregulated gene transcription, promoting unchecked cell proliferation, evading apoptosis, and fostering metastatic potential. Consequently, the strategic targeting of these enzymes with specific inhibitors has emerged as a major and promising focus within the expansive field of anticancer drug research. While a limited number of synthetic HDAC inhibitors have successfully navigated rigorous clinical trials and gained approval for the treatment of certain malignancies, their clinical utility is often constrained by significant and undesirable side effects, which can range from myelosuppression and cardiotoxicity to gastrointestinal disturbances and fatigue. These limitations underscore a persistent need for safer and more tolerable therapeutic options.
In response to the challenges posed by synthetic HDAC inhibitors, there has been a growing emphasis and concerted effort to explore and identify novel HDAC inhibitory compounds originating from natural sources, positioning them as potentially advantageous substitutes. Natural compounds often possess unique chemical scaffolds, inherent bioavailability, and a historically lower propensity for toxicity, making them attractive candidates for drug discovery. In a focused endeavor to expand the repertoire of potential HDAC inhibitors, this study embarked on a detailed computational investigation. Our primary objective was to meticulously evaluate the binding tendency of a diverse array of compounds derived from *Morinda lucida Benth.*, a plant with documented ethnomedicinal use and reported anticancer potentials from its extracts and isolated constituents. The aim was to identify potent HDAC inhibitors that could serve as promising candidates for the development of novel anticancer therapeutics.
The investigative methodology employed a sophisticated *in silico* approach, specifically molecular docking. A library comprising 49 compounds isolated or putatively derived from *Morinda lucida Benth.* was subjected to docking simulations against selected HDAC isoforms, allowing for the prediction of their binding interactions and affinities. Givinostat, a well-known synthetic HDAC inhibitor currently in clinical development, was included in these simulations to serve as a crucial reference and positive control for comparative assessment. The docking simulations were performed using AutodockVina, a widely recognized computational tool for molecular docking, and the resulting binding interactions, including specific amino acid residues involved in binding and predicted binding modes, were meticulously visualized and analyzed using Discovery Studio Visualizer, BIOVIA, 2016. To further refine our selection of promising candidates, the top seven compounds, based on their predicted binding affinities, underwent a comprehensive evaluation of their druglikeness properties and key Absorption, Distribution, Metabolism, and Excretion (ADME) parameters. This essential pharmacokinetic assessment was performed using the Swiss online ADME web tool, which provides predictive insights into how a compound might behave within a biological system, including its oral bioavailability, permeability, and potential for metabolic liabilities.
The results of our thorough computational screening revealed compelling findings. Out of the 49 *Morinda lucida*-derived compounds analyzed, a select group of five compounds consistently exhibited high predicted HDAC inhibitory activity, demonstrating remarkably strong binding affinities toward the selected HDAC isoforms when compared directly to the reference compound, givinostat. This promising subset included three distinct phytosterols: campesterol, cycloartenol, and stigmasterol. Additionally, two triterpenes, oleanolic acid and ursolic acid, were identified within this highly active group. Beyond their strong predicted inhibitory potential, these five lead compounds also successfully fulfilled the stringent criteria for oral drugability, adhering to Lipinski’s Rule of Five, a widely accepted guideline for assessing the drug-like properties of a molecule based on its physicochemical characteristics. This adherence suggests they possess favorable properties for oral administration, including appropriate molecular weight, lipophilicity, and hydrogen bond donor/acceptor counts, which are critical for absorption and distribution *in vivo*.
In conclusion, the phytosterols and triterpenes derived from *Morinda lucida* demonstrate a high binding tendency toward the selected HDAC isoforms. Moreover, these compounds exhibit excellent drugability characteristics, making them exceptionally promising candidates for subsequent and more extensive *in vitro* and *in vivo* experimental studies. Their identification represents a significant step forward in the ongoing search for novel and potent therapies specifically designed to counteract abnormalities linked with the over-activity of HDACs, particularly in the context of anticancer drug development and potentially other conditions where HDAC dysregulation plays a pathological role.
Introduction
Histone deacetylases (HDACs) represent a crucial family of enzymes that play a pivotal role in the epigenetic modulation of genomic activity. Their profound influence on gene expression, through the dynamic regulation of histone and non-histone protein acetylation, positions them as highly attractive and exploitable targets for the therapeutic intervention in a myriad of human diseases, most notably cancer. Cancer, indeed, remains one of the most formidable and pressing global health challenges of our time. Devastatingly, it was responsible for an estimated 9 million fatalities in 2015, translating to a staggering statistic where cancer accounts for approximately one in every six deaths reported worldwide. A disproportionate burden of this global mortality, roughly 70%, falls upon low and middle-income nations, highlighting significant disparities in cancer care and prevention. Fundamentally, cancer is characterized by deeply altered cellular chemistry that disrupts the delicate and tightly regulated control mechanisms governing essential cellular processes, including the cell cycle, angiogenesis (the formation of new blood vessels crucial for tumor growth), apoptosis (programmed cell death), and other vital cellular functions that maintain tissue homeostasis. It has now been unequivocally confirmed that beyond direct genetic mutations in DNA sequences, the relentless progression of cancer can also be driven by profound alterations in epigenetic processes. These epigenetic modifications occur without any change in the underlying DNA sequence itself, yet they critically impact gene expression patterns. Key epigenetic mechanisms implicated in carcinogenesis include the methylation of DNA, various modifications of histone proteins (the spools around which DNA is wrapped), and the deregulation of non-coding RNAs. The indispensable role of these epigenetic mechanisms in the complex trajectory of cancer development has been consistently demonstrated by numerous scientific investigations, with histone acetylation standing out as one such key and highly dynamic epigenetic process.
Histone acetylation is maintained in a finely balanced and reversible equilibrium within the cell, meticulously orchestrated by the opposing enzymatic activities of histone acetyltransferases (HATs), which add acetyl groups, and histone deacetylases (HDACs), which catalyze the removal of these acetyl groups from the N-terminal lysine side chains of both histone and a growing number of non-histone proteins. This reversible post-translational modification is one of the most common and critical events in regulating chromatin structure and gene transcription. Aberrant protein acetylation has been directly and consistently linked to oncogenesis, while the abnormal expression or dysregulated activity of HDACs has been firmly established in a wide array of various cancer types. The central and critical role of HDACs in this important epigenetic regulatory process has, therefore, propelled them to the forefront as promising and actionable targets for anticancer drug discovery. Consequently, the development and identification of potent HDAC inhibitors (HDIs) have become a major and highly active area of contemporary biomedical research.
HDACs are broadly categorized into four distinct classes based on their sequence homology and catalytic mechanisms. Class I HDACs comprise four members: HDAC1, HDAC2, HDAC3, and HDAC8. These isoforms are typically nuclear and have been extensively implicated in cancer progression and cell proliferation, making them prime targets for therapeutic intervention. Class II HDACs consist of six isoforms, further subdivided into class IIa (HDAC4, HDAC5, HDAC7, and HDAC9) and class IIb (HDAC6 and HDAC10). These are generally found in both the nucleus and cytoplasm and regulate a broader range of cellular processes. Class III HDACs are mechanistically distinct and are referred to as sirtuins (SIRT1-7). Unlike the other classes, sirtuins are nicotinamide adenine dinucleotide (NAD+)-dependent for their enzymatic activity, connecting them to cellular metabolism. Finally, HDAC11 is the sole member of Class IV, exhibiting unique characteristics. Importantly, Class I, II, and IV HDACs are collectively referred to as “classical HDACs” due to their shared dependency on zinc ions (Zn2+) for their enzymatic function.
A typical and pervasive observation in human cancer is the widespread deregulation of both DNA methylation patterns and post-translational histone modifications. This epigenetic imbalance has grave consequences, leading to profound deregulation of gene transcription, which drives the uncontrolled proliferation and survival characteristics of malignant cells. Numerous specific examples highlight the vital role played by particular HDAC isoforms in the progression of various cancers. For instance, HDAC1, HDAC2, and HDAC3 are known to meticulously regulate the cell cycle by influencing the expression of genes crucial for cell cycle progression, such as p21. Overexpression of HDAC1 has been widely reported in aggressive human cancers, including gastric, breast, prostate, and colon cancers, correlating with poorer prognoses. Similarly, the overexpression of HDAC2 has been implicated in gastric and cervical cancer, while its induction following the loss of the APC (adenomatous polyposis coli) gene has significant implications for colorectal tumor development. Elevated levels of HDAC3 have been consistently reported in colon cancer specimens. Moreover, the over-activity of both HDAC2 and HDAC3 has been directly associated with the active proliferation observed in breast tumors, further highlighting their oncogenic potential. Studies have also firmly linked HDAC3 to the precise regulation of cell proliferation and differentiation in various colon cancer models. Overexpression of HDAC8 has been specifically reported in childhood neuroblastoma, a highly aggressive pediatric cancer, while the involvement of both HDAC1 and HDAC8 has been implicated in breast cancer metastasis, a critical process that determines patient outcome. Furthermore, it has been found that the expression of HDAC7 was selectively reduced by apicidin, a confirmed HDAC inhibitor, in salivary mucoepidermoid carcinoma cells, demonstrating that the suppression of HDAC7 exerts a potent anti-tumor effect in these specific cancer cells.
The therapeutic potential of HDIs is multifaceted. By inhibiting HDACs, these compounds effectively increase the level of acetylated lysine residues on core histones. This crucial epigenetic modification then leads to the re-expression of previously silenced tumor suppressor genes and other regulatory genes in cancerous cells, thereby initiating a program of differentiation, cell cycle arrest, or apoptosis. Additionally, HDIs exhibit significant antiangiogenic activities by downregulating the expression of proangiogenic genes, such as basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), and hypoxia-inducible factor 1-α, thereby starving tumors of their blood supply. It is thus unequivocally evident that HDACs are profoundly and intrinsically linked with the development and progression of cancers. Consequently, the strategic targeting of these enzymes with specific inhibitors represents a major and highly active focus in contemporary anticancer drug research. Encouragingly, a few HDIs have successfully gained approval from the FDA for cancer treatment, a significant milestone that has further fueled intense interest in identifying new and more effective HDIs as potential anti-cancer agents. However, a persistent challenge in targeting these HDACs with synthetic HDIs lies in their significant and often dose-limiting side effects, which can include atrial fibrillation, thrombocytopenia, QT prolongation, pervasive tiredness, profound fatigue, anorexia, diarrhea, and nausea. These adverse effects underscore the urgent clinical need for safer and more tolerable alternatives.
In light of these limitations, substantial emphasis has been placed on the discovery and development of natural product-derived HDIs, positioning them as potentially superior substitutes for their synthetic counterparts. Natural compounds often offer unique chemical diversity, inherent bioavailability, and a generally lower propensity for toxicity, making them attractive candidates for novel drug development. For instance, the HDAC-inhibitory activity of apigenin, a plant-derived flavone, has been thoroughly reported, with its antiproliferative activity against breast cancer cells attributed to its ability to induce cell cycle arrest and interfere with HDAC function. Similarly, the HDAC inhibitory activity of luteolin, another plant-derived flavonoid, has been demonstrated in epithelioid cancer cells. Furthermore, apicidin, a fungal metabolite, is a well-established and confirmed HDAC inhibitor. Other notable HDAC inhibitors derived from natural sources include curcumin, linoleic acid, pomiferin, psammaplin A, romidepsin, stigmasterol, sulforaphane, trapoxin A, trapoxin B, and trichostatin A. Therefore, in a concerted effort to identify a broader spectrum of HDIs with therapeutic potential for anticancer drug development, our current study meticulously screened a diverse array of compounds isolated from *Morinda lucida Benth.* (a plant belonging to the Rubiaceae family) for their prospective HDAC inhibitory activities, employing advanced *in silico* methods.
The anticancer potentials of extracts and compounds derived from *Morinda lucida Benth.* have been corroborated by numerous scientific investigations. *Morinda lucida Benth.* is a highly valued and popular plant in African ethnomedicine, frequently cited in traditional recipes for the management of various cancers and cancer-related ailments, as revealed by ethnobotanical surveys of traditional medical practitioners in South-Western Nigeria. Specifically, it has been reported that infusions and decoctions prepared from the leaves and stem bark of *Morinda lucida* are traditionally employed in this region for cancer treatment. Beyond anecdotal evidence, scientific studies have provided compelling support for its anticancer properties. A crystalline substance meticulously isolated from the stem bark of *Morinda lucida* was observed to exhibit a significant carcinostatic effect on sarcoma 180 ascites in murine models. Furthermore, an aqueous extract derived from *Morinda lucida* root bark demonstrated clear inhibition in conventional telomerase assays, a strong indicator of potential anticancer activity since telomerase activity is often upregulated in cancer cells. The aqueous extract of *Morinda lucida* leaf also exhibited antiproliferative activity in HL-60 cells, a human leukemia cell line. Moreover, the cytotoxic activity of β-sitosterol, a phytosterol isolated from *Morinda lucida* leaves, has been specifically reported against various prostatic carcinoma cell lines (DU145, PC-3, and LNCAP AS0). The methanolic extracts of *Morinda lucida* leaves were found to be cytotoxic to tested monkey kidney epithelial cells (LLCMK2), indicating general cellular toxicity. Additionally, Molucidin, a compound isolated from *Morinda lucida* leaf, demonstrated cytotoxicity against LoVo (colon cancer) and KATO III (stomach cancer) cell lines. However, despite this growing body of evidence supporting the anticancer activity of *Morinda lucida*-derived compounds, studies elucidating the precise molecular mechanisms by which they exert these effects remain relatively sparse, including the intriguing possibility of HDAC inhibition.
Considering these established facts and the existing knowledge gaps, this study was meticulously designed to investigate the binding tendency of compounds derived from *Morinda lucida* towards specific HDAC isoforms. We focused on class I HDAC isoforms (HDAC1, HDAC2, HDAC3, and HDAC8) due to their well-documented involvement in cancer, and a representative class II isoform, HDAC7. This computational approach was chosen as *in silico* studies have proven to be highly valuable in the rapid and efficient discovery of HDAC inhibitors for potential anticancer therapy. We employed a sophisticated combinatorial *in silico* strategy to predict the precise orientation and binding affinity of various ligands isolated from *Morinda lucida* to these selected HDACs. This approach allowed us to gain crucial mechanistic insights into the preferential binding inclination of *Morinda lucida*-derived compounds towards specific HDAC isoforms. Furthermore, a key objective of this work was to elucidate whether these natural compounds exhibit a similar interaction profile, particularly concerning their binding site and mode of action, when compared to the well-characterized synthetic HDAC inhibitor, givinostat. Moreover, this study aimed to provide granular details regarding the critical active site residues within the HDACs that contribute significantly toward the stability of these compound-enzyme complexes. This structural insight is invaluable for supporting further comprehensive studies on *Morinda lucida*-derived compounds and for guiding their potential development as novel therapies against abnormalities linked with the over-activity of the selected HDAC isoforms.
Materials and Methods
Protein Preparation
The fundamental success of any structure-based computational modeling approach, such as molecular docking, is inherently and highly dependent on the accuracy and correctness of the starting protein structures. Therefore, for this study, the crystal structures of the target HDAC isoforms—HDAC1, HDAC2, HDAC3, HDAC7, and HDAC8—were meticulously retrieved from the widely recognized Protein Data Bank (PDB), identified by their respective PDB IDs: 4BKX, 4LY1, 4A69, 3C10, and 5FCW. Prior to docking, each retrieved crystal structure underwent a rigorous individual preparation process to ensure optimal quality for simulation. This involved the removal of any existing co-crystallized ligands and extraneous water molecules, which might interfere with subsequent docking calculations. Crucially, any missing hydrogen atoms in the protein structures were computationally added using the Autodock v4.2 program, developed by the Scripps Research Institute, to ensure accurate representation of atomic charges and interactions. Following this, non-polar hydrogens were merged, while polar hydrogens were specifically added to each enzyme, essential for correct electrostatic interactions. This entire preparation process was systematically repeated for each HDAC isoform, and the fully prepared structures were then saved into the dockable pdbqt format, ready for molecular docking.
Ligand Preparation
The chemical structures of the ligands used in this study were meticulously prepared to ensure compatibility with the docking software and accurate representation of their molecular properties. The SDF (Structure-Data File) structures of givinostat, which served as our standard reference HDAC inhibitor, and 49 distinct compounds reported to be derived from *Morinda lucida*, were systematically retrieved from the PubChem database, a publicly accessible repository of chemical information. To ensure that these compounds were properly recognized and processed by Autodock tools, their SDF formats were converted into the mol2 chemical format using Openbabel, a versatile cheminformatics toolkit. For each ligand, specific Gasteiger-type polar hydrogen charges were assigned, which is crucial for accurately representing electrostatic interactions during docking. Additionally, non-polar hydrogens were merged with their respective carbons, and the internal degrees of freedom and torsions for each molecule were precisely set, allowing for conformational flexibility during the docking process. Both the prepared protein structures and ligand molecules were then further converted into the dockable pdbqt format using Autodock tools, completing the preparation phase for molecular docking.
Molecular Docking
The core of our computational investigation involved molecular docking simulations, which predict the binding orientation and affinity of ligands to various protein targets. This was meticulously carried out using AutodockVina, a widely-accepted and robust molecular docking software. The prepared pdbqt files of individual enzyme targets and ligand molecules were systematically loaded into their respective columns within the software interface, and the docking simulations were then executed. Following the completion of the docking runs, a comprehensive cluster analysis was performed. This analysis grouped the predicted binding poses based on root mean square deviation (RMSD) values, which quantify the structural similarity between different conformations. The lowest energy conformation within the most populated cluster was consistently considered to be the most reliable and probable binding solution, reflecting the energetically most favorable and stable interaction. The binding affinities of all compounds for the five selected HDAC isoforms (HDAC1, HDAC2, HDAC3, HDAC7, and HDAC8) were meticulously recorded. Compounds were then ranked based on their calculated affinity scores, with more negative values indicating higher predicted binding affinity. For the purpose of direct comparison with the established synthetic HDAC inhibitor, givinostat, molecular interactions were visualized. Specifically, the binding interactions between HDACs and compounds that exhibited binding affinities equal to or greater than that of givinostat were viewed with Discovery Studio Visualizer, BIOVIA, 2016. The specific binding pockets within each receptor were systematically determined by identifying receptor cavities using Discovery Studio Visualizer. Subsequently, these binding pockets were highlighted using pink spheres, and the various amino acid residues contributing significantly to each pocket’s formation and interaction were precisely labeled, providing granular structural insights.
Druglikeness and Prediction of Absorption–Distribution–Metabolism–Excretion
The assessment of Absorption, Distribution, Metabolism, and Excretion (ADME) parameters is of paramount and critical importance in the early stages of new drug molecule development. These parameters play a decisive role in defining not only the appropriate dose amount and dosing interval but also the overall safety margins of a potential therapeutic agent. *In silico* methods for the determination of ADME parameters are highly valuable due to their low resource requirements and ability to rapidly screen numerous compounds. These methods rely on theoretically derived statistical models, which are generated by establishing relationships between the structural characteristics of compounds (that have been experimentally measured in a given assay) and their observed biological responses. This computational approach has gained widespread acceptance and utility in drug discovery. Therefore, based on their superior negative binding energy compared to givinostat and their ability to dock at the same catalytic pocket as givinostat, the top seven compounds identified from the molecular docking simulations were carefully selected for further assessment. These seven compounds included campesterol, cycloartenol, lucidin 3-0-ß-primveroside, morindin, oleanolic acid, stigmasterol, and ursolic acid. A comprehensive druglikeness study and ADME prediction were carried out using the Swiss online ADME web tool, a robust and widely used platform. This predictive analysis aimed to evaluate the drug-like properties of these selected phytochemicals, particularly focusing on those that demonstrated selective binding to more than one protein drug target, thereby providing crucial insights into their potential pharmacokinetic behavior *in vivo*.
Results
Our extensive molecular docking analysis provided a comprehensive binding profile of the 49 *Morinda lucida*-derived compounds against the five selected HDAC isoforms (HDAC1, HDAC2, HDCA3, HDCA7, and HDCA8), which was then systematically compared with the binding affinity of givinostat, our reference HDAC inhibitor. The compounds demonstrating the highest binding affinity to each corresponding HDAC isoform were specifically highlighted. Among the entire panel of tested compounds, a subset of 14 natural products exhibited notable binding potential. These included asperuloside, asperulosidic acid, campesterol, cycloartenol, damnacanthol, lucidin, lucidin 3-0-ß-primveroside, morindin, munjistin, oleanolic acid, oruwacin, ß-sitosterol, stigmasterol, and ursolic acid. Each of these compounds displayed at least one binding affinity value that was either higher than or equal to the binding affinity of givinostat for at least one of the five HDACs, indicating their potential as inhibitors.
A more focused analysis on the top performers revealed that a select group of seven compounds consistently exhibited higher binding affinities than givinostat across *all five* HDAC isoforms tested. These exceptional compounds were campesterol, cycloartenol, lucidin 3-0-ß-primveroside, morindin, oleanolic acid, stigmasterol, and ursolic acid. It is well-established in computational chemistry that a more negative docking score indicates a higher binding affinity of the ligand toward its receptor. In this context, oleanolic acid emerged as a standout, demonstrating the highest negative binding affinity value of −16.1 Kcal/mol for HDAC1, significantly surpassing givinostat’s affinity of −11.9 Kcal/mol. Other compounds with notably higher or equal binding affinities for HDAC1 compared to givinostat included: lucidin 3-0-ß-primveroside (−15.5 Kcal/mol), morindin (−15.4 Kcal/mol), ursolic acid (−15.4 Kcal/mol), ß-sitosterol (−15.0 Kcal/mol), cycloartenol (−14.9 Kcal/mol), campesterol (−14.5 Kcal/mol), stigmasterol (−13.9 Kcal/mol), oruwacin (−12.5 Kcal/mol), asperulosidic acid (−12.4 Kcal/mol), munjistin (−12.3 Kcal/mol), and asperuloside (−11.9 Kcal/mol). Notably, lucidin 3-0-ß-primveroside, morindin, ursolic acid, cycloartenol, campesterol, and stigmasterol were predicted to occupy the same catalytic binding pocket as givinostat in HDAC1. The amino acids comprising this shared binding pocket included TYR15, ASP16, GLY17, ARG36, HIS39, ASN40, THR196, VAL198, ASP248, ILE249, PHE252, ASP256, LYS260, ASP248 (repeated, likely an error in original text), ASP332, MET329, and TYR333. All of these compounds were found to interact with two or more of these critical amino acids through various types of bonds, including hydrogen bonds, alkyl interactions, or covalent bonds.
Turning to HDAC2, a range of *Morinda lucida*-derived compounds displayed superior binding affinities compared to givinostat (−11.6 Kcal/mol). These included asperulosidic acid (−14.7 Kcal/mol), campesterol (−13.0 Kcal/mol), cycloartenol (−15.0 Kcal/mol), damnacanthol (−12.0 Kcal/mol), lucidin (−12.2 Kcal/mol), lucidin 3-0-ß-primveroside (−16.9 Kcal/mol), morindin (−16.0 Kcal/mol), oleanolic acid (−15.1 Kcal/mol), oruwacin (−12.8 Kcal/mol), ß-sitosterol (−13.7 Kcal/mol), stigmasterol (−13.3 Kcal/mol), and ursolic acid (−14.3 Kcal/mol). However, the docking analysis revealed that these compounds occupied three distinct binding pockets within HDAC2. Givinostat, along with campesterol, cycloartenol, lucidin 3-0-ß-primveroside, and morindin, occupied the same primary binding pocket. The amino acids constituting this pocket involved LYS148, AP186 (likely an error for ASP186 or ALA186), GLY187, GLU190, TYR193, THR213, ARG217, GLY220, ALA221, ASP239, GLU240, GLY243, GLN244, PRO248, LYS288, GLU365, GLN369, LYS368, and PHE372, forming interactions primarily through alkyl contacts and potential covalent bonds. A second distinct binding pocket was occupied by ursolic acid, demonstrating hydrogen bond interactions with ARG311 and ILE310, along with covalent bonding to LYS36 and TYR341. Other amino acids within this pocket included ARG275, TYR308, THR309, ASN312, GLU340, GLN358, THR360. Stigmasterol, in turn, occupied a third unique binding pocket within HDAC2, characterized by interactions with ASP104, HIS145, HIS146, GLY154, PHE155, HIS183, PHE210, ASP269, LEU276, and GLY306, with observed covalent bond interactions.
The binding profiles of the compounds with HDAC3 revealed four distinct binding pockets. Givinostat and cycloartenol were predicted to occupy similar binding pockets, interacting with key residues such as THR116, ASN119, GLU156, LYS159, TYR160, LYS437, PHE440, ARG441, MET445, PHE444, TYR469, TYR470, THR473, and LYS474 through a combination of covalent bonds and conventional hydrogen bonds. A second distinct binding pocket was occupied by morindin and ursolic acid. Morindin formed two hydrogen bonds with ASP57 and ARG301, while ursolic acid similarly formed two hydrogen bonds with ASN302 and PRO340 within HDAC3. Campesterol, oleanolic acid, and stigmasterol were predicted to occupy a third binding pocket in HDAC3. Evidence of hydrogen bonding was observed between GLY143, TYR298 and both campesterol and stigmasterol, alongside alkyl bonding to PHE144, HIS172, and PHE200. Lucidin 3-0-ß-primveroside occupied a fourth distinct binding pocket in HDAC3 and exhibited the highest negative binding affinity value (−15.2 Kcal/mol) for this isoform, surpassing givinostat (−13.3 Kcal/mol). Lucidin 3-0-ß-primveroside demonstrated multiple hydrogen bond interactions with GLU347, GLN349, SER351, GLU459, and THR462.
For HDAC7, ursolic acid displayed the highest negative binding affinity (−17.1 Kcal/mol), significantly outperforming givinostat (−14.0 Kcal/mol). Ursolic acid was predicted to occupy the same binding pocket as cycloartenol (−15.0 Kcal/mol). Both compounds interacted with PHE679, HIS709, and PHE738 through covalent and conventional hydrogen bonds. Givinostat, campesterol, oleanolic acid, and stigmasterol shared similar binding pockets within HDAC7, composed of residues LEU588, TYR589, ASN592, PRO593, SER595, LEU599, LEU604, LEU607, VAL614, VAL621, VAL623, GLY622, and LEU681. Campesterol, oleanolic acid, and stigmasterol formed covalent bond interactions with PRO593, LEU594, LEU604, and LEU607, while oleanolic acid also engaged in a hydrogen bond interaction with LEU597 within the binding pocket. A third distinct binding pocket in HDAC7 was occupied by lucidin 3-0-ß-primveroside and morindin, both of which were observed to interact with LEU516, PHE518, ASP525, SER526, CYS566, ARG568, and ARG655 via hydrogen bonds.
Finally, in the case of HDAC8, oleanolic acid, lucidin 3-0-ß-primveroside, ursolic acid, stigmasterol, cycloartenol, campesterol, ß-sitosterol, and morindin exhibited superior binding affinities compared to givinostat (−13.4 Kcal/mol), with values of −16.1, −16.0, −15.5, −15.3, −14.9, −14.8, −14.8, and −14.6 Kcal/mol, respectively. All these compounds were predicted to occupy the same catalytic binding pocket as givinostat in HDAC8. Givinostat, along with campesterol, cycloartenol, oleanolic acid, stigmasterol, and ursolic acid, showed alkyl interactions with TYR100 and PHE200. Conversely, lucidin 3-0-ß-primveroside and morindin formed hydrogen bond interactions with ALA32 and LYS33 within the HDAC8 binding pocket.
A noteworthy observation from our detailed docking analysis was the remarkable consistency in binding pocket occupancy between several *Morinda lucida*-derived compounds and givinostat across multiple HDAC isoforms. Specifically, campesterol was predicted to occupy the same binding pockets as givinostat in HDAC1, HDAC2, HDAC7, and HDAC8. Cycloartenol displayed similar binding in HDAC1, HDAC2, HDAC3, and HDAC8. Lucidin 3-0-ß-primveroside shared binding pockets with givinostat in HDAC1, HDAC2, and HDAC8, while morindin also occupied the same pockets in HDAC1, HDAC2, and HDAC8. Oleanolic acid shared binding pockets with givinostat in HDAC7 and HDAC8. Stigmasterol showed consistent binding in the same pockets as givinostat in HDAC1 and HDAC7, and ursolic acid similarly in HDAC1 and HDAC8. Notably, campesterol, cycloartenol, lucidin 3-0-ß-primveroside, morindin, stigmasterol, and ursolic acid not only occupied the same binding pocket as givinostat in HDAC1 but also exhibited higher negative binding affinities for HDAC1 than givinostat, indicating stronger predicted interactions. While oleanolic acid displayed the highest binding affinity for HDAC1, it was predicted to bind at a different pocket than givinostat. For HDAC2, campesterol, cycloartenol, lucidin 3-0-ß-primveroside, and morindin all occupied the same binding pocket as givinostat and also demonstrated higher binding affinities for this HDAC. Cycloartenol was unique in occupying the same binding pocket as givinostat in HDAC3, while also displaying a higher affinity. In HDAC7, campesterol, oleanolic acid, and stigmasterol all shared the same binding pocket as givinostat and exhibited higher negative binding values for this enzyme. Furthermore, campesterol, cycloartenol, lucidin 3-0-ß-primveroside, morindin, oleanolic acid, stigmasterol, and ursolic acid consistently occupied the same binding pocket as givinostat in HDAC8 and all displayed higher binding affinities for this enzyme.
Following the identification of these promising compounds through molecular docking, the top seven candidates (campesterol, cycloartenol, lucidin 3-0-ß-primveroside, morindin, oleanolic acid, stigmasterol, and ursolic acid) were subjected to ADME (Absorption, Distribution, Metabolism, Excretion) prediction to assess their drug-likeness and pharmacokinetic properties. The results from this ADME study provided crucial insights into their potential oral bioavailability and overall suitability as drug candidates. Oleanolic acid and ursolic acid exhibited remarkably similar physicochemical properties, including molecular weight, number of rotatable bonds, hydrogen bond acceptors, hydrogen bond donors, and octanol-water partition coefficient (MLOGP), each displaying only one Lipinski rule violation. Other compounds with a single Lipinski violation were campesterol, cycloartenol, and stigmasterol (all due to MLOGP > 5). In contrast, lucidin 3-0-ß-primveroside and morindin had a total of three Lipinski rule violations, primarily due to their molecular weight, number of rotatable bonds, and number of hydrogen bond acceptors, suggesting potentially poor oral drugability.
Discussion
Molecular docking stands as a widely accepted and indispensable computational tool in modern drug discovery, primarily utilized to study molecular recognition. Its fundamental aim is to predict with high accuracy the binding mode and binding affinity of a complex formed between two or more constituent molecules with known three-dimensional structures. Such docking studies provide crucial and granular information regarding the precise orientation of inhibitors within the binding pocket of their target protein, offering invaluable insights into the atomic-level interactions. Consequently, molecular docking simulations have been instrumental in the identification of numerous potential therapeutic inhibitors. In line with this established utility, our study identified a select group of seven ligands—specifically campesterol, cycloartenol, lucidin 3-0-ß-primveroside, morindin, oleanolic acid, stigmasterol, and ursolic acid—that consistently demonstrated more negative binding affinities for all five of the HDAC isoforms (HDAC1, HDAC2, HDAC3, HDAC7, and HDAC8) compared to the reference synthetic inhibitor, givinostat.
The consistently higher binding affinities observed for these natural compounds may be attributed to their unique ability to favorably interact with key amino acid residues located within the catalytic site of these HDAC isoforms. This interaction is hypothesized to effectively prevent the binding of the essential zinc ion (Zn2+), which serves as a crucial cofactor required by Class I, Class II, and Class IV HDACs for their catalytic activity. By chelating or otherwise disrupting the zinc ion coordination, these compounds could effectively inhibit HDAC enzymatic function. This mechanism of action could be highly beneficial in the treatment of cancer, given the substantial evidence highlighting the pervasive role of HDACs in promoting uncontrolled proliferative cell growth, fostering undifferentiated phenotypes, driving aberrant angiogenesis, enhancing metastatic potential, and contributing to the genetic instability that fundamentally characterizes cancerous cells.
It is noteworthy that plant-derived flavones such as apigenin and luteolin have also been previously reported to exhibit similar binding patterns to selected HDACs, positioning them as potential anticancer agents due to their favorable negative binding energies and demonstrated stability within their respective binding pockets. Both apigenin and luteolin were specifically reported to bind to the HDAC2 pocket, interacting with residues such as HIS145, PHE155, and TYR308, mirroring interactions observed for some of the *Morinda lucida*-derived compounds in our current study. This striking consistency further strengthens the claim that compounds derived from *Morinda lucida* are potent inhibitors of various HDAC isoforms, as predicted *in silico*. Similarly, many of the *Morinda lucida*-derived compounds displayed more negative binding affinities when compared to the value of −7.07 Kcal/mol obtained for suberoyl anilide hydroxamic acid (SAHA), a widely known and potent inhibitor of Class II HDACs, further underscoring their predicted inhibitory strength.
Among the promising compounds identified, campesterol and stigmasterol are prominent examples of phytosterols, which are plant sterols structurally similar to cholesterol. Cycloartenol is also a phytosterol, sharing structural and functional similarities with campesterol and stigmasterol. The anticancer capabilities of these three phytosterols against various tumors have been consistently confirmed by both extensive epidemiological studies and rigorous experimental investigations. The proposed mechanisms underlying the anticancer activity of phytosterols are diverse and multifaceted. These include their ability to influence cell membrane structure and function, thereby impacting critical signaling events; their effects on signal transduction pathways that regulate cell proliferation and apoptosis, often shifting the balance towards cell death; their modulation of immune function; and their influence on cholesterol metabolism within cancer cells. Furthermore, the antiangiogenic activity of both campesterol and stigmasterol, meaning their ability to inhibit the formation of new blood vessels that feed tumors, has also been reported. However, reports specifically detailing the inhibition of HDACs by these phytosterols have been remarkably scanty in the scientific literature. The compelling results from our molecular docking study now provide strong *in silico* evidence indicating that the inhibition of HDACs could indeed represent yet another crucial mechanism by which these phytosterols exert their anticancer activity. This suggests that these phytosterols may have significantly contributed to the previously reported antiproliferative activity of *Morinda lucida* extracts, possibly through their direct inhibition of HDACs. This intriguing observation certainly necessitates further experimental studies, utilizing other complementary *in vitro* and *in vivo* models, for definitive confirmation.
Moreover, numerous studies have confirmed that various triterpenoids, including oleanolic acid and ursolic acid, are widely distributed in different plant species and consistently possess notable antitumor properties. While the mechanism of antitumor activity of triterpenoids has been broadly attributed to their ability to induce apoptosis, block the activation of nuclear factor-κB (NF-κB), activate specific transcription factors, inhibit angiogenesis, and interfere with various signal transduction pathways, the inhibition of HDACs has also been reported for some triterpenoids. Specifically, previous research has shown that treatment with ursolic acid significantly increased histone H3 acetylation and markedly decreased the activities of HDAC1, HDAC3, HDAC4, HDAC5, and HDAC6 in HL60 cancer cells, directly linking it to HDAC inhibition. Building upon this, the results from our current docking study indicate that the anticancer activity of *Morinda lucida Benth.* might be partly attributed to the actions of oleanolic acid and ursolic acid, quite possibly through their ability to inhibit histone deacetylases. However, analogous to the phytosterols, further rigorous evaluation using other experimental models is required to confirm this proposed mechanism and its contribution *in vivo*.
The Lipinski Rule of Five is a widely accepted set of guidelines that describes the relationship between key pharmacokinetic and physicochemical parameters, serving as a rapid tool to determine the “druglikeness” of compounds, particularly for oral bioavailability. Lipinski’s rule generally states that an orally active drug is likely to have no more than one violation of the following four criteria: (1) it should have no more than 5 hydrogen bond donors (defined as nitrogen or oxygen atoms with one or more hydrogen atoms); (2) it should have no more than 10 hydrogen bond acceptors (defined as nitrogen or oxygen atoms); (3) its molecular mass should be less than 500 Daltons; and (4) its calculated octanol–water partition coefficient (log P) should not be greater than 5. From the results obtained in our ADME study, lucidin 3-0-ß-primveroside and morindin each had a total of three Lipinski violations. These violations were primarily due to their higher molecular weight, a greater number of rotatable bonds, and a higher count of hydrogen bond acceptors, indicating potential issues with oral absorption and permeability. As such, despite their demonstrated superior HDAC inhibitory ability for all five HDAC isoforms, lucidin 3-0-ß-primveroside and morindin might not be optimal candidates for oral drug development due to their unfavorable pharmacokinetic profiles. However, campesterol, cycloartenol, oleanolic acid, stigmasterol, and ursolic acid all exhibited only a single Lipinski violation (specifically, MLOGP > 5). Thus, based on this ADME study, these five compounds—campesterol, cycloartenol, stigmasterol, oleanolic acid, and ursolic acid—largely fulfill the criteria for oral drugability as defined by Lipinski’s Rule of Five. This finding aligns with earlier reports suggesting that oleanolic acid and ursolic acid are generally considered quite non-toxic and have been widely incorporated into various health and cosmetics products, further supporting their safety profile. Taken together, these results strongly suggest that these five compounds represent promising candidates for oral HDAC inhibitory drug development. This assertion, however, is subject to further critical studies, including comprehensive bioavailability assessments and rigorous evaluation in other relevant experimental models to confirm their efficacy and safety *in vivo*.
Conclusion
The judicious application of computational methods has proven to be of immense value in the expedited identification and development of novel and promising therapeutic compounds. This approach is particularly advantageous due to its inherently low resource requirements and its ability to rapidly screen a vast number of molecules, thereby significantly streamlining the early stages of drug discovery compared to traditional, resource-intensive experimental procedures. In this comprehensive study, we meticulously evaluated 49 distinct compounds derived from *Morinda lucida*. Our rigorous *in silico* analysis unequivocally identified five prominent natural products: campesterol, cycloartenol, oleanolic acid, stigmasterol, and ursolic acid. These compounds not only demonstrated a high potential as potent inhibitors for all five of the selected HDAC isoforms (HDAC1, HDAC2, HDAC3, HDAC7, and HDAC8), exhibiting superior binding affinities compared to a known synthetic inhibitor, but they also emerged as excellent candidates for further drug development. This dual promise is underscored by their favorable performance in the assessment of oral drugability, adhering largely to Lipinski’s Rule of Five. Consequently, these phytosterols and triterpenes warrant intensive further studies, including rigorous *in vitro* and *in vivo* experimental validation, to fully elucidate their therapeutic potential in the search for effective therapies against abnormalities linked with the over-activity of HDACs, particularly in the context of cancer and other epigenetic disorders.