Introduction

Numerous blockbuster medicines are produced, either directly or indirectly, from plants, which are the major source of novel pharmacologically active chemicals (Dehelean et al. 2021). Plants play a vital role in the treatment and prevention of diseases, even in conjunction with synthetic chemistry as a technique of drug discovery and production from plant scaffold molecules. Natural products will remain a vital source of therapeutic medicines. Many other natural products can be used as chemical models or as models for synthesis, and semi-synthesis of new molecules meant to cure human ailments, in addition to the ones that have been shown to have direct medicinal applications (Emhemmed et al. 2022).

The biological process known as the epithelial-to-mesenchymal transition (EMT) is important for several physiological and pathological circumstances, including cancer and tissue homeostasis. It converts epithelial cells into mesenchymal cells, which enhances their migratory and invasion potential while decreasing their ability to adhere and undergo apoptosis (Beach et al. 2011; Hamidi et al. 2022). While the EMT process increases cell motility and the production of mesenchymal markers (N-cadherin, fibronectin, and vimentin), it also attenuates cell–cell adhesion and downregulates epithelial indicators (E-cadherin) (Serrano-Gomez et al. 2016). Moreover, it is connected to drug treatment resistance, metastasis, and tumor growth. Through EMT, tumor cells at the primary tumor site can become migratory and invasive, which helps them spread to other organs and eventually metastasis (Pan et al. 2021; Huang et al. 2022). The signaling pathway RAF/MEK/ERK is responsible for controlling various cellular functions, such as cell division, proliferation, motility, and survival. In general, ERK1/2 activation stimulates cell proliferation, and many malignancies are characterized by its dysregulated activity (Sugiura et al. 2021). Additionally, the PI3K/Akt pathway plays a critical role in EMT by triggering downstream effectors that control cellular functions such as invasion, migration, and cell survival (Wei et al. 2019; Navaei et al. 2021). Furthermore, a variety of downstream targets, including transcription factors (FOXOs), cell cycle regulators (p21 and p27), and elements of the mTOR pathway (mTOR and p70S6K), are phosphorylated and regulated by active Akt kinase (Johnson et al. 2010; Farhan et al. 2017).

The usage of innovative psychoactive drugs that contain synthetic cannabinoids is on the rise (Castaneto et al. 2014). Products containing synthetic cannabinoids have effects resembling those of natural cannabis, but they are stronger, more hazardous, and have been linked to harmful side effects. A variety of psychotropic substances, primarily with high-potency cannabinoid receptor binding, are included in synthetic pharmaceuticals. The effects of natural cannabis and Δ9-tetrahydrocannabinol are mimicked by these synthetic drugs, but they cause more severe side effects, such as chest pain, tachycardia, anxiety, cognitive impairment, agitation, respiratory difficulties, muscle twitches, acute renal failure, and psychosis (Cohen and Weinstein 2018).

Numerous studies have been conducted on the application of cannabinoids as an anti-cancer treatment. It was found that it generally has beneficial and protective effects, preventing the growth and spread of tumors and reestablishing homeostasis. Therapeutic trials on the use of cannabinoids as an anti-cancer medication are currently being conducted, even though their therapeutic use in palliative care is well documented (Tomko et al. 2020).

It is anticipated that the pharmacokinetic and molecular docking data of cannabinoids and the proteins related to MAPK-ERK signaling pathways will help ensure that these drugs are successfully deciphered and developed into oncological healthcare since drug repurposing is a much faster and more cost-effective process than the de novo introduction of a new drug into the clinic.

Materials and methods

Ligand preparation

Through the PubChem database, we downloaded cannabinoids as SDF file and saved in “.pdbqt” format using BIOVIA/ Discovery Studio 2021 (Table 1).

Table 1 PubChem CIDs of ligands
Full size table

Optimization of the proteins

The 3D crystal structure of all the target proteins was downloaded as the Protein Data Bank (PDB) from https://www.rcsb.org/). Discovery Studio Visualizer 2021 was run to crystallize the target proteins with ligands. For this, all water molecules, small molecules, and ligands were deleted from the protein crystal structure and the optimized proteins were saved in pdbqt format (Table 2).

Table 2 PDB ID of selected proteins
Full size table

Molecular docking analysis

The molecular docking analysis was performed on the PyRx Virtual screening tool (version 0.9) that assesses the suitable binding alignments of the ligands as well as the targeted proteins. The ligands are docked to the protein surface. Using this information, the optimal orientations of the ligand with the best binding affinities for the protein active sites were determined. Discovery Studio Visualizer 2021 was run to exhibit the ligands and protein binding with the respective amino acid residues. Based on the lowest binding affinities needed for the binding ligand to attach to the protein, the optimal poses were selected. The binding box of protein-ligands was built in Auto Dock software. The binding box-related files were analyzed in Auto Dock Vina built-in PyRx software, and PDBQT files were created.

Homology modeling for Protein structure validation

Homology modeling was performed to validate the structure of the optimized protein data bank before molecular docking. A program called PROCHECK was used to validate modeled proteins. PROCHECK generates a Ramachandran plot and assesses the atomic distances, surface area, bond angle, and torsion angles (Vyas et al. 2012). The Ramachandran Plot was provided information on stable conformations of amino acid residues in term of phi (φ) and psi (ψ) angels as well as allowed and disallowed region for amino acid residues in high resolution, non-homologus protein crystal structures. Plot points represented the torsion angles of amino acid residues in a three-dimensional protein model.

Pharmacokinetics and drug-likeness predictions

In computer-based drug development, pharmacokinetic, pharmacochemical, and drug-likeness studies have gained a lot of attention; they are used to determine the pharmacological structure by using the website (https://www.swissadme.ch). To create SMILES, the chemical structure of cannabinoids was drawn on Marvin and then immediately entered into the webpage to start the prediction process (Daina et al. 2017).

Results and discussion

Binding affinities of protein-ligands interactions

Molecular docking of cannabinoids with some of the proteins (ERK/MEK, and P13K/Akt/mTOR) of cancer pathways and EMT-related proteins (E-cadherin and vimentin) was determined using PyRx software. The maximum binding affinity was exerted by the tetrahydrocannabinol-MEK complex (− 0.8.8 kcal/mol), and the second highest binding affinity was observed in the cannabidiol-P13K complex (− 0.8.5 kcal/mol). The binding affinities were increased in order: cannabigerol < cannabidiol = cannabichromene < tetrahydrocannabivarin (Table 3). Few studies have evidently shown that cannabinoids interact with the MAPK-ERK -ERK signaling pathway and one earlier study depicted that cannabinoids interacted with and downregulated the MAPK-ERK signaling pathway to induce apoptosis in glioma cells (Ellert-Miklaszewska et al. 2005).

Table 3 Molecular docking results of cannabinoids-proteins of Pathways of MEK/ERK and P13K/Akt/mTOR, and proteins related to EMT induction
Full size table

Protein–ligand interactions

The current study of C. sativa compounds provides promising information on the possible effectiveness of these phytochemicals against cancer. Protein–ligand interactions are crucial to drug development and offer an excellent understanding of the simulation. These protein–ligand interactions fall into four categories: ionic, hydrophobic, hydrogen bonds, and water bridges (ur Rashid et al. 2022). Hydrogen bonding, or H-bonds, is crucial for protein folding and interactions with ligands (Yunta 2017). Table 4 presents the binding interaction of cannabinoids (cannabidiol, tetrahydrocannabivarin, cannabigerol, cannabinol, cannabichromene, tetrahydrocannabinol) with proteins of the MAPK-ERK signaling pathway. ERK1- tetrahydrocannabinol complex was interacted by alkyl bonding with amino acid residues: Cys183, Ile48, Leu173, Val56, and Ala69 at the binding pocket (Fig. 1). Hydrogen bonding interactions were developed by MEK-tetrahydrocannabinol complex with amino acid residues of Lys97 and residues (Phe129, val127,leu118, Ile141, Met143, and Ala220) were developed alky bonding at the active site. Amongst the hydrophobic interactions, pi anion was developed between the MEK-tetrahydrocannabinol complex by Ala220 (Fig. 2). Hydrophobic and hydrogen interactions of tetrahydrocannabinol at binding sites of cannabinoid receptors were reported in the former study and developed interaction by Val, Phe, and Tyr residues at the active site of acetylcholinesterase receptor (Furqan et al. 2020; Aviz-Amador et al. 2021).

Table 4 Amino acid interactions of cannabinoids with target proteins at their active sites
Full size table
Fig. 1
figure 1

3 D visualization of ERK1-Tetrahydrocannabinol

Full size image
Fig. 2
figure 2

3D visualization of MEK-Tetrahydrocannabinol

Full size image

P13k-cannabinol was developed H-interaction with amino acid residues Try836, Ile848,cys838, Ile932, and val850 and interacted by covalent bonding (Pi sigma bond) with Met922 residue at binding pockets (Fig. 3). ERK1-cannabinol was developed noncovalent molecular interaction (Pi cation) by Arg189 residue and interacted by covalent bond (alkyl bond) by residues Leu352, Leu86, Arg8, Ile190 at the binding pockets (Fig. 4). Only Akt-cannabinol was found to be interacting via H-bond at the binding site by pro349 residue. An earlier study reported that cannabinol was developed hydrogen and alkyl bonding interaction with cannabinoid receptors (Aviz-Amador et al. 2021).

Fig. 3
figure 3

3D visualization of P13K-Cannabinol

Full size image
Fig. 4
figure 4

3D visualization of ERK1-Cannabinol

Full size image

Validation of protein–ligand complex

The protocol of docking was validated by RMSD values and calculated using the PyRx Virtual screening tool. The Root mean square deviation (RMSD) values were calculated by the mean distance between atoms of a position concerning the best fitting position are measured through only movable heavy atoms. The rmsd/lb (the RMSD lower bound) and rmsd/ub (the RMSD upper bound) are two different RMSD metrics that were provided. Two alternatives of RMSD metrics are given, rmsd/ub (RMSD upper bound)) and rmsd/lb ((RMSD lower bound), rmsd/ub matches every atom in one conformation with itself in the other conformation. The rmsd/ub values range from 8.6 to 2.0 Å, the computed rmsd/lb values fluctuate between 1.0 and 2.0 Å. These findings suggested that binding complex of cannabinoids (cannabidiol, cannabinol, cannabigerol, tetrahydrocannabinol, tetrahydrocannabivarin, and cannabichromene) with target proteins (Vimentin, mTOR, MEK, Akt, JNK, ERK1/2, P13K, P38, and E-cadherin), who gave the maximum binding energies was the accurate binding complex. The RMSD values suggested that cannabinoids and protein complexes are valid and accurate and did not face any damage after binding (Table 5). The present results of rmsd/lb also showed the accuracy of the docking methodology employed in this study by the fact that the RMSD value is less than the threshold value of 2.0 established to assess reliability (Benhander and Abdusalam 2022).

Table 5 Root mean square deviation (RMSD) values, which indicate the average distance between atoms at a point in relation to the best fitting position
Full size table

Protein structure validation

Plots illustrate specific low energy conformations for ϕ (phi) and (psi) or stable conformations of amino acid residues along with favorable and unfavorable regions for amino acid residues in the plot. The result shows that the number of residues of protein in the allowed region is more than 80%. The number of residues is less than 1% in the disallowed region except ERK1 (1.4%). Some proteins such as E-cadherin, vimentin, Akt, JNK, and P38 have zero residues in the disallowed region (Table 6). This suggests that the optimized protein structure of MEK, ERK1/2/, P38, P13K, mTOR, Akt, E-cadherin, and vimentin are suitable for molecular docking and the respective ligands can interact at their binding pockets with stable binding. The majority of points of the Ramachandran plots are situated in favorable regions, suggesting that the majority of the dihedral angles of amino acid residues are in appropriate ranges that assisted to develop stable protein–ligand complex (Hollingsworth and Karplus 2010).

Table 6 Ramachandran plot of each protein shows the overview of allowed and disallowed regions of torsion angle values,
Full size table

The regions in the red indicates favoured region, yellow for allowed region, light yellow for generously allowed region of amino ac

Drug-likeness predictions and pharmacokinetics

The in silico pharmacokinetic and drug-likeness properties of the six cannabinoids from the leaves of C. sativa are reported (Table 7) and the values were predicted by the SwissADME. Lipinski's rule of five, which states that if any small molecule violates more than two of these criteria (molecular weight ≤ 500 g/mol, number of hydrogen bond donors ≤ 5, number of hydrogen bond acceptors ≤ 10, calculated logP ≤ 5), the molecule is said to be impermeable or badly absorbed (Lipinski et al. 1997). This rule was used to predict the drug-likeness of the cannabinoid compounds (Table 7). The cannabinoids have 1 violation (MLOGP > 4.15) of the Lipinski rules of five except tetrahydrocannabivarin (0 violation).All cannabinoids have a good bioavailability score (0.55) that verifies their drug-likeness properties of cannabinoids (Ibrahim et al. 2021). Six cannabinoids meet the requirements for leadlikeness that is, a molecular entity that can be optimized (Teague et al. 1999). However, the synthetic accessibility value of these cannabinoids is less than 6, which suggests their possibility to be synthesized. The primary concept of the SwissADME Synthetic Accessibility (SA) Score is that synthesis ease is correlated with the frequency of molecular fragments in "really" attainable compounds and the score ranges from 1 (very easy to synthesize) to 10 (very hard to synthesize) (Ertl and Schuffenhauer 2009). Furthermore, only tetrahydrocannabivarin has less than 5 values of Consensus Log Po/w which is an average of the five lipophilicity predictions, falling within an acceptable range (Alminderej et al. 2020). According to Lipinski’s rule of five, the range of total polar surface area (TPSA) should be of 0–140 that is benchmark for anticancer drugs and cannabinoids are compliance with Lipinski’s rule of five. The TPSA of all the compounds was between 29.46 and 40.46 Å2 (Jagannathan 2019). Cannabigerol, cannabichromene, tetrahydrocannabivarin, cannabinol, cannabidiol, and tetrahydrocannabinol have numerous rotatable bonds, which are less than 10. This indicates that these compounds have stable conformation and bioavailable if consumed orally (Rai et al. 2023).

Table 7 Physicochemical parameters of cannabinoids leaves of Cannabis sativa using SwissADME
Full size table

The current state of drug development is characterized by pharmacokinetic and safety profiles of novel chemical entities (NCE) (Chen et al. 2018). In the early stages of drug development, several computational techniques could assist us in making predictions about the drug-likeness activity and possible toxicity of novel molecules. The ADMESwiss predicted that all six cannabinoid compounds would be highly absorbed via the intestine.

Predictions were made for the pharmacokinetic characteristics of absorption, distribution, skin penetration, metabolism, biotransformation, and excretion. This prediction tool suggests that tetrahydrocannabinol, tetrahydrocannabivarin, cannabidiol, and tetrahydrocannabinol except cannabichromene and cannabigerol can cross the blood–brain barrier. This prediction makes it clear that every component had high GI absorption, as seen in Table 7. The chemicals that are blood–brain permeant may, upon metabolism, produce toxicants that are damaging to the brain and bloodstream. Potential medications for transdermal and oral delivery can be identified and predicted using the skin permeability model. Cannabigerol was discovered to be the cannabinoid more skin permeant. According to the model, a molecule is considered less skin permeant if its log Kp value is more negative which supports the present findings (Daina et al. 2017).

The potential for cannabinoids to function as either a substrate or an inhibitor of P-gp was assessed; the findings showed that cannabinoids are not substrates of P-gp. An earlier study reported if there is no P-glycoprotein (P-gp) substrate that suggests the compound has high bioavailability and intestinal absorption, this supports present findings (Montanari and Ecker 2015). Prediction reveals that cannabinoids are substrate of CYP1A2 (except cannabigerol), CYP2C19 (except cannabigerol and cannabidiol), CYP2C9 (except cannabichromene and cannabidiol), CYP2D6, and CYP3A4 (Table 8). Previous Cytochrome P450 (CYP) enzymes play a crucial role in drug removal through metabolic transformation, making molecule-enzyme interactions critical (Daina et al. 2017). Because inhibition of these isoenzymes reduces the solubility and accumulation of the drug or its metabolites, it may have unintended negative side effects (Mishra and Dahima 2019).

Table 8 Pharmacokinetics parameters of cannabinoids leaves of Cannabis sativa using by SwissADME
Full size table

All characteristics, including XLOGP3 (-0.7 to + 5.0), Molecular weight (150 to 500 g/mol), Solubility (log S not exceeding 6), and Flexibility (FLEX) (not exceeding 9 rotatable bonds), are within the permissible range, according to the bioavailability radar (Fig. 5). The present finding fall within the standard drug flexibility and solubility criteria according to that Log S for solubility should not be greater than 6. XLOGP3 for lipophilicity should fall between -0.7 to + 6.0. The molecule should have no more than 9 rotatable bonds for flexibility (Cheng et al. 2012). The drug-likeness and drug score of cannabinoids suggest that they are better suited for usage as medicines.

Fig. 5
figure 5

Bioavailability radars (a–f)

Full size image

Conclusion

This study docked six cannabinoids from Cannabis sativa with proteins related to MAPK-ERK signaling pathways and proteins related to epithelial-mesenchymal transition (EMT) induction. The pharmacokinetics and drug-likeness properties of six cannabinoids (tetrahydrocannabivarin, cannabigerol, cannabinol, cannabidiol, tetrahydrocannabinol, and cannabichromene) from C. sativa leaves were also analyzed. All the cannabinoids had expressed good binding affinities and their drug-likeness as well as pharmacokinetics elucidated that they may be used as active drugs or inhibitors to downregulate Akt, mTOR, JNK, MEK, P38, P13K, ERK1/2, vimentin, E-cadherin. According to drug-likeness, pharmacokinetic, and binding affinities, out of the six cannabinoids,. tetrahydrocannabinol and cannabinol may be the best inhibitors of proteins related to the MAPK-ERK signaling pathway and EMT induction. This shows that cannabinoids may be used to formulate excellent anticancer medications. The study suggests that cannabinoids are better suited as drugs and to acquire a better knowledge of these medications, improve safe use, and effective prescribing, more clinical research in the actual patient groups for whom prescribing may be considered is required.