Chrysopogon zizanioides (Vetiver) Essential Oil from Qatar Targets AKT1 and STAT3 in Colorectal and Lung Cancer: GC-MS Profiling, In Vitro Antiproliferative Activity, and In Silico Analyses

Abstract

Background: Chrysopogon zizanioides (L.) Roberty (vetiver) is a perennial medicinal grass with deep aromatic roots traditionally used for several ailments. Its root essential oil (CZEO) is rich in phytochemicals with documented antimicrobial, anti-inflammatory, and antioxidant activities. Although its anticancer potential remains underexplored, the complex phytochemical profile of CZEO positions it as a promising multi-target therapy, particularly for colorectal (CRC) and lung cancers where resistance and pathway redundancy often limit conventional treatments. Therefore, this study aimed to investigate the phytochemical composition and antiproliferative activity of CZEO from Qatar against colorectal (HCT-116) and lung (A549) cancer cells and to elucidate its molecular targets and mechanisms of action in CRC and lung cancer using network pharmacology and in silico approaches. Methods: CZEO was extracted by steam distillation and characterized using GC–MS. In vitro proliferation assays with HCT-116 colorectal and A549 lung cancer cells were conducted using the Alamar Blue assay. The ten most abundant phytochemicals identified by GC–MS were assessed for drug-likeness and ADMET properties and further analyzed through network pharmacology, molecular docking, and molecular dynamics (MD) simulations to elucidate the molecular targets and mechanisms underlying CZEO’s anticancer activity. Results: GC-MS profiling identified 40 compounds, predominantly sesquiterpenoids (93%), including khusimol, β-eudesmol, α-vetivone, and rosifoliol. CZEO inhibited cancer cell viability in a dose-dependent manner, with IC₅₀ values of 62.95 ± 2.19 µg/mL for HCT-116 and 167.82 ± 6.51 µg/mL for A549 cells, demonstrating greater potency against colorectal cancer. CZEO did not affect the growth of normal human neonatal fibroblasts (HDFn), suggesting potential selectivity for cancerous cells. ADMET predictions indicated favorable pharmacokinetics and low toxicity of CZEO’s top 10 abundant compounds (TACs). Network pharmacology revealed 373 and 394 overlapping gene targets between TACs and lung and colorectal cancer, respectively. The overlapping genes were used to construct a protein–protein interaction (PPI) network to identify hub genes. STAT3 and AKT1 consistently emerged as common top-scoring hub genes in both cancers. Molecular docking of TACs showed strong binding affinities of rosifoliol and α-vetivone to AKT1 (−6.20 and −5.93 kcal/mol, respectively) and STAT3 (−5.19 and −5.09 kcal/mol, respectively), surpassing reference inhibitors. MD simulations confirmed stable ligand–protein interactions and structural stabilization, particularly with α-vetivone. Conclusions: CZEO from Qatar exhibits potent antiproliferative activity against colorectal and lung cancer cells, supported by a sesquiterpenoid-rich phytochemical profile. Integrative computational analyses highlight AKT1 and STAT3 as key molecular targets, with rosifoliol and α-vetivone emerging as promising lead compounds. These findings support CZEO as a natural, multi-target anticancer agent, warranting further mechanistic and in vitro and in vivo validation.

1. Introduction

Cancer remains one of the leading global health challenges, accounting for nearly 10 million deaths annually worldwide [1,2]. Lung cancer is the foremost cause of cancer-related mortality, with non-small cell lung cancer (NSCLC) representing the predominant subtype and still exhibiting poor long-term survival despite advances in targeted and immunotherapeutic strategies [3,4]. Colorectal cancer (CRC) likewise ranks among the most frequently diagnosed malignancies and major causes of cancer death, with its molecular complexity and diverse subtypes contributing to variable responses to therapy [5,6,7]. Across solid tumors, intratumoral heterogeneity and the emergence of intrinsic and acquired drug resistance remain key barriers to durable treatment responses [8,9]. These limitations underscore the need for novel, multitarget and less toxic therapeutic options, with increasing attention directed toward natural-product-based agents and plant-derived chemotherapeutic compounds as promising leads for cancer drug discovery [8,10,11].

Natural products have historically served as vital sources of anticancer drug discovery, with nearly 60% of clinically used chemotherapeutics derived from or inspired by plant-based molecules [12]. Relatedly, essential oils (EOs), complex mixtures of mainly volatile compounds, extracted from medicinal or aromatic plants, harbor a wide range of bioactive natural compounds with a plethora of structures and chemistries that could serve as promising chemotherapeutics [13]. Recent studies demonstrate that many EO constituents exert antiproliferative, pro-apoptotic, anti-inflammatory, antioxidant, and anti-angiogenic effects, acting through mechanisms involving mitochondrial dysfunction, oxidative stress modulation, and interference with key oncogenic signaling pathways [13,14,15]. These multifaceted bioactivities highlight EOs as promising candidates for multitarget anticancer intervention and support further investigation of specific EOs such as the root-derived EOs of Chrysopogon zizanioides (L.) Roberty (Poaceae family).

C. zizanioides, commonly known as vetiver or khus, is a perennial grass cultivated for its robust therapeutic, ecological, and economic value. The plant grows in dense clumps with long, narrow, and rigid leaves, typically reaching a height of 1.5–2 m, and its extensive root system, penetrating up to 3–4 m, provides exceptional drought resistance, erosion control, and soil stabilization [16]. Because of its deep, fibrous root architecture, vetiver has been widely adopted in tropical and subtropical regions for slope stabilization, riverbank protection, rehabilitation of degraded lands, and phytoremediation of contaminated soils, where it can tolerate flooding, heavy metals, salinity, and wide temperature fluctuations [16,17,18]. In traditional medical systems, including Ayurveda, Unani, and Traditional Chinese Medicine (TCM), vetiver roots have long been employed in the treatment of fevers, inflammatory disorders, dermatological conditions, and gastrointestinal disturbances, among other conditions [17]. Vetiver-infused water is traditionally consumed for its purported cooling, detoxifying, and carminative effects [17,19] while root decoctions have been applied for antioxidant, analgesic, antirheumatic, and antimicrobial purposes, as well as the management of sprains, headaches, toothaches, oral ulcers, malaria, and urinary tract infections [18]. These uses underscore the relevance of C. zizanioides and provide a rationale for investigating its root-derived metabolites.

C. zizanioides roots essential oil (CZEO), also called ruh khus, is a critical ingredient in the cosmetics and fragrance industry due to its earthy aroma, stability, and antimicrobial and fixative properties [17,20,21], with a global market value exceeding USD 750 million in 2024 [22]. CZEO is characterized by a uniquely persistent, smoky-woody odor profile that makes it a classic base note in fine perfumery, and it is frequently used as a natural fixative to enhance the longevity and harmony of complex fragrance blends [20,23,24]. In addition to perfumery, CZEO is incorporated into aromatherapy formulations, topical preparations, and personal-care products, where its reported calming, sedative, and grounding effects have driven increasing consumer interest in “functional” fragrances [20,21,24]. Apart from its commercial utility, CZEO exhibits a range of pharmacological activities, including antioxidant [17,20,21,23,24,25], antimicrobial [17,20,25,26], anti-inflammatory [20,21,27], and evolving anticancer potential [28]. CZEO bioactivity is mainly attributed to its sesquiterpenoid-rich composition, dominated by khusimol, β-eudesmol, α-/β-vetivone, isovalencenol, and vetiselinenol [20,29]. These metabolites have been implicated in modulating cancer-associated signaling pathways, including PI3K/AKT/mTOR, JAK/STAT, MAPK, and NF-κB [20,30,31], which are especially relevant in colorectal and lung cancers, where AKT1 and STAT3 activation promote tumor growth and therapy resistance [31,32,33,34,35].

Previous studies have reported the cytotoxic effects of CZEO against breast (MCF-7), cervical (HeLa), colon (HCT-116), liver (HepG2), and lung (A549) cancer cells, focusing mainly on apoptosis induction, proliferation inhibition, or disruption of tumor signaling pathways [29,36]. However, the cytotoxic potential of CZEO in colorectal and lung cancer remains underexplored, particularly in studies integrating both experimental and computational approaches. Therefore, we employed a multidisciplinary strategy combining GC-MS chemical profiling, in vitro cytotoxicity assays in HCT-116 and A549 cells, and in silico analyses, encompassing ADMET predictions, network pharmacology, molecular docking, and molecular dynamics (MD) simulation, to provide mechanistic insights into CZEO’s activity, reinforcing its promise as a multi-target natural therapeutic candidate against colorectal and lung cancer.

2. Materials and Methods

2.1. Collection of Plant Material and Essential Oil Extraction

The hydrodistillation extraction process, traditionally used in India to prepare vetiver essential oils known as ruh khus, was employed in this study. Fresh vetiver roots were collected in the early morning (6:00–9:00 a.m.) from the fields of Torba Farm, Al Khor, Qatar (25°34′43.2″ N 51°18′44.3″). To maintain the integrity of the roots, the entire plant is carefully uprooted to preserve the roots. The roots are cut and soaked in water to ensure their cleanliness and maintain their freshness. The roots (10 Kg in 20 L of water) were subjected to steam distillation in a copper distillation system at an extraction temperature of 95–100 °C for 3 h. The hydrosol resulting from the initial extraction process is redistilled twice. The essential oils are extracted from the resulting hydrosols using phase separation with n-hexane. Following the evaporation of the leftover n-hexane, using a rotary evaporator at 45 °C, the C. zizanioides essential oils (CZEOs) were dried over anhydrous MgSO₄ to remove residual moisture, thereby enhancing the oils’ purity and shelf life. For cell treatment, the essential oils were first diluted 1:1 in 20% DMSO, and working dilutions were prepared in cell culture media. DMSO concentration was less than 0.5% in all treatments.

2.2. Gas Chromatography-Mass Spectrometry (GC–MS) Analysis

GC-MS analysis of CZEO was conducted using a SHIMADZU QP2010 Ultra GC-MS (Shimadzu, Kyoto, Japan) equipped with SH-Rxi-5Sil MS capillary column (30 m × 0.25 mm, film thickness 0.25 μm). A 0.5 μL aliquot of CZEO diluted in hexane was injected in split mode. The injector and transfer line temperatures were set at 320 °C and 200 °C, respectively. The column oven temperature was initially held at 50 °C for 3 min, followed by a gradual increase of 2 °C per minute until it reached 250 °C, which was then maintained for an additional 10 min. Helium was the carrier gas at a constant flow rate of 1.19 mL/min. The mass spectrometer was operated with an ion source temperature of 250 °C, an ionization voltage of 0.1 kV, and electron ionization mode, and recording of mass spectra within a scan range of 50–600 m/z. A homologous alkane (C5–C36) series was analyzed under identical conditions to facilitate retention index (RI) calculation. The essential oils constituents were identified by comparing their mass spectra and retention indices (RI) with reference compounds from libraries (NIST14.lib, NIST21.lib, and Wiley229.LIB) and, in some cases, by comparing the acquired mass spectra and retention indices with reference standards run under the same conditions. The relative abundance of each constituent was determined by calculating the percentage peak area in the total ion chromatogram without response factor correction [37].

The GC-MS analysis was run 3 times, and the average peak area was reported ± standard deviation (SD).

2.3. Cell Culture and Cell Viability Assays

Human A549 lung adenocarcinoma, HCT-116 colorectal adenocarcinoma, and human normal neonatal fibroblast (HDFn) cells were procured from the American Type Culture Collection (ATCC; Manassas, VA, USA) and maintained in Dulbecco’s Modified Eagle’s Medium (DMEM; Sigma-Aldrich Co., St. Louis, MO, USA), supplemented with 10% fetal bovine serum (FBS; cat# F9665, Sigma-Aldrich) and antibiotics (penicillin at 100 U/mL and streptomycin at 0.1 mg/mL). The cells were cultured in a humidified incubator at 37 °C with 5% CO₂. The medium was refreshed every 2–3 days, and the cells were subcultured using 0.05% trypsin-EDTA once they reached 80–90% confluence, thereby maintaining exponential growth.

Assessment of cell viability was performed using the Alamar Blue cell viability assay (Invitrogen by Thermo Fisher Scientific, Waltham, MA, USA), as previously described [38,39,40,41,42]. Briefly, 5.0 × 10³ cells were seeded in each well of a 96-well cell culture plate for 24 h. The cells were then treated with various concentrations (6.25, 12.5, 25, 50, 75, 100, 200 μg/mL) of CZEO for 48 h. The cells were washed with 1X PBS and resupplied with 100 μL of complete growth medium containing 10% (v/v) of the Alamar Blue reagent. The cells were incubated for 4 h in the dark in the cell culture incubator at 37 °C. Cell fluorescence was measured at 560 nm (excitation) and 590 nm (emission) using a TECAN Infinite M200 plate reader (Männedorf, Switzerland). Cell viability was expressed relative to the viability of control DMSO-treated cells, the viability of which was set as 100% viability. Percent Cell viability was calculated as:

% Cell viability = (Fluorescence of treated cells/Fluorescence of control cells) × 100

Each treatment was tested in triplicate wells, and each experiment was independently repeated 3 times (n = 3).

2.4. Statistical Analysis

All cell-based experiments were performed as three independent biological replicates (n = 3), each carried out in triplicate wells. Statistical analyses were conducted and GraphPad Prism (version 9; GraphPad Software, Boston, MA, USA). Data are expressed as mean ± standard deviation (SD). Differences among treatment groups were assessed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc test for pairwise comparisons. A p-value < 0.05 was considered statistically significant.

The half-maximal inhibitory concentration (IC₅₀) values were determined by fitting Alamar Blue assay data to a non-linear dose–response curve in GraphPad Prism, plotting percentage growth inhibition against the log-transformed concentration of the extract.

2.5. Drug-Likeness and ADMET Analysis of CZEO Compounds

The 10 most abundant bioactive compounds in CZEO, identified through GC-MS analysis (Table 1), were virtually screened for their physiological properties, drug-likeness, and pharmacokinetic characteristics, focusing on their ADMET properties (Absorption, Distribution, Metabolism, Excretion, and Toxicity). The PubChem database (https://pubchem.ncbi.nlm.nih.gov/, accessed on 25 April 2025) was used to retrieve the canonical SMILES for all 10 compounds [43]. The compounds’ physicochemical properties, lipophilicity, and drug-likeness were analyzed by importing their canonical SMILES into the SwissADME tool (http://www.swissadme.ch/, accessed on 25 April 2025) [44]. The tool applied Lipinski’s rule of five (RO5), including molecular weight (MW), number of hydrogen bond donors (HBD), number of hydrogen bond acceptors (HBA), and logP. Additionally, the total polar surface area (TPSA), molar refractivity (MR), solubility, and bioavailability were predicted for all 10 compounds. The ADMET properties were assessed using the pkCSM pharmacokinetics web tool (https://biosig.lab.uq.edu.au/pkcsm/, accessed on 25 April 2025), predicts pharmacokinetic parameters such as absorption rates, distribution throughout the body, excretion, and the toxicity profiles of the compounds.

2.6. Prediction of CZEO Cellular Target Proteins

The 10 most abundant CZEO compounds in the GC-MS profile (Table 1) and predicted to have good oral bioavailability and exhibit no toxicity, as determined by ADMET analysis above, were selected for compounds’ target prediction. The canonical SMILES of the 10 compounds were input into the SwissTargetPrediction (http://www.swisstargetprediction.ch/, accessed on 28 April 2025) and SuperPred (https://prediction.charite.de/, accessed on 28 April 2025) databases with Homo sapiens set as the study species. In the SwissTargetPrediction database, targets with a probability score of 0 or higher were considered, while in the SuperPred database, only compounds with a probability score of 50% or greater were selected.

Having tested the in vitro antiproliferative activity of CZEO in lung and colorectal cancer cell lines, we aimed to identify the disease target genes in these two cancers. The cancer-related targets were obtained from the Genecards (https://www.genecards.org/, accessed on 28 April 2025), NCBI (https://www.ncbi.nlm.nih.gov/gene/, accessed on 28 April 2025), and MalaCards (https://www.malacards.org/, accessed on 28 April 2025) databases by searching for the terms “lung cancer” and “colorectal cancer” as keywords. Genes from the 3 databases were combined for each cancer, and duplicate gene entries were removed.

2.7. Overlapping of Compound-Disease Target Genes

Further, the overlapping of targets between the CZEO compounds’ targets and the two cancer types was achieved using the Venn tool at the bioinformatics and evolutionary genomics platform (https://bioinformatics.psb.ugent.be/webtools/Venn/, accessed on 28 April 2025). These common genes were considered for the construction of the protein–protein interaction (PPI) network and further analyses.

2.8. Construction of the Protein–Protein Interaction (PPI) Network

The compound-cancer overlapping target genes identified by the Venn tool were used as input for PPI network construction. The intersecting genes were submitted to STRING version 12.0 (http://www.string-db.org/ (accessed on 5 May 2025)) to analyze the direct and indirect connections between these proteins. STRING database creates a PPI network using various types of evidence, such as experimental data, co-expression patterns, and curated database annotations. The PPI network was constructed by filtering Homo sapiens and applying a high confidence of 0.7 to ensure relevance and reliability. Data from the STRING database was downloaded as TSV format and imported into Cytoscape version 3.10.3 (http://www.cytoscape.org/ (accessed on 5 May 2025)) for visualization and further analysis. Cytoscape reveals topology parameters of nodes in the network, including betweenness centrality (BC), closeness centrality (CC), and degree centrality (DC) values. Using the degree score ranking, the Cytohubba plug-in in Cytoscape was utilized to identify the top 10 hub genes in the network. The hub genes were ranked from the highest to the lowest degree scores for each cancer type [45].

2.9. Analysis of the KEGG Pathway and Gene Ontology (GO) Function Enrichment

To understand the biological processes (BPs), molecular functions (MFs), cellular components (CCs), as well as the pathways associated with the identified 10 hub genes, we performed a functional enrichment analysis using the ShinyGO web server (http://bioinformatics.sdstate.edu/go/, accessed on 5 May 2025). The results of the enrichment analysis were presented, with the top 10 enriched GO terms and KEGG pathways (False Discovery Rate cut-off of 0.05), and then displayed as a bubble plot with fold enrichment and gene counts [46,47,48].

2.10. Molecular Docking

Molecular docking studies investigate the interactions within the protein-ligand complexes. This method reveals insights into the intermolecular interactions between small molecules and proteins, revealing interaction modes that can lead to protein inhibition [49]. The protein structures of Serine/Threonine Kinase 1 (AKT1; PDB ID: 4EKL) and Signal Transducer and Activator of Transcription 3 (STAT3; PDB ID: 6NUQ) were downloaded from the Protein Data Bank (https://www.rcsb.org (accessed on 5 May 2025)) in PDB file format [50]. Before docking, each protein was prepared using AutoDock Vina 1.2.0 [50,51,52] by removing water molecules and adding Gasteiger charges and polar hydrogens. The prepared structures were saved in PDBQT file format.

The three-dimensional structures of the 10 phytochemicals and the reference inhibitors of AKT1 (MK-2206; DB16828) and STAT3 (Stattic, PubChem CID 2779853) were downloaded from DrugBank (https://go.drugbank.com/ (accessed on 5 May 2025)) and PubChem, respectively, in SDF file format. The 2D chemical structures of MK-2206 and Stattic are displayed in Figure 3. These ligands were prepared for docking and saved in PDBQT format. A grid box was generated around the binding sites of each protein target to define the docking site, and the exhaustiveness parameter was set to 32 for accurate docking results. To validate the docking protocol, the known inhibitors of AKT1 and STAT3 were redocked to the binding site. The number of poses considered for each ligand was set to 10. Following the docking, the binding scores (in kcal/mol) and interactions such as hydrogen bonds and hydrophobic contacts for the docked complexes were noted to assess the affinity of the protein-ligand complexes. The protein-ligand complexes and docking poses were visualized using the PyMOL Molecular Graphics System, Version 3.0 (Schrödinger, LLC, New York, NY, USA).

2.11. Molecular Dynamics Simulation

Molecular dynamics (MD) simulations analyses commenced with the preparation of coordinate and topology files utilizing the “tLeap” module embedded within the AmberTools21 software suite AMBER21 software suite (version 21, 2021 release) [53,54]. To ensure charge neutrality, the system was immersed within an OPC solvent box and supplemented with ions. Parameterization of the ligand molecules ensued, employing the GAFF2 force field. Energy minimization was then performed using iterative algorithms such as steepest descent and conjugate gradient. Equilibration procedures involved a gradual increase in temperature, facilitated by a thermostat. The computation of long-range electrostatic interactions was achieved using the Particle Mesh Ewald (PME) method, while the Lennard-Jones potential was employed to account for Van der Waals forces [55]. During the equilibration process, we ensured stability through a series of meticulous steps. Initially, we restrained the positions of molecules, gradually increased their temperature, and allowed them to equilibrate freely. To maintain the integrity of covalent bonds and angles, we applied the SHAKE algorithm, while pressure levels were regulated using a barostat algorithm. Once equilibration was achieved, we ran a 200 ns simulation [56]. Post simulation analysis was carried out by using either the CPPTRAJ module [57]. Through these analyses, we investigated various molecular properties, including Root Mean Square Deviation (RMSD) in Å to monitor the structural stability and conformational divergence of the system throughout the trajectory, Root Mean Square Fluctuation (RMSF) representing the atomic positional flexibility, Radius of Gyration (Rg) expressing the overall structural compactness, and hydrogen bonding patterns. These comprehensive evaluations provided invaluable insights into the dynamics and stability of each ligand-protein target system under study [58,59,60].

The following equation was used to calculate the unweighted RMSD:

$$\mathrm{RMSD}=\sqrt{{\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\mathrm{d}}_{\mathrm{i}}^2}{\mathrm{N}}}}$$

(1)

where N is the total number of atoms being considered for the calculation, i is the atom index running from 1 to N, and d_i is the atomic displacement representing the Euclidean distance between the position of atom i in the target frame and its corresponding position in the reference frame.

RMSF was calculated by using the following mathematical expression [33].

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{F}=\sqrt{{\displaystyle \frac{3\mathrm{B}}{8{\mathsf{\pi }}^2}}}$$

(2)

where the thermal factor B is defined as

$$B={\displaystyle \frac{8{\pi }^2}{3}}⟨\mathsf{\Delta }{\mathrm{r}}^2⟩$$

(3)

where ⟨Δr²⟩ denotes the mean square fluctuation of atomic positions about their time-averaged coordinates.

Furthermore, Rg which represents the overall structural compactness and folding state of a protein-ligand complex, was calculated by the following equation:

$${\mathrm{R}}_{\mathrm{g}}^2={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\mathrm{m}}_{\mathrm{i}}{\left({\mathrm{r}}_{\mathrm{i}}-{\mathrm{r}}_{\mathrm{C}\mathrm{M}}\right)}^2}{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\mathrm{m}}_{\mathrm{i}}}}$$

(4)

where m_i is the mass of atom (in amu), r is the position vector of atom i at a given simulation frame, r_CM is the center-of-mass position vector of the selected N atoms of the protein, defined as:

$${\mathrm{r}}_{\mathrm{C}\mathrm{M}}={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\mathrm{m}}_{\mathrm{i}}{\mathrm{r}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\mathrm{m}}_{\mathrm{i}}}}$$

(5)

2.12. Binding Free Energies Calculation of Lead Complexes

To further evaluate the binding affinity of the lead compound–target complexes, binding free energies (BFE) were calculated using the Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) approach. This post-processing method incorporates molecular mechanics energy parameters with continuum solvation models, accounting for van der Waals interactions, electrostatic contributions, and solvation effects, providing more reliable estimates of ligand–protein binding affinity than docking scores alone [61,62].

From each equilibrated molecular dynamics (MD) simulation, 300 snapshots were extracted from the stable region of the simulation trajectory and subjected to MM/GBSA post-processing. The binding free energy (ΔG_bind) was computed for each frame, and the final value was reported as the mean value across all frames. Statistical reliability of the MM/GBSA estimates was quantified by computing the standard error (SE) as the standard deviation of the per-frame binding free energies divided by √N, where N= 300.

The MM/GBSA binding free energy is given by:

$${∆\mathrm{G}}_{\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{d}}={\mathrm{G}}_{(\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{x},\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d})}-{\mathrm{G}}_{\left(\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{a}\mathrm{n}\mathrm{d},\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\right)}-{\mathrm{G}}_{(\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n},\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d})}$$

(6)

The free energy of each species was decomposed as:

$$\mathrm{G}={\mathrm{E}}_{\mathrm{M}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{M}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c}\mathrm{s}}+{\mathrm{G}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}-\mathrm{T}\mathrm{S}$$

(7)

where G_solvated is the solvation free energy, and TS denotes the entropic contribution. Thus,

$${∆\mathrm{G}}_{\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{d}}={∆\mathrm{E}}_{\mathrm{M}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{M}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c}\mathrm{s}}+{∆\mathrm{G}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}-∆\mathrm{T}\mathrm{S}={∆\mathrm{G}}_{\mathrm{v}\mathrm{a}\mathrm{c}\mathrm{c}\mathrm{u}\mathrm{m}}+{∆\mathrm{G}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}$$

(8)

The molecular mechanics energy term was decomposed into:

$${∆\mathrm{E}}_{\mathrm{M}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{M}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c}\mathrm{s}}={∆\mathrm{E}}_{\mathrm{i}\mathrm{n}\mathrm{t}}+{∆\mathrm{E}}_{\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{c}}+{∆\mathrm{E}}_{\mathrm{v}\mathrm{d}\mathrm{W}}$$

(9)

The solvation free energy term was computed as:

$$∆{\mathrm{G}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}=∆{\mathrm{G}}_{\mathrm{G}\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{z}\mathrm{e}\mathrm{d} \mathrm{B}\mathrm{o}\mathrm{r}\mathrm{n}}+∆{\mathrm{G}}_{\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{f}\mathrm{a}\mathrm{c}\mathrm{e} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}}$$

(10)

where ΔG_{Generalized Born} represents the polar solvation energy estimated using the generalized Born (GB) model, and ΔG_{surface area} corresponds to the nonpolar solvation contribution, approximated as:

$$∆{\mathrm{G}}_{\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{f}\mathrm{a}\mathrm{c}\mathrm{e} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}}=\mathsf{\gamma }.\mathrm{S}\mathrm{A}\mathrm{S}\mathrm{A}+\mathrm{b}$$

(11)

with SASA denoting the solvent-accessible surface area, γ the surface tension coefficient, and b an empirical offset constant.

The vacuum-phase contribution was defined as:

$$∆{\mathrm{G}}_{\mathrm{v}\mathrm{a}\mathrm{c}\mathrm{c}\mathrm{u}\mathrm{m}}={∆\mathrm{E}}_{\mathrm{M}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{M}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c}\mathrm{s}}-\mathrm{T}∆\mathrm{S}$$

(12)

3. Results and Discussion

3.1. Identification of CZEO Compounds by GC–MS Analysis

GC–MS profiling was utilized to identify the phytochemicals in CZEO, identifying 40 CZEO metabolites, representing ~87% of the oil’s composition (Figure 1 and Table 1). The majority of the metabolites identified were sesquiterpenoids (93%), followed by sesquiterpenes (6%) and a single monoterpenoid compound (1%). Among the 40 metabolites, 28 were sesquiterpenoids (e.g., isovalencenol, α-vetivol, and khusimol), 11 sesquiterpenes (e.g., d-selinene and α-muurolene, and cyperene), and only one monoterpenoid (pulegone). Studies have reported fingerprint markers of vetiver oil, including α-vetivone, β-vetivone, vetiselinenol, and khusimol [18,63], all of which were consistently detected in our study. Additionally, our extraction method followed the traditional extraction methods using a copper vessel for boiling and re-extracted the hydrosol, and revealed the major compounds as sesquiterpenoids and sesquiterpenes, consistent with earlier reports using comparable extraction methods [64,65,66,67].

Figure 1. An expanded view of a region (40–65 min) of the total ion chromatogram (TIC) of C. zizanioides essential oil showing detected constituents. The full TIC is presented in Figure S1. Time is in minutes.

Figure 1. An expanded view of a region (40–65 min) of the total ion chromatogram (TIC) of C. zizanioides essential oil showing detected constituents. The full TIC is presented in Figure S1. Time is in minutes.

Table 1. Identified compounds in the C. zizanioides essential oil by GC-MS. The 10 most abundant compounds are displayed in bold.

No.	RT	RILit.	RICal.	Identification a	Chemical Class	Area%
1	40.342	1450	1448	Prezizaene	Sesquiterpene	0.38 ± 0.02
2	40.618	1597	1593	Khusimene	Sesquiterpene	0.67 ± 0.01
3	40.925	1474	1472	Valencene	Sesquiterpene	0.14 ± 0.02
4	42.328	1440	1438	α-Muurolene	Sesquiterpene	0.91 ± 0.02
5	42.817	1432	1430	Cyperene	Sesquiterpene	0.71 ± 0.01
6	43.124	1435	1433	g-Muurolene	Sesquiterpene	0.23 ± 0.02
7	43.789	1469	1468	d-Codinene	Sesquiterpene	0.37 ± 0.01
8	44.452	1386	1389	Aromandendrene	Sesquiterpene	0.22 ± 0.02
9	44.601	1523	1520	β-Guaiene	Sesquiterpene	0.53 ± 0.03
10	45.501	1402	1400	6,10-Dimethylbicyclo [4.4.0]decan-1-en-3-one	Sesquiterpenoid	0.82 ± 0.03
11	46.544	1489	1491	β-Vatirenene	Sesquiterpenoid	2.51 ± 0.02
12	47.194	1598	1597	Rosifoliol	Sesquiterpenoid	2.82 ± 0.03
13	47.741	1212	1210	Pulegone	Monoterpenoid	0.44 ± 0.02
14	48.51	1500	1499	3,5,11-Eudesmatriene	Sesquiterpene	0.30 ± 0.02
15	48.948	1580	1577	d-Cadinol	Sesquiterpenoid	0.41 ± 0.02
16	49.312	-	1442	Khusimone	Sesquiterpenoid	1.32 ± 0.01
17	50.435	1626	1628	γ-Eudesmol b	Sesquiterpenoid	1.20 ± 0.01
18	50.567	1615	1613	Germacrene D-4-ol	Sesquiterpenoid	3.04 ± 0.02
19	51.009	1593	1595	β-Eudesmol	Sesquiterpenoid	1.63 ± 0.03
20	52.368	1582	1581	α-epi-Muurolol	Sesquiterpenoid	2.63 ± 0.02
21	52.526	-	2305	Cyclocopacamphenol 1 c	Sesquiterpenoid	1.24 ± 0.03
22	52.71	-	2307	Cyclocopacamphenol 2 c	Sesquiterpenoid	1.51 ± 0.01
23	52.922	-	1669	Epizizanone	Sesquiterpenoid	1.59 ± 0.03
24	53.056	-	1650	Widdrol	Sesquiterpenoid	1.03 ± 0.03
25	53.496	-	1677	Zizanol	Sesquiterpenoid	2.59 ± 0.04
26	53.963	-	1727	Khusiol	Sesquiterpenoid	1.74 ± 0.03
27	54.697	-	1778	Costol 1 c	Sesquiterpenoid	2.37 ± 0.04
28	55.331	1481	1480	d-Selinene	Sesquiterpene	1.11 ± 0.01
29	55.624	-	1571	Isoshyobunone	Sesquiterpenoid	0.81 ± 0.02
30	56.11	1704	1701	Vetiselinenol	Sesquiterpenoid	4.01 ± 0.02
31	56.6	-	1736	α-Valerenol	Sesquiterpenoid	2.11 ± 0.03
32	57.124	1802	1802	α-Vetivol	Sesquiterpenoid	11.13 ± 0.02
33	57.27	1727	1724	Khusimol	Sesquiterpenoid	10.52 ± 0.03
34	58.085	-	1831	Valerenic acid	Sesquiterpenoid	0.71 ± 0.02
35	59.417	1765	1762	Isovalencenol	Sesquiterpenoid	12.05 ± 0.01
36	59.745	-	1780	Costol 2 c	Sesquiterpenoid	1.04 ± 0.03
37	60.446	-	1798	Isovalencenal 1 c	Sesquiterpenoid	0.98 ± 0.02
38	60.62	1792	1790	β-Vetivone	Sesquiterpenoid	3.13 ± 0.01
39	61.235	1800	1797	Isovalencenal 2 c	Sesquiterpenoid	2.73 ± 0.02
40	61.826	1812	1809	α-Vetivone	Sesquiterpenoid	3.35 ± 0.01
				Number of identified compounds	40
				Percentage of identified compounds	87.03 ± 0.06%
				Monoterpenoid	1%
				Sesquiterpenes	6%
				Sesquiterpenoids	93%

RT = retention time; RI_Lit. = NIST 14 mass spectral data retention index; RI_Cal. = retention index relative to alkanes (C5–C36) on SH-Rxi-5Sil MS; Compounds were identified by: ^a comparison with GC-MS spectral libraries NIST14.lib, NIST21.lib and WILEY229.LIB, comparison with spectra found in the literature, or ^b identification confirmed with pure standards injected in the same conditions; ^c Isomers that could not be distinguished.

Table 1. Identified compounds in the C. zizanioides essential oil by GC-MS. The 10 most abundant compounds are displayed in bold.

No.	RT	RILit.	RICal.	Identification a	Chemical Class	Area%
1	40.342	1450	1448	Prezizaene	Sesquiterpene	0.38 ± 0.02
2	40.618	1597	1593	Khusimene	Sesquiterpene	0.67 ± 0.01
3	40.925	1474	1472	Valencene	Sesquiterpene	0.14 ± 0.02
4	42.328	1440	1438	α-Muurolene	Sesquiterpene	0.91 ± 0.02
5	42.817	1432	1430	Cyperene	Sesquiterpene	0.71 ± 0.01
6	43.124	1435	1433	g-Muurolene	Sesquiterpene	0.23 ± 0.02
7	43.789	1469	1468	d-Codinene	Sesquiterpene	0.37 ± 0.01
8	44.452	1386	1389	Aromandendrene	Sesquiterpene	0.22 ± 0.02
9	44.601	1523	1520	β-Guaiene	Sesquiterpene	0.53 ± 0.03
10	45.501	1402	1400	6,10-Dimethylbicyclo [4.4.0]decan-1-en-3-one	Sesquiterpenoid	0.82 ± 0.03
11	46.544	1489	1491	β-Vatirenene	Sesquiterpenoid	2.51 ± 0.02
12	47.194	1598	1597	Rosifoliol	Sesquiterpenoid	2.82 ± 0.03
13	47.741	1212	1210	Pulegone	Monoterpenoid	0.44 ± 0.02
14	48.51	1500	1499	3,5,11-Eudesmatriene	Sesquiterpene	0.30 ± 0.02
15	48.948	1580	1577	d-Cadinol	Sesquiterpenoid	0.41 ± 0.02
16	49.312	-	1442	Khusimone	Sesquiterpenoid	1.32 ± 0.01
17	50.435	1626	1628	γ-Eudesmol b	Sesquiterpenoid	1.20 ± 0.01
18	50.567	1615	1613	Germacrene D-4-ol	Sesquiterpenoid	3.04 ± 0.02
19	51.009	1593	1595	β-Eudesmol	Sesquiterpenoid	1.63 ± 0.03
20	52.368	1582	1581	α-epi-Muurolol	Sesquiterpenoid	2.63 ± 0.02
21	52.526	-	2305	Cyclocopacamphenol 1 c	Sesquiterpenoid	1.24 ± 0.03
22	52.71	-	2307	Cyclocopacamphenol 2 c	Sesquiterpenoid	1.51 ± 0.01
23	52.922	-	1669	Epizizanone	Sesquiterpenoid	1.59 ± 0.03
24	53.056	-	1650	Widdrol	Sesquiterpenoid	1.03 ± 0.03
25	53.496	-	1677	Zizanol	Sesquiterpenoid	2.59 ± 0.04
26	53.963	-	1727	Khusiol	Sesquiterpenoid	1.74 ± 0.03
27	54.697	-	1778	Costol 1 c	Sesquiterpenoid	2.37 ± 0.04
28	55.331	1481	1480	d-Selinene	Sesquiterpene	1.11 ± 0.01
29	55.624	-	1571	Isoshyobunone	Sesquiterpenoid	0.81 ± 0.02
30	56.11	1704	1701	Vetiselinenol	Sesquiterpenoid	4.01 ± 0.02
31	56.6	-	1736	α-Valerenol	Sesquiterpenoid	2.11 ± 0.03
32	57.124	1802	1802	α-Vetivol	Sesquiterpenoid	11.13 ± 0.02
33	57.27	1727	1724	Khusimol	Sesquiterpenoid	10.52 ± 0.03
34	58.085	-	1831	Valerenic acid	Sesquiterpenoid	0.71 ± 0.02
35	59.417	1765	1762	Isovalencenol	Sesquiterpenoid	12.05 ± 0.01
36	59.745	-	1780	Costol 2 c	Sesquiterpenoid	1.04 ± 0.03
37	60.446	-	1798	Isovalencenal 1 c	Sesquiterpenoid	0.98 ± 0.02
38	60.62	1792	1790	β-Vetivone	Sesquiterpenoid	3.13 ± 0.01
39	61.235	1800	1797	Isovalencenal 2 c	Sesquiterpenoid	2.73 ± 0.02
40	61.826	1812	1809	α-Vetivone	Sesquiterpenoid	3.35 ± 0.01
				Number of identified compounds	40
				Percentage of identified compounds	87.03 ± 0.06%
				Monoterpenoid	1%
				Sesquiterpenes	6%
				Sesquiterpenoids	93%

RT = retention time; RI_Lit. = NIST 14 mass spectral data retention index; RI_Cal. = retention index relative to alkanes (C5–C36) on SH-Rxi-5Sil MS; Compounds were identified by: ^a comparison with GC-MS spectral libraries NIST14.lib, NIST21.lib and WILEY229.LIB, comparison with spectra found in the literature, or ^b identification confirmed with pure standards injected in the same conditions; ^c Isomers that could not be distinguished.

Although the major constituents of CZEO are broadly conserved across studies, their relative abundance varies considerably due to geographic, edaphic, and climatic factors [64,65,66,67]. Environmental factors such as temperature, rainfall, and seasonal fluctuations are known to shape the phytochemical composition of a plant. For instance, Kotoky et al. demonstrated that the quality of vetiver oil is closely linked to root metabolic activity, which is strongly modulated by agro-climatic conditions [68]. Cultivation practices also exert a profound impact on both yield and chemical profile. Microbe-assisted cultivation was reported to produce the highest oil content, enriched in low-molecular-weight volatiles, whereas semi-hydroponic cultivation, although associated with lower yields, generated a volatile profile more comparable to that of conventional soil-grown plants [69].

Notably, the same major CZEO constituents could be obtained using non-conventional extraction techniques such as microwave hydrodistillation, supercritical fluid extraction, carbon dioxide expanded-ethanol extraction (CXE), and pressurized-liquid extraction [17]; however, the relative representation of these compounds varied significantly across methods [70]. Classical hydrodistillation tends to rapidly recover highly volatile compounds (up to 95% within the first 30 min), thereby favoring lighter constituents. This explains why David et al. observed a higher relative abundance of light volatiles when analyzing early distillation fractions [71]. In contrast, CXE broadens recovery to include moderately polar and less volatile metabolites by enhancing mass transfer and preserving thermolabile [17]. Similarly, our extended hydrodistillation approach captured both volatile and less volatile constituents, yielding a more comprehensive chemical profile that better reflects the full complexity of CZEO.

3.2. CZEO Reduced the Viability of HCT-116 Colorectal and A549 Lung Cancer Cells

The cytotoxic activity of CZEO was evaluated against HCT-116 colorectal cancer and A549 lung cancer cells using the Alamar Blue viability assay. After 48 h of treatment, CZEO displayed a more potent antiproliferative effect on HCT-116 cells than A549 cells (Figure 2). A concentration-dependent reduction in HCT-116 viability was observed starting at 6.25 µg/mL, with a statistically significant effect (p < 0.001) at 75 µg/mL. In contrast, A549 viability was significantly reduced only at the highest tested concentration of 200 µg/mL (Figure 2). The IC₅₀ values of 62.95 ± 2.19 µg/mL for HCT-116 and 167.82 ± 6.51 µg/mL for A549, confirmed the higher sensitivity of CRC cells. Notably, CZEO did not show cytotoxicity against the normal human fibroblasts (HDFn), suggesting potential selectivity for cancer cells. Indeed, HDFn cells tolerated the highest treatment concentration of 300 µg/mL, whereas a concentration that markedly decreased the survival of HCT116 and A549 cells (Figure 2).

These results are consistent with previous studies demonstrating the cytotoxic potential of CZEO against several cancer cell lines, including MCF-7, HeLa, HCT-116, HepG2, A549, and WiDr (colorectal) cancer cells [28,29,36,71,72,73]. Similarly to our study, Duguluri and Selvakumar observed that four purified compounds from the chloroform extract of C. zizanioides displayed a substantial cytotoxic impact against colorectal cancer cell lines with IC₅₀ values ranging from 68.47 to 94.36 µg/mL. Arafat et al. reported that CZEO reduced the viability of A549 lung cancer cells and HepG2 liver hepatocellular cancer cells with IC₅₀ values of 32.16 and 37.63 µg/mL, respectively [36]. The reason for the discrepancy between our results and those of Arafat et al., who reported greater potency in A549 than in CRC cells, could be explained by geographical and methodological differences. While their C. zizanioides roots were collected in Egypt and extracted by hydrodistillation followed by drying over anhydrous sodium sulfate, we obtained CZEO from Qatar-grown roots using steam distillation with n-hexane phase separation, followed by drying over anhydrous MgSO₄.

The chemical composition of CZEO in our study further reinforces the notion that its biological activity is largely driven by its sesquiterpenoid-rich fraction, in agreement with previous reports [71,73,74,75,76]. Powers et al. highlighted the role of these compounds as functional drivers of vetiver oil activity [73]. Similarly, the sesquiterpenoids we identified by GC-MS analysis could drive CZEO cytotoxic activity, such as β-eudesmol, β-eudesmol, which has been well characterized for anticancer effects, including suppression of proliferation, migration, and angiogenesis in cholangiocarcinoma models (in vitro and in vivo) [77] and other tumor models [75], as well as regulation of apoptosis and ferroptosis via MAPK signaling [78]. Moreover, widdrol, another CZEO metabolite we report in the GC-MS analysis, has been shown to induce cell-cycle arrest and apoptosis in HT-29 colon cancer cells and suppress tumor growth in a xenograft animal model [76].

3.3. ADMET Analysis, Drug-Likeness, and Pharmacokinetics Screening of CZEO Most Abundant Compounds

Evaluation of drug-likeness and ADMET properties is a significant criterion for developing a molecule as a drug [79]. The 10 most abundant CZEO metabolites identified by GC–MS (Figure 3 and Table 1) were evaluated using the SwissADME tool for physicochemical properties (molecular weight [MW], molecular refractivity [MR], H-bond acceptors [HBA], H-bond donors [HBD], topological polar surface area [TPSA]), lipophilicity, solubility, and compliance with Lipinski’s rule of five. SwissADME analysis revealed that all 10 compounds satisfied Lipinski’s criteria, with MW between 150 and 500 g/mol, HBA ≤ 10, HBD ≤ 5, and MR within 40–130 (Table S1). All had LogP values ≤ 5, no more than one Lipinski rules violation, and a predicted oral bioavailability score of 0.55, suggesting good membrane permeability and oral availability (Table S1); the ADMET properties of all 10 compounds are listed in Table S1.

Figure 3. The chemical structure of the 10 most abundant compounds identified by GC-MS in CZEO and reference drug for AKT1 (MK-2206) and STAT-3 (Stattic).

Figure 3. The chemical structure of the 10 most abundant compounds identified by GC-MS in CZEO and reference drug for AKT1 (MK-2206) and STAT-3 (Stattic).

Pharmacokinetics and pharmacodynamics parameters were further predicted using the pkCSM web tool, demonstrating favorable pharmacokinetics (Table S1). All compounds exhibited high intestinal absorption (>90%) and acceptable permeability in CaCo2 cells and skin. The predicted log steady-state volume of distribution (log VDss) values were generally high (0.3–0.5 L/kg), indicating good tissue distribution, although β-Vetivone and Germacrene D-4-ol displayed relatively lower distribution (log VDss < 0.4 L/kg). Total clearance values ranged from 0.99 to 1.32 log ml/min/kg, with Germacrene D-4-ol showing the highest clearance (1.3 log mL/min/kg), suggesting faster elimination (Table S1).

Metabolism predictions indicated that khusimol, vetiselinenol, rosifoliol, and isovalencenal may inhibit CYP2C19, while khusimol (along with reference inhibitors MK-2206 and Stattic) could inhibit CYP1A2 (Table S1). Excretion parameters, including total clearance and Renal OCT2 substrate status, indicated moderate clearance across all compounds. None were predicted to be OCT2 substrates, supporting favorable excretion profiles. Total clearance values of 0.99–1.32 log ml/min/kg for all 10 metabolites indicated relatively consistent elimination rates. Germacrene D-4-ol had the highest clearance (1.3 log ml/min/kg), suggesting faster elimination (Table S1).

Toxicity predictions supported a favorable safety profile (Table S1). None of the compounds exhibited AMES mutagenicity, hepatotoxicity, or inhibition of hERG I/II potassium channels, indicating low genotoxic, hepatotoxic, and cardiotoxic risk. Predicted oral rat acute toxicity (LD₅₀) values ranged from 1.6 to 2.7 log mg/kg, consistent with moderate acute toxicity typical of bioactive natural products.

3.4. Identification of Potential Targets for CZEO Compounds in Lung and Colorectal Cancer

A total of 2180 putative targets were predicted for the 10 most abundant CZEO compounds using the SwissTargetPrediction and SuperPred databases. After removing duplicates, 470 unique targets were retained (Table S2).

For disease-related genes, 26,645 lung cancer targets were retrieved from GeneCards, 4315 from NCBI, and 1407 from MalaCards. Similarly, 9557, 4321, and 1437 CRC-related genes were identified from the same databases. Following filtering, deduplication and merging across all sources, the final consolidated lists comprised 10,804 unique lung cancer targets and 10,674 unique CRC targets (Tables S3 and S4).

Comparison of compound- and disease-related targets using Venn analysis revealed 373 and 394 overlapping genes for lung cancer and CRC, respectively (Figure 4). These intersecting genes represent potential mediators of CZEO’s anticancer activity and were subsequently subjected to protein–protein interaction (PPI) network construction and hub-gene analysis.

3.5. Protein–Protein Interaction (PPI) Network Analysis Reveals Key Hub Genes Linking CZEO Compounds to Cancer-Related Pathways

The overlapping gene sets (373 for lung cancer; 394 for CRC) were submitted to the STRING database to construct PPI networks (Figure 5A and Figure 6A). Visualization in Cytoscape and analysis with the CytoHubba plug-in identified the top 10 hub genes for each cancer type, based on degree, closeness, and betweenness centrality (Figure 5B and Figure 6B).

For lung cancer, the hub genes were AKT1, PIK3R1, STAT3, PIK3CA, MAPK1, HSP90AA1, PIK3CB, PIK3CD, MAPK3, and ESR1 (

Table 2. The top 10 hub genes for CZEO compounds-lung cancer targets, ranked by degree.

Rank	Gene	Score	Family
1	AKT1	31	Protein kinase
2	PIK3R1	27	Protein kinase
2	STAT3	27	Transcription Factor
4	PIK3CA	26	Protein kinase
4	MAPK1	26	Protein kinase
6	HSP90AA1	25	Molecular chaperone
7	PIK3CB	24	Protein kinase
7	PIK3CD	24	Protein kinase
7	MAPK3	24	Protein kinase
10	ESR1	23	Nuclear Receptor

). While for CRC, the hub genes were STAT3, AKT1, HSP90AA1, TNF, HIF1A, ESR1, MAPK3, MAPK1, HSP90AB1, and SIRT1 (

Table 3. The top 10 hub genes for CZEO compounds-CRC targets, ranked by degree.

Rank	Gene	Score	Family
1	STAT3	59	Transcription Factor
2	AKT1	56	Protein kinase
3	HSP90AA1	49	Molecular chaperone
4	TNF	45	Cytokines
5	HIF1A	43	Transcription Factor
6	ESR1	40	Nuclear Receptor
6	MAPK3	40	Protein kinase
8	MAPK1	38	Protein kinase
9	HSP90AB1	37	Molecular chaperone
10	SIRT1	34	Epigenetic regulator

).

Cross-referencing both sets revealed six hub genes common to both cancer types: STAT3, AKT1, HSP90AA1, ESR1, MAPK3, and MAPK1 (

Table 4. The Common hub genes between CZEO-lung cancer and CZEO-CRC targets.

No.	Gene	Target Family
1	STAT3	Transcription Factor
2	AKT1	Protein kinase
3	HSP90AA1	Molecular chaperone
4	ESR1	Nuclear Receptor
5	MAPK3	Protein kinase
6	MAPK1	Protein kinase

). Notably, AKT1 and STAT3 emerged as highly ranked hub genes across both cancers, highlighting their potential central role in mediating CZEO’s anticancer effects.

3.6. GO and KEGG Functional Enrichment Link CZEO to Oncogenic Signaling, Motility, and Therapy Resistance Pathways in Lung and Colorectal Cancer

Gene Ontology (GO) and KEGG analyses were performed using the ShinyGO online tool to explore the biological roles and signaling pathways associated with the hub genes. For lung cancer, we identified 1000 biological processes (BPs), 98 cellular components (CCs), and 155 molecular functions (MFs). Significant enrichment (FDR < 0.05) was observed in BPs such as positive regulation of cell migration, cellular component movement, and locomotion (Figure 7A), suggesting a role in tumor motility, invasion, and metastatic spread [80,81]. For cellular components, PI3K complexes, pseudopodia, caveolae, and membrane-associated sites were enriched (Figure 7B), highlighting the involvement of hub genes in cytoskeletal remodeling and receptor-mediated signaling [82,83]. Among molecular functions, hub genes were enriched in nitric oxide synthase regulation, PI3K activity, insulin receptor substrate binding, kinase binding, and phosphatase binding (Figure 7C), consistent with oncogenic kinase-driven signaling programs [81,82].

For colorectal CRC, GO enrichment of the top 10 hub genes identified 1000 BPs, 125 CCs, and 175 MFs. Significant enrichment (FDR < 0.05) was observed in BPs, including positive regulation of catalytic activity, macromolecule biosynthetic processes, nucleobase metabolic processes, and responses to oxygen-containing compounds (Figure 8A), suggesting a role in metabolic reprogramming and oxidative stress adaptation, both hallmarks of CRC progression [80,82]. For cellular components, hub genes were significantly enriched in the nuclear lumen, nucleoplasm, and mitochondria (Figure 8B), pointing to transcriptional regulation and mitochondrial functions central to CRC biology [82,83]. At the molecular function level, enrichment was observed in enzyme binding, protein–protein interactions, and kinase binding (Figure 8C), reflecting the centrality of kinase- and transcription factor-mediated signaling in tumor growth and survival.

KEGG pathway enrichment analysis of CZEO–lung cancer hub genes (AKT1, PIK3R1, STAT3, PIK3CA, MAPK1, HSP90AA1, PIK3CB, PIK3CD, MAPK3, and ESR1) revealed strong enrichment across several cancer types, myeloid leukemia, pancreatic cancer, and NSCLC. Importantly, eight hub genes (AKT1, STAT3, PIK3CA, PIK3CB, PIK3CD, PIK3R1, MAPK1, and MAPK3) mapped directly to NSCLC-related signaling pathways (Figure 9A), underscoring their functional relevance to lung cancer progression and potential therapeutic value.

These hub genes clustered within the PI3K/AKT, STAT3, and MAPK cascades, while also intersecting with pathways related to EGFR tyrosine kinase inhibitor resistance, prolactin (PRL) signaling, PD-1/PD-L1 immune evasion, and endocrine resistance (Figure 9A). PRL signaling was prominently enriched (Figure 9A), consistent with evidence that PRL binding to the prolactin receptor (PRLR) activates JAK2, STAT3, and ERK1/2 pathways [84]. These findings suggest that CZEO compounds may act by disrupting critical oncogenic signaling cascades that drive tumor growth, invasion, and drug resistance in lung cancer [81,83,85].

AKT1 and STAT3 emerged as central nodes in the enriched pathways (Figure 9A). AKT1 functions as a master regulator within the PI3K/AKT/mTOR pathway, frequently hyperactivated across lung cancer subtypes, where it facilitates tumor initiation, proliferation, and survival [30,32,33]. Likewise, STAT3 is constitutively activated in over 50% of NSCLC patients, with elevated expression associated with poor differentiation, advanced clinical stage, lymph node metastasis, and resistance to therapy [31].

KEGG pathway enrichment analysis of the CZEO–CRC hub genes revealed significant involvement in multiple signaling cascades directly or indirectly linked to CRC, including the prolactin signaling pathway, estrogen signaling, proteoglycans in cancer, chemical carcinogenesis–receptor activation, and lipid and atherosclerosis (Figure 9B). As was observed in lung cancer, the PRL signaling pathway was prominently enriched. Both PRL and its receptor (PRLR) are overexpressed in CRC tissues, sera, and cell lines compared with normal colonic epithelium. Mechanistically, PRL signaling activates the JAK2/STAT3/ERK1/2 axis, which in turn modulates Notch signaling and enhances cancer cell stemness [35], indicating that CZEO compounds may attenuate tumor progression by interfering with PRL–STAT3–Notch crosstalk.

In addition, enrichment of the PD-L1 expression and PD-1 checkpoint pathway indicates that CZEO metabolites may influence CRC immune evasion, a hallmark of disease persistence and progression. Hub genes were also significantly enriched in Th17 cell differentiation and IL-17 signaling pathways, both of which have been implicated in CRC initiation, chronic inflammation, and tumor progression, and are reported to be elevated in CRC tissues and sera compared with healthy controls [86,87]. More importantly, STAT3 could activate these pathways in colonic tissues [88].

Our findings align with previous research demonstrating the evaluation of multiple AKT and STAT3 inhibitors in both preclinical and clinical settings [89,90,91], underscoring the therapeutic significance of these pathways in the treatment of lung and colorectal cancers. In NSCLC, dual inhibition of PI3K/AKT and MET/STAT3 signaling markedly suppressed tumor growth in vivo [92]. Similarly, AKT inhibition has shown promise in CRC: compounds such as MK-2206 have displayed significant antitumor activity in preclinical models and advanced to clinical trials for CRC management [90]. Furthermore, the development and use of STAT3 inhibitors offer a new approach to cancer treatment [31]. For instance, JAK–STAT inhibitors like ruxolitinib significantly reduced cancer cell viability in stromal-rich CRC tumors [93], highlighting the translational potential of targeting this axis.

Given that STAT3 and AKT1 consistently ranked as top hub genes for CZEO in both lung and colorectal cancer networks, showed significant enrichment in KEGG pathways, and are supported by substantial evidence as therapeutic targets, we selected them for molecular docking and molecular dynamics simulations to assess their binding affinities with the most abundant CZEO compounds.

3.7. Molecular Docking Analysis Reveals Strong Binding of CZEO Phytocompounds to AKT1 and STAT3

Molecular docking was performed to evaluate the binding mechanisms of the most abundant phytochemical compounds in CZEO with the common hub targets AKT1 and STAT3.

For AKT1, docking scores ranged from −5.0 to −6.2 kcal/mol (Table S7). Rosifoliol showed the strongest affinity (−6.20 kcal/mol), followed by α-vetivone (−5.93 kcal/mol). By comparison, the AKT1 reference inhibitor MK-2206 displayed a lower docking score (−3.68 kcal/mol) (Table 5 and Table S7). Interaction mapping revealed that the AKT1–rosifoliol complex was stabilized by a hydrogen bond with GLU234, multiple hydrophobic contacts (e.g., ALA230, PHE438, LEU156, MET227, VAL164, LYS179), and van der Waals interactions with residues including TYR229, THR291, and GLU278 (Figure 10A). α-Vetivone engaged mainly hydrophobic residues (LEU202, ALA230, MET227, VAL164) and van der Waals interactions with PHE438, MET281, and GLU198 (Figure 10B). In contrast, MK-2206 formed only two hydrogen bonds (ASN279, THR160) and fewer stabilizing contacts overall (Figure 10C).

Figure 10. Molecular docking interactions between CZEO compounds and AKT1. Protein–ligand interaction diagrams showing hydrogen bonds, hydrophobic (alkyl, π-alkyl), and van der Waals contacts. (A) AKT1–rosifoliol. (B) AKT1–α-vetivone. (C) AKT1-MK-2206 (reference inhibitor). Ligands are shown in stick representation: rosifoliol (teal), α-vetivone (red), and MK-2206 (indigo).

Figure 10. Molecular docking interactions between CZEO compounds and AKT1. Protein–ligand interaction diagrams showing hydrogen bonds, hydrophobic (alkyl, π-alkyl), and van der Waals contacts. (A) AKT1–rosifoliol. (B) AKT1–α-vetivone. (C) AKT1-MK-2206 (reference inhibitor). Ligands are shown in stick representation: rosifoliol (teal), α-vetivone (red), and MK-2206 (indigo).

Table 5. Binding energies and interaction profiles of lead CZEO compounds with the hub protein targets AKT1 and STAT3. Docking scores (kcal/mol), hydrogen bonds, hydrophobic (alkyl, π-alkyl), van der Waals, and other key interactions are summarized alongside reference inhibitors (MK-2206 for AKT1 and Stattic for STAT3).

Target	Compound	Docking Score	Hydrogen Bond/Salt Bridge	Hydrophobic Interaction (Alkyl, Pi-Alkyl)	Van der Waals Interaction	Other Interactions
AKT1	Rosifoliol	−6.20	GLU234	VAL164, ALA230, LYS179, PHE438, MET281, MET227, LEU156, ALA177,	ASP292, THR211, THR291, LY158, GLY157, GLU278, THR229
	α-Vetivone	−5.93	-	LEU202, ALA230, LYS179, VAL164, MET227, ALA177	PHE438, LEU156, MET281, GLU228, THR291ASP292, GLU198, THR211
	MK-2206 (reference)	−3.68	ASN279, THR160	LYS158, VAL164	GLY162, GLU234, LYS163, MET281, PHE438, ALA230, THR211, ALA177, THR291, PHE161, ASP274, GLY157, GLY159, LYS179	LYS276, GLU278 (Electrostatic)
STAT3	Rosifoliol	−5.19	ARG609 GLU612	LYS591, PRO639	GLU594, GLU638, SER638, VAL637, THR620, SER611, SER613, SER614, GLU612
	α-Vetivone	−5.09	SER613, GLU612, SER611	PRO639, LYS591	ARG609, GLN635, THR620, VAL637, GLU638
	Stattic (reference)	−3.56	SER613, GLU612, SER611, ARG609 (Salt bridge)		SER639, THR620, VAL637, SER636	LYS591 (Electrostati)

Table 5. Binding energies and interaction profiles of lead CZEO compounds with the hub protein targets AKT1 and STAT3. Docking scores (kcal/mol), hydrogen bonds, hydrophobic (alkyl, π-alkyl), van der Waals, and other key interactions are summarized alongside reference inhibitors (MK-2206 for AKT1 and Stattic for STAT3).

Target	Compound	Docking Score	Hydrogen Bond/Salt Bridge	Hydrophobic Interaction (Alkyl, Pi-Alkyl)	Van der Waals Interaction	Other Interactions
AKT1	Rosifoliol	−6.20	GLU234	VAL164, ALA230, LYS179, PHE438, MET281, MET227, LEU156, ALA177,	ASP292, THR211, THR291, LY158, GLY157, GLU278, THR229
	α-Vetivone	−5.93	-	LEU202, ALA230, LYS179, VAL164, MET227, ALA177	PHE438, LEU156, MET281, GLU228, THR291ASP292, GLU198, THR211
	MK-2206 (reference)	−3.68	ASN279, THR160	LYS158, VAL164	GLY162, GLU234, LYS163, MET281, PHE438, ALA230, THR211, ALA177, THR291, PHE161, ASP274, GLY157, GLY159, LYS179	LYS276, GLU278 (Electrostatic)
STAT3	Rosifoliol	−5.19	ARG609 GLU612	LYS591, PRO639	GLU594, GLU638, SER638, VAL637, THR620, SER611, SER613, SER614, GLU612
	α-Vetivone	−5.09	SER613, GLU612, SER611	PRO639, LYS591	ARG609, GLN635, THR620, VAL637, GLU638
	Stattic (reference)	−3.56	SER613, GLU612, SER611, ARG609 (Salt bridge)		SER639, THR620, VAL637, SER636	LYS591 (Electrostati)

For STAT3, rosifoliol (docking score −5.19 kcal/mol) and α-vetivone (−5.09 kcal/mol) again achieved the most favorable binding energies among the 10 CZEO compounds tested (Table 5 and Table S7). Rosifoliol formed two hydrogen bonds with ARG609 and GLU612, alkyl contacts with LYS591 and PRO639, and van der Waals interactions with GLU594, SER611, SER613, SER636, THR620, VAL637, and GLU638 (Figure 11A). α-Vetivone bound via three hydrogen bonds with SER611, SER613, and GLU612 within the SH2 domain of STAT3, and additional hydrophobic and van der Waals interactions with LYS591, PRO639, ARG609, GLN635, SER636, VAL637, GLN638, and THR620 (Figure 11B). By comparison, the reference inhibitor Stattic had a weaker docking score (−3.56 kcal/mol), although it engaged similar critical residues (SER611, SER613, GLU612, ARG609) through hydrogen bonds, salt bridges, and van der Waals interactions (Figure 11C).

Importantly, both rosifoliol and α-vetivone recapitulated the binding pattern of Stattic in the phosphotyrosine-binding pocket of STAT3, a region essential for dimerization and transcriptional activation. (Figure 11 and Figure 12), suggesting that CZEO compounds may inhibit STAT3 activity by interfering with STAT3 dimerization and transcriptional activation [94]. Superimposed docking poses confirmed that CZEO compounds aligned well with the binding orientation of reference inhibitors of both AKT1 and STAT3 (Figure 12).

These results support the hypothesis that CZEO phytocompounds can act as natural modulators of oncogenic kinases, with rosifoliol and α-vetivone showing the greatest potential to serve as lead scaffolds for inhibiting AKT1 and STAT3.

In retrospect, we performed an expanded in silico analysis that included all CZEO compounds with a relative abundance greater than 1%, rather than limiting the analysis to the ten most abundant constituents. This broader dataset encompassed additional sesquiterpenoids such as α-valerenol, costol 1c, zizanol, β-vatirenene, and others identified in the GC–MS profile. Incorporating these compounds did not change the major network-pharmacology outcomes. As shown in Supplementary Figures S2–S4 and Supplementary Tables S8–S12, the overlapping target genes, PPI network topologies, and hub-gene rankings remained highly consistent with our original analysis. AKT1 and STAT3 continued to emerge as the top hub genes associated with CZEO activity in both colorectal and lung cancer (Tables S10 and S11).

Moreover, even after docking all compounds with >1% relative abundance to AKT1 and STAT3, α-vetivone and rosifoliol still achieved the highest docking scores (Table S12), reaffirming their status as the dominant predicted bioactive constituents of CZEO. These convergent findings confirm that expanding the chemical dataset does not alter the mechanistic predictions: CZEO is likely to exert anticancer activity primarily through modulation of oncogenic kinase pathways involving AKT1 and STAT3.

This expanded analysis strengthens the robustness of our computational pipeline and supports the selection of these two targets for subsequent molecular dynamics simulations.

3.8. Molecular Dynamics Simulations Confirm Stable Interactions of CZEO Lead Compounds with AKT1 and STAT3

3.8.1. Dynamic Stability Analysis Highlights Conformational Stabilization of AKT1 and STAT3 by Rosifoliol and α-Vetivone

Furthermore, the two lead compounds, rosifoliol and α-vetivone, which showed the most favorable binding affinities with STAT3 and AKT1, were subjected to molecular dynamics (MD) simulations to evaluate the stability of their interactions. RMSD (Root Mean Square Deviation) analysis (Figure 13) provided insights into the conformational stability of AKT1 and STAT3 in complex with rosifoliol and α-vetivone compared to their respective reference inhibitors (MK-2206 for AKT1 and Stattic for STAT3) during a 200 ns molecular dynamics simulation.

For AKT1, the control MK-2206/AKT1 complex exhibited a gradual increase in RMSD from ~1.0 Å to higher values across the simulation, suggesting progressive weakening of complex stability over time (Figure 13a). In contrast, the rosifoliol–AKT1 complex showed a sharp increase in RMSD during the first 50 ns followed by stabilization with only minor fluctuations, indicating an adaptive conformational adjustment before reaching equilibrium (Figure 13b). The α-vetivone–AKT1 complex displayed a more restrained and consistent RMSD profile, stabilizing between 1.5 and 2.0 Å, suggesting that α-vetivone induced moderate conformational changes while preserving the structural integrity of AKT1 (Figure 13c).

For STAT3, the control Stattic/STAT3 complex showed the highest RMSD values (greatest instability), with RMSD values fluctuating widely between 2.0 and 3.5 Å (Figure 13d). By comparison, rosifoliol–STAT3 RMSD values stabilized within 1.5–2.5 Å, exhibiting reduced fluctuations relative to the control, suggesting that rosifoliol binding enhances STAT3 stability, likely by restricting flexible domain motion or inducing a more compact conformation (Figure 13e). The α-vetivone–STAT3 complex showed a similar trend, with RMSD values stabilizing between 2.0 and 2.5 Å after an initial rise and remaining consistent thereafter throughout the simulation (Figure 13f). Although rosifoliol exhibited slightly greater fluctuations than α-vetivone, both ligands provided more stable interaction profiles with STAT3 than the control inhibitor.

Taken together, these RMSD results indicate that both rosifoliol and α-vetivone stabilize AKT1 and STAT3 compared to reference inhibitors, with α-vetivone showing the most consistent stabilizing effect, potentially locking the proteins into a more rigid and functionally constrained conformation.

3.8.2. Residual Fluctuation Analysis of CZEO Compounds-Protein Complexes

The Root Mean Square Fluctuation (RMSF) analysis (Figure 14) provides further insights into the dynamic behavior and flexibility of AKT1 and STAT3 proteins upon interaction with rosifoliol and α-vetivone compared to the reference inhibitors.

For AKT1 (330 residues), the control complex (MK-2206–AKT1; green) displays a baseline pattern of residue flexibility, with moderate fluctuations around residues 150–200 and a prominent and pronounced mobility in the C-terminal region (residues 300–330), consistent with looped or unstructured segments. Rosifoliol (purple) induced higher fluctuations in these same regions, suggesting increased conformational flexibility and potential destabilization of local domains. In contrast, α-vetivone (golden brown) closely followed the baseline pattern of the control-AKT1 complex, with slightly reduced fluctuations at residues 150–200 and the C-terminus, implying that α-vetivone binding stabilizes AKT1 by reinforcing structural rigidity.

For STAT3, both the green line (Stattic–STAT3) and α-vetivone (maroon) complexes displayed similar RMSF trends with peak fluctuations at residues 35–60 and ~80–90, corresponding to flexible loop regions or solvent-exposed regions. Rosifoliol (purple) induced elevated fluctuations in these regions compared with the control, suggesting enhanced conformational flexibility that could destabilize protein–protein interactions. In contrast, α-vetivone consistently dampened fluctuations relative to both Stattic and rosifoliol, particularly in the flexible loop domains, indicating a stabilizing effect on STAT3’s backbone.

Taken together, these results indicate that rosifoliol increases conformational flexibility, while α-vetivone mirrors the baseline control, and further reduces local fluctuations, thereby stabilizing AKT1 and STAT3. This stabilizing effect may enhance ligand–protein binding affinity and functional efficacy, underscoring α-vetivone’s promise as a lead compound.

3.8.3. Compactness Analysis of Ligand-Protein Complexes

The Radius of Gyration (Rg) analysis (Figure 15) was used to assess protein compactness during the 200 ns MD simulations of AKT1 and STAT3 complexes with control inhibitors and the lead CZEO compounds rosifoliol and α-vetivone.

For AKT1, the MK-2206 complex showed a gradual Rg increase from ~20.6 Å to 21.2 Å, indicating reduced compactness (Figure 15a). Rosifoliol binding induced moderate expansion (up to 21.4 Å by 100 ns) before partial stabilization, suggesting ligand-driven structural loosening (Figure 15b). In contrast, the α-vetivone–AKT1 complex was the most stable, with Rg values consistently between 20.6 and 20.9 Å, maintaining compactness throughout the simulation (Figure 15c).

In the STAT3 systems, the control (Stattic) complex exhibited pronounced Rg fluctuations (13.8–15.2 Å), reflecting significant conformational flexibility (Figure 15d). Rosifoliol binding reduced these fluctuations, stabilizing Rg at ~14.2–14.5 Å and promoting a more compact fold (Figure 15e). The α-vetivone–STAT3 complex showed even greater stability, with Rg values of 13.8–14.2 Å and an early downward trend, suggesting enhanced compaction compared to both rosifoliol and Stattic (Figure 15f).

Overall, Rg analysis supports the RMSD and RMSF results, indicating that α-vetivone confers stronger structural stabilization of both AKT1 and STAT3, highlighting its potential as a lead anticancer compound.

3.8.4. Post-Simulation Hydrogen Bonds Analysis of Ligand-Protein Complexes

Hydrogen bond (H-bond) analysis (Figure 16) offered further insight into protein stability during the 200 ns MD simulations. The MK-2206–AKT complex maintained ~140–180 H-bonds consistently, reflecting a stable tertiary structure. Rosifoliol–AKT showed a reduced and more variable profile (~120–170 H-bonds), with more frequent decreases in the number of H-bonds, indicating partial destabilization of AKT’s H-bonding network upon binding of rosifoliol. In contrast, the α-vetivone–AKT complex exhibited a stable pattern similar to the control (~120–170 H-bonds) but with fewer fluctuations, pointing to more structural stabilization than rosifoliol.

For STAT3, the static-STAT3 complex sustained ~35–55 H-bonds, confirming a stable architecture. Both rosifoliol– and α-vetivone–STAT3 complexes averaged ~40 H-bonds, but with different patterns. Rosifoliol induced greater fluctuations in flexible loop regions, suggesting increased conformational flexibility. By contrast, α-vetivone maintained a steadier profile, reinforcing STAT3’s stability by reducing dynamic backbone movement, indicating enhanced stabilization of STAT3.

Together, these results demonstrate that while both phytocompounds engage stabilizing hydrogen bonding networks, α-vetivone provides more consistent stabilization for both AKT and STAT3 compared to rosifoliol, complementing the RMSD, RMSF, and Rg analyses.

3.8.5. Binding Free Energies Calculation

Table 6. Binding free energies of the ligand–protein complexes calculated using the MM/GBSA approach. Values are presented as mean ± SD, where the standard deviation reflects fluctuations across the MD simulation frames used for the MM/GBSA analysis.

Parameter	Control-AKT	Rosifoliol-AKT	α-Vetivone-AKT
ΔEvdw	−35.15 ± 0.18	−33.37 ± 0.22	−53.26 ± 0.32
ΔEele	25.16 ± 1.77	216.25 ± 2.06	61.41 ± 0.92
EGB	−4.73 ± 1.56	−193.33 ± 1.80	−43.60 ± 0.87
ESURF	−4.43 ± 0.01	−3.68 ± 0.01	−5.37 ± 0.02
Delta G Gas	−9.99 ± 1.74	212.88 ± 1.99	8.15 ± 1.02
Delta G Solv	−9.17 ± 1.56	−197.02 ± 1.81	−48.98 ± 0.87
∆G total	−19.16 ± 0.34	−14.13 ± 0.36	−40.83 ± 0.29
Parameter	Control-STAT3	Rosifoliol-STAT3	Vetivone-STAT3
ΔEvdw	−15.87 ± 0.01	−23.13 ± 0.24	−15.97 ± 0.26
ΔEele	−41.72 ± 1.04	−132.22 ± 1.79	−136.81 ± 1.55
EGB	49.57 ± 0.93	132.00 ± 1.54	131.64 ± 1.39
ESURF	−2.46 ± 0.02	−3.38 ± 0.01	−2.59 ± 0.00
Delta G Gas	−57.59 ± 1.04	−155.35 ± 1.73	−152.78 ± 1.50
Delta G Solv	47.10 ± 0.93	128.62 ± 1.55	129.04 ± 1.40
∆G total	−10.48 ± 0.42	−26.72 ± 0.30	−23.73 ± 0.25

summarizes the MM/GBSA binding free energy calculations for rosifoliol, α-vetivone, and control inhibitors against AKT and STAT3, offering a quantitative comparison of ligand affinities. Binding free energies for all ligand–protein complexes were computed using the MM/GBSA approach, based on snapshots extracted from the equilibrated portion of the MD simulation trajectory. A total of 300 evenly spaced frames were used for post-processing. The reported values represent the mean binding free energy across all sampled frames, while the associated errors correspond to the standard deviation (SD), reflecting fluctuations in interaction energies across the sampled MD ensemble and representing the intrinsic sampling variability of the MM/GBSA method.

For AKT1, α-vetivone showed the most favorable total binding free energy (−40.83 kcal/mol), reflecting a strong and stable interaction. This binding was mainly driven by highly favorable van der Waals forces (−53.26 kcal/mol) and non-polar solvation energy (ESURF: −5.37 kcal/mol), which compensated for the moderate electrostatic penalty (+61.41 kcal/mol). Rosifoliol exhibited a weaker affinity (−14.13 kcal/mol), largely due to an extremely unfavorable electrostatic contribution (+216.25 kcal/mol) and gas-phase energy (+212.88 kcal/mol), despite strong polar solvation compensation (EGB: −193.33 kcal/mol). The reference AKT1 inhibitor MK-2206 had a moderate binding energy (−19.16 kcal/mol), positioning α-vetivone as the better AKT1 binder.

For STAT3, rosifoliol displayed the most favorable binding energy (−26.72 kcal/mol), slightly outperforming α-vetivone (−23.73 kcal/mol), with both far exceeding the affinity of the control inhibitor Stattic (−10.48 kcal/mol). These strong affinities arose from highly favorable electrostatic contributions (ΔEele: −132.22 and −136.81 kcal/mol for rosifoliol and α-vetivone, respectively) and van der Waals interactions, only partly offset by solvation penalties.

Overall, MD simulations confirmed that rosifoliol and α-vetivone form stable interactions with AKT1 and STAT3. RMSD, RMSF, and Rg analyses showed that α-vetivone stabilizes both proteins by maintaining compact conformations, while rosifoliol induces moderate flexibility, particularly in STAT3 loop regions. Post-simulation hydrogen bond and MM/GBSA analyses further supported these findings, identifying α-vetivone as the better AKT1 stabilizer and rosifoliol as the preferred STAT3 ligand. The two compounds emerge as promising natural scaffolds for dual targeting of oncogenic kinases in colorectal and lung cancer, underpinning CZEO’s anticancer activity.

4. Conclusions

This study demonstrates the anticancer potential of Chrysopogon zizanioides essential oil (CZEO), defined by a sesquiterpenoid-rich phytochemical profile. GC–MS profiling identified abundant metabolites such as α-vetivone, rosifoliol, khusimol, and β-eudesmol with favorable drug-likeness and pharmacokinetic properties. In vitro assays showed that CZEO reduced the viability of colorectal (HCT-116, IC₅₀ = 62.95 ± 2.19 µg/mL) and lung (A549, IC₅₀ = 167.82 ± 6.51 µg/mL) cancer cells, confirming its cytotoxic activity with greater potency against colorectal cancer. Network pharmacology and protein–protein interaction analyses highlighted AKT1 and STAT3 as central hub targets, while docking and MD simulation analyses confirmed stable interactions, with α-vetivone preferentially stabilizing AKT1 and rosifoliol exhibiting superior affinity for STAT3. These findings provide the first evidence that α-vetivone and rosifoliol act as functional contributors to CZEO’s anticancer activity, reinforcing their value as novel natural scaffolds for CRC and lung cancer therapy.

Despite these promising insights, challenges remain regarding the gap between computational models and clinical applicability. Future studies should integrate high-throughput multi-omics, systems biology, and AI-driven modeling with rigorous in vitro, in vivo, and clinical validation. Collectively, this study positions CZEO as a promising multitarget natural source for the development of anticancer therapeutics.