Network pharmacology, molecular docking, and in vitro study on Aspilia pluriseta against prostate cancer

Abstract

Background

Current prostate cancer treatments are associated with life-threatening side effects, prompting the search for effective and safer alternatives. Aspilia pluriseta Schweinf. ex Engl. has previously shown anticancer activity in lung and liver cancer cell lines. This study investigated its potential for prostate cancer.

Methods

A crude extract of A. pluriseta root was prepared using dichloromethane/methanol (1:1 v/v) and partitioned into hexane, ethyl acetate, and water fractions. The MTT assay was used to assess the antiproliferative activity of the fractions. The active fractions were tested at 6.25–200 µg/ml on human prostate cancer DU-145 cells and non-cancerous Vero E6 cells. Qualitative phytochemical and gas chromatography-mass spectrometry (GC-MS) analyses were conducted to identify chemical compounds. Network pharmacology was employed to predict molecular targets and modes of action of the identified chemical compounds, with subsequent validation through molecular docking and real-time PCR.

Results

Active extracts included crude dichloromethane/methanol, hexane, and ethyl acetate fractions, inhibiting DU-145 cell proliferation with IC₅₀ values of 16.94, 20.06, and 24.14 µg/ml, respectively. Selectivity indices were determined to be 6.04 (crude), 3.62 (hexane), and 6.68 (ethyl acetate). Identified phytochemicals comprised phenols, terpenoids, flavonoids, tannins, sterols, and saponins. GC-MS analysis revealed seventy-nine (79) compounds, with seven (7) meeting ideal drug candidate parameters; their hub gene targets included MAPK3, MAPK1, IL6, TP53, ESR1, PTGS2, MMP9, MDM2, AR, and MAP2K1, implicating regulation of PI3K/Akt, MAPK, and p53 signaling pathways as potential modes of action. Core compounds such as 1-heneicosanol, lanosterol, andrographolide, and retinoic acid exhibited strong binding activities, particularly lanosterol with MAPK21 (-9.7 kcal/mol), ESR1 (-8.9 kcal/mol), and MAPK3 (-8.8 kcal/mol). Treatment with A. pluriseta downregulated AR expression and upregulated p53, while also downregulating CDK1 and BCL-2 and upregulating caspase-3.

Conclusions

A. pluriseta extracts inhibited DU-145 cell growth without causing cellular toxicity, suggesting great potential for development as an anti-prostate cancer agent. However, further in vitro and in vivo experiments are recommended.

Background

Globally, there were an estimated 20 million annual incidences of cancer in 2020, with about 10 million deaths; in more than 60% of countries, cancer is the first or second cause of death [1, 2]. There is a wide range of cancer types. Cancer of the prostate, with 1.41 million annual incidences and 375,000 deaths worldwide in 2020, is the second most prevalent cancer among men after the age of 65 [3, 4]. Prostate cancer forms in the glandular prostate, where it can remain localized or advance by spreading outside the prostate [5]. The former, when detected early, can be effectively treated with surgery and radiotherapy [6]. However, current treatment options for the latter, namely hormonal and chemotherapy (including docetaxel, mitoxantrone, and cabazitaxel), are linked with the reoccurrence of cancer, treatment resistance, and adverse side effects such as damaging growing normal cells, causing fatigue, hypertension, hot flushes, arthralgia, fractures, peripheral oedema, and rashes [2, 7]. Unfortunately, prostate cancer is mainly diagnosed at the advanced stage, when the disease is usually incurable [8]. In Africa, this is particularly attributed to a lack of effective preventive and treatment strategies [9]. It is, therefore, necessary to search for newer chemotherapeutic agents that are more effective, have a lower chance of causing side effects, and are more tolerable than the current ones.

The utilization of medicinal plants for cancer treatment has become increasingly prevalent and accepted as they have lower chances of toxicity and side effects and are cost-effective [10]. Medicinal plants contain various secondary metabolites known as phytochemicals, which inhibit the development and progression of prostate cancer [11, 12]. Aspilia pluriseta Schweinf. ex Engl. belongs to the plant family Asteraceae (commonly called the aster, daisy, composite, or sunflower family). Aspilia is native to Africa and Latin America; A. pluriseta is widely spread in East, Central, and Southern Africa, especially in open woodlands and grasslands [13, 14]. This plant has been used for a long time to treat coughs, stomach infections, burns, bruises, cuts, wounds, pimples, infections of the ears, eyes, and nose, kwashiorkor, fever, worms, and diabetes mellitus [15, 16]. Furthermore, studies have shown that A. pluriseta has antiviral, molluscicidal, antihelmintic, hypoglycemic, antituberculosis, and antimicrobial properties [17,18,19,20,21]. A phytochemical investigation of A. pluriseta revealed the presence of several phytochemicals, including flavonoids and terpenoids; these secondary metabolites suppressed the growth of lung and liver cancer cells [16, 22]. Despite this, information on the anticancer activity of the plant is limited. Hence, in the search for alternative treatments for prostate cancer, the potential clinical applications of A. pluriseta in the management or treatment of prostate cancer were explored using both in vitro and in silico tools.

Network pharmacology is an innovative system-based approach combining network biology and polypharmacology to identify multiple target proteins and signaling pathways of bioactive compounds in plant extracts [23]. The multi-target pathway application of network pharmacology is widely adopted to study the mechanisms of action of traditional medicine [24]. It predicts gene or protein targets of plant chemical compounds and combines them with disease targets to generate a presentable drug-target-disease relationship. On the other hand, molecular docking is an established computational structure-based technique that can be used to predict the nature of interactions between plant chemical compounds and their molecular targets [25]. Network pharmacology and molecular docking are currently used in cancer therapy to develop new drugs; for example, these two techniques have been handy in the development of traditional Chinese medicine [26, 27].

In this study, we asked whether A. pluriseta has an inhibitory effect on the growth of prostate cancer cells. Not only that, but we also evaluated the safety of the plant extract at the cellular level. The antiproliferative activity of A. pluriseta root crude dichloromethane/methanol extract (ApC), hexane (ApH), ethyl acetate (ApE), and water (ApW) fractions against DU-145, a human prostate cancer cell line characterized by moderate metastatic potential, was therefore evaluated. The chemical compounds in the fractions that demonstrated antiproliferative properties were identified through qualitative phytochemical and gas chromatography-mass spectrometry (GC-MS) analyses. Network pharmacology was then explored to identify the putative molecular targets and mechanisms through which the active fractions induced antiproliferation. Molecular docking evaluations were conducted to study the strength and nature of the interactions between the top molecular targets and the core plant chemical compounds. We further validated the network pharmacology results through gene expression studies using real-time polymerase chain reaction (PCR).

Methods

Plant material

The roots were obtained from A. pluriseta plants from Mbeere South sub-county, Embu County, Kenya (0° 46’ 27.0” South, 37° 40’ 54.9” East) on the 26th of March 2022 and transported to the Centre for Traditional Medicine and Drug Research (CTDMR), Kenya Medical Research Institute (KEMRI), Kenya, in sterile bags. Identification and authentication were done by a trained botanist at Egerton University, Kenya, where voucher specimen number NSN2 was deposited and the name checked as acceptable (http://www.worldfloraonline.org/; https://powo.science.kew.org/).

Preparation of crude extract and fractions

The roots of A. pluriseta were washed with distilled water, cut into small pieces, dried in the shade at room temperature for 21 days, and powdered using a grinder (Christy 8 MILL, serial number 51474). The powder (201 g) was percolated exhaustively in a 1 L mixture of dichloromethane and methanol (1:1, v/v) [28]. The extract was filtered through Whatman number 1 filter paper and concentrated at 57 °C using a rotary vacuum evaporator (Rotavapor R-300; Buchi, Switzerland). The concentrated extract was then transferred into a Petri dish and left in the laboratory at room temperature until the dichloromethane and methanol solvents were completely evaporated. The yield of dry extract obtained from 201 g of powder was 38 g, which implies that the percentage yield is 19% w/w dry matter.

20 g of the dried crude dichloromethane and methanol extract was dissolved in 50 ml of a 1:1 solution of dichloromethane and methanol. The mixture was then defatted with 300 ml of hexane in a separating funnel, and the top hexane fraction was decanted. Another 300 ml was added to the remaining lower portion in the separating funnel; this was done two more times to ensure exhaustive collection of the hexane fraction. After that, 300 ml of ethyl acetate and distilled water (1:1) were introduced into the separating funnel; the mixture was shaken vigorously and allowed to stay undisturbed for 1 h. The ethyl acetate fraction settled atop the distilled water part, and the two layers were collected separately in beakers. The collected hexane and ethyl acetate fractions were concentrated using the rotary evaporator, while the water fraction was freeze-dried. The ApC, ApH, ApE, and ApW fractions were stored in a freezer at -20 ˚C until further analysis was needed [29].

Cell lines and culture medium

Human prostate cancer cell line (DU-145) and a non-cancerous mammalian kidney epithelial cell (Vero E6), both purchased from the American Type Culture Collection (ATCC), were cultured in sterile culture T-75 flasks containing Modified Eagle’s Medium (MEM, Sigma-Aldrich, USA) under standard conditions at 37 ˚C and 5% CO₂ at CTDMR-KEMRI, Kenya. The culture medium was supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco, USA), 1.5% sodium bicarbonate (Loba Chemie, India), 1% HEPES (1 M) (GoldBio, USA), and 1% mixture of penicillin and streptomycin (Sigma-Aldrich, USA). The passage number of the DU-145 was 30 (DU145HTB-81/P-30/06/22) while that of the Vero E6 was 14 (VeroE6/P-14/08/22). Both cell lines were passaged two times a week [30].

Antiproliferation and cytotoxicity assay

The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to assess the antiproliferative potential of A. pluriseta fractions. This assay is based on the ability of a mitochondrial enzyme, succinate dehydrogenase, in metabolically active cells to reduce water-soluble, yellow-colored tetrazolium salt MTT to water-insoluble purple formazan crystals. The intensity of formazan purple coloration is equivalent to the number of viable cells [31].

An initial growth inhibitory screening at a single concentration of 200 µg/ml ApC, ApH, ApE, and ApW fractions was carried out on DU-145 cells. The fractions that inhibited DU-145 cell growth by 50% or more after 48 h were selected for further testing at concentrations ranging from 6.25 to 200 µg/ml.

The DU-145 cell was plated at a density of 1 × 10⁴ in 100 µl per well in a 96-well cell culture plate [32]. Cells were incubated at 37 ˚C and 5% CO₂ for 24 h, and thereafter cells received treatment with fresh medium containing 200 µg/ml ApC, ApH, ApE, and ApW fractions for 48 h. Dimethyl sulfoxide (DMSO, 0.2% v/v, Finar Chemicals, India) acted as our negative control, while 200 µg/ml doxorubicin (DXR, Solarbio, China) served as our positive control. After treatment and incubation for 48 h, 10 µl of MTT (Solarbio, China) at 5 mg/ml in phosphate-buffered saline (PBS, Sigma-Aldrich, USA) was added to each well, and the plate was further incubated for 4 hours at 37 °C [32]. The formazan crystals were dissolved in DMSO (100 µl, 100% v/v), and the absorbance was read at 570 nm by a microplate reader (Infinite M1000 by Tecan) [32]. Three wells received the same treatment (technical replicates), and at least three separate experiments were conducted. With 100% being the viability assigned for the negative control cells that had received 0.2% DMSO treatment, growth inhibition was measured as a percentage of viable cells using the following formula:

$$\eqalign{& {\rm{Percentage cell viability}}\;{\rm{ = }}\;{\rm{100}}\; \times \cr & {{{\rm{Absorbance}}\;{\rm{of}}\;{\rm{treated}}\;{\rm{cells}}\;{\rm{ - }}\;{\rm{Absorbance}}\;{\rm{of}}\;{\rm{culture}}\;{\rm{medium}}} \over {{\rm{Absorbance}}\;{\rm{of}}\;{\rm{untreated}}\;{\rm{cells}}\; - \;{\rm{Absorbance}}\;{\rm{of}}\;{\rm{culture}}\;{\rm{medium}}}} \cr}$$

Subsequently, active fractions and DXR at concentrations ranging from 6.25 to 200 µg/ml were used to treat the DU-145 cells as described earlier. The antiproliferative potential was expressed as IC₅₀ (the concentration of the treatments that inhibited cell proliferation by 50%) and presented as the mean ± standard error of the mean (SEM). The IC₅₀ value is a measure of a compound’s ability to inhibit a biological function; hence, the lower the IC₅₀, the better the inhibitory ability of a compound. The MTT assay was also used to test the cytotoxicity of the active A. pluriseta crude extract or fractions (6.25 to 200 g/ml) on Vero E6 cells to see if they were safe for the cells. The concentration of the treatments that inhibited 50% of the Vero E6 cells (CC₅₀) was calculated.

Determination of selectivity index

The selectivity index (SI) is the ratio of the toxic concentration of a therapeutic agent to its effective concentration. The SI for each treatment was calculated by dividing its CC₅₀ by the corresponding IC₅₀. The higher the SI of an herbal drug, the better the indicator that it should be considered for further studies [33].

Phytochemical studies

To identify the bioactive compounds that might have contributed to the antiproliferative activity of A. pluriseta fractions, qualitative tests following the standardized method previously described by [34] were carried out to determine the presence or absence of sterols, phenols, flavonoids, alkaloids, terpenoids, tannins, saponins, and quinones. Gas chromatography-mass spectrometry (GC-MS) analysis was also done to identify the specific phytochemicals in the fractions.

Qualitative phytochemical screening

Sterol test

The Liebermann-Burchard test was conducted to ascertain the presence of sterols. A total of 2 ml of plant sample was mixed with 2 ml of chloroform (Scharlau, Spain), followed by the addition of 10 drops of acetic anhydride and 2 drops of concentrated sulfuric acid (H₂SO₄, HPLC, India). A positive indication was determined by the development of a dark pink coloration.

Phenol test

For the detection of phenols, a ferric chloride test was employed. 2 ml of the plant sample was treated with an aqueous 5% ferric chloride solution (GriffChem, India). The presence of phenols was confirmed by the formation of a deep blue color.

Flavonoid test

To identify flavonoids, the alkaline reagent test was utilized. A mixture comprising 2–3 drops of sodium hydroxide (NaOH, BDH Chemicals Ltd., Poole, England) and 2 ml of plant sample was prepared. A positive outcome was indicated by the development of a deep yellow color, which turned colorless upon the addition of a few drops of dilute (5%) hydrochloric acid (HCl).

Alkaloid test

The presence of alkaloids was determined using Dragendorff’s test. 1 ml of Dragendorff’s reagent was added to 2 ml of the plant sample. A positive result, indicative of alkaloid presence, was denoted by the appearance of a brownish-red color in the mixture.

Terpenoid test

For terpenoid detection, Salkowski’s test was employed. A mixture of 1 ml of chloroform and 2 ml of plant sample was prepared, followed by the addition of a few drops of concentrated H₂SO₄. A positive result was evidenced by the generation of a gray color.

Tannin test

Braymer’s test was conducted to identify tannins. 2 ml of plant sample was treated with a methanolic solution of ferric chloride (10%). The presence of tannins was confirmed by the appearance of a greenish-colored solution.

Saponin test

Saponins were detected using the foam test. 2 ml of plant sample was added to 6 ml of water in a test tube and vigorously shaken. A persistent foam formation indicated a positive result for the presence of saponins.

Quinone test

The presence of quinones was determined through the concentrated HCl test. 2 ml of the fraction was treated with concentrated (98%) HCl. A positive result was indicated by the appearance of a green coloration.

Gas chromatography-mass spectrometry analysis

The equipment used was a Model QP-2010SE from Shimadzu, Tokyo, Japan. The heat was programmed to rise from 55 ˚C to 280 ˚C with a 10 ˚C rise per minute. The injector was maintained at 200 ˚C; at a constant rate of 1.08 ml/min, carrier gas helium was used. The solvent delay was 4 min, and one microliter of source from A. pluriseta was automatically injected using an AS3000 autosampler coupled with GC in split mode, split ratio (10:1). 200 ˚C and 250 ˚C were respectively set for the ion source and interface temperature. Over the range of m/z 35–550, sample components were ionized and collected at 70 eV in full scan mode. For the qualitative identification of compounds found in the sample, the NIST mass spectral database was utilized.

Network pharmacology analysis

Screening of active compounds of A. pluriseta

Network pharmacology was used to identify the molecular targets and signaling pathways in the prostate cancer cells that A. pluriseta bioactive compounds must have acted upon to elicit their antiproliferative activity. To do this, the active compounds in the fractions that were identified through gas chromatography-mass spectrometry (GC-MS) analysis were pulled together, and duplicates were removed, leaving a single compound library of A. pluriseta. The PubChem ID (PCID) and the Canonical SMILES (Simplified Molecular Input Line Entry System) of the compounds were retrieved from PubChem (https://pubchem.ncbi.nlm.nih.gov/). Screening was carried out to identify the valid compounds with respect to their drug-likeness, physicochemical properties, and pharmacokinetics by submitting their SMILES to the Swiss ADME tool (http://www.swissadme.ch/index.php). The tool facilitates drug discovery, computing physicochemical properties, and making predictions on ADME (absorption, distribution, metabolism, and excretion) parameters, pharmacokinetic characteristics, drug-likeness, and medicinal chemistry suitability for one or several compounds [35]. Lipinski’s rule of five (RO5) was used to determine drug-likeness. According to the RO5, there is good absorption or permeation if a compound has a molecular weight (MW) ˂ 500 g/mol, a number of rotatable bonds ˂ 10, a number of hydrogen bond donors ˂ 5, a number of hydrogen bond acceptors ˂ 10, and a lipophilicity (MLOGP) ˂ 4.15 [36]. It is predicted that a violation of more than one of these rules would make the compound a non-orally accessible medication [37]. We considered compounds with no more than one violation of the RO5, MW between 150 and 500 g/mol, those that cannot cross the blood-brain barrier (BBB), those having a topological polar surface area (TPSA) between 20 and 130 Ų [38], and non-inhibitors of cytochrome P450s (CYP2D6 and CYP3A4) valid for network pharmacology analysis; others were excluded.

The TPSA is the sum of the surfaces of polar atoms in a compound, and it can be used to forecast the compound’s transport nature [39]. The BBB penetration and inhibition of CYP2D6 and CYP3A4 were examined for pharmacokinetics. This gives information on drug potential in terms of oral bioavailability or membrane permeability. The permeability of the BBB reveals how a compound moves between the hydrophilic blood and the lipophilic brain; substances that are able to cross the BBB can be toxic to the nervous system. Drug metabolism relies heavily on cytochrome P450 enzymes, especially CYP2D6 and CYP3A4, and a drug candidate should not inhibit the enzymes [39].

Prediction of potential targets of A. pluriseta for prostate cancer cell antiproliferation

The SMILES of the valid active compounds were submitted to BindingDB (https://bindingdb.org/rwd/bind/chemsearch/marvin/FMCT.jsp) and Swiss TargetPrediction (http://www.swisstargetprediction.ch/) databases to predict their targets. “Humans” (Homo sapiens) was set as the study species. Target collection from BindingDB and Swiss TargetPrediction was based on having a minimum similarity of > 0.7 and a probability greater than zero, respectively [40, 41]. The collected targets were combined and de-duplicated, then standardized on the Universal Protein database (https://www.uniprot.org/) and converted into the official gene symbol of Homo sapiens. Prostate cancer-related targets were obtained from the DisGeNET database (https://www.disgenet.org/) and GeneCards database (https://www.genecards.org/) by searching “prostate cancer” as the keyword and removing duplicated genes [42]. Then the overlapping targets of A. pluriseta and prostate cancer, considered potential targets of A. pluriseta in the inhibition of prostate cancer cell growth, were identified using the bioinformatics and evolutionary genomics platform (https://bioinformatics.psb.ugent.be/webtools/Venn/). The overlapped targets-plant components network was visualized using Cytoscape software (version 3.9.1).

Construction of A. pluriseta-prostate cancer targets protein-protein interaction (PPI) network

To further investigate the interaction of A. pluriseta with prostate cancer-related targets, overlapping targets were imported into the STRING (Search Tool for the Retrieval of Interaction Gene/Proteins) 11.5 database (https://string-db.org/), the set condition was human (homo sapiens), and a 0.4 minimum interaction threshold was used to construct the PPI network. The PPI network diagram was sent to Cytoscape, where the Cytohubba plug-in was used to identify the top 10 core target genes in the network based on the Maximal Clique Centrality (MCC) algorithm, from the highest to the lowest MCC [43, 44].

Gene Ontology (GO) and Kyoto Encyclopaedia of genes and genomes (KEGG) enrichment analysis of the A. pluriseta-prostate cancer targets

The gene IDs of the overlapping target genes were submitted to ShinyGO version 0.77 (http://ge-lab.org/go/) to perform gene ontology and enrichment analysis [45]. “Human” was selected as the best-matching species. GO was carried out to discover the biological processes, cellular components, and molecular function that are affected by the target genes, while the KEGG enrichment analysis was to identify the key pathways of the target genes, hence the mechanisms of action of A. pluriseta in prostate cancer cell inhibition. The results of the analysis were presented, with the top 10 from the GO analysis and 20 from the KEGG analysis (False Discovery Rate cut-off of 0.05) selected based on fold enrichment, and they were then displayed as bubble charts with the adjusted p-values and gene counts.

Molecular docking

Based on the network pharmacology analysis results, molecular docking was used to determine the strength and type of interactions between the top 10 core targets of A. pluriseta in prostate cancer intervention with A. pluriseta chemical compounds. The three-dimensional (3D) structure data file (SDF) format of the compounds (ligands) was downloaded from PubChem; the compounds library was prepared in BIOVIA Discovery Studio Visualizer 2021 and saved in PDB (Protein Data Bank) format [41]. The PDB format of the core targets was retrieved from the PDB database (https://www.rcsb.org/) and imported into the BIOVIA Discovery Studio Visualizer 2021 software; available ligands as well as water were removed from the core targets, and polar hydrogen was added. The prepared compound structures and those of target proteins were imported into Autodock Vina in the PyRx 0.8 software for compound-protein docking [46]; they were subsequently converted to the PDBQT (Protein Data Bank, Partial Charge (Q), and Atom Type (T)) files. Compounds’ energy minimization was done with the same software; the grid box was maximized to find the best binding site, and the docking was then run at the default exhaustiveness of eight. When the runs were finished, a CSV (comma-separated values) output file with binding energies generated for each compound was on an Excel sheet, and output data was analyzed to form the complexes in Notepad. This was saved as PDB files before opening the BIOVIA Discovery Studio Visualizer 2021 software for the visualization.

RNA isolation, complementary DNA synthesis, and real-time PCR analysis

The fraction of A. pluriseta with the highest selectivity index (SI) was selected as the representative sample to study gene expression levels. DU-145 cells were seeded in sterile culture T-25 flasks, and when grown to 80%, they were washed with sterile PBS and treated with fresh medium containing the calculated IC₅₀s of the selected fraction and doxorubicin (positive control). The negative, untreated control cells had 0.2% DMSO in free medium. The treated and untreated cells were washed three times with PBS and detached using a 0.25% trypsin-ethylenediaminetetraacetic acid (trypsin-EDTA) solution (Solarbio, China) following 48 h of incubation. The total RNA was extracted from cells by direct lysis using the RNeasy Mini Kit (Qiagen, Germany), and its concentration, purity, and integrity were assessed by nanodrop spectrophotometry (Thermofisher) and 1% agarose gel electrophoresis. The RNA concentration was adjusted to a baseline of 0.5 µg/µl [47], and the RNA was transcribed into complementary DNA (cDNA) by the FIRE Script RT cDNA synthesis kit (Solis BioDyne, Estonia), following the manufacturer’s protocol. Briefly, 1 µl random primer (100 µM), 0.5 µl dNTP mix (20 mM), 2 µl 10x RT reaction buffer with DTT, 1 µl FIRE Script RT, 0.5 µl RNase inhibitor (40 U/µl), an appropriate volume of template RNA (0.5 µg/µl), and nuclease-free water were added to give a final volume of 20 µl reaction mixture. The reverse transcription process was carried out in a SimpliAmp thermal cycler (Applied Biosystems); conditions included primer annealing at 26 °C for 6 min, 42 °C for 60 min for reverse transcription, and 85 °C for 5 min for enzyme inactivation. Then, the cDNA was processed for real-time PCR via Luna Universal qPCR Master Mix (New England Biolabs, USA). Briefly, the final volume of the 20-µl reaction mixture contained 10 µl of Luna Universal qPCR Mix, 0.5 µl forward primer, 0.5 µl reverse primer (10 µM), 2 µl cDNA (1:5 dilutions with nuclease-free water), and 7 µl nuclease-free water. The real-time PCR was carried out in the QuantStudio 5 (Applied Biosystems) qPCR machine with the following cycling program: initial denaturation at 95 °C for 60 s, 1 cycle; denaturation at 95 °C for 15 s, 40 cycles; extension at 60 °C for 30 s, 40 cycles; and melt curve at 60 s, 1 cycle. Gene-specific primers (Macrogen Europe BV, Netherlands) were designed utilizing the National Centre for Biotechnology Information (NCBI) Primer Blast tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast) for each gene of interest listed in Table 1. The fragment sizes ranged from 70 to 250 bps and had a 40–60% guanine cytosine (GC) content and self-complementarity ≤ 2. The results were interpreted using the 2^−ΔΔCt method, and the threshold cycle (Ct) values were normalized to the expression rate of β-actin as a housekeeping gene [48]. All the cDNA and real-time PCR reactions were performed in technical triplicate, and negative controls were included in each experiment.

Table 1 Primers sequences for real-time PCR amplification

Statistical analysis

All experiments were undertaken in triplicates (technical replicates) and repeated at least three times independently. Data is presented as mean ± SEM, and statistical analysis was performed using Prism 8.4.0 (GraphPad, San Diego, CA). A t-test and one-way ANOVA followed by the Tukey HSD post-hoc test were used. In all analyses, p values < 0.05 were considered significant.

Results

MTT cell viability assay in cancerous and normal cell lines

In the preliminary single-dose antiproliferation screening, at 200 µg/ml, the ApC, ApH, and ApE fractions inhibited the growth of DU-145 prostate cancer cells by 50% or more after 48 h, whereas the ApW fraction failed to do so (Fig. 1A). The water fraction was then excluded from the succeeding study, in which a 48-hour incubation of the cells with ApC, ApH, ApE, and DXR (at a concentration gradient of 6.25–200 µg/ml) resulted in a concentration-dependent decrease in the proliferation of DU-145 cells (Fig. 1B–E). Images showing the morphological features and alterations after the treatments are shown in Additional File 1. The ApH fraction, with an IC₅₀ of 16.94 ± 1.35 µg/ml, was the most potent (Table 2). ApC and ApE fractions exhibited antiproliferative effects with IC₅₀ values of 20.06 ± 1.06 µg/ml and 24.14 ± 0.49 µg/ml, respectively, which were significantly lower than that of the positive control, DXR (IC₅₀ of 5.30 ± 0.11 µg/ml, p < 0.05; Table 2).

Fig. 1

Antiproliferation assay of A. pluriseta on DU-145 cells. A Screening of ApC, ApH, ApE, and ApW fractions at a fixed concentration of 200 µg/ml; 200 µg/ml DXR was used as the positive control. Graphs B-E, respectively, depict the antiproliferative activity of ApC, ApH, ApE, and DXR at concentrations ranging from 6.25 to 200 µg/ml. 0.2% DMSO was used as a negative control. The dotted line in graph A indicates the set 50% inhibition mark, while in graphs B-E, it represents the IC₅₀. Results were expressed as means ± SEM (n = 3)

In order to determine the safety of A. pluriseta for use as a potential therapeutic agent for prostate cancer, the cytotoxic effects of the plant samples (ApC, ApH, and ApE) on the non-cancerous Vero E6 cell line were also evaluated. We observed a concentration-response relationship when the cells were incubated with increasing concentrations of samples (6.25–200 µg/ml) for 48 h (Fig. 2A-C). A. pluriseta had low cytotoxicity towards non-cancerous cells, with CC₅₀ of the ApC, ApH, and ApE fractions of 121.7 ± 1.36, 61.33 ± 0.70, and 161.19 ± 4.89 µg/ml, respectively. DXR (Fig. 2D) had a CC₅₀ of 176.10 ± 8.09, which was not significantly different (p = 0.1364) when compared with that of the ApE fraction. However, there was a significant difference when the positive control drug was compared with ApC and ApH (Table 2).

Fig. 2

Safety assessment of A. pluriseta in non-cancerous Vero E6 cells. ApC (A), ApH (B), ApE (C), and DXR (D) at increasing concentrations (6.25–200 µg/ml) were exposed to non-cancerous Vero cells, and cell viability was evaluated by the MTT assay after 48 h. 0.2% DMSO was used as a negative control. The dotted line in the graphs indicates the CC₅₀. Results were expressed as means ± SEM (n = 3)

Selectivity index

The SI of ApC, ApH, ApE fractions, and DXR was determined by dividing their CC₅₀ on normal cells (Fig. 2A-D) by their IC₅₀ on prostate cancer cells (Fig. 1B-E) (Table 2). Of the plant’s three fractions, ApE showed the highest selectivity for prostate cancer cells compared to ApC and ApH. The SI of DXR was 33.23.

Table 2 Summary of IC₅₀, CC₅₀, and SI values of the different plant fractions

Values that bear another superscript different from that of DXR in a column differ significantly from DXR. Analyzed by the post-hoc Tukey test (p < 0.05). Values were expressed as mean ± SEM); all treatments were done in triplicate (n = 3).

Phytochemical studies

Qualitative tests showed that ApC, ApH, and ApE contain six classes of compounds, namely terpenoids, flavonoids, tannins, saponins, sterols, and phenols. Alkaloids and quinones were absent in all the fractions. Specifically, terpenoids and flavonoids were present in all fractions; the same was true of tannins and saponins, except that they were absent in ApH. Sterols were present in the ApH and ApE fractions but absent in ApC, whereas phenols were present in ApC but absent in ApH and ApE. Table 3 summarizes the results of the qualitative phytochemical screening.

Table 3 Quantitative phytochemical screening results of Aspilia pluriseta fractions

The chromatograms depicting the GC-MS spectra of chemical compounds within each fraction are illustrated in Fig. 3. The GC-MS analysis revealed 38 peaks in ApC, 32 in ApH, and 49 in ApE. It is noteworthy that certain compounds exhibited multiple peaks; hence, the actual count of compounds in each fraction does not precisely align with the peak count. Identification of individual compounds was established based on peak area and retention time (Tables 4, 5 and 6). In ApC (Table 4), isosteviol methyl ester and lanosterol were each detected twice, resulting in a total of 36 identified compounds; among these, 17 were exclusive to ApC, while 9 were also found in ApH and 1 in ApE. In ApH (Table 5), andrographolide and androstan-17-one were duplicated, yielding a total of 30 compounds; of these, 10 were unique to the ApH fraction, 9 were shared with ApC, and 2 with ApE. Within the ApE fraction (Table 5), 10,12-tricosadiynoic acid, 1-octadecene, and 4-camphenylbutan-2-one were each observed twice, while 1-heneicosanol appeared three times, resulting in a total of 44 identified compounds; notably, 32 were exclusive to this fraction, with 1 compound overlapping with ApC and 2 with ApH. Remarkably, nine compounds, namely andrographolide, 1 H-naphtho[2,1-b]pyran, cis-verbenol, isosteviol methyl ester, kaur-16-en-18-oic acid, naphthalene, retinoic acid, trans-carveol, and α-campholenal, were detected across all fractions of A. pluriseta.

Fig. 3

GC-MS chromatograms of Aspilia pluriseta fractions. A Chromatogram of detected compounds in ApC. B Chromatogram of detected compounds in ApH. C Chromatogram of detected compounds in ApE

Table 4 Chemical compounds of ApC identified by GC-MS

Table 5 Chemical compounds of ApH identified by GC-MS

Table 6 Chemical compounds of ApE identified by GC-MS

Acquisition of potential A. pluriseta action targets for prostate cancer cell growth inhibition

The GC-MS-identified chemical compounds in the ApC [36], ApH [30], and ApE [44] fractions were 110 in total. After removing duplicates and choosing one of the synonyms, 79 were retained. Among the 79, the Canonical SMILES of α-Campholenal, α-Copaene, and β-pinene were unavailable on the PubChem database at the time of this study, and so 76 compounds were screened through the Swiss ADME tool (accessed on the 29th of March 2023) (Additional File 2). 7 out of the 76 chemical compounds passed the set screening parameters of 0 or 1 violation of Lipinski’s rule of five: molecular weight between 150 and 500 g/mol, topological polar surface area between 20 and 130 Ų, non-inhibitors of cytochrome P450s (CYP2D6 and CYP3A4), and inability to cross the blood-brain barrier. The compounds included 1-heneicosanol, lanosterol, andrographolide, 2-pyrrolidinone, retinoic acid, ethanol, and 10,12-tricosadiynoic acid (Table 7). After filtering out redundant targets, 188 potential targets were predicted for the seven chemical compounds from the SWISS TargetPrediction and BindingDB databases (accessed on the 29th of March 2023) (Additional File 3).

To search for and integrate prostate cancer-related targets, the DisGeNET and GeneCards databases were accessed on the 22nd of February 2023. Redundant targets were filtered out, and 12,889 prostate cancer-related target proteins were obtained to be closely related to prostate cancer disease (Additional File 4). The targets of the 7 chemical compounds of A. pluriseta overlapped with prostate cancer disease, yielding 163 target genes (Fig. 4A); details are provided in Additional File 5. The overlapping targets and their corresponding compounds were imported into the Cytoscape software in order to generate a compound-target network diagram. Among the overlapped 163 target genes, which were considered the potential targets for A. pluriseta in inhibiting prostate cancer cells, 21 were targets of 1-heneicosanol. Lanosterol, andrographolide, 2-pyrrolidinone, retinoic acid, and 10,12-tricosadiynoic acid have 65, 41, 2, 65, and 1 targets, respectively (Additional File 6). There was no target for ethanol. Each active compound corresponded to multiple targets, and each A. pluriseta-prostate cancer target was linked to several compounds. This shows the potential A. pluriseta mechanism of polypharmacology, multi-component, and multi-targeted treatment for prostate cancer.

Table 7 Physiochemical and pharmacokinetic properties of A. pluriseta root compounds selected as bearing putative druglike properties

Protein-protein interaction (PPI) network analysis

Through the STRING database, with confidence ≥ 0.4, a PPI network with 163 nodes and 1180 edges was obtained to show the interconnection between those 163 overlapping targets (Fig. 4B). The average node degree was 14.5 and the enrichment P-value < 1.0e-16. TTL (tubulin tyrosine ligase), CDC42BPA (CDC42 binding protein kinase alpha), and ACVRL1 (activin A receptor-like type 1) were not analyzed in the PPI network as they do not interact with other proteins. The top ten hub genes were selected from the PPI network using the Maximal Clique Centrality (MCC) algorithm and the CytoHubba plugin. They are considered the key molecular targets of A. pluriseta druglike compounds for inhibition of prostate cancer cell growth (Fig. 4C). The hub genes, from the highest MCC score to the least, include mitogen-activated protein kinase 3 (MAPK3), mitogen-activated protein kinase 1 (MAPK1), interleukin-6 (IL6), tumor protein p53 (TP53), estrogen receptor alpha (ESR1), prostaglandin-endoperoxide synthase 2 (PTGS2), matrix metalloproteinase 9 (MMP9), mouse double minute 2 proto-oncogene (MDM2), androgen receptor (AR), and mitogen-activated protein kinase kinase 1 (MAP2K1). The corresponding chemical compounds that target these hub genes are 1-heneicosanol, lanosterol, andrographolide, and retinoic acid. We consider them the core components of A. pluriseta antiproliferative activity against prostate cancer.

Fig. 4

Analysis of potential A. pluriseta action targets for prostate cancer cell growth inhibition. A Venn diagram of potential A. pluriseta action targets for growth inhibition of prostate cancer cells. B PPI network of potential A. pluriseta action targets. C Extracted from (B), the PPI network of the top ten targets of A. pluriseta action targets for prostate cancer cell growth inhibition with the highest MCC values. The larger the node is, the more critical the target is in the network

GO and KEGG analysis

The biological properties of A. pluriseta-prostate cancer targets and the signaling pathways involved were investigated using GO and KEGG analyses. A total of 1000 biological processes, 175 cellular components, and 500 molecular functions were enriched by GO analysis, and the top 10 of each were chosen for presentation after being ranked by corrected p-values (Fig. 5). According to the GO analysis, A. pluriseta-prostate cancer targets may participate in biological processes such as response to oxygen-containing compounds, cellular response to oxygen-containing compounds, and homeostatic processes; influence the cellular components, including the nuclear outer membrane-endoplasmic reticulum membrane network and endoplasmic reticulum subcompartment; and exert molecular functions like nuclear receptor activity, ligand-activated transcription factor activity, and monocarboxylic acid binding. Subsequently, KEGG pathway analysis was carried out to deduce the molecular mechanisms by which A. pluriseta inhibits prostate cancer cell growth. A total of 202 pathways, each with a p-value < 0.05, were identified after uploading 163 anti-prostate cancer targets of A. pluriseta to ShinyGO. The top 20 significant KEGG pathways are presented in Fig. 6. The top three pathways by which anti-prostate cancer targets of A. pluriseta appear to be mainly involved were pathways in cancer, neuroactive ligand-receptor interaction, and microRNAs in cancer. We selected and analyzed the pathway in prostate cancer as the most relevant pathway to the study. The prostate cancer pathway was further visualized in KEGG via Pathview (Fig. 7). A visual pathway map showed that the pathway was made up of signaling pathways including PI3K/Akt, MAPK, and p53 signaling pathways. These pathways appear likely to potentially play roles in the molecular mechanisms of A. pluriseta in inhibiting prostate cancer cell growth.

Fig. 5

GO analysis, in the top 10 significantly enriched terms, the y-axis represents the enriched category, the x-axis represents the number of enrichments and the order of importance was ranked from top to bottom by − Log10 (p value). A Biological process. B Cellular component. C Molecular function

Fig. 6

The top 20 significantly enriched pathways. The y-axis represents the enriched categories, and the x-axis represents the number of enrichments. The order of importance was ranked from top to bottom by − Log10 (p value). Red circles represent the main signaling pathways explored in this study

Fig. 7

Pathway maps of potential targets for KEGG analysis. The red part represents the targets of A. pluriseta

Molecular docking

Molecular docking was used to explore the possible interaction mode of A. pluriseta druglike compounds with the ten hub protein targets (MAPK3, MAPK1, IL6, TP53, ESR1, PTGS2, MMP9, MDM2, AR, and MAP2K1). A binding energy that is less than 0 kcal/mol is an indication that a ligand and its target can combine spontaneously; the interaction is better when the binding energy is less than − 5 kcal/mol, and when the binding energy is less than − 7 kcal/mol, the binding configuration has strong affinity [49]. The hub protein targets were chosen for this study because they were considered key molecular targets of A. pluriseta for inhibition of prostate cancer cell growth. The study adopted the method of [50] by choosing the core compounds that have the ten hub proteins as their targets. These compounds are 1-heneicosanol, lanosterol, andrographolide, and retinoic acid. The binding energies of the chemical compounds and the ten hub proteins are shown in Table 8. All binding energies were less than 0 kcal/mol, with 22 strong binding affinity values of less than − 7 kcal/mol and 13 binding energies that were between − 5 and − 7 kcal/mol. There were multiple hydrophobic interactions and hydrogen bonding between the compounds and target protein residues; the visualized docking modes of the five target-compound pairs with the highest affinity: MAPK21-lanosterol (-9.7 kcal/mol), ESR1-lanosterol (-8.9 kcal/mol), MAPK3-lanosterol (-8.8 kcal/mol), MAPK1-retinoic acid (-8.6 kcal/mol), and MMP9-retinoic acid (-8.5 kcal/mol) are shown in Fig. 8.

Table 8 The binding energy between the essential chemical compounds of A. pluriseta and the top ten hub protein targets

Fig. 8

Molecular docking modes of the five target-compound pairs with the highest affinity. A MAPK21-lanosterol. B ESR1-lanosterol. C MAPK3-lanosterol. D MAPK1-retinoic acid. E MMP9-retinoic acid

Gene expression levels

The ApE fraction was selected as the representative sample for A. pluriseta for gene expression studies. This is because the fraction had the best selectivity index (6.68) compared to the ApC extract (SI = 6.04) and the ApH fraction (SI = 3.62) (Table 2). The mRNA expression of AR, p53, CDK1, BCL-2, and caspase-3 was measured by real-time PCR to validate the network pharmacology-predicted A. pluriseta action targets for prostate cancer cell growth inhibition. AR and p53 were selected as they were among the top ten hub genes as well as being enriched in our study-relevant prostate cancer pathway; CDK1 was selected for its role in the cell cycle, while BCL-2 and caspase-3 were picked to verify the apoptotic role of A. pluriseta (Fig. 7). The threshold cycle (Ct) was calculated, and the relative mRNA expression levels of tested genes were normalized to β-actin. As shown in Fig. 9, treatment of the DU-145 prostate cancer cell line with ApE for 48 h resulted in significant upregulation of caspase-3 and p53. On the other hand, the expression level of BCL-2 decreased; similarly, AR and CDK1 decreased significantly. The relative fold change recorded for AR was 2.9, while CDK1 was 5.9.

Fig. 9

Expression levels of A AR, B BCL-2, C CDK1, D caspase-3, and E p53 in DU-145 cells after treatment with 24.14 µg/ml (IC₅₀) ApE for 48 h. Target genes were normalized to β-actin as a housekeeping control gene (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ns: not significant (p ˃ 0.05) as compared to 0.2% DMSO negative control)

Discussion

Prostate cancer is a major contributor to cancer-related deaths in men, and medicinal plants provide a great landscape in the search for new, effective, and safe treatment strategies for the disease. Through the MTT assay, we found that the crude dichloromethane/methanol, hexane, and ethyl acetate fractions of A. pluriseta lowered the viability of DU-145 cells with half-maximum inhibitory concentrations (IC₅₀s) that were less than the US National Cancer Institute (NCI) established 30 µg/ml criteria for crude extract [51]. This finding is in line with previous work that demonstrated that A. pluriseta diterpenoid had cytotoxic effects in the hepatocellular carcinoma (Hep-G2) cell line [16]. Genetic mutations render cancer cells more vulnerable to plant chemical compounds than non-cancerous cells [52]. The growth of normal control group Vero E6 cells was unaffected by A. pluriseta fractions at the same concentrations as the ones that inhibited the growth of prostate cancer cells. It implies that the plant compounds were selective between normal, non-cancerous cells and cancer cells. Our results on the cellular safety of A. pluriseta root extract and fractions are consistent with earlier studies [16, 31]. The selectivity indices (SI) of the active A. pluriseta fractions are greater than 3. These are considered high indices [33]; hence, the plant is a potential source of compounds that could be exploited in the development of selective antiprostate cancer leads.

Research has demonstrated that flavonoids possess antioxidant properties in normal conditions but exhibit potent pro-oxidant effects in cancer cells. They have been observed to initiate apoptotic pathways and downregulate pro-inflammatory signaling pathways [53, 54]. Terpenoids, on the other hand, exert their anticancer effects at various stages of tumor development and demonstrate the ability to induce autophagy in cancer cells [55, 56]. Additionally, phenols, saponins, sterols, and tannins have been found to arrest the cell cycle, induce apoptosis, regulate angiogenesis, and modulate proliferation and invasiveness [54, 57,58,59]. Altogether, we reckon that these phytochemicals are linked to the antiproliferative effects observed in A. pluriseta samples against DU-145 cells.

Despite containing only flavonoids, terpenoids, and sterols, ApH demonstrated the most potent inhibitory activity. This suggests that the antiproliferative effect may be more dependent on specific chemical compounds than on the number of chemical compound classes present. It is well established that extraction solvents significantly influence the composition of chemical compounds in plant extracts, thereby affecting their biological activity [60]. Gas chromatography-mass spectrometry (GC-MS) analysis enabled the identification of various unique compounds in each solvent (dichloromethane/methanol, hexane, and ethyl acetate). The combination of dichloromethane and methanol facilitated the effective extraction of both polar and non-polar chemical compounds. Hexane, being a non-polar solvent, predominantly extracted non-polar compounds, while polar compounds were expected to be partitioned and concentrated in ethyl acetate [61]. Among the identified compound components of A. pluriseta, only lanosterol had been previously reported in the plant [16, 19].

A particular interest of this study was to identify compounds from A. pluriseta that could be considered for the future development of anti-prostate cancer drugs. Corollary to that, we selected compounds with ideal drug properties from the pool of GC-MS-identified chemical compounds, and by exploring network pharmacology, we found out that 1-heneicosanol, lanosterol, andrographolide, 2-pyrrolidinone, retinoic acid, and 10,12-tricosadiynoic acid target protein genes that could potentially be responsible for playing a relevant role in prostate cancer cell proliferation. The most important ten among the protein genes as ranked by maximal clique centrality were MAPK3, MAPK1, IL6, TP53, ESR1, PTGS2, MMP9, MDM2, AR, and MAP2K1 [62,63,64]. The corresponding chemical compounds targeting the ten protein genes are 1-heneicosanol, lanosterol, andrographolide, and retinoic acid; we consider them to be the core chemical components in the antiproliferative activity of A. pluriseta in prostate cancer cells.

In advanced prostate cancer stages, AR mutations and overexpression contribute to sustained proliferation [65]. TP53 encodes the p53 tumor suppressor protein and acts as a vital mechanism for cellular anticancer defenses. TP53 restricts prostate cancer development through cell cycle progression control, senescence, DNA repair, and induction of apoptosis; mutation of TP53 is clinically used as a biomarker in the prediction of primary hormonal prostate cancer [66]. In high concentration, under normal physiology, MDM2 binds and degrades p53, acting as the negative regulator of the tumor suppressor. Recent findings demonstrated that prostate cancer progression was halted by inhibiting the MDM2-p53 interaction [67]. PTGS2, also referred to as cyclooxygenase 2 (COX-2), catalyzes the conversion of arachidonic acid to prostaglandins. Overexpression of PTGS2 has been reported in prostate cancer, and its inhibition improved the treatment of localized prostate cancer [68]. These reports suggest that A. pluriseta might have inhibited prostate cancer cell proliferation by acting on these molecular gene targets.

The prostate cancer pathway emerged as the most pertinent pathway in our study through KEGG analysis. This pathway encompasses crucial signaling pathways such as the PI3K/Akt, MAPK, and p53 pathways, all of which play pivotal roles in the initiation and progression of prostate cancer [69]. Saponin, found to be present in all fractions of A. pluriseta in this present study, has been reported to have mediated the inhibition of proliferation and metastasis in PC3 and DU-145 prostate cancer cell lines by blocking the PI3K/Akt pathway [70]. Also, chemical compounds from the flavonoids, saponin, terpenoid, tannin, phenol, and sterol classes have been reported to have inhibited cancer cells from proliferating by modulating the MAPK pathway [71]. We speculate that the modification of these signaling pathways is the most likely molecular mechanism of A. pluriseta druglike chemical components in prostate cancer cell antiproliferation. Future experimental studies with isolated compounds are therefore recommended to validate this idea.

Based on molecular docking, it seems likely that the core compounds of A. pluriseta bind to the active site of protein targets to modulate them and their downstream pathways. Lanosterol docked with MAPK21, ESR1, and MAPK3 with the lowest binding energies (ranging from − 9.7 to -8.8 kcal/mol). Retinoic acid also demonstrated strong binding activity with MAPK1 and MMP9. The nature of their interaction with the molecular targets includes conventional hydrogen bonds, hydrophobic interactions, and van der Waals forces. These bonds play important roles in establishing and maintaining the stability of protein-ligand interactions. Hydrogen bonds are important in ligand binding specificity; an increased number of hydrophobic contacts shared between a drug and its target has been said to increase the biological effects of the drug lead [72, 73]. Prior studies isolated lanosterol from A. pluriseta and demonstrated its anticancer activity in hepatocellular carcinoma cells [16]. Similarly, retinoic acid induced apoptosis in DU-145 [74]. Our study presents a novel natural source of retinoic acid, and together with lanosterol, we recommend further studies on target-specific evaluation of the compound for anti-prostate cancer activity.

Next, we evaluated the gene expression levels of AR and p53 to further validate the network pharmacology results. Given that defective PI3K/Akt, MAPK, and p53 signaling pathways promote cell cycle and suppress apoptosis in prostate cancer cell growth and survival [69], we additionally assessed the expression levels of CDK1, caspase-3, and BCL-2 to study the effect of A. pluriseta on apoptosis and cell cycle. In prostate cancer, the loss of apoptotic control such as executioner caspase-3 coupled with excessive cell proliferation are important factors that drive cancer initiation and progression [75]. Elevated expression of CDK1 contributes to the advancement of prostate cancer. Hyperactivation of CDK1 leads to phosphorylation events that activate the AR in the absence of ligands [76, 77]. An increased level of BCL-2 expression is considered a counter-response by prostate epithelial cells to evade cell death, and this subsequently facilitates the in vitro and in vivo progression of prostate cancer cells to androgen independence [78]. We observed an increase in the expression level of p53 and a decrease in AR in A. pluriseta-treated prostate cancer cells compared to untreated negative control cells. Recall that molecular docking studies showed a strong binding energy of less than − 5 kcal/mol between AR or p53 and lanosterol, andrographolide, and retinoic acid. Hence, these gene expression results further validate network pharmacology prediction, indicating that the chemical compounds of A. pluriseta may have inhibited the growth of prostate cancer cells by restoring the molecular integrity and physiology of AR and p53. Furthermore, we also observed a significant upregulation in the expression of caspase-3 and a downregulation of BCL-2 and CDK1, suggesting a potential involvement of the plant in inducing apoptosis and inhibiting the cell cycle. Similar investigations into AR, p53, CDK1, BCL-2, and caspase-3 expression with other plant compounds have yielded comparable results to our study [79,80,81]. Nonetheless, functional studies such as cell cycle analysis and apoptosis assays by flow cytometry are recommended.

While this study is the first comprehensive examination of the anticancer properties of A. pluriseta, several limitations warrant consideration. Firstly, while cancer cells share common features, differences in molecular architecture exist among cancer types. Therefore, the inclusion of additional cancer cell lines could strengthen the evidence for the anticancer activity of A. pluriseta. Secondly, although Vero cells from monkeys are commonly used to assess cellular toxicity due to their similarities with healthy human cells [82], employing healthy human prostate cells would have provided more direct relevance to the study of prostate cancer. Additionally, seasonal variations and environmental factors can influence the chemical composition of plants, suggesting that the reported compounds may not represent the entirety of A. pluriseta’s natural constituents. Furthermore, the exclusive use of gas chromatography-mass spectrometry due to financial constraints may have limited the detection of non-volatile, potentially bioactive compounds in the DU-145 growth inhibition. Therefore, we recommend further phytochemical characterization of ApC, ApH, and ApE fractions using techniques such as liquid chromatography-mass spectrometry, high-performance liquid chromatography, or nuclear magnetic resonance to identify additional chemical constituents.

Conclusions

The present investigation demonstrates the selective inhibitory effect of A. pluriseta on the proliferation of the DU-145 prostate cancer cell line. Network pharmacology analysis identified 1-heneicosanol, lanosterol, andrographolide, 2-pyrrolidinone, retinoic acid, and 10,12-tricosadiynoic acid as the druglike compounds within A. pluriseta. These compounds were predicted to target key proteins such as MAPK3, MAPK1, IL6, TP53, ESR1, PTGS2, MMP9, MDM2, AR, and MAP2K and modulate pathways such as PI3K/Akt, MAPK, and p53 signaling to inhibit prostate cancer cell proliferation. Molecular docking studies indicated strong binding affinities between these compounds and their predicted targets, with notable interactions observed, particularly lanosterol with MAPK21, ESR1, and MAPK1, and retinoic acid with MAPK1 and MMP9. Gene expression analysis corroborated these findings, revealing decreased AR expression and increased p53 expression in A. pluriseta-treated prostate cancer cells, along with decreased CDK1 and BCL-2 expression and increased caspase-3 activity, suggesting potential roles in cell cycle regulation and apoptosis induction. Our comprehensive investigation sheds light on the molecular targets and mechanisms underlying A. pluriseta’s antiproliferative effects in prostate cancer cells, positioning it as a promising source of phytocompounds for prostate cancer prevention and treatment. However, further validation through in vitro cell cycle and apoptosis assays, as well as in vivo studies, is warranted to fully elucidate its therapeutic potential.