Bioactivity-guided fractionation of Withania somnifera (L.) Dunal roots extract: evaluation of the anti-progressive potential on human Kelly neuroblastoma cell line

Abstract

Background

Withania somnifera (L.) Dunal, commonly known as ashwagandha, is a cornerstone of Ayurvedic medicine and has demonstrated anti-metastatic properties, including the ability to mitigate the cytotoxic effects of carcinogens and chemotherapeutic agents. Neuroblastoma (NB), a highly aggressive paediatric cancer, accounts for approximately 15% of childhood cancer-related deaths. Despite intensive treatment, over 50% of NB cases experience tumor recurrence and debilitating long-term effects. This study aimed to evaluate the anti-progression effects of W. somnifera root fractions on the human NB Kelly cell line at sub-cytotoxic concentrations and to identify the active bioactive constituents.

Methods

W. somnifera roots were extracted using 95% ethanol and subsequently fractionated via vacuum liquid chromatography with a methanol-water gradient elution, yielding twelve fractions. Kelly cells were treated with each fraction at sub-cytotoxic concentrations, as determined by MTT assay. Treated cells were then subjected to transwell extracellular matrix invasion and fibronectin adhesion assays. Data were analysed using one-way ANOVA (GraphPad Prism), with statistical significance set at P ≤ 0.05. Bioactive fractions were further subfractionated by preparative HPLC, and major constituents were tentatively identified using GC-MS.

Results

Fraction 9 (eluted with 70% methanol) exhibited the highest anti-invasive activity, whereas Fraction 10 (eluted with 80% methanol) demonstrated the most potent and statistically significant (P = 0.0409) anti-adhesive effect compared to vehicle-treated cells (0.5% DMSO). Subfraction analysis revealed that Subfraction 10/1 had a significant anti-adhesive effect (P = 0.0482), while subfractions 10/3 and 9/2 showed non-significant anti-adhesive effects. GC-MS analysis of subfractions 9/2, 10/1, and 10/3 revealed the presence of four previously unreported compounds in W. somnifera.

Conclusions

Constituents of W. somnifera roots exhibit promising anti-metastatic activity against neuroblastoma cells, highlighting their potential to complement existing chemotherapeutic regimens and reduce associated long-term side effects.

Clinical trial number

Not applicable.

Background

Neuroblastoma (NB) is the most common extracranial solid tumor in children [1]. It accounts for approximately 10% of all malignant childhood tumors and is responsible for up to 15% of childhood cancer-related deaths [2]. In 2021, the global incidence rate of NB was 0.28 per 100,000 among children aged 0–14 years, with a higher incidence and mortality observed in males compared to females (ratios of 1.96:1 and 2.02:1, respectively) [3]. The World Health Organization has recently highlighted NB as one of the most prevalent childhood cancers, with approximately 10% of cases thought to be attributable to genetic predisposition [4]. While the majority of NB cases are sporadic, familial forms represent 1–2% of all cases [5].

NB is believed to originate from the sympathoadrenal lineage of the neural crest, based on its primary anatomical sites and its cellular and neurochemical characteristics [6]. It is a highly heterogeneous disease, displaying considerable variability in both biological and clinical characteristics. Presentations can range from localized tumors that spontaneously regress, intermediate-risk tumors that are amenable to surgical resection, with or without post-operative chemotherapy, to high-risk disease with widespread dissemination and poor prognosis [7]. At diagnosis, more than 50% of NB cases are already metastatic [8].

The International Neuroblastoma Risk Group (INRG) classification system categorizes patients into low-, intermediate-, and high-risk groups. While low- and intermediate-risk patients generally have excellent outcomes, the prognosis for high-risk NB remains poor despite intensive treatment, with a long-term survival rate of approximately 50% [9, 10]. Consequently, there is an urgent need for novel therapeutic strategies that not only improve survival but also reduce the long-term toxic side effects associated with current treatments [1].

Withania somnifera (L.) Dunal, commonly known as ashwagandha, Indian ginseng, or Indian winter cherry, is a perennial shrub belonging to the Solanaceae family. Its roots have been used in Ayurvedic medicine for over 3,000 years [11]. Traditionally, extracts of W. somnifera have been employed as adaptogens and tonics, believed to relieve stress, rejuvenate the body, and promote longevity and vitality [12, 13]. Various parts of the plant, including leaves, roots, fruits, and stems, are used in folk medicine to treat conditions such as cardiovascular disorders, pain, liver disease, fever, respiratory infections, wounds, and ulcers. Additionally, W. somnifera is still used by traditional healers to manage arthritis, eye disorders, hemorrhoids, and certain cancers [14].

Pharmacologically, W. somnifera has been reported to be beneficial in neurological conditions such as epilepsy, Alzheimer’s disease, and Parkinson’s disease. Beyond its neurological effects, it also exhibits anti-diabetic, anti-inflammatory, antimicrobial, and anti-neoplastic properties [15]. These therapeutic activities are mediated through multiple molecular targets [11].

The aim of the present study is to investigate the potential anti-progression effects of W. somnifera root constituents on the human NB Kelly cell line and to identify the potential bioactive compounds responsible for these effects.

The significance of this study lies in the potential discovery of new anti-metastatic agents derived from W. somnifera roots. These agents may complement existing chemotherapeutic regimens, potentially limiting the spread of aggressive NB cells while reducing the late-stage toxicities associated with conventional chemotherapy.

Materials and methods

Materials

The dried, coarsely powdered roots of Withania somnifera were purchased from Starwest Botanicals (Sacramento, CA, USA). The human Kelly neuroblastoma cell line was obtained from the Health Protection Agency Culture Collections (HPACC, Salisbury, UK). Unless otherwise specified, all chemicals, reagents, and consumables were purchased from Sigma-Aldrich (Burlington, MA, USA). Cell culture plates were sourced from Corning Inc. (One Riverfront Plaza, Corning, NY, USA).

Extraction and fractionation of W. somnifera roots

Approximately one kilogram of coarsely powdered Withania somnifera roots was divided into two batches. Each batch was percolated with 95% ethanol (4 L 5 times). The ethanol extracts (totalling 40 L) were concentrated, separately, under reduced pressure. The resulting syrupy residues were lyophilized to obtain a brownish powder, which was stored at 4 °C for subsequent processing.

A portion of the dried extract (20 g) was mixed with 40 g of Celite and subjected to vacuum liquid chromatography on a C18 silica gel column (250 g, 4.5 cm 37 cm). Elution was performed sequentially with 100% water, followed by gradual increasing increments of 10% methanol in water, up to 100% methanol. Fractions were collected at 750 mL per solvent composition, yielding a total of twelve fractions of varying polarities.

This fractionation process was repeated using an additional 17.59 gm of the original extract, producing another twelve fractions. All fractions were evaporated to dryness, weighed, and stored in airtight containers at 4 °C for further processing.

Sub-fractionation, purification, and identification of compounds present in the bioactive fractions

Sub-fractionation of the bioactive fractions

To optimize the isolation and purification conditions for compounds present in the bioactive fractions, thin-layer chromatography (TLC) was initially performed using precoated normal phase silica gel F₂₅₄ glass plates. The plates were developed in 5% methanol in chloroform and visualized under UV light at wavelengths of 254 nm and 318 nm. Subsequent visualization was performed by spraying with p-anisaldehyde/sulfuric acid reagent followed by heating.

For further analysis and method validation, the bioactive fractions were analyzed using a Waters^® Acquity UPLC system equipped with a quaternary solvent manager (H-Class), sample manager, and photodiode array detector. Chromatographic separation was achieved on a Waters^® Acquity UPLC BEH C4 column (100 mm 2.1 mm, 1.7 μm) using isocratic elution with 30% acetonitrile in water at a flow rate of 0.5 mL/min. Data acquisition and processing were performed using Empower^® software.

Further fractionation of the bioactive samples was conducted using a Waters preparative HPLC system equipped with a 2998 photodiode array detector. Separation was carried out on a Waters XBridge™ Prep C18-column (10 150 mm, 5 μm) with 40% acetonitrile in water as the mobile phase at a flow rate of 2 mL/min. Detection was performed at λ_max = 254 nm. For each preparative injection, 100 µL of sample was loaded, with 20 injections performed per fraction. Peaks with identical retention times were pooled and evaporated to dryness to yield subfractions, which were stored at 4 °C for subsequent analysis.

Tentative identification of some constituents in W. somnifera roots

To tentatively identify the constituents present in the most bioactive subfractions, GC-MS analysis was performed. A volume of 1 µL from each bioactive subfraction (9/2, 10/1, and 10/3), dissolved in methanol, was analysed using a Thermo high-resolution gas chromatography-mass spectrometry double focusing sector system (GC-MS DFS). The system was equipped with a DB-5MS capillary column (30 m length 0.25 mm inner diameter, 0.25 μm film thickness), with helium as the carrier gas at a flow rate of 0.8 mL/min.

The chromatographic conditions were as follows: splitless injection mode with an injector temperature of 250 °C, detector temperature of 280 °C, and a temperature gradient from 40 °C to 230 °C at a rate of 5 °C/min, followed by a 10-min hold at the final temperature. The mass spectrometer was operated in electron impact ionization mode at 70 eV, with an ion source temperature of 175 °C and a scan range of m/z 40–900 Da.

Compound identification was performed by comparing the acquired mass spectra with entries in the National Institute of Standards and Technology (NIST) MS Search 2.0 library, and comparison with previously reported literature data [16].

Growth and maintenance of Kelly cell line culture and cell count

Kelly cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM)/Ham’s F12 Nutrient Mixture, supplemented with 10% foetal calf serum (FCS) and 1% L-glutamine. Cells were routinely maintained, subcultured, and counted as described in a previous study [17]. Cell counts were expressed as the mean number of cells 10⁴ cells/mL. For each experiment, the required seeding density was calculated by dividing the desired number of cells by the actual cell number, and then multiplying by the total required volume. The resulting cell suspension volume was subsequently diluted with culture medium to achieve the final required volume.

Determination of the sub-cytotoxic concentration of W. somnifera extract, fractions and subfractions on Kelly cells

Kelly cells were incubated with different concentrations of W. somnifera extract, fractions and subfractions to determine their sub-cytotoxic concentrations. This was to investigate the anti-progressive effects of W. somnifera on the viable treated cells.

Treatment of Kelly cells

Stock solutions of the Withania somnifera extract, fractions, and subfractions were prepared by dissolving each in DMSO. Working solutions were subsequently prepared from these stock solutions to ensure a final DMSO concentration of 0.5% in the culture medium.

To determine the sub-cytotoxic concentrations of W. somnifera on the Kelly NB cell line, cells were treated with the extract at concentrations ranging from 500 to 5 µg/mL, with the fractions at 250 to 0.5 µg/mL, and the subfractions at either 10 to 0.5 µg/mL or 0.25 to 0.01 µg/mL.

MTT assay (cell viability)

The cytotoxic effects of Withania somnifera extract, fractions, and subfractions on Kelly cell viability were assessed using the known 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay [18]. Kelly cells (200 µL of 1 10⁵ cells/mL) were seeded in 96-well flat-bottom microtiter plates and incubated overnight at 37 °C. One column of wells containing only culture medium (no cells) was included as a blank.

The following day, media were aspirated and replaced with one of the following: fresh medium (untreated control), medium containing 0.5% DMSO (vehicle control), or medium containing varying concentrations of W. somnifera extract, fractions, or subfractions (dissolved in the DMSO). Treated cells, along with controls and blank wells, were incubated for 96 h at 37 °C.

Following treatment, the MTT assay was performed as previously described [17]. Absorbance of the resulting formazan product was measured at 540 nm using a SpectraMax^® iD3 microplate reader (Molecular Devices, CA, USA). Cell viability was calculated by subtracting the mean absorbance of the blank from all experimental values. The corrected absorbance values of treated cells were then expressed as a percentage of the vehicle control using the following equation:

Graphs were established by plotting the mean percentages of cell viability ± SD versus concentrations.

Anti-progressive effect of W. somnifera

Kelly cells were incubated with medium containing the predetermined sub-cytotoxic concentrations of Withania somnifera extract, fractions, and subfractions, as established by the MTT assay. The following assays were then performed to evaluate the potential effects of the plant constituents on the progression of Kelly cells.

Cell invasion: transwell assay

The invasion assay is a standard in vitro method used to quantify the ability of cancer cells to invade through a basement membrane extracellular matrix (ECM), simulating metastatic behaviour. In this study, the invasive potential of Withania somnifera-treated Kelly cells was assessed using the Cultrex^® 24-well Basement Membrane Extract (BME) Cell Invasion Assay (R&D Systems, MN, USA) [19]. Invading cells were quantified using calcein fluorescence.

To validate the linearity of calcein fluorescence in relation to cell density, Kelly cells were seeded at densities ranging from 1 10⁴ to 5 10⁵ cells/mL in a 96-well plate and incubated overnight at 37 °C. One lane containing only medium was used as a blank. The following day, the medium was aspirated, and 200 µL of Cultrex Cell Dissociation/Calcein-AM solution was added to each well. After incubation for one hour at 37 °C, fluorescence was measured using a SpectraMax^® iD3 plate reader at excitation/emission wavelengths of 485/520 nm. Fluorescence values were background-corrected by subtracting the mean blank signal. The relationship between cell density and fluorescence intensity was analysed using non-linear regression (curve fit) in GraphPad Prism.

To evaluate the anti-invasive effect of W. somnifera, Kelly cells were seeded at a density of 2.5 10⁵ cells/mL in 6-well plates and incubated overnight. The next day, cells were treated with either medium (control), 0.5% DMSO (vehicle control), or sub-cytotoxic concentrations of W. somnifera extract, or fractions, and incubated for 72 h at 37 °C. During the final 24 h of the 96-hour total treatment period, cells were maintained in serum-free medium to starve them from serum-containing chemoattractants.

Simultaneously, 24-well invasion assay inserts (8 μm pore size, PET membrane) were coated with 100 µL of BME and incubated at 37 °C. These inserts served as the upper chambers, while the corresponding wells served as the lower chambers.

After the 96-hour treatment, cells were harvested and resuspended in serum-free medium containing their respective treatments. A volume of 100 µL of each suspension was added to the upper chambers. A blank insert received 100 µL of serum-free medium only. The lower chambers were filled with 500 µL of complete medium containing serum (chemoattractants). Cells were allowed to invade through the BME-coated membrane toward the chemoattractant for 24 h at 37 °C.

Post-invasion, the contents of the upper chambers were aspirated, and the inserts were gently washed with Cultrex buffer. Invading cells on the underside of the membrane were detached by placing the inserts into wells containing 500 µL of Cultrex Cell Dissociation/Calcein-AM solution. Calcein fluorescence, indicative of the number of invading cells, was measured using the SpectraMax^® iD3 plate reader.

Percent invasion in vehicle-treated cells was calculated relative to the untreated control, while invasion in W. somnifera-treated groups was expressed relative to vehicle-treated cells. Data are presented as mean ± SD from two independent experiments.

Cell adhesion: fibronectin assay

The adhesion assay is a widely used in vitro method to assess the ability of cells to adhere to specific ECM components. In this study, a previously established adhesion assay [17] was used to evaluate the adhesion potential of Kelly cells to fibronectin, a major ECM glycoprotein. Adhesion experiments were conducted in 24-well plates pre-coated with fibronectin (Corning, One Riverfront Plaza, NY, USA).

To validate the linearity of Kelly cell adhesion to fibronectin, 1 mL aliquots of cell suspensions at densities ranging from 0.625 10⁵ to 10 10⁵ cells/mL were added in duplicate to fibronectin-coated wells and incubated for 40 min at 37 °C. A well containing only medium served as a blank. Following incubation, non-adherent cells were removed, and adherent cells were subjected to a crystal violet-based adhesion assay. Absorbance of cell lysates and blanks was measured at 595 nm using a SpectraMax^® iD3 plate reader. Background-corrected absorbance values were obtained by subtracting the mean blank absorbance from each sample. Adhesion linearity was evaluated by plotting mean ± SD absorbance versus cell density and analyzed via non-linear regression (curve fit) using GraphPad Prism.

To assess the effects of Withania somnifera on cell adhesion, Kelly cells were seeded at a density of 1 10⁵ cells/mL in 6-well plates (2 mL per well) and incubated overnight. The following day, cells were treated with either medium (control), 0.5% DMSO (vehicle control), or sub-cytotoxic concentrations of W. somnifera extract, fractions, or subfractions. After 96 h of incubation at 37 °C, cells were harvested and resuspended in a final volume of 9 mL of medium. From this, 1 mL aliquots were seeded into fibronectin-coated wells and subjected to the adhesion assay as described above.

The percentage adhesion of vehicle-treated cells was calculated relative to the untreated control, while the adhesion of W. somnifera-treated cells was expressed relative to the vehicle-treated group. Data are presented as mean ± SD from three independent experiments.

Statistical analysis

Data were analysed by the one-way analysis of variants (ANOVA) followed by Dunnett’s multiple comparison test, or by Student’s t-test, using the GraphPad prism software package (version 9). P values of ≤ 0.05 were considered significant.

Results

Extraction and fractionation of W. somnifera

Two independent batches of approximately 500 g each of coarsely powdered Withania somnifera roots were extracted with 95% ethanol. The dried brownish extracts obtained weighed 26.6 g for batch I and 25.3 g for batch II. Portions of these extracts, 20 g from batch I and 17.59 g from batch II, were subjected to fractionation using vacuum liquid chromatography on a C18 silica gel column, employing gradient elution with increasing concentrations of methanol in water. This process yielded twelve fractions of varying polarities from each batch (Table 1). Notably, two fractions (F6 and F7) were eluted using 50% methanol in water, with fraction F7 selected for further investigation. The crude extracts and resulting fractions were subsequently evaluated for their potential anti-proliferative effects on the Kelly NB cell line.

Table 1.

Fractions obtained from W. somnifera extracts

Fraction	Eluent	Weight (gm)
F1	100% water	14.942
F2	10% methanol / water	0.241
F3	20% methanol / water	0.095
F4	30% methanol / water	0.197
F5	40% methanol / water	0.298
F6	50% methanol / water	0.21
*F7	50% methanol / water	0.112
F8	60% methanol / water	0.161
F9	70% methanol / water	0.397
F10	80% methanol / water	0.243
F11	90% methanol / water	0.139
F12	100% methanol	0.682

* F6 and F7 were eluted with the same solvent system, however F7 was used in the present studies

Determination of the sub-cytotoxic concentrations of W. somnifera extract and fractions on Kelly cells

To confirm the non-toxic effect of the vehicle DMSO at the concentration used, Kelly cells were concurrently incubated with 0.5% DMSO. As shown in Fig. 1, this concentration had no significant impact on cell viability compared to untreated control cells.

Fig. 1.

Percentage viability of Kelly cells treated with the vehicle. Cells were incubated in medium alone (control) or medium containing 0.5% DMSO (vehicle) for 96 h in 96-well plates at 37 °C. Following incubation, cell viability was assessed using the MTT assay, and absorbance of the resulting formazan product was measured at 450 nm with a plate reader. Viability values for vehicle-treated cells were calculated as percentages relative to the untreated control cells. Bars represent mean ± SD (n = 4). Data were analysed using an unpaired t-test (GraphPad Prism). ns, not significant

Incubation of Kelly cells with the plant extract at concentrations of 500, 250, 100, 75, and 50 µg/mL resulted in marked cytotoxicity, significantly reducing cell viability compared to vehicle-treated controls (Fig. 2A). In contrast, treatment with 25 µg/mL of the extract did not cause a significant decrease in viability. Given the slight increase in viability observed at 10 µg/mL (109.84% ± 21.23; indicated by an asterisk in Fig. 2A), this concentration of W. somnifera extract was selected for further experiments assessing its potential anti-progressive effects on Kelly cells.

Fig. 2.

Percentage viability of Kelly cells treated with W. somnifera. Cells were incubated for 96 h in 96-well plates with medium containing the vehicle (vehicle control) or vehicle supplemented with various concentrations of W. somnifera extract (A) or fractions 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, and 12 (B to L). Cell viability was assessed by MTT assay, and absorbance of the resulting colour was measured at 450 nm with a plate reader. Viability values for treated cells were expressed as percentages relative to vehicle-treated controls. Bars represent mean ± SD (n = 4). Data were analysed by one-way ANOVA followed by Dunnett’s test (GraphPad Prism). P values ≤ 0.05 were considered statistically significant; ns indicates not significant. Asterisks below the x-axes denote the determined sub-cytotoxic concentrations

Figure 2B-L show the effects of W. somnifera fractions (F1 - F12, skipping F6) on Kelly cell viability across a range of concentrations (250, 100, 75, 50, 25, 10, 7.5, 5, 2.5, 1, 0.75, and 0.5 µg/mL). As depicted, the fractions exhibited variable cytotoxicity at these concentrations. Sub-cytotoxic concentrations, indicated by asterisks in Fig. 2B-L, are summarized in Table 2. These sub-cytotoxic doses of W. somnifera fractions were subsequently employed to evaluate their potential anti-progressive effects on Kelly cells, in a manner analogous to the crude extract.

Table 2.

The sub-cytotoxic concentrations of W. somnifera fractions

Fraction	Sub-cytotoxic concentration
F1	50 µg/mL
F2	25 µg/mL
F3	25 µg/mL
F4	25 µg/mL
F5	0.75 µg/mL
F7	7.5 µg/mL
F8	7.5 µg/mL
F9	1 µg/mL
F10	5 µg/mL
F11	10 µg/mL
F12	50 µg/mL

The anti-invasive effect of W. somnifera on Kelly cells

To evaluate the anti-invasive potential of W. somnifera on Kelly cells, the Cultrex 24-well BME cell invasion assay was employed. In this assay, invading cells are quantified based on the intracellular conversion of calcein-AM to fluorescent calcein by cellular esterases. Prior to assessing plant-treated cells, calcein fluorescence was validated across a range of Kelly cell numbers to confirm assay linearity. Figure 3 illustrates the linear correlation between calcein fluorescence intensity and Kelly cell number.

Fig. 3.

Linearity of liberated calcein fluorescence versus Kelly cell number. Kelly cells were seeded overnight in 96-well plates at 37 °C. Cells were subsequently incubated for one hour with Cultrex cell dissociation buffer containing calcein-AM at 37 °C. Liberated calcein fluorescence was measured using a plate reader at an excitation wavelength of 485 nm and an emission wavelength of 520 nm. Data points represent mean ± SD (n = 4). The coefficient of determination (R²) was calculated using nonlinear regression analysis (GraphPad Prism)

To assess the effect of the vehicle DMSO on cell invasion, Kelly cells were incubated with the vehicle alone. As shown in Fig. 4A, DMSO treatment resulted in an increased invasion of Kelly cells (234.27% ± 101.58) compared to untreated control cells; however, this increase was not statistically significant. Despite the high variability observed (reflected by the large standard deviation), the invasion levels in vehicle-treated cells consistently exceeded those of cells incubated in medium alone. In contrast, treatment with the sub-cytotoxic concentration of W. somnifera extract (10 µg/mL) remarkably reduced cell invasion to 65.35% ± 3.46 relative to vehicle-treated cells (Fig. 4B).

Fig. 4.

Percentage invasion of Kelly cells treated with W. somnifera. Kelly cells were seeded at a density of 2.5 10⁵ cells/mL in 6-well plates. After 24 h, cells were incubated for 72 h with either medium alone (control), medium with vehicle (A), or vehicle containing sub-cytotoxic concentrations of W. somnifera extract or fractions (F1-F12, skipping F6) (B). Following this, cells were re-incubated for 24 h in the respective treatments without serum. Cells were then harvested, resuspended in the corresponding serum-free treatment solutions, and subjected to the invasion assay for 24 h at 37°C. Invasion was quantified by measuring calcein fluorescence released from invading cells using a plate reader. Invasion values for W. somnifera-treated cells were expressed as percentages relative to vehicle-treated cells. Bars represent mean ± SD from two independent experiments. Data were analysed using unpaired t-test (A) and one-way ANOVA followed by Dunnett’s test (B). ns indicates not significant

Treatment of Kelly cells with sub-cytotoxic concentrations of W. somnifera fractions (Table 2) produced variable effects on invasion, ranging from inhibition to slight promotion compared to vehicle controls. Specifically, invasion percentages ranged from 40.38% ± 29.11 with fraction F9 to 130.72% ± 59.80 with fraction F5 (Fig. 4B). Although the reduction observed with F9 was not statistically significant, these results suggest that F9 exerts the strongest anti-invasive effect among the tested fractions.

The anti-adhesive effect of W. somnifera on Kelly cells

The cell adhesion assay was initially validated by assessing the adhesion of varying densities of Kelly cells to fibronectin. Cell adhesion was quantified by measuring the absorbance of crystal violet dye, which binds to cellular DNA. Figure 5 illustrates the linear correlation between crystal violet absorbance and the number of adherent Kelly cells.

Fig. 5.

Linearity of Kelly cell adhesion to fibronectin. Adhesion assays were performed in 24-well plates pre-coated with fibronectin. After blocking non-specific binding sites, Kelly cells were seeded at densities ranging from 0.625 to 10 10⁵ cells/mL and incubated for 40 min at 37 °C. Adherent cells were stained with crystal violet dye, followed by lysis, and absorbance of the lysates was measured at 595 nm using a plate reader. Data points represent mean ± SD (n = 4). The coefficient of determination (R²) was calculated using nonlinear regression analysis (GraphPad Prism)

Incubation of Kelly cells with the vehicle reduced their adhesion to fibronectin to 68.44% ± 21.60 compared to untreated control cells (Fig. 6A). In contrast, treatment with the sub-cytotoxic concentration of W. somnifera extract (10 µg/mL) enhanced cell adhesion to the fibronectin-coated surface, increasing it to 114.76% ± 9.92 relative to vehicle-treated cells (Fig. 6B). Treatment with sub-cytotoxic concentrations of W. somnifera fractions (Table 2) resulted in variable modulation of cell adhesion. The lowest and statistically significant adhesion was observed with fraction F10 (38.31% ± 27.99), while the highest was observed with fraction F4 (116.71% ± 22.45) compared to vehicle-treated cells (Fig. 6B). These results suggest that fraction F10 exhibits the strongest anti-adhesive effect on Kelly cells among the tested fractions.

Fig. 6.

Percentage adhesion of Kelly cells treated with W. somnifera to fibronectin. Kelly cells were seeded at a density of 1 10⁵ cells/mL in 6-well plates. After 24 h, cells were incubated for 96 h with medium alone (control), medium with vehicle (A), or vehicle containing sub-cytotoxic concentrations of W. somnifera extract or fractions (F1-F12, skipping F6) (B). Adhesion assays were performed in 24-well plates pre-coated with fibronectin. Following blocking of non-specific binding sites, 1 mL of a 9 mL cell suspension was added per well and incubated for 40 min at 37°C. Adherent cells were stained with crystal violet dye, lysed, and absorbance of the lysates was measured at 595 nm using a plate reader. Adhesion values for treated cells were expressed as percentages relative to vehicle-treated controls. Bars represent mean ± SD (n = 3). Data were analysed using an unpaired t-test (A) and one-way ANOVA with Dunnett’s test (B). P value ≤ 0.05 were considered statistically significant; ns indicates not significant

Sub-fractionation and purification of W. somnifera bioactive fractions

The above results indicate that among the W. somnifera fractions, treatment of Kelly cells with fraction F9 produced the strongest anti-invasive effect as well as a remarkable anti-adhesive effect. Similarly, fraction F10 exhibited the highest and a statistically significant anti-adhesive activity along with a substantial anti-invasive effect. Consequently, both fractions were subjected to sub-fractionation, evaluation, and analysis to identify the bioactive constituent(s) responsible for these effects.

Thin-layer chromatography analysis

Initially, F9 and F10 were analysed by TLC, revealing the presence of several UV-active compounds in each fraction (Fig. 7, compounds A-D). Table 3 summarizes these compounds, including the produced colours after spraying with p-anisaldehyde/sulfuric acid reagent and their respective R_f values. F9, eluted with 30% water in methanol, and F10, eluted with 20% water in methanol, contained largely similar compounds, which is expected given the nearly similar elution strengths of the used solvent systems. Additionally, F10 exhibited a greater proportion of nonpolar compounds compared to F9, consistent with the use of a less polar solvent system for F10 relative to F9.

Fig. 7.

Thin-layer chromatogram of F9 and F10. F9 and F10 were applied to a silica gel TLC plate coated with a fluorescence indicator (F₂₅₄) on a glass backing. The plate was developed using 5% methanol in chloroform as the mobile phase. Visualization was performed under a UV lamp at 254 nm (A), followed by spraying with p-anisaldehyde/sulfuric acid reagent and heating to reveal coloured spots (B). Retention factors (R_f) for the major compounds were calculated and are presented in Table 3

Table 3.

TLC-detected compounds in F9 and F10

Compound	UV Detection	Colour*	Rf value
A	+ ve	Blue	0.34
B	+ ve	Bluish Violet	0.43
C	+ ve	Violet	0.50
D	+ ve	Bluish Violet	0.77

* After spraying with p-anisaldehyde/sulfuric acid and heating

Analytical UPLC

Analytical UPLC analysis determined that the optimal eluent solvent system consists of 30–40% acetonitrile in water. The chromatogram of fraction 9 revealed four major compounds with retention times ranging from 6.95 to 12.00 min, whereas fraction 10 exhibited five distinct compounds eluting between 20.00 and 30.77 min (Fig. 8).

Fig. 8.

Chromatographic analysis of F9 and F10. Analytical UPLC was performed using a C18 column with isocratic elution of 30% acetonitrile in water at a flow rate of 0.5 mL/min. Compounds were detected at 254 nm. F9 exhibited four main peaks (A), while F10 displayed five distinct peaks (B)

Preparative HPLC analyses

Based on the optimized conditions established from the analytical UPLC analysis, subfractionation of F9 and F10 was performed using a preparative HPLC system. Figure 9 presents the HPLC chromatograms of F9 and F10. As shown in Fig. 9A, subfractionation of F9 yielded several UV-active compounds, with three major subfractions designated as 9/1, 9/2, and 9/3. In contrast, F10 produced one major subfraction, 10/2, and two minor subfractions, 10/1 and 10/3 (Fig. 9B). Table 4 summarizes the isolated subfractions.

Fig. 9.

Representative HPLC chromatograms of F9 and F10 subfractionation. Subfractionation was performed on an XBridge™ Prep C18 column using isocratic elution with 40% acetonitrile in water at a flow rate of 2 mL/min. Compounds were detected at 254 nm. Three subfractions, 9/1, 9/2, and 9/3, were isolated from fraction F9 (A), and three subfractions, 10/1, 10/2, and 10/3, were obtained from fraction F10 (B)

Table 4.

Subfractions obtained from the bioactive W. somnifera fractions F9 and F10

Fraction	Subfraction	Weight (mg)
F9	9/1	31.1
	9/2	7.2
	9/3	2.3
F10	10/1	4.75
	10/2	8.7
	10/3	1.35

Anti-adhesive effects of W. somnifera of F9 and F10 subfractions

The bioactivity of the produced subfractions was assessed using cell adhesion assay on Kelly cells at sub-cytotoxic concentrations.

Sub-cytotoxic concentration of W. somnifera subfractions on Kelly cells

The sub-cytotoxic concentrations of W. somnifera subfractions on Kelly cells were determined, as shown in Fig. 10 and summarized in Table 5. It should be noted that the vehicle exhibited no cytotoxic effect on Kelly cells, as demonstrated in Fig. 1 relative to control untreated cells.

Fig. 10.

Viability of Kelly Cells treated with W. somnifera subfractions 9 and 10. Kelly cells were incubated for 96 h in 96-well plates with either medium containing 0.5% DMSO (vehicle) or the vehicle supplemented with various concentrations of W. somnifera subfractions: 9/1 (A), 9/2 (B), 9/3 (C), 10/1 (D), 10/2 (E), and 10/3 (F). Following incubation, cell viability was assessed using the MTT assay. Absorbance of the resulting colour was measured at 450 nm using a microplate reader. Viability percentages were calculated relative to vehicle-treated cells. Data are presented as mean ± SD (n = 4). Statistical analysis was performed using ANOVA followed by Dunnett’s test (GraphPad Prism). P ≤ 0.05 were considered statistically significant; “ns” indicates non-significance. Asterisk below the x-axes indicates sub-cytotoxic concentrations

Table 5.

The sub-cytotoxic concentrations of W. somnifera subfractions

Subfraction	Sub-cytotoxic concentration
F9/1	10 µg/mL
F9/2	0.1 µg/mL
F9/3	5 µg/mL
F10/1	10 µg/mL
F10/2	10 µg/mL
F10/3	5 µg/mL

Anti-adhesive effect of W. somnifera subfractions on Kelly cells

The anti-adhesive effects of the subfractions derived from the bioactive F9 and F10 were further evaluated. Given that both F9 and F10 exhibited anti-invasive activity, it is plausible that their anti-adhesive effects may be associated with their anti-invasive properties.

Treatment of Kelly cells with sub-cytotoxic concentrations of W. somnifera subfractions (Table 5) produced reduced adhesion to fibronectin compared to vehicle-treated cells. The observed reduction ranged from 50.89% ± 37.54 for subfraction 10/1 to 88.11% ± 22.41 for subfraction 9/1 (Fig. 11). Among all tested subfractions, 10/1 showed the most significant anti-adhesive effect on Kelly cells. This was followed by non-significant reductions observed with subfractions 10/3 and 9/2, which reduced adhesion to 53.68% ± 11.36 and 57.00% ± 5.00, respectively.

Fig. 11.

Adhesion percentages of Kelly cells treated with W. somnifera subfractions to fibronectin. Kelly cells were seeded at a density of 1 10⁵ cells/mL in 6-well plates. The following day, cells were treated for 96 h with either medium containing vehicle (0.5% DMSO) or sub-cytotoxic concentrations of W. somnifera subfractions (9/1, 9/2, 9/3, 10/1, 10/2, and 10/3). Adhesion assays were performed in 24-well plates precoated with fibronectin. After blocking non-specific binding sites, 1 mL of the 9 mL cell suspension was added per well and incubated for 40 min at 37 °C. Adherent cells were stained with crystal violet, lysed, and absorbance was measured at 595 nm using a plate reader. Adhesion values of treated cells were expressed as percentages relative to vehicle-treated controls. Bars represent mean ± SD (n = 3). Statistical analysis was performed using one-way ANOVA followed by Dunnett’s test. P ≤ 0.05 was considered statistically significant; ns = not significant

It is noteworthy that the vehicle (0.5% DMSO) also exhibited a reducing effect on cell adhesion to fibronectin compared to untreated control cells, as shown in Fig. 6A.

Tentative identification of constituent(s) present in the bioactive subfractions 9/2, 10/1 and 10/3

Based on their observed anti-adhesive potential, subfractions 9/2, 10/1, and 10/3 were subjected to GC-MS analysis to tentatively identify constituents potentially responsible for this activity. Figure 12 displays the total ion chromatogram (TIC) of subfraction 9/2, revealing six distinct peaks with their corresponding mass fragmentation patterns. The predominant compound in this subfraction eluted at a retention time (t_r) of 22.49 min, accounting for 48% of the total peak area. Another major compound eluted at t_r = 21.18 min, representing 25.61% of the total peak area.

Fig. 12.

GC-MS total ion chromatogram (TIC) and mass spectra of compounds detected in subfraction 9/2. Peaks were tentatively identified by comparing their mass spectral data with entries in the NIST MS Search 2.0 library stored in the GC-MS system. Corresponding compound information is summarized in Table 6

Similarly, the TIC of subfraction 10/1 (Fig. 13) exhibited a single dominant peak at t_r = 30.75 min, with a peak area representing 100% of the total ion signal. This compound was tentatively identified as bis(2-ethylhexyl) hexanedioate, based on spectral matching with the NIST MS Search 2.0 library integrated into the GC-MS system.

Fig. 13.

GC-MS total ion chromatogram (TIC) and mass spectra of compounds detected in subfraction 10/1. Peaks were tentatively identified by comparing their mass spectral data with entries in the NIST MS Search 2.0 library stored in the GC-MS system. Corresponding compound information is summarized in Table 6

In contrast, the TIC of subfraction 10/3 (Fig. 14) showed seven distinct peaks along with their corresponding fragmentation patterns. One of the major constituents eluted at t_r = 20.51 min, contributing 25.21% of the total peak area, while another prominent compound eluted at t_r = 22.03 min, accounting for 49.6% of the total area.

Fig. 14.

GC-MS total ion chromatogram (TIC) and mass spectra of compounds detected in subfraction 10/3. Peaks were tentatively identified by comparing their mass spectral data with entries in the NIST MS Search 2.0 library stored in the GC-MS system. Corresponding compound information is summarized in Table 6

Other detected compounds were present in low abundance and are unlikely to contribute significantly to the observed bioactivities of the subfractions. Table 6 summarizes the identified compounds in each subfraction, based on spectral matching with the NIST MS Search 2.0 library.

Table 6.

Compounds identified in W. somnifera bioactive subfractions

Subfraction #	% Area	tr (min)	a Identified Compounds
9/2	4.05	18.85	[1-(8-methyl)phenanthryl] methyl triphenylphosphonium bromide
	13.31	20.76	methyl-3-(3,5-ditertbutyl-4-hydroxyphenyl) propionate
	25.61	21.18	2,4,5-trimethyl-2-(2,3-dimethylcyclopenten-3-yl)-1,3-dioxolane
	7.11	21.41	(3R,4S)-3-(2-nitro-4-methoxyphenyl)-4-(4-hydroxyphenyl)hexane
	47.99	22.49	2-(4-octen-4-yl)-5-cycloocten-1-one
10/1	100	30.75	bis(2-ethylhexyl) hexanedioate
10/3	25.21	20.51	4,8a-dimethyl-6-(prop-1-en-2-yl)-1,2,3,5,6,7,8,8a-octahydronaphthalene-2,3-diol (known as britanlin E)
	5.05	20.71	4,5-dimethyl-1-(1-hydroxy-2-propyl)-5-(3-methyl-2-pentenoyl)- bicyclo[4.3.0]nonane
	1.63	20.84	eicosamethylcyclodecasiloxane
	5.31	21.41	(3R,4S)-3-(2-nitro-4-methoxyphenyl)-4-(4-hydroxyphenyl)hexane
	49.6	22.03	8-hydroxy-6-methoxy-3-methyl-1-oxo-1 H-2-benzopyran-7-carboxaldehyde (known as 7-formyl-8-hydroxy-6-methoxy-3-methylisocoumarin)
	2.23	23.14	tetracosamethylcyclododecasiloxane
	10.97	30.77	di-(2-ethylhexyl)adipate

^a Identified by matching with mass spectral data in the NIST MS Search 2.0 library stored in the GC-MS system, and comparison with data reported in the literature

Discussion

Neuroblastoma (NB) is the most common extracranial malignant childhood tumor and accounts for approximately 15% of cancer-related childhood deaths. Despite aggressive therapeutic regimens, more than 50% of high-risk NB patients experience tumor recurrence and late adverse effects. Withania somnifera, commonly known as ashwagandha, is widely used in Ayurvedic traditional medicine, with several preclinical studies demonstrating its potential to prevent or slow cancer progression [20]. Additionally, W. somnifera was reported to attenuate adverse effects associated with conventional chemotherapy [21–25]. This study aimed to evaluate the anti-progression effects of W. somnifera on human NB Kelly cells at sub-cytotoxic concentrations and to tentatively identify bioactive constituents responsible for these effects.

Roots of W. somnifera were extracted and subjected to extensive fractionation, yielding twelve fractions of varying polarity. Sub-cytotoxic concentrations of the crude extract and fractions were determined to assess their potential anti-invasive and anti-adhesive effects on viable Kelly cells. In vitro invasion assays through ECM revealed that the crude extract remarkably reduced Kelly cell invasion, while individual fractions exerted either inhibitory or stimulatory effects compared to vehicle-treated cells. Among the fractions, fraction 9 (F9), eluted with 70% methanol, exhibited the most pronounced anti-invasive effect, suggesting its potential as a source of promising anti-metastatic agents. This finding is consistent with previous studies demonstrating that intraperitoneal administration of 70% alcoholic extracts of W. somnifera roots significantly inhibited lung metastases in mice bearing highly metastatic B16F-10 melanoma cells [26].

Regarding cell adhesion, treatment with the crude ethanolic extract increased Kelly cell adhesion to fibronectin, whereas fractions elicited varied effects. Fraction 10 (F10), eluted with 80% methanol, showed the greatest and statistically significant reduction in adhesion to fibronectin. Together, F9 exhibited the highest anti-invasive activity alongside substantial anti-adhesive effect, while F10 showed the strongest anti-adhesive activity coupled with notable anti-invasive potential.

The relationship between cancer cell adhesion to ECM components and metastatic potential is complex and dependent on cell type and oncogenic drivers. Increased or decreased adhesion to fibronectin may correlate with metastatic behaviour [27]. To our knowledge, no prior studies have assessed Kelly cell adhesion to fibronectin or other ECM components in relation to aggressiveness. Fibronectin, a key glycoprotein within the ECM, is integral to cell proliferation, embryonic development, tissue integrity, and cancer metastasis. Metastatic cancer cells typically express highly branched, sialylated glycoproteins on their surfaces, which mediate adhesion to ECM substrates and influence metastatic potential [28]. Previously, we identified polysialic acid (PSA) glycan, a polymer of α-2,8-linked sialic acid residues, on the surface of Kelly cells [29].

In the current study, several W. somnifera fractions reduced Kelly cell adhesion to fibronectin, potentially mediated via structural alterations of PSA attached to the neural cell adhesion molecule (NCAM) on the cell surface. Consistent with our results, W. somnifera leaf extracts have been shown to exert anti-metastatic effects in rat C6 glioma cells, evidenced by increased NCAM expression and reduced cellular motility [30, 31]. Similarly, treatment of IMR-32 NB cells with a 5% aqueous leaf extract of W. somnifera resulted in approximately 85% reduction in polysialylated NCAM surface expression, decreased migration rates, inhibited proliferation, and induced differentiation [32].

Following identification of bioactive fractions F9 and F10, their subfractions were further investigated. Sub-cytotoxic concentrations of these subfractions were tested for anti-adhesive effects, revealing that subfraction 10/1 significantly inhibited Kelly cell adhesion to fibronectin, while subfractions 10/3 and 9/2 showed non-significant but notable reductions. Consequently, GC-MS analysis was performed on subfractions 9/2, 10/1, and 10/3 to tentatively identify bioactive constituents.

In subfraction 9/2, the major compound identified was 2-(4-octen-4-yl)-5-cycloocten-1-one (Fig. 15), previously reported only in the volatile oil of Serissa serissoides Druce [33], marking its first report from W. somnifera. Another major compound in 9/2 was 2,4,5-trimethyl-2-(2,3-dimethylcyclopenten-3-yl)-1,3-dioxolane, which has not been previously identified as a natural product and is likely an artifact generated during isolation or analysis.

Fig. 15.

Key tentatively identified compounds in subfractions 9/2, 101/1, and 10/3

One major compound detected in subfraction 10/1 (Fig. 15) was tentatively identified as bis(2-ethylhexyl) hexanedioate. This compound has previously been reported in several plants, including Thymus schimperi and Rhamnus prinoides. Dichloromethane extracts of these plants have demonstrated antioxidant and antibacterial activities [34]. However, this is the first report of its presence in Withania somnifera.

Finally, analysis of subfraction 10/3 led to the tentative identification of the known sesquiterpene britanlin E (Fig. 15). This compound has previously been reported in the flowers of Inula britannica L [35]. and the rhizomes of Cyperus rotundus L [36]. To the best of our knowledge, this is the first report of britanlin E in Withania somnifera. Sesquiterpenes, which are 15-carbon isoprenoid compounds, are widely distributed in plants and marine organisms. They are known for their therapeutic potential, including anti-cancer activity by inhibiting cancer progression, and are considered promising candidates for chemotherapeutic development [37]. Specifically, britanlin E has shown cytotoxic activity against human ovarian A2780 and endometrial Ishikawa adenocarcinoma cell lines [38].

Another major compound identified in subfraction 10/3 was 8-hydroxy-6-methoxy-3-methyl-1-oxo-1 H-2-benzopyran-7-carboxaldehyde, also known as 7-formyl-8-hydroxy-6-methoxy-3-methylisocoumarin (Fig. 15). This compound has previously been reported as a biosynthetic intermediate in the production of the antibiotic canescin by the actinomycete Aspergillus malignus [39]. However, it has not been previously identified from a plant source.

In summary, four key natural products were tentatively identified in subfractions 9/2, 10/1, and 10/3. Three of these compounds have been previously reported from plant sources and one from actinomycetes; however, none have been previously reported in Withania somnifera. It is important to note that these identifications are tentative and require further confirmation through purification and comprehensive structural elucidation.

Limitations of the study

The identification of bioactive constituents in Withania somnifera root subfractions was tentatively conducted using GC-MS. It is important to acknowledge that GC-MS, particularly when using electron ionization, is inherently biased toward the detection of low molecular weight, volatile, and thermally stable compounds, typically those represented in standard spectral libraries. Consequently, the absence of well-known W. somnifera constituents such as withanolides can be attributed to their high molecular weight, thermal instability, and non-volatility, which make them unsuitable for GC-MS analysis without prior derivatization.

As the current methodology did not include a derivatization step, such compounds were not detected. This limitation is recognized, and future studies are planned to employ complementary techniques such as NMR spectroscopy and LC-MS, which are more suitable for the detection and structural elucidation of higher molecular weight and thermolabile phytoconstituents.

Conclusions

This study demonstrated the anti-progressive potential of bioactive constituents derived from Withania somnifera roots using the human neuroblastoma (NB) Kelly cell line as a model. The tentative identification of these compounds highlights their potential as lead candidates for the development of anti-metastatic agents targeting aggressive NB. These compounds may serve as valuable adjuncts to current chemotherapeutic regimens in high-risk NB, potentially enabling dose reduction of cytotoxic drugs and thereby minimizing associated adverse and long-term effects. This integrative approach could help limit NB progression by impairing critical processes such as cellular adhesion and invasion, while maintaining cytotoxic efficacy.

Future studies should focus on the in vivo evaluation of these bioactive constituents to confirm their anti-metastatic activity, alongside investigations into their pharmacokinetic profiles, metabolic stability, and safety. Such work may ultimately support the development of novel adjuvant therapies derived from W. somnifera.

Abbreviations

NB

Neuroblastoma

W. somnifera

Withania somnifera

TLC

Thin layer chromatography

UV

Ultra-violet

UPLC

Ultraperformance liquid chromatography

HPLC

High performance liquid chromatography

GC-MS

Gas chromatography-mass spectrometry

DMSO

Dimethyl sulfoxide

MTT

3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide

SD

Standard deviation

BMT

Basement membrane extract

ECM

Extracellular matrix

PSA

Polysialic acid

NCAM

Neural cell adhesion molecule

Author contributions

N. AA. Principal Investigator carried out the project funding acquisition and administration, contributed to conception, design of the investigation, methodology, data curation and analysis, writing the original draft, reviewing and editing the manuscript. N. FA. co-author contributed to design of the investigation, methodology, data curation and analysis, reviewing and editing the manuscript. J. J contributed to the methodology. K. YO. co-author contributed to conception and design of the investigation, methodology, data acquisition and analysis, writing the original draft, reviewing and editing the manuscript. All authors read and approved the final version of the manuscript.

Funding

This work was financially supported by the Research Administration, Kuwait University (Grant No. RP01/16). The GC-MS analyses were done at the Research Sector Project Unit, College of Science, Kuwait University, supported by Grant No. GS01/03.

Data availability

All data generated or analyzed during this study are included in the article and there is no supplementary information files of uncropped Gels and Blots images.

Declarations

Ethical approval

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.