Different parts (leaves, fruits, and barks) of S. pobeguinii were harvested in Ezezan (Nyom II), a neighbouring locality situated at 40 km from Yaoundé (Cameroon). A voucher specimen was prepared and the authentication was done by Mr. Ngansop Eric, a plant taxonomist, after comparison with the specimen number N°32,567 BRF/CAM already available in the library of the National Herbarium of Cameroon. Our plant material was then registered under the number Letouzey R.12493 (YA).

The powder (907 g) obtained from the air dried and grounded fruits of S. pobeguinii was macerated in the mixture of methanol and methylene chloride (1:1) at room temperature for 48 h with occasional stirring, and the crude extract was obtained after filtration with Whatman N⁰1, followed by evaporation of the solvent to dryness in vacuo. The dried extract (76.5 g) was re-dissolved in MeOH and partitioned with ethyl acetate (EtOAc) to yield the EtOAc fraction (10.2 g) and the MeOH residue (54.7 g). The MeOH soluble fraction was then subjected to silica gel column chromatography (CC) eluted with a gradient mixture of n-hexane-EtOAc (9:1; 7:3; 6:4; 0:1). The collected sub-fractions were pooled based on their thin layer chromatography (TLC) profiles and yielded six sub-fractions (F1-F6) after evaporation. Sub-fraction F₃ (8 g) was submitted to silica gel CC, and eluted with the mixture of n-hexane-EtOAc (2:1) to yield the following compounds: (1) (6 mg) identified as a mixture of nauclealatifoline G and naucleofficine D (Agomuoh et al., 2013); (2) (7 mg) known as hederagenin (Joshi et al., 1999); and (3) (6 mg) characterized as chletric acid (Takahashi and Takani, 1978).

The powder of air-dried leaves of S. pobeguinii (1 kg) was extracted in MeOH and heated at 40°C–50°C for 2 h. The methanolic extract was concentrated in vacuo to yield a dark greenish mass (66 g) which was partitioned with n-hexane. The n-hexane soluble part and the methanolic residue were evaporated under pressure to obtain respectively a black residue (26.5 g) and a greenish brown extract (37.3 g). The latter was subjected to silica gel flash chromatography, and eluted respectively with CH₂Cl₂; CH₂Cl₂/MeOH (1:1) and MeOH. The n-hexane soluble part and the CH₂Cl₂ fraction were mixed due to their similar TLC profiles, and this mixed fraction was submitted to silica gel CC eluted with the gradient mixture of n-hexane and EtOAc (9:1; 3:1; 3:2; 1:1; 1:3) to afford compound (4) (5 mg) identified as taraxerol (Yen et al., 2013) collected in n-hexane/EtOAc (3:1); compound (5) (4 mg) known as α-amyrin (3β-hydroxy-urs-12-en-3-ol) (Viet et al., 2021) collected in n-hexane/EtOAc (1:1); and (6) (6 mg) known as quinovic acid 3-O-[α-D-quinovopyranoside] (Mohamed, 1999) collected in n-hexane/EtOAc (1:3).

The powder of dried bark of S. pobeguinii (1 kg) was extracted in MeOH and heated at 40°C–50°C for 2 h. The crude extract was obtained after filtration with Whatman N⁰1, followed by evaporation under reduced pressure to obtain a yellow gummy residue (77 g). This methanolic extract was partitioned with n-hexane to yield the n-hexane fraction (15 g) and a residual MeOH fraction (54 g). The MeOH fraction was then subjected to silica gel CC, and eluted with a gradient of n-hexane/EtOAc yielding the following compounds: (7) (4 mg) identified as erythrodiol and (8) (8 mg) known as quinovic acid (Fatima et al., 2002; Aktar et al., 2009) both collected in the n-hexane/EtOAc (9:1); then compound (9) (8 mg) characterized as quinovic acid 3-O-[β-D-quinovopyranoside] (Yepez et al., 1991) collected in n-hexane/EtOAc (3:1); and finally compound (10) (4 mg) identified as latifoliamide C (Agomuoh et al., 2013) collected in n-hexane/EtOAc (1:9).

The chemical constituents of S. pobeguinii were purified using open silica gel column chromatography (Merck, [Darmstadt, Germany]). TLC was done on Alu g R; SIL G/UV₂₅₄ silica gel plates (Merck, [Darmstadt, Germany]), and visualization of the spots on TLC plates was achieved either by exposure to iodine vapour, UV light or by spraying sulphuric acid and heating the plate at 75°C. Melting points were recorded on a Buchi B-54 apparatus. ¹H and ¹³C NMR spectra as well as 2D NMR experiments (see Supplementary Material, Supplementary Figures S1–S22) were recorded in CDCl₃, MeOH- _d4 , DMSO- _d6 and pyridine- _d5 in a JEOL ECX 500 spectrometer (Akishima, Japan) and on a Bruker ARX 400. Chemical shifts were expressed in part per million δ) relative to tetramethylsilane as internal standard. The spectroscopy data recorded were compared with data from literature, for the characterization of the chemical structures of isolated compounds which are presented in Figure 1.

Cancerous cell lines (MCF-7: human breast adenocarcinoma cells; HepG2: human hepatocellular carcinoma cells; Caco-2: human epithelial colorectal adenocarcinoma cells; A549: human epithelial lung adenocarcinoma cells), obtained from the American Type Culture Collection (ATCC) (Rockville, MD, United States of America), were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) high glucose (4.5 g/L) containing L-glutamine (4 mM) and sodium-pyruvate (Hyclone™) supplemented with 10% (v/v) fetal bovine serum (FBS) (Capricorn Scientific GmbH, South America), and incubated at 37°C with 5% CO₂ in a humidified environment. African green monkey (Vero) kidney cells (also obtained from ATCC), a non-cancerous cell line, were grown in DMEM high glucose (4.5 g/L) containing L-glutamine (Lonza, Belgium) and supplemented with 5% FBS (Capricorn Scientific GmbH, South America) and 1% gentamicin (Virbac, RSA) in the same environment as cancer cells. The RAW 264.7 murine macrophage cells (obtained from ATCC) were cultured in DMEM high glucose (4.5 g/L) containing L-glutamine (Lonza, Belgium), supplemented with 10% FBS (Capricorn Scientific GmbH, South America) and 1% penicillin/streptomycin/fungizone (PSF) solution, and kept at 37°C in a 5% CO₂ humidified environment.

The cancerous and non-cancerous cells were seeded at a density of 10⁴ cells per well in 96-well microtiter plates, and they were incubated overnight at 37°C with 5% CO₂ in a 5% CO₂ humidified environment in order to allow the attachment of cells at the bottom of the plates. Then, the cells were exposed to increasing concentrations of compounds (100, 50, 25, 10 and 5 μg/mL) dissolved in dimethyl sulfoxide (DMSO) and further diluted in fresh culture medium. In this assay, the final concentration of DMSO in the culture medium was 0.5% used as negative control while doxorubicin hydrochloride (Pfizer, United States of America) was used as a positive control. The culture plates were incubated for 48 h at 37°C in a 5% CO₂ humidified environment, after which the culture medium was discarded, and replaced by 200 µL of fresh culture medium with 30 µL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) (5 mg/mL) dissolved in phosphate buffered saline (Mosmann, 1983). After 4 h of incubation, the medium was gently removed, and the formazan crystals were solubilized in 50 µL of DMSO. The absorbance was measured at 570 nm after shaking for 1 min on a microplate reader (Synergy Multi-Mode Reader, BioTek, Winooski, United States of America).

The cell viability rate was determined at each concentration of the compound as a percentage of cells treated with DMSO at 0.5% used as negative control. The 50% inhibitory concentrations (IC₅₀) were determined by using the non-linear regression graphical analysis of cell viability rate against the logarithm (log10) of compound concentrations with the software GraphPad Prism 6.0 (GraphPad software Inc., United States of America). The selectivity index (Supporting Material) values were calculated for each compound by dividing the IC₅₀ of non-cancerous cells by the IC₅₀ of cancerous cells in the same units (Mfotie Njoya et al., 2018; Mfotie Njoya et al., 2020).

The effect of the most bioactive compounds (2), (6) and (9) were used for the analysis of caspase-3 and caspase-7 activities on MCF-7 cells by using the Caspase-Glo^® 3/7 kit (Promega, Germany). In fact, MCF-7 cells were seeded at a density of 10⁴ cells per well on 96-well microtiter plates, and the plates were incubated overnight at 37 °C in a 5% CO₂ humidified environment. Then, the cells were exposed to the active compounds at different concentrations (½×IC₅₀, IC₅₀ and 2×IC₅₀) or DMSO (0.5%) used as negative control, and further incubated for 18 h at 37°C in a 5% CO₂ humidified environment. Thereafter, 100 µL of Caspase-Glo^® 3/7 reagent was added to each well, mixed and incubated in the dark for 1 h at room temperature. The luminescence was then measured on a microplate reader (Synergy Multi-Mode Reader, BioTek, Winooski, United States of America). The caspase-3/-7 activity was expressed as fold change of cells treated with 0.5% DMSO (control).

All experiments were performed in triplicate, and the results are presented as mean ± standard deviation (SD) values. The statistical analysis was done with the software GraphPad Prism 6.0 (GraphPad software Inc., United States of America) on which one-way analysis of variance (ANOVA) and Student–Newman–Keuls or Dunnett’s tests were used for the comparison of data among tested samples and/or controls. Results were considered significantly different when the p-value was greater than 0.05.

The X-ray crystal structures of putative target proteins in complex with inhibitors were retrieved from Protein Data Bank (https://www.rcsb.org/, accessed on 14 March 2023): Cyclooxygenase-2: COX-2 (PDB ID: 5KIR); Nuclear factor NF-kappa-B (p65 subunit): NFKB-p65 (PDB ID: 2RAM); Janus kinase 2: JAK2 (PDB ID: 6BBV); Signal transducer and activator of transcription 3: STAT3 (PDB ID: 6TLC); CASPASE-3 (PDB ID: 3DEI), CASPASE-7 (PDB ID: 1SHJ). PyMol v 2.0.7, was used to prepare the proteins for simulation by removing the inhibitors, water molecules, multi-chains and heteroatoms. MarvinSketch was used to draw the chemical structures of the ligands, and the Merck Molecular Force Field (MMFF94) as provided by MarvinSketch was used for energy minimization on the ligands (compounds) (Novichikhina et al., 2020). The energy minimization was carried out to make the ligands more stable near their initial states during molecular docking process. AutoDock Vina v 1.5.6 was used to add hydrogen atoms and Gasteiger charges accordingly to the ligands prior to molecular docking (Trott and Olson, 2010). The active sites of the target proteins were identified using CASTp 2.0 web-based tool (Tian et al., 2018). Further, AutoDock Vina v 1.5.6 was used to create a grid box around the identified active sites of the target proteins. Thereafter, standard precision ligand docking was performed using the same software. The best of nine binding poses based on the docking scores were visualized and captured in BIOVIA Discovery Studio Visualizer 2021 (Mfotie Njoya et al., 2023).

The growth inhibitory effect of natural compounds isolated from leaf, fruits and bark of Sarcocephalus pobeguinii (Rubiaceae) against four cancerous cells (MCF-7, HepG2, Caco-2 and A549) and non-cancerous Vero cells was expressed as percentage of respective cells treated with DMSO (0.5%) used as negative controls. Based on the results obtained for each compound tested at different concentrations, the 50% inhibitory concentrations (IC₅₀) were determined and presented in Table 1. It was observed that hederagenin (2), quinovic acid 3-O-[α-D-quinovopyranoside] (6) and quinovic acid 3-O-[β-D-quinovopyranoside] (9) exhibited significant (p < 0.05) antiproliferative effect against all cancerous cells, especially for (6) which showed the highest cytotoxic effect on A549 cells with IC₅₀ of 1.96 μg/mL (3.08 µM). Additionally, (6) was also toxic to the non-cancerous Vero cells with IC₅₀ of 6.21 μg/mL (9.81 µM). By calculating the selectivity indexes (SI) as presented in Table 2, we found that (6), isolated from leaves of S. pobeguinii, had a poor selectivity (SI ≤ 1) (except for A549 cells) towards non-cancerous Vero cells where SI was 3.17. Similarly, (9), isolated from bark of S. pobeguinii, also had poor selectivity (SI ≤ 1) for all cancer types. On the other hand, (2), isolated from fruits of S. pobeguinii, was the only compound which had acceptable selectivity (1.24 ≤ SI ≤ 5.11) for all cancer cells towards non-cancerous Vero cells thereby suggesting that this compound can selectively kill cancer cells at the IC₅₀ without causing harmful effects on non-cancerous cells. Therefore, regarding the fact that hederagenin (2) exhibited good selectivity, this compound was tested at 10 μg/mL on all cancer cells and non-cancerous cells for 12, 24 and 48 h in order to evaluate the time-dependent cytotoxic effect on cell growth. It was observed that (2) induced the growth inhibition of different cancer cells in a time-dependent manner, and the cytotoxic effect was more pronounced on cancer cells compared to non-cancerous cells (Figure 2). In fact, the MCF-7 cells followed by A549 cells were the most sensitive to the cytotoxic effect of hederagenin (2) which was less toxic to non-cancerous Vero cells. Interestingly, hederagenin, isolated from several plants, has been previously reported to potentially inhibit the proliferation of human lung cancer cells A549 (IC₅₀ values of 26.3 and 39 μM) (Gauthier et al., 2009; Gao et al., 2016). In the current study, hederagenin inhibits A549 growth with an IC₅₀ of 40.11 µM which is comparable to results obtained from previous works. Additionally, hederagenin was found to be less cytotoxic to human normal skin fibroblasts (WS1) cell lines (IC₅₀ of 77 μM) (Gauthier et al., 2009) which also confirms the fact that this compound selectively kills cancer cells with least toxic effect to normal cells. Further, hederagenin, extracted in high quantities from the fruits of Sapindus saponaria L., (Sapindaceae), has been reported to be cytotoxic to different cancer cells, and structural modifications resulted to the development of potent anti-tumour compounds (Rodriguez-Hernandez et al., 2015; Rodriguez-Hernandez et al., 2016). Overall, hederagenin was isolated for the first time from fruits of S. pobeguinii (Rubiaceae), and our data corroborates its use as a potential candidate for the development of anticancer agents.

Three triterpenoids including two saponins isolated from S. pobeguinii were cytotoxic to cancer cells, and their efficacy varied according to their structural configuration or the presence of some functional groups. When compared to other triterpenoids, the cytotoxic effect of the two bioactive saponins, namely, quinovic acid 3-O-[α-D-quinovopyranoside] (6) and quinovic acid 3-O-[β-D-quinovopyranoside] (9) can be attributed to the presence of a pyranose moiety attached at position C-3. Moreover, the α anomer (6) is highly cytotoxic compared to the β anomer (9) which suggests that the configuration of the glycosidic bond (C-O-sugar bond) has an impact on the cytotoxic effect. According to several reports, the anticancer activity of triterpenoid saponins is strongly linked to the presence of functional carboxylic and hydroxyl groups on the aglycone chain, the stereo-selectivity and the type of sugar molecule attached (Nag et al., 2012; Xu et al., 2013; Elekofehinti et al., 2021; Podolak et al., 2023). In contrast, we observed from our study that despite the presence carboxylic and hydroxyl groups, other triterpenoids (3), (4), (5), (7) and (8) were less cytotoxic or inactive compared to hederagenin (2) which strongly inhibit the growth of several cancer cells. We therefore suggest that the efficacy of hederagenin might be attributed to the asymmetric carbon at position C-4 which yield a different conformation due to irregular spatial arrangements of chemical groups. This conformation might ease the interaction with molecular targets involved in the activation of cancer cell death pathways. Further, the solubility of hederagenin in culture media may also be responsible for its efficacy on cancer cells.

The potential mechanism of action of bioactive compounds on cellular viability was explored by quantifying the caspase-3/-7 activity in MCF-7 cells. It was found that the caspase-3/-7 activity was significantly (p < 0.05) induced in a concentration-dependent manner compared to control cells treated with DMSO 0.5% (Figure 3). Additionally, the optimal effect was observed with hederagenin (2) and quinovic acid 3-O-[α-D-quinovopyranoside] (6) tested at 2×IC₅₀ where an induction up to 2-fold change in the activation of caspase-3/-7 activity was recorded. Caspases 3 and 7 are known as “executioners” of apoptosis, and the combined role of both caspases is crucial for the activation of apoptotic pathways (Olsson and Zhivotovsky, 2011). In fact, caspase 3 controls DNA fragmentation and morphologic changes of apoptosis, while caspase 7 is more important for the loss of cellular viability (Lakhani et al., 2006). Therefore, the induction of caspase-3/-7 activity after treatment with bioactive compounds implied that the activation of apoptotic pathways is involved in the mechanism of induced cell death in MCF-7 cells.

Cancer development is known as a multistep process during which inflammation is one of the factors contributing to the promotion of cell proliferation, invasion, angiogenesis, and metastasis. Moreover, it has been proven that patients on Nonsteroidal anti-inflammatory drugs (NSAIDs) are at reduced risk of cancer development (Singh et al., 2019). One of the possible reasons is that anti-inflammatory drugs might also interfere with signaling pathways modulated in both inflammation and cancer. Therefore, we decided to investigate the anti-inflammatory effect of compounds isolated from S. pobeguinii on two experimental models. We found that all compounds (tested at 100 μg/mL) inhibited the NO production in LPS-stimulated RAW 264.7 cells, with the mixture nauclealatifoline G and naucleofficine D (1), hederagenin (2), quinovic acid 3-O-[α-D-quinovopyranoside] (6) and quinovic acid 3-O-[β-D-quinovopyranoside] (9) being the most active by reducing at least 50% of NO production (Figure 4A). However, the NO production inhibitory effect of compounds (1), (2), (6) and (9) might be attributed to their cytotoxic effect on LPS-stimulated RAW 264.7 cells as it was observed that these compounds significantly (p < 0.05) reduced the cell viability (Figure 4B). Besides, by determining the IC₅₀ values as presented in Table 3, we found that (6) and (9) strongly inhibits NO production in a concentration-dependent manner with IC₅₀ values of 28.71 and 77.36 µM, respectively. On the other hand, we examined the inhibitory effect of isolated compounds on the activity of 15-LOX, an enzyme which regulate the inflammatory responses via the generation of pro-inflammatory mediators known as leukotrienes. It resulted that the mixture of nauclealatifoline G and naucleofficine D (1), hederagenin (2) and chletric acid (3) strongly inhibited the activity of 15-LOX with IC₅₀ values of 16.51, 28.20 and 41.17 µM, respectively (see Table 3). These compounds were even more active than quercetin, which had an IC₅₀ value of 61.94 µM. Based on the fact that 15-LOX and NO are among the mediators involved in the progression of various inflammation-related diseases including cancer (Zhao et al., 2021), our study therefore identified compounds which can be considered as potential anticancer agents with additional anti-inflammatory effect for the effective management of both diseases.

Six putative proteins, namely, COX-2, NFκB-p65, JAK2, STAT3, CASPASE-3 and CASPASE-7 were chosen for molecular docking based on the fact that they are frequently used as therapeutic targets to regulate inflammatory signaling pathways in cancer treatment (Jha et al., 2022). As presented in Table 4, the docking score (kcal/mol) of the best binding pose was determined for each ligand against the selected proteins. According to these results, hederagenin (2), quinovic acid 3-O-[α-D-quinovopyranoside] (6) and quinovic acid 3-O-[β-D-quinovopyranoside] (9) exhibited activity within the range of −4.3 to −6.8 kcal/mol, where −6.8 kcal/mol was the most effective interaction between JAK2 and (9), and −4.3 kcal/mol was the lowest binding score between CASPASE-3 and (2). Taken individually, each of the three bioactive compounds exhibited the greatest binding efficacy (binding score above −6 kcal/mol) against JAK2 and COX-2 which suggests that these proteins are the potential molecular targets involved in their antiproliferative and anti-inflammatory effects (Figure 5). In addition, STAT3 and NFKB-p65 also interacted well with bioactive compounds. According to the literature reports, the high activation of JAK2/STAT3 signaling pathway, frequently detected in various tumors, has recently emerged as a new site for the development of novel anti-tumor agents, and these proteins (JAK2 and STAT3) are promising therapeutic targets for the treatment of many solid tumors (Huang et al., 2022; Mengie Ayele et al., 2022). Moreover, NFκB-p65 is constitutively activated in many human cancers, where it contributes to almost all steps of tumorigenesis including sustained proliferation, cell death resistance, tumor-promoting inflammation, tissue invasion, angiogenesis, and metastasis (Lin et al., 2010). As such, the NF-κB pathway is an attractive therapeutic target in a broad range of human cancers, as well as in numerous non-malignant diseases. Furthermore, COX-2, an enzyme that catalyzes the first step in the synthesis of prostanoids, is associated with inflammatory diseases and carcinogenesis (Liu et al., 2015). An overexpression of COX-2 is observed in cancers of the pancreas, breast, colorectal, stomach, and lung carcinoma. Therefore, COX-2 is considered as a significant target for the development of new anticancer agents (Mohsin et al., 2022). Among the bioactive compounds investigated, the mechanism of action of hederagenin has been extensively studied in several studies. In fact, hederagenin has been reported to have anti-tumour and anti-inflammatory effects by regulating the NFκB, PI3K/AKT and JAK2/STAT3/MAPK signalling pathways, and by reducing the expression of pro-inflammatory cytokines or enzyme production, including Tumor Necrosis Factor- α (TNF-α), interleukin 6 (IL-6), NO, prostaglandin E2 (PGE₂), inducible nitric oxide synthase (iNOS) and COX-2 (Zhang et al., 2012; Lee et al., 2015; Kim et al., 2017; Yu et al., 2020; Shen et al., 2023). Our results indicate the accordance between the molecular docking and the previous studies thereby confirming the efficacy of hederagenin as a prominent lead compound for the management of inflammation and cancer development. Further investigations are required for quinovic acid 3-O-[α-D-quinovopyranoside] and quinovic acid 3-O-[β-D-quinovopyranoside] which modes of action are still not experimentally investigated.