Cytotoxicity profiles against RAW 264.7 murine macrophages and human fibroblast (HDF) cell lines

The cytotoxicity of the T. ferdinandiana fruit and leaf extracts was initially screened at 5 mg/mL against human dermal fibroblasts (HDF; Fig. 1a) to evaluate their suitability for therapeutic use as fibroblasts are one of the most widely used cell lines for cytotoxicity evaluations (McGaw et al. 2014). The extracts were also screened against murine RAW 264.7 macrophage to determine the concentrations to be used for testing in the cytokine, COX-2, PGE2, LTB4 and NF-κB assays (Fig. 1b). Macrophages produce inflammatory mediators during an immune response (Fullerton and Gilroy 2016). Therefore, HDF and RAW 264.7 cells were considered relevant in vitro models for the toxicity evaluations in this study.

Fig. 1

The cytotoxic effects of T. ferdinandiana fruit and leaf extracts and a turmeric control on a human dermal fibroblasts and b RAW 264.7 cell lines at 5 mg/mL as determined by MTS cell viability assays. NC negative control (media); FM = methanolic fruit extract; FW = aqueous fruit extract; LM = methanolic leaf extract; LW = aqueous leaf extract; Tum = turmeric control (1.25 mg/mL); PC = 10% DMSO-positive control. Data are represented as mean values of triplicate results ± standard deviation. ## represents results that are significantly different to the negative control (p < 0.005)

The aqueous extracts were substantially less cytotoxic towards both cell lines, compared to the corresponding methanolic extracts. Indeed, exposure to the aqueous fruit and leaf extracts (5 mg/mL) decreased HDF cell viability by approximately 23% and 33%, respectively, compared to the untreated (media) control (Fig. 1a). The RAW 264.7 cells were even less affected by the aqueous fruit and leaf extracts, with decreases in cell viability of approximately 5% and 16%, respectively. Whilst these decreases in cell viability are statistically significant, the viability for all cells exposed to the aqueous extracts were substantially greater than 50%. Therefore, the aqueous T. ferdinandiana fruit and leaf extracts were considered nontoxic at 5 mg/mL and were not tested at further dilutions. In contrast, both the fruit and leaf methanolic extracts induced substantially greater decreases in cell viability in both cell lines. Indeed, exposure to the methanolic fruit and leaf extracts resulted in HDF cell viabilities of approximately 36% and 28%, respectively (Fig. 1a). Similar decreases in RAW 264.7 viability were also noted for the fruit and leaf methanolic extracts (~ 34% and 27% of the untreated control cell viability, respectively). The turmeric control decreased cell viability in the HDF and RAW 264.7 cells by ~ 36% and 38%, respectively.

Due to the greater decrease in cell viability noted for the methanolic extracts, their toxicity was further evaluated by screening a dilution series for each extract and calculating LC50 values (Table 1). Plant extracts with LC50 values < 1 mg/mL were considered to be toxic in this study. Notably, the LC50 values were ≥ 2.5 mg/mL for all extracts. Therefore, all the T. ferdinandiana extracts were deemed to be nontoxic towards HDF and RAW 264.7 cells and 2.5 mg/mL was selected as the test extract concentration for the cytokine, COX-2, PGE2, LTB4 and NF-κB inhibition assays.

Table 1 The LC50 values (μg/mL) of T. ferdinandiana extracts against RAW 264.7 murine macrophage and human dermal fibroblasts (HDF)Modulation of cytokine secretion in RAW 264.7 cells by T. ferdinandiana extracts

The extracts were screened for their effects on cytokine secretion in both unstimulated and LPS-stimulated RAW 264.7 cells to evaluate their immune-modulatory effects. The cytokines screened in our study were selected to test the effects on both anti-inflammatory (IL-10) and pro-inflammatory cytokines (IFNγ, IL-1β, IL-6, and TNF-α) (Kany et al. 2019). IL-2 was included as it has both anti- and pro-inflammatory effects under different conditions (Kolios et al. 2021; Sharma et al. 2011). In addition, the chemokines MCP-1 and MIP-2a were included to evaluate the potential of the extracts to regulate the recruitment of monocytes to the site of the inflammation (Wong et al. 2010).

Effect of the extracts on cytokine secretion in unstimulated RAW 264.7 cells

To investigate whether the T. ferdinandiana fruit and leaf extracts modulate inflammatory mediators in the absence of a pro-inflammatory stimulus, RAW 264.7 macrophages were treated with the extracts, without prior LPS stimulus (Fig. 2). Interestingly, treatment with the extracts decreased the secretion of both pro-inflammatory and anti-inflammatory cytokines, indicating that they may be affecting a general regulatory mechanism. IL-2 (Fig. 1a) and IL-10 (Fig. 1b) were included in this study as they both have anti-inflammatory effects (IL-2 also has pro-inflammatory properties) (Kolios et al. 2021; Sharma et al. 2011). Therefore, increased secretion of these cytokines is associated with anti-inflammatory activity. Interestingly, all extracts significantly inhibited the secretion of these cytokines. The extracts were potent inhibitors of IL-10, with secretion inhibited by > 95% by all extracts. Similarly, IL-2 was also inhibited, albeit to a lesser extent (50–73% inhibition compared to the untreated control). Whilst this may not result in pro-inflammatory effects, the decreased levels of anti-inflammatory cytokines cells may confer a decreased ability for the cells to respond to inflammation. In contrast, the turmeric control induces substantial increases in IL-10 secretion (to ~ 122% of the untreated control), indicating that turmeric treatment increases the cells ability to respond to inflammation.

Fig. 2

The effects of the T. ferdinandiana fruit and leaf extracts (2.5 mg/mL) on secretion of a IL-2; b IL-10; c IFNγ; d IL-1β; e IL-6; f MCP-1; g MIP-2α; and h TNFα in unstimulated RAW 264.7 cells. Arrows indicate plant extracts that induce cytokine levels substantially greater than 120%. NC = untreated negative control (media); FM = methanolic fruit extract; FW = aqueous fruit extract; LM = methanolic leaf extract; LW = aqueous leaf extract; PC = turmeric positive control (1.25 mg/mL). Data are represented as mean values of duplicate results ± standard deviation. # and ## represent results that are significantly different to the negative control at p < 0.01 and p < 0.005, respectively

The T. ferdinandiana extracts also significantly (p < 0.005) down-regulated the secretion of the pro-inflammatory cytokines IFN-γ (Fig. 1c), IL-1β (Fig. 1d), IL-6 (Fig. 1e) and TNF-α (Fig. 1h), demonstrating that the T. ferdinandiana extracts have immune-modulatory and anti-inflammatory effects. The extracts were particularly good inhibitors of IL-6 secretion (Fig. 1e), with between 93 and 97% inhibition noted, in comparison to the untreated control. The extracts also significantly inhibited the secretion of all other pro-inflammatory cytokines (generally inhibited by 65–83% of the untreated control levels). This indicates that these extracts have anti-inflammatory effects. Notably, the turmeric control had the opposite effect to that of the T. ferdinandiana extracts. Indeed, turmeric stimulated pro-inflammatory cytokine secretion in RAW 264.7 macrophages by up to 163% of the untreated control value (for IFN-γ). Exposure to the T. ferdinandiana extracts also reduced the secretion of the MCP-1 (Fig. 1f) and MIP-2a (Fig. 1g) chemokines. Notably the turmeric control was a potent inhibitor of chemokine secretion, with a complete inhibition of MIP-2a noted.

Effect of the extracts on cytokine secretion in LPS-stimulated RAW 264.7 cells

The T. ferdinandiana extracts were also tested against LPS-stimulated RAW 264.7 macrophages to evaluate their ability to modulate immune-modulatory markers during inflammation (Fig. 3). Notably, the T. ferdinandiana extracts were potent inhibitors of the secretion of all anti-inflammatory and pro-inflammatory cytokines, as well as the chemokines. The methanolic extracts were substantially more potent inhibitors of cytokine secretion than the aqueous extracts were. Interestingly, the methanolic fruit and leaf extracts blocked 41–52% of IL-2 (Fig. 1a) and 96–98% of the secretion of IL-10 (Fig. 2b), respectively, from the RAW 264.7 cells, indicating they may potentiate the inflammatory effects of LPS. The turmeric control also significantly inhibited IL-10 secretion. In contrast, turmeric stimulated IL-2 secretion compared to the untreated LPS-stimulated control. The aqueous extracts also significantly inhibited the secretion of IL-2 and IL-10, although the inhibition was substantially less potent.

Fig. 3

The effects of the T. ferdinandiana fruit and leaf extracts (2.5 mg/mL) on secretion of a IL-2; b IL-10; c IFNγ; d IL-1β; e IL-6; f MCP-1; g MIP-2α; and h TNFα in LPS-stimulated (100 ng/mL) RAW 264.7 cells. Arrows indicate percentage inhibition of plant extracts that is substantially greater than 120% scale of the graphs. NC = untreated negative control (media); FM = methanolic fruit extract; FW = aqueous fruit extract; LM = methanolic leaf extract; LW = aqueous leaf extract; PC = turmeric positive control (1.25 mg/mL). Data are represented as mean values of duplicate results ± standard deviation. # and ## represent results that are significantly different to the negative control at p < 0.01 and p < 0.005, respectively

The T. ferdinandiana fruit and leaf extracts were substantially more potent inhibitors of secretion of the pro-inflammatory cytokines and the chemokines from RAW 264.7 cells. Indeed, the methanolic fruit and leaf extracts completely inhibited the secretion of IFN-γ (Fig. 2c), IL-1β (Fig. 2d), IL-6 (Fig. 2e), as well as inhibiting TNF-α (Fig. 2h) secretion by ~ 98%. Notably, the inhibition of pro-inflammatory cytokine secretion by the methanolic extracts was similar to that determined for the turmeric control. The aqueous extracts also significantly inhibited pro-inflammatory cytokine secretion (p < 0.005), albeit by a substantially lower percentage than noted for the methanolic extracts. The potent inhibition of the pro-inflammatory cytokines indicates that the T. ferdinandiana extracts have anti-inflammatory effects. Similarly, the T. ferdinandiana extracts (and the turmeric control) completely blocked the secretion of the MCP-1 (Fig. 2f) and MIP-2a chemokines (Fig. 2g) from RAW 264.7 macrophages.

The effect of T. ferdinandiana extracts on COX-2 levels in RAW 264.7 cells

The cytosolic levels of COX-2 in the unstimulated RAW 264.7 macrophages were unaffected by treatment with the T. ferdinandiana leaf and fruit extracts when tested at a concentration of 2.5 mg/mL (Fig. 4a). Indeed, no significant differences were noted between the cells exposed to the T. ferdinandiana extracts and the untreated control cells. This is perhaps not surprising as the COX-2 enzyme is inducible and its synthesis is up-regulated during inflammation, whilst its levels are generally low in unstimulated cells (Ju et al. 2022; Mohsin and Irfan 2019). In contrast, the turmeric positive control significantly down-regulated the cytosolic COX-2 level in the RAW 264.7 macrophages, although the COX-2 levels were only decreased by approximately 9% (p < 0.01).

Fig. 4

The effects of the T. ferdinandiana fruit and leaf extracts on COX-2 levels in a unstimulated and b LPS-stimulated (100 ng/mL) RAW 264.7 cells, as well as PGE2 levels in c unstimulated and d LPS-stimulated (100 ng/mL) RAW 264.7 cells. NC negative control (media); FM = methanolic fruit extract; FW = aqueous fruit extract; LM = methanolic leaf extract; LW = aqueous leaf extract; PC = turmeric positive control (1.25 mg/mL). Data are represented as mean values of triplicate results ± standard deviation. # and ## represent results that are significantly different to the negative control at p < 0.01 and p < 0.005, respectively

In contrast, all the T. ferdinandiana extracts substantially reduced COX-2 levels in LPS-stimulated RAW 264.7 macrophages (Fig. 4b). The methanolic fruit and leaf extracts were particularly good at reducing COX-2 levels, decreasing cytosolic COX-2 levels by approximately 87% and 95% compared to the untreated LPS-stimulated control, respectively. The aqueous extracts also significantly reduced COX-2 levels in RAW 264.7 cells (p < 0.005), with reductions of approximately 47% and 43%, respectively, compared to the untreated LPS-stimulated control. The turmeric positive control had profound effects on COX-2 levels, indicating that the assay was functioning correctly. Indeed, turmeric treatment resulted in a 98% reduction in COX-2 levels compared to the untreated control.

The effect of T. ferdinandiana extracts on PGE2 secretion in RAW 264.7 cells

Prostaglandin E2 is produced from arachidonic acid in a series of reactions, starting with an oxygenation reaction catalysed by the COX enzymes (Park et al. 2006). Therefore, we quantified the levels of PGE2 secreted by both the unstimulated and the LPS-stimulated RAW 264.7 macrophages. As noted for COX-2, the T. ferdinandiana fruit and leaf extracts had minimal effects on PGE2 secretion in non-LPS-stimulated RAW 264.7 macrophages (Fig. 4c). Indeed, neither the aqueous fruit nor the leaf extracts significantly affected the secretion of PGE2 in the unstimulated cells. Whilst significant reductions (p < 0.01) in PGE2 secretion were noted in response to treatment with the methanolic extracts, the levels were only reduced by approximately 8% compared to the untreated control. Similar decreases in PGE2 secretion were also noted following turmeric treatment.

In contrast, the T. ferdinandiana extracts had profound effects on the secretion of PGE2 in LPS-stimulated cells (Fig. 4d). The methanolic fruit and leaf extracts were particularly effective, inhibiting PGE2 secretion by approximately 97% and 99%, respectively, compared to the untreated LPS-stimulated control. The inhibition by the methanolic extracts compared favourably to the inhibition noted for the turmeric positive control (~ 93%). The aqueous fruit (~ 55% inhibition) and leaf extracts (~ 65% inhibition) were also good inhibitors of PGE2 secretion, although substantially less potent than the methanolic extracts (based on the % inhibition compared to the untreated LPS-stimulated control).

The effect of T. ferdinandiana extracts on LTB4 secretion in RAW 264.7 cells

LTB4, another pro-inflammatory mediator, is secreted by macrophages following inflammatory stimuli (as reviewed in Haeggstrom and Funk 2011). It is a potent chemotactic agent, with activity more than three orders of magnitude higher than histamine. LTB4 is a mediator of pathogenesis for multiple acute and chronic inflammatory conditions including asthma, atherosclerosis, cancer, dermatitis, inflammatory bowel disease, nephritis, psoriasis and rheumatoid arthritis (He et al. 2020; Liu and Yokomizo 2015). For this reason, the production and secretion of LTB4 is a target for the treatment of inflammation. Furthermore, the activity of lipoxygenase (LOX) enzymes (and, therefore, the synthesis of leukotrienes, including LTB4) is linked with COX enzyme activity. The inhibition of COX activity makes more arachidonic acid (AA) available for LOX catalysis, which may result in increased secretion of leukotrienes, including LTB4 (Calder 2020; Saini et al. 2020). It was, therefore, deemed relevant to assess the effects of the T. ferdinandiana extracts on LTB4 secretion in RAW 264.7 cells.

Only the aqueous leaf T. ferdinandiana extract had significant effects on LTB4 secretion in unstimulated RAW 264.7 macrophages (Fig. 5a). Although the aqueous leaf extract did significantly inhibit LTB4 secretion (p < 0.01), relatively low inhibitory activity (~ 8%) was noted. Furthermore, the turmeric control also had relatively minor effects on unstimulated RAW 264.7 cells, increasing LTB4 secretion by approximately 7%. It is noteworthy that the levels of LTB4 secreted in the untreated cells was low and near the detection threshold of this assay (results not shown), and therefore, these differences may not be indicative of inflammo-modulatory activity of the aqueous leaf extract and the turmeric control, and may instead be due to minor fluctuations in the low levels of this lipid. Further work is required to verify whether the trends noted herein demonstrate inflammo-modulation.

Fig. 5

The effects of the T. ferdinandiana fruit and leaf extracts on LTB4 levels in a unstimulated and b LPS-stimulated (100 ng/mL) RAW 264.7 cells. NC negative control (media); FM = methanolic fruit extract; FW = aqueous fruit extract; LM = methanolic leaf extract; LW = aqueous leaf extract; PC = turmeric positive control (1.25 mg/mL). Data is represented as mean values of triplicate results ± standard deviation. # and ## represent results that are significantly different to the negative control at p < 0.01 and p < 0.005, respectively

The LPS-stimulated RAW 264.7 macrophages secreted substantially higher amounts of LTB4 (results not shown), and therefore the effects of the T. ferdinandiana extracts were more pronounced (Fig. 5b). An interesting trend was evident: treatment with the methanolic extracts and the turmeric control significantly increased LTB4 secretion (p < 0.005), indicating pro-inflammatory effects via increased LTB4 production. Interestingly, these findings contrast to the effects of the extracts (and turmeric) against cytosolic COX-2 levels (Fig. 4b) and PGE2 secretion (Fig. 4d), where substantial decreases were evident. As both COX and LOX enzymes use AA as a substrate, decreased COX-2 synthesis (and the subsequent decreased synthesis of PGE2) may make more AA available to be converted to LTB4 by LOX, thereby accounting for the increased secretion of LTB4. The effects of the aqueous extracts were substantially different, with significant inhibition of LTB4 secretion (p < 0.005) noted for the aqueous fruit and leaf extracts, with decreases of approximately 27% and 25%, respectively, in comparison with the LPS-stimulated untreated control. Interestingly, the aqueous extracts were substantially less effective inhibitors of COX-2 levels and PGE2 secretion in LPS-stimulated RAW 264.7 cells (Fig. 4), and this may decrease the amount of AA available to the cells (compared to the methanolic extracts), thereby reducing the production and secretion of LTB4.

Effect of the extracts on cytosolic NF-κB levels in LPS-stimulated RAW 264.7 cells

The production and secretion of cytokines is regulated by the transcription factor NF-κB (Serasanambati and Chilakapati 2016; Tripathi and Aggarwal 2006). Similarly, COX-2 synthesis (and, therefore, PGE2 secretion) are controlled by NF-κB (Serasanambati and Chilakapati 2016; Zarghi and Arfaei 2011). Therefore, the inhibition of cytokine, chemokine, COX-2 and PGE2 by the extracts may result from inhibition of NF-κB function. In contrast, LOX transcription is independent of NF-κB, although several LOX products are known to regulate NF-κB production, and thus affect the levels of other inflammatory mediators (Lorenzetti et al. 2019; Tavares et al. 2014). Therefore, the effects of the T. ferdinandiana extracts on cytosolic NF-κB levels in RAW 264.7 macrophages was also investigated. Interestingly, all the extracts significantly inhibited the cytosolic levels of NF-κB in the unstimulated RAW 264.7 macrophages (Fig. 6a). Similarly, the turmeric control also decreased cytosolic NF-κB significantly in the unstimulated cells. Notably, the methanolic extracts were substantially stronger inhibitors of NF-κB production than for the corresponding aqueous extracts.

Fig. 6

The effects of the T. ferdinandiana fruit and leaf extracts on NF-κB levels in a unstimulated and b LPS-stimulated (100 ng/mL) RAW 264.7 cells. NC negative control (media); FM = methanolic fruit extract; FW = aqueous fruit extract; LM = methanolic leaf extract; LW = aqueous leaf extract; PC = turmeric positive control (1.25 mg/mL). Data is represented as mean values of triplicate results ± standard deviation. # and ## represent results that are significantly different to the negative control at p < 0.01 and p < 0.005, respectively

The methanolic fruit and leaf extracts also significantly inhibited cytosolic NF-κB levels in LPS-stimulated RAW 264.7 macrophages, with decreases of approximately 33% and 44%, respectively, compared to the LPS-stimulated control (Fig. 6b). This compares favourably to the decrease in cytosolic NF-κB induced by the turmeric positive control (~ 34%). Curcumin (the main bioactive compound in turmeric) potently inhibits NF-κB transcription pathways (Ghasemi et al. 2019), and thereby down-regulates multiple pro-inflammatory mediator levels. It is, therefore, likely that the anti-inflammatory properties of the T. ferdinandiana extracts may also be mediated via suppression of NF-κB transcription. In contrast, the aqueous extracts did not significantly down-regulate the cytosolic levels of NF-κB in RAW 264.7 macrophages.

Qualitative HPLC–MS/MS analysis of the methanolic T. ferdinandiana extracts

As the T. ferdinandiana methanolic fruit and leaf extracts displayed the greatest effects against the immune-modulatory and inflammatory mediators screened in this study, they were further examined by LC–MS analysis to highlight noteworthy components. Previous studies in our group developed analysis parameters to profile T. ferdinandiana extracts and identify noteworthy flavonoid and tannin components (Shalom and Cock 2018). The current study utilised a metabolomics fingerprint approach to detect these phytoconstituents, and to determine their relative abundances, which allows for a comparison between the constituents in the methanolic extracts. The presence of 12 previously highlighted compounds was verified in the methanolic fruit and leaf extracts (Table 2). The tannin components ellagic acid (and its dehydrate derivative), gallic acid, chebulic acid, corilagen and exifone were particularly abundant in both the fruit and leaf extracts. In addition, castalagin and chebulagic acid were abundant in the methanolic fruit extract, although they were detected in lower relative abundancies in the methanolic leaf extract, whilst punicalin and chebulinic acid were present in higher relative abundance in the leaf extract than in the fruit extract. The flavonoid rutin was abundant in both the fruit and leaf extracts.

Table 2 Qualitative HPLC–MS/MS analysis of the methanolic T. ferdinandiana fruit and leaf extracts in negative ionisation mode, elucidation of empirical formulas and identification of the compound