1.1. Botany

B. pilosa was first collected and named by Carl Linnaeus in 1753 [3]. Taxonomically, it is assigned to the Bidens genus (Asteraceae) as shown in Table 1. This genus is estimated to include 230 to 240 species worldwide [3, 4]. B. pilosa has several varieties such as B. pilosa var. radiata, var. minor, var. pilosa, and var. bisetosa. Alongside examination of morphological traits, authentication of B. pilosa can be aided by chemotaxonomy and molecular characterization [7].

B. pilosa is an erect, perennial herb widely distributed across temperate and tropical regions. B. pilosa is either glabrous or hairy, with green opposite leaves that are serrate, lobed, or dissected. It has white or yellow flowers, and long narrow ribbed black achenes (seeds). It grows to an average height of 60cm and a maximum of 150cm in favorable environments [8] (Figure 1). B. pilosa prefers full sun and moderately dry soil. However, it can grow in arid and barren land from low to high elevations. With the advantage of being fast-growing, in the 1970s, the Food and Agricultural Organization actively promoted the cultivation of B. pilosa in Africa [9]. B. pilosa propagates via seeds. A single plant can produce 3000–6000 seeds. Dry mature seeds from B. pilosa can be germinated in 3 to 4 days in moist soil or after being soaking in water. Seeds are viable for at least 3 years [10]. Minimal agricultural techniques are required for B. pilosa cultivation. Due to its invasive tendencies, B. pilosa is generally considered to be a weed [11].

Open in a separate window

Figure 1

B. pilosa (a) and its flowers (b) and achenes (c).

B. pilosa is thought to have originated in South America and subsequently spread all over the world [12]. Bidens species and their varieties bear vernacular names based on their characteristics. For example, Bidens species are known by such names as Spanish needles, beggar's ticks, devil's needles, cobbler's pegs, broom stick, pitchforks, and farmers' friends in English and some other languages because of their sticky achenes [8] and are sometimes known as xian feng cao (“all bountiful grass”) in Chinese because of their prosperous growth.

3.1. Anticancer Activity

Folkloric reports revealed the possible antitumor efficacy of B. pilosa, and several scientific in vitro studies have supported the claim that B. pilosa extracts and isolated compounds possess anti-cancer activities against a variety of cancer cells. Several studies have used bioassay guided isolation and fractionation methods to discover new compounds from B. pilosa. For example, Kviecinski and colleagues tested hydroalcoholic crude extracts, chloroform, ethyl acetate, and methanol fractions for anti-tumor activity [35]. The cytotoxicity of the extracts was assessed using brine shrimp, hemolytic, MTT, and neutral red uptake (NRU) assays. In vivo studies were performed using Ehrlich ascites carcinoma in isogenic BALB/c mice. Among them, the chloroform fraction was the most toxic with a half maximal inhibitory concentration (IC₅₀) of 97 ± 7.2 and 83 ± 5.2μg/mL in NRU and MTT, respectively [35]. Kumari and colleagues also reported the anti-cancer and anti-malarial activities of B. pilosa leaves [30]. Based on a cytotoxicity-directed fractionation strategy, they identified phenyl-1,3,5-heptatriene with IC₅₀ values of 8 ± 0.01, 0.49 ± 0.45, 0.7 ± 0.01, and 10 ± 0.01μg/mL against human oral, liver, colon, and breast cancer cell lines, respectively. However, phenyl-1,3,5-heptatriyne showed lower activity against breast cancer cell lines than the chloroform leaf extract which had an IC₅₀ value of 6.5 ± 0.01μg/mL. Moreover, the positive control, taxol, showed higher activity than phenyl-1,3,5-heptatriyne [30]. Furthermore, in vitro comet assays were performed to evaluate the toxicity of n-hexane, chloroform, and methanol extracts of B. pilosa and its ethyl acetate, acetone, and water fractions on Hela and KB cells. The ethyl acetate fraction from the methanol extract exhibited the highest activity with half maximal cytotoxic concentrations (CTC₅₀) of 965.2μg/mL and 586.2μg/mL against Hela and KB cells, respectively. Despite the moderate toxicity, these findings suggest that these B. pilosa extracts/fractions could be useful for future studies [36]. Hot water extracts of B. pilosa var. minor Sheriff were also assessed for its antileukemic effects on leukemic cell lines L1210, U937, K562, Raji, and P3HR1 using XTT-based colorimetric assays. The extract inhibited the five cell lines with IC₅₀ values ranging from 145μg/mL to 586μg/mL. L1210, K562, Raji, and P3HR1 were more sensitive to B. pilosa extract with IC₅₀ values below 200μg/mL [101].

Consistent with the antitumor activities of B. pilosa extracts and fractions, some of its phytochemicals also showed anticancer activity as outlined in Table 11. Among them, luteolin (103), a well-studied flavonoid with multiple bioactivities, was more effective against tumor cell proliferation than its derivatives with IC₅₀ values ranging from 3μM to 50μM in cells, and 5 to 10mg/kg in animals. Luteolin was also found to fight cancer as a food additive at concentrations of 50 to 200ppm [31] and prevent skin cancer [31] and cancer invasion [102]. Lee and colleagues reported that luteolin prevents cancer by inhibiting cell adhesion and invasion [102]. Significant inhibition concentration was reported to be 5μM and complete inhibition concentration was reported to be 40μM. Moreover, luteolin was reported to inhibit hepatocyte growth factor (HGF)-induced cell scattering as well as cytoskeleton changes such as filopodia and lamellipodia which was determined using phase-contrast and fluorescence microscopy. Furthermore, luteolin also inhibited the HGF-induced phosphorylation of c-Met, ERK 1/2, and Akt as well as the MAPK/ERK and P13K-Akt pathways [31, 102]. Other mechanisms underlying the anticancer activities of luteolin are the inhibition of topoisomerase I and II, which inhibits cell replication and DNA repair thus promoting apoptosis, regulation of PI-3-Kinase/Akt/MAPK/ERK/JNK, activation of apoptosis in the mitochondrial pathway by activating caspase 9 and caspase 3, which were found in malignant cells but not in normal human peripheral blood mononuclear cells, death receptor-induced apoptosis, and a cell-cycle arrest mechanism, inhibition of fatty acid synthase which is upregulated in many cancer cells, and sensitization to chemotherapy whereby luteolin increases the susceptibility of cancer cells to chemotherapy [31]. Butein (82) is another flavonoid that showed a cytotoxic effect on human colon adenocarcinoma cell proliferation with a reported IC₅₀ value of 1.75μM. Butein at 2μM affected the incorporation of [¹⁴C]-labeled leucine, thymidine, and uridine which can cause the inhibition of DNA, RNA, and protein synthesis of human colon cancer cells. Moreover, butein also exhibited noncompetitive inhibition of 1-chloro-2,4-dinitrobenzene (CDNB) in glutathione S-transferase (GST) activity. Tumor resistance was correlated with high levels of GST, thus, butein inhibited proliferation of cancer cells [33]. Another flavonoid present in B. pilosa, centaureidin (109), also showed anti-cancer activity in B lymphoma cells. Centaureidin, isolated from Polymnia fruticosa, inhibited tubulin polymerization in vitro and induced mitotic figure formation in CA46 Burkitt lymphoma cells. Using turbidimetric assay, the IC₅₀ value of centaureidin in inhibition of mitosis was 3μM [103]. Cytotoxicity of centaureidin was further analyzed using American National Cancer Institute (NCI) 60 human tumor cell lines. The cytotoxicity potency of centaureidin, expressed as GI₅₀ (50% growth inhibition in the NCI tumor line panel), was 0.24μM [34]. These data mark centaureidin as a promising antimitotic agent for tumor therapy.

In addition to anti-tumor flavones, polyynes found in B. pilosa have also been shown to possess anti-tumor properties. Based on a bioactivity-directed isolation approach, Wu and colleagues identified two polyyne aglycones from the ethyl acetate fraction of B. pilosa [70]. 1,2-Dihydroxytrideca-5,7,9,11-tetrayne (48) and 1,3-Dihydroxy-6(E)-tetradecene-8,10,12-triyne (46) exhibited significant anticell proliferation activity in primary human umbilical vein endothelium cells (HUVEC) with IC₅₀ values of 12.5μM and 1.73μM, respectively. They also decreased angiogenesis and promoted apoptosis in human endothelial cells. Their anti-angiogenic and cytotoxic effects correlated with activation of the CDK inhibitors and caspase-7 [70]. In addition, 1,2-Dihyroxy-5(E)-tridecene-7,9,11-triyne (45) showed antiangiogenic effects in HUVECs with an IC₅₀ value of 12.4μM as evidenced by a decrease in the tube formation and migration of HUVECs [28]. The IC₅₀ value of compound 45 in the inhibition of basic fibroblast growth factor-induced HUVEC growth was 28.2μM. However, it had higher IC₅₀ values than those for lung carcinoma cells and keratinocytes. This compound could also inhibit cell proliferation of HUVECs, lung carcinoma A549 cells and HACAT keratinocytes. The mechanism by which compound 45 inhibits HUVEC growth and angiogenesis is complicated and includes decreasing the expression of cell cycle regulators (CDK4, cyclins D1 and A, retinoblastoma (Rb), and vascular endothelial growth factor receptor 1), caspase-mediated activation of CDK inhibitors p21 (Cip1) and p27 (Kip), upregulation of Fas ligand expression, downregulation of Bcl-2 expression, and activation of caspase-7 and poly (ADP-ribose) polymerase [28].

3.2. Anti-Inflammatory Activity

B. pilosa is commonly used to treat inflammatory disorders. The anti-inflammatory phytochemicals present in B. pilosa are listed in Table 11. Cyclooxygenase-2 (COX-2) is a physiologically important enzyme that converts arachidonic acid to prostaglandin (PGE₂). Its expression is induced by a wide variety of external stimuli indicating its involvement in inflammatory diseases, and it is used as an inflammatory marker [42]. Yoshida and colleagues studied the effects of the aqueous extracts of B. pilosa aerial parts in the production of COX-2 and PGE₂ as well as on the activation of mitogen activated protein kinases (MAPKs) in normal human dermal fibroblasts (HDFs) in response to inflammatory cytokine, IL-1β. This work showed that IL-1β activated MAPKs such as ERK1/2, p38, and JNK to different extents and induced COX-2 expression. The COX-2 expression in HDFs was regulated mainly by p38 following IL-1β stimulation. Consistently, the p38 inhibitor SB203580 blocked this expression. Using this cell platform, B. pilosa extracts were tested for inhibition of inflammation. The extract dose-dependently suppressed the activation of p38 and JNK and moderately suppressed ERK1/2, as well as suppressing COX-2 expression and PGE₂ production [42]. This work supports the use of B. pilosa as an anti-inflammatory agent; however, no compounds responsible for the anti-inflammatory activity of B. pilosa were identified.

A further study also reported the anti-inflammatory activity as well as the antiallergic activity of B. pilosa [37]. In this study, dried powder of the aerial part of B. pilosa, which had been pretreated with the enzyme cellulosine, was used for further tests. The results showed that oral administration of the cellulosine-treated B. pilosa lowered the level of serum IgE in mice 10 days after immunization with DNP (2,4-dintrophenyl)-Ascaris as an antigen. This treatment also reduced dye exudation in skin induced by passive cutaneous anaphylaxis and production of inflammatory mediators, histamine, and substance P in rats [37]. Phytochemical analysis showed that cellulosine treatment increased the percentage of caffeic acid and flavonoids. This study suggests that B. pilosa and its phenolics have anti-inflammatory functions.

Phenolics and polyynes are major anti-inflammatory phytochemicals present in B. pilosa (Table 11). Unsurprisingly, phenolics such as luteolin (103) and ethyl caffeate (161) that are major constituents of B. pilosa have also been reported to possess anti-inflammatory activity. Luteolin was reported to exhibit anti-inflammatory activity in macrophages. Xagorari and colleagues showed that luteolin inhibited the release of inflammatory cytokines, TNF-α and interleukin-6, in RAW 264.7 cells following LPS stimulation [41]. It inhibited TNF-α production with an IC₅₀ value of 1μM. The underlying anti-inflammatory mechanism of luteolin was reported to be the inactivation of Akt and NF-κB activation [41]. In addition, luteolin was reported to confer anti-inflammatory activity through inhibition of LPS-stimulated iNOS expression in BV-2 microglial cells. It inhibited LPS-activated microglia in a dose-dependent manner with an IC₅₀ value of 6.9μM. Moreover, immunoblot and RT-PCR data proved that luteolin suppressed IκB-α degradation and iNOS expression in LPS-activated microglia [39]. Kim and colleagues stated that luteolin may have beneficial effects on inflammatory neural diseases through inhibition of iNOS expression [39]. A related study revealed that luteolin decreased the transcriptional activity of NF-κB RelA via partial inhibition of TNF-mediated NF-κB DNA binding activity. Luteolin also inhibited Akt phosphorylation and induced degradation of a transcription factor, interferon regulatory factor (IRF) [40].

Chiang and colleagues showed that ethyl caffeate (161) significantly inhibited NO production in mouse macrophages, RAW 264.7 cells [38]. Based on MTT assays, they concluded that this inhibition was not due to the cytotoxicity of ethyl caffeate. The IC₅₀ value of ethyl caffeate in the inhibition of NO production was 5.5μg/mL, slightly lower than curcumin (positive control) which has an IC₅₀ value of 6.5μg/mL. They demonstrated that ethyl caffeate exerted anti-inflammatory activity via the reduced transcription and translation of iNOS (inducible nitric oxide synthase) in RAW 246.7 cells. In addition, this compound also suppressed COX-2 expression in RAW 246.7 cells and MCF-7 cells. The in vivo anti-inflammatory effect of ethyl caffeate was verified by testing in TPA-treated mouse skin. Like celecoxib, the positive control, ethyl caffeate significantly abolished COX-2 expression in a dose-dependent manner. Ethyl caffeate at 1mg/200μL/site (24mM) inhibited COX-2 expression at a level comparable to celecoxib at 1mg/200μL/site (13mM). Remarkably, ethyl caffeate at 48mM (2mg/200μL/site) was more effective than celecoxib at 131mM (10mg/200μL/site). Moreover, this compound inhibited the activation of nuclear factor-κB (NF-κB) by LPS via the prevention of NF-κB binding to DNA [38]. In addition, 3 ethyl caffeate analogs (ethyl 3,4-dihydroxyhydrocinnamate, ethyl cinnamate, and catechol) also showed different degrees of NF-κB binding to DNA as listed in Table 12. Ethyl cinnamate lacks a catechol moiety which results in ineffective inhibition of NF-κB binding to DNA [38].

Pereira and colleagues assessed the anti-inflammatory and immunomodulatory activities of B. pilosa methanol extract as well as one polyyne, 2-O-β-glucosyltrideca-11(E)-en-3,5,7,9-tetrayn-1,2-diol (54) in T lymphocytes and a zymosan-induced arthritis mouse model [18]. They first examined the in vitro effect of the B. pilosa extract and compound 54 on cell proliferation of human T cells stimulated with 5μg/mL phytohemagglutinin (PHA) or 100nM 12-O-tetradecanoyl phorbol-13-acetate (TPA) plus 15μM ionomycin and on cell proliferation of mouse T cells stimulated with 5μg/mL concanavalin A (Con A). The data demonstrated that both methanol extract and compound 54 suppressed T-cell proliferation in a dose-dependent manner. The estimated IC₅₀ values of the B. pilosa extract against human T cells stimulated with 5μg/mL PHA and 100nM TPA plus 15μM ionomycin were 12.5 and 25μg/mL, respectively. In comparison with the methanol extract, compound 54 showed 10-fold more inhibition of human T-cell proliferation with an estimated IC₅₀ value of 1.5μg/mL. Accordingly, the B. pilosa extract and compound 54 dose-dependently suppressed mouse T-cell proliferation with estimated IC₅₀ values of 30 and 2.5μg/mL, respectively. Taken together, the data indicate that the B. pilosa extract and compound 54 act on human and mouse T cells. To test the in vivo effect of the B. pilosa extract and compound 54, a zymosan-induced arthritis mouse model was used. This model was established from B10.A/SgSnJ mice with an injection of zymosan (0.15mg). The zymosan-injected mice received an intraperitoneal injection of the B. pilosa extract (1, 5, or 10mg) at one dose a day for 5 days. Popliteal lymph node (PLN) weight was monitored to check the development of arthritis. The results revealed that 10mg of the methanol extract of B. pilosa extract could significantly diminish inflammation as evidenced by PLN weight [18]. This work suggests that B. pilosa (and compound 54) can suppress immune response and inflammation.

3.3. Antidiabetic Activity

Anti-diabetic agents are primarily developed from plants and other natural resources [7, 43–46].B. pilosa is one of 1,200 plant species that have been investigated for antidiabetic activity [119, 120]. B. pilosa is used as an anti-diabetic herb in America, Africa, and Asia [45, 119, 121]. Many studies have indicated that B. pilosa could treat type 1 diabetes (T1D) and type 2 diabetes (T2D) in animals.

Etiologically speaking, T1D is caused by autoimmune-mediated destruction of pancreatic β cells, leading to insulin deficiency, hyperglycemia, and complications. Currently, there is no cure for T1D. Polarization of Th cell differentiation controls the development of T1D. Suppression of Th1 cell differentiation and promotion of Th2 cell differentiation ameliorate T1D [122]. One study showed that the butanol fraction of B. pilosa inhibited T-cell proliferation, decreased Th1 cells and cytokines, and increased Th2 cells and cytokines, leading to prevention of T1D in nonobese diabetic (NOD) mice [44]. Based on a bioactivity-directed isolation strategy, 3 polyynes, 2-β-D-Glucopyranosyloxy-1-hydroxytrideca-5,7,9,11-tetrayne (53), also known as cytopiloyne, 3-β-D-Glucopyranosyl-1-hydroxy-6(E)-tetradecene-8,10,12-triyne (69), 2-β-D-Glucopyranosyloxy-1-hydroxy-5(E)-tridecene-7,9,11-triyne (50) were identified from B. pilosa [44, 46]. The IC₅₀ value of the butanol fraction was 200μg/mL. This inhibition was reported to be partially attributed to cytotoxicity because the butanol fraction at 180μg/mL could cause 50% death of Th1 cells. Moreover, this study suggested that the butanol fraction may prevent diabetes in NOD mice in vivo via downregulation of Th1 cells or upregulation Th2 cells that have effects which are antagonistic of those of Th1 cells [44]. This was proven by intraperitoneal injection of the butanol fraction at a dose of 3mg/kg BW, 3 times a week, to NOD mice from 4 to 27 weeks. This dosage resulted in lower incidence of diabetes (33%). At a dose of 10mg/kg, the butanol fraction of B. pilosa totally eliminated (0%) the initiation of the disease. To further support this result, assessment of IgG2a and IgE production was performed in the serum of NOD mice. As in vivo results obtained from intracellular cytokine staining, experiments were not very conclusive levels of IgG2a and IgE were measured since Th1 cytokine IFNγ; and Th2 cytokine IL-4 favor the production of IgG2a and IgE, respectively. As expected, high levels of IgE and some decline in the levels of IgG2a were observed in the serum. Profiling of the butanol extract revealed five compounds, 3-β-D-Glucopyranosyl-1-hydroxy-6(E)-tetradecene-8,10,12-triyne (69), 2-β-D-Glucopyranosyloxy-1-hydroxy-5(E)-tridecene-7,9,11-triyne (50), 4,5-Di-O-caffeoylquinic acid, 3,5-Di-O-caffeoylquinic acid, and 3,4-Di-O-caffeoylquinic acid. Only the first two compounds showed similar effects on the prevention of diabetes in NOD mice as the B. pilosa butanol fraction. Moreover, compound 50 showed greater activity than compound 69 in terms of enhancement (by 34% compared to 8%) of differentiation of Th0 to Th2 at 15μg/mL (both compounds) and inhibition (by 40% compared to 10%) of differentiation to Th1 at the same concentration [44].

Among the three polyynes found in B. pilosa, cytopiloyne (53) had the most potent anti-T1D activity [46]. To test the in vivo effect of cytopiloyne, NOD mice received intraperitoneal or intramuscular injection of cytopiloyne at 25μg/kg BW, 3 times per week. Twelve-week-old NOD mice started to develop T1D, and 70% of NOD mice aged 23 weeks and over developed T1D. Remarkably, 12- to 30-week-old NOD mice treated with cytopiloyne showed normal levels of blood glucose (<200mg/dL) and insulin (1-2ng/mL). Consistent with T1D incidence, cytopiloyne delayed and reduced the invasion of CD4⁺ T cells into the pancreatic islets [46].

In vitro study showed that cytopiloyne (53) inhibited the differentiation of naïve Th (Th0) cells (i.e., CD4⁺ T cells) into Th1 cells and promoted differentiation of Th0 cells into Th2 cells [50]. The in vitro data are consistent with the in vivo results indicating that cytopiloyne reduced Th1 differentiation and increased Th2 differentiation as shown by intracellular cytokine staining and FACS analysis [46]. In line with the skewing of Th differentiation, the level of serum IFN-γ and IgG2c decreased while that of serum IL-4 and serum IgE increased compared to the negative controls (PBS-treated mice). Cytopiloyne also enhanced the expression of GATA-3, a master gene for Th2 cell differentiation, but not the expression of T-bet, a master gene for Th1 cell differentiation, further supporting its role in skewing Th differentiation [46].

Also importantly, cytopiloyne partially depleted CD4⁺ rather than CD8⁺ T cells in NOD mice [46]. As shown in Table 13, coculture assays showed that the depletion of CD4⁺ T cells was mediated through the induction of Fas ligand expression on pancreatic islet cells by cytopiloyne, leading to apoptosis of infiltrating CD4⁺ T cells in the pancreas via the Fas and Fas ligand pathways. However, cytopiloyne did not induce the expression of TNF-α in pancreatic islet cells and, thus, had no effect on CD8⁺ T cells [46].

In addition, Chang and colleagues showed that cytopiloyne dose-dependently inhibited T-cell proliferation stimulated by IL-2 plus Con A or anti-CD3 antibody, using [³H] thymidine incorporation assay [46].

Overall, the mechanism of action of cytopiloyne and, probably, its derivatives in T1D includes inhibition of T-cell proliferation, skewing of Th cell differentiation, and partial depletion of Th cells. Due to the anti-diabetic mechanisms of action, it was hypothesized that cytopiloyne protects NOD mice from diabetes by a generalized suppression of adaptive immunity. To evaluate this hypothesis, ovalbumin (Ova) was used as a T-cell dependent antigen to prime NOD mice, which had already received cytopiloyne or PBS vehicle. Ova priming boosted similar anti-Ova titers in cytopiloyne-treated mice and PBS-treated mice, but a difference in immunoglobulin isotype was observed in the two groups. Thus, it was concluded that cytopiloyne is an immunomodulatory compound rather than an immunosuppressive compound [46, 50].

T2D is a chronic metabolic disease with serious complications resulting from defects in either insulin secretion, insulin action, or both [123]. A study by Ubillas et al. showed that the aqueous ethanol extract of the aerial part of B. pilosa at 1g/kg body weight (BW) lowered blood glucose in db/db mice, a T2D mouse model [45]. Based on a bioactivity-guided identification, compounds 69 and 50 were identified. Further, the mixture of the compounds (69:50) in a 2:3 ratio significantly decreased blood glucose concentration and reduced food intake on the second day of treatment when administered at doses of 250mg/kg twice a day to C5BL/Ks-db/db mice. When tested at 500mg/kg, a more substantial drop in blood glucose level as well as the stronger anorexic effect (food intake reduced from 5.8g/mouse/day to 2.5g/mouse/day) was observed [45]. In this study, it was suggested that the blood glucose lowering effect of B. pilosa was caused, in part, by the hunger suppressing effect of its polyynes [45]. However, the hunger suppressing effect of the ethanol extract of B. pilosa was not found in the studies described below. In another study [43], water extracts of B. pilosa (BPWE) were used in diabetic db/db mice, aged 6–8 weeks, with postprandial blood glucose levels of 350 to 400mg/dL. Like oral anti-diabetic glimepiride, which stimulates insulin release, one single dose of BPWE reduced blood glucose levels from 374 to 144mg/dL. The antihyperglycemic effect of BPWE was inversely correlated to an increase in serum insulin levels, suggesting that BPWE acts to lower blood glucose via increased insulin production. However, BPWE had different insulin secretion kinetics to glimepiride [43]. One flaw in current anti-diabetics is their decreasing efficacy over time. The authors investigated the long term anti-diabetic effect of BPWE in db/db mice. BPWE reduced blood glucose, increased blood insulin, improved glucose tolerance, and reduced the percentage of glycosylated hemoglobin (HbA1c). Both long-term and one-time experiments strongly support the anti-diabetic action of BPWE [43]. In sharp contrast to glimepiride, BPWE protected against islet atrophy in mouse pancreas. The investigators further evaluated anti-diabetic properties of 3 B. pilosa varieties, B. pilosa L. var. radiate (BPR), B.pilosa L. var. pilosa (BPP), and B. pilosa L. var. minor (BPM) in db/db mice [7]. One single oral dose (10, 50 and 250mg/kg body weight) of BPR, BPP, or BPM crude extracts decreased postprandial blood glucose levels in db/db mice for up to four hours, and the reduction of glucose levels in the blood appeared to be dose-dependent. Comparing the three variants, BPR extract resulted in a higher reduction in blood glucose levels when administered at the same dose as the other two varieties. In terms of serum insulin levels, a dose of 50mg/kg of each extract was used, and the BPR extract, together with the three polyynes, significantly increased the serum insulin level in db/db mice. Long-term experiments (28-day treatment) were then conducted using diabetic mice with postprandial glucose levels from 370 to 420mg/dL, and glimepiride was used as positive control. The range of dosages applied was from 10mg/kg BW to 250mg/kg BW. Results showed that the positive control as well as the crude extracts of the three varieties lowered the blood glucose levels in db/db mice. However, only BPR extract, containing a higher percentage of cytopiloyne (53), reduced blood glucose levels and augmented blood insulin levels more than BPP and BPM. The percentage of glycosylated hemoglobin A1c (HbA1c) was also measured and found to be 7.9% ± 0.5% in mice aged 10–12 weeks, and 6.6% ± 0.2%, 6.1% ± 0.3% and 6.2% ± 0.3% in the blood of age-matched mice following treatment with BPR crude extract (50mg/kg), glimepiride (1mg/kg), and compound 53 (0.5mg/kg), respectively [7]. Among the polyynes found in B. pilosa, cytopiloyne was the most effective against T2D. Hence, cytopiloyne was used for further study on anti-diabetic action and mechanism [124]. The data confirmed that cytopiloyne reduced postprandial blood glucose levels, increased blood insulin, improved glucose tolerance, suppressed HbA1c level, and protected pancreatic islets in db/db mice. Nevertheless, cytopiloyne failed to decrease blood glucose in streptoztocin (STZ)-treated mice whose b cells were already destroyed. Additionally, cytopiloyne dose-dependently increased insulin secretion and expression in b cells as well as calcium influx, diacylglycerol, and activation of protein kinase Cα. Collectively, the mechanistic studies suggest that cytopiloyne treats T2D via regulation of insulin production involving the calcium/DAG/PKCα cascade in b cells.

The studies detailed above point to the conclusion that cytopiloyne and related polyynes (compounds 69 and 49) are anti-diabetics in animal models. The data uncover a new biological action of polyynes. It should be noted that, like all anti-diabetic drugs, cytopiloyne failed to prevent or cure diabetes completely but reduced diabetic complications [124]. Intriguingly, 34 polyynes have been found in B. pilosa so far. It remains to be seen whether all the polyynes present in this plant have anti-diabetic activities.

3.4. Antioxidant Activity

Free radicals can damage cellular components via a series of chemical reactions [48] leading to development and progression of cardiovascular disease, cancer, neurodegenerative diseases and ageing [47]. Free radicals, nitric oxide (NO), and superoxide anions can be produced in macrophages to kill microbes. However, an excessive generation of the free radicals under pathological conditions is associated with a wide range of illnesses. Plants are known to be rich in antioxidant phytochemicals. Chiang and colleagues evaluated the free radical scavenging activity of crude extract, fractions, and compounds of B. pilosa using 1,1-diphenyl-2-picrylhydrazyl (DPPH) and hypoxanthine/xanthine oxidase assays [47]. Using DPPH and hypoxanthine/xanthine oxidase assays, they found that the B. pilosa crude extract and the ethyl acetate, butanol, and water fractions had free radical scavenging activity. Nine compounds, Heptyl-2-O-β-xylofuranosyl-(1→6)-β-glucopyranoside (199), 3-O-Rabinobioside (124), Quercetin 3-O-rutinoside (130), Chlorogenic acid (167), 3,4-Di-O-caffeoylquinic acid (169), 3,5-Di-O-caffeoylquinic acid (170), 4,5-Di-O-caffeoylquinic acid (171), Jacein (119), and Centaurein (110) had DPPH radical scavenging activity [47]. The IC₅₀ values of the B. pilosa crude extract/fractions and compounds are summarized in Tables ​Tables1414 and ​and15,15, respectively.

Measurement of free radical scavenging activities is one way of assessing the antioxidant activities of B. pilosa and its fractions and compounds. It is interesting that the ethyl acetate and butanol fractions are more active than the water fraction and B. pilosa crude extract [47]. Of the secondary metabolites, only phenolic compounds 124, 130, 167, 169, and 171 showed significant DPPH-radical scavenging activities. Further analysis of the structure-activity relationship of the compounds suggested that substitution of the C3 hydroxyl group with glycosides increased the activity approximately 2-fold (for example, in compounds 124 and 130) relative to quercetin (–OH in its C3) [47]. Further modification of the structures of the active compounds needs to be performed to test the effects of various substituents on activity. The reason why most of the antioxidant compounds contain phenol moieties in their structure could be that the reduction-oxidation (redox) properties of phenols allow them to act as reducing agents, singlet oxygen quenchers, and hydrogen donors.

A complementary study by Muchuweti and colleagues determined phenolic content, antioxidant activity, and the phenolic profile of B. pilosa methanol extract [48]. They estimated that the phenolic content of the methanol extract of B. pilosa was 1102.8 ± 2.2mg/g [48]. Vanillin, hydroxybenzaldehyde, caffeic acid, coumaric acid, and ferulic acid were found in this extract. The B. pilosa extract also showed DPPH radical scavenging activity. Furthermore, the antioxidant activity of the flavonoids found in B. pilosa was correlated with its hepatoprotective effects through their inhibition of NF-κB activation which may lessen the oxidative stress caused by the production of free radicals during liver injury [81]. This activity might also be due to the anti-inflammatory effects of the aqueous extracts of B. pilosa aerial parts on the inhibition of COX-2 and PGE₂ production [42].

Essential oils from B. pilosa flowers and leaves are also reported to possess antioxidant activity. With the aim of replacing chemically synthesized additives, Deba and colleagues [53] worked on the antioxidant, antibacterial, and antifungal activities of essential oils and water extracts of B. pilosa's leaves and flowers. Table 16 summarizes the results obtained from DPPH free radical scavenging assay.

It can be inferred from Table 16 that essential oils from the leaves possessed the highest activity. It is reported elsewhere that monoterpenes present in essential oils such that of B. pilosa have protective effects and antioxidant properties [53]. Beta-carotene bleaching method was also performed. Leaves essential oils and aqueous extracts of the leaves and flowers showed higher activity than the flower essential oils. This is due to the volatility of the flower essential oils. The activity exhibited by the aqueous extracts was accounted to the presence of phenolic compounds that are reported to donate a hydrogen atom to free radicals such that the propagation of the chain reaction during lipid oxidation is terminated [53]. Overall, essential oils and phenolics present in B. pilosa can be thought of as major antioxidant compounds.

3.5. Immunomodulatory Activity

B. pilosa is thought to be an immunomodulatory plant and is reported to be effective in the treatment of immune disorders such as allergy [37], arthritis [37], and T1D [46, 50, 73].

As pointed out in the discussion of its anti-diabetic activities (Section 3.3), a combination of phytochemicals method and T-cell activation assays was used to study immunomodulatory properties of B. pilosa. IFN-γ is a key cytokine released by T and NK cells that mediates immune cells and sustains immunity against pathogens. Defects in IFN-γ expression, regulation, and activation result in vulnerability to diseases caused by bacteria and viruses [29]. An elegant study, performed by Chang and colleagues using IFN-γ promoter-driven luciferase reporter construct in Jurkat T cells, showed that hot water crude extracts of B. pilosa increased IFN-γ promoter activity two-fold [29]. Out of the subfractions of this extract, the butanol fraction, but not the water or ethyl acetate, fractions, increased IFN-γ promoter activity six-fold. Centaurein (110) and centaureidin (109) were identified from the butanol fraction and were stated to cause a four-fold increase in IFN-γ promoter activity with EC₅₀ values of 75μg/mL and 0.9μg/mL, respectively. The mechanism of action of centaurein was determined using transcription factors such as AP-1, NFAT, and NFκB which are reported to bind to IFN-γ promoter and regulate IFN-γ transcription. Unlike with the activity of the positive control, PHA, centaurein caused a four-fold increase in NFAT, a 3-fold increase in NFκB and had little, if any, effect on AP-1 enhancer activities (23-fold, three-fold, and ten-fold increases, respectively, were seen with PHA) [29]. The authors concluded that centaurein modulates IFN-γ expression by the NFAT and NFκB pathways. The article only determined the mechanism of action of centaurein. Its aglycone centaureidin may act through the same mechanism though this conclusion needs to be further verified.

B. pilosa extract and its compounds are reported to inhibit differentiation of naïve CD4⁺ helper T (Th0) cells into Th1 cells [44]. Using the Th cell differentiation assay as a screening platform, 3 polyynes, 2-β-D-glucopyranosyloxy-1-hydroxytrideca-5,7,9,11-tetrayne (53), 3-β-D-glucopyranosyl-1-hydroxy-6(E)-tetradecene-8,10,12-triyne (69), and 2-β-D-glucopyranosyloxy-1-hydroxy-5(E)-tridecene-7,9,11-triyne (49) were discovered from B. pilosa [44, 46]. The data shows that cytopiloyne and other two polyynes suppressed the differentiation of type 1 helper T (Th1) cells and production of Th1 cytokines and promoted that of type 2 helper T (Th2) cells and production of Th2 cytokines, thus explaining the immunomodulatory and anti-inflammatory effects of B. pilosa and its polyynes.

Chang and colleagues were the first to report the effect of the butanol extract of B. pilosa on the autoimmune diabetes and airway inflammation in mice [73]. Imbalance in the levels of Th1 and Th2 and of various cytokines leads to autoimmune diseases. T1D and other autoimmune diseases (rheumatoid arthritis, Crohn's disease, among others) are exacerbated by an increase Th1 levels (specifically, CD4⁺ Th1 cells) while Th2 cells antagonize this effect [56]. Moreover, Th2 cells mediate asthma in ovalbumin-induced hypersensitivity in BALB/c mice [73]. In their work [73], Chang and colleagues showed that 10mg/kg butanol extracts with 1.5% (w/w) compound 69 and 1.1% (w/w) compound 49 (Table 11) ameliorated the development of Th1-mediated diabetes in NOD mice through inhibition of β cell death and leukocyte infiltration. At the same dosage, the butanol extracts also exacerbated ovalbumin-induced pulmonary inflammation in BALB/c mice with an increase in the infiltration of eosinophils and mast cells into the airway of the mice [73]. Despite the different outcomes, both mouse models proved the concept that control over the Th1/Th2 shift is associated with autoimmune diseases and that the B. pilosa butanol extract can shift the differentiation of Th0 cells to Th2 cells [44, 73].

An extended study presented by Chiang and colleagues showed that compound 53 modulates T-cell functions [50]. Using CD4⁺ T cells from BALB/c mice, they demonstrated that cytopiloyne decreased levels of IFN-γ producing cells (Th1) by 12.2% (from 72% to 59.8%). Since Th1 and Th2 cell differentiation is antagonistic, it was expected that compound 53 increased the percentage of mouse IL-4 producing cells (Th2) by 7.2% (from 23.7% to 30.9%). Subsequent assessment of effect of compound 53 on the modulation of the transcription of IL-4 and IFN-γ showed that cytopiloyne, as expected, decreased the splenocyte levels of IFN-γ mRNA and increasing that of IL-4 in a dose-dependent manner. In the range of 0.1 to 3μg/mL of compound 53, the effects were not attributed to its cytotoxicity. Consequently, using 3μg/mL cytopiloyne for 72 and 96 hours, the protein concentration of IFN-γ decreased to 18.6% and 44.4%, respectively. Under the same conditions, cytopiloyne increased IL-4 concentrations to 198.5% and 247.0% (for 72 and 96 hours, resp.). This modulation of T-cell differentiation exhibited by compound 53 was used to explain its anti-diabetic activity [50].

The anti-diabetic role of cytopiloyne was extensively discussed above (Section 3.3, anti-diabetic activity). The molecular basis of the regulation of cytokine expression by cytopiloyne has been described. Cytopiloyne directly elevated the expression level of IL-4 via GATA-3 upregulation in T cells [46]. However, reduction of IFN-γ expression in T cells seemed to come from the indirect opposing effect IL-4 cytokine because the expression level of T-bet was unaltered [46]. In this way, cytopiloyne skewed Th1 polarization into the Th2 state, conferring protection against T1D in NOD mice. Aside from polarization of Th cell differentiation, cytopiloyne also activated the expression of Fas ligand in pancreatic b cells, this increase leading to the partial depletion of T cells and reduction of immune response in local areas such as the pancreas. Of note, cytopiloyne also inhibited T-cell proliferation and activation. By targeting T cells from three immunomodulatory actions, cytopiloyne protects against T1D and probably other Th1-mediated autoimmune diseases [56].

The phytochemical constituents of B. pilosa exert their functions on different immune cells to modulate immune response. It is possible that some of the compounds may have agonistic or antagonistic effects on immune response. Immune function of B. pilosa may depend on its composition and amount of compounds, which could explain the apparently conflicting report of B. pilosa butanol extract aggravating allergy in mice [73] while cellulosine-treated extract ameliorated allergy [37, 73]. IFN-γ promoter reporter assays and T-cell differentiation assays were used to isolate 2 flavonoids [29] and 3 polyynes [44, 50] as immunomodulatory compounds from the butanol fraction of B. pilosa. Interestingly, the flavonoids promote IFN-γ expression in NK and T cells. In marked contrast, the polyynes promote IL-4 expression and indirectly inhibit IFN-γ expression in differentiating T cells. Sensitivity appears to be the key to identifying structure- and bioactivity-related phytochemicals from medicinal plants.

3.6. Antimalarial Activity

The use of chemical drugs against pathogens has resulted in drug-resistant mutants. Examples of drug resistance can be found in the species of the Plasmodium that cause malaria. It is important to search for new compounds to combat Plasmodium parasites [51]. A study of the anti-malarial activity of the leaf extracts of B. pilosa using a combination of phytochemistry and bioassays showed that compound 49 (Table 11) showed activity against a malaria parasite (P. falciparum NF54 strain) with an IC₅₀ value of 6.0μg/mL [30]. In addition, compound 49, isolated from the aerial parts of B. pilosa, inhibited growth of the P. falciparum FCR-3 strain with an IC₅₀ value of 0.35μg/mL. This compound was tested for its in vivo effect in mice infected with P. berghei NK-65 strain. Results showed that compound 49 decreased the average parasitemia in the red blood cells by 20.7 (from 32.8% of that of the control to 12.1%) after an intravenous injection of 0.8mg/kg BW/day for four days [51]. Further studies addressing the anti-malarial mechanism underlying both polyynes and clinical studies are needed.

3.7. Antibacterial Activity

Emergence of multiple antibiotic-resistant microbes is becoming a global threat to public health and a challenge to disease treatment. For instance, penicillin is commonly used to combat a food-borne intracellular bacterium Listeria; however, penicillin-resistant bacteria have been discovered recently [29]. Chang and coworkers isolated centaurein (110) and centaureidin (109) from B. pilosa extract [29, 47]. Centaurein enhances expression of IFN-γ, a key cytokine for macrophage activation and, consequently, enhances bactericidal activity in macrophages [29, 47]. In agreement with observed in vitro effects, centaurein was reported to prevent and treat Listeria infection in C57BL/6J mice [52]. Mechanistic studies confirmed that centaurein exerted antilisterial action via IFN-γ expression in wild-type mice but not IFN-γ knockout mice [52]. Further in vitro studies on centaurein showed that this compound increased IFN-γ expression by 13% (from 17% to 20%), 20% (from 21% to 41%), and 11% (from 6% to 17%) in CD4⁺ T cells, CD8⁺T cells, and NK cells, respectively. That is to say, there was an increase in IFN-γ producing immune cells. As expected, centaurein also enhanced the expression level of T-bet, a key nuclear factor for IFN-γ expression. Consistently, centaurein augmented the serum IFN-γ levels in C57BL/6J mice, and this augmentation peaked 24 hours after compound injection. The quantity of mouse serum IFN-γ was sufficient to activate macrophages in vitro and eradicated GFP-producing Listeria inside macrophages. However, the entry of Listeria into macrophages was not affected by centaurein-treated mouse sera. Centaurein treatment at 20μg per mouse rescued 30% of the mice infected with a lethal dose of Listeria (2 × 10⁶CFU). It is noteworthy that in the presence of ampicillin (5μg/mouse), centaurein (20μg/mouse) rescued 70% of the mice, suggesting an additive effect between ampicillin and centaurein [52]. Despite lower abundance, centaureidin was 30 times more active than centaurein in terms of IFN-γ production [52].

Aside from the indirect antibacterial action mentioned above, extract and/or compounds of B. pilosa also showed direct bacteriostatic and/or bactericidal action. One study reported that essential oils and leaf/flower extracts of B. pilosa could suppress the growth of gram positive and gram negative bacteria as evidenced by zone of inhibition assays. In this study, antibacterial activity of the essential oils from the leaves and flowers of B. pilosa was determined in an attempt to identify natural products as food preservatives for prevention of microbial multiplication and food oxidation. The essential oils and extracts of B. pilosa leaves and flowers showed moderate but different extents of antibacterial activity (Table 17). In general, essential oils had higher antibacterial activity than crude extracts. One explanation for this could be that monoterpenes in the essential oils destroy cellular integrity and, subsequently, inhibit the respiration and ion transport processes. The presence of antibacterial β-caryophyllene could be another explanation as reported elsewhere [53]. Another study reported that the methanol and acetone extract of B. pilosa roots displayed antibacterial activities against the bacteria listed in Table 18 [54] and methanol extracts from the roots seemed to be the most effective.

Another study indicated that the polyyne, (R)-1,2-dihydroxytrideca-3,5,7,9,11-pentayne (49), from this plant also suppressed bacterial growth as shown by the minimum inhibitory concentration required to inhibit 50% bacterial growth (MIC₅₀) in Table 19. This compound was highly effective against several Gram positive and Gram negative bacteria including the drug-resistant bacteria Staphylococcus aureus N315 (MRSA) and Enterococcus faecalis NCTC12201 (VRE) [51]. Strikingly, compound 49 had a similar MIC₅₀ value to antibiotics (ampicillin, tetracycline, norfloxacin, and amphotericin B) in most of the bacteria tested.

Antibacterial activity of B. pilosa extracts and components, expressed as MIC₅₀ and the mean zone of inhibition, is tabulated in Tables ​Tables17,17, ​,18,18, and ​and19.19. The zone of inhibition for Ampicilline (positive control) ranges from 15.3 ± 0.3mm to 44.3 ± 0.2mm [53].

3.8. Antifungal Activity

B. pilosa has traditionally been used to treat microbial infection. Recently, different parts of B. pilosa have been tested for antifungal activities. Deba and colleagues first evaluated the antifungal effect of the hot water extracts of the B. pilosa roots, stems, and leaves against Corticium rolfsii, Fusarium solani, and Fusarium oxysporum. They discovered that C. rolfsii was most suppressed by treatment with B. pilosa as its growth was reduced at almost all the tested doses, followed by F. oxysporum and F. solani [97]. However, the fungicidal activities of the stems, and roots were greater than the leaves [97]. Moreover, the same group assessed the antifungal activity of the essential oils and aqueous extracts from B. pilosa flowers and leaves [53]. They showed that the extracts and oils had antifungal activity against C. rolfsii, F. solani, and F. oxysporum. Essential oils appeared to have better fungicidal activity than water extracts as summarized in Table 20.

Another study by Ashafa and colleagues showed that acetone, methanol, and water extracts of the B. pilosa roots showed antifungal activities against Aspergillus niger, A. flavus, and Penicillium notatum using the agar dilution method. The results are tabulated in Table 21 [54]. Negative controls showed 0% growth inhibition. The methanol extract of the B. pilosa roots at 10mg/mL was also effective against Candida albicans [54]. Of note, B. pilosa obtained from Papua New Guinea had no activity against A. niger and C. albicans [125], but the South African (Eastern Cape) ecotype exhibited moderate activity against C. albicans [54]. This discrepancy may depend on extraction solvents, extraction procedure, assay techniques, different plant parts, and abundance of active compounds.

B. pilosa produces a variety of secondary metabolites such as flavonoids, phenylacetylenes, alkaloids, sterols, terpenoids, and tannis [53, 54]. However, none of them have been confirmed as active compounds against fungi. Further investigation of active compounds from B. pilosa is necessary to further understand the antifungal efficacy of this plant.

3.9. Hypotensive and Vasodilatory Activities

In early studies, Dimo and colleagues used three rat models, normotensive Wistar rats (NTR), salt-loading hypertensive rats (SLHR), and spontaneous hypertensive rats (SHR) to investigate the hypotensive effect of the methanol crude extract of B. pilosa leaves [19, 126]. The extract lowered systolic blood pressure in hypertensive rats (SLHR and SHR) to a greater degree than NTR [19, 126]. In addition, a decrease in urinary sodium ions and an increase in urinary potassium ions were observed after treatment with the methanol extract of B. pilosa leaves although neither differences were statistically significant [19, 126]. Taking the data together, the study proposed that B. pilosa leaf extract reduced blood pressure via vasodilation [19, 126]. The same group continued to test the antihypertensive effect of aqueous and methylene chloride extracts of B. pilosa leaves in a hypertensive rat model [19, 126]. To establish a fructose-induced hypertension model, male Wistar rats were given 10% fructose solution to drink ad libitum for three weeks. In addition to free access to 10% fructose, the rats were treated with the aqueous (150mg/kg) or methyl chloride (350mg/kg) extracts of B. pilosa for additional three weeks [19, 126]. Both extracts of B. pilosa leaves had a hypotensive effect on rats. However, neither extracts reversed the elevation of serum insulin in fructose-fed rats. Therefore, B. pilosa lowered blood pressure irrespective of insulin [19, 126]. To better understand the hypotensive mechanism, the authors investigated the effect of a neutral extract of B. pilosa (NBP), a mixture of methanol and methylene chloride (1:1) extract after neutralization with NaOH and HCl, on the heart and the blood pressure of NTR and SHR [109]. This study showed that an intravenous injection of the NBP resulted in a biphasic reduction in systolic blood pressure. In addition, one intravenous dose of the extract at 10, 20, and 30mg/kg BW decreased systolic blood pressure in normal rats by 18.3%, 42.5%, and 30%, respectively, and the same doses reduced the blood pressure in hypertensive rats by 25.8%, 38.9%, and 28.6%, respectively. Only the highest dose (30mg/kg) affected the force of the contraction of the heart. Atropine and propranolol were used to interfere with the hypotensive action of the NBP. Atropine reduced the initial phase of the hypotensive response in NBP and completely abolished the second phase of hypotensive response in NBP. In contrast, propranolol increased the first hypotensive response but partially abolished the second hypotensive response provoked by the NBP [109]. This mechanistic study suggested that B. pilosa invokes the biphasic hypotensive responses via targeting cardiac pump efficiency during the first phase and vasodilation at the second phase [109].

A further study was performed to investigate the relaxing effect of a neutral extract of B. pilosa (NBP) on rat aorta contracted with KCl (60mM) and norepinephrine (0.1mM) [55]. Cumulative addition of NBP relaxed the rat aorta previously contracted by KCl in a dose-dependent manner. The EC₅₀ value of the NBP for vasorelaxation was 0.32mg/mL. The data also showed that the NBP reduced the contraction of aorta previously contracted by KCl irrespective of the presence of aortic endothelium [55].

Pretreatment with glibenclamide, an ATP-dependent K⁺ channel blocker, did not considerably affect the relaxant effect of the NBP on KCl-induced contraction, suggesting that the vasodilatory effect of B. pilosa was not related to the opening of this ATP-dependent K⁺ channel [55]. On the other hand, in the presence of indomethacin or pyrilamine maleate, the relaxant response induced by the plant extract was significantly inhibited at the lower concentrations. The plant extract was able to reduce the aorta resting tone, inhibit the KCl-induced contractions by 90% at 1.5mg/mL and the CaCl₂-induced contractions by 95% at 0.75mg/mL. These results demonstrate that B. pilosa can act as a vasodilator probably via acting as a calcium antagonist [55].

However, no specific compound for the above activity has been identified from B. pilosa to date. A bioactivity-guided identification approach may be adopted to identify the active compounds in B. pilosa that possess hypotensive and vasodilatory effects and understand their mechanism of action.

The authors thank their laboratory members for constructive suggestions and the authors whose publications they cited for their contributions. The authors also thank Ms Miranda Loney of Academia Sinica for English editing of this paper. This work was supported by Grants 99-CDA-L11 and 101S0010056 from Academia Sinica, Taiwan. Arlene P. Bartolome is a recipient of MECO-TECO Sandwich Scholarship Program between Taiwan and the Philippines.