A Pharmacological Overview of Alpinumisoflavone, a Natural Prenylated Isoflavonoid

Over the last decade, several studies demonstrated that prenylation of flavonoids enhances various biological activities as compared to the respective nonprenylated compounds. In line with this, the natural prenylated isoflavonoid alpinumisoflavone (AIF) has been explored for a number of biological and pharmacological effects (therapeutic potential). In this review, we summarize the current information on health-promoting properties of AIF. Reported data evidenced that AIF has a multitherapeutic potential with antiosteoporotic, antioxidant and anti-inflammatory, antimicrobial, anticancer, estrogenic and antiestrogenic, antidiabetic, and neuroprotective properties. However, research on these aspects of AIF is not sufficient and needs to be reevaluated using more appropriate methods and methodology. Further series of studies are needed to confirm these pharmacological effects, and this review should lay the basis for the design of respective investigations. Overall, despite the drawbacks of studies recorded, AIF exhibits a potential as drug candidate.


INTRODUCTION
In the drug discovery process, plants still remain an invaluable source of drugs and drug leads. They possess enormous structural and chemical diversity that is not matched by any synthetic libraries of small molecules (Shen, 2015). As pharmacological activities of chemicals are generally structure dependent, the structural and chemical diversity is obviously an advantage. Over the last decade, the interest in (iso) flavonoids strongly increased. Especially the prenylated forms moved into the focus because of their versatile and promising pharmacological properties and health benefits on multitarget tissues (Kumar and Pandey, 2013;Chen et al., 2014). Prenylated isoflavonoids have increased lipophilicity as compared to nonprenylated forms, leading to high affinity with cell membranes and enhanced biological activities or significant pharmacological effects (Chen et al., 2014;Sherif et al., 2015). These compounds offer a multitude of biological activities, which justify major and much deeper pharmacological investigation (Botta et al., 2009). Accordingly, there is a recent in-depth investigation of prenylated flavonoids as promising anticancer, anti-inflammatory, antioxidant, and neuroprotective nutraceuticals (Yang et al., 2015;Venturelli et al., 2016), with the prenyl substituent playing a key role in the molecular activity. Prenylated flavonoids are found predominantly in the Leguminosae family, although the September 2019 | Volume 10 | Article 952 Frontiers in Pharmacology | www.frontiersin.org phenyl-propanoid pathway-necessary for their production-is ubiquitously present in plants including nonleguminous families (Reynaud et al., 2005;Lapčík, 2007).
Alpinumisoflavone (AIF) or [5-hydroxy-7-(p-hydroxyphenyl)-2,2-dimethyl-2H-6H-benzo- [1,2-b:5,4-b]dipyran-6-one] is a dimethylpyrano derivative prenylated at ring A of genistein (Figure 1). It is a major constituent of Derris eriocarpa F.C. How, commonly referred as "Tugancao" in "Zhuang" and "Dai" ethnomedicine in Guangxi and Yunnan Province of China (Guangxi Institute of Chinese Medicine, 1986). A high content of AIF was reported in fully mature fruits (mandarin melon berry) of Cudrania tricuspidata Bur. ex Lavallee (syn. Maclura tricuspidata Carrière) (Shin et al., 2015), a crop cultivated in East Asia (Xiong et al., 1993;Shi, 2010), Europe and America (Markovski, 2016) for its fruits and timber, and with an immense medicinal and economic value (Xin et al., 2017). Isolated for the first time by Jackson et al. (1971), AIF was identified in many medicinal plants widely used over the world (Table 1). Although data depicted in this table are not exhaustive, the global trend is in accordance with Botta et al. (2005) who reported that prenylated flavonoids occur mostly in Leguminosae and Moraceae, with few detected in other families. Over the last two decades, the body of literature of AIF and its pharmacological potential is steadily growing. This review summarizes and gives a critical look on the current knowledge of the biological activities, therapeutic potential, and mechanism of action of AIF.

Estrogenic and Antiestrogenic Activities
Estrogenic plant-derived products act via binding to human estrogen receptors (ERs). AIF was found to be a weak ERα and ERβ binder with conflicting results concerning the preference for ERβ versus ERα (Djiogue et al., 2009;Magne Nde et al., 2012;Mvondo et al., 2012). The authors used the same estrogen receptor competitor assay based on fluorescence polarization in the same laboratory and according to the instructions of the same manufacturer. The discrepancies can probably be ascribed to the purity of compound. The ER competitive ligand binding assay cannot distinguish between estrogenic and antiestrogenic substances and does not provide insight into the ability of a substance to initiate the molecular cascade leading to altered gene expression (Legler et al., 1999). To overcome this disadvantage, reporter gene assays such as the ER-mediated chemically activated luciferase gene expression assay (ER-CALUX) and the yeast estrogen screen (YES) based on stably transfected cell lines are usually applied. In an ER-CALUX assay using human osteosarcoma U2OS cells stably transfected with ERα and transiently transfected with ERβ, AIF stimulated the endogenous ER-estrogen response element (ERE) interaction and, thus, the luciferase reporter gene activity (Djiogue et al., 2009;Magne Nde et al., 2012). However, in a yeast two-hybrid β-galactosidase assay, AIF failed to induce the ligand-dependent interaction of ERα and coactivator TIF2 as determined by the expression of a reporter gene, β-galactosidase (Okamoto et al., 2006). Although ER-CALUX and YES assays rely on the same principle and use the same receptors, the yeast cell wall is usually less permeable to compounds compared to mammalian cell membranes (Legler et al., 1999). This makes the ER-CALUX assay robust, more sensitive and more predictable than the YES assay (Leusch et al., 2010). In MCF-7 cells, AIF upregulated ERα target genes such as proliferating cell nuclear antigen (PCNA), cyclin D1, cyclin E1, cMyc (myelocytomatosis viral oncogene homologue), and liver receptor homologue 1 (LRH-1), and downregulated growth regulation by estrogen in breast cancer 1 (GREB1) (Magne Nde et al., 2012). On the other hand, AIF suppressed estradiol (E2)induced activity in U2OS-ERβ cells but not in U2OS-ERα cells and ERα yeast two-hybrid systems (Okamoto et al., 2006;Magne Nde et al., 2012). Antagonizing the ERβ-mediated signaling pathway in the presence or absence of E2 is not promising as ERβ is known to counteract the proliferative responses of ERα involved in estrogen-related cancers, osteoporosis, and cardiovascular diseases.
In in vivo studies AIF induced estrogen-like effects by increasing uterine wet weight as well as uterine and vaginal epithelial height in ovariectomized Wistar rats (Mvondo et al., 2011, Mvondo et al., 2012. In this model, AIF also reduced the hot flush index by increasing the FSH/LH ratio. It displayed atheroprotective effects by an augmentation of HDL-cholesterol levels, a reduction in the atherogenic index of plasma (Mvondo et al., 2011), and by upregulating the expression of estrogensensitive genes associated with bile acid formation (Cyp7a1) (Mvondo et al., 2015). Taken together, the in vitro and in vivo systems/models used to study estrogenic effects of AIF are quite suitable. The investigations demonstrated that AIF,

Antiosteoporotic Activity
Isoflavonoids are increasingly considered as a promising first-line prophylaxis for osteoporosis in clinical settings (Ma et al., 2008;Lambert and Jeppesen, 2018). The therapeutic strategies globally emphasize the inhibition of "osteoclast-mediated bone resorption" and/or the prevention of the apoptosis of osteoblasts and osteocytes.
Using experimental protocols of postmenopausal or glucocorticoid-induced osteoporosis, AIF exhibited an antiosteoporotic activity both in vitro and in vivo. A 6-week oral treatment with AIF (10 and 25 mg/kg) prevented ovariectomyinduced osteoporosis in mice by suppressing osteoclast differentiation (Cong et al., 2017). Despite its limits (Egermann et al., 2005;Lelovas et al., 2008), the ovariectomized rat/mouse model is the most widely used animal model in research on postmenopausal osteoporosis. Long-term exposure to glucocorticoids, e.g., in the treatment of chronic autoimmune and pulmonary disorders, cancers of the lymphoid system, as well as in the prevention of transplant rejection (Oakley and Cidlowski, 2013), is the primary cause of the secondary osteoporosis (Tanaka, 2014). In dexamethasone-induced osteoporosis in rats, 20 and 40 mg/kg AIF p.o. prevented bone loss (Wang et al., 2017b). In vitro, 5-20 µM AIF abrogated the dexamethasone-induced cytotoxicity and proapoptotic effects on osteoblasts and osteocytes (MC3T3-E1 and MLO-Y4 cells) through the inhibition of ROS production as well as through the activation of nuclear factor erythroid 2-related factor 2 (Nrf2) and AMPK-dependent NAD(P)H oxidase 2 (Nox2) signaling pathways (Wang et al., 2017b;Yin et al., 2018). Osteoporosis remains an important target of research (Hendrickx et al., 2015), and, despite some limits, the two described animal models are appropriate and closer to the human situation than other models. Although further investigation is needed, the available studies showed that AIF, by suppressing osteoclast differentiation or osteoblasts and osteocytes apoptosis, could have beneficial effects on postmenopausal-and glucocorticoidinduced bone damage.

Antioxidant and Anti-Inflammatory Activities
Through free radical-scavenging and antioxidative effects, antioxidants constitute the first line of defense against the pathogenesis of several diseases (Rani, 2017).
There is a long and ever-growing list of in vitro antioxidant assays. In the DPPH free radical scavenging assay, quite different IC 50 values were obtained with AIF probably due to the differences in assay conditions (Table 3). Variable DPPH concentrations, incubation times, sample volumes, solvent systems, and pH clearly result in large differences in IC 50 values (reviewed by Tan and Lim, 2015). To standardize the methodology and ensure comparability between studies or laboratories, a DPPH concentration of 50 µM (for good accuracy), an incubation time of 30 min, and methanol as solvent for less polar samples or buffered methanol for more polar samples have been proposed (Sharma and Bhat, 2009;Mishra et al., 2012). Results expressed in different units additionally impede cross-comparison in many cases. The DPPH assay does not actually measure the antioxidant activity but the reducing capacity of the sample (Benzie and Strain, 1999). Moreover, there is no linear relationship between the antioxidant concentration and the radical scavenging activity. The numerous drawbacks of this assay underline its ineptitude to evaluate the antioxidant capacity. Among the other single electron transfer (SET)-based assays such as the trolox equivalent antioxidant capacity (TEAC), ferric reducing antioxidant power (FRAP), and thiobarbituric acid reactive substances (TBARS), TEAC assay is most popular due to its convenient application which better reflects the antioxidant activity (Floegel et al., 2011). More recently, hydrogen atom transfer (HAT)-based assays like oxygen radical absorbance capacity (ORAC), total radical-trapping antioxidant parameter (TRAP), and crocinbleaching assays provide better analogies to in vivo action . The main limits of TRAP (which relies only on the lag phase of the kinetic curve for quantitation) and the crocin-bleaching (easily disturbed by compounds absorbing at the monitored wavelength of 450 nm, crocin is not sold as pure compound but as an extract of saffron) assays are probably the reason why the ORAC assay is currently preferred in food and pharmaceutical industries (Huang et al., 2005;Číž et al., 2010;Power et al., 2013). According to Tan and Lim (2015), a mix of SET and HAT-based assays, encompassing several different radical types is recommended to better estimate the overall antioxidant activity of a sample. In summary, the DPPH assay is not appropriate to evaluate the antioxidant activity of a sample. Moreover, the studies recorded in this review did not use the standardized protocols. However, results from the LPS-induced ALI protocol, well known to be associated with the production of ROS and oxidative stress (Su et al., 2014;Yeh et al., 2014), indicate that, via an activation of antioxidative enzymes, AIF could be beneficial in the treatment of diseases associated with oxidative stress. The overproduction of free radicals is usually associated with excessive or sustained inflammatory reactions (Dandekar et al., 2015). Administered 1 h before LPS challenge, AIF (10 mg/kg i.p.) alleviated LPS-induced lung lesions, pulmonary edema, and hemorrhages by reducing the activity of myeloperoxidase . In this model and in LPS-induced RAW264.7 cells, 5 and 10 µg/ml AIF inhibited the production of proinflammatory mediators including tumor necrosis factor (TNF)-α, interleukin (IL)-6, IL-1b, intercellular adhesion molecule-1 (ICAM-1), and nitric oxide (NO). The mechanisms underlying these activities include the suppression of nuclear factor-kappa B (NF-κB), mitogen-activated protein kinases (MAPKs), the nucleotide-binding domain-like receptor protein 3 (NLRP3) inflammasome, and IL-17 signaling pathways . A similar effect on LPS-induced NO production in RAW264.7 cells was observed by Namkoong et al. (2011) with an IC 50 value of 15.97 µM. In GOLD docking fitness, AIF displayed a fitness score of 35.42 against cyclooxygenase (COX)-2, an inducible enzyme only expressed after an inflammatory stimulus (Uddin et al., 2014). Four hours after administration, AIF (25 mg/kg i.p.) reduced the carrageenan-induced rat paw edema by 29% in male Wistar rats, whereas the positive control indomethacin (10 mg/kg i.p.) reduced it by 41.67% (Amen et al., 2013). The models used in studies recorded in this review including the carrageenan-induced rat paw edema and the LPS-induced ALI are well-established to study the anti-inflammatory potential of chemicals. Therefore, results from these models provide a strong evidence of the anti-inflammatory potential of AIF.
The antibacterial activity of flavonoids is often widely conflicting mainly due to the use of different nonstandardized techniques. To overcome this issue, the Clinical and Laboratory Standards Institute (CLSI) and the European Committee for Antimicrobial susceptibility testing (EUCAST) have approved and published some guidelines over the last two decades (CLSI, 1998;CLSI, 2002;CLSI, 2004;CLSI, 2008;CLSI, 2010a;CLSI, 2010b;CLSI, 2012a;CLSI, 2012b;EUCAST Definitive Document, 2000;EUCAST Discussion Document, 2003). However, despite these guidelines for agar dilution, broth microdilution, and broth macrodilution, results from nonstandardized protocols are still published even in highly reputable journals. Although the disk diffusion technique is easy to apply without specialized equipment, and cheap, the determined IZ value is not related to the antibacterial activity but depends on polarity, concentration, and molecular weight of compounds (Tan and Lim, 2015). Thus, highly polar compounds display a high IZ, and many compounds with the same diffusion rate result in quite different antimicrobial activities. This method is only useful for a simple qualitative screening. It does not allow the quantification of the amount of the antimicrobial agent diffused into the agar, impeding the determination of MICs and MBCs (Ncube et al., 2008;Balouiri et al., 2016). The broth macrodilution or microdilution assays are among the most appropriate methods to determine MIC and MBC values despite the fact that they are unsuitable for highly nonpolar compounds (Tan and Lim, 2015). The reproducibility and the low price due to small amounts of reagents are the main advantages of the microdilution assay over the macrodilution assay. The latter is tedious to perform, requires a lot of manual handling, and is associated with a risk of errors in the preparation of antimicrobial solutions (Jorgensen and Ferraro, 2009). Accordingly, the microdilution method appears to be more accurate. In general, the interpretation of the efficacy depends on the profound knowledge of the model and the used protocol. Nevertheless, stringent endpoint criteria have been set to MIC values of <10 µg/ml or <25 µM for promising plant compounds (Ríos and Recio, 2005;Cos et al., 2006). According to this criterion, AIF could be considered promising only against S. aureus ATCC12600, E. coli ATCC11775, K. pneumonia ATCC13883, B. subtilis ATCC6051, and wild and ATCC18804 strains of C. albicans. Overall, despite the limits of the used assays/ protocols and the discrepancies in results, recorded data suggest the potential of AIF to act as an antimicrobial drug against few microorganisms.

Anticancer Activity
Different studies reported promising anticancer activities of plant-derived (iso)flavonoids (Magne Nde et al., 2015;Nwodo et al., 2016;Patil and Masand, 2019) including the suppression of proliferation, migration/invasion, tumor angiogenesis and metastasis, and the promotion of apoptosis in various cancers.
In several studies the 3-(4,5-dimethylthiazole-2-yl)-2,5diphenyltetrazolium bromide (MTT) assay was applied to determine the cytotoxic effects of AIF: The compound exhibited strong cytotoxicity against human oral epidermoid carcinoma KB cells (IC 50 = 4.13 μg/ml) (Nkengfack et al., 2001) and murine leukemia P-388 (IC 50 = 4.31 μg/ml) cells (Tjahjandarie and Tanjung, 2015a). IC 50 values of 19, 34, and 41 μM, respectively, were observed against human leukemia HL-60, K-562, and MOLT-4 cell lines . In human lung H2108, H1299, and MRC-5 cancer cell lines, AIF displayed moderate cytotoxicity with IC 50 values of 33.5, 38.8, and 52.5 µM, respectively (Namkoong et al., 2011). At the concentration of 10 µM, the growth and invasion of the human clear-cell renal carcinoma ccRCC 786-O and Caki1 cells were suppressed by 40 and 50-60%, respectively (Wang et al., 2017a). An IC 50 value of >25 μM was obtained against human prostate PC-3 (Nana et al., 2012) and neuroblastoma SH-SY5Y (Hong et al., 2018) cancer cells. At the concentration of 10 −5 M, AIF inhibited the growth of the renal SN12C cancer cells by 32.67% (Amen et al., 2013). AIF (10 μg/ml) displayed a low antiproliferative activity (30-40% inhibition) against the human melanoma A375 and SK-MEL-1 cells after a 24-h incubation and suppressed the migration and invasion of these cell lines (Gao et al., 2017). However, after 48 h of incubation, AIF did not exhibit inhibitory effects against SK-MEL-28 cells in a study by Hu et al. (2017). The degree of cytotoxicity in MTT assay increases with the cell number, the concentration of MTT, and the incubation time (van Tonder et al., 2015). The concentration of MTT was not indicated in the two latter studies, and the cell number (5 × 10 3 cells/well) was only indicated by Hu et al. (2017). Comparison of the incubation times showed that the higher incubation time (Hu et al., 2017) was associated with lower antiproliferative activity. Although neglected in the vast majority of studies, long incubation times are often associated with the decomposition, metabolism, or precipitation of compounds (Ateba et al., 2018).
Under use of a cell counting kit-8 (CCK-8) assay, AIF exhibited a moderate antiproliferative activity against human esophageal squamous carcinoma (ESCC) Eca109 and KYSE30 cells and at 5 µM enhanced the sensitivity of these cell lines to irradiation . In the same model, it also reduced the viability of colorectal HCT-116 and SW480 cancer cells (IC 50 of 10 and 5 µM, respectively) (Li et al., 2019). The enzyme-based methods including CCK-8 and MTT assays are easy to use, safe, and have a high reproducibility. However, the toxicity of MTT as well as interference of polyphenols with the tetrazolium MTT dye has to be taken into consideration (Wang et al., 2010). A further advantage of the CCK-8 method is its far higher sensitivity (https://www.dojindo.eu.com/Shared/Flyers/ Flyer_CCK-8-Rev.pdf) At 10 −5 M, AIF inhibited the growth of leukemia CCRF-CEM, MOLT-4, and HL-60 cancer cells by 51.17, 26.15, and 15.49%, respectively. In this study, the type of assay was not specified (Amen et al., 2013).
Induction of apoptosis is a very important property of anticancer drug candidates. In 786-O and Caki1 cells, AIF led to apoptosis by modulating the miR-101/RLIP76 signaling pathway through the inhibition of Akt (Wang et al., 2017a). In addition to the induction of DNA damage and cell cycle arrest , AIF induced apoptosis in ESCC cells by upregulating the miR-370/Pim family kinases 1 (PIM1) signaling (Han et al., 2016) and by suppressing the expression of Nrf2, HO-1 and NADPH:quinine oxidoreductase-1 . In HL-60 leukemia cells, apoptotic cell death was observed via the suppression of NF-κB and the signal transducer and activator of transcription (STAT) signaling pathway . AIF induced lung tumor apoptotic cell death by repressing both the ERK/MAPK and NF-κB pathways (Namkoong et al., 2011). In HCT-116 and SW480 cells, it triggered apoptosis by blocking DNA damage repair mediated by the DNA double-strand break repair gene RAD51 (Li et al., 2019) and in CCRF-CEM cells through the loss of MMP and production of ROS (Kuete et al., 2016).
The role of increased activity of hypoxia-inducible factor-1 (HIF-1), especially HIF-1α is well known in cancer progression (Massoud and Li, 2015;Schito and Semenza, 2016). Hypoxic cancer cells seem to be resistant to radiation and chemotherapy (Rohwer and Cramer, 2011;Zhang et al., 2015). Therefore, targeting HIF-1 is an important approach for cancer prevention and treatment. AIF suppressed both hypoxia-induced and iron chelator-induced HIF-1 activation in T47D human breast cancer cells as well as MDA-MB-231 cell migration (Liu et al., 2009).
The antiproliferative or cytotoxic activity associated with apoptosis in malignant cells is a highly important target in the screening of anticancer drugs. Given the severe adverse reactions in normal tissues by tumoricidal doses of chemotherapeutic agents, the cytotoxic activity of drug candidates should also be evaluated against normal cells. In addition to a strong antiproliferative activity (IC 50 <4 µg/ml or <10 µM for a pure compound after 48-72 h incubation) (Boik, 2001), a high selectivity (selectivity index ≥3) towards malignant cells is needed. The inclusion of positive controls in respective studies of natural compounds is indispensable in good experimental practice (Ateba et al., 2018). Among the 13 in vitro studies recorded in this review, only one investigated the effects against normal cells  and four used a positive control (Nana et al., 2012;Tjahjandarie and Tanjng, 2015a;Hu et al., 2017;Hong et al., 2018). In addition to this deficit, human tumor cell lines as the workhorse of cancer research are cultured since decades and do not adequately mirror (different tumor environment) the biology of human tumors (Ben-David et al., 2017). Therefore, in vivo models with better and more clinically predictive power of human cancers are an imperative.
In vivo, AIF has been tested in various xenograft mouse models. After 30 consecutive days of treatment, the compound reduced tumor growth in KYSE30 (50 and 100 mg/kg/day) and Eca109 (20 mg/kg/day) xenograft mouse models (Han et al., 2016;Zhang et al., 2017). It also suppressed the tumor growth in an HCT-116 xenograft mouse model after 24 days treatment (25 and 50 mg/kg/day AIF i.p.) (Li et al., 2019). In a B16-F10 mouse lung model of metastasis, 24-day intragastrical administration of 20 and 50 mg/kg/day AIF decreased the number of metastatic pulmonary nodules. The reduction in COX-2 through modulating miR-124/SPHK1 axis was the underlying mechanism involved (Gao et al., 2017). The dose of 40 and 80 mg/kg/day for 24 days suppressed the growth and pulmonary metastatic nodules in a 786-O xenograft mouse model by modulating miR-101/RLIP76 signaling (Wang et al., 2017a). In a study by Zhang et al. (2017), combination of AIF (20 mg/kg/day for 30 days) with irradiation induced a more profound tumor regression than single treatments. All these in vivo activities occurred without affecting the body weight of the mice. Despite the drawbacks of the majority of the current cell-line-derived or patient-derived mouse xenograft models reviewed by Landgraf et al. (2018), they have become a prominent cancer model system over decades. For research in pharmaceutical industry, the accurate description of materials and methodology is indispensable to assure that experiments can be accurately replicated. However, the route of administration of AIF, an extremely important parameter, is not mentioned in the studies by Zhang et al. (2017) and Wang et al. (2017a), published in "high-impact" journals. As a different route of administration leads to different results, this underlines the importance of an accurate review of such papers. Nevertheless, all data reported in this section demonstrate that AIF could have a potential to suppress some tumor growth in vivo. September 2019 | Volume 10 | Article 952 Frontiers in Pharmacology | www.frontiersin.org

Antidiabetic Activity
Adequate glycemic control remains the main foundation of managing diabetes mellitus (DM) (Chaudhury et al., 2017). Retarding the release of D-glucose from dietary carbohydrates and delaying its absorption through the inhibition of α-glucosidase is an attractive therapeutic target for the treatment of DM, obesity, and other related complications (van de Laar et al., 2005). In vitro, AIF exhibited a moderate α-glucosidase inhibitory activity with an IC 50 value of 73.3 ± 12.9 μM (Fu et al., 2018).
Protein tyrosine phosphatase 1B (PTP1B) is a negative key regulator of insulin signaling pathways that leads to insulin resistance. Thus, it is a promising molecular-level therapeutic target in the management of type 2 DM and obesity (Wang et al., 2015). In a study by Na et al. (2006), AIF exhibited in vitro PTP1B inhibitory activity with an IC 50 value of 42 μM as compared to the positive controls RK-682 (IC 50 = 4.5 ± 0.5 μM) and ursolic acid (IC 50 = 3.6 ± 0.2 μM) (Trinh et al., 2017). By increasing the AMPK activation and the expression of glucose transporters' (GLUT-4 and -1) mRNA as well as by inhibiting the PTP1B activity (IC 50 = 37.52 µM vs. ursolic acid-5.13 µM), AIF significantly stimulated the glucose uptake in L6 myotubes (Lee et al., 2009b). These differences in IC 50 values can be explained by the application of different experimental conditions. Using a nonkinetic method to estimate the amount of produced p-nitrophenol at 405 nm, Na et al. (2006) added 10 M NaOH to stop the reaction, while in the study of Trinh et al. (2017), the release rate of p-nitrophenol (kinetic method) was determined by measuring the absorbance at 405 nm every 30 s for 10 min. Moreover, these studies used different concentrations of PTP1B and the substrate p-nitrophenyl phosphate.
Acyl-CoA:diacylglycerol acyltransferase (DGAT) is a key enzyme in the synthesis of triglycerides, the imbalance of which usually leads to insulin resistance and type 2 DM. At the concentration of 12.5 μg/ml, AIF induced 23% inhibition of the activity of this enzyme, while the positive control displayed an IC 50 value of 4.8 µg/ml (Oh et al., 2009).
Overall, these preliminary results suggest that AIF could exhibit a potential for the treatment of type 2 DM by retarding the glucose absorption from small intestine, by increasing the insulin sensitivity and the glucose transport into cells, and by improving triglycerides' profile. But most important, this hypothesis has to be confirmed by respective meaningful in vivo models.

Neuroprotective Activity
Elevation of the activity of brain monoamine oxidases (MAOs), especially MAO-B, contributes to chronic neurodegeneration and brain atrophy (Naoi et al., 2018;Tong et al., 2017). AIF inhibited the mixed type of mouse total brain MAO with an IC 50 value of 25.8 μM. Its activity on MAO-B (IC 50 = 16.8 μM) was 3.1-fold higher than that on MAO-A (Han et al., 2005). Globally, AIF was more active than the positive control amitriptyline on mixed MAO, MAO-A, and MAO-B. By destroying dopaminergic and noradrenergic neurons in the brain through excessive production of ROS such as superoxide radicals, the neurotoxin 6-hydroxydopamine (6-OHDA) induces neuronal cell death and Parkinson's disease in rats (Heikkila et al., 1989;Perese et al., 1989;Schober, 2004). At noncytotoxic concentrations, AIF attenuated (IC 50 > 25 μM) the 6-OHDA-induced neurotoxicity and ROS generation in SH-SY5Y cells .
The relatively high MAO inhibitory activity of AIF compared to amitriptyline and its capacity to protect against 6-OHDAinduced neurotoxicity justifies further in-depth investigations of AIF for its potential in neurodegenerative diseases such as Parkinson's and Alzheimer's.

Other Activities
With 216 million cases and 445,000 deaths in 2016, malaria remains a major cause of death worldwide, especially in Africa (http:// www.who.int/malaria/en/). AIF has shown strong antiplasmodial properties against Plasmodium falciparum with an IC 50 value of 1.98 µg/ml as compared with the positive control chloroquine (IC 50 = 1.02 µg/ml) (Tjahjandarie and Tanjung, 2015b).
HIV-1 protease and reverse transcriptase are the most important targets in the search for anti-HIV agents. In vitro, AIF showed a low inhibitory activity against HIV-1 protease with an IC 50 value of 30.1 μM (Lee et al., 2009a).
The differences in the activities of some compounds closely related to AIF, preclinically tested in comparable assays but not directly compared with AIF, are difficult to interpret due to the limitations as discussed above. Nevertheless, we include those results, which were obtained in the most similar experimental setups: i) The 2′-OH derivative parvisoflavone B showed a stronger α-glucosidase inhibition (IC 50 = 12.2 µM; Dendup et al., 2014) than AIF (IC 50 = 73.3 µM; Fu et al., 2018). Nevertheless, due to the lack of a positive control in the study with AIF, a comparison of the results remains questionable. Parvisoflavone B resulted also in a better cytotoxic effect against MDA-MB-231 breast cancer cells (EC 50 = 16.9 µM) (Nyandoro et al., 2017). The weak antimycobacterial effect (MIC = 90.9 µM) of parvisoflavone B against M. tuberculosis H37rV) (Nyandoro et al., 2017) differed from the inactive AIF (Kuete et al., 2008). ii) Cudraisoflavon M-carrying an additional 2,3-dihydroxyprenyl group at C-8 of AIF-did not show any activity against 6-OHDA-induced cell death in SH-SY5Y cells (Hiep et al., 2017), whereas cudraisoflavon H with a prenyl group at C-8 and an additional hydroxy group at C-2″ of AIF resulted in an IC 50 value of 4.5 µM (Hiep et al., 2015). iii) The comparison of the antimicrobial effects of derrone with AIF is extremely difficult due to the differences in the experimental setup as discussed above. Nevertheless, the activity of derrone against E. coli, S. aureus, and C. albicans seems lower than the one of AIF (Edziri et al., 2012). The antiproliferative potential of derrone and AIF differ in dependence of the cell line: In SW480 cells, AIF with an IC 50 value of 5 µM inhibits the proliferation (Li et al., 2019), whereas derrone remains inactive . In MCF-7 cells, AIF was inactive (Sudanich et al., 2017) and derrone at 10 µM inhibited this cell line by 13.6% . HepG2 cells seem to be similarly sensitive to the two compounds [AIF-IC 50 37.99 µM (Kuete et al., 2016); derrone-23.7% inhibition at 10 µM ].
iv) 2′-Hydroxyerythrin A with the OH group from C-5 shifted to C-2′ showed good activity against several Grampositive and Gram-negative bacteria (Wang et al., 2018). The magnitude of the DPPH radical scavenging effect was in the same range (Wang et al., 2018) as in some studies with AIF (Rahman et al., 2010;Bórquez et al., 2013;Fu et al., 2018).
The infrequence of studies and the low number of different substitution patterns and of investigated activities are the main drawback in the deduction of structure-activity relationships (SAR) of AIF and its derivatives. Continued efforts are needed to further synthesize or isolate new derivatives of AIF to expand SAR. Nevertheless, published data indicate that the C4′-Omethylation and the C8-prenylation increase the activity of AIF in cancer and neurodegenerative conditions. This is in accordance with Bernini et al. (2011) who indicated that O-methylation of flavonoids ensures a superior anticancer activity as compared with the corresponding hydroxylated derivatives, since such compounds are more resistant to hepatic metabolism and show higher intestinal absorption. In addition, Walle et al. (2007) suggested that O-methylation enhances the stability of flavonoids to metabolic degradation and increases their bioavailability as well as a higher tissue distribution as compared to unmethylated forms.

POINT OF VIEW AND FUTURE PERSPECTIVE
Prenylated (iso)flavonoids are attracting more and more attention due to a series of promising biological activities ascribed to their increased lipophilicity and a strong affinity to biological membranes as compared to the respective unprenylated compounds (Botta et al., 2005;Botta et al., 2009;Chen et al., 2014;Sherif et al., 2015;Mukai, 2018). In this context, numerous pharmacological investigations of alpinumisoflavone, extracted from various medicinal plants, were carried out over the last decades. Data recorded in this review evidence a wide array of activities such as antiosteoporotic, antioxidant, anti-inflammatory, antimicrobial, anticancer, estrogenic and antiestrogenic, antidiabetic, and neuroprotective. Discrepancies between results were usually attributed to the purity of the tested compound, the experimental setup, the operator's experience, or other experimental parameters (Ateba et al., 2018). Many of related pathologies or conditions such as antimicrobial resistance, cancer, diabetes mellitus, and neurodegenerative diseases are becoming pivotal concerns for public health over the world. However, although AIF might be considered a promising preventive and/or therapeutic agent for such ailments, these investigations are only at the beginning. Using suitable and well-designed standardized models or assays, further and thorough studies related to the above mentioned or other pathologies/conditions are needed to confirm this potential. In vitro evaluation is an important primary screen and due to its rapidity common practice in many research laboratories. Nevertheless, many in vitro studies are not necessarily optimal due to poor standardization, redundancy, and/or outdated methodology (Tan and Lim, 2015). Clearly, compounds exhibiting promising activity require further studies to validate or confirm their therapeutic potential (Kenny et al., 2015). Accordingly, the correlation with in vivo data using appropriate models is an indispensable prerequisite. The analysis of structure-activity relationships provides information on the preferential conformation to maintain high activities. Studies of AIF until now revealed that the free -OH groups at C-4′ and C-5 are important for the fungicidal activity towards C. albicans (Ayine-Tora et al., 2016). 4′-O-Methylation and the presence of a prenyl group at C-8 enhanced anticancer activities (Liu et al., 2009;Hu et al., 2017). Studies with diversified substituents would be ideal for the investigation of SARs. They might allow the identification of important structures with reduced toxicity and increased therapeutic efficacy that can guide the design of novel leads or drug candidates. However, such studies on AIF and its derivatives are scarce until now, and this underlines the necessity of further well-performed investigations.
Besides the efficacy, extensive safety and pharmacokinetic data are required for potential drug candidates as an important aspect in the drug development process. However, till today, no study dealing with the toxicity or pharmacokinetics of AIF has been reported.