Alpinumisoflavone against cancer pro-angiogenic targets: In silico, In vitro, and In ovo evaluation

Abstract

Background

Breast cancer is currently the world’s most predominant malignancy. In cancer progression, angiogenesis is a requirement for tumor growth and metastasis.Alpinumisoflavone (AIF), a bioactive isoflavonoid, exhibited good binding affinity with the angiogenesis pathway’s druggable target through molecular docking.

Objectives

To confirm AIF’s angiogenesis inhibitory activity, cytotoxic potential toward breast cancer cells, and druggability.

Methods

Antiangiogenic activity was evaluated in six pro-angiogenic proteins in vitro, duck chorioallantoic membrane (CAM) in ovo, molecular docking and druggability in silico.

Results

Findings showed that AIF significantly inhibited (p =  < 0.001) the HER2(IC₅₀ = 2.96 µM), VEGFR-2(IC₅₀ = 4.80 µM), MMP-9(IC₅₀ = 23.00 µM), FGFR4(IC₅₀ = 57.65 µM), EGFR(IC₅₀ = 92.06 µM) and RET(IC₅₀ =  > 200 µM) activity in vitro.AIF at 25 µM-200 µM significantly inhibited (p =  < 0.001) the total number of branch points (IC₅₀ = 14.25 μM) and mean length of tubule complexes (IC₅₀ = 3.52 μM) of duck CAM comparable (p =  > 0.001) with the positive control 200 µM celecoxib on both parameters.AIF inhibited the growth of the estrogen-receptor-positive (ER +) human breast cancer cells (MCF-7) by 44.92 ± 1.79% at 100 µM while presenting less toxicity to human dermal fibroblast neonatal (HDFn) normal cells.The positive control 100 µM doxorubicin showed 86.66 ± 0.93% and 92.97 ± 1.27% inhibition with MCF-7 (IC₅₀ = 3.62 μM) and HDFn, (IC₅₀ = 27.16 μM) respectively.In docking, AIF has the greatest in silico binding affinity on HER2 (-10.9 kcal/mol) among the key angiogenic molecules tested. In silico rat oral LD₅₀ calculation indicates that AIF is moderate to slightly toxic at 146.4 mg/kg with 1.1 g/kg and 20.1 mg/kg upper and lower 95% confidence limits. Lastly, it sufficiently complies with Lipinski’s, Veber’s, Egan’s, Ghose’s, and Muegge’s Rule, supporting its oral drug-like property.

Conclusion

This study revealed that AIF possesses characteristics of a phytoestrogen compound with significant binding affinity, inhibitory activity against pro-angiogenic proteins, and cytotoxic potential against ER + breast cancer cells.The acceptable and considerable safety and drug-likeness profiles of AIF are worthy of further confirmation in vivo and advanced pre-clinical studies so that AIF can be elevated as a promising molecule for breast cancer therapy.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s40199-022-00445-9.

Introduction

Cancer remains among the top and an unceasing alarming burden to public health globally [1]. There is a projection that the second prime cause of death, cancer, with new cases in 2020 of around 19.3 million, will reach 28.4 million by 2040. In the Philippines, there were 92,606 reported cancer mortalities in 2020. Among the different types of malignancy, breast cancer has been the top prevailing case globally and in the Philippines for the past five years [2]. These reports call for urgent action to address this incessantly rising health crisis.

The main driving force of cancer mortality is metastasis, responsible for over 90% of all cancer fatalities [3]. Sustained angiogenesis is crucial in cancer stage advancement by promoting tumor growth and metastasis [4]. Many pro-angiogenic molecules are highly active during tumor neovascularization. These include the most reported ones such as Vascular endothelial growth factor-A (VEGF-A), Vascular endothelial growth factor receptor-1 and -2 (VEGFR-1 & VEGFR-2), Epidermal growth factor receptor (EGFR), Human epidermal growth factor receptor 2 (HER2), Fibroblast growth factor receptor 1 and 4 (FGFR1 & FGFR4), Matrix Metalloproteinase-9 (MMP-9), and Rearranged during transfection (RET) receptor tyrosine kinase among others. They stimulate the endothelial cells lining the inside wall of the pre-existing blood vessels to proliferate, migrate, invade the surrounding tissues, and form tubes, creating a new blood vessel [5, 6]. As such, a compound that could inhibit the activity or expression of these critical molecules in vascularization has been the aim of most research endeavors searching for an ideal lead for cancer therapy.

Auspiciously, combining molecularly targeted drugs with chemotherapeutic agents improves anticancer efficacy. So, it is now one of the standard-of-care therapies [7]. Although there are a variety of antiangiogenesis medications on the market, continued research of lead compounds is necessary due to the limitations of existing therapies, such as the development of resistance, insufficient efficacy, and toxicity [8].

Alpinumisoflavone (AIF), a prenylated isoflavonoid, has been documented to have remarkable anticancer properties through its antioxidant, anti-inflammatory, apoptotic, pyroptotic, and antitumor activities [9–16]. AIF has also been shown to possess multi-kinase inhibitory potential against various pro-angiogenic molecules in silico screening. Particularly, the AIF showed the strongest binding affinity with MMP-9, VEGFR-2, and RET [17].

In this work, we presented the potential of AIF as a safe antiangiogenic compound. Reverse docking was done to evaluate the binding mode of AIF on various pro-angiogenic proteins. The in silico predicted proteins were prioritized based on docking scores and tested in vitro utilizing enzyme-based inhibitory kits. Its cytotoxic effect was evaluated in MCF-7 in vitro. The potential to inhibit blood vessel formation was determined in ovo CAM duck assay. The in silico investigation of toxicity and profiling of drug-likeness were also conducted to examine its druggability for development into a clinically applicable oral drug for breast cancer.

Material and methods

Chemical reagents and kits

AIF, regorafenib, sorafenib, lapatinib, gefitinib, and lenvatinib (≥ 98.0% purity) were from ChemFaces Biochemical CO., Ltd. (Wuhan, Hubei, China). The gelatinase (MMP-9) and kinase (VEGFR-2, HER2, RET, EGFR, and FGFR4) activity assay kits were from BioVision (Milpitas, CA, USA) and BPS Bioscience (San Diego, CA, USA), respectively. The Kinase-Glo™ Max luminescence kinase reagent was from Promega (Madison, WI, USA). The MTT dye and celecoxib were from Sigma Aldrich (St. Louis, MO, USA), while the doxorubicin hydrochloride was from the University of Santo Tomas-Hospital. Dimethylsulfoxide (DMSO) was from J.T. Baker (Center Valley, PA, USA), while the MCF-7 (ATCC HTB-22) and HDFn (PCS-201–010) cells were from the American Type Culture Collection (ATCC) (Rockville, MD, USA). The Gibco™ Dulbecco’s modified Eagle medium (DMEM) supplemented with D-glucose and L-Glutamine, Fetal Bovine Serum (FBS), 1% antibiotic–antimycotic solution (PenStrep-Ampothericin B), Dulbecco’s Phosphate Buffered Saline (DPBS), 0.25% Trypsin–EDTA (1X), and 0.4% Trypan blue were obtained from Thermo Fisher Scientific (Waltham, MA, USA). The sterile paper discs were bought from Hi Media® Laboratories (Vadhani, India). Lastly, the fertilized duck eggs were from an egg farm in Nueva Ecija, Philippines.

Screening for the antiangiogenesis property of AIF

In silico reverse molecular docking

The AutoDock Vina wizard v1.1.2 in the Python Prescription (PyRx) virtual screening software (https://pyrx.sourceforge.io) was used for reverse docking [18].

The structures of the ligands were accessed from the PubChem (https://pubchem.ncbi.nlm.nih.gov/) while the receptors (angiogenic proteins) were obtained from the RCSB protein data bank (https://www.rcsb.org/). Angiogenic proteins HER2, PDGFR, DHFR, FGFR1, FGFR4, BRAF, PAI-1, IGF1R, and FLT3, were the selected targets in the reverse docking. The PyRx’s Open Babel was utilized to prepare structures of the ligands, while UCSF Chimera v1.13 (https://www.cgl.ucsf.edu/chimera/) and PyRx’s AutoDock Tool were used to process and prepare the angiogenic proteins. Using the AutoDock Tools in PyRx, a PDBQT file of the ligand and receptor required for docking was prepared. A receptor grid was generated around the angiogenic protein’s target region to allow ligands to dock in a specified binding location. The grid box size (Table 1) was optimized so that the in silico docked conformation agreed with the original crystallized structure. The prepared ligands (AIF, original co-crystallized ligand, and standard inhibitors) were docked using PyRx’s AutoDock Vina wizard using the Assisted Model Building with Energy Refinement (AMBER) force field as the scoring parameter. A chimera software was used to examine the best pose of the in silico re-docked ligand by calculating the Root-Mean-Square Deviation (RMSD) vis-a-vis the conformation of the ligand in the original crystallized structure of the complex. If the RMSD value was less than 2.0 Å, the docking method was considered valid and successful [19, 20]. The interactions were graphically visualized, analyzed, and generated using the Discovery Studio (DS) 2021 Client software (https://www.3dsbiovia.com).

Table 1.

Optimized dimension of the grid box that locates and encloses the binding site

Angiogenic protein of interest	Protein Data Bank (PDB) ID	Resolution of Crystal Structure	Grid box parameters (x,y,z coordinates)
DHFR	3EIG	1.7 Å	Center: x = 11.676, y = -5.566, z = -14.736 Dimensions (Å): x = 23.793, y = 18.489, z = 25.869
FGFR1	5EW8	1.63 Å	Center: x = 89.991, y = 1.066, z = 14.992 Dimensions (Å): x = 16.710, y = 16.695, z = 22.565
FGFR4	4XUQ	1.95 Å	Center: x = -0.311, y = -2.860, z = 14.366 Dimensions (Å): x = 20.427, y = 20.385, z = 25.404
FLT3	4XUF	3.20 Å	Center: x = 24.246, y = 18.877, z = -15.819 Dimensions (Å): x = 22.168, y = 19.611, z = 29.155
HER2	3PP0	2.25 Å	Center: x = 18.996, y = 18.044, z = 26.791 Dimensions (Å): x = 23.995, y = 24.423, z = 20.457
IGF1-R	2OJ9	2.0 Å	Center: x = 4.205, y = -4.459, z = 18.776 Dimensions (Å): x = 16.335, y = 19.042, z = 21.513
PAI-1	4AQH	2.4 Å	Center: x = -26.227, y = 3.316, z = 0.273 Dimensions (Å): x = 24.499, y = 21.014, z = 20.056
PDGFR	6JOL	1.9 Å	Center: x = -38.810, y = 160.687, z = 1.294 Dimensions (Å): x = 22.986, y = 22.917, z = 27.375
B-RAF	5ITA	1.95 Å	Center: x = 25.516, y = -2.517, z = -19.993 Dimensions (Å): x = 25.552, y = 16.172, z = 16.445

In vitro inhibitory activity assay of AIF against selected angiogenic molecules

The MMP-9 inhibitory activity of AIF was investigated using a Gelatinase assay kit. The inhibitory property of AIF against VEGFR2, HER2, RET, EGFR, and FGFR4 activity was evaluated using Kinase assay kits with the Kinase-Glo™ Max luminescence kinase reagent. The assays were carried out following the manufacturer’s protocol.

In ovo CAM Assay

Roldan et al.’s (2018) [21] method was followed with modifications on the concentration of the positive control. Six fertilized duck eggs assigned in each group (zero-day old) were incubated for eight days at 37 °C ± 2 °C in a humidified Incubox automatic turning incubator. On the eighth day of incubation, after sanitizing the marked surface with 70% alcohol, a 1–2 cm window was cut in the eggshell to expose the CAM for direct access. Sterile filter paper discs about 10 mm in diameter were saturated with the freshly prepared test sample-AIF, positive control-celecoxib, and negative control-0.3% DMSO in DPBS and were placed in contact with the CAM using sterilized forceps. Then, the treated eggs were sealed using sterile adhesive tape and further incubated for 48 h. The entire experiment was carried out under sterile conditions. The hard eggshell covering was removed two days after incubation, and the CAM of the duck embryo was photographed using a Fujifilm X-A2 camera at a uniform distance (7 inches from the lens). The AngioQuant software v1.33 (MATLAB Inc., Tampere, Finland), an automated image analyzer, was used to quantify the antiangiogenesis activity based on the inhibition of the total number of branch points and the mean length of tubule complexes [22].

Test for the cytotoxic activity of AIF (MTT assay)

The cytotoxic activity of AIF on the MCF-7 (ATCC, passage 5–7) and HDFn (PCS-201–010, passage 3) cells was examined using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium-bromide) colorimetric assay. The cells were grown separately in a tissue culture-25 cm² polystyrene flask containing DMEM supplemented with D-glucose and L-glutamine, 10% FBS, and 1% PenStrep-Ampothercin B, in a humidified 5% CO₂ atmosphere at 37 °C incubator. After reaching 75–85% confluency, the cells were harvested, and the sterile 96-well microtiter plate was seeded with 1 × 10⁴ cells per well and was incubated for adhesion and formation of a confluent monolayer. Thereafter, cells were treated with AIF or doxorubicin. Doxorubicin served as the positive control, while 0.3% DMSO served as the negative control. Then, the treated cells were incubated for an additional 24 h or 48 h. Subsequently, the media were discarded, followed by adding 5 mg/mL MTT dye in PBS in each well. For another 4 h, the cells were incubated at 37 °C with 5% CO_2. Thereafter, DMSO was used to dissolve the formazan crystals, and the absorbance was read using a Thermo Scientific microplate reader at 570 nm. The assays were done in three trials with three replicates for each group for MCF-7, while one trial with three replicates per group was performed with the HDFn.

In silico Toxicity and drug-likeness profiling of AIF

In silico rat oral median lethal dose (LD₅₀) and ocular irritancy prediction

AIF was subjected to in silico rat oral LD_50, and ocular irritancy identification utilizing the Toxicity Prediction by Komputer Assisted Technology (TOPKAT) protocol in Discovery Studio 4.0 (DS4.0; BIOVIA-Dassault Systèmes, formerly Accelrys).

In silico drug-likeness profiling

SwissADME software (http://www.swissadme.ch/) was used to examine the physicochemical properties [23] of AIF in compliance with Lipinski’s [24], Veber’s rule [25], Ghose’s rule [26], Muegge’s rule [27], and Egan’s rule [28] to predict AIF’s druggability.

Statistical analysis

All data were analyzed by one-way analysis of variance (ANOVA) followed by post hoc Tukey’s HSD multiple comparison tests using IBM Statistical Package for Social Sciences version 21 (SPSS) for Windows. Data were presented as mean ± standard deviation (SD), and significant differences were established at p < 0.001. The half-maximal inhibitory concentration (IC₅₀) values were determined using the four-parameter logistic equation of GraphPad Prism Software version 9.2. See Supplementary Information for the formula on the data analysis of the percentage inhibition for the assays.

Results and Discussion

In silico reverse molecular docking

Reverse molecular docking

The tested proteins are the vital molecules and druggable targets in the angiogenesis signaling pathways, highly associated with aggressive malignant phenotype and poor cancer prognosis [5]. An RMSD value ranging from 0.477 Å to 1.938 Å (Fig. 1) was obtained, which is within the acceptable range of 0 Å-2.0 Å, indicating the protocol’s validity [29].

Fig. 1.

Superimposition of originally bound (white) and in silico re-docked co-crystallized (yellow) ligands or protein for the different target angiogenic proteins

The results showed that AIF had negative binding energy (Table 2) on the angiogenic proteins’ active site, indicating a favorable capacity to interact and bind to the targets with a good fit. A ligand binding to the active site would mean enzyme inhibition [30]. Thus, this gives an insight into AIF’s potential multi-targeted antiangiogenesis property. Findings revealed that AIF has a stronger binding affinity against the FDA-approved inhibitors of HER2, PDGFR, FGFR1, FGFR4, PAI-1, FLT3, and comparable binding affinity with DHFR as well as with IGF1-R.

Table 2.

Binding energy and key residues involved in ligand-target protein interaction

Angiogenic Proteins/PDB ID	Ligands	Binding energy (kcal/mol)	Key residues of interaction
1.HER2 (3PP0)	Test compound Alpinumisoflavone	–10.9	Leu726, Val734, ALa751, Lys753, Leu796, Cys805, Leu852, Phe864
	Standard inhibitors Lapatinib	–9.0	Ser728, Val734, Lys753, Leu796, Cys805, Asp808, Arg849, Leu852, Thr862, Asp863
	Neratinib	–8.4	Leu726, Val734, Ala751, Lys753, Leu796, Met801, Gly804, Cys805, Asp808, Leu852, Asp863
	Co-crystallized ligand 03Q1	–11.0	Leu726, Ser728, Val734, Ala751, Lys753, Glu770, Met774, Leu785, Leu796, Gln799, Met801, Leu852, Phe864
2.PDGFR (6JOL)	Test compound Alpinumisoflavone	–10.5	Leu599, Val607, Ala625, Tyr676, Cys677, Leu825, Phe837
	Standard inhibitors Regorafenib	–11.5	Leu599, Val607, Ala625, Lys627, Glu644, Met648, Val658, Thr674, Tyr676, Cys677, Leu809, Cys814, His816, Leu825, Cys835, Asp836
	Sorafenib	–11.2	Leu599, Val607, Ala625, Lys627, Glu644, Met648, Ile657, Val658, Tyr676, Cys677, Leu809, His816, Leu825, Ile834, Asp836
	Axitinib	–11.1	Leu599, Val607, Ala625, Lys627, Glu644, Met648, Ile672, Thr674, Glu675, Tyr676, Cys677, Leu825, Asp836, Phe837
	Pazopanib	–10.9	Leu599, Val607, Ala625, Lys627, Glu644, Met648, Ile657, Val658, Tyr676, Cys677, Leu809, Cys814, His816, Leu825, Cys835, Asp836, Phe837
	Lenvatinib	–10.3	Leu599, Val607, Ala625, Lys627, Glu644, Met648, Glu675, Tyr676, Cys677, Gly680, Leu825, Cys835, Asp836, Phe837
	Vandetanib	–9.8	Leu599, Val607, Ala625, Met648, Val658, Cys677, Asp681, Leu825, Phe837
	Sunitinib	–9.7	Leu599, Val607, Ala625, Lys627, Glu644, Met648, Ile657, Ile672, Cys835, Cys677, His816, Leu825, Phe837
	Co-crystallized ligand Imatinib	–12.5	Leu599, Val607, Ala625, Lys627, Met648, Ile672, Thr674, Tyr676, Cys677, His816, Leu825, Asp836, Phe837
3.DHFR (3EIG)	Test compound Alpinumisoflavone	–10.5	Asp21, Leu22, Arg31, Phe34, Lys55, Thr56, Ser59, Thr146
	Standard inhibitor Pralatrexate	–10.5	Ile7, Ala9, Ile16, Arg31, Phe34, Glu35, Arg70
	Co-crystallized ligand Methotrexate	–10.0	Ile7, Ala9, Leu22, Glu30, Arg31, Glu35, Ile60, Pro61, Asn64, Arg70, Val115
4.FGFR1 (5EW8)	Test compound Alpinumisoflavone	–9.9	Leu484, Val492, Ala512, Lys514, Met535, Val561, Leu630, Asp641
	Standard inhibitors Pazopanib	–8.9	Leu484, Val492, Ala512, Lys514, Glu531, Met535, Ile545, Val561, Glu562, Leu630, Ala640, Asp641
	Regorafenib	–8.8	Leu484, Val492, Ala512, Lys514, Glu531, Met535, Ile545, Val561, Gly567, Asn568, Glu571, Leu630, Asp641
	Lenvatinib	–8.7	Leu484, Val492, Ala512, Lys514, Glu531, Met535, Val559, Val561, Ser565, Leu630, Asp641
	Co-crystallized ligand Erdafitinib	–8.7	Leu484, Val492, Ala512, Lys514, Met535, Ile545, Val559, Val561, Ala564, Leu630, Ala640, Asp641, Phe642
5.BRAF (5ITA)	Test compound Alpinumisoflavone	–9.8	Ile463, Val471, Ala481, Lys483, Leu505, Ile527, Thr529, Trp531, Phe583
	Standard inhibitors Regorafenib	–10.5	Ala481, Lys483, Leu505, Ile463, Ile527, Gln530, Trp531, Cys532, Ser536, Phe583
	Sorafenib	–10.3	Ile463, Ala481, Lys483, Leu505, Ile527, Trp531, Cys532, Ser536, Phe583
	Co-crystallized ligand 6DC	–11.2	Ile463, Ala481, Val471, Lys483, Leu505, Leu514, Trp531, Cys532, Phe583, Phe595, Gly596
6.FGFR4 (4UXQ)	Test compound Alpinumisoflavone	–8.7	Asp516, Ser519, Glu520, Val523, Ile527, Ile533, Leu603, Cys608, His610, Ile628, Asp630
	Standard inhibitors Regorafenib	–11.2	Leu473, Val481, Ala501, Glu520, Met524, Ile527, Ile534, Val550, Leu619, Ile628, Ala629, Asp630, Phe631,
	Lenvatinib	–8.2	Ser519, Glu520, Val523, Leu603, His610, Ile628, Asp630, Arg635, Ile640
	Co-crystallized ligand Ponatinib	–12.6	Leu473, Val481, Ala501, Lys503, Glu520, Met524, Ile527, Ile534, Val550, Cys552, Ala553, His610, Leu619, Ile628, Ala629, Asp630,
7.PAI-1 (4AQH)	Test compound Alpinumisoflavone	–8.7	Tyr37, Leu75, Arg76, Tyr79, Thr93, Phe117, Arg118
	Co-crystallized ligand TB71380	–8.6	Tyr37, Leu75, Tyr79, Asp95, Phe117, Arg118
8.IGF1-R (2OJ9)	Test compound Alpinumisoflavone	–8.4	Val983, Ala1001, Lys1003, Asp1056, Met1112, Ile1130
	Standard inhibitors Ceritinib	–8.6	Leu975, Gln977, Val983, Ala1001, Lys1003, Met1112, Met1126
	Linsitinib	–8.4	Leu975, Gly976, Val983, Ala1001, Lys1003, Met1049, Asp1056, Ser1059, Met1112
	Co-crystallized ligand BMI	–8.9	Leu975, Gln977, Val983, Lys1003, Asp1056, Met1112, Met1126, Ile1130
9.FLT3 (4XUF)	Test compound Alpinumisoflavone	–8.3	Leu616, Val624, Ala642, Lys644, Val675, Phe691, Tyr693, Cys694, Leu818, Cys828, Phe830
	Standard inhibitors Sorafenib	–10.1	Leu616, Val624, , Lys644, Glu661, Met664, Ile674, Val675, Phe691, Tyr693, Cys694, Leu802, His809, Leu818, Ile827, Cys828, Asp829
	Sunitinib	–7.4	Leu616, Val624, Ala642, Cys694, Leu818, Phe830
	Co-crystallized ligand Quizartinib	–11.0	Leu616, Val624, Ala642, Glu661, Met664, Met665, Val675, Phe691, Cys694, Cys695, Leu802, His809, Leu818, Cys828, Asp829, Phe830

AIF has the greatest in silico binding affinity on HER2 (-10.9 kcal/mol) among the key angiogenic molecules tested in this current study, together with stronger binding affinity with MMP-9 (-11.4 kcal/mol) and VEGFR-2 (-11.0 kcal/mol) and comparable binding affinity with RET (-10.9 kcal/mol) as obtained in the previous in silico investigation of AIF against angiogenic molecules [17]. Therefore, these were chosen for the in vitro investigation of AIF’s inhibitory activity through enzyme-based assays. Since the angiogenic enzymes next to the in silico ranking of RET or HER2 were not available at the time of the experiment, the EGFR (-9.4 kcal/mol) and FGFR4 (-8.7 kcal/mol) were tested instead.

Insights on the binding of AIF on HER2

AIF has shown comparative binding on HER2 with the standard inhibitors (Fig. 2a). Notably, out of eight AIF interactions with the HER2 active site, seven were observed with neratinib and five with lapatinib (Table 2).

Fig. 2.

a Superimposition of the docked AIF (violet) and docked standard inhibitors (green) in the HER2 active site. b 2D map interaction (right) diagram of the docked Alpinumisoflavone in HER2 active site

Hydrophobic bonds govern the binding of AIF to HER2 (Fig. 2b). It includes multiple pi-alkyl interactions with Leu726, Val734, Ala751 (2 pi-alkyl bonds), Lys753, Cys805, Leu852, and Phe864. In addition, AIF formed alkyl interactions with Val734, Lys753 (2 alkyl bonds), and Leu796. Lastly, Pi-sigma interactions with Val734 and Leu852.

Mainly, the AIF comes in contact with the glycine-rich nucleotide phosphate-binding loop (Leu726, Val734) at the ATP binding pocket, phosphate/hydrophobic I pocket (Lys7530), phosphate/hydrophobic II pocket (Cys805), and DGF (aspartic acid–glycine–phenylalanine) motif region activation loop (Phe864) of the active site of HER2 tyrosine kinase. Since the interactions at the binding pocket are necessary to increase affinity for HER2 inhibitors [31, 32], the predicted interactions of AIF with the abovementioned key amino acid residues indicate AIF’s potential to inhibit the angiogenic molecule.

In vitro inhibitory activity assay on angiogenic proteins

When compared to the negative control (1% DMSO), AIF at all concentrations tested showed significant inhibition (p =  < 0.001) of HER2 (IC₅₀ = 2.96 µM), MMP-9 (IC₅₀ = 23.00 µM), VEGFR-2 (IC₅₀ = 4.80 µM), RET (IC₅₀ =  > 200 µM), and EGFR (IC₅₀ = 92.06 µM) activity (Fig. 3a–e) while a significant inhibition for FGFR4 (IC₅₀ = 57.65 µM) activity was observed at 6.25 µM-200 µM (Fig. 3f). Notably, AIF at 12.5 µM-200 µM has comparable HER2 inhibitory activity with (p =  > 0.001) with the positive control 10 µM lapatinib.

Fig. 3.

Inhibitory activity of AIF on angiogenic proteins. a HER2 tyrosine kinase, b MMP-9 gelatinase, c VEGFR-2 tyrosine kinase, d RET tyrosine kinase, e EGFR tyrosine kinase and, f FGFR4 tyrosine kinase. The data were represented as mean ± SD, n = 3 for kinases, n = 6 for gelatinase. *Denotes a significant difference against the negative control group (1% DMSO); ** Denotes a significant difference against the positive control group (p < 0.001) by One-way ANOVA and Tukey’s Test

Despite the good binding energy of AIF with MMP-9 in the in silico docking, the high IC₅₀ in the in vitro assay could be due to its potential lack of binding interaction with the MMP-9’s active site catalytic Zn²⁺ ion, as observed in the docking analysis [33, 34]. Nevertheless, the absence of in silico interaction of AIF with Zinc could be an advantage in avoiding severe adverse effects such as musculoskeletal syndrome (MSS) associated with a lack of inhibitor specificity to homogenous MMPs. Because the metal site is the most conserved feature in all MMPs, recent innovative approaches to MMP inhibitor selectivity emphasize the necessity of limiting or even preventing interactions with the catalytic Zn²⁺ [29, 35]. The FDA has approved only one MMP inhibitor, the periostat, a broad-spectrum MMP inhibitor with IC₅₀ = 2 μM–50 μM [33] as most drug candidates failed in clinical trials due to a lack of selectivity for homogenous MMPs, causing adverse reactions like MSS [36, 37]. Further experimentation is essential to establish the non-ZBG MMP-9 inhibitory activity of AIF.

For the VEGFR-2 (Fig. 3c), AIF at 200 µM with a percentage inhibition of 98.13 ± 1.52% showed comparable (p = 1.000) activity with the positive control 10 µM regorafenib. The good binding energy and ability of AIF to interact with the key amino acid in the ATP binding pocket, neighboring region, and gatekeeper of the adjacent allosteric hydrophobic back pocket of the VEGFR-2 active site in the molecular docking provides insight into AIF’s good inhibitory activity on VEGFR-2 [38].

Regardless of the high binding energy of AIF with RET in the in silico molecular docking, in vitro results (Fig. 3d) showed weak inhibitory activity. The highest concentration tested, 200 µM, exhibited 43.64 ± 0.77% inhibition only. In previous x-ray studies, Ala807 was a critical residue for the inhibitory activity of vandetanib, which binds within the RET’s ATP binding pocket [39, 40]. In our previous docking study, Ala807 interaction was observed with vandetanib, sorafenib, regorafenib, lenvatinib, cabozantinib, vandetanib, and sunitinib [17]. These compounds are FDA-approved antiangiogenic agents (multi-kinase inhibitors) targeting RET to treat cancer [40].

AIF at 200 µM with a percentage inhibition of 63.75 ± 3.02% showed comparable (p = 0.027) activity with the positive control 10 µM gefitinib against EGFR (Fig. 3e). In the previous molecular docking study, AIF interacted in silico with the glycine-rich nucleotide phosphate-binding loop (Leu718, Val726) of the N-lobe and other key amino acid residues (Ala743, Lys745, Leu788, Leu844, Phe997) of the EGFR active site [31], providing insight into AIF’s inhibitory activity.

The 200 µM AIF with a percentage inhibition of 70.13 ± 4.54% has comparable activity (p = 0.018) with the positive control 10 µM lenvatinib against FGFR4 (Fig. 3f). AIF’s good binding energy and ability to interact with the key amino acid (Table 2) in the αC helix (Glu520), activation loop (Ile628, Asp630), and other key residues (Asp516, Ser519, Val523, Ile527, Ile533, Leu603, Cys608, His610) of the FGFR4 active site [41] provides insight into AIF’s inhibitory effect.

Overall, findings showed that AIF displayed a concentration-dependent inhibition on the tested angiogenic molecules. In vitro, enzyme-based assay results substantiate the in silico molecular docking results of AIF as an inhibitor of MMP-9, VEGFR-2, RET, HER2, EGFR, and FGFR4, which are pivotal stimuli of angiogenesis.

The proto-oncogene RET and HER2 promote angiogenesis through its downstream signaling pathways Ras, PI3K, and mTOR that increase the release of growth factors (VEGF,EGF,FGF), cytokines, and extracellular matrix proteases (MMP-9) [42, 43]. MMP-9 degrades the ECM and basement membrane components, releasing the sequestered angiogenic growth factors, mainly VEGF, set to bind with their respective receptors. Moreover, the degradation of ECM caused by MMP-9 provides an avenue for cells to migrate and invade the adjacent tissue, highlighting MMP-9’s involvement in preparing a pre-metastatic niche [44]. While VEGF is the chief growth factor, the principal mediator of angiogenesis, and is the strongest and most potent inducer of vascularization, the binding to its primary receptor, VEGFR-2, is required for carrying out its activities in prompting the events vital for angiogenesis, such as endothelial cell migration, invasion, and tube formation [45]. Likewise, the angiogenic growth factors EGF and FGF binding to their receptors, such as EGFR and FGFR4, are necessary for vascularization [42, 46]. The EGFR regulates steps to sustain tumor cell intravasation such as cellular proliferation, chemotactic migration, invasion, and anti-apoptosis through its downstream signaling networks [47]. FGFR4 activation supports endothelial cell adherens junction stabilization, proliferation, migration, vascular integrity, and homeostasis [48].

Hence, inhibition of HER2, RET, MMP-9, VEGFR-2, EGFR, and FGFR4 activity by AIF in vitro indicates its multi-system action as an antiangiogenic compound.

In ovo CAM assay

After 48 h of treatment incubation, numerous blood vessel branch points and elongated tubule complexes were observed in the CAM of negative control (0.3% DMSO), indicating rich vascularization. In contrast, noticeable suppression was detected in AIF and celecoxib (positive control) treated groups (Fig. 4a).

Fig. 4.

a Representative images of duck embryonic chorioallantoic membrane (CAM) after being treated with various concentrations of alpinumisoflavone (AIF), positive control (celecoxib), or negative control (0.3% DMSO) for 48 h. Scale bar: 2 mm. White arrow indicates the main blood vessel while Black arrow indicates the branch point (vessel junction). Percentage inhibition on b the total number of branch points and c the mean length of tubule complexes relative to the negative control (0.3% DMSO) measured by AngioQuant™. The data were represented as mean ± SD, n = 6. *Denotes a significant difference against the negative control group (0.3% DMSO); ** Denotes a significant difference against the positive control group (celecoxib) (p < 0.001) by One-way ANOVA and Tukey’s Test

Quantitative measurement using Angioquant™ image analysis software showed that 3.125 μM-200 μM AIF treated CAM exhibited a significant decrease (p =  < 0.001) in the total number of branch points (Fig. 4b) and mean length of tubule complexes (Fig. 4c), respectively, as compared with the negative control, suggesting a wide range of effective concentration.

The inhibitory activity was more evident with the increase in AIF concentration. This concentration-dependent antiangiogenic activity has also been identified with other flavonoids and their subclasses, including the isoflavones [49]. Even at low concentrations, AIF inhibited 50% of the total number of branch points (IC₅₀ = 14.25 µM) and the mean length of tubule complexes (IC₅₀ = 3.52 µM), suggesting AIF’s potency in suppression of neovascularization in ovo. Also, AIF inhibited new embryonic blood vessel formation without any sign of hemolysis or toxic effects.

Isoflavones are among the most active antiangiogenic flavonoids, demonstrating a reduction of blood vessels in CAM assay for 64% to 80% at 100 µM [49]. In this study, 100 µM AIF presented 68.14 ± 3.10% inhibition of the total number of branch points and 87.79 ± 1.21% inhibition of the mean length of tubule complexes of CAM relative to the negative control. The antiangiogenic activity of AIF is attributed to the apparent critical structural features such as hydroxyl (-OH) groups at C-5 and C-4’, the unsaturation of the C-2 & C-3 bond, and carbonyl (C = O) moiety at C-4 [49]. Furthermore, the cyclic prenyl moiety attached to ring A of AIF is responsible for its lipophilicity, enabling it to diffuse readily on the CAM [50].

In addition, the total number of branch points (Fig. 4b) and the mean length of tubule complexes (Fig. 4c) of duck CAM treated with 25 μM-200 μM AIF showed comparable (p =  > 0.001) activity to the positive control 200 μM celecoxib. Several studies support the use of celecoxib as a positive control in CAM assay because it can inhibit COX-2, VEGF-A, VEGFR-2, MMP-9, HER2, and EGFR [51–54]. Present results show that AIF can suppress blood vessel formation in the embryo of fertilized duck eggs in ovo.

In vitro MTT assay for cell viability

AIF inhibited the growth of MCF-7 in a concentration-dependent manner (Table 3). However, inhibition did not reach more than 50% for all tested concentrations (Fig. 5). As expected, the positive control doxorubicin significantly decreased the viability of MCF-7 (24 h IC₅₀ = 7.05 µM and 48 h IC₅₀ = 3.62 µM) and HDFn (48 h IC₅₀ = 27.16 µM).

Table 3.

The percentage inhibitory activity of alpinumisoflavone (AIF) and doxorubicin on the viability of MCF-7 and HDFn cells after 24 or 48 h of treatment incubation

Concentrations	MCF-7				HDFn
	AIF		Doxorubicin		AIF	Doxorubicin
	24 h	48 h	24 h	48 h	48 h	48 h
3.125 µM	14.37 ± 3.86	22.10 ± 3.18	44.594 ± 2.29	52.13 ± 2.23	-40.33 ± 9.85	-40.21 ± 4.80
6.25 µM	17.42 ± 4.80	23.99 ± 2.82	46.995 ± 1.32	56.13 ± 1.23	-30.01 ± 6.51	-27.78 ± 2.82
12.5 µM	22.80 ± 3.49	26.06 ± 2.47	50.982 ± 0.80	58.50 ± 1.28	-16.65 ± 6.89	-2.46 ± 2.64
25 µM	27.63 ± 4.03	33.34 ± 4.57	51.888 ± 1.62	59.65 ± 0.30	-8.56 ± 2.99	5.51 ± 5.10
50 µM	30.01 ± 4.68	40.96 ± 1.58	60.376 ± 2.98	66.37 ± 0.43	3.40 ± 2.64	65.18 ± 4.15
100 µM	42.30 ± 2.72	44.92 ± 1.79	85.467 ± 0.99	86.66 ± 0.93	20.87 ± 6.33	92.97 ± 1.27

Fig. 5.

Effects of alpinumisoflavone (AIF) on the viability of MCF-7 and HDFn. Both cells (1 × 104 cells/well) were treated with AIF at the indicated concentrations, and the percentage inhibition of cell viability was determined by MTT assay after 24 h or 48 h. The data were represented as mean ± SD, of three trials in triplicates for MCF-7 and one trial in triplicates for HDFn. *Denotes a significant difference against the negative control group (0.3% DMSO) (p < 0.001) by Oneway ANOVA and Tukey’s Test

AIF significantly inhibited (p =  < 0.001) the viability of MCF-7 compared with the negative control (0.3% DMSO) in 24 h (6.25 µM-100 µM) and 48 h (3.125 µM-100 µM) of treatment incubation. The highest concentration-100 µM, showed 42.30 ± 2.72% and 44.92 ± 1.79% mean percentage inhibition after 24 h and 48 h of AIF treatment, respectively.

AIF has been toxic against the more aggressive and challenging type of breast cancer, invasive ductal carcinoma, accounting for 70–80% of all invasive breast cancer diagnoses [55, 56]. The previous study of Kuete et al. (2016) [12] on MDA-MB-231-pcDNA and multi-drug resistant MDA-MB-231-BCRP basal type human breast cancer cell lines showed higher susceptibility on AIF with an IC₅₀ = 42.57 ± 3.81 µM and IC₅₀ = 65.65 ± 6.04 µM, respectively. The different features of the MCF-7 (hormonal therapy model) and MDA-MB-231 (chemotherapy model) could explain their differential response to AIF. The non-invasive MCF-7 is an ER + , Progesterone Receptor (PR) positive, and HER2 negative (HER2-) breast cancer model, while the invasive MDA-MB-231 is a triple-negative breast cancer model characterized by ER-negative, PR-negative, and HER2- expression [57, 58] Previous findings on AIF support the reports that AIF can inhibit the growth and development of breast cancer both via estrogen-dependent and -independent signaling, but its action is highly selective. Accordingly, data obtained on AIF’s activity support its potential as a systemic anticancer agent (Hormonal therapy, Chemotherapy, Targeted Therapy) for several types of breast cancer.

AIF displayed a weak cytotoxic activity against HDFn with an IC₅₀ =  > 100 μM. Interestingly, AIF exhibited a biphasic effect on HDFn as it induced proliferation at low concentrations (3.125 µM-25 µM) (Fig. 5) but decreased viability at high concentrations (50 µM-100 µM). This biphasic dose–response in cultured cell viability showing an inhibitory effect at high concentrations and a stimulatory effect at low concentrations has been observed in many compounds. However, most are with phytoestrogen compounds such as isoflavones [59]^. A report on genistein (1 µM-100 µM), the most studied isoflavone, showed a biphasic effect in HDFn cells [60]. Often, a 30–60% increase over control is observed as the maximum stimulation in the biphasic response [59], which we observed in the lowest concentration tested on AIF (3.125 µM), showing a proliferative effect of 40% relative to the control (Fig. 5). One of the proposed mechanisms of the phytoestrogen compounds includes the estrogen receptor pathway being involved in the proliferative effect at lower concentrations. Another mechanism is linked to antioxidant properties at low concentrations and pro-oxidant properties at higher ones, influencing cell growth [60]. We hypothesized that the proliferative effect of AIF at low concentrations is due to its isoflavone structure and attributes.

This finding in HDFn is relevant because dermatologic conditions are among the most frequently encountered side effects in patients receiving systemic cancer therapy and radiation therapy [61, 62]. Besides, Lee et al. (2021) [9] study reported a protective effect of AIF against skin damage. Accordingly, a compound that presents good anticancer properties with less toxicity on skin cells may alleviate the unwanted effects and help improve patients’ quality of life.

Amen et al., 2013 [63] reported that AIF at 10 µM inhibited MCF-7 by 17.18% after 48 h. In this study, AIF exhibited higher activity of 22.10 ± 3.18% inhibition at 3.125 µM and 26.06 ± 2.47% at 12.5 µM. The discrepancies between results may be attributed to several factors, such as the differences in the assay method used (Amen et al., 2013 utilized Sulforhodamine B (SRB) assay while MTT assay was used in this present study), the tested compound’s purity, and the experimental setup, or other experimental factors [50]. However, despite the difference in results, recorded data agree on the cytotoxic property of AIF on MCF-7 breast cancer cells.

Despite the good in vitro inhibitory activity of AIF against the angiogenic molecules, it could not entirely inhibit the growth of the MCF-7 breast cancer cells. Antiangiogenic drugs do not directly block the growth of tumor cells themselves. Instead, it hinders the development of new blood vessels that support tumor growth and metastasis [64]. The cytotoxic mechanism of antineoplastic drugs such as doxorubicin is different from the proposed mechanisms of most FDA-tested antiangiogenic agents. In this case, most antiangiogenic drugs are used as an adjuvant to other anticancer treatments, such as chemotherapy, surgery, and radiation therapy [42]. In addition, single-targeted pathways are thought not to be sufficiently effective given the complexity of the disease-pathway of cancer [65]. Nevertheless, AIF has the potential to be an adjunct to other cancer treatments as suggested by significant inhibition of angiogenic proteins by AIF with moderate inhibition of breast cancer cell MCF-7.

In silico rat oral LD₅₀ prediction

A computed rat oral LD₅₀ of 146.4 mg/kg with lower and upper 95% confidence limits of 20.1 mg/kg and 1.1 g/kg was obtained (Table 4). Based on the Hodge and Sterner Scale for toxicity [66], AIF could be moderately toxic (Rat Oral LD₅₀ 50–500 mg/kg) to slightly toxic (Rat Oral LD₅₀ 500–5000 mg/kg), and its probable Lethal Dose for Human ranges from 30 mL (moderately toxic) to 600 mL (slightly toxic). Further, in vivo investigations are recommended to confirm the predicted doses obtained in this study, mainly to test above 100 mg/kg AIF. The highest recorded animal dosing studied was 100 mg/kg in mice [11].

Table 4.

In silico Rat Oral LD₅₀ profile of alpinumisoflavone (AIF)

Model: Rat Oral LD50 (v3.1-FDA; DS 4.0 software)	AIF
Computed Rat Oral LD50 Log (1/Moles) Computed Rat Oral LD50 Lower 95% Confidence Limits Upper 95% Confidence Limits	3.361 146.4 mg/kg 20.1 mg/kg 1.1 g/kg

Previous animal studies showed that AIF elicited its target activity effectively at lower doses ranging from 20–100 mg/kg/day while presenting a good safety profile [11, 14–16]. Administration of 20–50 mg/kg/day of AIF for 24–56 days in the in vivo mice animal models showed no adverse effects or marked alteration in the histology of cardiac, pulmonary, renal, splenic, and hepatic tissues, indicating that it was well tolerated [14, 15]. Treatment with AIF from 20–100 mg/kg/day for 24–30 days did not cause significant differences (p < 0.01) in the body weight of the treated mice, proposing a good safety profile of the AIF in vivo [11, 16].

In silico ocular irritancy

The non-ocular irritant potential of AlF, as presented in Table 5, is favorable, particularly if given a chance to be developed as a major active ingredient in an eye formulation. The success of the eye injectable antiangiogenesis drugs Aflibercept, Pegaptanib sodium, Bevacizumab, and Ranibizumab for the treatment of wet macular degeneration, an angiogenesis dependent disease and one of the leading causes of blindness, has been encouraging scientists to develop a non-invasive eye drop preparation to improve patient adherence [67, 68].

Table 5.

In silico Ocular Irritancy profile of alpinumisoflavone (AIF)

TOPKAT (Toxicity Prediction by Komputer Assisted Technology) (DS 4.0)	Result	AIF (Description)
Model: Ocular Irritancy SEV/MOD vs MLD/NON (v5.1)	0.000*	Non-ocular irritant
Model: Ocular Irritancy SEV vs MOD (v5.1)	0.016*	Non-ocular irritant
Model: Ocular Irritancy MLD vs NON (v5.1)	0.004*	Non-ocular irritant

TOPKAT values: 0–0.29: low probability; 0.30–0.69; indeterminate; 0.70–1.00: high probability. *Within OPS- The optimum prediction space (OPS) is unique multivariate descriptor space in which the model is applicable. Assessment of this is needed to determine if the chemical structure being examined is within the OPS of a model; thus the probability results may be accepted with confidence, subject to the results obtained from hypothesis testing

In silico drug-likeness profiling

The predicted high gastrointestinal absorption and compliance to Lipinski’s, Veber’s, Ghose’s, Egan’s, and Muegge’s Rule of the AIF (Table 6) theoretically propose it can be active after oral administration. Based on the computational prediction of druggability, AIF presented zero (0) violations on the five different rule-based filters. The overall conformity of AIF to the rules fortifies its drug-like property and would be beneficial in oral drug delivery formulation. Foremost, patients prefer oral dosage forms due to ease of administration and cost-effectiveness, resulting in higher patient compliance and improved health outcomes [69].

Table 6.

In silico Drug-likeness profile of alpinumisoflavone (AIF)

Lipinski's (Pfizer) Rule (SwissADME software)	Acceptable range	AIF (Result)
Molecular Weight (g/mol)	≤ 500	336.34 g/mol
Lipophilicity (MLogP)	≤ 5	1.64
Hydrogen bond donors (NH or OH)	≤ 5	2
Hydrogen bond acceptors (N or O)	≤ 10	5
Rule Violation	≤ 1	0
MOLECULAR DRUG-LIKENESS	YES, 0 violation
Veber’s (GSK) Rule (SwissADME software)	Acceptable Range	AIF (Result)
Polar Surface Area	≤ 140 Å2	79.90 Å2
Number of Rotatable Bonds	≤ 10	1
Rule Violation	0	0
MOLECULAR DRUG-LIKENESS	YES, 0 violation
Ghose’s (Amgen) Rule (SwissADME software)	Acceptable Range	AIF (Result)
Molecular Weight	160–480	336.34 g/mol
Lipophilicity (WlogP)	-0.4–5.6	3.95
Molecular Refractivity	40–130	96.09
Number of atoms	20–70	41
Rule Violation	0	0
MOLECULAR DRUG-LIKENESS	YES, 0 violation
Egan’s (Pharmacia) Rule (SwissADME software)	Acceptable Range	AIF (Result)
Lipophilicity (WlogP)	≤ 5.88	3.95
Polar surface area	≤ 131.6 Å2	79.90 Å2
Rule Violation	0	0
MOLECULAR DRUG-LIKENESS	YES, 0 violation
Muegge’s (Bayer) Rule (SwissADME software)	Acceptable Range	AIF (Result)
Molecular Weight	200–600	336.34 g/mol
Lipophilicity (XlogP)	-2.0–5.0	3.94
Polar Surface Area	< 150 Å2	79.90 Å2
Number of rings	< 7	4
Number of carbons	> 4	20
Number of heteroatoms	> 1	5
Number of Rotatable Bonds	< 10	1
Hydrogen bond donors (NH or OH)	< 5	2
Hydrogen bond acceptors (N or O)	< 10	5
Rule Violation	0	0
MOLECULAR DRUG-LIKENESS	YES, 0 violation

Considering the in silico results, AIF might be categorized under Biopharmaceutical Classification System (BCS) II, a compound with low aqueous solubility but high permeability. According to Nikolakakis and Partheniadis [70], 30% of drugs present in the market were under BCS II, comprising the majority (60%-70%) of the drug molecules currently under development. The lipophilic molecules are primarily selected as potential therapeutic candidates for development based on their ability to attach to cell receptors, mainly involving hydrophobic interactions. Correlating with the in silico molecular docking studies, the AIF’s lipophilic nature explains the dominant hydrophobic interactions against the MMP-9, VEGFR-2, RET, HER2, EGFR, and FGFR4. Overall, the in silico predicted AIF’s non-carcinogenicity, non-mutagenicity, and non-toxicity to a fetus are noteworthy for further investigations or evaluations using appropriate animal models.

Conclusion

AIF inhibited HER2, MMP-9, VEGFR-2, RET, EGFR, FGFR4, and HER2 in vitro, which confirmed the significant bindings or interaction of AIF to these angiogenic proteins in silico. In ovo CAM assay revealed that AIF significantly inhibited the total number of branch points and mean length of tubule complexes of duck CAM, having comparable inhibitory activity with the positive control celecoxib on both parameters. AIF significantly inhibited the viability of MCF-7 breast cancer cells compared with the negative control (0.3% DMSO) while displaying a biphasic dose–response in cultured HDFn cells showing an inhibitory effect at high concentrations and a stimulatory effect at low concentrations. Based on the computational prediction of druggability, AIF presented zero (0) violations on the five different rule-based filters. The overall conformity of AIF to the rules fortifies its drug-like property and would be beneficial in oral drug delivery formulation. The acceptable safety, drug-likeness profiles, and multi-kinase inhibitory properties of AIF are worthy of further validation in animal models and other pre-clinical studies to be elevated as a promising compound for the development of a multi-kinase inhibitor for ER + breast cancer cells.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors contribution

All authors contributed to the study’s conception and design. The grant application, material preparation, data collection, data analysis, and interpretation of results were performed by Ross D. Vasquez, Agnes L. Castillo, and Honeymae C. Alos. The first draft of the manuscript was written by Honeymae C. Alos, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The study was supported by the Department of Science and Technology-National Research Council of the Philippines under the Institutional Grant for Invigorating Basic Research on Health Sciences Phase II Research Program (DOST-NRCP Project No. Q-006).

Code availability

Not Applicable.

Declarations

Ethics approval

This article does not contain any studies with human participants or animals performed by any authors.

Informed consent

The manuscript does not contain clinical studies or patient data.

Conflicts of interest

The authors have no relevant financial or non-financial interests to disclose.