Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

Contents lists available at ScienceDirect

Pharmacological Research - Modern Chinese Medicine

journal homepage: www.elsevier.com/locate/prmcm

Network pharmacology-based elucidation of bioactive compounds in

propolis and putative underlying mechanisms against type-2 diabetes
mellitus✰
Emmanuel I. Ugwor a,b,∗, Adewale S. James c, Adekunle I. Amuzat b, Emmanuel O. Ezenandu b,
Victory C. Ugbaja b, Regina N. Ugbaja b
a
Department of Biochemistry and Molecular Biology, Federal University of Viçosa, Viçosa, Minas Gerais, Brazil
b
Department of Biochemistry, College of Biosciences, Federal University of Agriculture, Abeokuta, Ogun State, Nigeria
c
Department of Chemical Sciences, Faculty of Science, Augustine University, Ilara-Epe, Lagos State, Nigeria

a r t i c l e i n f o a b s t r a c t

Keywords: Propolis is a common remedy in Chinese medicine. We have previously demonstrated the anti-diabetic poten-
Nigerian propolis tials of Nigerian propolis. However, the underlying mechanisms against type-2 diabetes mellitus (T2DM) remain
Network pharmacology unclear. This study used network pharmacology-based analysis to unravel the possible mechanisms of NP’s anti-
Type-2 diabetes
T2DM activity. Previously identified compounds in NP were retrieved from literature, screened for drug-likeness,
Bioactive compounds
and used to retrieve targets associated with T2DM, from which compound-target, protein-protein interaction,
and target-pathways networks were constructed. Network pharmacology-based and enrichment analyses were
conducted on the networks. Further, NP’s inhibitory activity against targets identified in network analyses was
assessed in-vitro. Library search revealed 202 previously reported compounds in NP; 96 were retained after
screening for drug-likeness. However, only 44 NP compounds interacted with T2DM-associated targets, with
2-ethyl-1,4-dimethoxybenzene as the most active component. Network analyses revealed 167 putative targets,
with HSP90AA1, JUN, ESR1, STAT3, CYP3A4, PTGS2, RELA, VEGFA, CYP2C9, and PPARG as the core anti-T2DM
targets of NP compounds. Gene ontology analyses indicated that these targets were predominantly localised in
the plasma membrane and cytoplasm and primarily involved in regulatory, signal transduction and cellular re-
sponse processes. KEGG pathway enrichment implicated metabolic pathways (involving lipids), AGE-RAGE, and
calcium/cAMP, among others, in the anti-T2DM effects of the compounds. Furthermore, in vitro pharmacologi-
cal assessment demonstrated appreciable inhibitory effects of 2-ethyl-1,4-dimethoxybenzene against 𝛼-amylase,
𝛼-glucosidase, and HMG-CoA reductase. This study provides holistic mechanistic insights into the anti-T2DM
activities of the constituents of Nigerian propolis.

1. Introduction diseases, such as obesity and overweight. It is also a risk factor for heart
disease – the leading cause of death globally [5].
Type-2 diabetes mellitus (T2DM) is one of the four major non- Although T2DM prognosis has improved and several anti-T2DM ther-
communicable diseases that has continually afflicted the global popu- apies have been formulated over the years, its prevalence continues to
lace and has been hallmarked as the epidemic of the millennium [1,2]. surge [6]. Pharmacological agents, such as sulfonylureas, glucagon-like
Globally, the prevalence of T2DM has increased by 49% since 1990. In peptide1 agonists, HMG-CoA reductase inhibitors, and 𝛼-glucosidase
2019, there were 437.9 million reported cases of T2DM, with the highest inhibitors, and lifestyle modification have been the mainstream anti-
proportion in low- and middle-income countries [3]. T2DM is a chronic T2DM therapies [6]. However, these agents are limited by adverse side
metabolic condition brought on by impaired insulin production by pan- effects and high costs. Besides, they often target one aspect of the other-
creatic beta cells and insulin resistance in peripheral insulin-sensitive wise multifaceted complications associated with T2DM [7]. Conversely,
organs [4]. The aetiology of T2DM stems from various environmental, natural products, encompassing herbal remedies, possess few side ef-
genetic and epigenetic factors and is associated with several metabolic fects and have increasingly been exploited by researchers in developing
therapeutic agents against T2DM [8].

✰
Type of Paper: Original article.
∗
Corresponding author at: Department of Biochemistry, College of Biosciences, Federal University of Agriculture, Abeokuta, Ogun State, Nigeria.
E-mail address: [email protected] (E.I. Ugwor).

https://doi.org/10.1016/j.prmcm.2022.100183
Received 28 August 2022; Received in revised form 12 October 2022; Accepted 20 October 2022
Available online 21 October 2022
2667-1425/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

Propolis is one of the natural products investigated for its efficacy 2.3. Compounds- and disease-associated targets
against T2DM [9–12] and is a common remedy in Chinese medicine
[9]. It is a resinous substance produced by honey bees from mixing their On the one hand, the SMILES of the screened compounds
salivary and enzymatic secretions with plant sap, buds, gums, and other were used to retrieve linked target genes from the Similarity
vegetal materials [13]. Propolis is abundant in bioactive compounds, Ensemble Approach (SEA; http://sea.bkslab.org/) and (Swiss Tar-
such as flavonoids, terpenes, aromatic acids and their esters, which en- get Prediction (STP; http://www.swisstargetprediction.ch/) databases
dow it with an extensive range of biological properties, including an- [23]. On the other hand, using the Homo sapiens mode and
tioxidant, antibacterial, anti-inflammatory, anti-diabetic and anticancer “type 2 diabetes mellitus” as the keyword, T2DM-linked targets
properties [11,12]. However, the chemical constituent of propolis is were recovered from the DisGeNeT (https://www.disgenet.org/),
highly divergent, depending heavily on the bee species, the flora sur- Malacards (https://www.malacards.org/), and Kyoto Encyclopedia
rounding the collection site, and geographical zones [13]. In Nigeria, of Genes and Genomes (KEGG; https://www.kegg.jp) databases
propolis has purportedly been used to treat ringworm, chickenpox, and [18]. Subsequently, overlapping targets between the compound-
measles and manage diabetes [12]. We have previously demonstrated associated and ND-associated targets were identified using VENNY 2.1
the anti-diabetic potentials of Nigerian propolis (NP), including hypo- (https://bioinfogp.cnb.csic.es/tools/venny/) and used for network con-
glycemic, hypolipidemic, and hepatoprotective properties, in rats with struction.
streptozotocin-induced diabetes [13]. Other studies have reported sim-
ilar anti-diabetic properties of NP [10,12,15]. However, the underlying 2.4. Network construction and analyses
mechanisms of NP anti-T2DM effects have not been elucidated.
The varied bioactive components in natural products and their po- Three networks – compound-target (CT), protein-protein interaction
tential to act synergistically on multiple targets pose limitations to elu- (PPI), and target-pathways (TP) networks) – were constructed using Cy-
cidating their mechanisms of action [16]. However, recent advances in toscape 3.9.1 software (http://www.cytoscape.org/) to visualise the in-
computer-aided drug designs, such as network pharmacology, now al- teraction between the screened bioactive compounds, selected targets,
low a holistic view of the multicomponent-multi target paradigm of nat- and associated pathways [24]. First, the screened compounds and the
ural products, thus revealing the putative mechanisms of action against associated overlapping targets were used to construct the CT network.
different diseases via interacting targets and associated pathways [17– Next, the overlapping targets were uploaded onto the STRING database
19]. Network pharmacology is a scope in systems biology that visualises (Version 11.5; (https://string-db.org/) to predict the PPI [25]. Protein
the interactions of the bioactive compound(s) with both the interactome interactions with confidence level > 0.7 within the Homo sapiens mode
and the diseasome, thus heightening the understanding of the multi- were then used to construct the PPI network in Cytoscape. Core com-
target mechanism of the phytochemical constituents of natural products pounds and targets were identified using the degree algorithm of the
via a compound–target–pathway-disease network constructs [16,19]. CytoHubba plugin in the Cytoscape to analyse the CT and PPI networks,
Therefore, this study utilised a network-based pharmacology anal- respectively. Further, the nodes in the PPI network were aggregated into
ysis to unravel the possible mechanisms of NP anti-T2DM activity and clusters using the MCODE plugin within Cytoscape with the following
identify responsible bioactive components. In addition, the inhibitory parameters: Degree cutoff = 2, Node Score Cutoff = 0.2, and K-Core = 2.
activity of NP’s bioactive component against some targets identified was Information about each cluster was fetched directly from the STRING
also evaluated. database using its application programming interface domiciled in Cy-
toscape.
The overlapping targets were mapped onto the Comparative Tox-
2. Methods icogenomics Database (CTD; http://ctdbase.org/tools/) to extract the
canonical pathways highly associated with these proteins. Within CTD,
2.1. Assembling of compounds database for Nigerian propolis gene inputs identify pathways involving these genes based on annota-
tion from the KEGG and REACTOME databases. Non-specific and non-
A database of ligands in Nigerian propolis was compiled from pre- diabetes-related terms were removed, and the TP network was con-
viously reported works conducted on the characterisation of propolis structed in Cytoscape to present a global view of the interactions be-
within Nigeria. The canonical SMILES of each compound were retrieved tween targets and associated pathways.
from PubChem (https://pubchem.ncbi.nlm.nih.gov/). One compound
(riverinol) had no deposition in PubChem, so its SMILES was deduced 2.5. Gene ontology (GO) and KEGG pathway enrichment analyses
from its structure using Marvin JS plugin in SwissADME online server
(http://www.swissadme.ch/index.php). The Database for Annotation, Visualization and Integrated Discovery
(DAVID; https://david.ncifcrf.gov/tools.jsp) server was used for enrich-
ment analyses [16]. The overlapping genes were uploaded onto DAVID,
2.2. Screening of compounds the identifier was OFFICIAL GENE SYMBOL, species was set as Homo
sapiens, and information on the cellular component (CC), biological pro-
Compounds were screened based on their absorption, distribution, cess (BP), molecular function (MF), and KEGG pathways pertinent to
metabolism and excretion (ADME) properties, bioavailability score, and each target were retrieved. Only GO terms and KEGG pathways with
violations of the rules of the Lipinski et al. [20], Veber et al. [21], and a p-value of less than 0.05 and Benjamini value of less than 0.5 were
Ghose et al. [22] drug-likeness approaches, which was computed on retrieved. Within the DAVID server, the p-values were adjusted for the
the SwissADME server. Compounds with two or more violations and false discovery rate (i.e., Q-values) [23]. GraphPad Prism 9.3.1 was used
estimated oral bioavailability < 50% were excluded from further analy- to visualise the top ten most enriched components of the GO and KEGG
ses. In addition, 37 FDA-approved drugs for T2DM were from DrugBank enrichment analyses.
(https://go.drugbank.com), and their pharmacological properties were
computed with SwissADME (Supplementary File 1). Principal compo- 2.6. In-vitro pharmacological validation
nent analysis (PCA) was conducted on the calculated properties [17];
the two most significant principal components were used to compare The anti-diabetic potentials 2-ethyl-1,4-dimethoxybenzene (identi-
the chemical space between the screened compounds and drugs in a fied from network analyses as the most active NP component) were
scatterplot. evaluated by assaying for the inhibitory activity against selected targets

2
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

identified by network analyses: 𝛼-amylase, 𝛼-glucosidase, and HMG-

CoA reductase (HMGR). While 2-ethyl-1,4-dimethoxybenzene (Cat. NO.:
3452AA) was procured from Ak Scientific, Inc. (Union City, CA, USA),
the enzymes and other reagents used for the in-vitro assays were prod-
ucts of Sigma-Aldrich (St. Louis, MO, USA).

2.6.1. Anti-diabetic activity

Briefly, 100 𝜇L of the different concentrations (20, 40, 60, 80,
100 𝜇g/ml) of 2-ethyl-1,4-dimethoxybenzene or acarbose (Cat. NO.:
A8980) and 500 𝜇l of 20 mM sodium phosphate buffer (pH 6.9) contain-
ing 6 mM NaCl and pancreatic 𝛼-amylase (10 U/ml; Cat. No.: A3176)
were incubated at 25 °C for 10 min. The reaction was initiated by adding
1% starch solution (500 𝜇l, prepared in the same phosphate buffer) and
monitored at 540 nm for 10 min [26].
For the 𝛼-glucosidase assay, 100 𝜇L of the 𝛼-glucosidase solution
(Cat. No.: G5003) was allowed to equilibrate with 50 𝜇L of 2-ethyl-1,4-
Fig. 1. The distribution in chemical space according to principal component
dimethoxybenzene or acarbose (20, 40, 60, 80, 100 𝜇g/ml) at 25 °C for analysis of screened compounds in Nigerian propolis (n = 96) and FDA-approved
10 min. Then, 50 𝜇L of 5 M p-nitrophenyl-𝛼-D-glucopyranoside solution drugs (n = 37). The two most significant principal components (PC1 and PC2)
(in 0.1 M phosphate buffer; pH 6.9) was added, and the reaction was were used to compare the chemical space.
monitored at 405 nm for 5 min [27].
For HMGR, assay kit (Sigma; Cat. No.: CS1090) containing the cat-
alytic domain of the human enzyme and other reagents was used for The chemical space occupied by these screened compounds was then
the HMGR inhibitory assay, as previously described [28]. Briefly, 20 𝜇L compared to those of FDA-approved drugs using the two most signifi-
of the enzyme (0.5 mg/ml) and 10 𝜇L of 2-ethyl-1,4-dimethoxybenzene cant principal components, which explained 53.0% and 26.0% of vari-
or simvastatin (Cat. NO.: S6196) were added to the reaction mixture ances, respectively (Fig. 1). Although both groups occupied two distinct
containing 400 𝜇M NADPH and 400 𝜇M HMG-CoA substrate in a final clusters, there were clear overlaps between the two groups, further il-
volume of 2 mL of 100 mM potassium phosphate buffer, pH 7.4 (con- lustrating the drug-like potentials of the compounds.
taining 120 mM KCl, 1 mM EDTA, and 5 mM DTT). The rate of NADPH
disappearance was monitored at 340 nm for 15 min. The blanks con- 3.2. Elucidation of T2DM-associated targets of screened Nigeria propolis
tained neither 2-ethyl-1,4-dimethoxybenzene nor reference drugs. compounds
Tests were performed in triplicates, and the percentage inhibition for
each enzyme was calculated using the formula below: Network pharmacology-based analysis was used to identify T2DM-
associated targets of the screened compounds. Targets associated with
Δabsorbence blank − Δabsorbence sample each compound were retrieved from SEA (580 targets and STP (927
%Inhibition = × 100
Δabsorbence blank targets), while T2DM-associated targets (n = 3395) were retrieved from
DisGeNeT, Malacards, and KEGG databases. Comparing these three data
2.7. Data management sets, we uncovered 167 overlapping targets (Fig. 2) common to both
compounds- and disease-associated targets (detailed in Supplementary
Data retrieved or obtained in this study were managed with Mi- File 3).
crosoft Excel (2019). Graphs were graphs using GraphPad Prism 9.3.1 These common targets were then entered into the STRING database
(GraphPad Software Inc., CA, USA). GraphPad was used to fit results ob- to analyse the protein-protein interactions, which comprise direct (phys-
tained from the in-vitro analyses into a dose-response curve to determine ical) and indirect (functional) associations, obtained from computa-
IC50 values. tional prediction, inter-organismal knowledge transfer, and associations
culled from other databases [25]. The significant interactions (with a
3. Results and discussion high confidence level of >0.7) were visualised as PPI-network (Fig. 3A).
The PPI network included 135 nodes connected by 371 edges after elim-
3.1. Nigerian propolis compound database inating isolated nodes and edges, with an average node degree of 4.49.
The substantial number of interactions between the targets suggests that
A literature search revealed five previous original researches that they may be biologically related, and the anti-T2DM effects of the com-
characterised the phytochemical composition of Nigerian propolis pounds may involve multiple targets. The top 10 targets (hereafter re-
[12,29–32], from which we extracted 202 compounds, compiled in Sup- ferred to as core targets) in the PPI network were identified based on
plementary File 1. The compounds were then screened using the ADME degrees, betweenness, and closeness parameters, following the analysis
properties, three drug-likeness approaches, and projected oral bioavail- of the network with the CytoHubba algorithm (Supplementary File 4).
ability. A compound’s ADME features predict its disposition inside an The core targets were HSP90AA1, JUN, ESR1, STAT3, CYP3A4, PTGS2,
organism, contributing to its pharmacological (or toxicological) action RELA, VEGFA, CYP2C9, and PPARG (Fig. 3B); these targets accounted
[33]. The oral bioavailability indicates the fraction of the compound for 47.8% of the total protein-protein interactions (167 of 371 PPI). Be-
that will reach systemic circulation unchanged following oral admin- sides, several studies have highlighted the importance of these targets
istration, which correlates with its efficacy. Furthermore, the rules of to T2DM. Lazaro et al. [34] reported that pharmacological inhibition of
Lipinski, Ghose, and Verber are drug-likeness approaches used to iden- HSP90 ameliorated diabetes-associated complications, while Yang et al.
tify whether a chemical molecule with a particular pharmacological or [35] revealed decreased Hsp90 expression in isolated rat islets exposed
biological activity has chemical and physical qualities, such as molecu- to high glucose. They surmised that Hsp90 might be involved in high
lar weight, partition coefficient, and the number of hydrogen donors and glucose-induced islet dysfunction. Meanwhile, the cross-talk between
acceptors, which would make it a potential orally active drug [20–22]. JUN, RELA, and STAT3 has been implicated in the progression of T2DM
Only compounds meeting these criteria were used for further analyses. [36,37]. Clinical and pre-clinical studies have reported the roles of CYP,
Ninety-six compounds were included (Supplementary File 2), while 106 PTGS2, and VEGFA in diabetes [38–40]. ESR1 mediates the effects of es-
compounds violated these criteria and were excluded. trogens on glucose homeostasis [41].

3
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

eicosanoids) metabolism; cluster 2 nodes are primarily involved in cellu-

lar responses to stimuli, such as hypoxia (HIF-1), inflammation (NFKB1,
STAT3, and RELA), heat (HSP90), proliferation (VEGF, PDGFRB, PRKCZ
and PPARG), and tissue remodelling (MMPs and VEGFA). Clusters 3 and
4 were involved in signal transduction and alcohol metabolism, respec-
tively. Nodes in cluster 5 have been correlated to amyloid deposition,
while those in clusters 6 and 7 regulate adenylate cyclase/cGMP down-
stream signalling and neurotransmitter transport/activity, respectively.
Nodes in cluster 8 share phosphor-amino acids (e.g., serine, threonine,
or tyrosine) motifs and function as adapter proteins.
Interestingly, of the 96 compounds, only 44 had targets common to
T2DM-associated targets (Table 1), whereas 52 compounds had no over-
lap with T2DM targets. The interactions between overlapping targets
and the screened compounds were visualised in a CT network (Fig. 5A),
consisting of 221 nodes (44 compounds and 167 targets) and 620 inter-
actions, with an average of 3.52 targets/compound. The overlap of tar-
gets reflects the multi-target characteristics of the bioactive compounds.
The CT network was subsequently analysed with CytoHubba to identify
hit compounds (Supplementary File 4). The top 10 compounds, based
on degree, are presented in Fig. 5B. The most active compound – 2-
ethyl-1,4-dimethoxybenzene – interacted with 63 targets, followed by
borneol (32). These top compounds accounted for 73.2% of compound-
target interactions (227 of 310).
The overlapping targets were then linked to biochemical pathways
using the CTD online server; the highly significant T2DM-linked path-
Fig. 2. Venn diagram illustrating the overlapping targets (central) between ways were visualised as a network with the targets and associated path-
compound-associated disease targets (CAT) obtained from Similarity Ensemble ways represented as nodes and the associations as edges (Fig. 6). In this
Approach (SEA) and Swiss Target Prediction (STP) databases, as well as the way, the 143 targets were mapped to 43 canonical pathways, with 576
disease-associated targets (DAT). interactions between these components, further highlighting the diverse
mechanisms at play in propolis anti-T2DM activity.
Next, the overlapping targets were annotated using DAVID and re-
To provide more mechanistic insights, we clustered the PPI-network trieved three GO terms – CC, BP, and MF (Supplementary File 5). The top
nodes using the MCODE plugin in Cytoscape (Fig. 4); information about 10 terms for each category are presented in Fig. 7. Almost half (49.4%)
each cluster was fetched directly from the STRING database. Nodes of the targets can be found in the plasma membrane; 42.2, 38, and 29.5%
in cluster 1 were involved in lipids (phospholipids arachidonate, and were within the cytoplasm, integral component of the membrane, and

Fig. 3. Protein-protein interaction network (A.) and the top ten (core) type-2 diabetes disease targets (B.) associated with screened compounds in Nigerian propolis.
Core targets were identified based on the number of degrees, betweenness, and closeness using the degree algorithm of the CytoHubba plugin within Cytoscape
(version 3.9.1).

4
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

Fig. 4. Cluster analysis of the targets in the protein-protein interaction network. Cluster analysis was performed with the MCODE plugin within Cytoscape (version
3.9.1) with the following parameters: Degree cutoff = 2, Node Score Cutoff = 0.2, and K-Core = 2.

Fig. 5. Compound-target network (A.) and top ten compounds in Nigerian propolis (B.) against overlapping type-2 diabetes disease targets. Compounds are ranked
for bioactivity based on the number of degrees, betweenness, and closeness using the degree algorithm of the CytoHubba plugin within Cytoscape (version 3.9.1).
Green nodes = compounds, blue nodes = targets.

5
 E.I. Ugwor, A.S. James, A.I. Amuzat et al.
Table 1
Bioactive compounds in Nigerian propolis against type-2 diabetes mellitus, as identified by network pharmacology analysis.

Rank Compound Parameters∗ Rank Compound Parameters∗

Degree Betweenness Closeness Degree Betweenness Closeness

1 2-Ethyl-1,4-dimethoxybenzene 63 22,890.92 105.53 23 Gerontoxanthone H 2 647.62 69.66

2 Borneol 32 8453.39 80.53 24 Vesticarpan 2 334.41 71.17
3 3,8-dihydroxy-9-methoxy-pterocarpan 27 7772.28 79.23 25 Calycosin 2 67.46 54.95
4 Kaempferol 25 5349.30 72.02 26 6-Prenylnaringenin 2 56.79 51.70
5 Decanoic acid, methyl ester 24 6478.51 76.83 27 Octanoic acid 1 0.00 64.96
6 Resveratol 15 4971.80 70.78 28 a-Copaene 1 0.00 53.89
7 13-Tetradece-11-yn-1-ol 13 2576.47 63.52 29 a-melonol 1 0.00 53.89
8 Broussonin B 10 2469.24 62.07 30 cis-Geraniol 1 0.00 53.89
9 1,5,9,9-tetramethyl-1,4,7-Cycloundecatriene 9 2062.24 61.59 31 Vestitol 1 0.00 53.89
6

10 2-(7-Octenyl)oxirane 9 1824.00 63.34 32 Quinine 1 0.00 46.30

11 4-epi-a-Acoradiene 8 1577.66 55.85 33 Ribalinidine 1 0.00 46.30
12 Dodecanoic acid 7 1566.00 57.46 34 Linalool oxide 1 0.00 45.88
13 Pterocarpan 6 1825.46 55.89 35 Cyclosativene 1 0.00 44.30
14 Lunamarin 6 1737.42 57.18 36 Caryophyllene oxide 1 0.00 43.51
15 Catechin 5 1048.04 51.43 37 Methyl chavicol 1 0.00 42.72
16 4,4-Dipropylheptane 5 127.81 50.29 38 Medicarpin 1 0.00 39.30
17 Benzene, 1-methyl-4-(2-pentenyloxy) 4 989.00 45.42 39 Citral 1 0.00 35.88

Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

18 Naringenin 4 608.45 55.59 40 Flavan-3-ol 1 0.00 34.30
19 Pinocembrin 4 6.00 50.57 41 Humulene-1,2-epoxide 1 0.00 30.86
20 a-Calacorene 3 689.90 55.28 42 Dodecane 1 0.00 1.50
21 Epihedrine 3 252.37 54.36 43 Ethyl 2-(5-methyl-5-vinyltetrahydrofuran-2-yl) propan-2-yl carbonate 1 0.00 1.00
22 b-Methylnaphthalene 2 1188.00 54.94 44 Heptane, 2,4,6-trimethyl- 1 0.00 1.00
∗
Parameters were calculated from analysing the compound-target network with the CytoHubba in Cytoscape (version 3.9.1).
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

Fig. 6. Target-pathway network. Dark green nodes = path-

way, blue nodes = targets.

Fig. 7. Top 10 components of Gene Ontology

(GO) enrichment analyses of Nigerian propo-
lis compounds’ anti- type-2 diabetes disease
targets. (A) Cellular component; (B) Biological
process; and (C) Molecular function.

7
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

in G-protein coupled receptor (GPCR) signalling and signal transduc-

tion, while 12.7% and 12.0% were associated with inflammatory and
drug response, respectively. These biological processes correspond with
MP annotation, revealing that 75.3% of the targets’ functions involved
protein binding, a central regulatory and signalling mechanism. Other
functions include binding to DNA, zinc, and ATP, protein homodimerisa-
tion, kinase, and GPCR activities, all of which are involved in signalling
and regulation. Indeed, T2DM develops from dysfunctional insulin sig-
nalling and has been associated with impaired regulatory and signalling
mechanisms [42,43].
KEGG pathways enriched for the overlapping targets were retrieved
from the DAVID online tool; pathways not associated with T2DM (e.g.,
chemical carcinogenesis) or not having P-values less than 0.05 were ex-
cluded (Supplementary File 5). The ten most enriched pathways are pre-
sented in a bubble plot against the fold enrichment (Fig. 8). 31.9% of
Fig. 8. Bubble plot showing top 10 highly enriched pathways obtained from
the targets were involved in metabolic pathways, which is unsurprising
Kyoto encyclopaedia of Genes and Genomes (KEGG) pathway assessment. given that T2DM is a metabolic disease [43]. Differentially expressed
NLRI = neuroactive ligand-receptor interaction. genes linked to neuroactive ligand-receptor interaction (12.7% of tar-
gets) have been reported in diabetic patients [44,45]. Advanced glyca-
tion end products (AGEs) interact with their receptor (RAGE) to potenti-
ate crosstalk that amplifies T2DM-associated complications [36]. In con-
nucleoplasm, while 14.5% can be found in the mitochondrion. Regula- trast, cAMP/adenylate cyclase/calcium crosstalk has been posited as a
tory, signal transduction and cellular response processes accounted for pharmacological target to enhance 𝛽-cell survival/function and improve
the top 10 biological processes. These processes are closely related. For glucose homeostasis [46]. Ras and MAPK8 (c-Jun N-terminal kinase)
example, external/internal stimuli (e.g., insulin or low glucose levels) signalling have been implicated in the orchestration of intricate events
can act as molecular signals, which the cell amplifies or transmits, usu- that lead to chronic low-grade inflammation characteristic of T2DM and
ally involving signalling cascades of protein interactions, and leads to its microvascular complications [47]. The PI3K-Akt pathway is crucial
the regulation of several aspects of the biological system, in response to in mediating insulin’s metabolic effects. Insulin binds to its receptor,
the stimuli. which phosphorylates insulin receptor substrates, promoting the activa-
Furthermore, BP annotation indicated that 18.1 and 10.8% of the tion of PI3K and Akt downstream. Activated Akt regulates the activity
overlapping targets were involved in the positive regulation of transcrip- of numerous targets, including kinases, transcription factors and other
tion and apoptosis, while 10.8% and 12.7% were negative regulators of regulatory molecules, one of which involves the translocation of glucose
these processes, respectively. Meanwhile, 16.9 and 13.3% were involved transporter-4 (GLUT4) to the membrane [48].

Fig. 9. Dose-response curve of 2-ethyl-1,4-

dimethoxybenzene (left) and reference drug
(right) inhibitory activities against 𝛼-amylase,
𝛼-glucosidase, and HMG-CoA reductase. The
reference drug for 𝛼-amylase and 𝛼-glucosidase
assays was acarbose, whereas simvastatin was
used for the HMG-CoA reductase inhibitory as-
say. Data are presented as mean ± standard de-
viation of three independent experiments.

8
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

These network-based analyses provide insights into the targets and Supplementary File 4 CytoHubba analysis of the compound-target
mechanisms potentially involved in the anti-T2DM effects of Nigerian and protein-protein interaction networks.
propolis. Supplementary File 5 GO and KEGG pathways information ex-
tracted from the DAVID bioinformatics resources.
3.3. Inhibitory activities of 2-ethyl-1,4-dimethoxybenzene against key
T2DM-associated targets Declaration of Competing Interest

Furthermore, to validate the anti-T2DM potential of Nige- The authors declare that they have no known competing financial
rian propolis, we evaluated the inhibitory activity of 2-ethyl-1,4- interests or personal relationships that could have appeared to influence
dimethoxybenzene against 𝛼-amylase, 𝛼-glucosidase, and HMG-CoA re- the work reported in this paper.
ductase, compared it to referenced drugs (Fig. 9). HMGR and 𝛼-amylase
are among the 167 overlapping targets associated with NP anti-T2DM
CRediT authorship contribution statement
activity; acarbose and simvastatin, FDA-approved drugs for manag-
ing T2DM, are inhibitors of 𝛼-glucosidase and HMGR, respectively.
Emmanuel I. Ugwor: Conceptualization, Writing – original draft,
2-ethyl-1,4-dimethoxybenzene showed high inhibitory activity against
Project administration, Methodology, Formal analysis, Visualization.
𝛼-amylase and 𝛼-glucosidase, with IC50 values lower than acarbose.
Adewale S. James: Conceptualization, Project administration, Method-
Although the IC50 value of simvastatin (reference drug for HMGR
ology, Writing – review & editing. Adekunle I. Amuzat: Project admin-
assay) was lower, 2-ethyl-1,4-dimethoxybenzene showed appreciable
istration, Methodology, Resources. Emmanuel O. Ezenandu: Writing
HMGR inhibitory activity (IC50 = 111.2 ± 1.76 μg/ml compared to
– original draft, Methodology, Resources. Victory C. Ugbaja: Writing
60.54 ± 3.22 μg/ml for simvastatin). These results are concordant with
– original draft, Methodology, Resources. Regina N. Ugbaja: Supervi-
our network analyses and previous reports of propolis from Nigeria
sion, Writing – review & editing, Resources.
[10,14, 15] and other regions such as China [9], Taiwan [11], Morocco
[49], and Brazil [50]. Data Availability

3.4. Limitations of the study All data generated or analysed during this study are included in this
published article and its Supplementary Files.
A significant limitation of our study is that we did not evaluate the
effects of the identified bioactive compounds on cell viability nor their Acknowledgments
effect at the gene/protein level. However, our study provides a pipeline
of propolis compounds that be appraised for their anti-diabetic poten- None.
tials via in vivo and clinical studies.

Supplementary materials
4. Conclusion
Supplementary material associated with this article can be found, in
The network pharmacology approach, which holistically consid-
the online version, at doi:10.1016/j.prmcm.2022.100183.
ers the complex interactions in biological systems, was employed to
elucidate the bioactive constituents of Nigerian propolis and provide
References
insights into the mechanism of their anti-T2DM activity. We iden-
tified 44 compounds (out of 202) that interacted with anti-T2DM [1] A.A. Gunaid, A.M. Assabri, Prevalence of type 2 diabetes and other cardiovascular
targets, with 2-Ethyl-1,4-dimethoxybenzene, borneol, 3,8-dihydroxy- risk factors in a semirural area in Yemen, East. Mediterr. Health J. 14 (1) (2008)
9‑methoxy-pterocarpan, kaempferol, decanoic acid, methyl ester, 42–56 https://apps.who.int/iris/handle/10665/117407.
[2] WHONoncommunicable Diseases, World Health Organization, 2020
resveratrol, 13-tetradece-11-yn-1-ol, broussonin B, 1,5,9,9-tetramethyl-
https://www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases
1,4,7-cycloundecatriene, and 2-(7-Octenyl)oxirane as the top ten com- Accessed June 27 2022.
pounds. On the other hand, HSP90AA1, JUN, ESR1, STAT3, CYP3A4, [3] S. Safiri, N. Karamzad, J.S. Kaufman, A.W. Bell, S.A. Nejadghaderi, M.J. Sullman,
et al., Prevalence, deaths and disability-adjusted-life-years (DALYs) due to type 2
PTGS2, RELA, VEGFA, CYP2C9, and PPARG were identified as the core
diabetes and its attributable risk factors in 204 countries and territories, 1990–2019:
anti-T2DM targets (n = 167) associated with Nigerian propolis con- results from the global burden of disease study 2019, Front. Endocrinol. 3 (2021)
stituents. The targets were primarily involved in pathways pertinent to 838027, doi:10.3389/fendo.2022.838027.
lipid metabolism, regulation, and signal transduction, while 49.3% of [4] U. Galicia-Garcia, A. Benito-Vicente, S. Jebari, A. Larrea-Sebal, H. Siddiqi,
K.B. Uribe, et al., Pathophysiology of type 2 diabetes mellitus, Int. J. Mol. Sci. 21
the targets were localised in the plasma membrane. Thus, we provide (17) (2020) 6275, doi:10.3390/ijms21176275.
thorough mechanistic insights into the potential anti- T2DM activities of [5] D. Himanshu, W. Ali, M. Wamique, Type 2 diabetes mellitus: pathogenesis
the Nigerian propolis’ constituents while revealing a pipeline of bioac- and genetic diagnosis, J. Diabetes Metab. Disord. 19 (2) (2020) 1959–1966,
doi:10.1007/s40200-020-00641-x.
tive compounds that can be developed for the therapeusis of type 2 di- [6] F. Ventura, C. Borghi, Diabetes in the third millennium: prognosis improves, but
abetes mellitus. juvenile forms are more frequent, Eur. Heart J. Suppl. 21 (Suppl B) (2019) B17 J
Eur Soc Cardiol, doi:10.1093/eurheartj/suz006.
[7] A.S. Wickramasinghe, P. Kalansuriya, A.P. Attanayake, Nanoformulation of plant-
Funding based natural products for type 2 diabetes mellitus: from formulation design
to therapeutic applications, Curr. Ther. Res. Clin. Exp. 96 (2022) 100672,
doi:10.1016/j.curtheres.2022.100672.
This research did not receive any specific grant from the public, com-
[8] S. Kumar, A. Mittal, D. Babu, A. Mittal, Herbal medicines for diabetes manage-
mercial, or not-for-profit funding agencies. ment and its secondary complications, Curr. Diabetes Rev. 17 (4) (2021) 437–456,
Supplementary File 1 Nigerian propolis compound database com- doi:10.2174/1573399816666201103143225.
[9] Y.Z. Shi, Y.C. Liu, Y.F. Zheng, Y.F. Chen, J.J. Si, M.L. Chen, et al., Ethanol extract
piled from previously reported works and FDA-approved diabetic drugs
of Chinese propolis attenuates early diabetic retinopathy by protecting the blood–
used in the present study. retinal barrier in streptozotocin-induced diabetic rats, J. Food Sci. 84 (2) (2019)
Supplementary File 2 Pharmacologically relevant characters of the 358–369, doi:10.1111/1750-3841.14435.
96 screened compounds in Nigerian propolis. [10] M.I. Oladayo, Nigerian propolis improves blood glucose, glycated hemoglobin
A1c, very low-density lipoprotein, and high-density lipoprotein levels in rat
Supplementary File 3 Comprehensive lists of targets retrieved from models of diabetes, J. Intercult. Ethnopharmacol. 5 (3) (2016) 233–238,
the databases. doi:10.5455/jice.20160502065029.

9
E.I. Ugwor, A.S. James, A.I. Amuzat et al. Pharmacological Research - Modern Chinese Medicine 5 (2022) 100183

[11] L.H. Chen, Y.W. Chien, M.L. Chang, C.C. Hou, C.H. Chan, H.W. Tang, et al., Tai- [31] I.A. Duru, Comparative phytochemical analysis of brown, green and red propo-
wanese green propolis ethanol extract delays the progression of type 2 diabetes mel- lis from Umudike, Abia State Nigeria, Adv. J. Chem. Sect. B 3 (1) (2020) 86–97,
litus in rats treated with streptozotocin/high-fat diet, Nutrients 10 (4) (2018) 503, doi:10.22034/ajcb.2021.121910.
doi:10.3390/nu10040503. [32] O.S. Balogun, A.O. Oluduro, Z. Liu, Chemical composition and in vitro biologi-
[12] C.S. Alaribe, T. Esposito, F. Sansone, A. Sunday, I. Pagano, A.L. Piccinelli, cal studies of volatile oils from Nigerian bee propolis, J. Apic. Res. (2020) 1–9,
et al., Nigerian propolis: chemical composition, antioxidant activity and 𝛼- doi:10.1080/00218839.2020.1862392.
amylase and 𝛼-glucosidase inhibition, Nat. Prod. Res. 35 (18) (2021) 3095–3099, [33] A. Daina, O. Michielin, V. Zoete, SwissADME: a free web tool to evaluate pharma-
doi:10.1080/14786419.2019.1682576. cokinetics, drug-likeness and medicinal chemistry friendliness of small molecules,
[13] S. Huang, C.P. Zhang, K. Wang, G.Q. Li, F.L Hu, Recent advances in the Sci. Rep. 7 (1) (2017) 1–3, doi:10.1038/srep42717.
chemical composition of propolis, Molecules 19 (12) (2014) 19610–19632, [34] I. Lazaro, A. Oguiza, C. Recio, B. Mallavia, J. Madrigal-Matute, J. Blanco, et al.,
doi:10.3390/molecules191219610. Targeting HSP90 ameliorates nephropathy and atherosclerosis through suppression
[14] R.N. Ugbaja, T.P. Fatokun, D.I. Akinloye, A.S. James, O.O. Onabanjo, O.A. Akinloye, of NF-𝜅B and STAT signaling pathways in diabetic mice, Diabetes 64 (10) (2015)
Propolis ethanol extract abrogates hyperglycemia, lipotoxicity, and lowered hepatic 3600–3613, doi:10.2337/db14-1926.
poly (ADP-ribose) polymerase protein level in male albino rats, J. Diabetes Metab. [35] X. Yang, Y. Zhang, W. Xu, R. Deng, Y. Liu, F. Li, et al., Potential role of Hsp90 in
Disord. 20 (1) (2021) 683–696, doi:10.1007/s40200-021-00800-8. rat islet function under the condition of high glucose, Acta Diabetol. 53 (4) (2016)
[15] I.R. Babatunde, A. Abdulbasit, M.I. Oladayo, O.I. Olasile, F.R. Olamide, B.W. Gbo- 621–628, doi:10.1007/s00592-016-0852-2.
lahan, Hepatoprotective and pancreatoprotective properties of the ethanolic extract [36] N.M. El-Nasr, D.O. Saleh, S.S. Mahmoud, S.M. Nofal, R.M. Abdelsalam, M.M. Safar,
of Nigerian propolis, J. Intercult. Ethnopharmacol. 4 (2) (2015) 102. et al., Olmesartan attenuates type 2 diabetes-associated liver injury: cross-talk of
[16] R.S. Ibrahim, A.A. El-Banna, Network pharmacology-based analysis for unraveling AGE/RAGE/JNK, STAT3/SCOS3 and RAS signaling pathways, Eur. J. Pharmacol.
potential cancer-related molecular targets of Egyptian propolis phytoconstituents 874 (2020) 173010, doi:10.1016/j.ejphar.2020.173010.
accompanied with molecular docking and in vitro studies, RSC Adv. 11 (19) (2021) [37] A. Kowalczuk, N. Bourebaba, K. Kornicka-Garbowska, E. Turlej, K. Marycz, L. Boure-
11610–11626, doi:10.1039/D1RA01390D. baba, Hyoscyamus albus nortropane alkaloids reduce hyperglycemia and hyperinsu-
[17] B. Gogoi, D. Gogoi, Y. Silla, B.B. Kakoti, B.S. Bhau, Network pharmacology- linemia induced in HepG2 cells through the regulation of SIRT1/NF-kB/JNK path-
based virtual screening of natural products from Clerodendrum species for iden- way, Cell Commun. Signal. 19 (1) (2021) 1–26, doi:10.1186/s12964-021-00735-w.
tification of novel anti-cancer therapeutics, Mol. Biosyst. 13 (2) (2017) 406–416, [38] S. Gravel, J.L. Chiasson, J. Turgeon, A. Grangeon, V. Michaud, Modulation of CYP
doi:10.1039/C6MB00807K. 450 activities in patients with type 2 diabetes, Clin. Pharmacol. Ther. 106 (6) (2019)
[18] V.M. Kutluay, N.Y. Diker, Constitution of a comprehensive phytochemical profile 1280–1289, doi:10.1002/cpt.1496.
and network pharmacology based investigation to decipher molecular mechanisms [39] N. Sellami, L.B. Lamine, A. Turki, S. Sarray, M. Jailani, A.K. Al-Ansari, M. Ghor-
of Teucrium polium L. in the treatment of type 2 diabetes mellitus, PeerJ 8 (2020) bel, T. Mahjoub, W.Y. Almawi, Association of VEGFA variants with altered VEGF
e10111, doi:10.7717/peerj.10111. secretion and type 2 diabetes: a case-control study, Cytokine 106 (2018) 29–34,
[19] M. Liao, H. Shang, Y. Li, T. Li, M. Wang, Y. Zheng, et al., An integrated ap- doi:10.1016/j.cyto.2018.03.003.
proach to uncover quality marker underlying the effects of Alisma orientale on lipid [40] E.M. Vázquez, N. Cobo-Vuilleumier, R.A. Legido, S. Marín-Cañas, E. Nola, A. Dor-
metabolism, using chemical analysis and network pharmacology, Phytomedicine 45 ronsoro, et al., NR5A2/LRH-1 regulates the PTGS2-PGE2-PTGER1 pathway con-
(2018) 93–104, doi:10.1016/j.phymed.2018.04.006. tributing to pancreatic islet survival and function, iScience 25 (5) (2022) 104345,
[20] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Experimental and com- doi:10.1016/j.isci.2022.104345.
putational approaches to estimate solubility and permeability in drug discov- [41] K.C. Gregorio, C.P. Laurindo, U.F. Machado, Estrogen and glycemic homeostasis: the
ery and development settings, Adv. Drug. Deliv. Rev. 23 (1–3) (1997) 3–25, fundamental role of nuclear estrogen receptors ESR1/ESR2 in glucose transporter
doi:10.1016/S0169-409X(96)00423-1. GLUT4 regulation, Cells 10 (1) (2021) 99, doi:10.3390/cells10010099.
[21] D.F. Veber, S.R. Johnson, H.Y. Cheng, B.R. Smith, K.W. Ward, K.D. Kopple, Molec- [42] S. Balakrishnan, S. Dhavamani, Prahalathan C. 𝛽-Cell specific transcription factors
ular properties that influence the oral bioavailability of drug candidates, J. Med. in the context of diabetes mellitus and 𝛽-cell regeneration, Mech. Dev. 163 (2020)
Chem. 45 (12) (2002) 2615–2623, doi:10.1021/jm020017n. 103634, doi:10.1016/j.mod.2020.103634.
[22] A.K. Ghose, T. Herbertz, J.M. Salvino, J.P. Mallamo, Knowledge-based chemoinfor- [43] T.M. Batista, N. Haider, C.R. Kahn, Defining the underlying defect in in-
matic approaches to drug discovery, Drug Discov. Today 11 (23–24) (2006) 1107– sulin action in type 2 diabetes, Diabetologia 4 (5) (2021) 994–1006,
1114, doi:10.1016/j.drudis.2006.10.012. doi:10.1007/s00125-021-05415-5.
[23] J. Wang, X. Chen, W. Bai, Z. Wang, W. Xiao, J Zhu, Study on mechanism [44] S. Kaviarasan, S. Muniandy, R. Qvist, I.S. Ismail, Gene expression profile in leuco-
of Ginkgo biloba L. leaves for the treatment of neurodegenerative diseases cytes of type 2 diabetic subjects, Biomed. Res. 20 (3) (2009) 158–165 (Aligarh),
based on network pharmacology, Neurochem. Res. 46 (7) (2021) 1881–1894, doi:10.4103/0970-938X.54834.
doi:10.1007/s11064-021-03315-z. [45] W. Qu, C. Han, M. Li, J. Zhang, L. Li, Revealing the underlying mechanism of diabetic
[24] P. Shannon, A. Markiel, O. Ozier, N.S. Baliga, J.T. Wang, D. Ramage, N. Amin, nephropathy viewed by microarray analysis, Exp. Clin. Endocrinol. Diabetes 123 (6)
B. Schwikowski, T. Ideker, Cytoscape: a software environment for integrated mod- (2015) 353–359, doi:10.1055/s-0035-1548849.
els of biomolecular interaction networks, Genome Res. 13 (11) (2003) 2498–2504, [46] D.S. Carvalho, A.A. de Almeida, A.F. Borges, D.V. Campos, Treatments for diabetes
doi:10.1101/gr.1239303. mellitus type II: new perspectives regarding the possible role of calcium and cAMP
[25] D. Szklarczyk, A.L. Gable, K.C. Nastou, D. Lyon, R. Kirsch, S. Pyysalo, et al., The interaction, Eur. J. Pharmacol. 830 (2018) 9–16, doi:10.1016/j.ejphar.2018.04.002.
STRING database in 2021: customizable protein-protein networks, and functional [47] S. Huang, J. Wang, L. Zhang, S. Tian, Y. Wang, X. Shao, et al., Ras guanine
characterization of user-uploaded gene/measurement sets, Nucleic Acids Res. 49 nucleotide-releasing protein-4 promotes renal inflammatory injury in type 2 diabetes
(D1) (2021) D605–D612, doi:10.1093/nar/gkaa1074. mellitus, Metabolism 131 (2022) 155177, doi:10.1016/j.metabol.2022.155177.
[26] P. Bernfeld, Enzymes of starch degradation and synthesis, Adv. Enzymol. Relat. Subj. [48] S.A. Khorami, A.R. Movahedi, K. Huzwah, A.M. Sokhini, PI3K/AKT pathway in mod-
Biochem. 12 (1951) 379–428, doi:10.1002/9780470122570.ch7. ulating glucose homeostasis and its alteration in diabetes, Biomed. Sci. Instrum. 1
[27] E. Apostolidis, Y.I. Kwon, K. Shetty, Inhibitory potential of herb, fruit, and fungal- (2) (2015) 46–55.
enriched cheese against key enzymes linked to type 2 diabetes and hypertension, In- [49] N. El Menyiy, N. Al-Wali, A. El Ghouizi, S. El-Guendouz, K. Salom, B Lyoussi, Po-
nov. Food Sci. Emerg. Technol. 8 (1) (2007) 46–54, doi:10.1016/j.ifset.2006.06.001. tential therapeutic effect of Moroccan propolis in hyperglycemia, dyslipidemia, and
[28] D. Iqbal, M.S. Khan, A. Khan, M. Khan, S. Ahmad, A.K. Srivastava, et al., In vitro hepatorenal dysfunction in diabetic rats, Iran. J. Basic Med. Sci. 22 (11) (2019) 1331
screening for 𝛽-hydroxy-𝛽-methylglutaryl-coa reductase inhibitory and antioxidant 10.22038%2Fijbms.2019.33549.8004.
activity of sequentially extracted fractions of Ficus palmata forsk, Biomed. Res. Int. [50] T. Fukuda, M. Fukui, M. Tanaka, T. Senmaru, H. Iwase, M. Yamazaki, et al.,
(2014) 762620, doi:10.1155/2014/762620. Effect of Brazilian green propolis in patients with type 2 diabetes: a double-
[29] B. Lawal, O.K. Shittu, A.N. Abubakar, I.A. Olalekan, A.M. Jimoh, A.K. Abdulazeez, blind randomized placebo-controlled study, Biomed. Rep. 3 (3) (2015) 355–360,
Drug leads agents from methanol extract of Nigerian bee (Apis mellifera) propolis, J. doi:10.3892/br.2015.436.
Intercult. Ethnopharmacol. 5 (1) (2016) 43–48, doi:10.5455/jice.20151208122127.
[30] R.M. Omar, J. Igoli, A.I. Gray, G.U. Ebiloma, C. Clements, J. Fearnley, et al., Chem-
ical characterisation of Nigerian red propolis and its biological activity against Try-
panosoma brucei, Phytochem. Anal. 27 (2) (2016) 107–115, doi:10.1002/pca.2605.

10