Abstract
Mulberry (Morus alba L.) leaves have long been considered beneficial in
traditional Chinese medicine to treat infectious and internal diseases.
Recently studies have discovered that the mulberry leaf’s total
flavonoids (MLF) display excellent hypoglycemia properties. However,
the active ingredients and their molecular mechanisms are still
uncharacterized. In this study, we explored the hypoglycemic effects of
MLF and mulberry leaf polysaccharides (MLP) on ob/ob mice, an animal
model of type 2 diabetes mellitus (T2DM), compared with Ramulus Mori
(Sangzhi) alkaloid (RMA). Network pharmacology was employed to identify
the potential available targets and active compounds of MLF and RMA
against hyperglycemia. Molecular docking, an insulin-resistant cell
model and qPCR were employed to verify the antidiabetic activity of the
critical compounds and the gene expression profiles of the top
molecular targets. Here, the results showed that MLF and MLP improved
glucose uptake in insulin-resistant hepatocytes. MLF, MLP and RMA
alleviated insulin resistance and glucose intolerance in ob/ob mice.
Unlike MLF and MLP, RMA administration did not influence the
accumulation of intrahepatic lipids. Network pharmacology analysis
revealed that morusin, kuwanon C and morusyunnansin L are the main
active compounds of MLF and that they amend insulin resistance and
glycemia via the PI3K- Akt signaling pathway, lipid and atherosclerosis
pathways, and the AGE-RAGE signaling pathway. Moreover,
1-deoxynojirimycin (DNJ), fagomine (FA), and
N-methyl-1-deoxynojirimycin are the primary active ingredients of RMA
and target carbohydrate metabolism and regulate alpha-glucosidase
activity to produce a potent anti-diabetic effect. The molecular
docking results indicated that morusin, kuwanon C and morusyunnansin L
are the critical bioactive compounds of MLF. They had high affinities
with the key targets adenosine A1 receptor (ADORA1), AKT
serine/threonine kinase 1 (AKT1), peroxisome proliferator-activated
receptor gamma (PPARγ), and glycogen synthase kinase 3 beta (GSK3β),
which play crucial roles in the MLF-mediated glucose-lowering effect.
Additionally, morusin plays a role in amending insulin resistance of
hepatocytes by repressing the expression of the ADORA1 and PPARG genes.
Our results shed light on the mechanism behind the glucose-lowering
effects of MLF, suggesting that morusin, kuwanon C, and morusyunnansin
L might be promising drug leads for the management of T2DM.
Keywords: mulberry leaf, flavonoids, type 2 diabetes, network
pharmacology, molecular docking
Introduction
Type 2 diabetes mellitus (T2DM) is a chronic metabolic disease that may
lead to multiple complications, such as cardiovascular, renal and
ophthalmic complications ([40]Ali et al., 2022). T2DM is relatively
heterogeneous and very complex, involving multiple pathophysiological
mechanisms that affect the pancreas and metabolic organs, making
effective treatment very challenging ([41]Demir et al., 2021).
Adenosine A1 receptor (ADORA1) is known to inhibit adenylate cyclase
and play a role in regulating cell metabolism and gene transcription.
Previous studies have shown that ADORA1 plays a vital role in
carcinogenesis and is an important drug target in tumors ([42]Liu et
al., 2020a; [43]Pan et al., 2021). Moreover, a recent study showed that
ADORA1 involved in maintaining glucose homeostasis and regulating
glucagon secretion as a G-protein-coupled receptor ([44]Cheng et al.,
2000). Meanwhile, activation of ADORA1 signaling in peripheral tissues
facilitates high-fat diet-induced obesity. Specific inhibition of
ADORA1 in the liver helps prevent body weight gain and alleviate
hepatic steatosis, suggesting that ADORA1 might be a promising drug
target for treating diabetes and obesity ([45]Hong et al., 2019).
Peroxisome proliferator-activated receptor gamma (PPARγ), a known
target for thiazolidinediones, belongs to the nuclear receptor family.
Activation of PPARγ results in increased insulin sensitivity in
skeletal muscle and liver and improves the secretory profile of adipose
tissue, favoring the release of insulin-sensitizing adipokines, such as
adiponectin, and reducing inflammatory cytokines ([46]Skat-Rordam et
al., 2019; [47]Wang et al., 2020a). However, thiazolidinediones cause
adverse effects such as weight gain, fluid retention, bone fractures,
and congestive heart failure, which impose a huge health burden
([48]Kahn and McGraw, 2010). Interestingly, full and partial activation
and antagonism of PPARγ can all improve insulin sensitivity
([49]Ahmadian et al., 2013). Therefore, discovering novel selective
modulators of PPARγ that evoke fewer side effects while possessing
insulin-sensitizing potential is a vital goal.
Natural products derived from medicinal plants provide multiple health
benefits ([50]Al-Ishaq et al., 2019). Over the past 20 years,
scientific attention has been given to natural compounds, that play
pivotal roles in drug or lead discovery, especially for infectious
diseases, diabetes, and cardiovascular disease ([51]Ong and Khoo, 2000;
[52]Atanasov et al., 2021). Flavonoids are a group of polyphenolic
compounds that are widely distributed in plants ([53]Cao et al., 2019)
and display various positive health effects on metabolic disorders.
Studies have shown that flavonoid intake may decrease the risk of
developing T2DM ([54]Liu et al., 2014; investigators, 2015) by
regulating targeted cellular signaling networks related to insulin
secretion, glucose metabolism, and glucose transport in pancreatic
β-cells, hepatocytes, skeletal myofibers, and adipocytes ([55]Hussain
et al., 2020). Therefore, developing and utilizing flavonoids are
essential for the therapy and prevention of metabolic disorders.
Mulberry (Morus alba L.) is a plant belonging to the family Moraceae
and the genus Morus ([56]Lee et al., 2020). Ramulus Mori (Sangzhi)
alkaloid (RMA), a group of effective polyhydroxy alkaloids derived from
Ramulus Mori ([57]Liu S. et al., 2019), is a novel inhibitor of
α-glucosidase that the China National Medical Products Administration
has approved for the treatment of T2DM ([58]Liu et al., 2021).
Therefore, mulberry leaves have been evaluated and have been found to
exhibit excellent hypoglycemic activity and reduce inflammation and
insulin resistance in T2DM ([59]Tian et al., 2019; [60]Li et al., 2020;
[61]Meng et al., 2020). Mulberry leaf flavonoids (MLF), polysaccharides
(MLP) and alkaloids are the main functional components of mulberry
leaves with various biological activities, such as antioxidation,
hypolipidemia and hypoglycemia ([62]Meng et al., 2020; [63]Zhong et
al., 2020). MLF ameliorates skeletal muscle insulin resistance
([64]Meng et al., 2020), reduces the accumulation of lipids and hepatic
steatosis, and whitens brown fat in diet- or gene deficiency-induced
obese mice ([65]Zhong et al., 2020). MLP effectively normalizes hepatic
glucose metabolism and insulin signaling and mitigates oxidative stress
in the livers of rats with T2DM induced by high fat diet and
streptozotocin ([66]Ren et al., 2015). RMA, as an inhibitor of
α-glucosidase, mainly acts on the gut and delays the intestinal
digestion of carbohydrates ([67]Li et al., 2016). Hence, we employed
ob/ob mice to compare the glucose -lowering effects of MLF, MLP and RMA
and performed network pharmacology analysis to discover their potential
active compounds and mechanisms, verifying the findings in human
hepatocytes. A total of 29 flavonoids of mulberry leaf and 4 alkaloids
of Ramulus Mori were collected from the TCMSP database and published
literature for network analysis. Our results indicated that
1-deoxynojirimycin (DNJ), fagomine (FA), and
N-methyl-1-deoxynojirimycin are the primary active compounds of RMA and
target maltase-glucoamylase (MGAM) and sucrase-isomaltase (SI) proteins
to lower glucose. Meanwhile, morusin, kuwanon C and morusyunnansin L
are probably the important ingredients of MLF in hypoglycemia, which
may function by regulating key targets, including ADORA1, AKT
serine/threonine kinase 1 (AKT1), PPARγ and glycogen synthase kinase-3
beta (GSK3β).
Methods
Preparation of the crude extract of mulberry leaf
Mulberry (Morus alba L.) leaves were purchased from Beijing Tong Ren
Tang Co., Ltd. (Beijing, China). The crude extract of mulberry leaf
(MLE) was prepared with the following procedures. Approximately 100 g
of mulberry leaves was refluxed with 1,400 ml water for 1 h. The
filtrate was collected by filtration using a Buchner funnel and
evaporated to obtain crude extracts (17.92 g).
Preparation and quality control of mulberry leaf extract flavonoids
The MLF was prepared with the following procedures. One kilogram of
mulberry leaves was refluxed with 60% ethanol (1:10, w/v) for 1 h, and
then the filtrate was collected. A further 10,000 ml of 60% ethanol was
added to the drug residue and refluxed for another 1 h. All the
filtrates were collected and decompressed to concentrate. Then, the
alcohol extract was purified on an AB-8 macroporous adsorption resin
column (Shanghai Macklin Biochemical Co., Ltd., Beijing, China), and
the elution solvent was a water-ethanol system (0%, 20%, and 70%).
Finally, the eluent of 70% ethanol was collected and then concentrated
to dryness.
The main ingredients rutin, isoquercitrin, and astragalin in MLF were
confirmed using an Agilent 1,260 liquid chromatography system ([68]Meng
et al., 2020). In brief, 15 μl MLF was injected into the apparatus with
an autosampler. Chromatographic separation was performed using an
Agilent C18 column (4.6 × 250 mm, 5 μm) with a flow rate of 1.0 ml/min.
The mobile phases were A-0.1% (v/v) formic acid in water and
B-acetonitrile. The gradient elution conditions are shown in
[69]Supplementary Table S1. The column temperature was 30°C, with a
detection wavelength of 365 nm. The rutin, isoquercitrin and astragalin
contents in MLF were 0.4954%, 0.8826%, and 0.3638%, respectively
([70]Figure 1A; [71]Table 1).
FIGURE 1.
[72]FIGURE 1
[73]Open in a new tab
HPLC chromatogram for determining the content of rutin, isoquercitrin
and astragalin in MLF and the content of DNJ in RMA. (A) Determination
of rutin, isoquercitrin and astragalin contents in MLFs by HPLC
analysis. The black, blue and purple lines represent rutin,
isoquercetin, and astragalin, respectively. The brown line represents
MLF. (B) Determination of DNJ content in RMA by HPLC analysis. The
upper half represents RMA, whereas the lower half represents DNJ.
TABLE 1.
Determination of rutin, isoquercitrin and astragalin contents in MLFs
by HPLC analysis.
Serial number Compound Retentiontime (min) Relative peak area
Concentration (mg/ml) Content (%)
1 Rutin 29.920 3.4953 0.01392 0.4954
2 Isoquercitrin 32.860 9.0638 0.02480 0.8826
3 Astragalin 45.043 3.2860 0.02045 0.3638
[74]Open in a new tab
Estimation of the polysaccharide content in MLP
MLP was purchased from Shanghai Yuanye Bio-Technology Co., Ltd.
(Shanghai, China), and the polysaccharide content in MLP was measured
using the phenol‒sulfuric acid method ([75]Dubios et al., 1956).
According to the glucose standard curve ([76]Supplementary Figure S2),
the polysaccharide content of MLP was calculated to be 78.28%.
Determination of the DNJ content in RMA
RMA was purchased from Beijing Wehandbio Co., Ltd. (Beijing, China),
and the DNJ content was determined using high-performance liquid
chromatography (HPLC) ([77]Piao et al., 2018; [78]Ma et al., 2019). The
following analysis conditions were used: column, Agilent C18 column
(4.6 × 250 mm, 5 μm); mobile phase, A-acetonitrile, B-75 mmol/L sodium
citrate (pH = 4.21); flow rate, 1.5 ml/min; column temperature, 32°C;
UV detector wavelength, 264 nm; and the injection volume, 10 µl. The
gradient elution conditions are shown in [79]Supplementary Table S2.
The DNJ content in RMA was 9.58% ([80]Figure 1B).
Cell culture and glucose uptake experiment
The human normal liver L02 cell line was kindly provided by Dr. Jiyan
Zhang (Academy of Military Sciences, Beijing, China). The insulin
resistance cell model was established according to our previous method
([81]Lv et al., 2019). Briefly, the drugs and extracts were first
dissolved in DMSO and diluted with 1640 RPMI medium to appropriate
experimental concentrations for cell exposure experiments. The final
concentration of DMSO was less than 0.1%. Cells were seeded in a
96-well plate containing 1,640 medium supplemented with 10% fetal
bovine serum (FBS, Gibco), 100 U/ml penicillin and 1% streptomycin
(Gibco). Cells were cultured at 37 °C in a humidified atmosphere
containing 5% CO[2]. Twenty-four hours after seeding, the cell medium
was changed to 1,640 containing 2% FBS and 25 μmol/L lithocholic acid
(LCA) for 24 h to induce insulin resistance. Next, different
concentrations of drugs (0.5, 1, 2 mg/ml MLE; 25, 50, 100 mg/L MLF; 25,
50, 100 μmol/L DNJ; 62.5, 125, 250 mg/L MLP) or 25 μmol/L metformin
(Met) were added and incubated for 24 h. After this incubation,
cellular glucose uptake was examined using fluorescent
2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]-D-glucose (2-NBDG,
Invitrogen). Cells were washed with PBS and incubated with 100 μmol/L
2-NBDG supplemented with 1 × 10^-7 mol/L insulin at 37 °C with 5% CO[2]
for 30 min. After the treatment, the cells were washed with PBS, and
fresh PBS was added to each well (100 μL per well). The fluorescence
was detected using a fluorescence microplate reader (excitation
wavelength 488 nm, emission wavelength 520 nm).
Cell viability assay
Cell viability was detected by MTT assay ([82]Zhang et al., 2017). The
protocol of drug administration was the same as that in the experiments
described above. Briefly, 24 h after drug treatment, the culture medium
was removed, and 100 μl of 1,640 medium containing 0.5 mg/ml MTT was
added, followed by incubation at 37°C in a humidified atmosphere of 5%
CO[2]. Four hours later, the culture medium was completely removed from
each well, and 150 μl DMSO was used to dissolve the insoluble formazan
crystals. The absorbance of the solvate of each well was detected by a
microplate reader at 570 nm.
The cell glucose uptake rate (%) was calculated according to the
following formula:
[MATH:
Cellula
r gluc
ose upt
ake ra
te=[OD (Dru
g grou
p cellu
lar gl
ucose u
ptake)÷OD(D<
mi>rug gro<
/mi>up cel<
mi>l viabi<
/mi>lity)][OD(Control group cellular glucose uptake)<
/mrow>÷OD (Control group cell viability)]×100% :MATH]
Cellular glucose consumption assessment
The cell culture conditions and reagents were the same as described
above. L02 cells were cultured in 96-well plates and divided into six
groups: control (Con), model (Mod), 1.25, 2.5, 5, and 10 μmol/L
morusin. Cells in the Mod and drug groups were treated with 25 μmol/L
LCA for 24 h. Then, cells in the drug groups were treated with morusin
at different concentrations (1.25, 2.5, 5, and 10 μmol/L) for 24 h.
Twenty-4 hours later, the culture medium was removed, and 100 μl of
1,640 medium was added. Cells were then incubated at 37°C for 6 h in a
CO[2] incubator. Subsequently, 5 μl of the supernatant from each well
or different concentrations of the standard glucose solution were added
to 100 μl of working solution. The absorbance was determined at 550 nm
after incubation for 20 min at 37°C. The glucose consumption was
calculated by subtracting the glucose concentration of blank wells from
that of cell-plated wells. After aspirating the supernatant to detect
glucose consumption, 20 μl MTT was added to each well, and the cells
were then incubated at 37°C for 4 h. Cell survival was then detected
through an MTT assay.
Animals and treatments
Five-week-old (B6/JGpt-LepemICd25/Gpt, Leptinmut/mut) ob/ob mice and
their leptinwt/wt (WT) littermates (GemPhamatech Co.,Ltd. Jiangsu,
China) were housed under a 12 h light/dark cycle with free access to
food and water. The animal experimental project was reviewed and
approved by the Guidelines and Policies for Animal Surgery under the
approval of the Chinese Academy of Medical Sciences and Peking Union
Medical College, Beijing, China (approval No: SLXD-20200827001), and
was approved by the Institutional Animal Use and Care Committee. After
7 days of acclimatization, the ob/ob mice were randomly divided into 4
groups containing 7 mice: the ob/ob, RMA, MLF, and MLP groups. Mice in
the WT and ob/ob groups were fed a standard chow diet, while MLF and
MLP mice received standard chow containing 1% (w/w) MLF or MLP,
respectively. Mice in the RMA group were administered RMA by gavage
(50 mg/kg). At 20 weeks of age, all mice were fasted for 12 h and
terminally anesthetized with 200 mg/kg tribromoethanol. The blood was
collected and centrifuged at 3,000 rpm for 10 min, and the serum was
used to determine serum biomarkers. The livers were fixed in 4%
paraformaldehyde or quickly frozen in liquid nitrogen and stored at
−80°C for subsequent analysis.
Oral glucose tolerance test and insulin tolerance test
At the 13th week, oral glucose tolerance tests (OGTTs) were performed
on 12 h fasted mice administered a glucose solution (2 g/kg). In the
14th week, all mice were fasted for 4 h before the insulin tolerance
tests (ITTs) were carried out, in which the mice were intraperitoneally
injected with recombinant human insulin (0.75 U/kg, provided by Novo
Nordisk). In the OGTTs and ITTs, the blood was taken from the tail
vein, and blood glucose levels were detected at 0, 30, 60, 90 and
120 min using a glucometer and test strips after the glucose and
insulin were given to the mice.
Blood biochemical analysis
Serum glucose levels were measured by Beckman Coulter AU480 Automatic
Biochemistry device using a glucose kit (ZHONGSHENG BEIKONG
BIO-TECHNOLOGY AND SCIENCE, INC.). Insulin contents in the serum were
detected using a mouse insulin ELISA kit (Beijing Sino-UK Institute of
Biological Technology) according to the manufacturer’s directions.
Homeostasis model assessment-insulin resistance (HOMA-IR) and
homeostasis model assessment-insulin sensitive index (HOMA-ISI) were
used to evaluate the insulin resistance from basal glucose and insulin.
The HOMA-IR and HOMA-ISI indices were calculated using the following
formulas:
[MATH:
HOMA−IR
=[fas
ting glucose (mmol/L)×fas
ting in
sulin
(μU/mL)]/22.5 :MATH]
[MATH:
HOMA−IS
I=1/[
fasting glucos
e(mmol/L)×fa<
/mi>sting <
mi>insulin(μU/mL<
/mrow>)] :MATH]
Hematoxylin-eosin staining
The liver samples were fixed in 4% formaldehyde, dehydrated, embedded
in paraffin, and sectioned (5 μm). For histopathological evaluation,
the paraffin-embedded liver sections were stained with hematoxylin and
eosin (H&E).
Compound target prediction and screening of disease targets
The flavonoids and alkaloids of mulberry leaves and Ramulus Mori were
obtained from the Traditional Chinese Medicine Systems Pharmacology
Database (TCMSP, [83]http://tcmspw.com/tcmsp.php) ([84]Ru et al., 2014)
and published literature ([85]Chen et al., 2000; [86]Yang, 2010;
[87]Chen, 2014; [88]Yang et al., 2015; [89]Li, 2017). The
SwissTargetPrediction database
([90]http://www.swisstargetprediction.ch) ([91]Daina et al., 2019) was
utilized to predict the potential targets of the active molecules in
MLF and RMA. Homo sapiens was selected as the target organism. The
keyword “Type 2 diabetes” was used to collect potential genes. The
T2DM-associated targets were acquired from the DrugBank database
([92]https://go.drugbank.com/drugs) ([93]Law et al., 2014), the
Therapeutic Target Database (TTD,
[94]http://bid.nus.edu.sg/group/cjttd/) ([95]Wang et al., 2020b), the
Online Mendelian Inheritance in Man database (OMIM,
[96]https://omim.org/) ([97]Amberger et al., 2015) and the human gene
database (GeneCards, [98]https://www.genecards.org/) ([99]Stelzer et
al., 2016). The targets of T2DM and predicted compound targets were
verified using the UniProt database ([100]https://www.uniprot.org/)
([101]The UniProt, 2017), which was also employed to obtain the protein
and gene names.
Network construction
The overlapping genes between compounds and T2DM target genes were
identified and visualized using Venn diagrams. The overlapping genes
and their corresponding active compounds were imported into Cytoscape
3.7.1 software to construct a compound anti-diabetes target network.
Furthermore, the Network Analyzer plug-in was utilized to analyze the
topological parameters associated with the target degree. The degree
value represents the number of nodes connected by a node. In the
network, the size of the nodes represents the degree value, so the
larger the node in the network is, the higher the degree value of the
node.
Construction of the PPI network
The targets of MLF and RMA for the treatment of T2DM were imported into
the STRING database ([102]https://www.string-db.org/) ([103]Szklarczyk
et al., 2019) to construct a protein‒protein interaction (PPI) network
and analyze the functional interactions between proteins. Analysis was
carried out with Homo sapiens as the organism option and with medium
confidence greater than 0.4. The visualization process was performed
using Cytoscape (Version 3.7.1), and the MCODE plugin in Cytoscape was
used to detect clusters in the PPI network ([104]Liu et al., 2020b).
The parameters were as follows: degree cutoff ≥2, K-core ≥ 4, node
score cutoff ≥0.2, and max depth = 100.
GO and KEGG pathway enrichment analysis
GO (Gene Ontology) function and KEGG (Kyoto Encyclopedia of Genes and
Genomes) pathway enrichment analyses were conducted to explore the core
mechanism and pathway of MLF and RMA anti-diabetes in the Metascape
database ([105]http://metascape.org) ([106]Zhou et al., 2019). We
searched the gene symbols of common targets in the Metascape database
by limiting the species to “Homo sapiens”, and setting the minimum
overlap as 3 and the cutoff p value as 0.01 for enrichment analysis,
including the GO (biological processes, cellular component, and
molecular function) and KEGG pathways. The bubble charts were generated
by the online bioinformatics tool
([107]http://www.bioinformatics.com.cn/), and the target-pathway
network was constructed using Cytoscape 3.7.1 software.
Molecular docking
Classic molecular dynamics were used to analyze the interactions
between compounds and target proteins using AutoDocTools-1.5.6,
PyMOL-1.7.2.1, and Discovery Studio-2020 to elucidate the mechanism of
the antidiabetic activity of these compounds ([108]Liu et al., 2020b).
Compound 3D structures were drawn using ChemDraw 20.0 and Chem3D 20.0
software. Crystal structures of target proteins were obtained from the
RCSB Protein Data Bank (PDB, [109]https://www.rcsb.org/) ([110]Berman
et al., 2000). AutoDockTools-1.5.6 and Discovery Studio-2020 were used
to prepare the cocrystalized ligands split from the receptors and the
active pocket of each target. The molecular docking simulations and
free binding energy calculations were performed using AutoDock
Vina-1.1.2. The binding interactions in the protein‒ligand complex were
analyzed and visualized using Discovery Studio-2020 software and
PyMOL-1.7.2.1.
Western blot
Protein was extracted from livers using RIPA lysis buffer containing
protease/phosphatase inhibitor cocktail (Beyotime Biotechnology).
Antibodies against p-AKT (phospho Ser473, ab66138, Abcam), AKT (#4691,
Cell Signaling Technology), p-GSK3β (phospho Ser9, #9322, Cell
Signaling Technology), and GSK3β (#12456, Cell Signaling Technology)
were used.
Cell preparation
Cells in logarithmic-growth phase were inoculated in four 60 mm culture
dishes and divided into four groups: Con group, Mod group and two drug
groups. When the available space in the cell culture vessel reached 80%
confluency, cells in the Mod and drug groups were treated with
25 μmol/L LCA for 24 h. Subsequently, cells in the drug groups were
administered morusin at final concentrations of 2.5 and 5 μmol/L for
24 h. The cells were then harvested following insulin (1 × 10^−7 mol/L)
stimulation for 30 min.
Real-time polymerase chain reaction analysis
Total RNA was isolated from liver and cells from the previous step
using a total RNA extraction kit (RNAiso Plus, TaKaRa) according to the
manufacturer’s instructions. First-strand cDNA was synthesized from
total RNA using PrimeScript^TM RT Master Mix (TaKaRa). Real-time PCR
was performed with a total volume of 20 μl, which contained 2 μl of
cDNA, 0.4 μl of each 10 μM forward and reverse primer, 5.2 ml of ddH2O
and 10 μl of 2×PerfectStart® Green qPCR SuperMix (TransGen Biotech) on
a CFX96^TM Real-Time PCR Detection System. PCR amplification was
performed using cycling conditions of 94°C for 30 s, followed by 45
cycles of 94°C for 5 s and 60°C for 30 s. Relative gene expression
changes were measured by the comparative Ct method, X = 2^-△△Ct
([111]Dong et al., 2022), using GAPDH as our housekeeping internal
control gene. The primers used for qPCR are listed in
[112]Supplementary Table S3.
Statistical analysis
GraphPad Prism 8.0 software was utilized for all data analyses. All
values are presented as the mean ± standard error of the mean. Multiple
groups or treatments were compared using one-way analysis of variance
(ANOVA). Post-ANOVA comparisons were made using Dunnett’s correction.
Differences were considered significant when p < 0.05.
Results
MLE, MLF, DNJ and MLP attenuated LCA-induced insulin resistance in vitro
The flow chart of this study is displayed in [113]Figure 2. Our
previous study established an insulin-resistant cell model using LCA to
verify it using antidiabetic drugs and to screen plant ingredients
([114]Lv et al., 2019). Our results demonstrated that cellular glucose
uptake in the Mod group was lower than that in the Con group (p <
0.05). At the same time, 25 μmol/L Met treatment reversed this effect
of LCA (p < 0.05), suggesting that the insulin-resistant model was
successfully constructed and can be used to screen hypoglycemic
compounds.
FIGURE 2.
[115]FIGURE 2
[116]Open in a new tab
Integrated workflow for elucidating the active compounds and the
underlying hypoglycemic mechanism of mulberry leaves.
Subsequently, we assessed the effects of MLE, MLF, DNJ, and MLP on
cellular glucose uptake at different concentrations. The results showed
that MLE at all treatment concentrations (0.5, 1 and 2 mg/ml) markedly
increased glucose uptake in the presence or absence of LCA ([117]Figure
3A). Compared with the Mod group, the 100 mg/L MLF, 50 μmol/L DNJ, and
250 mg/L MLP treatments all significantly alleviated cell insulin
resistance induced by LCA. Treatment of L02 hepatocytes with MLF, DNJ
and MLP also increased cellular glucose uptake stimulated by insulin
([118]Figure 3). The MTT assay results showed that cell viability was
obviously affected by MLE, MLF, DNJ or MLP treatment at the indicated
concentrations ([119]Figure 3). These results suggested the
antidiabetic activity of MLF, DNJ and MLP in amending insulin-resistant
hepatocytes.
FIGURE 3.
[120]FIGURE 3
[121]Open in a new tab
MLE, MLF, DNJ, and MLP treatment promotes glucose uptake in
insulin-resistant L02 cells. Glucose intake levels in human L02
hepatocytes treated with MLE (A), MLF (C), DNJ (E) and MLP (G) at
different doses. Viability of L02 hepatocytes treated with MLE (B), MLF
(D), DNJ (F) and MLP (H) at different doses. *p < 0.05, **p < 0.01,
***p < 0.001, ****p < 0.0001, compared with Con group; ^# p < 0.05, ^##
p < 0.01, ^### p < 0.001, ^#### p < 0.0001, compared with Mod group.
RMA, MLF and MLP alleviate glucose tolerance and insulin resistance in ob/ob
mice
To compare the hypoglycemic effects of RMA, MLF and MLP, we fed ob/ob
mice standard chow containing 1% MLF and 1% MLP for 14 weeks. Mice in
the RMA group were administered RMA (50 mg/kg) by gavage. Compared with
their lean counterparts, ob/ob mice exhibited marked increases in water
intake, food intake, body weight, and food utility, which were not
affected by RMA, MLF, or MLP treatments ([122]Supplementary Figure S1).
The OGTTs were performed on the 13th week of drug treatments, and the
results revealed that ob/ob mice exhibited higher glycemic values
following acute oral glucose than lean WT mice ([123]Figure 4B). MLF
significantly decreased blood glucose levels at 30 and 60 min after the
glucose load in ob/ob mice, consistent with the area under the curve
(AUC)-OGTT ([124]Figures 4B,C). RMA and MLP treatments for 13 weeks
significantly reduced fasting blood glucose levels and slightly
decreased AUC-OGTT with no remarkable difference relative to the ob/ob
group ([125]Figures 4B,C). The mice were subjected to ITTs on the 14th
week of drug treatment to assess insulin tolerance. We observed that
ob/ob mice showed insulin intolerance, and the AUC-ITT was
significantly higher than that of lean WT mice, which was obviously
mitigated in the RMA or MLF group. Similarly, RMA, MLF and MLP
treatment significantly decreased serum insulin levels in ob/ob mice
([126]Figures 4F,G). Subsequently, insulin sensitivity was assessed by
HOMA-IR and ISI. As expected, treatment with RMA, MLF and MLP in ob/ob
mice significantly reduced HOMA-IR and enhanced HOMA-ISI ([127]Figures
4H,I) indicating that RMA, MLF and MLP elevate insulin sensitivity in
ob/ob mice. Taken together, these results suggested that RMA and MLF
treatment both ameliorate glucose and insulin intolerance in ob/ob
mice, whereas MLF has a more significant effect.
FIGURE 4.
[128]FIGURE 4
[129]Open in a new tab
MLF and MLP treatment improved metabolic parameters of glucose
tolerance and insulin resistance, and alleviated hepatic steatosis in
ob/ob mice. (A) Experimental scheme of the ob/ob mouse protocol. (B)
OGTT (oral glucose tolerance tests) and (C) AUC of the OGTT on the 13th
week of treatment in ob/ob mice. (D) ITT (insulin tolerance tests) and
(E) AUC of the ITT on the 14th week of treatment in ob/ob mice. (F)
Fasting blood glucose in serum. (G) Serum insulin levels. (H) HOMA-IR
index. (I) HOMA-ISI index. (J) Representative pictures of liver stained
with H&E (magnification: ×40). *p < 0.05, **p < 0.01, ***p < 0.001,
****p < 0.0001, compared with WT group; ^# p < 0.05, ^## p < 0.01, ^###
p < 0.001, ^#### p < 0.0001, compared with the ob/ob group.
MLE and MLP ameliorate hepatic lipid accumulation in ob/ob mice
To evaluate the effects of RMA, MLF, and MLP on ob/ob mouse hepatic
lipid accumulation, we stained the liver tissues with H&E. We observed
that the livers of lean WT mice showed an utterly normal structure with
distinguishable edges and a clear outline. However, the livers of ob/ob
mice had severe hepatic lipid accumulation, the hepatic lobular
structure was unclear, and the cytoplasm was filled with a large number
of fat vacuoles. RMA treatment did not influence the accumulation of
intrahepatic lipids in ob/ob mice, while MLF and MLP treatments reduced
fat vacuoles and ameliorated hepatic lipid accumulation ([130]Figure
4J).
Screened compounds of MLF and RMA and T2DM targets
Network pharmacology was conducted to investigate the mechanism of MLF
and RMA in the treatment of T2DM. A total of 29 flavonoids of Morus
alba L. were collected from the TCMSP database and published literature
([131]Yang, 2010; [132]Chen, 2014; [133]Li, 2017) ([134]Table 2).
SwissTargetPrediction was used for target prediction, and the results
with a probability >0.1 were selected for subsequent analysis. With
repeat targets excluded, 240 drug targets were ultimately obtained. We
retrieved a total of 3,350 putative targets of T2DM from the DrugBank,
OMIM, TTD and GeneCards databases ([135]Figure 5A). We compared these
targets with the predicted MLF targets, and 135 common targets were
filtered as the key targets for testing the antidiabetic activity of
the MLF ([136]Figure 5B). Subsequently, we constructed a
compound-protein network based on the 135 overlapped targets and their
corresponding compounds, composed of 164 nodes and 489 edges
([137]Figure 5D). Network analysis revealed that the average degree of
the 29 flavonoids was 16.8. We obtained 8 compounds (morusin,
kaempferol, quercetin, norartocarpetin, kuwanon C, morusyunnansin L,
morin, and fisetin) that had degree values higher than the average
degree of 16.8. Therefore, these 8 compounds were regarded as potential
bioactive compounds of MLF against T2DM. Moreover, we obtained 4
alkaloids of Ramulus Mori ([138]Table 2) from published literature
([139]Chen et al., 2000; [140]Yang et al., 2015) and a total of 22
targets were predicted through a chemical similarity-based target
search. We constructed a Venn diagram based on the targets of RMA and
T2DM and obtained 14 overlapping targets that were considered the
potential targets of RMA against diabetes ([141]Figure 5C). A
compound-target network comprising 18 nodes and 42 edges was then
constructed ([142]Figure 5E). Network analysis showed that DNJ (RMA1,
degree = 11) and N-methyl-1-deoxynojirimycin (RMA4, degree = 14) had
the highest numbers of connections to different targets. Taken
together, we screened 135 targets and 8 compounds in MLF and 14 targets
and 2 compounds in RMA by network pharmacology analysis.
TABLE 2.
The 29 active compounds of mulberry (Morus alba L.) leaf flavonoids and
4 active compounds of Ramulus Mori alkaloids.
ID Name 2D structure Source
MLF1
2-(8-(2-hydroxypropan-2-yl)-3,4,8,9-tetrahydro-2H-furo[2,3-h]chromen-2-
yl)-5-methoxyphenol graphic file with name
FPHAR_fphar-2022-986931_wc_tfx1.jpg published literature
MLF2
2-(2-hydroxy-4-methoxyphenyl)-8,8-dimethyl-3,4,9,10-tetrahydro-2H,8H-py
rano[2,3-f]chromen-9-ol graphic file with name
FPHAR_fphar-2022-986931_wc_tfx2.jpg published literature
MLF4 Morusin graphic file with name FPHAR_fphar-2022-986931_wc_tfx3.jpg
published literature
MLF7 Kaempferol graphic file with name
FPHAR_fphar-2022-986931_wc_tfx4.jpg TCMSP
MLF8 Quercetin graphic file with name
FPHAR_fphar-2022-986931_wc_tfx5.jpg TCMSP
MLF11 Rutin graphic file with name FPHAR_fphar-2022-986931_wc_tfx6.jpg
published literature
MLF23 Norartocarpetin graphic file with name
FPHAR_fphar-2022-986931_wc_tfx7.jpg TCMSP
MLF24 Kuwanon C graphic file with name
FPHAR_fphar-2022-986931_wc_tfx8.jpg published literature
MLF26 Mornigrol F graphic file with name
FPHAR_fphar-2022-986931_wc_tfx9.jpg published literature
MLF27 Mornigrol G graphic file with name
FPHAR_fphar-2022-986931_wc_tfx10.jpg published literature
MLF28 6-geranylapigenin graphic file with name
FPHAR_fphar-2022-986931_wc_tfx11.jpg published literature
MLF30 (2S)-2′,4-dihydroxy-7-methoxy-8-yl butyrate flavan graphic file
with name FPHAR_fphar-2022-986931_wc_tfx12.jpg published literature
MLF31 isopentenyl-2′,4′-dihydroxy-7-methoxy flavan graphic file with
name FPHAR_fphar-2022-986931_wc_tfx13.jpg published literature
MLF32 isopentenyl-7,2c--dihydroxy-4′-methoxy flavan graphic file with
name FPHAR_fphar-2022-986931_wc_tfx14.jpg published literature
MLF33 Brosimine B graphic file with name
FPHAR_fphar-2022-986931_wc_tfx15.jpg published literature
MLF36 Morachalcone A graphic file with name
FPHAR_fphar-2022-986931_wc_tfx16.jpg published literature
MLF37 Isobavaehaleone graphic file with name
FPHAR_fphar-2022-986931_wc_tfx17.jpg published literature
MLF40 Morusyunnansin J graphic file with name
FPHAR_fphar-2022-986931_wc_tfx18.jpg published literature
MLF42 Morusyunnansin L graphic file with name
FPHAR_fphar-2022-986931_wc_tfx19.jpg published literature
MLF43 Morusyunnansin M graphic file with name
FPHAR_fphar-2022-986931_wc_tfx20.jpg published literature
MLF44 Morusyunnansin N graphic file with name
FPHAR_fphar-2022-986931_wc_tfx21.jpg published literature
MLF45 (2S)-7,2′-dihydroxy-4′-methoxy-8-prenylflavan graphic file with
name FPHAR_fphar-2022-986931_wc_tfx22.jpg published literature
MLF46 (2S)-2′,4′-dihydroxy-7-methoxy-8-prenylflavan graphic file with
name FPHAR_fphar-2022-986931_wc_tfx23.jpg published literature
MLF49 2,4,2′,4′-tetrahydroxychalcone graphic file with name
FPHAR_fphar-2022-986931_wc_tfx24.jpg published literature
MLF50 Euchrenone a7 graphic file with name
FPHAR_fphar-2022-986931_wc_tfx25.jpg published literature
MLF51 Morin graphic file with name FPHAR_fphar-2022-986931_wc_tfx26.jpg
published literature
MLF53 Iristectorigenin A graphic file with name
FPHAR_fphar-2022-986931_wc_tfx27.jpg published literature
MLF54 Tetramethoxyluteolin graphic file with name
FPHAR_fphar-2022-986931_wc_tfx28.jpg published literature
MLF55 Fisten graphic file with name
FPHAR_fphar-2022-986931_wc_tfx29.jpg published literature
RMA1 1-deoxynojirimycin graphic file with name
FPHAR_fphar-2022-986931_wc_tfx30.jpg published literature
RMA2 Fagomine graphic file with name
FPHAR_fphar-2022-986931_wc_tfx31.jpg published literature
RMA3 1,4-dideoxy-1,4-imino-D-arabinitol graphic file with name
FPHAR_fphar-2022-986931_wc_tfx32.jpg published literature
RMA4 N-methyl-1-deoxynojirimycin graphic file with name
FPHAR_fphar-2022-986931_wc_tfx33.jpg published literature
[143]Open in a new tab
FIGURE 5.
[144]FIGURE 5
[145]Open in a new tab
Target or active compound prediction and network construction. (A)
Putative targets for T2DM were retrieved from the Drugbank, OMIM, TTD
and GeneCards databases. (B) The 135 matched targets common between the
predicted mulberry leaf flavonoid targets and the type 2 diabetes
targets. (C) The 14 matched targets common between the predicted
Ramulus Mori alkaloid targets and the type 2 diabetes targets. (D) The
compound-target network implicated in type 2 diabetes using mulberry
leaf flavonoids. The red nodes represent the active mulberry leaf
flavonoids, whereas the blue nodes represent the antidiabetic targets
of the active compounds. (E) The compound-target network implicated in
type 2 diabetes using the Ramulus Mori alkaloids. The orange nodes
represent the active Ramulus Mori alkaloids, whereas the blue nodes
represent the antidiabetic targets of the active compounds. The edges
represent the interactions between compounds and targets, and the node
size is proportional to the degree of interaction.
PPI network of the anti-diabetic targets of MLF and RMA
To explore the PPI relationships of 135 potential protein targets of
MLF and 14 targets of RMA related to the treatment of T2DM, we imported
these data into the STRING database for analysis. Then, the .tsv file
of the PPI data generated in STRING was input into Cytoscape (version
3.7.1) to construct a more intuitive network ([146]Figures 6A,C). In
the PPI network, a node with a larger size and deeper color possesses a
higher degree value. The PPI network of MLF comprised 134 nodes (1
disconnected node was deleted) and 1,552 edges ([147]Figure 6A),
whereas the PPI network of RMA involved 14 nodes and 40 edges
([148]Figure 6C). Similar functional clusters of the PPI network were
selected by MCODE analysis using Cytoscape 3.7.1 software, and the
attribute values of the cluster are listed in [149]Table 3. The MLF
cluster contained 34 nodes and 446 edges ([150]Figure 6B). The average
values of degree centrality, betweenness centrality, and closeness
centrality were 26.23529412, 0.006405972, and 0.842350158,
respectively. We found 16 targets whose degree centrality, betweenness
centrality, and closeness centrality values were greater than the
average. The 16 targets were PTGS2, SRC, MDM2, ESR1, AKT1, VEGFA,
CASP8, MMP9, MAPK1, PPARG, STAT3, ERBB2, EGFR, CASP3, HSP90AA1 and
CTNNB1 ([151]Figure 6B). Moreover, for RMA, the cluster consisted of 6
nodes and 13 edges, and the average values of degree centrality,
betweenness centrality, and closeness centrality were 4.333333,
0.033333335, and 0.896825395, respectively. The top 3 hub targets of
RMA were MGAM, GLB1 and SI ([152]Figure 6D). It is believed that these
hub targets play a major role in treating T2DM by MLF and RMA.
FIGURE 6.
[153]FIGURE 6
[154]Open in a new tab
Construction of the PPI network and core targets. (A) The PPI network
of potential protein targets of MLF in the treatment of T2DM
constructed using Cytoscape and analyzed using NetworkAnalyzer. (B)
MCODE cluster generated from (A). (C) PPI network of potential targets
of RMA in the treatment of T2DM. (D) MCODE cluster generated from (C).
The depth of color represents the degree value, and the node size is
proportional to the degree of interaction.
TABLE 3.
The cluster network parameters of mulberry (Morus alba L.) leaf
flavonoids and Ramulus Mori alkaloids.
Network parameters Value
Number of nodes 34
Number of edges 446
Clustering coefficient 0.850
Network diameter 2
Network radius 1
Network centralization 0.218
Network density 0.795
Shortest paths 1,122 (100%)
Characteristic path length 1.205
Avg. number of neighbors 26.235
Number of nodes 6
Number of edges 13
Clustering coefficient 0.9
Network diameter 2
Network radius 1
Network centralization 0.200
Network density 0.867
Shortest paths 30 (100%)
Characteristic path length 1.133
Avg. number of neighbors 4.333
[155]Open in a new tab
GO enrichment analysis
To further explore the mechanisms of MLF and RMA in T2DM, we performed
a GO enrichment analysis of the 135 predicted targets of MLF and the 14
potential targets of RMA. Our results revealed the top 20 enriched GO
terms of biological process (BP), cellular component (CC), and
molecular function (MF) of MLF and RMA ([156]Figure 7). According to
the BP results ([157]Figure 7A), the functions of active compounds of
MLF in T2DM mainly focused on cell migration, lipid metabolic process
and hydrolase activity, response to hormone stimulus, oxidative stress
and lipid, and were involved in rhythmic processes and immune system
development. The CC results mainly included receptor complex,
perinuclear region of cytoplasm, caveola, postsynapse, and the external
side of the plasma membrane ([158]Figure 7B). For MF ([159]Figure 7C),
the targets mostly involved transmembrane receptor protein tyrosine
kinase activity, protein serine/threonine kinase activity,
transcription coregulator binding, nuclear receptor activity and
insulin receptor substrate binding. The analyses above showed that
these targets are closely related to the processes of regulating kinase
activity, lipid metabolism, and the insulin signaling pathway. The top
three BP terms of RMA were carbohydrate metabolic process, glycoside
catabolic process and glycolipid catabolic process ([160]Figure 7D).
The CC results included lysosomal lumen, azurophil granule lumen and
ficolin-1-rich granule ([161]Figure 7D). For MF, the top 3 terms were
hydrolase activity, hydrolyzing O-glycosyl compounds, glucosidase
activity and alpha-glucosidase activity ([162]Figure 7D). These results
suggested that the potential targets of RMA are highly associated with
carbohydrate metabolism and the regulation of alpha-glucosidase
activity.
FIGURE 7.
[163]FIGURE 7
[164]Open in a new tab
GO enrichment analysis of the antidiabetic targets of MLF and RMA. (A)
Biological processes of MLF; (B) cellular components of MLF; (C)
molecular function of MLF; (D) GO analysis of RMA.
KEGG enrichment analysis
KEGG pathway enrichment analysis was applied to explore the functions
and signaling pathways of MLF and RMA antidiabetic targets. The top 20
KEGG pathways of MLF targets include the PI3K- Akt signaling pathway
(hsa04151), lipid and atherosclerosis (hsa05417), the AGE-RAGE
signaling pathway in diabetic complications (hsa04933), insulin
resistance (hsa04931), and type 2 diabetes mellitus (hsa04930).
Consistent with the above GO analysis, they were closely related to
glucose and lipid metabolism, insulin signaling and oxidative stress
([165]Figure 8A). Then, based on the number of targets involved in each
pathway, a target-pathway network was constructed using Cytoscape
(version 3.7.1) ([166]Figure 8B). Most targets were mainly enriched in
pathways in cancer, PI3K- Akt signaling, and lipid and atherosclerosis.
In addition, AKT1, MAPK1, PIK3R1 and INSR participated in the greatest
numbers of pathways. The top three significant KEGG pathways of RMA
were galactose metabolism (hsa00052), other glycan degradation
(hsa00511), and starch and sucrose metabolism (hsa00500) ([167]Figure
8C; [168]Table 4). The target-pathway network of RMA showed that GLB1,
GLA, GBA and GAA are the core targets involved in the majority of
pathways ([169]Figure 8D). Combined with the GO analysis results, these
results suggest that the mechanism of action of RMA improves insulin
resistance and glucose tolerance through influencing carbohydrate
metabolism by regulating alpha-glucosidase activity.
FIGURE 8.
[170]FIGURE 8
[171]Open in a new tab
KEGG pathway enrichment analysis of the anti-diabetes targets of MLF
and RMA. (A) Pathway enrichment results of MLF at p < 0.01. (B) The
target–pathway network implicated in the mechanism of MLF in type 2
diabetes treatment. (C) Pathway enrichment results of RMA at p < 0.01.
(D) The target–pathway network implicated in the mechanism of RMA in
type 2 diabetes treatment. The red nodes represent the pathways,
whereas the blue nodes represent the targets involved in these
pathways. The edges represent the interactions between the targets and
the pathways, and the node size is proportional to the degree of
interaction.
TABLE 4.
Annotation of KEGG pathways of mulberry (Morus alba L.) leaf flavonoids
and Ramulus Mori alkaloids.
Term ID Description Count p Value Genes
hsa05200 Pathways in cancer 52 9.87066E-56 AGTR1, AKT1, ALK, AR, BRAF,
CASP3, CASP8, CDK2, CDK6, CSF1R, CTNNB1, EDNRB, EGFR, ERBB2, ESR1,
ESR2, F2, FLT3, GSK3B, HDAC1, HDAC2, HSP90AA1, IGF1R, IL2, JAK1, JAK2,
JAK3, KIT, MDM2, MET, MMP1, MMP2, MMP9, NFKB1, NOS2, NTRK1, PDGFRA,
PGF, PIK3R1, PPARG, PRKCA, PRKCB, PRKCG, MAPK1, MAP2K1, PTGS2, RELA,
RXRA, STAT3, TERT, VEGFA, WNT3A
hsa04920 Adipocytokine signaling pathway 7 2.6053E-08 AKT1, JAK2,
NFKB1, PPARA, RELA, RXRA, STAT3
hsa01521 EGFR tyrosine kinase inhibitor resistance 20 4.93654E-30 AKT1,
BRAF, EGFR, ERBB2, GSK3B, IGF1R, JAK1, JAK2, KDR, MET, PDGFRA, PIK3R1,
PRKCA, PRKCB, PRKCG, MAPK1, MAP2K1, SRC, STAT3, VEGFA
hsa04932 Non-alcoholic fatty liver disease 12 5.81113E-12 AKT1, CASP3,
CASP8, GSK3B, INSR, NFKB1, PIK3R1, PPARA, PPARG, RELA, RXRA, EIF2AK3
hsa05216 Thyroid cancer 7 2.7995E-10 BRAF, CTNNB1, NTRK1, PPARG, MAPK1,
MAP2K1, RXRA
hsa04910 Insulin signaling pathway 10 6.41988E-10 AKT1, BRAF, FASN,
GSK3B, INSR, PDPK1, PIK3R1, MAPK1, MAP2K1, PTPN1
hsa05417 Lipid and atherosclerosis 23 3.16371E-25 AKT1, CASP1, CASP3,
CASP8, CYP1A1, GSK3B, HSP90AA1, JAK2, MMP1, MMP3, MMP9, NFKB1, PDPK1,
PIK3R1, PPARG, PRKCA, MAPK1, RELA, RXRA, SELP, SRC, STAT3, EIF2AK3
hsa04933 AGE-RAGE signaling pathway in diabetic complications 17
1.93973E-22 AGTR1, AKT1, CASP3, F3, JAK2, MMP2, NFKB1, PIK3R1, PRKCA,
PRKCB, PRKCD, PRKCE, MAPK1, RELA, STAT3, VEGFA, NOX4
hsa04931 Insulin resistance 13 2.25338E-15 AKT1, GSK3B, INSR, NFKB1,
PDPK1, PIK3R1, PPARA, PRKCB, PRKCD, PRKCE, PTPN1, RELA, STAT3
hsa04520 Adherens junction 10 8.29E-13 ACP1, CTNNB1, EGFR, ERBB2,
IGF1R, INSR, MET, MAPK1, PTPN1, SRC
hsa04913 Ovarian steroidogenesis 6 1.08695E-07 CYP1A1, CYP1B1, CYP19A1,
IGF1R, INSR, PTGS2
hsa04151 PI3K-Akt signaling pathway 32 1.3511E-32 AKT1, CDK2, CDK6,
CSF1R, EGFR, ERBB2, FLT3, GSK3B, HSP90AA1, IGF1R, IL2, INSR, JAK1,
JAK2, JAK3, KDR, KIT, MDM2, MET, NFKB1, NTRK1, PDGFRA, PDPK1, PGF,
PIK3CG, PIK3R1, PRKCA, MAPK1, MAP2K1, RELA, RXRA, VEGFA
hsa04072 Phospholipase D signaling pathway 15 2.02471E-16 AGTR1, AKT1,
AVPR2, EGFR, F2, CXCR1, CXCR2, INSR, KIT, PDGFRA, PIK3CG, PIK3R1,
PRKCA, MAPK1, MAP2K1
hsa04916 Melanogenesis 11 1.05866E-12 CTNNB1, EDNRB, GSK3B, KIT, PRKCA,
PRKCB, PRKCG, MAPK1, MAP2K1, TYR, WNT3A
hsa05222 Small cell lung cancer 10 1.19614E-11 AKT1, CASP3, CDK2, CDK6,
NFKB1, NOS2, PIK3R1, PTGS2, RELA, RXRA
hsa04152 AMPK signaling pathway 10 1.73256E-10 AKT1, CFTR, FASN, HMGCR,
IGF1R, INSR, PDPK1, PIK3R1, PPARG, SIRT1
hsa04930 Type II diabetes mellitus 6 5.75767E-08 CACNA1C, INSR, PIK3R1,
PRKCD, PRKCE, MAPK1
hsa04923 Regulation of lipolysis in adipocytes 6 1.92408E-07 ADORA1,
AKT1, INSR, PIK3R1, PTGS1, PTGS2
hsa00910 Nitrogen metabolism 4 8.67479E-07 CA1, CA2, CA3, CA4
hsa00052 Galactose metabolism 6 2.08301E-15 GAA, GANC, GLA, GLB1, SI,
MGAM
hsa00500 Starch and sucrose metabolism 5 3.57233E-12 AGL, GAA, GANC,
SI, MGAM
hsa00511 Other glycan degradation 5 8.15532E-14 FUCA1, GBA, GLB1,
MAN2B1, GBA2
hsa04142 Lysosome 6 1.81021E-11 FUCA1, GAA, GBA, GLA, GLB1, MAN2B1
hsa00600 Sphingolipid metabolism 4 6.03913E-09 GBA, GLA, GLB1, GBA2
[172]Open in a new tab
Molecular docking findings
We employed molecular docking to analyze the possibility of binding
between the core targets and the active compounds via AutoDockTools. A
previous study proved that a binding affinity < −7.0 kcal/mol indicated
that the two molecules had strong binding activity ([173]Trott and
Olson, 2010). In the current study, we docked two top targets in the
insulin signaling pathway (serine/threonine-protein kinase, AKT and
glycogen synthase kinase-3 beta, GSK3β), PPARγ and ADORA1 with active
compounds of MLF. The results illustrated that most of the binding
energies were < −7 kcal/mol, and the binding energies between PPARγ and
MLF24 as well as ADORA1 and MLF4 were < −8.9 kcal/mol ([174]Table 5).
Therefore, diabetes-associated targets (AKT1 and PPARγ) and two targets
involved in regulating glucolipid metabolism (GSK3β and ADORA1), which
have the lowest free energy binding with their compounds, were selected
for molecular docking and refined by exploring the specific binding
sites. A lower free binding energy value indicates stronger binding to
the target protein. The free binding energies of MLF24 with AKT1 and
PPARγ were −8.51 and −8.99 kcal/mol, respectively. In AKT1, MLF24 had
hydrogen bonding with the ASP-274, ARG-273, GLU-85 and ASN-54 residues;
a Pi-anion interaction with the GLU-17 residue; a Pi-Sigma interaction
with the ILT-84 residue; and hydrophobic interactions with the VAL-270,
TYR-18 and ARG-86 residues of AKT1([175]Figures 9A1,A2). In PPARγ, the
binding affinity was contributed by the following: hydrogen bonding
with the LYS-367, TYR-327, SER-342 and ARG-288 residues; a Pi-Sigma
interaction with the LEU-330 residue; a Pi-Sulfur interaction with the
MET-364 residue; and hydrophobic interactions with the ILE-326,
MET-329, ALA-292, ILE-341, MET-348, LEU-353, CYS-285 and VAL-339
residues of PPARγ ([176]Figures 9B1,B2). MLF42 showed −8.44 kcal/mol
with GSK3β and formed hydrogen binding, carbon‒hydrogen binding, and
Pi-Pi stacking with residues VAL-135, VAL-61, and TYR-134,
respectively. Through hydrophobic interactions MLF42 interacts with the
ALA-83, LEU-188, VAL-70, ILE-62 and LYS-60 residues of GSK3β. MLF4 has
solid binding interactions with ADORA1 (binding energy =
-9.31 kcal/mol). MLF4 docked with six residues to form hydrophobic
interactions in ADORA1 (LEU-253, MET-177, ALA-84, VAL-87, VAL-62 and
HIS-278) and two Pi-Sigma interactions with the LEU-250 and ILE-274
residues, as well as Pi-Pi stacking interactions with the PHE-171
residue ([177]Figures 9D1,D2). The molecular docking results suggested
that hydrogen bonding and hydrophobic interactions were the main forms
of interaction. Collectively, these results implied that kuwanon C,
morusin and morusyunnansin L are the main compounds of MLF, which exert
antidiabetic effects by regulating ATK1, PPARγ, ADORA1, and GSK3β,
respectively.
TABLE 5.
Free binding energies of AKT1, PPARG, GSK3β and ADORA1 with their
corresponding active compounds.
Target Compound Free binding energy (kcal/mol)
AKT1 MLF24 −8.51
MLF42 −8.05
MLF8 −7.59
MLF55 −7.47
MLF7 −7.37
MLF23 −6.89
PPARG MLF24 −8.99
GSK3β MLF42 −8.44
MLF55 −7.71
MLF8 −7.6
MLF7 −7.37
MLF23 −7.36
MLF51 −7.36
MLF54 −7.12
ADORA1 MLF4 −9.31
MLF24 −8.12
MLF54 −7.53
MLF55 −7.25
MLF8 −6.89
MLF51 −6.80
MLF23 −6.72
MLF7 −6.37
[178]Open in a new tab
FIGURE 9.
[179]FIGURE 9
[180]Open in a new tab
Schematic 2D and 3D representations of the molecular docking model and
active sites. Binding modes of kuwanon C (MLF24) to AKT1 (A1), kuwanon
C (MLF24) to PPARγ (B1), morusyunnansin L (MLF42) to GSK3β (C1), and
morusin (MLF4) to ADORA1(D1). (A2, B2, C2, D2): Two-dimensional
patterns of bonds.
Validation of the key targets in ob/ob mice
Validation of the key targets was carried out in ob/ob mice. Compared
with WT mice, the phosphorylated protein levels of AKT (Ser473) and
GSK3β (Ser9) were reduced in the livers of ob/ob mice ([181]Figure
10A). However, MLF treatment increased the protein expression levels of
p-AKT and p-GSK3β in ob/ob mice ([182]Figure 10A). In addition, the
protein expression levels of p-AKT and p-GSK3β were not obviously
influenced by RMA treatment ([183]Figure 10A). MLP treatment increased
the phosphorylation of AKT, but had no apparent effect on the protein
expression levels of p-GSK3β ([184]Figure 10A). In addition, compared
with WT mice, the transcriptional level of PPARG in the livers of ob/ob
mice was significantly increased ([185]Figure 10B). Surprisingly, RMA
treatment increased the expression level of PPARG, while MLF and MLP
treatment had no effect on the expression of the PPARG gene
([186]Figure 10B). Moreover, there was no apparent difference in the
gene expression level of ADORA1 between ob/ob mice and WT mice
([187]Figure 10C). However, RMA and MLP treatments significantly
increased the expression level of the ADORA1 gene, and MLF treatment
markedly reduced the gene expression level of ADORA1 in the livers of
ob/ob mice ([188]Figure 10C).
FIGURE 10.
[189]FIGURE 10
[190]Open in a new tab
Validation of key targets in ob/ob mice and insulin-resistant L02
cells. (A) Protein levels of p-AKT, AKT, p-GSK3β and GSK3β in the
livers of ob/ob mice. mRNA levels of PPARG (B) and ADORA1 (C) in the
livers of ob/ob mice. *p < 0.05, ***p < 0.001, compared with WT group;
^# p < 0.05, ^### p < 0.001, ^#### p < 0.0001, compared with the ob/ob
group. Glucose consumption (D) and cell viability (E) were determined
following treatment with different concentrations of morusin in L02
cells. RT‒qPCR analysis of mRNA levels of ADORA1 (F) and PPARG (G) in
insulin‒resistant L02 cells stimulated with morusin for 24 h *p < 0.05,
**p < 0.01, compared with Con group; ^## p < 0.01, ^### p < 0.001,
^#### p < 0.0001, compared with Mod group.
Morusin facilitates glucose consumption and represses the gene expression of
ADORA1 and PPARG in L02 cells
Our results indicated that morusin (2.5 and 5 μmol/L) markedly reversed
cellular insulin resistance and reliably facilitated glucose
consumption ([191]Figure 10D). In addition, morusin significantly
decreased the cell survival rate in a concentration range (1.25, 2.5, 5
and 10 μmol/L) ([192]Figure 10E). We next examined the potential
effects of morusin on ADORA1 and PPARG gene expression. The expression
level of ADORA1 was slightly increased in the Mod group compared with
the Con group, while 2.5 and 5 μmol/L morusin treatment observably
repressed the expression of ADORA1 ([193]Figure 10F). Similarly,
5 μmol/L morusin also markably inhibited PPARG expression induced by
insulin resistance ([194]Figure 10G).
Discussion
Natural products, including herbal formulas and their extracts, have
been used to treat human diseases through unique systems of theories
and therapies for thousands of years, and have also been increasingly
applied to treat T2DM ([195]Xu et al., 2018). Mulberry leaves, a
traditional Chinese medicine, have been reported to reduce cholesterol
levels, enhance high-density lipoprotein cholesterol, and decrease
serum triglyceride and low-density lipoprotein cholesterol levels in
patients with mild dyslipidemia ([196]Aramwit et al., 2011;
[197]Aramwit et al., 2013). The functional components of mulberry
leaves, mainly flavonoids, alkaloids and polysaccharides, account for
approximately 0.7%–1.3%, 0.09–0.28%, and 1.2%–3.1% of their total dry
weight, respectively ([198]Liao et al., 2008; [199]Lin et al., 2008;
[200]Pu, 2016; [201]Ju et al., 2018). Our present results demonstrated
that the glucose-lowering efficacy of MLF is comparable to those of RMA
and MLP, while the lipid-lowering ability of MLF is superior to that of
RMA after 14 weeks of treatment in ob/ob mice, suggesting that MLF
holds better potential in the treatment of diabetes. Therefore, it is
essential to explore the bioactive components and mechanisms of MLF and
RMA for treating diabetes. We employed network pharmacology and
molecular docking techniques to uncover the active ingredients, their
potential targets, and the signaling pathways of MLF and RMA for the
treatment of diabetes.
Network pharmacology was conducted on the 4 main ingredients (DNJ, FA,
1,4-dideoxy-1, 4-iminod-D-arabinitol and N-methyl-1-deoxynojirimycin)
of RMA to predict the potential targets. Previous studies revealed that
RMA had a high affinity for the disaccharidase active site and could
selectively inhibit it ([202]Liu Z. et al., 2019). Consistently, our
PPI results suggested that MGAM and SI might be the key antidiabetic
targets of RMA. Similarly, our GO analysis and KEGG pathway analysis
results showed that RMA targeted hydrolase, glucosidase and
alpha-glucosidase activity, and was involved in galactose, starch and
sucrose metabolism, and other glycan degradation. Therefore, our
current study suggests that RMA might exert hypoglycemic effects by
inhibiting MGAM and SI to reduce alpha-glucosidase activity.
In contrast, MLF acts through a different mechanism. In the present
study, we screened 26 active components of MLF and 135 hub targets via
network pharmacology. The MLF compound–target network analysis
indicated that morusin, kaempferol, quercetin, norartocarpetin, kuwanon
C, morusyunnansin L, morin, and fisetin are the main antidiabetic
active compounds of MLF. The PPI analysis shows that AKT1 and PPARγ are
the targets with higher degrees in the cluster network. The
serine/threonine kinase Akt, also known as protein kinase B, is a
downstream effector of PI3K. Activated AKT modulates downstream targets
and is involved in energy metabolism in the liver, skeletal muscle, and
adipose tissue ([203]Mora et al., 2004; [204]Batista et al., 2021).
PPARγ is highly expressed in adipose tissue and plays a vital role in
maintaining glucose and lipid homeostasis by regulating genes involved
in fatty acid transport and the triglyceride synthesis pathway
([205]Lee et al., 2012). Glucose and lipid metabolic disorders are key
to T2DM, while one of the main benefits of MLF is alleviating this
disorder.
The GO enrichment analysis further revealed that MLF might regulate
lipid metabolism, especially lipid biosynthetic processes, oxidative
stress, inflammatory responses, and insulin signaling to improve
metabolic disorders and exert antidiabetic effects. Abnormal lipid
metabolism and inflammation are tightly associated with T2DM
([206]Esser et al., 2014; [207]Athyros et al., 2018). A previous study
indicated that flavonoids from mulberry leaves could attenuate
adiposity and regulate lipid metabolism in HFD-fed ICR mice ([208]Zhong
et al., 2020). Consistently, our present studies showed that MLF
alleviates hepatic steatosis in ob/ob mice. Oxidative stress, induced
by an abundance of reactive oxygen species or failure in the
antioxidative machinery, has been considered a significant hallmark for
the pathogenesis and development of T2DM ([209]Rehman and Akash, 2017).
Interestingly, numerous statistically significant BP terms, such as
glucose metabolic process, glucose homeostasis, and response to
glucose, were not at the top of the list. Taken together, we posit that
the benefits of MLF in hypoglycemia are due not to lowering glucose
profiles directly but to its effects on the insulin signaling pathway,
lipid metabolism, and the inflammatory response.
To gain further insight into the mechanism, we obtained 20 signaling
pathways and targets that MLF might correlate with the development of
T2DM and its complications by KEGG pathway analysis. MLF impacted
targets are widely involved in cancer, lipid metabolism, nonalcoholic
fatty liver, insulin resistance and type 2 diabetes. For example,
ADORA1 has been proven to be associated with lipid metabolism. ADORA1
is a G protein-coupled receptor family member and an important drug
target for numerous diseases ([210]Nieto Gutierrez and McDonald, 2018).
Previous studies have shown that activation of ADORA1 in the central
nervous system prevents body weight gain by enhancing adipose
sympathetic innervations to augment adipose tissue lipolysis
([211]Zhang et al., 2021). In contrast, activation of ADORA1 in
peripheral tissues could facilitate HFD-induced obesity in C57BL/6J
mice ([212]Zhang et al., 2021). Both mice fed the HFD diet and patients
with hepatic steatosis showed increased hepatic ADORA1 expression.
Specific inhibition of ADORA1 in the liver helps prevent body weight
gain and alleviate hepatic steatosis ([213]Hong et al., 2019). In the
current study, we found that MLF treatment significantly reduced ADORA1
mRNA expression in the livers of ob/ob mice. Therefore, we speculated
that MLF might regulate lipid metabolism and alleviate hepatic
steatosis by modulating the expression of ADORA1. To determine the
interactions between compounds and their corresponding hub targets, we
calculated their free binding energy and employed molecular docking to
determine their binding mode. The results revealed that morusin (MLF4)
displayed the highest affinity for the ADORA1 protein. Our verified
experiments in human L02 hepatocytes revealed that the expression
levels of ADORA1 genes were upregulated in insulin resistance, while
morusin treatment markedly increased cellular glucose consumption
stimulated by insulin and downregulated ADORA1 expression. Morusin is a
prenylated flavone that allows for versatile salutary effects,
including antioxidant, antitumor, and anti-inflammatory activities
([214]Choi et al., 2020; [215]Panek-Krzysko and Stompor-Goracy, 2021).
Recently, the metabolically beneficial effects of morusin have been
gradually recognized. Morusin downregulates the expression level of
adipogenic transcription factors (PPARγ and C/EBPα) to inhibit lipid
accumulation in 3T3-L1 adipocytes ([216]Lee et al., 2018). MLF
treatment had no obvious effect on PPARG gene expression in ob/ob mice.
Notably, our results revealed that morusin showed higher binding
activity with PPARγ, a regulator of adipocyte differentiation, lipid
storage, glucose metabolism, and insulin sensitivity. Our in vitro
results confirmed that morusin blocked the overexpression of PPARG
caused by insulin resistance. Additionally, we found that kuwanon C
(MLF24) also has a strong bond with PPARγ and ADORA1, suggesting that
kuwanon C might be a potential selective modulator of PPARγ. A recent
study demonstrated that mulberry leaves inhibited adipocyte
differentiation and triglyceride synthesis by regulating the
PPAR-γ-C/EBP-α (CCAAT/enhancer-binding protein type α) signaling
pathway ([217]Liao et al., 2021). Moreover, our current results
revealed that morusin and kuwanon C might be important compounds of MLF
that lower the levels of glucose and lipids, which play their role by
targeting ADORA1 and PPARG.
Impaired insulin signaling is also central to the development of T2DM.
Insulin binding to its receptor is the first step in activating the
insulin signaling pathway, leading to tyrosine phosphorylation of IRS,
which activates IRS and recruits PI3K to tyrosine-phosphorylated IRS
([218]Saltiel and Kahn, 2001; [219]Batista et al., 2021). Subsequently,
phosphatidylinositol (3,4,5)-triphosphate (PIP3) was formed, and
3-phosphoinositide dependent protein kinase (PDK)-dependent AKT was
activated, which then modulated the downstream targets and was involved
in energy metabolism in liver, skeletal muscle and adipose tissue
([220]Mora et al., 2004; [221]Batista et al., 2021). In this study,
treatment with MLF enhanced the phosphorylated protein levels of AKT
(Ser473) and GSK3β (Ser9) in ob/ob mice. Moreover, we observed that
kuwanon C also showed the highest affinity with AKT1, indicating that
kuwanon C might target AKT1, regulate insulin signaling, and be
involved in glucose metabolism. Notably, there has been no report about
the bioactivity of morusyunnansin L. Our results revealed that
morusyunnansin L exhibited a favorable molecular interaction with
GSK3β. Therefore, we speculate that morusyunnansin L (MLF42) plays a
hypoglycemic role by targeting GSK3β to modulate glycogen synthesis and
increase glucose utilization.
Conclusion
In the present study, we have experimentally demonstrated that MLF and
MLP predominantly enhance glucose uptake in insulin-resistant human L02
hepatocytes. The glucose-lowering efficacies of MLF and MLP in ob/ob
mice are comparable to that of RMA, while the lipid-lowering effects of
MLF and MLP are superior to that of RMA, suggesting the potential of
MLF and MLP in antidiabetes and antiobesity. A recent study has shown
that MLF improves high-fat -diet-induced glycolipid metabolic
abnormalities in mice by mediating gut microbiota, although the
specific hypoglycemic active ingredients and targets of MLF remain
unclear ([222]Zhong et al., 2020). Here, our study revealed that DNJ,
FA, and N-methyl-1-deoxynojirimycin are the primary active ingredients
of RMA and target MGAM and SI proteins to lower glucose, whereas
morusin, kuwanon C and morusyunnansin L are the key active compounds of
MLF and play their hypoglycemic roles by targeting key proteins
involved in lipid metabolism (ADORA1 and PPARγ) and insulin signaling
(AKT1 and GSK3β). Additionally, we validated the hypoglycemic effects
of morusin on repressing the expression of the ADORA1 and PPARG genes
to improve insulin resistance in L02 cells. Collectively, our results
shed light on the mechanisms behind the glucose-lowering effects of
MLF, suggesting that morusin and kuwanon C might be selective PPARγ
modulators and possess broad prospects as new drugs or leads against
diabetes.
Acknowledgments