Keywords
mulberry leaf polysaccharides - DKD - meta-analysis - systematic review - renal protection
-
Moraceae
-
Morus
Introduction
Diabetic kidney disease (DKD) emerges as one of the prevailing microvascular complications
of diabetes mellitus. It serves as a significant driver of end-stage renal disease
(ESRD) on a global scale, potentially truncating a patientʼs lifespan by up to 16.9
years [1], while also imposing remarkable financial and social burdens on families and communities.
Fundamental clinical features of DKD frequently entail a decline in the estimated
glomerular filtration rate and/or an escalation in the urinary albumin excretion rate
[2]. Concurrently, typical pathological manifestations encompass the thickening of the
glomerular basement membrane, the proliferation of the mesangial matrix, and the onset
of glomerulosclerosis [3]. Due to its insidious onset, rapid progression, and complex pathogenesis, conventional
therapies targeting glycemic control, blood pressure, and proteinuria have limited
efficacy in
halting the progression of renal disease. Moreover, novel targeted therapies and biological
treatments are still in the research and development phase and are costly, contributing
to the continued rise in the incidence of end-stage DKD in recent years. Consequently,
finding safe, economical, and efficient DKD treatments is crucial.
Natural plants represent a readily available resource for drug development. Morus alba L. (family Moraceae) has both medicinal and dietary uses. The total alkaloids extracted
from its branches effectively treat hyperglycemia and have been approved as the first
natural anti-diabetic drug marketed in China [4]. Mulberry leaf polysaccharides (MLPs), a major natural component of mulberry leaves,
represent a compound polysaccharide mainly composed of glucose, gluconic acid, galacturonic
acid, fructose, xylose, arabinose, galactose, rhamnose, and mannose. The techniques
used for MLP extraction mainly included hot water extraction (HWE), ultrasound-assisted
extraction (UAE), enzyme-assisted extraction (EAE), and microwave-assisted extraction
(MAE). A comparison of these extraction methods and conditions for polysaccharides
from mulberry leaves are summarized in [Table 1]
[5], [6], [7], [8], [9], [10]. The structural characteristics of some of the polysaccharides extracted from mulberry
leaves are shown in [Fig. 1]
[5], [11], [12], [13]. Current research on MLPs has focused primarily on their potential activities in
treating hyperglycemia, obesity prevention, immunomodulatory activity, antioxidant
activity, and ability to regulate the gut microbiota [14], [15]. Despite the positive effects of MLPs on DKD demonstrated by preliminary preclinical
studies, these studies are limited due to their small sample sizes, dispersed evidence,
different experimental conditions used, and inadequate knowledge of the mechanisms
involved. As a result, it is difficult to come to reliable conclusions, which has
limited the use of MLPs in clinical settings. Systematic reviews of preclinical studies
are highly regarded as essential tools for gaining knowledge and identifying pathways
to guide the design of animal experiments [16]. Hence, this study utilized meta-analysis to assess the effectiveness and underlying
mechanisms of MLPs in DKD models, offering preclinical proof and endorsing future
clinical treatment and drug advancement.
Table 1 Extraction conditions, techniques, and yields of MLPs.
|
Extraction method
|
Conditions
|
Yield (%)
|
Ref
|
|
Temperature (℃)
|
Time (min)
|
Solid–liquid ratio (mg/L)
|
Power/Enzyme/pH
|
|
HWE: hot water extraction; UAE: ultrasound-assisted extraction; MAE: microwave-assisted
extraction; EAE: enzyme-assisted extraction
|
|
HWE
|
100
|
180
|
15 : 1
|
–
|
7.2
|
[5]
|
|
HWE
|
80
|
60
|
40 : 1
|
–
|
11.3
|
[6]
|
|
UAE
|
60
|
20
|
15 : 1
|
Ultrasonic power 60 W
|
10.79
|
[9]
|
|
UAE
|
57
|
80
|
53 : 1
|
Ultrasonic power 100 W
|
6.92
|
[10]
|
|
MAE
|
88
|
10
|
–
|
Sample mass 20 g, microwave power 170 W
|
9.41
|
[8]
|
|
HWE-EAE
|
HWE: 85℃, EAE: 45℃
|
HWE:60 min, EAE:50 min
|
30 : 1
|
pH 6.5; Enzyme: pectinase, protease
|
24.04
|
[7]
|
Fig. 1 Structural characteristics of some extracted polysaccharides from mulberry leaves.
Results
The preliminary search produced 489 articles based on the predefined retrieval strategy.
After EndNoteX9.1 software was used to eliminate duplicates, 179 articles were excluded.
Following an examination of the titles and abstracts, an additional 221 papers were
disqualified for several factors, as follows: 1) not animal studies; 2) not MLP-related
studies, or the subjects were not diabetic animal models; 3) reviews, case analyses,
and comments, among others, as specified in the exclusion criteria. Following an additional
screening that included a thorough full-text assessment, a further 81 articles were
disqualified for reasons including 1) combined MLP therapy, 2) subjects not DKD animal
models, 3) lack of predefined outcome indicators, 4) duplicate publication, 5) absence
of controls, and 6) absence of full text. Eventually, the systematic review included
eight qualifying studies [17], [18], [19], [20], [21], [22], [23], [24] ([Fig. 2]).
Fig. 2 The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelinesʼ
methodological flowchart.
The meta-analysis encompassed eight studies, incorporating a collective cohort of
270 animals, distributed between 191 in the treatment group and 79 in the control
group. Individual study sample sizes ranged from 16 to 60 animals. The animal species
included rats and mice, with four studies using male Wistar rats [17], [20], [23], [24], two studies using male Sprague-Dawley (SD) rats [19], [21], one study using male db/db mice [18], and one study using equal numbers of male and female Kunming mice [22]. The weight range of the mice was 18 – 40 g, and the weights of the SD and Wistar
rats were 150 – 270 g. Four studies employed intravenous administration of streptozotocin
(STZ) (30 – 120 mg/kg) to induce models [17], [21], [22], [23], while three more studies used an intraperitoneal injection of STZ (50 – 65 mg/kg)
[19], [20], [24]. One study used spontaneously diabetic db/db mice to generate the animal models
[18]. For anesthesia induction, two studies used ether [17], [23], one study used chloral hydrate [20], and one study used urethane [21], while four studies did not report the anesthetic agents used [18], [19], [22], [24]. Five studies administered drugs orally by gavage (0.1 – 1.0 g/kg/day) [18], [19], [20], [21], [22], while three studies administered drugs by intraperitoneal injection (0.2 – 1.2 g/kg/day)
[17], [23], [24]. Five studies implemented a dose gradient of MLPs (0.1 – 1.2 g/kg/day) [17], [20], [22], [23], [24]. The animals in the control groups were given distilled water treatment or an equal
amount of physiological saline, and throughout the intervention, the total period
ranged from 35 days to 8 weeks. In terms of outcomes, five studies reported primary
outcomes, including blood urea nitrogen (BUN) and serum creatinine (Scr) [19], [20], [21], [22], [24], five studies reported 24-hour urinary protein (UP) [17], [19], [21], [22], [23], and five studies reported urinary microalbumin (UAlb) [17], [18], [20], [23], [24]. Secondary outcomes included fasting blood glucose (FBG), reported in seven studies
[17], [18], [19], [20], [21], [22], [24], and total cholesterol (TC), reported in three studies [18], [19], [22]. Triglyceride (TG) was reported in two studies [18], [19]. In comparison, transforming growth factor beta1 (TGF-β1) protein expression in renal tissues was reported in three studies [18], [20], [22], connective tissue growth factor (CTGF) mRNA expression in renal tissues was reported
in two studies [17], [23], and the kidney index was described in two studies [18], [21]. Several studies also reported other related indicators, such as TGF-β1 mRNA, insulin receptor substrate-1 (IRS-1) mRNA, renal insulin-like growth factor-1/IGF-binding
protein-3 (IGF-1/IGFBP-3) mRNA, renal CTGF protein, C-reactive protein (CRP), tumor
necrosis factor-alpha (TNF-α), serum
IGF-1/IGFBP-3, nuclear factor kappa B (NF-κB) protein, and renal Smad2, Smad3, and Smad4 protein expression. Additionally, some
studies described renal tissue morphology, with some results presented semi-quantitatively,
such as the relative area of the extracellular matrix (ECM) or the average optical
density value (see [Table 2]).
Table 2 Features of the involved studies.
|
Author (Ref.)
|
Species (sex, N)
|
Weight (g)
|
Model method
|
Criteria for modeling
|
Anesthetic
|
Administration Mode
|
Intervention
|
Duration of the Treatment
|
Outcome index
|
|
Treatment group (MLPs)
|
Control group
|
|
Treatment group (MLPs)
|
Control group
|
|
*MLPs: mulberry leaf polysaccharides; BUN: blood urea nitrogen; Scr: serum creatinine;
UP:24-hour urinary protein; UAlb: urinary microalbumin; FBG: fasting blood glucose;
TC: total cholesterol; TG: triglyceride; TGF-β1: transforming growth factor beta1; CTGF: connective tissue growth factor; IRS-1:
insulin receptor substrate-1; IGF-1: insulin-like growth factor-1; IGFBP-3: IGF-binding
protein-3; CRP: C-reactive protein; TNF-α: tumor necrosis factor-alpha; NF-κB: nuclear factor kappa B; ALT: alanine transaminase; AST: aspartate transaminase;
SD rats: Sprague-Dawley rats; STZ: streptozotocin; NM: not mentioned; i. p.: intraperitoneal;
i. v.: intravenous; NS: normal saline
|
|
Liu [24]
|
Wistar rats (male, 45/15)
|
170 – 270
|
By i. p. injection of
STZ (65 mg/kg)
|
Rats with a blood glucose
level over 16.7 mmol/L in 3
different times after 72 h of
STZ injection
|
NM
|
i. p. injection
|
0.8/0.4/0.2 g/kg/d
|
same volume of saline
|
8 weeks
|
-
BUN and SCr
-
UAlb
-
FBG
-
TGF-β1 mRNA
|
|
Song [23]
|
Wistar rats (male, 30/10)
|
170 – 270
|
By i. v. injection of
STZ (65 mg/kg)
|
Rats with a blood glucose
level over 16.7 mmol/L in 3
different times after 72 h of
STZ injection; UAlb>15 ug/ml
|
diethyl ether
|
i. p. injection
|
0.8/0.4/0.2 g/kg/d
|
same volume of saline
|
8 weeks
|
-
UAlb and UP
-
CTGF mRNA
|
|
Huang [21]
|
SD rats (male, 8/8)
|
200 – 240
|
By. i. v. injection of
STZ (30 mg/kg)
|
Rats with a blood glucose
level over 16.7 mmol/L in 3 consecutive days
|
20% urethane,
1.5 g/kg
|
oral gavage
|
0.15
g/kg/d
|
same volume of distilled water
|
8 weeks
|
-
BUN and SCr
-
UP
-
FBG
-
kidney index
-
TNF-α and CRP
|
|
Zhang [20]
|
Wistar rats (male, 32/10)
|
150 – 250
|
By i. p. injection of
STZ (65 mg/kg)
|
Rats with a blood glucose
level over 16.7 mmol/L after 72 h of STZ injection; UAlb>15 ug/ml
|
10% Chloral
hydrate,
0.6 – 0.8 ml/100 g
|
oral gavage
|
0.4/0.2/0.1
g/kg/d
|
same volume of saline
|
8 weeks
|
-
BUN and SCr
-
UAlb
-
FBG
-
TGF-β1 protein and IRS-1 mRNA
|
|
Chen [22]
|
Kunming mice
(male/female,
30/10)
|
18 – 22
|
By i. v. injection of
STZ (120 mg/kg)
|
Rats with a blood glucose
level over 11.1 mmol/L after 72 h of STZ injection;
|
NM
|
oral gavage
|
1/0.5/0.25
g/kg/d
|
same volume of saline
|
35 days
|
-
BUN and SCr
-
UP
-
FBG and.TC
-
TGF-β1 protein and NF-κB protein
-
Renal histopathology
|
|
Zhang [19]
|
SD rats
(male, 8/8)
|
190 – 210
|
By i. p. injection of
STZ (50 mg/kg)
|
Rats with a blood glucose
level over 16.7 mmol/L, with 24H. UP>150% of before modeling, after 2 weeks of STZ
injection
|
NM
|
oral gavage
|
0.2
g/kg/d
|
same volume of distilled water
|
8 weeks
|
-
BUN and SCr
-
UP
-
FBG, TC and TG
-
Serum IGF-1 and IGFBP-3
-
Renal IGF-1 and IGFBP-3 mRNA
-
Renal histopathology
|
|
Zhang [18]
|
db/db and db/m mice
(male, 8/8)
|
35 – 40 (db/db),
15 – 20 (db/m)
|
spontaneous
disease model
|
Rats with a blood glucose
level over 16.7 mmol/L
|
NM
|
oral gavage
|
0.4
g/kg/d
|
same volume of saline
|
6 weeks
|
-
UAlb
-
FBG
-
TC, TG, ALT/AST and HOMA-IR index
-
Renal TGF-β1, CTGF, Smad2, Smad3, Smad4 protein
-
Kidney index
-
Renal histopathology
|
|
Wu [17]
|
Wistar rats (male, 30/10)
|
NM
|
By i. v. injection of STZ (65 mg/kg)
|
Rats with a blood glucose
level over 16.7 mmol/L in 3 consecutive days,
UAlb>15 ug/ml
|
diethyl ether
|
i. p. injection
|
1.2/0.8/0.4
g/kg/d
|
same volume of saline
|
8 weeks
|
-
UAlb and UP
-
FBG
-
CTGF mRNA
|
All studies employed random allocation to the control and intervention groups. Out
of the total, two studies (25%) provided sufficient details about the process of random
sequence generation [19], [22]. One study used a random number table approach, while another utilized computer-generated
random numbers. However, the six remaining studies failed to provide detailed descriptions
of the specific procedures used for random allocation, indicating a high risk of bias
in developing random sequences (selection bias). Out of all the studies included,
two (25%) mentioned that animals were randomly housed [18], [21]. Nevertheless, none of the studies provided information on the baseline characteristics
of the animals, allocation concealment, blinding of animal caregivers, researchers,
and outcome assessors, or random outcome assessment. Consequently, evaluating the
potential bias resulting from the concealment of group allocation and blinding of
subjects is still uncertain. Nearly all studies provided complete data except for
one study, which failed to indicate whether the missing animals affected the final
results [20]. None of the studies discovered any additional sources of bias. The risk of bias
summary for each study is presented in [Fig. 3].
Fig. 3 Risk of bias evaluation. a Risk of bias graph. b Risk of bias summary. + = low risk of bias, – = high risk of bias,? = unclear risk
of bias
Five research studies examined the influence of MLPs on Scr. Compared to the control
group, MLPs substantially decreased Scr (n = 174, SMD = − 1.45, 95% CI [− 2.27, − 0.63],
P = 0.0005; heterogeneity: Chi² = 15.88, P = 0.003; I²=75%; [Fig. 4 a]). Five studies also reported the effect of MLPs on BUN. The presence of a significant
difference in BUN level was noteworthy between the treatment group and control group
(n = 174, SMD = − 1.19, 95% CI [− 2.22, − 0.16], P = 0.02; heterogeneity: Chi² = 26.58,
P < 0.0001; I² = 85%; [Fig. 4 b]). Notably, UP was evaluated in five studies, revealing significant heterogeneity
in their outcomes. Utilizing a random-effects model, a meta-analysis elucidated that
the experimental group manifested a lowered UP levels relative to the control group
(n = 152, SMD = − 2.29, 95% CI [− 3.65, − 0.92], P = 0.001; heterogeneity: Chi² = 29.10,
P < 0.00 001; I² = 86%; [Fig. 5 a]). There have been five studies that discussed how MLPs affect UAlb. The data indicated
that the experimental group had significantly lower UAlb levels following MLP treatment
in comparison to the control group (n = 198, SMD = − 1.51, 95% CI [− 2.18, − 0.83],
P < 0.0001; heterogeneity: Chi² = 12.85, P = 0.01; I² = 69%; [Fig. 5 b]). Given the notable statistical heterogeneity observed across these studies, a random-effects
model was adopted alongside subgroup analysis to elucidate the underlying factors
contributing to the heterogeneity further.
Fig. 4 Forest plot: effect of MLPs on serum serum creatinine (a) and blood urea nitrogen (b) levels vs. control.
Fig. 5 Forest plot: effect of MLPs on serum 24-hour urinary protein (a) and urinary microalbumin (b) levels vs. control.
FBG served as an outcome measure in seven studies. Relative to the control group,
the MLP group exhibited a notably lower FBG (n = 230, SMD = − 2.34, 95% CI [− 2.79,
− 1.88], P < 0.00 001; heterogeneity: Chi² = 8.22, P = 0.22; I² = 27%; [Fig. 6 a]). TC values were documented in three investigations. The meta-analysis unveiled
that relative to the control group, the experimental group displayed a significantly
diminished TC level (n = 72, SMD = − 1.54, 95% CI [− 2.86, − 0.22], P = 0.02; heterogeneity:
Chi² = 8.77, P = 0.01; I² = 77%; [Fig. 6 b]). TG levels were reported in two studies. The pooled results did not differ significantly
from the control group, despite the individual study results suggesting statistically
significant differences (n = 32, SMD = − 6.42, 95% CI [− 13.90, 1.07], P = 0.09; heterogeneity:
Chi² = 10.79, P = 0.001; I² = 91%; [Fig. 6 c]). The meta-analysis of the three
included studies that assessed the expression of TGF-β1 protein showed that the experimental groupʼs TGF-β1 protein levels were significantly lower than those of the control group (n = 98,
SMD = − 2.32, 95% CI [− 3.99, − 0.65], P = 0.007; heterogeneity: Chi² = 14.16, P = 0.0008;
I² = 86%; [Fig. 7 a]). The impact of MLP therapy on CTGF mRNA expression was documented in two studies.
Relative to the control cohort, the experimental group displayed a significant reduction
in CTGF mRNA level (n = 80, SMD = − 1.20, 95% CI [− 1.79, − 0.61], P < 0.0001; heterogeneity:
Chi² = 1.16, P = 0.28; I² = 13%; [Fig. 7 b]). Two studies examined the influences of MLPs on the kidney index. Following MLP
treatment, the experimental group exhibited a significantly lowered kidney index relative
to the control group (n = 32, SMD = − 1.54, 95% CI [− 2.37, − 0.72], P = 0.0002; heterogeneity:
Chi² = 0.30, P = 0.58; I² = 0%; [Fig. 7 c]).
Fig. 6 Forest plot: effect of MLPs on serum fasting blood glucose (a), total cholesterol (b), and triglyceride (c) levels vs. control.
Fig. 7 Forest plot: effect of MLPs on serum TGF-β1 protein (a), CTGF mRNA (b), and kidney index (c) vs. control.
Three studies evaluated pathological changes in renal tissues [18], [19], [22]. Following H&E or periodic acid-Schiff (PAS) staining, the MLP treatment group demonstrated
much less proliferation of the mesangial matrix and mesangial cells than the control
group, with a remarkable improvement of basement membrane thickening. The electron
microscope verified these findings [22]. Optical microscopy results reported in two studies demonstrated improvements in
epithelial cell swelling and reduced interstitial inflammatory cell infiltration in
the drug-treated group [19], [22]. Additionally, it was discovered that the relative area of the ECM was substantially
smaller than that of the control group [19]. The average optical density value (IOD)/AREA in the drug group was
significantly lower than in the control group, according to the analysis of the IOD
value of the PAS-positive basement membrane and the pixel area of the glomerular vascular
tuft (AREA) [18].
Subgroup analysis of Scr revealed significant heterogeneity attributable to differences
in the injected dose of STZ (I² = 93.4%; P < 0.05) and species variation (I² = 86.9%;
P < 0.05). A comparison of Scr levels between the different rodent types indicated
that MLPs exhibited superior efficacy in SD rats (SMD, − 3.20; 95% CI, [− 4.34, − 2.05])
compared to mice (SMD, − 0.82; 95% CI, [− 1.56, − 0.08]) and Wistar rats (SMD, − 0.78;
95% CI, [− 1.24, − 0.31]). Furthermore, establishment of the animal models using low-dose
STZ (STZ ≤ 60 mg/kg) led to better improvements in the Scr levels compared to those
given high STZ doses (STZ > 60 mg/kg) (SMD, − 3.20; 95% CI, [− 4.34, − 2.05] vs. SMD,
− 0.79; 95% CI, [− 1.18, − 0.39]) (See Table 1S, Supporting Information). In the UAlb subgroup analysis, significant heterogeneity
was observed, possibly due to species variation (I² = 80.4%; P < 0.05), duration of
drug administration (I² = 80.4%; P < 0.05), and different
methods used for model establishment (I² = 80.4%; P < 0.05). Comparative analysis
of UAlb among the different rodent types showed that MLP treatment was more effective
in db/db mice (SMD, − 3.3; 95% CI, [− 4.94, − 1.66]) compared to Wistar rats (SMD,
− 1.29; 95% CI, [− 1.87, − 0.71]). When MLPs were administered for less than 8 weeks,
a more significant reduction in UAlb was observed compared to treatment for longer
than 8 weeks (SMD, − 3.3; 95% CI, [− 4.94, − 1.66] vs. SMD, − 1.29; 95% CI, [− 1.87,
− 0.71]). Furthermore, the use of spontaneous db/db mice resulted in better improvement
in UAlb compared to STZ-induced diabetic models (SMD, − 3.3; 95% CI, [− 4.94, − 1.66]
vs. SMD, − 1.29; 95% CI, [− 1.87, − 0.71]) (see Table 2S, Supporting Information). Species stratification in the BUN subgroup analysis substantially
reduced the heterogeneity of the results (I² = 81.2%; P < 0.05). Comparative study
of BUN among the different rodent types showed that MLP treatment
was more effective in SD rats (SMD, − 3.03; 95% CI, [− 5.99, − 0.08]) compared to
mice (SMD, − 1.28; 95% CI, [− 2.06, − 0.51]) and Wistar rats (SMD, − 0.03; 95% CI,
[− 0.48, 0.42]) (see Table 3S, Supporting Information). The UP subgroup analysis did not reveal any significant
impacts on the heterogeneity of the data concerning the amount of STZ injection, type
of rodent, dosage of the drugs, or duration of drug administration (see Table 4S, Supporting Information).
Sensitivity analysis of Scr, BUN, UAlb, and UP did not reveal any significant bias,
indicating the stability of the meta-analysis results (Fig. 1S, Supporting Information).
Notably, none of the studies included in the analysis reported any adverse events,
making it difficult to determine whether or not the MLP treatment was related to any
complications or adverse events.
Discussion
To our knowledge, this study represents the first meta-analysis of preclinical evidence
that examines the protective benefits of MLPs on DKD. The results indicated that MLP
treatment offers multifaceted protection against DKD, suggesting its potential as
a therapeutic agent for DKD. Nevertheless, the heterogeneity in the outcomes must
not be disregarded. This heterogeneity may partially stem from differences among the
included studies, including pharmacological variations, differences in the type of
rodent used as models, and the study designs. Sensitivity analysis suggested the stability
of the results. For further exploration, a random-effects model was cautiously employed.
We conducted subgroup analyses of outcome measures, specifically, different doses
of STZ injection, rodent type, drug doses, and duration of drug administration. The
results indicated significant differences in the efficacy of MLP treatment associated
with the animal model used, with more substantial
improvements in the Scr and BUN levels observed in SD rats compared to the other animal
models (P < 0.05). However, improvements in UAlb were more marked in db/db mice (P < 0.05).
It is thus suggested that the animal model should be specifically selected to observe
more significant differences in outcomes and treatment efficacy. The selection of
the STZ dosage is a crucial determinant for successfully establishing the DKD model.
Lower dosages of STZ may not effectively cause diabetes as anticipated. However, high
dosages of STZ can lead to either the mortality of the animal or nephrotoxicity [25]. The subgroup analysis in this study revealed notable disparities in the improvement
of Scr and UAlb between high-dose STZ (≥ 60 mg/kg) and low-dose STZ (< 60 mg/kg),
with animals administered low-dose STZ demonstrating superior effectiveness. A possible
explanation is that besides inducing hyperglycemia, high-dose STZ has toxic effects
on the
kidneys, which can act as a confounding factor in animal models of DKD. Considering
that many studies still use high doses of STZ for modeling, it is recommended that
future DKD research employ appropriate STZ doses (40 – 60 mg/kg) for successful modeling
without the induction of nephrotoxicity. In clinical practice, the dose-response and
time-response interactions of drugs are crucial. The subgroup analysis of UAlb indicated
that the duration of drug action had a marked effect on the treatment efficacy. In
contrast, the dosage did not significantly impact the results. The effectiveness of
MLP treatment did not improve with time. Nonetheless, it revealed a negative tendency
in the later stages, indicating that treatment duration may be a source of heterogeneity.
We tentatively attribute this to the progressive and irreversible nature of DKD, where
prolongation of MLP treatment only delays rather than reverses DKD progression. As
the disease progresses, UAlb levels may not
accurately reflect the extent of kidney damage, and concurrent evaluation of kidney
function using indicators such as Scr, BUN, and creatinine clearance is thus recommended.
The results of the subgroup analysis should be considered carefully due to the small
sample size. Because of the limited number of articles, we could not perform a meta-regression
analysis.
For diabetics, hyperglycemia has traditionally been regarded as the triggering factor
for DKD, where dysfunctions in glucose and lipid metabolism contribute to hemodynamic
disturbances in renal blood flow, thereby constituting the primary pathological mechanisms
underlying DKD. Numerous studies show that keeping blood glucose levels close to normal
can prevent the development of overt proteinuria and lower the urinary albumin excretion
rate [26]. Therefore, improving blood glucose control is believed to be effective for renal
protection. Numerous preclinical investigations have demonstrated the strong hypoglycemic
bioactivity of MLPs [14], [15]. MLPs can reduce the levels of free fatty acids in mice with type 2 diabetes [27]. According to recent research, isolated and purified MLP-2C can decrease cholesterol
levels by enhancing the conversion of cholesterol to
bile acids, reducing the generation of endogenous cholesterol, and raising cholesterol
efflux [5]. In comparison to the control group, the current meta-analysisʼs findings likewise
demonstrated a significant drop in FBG and an improvement in TC following MLP treatment.
Two studies involving TG levels suggested that MLP treatment could lower TG, although
no positive results were observed after the combination. The heterogeneity of these
two studies (P = 0.001, I² = 91%) may have led to wide confidence intervals in the
random-effects model and produced negative results (P = 0.09). Confirmation of the
kidney protective effects of MLPs concerning improvements in glucose and lipid metabolism
necessitates additional investigation. Considering the high variability in FBG measurements,
future evaluations could include the measurement of HbA1c or glycosylated albumin
better to assess the impact of glucose metabolism on renal function.
The molecular and biological mechanisms underlying the renoprotective effects of MLPs
have not been fully elucidated. The possible mechanisms by which MLPs treat DKD can
be summarized as follows ([Fig. 8]): (1) anti-fibrosis: renal interstitial fibrosis is seen as the last common pathway
that leads to renal failure in DKD. Along with the downstream transcription factor
Smad, TGF-β1 is an essential mediator of fibrosis [28]. Dysregulation of the TGF-β1/Smad pathway may either promote ECM deposition directly or induce epithelial cells
to transform into myofibroblasts through the epithelial-mesenchymal transition (EMT),
resulting in renal fibrosis [29]. CTGF, as a downstream effector of TGF-β1, accelerates the progression of renal fibrosis induced by TGF-β
[30]. The comprehensive findings from prior research [17], [18], [20], [22], [23], [24] unveiled that MLP therapy markedly reduced the levels of both TGF-β1 protein and CTGF mRNA, alongside diminished expression of TGF-β1 mRNA and Smad protein relative to the control group. These findings indicated that
MLPs may reduce activation of the TGF-β1/Smad signaling pathway, and downregulate the mRNA expression of CTGF, thus mitigating
the progression of diabetic kidney fibrosis. (2) Anti-inflammation: earlier research
has demonstrated that one of the critical factors in the development of DKD is inflammation.
The transcription factor NF-κB is essential for initiating and regulating inflammatory responses. It can cause
over-activation of pathways linked to TNF-α, IL-6, hs-CRP, and MCP-1 (monocyte chemotactic protein-1), setting off a chain reaction
of
inflammatory responses and worsening damage to renal tissue [31]. Animal experiments [21], [22] have shown that MLPs can downregulate NF-κB expression in the kidney tissues of DKD mice and rats and reduce serum TNF-α and hs-CRP levels. This implies that MLPs may modulate the protein expression of
NF-κB in the kidney, limit the production of inflammatory factors, alleviate renal tissue
inflammation, reduce mesangial matrix proliferation, and thus mitigate DKD. (3) Dysregulation
of the growth hormone–insulin-like growth factor–insulin-like growth factor binding
protein (GH–IGF–IGFBP) axis: another element in the pathophysiology of DKD is the
GH–IGF–IGFBP axis [32]. Serum IGF-1 induces renal interstitial cell proliferation and has been linked to
proliferative alterations in the kidneyʼs vasculature. Studies have shown that increased
expression of
renal IGF-1 receptors may lead to kidney hypertrophy, a hallmark of DKD. A meta-analysis
of clinical studies revealed a correlation between high levels of IGF-1 or an elevated
IGF-1/IGFBP-3 ratio and the advancement of renal disease, leading to a higher all-cause
mortality rate [33]. It was found that MLP treatment decreased the levels of IGF-1/IGFBP-3 mRNA in both
the serum and kidney [19]. This suggests that the therapeutic impact of MLPs on DKD may be attributed to the
inhibition of the IGF-1/IGFBP-3 signaling pathway. However, the specific mechanism
awaits further investigations. (4) Regulation of metabolism: as described, dysregulation
of glucolipid metabolism is a crucial pathological process involved in DKD. In addition
to exerting renoprotective effects by the direct lowering of glucose and lipid levels,
MLPs were shown to inhibit the development of insulin resistance [34].
Preliminary studies have found that MLPs can regulate the expression of IRS-1, AdipoR1,
and resistin mRNA in DKD rats [20], [35], [36], among which, improving insulin sensitivity may play an important role in stabilizing
glucose–lipid metabolism and thus mitigating the progression of DKD. However, further
in-depth molecular studies are required to assess the mechanism underlying the effects
of MLPs.
Fig. 8 A schematic representation of MLP renoprotective mechanisms in DKD (the red solid
arrow indicates the path and mechanism leading to renal damage in the diabetic animal
model, while the green dovetail arrow indicates the path and mechanism of MLPs treatment).
Before researching the biological activities and applications of polysaccharides,
extracting and purifying them is necessary. Variations in extraction procedures, separation
techniques, and purification processes can result in discrepancies in the chemical
makeup of polysaccharide extracts, which can impact the reliability of subsequent
evaluations of biological activity. Currently, due to their significant diversity
and complex chemical compositions and structures, there is no standardized, simple,
low-cost, and effective separation and purification system for MLPs [15]. The different extraction methods lead to differences in the yield of MLPs, as is
shown in [Table 1]. Among the various reported extraction processes on MLPs, Yang et al. optimized
the extraction conditions under the synergy of HAE and EAE by Box–Behnken design and
obtained the best collaborative extraction processes with the extraction rate of MLPs
reaching as high as 24.04 ± 0.98% [7]. It is thus recommended that future efforts focus on quality control and the establishment
of standardization in the extraction and purification process, fully considering the
samplesʼ purity to ensure the productʼs safety and effectiveness.
Compared to clinical trials, animal experiments are more exploratory in their research
protocols and intervention processes. The quality of this preclinical systematic review
was rated as moderate. The risk of bias tool for animal studies (RoBT), which was
established by the Systematic Review Center for Laboratory Animal Experimentation
(SYRCLE), has been widely used in preclinical systematic reviews and recommended by
other organizations as a standard for methodological quality assessment [37]. We recommend that future researchers adhere to SYRCLEʼs RoBT for evaluating the
methodological quality of animal experiments and complying with the reporting criteria
of preclinical studies as defined in the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines [38]. It is crucial to record both the characteristics of animals and baseline data carefully.
Additionally, providing thorough descriptions of the
randomization and allocation concealment processes, estimating appropriate sample
sizes, maintaining rigorous blinding protocols, and ensuring outcome blinding are
imperative to elevate the quality of animal experiments and promote reporting standards.
The investigation into MLPs is still in its initial phases and has concentrated chiefly
on cellular and animal models. There is a lack of enough clinical research to confirm
the effectiveness and safety of these substances in humans. Hence, we propose enhancing
interdisciplinary cooperation and carrying out controlled human trials under appropriate
conditions integrating pharmacokinetics, toxicology, and clinical experiments. This
approach will enable a comprehensive investigation of the clinical effectiveness,
optimal dosage, and treatment duration of MLPs, thereby thoroughly assessing their
efficacy and potential risks.
This systematic review may lack toxicological information. While a few of the included
studies described normal levels of alanine transaminase (ALT) and aspartate transaminase
(AST) following treatment with MLPs, this is still far from actual toxicological reports.
We advise that future researchers enhance the documentation of toxicological findings.
The limitations of our study may include the following. (1) Research on natural materials
associated with traditional Chinese medicine primarily focuses on Southeast Asia.
Due to probable language barriers, only papers written in Chinese or English were
included in the search, leading to an unavoidable selection bias. (2) The included
studies lacked negative results, and the animal models did not incorporate animals
with relevant comorbidities (the animals used in the studies lacked relevant comorbidities
or risk factors), potentially leading to overestimating the treatment efficacy [39]. Animal models having comorbidities, such as obesity, hyperlipidemia, hypertension,
or other risk factors, should be used when conditions allow. This may better reflect
the physiology of diabetic patients and assist in the therapeutic translation of experimental
outcomes [40]. (3) For the multiple-dose group, we adopted the strategy
recommended by CHSRI, combining them to form a single-dose group. However, this approach
also loses some information regarding dose-response relationships. (4) Mechanism-related
indicators are insufficient for the comprehensive description of the effects of MLPs
on renal protection mechanisms. Moreover, the studies on relevant indicators were
independent, lacking connections between mechanisms, and are thus unable to reflect
the relationships between different mechanisms. Future studies are advised to consider
high-throughput analysis to assist in revealing the primary targets or critical pathways
linked with the impacts of MLPs.
In conclusion, preclinical in vivo evidence indicates that MLPs protect the kidneys and are thus potential candidates
for treating DKD. Further high-quality, large-sample, multi-mechanism studies are
required to assess the effectiveness, safety, and renal protective mechanisms of MLPs.
Additionally, it is unknown whether the results observed in animal models apply to
humans, necessitating further evaluation by clinical trials.
Material and Methods
The study was registered at PROSPERO under registration number CRD42022309117. The
systematic review and meta-analysis adhered to the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) standards, the preferred reporting items for systematic
reviews and meta-analyses.
A computerized search was performed in the Web of Science, Cochrane Central Register
of Controlled Trials, EMBASE, and PubMed databases, as well as in Chinese databases,
including SinoMed, Wanfang, VIP, and CNKI, to identify studies investigating MLP treatment
of DKD in animal models. Every search strategy was used from the databaseʼs creation
until January 2024. The search encompassed articles published solely in Chinese and
English. In addition, the bibliographies of the incorporated articles underwent manual
scrutiny to uncover additional relevant research. The full details regarding the literature
search strategy are to be found in Table 5S (Supporting information).
Two authors (Y. W. and J. Z.) independently reviewed the research abstracts and titles
and then conducted full-text reviews to determine which papers should be included
or excluded. The inclusion criteria were as follows. (1) Selection of diabetic animal
models as the study model. Models can be constructed using many methods, with the
criterion for success being a FBG level of more than 11.1 mmol/L. (2) The treatment
group received exclusively MLPs at any dosage, whereas the control group received
non-functional fluids of the same volume or no therapy. There were no limitations
on the delivery method and the formulation. (3) The primary outcome measures were
UAlb, UP, BUN, and Scr. For a more comprehensive assessment of renal damage, at least
one of four parameters, namely, Scr, BUN, UP, or UAlb, had to be included. The secondary
outcomes assessed in this study were FBG, TC, and TG levels. Additionally, the study
investigated the protective mechanism of MLPs in DKD by
measuring TGF-β1 protein levels and CTGF mRNA expression, among other factors. (4) Randomized controlled
trials. The following were the criteria for exclusion: (1) studies that are not in vivo, including in vitro investigations, clinical trials, case reports, reviews, editorials, and abstracts;
(2) alternative animal models; (3) MLP treatment combined with other medications;
(4) absence of predetermined primary outcome measures; (5) lack of a control group;
(6) duplicate publications.
The following information was independently extracted by two authors, identified as
B. C. and D. W.: (1) the year the study was published and the name of the first author;
(2) specific information about the animals used in each study, including the number,
species, sex, and weight; (3) how the animal models were developed, including the
drug dosages needed, the method of administration, and the criteria used to determine
if the model was successful, as well as details about the anesthesia protocols used;
(4) the intervention procedures for both the treatment and control groups will include
details such as the method of administration, the dosages of drug given, and the length
of time the intervention will last; (5) the primary and secondary outcome measures
will be used to evaluate the effectiveness of the intervention. The variables were
retrieved from the final time if the results were acquired at several time points.
All data were collected if results were obtained from
interventions in subgroups with varying dosages. Efforts were made to reach out to
the authors for more information, but without a response, the data were obtained through
WebPlotDigitizer. During the data extraction procedure, discrepancies were handled
by conversation or discussion with a third party (Y. R.).
The assessment of study quality for each included study was independently undertaken
through two authors (Y. W. and B. C.) using the SYRCLEʼs RoBT [37]. This tool comprises 10 components, encompassing aspects, such as sequence generation,
baseline characteristics, allocation concealment, random housing of animals, blinding
for caregivers and researchers, random outcome assessment, blinding for outcome assessors,
incomplete outcome data reporting, selective outcome reporting, and identification
of other potential sources of bias. Each component is assigned a score of 1 point,
yielding a total score out of 10. In cases of disagreement, a third party (Y. R.)
intervened to resolve discrepancies through arbitration.
When there were differences in the dosage of MLP among subgroups, the Cochrane Handbook for Systematic Reviews of Interventions (CHSRI Version 6.4, 2023) [41] suggested a technique to handle this by merging the results of the subgroups with
varying dosages into a single treatment group. The formula employed for merging continuous
variables was as follows:
The predetermined outcomes were regarded as continuous variables, and their evaluation
was implemented utilizing the mean difference (MD) with a 95% confidence interval
(CI). In cases where there were discrepancies in measurement methods or units used
for outcome reporting, the standardized MD (SMD) was employed rather than MD to compute
the overall effect size. It was attempted to evaluate the statistical heterogeneity
through the Cochrane Q-test and the I2 statistic; a value of I2 below 50% was suggestive of the adoption of a fixed-effects model, while I2 surpassing 50% led to the utilization of a random-effects model. It was attempted
to implement subgroup analysis to figure out the influences of variables involving
animal species, STZ injection dosage, drug dosage levels, and treatment duration on
the outcomes, revealing potential sources of clinical heterogeneity. For the purpose
of conducting sensitivity analysis, STATA 16.0 was employed
to ascertain the robustness and consistency of the primary outcomes. Meta-regression
was precluded due to the restricted number of studies. The RevMan 5.4.1 was utilized
for meta-analysis. Given the inclusion of less than 10 studies, no publication bias
was analyzed. P subordinate 0.05 was regarded to signify statistical significance.
Contributorsʼ Statement
Data collection: Y.Wang, J.Zhang; design of the study: Y.Wang, Y.Ruan; statistical
analysis: D.Wang, B.Chen; analysis and interpretation of the data: B.Chen, J.Zhang;
drafting the manuscript: Y.Wang, B.Chen; critical revision of the manuscript: Y.Wang.
