Angiopoietin‐like 4 is a potential biomarker for diabetic kidney disease in type 2 diabetes patients

Abstract

Aims/Introduction

The association between serum angiopoietin‐like 4 (ANGPTL4) levels and the severity of diabetic kidney disease (DKD) in patients with type 2 diabetes mellitus remains unclear.

Methods

A total of 1,115 type 2 diabetes mellitus patients were analyzed in this cross‐sectional study. DKD index included DKD stages defined by estimated glomerular filtration rate, the albuminuria grades and DKD risk management grades. Serum levels of ANGPTL4 and other biomarkers were detected. Multivariable‐adjusted linear and logistic analyses were used to study the association between ANGPTL4 and DKD. The protein levels of ANGPTL4 were assessed in the kidney. Renal tubular cells were stimulated with glucose to study ANGPTL4 expression.

Results

Compared with the participants in the third or fourth quantile of ANGPTL4, those in the first or second quantile of ANGPTL4 were younger, with lower glycated hemoglobin, triglycerides and urinary albumin creatinine ratio (all P < 0.05). There was a negative nonlinear relationship between ANGPTL4 and estimated glomerular filtration rate in type 2 diabetes mellitus patients. One standard deviation increased serum ANGPTL4 levels, the odds ratio of having DKD was 1.40 (95% confidence interval 1.08–1.80). The mediation analysis showed that triglycerides did not mediate the association between ANGPTL4 and DKD. Furthermore, ANGPTL4 could be the strongest among multiple panels of biomarkers in its association of DKD. Compared with mice at 8 weeks‐of‐age, db/db mice at 18 weeks‐of‐age had increased ANGPTL4 expression in glomeruli and tubular segments. In vitro, glucose could stimulate ANGPTL4 expression in tubular cells in a dose‐dependent manner.

Conclusions

ANGPTL4 could be a potential marker and therapeutic target for DKD treatment.

This study reported that increased serum angiopoietin‐like 4 levels were related to the increased severity of diabetic kidney disease in type 2 diabetes mellitus patients.

INTRODUCTION

Diabetic kidney disease (DKD) is one of the main complications in type 1 and 2 diabetic patients, as well as the major cause of end‐stage kidney disease. Among those, approximately 30–40% of diabetes patients progress to DKD, and approximately 30% finally progress to end‐stage kidney disease ¹ , ² . The pathological development of DKD is complicated, and consists of multiple stages with injury sites spanning from glomeruli to tubular segments. Although elevated levels of microalbuminuria could predict the progression of DKD ³ , the regression of microalbuminuria has frequently occurred in diabetes patients after years of follow up ⁴ , ⁵ , and patients can progress without albuminuria ⁶ , implying that other markers in combination with microalbuminuria might assess DKD progression more precisely.

From animal and human studies, a variety of biomarkers have been identified that are associated with the injury of the glomerular, proximal tubule or distal convoluted tubule, respectively ⁷ , ⁸ , ⁹ . Nevertheless, whether glomeruli and tubule epithelial cells share any common injury biomarker is entirely unknown. If there is any, it would be a very interesting and potential surrogate for DKD diagnosis and intervention.

ANGPTL4 is a member of an angiopoietin‐like protein family. It is expressed in a variety of organs, such as the liver, adipose tissue and intestine ¹⁰ . Once secreted into circulation, it inhibits lipoprotein lipase (LPL) activity and regulates angiogenesis ¹¹ , ¹² . Individuals with loss‐of‐function variants of ANGPTL4 are associated with improved glucose homeostasis and a lower odd ratio of type 2 diabetes mellitus compared with controls ¹³ . Previously, we and other groups showed that ANGPTL4 levels were increased in the blood and aqueous and vitreous in type 2 diabetes mellitus patients compared with nondiabetic individuals ¹⁴ , ¹⁵ . In parallel, urinary ANGPTL4 levels were positively correlated with albuminuria levels in diabetic rats and type 2 diabetes mellitus patients ¹⁶ . Podocytes are the main source of ANGPTL4 production in the kidney ¹⁶ , by real‐time polymerase chain reaction, ANGPTL4 has been identified as the target gene of hypoxia‐inducible factor in tubular epithelial cells in vitro ¹⁷ . Therefore, we hypothesized that ANGPTL4 could represent glomeruli and tubular injury, and thus behave as a potential biomarker of DKD.

In the present study, we compared the associations of serum ANGPTL4 levels and other biomarkers with DKD index, including estimated glomerular filtration (eGFR), chronic kidney disease stages and the grades of urinary albuminuria‐to‐creatinine in 1,115 type 2 diabetes mellitus patients. We evaluated ANGPTL4 expression levels and distribution patterns in the kidneys of db/db mice at 8 or 18 weeks‐of‐age. Finally, we studied ANGPTL4 production in tubular epithelial cells exposed to different glucose concentrations.

MATERIALS AND METHODS

Human study

Study population

This was a cross‐sectional study. A total of 1,207 inpatients diagnosed with type 2 diabetes mellitus were recruited from Beijing Luhe Hospital, Beijing, China, from July 2015 to January 2017. The criteria of type 2 diabetes mellitus included fasting blood glucose ≥7.0 mmol/L and/or random blood glucose ≥11.0 mmol/L and/or 2 h blood glucose after an orally glucose tolerance test ≥11.0 mmol/L. Patients with type 1 diabetes mellitus, gestational diabetes mellitus or cancer were excluded.

Data cleaning procedure

After excluding missing data (n = 12) or repeated recruitment (n = 41) or outlier value (n = 39), a total of 1,115 participants enrolled in this study (Figure S1).

Clinical measurements

Systolic and diastolic blood pressure were measured using a standard mercury sphygmomanometer as described before ¹⁸ , ¹⁹ . Hypertension was diagnosed as the blood pressure of at least 140 mmHg systolic or 90 mmHg diastolic or using antihypertensive drugs. Information on height and weight, medical history, smoking and drinking habits, and intake of medications were collected. Waist‐to‐hip ratio was the ratio of waist circumference versus hip circumference.

Biochemical measurements

After overnight fasting, venous blood samples were drawn to measure the levels of white blood cell count, fasting blood glucose, total cholesterol, triglycerides (TG), high‐density lipoprotein cholesterol, creatinine, uric acid and HbA1c. Urinary samples were collected to measure the levels of creatinine and albumin. Low‐density lipoprotein cholesterol was calculated by the Friedewald equation.

Assessment of diabetic kidney disease

GFR was derived by the Chronic Kidney Disease Epidemiology Collaboration equation ²⁰ . The DKD risk management grades were based on the National technical guidelines of China for the prevention and treatment of diabetic kidney disease in primary care (2023).

DKD was categorized into stages 1, 2, 3, 4 and 5 based on eGFR values of ≥90, 60–89, 30–59, 15–29 and <15 mL/min/1.73 m² according to the National Kidney Foundation K/DOQI guideline. The normal‐to‐mild, moderate and severe albuminuria were defined when urinary albumin‐to‐creatinine was <30, 30–300 or >300 mg/g, respectively ²¹ .

Enzyme‐linked immunosorbent assay

Serum levels of ANGPTL2, ANGPTL3, ANGPTL4, vascular adhesion molecule–1 (VCAM‐1), intracellular adhesion molecule‐1 (ICAM‐1), high‐sensitivity C‐reactive protein (CRP), galectin‐3, resistin, vascular endothelial growth factor (VEGF), transforming growth factor β1 (TGF‐β1), neutrophil gelatinase‐associated lipocalin (NGAL) were measured following the instructions of enzyme‐linked immunosorbent assay kits. Intra‐assay coefficients of variation for ANGPTL2, ANGPTL3, ANGPTL4, VCAM‐1, ICAM‐1, VEGF, CRP, galectin‐3, resistin, TGF‐β1 and NGAL were 5.2, 5.9, 9.4, 5.2, 5.4, 4.6, 5.8, 6.5, 6.6, 5.7 and 6.2, respectively. Interassay coefficients of variation for ANGPTL2, ANGPTL3, ANGPTL4, VCAM‐1, ICAM‐1, VEGF, CRP, galectin‐3, resistin, TGF‐β1 and NGAL were 7.8, 6.8, 7.8, 7.3, 7.3, 7.3, 6.5, 7.3, 8.4, 8.5, 7.1 and 7.7, respectively.

Animal study

Male db/db mice aged 8 and 18 weeks (Model Animal Research Center of Nanjing University) were used in the present study. All mice were maintained at a constant temperature and humidity with free access to food and water.

Glucose tolerance test

Mice were fasted overnight, and then injected with 10% glucose at 10 μL/g bodyweight, and the blood glucose levels in venous blood were measured at 0, 15, 30, 60, 90 and 120 min after glucose administration.

Glomeruli isolation

After anesthesia, db/db mice were perfused with 8 × 10⁷ Dynabeads (Thermo Fisher Scientific, Waltham, MA, USA) diluted in 40 mL of phosphate‐buggered saline through the heart. The renal cortex was minced into 1‐mm³ pieces, and digested in collagenase (1 mg/mL collagenase A; R&D Systems, Minneapolis, MN, USA) at 37°C for 30 min. The digested tissue was filtered by using a 100‐μm cell strainer, and then the cell strainer was washed with Hanks’ balanced salt solution. The filtered cells were re‐passed through a new cell strainer, and then centrifuged at 200 g for 5 min. The supernatant was discarded, and the cell pellet was resuspended in Hanks’ balanced salt solution. Finally, glomeruli containing Dynabeads were gathered by a magnetic particle concentrator and washed with Hanks’ balanced salt solution for further experiments.

Cell culture

Renal tubular cells (HK‐2) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Cells were starved for 16 h, and then stimulated with glucose ranging from 0 to 20 mmol/L for 24 h. The whole cell lysate was detected by western blot.

Histology

The renal tissue was fixed in 4% paraformaldehyde, and embedded in paraffin. The paraffin‐embedded kidney tissues were sliced into 5‐μm sections, and then the sections were stained with hematoxylin–eosin or periodic acid‐Schiff. Images were captured using Leica Microsystems (Wetzlar, Germany) and analyzed by ImageJ 1.62 (National Institutes of Health, Bethesda, MD, USA).

Immunostaining

Kidney cryosections were incubated with anti‐ANGPTL4 (1:100; Santa Cruz Biotech, Santa Cruz, CA, USA) and anti‐nephrin (1:100; Abcam) overnight. The sections were probed with Alexa Fluor 488‐conjugated goat anti‐mouse immunoglobulin G (IgG; 1:200; Abcam) and Alexa Fluor 594‐conjugated goat anti‐rabbit IgG (1:200; Abcam, Cambridge, UK) for 1 h at room temperature. The sections were observed and pictures were taken by fluorescent microscope (Axioimager Z1 microscope; Carl Zeiss, Württemberg, Germany).

Western blot

The protein lysates were denatured at 95°C for 5 min and separated in 10% polyacrylamide sodium dodecyl sulfate gel, and transferred onto a polyvinylidene fluoride membrane. The membranes were blocked at room temperature for 1 h with 5% powdered milk in Tris‐buffered saline with Tween® 20 Detergent, and then incubated with primary antibodies overnight at 4°C followed by incubation with secondary antibodies (horseradish peroxidase conjugated goat anti‐mouse IgG or goat anti‐rabbit IgG; ZSGB‐BIO, Beijing, China) at room temperature for 1 h. Immunoreactive bands were visualized by ECL detection reagents (Applygen Technologies, Beijing, China) and analyzed with Image J 1.62.

Statistical analysis

Data are expressed as the mean ± standard deviation (SD) or median [interquartile range (IQR)] for continuous variables, and as the frequency (%) for categorical variables. P‐values were calculated using the Cochran–Armitage trend test and linear regression analyses for categorical and continuous variables across the three groups, respectively. The generalized additive models were used to analyze the nonlinear relationship between eGFR and ANGPTL4. Multivariable logistic regressions were used to investigate the association between ANGPTL4 and the DKD risk management grades. More logistic regression models were established to evaluate the association of ANGPTL4 and DKD according to different guidelines of DKD. The mediation analysis was carried out to detect whether TG mediated the association between ANGPTL4 and DKD.

During the analysis, the level of ANGPTL2, ANGPTL3, ANGPTL4, VCAM‐1, ICAM‐1, VEGF, CRP, galectin‐3, resistin, TGF‐β1 and NGAL were treated by z transformation. AP‐value <0.05 was considered statistically significant, and a P‐value was adjusted by the Benjamini–Hochberg method in the analysis of biomarker's multiple testing. All statistical analyses were carried out using R software (version 4.2.2, https://www.r‐project.org/).

RESULTS

General characteristics of type 2 diabetes mellitus patients

A cross‐sectional study was carried out with 1,115 type 2 diabetes mellitus patients. Among them, 527 (47.3%) were women. The mean age and mean duration of diabetes were 56.7 ± 13.4 and 9.16 ± 7.54 years, respectively. Among all participants, 590 (53.0%) had hypertension. Compared with the participants in the third or fourth quantile of ANGPTL‐4, those in the first or second quantile of ANGPTL‐4 were younger, with lower HbA1c, fasting blood glucose, TG, urinary albumin creatinine ratio (UACR) and uric acid, and higher eGFR (all P < 0.05). More demographic characteristics of participants are shown in Table 1.

Table 1.

Demographic characteristics and clinical characteristics of participants of this study

Group of serum ANGPTL 4	All	Quantile 1 n = 279 (9.96–<18.57)	Quantile 2 n = 279 (18.57–<23.56)	Quantile 3 n = 278 (23.56–<30.49)	Quantile 4 n = 279 (30.54–<100.97)	P trend
Demographic characteristic
Female, n (%)	527 (47.3)	145 (52.0)	121 (43.4)	131 (47.1)	130 (46.6)	0.355
Age (year), mean (SD)	56.7 (13.4)	55.0 (12.6)	56.0 (13.9)	57.3 (12.7)	58.5 (14.0)	0.001
Duration (year), mean (SD)	9.16 (7.54)	8.97 (6.86)	9.57 (7.40)	8.68 (7.36)	9.43 (8.46)	0.806
Smoke, n (%)	343 (31.4)	84 (30.5)	100 (36.8)	82 (29.8)	77 (28.4)	0.292
Drink, n (%)	336 (30.4)	100 (35.8)	86 (31.0)	83 (30.1)	67 (24.5)	0.005
Hypertension, n (%)	590 (53.0)	137 (49.1)	154 (55.2)	142 (51.1)	157 (56.5)	0.179
BMI (kg/m2), mean (SD)	26.2 (3.73)	26.2 (3.51)	26.2 (3.92)	25.9 (3.72)	26.5 (3.77)	0.472
WHR, mean (SD)	0.94 (0.08)	0.93 (0.08)	0.94 (0.09)	0.94 (0.07)	0.95 (0.08)	0.033
Laboratory Index of blood and urine
HbA1c (%), mean (SD)	10.1 (2.19)	9.82 (2.10)	10.0 (2.20)	10.2 (2.23)	10.4 (2.18)	0.001
FBG (mmol/L), mean (SD)	9.28 (4.14)	9.07 (3.92)	9.61 (4.42)	9.14 (3.88)	9.31 (4.31)	0.834
UA (μmol/L), mean (SD)	308 (92.4)	286 (84.8)	304 (87.8)	315 (96.5)	325 (96.0)	<0.001
SBP (mmHg), mean (SD)	132 (17.7)	130 (16.5)	131 (16.4)	132 (17.3)	135 (20.2)	0.002
TG (mmol/L), mean (SD)	2.15 (2.57)	1.93 (1.87)	1.97 (2.19)	2.16 (2.24)	2.52 (3.60)	0.004
TC (mmol/L), mean (SD)	4.68 (1.27)	4.57 (1.09)	4.59 (1.27)	4.73 (1.29)	4.81 (1.39)	0.011
HDL (mmol/L), mean (SD)	1.07 (0.30)	1.10 (0.33)	1.07 (0.28)	1.08 (0.29)	1.04 (0.29)	0.031
eGFR (mL/min/l.73 m2), mean (SD)	98.7 (18.1)	103 (15.9)	99.8 (17.7)	98.7 (17.9)	93.5 (19.7)	<0.001
UACR (mg/g), mean (SD)	64.41 (155.64)	36.42 (62.15)	51.65 (113.53)	70.41 (161.86)	99.18 (228.08)	<0.001
DKD risk management grades, n (%)
Level 1	589 (52.8)	172 (61.6)	155 (55.6)	147 (52.9)	115 (41.2)	<0.001
Level 2	440 (39.5)	101 (36.2)	110 (39.4)	105 (37.8)	124 (44.4)
Level 3	68 (6.10)	4 (1.43)	10 (3.58)	22 (7.91)	32 (11.5)
Level 4	18 (1.61)	2 (0.72)	4 (1.43)	4 (1.44)	8 (2.87)

The risk of diabetic kidney disease progression was defined by the clinical guideline for the prevention and treatment of diabetic kidney disease in China.

BMI, body mass index; DKD, diabetic kidney disease; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; HbA1c, glycosylated hemoglobin; HDL, high‐density lipoprotein; SBP, systolic blood pressure; TC, total cholesterol; TG, triglyceride; UA, uric acid; WHR, waist‐to‐hip rate.

Table S1 summarizes the clinical and biological characteristics of the study participants by the severity of albuminuria. Type 2 diabetes mellitus patients were categorized into normal‐to‐mild, moderate or severe albuminuria when the UACR was <30, 30–300 or >300 mg/mmol, respectively. Compared with those with mild and moderate albuminuria, patients with severe albuminuria featured a higher prevalence of hypertension, longer diabetic duration and lower eGFR. Among all biomarkers examined, serum levels of ANGPTL3 and ANGPTL4 were increased, and peaked in patients with severe albuminuria.

Association between ANGPTL4 and DKD

As shown in Figure 1, there was a negative nonlinear relationship between ANGPTL4 and eGFR in diabetes patients. The curve of Figure 1 was fitted by the GAMs. Furthermore, we carried out the logistic model adjusted for confounding variables, the models adjusted for age, sex, duration of diabetes, smoking, body mass index, waist‐to‐hip ratio, drinking, high‐density lipoprotein cholesterol, Low‐density lipoprotein cholesterol, TG, hypertension, HbA1c, fibrates, insulin and metformin. The results of the association between ANGPTL4 and DKD stage is shown in Table 2, which shows that per 1‐SD increase of serum ANGPTL4 levels, the odds ratio (OR) of having DKD (stage ≥3) was 1.40 (95% confidence interval [CI] 1.08–1.80).

Figure 1.

The nonlinear relationship of estimated glomerular filtration rate (eGFR) and angiopoietin‐like 4 (ANGPTL4) in sex group. The curves were fitted by generalized additive models method.

Table 2.

Association between serum angiopoietin‐like 4 levels and diabetic kidney disease and the progression risk of diabetic kidney disease

	OR (95% confidence interval)	P‐value
DKD stage (defined by eGFR)
Stage <3 vs stage ≥3	1.40 (1.08, 1.80)	0.009
DKD risk management grades
Level <3 vs Level ≥3	1.40 (1.15, 1.69)	0.001
Albuminuria grade
Normal‐to‐mild vs moderate–severe	1.20 (1.05, 1.37)	0.008
Normal‐moderate vs severe	1.24 (1.00, 1.52)	0.041

Diabetic kidney disease (DKD) stage <3 and ≥3 were determined by estimated glomerular filtration rate values ≥60 or <60 mL/min/1.73 m², respectively, according to the National Kidney Foundation K/DOQI guideline. Normal/to‐mild, moderate and severe albuminuria were defined as urinary albumin‐to creatinine ratio <30, 30–300, or >300 mg/g, respectively. DKD risk management grades was defined by National technical guidelines of China for the prevention and treatment of diabetic kidney disease in primary care (2023). Models was adjusted for age, sex, duration of diabetes, smoking, body mass index, waist‐to‐hip ratio, drinking, high‐density lipoprotein, low‐density lipoprotein, triglyceride, hypertension, glycated hemoglobin, fibrates, insulin and metformin.

We also estimated the association between ANGPTL 4 and the DKD risk management grades, that per 1‐SD increase of serum ANGPTL4 levels, the OR of raising for DKD risk management grades was 1.40 (95% CI 1.15–1.69).

In addition, we evaluated the relationship between serum ANGPTL4 levels and the grades of albuminuria. Overall, for a 1‐SD increase of serum ANGPTL4 levels, the OR of having moderate–severe albuminuria and severe albuminuria was 1.20 (95% CI 1.05–1.37) and 1.24 (95% CI 1.00–1.52), respectively.

To detect whether TG mediated the association between ANGPTL4 and DKD stage, the mediation analysis was carried out. As shown in Table S2, the estimated natural indirect effect was 0.000(95%CI：‐0.002,0.002), meaning that TG did not mediate the association between ANGPTL4 and DKD. Furthermore, TG did not mediate the association between ANGPTL4 and the DKD risk management grades and albuminuria (Table S2).

Serum ANGPTL4 levels as the potential indicator for DKD in type 2 diabetes mellitus patients

A series of biomarkers have been identified to present either glomerular or tubular dysfunction in type 2 diabetes mellitus patients. We also compared the association of the reported biomarkers with the DKD risk management grades, DKD stages or the grades of albuminuria. P‐values were adjusted by the Benjamini–Hochberg method in the analysis of biomarkers’ multiple testing.

Figure 2 shows the ORs of various biomarkers with DKD, which shows that the increase of ANGPTL4 and ANGPTL2 were the risk factors for DKD (stage ≥3). Furthermore, except for ANGPTL4, none of the biomarkers were significantly associated with the DKD risk management grades. In addition, except for ANGPTL4, none of the biomarkers were significantly associated with albuminuria.

Figure 2.

The association between diabetic kidney disease (DKD) and biomarkers of serum. UACR‐1, normal‐to‐mild albuminuria versus moderate–severe albuminuria; UACR‐2, normal‐moderate albuminuria versus severe albuminuria. Normal‐to‐mild, moderate and severe albuminuria were defined as urinary albumin‐to creatinine ratio <30, 30–300 or >300 mg/g, respectively. P‐value was adjusted by Benjamini–Hochberg method in the analysis of biomarker's multiple testing.

ANGPTL4 expression levels and patterns in kidneys of db/db mice

Finally, we thoroughly dissected ANGPTL4 expression levels and distribution patterns in the kidney of db/db mice at 8 and 18 weeks‐of‐age. At the age of 18 weeks, db/db mice showed higher blood glucose levels in glucose tolerance tests (Figure 3a), and pathological changes in kidneys with periodic acid‐Schiff staining (Figure 3b). When kidney homogenates were subjected to western blot, ANGPTL4 expression was increased in kidneys in db/db mice at 18 weeks‐of‐age when compared with those at 8 weeks‐of‐age (Figure 3c). When kidney sections were probed with anti‐mouse ANGPTL4 antibody, ANGPTL4 expression was increased inside and outside of glomeruli in 18‐week‐old db/db mice (Figure 3d).

Figure 3.

Increased renal angiopoietin‐like 4 (ANGPTL4) expression in diabetic db/db mice. (a) Oral glucose tolerance test results and (b) hematoxylin–eosin and periodic acid‐Schiff staining of renal tissues from the db/db mice of 8 and 18 week‐of‐age (×400), n = 10–15 per group. (c) Western blot and quantification of ANGPTL4 expression in kidney homogenates of db/db mice at 8 and 18 weeks‐of‐age, n = 10 per group. (d) ANGPTL4 expression in kidney sections of db/db mice at 8 and 18 weeks‐of‐age. Scale bars: 50 μm, n = 6 per group. (e) ANGPTL4 expression in ex vivo isolated glomeruli of 8 and 18 weeks‐of‐age db/db mice, n = 6 per group. (f) Renal tubular cells (HK‐2) were starved for 16 h, then stimulated by (0, 5, 10, 20 mmol/L) glucose for 24 h. Western blot of ANGPTL4 and glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) expression.

To further elucidate the ANGPTL4 expression pattern in diabetic kidneys, glomeruli were isolated from db/db mice. By western blot, ANGPTL4 was elevated in the glomeruli of db/db mice at 18 weeks‐of‐age compared with those at 8 weeks‐of‐age (Figure 3e). Likewise, glucose could stimulate ANGPTL4 expression in tubular epithelial cells in a dose‐dependent manner (Figure 3f). Taken together, ANGPTL4 was induced and increased in both glomeruli and renal tubular cells under the diabetic condition.

DISCUSSION

Previous small sample cross‐sectional studies have reported the diagnostic potential role of ANGPTL4 for DKD ²² , ²³ . However, there is a need for further comprehensive investigation into the relationship between ANGPTL4 levels and the severity of DKD, which will help generate novel insights for the clinical diagnosis and treatment of DKD. In the present cross‐sectional study containing 1,115 type 2 diabetes mellitus patients, we showed that serum ANGPTL4 level could be a potential biomarker in reflecting the severity of DKD, and the increase of serum ANGPTL4 predicted the higher DKD risk management grades. In addition, we also confirmed that ANGPTL4 could induce overexpression by hyperglycemia in glomeruli and renal tubules, which was detected both in vivo and in vitro.

The prevalence of DKD is increasing, and has become one of the major reasons for end‐stage kidney disease worldwide. The identification of risk factors and molecular targets that are critical in DKD progression is crucial in disease control. The nephron is the structural and functional unit of the kidney, which consists of glomeruli, renal sacculus and tubules. Through filtration in glomeruli, and reabsorption, secretion and excretion in tubules, the kidney participates in whole‐body homeostasis and blood pressure. On diabetic stimulus, damage of podocytes, glomerular endothelial cells and tubular cells in nephrons have been extensively reported ²⁴ , ²⁵ , ²⁶ . Meanwhile, a panel of biomarkers have been found in diabetes patients, which are localized in different nephron segments, and orchestrated in inflammation, fibrosis, cholesterol and lipid regulation, and ion imbalance ⁷ , ²⁷ . Despite that, it is unclear whether there is any common biomarker to represent both glomeruli and tubular injury in diabetes. Nevertheless, this issue could be interesting and clinically relevant.

Recently, ANGPTL4 has been found to be a key factor in nephrotic syndrome. It is known that podocytes produce two types of ANGPTL4 in hyposialylated and normosialylated forms ²⁸ . By immunogold electron microscopy, Clement et al. ²⁹ , ³⁰ showed that normosialylated ANGTL4 could protect glomerular endothelial cells from oxidative injury and preserve glomeruli structure. In rats with minimal change disease, podocytes produce more hyposialylated ANGPTL4 that could enhance endothelial cell injury and mobility, disrupt glomerular structure, and promote albuminuria. Feeding minimal change disease rats with sialic acid precursor could recapitulate sialylated ANGPTL4, improving glomeruli integrity and reducing albuminuria ²⁹ , ³⁰ . Similar to the minimal change disease model, in streptozotocin‐injected diabetic rats, ANGPTL4 expression was increased in the glomeruli, and urinary ANGPTL4 levels were positively associated with urinary albuminuria in rats and patients with DKD ¹⁶ . In addition, ANGPTL4 expression was increased in high‐glucose‐stimulated glomerular mesangial cells and podocytes ²⁸ , ³¹ . Apart from these reports, hypoxia‐inducible factor‐1α is a master in tubulointerstitial injury. By real‐time polymerase chain reaction, ANGPTL4 was shown to be a target gene of hypoxia‐inducible factor‐1α in tubular epithelial cells ¹⁷ . These brought us to initiate this study to address: (1) whether ANGPTL4 is a potential biomarker in predicting the severity of DKD; and (2) whether ANGPTL4 protein could be induced in both glomeruli and tubules by hyperglycemia.

In the present study, we detected the serum ANGPTL4 levels of 1,115 type 2 diabetes mellitus patients, and 1,115 type 2 diabetes mellitus patients were divided into four groups according to the level of serum ANGPTL4 from low to high. We found that UACR levels in ANGPTL4 high‐level group were significantly higher than that of ANGPTL4 low‐level group, and eGFR levels showed the opposite trend (Table 1). Furthermore, compared with those with mild and moderate albuminuria, patients with severe albuminuria featured higher serum ANGPTL4 levels (Table S1), and there was a negative nonlinear relationship between ANGPTL4 and eGFR in diabetes patients (Figure 1). Previous small sample cross‐sectional studies showed that urinary or circulating ANGPTL 4 levels were significantly higher in type 2 diabetes mellitus patients with DKD than in the type 2 diabetes mellitus patients without DKD, and negatively correlated with eGFR ²² , ²³ , which were consistent with the present findings. Additionally, we dissected the DKD index by DKD stage (defined by eGFR), DKD risk management grades and albuminuria grades, and analyzed serum ANGPTL4 with each index individually. After adjusting for confounding factors, serum ANGPTL4 remained highly correlated with DKD (Table 2). These results confirmed a positive relationship between ANGPTL4 levels and severity of DKD.

ANGPTL4 has been found to be a potent inhibitor of LPL that regulates LPL activity to suppress the clearance of TG from the circulation, and increased ANGPTL4 expression would elevate circulating TG ¹² . In the present study, we showed that TG levels in ANGPTL4 high‐level group were significantly higher than that of ANGPTL4 low‐level group in type 2 diabetes mellitus patients (Table 1). Increased TG is an important feature of dyslipidemia in diabetes patients, and is associated with the development of diabetic complications, including DKD, cardiovascular disease and so on. In the present study, the mediation analysis was carried out to diagnose whether TG was an intermediate variable in the relationship between ANGPTL4 and DKD. Although the participants in this study were restricted to hospitalized patients with higher levels of HbA1c, the results of the mediation analysis showed that TG did not mediate the association between ANGPTL4 and DKD in hospitalized patients (Table S2), which needs further analysis based on the cohort data in future.

Most previous studies were carried out with a limited number or focus on selected populations, such as individuals with eGFR ranging from 30 to 60 mL/min/1.73 m², having normoalbuminuria and/or microalbuminuria, or who underwent hemodialysis ³² , ³³ , ³⁴ , ³⁵ . Limited studies have sufficiently validated the strength of the association between multiple biomarkers and the DKD. Thus, the present study, which evaluated 1,115 type 2 diabetes mellitus patients, provides a thorough analysis and comparison of multiple panels of biomarkers in their relationship to DKD. We further showed that ANGPTL4 could be the strongest among all in its association with DKD (Figure 2). ANGPTL3 and ANGPTL4 belong to the same family, and share some common features in LPL inhibition and regulating angiogenesis. Previously, we reported that ANGPTL3, but not ANGPTL4, was positively associated with the stages of diabetic retinopathy in type 2 diabetes mellitus patients ¹⁸ . In the present study, although ANGPTL3 levels increased and peaked in patients with lowest eGFR, it was neither associated with eGFR nor with the grades of albuminuria, underlying different pathological mechanisms between these two proteins in diabetic microvascularpathy. The different interacting proteins of ANGPTL3 and ANGPTL4 shown by network STRING analysis might contribute to our findings in the study.

ANGPTL4 has been identified as a secreted glycoprotein produced by a variety of tissues, such as the kidney, liver and adipose tissue. To explore whether the levels of ANGPTL4 directly related to kidney damage in diabetes, we detected the expression level of ANGPTL4 in kidney tissue and cells, and found that ANGPTL4 could be upregulated in the kidney tissue of db/db mice, as well as diabetic glomeruli and high‐glucose‐treated tubular epithelial cells (Figure 3), indicating that hyperglycemia could induce ANGPTL4 expression in kidney tissue and cells. However, type 2 diabetes mellitus patients or mice often have complex abnormal metabolism issues, including hyperglycemia, hypertension and hypertriglyceridemia. Previous studies have shown that ANGPTL4 levels in both plasma and tissues were increased in the condition of hypertension or hypertriglyceridemia ³⁶ , ³⁷ , two critical risk factors of DKD. It was also found that elevation of plasma free fatty acids could promote ANGPTL4 circulation ³⁰ . Additionally, lipopolysaccharide increased ANGPTL4 expression in podocytes, but not in other glomerular cells, such as mesangial cells, glomerular endothelial cells and tubular cells ³⁸ . Furthermore, ANGPTL4 expression was significantly increased in high‐glucose‐stimulated glomerular mesangial cells ³¹ and podocytes ²⁸ . It was difficult to determine which was the most key factor that regulates ANGPTL4 expression and diabetic kidney from the present research, which requires further exploration. In addition, several previous studies also paid attention to the potential mechanism linking ANGPTL4 and DKD. Qin et al. ³¹ reported that knockdown of ANGPTL4 suppressed inflammatory response, and extracellular matrix proteins accumulation through inhibiting the nuclear factor‐κB signaling pathway in DKD. Increased expression of ANGPTL4 contributed to podocyte apoptosis and actin cytoskeleton rearrangement through the integrin/FAK pathway under high‐glucose conditions ²⁸ . In addition, overexpression of ANGPTL4 increased high‐glucose‐induced podocyte injury through the ROS/NLRP3 signaling pathway ³⁹ . It was also reported that hyposialylated ANGPTL4 played an important role in the formation of diabetic kidney injury ²⁹ . Furthermore, overexpression of ANGPTL4 led to increased lipid deposition and senescence in renal tubular epithelial cells ⁴⁰ . Further experimental study is required to clarify the specific relationship between ANGPTL4 and DKD.

The current study had some potential limitations. First, we cannot infer causality and highlight reverse causality as a possibility due to the cross‐sectional study design. A cohort study should be designed to explore the relationship between ANGPTL4 and DKD. Second, the sample size was relatively insufficient for subgroup analysis, so we could not explain the stability of results in different characteristics of participants. Last, all the study participants were inpatients, so the HbA1c was higher compared with other studies. Shawaf et al. ²² found that the elevation in ANGPTL4 correlated with clinical markers of diabetic nephropathy, such as albumin creatinine ratio, serum creatinine and eGFR in 122 Kuwaiti participants including 37 type 2 diabetes mellitus patients with normal kidney function and 49 type 2 diabetes mellitus patients with DN, with an HbA1c of 7–8%, showing that the association between ANGPTL4 and kidney damage also appeared in diabetes patients with low HbA1c. In future studies, we will collect community population to verify the results of the present study.

In conclusion, we found serum ANGPTL4 levels were significantly and negatively associated with eGFR in type 2 diabetes mellitus patients. Increased ANGPTL4 levels were related to the severity of DKD and progressive albuminuria in type 2 diabetes mellitus patients. ANGPTL4 could serve as a novel therapeutic candidate for the treatment of DKD.

DISCLOSURE

The authors declare no conflict of interest.

Approval of the research protocol: This study was performed in accordance with the principles of the Helsinki Declaration for investigation of human subjects approved by Institutional Review Boards of the Beijing Luhe Hospital and Capital Medical University.

Informed consent: All patients provided written informed consent.

Registry and the registration no. of the study/trial: N/A.

Animal studies: The animal experiments were performed in strict accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and approved by the Animal Experiments Committee of Capital Medical University.

Supporting information

Figure S1. Data cleaning procedure.

Table S1. Serum levels of biomarkers according to the severity of urinary albumin creatinine ratio.

Table S2. The results of mediation analyses.

DATA AVAILABILITY STATEMENT

The data used and/or analyzed during the current study are available from the corresponding author on reasonable request.