Gene expression profiling of KLF4 and KLF5 in visceral adipose tissue of obese women: insights into adipogenic and metabolic regulation

Abstract

Background

As global obesity rates rise, women are disproportionately affected, placing them at elevated risk for both metabolic dysfunction and cardiovascular complications. This study investigated the expression of Krüppel-like transcription factors KLF4 and KLF5 in visceral adipose tissue (VAT) from obese women to explore their potential roles in the pathogenesis of metabolic dysfunction.

Methods

In this case–control study, 46 women undergoing laparoscopic surgery were categorized into obese (n = 24) and non-obese control (n = 22) groups. VAT samples were analyzed for KLF4 and KLF5 mRNA expression using quantitative RT-PCR, normalized to GAPDH. Serum biomarkers related to glycemic status, lipid metabolism, adiposity indices, and cardiovascular risk were measured with automated clinical analyzers. Gene expression was correlated with metabolic parameters using Pearson and Spearman tests, with false discovery rate (FDR) correction for multiple comparisons.

Results

KLF4 and KLF5 expression levels did not differ significantly between groups. However, in the obese group, KLF4 expression showed positive correlations with HOMA-IR, HbA1c, fasting insulin, and glucose—indicating links to insulin resistance and glycemic regulation. In contrast, KLF5 expression was associated with lipid-related and cardiovascular parameters, including LDL-C, total cholesterol, CK-MB, and waist-to-hip ratio. In non-obese women, KLF4 and KLF5 exhibited coordinated expression, though their associations with metabolic traits were less pronounced. These findings indicate a BMI-dependent modulation of the KLF4/KLF5 regulatory axis, with obesity amplifying its links to dysglycemia and dyslipidemia.

Conclusion

The KLF4/KLF5 transcriptional axis exhibits obesity-related regulatory shifts that may contribute to cardiometabolic dysfunction. These patterns highlight the potential utility of KLF4 and KLF5 as biomarkers for risk stratification, informing the development of tissue-specific therapeutic strategies aimed at improving metabolic outcomes in obese women, while ensuring that translational approaches remain safe, equitable, and targeted.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12944-025-02734-z.

Introduction

Being overweight and obesity are increasingly problematic for public health on a global scale and are still on the rise in many populations, according to several published sources [1, 2]. Individuals with a body mass index (BMI) above 25 kg/m² are classified as overweight, while those with a BMI above 30 kg/m² are classified as obese [3]. These conditions stem from a sustained energy imbalance, typically driven by excessive caloric intake and reduced physical activity, resulting in weight gain that exceeds the physiologic threshold for a given height. National Surveys and pooled data suggest that between 60% and 70% of Iranian adults fall within the overweight or obese range [4–10].

Obesity is a chronic, noncommunicable, multisystem disorder influenced by metabolic, behavioral, environmental, and genetic factors. Anthropometric indices such as BMI, waist circumference (WC), waist-to-hip ratio (WHR), and body fat percentage serve as primary tools for evaluating obesity status and related health risks [11, 12]. Obesity is closely linked to chronic low-grade inflammation, which disrupts normal metabolic pathways and contributes to insulin resistance, type 2 diabetes, and cardiovascular disease [13]. Accordingly, routine clinical assessments in individuals with obesity typically include lipid panels, fasting blood glucose levels, blood pressure measurements, and other metabolic markers. Elevated levels of C-reactive protein (CRP) and additional proinflammatory biomarkers are frequently observed, indicating persistent systemic inflammation. Mental health screening may also be warranted, given the association between obesity and increased susceptibility to depression, anxiety, and other psychiatric conditions [14].

Although not included among traditional metabolic biomarkers such as the homeostatic model assessment of insulin resistance (HOMA-IR), lipid levels, or inflammatory indices, Krüppel-like factors—particularly KLF4 and KLF5—have emerged as critical regulators of adipogenesis [15, 16] and obesity-related pathophysiology [17]. These transcription factors modulate the expression of genes involved in a wide array of cellular processes, including stemness, proliferation, apoptosis, autophagy, and cell migration. Both KLF4 and KLF5 contribute to insulin signaling and glucose metabolism, and their dysregulation has been implicated in various metabolic disorders [18].

The mammalian Krüppel-like factor family consists of 18 members (KLF1–KLF18) [19]. Among these, specific members act as pro-adipogenic factors—particularly KLF4 and KLF5—while others, such as KLF2, KLF3, and KLF7, exert inhibitory effects. Based on structural and transcriptional characteristics, KLF4 and KLF5 are classified within group 2 KLFs.

Globally, obesity disproportionately affects women, a disparity attributed in part to sex-specific differences in fat metabolism and the hormonal regulation of key metabolic genes. Estrogen, for instance, has been shown to upregulate KLF4 expression in selected tissues, potentially contributing to differential gene expression patterns between sexes [20–23].

Despite accumulating evidence from in vitro and animal studies, research on KLF4 and KLF5 in human adipose tissue has been limited, particularly in relation to metabolic traits. To address this gap, the present study quantified the transcriptional levels of KLF4 and KLF5 in visceral adipose tissue (VAT) from a homogenous cohort of premenopausal women. The study also investigated associations between these transcription factors and key anthropometric indices (body weight, height, BMI, WC, and WHR), metabolic parameters [including HOMA-IR, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and total cholesterol (TC)], and inflammatory markers such as CRP.

Methods

Study design and population

The study population has been described in detail elsewhere [24]. Sample size was determined based on prior literature [25–28] and calculated using G*Power (v3.1) for a two-tailed independent-samples t test. A priori power analysis indicated that a minimum of 21 participants per group would be required to detect a medium-to-large effect size (Cohen’s d = 0.7) with 80% power and α = 0.05. The calculation was based on the following formula:

A case–control design was implemented, including 46 premenopausal women divided into two groups. The case group (n = 24) included women with obesity referred to Imam Khomeini Hospital for bariatric surgery, all of whom underwent one of the following procedures: sleeve gastrectomy (SG), Roux-en-Y gastric bypass (RYGB), or laparoscopic adjustable gastric bypass (LAGB). The control group (n = 22) consisted of individuals with normal BMI undergoing elective non-metabolic surgical procedures. These included cholecystectomy (15 individuals; 68.2%), appendectomy (2 individuals; 9%), splenectomy (1 individual; 4.5%), and hernia repair (4 individuals; 18.2%).

Participants in both groups were selected to minimize confounding effects from surgery-related stress responses, inflammation, or underlying pathological conditions. Efforts were made to match surgical approaches across groups where possible.

Exclusion criteria were established to ensure baseline comparability between groups. Individuals with known endocrine disorders, irregular menstrual cycles, or use of medications affecting inflammation, lipid metabolism, or insulin sensitivity (e.g., corticosteroids, nonsteroidal anti-inflammatory drugs [NSAIDs], or lipid-lowering agents) were excluded. All participants had abstained from medication use for at least six months before enrollment, except for non-hormonal supplements and occasional analgesics.

Eligible participants were between 18 and 45 years of age and selected to ensure consistency in hormonal status. This age range aligns with the standard reproductive age, as natural menopause typically occurs between 49 and 51 years in the studied population [29]. None of the participants used hormonal contraception or hormone replacement therapy. Menstrual regularity was employed as a surrogate marker for hormonal stability due to the absence of direct hormonal profiling. Additional exclusion criteria included menopause, pregnancy, lactation, and any surgical procedure within the six months preceding enrollment.

Data on physical activity and dietary intake were not collected due to the homogeneity of the study sample and the narrow inclusion parameters; this limitation is acknowledged.

Sample collection

As previously described [24], 10 mL of peripheral venous blood was collected from each participant into sterile BD Vacutainer tubes following an overnight fast of 8–10 h before surgery. After processing, serum was separated and stored at − 80 °C to preserve biochemical integrity. Additional whole blood samples were drawn into ethylenediaminetetraacetic acid (EDTA)-containing tubes for the analysis of glycated hemoglobin (HbA1c).

All Subjects Submitted written informed consent prior to inclusion in the study. During laparoscopic surgery, approximately 1cm³ (0.4 g) of VAT was excised. Each sample was immediately transferred into individual pre-cooled Falcon tubes and subsequently preserved in liquid nitrogen and RNAlater solution until further molecular analysis was performed.

Estimation of biochemical parameters

Biochemical parameters were assessed using standardized laboratory protocols. Fasting blood sugar (FBS), uric acid, urea, creatinine, LDL-C, HDL-C, TC, triglycerides (TG), alkaline phosphatase (ALP), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), creatine kinase (CK), creatine kinase–myocardial band (CK-MB), amylase, lipase, lactate dehydrogenase (LDH), and alanine aminotransferase (ALT) were measured using commercial assay kits (Pars Azmoon, Tehran, Iran) and an automated chemistry analyzer (Hitachi, Japan). High-sensitivity C-reactive protein (hs-CRP) was quantified via immunoturbidimetry on the Roche Integra platform (Roche Diagnostics, Mannheim, Germany). Fasting plasma insulin levels were determined using the Cobas 6000 E601 analyzer with electrochemiluminescence (ECL) technology.

Systolic and diastolic blood pressure were measured twice on the right arm after a 10-minute seated rest using a calibrated mercury sphygmomanometer to ensure consistency and accuracy [30, 31].

Several indices were calculated as previously described [30]. BMI was derived by applying the standard equation, where weight in kilograms is divided by height in meters squared. The conicity index (CI) was determined using its established mathematical expression as follows:

This method, originally proposed by Valdez and later refined by Kapoor and Faubion [32, 33], evaluates central adiposity. The body adiposity index (BAI), calculated as follows—BAI = hip circumference (HC in cm) ÷ [(height in m) ¹.⁵ − 18]—has demonstrated reasonable correlation with body fat percentage (% fat) [13]. To assess abdominal obesity, the abdominal volume index (AVI) was computed using the equation:

This formula estimates abdominal volume extending from the symphysis pubis to the xiphoid process, integrating both intra-abdominal and subcutaneous fat compartments.

The lipid accumulation product (LAP) was calculated based on sex-specific formulas. For women:

The visceral adiposity index (VAI) for females was computed using:

[34].

Weight-adjusted waist index (WWI) was derived by dividing WC by the square root of body weight (kg). Insulin resistance was estimated using the homeostatic model assessment for insulin resistance (HOMA-IR):

Based on the thresholds proposed by Jiménez-Maldonado and García-Suárez [35], HOMA-IR values between 0.5 and 1.4 are considered normal, values ≥ 1.9 suggest early insulin resistance (IR), and values ≥ 2.9 indicate established IR.

β-cell function was evaluated using the HOMA-β formula:

[35].

Analysis of mRNA Levels of KLF4 and KLF5 in VAT

As previously described [36], frozen VAT specimens were homogenized in liquid nitrogen, and total RNA was extracted using the RNeasy Lipid Tissue Mini Kit (Qiagen GmbH, Germany). RNA purity and integrity were assessed by measuring the absorbance ratio at 260/280 nm and by performing agarose gel electrophoresis.

Complementary DNA (cDNA) was synthesized from 1000 ng of DNase-treated RNA using the PrimeScript 1 st Strand cDNA Synthesis Kit (Takara, Japan). The resulting cDNA was used to quantify the expression of KLF4 and KLF5 by real-time polymerase chain reaction (RT-PCR), with GAPDH serving as the internal control or reference gene. Pishgam Biotech supplied primers.

Primer sequences are presented in Table 1. Primer design was performed using the Basic Local Alignment Search Tool (BLAST) suite, provided by the National Center for Biotechnology Information (NCBI). The characteristics of the primers—such as melting temperature, GC content, and potential secondary structures—were evaluated using OligoAnalyzer. Primer specificity was verified via BLAST alignment. This multistep approach ensured that each primer set met the analytical requirements for downstream quantification.

Table 1.

Primer information

Gene name		Oligo Name Sequence 5’−3’
KLF4	Forward	CGG ACA TCA ACG ACG TGA G
	Reverse	GAC GCC TTC AGC ACG AAC T
KLF5	Forward	CCT GGT CCA GAC AAG ATG TGA
	Reverse	GAA CTG GTC TAC GAC TGA GGC
GAPDH	Forward	ACA ACT TTG GTA TCG TGG AA GG
	Reverse	GCC ATC ACG CCA CAG TTT C

Primer efficiencies were determined using LinRegPCR software: KLF5 (E = 1.87), KLF4 (E = 1.83), and GAPDH (E = 1.62). Relative gene expression was calculated using the Pfaffl method, adjusting for variations in amplification efficiency between primer sets [37].

Statistical analysis

Missing data were addressed using the multiple imputation (regression method) function in the Statistical Package for the Social Sciences (SPSS, version 26.0.0) [38]. The distribution of variables within the obese and control groups was evaluated using the Shapiro–Wilk test. Variables that followed a normal distribution included age, height, waist-to-height ratio (WHtR), BAI, WWI, FBS, urea, creatinine, uric acid, HDL-C, LDL-C, TC, CI, ALP, total iron-binding capacity (TIBC), LDH activity, amylase activity, and fasting insulin. Several other variables exhibited non-normal distributions in at least one of the study groups.

Non-normally distributed variables were transformed using a two-step normalization process. This method involved ranking the data (fraction rank) and applying statistical normalization to generate standardized values labeled as “Normal_Variable” [39].

Descriptive statistics, including means and standard deviations, were calculated separately for the obese and control groups using SPSS. Group comparisons were conducted using the Chi-square test for categorical variables and the Student t-test for continuous variables. Statistical significance was defined as p < 0.05. Analysis of covariance (ANCOVA) was used to adjust for potential confounding factors, including age, smoking status, and anti-Müllerian hormone (AMH).

Correlation analyses were performed using Python within a Jupyter Notebook environment. Both Pearson and Spearman correlation coefficients were calculated to assess associations among continuous variables. Analyses were conducted on the entire dataset and stratified by group (obese vs. control), with and without adjustment for identified confounders. Given the limited sample size, the term “association” was preferred over “correlation” when describing the relationships in the results.

To adjust for multiple comparisons, the Benjamini–Hochberg (BH) procedure was applied to all 60 statistical tests. This method controls the false discovery rate (FDR) by converting p-values to q-values. The q-values were computed by ranking the unadjusted p-values in ascending order and using the formula AdjPi = pi × m/i, where m is the total number of tests, i is the p-value’s rank, and Pi is the corresponding unadjusted p-value. The q-values were verified in Python. A threshold of q ≤ 0.10 was used to determine statistical significance. Under this criterion, up to 10% of reported significant findings may represent false positives. Given the exploratory nature of this study, this threshold was selected to optimize sensitivity while maintaining control over Type I error.

Results

Descriptive statistics for anthropometric and biochemical parameters are summarized in Table 2. The study population comprised two groups: 24 individuals classified as obese and 22 non-obese controls. Significant between-group differences were observed in several anthropometric indices. However, no significant differences were found in height, age, WHR, VAI, WWI, or AMH (P > 0.05). In contrast, BMI, body weight, WC, CI, HC, WHtR, BAI, AVI, and LAP differed significantly between the groups (P < 0.05).

Table 2.

Anthropometric, clinical, and metabolic attributes of participants

Variable	Control (n = 22)	Obese (n = 24)	P-Val
Age, years	37.68 (9.06)	34.37 (6.61)	0.169
BMI, kg/m2	23.31 (1.39)	42.58 (9.67)	0.000
SBP, mmHg	120.0 (10.0)	120.0 (16.25)	0.234
DBP, mmHg	80.0 (10.0)	80.0 (20.0)	0.382
TG, mg/dL	91.55 (56.17)	105.95 (69.6)	0.839
LDL-C, mg/dL	88.69 (28.93)	112.05 (20.20)	0.003
HDL-C, mg/dL	44.02 (7.33)	44.79 (7.15)	0.721
VLDL-C, mg/dL	20.0 (15.25)	20.5 (12.0)	0.196
LDL-C/HDL-C, -	1.76 (1.38)	2.40 (0.66)	0.013
TC, mg/dL	147.40 (37.60)	178.75 (25.99)	0.002
AST, U/L	16.7 (7.82)	20.65 (7.25)	0.044
ALT, U/L	12.55 (8.37)	20.3 (13.35)	0.032
ALP, U/L	67.29 (26.60)	74.24 (16.61)	0.301
GGT, U/L	21.8 (40.97)	22.7 (16.58)	0.878
Amylase, U/L	43.83 (14.58)	41.431 (9.66)	0.518
Lipase, U/L	21.7 (9.65)	22.77 (7.71)	0.158
CK, U/L	71.15 (54.25)	90.49 (66.35)	0.005
CK-MB, U/L	14.11 (2.61)	14.94 (6.11)	0.664
LDH, U/L	195.2 (66.03)	213.73 (49.75)	0.027
FBS, mg/dL	85.209 (7.29)	88.99 (8.71)	0.116
Insulin, µU/mL	8.02 (3.57)	19.65 (4.68)	0.000
HOMA-IR, -	1.69 (0.86)	3.92 (1.69)	0.000
HOMA-beta, %	115.04 (72.35)	293.73 (181.62)	0.000
HbA1c, %	5.2 (0.42)	5.5 (0.45)	0.002
Uric acid, mg/dL	4.00 (0.83)	5.55 (1.10)	0.000
Urea, mg/dL	22.6 (7.54)	26.08 (5.63)	0.086
Creatinine, mg/dL	0.58 (0.16)	0.73 (0.12)	0.001
VAI, -	1.76 (0.92)	1.84 (1.66)	0.146
BAI, -	27.81 (3.21)	43.35 (7.12)	0.000
WWI, -	10.66 (0.70)	10.97 (0.95)	0.213
AVI, -	14.54 (1.30)	26.22 (3.19)	0.000
LAP, -	24.14 (18.65)	73.94 (56.35)	0.000
CI, -	0.89 (0.09)	1.09 (0.11)	0.000
WC, cm	85.0 (4.0)	114.0 (6.75)	0.000
HC, cm	95.0 (6.75)	128.0 (12.5)	0.000
WHtR, -	0.51 (0.04)	0.71 (0.05)	0.000
WHR, -	0.89 (0.06)	0.93 (0.07)	0.074
AMH, ng/ml	1.752 ± 1.584	3.11 ± 2.514	0.060
Hs-CRP, mg/L	1.8 (1.47)	5.69 (7.74)	0.000
TIBC, µg/dL	308.31 (59.02)	363.13 (55.52)	0.003
Iron, µg/dL	64.92 (44.72)	72.9 (33.0)	0.213
Albumin, g/dL	3.70 (0.54)	4.32 (0.30)	0.000
TP, g/dL	5.89 (0.91)	6.96 (0.46)	0.000

Data are presented as mean (standard deviation) or median (interquartile range). Statistically significant differences between the control and obese groups are indicated in bold

Systolic blood pressure (SBP) and diastolic blood pressure (DBP) did not differ significantly between groups. FBS was higher in the obese cohort, though this difference was not statistically significant (P > 0.05). Serum concentrations of creatinine and uric acid were significantly elevated in the obese group (P < 0.001).

Liver enzyme activity was also higher among obese participants, although no significant differences were observed in ALP or GGT levels. Total cholesterol (TC) and LDL-C were significantly elevated in the obese group, whereas HDL-C, TG, and very low-density lipoprotein cholesterol (VLDL-C) showed no statistically significant differences. High-sensitivity C-reactive protein (hs-CRP) was significantly elevated in the obese cohort.

Markers of insulin resistance and β-cell function also differed significantly between groups. Obese participants exhibited higher mean values for fasting insulin, HbA1c, HOMA-IR, and HOMA-β (P < 0.001; Table 2).

Circulating levels of CK and TIBC were markedly higher in the obese group (P < 0.05). Although serum iron concentrations tended to be elevated in this group, the difference was not statistically significant (P > 0.05). CK-MB levels showed no meaningful variation between groups. In contrast, both total protein (TP) and serum albumin concentrations were significantly increased in obese individuals.

As previously noted, SBP and DBP did not differ significantly between groups. Antihypertensive medication use was slightly more frequent in the obese cohort (P = 0.086). The prevalence of hyperlipidemia was comparable between the groups (P > 0.05). Although the obese group demonstrated a higher incidence of diabetes, this difference did not reach statistical significance (P > 0.05). Similarly, the use of metformin was not significantly different between the two cohorts (P > 0.05). Levothyroxine was used exclusively by participants in the obese group (P > 0.05), while neither group reported statin use. A statistically significant difference in smoking status was observed, with 8 individuals in the obese group (33.3%) reporting active smoking (Supplementary Table S1).

Expression levels of KLF4 and KLF5 were modestly higher in the obese group, and the KLF4:KLF5 expression ratio was also elevated. The observed differences were not significant (P > 0.05) (Fig. 1).

Fig. 1.

Relative mRNA expression of KLF4 and KLF5.

The mean expression levels of KLF4 were higher in the Obese group (2.460 ± 2.945) compared to Controls (1.365 ± 2.701), although this difference was not statistically (P = 0.201). Similarly, the expression of KLF5 was modestly elevated in the Obese group (5.253 ± 7.520) versus Controls (2.513 ± 7.520), without reaching statistical significance (P = 0.229). The KLF4:KLF5 expression ratio was also higher in the Obese group (0.340 ± 0.557) than in Controls (0.012 ± 0.499), with no statistically significant difference (P = 0.128)

Correlation analyses

In the full dataset (i.e., before stratifying participants by group), KLF4 expression demonstrated positive associations with several variables, including LDL-C (FDR < 0.1), TC, amylase, CK, CI, WWI, and the LDL-C: HDL-C ratio (Supplementary Table S2). Negative associations were observed between KLF4 and ALP, fasting insulin, and HbA1c. Furthermore, normalized KLF4 (Normal_KLF4) exhibited significant inverse associations with HOMA-IR and HOMA-beta.

KLF5 expression showed a statistically significant negative association with CK-MB (FDR < 0.1) and a non-significant inverse relationship with SBP (FDR > 0.1). Positive associations were observed with LDL-C, TIBC, TC, and ALT.

Importantly, KLF4 and KLF5 displayed a strong positive association with each other. After controlling for the false discovery rate, associations between KLF4 and KLF5, the KLF4/KLF5 ratio, and LDL-C remained statistically significant. In addition, associations between KLF5 and CK-MB (FDR = 0.019) as well as LDL-C (FDR = 0.065) retained significance after FDR correction.

After stratifying the dataset, the Obese group exhibited a significant inverse association between KLF4 expression and BMI (r = − 0.581, FDR = 0.046), and between KLF4 and AST (r = − 0.550, FDR = 0.068), alongside a positive association with amylase (r = 0.602, FDR = 0.036) (Table 3). The KLF4–AST relationship lost significance after adjusting for age, and the KLF4–BMI association became non-significant after controlling for both smoking and AMH. However, the association between Normal_KLF4 and BMI remained statistically significant even after simultaneous adjustment for age, smoking, and Normal_AMH (r = 0.0061, FDR = 0.0975).

Table 3.

Notable associations between KLF4 and KLF5 and various obesity indices within the control (non-obese) group

KLF4; Control (non-obese)
Original_KLF4	Spearman			Normal_KLF4	Pearson
Variable	r	P-Val	Adj.P	Variable	r	P-Val	Adj.P
KLF5	0.718	0.000	0.005	Normal_KLF5	0.652	0.001	0.021
KLF4/KLF5	0.682	0.000	0.007	Normal_KLF4/KLF5	0.640	0.001	0.021
Amylase	0.602	0.003	0.036	Normal_BMI •	−0.562	0.007	0.078
BMI •	−0.581	0.005	0.046	Normal_BMI †‡₸	−0.587	0.004	0.082
BMI ‡ǂ	−0.165	0.462	0.816	Normal_Iron •	−0.376	0.085	0.480
AST •	−0.550	0.008	0.068	Normal_Iron †	−0.554	0.007	0.089
AST †	−0.287	0.195	0.638	Normal_HOMAbeta	−0.548	0.008	0.125
LDH	−0.466	0.029	0.217	Amylase	0.536	0.010	0.102
HOMAbeta	−0.451	0.035	0.351	Insulin	−0.501	0.018	0.176
				LDH	−0.457	0.032	0.253
				Normal_AST	−0.454	0.034	0.253

KLF5; Control (non-obese)
Original_KLF5	Spearman			Normal_KLF5	Pearson
Variable	r	P-Val	Adj.P	Variable	r	P-Val	Adj.P
KLF4	0.718	0.000	0.005	Normal_KLF4	0.652	0.001	0.030
KLF4/KLF5	−0.639	0.001	0.027	Normal_KLF4/KLF5	−0.552	0.008	0.155
CK-MB	−0.596	0.003	0.068	CK-MB	−0.531	0.011	0.219
BAI	0.505	0.017	0.248	TIBC	0.493	0.020	0.296
TIBC	0.480	0.024	0.355	Amylase	0.456	0.033	0.409
Weight •	−0.478	0.024	0.262	ALP	0.455	0.033	0.401
Weight ‡	−0.478	0.024	0.258	normalWeight •	−0.448	0.037	0.409
Amylase	0.473	0.026	0.262	normalWeight ‡	−0.448	0.037	0.409
sUA/sCr •	0.462	0.031	0.367	normalBMI †‡₸	−0.435	0.043	0.397
sUA/sCr ǂ	0.389	0.074	0.706	normalBMI •	−0.411	0.057	0.409
BMI •	−0.429	0.046	0.312	LDL-C	0.411	0.057	0.409
BMI ‡ǂ	−0.159	0.481	0.904
Iron	−0.428	0.047	0.312
ALP	0.424	0.049	0.494

KLF4 and KLF5 refer to Krüppel-like transcription factors 4 and 5, respectively

Original_KLF4 and Original_KLF5 denote raw, non-normalized expression values

Normal_KLF4 and Normal_KLF5 represent values normalized using a two-step transformation protocol [39]

Associations marked with • were unadjusted; †, ‡, ǂ, and ₸ indicate results adjusted for age, smoking status, AMH, and Normal_AMH, respectively

Notably, the association between Normal_KLF4 and Normal_Iron became statistically significant following adjustment for age

After FDR correction, associations between KLF4 and LDH, HOMA-beta, and between Normal_KLF4 and Normal_HOMA-beta, insulin, amylase, LDH, and Normal_AST in the Control group did not remain statistically significant.

In the Control cohort, none of the correlations between KLF5 and BAI, TIBC, body weight, amylase, serum uric acid-to-creatinine ratio (sUA/sCr), BMI, iron, or ALP reached significance after FDR adjustment (FDR > 0.1). Only the association between KLF5 and CK-MB remained statistically significant after controlling for multiple comparisons.

Within the Obese group, no statistically significant association was identified between KLF4 and BMI, as evidenced by the absence of BMI-related correlations in Table 4. A comprehensive list of BMI-related associations is provided in Supplementary Table S3. As presented in Table 4, KLF4 expression exhibited strong inverse associations with HOMA-IR (r = − 0.622, FDR = 0.035), fasting insulin (r = − 0.582, FDR = 0.047), and HbA1c (r = − 0.559, FDR = 0.054). A weaker negative association with the LDL-C: HDL-C ratio was also noted (r = − 0.434, FDR > 0.1). In contrast, KLF4 showed modest positive correlations with LDL-C, CK, and TC (FDR > 0.1), and a more pronounced positive association with HDL-C (r = 0.523, FDR = 0.088).

Table 4.

Notable associations between KLF4 and KLF5 with obesity indices in the obese cohort

KLF4; Obese
Original_KLF4	Spearman			Normal_KLF4	Pearson
Variable	r	P-Val	Adj.P	Variable	r	P-Val	Adj.P
KLF5	0.802	0.000	0.000	Normal_KLF5	0.767	0.000	0.000
HOMAIR	−0.622	0.001	0.035	Insulin	−0.586	0.003	0.077
Insulin	−0.582	0.003	0.047	Normal_KLF4/KLF5	0.563	0.004	0.077
KLF4/KLF5	0.577	0.003	0.047	HDL-C	0.552	0.005	0.077
HbA1c	−0.559	0.005	0.054	Normal_HOMAIR	−0.528	0.008	0.092
LDL-C	0.553	0.005	0.101	Normal_HbA1c	−0.520	0.009	0.092
HDL-C	0.523	0.009	0.088	Normal_CK	0.413	0.045	0.669
CK	0.521	0.009	0.136
TC	0.496	0.014	0.165
LDL-C/HDL-C	−0.434	0.034	0.293

KLF5; Obese
Original_KLF5	Spearman			Normal_KLF5	Pearson
Variable	r	P-Val	Adj.P	Variable	r	P-Val	Adj.P
KLF4	0.802	0.000	0.000	Normal_KLF4	0.767	0.000	0.000
KLF4/KLF5	−0.513	0.010	0.308	Normal_KLF4/KLF5	−0.533	0.007	0.182
Chloride •	0.488	0.016	0.312	FBS	−0.521	0.009	0.182
Chloride †	0.496	0.014	0.413	CK-MB	−0.428	0.037	0.554
HOMAIR	−0.432	0.035	0.362	Normal_BMI †‡₸	−0.417	0.042	0.364
CK-MB	−0.418	0.042	0.503	Normal_BMI •	−0.297	0.159	0.697
LDL-C	0.417	0.043	0.362	Normal_SBP •	−0.400	0.053	0.697
TC	0.415	0.044	0.362	Normal_SBP †	−0.384	0.064	0.770
WHR	0.414	0.044	0.362
HC	−0.407	0.048	0.362

KLF4 and KLF5 refer to Krüppel-like transcription factors 4 and 5, respectively

Original_KLF4 and Original_KLF5 denote unnormalized expression values, whereas Normal_KLF4 and Normal_KLF5 represent values normalized through a two-step transformation process [39]

Associations marked by • are unadjusted; †, ‡, ǂ, and ₸ indicate adjustment for age, smoking status, AMH, and Normal_AMH, respectively

For KLF5, several preliminary associations with variables including chloride, HOMA-IR, CK-MB, LDL-C, TC, WHR, and HC did not retain statistical significance following adjustment for multiple comparisons (FDR > 0.1). Normal_KLF5 was weakly associated with FBS and Normal_BMI, though these relationships did not meet the significance threshold (FDR > 0.1) (Table 4).

Consistent with the findings in the Control group, strong positive associations were observed between KLF4 and KLF5 in the Obese cohort. Additionally, both factors demonstrated the expected correlations with the KLF4/KLF5 ratio—KLF4 showing positive associations and KLF5 showing inverse associations (Tables 3 and 4).

Further insights regarding the associations between metabolic regulators and HOMA-IR, HOMA-beta, and HbA1c are available in Supplementary Tables S4-S6.

Discussion

This study evaluated KLF4 and KLF5 gene expression in visceral adipose tissue from obese and normal-weight women and explored their associations with diverse adiposity indices.

Elevated KLF4 and KLF5 gene expressions in obesity

Expression levels of KLF4 and KLF5 were marginally increased in the Obese group, though without statistical significance (Fig. 1). In a secondary analysis using GEO dataset GSE88837 and the R programming environment, comparisons among Black, White, and Hispanic women with high versus low BMI (n = 15 per group) demonstrated a 1.63-fold upregulation in KLF4 expression (P = 0.00011; adjusted P = 0.107) and a − 0.83-fold non-significant downregulation in KLF5 expression (P = 0.228; adjusted P = 0.803) [40]. Although the unadjusted P value for KLF4 indicated high statistical significance, this significance diminished after FDR correction. Both the present study and GEO-based findings suggest a trend toward elevated KLF4 expression in individuals with obesity, with no meaningful differential expression of KLF5.

Interpretation of these results requires attention to the basal expression levels of KLF4. If baseline KLF4 expression is inherently low, a 1.63-fold increase may still exert substantial biological effects, despite lacking statistical significance. The disconnect between fold change and adjusted significance likely reflects limited statistical power or interindividual variability. Additionally, inconsistencies across studies may stem from differences in sample size, data acquisition methods, and population-level variables, including genetics, diet, and environmental exposures.

The generally low expression of KLF4 and KLF5, even in metabolically altered states, may explain the high variability and frequent absence of statistical significance. Alternatively, stringent post-transcriptional regulation, such as degradation of excess mRNA via microRNAs or RNA-binding proteins, may contribute to the observed homeostasis in expression levels despite the presence of metabolic stressors such as obesity [41].

Correlation analyses

Negative associations between BMI and KLF4 in control but not obese

Tables 3 and 4 detail the associations of KLF4 and KLF5 with metabolic and anthropometric variables in the Control (non-obese) and Obese groups, respectively.

Normal_BMI was inversely correlated with Normal_KLF4 in the Control cohort (r = − 0.587, FDR = 0.082), and this association remained statistically significant following adjustment for age, smoking status, and Normal_AMH. A marginally significant inverse association was also observed between Normal_BMI and KLF5 (r = − 0.435, P = 0.043), suggesting reduced expression of KLF4 and KLF5 in metabolically healthy individuals with higher BMI values.

Krüppel-like factors such as KLF4 and KLF5 are integral to adipogenesis. KLF4 has been shown to regulate key pathways that mediate adipocyte differentiation, including synergistic activation of peroxisome proliferator-activated receptor gamma (PPARγ) and the CCAAT/enhancer-binding protein (C/EBP) family, suppression of Wnt signaling, and modulation of lipid metabolism through the sterol regulatory element-binding protein (SREBP) pathway. In addition to regulating adipocyte formation, KLF4 influences inflammatory signaling, insulin sensitivity, and systemic energy balance—critical factors implicated in the pathophysiology of obesity [42, 43].

KLF5 also plays a crucial role in adipocyte development by regulating genes involved in lipid metabolism, such as fatty acid synthase (FAS) and acetyl-CoA carboxylase (ACC), and modulating insulin sensitivity through its interaction with PPARγ and transcription factors like FOXO and NF-κB [44, 45].

In individuals with a normal BMI, higher expression of KLF4 and KLF5 may promote physiological adipogenesis and efficient fat storage [46]. However, in obesity, various compensatory or maladaptive mechanisms may suppress the expression of KLF4 and KLF5, despite the presence of increased adiposity. These include adipocyte hypertrophy, chronic inflammation, endocrine dysregulation, metabolic feedback loops, and adipose tissue remodeling.

As adiposity increases, hypertrophic adipocytes may downregulate adipogenic transcription factors due to mechanical and metabolic stress, thereby reducing the expression of KLF4 and KLF5 [1]. Moreover, the persistent low-grade inflammation associated with obesity—characterized by elevated circulating cytokines—may further suppress these genes’ expression via inflammatory signaling pathways [47]. Endocrine disruptions involving insulin, leptin, or ghrelin may also interfere with KLF4 and KLF5 transcriptional activity [48, 49], particularly in insulin-resistant states where key intracellular signaling cascades are altered.

As a metabolic adaptation, the organism may shift lipid handling strategies away from classical adipogenic pathways toward alternative lipid storage mechanisms. This may occur without further recruitment of KLF4 or KLF5, especially once adipose tissue reaches a threshold of expansion. Functional redundancy among KLF family members may further permit such adaptation through compensatory expression of other transcriptionally similar proteins [50].

A negative feedback loop may also suppress KLF4 and KLF5 expression in response to excessive adiposity, effectively restraining further adipocyte formation and promoting metabolic conservation [51, 52]. Additionally, adipose tissue composition shifts in obesity—namely, a higher proportion of dysfunctional white adipose tissue relative to metabolically active brown adipose tissue—may also impact the expression profiles of these transcription factors [53].

The literature provides mixed evidence regarding the direction of KLF4–BMI associations. Baran et al. reported a positive correlation between KLF4 expression and BMI, whereas Wang et al. found that KLF4 expression was significantly reduced in individuals with obesity and was negatively correlated with BMI, triglyceride levels, and tumor necrosis factor-alpha (TNF-α) concentrations [54, 55].

These conflicting findings likely reflect tissue-specific expression (e.g., omental vs. visceral fat), differences in disease stage or severity, and variations in study populations.

Furthermore, Cervantes-Camacho et al. demonstrated that transient rather than sustained expression of KLF4 and KLF5 is essential during early adipocyte differentiation in 3T3-L1 cells [56], suggesting temporal dynamics in their function. Physiological states such as fasting, postprandial conditions, or weight loss may further modulate KLF4 expression, thereby influencing its associations with insulin sensitivity and other metabolic markers.

Associations between KLF4 and KLF5 with metabolic parameters

In the obese cohort, KLF4 showed notable inverse associations with several key metabolic indicators, including HOMA-IR, fasting insulin, and HbA1c (FDR < 0.1) (Table 4). In the Control group, a negative association was observed between KLF4 and HOMA-β, a marker of pancreatic β-cell function; however, this association did not remain significant after FDR correction (Table 3). These findings support the hypothesis that KLF4 plays a regulatory role in glucose homeostasis and insulin signaling.

The observed inverse relationship between KLF4 expression and insulin-related markers may reflect its dynamic and stage-dependent activity within visceral adipose tissue (VAT). KLF4 is known to function transiently during adipogenesis and metabolic transitions, which could explain the fluctuating expression patterns seen in individuals with obesity. While previous studies have reported chronically reduced KLF4 expression in other tissues or disease models [55], the present data suggest that in VAT, KLF4 activity may vary in response to metabolic stressors. Therefore, interpreting its expression based solely on steady-state levels may oversimplify its context-specific role in metabolic regulation.

In contrast, within the Obese group, KLF5 showed a weak but negative association with fasting blood sugar (FBS) (r = −0.521, FDR > 0.1), suggesting that lower expression of KLF5 may be associated with elevated glucose levels. Moreover, KLF5 expression was negatively correlated with HOMA-IR (Table 4), suggesting a potential link between KLF5 and insulin sensitivity in obesity. In the Control group, strong associations were observed between both KLF4 and KLF5 and serum amylase (Table 3), indicating a possible connection between these transcription factors and exocrine pancreatic function or regulation of digestive enzymes.

A growing body of evidence highlights the involvement of KLF4 and KLF5 in glucose metabolism. KLF4 promotes the phosphorylation of insulin receptor substrate (IRS) proteins, thereby facilitating downstream signaling pathways critical for glucose uptake [57, 58]. During ischemic conditions, KLF5 has been shown to upregulate key glycolytic enzymes and glucose transporters, including GLUT1 and GLUT4 [59]. Similarly, KLF4 has been reported to enhance GLUT1 expression during fasting [60]. Both transcription factors contribute to cellular protection during metabolic stress by preserving mitochondrial function and limiting the production of reactive oxygen species (ROS) [61, 62].

Insulin resistance is closely linked to chronic low-grade inflammation, in which cytokines such as interleukin-6 (IL-6) and TNF-α are major players. Both KLF4 and KLF5 regulate the transcription of these pro-inflammatory cytokines [51, 63]. Furthermore, they interact with other key transcriptional regulators, including FOXO1 and NF-κB, which are implicated in both immune modulation and metabolic control [45, 64, 65]. Notably, several intracellular signaling pathways modulated by KLF4 and KLF5, such as the PI3K/Akt/NF-κB axis, overlap with canonical insulin signaling cascades [66–73].

Recent investigations have further uncovered interactions between KLF4 and metformin, a first-line pharmacological treatment for type 2 diabetes [74]. In addition, emerging evidence suggests that KLF4 and KLF5 may be responsive to alterations in the gut microbiota, potentially linking the gut–adipose axis to systemic metabolic inflammation [75]. Variations in microbial composition could, therefore, indirectly influence KLF expression and downstream metabolic effects.

Associations between KLF4 and KLF5 with liver function

Markers indicative of hepatic dysfunction—ALT, AST, and ALP—were elevated in the obese cohort, suggesting possible liver impairment (Table 2). In the Control group, an inverse relationship was observed between AST and KLF4, indicating that higher KLF4 expression may be associated with lower transaminase activity. A modest positive association was also identified between KLF5 and ALP in the same group (Table 3).

Supplementary Table S7 details the specific associations observed between AST and multiple clinical variables. In addition, within the Control cohort, ALT was negatively correlated with fasting blood sugar (FBS) (r = −0.423, FDR > 0.1) and showed positive associations with both GGT (r = 0.703, FDR < 0.01) and ALP (r = 0.525, FDR > 0.1), as presented(Supplementary Table S8).

Although extensively studied in the context of adipogenesis, KLF4 is also expressed in hepatic tissue and plays diverse roles in liver biology [76]. It modulates lipid metabolism, mediates anti-inflammatory responses, regulates oxidative stress pathways, and interacts with transcription factors central to hepatic homeostasis and hepatocyte function [77]. The negative association between normalized KLF4 expression and AST levels in the Control group may reflect a protective role for KLF4 in maintaining liver integrity, possibly through its regulation of hepatic enzyme activity and anti-inflammatory mechanisms [78–80].

Associations between KLF4 and KLF5 with cholesterol

In the Obese group, a significant positive association was identified between KLF4 expression and HDL-C levels (Table 4). This relationship suggests that elevated KLF4 may be linked to increased HDL-C, a lipoprotein widely regarded as cardioprotective. This finding highlights a potential protective role for KLF4-mediated pathways in lipid regulation among individuals with obesity.

Obesity is commonly associated with dyslipidemia, as reflected in the significantly elevated LDL-C and total cholesterol (TC) levels observed in the Obese cohort compared to the Controls (Table 2). High LDL-C levels are well-documented contributors to cardiovascular risk, and KLF5 has been implicated in promoting these adverse lipid profiles [81–83]. In the Obese group, both KLF4 and KLF5 showed notable associations with LDL-C levels (Table 4). In the Control group, LDL-C was positively associated with insulin concentrations (r = 0.487, FDR > 0.1), as shown in (Supplementary Table S9).

The positive correlation between KLF4 and HDL-C became more pronounced after normalization (r = 0.523, FDR = 0.088), reinforcing the potential of KLF4 to support beneficial lipid metabolism. In contrast, the association between KLF4 and LDL-C, though initially moderate (r = 0.553), lost significance following FDR adjustment (FDR > 0.1), suggesting that this relationship may be more nuanced or indirect (Table 4) [84].

Regulation of lipid homeostasis by KLF4 in obesity likely reflects a multifactorial process involving interactions with insulin signaling, adipokines, and inflammatory cytokines that collectively influence lipoprotein synthesis and clearance. While KLF5 has been linked to worsening lipid parameters, KLF4 may exert an opposing effect, particularly through its association with HDL-C. It is essential to acknowledge that this study did not stratify LDL-C by subtype, which limits conclusions about the relative atherogenicity of different LDL particles.

Positive, strong associations between KLF4 and KLF5

A robust positive association was identified between KLF4 and KLF5 in both the Obese and Control groups, indicating a shared regulatory axis in adipose tissue. While KLF5 is predominantly implicated in terminal adipocyte differentiation, KLF4 plays a key role in initiating adipogenesis. Their co-expression reflects the multifaceted nature of transcriptional control in metabolic regulation.

Both transcription factors contribute to adipocyte differentiation and lipogenesis, potentially through direct molecular interactions such as heterodimerization. This cooperative regulation has also been observed in tumor biology, suggesting a conserved mechanism across tissue types [85–89]

The functional interaction between KLF4 and KLF5 is context-dependent—sometimes antagonistic, sometimes synergistic—depending on cellular environment, expression levels, and adipogenesis stage [90]. Their activity may differ between subcutaneous and visceral adipose depots and across adipocyte maturation phases. The concurrent upregulation of both in obesity likely reflects a unified transcriptional response to metabolic and hormonal cues associated with excess adiposity. Such coordinated expression may promote lipid accumulation and adipocyte expansion while maintaining cell viability under stress conditions.

The possibility of compensatory regulation also emerges: increased expression of one factor may partially offset downregulation or dysfunction of the other. This is consistent with their distinct temporal roles—KLF4 is predominantly active in early differentiation, and KLF5 sustains later stages.

Analysis of the KLF4:KLF5 ratio highlights this complexity. The ratio was substantially higher in the Obese group (mean = 0.340) relative to the Control group (∼0.01), as shown in Fig. 1. This shift may reflect a compensatory mechanism wherein increased KLF4 expression acts to buffer against the potentially pro-inflammatory effects of elevated KLF5 levels—a characteristic feature of obesity-related chronic inflammation.

From a therapeutic standpoint, KLF5 emerges as a potential target. Although both KLF4 and KLF5 are upregulated in obesity, the increase in KLF5 is more pronounced. Suppressing KLF5 expression or function could reduce inflammation while enhancing the relative activity of KLF4, thereby increasing the KLF4:KLF5 ratio. This modulation may yield protective metabolic effects in obese individuals [91]. Supplementary Table S10 outlines the associations between the KLF4:KLF5 ratio and various biochemical and metabolic parameters. Collectively, the findings support a model in which KLF5 assumes a more dominant and potentially maladaptive role in obesity, while KLF4 may exert a moderating or protective influence.

Other significant associations

Several additional associations emerged that warrant further investigation. An inverse relationship between KLF4 and LDH (Table 3) may suggest a role for KLF4 in preserving cellular integrity and metabolic function. The observed associations of KLF4 and KLF5 with CK, CK-MB, TIBC, iron, chloride, magnesium (Mg), and the sUA/sCr ratio (Tables 3 and 4, Supplementary Table S2, and Supplementary Table S10) point to their possible involvement in cardiac stress response [92], skeletal muscle health or injury [93, 94], iron homeostasis [95], electrolyte regulation, and renal function [50, 96, 97].

Limitations of the study

While the present findings provide preliminary insights into the associations of KLF4 and KLF5 with a range of biochemical and physiological markers, the study is limited by its relatively small sample size. The exclusive focus on women provides valuable insights into sex-specific metabolic adaptations and regulatory mechanisms, particularly within female cohorts, such as the Iranian population. Nonetheless, extrapolation to male populations requires dedicated investigation. Another limitation of the study lies in its scope, which did not extend to potential interactions between KLF4, KLF5, and other members of the KLF family. Future research should incorporate broader transcriptomic profiling to investigate the distinct and potentially synergistic roles of the KLF gene family in adipogenesis, insulin signaling, and metabolic dysregulation. In addition, functional validation through pathway analyses and in vitro or in vivo assays is necessary to clarify the mechanistic contributions of KLF4 and KLF5 in the metabolic context of obesity.

Conclusion

In adults with obesity, KLF4 and KLF5 exhibit a robust and coordinated expression pattern, with Original_KLF4 and Original_KLF5 showing strong covariation (r ≈ 0.80, p < 0.001). Moreover, the KLF4:KLF5 ratio is associated with key metabolic risk markers. Higher KLF4 levels are linked to improved insulin sensitivity, as reflected by lower HOMA-IR and HbA1c values, while KLF5 expression aligns more closely with dyslipidemia and elevated glycemic risk. This regulatory axis appears more pronounced in obese individuals, where associations between KLF expression and metabolic traits are stronger compared to non-obese controls, where such correlations are typically weaker or nonsignificant.

These findings suggest that the balance between KLF4 and KLF5 may serve as a potential biomarker for identifying individuals with obesity who are at heightened metabolic risk and may benefit from targeted interventions. Future therapeutic approaches should consider tissue-specific modulation of KLF activity to optimize metabolic outcomes while avoiding undesirable effects on lipid metabolism or adipose tissue expansion.

Before clinical implementation, these results require validation in larger, more diverse cohorts of obese individuals. Additionally, the development of reliable surrogate markers for KLF4/KLF5 activity is essential, along with stringent monitoring of glucose, lipid, and cardiovascular profiles to ensure safety and efficacy in any translational application.

Abbreviations

ACC

Acetyl-CoA carboxylase

ALP

Alkaline phosphatase

ALT

Alanine aminotransferase

AMH

Anti-Müllerian hormone

ANCOVA

Analysis of covariance

AST

Aspartate aminotransferase

AVI

Abdominal volume index

BAI

Body adiposity index

BH procedure

Benjamini-hochberg procedure

BLAST

Basic local alignment search tool

BMI

Body mass index

C/EBPs

CCAAT/enhancer binding protein family

CI

Conicity index

CK

Creatine kinase

CK-MB

Creatine kinase myocardial band

DBP

Diastolic blood pressure

ECL

Electrochemiluminescence

EDTA

Ethylenediaminetetraacetic acid

FAS

Fatty acid synthase

FBS

Fasting blood sugar

FDR

False discovery rate

FOXO

Forkhead box protein O1

GGT

Gamma-glutamyl transferase

GLUT1

Glucose Transporter Type 1

GLUT4

Glucose Transporter Type 1

HbA1c

Hemoglobin A1C

HC

Hip circumference

HDL-C

High-density lipoprotein cholesterol

HOMA-beta

Homeostatic model assessment β-Cell function

HOMA-IR

Homeostatic model assessment of insulin resistance

hs-CRP

high-sensitivity C-reactive protein

IL-6

Interleukin 6

Insulin

Fasting insulin

IR

Insulin resistance

IRS

Insulin receptor substrate

KLF4 and KLF5

Krüppel-like transcription factors four and five

KLF4/KLF5

KLF4 to KLF5 ratio

LAGB

Laparoscopic adjustable gastric bypass

LAP

Lipid accumulation product index

LDH

Lactate dehydrogenase

LDL-C

Low-density lipoprotein cholesterol

LDL-C/HDL-C

LDL-C to HDL-C ratio

miRNAs

MicroRNAs

Mg

Magnesium

NCBI

National center for biotechnology information

Normal_KLF4 and normal_KLF5

Normalized data by a two-step process

NSAID

Non-steroidal anti-inflammatory drug

NF-κB

Nuclear factor kappa B

PPARγ

Peroxisome proliferator-activated receptor gamma

ROS

Reactive oxygen species

RYGB

Roux-en-Y gastric bypass

SBP

Systolic blood pressure

SG

Sleeve gastrectomy

SREBP

Sterol regulatory element-binding protein

TC

Total cholesterol

TG

Tryglyceride

TIBC

Total iron-binding capacity

TNF-α

Tumor necrosis factor-alpha

TP

Total protein

sUA/sCr

Serum uric acid to serum creatinine ratio

VAI

Visceral adiposity index

VLDL-C

Very low-density lipoprotein cholesterol

WC

Waist circumference

WHR

Waist-to-hip ratio

WHtR

Waist-to-height ratio

WWI

Weight-adjusted waist circumference index

Authors’ contributions

S.E. and A.N. conceived and designed the study. H.Z.M. provided the tissue samples. D.J.M., G.P., S.E., and A.N. conducted the laboratory experiments. M.M.M. and A.N. performed the data analysis. D.J.M. drafted the manuscript, which was critically reviewed and revised by A.N. and S.E. All authors contributed to the manuscript’s development and approved the final version for submission.

Funding

This research was supported by the Tehran University of Medical Sciences and Health Services under grant number 1402-4-368-68185.

Data availability

The data sets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of Tehran University of Medical Sciences (approval ID: IR.TUMS.Medicine.REC.1402.414). Written informed consent was obtained from all participants after they were provided with comprehensive information about the study objectives and procedures.

Competing interests

The authors declare no competing interests.

Conflict of interest

The authors declare no conflicts of interest related to the publication of this work.