Thioredoxin-interacting protein is associated with obesity-induced insulin resistance in PCOS patients: a large-scale case-control study

Abstract

Purpose: Polycystic ovary syndrome (PCOS) is a common endocrine disorder often associated with obesity and insulin resistance (IR), though the role of thioredoxin-interacting protein (TXNIP) in obesity-induced IR in PCOS remains unclear. This study explores the relationship between TXNIP levels and obesity-associated IR in women with PCOS.

Methods: A case-control study was conducted from January 2019 to December 2020, including 161 women with PCOS and 107 healthy controls. PCOS patients were categorized into insulin-resistant (IR) and non-IR subgroups, further divided by BMI into obese, overweight, and normal weight groups. The metabolic parameters such as cholesterol, triglycerides, fasting blood glucose, homocysteine, and serum TXNIP levels were measured. Logistic regression assessed the relationship between TXNIP expression and metabolic dysfunction.

Results: TXNIP levels were significantly higher in the PCOS group compared to controls (67% increase), with a further 56% increase in the IR subgroup. The TXNIP levels were elevated in the obese group compared to overweight and normal weight groups (P < 0.05). TXNIP expression was negatively correlated with obesity (R = - 0.116, P = 0.007) and HDL cholesterol (R = - 0.196, P = 0.001), but positively associated with triglycerides (R = 0.181, P = 0.003) and homocysteine (R = 0.130, P = 0.034). After adjusting for confounders, TXNIP remained significantly associated with IR (P < 0.05). TXNIP demonstrated excellent diagnostic performance in distinguishing IR from non-IR PCOS patients, with an AUC of 0.89 (95% CI 0.84-0.94; P < 0.001).

Conclusions: TXNIP is significantly correlated with IR in women with PCOS, highlighting its potential as a biomarker for metabolic abnormalities. Further research is needed to fully understand its role in obesity-induced IR in PCOS.

Keywords: Insulin resistance; Metabolic; Obesity; Polycystic ovary syndrome; Thioredoxin-interacting protein.

Fig. 1

TXNIP expression in PCOS and association with insulin resistance and obesity. A TXNIP expression is elevated in PCOS compared to healthy controls. TXNIP expression (ng/mL) was significantly higher in individuals with PCOS compared to healthy controls, suggesting a potential role of TXNIP in PCOS pathophysiology. B This graph shows a TXNIP expression is increased in insulin-resistant (IR) PCOS compared to non-insulin-resistant (non-IR) PCOS. TXNIP levels were significantly higher in IR-PCOS individuals than in non-IR-PCOS individuals, indicating an association between TXNIP and insulin resistance in PCOS. C TXNIP expression differs among PCOS subgroups based on obesity status. TXNIP levels were significantly lower in overweight (OW-PCOS) and obese (OB-PCOS) individuals compared to controls (Ctrl), with p-values of 0.034 and 0.012, respectively. These findings suggest that TXNIP expression may be influenced by BMI in PCOS patients. Each data point represents an individual, and horizontal bars indicate the mean ± standard error of the mean (SEM). Statistical significance was determined using independent t-tests

Fig. 2

Collection of the TXNIP expression with metabolic and hormonal parameters in PCOS. A Negative correlation between TXNIP expression and BMI in control (Ctrl) and PCOS groups. TXNIP expression levels (ng/mL) were plotted against BMI (kg/m²) for individuals with PCOS (red) and healthy controls (green). A weak but significant negative correlation was observed, suggesting that TXNIP levels tend to decrease with increasing BMI in PCOS. B Negative correlation between TXNIP expression and BMI in insulin-resistant (IR) and non-insulin-resistant (non-IR) PCOS groups. TXNIP expression (ng/mL) was plotted against BMI (kg/m²) for IR-PCOS (red) and non-IR-PCOS (blue) groups, revealing a significant negative correlation, further supporting an association between TXNIP and metabolic alterations in PCOS. C Correlation matrix depicting relationships between TXNIP and key metabolic and reproductive markers. Scatter plots show pairwise correlations between TXNIP, anti-Müllerian hormone (AMH), triglycerides (TG), high-density lipoprotein (HDL), and homocysteine (HCY), with Pearson correlation coefficients (r) and p-values indicated. TXNIP was positively correlated with TG and HCY but negatively correlated with HDL. Additionally, TXNIP exhibited a weak but significant positive correlation with AMH, indicating a potential role in ovarian function. Strong negative correlations were observed between HDL and TG, reinforcing the established inverse relationship between these lipid markers. The size and color intensity of the scatter points represent data density

Fig. 3

Diagnostic performance of TXNIP expression across different groups. A ROC curve analysis comparing TXNIP expression between PCOS patients and age- and BMI-matched healthy controls. The AUC was 0.5667 indicating a non-significant ability to distinguish between groups. B ROC curve comparing insulin-resistant PCOS (IR-PCOS) versus non-insulin-resistant PCOS (non-IR-PCOS) patients. TXNIP demonstrated strong discriminative power with an AUC of 0.8950 (95% CI 0.8435–0.9465; P < 0.0001). C ROC analysis comparing obese PCOS patients with healthy controls, yielding a low AUC of 0.5366, indicating poor diagnostic utility in this subgroup. The red and black lines represent the reference and observed performance curves, respectively