Fig 1: Correlations between plasma CES1 concentrations and enalapril pharmacokinetics in healthy individuals. Measured plasma CES1 versus: (A) enalapril area under the curve (AUC E), (B) enalaprilat area under the curve (AUC ET), and enalaprilat-to-enalapril AUC ratio (ET/E AUC) (C). (D–F) G143E genotype–adjusted plasma CES1 versus the same pharmacokinetic parameters as panels (A–C), where plasma CES1 concentrations were reduced by 50% in G143E heterozygous carriers to reflect reduced hepatic enzymatic function. Dashed lines represent the linear regression fit with 95% confidence bands (gray shading). Sample size (n) is indicated in each panel.
Fig 2: Association between plasma CES1 and liver transaminase levels stratified by disease status. (A) Plasma CES1 vs ALT in healthy subjects. (B) Plasma CES1 vs AST in healthy subjects. (C) Plasma CES1 vs ALT in MASLD patients. (D) Plasma CES1 vs AST in MASLD patients. (E) Plasma CES1 vs ALT in all subjects. (F) Plasma CES1 vs AST in all subjects. Each point represents an individual subject (green = healthy controls, red = MASLD patients). Lines indicate linear regression fits. Text boxes display the coefficient of determination (R2), P value, and sample size (n) for each analysis.
Fig 3: Plasma and hepatic CES1 levels in healthy and MASLD subjects. (A) Correlation between plasma and hepatic CES1 protein concentrations in MASLD patients. The scatter plot compares plasma CES1 concentrations (ng/mL) with hepatic CES1 expression measured in liver S9 fractions (pmol/mg protein) in 17 patients with biopsy-confirmed MASLD. The red line indicates the linear regression line of best fit. Comparison of CES1 concentrations in (B) plasma (ng/mL) and (C) liver S9 fractions (pmol/mg protein) between healthy subjects and patients with MASLD. Box plots display the median (horizontal line), interquartile range (box), whiskers (data range within 1.5 × IQR), mean (green triangle), and individual data points (dots). Statistical significance was assessed using the Mann-Whitney U test. ∗∗P < .0001.
Fig 4: Distribution of plasma CES1 levels by MASLD activity score. (A) Stacked histogram showing the distribution of plasma CES1 concentrations across NAS categories. (B) Box plot of plasma CES1 levels stratified by NAS categories, showing median, interquartile range, mean (dashed line), and individual data points. Significance brackets indicate pairwise Mann-Whitney U tests with Bonferroni correction for multiple comparisons (3 comparisons, corrected α = 0.0167). ∗∗P < .01. An overall comparison was performed using the Kruskal-Wallis test.
Fig 5: Lasso logistic regression classification of MASLD and disease severity using plasma CES1 and clinical lab features. A Lasso logistic regression model used 3 feature sets: CES1 Only (plasma CES1), Labs Only (AST, ALT, bilirubin, WBC, BMI), and CES1 and Labs combined; model performance was evaluated by LOOCV. LOOCV accuracy, F1, and AUC for each feature set (A, E, I); LOOCV ROC curves with AUC values in the legends (B, F, J); display LOOCV confusion matrices for the CES1 only feature set (C, G, K); feature importance analysis (D, H, L). Row 1 (A–D) presents the binary task (healthy vs MASLD), row 2 (E–H) presents staging within biopsy-positive disease (NAS 1–4 vs NAS ≥ 5), and row 3 (I–L) presents the multiclass task (healthy, NAS 1–4, NAS ≥ 5).
Supplier Page from BioLegend for Recombinant Human CES1 (carrier-free)