Background and objectives: There is a need to develop noninvasive diagnostic and monitoring tests for nonalcoholic fatty liver disease (NAFLD).
Methods: Using modified-aptamer proteomics, we analyzed 5220 proteins in each of 2852 serum samples from 636 patients with histologically confirmed NAFLD. We created and validated dichotomous protein-phenotype models for clinically relevant severity of steatosis (grade 0 vs. 1-3), hepatocellular ballooning (grade 0 vs. 1 or 2), lobular inflammation (grade 0-1 vs. 2-3), and fibrosis (stages 0-1 vs. 2-4).
In training/paired validation, the results for the four protein models based on 37 analytes (18 not previously associated with NAFLD) were as follows: fibrosis (AUC 0.92/0.85); steatosis (AUC 0.95/0.79); inflammation (AUC 0.83/0.72); and ballooning (AUC 0.87/0.83). Multiplying the outputs of each component model (AUC 0.93/0.85) yielded a different outcome, at-risk NASH, defined as steatohepatitis with NAS 4 and at least one point in each component, as well as fibrosis stage 2 or higher. In addition, their ability to detect changes in histology following treatment with a placebo, pioglitazone, vitamin E, or obeticholic acid was evaluated (OCA). Component model scores improved significantly in the active therapies compared to the placebo group, and distinct effects of vitamin E, pioglitazone, and OCA were identified.
Conclusions: Serum protein scanning revealed signatures corresponding to NAFLD's key liver biopsy components. The developed models were sensitive enough to characterize longitudinal changes for three distinct drug interventions. These findings support the ongoing validation of these proteomic models to enable a "liquid biopsy"-based evaluation of NAFLD.
Lay summary: Liver biopsies can determine the presence of liver inflammation, scarring, and fat in NASH patients. In this study, we use an innovative protein scanning technique to develop and validate blood-based signatures of liver damage that have the potential to replace liver biopsies in diagnosing NASH and monitoring treatment response. These signatures can also provide insights into treatment-mediated responses and mechanisms of action for NASH drug developers.
Effects and repercussions: An aptamer-based protein scan of serum proteins was conducted to identify diagnostic signatures of the critical histological features of nonalcoholic steatohepatitis, for which there are no currently approved noninvasive diagnostic tools. Specific protein signatures correlated with the presence and severity of NASH and its histological components and were also time-sensitive. These are fundamental initial steps in establishing a serum-proteome-based diagnostic signature of NASH and provide a rationale for using these signatures to test treatment response and identify several novel targets to evaluate in the pathogenesis of NASH.
Methods: Using modified-aptamer proteomics, we analyzed 5220 proteins in each of 2852 serum samples from 636 patients with histologically confirmed NAFLD. We created and validated dichotomous protein-phenotype models for clinically relevant severity of steatosis (grade 0 vs. 1-3), hepatocellular ballooning (grade 0 vs. 1 or 2), lobular inflammation (grade 0-1 vs. 2-3), and fibrosis (stages 0-1 vs. 2-4).
In training/paired validation, the results for the four protein models based on 37 analytes (18 not previously associated with NAFLD) were as follows: fibrosis (AUC 0.92/0.85); steatosis (AUC 0.95/0.79); inflammation (AUC 0.83/0.72); and ballooning (AUC 0.87/0.83). Multiplying the outputs of each component model (AUC 0.93/0.85) yielded a different outcome, at-risk NASH, defined as steatohepatitis with NAS 4 and at least one point in each component, as well as fibrosis stage 2 or higher. In addition, their ability to detect changes in histology following treatment with a placebo, pioglitazone, vitamin E, or obeticholic acid was evaluated (OCA). Component model scores improved significantly in the active therapies compared to the placebo group, and distinct effects of vitamin E, pioglitazone, and OCA were identified.
Conclusions: Serum protein scanning revealed signatures corresponding to NAFLD's key liver biopsy components. The developed models were sensitive enough to characterize longitudinal changes for three distinct drug interventions. These findings support the ongoing validation of these proteomic models to enable a "liquid biopsy"-based evaluation of NAFLD.
Lay summary: Liver biopsies can determine the presence of liver inflammation, scarring, and fat in NASH patients. In this study, we use an innovative protein scanning technique to develop and validate blood-based signatures of liver damage that have the potential to replace liver biopsies in diagnosing NASH and monitoring treatment response. These signatures can also provide insights into treatment-mediated responses and mechanisms of action for NASH drug developers.
Effects and repercussions: An aptamer-based protein scan of serum proteins was conducted to identify diagnostic signatures of the critical histological features of nonalcoholic steatohepatitis, for which there are no currently approved noninvasive diagnostic tools. Specific protein signatures correlated with the presence and severity of NASH and its histological components and were also time-sensitive. These are fundamental initial steps in establishing a serum-proteome-based diagnostic signature of NASH and provide a rationale for using these signatures to test treatment response and identify several novel targets to evaluate in the pathogenesis of NASH.