Identification and Validation of a Common Molecular Signature of Progressive Fibrosis in Human Livers

• Author / Creator

  Hsu, Mu Hao

• Over 25% of the world population suffer from chronic liver disease. Most common causes are long-term viral infections, alcohol abuse, and non-alcoholic fatty liver disease (NAFLD) associated with obesity, diabetes or hyperlipidemia. Chronic liver disease due to any cause is characterized by repetitive or long-lasting liver injury which may result in liver fibrosis. Regardless of underlying etiology, advanced liver fibrosis is the final common pathway that eventually leads to poor clinical outcomes. Histopathology assessment of liver biopsy plays a central role in evaluating liver fibrosis, but it is subjective, suffers from suboptimal reproducibility, and, more importantly, cannot predict fibrosis progression. The natural history of fibrosis progression varies, with a high degree of heterogeneity among patients. The factors that drive this diversity are unknown. In addition, it is unclear if there is a common molecular signature for progressive fibrosis independent of its etiology. This thesis aims to identify a common molecular signature associated with advanced fibrosis that is shared by different liver diseases in humans, and to translate this knowledge to develop a surrogate biomarker test for clinical use.
  To identify a common molecular signature of advanced fibrosis, I analyzed 304 Affymetrix genome-wide microarrays of fresh livers from 312 patients with various chronic liver diseases (NAFLD, alcoholic liver disease, viral hepatitis B or C) available in open access databases. Using a discovery set (n=70), I defined a 48-gene signature with the highest discriminate ability for advanced fibrosis using multiple machine learning class prediction algorithms and class comparison methods. The 48-gene signature was validated in three independent validation sets (n=70, 91, and 73) with over 0.96 area under the receiver operating characteristic curve and over 93% accuracy for histologically-proven advanced fibrosis.
  As a clinical translational approach, the 48-gene signature was then analyzed in 348 formalin-fixed paraffin-embedded (FFPE) clinical liver biopsies and 15 paired fresh and FFPE explant livers by a digital direct nucleic acid profiling method (NanoString assay) to understand if multiplex gene expression quantification in clinical liver biopsies is a feasible and reliable approach. NanoString gene expression levels were strongly correlated between paired fresh and FFPE livers (r=0.944, p<0.001). NanoString gene expression measurements were highly reproducible in repeat runs with different RNA input quantities (r=0.946-0.995, p<0.001), different operators (r=0.949-0.992, p<0.001), and different lots of reagents (r=0.998-0.999, p<0.001).
  To see if gene signatures in liver biopsies can predict fibrosis progression, the 48-gene signature in FFPE clinical liver biopsies was analyzed in 299 patients with different liver diseases (122 patients with recurrent viral hepatitis C, 76 with autoimmune hepatitis, and 101 with NAFLD). Patients with a high 48-gene signature had a significantly higher probability of fibrosis progression compared to patients with a low 48-gene signature in all three liver diseases, but the probability of fibrosis progression was similar across different histological fibrosis stages. The results supported that the 48-gene signature can predict progressive fibrosis earlier than histology in different liver diseases.
  For protein level validation, two of the genes in the 48-gene signature were analyzed by immunohistochemistry in 94 patients with available liver tissue after NanoString assay. The increase in mRNA expression of the two genes led to an increased expression of the corresponding protein. Patients with a high protein expression of the two genes had significantly higher probability of fibrosis progression compared to patients with a low protein expression.
  The research presented in this thesis is the first comprehensive collection of a gene expression signature that drives advanced liver fibrosis in humans in the context of various common causes of chronic liver disease. The 48-gene signature may serve as a prognostic biomarker for fibrosis progression in patients with different chronic liver diseases. This biomarker test can be easily implemented as a clinical assay.

• Subjects / Keywords

  • liver fibrosis
  • gene signature
  • NAFLD
  • gene expression
  • liver biopsy
  • chronic liver disease
  • nonalcoholic fatty liver disease
  • microarray
  • Nanostring

• Graduation date

  Spring 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-sy68-hg63

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.