Author ORCiD
Date of Award
2023
Document Type
Thesis
Degree Name
Doctor of Philosophy (PhD)
Thesis Advisor
Sanford M. Simon
Keywords
Fibrolamellar Hepatocellular Carcinoma (FLC), DNAJB1::PRKACA chimera, transcriptomic signatures, RNA-seq analysis, PKA activation, liver cancer biomarkers
Abstract
Fibrolamellar Hepatocellular Carcinoma (FLC) is a rare liver cancer affecting adolescents and young adults. Almost all FLC patients present a somatic heterozygous deletion in chromosome 19p13.12, producing the chimeric transcript DNAJB1::PRKACA, which is sufficient to drive FLC in mice. Nevertheless, a few cases of liver tumors presenting FLC histopathological characteristics but missing the chimera DNAJB1::PRKACA have been reported. They have either 1) a lack of protein expression of PRKAR1A, 2) the presence of the chimeras ATP1B1::PRKACB or ATP1B1::PRKACA, or 3) a chromosome gain of PRKACA or loss of PRKAR2A, accompanied with inactivation of BAP1. A few RNA-seq studies compared FLC tumors to adjacent nontransformed liver (normal) samples, revealing many differences. However, they present experimental and methodological problems and limitations, including using tumors categorized as FLC based only on histopathology, studying small datasets, using a high number of unpaired samples and surrogate normal samples, aggregating samples from different studies without a proper assessment of the variability, and using outdated algorithms to estimate gene expression. These problems resulted in 16-47% agreement between the differentially expressed genes obtained in them. In this thesis, I addressed these issues and comprehensively characterized the transcriptome of FLC. To accomplish this, I performed RNA-seq on 143 frozen samples of FLC patients, the largest RNA-seq dataset of FLC to date. This was supplemented with 100 FLC samples from other studies. All were reanalyzed using state-of-the-art bioinformatic methods, checking for detectability and consistency among datasets. To identify the transcriptional FLC signature, only FLC tumors harboring the chimera DNAJB1::PRKACA and having a normal paired sample were selected and divided into testing, perfecting, and validation groups. This resulted in 287 up- and 406 downregulated genes, comprising a transcriptomic signature which demonstrated to be independent of the experimental or sequencing procedure. It helped to identify PKA activation as a unifying phenotype for all FLC tumors. Then, I explored whether and to what extent the increased PKA activation could be controlling the transcriptomic FLC signature by impacting transcription factors, identifying 82 transcription factors linked to at least one gene of the transcriptomic FLC signature in liver ChIP-seq data. This revealed that CEBPB and NFIC3 are notably related to genes in the transcriptomic FLC signature. I also studied if the changes observed in the transcriptomic FLC signature can be explained by somatic mutations or methylation events, finding little explanation. We also explore if the transcriptional changes are reflected in downstream omic levels, finding a very high correlation with the changes obtained in the proteome and phosphome of FLC tumors. The transcriptomic FLC signature can distinguish FLC from other tumors. It was further characterized by comparing it with the genes differentially expressed in 929 RNA-seq paired samples from other liver cancers, including hepatocellular carcinoma (HCC, N=642), hepatoblastoma (HBL, N=148), and intrahepatic cholangiocarcinoma (iCCA, N=139). We found 160 up- and 48 down-regulated genes in common in all these liver cancers, but 20 up- and 121 down-regulated only in FLC. The contribution of the different cell types in an FLC tumor to the transcriptomic FLC signature was also studied, providing for the first time a single-cell spatial transcriptomic characterization of FLC. It showed clear differential expression patterns in tumor, normal, and stromal cells. By performing the largest transcriptomic study on liver cancers, I identified a transcriptomic FLC signature that transcends processing or analysis methods and can distinguish FLC tumors from normal samples or other liver tumors. I present its utility by 1) validating in vitro and in vivo models and 2) exploring biological questions about FLC. This tool, the methodology designed, and the large amount of multi-omic data generated represent an important source of information for the study of all liver cancers. They will be released on an interactive website to facilitate access without requiring bioinformatics expertise.
DOI
10.48496/9jv2-bf55
License and Reuse Information
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.
Recommended Citation
Requena Anicama, David R., "Multiomics for Elucidating the Pathogenesis of Liver Cancers: a Transcriptomic Signature of Fibrolamellar Hepatocellular Carcinoma" (2023). Student Theses and Dissertations. 851.
https://doi.org/10.48496/9jv2-bf55
Comments
A Thesis Presented to the Faculty of The Rockefeller University in Partial Fulfillment of the Requirements for the degree of Doctor of Philosophy