Transcriptomic analysis of hepatocellular carcinoma reveals molecular features of disease progression and tumour immune biology

Immunotherapies for the treatment of a variety of malignancies have become one of the most exciting developments in cancer treatment of recent years. However, it is becoming clear that the immune cell composition of a number of solid tumours and their surrounding stromal tissue is important in their response to PD-1/PD-L1 checkpoint inhibitor treatment, and may in part be responsible for the emergence of inhibitor-refractile tumours and therapeutic failure.

Hepatocellular carcinoma (HCC) is a common cause of cancer deaths with a particularly poor prognosis. The majority of HCC cases arise as a result of chronic inflammatory liver damage as a progression from liver fibrosis or cirrhosis.

A recent publication by Okrah et al., describing molecular aspects of HCC progression illustrates the power of multigene transcriptomic signatures in describing the etiology of disease and indicating potential treatment modalities.

Whole transcriptome RNA-Seq analysis was performed on RNA extracted from formalin fixed, paraffin embedded tissue sections from a set of clinically graded human HCC tumour samples (T1, T2, T3 n=98) and patient matched non tumour cirrhotic samples (n=78/98).

Clustering analysis of the data revealed three distinct gene signatures whose expression levels were directionally associated with disease stage. Cluster 1 contained genes involved in immune stimulation, and was consistently down regulated from cirrhosis through tumour stage T1 to stage T3. Cluster 3 was dominated by genes associated with WNT pathway activation. Expression of this cluster increased from cirrhosis to T3.

CD8+ T lymphocyte immunohistochemistry (IHC) using an anti CD8A antibody was also performed on tumour sections. An expression signature for cytotoxic T effector cells (Teff sig) correlated strongly with the expression of CD8A in the HCC samples. This Teff signature also correlated positively with the expression of genes in cluster 1 and negatively with genes in cluster 3. Together these data suggested an increasingly immunosuppressive tumor environment during HCC progression.

CD8+ T lymphocyte IHC also allowed tumour sections to be classified into intra-tumoural (“inflamed”) or peri-tumoural (“excluded”) tumour types, depending on the localization of the CD8A positive cells. These classes comprised 27% and 23% respectively of the total HCC samples.

Differential gene expression analysis of these two populations produced a signature of 85 genes shown to be upregulated in specimens of the peri-tumoural phenotype. Top scoring genes in this “CD8 exclusion” signature represented collagens, extracellular matrix (ECM) and pathways known to be involved in upregulating fibrosis (TGFβ, PDGFR, SHH and Notch).

This fibrotic gene set was further refined to set of 23 genes referred to as “Module 1”, mostly representing collagens and ECM components. Module 1 correlated with gene expression in multiple solid tumour types, and was able to identify specific molecular subtypes in previously published data sets from both gastric (EMT subtype) and colorectal (CMS4 subtype) cancer.

When applied back to the HCC sample set from the study, the Module 1 signature could distinguish specimens with intra-tumoural from peri-tumoural distribution of CD8+ T cells with a reasonable degree of statistical significance (76% specificity and 85% sensitivity). This analysis was also prospectively validated in a separate, independent cohort of HCC specimens.

These findings suggested a potential causal connection between elevated levels of fibrosis (possibly modulated by TGFβ, PDGFR, SHH or Notch pathways) and the CD8+ T cell excluded tumour phenotype.

To investigate this idea further in a STAM™ model of murine HCC, a TGFβ neutralizing antibody was able to alter the distribution of CD8+ T cells in the tumours from excluded to a more infiltrated phenotype. Treatment also reduced the degree of fibrosis and reduced the tumour burden in the livers of treated animals both reducing the number and size of tumour nodules.

This finding is interesting given that in another similar recent study (Mariathasan et al. (2018) Nature 554, 544-548) in an EMT6 mammary carcinoma cell tumor model, both PD-L1 checkpoint blockade and inhibition of TGFβ signaling were required for tumor regression.

Overall, the data presented in this paper suggest a rationale for immunotherapy early in HCC disease progression and potentially in combination with an anti-fibrotic strategy.

Get in touch

Find out how Epistem can assist with your research…

Thank you

We will be in touch as soon as possible.

Epistem - Your Ideal Partner for Biomarker Discovery Research

Epistem is your ideal partner for biomarker discovery and validation for use in clinical and pre-clinical studies. We have extensive experience in gene expression analysis with qPCR and microarrays in our GCLP-accredited labs and have recently expanded our next generation sequencing capabilities with the purchase of a NextSeq550 machine. We have considerable experience in analysis of tissues with low RNA input such as liquid biopsies, single cells and hair bulbs and can provide full in house bioinformatics support for all of our studies.

Our preclinical services include orthotopic imaging models (including xenografts), for a number of different cancers. All our models are supported by GCLP-accredited histology/IHC analytical services, including optimized IHC protocols for PD1/PD-L1, with a demonstrated ability to provide analyses for early phase clinical studies.

Okrah et al. (2018) Transcriptomic analysis of hepatocellular carcinoma reveals molecular features of disease progression and tumor immune biology. NPJ Precis Oncol. 2, Article number: 25

About Epistem

Epistem's contract research service is committed to providing reliable, innovative and transferable pre-clinical models and services to support decision making throughout the drug discovery and development pipeline.

Tel: +44 (0)161 850 7600  Email: