The expression signature of cancer‐associated <scp>KRAB</scp>‐<scp>ZNF</scp> factors identified in <scp>TCGA</scp> pan‐cancer transcriptomic data

Marta Machnik, Rafał Cylwa, Kornel Kiełczewski, Przemysław Biecek, Triantafillos Liloglou, Andrzej Maćkiewicz, Urszula OleksiewiczMolecular 2018 · DOI

TL;DR

This study identifies a specific group of KRAB-ZNF genes that are consistently overexpressed in a wide range of cancers compared to normal tissues. These genes and their splicing variants show correlations with tumor characteristics and patient survival, suggesting they could be valuable biomarkers or therapeutic targets in cancer.

What Problem Does This Paper Address?

The KRAB-ZNF (Krüppel-associated box domain zinc finger) gene family is a large and complex group of genes involved in epigenetic regulation. While known to play a role in normal development and cellular processes, their specific involvement in cancer has remained largely unexplored due to their sheer number and the complexity of their isoforms. This paper aims to systematically investigate the expression patterns of KRAB-ZNFs across various cancer types to identify those that are dysregulated in cancer and explore their potential roles as biomarkers or therapeutic targets.

Methods

The researchers utilized a multi-pronged approach:

  • TCGA Pan-Cancer Analysis: They analyzed RNA-sequencing data from over 6,700 tumor and normal tissue samples across 16 different cancer types from The Cancer Genome Atlas (TCGA) database. This allowed for a broad, pan-cancer view of KRAB-ZNF expression.
  • Differential Expression Analysis: Standard statistical methods (DESeq, binomial test, t-tests with FDR correction) were employed to identify KRAB-ZNFs with significantly altered expression levels between tumor and normal tissues.
  • Independent Validation: To confirm their findings, the study used independent datasets:
  • Cell Lines: RT-qPCR was used to assess KRAB-ZNF expression in lung and breast cancer cell lines.
  • Tissue Microarrays: Immunohistochemistry was performed on lung and breast cancer tissue microarrays to validate protein expression.
  • Isoform Analysis: The study delved into the complexity of KRAB-ZNF splicing variants, analyzing their differential expression and potential "variant switches" between normal and tumor tissues.
  • Clinicopathological Correlation: Statistical analyses were performed to correlate KRAB-ZNF expression with various clinical parameters, including tumor histology, molecular subtypes, patient survival, and receptor status (ER, PR, HER2).

Datasets

  • The Cancer Genome Atlas (TCGA): This was the primary dataset, providing RNA-seq and clinicopathological data for 6,727 tumor and normal tissue samples from 16 cancer types.
  • Independent Lung and Breast Cancer Tissue Samples: A cohort of 58 matched lung cancer and normal tissue pairs, and unmatched breast cancer tissues were used for validation.
  • Breast Cancer cDNA Array II: This dataset provided clinical and molecular information for breast cancer samples.
  • The Human Protein Atlas: Used for immunohistochemical validation of protein expression.
  • Various Lung and Breast Cancer Cell Lines: A panel of cell lines representing different subtypes was used for validation.

Key Insights and Ideas

  • A Distinct Cancer-Associated Signature: Despite the vastness of the KRAB-ZNF family, a small, specific subset of 16 genes consistently shows upregulation across multiple cancer types. This suggests a targeted role for these genes in tumorigenesis.
  • Isoform Complexity Matters: The study highlights that not only the gene but also its specific splicing isoforms can be differentially expressed in cancer, and these variants often show coordinated upregulation.
  • Beyond Simple Upregulation: While most KRAB-ZNFs are upregulated, the paper also notes instances where isoforms might be switched or have altered expression patterns, hinting at complex regulatory mechanisms.
  • Clinical Relevance: The identified KRAB-ZNFs are not just molecular curiosities; their expression levels correlate with key clinical features like tumor histology, molecular subtypes, and, crucially, patient survival. This points to their potential as prognostic markers.

Key Results

  • Out of 381 analyzed KRAB-ZNFs, 56 showed differential expression in tumors compared to normal tissues (14.7%).
  • A distinct cluster of 16 KRAB-ZNFs was commonly upregulated across multiple cancer cohorts.
  • This upregulation was confirmed in independent lung and breast cancer cell lines and tissues.
  • The majority of KRAB-ZNF splicing variants (84.5%) were overexpressed in cancer tissues.
  • In LUAD and LUSC, the expression of certain KRAB-ZNFs correlated with histological subtypes, molecular subtypes, and patient survival. For example, high ZNF695 expression was associated with poorer survival in LUSC.
  • In breast cancer, KRAB-ZNF expression correlated with histological subgroups (basal-like, HER2+), receptor status (ER+, PR+), and methylation clusters.
  • Several KRAB-ZNFs were identified as potential prognostic markers, with high expression correlating with either shorter or longer survival depending on the gene and cancer type.

Limitations and Future Work

  • Functional Elucidation: While the study identifies cancer-associated KRAB-ZNFs and their correlations, their precise molecular functions in cancer remain largely unknown. The authors emphasize the need for further functional studies to understand their mechanisms of action.
  • Isoform-Specific Roles: The study acknowledges that while many isoforms are upregulated, some might be truncated or lack functional domains, and their specific roles require further investigation.
  • Experimental Validation: The findings are primarily based on transcriptomic data. Further experimental validation of the identified KRAB-ZNFs as oncogenic drivers or tumor suppressors is necessary.
  • Therapeutic Potential: The authors suggest that these KRAB-ZNFs could serve as novel oncologic biomarkers and treatment targets, but this hypothesis requires rigorous experimental testing.

Summary updated

Abstract

The KRAB-ZNF (Krüppel-associated box domain zinc finger) gene family is composed of a large number of highly homologous genes, gene isoforms, and pseudogenes. The proteins encoded by these genes, whose expression is often tissue-specific, act as epigenetic suppressors contributing to the addition of repressive chromatin marks and DNA methylation. Due to its high complexity, the KRAB-ZNF family has not been studied in sufficient detail, and the involvement of its members in carcinogenesis remains mostly unexplored. In this study, we aimed to provide a comprehensive description of cancer-associated KRAB-ZNFs using publicly available The Cancer Genome Atlas pan-cancer datasets. We analyzed 6727 tumor and normal tissue samples from 16 cancer types. Here, we showed that a small but distinctive cluster of 16 KRAB-ZNFs is commonly upregulated across multiple cancer cohorts in comparison to normal samples. We confirmed these observations in the independent panels of lung and breast cancer cell lines and tissues. This upregulation was also observed for most of the KRAB-ZNF splicing variants, whose expression is simultaneously upregulated in tumors compared to normal tissues. Finally, by analyzing the clinicopathological data for breast and lung cancers, we demonstrated that the expression of cancer-associated KRAB-ZNFs correlates with patient survival, tumor histology, and molecular subtyping. Altogether, our study allowed the identification and characterization of KRAB-ZNF factors that may have an essential function in cancer biology and thus potential to become novel oncologic biomarkers and treatment targets.
BibTeX
@inproceedings{machnik2018expression,
  author = {Marta Machnik and Rafał Cylwa and Kornel Kiełczewski and Przemysław Biecek and Triantafillos Liloglou and Andrzej Maćkiewicz and Urszula Oleksiewicz},
  title = {{The expression signature of cancer‐associated <scp>KRAB</scp>‐<scp>ZNF</scp> factors identified in <scp>TCGA</scp> pan‐cancer transcriptomic data}},
  booktitle = {Molecular},
  year = {2018},
  doi = {10.1002/1878-0261.12407},
}
Figure