The Cancer Genome Atlas or TCGA, is a comprehensive and collaborative tumor profiling effort spearheaded by the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI). This effort is designed to accelerate our understanding of the molecular basis of cancer through the application of genome analysis technologies. In the same way that the voyages and expeditions of old opened up new routes for trade in previously uncharted territories, every publication resulting from this tumor profiling effort opens up new avenues for understanding the underlying complexity of cancer.
Cancer is a complex disease resulting from the dysregulation of multiple pathways that govern cell growth and survival. Each of these pathways, acting alone or in concert, contribute variably to the overall disease manifestation, from the initiation of tumorigenesis to the aggressiveness with which it spreads. This variability depends on the collection of variants, ranging from nucleotide variants (SNPs or small indels), larger structural variants, or copy number alterations, present within the tumors. The presence of both large-scale chromosomal abnormalities and smaller changes at the nucleotide level highlight the inherent complexity of the cancer genome as well as the importance of integrating various data types in profiling these tumors and defining the molecular basis of various cancers—a major objective of the TCGA effort.
With the central dogma of molecular biology as the guiding concept, somatic mutations are catalogued from the nucleotide level of tumor-normal pairs using whole genome sequencing or exome studies, validated on the RNA level by assessing gene expression changes as well as changes in methylation patterns that affect gene regulation through RNA sequencing (RNA-Seq) and methylation arrays, respectively, and confirmed on the protein level as well. By overlaying these pieces of information generated by different sequencing techniques, a more complete picture of the disease emerges leading to more specific insights for targeted therapeutics.
Last month, two such tumor profiling studies were published. In Nature, The Cancer Genome Atlas Network reported on the “Integrated Genomic Characterization of Endometrial Carcinomas” , and in the New England Journal of Medicine, they reported on the “Genomic and Epigenomic Landscapes of Adult De Novo Acute Myeloid Leukemia”. Both papers resulted from integrated molecular characterization efforts on tumors which catalogued the various changes within the samples from 1) the DNA level, using whole-genome, exome, and methylation assays, 2) RNA level, using RNA-Seq and, in the case of endometrial carcinomas, 3) protein level using reverse-phase protein array (RPPA).
The first study targeted both type I and type II endometrial carcinomas. Endometrial cancer is the most common gynecological cancer in the US with approximately 50,000 cases diagnosed each year. Despite the high frequency of early detection and the good chances for successful treatment, there is still a 16% mortality rate. Type I endometrioid tumors have been linked to excess estrogen, obesity, and hormone receptor expression and are usually treated with adjuvant radiotherapy. Type II tumors on the other hand, are primarily serous, have been linked with old age and non-obese women and are typically treated with chemotherapy. Integrated multiplatform tumor profiling analysis provided key molecular insights which facilitated the classification of endometrial carcinomas into four new genomic-based categories, each with its own molecular signature:
- POLE ultramutated: high frequency of mutations in POLE, which encodes the catalytic domain of DNA polymerase epsilon involved in nuclear DNA replication and repair
- MSI: characterized by microsatellite instability and high mutation frequency
- MSS/copy number low: microsatellite stab le with low frequency of copy number alterations and high frequency of mutations in CTNNB1 which encodes B-catenin, a key mediator of the Wnt signaling pathway.
- MSS/copy number high: microsatellite stable with high frequency of copy number alterations and low frequency of mutations.
Upon looking at the composition of tumors that fall within the respective categories, the first three categories consist mainly of endometrioid tumors. Interestingly, the fourth category consists mostly of uterine serous cases and 25% of the grade 3 endometrioid cases included in the study, subtypes which are currently treated with different approaches. This study therefore shifts the paradigm by highlighting the overlapping molecular profile between a fraction of grade 3 endometrioid cases and the uterine serous cases, which also shares genomic features with ovarian serous and basal-like breast carcinomas. This result suggests the potential for improved management of these select endometrioid cases with chemotherapy instead of adjuvant therapy. Furthermore, the distinct separation in molecular features between endometrioid and uterine serous carcinomas suggest that these tumor types merit different therapeutic approaches to improve patient outcome highlighting the importance of genomic-based classification.
In a manner similar to the study on endometrial carcinoma, the TCGA Network performed a detailed molecular tumor profiling study of AML. Unlike most cancers, AML genomes have low mutational burden. The most common abnormalities are structural in nature, potentially due to the limitations of cytogenetic techniques in identifying single nucleotide variations. The group cites 50% of patients with normal karyotypes and lacking structural aberrations highlighting the need for more detailed characterization to enable better classification of tumors, the associated risks, and enable targeted modes of treatment. The group performed whole-genome and exome sequencing and combined this data with mRNA and miRNA sequencing, as well as methylation analysis to enable a more comprehensive understanding of the key players in AML pathogenesis. The integrated analysis of these data types facilitated the identification of nine categories based on the tumors’ unique molecular profiles, which strongly correlate with specific mRNA and miRNA expression and methylation signatures.
1) Transcription factor fusions which correlated with specific mRNA expression patterns; in some cases, specific miRNA and methylation signatures as well
2) Mutated NPM1 which has multiple functions among which is maintaining genomic stability and facilitating DNA repair through its interaction with p53
3) Mutations within tumor suppressor genes
4) Mutations within DNA methylation-related genes
5) Changes that activate signaling genes
6) Changes within chromatin modifying genes
7) Changes within myeloid transcription factor genes
8) Changes within cohesion complex genes
9) Changes within spliceosome genes
Multiple genomic-based AML categories associated with unique molecular profiles highlight the previously unappreciated complexity of the disease. Given this, proper subtype classification of tumors, especially for the majority of AML cases that lack structural aberrations, becomes very crucial due to the differences in molecular signatures, thereby impacting the prediction of disease progression as well as selection of appropriate therapy.
Each comprehensive tumor profiling study from the TCGA Network charts new and unique molecular pathways that distinguish one tumor type from another. These pathways help physicians and researchers navigate the complex landscape of cancer by highlighting distinct signatures that act as molecular landmarks. These aid in identifying potential targets, driving the practice of precision medicine and leading to better patient outcomes and an improved quality of life.
Cancer Gene Expression Changes
Epigenetic Changes in Cancer
Cancer Exome Sequencing
Targeted Cancer Sequencing