SEARCH

SEARCH BY CITATION

Keywords:

  • cancer genomes;
  • human genome;
  • somatic mutations;
  • cancer genetics;
  • pediatric cancer;
  • adolescent cancer;
  • young adult cancer;
  • pharmacogenomics;
  • pharmacogenetics;
  • whole genome sequencing

Abstract

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES

This mini-review describes the rapid changes in genome technologies that are leading to comprehensive views of genetic alterations in cancer, and presents high-level thoughts on ways to accelerate translation into clinical medicine. Issues that are more relevant to children, adolescents, and young adult patients with cancer are highlighted. Cancer 2011;117(10 suppl):2262–7. © 2011 American Cancer Society.

Each cancer patient is distinct with regard to clinical, familial, socioeconomic, environmental, and many other macroscopic factors. At the microscopic level, all cancer patients vary with regard to the DNA inherited from their parents and DNA within their tumors, the latter resulting from acquired (or somatic) mutations because of carcinogens and abnormal DNA repair mechanisms that frequently occur in cancer cells. Although the individuality of patients with cancer has always been recognized, the fundamental uniqueness of each cancer genome, and the importance that this may have in the clinical management of patients, has only been recognized in the latter half of the 20th century, initially as a result of cytogenetics1, 2 and then gradually at the molecular level based on the study of specific genes,3, 4 and today as a result of whole cancer genome sequencing. The implications of this increased understanding of cancer as a complex and heterogenenous disease involving mutations in hundreds or thousands of genes has led to a new paradigm in cancer research and management coined “personalized medicine.” The National Cancer Institute (NCI) in the United States defines personalized medicine as “a form of medicine that uses information about a person's genes, proteins, and environment to prevent, diagnose, and treat disease.” A few examples of personalized medicine have already reached the clinical setting, such as trastuzumab for some forms of breast cancer5 and imatinib for leukemias6, 7 that display specific genetic abnormalities.

Technological advances and large directed efforts allowing comprehensive studies of genomes in cancer patients and/or tumors have accelerated the search for genes and mutations implicated in the etiology of cancer and progression toward aggressive and metastatic disease. In the 1990s, DNA sequencing using Sanger-based chemistry was used on different platforms that evolved from vertical gel electrophoresis with a capacity of approximately 1000 nucleotide bases per day to capillary instruments that yielded over 1 million bases per day. Since the release of the first draft of the human genome sequence,8 the capacity has kept improving at a breathtaking pace, with several instruments currently generating in excess of 10 billion bases per day. These advances have been accompanied by drastic decreases in the cost of human genome sequencing. Current estimates of reagent costs for a whole genome sequence in 2010 are approximately $10,000 US dollars, a > 10,000-fold decrease over approximately 8 years. New technologies suggest that whole genomes will be sequenced at a cost of < $1000 in a few years. Other technologies to study genomes have also evolved rapidly.

Germline Genome Variation and Cancer

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES

Before the release of the first human genome reference sequence in 2001, approximately 30 tumor suppressor genes and 100 dominant oncogenes had been identified.9 Many of these were identified by mapping strategies in families having multiple relatives affected with similar tumors. Notwithstanding notable examples of breast and colon cancer genes implicated in adult-onset cancers (ie, BRCA1, BRCA2, mutS homolog 2 [MSH2], mutL homolog 1 [MLH1], and other DNA repair genes), many of the first tumor suppressors and oncogenes to be identified were involved in childhood cancers (retinoblastoma 1 [RB1] and retinoblastoma, Wilms tumor 1 [WT1] and familial Wilms tumor, etc). Many of the DNA mutations in these cancer genes affect protein structure and are highly penetrant.

In recent years, genome-wide studies of inherited genomes have allowed comprehensive coverage of common polymorphisms that are shared in human populations. The rationale stems from the finding that most patients with cancer (and other common diseases) do not exhibit strong familial histories, nor have detectable mutations in previously characterized cancer genes. The approach of identifying common alleles in human populations predisposing to cancer is based on correlations (or linkage disequilibrium) between nearby genetic variants in the genome. Catalogues of common alleles (most often comprised of single nucleotide polymorphisms [SNPs]) and linkage disequilibrium patterns were generated by the International HapMap Consortium.10 Array-based technologies to rapidly analyze hundreds of thousands of SNPs tagging most of the human genome in thousands of individuals with and without cancer are now used extensively to find new cancer loci. The following examples of genome-wide association studies (GWAS) are provided to illustrate the approach and results implicating novel cancer-predisposing loci.

DNA samples from 317 children with acute lymphoblastic leukemia (ALL) were genotyped using a SNP mapping array. Genotype frequencies were compared at > 300,000 polymorphic sites across the genome with frequencies determined in independent control groups.11 Genotypes for 18 SNPs at 12 gene loci were found to be statistically different (P < 1 × 10−5) between cases and controls. The 18 SNPs had odds ratios (ORs) of 1.43 to 3.62. In the second stage of the study, the frequencies of the 18 SNPs were compared in 4 subtypes of ALL (B-other, B-hyperdiploid, t(12;21)/ETV6-RUNXI, and T-cell ALL). SNP frequencies in AT rich interactive domain 5B (MRF1-like) (ARID5B) were shown to distinguish B-hyperdiploid ALL (61%) from other subtypes (42%) and controls (33%). An independent study using a similar approach to compare 907 ALL cases and 2398 controls identified 3 risk loci for ALL, one of which was ARID5B.12 The results of these recent reports are similar to many other GWAS studies investigating cancer and other common diseases; the majority of alleles confer relatively small increments in risk. The potential for using this information to predict disease in individuals remains limited, although the discovery of many more risk loci may increase the usefulness of germline genomic profiling.13 Notwithstanding their limited use as predictive markers, the biological insight gained from new cancer loci is interesting; for example, the germline differences in ARID5B for different subtypes of cancer suggest a role for this gene in B-lineage development.

Large-scale genotyping of SNPs in candidate genes or GWAS methods can been applied to other important cancer traits, including pharmacogenetics (ie, the study of genetic variation that gives rise to differing responses to or adverse effects of drugs). Ross et al compared genetic variants in children with serious cisplatin ototoxicity with children receiving cisplatin without experiencing hearing loss.14 Nearly 2000 SNPs within or near 299 genes involved in the adsorption, distribution, metabolism, and elimination of drugs were evaluated. SNPs in thiopurine S-methyltransferase (TPMT) and catechol O-methyltransferase (COMT) were associated with hearing loss, with ORs of 17.0 and 5.5, respectively. Given the strong association between the risk alleles and cisplatin ototoxicity, further validation is warranted, because genetic testing of TPMT and COMT variants could be used to inform treatment options that would avoid using cisplatin in children at risk of treatment-induced hearing loss.

Somatic Genome Variation and Cancer

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES

Recent large-scale studies of cancer genomes from different tumor types all revealed tremendous heterogeneity and suggested the existence of a large repertoire of unknown cancer genes. A mutational screen of the complete family of 518 protein kinases in 210 human cancers of multiple types indicated that there exist many unidentified mutated genes in cancer.15 An extensive transcriptome, genome, and epigenome characterization of 209 glioblastomas as well as sequencing of 601 candidate cancer genes in a subset of these tumors revealed that 3 signaling pathways are important in this disease.16 A survey of 623 genes in 188 human lung adenocarcinomas revealed > 1000 somatic mutations.17 An analysis of 20,857 transcripts from 18,191 genes in DNA from 11 breast and 11 colorectal tumors characterized the frequency of somatic mutations as a few commonly mutated gene “mountains” and a much larger number of gene “hills” that are mutated at low frequency.18 Similar findings were observed in pancreatic cancer and glioblastoma.19, 20 With advances in massively parallel sequencing technology, whole cancer genome sequencing has recently been initiated.21-24 The latter projects are impressive not only because of the sheer sizes of the datasets that were generated, but because of the high abundance of somatic mutations detected in each cancer genome, with reports describing > 10,000 acquired mutations in single tumors.

Although the majority of the variants discovered in large cancer genome studies are likely neutral changes that arose in cancer cells having defective DNA repair mechanisms, dozens to hundreds of mutations in each cancer likely confer a biological advantage to the tumor. Although the latter will require validation and functional studies, many genes identified in these projects constitute potential therapeutic targets, including phosphoinositide-3-kinase, catalytic, α polypeptide (PIK3CA); v-raf murine sarcoma viral oncogene homolog B1 (BRAF); neurofibromin 1 (NF1); kinase insert domain receptor (KDR); and phosphoinositide-3-kinase, regulatory subunit 1 (α) (PIK3R1). Additional insight derived from these projects includes correlations between cancer mutations and prognosis, such as isocitrate dehydrogenase 1 (nicotinamide adenine dinucleotide phosphate [NADP]+), soluble (IDH1) and IDH2 mutations in several types of gliomas,20, 25 and extremely high mutation rates observed in 2 glioblastomas previously treated with temozolamide.15, 16 In a lung cancer genome study,17 clinical correlations were found to exist between the number of mutations and clinical grade, mutations in specific genes and risk factors (ie, epidermal growth factor receptor [EGFR] mutations and smoking), and pathological subtypes (ie, negative correlation for LRPB1, tumor protein p53 [TP53], and inhibin, β A [INHBA], with acinar, papillary, and bronchioloalveolar carcinoma subtypes). These preliminary studies that globally evaluate cancer genomes have generated new knowledge that points to challenges as well as new opportunities that will translate into clinical benefits. Furthermore, these reports emphasize that, in addition to the more easily recognizable mutations that affect protein structure, noncoding and nongenic variants may have important roles in the pathogenesis of the tumors, and that sequencing hundreds (if not thousands) of genomes will be necessary to discover and validate all the important mutations in each tumor type. The large number of tumors that need to be studied is because of the high degree of heterogeneity within most tumor types as well as the observation that several driver genes have been observed at low frequencies (< 5%). There appears to be a wide variation in the number of mutations between tumors, from the low numbers observed in some cytogenetically normal acute myeloid leukemia (AML) genomes22 to sarcomas having highly complex patterns of chromosomal abnormalities. The abundance of mutations per tumor is still generally ill-defined and will be better known as more cancer genomes are analyzed.

One can conclude from these (and other) studies that, although we may believe that a great deal is already known concerning the molecular basis of cancer, there is considerably more knowledge to gain by moving toward systematic studies of cancer genomes at the genomic, epigenomic, and transcriptomic levels to reveal the full repertoire of oncogenic mutations and clinically relevant subtypes for prognosis and therapeutic management, and the development of new therapeutics that target specific cancer genome abnormalities.

It is interesting to note that the recent literature describing global analyses of cancer genomes has focused mostly on adult cancers. At the time of preparing this article, no comparable study involving large-scale resequencing of multiple childhood-onset tumors had been published. Less comprehensive surveys of cancer genomes using lower resolution technologies have provided valuable insight regarding cancer genomes in samples obtained from younger patients. There are some intriguing differences in mutation loads noted among tumor types that have been studied. For example, a study of > 200 medulloblastomas evaluated for copy number changes using high-resolution SNP genotyping arrays26 identified numerous amplifications and homozygous deletions, including recurrent events in genes targeting histone lysine methylation. A different study of SNP arrays and candidate gene sequencing of samples from 111 children with de novo AML displayed a very low burden of genomic alterations in most samples, and an absence of any identifiable copy number alterations in 34% of the leukemias.27 The study of more subtypes of pediatric cancers with higher resolution genome technologies will be needed to better understand the spectrum of mutation patterns in cancers affecting young patients, and to detect additional cancer targets that would lead to new treatment strategies.

Systematic Cataloguing of Somatic Mutations in Cancer Genomes

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES

Two pioneer initiatives were launched in the United Kingdom and the United States to catalogue somatic mutations: the Cancer Genome Project (CGP)28 and The Cancer Genome Atlas (TCGA),29 respectively. Valuable lessons obtained from these pilot studies were instrumental in developing the framework for an international consortium. The International Cancer Genome Consortium (ICGC) was formed to coordinate an international research effort to obtain a comprehensive description of genomic, transcriptomic, and epigenomic changes in the major forms of cancer affecting children and adults. Over the next 10 years, the ICGC expects to produce comprehensive catalogues of the full range of genetic mutations in 50 types of cancer, with each study involving specimens of tumor plus normal tissue from approximately 500 patients (unless the tumor type is rare). The ICGC will thus generate data for 25,000 cancer and 25,000 germline genomes datasets having high resolution to capture nucleotide-level alterations.

Several technologies will be used to detect somatic alterations. Whole genome sequencing studies that include high coverage whole genome shotguns of cancer and normal genomes will obviously be more comprehensive than datasets generated through targeted resequencing of the exome (ie, part of the genome formed by exons and other functionally important elements). However, whole genome sequencing is more expensive (currently by a factor of 5 to 10). It is also argued that the scientific community lacks the technologies to efficiently characterize noncoding sequences, which leads to a greater use of the exomic sequence, whether it is derived by whole genome sequencing or targeting strategies. Various other approaches are used to identify structural variants, including high-resolution genotyping or comparative genomic hybridization arrays for chromosomal deletions or amplifications, and paired end sequencing methods that sequence both ends of the same molecule, from which translocations, inversions, and other chromosomal rearrangements can be inferred. Finally, the majority of ICGC projects will generate DNA methylation and expression datasets from the same tumor samples that are analyzed for DNA mutations.

ICGC projects use common standards of data collection and analysis, and have adopted a rapid prepublication data release policy. In April 2010, the ICGC launched its data portal at www.icgc.org, with the release of several cancer genome datasets that are freely available to the global research community.30

As of April 2010, the ICGC had received commitments from funding organizations in Asia, Australia, Europe, and North America for genome analyses of > 10,000 cancers as part of 20 ICGC projects. Of these, only 1 currently targets a pediatric form of cancer. The German Cancer Aid and the German Ministry for Education and Research are supporting an ICGC project regarding pediatric brain cancer that is being led by Professor Peter Lichter from the German Cancer Research Center in Heidelberg. Current plans involve the analysis of 300 medulloblastomas, a common malignant brain tumor in childhood with a 5-year overall survival rate of approximately 60% to 70%, and 300 pediatric pilocytic astrocytomas, the most frequent pediatric brain tumor. Approximately two-thirds of specimens will be collected prospectively and obtained before any radiotherapy or chemotherapy exposure. The low number of projects for childhood and other early onset cancers is a concern to members of the ICGC that is being addressed by further coordination among funders and the engagement of the pediatric cancer research community.

The NCI has launched a significant program to discover targets in cancers of childhood. Named the “Therapeutically Applicable Research to Generate Effective Treatments (TARGET)” initiative, the program is investigating the genome-related changes associated with ALL, AML, neuroblastoma, osteosarcoma, and high-risk Wilms tumor, 5 common types of childhood cancer. Further information concerning the TARGET initiative is available at http://target.cancer.gov/. Finally, researchers at St. Jude Children's Research Hospital and the Washington University School of Medicine recently announced a collaboration that will sequence the genomes of at least 600 pediatric cancers, including leukemias, brain tumors, and sarcomas (available at: http://www.cancer.gov/ncicancerbulletin/012610/).

Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES

One could reasonably conclude that cancer genome studies for pediatric and other early onset cancers are relatively understudied compared with many adult-onset cancers. Some causes for this are: 1) the higher cure rates for several pediatric cancers compared with adult-onset cancers, which may diminish the interest in funding expensive genome studies seeking new therapeutic targets; 2) the rarity of many forms of pediatric cancers, which limits the generation of sample panels that are sufficiently large to provide the statistical power needed to address the multiple hypothesis testing that occurs in large-scale genome research projects; 3) the relative lack of communication and advocacy regarding pediatric cancer; and 4) the increased use of research-related resources toward many previously neglected biomedical and psychosocial needs of long-term survivors.

None of these issues mentioned justifies a gap between how cancer genome research is conducted in early versus adult-onset forms of cancer. Several forms of childhood cancer having poor survival would benefit from the identification of new cancer targets and the development of targeted therapies such as trastuzumab and imatinib. Cancer survivors often experience significant morbidity after the disease is cured, including growth and developmental disorders, and long-term toxicities appearing several years after the discontinuation of initial treatments. Research that leads to the prevention of serious side effects, such as pharmacogenetic markers that would predict ototoxicity to cisplatin, or new targeted agents that would decrease or replace the traditional chemotherapies and radiotherapies used in most childhood cancers could lead to a better quality of life in long-term survivors.

The path from cancer targets to cancer therapies is long and difficult. Experience gained from numerous genomic studies of adult cancer has demonstrated that the discovery of cancer-specific mutations is becoming relatively simple, but understanding which cancer mutations and genes will lead to clinical impact is challenging. Biological validation using high-throughput RNA interference (RNAi) and chemical screens in cell lines and xenograft models will help move from hypothesis-driven approaches to systematic evaluations of numerous potential cancer mutations and genes. The evaluation of pharmacogenetic markers, and markers used to classify tumor subtypes and select treatment options, will require clinical trials in age-appropriate patient populations, even if drugs are first developed and shown to be useful and safe in adult cancers. Given the relative rarity of many forms of early onset cancers, the sufficient sample size will only be possible through national and international networks of cancer centers that have the essential infrastructure (with biospecimen collections and molecular pathology services) needed for conducting such trials.

The promises of cancer genome research are exciting. Successful translation for the benefit of children, adolescents, and young adult patients with cancer and survivors will require a coordinated approach linking basic and clinical research communities in multiple institutions and nations, and tighter communication with cancer patients, advocates, and funders.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES

Funding for the workshop was provided by C17; the Advisory Board of the Institute for Cancer Research at the Canadian Institutes for Health Research (CIHR); the Public Health Agency of Canada; the Ontario Institute for Cancer Research; the Meetings, Planning and Dissemination Grants program of the CIHR; the Terry Fox Research Institute; LIVESTRONG, formerly the Lance Armstrong Foundation; the Canadian Cancer Society Research Institute; Young Adult Cancer Canada; Hope and Cope; and the Comprehensive Cancer Centre at the Hospital for Sick Children, Toronto, in addition to the support provided by the Canadian Partnership Against Cancer to the Task Force on AYA Cancer. This study was supported by the Ontario Institute for Cancer Research and the Ontario Ministry of Research and Innovation.

CONFLICT OF INTEREST DISCLOSURES

The author made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. Germline Genome Variation and Cancer
  4. Somatic Genome Variation and Cancer
  5. Systematic Cataloguing of Somatic Mutations in Cancer Genomes
  6. Challenges and Opportunities to Accelerate Cancer Genome Research in Cancers Affecting Children, Adolescents, and Young Adults
  7. FUNDING SOURCES
  8. REFERENCES