Genetic diagnosis in malignant hemopathies: from cytogenetics to next-generation sequencing

Abstract

Since the first specific chromosomal abnormality was identified in leukemia more than 50 years ago, technology has much evolved, now allowing the deciphering of cancer genomes in ever-greater detail. However, much has still to be learned as we have not yet completely dissected all the genomic aberrations driving the genesis and the evolution of malignant hemopathies. The first techniques that have been developed allowed ‘gross’ chromosomal abnormalities to be identified. They include conventional and molecular cytogenetics and microarray-based techniques. However, these techniques can only reveal part of the problem, as genes can be altered in a number of ways (mutations, methylation and so on). This led to the development of what is now known as next-generation sequencing (NGS). Each method has advantages and limits. At present, no single method can decipher all the mechanisms involved in leukemogenesis. Therefore, in our view, it is unlikely that a particular technique will become the ‘gold standard’.

Keywords::

• cytogenetics
• fusion genes
• malignant hemopathies
• methylation
• next-generation sequencing

Although Boveri, as early as in 1914, suggested a role for chromosomal imbalance in carcinogenesis [1], it was only in 1960 that the first specific chromosomal abnormality was identified in chronic myeloid leukemia, the so-called Philadelphia chromosome [2].

Over the last 50 years, technology has much evolved, now allowing the deciphering of cancer genomes in ever greater detail. This led to the launch of the International Cancer Genome Consortium, whose goal is to coordinate a large number of research projects aimed at elucidating comprehensively the genomic changes (somatic mutations, abnormal expression of genes, epigenetic modifications and so on) present in many types of cancers [3]. Indeed, the genetic landscape of cancer and malignant hemopathies includes mutations in important regulatory genes and epigenetic factors in addition to structural changes such as translocations, deletions and amplifications.

Furthermore, the use of ‘omics’ technologies, such as transcriptomics (gene expression profiling) and methylomics (DNA methylation profiling), has dramatically increased in recent years. They are now considered standard tools in cancer research, but they have also attracted the attention of clinicians who found these methods of investigation not only for better and more accurate diagnosis but also for indicators of prognosis and treatment response. However, much has still to be learned at the genome level as we have not yet completely dissected all the aberrations driving the genesis and the evolution of cancer, notably of malignant hemopathies.

Leukemia is characterized by abnormal proliferation of hematopoietic cells. As in other tumors, the number and complexity of genetic aberrations tend to increase during disease evolution. However, to the contrary of other cancers that are mostly associated with gene mutations, copy number variations (deletions and/or amplifications) and loss of heterozygosity (LOH), malignant hemopathies are also characterized by the generation of fusion genes due to chromosomal translocations, inversions or insertions [4,5].

Therefore, given the limitations of the techniques so far available to study the genome of the malignant hemopathies, one has to resort to several methods, from cytogenetics to whole-genome sequencing to have a better, yet incomplete, picture of the genetic aberrations [6].

Cytogenetics

For a long time, conventional cytogenetics was the standard tool to study chromosomal abnormalities in malignant hemopathies. In the last 40 years, karyotyping has allowed identification of most of the recurrent numerical and structural chromosomal abnormalities. This method is particularly useful in detecting balanced rearrangements (mainly translocations but also insertions and inversions) associated with chimeric (fusion) genes. However, this methodology has limitations. It relies on the presence of dividing cells that can be blocked at the metaphase stage of mitosis, but some disorders have a low mitotic index. Furthermore, although conventional cytogenetics provides a whole scan of the chromosomes, its resolution is low (about 10 Mb). Therefore, some rearrangements, such as small deletions or cryptic translocations, can escape the analysis. Minor clones can also be missed even if a large number of metaphases are karyotyped [6].

In order to overcome some of these difficulties, FISH has been applied to hematological oncology. Different types of DNA probes have been developed, ranging from Locus-specific Identifier probes to Whole Chromosome Paint probes. FISH using Locus-specific Identifier probes has two major advantages over conventional cytogenetics. It does not rely on the presence of dividing cells, but can be performed on cells in interphase, and it detects cryptic deletions and balanced translocations by targeting the gene(s) involved in the deletion or the fusion. However, it does not provide a genome-wide assessment, to the contrary of multiplex-FISH, which uses different Whole Chromosome Paint probes to generate a 24-color karyotype. Although this latter technique can be useful in revealing the constitution of complex karyotypes, it has the same limitations as conventional cytogenetics (dividing cells, low resolution) [6].

Microarray-based techniques

Array comparative genomic hybridization (array CGH or aCGH) is a molecular cytogenetic technique used to compare the genetic material of a test sample to that of a reference sample [7]. It explores all 46 human chromosomes in a single test at a high-resolution scale without the need for culturing cells. It is a method of choice for searching for chromosomal copy number changes, either losses or gains of whole or segmental chromosomes [8]. Copy number changes of 5–10 kb of DNA can be identified. However, a major setback of the technique is its inability to detect balanced chromosomal abnormalities, such as reciprocal translocations or inversions that are common features in malignant hemopathies, and to analyze small populations of malignant cells.

Two other phenomena are also frequent events in malignant hemopathies. These are LOH and the copy number neutral LOH (CN-LOH). LOH occurs when a deleterious mutation takes place in one allele, and the normally functioning allele at the same locus is lost. CN-LOH (called acquired isodisomy or acquired uniparental disomy) consists of the loss of all or part of a chromosome and doubling of the remaining homologous material. Solely, single nucleotide polymorphism array can identify CN-LOH, which cannot be detected by cytogenetics (conventional or FISH) or aCGH [8]. As for aCGH, single nucleotide polymorphism array does not identify balanced chromosomal rearrangements.

Next-generation sequencing

The use of the techniques described above allows ‘gross’ chromosomal abnormalities such as translocations, amplifications and deletions to be identified, and an overwhelming amount of papers have appeared in the literature. However, this is only the tip of the iceberg as genes can be altered in a number of ways (mutations, methylation and so on) that could be critical to the onset and/or progression of malignant hemopathies.

Initially, mutations in single genes were searched for using PCR followed by Sanger sequencing. In the past decade, technological advances have revolutionized our capacities to question the genome at a single base pair resolution. Several techniques, referred to as next-generation sequencing (NGS), have been developed [9,10].

Whole-exome sequencing (WES) was designed to selectively sequence the coding regions of the genome [11]. As the exons constitute about 1% of the human genome, this method has a reduced cost compared with other NGS methods, but it cannot identify mutations outside the coding regions (which represent 99% of the human genome), nor chromosomal abnormalities, such as translocations and inversions, with breakpoints located in the introns. Therefore, the question has been raised whether WES was a transient technology in the landscape of NGS [12].

Whole-genome sequencing can overcome these difficulties in that it covers the entire genome, allowing not only gene mutations but also structural abnormalities, such as deletions, amplifications and translocations, to be detected [10]. Although whole-genome sequencing appears to be the method of choice for mutation detection, its major drawbacks are the high cost, the large amount of data generated and the complexity of the analysis. This complexity is even more increased as the reliability of the technique is dependent on the number of hits, which expands the information to deal with. Furthermore, it does not provide information about alterations in gene expression and epigenetic modifications.

Therefore, an alternative method is transcriptome sequencing (also known as RNA sequencing or RNA-seq). Both coding and noncoding RNAs are explored. The interest of this technique is that it allows post-transcriptional changes and fusion transcripts to be analyzed [13,14]. However, as for WES, it cannot detect mutations in noncoding regions of the genome. Furthermore, mutation discovery remains difficult due to mapping complexities across intron–exon boundaries and splice variation. Finally, it remains computationally more complex and expensive than other NGS methods. Still, RNA-seq is the best tool for identifying novel fusion events.

The methods described above explore the genome at different levels. However, it is now well-known that epigenetic mechanisms contribute to leukemogenesis. Therefore, genome scale studies using whole-genome bisulfite sequencing to investigate DNA methylation and chromatin immunoprecipitation (ChIP)-Chip or ChIP-seq followed by NGS to analyze chromatin structure are much needed [15–17]. Leukemogenesis is a multistep process involving both genetic and epigenetic mechanisms. NGS is a major advance in elucidating the molecular basis of leukemia, as we can now classify malignant hemopathies at the molecular level, which is superior to a cytological-based classification. Although whole-genome sequencing will undoubtedly become part of the standard diagnostic evaluation within the next few years, at present however, no single method can decipher all the mechanisms involved in leukemogenesis [18]. Therefore, in our view, it is unlikely that a particular technique will become the ‘gold standard’.

Transferring the NGS methods from the research laboratory to the clinical setting could be very challenging [13,19]. The bioinformatics infrastructure and expertise needed to analyze the huge amount of data generated are likely to remain a limiting factor. Furthermore, during their evolution, malignant hemopathies tend to accumulate a number of genetic abnormalities that may not have clinical and/or therapeutic significance. At present, only a small number of genetic abnormalities are used to predict prognosis and orientate therapy. Separating the primary abnormalities from the redundant anomalies, due to the intricate metabolic pathways used by the malignant cells, could be very arduous.

Furthermore, in the near future, it is likely that these techniques would be performed on cells selected according to their immunophenotypic protein expression profile or on circulating malignant cells.

The ultimate purpose is, first, to relate structural changes in genomes of malignant hemopathies to the underlying disease mechanisms and, second, to reveal new opportunities to develop strategies for treatment.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.