What lies beyond del(5q) in myelodysplastic syndrome?

Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal hematopoietic stem cell malignancies characterized by ineffective differentiation of one or more bone marrow cell lineages. Much of the phenotypic variability is likely explained by the diverse set of genetic abnormalities responsible for the development and progression of these disorders. However, current clinical decision-making for MDS is based on diagnostic and prognostic criteria that do not include any molecular genetic information. In fact, the only subtype of MDS to be defined by a genetic abnormality is the group with isolated deletion of chromosome 5q [del(5q)].1 Just under 50% of patients with de novo MDS will be found to have cytogenetic abnormalities, of which del(5q) is the most common. In the 10% of cases with the del(5q) as a sole abnormality, this lesion is associated with a more favorable prognosis.2 In another 5–10% of cases, del(5q) is found as part of a complex karyotype (3 or more cytogenetic abnormalities): the prognosis in these cases is poorer.

Substantial effort has gone into understanding how the del(5q) abnormality contributes to the pathogenesis of MDS.3 Early studies tried and failed to find recurrently mutated genes in the remaining intact genes that mapped to the commonly deleted region of chromosome 5. Instead, it was discovered that haploinsufficiency of several genes located in this region were capable of generating the clinical phenotype seen in patients with MDS. Loss of one RPS14 allele for example, can recapitulate the dyserythropoiesis seen in MDS patients with del(5q). Loss of one copy of the microRNA miR-145 and miR146 may confer a clonal advantage and contribute to preserved or increased platelet counts observed in patients with the 5q-minus syndrome.4 Haploinsufficiency of several other genes of the commonly deleted region, including HSPA9, CTNNA1, and EGR1, may also cooperate to promote the development of disease. Finally, loss of the more proximal APC gene and the more distal NPM1 gene may play a role in higher risk MDS cases with more adverse outcomes since this group tends to have larger 5q deletions that extend well beyond the commonly deleted region.3

Identification of these disease-related haploinsufficient genes has informed our biological understanding of del(5q) MDS. Loss of ribosomal protein genes, such as RPS14, has been shown to increase levels of p53, primarily in erythro-blasts, and promote their apoptosis – both elements observed in patients with isolated del(5q) MDS. If p53 activity checks the growth of more primitive del(5q) disease cells, there may be a selective pressure for them to mutate or lose the TP53 gene. In fact, the del(5q) abnormality and TP53 mutations have been found to co-exist more often than their base occurrence rates alone would predict, indicating a likely synergy between these lesions.5 More importantly, this is not restricted to those patients with complex karyotypes in whom TP53 mutations are most common. Even patients with isolated del(5q) and presumed lower risk MDS appear more likely to harbor a concurrent, and often subclonal, TP53 mutation. When this occurs, patients may have poorer response to therapy and a higher than predicted risk of transformation to acute myeloid leukemia (AML).6

This variable relationship between del(5q) and somatic mutations highlights how complex the molecular pathophysiology of MDS actually is. Patients that we group together based on a shared cytogenetic finding may actually have little in common if we were to examine the full range of molecular abnormalities they contain. We now know that recurrent mutations in over 40 genes can occur in patients with MDS in a wide variety of combinatorial and subclonal relationships.7 Understanding how these mutations cooperate could provide mechanistic insight into the pathophysiology of MDS and help us better classify, risk stratify and, treat these patients.

In their article, Fernandez-Mercado et al. describe the application of targeted next-generation sequencing in patients with MDS and del(5q).8 They designed a panel of 25 frequently mutated myeloid malignancy genes (ASXL1, ATRX, CBL, CBLB, CBLC, DNMT3A, ETV6, EZH2, FLT3, IDH1, IDH2, JAK2, KIT, MPL, NPM1, NRAS, PDGFRA, RUNX1, SF3B1, SRSF2, TET2, TP53, U2AF1, WT1, ZRSR2), covering a total of 46 kilobases.

Figure 1.New dimension to the spectrum of genetic abnormalities identified in del(5q) MDS and their impact on disease risk. Increasing chromosomal complexity is recognized as an adverse risk factor in del(5q) MDS. Fernandez-Mercado et al.8 demonstrate the molecular genetic heterogeneity of this group which include additional mutations in many genes, including TP53 which has been associated with a very poor prognosis. This may lead us to reconsider how we assess MDS patients we previously put into discrete risk categories based on cytogenetics alone.

This panel was used to study samples from 43 patients with del(5q) MDS [22 with del(5q) syndrome, 9 with refractory anemia and del(5q) plus additional alterations, 11 with refractory anemia with excess of blasts and 1 with chronic myelomonocytic leukemia in transformation]. Thirty-three patients were also studied by single nucleotide polymorphism arrays. In total 29 non-synonymous alterations, distributed over ten genes were found. The most frequently mutated genes were TP53 (7 cases), ASXL1 (6 cases), and TET2 (5 cases). Only six mutations were detected at a frequency less than 20%, which may have been missed by traditional Sanger sequencing.

The first conclusion that can be drawn from these results is that the spectrum of mutations in patients with del(5q) abnormalities is not dramatically different from that in patients with other forms of MDS.7,9 This is not surprising given data that del(5q) is not necessarily a primary pathogenic abnormality and may instead be acquired after other disease-initiating mutations.5 The second finding is that patients with more clinically advanced disease tend to have a greater number of mutations, a pattern that is seen in patients without del(5q) as well.

Interestingly, TP53 mutations were particularly concentrated in del(5q) MDS patients with additional chromosomal abnormalities [4.5% for del(5q) alone versus 41.7% for complex karyotype with del(5q)]. TP53 mutations have been described in patients with MDS and AML patients with complex karyotypes that often include del(5q). However, they have also been identified in as many as 17% of patients with low-risk MDS with isolated del(5q) in whom they have been associated with resistance or relapse during lenalidomide treatment.10,11 This highly adverse mutation may partially explain why patients with multiple chromosomal abnormalities have a worse prognosis and lower likelihood of response to lenalidomide, but may also help to refine risk prediction in patients presumed to have lower risk disease.

MDS risk stratification is routinely performed with models such as the International Prognostic Scoring System-Revised (IPSS-R)12 which considers cytogenetic abnormalities, percentage of bone marrow blasts, and the severity of cytopenias as risk factors. Nevertheless, the IPSS-R does not include molecular genetic criteria. More than half of MDS patients present with a normal karyotype when analyzed with conventional G-banding cytogenetics and two-thirds fall into the ‘good risk’ cytogenetic category which includes those with isolated del(5q). For these cases, new biomarkers are needed to refine the risk stratification.13

Next-generation sequencing platforms sequence many DNA strands in parallel enabling greater coverage at less cost while providing quantitative information about mutation abundance. This allows for sequencing of many more genes, including those that make rare contributions to a particular phenotype.14,15 The greater sensitivity and digital nature of next-generation sequencing also provide information about the clonal architecture of the disease. In their study Fernandez-Mercado et al. assessed clonal architecture according to the proportion of mutant sequencing reads identified. They assume that mutations only occur once during clonal evolution and therefore they can establish mutation timing. For example, relative clonality was established between a DNMT3A mutation (44% abundance) and a presumably subclonal JAK2 mutation (7%). In another case, 80% of cells appeared to carry an SF3B1 mutation, 20% an ASXL1 mutation, and 10% a TET2 mutation. This highlights the potential complexity associated with interpreting molecular genetic tests. Additional studies will be needed to determine whether subclonal complexity and order of acquisition affect clinically meaningful endpoints. However, even low abundance clones defined by adverse mutations such as those in TP53 may be harbingers of aggressive disease and important to detect at the time of diagnosis. This is particularly true in patients perceived as having lower risk disease or if treatments actually select for the growth of the adverse clone.

In summary, the report by Fernandez-Mercado et al. represents an important, focused look into the genetic makeup of MDS with del(5q).8 More will be learned as additional patients are examined and as more agnostic approaches, such as exome and whole genome techniques, are applied. Next-generation sequencing is rapidly moving from research laboratories into clinical settings. For MDS patients, the identification of gene mutations in a wide set of genes will provide relevant information for diagnosis, accurate risk stratification, assessment of therapy, development of minimal residual disease strategies, characterization of progression mechanisms and identification of molecular targets. Despite its complexity, we welcome this new era of molecular genetic medicine.

Footnotes

• Vera Adema is a PhD candidate in the laboratory of Dr. Francesc Sole at the Josep Carreras Leukemia Research Institute (Badalona, Spain). Her main focus of study is MDS with del(5q) and other cytogenetic abnormalities.
• Rafael Bejar is an Assistant Professor in Hematology and Oncology at the UCSD Moores Cancer Center (La Jolla, USA). His main area of interest is the molecular genetics of myelodysplastic syndromes.
• Financial and other disclosures provided by the author using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are available with the full text of this paper at www.haematologica.org.