Different boundaries characterize isolated and non-isolated 5q deletions in myelodysplastic syndromes and acute myeloid leukemias

Chromosome 5 long arm deletions, del(5q), are heterogeneous cytogenetic changes. An isolated 5q interstitial deletion is typically associated with erythroid hypoplasia, hypolobated micromegakaryocytes, less than 5% bone marrow blasts, macrocytic anemia, mild leukopenia, normal or increased thrombocytes, female preponderance and a relatively good prognosis.1 In contrast, a 5q loss in a complex karyotype is associated with high-risk myelodysplastic syndromes (MDS) and acute myeloid leukemias (AML), including cases arising after radio- and/or chemotherapy.2 Much emphasis has been given to common deleted regions (CDRs) and genes.3 We recently showed that telomeric breakpoints were significantly different in cases with isolated and non-isolated del(5q). Non-isolated del(5q) showed NPM1/5q35.1 monoallelic loss in 38 of 84 cases (45.2%) versus one of 43 (2.3%) of isolated del(5q) (Fisher’s exact test P<0.001)4. Moreover, gross chromosomal anomalies and monosomies, as seen in high-risk MDS/AML phenotype, were significantly related to NPM1 haploinsufficiency.4

In the present study, we extended our deletion mapping analysis to 170 cases, including a new prospective series of 64 patients with MDS/AML; namely 24 cases with isolated del(5q) and 40 cases with non-isolated del(5q). In the last group, FISH with clone RP11-117L6/NPM1 (5q35.1) detected NPM1 monoallelic deletion in 15 of 40 (37.5%) cases versus one of 24 (4.1%) cases with isolated del(5q) (Fisher’s exact test P=0.002).

Twenty-four cases with complex karyotypes and monosomy 5 were also included after FISH correction (Vysis LSI EGR1/D5S23,D5S721 Dual Color Probe and/or LSI CSF1R/D5S23,D5S721; Chromosome 1, 5, 19, α-satellite, Qbiogene) which demonstrated two copies of 5p15.2 and/or of chromosome 5 centromere, thus indicating that only 5q genomic material was missing.

Centromeric breakpoints were narrowed with 24 genomic clones mapping from 5q11 to q14 (Table 1). 5q11: RP11-648C7, RP11-317O24, RP11-97O9, RP11-640L3, RP11-20I7, RP11-528L24, RP11-175M2, RP11-479O16, RP11-266N13. 5q12: RP11-489L13, RP11-298P6, RP11-79C20. 5q13: RP11-170N5, RP11-195G20, RP11-195E2, RP11-771B3, RP11-79P5, RP11-229C3, RP11-469J18, RP11-996M9. 5q14: RP11-168A11, RP11-80K5, RP11-1089B2, RP11-885P10. Telomeric breakpoints distally to NPM1 were investigated with the following 3 genomic clones: 5q35.2-q35.3: RP11-286C20, RP11-549A4. 5q35.3: RP1-240G13 (subtelomeric probe) (Table 1). DNA clones were selected from http://www.ncbi.nlm.nih.gov, http://genome.ucsc.edu/ and http://projects.tcag.ca/variation/ and were kindly provided by Mariano Rocchi (Department of Genetics and Microbiology, University of Bari, Italy). Furthermore, SNP-aCGH (Affymetrix Cytogenetics Whole-Genome 2.7M Array) data were available in 2 cases.

Figure 1.Characterization of centromeric breakpoints in MDS/AML with del(5q). (A) A proximal breakpoint centromeric to RP11-80K5 was found in none of the 62 patients with isolated del(5q) (a), in 17 of 65 (26.1%) cases of non-isolated del(5q) NPM1+/+ (b) and in 29 of 43 (67.4%) cases of non-isolated del(5q) NPM1+/− (c). (B) Distribution of breakpoints centromeric to RP11-80K5 (indicated in bold) in cases maintaining (●) or losing (○) the NPM1 gene. (C) SNP-aCGH (Affymetrix Cytogenetics Whole-Genome 2.7M Array) analysis in 2 patients with isolated del(5q). Arrows indicate start (→) and end (←) of deletions.

On the centromeric side, the RP11-80K5 clone emerged as an interesting tool to differentiate simple and complex karyotypes by FISH. In fact, breakpoints fell distally to RP11-80K5 (band 5q14.1) in all the 62 cases with isolated del(5q) (24 from the prospective study and 38 from the previous series; Figure 1A, a) including the 2 cases with loss of the NPM1 gene. In this study, centromeric breakpoints were evaluated in 26 cases with isolated del(5q). Results showed a cluster in a ~1.9 megabases region between RP11-80K5 and RP11-1089B2 in 6 of 26 (23%). Within this region, SNPs further refined the molecular heterogeneity (Figure 1C). All the remaining breakpoints (76%) were located distal to RP11-885P10. Interestingly, in SNP studies, Gondek et al. showed similar results in 9 cases with isolated del(5q) at karyotype.5 As far as we know, only one case with isolated del(5q) and a centromeric breakpoint above RP11-80K5, at 5q12-q13, has been reported by Royer-Pokora and co-workers.6

Table 1.Mapping of FISH clones used to characterize centromeric and telomeric boundaries of 5q deletions.

In contrast, in non-isolated del(5q), centromeric breakpoints were significantly more frequently proximal to RP11-80K5 (46 of 108, 42.5%, vs. no cases in isolated del(5q) group; Fisher’s exact test: P<0.001). However, depending on concomitant loss of NPM1 at the telomeric side, they spread differently. Breaks were upstream RP11-80K5 in 17 of 65 (26.1%) NPM1+/+ and in 29 of 43 (67.4%) NPM+/− cases (Fisher’s exact test P<0.001) (Figure 1A, b and c). Notably, from band q11.2 to q14.1, four clusters of breakpoints (each found in ≥4 cases) emerged: two at band 5q11.2 (between RP11-20I7 and RP11-528L24, ~1.9 megabases and between RP11-175M2 and RP11-479O16, ~2.7 megabases), one at band 5q12-q13 (between RP11-79C20 and RP11-170N5 ~2.7 megabases), and one within band 5q13 (between RP11-229C3 and RP11-469J18, ~1.4 megabases) (Figure 1B).

The most telomeric boundaries were defined in 31 of 43 NPM1+/− cases. In 5 of 31 (16.1%) cases, breakpoints fell between RP11-117L6 (NPM1/5q35.1) and RP11-286C20/5q35.2-q35.3 (~5.7 megabases), while subtelomeric sequences were lost in 26 of 31 (83.8%) cases suggesting that, in the large majority of cases, NPM1 deletion may reflect unbalanced 5q rearrangements.

In conclusion, we demonstrated that, in MDS/AML, the del(5q) boundaries are strongly correlated to karyotype complexity. Indeed, in our cohort of patients, and in previously published studies,5 large 5q deletions involving both the RP11-80K5 clone and the NPM1 gene have been found only in non-isolated 5q deletions (approx. 26% of cases in this series). Altogether these results suggest that the identification of the number and type of haploinsufficient genes outside the CDRs might contribute to our understanding of the clinical phenotype associated with different 5q deletions in good or poor risk MDS/AML.

Acknowledgments

The authors wish to thank Dr Geraldine Boyd for her assistance in the preparation of the manuscript. This study was partly supported by AIRC (Associazione Italiana Ricerca sul Cancro), IG-11512; MIUR (Ministero per l'Istruzione, l'Università e la Ricerca Scientifica); Fondazione Cassa di Risparmio, Perugia (Grant n. 2011.0159.021); AULL (Associazione Umbra per la Lotta Contro le Leucemie), Perugia; IAP (Interuniversity Attraction Poles, University of Leuven), Belgium, “Associazione Sergio Luciani”, Fabriano; Ricerca finalizzata 2008, regione Umbria, Italy. RP11 clones belong to the Roswell Park Cancer Institute libraries (http://www.chori.org/BAC-PAC) and were kindly provided by Dr Mariano Rocchi (DAPEG, University of Bari, Italy).

Footnotes

• The information provided by the authors about contributions from persons listed as authors and in acknowledgments is available with the full text of this paper at www.haematologica.org.
• Financial and other disclosures provided by the authors using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are also available at www.haematologica.org.