Author Affiliations

  1. Ana Rio-Machín1,*,
  2. Juliane Menezes1,*,
  3. Alba Maiques-Diaz1,
  4. Xabier Agirre2,
  5. Bibiana I. Ferreira1,
  6. Francesco Acquadro1,
  7. Sandra Rodriguez-Perales1,
  8. Karmele A. Juaristi3,
  9. Sara Álvarez1 and
  10. Juan C. Cigudosa1
  1. 1Molecular Cytogenetics Group, Human Cancer Genetics Program, Centro Nacional Investigaciones Oncológicas (CNIO) - Centro de Investigación de Enfermedades Raras (CIBERER), Madrid
  2. 2Division of Oncology, Department of Hematology Service and Area of Cell Therapy, Center for Applied Medical Research (CIMA), University of Navarra, Pamplona, Navarra:
  3. 3Hematology Service, Hospital Alcorcón Foundation, Madrid, Spain
  1. Correspondence: Juan C. Cigudosa, Molecular Cytogenetics Group, Human Cancer Genetics Program, Spanish National Cancer Research Centre – CNIO, Melchor Fernández Almagro, 3, 28029, Madrid, Spain. E-mail: jccigudosa{at}cnio.es
View Abstract

Abstract

The disruption of RUNX1 function is one of the main mechanisms of disease observed in hematopoietic malignancies and the description of novel genetic events that lead to a RUNX1 loss of function has been accelerated with the development of genomic technologies. Here we describe the molecular characterization of a new t(4;21)(q21;q22) in a de novo myelodysplastic syndrome that resulted in the deletion of the RUNX1 gene. We demonstrated by quantitative real-time RT-PCR an almost complete depletion of the expression of the RUNX1 gene in our t(4;21) case compared with CD34+ cells that was independent of mutation or DNA methylation. More importantly, we explored and confirmed the possibility that this abrogation also prevented transactivation of RUNX1 target genes, perhaps confirming the genetic origin of the thrombocytopenia and the myelodysplastic features observed in our patient, and certainly mimicking what has been observed in the presence of the RUNX1/ETO fusion protein.

Introduction

The human RUNX1 gene (also known as AML1, CBFA2 or PEBP2), located in the 21q22 chromosomal band, encodes for one of the two subunits forming the heterodimeric transcription factor complex, human core-binding factor (CBF), a key regulator in the hematopoietic system.1 This transcription factor is one of the genes most frequently deregulated in hematologic malignancies mainly through point mutations and chromosomal translocations. RUNX1 mutations that lead to a RUNX1 haploinsufficiency or loss of function are found as germ line heterozygous mutations in familial platelet disorder2 and familial myelodysplastic syndromes (MDS), and as sporadic point mutations in about 20% of acquired MDS patients,3 both diseases predisposing to acute myeloid leukemia (AML). On the other hand, the most frequent chromosome translocation observed in AML (10–20%) is the t(8;21) which generates a fusion protein formed by the N-terminal portion of RUNX1 with the C-terminal portion of the heterologous partner protein ETO. This and other variant fusion proteins involving RUNX1 always lose their transactivation potential, resulting in a disruption of the normal hematopoietic differentiation program secondary to a wild-type RUNX1 haploinsufficiency.4 On a clinical level, RUNX1 mutations are more likely to occur in patients with secondary exposure-related MDS (t-MDS) due to a mutagenic effect of the chemotherapy or ionizing radiation and its association with a worse prognosis has been established.3 In summary, the disruption of RUNX1 function is one of the main mechanisms of disease observed in hematopoietic malignancies and the development of genomic technologies has allowed novel genetic events that lead to a RUNX1 loss of function to be identified and described.

We present the genomic characterization of a unique case with an unbalanced t(4;21)(q21;q22) that resulted in the heterozygous loss of the RUNX1 gene in a male patient who showed refractory cytopenia with multilineage dysplasia (RCMD), characterized by a severe thrombocytopenia.

Design and Methods

Patient sample

A bone marrow (BM) sample was collected from the patient for diagnostic purposes after obtaining informed consent. The study was approved by the ethical research and animal care committee from the Carlos III Institute of Health (CEi PI 32_2009).

Cytogenetic and FISH assay

BM cells obtained at diagnosis were cultivated for 24 h without mitogenic stimulation and processed according to standard protocols.5 GTG-banding was performed and 20 metaphases were analyzed. To evaluate a possible rearrangement of the RUNX1 gene in our case, we designed a fluorescent in situ hybridization (FISH) break-apart probe flanking the gene: bacterial artificial chromosomes (BACs) mapping 3′ of the RUNX1 locus (RP11-8P19, RP11-639G23 and RP11-272A03) were labeled using Spectrum Green d-UTP, while the 5′ BACs (RP11-154L23, RP11-396G11 and RP11-768B16) were labeled with Spectrum Orange d-UTP. FISH assays were performed with standard methods.5 After hybridization, the resulting patterns for a single cell should be: two fusion (2F) signals corresponding to normal RUNX1 or one red, one green, and one fusion (1R1G1F) signal if the cells are affected by a RUNX1 rearrangement.

Comparative genome hybridization array

The DNA sample was hybridized against human genome CGH 44K microarrays (Agilent Technologies, Palo Alto, CA, USA) spanning the entire human genome at a median resolution of ~75Kb. Hybridizations and all data analysis were carried out according to the manufacturer’s protocols. The slide was scanned with an Agilent scanner and data were analyzed with Agilent Feature Extraction and CGH Analytic software 3.5.14.

RUNX1 screening

For RUNX1 mutation screening by reverse-transcription poly-merase chain reaction (RT-PCR) and direct sequencing, the following forward/reverse (F/R) overlapping primer pairs were used: 1F/R (5’– GATGCGTATCCCCGTAGATG/5’-AGCACG-GAGCAGAGGAAGT), 2 F/R (5’–AGCATGGTGGAGGT-GCTG/5’-CAGCCATCACAGTGACCAGA), 3 F/R (5’– CCC-TAGGGGATGTTCCAGAT/5’-TGAAGCTTTTCCCTCTTC-CA), 4 F/R (5’ GGCTGGCAATGATGAAAACT/5’-CCGAT-GTCTTCGAGGTTCTC), 5 F/R (5’– CACCTACCACAGAGC-CATCA/5’-GCTGAGGGTTAAAGGCAGTG), 6 F/R (5’–AACC-CTCAGCCTCAGAGTCA/5’-TTCTGCAGAGAGGGTTGT-CA), 7 F/R (5’– CCAATACCTGGGATCCATTG/5’-GTGAAG-GCGCCTGGATAGT), 8 F/R (5’– TCCATTGCCTCTCCTTCT-GT/5’-CACCATGGAGAACTGGTAGGA) and 9 F/R (5’–AGCTCGCCCTCCTACCAC/5’-CCTCAGTAGGGCCTCCA-CAC). We used the following PCR conditions: denaturation at 95ºC for 5 min; 34 subsequent amplification cycles performed at 95ºC for 30 sec, at 58ºC for 2 min, and at 72ºC for 1 min; a final step was performed at 72ºC for 10 min. PCR products were purified with QIAquick PCR purification kit (Qiagen, Hilden, Germany) and sequenced. Both F and R strands were analyzed by BLAST (http://www.ncbi.nlm.nih.gov/BLAST/).

Quantitative real-time PCR

Hematopoietic stem cells were purified from cord blood samples and from the bone marrow of the patient using CD34+ immunomagnetic microbeads (CD34 MicroBead Kit human, MACS Miltenyi Biotec, Cologne, Germany). RUNX1 isoform b and c expression was evaluated in a single assay by quantitative real-time RT-PCR starting with total RNA that was reverse-transcribed using random hexamer primers with the TaqMan® Gold RT-PCR kit (Applied Biosystems, Foster City, CA, USA). The calibrated normalized relative quantity, taking into account target-and run-specific amplification efficiencies, was calculated using endogenous GAPDH expression with the Qbase software application. The same approach was used to explore the level of expression of some other genes that were previously found down-regulated by the presence of the RUNX1/ETO fusion (YES1, MLLT3, CTCF and MAPK1),6 both in our patient and in Kasumi-1 (a cell line carrying the RUNX1/ETO fusion gene).

Results and Discussion

This 57-year old male patient showed refractory cytopenia with multilineage dysplasia (RCMD), characterized by a severe thrombocytopenia. BM aspirate and core biopsy showed 4% of blasts with dysplastic features in over 10% on all myeloid lineages. His International Prognostic Score System was 0.5, classifying him as an intermediate risk-1 MDS patient. Twenty-four months after diagnosis, the patient (by that time with a normal karyotype: 46,XY) underwent allogeneic BM transplantation. Unfortunately, he died on Day +30 due to graft-versus-host disease.

GTG-banding chromosome analysis revealed, in addition to normal metaphases, the presence of a cellular clone showing an apparently balanced translocation, 46,XY,t(4;21)(q21;q22) in 17 out of 20 analyzed metaphases (Figure 1A). This translocation occurred only in the BM cells and it was not observed in the constitutional karyotype of the patient obtained after a conventional peripheral blood cytogenetic analysis. Interphase FISH showed 85% nuclei with an unexpected abnormal pattern: one fusion signal (red/green) plus one small green signal. On metaphases, the fusion signal mapped to the normal copy of the gene on chromosome 21, and the green signal mapped to the rearranged RUNX1 locus on the derivative chromosome 21 of the t(4;21)(q21;q22) translocation. The reciprocal red signal (expected on the derivate chromosome 4) was missing (Figure 1B). These results indicated that our patient had a 21q22 unbalanced chromosomal rearrangement with a breakpoint centromeric to the RUNX1 gene and with a concomitant loss of, at least, the 5′RUNX1 region.

To better define the nature of DNA copy number variations (CNV) in the sample, we performed high-resolution array CGH. We found a mono-allelic loss of 6.6Mb at chromosome 4q21.21 and a mono-allelic loss of 2.3Mb at chromosome 21q22.12 (Figure 1C). The genes included on these regions are listed in the Online Supplementary Table S1. Among the 32 genes deleted at chromosome 4q21.21, we could not identify relevant biological information related to hematologic disease. On the other hand, RUNX1 was among the 12 deleted genes at the 21q22.12 breakpoint. To the best of our knowledge, this is the first report in the literature of an MDS case with an isolated reciprocal translocation, t(4;21)(q21;q22), that results in a RUNX1 deletion. So far only a few RUNX1 deletions have been described in MDS.7

Since one of the copies of RUNX1 was lost as a consequence of the unbalanced translocation, we explored its expression in the CD34+ BM cells of the patient by quantitative real-time RT-PCR. We observed an almost complete depletion of the expression of RUNX1 gene in our t(4;21) case compared with CD34+ cells (Figure 1D). To explain this lack of expression, we examined whether RUNX1 had been inactivated through point mutations, since it has been reported to be an important pathogenic mechanism in AML and MDS. We amplified and sequenced all RUNX1 exons in one single PCR and found no mutations in the retained RUNX1 allele. We then decided to explore epigenetic inactivation by assessing the methylation status of its promoter region. Although transcription of RUNX1 may be initiated at two distinct 5’ promoter regions (distal and proximal),8 only the proximal region has a significant CpG island. This was further studied by Methylation Specific PCR (MSP) and pyrosequencing.9 We also ruled out this molecular mechanism as an alternative explanation for the absence of RUNX1 expression (Online Supplementary Figure S1).

Functional in vivo studies have shown that the fusion RUNX1/ETO acts mainly as a dominant-negative suppressor of the wild-type RUNX1, with a similar phenotype as that observed in the RUNX1 knockout model. To evaluate this hypothesis and the putative leukemic role of the RUNX1 abrogation in the context of our patient, we decided to explore the level of expression of some other genes that had previously been found to be down-regulated by the presence of the RUNX1/ETO fusion.6 In detail, we used quantitative real-time RT-PCR to study the expression of 4 genes (YES1, MLLT3, CTCF and MAPK1) and we observed a significant downregulation in all of them, both in our patient and in Kasumi-1 (Figure 1D). Taken together, we observed that these 5 genes, which are undoubtedly involved in cell differentiation, are significantly down-regulated, indicating that the abrogation of RUNX1 observed in our patient results in the loss of its capacity to transactivate RUNX1 target genes, mimicking what has been observed in the presence of the RUNX1/ETO fusion protein.

Figure 1.

Genomic and expression charactarization of the unbalanced t(4;21)(q21;q22) translocation. (A) GTG-banding revealed a 46,XY, t(4;21)(q21;q22) [17/20] karyogram. The arrows indicate the aberrant chromosomes. (B) Metaphase plate hybridized with the RUNX1 break-apart probe. A fusion signal indicates an intact RUNX1 gene (white asterisk). The green signal indicates a break within the centromeric region outside the RUNX1 gene (white arrow). (C) ArrayCGH representative ideogram showing a mono-allelic deletion of 2.3Mb in chromosome 21q22.12, that results in the loss of RUNX1 gene (blue arrow), and the localization of the BACs used in the RUNX1 FISH break-apart probe. (D) Analysis of the expression levels of RUNX1 gene and 4 RUNX1/ETO target genes in CD34+ bone marrow cells, Kasumi-1 cell line and our t(4;21). This experiment shows that RUNX1 and all RUNX1/ETO target genes are down-regulated on Kasumi-1 cell line and in our t(4;21) case, these differences are statistically significant in both comparisons.

MDS are clonal disorders of the hematopoietic system, with morphological dysplasia, ineffective hematopoiesis and peripheral blood cytopenias. Our patient presented a RCMD MDS subtype that was characterized by a severe thrombocytopenia (29.0×109/L; reference values 130-400×109/L). It has been reported that haploinsufficiency of RUNX1 causes familial thrombocytopenia with propensity to develop AML2 and patients with Down syndrome, in which there is an increased gene dosage of RUNX1, have an increased risk of developing acute megakaryoblastic leukemia (AML-M7).10 All this evidence indicates that RUNX1 dosage has an important role in the development and maturation of the megakaryocytic lineage.

In summary, this work demonstrates that unbalanced translocations involving 21q22 can lead to the complete abrogation of RUNX1 expression and, probably due to the subsequent downregulation of its target genes, may be the genetic origin of the thrombocytopenia and the myelodysplastic features observed in our patient. It is safe to conclude that a more complete knowledge of the RUNX1 pathways will be vital to improve leukemia diagnosis and therapeutics.

Acknowledgments

The authors would like to thank all the co-workers in our laboratory for their excellent technical assistance. This work was supported by an INTRASALUD project PI 08-0440 grant to JCC, a PhD fellowship from Obra Social-Fundación “La Caixa” to JM, and a PhD fellowship from the Ministry of Science and Innovation to AdR-M.

Footnotes

  • * AdR-M and JM contributed equally to this manuscript.

  • The online version of this article has a Supplementary Appendix.

  • Authorship and Disclosures

    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.

  • Received June 29, 2011.
  • Revision received November 7, 2011.
  • Accepted November 9, 2011.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
View Abstract