Endothelial colony-forming cells from patients with chronic myeloproliferative disorders lack the disease-specific molecular clonality marker

Recently, a role for circulating endothelial progenitor cells (EPCs) in the pathogenesis of bone marrow and spleen neoangiogenesis of patients with chronic myeloproliferative disorders (CMPDs) has been postulated, opening the way to the hypothesis that a common progenitor to both hematopoietic and vascular endothelial lineage, the so-called hemangioblast, could be the malignant stem cell of these disorders. The presence of a hemangioblast was initially suggested in chronic myeloid leukemia (CML),¹  when the BCR-ABL fusion gene was detected in a proportion of endothelial-like cells generated in vitro from mononuclear cells (MNCs). Since then, this hypothesis was tested in CML and Ph-negative (Ph⁻) CMPDs,^2,3  reinforcing the possibility that either the BCR-ABL rearrangement or the JAK2-V617F mutation, respectively, might happen at the level of a hemangioblastic progenitor cell.⁴ 

Several theories have been proposed about the origin and classification of EPCs.^5–8  The scene has been further entangled by the finding that 2 types of EPCs, namely, colony-forming unit–endothelial cells (CFU-ECs) and endothelial colony-forming cells (ECFCs) can be detected in human peripheral blood.^9–11  More recently, Yoder et al^12–14  clearly settled that CFU-ECs, acknowledged by many authors as endothelial in origin,^7,10  derive from the hematopoietic system, with no ability to form secondary endothelial colonies in vitro or new vessels in vivo. In contrast, ECFCs were shown to belong to the endothelial lineage.¹⁴  They have a robust proliferative potential, and vessel-forming capacity in vivo. Therefore, CFU-ECs and ECFCs are 2 distinct progenitor cells.

We have readdressed the issue of endothelial cell origin in CMPDs by investigating whether circulating EPCs, isolated as both CFU-ECs and ECFCs, bear the molecular clonality marker of the disease. For CML, the BCR-ABL fusion gene was used as such a marker, whereas for Ph⁻ CMPDs the JAK2-V617F mutation and chromosomal aberrations detected by comparative genomic hybridization (CGH) were used.

Seventy-six patients with CMPDs and 27 healthy subjects were studied. All patients were out of therapy, and their characteristics are reported in Table 1. The Institutional Review Board at Ospedale San Martino in Genoa and Istituto di Ricovero e Cura a Carattere Scientifico Policlinico San Matteo Foundation in Pavia approved all protocols. Informed consent was obtained according to the Declaration of Helsinki.

Fresh tissue samples were obtained from the spleen of 3 patients with primary myelofibrosis (PMF) who underwent therapeutic splenectomy. After extensive washing, single-cell suspension was obtained by mechanical disaggregation. Cells were then used for the ECFC assay, as described in “EPC culture and characterization.”

Circulating ECFCs were isolated according to Ingram et al,¹¹  whereas CFU-ECs were grown following the method of Hill et al.¹⁰  When appropriate, single ECFC-derived colonies were expanded for 2 or 3 passages to obtain an amount of cells suitable for further experiments, whereas CFU-ECs could not be expanded and were pooled. Immunophenotypic characterization was performed by flow cytometry analysis.^11,15  Capillary formation assay was performed as described¹¹  using ECFCs at passage 2, or pooled CFU-ECs.

Total RNA was extracted using the PicoPure RNA Isolation Kit (Arcturus Bioscience). Quantitative reverse transcription–polymerase chain reaction (RT-PCR) for BCR-ABL and ABL were performed using the TaqMan system (7900 HT Fast Real-Time PCR System; Applied Biosystems).¹⁶ 

ECFCs and CFU-ECs from 7 patients with CML were studied by interphase fluorescence in situ hybridization (FISH) with the LSI BCR-ABL dual-color, dual-fusion DNA probe from Vysis.

Genomic DNA from granulocytes was extracted with the Purgene Genomic DNA Purification Kit (Gentra Systems) and amplified in an allele-specific PCR.¹⁷  DNA from individual colonies was obtained by cellular lysis at 90°C for 5 minutes and amplified by a seminested PCR.¹⁸ 

Array-CGH was performed with the Human Genome CGH Microarray kit (Agilent Technologies)¹⁹  on 7 individual ECFC-derived colonies from 5 PMF patients. Every array-CGH experiment (ECFCs vs polymorphonuclear leukocytes [PMNs]) was performed 3 times to avoid possible artifacts. ECFC and PMN cells were hybridized against a normal control (Human Genomic DNA; Promega) to confirm the presence/absence of imbalances.

The frequency of ECFCs and CFU-ECs was highly variable both in CMPD patients and healthy subjects (Figure 1A-B). In both groups, ECFCs had a comparable endothelial phenotype (not shown) and were able to form capillary-like structures in vitro. CFU-ECs from patients and controls had also similar phenotypes, but they differed from ECFCs for the positive expression of CD45 and CD14 (not shown) and for the formation of an unshaped capillary network in vitro (Figure 1C-D). Quantitative RT-PCR on pooled CFU-ECs from CML patients invariably showed the presence of the specific p210 transcript, which, on the contrary, was never observed in 51 ECFC-derived colonies. In addition, ECFCs from 7 CML patients resulted negative for the presence of the t(9;22)(q34;q11) at interphase FISH, whereas all the CFU-EC-derived cells showed the fusion signal. The JAK2-V617F mutation was present in the genomic DNA from granulocytes of all Ph⁻ CMPD samples. All ECFCs (n = 75) showed the JAK2 wild-type allele, whereas all CFU-ECs showed the JAK2-V617F allele (Figure 1E-G). In addition, 6 ECFCs, obtained from spleen samples of 3 PMF patients, showed the absence of the JAK-2V617F mutation. Consistent with these results, array-CGH showed the absence of chromosomal abnormalities in all ECFC-derived colonies from 5 PMF patients, although the PMN cells of 2 patients were characterized by chromosomal imbalances. One patient presented 3 deletions in his granulocytes: del(4)(21.21-q25) of 29 Mb, del(7)(pter-p13) of 47 Mb, and del(7)(q21.11-qter) of 75.3 Mb. The second patient was characterized by chromosome 9 trisomy. All anomalies were found with a medium log₂ ratio representative of a cellular mosaic between an anomalous cancer clone and normal cells of approximately 90% and 50%, respectively. Thus, FISH and quantitative RT-PCR results obtained in ECFCs from CML patients as well as PCR and array-CGH results in ECFCs from Ph⁻ CMPD patients clearly document that circulating EPCs are not clonally related to the cell giving rise to the hematopoietic malignancy. On the other hand, the presence of the disease marker in all CFU-ECs confirms that they belong to the malignant clone, as expected because they derive from myeloid cells.^13,14  That the clone giving rise to CMPD does not encompass all lineages is known for a long time; indeed, often T cells in CML do not belong to the clone bearing the BCR-ABL transcript.²⁰  Moreover, in keeping with our findings, Della Porta et al²¹  reported that myelodysplastic syndrome patients carrying chromosomal abnormalities in their hematopoietic cells had normal karyotype in their ECFCs both by FISH and by genotyping of polymorphic loci. Previously, Oppliger Leibundgut et al reported the presence of either trisomy 8 or JAK2-V617F mutation in the endothelial cells of CMPD patients with the same abnormalities in their granulocytes,³  whereas Gunsilius et al reported that endothelial-like cells grown in vitro from MNCs of CML patients had the BCR-ABL translocation.¹  These results, interpreted as evidence that the endothelial lineage was part of the malignant clone, were probably obtained using an in vitro assay similar to that used for the growth of CFU-ECs. Taken together, our results challenge the hypothesis of a malignant common progenitor cell for hematopoietic and endothelial lineages and indicate that the transforming event occurs in a progenitor cell devoid of angiogenetic potential.

Contribution: G.P., V.R., F.F., and G. Barosi designed and performed research, collected and analyzed data, and wrote the paper; F.B., G. Bergamaschi, L.V., S.P., D.I., and A.G. performed quantitative RT-PCR analysis; R.G. and A.B. designed research and collected and analyzed data; M.C., E.B., and B.C. grew the endothelial cells; M.S. performed cytogenetic analysis; R.C., G.V., and M.M. performed FACS analysis; A.P. performed cytochemical studies; and F.N. and O.Z. performed CGH experiments.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Francesco Frassoni, Centro Cellule Staminali e Terapia Cellulare, Ospedale San Martino, Lgo R Benzi, 10, 16132 Genova, Italy; e-mail: francesco.frassoni@hsanmartino.it.