TET2 gene mutation is a frequent and adverse event in chronic myelomonocytic leukemia
Olivier Kosmider, Véronique Gelsi-Boyer, Marion Ciudad, Cindy Racoeur, Valérie Jooste, Norbert Vey, Bruno Quesnel, Pierre Fenaux, Jean-Noël Bastie, Odile Beyne-Rauzy, Aspasia Stamatoulas, François Dreyfus, Norbert Ifrah, Stéphane de Botton, William Vainchenker, Oliver A. Bernard, Daniel Birnbaum, Michaëla Fontenay, Eric Solary

Author Affiliations

  1. Olivier Kosmider1,2,3,4,
  2. Véronique Gelsi-Boyer5,
  3. Marion Ciudad6,7,
  4. Cindy Racoeur6,7,8,
  5. Valérie Jooste6,7,
  6. Norbert Vey5,
  7. Bruno Quesnel9,
  8. Pierre Fenaux10,11,
  9. Jean-Noël Bastie6,7,8,
  10. Odile Beyne-Rauzy12,
  11. Aspasia Stamatoulas13,
  12. François Dreyfus1,2,3,4,
  13. Norbert Ifrah14,
  14. Stéphane de Botton15,
  15. William Vainchenker15,
  16. Oliver A. Bernard16,
  17. Daniel Birnbaum5,
  18. Michaëla Fontenay1,2,3,4 and
  19. Eric Solary6,7,8,15
  20. on behalf of the Groupe Francophone des Myélodysplasies
  1. 1 Hematology Department, Hôpital Cochin (APHP), Paris
  2. 2 Inserm U567, Paris
  3. 3 CNRS UMR 8104, Paris
  4. 4 University Paris 5, Faculty of Medicine René Descartes, UM 3, Paris
  5. 5 Hematology Department, Institut Paoli-Calmettes, Marseille
  6. 6 Inserm UMR866, 7 Boulevard Jeanne d’Arc, Dijon
  7. 7 University of Burgundy, Faculty of Medicine, 7 Boulevard Jeanne d’Arc, Dijon
  8. 8 CHU Le Bocage, Dijon
  9. 9 CHU Lille, Lille
  10. 10 Hôpital Avicenne (APHP)/ University Paris 13, Bobigny
  11. 11 Inserm U848, Institut Gustave Roussy/University Paris 11, Villejuif
  12. 12 CHU Purpan, Toulouse
  13. 13 CHU Rouen, Rouen
  14. 14 CHU Angers, Angers
  15. 15 Inserm U790, Institut Gustave Roussy/University Paris 11, Villejuif
  16. 16 Inserm E010, Hôpital Necker/University Paris 5, Paris, France
  1. Correspondence: Eric Solary, Inserm UMR866, Faculty of Medicine, 7 boulevard Jeanne d’Arc, 21000 Dijon, France. E-mail: esolary{at}u-bourgogne.fr
View Abstract


Background Acquired somatic deletions and loss-of-function mutations in one or several codons of the TET2 (Ten-Eleven Translocation-2) gene were recently identified in hematopoietic cells from patients with myeloid malignancies, including myeloproliferative disorders and myelodys-plastic syndromes. The present study was designed to determine the prevalence of TET2 gene alterations in chronic myelomonocytic leukemias.

Design and Methods Blood and bone marrow cells were collected from 88 patients with chronic phase chronic myelomonocytic leukemia and from 14 with acute transformation of a previously identified disease. Polymerase chain reaction analysis and direct sequencing were used to sequence exons 3 to 11 of the TET2 gene. Annotated single nucleotide polymorphisms were excluded. Survival curves were constructed by the Kaplan-Meier method.

Results We detected TET2 mutations in 44 of 88 (50%) patients with chronic myelomonocytic leukemia, which suggests that TET2 gene mutations are especially frequent in this myeloid disease. A TET2 gene alteration was identified in 18 of the 43 patients studied at diagnosis and was associated with a trend to a lower overall survival rate; confining the analysis to the 29 patients with chronic myelomonocytic leukemia-1, according to the WHO classification, the difference in overall survival between patients with or without TET2 gene mutations became statistically significant.

Conclusions TET2 gene alterations are more frequent in chronic myelomonocytic leukemia than in other subgroups of hematopoietic diseases studied so far and could negatively affect the patients’ outcome. The striking association between TET2 gene alterations and monocytosis, already observed in patients with systemic mastocytosis, could indicate a negative role of TET2 in the control of monocytic lineage determination.


Chronic myelomonocytic leukemia (CMML) is a clonal myelodysplastic/myeloproliferative disorder observed in the elderly.1 The most frequent genetic abnormalities identified in this disease include mutations in RUNX12,3 and in the polycomb-associated gene ASXL1.4 A copy-neutral uniparental disomy is also frequent and can be associated with mutations of the CBL gene.5,6 RAS gene mutations are observed in one third of CMML,7 while other acquired genetic abnormalities are limited to small subgroups.8 It remains difficult to distinguish among these somatic events those that drive the disease pathogenesis from those that are acquired as a consequence of disease progression. Using various genetic approaches, acquired somatic mutations (deletions, insertions, nonsense and missense point mutations) in the coding sequence of TET2 (Ten-Eleven Translocation-2) gene were recently identified in hematopoietic cells from patients with myeloid malignancies, including myeloproliferative disorders and myelodysplastic syndromes.917 Colony studies in informative cases of myeloproliferative disorders suggested that TET2 mutations could precede the JAKV617F mutation and may endow cells with an increased ability to repopulate the bone marrow of NOD/SCID mice, with respect to TET2 wild-type hematopoietic stem cells.14 The incidence of TET2 gene alterations in various myeloid diseases has been suggested to range between 10 and 25%.917 In patients with systemic mastocytosis, a significantly higher incidence of TET2 mutations was associated with monocytosis.10 The aim of this study was to determine the prevalence and prognostic impact of TET2 gene mutations in CMML.

Design and methods


Peripheral blood and/or bone marrow cells were collected between February 2005 and October 2008 in Dijon (Inserm UMR866) and Marseille (IPC) from 88 patients with CMML1 (n=70) or CMML2 (n=18) according to the World Health Organization (WHO) criteria1 and from 14 patients with acute blastic transformation of a previously identified CMML. Patients signed informed consent to participation in the study in accordance with current ethical regulations. Six of the nine CMML patients included in a previous study14 are part of the present series (the other three were excluded because of lack of sufficient clinical or biological information). We also included 46 patients in whom the ALXL1 sequence had been previously examined.4 None of the other cases has ever been reported. Samples were collected from consecutive patients seen in the different centers, pending the collection of sufficient biological material and annotations. Patients with CMML in chronic phase were either newly diagnosed (n=43) or known to have this hematopoietic disease and were being followed up every 3 months in the absence of active therapy, while receiving supportive care or during cytotoxic treatment (n=45), in most cases with hydroxyurea. Acute transformation was considered to have occurred when the blastic phase was identified. The main characteristics of the patients studied are summarized in Table 1.

Table 1.

Characteristics of the studied patients according to the pressence or absence of TET2 mutations. The only significant difference (higher count of peripheral blood monocytes in the TET2 mutated group) was not confirmed when analysis was limited to patients at diagnosis.

Nucleic acid methods

Blood and bone marrow samples were collected on EDTA and mononuclear cells were selected by Ficoll Hypaque. DNA was extracted using commercial kits (Qiagen, Hilden, Germany). Polymerase chain reaction (PCR) analysis and direct sequencing were performed using standard conditions with gene-specific primers designed to amplify coding sequences spanning from exon 3 to exon 11 of the TET2 gene, as described elsewhere.14,17 For each PCR reaction, 20 ng of genomic DNA were used for the PCR amplification followed by magnetic bead purification and bidirectional sequencing using ABI 3300 capillary sequencers (Agencourt Bioscience, Beverly, MA, USA). Mutation Surveyor (Softgenetics, Inc., Stat College, PA, USA) was used to detect nonsense and missense mutations located in conserved regions spanning from 1134–1444 and 1842–1921 and sequences were reviewed manually to detect frameshift mutations. TET2 abnormalities were numbered according to the FM 992369 EMBL nucleotide sequence database. Previously annotated single nucleotide polymorphisms (http//www.hapmap.org) were not considered pathogenic. The ASXL1 sequence was determined in 49 samples, as described previously.4

Table 2.

List of the 68 TET2 mutations identified in 46 CMML patients, of whom 44 were in chronic phase and 2 in blastic transformation. UPN are arbitrary numbers given to each patient.

Figure 1.

Schematic representation of the location of the 68 mutations identified in 46 patients with CMLL patients, of whom 44 in chronic phase CMML and 2 in blastic transformation. Two distinct mutations were identified in 19 of these patients (see Table 2); 29 frameshifts (white triangles); 20 nonsense mutations (hatched triangles), 16 missense mutations (gray triangles) and 3 mutations in splice sites (black triangles) are indicated. Gray bars indicate conserved domains (AA 1134–1444 and 1842–1921). TET2 abnormalities were numbered according to the FM 992369 EMBL nucleotide sequence database. Previously annotated single nucleotide polymorphisms were not considered pathologic.

Comparative genomic hybridization arrays

Comparative genomic hybridization (CGH) was performed using 244K CGH Microarrays (Hu-244A, Agilent Technologies, Massy, France) with a resolution up to 6 Kb. Scanning was done with an Agilent Autofocus Dynamic Scanner (G2565BA, Agilent Technologies).2,4 Copy number changes in the 4q24 region between 104,680 and 106,960 according to http//www.genome.ucsc.edu were characterized.

Statistical analysis

Statistical analyses were performed using Stata 10TM. All p values were two-tailed and the threshold of statistical significance was p less than 0.05. Clinical and biological parameters were recorded at the time of diagnosis or referral to the medical center. Categorical variables are reported as counts and relative frequencies (%) and compared between groups by χ2 or exact Fisher’s statistics. Continuous variables are indicated as medians and ranges. We used the Mann-Whitney U test to compare continuous variables. Survival curves were constructed by the Kaplan-Meier method using the interval from the date of diagnosis to the date of last contact or death and compared using the log-rank test. A multivariable Cox model was fitted in order to take age at diagnosis into account.

Figure 2.

Kaplan-Meier analysis of overall survival in patients in whom the TET2 mutation was analyzed at diagnosis. (A) Overall survival of 43 patients with either CMML1 or CMML2 according to the WHO classification (mutated TET2: 18 cases; non-mutated TET2: 25 cases). (B) Overall survival of the 29 patients with CMML1 in chronic phase according to the WHO classification (mutated TET2: 10 cases; non-mutated TET2: 19 cases).


The nature and frequency of somatic mutations affecting the TET2 coding sequences were studied in bone marrow or peripheral blood collected from 88 patients with chronic phase CMML according to the WHO criteria. A mutation of the TET2 gene was detected in 44 out of these 88 (50%) patients. The TET2 gene was mutated in 18 (42%) of the 43 patients studied at diagnosis, and in 26 of the 45 patients (58%) studied during the course of their disease. These results suggest that the prevalence of TET2 mutations is higher in CMML than in any other myeloid disease studied.917 The broad range of myeloid disorders in which mutations in the TET2 gene have been identified suggests that the gene has a pleiotropic role in these diseases. It remains to be established what role the mutation plays in the phenotype of the disorders. The striking association between the presence of a TET2 mutation and monocytosis in patients with systemic mastocytosis10 and the very high incidence of TET2 mutations in CMML1 found in this and other studies114 lends support to a phenotypic association.

Among an additional series of 14 CMML patients who had 20% or more blast cells in the bone marrow, indicating blastic transformation, a TET2 mutation was identified in two patients (14%) (Table 2). It was shown recently that the JAK2V617F mutation is frequently absent in leukemic blast cells from patients with transformed JAK2V617F-positive myeloproliferative disorders. In these patients, leukemic transformation could arise from a JAK2V617F-negative ancestor cell.18 Larger series and follow-up of individual cases will be needed to determine whether the TET2 mutation, when present at diagnosis, can be lost upon leukemic transformation.

Two distinct mutations in the sequence of the TET2 gene were identified in 23 out of the 44 (52%) TET2-mutated patients with chronic phase CMML, including 6 out of the 18 (33%) patients whose mutations were identified at diagnosis, and 16 out of the 26 (61%) patients whose mutations were identified during the course of their disease. Two distinct mutations were also identified in one of the patients in blastic transformation of CMML. Altogether, 68 mutations were identified, including 29 frameshift mutations, 20 nonsense mutations, 16 missense mutations and 3 mutations targeting a splice site. The mutations most frequently involved exon 3 (22 events), exon 10 (9 events) and exon 11 (13 events). Missense mutations were considered if located in conserved domains, most of them leading to modifications in potentially important amino-acids in the protein19 (Table 2, Figure 1).

Conventional cytogenetic analysis of bone marrow was performed in 72 of the 88 patients with chronic phase CMML and detected abnormalities in 17 cases but never identified any deletion of the 4q24 band, in either the mutated or the non-mutated group of patients. Genome-wide high-density arrays (CGH) compared the leukemic cell profile to normal DNA in 28 of the 88 cases and detected a TET2 deletion (according to the http://genome.uscs.edu) in one of the ten studied patients with a mutated TET2 copy. Thus, copy number alterations and deletion of the wild-type TET2 copy in TET2-mutant CMML cases appears to be uncommon, although this remains to be proven in larger series.20 It is too early to determine whether the numbers and types of mutation (point mutation or frameshift) and the gene dosage (loss of one or two copies) differ among the various myeloid disorders and contribute to the disease phenotype. When the ASXL1 gene sequence could be analyzed simultaneously (n=49), a mutation was found in patients with wild-type as well as mutated TET2 (7 of 14 samples and 14 of 35 respectively; p=ns).

The clinical and biological features of the 88 patients with chronic phase CMML are presented in Table 1. The presence of a TET2 mutation was associated with a trend towards higher monocyte and lower platelet counts. Analysis of overall survival was performed in the 43 patients whose TET2 status was determined at diagnosis and indicated a lower 1-year overall survival rate in the 18 patients of this cohort with a TET2 mutation, but the difference was not statistically significant (Figure 2A). When the overall survival analysis was limited to the 29 patients with CMML1, according to the WHO classification, and a follow-up of at least 2 months, the difference between those with and without TET2 mutations became significant (p<0.01 Figure 2B). The survival of CMML1 and CMML2 patients was not significantly different at 12 months but all patients with CMML2 died within 28 months of diagnosis whereas half of the CMML1 patients were still alive (data not shown). The survival of patients with secondary acute myeloid leukemia was significantly shorter than that of those with CMML1 or CMML2 (p<0.04, data not shown). None of the other tested parameters, including age, sex and FAB classification, affected survival. Given the low number of patients in each group, only age was introduced in the Cox model and did not affect the trend for a negative effect of TET2 mutation on survival (p=0.08).


The present study indicates that, in addition to RUNX1,2,3 ASXL1,4 and RAS genes,6 TET2 is a commonly mutated gene in patients with CMML. Half of the patients with mutated TET2 had two distinct gene alterations, suggesting that the two gene copies were affected. In four cases, sequencing results showed a 100% mutant sequence, which could indicate a combination of TET2 mutation and either uniparental disomy or 4q24 deletion or a homozygous mutation. We did not collect enough biological material to perform CGH, which would have been able to identify 4q24 deletions involving TET2, and single nucleotide polymorphism analysis, which would have been able to detect segmental acquired uniparental disomy resulting in loss of heterozygosity.5 Based on the series published to date, the frequency of TET2 mutations in chronic phase CMML ranges between 35 and 42%.1113 TET2 gene mutations were associated with increased monocytosis, as observed in mastocytosis.10 We also noted a higher number of immature dysplastic granulocytes in the peripheral blood of CMLL patients with mutated TET2 gene (data not shown). Identification of other surrogate markers of the TET2 mutation may be useful as the mutations in this gene as well as in other genes recently identified to be mutated in CMML, such as ASXL1,4 are spread over the full length of the genes.

The TET family includes three genes (TET1, TET2 and TET3) with highly conserved regions.13 TET1, standing for ten-eleven translocation 1, is also known as CXX6 or LCX and was identified as a fusion partner of the MLL gene in the acute myeloid leukemia-associated translocation t(10;11)(q22;q23).21,22 The MLL-TET1 fusion gene was also detected in two adults with CD10-negative B-cell precursor acute lymphoblastic leukemia23 and a single nucleotide polymorphism in the TET1 gene coding region has been associated with late-onset Alzheimer's disease.24 Recent evidence indicates that TET1, and possibly other proteins of the family, encodes an enzyme responsible for the conversion of 5-methylcytosine to 5-hydroxymethylcytosine,19 thus having potential roles in CpG methylation pattern and epigenetic regulation. TET2 was suggested to have tumor suppressor function14 but the physiological functions of the TET2 protein in hematopoiesis are yet to be identified, for example, to determine whether TET2 could negatively regulate monocyte lineage determination, and to analyze how the loss of TET2 protein through a variety of genetic mechanisms leads to myeloid cell proliferation and dysplasia.

The negative prognostic impact of TET2 mutations in chronic phase CMML needs to be corroborated in a prospective study determining the combined clinical impact of recently identified frequent mutations in TET2, ASXL1, RUNX1 and RAS genes.26 Such a prospective study has been initiated by the Groupe Francophone des Myelodysplasies in the context of an ongoing phase II trial testing decitabine in CMML patients.


  • OK and VG-B contributed equally to the work.

  • Funding: this work was supported by grants from the Programme Hospitalier de Recherche Clinique (PHRC national MAD06 to MF and ES), the Ligue Nationale Contre le Cancer (équipes labellisées, ES, OB, VW), the association «cent-poursanglavie», the Association Nationale de la Recherche (ES), the National Institute of Cancer (INCa – ES, MF, OB, WV), the Canceropole Ile-de-France (AAP INCA 2008) (MF), and the Fondation de France, Fondation contre la leucémie (OB, DB).

  • Authorship and Disclosures

    OK and VGB performed genetic analyses, MC and CR received the samples and sorted the cells, VJ performed statistical analyses, NV, BQ, PF, NJB; OBR, AS, FD, NI, and SdB provided samples, WV and OAB discovered TET2 mutations, DB, MF and ES designed the study, ES wrote the paper.

  • The authors reported no potential conflicts of interest.

  • Received May 12, 2009.
  • Revision received July 1, 2009.
  • Accepted July 1, 2009.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
View Abstract