Annals of Hematology

, Volume 91, Issue 4, pp 533–541

TET2, ASXL1, IDH1, IDH2, and c-CBL genes in JAK2- and MPL-negative myeloproliferative neoplasms

Authors

  • Luz Martínez-Avilés
    • Department of PathologyHospital del Mar
    • Universitat Autònoma de Barcelona, IMIM-Hospital del Mar
    • Grup de Recerca Aplicada en Neoplasies Hematològiques, IMIM-Hospital del Mar
    • Department of Clinical HematologyHospital del Mar
    • Grup de Recerca Aplicada en Neoplasies Hematològiques, IMIM-Hospital del Mar
  • Alberto Álvarez-Larrán
    • Department of Clinical HematologyHospital del Mar
    • Grup de Recerca Aplicada en Neoplasies Hematològiques, IMIM-Hospital del Mar
  • Erica Torres
    • Department of PathologyHospital del Mar
  • Sergi Serrano
    • Department of PathologyHospital del Mar
    • Universitat Autònoma de Barcelona, IMIM-Hospital del Mar
  • Beatriz Bellosillo
    • Department of PathologyHospital del Mar
    • Universitat Pompeu Fabra, IMIM-Hospital del Mar
    • Grup de Recerca Aplicada en Neoplasies Hematològiques, IMIM-Hospital del Mar
Original Article

DOI: 10.1007/s00277-011-1330-0

Cite this article as:
Martínez-Avilés, L., Besses, C., Álvarez-Larrán, A. et al. Ann Hematol (2012) 91: 533. doi:10.1007/s00277-011-1330-0

Abstract

Mutations in the TET2 and ASXL1 genes have been described in approximately 14% and 8% of patients, respectively, with classic myeloproliferative neoplasms (MPN), but their role as possible new diagnostic molecular markers is still inconclusive. In addition, other genes such as IDH1, IDH2, and c-CBL have also been reported in several myeloid neoplasms. We have studied the mutational status of TET2 (complete coding region), ASXL1 (exon12), IDH1 (R132), IDH2 (R140 and R172), and c-CBL (exons 8 and 9) in 62 MPN patients (52 essential thrombocythemia (ET), five polycythemia vera (PV), and five primary myelofibrosis (PMF)) negative for both JAK2 (V617F and exon 12) and MPL (exon 10) mutations. Pathogenic alterations in the TET2 gene were detected in three out 52 ET cases (4.8%). ASXL1 gene pathogenic mutations were also detected in three cases (two ET and one PMF). One ET patient harbored, simultaneously, one TET2 and one ASXL1 mutations. Mutations in the TET2 and ASXL1 genes showed no association with the JAK2 46/1 haplotype. Analysis of a JAK2V617F-positive cohort of 50 ET patients showed no mutations in either the TET2 or ASXL1 genes. Regarding IDH1, IDH2, and c-CBL genes, no mutations were found in any patient. In conclusion, TET2 and ASXL1 pathogenic mutations are found in 8% of MPN lacking JAK2 and MPL mutations, whereas IDH1, IDH2, and c-CBL mutations are not detected in this subset of patients.

Keywords

TET2ASXL1IDH1IDH2c-CBLMutations

Introduction

Myeloproliferative neoplasms (MPN) are a spectrum of clonal disorders of the hematopoietic system characterized by overproduction of mature blood elements, a trend to thrombotic and/or hemorrhagic complications and variable rates of transformation to myelofibrosis and acute leukemia. The presence of JAK2 and MPL mutations is a major diagnostic criteria in the WHO classification of classic MPN. However, a variable percentage of patients lack both molecular markers. The molecular basis of JAK2- and MPL-negative MPN remains largely unexplained. Recently, new molecular markers have been described in a vast array of myeloid cancers. Alterations in the TET2 gene, a putative tumor suppressor gene located at chromosomal region 4q24, have been identified in 7–13% of MPN patients, in 19–26% myelodysplastic syndromes (MDS), in 12–24% of acute myeloid leukemia (AML), in 20–40% of chronic myelomonocytic leukemia (CMML), and in 29% of systemic mastocytosis [18].

More recently, mutations in the ASXL1 gene which is located in chromosomal region 20q11 have also been reported in several myeloid neoplasms such as MDS (10%), CMML (40%), and MPN (8%). ASXL1 mutations mainly occur in exon 12, leading to a premature truncated protein that lacks its PHD domain (C-terminal plant homeodomain) involved in the interaction between proteins, and therefore compromising the formation of chromatin modifier complexes. The prognostic relevance of these alterations remains unknown [811]. In addition, mutations in the IDH1 and IDH2 genes that encode for enzymes responsible of the oxidative decarboxylation of isocitrate to α-ketoglutarate have been described in around 10% of de novo AML, in 20% AML secondary to a MPN, and more rarely in chronic-phase MPN [8, 1217]. Finally, alterations in the c-CBL gene, that gives rise to the c-Cbl protein with ubiquitin ligase activity and targets a variety of tyrosine kinases for degradation by ubiquitination, have been identified in variable percentages in AML, MDS/MPN, juvenile myelomonocytic leukemia, and CMML patients [1822].

On the other hand, different studies have identified a JAK2 haplotype named as 46/1 that seems to predispose to the acquisition of the JAK2V617F mutation in the same allele carrying the haplotype, although the underlying mechanism remains unknown. This JAK2 haplotype has also been associated with the acquisition of JAK2 exon12 and MPL mutations [2327], but little is known about the correlation between the JAK2 46/1haplotype and the aforementioned genes.

The role of TET2, ASXL1, IDH1, IDH2, and c-CBL mutations in polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF) patients who lack JAK2 and MPL mutations has not been widely studied in terms of diagnostic relevance. In addition, their incidence in this subset of patients and the correlation with the presence of JAK2 46/1 haplotype is not well known. The aim of this study was to determine the incidence and diagnostic significance of mutations in this group of genes in a cohort of MPN negative for both JAK2 and MPL mutations.

Materials and methods

Patients

From a whole cohort of 241 patients with classic MPN—93 PV, 132 ET, and 16 PMF patients—from our institution and diagnosed according to the WHO criteria, we selected those patients lacking the following genetic alterations: JAK2V617F, JAK2 exon 12 mutations, or MPL exon 10 mutations (S505N or W515K/L). A group of 62 patients (five PV, five PMF, and 52 ET) lacking JAK2 and MPL mutations were analyzed for TET2, ASXL1, IDH1, IDH2, and c-CBL mutations (Table 1). Sixty-one out of 62 patients were analyzed at the chronic phase of the disease. Eight patients were analyzed at the moment of diagnosis whereas the remaining 54 were analyzed during follow-up. The median interval time between the date of diagnosis and the sample collection was 78.2 months (0–367.9). In addition, we analyzed the presence of TET2 or ASXL1 mutations in a group of 50 JAK2V617F-positive ET patients in order to compare the frequency of mutations in these genes between ET patients positive and negative for the JAK2V617F mutation. The study was approved by the local ethics committee, and all patients provided written informed consent according to the Declaration of Helsinki. Biological samples were obtained from Parc de Salut MAR Biobank (MARBiobanc; Barcelona).
Table 1

Main characteristics of JAK2V617F-negative MPN patients

 

ET

PV

PMF

N = 52

N = 5

N = 5

Agea

56 (28–86)

51 (32–70)

63 (53–81)

Gender (M/F)

14:38

4:1

1:4

Hb (g/La)

138 (113–166)

194 (183–203)

116 (101–128)

WBC × 109/La

8 (5–13)

8 (6–10)

9 (6–12)

Platelet count × 109/La

791 (493–1,471)

347 (167–597)

221 (71–506)

TET2 pathogenic mutations

3b/52

0/5

0/5

ASXL1 pathogenic mutations

2b/52

0/5

1/5

Status at last follow-up

 Alive in CP

46

5

4

 Dead

4

0

1

 Transformed

2

0

0

Follow-up (monthsa)

118 (4–296)

215 (95–454)

65 (6–125)

ET essential thrombocythemia, PV polycythemia vera, PMF primary myelofibrosis, M male, F female, WBC white blood cells, CP chronic phase

aMedian (range)

bOne patient harbored simultaneously mutations in TET2 and ASXL1

Cell fractionation and RNA extraction

Peripheral blood was collected on ethylendiaminetetracetic acid. Granulocytes were isolated by Lymphoprep (1.077 g/ml) density gradient, followed by dextran sedimentation. Total RNA was isolated from granulocytes using guanidium thiocyanate method (Ultraspec™ Biotecx Laboratories, Houston, TX). cDNA was reverse transcribed from 1 μg of total RNA with murine Moloney leukemia virus reverse transcriptase (Invitrogen, Paisley, UK) according to standard procedures with random hexamers.

TET2, ASXL1, IDH1, IDH2, and c-CBL analysis by direct sequencing

The mutational analysis of TET2 (complete coding region), ASXL1 (exon 12), IDH1 (R132), IDH2 (R140 and R172), and c-CBL (exons 8 and 9) was performed by direct sequencing using cDNA from granulocytes. Primers were designed using the Primer Express software (Applied Biosystems, Foster City, CA). PCR conditions and primer sequences used for PCR amplification and sequencing are available in the Electronic supplementary material. Sequencing was performed with BigDye v3.1 (Applied Biosystems) following the manufacturer´s instructions and analyzed on an ABI3730XL Sequencer (Applied Biosystems). The sequence data files were analyzed using SeqScape software (Applied Biosystems), and all mutations were confirmed with an independent PCR.

Determination of the JAK2 46/1 haplotype

Granulocyte DNA samples were genotyped using the TaqMan SNP genotyping assays for the rs12340895 and rs12343867 SNPs (Applied Biosystems). Genotyping was performed in an Applied Biosystems 7500 Fast Real-Time PCR System.

High-resolution melting analysis for TET2 mutation detection

High-resolution melting (HRM) analysis was performed using the Light Cycler® High-Resolution Melting Master reagents in a Light Cycler® 480 real-time PCR system (Roche Applied Science, Basel, Switzerland) using the manufacturer’s instructions, with the same primers used in the sequencing analysis.

Cloning assays for the characterization of TET2 variants

Amplification of the TET2 cDNA segment comprising exons 3 to 10 (where a fragment of 1,097 base pairs was expected) was performed using the following primers: forward 5′-TGAAATGTCAGGGCCAGTCA-3′ and reverse 5′-TGTATAAAGGCAGAACGTGAAGCT-3′.

PCR products obtained were cloned in the pCR4 sequencing vector (Invitrogen) and transformed in TOP10 bacteria (Invitrogen). Individual bacterial colonies were picked and grown overnight in LB medium containing 100 μg/ml of ampicillin. Plasmids were extracted using the Purelink™ Quick Plasmid Miniprep Kit (Invitrogen) and sequencing of the insert was performed to confirm the different splicing variants.

Results

TET2 mutational status

Three pathogenic mutations in the TET2 gene were detected in 3 out of 62 MPN patients as shown in Table 2. One alteration consisted in a deletion of 10 nucleotides that led to a frameshift mutation (p.V1395fs) whereas in the other two cases a stop codon amino acid substitution (p.Q706X and p.S1848X) producing a premature truncated protein was observed (Table 2). The three mutations were not present in matched normal DNA from mononuclear cells indicating that these mutations were somatically acquired since they only affected myeloid cells. Of note, two of the mutations (p.V1395fs and p.S1848X) were localized in highly conserved regions of the TET2 gene.
Table 2

TET2 and ASXL1 mutations in blood cells subpopulations

Patient no.

Diagnosis

TET2 mutation

ASXL1 mutation

JAK2 haplotype 46/1

Granulocytes

Mononuclear cells

Granulocytes

Mononuclear cells

1

MF post-ET

p.Q706X

Wild typea

Wild type

Wild type

2

ET

p.S1848X

Wild type

Wild type

Wild type

+

3

ET

p.V1395fs

Wild type

p.R634fs

p.R634fs

4

PMF

Wild type

Wild type

p.N893fs

p.N893fs

+

5

ET

Wild type

Wild type

p.G966_G967del

p.G966_G967dela

aMutational analysis performed in purified T lymphocytes

In 44 out of 62 (71%) patients, a total of 14 different missense mutations and one silent mutation in the coding sequence of the gene were observed. The most frequent missense mutation was the I1762V, which was detected in 28 patients (Table 3). In order to analyze whether these mutations were somatically acquired in myeloid cells, we studied the presence of these variations in DNA obtained from T lymphocytes in 21 cases. All missense and silent mutations were detected in normal DNA suggesting that they were single-nucleotide polymorphisms (SNPs) of the TET2 sequence.
Table 3

Missense and silent mutations detected in the coding sequence of TET2 and ASXL1 genes

Variants

Number of patients

TET2

 P29Ra

2

 L34Fa

4

 V218Ma

3

 P363La

5

 G355Da

3

 Y867Ha

2

 Q810Ra

1

 S1039Sa

1

 V1718La

2

 M1701Ia

2

 L1721Wa

12

 P1723Sa

2

 I1762Va

28

 H1778Ra

5

 A1831S

1

ASXL1

 P750Pa

1

 H995Ha

2

 E1102Da

1

 R1171R

6

 S1253Sa

30

 L1325Fa

1

All these missense and silent mutations have been detected in control DNA suggesting that they probably are SNPs of the coding sequence of TET2 and ASXL1

aPreviously reported TET2 and ASXL1 variants

The patient with the TET2 p.Q706X mutation developed myelofibrosis 12 years after ET diagnosis. The mutation was detected by direct sequencing in a sample obtained at an advanced phase of the disease, but not in a previous sample corresponding to an initial phase of the myelofibrotic transformation (3 years before). In order to confirm that this mutation was absent at this moment, we performed HRM analysis, which has a higher sensitivity than direct sequencing. With this technique, we detected in the first sample (initial phase of myelofibrosis) a similar melting profile, but with a lower mutant allele burden, to the one observed in the sample obtained three years later, where the TET2 mutation was detected both by direct sequencing and HRM, suggesting the presence of a small clone with the TET2 mutation that expanded during the progression of the disease (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1330-0/MediaObjects/277_2011_1330_Fig1_HTML.gif
Fig. 1

Mutational analysis of the TET2 gene segment containing the amino acid Q706. a Sequence chromatograms showing the wild-type amino acid Q706 in an early phase of myelofibrosis (left) and the TET2 p.Q706X mutation at an advanced phase of myelofibrosis in a TET2-positive MF post-ET patient. b HRM profiles of the aforementioned samples showing that the TET2 mutation was already present in the early phase of myelofibrosis but with a lower mutant allele burden

Moreover, analysis of the TET2 gene showed the presence of two additional cDNA variants to the one corresponding to the whole cDNA, resulting from alternative splicing of mRNA. To confirm the presence of these two variants, amplification of a long cDNA fragment comprising partially exons 3 to 10 was performed. PCR products were cloned in a pCR4 sequencing vector and transformed to competent bacteria. The sequencing analysis of the construct extracted from the positive transformants confirmed the presence of the two additional TET2 variants. One of them consisted in exclusion of exon 4 causing the direct union between exons 3 and 5, and a second variant consisting in the exclusion of exons 4, 5, and 6, causing the direct union between exons 3 and 7. These results were obtained not only with cDNA from granulocytes and lymphocytes from MPN patients, but also from healthy controls, indicating that the presence of these variants is not associated with the disease since they are also present in healthy controls (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1330-0/MediaObjects/277_2011_1330_Fig2_HTML.gif
Fig. 2

Sequencing chromatograms and schematic representation of the TET2 variants. a Sequencing chromatogram showing the expected union between exons 3 and 4 corresponding to the whole coding sequence (GenBank accession number NM_1127208.2; Ensembl accession number ENST00000380013). b Sequencing chromatogram showing the direct union between exons 3 and 5 due to exon 4 skipping (not reported in GenBank; Ensembl accession number ENST00000545826). c Sequencing chromatogram showing the direct union between exons 3 and 7 due to exons 4, 5, and 6 skipping. d Schematic representation of the TET2 variants found: 1, Transcript variant including all 11 exons (GenBank accession number NM_1127208.2; Ensembl accession number ENST00000380013); 2, transcript variant with exclusion of exon 4 (Ensembl accession number ENST00000545826); 3, transcript variant with exclusion of exons 4, 5, and 6

ASXL1 mutational status

The analysis of the mutational status of ASXL1 exon 12 in the same cohort of patients, showed the presence of two frameshift mutations leading to premature truncated proteins in one ET and one PMF patient. The two pathogenic mutations were as follows: p.N893fs and p.R634fs. The p.R634fs mutation was found in the same patient in whom the pathogenic p.V1395fs TET2 mutation was detected. In a third patient, one in-frame deletion of three nucleotides affecting two glycines at amino acid positions 966 and 967 was detected (Table 2). The pathogenic effect of this mutation (p.G966_G967del) resulting only the deletion of one amino acid is currently unknown.

Analysis of normal DNA of the patient who simultaneously carried the TET2 and the ASXL1 deleterious mutation showed that the TET2 mutation (p.V1395fs) was not detected in DNA from peripheral blood mononuclear cells whereas ASXL1 mutation (p.R634fs) was present in both granulocytes and mononuclear cells. In a similar way, we detected the ASXL1 p.N893fs and p.G966_G967del mutations in DNA from mononuclear cells in patients 4 and 5, respectively.

In addition, we also detected two missense and four silent mutations in the exon 12 sequence of ASXL1 in 63% of patients (Table 3). We analyzed the presence of these missense mutations in T lymphocytes DNA being all of them present in non-myeloid cells suggesting that they probably correspond to SNPs of the ASXL1 coding sequence.

IDH1, IDH2, and c-CBL mutational status

Regarding the mutational analysis of the IDH1 (R132), IDH2 (R140 and R172), and c-CBL (exons 8 and 9) genes, no mutations were detected in any of the 62 patients analyzed, suggesting that these genes might not be involved in the pathogenesis of a myeloid neoplasm in chronic phase.

JAK2 46/1 haplotype

We assessed rs12340895 and rs12343867 SNPs genotypes in order to determine the presence of the JAK2 46/1 haplotype in our cohort of MPN without JAK2 and MPL mutations. Twenty-nine patients did not present the JAK2 46/1 haplotype, whereas 33 patients presented this haplotype (29 heterozygous; 4 homozygous). One of the patients with a TET2 mutation presented a heterozygous JAK2 46/1 haplotype, whereas the second one did not present the haplotype. The same results were found in two of the ASXL1 positive patients. Finally, the patient harboring simultaneously mutations in TET2 and ASXL1 did not present the JAK2 46/1 haplotype.

TET2 and ASXL1 mutational status in JAK2V617F-positive ET patients

In order to compare the incidence of mutations in TET2 and ASXL1 in ET patients according to their JAK2 mutational status, we analyzed an additional group of 50 JAK2V617F-positive ET patients. We did not detect any TET2 or ASXL1 pathogenic mutations in this group of patients, suggesting that the presence of these mutations in JAK2V617F-positive ET patients is less frequent than in JAK2V617F-negative ET patients.

Discussion

In the present study we have analyzed the incidence of TET2, ASXL1, IDH1, IDH2 and c-CBL mutations in 62 MPN patients lacking JAK2 (V617F or exon 12) and MPL (W515K/L and S505N) mutations.

TET2 mutations were observed in three (4.8%) patients, a frequency similar to other studies reporting results on JAK2V617F-negative cohorts (2.5–8%) [1, 2, 5]. Regarding disease specific rates, three out of 52 (5.8%) ET patients harbored TET2 mutations, a similar incidence found by other authors [1, 5, 8]. To compare the incidence of mutations in ET patients according to their JAK2 mutational status, we analyzed a group of 50 JAK2V617F positive ET patients, but none of them presented mutations in either TET2 or ASXL1 indicating that mutations in these genes are less frequent in our series. These results suggest that the study of these genes would mainly be useful in the diagnosis of patients lacking JAK2 or MPL mutations although this finding should be corroborated in more studies.

Mutations in TET2 were initially described as a genetic event previous to the JAK2V617F acquisition [2] but more recently, some studies have described TET2 mutations as a late event in the progression of MPN to leukemic transformation [2831]. In our study, one ET patient carrying a TET2 mutation developed myelofibrotic transformation. The mutation was detected by direct sequencing in the advanced phase of the myelofibrosis, but analysis by more sensitive techniques such as HRM showed a melting profile corresponding to the TET2 mutation in the sample from the initial phase of the myelofibrotic transformation, suggesting the presence of a small TET2-positive clone that expanded during the progression of the disease. Similarly, Delhommeau et al. [2] reported a patient diagnosed with a TET2-negative first phase refractory anemia with excess of blasts (RAEB1) showing a TET2 mutation when the disease evolved to phase two refractory anemia (RAEB2). These findings support the possibility that TET2 mutations may not only contribute to the development of a MPN but also represent a late event in the progression of a MPN to leukemic transformation or myelofibrotic transformation [2831].

Concerning the TET2 mutational analysis, we detected three different TET2 isoforms. One of them contained all eleven exons with the expected fusion between exons 3 and 4 whereas the second variant lacked exon 4 and the third one lacked exons 4, 5, and 6. Of note, TET2 exon 4 partially encodes for one of the highly conserved domains of the TET2 gene so these additional variants would not contain this conserved domain and the function of these potentially new TET2 proteins should be further studied. We showed that these variants are also present in healthy controls, so they are less likely to be pathogenic variants.

Mutations affecting exon 12 of the ASXL1 gene have also been described in different myeloid neoplasms as well as in 11% of MDS, 43% of CMML, and 8% of MPN. In our study, three patients harbored ASXL1 mutations (4.8%), a lower frequency than the one reported by other studies [10]. Of note, one ET patient harbored simultaneously a TET2 and an ASXL1 mutation. Four different silent mutations and two missense mutations were detected in 63% of patients; all of them present in control DNA.

When we analyzed the pathogenic TET2 and ASXL1 mutations in those patients whose control DNA was available, we observed that TET2 mutations were not present in mononuclear cells DNA, indicating that TET2 mutations mainly affect the myeloid lineage. By contrast, all ASXL1 mutations were detected in DNA from mononuclear cells suggesting that ASXL1 mutations may affect an early hematopoietic progenitor or might be present in germline DNA.

We also looked for mutations in the IDH1, IDH2, and c-CBL genes that have also been described in myeloid neoplasms, mainly in the novo or secondary acute myeloid leukemia. Similarly to other studies, no mutations were detected in any of the patients with MPN, indicating that these genes are rarely mutated in the chronic phase of a MPN although they might be more relevant in an advanced phase of the disease [1217].

In ET, around 60% of patients present the JAK2V617F mutation and approximately 8% of JAK2V617-negative ET patients show MPL mutations, but there still remains a significant percentage of ET patients without a molecular marker. According to our results, analysis of the TET2 and the ASXL1 genes would increase in an additional 8% the percentage of ET patients with a molecular marker of clonality. The incorporation of these determinations in the diagnostic routine is labor intensive if classic Sanger sequencing is to be used, but the implementation of new technologies such as next generation sequencing, can be helpful to incorporate the mutational analysis of these genes in the molecular diagnosis of MPN. These new high-throughput technologies can analyze a large number of genes at the same time with a sensitivity of down to 1%, allowing the detection of mutations in small clones. Recently, several studies have shown that next generation sequencing is a useful tool to detect mutations in myeloid malignancies. The extensive information obtained by these techniques can be used to characterize the molecular pattern of each patient and to better understand the pathogenesis of the disease and its evolution [32, 33]. In this regard, Jädersten et al. showed that the detection of TP53 mutations in 18% of low-risk MDS were associated with a higher risk of leukemic transformation. In some patients, these mutations were already detectable at an early phase of the disease, with a low mutant allelle burden, showing that early detection of mutations could be of prognostic value [34]. Accordingly, next generation sequencing can provide a large amount of molecular information that can be used for a more accurate diagnosis, and potentially as prognostic factors.

Finally, the present study is, to the best of our knowledge, the first one that has assessed the correlation between the JAK2 46/1 haplotype and mutations in TET2 and ASXL1 genes. In this regard, we could not establish any association between the presence of mutations in these genes and the JAK2 46/1 haplotype.

In summary, TET2 and ASXL1 pathogenic mutations are found in a low percentage (8%) of MPN lacking JAK2 and MPL mutations, whereas IDH1, IDH2, and c-CBL mutations are not detected in this subset of patients.

Acknowledgments

This work was supported by grants from the Spanish Health Ministry “Fondo de Investigación Sanitaria” EC 07/90791, PI10/01807, Instituto de Salud Carlos III FEDER (RD09/0076/00036), and the “Xarxa de Bancs de Tumors sponsored by Pla Director d’Oncologia de Catalunya (XBTC)”. Luz Martínez-Avilés is recipient of a fellowship from the “Comissionat per a Universitats i Recerca del department d’ Innovació, Universitats i Empresa de la Generalitat de Catalunya i del Fons Social Europeu.”

Supplementary material

277_2011_1330_MOESM1_ESM.doc (33 kb)
Supplementary Table 1(DOC 33 kb)

Copyright information

© Springer-Verlag 2011