IDH1 mutations in patients with myelodysplastic syndromes are associated with an unfavorable prognosis

Abstract

Background Myelodysplastic syndromes are a heterogeneous group of hematopoietic stem cell disorders with a high propensity to transform into acute myeloid leukemia. Heterozygous missense mutations in IDH1 at position R132 and in IDH2 at positions R140 and R172 have recently been reported in acute myeloid leukemia. However, little is known about the incidence and prognostic impact of IDH1 and IDH2 mutations in myelodysplastic syndromes.Design and Methods We examined 193 patients with myelodysplastic syndromes and 53 patients with acute myeloid leukemia arising from myelodysplastic syndromes for mutations in IDH1 (R132), IDH2 (R172 and R140), and NPM1 by direct sequencing.Results We found that mutations in IDH1 occurred with a frequency of 3.6% in myelodysplastic syndromes (7 mutations in 193 patients) and 7.5% in acute myeloid leukemia following myelodysplastic syndromes (4 mutations in 53 patients). Three mutations in codon R140 of IDH2 and one mutation in codon R172 were found in patients with acute myeloid leukemia following myelodysplastic syndromes (7.5%). No IDH2 R140 or R172 mutations were identified in patients with myelodysplastic syndromes. The presence of IDH1 mutations was associated with a shorter overall survival (HR 3.20; 95% CI 1.47–6.99) and a higher rate of transformation into acute myeloid leukemia (67% versus 28%, P=0.04). In multivariate analysis when considering karyotype, transfusion dependence and International Prognostic Scoring System score, IDH1 mutations remained an independent prognostic marker in myelodysplastic syndromes (HR 3.57; 95% CI 1.59–8.02; P=0.002).Conclusions These results suggest that IDH1 mutations are recurrent molecular aberrations in patients with myelodysplastic syndromes, and may become useful as a poor risk marker in these patients. These findings await validation in prospective trials.

Introduction

Myelodysplastic syndromes (MDS) are a heterogeneous group of hematopoietic stem cell disorders. They are characterized by two cardinal features: ineffective hematopoiesis leading to bone marrow failure and a propensity to transform into acute myeloid leukemia (AML). While significant progress has been made in revealing cytogenetic and molecular changes in AML1,2 less is known about the molecular changes that can lead to MDS. A number of non-specific recurrent mutations have been described in MDS including NRAS,3 TP53,4 RUNX1,5,6 and FMS,7,8 with the recent additions of TET29,10 and ASXL1.11,12 However, given the heterogeneity of the disease it is of major importance to characterize MDS patients better at the molecular level, and to evaluate the prognostic relevance of new mutations.

While clinical scoring systems such as the World Health Organization (WHO) adapted Prognostic Scoring System (WPSS) and the International Prognostic Scoring System (IPSS)13,14 can help to stratify patients according to their risk of death and leukemic progression, good molecular prognostic markers are still lacking.15

Isocitrate dehydrogenase 1 gene (IDH1) encodes for the protein isocitrate dehydrogenase 1, an enzyme that participates in the citric acid cycle. It catalyzes the carboxylation of isocitrate to alpha-ketoglutarate. Recurrent mutations in IDH1 have been described in 12% of patients with glioblastomas16 as well as in 70% of patients with WHO grade II and III astrocytomas and oligodendrogliomas.17 In these entities IDH1 mutations affect one single amino acid residue in position 132 leading to a switch from arginine to histidine (R132H).17 At a lower frequency, mutations in IDH2 affecting the analogous amino acid (R172) have also been described in oligodendrogliomas and astrocytomas.17 Interestingly, the same IDH1 mutation at position 132 was found when sequencing the genome of a patient with AML-M1.18 Since this first description of IDH1 mutations in AML several reports have confirmed that IDH1 mutations occur in patients with cytogenetically normal AML with a frequency of 5.5–11%.19–21 Additionally, a strong association between IDH1 mutations with intermediate risk karyotype and concurrent NPM1 mutations was found.19–21 A recent study in AML patients showed a low frequency of mutations of codon R172 of IDH2.21 A novel mutation in codon R140 of IDH2 was identified in two patients with leukemic transformation of myeloproliferative neoplasms as well as in patients with AML.22–24

Little is known about the incidence and prognostic impact of IDH1 and IDH2 mutations in MDS patients. In this study, we examined the DNA of 193 patients with MDS for the presence of mutations in IDH1 (R132), IDH2 (R172 and R140), and NPM1 by direct sequencing, and evaluated the prognostic impact of these mutations.

Design and Methods

Patients

Samples from 193 patients with MDS and 53 with AML with a prior history of MDS were collected at the time of enrollment in clinical trials. All patients with MDS were enrolled in multicenter treatment trials that investigated the use of all-trans retinoic acid,25 antithymocyte globulin,26 deferasirox,27 lenalidomide, or thalidomide for the treatment of MDS while demethylating agents were not employed in this cohort of patients.

The diagnosis of AML arising from MDS was based on history, cytogenetics and morphology. All patients with secondary AML were treated within a trial of 5-aza-2′-deoxycytidine (decitabine) (Clinical Trials Identifier NCT00866073) or multicenter treatment trials AML SHG 0295 and AML SHG 0199 (ClinicalTrials Identifier NCT00209833). Details of the treatment protocols have been reported previously.28,29 Clinical and hematologic data were recorded after patients had given their informed consent in accordance with the Declaration of Helsinki, and the scientific analysis of the samples was approved by the institutional review board of Hannover Medical School (n. 2467). According to the WHO classification, patients were classified as having refractory anemia (n=38), refractory anemia with ringed sideroblasts (n= 20), MDS with isolated del(5q) (n= 18); refractory cytopenia with multilin-eage dysplasia (n=30); refractory anemia with excess blasts-1 (n=22), refractory anemia with excess blasts-2 (n=31) and MDS-unclassifiable (n=7). Information on WHO subtype was not available for 27 patients. The IPSS stratification was low in 39 patients, intermediate-1 in 57, intermediate-2 in 38, and high in 13 (information on IPSS score was not available for 46 patients). Follow-up samples were available for 35 patients (median follow-up, 226 days; range, 13–988 days). Seven patients had had samples taken while they had MDS and also after their disease had progressed to AML (Figure 1). Follow-up information was available for 153 of the 193 patients with MDS. The follow-up information was updated by means of clinic visits as well as telephone calls to patients, their doctors, and local registry offices.

Cytogenetic analysis and mutation analysis of IDH1/2 and NPM1

Cytogenetic analysis was performed centrally by G- and R-banding analysis. Mutation analysis was performed as described previously.30 Mononuclear cells from patients’ samples were enriched by Ficoll density gradient centrifugation and were stored at −196°C in liquid nitrogen until use. Genomic DNA was extracted from samples using the All Prep DNA/RNA Kit (Qiagen, Hilden, Germany) according to the manufacturer’s recommendations. The genomic region that spans the wild-type R132 of IDH1 (exon 4) was amplified using polymerase chain reaction (PCR) with the following primers: 5′TGTGTTGAGATGGACGCC-TATTTG and 5′TGCCACCAACGACCAAGTCA as previously described.16 The following primers were used for amplification of the genomic region that spans wild-type R140 and R172 of IDH2 (exon 4) using PCR: 5′GGGGTTCAAATTCTGGTTGA and 5′CTAGGCGAGGAGCTCCAGT. PCR fragments were directly sequenced, and were analyzed using Sequencing Analysis 5.3.1 software (Applied Biosystems, Darmstadt, Germany) and Vector NTI Advance 10 software (Invitrogen, Karlsruhe, Germany). Point mutations were confirmed in independent experiments. Genomic DNA was analyzed for NPM1 mutations as previously described.31

Statistical analysis

Overall survival end-points, measured from the date of first sample collection, were death (failure) and alive at last follow-up (censored). Event-free survival end-points, measured from the date of first sample collection, were progression to AML or death (failure), and alive without progression of disease to AML at time of last follow-up (censored). Progression to AML was defined according to the 2008 WHO classification The median follow-up times for overall survival and event-free survival were calculated according to the method of Korn.32 The primary analysis was performed on overall survival. Sensitivity analysis was performed on event-free survival, and the results are displayed for exploratory purposes. Pair-wise comparisons were performed using two-sided Kolmogorov-Smirnov tests for continuous variables and by two-sided χ tests for categorical variables, and are provided for exploratory purposes. For multivariate analysis, a Cox proportional hazards model was constructed for overall survival and event-free survival, adjusting for potential confounding covariates.33 Variables considered for inclusion in the model were karyotype (favorable versus intermediate risk versus high risk), IPSS (low/int-1 versus int-2/high), transfusion dependence (yes versus no), ferritin level (above or below 1000 μg/L), age (below versus above median), number of therapies (best supportive care versus at least one other treatment), and IDH1 mutation status. Variables with a P value of 0.05 or less in the univariate analysis for overall survival or event-free survival were included in the model. The two-sided level of significance was set at P less than 0.05. The statistical analyses were performed with the statistical software package SPSS 17.0 (SPSS Science, Chicago, IL, USA).

Figure 1.Description of patients’ samples. MDS, myelodysplastic syndrome; AML, acute myeloid leukemia

Results

Mutation status of IDH1/2 in patients with myelodys-plastic syndromes

Among 193 patients with MDS, seven patients (3.6%) had a heterozygous mutation in codon 132 of IDH1. Six patients showed conversion of CGT to TGT leading to an Arg132Cys substitution and one patient had a CGT to CAT conversion leading to an Arg132His substitution (Table 1). Follow-up samples were available for 35 of these 193 patients (median follow-up, 226 days) and examined in addition to the first sample. One of these patients had an identical mutation in IDH1 codon 132 at the time of MDS (refractory anemia) and in the follow-up sample at the time of AML transformation 2.7 years later. In the remaining 34 cases with follow-up samples, including six patients for whom we had a follow-up sample at the time of progression to AML, no IDH1 mutation was found. Mutations in codon 140 or 172 of IDH2 were not identified in any of the 193 MDS patients. Only one NPM1 mutation was found in a MDS patient with wild-type IDH1 and IDH2.

Mutation status of IDH1/2 in patients with acute myeloid leukemia following myelodysplastic syndrome

Among the 53 AML patients with a previous history of MDS, four patients (7.5%) had IDH1 mutations in codon 132. Two patients showed conversion of CGT to CAT, one patient had conversion of CGT to TGT, and one patient had conversion of CGT to AGT leading to an Arg132Ser substitution. Three out of 53 patients with AML following MDS were found to have mutations in codon R140 of IDH2. All showed a conversion of CGG to CAG leading to an arginine to glutamine substitution. One patient with AML following MDS had a mutation in codon 172 of IDH2 causing a conversion of AGG to AGT which leads to an Arg172Ser substitution (Table 2). Thus, 15% of AML patients with a history of MDS had mutated IDH1 or IDH2. Mutated NPM1 was found in three of 33 studied patients with AML following MDS; of these, one also had mutated IDH1.

Patients’ characteristics in relation to IDH1/2 mutations

The clinical and hematologic characteristics of patients with and without mutations are compared in Table 3. There were no differences in age, sex, WHO classification, karyotype, bone marrow blasts, hemoglobin, transfusion dependence, ferritin, IPSS score, or number of treatments between the patients with IDH1 mutations and those with the wild-type gene.

Prognostic impact of IDH1 mutations

The prognostic impact of IDH1 mutations was evaluated in MDS patients for whom follow-up information was available (n=153). The median follow-up of patients alive was 3 years. In univariate analysis the overall survival of patients with IDH1 mutation was significantly shorter than that of patients with wild-type IDH1 (HR 3.20; 95% CI 1.47–6.99; Figure 2A). The 2-year survival rate was 52% and 14% for MDS patients with wild-type and mutated R132 IDH1, respectively. In univariate analysis IPSS score (high/int-2 versus low/int-1; HR 2.05; 95% CI 1.28–3.74; P=0.003), transfusion dependence (dependent versus independent; HR 3.62; 95% CI 1.65–7.93; P=0.001) and karyotype (high versus intermediate versus low risk; HR 1.91; 95% CI 1.44–2.52; P<0.001) were also identified as prognostic factors for overall survival. In multivariate analysis including IDH1 mutation status, IPSS score, transfusion dependence, and karyotype, the presence of an IDH1 mutation was found to be an independent unfavorable prognostic factor for overall survival (Table 4). Among the 186 MDS patients with wild-type IDH1, information about progression to AML was available for 145 patients. Among these, 41 developed AML (28.4%). Of the seven MDS patients with IDH1 mutations, information about progression to AML was available for six patients; of these, four developed documented AML (67%, P=0.04). Among the remaining two patients without documented progression to AML, one patient had a leukocyte count of 70.2×10/L with 1% blasts in the differential count at the time of death. Due to critical illness a bone marrow biopsy was not obtained to confirm the diagnosis of AML, and the patient died shortly thereafter. In univariate analysis patients with mutations in codon R132 of IDH1 had significantly lower event-free survival (HR 2.37; 95%CI 1.10–5.11) (Figure 2B). Furthermore, in univariate analysis for event-free survival, IPSS score (high/int-2 versus low/int-1; HR 2.00; 95% CI 1.39–2.86; P<0.001), transfusion dependence (dependent versus independent; HR 2.29; 95% CI 1.46–3.52; P<0.001), and karyotype (high versus intermediate versus low risk; HR 1.68; 95% CI 1.33–2.12; P<0.001) were identified as prognostic factors for event-free survival. In multivariate analysis including IDH1 mutation status, IPSS score, transfusion dependence, and karyotype, the presence of IDH1 mutations was found to be an independent unfavorable prognostic factor for event-free survival (Table 4).

Table 1.Details of IDH1 mutations in patients with myelodysplastic syndromes.

Table 2.Details of IDH1/2 mutations in acute myeloid leukemia patients with a previous history of myelodysplastic syndromes.

Discussion

In the present study we found that IDH1 mutations were a recurring molecular aberration in MDS patients, occurring with a frequency of 3.6%. Moreover, mutated IDH1 was associated with a high rate of leukemic transformation, and poor event-free and overall survival rates. The rate of IDH1 R132 mutations in MDS patients was lower than the rate in patients with AML arising from MDS or the reported rate in patients with de novo AML.18,20,34 We did not identify any IDH2 R172 or R140 mutations in MDS patients. However, we demonstrated that mutations of IDH2 occur in AML patients with a prior history of MDS. We, therefore, showed that IDH1 mutations are rare but recurrent molecular aberrations in MDS patients, and establish IDH1/2 mutations as one of the most frequent mutations in AML arising from MDS (15%). In our cohort of MDS patients, NPM1 mutations were not identified in patients with IDH1 mutations, and the rate of NPM1 mutations was very low. Given the low frequency of NPM1 mutations in our cohort of patients we could not evaluate an association of NPM1 and IDH mutations previously found in AML patients.

Figure 2.Kaplan-Meier curves for overall and event-free survival. (A) Overall survival in MDS patients with mutated (n= 7) and unmutated (n=146) IDH1 (log-rank test, P=0.002). (B) Event-free survival in MDS patients (n=153) with mutated (n= 7) and unmutated (n=146) IDH1 (log-rank test, P=0.02).

Table 3.Comparison of clinical and molecular characteristics of MDS patients with mutated or wild-type IDH1.

In this study, mutated IDH1 was an independent unfavorable prognostic marker for both event-free survival and overall survival of patients with MDS. Different treatment regimens did not influence prognosis in our cohort (data not shown), suggesting that treatment differences between patients were unlikely to have had a confounding effect on our analysis. The present study illustrates a potential new unfavorable prognostic marker in MDS patients, especially in patients with normal cytogenetics, which may become useful for treatment stratification in the future. Additional studies in larger cohorts of patients are warranted.

Table 4.Cox regression analysis for overall survival and event-free survival in MDS patients with mutated (n=7) or unmutated (n=146) IDH1.

The intriguing finding that IDH1 and IDH2 mutations occur in the leukemic transformation of myeloproliferative neoplasms, but not in patients in chronic-phase poly-cythemia vera or essential thrombocythemia,22 suggests that these mutations play an important role in leukemo-genesis. Our results indicate that MDS patients with mutated IDH1 undergo a high rate of leukemic transformation, which is in accordance with the data in myeloproliferative neoplasms.22 Interestingly, no IDH2 mutations were observed in MDS patients while the incidence of IDH2 mutations in patients with AML arising from MDS was found to be 7.5%. Functional studies may clarify whether mutations in IDH1 and IDH2 have distinct effects on leukemic progression from MDS to AML.

In summary, we identified IDH1 mutations of amino acid 132 in 3.6% of MDS patients, and found a strong correlation of mutated IDH1 with unfavorable outcome in these patients. Because of the low frequency of IDH1 mutations occurring in MDS the prognostic impact of the mutation should be confirmed in larger groups of uniformly treated MDS patients and put in context with other novel markers such as TET2, ASXL1 and RUNX1. Our study also provides evidence that mutations of codons R140 and R172 of IDH2 occur in patients with AML arising from MDS. Mutation analysis of IDH1 in MDS patients may become useful for risk and treatment stratification in the future.

Acknowledgments

we are indebted to all patients and contributing doctors. We thank Kerstin Görlich, Elvira Lux, Sylvia Horter, Susanne Luther, Jana König, and Maren Herten for their excellent support in sample and data acquisition.

Footnotes

• Funding: this study was supported by the Fellowship 2007/04 awarded to MH by the European Hematology Association, the Hannelore-Munke Fellowship (MH), and the Dieter-Schlag Stiftung (MH, FT).
• 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 March 22, 2010.
• Revision received April 23, 2010.
• Accepted May 17, 2010.