Prognostic impact of the isocitrate dehydrogenase 1 single-nucleotide polymorphism rs11554137 in malignant gliomas

Authors

  • Xiao-Wei Wang MD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    Search for more papers by this author
  • Blandine Boisselier MSc,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    Search for more papers by this author
  • Marta Rossetto MD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    Search for more papers by this author
  • Yannick Marie PhD,

    1. Genotyping and Sequencing Platform, ICM, Paris, France
    Search for more papers by this author
  • Ahmed Idbaih MD, PhD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    4. Department of Neurology, Pitie-Salpetriere Hospital, APHP, Paris, France
    Search for more papers by this author
  • Karima Mokhtari MD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    4. R. Escourolle Laboratory of Neuropathology, Pitie-Salpetriere Hospital, APHP, Paris, France
    Search for more papers by this author
  • Konstantinos Gousias MD,

    1. Department of Neurosurgery, University Clinic Bonn, Bonn, Germany
    Search for more papers by this author
  • Khê Hoang-Xuan MD, PhD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    4. Department of Neurology, Pitie-Salpetriere Hospital, APHP, Paris, France
    Search for more papers by this author
  • Jean-Yves Delattre MD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    4. Department of Neurology, Pitie-Salpetriere Hospital, APHP, Paris, France
    Search for more papers by this author
  • Matthias Simon MD,

    1. Department of Neurosurgery, University Clinic Bonn, Bonn, Germany
    Search for more papers by this author
  • Marianne Labussière PharmD, PhD,

    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    Search for more papers by this author
  • Marc Sanson MD, PhD

    Corresponding author
    1. CRICM, Pierre and Marie Curie University, Paris, France
    2. INSERM U 975, Paris, France
    3. UMR 7225, CNRS, Paris, France
    4. Department of Neurology, Pitie-Salpetriere Hospital, APHP, Paris, France
    • Fédération de Neurologie Mazarin, Groupe Hospitalier Pitié-Salpêtrière, 75651, Paris Cedex 13, France

    Search for more papers by this author
    • Fax: (011) 33 1 42 16 03 75


  • We are indebted to the patients who agreed to participate in this study, to Anne-Marie Lekieffre and Muriel Brandel for their assistance with the study, and to the medical staff for collecting blood samples.

Abstract

BACKGROUND.

The IDH1 gene, which encodes isocitrate dehydrogenase 1, is frequently mutated in gliomas and acute myeloid leukemia. The single-nucleotide polymorphism (SNP) (reference SNP no. rs11554137:C>T) located on IDH1 codon 105 has been associated with a poor outcome in patients with acute myeloid leukemia but has not been investigated in patients with gliomas.

METHODS.

The IDH1 codon 105 SNP was genotyped first in a series of 952 patients with grade 2 through 4 gliomas and was correlated with outcomes and tumor genomic profile. Then, it was genotyped in 2 validations sets of 306 patients with glioblastoma (GBM) and 591 patients with glioma.

RESULTS.

The minor allele codon 105 glycine (GGT) SNP (IDH1105GGT) was identified in 98 of 952 patients (10.3%) and was not associated with the codon 132 (IDH1132) mutation. Patients who had GMB with the IDH1105GGT variant had a poorer outcome than patients without the variant (median overall survival [OS], 10.7 months vs 15.5 months; P = .001; median progression-free survival [PFS], 6.4 months vs 8.5 months; P = .003). The prognostic impact was confirmed in an independent validation set of 306 GBMs from the same center (median PFS, 6.8 months vs 9.7 months; P = .006; median OS, 13.9 months vs 18.8 months; P = .0187). In the second validation cohort (591 grade 2-4 gliomas), a significant association was observed between IDH1105GGT and an adverse prognosis for the overall series and for patients with World Health Organization grade 3 gliomas, but the difference did not reach significance in patients with GBM.

CONCLUSIONS.

Taken together, the current data strongly suggested an association between the SNP rs11554137:C>T polymorphism and adverse outcomes in patients with malignant glioma. A single-nucleotide polymorphism (SNP) located on codon 105 of the isocitrate dehydrogenase 1 (IDH1) gene (reference SNP rs11554137) is analyzed in 3 independent series of patients with gliomas. The SNP rs11554137 is independent of the occurrence of somatic mutation on IDH1 codon 132, but, per se, has a prognostic impact in malignant gliomas. Cancer 2013. © 2012 American Cancer Society.

INTRODUCTION

Mutations in the isocitrate dehydrogenase 1 (IDH1) gene (Mendelian Inheritance in Man gene no. 147700) affect nearly 40% of gliomas and are associated with a better outcome.1–3 IDH1 encodes isocitrate dehydrogenase 1, which catalyzes the oxidative decarboxylation of isocitrate to α-ketoglutarate. Mutations that affect codon 132 lead not only to a dramatic decrease in IDH1 activity but also to a gain of enzyme function for the nicotinamide adenine dinucleotide phosphate (NADPH)-dependent reduction of α-ketoglutarate to 2-hydroxyglutarate, which accumulates in the IDH1-mutated cells.4 IDH1 mutations are inversely associated with tumor grade, affecting nearly 75% of World Health Organization (WHO) grade 2, 50% of grade 3, and only 5% of primary grade 4 gliomas2; whereas 80% of secondary glioblastomas have IDH1 mutations.1,5

Along with gliomas, IDH1 mutations also affect 8% to 15% of acute myeloid leukemias (AMLs) (mostly with normal karyotype),6,7 cartilaginous neoplasms (chondromas and chondrosarcomas),8 and cholangiocarcinomas.9 In contrast to patients with glioma, IDH1 mutations in patients with AML who have a normal karyotype are associated with a poor prognosis.7 Recently, a synonymous single-nucleotide polymorphism (SNP), reference SNP 11554137:C>T (rs11554137:C>T), which is located in codon 105 in the same exon as the IDH1 codon 132 arginine (IDH1R132) mutation, has been described in patients with AML and has been identified as an adverse prognostic factor.10 In the current study, we screened 952 gliomas samples for presence of the IDH1 SNP rs11554137:C>T, and we correlated the findings with IDH1 mutation status, genomic profile (deletion of cyclin-dependent kinase inhibitor 2A [p16], amplification of epidermal growth factor receptor [EGFR], loss of the long arm of chromosome 10 [10q loss], and codeletion of the short arm of chromosome 1 and the long arm of chromosome 19 [chromosome 1p/19q codeletion]), methyl guanine methyl transferase (MGMT) promoter methylation status, and outcome. In addition, we investigated the prognostic role of the rs11554137:C>T polymorphism in 2 independent patient cohorts (a Paris validation series of 309 patients with glioblastoma and a Bonn series of 591 patients with WHO grade 2-4 gliomas).

MATERIALS AND METHODS

Patients

Two series of patients with glioma were selected from the neuro-oncology database at the Pitie-Salpetriere Hospital (Paris, France): The first 1 included 952 patients with a histologic diagnosis of glioma (grade 2-4), and the second was an additional series of 309 patients with glioblastoma multiforme (GBM). We also analyzed an independent data set of 591 grade 2 through 4 gliomas that were collected in the Department of Neurosurgery at the University of Bonn (Bonn, Germany) from 1996 to 2007. Collection of blood samples and clinicopathologic information was undertaken with informed consent and relevant ethical board approval in accordance with the tenets of the Declaration of Helsinki.

Tumor and Blood DNA Analysis

Tumor and blood DNA were extracted from frozen tumors using the QIAmp DNA Mini Kit, as described by the manufacturer (Qiagen, Hilden, Germany). IDH1 SNP rs11554137:C>T was characterized in blood DNA using a Taqman SNP Genotyping Assay (no. C_42648573_10; Applied Biosystems, Foster City, Calif).

The genomic region of the IDH1 gene (exon 4) that contains the mutation hotspot codon arginine 132 (Arg132 or R132) and the SNP rs11554137:C>T was amplified from tumor DNA, as previously described,2 using polymerase chain reaction (PCR) with the following primers: IDH1 forward, 5-TGTGTTGAGATGGACGCCTATTTG; and IDH1 reverse, 5-ACTGAACCAGCAACCACCGT. Purified PCR fragments were directly sequenced using the BigDye Terminator Sequencing Kit (version 3.1; Applied Biosystems), and both forward and reverse chains were analyzed on an ABI prism 3730 DNA analyzer (Perkin Elmer, Waltham, Mass). The comparative genomic hybridization (CGH) array was performed as previously described11; or, if the CGH array was not available, then we used loss of heterozygosity analysis (to assess loss of the long arm of chromosome 10 [10q loss]) and quantitative PCR (to determine p16 and EGFR status).12 MGMT promoter methylation status was determined as previously described.13

Independent Data Sets

After the analysis of these 952 tumors, we studied 2 independent patient series to confirm a potential prognostic role for rs11554137:C>T. In both series, IDH1 SNP rs11554137:C>T was characterized using a Taqman SNP Genotyping Assay (no. C_42648573_10; Applied Biosystems).

Statistical Analysis

The independence of alleles (Hardy-Weinberg equilibrium) was confirmed using the chi-square test with 1 degree of freedom for each polymorphism. The chi-square test (or the Fisher exact test when 1 subgroup included <5) was used to compare the genotype distribution. Comparison of IDH1 expression was performed using unpaired t tests.

Overall survival (OS) was calculated from the date of diagnosis to the date of either death or last follow-up. Patients who remained alive at last follow-up were considered censored events in our analysis. Progression-free survival (PFS) was calculated from the date of diagnosis to the date of either recurrence or last follow-up. Patients who were recurrence free at last follow-up were considered censored events in our analysis. To identify clinical and/or genomic factors related to OS (or PFS), survival curves were calculated according to the Kaplan-Meier method, and differences between curves were assessed using the log-rank test. Variables with significant P values (< .05) were used to build a multivariate Cox model.

RESULTS

Mutational Analysis of IDH1

Tumor DNA samples were analyzed from 952 adult patients with glioma (239 WHO grade 2 tumors, 264 grade 3 tumors, and 449 grade 4 tumors) for the presence of IDH1 rs11554137:C>T. The sex ratio was 1.3 (533 men and 419 women), and the median patient age was 50.5 years (range, 16.1–86.5 years) (Table 1).

Table 1. Tumor Population Characteristics
 Glioma by Grade
 234
No. of Patients239264449
  1. Abbreviations: OS, overall survival; PFS, progression-free survival.

Sex   
 Ratio of men to women1.21.31.3
 No. of men132149252
 No. of women107115197
Age at diagnosis, y   
 Median38.147.058.5
 Range16.1–76.719.1–84.118.2–86.5
Median survival, mo   
 OS110.036.315.0
 PFS37.919.89.8

In total, 101 patients were positive for SNPs, including 98 who were heterozygous (10.3%) and 3 (0.3%) who were homozygous for the minor allele glycine (GGT) on codon 105 (IDH1105GGT) (Table 2). The distribution of allele frequencies in our population met Hardy-Weinberg equilibrium (P = .92). Allele frequencies did not differ between the 3 tumor grades: Heterozygous IDH1 SNP rs11554137:C>T was identified in 25 (10.4%) WHO grade 2 gliomas, in 27 (10.2%) grade 3 gliomas, and in 46 (10.2%) grade 4 gliomas. The variant allele was more frequent in women (41 men and 57 women; ratio, 0.7 vs 1.4; P = .004).

Table 2. Characteristics of Patients With or Without the Isocitrate Dehydrogenase 1 Single-Nucleotide Polymorphism rs11554137:C>T According to Glioma Grade
 Glioma Grade
 234
IDH1 rs11554137 StatusGGC/GGCGGC/GGT and GGT/GGTGGC/GGCGGC/GGT and GGT/GGTGGC/GGCGGC/GGT and GGT/GGT
  1. Abbreviations: 1p, short arm of chromosome 1; 10q, long arm of chromosome 10; 19q, long arm of chromosome 19; Arg132His, substitution of arginine by histidine at codon 132; EGFR, epidermal growth factor receptor; GGC and GGT, glycine genotypes; IDH1, isocitrate dehydrogenase 1; MGMT, methyl guanine methyl transferase; NR, not reached; OS, overall survival; p16, cyclin-dependent kinase inhibitor 2A; PFS, progression-free survival; R132, codon 132 arginine.

Total no. of patients213262352940346
Sex      
 Ratio of men to women1.40.51.40.81.40.8
 Men12391361323220
 Women9017991617126
Age at diagnosis, y      
 Median38.236.847.052.258.559.8
 Range16.1–77.017.3–64.819.1–84.123.7–73.718.2–86.523.4–79.7
Prognosis      
 Median OS, mo110.0NR36.531.615.510.7
 Median PFS, mo33.335.916.514.58.56.4
Genetic alterations: Percentage of patients (no./total no.)      
 IDH1 R132 mutation65.1 (138/212)72 (18/25)56.4 (132/234)55.6 (15/27)6 (24/402)0 (0/46)
 Arg132His91.3 (126/138)88.9 (16/18)92.4 (122/132)80 (12/15)91.7 (22/24)0 (0/46)
 Other 132 mutations8.7 (12/138)11.1 (2/18)7.6 (10/132)20 (3/15)8.3 (2/24)0 (0/46)
 EGFR amplification0.5 (1/205)0 (0/25)14.7 (34/232)7.4 (2/27)38.1 (139/365)21.4 (9/42)
 Chromosome 10q loss15.1 (30/199)16.7 (4/24)36.3 (81/223)36 (9/25)80.4 (263/327)81.1 (30/37)
 1p/19q Codeletion27.3 (36/132)31.6 (6/19)35.6 (67/188)20 (4/20)0 (0/254)0 (0/28)
 MGMT methylation87 (80/92)92.3 (12/13)65.3 (79/121)53.8 (7/13)53.2 (124/233)51.8 (14/27)
 Homozygous p16 deletion6.3 (12/192)20 (5/25)12.1 (28/232)19.2 (5/26)39.4 (143/363)31 (13/42)

Of 946 tumors, 330 (34.9%) had codon 132 mutations, including 157 (65.7%). WHO grade 2 gliomas, 149 (56.4%) grade 3 gliomas, and 24 (5.3%) grade 4 gliomas. An arginine-to-histidine substitution at codon 132 (Arg132His; CGT→CAT) was identified in 300 tumors (91%), and other mutations were identified in 30 tumors (9%). We observed no association between the SNP rs11554137:C>T polymorphism and the global incidence of codon 132 mutations: Of all 327 mutated tumors, 33 were heterozygous for the rs11554137:C>T minor allele (10%) compared with 65 of 619 nonmutated tumors (10.5%). We also analyzed the relation between SNP rs11554137:C>T and each type of IDH1 mutation: Thirty of 300 tumors with the Arg132His (CAT) mutation had the variant SNP rs11554137:C>T versus 6 of 30 tumors with other IDH1R132 mutations (10% vs 20%; P = .12), including 1 of 5 tumors with a CGT→AGT mutation (an arginine-to-serine substitution [Arg132Ser]), 1 of 3 tumors with a CGT→CTT mutation (an arginine-to-leucine substitution [Arg132Leu]), 2 of 12 tumors with a CGT→GGT mutation (an arginine-to-glycine substitution [Arg132Gly]), and 2 of 10 tumors with a CGT→TGT mutation (an arginine-to-cysteine substitution [Arg132Cys]). Of the 26 IDH2 mutations that were identified in the whole series, only 1 had the variant SNP rs11554137:C>T.

Genomic Profiling

Genomic analysis was performed on the 896 tumor samples for which all data were available, including 643 samples that were analyzed by CGH array. MGMT promoter methylation status was determined in 499 samples (Table 3). There was an inverse correlation between the SNP rs11554137:C>T polymorphism and EGFR amplification (11 of 94 samples vs 174 of 802 samples; P = .02); however, no correlation was observed with p16 deletion, chromosome 10q loss, codeletion of the short arm of chromosome 1 and the long arm of chromosome 19 (1p/19q codeletion), or MGMT methylation status.

Table 3. Multivariate Cox Proportional Hazards Regression Model Analysis of Survival in Patients Glioblastoma
VariableHR95% CIP
  1. Abbreviations: CI, confidence interval; HR, hazard ratio; IDH1, isocitrate dehydrogenase 1 gene; rs11554137, reference single-nucleotide polymorphism 11554137; SNP, single-nucleotide polymorphism.

Progression-free survival   
 Age1.411.15–1.7.0008
 IDH1 SNP rs115541371.641.19–2.27.003
 IDH1 mutation0.610.39–0.97.04
Overall survival   
 Age1.601.28–2.00.0000
 IDH1 SNP rs115541371.651.17–2.33.004
 IDH1 mutation0.710.43–1.18.19

Prognostic Impact of IDH1 Single-Nucleotide Polymorphism rs11554137

We investigated the prognostic impact of IDH1 SNP rs11554137:C>T in grade 2, grade 3, and grade 4 gliomas (Table 2, Fig. 1). In grade 4 gliomas, both PFS and OS were shorter for patients who were positive for the SNP compared with those who were negative for the SNP (median OS, 10.7 months vs 15.5 months, respectively; P = .001; median PFS, 6.4 months vs 8.5 months, respectively; P = .003). We observed a similar trend in grade 3 gliomas (median OS, 31.6 months vs 36.5 months, respectively; median PFS, 14.5 months vs 16.5 months, respectively; P value, nonsignificant) but not in grade 2 gliomas.

Figure 1.

Survival curves illustrate (Left) progression-free survival and overall survival in patients with glioblastoma according to isocitrate dehydrogenase 1 (IDH1) reference single-nucleotide polymorphism rs11554137:C>T status.

The following parameters were entered as candidate variables into a multivariate Cox proportional hazards regression model analysis of OS for patients with glioblastoma: age, EGFR amplification status, IDH1132 mutation status, and IDH1 SNP rs11554137:C>T mutation status (Table 3). For both PFS and OS, the IDH1 SNP rs11554137:C>T mutation was a strong and independent predictor of outcome (PFS: hazard ratio, 1.64; 95% confidence interval [CI], 1.1873–2.2651; P = .003; OS: hazard ratio, 1.65; 95% CI, 1.1725–2.3342; P = .004).

Impact of the IDH1 Single-Nucleotide Polymorphism rs11554137:C>T on Prognosis in an Independent Data Set of 306 Glioblastomas

To confirm the apparent major prognostic impact of IDH1 SNP rs11554137:C>T in glioblastomas, we analyzed this polymorphism in an independent series of 306 patients with glioblastomas who were treated at the Pitie-Salpetriere Hospital (Table 4). In this validation series, the presence of IDH1 SNP rs11554137:C>T was identified in 35 patients (11.4%), including 33 heterozygotes and 2 homozygotes, who were considered as a single group for subsequent survival analyses. Both PFS (6.8 months vs 9.7 months; P = .006) and OS (13.9 months vs 18.8 months; P = .02) were reduced significantly for SNP-positive patients (Fig. 2). In 125 of these patients who had tumor DNA available, the inverse association between IDH1 SNP rs11554137:C>T variant and tumor EGFR amplification was not confirmed (4 of 15 patients with the SNP vs 32 of 110 patients without the SNP had an EGFR amplification in tumor DNA; P value, nonsignificant).

Figure 2.

Survival curves illustrate (Left) progression-free survival and (Right) overall survival in the validation series of 306 patients with glioblastoma according to isocitrate dehydrogenase 1 (IDH1) reference single-nucleotide polymorphism rs11554137:C>T status.

Table 4. Characteristics of the Patients Included in the First Validation Series: Paris
 Genotype 
 GGC/GGCGGC/GGT or GGT/GGTP
  1. Abbreviations: GGC and GGT, glycine genotypes; OS, overall survival; PFS, progression-free survival; NS, nonsignificant.

Total no. of patients27135 
Sex   
 Ratio of men to women1.71.9NS
 No. of men16923 
 No. of women10212 
Age at diagnosis, y   
 Median57.860.0NS
 Range19.9–86.536.5–84.6 
Median survival, mo   
 OS14.510.1.02
 PFS8.76.3.006

Impact of the IDH1 Single-Nucleotide Polymorphism rs11554137:C>T on Prognosis in an Independent Data Set of 591 Grade 2 Through Grade 4 Gliomas

The IDH1 SNP rs11554137 polymorphism was analyzed in a cohort of 591 patients with gliomas patients (125 with grade 2 tumors, 129 with grade 3 tumors, and 337 with grade 4 tumors) who were treated at Bonn University Hospital for which survival data were available (Table 5). IDH1 SNP rs11554137:C>T was present in 47 patients (8%; all heterozygous), including 8 of 125 patients with grade 2 tumors (6.4%), 11 of 130 patients with grade 3 tumors (8.5%), and 28 of 361 patients (8.3%) with grade 4 tumors. The C>T variant was associated with poorer outcomes in patients with grade 3 glioma, but the trend was not significant in patients who had WHO grade 2 gliomas and GBM (Tables 6 and 7).

Table 5. Characteristics of the Patients Included in the Second Validation Series: Bonn Series
 Glioma Grade
Variable234
  1. Abbreviations: CI, confidence interval; GGC and GGT, glycine genotypes; GTR, gross total resection; NS, nonsignificant; OS, overall survival; PFS, progression-free survival; RT: radiotherapy; SD, standard deviation; STR, subtotal resection.

Total no. of patients125129337
Sex   
 Ratio of men to women1.451.301.57
 No. of men7473206
 No. of women5156131
Age at diagnosis, y   
 Median344161
 Range13–7417–8013–83
Karnofsky performance index: Mean±SD84.6±13.585.5±11.775.5±13.5
Degree of resection, %   
 Biopsy82.43.9
 STR5240.938.1
 GTR456.758
Postoperative RT, %24.889.991.4
Postoperative chemotherapy, %16.875.237.1
Median survival (95% CI), mo   
 OS178.0 (155.4–200.6)103.0 (71.2–134.8)11.0 (10.4–13.6)
 PFS60.0 (46.0–70.0)42.0 (39.0–62.0)8.0 (6.0–11.0)
Table 6. Isocitrate Dehydrogenase 1 Single-Nucleotide Polymorphism rs11554137:C>T Genotype and Patient Characteristics: Bonn Series
 Genotype 
VariableGGC/GGCGGC/GGTP
  1. Abbreviations: CI, confidence interval; GGC and GGT, glycine genotypes; GTR, gross total resection; NS, nonsignificant; OS, overall survival; PFS, progression-free survival; RT, radiotherapy; SD, standard deviation; STR, subtotal resection.

Total no. of patients54447 
World Health Organization grade, %  NS
 293.66.4 
 391.58.5 
 491.78.3 
Sex   
 Ratio of men to women1.451.94NS
 No. of men32231 
 No. of women22216 
Age at diagnosis, y   
 Median5052NS
 Range17–8319–81 
Karnofsky performance index: Mean±SD79.9±13.975.7±13.9NS
Degree of resection, %   
 Biopsy4.18.7NS
 STR41.643.5 
 GTR54.447.8 
Postoperative therapy, %   
 RT76.780.9NS
 Chemotherapy41.734.0NS
Median survival (95% CI), mo   
 OS33.0 (18.0–48.0)16.0 (7.4–24.6).037
 PFS13.0 (11.0–15.0)10.0 (6.0–19.0)NS: .093
Table 7. Overall and Progression-Free Survival According to Isocitrate Dehydrogenase 1 Single-Nucleotide Polymorphism rs11554137:C>T Status: Bonn Series
 Genotype: HR (95% CI), mo 
Glioma GradeGGC/GGCGGC/GGTP
  1. Abbreviations: CI, confidence interval; GGC and GGT, glycine genotypes; HR, hazard ratio; NR, not reached; NS, nonsignificant; OS, overall survival; PFS, progression-free survival.

Total no. of patients54447 
OS   
 2184.0 147.8–220.2160.0 (NR-339.0)NS: .072
 3123.0 70.0–176.016.0 (13.1–18.9).001
 412.0 10.3–13.710.0 (3.6–16.4)NS
PFS   
 261.0 50.0–74.037.0 (21.0–53.0)NS
 349.0 48.0–71.013.0 (NR-24.0)NS: .066
 410.0 9.0–11.010.0 (6.0–19.0)NS

Impact of IDH1 Single-Nucleotide Polymorphism rs11554137:C>T on the Response to Radiotherapy and Radiochemotherapy

To determine whether the impact of IDH1 SNP rs11554137:C>T had any correlation with treatment, we focused on homogeneously treated patients in the French series. Again, we observed a difference in PFS for patients who received radiotherapy (RT) (n = 475 patients; 8.2 months for SNP-positive patients vs 10.2 for SNP-negative patients; P = .0001) (Fig. 3A) and for patients who received combined RT plus temozolomide (n = 225 patients; 8.2 months for SNP-positive patients vs 11.5 for SNP-negative patients; P = .0003) (Fig. 3B).

Figure 3.

Progression-free survival is illustrated in patients with glioblastoma who received either (A) radiotherapy or (B) or a radiotherapy plus temozolomide regimen according to isocitrate dehydrogenase 1 (IDH1) reference single-nucleotide polymorphism rs11554137:C>T status.

DISCUSSION

The synonymous IDH1 codon 105 SNP rs11554137:C>T (IDH1105GGT) is located in the 5′ region of exon 4 and, thus, is very close to IDH1 codon 132. It has been identified as an adverse prognostic factor in AML.10 In line with this finding, we have demonstrated in a cohort of 952 patients with glioma that rs11554137:C>T also has a negative impact on patients with glioblastoma in terms of both PFS and OS. We confirmed these findings in an independent series of 306 patients with GBM from the same center and achieved similar results. Our data also suggest that SNP rs11554137:C>T is associated with a poor response to either RT alone or combined RT and temozolomide.14

In a second independent cohort from Germany, we also observed a negative impact of rs11554137:C>T on OS. However, this trend was significant for the whole series and for patients with WHO grade 3 gliomas subset, but not for patients with glioblastoma or WHO grade 2 gliomas. Taken together, these findings generally support the negative prognostic impact of rs11554137:C>T. Possible explanations for the discrepancy between the 2 French series and the German series include the lack of a central neuropathologic review, which may be particularly critical for grade 3 gliomas and glioblastomas, and the nature of the German series, which was a surgical series that included a substantial rate of early death (mostly patients in poor condition who were not treated after surgery). In addition, the German series contained a higher rate of censored events (52%) than the French series (27%) and, thus, the results may be more fragile (when excluding the censored events, the difference remained significant in the 2 French series but not in the German series)

The biologic consequences of this synonymous polymorphism remain speculative. One possibility is that SNP rs11554137:C>T may impact the stability of mRNA. Like what has been observed in adult patients with AML,10,15 we observed higher IDH1 mRNA levels in the IDH1105GGT variant glioma samples (data not shown). However, IDH1 mRNA levels were not related to outcome in glioblastoma; thus, it is unlikely that they explain the prognostic impact of SNP rs11554137:C>T (data not shown).

Although SNP rs11554137:C>T is very close to codon 132, we observed no correlation between the SNP rs11554137:C>T mutation and the overall occurrence of IDH1 codon 132 mutations. Similarly, we observed no association between SNP rs11554137:C>T and the other main genetic alterations that have been identified in gliomas (the inverse relation with EGFR amplification was not confirmed by the independent series).

In conclusion, our data suggest that the IDH1105GGT variant is associated with a negative prognostic impact in patients with high-grade gliomas, and particularly in those with glioblastomas. Further independent studies are warranted to define this potential prognostic role more clearly. An in vitro analysis will be required to better understand the functional impact of the IDH1105GGT variant.

FUNDING SOURCES

This work was supported by grants from the Association pour la Recherche sur le Cancer and the Canceropole Ile de France.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary