IDH1 Mutation Is an Independent Inferior Prognostic Indicator for Patients with Myelodysplastic Syndromes

Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal haematopoietic stem cell diseases occurring mostly in the elderly. They are characterized by peripheral cytopenias, dysplasia, and ineffective haematopoiesis, and an increased risk of leukaemia transformation [1, 2]. The median survival of MDS patients ranges from <6 months to several years [3]. Environmental exposures to smoking, benzene, and ionizing radiation have been considered risk factors for developing MDS [4]. With the ageing of the population and increasingly severe environmental pollution, the incidence of MDS has been increasing all over the world.

As a result of genomic sequencing technology, a set of recurrently mutated genes involved in biological and prognostic features have been identified in malignant haematological disorders [5, 6, 7, 8, 9]. Among these mutations, the prognostic implications of somatic isocitrate dehydrogenase (IDH) mutation (mIDH) remain controversial. Several studies have recorded adverse outcomes, while others have led to no definitive conclusions or have shown improved results [10, 11, 12, 13, 14, 15, 16]. IDH plays an essential role in diverse cellular processes, including metabolism, DNA synthesis, and adaptation to hypoxia [17]. The IDH protein is a key rate-limiting enzyme of the tricarboxylic acid cycle, which normally catalyses the oxidative decarboxylation of isocitrate to α-ketoglutarate (α-KG) to produce nicotinamide adenine dinucleotide phosphate (NADPH), while mutated forms of the protein instead convert α-KG into 2-hydroxyglutarate (2-HG) [18, 19]. Recent work has shown that excess 2-HG is involved in changes in methylation patterns in vitro and leads to impaired haematopoietic differentiation [20]. Common recurrent IDH mutations are located in codon 132 in exon 4 of the IDH1 gene and in codon 140 and codon 172 in exon 4 of the IDH2 gene. IDH1 encodes an enzyme that localizes to the cytosol and peroxisomes, while IDH2 encodes an enzyme that localizes to the mitochondria. Although the frequency of IDH mutations in MDS varies from 4 to 12%, IDH mutations are generally less common in MDS than in acute myeloid leukaemia (AML) [9, 10, 11, 21, 22, 23]. Additionally, IDH mutations are usually less common in Asian MDS patients than in the Caucasian population [23, 24]. In one study, mIDH was found to be less frequent in MDS (5%) and myeloproliferative neoplasms, but the frequency increased to 20% of patients at leukaemic transformation. This finding could indicate that mIDH may affect leukaemia transformation during the progression of haematological malignancies, which might then escape from chemotherapy, eventually leading to the relapse or progression of the malignancy. Similarly, we assume that mIDH may play an important role in MDS progression.

In this study, we examined mutations in exon 4 of the IDH1 and IDH2 genes in 97 Chinese patients with MDS, and explored the correlation between IDH1 or IDH2 gene mutations and the survival and progression of MDS.

A total of 97 patients with MDS (60 males and 37 females) diagnosed according to the World Health Organization criteria (2008), from the Department of Haematology, Shandong Provincial Hospital Affiliated to Shandong University, were included in this study from January 2013 to July 2016. The study was approved by the Medical Ethics Committee of the hospital, and informed consent documents were obtained from each patient in accordance with the Declaration of Helsinki. Clinical information was carefully obtained, including baseline demographics, peripheral blood cell counts, bone marrow (BM) blast counts, diagnosis, therapy, leukaemic transformation, the date of the last contact with the patient, and the patient's status at the last follow-up. We calculated the overall survival (OS) of patients with MDS using the date of diagnosis and the date of the last contact with the patient or their death. For patients with MDS, progression-free survival (PFS) was calculated from the date of diagnosis to the date of leukaemia transformation, death, or the last contact with the patient without disease progression. The International Prognostic Scoring System (IPSS) risk classification system was applied to make therapeutic decisions. Decitabine treatment was recommended for patients in the intermediate- and high-risk groups, but only 31 patients consented. The others accepted supportive treatment, with the exception of 4 who underwent chemotherapy. The low-risk patients received conservative and supportive care. Only 5 patients received allo-transplantation.

DNA was extracted from samples from the MDS patients using the TIANamp blood DNA kit (TIANGEN, Beijing, China) following the manufacturer's instructions. Exon 4 of the IDH1and IDH2 genes was amplified from each patient's DNA using a polymerase chain reaction (PCR) assay, and sequenced for somatic mutations. The PCR assay to detect IDH1 exon 4 mutations was conducted using the following primer sequences: sense 5′-CGGTCTTCAGAGAAGCCATT-3′ and anti-sense 5′-CACATTATTGCCAACATGAC-3′[14]. IDH2 exon 4 was done using sense 5′-CCACTATTATCTCTGTCCTC-3′ and anti-sense 5′-GCTAGGCGAGGAGCTCCAGT-3′[25]. PCR was performed in a 50-μL volume, containing 25 µL of PrimeSTAR Max Premix (2×) (Takara, Dalian, China), 1 µL (100 ng) of DNA, 1 µL (10 µM) of each of the sense and anti-sense primers, and 22 µL of dH₂O. The thermal cycling programme was as follows: 3 min at 94°C, followed by 35 cycles of denaturing at 94°C for 15 s, annealing at 57°C for 5 s, and extension at 72°C for 40 s. The resulting PCR products were sent to a sequencing company (BioSune Biotechnology Co., Ltd., Shanghai, China) for purification and the detection of IDH1 and IDH2 gene mutations via direct DNA sequencing techniques. Abnormal sequencing results were confirmed through at least 2 repeated analyses.

Bone marrow cells were collected at the time of the diagnosis of MDS, and metaphase chromosomes were identified according to the guidelines of the International Society for Cytogenetic Nomenclature (ISCN) [26]. According to the ISCN, at least 20 metaphase cells must be analysed and at least 2 metaphases must exhibit the same aberration in terms of a structural abnormality or extra chromosome. Karyotype risk stratification was defined as follows: good, normal karyotype, isolated -Y, del (5q) or del (20q-); poor, complex (≥3 abnormalities) or chromosome 7 anomalies; intermediate, other abnormalities.

Statistical analysis was performed with SPSS v17.0 (SPSS, Chicago, IL, USA). The differences between patients according to their IDH1 and IDH2 mutational status were tested using the Pearson χ² test or the Fisher exact test for categorical variables. Differences in continuous variables were compared with the Mann-Whitney U test for 2 groups. Kaplan-Meier survival curves were plotted for the estimated probabilities of disease-free survival and OS, and the survival distributions were compared using the log-rank test. The Cox proportional-hazards regression model was applied for multivariate analysis to analyse independent prognostic factors for OS and PFS. The reported p values are 2-sided, and p < 0.05 was considered to indicate statistical significance.

We detected 7 IDH mutations in the total cohort of 97 patients (7.2%). Sequence analysis of IDH1 mutations in the 97 samples revealed 3 somatic mutations (3.1%, 3/97), all at residue R132C (Arg132Cys). Sequencing evaluation of IDH2 mutations revealed 4 somatic mutations in IDH2 (4.1%, 4/97), including R140Q (Arg140Gln) in 3 patients and R140L (Arg140Leu) in the other patient. All of the patients carrying the missense mutation were heterozygous and mutually exclusive. All of the IDH mutations occurred in patients with refractory anaemia with excess blasts (RAEB). The frequency of IDH1 mutations was 16.7% (2/12) in the RAEB-1 patient cohort, while the frequency of IDH2 mutations was 0. Additionally, among the RAEB-2 patients, 2.9% (1/34) of the patients exhibited an IDH1 mutation, and 11.8% of patients exhibited IDH2 mutations (4/34).

The characteristics of the 60 male and 37 female MDS patients are summarized in Table 1 according to their IDH1/2 mutation status. The patients' ages ranged from 13 to 80 years, with a median of 59 years. Among the 97 patients, 39 (40.2%) had died by the last follow-up, and 18 patients (18.6%) had developed AML. Patients with IDH mutations presented higher BM blast counts (median 10.0 vs. 2.3%; p = 0.019), lower absolute neutrophil counts (median 0.44 × 10⁹/L vs. 1.21 × 10⁹/L; p = 0.027) and a higher ratio of decitabine treatment (21.4 vs. 28.9%; p = 0.032). There was no significant difference in other clinical parameters, including age, sex, haemoglobin, platelets, karyotype risk, transplantation rate, 2008 WHO classification, or IPSS [27] or revised IPSS (IPSS-R) risk stratification [3] between the patients with IDH mutations and those without such mutations. Patients with IDH1 mutations presented lower white blood cell counts (median 1.68 × 10⁹/L vs. 2.66 × 10⁹/L; p = 0.048) and lower absolute neutrophil counts (median 0.46 × 10⁹/L vs. 1.21 × 10⁹/L; p = 0.016). Patients with IDH2 mutations presented higher BM blast counts (median 12.0 vs. 2.3%; p = 0.013) and lower absolute neutrophil counts (median 0.43 × 10⁹/L vs. 1.21 × 10⁹/L; p = 0.016) and IPSS risk stratification (p = 0.044). Notably, among the 18 patients showing leukaemic transformation, only 1 exhibited an IDH1 mutation and all (100%) of the patients showing leukaemic transformation had died by the last follow-up. Additionally, 5 (71.43%) patients with IDH mutations had died, including 3 with IDH1 mutations and 2 with IDH2 mutations. Forty-seven of the 51 non-RAEB MDS patients received conservative and supportive care and the other 4 received decitabine treatment. Of the 46 MDS-RAEB patients included in this study, 27 were treated with decitabine, 4 underwent chemotherapy, and the others received supportive treatment for various reasons.

All chromosomal information was obtained from the 97 patients at diagnosis, and clonal chromosomal abnormalities were detected in 51 patients (52.6%). The frequency of IDH mutations associated with abnormal karyotypes (4/51, 7.8%) was similar to the frequency associated with normal karyotypes (3/46, 6.5%), with no statistically significant difference (p = 0.802). There was no significant association of IDH mutations with any special karyotype. Notably, the karyotypes of the chromosomes of the patients carrying IDH1 mutations were all normal. However, there was no significant difference in karyotype between IDH1 or IDH2 mutations and the wild type. Among the 4 patients with IDH2 mutations, 2 patients exhibited trisomy 8, and 2 displayed complex chromosome abnormalities, 1 of which included the chromosome 17 deletion (-17) and the chromosome 20 deletion (-20) (Table 2).

For the assessment of the influence of IDH1 and IDH2 mutations on the prognosis, follow-up data were available for all 97 MDS patients. The median follow-up duration was 10.0 months (range 1.0-41.0 months). The median OS was 19.8 months, while the median PFS was 16.9 months. Patients with IDH1 mutations exhibited shorter OS than those without such mutations (median 11.5 vs. 22.3 months; p = 0.039; Fig. 1a), and the 1-year survival rates were 0 and 68% for patients with mIDH1 and the wild type, respectively. Patients harbouring IDH2 mutations showed no difference in OS compared with those without IDH2 mutations (median 18.0 vs. 19.7 months; p = 0.560; Fig. 1b), and the 1-year survival rates were 50 and 66% for patients with mIDH2 and the wild-type, respectively.

The presence of IDH1 mutations in patients predicted reduced PFS when compared with the absence of IDH1 mutations (median 7.5 vs. 19.6 months; p = 0.042; Fig. 2a). Patients with IDH2 mutations exhibited no difference in PFS when compared with those without IDH2 mutations (median 12.0 vs. 19.4 months; p = 0.218; Fig. 2b).

Multivariate analysis demonstrated that IDH1 mutation (p = 0.018; hazard ratio [HR] 4.735; 95% confidence interval [CI] 1.299-17.264), karyotype risk (p = 0.036; HR 1.619; 95% CI 1.033-2.539), and IPSS-R risk category (p < 0.0001; HR 2.122; 95% CI 1.401-3.213) were independent inferior prognostic factors for OS (Table 3).

In this study, the incidence of mutations of IDH identified in MDS patients from northern China was 7.2% (3.1% IDH1 mutations and 4.1% IDH2 mutations). Thus far, there have been few studies concerning IDH mutations in Asian MDS patients. The frequency of IDH mutations found in this study was similar to that identified in 2 previously published studies with southern Chinese MDS patients and 1 small study of 22 MDS patients in Thailand [22, 23, 28]. However, Lin et al. [24]reported that the rate of IDH mutations in Taiwanese patients with MDS may be as low as 4.6% (22/477). Some studies have shown that the mutation rate of IDH in Caucasian MDS patients is between 4 and 12% [29]. In a study from the Mayo Clinic, it was shown that MDS patients with refractory cytopenia with multilineage dysplasia or RAEB exhibited higher IDH mutation rates than those with refractory anaemia [21]. We found that the MDS patients with relatively higher risk (the RAEB patients) exhibited a higher IDH mutation rate, as observed in Caucasian patients. The frequency of IDH1 and IDH2 mutations was much higher in patients with AML than in MDS patients. One review indicated that IDH mutations are among the most common mutations found in AML patients, with IDH1 mutations occurring in 7-14% and IDH2 mutations in 8-19% [30]. Our research also showed that IDH mutations are closely associated with higher BM blast counts and lower absolute neutrophil counts. All of the above findings suggest that IDH mutations may be associated with the progression of MDS.

In accordance with our findings, all previous studies have determined that the frequency of IDH1 mutations is lower than that of IDH2 mutations in MDS. The reasons for the variation in mIDH frequency in different studies are related to the limitations in sequencing technology, selection bias, and ethnic and geographical differences. Our study showed that all of the identified mutations were mutually exclusive. Notably, a recent study identified dual mutations in the IDH1 and IDH2 genes in MDS via targeted next-generation sequencing [31].

Some recent publications have shown that normal karyotypes in patients harbouring IDH mutations indicate adverse outcomes in AML [32]. Nevertheless, no significant relationship between IDH mutations and any special karyotype was observed in our study, suggesting that IDH mutations may influence MDS via mechanisms different from those in AML. In addition, 1 study showed that the frequency of IDH mutations in patients with trisomy-8-associated MDS and AML was much higher than in those with MDS or AML without isolated trisomy 8, although in patients with trisomy-8-associated MDS, the prognosis was not significantly different between those with and those without IDH mutations [33]. In this study, we identified 3 mIDH patients who exhibited karyotype abnormalities typical of MDS, such as +8, -17, and 20q-. Thus, we have reason to believe that IDH2 mutations are related to the pathogenesis and progression of MDS. Due to the limited number of cases examined and the low frequency of typical MDS karyotypes, we did not statistically calculate the effect of IDH2 mutations in these abnormal karyotypes on the prognosis of MDS. Additional investigations are needed to infer the relationship between IDH mutations and typical karyotypes of MDS.

Two meta-analyses have shown that patients with mIDH1 CN-AML exhibit lower rates of complete remission and inferior OS when compared with those without IDH1 mutations [34, 35]. Another meta-analysis revealed that the IDH1 mutation may be associated with inferior event-free survival, while the IDH2 mutation was observed to improve OS [36]. However, the prognostic impact of IDH1 and IDH2 mutations in MDS has always been controversial. Some studies suggest that IDH1 mutations in patients with MDS are associated with an unfavourable prognosis [11, 21, 23]. Conversely, 2 other studies showed that IDH2 mutations, rather than IDH1 mutations, were associated with significantly shorter OS in an MDS cohort [9, 24]. A few studies have also found no influence of IDH mutations on OS in MDS patients [10, 22, 37]. In our MDS cohort, we observed that IDH1 mutations were associated with shorter OS and PFS, whereas IDH2 mutations had no impact. Interestingly, 1 study showed that IDH1 R132 mutations could increase hypoxia-inducible factor 1α (HIF-1α) levels, which promotes tumour growth when oxygen levels decrease, and the stability of HIF-1α is regulated by α-KG [38]. Moreover, in 2 other studies, it was shown that DNA hypermethylation and HIF-1α expression predict a poor prognosis in MDS [39, 40]. Notably, biochemical studies have demonstrated that the capacity of different IDH mutations differs. The most effective IDH mutation is IDH2R172K, followed by IDH1R132H, IDH1R132C, and IDH2R140Q [41]. In our study, 3 IDH1R132C mutations, 3 IDH2R140Q mutations, and 1 R140L mutation were identified. Thus, the inhibition ability of α-KG-dependent enzymes is significantly different among different IDH1/2mutations, which may partially explain the difference in prognosis of MDS between IDH1 mutations and IDH2 mutations. The occurrence of the IDH2 R140L mutation is rare, and the associated OS has not been reported in the literature. The OS of the patient in our cohort harbouring an IDH2 R140L mutation was only 5 months. This patient also exhibited lower IDH2 expression than the patients with the IDH2 R140Q mutation, according to real-time fluorescence quantitative PCR analysis. This finding may indicate that the IDH2 R140L mutation exhibits a stronger α-KG-dependent enzyme-suppressing ability than the IDH2R140Q mutation; however, due to our limited data, we could not statistically confirm this result.

In addition, IDH1 mutations, rather than IDH2 mutations, represent an adverse prognostic factor in MDS, which may be due to the close association of IDH1 mutations with DNMT3A and NPM1 mutations, while IDH2 mutations are closely associated with SRSF2, ASXL1, STAG2, and RUNX1 mutations [42]. A series of IDH-mutated inhibitors are currently under study, including first-generation inhibitors of IDH1 (AG120, IDH-305, and FT-2102) and IDH2 (AG221) mutants, and a second-generation pan-inhibitor (AG881). Both the IDH1 inhibitors (AG120 and FT305) and the IDH2 inhibitor (enasidenib) have shown promising effects in relapsed and refractory AML patients [43, 44]. We believe that IDH1-mutant inhibitors will be widely applied in the clinical therapy of MDS as a promising therapeutic option in the future.

In conclusion, the incidence of IDH mutations in northern Chinese MDS patients is 7.2%. IDH mutations were closely associated with higher BM blast counts and lower absolute neutrophil counts, and showed no relationship with any specific karyotype. IDH1 mutation, karyotype risk, and IPSS-R risk category were found to be independent inferior prognostic factors for MDS. Due to the limited sample size, the results of our study should be treated cautiously, and the adverse prognostic impact of IDH1 mutations should be validated by large samples. We believe that if a larger study is conducted, it will be demonstrated that IDH1 mutations can serve as a useful biomarker for the timely and proper treatment of MDS.

This work was supported by grants from the Natural Science Foundation of Shandong Province (No. ZR2015PH028), the National Science Foundation of China (No. 81570104), the Medicine and Health Science Technology Development Project in Shandong Province (No. 2014WS0348), and the Key Research and Development Project of Shandong Province (No. 2015GSF118058).

The authors declare that they have no competing interests.