Molecular Evaluation of DNMT3A and IDH1/2 Gene Mutation: Frequency, Distribution Pattern and Associations with Additional Molecular Markers in Normal Karyotype Indian Acute Myeloid Leukemia Patients

Mutations in the DNMT3A and IDH genes represent the most common genetic alteration after FLT3/NPM1 in acute myeloid leukemia (AML). We here analyzed the frequency and distribution pattern of DNMT3A and IDH mutations and their associations with other molecular markers in normal karyotype AML patients. Forty-five patients were screened for mutations in DNMT3A (R882), IDH1 (R132) and IDH2 (R140 and R172) genes by direct sequencing. Of the 45 patients screened, DNMT3A and IDH mutations were observed in 6 (13.3%) and 7 (15.4%), respectively. Patients with isolated DNMT3A mutations were seen in 4 cases (9%), isolated IDH mutations in 5 (11.1%), while interestingly, two cases showed both DNMT3A and IDH mutations (4.3%). Nucleotide sequencing of DNMT3A revealed missense mutations (R882H and R882C), while that of IDH revealed R172K, R140Q, R132H and R132S. Both DNMT3A and IDH mutations were observed only in adults, with a higher frequency in males. DNMT3A and IDH mutations were significantly associated with NPM1, while trends towards higher coexistence with FLT3 mutations were observed. This is the first study to evaluate DNMT3A/ IDH mutations in Indian patients. Significant associations among the various molecular markers was observed, that highlights cooperation between them and possible roles in improved risk stratification.


Introduction
Acute myeloid leukemia (AML) represents one of the most biologically and clinically diverse subtypes among various hematological malignancies.The diagnostic karyotype of AML patients still remains to be the most reliable prognostic marker deciphering patients into different risk group (Grimwade et al., 1998) however the major challenge for prognosticfication occurs when the patient shows apparently normal karyotype (NK-AML).These NK-AML patient accounts for nearly 45% of all cases and have been assigned into "intermediate risk group" because of their uncertain clinical outcome (Slovak et al., 2000).The absence of cytogenetic abnormalities in a considerable proportion of patients argues for refinement of the classification of AML.Infact, identification of molecular markers in these NK-AML patients have been the focus of most of the research groups across the globe.At the present time, the most well known molecular RESEARCH ARTICLE Molecular Evaluation of DNMT3A and IDH1/2 Gene Mutation: Frequency, Distribution Pattern and Associations with Additional Molecular Markers in Normal Karyotype Indian Acute Myeloid Leukemia Patients Firoz Ahmad 1 , Rupali Mohota 1 , Savita Sanap 1 , Swarna Mandava 2 , Bibhu Ranjan Das 1 * different clinical outcome (Gaidzik et al., 2008;Mrozek et al., 2008).Interestingly, few of these markers such as FLT3, NPM1 and CEBPA have been already included in the revised version of World Health Organization classification of leukemias (Swerdlow et al., 2008).
In search of additional molecular markers in AML, several breakthrough researches leveraging the power of recently identified novel technologies, such as massively parallel DNA sequencing and next-generation sequence techniques have eased the systematic efforts to identify molecular markers with valuable prognostic significance (Mardis et al., 2009;Ley et al., 2010).The outcome of such advanced studies has led to the identification of DNA methyltransferase 3a (DNMT3A) mutations (Ley et al., 2010), ten-eleven-translocation oncogene family member 2 (TET2) (Delhommeau et al., 2009), and mutations and isocitrate dehydrogenase (IDH) 1 gene mutations (Mardis et al., 2009) in AML patients.Of these mutations, two markers viz., DNMT3A and IDH mutations have gained tremendous attention in recent times, and have been extensively evaluated by different research groups across the globe (Paschka et al., 2010;Thol et al., 2011;Chotirat et al., 2012;Marcucci et al., 2012).
The gene DNMT3A is located on chromosome 2p23 locus, encodes the enzyme DNA methyltransferase (DNMT) 3A, which catalyzes the addition of methyl groups to the cytosine residue of CpG dinucleotides in DNA (Brenner and Fuks, 2006;Chen and Li, 2006).Recently, somatic mutations in DNMT3A gene were first discovered by massively parallel DNA sequencing (99.6%diploid coverage) from the genome of cells of a NK-AML patient.Different types of mutations such as missense, nonsense, frameshift, splice-site mutations and deletions were detected, of which missense mutations especially R882 was the most recurrent mutations occurring within dimer or tetramer interface of DNMT3A (Ley et al., 2010).Overall, the frequency of DNMT3A mutations varies from 14% to 35% in AML globally (Ley et al., 2010;Marcucci et al., 2012;Hou et al., 2012;Renneville et al., 2012) Ribeiro et al., 2012;Patel et al., 2012).More interestingly, similar to FLT3 mutations, several recent studies have consistently reported association of DNMT3A mutation with poor clinical outcome in AML patients (Ley et al., 2010;Thol et al., 2011;Marcucci et al., 2012).
Furthermore, the gene IDH encodes a nicotinamide adenine dinucleotide phosphate+-dependent enzyme for oxidative decarboxylation of isocitrate, and has an essential role in the tricarboxylic acid cycle as well as involved in the cellular defence of oxidative damage (Lee et al., 2004).Similar to the discovery of DNMT3A mutations, molecular alterations of IDH1 were first reported in AML by massively parallel DNA sequencing of an entire NK-AML genome (Mardis et al., 2009).Further studies found recurrent mutations in IDH2 gene indicating that both IDH1 and 2 are frequently mutated in AML patients (Paschka et al., 2010;Marcucci et al., 2010).The overall combined frequency of IDH mutations varies from 15-22% of all AML patients (Patel et al., 2011;Abbas et al., 2010;Zou et al., 2010) and 25-30% of NK-AML patients have either of IDH1 or IDH2 mutations (Marcucci et al., 2010;Paschka et al., 2010).Nevertheless, the true clinical impact of IDH mutations is still debatable, and it varies considerably among the different type of mutations detected, and also by the presence of concurrent mutation in other clinically relevant genes.
Most reports of DNMT3A and IDH mutations come from Western countries, while there are few reports from Asia (Zou et al., 2010;Lin et al., 2011;Chotirat et al., 2012;Hou et al., 2012).At the present time, no precise published data exist in India with respect to the frequency and distribution patterns of DNMT3A and IDH mutations, despite the fact that AML is one of the most common hematologic malignancies associated with the greatest mortality and morbidity in this country.In this study, we performed mutation analysis of the DNMT3A and IDH genes in Indian AML patients with an objective to explore their frequency and distribution patterns, and correlate results with the clinical features and cytogenetic findings and other molecular markers.To the best of our knowledge, this study represents the first series of DNMT3A and IDH mutations ever reported from India.

Materials and Methods
The present study was conducted at the Research and Development Division of SRL Ltd., Mumbai, India.The study included 45 de novo normal karyotype AML (NK-AML) samples, of whome 36 were described in our recent report on CEBPA gene mutations (Ahmad et al., 2012) and an additional 9 newly diagnosed NK-AML patients were included.Genomic DNA was available from mononuclear cells from the bone marrow and/or whole blood.DNA yield and purity were determined by measuring the absorbance at 260/280 nm.The presence of other molecular markers, such as FLT3, NPM1, WT1, KIT and CEBPA gene mutations were screened in additional nine cases as per our previous reports (Ahmad et al., 2009;Ahmad et al., 2010;Ahmad et al., 2011).The study was carried out in accordance with the Declaration of Helsinki, and written consent was obtained from each subject.Treatment and outcome were not analyzed in the current study.

Screening of DNMT3A gene mutation
High molecular weight genomic DNA was amplified for missense mutation R882 of the DNMT3A gene using the following primer sets: DNMT3A_F: 5'-TCCTGCTGTGTGGTTAGACG-3' and DNMT3A_R: 5'-TCTCTCCATCCTCATGTTCTTG-3' as described earlier (Fiied et al., 2012).Briefly, the polymerase chain reaction (PCR) was performed in a 50 µl volume containing 50 ng of genomic DNA, 1.5 mmol/L MgCl 2 , 0.2 mM dNTPs, 20 pmol of each oligonucleotide primer, and 1.5 U of Taq polymerase (Invitrogen Life Technology, Sao Paulo, Brazil).The PCR conditions consisted of an initial denaturation step at 98ºC for 3 min followed by 35 cycles at 94ºC for 30 s, 58ºC for 30 s, 72ºC for 45 s, and a final step at 72ºC for 10 min.Amplified products were visualized on 2% agarose gel and directly sequenced.

Statistical analysis
Data were analyzed using Student's t-test and chisquare test to calculate the significance of association DOI:http://dx.doi.org/10.7314/APJCP.2014.15.3.1247 DNMT3A andIDH Mutations in Normal Karyotype Indian AML patients between DNMT3A/IDH mutations and other discrete variables among subgroups of patients.p-values <0.05 were considered to be significant.

DNMT3A and IDH mutations and their relation to clinical characteristics
The presence of DNMT3A and IDH mutations were correlated with clinical and laboratory parameters.Both DNMT3A and IDH mutations were predominantly present in male patients in comparison to female patients (15.2% vs 8.3% for DNMT3A and 18.1% vs 8.3% for IDH) (p>0.05).The current study included pediatric (n=8) as well as adult AML patients (n=37), of which both DNMT3A and IDH mutations were observed exclusively in adults (16.2% (6/37) for DNMT3A and 23.3% (7/37) for IDH), while none of the eight pediatric cases showed any evidence of mutations indicating that these mutations are less frequent in pediatric groups.Nevertheless, none of these differences were statistically significant (p>0.05).Interestingly, the median age of patients with either DNMT3A or IDH or both mutations (n=11) were significantly higher with respect to those without these mutations (n=34) (45 years vs 32.5 years, p<0.05).Furthermore, DNMT3A and IDH mutations were not associated with any FAB subtypes (p>0.05).There was no statistical correlation with other clinical parameters, including white blood cell (WBC) count, hemoglobin, and platelet count between patients with and without either of the mutations (p>0.05).

DNMT3A and IDH mutations and their association with other molecular markers
The presence of mutations in NPM1, FLT3, WT1, KIT and CEBPA genes were known in 36 of these cases, as per our previous report (Ahmad et al., 2012), while additional 9 cases were screened for these mutations in this study as per our previous reports (Ahmad et al., 2009;Ahmad et al., 2010;Ahmad et al., 2011).Table 2 shows the overall molecular characteristics of the entire cohort, while Figure 2 shows the interrelationships among the various mutations, as represented visually with the use of a Circos plot.The length of the arc corresponds to the frequency of mutations in the first gene, and the width of the ribbon corresponds to the percentage of patients who also had a mutation in the second gene.Pairwise co-occurrence of mutations is denoted only once, beginning with the first gene in the clockwise direction.It is worth noting that DNTM3A mutations were significantly more prevalent in NPM1 positive cases when compared to NPM1 wild type cases (28.6% vs 6.4%, p<0.05) (Table 2).In addition, when correlated with FLT3 mutations, though DNTM3A mutations were more dominant in FLT3 positive cases than wild type (25% vs 6.9%), this difference was statistically insignificant (p>0.05)(Table 2).Furthermore, the IDH mutations were present in 5 of 14 (35.7%)cases with NPM1 mutation, when compared with only 2 of 31 (6.4%) cases with NPM1 wild type genotype (p<0.05).Likewise DNMT3A mutation, IDH mutations were also    doi.org/10.7314/APJCP.2014.15.3.1247 DNMT3A andIDH Mutations in Normal Karyotype Indian AML patients preponderant in FLT3 positive cases in comparison to those with normal FLT3 (25% vs 10.3%) cases, however this association failed to achieve statistical significance (p<0.05).Thus, it appears that NPM1 mutation appears as a major significant coexisting genetic mutation in DNMT3A and IDH-mutated cases.
In contrast, no DNMT3A and IDH mutations were observed in cases with WT1, KIT or CEBPA mutation indicating that these mutations cane be mutually exclusive in nature.Both DNMT3A and IDH mutations were observed only in WT1, KIT and CEBPA wild type cases (p>0.05)(Table 2).

Discussion
Research in the last decade has shown tremendous improvement in terms of identification of new molecular lesions which are of high clinical relevance in AML especially those with NK-AML.A subset of these NK-AMLs are found to harbor mutations of genes that normally function in cell proliferation, differentiation, and survival such as FLT3, NPM1, CEBPA, KIT, WT1, RAS and AML1 (Takahashi et al., 2011a;2011b).The mutations are already known to play a significant role in the clinical outcome of AML patients.Furthermore, identification of additional molecular markers is a need, which can allow further risk stratification and effective patient management.
Mutations of DNMT3A and IDH were recently described in AML through a rapid whole genome sequencing approach (Mardis et al., 2009;Ley et al., 2010).The reported frequency of DNMT3A mutations varies between 14% to 35% (Ley et al., 2010;Hou et al., 2012;Marcucci et al., 2012;Patel et al., 2012;Renneville et al., 2012;Ribeiro et al., 2012;), while that of IDH mutations varies between 25-30% of NK-AML globally (Paschka et al., 2010;Marcuccii et al., 2010) (Table 3).However, most of these available reports on DNMT3A and IDH mutations are from Western countries, while those from Asian countries are few (Lin et al., 2011;Chotirat et al., 2012;Hou et al., 2012), and in India to date there is no precise published data (Table 3).To the best of our knowledge, this is the first study to evaluate the mutation status of DNMT3A and IDH genes in an Indian NK-AML patient.We have found the frequency of DNMT3A and IDH gene alterations in 13.3% and 15.4% of cases respectively, suggesting that alterations of the IDH gene are more common in comparison to mutations of the DNMT3A gene (p>0.05).The frequency of DNMT3A mutation (15.4%) in our NK-AML cohort was lower in comparison to some of the reported frequency from France (29%) (Renneville et al., 2012), Taiwan (22.9%) (Hou et al., 2012), Germany (27.2%) (Thol et al., 2011).The difference in the frequency of DNTM3A mutation may be explained by the fact that in the current study only R882 mutations were evaluated and mutation other than R882 were not identified.In addition to this, variable sensitivity of the detection assays, or the racial differences can also be taken into consideration.
Both DNMT3A and IDH mutations were observed only in adults (>18 years), while no evidence of mutation in either of the genes were seen in pediatric patients (≤ 18 years), though the no. of pediatric patients were lesser in the current study (n=8).This is in agreement with a recent study (Ho et al., 2011), and highlights the fact that childhood leuekemias can be very different from that of adults in pathogenesis and biological characteristics (Jeha et al., 2002), and raises the possibility that there are significant differences in the mechanisms responsible for the introduction of DNMT3A or IDH mutations in the two populations.There appears to be higher frequency of DNMT3A/IDH mutations in male patients in comparison to female patients, however the difference was statistically insignificant (p>0.05).No statistical association with any FAB classification was observed which is in agreement with recent literature (Qiao et al., 2011;Marková et al., 2012).With respect to correlation with laboratory parameters, both DNMT3A and IDH mutations were not associated with Hb, WBC and PC (p>0.05) which tallies with previous finding (Qiao et al., 2011).
Sequencing analysis of all the cases for DNMT3A R882 mutation revealed R882H (5 of 6, 83.4%) as the predominant mutation type, while R882C (1 of 6, 16.6%) mutation type was less frequently observed, which is in agreement with a recent study from Germany (Thol et al., 2011).Furthermore, sequencing of IDH gene revealed two types of amino acid exchanges in IDH1 gene (R132H and R132S), one each in two different cases.Likewise, two types of IDH2 mutations were detected, R172K (3 of 5, 60%) which was more frequent in comparison to R140Q (2 of 5, 40%).However, another study reported R140 mutations as the prevalent IDH2 mutation in comparison to R172 mutations (Marcuccii et al., 2010).Nevertheless, none of our cases showed simultaneous mutations of both IDH1 and IDH2 together which tallies with previous finding (Marcuccii et al., 2010).
To explore if any possible correlation of DNM3A and IDH mutations coexist with other clinically relevant molecular markers, we had data for FLT3, NPM1, CEBPA, WT1 and KIT genes as per our previous findings (Ahmad et al., 2009;Ahmad et al., 2010;Ahmad et al., 2011).Both DNMT3A and IDH mutations were significantly more frequent in NPM1 mutant cases in comparison to NPM1 wild type cases (28.6% vs 6.4% and 35.7% vs 6.4%, p<0.05) respectively (Table 2).This is in agreement with previous reports, wherein, DNMT3A and IDH mutations were significantly associated with NPM1 mutations (Marcuccii et al., 2010;Thol et al., 2011).The circos plot (Figure 2) shows the inter relationships among the various mutations, and it looks like NPM1 mutation appears as a major significant coexisting genetic mutation in DNMT3A and IDH-mutated cases.It is interesting to note that, unlike previous report (Marcuccii et al., 2010) , wherein, R172 IDH2 mutations were mutually exclusive with all other prognostic mutations (NPM1, FLT3, WT1, CEBPA and MLL-PTD), in the present study, of the three R172 IDH2 mutations detected, two of them also showed presence of NPM1 mutations and remaining one showed FLT3-ITD mutations.Additionally, cases with DNMT3A and IDH mutations were more likely to have mutations in FLT3 in comparison to FLT3 wild types cases which is in agreement with previous findings (Marcuccii et al., 2010;Thol et al., 2011), though the differences in the current study was statistically insignificant (p>0.05).In contrast to NPM1 and FLT3 mutation, both DNMT3A and IDH mutations were exclusively found in CEBPA, WT1 and KIT wild type cases (p>0.05), which corroborates previous finding (Marcuccii et al., 2010;Chotirat et al., 2012) (Table 2).Nevertheless, additional number of cases would have certainly helped to get more statistically defined information in terms of the co-occurrence and their association.
As far as the possible mechanism of DNMT3A mutation in oncogenesis is concerned, Holz-Schietinger and colleagues, recently demonstrated that the common mutations such as R882H which occurs within the dimer or tetramer interface domain disrupts the tetramerization activity of DNMT3A.The processive methylation of multiple CpG sites is disrupted when tetramerization is eliminated which may contribute to initiation of oncogenesis and its progression (Holz-Schietinger et al., 2012).Furthermore, in case of IDH mutations and its role in cancer formation, it is thought that these mutations prevent oxidative decarboxylation of isocitrate to α-ketoglutarate and confer novel enzymatic activity, facilitating the reduction of α-ketoglutarate to 2-hydroxyglutarate, a putative oncometabolite (Rakheja et al., 2012).
Finally, though the current study did not evaluated the effect of DNMT3A and IDH mutations in the clinical outcome, numerous studies have consistently reported association of DNMT3A mutation with poor clinical outcome in AML patients (Renneville et al., 2012;Marcuccii et al., 2010), usually independent of the presence of either FLT3 or NPM1 mutations (Ley et al., 2010).More recently, another interesting study reported no impact of DNMT3A mutation on outcome, but could be a predictive factor for response to idarubicin and thus, could have a direct influence in the way AML patients should be managed (LaRochelle et al., 2011).In contrast, the true clinical impact of IDH mutations is still debatable, and it varies considerably among the different type of mutations detected, and also by the presence of concurrent mutation in other clinically relevant genes (Paschka et al., 2010).
In conclusion, this is the first study to evaluate DNMT3A and IDH mutations in Indian patients, and has demonstrated that the frequency of DNMT3A and IDH1 mutation is lower to the worldwide incidence, while that of IDH2 is comparable.The present study depicts for the first time in Indian patients the involvement of DNMT3A/ IDH in AML.Clearly, the list of acquired mutations with clinical value in AML is growing each day.The true clinical implementation of this multitude of markers will require integrated approaches and selection of markers to facilitate AML risk stratification in the future, especially in NK-AML.

Figure 2 .
Figure 2. Circos Plot Showing the Overall Frequency of Somatic Mutations in The Entire Cohort (N=45) and the Interrelationships among the Various Mutations.The length of the arc corresponds to the frequency of mutations in the first gene, and the width of the ribbon corresponds to the percentage of patients who also had a mutation in the second gene.Pairwise co-occurrence of mutations is denoted only once, beginning with the first gene in the clockwise direction

Table 3 . Comparison of Some of the Recent worldwide Incidence of DNMT3A and IDH Mutations in de Novo AML
*ND: Not done; NA: Not applicable