Characteristics of DNMT3a mutation in acute myeloid leukemia and its prognostic implication

Background Acute myeloid leukemia (AML) is a clonal disorder arising from the differentiation arrest of myeloid pre‑ cursor and malignant proliferation of a bone marrow derived, self‑renewing stem or progenitor cells inside the bone marrow (BM) and blood due to numerous genetic mutations. Some mutations can also adjust DNA methylation and may play a critical function in pathogenesis in Cytogenetically Normal Acute Myeloid Leukemia (CN‑AML). Somatic mutations in DNMT3a were pronounced in approximately 20% and ∼ 30–35% of overall AML and CN‑AML, respectively. Most DNMT3a mutations in AML have been observed to be heterozygous, A missense mutation, R882, located inside Hot spot exon 23, has been found to be the maximum common mutation. This is a preliminary study conducted on 20 adult Egyptian patients newly diagnosed as AML where Sanger sequencing of Hotspot Exon 23 of DNMT3a gene was performed on their initial bone marrow samples and were followed up to 3 months post‑induc‑ tion therapy. Only De Novo AML patients were included in our study. Results Our results revealed that overall DNMT3a mutations were present in 25% of our patients, 10% having the R882 (rs147001633) mutation being 5% R882C and 5% R882H. Immunophenotyping analysis among Mutated DNMT3a (R882 and Non R882) and Wild DNMT3a revealed that AML markers exhibited no significant differences except for myeloperoxidase positivity which was significant among the groups (0.050). Regarding cytogenetics, only one case of the mutated DNMT3a had positive FISH inv (16), where the rest were FISH negative. After 28 days of induction, 75% of all our patients achieved complete response (CR), 20% achieve partial response (PR) out of which 75% are DNMT3a mutated. After 3 months follow‑up, 10% of all patients faced mortality where 5% was DNMT3a wild type (died due to treatment‑related mortality) and 5% was R882 mutated DNMT3a. Conclusion DNMT3a mutations are present in 25% (5/20) of our AML patients, with 10% (2/20) having the R882 mutation being 5% (1/20) R882C and 5% (1/20) R882H. R882 mutation is associated with resistance to chemotherapy, and poorer outcomes, highlighting its poorer prognostic significance in AML.


Background
Acute myeloid leukemia (AML) is a clonal disorder arising from the arrest of differentiation of myeloid precursors and the malignant proliferation of self-renewing stem or progenitor cells derived from the bone marrow (BM) and blood.It can be classified into different subtypes based on the cell type and the degree of maturity according to WHO classification of hematological diseases [1].
In hematologic malignancies, recurrent chromosomal abnormalities play a significant role in leukemogenesis [2].Gene fusions and mutations, including those in CEBPA and NPM1, are major recurrent genetic abnormalities in AML and have been established as major indicators for classification [3].
Furthermore, genes involved in DNA methylation, including DNA methyltransferase 3 a (DNMT3a), Isocitrate dehydrogenase 1 and 2 (IDH1 & IDH2) and Ten eleven translocation 2 (TET2) were found to be frequently mutated, especially in the cytogenetically normal AML (CN-AML) [4].DNA methylation, involved in the silencing of tumor suppressor genes, has been associated not only with leukemogenesis but also with the clonal evolution of the myelodysplastic syndrome to AML [5].
Therefore, mutations in these genes play an important role in disease pathogenesis in CN-AML.Somatic mutations in DNMT3a have been reported in approximately 20% and ∼30-35% of total AML and CN-AML, respectively [6].
One of the DNMT3a mutations is rs147001633 which is a missense mutation located in the methyltransferase domain, often referred to as R882 mutation, has been found to be the most common mutation [7].R882 DNMT3a inhibits the expression of the wild-type DNMT3a and reduces the de-novo methyltransferase activity, which results in the focal hypomethylation at specific CpGs throughout the genomes of AML cells [8].To our knowledge, limited number of studies characterized DNMT3a mutation in Egyptian population using Sanger sequencing.We aim to study the DNMT3a Mutation in Egyptian patients with AML to be used as a routine test at initial diagnosis and to correlate its prognostic implication with the immunophenotyping, cytogenetic FISH analysis, peripheral blood count and radiological studies of the patients.

Subjects
This preliminary study was conducted on 20 adult Egyptian patients newly diagnosed as AML.Patients were recruited from hematology unit of Internal Medicine department, Ain shams University Hospitals during the period from July 2022 to April 2023.The diagnosis of AML was based on morphologic, cytochemical evaluation, immunophenotyping and complementary cytogenetics according to updated WHO 2016 diagnostic criteria.Only De Novo AML patients were included in our study.Relapsed AML were excluded from our study.Faculty of medicine Ain Shams University Research Ethical committee (REC) approval was granted under number FMASU MD128/2022.Informed Written Consents was taken from all patients.
All Patients were subjected to the following: Full history taking and imaging studies for detection of organomegaly for AML cases.
Cytogenetic analysis carried on bone marrow or peripheral blood samples using FISH or karyotyping techniques.At least 100 interphase nuclei were scanned in every case under the chromo-scan.For scanning, lowpower objective lens was used, followed by capture using oil immersion objective lens for the detection of the signals by CytoVision automated cytogenetics platform [Leica Biosystems Richmond, USA].A Cut off value for diagnosis of positive results was > 10% for single fusion probe and > 3% for double fusion probes.The used probes: Vysis RUNX1/RUNX1T1 double fusion probe for t (8;21).Vysis PML/RARA single fusion probe for t (15;17).Vysis CBFB break apart probe for inv (16).Vysis LSI MLL Dual Color break apart rearrangement probe for 11q23 rearrangement.Vysis LSI BCR/ABL single fusion probe for t (9;22).
Follow up was done for all studied patients to detect response to induction therapy at day 28 and 3 months intervals.

Sample collection
Two mL of whole venous blood were collected, on K2-EDTA (1.2 mg/mL) (Greiner) for CBC.Three mL of fresh BM samples (on two separate tubes) were collected on K2-EDTA: 0.5 mL for preparation of BM smears and cytochemical staining for preliminary determination of leukemic lineage, 1 mL for IPT by FCM and 1.5 mL for Sanger Sequencing.(Samples were preserved at − 20 °C until analysis) One mL BM aspirate samples were collected on sterile lithium heparin coated vacutainer tubes for cytogenetic analysis.Conventional Karyotyping results were retrieved from patients' medical records.
Sanger Sequencing of Hotspot Exon 23 of DNMT3a gene:

Primer designing
Primers used for the amplification of Hotspot exon 23 of the DNMT3a gene were designed using National Center for Biotechnology Information (NCBI) Pick primer Tool and were ordered and purchased from ThermoFisher [Massachusetts, USA].

DNA extraction
The QIAamp Blood Kit (Cat.No. 51106; Qiagen Inc., Valencia, CA) was used for performing DNA extraction.

Initial DNA amplification
Master Mixing and conventional PCR was done using PCR Master Mix Kit by (ThermoFisher, Massachusetts, USA: cat no.#K0171.Amplification protocol was: Initial denaturation at 95° for 1 min, 40 cycles of denaturation at 95° for 30 s, annealing for 30 s at 60°, Extension at 72° for 1 min and final extension at 72° for 15 min for a total of 40 cycles.After gel electrophoresis was done, PCR cleanup was performed using Thermofisher Exo-SAP IT Express PCR product cleanup kit (Cat.No. 75001.1.EA).

Sanger sequencing
DNA sequencing using Sanger's dideoxy method was performed using BigDye ® Terminator cycle sequencing kit v3.1 (Massachusetts, USA) compatible with ABI 3500 ® Genetic analyzer.The original reaction setup, recommended by the manufacture, was optimized with the following steps: In a total assay volume of 20 μL, BigDye ® Terminator mix 4 uL, Sequencing buffer 2 uL, forward or reverse primer 1 uL, Amplified DNA 2 uL were added to the respective reaction assays and 11 uL RNase free water.Assay conditions were setup by the manufacturer recommendations with the following settings: Initial denaturation at 96 °C for 1 min, 15 cycles of denaturation (at 96 °C for 10 s), annealing (at 50 °C for 5 s) and extension (at 60 °C for 75 s), followed by final extension (2 set of 5 cycles with expanding the extension phase 60 °C to 90 s and 2 min subsequently).Same steps were repeated for reverse sequencing.Post-sequencing cleanup was done using Big Dye X Terminator Purification kit (Massachusetts, USA:Cat.No. 4376485).Capillary electrophoresis was performed using ABI 3500 ® Genetic analyzer (Thermo Fisher Scientific Inc. USA).The electropherogram obtained from the ABI 3500 Genetic Analyzer is exported to the sequence analysis software, Codoncode ® aligner program version 7.0.

Comparison of sequences by Basic Local Alignment
Search Tool (BLAST) from National Center for Biotechnology Information (NCBI).Pathogenicity of variants was identified from different public databases such as dbSNP (NCBI, NIH USA), ClinVar (NCBI, NIH USA), and/or COSMIC database (Sanger Institute, UK).

Data management and analysis
Recorded data were analyzed using the statistical package for social sciences, version 23.0 (SPSS Inc., Chicago, Illinois, USA).Shapiro-Wilk test was done to test the normality of data distribution.Data are presented according to their type, qualitative data presented as frequency and percentage, quantitative continues group expressed by mean ± SD.We used fisher exact test to examine the relation between qualitative variables.For quantitative data, differences between quantitative independent 3 groups were done by parametric One-Way ANOVA test which used for normally distributed quantitative data.Survival analysis was done using Kaplan Meier method.A p value < 0.05 was considered significant.

Results
A total number of 20 newly diagnosed (De-novo) AML were enrolled in our study.The demographic profile of our 20 patients showed a slight female predominance, with 40% males and 60% females.The age range is broad, spanning from 17 to 70 years, with a mean age of 43.6 ± 15.6 years (Table 1).
Hepatomegaly was observed in a minimal proportion, with only 5% of patients showing positive results.Similarly, splenomegaly is identified in 10% of patients as shown in Table 2. CBC, BM examinations parameters and Immunophenotyping analysis are shown in Tables 3  and 4.
Cytogenetic analysis revealed positivity in 35% of cases; while, French-American-British classification system of AML (FAB) indicated diverse subtypes, emphasizing the heterogeneity of the mini-cohort (Figs. 1, 2).
Our study aimed to characterization of the DNMT3a exon 23 mutations and showing its types among our De-novo AML patients.Overall, DNMT3a mutations were present in 25% of our patients.The R882 mutation was found to be positive in 10% of cases.Meanwhile, non-R882 mutations were identified in 15% of patients.(Tables 6, 7).
Rs47001633 missense mutation at chr2:25234373 (R882) position was found in two patients, case 1 and case 10.Case 1 was heterozygous C > G and in Case 10 was homozygous C > A leading to replacement of Arginine amino-acids at position 882 with Cysteine and Histidine, respectively (Fig. 3).
In cases 5, 6 and 16, non-R882 mutations were detected which were all found to be intron variants.SNP Mutation rs11695471 (chr2:25234839 T > A) was detected in cases 5 and 6 being homozygous in case 5 and heterozygous in case 6; while, mutation rs143635198 (chr2:25234772 T > C) was detected in cases 6 and 16 being heterozygous in case 6 and homozygous in case 16 (Figs.3, 4).
Hence, our 20 patients will be divided into 3 groups according to type of mutation: Group A: Non-mutated patients.Group B: Patients with Mutation R882.Group C: Patients with Mutation non-R882.When comparing the three groups together, there were no statistically significant difference between demographic/organomegaly groups (Tables 8, 9).
As for comparing laboratory results among the three mutation groups, Mean hemoglobin levels (Hb) showed no significant difference (p = 0.976) among the groups.Total leukocyte count (TLC) and peripheral blast percentage also exhibited no significant variance (p = 0.701 and p = 0.452, respectively).Platelet count (PLT) demonstrated a significant difference (p = 0.024*), with Group B showing higher levels which may give a link between R882 mutation and higher platelet count (Table 10, Fig. 5).
In terms of bone marrow blast percentage, FAB classification, and cytogenetic analysis, no statistically significant differences were observed among the groups which may be due to small sample size.However, FAB classification indicated a trend toward significance (p = 0.248) which needs further studies.Moreover, DNMT3a mutations were found in 30% of CN-AML patients (Table 10).
Comparison of immunophenotyping analysis among the three mutation groups exhibited no significant differences among them except for Myeloperoxidase positivity which was significant among the groups where only one patient in the non-R882 group was MPO negative (p = 0.05*) (Table 11).
Comparing the response to induction therapy among the groups showed the mean bone marrow (BM) blast percentage post-treatment is significantly different among the groups (p < 0.001), with Group B (R882 Mutation) exhibited a notably higher mean (6.0%) compared to Groups A (1.3%) and C (2.3%).Minimal residual disease (MRD) blast percentage did not show a significant difference among the groups (Table 12).
In terms of treatment response, the distribution of patients achieving partial remission (PR) was significantly different (p = 0.038), with 100% in Group B (R882 Mutation) achieved PR; while, Groups A and C showed lower PR rates.Similarly, the follow-up after 3 months indicated a significant difference in PR rates among the groups (p = 0.034), with Group B (R882 Mutation) having higher PR rates (Table 12).
Kaplan-Meier curve for Response to induction and follow-up after 3 months (Fig. 6) showed great effect of R882 mutation on the OS other than Wild type DNMT3a and non-R882 mutation.

Discussion
Acute myeloid leukemia (AML) is a clonal hematopoietic stem cell malignancy characterized by accumulation of immature progenitor cells with arrested differentiation leading to suppression of hematopoiesis.AML is heterogeneous with respect to morphology, immunophenotype, (cyto) genetic and epigenetic signatures and responses to treatment, including patient outcomes [9].
DNMT3a encodes a DNA methyltransferase that regulates the epigenetic modification of gene expression by catalyzing the addition of a methyl group to the cytosine residue of cytosine guanine dinucleotides.DNMT3a mediates DNA methylation involved in the differentiation of hematopoietic stem cells into a predominantly granulocytic lineage.DNMT3a mutations are relatively common in myeloid neoplasia and are detected in 20% of AML patients [10].
DNMT3a mutation is the commonest aberration in clonal hematopoiesis of indetermined potential (CHIP), a pre-leukemic state associated with a very increased risk of developing myeloid neoplasm, including AML.The broad clinical context in which DNMT3a mutation occurs has made it difficult to study the prognostic effect of DNMT3a mutations in AML specifically; moreover, numerous studies have revealed varied clinical correlations among DNTM3A-mutated AML patients associated with specific variant type (e.g., R882 vs. non-R882) [11].
We aimed to study DNMT3a mutations in 20 Egyptian AML patients free from either solid or hematological malignancies and to elucidate its prognostic influence on AML disease course.Our patients were recruited from hematology units of Internal Medicine department, Ain shams University Hospitals during the period from July 2022 to April 2023.
Diagnosis was based on updated WHO 2016 diagnostic criteria and its 5th edition paper.Our AML patients were then followed up to evaluate their response to induction after 28 days of chemotherapy and after 3 months both morphologically and via MRD quantification by multiparameter flow cytometry.DNMT3a hot spot exon 23 mutation was characterized using Sanger Sequencing technique.
Our results revealed that overall DNMT3a mutations were present in 25% of our patient consistent with Park et al. [6] findings who detected DNMT3a mutations in approximately 20% of AML cases.While it was present in 20% among cytogenetically normal AML in our study, Park et al. [6] found that it increased to 29.5% when  analysis was restricted to CN-AML which might be due to bigger sample size of his study (n = 142).
In addition, Ley et al. detected DNMT3a mutations in approximately 20% of AML cases [12].Interestingly, Loghavi et al. found DNMT3a mutations in up to 48% of their studied population which may be due to the variety in the ethnic group of her study which included Hispanics, African and Asian individuals [13].
In our study R882 mutation was positive in 10% of cases which is almost similar to Yuan et al.where 8.51% carried the DNMT3a R882 mutation [14] and Blau et al.where they found DNMT3a R882 in 11% of their subjects [2].Fifty percent of the R882 group were R882C mutated and 50% R882H mutated, where Yuan et al. found that 72% of R882 mutated cases were positive for the R882H mutation, 25% with the R882C mutation, and 3% cases with the R882P where [14].Ethnicity mostly played an important role in that difference where Yuan studied Chinese population.
Regarding Gender and age, our results showed no significant difference in the distribution of gender within the groups.Similarly, age analysis indicated no significant variance (p = 0.461) among the group.In contrast, Loghavi et al. found that patients with DNMT3a mutation were significantly younger (median 56.0 vs. 62.0 years; p = 0.025) and mostly were women (65.7% vs. 46.9%;p = 0.045), similar to what was also found by El-Rhman et al. whose patients were significantly younger (p = 0.005) [13,15].Our smaller sampling size might have played role in this, where Loghavi et al. and Rhman et al. bigger sample sizes might have marked that difference (n = 178 and 100, respectively) [13,15] Similar to El-Rahman et al., our study showed no significant difference with respect to liver and spleen enlargement [15].
In our study, platelet count (PLT) demonstrated a significant difference (p = 0.024), with R882 mutation showing higher levels which may give a link between R882 mutation and higher platelet count, which is similar to what Veninga et al. found and attributed it to methylation error of the mutation affecting megakaryopoiesis [16].Mean hemoglobin levels (Hb), total leukocyte count (TLC) and peripheral blast percentage exhibited no significant variance.In this aspect, Hou et al. had higher WBC, blast, and platelet counts than DNMT3a-wild patients (P value 0.0018, 0.0012, and 0.0001, respectively) where Chen et al. showed only highly significant difference in TLC (< 0.0001) with no significance Hb, PLT and peripheral blast percentage.El-Rhman et al. showed only significance in TLC and peripheral blast count (0.03 and 0.02, respectively) [15,17,18].Immunophenotyping analysis among the three mutation groups revealed that most markers, including CD19, CD36, CD7, CD56, CD33, CD13, CD14, CD117, CD64, CD34, and HLA-DR, exhibited no significant differences among the mutation groups.Obviously, Myeloperoxidase positivity was significant among the groups (0.05), which may be due to small sample size.However, it is similar to Kamijo et al. who found that DNMT3A mutation was more likely to be present in the MPO-low group (p = 0.001) [19].
Cytogenetic analysis revealed positivity in 35% of all our AML cases with t(15:17) and inv( 16) each representing 10% of the cases, t(8;21) representing 15% of cases and rest of cases are negative cytoFISH analysis.As regards DNMT3a mutated cases, only one case of the mutated DNMT3a had positive inv (16) by FISH, where the rest were FISH negative.To our knowledge, AML with inv (16), and inv(3)/t(3;3) rarely harbored the DNMT3a mutation, and this is the first reported case of DNMT3a mutation in AML with inv (16).This is similar to Park et al. who stated that DNMT3a mutations were more frequently identified in patients with cytogenetically normal (CN)-AML (p = 0.0112).In addition,  Laghavi et al. notably found that none of their patients had recurrent AML-associated cytogenetic abnormalities.Similarly El-Rhamn et al. concluded that DNMT3a occurred exclusively in CN-AML.Yet, cytogenetics FISH revealed no significant difference among our study groups (0.536%) [6,13,15] As for response to induction therapy, the mean bone marrow (BM) blast percentage was significantly different among the groups (p < 0.001), with R882 Mutation group exhibiting a notably higher mean (6.0%) compared to DNMT3a wild (1.3%) and non-R882 mutation (2.3%) implementing resistance to standard chemotherapy.This was reflected on treatment response, where the distribution of patients achieving partial remission (PR) was significantly different (p = 0.038), with 100% in R882 Mutation group achieving PR; while, Groups DNMT3a wild who stated that pre-leukemic hematopoietic clones with DNMT3a mutations may be resistant to leukemic therapy and may lead to further clonal expansion during remission, and may eventually cause recurrent disease [6,14].
After 28 days of follow-up post-induction, 75% of all our patients achieved complete response (CR), 20% achieved partial response (PR) out of which 75% are DNMT3a mutated, and 5% of all patients succumbed to treatment-related mortality which was DNMT3a wild type.
After a 3-month follow-up, 50% of all our patients maintained CR, 40% exhibited PR where half of them were DNMT3a mutated patients and 10% faced mortality where half of them was R882 mutated which represents 20% of all DNMT3a mutated patients.The follow-up after 3 months indicated a significant difference in PR rates among our groups (p = 0.034), with R882 Mutation group having higher relapse rates.In addition, Kaplan Meier survival analysis showed poor survival rate for R882 mutated group.Coinciding with our results, Park et al. concluded that DNMT3a mutated AML showed poor OS and eventfree survival (EFS), compared to that from the DNMT3a wild-type AML (p = 0.0484 and p = 0.0012, respectively) and they analyzed the prognostic effect of DNMT3a mutations in CN-AML and found that the DNMT3a-mutated CN-AML showed poorer OS and EFS compared to that shown by the DNMT3a wild-type CN-AML (p = 0.0376 and p = 0.0019).Similarly, Loghavi et al. concluded that in their study group, patients with DNMT3a mutations had worse outcomes than those with wild-type DNMT3a.El-Rhaman et al. suggested in their study that DNMT3a mutations were accompanied by worse outcome including significantly shorter OS and EFS and are an independent factor of worse outcome in younger patients with cytogenetic normal AML (CN-AML) [6,13,15].However, Yuan et al. indicated that in spite of the association between DNMT3a R882 mutations and worse outcome, it is obscure whether the mutations are associated with response to anti-leukemic therapies, and whether DNMT3a mutant types at amino acid 882 or allele burden influence prognosis of AML which had been rooted to dependency on DNMT3A R882 mutant-allele ratio [14].The R882 mutation is associated with higher platelet counts, resistance to chemotherapy, and poorer outcomes, highlighting its prognostic significance in AML.
Fig. 1 FAB classification among the patients

Fig. 2
Fig. 2 Cytogenetics FISH distribution among the patients

Table 7 Fig. 3 Fig. 4
Fig. 3 Displaying the R882 mutation in the Sanger Sequencing raw data, where A is forward Sequencing and B is Reverse Sequencing, where Case1 showed: Heterozygous mutation at R882 position (rs147001633 C > G), Case10: Homozygous mutation R882 position (rs147001633 C > A)

Fig. 5
Fig. 5 Comparison of lab results

Table 1
Demographic data of the patients

Table 2
Organomegaly among the patients

Table 3
Laboratory results among all our AML patients

Table 4
Showing Immunophenotyping analysis among our AML Patients

Table 5
Response to induction therapy among the Patient

Table 6
Showing distribution of DNMT3a among our patients

Table 8
Comparison of demographic data among mutationsUsing: § = One-way ANOVA test t for Mean ± SD; * = Fisher exact test, when appropriate p value > 0.05 is insignificant; *p value < 0.05 is significant

Table 9
Comparison of Organomegaly among mutations Using: Fisher's exact test, when appropriate p value > 0.05 is insignificant; *p value < 0.05 is significant

Table 10
Comparison of lab results among mutationsUsing: § One-way ANOVA test t for Mean ± SD; * = Fisher exact test, when appropriate p value > 0.05 is insignificant; *p value < 0.05 is significant

Table 11
Comparison of Immunophenotyping analysis among mutations Fisher exact test, when appropriate p value > 0.05 is insignificant; *p value < 0.05 is significant * =and non-R882 mutation showed lower PR rates.Minimal residual disease (MRD) blast percentage did not show a significant difference among our groups.This is coinciding with Yuan et al.where the pointed that DNMT3a R882 mutations might enhance chemoresistance to induction regimens including anthracyclines.Compared to patients without R882 mutations, those DNMT3a R882 mutations positive patients showed a significantly lower CR rate after the first cycle of induction therapy (28.77% vs. 42.48%,p = 0.023).This could be explained by Park et al.
Fig. 6 Kaplan-Meier curve for response and follow-up among patients Conclusion