Skip to main content
NCBI home page
Cancer Reports logoLink to Cancer Reports
. 2018 Jul 27;1(2):e1111. doi: 10.1002/cnr2.1111

Aberrant DNA methylation at HOXA4 and HOXA5 genes are associated with resistance to imatinib mesylate among chronic myeloid leukemia patients

Marjanu Hikmah Elias 1, Husin Azlan 2, Sarina Sulong 4, Abdul Aziz Baba 3, Ravindran Ankathil 4,✉
PMCID: PMC7941502  PMID: 32721103

Abstract

Background

Imatinib mesylate is a molecularly targeted tyrosine kinase inhibitor drug. It is effectively used in the treatment of chronic myeloid leukemia (CML) patients. However, development of resistance to imatinib mesylate as a result of BCR‐ABL dependent and BCR‐ABL independent mechanisms has emerged as a daunting problem in the management of CML patients. Between these mechanisms, BCR‐ABL independent mechanisms are still not robustly understood.

Aim

To investigate the correlation of HOXA4 and HOXA5 promoter DNA hypermethylation with imatinib resistance among CML patients.

Methods and results

Samples from 175 Philadelphia positive CML patients (83 good response and 92 BCR‐ABL non‐mutated imatinib resistant patients) were subjected to Methylation Specific High Resolution Melt Analysis for methylation levels quantification of the HOXA4 and HOXA5 promoter regions. Receiver operating characteristic curve analysis was done to elucidate the optimal methylation cut‐off point followed by multiple logistic regression analysis. Log‐Rank analysis was done to measure the overall survival difference between CML groups. The optimal methylation cut‐off point was found to be at 62.5% for both HOXA4 and HOXA5. Chronic myeloid leukemia patients with ≥63% HOXA4 and HOXA5 methylation level were shown to have 3.78 and 3.95 times the odds, respectively, to acquire resistance to imatinib. However, overall survival of CML patients that have ≤62% and ≥ 63% methylation levels of HOXA4 and HOXA5 genes were found to be not significant (P‐value = 0.126 for HOXA4; P‐value = 0.217 for HOXA5).

Conclusion

Hypermethylation of the HOXA4 and HOXA5 promoter is correlated with imatinib resistance and with further investigation, it could be a potential epigenetic biomarker in supplement to the BCR‐ABL gene mutation in predicting imatinib treatment response among CML patients but could not be considered as a prognostic marker.

Keywords: chronic myeloid leukemia, high‐resolution melt analysis, hypermethylation, imatinib mesylate, resistance

1. INTRODUCTION

Chronic myeloid leukemia (CML) accounts for 14% of all leukemias and 20% of all adult leukemias.1 CML results from the aberrant tyrosine kinase activity of the BCR‐ABL oncoprotein, which is encoded by the BCR‐ABL oncogene, formed as a result of Philadelphia chromosome translocation t(9;22)(q34;q11). Imatinib mesylate (IM), the molecularly targeted drug, has shown remarkable efficacy against BCR‐ABL oncogene protein and has become the current first line treatment for CML. Despite its remarkable efficacy, a significant proportion of CML patients treated with IM have been noted to have an unsatisfactory effect because of failure to respond, or to relapse because of primary or acquired resistance.2, 3 CML patients are judged to be IM resistant when response is lost or is not seen with a daily dose of >400 mg IM based on European LeukemiaNet.4, 5 Development of resistance to IM has emerged as a daunting problem in the management of CML patients.

Several pathways have been proposed and studied worldwide to understand the development of IM resistance in CML patients. Causative mechanisms of resistance to IM have been classified under two broad categories1: resistance involving BCR‐ABL dependent pathways and2 resistance involving BCR‐ABL independent pathways. BCR‐ABL dependent pathways, involving various BCR‐ABL mutations,6, 7 were reported to be the most common cause of IM resistance.8 However, several reports including ours, indicate that BCR‐ABL mutations account for only 19% to 63% of resistance to IM in CML patients.9, 10, 11, 12 The cause of IM resistance among BCR‐ABL mutation‐free CML patients can be postulated to be due to aberrant BCR‐ABL independent mechanisms involving other alternative signaling or epigenetic pathways. This is because, apart from BCR‐ABL1 protein, several other proteins like transcription factors, transporter protein, and many other proteins are also important in regulating the high rate of proliferation and the suppression of apoptosis.13, 14

Abnormal epigenetic regulation of the expression of CML‐associated genes has been postulated15 to play a critical role in its pathogenesis and in the mechanisms modulating therapeutic responsiveness.16 Epigenetic silencing is a phenomenon whereby gene transcription may be suppressed through DNA methylation (a process that may regulate gene use) resulting in decreased protein expression. The methylation of the promoter is an alternative to coding region mutations in silencing tumor suppressor gene function and is observed in a non‐random tumor‐type specific manner. Aberrant hypermethylation has been observed in leukemia including CML.16, 17 Few studies18, 19, 20 have suggested that hypermethylation might play a role in disease progression in CML as hypermethylation of several genes were noted in progression of CML. The observation that increased epigenetic silencing of potential tumor suppressor genes correlates with disease progression in a small proportion of patients treated with IM by Issa et al (2007) suggested a relationship between epigenetic silencing and development of IM resistance.21

The homeobox (HOX) genes are a family of homedomain containing transcription factors present in four separate clusters (HOXA, HOXB, HOXC, and HOXD) which are key regulators of embryonic development,22 haematopoitic differentiation, and leukemogenesis.23, 24, 25, 26 The HOX gene is a 183 bp DNA sequence and codes a 61 amino acid domain known as homeodomain. It is typically located at a terminal or sub‐terminal position of the corresponding homeoprotein and is responsible for recognizing and binding to the sequence specific targeted DNA.25 The HOX gene networks are believed to encode master regulators in haematopoiesis and are associated with the development of haematologic malignancies, prognosis, and treatment response.27 Interestingly, DNA methylation was reported to have an important role in aberrant control of HOX gene expression during the development of leukemia. DNA hypermethylation of several HOX genes was believed to be associated with progression of CML into blast phase.28, 29 We hypothesized that DNA hypermethylation of several HOXA genes is correlated with resistance to IM in CML patients. Thus, 2 HOXA genes, the HOXA4 gene (OMIM#142953) and HOXA5 gene (OMIM#142952), that are located at 7p14.2 to 7p15 were selected as candidate genes to be investigated for their involvement in mediating IM resistance. In our previous study,30 we only reported on the correlation between HOXA4 promoter hypermethylation and resistance to IM in CML patients. Further to that, the present study incorporates (1) data of HOXA4 with more number of patients, (2) original data on methylation levels of HOXA5, (3) the correlation of HOXA4/HOXA5 promoter hypermethylation (alone or combination) with IM resistance, and (4) the correlation of overall survival with HOXA4/HOXA5 promoter hypermethylation.

2. METHODOLOGY

The current study was done at Human Genome Centre, Universiti Sains Malaysia. This study has been approved by the Ministry of Health, Malaysia (NMRR‐10‐1206‐7127) as well as Research and Ethics Committee of Universiti Sains Malaysia. A total of 175 Philadelphia chromosome‐positive CML patients were recruited. Those CML patients were in chronic, accelerated, or blast phase that has been treated for at least 12 months with IM as a frontline treatment. The diagnosis of each patient was confirmed hematologically, cytogenetically as well as via molecular analysis.

The CML patient's response to IM treatment was assessed based on the hematologic, cytogenetic, and molecular responses results according to European LeukemiaNet 2010.31 In European LeukemiaNet 2010, an IM resistant CML patient was defined as showing lack of complete cytogenetic response (CCyR) by 12 months and/or major molecular response by 18 months after started the treatment.31 Based on these assessment criteria, the CML patients were categorized according to their response to IM (good response and resistance).

2.1. Sample collection and DNA extraction

Peripheral blood (3 mL) from CML patient was collected in EDTA tube. Genomic DNA was extracted and was kept in −40°C before subjected to bisulfite treatment. Genomic DNA of patients and controls was extracted and purified using the GENTRA PUREGENE Blood Kit (Qiagen, Germany) according to the recommendation. DNA concentration was quantitated by using NanoQuant Infinite M200 (Tecan, Switzerland), and the quality of the DNA was confirmed using agarose gel electrophoresis.

2.2. Bisulfite treatment

Five hundred nanograms of the DNA was treated with bisulfite using EZ DNA Methylation‐Gold Kit (ZYMO Research, USA) following manufacturer's recommendation. The concentration of the bisulfite treated DNA was measured as described previously.30

The 100% and 0% methylation DNA control used were universal methylated DNA and unmethylated DNA (ZYMO research, USA). Both DNA methylation control were treated with bisulfite, and a series of methylation dilution were subsequently prepared according to the ratio of 10%, 25%, 50%, and 75%. This set of serial DNA methylation percentages was included in each experimental run.

2.3. MS‐HRM analysis

MS‐HRM primers (Table 1) were designed using Methyl Primer Express v1.0 Software (Applied Biosystem, USA) based on criteria suggested by Wojdacz et al (2008) with some modifications.

Table 1.

Primer sequences used for MS‐HRM of HOXA4 and HOXA5 promoter amplification

Gene Name Sequence Size of Amplicon Number of CpG Annealing Temperature
HOXA4 HOXA4_fw 5′ TTTTGAAGGATACGAAGTTTGA 3′ 115 bp (−519 to −399) 9 CpG 50.0 °C
HOXA4_rv 5′ TCCTCTCGAAAACCCTCTAC 3′
HOXA5 HOXA5_1fw 5′ AGAGCGGTAGGGATATTGTTT 3′ 169 bp (−754 to −586) 7 CpG 51.2 °C
HOXA5_1rv 5′ ACTAACCGTCCTACCAACAACT 3′
HOXA5_2fw 5′ AGGTAGGATTTATGATTGGATAAT 3′ 140 bp (−186 to −47) 6 CpG 54.5 °C
HOXA5_2rv 5′ ACATCTATAACACCCTTACACAAT 3′
Open in a new tab

MS‐HRM analysis was done using CFX Real Time PCR Detection System (Bio‐Rad Laboratories, USA). MS‐PCR amplification was monitored using CFX Manager Software, and the HRM data was analysed using Bio‐Rad Precision Melt Analysis Software. Total volume of PCR reaction used was 10 μL, comprising of 1X Precision Melt Supermix (Bio‐Rad Laboratories, USA), 200 nM of each primer, and 20 ng of bisulfite‐treated DNA. Technical replication was performed in triplicate. PCR condition used was as follows: (1) initial denaturation at 95°C for 2 minutes and 50 cycles of (2) denaturation at 95°C for 10 seconds, (3) annealing at 50°C to 54.5°C (as shown in Table 1) for 30 seconds, and (4) extension at 72°C for 30 seconds. Heteroduplex formation was done after PCR amplification at 95°C for 30 seconds and subsequently at 60°C for a minute. HRM analysis was performed immediately after heteroduplex formation by 0.2°C increments, and the temperature was held for 10 seconds of each increment. The increments were done from 65°C to 95°C. A no template control and set of serial percentage control (0%, 10%, 25%, 50%, 75%, and 100%) in triplicate were included in each experimental run.

MS‐HRM results were then used to construct difference curve. Normalized melt curve was then developed by choosing the 0% melt curve as the reference relative fluorescent unit (RFU), using Bio‐Rad Precision Melt Analysis Software. The peak reading of difference‐RFU from the difference curve for all the 0%, 25%, 50%, 75%, and 100% methylation control levels was plotted, and a “best‐fit” straight line through the data was drawn, and the linear regression analysis was applied to the data using excel (Figure 1). Methylation level percentages of all IM treated CML patients were subsequently calculated based on the linear regression equation of the difference‐RFU peaks.

Figure 1.

Figure 1

Open in a new tab

A, Normalized melt curve, B, difference curve of the 0%, 25%, 50%, 75%, and 100% methylation control, and C, standard curve from difference curve peak of all methylation control that was obtained from MS‐HRM data, plotted, and the equation generated was used for the samples methylation percentage calculation of the samples

2.4. Statistical analysis

Statistical Package for the Social Sciences (SPSS) Software version 20.0 was used for statistical analysis. ROC curve was generated from the numerical data to determine the optimal cutoff point of HOXA4 and HOXA5 methylation levels in discriminating the IM good response and resistance CML patients. From the ROC curve, an optimal cutoff point was identified by employing the “points on curve closest to the (0, 1)” method.32 Later, the numerical data of HOXA4 and HOXA5 methylation levels were divided into 2 groups according to the cut‐off point calculated. Simple logistic regression was then used to identify the risk of developing IM resistance by determining the crude odd ratios (ORs) and 95% confidence interval (CI). Multiple logistic regression analysis was used to evaluate the relationship between HOXA4 as well as HOXA5 promoter methylation levels and IM response of CML patients by calculating the ORs and 95% CI. The analysis was performed by SPSS software (version 20) with all P‐values as 2‐sided. A 5‐year survival analysis was performed by using Cox proportional hazard model involving simple Cox regression deriving hazard ratio for univariate analysis and Kaplan Meier curves were plotted. The survival distributions for univariate were compared using log‐rank test. A P‐value of less than 0.05 was considered as statistically significant.

3. RESULTS

3.1. Demographic profile

In this study, 175 Philadelphia chromosome positive CML patients comprising both IM good response (n = 83) as well as IM resistant (n = 92) groups were recruited. Among the 83 IM good response CML patients, 42 are male and 41 are female with mean age of 42 (SD = 14.33) while among 92 IM resistant CML patients, 43 are male and 49 are female with mean age of 43 (SD = 14.54). Among the BCR‐ABL non‐mutated IM resistant CML patients, 32 (34.8%) were in warning group, 40 (43.5%) in primary resistance group, and 20 (21.7%) in secondary resistance group.

3.2. Correlation between methylation level and IM response

In the MS‐HRM analysis, the well‐separated and nearly equal deviation between all percentages of the methylation control in the normalized melt curve (Figure 1A) and the difference curve (Figure 1B) showed an acceptable sensitivity of the analysis. In each run, the difference curve peak of all methylation controls was obtained, plotted, and the coefficient of determination (r 2) was found to be higher than 0.9 (Figure 1C).

From the MS‐HRM analysis of IM good response and resistant CML patients, the methylation percentage of the HOXA4 (519 to −399 from transcription start site [TSS] of HOXA4) was found to be scattered from 10% to 100%. However, for HOXA5, all patients regardless of their response to IM showed methylation level of more than 75% when analysed using HOXA5_1 primers set (Table 1). This indicated that promoter region at −754 to −586 of HOXA5 TSS is highly methylated regardless of their response to IM. Thus, this region was regarded as not appropriate to be an epigenetic marker for IM resistance identification. Interestingly, HOXA5_2 primer set (Table 1) used to amplify −186 to −47 from HOXA5 TSS showed the methylations percentage were scattered from 0% to 100%.

In HOXA4, the mean methylation levels were found to be higher in IM‐resistant CML patients (mean = 64.52; SD = 19.51) compared with IM good response CML patients (mean = 52.25; SD = 17.04). Similarly in HOXA5, the mean methylation levels were found to be higher in IM resistant CML patients (mean = 71.42; SD = 20.98) compared with IM good response CML patients (mean = 52.34; SD = 27.28). Subsequently, from the ROC curve analysis, the methylation levels of HOXA4 and HOXA5 promoter between IM good response and IM resistance CML patients were discriminated, considering the optimal methylation level cutoff point as 62.5 for both HOXA4 (0.644 sensitivity, 0.723 specificity) and HOXA5 (0.741 sensitivity, 0.402 specificity). Figure 2A,B shows the ROC curve for HOXA4 and HOXA5, respectively. Based on the methylation cutoff point, 60.0% of the CML patients showed a parallel methylation level in HOXA4 and HOXA5. A total of 45.8% of the IM good response CML patients had methylation lower than 62.5% in both HOXA4 and HOXA5, while 50.0% of the IM resistant CML patients had methylation higher than 62.5% in both HOXA4 and HOXA5. Others showed different methylation status between HOXA4 and HOXA5.

Figure 2.

Figure 2

Open in a new tab

ROC curve of A, HOXA4 with optimal cut‐off point at 62.5% (0.644 sensitivity, 0.723 specificity, 95% CI 0.624, 0.782) and B, HOXA5 with optimal cut‐off point at 62.5% (0.741 sensitivity, 0.598 specificity, 95% CI 0.63, 0.79)

Based on the final model of multiple logistic regression, a CML patient with ≥63% HOXA4 and HOXA5 methylation level had 3.78 and 3.95 times, respectively, the odds to acquire resistance to IM when adjusted for each other (Table 2). For a better clinical point of view, European LeukemiaNet 2013 classification was employed and the risk association of HOXA4 and HOXA5 methylation levels with the IM warning, primary resistant, and secondary resistant groups of CML patients was evaluated. The mean promoter methylation level of HOXA4 among the IM optimal response group was significantly different from IM primary (P < 0.001) and secondary resistant (P < 0.001) CML patients, but not from the IM warning response group (P = 0.093). Interestingly, unlike in HOXA4, the mean promoter methylation level of HOXA5 among the IM optimal response CML patients was significantly different from all IM warning group (P = 0.005), IM primary (P = 0.001), and secondary resistant (P = 0.006) CML patients (Table 3).

Table 2.

Risk association of HOXA4 and HOXA5 methylation levels with the development of IM resistance in CML patients

Genes Methylation Percentage Simple Logistic Regression Multiple Logistic Regressiona
Regression coefficient (b) Crude OR (95% CI) P value Regression coefficient (b) Adjusted OR (95% CI) P value
HOXA4 0–62% 0 1.00 0 1.00
63–100% 1.55 4.71 (2.46,9.03) <0.001* 1.15 3.78 (1.89,7.55) <0.001*
HOXA5 0–62% 0 1.00 0 1.00
63–100% 1.45 4.26 (2.22,8.17) <0.001* 1.38 3.95 (2.00,7.82) <0.001*
Open in a new tab
*

P value <0.05 is significant.

a

Forward LR multiple logistic regression model was applied.

Multicollinearity and interaction term checked and not found.

Hosmer‐Lemeshow test, (P‐value = 0.940), classification table (overall correctly classified percentage = 68%) and area under the ROC curve (74.9%) were applied to check the model fitness.

Table 3.

Risk association of the HOXA4 and HOXA5 methylation groups, divided based on the 62.5% cutoff point towards the 4 response groups to IM treatment

IM response Groups HOXA4 HOXA5
Regression coefficient (B) Crude odds ratio (95% CI) P‐value Regression coefficient (B) Crude odds ratio (95% CI) P‐value
Optimal 0 1.00 ‐ 0 1.00 ‐
Warning 0.75 2.12 (0.88,5.09) 0.093 1.33 3.78 (1.50,9.54) 0.005*
Primary 1.95 7.04 (2.95,16.82) <0.001* 1.50 4.48 (1.88,10.69) 0.001*
Secondary 2.12 8.35 (2.74,25.41) <0.001* 1.53 4.61 (1.54,13.79) 0.006*
Open in a new tab
*

P < 0.05, significant at 95% CI.

3.3. Correlation between methylation level and overall survival

Overall survival of IM treated CML patients was also evaluated based on their HOXA4 and HOXA5 promoter hypermethylation levels. Figure 3A,B shows the Kaplan Meier curves for HOXA4 and HOXA5, respectively. However, from the Log‐Rank analysis, there was no significant difference in the overall survival between CML patients in group that have ≤62% and ≥63% methylation levels (P‐value = 0.126 for HOXA4; P‐value = 0.217 for HOXA5). Table 4 shows the Cox regression analyses for the HOXA4 and HOXA5 methylation levels.

Figure 3.

Figure 3

Open in a new tab

Kaplan Meier curves of A, HOXA4 and B, HOXA5

Table 4.

Cox regression analyses showing the value of HOXA4 and HOXA5 methylation level in predicting the risk of mortality in IM treated CML patients

Variables Regression Coefficient (b) Crude Hazards Ratio (95% CI) Wald Statistic P‐Value
HOXA4 0%–62% 0 1 ‐ ‐
63%–100% 1.53 4.60 (0.54, 39.38) 1.94 0.164
HOXA5 0%–62% 0 1 ‐ ‐
63%–100% 1.27 3.55 (0.42, 30.39) 1.34 0.248
Open in a new tab
*

P value <0.05 was significant.

4. DISCUSSION

It is well documented that hypermethylation of gene promoter associated CpG islands leads to inactivation of gene expression. The HOX genes are known to play crucial roles in the control of differentiation of adult haematopoietic cells.25 According to Shah and Sukumar (2010), normal HOX gene expression disruption in cancer affects various pathways that promote tumorigenesis, and metastasis, including the activation of anti‐apoptotic pathways, suppression of differentiation,33, 34 and necroptosis.35 From what is known so far, homeobox proteins interact with numerous regulatory pathways, including fibroblast growth factor, bone morphogenic protein, retinoic acid, sex steroid signaling, and also proteins involved in cell matrix interaction, such as integrins and intracellular adhesion molecule.36 From all clusters of HOX genes, aberrant methylation was reported to be most frequently occurring in the promoter regions of six HOXA cluster genes, with some of them showing leukemia type‐specific hypermethylation in all leukemia cell lines.37 Among the HOXA cluster genes, HOXA4 and HOXA5 genes are transcription factors that are known to regulate haematopoiesis38 and are associated with acute myeloid leukemia development.39 HOXA4 was proven to produce mature myeloid and lymphoid progeny in hematopoietic stem cells and was shown to be important in the regulatory mechanisms of controlling hematopoiesis within its natural context.40 HOXA5, on the other hand, works by regulating normal proliferation and apoptosis, and hypermethylation of HOXA5 is also known to regulate myeloid differentiation29 and necroptosis.35 Studies focusing on the potential oncogenic and tumour suppressive role of HOX gene family members in leukemia development have shown that inactivation of HOXA4 is associated with poor prognosis. In both adult myeloid and lymphoid leukemias, HOXA4 and HOXA5 were found to be frequently inactivated by promoter hypermethylation.

Increased level of DNA methylation was observed in HOXA4 gene and was reported to be associated with transcriptional repression in chronic lymphocytic leukemia41 and acute myeloid leukemia.29 In CML, inactivation of HOXA4 and HOXA5 genes by hypermethylation is strongly correlated with progression into blast crisis.29 Thus, the changes in level of HOXA4 and HOXA5 protein production will perturb the normal regulation of haematopoiesis and subsequently promote leukemogenesis.29 Even though HOXA4 and HOXA5 were reported to be associated with the poor prognosis in CML patients, their involvement in mediating IM resistance among IM‐treated CML patients has not been explored earlier. To assess the possible epigenetic component of IM resistance, we compared the methylation levels of HOXA4 and HOXA5 in Malaysian CML patients responsive and resistant to IM.29, 41

The present finding involving 175 CML patients also suggested that the methylation level of HOXA4 has a strong correlation with IM resistance among CML patients, which is in agreement with our earlier report on 95 CML patients.30 Additionally, in the present study, the involvement of HOXA5 methylation level towards IM resistance was also evaluated. Interestingly, instead of 50% cutoff point, a new optimal methylation level cutoff point of 62.5% was considered for both HOXA4 and HOXA5. This new cutoff point was determined via ROC curve which is more reliable than the methylation cutoff point used in our previous study. Based on the 62.5% methylation cutoff point, the response towards IM treatment among CML patients was more significantly differentiated (P < 0.001) for both HOXA4 and HOXA5 promoter methylation levels. Interestingly, from the association study, HOXA4 promoter methylation level was able to differentiate IM primary and secondary resistant CML patients from the IM optimal response but not from the IM warning response. However, apart from IM primary and secondary resistant CML patients, HOXA5 promoter methylation level was able to differentiate IM warning group also, from the IM optimal response group of CML patients. This present study demonstrated that DNA hypermethylation of HOXA4 and HOXA5 promoters could be a biomarker for predicting resistance towards IM treatment in CML patients. To the best of available knowledge, there are no other studies till date that had reported the contribution of HOXA4 and HOXA5 in mediating IM resistance in CML patients.

Several studies have proven that hypermethylation of HOXA4 is able to promote inactivation of its gene expression.39, 41, 42 HOXA4 and HOXA5 proteins are known as DNA‐binding transcription factors which may regulate gene expression, morphogenesis, and differentiation. Thus, it is reasonable to suggest that the suppression of HOXA4 and HOXA5 protein production by hypermethylation‐induced gene silencing might be promoting resistance to IM. Hypermethylation of HOXA4 was reported to be involved with the development of leukemia in several studies.39 In 2009, Zangenberg et al reported that 77% of their acute myeloid leukemia (AML) patients exhibited hypermethylation of HOXA4 promoter region.42

It is known that p53 gene plays an important role in regulating apoptosis. This programmed cell death associated gene was reported to be transcriptionally regulated by the HOXA5.43 HOXA5 has been shown to function as a transactivator of the P53 gene43 and to induce apoptosis by P53 dependent and P53 independent mechanisms.44 DNA methylation of the CpG island located in the 5′ end of the HOXA5 gene has been identified in breast43 and lung cancer45 and has been associated with loss of HOXA5 expression.43 As in breast cancer, where overexpression of HOXA5 has been found to induce apoptosis by upregulating the expression of p53,44 HOXA5 may also behave as a tumor suppressor gene in CML. Likewise hypermethylation of HOXA5 results in downregulation of its expression. Thus, downregulation of HOXA5 might be consequently reducing p53 protein and blocking the guardian effects of p53 on the genome which may lead to uncontrolled cell growth and disease progression.

Furthermore, Chen et al (2005) had demonstrated that HOXA5 may trigger the receptor‐mediated apoptotic pathways via the activation of NF‐κB signaling pathway. During leukemogenesis, NF‐κB signaling is blocked by the activation of PI3K‐Akt signaling pathway, a downstream pathway of BCR‐ABL, subsequently leading to reduced apoptosis. Therefore, it is reasonable to suggest that even though IM blocks BCR‐ABL tyrosine kinase activity, the disease continues to progress via the inactivation of NF‐κB by the down regulation of HOXA5. Thus, apart from BCR‐ABL gene, the failure in the regulation of normal HOXA5 expression, most probably due to hypermethylation, will also lead to the uncontrolled leukemic cell accumulation in leukemogenesis of CML.44

From the present study, it is reasonable to speculate that DNA hypermethylation induced gene silencing can provide an alternative to BCR‐ABL mutations in mediating resistance to IM in CML patients. Thus, further studies are needed for hypermethylation status of HOXA4 and HOXA5 to be applied as an additional marker of resistance to IM in supplement to BCR‐ABL mutation analysis. This is because, no known mechanism of IM activity towards HOXA4 and HOXA5 hypermethylation has been reported so far. Hence, mechanistic studies are still needed to prove how hypermethylation of HOXA4 and HOXA5 causes poor response to IM. Furthermore, the cause of hypermethylation particularly in both HOXA4 and HOXA5 is still not fully revealed. Arising‐mutation of DMNT3a could be eliminated as one of the causes of DNA hypermethylation as DMNT3a has been associated with lower level of DNA methylation in many genes, including HOXA4 and HOXA5.46 Another gene associated with epigenetic regulation is TET2. This TET2 gene is commonly mutated in CML patients; thus, it is related to the pathogenesis of CML as well as resistance towards TKI treatment.47 However, no studies have shown the association of TET2 mutation and HOXA4 and HOXA5 methylation level.

The association of HOXA4 and HOXA5 gene expression levels with overall survival of these CML patients was also investigated in the present study. From the overall survival analysis, no significant difference in the overall survival between patients with ≤62% and ≥63% of HOXA4 and HOXA5 methylation levels was observed. Thus, the levels of HOXA4 and HOXA5 promoter methylation could be considered as epigenetic markers for predicting the response to IM but could not be considered as a prognostic marker as the methylation level was not significantly associated with the poor prognosis among imatinib‐treated CML patients. There might be differences between particular genes being hypermethylated and their prognostic impact. We can hypothesize that hypermethylation of genes involved in specific cellular pathways may make these cells more prone and sensitive to chemotherapy treatment.

However, recent studies on myeloproliferative neoplasm have revealed that 5‐mC is converted to an intermediate state of 5‐hmC before further demethylated into C.48, 49 Unfortunately, MS‐HRM which was used in this study utilized the principle of bisulfite analysis that recognizes both 5‐mC and 5‐hmC as C. Thus, the level of expression could be affected in all methylation‐positive samples as those samples also may contain 5‐hmC. Oxidative MS‐HRM together with MS‐HRM will help to distinguish 5‐mC and 5‐hmC for more accurate determination of methylation level that may reflect the expression level of HOXA4 and HOXA5.

In conclusion, the present study result suggests that promoter hypermethylation of HOXA4 and HOXA5 genes is correlated with IM resistance and could be a potential epigenetic biomarker in supplement to the BCR‐ABL gene mutation in predicting IM treatment response among CML patients.

AUTHORS' CONTRIBUTIONS

All authors had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Conceptualization, R.A.; Methodology, M.H.E.; Investigation, M.H.E.; Formal Analysis, M.H.E.; Resources, M.H.E.; Writing, M.H.E.; Original Draft, M.H.E.; Writing‐Review & Editing, R.A., H.A., S.S., A.A.B.; Visualization, R.A.; Supervision, R.A.; Funding Acquisition, R.A.

ACKNOWLEDGEMENT

This study was supported by Universiti Sains Malaysia, Malaysia, Research University (USM‐RU) Grant/Award Number: 1001/PPSP/812070. Exceptional appreciation to Dr Hoh Boon Peng and staff of Institute of Medical Molecular Biotechnology, Faculty of Medicine, Universiti Teknologi MARA for permitting to use the facilities used in this study. The co‐operation and support of staff, Human Genome Center, Universiti Sains Malaysia, and all the patients who have participated in this study are also gratefully acknowledged.

Elias MH, Azlan H, Sulong S, Baba AA, Ankathil R. Aberrant DNA methylation at HOXA4 and HOXA5 genes are associated with resistance to imatinib mesylate among chronic myeloid leukemia patients. Cancer Reports. 2018;1:e1111. 10.1002/cnr2.1111

REFERENCES

  • 1. Fehniger TA, Welch JS, Pluard TJ. The Washington Manual of Oncology: Lippincott Williams & Wilkins; 2007.
  • 2. Quintas‐Cardama A, Kantarjian HM, Cortes JE. Mechanisms of primary and secondary resistance to imatinib in chronic myeloid leukemia. Cancer Control. 2009;16(2):122‐131. [DOI] [PubMed] [Google Scholar]
  • 3. Zhang WW, Cortes JE, Yao H, et al. Predictors of primary imatinib resistance in chronic myelogenous leukemia are distinct from those in secondary imatinib resistance. J Clin Oncol. 2009;27(22):3642‐3649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Baccarani M, Saglio G, Goldman J, et al. Evolving concepts in the management of chronic myeloid leukemia: recommendations from an expert panel on behalf of the European LeukemiaNet. Blood. 2006;108(6):1809‐1820. [DOI] [PubMed] [Google Scholar]
  • 5. Kantarjian HM, Talpaz M, O'Brien S, et al. Dose escalation of imatinib mesylate can overcome resistance to standard‐dose therapy in patients with chronic myelogenous leukemia. Blood. 2003;101(2):473‐475. [DOI] [PubMed] [Google Scholar]
  • 6. Azevedo AP, Reichert A, Afonso C, Alberca MD, Tavares P, Lima F. BCR‐ABL V280G mutation, potential role in Imatinib resistance: first case report. Clin Med Insights Oncol. 2017;11:1179554917702870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Kaplan JB, Platanias LC. Another tyrosine kinase inhibitor‐resistance mutation within the BCR‐ABL kinase domain: chasing our tails? Leuk Lymphoma. 2017;58(7):1526‐1527. [DOI] [PubMed] [Google Scholar]
  • 8. Kantarjian HM, Talpaz M, Giles F, O'Brien S, Cortes J. New insights into the pathophysiology of chronic myeloid leukemia and imatinib resistance. Ann Intern Med. 2006;145(12):913‐923. [DOI] [PubMed] [Google Scholar]
  • 9. Elias MH, Baba AA, Husin A, et al. Contribution of BCR‐ABL kinase domain mutations to imatinib mesylate resistance in Philadelphia chromosome positive Malaysian chronic myeloid leukemia patients. Hematol Rep. 2012;4(4):e23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Kim SH, Kim D, Kim DW, et al. Analysis of Bcr‐Abl kinase domain mutations in Korean chronic myeloid leukaemia patients: poor clinical outcome of P‐loop and T315I mutation is disease phase dependent. Hematol Oncol. 2009;27(4):190‐197. [DOI] [PubMed] [Google Scholar]
  • 11. Qin Y, Chen S, Jiang B, et al. Characteristics of BCR‐ABL kinase domain point mutations in Chinese imatinib‐resistant chronic myeloid leukemia patients. Ann Hematol. 2011;90(1):47‐52. [DOI] [PubMed] [Google Scholar]
  • 12. Sherbenou DW, Hantschel O, Kaupe I, et al. BCR‐ABL SH3‐SH2 domain mutations in chronic myeloid leukemia patients on imatinib. Blood. 2010;116(17):3278‐3285. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Kalle AM, Sachchidanand S, Pallu R. Bcr‐Abl‐independent mechanism of resistance to imatinib in K562 cells: induction of cyclooxygenase‐2 (COX‐2) by histone deacetylases (HDACs). Leuk Res. 2010;34(9):1132‐1138. [DOI] [PubMed] [Google Scholar]
  • 14. Nambu T, Araki N, Nakagawa A, et al. Contribution of BCR‐ABL‐independent activation of ERK1/2 to acquired imatinib resistance in K562 chronic myeloid leukemia cells. Cancer Sci. 2010;101(1):137‐142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Machova Polakova K, Koblihova J, Stopka T. Role of epigenetics in chronic myeloid leukemia. Curr Hematol Malig Rep. 2013;8(1):28‐36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Heller G, Topakian T, Altenberger C, et al. Next‐generation sequencing identifies major DNA methylation changes during progression of Ph+ chronic myeloid leukemia. Leukemia. 2016;30(9):1861‐1868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Toyota M, Kopecky KJ, Toyota MO, Jair KW, Willman CL, Issa JP. Methylation profiling in acute myeloid leukemia. Blood. 2001;97(9):2823‐2829. [DOI] [PubMed] [Google Scholar]
  • 18. Jelinek J, Gharibyan V, Estecio MR, et al. Aberrant DNA methylation is associated with disease progression, resistance to imatinib and shortened survival in chronic myelogenous leukemia. PLoS One. 2011;6(7):e22110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Wang XJ, Li J, Fu BJ, Guo LL, Zhang JH, Huang SA. Methylation status of JunB and CDH13 gene promoter in CD34(+)CD38(−) chronic myelogenous leukemia cells. Zhongguo Shi Yan Xue Ye Xue Za Zhi. 2009;17(6):1405‐1408. [PubMed] [Google Scholar]
  • 20. Wang YL, Qian J, Lin J, et al. Methylation status of DDIT3 gene in chronic myeloid leukemia. J Exp Clin Cancer Res. 2010;29(1):54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Issa JP. DNA methylation as a therapeutic target in cancer. Clin Cancer Res. 2007;13(6):1634‐1637. [DOI] [PubMed] [Google Scholar]
  • 22. Deschamps J, Duboule D. Embryonic timing, axial stem cells, chromatin dynamics, and the Hox clock. Genes Dev. 2017;31(14):1406‐1416. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Abramovich C, Humphries RK. Hox regulation of normal and leukemic hematopoietic stem cells. Curr Opin Hematol. 2005;12(3):210‐216. [DOI] [PubMed] [Google Scholar]
  • 24. Buske C, Humphries RK. Homeobox genes in leukemogenesis. Int J Hematol. 2000;71(4):301‐308. [PubMed] [Google Scholar]
  • 25. Cillo C, Cantile M, Faiella A, Boncinelli E. Homeobox genes in normal and malignant cells. J Cell Physiol. 2001;188(2):161‐169. [DOI] [PubMed] [Google Scholar]
  • 26. Grier DG, Thompson A, Kwasniewska A, McGonigle GJ, Halliday HL, Lappin TR. The pathophysiology of HOX genes and their role in cancer. J Pathol. 2005;205(2):154‐171. [DOI] [PubMed] [Google Scholar]
  • 27. Kontro M, Kumar A, Majumder MM, et al. HOX gene expression predicts response to BCL‐2 inhibition in acute myeloid leukemia. Leukemia. 2017;31(2):301‐309. [DOI] [PubMed] [Google Scholar]
  • 28. Esteller M. Cancer epigenetics: DNA methylation and chromatin alterations in human cancer. Adv Exp Med Biol. 2003;532:39‐49. [DOI] [PubMed] [Google Scholar]
  • 29. Strathdee G, Holyoake TL, Sim A, et al. Inactivation of HOXA genes by hypermethylation in myeloid and lymphoid malignancy is frequent and associated with poor prognosis. Clin Cancer Res. 2007;13(17):5048‐5055. [DOI] [PubMed] [Google Scholar]
  • 30. Elias MH, Baba AA, Husin A, et al. HOXA4 gene promoter hypermethylation as an epigenetic mechanism mediating resistance to imatinib mesylate in chronic myeloid leukemia patients. Biomed Res Int. 2013;2013:129715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. LeukemiaNet E . Recommendations from the European LeukemiaNet for the Management of chronic myeloid leukemia (CML). update 2010 ed: European LeukemiaNet; 2010.
  • 32. Kumar R, Indrayan A. Receiver operating characteristic (ROC) curve for medical researchers. Indian Pediatr. 2011;48(4):277‐287. [DOI] [PubMed] [Google Scholar]
  • 33. Shah N, Sukumar S. The Hox genes and their roles in oncogenesis. Nat Rev Cancer. 2010;10(5):361‐371. [DOI] [PubMed] [Google Scholar]
  • 34. Domsch K, Papagiannouli F, Lohmann I. The HOX‐apoptosis regulatory interplay in development and disease. Curr Top Dev Biol. 2015;114:121‐158. [DOI] [PubMed] [Google Scholar]
  • 35. Alharbi RA, Pandha HS, Simpson GR, et al. Inhibition of HOX/PBX dimer formation leads to necroptosis in acute myeloid leukemia cells. Oncotarget. 2017;8(52):89566‐89579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Rodini CO, Xavier FC, Paiva KB, et al. Homeobox gene expression profile indicates HOXA5 as a candidate prognostic marker in oral squamous cell carcinoma. Int J Oncol. 2012;40(4):1180‐1188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Fang MH, Liu WL, Meng FK, Sun HY. Aberrant methylation at promoter region of HOX a gene cluster in leukemia cells. Zhonghua Xue Ye Xue Za Zhi. 2009;30(7):468‐472. [PubMed] [Google Scholar]
  • 38. Lawrence HJ, Largman C. Homeobox genes in normal hematopoiesis and leukemia. Blood. 1992;80(10):2445‐2453. [PubMed] [Google Scholar]
  • 39. Musialik E, Bujko M, Kober P, et al. Promoter DNA methylation and expression levels of HOXA4, HOXA5 and MEIS1 in acute myeloid leukemia. Mol Med Rep. 2015;11(5):3948‐3954. [DOI] [PubMed] [Google Scholar]
  • 40. Fournier M, Lebert‐Ghali CE, Krosl G, Bijl JJ. HOXA4 induces expansion of hematopoietic stem cells in vitro and confers enhancement of pro‐B‐cells in vivo. Stem Cells Dev. 2012;21(1):133‐142. [DOI] [PubMed] [Google Scholar]
  • 41. Strathdee G, Sim A, Parker A, Oscier D, Brown R. Promoter hypermethylation silences expression of the HoxA4 gene and correlates with IgVh mutational status in CLL. Leukemia. 2006;20(7):1326‐1329. [DOI] [PubMed] [Google Scholar]
  • 42. Zangenberg M, Grubach L, Aggerholm A, et al. The combined expression of HOXA4 and MEIS1 is an independent prognostic factor in patients with AML. Eur J Haematol. 2009;83(5):439‐448. [DOI] [PubMed] [Google Scholar]
  • 43. Raman V, Martensen SA, Reisman D, et al. Compromised HOXA5 function can limit p53 expression in human breast tumours. Nature. 2000;405(6789):974‐978. [DOI] [PubMed] [Google Scholar]
  • 44. Chen H, Rubin E, Zhang H, et al. Identification of transcriptional targets of HOXA5. J Biol Chem. 2005;280(19):19373‐19380. [DOI] [PubMed] [Google Scholar]
  • 45. Shiraishi M, Sekiguchi A, Oates AJ, Terry MJ, Miyamoto Y. HOX gene clusters are hotspots of de novo methylation in CpG islands of human lung adenocarcinomas. Oncogene. 2002;21(22):3659‐3662. [DOI] [PubMed] [Google Scholar]
  • 46. Hajkova H, Markova J, Haskovec C, et al. DNMT3A mutations are connected with lower DNA methylation levels and numbers of concurrently Hypermethylated genes in AML patients. Blood. 2011;118(21):4629. [Google Scholar]
  • 47. Kim T, Tyndel MS, Kim HJ, et al. Spectrum of somatic mutation dynamics in chronic myeloid leukemia following tyrosine kinase inhibitor therapy. Blood. 2017;129(1):38‐47. [DOI] [PubMed] [Google Scholar]
  • 48. Mori N, Ohwashi‐Miyazaki M, Yoshinaga K, et al. Tumor suppressor gene methylation on the short arm of chromosome 1 in chronic myelogenous leukemia. Eur J Haematol. 2017;98(5):467‐477. [DOI] [PubMed] [Google Scholar]
  • 49. Pronier E, Almire C, Mokrani H, et al. Inhibition of TET2‐mediated conversion of 5‐methylcytosine to 5‐hydroxymethylcytosine disturbs erythroid and granulomonocytic differentiation of human hematopoietic progenitors. Blood. 2011;118(9):2551‐2555. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Cancer Reports are provided here courtesy of Wiley

RESOURCES