|Home | About | Journals | Submit | Contact Us | Français|
TET2 enzymatically converts 5-methyl-cytosine to 5-hydroxymethyl-cytosine, possibly leading to loss of DNA methylation. TET2 mutations are common in myeloid leukemia and were proposed to contribute to leukemogenesis through DNA methylation. To expand on this concept, we studied chronic myelomonocytic leukemia (CMML) samples. TET2 missense or nonsense mutations were detected in 53% (16/30) of patients. In contrast, only 1/30 patient had a mutation in IDH1 or IDH2, and none of them had a mutation in DNMT3A in the sites most frequently mutated in leukemia. Using bisulfite pyrosequencing, global methylation measured by the LINE-1 assay and DNA methylation levels of 10 promoter CpG islands frequently abnormal in myeloid leukemia were not different between TET2 mutants and wild-type CMML cases. This was also true for 9 out of 11 gene promoters reported by others as differentially methylated by TET2 mutations. We found that two non-CpG island promoters, AIM2 and SP140, were hypermethylated in patients with mutant TET2. These were the only two gene promoters (out of 14,475 genes) previously found to be hypermethylated in TET2 mutant cases. However, total 5-methyl-cytosine levels in TET2 mutant cases were significantly higher than TET2 wild-type cases (median = 14.0% and 9.8%, respectively) (p = 0.016). Thus, TET2 mutations affect global methylation in CMML but most of the changes are likely to be outside gene promoters.
TET2 [ten-eleven translocation (TET) oncogene family member 2] is a tumor suppressor gene on chromosome 4q24.1 TET2 mutations were first described in myeloproliferative neoplasms (MPN),1 and were later also described in systemic mastocytosis,2 chronic myelomonocytic leukemia (CMML),3 myelodysplastic syndrome (MDS),4 MDS/MPN5 and acute myeloid leukemia (AML).6 The incidence of TET2 gene alterations ranges from 10–25% in these myeloid malignancies, with the highest frequency of mutation found in CMML where TET2 mutations are found in 35–50% of cases.5–7 As reported for TET1,8 TET2 also converts 5-methyl-cytosine to 5-hydroxymethyl-cytosine9 in embryonic stem cells, and thus mutations of TET2 were proposed to contribute to leukemogenesis by altering epigenetic regulation of transcription through DNA methylation. In fact, TET2 has been shown to be the sole gene found to be frequently mutated in myeloid malignancies among the three TET family gene members (TET1, TET2 and TET3),6 disrupting hematopoietic differentiation.10,11 Furthermore, in murine models, TET2 deficiency impairs hematopoietic differentiation with expansion of myeloid precursors.12,13 The exact mechanism and the extent to which TET2 mutations affect DNA methylation remain in question. Ko et al. reported that loss of 5-methyl-cytosine (hypomethylation),10 was a remarkable characteristic in CMML patients with TET2 mutations, and found 2510 differentially hypomethylated regions and only two hypermethylated regions. In contrast, Figueroa et al. studied TET2 mutant AMLs and identified a hypermethylation phenotype, including 129 differentially methylated regions.11 These studies were conducted using microarray-based screening methods for DNA methylation analysis, which might have false positive and negative findings. To examine this issue in more detail, we used bisulfite pyrosequencing, which is one of the most reliable ways to analyze DNA methylation for individual genes, and probed the DNA methylation status of 21 promoters of interest as well as global DNA methylation levels in a cohort of 30 CMML patients. Our data suggest that effects of TET2 mutations on DNA methylation are primarily outside gene promoters.
We analyzed the nature and frequency of somatic mutations affecting the TET2 coding sequence (exons 3–11) in a cohort of 30 patients with CMML according to WHO criteria. TET2 missense or nonsense mutations were detected in 16 out of 30 (53%) patients. Ten patients had a single heterozygous mutation, two had a biallelic or homozygous mutations, three had two mutations, and one patient had three distinct mutations. Altogether, 21 mutations were identified, including 7 missense, 7 nonsense and 7 frame shift mutations. Detailed mutation information is shown in Table 1 and Figure S1. Mutations were observed in exon 3 (8 events), exon 5 (1 event), exon 6 (3 events), exon 7 (1 event), exon 9 (1 event), exon 10 (2 events) and exon 11 (5 events). Only 5 out of 21 identified mutations have been reported previously. Six out of seven identified missense mutations were predicted to affect protein function by using SIFT software. We identified IDH2 R140Q mutation in 1 out of 30 CMML patients (3%). Mutation at the R882 residue in DNMT3A was found in none of the 30 CMML patients. The clinical characteristics (age, gender, blood count means, bone marrow blast cell count and cytogenetic status defined by the IPSS score14) of the 16 patients with TET2 mutations were similar to that of the 14 patients without mutations, except for hemoglobin (Table 2), though the difference did not reach statistical significance after p value correction. We compared the overall and progression free survival in patients with vs. without mutations and observed no significant differences.
Next, we performed bisulfite pyrosequencing to compare DNA methylation status between patients with mutant vs. wild-type TET2 genes (Table 3 and Figs. S2 and S3). Bisulfite pyrosequencing is a highly quantitative and reliable method for methylation analysis of individual CpG sites. We compared TET2 mutant to TET2 wild-type cases to distinguish the effects of TET2 on methylation from the effects of CMML transformation. First, we studied DNA methylation levels of 10 promoter CpG islands frequently abnormal in MDS15 since these genes are assumed to be most likely to show abnormal DNA methylation levels in their promoters when the DNA methylation machinery is altered. Overall, there was no significant difference between TET2 mutant and wild-type cases (Fig. S2). LINE-1 methylation, a marker of global repetitive element methylation level, was also not affected by TET2 mutations (Fig. S2).
Figueroa et al. previously reported that TET2 mutant AMLs had 129 differentially hypermethylated regions compared with normal CD34+ cells obtained from the bone marrows of healthy individuals. On the other hand, Ko et al. reported 2510 hypomethylated genes and only two hypermethylated genes out of 14,475 genes studied in TET2 mutant/low 5-hydroxymethyl-cytosine CMML cases compared with bone marrow/peripheral blood samples from healthy controls. To confirm this, we selected four genes from the study by Figueroa et al. (one each on chromosome 1, 3, 4 and 8) and five hypomethylated genes as well as two hypermethylated genes from the study by Ko et al. (We selected the genes shown in the paper). We studied the same regions as in these reports by bisulfite pyrosequencing. Nine out of eleven genes showed no difference in TET2 mutant cases compared with TET2 wild-type cases (Fig. S3). SP140, was hypermethylated in patients with mutant TET2 (p = 0.0011) (Fig. S3). AIM2 also showed a trend for hypermethylation in TET2 mutant cases, though the difference did not reach statistical significance after p value correction (p = 0.007). Both promoters are not in CpG islands. These results suggest that DNA methylation of SP140 (and possibly AIM2) might be good markers for detecting TET2 mutation in CMML.
To confirm that DNA methylation levels of AIM2 and SP140 are good markers for detecting TET2 mutations, we added 13 CMML (5 mutant and 8 wild-type cases) samples from another cohort. In the validation data set, methylation of both AIM2 and SP140 was higher in mutant cases (median methylation = 65%; range 53% to 88% for AIM2, median methylation = 80%; range 35–88% for SP140) compared with wild-type cases (median methylation = 51%; range 32–81% for AIM2, median methylation = 47%; range 27–78% for SP140). When the two cohorts were combined together, both AIM2 and SP140 were found to show significant differences between TET2 mutant (median methylation = 79%; range 21–88% for AIM2, median methylation = 80%; range 35–93% for SP140) and wild-type cases (median methylation = 40%; range 3–94% for AIM2, median methylation = 48%; range 11–78% for SP140) (p = 0.0017 and 0.0002, respectively) (Fig. 1), which remained significant after p value correction.
We next measured total 5-methyl-cytosine levels in the patient samples where a sufficient quantity of genomic DNA was available for mass spectrometry (12 mutant and 7 wild-type cases). This analysis revealed a higher 5-methyl-cytosine level in TET2 mutant cases (median = 14.0%; range 8.6–17.2%) compared with TET2 wild-type cases (median = 9.8%; range 8.1–11.9%) (p = 0.016) (Fig. 2). Thus, in CMML, TET2 mutations are associated with higher genomic 5-methyl-cytosine levels, but this is not clearly reflected in promoter DNA methylation levels (at least for the genes studied).
In this cohort of 30 CMML patients, we found that missense or nonsense mutations of TET2 were detected in 16 out of 30 (53%) patients. This observed rate is almost identical to previous reports5–7, and slightly higher than in the other myeloid diseases reported so far.3–5 Mutations were found to be distributed broadly from exon 3 to exon 11. Furthermore, only 5 out of 21 mutations were the same as previously reported, confirming the marked heterogeneity in mutational status. Overall, in addition to the frequency of mutations, the characteristics of the mutations in this study are in good agreement with what has been reported so far. Missense mutations and frame shift mutations are mainly found in exon 3 of TET2, whereas point mutations are found in exons 4 to 11 (Fig. S1). Although these analyses revealed that TET2 does not have mutation “hot spot(s)” as seen for IDH1/2 and DNMT3A in MDS, some locations in TET2 were found to have high frequencies of mutations. We identified IDH2 R140Q mutation only in 1 out of 30 CMML patients. This rate is almost identical to a previous report in reference 16. Furthermore, this patient did not have a mutation of TET2, which is in agreement with the finding that TET2 and IDH1/2 gene mutations are mutually exclusive.11,17 We also confirmed that none of 30 CMML patients had mutation at the R882 residue in DNMT3A. In fact, DNMT3A gene mutations have been reported to be infrequent (8%) in MDS.18 However, we cannot rule out the possibility that mutation(s) in other locations of these genes might affect methylation status. We could not find significant differences in overall and progression free survival between TET2 mutant and wild-type cases in this cohort. However, correlation of TET2 mutation and survival is still in question; reported by different studies as superior in MDS19 and CMML,17 inferior in AML6 and CMML,7 and not different in MDS.4 Larger studies will be needed to confirm the effect of TET2 mutations on survival.
Bisulfite pyrosequencing is one of the most reliable ways to analyze DNA methylation for individual genes, and we find that only two genes, AIM2 and SP140, were hypermethylated in patients with mutant TET2 compared with wild-type TET2. These genes are the only two genes found to be hypermethylated in a previous report that studied 14,475 genes.10 Recently, AIM2 was reported to have a putative role in reduction of cell proliferation by cell cycle arrest;20 therefore, methylation of the promoter might provide a growth advantage to cancer cells. SP140 is known to be a lymphoid-restricted gene with a potential role in gene transcription.21 Polymorphisms of SP140 were recently reported to be correlated with chronic lymphoid leukemia.22 Methylation of SP140 might have an effect on differentiation to specific lineages. Overall, we find rare promoter methylation differences in TET2 mutant cases, but hypermethylation of AIM2 and SP140 may be useful biomarkers of TET2 mutations in CMML.
TET2 mutation has been shown to lead to inefficient conversion of 5-methyl-cytosine to 5-hydroxymethyl-cytosine. Consistent with this, we found that 5-methyl-cytosine levels of TET2 mutant cases are higher than TET2 wild-type cases. However, this does not seem to translate to increased promoter methylation, with AIM2 and SP140 being notable exceptions. While we did not study the whole genome to be completely confident of this fact, we did investigate the most frequently hypermethylated genes in MDS, and others studied genome-wide methylation with similar findings (hypermethylation of only two out of 14,475 genes). We could not confirm hypomethylation in TET2 mutant CMML cases. Given the above, our findings suggest that the total methylation level increase in TET2 mutant cases is mostly outside CpG islands and promoters examined so far. Interestingly, LINE-1 also did not show significant differences between TET2 mutant and wild-type cases in our study, suggesting that other sequences need to be investigated.
There are several possible explanations for the findings in this study about the impact of TET2 mutations on promoter methylation. First, different TET2 mutations might affect DNA methylation in divergent ways; however, most of the mutations found in this report are predicted to negatively affect protein function. We found no difference in DNA methylation in patients with homozygous, biallelic or frame-shift mutations. In addition, AIM2 and SP140 were almost universally more methylated in TET2 mutant cases (Fig. 1), suggesting a fairly uniform effect. Second, the effect of TET2 mutations might be different in each disease type. To support this, the only two differentially methylated genes in this study, AIM2 and SP140, were selected from a prior CMML study,10 but not in an AML study.11 It is also possible that the effect on promoter DNA methylation of TET2 is not global but very restricted to a few genes such as AIM2 and SP140. However, the TET1 protein has been found to be enriched at most CpG-rich sequences23,24 and there is no mechanism to explain selectivity. Because the promoters of AIM2 and SP140 are not in CpG islands, the observed effect on DNA methylation could be secondary to other effects of TET2 on gene expression. Indeed, TET1 protein has been found to affect gene expression independent of DNA methylation.23 Altogether our data suggest that TET2 mutations have effects on global DNA methylation, but we have not been able to detect major effects on promoter methylation (with the limitations previously discussed). It appears likely that TET2 mutations affect DNA methylation in other regions such as gene bodies or intergenic areas. Larger and genome wide studies will be needed to confirm the precise relationship between TET2 mutations and DNA methylation.
We analyzed whole bone marrow or peripheral blood (bone marrow was not available in one TET2 wild-type patient) samples prior to treatment from 30 patients with CMML referred to The University of Texas MD Anderson Cancer Center or enrolled in a multi-institution Phase III trial comparing decitabine with supportive care.25 For validation of DNA methylation levels of AIM2 and SP140, we also analyzed bone marrow samples from 13 patients with CMML including 5 mutant and 8 wild-type cases from The University of Chicago. The Institutional Review Board at The University of Texas MD Anderson Cancer Center and The University of Chicago approved each institution's respective protocols, and all patients gave informed consent for the collection of residual tissues as per institutional guidelines and in accordance with the Declaration of Helsinki.
For TET2 gene analysis, polymerase chain reaction (PCR) and direct sequencing of exon 3–11 was performed starting from 20 ng of genomic DNA, as previously described1. PCR amplicons were sequenced by Beckman Coulter Genomics (Danvers, MA). All TET2 mutations were scored on both strands. Sequence traces were analyzed with SeqMan Pro (DNASTAR, Inc., WI) and reviewed visually. TET2 anomalies were numbered according to European Molecular Biology Laboratory nucleotide sequence reference FM992369. Previously annotated single nucleotide polymorphisms in the HapMap database (www.hapmap.org) were discarded. SIFT software26 was used to determine the probability that a particular amino acid substitution is tolerated. For a cohort of 30 CMML cases, we used pyrosequencing to analyze mutations of the R132 residue in IDH1, and residues R140 and R173 in IDH2 which have been reported in MDS27 and glioblastoma.28 Mutations encoding amino acid R882 residue in DNMT3A gene18,29 were analyzed by pyrosequencing. Primer sequences are listed in Table S1.
We used bisulfite pyrosequencing to quantitatively assess DNA methylation30 for 10 promoter CpG islands frequently abnormal in MDS [ER (ESR1), NOR1 (OSCP1), p15 (CDKN2B), NPM2, ECAD (CDH1), CDH13, OLIG2, PGRB, PGRA and RIL (PDLIM4)],15 and 11 promoter regions of genes reported by others to be differentially methylated by TET2 mutations (hypomethylated: C9orf16, PSMD6, LRRC32, TMEM34, FSD1NL; hypermethylated: AIM2 and SP140 from Ko et al. and ACOX3, SLC39A14, ZNF662 and DNM3 from Figueroa et al.). For the genes reported to be differentially methylated in the previous reports, we analyzed the same sites of these genes analyzed in the prior reports (Table S2). Long interspersed nuclear element-1 (LINE-1) was also analyzed to measure global repeat element methylation. We tested samples for which a sufficient amount of DNA was available for bisulfite treatment. The number of patients with successful results (mostly >90% success rate) varied slightly for each gene. Primer sequences are listed in Table S1.
DNA hydrolysis was performed as previously described by Song et al. with minor modifications.11,31 Briefly, one microgram of genomic DNA was first denatured by heating at 100°C. Five units of Nuclease P1 (Sigma-Aldrich, Cat # N8630, MO), were added and the mixture incubated at 45°C for 1 h. A 1/10 volume of 1 M ammonium bicarbonate and 0.002 units of venom phosphodiesterase 1 (Sigma-Aldrich, Cat # P3243, MO) were added to the mixture and the incubation continued for 2 h at 37°C. Next, 0.5 units of alkaline phosphatase (Invitrogen, Cat # 18009-027, CA) were added, and the mixture incubated for 1 h at 37°C. Before injection into the Zorbax XDB-C18 2.1 mm x 50 mm column (1.8 µm particle size) (Agilent Cat # 927700-902, CA), the reactions were diluted 10-fold to dilute out the salts and the enzymes. Samples were run on an Agilent 1200 Series liquid chromatography machine in tandem with the Agilent 6410 Triple Quad Mass Spectrometer. LC separation was performed at a flow rate of 220 µL/min. Quantification was done using a LC-ESI-MS/MS system in the multiple reaction monitoring (MRM) mode. We measured 5-methyl-cytosine levels in genomic DNA of TET2 mutant and wild-type cases where sufficient amount of samples are available in quantity for mass spectrometry (12 mutant and 7 wild-type cases).
Statistical analyses were performed using PRISM (GraphPad Software, Inc., CA). Differences in clinical characteristics of patients with or without TET2 mutations were assessed using the Fisher's exact test, the Mann-Whitney or log-rank analysis. We used Kaplan-Meier tests to calculate and generate survival curves and used the log-rank test to determine significance between the group of TET2 mutant and wild-type. Progression-free survival was defined as time to development of acute myelogenous leukemia. We used the Mann-Whitney test to compare continuous variables of DNA methylation and 5-methyl-cytosine levels between TET2 mutant and wild-type cases. All p values were two-tailed and the threshold of statistical significance was p less than 0.05 followed by Bonferroni's correction when multiple analyses were performed.
The authors thank the patients who have contributed to our understanding of these disorders. This work was supported by National Institutes of Health grants CA100632, CA121104 and CA046939 (to J.P.I.) and CA129831 and CA129831-03S1 (to L.A.G.) and supported by a Stand Up to Cancer grant from the American Association for Cancer Research. J.P.I. is an American Cancer Society Clinical Research professor supported by a generous gift from the F.M. Kirby Foundation.
No potential conflicts of interest were disclosed.
This work was supported by National Institutes of Health grants CA100632, CA121104 and CA046939 (to J.P.I.) and CA129831 and CA129831-03S1 (to L.A.G.) and supported by a Stand Up to Cancer grant from the American Association for Cancer Research.