Epigenetic regulation of human cancer/testis antigen gene, HAGE, in chronic myeloid leukemia

Abstract

Background and Objectives Cancer testis antigens (CTA) provide attractive targets for cancer-specific immunotherapy. Although CTA genes are expressed in some normal tissues, such as the testis, this immunologically protected site lacks MHC I expression and as such, does not present self antigens to T cells. To date, CTA genes have been shown to be expressed in a range of solid tumors via demethylation of their promoter CpG islands, but rarely in chronic myeloid leukemia (CML) or other hematologic malignancies.Design and Methods In this study, the methylation status of the HAGE CTA gene promoter was analyzed by quantitative methylation-specific polymerase chain reaction (MSP) and sequencing in four Philadelphia-positive cell lines (TCC-S, K562, KU812 and KYO-1) and in CML samples taken from patients in chronic phase (CP n=215) or blast crisis (BC n=47). HAGE expression was assessed by quantitative reverse transcriptase-polymerase chain reaction.Results The TCC-S cell line showed demethylation of HAGE that was associated with over-expression of this gene. HAGE hypomethylation was significantly more frequent in BC (46%) than in CP (22%) (p=0.01) and was correlated with high expression levels of HAGE transcripts (p<0.0001). Of note, in CP-CML, extensive HAGE hypomethylation was associated with poorer prognosis in terms of cytogenetic response to interferon (p=0.01) or imatinib (p=0.01), molecular response to imatinib (p=0.003) and progression-free survival (p=0.05).Interpretations and Conclusion The methylation status of the HAGE promoter directly correlates with its expression in both CML cell lines and patients and is associated with advanced disease and poor outcome.

Chronic myeloid leukemia (CML) is a clonal disease of the hematopoietic stem cell, in which a reciprocal translocation, t(9;22)(q34;q11), forms the Philadelphia chromosome (Ph′) and creates a new fusion gene, BCR-ABL. This chimeric gene is translated into a 210 kDa protein (p210), which has abnormal tyrosine-kinase activity that is central to the pathogenesis of the disease.1 Treatment of CML has been notably improved by imatinib mesylate, a potent tyrosine-kinase inhibitor that blocks the kinase activity of p210, thus inhibiting proliferation of Ph′-positive progenitors.2 In the past 5 years, the results of imatinib treatment of CML have established this drug as the first-line therapy for CML patients.3 However, both the persistence of molecular disease in most imatinib-treated patients, as well as the observation that discontinuation of the drug usually results in a rapid loss of the response, indicate that it is unlikely that imatinib alone can cure CML.2,3

An alternative attempt to target CML cells is an active, specific immunotherapy (e.g. a vaccine). In fact, because of the unique amino acid sequence of p210 at the fusion point, the protein is a tumor-specific antigen to which an immune response can be induced. In some studies it has been shown that peptides derived from the p210 fusion point (b3a2) can bind to several HLA class I and class II molecules and thus generate peptide-specific dendritic cells (DC) and cytotoxic T lymphocyte (CTL) responses in vitro.4,5 Preliminary clinical data suggest that the addition of a b3a2-specific vaccine to b3a2- CML patients treated with conventional treatment might favor a further reduction of residual disease, therefore increasing the number of patients who reach a molecular response.6 In disagreement with these findings, it has been recently found that the CTL induced by DC transfected with RNA extracted from BCR-ABL-positive K-562 cells or CML blasts do not recognize epitopes derived from the chimeric BCR-ABL fusion protein. In contrast, they were able to lyse autologous DC electroporated with RNA from patients with acute myeloid leukemia (AML), indicating that some antigens shared by these malignant cells are involved and recognized by these CTL.7 Therefore, BCR-ABL is processed and presented by Ph′-positive cells but is not the immunodominant antigen that induces CD8 T lymphocytes in competition with other tumor-associated antigens. However, the nature of these possible antigens is unknown.

Among the tumor antigens that are presented to human CTL by HLA class I molecules, several constitute safe targets for immunotherapy because they are absent in normal tissues. An important category of such tumor-specific antigens are those encoded by the cancer testis antigen (CTA) genes, such as the MAGE, BAGE, CAGE, GAGE, HAGE, LAGE, PAGE, NY-ESO-1, SSX, and SCP gene families.8 These genes are expressed in several types of tumors but not in adult tissues with the exception of spermatogonia, which do not carry HLA molecules and therefore cannot present antigens to the T cells. Despite the fact that these antigens are expressed in a broad range of human tumors, few CTA genes are expressed in hematologic malignancies, although SCP1 and, specially, HAGE have been found in a proportion of samples from both AML and CML patients.9,10 Epigenetic events appear to represent the unique mechanism regulating the expression of CTA in cancer cells, and DNA methylation seems to play the major role; in fact, a correlation between hypomethylated CpG dinucleotides in CTA promoters and their expression has been found in neoplastic cell lines and tissues.11–14 Moreover, unmethylated MAGE promoters drive the transcription of reporter genes in CTA-negative neoplastic cells, suggesting that availability of transcriptional factors is not involved in the heterogeneity of CTA expression and that DNA methylation is the main restricting factor for CTA expression.15 Consistently, in vitro methylation of these reporter constructs is sufficient to block their transcriptional activity in MAGE-positive cells.16 In addition, all CTA genes that are expressed in tumors or testis can be induced in vitro by DNA demethylation or by inhibitors of histone deacetylation.17

We have recently shown that sense and antisense transcription of the LINE1 retrotransposon (one of the most frequent repetitive sequences in the genome) is activated by promoter hypomethylation in CML and that this event is frequently associated with evolution of the disease to the advanced phase.18 Since global hypomethylation is consistently associated with hypomethylation of the LINE1 elements,19 it can be tentatively speculated that this genome-wide hypomethylation could lead to the reactivation of genes silenced by DNA methylation in CML patients, as for example, CTA genes. In the present study, we provide the first evidence showing the role of promoter methylation in the primary regulation of HAGE and its role in the progression and clinical behavior of CML.

Design and Methods

Cell lines and samples

Four human Ph′-positive CML cell lines (K562, KU812, KYO-1 and TCC-S) were obtained from the American Type Culture Collection (Manassas, VA, USA). The cells were cultured in appropriate medium until harvested for extraction of DNA and RNA. Heparinized bone marrow cells were collected from patients with chronic phase CML (CP-CML), and from healthy marrow donors. Immediately after harvest, total white blood cells were obtained by dextran sedimentation or by red cell lysis of centrifuged buffy coat preparations. Mononuclear cells were isolated from patients with CML in blast crisis (BC-CML) and donors by sedimentation on Ficoll-Hypaque gradients. More than 90% of the mononuclear cell populations from patients with acute phase CML were leukemic blasts. We studied 215 patients with Ph-positive CP-CML, diagnosed between August 1982 and December 2005. The patients were unselected with regard to the type of front-line therapy (39 patients received oral chemotherapy alone, 76 interferon plus cytarabine and 100 imatinib). A diagnostic sample in CP was available for all patients. Paired samples, when both a CP diagnostic sample and a BC sample were available, were analyzed in 47 patients (34 in myeloid BC and 13 in lymphoid BC). BC was defined by the presence of at least 30% blasts in the blood or bone marrow or extramedullary involvement. Risk categories according to the Sokal and Hasford score systems were determined as previously described.20,21 Hematologic, cytogenetic and molecular responses to interferon and imatinib were evaluated. A complete hematologic response (CHR) was defined by a white cell count of less than 101×10/L, with no immature cells and less than 5% basophils in the peripheral blood, a platelet count of less than 450×10/L, and the absence of palpable splenomegaly or extramedullary involvement. Cytogenetic response (CgR) was assessed by G-banding in at least 20 cells in metaphase per sample and was defined as major CgR, including complete responders (CCgR, 0% Ph′-positive metaphases) and partial responders (PCgR, 1%–34% Ph′-positive metaphases) or poor CgR, including minor (MinCgR, 35%–94% Ph′-positive metaphases) or no response (NCgR, 95%–100% Ph′-positive metaphases).

Disease progression was defined by any of the following events, whichever came first: (i) death from any disease-related cause during treatment, (ii) the development of accelerated-phase CML (defined by the presence of at least 15%blasts in the blood or bone marrow, at least 30% blasts plus promyelocytes in the blood or bone marrow, ≥20% peripheral basophils, or thrombocytopenia <100×10/L unrelated to treatment), (iii) blastphase CML (defined by the presence of at least 30% blasts in the blood or bone marrow or extramedullary blastic involvement), iv) loss of complete hematologic response (defined by the appearance of any of the following findings in two blood samples obtained at least 1 month apart: a white-cell count >20×10/L, a platelet count ≥600×10/L, the appearance of extramedullary disease, the appearance of at least 5% myelocytes and metamyelocytes in the peripheral blood, or the appearance of blasts or promyelocytes in the peripheral blood), v) loss of major cytogenetic response (defined as ≥30% increase in the Ph′-positive cells in two studies performed at least 1 month apart), or vi) an increasing white-cell count (defined as a doubling of the count to more than 2×10/L on two occasions at least 1 month apart in a patient who had never strictly had a complete hematologic response despite the maximum tolerated doses of therapy).

Minimal residual disease was determined during follow-up using a standardized real-time quantitative reverse transcriptase–polymerase chain reaction (PCR) analysis on peripheral blood and/or marrow aspirate, as established in the framework of the EU concerted action.22 Complete molecular remission (CMR) was defined as negative quantitative PCR confirmed by nested PCR for BCR-ABL. Major molecular response (MMR) was defined as a reduction in the BCR-ABL/ABL ratio of at least three logs with regard to the diagnostic baseline value. Some of the patients in the present study (n=140) had been previously typed for hypomethylation of LINE1 retrotransposons.18

Semiquantitative real time methylation-specific PCR (qrt-MSP) of HAGE promoter and exon 1

Analysis of the HAGE promoter and exon 1 (GeneBank: NM_018665) has revealed that HAGE possesses a 383bp CpG island located between nt48 and nt430, showing >60% C+G content and an observed-over-expected CpG frequency of >0.6. Promoter methylation of the HAGE gene was determined by the MSP method, as previously reported.23 Briefly, 1 μg of genomic DNA was denatured by treatment with NaOH and modified by sodium bisulfite. Qrt-MSP was performed as previously reported by our group18 in a rapid fluorescent thermal cycler with three-color fluorescence monitoring capability (LightCycler, Roche), using 1 μL of bisulfite-modified DNA in a 10 μL reaction volume with 0.4 μmol/L of each primer, and 1 μL of 1× LightCycler FastStar DNA Master SYBR Green I (Roche Molecular Biochemicals). The final Mg concentration in the reaction mixture was adjusted to 3.5 mmol/L. Amplification of the HAGE methylated sequences was used as target sequence (forward, 5′-GGAGGAGTTTTTAAGGTTTTTACGT-3′; reverse, 5′-GACAATTCCTCGTAACCAACG-3′). The following program conditions were applied for running the qrt-MSP: denaturation program, one cycle at 95°C for 10minutes; amplification program, 45 cycles at 95°C for 10 s, 65°C for 10 s and 72°C for 10 s; melting program, one cycle at 95°C for 0 s, 40°C for 60 s and 90°C for 0 s; and cooling program, one cycle at 40°C for 60 s. The temperature transition rate was 20°C/s, except in the melting program, in which it was 0.4°C/s between 40°C and 90°C. Amplification of the HAGE unmethylated sequences for each sample was performed as the reference sequence (forward, 5′-GGAGGAGTTTTTAAGGTTTTTATGT- 3′; reverse, 5′-ACAACAATTCCTCATAACCAACAA- 3′). It was amplified in the same run and following the same procedure described above for methylated sequences. A procedure based on the relative quantification of target sequence (methylated sequences) vs their controls/calibrators in relation to the reference sequence (unmethylated sequences) was used to assesses the degree of HAGE promoter methylation. Calculations were automatically performed by the LightCycler software (RealQuant, version 1.0, Roche). The normalized ratio was obtained from the following equation and expressed as a percentage of the control/calibrator:

The efficiency (E) of each gene was calculated from the slope of the crossover points (Cp) versus DNA concentration plot, according to the formula E=10. ΔCp corresponded to the difference between control/calibrator Cp and sample Cp, for either the target or the reference sequences. The selected control/calibrator was the bone marrow specimen from a healthy donor, which was considered as 100% (this is not an absolute value indicating a fully methylated patient but a relative value used as a measure for the relative level of HAGE methylation in the particular sample). Water blanks were included with each assay. Results were confirmed by repeating bisulfite treatment and MSP assays for all samples. Occasionally, equal amounts of PCR products were separated on a 2% agarose gel, stained with ethidium bromide, and visualized under UV light.

Expression of HAGE transcripts

Qrt-PCR for HAGE expression was performed with the LightCycler technology, using 1 μL of cDNA in a 20 μL reaction volume with 0.4 μmol/L of each primer (forward: 5′-CCTTTCAATGTTATCCTGAG-3′; reverse: 5′-TATTCTTCAGATTGACGAAG-3′), and 2 μL of 10× LightCycler FastStar DNA Master SYBR Green I (Roche Molecular Biochemicals). The Abelson gene (ABL1) was employed as the reference gene, and it was amplified in the same run and following the same procedure described above (forward: 5′-CCCAACCTTTTCGTTGCACTGT- 3′; reverse: 5′-CGGCTCTCGGAGGAGACGTAGA- 3′). In order to reduce the variation between different assays and samples, a procedure based on the relative quantification of target genes versus their controls/calibrators in relation to the reference gene was used. Calculations were automatically performed by LightCycler software (RealQuant, version 1.0, Roche). The normalized ratio (N_{HAGE-EXPRESSION}), expressed as a percentage of the control/calibrator, was obtained by RealQuant software as described above for qrt-MSP of the HAGE promoter. The selected controls/calibrators were bone marrow specimens from healthy donors, which were considered as having 100% expression.

5-aza-2′-deoxycytidine treatment

CML-derived K562, KU812, KYO-1 and TCC-S cell lines were grown at a density of 750,000 cells/mL in 25 cm flasks with 8 mL of RPMI 1640 medium supplemented with 20% fetal bovine serum and maintained at 37°C in a humid atmosphere containing 5% CO₂. Each of the cell lines was treated with 2 and 4 μM of 5-aza-2′-deoxycytidine (Sigma-Aldrich, Steinheim, Germany) for 4 days.

HAGE promoter and exon 1 analysis by sequencing after sodium bisulfite modification

The methylation status of HAGE promoter and exon 1 was analyzed by bisulfite genomic sequencing of its CpG island in TCC-S and KU812 CML derived cell lines. One microgram of genomic DNA was treated and modified using the CpGenomic DNA Modification Kit (Intergen Company, Purchase, NY, USA). Bone marrow DNA from healthy donors was used as a methylated or positive control for sequencing analysis. After bisulfite modification, the HAGE promoter was amplified by seminested-PCR. The first PCR was performed using 5 μL of modified DNA and HAGE1-SB1 (5′-TTTTTTTTGGAATAATGTTTTATTA- 3′) and HAGE1-SB2 (5′-TAACCCCACCTATCCTACCCTAC- 3′) primers under the following conditions: 94°C for 10 min, 35 cycles at 94°C for 1 min, 52°C for 1 min and 72°C for 1 min, and a final elongation cycle at 72ºC for 10 min. The second PCR was performed with 5 μL of the first PCR product using HAGE1-SB3 (5′-GGAGGAGTTTTTAAGGTTTTTA-3′) and HAGE1-SB2 primers under the following conditions: 94°C for 10 min, 20 cycles at 94°C for 1 min, 52°C for 1 min and 72°C for 1 min, and a final elongation cycle at 72ºC for 10 min. Two PCR reactions were carried out in a total volume of 25 μL, with 1 U high fidelity Platinum Taq DNA polymerase (Invitrogen Life Technologies, Paisley, UK), 1.5 mM MgCl₂, 0.2 mM dNTP and 50 pmol of each primer. The second 280 bp PCR products were separated on a 2% agarose gel, stained with ethidium bromide and visualized under UV light. Amplification products obtained in the second PCR reaction were subcloned into pCR 4- TOPO plasmid using TOPO TA Cloning Kit for sequencing (Invitrogen Life Technologies, Paisley, UK) and transformed into Escherichia coli according to the manufacturer’s recommendations. Colonies with recombinant plasmids containing the described PCR products were screened by digestion with Eco I (Amersham Biosciences, Buckinghamshire, UK). Candidate plasmid clones were sequenced with ABI-PRISM d-Rhodamine Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) in an ABI PRISM 377 DNA Sequencer (Applied Biosystems, Foster City, CA, USA) using T7 and T3 universal forward and reverse primers.

Statistical analysis

All calculations were performed with the SPSS statistical package (SPSS, Chicago, IL, USA). The medians, standard deviations (SD) and interquartile ranges for age and the most relevant clinical and laboratory findings at diagnosis were calculated for patients with and without HAGE hypomethylation and tested for any significant differences with the Mann-Whitney U test (for continuous variables) or χ analysis and Fisher’s exact test (for categorical variables). Overall survival (OS) was calculated from time of diagnosis to death from any cause and was censored only for patients known to be alive at last contact. Progression-free survival (PFS) was measured from the diagnosis of CML to the appearance of BC or death without disease progression, and was censored only for those patients alive and without evidence of progression at last follow-up. For calculations of both OS and PFS, bone marrow transplant recipients were censored at the time of transplantation. Distributions of OS and PFS curves were estimated by the method of Kaplan and Meier, with the 95% confidence intervals being calculated using Greenwood’s formula. Comparisons of OS and PFS between groups were based on the log-rank test. Comparisons adjusted for significant prognostic factors were based on Cox regression models and hazard regression models. All progression and survival data were updated as of March 31, 2006, and all follow-up data were censored at that date.

Results

HAGE is hypermethylated in normal bone marrow cells

We analyzed HAGE methylation in 30 normal bone marrow samples. As expected, HAGE promoter DNA sequences were strongly methylated in normal bone marrow cells. This is illustrated by the amplification of the methylated sequences with the complete lack of non-methylated sequence amplification in half of these samples (Figure 1A). However, 50%of the non-neoplastic marrow specimens displayed a slight amount of HAGE hypomethylation, as defined by the presence of specific products in the melting curves from both methylated (melting temperature 76.1ºC) and unmethylated sequences (melting temperature 74.6ºC, Figure 1B and 1C). Based on the background fluorescence intensity, a cut-off level was determined for specific fluorescence for both unmethylated and methylated sequences in normal individuals. This threshold was used to calculate the cycle threshold or crossing point (C_P) of each sample. The C_P value was directly proportional to the amount of target sequence present in the sample. The normal sample which showed the lowest difference in C_P between the target (methylated sequences) and the reference (unmethylated sequence) sequences was used as the control/calibrator sample for quantification of HAGE promoter methylation in both healthy individuals and CML samples. It was considered as 100%.

Figure 1.Methylation status of the HAGE promoter in healthy individuals and in a patient with CML. (panel A) MSP analysis of the CpG island within the HAGE promoter in healthy individuals and CML patient. BC indicates blast crisis CML; UM, unmethylated sequences; M, methylated sequences. (panel B) qrt-MSP melting curve analysis of a sample from a CML patient showing the presence of two specific PCR products: unmethylated (dotted line, melting temperature 74.6ºC) and methylated HAGE sequences (solid line, melting temperature 76.1ºC) (panel C) HAGE promoter methylation level of a representative CML case (lines with triangles) and a control/healthy individual (lines with circles) as measured by qrt-MSP. The difference between the crossing points of the methylated (bold lines) and unmethylated (fine lines) HAGE sequences was smaller for the CML patient than for healthy individual, indicating a higher level of hypomethylation in the CML patient.

Based on these prerequisites and to determine the cut-off value for altered HAGE promoter methylation in CML samples, the methylation status of the HAGE promoter was quantified by means of qrt-MSP in the healthy donors showing some degree of unmethylated sequence amplification. N_{HAGE-METHYLATION} ratios fell between 100%and 217%(160± 40%). A N_{HAGE-METHYLATION} ratio equal or below 80% (determined as the mean minus 2 SD) was chosen to define hypomethylation of HAGE promoter in CML DNA samples.

HAGE is hypomethylated in CML samples

By qrt-MSP, the CpG island of the HAGE promoter was revealed to be highly hypomethylated in the TCC-S Ph-positive CML cell line (N_{HAGE-METHYLATION} ratio: 10%) whereas K562, KYO-1 and KU812 cell lines showed normal levels of HAGE methylation (median N_{HAGE-METHYLATION} ratio: 98%, range: 100–97%, Figure 2 and Table 1). In order to confirm the results of the qrt-MSP, we screened bone marrow DNA from healthy donors and two CML-derived cell lines (TCC-S and KU812) using bisulfite genomic sequencing. As described in the Design and Methods section, the 280 bp amplification products of the HAGE promoter obtained in a second PCR, which are composed of 20 CpG dinucleotides, were subcloned and sequenced. Genomic sequencing after bisulfite modification revealed that bone marrow DNA from healthy donors and DNA from the KU812 CML-derived cell line were almost completely methylated, showing methylation of the 20 analyzed CpG dinucleotides in all analyzed clones. In contrast, the TCC-S cell line showed a complete lack of methylation of HAGE promoter in all analyzed CpG dinucleotides and clones (Figure 3).

Figure 2.Methylation status of the HAGE promoter in Ph′-positive CML cell lines. MSP analysis of the CpG island within the HAGE promoter in four Ph′-positive CML cell lines. UM, unmethylated sequences; M, methylated sequences. Only the TCC-S cell line showed HAGE hypomethylation.

Figure 3.Analysis of HAGE1 CpG island methylation status by bisulfite sequencing in bone marrow of healthy donors, TCC-S and KU812 CML derived cell lines. (Panel A) Schematic description of the HAGE CpG island. Long black arrow indicates the HAGE translation start site and each vertical bar represents a CpG dinucleotide. The gray arrows show the location of the MSP primers and the black arrows the location of bisulfite sequencing primers (A: HAGE1-SB1 primer; B: HAGE1-SB2 primer; C: HAGE1-SB3 primer). (Panel B) Bisulfite sequencing of the HAGE CpG island. Each box indicates a CpG dinucleotide (white box: unmethylated, black box: methylated) and each line represents the analysis of 20 CpG dinucleotides of a single clone of an HAGE analyzed region. 1: non-modified sequence of HAGE1 CpG island; 2: methylated sequence of HAGE CpG island after bisulfite modification; 3: non-methylated sequence of HAGE CpG island after bisulfite modification. *CpG dinucleotide position.

Table 1.HAGE methylation status and mRNA expression in Ph′-positive CML cell lines before and after treatment with 5-aza-2′-deoxycytidine.

Among CML patients, hypomethylation of HAGE promoter was more frequently observed in BC (22/47, 46%) than during C_P disease (47/215, 22%), with this difference being statistically significant (p=0.01). Furthermore, a more profound level of hypomethylation was observed in BC than in C_P samples (mean N_{HAGE-METHYLATED} ratios: 16%±5% vs 32%±8%, p=0.06). There was a significant concordance between HAGE hypomethylation and presence of LINE1 hypomethylation; HAGE hypomethylation was present in 23 of 93 (25%) LINE1 methylated tumors versus 24 of 47 (52%) LINE1 hypomethylated tumors (p=0.01).

HAGE hypomethylation in CML is associated with over-expression of HAGE transcripts

Quantitative expression of HAGE transcripts was assessed by means of qrt-PCR using cDNA from a healthy donor as the control (considered to have a N_{HAGE-EXPRESSION} ratio of 100%). Normalized ratios for HAGE expression were determined in bone marrow specimens from 30 healthy individuals. N_{HAGE-EXPRESSION} ratios fell between 20%and 100%(mean N_{HAGE-EXPRESSION}: 35±25%). Among CML patients, those with methylated HAGE promoter showed a mean N_{HAGE-EXPRESSION} of 34%±30%, similar to that found in healthy individuals; however, the mean N_{HAGE-EXPRESSION} was significantly higher in those CML patients who showed HAGE promoter hypomethylation (319%±118%, p<0.0001). An N_{HAGE-EXPRESSION} value equal or above 160% (determined as the mean N_{HAGE-EXPRESSION} from normal individuals plus 5 SD) was chosen to define over-expression of HAGE in CML RNA samples. Using this cut-off value, over-expression of HAGE was found in 100% of CML patients with HAGE hypomethylation and in none of the CML patients with methylated HAGE promoter (p<0.0001). High levels of HAGE expression were observed in the HAGE hypomethylated TCC-S cell line (N_{HAGE-EXPRESSION} ratio: 1045%) whereas HAGE hypermethylated Ph-positive CML (K562, KU812 and KYO-1) cell lines showed low levels of HAGE expression (mean N_{HAGE-EXPRESSION} ratio: 10%, range: 0–20%, Table 1).

Exposure of CML cell lines to the demethylating agent 5-aza-2′-deoxycytidine restored expression of HAGE mRNA in cell lines that show hypermethylation of the HAGE promoter, such as K562, KU812 and KYO- 1 (mean N_{HAGE-EXPRESSION} ratio: 160%, range: 142–178%) but had no effect on HAGE mRNA expression in the hypomethylated CML-derived TCC-S cell line (Table 1). These results indicate that hypermethylation is a major mechanism through which HAGE expression is silenced in CML cells.

HAGE hypomethylation, response to treatment and clinical outcome

The clinical and laboratory characteristics at diagnosis of CP-CML patients with hypomethylated and normal HAGE at diagnosis are shown in Table 2. Individual factors such as sex, age, spleen size, percentage of blast cells in the peripheral blood, platelet count, hemoglobin level and white cell count were not significantly associated with the HAGE methylation status. When HAGE methylation status was correlated with pretreatment risk groups, no significant association between HAGE hypomethylation and high-risk patients, as assessed by the Sokal and Hasford scoring systems, was found. In the present study, CML patients were treated with oral chemotherapy (mainly, hydroxyurea, n=39), interferon-based regimens (n=76) or imatinib (n=100). Thirty-seven patients received stem cell transplantation (5 autologous, 32 allogeneic). The treatment modality and number of patients who received transplantation were similarly distributed between the two HAGE methylation groups (Table 2). Moreover, mean time of interferon administration was similar in methylated (31 months) and hypomethylated (30.4 months) patients. Hypomethylation of the HAGE promoter was correlated with a poor response to treatment (Table 2). Thus, the CHR rate was significantly lower among hypomethylated patients than in patients with methylated HAGE promoter (72% versus 89%, p=0.05). Among 76 patients under interferon therapy, a CCgR was observed in 28% of those with methylated HAGE and in none of the hypomethylated patients (p=0.01). Among the 100 CP-CML patients receiving imatinib as front-line therapy, the rates of complete cytogenetic response at 6 and 12 months were 66% and 90% in the HAGE methylated group, as compared with 23% and 55% in the HAGE hypomethylated group (p=0.01 and p=0.02, respectively). In addition, the major molecular response rate at 12 months was better for patients with methylated HAGE (58%) than for HAGE hypomethylated patients (18%, p=0.003). Moreover, the time to achieve a CCgR was significantly longer for hypomethylated patients (mean: 11.1±7.4 months) than for methylated patients (mean: 7.3±3.3 months, p=0.01).

Table 2.Clinical characteristics at diagnosis and outcome of 215 chronic myeloid leukemia patients divided according to HAGE methylation status.

Survival data were available from all patients. Kaplan- Meier analysis revealed differences in the duration of the CP among patients under interferon therapy (Figure 4). The estimated median PFS time for hypomethylated patients was 72.2months (95%CI [confidence interval], 42.6–101.5) compared to 112.6 months (95% CI, 87.8–137.4) for patients with methylated HAGE (p=0.05). Among patients receiving imatinib therapy, only two patients progressed, both of whom had hypomethylation of the HAGE gene.

Figure 4.Kaplan-Meier progression-free survival curve for CML patients. PFS curve according to the HAGE methylation status for 76 CML patients treated with interferon. Solid line, patients with methylated HAGE; dashed line, patients with hypomethylated HAGE.

Discussion

In this study, we provide the first evidence that promoter hypomethylation is the molecular mechanism directly responsible for the high expression levels of the HAGE gene in CML. We have demonstrated that HAGE mRNA is expressed in 22 and 46% of CML patients in chronic phase and blast crisis, respectively. HAGE mRNA was previously reported to be expressed in half of 43 CML patients at presentation.10 However, this study did not verify the methylation status of the HAGE promoter CpG islands. The evidence of HAGE expression, induced by hypomethylation, was provided by the re-expression of HAGE, after treatment with 5-aza-2′-deoxycytidine, in CML cell lines not expressing HAGE. In addition, our results clearly indicate a significant correlation between the expression and hypomethylation of the CpG sites of HAGE. The methylation status analysis of the HAGE promoter in CML cell lines and patients showed a perfect correlation between the expression of HAGE and hypomethylation. Interestingly, the key finding of increased HAGE expression in BC-CML with hypomethylated HAGE promoter seems to be contradicted by the fact that three out of four CML cell lines had low or no detectable HAGE expression. However, we have observed18 that cancer cell lines (including CML-derived cell lines) have an increased rate of hypermethylation of certain CpG islands and less global genomic hypomethylation. Fortunately, some of them (TCC-S) retain the specific epigenetic profile of each tumor cell type and are, therefore, very useful for in vitro experiments while not contradicting the results obtained in clinical samples.

Recently, we showed that LINE1 sense and antisense transcription was activated by promoter hypomethylation in CML and that this event was frequently associated with the evolution of the disease to an advanced phase.18 Because LINE1 hypomethylation is an indicator of global genomic hypomethylation and LINE1 under-methylation is closely correlated with that observed in the HAGE gene, we hypothesize that in CML, demethylation of the HAGE gene occurs during progressive stages of leukemogenesis in CML, probably after global DNA hypomethylation, as seen in many advanced cancers. Moreover, all these findings indicate that hypomethylation is an important feature in CML. Firstly, hypomethylation increased from non-neoplastic marrow cells toward CML cells; secondly, an increase in the hypomethylation status was observed in CML patients with advanced disease; and finally, high levels of hypomethylation would be a marker of poor prognosis in a subset of patients with CP-CML.

What could be the functional significance underlying the hypomethylation of HAGE in CML? The transition of CML from CP to BC is characterized by genomic instability leading to the accumulation of molecular and chromosomal abnormalities in addition to the Ph′-chromosome. 24 Several reports have shown that mammalian DNA methylation plays an important role in maintaining genomic stability unrelated to the effects of DNA methylation on gene expression and that hypomethylation of DNA in tumor cells is associated with genomic instability.25 Moreover, genomic hypomethylation causes tumorigenesis in mice associated with the acquisition of additional genomic changes.26 Our results suggest that hypomethylation affecting not only repetitive DNA sequences but also specific gene promoters (i.e, CTA genes) could be one of the mechanisms employed by CML cells to generate an unstable genome.

It is evident that HAGE hypomethylation occurs in many BC-CML patients but whether this epigenetic change is a cause or a consequence of the progression is unclear. Our data argue against the second alternative because HAGE hypomethylation was not only associated with BC but was also observed in a proportion of CML patients in CP. Interestingly, these patients responded slowly to treatment and progressed to BC much more rapidly than CP patients lacking hypomethylation, suggesting that hypomethylated patients ab initio may be more prone to genomic instability, similarly to patients with deletions of the derivative chromosome 9.27 In support of the role of this epigenetic event in the progression of cancer, genome-wide hypomethylation has been shown to increase progressively in parallel to advanced grade in breast, ovarian, cervical, and neural cancers28–30 and it has also been associated with the progression of lung cancer.31 Therefore, it is not surprising that CTA gene expression has also been reported to be associated with less-differentiated, higher-grade tumors, later stages of cancer, and worse outcome.32–34 In fact, MAGEA3, MAGE-A10, andMAGE-A1 promoters were shown to undergo progressive demethylation in gastric cancer in parallel to disease progression.35

On the other hand, although HAGE expression could be a passive bystander in a more hypomethylated genome, the HAGE gene (located on 6q12–q13) encodes a protein that shows 55% similarity with the human p68 protein, a DEAD-box protein which has been demonstrated to have ATP-dependent RNA helicase activity.36 The protein encoded by the HAGE gene seems to be a new member of the family of DEAD-box proteins,37 which contain the highly conserved Asp-Glu-Ala-Asp (D-E-A-D) motif. These proteins are involved in many aspects of RNA metabolism, spermatogenesis, embryogenesis, and cell growth, functioning as important transcriptional regulators. The amino acids that are highly conserved in all of the DEAD-box proteins are also present in the HAGE protein, suggesting that HAGE may also be an ATP-dependent RNA helicase. DEAD-box proteins are up-regulated in cancer cells but not in the corresponding normal cells/tissues.38 In addition, it has been recently reported that expression of an uncharacteristically large proportion of genes involved in pre-mRNA processing (including RNA helicases) is increased in primary p210BCR/ABL-positive CD34 cells, which correlates with altered splicing of gene products.39 Taken together, these data suggest that HAGE is involved in both altered gene expression and RNA editing. One can, therefore, speculate that CML cells that overexpress HAGE may have a selective advantage over HAGE-negative cells, which would provide an explanation for why HAGE expression is positively selected during CML progression, even though its presence elicits a cytotoxic T-cell-mediated immune response. However, we have not performed functional analyses and therefore, the possible role of HAGE in the transformed status of CML cells should be clarified in further studies.

Besides the possible role of the HAGE gene in the pathogenesis and prognosis of CML, most CTA antigens are immunogenic, and their use as therapeutic cancer vaccines is currently being evaluated.40,41 The DNA methyltransferase inhibitor 5-aza-2′-deoxycytidine (decitabine) has been used as part of chemotherapy protocols, primarily to reverse hypermethylation of tumor suppressor gene promoters, and has been shown to be beneficial in CML.42 This drug, capable of inducing HAGE expression, could increase patient eligibility and treatment effectiveness in CTA-targeted immunotherapy. Our results suggest that the level of HAGE gene expression increases as CML progresses. This in turn suggests a potentially important therapeutic role for CTA-based cancer vaccines in the management of the late stages of CML that should be addressed in future studies.

In conclusion, the results of the present study show that the methylation status of the HAGE promoter directly correlates with its expression status in CML cell lines and patients and that in CML high levels of HAGE transcripts are associated with advanced disease and poor outcome.

Footnotes

• Authors’ Contributions JR-G, AJ-V and XA contributed to the conception of this study. JAC, GN, FC, FP, AH and AT were responsible for the patients studied and clinical data. JR-G, AJ-V and XA performed and processed most experiments, assisted by ESJ-E, LG and LC. JR-G, AJ-V and XA contributed to study design, data analysis and the final version of the manuscript, which was drafted by JR-G and XA, and approved by all authors. All the figures and tables of the paper have been created by JR-G.
• Conflict of Interest The authors reported no potential conflicts of interest.
• Funding: Supported by grants from Fondo de Investigacion Sanitaria (FIS, Spain) 06/0003, 03/0141, 01/0013-01, 01/F018, 02/1299; Navarra Goverment (31/2002); RETIC C03/10, Junta de Andalucia 03/143; 03/144 and funds from IMABIS (Malaga. Spain), “UTE project CIMA”, Fundación de Investigación Médica Mutua Madrileña Automovilista, and Asociacion Medicina e Investigacion (AMI).

• Received September 18, 2006.
• Accepted January 2, 2007.