Progressive Chromatin Repression and Promoter Methylation of a-catenin Correlates with AML Transformation in Patients with and without 5q Deletions.

Background: Loss of the long arm of chromosome (Ch) 5 or complete loss of Ch 5 is frequent in de novo MDS and AML. Epigenetic modifications of tumor suppressor genes, including aberrant DNA methylation, may play an important role in the progression of MDS and AML. Cell signaling is regulated by a-catenin, which forms a trimolecular complex with E-cadherin (ECAD) and b-catenin that links with actin-containing filaments of the cytoskeleton. Studies have reported reduced expression of a-catenin in MDS and AML patients with 5q deletion as compared to those without 5q deletion. Whether promoter methylation of the a-catenin gene is responsible for this decreased expression is controversial. To explore the potential role of a-catenin in the pathogenesis of AML transformation, we (a) determined the tumor specificity of methylation in AML and non- myeloid malignancies and (b) frequency of methylation in patients with −5/del(5q) MDS/AML and those with normal cytogenetics; (c) performed bioinformatics and experimental analysis of the a-catenin adhesion complex and local 5q31.1 genes to investigate mechanisms of epigenetic silencing; and (d) performed a detailed analysis of primary AML samples correlating promoter methylation, a-catenin expression, and chromatin conformation.

Methods and Results: Using methylation sensitive PCR, we found that methylation of the a-catenin promoter gene was specific for myeloid malignancy. No methylation was observed in 19 acute lymphocytic leukemia cases, 20 chronic myelogenous leukemia cases, or in 99 primary gastric and esophageal samples where a-catenin has been implicated as a tumor suppressor gene. In those patients with AML and an associated 5q deletion the frequency was 31% (8/26) as compared to those without a 5q deletion with 13% (16/120). Bioinformatics analyses of the Valk et al., 2004 leukemia database provide supportive data for under expression of a-catenin in non-5/del (5q) AML cases. We quantitated a-catenin mRNA expression by Q-PCR in our cohort. Expression was lowest in AML patients with a-catenin methylation (n=9), but also in a subset of patients without promoter methylation (n=17), suggesting alternative mechanisms of inactivation. In contrast to AML, methylation of a-catenin was rare in myelodysplastic syndrome (MDS). Although p15 was methylated in over 50% of these cases as a positive control, only 2/18 MDS cases with 5q deletion (11%) and 1/13 MDS cases with 5q intact (8%) were methylated at the a-catenin promoter. The three positive cases were RAEB-2 (1) or RAEB-t (2), suggesting that a-catenin methylation may be most important in promoting transformation from MDS to AML. To explain a potential lack of correlation of methylation with decreased a-catenin in MDS and AML, we investigated a-catenin chromatin in a myeloid stem cell progression model and in primary AML samples. We performed chromatin immunoprecipitation on the CTNNA1 promoter using two active chromatin histone marks, H3K9Ac and H3K4me2 and two inactive marks, H3K9me2 and H3K27me3. In cell lines and primary leukemia samples where a-catenin was highly expressed (N=4), activation marks were present and repression marks absent. In cell lines and primary samples with low a-catenin expression and methylation of the promoter (N=4), the opposite pattern was observed. So called “bivalent” chromatin with mixed marks, no a-catenin methylation, and intermediated mRNA expression levels were observed with 7 additional cases (2 cell lines, 5 primary AML).

Conclusions: Our data indicate that methylation of a-catenin is common in AML patients with 5q deletion but also observed in cases with normal chromosome 5 copy number. We propose a model of progressive inactivation of the a-catenin locus with AML transformation, with methylation representing a late stage event. The tissue-specificity of our results and in vivo chromatin observations in primary AML samples have implications for the timing and combinatorial therapy of MDS/AML with HDAC inhibitors and methylation inhibitors. Additionally, it appears that inactivation of adhesion molecules (ECAD, a-catenin) are frequent events overall in AML, suggesting a new pathway of transformation from MDS to AML.