Research Highlights:Highlights from the latest articles in epigenomics

Can promoter methylation in peripheral blood cells predict the development of cancer?

Evaluation of: Iwamoto T, Yamamoto N, Taguchi T, Tamaki Y, Noguchi S: BRCA1 promoter methylation in peripheral blood cells is associated with increased risk of breast cancer with BRCA1 promoter methylation. Breast Cancer Res. Treat. (2010) (Epub ahead of print).

The tumor suppressor gene BRCA1 displays germline mutations in a high proportion of hereditary breast cancers and in a smaller percentage in sporadic breast cancers. The resulting tumor phenotypes are similar with a preponderance of estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) negativity. Based on earlier research [1,2], Iwamoto T et al., investigated the hypothesis that BRCA1 promoter methylation in peripheral blood cells (PBCs) may represent a risk factor for breast cancer [3]. The study involved 200 breast cancer cases and 200 controls which were evenly distributed over four age decades (30–60 years of age) to minimize age effects. BRCA1 promoter methylation was examined with real-time methylation-specific PCR. Tumor tissue was obtained from 162 of the breast cancer cases and evaluated for BRCA1, ER, PR and HER2 protein expression by immunohistochemistry and for BRCA1 promoter methylation by methylation-specific PCR. BRCA1 promoter methylation was identified in PBCs from 21.5% of breast cancer patients and 13.5% of controls. By logistic regression analysis, the authors calculated an overall risk association with breast cancer for PBC BRCA1 methylation with an odds ratio of 1.73 (p = 0.045). Stratification by tumor promoter methylation status showed an odds ratio of 0.84 (p = 0.61) and 17.78 (p = 0.001) for BRCA1 promoter methylation-negative and -positive tumors, respectively, in association with PBC methylation. Thus, the authors concluded that the presence of BRCA1 promoter methylation in PBCs is associated with a significant risk for BRCA1 methylation-positive but not for methylation-negative tumors. There does in fact appear to be an association between BRCA1 promoter methylation in PBCs and BRCA1 promoter methylation in primary tumors.

Appropriately, the authors note that their findings are discordant with those of Chen et al. who found no evidence of differential PBC methylation between women with a history of breast cancer with a high risk of BRCA1 mutations by family history and negative controls considered to show no evidence of BRCA1 mutations [3]. As noted by the authors, the Chen et al. study did not select for BRCA1-positive tumors and provides no information about the incidence of such tumors in their study group [4]. Another study by Cho et al. also found no evidence of significant promoter methylation of eight tumor suppressor genes, including BRCA1, in PBCs of patients with invasive ductal carcinoma of no specific subtype [5]. Since Iwamoto et al. found a statistically significant correlation between PBC BRCA1 methylation only for BRCA1 methylation-positive tumors, the methodology of these negative studies do not appear comparable. By contrast, Wong et al., found increased BRCA1 promoter methylation in PBCs of patients with breast carcinomas displaying a range of morphologic features consistent with BRCA1 mutation-associated tumors compared with the normal control group (p = 0.004) [6]. Methylation was associated with a 3.5-fold increased risk of having early onset breast cancer.

In summary, this is an intriguing study in which the authors identified an association between the presence of BRCA1 promoter methylation in PBCs and BRCA1 promoter methylation-positive breast cancers. They propose that this might represent a significant risk factor for the development of breast cancer in general, but with a far higher risk for developing breast cancer with BRCA1 promoter methylation. An extensive study from Wong et al. confirms these findings and both papers provide strong support for the identification of epigenetic mutations in PBCs as biomarkers for BRCA1 methylation-associated breast carcinoma [6]. However, what is lacking from these association studies is a discussion of the significance of BRCA1 methylation in PBCs of the nonaffected control cases, where 13.5% of controls also demonstrated promoter methylation. Finally, it has previously been reported that DNA from tumors is bound to the extracellular surface of leukocytes and shows a similar pattern of methylation as found in the primary breast tumor [7]. This could be one reason for the much greater amount of BRCA1 methylation found associated with PBCs from patients with BRCA1 methylation-positive tumors compared with BRCA1 methylation-negative tumors. While this is an exciting area of investigation, clearly more work is needed to fully understand and appreciate the findings of Iwamoto and others, as an association is different than a causation.

Financial & competing interests disclosure

Charles W Caldwell is supported by the National Library of Medicine (5T15LM007089-19) and the CRC Missouri Chair in Cancer Research endowment. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

Methylation of CD44 as a candidate biomarker in lymphomas

Evaluation of: Eberth S, Schneider B, Rosenwald A et al.: Epigenetic regulation of CD44 in Hodgkin and non-Hodgkin lymphoma. BMC Cancer 10(1), 517 (2010).

Increasing numbers of malignancies have been shown to be associated with epigenetic inactivation of tumor suppressor genes (TSG) through the mechanism of promoter CpG island hypermethylation. Eberth et al. investigated the methylation status of 24 TSGs in 40 lymphoma-derived cell lines and 50 primary patient lymphoma samples using methylation-specific multiplex ligation-dependent probe amplification [1]. They demonstrated that an average of 8 ± 2.8 of TSGs in lymphoma cell lines were methylated versus only 2.4 ± 2 in primary lymphomas in concordance with the general finding of hypermethylation in cultured cell lines. In contrast, no TSG methylation was identified in DNA from tonsils and peripheral blood mononuclear cells of three healthy donors. Interestingly, in regard to incidence of potential pathogenetic mechanisms, TSG deletions averaged only 2.4 ± 1.5 per cell line. The extensive methylation of several TSGs over the entire range of 40 lymphoma cell lines is also worth noting: DAPK1; 37/40, RARβ; 36/40, CDH13; 36/40, IGSF4; 36/40, TIMP3; 35/40 and ESR1; 31/40. These genes were also identified as the most commonly methylated in primary tumor tissue.

The authors focused on one TSG, CD44, which has previously been shown to be hypermethylated and silenced in prostate carcinoma and neuroblastoma [2,3]. In this study, CD44 displayed differential methylation among several subtypes of lymphoma, being frequently hypermethylated and transcriptionally silenced in anaplastic large cell lymphoma, Burkitt‘s lymphoma (BL), diffuse large B-cell and follicular lymphoma cell lines. By contrast, it was unmethylated and expressed in several Hodgkin‘s lymphoma and most mantle cell lymphoma (MCL) cell lines. Concordant methylation results were confirmed in primary MCL (0/11) and BL (18/29). Evaluation of CD44 transcription using quantitative real-time PCR demonstrated a good correlation as well with expression shown in all MCL (7/7) and a majority of Hodgkin‘s lymphoma (5/7) and anaplastic large cell lymphoma (3/7) but rarely seen in BL, follicular lymphoma and diffuse large B-cell lymphoma. Treatment of hypermethylated cell lines with the demethylating agent 5-aza-2‘-deoxycytidine re-established mRNA and protein expression.

This study provides persuasive evidence for the use of CD44 methylation as a potential epigenetic biomarker and therapeutic target for selected lymphomas as well as other types of cancer that require further investigation.

DNA hypomethylation in a spectrum of plasma cell disorders

Evaluation of: Salhia B, Baker A, Ahmann G, Auclair D, Fonseca R, Carpten J: DNA methylation analysis determines the high frequency of genic hypomethylation and low frequency of hypermethylation events in plasma cell tumors. Cancer Res. 70(17), 6934–6944 (2010); Park G, Kang SH, Lee JH et al.: Concurrent p16 methylation pattern as an adverse prognostic factor in multiple myeloma: a methylation-specific polymerase chain reaction study using two different primer sets. Ann. Hematol. 90(1), 73–79 (2011).

Plasmacytic malignancies arise through a series of transformational steps from normal plasma cells (NPC) and pass through the preneoplastic stage, referred to as monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM) and then the full-bore malignancy termed multiple myeloma (MM). In some cases, MM can progress to very aggressive plasma cell leukemia (PCL). While both genetic and epigenetic alterations likely impact the underlying biology of these changes, they are not well understood at this time.

One recent study by Salhia et al. examined DNA methylation in CD138-purified NPC, SMM and MM cells using the Illumina^® GoldenGate Methylation Cancer Panel I^® microarray and found a large number of differentially methylated loci that increased with tumor grade [1]. This array is comprised of 1505 CpG sites of interrogation representing 807 genes. For appropriate validations, candidate genes were further examined by methylation-specific PCR and pyrosequencing. Interestingly, the majority of these differentially methylated loci were considered hypomethylated when compared with NPC, with fewer showing gene-specific hypermethylation. It was suggested that hypomethylation is an early event in development of MM, as even at the MGUS stage (and SMM), this pattern was dominant. This is an important concept, as most cancer epigenetic studies have focused on hypermethylation. Nevertheless, genomic hypomethylation is known to accompany, or even precede, gene-specific hypermethylation in several tumor types. While this is an insightful investigation, the authors appropriately point out a few differences with other studies. The main limitation to this study, which does not diminish the observations, is that the array itself represents only a fraction of the genome and only one or a few probes for each gene, so larger-scale studies are certainly warranted.

In this context, Walker et al. more recently used the Infinium^® BeadArray™ that interrogates 27,578 CpG dinucleotides (mainly in CpG islands) across 14,495 genes to analyze methylation patterns in normal B cells, NPC, MGUS, MM and PCL [2]. They found that methylation patterns distinguished NPC from malignant cells. The main reason for this is hypomethylation of the genome at the transition from MGUS to MM. This observations also fully supports the findings of Salhia et al.[1]. Gene-specific hypermethylation was present at the MM stage, but even more striking, they found remethylation of the genome at the transition from MM to PCL, particularly of genes involved in cell–cell signaling and cell adhesion, which may related to independence from the bone marrow microenvironment. They also found a high degree of methylation variability within MM samples associated with cytogenetic differences. Two groups of hyperdiploid samples were identified, on the basis of unsupervised clustering, which had an impact on overall survival. Thus, DNA hypo- and hyper-methylation changes significantly during plasma cell disease progression and between clinical subgroups.

Certainly, it would be quite useful to be able to predict the clinical behaviors of all malignancies based on biological attributes. As one example, Park et al. studied promoter DNA methylation of the cell cycle control gene, p16 and found potential usefulness of a differential methylation pattern as a discriminator of prognosis in MM [3]. They used a fairly standard method, methylation-specific PCR, but included two different sets of primers to interrogate the promoter region. Their patient cohort was also instructive. It included patients with well defined MM that were treated on one of three chemotherapy regimens. When examined with respect to cytogenetic and clinic–laboratory findings, there was no significant correlation with p16 methylation. However, when both p16 primers delivered positive results (p16P1P2), patients had a poor overall survival regardless of their treatment regimen, age or other laboratory features. If this finding holds up in larger prospective trials, it may be a very useful biomarker for predicting prognosis.

In summary, Salhia et al.[1] and others [2] have contributed to larger scale epigenetic insights into potential biology of plasma cell diseases, while Park et al. have provided a useful example of an epigenetic biomarker of disease prognosis in MM, regardless of treatment regimen [3]. It is increasingly clear that no single method, short of genomic bisulfite sequencing, is likely to provide the complete cellular methylome. Nevertheless, each method can contribute to the eventual compilation of a cell type-specific methylome.

Additional information

Funding

Charles W Caldwell is supported by the National Library of Medicine (5T15LM007089-19) and the CRC Missouri Chair in Cancer Research endowment. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.