An important question in the field of cancer epigenetics involves the causes of CpG island hypermethylation in tumor suppressor genes leading to transcriptional silencing. L. Di Croce et al. recently shed some light on this subject (“Methyltransferase recruitment and DNA hypermethylation of target promoters by an oncogenic transcription factor,” Reports, 8 Feb., p. [1079][1]). They report that the leukemia-promoting PML-RAR fusion protein may be an important cause of this epigenetic aberration. However, several lines of evidence suggest that this experimental model does not explain the accumulating data concerning promoter region methylation.
In this experimental model, expression of the PML-RAR transcript with an exogenous construct leads to recruitment of DNMT1 and DNMT3a and increased methylation of a RARβ-Luc fusion construct or the endogenous promoter. Consistent with this, seven of nine primary acute promyelocytic leukemias (APLs) had methylation of the 5′ region of the endogenous RARβ βgene. Our own analysis suggests that the frequency of methylation of RARβ2 is the same in APLs that have the PML-RAR translocation as it is in other subtypes of acute myelogenous leukemia (AML) that do not have this alteration (see figure). This suggests that the presence of the PML-RAR translocation is neither necessary nor sufficient to induce RARβ2 methylation. In fact, many other malignancies have RARβ2 methylation without this translocation (see figure), in some cases more commonly than APL. We have also observed that APL patients with the PML-RAR translocations have the same frequency of CpG island hypermethylation of p15INK4b, CDH1, and p73 (all of them with potential RAR elements in their promoters) and global genomic methylation as other AML subtypes without the translocation (see figure). These observations prompt us to caution the extension of the experimental studies described into the more complex genetic and epigenetic alterations observed in primary human malignancies.
# Response {#article-title-2}
We recently demonstrated that, in APLS, PML-RAR promotes specific methylation of its target gene RARβ2 by recruiting DNA methyltransferases (DNMTs) to the promoter region. On the basis of these findings, we have proposed a general mechanism for the specificity of DNA methylation in cancer cells, e.g., aberrant recruitment of DNMTs by oncogenic transcription factors to specific regulative loci. Esteller et al . now report that the frequency of RARβ2 methylation (and other RA-target genes) is similar for APLs, which express PML-RAR, and other subtypes of AMLs, which do not express this fusion protein, and they question the importance of the role of PML-RAR in RARβ2 methylation.
The mechanism(s) responsible for RARβ2 methylation in AMLs is presently unknown. On the basis of our proposed model, RARβ2 methylation in AMLs might be triggered by AML-specific fusion proteins. Indeed, one of the two components of each fusion protein is generally a transcription factor, so that AML-associated fusion proteins function as aberrant transcriptional regulators (similar to PML-RAR in APLs). We have recently shown that transcriptional repression of RA signaling is a common feature of AMLs. In particular, we have shown that AML1-ETO , the most common AML-associated fusion protein, is an HDAC-dependent repressor of RA signaling, thereby suggesting that RA target genes (such as RARβ2 and potentially the p15INK4b, CDH1, and p73 genes mentioned by Esteller et al .) are deregulated by AML-fusion proteins ([1][2]). Thus, the data reported by Esteller et al . do not contradict our model. Rather, they suggest that methylation of RA target genes is a frequent event in leukemias. Further investigation is required to decipher the mechanistic roles of AML-associated fusion proteins in establishing specific DNA methylation in AMLs.
A similar scenario (aberrant recruitment of DNMTs by oncogenic transcription factors) can be envisioned for RARβ2 methylation in other cancers. Indeed, we are currently investigating the interactions between DNMTs and general transcription factors, both in normal and transformed cells. However, we cannot rule out the possibility that, in some tumors, RARβ2 methylation is caused by secondary mechanism(s).
1. [↵][3]1. F. F. Ferrara 2. et al.
, Cancer Res. 61, 2 (2001).
[OpenUrl][4][Abstract/FREE Full Text][5]
[1]: /lookup/doi/10.1126/science.1065173
[2]: #ref-1
[3]: #xref-ref-1-1 "View reference 1 in text"
[4]: {openurl}?query=rft.jtitle%253DCancer%2BResearch%26rft.stitle%253DCancer%2BRes.%26rft.issn%253D0008-5472%26rft.aulast%253DFerrara%26rft.auinit1%253DF.%2BF.%26rft.volume%253D61%26rft.issue%253D1%26rft.spage%253D2%26rft.epage%253D7%26rft.atitle%253DHistone%2BDeacetylase-targeted%2BTreatment%2BRestores%2BRetinoic%2BAcid%2BSignaling%2Band%2BDifferentiation%2Bin%2BAcute%2BMyeloid%2BLeukemia%26rft_id%253Dinfo%253Apmid%252F11196162%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx
[5]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NjoiY2FucmVzIjtzOjU6InJlc2lkIjtzOjY6IjYxLzEvMiI7czo0OiJhdG9tIjtzOjI1OiIvc2NpLzI5Ny81NTg4LzE4MDcuNC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=