The nucleolar size is associated to the methylation status of ribosomal DNA in breast carcinomas

Abstract

Background: There is a body of evidence that shows a link between tumorigenesis and ribosome biogenesis. The precursor of mature 18S, 28S and 5.8S ribosomal RNAs is transcribed from the ribosomal DNA gene (rDNA), which exists as 300-400 copies in the human diploid genome. Approximately one half of these copies are epigenetically silenced, but the exact role of epigenetic regulation on ribosome biogenesis is not completely understood. In this study we analyzed the methylation profiles of the rDNA promoter and of the 5' regions of 18S and 28S in breast cancer.

Methods: We analyzed rDNA methylation in 68 breast cancer tissues of which the normal counterpart was partially available (45/68 samples) using the MassARRAY EpiTYPER assay, a sensitive and quantitative method with single base resolution.

Results: We found that rDNA locus tended to be hypermethylated in tumor compared to matched normal breast tissues and that the DNA methylation level of several CpG units within the rDNA locus was associated to nuclear grade and to nucleolar size of tumor tissues. In addition we identified a subgroup of samples in which large nucleoli were associated with very limited or absent rDNA hypermethylation in tumor respect to matched normal tissue.

Conclusions: In conclusion, we suggest that rDNA is an important target of epigenetic regulation in breast tumors and that rDNA methylation level is associated to nucleolar size.

Figure 1

Location of the target regions selected for DNA methylation analysis within the rDNA locus. The picture reports a schematic representation of the rDNA locus and the location of the 3 target regions (RiboPromoter, 18S and 28S) that are amplified and analyzed by the MassARRAY EpiTYPER assay. Base positions are relative to the transcription starting site (+1) of rRNA primary transcript. For each amplicon the amplified strand is indicated, together with the sequence of the unconverted target region. The CpG sites whose methylation status can be assessed by the MassARRAY EpiTYPER assay are reported in bold. Abbreviations: 5’-ETS, 5’ external transcribed spacer; ITS1, Internal transcribed spacer 1; ITS2, Internal transcribed spacer 2; 3’-ETS, 3’ External transcribed spacer.

Figure 2

DNA methylation of rDNA locus in pair-matched normal and tumor tissues. (A) The correlation matrices of CpG sites analyzed in the 3 target regions (RiboPromoter, 18S and 28S) are reported for normal (left panel) and tumor (right panel) tissues. (B) The methylation levels of rDNA CpG sites are reported for 45 pair-matched normal (left panel) and tumor (right panel) tissues. (C) The boxplot compares, for each CpG site included in the analysis, the DNA methylation levels in 45 normal and tumor tissues.

Figure 3

Relationship between rDNA methylation and tumor parameters. (A) Mean methylation levels of rDNA CpG sites in tumor samples divided for nuclear grade (NG). Standard deviation bars are reported. (B) Silver staining of two breast carcinomas. Note the higher quantity of silver stained nucleolar structures in left panel compared with those in right panel. (C) Mean methylation levels of rDNA CpG sites in tumor samples divided for nucleolar size values. Standard deviation bars are reported.

Figure 4

Relationship between nuclear grade, nucleolar size and rDNA methylation. (A) Classification of the analyzed breast carcinoma samples depending on NG and nucleolar size values. (B) Mean methylation levels of rDNA CpG sites in tumor samples divided in four classes depending on NG and nucleolar size values. Standard deviation bars are reported.

Figure 5

Relationship between ribosome biogenesis and rDNA methylation differences in tumor-normal tissue pairs. (A) For each normal-tumor tissue pair, DNA methylation differences were calculated and subjected to hierarchical clustering using complete linkage method and a euclidean distance measure. (B) The boxplot compares nucleolar size values between normal-tumor tissue pairs, subdivided in two groups on the basis of the results of hierarchical clustering.