Skip to main content
Premium Trial:

Request an Annual Quote

MicroRNA Fine Tune Gene Expression in Breast Cancer, Study Says

NEW YORK (GenomeWeb News) – At the whole-genome level, microRNAs mostly act as fine-tuners of gene expression in breast cancer, a study appearing online yesterday in Nature reported.

MiRNAs, as regulators of gene expression, have been thought to play a role in cancer formation and development, and have been identified as possible markers for diagnosis or prognosis. In this study, researchers led by Cancer Research UK's Carlos Caldas examined the miRNA expression profiles of more than 1,300 breast tumors as well as matched clinical, genomic, and messenger RNA expression data. Drawing on this, the researchers were able to get a picture of how different genomic, transcriptional, and post-transcriptional events affect miRNA expression architecture. Caldas and his team found that miRNAs play an important role in modulating immune response in a certain type of breast tumor.

"We demonstrate such an important modulatory role for miRNAs in the biology of [copy-number aberration]-devoid breast cancers, a common subtype in which the immune response is prominent," the researchers wrote.

To get a sense of miRNA expression levels in breast tumors, the researchers quantified known and putative miRNAs using Agilent microarrays in 1,302 breast tumors — including ER+, ER-, and other tumor subtypes — 116 adjacent normal breast tissue, and 28 breast cancer cell lines. While the researchers saw an overall decrease of miRNA expression in tumors, as compared to normal tissues, they were able to detect 853 miRNAs.

In about 10 percent of samples, they identified 226 miRNAs in regions affected by somatic copy number aberrations, though few miRNAs had CNA-driven expression. Another 133 miRNAs were spotted in minimal common regions.

Additionally, by comparing miRNAs to their closest protein coding genes, researchers found that epigenetic effects contribute to only 1 percent of the expression variation in miRNAs; the density of CpG islands near miRNA genes is lower than it is near mRNA promoters. Further, there is likely limited miRNA-host mRNA co-transcription, though some miRNA-host interactions are breast cancer subtype specific. For example, miR-10a and miR-505 are co-transcribed with HOXB3 and HOXB4 or ATP11C, but only in ER+ or ER- samples, respectively, the researchers noted.

Caldas and his colleagues also examined the relationship between this miRNA landscape and the clinical, histological, and molecular data they gathered using a generalized additive model. That model looked at 302 variable miRNAs or 5,367 variable mRNAs. Clinical variables included age at diagnosis, tumor size, and lymph node status, among other factors.

While none of the clinical variables stuck out as an important covariate, the researchers noted that a subset of mRNA varied between ER and PAM50 breast cancer subtypes.

By clustering the miRNA-mRNA GAM values, the researchers uncovered 10 mRNA branches, including ones that appeared to be involved in processes like cell adhesion, cell cycle, or sex hormone activity.

They zoomed in particularly on miRNA-mRNA relationships in ER+ and ER- tumors, and found that ER- tumors were enriched for ECM/cell adhesion and immune response Gene Onotology terms while ER+ tumors were enriched for components of the Wnt pathway, which is involved in the epithelial-mesenchymal transition.

For instance, they noted that SFRP1 is associated with the tumor suppressor miR-99a in ER+, though not in ER-, samples. MiR-342, by contrast, was linked to immune-origin mRNAs in ER- samples.

"The miRNA links of mRNAs pertaining to mammary gland epithelial proliferation were almost mutually exclusive between ER+ and ER- cancers," the researchers wrote.

While a number of miRNAs have been thought to be prognostic, Caldas and his colleagues found that most prognostic miRNAs that were identified through a univariate model lost their significance in a multivariate model that incorporated clinical parameters. Additionally, most of those that retained significance could not be validated in a separate dataset.

Using a principal component analysis approach, though, the research team stratified iClust4 patients into prognostic groups using a 41-miRNA signature. Signatures derived for other subtypes, the group noted, were not consistently prognostic, suggesting to the team that "the iClust4 miRNA landscape is consistently linked with survival."

"Motivated by this observation, we proposed that miRNAs have a pathogenic role in these tumors," the team wrote, adding that it suspected "that rather than acting as on-off switches of particular mRNAs, most miRNAs exert their effect by modulating the relationship between effector and target mRNAs."

To test that idea, Caldas and his team looked at about 1,000 possible mRNA effectors and about 5,000 mRNA targets, and they tested the effector-targets in populations with high or low levels of potential miRNA modulators. ER+ tumors had the lowest number of effector-target pairs, with about 17, and, therefore, the lowest miRNA modulator activity; iClust4 tumors had the highest numbers of pairs per modulator.

"Intermediate counts of modulators, effectors and targets were measured for ER2 tumors and iClust4mimic, suggesting that at least some of the heterogeneity among these tumors may be related to altered miRNA activity," the researchers added.

Many of the potential target mRNAs in the iClust4 samples were linked to lymphopcytic infiltration and immune response, the researchers noted. In a modulation network focused on immune response that they developed, the researchers found members of the miR-27 and miR-29 families, which have been linked to Toll-like-receptor-mediated induction of immune response. Mi-301a, they added, also formed a hub in the network, and its targets are enriched for T-cell-related processes.

"This panoramic view shows the details of individual miRNAs that are expressed as classical oncogenes or tumor-suppressor genes, but mostly reveals that at the whole-genome level miRNAs behave more consistently as fine-tuners/modulators of gene expression," the researchers wrote. "This is especially evident in the common subtype of breast cancer, in which copy number events have no major role and miRNAs seem to modulate the immune response that characterizes these tumors."