Despite recent advances in molecular characterisation of breast cancer, the underlying mechanisms of dissemination of primary tumour cells remain elusive. By utilising an integrated genomics and expression profiling approach with multiple independent breast cancer isolates encompassing a spectrum of metastatic capabilities, we sought to identify genes relevant to metastasis with greater confidence.
We have generated multiple sublines of the clinically relevant 4T1 model of breast cancer metastasis with a spectrum of spontaneous metastatic phenotypes when transplanted orthotopically into syngeneic hosts. Tumour cells were purified from primary tumours using FACS and interrogated for DNA copy number variations with the Agilent 105K mouse CGH array, and expression profiled with the Affymetrix exon array for gene and exon level analysis.
Significantly, our top upregulated candidate from the exon array in the aggressive metastatic tumour lines was the imprinted long non-coding RNA (lncRNA) H19. While H19 can act epigenetically to repress transcription, it can also associate with polyribosomes and therefore be capable of regulating translation. Hence, the precise mechanism by which H19 modulates metastatic potential remains unresolved and is under investigation.
We also identified DNA amplification of a region harbouring Neat1, an architectural lncRNA critical in the biogenesis of subnuclear organelles known as paraspeckles. Subsequent analyses revealed significant upregulation of this lncRNA transcript in highly metastatic tumours. In addition to its role as a molecular scaffold, Neat1 participates in the selective retention of hyperedited transcripts within the nucleus to prevent their translation. It is therefore possible that Neat1 facilitates metastasis by preventing the translation of yet to be discovered metastasis suppressors that we are seeking to identify through an RNA-seq approach by comparing total and nuclear RNA fractions in tumours with and without elevated Neat1 expression.
To compensate for the paucity of annotated lncRNA probesets on the exon array, we also interrogated our raw data using a custom ncRNA chip definition file devised by Risueño and colleagues (GATExplorer) to unambiguously re-map all intronic and un-assigned probes to bona fide lncRNA entities with accessions in RNAdb. We discovered over 400 unique long non-coding RNA transcripts associated with metastasis, nine of which exhibited corresponding DNA copy number variations.
In summary, we have implicated H19 and Neat1 and revealed a large set of long non-coding RNAs associated with breast cancer metastasis. Efforts are underway to explore the functional role of these lncRNAs in our metastasis models.