Bevacizumab-containing therapy improves progression-free survival (PFS) in human epidermal growth factor receptor (HER) 2-negative metastatic breast cancer (mBC), but its use has been questioned due to the lack of benefit in overall survival (OS). To date, biomarkers to predict its positive effect are not available. Interestingly, miRNAs have emerged as regulators of most processes, forming tight interconnected feedback loops with genes under their regulation. Currently, different analyses, such as microarray, are performed in order to identify miRNA biomarkers, using formalin-fixed and paraffin-embedded (FFPE) tissue samples.
In the present study we recorded clinical data from 57 mBC patients, and selected two (4 + 4) PFS extreme groups for the analysis of the miRnome. Three miRNAs were used for normalization (U6, 191-5p and 103-a-3p). miRNAs for model construction were selected by differential expression between the two groups. Candidates were measured in the remaining 49 cases, and stepwise based Akaike criterion was used for profile generation. Additionally, integrative miRNA and mRNA analyses were done to reveal markers and pathways with potential clinical impact.
We selected two groups of patients with extreme differences on PFS (2.48 ± 1.85 vs 35.43 ±8.03 months) for their miRnome study. The expression profiles of miRNAs in both groups were highly correlated, except for 13 miRNAs where statistical differences arised. These miRNAs were selected as candidates for profile generation on the 49 additional cases, and a combination of five of them (miR-362-3p, miR-150b-5p, miR-671-3p, miR-744-3p and miR-941) was able to accurately discriminate two PFS groups. Additionally, Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses on miRNA possible target genes revealed interesting pathways to explore in these patients, such as cellular adhesion.
By combining experimental approaches and computational biology, we have identified candidate markers of outcome for bevacizumab-containing therapy. The five miRNAs included in the prognostic profile and cellular adhesion related genes should be explored as potential biomarkers.
Clinical trial identification
Legal entity responsible for the study
Fundación para la Investigación del Hospital Universitario La Paz
All authors have declared no conflicts of interest.