(P 17) IDENTIFICATION OF MARKERS PREDICTIVE OF RESPONSE TO GEMCITABINE AND DOCETAXEL IN LEIOMYOSARCOMA THROUGH MIRNA SEQUENCING

Objective: The combination of gemcitabine and docetaxel (Gem/Tax) is one of the most commonly used systemic treatments for leiomyosarcoma (LMS); however, only 25-36% of patients show response to this treatment (PMID: 29220301). Despite the widespread use of this therapy, the underlying biological basis for this variability of response remains unknown, and there are no predictive biomarkers for the response to Gem/Tax in LMS. Identification of such biomarkers would facilitate personalized treatment for patients who are most likely to benefit from Gem/Tax therapy, while mitigating unnecessary drug exposure and associated toxicities in patients who are unlikely to benefit from this therapy. We previously presented the candidate markers associated with response to Gem/Tax therapy based on transcriptomic and epigenomic profiling of LMS tumors. Here, we sought to further expand this multi-omics study by miRNA sequencing of the same specimens.

Methods: miRNA expression was profiled in 35 archival untreated primary tumor specimens from patients who then underwent adjuvant treatment with Gem/Tax. These specimens were obtained from the Leiomyosarcoma Direct Research Foundation Tumor Bank hosted at Stanford University. Sixteen patients developed recurrence following the Gem/Tax treatment (non-responders), and 19 patients did not develop recurrence (responders). All patients had localized disease at the time of diagnosis. Libraries for miRNA sequencing were prepared using QIAseq miRNA library protocol (Qiagen) and sequenced on the Illumina NovaSeq 6000 instrument. The sequencing reads were mapped and filtered using dedicated GeneGlobe software (Qiagen). We performed differential expression analysis using SAMseq to identify miRNAs up-/down-regulated in responders and non-responders (with adjusted p < 0.001, fold change ≥ 1.5 and ≤ -1.5, respectively). Kaplan–Meier analysis was applied for survival analysis. Whole transcriptome RNA-seq data from the same specimens was used to investigate correlation between differentially expressed miRNAs and their candidate target genes. miRTarBase database (release 9.0) was used to identify known target genes of selected miRNAs. Pearson correlation was applied to identify genes that may be positively and negatively regulated by selected miRNAs.

Results: We identified three significantly up-regulated miRNAs in non-responders to Gem/Tax, including hsa-miR-196a-5p, hsa-miR-625-3p and hsa-miR-7977 (adjusted p < 0.001, fold change = 3.12, 1.87, and 1.74, respectively). For survival analysis, patients were classified into low and high expression groups using the 20th percentile of normalized expression values as the cut-off. Patients with low expression of these miRNAs showed a trend for favorable recurrence-free survival compared to patients with high expression of these miRNAs. Next, using the miRTarBase database, we identified 304 genes as candidate targets regulated by hsa-miR-196a-5p, 34 targets for hsa-miR-625-3p, and 643 targets for hsa-miR-7977. However, we have not identified any significant correlation between these miRNAs and their target genes that were previously reported in different types of tumors and normal tissues. This warrants further studies to elucidate the LMS-specific role and target genes of these differentially expressed miRNAs.

Conclusion: This is the first study that identifies miRNAs as potential markers linked to the response to Gem/Tax therapy in LMS patients. Specifically, we identify three miRNAs up-regulated in non-responders to Gem/Tax therapy, which may potentially contribute to resistance to this treatment. Our findings warrant further studies on the role of these miRNAs and their target genes in LMS.