Poster M11 - Genetic network for discovering prediction marker of relapse of chemotherapy resistant TNBC

Introduction: Chemotherapy stands as the cornerstone of treatment for over 70% of cancer types, yet its efficacy is often compromised by the development of drug resistance due to the targeting of rapid replicating cells. Despite attempts to combat resistance through combination therapy, challenges persist. Molecular-based approaches, such as immune checkpoint therapy, offer promising alternatives, however certain cancers like TNBC which is lack viable targeted therapies due to their unique molecular profiles. To identify target genes, differential expressed genes (DEGs) analysis has shown promise results for selecting highly expressed genes compared to normal cohort. However, their clinical translation has been limited. This is attributed to single gene identification, contrasting with the broad-scale phenomenon of drug resistance, which involved multiple genes and complementary biological functions. Recognizing resistance as a multifactorial issue necessitates an approach focusing on resistance cohorts’ genes rather than individual genes. This underscores the need for a genetic approach in addressing chemotherapy resistant TNBC, prioritizing the identification of core nodes within genetic networks to maximize therapeutic impact. Our research has focused on constructing co-interact genes networks from public RNA-seq data of TNBC breast cancer, revealing candidate central gene of genetic network capable of re-sensitizing chemotherapy-resistant cells. Validation through doxorubicin-resistant cell lines confirmed drug responsiveness, using these candidates it is enabling proactive differentiation between relapse and NED status in TNBC patients treated with neoadjuvant chemotherapy. This genetic network analysis method offers promise in guiding drug target selection and predicting early relapse, providing a critical advancement in combating chemotherapy resistance.

Materials and

Methods: Genetic network
To construct the co-interact genetic network, we collected the TNBC breast cancer data from public microarray data of METABRIC. The total genes of the microarray data are 20,386 without any duplicated genes. We identified with 19,710 extreme expression genes among TNBC cohort. Genes with more than 10 % of sample were selected to build input data of extreme expression pattern analysis calculation. After construct the co-interact genetic network, by applying Girvan-Newman algorithm to find the hubness of network. In the hub, the central node of hubs was identified and noted as central genes.
Gene knockout
To assess the re-sensitizing chemotherapy sensitivity, two siRNA (BIONEER, 8614-1, 221061-1) identified from genetic network as central genes were separately treated with RNAiMAX (Invitrogen, 13778075) in 96 well plates. After 24 hour treating siRNA to cells, serial concentration up to 200 M of doxorubicin were treated and assess the cell viability assay with CCK-8 assay.
RNA sequencing
We sequenced triplicate of primary TNBC breast cancer cell line (MDA-MB-231) and doxorubicin resistant TNBC cell line (MDA-MB-231-R). We also sequenced a total of 15 chemotherapy treated TNBC with relapse after surgery and 12 neoadjuvant chemotherapy treated TNBC with no evidence of disease after surgery. Overall, there were 27 TNBC cohort RNA-seq data in our data set. All clinical derived samples were obtained through a biobank in Seoul Asan Medical Center with the consent from patients and appropriate approval from ethic committees.

Results, Conclusions, and Discussions: To investigate TNBC breast cancer specific genetic network, we combined the public microarray transcriptome data and constructed the genetic network with cohort’s extreme gene expression pattern among TNBC breast cancer. TNBC specific genetic network has functionality focused features such as ATP binding genes, transmembrane transport, cell to cell signaling, and UDP-glucose dehydrogenase. The network featured gene relationships are showing network specific featured gene relationship especially in TNBC breast cancer cohort. To figure out the functionality of genetic hub, we applied clustering algorithm to make central connectivity by repeated random gene selection. From the different hub, we can get the central node genes that are related to TNBC phenotype. To figure out the featured genes can predict the relapse of TNBC, we sequenced the neoadjuvant chemotherapy treated TNBC breast cancer patient’s surgery tissues and compared between relapse group and disease-free survival group. We can cluster those two groups with co-interact genetic network selected genes without giving information of relapse.
Through the gene analysis method we propose, we have identified a central node gene that regulates the biological functions in chemotherapy resistant TNBC patients. This gene accelerates the phenotype of resistance, as confirmed by observing re-inhibited cell viability upon knockout of this gene when exposed to the same drug. Furthermore, by assessing this gene in samples from 27 TNBC breast cancer patients who underwent chemotherapy, we have validated its potential as a prognostic factor for distinguishing between relapse and disease-free survival. This suggests its utility as a predictive factor for treatment outcomes.

Acknowledgements (Optional):