Interrogating HER2+ Breast Cancer Through Multi-omic Bioinformatic Approaches

The development of numerous high-throughput technologies has enabled us to interrogate human epidermal growth factor 2 (HER2) positive breast cancer on a “panomic” level. These technologies allow us to examine the exome, transcriptome, proteome, and epigenome simultaneously. To demonstrate the effectiveness of research projects across multiple “omes” three studies were performed, each encompassing at least two different “omes”.

The first project looked at single nucleotide polymorphism (SNPs) in the homologous recombination repair (HRR) and HER2 pathways, compared them to proteomic data, and linked the findings with clinical outcome. I found 6 SNPs that correlate with recurrence free survival (RFS) when comparing patients who received TCH-based (docetaxel (T), carboplatin (C), and trastuzumab (H)) treatment versus those who received other treatment regimens, and an ERBB3 SNP that has an impact on protein phosphorylation in the PI3K/AKT and MAPK/ERK signaling pathways.

For the second and third projects, I focused on the phenomenon known as bivalency, which encompasses two histone 3 trimethylations, located at lysine 4 (H3K4me3) and lysine 27 (H3K27me3). In the second project, I looked at how they differ between estrogen receptor (ER) positive and negative HER2+ breast cancer cell lines, and what impact they had on gene expression. I confirmed that there was a strong correlation between H3K4me3 with higher gene expression and H3K27me3 with low to no gene expression, and that there are several differences in genes within the HER and ER pathways. Additionally, I compared the bivalency profiles of several downstream genes of these two pathways with gene expression in patients, and found genes differentially expressed, and significantly correlated with RFS.

The third project continued the bivalency studies by looking and how bivalency status in ER+/- HER2+ cell lines change in the presence of a positive and a negative environmental stimulus, and how that changes gene expression. E2, the ligand for the ER, was chosen as the positive stimulus, and trastuzumab, an anti-HER2 drug, was chosen as the negative stimulus. For this project, a new method was adapted by taking the framework of both the sequential chromatin immunoprecipitation followed by sequencing (reChIP-seq) and Cleavage Under Targets and Release Using Nuclease (CUT&RUN) to form what we call reCUT&RUN. After the reCUT&RUN protocol was developed, several workflows were tested to optimize the analysis of the sequencing data generated. KEGG pathway analysis revealed that treatment had little impact on the distribution of the H3K4me3 and H3K27me3 marks. However, those found to be truly bivalent using reCUT&RUN differed between the control, E2, and trastuzumab-treated groups. Finally, the inclusion of RNA-seq and differential expression analysis revealed that over 70% of genes that were significantly differentially expressed had the H3K4me3 mark and less than 1% were H3K27me3. This indicates that the bivalent marks may play a role in which genes could have their expression impacted by treatment, including several genes involved in epithelial-mesenchymal transition and the regulation of H3K4me3 and H3K27me3 enrichment. This chapter not only demonstrated the usefulness of the reCUT&RUN protocol over other conventional methods, but it also gives new insight to the role that bivalent markers may play in treatment response.

Overall, this thesis demonstrates the power of interrogating HER2+ breast cancer on a panomic level. By doing so, we can better characterize the complex interactions that take place across the exome, transcriptome, proteome and epigenome, allowing us to develop a better model to understand the complexities of oncogenesis and how patients respond to therapy.