Integrated genomics and comprehensive validation reveal drivers of genomic evolution in esophageal adenocarcinoma

Esophageal adenocarcinoma (EAC) is associated with a marked genomic instability, which underlies disease progression and development of resistance to treatment. In this study, we used an integrated genomics approach to identify a genomic instability signature. Here we show that elevated expression of this signature correlates with poor survival in EAC as well as three other cancers. Knockout and overexpression screens establish the relevance of these genes to genomic instability. Indepth evaluation of three genes (TTK, TPX2 and RAD54B) confirms their role in genomic instability and tumor growth. Mutational signatures identified by whole genome sequencing and functional studies demonstrate that DNA damage and homologous recombination are common mechanisms of genomic instability induced by these genes. Our data suggest that the inhibitors of TTK and possibly other genes identified in this study have potential to inhibit/reduce growth and spontaneous as well as chemotherapy-induced genomic instability in EAC and possibly other cancers.

The authors utilised existing TCGA datasets to perform a customised intergrative analysis that takes into account copy number variation, transcriptomic changes and survival, and found genomic instability gene signature to be overexpressed in the EACs samples. The observation of genomic instability is consistent with other publications (PMID: 28930282,PMID: 23604115) Of all the genes identified that supposedly contribute to genomic instability, the authors chose 3 for further in vitro and in vivo studies. The data established that these 3 genes accelerate tumour growth in EAC.
Of all the genes found to contribute to genomic instability , it will be informative if the authors can classify them according to the roles they play in genomic instability -which aspect of genomic instability they are enriched in EAC? In addition, the authors should justify why the 3 genes -TTK, TPX2, RAD54B are chosen for further studies. The authors have selected 2 EAC cell lines for their downstream studies, but the expression of TTK, TPX2, RAD54B are known to be high in these cell lines to begin with, which may compromise the strength of subsequent functional genomic analysis. It will be more powerful if the authors could involve and screen more EAC cell lines (which are available), and then select cell lines with low and high expression of these 3 genes for their downstream studies. Another question is whether overexpression of these 3 genes in normal non-cancerous cells also increases genomic instability, or their function is dependent on the context of cancer. In vivo TTK-I treatment does reduce tumour volume, but again, addition of cell lines with low expression of the 3 genes will support evidence of the efficacy of the inhibitor.
It will be helpful to the readers if the authors can reorganise the data and figures to improve readibility, flow and clarity Reviewer #2 (Remarks to the Author): Through integrative genomic data analyses and in-depth in vitro and in vivo functional validations, Kumar and colleagues have identified several driver genes (e.g. TTK, TPX2 and RAD54B) whose overexpression are highly associated with genomic instability in esophageal adenocarcinoma. Interestingly and importantly, they also showed that TTK inhibitor synergistically increases chemotherapy-induced cytotoxicity while inhibiting rather than promoting genomic instability in surviving cells. This is a well-designed study providing strong and clean evidences on the drivers of genomic instability in EAC. I have one major comment and several minor ones.
Major comments: 1. In the last section of Results, the authors performed experimental assays on HR activity and DNA breaks to show that TTK inhibitor inhibits spontaneous DNA damage and HR activity, and reverses genomic instability caused by chemotherapeutic agent. Although these experiments are essential and important, genomic sequencing (e.g. WGS/WXS) experiments are also requisite in order to verify these important findings. For instance, would the copy number/mutational burden be decreased in combined TTK inhibitor and chemo treatments as compared to chemo agent alone?
Minor comments: 1. Fig. 5C -II, the mutational signatures in normal control should also be shown.

Point by point response to reviewers' comments and details of revisions.
Reviewer # 1 #1: Comment 1: Of all the genes Of all the genes found to contribute to genomic instability , it will be informative if the authors can classify them according to the roles they play in genomic instabilitywhich aspect of genomic instability they are enriched in EAC?

Response:
We have now added a Response: For practical reasons, in most screens, usually one or two hits are investigated in depth in a single paper. We chose three genes which belonged to diverse pathways of genome stability/growth. These included TTK, a kinase; TPX2, a spindle assembly factor; and RAD54B, a homologous recombination protein. This information is now clearly provided in the paper (Revised manuscript: from last two lines of page 4 to first two lines of page 5).

Comments 3 and 4:
The authors have selected 2 EAC cell lines for their downstream studies, but the expression of TTK, TPX2, RAD54B are known to be high in these cell lines to beginSuppl emen ry Figurwith, 4A which may compromise the strength of subsequent functional genomic analysis. It will be more powerful if the authors could involve and screen more EAC cell lines (which are available), and then select cell lines with low and high expression of these 3 genes for their downstream studies. Another question is whether overexpression of these 3 genes in normal non-cancerous cells also increases genomic instability, or their function is dependent on the context of cancer. In vivo TTK-I treatment does reduce tumour volume, but again, addition of cell lines with low expression of the 3 genes will support evidence of the efficacy of the inhibitor.

Response:
We initially used one normal esophageal cell type to study the impact of overexpression of these genes, one EAC cell type to study the impact of further increase (overexpression) of these genes and two EAC cell lines to study the impact of knockdown of these  We have now also shown the impact of the overexpression of these genes in EAC cell line (OE19) on spontaneous DNA breaks, DNA end resection, and HR activity (New Supplementary Figures 4 A and B) as well as genomic instability (New Supplementary Figure  6). The impact of the overexpression on spontaneous DNA breaks and DNA end resection, a distinct step in the initiation of HR, is now also shown in EAC (FLO-1) cells (New Supplementary Figure  5). We also suppressed these genes in normal cells (fibroblasts). However, since normal cells have very low levels of expression of these genes and relevant activities, the knockdown did not produce any conclusive data, which was expected (not shown). So now, we have done overexpression of these genes in both the normal cells (which have low expression of these genes) and cancer cell lines (with already high expression of these genes) as well as suppression of these genes in cancer cell lines. Moreover, we have used multiple approaches and methods including the evaluation of spontaneous DNA breaks, DNA end resection, HR activity, micronuclei, single nucleotide polymorphism arrays and whole genome sequencing to demonstrate that increased expression of TTK, TPX2 and RAD54B disrupts genome stability (This information is provided in Lines, 5 -16, Page 6 of revised manuscript and data shown in Supplementary Material).

Reviewer #2.
Major comments: 1. In the last section of Results, the authors performed experimental assays on HR activity and DNA breaks to show that TTK inhibitor inhibits spontaneous DNA damage and HR activity, and reverses genomic instability caused by chemotherapeutic agent. Although these experiments are essential and important, genomic sequencing (e.g. WGS/WXS) experiments are also requisite in order to verify these important findings. For instance, would the copy number/mutational burden be decreased in combined TTK inhibitor and chemo treatments as compared to chemo agent alone?
Response: 1) We used both the WGS and SNP arrays to confirm impact of the overexpression of these genes in normal cells. For combination experiments, we demonstrated impact on genome stability by evaluating micronuclei (marker of ongoing genome stability), DNA breaks, and HR (a mechanism of ongoing copy number and LOH events in cancer; Shammas et al. 2009;Pal et al. ).
To further demonstrate the impact on copy number changes, we now demonstrate this using AxiomTM Precision Medicine Diversity Arrays. We show that etoposide increased the