NEW YORK – Using a combination of transcriptomic, proteomic, and genomic data, a team led by researchers in the UK has identified a signature linked to tumors that "phenocopy" features of cancers with defects in the tumor suppressor gene RB1, despite containing a wild-type version of the gene themselves.
"[W]e propose an integrative proteogenomic approach that predicts cancers with dysregulation in the RB pathway," co-senior and co-corresponding authors Syed Haider and Christopher Lord, researchers at the Institute of Cancer Research's Breast Cancer Now Toby Robins Research Centre, and their colleagues wrote in a paper appearing in Science Advances on Wednesday, noting that "dysregulation of the RB pathway in cancers lacking RB1 genomic defects provides a molecular rationale for how these cancers could be treated."
For their study, the investigators brought together multiomic and clinical data for 1,093 breast cancer cases profiled for the Cancer Genome Atlas (TCGA) project, identifying protein abundance, phosphorylation, messenger RNA abundance, and other features found in tumors known to be RB1-defective or RB1-proficient.
Bringing in available transcriptomic data, the team identified a so-called "RBness" signature (RBNSig) for breast cancer, which was used to classify TCGA samples without clear RB1 proficiency or defects. Along with tumors known to contain RB1 alterations, the RBNSig flagged more than 300 TCGA breast cancer cases that mimicked, or phenocopied, the features found in RB1-defective tumors.
That, in turn, prompted the investigators to take a closer look at the clinical, prognostic, and functional effects of RBness in still other breast or ovarian cancer cases and in cell lines from several other cancer types.
"[T]he identification of an RBness group has translational implications for rationalizing clinical heterogeneity and understanding the unique molecular drivers of treatment resistance and sensitivity," the authors explained, arguing that "this group, alongside bona fide RB1-defective patients, also warrants inclusion in preclinical and clinical studies, thereby increasing both statistical power and opportunities to stratify patients into treatment-matched groups."
After using data for 74 breast cancer samples with matched DNA and protein profiles to compare features in RBness cases linked to RB1 mutations, altered levels of RB protein, or protein hyperphosphorylation, the investigators validated their breast cancer RBNSig in thousands more samples from the Metabric and SCAN-B cohorts.
Together, the results suggested that breast cancer cases marked by RBness tended to be aggressive, with sharper declines in overall survival than RB1-proficient cases, despite enhanced sensitivity to chemotherapy. On the other hand, the team's analysis of expression data from a Phase II clinical trial of the CDK4/6 inhibitor palbociclib hinted that RB1 mutations and RBness both coincided with decreased response to CDK4/6 inhibition in estrogen receptor-positive breast cancer.
Building on these findings, the investigators used multiomic TCGA data on 89 high-grade ovarian adenocarcinoma cases to establish a transcriptomic signature for RBness in ovarian cancer, identifying 41 genes linked to RBness in a broader TCGA ovarian cancer collection and in a set of 81 samples from a TCGA Clinical Proteomic Tumor Analysis Consortium.
When they applied the RBNSig scoring approach to 563 previously characterized pan-cancer cell line models, meanwhile, the investigators found that RB1-defective lines were overrepresented in the high scoring group.
But even in the absence of an RB1 defect, the RBness lines shared synthetic lethality patterns in published CRISPR-Cas9 experiments on the lines — results that the team confirmed using their own RNA interference experiments on RB1-wild type or RB1-deleted breast cancer cell lines.
Likewise, the team found that RB1-defective and RBness cell lines tended to show similar drug sensitivity patterns in analyses that used RBNSig scores to classify cell lines used in published drug response studies spanning 461 pan-cancer cell line responses to 481 compounds.
"We report RBness as a pan-cancer phenomenon, associated with patient outcome and chemotherapy response in multiple cancer types, and predictive of CDK4/6 inhibitor response in estrogen-positive breast cancer," the authors wrote, noting that "RBness cancers have a transcriptional program that is correlated with genes that are known to be synthetic lethal with RB1 mutation, suggesting new ways these cancers that phenocopy RB1 defects could be treated."