NEW YORK – With transcriptome profiling on patient-derived xenograft (PDX) samples, a Baylor College of Medicine research team has tracked down four expression-based metastatic subtypes that turn up in metastases from patients with a wide range of primary cancer diagnoses.
"These subtypes facilitate understanding of the molecular underpinnings of metastasis beyond tissue-oriented domains, with therapeutic implications," senior and corresponding author Chad Creighton, at Baylor College of Medicine, and his colleagues wrote.
As they reported in Cell Reports Medicine, the researchers first compiled RNA sequence data from 14 prior studies, representing 2,371 mouse-grown metastatic PDX tumors originating from 973 patients with more than 17 tissue-of-origin-based cancer types. That set encompassed 1,955 PDX models established using samples from 557 patients with two or more tumors.
The team analyzed the PDX expression data in concert with expression profiles for more than 10,200 tumors assessed for the Cancer Genome Atlas project and RNA-seq data for another 2,405 metastatic tumors from 24 studies focused on individuals with more than 26 cancer types — including 307 metastatic tumors with matched primary samples and tumor samples from 131 individuals with multiple metastases.
Together, the profiles pointed to a handful of potentially informative pan-cancer metastatic tumor molecular subtypes.
"By our analytical approach, the molecular subtypes spanned tumors of diverse lineages and tissues of origin and multiple datasets from independent laboratories," the authors explained. "The idea that cancer metastases can be categorized into a handful of distinct groups would have important implications for understanding the biology of metastasis."
Based on the somatic mutation, pathway, drug response, and transcription factor binding features found in the PDX samples, analyzed in relation to expression profiles from cancer-free cells or tissues and matched primary tumor samples (when possible), the team highlighted four molecular subtypes in metastatic cancer, known as subtypes s1 to s4.
"The metastasis subtypes reflected expression differences from paired primaries, with subtype switching being common," the authors wrote, suggesting that "[m]etastatic cells that escape from the primary tumor may develop into tumors of a different molecular subtype from that of the primary, while still falling within one of a discrete set of subtypes."
In the s1 subtype, for example, the investigators saw rampant copy number changes, including gains affecting the MYC gene, along with DNA repair gene alterations and an apparent propensity to respond to bromodomain inhibitor treatment.
Higher-than-usual metabolism and prostaglandin synthesis or regulation were marks of the s2 subtype, they noted, while the s3 subtype had enhanced expression of genes involved in DNA or histone methylation, increased response to BCL2 inhibitors, and altered neuronal differentiation and EZH2 transcriptional target activity. Finally, the team described an s4 metastatic tumor subtype with elevated immune checkpoint component expression and ramped-up expression of Notch 2 pathway players.
Based on data for cases with matched primary tumor and metastatic samples, the team saw signs that metastatic tumors may fall into subtypes with molecular features that are distinct from those found in corresponding primary cancers, consistent with so-called subtype switching patterns that may influence targeted treatment strategies and responses.
"Our molecular subtypes could have important implications for applying existing therapies or developing alternate therapeutic approaches," the authors wrote, adding that the results so far indicate that "no single therapeutic approach would be effective for all cancers but that the gene expression profile and associated molecular subtype of the tumor could help maximize precision medicine approaches."