Comprehensive genomic characterization of matched primary-metastatic lung adenocarcinomas using a multiparameter nuclei flow-sorting approach

Lung cancer is the most frequent cancer worldwide and is responsible for more death than any other tumor type. Genomic intratumor heterogeneity (ITH) is thought to be implicated in the resistance to current therapies, but in spite of its high prevalence, only a few studies have investigated genomic ITH in non-small cell lung cancer (NSCLC). Moreover, these studies have mainly focused on primary tumors. Yet, it is the progression to metastatic disease that makes lung cancer so lethal.
We sought to shed light on the evolution of metastasizing lung adenocarcinoma (LUAD) by investigating the extent of genomic ITH in clonally related primary and metastatic sites in a cohort of 16 LUAD patients. We established and validated the use of a refined multiparameter nuclei flow-sorting approach and demonstrated that it substantially increases the purity of tumor DNA for subsequent genomic analyses. We applied this method to sort tumor populations from bulk tissues in order to determine somatic copy number aberrations (SCNAs) by array-comparative genomic hybridization (aCGH) and single nucleotide variants (SNVs) in 409 well-known cancer genes by targeted ultra-deep sequencing.
This comprehensive genomic analysis revealed that 88% of SCNAs and 78% of SNVs were propagated from primary tumors to metastases, which indicated their accumulation in the primary tumors before metastatic dissemination. The fact that SCNAs were not increased at metastatic sites resulted in stable ploidies across our cohort even over long time periods. Despite the fact that the pure presence of SCNAs implies chromosomal instability (CIN) at some point during tumorigenesis, our data suggested that there is no or little increase in CIN during metastatic progression. Nevertheless, the analysis of shared and private alterations demonstrated a broad continuum from linear to parallel progression, indicating a late and early dissemination of metastases from their primary tumors, respectively. Furthermore, we confirmed that both primary tumors and metastases can be the source of further metastases. However, no recurrent alterations enriched in metastases were detected, which suggested that LUADs obtain their metastatic phenotype via additional mechanisms that were not investigated by this work, such as modification on an epigenetic and transcriptomic level in interaction with the tumor microenvironment and the immune system.
Our data might have important clinical implications, because they indicate that most of the SCNAs and driver mutations can be detected in the primary tumor. However, the observation that established metastases spread to other sites in a cascading manner argues in favor of aggressive local treatment of all metastatic sites in patients with oligometastatic disease.