Simultaneous evolutionary expansion and constraint of genomic heterogeneity in multifocal lung cancer

Abstract

Recent genomic analyses have revealed substantial tumor heterogeneity across various cancers. However, it remains unclear whether and how genomic heterogeneity is constrained during tumor evolution. Here, we sequence a unique cohort of multiple synchronous lung cancers (MSLCs) to determine the relative diversity and uniformity of genetic drivers upon identical germline and environmental background. We find that each multicentric primary tumor harbors distinct oncogenic alterations, including novel mutations that are experimentally demonstrated to be functional and therapeutically targetable. However, functional studies show a strikingly constrained tumorigenic pathway underlying heterogeneous genetic variants. These results suggest that although the mutation-specific routes that cells take during oncogenesis are stochastic, genetic trajectories may be constrained by selection for functional convergence on key signaling pathways. Our findings highlight the robust evolutionary pressures that simultaneously shape the expansion and constraint of genomic diversity, a principle that holds important implications for understanding tumor evolution and optimizing therapeutic strategies.Across cancer types tumor heterogeneity has been observed, but how this relates to tumor evolution is unclear. Here, the authors sequence multiple synchronous lung cancers, highlighting the evolutionary pressures that simultaneously shape the expansion and constraint of genomic heterogeneity.

Conflict of interest statement

The authors declare no competing financial interests.

Figures

Fig. 1

Clonal architecture and mutational landscape of MSLC. a Computed tomography (CT) diagnosis and clonal architecture of multifocal tumors in four MSLC patients. Heat maps showed the presence (blue) or absence (gray) of all non-silent somatic mutations in every tumor region. Phylogenetic trees indicated the clonal structure of sequenced tumor regions in each patient. Scale bar, 1 cm. b Mutational landscape of all 16 sequenced tumor regions. Putative driver genes with somatic mutations were classified according to the functional categories. c Frequencies for each of the six substitutions at all 16 possible trinucleotide contexts were displayed in a heat map. All specimens were separated into two major clusters based on the mutational signatures

Fig. 2

Interfocal and intrafocal genetic heterogeneity of MSLC. a A schematic overview of tumor heterogeneity analysis at interpatient, interfocal and intrafocal levels. b Mutation spectra of four MSLC patients were summarized by pie charts and shared mutations between different subjects were showed in Venn diagram. c Overexpression of indicated wild-type or mutant genes in BEAS-2B or LXF-289 cells. Gene expression and ERK phosphorylation were measured by immunoblotting. d Indicated wild-type or mutant genes were lentivirally introduced into BEAS-2B or LXF-289 cells. Cells were cultured for two weeks and stained with crystal violet. Scale bar, 5 mm. e Tumor growth of BEAS-2B xenografts that ectopically expressed indicated mutant genes. Each line represented mean tumor volume of the respective group, and error bars indicated standard deviation (10 mice per group). Scale bar, 10 mm. f A heat map presented the CCF of putative driver mutations in each sequenced region of the MSLC tumors. g Mutation spectra of early mutations (trunk) and late mutations (branch), and mutational signatures of different tumor regions in M-seq samples of RJLC1-T1 and RJLC4-T1

Fig. 3

Constrained tumorigenic pathways among multicentric lesions. a A summary of the most prominent driver mutation identified in each primary tumor of MSLC. b EGFR or KRAS was knocked out using CRISPR-Cas9 system in lung cancer cell lines and replaced with HER2, c-MET, ARAF, BRAF, or MEK mutants. c EGFR was knocked out using CRISPR-Cas9 system in PC9 cells and replaced with HER2YMVA. AKT and ERK phosphorylation were measured by western blot analysis. d Cells were treated with a serial dilution of indicated inhibitors for a week and stained with crystal violet. Scale bar, 5 mm. The corresponding cell lysates were analyzed by immunoblotting. e KRAS was knocked out using CRISPR-Cas9 system in NCI-H1944 cells and replaced with BRAFD594G, c-METΔE14, ARAFS214C, or MEK1E102-I103 del. ERK phosphorylation was measured by western blot analysis. Indicated cells were treated with a serial dilution of trametinib for a week and stained with crystal violet. Scale bar, 5 mm. The corresponding cell lysates were analyzed by immunoblotting

Fig. 4

Independent validation of oncogenic pathway convergence. a CT diagnosis of synchronous lung and breast tumors in RJLC5 and RJLC6. Scale bar, 1 cm. b Circos plots highlighting different structural variations of lung and breast tumors in RJLC5 and RJLC6. RJLC5-Lung harbored c-METΔE14 and RJLC5-Breast harbored HER2Y777L; RJLC6-Lung harbored EGFR-KDD and RJLC6-Breast harbored HER4N855K. c CT diagnosis of MSLC and circos plots representation of somatic mutations in RJLC7. EGFRL858R identified in RJLC7-T1 was confirmed using ARMS PCR. Scale bar, 1 cm. d Time line of RJLC7 MSLC diagnosis and treatment. Baseline and 21-month CT scan for RJLC7 following treatment with gefitinib. Scale bar, 1 cm. e A schematic model showing that simultaneous evolutionary expansion and constraint of genomic heterogeneity collaboratively shape tumorigenic process and clonal architecture