Fragmentation of Small-cell Lung Cancer Regulatory States in Heterotypic Microenvironments

Abstract

Small-cell lung cancers derive from pulmonary neuroendocrine cells, which have stemlike properties to reprogram into other cell types upon lung injury. It is difficult to uncouple the plasticity of these transformed cells from heritable changes that evolve in primary tumors or select in metastases to distant organs. Approaches to single-cell profiling are also problematic if the required sample dissociation activates injury-like signaling and reprogramming. Here, we defined cell-state heterogeneities in situ through laser capture microdissection-based 10-cell transcriptomics coupled with stochastic-profiling fluctuation analysis. Using labeled cells from a small-cell lung cancer mouse model initiated by neuroendocrine deletion of p53 and Rb, we profiled cell-to-cell transcriptional-regulatory heterogeneity in spheroid cultures and liver colonies seeded intravenously. Fluctuating transcripts in vitro were partly shared with other epithelial-spheroid models, and candidate heterogeneities increased considerably when cells were delivered to the liver. Colonization of immunocompromised animals drove the fractional appearance of alveolar type II-like markers and poised cells for paracrine stimulation from immune cells and hepatocytes. Immunocompetency further exaggerated the fragmentation of tumor states in the liver, yielding mixed stromal signatures evident in bulk sequencing from autochthonous tumors and metastases. We identified dozens of transcript heterogeneities that recur irrespective of biological context; their mapped orthologs brought together observations of murine and human small-cell lung cancer. Candidate heterogeneities recurrent in the liver also stratified primary human tumors into discrete groups not readily explained by molecular subtype.We conclude that heterotypic interactions in the liver and lung are an accelerant for intratumor heterogeneity in small-cell lung cancer. Statement of significanceThe single-cell regulatory heterogeneity of small-cell lung cancer becomes increasingly elaborate in the liver, a common metastatic site for the disease.

Pan-cancer Drivers are Recurrent Transcriptional Regulatory Heterogeneities in Early-stage Luminal Breast Cancer

“…4B and Supplementary Table ST1). RHEGs provide a conceptual framework for prioritizing cell-state regulatory heterogeneities identified in vivo (49,50).…”

“…Regardless of the output, all leading approaches dissociate tumors into single-cell suspensions, requiring up to an hour of tissue processing and yielding a range of carcinoma proportions depending on cancer type ( Table 1). The impact of these preparative steps on the transcriptomes of live cells is recognized (11) [and see also the accompanying study (12)], but they are considered to be an unavoidable tradeoff of the scRNA-seq approach.…”

PremalignantNf1, Trp53-null Oligodendrocyte Precursor Cells Become Stalled in a Heterogeneous State of Replication Stress Before Gliomagenesis

“…Half of the 90 dpi mutant samples were similar to 12 dpi bulk observations, indicating an averaged population that was still undergoing adaptation, while the other half mapped toward the 150 dpi tumors, suggesting a fraction of cells in those bulk samples were transitioning to a gliomagenic state. The confused (de)differentiation of mutant OPCs during premalignancy [conceptually similar to that in the accompanying contribution (20)] undoubtedly contributed intrinsic variance to the very small number of differential transcripts detected at 90 dpi (Fig. 4B).…”

“…S2). Control OPCs were singular and much more entangled with unlabeled neighbors compared to prior applications of fluorescence-guided LCM by our group (20,34). To assess overall purity of normal controls, we used a collection of brain cell typeenriched marker transcripts (25) to construct a signature matrix for bulk sample deconvolution using CIBERSORT (24).…”

“…Here, we deconstructed the glioma GEMM's premalignant phases by a combination of bulk transcriptomics and targeted analysis of heterogeneous cell states through 10-cell RNA sequencing (10cRNA-seq) (17)(18)(19). Both approaches used laser-capture microdissection (LCM) to isolate labeled cells in situ to avoid tissue dissociation-induced changes of gene expression would confound our analysis (20). Shortly after deletion-labeling, bulk transcriptomic differences largely indicated generic adaptations to Nf1-Trp53 loss, whereas the heterogeneity analysis of transcriptional variation gave clues about the early states of labeled OPCs.…”