TL;DR
This study uses single-cell RNA sequencing to characterize circulating tumor cells (CTCs) in the cerebrospinal fluid (CSF) of lung cancer patients with brain metastases. They identified unique gene expression profiles, significant heterogeneity, and potential biomarkers, paving the way for improved diagnostics and targeted therapies for this challenging condition.
What Problem Does This Paper Address?
Lung cancer is a leading cause of cancer death, and brain metastases (BMs) are a common and often fatal complication, particularly in lung adenocarcinoma (LUAD). Leptomeningeal metastases (LMs), where cancer cells spread to the membranes surrounding the brain and spinal cord, are associated with poor prognosis and limited treatment options due to poor drug penetration into the central nervous system (CNS). While CSF cytology is the gold standard for diagnosing LMs, the molecular and transcriptional characteristics of these circulating tumor cells (CTCs) in the CSF (CSF-CTCs) have remained largely undefined. This lack of understanding hinders the development of effective diagnostic tools and targeted therapies. This paper aims to fill this knowledge gap by performing a comprehensive single-cell transcriptome analysis of CSF-CTCs from LUAD-LM patients.
Methods
The researchers employed a multi-step approach:
- CSF-CTC Isolation: They developed an effective protocol to collect and isolate individual CSF cells from patients. This involved fluorescence-activated cell sorting (FACS) to enrich for tumor cells (CD45-negative, live cells) and exclude immune cells.
- Single-Cell RNA Sequencing (scRNA-seq): A total of 3,792 single cells were sequenced using the Smart-seq2 protocol. This technique allows for the detailed analysis of gene expression at the individual cell level, crucial for capturing cellular heterogeneity.
- Data Analysis:
- Quality Control and Filtering: Raw sequencing data underwent rigorous quality control, including adapter trimming, alignment to the human genome, and filtering of low-quality cells and genes.
- Cell Type Identification: Unsupervised clustering using t-distributed stochastic neighbor embedding (t-SNE) was performed to group cells based on their gene expression profiles. Marker genes were used to identify different cell types, including T cells, B cells, monocytes, and CTCs.
- Differential Gene Expression Analysis: DESeq2 was used to identify differentially expressed genes (DEGs) between CSF-CTCs and leukocytes.
- Pathway Enrichment Analysis: Gene Set Enrichment Analysis (GSEA) was employed to identify enriched biological pathways in CSF-CTCs.
- Cell Cycle Analysis: The CellCycleScoring function in Seurat was used to determine the cell cycle phase of individual cells.
- Cancer Stem Cell (CSC) and Epithelial-to-Mesenchymal Transition (EMT) Analysis: Expression of known CSC markers and genes associated with EMT was investigated.
- Cancer-Testis Antigen (CTA) Analysis: The expression of a panel of CTAs was examined to assess their contribution to heterogeneity.
Datasets
- Patient Samples: CSF samples were collected from five patients with LUAD-LM and one patient with Cancer of Unknown Primary site (CUP)-LM. Three CSF samples from healthy individuals served as controls.
- Cell Lines/Xenografts: For comparison, single-cell RNA-seq data from a human NSCLC cell line (H358) and two human LUAD patient-derived xenograft (PDX) samples (MBT15 and PT45) were also analyzed.
- Publicly Available Data: Gene expression data for PDX samples and the H358 cell line were obtained from the NCBI Sequence Read Archive (accession number GSE69405).
Key Insights and Ideas
- CSF-CTCs are Distinct and Heterogeneous: The study clearly defines CSF-CTCs based on their unique transcriptome signatures, differentiating them from normal CSF cells. Crucially, it reveals substantial heterogeneity not only between patients but also within individual patients' CTC populations.
- Metabolic and Adhesion Pathways are Key: Enriched pathways in CSF-CTCs point to enhanced energy metabolism (glycolysis, TCA cycle) and cell adhesion mechanisms. These are vital for tumor cell survival, migration, and invasion in the challenging brain microenvironment.
- Partial EMT is Present: While CSF-CTCs retain epithelial characteristics, a subset exhibits features of partial epithelial-to-mesenchymal transition (EMT), suggesting a more adaptable and invasive phenotype.
- Cancer-Testis Antigens (CTAs) Contribute to Heterogeneity: The diverse expression of CTAs across CTCs from different patients highlights their role in driving inter-patient variability.
- CUP Case Provides Diagnostic Clues: The analysis of a CUP-LM patient's CSF-CTCs demonstrated the potential of scRNA-seq to aid in identifying the primary tumor origin, even when traditional methods are inconclusive.
Key Results
- Cell Composition: Normal CSF samples primarily consisted of T cells (80.3%) and monocytes (19.7%), with no B cells detected.
- CSF-CTC Identification: The study successfully identified and characterized CSF-CTCs from LUAD-LM patients, defining them by epithelial and lung origin markers, and proliferation markers, while lacking immune cell markers.
- Transcriptome Signatures: 290 genes were significantly upregulated in CSF-CTCs compared to normal CSF cells. Key enriched pathways included energy metabolism and cell adhesion.
- Heterogeneity: Inter-patient heterogeneity was significantly greater than intra-patient heterogeneity. This was linked to spatial distribution of metastases, cell cycle, CTA expression, and partial EMT.
- Cell Cycle: The majority of LUAD-LM CSF-CTCs were in a non-cycling state (average 7.2% high-cycling cells), a lower proportion compared to cell lines or PDX models.
- Cancer Stemness and EMT: While epithelial markers were abundant, a subset of CSF-CTCs showed partial EMT features. Stem cell markers like CD44 were detected, but not always correlated with EMT markers.
- CTAs: Significant heterogeneity in CTA expression was observed among patients, with some CTAs like SPAG9 showing broad expression and potential as therapeutic targets.
- CUP Case: For the CUP-LM patient, scRNA-seq data, combined with IHC, helped narrow down the potential primary origin to breast, sweat/salivary gland, or prostate cancer, with specific genes like PIP and ANKRD30A being P8-specific.
Limitations and Future Work
- Sample Size: While the study analyzed a significant number of single cells, the number of patients (five LUAD-LM, one CUP-LM) is relatively small. This limits the generalizability of some findings, particularly regarding inter-patient heterogeneity.
- Temporal Heterogeneity: While temporal differences were observed between samples collected 6 months apart from the CUP patient, the LUAD-LM patient samples collected within a 2-month interval showed similar patterns. More extensive longitudinal sampling would be beneficial to fully understand temporal heterogeneity.
- CUP Diagnosis: Despite promising scRNA-seq insights, a definitive primary tumor origin for the CUP-LM patient could not be established due to the patient's passing and refusal of autopsy.
- Future Directions: The authors propose developing an RNA-based digital detection method for CSF-CTCs to aid in LUAD-LM diagnosis. They also aim to investigate the interactions between CSF-CTCs and the CSF microenvironment, which could reveal further therapeutic targets. The potential of identified genes (e.g., CEACAM6, SCGB3A2, C3, MMP7, CLDN7, ICAM1) as therapeutic targets for decreasing metastatic ability is also highlighted.
Summary updated