TL;DR
This study demonstrates that human lung, spleen, and oesophagus tissues remain remarkably stable for single-cell RNA sequencing (scRNAseq) for up to 72 hours when stored cold. This finding is crucial for large-scale projects like the Human Cell Atlas, enabling more flexible sample collection and processing logistics without compromising data quality.
What Problem Does This Paper Address?
The Human Cell Atlas (HCA) aims to create a comprehensive map of all human cell types, relying heavily on single-cell RNA sequencing (scRNAseq) to generate high-resolution transcriptional data. However, a significant logistical challenge in obtaining human tissue samples is the delay between collection and processing. These delays, often due to distance between clinics and labs or scheduling issues, can lead to sample degradation and compromised scRNAseq data quality. While cold storage is a common practice, there hasn't been a systematic evaluation of its impact on the quality of scRNAseq data from key human organs like the lung, spleen, and oesophagus over extended periods. This paper addresses this gap by investigating how cold storage up to 72 hours affects the integrity and quality of scRNAseq data from these tissues.
Methods
The researchers employed a multi-pronged approach to assess tissue stability:
- Sample Collection and Storage: Fresh, healthy lung, spleen, and oesophagus tissues were obtained from organ donors. Samples were immediately placed in cold HypoThermasol FRS preservation solution at 4°C and processed at various time points: immediately (T0), 12 hours, 24 hours, and 72 hours.
- Single-Cell RNA Sequencing (scRNAseq): The 10x Genomics 3' v2 platform was used to generate scRNAseq libraries from the processed tissues. This is a standard, high-throughput method for capturing transcriptomes at the single-cell level.
- Bulk RNA Sequencing (bulk RNA-seq): In parallel, bulk RNA-seq was performed on some samples to compare with single-cell data and assess overall gene expression changes.
- Whole Genome Sequencing (WGS): WGS was conducted on donor samples to confirm the absence of gross genomic abnormalities.
- Histology: Hematoxylin and eosin staining was used to assess tissue health and morphology at different time points.
- Computational Analysis: Standard bioinformatics pipelines were used for read mapping, quality control, cell filtering (e.g., EmptyDrops algorithm), and dimensionality reduction (UMAP). Novel aspects included specific analyses to assess the impact of storage time on various quality metrics, mitochondrial read percentages, ambient RNA contamination, and cell type distribution.
Datasets
The study utilized datasets generated from:
- Human Tissues: Lung, spleen, and oesophagus samples from 12 organ donors.
- scRNAseq Data: Approximately 240,000 high-quality single-cell transcriptomes were generated.
- Bulk RNA-seq Data: Generated from some samples for comparative analysis.
- Whole Genome Sequencing (WGS) Data: From 13 donors.
The generated datasets, including detailed annotations, are made available through the Human Cell Atlas Data Co-ordination Platform and a dedicated website (www.tissuestabilitycellatlas.org).
Key Insights and Ideas
- Cold Storage is a Viable Strategy: The most significant takeaway is that cold storage at 4°C is a robust method for preserving human tissue for scRNAseq for at least 24 hours, and often up to 72 hours, without major data quality degradation. This is a critical enabler for large-scale human genomics projects.
- Tissue-Specific Differences Emerge: While generally stable, the spleen showed some specific changes at 72 hours, including increased mitochondrial reads (indicating cellular stress) and ambient RNA contamination. This highlights that while the method is broadly applicable, subtle tissue-specific responses to prolonged cold storage can occur.
- Cellular Stress vs. Degradation: The increase in mitochondrial reads at 72 hours in spleen suggests cellular stress rather than widespread RNA degradation, which would manifest as a broader loss of transcript integrity.
- Ambient RNA is a Key Contaminant: The study emphasizes the role of ambient RNA (background RNA in droplets) as a source of contamination, particularly at later time points in spleen, and notes that genes driving this contamination are often related to cell death.
Key Results
- Overall Stability: Lung and oesophagus tissues remained stable across all tested metrics for up to 72 hours of cold storage.
- Spleen at 72h: In spleen, at 72 hours, there was a decrease in reads mapped to exons and an increase in intronic reads, suggesting a potential skewing of transcript representation. Additionally, a higher percentage of mitochondrial reads was observed, indicative of cellular stress.
- No Major Impact on Cell Type Diversity: The diversity of cell populations identified by scRNAseq was largely maintained over the storage period across all tissues.
- Minimal Impact on Bulk RNA-seq: Bulk RNA-seq data showed no significant changes associated with the length of cold storage.
- Donor Variability is Significant: The study highlights that donor-to-donor variability in cell type composition and gene expression can be greater than the variation introduced by cold storage time, a crucial consideration for future HCA study design.
Limitations and Future Work
- Spleen-Specific Changes: The observed changes in spleen at 72 hours, particularly increased mitochondrial reads and ambient RNA contamination, warrant further investigation. While not a complete loss of data, it suggests that for spleen, a 24-hour window might be optimal for the highest quality data, or that specific cell types within the spleen are more sensitive.
- Limited Number of Donors for Viability: While cell viability was assessed, the statistical significance of observed decreases was limited by the small number of donors, suggesting that more extensive viability studies might be beneficial.
- Future Work: The authors suggest that this method is easily adoptable and will greatly facilitate primary sample collections for Human Cell Atlas studies. They also highlight the importance of accounting for anatomical heterogeneity and significant donor heterogeneity in future study designs. The generated dataset is a valuable resource for further investigation into the biology of these tissues and for future tissue-specific, single-cell eQTL studies.
Summary updated