Can post-fixation samples be sorted in the Flex Gene Expression assay?

Question: Can post-fixation samples be sorted in the Flex Gene Expression assay?

Answer: Post-fixation samples, including those that have undergone storage, can be sorted using FACS for advanced sample clean-up, as well as for the enrichment of specific populations in the Flex Gene Expression assay.  

Enrichment of Specific Populations, either before or after storage:

Fixation is not expected to preclude FACS isolation for most fluorophores. Some fluorophores may be impacted by longer fixation.  We have performed limited testing for the enrichment of specific populations post-fixation (1 h fixation) and after fixation and storage (24 h fixation) using FACS. Cells can be labeled with fluorescent antibodies before fixation & storage (labeled live), or after storage (labeled after fixation). For highest quality labeling of cell surface antigens, cells should be labeled prior to fixation following the Cell Surface & Intracellular Protein Labeling for GEM-X Flex Gene Expression Demonstrated Protocol (CG000781).  Following cell surface protein labeling, the cells should be immediately fixed following the guidance in the Fixation of Cells & Nuclei Demonstrated Protocol. Labeling cells with fluorescent antibodies after fixation may still be possible, but the compatibility of an antibody with fixed samples needs to be tested. Labeling of intracellular antigens with fluorescent antibodies for isolation by FACS is also possible, see protocol Cell Surface & Intracellular Protein Labeling for GEM-X Flex Gene Expression Demonstrated Protocol (CG000781) for the intracellular protein labeling protocol.  

Limited testing for the enrichment of specific populations post-fixation using FACS found that although slight shifts in the SSC and FSC profiles were observed, both positive and negative populations could be distinguished in samples that were sorted post-fixation. Fluorescent reporter proteins such as EGFP and RFP may be dimmer after fixation, but -/+ populations could still be distinguished in samples that were sorted post-fixation. Additional optimization may be required for sorting samples post-fixation. A pilot experiment is recommended prior to committing to a larger study.

Note: tandem dyes should be avoided if possible, as they can degrade after fixation, leading to increased spillover. However, we have found that APC-Cy7 and PE-Cy7 are likely still compatible if their use is unavoidable.

Below is a demonstration for sorting labeled fixed cells. Human PBMCs were stained with two fluorescent antibodies targeting T cells and B Cells following Cell Surface & Intracellular Protein Labeling for GEM-X Flex Gene Expression Demonstrated Protocol (CG00078), and then fixed for either 1 h at room temperature, or overnight at 4C. After fixation, a portion of these cells were sorted for CD3 positive cells or CD19 positive cells. These sorted samples were then used in the Flex assay. The non-sorted sample gave comparable data to the sample sorted for all populations, suggesting that flow sorting does not seem to impact Flex data quality. The purity of each cell type, as well as the representative gene expression, for both 1 h fixed and overnight fixed samples for the individual sorted T cells and B cells is shown on the right.

Demonstration:

Below is a demonstration for sorting labeled fixed cells after storage. A mixture of HEK293 (CD45-) and Ramos (CD45+) cell lines were mixed together to generate a single suspension. This sample was then either labeled with a fluorescent anti-human CD45 Antibody (clone HI30, APC) before fixation & storage (live cells), or fixed and stored without antibody labeling. After storage at either 4C or -80C for one week, the stored samples were thawed and the unlabelled cells were then labeled with the anti-CD45 antibody (fixed cells). Note that after fixation and permeabilization, the cells are susceptible to degradation by RNases. Therefore, the antibody labeling buffer and washes were all performed with PBS + 1% nuclease- free BSA, supplemented with 0.2 U/ul RNase inhibitor (Protector RNase inhibitor from Sigma, PN-3335399001). Prior to undertaking the experiment, the ability of different anti-human CD45 clones to label fixed cells as well as unfixed cells was tested, finding the clone HI30 that was compatible with fixation. 

LIVE/DEAD™ Fixable Stains from Thermo Fisher Scientific can be used to determine the viability of fixed cells upstream of the Flex Gene Expression assay. This dye enables users to stain their unfixed cells, fix their cells, and then sort the labeled cells. Together, this provides the benefit of maximizing the likelihood of preserving sample viability and quality during FACS isolation. However, complex samples such as dissociated tumor cells with a mix of large and small cells may be trickier, we recommend a pilot experiment to optimize conditions and to ensure successful isolation of the correct populations.

We have performed limited testing with human PBMCs and found that the fixable dye (catalog number: L34971 used from Fixable Dead Cell Stain Sampler Kit catalog number: L34960) successfully allows for sorting fixed cells. To perform this experiment, a subset of PBMCs were heat-killed (incubated at 65°C for 1 min) and then mixed (at 1:1 ratio) with fresh human PBMCs to generate a sample with mixed viability. The PBMCs were then split and either: 

1. Labeled with Protein Feature Barcode-compatible antibodies, washed, stained with 7-AAD, sorted, and then fixed (see: Advanced Sample Cleanup below) or
2. Labeled with Protein Feature Barcode-compatible antibodies, washed, stained with LIVE/DEAD™ Fixable Stain, fixed, and then sorted or
3. Labeled with Protein Feature Barcode-compatible antibodies, washed, and fixed

In this experiment, the Gene Expression library quality (Valid Barcodes/UMIs), complexity (UMIs/Genes per Cell), and sensitivity (cell type distribution comparable) were largely unaffected by sorting (7-AAD or LIVE/DEAD™ stained cells) as compared with unsorted PBMCs. This is somewhat expected as the Gene Expression Flex application is fairly robust to lower-viability samples (see: What are the sample quality recommendations for optimal performance of the Flex Gene Expression assay?). When reviewing Protein Feature Barcode data, there were more significant differences between sorted and unsorted data as noise (background signal) and antibody aggregates were drastically reduced when implementing sorting to enrich for live cells. Furthermore, when comparing samples with sorted live cells vs unsorted, there were stronger gene expression and protein correlations.    

We expect that when users scale up to work with more samples or use more complex samples than PBMCs, that using LIVE/DEAD™ Fixable Stains will likely yield better data quality than sorting live cells.        

Advanced Sample Cleanup:

Post-fixation samples can be flow sorted for advanced sample clean-up to remove cellular debris from single-cell or nuclei suspensions using 7-AAD. DAPI, Hoechst, and other DNA fluorophores are also expected to be compatible with staining fixed and nucleated cells.

Debris heavy tissues such as brain may benefit from sorting to remove cellular debris after tissue fixation and dissociation following the Tissue Fixation & Dissociation Demonstrated Protocol. However, it is important to count post-sorting before proceeding with probe hybridization, given that the cell recovery will be reduced.

Best Practices for Sorting Fixed Samples:

• If storing samples, Consult Fixation of Cells & Nuclei for Chromium Fixed RNA Profiling Demonstrated Protocol (CG000478) for details on storage conditions and reagents.
• Protect samples from light during processing and storage
• Prior to sorting, place samples in Quench buffer
• Samples should be collected in PBS + 1% nuclease- free BSA supplemented with RNase inhibitor (Protector RNase inhibitor from Sigma, PN-3335399001).  We recommend a final concentration of 0.2 U/ul of RNAse inhibitor, accounting for dilution occurring during the sorting process.  
• Samples should be spun down and resuspended in 0.5X PBS + 0.02% nuclease-free BSA before proceeding with probe hybridization.

It is always recommended to consult a core facility or an experienced flow user before initiating a flow sorting experiment.

Products: Flex Gene Expression