Human brain tissues are challenging samples to prepare due to considerable variability in terms of tissue types and logistic challenges in sample handling, potentially leading to delayed post mortem interval. For instance, there are reports of the Cerebellum undergoing autolysis much quicker than other regions. These challenges can lead to unpredictable quality in fresh frozen and FFPE blocks.
While we have achieved good results with healthy human brain FFPE samples using the Xenium FFPE - Tissue Preparation Guide (CG000578), we have observed variability across blocks, with some showing low sensitivity. These blocks yielding lower sensitivity tended to be of lower sample quality. While our testing is primarily focused on human FFPE samples, we present theoretical fresh frozen considerations here as well.
Fresh Frozen Sample Processing Considerations
It is necessary to preserve the brain’s structural and cellular integrity during fresh-frozen block preparation. Key areas to focus on include tissue handling, freezing methods, temperature control, and tissue quality assessment.
-
Tissue handling:
- Reducing Postmortem Delay: Process the brain as quickly as possible to reduce cell autolysis and RNA degradation.
- Maintaining Cold Chain: Keep tissue cold with ice or chilled saline, with a short time between embedding in OCT. 10x has not tested internally, but a theoretical recommendation of less than an hour is a starting point.
- Moisture-free samples: If samples contain moisture prior to freezing, this can lead to freezing artifacts during freezing and embedding in OCT.
Failure to observe these considerations may result in decreased sample quality.
- Freezing methods: Follow the OCT methods outlined in the Xenium FF - Tissue Preparation Guide (CG000579) and consider using cryomolds shaped to the size of the Xenium capture area to ensure minimal post-mold construction processing.
- Temperature control: Ensure uniform cooling by surrounding tissue with an even layer of pre-chilled embedding media. Post-freezing, store blocks at -80°C and minimize freeze-thaw cycles to maintain optimal RNA and tissue integrity.
- Tissue Quality: The quality of the tissue is crucial. Fresh frozen human brain tissues can be challenging to work with due to potential freezing artifacts, which can negatively impact cell segmentation and overall data quality. It is recommended to assess tissue quality through histological staining (e.g., H&E staining and DAPI) and RIN to check for any artifacts. While we do not have guidelines for RIN for Xenium, a RIN of more than 4 (aligned with Visium) is a good starting point for brain tissue. Even with blocks having a RIN of 4+, suboptimal performance is still possible. DAPI QC is outlined in the Tissue Preparation Guides.
FFPE Sample Processing Considerations
A challenge with brain tissue is its size, as often large pieces (e.g., entire brains) are placed in fixatives prior to cutting. This can be challenging as fixative oftentimes does not penetrate the tissue optimally. In this article, it is stated that “it can take 20 to 46 days for a sufficient amount of formaldehyde to diffuse to the innermost parts of a brain hemisphere and begin fixation. During this time, tissue in the inner regions of the brain will undergo microbial degradation, autolysis, breakdown of cellular membranes, and stochastic diffusion of molecules. As a result, immersion fixation causes gradients in fixation quality, whereby the surface regions where the fixation was initially applied have substantially better tissue preservation quality than deeper regions.”
Therefore, it is recommended to cut the brain into pieces prior to fixation to ensure optimal fixation. Additional fixation best practices can be found in the following KB articles:
- Do you have recommendations for FFPE tissue fixation for Xenium?
- Do you have tips and tricks for pre-fixation tissue handling for Xenium?
- How does delayed fixation impact Xenium performance?
After fixation, it is important to perform Tissue QC. Like fresh frozen sections, we recommend histological staining (e.g., H&E staining and DAPI) and DV200 to check for any artifacts. A DV200 of more than 30% is recommended. Even with blocks having a DV200 of 30%+, suboptimal performance is still possible. In general, DAPI is weaker in sub-optimally fixed tissues as outlined in the FFPE Tissue Preparation Guide tips and tricks and DAPI QC sections (FFPE Tissue Preparation).
Figure 1. DAPI images of FFPE hBreast, hColor, hLung, and hKidney imaged under the same exposure. hKidney shows washed out DAPI, indicating suboptimal tissue quality and most likely poor Xenium data.
Generally, poor DAPI staining in under or over-fixed brain tissue is correlated with very poor sample performance (no Tx/Cells).
Can over- or under fixed tissue be rescued?
Rescuing over-fixed or under-fixed tissue can be challenging and is unlikely to improve performance. This said, here are some considerations:
- Over-fixed tissues - Extended Decrosslinking Time: Extending the decrosslinking time (up to 2 hours from the 30 minutes outlined in the SOP) has been tested to try to rescue over-fixed datasets. However, it has been noted that this approach does not rescue transcripts significantly.
- Under-fixed tissues - Re-fixation: If RNA has already been degraded, there is no rescue as RNA is degraded and cellular integrity compromised. In addition, under-fixed samples often have much higher gDNA.
It is generally recommended to ensure proper fixation protocols are followed to avoid these issues.
Can decrosslinking temperature be modified to improve sensitivity?
For Xenium, we have tested modifying decrosslinking conditions across a large range of temperatures for a tissue array containing FFPE human brain samples. Results indicate that both negative control probes, gDNA, and transcript counts are sensitive to decrosslinking temperature. Increasing the decrosslinking temperature above 80°C would likely increase gDNA. In contrast, decreasing the decrosslinking temperature below 80°C is likely to lead to lower gDNA and negative controls, which may decrease transcript accessibility and result in lower transcripts per cell. Therefore, the recommendation for Xenium is to follow the Sample Preparation Demonstrated Protocol as written.
Cell Segmentation Staining Considerations
Oftentimes in brain datasets, the fraction of transcripts within cells is low, and there may be an analysis summary alert indicating this issue. Guidance on transcript rescue with third-party applications is outlined in this article.
Data Highlights
Here are two human brain preview datasets for Xenium v1:
- Human brain preview on hBrain panel (Xenium v1)(linked here)
- Human brain cancer preview on IO panel (Xenium v1)(linked here)
For Xenium Prime, we present data highlights from FFPE healthy human brain (Figure 2) and FFPE human glioma (Figure 3).
Figure 2. High level Xenium 5k metrics for FFPE human brain with the human multi panel.
Figure 3. High level Xenium 5k metrics for FFPE human glioma with the human multi panel.
Transcript count vary depending on the disease state. If we dig deeper in the Xenium Explorer, we can observe different cell types.
Figure 4. Cell clusters and transcript view of FFPE human glioma in Xenium Explorer.
Here we zoom in on astrocyte, neuronal, and oligodendrocyte markers.
Figure 5. Cell clusters and transcript view of FFPE healthy human brain in Xenium Explorer.
And here we zoom in on glial and neuronal markers.
Figure 6. Cell clusters and transcript view of FFPE healthy human brain in Xenium Explorer.
Products: Xenium In Situ, Xenium Prime In Situ Gene Expression, Multimodal Cell Segmentation