What should I consider when processing Healthy Human Brain FFPE samples on Visium HD?

Question: What should I consider when processing Healthy Human Brain FFPE samples on Visium HD?

Answer: Healthy human brain tissues are challenging samples to prepare due to considerable variability in the sample handling as well as sample processing steps including the ischemic and post-mortem interval time. The challenges faced with healthy human brain sample handling and processing may lead to unpredictable FFPE block quality. Please see the associated KBA, What best practices should be considered when preparing FFPE samples for the Visium HD Spatial Gene Expression assay?, for optional recommendations on determining your tissue’s quality, including DV200 measurement as well as H&E and DAPI morphological assessment. 

While in house we've had good results with healthy human brain FFPE samples using protocol CG000684 (Visium HD FFPE Tissue Preparation Handbook), we have seen variability across blocks, with some showing low sensitivity. These blocks yielding lower sensitivity tended to be of lower sample quality. In some cases, we found that a modified protocol was able to deliver moderately increased sensitivity for these blocks. This modified Visium HD protocol includes increasing the decrosslinking temperature to 95°C for 30 minutes as outlined below and may be a way to modestly increase sensitivity if you experience lower than desired sensitivity for a healthy human brain sample with Visium HD. 

Modifying the decrosslinking temperature to 95°C may result in increased probe access to genomic DNA sequences (% UMIs from Genomic DNA). It has been observed that in sensitivity troubleshooting cases, when healthy human brain FFPE samples are run on Visium HD with 95°C decrosslinking, the modified protocol can result in % gDNA that is higher than the SOP Visium HD conditions, but similar to Visium V2 performance. Please review Technical Note, CG000605 (Visium CytAssist Spatial Gene Expression for FFPE: Robust Data Analysis with Minimal Impact of Genomic DNA) for more information on expected gDNA impact. 

The example data below showcases two scenarios where an increase in sensitivity between decrosslinking temperatures was observed but leads to variable % gDNA between tissue blocks.

Figure 1 exemplifies a favorable sensitivity and % gDNA tradeoff in a case utilizing proximal sections from the healthy human brain FFPE Block 1. Increasing the decrosslinking temperature from 80°C to 95°C resulted in ~2-fold increased sensitivity (15.7 mean UMIs per 8 µm bin increased to 34.2 mean UMIs per 8 µm bin, respectively), while % gDNA remains within an expected range for Visium V2 and HD performance (1.0% gDNA UMIs to 3.0% gDNA UMIs, respectively). Additionally, it is observed that Loupe-derived graph-based clustering better resolves the anatomical regions in the 95°C section compared to the 80°C section. This refinement may be attributed to the moderate boost in sensitivity provided by the decrosslinking modification for this sample.

Figure 2 shows proximal sections from the healthy human brain FFPE Block 2, where increasing the decrosslinking temperature from 80°C to 95°C resulted in ~1.5-fold increased sensitivity (56.8 mean UMIs per 8 µm bin increased to 84.5 mean UMIs per 8 µm bin, respectively); however, the % gDNA was increased by 8-fold (1.6% gDNA UMIs to 12.3% gDNA UMIs, respectively). The increased sensitivity from 95°C decrosslinking on Block 2 in figure 2 could be attributed to the % UMIs from gDNA, rather than increased access to RNA, which may not contribute to biological interpretations (no significant differences were observed in the Loupe-derived graph-based clustering between the 80°C and 95°C sections). There does not appear to be a favorable sensitivity - % gDNA tradeoff to run the modified protocol on Block 2.

Example Data:

Proximal sections from two Human Brain Normal FFPE blocks (Block 1 and Block 2) were processed through the Visium HD Spatial Gene Expression assay workflow with and without a modification to the decrosslinking temperature.

Figure 1. Human Brain Normal FFPE Block 1
80°C Decrosslinking Temperature	95°C Decrosslinking Temperature
15.7 Mean UMIs per 8 µm Bin	34.2 Mean UMIs per 8 µm Bin
1.0% Estimated UMIs from gDNA	3.0% Estimated UMIs from gDNA

Figure 1. Proximal sections from the healthy human brain FFPE Block 1 were run with 80°C and 95°C decrosslinking conditions. Increasing the decrosslinking temperature from 80°C to 95°C resulted in ~2-fold increased sensitivity (15.7 mean UMIs per 8 µm bin increased to 34.2 mean UMIs per 8 µm bin, respectively). The % gDNA remains in an expected range for Visium V2 and HD performance (1.0% gDNA UMIs to 3.0% gDNA UMIs, respectively). Loupe-derived graph-based clustering better resolves anatomical regions in the 95°C Block 1 section compared to the 80°C Block 1 section.

Figure 2. Human Brain Normal FFPE Block 2
80°C Decrosslinking Temperature	95°C Decrosslinking Temperature
56.8 Mean UMIs per 8 µm Bin	84.5 Mean UMIs per 8 µm Bin
1.6% Estimated UMIs from gDNA	12.3% Estimated UMIs from gDNA

Figure 2. Proximal sections from the healthy human brain FFPE Block 2 were run with 80°C to 95°C decrosslinking conditions. Increasing the decrosslinking temperature from 80°C to 95°C resulted in ~1.5-fold increased sensitivity (56.8 mean UMIs per 8 µm bin increased to 84.5 mean UMIs per 8 µm bin, respectively). The % gDNA was increased by 8-fold (1.6% gDNA UMIs to 12.3% gDNA UMIs, respectively). No significant differences in Loupe-derived graph-based clustering were observed between the 80°C and 95°C Block 2 sections.

Low sensitivity, healthy human brain FFPE Visium HD protocol modification:

Our internal testing demonstrates that 95°C decrosslinking on healthy human brain FFPE samples can improve sensitivity. If you determine that your healthy human brain FFPE block may benefit from sensitivity troubleshooting, we suggest the protocol modification outlined below. 

Please assess your healthy human brain sample quality as outlined in this KBA, and keep in mind that the increased decrosslinking temperature run can result in a % gDNA tradeoff that is not recommended for Visium HD data analysis. Therefore, the CG000684 protocol (80°C decrosslinking for 30 minutes) should be followed for cases when healthy human brain FFPE samples do not require sensitivity troubleshooting. However, if you believe your healthy human brain FFPE samples may have lower sensitivity due to tissue quality (have low DV200, are over/under-fixed, stored improperly, have poor tissue morphology), you may want to consider starting with this modified protocol.

Modified Step 4.2 Decrosslinking (CG000684 Visium HD FFPE Handbook):    

A. Prepare the thermal cycler with the following incubation protocol and start the program. Note decrosslinking temp is modified to 95°C.

B. Please refer back to CG000684 (Visium HD FFPE Tissue Preparation Handbook) and proceed with step 4.2B.

If you have any questions regarding the protocol modification for low quality, healthy human brain FFPE samples, please contact support@10xgenomics.com. 

Product: Visium HD