Liver metabolic function, dissecting one cell at a time Q&A

The Jupyter Notebook code for analysis is available for download here and in Zenodo. The data is now publicly available here and here.

To receive the mosquito scripts for cDNA synthesis and bead clean-ups, please get in touch with the SPT Labtech Field Applications Team.

In this case, we would recommend the the SMART-Seq v4 Kit.  

The SMART-Seq v4 Kit (SSv4) and the SMART-Seq v4 PLUS Kit (SSv4 PLUS) are the fourth generation of our SMARTer ultra-low solutions and our most sensitive kits for ultra-low inputs. These kits use oligo(dT) priming to generate high-quality, full-length cDNA directly from multiple intact cells (up to 1,000 cells) or 10 pg–10 ng of total RNA.

With the transcriptional profile alone, it is difficult to distinguish between 2n nuclei from mononucleated hepatocytes and 2n nuclei from tetraploid binucleated hepatocytes (2n x 2).

Technologies such as RNAScope (in situ hybridization assay for detection of target RNA within intact cells) can be used to distinguish this hepatocyte ploidy. Spatial transcriptomics approaches in the future will be very useful to address these questions genome-wide.

Taking advantage of the full-length transcript information that snRNA-seq2 offers, we are now investigating the potential splicing variant differences between hepatocyte nuclei.

In our lab we have used snRNA-seq2 method with liver, gut, brain and muscle tissue.

For the single nuclei isolation, the protocol described by Krishnaswami et al. was followed with several modifications.

In summary, fresh frozen liver tissues (~3 mm³) were homogenized and the final nuclei suspension was used for FACS sorting of 1 nuclei per well containing 0.9 µL final volume of the Takara Bio lysis buffer.

We have used the nuclei isolation protocol described in snRNA-seq2 with various tissues from mice, including lung and we obtain a clean suspension. Similarly, we have used it for human tissues, liver and bone marrow but not with human lung. We are positive that it would work although it might require additional optimization to increase the purity of the single-nuclei suspension.

We have not explored the differences on stellate cell ploidy status. We focused on hepatocytes in this study.

According to analysis via Gene Ontology between 2n and 4n hepatocytes, an enrichment of genes related to lipid and xenobiotic metabolism as well as oxidative stress pathways is oberserved in 4n hepatocytes. Other pathways can be explored and investigated under different treatments and conditions.

Starting from single-nuclei suspensions, we can only distinguish DNA content per nuclei using Hoescht and FACs sorting. We found no significant differences in the mRNA levels between 2n nuclei in diploid and tetraploid binucleated cells using RNAscope, and those populations are not able to be distinguished by snRNA-seq2 computationally.

Both fresh and frozen tissues can be used performing the with the snRNA-seq2 using the SMART-Seq Single Cell Kit. When possible, fresh tissue and whole cell approaches might be preferred, depending on the experimental design, dissociation methods, and accessibility to fresh samples.

There are several recent publications using lineage tracing strategies to investigate the regenerative potential of tetraploid hepatocytes (Chen F, et al., 2020; Sun T, et al. 2020; Matsumoto T, et al., 2020).

Their work and our work conclude that, both 2n and 4n can contribute to liver regeneration, although 4n cells need to divide into 2n first.

Although we have not used this methodology for single cell/nuclei RNAseq from laser microdissection, the addition of lysis buffer 2 is compatible with downstream SMARTer chemistry. As long as a strategy for placing the cell in wells with lysis buffer is in place, the cDNA generation using the presented pipeline should be straight-forward.

For our snRNA-seq2 approach, we use FACS sorting to address ploidy levels in our nuclei suspension and sort single nuclei into wells. The FACS sorting step though is not always necessary for downstream RNAseq, as long as the suspension is clean enough. For single-cell RNAseq based on well-plates, FACS sorting or other instruments for dispensing single-cells/nuclei into weeks will be needed.

Using snRNA-seq2, only the nuclear mRNA is captured, loosing partial information about the mature cytoplasmic mRNAs. The high correlation between nuclear and whole-cell transcriptome shows that the analysis of nuclear transcripts is a robust and reliable approach for cell-type identification.

We have not investigated gender bias in hepatocyte ploidy. It will certainly be an interesting question to explore.

Our BioRxiv manuscript was recently accepted in Nature Communications, and all data will be publicly available shortly.

In young C57Bl6/J, the expression levels of mononucleated and binucleated hepatocytes are conditioned by their position in the hepatic lobule (confirmed with RNAScope). The global expression profiles of 2n nuclei and 4n nuclei is very similar and they are not easily distinguished without previous knowledge on their genome content (e.g. by DAPI or Hoescht).

We have not tried our pipeline with RNAlater or other cryopreservation methods. We do not know if cryopreservation reagents would have a major impact in the in nuclear transriptome.

In our data, we are not able to detect high DNA contamination sequences. In the case of high DNA content, we can follow the manufacturer's suggestion to pre-treat the samples with DNAse I.

To acquire clean nuclei suspension, we use an OptiPrep gradient and inspect the quality under a microscope. For further improvements, washing and filtering steps can be used. To maintain integrity, the appropriate nuclei storage buffer needs to be used for each tissue and all manipulations needs to be done with caution regarding shearing forces.

We are now further investigating the changes in the transcriptome of human chronic liver diseases and whether the percentages of 2n to 4n change as well.