Profiling cellular heterogeneity in formalin-fixed paraffin-embedded (FFPE) tissues is key to characterizing clinical specimens for biomarkers, therapeutic targets, and drug responses. Recent advancements in single-nucleus RNA sequencing (snRNA-seq) techniques tailored for FFPE tissues have demonstrated their feasibility. However, isolation of high-quality nuclei from FFPE tissue with current methods remains challenging due to RNA cross-linking. We, therefore, proposed a novel strategy for the preparation of high-fidelity nuclei from FFPE samples, cryogenic enzymatic dissociation (CED) method, and performed snRandom-seq (snCED-seq) for polyformaldehyde (PFA) -fixed and FFPE brains to verify its applicability. The method is compatible with both PFA-based and FFPE brains or other organs with less hands-on time and lower reagent costs, and produced 10 times more nuclei than the homogenate method, without secondary degradation of RNA, and maximized the retention of RNA molecules within nuclei. snCED-seq shows 1.5-2 times gene and UMI numbers per nucleus, higher gene detection sensitivity and RNA coverage, and a minor rate of mitochondrial and ribosomal genes, compared with the nuclei from traditional method. The correlation gene expression of nucleus from the post-fixed and the frozen sample can be up to 94 %, and the gene expression of our nuclei was more abundant. Moreover, we applied snCED-seq to cellular heterogeneity study of the specimen on Alzheimer's Disease (AD) to demonstrate a pilot application. Scarce Cajal Retzius cells in older mice were robustly detected in our data, and we successfully identified two subpopulations of disease-associated in astrocytes, microglia and oligodendrocytes, respectively. Meanwhile, we found that most cell types are affected at the transcriptional level by AD pathology, and there is a disease susceptibility gene set that affects these cell types similarly. Our method provides powerful nuclei for snRNA-seq studies for FFPE specimens, and even helps to reveal multi-omics information of clinical samples.
