Study participants
Samples were acquired through the Stanford Brain Rejuvenation Program, the NIA funded Stanford Alzheimer’s Disease Research Center (ADRC), the University of California at San Francisco ADRC and the University of California at San Diego ADRC. Collection of CSF was approved by the Institutional Review Board of each university; written consent was obtained from all subjects. A total of 34 living subjects were used in this study, 24 of which were used for scRNA-TCRseq. The 24 subjects included 8 healthy controls, 4 patients with AD, 5 with MCI, and 7 with PD.
Cryopreservation of CSF cells
CSF was collected by lumbar puncture, then centrifuged at 300 rcf for 10 minutes at 4 ◦C to pellet immune cells. Importantly, CSF samples were checked for blood contamination by examining the pellet for the presence of red blood cells by eye. An example of a CSF sample contaminated with blood is shown in Supplemental Figure 1a. Note that cells should remain at 4 ◦C until they are further processed, but it is best to freeze the cells as quickly as possible to limit cell death. The supernatant (cell free CSF) was aliquoted, carefully leaving behind 100 μl of CSF with the pelleted cells. 100 μl of CSF was left so that cells were concentrated enough for counting and viability measurements. The pelleted cells were then gently resuspended in the 100 μl CSF and 10 μl of resuspended cells were then removed for counting. Importantly, cells were gently resuspended by first tapping the bottom of the tube and then gently triturating 10 times, making sure not to touch the pipette tip to the edge of the tube. Then 10 μl CSF was removed and mixed with 10 μl trypan blue to assess red blood cell content and viability. Cells were then visualized on a TC20 automated cell counter (BioRad) and cell number, viability and the presence or absence of red blood cells was recorded. CSF samples contaminated with blood were discarded. The resuspended cells were then mixed with 900 μl Recovery Cell Culture Freezing Medium (Thermo Fisher Scientific). This medium is an optimized version of the typical freezing medium, containing high-glucose Dulbecco's Modified Eagle Medium with 10% serum and 10% dimethyl sulfoxide. We utilized this medium because it is quality tested for pH, osmolality, sterility, and endotoxin and each lot is quality tested on CHO-K1 cells. The freezing medium was first thawed at 37 ◦C, aliquoted, and stored at -20 ◦C. Before use, the medium was thawed at 37 ◦C and kept on ice. After each aliquot was thawed, the freezing medium was stored at 4 ◦C for up to one month. All samples were frozen overnight at −80 ◦C in a Mr. Frosty freezing container (Thermo Fisher Scientific) and transferred the following day to liquid nitrogen for storage. CSF cells were stored in liquid nitrogen on average 266 days.
Preparation of frozen CSF cells for analysis
CSF cells were thawed at 37 ◦C in a water bath with the media submerged and the top of the tube out of the water. Cells were kept in the water bath for as little as possible and removed when the media was nearly completely thawed. The cells were then removed and gently pipetted into a 5 mL flow cytometry tube containing 3 mL pre-warmed (37 ◦C) sorting buffer (PBS with 0.04% bovine serum albumin (BSA)). The tube was then rinsed once with the sorting buffer and placed into the same flow cytometry tube. Cells were then centrifuged at 350 rcf for ten minutes. The supernatant was removed, and cells were resuspended in 500 μl sorting buffer. ½ μl of Sytox Red (Thermo Fisher Scientific) was added to the sample immediately before sorting by flow cytometry. Live cells were sorted into 1.5 mL Eppendorf tubes containing 750 uL sorting buffer. Once all the samples were sorted, cells were spun at 350 rcf at 4 ◦C in a spinning bucket centrifuge for 7 minutes. The supernatant was then removed, leaving behind 10 μl. 10 μl was left behind to resuspend the CSF cells and load the entire volume for droplet scRNA-seq.
Drop-seq of CSF cells
Chromium Single Cell 5’ Library & Gel Bead kit, Chromium Single Cell 5’ Library construction kit, Chromium Single Cell A Chip Kit, and Chromium i7 Multiplex kit (10X Genomics) were used for scRNAseq of CSF cells. We followed 10x Genomic’s User Guide for library construction. The only change we made to their protocol was in Step 1, GEM Generation & Barcoding. The user guide recommends loading a certain volume of cell suspension stock depending on the concentration of the cell suspension and the user’s desired cell recovery. However, because CSF contains such low cell numbers, we loaded all the cells that were resuspended in 10 μl of sorting buffer. We then added the 10 μl of cell suspension and 21.7 μl of nuclease free water, which results in the same total volume of cell suspension/water that the user guide recommends. After library construction, libraries were sequenced by Novogene on a Novoseq S4 sequencer and FASTQ files were generated by Novogene. Cell Ranger v.3.0.2 was used to generate gene-expression matrices for CSF cells. Reads from the 10X v.2 5′ paired library were mapped to the human genome build GRCh38 3.0.0. The 5′ gene-expression libraries were then analyzed with the Cell Ranger count pipeline and the resulting expression matrix was used for further analysis in the Seurat package v.3.0.
Seurat clustering of CSF cells
Individual sample expression matrices were loaded into R using the function Read10x under the Matrix package v.1.2-15. The expression matrix for each sample was merged into one Seurat object using the CreateSeuratObject and MergeSeurat functions. The Seurat package v.3.0 was used for filtering, variable gene selection, normalization, scaling, dimensionality reduction, clustering and visualization. Genes were excluded if they were expressed in fewer than 10 cells and cells were excluded if they expressed fewer than 200 genes. Cells that expressed more than 1,600 genes, more than 6,000 UMIs and more than 10% mitochondrial genes were excluded from the analysis. The sctransform normalization method was used to normalize, scale, select variable genes and regress out sequencing and experimental batch, mitochondrial mapping percentage, number of UMIs, and number of genes. After filtering and normalization, there were 26,797 cells and 14,953 genes. Following PCA, 5 principle components were selected for clustering tSNE dimensionality reduction.
Narcolepsy patient TCR sequences
All Narcolepsy patient TCR sequences were acquired from the Latorre D, et al 2018 study (15).
Calculation of Levenshtein similarities
L-sim scores were calculated using the levenshteinSim function in the RecordLinkage package for R (24). L-sim calculation incorporates the Levenshtein distance algorithm, which quantifies the number of edits, deletions, or insertions required to make two strings identical. L-sim includes a string length normalization and transformation of the final value to be between 0 and 1, with 1 representing identical TCR sequences:
TCR network visualization
TCR networks that show connections between similar TCRs organized by patient IDs and diagnosis groups were generated using the qgraph function in the qgraph package for R (25). Analysis from Figure 3b-d include only clonal TCRs with unambiguous alpha and beta chains. Supplementary Figures 4 and 5 include only clonal TCRs with unambiguous beta chains.
TCR motif analysis
A custom script was used to identify motifs in our TCR beta chains. Only clonal TCRs with unambiguous beta chains were included for analysis. Identified motifs were searched in the McPAS-TCR database, a manually curated database of TCR sequences found to be associated with pathological conditions in mice and humans (21).