GEO Accession viewer

NCBI > GEO > Accession Display

Not logged in | Login

GEO help: Mouse over screen elements for information.

Sample GSM7842416

Query DataSets for GSM7842416

Status

Public on Apr 05, 2024

Title

R.301_Day28_CARPos_GEX

Sample type

SRA

Source name

peripheral blood

Organism

Macaca mulatta

Characteristics

car: CARPos
day: Day28
timepoint: Contraction
library type: GEX
nhp: R.301
tissue: peripheral blood

Treatment protocol

NHP CD20 CAR-T cells were transduced and expanded as previously described in (Taraseviciute et al 2018).

Extracted molecule

total RNA

Extraction protocol

Peripheral blood mononuclear cells (PBMC) were isolated from adult NHP peripheral blood by Ficoll-PaquePLUS (GE Healthcare Bio-Sciences) or SepMate-50 (Stemcell). Total T cells were isolated from PBMCs using a T cell isolation kit per manufacturer’s instructions (Miltenyi Biotech). Polyclonal T cells were activated with Miltenyi NHP T cell Activation/Expansion kit (Miltenyi Biotec) in NHP media (X-Vivo 15 medium (Lonza) supplemented with 10% FBS (HyClone/Gibco), and recombinant human IL-2 (rhIL2,50 U/mL; R&D Systems/CellGenix)) with or without 1% penicillin/streptomycin/l-glutamine (Invitrogen), with or without 50 pmol/L p-mercaptoethanol (Sigma). Lentiviral transduction with spinoculation was performed 24-48 hours post stimulation. At the end of the stimulation cycle, stimulation beads were removed using Miltenyi LS columns (Miltenyi Biotec) and cryopreserved. Cryopreserved NHP cells were thawed, stained with fixable Live/Dead stain (Invitrogen), CD4 (BD bioscience, Cat# 563914, RRID:AB_2738485), CD8 (BD bioscience Cat# 563795, RRID:AB_2722501), CD20 (BD Bioscience, Cat# 566988, RRID:AB_2869992) and CD14 (Biolegend, Cat# 301834, RRID:AB_11126983) antibodies. The CAR was detected by staining with an anti-rituximab conjugated flow antibody (GeneTex, Cat# GTX43347, RRID:AB_11166924) in animal R.315 and with an anti-EGFR antibody (R&D systems, Cat# FAB9577G-100) for animals R.301-304. CD3 flow cytometric antibodies were not utilized, to avoid in vitro stimulation, which can result in alteration of the transcriptional profile. Live CARpos and CARneg, CD4/CD8pos and CD20/CD14neg cells were flow cytometrically sorted and immediately processed for single cell sequencing.
All products used in the preparation of these samples, excluding the NHP TCR specific custom primer set, are available from 10x Genomics. Samples in this publication were prepared using the Next GEM 5’ v2 Gel Beads with the i7 Multiplex Plate (Single Index) or the Next GEM Single Cell 5’ v2 Sample Index Plate TT (Dual Index). GEMs were generated using v2 Gel Beads and the v1 Target Enrichment kit utilizing the off-the-shelf human/mouse T Cell Mix 1 and 2 premixed primers for the human samples and substituting the provided primers with custom designed primers as previously described in (Gerdemann et al 2021) for NHP samples.

Library strategy

RNA-Seq

Library source

transcriptomic single cell

Library selection

cDNA

Instrument model

Illumina NovaSeq 6000

Description

10x Genomics
Remaining_NHPs.hdf5
Remaining_NHPs_scVI_model.pt

Data processing

Paired GEX and VDJ libraries were aligned using the ‘Cellranger multi’ pipeline (v7.0.0) with 32 CPU’s and 128 GB of RAM, with VDJ libraries listed as “vdj-t” to indicate T cell libraries. Samples were aggregated with the ‘Cellranger aggr’ pipeline (v7.0.0) with arguments ‘--normalize=none --nosecondary’. Our reference transcriptome was created with the Cellranger makeref command, and used genome assembly Mmul_10 as the reference genome and Ensembl v107 as the reference set of genomic annotations for the NHP samples, with CAR sequences added. For the scTCR-Seq libraries, we used the reference VDJ set created in (Gerdemann et al 2021) for NHPs
The “filtered_feature_bc_matrix.h5” file and “filtered_contig_annotations.csv” output files from Cellranger were combined into an AnnData object and loaded into scanpy. scTCR-Seq for R.301-R.304 was removed due to low quality. Since the libraries for R.315 and the libraries for R.301-R.304 were generated under different conditions, we decided to keep the two datasets separate rather than attempt to integrate them.
For the Initial NHP dataset, we retained cells that met the following criteria: <10% of the transcripts mapped to the mtDNA, genes detected > 900, and total transcripts between 2,500 and 25,000. For the validation dataset, we retained cells that met the following criteria: <10% of the transcripts mapped to the mtDNA, genes detected > 200, and total transcripts between 1000 and 25,000.
We then identified the 7,500 most variable genes in each of the two datasets and filtered the datasets to these genes along with CD3E, TRAC, CD4, CD8A, and the two CAR transcripts. We then applied PCA, and the neighbors, Leiden (clustering), and UMAP commands from Scanpy with default parameters on each dataset and retained clusters enriched for CD3E and TRAC, and removed those enriched for monocyte and other non-T cell markers.
We then used scvi-tools (v0.19.0) to fit a variational autoencoder (VAE) to each of the datasets. For each dataset, we trained the variational autoencoder on our raw transcript counts using the parameters n_layers=2, n_latent=30, gene_likelihood="nb". We found a small monocyte cluster in the initial NHP dataset, removed these cells, and fit a new VAE to the data.
We used the “get_latent_representation” function from scVI to get the latent space from the VAE for each dataset. We then applied the neighbors, UMAP, and Leiden functions from Scanpy with default settings to obtain clusters and the UMAP dimensionality reduction for the two datasets. We used the “get_normalized_expression” function from scVI with library size set to 10,000 to obtain normalized counts for the cells.
Assembly: Mmul10
Supplementary files format and content: “*.hdf5” files of filtered single-cell data that can be loaded into scanpy, “*.pt” files containing parameters for scVI VAEs .

Submission date

Oct 15, 2023

Last update date

Apr 05, 2024

Contact name

Jim Kaminski

E-mail(s)

James.Kaminski@childrens.harvard.edu

Organization name

Boston Children's Hospital

Street address

1 Blackfan Circle