Summary: Multiple myeloma (MM) is a malignant plasma cell disorder with well-defined clonal genetic/cytogenetic abnormalities. However, cellular heterogeneity is a key factor in MM progression, therapeutic decision, and response to treatment. Single cell whole transcriptome profiling (scRNA-Seq) offers an opportunity to dissect this molecular heterogeneity during MM progression to better understand the disease and guide rational therapy. Here, we examined 597 CD138 positive cells from 15 patients at different stages of MM progression using scRNA-Seq. We selected 790 genes based on a Coefficient of Variation (CV) approach which organized cells into four clusters (L1-L4) based on unsupervised clustering. Plasma cells from each patient contained a mixed population of plasma cells at different state of aggressiveness based on gene expression signature reflecting the inter-cellular heterogeneous nature of MM. Cells in the L1 group is characterized by low level expression of genes involved in the oxidative phosphorylation, Myc targets, and mTORC1 signaling pathway having most cells from MGUS patients (p < 1.2x10-14). In contrast, low level of these genes in L1 group increased progressively and were the highest in the L4 group containing only cells from high-risk MM patients with t(4;14) translocations. Furthermore, 44 genes consistently overexpressed by pair-wised comparisons of the four groups strongly associated with a reduced overall survival in MM patients (APEX trial, p < 0.0001; Hazard Ratio (HR), 1.83; 95% CI, 1.33 to 2.52), particularly those in the bortezomib treated group (p < 0.0001; HR, 2.00; 95% CI, 1.39 to 2.89). No survival significance was observed for the dexamethasone treated group. Our study at the resolution of single cells showed that there is a mixed population of cells in each patient at different stages of MM progression and these cells can be organized into four different subgroups (L1 to L4). Consistent overexpression of the 44 genes from L1 to L4 groups is associated with patient outcome and treatment response. Our results show that oxidative phosphorylation, Myc target, and mTORC1 signaling genes are significant pathways for MM progression and affect MM prognosis and treatment stratification.
Overall Design: 597 single cell libraries passed QC and were included in the downstream analysis
Strategy: |
|
Species: |
|
Tissue: |
|
Healthy Condition: |
|
Growth Protocol: | - |
Treatment Protocol: | - |
Extract Protocol: | Bone marrow aspirates were collected from 15 patients after informed consent and subjected to ACK lysis and mononuclear cell isolation. Plasma cells were separated by positive selection using CD138 coated magnetic beads (MACS; Miltenyi Biotec, CA) in a RoboSep system (STEMCELL Technology, Canada). CD138 positive cells were examined using Vi-CELL XR (Beckman Coulter, CA) to determine cell number, viability and average size. A microfluidic mRNA-Seq chip (Fluidigm, CA) was used for capturing cells from each sample at a concentration of 500ells/l and run in the Fluidigm C1 system to generate double stranded cDNA using SMARTer Ultra Low RNA kit for Illumina (Takara, CA). All samples were assessed for cell capture in a C1 chip by direct observation under a microscope and for cDNA quality using Fragment Analyzer (HS Large Fragment kit, Advanced Analytical Technologies, IA). |
Library Construction Protocol: | - |
Molecule Type: | poly(A)+ RNA |
Library Source: | |
Library Layout: | PAIRED |
Library Strand: | - |
Platform: | ILLUMINA |
Instrument Model: | Illumina HiSeq 2500 |
Strand-Specific: | Unspecific |
Data Resource | GEN Sample ID | GEN Dataset ID | Project ID | BioProject ID | Sample ID | Sample Name | BioSample ID | Sample Accession | Experiment Accession | Release Date | Submission Date | Update Date | Species | Race | Ethnicity | Age | Age Unit | Gender | Source Name | Tissue | Cell Type | Cell Subtype | Cell Line | Disease | Disease State | Development Stage | Mutation | Phenotype | Case Detail | Control Detail | Growth Protocol | Treatment Protocol | Extract Protocol | Library Construction Protocol | Molecule Type | Library Layout | Strand-Specific | Library Strand | Spike-In | Strategy | Platform | Instrument Model | Cell Number | Reads Number | Gbases | AvgSpotLen1 | AvgSpotLen2 | Uniq Mapping Rate | Multiple Mapping Rate | Coverage Rate |
---|