Single Nuclei Multiomic Profiling of Transcriptional and Chromatin Accessibility of Tumor Cells Underlines Extensive Cis-Regulatory Interaction during Multiple Myeloma Progression

Introduction

Multiple myeloma (MM) is a complex hematological disease with high heterogeneity resulting from abnormal proliferation of plasma cells in bone marrow. The progression of the disease from Monoclonal Gammopathy of Undetermined Significance (MGUS), Smoldering Multiple Myeloma (SMM), to MM involves multiple drivers and, moderately, by dynamic interactions of cis-regulatory elements in premalignant/malignant cells. Using single-nuclei multiome technology, the same-cell readouts of nucleus RNA sequencing (snRNA-seq) and Assay for Transposase-Accessible Chromatin sequencing (snATAC-seq) provide a less biased measurement of gene expressions and chromatin accessible regions. Profiling and analyzing both omics data at different stages of MM may reveal essential molecular signatures and regulatory elements for disease progression. These insights could benefit MM patients as novel targets for early therapeutic intervention at different stages of progression.

Methods

Bone marrow aspirates were collected from 33 individuals: Healthy donors (HD; n=3), patients with MGUS (n=2), SMM (n=5), MM (n=21), and Plasma Cell Leukemia (PCL; n=2). Viable mononuclear cells underwent CD138+ sorting and nuclei isolation prior to library preparation. Libraries were prepared with the Chromium Multiome kit (10X Genomics) and sequenced on NovaSeq S4 flow cells at UAMS. The clonal V(D)J arrangements of each sample and expression level from HD samples were used to distinguish malignant (MM) from normal plasma cells (PC). Data of the paired modalities were integrated using Signac with the weighted nearest neighbor (WNN) analysis. All inter- and intra-tumor clusters were evaluated for markers of molecular subtype, differential gene expression, and differentially accessible regions. A variance component analysis of proximal, distal chromatin accessibility, and gene expression, was estimated by the average information restricted likelihood estimation.

Results

We performed the single nuclei multiomic sequencing to generate multi-stage chromatin accessibility and transcriptional profiles of plasma cell neoplasms comprising 439,134 cells across 33 samples. The malignant plasma cells for each sample were extracted by the unique clonotypic signature of the immunoglobulin gene rearrangement. Using our previous microarray data, we detected the overexpression of molecular signatures for most cells in each molecular subgroup, confirming the molecular heterogeneity in MM. The multiomics integration generated less distinct clusters between MGUS and SMM, indicating the gradual complex alterations of gene expression and chromatin accessibility in malignant transformation. We then applied a variance component model to quantify the relative contribution of accessibility to transcriptional variance globally and corrected for inter-sample and disease stage effects. The results showed that more than 70% of gene expression variance was attributed to the accessibility of promoter and enhancer. However, there were 76 genes for which >60% of the gene expression variance could be explained by the disease stage covariance. Gene enrichment analysis revealed enrichment in the cell developmental process and differentiation, suggesting the increase of disease aggressiveness during the progression. Also, we have identified domains of regulatory chromatin (DORCs), the large clusters of regions of significant peak-gene links overlapped with super-enhancers that might play critical regulatory roles in disease progression.

Conclusion

In this large multiomics study of CD138+ cells, we demonstrate the cis-regulatory associations between chromatin accessibility and target genes from patients with early-stage disease to highly aggressive form. Throughout the disease progression, we found that transcriptional heterogeneity is predominantly associated with the variation of chromatin accessibility. Also, we uncover unique linked peak-gene pairs for each disease stage that can further refine our understanding of cis regulation patterns in multiple myeloma. Collectively, our joint single nuclei profiling reveals additional insights into cis-regulatory interactions and investigates novel regulatory domains involved in disease progression.