Immunophenotypic correlates of sustained MRD negativity in patients with multiple myeloma

Abstract

The role of the immune microenvironment in maintaining disease remission in patients with multiple myeloma (MM) is not well understood. In this study, we comprehensively profile the immune system in patients with newly diagnosed MM receiving continuous lenalidomide maintenance therapy with the aim of discovering correlates of long-term treatment response. Leveraging single-cell RNA sequencing and T cell receptor β sequencing of the peripheral blood and CyTOF mass cytometry of the bone marrow, we longitudinally characterize the immune landscape in 23 patients before and one year after lenalidomide exposure. We compare patients achieving sustained minimal residual disease (MRD) negativity to patients who never achieved or were unable to maintain MRD negativity. We observe that the composition of the immune microenvironment in both the blood and the marrow varied substantially according to both MRD negative status and history of autologous stem cell transplant, supporting the hypothesis that the immune microenvironment influences the depth and duration of treatment response.

Fig. 1. Study design and treatment outcomes.

a Immune profiling was performed on blood and bone marrow collected from 23 patients with multiple myeloma receiving lenalidomide as maintenance therapy (created with BioRender.com). b Swimmer plot showing duration and depth of response to lenalidomide in relation to the sample collection time points and immune assays performed. c Kaplan–Meier survival analysis of progression-free survival from the start of the initiation of maintenance therapy for all patients and d stratified according to the minimal residual disease (MRD) negative status. PR partial response, VGPR very good partial response, CR complete response, TCR T-cell receptor, BMMC bone marrow mononuclear cells, PBMC peripheral blood mononuclear cells. Source data are provided as a Supplementary Data file.

Fig. 2. scRNAseq of PBMC reveals differences in immune cell frequency according to minimal residual disease (MRD)-negative status and transplant history.

a Aggregate UMAP and trajectory analysis of all cells analyzed color-coded by cell type and b presence of B-cell receptor (BCR) or T-cell receptor (TCR) by V(D)J sequencing. c Area plot showing the mean frequency of T cells, d B cells, and e myeloid cells at baseline and follow-up according to the patients’ MRD-negative status and transplant history. f Volcano plot comparing differentially abundant circulating immune cells between patients achieving sustained and non-sustained MRD negativity at baseline (left) and follow-up (right). g Volcano plot comparing differentially abundant circulating immune cells between patients with prior history of transplant or not at baseline (left) and follow-up (right). h Spider plot showing the percent change in cell frequency between baseline and follow-up time points according to the patients’ MRD-negative status (top row) and transplant history (bottom row). Only cell types that are significantly increased or decreased over time (Wilcoxon rank-sum test P < 0.05) are shown. N = 20 sustained MRD- (16 no prior HDM-ASCT, 4 prior HDM-ASCT) and 20 non-sustained MRD- (6 no prior HDM-ASCT, 14 prior HDM-ASCT) patient samples. Source data are provided as a Supplementary Data file. HDM-ASCT high-dose melphalan autologous stem cell transplant.

Fig. 3. Analysis of circulating T cells expressing markers of exhaustion and T-cell receptor repertoire diversity.

a Abundance of terminal effector memory T cells (T_EM) expressing known exhaustion markers in cancer. b The mean percentage of TOX + T_EM cells and standard error at baseline and follow-up according to minimal residual disease (MRD) negative status and transplant history. c Single-cell T-cell receptor mean Gini index and standard error showing the diversity of CD4⁺ and CD8⁺ T cells at baseline and follow-up according to MRD-negative status and transplant history. A higher Gini index corresponds to greater receptor diversity. d Circle packing plot showing individual T-cell phenotype (colored circles) in relation to clonotype (light-gray circles) and shared antigen specificity group (dark gray circles). Asterisks indicate a two-sided Wilcoxon rank-sum test P < 0.05. N = 20 sustained MRD- (16 no prior HDM-ASCT, 4 prior HDM-ASCT) and 20 non-sustained MRD- (6 no prior HDM-ASCT, 14 prior HDM-ASCT) patient samples. In the box plot, the ends of the whiskers represent 1.5 times the interquartile range, the center represents the median, and the bounds of the box represent the first and third quartiles. Source data are provided as a Supplementary Data file. HDM-ASCT high-dose melphalan autologous stem cell transplant.

Fig. 4. CyTOF of bone marrow reveals differences in immune cell frequency according to minimal residual disease (MRD)-negative status and transplant history.

a Heatmap showing the frequency of immune cell subsets per patient. b Area plot showing the mean frequency of T cells, c B cells, and d myeloid cells at baseline and follow-up according to the patients’ MRD-negative status and transplant history. e Volcano plot comparing differentially abundant bone marrow cells between patients achieving sustained and non-sustained MRD negativity at baseline (left) and follow-up (right). f Volcano plot comparing differentially abundant bone marrow cells between patients with prior history of transplant or not at baseline (left) and follow-up (right). g Spider plot showing the percent change in cell frequency between baseline and follow-up time points according to the patients’ transplant history. Only cell types that are significantly increased or decreased over time (Wilcoxon rank-sum test P < 0.05) are shown (there were no significant changes in cell frequency in relation to MRD-negative status). N = 7 sustained MRD- (5 no prior HDM-ASCT, 2 prior HDM-ASCT) and seven non-sustained MRD- (4 no prior HDM-ASCT, 3 prior HDM-ASCT) patient samples. Source data are provided as a Supplementary Data file. HDM-ASCT high-dose melphalan autologous stem cell transplant.

Fig. 5. T-cell receptor β (TCRβ) sequencing of blood reveals unique T-cell receptors to patients with sustained minimal residual disease (MRD) negativity.

a Box plots comparing the TCRβ diversity according to MRD and transplant status using the Gini index, where 1 equals the most clonal and 0 indicates the most diverse. b Spider plot showing the percent change in the Gini index across baseline and follow-up time points. c Heatmap showing the frequency of TCRβ sequences specific to patients with sustained MRD negativity (n = 74 sequences) or non-sustained MRD negativity (n = 23) that are not expected by chance (two-sided Fisher’s exact test <0.05). Pgen, progability of generating CDR3β sequence (Olga); Prev, prevalence of CDR3β sequence in healthy population (LymphoSeqDB). Two-sided Wilcoxon rank-sum test P values are shown. N = 6 sustained MRD- (4 no prior HDM-ASCT, 2 prior HDM-ASCT) and 18 non-sustained MRD- (6 no prior HDM-ASCT, 12 prior HDM-ASCT) patient samples. In the box plot, the ends of the whiskers represent 1.5 times the interquartile range, the center represents the median, and the bounds of the box represent the first and third quartiles. Source data are provided as a Supplementary Data file. HDM-ASCT high-dose melphalan autologous stem cell transplant.

Fig. 6. Comparison of immune cell frequency in the bone marrow and blood among healthy donors as well as patients with various stages of myeloma measured by scRNAseq.

a Line plot showing the mean and standard error of the cell frequency by cell type (plasmablasts excluded) in healthy bone marrow, precursor myeloma, untreated myeloma as well as blood from patients after receiving lenalidomide maintenance. b Heatmap and unsupervised hierarchical of cell frequency for each cell type averaged across subgroup. Samples are grouped according to K means clustering (k = 2). c Pairwise Pearson’s correlation matrix comparing the mean cell frequency of each of the measured immune cell subsets. Higher numbers (red) indicate greater similarity of samples derived from patients with the shown diagnosis. d Scatter plot and linear regression analysis showing correlation between immune cell frequency in healthy BMMC versus myeloma without prior HDM-ASCT achieving sustained MRD negativity. Line of best fit is shown in red and 95% confidence interval is indicated in gray. The linear regression P value and coefficient of determination are shown. e Summary of major differences observed between sustained and non-sustained MRD-negative patients (created with BioRender.com). N = 9 healthy donors, 5 patients with MGUS, 11 patients with SMM, and 7 patients with untreated MM, 20 patients with sustained MRD- (16 no prior HDM-ASCT, 4 prior HDM-ASCT), and 20 patients with non-sustained MRD- (6 no prior HDM-ASCT, 14 prior HDM-ASCT). Source data are provided as a Supplementary Data file. Non-sus non-sustained MRD negative, Sus sustained MRD negative, ASCT autologous stem cell transplant, BMMC bone marrow mononuclear cells, PBMC peripheral blood mononuclear cells.