Revised International Staging System (R-ISS) stage-dependent analysis uncovers oncogenes and potential immunotherapeutic targets in multiple myeloma (MM)

Article activity feed

1. 

   Version published to 10.7554/elife.75340 on eLife

   Oct 31, 2022
2. 

   eLife

   Sep 7, 2022

   Author Response

   Reviewer #2 (Public Review):

   Zhong et al conducted a scRNA-seq analysis to uncover the features in multiple myeloma (MM) based on the Revised International Staging System (R-ISS) stage. They contributed 11 scRNA-seq datasets, including 9 MM samples and 2 healthy BMMC. And validated their findings using the deconvolution method in large cohorts.

   In addition, the newly identified and validated a subset of GZMA+ cytotoxic multiple myeloma cells. The experiments were nicely conducted and the datasets generated in this study might benefit many other studies. Major comments:

   1. Several studies on scRNA-seq in MM have been reported, but different from that reported in this study. The authors might discuss the insight gained from their study.

   Thanks for your comments. Several studies on scRNA-seq in MM have been disclosed …

   Author Response

   Reviewer #2 (Public Review):

   Zhong et al conducted a scRNA-seq analysis to uncover the features in multiple myeloma (MM) based on the Revised International Staging System (R-ISS) stage. They contributed 11 scRNA-seq datasets, including 9 MM samples and 2 healthy BMMC. And validated their findings using the deconvolution method in large cohorts.

   In addition, the newly identified and validated a subset of GZMA+ cytotoxic multiple myeloma cells. The experiments were nicely conducted and the datasets generated in this study might benefit many other studies. Major comments:

   1. Several studies on scRNA-seq in MM have been reported, but different from that reported in this study. The authors might discuss the insight gained from their study.

   Thanks for your comments. Several studies on scRNA-seq in MM have been disclosed some heterogeneity of MM. For example, Jang JS et al identified the molecular pathways during MM progression (MGUS, SMM, NDMM, and RRMM) [Blood Cancer J. 2019 Jan 3;9(1):2.]. Jean Fan et al devised a computational approach called HoneyBADGER to identify copy number variation and loss of heterozygosity in individual cells from single-cell RNA-sequencing data [Genome Res. 2018 Aug; 28(8):1217-1227.]. These studies verified the high heterogeneities existed in MM. But the specific the mechanism was not clear. Furthermore, these studies didn’t specify the heterogeneity among different stages in R-ISS staging system, which has been an international wide used prognostic stratification system. Therefore, we focused on the specific cluster, marker, and cross-talk pattern among the three stages of MM to reveal the potential mechanism of heterogeneity.

   1. The author claimed Proliferating plasma cells were increased in EBV-positive MM patients. It would be interesting to examine the abundance of EBV RNA levels in the scRNA-seq datasets. Several tools, such as viral-track or PathogenTrack, might be used to conduct such analysis.

   Thanks for the reviewer’s great suggestions and comments. According to your suggestion, we used PathogenTrack to identify pathogens in MM patients and added this analysis results in the file ‘Data for reviewers-1(PathogenTrack).xlsx’. However, the algorithm did not identify EBV reads in the scRNA-seq datasets. In order to verify our conclusion, we collected more MM patients’ samples and examined EBV, MKI67, and PCNA. Our result showed that EBV positive samples had significantly higher MKI67 and PCNA expression, compared with EBV negative samples on Lines 193 to 195, Page 6 (in Figure 5B and 5C).

   1. Methods used for deconvolution are missing.

   We thank the reviewer’s comments and suggestions. In our study, we didn’t use an analytical tool named CIBERSORT, thus we didn’t use deconvolution either in the manuscript. It may cause you a misunderstanding because of our unclear description.

   Reviewer #3 (Public Review):

   The authors constructed a single-cell transcriptome atlas of bone marrow in normal and R-ISS-staged MM patients. A group of malignant PC populations with high proliferation capability (proliferating PCs) was identified. Some intercellular ligand receptors and potential immunotargets such as SIRPA-CD47 and TIGIT-NECTIN3 were discovered by cell-cell communication. A small set of GZMA+ cytotoxic PCs was reported and validated using public data.

   For scRNA-seq data analysis, the authors did QC and filtering and removed low quality cells, including some doublets and followed by batch effect correction. Malignant PC populations were identified using the copy number analysis tool "inferCNV".

   The authors have done lots of analysis. But I think the results can be improved if they can do more analyses. I would recommend to 1) analyze doublets; 2) remove cell cycle effect; 3) GO and pathway analysis for genes with copy number change; 4) do cell-cell communication with more cell type/clusters.

   Thanks for your suggestion and comment.

   1. We applied Scrublet to computationally infer and remove doublets in each sample individually, with an expected doublet rate of 0.06 and default parameters used otherwise. The doublet score threshold was set by visual inspection of the histogram in combination with automatic detection. Information about this description was added to material and methods section as ‘We applied Scrublet [74] to computationally infer and remove doublets in each sample individually, with an expected doublet rate of 0.06 and default parameters used otherwise. The doublet score threshold was set by visual inspection of the histogram in combination with automatic detection.’ accordingly in Lines 731-734, Page 27.

   2. As we focused on the differences in proliferative capacity of myeloma cells, the cell cycle could reflect the difference well. Therefore, the cell cycle data was provided accordingly. Information about this description was added into main text as ‘Next, we analysed the cell cycle of six PC clusters, and distinguished them from other clusters, PCs in cluster 6 (PCC6) were presumably enriched in G2/M stage (Figure. 3B)’ in Lines 142-144, Page 5.

   3. We have analyzed the GO and pathway analysis for genes with copy number changes, and provided the file ‘Data for reviewers-2 and 3 (InferCNV for PCC4 and PCC6)’. Based on this, we found that oxidative phosphorylation was the most significant enriched pathways for PCC4 and PCC6, respectively. Cell-cell communication with more cell type/clusters was provided with the supplementary data in the file ‘Data for reviewers-3 (Overall T cells interaction ligand-receptor pairs dotplot, Overall T cells interaction ligand-receptor, Overall T cells interaction map)’.

   Data analysis of public data was sufficient to prove the small set of GZMA+ cytotoxic PCs. More data analysis or wet experiment proof is required.

   Thanks for your suggestion. The subset of cytotoxic PCs was identified in this study. These PCs exhibited NKG7 and GZMA. Furthermore, NKG7 showed the higher expression level than NKG7. Therefore, we validated it using Multi-parameter Flow Cytometry (MFC) and Immunofluorescence in MM samples. We identified a new subset of NKG7+ cytotoxic PCs and found that the percentage of NKG7+ PCs displayed obvious diversities among stage I, II and III groups. Information about this description was added in the main text as ‘In another MM single-cell dataset focusing on PC heterogeneity of symptomatic and asymptomatic myeloma (dataset GSE117156) [19], one cluster, C21, exclusively expressing NKG7 corresponded to PC18 in our dataset (Fig 2C-2D). In GSE117156 of all 42 samples, the cell proportion varied from 0% to 30.95% of all PCs, with an average percentage of 4.28% (Figure. 2E).Next, immunofluorescence confirmed the expression of NKG7 in cytoplasm of PCs (CD138 positive) from patients with MM (Figure. 2F). Finally, twenty MM patients (stage I: three patients, stage II: 10 patients and stage III: seven patients) were enrolled for multi-parameter flow cytometric (MFC) analysis. The results showed that the percentage of NKG7+ PCs displayed obvious diversities among stage I, II and III groups (Figure. 2G and Figure. S2). The average percentage of NKG7+ population was 2.73% in stage I, 8.89% in stage II and 0.58% in stage III (Figure. 2G and Figure. S3). In summary, we characterized a NKG7+ PC population (PC18), which may provide a novel perspective for the cytotherapy of MM.’ in Figure 2 and S3 and Lines 118-130, Page 4-5.

   Read the original source
3. 

   eLife

   Sep 7, 2022

   Evaluation Summary:

   This paper is of potential interest to a broad audience across myeloma study and single cell technology, as it implies a major adjustment to our current understanding of pathogenesis and treatment of myeloma. Overall the data quality is good, although reasonable alternative explanations of the data can be identified.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)

   Read the original source
4. 

   eLife

   Sep 7, 2022

   Reviewer #1 (Public Review):

   This study analyzes the R-ISS-related plasma cell (PC) heterogeneity by 10X Genomics ScRNA sequencing and identified the two subsets of PCs(GZMA+ cytotoxic PCs and proliferating PCs). Three R-ISS-dependent gene modules in cytotoxic CD8+ T and NKT cells were also functionally analyzed. Potential immuno cell-cell communication such as SIRPA-CD47 and TIGIT-NECTIN3 were explored for the potential immunotargets which is an important direction for treating R/R MM. The work holds a promising way to study the drug resistance of R/R myeloma. However, the cost and complexity of the experimental method make it difficult to be widely used.

   Read the original source
5. 

   eLife

   Sep 7, 2022

   Reviewer #2 (Public Review):

   Zhong et al conducted a scRNA-seq analysis to uncover the features in multiple myeloma (MM) based on the Revised International Staging System (R-ISS) stage. They contributed 11 scRNA-seq datasets, including 9 MM samples and 2 healthy BMMC. And validated their findings using the deconvolution method in large cohorts. In addition, the newly identified and validated a subset of GZMA+ cytotoxic multiple myeloma cells. The experiments were nicely conducted and the datasets generated in this study might benefit many other studies.

   Major comments:

   1. Several studies on scRNA-seq in MM have been reported, but different from that reported in this study. The authors might discuss the insight gained from their study.

   2. The author claimed Proliferating plasma cells were increased in EBV-positive MM patients. It would …

   Reviewer #2 (Public Review):

   Zhong et al conducted a scRNA-seq analysis to uncover the features in multiple myeloma (MM) based on the Revised International Staging System (R-ISS) stage. They contributed 11 scRNA-seq datasets, including 9 MM samples and 2 healthy BMMC. And validated their findings using the deconvolution method in large cohorts. In addition, the newly identified and validated a subset of GZMA+ cytotoxic multiple myeloma cells. The experiments were nicely conducted and the datasets generated in this study might benefit many other studies.

   Major comments:

   1. Several studies on scRNA-seq in MM have been reported, but different from that reported in this study. The authors might discuss the insight gained from their study.

   2. The author claimed Proliferating plasma cells were increased in EBV-positive MM patients. It would be interesting to examine the abundance of EBV RNA levels in the scRNA-seq datasets. Several tools, such as viral-track or PathogenTrack, might be used to conduct such analysis.

   3. Methods used for deconvolution are missing.

   Read the original source
6. 

   eLife

   Sep 7, 2022

   Reviewer #3 (Public Review):

   The authors constructed a single-cell transcriptome atlas of bone marrow in normal and R-ISS-staged MM patients. A group of malignant PC populations with high proliferation capability (proliferating PCs) was identified. Some intercellular ligand receptors and potential immunotargets such as SIRPA-CD47 and TIGIT-NECTIN3 were discovered by cell-cell communication. A small set of GZMA+ cytotoxic PCs was reported and validated using public data.

   For scRNA-seq data analysis, the authors did QC and filtering and removed low quality cells, including some doublets and followed by batch effect correction. Malignant PC populations were identified using the copy number analysis tool - "inferCNV".

   The authors have done lots of analysis. But I think the results can be improved if they can do more analyses. I would …

   Reviewer #3 (Public Review):

   The authors constructed a single-cell transcriptome atlas of bone marrow in normal and R-ISS-staged MM patients. A group of malignant PC populations with high proliferation capability (proliferating PCs) was identified. Some intercellular ligand receptors and potential immunotargets such as SIRPA-CD47 and TIGIT-NECTIN3 were discovered by cell-cell communication. A small set of GZMA+ cytotoxic PCs was reported and validated using public data.

   For scRNA-seq data analysis, the authors did QC and filtering and removed low quality cells, including some doublets and followed by batch effect correction. Malignant PC populations were identified using the copy number analysis tool - "inferCNV".

   The authors have done lots of analysis. But I think the results can be improved if they can do more analyses. I would recommend to 1) analyze doublets; 2) remove cell cycle effect; 3) GO and pathway analysis for genes with copy number change; 4) do cell-cell communication with more cell type/clusters.

   Data analysis of public data was sufficient to prove the small set of GZMA+ cytotoxic PCs. More data analysis or wet experiment proof is required.

   Read the original source
7. 

   Version published to 10.1101/2021.12.06.471423v1 on bioRxiv

   Dec 7, 2021