BCR-ABL1-driven exosome-miR130b-3p-mediated gap-junction Cx43 MSC intercellular communications imply therapies of leukemic subclonal evolution

Abstract

Rationale: In the bone marrow microenvironment (BMME), mesenchymal stem/stromal cells (MSCs) control the self-renewal of both healthy and cancerous hematopoietic stem/progenitor cells (HSPCs). We previously showed that in vivo leukemia-derived MSCs change neighbor MSCs into leukemia-permissive states and boost leukemia cell proliferation, survival, and chemotherapy resistance. But the mechanisms behind how the state changes are still not fully understood. Methods: Here, we took a reverse engineering approach to determine BCR-ABL1+ leukemia cells activated transcriptional factor C/EBPβ, resulting in miR130a/b-3p production. Then, we back-tracked from clinical specimen transcriptome sequencing to cell co-culture, molecular and cellular assays, flow cytometry, single-cell transcriptome, and transcriptional regulation to determine the molecular mechanisms of BCR-ABL1-driven exosome-miR130b-3p-mediated gap-junction Cx43 MSC intercellular communications. Results: BCR-ABL1-driven exosome-miR130a/b-3p mediated gap-junction Cx43 (a.k.a., GJA1) BMSC intercellular communications for subclonal evolution in leukemic microenvironment by targeting BMSCs-expressed HLAs, thereby potentially maintaining BMSCs with self-renewal properties and reduced BMSC immunogenicity. The Cx43^low and miR-130a/b^high subclonal MSCs subsets of differentiation state could be reversed to Cx43^high and miR-130a/b^low subclones of the higher stemness state in Cx43-overexpressed subclonal MSCs. Both miR-130a and miR-130b might only inhibit Cx43 translation or degrade Cx43 proteins and did not affect Cx43 mRNA stability. The subclonal evolution was further confirmed by single-cell transcriptome profiling of MSCs, which suggested that Cx43 regulated their stemness and played normal roles in immunomodulation antigen processing. Thus, upregulated miR-130a/b promoted osteogenesis and adipogenesis from BMSCs, thereby decreasing cancer progression. Our clinical data validated that the expression of many genes in human major histocompatibility was negatively associated with the stemness of MSCs, and several immune checkpoint proteins contributing to immune escape in tumors were overexpressed after either miR-130a or miR-130b overexpression, such as CD274, LAG3, PDCD1, and TNFRSF4. Not only did immune response-related cytokine-cytokine receptor interactions and PI3K-AKT pathways, including EGR3, TNFRSF1B, but also NDRG2 leukemic-associated inflammatory factors, such as IFNB1, CXCL1, CXCL10, and CCL7 manifest upon miR-130a/b overexpression. Either BCR siRNAs or ABL1 siRNAs assay showed significantly decreased miR-130a and miR-130b expression, and chromatin immunoprecipitation sequencing confirmed that the regulation of miR-130a and miR-130b expression is BCR-ABL1-dependent. BCR-ABL1 induces miR-130a/b expression through the upregulation of transcriptional factor C/EBPβ. C/EBPβ could bind directly to the promoter region of miR-130b-3p, not miR-130a-3p. BCR-ABL1-driven exosome-miR130a-3p could interact with Cx43, and further impact GJIC in TME. Conclusion: Our findings shed light on how leukemia BCR-ABL1-driven exosome-miR130b-3p could interact with gap-junction Cx43, and further impact GJIC in TME, implications for leukemic therapies of subclonal evolution.

Keywords: B progenitor acute lymphoblastic leukemia; BCR-ABL1; Cx43; MSCs; acute lymphoblastic leukemia; chemotherapy resistance; gap junctions; leukemic microenvironment; miR-130a and miR-130b; stemness.

Figure 1

miR-130a/b expression in BCR-ABL1 positive B-ALL patients. a RT-qPCR analysis of miR130a-3p, miR130a-5p, miR130b-3p and miR130b-5p in BM aspirates from B-ALL leukemia. BCR-ABL1 positive (n=19) or negative B-ALL (n=30). b Pairwise miRNA correlation analysis for these miRNA expressions from total B-ALL samples (n=49). c The relative expression of miRNAs in paired plasma and PBMC samples from BCR-ABL1 positive samples (n=12). d The relative expression of miR130a/b-3p in cells versus exosome from Sup-b15 and K562 cell lines. The expression level of miRNAs was normalized against U6. Fold changes were normalized by expression levels in exosomes. The miRNA expression level was normalized against U6. Data are shown as mean ± SD representing three biological replicates. *P < 0.05, **P < 0.01, ***P < 0.001, by student's paired t-test.

Figure 2

Exosomes derived from leukemia cells impaired GJIC of BMSCs. a Nanoparticle tracking analysis showed the size distribution of exosomes derived from Sup-B15 cells. b Exosomes were imaged by Electron microscopy, revealing the typical morphology and size. Scale bar=200 nm. c Western blotting analysis of TSG101, CD81 and Calnexin in exosomes from Sup-B15 cells. d Representative images of gap junction communication of BMSCs were assessed by measuring FRAP. Photobleached BMSC was marked with red arrows. e Statistical analysis of FRAP results for BMSCs treated by exosomes derived from HL60, K562, Ball-1, Jurkat, and Sup-B15, respectively. f, g showed representative images of western blotting. The expression levels of Cx43 in BMSCs treated by exosomes derived from multiple leukemia cell lines, respectively. h Fold changes were normalized by the blank group. I, j showed representative images of western blotting after co-culture treatment. The relative levels of Cx43 in BMSCs co-culturing with multiple leukemia cell lines, respectively. Mean ± SD representing more than three biological replicates. *P < 0.05, **P < 0.01, ***P < 0.001, by student's unpaired t-test.

Figure 3

Both miR-130a and miR-130b target Cx43 and impair the GJIC of BMSCs. a Bioinformatics pipeline prediction showing miR-130a/b targeting mRNA of Cx43. b-e Western blotting was applied to measure the relative protein levels of Cx43 in BMSCs after transfection with indicated or miR-130b-3p mimics or inhibitors for 48 hs (b and c). The results were quantified and normalized to NC groups and TUBULIN (d and e). mean ± SD representing 4 different donors for biological replicates. f The luciferase reporter vectors were constructed. Cx43-3' UTR WT or mutant sequences were inserted into psiCHECK2 plasmids, respectively. g Luciferase activity assay was performed and normalized to the luciferase activity of Cx43-3' UTR WT. h FRAP assays showed that miR-130a-3p and miR-130b-3p had an impact on the gap junction activity of BMSCs, respectively. Each dot represents one observed BMSC. *P < 0.05, **P < 0.01, ***P < 0.001, by student's unpaired t-test.

Figure 4

Both miR-130a and miR-130b inhibits osteogenic differentiation and promotes adipogenic differentiation of BMSCs. a, b Representative images of Alizarin Red S (a) or Oil Red O (b). Staining of mock and miRNA overexpressing BMSCs, respectively. BMSCs were transduced with the lentiviral vectors pLV-EGFP-miR-130a/b or pLV-EGFP-mock vectors, respectively. The positive BMSCs were purified by puromycin screening. After purification, BMSCs were induced for 14 or 21 days in an osteogenic or adipogenic medium, respectively. Scale bars=100um. The positive stainings were quantitively analyzed in 8 random areas in each biological replicate using image pro plus 6.0. All experiments were replicated more than 3 times independently. c, d Real-time PCR analysis of osteogenic (c) and adipogenic (d) expression levels after osteogenic or adipogenic differentiation of BMSCs for 14 or 21 days, respectively. e, f BMSCs were co-infected with Cx43 overexpressing or control lentivirus with the mixed pools of miR-130a/b lentivirus in osteogenic or adipogenic medium for 14 days or 21 days, respectively. Representative images of Alizarin Red S (e) or Oil Red O (f) staining were shown, respectively. Scale bars=100 um. The positive stainings were quantitively analyzed in 8 random areas in each biological replicate using image pro plus 6.0. All experiments were replicated more than 3 times independently. g, h The relative expression levels of osteogenic (g) or adipogenic (h) genes were measured by Realtime PCR at the indicated time points, respectively. The relative mRNA levels of these genes were normalized by NC group and internal GAPDH expression, and mean ± SD represents 3 independent replicates in all the Real-time PCR experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by student's unpaired t-test.

Figure 5

Single-cell transcriptome exploring the connections between MSC stemness and Cx43. a Heatmap of marker genes related to gap-junction in each BMSC subtype. b tSNE plot showing MSC subtypes. Each subtype was annotated according to the relative Cx43 expression, in addition to cycling BMSCs. c Pseudotime ordering of each MSC subtype. d Heatmap showing the active gene-regulatory networks across MSC subtypes predicted by the SCENIC package. e tSNE plot of HLA-related marker gene expression in BMSCs. f Dynamics of gene expression along the pseudotime. The bold line indicated mean expression across pseudotime.

Figure 6

miR-130a/b increased the immunosuppressive capacity of BMSCs. a Venn diagram of DEGs in miR-130a and miR-130b overexpressed BMSCs. b Principal-component analysis (PCA) plot on the transcriptome to visualize the similarity among all the samples. c, d Volcano plots showing representative DEGs related to immune responses in BMSCs after miR-130a or miR-130b overexpression, respectively. e Protein-protein interaction network of DEGs involved in KEGG immune-related pathways. f, g The indicated immune checkpoint genes (f) and inflammatory-related genes (g) are assessed by Realtime PCR in BMSCs transfected with miR-130a, or miR-130b mimics at 48 hs after transfection. The relative expression levels of these genes were normalized by the NC group and GAPDH expression. Mean ± SD represents 3 independent experiments. h, i BMSCs were transfected with miR-130a/b mimics or NC prior to co-culture with 1x10⁵ activated CIK (CD3⁺CD56⁺) in the indicated amounts. After 24 h and 48h of co-cultures, Representative flow cytometry plots of the apoptosis marker Annexin-V and 7-AAD staining of CIK (h), the apoptosis percentages of CIK cells were assessed by flow cytometry(n=3) (i).

Figure 7

BCR-ABL1 positively regulates the expression of miR-130a/b through transcription factor C/EBPβ. a-c Western blot analysis of the inhibition of BCR-ABL1 protein in Sup-B15 transfected with BCR siRNA (a), c-ABL siRNA (b) and treated with imatinib (c) (0.5uM and 1uM) for 72h, respectively. qRT-PCR analysis of miR-130a/b in Sup-B15 treated by BCR siRNA, c-ABL siRNA, and imatinib in Sup-B15. Mean ± SD represents 3 independent replicates in all the Real-time PCR experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by student's unpaired t-test. d Heatmap of ATAC-seq data suggesting a few differentially accessible chromatin regions between WT BMSCs and miR-130a or miR-130b overexpressing BMSCs. e Piechart showing the proportion of total peaks in the indicated regions.

Figure 8

BCR-ABL1 positively regulates the expression of miR-130a/b through transcription factor C/EBPβ. a-c Protein levels of C/EBPβ after inhibition of BCR-ABL1 expression by BCR siRNA (a), c-ABL siRNA (b), and imatinib (c) in Sup-B15 for 72h, respectively. d Transfection of C/EBPβ siRNA knocked down C/EBPβ expression in Sup-B15. The expression of miR-130a/b was assessed in Sup-B15 transfected with C/EBPβ siRNA. e Sup-B15 were transduced with lentiviruses expressing C/EBPβ or control (Ctrl); Western blotting analysis identified that C/EBPβ was overexpression in Sup-B15(n=3). The expression of miR-130a/b was assessed in Sup-B15 overexpressing C/EBPβ (n=4). f, g Visualization of C/EBPβ ChIP-seq peaks overlapping with miR-130b promoter region in Thp1 (f) and K562 (g). h, i CHIP-qPCR was further validated for specific regions for both miR-130a and miR-130b promoters. Mean ± SD represents 2 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by student's unpaired t-test.