Tissue-resident macrophages promote early dissemination of multiple myeloma via IL-6 and TNFα

Abstract

Multiple myeloma (MM) is a plasma cell malignancy characterized by the presence of multiple foci in the skeleton. These distinct tumor foci represent cycles of tumor growth and dissemination that seed new clusters and drive disease progression. By using an intratibial Vk*MYC murine myeloma model, we found that CD169+ radiation-resistant tissue-resident macrophages (MPs) were critical for early dissemination of myeloma and disease progression. Depletion of these MPs had no effect on tumor proliferation, but it did reduce egress of myeloma from bone marrow (BM) and its spread to other bones. Depletion of MPs as a single therapy and in combination with BM transplantation improved overall survival. Dissemination of myeloma was correlated with an increased inflammatory signature in BM MPs. It was also correlated with the production of interleukin-6 (IL-6) and tumor necrosis factor α (TNFα) by tumor-associated MPs. Exogenous intravenous IL-6 and TNFα can trigger myeloma intravasation in the BM by increasing vascular permeability in the BM and by enhancing the motility of myeloma cells by reducing the adhesion of CD138. Moreover, mice that lacked IL-6 had defects in disseminating myeloma similar to those in MP-depleted recipients. Mice that were deficient in TNFα or TNFα receptor (TNFR) had defects in disseminating MM, and engraftment was also impaired. These effects on dissemination of myeloma required production of cytokines in the radiation-resistant compartment that contained these radiation-resistant BM MPs. Taken together, we propose that egress of myeloma cells from BM is regulated by localized inflammation in foci, driven in part by CD169+ MPs.

Graphical abstract

Figure 1.

Tissue -resident MPs promote dissemination of myeloma and disease progression. (A) Sample plots and analysis of MP (CD11b^lowF4/80⁺) frequencies in the BM in CD169-DTR mice before and after treatment with DT; data were pooled from total myeloid cells (CD11b⁺), monocytes (CD11b⁺Ly6C^hi), neutrophils (CD11b⁺Ly6G^hi), and MPs normalized to untreated controls. (B) Tumor burden analyzed in paired injected (inj; blue) and contralateral (CL; green) tibias of DT-treated CD169-DTR mice and control mice at 5 weeks after intratibial tumor inoculation and calculated dissemination index (ratio of myeloma burden in contralateral BM to that in injected BM). (C) Experiment setup as in panel B, with DT treatments starting at 2 weeks post tumor challenge (ptc). (D) Survival analysis (using Mantel-Cox test) of DT-treated or untreated CD169-DTR mice after intratibial tumor inoculation as in panel B. (E) Analysis of M-spike levels at 5 weeks and 20 weeks post tumor challenge in survival study shown in panel D. (F) Tumor burden was analyzed in DT-treated CD169-DTR (DTR+DT) mice and control (WT+DT) mice at 5 weeks after intravenous inoculation. (G) Analysis of tumor burden and dissemination index in chimeric hosts generated from lethal irradiation of recipient mice and reconstitution with donor BM cells, as labeled. Data from multiple experiments were pooled. All experiments were independently repeated 2 to 5 times, and each dot presents an individual mouse. Data comparisons were analyzed by using a Mann-Whitney t test. Error bars represent standard deviation. *P < .05; **P < .01; ***P < .001; ****P < .0001. n.s., not significant.

Figure 1.

Tissue -resident MPs promote dissemination of myeloma and disease progression. (A) Sample plots and analysis of MP (CD11b^lowF4/80⁺) frequencies in the BM in CD169-DTR mice before and after treatment with DT; data were pooled from total myeloid cells (CD11b⁺), monocytes (CD11b⁺Ly6C^hi), neutrophils (CD11b⁺Ly6G^hi), and MPs normalized to untreated controls. (B) Tumor burden analyzed in paired injected (inj; blue) and contralateral (CL; green) tibias of DT-treated CD169-DTR mice and control mice at 5 weeks after intratibial tumor inoculation and calculated dissemination index (ratio of myeloma burden in contralateral BM to that in injected BM). (C) Experiment setup as in panel B, with DT treatments starting at 2 weeks post tumor challenge (ptc). (D) Survival analysis (using Mantel-Cox test) of DT-treated or untreated CD169-DTR mice after intratibial tumor inoculation as in panel B. (E) Analysis of M-spike levels at 5 weeks and 20 weeks post tumor challenge in survival study shown in panel D. (F) Tumor burden was analyzed in DT-treated CD169-DTR (DTR+DT) mice and control (WT+DT) mice at 5 weeks after intravenous inoculation. (G) Analysis of tumor burden and dissemination index in chimeric hosts generated from lethal irradiation of recipient mice and reconstitution with donor BM cells, as labeled. Data from multiple experiments were pooled. All experiments were independently repeated 2 to 5 times, and each dot presents an individual mouse. Data comparisons were analyzed by using a Mann-Whitney t test. Error bars represent standard deviation. *P < .05; **P < .01; ***P < .001; ****P < .0001. n.s., not significant.

Figure 2.

TAMs exhibit proinflammatory phenotype in BM cluster of MM cells. (A-B) Intravital live imaging of GFP⁺ myeloma cells (green) in the injected tibia at 2 weeks after inoculation (n = 2 mice). (A) 3D analysis of distances between myeloma cells and anti-CD169-PE–labeled MPs (red) within a small cluster. (B) Examples of myeloma nanotube structures extending to BM autofluorescent (MP) cells (red). Yellow arrows show transfer of GFP signals from MM cells to surrounded microenvironment cells. (C) Gating strategy to identify GFP⁺ and GFP^– BM MPs and the CD206 subset and analysis of CD206^– (M1-like) subset of GFP⁺ and GFP^– BM MPs in tumor-bearing mice. (D) Kinetics of CD206^– MP subset as a function of tumor burden in the BM. (E) Samples of frequency of intracellular IL-6 and TNFα production in MPs, with fold increase in tumor-bearing mice normalized to that of naïve mice, analyzed by a Wilcoxon test. (F) Plot of TNFα-producing BM MPs vs tumor burden in the BM. All experiments were independently repeated at least 2 times; data were pooled and comparisons were analyzed by using a Mann-Whitney t test. Scale bars, 20 μm. *** is P < .001. Errors bars in panel C represent standard deviation. Horizontal lines in panels A and E are mean values.

Figure 2.

TAMs exhibit proinflammatory phenotype in BM cluster of MM cells. (A-B) Intravital live imaging of GFP⁺ myeloma cells (green) in the injected tibia at 2 weeks after inoculation (n = 2 mice). (A) 3D analysis of distances between myeloma cells and anti-CD169-PE–labeled MPs (red) within a small cluster. (B) Examples of myeloma nanotube structures extending to BM autofluorescent (MP) cells (red). Yellow arrows show transfer of GFP signals from MM cells to surrounded microenvironment cells. (C) Gating strategy to identify GFP⁺ and GFP^– BM MPs and the CD206 subset and analysis of CD206^– (M1-like) subset of GFP⁺ and GFP^– BM MPs in tumor-bearing mice. (D) Kinetics of CD206^– MP subset as a function of tumor burden in the BM. (E) Samples of frequency of intracellular IL-6 and TNFα production in MPs, with fold increase in tumor-bearing mice normalized to that of naïve mice, analyzed by a Wilcoxon test. (F) Plot of TNFα-producing BM MPs vs tumor burden in the BM. All experiments were independently repeated at least 2 times; data were pooled and comparisons were analyzed by using a Mann-Whitney t test. Scale bars, 20 μm. *** is P < .001. Errors bars in panel C represent standard deviation. Horizontal lines in panels A and E are mean values.

Figure 3.

Proinflammatory cytokines TNFα and IL-6 promote dissemination of MM. (A) Analysis of circulating tumor cells in tumor-bearing mice (4-6 weeks post tumor challenge), before (pre) and 2 hours (hr) after (post) treatment for 2 days with recombinant cytokines rTNFα or rIL-6. (B) Analysis of circulating tumor cells in tumor-bearing mice (4-6 weeks post tumor challenge) at 24 hours after treatment for 2 days with rTNFα or rIL-6. (C) Cell tracks of myeloma cells (n = 99) after time-lapse imaging in the BM focus before and 0 to 2 hours after treatment with rTNFα (1 µg) and analysis of their track speeds. (D-E) Analysis of tumor burdens in tibias and dissemination index in IL-6 and WT hosts after intratibial tumor inoculation. (F) Analysis of tumor burden in bones after intravenous (i.v.) tumor inoculation. (G-H) Comparison of tumor burdens (as in panel D) in TNFα^−/−, TNFR^−/−, and WT control mice. (I-J) Analysis of tumor burden and dissemination index in chimeric hosts, generated from lethal irradiation of recipient mice and reconstitution with donor BM cells as labeled. All experiments were independently repeated at least 2 times. Data points represent individual mice, and data from multiple experiments were pooled; comparisons were analyzed by using a Mann-Whitney t test. *P < .05; **P < .01; ***P < .001; ****P < .0001. Thick horizontal lines are means throughout; error bars are standard deviation in panels C and F and SEM in panel E.

Figure 3.

Proinflammatory cytokines TNFα and IL-6 promote dissemination of MM. (A) Analysis of circulating tumor cells in tumor-bearing mice (4-6 weeks post tumor challenge), before (pre) and 2 hours (hr) after (post) treatment for 2 days with recombinant cytokines rTNFα or rIL-6. (B) Analysis of circulating tumor cells in tumor-bearing mice (4-6 weeks post tumor challenge) at 24 hours after treatment for 2 days with rTNFα or rIL-6. (C) Cell tracks of myeloma cells (n = 99) after time-lapse imaging in the BM focus before and 0 to 2 hours after treatment with rTNFα (1 µg) and analysis of their track speeds. (D-E) Analysis of tumor burdens in tibias and dissemination index in IL-6 and WT hosts after intratibial tumor inoculation. (F) Analysis of tumor burden in bones after intravenous (i.v.) tumor inoculation. (G-H) Comparison of tumor burdens (as in panel D) in TNFα^−/−, TNFR^−/−, and WT control mice. (I-J) Analysis of tumor burden and dissemination index in chimeric hosts, generated from lethal irradiation of recipient mice and reconstitution with donor BM cells as labeled. All experiments were independently repeated at least 2 times. Data points represent individual mice, and data from multiple experiments were pooled; comparisons were analyzed by using a Mann-Whitney t test. *P < .05; **P < .01; ***P < .001; ****P < .0001. Thick horizontal lines are means throughout; error bars are standard deviation in panels C and F and SEM in panel E.

Figure 4.

Increased vascular leakage and reduced surface CD138 contribute to inflammation-enhanced dissemination. (A) Snapshots from intravital time-lapse images taken before, 2 minutes after, and 2 hours after intravenous administration of Texas Red dextran in the myeloma focus (yellow dashed outline) in the tibia. Insets show dextran leakage in regions within the focus (green) and outside the focus (cyan). (B) Flow cytometric analysis of anti-CD138-PE in vivo labeling of polyclonal normal PCs (nPCs) and myeloma cells (MM) in the tibia BM, with ex vivo co-staining with anti-CD138-APC. (C) Flow cytometric analysis of surface expression of CD138 on myeloma cells with and without in vitro treatment with rTNFα. (D) Comparison of surface expression of CD138 on myeloma cells in the BM of DT-treated or untreated CD169-DTR recipients. Scale bars represent 22 μm. All error bars are standard deviation, with horizontal bars reflecting the mean. *P < .05. APC, allophycocyanin; gMFI, geometric mean fluorescence intensity.

Figure 5.

Tissue-resident MPs contribute to MM relapse after irradiation therapy. (A) Experimental design and groups . (B) Analysis of systemic tumor burden based on serum M-spike 8 weeks after irradiation normalized to serum from naïve mice. Bars represent mean. (C) Survival curve calculated by using the log-rank test (n = 8-9 mice per group).

Figure 6.

TAMs have unique gene expression signatures. (A) Principal component analysis plot comparing transcriptomes of naïve MPs, TAM-ICs, and TAM-NICs. (B) Venn diagrams comparing DEGs between 3 MP subsets. (C) Unsupervised clustering of all DEGs from 3 pair-wise comparisons: TAM IC vs TAM NIC, TAM IC vs MP, and TAM NIC vs MP. (D) GO analysis (biological process terms) for gene clusters upregulated commonly or uniquely in the 2 TAM subsets vs MPs.

Figure 6.

TAMs have unique gene expression signatures. (A) Principal component analysis plot comparing transcriptomes of naïve MPs, TAM-ICs, and TAM-NICs. (B) Venn diagrams comparing DEGs between 3 MP subsets. (C) Unsupervised clustering of all DEGs from 3 pair-wise comparisons: TAM IC vs TAM NIC, TAM IC vs MP, and TAM NIC vs MP. (D) GO analysis (biological process terms) for gene clusters upregulated commonly or uniquely in the 2 TAM subsets vs MPs.

Figure 7.

TAMs and MM have unique cell-cell interactions . (A) Upregulated MP ligand-MM receptor pairs clustered by interaction scores. (B) GO-term analysis of the 3 clusters of ligand-receptor pairs in panel A. (C) Upregulated MM ligand-MP receptor pairs clustered by interaction scores. (D) GO-term analysis of the 3 clusters of ligand-receptor pairs in panel C. (E) Flow cytometric analysis of surface expression of CD324, CD71, and CD29 on MM cells (CD138^hiGFP^hi) vs PCs (CD138^hiB220^–GFP^–), and VCAM-1 expression on TAM-IC vs TAM-NIC cells using Mann-Whitney t tests (n = 3 mice) with similar results from a second experiment (not shown). *P < .05; **P < .01; ***P < .001; ****P < .0001. Error bars represent standard deviation.

Figure 7.

TAMs and MM have unique cell-cell interactions . (A) Upregulated MP ligand-MM receptor pairs clustered by interaction scores. (B) GO-term analysis of the 3 clusters of ligand-receptor pairs in panel A. (C) Upregulated MM ligand-MP receptor pairs clustered by interaction scores. (D) GO-term analysis of the 3 clusters of ligand-receptor pairs in panel C. (E) Flow cytometric analysis of surface expression of CD324, CD71, and CD29 on MM cells (CD138^hiGFP^hi) vs PCs (CD138^hiB220^–GFP^–), and VCAM-1 expression on TAM-IC vs TAM-NIC cells using Mann-Whitney t tests (n = 3 mice) with similar results from a second experiment (not shown). *P < .05; **P < .01; ***P < .001; ****P < .0001. Error bars represent standard deviation.