Immunoregulatory roles of versican proteolysis in the myeloma microenvironment

Abstract

Myeloma immunosurveillance remains incompletely understood. We have demonstrated proteolytic processing of the matrix proteoglycan, versican (VCAN), in myeloma tumors. Whereas intact VCAN exerts tolerogenic activities through Toll-like receptor 2 (TLR2) binding, the immunoregulatory consequences of VCAN proteolysis remain unknown. Here we show that human myeloma tumors displaying CD8(+) infiltration/aggregates underwent VCAN proteolysis at a site predicted to generate a glycosaminoglycan-bereft N-terminal fragment, versikine Myeloma-associated macrophages (MAMs), rather than tumor cells, chiefly produced V1-VCAN, the precursor to versikine, whereas stromal cell-derived ADAMTS1 was the most robustly expressed VCAN-degrading protease. Purified versikine induced early expression of inflammatory cytokines interleukin 1β (IL-1β) and IL-6 by human myeloma marrow-derived MAMs. We show that versikine signals through pathways both dependent and independent of Tpl2 kinase, a key regulator of nuclear factor κB1-mediated MAPK activation in macrophages. Unlike intact VCAN, versikine-induced Il-6 production was partially independent of Tlr2. In a model of macrophage-myeloma cell crosstalk, versikine induced components of "T-cell inflammation," including IRF8-dependent type I interferon transcriptional signatures and T-cell chemoattractant CCL2. Thus the interplay between stromal cells and myeloid cells in the myeloma microenvironment generates versikine, a novel bioactive damage-associated molecular pattern that may facilitate immune sensing of myeloma tumors and modulate the tolerogenic consequences of intact VCAN accumulation. Therapeutic versikine administration may potentiate T-cell-activating immunotherapies.

Figure 1

Versikine, a novel DAMP with immunoregulatory roles in the myeloma microenvironment. (A-B) Freshly explanted MAM from the human myeloma marrow cases indicated were exposed to 1 μM versikine (Vkine) for 12 hours. Relative expression of IL-1β and IL-6 transcripts is shown. Black bars, vehicle; gray bars, versikine-treated; NS, not significant; veh, vehicle. (C-D) Wild-type (WT) and Tpl2^−/− BMDMs were treated with 1 μM versikine, and cytokine concentrations were measured in the culture supernatant at 12 hours postexposure. (C) Il-1β (left), Il-6 (right); before Il-1β assay, cells were treated with 5 mM ATP for 20 minutes. (D) Il-10 (left); Il-12p40 (right). (E) Versikine modulates macrophage polarization. BMDMs were exposed to vehicle, versikine alone, or versikine + OVA/anti-OVA immune complexes (IC), as previously described. Versikine exposure resulted in M1-like phenotype (Il-12^hi, Il-10^lo) in the absence of concurrent Fcγ ligation. Versikine + IC promoted macrophage polarization toward an M2b-like, immunoregulatory phenotype (Il-12^lo, Il-10^hi). (F) WT and Tlr2^−/− BMDM were stimulated by versikine for 12 hours and Il-6 protein was measured in the supernatant. (G) Signaling mediators induced by versikine stimulation of WT and Tpl2^−/− BMDMs. BMDMs were collected following stimulation with versikine at designed time-points (each number reflects minutes) and subjected to immunoblot analysis with the antibodies shown. (H) RNA-seq analysis of MM1.S myeloma cells exposed to versikine-transduced macrophages for 48 hours. Only 23 genes were differentially expressed and all were upregulated. Thirteen of 23 upregulated genes were ISGs (highlighted in gray). VCAN gene transcription changes are underlined. (I-J) Myeloma cell-macrophage cocultures were exposed to 0.5 μM purified versikine for 4, 18, or 48 hours. Representative ISG transcription is shown for THP-1 (I) and MM1.S cells (J). Relative mRNA transcription is normalized to vehicle-only control at each time point. *P < .05, **P < .01, ***P < .001, ****P < .0001.

Figure 2

Versikine acts through IRF8 to promote transcription of ISGs. (A) IRF transcription in MM1.S cells, following 48 hours of coculture with THP-1 macrophages in the presence of 0.5 μM versikine (Vkine) or vehicle (veh). Expression is normalized to veh-only levels at 4 hours. (B) IRF9 mRNA levels in MM1.S cells cocultured with THP-1 macrophages in the presence of 0.5 μM versikine (gray bars) or vehicle (black bars) for indicated time lengths. (C) THP-1 cells expressing Vkine or an empty-vector (EV) control were transduced with control lentivirus (NT) or lentivirus expressing short hairpin (shRNA) targeting IRF8 (shIRF8) (see supplemental Figure 3 for validation of IRF8 knockdown at the protein level). Versikine-mediated induction of 3 ISGs shown was measured in the presence and absence of IRF8. (D) EBI3 transcription in MM1.S cells cocultured with THP-1 macrophages and treated with 0.5 μM versikine (gray bars) or vehicle (black bars) for indicated time lengths. (E) RT-PCR analysis for EBI3 transcripts in patient-derived, freshly explanted MAM treated with 1 μM versikine for 12 hours. Relative expression is normalized to vehicle-only control (= 1). (F) Staining of human myeloma bone marrow core biopsy consecutive sections with antibodies against neoepitope DPEAAE generated by V1-VCAN cleavage at Glu⁴⁴¹-Ala⁴⁴² and T-cell marker CD8. DPEAAE constitutes the C terminus of versikine. Four patterns of staining were observed in 19 informative punches: Pattern (a): intense/moderate VCAN proteolysis-CD8 infiltration/aggregates (>5 CD8⁺ cells in cluster). Pattern (b): intense/moderate VCAN proteolysis-CD8 poor (single/doublet CD8⁺ cells sparsely distributed within tumor). Pattern (c): weak/focal VCAN proteolysis-CD8 poor (single/doublet CD8⁺ cells sparsely distributed within tumor). Pattern (d): absent VCAN proteolysis-CD8 poor (single/doublet CD8⁺ cells sparsely distributed within tumor). (G) Proposed immunomodulatory roles of VCAN proteolysis in the myeloma microenvironment. Whereas intact VCAN is thought to exert tolerogenic activities through TLR2 binding on antigen-presenting cells, its proteolytic product, versikine, may promote immunosurveillance through IRF8-mediated effects on antigen-presenting cells and tumor cells. Currently untested hypotheses are represented by broken lines. *P < .05, **P < .01, ***P < .001, ****P < .0001.