DC-SIGN binding to mannosylated B-cell receptors in follicular lymphoma down-modulates receptor signaling capacity

Abstract

In follicular lymphoma (FL), surface immunoglobulin (sIg) carries mandatory N-glycosylation sites in the variable regions, inserted during somatic hypermutation. These glycosylation sites are tumor-specific, indicating a critical function in FL. Added glycan unexpectedly terminates at high mannose (Mann) and confers capability for sIg-mediated interaction with local macrophage-expressed DC-SIGN lectin resulting in low-level activation of upstream B-cell receptor signaling responses. Here we show that despite being of low-level, DC-SIGN induces a similar downstream transcriptional response to anti-IgM in primary FL cells, characterized by activation of pathways associated with B-cell survival, proliferation and cell-cell communication. Lectin binding was also able to engage post-transcriptional receptor cross-talk pathways since, like anti-IgM, DC-SIGN down-modulated cell surface expression of CXCR4. Importantly, pre-exposure of a FL-derived cell line expressing sIgM-Mann or primary FL cells to DC-SIGN, which does not block anti-IgM binding, reversibly paralyzed the subsequent Ca²⁺ response to anti-IgM. These novel findings indicate that modulation of sIg function occurs in FL via lectin binding to acquired mannoses. The B-cell receptor alternative engagement described here provides two advantages to lymphoma cells: (i) activation of signaling, which, albeit of low-level, is sufficient to trigger canonical lymphoma-promoting responses, and (ii) protection from exogenous antigen by paralyzing anti-IgM-induced signaling. Blockade of this alternative engagement could offer a new therapeutic strategy.

Figure 1

Effect of DC-SIGN on gene expression in primary FL samples. (a–d) RNA-seq analysis in FL samples FL-B536, FL-LY86 and FL-LY221 stimulated with anti-IgM or DC-SIGN-Fc. (a) Volcano plots showing effect of anti-IgM (left) or DC-SIGN-Fc (right) on gene expression in anti-IgM/DC-SIGN-Fc-treated vs. non treated (NT) cells. Genes considered to be significantly upregulated or downregulated (log₂FC ≥ 1.0/ ≤ − 1.0, FDR Q ≤ 0.05) are colored red or blue, respectively. (b) Venn diagrams showing the degree of overlap between the anti-IgM and DC-SIGN-Fc upregulated (top panel) and downregulated (bottom panel) genes. (c) Graphs show effect of DC-SIGN-Fc on genes significantly (i) induced (log₂FC ≥ 1.0, FDR Q ≤ 0.05) or (ii) repressed (log₂FC ≤ 1.0, FDR Q ≤ 0.05) by anti-IgM, and of anti-IgM on genes significantly (i) induced (log₂FC ≥ 1.0, FDR Q ≤ 0.05) or (ii) repressed (log₂FC ≤ 1.0, FDR Q ≤ 0.05) by DC-SIGN-Fc. The black lines show regression line with 95% confidence interval (grey) fitted by linear regression between the log₂FC datasets. The blue dotted line is a line of equality (x = y). Results of Pearson’s correlation analysis are shown. (d) Gene set enrichment analysis (GSEA). Left panel shows NES scores for all gene sets for anti-IgM or DC-SIGN-Fc-treated FL samples, with gene sets with enrichment FDR Q < 0.01 colored as indicated. Right panel is an expanded view of the region containing positively correlated gene sets with enrichment FDR Q < 0.01 for both anti-IgM and DC-SIGN-Fc with identity of selected pathways detailed. Full details of GSEA outputs are provided in Supplementary file 2 online.

Figure 2

Regulation of CXCR4 expression in primary FL samples. (a, b) CXCR4 expression in CD10⁺/CD20⁺ cells of FL samples FL-B536, FL-LY86 and FL-LY221 at the start of the experiment, or after 6 h of treatment with anti-IgM or DC-SIGN-Fc, or left untreated (Ut) as a control. (a) Representative results obtained using sample FL-B536 and (b) summary for all samples (n = 3) analyzed. The graph shows mean (± SD) relative CXCR4 expression with values for control cells at 0 h set to 100%. The statistical significance of the indicated differences is shown (Paired t-test).

Figure 3

DC-SIGN binding and signaling in WSU-FSCCL cells. (a) Binding of DC-SIGN-Fc or DC-SIGN-HA to WSU-FSCCL and Ramos cells. Cells with no addition were analyzed as a control. Representative of 5 independent experiments. (b) WSU-FSCCL cells were incubated with DC-SIGN-Fc, DC-SIGN-HA, anti-IgM or left untreated as a control for 30 min at 4 °C and then warmed rapidly to 37 °C for the indicated times. Expression of total and phosphorylated AKT, ERK1/2 and GAPDH (loading control) was analyzed by immunoblotting. The gels were run under the same experimental conditions. Molecular weight of proteins markers shown as KDa. Results shown are representative of 3 independent experiments. Full-length blots are shown in Supplementary Fig. 3. (c) Representative iCa²⁺ flux analysis after addition (arrow) of anti-IgM, DC-SIGN-Fc, DC-SIGN-HA or control antibody (control ab) and summary of data for 5 separate experiments. Graph shows mean (± SD) response with values for anti-IgM-treated cells set to 100%. The statistical significance of the indicated differences is shown (paired t-test) (d) Confocal imaging of bound DC-SIGN-Fc (green) and sIgM (red) on fixed WSU-FSCCL cells. Cells were incubated with DC-SIGN-Fc at 4 °C for 1 h and anti-IgM was added for the last 30 min. DAPI staining is shown in blue. Two representative cells shown. (e) Anti-IgM accessibility with DC-SIGN bound to BCR. FACS analysis of WSU-FSCCL cells incubated with DC-SIGN-Fc (left), DC-SIGN-HA (right) or no addition for 30 min at 4 °C and stained with FITC anti-IgM or left unstained as control.

Figure 4

Effect of DC-SIGN pre-exposure on anti-IgM-induced signaling in WSU-FSCCL cells. (a, b) Anti-IgM-induced iCa²⁺ flux in WSU-FSCCL cells following pre-exposure to DC-SIGN-Fc/HA. (a) Representative results for DC-SIGN-Fc (left) and DC-SIGN-HA (right) pre-treatment follow by addition (arrow) of the stimulation as stated and (b) summary of data mean for 6 (DC-SIGN-Fc) or 3 (DC-SIGN-HA) individual experiments. Graphs show mean (± SD) response with values for anti-IgM only cells set to 100%. Statistical significance of the indicated differences is shown (paired t-test). (c) DC-SIGN displacement after mannan incubation. FACS analysis of WSU-FSCCL cells left untreated (grey), incubated with DC-SIGN-Fc 30 min at 37 °C (blue) or incubated with DC-SIGN-Fc 30 min at 37 °C, washed and left 15 min with 40 μg/ml of Mannan at 37 °C (red). Cells were stained with APC Cy-7 F(ab’)₂ anti-hIgG Fc (α-IgG Fc) to detect DC-SIGN-Fc binding. (d, e) Effect of pre-treatment with DC-SIGN-Fc ± mannan for 15 or 30 min at 37 °C on iCa²⁺ flux. (d) Representative result and (e) summary of data mean for 3 individual experiments. Graphs show mean (± SD) response with values for anti-IgM only/no pretreatment cells set to 100%. Statistical significance of the indicated differences is shown (paired t-test).

Figure 5

Effect of DC-SIGN pre-exposure on anti-IgM-induced endocytosis. (a) WSU-FSCCL cells incubated with AF488-conjugated anti-IgM (5 μg/ml) at 4 or 37 °C or were left untreated for 30 min. Then cells were incubated with anti-AF488 antibody (30 min at 4 °C to quench AF488 remaining on the surface), or left untreated. Graph on the left shows cells with no pre-treatment, whereas graphs in the middle and right show results for cells that were pre-treated for 30 min at 4 °C with DC-SIGN-Fc or DC-SIGN-HA respectively, prior to anti-IgM stimulation.

Figure 6

Effect of DC-SIGN on anti-IgM signaling in primary FL samples. (a, b) Anti-IgM-induced iCa²⁺ flux in primary FL samples following pre-treatment with DC-SIGN-HA. (a) Representative results of FL-B536 pre-treated with DC-SIGN-HA follow by addition (arrow) of the stimulation as stated and (b) summary of data for 4 samples. Graphs show mean (± SD) response with values for anti-IgM only cells set to 100%. Statistical significance of the indicated differences is shown (paired t-test). (c, d) Immunoblot analysis showing effect of pre-treatment with DC-SIGN-HA on total and phosphorylated LYN and SYK expression in anti-IgM-stimulated FL cells. Cells were pre-exposed to DC-SIGN-HA for 30 min at 4 °C and then stimulated for 30 s with control antibody, DC-SIGN-HA or anti-IgM at 37 °C. GAPDH expression was analyzed as loading control. (c) Representative results (FL-B536). Gels were run under the same experimental conditions and molecular weight of proteins markers shown as KDa. Full-length blots are shown in Supplementary Fig. 5. (d) Summary of data for 4 samples. Graphs show mean (± SD) response with values for no pre-treat (ie anti-IgM treated only) cells set to 100%. The statistical significance of the indicated differences is shown (paired t-test).