Discoidin domain receptor 2 regulates neutrophil chemotaxis in 3D collagen matrices

Abstract

Neutrophils express a variety of collagen receptors at their surface, yet their functional significance remains unclear. Although integrins are essential for neutrophil adhesion and migration on 2-dimensional (2D) surfaces, neutrophils can compensate for the absence of integrins in 3-dimensional (3D) lattices. In contrast, we demonstrate that the inhibition of the tyrosine-kinase collagen receptor discoidin domain receptor 2 (DDR2) has no impact on human primary neutrophil migration on 2D surfaces but is an important regulator of neutrophil chemotaxis in 3D collagen matrices. In this context, we show that DDR2 activation specifically regulates the directional migration of neutrophils in chemoattractant gradients. We further demonstrate that DDR2 regulates directionality through its ability to increase secretion of metalloproteinases and local generation of collagen-derived chemotactic peptide gradients. Our findings highlight the importance of collagen-derived extracellular signaling during neutrophil chemotaxis in 3D matrices.

Figure 1

DDR2 is expressed and functional in human primary neutrophils. (A) DDR2 is phosphorylated upon activation by collagen I. Human primary neutrophils were cultured in the presence of collagen I (pretreated with or without rDDR). DDR2 was immunoprecipitated from cell lysates at different times of collagen treatment. Results are representative of 3 independent experiments. (B) DDR2 phosphorylation is inhibited by DDR2 blocking-antibodies. Human primary neutrophils were pretreated for 30 minutes with blocking DDR2 antibodies (50 μg/mL) and cultured in collagen I for 1 hour. DDR2 was immunoprecipitated from cell lysates. Results are representative of 3 independent experiments. (C) MMP secretion is amplified after DDR2 activation. Human primary neutrophils were embedded in collagen I (treated with or without rDDR) and stimulated with 500nM IL-8. The supernatant was collected and added to a solution of DQ collagen. The fluorescence generated by collagen hydrolysis was measured at 515 nm. The pan-MMP inhibitor GM6001 (25μM) was used as control. Results represent the data from 4 independent experiments. *P < .05; Friedman test; Dunn posthoc test. (D) MMP-8, but not MMP-9, secretion is amplified on DDR2 activation. Human primary neutrophils were embedded in collagen and stimulated with 500nM IL-8. DDR2 activity was inhibited either by collagen pretreatment with rDDR or with blocking DDR2 antibodies (pretreatment for 30 minutes). The levels or MMP-8 and MMP-9 in the supernatant were determined by ELISA. Results represent the data from 3 independent experiments. *P < .05; Friedman test; Dunn posthoc test.

Figure 2

DDR2 regulates neutrophil migration in 3D. (A) Neutrophil adhesion to a collagen-coated surface is not altered by rDDR treatment Human primary neutrophils were plated on collagen I–coated surfaces (treated with or without rDDR) for 30 minutes under IL-8 stimulation. Results represent the relative percentage of adhering cells (average ± SEM) of 3 independent experiments. (B) DDR inhibition does not affect neutrophil migration in 2D. Neutrophil migration to 500nM IL-8 on a collagen I–coated surface (pretreated with or without rDDR) was measured in a EZ-Taxiscan assay. Images were taken every 30 seconds for 30 minutes; 0-, 10-, and 20-minute images are presented. Also see supplemental Video 1. Results are representative of 3 independent experiments. Quantification is presented in supplemental Figure 2A. (C) rDDR treatment inhibits neutrophil migration through collagen-coated Transwells. Migration of human primary neutrophils to 500nM IL-8 or 1μM LTB₄ through collagen I–coated Transwells (pretreated of not with rDDR) was determined after 6 hours. Results represent the relative percentage of migrating cells after treatment (average ± SEM) of 4 independent experiments. *P < .05, Friedman test; Dunn posthoc test. (D) Neutrophil treatment with DDR2-blocking antibodies inhibits neutrophil migration through collagen-coated Transwells. Human primary neutrophils were pretreated for 30 minutes with blocking antibodies (or a control antibody). The number of neutrophils migrating to 500nM IL-8 through collagen I–coated Transwells (pretreated with or without rDDR) was determined after 6 hours. Results represent the relative percentage of migrating cells after treatment (average ± SEM) of 4 independent experiments. *P < .05, Friedman test; Dunn posthoc test.

Figure 3

DDR2 regulates neutrophil directionality in 3D collagen matrices. (A) rDDR treatment does not affect neutrophil chemotaxis in a 3D fibrin gel. Human primary neutrophils were embedded in fibrin (pretreated with or without rDDR). Migration to 50nM IL-8 was recorded (images were taken every 30 seconds with a 5X objective). Migrating cells were tracked for 20 minutes. The percentage, average speed, and CI of migrating neutrophils were determined. Results are representative of 5 independent experiments. (B) rDDR treatment inhibits neutrophil directionality in a 3D collagen matrix. Human primary neutrophils were embedded in collagen I (pretreated with rDDR or not). Migration of neutrophils to 50nM IL-8 was recorded as in panel A. Results are representative of 5 independent experiments. *P = .03, Wilcoxon test. (C) DDR2 inhibition with Ab1 treatment reduces neutrophil directionality in a 3D collagen matrix. Human primary neutrophils were pretreated with blocking DDR2 antibodies or a control IgG for 30 minutes and embedded in collagen I. Migration of neutrophils to 50nM IL-8 was recorded as in panel A. Results are representative of 5 independent experiments. *P = .03, Wilcoxon test.

Figure 4

MMP activity and PGP generation mediate the effects of DDR2 on neutrophil chemotaxis. (A) MMP and DDR2 inhibition affect neutrophil directionality in 3D collagen matrix. Human primary neutrophils were embedded in collagen I (pretreated with or without rDDR). Migration of neutrophils to 50nM IL-8 was recorded as described in Figure 3A. Results are representative of 5 independent experiments. *P < .05, Friedman test; Dunn posthoc test. (B) Neutrophil directionality in a collagen matrix is reduced upon DDR2 or MMP inhibition. The images show paths of individual cells migrating in a 3D collagen I matrix to IL-8 in the presence of the different inhibitors. Data are representative of 5 independent experiments. Also see supplemental Videos 2-4. (C) DDR2- and MMP-dependent migration is dependent on PGP production. Human primary neutrophils were pretreated for 30 minutes with 350 μg/mL RTR, 25μM GM6001, or 100μM Z-PP-CHO. The number of neutrophils migrating to 1μM LTB₄ (in the presence of the inhibitors) through collagen I–coated Transwells was determined after 6 hours. Results represent the relative percentage of migrating cells after treatment (average ± SEM) of 4 independent experiments. *P < .05, Friedman test; Dunn posthoc test.

Figure 5

DDR2 amplifies neutrophil migration at the single-cell level. (A) MMP regulates neutrophil persistence during chemokenesis in a 3D collagen matrix. Human primary neutrophils were pretreated for 30 minutes with 25μM GM6001 and embedded in collagen I. Migration of neutrophils after a uniform10nM IL-8 stimulation was recorded. The percentage, average speed, and persistence (see “Methods” for details) of migrating neutrophils were determined. Percentage and average speed are from 4 independent experiments. The persistence graph is representative of the 4 experiments. (B) rDDR inhibition is independent of cell density. Human primary neutrophils were embedded at different cell density in collagen I (pretreated with or without rDDR). Migration of neutrophils to 50nM IL-8 was recorded. Results are representative of 5 independent experiments. *P = .03, Wilcoxon test.

Figure 6

Model depicting DDR2 regulation of neutrophil chemotaxis in a 3D collagen matrix. In response to a gradient of IL-8, neutrophils embedded in collagen (gray lines) release low levels of MMPs (orange circles) and migrate with a low persistence. After DDR2 activation (depicted by the change from black to green), MMP release, mainly MMP-8, is increased resulting in substantial cleavage of collagen and generation of PGP-containing peptides, which are further processed into PGP by PE (blue halo). As a consequence, a local gradient of PGP is produced stabilizing neutrophil persistence and directionality.