Regulation of macrophage motility by Irgm1

Abstract

IRG are a family of IFN-regulated proteins that are critical for resistance to infection. Mouse IRG proteins are divided into GMS and GKS subfamilies, based on a sequence within the G1 GTP-binding motif. The GMS proteins have a particularly profound impact on immunity, as typified by Irgm1, of which absence leads to a complete loss of resistance to a variety of intracellular bacteria and protozoa. The underlying molecular and cellular mechanisms are not clear. Here, we use time-lapse microscopy and cell-tracking analysis to demonstrate that Irgm1 is required for motility of IFN-gamma-activated macrophages. The absence of Irgm1 led to decreased actin remodeling at the leading edge of migrating macrophages, as well as decreased Rac activation. Although Irgm1 did not localize to the leading edge of migrating macrophages, it was found to regulate the localization of a GKS IRG protein, Irgb6, which in turn, concentrated on the plasma membrane in the advancing lamellipodia, in close apposition to molecular components that regulate membrane remodeling, including Rac, paxillin, and actin. Thus, Irgm1 likely controls macrophage motility by regulating the positioning of specific GKS IRG proteins to the plasma membrane, which in turn, modulate cytoskeletal remodeling and membrane dynamics.

Figure 1.

Reduced motility of macrophages lacking Irgm1. WT and Irgm1 KO BMM were plated onto polylysine-coated glass coverslips, activated with 100 U/mL IFN-γ for 24 h, and exposed to CSF-1 to stimulate motility. The cells were then used for time-lapse video microscopy isolating images every 5 min. A 150-min time period was used for cell-tracking analysis, as detailed in the text. Shown are the average velocities, with error bars indicating sd, and a histogram stratifying the variation in velocities for each of three separate experiments. WT cells are represented by gray bars and Irgm1 KO cells by black bars. The differences in average velocities were statistically significant in each study, with P values <0.000005. In Experiment 1, 47 WT and 44 Irgm1 KO cells were tracked; in Experiment 2, 59 WT and 59 Irgm1 KO cells; and in Experiment 3, 177 WT and 179 Irgm1 KO.

Figure 2.

Decreased membrane ruffling in macrophages lacking Irgm1. IFN-γ-activated WT and Irgm1 KO BMM were exposed to CSF-1 or control conditions for 5 min and then used for staining with AlexaFluor-594-phalloidin to stain F-actin. The cells were analyzed in a blinded manner to score their degree of membrane ruffling using a 4-point index, with 4 representing the highest degree of ruffling. (A) Representative images of CSF-1-stimulated WT and Irgm1 KO BMM. The cells were originally magnified ×1000. Note the higher degree of ruffling at the ventral and dorsal surfaces of the WT cells. (B) The average ruffling indices for WT and Irgm1 KO BMM, with and without [control (con)] activation by CSF-1. The average indices were derived from three separate experiments with at least 60 cells analyzed per group in each experiment. The error bars represent sd based on the average results from the three experiments. Note that the difference between WT/CSF-1 and Irgm1 KO/CSF-1 was statistically significant (P=0.037).

Figure 3.

Decreased CSF-1-induced Rac activation in Irgm1 KO macrophages. BMM were maintained under control conditions or were exposed to IFN-γ for 24 h and then stimulated with CSF-1 for an additional 5 min. The cells were lysed and used for quantification of active Rac using an ELISA method described in the text that specifically measures GTP-bound Rac. Shown are (A) a representative study and (B) the CSF-1-stimulated increase in Rac activity averaged across four experiments. The error bars represent sd. Note that there is nearly equal basal Rac in activated WT and Irgm1 KO BMM under control conditions, and the CSF-1-stimulated increase in active Rac is reduced substantially in Irgm1 KO cells (P=0.0014). Abs, Absorbance.

Figure 4.

Lack of Irgm1 localization at the leading edge of motile macrophages. WT BMM were activated with IFN-γ for 24 h and then exposed to CSF-1 for 5 min to stimulate motility. The cells were then used for immunostaining with anti-Irgm1 antibodies, anti-Irgm3 antibodies, or AlexaFluor-594-phalloidin, as indicated. The cells were originally magnified ×1000. These images are representative of those from three independent experiments. Note that the cells lack discernable Irgm1 or Irgm3 staining along the plasma membrane in the advancing lamellipodia.

Figure 5.

Irgb6 and Irga6 localization at the leading edge of motile macrophages. WT BMM were activated with IFN-γ for 24 h and then exposed to CSF-1 for 5 min to stimulate motility. The cells were then used for immunostaining with anti-Irgb6 antibodies, anti-Irga6 antibodies, or AlexaFluor-594-phalloidin, as indicated. The cells were originally magnified ×1000. These images are representative of those from three independent experiments. Note that the cells display strong Irgb6 and Irga6 staining on the plasma membrane in the advancing lamellipodia, which is closely apposed to actin-stained areas.

Figure 6.

Irgb6 and Irga6 costaining with paxillin in motile macrophages. WT BMM were activated with IFN-γ for 24 h and then exposed to CSF-1 for 5 min to stimulate motility. The cells were then used for immunostaining with anti-Irgb6, Irga6, paxillin, and/or vinculin antibodies, as indicated. The cells were originally magnified ×1000. These images are representative of those from three independent experiments. Note that Irgb6- and Irga6-stained areas in the advancing lamellipodia are closely apposed or coincident with paxillin-stained areas.

Figure 7.

Loss of Irgb6, but not Irga6, at the leading edge of motile macrophages that lack Irgm1 expression. WT or Irgm1 KO BMM were activated with IFN-γ for 24 h and then exposed to CSF-1 for 5 min to stimulate motility. The cells were then used for immunostaining with (A) anti-Irgb6 antibodies, (B) anti-Irga6 antibodies, and/or AlexaFluor-594-phalloidin, as indicated. Exposure times were equal for all cells. The cells were originally magnified ×1000. These images are representative of those from three independent experiments. Note that in Irgm1 KO cells, there is a decrease in the intensity of actin staining and a substantial loss of Irgb6 staining on the plasma membrane in the advancing lamellipodia. In contrast, strong Irga6 staining is maintained in Irgm1 KO cells.

Figure 8.

Irgb6 and Rac localization in motile macrophages. WT BMM were activated with IFN-γ for 24 h and then exposed to CSF-1 for 5 min to stimulate motility. The cells were then used for immunostaining with anti-Irgb6 and Rac antibodies, as indicated. The cells were originally magnified ×1000. Note the closely apposed Irgb6 and Rac staining on the plasma membrane in the advancing lamellipodia, as well as coincident Irgb6 and Rac costaining in dorsal membrane ruffles.

Figure 9.

Hypothetical model for Irgm1 regulation of macrophage motility. Binding of CSF-1 to its receptor is known to initiate a signaling cascade that includes activation of paxillin and Rac and ultimately, increases actin remodeling at the leading edge of motile macrophages. In the current work, Irgm1 has been found to control the localization of Irgb6 and likely other GKS IRG proteins to the plasma membrane at the leading edge of migrating IFN-γ-activated macrophages. On the membrane, Irgb6 is in position to interact with Rac, paxillin, and other molecular components of the actin machinery. The presence of Irgb6 at these sites leads to increased Rac activation, actin remodeling, and cellular movement.