Integrin-Kindlin3 requirements for microglial motility in vivo are distinct from those for macrophages

Abstract

Microglia play a critical role in the development and homeostasis of the CNS. While mobilization of microglia is critical for a number of pathologies, understanding of the mechanisms of their migration in vivo is limited and often based on similarities to macrophages. Kindlin3 deficiency as well as Kindlin3 mutations of integrin-binding sites abolish both integrin inside-out and outside-in signaling in microglia, thereby resulting in severe deficiencies in cell adhesion, polarization, and migration in vitro, which are similar to the defects observed in macrophages. In contrast, while Kindlin3 mutations impaired macrophage mobilization in vivo, they had no effect either on the population of microglia in the CNS during development or on mobilization of microglia and subsequent microgliosis in a model of multiple sclerosis. At the same time, acute microglial response to laser-induced injury was impaired by the lack of Kindlin3-integrin interactions. Based on 2-photon imaging of microglia in the brain, Kindlin3 is required for elongation of microglial processes toward the injury site and formation of phagosomes in response to brain injury. Thus, while Kindlin3 deficiency in human subjects is not expected to diminish the presence of microglia within CNS, it might delay the recovery process after injury, thereby exacerbating its complications.

Keywords: Cell Biology; Neuroscience.

Figure 1. Kindlin3 is expressed in microglia in vitro and in vivo.

(A) Brain tissue from P9 Cx3cr1^GFP/+ mice was stained for Kindlin3. Representative confocal image from the cortical region showing expression of Kindlin3 (red), GFP (green), and overlay (right), with nuclei shown in blue. Scale bar: 10 μm. (B) Projection of z-stack images showing Kindlin3 in whole-mount retinas from P9 Cx3cr1^GFP/+ mice (Kindlin3, red; GFP, green; isolectin, blue). Scale bar: 100 μm. (C) 3D reconstitution of an image stack from P7 whole-mount retinas showing coexpression of CD11b (red) and Kindlin3 (green) within the same cells. Blood vessels are in blue (isolectin). Scale bar: 100 μm. (D) 3D reconstitution of a brain slice from the cortical region (left)showing lack of Kindlin3 expression (green) in GFAP-positive cells (red). Staining of cultured astrocytes and microglial cells (right) showing lack of Kindlin3 expression (green) in astrocytes (actin-rich cells predominantly stained with phalloidin, red). Nuclei are shown in blue. Scale bar: 10 μm. (E) P7 whole-mount retinas stained for Kindlin3 (green) and blood vessels (isolectin, red). Scale bar: 100 μm.

Figure 2. Impaired cell spreading and front-to-back cell polarity in Kindlin3 mutant microglia.

(A) Mixed astrocytes and microglia from WT and Kindlin3 mutant knockin (K3KI) mice were plated on fibronectin or laminin for 24 hours and stained for F4/80 (green) and filamentous actin (phalloidin, red). Images show impaired cell spreading of microglia from K3KI mice. Note that spreading of astrocytes is not affected by Kindlin3 mutation. Scale bar: 10 μm. (B) Phase-contrast image showing brain microglia from WT and K3KI mice. Microglial cells were isolated from mixed brain culture by removal of astrocytes with trypsin, as described in Methods. Kindlin3-deficient microglia adherent to astrocyte-deposited ECM display lack of lamellipodia, reduced cell spreading, and loss of front-to-back cell polarity. Scale bar: 10 μm. (C–E) Quantification of cell-spreading area, circularity, and longest axis in purified microglial cells (n = 15 per group from 3 independent experiments). ***P < 0.001, 1-tailed Student’s t test. Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars).

Figure 3. Kindlin3 excision in microglia leads to the loss of cell adhesion and spreading.

(A) Scheme illustrating generation of Kindlin3^lox/lox mice (top). Microglial cells isolated from Kindlin3^lox/lox mice were infected with either control adenovirus (bottom left) or Cre-encoding adenovirus (bottom right) to delete Kindlin3. Images show cell-spreading impairment in Cre-expressing microglia. Scale bar: 10 μm. (B and C) Quantification of longest axis and circularity in virus-infected microglial cells (n = 10 per group from 2 independent experiments). ***P < 0.001, 1-tailed Student’s t test. Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars).

Figure 4. Mn²⁺ activates integrins in Kindlin3-knockin microglia but fails to induce outside-in signaling.

(A) Microglial cells from WT and Kindlin3 mutant knockin (K3KI) mice were tested for ability to bind soluble fibrinogen (FG) with or without MnCl₂ (1 mM) as described in Methods. Flow cytometry analysis of FITC-conjugated FG binding shows a 4-fold increase in the percentage of cells positive for FG binding in both WT and K3KI mice after stimulation with Mn²⁺ (mean ± SEM from n = 3 independent experiments using a total of 8 mice per group). (B) Representative quantification of 9EG7 binding with and without 5-minute MnCl₂ stimulation (mean ± SEM from n = 5 cells per group of 4 independent experiments). (C) Analysis of cell-spreading area. Microglia from WT and K3KI mice were plated on fibronectin and incubated with C5a (10 ng/ml), fractalkine (200 ng/ml), or MnCl₂ (1 mM) for 50 minutes. Both chemokines and Mn²⁺ enhanced cell spreading in WT microglia but not in K3KI microglia (n = 10 cells per group from 3 independent experiments). (D) Microglial cells from WT and K3KI mice were plated on fibronectin for 2 hours and stained for active β1 integrin (clone 9EG7, green) and total β1 integrin (magenta). Scale bar: 5 μm. Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars). *P < 0.05, **P < 0.01, ***P < 0.001, 1-tailed Student’s t test.

Figure 5. Integrin activation fails to induce focal adhesion kinase phosphorylation in K3KI microglia.

(A) Staining for phospho-Y397 focal adhesion kinase (pY397FAK) and total FAK of microglia plated on fibronectin (FN) for 1 hour. Scale bar: 10 μm. (B) Quantification of pY397 FAK to total FAK ratio in microglia plated on FN that were either stimulated with Mn²⁺ or left untreated (No stim) (n = 30 from 2 experiments). Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars). **P < 0.01, ***P < 0.001, 1-tailed Student’s t test.

Figure 6. Kindlin3 is required for focal adhesion kinase recruitment and phosphorylation.

(A) Microglial cells stably adherent to fibronectin (FN) from WT and Kindlin3 mutant knockin (K3KI) mice stained for phospho-Y397 focal adhesion kinase (pY397FAK) (red) and total FAK (green). Scale bar: 10 μm. (B) Over 95% of the WT microglia are polarized (measured as the ratio of the longest axis of the cell to its shortest axis) in comparison to 10% of K3KI microglia (mean ± SEM from n = 6 independent experiments with 30 cells counted/group/experiment). ***P < 0.001, 1-tailed Student’s t test. (C) Average fluorescent intensities for WT and K3KI mice measured over a line profile for each antigen (as described in Methods), starting from the leading edge (pixel 1) (n = 3). In WT cells, FAK is predominantly phosphorylated on Y397 at the leading edge, but K3KI cells show diffused Y397 FAK phosphorylation.

Figure 7. Absence of phospho–myosin light chain polarization and Iba1-positive podosome formation in Kindlin3 mutant microglia.

(A) Phospho–myosin light chain (p-MLC) staining of stably adherent microglia. Representative line profiles (white lines labelled with their length in pixels) were used to measure fluorescent intensity from leading edge to trailing edge of cells. Scale bar: 10 μm. (B) Average fluorescent intensity for p-MLC staining from representative cells measured (n = 3) over a line profile for WT and K3KI mice from the leading edge (pixel 1) (as described in Methods). (C) Stably adherent microglial cells from WT and K3KI mice were stained for Iba1 6 hours after plating on fibronectin. Confocal images from the bottom cell surfaces show Iba1 localization to podosomal rings in WT cells but not in K3KI cells. Scale bar: 10 μm.

Figure 8. Kindlin3 is required for migration of microglial cells in vitro.

(A) Random migration of microglia isolated from WT and Kindlin3 mutant knockin (K3KI) mice observed using differential interference microscopy. Time-lapse image sequence showing actively migrating WT microglia but not K3KI microglia (for the whole video, see Supplemental Videos 1 and 2). Scale bar: 10 μm. (B) Velocity of randomly migrating microglia was quantified with ImageJ software (n = 10 cells per group from 2 experiments). (C) C5a-induced chemotaxis of microglia was analyzed in a Transwell assay, as described in Methods (n = 4–9 fields per group from 3 experiments). Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars). ***P < 0.001, 1-tailed Student’s t test.

Figure 9. Kindlin3 mutant knockin macrophages exhibit defective spreading and migration in vitro as well as impaired recruitment to inflammatory sites in vivo.

(A) 5 × 10⁴ peritoneal macrophages were seeded per well, precoated with 10 μg/ml fibronectin for 24 hours, and stained with hematoxylin. The images show impaired spreading and lack of polarity in the Kindlin3 mutant knockin (K3KI) macrophages. Scale bar: 10 μm. (B) The number of polarized cells (cells with a ratio of the longest axis to shortest axis of 2:1 or greater) in both WT and K3KI mice was counted and is represented as the percentage of spread cells from n = 5 independent experiments, with 100 cells counted/group/experiment. (C) The average polarity of the K3KI cells shown as a measurement of their longest axis relative to WT cells (n = 5; mean ± SEM). (D) Spreading of Kindlin3-deficient macrophages was rescued by transfection with GFP-Kindlin3 vector but not with the GFP vector alone. Scale bar: 10 μm. (E) Quantification showing a significant increase in the area of spreading by Kindlin3-deficient macrophages transfected with GFP-Kindlin3 vector (n = 22 cells per group from 3 experiments). (F) In vitro cell migration assessed by the Oris cell migration assay. The quantification of cell migration represented as fold change in migration area, showing a significant reduction in migration by K3KI macrophages in comparison to WT macrophages (mean ± SEM, n = 3 independent experiments). (G) Peritoneal macrophages from WT and K3KI mice were collected after 72 hours of thioglycollate injection and counted with a hematocytometer. The graph shows a significant reduction in the number of cells accumulated in peritoneum of K3KI mice compared with WT mice (mean ± SEM, n = 6 per group). Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars). **P < 0.01, ***P < 0.001, 1-tailed Student’s t test.

Figure 10. Kindlin3 is not required for colonization of the brain by microglial cells during development.

(A) Cross section of day 13.5 embryos showing comparable numbers of microglia present in WT and Kindlin3 mutant knockin (K3KI) brain and eye compartments. Embryos were stained with Iba1 for microglia. Scale bar: 100 μm. The graph shows the number of microglia per field (n = 6) from 3 mice per group (mean ± SEM). (B) Cortical microglia in adult WT and K3KI mice expressing CX3CR1-GFP. Projection from a stack of multiphoton images, 80–100 μm below the pia, acquired through the cranial window. Blood vessels are labeled in red by fluorescently conjugated dextran. Scale bar: 50 μm. The graph shows an insignificant difference in the number of microglia (mean ± SEM, n = 6 fields from 3 mice per group). n.s. indicates no significance using 1-tailed Student’s t test.

Figure 11. Kindlin3 mutant knockin microglia migrate and populate the retina, similar to WT microglia.

(A) Cross section of E13.5 retina expressing CX3CR1-GFP stained with DAPI (blue) and isolectin (red). Scale bar: 200 μm. No significant difference was observed in microglia number, as shown by graph (mean ± SEM, n = 6 fields from 4 mice per group). (B) Confocal images of p12 and p60 whole-mount retinas from WT and K3KI mice expressing CX3CR1-GFP. Scale bar: 100 μm. The graph (right) shows an insignificant difference in number of microglia between the two groups at p12 and p60 (mean ± SEM, n = 10 fields from 5 mice per group). n.s. indicates no significance using 1-tailed Student’s t test.

Figure 12. WT and Kindlin3 mutant knockin microglia show similar responses in a model of multiple sclerosis.

(A) Images of brain sections stained with myelin and DAPI (top) or images of CX3CR1-GFP–expressing microglia (bottom) show the even distribution of microglia in control brains of WT/Cx3cr1^GFP/+ and K3KI/Cx3cr1^GFP/+ mice and uneven accumulation in the region of corpus callosum of brains from mice fed with cuprizone for a period of 6 weeks. Scale bar: 100 μm. (B) The microglia distribution pattern in healthy brains (control) of WT/Cx3cr1^GFP/+ and K3KI/Cx3cr1^GFP/+ mice is similar, with one peak with 1 cell/field within 500 μM² area. (C) Cuprizone induces two distinct peaks of microglia distribution representing microglia-free and microglia clusters areas.

Figure 13. Normal migration and accumulation of Iba1-positive microglia of WT and K3KI mice in a multiple sclerosis model.

Iba1 was immunostained on brain sections from WT/Cx3cr1^GFP/+ and K3KI/Cx3cr1^GFP/+ mice after 6 weeks of normal or cuprizone-supplemented diet. Representative confocal images of the corpus callosum area of control and cuprizone-treated brains reveal no defects in migration and accumulation of Iba1-positive cells (magenta) in K3KI/Cx3cr1^GFP/+ mice compared with WT/Cx3cr1^GFP/+ mice. Scale bar: 20 μm.

Figure 14. Kindlin3 is involved in microglial response to laser-induced brain injury.

(A) Selected projections from a stack of multiphoton images taken 80–100 μm below the pia, acquired through the cranial window in vivo. Microglia projections labeled in green by GFP expressed under Cx3cr1 promoter coalesce around laser injury (center). Blood vessels are labeled in red by retro-orbital injection of fluorescently conjugated dextran. Numbers above images represent time in minutes. Scale bar: 20 μm. (B) Representative kymographs showing dynamics of microglial processes toward the ablation site (indicated by arrow). The x axis represents the distance (75-μm field); the y axis represents the time from top to bottom (90 minutes, starting at 45 seconds after the injury). (C) Velocity of microglial process toward the site of injury was determined from kymographs, as described in Methods (n = 8 kymographs per group from total 6 animals). Box-and-whisker plots show median (line within box), upper and lower quartiles (bounds of box), and minimum and maximum values (bars). ***P < 0.001, 1-tailed Student’s t test.

Figure 15. Kindlin3 is involved in microglial phagosome formation in proximity of injury scars.

(A) Formation of phagosomes between 2.5 hours and 9 hours after laser-induced injury. Arrowheads indicate individual phagosomes. Scale bar: 50 μm. (B) Tracking of a single of phagosome (white arrows) for 30 minutes in a representative cell. Scale bar: 10 μm. (C) The number of phagosomes at 6 hours after injury, quantified in Z-stack image (mean ± SEM; n = 6 fields per group from total 6 animals). ***P < 0.001, 1-tailed Student’s t test.