Progranulin expression is upregulated after spinal contusion in mice

Abstract

Progranulin (proepithelin) is a pleiotropic growth-factor associated with inflammation and wound repair in peripheral tissues. It also has been implicated in the response to acute traumatic brain injury as well as to chronic neurodegenerative diseases. To determine whether changes in progranulin expression also accompany acute spinal cord injury, C57BL/6 mice were subjected to mid-thoracic (T9 level) contusion spinal cord injury and analyzed by immunohistochemical and biochemical methods. Whereas spinal cord sections prepared from non-injured laminectomy control animals contained low basal levels of progranulin immunoreactivity in gray matter, sections from injured animals contained intense immunoreactivity throughout the injury epicenter that peaked 7-14 days post injury. Progranulin immunoreactivity colocalized with myeloid cell markers CD11b and CD68, indicating that expression increased primarily in activated microglia and macrophages. Immunoblot analysis confirmed that progranulin protein levels rose after injury. On the basis of quantitative polymerase chain reaction analysis, increased protein levels resulted from a tenfold rise in progranulin transcripts. These data demonstrate that progranulin is dramatically induced in myeloid cells after experimental spinal cord injury and is positioned appropriately both spatially and temporally to influence recovery after injury.

Fig. 1. Anti-GRN antibody detects full-length GRN

(a) Whole-cell lysates (Lys) and conditioned media (CM) were prepared from HEK293 cells transfected with expression vector pcDNA3.1 alone (GRN −) or pcDNA3.1-GRN (GRN +). Samples were then incubated in the presence or absence of Peptide N-glycosidase F (PNGase F) followed by immunoblot analysis (10% acrylamide gel) with polyclonal anti-mouse GRN antibody (1:1000). The resultant staining pattern was calibrated with prestained molecular mass standards (shown in units of kDa on left). Antibody reactivity was detected only in lysates and conditioned media prepared from cells transfected with pcDNA3.1-GRN, and was retained after deglycosylation with Peptide N-glycosidase F. (b) Recombinant mouse GRN (5.0 µl conditioned media) was incubated (37°C) for up to 18 h in the presence of human neutrophil elastase (0.3, 1, 3, or 10 U/ml), then subjected to immunoblot analysis (10 – 20% gradient acrylamide gel) with anti-GRN antibody. Control reactions included C1 (no elastase, incubation at 0°C); C2 (elastase was added to the reaction mixture and sample was boiled immediately); and C3 (no elastase, incubation at 37°C for 60 min). Elastase-mediated cleavage of GRN greatly attenuated antibody reactivity relative to control reactions.

Fig. 2. GRN immunoreactivity increases after contusion SCI

Transverse sections prepared from a laminectomy control (a, b) and injured (c–h) C57BL/6 mice were subjected to immunohistochemistry using anti-GRN antibody and chromogenic substrate diaminobenzidine. Labeled sections are shown at the level of the injury epicenter at low magnification in panels a, c, e, g (scale bar = 250 µm), with the boxed regions shown at 4-fold higher magnification in panels b, d, f, h (scale bar = 125 µm). In the absence of contusion injury (a, b), basal GRN immunoreactivity was diffuse and appeared primarily in gray matter. After contusion injury (c – h), GRN immunoreactivity greatly increased with time at the lesion epicenter, and by 14 dpi occupied most of the cross-sectional area. At later time points (g, h), after formation of the glial scar (arrowhead in g), GRN labeling was mostly confined to the lesion core with modest labeling in the spared peripheral white matter. (i) The proportional area (PA) of each tissue section occupied by GRN immunoreactivity was quantified ± SD and plotted as a function of time (three sections/animal; three animals/time point). Proportional area increased significantly after injury and was maximal between 7 and 14 dpi (**, p < 0.01; ***, p < 0.001 compared to the laminectomy control group; ANOVA with Dunnett’s post hoc test).

Fig. 3. Post-injury GRN expression increases at the protein level

(a) Levels of GRN protein in homogenates of injured and laminectomy control spinal cords were assessed by immunoblot analysis (10 – 20% gradient acrylamide gel) with anti-GRN and antibody. The resulting staining pattern was calibrated with prestained markers (shown in units of kDa on left) and conditioned medium from HEK293 cells expressing GRN (standard). β3-tubulin immunoreactivity served as loading control [34]. (b) GRN immunoreactivity was quantified by densitometry from triplicate immunoblots and normalized for β3-tubulin immunoreactivity. Results were then normalized to the laminectomy control and plotted as relative GRN level ± SD as a function of dpi. GRN levels increased relative to the laminectomy control (C, hollow bar) by 3 dpi and reached a maximum near 7 dpi. GRN levels decreased thereafter but remained elevated above the laminectomy control. (*, p < 0.05; ***, p < 0.001 compared to the laminectomy control group; ANOVA with Dunnett’s post hoc test). (c) Spinal cord homogenate prepared 7 dpi was incubated in the presence or absence of Peptide N-glycosidase F (PNGase F) followed by immunoblot analysis (10% acrylamide gel) with polyclonal anti-mouse GRN antibody. Tissue-derived GRN shifts to lower molecular mass in the presence of Peptide N-glycosidase F.

Fig. 4. Post-injury GRN expression occurs predominantly in macrophages and microglia

Double-label confocal images of transverse sections (at injury epicenter) labeled with CD11b to detect leukocytes (a, d) and anti-GRN antibody (b, e) at 14 dpi (gray matter) and 56 dpi (spared white matter). Merged images (c, f) reveal that nearly all GRN immunoreactive cells label positively for CD11b. Scale bar = 50 µm for panels a–c and 25 µm for panels d–f.

Fig. 5. Post-injury GRN expression colocalizes with lysosome-associated antigen CD68

Double-label confocal images of transverse sections (gray matter at injury epicenter) labeled with CD68 (a, d) to detect activated macrophages and microglia and anti-GRN antibody (b, e) at 28 dpi. Merged images (c, f) reveal that GRN associates with lysosomal-associated membrane protein CD68 in macrophage/microglia phagocytic vesicles. Panels d, e and f (scale bar = 25 µm) are magnified insets from images a, b, and c (scale bar = 100 µm), respectively.

Fig. 6. Post-injury GRN expression increases at the mRNA level

Levels of GRN mRNA in homogenates of injured and laminectomy control (C, hollow bar) spinal cords were assessed by microarray (a) and quantitative RT-PCR (b) analyses, and plotted as log 2 ratio and 2^−ΔΔC_T, respectively. Both methods reveal significant time-dependent increases in GRN gene expression after injury (**, p < 0.01; ***, p < 0.001 compared to the laminectomy control group; ANOVA with Dunnett’s post hoc test).