Platelet-derived growth factor-beta receptor activation is essential for fibroblast and pericyte recruitment during cutaneous wound healing

Abstract

Connective tissue remodeling provides mammals with a rapid mechanism to repair wounds after injury. Inappropriate activation of this reparative process leads to scarring and fibrosis. Here, we studied the effects of platelet-derived growth factor receptor-beta blockade in vivo using the platelet-derived growth factor receptor (PDGFR)-beta inhibitor imatinib mesylate on tissue repair. After 7 days, healing of wounds was delayed with significantly reduced wound closure and concomitant reduction in myofibroblast frequency, expression of fibronectin ED-A, and collagen type I. Using a collagen type I transgenic reporter mouse, we showed that inhibiting PDGFR-beta activation restricted the distribution of collagen-synthesizing cells to wound margins and dramatically reduced cell proliferation in vivo. By 14 days, treated wounds were fully closed. Blocking PDGFR-beta signaling did not prevent the differentiation of myofibroblasts in vitro but potently inhibited fibroblast proliferation and migration. In addition, PDGFR-beta inhibition in vivo was accompanied by abnormal microvascular morphogenesis reminiscent of that observed in PDGFR-beta-/- mice with significantly reduced immunostaining of the pericyte marker NG2. Imatinib treatment also inhibited pericyte proliferation and migration in vitro. This study highlights the significance of PDGFR-beta signaling for the recruitment, proliferation, and functional activities of fibro-blasts and pericytes during the early phases of wound healing.

Figure 1

Wound healing is impaired in mice treated with imatinib. A: Punch wounds (4 mm³) were made on the back of anesthetized mice. After 3 days, wound diameters are noticeably smaller in control animals (B) in comparison with imatinib-treated animals (D). After 7 days, wound diameters in control wounds (C) are clearly reduced compared with imatinib-treated wounds (E). Analysis of sections stained with H&E confirmed that after 7 days, distance between wound margins was greater in imatinib-treated animals (G) compared with control animals (F). Staining of sections with Masson’s trichrome confirmed that granulation tissue in treated animals was poorly formed and relatively hypocellular (I, arrows) compared with control animals (H, arrows). Granulation tissue in control animals was characterized by the formation of small blood vessels (H, inset, arrowhead), which were not present in imatinib-treated wounds (I, inset, arrowhead). Quantification of wound diameter throughout 10 days after injury. Results represent the mean ± SEM. Between 3 and 7 days after wounding, the difference in wound diameter was significant (P < 0.05). Scale bars = 100 μm (F, G); 50 μm (H, I); 10 μm (insets).

Figure 2

Cell proliferation is inhibited in imatinib-treated mice. Mice were injected with BrdU 2 hours before sacrifice. Sections were subsequently stained with an anti-BrdU antibody. Three days after injury, immunostaining for BrdU was significantly increased in control animals (A, C) compared with treated animals (B, D). In control animals, BrdU immunostaining was predominant at the wound margins (A, arrow), which was significantly reduced in imatinib-treated animals (B, arrow). In control animals, BrdU immunostaining was present in pericytes (C, inset, arrowhead) in contrast to imatinib-treated animals (D, inset, arrowhead). E and F: After 7 days, BrdU immunostaining was present in the granulation tissue of both control and imatinib-treated animals. G: Quantification of BrdU-stained cells confirming that imatinib treatment produced a significant reduction in cell proliferation after 3 and 7 days. Results represent the mean ± SD. *P < 0.005. Inhibition of pericyte (H) and fibroblast (I) proliferation by imatinib was confirmed by in vitro analyses. Ten percent of FCS and PDGF-BB-induced stimulation of pericyte and fibroblast proliferation was inhibited by treatment with imatinib (2 μm). Results represent the mean ± SD. *P < 0.05. Scale bars = 100 μm (A, B); 50 μm (C–F); 10 μm (insets).

Figure 3

Imatinib treatment does not affect apoptotic cell death in vivo. Apoptotic cell death in day 7 wounds was assessed by TUNEL staining. Apoptotic nuclei (arrows, red) were detected in both control (A) and imatinib-treated wounds (C). Cell nuclei were counterstained with DAPI (B, D). Analysis of the percentage of apoptotic cell nuclei in control and imatinib-treated sections revealed no significant difference (E) (P = 0.8). Quantification of cell nuclei revealed 10% fewer cells in imatinib-treated wounds compared with controls (F). Data shown are mean ± SD. Original magnifications, ×20.

Figure 4

Imatinib treatment impairs migration of pericytes and fibroblasts in vitro. A: Scratch wounds were made in confluent cell monolayers. In response to both 10% FCS and 10 ng/ml PDGF-BB, fibroblasts (B, D) and pericytes (F, H) completely fill a scratch wound after 72 hours. The addition of imatinib (2 μm) completely abrogates migration induced by both 10% FCS and PDGF-BB in fibroblasts (C, E) and pericytes (G, I). Impairment of cell migration was confirmed and quantified using free-floating collagen matrices. Representative images of pericyte-seeded gels are shown. Pericytes were seeded into a collagen gel matrix, and after gel polymerization, the gel was detached from the tissue culture plate and incubated in the presence or absence of PDGF-BB (10 ng/ml). J: Pericyte contraction of gels in response to PDGF-BB is inhibited in the presence of imatinib. K: Quantification of gel weight after contraction. Results represent the mean ± SD.

Figure 5

Imatinib treatment delays the appearance of myofibroblasts. A: Three days after injury, α-SMA immunostaining is present in myofibroblasts in the wound margins of control animals (arrow). B: In imatinib-treated animals, α-SMA immunostaining is restricted to microvessels (arrow). After 7 days, α-SMA immunostaining is present throughout the granulation tissue of control (C, arrows) and present in imatinib-treated wounds, particularly at the epidermal/dermal junction (D, arrows). Quantification of the number α-SMA-expressing myofibroblasts by image analysis confirmed a significant reduction in myofibroblast numbers in imatinib-treated animals 3 days after injury. E: After 7 days, myofibroblast numbers were still lower in imatinib-treated animals; however, the difference was not significant. Expression of ED-A FN was also reduced as a result of imatinib treatment (G, arrow) compared with control tissue (F, arrows). Results represent the mean ± SD. *P < 0.01. TGF-β (2 ng/ml) treatment of cultured fibroblasts produced prominent α-SMA stress fibers (H, arrowhead), which was not inhibited by imatinib treatment (I, arrowhead). J: Quantification of myofibroblast numbers after TGF-β and imatinib treatment. Results represent the mean ± SD. Scale bars = 50 μm (A–G); 25 μm (H, I).

Figure 6

Imatinib mesylate does not block serum-induced contraction of tethered collagen gels. The ability of fibroblasts to contract tethered collagen lattices was assessed. Cells were seeded into tethered collagen gels. Mechanical tension was generated throughout a 48-hour period after which the gels were released. DMEM containing 10% FCS (C) stimulated contraction in comparison to serum-free DMEM (A) and PDGF-BB (10 ng/ml) (E). Imatinib did not affect contraction by serum-free DMEM (B), 10% FCS (D), and PDGF-BB (F). G: Quantification of gel weight after contraction. Results represent the mean ± SD. *P < 0.05.

Figure 7

Collagen type I gene promoter activity is reduced in imatinib-treated animals. Seven days after injury, wounds were excised from mice and stained with 1 mg/ml X-gal. Control wounds show uniformly intense X-gal staining (A, arrow), whereas staining in tissue from imatinib-treated mice was significantly weaker and confined to the wound margins (B, arrow). C: In histological sections, X-gal staining could be observed in fibroblastic cells throughout the granulation tissue (arrowheads). D: In sections from imatinib-treated animals, relatively fewer blue cells were present and were restricted to the wound margins (arrowhead). G: Imatinib-mediated reduction in collagen transgene activity was confirmed by quantification of transgene activity in whole wounds. Results represent the mean ± SD. *P < 0.01. Immunofluorescence staining of 14-day-old wounds confirming that expression of collagen protein was concordantly reduced in imatinib-treated wounds (F, arrow) relative to control tissue (E, arrow). Scale bar = 50 μm.

Figure 8

Imatinib treatment impairs angiogenesis during wound healing. Masson’s trichrome staining demonstrating the presence of abnormally dilated microvessels in day 7 granulation tissue of imatinib-treated animals (B, arrows) compared with normal microvascular architecture in control wounds (A, arrows). Cryosections from control and imatinib-treated wounds were stained with antibodies recognizing endothelial cells (CD31) and pericytes (NG2). In control wounds, 7 days after injury, immunostaining for both CD31 (C, arrows) and NG2 (E, arrows) was prominent throughout the granulation tissue. In imatinib-treated wounds, immunostaining for both CD31 (D, arrows) and NG2 (F, arrows) was comparatively reduced. Scale bar = 25 μm.