Communicated by David M. Kipnis, August 22, 1988. ABSTRACT. Transforming growth factor ,B ..... rPDGF-B is 360 g at day 7. 2230. Cell Biology: Pierce et al.
Proc. Nati. Acad. Sci. USA Vol. 86, pp. 2229-2233, April 1989 Cell Biology
Transforming growth factor /3 reverses the glucocorticoid-induced wound-healing deficit in rats: Possible regulation in macrophages by platelet-derived growth factor (procoflagen/chemotaxis/fibroblasts)
GLENN F. PIERCE*t*, THOMAS A. MUSTOE§, JANE LINGELBACHt, VICTORIA R. MASAKOWSKIt¶ PEGGY GRAMATESt§, AND THOMAS F. DEUELt¶ Departments of *Pathology, tMedicine, §Surgery, and IBiological Chemistry, Jewish Hospital at Washington University Medical Center, Saint Louis, MO 63110
Communicated by David M. Kipnis, August 22, 1988
ABSTRACT Transforming growth factor ,B (TGF-fi) and the platelet-derived growth factor (PDGF) are potent mitogenic polypeptides which enhance rates of wound healing in experimental animals; in contrast, glucocorticoids inhibit wound repair. The potential of TGF-fi and PDGF to reverse this inhibition in healing was tested in methylprednisolone-treated rats with deficits in skin wound strength of 50%. Single applications of TGF-fi (10-40 pmol per wound, 0.25-1 ,ug) applied locally at the time of wounding fully reversed this deficit in a concentration-dependent and highly reproducible manner. Wounds in glucocorticoid-treated animals were characterized by a near total absence of neutrophils and macrophages and by a delayed influx and reduced density of fibroblasts; however, such wounds treated with TGF-13 showed significant increases in wound fibroblasts and in intracellular procollagen type I. PDGF did not reverse the deficit in wound breaking strength in glucocorticoid-treated rats; there were more fibroblasts in the PDGF-treated wounds, but these fibroblasts lacked the enhanced expression of procollagen type I found in TGF-,-treated wounds. The wound macrophages, required for normal tissue repair, remained absent from both PDGF- and TGF-,B-treated wounds in glucocorticoid-treated animals. This result suggested that macrophages might normally act as an intermediate in the induction of procollagen synthesis in fibroblasts of PDGF-treated wounds and that TGF-,B might bypass the macrophage through its capacity to stimulate directly new synthesis of procollagen type I in fibroblasts. Whereas PDGF does not stimulate procollagen synthesis, in a rodent macrophage cell line, PDGF induced a highly significant, time-dependent enhancement of expression of TGF-fi.
Polypeptide growth factors derived from platelets and inflammatory cells are important mediators of tissue repair processes (1). Transforming growth factor f (TGF-f3) is a highly conserved 25-kDa homodimeric polypeptide ubiquitously expressed by mammalian cells and recognized by highly specific receptors on virtually every cell type analyzed (2, 3). TGF-,8 is released by cells that are localized at sites of tissue repair, such as platelets, activated macrophages, lymphocytes, and possibly fibroblasts (2-4). TGF-f3 is also a potent chemotactic agent for monocytes and fibroblasts (5, 6); it induces the synthesis of extracellular matrix proteins by fibroblasts (7-10), and it induces reversible fibroplasia in vivo (11). TGF-,8 has been found to directly accelerate the rate of healing of linear incisional wounds in rats (12, 13). Another polypeptide growth factor, platelet-derived growth factor (PDGF), is also a candidate wound hormone
(1). It is released by activated platelets, macrophages, and fibroblasts. PDGF has potent chemotactic activity for neutrophils, monocytes, fibroblasts, and smooth muscle cells (14-16). PDGF stimulates these cell types to express activities associated with inflammation and repair (17-19), recruits inflammatory cells and fibroblasts in vivo, and accelerates the healing of linear incisions in rats (20). Glucocorticoids, in contrast, inhibit wound healing in humans and in animal models of tissue repair (21, 22). The inhibition of procollagen synthesis in fibroblasts and the decrease in circulating monocyte levels associated with systemic administration of glucocorticoids sharply reduce the potential host responses for wound repair in vivo (21-25). In the present experiments, the potential roles of TGF-/3 and PDGF in reversing the glucocorticoid-induced deficit in wound strength of rat linear skin incisions have been investigated.
METHODS Study Design and Model. A full-thickness linear skin incision model in young adult male Sprague-Dawley rats (Sasco, Omaha, NB) was analyzed as described (12). From each rat, two to three pairs of dorsal skin strips were obtained for tensometry analysis, and one to two paired samples were obtained for histologic analysis. Methylprednisolone acetate (Depomedrol, Upjohn) was injected intramuscularly into rats 2 days prior to wounding. Rats were weighed at times of injection, wounding, and excision. Complete blood counts were analyzed by using a Coulter S Plus 4 automated analyzer. Highly purified human platelet-derived TGF-,f1 (the kind gift of M. Sporn, National Cancer Institute) (2) and rPDGF-B, the recombinant homodimeric product of the human c-sis gene (the kind gift of A. Thomason, Amgen, Thousand Oaks, CA) were applied as described (12). Collagen alone (control wounds) did not influence the rate of healing (12, 13). Data Analysis. Eight rats were used for each variable unless otherwise stated. Analyses of variance and paired t tests of breaking strength scores and of differences between matched experimental and control values were performed using the SAS data analysis system (Division of Biostatistics, Washington Univ.). Breaking strength measurements (grams of force; 1 g = 9.8 millinewtons) were performed on a tensiometer (Tensiometer 10, Monsanto) blindly on coded samples. Histologic Analysis and Immunoperoxidase Staining. Matched paired samples of wounds were placed immediately Abbreviations: rPDGF-B, the recombinant-DNA-produced B-B homodimeric form of platelet-derived growth factor; TGF-,B, transforming growth factor (3. tPresent address: Amgen, 1900 Oak Terrace Lane, Thousand Oaks, CA 91320.
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 2229
2230
Cell Biology: Pierce
et
Proc. Natl. Acad. Sci. USA 86 (1989)
al.
into acetic acid/ethanol (1:99, vol/vol) fixative and processed as described (20). Paired sections were microscopically analyzed by two independent observers who did not know which wound treatment the samples represented, using an arbitrary scale from 0 to 4 for degree of cellularity and for cell types. Immunoperoxidase staining of samples was performed by established methods, using rabbit anti-rat procollagen type I monospecific antiserum (a generous gift of K. Cutroneo, Univ. of Vermont) at 2 gg/ml (24) and biotinconjugated goat anti-rabbit IgG (Bethesda Research Laboratories) followed by streptavidin-horseradish peroxidase (Bethesda Research Laboratories) (26). Negative controls consisted of parallel sections incubated with comparable dilutions of irrelevant primary antisera. Paired wound sections were analyzed microscopically by two independent observers as described above. RNA Isolation and Northern Blot Analysis. The BALB/c murine macrophage cell line J774A.1 (American Type Culture Collection) was treated with rPDGF-B at 50 ng/ml for up to 52 hr. Total cellular RNA was isolated by a modified guanidine hydrochloride procedure (27). Equal amounts (15 tkg) of RNA were separated on standard denaturing formaldehyde agarose gels, blotted onto nitrocellulose (Schleicher & Schuell), prehybridized for 2 hr at 420C, and transferred to fresh hybridization solution with the addition of 1 x 106 cpm/ml of nick-translated TGF-P probe [1.6-kilobase (kb) insert of the sp65-Murf3a.s. plasmid, gift of R. Derynck, Genentech] labeled to a specific activity of 5 x 108 cpm//ug. After an overnight hybridization at 420C, the washed blot was exposed to Kodak X-Omat film for 1 hr and quantitated on a scanning laser densitometer (LKB).
RESULTS Establishment of Significant Impairment of Wound Healing by Glucocorticoid Administration. Previous experiments have shown that growth factors increase by 2-fold the force
required to rupture incisional wounds in normal rats, thereby accelerating wound healing by nearly 3 days in the first week of wounding (12, 20). However, in severely impaired wound healing, the effects of growth factors have not been studied. A reproducible and significant inhibition of wound healing was established by testing increasing doses of methylprednisolone (0, 12, 30, and 60 mg/kg) in a slow release vehicle to groups of four rats by a single injection 2 days prior to wounding. The breaking strength of wounds was an accurate and highly reproducible measure of impairment in wound healing. Methylprednisolone at 30 mg/kg decreased wound breaking strength by 74% and 47% at days 5 and 7 after wounding, respectively. Higher doses of methylprednisolone induced too profound a reduction of wound breaking strength for reproducible analyses. TGF-13 Reverses Healing Impairment in Methylprednisolone-Treated Rats. Since previous investigations showed that TGF-,B applied at the time of wounding doubled breaking strength of day 7 wounds in normal rats (12, 13), day 7 was chosen to measure the effect of increasing concentrations of TGF-,B in glucocorticoid-treated rats (Fig. 1). A positive dose-dependent effect of TGF-f3 was observed on wound breaking strength, beginning at 0.2 ,ug of TGF-,B per incision. Full reversal of the wound-healing deficit resulted when 0.51 ,ug of TGF-P per incision was applied. Wounds treated with 1 Mg of TGF-P had breaking strength values nearly 200% of control values after 1 week (Fig. 1); the reversal of the wound-healing deficit was transient, however (Table 1). Statistically significant reversal was demonstrated at days 5 and 7 after wounding but not at days 10-14. By this time, the inhibitory effects of the single pretreatment with methylpred-
nisolone were diminished (Table 1) and, importantly, the augmented healing in normal rats induced by TGF-f3 was
O GC* hTGF-p3
0.001
z
J'C
0.05 0.01
2000
0.5
E
LU
1001
0 0
0
0.1
0
0.2
hTGF-P
0 0.5
0
1.0
0
1.5
(,ug/incision)
FIG. 1. Dose-dependent increase in breaking strength of TGF(3-treated wounds 7 days after wounding. Rats were injected with the glucocorticoid (GC) methylprednisolone (30 mg/kg) 2 days prior to wounding. TGF-/3 was applied locally to one incision, at the time of surgery; the other incision served as a matched, paired control. Wounds excised on the seventh day showed a maximal TGF-/3 effect in normal rats (12). SEM and P relative to control are given above the bars.
declining (12). TGF-f3 thus fully reversed the deficit in wound healing induced by glucocorticoids in rats. Effects of rPDGF-B on the Methylprednisolone-Induced Wound-Healing Deficit. rPDGF-B at 20 ug per incision was ineffective in augmenting the breaking strength of wounds in the methylprednisolone-treated animals when tested at days 7 and 10 after wounding (Table 1). In contrast, highly significant stimulatory effects of rPDGF-B were reproducibly observed in normal rats at days 7 to 10 after wounding, enhancing the wound breaking strength to values 170% of control wounds (Table 1; ref. 20). Analysis of Cellular Migration into Growth Factor-Treated Wounds. A striking influx of cells into TGF-p- and rPDGFB-treated wounds in normal rats was observed previously (12, 13, 20). Histologically, these wounds resembled a highly exaggerated normal inflammatory component of tissue reTable 1. Effect of TGF-,B or rPDGF-13 applied at the time of wounding on the glucocorticoid-induced healing impairment
Breaking strength, g P Control 34 ± 3 (29%7o) 0.01 49 ± 5 (42%) TGF-j3 7 TGF-f3 187 ± 22 (97%) 132 ± 20 (68%) 0.02 10 TGF-,8 267 ± 25 (75%) 240 ± 24 (68%) NS 14 TGF-f3 523 ± 71 (92%) 474 ± 43 (84%) NS 7 rPDGF-/3 124 ± 14 (74%) 118 ± 7 (701%) 10 rPDGF-P 216 ± 18 (49%7b) 233 ± 24 (53%) Groups of eight rats were injected with methylprednisolone at 30 mg/kg 2 days before wounding and either 1 ,ug of TGF-f3 or the collagen vehicle alone was applied opposite to paired incisions on each rat. Wounds were tested for breaking strength on the days on which TGF-,f- or rPDGF-,8-treated wounds were previously shown to have significantly stronger breaking strength measurements compared to matched controls (12, 20). Results are presented as mean ± SEM. P is for two-tailed paired t test on glucocorticoid-treated samples, growth factor-treated vs. controls. NS, not significant. The percentage of the normal control wound breaking strength values for each day is shown in parentheses. The breaking strength of normal wounds treated with 2 ,&g of TGF-f3 is 180 g at day 5 and 310 g at day 7. The breaking strength of normal wounds treated with 20 jig of rPDGF-B is 360 g at day 7.
Day after wounding 5
Growth factor
TGF-.3-treated
Cell
Proc. Nati. Acad. Sci. USA 86 (1989)
Biology: Pierce et A
~~~~~~~~~~~.05
4
~
1
2
~
3
~
0
5
7
10
DAYS POSTWOUNDING
FIG. 2. Cellular influx into wounds. Rats were pretreated with the glucocorticoid (GC), and either TGF-,B (1 ,ug) or rPDGF-B (20 Ag) was applied locally at the time of surgery. Paired wounds from four to eight rats per time point were excised and two separate histologic sections were obtained and rated for cellularity on a scale from 0 to 4 by two independent observers who did not know which treatments the samples represented. A score of 0 represented baseline cellularity in unwounded dermis, and a score of 4 represented the degree of cellularity observed 2-3 days after wounding in normal (untreated) wounds. Results from both observers were averaged and analyzed. Although some neutrophils were observed in wounds at days 1 and 2, macrophages were not observed, and the vast majority of the increased cellular influx from day 3 onward was due to fibroblasts, which peaked at days 5-7 after wounding in glucocorticoid-treated rats compared to days 3-5 after wounding in normal rats. Both rPDGF-B and TGF-,8 enhanced the degree of fibroblastic influx but did not alter the delayed kinetics induced by glucocorticoid treatment. The fibroblastic influx induced by rPDGF-B was more sustained than that found in TGF-p8-treated wounds. In contrast, granulation tissue was increased only in TGF-p-treated wounds (Fig. 3).
pair. The cellular influx correlated directly with the potent influence of growth factors as chemotactic factors in vitro. Normal wound healing is characterized by the timedependent influx of neutrophils within the first day after wounding, followed by an influx of macrophages within 2-3 __ -,.
2231
days of wounding, and by an influx of fibroblasts by days 34 (13, 20). In contrast, in methylprednisolone-treated animals, the wounds contained few neutrophils at days 1 and 2 and were essentially devoid of macrophages (Fig. 2). Fibroblast influx was delayed until days 5-7 after wounding. TGF-f3 and rPDGF-B applied locally had no effect on restoring neutrophil and macrophage influx; both growth factors significantly augmented the number of fibroblasts entering wounds (PDGF-B > TGF-f8), though entry was still delayed (see Fig. 2 legend). TGF-/3-treated wounds from glucocorticoid-treated animals had increased granulation tissue 7 days after wounding (Fig. 3A), compared to matched control wounds (Fig. 3B), whereas rPDGF-B-treated wounds did not. Thus, in the wounds from PDGF-treated rats, the more enhanced fibroblastic response was associated with a marked reduction in the formation of granulation tissue. TGF-P- and rPDGF-B-treated wounds in glucocorticoidtreated animals lacked the influx of monocytes and neutrophils which is characteristic in wounds in nonglucocorticoidtreated animals. A drop in the leukocyte count (WBC) reproducibly occurred after methylprednisolone treatment (normal, 9820 cells per mm3 vs. 3200 cells per mm3 2 days after glucocorticoid treatment; P < 0.01). Monocyte counts decreased to nearly undetectable levels within 1 day after wounding (normal, 680 cells per mm3 vs. glucocorticoidtreated, 80 cells per mm3; P < 0.001) before returning over the next week to nontreated control levels. The loss of circulating leukocytes correlated directly with the lack of neutrophils and tissue macrophages in wounds in glucocorticoid-treated rats. Since the monocyte/macrophage, but not the neutrophil, is required for normal incisional repair (21), these results suggested that TGF-13 may enhance wound healing in this model by a direct influence on the fibroblast; PDGF, in contrast, may require the macrophage to mediate a significant component of the tissue repair activities it has in normal rats. Enhanced Procollagen Type I in TGF-i3-Treated Wounds. In normal wounds, intracellular procollagen type I may be detected as early as day 2 after wounding in fibroblasts migrating into the wound (unpublished observations). However, in control, TGF-.8-treated, and rPDGF-B-treated wounds from methylprednisolone-treated animals, intracellular procollagen type I was not detected until day 5 after wounding, correlating directly with delayed fibroblast influx
FIG. 3.
Photomicrographs
showing fibroblast and granulation tissue content in TGF-Ptreated (A) and matched, paired I6
'* *;K
control (B) wounds after 7 days. gAt the time of surgery, TGF-,B, 1 or vehicle alone was applied to incisions. Wounds were harvested and processed for histologic anal$ s -
-ug,
N i-t; b
-Xx ;
b _
n
ysis. (Original magnification, x 100; printed at x70). The arrows
demarcate the boundaries of new
granulation tissue through the der-
mis. Note the increased numbers of fibroblasts in TGF-f-treated wounds compared to controls. rPDGF-B induced increased numbers of wound fibroblasts without the increased granulation tissue observed in TGF-,3-treated wounds (data not shown).
2232
He~ ~ lo
Cell Biology: Pierce et al.
Proc. Natl. Acad. Sci. USA 86 (1989) *
,
.
L
49'4
A..:.
*:qa.+:
-sk..,..,. , !0 t
Ie
FIG. 4. Photomicrographs showing fibroblast influx and procollagen type I staining in TGF-fi (1 ,ug) (A) and matched, paired control (B) wounds 5 days after wounding. Both fibroblast content and procollagen type I staining (brown) were assessed on coded slides. (Original magnification, x200; printed at x 140). These slides showed the wound dermis at the level of the panniculus carnosus and were representative of increased fibroblast content and increased procollagen type I staining observed in TGF,B-treated wounds compared to controls.
A.rAP-¼hAp;- t MO ~ ~ ~
tve~l F- _
'~'
~
~
~
~
~
~~~-
4p
77 Mu; ~*
of -.
-C
.401X
:*
