Platelet-Derived Growth Factor BB - Europe PMC

American Journal of Pathology, Vol. 145, No. 6, December 1994 Copyright C) Amenrican Society for Investigative Pathology

Tissue Repair Processes in Healing Chronic Pressure Ulcers Treated with Recombinant Platelet-Derived Growth Factor BB

Glenn F. Pierce,* John E. Tarpley,* Richard M. Allman,t Patricia S. Goode,t Cuneyt M. Serdar,A Barry Morris,* Thomas A. Mustoe,§ and Jerry Vande Berg"l From the Department of Experimental Pathology,* Amgen Inc., T7housand Oaks, California; Centerfor Aging,

Department of Medicine,t University of Alabama at Birmingham and the Birmingham Veterans Affairs Medical Center, Birmingham, Alabama; Department of Clinical Affairs,* Amgen Inc., Thousand Oaks, California; Division of Plastic and Reconstructive Surgery,5 Northwestern University, Chicago, Illinois; and Core Electron Microscopy Lab,1" Veteran's Affairs Medical Center, University of California, San Diego, California

Celular and molecular mechanisms responsible for the observed vulnerary effects of recombinant human platelet-derived growth factor BB (rPDGF-BB) in man have not been elucidated. In a double-blinded trial, patients having chronic pressure ulcers were treated topicaly with either rPDGF-BB orplacebofor28 days. To explore how rPDGF-BB may induce chronic wounds to heal, biopsies were takenfrom the ulcers of a cohort of 20 patients from the ttlal and evaluated in a blindedfashion by light microscopyfor 1),fibroblast content, 2), neovesselformation, and 3), collagen deposition. Electron microscopy also was used to assess fibroblast activation and coUagen deposition. Before initiation of therapy most wounds had few fibroblasts and most of those present were not activated When mean scoresfor the total active treatment phase (days 8, 15, and 29) for rPDGF-BB-treated ulcers were compared with the scores for placebo-treated ulcers, fibroblast content was significantly higher for the rPDGF-BB-treated ulcers (P = 0.03, KruskalWaUis test). More significant differences in fibroblast and neovessel content were observed when six nonhealing wounds were eliminatedfrom the analysis (three placebo, three treatment). Thus, in aU healing wounds, rPDGF-BB therapy signifi-

cantly increased fibroblast (P = 0.0007) and neovessel (P = 0.02) content. These results were correlated with increased collagenfilbrillogenesis by fibroblasts from healing rPDGF-BB-treated wounds, as assessed by intracelularprocollagen type I immunostaining, and by electron microscopy, and were concordant with clinical measurements (eg, area of ulcer opening and ulcer volume) which showed greater healing in rPDGF-BBtreated wounds. These results suggest induction offibroblast proliferation and differentiation is one mechanism by which rPDGF-BB can accelerate wound healing and that rPDGF-BB can augment healing responses within a majority of, but not aUl, nonhealing chronic pressure ulcers in man. (Am JPathol 1994, 145:1399-1410) Chronic pressure ulcers have a complex etiology. Although they develop from pressure and shearing forces on skin overlying bony prominences, resulting in tissue ischemia and necrosis, the reasons that chronic pressure ulcers heal poorly are not understood,1'2 Chronic pressure ulcers frequently have poorly vascularized, thick fibrotic scar tissue surrounding the bed and keratinocytes at the ulcer margins incapable of adequate proliferation and migration.3 Fibroblasts are sparse and appear morphologically inactive. Thus, the regulatory signals normally found in the repair of acute wounds do not appear to be operational in chronic pressure ulcers. In addition, lack of adequate pressure relief results in ongoing ischemia and resultant tissue necrosis. Many wounds are contaminated with bacteria, which may also prevent the progression of normal tissue repair processes.3'4 PDGF is capable of stimulating proliferation, chemotaxis, and gene activation within macrophages Accepted for publication August 11, 1994. Address reprint requests to Dr. Glenn F. Pierce, Department of Preclinical Science, PRIZM Pharmaceuticals, 11035 Roselle Street, San Diego, CA 92121.

1399

1400

Pierce et al

AJP December 1994, Vol. 145, No. 6

and fibroblasts.5'6 PDGF induces accelerated deposition of provisional matrix and subsequent collagen formation in experimental models of tissue repair.7 -14 Thus, in rodent, lapine, and porcine wound healing models, locally applied PDGF has been found to exaggerate acute inflammatory and matrix deposition phases of tissue repair, resulting in more rapid healing.6'15 PDGF can also reverse the healing impairment in irradiated, diabetic, and hypoxic wounds,716 18 although the relevance of these models to chronic ulcers remains to be established. In man, rPDGF-BB has been shown to accelerate healing in chronic pressure ulcers19-21 although the responsible mechanism(s) have not been elucidated at the cellular level. In an initial small double-blinded clinical trial performed in the young, healthy spinal cord injury patients who were hospitalized for the entire treatment period, all ulcers showed substantial healing; however, ulcers treated with rPDGF-BB (100 pg/ml/day) had somewhat greater healing than placebo or lower dose rPDGF-BB-treated ulcers.19,20 In a subsequent double-blinded study performed on elderly debilitated patients who predominantly resided in nursing homes, placebo-treated ulcers healed approximately 25% over 28 days as assessed by ulcer volume and area of opening, whereas rPDGF-BB-treated ulcers (100 or 300 pg/ml/day) healed approximately 50% during the treatment period.21 From a cohort of patients in this latter study, we studied ulcer biopsies that were obtained during the treatment period to evaluate the effect of rPDGF-BB on fibroblast content, neovessel formation, and collagen deposition in pressure ulcers. Our findings provide insight into the mechanisms by which rPDGF-BB stimulates healing of such chronic wounds.

Materials and Methods

Patients and Study Design All patients had chronic pressure ulcers for at least 2 months and were an average age of approximately 70 years. The multicenter trial was double-blinded; patients were randomized and treated daily with either placebo, rPDGF-BB (100 pg/ml (1 pg/cm2) or 300 pg/ml (3 pg/cm2)) for 28 days, as reported.21 Volumetric and planimetric measurements were obtained weekly and were used to determine whether ulcers were healing (both measurements decreasing in size) or nonhealing (one or both measurements remaining static or increasing in size during the treatment period) in our cohort of patients. The clinical findings

from all patients enrolled in the trial have been published elsewhere.

Biopsies Three-millimeter full thickness punch biopsies were collected before treatment on day 0 and on days 8, 15, and 29 (day after last treatment) from approximately half the patients in the clinical trial (two of the three treatment sites). Serial biopsies were obtained from seven placebo patients, seven 100 pg/ml rPDGF-BB-treated patients, and six 300 pg/ml rPDGF-BB-treated patients. On days 0 and 8, punch biopsies were taken from the leading edge of the ulcer whereas biopsies from the remaining treatment days were collected from the same radius as the day 8 biopsy. Thus, tissue samples from healing ulcers were from an area of maturing scar on days 15 and 29, whereas biopsies from nonhealing ulcers were from the initial wound margin at these time points.

Light and Electron Microscopy Biopsies were collected for light microscopy at all time points. Those collected solely for light microscopy were fixed in OmniFix 11 (An-Con Genetics, Inc., Melville, NY), processed, and paraffin embedded. Biopsies obtained for both light and transmission electron microscopy (approximately half the patients in the cohort) were fixed in a modified Trump's fixative (1.5% paraformaldehyde/0.5% glutaraldehyde in phosphate buffer, pH 7.4) and then bisected longitudinally. The half used for light midroscopy were processed and paraffin embedded, and the half for electron microscopy were minced, then post-fixed in phosphate buffered osmium tetroxide, pH 7.4. Ulcer tissues were dehydrated through a graded series of alcohol and propylene oxide and embedded in SciPoxy 812 (Energy Beam Sciences, Agawam, MA). All tissue blocks were thick sectioned (1.0 p), stained with toluidine blue, and examined by light microscopy for tissue orientation. Thin sections (60 nm) were stained with uranyl acetate and bismuth nitrate, mounted on unsupported 200-mesh copper grids, before examining in a Zeiss EM-10B electron microscope.

Histochemical Stains Masson's trichrome staining was done on each biopsy to evaluate fibroblast content and size, neovessel formation, and collagen staining. To further study

._

Healing Processes in Chronic Pressure Ulcers 1401 AJP December 1994, Vol. 145, No. 6

Figure 1. Appcearance of chronic ulcers beo/bre treatment. Many fibroblasts (F) in chronic lulcers often appeared necrotic. These cells contained vacuolated nuclei and a cytoplasm that uas vacuolated and/or filled with filaments and fragmented organelles. Arrowheads, vacuoles; arrow, microfilaments. Original magnification, x 13,500.

Table 1. Area of Ulcer Opening in the Cohort of Patients Biopsied

Treatment Placebo

1 pg/cm2 3 pg/cm2 *

t

w

0 CL)

n)

n 7 7 6

Day 0 6.9 5.1 5.4

+ + +

4.1t 1.8 1.6

Ulcer Area (cm2)* Day 8 Day 15 6.7 + 4.1 4.2 + 1.4 4.5 + 1.4

6.1 + 3.7 2.7 + 1.0 4.2 + 1.4

Day 29 5.1 + 3.1 1.6 + 0.4 2.6 + 0.7

Area of ulcer opening as measured by planimetry.21 Mean + SE.

Figure 2. Fibroblast and neovessel content in rPDGF-BB- and placebo-treated ulcers. Serial biopsies from all ulcers, healing ulcers, and nonhealing ulcers were analyzed separately for both treatment groups (number of biopsies eLaluated at bottom of bar).

the collagen content, Sirius red staining was performed and evaluated under polarization optics.15 Sirius red binds linearly to collagen and enhances its natural birefringence. The color of birefringence does not correlate with collagen type but does correlate to fiber size and degree of crosslinking.

Assessment of Fibroblast and Neovessel Content in Biopsy Photomicrographs For photo rankings, photomicrographs of both Masson trichrome and Sirius red stains were taken of all biopsies at low power (x4; Nikon Optiphot). Sirius red

1402

Pierce et al


A. ALL ULCERS

3pd.1

w 20

CD) 1I

n 11-

Figure 3. Fibroblast and neovessel content in rPDGF-BB- andplacebo-treated ulcers through-

8

out the treatment phase.

15

29

8

15

29

TREATMENT DAYS

L

....i

i i 11

i: ..-j

Figure 4. Histological picture of healing and nonhealing ulcers on treatment day 15. A, rPDGF-BB-treated, healing; B, placebo-treated, healing; C, placebo-treated, nonhealing. Note activated fibroblasts and neovessel formation in healing ukers. Ep, epidermis; arrowheads, developing neovessels; arrows, fibroblasts. Masson-trichrome, original magnification, x 160.

stains were photographed with a polarization attachment. Two observers blinded to the treatments evaluated prints for fibroblast and neovessel content, as well as collagen deposition, using a scale of 0 to 4 with 0 indicating no activity and 4 indicating marked activity. Standard photos for each rating point were chosen before the evaluation and each photo was referenced to the standards during evaluation. Scores from the two evaluators were averaged and usually agreed to within 0.5 units. Fibroblasts were considered activated if they were enlarged and contained a prominent nucleus and cytoplasm. Neovessels were easily delineated on Masson-trichrome-stained sections.

Immunohistochemical Stains To further investigate fibroblast activation and new collagen production, sections were stained with an antibody to the N-terminus of the type procollagen (PC-I) molecule (Chemicon, Temecula, CA) by using conventional techniques.22

Quantitation of Fibroblast Size The cytoplasmic area of PC-I-stained fibroblasts was measured23 with a Quantimet 520 system (Leica, Deerfield, IL). Slides were imaged using a Nikon Optiphot microscope, a Cohu video camera (Cohu, San

Healing Processes in Chronic Pressure Ulcers 1403 AJP December 1994, Vol. 145, No. 6

Figure 5. Appearance of placebo-treated chronic ulcers on day 8. Placebo-treated wound fibroblasts were often similar morphologically to fibroblasts in untreated wounds (see Figure 1). Fragmented cell membranes were commonly observed in the extracellular matrix. Circumscribing endothelial cells (EN) often surrounded lumina occluded with fibrin (arrowhead), cell membranes, chromatin (*) and other subcellular debris. Magnification, x 6300.

Diego, CA) and a 540-nm filter with a 10-nm band width (Oriel, Stratford, CT). Image frames were set to measure across the biopsy width from the base of the biopsy covering (exudate, epithelium, or no covering) to a depth of 0.5 mm. Measurements were collected from the entire biopsy width (total fibroblasts), from the area of the biopsy covered only by exudate (ulcer fibroblasts), and from the area covered by new epithelium (subepidermal fibroblasts). Thus, individual measurements were made of fibroblast areas in the total biopsy width, the ulcer bed only, and under the newly formed scar.

Data Analysis Data from both the 1- and 3-pg/cm2 rPDGF-BB treatment groups were pooled for most analyses, inasmuch as both groups previously showed comparable healing.21 Significance was tested by using KruskalWallis and Mann-Whitney U nonparametric tests (StatView IV, Abacus Concepts, Berkeley, CA). Twotailed P values were used. For fibroblast areas onetailed unpaired Student's t-tests were performed with the same software.

Results Fibroblast and Neovessel Content in Wounds Biopsies were obtained from each patient on days 0 and 8 and from the day 8 radius on days 15 and 29. Healing wounds were wounds in which the area of ulcer opening and the ulcer volume decreased over

the 28-day treatment period. Thus, in healing wounds, biopsies from days 15 and 29 were obtained from regenerated epidermis and maturing granulation tissue. Because there was no repair in nonhealing wounds, biopsies from days 15 and 29 continued to be obtained from the leading edge of the wound bed. In this cohort of patients, 10 of 13 ulcers in the rPDGF-BB treatment groups (6 of 7 1 pg/cm2; 4 of 6 3 pg/cm2) were healing, compared with 4 of 7 placebo-treated ulcers. On day 0, before treatment, most chronic wounds contained primarily inflammatory cells within a fibrin matrix. They contained fibroblasts that appeared inactive and were scattered at wound borders. In most instances, fibroblast cytoplasm was sparse and vacuolated and contained few organelles and fragmented mitochondria (Figure 1). Ulcer beds had a variable number of vessels, and some were thrombosed. Neutrophils tended to be localized to vessel walls and near the wound surface. Monocytes and macrophages were scattered throughout the ulcer bed. Lymphocytes were present in appreciable quantities in few wounds and, when present, appeared to be responding to foreign material or wound debris. No ulcers appeared to be infected. In the clinical trial, both rPDGF-BB dose groups showed increased healing compared with the placebo group (Table 1), and no dose-dependent differences were detected.21 In the histological analysis, both 1- and 3-pg/cm2 doses showed comparable increases above placebo for fibroblast content by multivariate analysis (P = 0.03, Kruskal-Wallis). Thus, given the small sample size for most subsequent

1404

Pierce et al


Figure 6. Appearance of rPDGF-BB-treated wounds after 8 days (A) or 15 days (B) of treatment. In (A), fibroblasts (F) at the leading edge of the wound are metabolically active and are surrounded by a provisional matrix lacking significant amounts of collagen. Macrophages (M) are also present Magnification, X 10,800. In (B), fibroblasts present in matur-

ing granulation tissue contained abundant mitochondria and rough endoplasmic reticulum. Collagen deposition is evident, and some collagen bundles were observed. Some inflammatory cells (M, mast cell; E, eosinophil) and a new vessel (EN, endothelial cells) were found. Arrowheads, collagen bundles. Original magnifi-

N.'t

cation, x 13,500.

Table 2. Score for Collagen Deposition in All Pressure Ulcers on Treatment Days 15 and 29

Treatment

n

Placebo rPDGF-BBt

7 13

t

Day 15 2.43 ± 0.41* 2.77 + 0.22

Day 29 2.57 ± 0.49 3.12 ± 0.20

Mean - SE. Both 1- and 3-pg/cm2 treatment groups.

analyses, both dose groups were pooled and compared with the placebo-treated group. An increase in fibroblast content was detected in all rPDGF-BB-treated ulcers, compared with placebo (2.81 ± 0.17 versus 2.05 ± 0.24, P = 0.01), when biopsies from treatment days 8, 15, and 29 were analyzed together in a blinded fashion for each patient

(Figure 2). Although marked increases in neovesselcontent also were detected in rPDGF-BB-treated ulcers, the increase was not statistically significant because of wound variability. When only ulcers that showed healing during the treatment phase were analyzed, significant increases in both fibroblast content and neovessel formation were observed in rPDGF-BB-treated ulcers compared with placebo-treated ulcers (Figure 2). These results indicate more pronounced cellular responses induced by rPDGF-BB in healing ulcers. In nonhealing ulcers, no differences were observed in fibroblast and neovessel content. For all ulcers, when treatment days 8, 15, and 29 were analyzed separately, differences in fibroblast

Healing Processes in Chronic Pressure Ulcers

1405


Figure 7. Sirius red staining of healing and nonhealing ulcers. A: rPDGF-BB-treated, healing; note the maturing red collagen fibrils and maturing collagen bundle formation. B: placebo-treated, nonhealing; note the lack of mature collagen bundles and the scattered fibil formation. The rPDGF-BB-treated wound biopsy uas taken from newly healed tissue and is entirely covered with new epithelium. Ep, epidermis, arrou, neuw collagen bundles. Original magnification, X 160.

content between treatment groups persisted and were most pronounced on day 29 (P = 0.045; Figure 3A). In healing ulcers, differences in fibroblast content between rPDGF-BB-treated and placebo groups were more pronounced (day 8, P = 0.015; day 29, P = 0.044; Figure 3B). In all ulcers as well as in only healing ulcers, increased neovessel content was observed in the rPDGF-BB treatment group, but vari-

ability precluded finding statistically significant differences (Figure 3). On day 8, a trend toward increased neovessel formation was observed in rPDGF-

BB-treated healing ulcers compared with placebo (2.88 + 0.43 versus 1.38 ± 0.31, P= 0.07; Figure 3B). Representative healing and nonhealing ulcers show differences in fibroblast and neovessel content (Figure 4). As seen by electron microscopy on day 8, fibroblasts in placebo-treated chronic ulcers were sparse and remained in a disorganized collagen-fibrin matrix. Vessels were frequently occluded (Figure 5). Increased numbers of viable fibroblasts were detected

in rPDGF-BB-treated healing wounds, compared with placebo-treated wounds on days 8 and 15 (Figure 6). These fibroblasts contained a cytoplasm with well developed rough endoplasmic reticulum and dilated cisternae. Golgi apparatus and polyribosomes were also detected. Inflammatory cells were still observed in most biopsies on days 8 and 15. In rPDGF-BBtreated ulcers, little collagen deposition was detected on day 8 (Figure 6A), in contrast to day 15 biopsies (Figure 6B), in which abundant collagen fibrillogenesis was detected.

Collagen and PC-I Content in Wounds Collagen deposition was assessed in ulcer biopsies on days 15 and 29 by using the Sirius red stain. Although some increases were consistently observed in rPDGF-BB-treated ulcers (Table 2), they were not significant. In this assessment, it was difficult to determine whether only newly formed collagen was being

1406

Pierce et al


4

-

s ^

D- ~~~ ~~~ ~~~~~~~~~~~

~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 1

1

"! 1|io-;

-.

1

Figure 8. PC-I staining of healing and nonhealing ulcers. A and B: rPDGF-BB-treated healing ulcer, C and D: placebo-treated nonhealing ulcer. Ep, epidermis. Original magnification, X32 (A and C); X320 (B and D).

Afl..

* ;..

i. I

....I

. 4' 4

Figure 9. Increased PC-I expression in rPDGF-BB-treated woundfibroblasts (A) compared with placebo-treatedfibroblasts (B) on day 15. Note the presence and increased amounts of PC-I in allfibroblasts in rPDGF-BB-treated wounds, compared with controls. Ep, epithelium. Original magnification, x160.

scored because in many cases old collagen also was present in the biopsy sample. Representative healing and nonhealing ulcers also showed differences in collagen deposition within the wound bed, with healing ulcers showing maturing collagen bundles located

under new epithelium whereas nonhealing ulcers showed only small, relatively unorganized collagen fibrils at the ulcer margin (Figure 7). Therefore, to identify only new collagen production, PC-I immunostaining was performed on samples


1407


Table 3. PC-I-Positive Fibroblast Areas in rPDGF-BB and Placebo-Treated Wounds*

Placebo rPDGF-BB P value

n

Total

5 11

26.6 ± 2.1t 31.6 ± 1.4 0.027t

PC-I-Positive Fibroblast Area (p)2 n Ulcer 1 5

23.8 30.2 ± 2.6

NA§

n

Subepidermal

5 8

27.2 ± 2.0 32.5 ± 1.6 0.035

Measurements of PC-I-positive fibroblast areas were made on fibroblasts greater than 13 p2 within the entire biopsy sample (total), within the ulcer bed (ulcer), or subjacent to the epidermis in the newly formed scar (subepidermal) fibroblasts. t Mean ± SE. t Unpaired Student's t-test. § NA, not applicable.

from treatment days 15 and 29. This antibody detects primarily newly synthesized intracellular procollagen before it is secreted and the N- and C-termini are cleaved in the pericellular space. Marked variations in PC-I-containing fibroblasts were observed between healing and nonhealing ulcers, with greater staining seen in rPDGF-BB-treated compared with placebo-treated ulcers (Figure 8). Additionally, almost all fibroblasts in rPDGF-BB-treated healing ulcers showed strong PC-I staining whereas fibroblasts in healing placebo-treated ulcers showed a weaker response (Figure 9). Thus, intense PC-I staining was closely associated with markedly enlarged fibroblasts. To further assess PC-I staining, fibroblast areas were measured by using image analysis. Fibroblasts having cytoplasmic areas greater than 13 p2 were used to eliminate small fragments of cells from the measurements. Measurements were made within the entire biopsy width, within the exposed ulcer bed, and subjacent to the newly formed epidermis (Table 3). In all three regions, rPDGF-BB-treated ulcer fibroblasts showed increases in PC-I-positive areas over placebo-treated ulcers. There were no differences within groups for measurements in the three regions. Electron microscopy of ulcers on day 29 confirmed the light microscopic observations on differences in wound collagen content (Figure 10A). rPDGF-BBtreated healing ulcers had better organized collagen fibril deposition near the perimeter of fibroblasts than ulcers that remained unhealed (Figure 1OB). Nonhealing ulcers had a loosely organized extracellular matrix containing scattered fibroblasts and few collagen fibrils. By day 29, rPDGF-BB-treated healing ulcers contained maturing collagen bundles. Fibroblasts in these ulcers continued to contain active protein synthetic machinery. Interestingly, on day 29, nonhealing rPDGF-BB-treated ulcers, like placebo-treated ulcers, continued to show inflammatory cells and a lack of protein-synthesizing fibroblasts when analyzed by electron microscopy.

Discussion Although rPDGF-BB can stimulate soft tissue repair in humans, the processes underlying this clinical effect have not been identified. In the present study, a small cohort of 20 elderly adults with chronic pressure ulcers were serially biopsied during a 28-day period of daily rPDGF-BB treatment to determine whether microscopic correlates of healing (or nonhealing) could be established. In biopsies from healing but not nonhealing ulcers, significant differences between rPDGF-BB-treated and placebo-treated ulcers were detected for fibroblast content and neovessel formation. Thus, within healing ulcers, rPDGF-BB was capable of exaggerating the healing response in a manner highly similar to what has been previously described in animal models of dermal repair.14'15 When treatment days were analyzed separately, a rPDGF-BB treatment effect was observed as early as treatment day 8, suggesting rPDGF-BB augmented the initial repair response and more rapidly induced provisional matrix deposition within the wound. In contrast, nonhealing ulcers from rPDGF-BBtreated and placebo-treated groups appeared similar histologically. Fibroblasts were less activated, and a rPDGF-BB treatment effect was not detected. The reasons for an absence of healing and an absence of a rPDGF-BB effect in these ulcers are not understood but may be due to inadequate delivery of the growth factor to the wound, inadequate pressure relief and continuing necrosis of the ulcer, or host or ulcer factors not currently recognized among nonhealing patients.24 Caution is required in the interpretation of the descriptive data from these biopsies. Treatment group sizes were small and became smaller when healing and nonhealing ulcers were analyzed separately, although each patient did have multiple sequential biopsies during the treatment phase. In addition, a 3-mm biopsy is likely an inadequate representation of the processes occurring within a 10- to 20-cm3 ulcer.

1408

Pierce et al


Figure 10. New collagen deposition in rPDGFBB-treated wounds after 28 days of treatment. A: fibroblasts (F) are contained within a dense collagen matrix Magnification, X 13,500. B: active collagen fibnillogenesis in a fibroblast with electron dense material in dilated cisternae. Pericellular deposition of new fibrils is associated with abundant collagen bundle formation. Magnification, X 21,600. Arrows, dilated cisternae; arrowheads, collagen bundles.

These factors undoubtedly accounted for some of the variability observed in the blinded photo rankings.25 In addition, the biopsies could not be used as independent predictors of the healing process but could only substantiate the macroscopic measurements. Thus, their primary utility has been to identify the potential mechanisms involved in growth factoraugmented repair and to delineate the normal healing course in chronic wounds. Recently, Herrick and colleagues26 have used this approach to better characterize venous stasis ulcers. PC-I staining is more specific for newly synthesized collagen inasmuch as it identifies only intracellular procollagen before the N- and C-termini are clipped

in the pericellular space.22 Thus PC-I-containing fibroblasts were associated with an activated, enlarged phenotype and were more prominent in rPDGF-BB-treated healing wounds. PC-I appears to be a good marker of activated fibroblasts and confirmed the morphological observations made using the Masson trichrome stain. Sirius red staining of new collagen was not as informative as the previous studies of dermal ulcer healing in the rabbit ear, in which wound boundaries are well defined, would have pre-

dicted.15 Why do nonhealing wounds fail to heal? The signals for the initiation of repair appear to involve the coordinated expression and release of growth factors


1409


that in turn stimulate proliferation, migration, and provisional matrix deposition.27-30 Diminished levels of fibronectin and growth factors have been reported in fluids obtained from chronic wounds in humans compared with acute surgical wounds,31'32 consistent with our histological observations. In addition, keratinocytes and cells within granulation tissue release growth factors, collagenases, and collagenase inhibitors, which contribute to repair.33-36 rPDGF-BB applied to nonhealing wounds appears to initiate a cascade of activities that contribute to a repair phenotype. PDGF clearly stimulates migration, activation, and proliferation of fibroblasts.5 10 Although it triggers provisional matrix production,14'15,37'38 PDGF is not known to directly induce collagen type synthesis.39 Thus, collagen deposition in rPDGF-BB-treated wounds may be induced by endogenous transforming growth factor f3 within the wound. PDGF can stimulate the production of transforming growth factor (31, as well as PDGF, in fibroblasts.' 1,40 Transforming growth factor (3 is widely distributed in healing wounds,41 further suggesting that pharmacological application of rPDGF-BB can stimulate a cascade of autocrine and paracrine growth factors that lead to an epithelialized collagen-containing scar. Whether rPDGF-BB directly or indirectly triggers neovascularization within wounds is not clear, although microvascular endothelial cells have been shown to respond to PDGF42-45 and PDGF stimulates neovessel formation within normal and ischemic wounds. 15'18

Acknowledgments We appreciate the assistance of Drs. M. Robson and R. Rudolph in providing biopsy material for the development of the staining techniques. We thank Diane Duryea, Donna Yanagihara, Joan Fare, and Judy Cribbs for outstanding technical assistance, Jennifer Keysor and Maureen Antonio for illustrations, and Joan Bennett for manuscript preparation.

References 1. Allman RM: Pressure ulcers among the elderly. N Engi J Med 1989, 320:850-853 2. Goode PS, Allman RM: The prevention and management of pressure ulcers. Med Clin N Am 1989, 73:1511 3. Seiler WO, Stahelin HB: Recent findings on decubitus ulcer pathology: implications for care. Geriatrics 1986, 41:47 4. Daltrey DC, Rhodes B, Chattwood JG: Investigation into the microbial flora of healing decubitus ulcers. J Clin Pathol 1981, 34:701

5. Deuel TF, Kawahara R, Mustoe TA, Pierce GF: Growth factors and wound healing: platelet-derived growth factor as a model cytokine. Annu Rev Med 1991, 42: 567-584 6. Pierce GF, Mustoe TA, Altrock B, Deuel TF, Thomason A: Role of platelet-derived growth factor in wound healing. J Cell Biochem 1991, 45:319-326 7. Grotendorst GR, Martin GR, Pencev D, Sodek J, Harvey AK: Stimulation of granulation tissue formation by platelet-derived growth factor in normal and diabetic rats. J Clin Invest 1985, 76:2323-2329 8. Lynch SI, Nixon JC, Colvin RB, Antoniades HN: Role of platelet-derived growth factor in wound healing: synergistic effects with other growth factors. Proc Natl Acad Sci USA 1987, 84:7696-7700 9. Sprugel KH, McPherson JM, Clowes AW, Ross R: Effects of growth factors in vivo. 1. Cell ingrowth into porous subcutaneous chambers. Am J Pathol 1987, 129: 601-613 10. Pierce GF, Mustoe TA, Senior RM, Reed J, Griffin GL, Thomason A, Deuel TF: In vivo incisional wound healing augmented by platelet-derived growth factor and recombinant c-sis gene homodimeric proteins. J Exp Med 1988, 167:974-987 11. Pierce GF, Mustoe TA, Lingelbach J, Masakowski VR, Griffin G, Senior RM, Deuel TF: Platelet-derived growth factor and transforming growth factor ,B induce in vivo and in vitro tissue repair activities by unique mechanisms. J Cell Biol 1989, 109:429-440 12. Lynch SE, Colvin RB, Antoniades NH: Growth factors in wound healing. J Clin Invest 1989, 84:640-646 13. Mustoe TA, Pierce GF, Morishima C, Deuel TF: Growth factor induced acceleration of tissue repair through direct and inductive activities in a rabbit dermal ulcer model. J Clin Invest 1991, 87:694-703 14. Pierce GF, Vande Berg J, Rudolph R, Tarpley J, Mustoe TA: PDGF-BB and TGF-p1 selectively modulate glycosaminoglycans, collagen, and myofibroblasts in excisional wounds. Am J Pathol 1991, 138:629-646 15. Pierce GF, Tarpley J, Yanagihara D, Mustoe TA, Fox GM, Thomason A: PDGF-BB, TGF-,B1, and basic FGF in dermal wound healing: neovessel and matrix formation and cessation of repair. Am J Pathol 1992, 140: 1375-1388 16. Mustoe TA, Purdy J, Gramates P, Deuel TF, Thomason A, Pierce GF: Reversal of impaired wound healing in irradiated rats by platelet-derived growth factor-BB: requirement of an active bone marrow. Am J Surg 1989, 158:345-350 17. Greenhalgh DG, Sprugel KH, Murray MJ, Ross R: PDGF and FGF stimulate wound healing in the genetically diabetic mouse. Am J Pathol 1990, 136:12351246 18. Mustoe TA, Ahn ST, Tarpley JE, Pierce GF: The effect of hypoxia on growth factor actions: differential response of basic fibroblast growth factor and plateletderived growth factor in an ischemic wound model. Wound Repair Regen, in press

1410 Pierce et al AJP December 1994, Vol. 145, No. 6

19. Robson MC, Phillips LG, Thomason A, Robson LF, GF Pierce: Platelet-derived growth factor-BB in chronic pressure ulcers. Lancet 1992, 339:23-25 20. Robson MC, Phillips LG, Thomason A, Altrock BA, Pence P, Heggers JP, Johnston AF, McHugh TP, Anthony MS, Robson LF, Odom LL, Yanagihara D, Pierce GF: Recombinant human platelet-derived growth factor-BB for the treatment of chronic pressure ulcers. Ann Plast Surg 1992, 29:193-201 21. Mustoe TA, Cutler NR, Allman RM, Goode PS, Deuel TF, Prause JA, Bear M, Serdar CM, Pierce GF: A phase 11 study to evaluate recombinant PDGF-BB in the treatment of stage III/IV pressure ulcers. Arch Surg 1994, 129:213-219 22. McDonald JA, Broekelmann TJ, Matheke ML, Crouch E, Koo M, Kuhn C: A monoclonal antibody to the carboxyterminal domain of procollagen type visualizes collagen synthesizing fibroblasts: detection of an altered fibroblast phenotype in lungs of patients with pulmonary fibrosis. J Clin Invest 1986, 78:1237-1244 23. Pierce GF, Brown D, Mustoe TA: Quantitative analysis of inflammatory cell influx, procollagen type synthesis, and collagen cross-linking in incisional wounds: influence of PDGF-BB and TGF-11 therapy. J Lab Clin Med 1991, 117:373-382 24. Falanga V, Grinnell F, Gilchrest G, Maddox YT, Moshell A: Workshop on the pathogenesis of chronic wounds. J Invest Dermatol 1994, 102:125-127 25. Vande Berg JS, Rudolph R: Pressure (decubitus) ulcer: variation in histopathology, a light and electron microscope study. Hum Pathol, 1995, in press 26. Herrick SE, Sloan P, McGurk M, Freak L, McCollum CN, Ferguson MWJ: Sequential changes in histological pattern and extracellular matrix deposition during the healing of chronic venous ulcers. Am J Pathol 1992, 141:1085-1095 27. McDonald JA, Kelley DG, Broekelmann TJ: Role of fibronectin in collagen deposition: Fab' to the gelatinbinding domain of fibronectin inhibits both fibronectin and collagen organization in fibroblast extracellular matrix. J Cell Biol 1982, 92:485-492 28. Welch MP, Odland GF, Clark RAF: Temporal relationships of F-actin bundle formation, collagen and fibronectin matrix assembly, and fibronectin receptor expression to wound contraction. J Cell Biol 1990, 110:133-145 29. Brown LF, Dubin D, Lavigne L, Logan B, Dvorak HF, Van de Water L: Macrophages and fibroblasts express embryonic fibronectins during cutaneous wound healing. Am J Pathol 1993, 142:793-801 30. Clark RAF: Regulation of fibroplasia in cutaneous wound repair. Am J Med Sci 1993, 306:42-48 31. Wysocki AB, Grinnell F: Fibronectin profiles in normal and chronic wound fluid. Lab Invest 1990, 63:825-831 32. Bennett NT, Schultz GS: Growth factors and wound healing: II. Role in normal and chronic wound healing. Am J Surg 1993, 166:74-81 33. Stricklin GP, Li L, Jancic V, Wenczak BA, Nanney LB:

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Localization of mRNAs representing collagenase and TIMP in sections of healing human burn wounds. Am J Pathol 1993, 143:1657-1666 Juhasz I, Murphy GF, Yan H-C, Herlyn M, Albelda SM: Regulation of extracellular matrix proteins and integrin cell substratum adhesion receptors on epithelium during cutaneous human wound healing in vivo. Am J Pathol 1993, 143:1458-1469 Marikovsky M, Breuing K, Liu PY, Eriksson E, Higashiyama S, Farber P, Abraham J, Klagsbrun M: Appearance of heparin-binding EGF-like growth factor in wound fluid as a response to injury. Proc Natl Acad Sci USA 1993, 90:3889-3893 Werner S, Peters KG, Longaker MT, Fuller-Pace F, Banda MJ, Williams LT: Large induction of keratinocyte growth factor expression in the dermis during wound healing. Proc Natl Acad Sci USA 1992, 89: 6896-6900 Blatti SP, Foster DN, Ranganathan G, Moses HL, Getz MJ: Induction of fibronectin gene transcription and mRNA is a primary response to growth-factor stimulation of AKR-2B cells. Proc Natl Acad Sci USA 1988, 85:1119-1123 Heldin P, Laurent TC, Heldin C-H: Effect of growth factors on hyaluronan synthesis in cultured human fibroblasts. J Biochem 1989, 258:919-922 Rossi R, Karsenty G, Roberts AB, Roche NS, Sporn MB, de Crombrugghe B: A nuclear factor 1 binding site mediates the transcriptional activation of a type collagen promoter by transforming growth factor f3. Cell 1988, 52:405-414 Paulsson Y, Hammacher A, Heldin C-H, Westermark B: Possible positive autocrine feedback in the prereplicative phase of human fibroblasts. Nature 1987, 328: 715-717 Levine JH, Moses HL, Gold Li, Nanney LB: Spatial and temporal patterns of immunoreactive transforming growth factor 31, ,B2, and f3 during excisional wound repair. Am J Pathol 1993, 143:368-380 Hermansson M, Nister M, Betsholtz C, Heldin C-H, Westermark B, Funa K: Endothelial cell hyperplasia in human glioblastoma: coexpression of mRNA for platelet-derived growth factor (PDGF) B chain and PDGF receptor suggests autocrine growth stimulation. Proc NatI Acad Sci USA 1988, 85:7748-7752 Beitz JG, Kim l-S, Calabresi P, Frackelton Jr AR: Human microvascular endothelial cells express receptors for platelet-derived growth factor. Proc Natl Acad Sci USA 1991, 88:2021-2025 Sato N, Beitz JG, Kato J, Yamamoto M, Clark JW, Calabresi P, Frackelton Jr AR: Platelet-derived growth factor indirectly stimulates angiogenesis in vitro. Am J Pathol 1993, 142:1119-1130 Khouri RK, Hong S-P, Deune EG, Tarpley JE, Song S-Z, Serdar CM, Pierce GF: De novo generation of permanent neovascularized soft tissue appendages by platelet-derived growth factor. J Clin Invest 1994, 94:1757-1763