EXPERIMENTAL
AND
MOLECULAR
PATHOLOGY
52, 132-M
(I!%@
Soluble Factor(s) in Rat Wound Fluid Inhibit Fibroblast Populated Lattice Contraction TONI RITTENBERG, Wound Healing Laboratory,
D. ANDREW
R. BURD, AND H. PAUL EHRLICH
Shriner Burns Institute, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts Received July 13, 1989
Closure of full-thickness open wounds in loose-skinned animals is accomplished by wound contraction. Fibroblast-populated collagen lattice (FPCL) contraction is an in vitro model for studying wound contraction. Fibroblasts suspended in a collagen matrix reorient the surrounding collagen fibers, resulting in a reduction in the size of the FPCL. The organization of collagen fibers by fibroblast-generated forces produces lattice contraction. An open wound in a rat begins to show contraction by 3 days, and its size will be reduced by 50% at 7 days. Fluid from 3- and ‘I-day-old rat wounds was examined for its ability to affect in vitro lattice contraction. Wound fluid was found to inhibit lattice contraction. The fractions which inhibit lattice contraction had molecular weights ranging between 10,000 and 20,000, as revealed by molecular sieve chromatography, and a high positive charge, as demonstrated by ion exchange chromatography. The factor(s) was only slightly affected by added indomethacin in the FPCL contraction model. This suggests a mechanism independent of the generation of prostaglandins. The factor(s) was tested in an ATP-induced model of fibroblast contraction where it was shown to be ineffective at altering cell contraction. The factor(s) did, however, prevent cell spreading and elongation on glass surfaces. Wound fluid has a factor(s) which hinders tibroblast spreading and elongation and which inhibits FPCL contraction. 8 1990 Academic press, Inc.
INTRODUCTION Cell-mediated organization of connective tissue matrix occurs under such conditions as growth, development, and wound healing. Wound contraction in the closure of full-thickness open wounds involves the organization of the newly deposited connective tissue (Ehrlich, 1988a). The predominant method of open wound healing in loose-skinned animals such as rats is wound contraction (Billingham and Medawar, 1955). Wound contraction is effected by forces generated by the resident fibroblasts in granulation tissue of healing wounds. These forces pull the surrounding normal skin over the defect. The rapid closure of open wounds in rats by wound contraction raises the possibility that some soluble factor(s) present in granulation tissue enhances the contractile process. Soluble factors released from rat wounds were examined in an in vitro model to test this hypothesis. The fibroblast-populated collagen lattice (FPCL) is manufactured from monolayer cultured fibroblasts, serum-rich culture medium, and soluble native collagen. When these components are mixed together rapidly, under physiological conditions, the collagen polymerizes quickly and traps fibroblasts in the newly formed matrix. The cells initially spread and elongate in the collagen lattice and they subsequently attempt to move through the lattice. The forces of fibroblast locomotion generate the power to reorganize the collagen matrix which causes the collagen fibers to become compact, reducing the size of the FPCL (Ehrlich, 1988b). This reduction in size is referred to as lattice contraction and it is assumed that the cell-matrix interactions studied in vitro are similar to those 132 001448OO/!XI $3.00 Copyright 0 1990 by Academic Press. Inc. AU rights of reproduction in any form reserved.
SOLUBLE
FACTOR(S)
IN
RAT
WOUND
FLUID
133
seen in in vivo wound contraction. Soluble factors released from granulation tissue can be isolated from subcutaneously implanted stainless steel mesh wound chambers (Hunt et al., 1967). Wound fluid, from the initial or lag phase of healing, can be collected from these chambers as early as 3 days after implantation. By 7 days, the wound has entered the proliferative phase of healing, and at this stage, an excised wound in a rat is expected to have contracted to about 50% of its original size. If soluble factors which stimulate wound contraction are present, one would expect to find them in the 7-day wound fluid. However, the FPCL has demonstrated that wound fluid inhibits lattice contraction. Wound fluid harvested from 3-day-old wound chambers also contains a soluble factor(s) which inhibits lattice contraction. This suggests that an inhibitor of lattice contraction may be produced by early inflammatory cells. An inhibitory factor(s) in wound fluid was isolated by molecular sieve and ion exchange chromatography. Three- and seven-day-old wound fluid seem to share the same inhibitory factor(s). The isolated factor(s) was tested for its capacity to affect ATP-induced cell contraction. When permeabilized fibroblast preparations from monolayer are treated with ATP and cofactors, the actin filaments aggregate, and the size of the tibroblasts is reduced (Goldman et al., 1976). PGE,, which enhances the elevation of intracellular levels of CAMP, has also been shown to inhibit lattice contraction and ATP-induced cell contraction (Ehrlich et al., 1986). Wound fluid factor(s) was tested for its ability to promote the synthesis and release of PGE*. MATERIALS
AND METHODS
Rats and Wound Fluid Three male 300- to 350-g SpragueDawley rats (Taconic Farms, Kingston, NY) were anesthetized with ether. Their backs were washed with 70% ethanol, and an incision was made to subcutaneously accommodate a 1.2 X 6.0-cm stainless steel mesh cylindrical chamber. The incision was closed using stainless steel sutures (Ehrlich, 1984). The rats were returned to their cages where they were allowed food and water ad lib., and they continued to gain weight and were active and alert. Three and seven days after implantation, wound fluid was removed from the chambers using a hypodermic syringe under sterile conditions. Between 2.5 and 3.5 ml of straw-colored wound fluid was removed and centrifuged to remove cells and debris. The cleared supematant was saved for study. Column
Chromatography
Pooled wound fluid was concentrated by vacuum dialysis; applied to an ACA 54 column, 1.6 x 80 cm, (LKB, Gaithersburg, MD); equilibrated at 4°C in physiological buffered saline (PBS); and eluted. The eluted fractions were assayed in an FPCL contraction model. Molecular weight standards (bovine serum albumin, egg albumin, carbonic anhydrase, and lysozyme) in PBS were used to standardize the column. Ion exchange chromatography was performed using the Mono Q column (Pharmacia, Piscataway, NJ). The column was equilibrated at room temperature with 40 mM sodium phosphate buffer, pH 7.3, at a flow rate of 1 ml/min. The sample was pumped into the column in the same buffer and eluted with a 0 to 1 M NaCl gradient over 28 min. Two-min 2-ml fractions were collected.
134 Fibroblast-Populated
RITTENBERG,
BURD,
AND
EHRLICH
Collagen Lattices
Normal dermal fibroblasts were a gift from Dr. David Wyler, Department of Medicine, Tufts New England Medical School. Fibroblasts were grown in Dulbecco’s modification of Eagle’s medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS) and 10 kg/ml gentamycin. All cells in these experiments were maintained in a 37°C incubator with 5% CO* and 95% air in a water-saturated atmosphere. Cells were passed by trypsinization when they reached confluence and were replated at a ratio of 2: 1. Cells in these experiments were between their 8th and 15th passages. The collagen used to make FPCL was taken from rat tail tendon by acetic acid extraction and sodium chloride salt precipitation (Buttle and Ehrlich, 1983). The purified collagen was dialyzed, frozen, lypholized, weighed, and stored at 5 mg/ml in 1 mM HCI at 4°C as a viscous solution. Trypsin-freed monolayer fibroblasts, 0.5 ml, 160,000 to 200,000 cells/ml, 0.8 ml of DMEM with serum, 0.2 ml of wound fluid fraction in PBS, and 0.5 ml of collagen solution were rapidly mixed in a 35-mm petri dish and placed in the incubator. The collagen polymerized in less than 90 sec. The diameter of each FPCL was measured daily with a ruler to the nearest 0.5 mm to determine lattice contraction. Part of the FPCL was transferred to glass coverslips for histology, to investigate possible changes in cell shape after treatment with wound fluid fractions. Cell were viewed under phase-contrast optics. Cell Contraction Approximately 5000 tibroblasts in 2 ml of DMEM were transferred onto a sterilized 22 x 22-mm glass coverslip placed in a 35-mm petri dish. All dishes were incubated for 48 hr. Cells were permeabilized by glycerol using the technique of Goldman et al. (1976), modified by Kreiss and Birchmeirer (1980). Coverslips were rinsed with PBS and then 50% (v/v) glycerol in 50 mM KCl, 5 mM MgCl,, 10 mM Tris-HCl, pH 7.5, buffer for 30 min at room temperature. Separate 30-min washes were made with 25, 12.5, and 5% glycerol in the same buffer. Cells remained in the 5% glycerol buffer until testing, usually before 60 min. Half the glycerol-treated coverslips were immediately fixed for 5 min in 4% paraformaldehyde in PBS buffered to pH 7.5 and rinsed three times in PBS. ATP was added to the other half. For cell contraction to commence, the 5% glycerol buffer was replaced in a contraction buffer of 1 mM ATP in 5 mM MgCl,, 30 miI4 KCI, 0.1 mM CaCl,, and 10 mM Tris-HCl, pH 7.0. After a lo-min incubation at room temperature the coverslips were fixed as described above. Changes in cell length were documented by measuring cell lengths using an eyepiece micrometer. Fifty cells for each of four coverslip preparations were measured and recorded. The fixed cells on coverslips were incubated at room temperature for 30 min with rhodamine phalloidin (Rh-phalloidin; Molecular Probes, Inc., Eugene, OR) diluted 1:200 with PBS. Rh-phahoidin is a fluorescently labeled mushroom toxin which specitically binds to and stains F-actin filaments (Barak et al., 1980; Weiland and Faulstich, 1978). After treatment, the coverslips were washed three times in PBS and mounted in 9:l PBS glycerol solution. Slides were viewed by epifluorescence with appropriate rhodamine filters.
SOLUBLE
FACTOR(S)
IN RAT WOUND
FLUID
135
RESULTS Rat wound fluid was harvested and pooled 3 and 7 days after implantation of stainless steel mesh wound chambers and centrifuged to remove cells and particulate matter prior to filter sterilization. All wound fluid collections were visibly free of infection. Sterile wound fluid was introduced to the manufactured FPCL at 10% v/v, that is, 0.2 ml of wound fluid was included in the 2 ml FPCL containing 80,000 cells. Wound fluid was observed to inhibit lattice contraction by 10%. Since wound fluid is a complex mixture, perhaps rich in antagonists and inhibitors, its fractionation may isolate a factor(s) producing greater inhibition of lattice contraction. Wound fluid from 3- and 7-day chambers was concentrated from 6 and 9 ml to 2 and 3 ml by vacuum dialysis and applied to a molecular sieve chromatography column. As shown in Fig. 1, wound fluid fractionation produced two major optical density peaks, a breakthrough, the void volume peak, and an included peak, of between 40,000 and 70,000 MW. All fractions in PBS were filter sterilized prior to testing at 10% v/v in FPCL contraction assay. Material which eluted between 10,000 and 20,000 MW inhibited lattice contraction (results not shown). These inhibitory fractions were identical whether fractionated from 3- or 7-day-old wound fluid. They were pooled and dialyzed into Mono Q column starting buffer prior to elution by Mono Q ion exchange chromatography. After the eluted fractions were filter sterilized, the Mono Q fractions were tested with FPCL (Fig. 2). Fractionated wound fluid from 3-day-old implants was tested; the results are presented in Fig. 2a. Three-day wound fluid fractions that displayed a sharp inhibitory activity appeared at fraction 10. The inhibitory effect of fraction 10 on lattice contraction was much less dramatic in 7-day fluid. The
FRACTION
NUMBER
1. Fractionation of ‘I-day rat wound fluid by molecular sieve chromatography. Concentrated wound fluid was applied to an ACA 54 column and the elution was monitored by absorption at Ezm Fractions were tested for additional effects on lattice contraction. Fractions 15-19 were pooled and further fractionated by ion exchange chromatography. Calibration of the column used molecular weight standards: a = 68,000; b = 43,000; c = 30,000; and d = 14,600. FIG.
136
RITTENBERG,
BURD,
AND
EHRLICH
250
a 200
i
150
d P IL I; (
100
AT
2 DAYS
E a AT
5 DAYS
50 AT
0
4 DAYS
5 MONO
10 0 FRACTION
1’5
0 MONO
0
FRACTION
FIG. 2. Fractionations of rat wound fluid effect on lattice contraction. (a) Three-day rat wound fluid was chromatographed and fractionated by ACA 54 column molecular sieve chromatography. The fractions between 10,000 and 20,000 MW were pooled and fractionated by ion exchange chromatography using a Mono Q column. Collected fractions were tested by a lattice contraction assay. Insert shows two FPCL at 4 days which were incubated with fraction 7 or 11. (b) Seven-day wound fluid initially fractionated by size was fractionated by Mono Q ion exchange chromatography. The fractions were tested at 10% v/v in FPCL. Lattice size was recorded at 2 and 5 days. The larger the area, the greater the inhibition to FPCL contraction.
fractionation of wound fluid derived from 7-day-old implants is shown in Fig. 2B. Like 3-day wound fluid, its strongest inhibitory activity is in fractions strongly bound to column fractions 10 and 11. The inhibitory effect of 7-day wound fluid is a broader peak of fractions 6 through 11 than fractions 8 to 11 from 3-day wound fluid. Wound fluid from early 3-day implants and 7-day implants both showed a low molecular weight inhibitory activity which bound strongly to the charged ion exchange resin. The inhibitory activity isolated at the two time points appears to be the same. The mechanism of inhibited FPCL contraction may be related to actin-myosin filament sliding. This possibility can be explored by ATP-induced cell contraction
SOLUBLE
FACTOR(S)
IN RAT WOUND
137
FLUID
and the aggregation of actin filaments. Human fibroblasts plated on glass coverslips were incubated for 48 hr with individual fractions 8 through 11 from 3-day wound fluid, as shown in Fig. 2a. Harvested coverslips were permabilized with glycerol and half were fixed and stained with Rh-phalloidin. The other half was incubated with ATP and cofactors for 10min, then fixed, and stained. Cell length, determined by stained actin filaments, was measured (see Table I).While these fractions did not inhibit cell contraction, they do prevent fibroblasts from spreading and elongating on glass surfaces. Cell elongation was inhibited between 29 and 45%. This implies that a mechanism of inhibited lattice contraction may be related to impaired cell spreading in the collagen matrix. Identical results were produced by 7-day wound fluid fractions (results not shown). Fibroblasts in collagen lattices were observed. When FPCL were viewed by phase-contrast light microscopy, the variety of cell lengths within was great and no clear observable difference was found between treated and untreated FPCL. Interleukin-1 (IL-l), a well-characterized product of macrophages, is also a known modulator of lattice contraction. It has been shown to inhibit lattice contraction (Ehrlich and Wyler, 1983). Adding indomethacin with IL-1 to the FPCL blocked IL-l inhibitory activity. Indomethacin added with 7-day wound fluid fractions to the FPCL produced a minor reduction in the potency of wound fluid inhibitory activity, as shown in Table II. In no case did it restore lattice contraction to normal. It appears, therefore, that the production and release of prostaglandins may have a minimal effect on the inhibition of lattice contraction, while other components of wound fluid are more potent inhibitors of lattice contraction. DISCUSSION Implanted stainless steel wound chambers allow the sampling of soluble factors released in the interstital spaceduring wound repair. It would be expected that the major cell populations within the chamber at 3 days would be inflammatory cells such as neutrophils and macrophages (Edwards and Dunphy, 1958). By Day 7 the number of neutrophils will be reduced, and macrophages and fibroblasts will become the predominant cell types. Since the effects of 3- and 7-day wound fluid are similar, one could postulate that macrophages are common to both. One TABLE I ATP-Induced Cell Contraction Cell length with ATP Treatment
0 time
10 min
PBS F-S F-9 F-10 F-11
115.4 72.4 81.7 63.0 62.6
14.5 10.8 8.9 9.8 9.2
% Contraction 87 85 89 85 85
Note. Fibroblasts growing on glass coverslips were incubated with Mono Q fractions 8 through 11 or PBS for 48 hr and then treated with glycerol. Half the coverslips were ftxed and stained, and their cell lengths were measured. The other coverslips were incubated with ATP and cofactors for 10 mitt, fixed, and stained, and the cell lengths were measured. ATP-induced actin-myosin filament sliding produced cell contraction.
138
RITTENBERG,
BURD, AND EHRLICH
TABLE II Addition of Indomethacin to Fibroblast-Populated
Collagen Lattice
Lattice area (mm2) Treatment PBS PBS F-8 F-8 F-9 F-9 F-10 F-10 F-11 F-11
Indomethacin
Day 2
Change (%)
Day 4
Change (o/o)
-
106 95 189 171 237 157 227 259 207 186
11 10 20 18 25 16 24 27 22 19
64 58 143 128 165 130 234 254 207 174
7 6 15 13 17 14 24 26 22 18
+ + + + +
Note. Mono Q fractions 8 to 11 from 3-day rat wound fluid initially fractionated by size were included in the manufacture of FPCL at 10% v/v. Indomethacin at 10 &ml was included in the manufacture of half the FPCL. Lattice size was measured daily for 4 days.
speculation is that macrophages produce and release soluble factors to alter fibroblast-generated contractile forces. One such factor, IL-l, has been studied in lattice contraction (Ehrlich and Wyler, 1983). In that study, soluble factors released by experimental mouse granulomas incubated for 16 hr in tissue culture were shown to inhibit lattice contraction. The lattice contraction inhibitory activity released by chronic inflammatory cells was lost when indomethacin which blocks cyclooxygenase activity was included in the manufacture of FPCL. In contrast, the results reported here show that indomethacin had a modest antagonistic effect on inhibition induced by wound fluid fractions and only slightly restored lattice contraction. For example, fraction 11 at 4 days inhibited lattice contraction. The area of the control lattice was reduced to 7% of its initial size; the addition of fraction 11 reduced it to only 22%. Adding indomethacin increased lattice contraction from 22 to 18%, only a 4% improvement. Indomethacin did not restore lattice contraction to anywhere near normal, however. That would require the lattice to contract by 93% to 7% of its initial area. It appears that one of the inhibitory factors in wound fluid promotes the synthesis and release of prostaglandins such as IL-l. Other factor(s) are present, however, and they block lattice contraction by a mechanism unrelated to the release of prostaglandins. The inhibitor(s) of lattice contraction has a molecular weight between 10,000 and 20,000. It is strongly bound to the Mono Q column and requires more than 700 mM NaCl to displace it from the charged resin. It appears to differ from IL-1 because of biological considerations. In unpublished work, IL-I and PGEz included in monolayer cultures caused inhibited ATP-induced cell contraction. The mechanism for this inhibition appears related to the accumulation of intracellular CAMP induced by PGE2 (Ehrlich and Griswold, 1984; Ehrlich et al., 1989). The other mechanism of wound fluid inhibitory factor(s) appears to be independent of PGE,-induced CAMP production. It appears that wound fluid factor(s) inhibits cell spreading. Cell spreading requires actin-myosin filaments. Its modulation may be related to the mechanism responsible for inhibited lattice contraction. The identity of the mechanism of inhibited lattice contraction is unknown. Work to identify it continues.
SOLUBLE
FACTOR(S)
IN RAT WOUND
FLUID
139
These results support the possibility that macrophages produce factors which irdluence wound contraction. Macrophages are required for fibroplasia, and fibroblasts produce the force required for wound contraction. The initial speculation was that macrophages would produce factors to promote wound contraction. Instead, soluble factor(s), apparently released by macrophages, inhibit wound contraction. This is a puzzling result. Some work exists which links macrophages and wound contraction. Cortisone therapy inhibits wound contraction (Hunt et al., 1969). Adding vitamin A promotes fibroplasia and collagen synthesis in the presence of cortisone, but does not restore wound contraction. Tetracholorodecaoxygen (TCDO), a compound which reportedly enhances macrophage migration at the site of damage will stimulate wound contraction in open wounds in cortisone-treated rats (Hatz et al., in press). TCDO reportedly promotes macrophage accumulation at the wound site (Hinz et al., 1986). It may be that adding TCDO restores macrophage infiltration, but the responding macrophages are different from those invading the dead space of a normal rat wound. Macrophages stimulated by TCDO to infiltrate wounds may produce products and factors different from those of macrophages produced by acute inflammation. The macrophages exposed to TCDO may produce a factor(s) which promotes lattice contraction. The concept that wound fluid contains inhibitors of wound contraction suggests that if these factors could be neutralized, wound contraction might proceed faster. In some special cases enhanced wound contraction would be desirable for wound closure. ACKNOWLEDGMENTS The authors thank Joseph B. M. Rajaratnam, Alice Gosselin, and Frances M. Mascolo for help in preparing this manuscript. This work was supported by grants from the Shriners of North America and National Institutes of Health GM-32705.
REFERENCES BARAK, L. S., YOCUM, R. R., NOJHWAGEL, E. A., and WEBB, W. W. (1980). Fluorescence staining of the actin cytoskeleton in living cells with 7-nitrobenz-2-oxa-1,3-diazole phalloidin. Proc. Natl. Acad. Sci. USA II, 98&984. BELL, E., IVARSSON, B., and MERRILL, C. (1979). Production of a tissue-like structure by contraction of a collagen lattice by human tibroblasts of diierent proliferative potential in vitro. Proc. Natl. Acad. Sci. USA 76, 1274-1278. BILLINGHAM, R. E., and MEDAWAR, P. B. (1955). Contracture and intussusceptive growth in the healing of extensive wounds in mammalian skin. J. Anat. 89, 114-l 19. BUTTLE, D. B., and EHRLICH, H. P. (1983). Comparative studies of collagen lattice contraction utilizing a normal and transformed cell line. J. Cell. Physiol. 116, 159-168. EDWARDS, L. C., and DUNPHY, J. E. (1958). Wound Healing. II. Injury and abnormal repair. N. Engl. 1. Med. 259, 275-284. EHRLICH, H. P. (1984). Anti-intlammatory drugs in vascular response to bum injury. J. Trauma 24, 311-318. EHRLICH, H. P. (1988a). The role of connective tissue matrix in wound healing. In “Growth Factors and Other Aspects of Wound Healing” (A. Barbul, E. Pines, M. Caldwell, and T. K. Hunt, Eds.). A. R. Liss, New York. EHRLICH, H. P. (1988b). Wound closure: Evidence of cooperation between tibroblasts and collagen matrix. Eye 2, 149-157. EHRLICH, H. P., BUTTLE, D. J., and BERNANKE, D. H. (1989). Physiological variables affecting collagen lattice contraction by human dermal tibroblasts. Exp. Mol. Pathol. 50, 220-229. EHRLICH, H. P., and GRISWOLD, T. R. (1984). Epidermolysis bullosa dystrophica recessive tibro-
140
RITTENBERG,
BURD, AND EHRLICH
blasts produce increased concentrations of CAMP within a collagen matrix. J. Invest. Dermatol. 83, 330-333. EHRLICH, H. P., GRISWOLD, T. R., and RAJARATNAM, J. B. M. (1986). ATP-induced cell contraction with epidermolysis bullosa dystrophica recessive and normal dermal tibroblasts. J. Invest. Dermatol. 86, !46-100. EHRLICH, H. P., and WYLER, D. J. (1983). Fibroblast contraction of collagen lattices in vitro: Inhibition by chronic infIammatory cell mediators. J. Cell. Physiol. 116, 345-351. GOLDMAN, R. D., SCHLOSS, J. A., and STARGER, J. M. (1976). Organizational changes of actin-like microfilaments during cell movement. In “Cell Mobility” (R. D. Goldman, J. Pollard, and J. Rosenbaum, Eds.). Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. HATZ, R. A., KELLEY, S. F., and EHEUCH, H. P. The tetrachlorodecaoxygen complex reverses the effect of cortisone on wound healing. Plast. Reconstr. Surg., in press. HINZ, J., HAUTZINGER, H., and STAHL, K.-W. (1986). Rationale for and results from a randomized double-blind trial of tetrachlorodecaoxygen, an ion complex in wound healing. Lancet 8485, 825828. HUNT, T. K., T~OMEY, P. ZEDENFELT, B., and DUNPHY, J. E. (1967). Respiratory gas tensions and pH in healing wounds. Amer. J. Surg. 114, 302-307. HUNT, T. K., EHRLICH, H. P., GARCIA, J. A., and DUNPHY, J. E. (1%9). Effect of vitamin A on reversing the inhibitory effect of cortisone on healing of open wounds in animals and man. Ann. Surg. 170, 633+tO. Kamss, T. E., and BIRCHMEIRER, W. (1980). Stress fiber sarcomeres of tibroblasts are contractile. Cell 22, 555-561. WEILAND, T., and FAULSTICH, H. (1978). Amatoxins, phallotoxin, phallolysin, and atamanide: The biologically active components of poisonous Amanita mushrooms. CRC Crit. Rev. Biochem. 5, 185-260.
AND
MOLECULAR
PATHOLOGY
52, 132-M
(I!%@
Soluble Factor(s) in Rat Wound Fluid Inhibit Fibroblast Populated Lattice Contraction TONI RITTENBERG, Wound Healing Laboratory,
D. ANDREW
R. BURD, AND H. PAUL EHRLICH
Shriner Burns Institute, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts Received July 13, 1989
Closure of full-thickness open wounds in loose-skinned animals is accomplished by wound contraction. Fibroblast-populated collagen lattice (FPCL) contraction is an in vitro model for studying wound contraction. Fibroblasts suspended in a collagen matrix reorient the surrounding collagen fibers, resulting in a reduction in the size of the FPCL. The organization of collagen fibers by fibroblast-generated forces produces lattice contraction. An open wound in a rat begins to show contraction by 3 days, and its size will be reduced by 50% at 7 days. Fluid from 3- and ‘I-day-old rat wounds was examined for its ability to affect in vitro lattice contraction. Wound fluid was found to inhibit lattice contraction. The fractions which inhibit lattice contraction had molecular weights ranging between 10,000 and 20,000, as revealed by molecular sieve chromatography, and a high positive charge, as demonstrated by ion exchange chromatography. The factor(s) was only slightly affected by added indomethacin in the FPCL contraction model. This suggests a mechanism independent of the generation of prostaglandins. The factor(s) was tested in an ATP-induced model of fibroblast contraction where it was shown to be ineffective at altering cell contraction. The factor(s) did, however, prevent cell spreading and elongation on glass surfaces. Wound fluid has a factor(s) which hinders tibroblast spreading and elongation and which inhibits FPCL contraction. 8 1990 Academic press, Inc.
INTRODUCTION Cell-mediated organization of connective tissue matrix occurs under such conditions as growth, development, and wound healing. Wound contraction in the closure of full-thickness open wounds involves the organization of the newly deposited connective tissue (Ehrlich, 1988a). The predominant method of open wound healing in loose-skinned animals such as rats is wound contraction (Billingham and Medawar, 1955). Wound contraction is effected by forces generated by the resident fibroblasts in granulation tissue of healing wounds. These forces pull the surrounding normal skin over the defect. The rapid closure of open wounds in rats by wound contraction raises the possibility that some soluble factor(s) present in granulation tissue enhances the contractile process. Soluble factors released from rat wounds were examined in an in vitro model to test this hypothesis. The fibroblast-populated collagen lattice (FPCL) is manufactured from monolayer cultured fibroblasts, serum-rich culture medium, and soluble native collagen. When these components are mixed together rapidly, under physiological conditions, the collagen polymerizes quickly and traps fibroblasts in the newly formed matrix. The cells initially spread and elongate in the collagen lattice and they subsequently attempt to move through the lattice. The forces of fibroblast locomotion generate the power to reorganize the collagen matrix which causes the collagen fibers to become compact, reducing the size of the FPCL (Ehrlich, 1988b). This reduction in size is referred to as lattice contraction and it is assumed that the cell-matrix interactions studied in vitro are similar to those 132 001448OO/!XI $3.00 Copyright 0 1990 by Academic Press. Inc. AU rights of reproduction in any form reserved.
SOLUBLE
FACTOR(S)
IN
RAT
WOUND
FLUID
133
seen in in vivo wound contraction. Soluble factors released from granulation tissue can be isolated from subcutaneously implanted stainless steel mesh wound chambers (Hunt et al., 1967). Wound fluid, from the initial or lag phase of healing, can be collected from these chambers as early as 3 days after implantation. By 7 days, the wound has entered the proliferative phase of healing, and at this stage, an excised wound in a rat is expected to have contracted to about 50% of its original size. If soluble factors which stimulate wound contraction are present, one would expect to find them in the 7-day wound fluid. However, the FPCL has demonstrated that wound fluid inhibits lattice contraction. Wound fluid harvested from 3-day-old wound chambers also contains a soluble factor(s) which inhibits lattice contraction. This suggests that an inhibitor of lattice contraction may be produced by early inflammatory cells. An inhibitory factor(s) in wound fluid was isolated by molecular sieve and ion exchange chromatography. Three- and seven-day-old wound fluid seem to share the same inhibitory factor(s). The isolated factor(s) was tested for its capacity to affect ATP-induced cell contraction. When permeabilized fibroblast preparations from monolayer are treated with ATP and cofactors, the actin filaments aggregate, and the size of the tibroblasts is reduced (Goldman et al., 1976). PGE,, which enhances the elevation of intracellular levels of CAMP, has also been shown to inhibit lattice contraction and ATP-induced cell contraction (Ehrlich et al., 1986). Wound fluid factor(s) was tested for its ability to promote the synthesis and release of PGE*. MATERIALS
AND METHODS
Rats and Wound Fluid Three male 300- to 350-g SpragueDawley rats (Taconic Farms, Kingston, NY) were anesthetized with ether. Their backs were washed with 70% ethanol, and an incision was made to subcutaneously accommodate a 1.2 X 6.0-cm stainless steel mesh cylindrical chamber. The incision was closed using stainless steel sutures (Ehrlich, 1984). The rats were returned to their cages where they were allowed food and water ad lib., and they continued to gain weight and were active and alert. Three and seven days after implantation, wound fluid was removed from the chambers using a hypodermic syringe under sterile conditions. Between 2.5 and 3.5 ml of straw-colored wound fluid was removed and centrifuged to remove cells and debris. The cleared supematant was saved for study. Column
Chromatography
Pooled wound fluid was concentrated by vacuum dialysis; applied to an ACA 54 column, 1.6 x 80 cm, (LKB, Gaithersburg, MD); equilibrated at 4°C in physiological buffered saline (PBS); and eluted. The eluted fractions were assayed in an FPCL contraction model. Molecular weight standards (bovine serum albumin, egg albumin, carbonic anhydrase, and lysozyme) in PBS were used to standardize the column. Ion exchange chromatography was performed using the Mono Q column (Pharmacia, Piscataway, NJ). The column was equilibrated at room temperature with 40 mM sodium phosphate buffer, pH 7.3, at a flow rate of 1 ml/min. The sample was pumped into the column in the same buffer and eluted with a 0 to 1 M NaCl gradient over 28 min. Two-min 2-ml fractions were collected.
134 Fibroblast-Populated
RITTENBERG,
BURD,
AND
EHRLICH
Collagen Lattices
Normal dermal fibroblasts were a gift from Dr. David Wyler, Department of Medicine, Tufts New England Medical School. Fibroblasts were grown in Dulbecco’s modification of Eagle’s medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS) and 10 kg/ml gentamycin. All cells in these experiments were maintained in a 37°C incubator with 5% CO* and 95% air in a water-saturated atmosphere. Cells were passed by trypsinization when they reached confluence and were replated at a ratio of 2: 1. Cells in these experiments were between their 8th and 15th passages. The collagen used to make FPCL was taken from rat tail tendon by acetic acid extraction and sodium chloride salt precipitation (Buttle and Ehrlich, 1983). The purified collagen was dialyzed, frozen, lypholized, weighed, and stored at 5 mg/ml in 1 mM HCI at 4°C as a viscous solution. Trypsin-freed monolayer fibroblasts, 0.5 ml, 160,000 to 200,000 cells/ml, 0.8 ml of DMEM with serum, 0.2 ml of wound fluid fraction in PBS, and 0.5 ml of collagen solution were rapidly mixed in a 35-mm petri dish and placed in the incubator. The collagen polymerized in less than 90 sec. The diameter of each FPCL was measured daily with a ruler to the nearest 0.5 mm to determine lattice contraction. Part of the FPCL was transferred to glass coverslips for histology, to investigate possible changes in cell shape after treatment with wound fluid fractions. Cell were viewed under phase-contrast optics. Cell Contraction Approximately 5000 tibroblasts in 2 ml of DMEM were transferred onto a sterilized 22 x 22-mm glass coverslip placed in a 35-mm petri dish. All dishes were incubated for 48 hr. Cells were permeabilized by glycerol using the technique of Goldman et al. (1976), modified by Kreiss and Birchmeirer (1980). Coverslips were rinsed with PBS and then 50% (v/v) glycerol in 50 mM KCl, 5 mM MgCl,, 10 mM Tris-HCl, pH 7.5, buffer for 30 min at room temperature. Separate 30-min washes were made with 25, 12.5, and 5% glycerol in the same buffer. Cells remained in the 5% glycerol buffer until testing, usually before 60 min. Half the glycerol-treated coverslips were immediately fixed for 5 min in 4% paraformaldehyde in PBS buffered to pH 7.5 and rinsed three times in PBS. ATP was added to the other half. For cell contraction to commence, the 5% glycerol buffer was replaced in a contraction buffer of 1 mM ATP in 5 mM MgCl,, 30 miI4 KCI, 0.1 mM CaCl,, and 10 mM Tris-HCl, pH 7.0. After a lo-min incubation at room temperature the coverslips were fixed as described above. Changes in cell length were documented by measuring cell lengths using an eyepiece micrometer. Fifty cells for each of four coverslip preparations were measured and recorded. The fixed cells on coverslips were incubated at room temperature for 30 min with rhodamine phalloidin (Rh-phalloidin; Molecular Probes, Inc., Eugene, OR) diluted 1:200 with PBS. Rh-phahoidin is a fluorescently labeled mushroom toxin which specitically binds to and stains F-actin filaments (Barak et al., 1980; Weiland and Faulstich, 1978). After treatment, the coverslips were washed three times in PBS and mounted in 9:l PBS glycerol solution. Slides were viewed by epifluorescence with appropriate rhodamine filters.
SOLUBLE
FACTOR(S)
IN RAT WOUND
FLUID
135
RESULTS Rat wound fluid was harvested and pooled 3 and 7 days after implantation of stainless steel mesh wound chambers and centrifuged to remove cells and particulate matter prior to filter sterilization. All wound fluid collections were visibly free of infection. Sterile wound fluid was introduced to the manufactured FPCL at 10% v/v, that is, 0.2 ml of wound fluid was included in the 2 ml FPCL containing 80,000 cells. Wound fluid was observed to inhibit lattice contraction by 10%. Since wound fluid is a complex mixture, perhaps rich in antagonists and inhibitors, its fractionation may isolate a factor(s) producing greater inhibition of lattice contraction. Wound fluid from 3- and 7-day chambers was concentrated from 6 and 9 ml to 2 and 3 ml by vacuum dialysis and applied to a molecular sieve chromatography column. As shown in Fig. 1, wound fluid fractionation produced two major optical density peaks, a breakthrough, the void volume peak, and an included peak, of between 40,000 and 70,000 MW. All fractions in PBS were filter sterilized prior to testing at 10% v/v in FPCL contraction assay. Material which eluted between 10,000 and 20,000 MW inhibited lattice contraction (results not shown). These inhibitory fractions were identical whether fractionated from 3- or 7-day-old wound fluid. They were pooled and dialyzed into Mono Q column starting buffer prior to elution by Mono Q ion exchange chromatography. After the eluted fractions were filter sterilized, the Mono Q fractions were tested with FPCL (Fig. 2). Fractionated wound fluid from 3-day-old implants was tested; the results are presented in Fig. 2a. Three-day wound fluid fractions that displayed a sharp inhibitory activity appeared at fraction 10. The inhibitory effect of fraction 10 on lattice contraction was much less dramatic in 7-day fluid. The
FRACTION
NUMBER
1. Fractionation of ‘I-day rat wound fluid by molecular sieve chromatography. Concentrated wound fluid was applied to an ACA 54 column and the elution was monitored by absorption at Ezm Fractions were tested for additional effects on lattice contraction. Fractions 15-19 were pooled and further fractionated by ion exchange chromatography. Calibration of the column used molecular weight standards: a = 68,000; b = 43,000; c = 30,000; and d = 14,600. FIG.
136
RITTENBERG,
BURD,
AND
EHRLICH
250
a 200
i
150
d P IL I; (
100
AT
2 DAYS
E a AT
5 DAYS
50 AT
0
4 DAYS
5 MONO
10 0 FRACTION
1’5
0 MONO
0
FRACTION
FIG. 2. Fractionations of rat wound fluid effect on lattice contraction. (a) Three-day rat wound fluid was chromatographed and fractionated by ACA 54 column molecular sieve chromatography. The fractions between 10,000 and 20,000 MW were pooled and fractionated by ion exchange chromatography using a Mono Q column. Collected fractions were tested by a lattice contraction assay. Insert shows two FPCL at 4 days which were incubated with fraction 7 or 11. (b) Seven-day wound fluid initially fractionated by size was fractionated by Mono Q ion exchange chromatography. The fractions were tested at 10% v/v in FPCL. Lattice size was recorded at 2 and 5 days. The larger the area, the greater the inhibition to FPCL contraction.
fractionation of wound fluid derived from 7-day-old implants is shown in Fig. 2B. Like 3-day wound fluid, its strongest inhibitory activity is in fractions strongly bound to column fractions 10 and 11. The inhibitory effect of 7-day wound fluid is a broader peak of fractions 6 through 11 than fractions 8 to 11 from 3-day wound fluid. Wound fluid from early 3-day implants and 7-day implants both showed a low molecular weight inhibitory activity which bound strongly to the charged ion exchange resin. The inhibitory activity isolated at the two time points appears to be the same. The mechanism of inhibited FPCL contraction may be related to actin-myosin filament sliding. This possibility can be explored by ATP-induced cell contraction
SOLUBLE
FACTOR(S)
IN RAT WOUND
137
FLUID
and the aggregation of actin filaments. Human fibroblasts plated on glass coverslips were incubated for 48 hr with individual fractions 8 through 11 from 3-day wound fluid, as shown in Fig. 2a. Harvested coverslips were permabilized with glycerol and half were fixed and stained with Rh-phalloidin. The other half was incubated with ATP and cofactors for 10min, then fixed, and stained. Cell length, determined by stained actin filaments, was measured (see Table I).While these fractions did not inhibit cell contraction, they do prevent fibroblasts from spreading and elongating on glass surfaces. Cell elongation was inhibited between 29 and 45%. This implies that a mechanism of inhibited lattice contraction may be related to impaired cell spreading in the collagen matrix. Identical results were produced by 7-day wound fluid fractions (results not shown). Fibroblasts in collagen lattices were observed. When FPCL were viewed by phase-contrast light microscopy, the variety of cell lengths within was great and no clear observable difference was found between treated and untreated FPCL. Interleukin-1 (IL-l), a well-characterized product of macrophages, is also a known modulator of lattice contraction. It has been shown to inhibit lattice contraction (Ehrlich and Wyler, 1983). Adding indomethacin with IL-1 to the FPCL blocked IL-l inhibitory activity. Indomethacin added with 7-day wound fluid fractions to the FPCL produced a minor reduction in the potency of wound fluid inhibitory activity, as shown in Table II. In no case did it restore lattice contraction to normal. It appears, therefore, that the production and release of prostaglandins may have a minimal effect on the inhibition of lattice contraction, while other components of wound fluid are more potent inhibitors of lattice contraction. DISCUSSION Implanted stainless steel wound chambers allow the sampling of soluble factors released in the interstital spaceduring wound repair. It would be expected that the major cell populations within the chamber at 3 days would be inflammatory cells such as neutrophils and macrophages (Edwards and Dunphy, 1958). By Day 7 the number of neutrophils will be reduced, and macrophages and fibroblasts will become the predominant cell types. Since the effects of 3- and 7-day wound fluid are similar, one could postulate that macrophages are common to both. One TABLE I ATP-Induced Cell Contraction Cell length with ATP Treatment
0 time
10 min
PBS F-S F-9 F-10 F-11
115.4 72.4 81.7 63.0 62.6
14.5 10.8 8.9 9.8 9.2
% Contraction 87 85 89 85 85
Note. Fibroblasts growing on glass coverslips were incubated with Mono Q fractions 8 through 11 or PBS for 48 hr and then treated with glycerol. Half the coverslips were ftxed and stained, and their cell lengths were measured. The other coverslips were incubated with ATP and cofactors for 10 mitt, fixed, and stained, and the cell lengths were measured. ATP-induced actin-myosin filament sliding produced cell contraction.
138
RITTENBERG,
BURD, AND EHRLICH
TABLE II Addition of Indomethacin to Fibroblast-Populated
Collagen Lattice
Lattice area (mm2) Treatment PBS PBS F-8 F-8 F-9 F-9 F-10 F-10 F-11 F-11
Indomethacin
Day 2
Change (%)
Day 4
Change (o/o)
-
106 95 189 171 237 157 227 259 207 186
11 10 20 18 25 16 24 27 22 19
64 58 143 128 165 130 234 254 207 174
7 6 15 13 17 14 24 26 22 18
+ + + + +
Note. Mono Q fractions 8 to 11 from 3-day rat wound fluid initially fractionated by size were included in the manufacture of FPCL at 10% v/v. Indomethacin at 10 &ml was included in the manufacture of half the FPCL. Lattice size was measured daily for 4 days.
speculation is that macrophages produce and release soluble factors to alter fibroblast-generated contractile forces. One such factor, IL-l, has been studied in lattice contraction (Ehrlich and Wyler, 1983). In that study, soluble factors released by experimental mouse granulomas incubated for 16 hr in tissue culture were shown to inhibit lattice contraction. The lattice contraction inhibitory activity released by chronic inflammatory cells was lost when indomethacin which blocks cyclooxygenase activity was included in the manufacture of FPCL. In contrast, the results reported here show that indomethacin had a modest antagonistic effect on inhibition induced by wound fluid fractions and only slightly restored lattice contraction. For example, fraction 11 at 4 days inhibited lattice contraction. The area of the control lattice was reduced to 7% of its initial size; the addition of fraction 11 reduced it to only 22%. Adding indomethacin increased lattice contraction from 22 to 18%, only a 4% improvement. Indomethacin did not restore lattice contraction to anywhere near normal, however. That would require the lattice to contract by 93% to 7% of its initial area. It appears that one of the inhibitory factors in wound fluid promotes the synthesis and release of prostaglandins such as IL-l. Other factor(s) are present, however, and they block lattice contraction by a mechanism unrelated to the release of prostaglandins. The inhibitor(s) of lattice contraction has a molecular weight between 10,000 and 20,000. It is strongly bound to the Mono Q column and requires more than 700 mM NaCl to displace it from the charged resin. It appears to differ from IL-1 because of biological considerations. In unpublished work, IL-I and PGEz included in monolayer cultures caused inhibited ATP-induced cell contraction. The mechanism for this inhibition appears related to the accumulation of intracellular CAMP induced by PGE2 (Ehrlich and Griswold, 1984; Ehrlich et al., 1989). The other mechanism of wound fluid inhibitory factor(s) appears to be independent of PGE,-induced CAMP production. It appears that wound fluid factor(s) inhibits cell spreading. Cell spreading requires actin-myosin filaments. Its modulation may be related to the mechanism responsible for inhibited lattice contraction. The identity of the mechanism of inhibited lattice contraction is unknown. Work to identify it continues.
SOLUBLE
FACTOR(S)
IN RAT WOUND
FLUID
139
These results support the possibility that macrophages produce factors which irdluence wound contraction. Macrophages are required for fibroplasia, and fibroblasts produce the force required for wound contraction. The initial speculation was that macrophages would produce factors to promote wound contraction. Instead, soluble factor(s), apparently released by macrophages, inhibit wound contraction. This is a puzzling result. Some work exists which links macrophages and wound contraction. Cortisone therapy inhibits wound contraction (Hunt et al., 1969). Adding vitamin A promotes fibroplasia and collagen synthesis in the presence of cortisone, but does not restore wound contraction. Tetracholorodecaoxygen (TCDO), a compound which reportedly enhances macrophage migration at the site of damage will stimulate wound contraction in open wounds in cortisone-treated rats (Hatz et al., in press). TCDO reportedly promotes macrophage accumulation at the wound site (Hinz et al., 1986). It may be that adding TCDO restores macrophage infiltration, but the responding macrophages are different from those invading the dead space of a normal rat wound. Macrophages stimulated by TCDO to infiltrate wounds may produce products and factors different from those of macrophages produced by acute inflammation. The macrophages exposed to TCDO may produce a factor(s) which promotes lattice contraction. The concept that wound fluid contains inhibitors of wound contraction suggests that if these factors could be neutralized, wound contraction might proceed faster. In some special cases enhanced wound contraction would be desirable for wound closure. ACKNOWLEDGMENTS The authors thank Joseph B. M. Rajaratnam, Alice Gosselin, and Frances M. Mascolo for help in preparing this manuscript. This work was supported by grants from the Shriners of North America and National Institutes of Health GM-32705.
REFERENCES BARAK, L. S., YOCUM, R. R., NOJHWAGEL, E. A., and WEBB, W. W. (1980). Fluorescence staining of the actin cytoskeleton in living cells with 7-nitrobenz-2-oxa-1,3-diazole phalloidin. Proc. Natl. Acad. Sci. USA II, 98&984. BELL, E., IVARSSON, B., and MERRILL, C. (1979). Production of a tissue-like structure by contraction of a collagen lattice by human tibroblasts of diierent proliferative potential in vitro. Proc. Natl. Acad. Sci. USA 76, 1274-1278. BILLINGHAM, R. E., and MEDAWAR, P. B. (1955). Contracture and intussusceptive growth in the healing of extensive wounds in mammalian skin. J. Anat. 89, 114-l 19. BUTTLE, D. B., and EHRLICH, H. P. (1983). Comparative studies of collagen lattice contraction utilizing a normal and transformed cell line. J. Cell. Physiol. 116, 159-168. EDWARDS, L. C., and DUNPHY, J. E. (1958). Wound Healing. II. Injury and abnormal repair. N. Engl. 1. Med. 259, 275-284. EHRLICH, H. P. (1984). Anti-intlammatory drugs in vascular response to bum injury. J. Trauma 24, 311-318. EHRLICH, H. P. (1988a). The role of connective tissue matrix in wound healing. In “Growth Factors and Other Aspects of Wound Healing” (A. Barbul, E. Pines, M. Caldwell, and T. K. Hunt, Eds.). A. R. Liss, New York. EHRLICH, H. P. (1988b). Wound closure: Evidence of cooperation between tibroblasts and collagen matrix. Eye 2, 149-157. EHRLICH, H. P., BUTTLE, D. J., and BERNANKE, D. H. (1989). Physiological variables affecting collagen lattice contraction by human dermal tibroblasts. Exp. Mol. Pathol. 50, 220-229. EHRLICH, H. P., and GRISWOLD, T. R. (1984). Epidermolysis bullosa dystrophica recessive tibro-
140
RITTENBERG,
BURD, AND EHRLICH
blasts produce increased concentrations of CAMP within a collagen matrix. J. Invest. Dermatol. 83, 330-333. EHRLICH, H. P., GRISWOLD, T. R., and RAJARATNAM, J. B. M. (1986). ATP-induced cell contraction with epidermolysis bullosa dystrophica recessive and normal dermal tibroblasts. J. Invest. Dermatol. 86, !46-100. EHRLICH, H. P., and WYLER, D. J. (1983). Fibroblast contraction of collagen lattices in vitro: Inhibition by chronic infIammatory cell mediators. J. Cell. Physiol. 116, 345-351. GOLDMAN, R. D., SCHLOSS, J. A., and STARGER, J. M. (1976). Organizational changes of actin-like microfilaments during cell movement. In “Cell Mobility” (R. D. Goldman, J. Pollard, and J. Rosenbaum, Eds.). Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. HATZ, R. A., KELLEY, S. F., and EHEUCH, H. P. The tetrachlorodecaoxygen complex reverses the effect of cortisone on wound healing. Plast. Reconstr. Surg., in press. HINZ, J., HAUTZINGER, H., and STAHL, K.-W. (1986). Rationale for and results from a randomized double-blind trial of tetrachlorodecaoxygen, an ion complex in wound healing. Lancet 8485, 825828. HUNT, T. K., T~OMEY, P. ZEDENFELT, B., and DUNPHY, J. E. (1967). Respiratory gas tensions and pH in healing wounds. Amer. J. Surg. 114, 302-307. HUNT, T. K., EHRLICH, H. P., GARCIA, J. A., and DUNPHY, J. E. (1%9). Effect of vitamin A on reversing the inhibitory effect of cortisone on healing of open wounds in animals and man. Ann. Surg. 170, 633+tO. Kamss, T. E., and BIRCHMEIRER, W. (1980). Stress fiber sarcomeres of tibroblasts are contractile. Cell 22, 555-561. WEILAND, T., and FAULSTICH, H. (1978). Amatoxins, phallotoxin, phallolysin, and atamanide: The biologically active components of poisonous Amanita mushrooms. CRC Crit. Rev. Biochem. 5, 185-260.
