|Click here for information on how to order reprints of this article.|
Influence of Single and Sequential Cytokine Therapy on the Cell Cycle of Pressure Ulcer Fibroblastsa
Jerry S. Vande Berg, PhD*,#
Paul D. Smith, MD+
Patricia L. Haywood-Reid, BA*
Alivia B. Munson, BSc*
Nicole R. Bradley, BSc*
Martin C. Robson, MD+
*San Diego Veterans Administration Medical Center
#Division of Plastic Surgery, University of California, San Diego
+Institute for Tissue Regeneration, Repair and Rehabilitation
Bay Pines Veterans Administration Medical Center
Department of Surgery, University of South Florida, Tampa
aThis study was supported in part by the Medical Research Service, Department of Veterans Affairs and by the National Institutes of Health Grant R01-AR42967 from NIAMS.
KEY WORDS: cell cycle, chronic wounds, senescence, p21, PCNA
Pressure ulcers were treated with (A) granulocyte macrophage colony stimulating factor (GM-CSF), (B) basic fibroblast growth factor (bFGF), (C) sequential GM-CSF and bFGF, or (D) placebo for 36 days. Each treatment was investigated to determine its ability to modulate changes in ulcer fibroblast populations that were capable of cell division and those that were arrested in the cell cycle. In an in vitro colony size distribution assay, cultured fibroblasts isolated from tissues of each treatment group determined the ability of each cytokine to stimulate colony growth of various sizes. With regard to this phase of the study, it was hypothesized that the cytokine(s) that stimulated cell proliferation to form the greatest number of colonies would also be the most effective in promoting wound closure.
Examination of pressure ulcer surgical specimens by immunostaining for specific cell cycle markers was used to identify the effect that each cytokine had on stimulating ulcer fibroblasts in vivo. Fibroblast nuclei, stained positively for p21, suggested that these ulcer fibroblasts were senescent and nonproliferative. Ulcer fibroblast nuclei, stained for proliferating cell nuclear antigen (PCNA), identified cells that were capable of synthesizing DNA, thereby contributing to wound repair. Other fibroblasts demonstrated co-localization of both antigens and were also considered arrested in cell cycle, possibly for repair of DNA. With regard to the second phase of this study, it was hypothesized that the treatment that stimulated the largest number of PCNA positive cells and fewest number of p21 positive cells would be the most effective in stimulating wound repair. Collectively, these in vitro and in vivo approaches were evaluated to measure the effectiveness of cytokine therapy in the repair process of pressure ulcers.
Previous clinical trials have shown that application of specific cytokines to chronic wounds induced cell proliferation and formation of extracellular matrix (ECM).1-5 Other studies employed a multifunctional cytokine that stimulated cell proliferation, promoted ECM development, and served as a bactericide.6 Unfortunately, the outcome from these studies was not consistent or predictable. Recently, chronic wound studies have focused on the utilization of cytokine cocktails to treat pressure ulcers. These investigations were predicated on the assumption that each of several cytokines comprising the cocktail would stimulate a specific phase in the repair process.
In the form of a standard healing curve, Robson and associates showed that GM-CSF appeared to accelerate healing during the early stages of chronic wound repair.7 Basic FGF was shown to exert its influence in the second phase of the contraction trajectory curve by overcoming the inhibition to contraction caused by infection.8 Based on these findings, it was postulated that the sequential use of GM-CSF and bFGF as a cytokine cocktail would accelerate contraction, resulting in a greater effect on wound healing than seen with either agent alone.
The present study is a collaborative effort between a clinical investigation conducted by the Robson research team at the Institute for Tissue Regeneration, Repair and Rehabilitation at the VA Medical Center, Bay Pines, Florida9 and cellular analyses performed in our laboratory. In the Robson clinical trial, pressure ulcers were treated with topically applied GM-CSF or bFGF alone or in sequence. To identify the ability of these cytokines to stimulate cell proliferation, our first objective was to apply a colony size distribution assay to cultured fibroblasts isolated from treated pressure ulcers.10 To further evaluate this process, immunocytochemical localization of specific cell cycle markers was used to additionally identify and measure the capability of these cytokines to modulate the fibroblast population in pressure ulcers. Because our previous studies showed that fibroblasts in pressure ulcers may become prematurely senescent, the significance of this study was to utilize the colony size distribution assay and immunocytochemical localization of cell cycle markers to determine the ability of these cytokines to modulate the remaining viable population of ulcer fibroblasts. Our hypotheses were that these in vitro and in vivo approaches would reflect the ongoing status of wound closure in pressure ulcers undergoing repair.
MATERIALS AND METHODS
Pressure ulcer specimens were obtained according to an approved protocol in a clinical trial conducted by Dr. M. Robson at the VAMC, Bay Pines. The clinical trial was composed of 61 patients from a four-treatment blinded randomized trial comparing placebo, sequential topical GM-CSF/bFGF therapy, as well as each cytokine alone.9 Only random pressure ulcer tissues were available during the trial from each treatment group.
Clinical Trial Treatment Data
Consecutive patients fulfilling the entry
criteria with pressure ulcers measuring 10 to 200 cm3 of
at least 8 weeks duration were randomized to one of four treatment
regimens: (A) 2.0 µg/cm2 GM-CSF topically applied daily
for 35 days; (B) 5.0 µg/cm2 bFGF applied daily for 35 days; (C)
Colony Size Distribution Assay
Fibroblast cultures were established from additional random portions of pressure ulcer biopsies after day 0 and 36 days of either placebo or growth factor treatment. Tissue pieces (0.5 to 1.0 mm3; 2 to 3 pieces/dish) were anchored by glass coverslips onto 60-mm2 culture dishes and incubated in DMEM/Hams F12 medium, pH 7.2, containing 10% fetal bovine serum, 50 IU/mL penicillin, 50 µg/mL streptomycin and 5.0 mM glutamine. Once fibroblast growth was observed from each explant, cells continued to appear and migrate across the culture-dish surface. Cells were fed every 4 to 5 days and maintained in a humidified atmosphere of 5% CO2 at 37ºC.
From first passage, subconfluent populations, fibroblasts were diluted in the above culture medium to an approximate concentration of 300 cells/100 mm2 culture dishes. Plating efficiency in most culture dishes appeared to be 50% to 60%, indicating that approximately 180 to 200 viable fibroblasts were attached to the dish surface to form cell colonies. Cultures were blinded and plated in triplicate. Once the cells were plated in the culture dishes, they were placed in the cell incubator and left undisturbed for 2 weeks. Cell colonies were then washed twice in phosphate buffered saline (pH, 7.4) then fixed in 2% buffered paraformaldehyde for 15 minutes. Following 1% methylene blue staining for 10 minutes, colonies were then washed in distilled water and dried in a forced air dryer chamber for 15 minutes. Methylene blue staining provided an easy distinction between large (250 to 600 cells) and small colonies (20 to 100 cells). Well-formed colonies that were intermediate in size were found to have between 100 and 250 cells. Scattered cells, localized in colonies less than 20 cells, were not counted.
Immunofluorescence Microscopy of Cell Cycle Markers
At the VAMC, San Diego, tissues were washed in 0.1 M sodium phosphate buffer (pH, 7.4) to remove fixative and placed in 70% ethanol overnight at 4ºC. Tissues were dehydrated through a graded series of ethanol washes before embedding in paraffin. Tissue blocks were cut into 5-µm sections and collected onto glass slides. Sections were deparaffinized and rehydrated through a graded ethanol series to distilled water. Sections were incubated in 3% hydrogen peroxide to block endogenous peroxidase activity and then incubated in 50-mL preheated Target Retrieval solution (pH, 6.0; Dako, Carpinteria, CA) for 40 minutes at 100ºC to reverse loss of antigenicity caused by fixation. After 10 minutes of cooling, tissues were transferred to TBS (tris buffered saline) for 5 minutes followed by 5% normal goat serum (GibcoBRL, Bethesda, MD) diluted in TBS for 5 minutes. Tissues were incubated overnight at room temperature with a primary antibody diluted in antibody diluent (Dako) to reduce background. Optimum dilutions were experimentally determined. Primary antibodies included a rabbit anti-p21 (Santa Cruz Biotechnology Inc., Santa Cruz, CA; concentration, 0.2 mg/mL) at a dilution of 1/100 and mouse anti-PCNA, clone PC10 (Biodesign; concentration, 0.1 mg/mL) at a dilution of 1/100. Negative control primary antibodies were rabbit IgG (Sigma Immunochemicals, St. Louis, MO; concentration, 10 mg/mL) with a dilution of 1/10,000 and mouse IgG, isotypic control (Dako; concentration, 0.1 mg/mL) with a dilution of 1/100. Slides were rinsed in TBS twice for 10 minutes (2X) at room temperature before incubation with a secondary antibody diluted in TBS for 2 hours at room temperature. Secondary antibodies included alkaline phosphatase labeled goat anti-rabbit IgG (GAR; Kirkegaard and Perry Laboratories) at a dilution of 1/250 and horseradish peroxidase labeled goat anti-mouse IgG (GAM; Kirkegaard and Perry Laboratories) at a dilution of 1/250. Reporter systems used for GAR and GAM secondary antibodies were HistoMark RED (Kirkegaard and Perry Laboratories) for alkaline phosphatase and TrueBlue peroxidase substrate (Kirkegaard and Perry Laboratories), respectively.
Counting Labeled Fibroblasts
All slides were blinded and coded by one individual (PHR). A second investigator (JV) photographed six random areas near the surface of the ulcer bed and six interior regions, approximately 3 to 5 mm below the ulcer bed surface. The rationale for counting near the surface was that fibroblasts in this area would likely be exposed to the adverse effects of low pH, sepsis, etc. within the pressure ulcer environment. On the other hand, measuring activity among fibroblasts in subsurface regions may be important to understanding the contribution of angiogenesis toward progression of these cells through G1. From an original magnification of 62.5X, labeled cells were counted from 4 X 6 micrographs, which represented 0.04 mm2 area of the pressure ulcer bed. Because ulcer bed tissues were highly variable in organization and cell number, absolute cell count as a background was not considered to be an objective denominator. Therefore, we adapted the procedure of Atropoulos and Williams13 to count cells within a specific area to normalize the denominator (i.e., cells/0.04 mm2). When the code was revealed, labeled cells were counted as percentage of total cells stained.
Colony Size Distribution Assay
From 15 patients in each treatment group, six individuals each from the GM-CSF, five patients each from bFGF and sequential GM-CSF/bFGF, and four patients from placebo therapies provided pressure ulcer tissues. Selection of patients from each treatment category was entirely random.
Fibroblasts cultured from cytokine-treated pressure ulcers were isolated as single cells to form small, medium, and large cell colonies. In 100-mm2 culture dishes, colonies of 20 to 100 cells were classified as small, 101 to 250 cells as medium, and 251 to 600 cells were considered large. Table 1 shows that before cytokine treatment at day 0, cell colonies from each group were relatively similar in size. In most instances after 36 days of treatment, the average number of small and medium colonies decreased. The average number of large colonies increased only in fibroblasts that were isolated from pressure ulcers treated with bFGF and sequential GMCSF/bFGF.
showed that some treatment groups made more colonies than others (P
= .02 ANOVA, two-way repeated measures). Post-hoc Tukey tests were
computed to clarify this effect. Irrespective of colony size, ulcer
fibroblasts in the GM-CSF/bFGF treatment group produced significantly
more colonies than cells in the GM-CSF group (P = .05). Ulcer fibroblasts
in the GM-CSF/bFGF treatment group produced more colonies than cells
from placebo treated wounds, although this difference only approached
Immunofluorescence of Cell Cycle Markers
From approximately 15 patients in each treatment category, four individuals each from the GMCSF and placebo treatment groups and five patients each from bFGF and sequential GM-CSF/bFGF therapies provided pressure ulcer tissues for study. The positive staining of ulcer fibroblast nuclei with anti-p21 appeared red (phosphatase) while cells labeled with blue (peroxidase) were positive for PCNA. Some cell nuclei demonstrated co-localization of both antigens. Uninjured, normal skin adjacent to the pressure ulcer demonstrated only a small percentage of p21 and PCNA positively stained fibroblasts. Likewise, fibroblast nuclei demonstrating co-localization of both antigens were rarely observed. Pardee suggested that in normal skin these observations were to be expected as many cells in vivo are in a quiescent state (Go).14 Although some endothelial cells, smooth muscle cells, and pericytes surrounding blood vasculature exhibited staining of both cell cycle markers, only fibroblast populations were counted and analyzed.
Placebo-Treated Pressure Ulcers
Twelve regions in each of four placebo-treated
pressure ulcers were examined. The arrested fibroblast population
(p21 and PCNA+p21) was commonly found along the surface of the ulcer
bed, while most of the cells capable of division (PCNA positive) were
observed in proximal regions near the blood vasculature. In the latter
areas, PCNA-labeled ulcer fibroblasts were organized in parallelan
arrangement that is commonly observed in healing acute wounds. Before
placebo treatment began, 57% of the ulcer fibroblast population was
arrested in the cell cycle. After 36 days, placebo wounds showed an
average closure of 71% and an average increase of 37% in fibroblasts
that were capable of cell division (Figure 1A).
GM-CSFTreated Pressure Ulcers
Data were collected from four patients
at 12 sites in each of the GM-CSFtreated pressure ulcers. Pressure
ulcer fibroblasts immunostained for p21 were found near the surface
of the ulcer bed where only reticulated collagen fibers were a common
feature in these wounds. Fibroblasts staining positive for PCNA and
capable of division were observed in regions near blood vasculature.
Fibroblasts demonstrating co-localization of both markers were found
randomly throughout the ECM. Following 36 days of GM-CSF treatment,
sample wounds from our study averaged 61% closure (Figure 1B).
BFGF-Treated Pressure Ulcers
Data were collected from five patients at 12 sites in each of the bFGF-treated pressure ulcers. After 36 days of treatment, average closure in our wound samples was 75% (Figure 1C). Most of these wounds exhibited a greater overall development and organization of the ECM than found in the placebo and GM-CSF treatment groups. Evidence of ongoing angiogenesis near the surface of the ulcer bed was identified by endothelial cells that stained positive for PCNA. Unlike placebo and GMCSF treated wounds, the surface of the ulcer bed also contained an average increase of 37% in the fibroblast population capable of cell division (PCNA positive). Similar to tissues from other treatment groups, most of the p21 positive senescent cells were observed near the surface of the ulcer bed with some cells scattered randomly in basal regions. Before the onset of bFGF treatment, approximately 63% of the ulcer fibroblast population was p21-positive; however, following treatment, the size of this arrested cell population decreased to 21%. The proportion of cells demonstrating co-localization of both antigens increased slightly and was observed primarily interior to the ulcer bed surface.
GM-CSF/bFGFTreated Pressure Ulcers
Data were collected from 5 patients at 12 separate sites in each of the pressure ulcer tissues. At the end of the sequential treatment, average wound closure in our wound sample was 70% (Figure 1D). Some of these pressure ulcers showed evidence of remodeling that was comparable to that observed in the bFGF treatment group. Immunostaining of random pressure ulcer sections showed an even distribution of ulcer fibroblasts that were capable of cell division (PCNA positive). Most of the p21-positive and double-labeled fibroblasts were observed near the surface of the ulcer bed. Following GM-CSF/bFGF treatment, there was a 20% increase in the proportion of cells that were capable of cell division and only a marginal decrease in the number of cells arrested in the cell cycle. Ulcer fibroblasts demonstrating co-localization of both antigens appeared to decrease 28%.
Cell cycle data were evaluated using two-way
ANOVAs with repeated measures. Group differences in the change over
time were not significant for ulcer fibroblasts labeled with p21 and
cells labeled with both antigens. After 36 days, a significant change
in cells capable of division (PCNA positive) was observed for bFGF
In a previous investigation, we showed for the first time that regardless of patient age, fibroblasts in pressure ulcers appeared to be prematurely senescent.15 These findings were significant because they provided an explanation as to why pressure ulcers, regardless of treatment, do not always heal consistently. In a subsequent study involving pressure ulcers treated with only quality care, we utilized cell cycle markers to evaluate changes in proliferating cell populations versus populations of cells with arrested growth in pressure ulcers.16
In this investigation, we utilized a colony size distribution assay plus cell cycle markers to examine the ability of GM-CSF or bFGF alone or combined GM-CSF/bFGF treatments to modulate pressure ulcer populations in vivo. Our hypothesis for the first approach was that the most effective cytokine in stimulating colony growth, particularly large colonies, would be the most effective in promoting wound closure. Our hypothesis for the second approach was that the best treatment would stimulate a decrease in the nondividing cell population and an increase in the proportion of ulcer fibroblasts capable of cell division. Based on this information, we theorized that both approaches would reflect the status of wound closure.
The Robson clinical data showed that significantly
more patient wounds receiving any of the cytokine treatment reached
85% closure compared with placebo treated wounds (P = .03). Wounds
receiving bFGF alone responded the best. Compared with placebo-treated
wounds, there were more wounds in the bFGF group that achieved 85%
closure (P = .02). The outcome of the sequential GM-CSF/bFGF treatment
PCNA, as an indicator of a cells ability to synthesize DNA and divide, was the only cell cycle marker that demonstrated significant differences with respect to 36 days of cytokine therapy. Before the onset of only placebo-treated wounds at day 0, it was noted that the majority of ulcer fibroblasts stained positive for p21 and were considered arrested in the cell cycle. At the same time, there were also fewer ulcer fibroblasts that were capable of cell division as indicated by PCNA staining. However, after 36 days of quality care, these pressure ulcers demonstrated a decrease in the arrested fibroblast population and a large increase in the cell population that was capable of cell division. This trend in modulation of fibroblast populations was evident in placebo, bFGF, and sequential GM-CSF/bFGFtreated wounds where the average wound closure was above 70% (Figures 1A, 1C, 1D). In the GM-CSFtreated pressure ulcers where the average wound closure was 61% (Figure 1B), the proliferating fibroblast population decreased and the arrested cell population remained unchanged.
In this regard, bFGF appeared to be significantly more effective than sequential GM-CSF/bFGF (P = .03) and GMCSF alone (P = .048) in stimulating cells to proliferate. An explanation for the success of bFGF in promoting ulcer fibroblast proliferation may be based on its capacity to activate competency early in G1 and its ability to enhance the response to progression factors like IGF-1 to the S phase.17 Additionally, midway through G1, bFGF has been shown to induce an immediate increase in the number of cyclin D1 transcripts and protein level in fibroblasts.17 The level of D-type cyclins is significant because of the ability of this protein to bind and hyperphosphorylate the retinoblastoma protein (pRb) results in the release of the E2F transcription factors and entry to the S phase.18 Because cells that are unable to phosphorylate pRb become senescent,19 the ability of bFGF to stimulate the remaining viable fibroblasts to divide becomes a significant factor for including this cytokine in future growth factor cocktails.
Our results demonstrated that when there was a significant viable fibroblast population, pressure ulcers began to heal and close. However, when a crucial portion of the fibroblast population was arrested and/or in a stage of DNA repair, these wounds demonstrated less closure. Colony formation and cell cycle data appeared to mirror the clinical trial data as suggested by the ability of bFGF treatment to (1) form more large colonies and (2) form the most sizable increase in proportion of dividing cells and the most extensive decrease in arrested cell population as shown by immunohistochemistry. Thus, results from the clinical and laboratory studies appear to support our hypothesis that the best treatment would demonstrate an increase in the proportion of ulcer fibroblasts capable of proliferation and a decrease in the number of cells arrested in the cell cycle. This conclusion must be accepted with caution as the differences in wound closure among our wound samples from all treatment groups were not significantly different. To fully evaluate these approaches to cytokine treatment, a larger sample of pressure ulcers is needed to demonstrate precise differences and to more clearly differentiate cytokine treatment effects.
1. Robson MC, Phillips LG, Thomason A, et al: Platelet derived growth factor-BB in chronic pressure ulcers. Lancet 339:2325, 1992.
2. Mustoe TA, Cutler NR, Allman RM, et al: A phase II study to evaluate recombinant PDGF-BB in the treatment stage III/IV pressure ulcers. Arch Surg 129:213219, 1994.
3. Rees RS, Robson MC, Smiell SM, et al: Becaplermin gel in the treatment of pressure ulcers: A phase II randomized, double-blind, placebo-controlled study. Wound Rep Reg 7(3):141147, 1999.
4. Pierce GF, Tarpley JE, Allman RM, et al: Tissue repair processes in healing chronic pressure ulcers treated with recombinant platelet derived growth factor BB. Am J Pathol 145:13991410, 1994.
5. Robson MC, Abdullah A, Burns BF, et al: S. Safety and effect of the topical recombinant human interleukin-1b in the management of pressure sores. Wound Rep Reg 2:177181, 1994.
6. Kucukcelebi A, Hui PS, Sahara K, et al: The effect of interleukin-1b on the inhibition of contraction of excisional wounds caused by bacterial contamination. Surg Forum 43:715716, 1992.
7. Kucukcelebi A, Carp SS, Hayward PG, et al: Granulocyte-macrophage colony stimulating factor reverses the inhibition of wound contraction caused by bacterial contamination. Wounds 4:241247, 1992.
8. Hayward PG, Hokanson J, Heggers JP, et al: Fibroblast growth factor reverses the bacterial retardation of wound contraction. Am J Surg 163:288293, 1992.
9. Robson MC, Hill DP, Smith PD, et al: Sequential cytokine therapy for pressure ulcers: Clinical and mechanistic response. Ann Surg 231(4):600611, 2000.
10. Smith JR, Pereira-Smith OM, Braunschweiger KI, et al: A general method for determining the replicative age of normal animal cell cultures. Mech Aging Devel 12:355365, 1980.
11. Resch CS, Kerner E, Heggers JP, Scherer M, et al: Pressure sore volume measurement: A technique to document and record wound healing. J Am Geriatr Soc 36:444446, 1988.
12. Robson MC: Infection in the surgical patient: An imbalance in normal equilibrium. Clin Plast Surg 6:493503, 1979.
13. Atropoulos MJ, Williams GM: Proliferation markers. Exp Toxic Pathol 48:175181, 1996.
14. Pardee AB: G1 events and regulation of cell proliferation. Science 246:603608, 1989.
15. Vande Berg JS, Rudolph R, Hollan C, et al: Fibroblast senescence in pressure ulcers. Wound Rep Reg 6:3849, 1998.
16. Vande Berg JS, Smith PD, Haywood-Reid PL, et al: Dynamic forces in the cell cycle affecting fibroblasts in pressure ulcers. Wound Rep Reg 9:1927, 2001.
17. Verrier B, Muller D, Bravo R, Muller R: Wounding a fibroblast monolayer results in the rapid induction of the c-fos proto-oncogene. EMBO J 5(5):913917, 1986.
18. Rao SS, Kohtz DS: Positive and negative regulation of D-type cyclin expression in skeletal myoblasts by basic fibroblast growth factor and transforming growth factor B. A role for cyclin D-1 in control of myoblast differentiation. J Biol Chem 270:40934100, 1993.
19. Campisi J: Replicative senescence and immortalization, in Stein GS, Baserga R, Giordano G, Denhardt D (eds): The Molecular Basis of Cell Cycle and Growth Control. New York, Wiley-Liss Inc., pp 348373, 1999.
Effects of Cytokine Treatment on Colony Size Distribution
Day 0 Day 36 Day 0 Day 36 Day 0 Day 36
Treatment Small Colonies Medium Colonies Large Colonies
Palcebo 45 (18.0) 27 (10.6) 34 (4.30) 6 (4.30) 19 (13.3) 14 (1.7)
GMCSF 43 (20.0) 16 (8.5) 37 (4.3) 5 (4.3) 19 (13.0) 2.5 (2.0)
bFGF 45 (18.0) 27 (10.6) 31 (12.0) 21 (14.2) 23 (14.30) 40 (13.4)
GMCS/bFGF 45 (14.1) 6 (7.8) 38 (8.5) 39 (32.3) 28 (15.1) 41 (3.5)
Table 1. Colonies in triplicate dishes from each patient were counted and expressed as an average percentage (standard deviation) of the total colonies. Basic FGF stimulated more ulcer fibroblasts to form more large than small- or medium-sized colonies. Statistical analyses showed that some treatment groups made more colonies than others (P 5 .02; ANOVA two-way repeated measures). Post-hoc Tukey tests were computed to clarify this effect. Irrespective of colony size, ulcer fibroblasts in the GM-CSF/bFGF treatment group produced significantly more colonies than cells in the GM-CSF group (P 5 .05). Ulcer fibroblasts in the GM-CSF/bFGF treatment group produced more colonies than cells from placebo-treated wounds, although this difference only approached statistical significance (P 5 .052). No significant differences were observed in colony formation of any size between GM-CSF/bFGF and bFGF-treated wounds.
©2000-2013. All Rights Reserved. Veterinary Solutions LLC