The Effect of Several Silver-Containing Wound Dressings on Fibroblast Function In Vitro Using the Collagen Lattice Contraction M

Christine A. Cochrane, PhD;1 Michael Walker, PhD;2 Phil Bowler, MPhil;2 David Parsons, PhD;2 Derek C. Knottenbelt, OBE, DVM&S1

Over the last 2 millennia, many civilizations have recognized silver as a treatment for preventing diseases.1 Within the last 50 years, following renewed interest in the use of silver nitrate solution for burn wound infections,2 silver became known as one of the least toxic antimicrobial agents and regained widespread acceptance in wound care.3
The primary purpose of a topical antimicrobial agent is to reduce bacterial contamination and opportunities for infection. Ideally, these agents should combine maximum antimicrobial function with minimal toxicity to either intact or damaged tissue. In particular, when a topical antimicrobial agent is applied to an open wound, it is important to understand the possible effects that this agent may have on cells that play a critical role in tissue regeneration (eg, fibroblasts, endothelial cells, and keratinocytes). Consequently, the concentration and availability of an antimicrobial agent to host tissue are important issues.
It is difficult to use in-vitro models to mimic the complex events that take place in wound healing, as they cannot duplicate many of the complex and uncontrolled events within a wound environment.4 However, the fibroblast-seeded collagen gel contraction model, first proposed by Bell et al, has been used extensively for more than 25 years.5 In a recent review, Carlson and Longaker suggested that this model’s simplicity (ie, the use of a single cell line embedded in a collagen gel) may be its strength.4 It was further stated that “If its limitations are respected then the fibroblast-populated collagen matrix can be used as a model of the healing wound.”4
In previous studies using this in-vitro model, the contraction rates of equine granulation tissue fibroblasts have been shown to be comparable to human granulation tissue fibroblasts,6 as this model allows fibroblasts to grow in a 3-dimensional collagen gel structure, a matrix component native to the wound environment. Previous studies have looked at the antibacterial, physical, and chemical characteristics of 7 proprietary silver-containing dressings.7 In this report, in-vitro studies of biological characteristics (eg, fibroblast function) of the same dressings were conducted using equine fibroblasts seeded within a collagen gel structure.

Materials and Methods

Dressings. This study assessed the characteristics of 7 proprietary silver-containing antimicrobial dressings that have all received an appropriate regulatory (eg, US Food and Drug Administration) risk-benefit profile: 3 fibrous dressings—AQUACEL® Ag (ConvaTec, Skillman, NJ, USA; referred to throughout this article as nonwoven A), Acticoat® Absorbent (Smith & Nephew, London, UK; referred to throughout this article as nonwoven B), and SILVERCEL® (Johnson & Johnson Wound Management, Somerville, NJ, USA; referred to throughout this article as nonwoven C); 2 foam dressings—Contreet® Foam (Coloplast, Holtedam, Denmark; referred to throughout this article as foam A) and PolyMem® Silver (Ferris Pharmaceuticals, Burr Ridge, Ill, USA; referred to throughout this article as foam B); a gauze dressing—Urgotul® S.Ag (Laboratoires Urgo, Chenôve, France; referred to throughout this article as gauze); and a nonadhesive polymer hydrogel sheet—SilvaSorb® (AcryMed/Medline, Mundelein, Ill, USA; referred to throughout this paper as hydrogel). The physical and chemical characteristics of these dressings have been previously described.7
Cell culture. Granulation tissue fibroblasts (GTFs) were taken from slow-healing areas of equine wounds, and the fibroblast-seeded collagen gels were prepared as described previously.6 Briefly, tissue samples were immediately transferred to a sterile Petri dish, washed in Hanks’ balanced salt solution (HBSS) (all cell culture materials were supplied by Gibco, UK unless otherwise stated), and cut into 3–5 mm2 pieces before being placed into 25 cm2 tissue culture flasks containing media (prepared by adding 10% fetal calf serum to Dulbecco’s Modified Eagle Medium [DMEM], supplemented with 20 mM Hepes buffer, 100 µg/mL gentamicin, and 0.5 µg/mL amphotericin B) and incubated in a 5% carbon dioxide/95% air environment at 37˚C.
Following incubation, the fibroblasts were harvested from stock dishes and plated in 35 mm 6-well plates at a concentration of 1 x 106 cells/mL in molten type I collagen (Sigma, UK) prior to the application of moistened dressings.
Gel contraction and viability assays. For gel contraction and viability assays, 0.5 g of each silver-containing wound dressing was pre-moistened with sterile saline (0.5 mL) and placed onto the surface of the fibroblast-seeded collagen gel, and 0.5 mL of sterile saline was applied directly to the control gel with no dressing. Gel contraction measurements were made using calipers (mm) at 24-hour intervals over a period of 96 hours. Quantitative evaluation of fibroblast viability was carried out at the end of the experiment (ie, after 96 hours) using the trypan blue exclusion assay. Six replicates were performed for each dressing plus controls.
Statistical analysis. A multivariate analysis of variance (Duncan’s multiple comparison test) was used to analyze the data.


To more accurately measure the individual effect of a dressing on fibroblast contraction, a percentage area of collagen gel contraction relative to the control was calculated for each dressing using the equations outlined as follows. Equation 1 shows the calculation for the absolute area of contraction (AAC) for the control and each individual dressing expressed as a percentage:

AAC= [initial radius of the gel (ie, 35 mm)]2 – (final radius)2 X 100 ÷ (initial radius)2

Equation 2 expresses the AAC of each dressing as a percentage contraction relative to the control:

% contraction relative to control =AAC dressing ÷ AAC control X 100

Consequently, a high percentage value in both calculations (ie, contraction rate similar to control) would suggest that there had been minimal dressing interaction by the active silver component and/or other dressing components over the experimental exposure period of 96 hours.
The results from these in-vitro studies show that 4 of the dressings had a minimal effect on collagen gel contraction while the other 3 dressings, nonwoven B, foam A, and hydrogel, induced minimal changes in percentage area contraction (Figure 1). The cell viability data showed a trend similar to the contraction data (Figure 2) with nonwoven A, nonwoven C, gauze, and foam B having the least effect on cell viability after 96 hours (ie, approximately 70% viability) compared to > 80% cell death for the remaining 3 dressings.


Since its introduction in the 1970s, the in-vitro fibroblast-seeded collagen gel model has been extensively utilized. In 1990, Ehrlich and Rajaratnam8 proposed a mechanism of wound closure based on fibroblasts using locomotion forces. According to this proposal, the fibroblasts rearrange the collagen fibrils by their movement, resulting in contraction of the collagen gel; however, other factors can also potentially influence contraction in this model (eg, cell density, viability, and proliferation rates9), and as a consequence, contraction rates will be greater in vitro than in vivo.8,9
In these investigations, silver-containing dressings were applied to a fibroblast-seeded collagen gel to determine the effect of silver on fibroblast gel contraction and viability. Using saline as the hydrating medium, the results from these studies showed that although there were marked differences between the dressings in their ability to allow the collagen gels to contract, this was unlikely to be related to silver availability alone but may have been due to other factors, such as the physicochemical characteristics of individual dressings or other dressing components. For example, the hydrogel dressing has a relatively low silver content (eg, < 10 mg/10 cm x 10 cm dressing),7 yet this showed minimal percentage area collagen gel reduction and was significantly different (P < 0.05) from all the other dressings tested throughout the experimental period. This dressing also caused a marked reduction in cell viability (eg, < 99%) over 96 hours, suggesting that there may be cellular/dressing interactions that were contributing to the observed effect.
The considerable differences in the fluid-handling capabilities of the dressings7 may have been another factor that influenced collagen gel contraction. For example, of the 2 foam dressings that were evaluated (ie, foam A and foam B), the latter was shown to induce a significantly greater (P < 0.05) percentage area collagen gel contraction and minimal effect on cell viability (Figure 2). One explanation could be due to the fact that foam A has been shown to handle approximately 4 times as much fluid as foam B.7 Under these experimental conditions, the increased absorbent capacity of foam A may lead to reduced fluid levels within the collagen gel, which in turn may reduce the ability of the fibroblasts to move, allowing contraction to take place. Consequently, this may explain the poor viability observed for this product at 96 hours (Figure 2).
Although 2 of the nonwoven dressings (B and C) have similar silver content,7 their influence on collagen gel contraction and subsequent cell viability were markedly different. Nonwoven C was shown to have a significantly greater area reduction and viability profile than nonwoven B, and again, this could be due to differences in the physical structure of each dressing. Nonwoven C is a blended calcium alginate-carboxymethylcellulose fiber with metallic silver coated onto nylon fiber, whereas nonwoven B is a calcium alginate dressing coated with metallic silver. In these experiments, it was observed that some of the metallic-coated fibers of nonwoven B became detached from the dressing. The detachment of fibers and possible displacement of metallic silver particles may explain the reduced area collagen gel contraction and low cell viability seen with nonwoven B compared with nonwoven C.
The remaining 2 dressings tested, nonwoven A and gauze, were associated with the best rates of collagen gel contraction and along with nonwoven C and foam B indicated a good cell viability count of > 70% after 96 hours. Nonwoven A and gauze were shown not to be significantly different from each other, although nonwoven A was significantly different from all the other dressings tested (P > 0.05), while nonwoven C, foam B, and gauze were shown to be significantly different from all other tested dressings. The gauze dressing is coated in a petrolatum paste that contains hydrocolloid and silver sulfadiazine particles, and while it has minimal fluid-handling capabilities,7 it is likely to conform well to the collagen gel surface without appearing to compromise the fibroblast activity. Nonwoven A is a 100% sodium carboxymethylcellulose dressing that has been shown to have good absorbency and fluid retention properties7 and when hydrated forms a cohesive gel structure that provides excellent conformability.10


The fibroblast-seeded collagen gel model has been shown to be effective in evaluating the effect of silver-containing dressings on fibroblast contraction and viability. Using a physiologically relevant hydrating medium (ie, saline), the in-vitro toxic effects exerted by silver-containing dressings have been shown to be variable. Such effects may be associated with both the active agent and the physicochemical properties of the different dressings. In particular, nonwoven A, nonwoven C, foam B, and the impregnated gauze dressing were shown to be least detrimental to cells involved in this in-vitro study.
These results have provided further evidence that dressing selection should not be based on one particular attribute but that dressing choice should be based on the overall characteristics of a dressing, such as its antimicrobial, fluid handling, physical, and chemical properties.


Back to top