Abstract
Objective. To assess whether blockade of tumor necrosis factor-α (TNF-α) influences surgical wound healing in a normal mouse experimental model.
Methods. Wound healing time course and degree of surgical wound collagenization were measured by morphological techniques and digital image analysis in 80 male SWISS-OF1 mice (40 received subcutaneous etanercept at a dose of 0.1 mg/25g/ at −7, 0, 7, and 14 days).
Results. No significant differences were observed between treated and untreated animals in wound healing, re-epithelialization, or formation of inflammatory infiltrate or granulation tissue at days 7, 15, or 20 after surgery. At 20 days, the collagen area was larger in treated versus untreated mice (109029 ± 28489 μm2 vs 79305 ± 19798 μm2, p = 0.026, Mann-Whitney U test).
Conclusion. Surgical wounds showed a higher degree of collagenization at 20 days in etanercept-treated versus untreated mice, with no differences in the time course of wound healing. These data suggest that biological therapies to block TNF-α do not affect wound healing and do not need to be suspended during the perioperative period.
Tumor necrosis factor-α (TNF-α) is a pivotal proinflammatory cytokine in rheumatoid arthritis (RA). Over the past decade, biological therapies against this cytokine have proven highly effective to treat patients with rheumatoid arthritis (RA) or other autoimmune diseases1,2. Various cytokines, including TNF-α and interleukin 1, participate in the wound-healing response to surgical incisions3. For this reason, anti-TNF-α treatment is usually suspended perioperatively to avoid possible cicatrization complications4,5. However, this practice is not supported by any clinical studies, and very little experimental research has been published on TNF-α–induced changes in wound-healing6,7.
We used digital microscopic image analysis to determine whether exposure to etanercept affects the time course of wound healing and formation of collagen in a murine experimental model.
MATERIALS AND METHODS
Experimental study in mice
Eighty male 35–40 g SWISS-OF1 mice kept in individual cages with food and drink ad libitum were randomly divided between a group treated with 0.1 mg/25 g of subcutaneous etanercept (Enbrel® Wyeth Europe Ltd.) on days 1, 2, 7, and 14 (N = 40) and a control group (N = 40). On day 0, all mice were anesthetized intraperitoneally, their backs were shaved, and a 5 × 5 mm incision was performed, withdrawing cutaneous and subcutaneous cell tissue. The study complied with current European legislation on research ethics and was approved by the ethical commission for animal research of our institution.
Measurement of wound-healing and collagen areas
Photographs of the wound were taken on days 0, 7, 15, and 20 using a Sony Cyber-Shot DSC-P120 digital camera at a constant distance and under identical conditions (Figure 1). The wound-healing area was measured from these photographs using the AutoCad2005® program (Tronix SRL). Wound-healing areas in each mouse were recorded as percentages of the area on day 0. Morphological changes in the widest section of the wound were semiquantitatively evaluated using hematoxylin and eosin and Masson trichrome methods. Re-epithelialization was scored on a 3-point scale: 0 = no epithelial lining, 1 = partial epithelial lining, and 2 = complete re-epithelialization. The thickness of the epithelial lining was measured (in microns) on light microscopy images (20×) using the AutoCad2005® program. A millimeter scale in the eyepiece of a BH2 microscope (Olympus Optical Company, Ltd., Tokyo, Japan) with 40× objective was used to count fibroblasts, granulocytes, and mononuclear cells per mm2. Results were expressed as cells/mm2. Connective tissue formation was evaluated by dying wound-tissue samples with Syrius red. The degree of collagenization of wound tissue and healthy tissue at the wound margin was measured from 5 images of each area taken with a video camera (20× objective) connected to an optic microscope, using the HR® fibrosis analysis program8 for automatic measurement of connective tissue.
Statistical analysis
Values are expressed as mean ± standard deviation (SD). Comparisons were evaluated with the Mann-Whitney U test and the Fischer test. SPSS 15.0 for Windows (SPSS Inc, Chicago, IL, USA was used for the statistical analyses, considering p < 0.05 to be significant.
RESULTS
Table 1 shows the time course of surgical wound healing (percentage of wound-healing area with respect to area on day 0) and the collagenization of wound and healthy tissue. No significant differences in wound healing time course or morphological changes (Figure 2) were observed between treated and untreated animals on days 7, 15, or 20. The mean collagen area was significantly larger in treated than in untreated mice (109029 ± 28489 μm2 vs 79305 ± 19798 μm2, p = 0.026 Mann-Whitney U test, Figure 3), with no significant differences in collagenization of the healthy margin of the wound.
Table 2 shows the mean number of fibroblasts, polymorphonuclear leukocyte cells, and mononuclear cells present and the re-epithelialization rate for each group of animals and each measurement time point. No statistically significant differences were found between groups or time points. Inflammatory cell counts were higher in non-treated animals, but statistical significance was not reached.
DISCUSSION
TNF-α blockade did not modify the time course of surgical wound healing in SWISS-OF1 mice, but blockade increased collagenization of the wound at 20 days after incision, with no change in collagenization of healthy areas. We can find no published experimental study on the effects of TNF-α blockade on the time course of wound healing and collagenization in healthy mice. Digital microscopic image analysis was used in the histopathological study of the wounds. This method offers highly precise and reproducible detection of differences in the degree of wound collagenization, since it is an automatic procedure in which the only intervention of the researcher is to select the areas for analysis. The surgical wound area was measured at baseline and on days 7, 15, and 20, selected to represent 3 phases of cicatrization: hemostasis/inflammation, proliferation, maturation/remodeling.
During wound repair, TNF-α overexpression is involved in neutrophil and monocyte chemotaxis, angiogenesis, fibroplasias, and metalloproteinase synthesis to destroy damaged extracellular matrix. Mori, et al6 found a higher amount of estimated neutrophils on day 1 and higher monocyte infiltrate on day 6 after TNF-α administration in surgically-wounded mice with p55 receptor than in p55-receptor knockout mice, demonstrating the role of TNF-α as neutrophil and monocyte chemoattractant in the wound-healing process. Although no statistical differences were found, our observation of a greater inflammatory response (granulocytes and mononuclear cells per mm2) in non-treated animals, especially on days 7 and 15, is compatible with the role of TNF-α described by Mori, et al.
We found no differences between treated and untreated mice in the percentage reduction in wound healing area on day 7. Mori, et al6 found a greater reduction in wound area in the knockout mice from day 3, suggesting an acceleration of cicatrization in the absence of TNF-α. We found no differences in the reduction of wound-healing area between treated and untreated animals on any day. These discrepancies are likely due to the difference in experimental models.
TNF-α appears to contribute to the formation of a collagen matrix of thicker fibers arranged parallel to stress lines2. However, we detected greater wound collagenization in the TNF-α-blocked mice, with no differences in healthy wound margins. Experimental studies of septic rats and diabetic mice with TNF-α overexpression7,9 concluded that an excess of TNF-α inhibits cicatrization. Although our mice were healthy, the TNF-α blockade may have had a selective fibrogenetic role in areas with overexpression, as after a surgical incision. The higher degree of collagenization in our treated mice might also have resulted from an imbalance between TNF-α and tumor growth factor ß, an antiinflammatory cytokine with fibroplastic effects10. This theory would explain the absence of significant differences in fibroblast count between treated and non-treated animals at any time point. Our histological findings do not demonstrate the known effect of TNF-α as inhibitor of fibroblast proliferation and promoter of fibroblast apoptosis.
In conclusion, this experimental study demonstrates greater collagenization of surgical wounds in mice treated with etanercept, with no differences in the time course of wound healing between treated and untreated mice. These data suggest that biological therapies to block TNF-α can be maintained during the perioperative period without increasing the risk of wound-healing complications.
Acknowledgments
The authors gratefully acknowledge the contribution of technicians María Dolores Rodríguez, Department of Pathology, and Serafín Vélez García and Carmen Ruiz Guzman, Department of Surgery, of Granada University.
Footnotes
- Accepted for publication May 27, 2009.