¹Departments of Biomedical Engineering, Dermatology, and Medicine, Stony Brook University, Stony Brook, New York, USA

Correspondence: Dr Richard AF Clark, Center for Tissue Engineering, Department of Biomedical Engineering, Stony Brook University, HSC T-16, Room 060, Stony Brook, New York 11794-8165, USA. E-mail: richard.clark@sunysb.edu

Abstract

In chronic wounds, fibroblast dysfunctions, such as increased apoptosis, premature senescence, senescence-like phenotype, or poor growth response in the absence of senescence markers, have been reported. Some of these differential dysfunctions may be secondary to differences in patient age or sex, ulcer size or duration, edge versus base sampling, or culture technique. Nevertheless, the entire spectrum of fibroblast dysfunction may exist and be secondary to, or a response to, different amounts of oxidative stress.

In their article entitled "Fibroblast dysfunction is a key factor in the non-healing of chronic venous leg ulcers," Wall et al. (2008, this issue) demonstrate that a telomere-independent fibroblast abnormality occurs in chronic venous leg ulcers and posit that oxidative stress underlies this dysfunctional behavior. Although it is dysfunctional, the authors have clearly emphasized both here and in a previous publication (Stephens et al., 2003) that the chronic wound fibroblast phenotype observed by them does not represent the replicative senescence-like phenotype as has been reported for chronic wound fibroblasts by a number of other investigators (Agren et al., 1999; Mendez et al., 1998; Stanley and Osler, 2001; Stanley et al., 1997; Vande Berg et al., 1998, 2005). However, Stephens's group isolated their chronic wound fibroblasts from biopsy specimens that come from the ulcer base while all other groups harvested chronic wound fibroblasts from the ulcer edge. Given the heterogeneity in growth response and other functions among human fibroblasts derived from either papillary or reticular dermis (Harper and Grove, 1979; Sorrell et al., 2008; Sorrell and Caplan, 2004), it is not surprising that Stephens's group has identified a different chronic wound fibroblast phenotype from the ulcer bed than the previously described fibroblast phenotype from the ulcer margin.

Nevertheless, the fibroblast phenotypes reported by all groups of investigators are consistent with the variety of functional changes that can be induced by oxidative stress on fibroblasts (Wlaschek and Scharffetter-Kochanek, 2005). As pointed out by Wall et al. (2008), the oxidative effects on fibroblast phenotype and function depend on the level and length of exposure to reactive oxygen species (ROS). Low-level, chronic exposure to ROS accelerates telomere shortening (von Zglinicki, 2002), whereas high levels of ROS induce telomere-independent premature senescence (Song et al., 2005). Another important aspect of the cell "senescence" response to oxidative stress is cell age; aged human dermal fibroblasts demonstrate increased entrance into a senescent state compared with their younger counterparts when exposed to oxidative stress (Gurjala et al., 2005). Furthermore, depending on the physiologic state of the cell, oxidative stress can induce apoptosis rather than senescence (Chen et al., 2000). Because chronic wounds demonstrate findings that indicate the presence of increased ROS (Wlaschek and Scharffetter-Kochanek, 2005), it is certainly possible that the dysfunctional fibroblast phenotypes reported at ulcer margins and the ulcer base are both attributable to oxidative stress.

ROS can be generated by a variety of sources, including H₂O₂ during purine catabolism (xanthine and hypoxanthine oxidase) and superoxide (O₂^-) during mitochondrial metabolism in all cells or the respiratory burst associated with microbiocidal activity in phagocytes (Wlaschek and Scharffetter-Kochanek, 2005; Wulf, 2002). Importantly, according to Wall et al. (2008), chronic wound fibroblasts generate elevated levels of ROS compared with normal dermal fibroblasts, setting up a positive feedback system for wound degeneration.

To understand the mechanisms of ROS-induced alteration and damage to cells, it is important to know that hydroxyl (OH) and peroxy radicals (OOH) are generated from O₂^- and H₂O₂ via superoxide dismutase (SOD) and the Fenton reaction, respectively (Wulf, 2002). SOD generates one oxygen molecule and one hydrogen peroxide molecule (H₂O₂) from two superoxide molecules (O₂^-) by the following reactions, where M is a transition metal that oscillates between its oxidized and reduced state:

The Fenton reaction converts H₂O₂ to hydroxyl and peroxy radicals via the following reactions:

In the Fenton reaction, the transition metal is iron, which oscillates between its reduced and oxidized states. Hydrogen peroxide (H₂O₂) is relatively long-lived in the cell until it is reduced by catalase in the perioxisome or by glutathione peroxidase in the cytoplasm by the following reactions, respectively, where GSH is reduced glutathione and GSSG is oxidized glutathione:

Because of its relatively long life span, H₂O₂ can act as an effector in signal transduction pathways (Wulf, 2002). Examples of genes induced by ROS (probably primarily H₂O₂) include peroxyredoxin I (a thioredoxin

Oxidative stress affects fibroblast phenotype.

peroxidase that is protective against oxidative stress and apoptosis), heme oxygenase-1 (HO-1), and the cystine transporter x_c^–. The redox control of the HO-1 gene is one of the best-studied models of redox regulation. For example, in dermal fibroblasts HO-1 induction may serve as an inducible defense pathway to remove oxidant-liberated heme, which is extremely toxic when free (Kumar and Bandyopadhyay, 2005). Interestingly, heme toxicity probably is attributable in large part to its iron core, which, when released in great abundance, overwhelms the iron-binding capacity in blood and tissue (i.e., ferritin) and therefore can act as a catalyst in the Fenton reaction to generate more ROS. Because HO-1 mRNA is inducible in many tissues and its expression is relatively stable, it is a useful marker for cellular oxidative stress.

As previously stated, ROS can induce cell apoptosis, and this appears to be mediated by H₂O₂ through its activation of the c-Jun N-terminal kinase (JNK) pathway, a subgroup of the mitogen-activated protein kinase family (Shen and Liu, 2006). In general, JNK is activated by environmental stresses, including osmotic shock, UV radiation, heat shock, oxidative stress, and chemotherapeutic agents, as well as the cell death ligands FasL and tumor necrosis factor- and proinflammatory cytokines such as interleukin-1. Among these inducers of JNK, however, oxidative stress seems particularly important. Although activated JNK induces a large family of genes, its pro-apoptotic activity is more direct (Shen and Liu, 2006). In stressed cells activated JNK translocates to the mitochondria, where it phosphorylates and inhibits the anti-apoptotic factor Bcl-2 and phosphorylates and activates the pro-apoptotic agents Bax, Bim, and Bmf. The alteration in balance between pro- and anti-apoptotic factor activities results in the release of cytochrome c from the mitochondria into the cytoplasm. Subsequently, free cytoplasmic cytochrome c activates the downstream effectors of the intrinsic (mitochondrial) apoptosis pathway.

Some years ago, it was thought that H₂O₂ was the primary positive effector of transcription factor NF-B, a central regulator of immunity, inflammation, and cell survival, and that many of the negative effects of ROS were mediated through NF-B, including expression of inflammatory cytokines such as IL-1, IL-6, IL-8, and tumor necrosis factor-. However, new evidence demonstrates that NF-B exerts negative control on ROS and JNK activities and may prevent apoptosis that ROS would otherwise induce through JNK (Bubici et al., 2006).

Not all signal transduction proteins are positively regulated by H₂O₂. Interestingly, hypoxia-inducible factor 1 (HIF-1) is negatively regulated by this ROS. HIF-1 and HIF-1 genes are constitutively expressed, but under normoxic conditions HIF-1 is rapidly degraded by proteasomes in an ROS-dependent manner (Wulf, 2002). Hypoxia decreases the ROS-mediated degradation of HIF-1 and thereby enhances the formation of the heterodimeric complex leading to HIF-1-dependent gene regulation. Many of the gene targets of HIF-1 are important positive regulators of wound healing. In particular, HIF-1 increases the transcription of key factors for blood flow, including vascular endothelial growth factor, HO-1, and both endothelial and inducible nitric oxide synthase (Hellwig-Bürgel et al., 2005). Thus, in the presence of oxidative stress these HIF-1-inducible genes will remain repressed. Parenthetically, it should be pointed out that nitric oxide, a reactive nitrogen species, is also a potent mediator of signal transduction systems, but a discussion of this important topic is beyond the scope of this Commentary.

Hydroxyl and peroxy radicals, in contrast to H₂O₂, are extremely short lived and essentially react with the first susceptible molecule with which they collide. Thus, these so-called free radicals damage cells directly by lipid peroxidation (Figure 1), protein oxidation (Figure 2), and scission of DNA single strands (Figure 3), although the figures oversimplify the actual processes. For example, hydroxyl radials can attack purines and pyrimidines, leading to gene mutations as well as the deoxyribosyl sugar in the DNA backbone (Wulf, 2002). It is important to emphasize that telomeres are particularly sensitive to DNA breaks because they have no DNA repair enzyme system (Wulf, 2002). Another important example of the biologic consequences of ROS-generated free radicals is hydroxyl radical attack on endothelial cell membranes. In addition to compromising the membrane itself, hydroxyl radicals appear to directly activate intracellular calcium-dependent adhesion molecule on the membrane, which makes the endothelium "stickier" for leukocytes passing in the circulation (Sellak et al., 1994). This both slows blood flow through the wound's microvascular circulation and leads to an influx of inflammatory cells into the wound.

Figure 1.

Lipid peroxidation refers to the oxidative degradation of lipids. It is the process whereby free radicals "steal" electrons from the lipids in cell membranes, resulting in cell damage. This process proceeds by a free radical chain reaction mechanism. It most often affects polyunsaturated fatty acids, because they contain multiple double bonds between which lie reactive hydrogens. As with any radical reaction, the reaction consists of three major steps: initiation, propagation, and termination. (From "Lipid peroxidation," Wikipedia, http://en.wikipedia.org/wiki/Lipid_peroxidation, with permission.)

Full figure and legend (9K)

Figure 2.

Protein oxidation by free radicals. (a) Hydrogen atom elimination at the -carbon of the protein backbone, resulting in backbone fragmentation, or (b) hydrogen atom elimination at side chains, resulting in different products, including peroxides, alcohols, and carbonyls. As with lipid perioxidation, protein peroxides are unstable and propagate further reactions. (Modified from the presentation "Protein Oxidation: Concepts, Mechanisms and New Insights," by M.J. Davies, Heart Research Institute, Sydney, Australia, with permission.)

Full figure and legend (12K)

Figure 3.

DNA single-strand breaks are the most common damage inflicted by ROS. The break results from collapse of the sugar that is illustrated in the process at the right. (Modified from the presentation "DNA Oxidation: A Simple Overview," by F.Q. Schafer, University of Iowa, with permission.)

Full figure and legend (16K)

Given all the adverse consequences of elevated ROS levels on cell machinery, it is no wonder that fibroblasts isolated from ROS-rich chronic wounds have a dysfunctional phenotype. These ongoing pathobiologic events in chronic wounds raise the question of how drug therapy can attenuate the impact of excessive ROS generation. Logically, the main target should be one or more of the pathways that generate ROS rather than dysfunctional chronic wound fibroblasts. If the hypothesis of Wall et al. (2008) is correct, relieving oxidative stress in chronic wounds will ameliorate fibroblast dysfunction. Here, clinicians and patients are fortunate to have a readily available drug that appears to attenuate oxidative stress in chronic wounds.

Pentoxifylline has been used for patients with venous leg ulcers for two decades, and a recent meta-analysis by the Cochrane Collaborative Group found that pentoxifylline was effective with a high level of certainty (Jull et al., 2007). Although the mechanisms of action of this drug have not been rigorously defined, they include scavenging hydroxyl radicals (Freitas and Filipe, 1995), inhibiting xanthine oxidase (Hammerman et al., 1999), and reducing circulating levels of tumor necrosis factor- (Fernandes et al., 2008; Zeni et al., 1996). Surprisingly, no other antioxidants have been tested in well-controlled clinical trials for efficacy in venous leg ulcer disease.