Harmful Waste Products as Novel Immune Modulators for Treating Inflammatory Arthritis?

Cope discusses a new study in rats suggesting that oxidative burst inducers might have a role to play in treating inflammatory arthritis.


The Role of Reactive Oxygen Species
For many years, reactive oxygen species (ROS), including superoxide, hydrogen peroxide, and hydroxyl radicals, and their reaction products, were classically described as harmful by-products of aerobic metabolism capable of causing DNA mutations, lipid peroxidation, and protein oxidation [1]. The identifi cation of enzymes such as superoxide dismutase, catalase, and peroxidase that served to eliminate these waste products rather substantiated this view. It soon became clear, however, that there existed a family of enzymes whose function it was to deliberately generate ROS. The fi rst of these was NADPH oxidase, which is responsible for the respiratory burst in neutrophils and macrophages in response to microbes or infl ammatory cytokines [2]. The catalytic centre of NADPH oxidase is the membrane-associated protein gp91 phox (see Glossary) complexed with p22 phox . Activation requires association with a phosphorylated form of p47 phox (also known as Ncf1), p67 phox , and the small GTPase Rac (Figure 1). p47 phox defi ciency in humans leads to neutrophil dysfunction and chronic granulomatous disease (CGD), a primary immunodefi ciency disorder characterised by the inability to eradicate bacterial infections [3]. Since the late 1990s, six human gp91phox homologues have been identifi ed, each with distinct functions [1].
Besides a role in phagocyte function and host defence, a large amount of evidence points to important roles for ROS in cell proliferation, apoptosis, angiogenesis, endocrinerelated functions, and oxidative modifi cation of the extracellular matrix. Indeed, increased ROS have been documented at sites of infl ammation, such as synovial joints of patients with infl ammatory arthritis (e.g., rheumatoid arthritis [RA]), and circulating neutrophils and monocytes from patients with RA have increased NADPH oxidase activity [4,5].
Interestingly, polyarthritis has also been described in patients with CGD. More recently, induction of arthritis in mice defi cient for the p47 phox subunit of NADPH oxidase was shown to induce granulomatous synovitis and exaggerated matrix destruction associated with enhanced expression of infl ammatory mediators [6]. Although the mechanism for this paradoxical relationship between defi cient NADPH Harmful Waste Products as Novel Immune Modulators for Treating Infl ammatory Arthritis?

Andrew P. Cope
The Perspectives section is for experts to discuss the clinical practice or public health implications of a published article that is freely available online.  Schematic of the molecular composition of the NADPH oxidase complex. The principal subunits are shown, together with some of the key cellular and molecular modifi cations that arise following activation of stress pathways and the generation of superoxide (O 2 − ). Two molecules of superoxide can react to generate hydrogen peroxide (H 2 O 2 ). In the presence of iron, superoxide and H 2 O 2 react to generate hydroxyl radicals (OH•). Through their effects on protein modifi cation and lipid peroxidation, reactive oxygen species exert pleiotropic effects on multiple molecular and cellular pathways.
oxidase activity and infl ammatory arthritis has not been elaborated until now, these fi ndings certainly imply a more complex role for the NADPH oxidase complex in chronic infl ammation than previously thought.

Oxidative Burst Capacity and Infl ammatory Arthritis
Chronic infl ammatory syndromes such as RA are complex polygenic diseases in which it is proposed that inheritance of gene polymorphism predisposes to distinct phenotypes such as the magnitude of the infl ammatory response, the development of autoantibodies, or destruction of cartilage and bone. One productive approach for screening for such gene polymorphism has been to defi ne in relevant rodent models chromosomal segments associated with each disease phenotype. Multiple backcrossing of these chromosomal segments to the disease-prone strain generates congenic animals, which are invaluable tools for further study. For example, linkage analysis in rats susceptible or resistant to pristane-induced arthritis (PIA, an infl ammatory arthritis induced with the alkane oil pristane) has defi ned quantitative trait loci (QTL) that associate with disease severity and cartilage destruction.
In 2003, Holmdahl and colleagues reported the results of a detailed analysis of one such QTL, pia4 , which is implicated not only with arthritis, but also in rodent models of multiple sclerosis and uveitis [7]. Introduction of the 20cM pia4 fragment from the arthritis-resistant E3 strain into arthritisprone DA rats by serial backcrosses substantially reduced arthritis severity. By positional cloning, two structural polymorphisms of the Ncf1 gene (encoding p47 phox ) were identifi ed in arthritis-susceptible DA rats characterised at the amino acid level by M106V and M153T substitutions. While the allelic variant had no effect on Ncf1 gene expression, the disease-associated allele conferred an unexpected reduction in oxidative burst upon stimulation of rat peritoneal cells with phorbol ester, presumably through the effects of the mutations on the function of the NADPH oxidase complex. Strikingly, this defective respiratory burst could be reversed in vivo by treating arthritis-prone DA rats with phytol, an alkane similar to pristane but that induces a robust reactive oxygen burst. Treatment not only restored the respiratory burst but also protected rats from developing arthritis.

A New Rodent Study
In a new rodent study in PLoS Medicine , Holmdahl and colleagues explore in more detail the relationship between oxidative burst capacity and predisposition to infl ammatory arthritis in an approach that has implications for better understanding this paradoxical relationship in humans [8]. First, they compared how alkane derivatives such as pristane and phytol function in terms of oxidative burst capacity and arthritogenicity, and discovered that arthritis induction was independent of oxidative burst. Thus, while both phytol and pristane were potent inducers of oxidative burst in granulocytes in vitro, pristane induced, whereas phytol protected rats from disease. Importantly, phytol treatment restored the oxidative burst of Ncf1 DA rats to levels observed in splenic granulocytes carrying the Ncf1 E3 protective allele, whilst phytol induced only very modest increases in ROS production in T cells.
Second, the authors explored the effects of phytol in different arthritis models, including collagen-induced arthritis, anti-collagen II antibodyinduced arthritis, and non-oil collageninduced arthritis, demonstrating both protective and therapeutic effects in all models; similar effects were observed in rats with normal oxidative burst capacity. Treatment resulted in reduction of the infl ammatory response, attenuation of cell-mediated immunity, and downregulation of markers of cartilage destruction, modifying clinical disease to an extent at least as good as methotrexate or TNF blockade (now considered goldstandard therapy for RA). Histological analysis of joint sections confi rmed the joint-protecting effects of phytol.
Finally, an elegant series of experiments involving the transfer of T cells from pristane-inoculated rats to unmanipulated donor animals demonstrated, among other things, that the effects of phytol on arthritogenic T cells were rapid, since treatment of the donor T cells with phytol for as little as three hours was suffi cient to prevent the development of arthritis in recipient rats. By contrast, adoptive transfer of a mixture of T cells from control and phytol-treated animals failed to inhibit arthritis, suggesting that the disease-modifying effects of phytol on T cells are cell intrinsic, affecting only those cells directly exposed to the therapeutic agent.

Study Implications
Why should enhancing oxidative burst and ROS production paradoxically ameliorate arthritis? There are several possible explanations, none of which would necessarily be mutually exclusive. We know, for example,

Glossary
Adoptive transfer: Transfer of cells from donor to recipient to induce or modify disease gp91 phox : 91-kD glycoprotein subunit of NADPH oxidase, a phagocyte oxidase (hence, "phox")

LAT: Linker for activation of T cells, an adaptor protein involved in transmitting receptor proximal signals from the TCR to downstream intracellular signal transduction pathways
Multiple backcrossing: Sequential breeding of F1 offspring with an animal derived from a particular inbred strain of mouse or rat; usually used to introduce a genetic fragment from one strain to another p22 phox : 22-kD subunit of NADPH oxidase p47 phox (Ncf1): 47-kD subunit of NADPH oxidase (also known as Ncf1) p67 phox : 67-kD subunit of NADPH oxidase (also known as Ncf2) Positional cloning: Sequencing genes in the vicinity of phenotype-linked microsatellites (in this case, Pia4 -linked) to identify the polymorphisms associated with the particular characteristic of interest (in this case, altered severity of arthritis) Pristane-induced arthritis: Induction of arthritis using the alkane oil pristane, commonly used as a model of RA Rac: A small GTP-binding protein involved in cell signalling TCR ζ chain: An invariant subunit of the T cell antigen receptor complex, also known as CD3ζ Thiols: Sulphhydryl groups (e.g., the -SH group is a reduced thiol) that defi ciency of Ncf1 in mice and humans predisposes to repeated insult with infectious pathogens [3,9]. Accordingly, chronic subclinical infection in Ncf1 DA rats could explain predisposition to disease through persistent activation of innate immune responses. However, the authors argue that arthritis occurs in both conventional and specifi c pathogenfree facilities with equivalent severity, implying that the impact of defi cient versus mutant Ncf1 gene products on host immunity might be distinct. Detailed analysis of the response of Ncf1 congenic rat strains to challenge with bacterial pathogens should now be undertaken to rule this out.
Another conundrum arising from this work is the strikingly different levels of ROS induced in granulocytes and T cells after treatment with phytol in vitro; effects on T cell ROS levels were modest at best. So how could phytol regulate T cell reactivity? There is literature pointing to indirect effects of granulocyte-or macrophagederived ROS on T cell phenotype and function. For example, coculture of T cells with ROS-producing neutrophils reduces T-cell reactivity as well as cytokine and proliferative responses [10]. At the molecular level, several mechanisms might be implicated, including alterations of expression or function of key T cell receptor (TCR)signalling molecules including the TCR ζ chain and the transmembrane adaptor protein linker for activation of T cells (LAT) [10][11][12]. The oxidation status of T cells exposed to extracellular ROS might also infl uence the function of phosphatases that are known to be exquisitely sensitive to the oxidation of cysteine in the catalytic site [13]. This in turn might alter thresholds of T-cell reactivity. Increased susceptibility to apoptosis associated with exaggerated oxidative burst has also been reported, but this does not appear to be playing a major role in Holmdahl and colleagues' rat models. Finally, we should not exclude the possibility that Ncf1 could play a role in as-yet undefi ned NADPH oxidase-independent pathways, perhaps exerting a negative regulatory role on intracellular signals involved in immune homeostasis.
One intriguing possibility may relate to recent fi ndings suggesting that cell surface thiols (-SH) are targets of redox regulation [14], which could directly infl uence thresholds of T-cell reactivity [15]. According to this model, a "normal" oxidising extracellular environment would favour oxidation of cell surface molecules and receptors, thereby maintaining T cells in check. Under conditions that predispose to a defective oxidative burst (which would include inheritance of Ncf1 mutations), increased surface thiols would lower thresholds of T cell activation, permitting uncontrolled expansion of effector T cells in vivo. This would presumably infl uence thymic development as well, since lowering TCR-signalling thresholds selects a repertoire of T cells that might otherwise die by neglect when failing to receive appropriate survival signals.
These intriguing results are important because they illustrate another means whereby redox reactions might be manipulated in the clinic. This would be especially attractive if, as this study demonstrates, the effects are rapid and reversible, because robust immunological and molecular tools are already available to monitor such interventions. This approach would be all the more appealing if genetic variation of subunits of the NADPH oxidase complex is associated with susceptibility to autoimmunity in humans. While the data provide unambiguous evidence for a dual role for the oxidative burst in innate and adaptive immune responses, it is now abundantly clear that waste products are not all bad.