INTRODUCTION

Whereas acquired autoimmunity is the sine qua non of lupus, the innate immune system plays an integral role in promulgating inflammatory responses and tissue destruction. An integral part of that innate immune response is the production of reactive nitrogen and oxygen intermediates (RNI and ROI). One of the most widely studied RNI, nitric oxide (NO), is overproduced in the setting of lupus activity. Its pathogenic potential in lupus or any other disease lies largely in the extent of its production and the proximity of its synthesis to ROI, such as superoxide (SO). NO and SO react to form peroxynitrite (ONOO ¹), a much more reactive and potentially pathogenic molecule.

There is convincing evidence in murine lupus nephritis that inducible nitric oxide synthase (iNOS) activity increases with the progression of disease and leads to glomerular, joint, and dermal pathology. In addition, ONOO ¹-mediated modifications of proteins and DNA may increase the immunogenicity of these self-antigens, leading to a break in immune tolerance. In humans, there are observational data suggesting that overexpression of iNOS and increased production of ONOO¹ leads to glomerular and vascular pathology. Therapies designed to target iNOS activity or scavenge ROI/RNI have not been tested in humans in part due to concerns over the specificity of many available compounds for their targets. However, several new compounds are in development that offer promise for human trials in the near future.

REACTIVE SPECIES IN INNATE IMMUNITY

Free radicals are highly reactive molecules with one or more unpaired electrons that are important to the innate immune response. In systemic lupus erythematosus (SLE), there is evidence to support the notion that overproduction of free radicals in the absence of infection leads to a break in immune tolerance, increased tissue damage, and altered enzyme function. For the purposes of this chapter, the discussion of reactive species is confined to a select group of reactive oxygen and nitrogen free radicals. Examples of ROI include SO, hydrogen peroxide, and hydroxyl radicals, whereas NO and ONOO ¹ are the RNI to be discussed. Reactive oxygen and nitrogen intermediates (RONI) play an important role in cellular signaling processes when produced at low levels. At higher levels, these molecules can cause direct toxicity to cells and induce modifications to lipids, amino acids, RNA, and DNA.

NO is a membrane-permeable free radical molecule derived from arginine through the catalytic activity of nitric oxide synthase (NOS). There are three isoforms of NOS that are transcribed from three separate genes. All isoforms require dimerization of identical monomers to become active. Each monomer contains a reductase and oxygenase domain. The reductase domain, with assistance from calmodulin, catalyzes the transfer of two electrons from the electron donor nicotinamide adenine dinucleotide phosphate (NADPH), through flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), to heme iron in the oxygenase domain. Heme and tetrahydrobiopterin (BH₄) interact in the oxygenase domain to catalyze a reaction of O₂ and L-arginine to form NO and citrulline (Fig. 26.1).

Fig. 26.1 NO synthesis from arginine by iNOS and cofactors. Electrons (e-) are donated by NADPH to FAD and FMN in the reductase domain. This step requires Ca²⁺ (much higher levels for eNOS and nNOS than for iNOS) and calmodulin. Two cycles of electrons are then transferred by these carriers to heme iron in the oxygenase domain of the adjacent dimmer. This reaction is similar to that in P450 enzymes. The role of tetrahydrobiopterin (BH₄) in this process is unclear, but it may assist in the coupling of NADPH oxidation and NO formation, thus preventing SO formation. With arginine and O2 as subtrates, donated electrons then catalyze two reaction steps, the formation of N^ω-hydroxy-L-arginine (NHA) followed by conversion of NHA to NO and citrulline.¹

Two isoforms (endothelial or eNOS and neuronal or nNOS) are generally constitutively expressed and are dependent on sufficient concentrations of calcium for activity. In the vascular system, NO (once termed endothelium-derived relaxing factor, or EDRF) produced by eNOS is a potent vasodilator and regulator of vascular tone in response to shear stress. Nitroglycerin mimics the activity of eNOS by acting as a donor of NO.¹ The beneficial effect of NO produced by the constitutively expressed NOS isoforms is blunted when NO is produced in or near cells producing high levels of ROI (as discussed later in the chapter).

A third NOS gene (NOS2) produces an inducible isoform (termed iNOS) that is primarily found in immune cells, most notably macrophages and macrophage-derived cells. iNOS is expressed in response to inflammatory stimuli that are well characterized in murine cells. iNOS is expressed during pathologic states in human endothelial cells, synovial fibroblasts, polymorphonuclear cells, lymphocytes, and natural killer cells.¹ In normal human tissue, expression is strong in myocytes, skeletal muscle, and Purkinje cells.¹ iNOS produces log-fold higher amounts of NO than the constitutively expressed isoforms and produces SO via the reductase domain when arginine substrate is less abundant.¹ NO (when combined with SO) forms peroxynitrite (ONOO¹), a more reactive and toxic molecule than NO itself. ONOO¹ produced by immune cells is capable of killing intracellular pathogens and tumors cells. Glutathione peroxidase, catalase, superoxide dismutase, and antioxidants serve to protect host cells during inflammatory states by reducing the total free radical burden.¹

iNOS aids in microbial defense during certain types of infection. Mice that lack the iNOS (NOS2) gene were less capable of inhibiting the growth of pathogens such as Mycobacterium tuberculosis¹ and developed more significant clinical sequelae of infection when challenged with coxsackievirus.¹ iNOS inhibitor therapy increased the clinical severity of infections in mice challenged with pathogens such as HSV-1¹ and Plasmodium chabaudi.¹ Certain polymorphisms of the NOS2 promoter were studied in young West African Ghanian children with Plasmodium falciparum infection. These polymorphisms were associated with high plasma levels of NO metabolites, lower levels of parasitemia, and a milder disease course in this population.¹ One of these polymorphisms has been associated with lower rates of malarial attacks in patients from Gabon and higher levels of iNOS activity in vitro,¹ an effect not seen in East African Tanzanian subjects.¹ Whereas chronic expression of ONOO¹ may lead to tumor formation by inducing chronic DNA damage, NO and ONOO¹ also have acute cytostatic, cytotoxic and pro-apoptotic effects on tumor cells.¹ Whether similar NOS2 polymorphisms lead to suppression of tumor growth is not known.

NO has the potential to induce both physiologic and pathologic effects, a dichotomy that pervades the literature. The ability of NO to induce cellular pathology is largely dependent on its conversion to more reactive nitrogen species such as ONOO ¹. In turn, the production of ONOO¹ is dependent on levels of SO in the cellular microenvironment in which NO is released. The following will serve as an example of this concept. When NO is produced in proximity to mitochondria in high redox states, it can react with SO to form ONOO⁻. This molecule can induce apoptosis of that cell via cytochrome c-mediated caspase activation. Mitochondria in this state can be found in activated T cells of lupus subjects more frequently than in healthy controls.¹ Alternatively, if NO is produced in low levels in the absence of ROI it promotes cell survival.¹ Thus, the fate of NO and its ultimate pathogenicity depends on levels of oxygen versus SO in its immediate cellular milieu (Fig. 26.2).

Fig. 26.2 Cellular microenvironment dictates the fate of NO. After synthesis by iNOS, eNOS, or nNOS, NO freely diffuses across membranes, forming a concentration gradient. Within this microenvironment also exists a redox gradient (represented by the bacl rectangle (formed by the presence of oxidant/reductant-coupled species. The redox state thus determines whether NO ultimately forms what are usually benign vs. pathogenix reactive nitrogen species (RNS). For instance, when formed in the presence of O₂, NO can oxidize NO₂ and NO₃. Wherease when formed in the presence of superoxide (SO or O₂⁻), NO oxidizes to form peroxynitrite (ONOO⁻).¹

NITRIC OXIDE BIOLOGY IN MURINE MODELS OF LUPUS

Observational Studies

Although iNOS activity may have beneficial effects in the setting of parasitemia or tumor growth, its overexpression in the setting of lupus disease activity appears to lead to organ damage and an altered immune response. Several studies involving murine models of lupus support this concept. Both MRL/MpJ-Fas^lpr/J (MRL/lpr) and (New Zealand Black × New Zealand White)_F1 (NZB/W) mice develop spontaneous proliferative lupus nephritis. MRL/lpr mice developed increasing levels of urine NO metabolites (nitrate + nitrite or NO_X) in parallel with clinical expression of glomerulonephritis.¹ This increase in iNOS activity was associated with post-translational modifications of proteins, specifically nitration of tyrosines (Tyr) to form 3-nitrotyrosine (3NTyr). Such modifications reduced the activity of catalase in the MRL/lpr kidney. Because catalase removes superoxide, its inactivation may expose cells to increased oxidative stress and accelerate tissue damage or modification.¹

Immune complex formation and tissue deposition appear to be proximal to increased iNOS activity in murine lupus. Supporting this hypothesis is the observation that iNOS inhibitor therapy, although improving renal histopathology, had no effect on glomerular immune complex deposition in MRL/lpr mice.¹ Autoantibodies increase markers of iNOS activity (3NTyr formation) in other antibody-mediated autoimmune diseases as well. For example, serum 3NTyr levels were increased after implantation of human β2-glycoprotein I antibody-producing hybridomas into mice with severe combined immunodeficiency syndrome.¹ A similar link between autoantibody deposition and 3NTyr formation has been observed in anti-glomerular basement membrane (GBM) and myeloperoxidase (MPO) antibody models of glomerulonephritis. In both models, autoantibodies were passively transferred to disease-free mice. This intervention was followed by up-regulation of iNOS protein and formation of 3NTyr in glomerular tissue.¹⁹^–²¹

Manipulation of iNOS in Murine Lupus

Several inhibitor studies suggest that iNOS activity is pathogenic in murine lupus. Blocking iNOS activity in MRL/lpr mice before disease onset with the nonspecific arginine analog L-N^G-monomethyl-L-arginine (L-NMMA) reduced 3NTyr formation in the kidney, partially restored renal catalase activity, and inhibited cellular proliferation and necrosis within the glomerulus.^16,^17,²² This effect occurred in the absence of a change in immunoglobulin or complement deposition in the glomerulus, suggesting that increased iNOS expression occurred after immune complex deposition and complement activation.¹ These results were confirmed using the partially selective iNOS inhibitor l-N6-(1-iminoethyl)lysine (L-NIL) to treat mice prior to disease onset. In the L-NIL-treated mice, glomerular histopathology was significantly improved over controls and slightly better than in L-NMMA-treated mice. However, proteinuria was only partially inhibited in the L-NIL-treated mice, whereas L-NMMA-treated mice developed no significant proteinuria. L-NMMA therapy in NZB/W mice that were already expressing clinical nephritis had a similar but less profound effect on proteinuria and renal histopathology than did preventative therapy. However, L-NMMA as monotherapy for the treatment of active disease was less effective in the rapidly progressive MRL/lpr model.¹

In conflict with the effectiveness of pharmacologic iNOS inhibition in murine lupus is the observation that iNOS-/- MRL/lpr mice, although having reduced signs of vasculitis and IgG rheumatoid factor production, had similar glomerular pathology in their MRL/lpr wild-type litter mates.¹ The mechanisms behind the discrepancy of treatment response between pharmacologic and genetic manipulation of iNOS activity are still under investigation. One possibility is that inhibition of iNOS with arginine analogs may reduce pathology through non-iNOS-mediated mechanisms. For instance, L-NMMA may inhibit the y⁺ amino acid transporter system that translocates L-arginine substrate into the cytoplasm.

Both L-arginine and L-NMMA were transported into activated RAW 264.7 macrophages via the same y+ transporter—an effect not seen in resting macrophages, suggesting that this transporter system is inducible. The transport of both molecules was inhibited by leucine.¹ Whether L-NMMA inhibits transport of L-arginine in a similar manner is now known. Thus, intra- and extracellular concentrations of L-arginine may vary according to the presence of other co-transported amino acid derivatives and the level of activation of the transporter system. Supporting the importance of L-arginine availability as a rate-limiting step in NO synthesis in lupus is a study in which supplementation with L-arginine increased renal fibrosis in the MRL/lpr model. Another study dampened enthusiasm for targeting L-arginine by demonstrating no clinical or histopathologic improvement in renal disease with an L-arginine-free diet in MRL/lpr mice.¹

Some interventions that do not directly inhibit iNOS activity may derive additional benefit by reducing expression of iNOS. For instance, chemical induction of heme oxygenase-1 and oral administration of mycophenolate mofetil are both effective therapies for treating glomerulonephritis in MRL/lpr mice (and both reduce iNOS expression in the kidney).²⁶^–²⁸ In contrast, cyclophosphamide therapy increases 3NTyr formation in the setting of rodent models of bone marrow transplantation¹ and cyclophosphamide-mediated bladder toxicity.¹ Although the data regarding NO_X formation in the setting of cyclophosphamide therapy are mixed, it is clear that oxidant stress is increased with its administration.¹ This increased oxidative state can change the fate of NO from NO_X to more toxic ONOO¹. Thus, reductions in iNOS expression and ONOO¹ production with MMF therapy may provide an additional therapeutic benefit over cyclophosphamide therapy for the treatment of lupus nephritis. This theory has not been tested in a rigorous fashion, however.

Potential Mechanisms for Pathogenicity of RNI in Lupus

The mechanisms through which iNOS activity may be pathogenic in SLE and vascular disease have been studied in animal models and in vitro (Table 26.1). See Table 26.1 for examples of the effects of ONOO⁻ on various cellular molecules. As mentioned previously, ONOO¹ (a by-product of iNOS activity) can nitrate protein amino acids and reduce the activity of enzymes. One such enzyme is catalase, which serves to protect host tissues from free radical attack.¹ In vascular tissue, prostacyclin synthase¹ and eNOS¹ are inactivated by ONOO¹, leading to vasoconstriction. These observations suggest that one mechanism through which iNOS activity is pathogenic is via deactivation of tissue-protective enzymes.