Cytokine-based immunointervention in the treatment of autoimmune diseases

INTRODUCTION

Activation of peripheral T cells by foreign and self-antigens is under stringent control by different mechanisms, both thymic and peripheral, and their failure may lead to autoimmune diseases. The progress in understanding the mechanisms of T cell activation and inactivation is currently being translated into strategies able to induce selective immunosuppression to treat different pathological situations, notably autoimmune diseases, as well as allergies and allograft rejection. The medical need for selective immunosuppression is very high, as the available immunosuppressive drugs are substantially inadequate because of limited efficacy, modest selectivity and considerable toxicity.

Key attack points for selective immunointervention have been identified: modulation of antigen recognition, co-stimulation blockade, induction of regulatory cells, deviation to non-pathogenic or protective responses, neutralization of proinflammatory cytokines, induction or administration of anti-inflammatory cytokines and modulation of leucocyte trafficking (Table 1). Therefore, to interfere selectively with the activation of pathogenic T cells, immunosuppressive therapy can be directed to three cellular targets: antigen-presenting cells, autoreactive T cells and suppressor/regulatory T cells, with the common goal to selectively inhibit the activation of pathogenic class II-restricted CD4⁺ T cells.

Table 1.

Selective immunointervention in autoimmune diseases

Targeting the MHC/TCR complex of pathogenic T cells

MHC blockade, deletion, altered peptide ligands

Co-receptors (CD4)

Co-stimulation blockade

Inhibition of interaction between CD28-CD80/CD86; CD154-CD40, LIGHT-HVEM; ICOS-ICOSL; CD134-CD134L

Up-regulation of negative co-regulators (CD152, PD-1)

Immune deviation

Skewing to Th2 via APC or direct T cell modulation

Cytokine-based immunointervention

Inhibition of proinflammatory cytokines (IL-1, IL-2, IFN-γ, IL-12, TNF-α)

Administration of anti-inflammatory cytokines (IL-4, IL-10, IFN-β, TGF-β)

Induction of regulatory T cells

T cell/TCR peptide vaccination, APC manipulation, cytokines (Il-10, TGF-β)

Targeting leucocyte trafficking

Adhesion molecules, chemokines, chemokine receptors

APC = antigen–presenting cells

All these forms of immunointervention have been used successfully to prevent and sometimes treat experimental autoimmune diseases. Based on these results, expectations have been raised for exploiting the same strategies to inhibit the activation of human autoreactive T cells. In this review, we will examine recent advances towards cytokine-based immunointervention in autoimmune diseases, highlighting their successes and failures in clinical applications. Indirect approaches to manipulate the cytokine network that rely on protocols based on oral tolerance or administration of altered peptide ligands to modify the pattern of cytokine production will not be discussed because of space constraints, but they have been reviewed recently [1–3].

CYTOKINE-BASED IMMUNOINTERVENTION

Cytokines are essential components of the immune response, and an imbalance in the cytokine network plays an important role in the initiation and perpetuation of autoimmune diseases [4]. Although preclinical models do not always predict accurately the efficacy of cytokine manipulation in human patients, considerable progress is being made in cytokine-based immunointervention. Essentially, this can be implemented by administration of ‘anti-inflammatory’ cytokines such as interferon (IFN)-β, transforming growth factor (TGF)-β, IL-4 and interleukin (IL)-10, or by neutralizing ‘proinflammatory’ cytokines such as IL-2, IFN-γ, IL-12, tumour necrosis factor (TNF)-α and IL-1.

Several approaches can be used to inhibit a given cytokine: the cytokine itself can be neutralized by specific monoclonal antibodies or by soluble cytokine receptors, whereas the cytokine receptor can be inhibited by monoclonal antibodies or by receptor antagonists competing with the ligand for the receptor binding site. In addition, vaccination against cytokines, e.g. TNF-α, can induce anticyokine autoantibodies that ameliorate autoimmune disease symptoms [5]. Novel technologies to target cytokines are also being developed, for example high-affinity blockers called ‘cytokine traps’, consisting of fusion proteins between the constant region of IgG and the extracellular domain of two distinct cytokine receptor components involved in cytokine binding, have been shown to potently block cytokines in vitro and in vivo[6].

Cytokine-based manipulation offers a unique possibility to interfere with autoimmune diseases. However, the potency of cytokines coupled to the complexity of the cytokine network can lead to severe side-effects, which can still occur after careful preclinical evaluation. The use of cytokine inducers could permit the avoidance of systemic administration of cytokines and their severe side-effects. An alternative approach might rely on the local delivery of anti-inflammatory cytokines by gene therapy [7,8].

NEUTRALIZATION OF PROINFLAMMATORY CYTOKINES

IL-1

IL-1 is a highly proinflammatory cytokine and agents that reduce the production and/or activity of IL-1 are likely to have an impact in the immunotherapy of autoimmune diseases. The production and activity of IL-1, particularly IL-1β, are regulated tightly by three types of inhibitors: IL-1 receptor antagonist (IL-1Ra), soluble IL-1R and membrane type II R, a non-signalling decoy receptor [9]. A human recombinant IL-1Ra (anakinra) has shown both efficacy and safety in a large cohort of patients with active and severe rheumatoid arthritis (RA) [10]. This agent, approved recently for the treatment of RA, has been the first biological to demonstrate a beneficial effect on the rate of joint erosion when used as monotherapy or together with methotrexate [11]. Local gene delivery of IL-1Ra or soluble IL-1R type I has therapeutic efficacy in animal models of RA [12] and has also been tested in a phase I clinical trial [13]. In addition, caspase-1, the IL-1β-converting enzyme that processes IL-1β and IL-18 to their mature forms, could represent a target to prevent autoimmune diseases [14].

IL-2

Targeting IL-2R-expressing lymphocytes is an effective strategy for the prevention of allograft rejection and antibodies specific for the IL-2Rα chain are now in clinical use for this indication. These agents are effective and safe, suggesting the possibility to use the same approach in autoimmune diseases. Humanized anti-IL-2R MoAb therapy, given intravenously with intervals of up to 4 weeks in lieu of standard immunosuppressive therapy, appeared to prevent the expression of uveitis in eight of 10 patients treated over a 12-month period [15]. These initial findings would suggest that anti-IL-2 receptor therapy may be an effective therapeutic approach for uveitis and, by implication, other disorders with a predominant Th1 profile.

DAB[389]IL-2 (denileukin diftitox, ONTAK, Seragen Inc., Hopkinton, MA, USA), an IL-2R-specific ligand fusion protein, has been examined in clinical trials of psoriasis [16] and rheumatoid arthritis, and has shown promising results. The potential utility in other autoimmune disorders is unknown, but diseases such as systemic lupus, scleroderma and vasculitis also may be effective candidates for such ligand fusion therapy [17].

IFN-γ

IFN-γ, produced by activated T and NK cells, is a potent activator of macrophages and monocytes and induces a variety of inflammatory mediators in these cells. It also up-regulates MHC class I and class II molecules, facilitating antigen presentation. Therefore, it is not surprising that a pilot trial of IFN-γ administration to multiple sclerosis (MS) patients resulted in a sharp increase in disease exacerbations [18]. However, the intrathecal delivery of the IFN-γ-encoding gene protects mice from chronic-progressive experimental autoimmune encephalomyelitis (EAE) by increasing apoptosis of central nervous system-infiltrating lymphocytes, challenging the exclusive detrimental role of IFN-γ in the central nervous system (CNS) and indicating that CNS-confined inflammation may induce protective immunological counter-mechanisms [19]. Although IFN-γ may be a reasonable target in itself, most of the ongoing work is concentrated on its inducers, in particular IL-12.

IL-12

IL-12 is a heterodimer composed of two covalently linked glycosylated chains, p35 and p40, encoded by distinct genes [20]. This cytokine, produced predominantly by activated monocytes and dendritic cells but also by other cell types such as microglia [21], enhances proliferation and cytolytic activity of NK and T cells, and stimulates their IFN-γ production [20]. Most importantly, IL-12 induces the development of Th1 cells in vitro and in vivo[20]. The important and non-redundant role of IL-12 in the induction of Th1 responses has been demonstrated in mice deficient for IL-12 [22], IL-12Rβ1 [23], or Stat4 [24].

IL-12-dependent Th1 responses have been implicated in a number of experimental autoimmune disorders, including insulin-dependent diabetes mellitus (IDDM) [25], EAE [26], collagen type II-induced arthritis (CIA) [27], experimental autoimmune uveitis (EAU) [28], granulomatous colitis [29], experimental autoimmune myasthenia gravis (EAMG) [30], and thyroiditis [31]. Neutralization of endogenous IL-12, by anti-IL-12 MoAb or by IL-12R antagonists, has significantly contributed to clarify the important role of this cytokine in the pathogenesis of IDDM [32,33], granulomatous colitis [29] and EAE [26]. Anti-IL-12 MoAb treatment also prevents superantigen-induced EAE and subsequent relapses [34]. Increased IL-12 production by monocytes and increased IFN-γ production by T cells associated with disease activity has been observed in MS patients [35,36]. A recent report demonstrates that neutralization of IL-12 prevents EAE in marmosets, with a protective effect on neuropathological parameters as well as on neurological dysfunction [37]. These findings suggest that targeting IL-12 may prove beneficial in some forms of MS, and it is likely that IL-12 antagonists can be useful in other autoimmune conditions, such as inflammatory bowel disease [38]. Given the critical role of IL-12 in the induction of Th1-mediated autoimmune diseases [25], IL-12 antagonists could be candidates for immunointervention [39]. In addition, small molecular weight agents able to inhibit IL-12 production in vivo and active in autoimmune diseases, such as vitamin D₃ analogues, are becoming attractive [40].

TNF-α

Consistent results from clinical trials have demonstrated the efficacy of TNF-α-neutralizing therapies in different autoimmune diseases, in particular RA. Although several drug targets have been identified for the treatment of RA (Table 2), TNF-α inhibitors have been most successful. Animal studies first documented clearly the important role of TNF-α in RA. Mice transgenic for the human TNF-α gene produce high levels of this cytokine and develop arthritis beginning at 4 weeks of age [41]. In addition, in a model of type II collagen-induced arthritis, administration of antimouse TNF-α, even after disease onset, significantly reduced inflammation and tissue destruction [42].

Table 2.

Major current drug targets in rheumatoid arthritis

TNF-α (inhibitors, antagonists, TACE inhibitors)

Cyclo-oxygenase 2 (COX-2)

Matrix metalloproteases (MMP)

Signal transduction pathways (MAP, p38)

T cells (immunosuppressive agents, MoAbs, vaccines)

Other cytokines (IL-1, IL-2, IL-12)

Adhesion molecules (VLA-4, ICAM)

Chemokine receptors (CXCR3, CCR5, CCR7)

Based on these results, chimeric anti-TNF-α MoAb was administered to RA patients [43]. Treatment with anti-TNF-α was safe and well tolerated, and led to significant clinical and laboratory improvements. After the first administration of anti-TNF-α MoAb remissions lasted, on average, about 3 months. Re-injection of the MoAb, however, induced a significant antiglobulin response in most patients, reducing considerably the efficacy of the treatment. Clinical improvement after anti-TNF-α MoAb therapy was also seen in active Crohn's disease, accompanied by significant healing of endoscopic lesions and disappearance of the mucosal inflammatory infiltrate [44].

A pivotal clinical trial administering multiple intravenous infusions of anti-TNF-α MoAb combined with low-dose weekly methotrexate in RA patients displayed efficacy and a lack of major side-effects [45]. Longitudinal analysis demonstrated rapid down-regulation of a spectrum of cytokines, cytokine inhibitors and acute-phase proteins [46]. IL-6 reached normal levels within 24 h. Serum levels of cytokine inhibitors, such as soluble p75 and p55 TNFR, were reduced, as was IL-1 receptor antagonist. Reduction in acute-phase proteins was also observed. These results are consistent with the concept of a cytokine-dependent cytokine cascade. The degree of clinical benefit noted after anti-TNF-α therapy is due probably to the reduction of many proinflammatory mediators apart from TNF-α.

An alternative approach, using the soluble TNFR p55 chain fused to the constant region of human IgG1 heavy chain (sTNFR-IgG1), has been demonstrated to be about 10-fold more effective than anti-TNF-α MoAb at neutralizing the activity of endogenous TNF, as assessed in a model of listeriosis [47] or in chronic relapsing EAE [48]. This fusion protein appears to achieve the same clinical effects as anti-TNF-α MoAb administration without strong induction of neutralizing antibodies. In a phase II randomized, double-blind, placebo-controlled trial, recombinant human TNFR(p75):Fc fusion protein safely produced rapid, significant and sustained dose-dependent improvement in RA patients [49].

The chimeric anti-TNF-α MoAb (infliximab, Remicade®, Centocor, Malrern, PA, USA) and recombinant human TNFR(p75):Fc fusion protein (Etanercept, Enbrel®, Amgen, Thousand Oaks, CA, USA), both approved by the FDA in 1998, are examples of a new class of disease-modifying anti-inflammatory drugs that interfere with the action of a prototypical proinflammatory cytokine and are effective in RA, psoriatic arthritis and Crohn's disease, besides showing very promising activity in other indications, such as psoriasis and spondiloarthropathies [50,51]. Targeting of cytokines is still in its infancy for therapy of skin diseases, but blocking TNF-α by infliximab or etanercept has shown good efficacy in the management of psoriasis [52]. Both agents show promise in treating a variety of additional autoimmune diseases, but the long-term risks and benefits of these drugs are not yet known. Curiously, these agents show different, although rare, side-effects: infliximab can exacerbate latent tuberculosis [53], and etanercept induces neurological symptoms [54]. In any case, their clear-cut efficacy and relatively modest toxicity demonstrate the power of appropriate immunointervention in autoimmune diseases.

Even though the clinical results of anti-TNF-α therapy in RA and Crohn's disease patients are very exciting, the role of TNF-α in other autoimmune diseases, such as IDDM and EAE/MS, is still puzzling. The fact that anti-TNF-α MoAb treatment initiated before 3 weeks of age prevents insulitis and IDDM suggests clearly that TNF-α may be an essential mediator for the generation and/or activation of autoreactive lymphocytes [55]. Intriguingly, administration of TNF-α to adult non-obese diabetic (NOD) mice could also prevent IDDM, but the mechanism is still unclear [55]. More recently, TNF-α has been shown to partially protect β cells in syngeneic islet grafts from recurrent autoimmune destruction by reducing CD4⁺ and CD8⁺ T cells and down-regulating type 1 cytokines, both systemically and locally in the islet graft [56]. TNF-α thus appears to have a distinct effect on the diabetogenic process depending upon the developmental stage of the immune system and of the target organ, perhaps in a manner analogous to IL-10. These results stress the importance of the time window of cytokine or anti-cytokine treatment to obtain the desired effect. If this concept cannot be translated to clinical practice, conditions to recreate a situation favouring the protective effects of the anti-cytokine treatment should be optimized.

A complex situation exists also in EAE/MS. Although TNF-α has a demyelinating effect in vitro[57] and TNF-α administration enhances EAE [58], TNF-α-deficient mice immunized with myelin oligodendrocyte glycoprotein (MOG) develop severe neurological impairment with extensive inflammation and demyelination leading to high mortality [59]. These results suggest that TNF-α may actually limit the extent and duration of severe CNS pathology. This view is consistent with increased gadolinium-enhancing lesions and lack of efficacy of anti-TNF-α MoAb treatment in MS patients [60]. In MS patients treated with a recombinant TNF receptor p55 immunoglobulin fusion protein (Lenercept Hoffman-La Roche, Basel, Switzerland), there were no significant differences between groups in any MRI measure, but the number of Lenercept-treated patients experiencing exacerbation was significantly increased compared with patients receiving placebo, and their exacerbation occurred earlier [61], closely resembling the more aggressive EAE seen in TNF-α-deficient mice. The surprising outcome of TNF-α targeting in MS, compared to the successful results obtained in other autoimmune diseases, stresses the fact that immunotherapies effective in a given autoimmune condition cannot be translated automatically to any autoimmune disease. In addition, these results raise important questions concerning the pathogenesis of MS, and highlight the apparently positive influence of TNF-α at least in some phases of the disease [62].

ADMINISTRATION OF ANTI-INFLAMMATORY CYTOKINES

IL-4

IL-4 is the cytokine which plays the most important role in Th2 cell development. Using TCR transgenic T cells, it has been shown that IL-4 drives the development of Th2 cells, and this effect is dominant over that of IL-12. An important function of Th2 cells, and hence of IL-4, could be the control of the tissue-damaging effects of proinflammatory cytokines secreted or induced by Th1 cells.

Transgenic NOD mice that express IL-4 in their pancreatic β cells are protected from insulitis and IDDM, indicating the feasibility of a peripheral approach to the treatment of autoimmunity [63]. IL-4 administration to adult NOD mice inhibits IDDM development, although insulitis is not blocked [64]. This is consistent with the hypothesis that islet infiltration is not necessarily associated to IDDM, if Th2-type cells predominate. For example, in NOD male mice, which develop insulitis but only low incidence of IDDM, IL-4 mRNA expression is associated with non-destructive insulitis [65]. In streptococcal-induced arthritis IL-4 administration suppresses the chronic destructive phase [66], and in EAE it inhibits considerably the clinical manifestations of the disease [67]. However, IL-4-deficient mice develop less acute but more chronic-relapsing collagen-induced arthritis, suggesting that IL-4 plays different roles in different phases of the disease [68].

Based on these promising preclinical resuts, a trial of continuous IL-4 administration has been conducted recently in patients affected by severe psoriasis, a Th1-mediated autoimmune disease of the skin. Three daily subcutaneous injections of 0·2–0·5 µg/kg rhIL-4 were well tolerated, improved psoriasis markedly, skewed intralesional cytokines towards an anti-inflammatory Th2 pattern, and induced a two- to threefold increase in IL-4-producing circulating CD4⁺ T cells [69].

Local delivery of IL-4 by gene therapy has also been shown be beneficial without major side-effects in mouse [70,71] and monkey [72] EAE, suggesting that intrathecal delivery of anti-inflammatory cytokine genes may play a role in the future therapeutic armamentarium of inflammatory CNS-confined de-myelinating diseases and, in particular, in the most fulminant forms where conventional therapeutic approaches have, so far, failed to achieve a satisfactory control of the disease evolution [73].

IL-10

IL-10 is a potent suppressor of several effector functions of macrophages, T cells and NK cells. In addition, it contributes to regulate proliferation and differentiation of B cells, mast cells and thymocytes [74]. The most important property of IL-10, from an immunotherapeutic perspective, is its capacity to inhibit Th1 cells. The inhibition of the Th1 cell pathway by IL-10 is mediated by several mechanisms, including inhibition of IL-12 production by antigen-presenting cells (APCs) and blocking of IFN-γ synthesis by differentiated Th1 cells [74]. In addition, IL-10 strongly inhibits production of proinflammatory monokines as IL-1, IL-6, IL-8, TNF-α and GM-CSF as well as of reactive oxygen and nitrogen species following activation of human or mouse macrophages. Thus, IL-10 has strong anti-inflammatory properties [74]. In addition, IL-10 can induce regulatory cells (Tr1) able to inhibit autoimmune diseases [75].

IL-10 injection in adult NOD mice has been shown to decrease insulitis and IDDM [76] indicating that systemic administration of this cytokine may affect the course of autoimmune diseases. However, this may not be true in any situation, as demonstrated by the observation that transgenic IL-10 expression under the insulin promoter does not prevent islet destruction and may actually enhance the inflammatory response [77,78]. The discrepancy between the protective effect of IL-10 administration on adult NOD mice vs. the precipitating effect of transgenic IL-10 expression in pancreatic islets points to the importance of whether cytokines are produced locally or given systemically. Alternatively, the different effects of IL-10 on IDDM induction could be related to the developmental stage of the immune system or of the target cell at the time of cytokine delivery.

It is possible that these inhibitory functions of IL-10 can be exploited clinically and its activity in inhibiting APC functions and Th1 cytokine synthesis suggests a possible use as a non-antigen-specific suppressor factor in the treatment of autoimmune diseases. This possibility would be in accord with the observation that mice failing to make IL-10 because of targeted disruption of its gene develop a severe inflammatory bowel disease [79]. The disease may reflect an overstimulation of Th1 cells, not controlled by Th2-derived IL-10, to gut antigenic stimulation. Supporting this hypothesis, severe colitis was abrogated in a model of inflammatory bowel disease by systemic administration of IL-10 but, interestingly, not of IL-4 [80]. In addition, IL-10 treatment can ameliorate mouse lupus via inhibition of pathogenic Th1 cytokines [81]. IL-10 administration is currently being tested in inflammatory bowel disease and in RA, but the results, so far, are inferior to expectations.

IFN-β

Interferons comprise a family of cytokines that interfere with viral replication. All IFNs increase expression of MHC class I molecules but, unlike IFN-γ, IFN-α and β (Type I IFNs) inhibit MHC class II expression [82]. This could be important to explain their immunosuppressive activity and also, at least in part, the disease enhancing effect of IFN-γ[83].

Cytokine-based immunointervention has been tested extensively in MS (Table 3), and different forms of IFN-β are now an established therapeutic option for relapsing remitting MS [84]. IFN-β treatment reduces attack frequency by 30% and major attacks by an even greater margin. Accumulating disease burden as measured by annual magnetic resonance imaging (MRI) is markedly lessened, and disease activity as measured by serial gadolinium-enhanced MRI scanning is reduced by over 80%. More recently, IFN-β has also been shown to slow down disease progression in secondary progressive MS [85]. The clinical effect of IFN-β is reflected in MRI studies demonstrating a dramatic effect in reducing disease activity. The drug is generally well tolerated, but its efficacy can be compromised in some patients by the emergence of neutralizing antibodies [86]. Although its mode of action in the treatment of MS is still unclear (Table 4), there is ample evidence from in vitro studies that IFN-β directly modulates the function of immune cells. IFN-β treatment was found to increase plasma IL-10 levels transiently, whereas IL-12p40 was not affected. A significantly lower ratio of Th1 versus Th2-type cells was observed in CD8⁺ but not in the CD8⁻ T cell subset [87]. Interestingly, an initial rise in the mean percentage of CD95⁺ T cells and a gradual increase in the mean level of soluble CD95 in plasma was seen [88]. Enhanced IL-10 secretion may have anti-inflammatory effects and the increased CD95 expression may directly interfere with T cell survival. Although a deviation to the Th2 pathway has been implicated, IFN-β treatment in MS patients has been found to induce a profound and persistent down-regulation in the number of circulating T cells secreting IFN-γ and IL-4 [89], thus suggesting a broader immunomodulatory effect of this treatment in MS.

Table 3.

Cytokine-based immunointervention in multiple sclerosis

Approach	Company	Status
Caspase 1 (ICE) inhibitors	BASF/Pfizer	Preclinical
Anti-IL-12	Centocor	Phase I
MMP/TNF inhibitor	British Biotech	Phase I
TGF-β2	NIH	Phase I/II (failed)
sTNFR	Roche	Phase II (failed)
IFN-β1a (Avonex)	Biogen	Phase IV
IFN-β1b (Betaseron)	Schering/Berlex/ emsp;Chiron/Serono	Phase IV

The clinical development status is subject to modifications. Phase IV refers to drugs on the market, labelled for the indication of multiple sclerosis.

Table 4.

Proposed mechanisms of action for IFN-β in multiple sclerosis

Anti-viral

Inhibition of MHC class II expression

Inhibition of IL-12 production

Inhibition of T cell proliferation and IL-2R expression

Up-regulation of IL-4, IL-10 and TGF-β (immune deviation)

Altered cell trafficking

IFN-β treatment was also evaluated in CIA in rhesus monkeys, and in patients with RA [90]. Rapid clinical improvement during IFN-β therapy was observed in three of the four rhesus monkeys with CIA. There was also a marked decrease in serum C-reactive protein (CRP) levels with a subsequent increase after discontinuation of the treatment in all monkeys. The 10 RA patients who completed the study exhibited on average gradual clinical improvement indicating that IFN-β treatment has a beneficial effect on arthritis, and suggesting its possible use as an antirheumatic agent [91].

TGF-β

TGF-β, a molecule known for its pleiotropic activities, can promote or inhibit cell growth and function. TGF-β1 is produced by every leucocyte lineage, including lymphocytes, macrophages, and dendritic cells. It can modulate expression of adhesion molecules, provide a chemotactic gradient for leucocytes and other cells participating in an inflammatory response, and inhibit them once they have become activated [92].

The important role of TGF-β in autoimmune diseases is shown by the massive autoimmune inflammation affecting multiple organs in mice deficient in TGF-β[93] or with induced disruption of the TGF-β type II receptor [94], as well as, for example, by the inhibition of EAE following TGF-β administration [95], and by enhancement of EAE upon its neutralization [96]. In addition, TGF-β is considered a major mediator in oral tolerance [97]. Although the disease-limiting properties of TGF-β in autoimmune diseases seem attractive, disruption of the balance between its opposing activities can contribute to aberrant development, malignancy, or pathogenic immune and inflammatory responses characterized by widespread tissue fibrosis and deposition of extracellular matrix [98]. The safety of TGF-β2 was tested in an open-label trial of 11 patients with secondary progressive MS. There was no change in expanded disability status scale or MRI lesions during treatment, but five patients experienced a reversible decline in the glomerular filtration rate [99]. Systemic TGF-β2 treatment thus appears to be associated with reversible nephrotoxicity, and further investigation of its therapeutic potential should be performed with caution.

PROSPECTS FOR CYTOKINE-BASED IMMUNOTHERAPY OF HUMAN AUTOIMMUNE DISEASES

The efficacy of cytokine-specific treatments in chronic inflammatory disease states, in particular anti-TNF-α therapies in RA and Crohn's disease patients, and IFN-β in MS patients, documents the coming-of-age of cytokine-based immunointervention in autoimmune diseases. Additional indications for the available cytokine-based therapies are being examined, and the prospects look bright. Targeted delivery by gene therapy of cytokines or cytokine antagonists should permit even more effective and less toxic treatments. We will certainly witness the application of more articulate strategies able to selectively target cytokine production by Th1 or Th2 cells or to modify the Th1/Th2 balance in clinical situations. An effective manipulation of pathogenic and protective cells in autoimmunity may eventually rely on cytokine-based approaches that divert autoreactive T cells from autoaggression, while enhancing the frequency and/or the activity of suppressor/regulatory T cells. Work towards this goal is well on its way.