MILLENNIUM REVIEW The application of gene therapy in autoimmune diseases

(2000) 7, 9 13 2000 Macmillan Publishers Ltd All rights reserved 0969-7128/00 $15.00 www.nature.com/gt MILLENNIUM REVIEW The application of gene therapy in autoimmune diseases Stanford University School of Medicine, Department of Medicine, Division of Immunology and Rheumatology, 300 Pasteur Drive, Room S021, Stanford, CA 94305-5111, USA The application of gene therapy in autoimmune disease represents a novel use of this technology. The goal of gene therapy in the treatment of autoimmune disease is to restore immune homeostasis by countering the pro-inflammatory effects of the CD4 + T cells in the lesions of autoimmunity. This can be accomplished by adoptive therapy with transduced T cells which can specifically home to the site of inflammation and secrete regulatory protein(s) to ameliorate the inflammation or by direct targeting of the retroviral vector to activated T cells in the sites of inflammation. Transduction of autoantigen recognizing CD4 + T cells, to secrete anti-inflammatory products, may become the magic bullet to combat the ravages of autoimmune inflammation and tissue destruction. (2000) 7, 9 13. Keywords: gene therapy; autoimmune disease; CD4 + T cell Introduction Current human gene therapy clinical trial protocols have primarily focused on somatic gene replacement and cancer immunotherapy. The application of gene therapy in autoimmune disease represents a novel use of this technology. Over the past year, clinical trials utilizing gene therapy for rheumatoid arthritis have begun. The first trial involved ex vivo infection of synoviocytes with a retroviral vector and re-injection into a single metacarpophalangeal joint. 1 This phase one clinical trial was designed to test for transgene expression, in this case the IL-1 receptor antagonist, (IL-1Ra), since the recipients were scheduled to have joint replacement 1 week after re-infusion of the transduced synoviocytes. As discussed below, several laboratories, including our own, have begun to study the use of targeted gene therapy by using antigen-specific CD4 + T cells as vehicles for local delivery of anti-inflammatory proteins in animal models of organspecific autoimmune diseases. The underlying immunopathogenesis of autoimmune diseases has not been fully elucidated. The precise role of B cells, antigen-presenting cells (APC) and T cells in various autoimmune diseases is unclear. However, CD4 + T cells have been shown to be critically important mediators of organ-specific autoimmune disease from studies in animal models, 2,3 as well as indirect evidence in humans. 4 6 The CD4 + T cells that appear to be important in most organ-specific autoimmune diseases have an effector function consistent with a T helper type 1 (Th 1) polarized cell. 7 Several studies have demonstrated the development of Th2-like autoantigen-specific T cells in the recovery phase of disease, suggesting that Th2 cytokines such as IL-4 and IL-10 may be important Correspondence: CM Seroogy for disease recovery. 8 10 This may represent restoration of a more balanced immune response to auto-antigen stimulation and result in disease amelioration. Although alteration of effector function of a differentiated CD4 + T cell appears to be unlikely, 11,12 it appears to be possible to modulate the effector function and differentiation of undifferentiated Th0 T cells and CD4 + T cells that are not completely differentiated. 13 Thus, a desirable approach to the treatment of autoimmune diseases in humans might be the selective delivery of regulatory proteins to sites of inflammation in autoimmune diseases. Several approaches have been taken utilizing animal models to provide a rationale for the application of gene therapy in human autoimmune diseases. Rationale Preliminary data in support of the rationale came from diverse areas. Autoantigen-specific therapies in animal models have studied the effects of altered peptide ligands (APLs) and oral tolerance regimens. These studies have demonstrated that there are important trans-acting effects at the sites of inflammation that modulate the disease outcome. Moreover, these effects did not require precise identification of the autoantigen for therapeutic effect, a critical point particularly for human disease where the autoantigen is frequently poorly characterized. For example, induction of experimental autoimmune encephalitis (EAE), an animal model of multiple sclerosis, with myelin basic protein (MBP) Ac1-11 specific T cell clones can be reversed by treatment with an altered peptide ligand modified for T cells specific for MBP p87-99. 14 Studies in vitro with the APL, using MBP p87-99 specific T cell clones, demonstrated an increase in the ratio of IL- 4:TNF in favor of a Th2 response. Reverse transcriptase

10 polymerase chain reaction (RT PCR) of spinal cord tissue from APL-treated animals revealed no detectable TNF mrna, suggesting that the APL might be working by a similar mechanism in vivo. These authors went on to demonstrate that the simultaneous treatment of the animal with IL-4 blocking antibodies abrogated the effects of the APL treatment. These findings provided support for the concept of gene therapy to replace peptide therapy based on the following. First, the immune modulating effects of IL-4 had a trans-effect: the pathogenic T cell clones were still present in the lesion of EAE as demonstrated by PCR for T cell receptor (TCR) V gene usage, but were silenced by the local production of IL-4 and possibly other unidentified immunomodulatory substances. Second, it was not necessary to use the exact epitope that incited the disease to lead to this ameliorating effect. Although APL or peptide administration may seem like a feasible approach for therapy of human autoimmune disease with little inherent risk, there are several problems. For most human autoimmune diseases, the autoantigen is not known. In animal models, where most of these studies have been performed, the disease is typically induced using a particular antigen: eg type II collagen for an animal model of rheumatoid arthritis. It is more likely that there are multiple autoantigens in human autoimmune disease. The diseases may start by an immune response to a particular antigen (either an autoantigen or a conventional antigen), followed by breaking tolerance and determinant spreading against multiple autoantigenic epitopes, thus perpetuating the immune response. Oral therapy (with peptides or proteins) has been studied in animal models and in humans with autoimmune diseases. Important to the concept of gene therapy approaches to treatment of autoimmune diseases, oral tolerance experiments in EAE have demonstrated the feasibility of generating regulatory T cells. Although, identical to the encephalitogenic CD4 + T cells with respect to TCR gene usage, MHC allele, and epitope recognition the regulatory T cells differed in their cytokine production. 15 The generation (or targeting) of these regulatory cells appeared to be local to the inflammatory lesions, since feedings with irrelevant proteins had no therapeutic effect, despite demonstration of the presence of regulatory cells for the irrelevant protein. Similar results have been obtained in other animal models of autoimmune disease. 16 19 The development of these regulatory T cells has been termed active suppression and is felt to be a vital immune mechanism in peripheral tolerance. Clinical trials are underway using this approach in human autoimmune diseases. While most studies are ongoing, completion of a phase II trial of oral myelin basic protein in multiple sclerosis did not show a significant clinical benefit over the placebo group. 20 Although, a phase two clinical trial (n = 60) studying oral chicken collagen type II (CII) in severe rheumatoid arthritis demonstrated a significant decrease in joint swelling in patients receiving CII compared with placebo, 21 the follow-up trial failed to demonstrate efficacy of this therapy. 22 These animal experiments have provided a basis for rational approaches to T cell-mediated gene therapy in autoimmune diseases. Transducing an antigen-specific T cell that could home to the site of inflammation with vectors containing genes that allow expression of regulatory proteins of the type identified in the studies outlined above, could provide local immunomodulatory effects that would bypass the problems inherent with systemic administration of these proteins. T cell-mediated gene therapy experiments using animal models of autoimmune diseases have been few in number because of the inability to transduce murine T cells efficiently. In addition, these experiments have required long incubation periods in vitro, as well as selection methods that may subsequently lead to alteration in the transduced T cells. CD4 + T cells as an approach to targeted gene therapy A T cell-based gene therapy approach, which targets sites of inflammation, could limit the degree of immunosuppression compared with that seen with systemic immunosuppressive agents. In particular, the local expression of regulatory proteins could be directed to sites of autoimmune inflammation to delivery regulatory proteins (ie cytokines or chemokines) with local immunomodulatory effects. This would be a step beyond current therapeutic modalities. It is unclear what effect various cytokines will have on disease processes emphasizing the importance of studies in animal models. One advantage of gene-modified cells is the ability to insert suicide genes into the vector construct allowing the possibility of killing the gene-modified cells if necessary. This was recently demonstrated in human trials using the thymidine kinase gene. 23 Currently, the most efficient way to introduce genes into lymphoid cells is via transduction, utilizing viral vectors. The most widely studied and utilized viral vectors in humans are murine retrovirus derived. The main advantages of retroviruses for transduction are chromosomal integration and stable gene expression. Furthermore, these vectors are not associated with inciting an antiviral inflammatory response, a major consideration for treatment of autoimmune diseases. The major disadvantages and technical hurdles of retroviral vectors are two-fold: (1) the cells must be dividing in order to be transduced; and (2) it is difficult to attain adequate levels of gene expression. In addition, there is some concern regarding insertional mutagenesis. However, the data from human clinical trials, animal studies, and experiments in vitro suggest that the rate of insertional mutagenesis leading to malignant transformation appears to be the same or less than the rate of spontaneous mutagenesis. 24,25 To date, almost all gene therapy trials have involved ex vivo infection of cells with reinfusion after extensive safety studies. The retroviral vectors have been modified to become replication-defective viruses. This ensures that the virus is capable of infecting only the target cell without further propagation of virus and resultant infection of other cells following transfer. This is a critical safety issue, since the presence of replication-competent virus would significantly increase the risk for insertional mutagenesis. Studies with retroviruses in vivo have been small in number and have been predominantly performed in animal models. The future of retroviral gene therapy will involve infusion of modified retroviruses directly into human tissues avoiding the expense and time of ex vivo applications. This approach

is currently hampered by the inability to target the retrovirus selectively to a particular cell, and the requirement for high-titer virus in order to have an acceptable transduction efficiency. Transduction of antigen-specific CD4 + T cells in animal models of autoimmune disease has been hampered by low transduction efficiencies. Several studies have been reported, including transduction of T cell hybridomas from our own laboratory. Collectively, these studies have demonstrated that transduced antigen-specific T cells can lead to disease amelioration. One study demonstrated an ameliorating effect in established disease which has major implications for treatment in human disease. 26 In these studies, the effect was indirectly demonstrated to be local or in the environment of autoantigen presentation. Our study utilizing MBP-specific T cell hybridomas in the animal model EAE, demonstrated a loss of the ameliorative effects of targeted gene therapy if T cell receptor (TCR) negative, IL-4 expressing hybridomas were adoptively transferred. 27 This indirectly suggested that it was not the systemic, but the local effects of the constitutively expressed IL-4. Furthermore, local production was demonstrated by detection of IL-4 mrna in spinal cord tissue of TCR +, IL-4 secreting hybridomas following successful adoptive gene therapy in EAE. Systemic levels of IL-4 were not detectable until well after the initial recovery phase of the disease. Mathisen et al 26 studied the role of IL-10 in preventing the onset of EAE and treating established disease. In this study, T cell clones specific for proteolipid protein (PLP) were transfected in vitro using plasmid DNA and a polybrene-dmso transfection method. The murine IL-10 gene was subcloned downstream of the murine IL-2 promoter to allow antigen-inducible, non-constitutive expression. After selection of the transfectants (requiring weeks in culture with low transfection efficiency), IL-10 producing cells were intravenously injected into mice before and after the onset of disease, resulting in protection or amelioration of disease, respectively. This was inferred to be an antigen-specific response, since adoptive transfer of an IL-10 expressing T cell clone without CNS antigen specificity demonstrated no clinical or histopathologic benefit. Confounding this result was the observation that the transfected and adoptively transferred T cell clone revealed endogenous IL-10 mrna on an RNase protection assay, that was the majority of the IL- 10 message (60% endogenous, 40% transgene). This suggested that the adoptively transferred T cell clone already had a Th2 phenotype. Previously, co-transfer experiments had established the ameliorating effects of Th2 PLP- or MBP-specific T cells in this disease, 28 and since this experiment did not have a nontransduced T cell clone control, the results must be interpreted carefully. Chen et al 29 recently described amelioration of EAE with latent TGF- transduced MBP-specific T cells using a constitutive retroviral vector. In this experiment, a Th1 T cell clone was utilized and was not altered with respect to Th1 mrna levels after transduction (in this study the authors had only a 3% transduction efficiency). Furthermore, the nontransduced control had slightly more severe disease than the immunized control group. The majority of gene therapy studies using antigenspecific T cells have been done in EAE, one other study has been reported that was in the neonatal NOD mouse. Moritani et al 30 used a murine retroviral vector containing the murine IL-10 gene downstream of a constitutive internal promoter to transduce Th1 islet reactive T cell clones in vitro. These cells were adoptively transferred to neonatal NOD mice alone or mixed with unmodified disease-accelerating Th1 clones. The animals that received the unmodified Th1 clones had a significantly higher disease grade (and insulitis) compared with genemodified Th1 clones (66% versus 8%). In addition, by mixing the two populations, the disease incidence was significantly reduced (66% versus 20%) suggesting that this may have therapeutic benefit in early disease. In this study, the presence of the neomycin reporter gene in genomic DNA from recipient islets suggested that the transferred cells migrated to the pancreas and possibly proliferated in that area. Notably, this result stands in contrast to the results published by the same group, and another group, on transgenic expression of IL-10 in NOD mice, 31,32 illustrating the care needed in interpreting the implications of various models of potential immunotherapy. Problems and potential solutions One major obstacle to T cell-mediated gene therapy studies in animal models of autoimmune disease has been the inability of investigators to transduce murine T cells efficiently. Previous studies have utilized T cell hybridomas or required weeks in culture under selection conditions that could potentially alter the T cell phenotype. Our laboratory, along with others, has optimized transduction conditions to allow more meaningful gene therapy studies. 33,34 In these studies, the efficiency of transduction has been increased up to 60% of the antigenspecific CD4 + T cells allowing the possibility of adoptive gene therapy with sufficient numbers of recently transduced CD4 + T cells. Another adaptation of these recent studies has been the incorporation of the marker protein, green fluorescent protein (GFP), as the basis of selection. This has allowed a titration of the delivered regulatory gene product based upon the linear relation demonstrable between GFP expression and the regulatory gene transcribed through use of an internal ribosomal entry site (IRES). Additional studies are being planned with a double IRES vector to allow the insertion of a gene encoding regulatory transcription factors (ie GATA3 11 or T- Bet (personal communication L Glimcher)) in addition to genes encoding the marker and immuno-regulatory proteins. These vectors have been designed to block the inherent TH1 or TH2 cytokine gene expression of the activated, transduced CD4 + T cells, and allow transgene regulatory protein to be expressed in the absence of the counter-regulatory effects of the endogenous pro-inflammatory cytokines. Summary The goal of gene therapy in autoimmune disease treatment is to restore immune homeostasis by countering the pro-inflammatory effects of the CD4 + T cells in the autoimmune lesions. We hypothesize this could be accomplished by adoptive therapy with transduced CD4 + T cells which can specifically home to the site of inflammation and secrete their regulatory protein(s) to ameli- 11

12 orate the inflammation. Transduction of autoantigen recognizing CD4 + T cells, which secrete anti-inflammatory products, may become the magic bullet to combat the ravages of autoimmune inflammation and tissue destruction. Many additional studies need to be performed before this therapy can be applied to human disease. Can these regulatory gene products, cytokines or chemokines or single chain antibodies, affect ongoing disease? The answer is unknown but the question must be addressed before this therapy is applicable to human disease. Is it possible to target the retroviral vector to the pro-inflammatory T cell in the autoimmune lesion, thus obviating the need for expensive and potentially dangerous T cell expansion ex vivo? Again, the answer is unknown but recent advances in re-targeting of retroviral vectors might support this possibility. 35 It may be possible to modify the coat protein of the replication deficient retroviral vector to target the activated CD4 + T cell through cell surface receptors which uniquely mark the pathogenic T cell in the lesion and have high affinity recognition of the modified envelope protein. Thus, the potential of gene therapy for human autoimmune disease has been demonstrated in animal models, but caution must be taken in extrapolating these early successes to human trials. Cautious optimism may well be the appropriate view of this expanding field. The transduction efficiency and homing properties of the autoantigen specific CD4 + T cells, as well as the immunologically silent nature of retroviral transduction should allow rapid advance in this exciting field of gene therapy for autoimmune disease. References 1 Evans CH et al. Clinical trial to assess the safety, feasibility, and efficacy of transferring a potentially anti-arthritic cytokine gene to human joints with rheumatoid arthritis. Hum Gene Ther 1996; 7: 1261 1280. 2 Linington C et al. Induction of persistently demyelinated lesions in the rat following the repeated adoptive transfer of encephalitogenic T cells and demyelinating antibody. J Neuroimmunol 1992; 40: 219 224. 3 Kakimoto K et al. Isolation of T cell line capable of protecting mice against collagen-induced arthritis. J Immunol 1988; 140: 78 83. 4 Almawi WY, Tamim H, Azar ST. Clinical review 103: T helper type 1 and 2 cytokines mediate the onset and progression of type I (insulin-dependent) diabetes. Clin Endocrinol Metab 1999; 84: 1497 1502. 5 Calabresi PA et al. Cytokine gene expression in cells derived from CSF of multiple sclerosis. J Neuroimmunol 1998; 89: 198 205. 6 Skapenko A et al. Altered memory T cell differentiation in patients with early rheumatoid arthritis. J Immunol 1999; 163: 491 499. 7 Liblau RS, Singer SM, McDevitt HO. Th1 and Th2 CD4 + T cells in the pathogenesis of organ-specific autoimmune diseases (see comments). Immunol Today 1995; 16: 34 38. 8 Khoury SJ, Hancock WW, Weiner HL. Oral tolerance to myelin basic protein and natural recovery from experimental autoimmune encephalomyelitis are associated with downregulation of inflammatory cytokines and differential upregulation of transforming growth factor beta, interleukin 4, and prostaglandin E expression in the brain. J Exp Med 1992; 176: 1355 1364. 9 Begolka WS et al. Differential expression of inflammatory cytokines parallels progression of central nervous system pathology in two clinically distinct models of multiple sclerosis. J Immunol 1998; 161: 4437 4446. 10 Mauri C et al. Relationship between Th1/Th2 cytokine patterns and the arthritogenic response in collagen-induced arthritis. Eur J Immunol 1996; 26: 1511 1518. 11 Ouyang W et al. Inhibition of Th1 development mediated by GATA-3 through an IL-4-independent mechanism. Immunity 1998; 9: 745 755. 12 Swain SL. Generation and in vivo persistence of polarized Th1 and Th2 memory cells. Immunity 1994; 1: 543 552. 13 Bradley LM, Yoshimoto K, Swain SL. The cytokines Il-4, IFNgamma and IL-12 regulate the development of subsets of memory effector helper T cells in vitro. J Immunol 1995; 155: 1713 1724. 14 Brocke S et al. Treatment of experimental encephalomyelitis with a peptide analogue of myelin basic protein. Nature 1996; 379: 343 346. 15 Chen Y et al. Regulatory T cell clones induced by oral tolerance: suppression of autoimmune encephalomyelitis. Science 1994; 265: 1237 1240. 16 Bergerot I et al. Active suppression of diabetes after oral administration of insulin is determined by antigen dosage. Ann NY Acad Sci 1996; 778: 362 367. 17 Nagler-Anderson C et al. Suppression of type II collageninduced arthritis by intragastric administration of soluble type II collagen. Proc Natl Acad Sci USA 1986; 83: 7443 7446. 18 Thompson HS, Staines NA. Gastric administration of type II collagen delays the onset and severity of collagen-induced arthritis in rats. Clin Exp Immunol 1986; 64: 581 586. 19 Wang ZY et al. Suppression of experimental autoimmune myasthenia gravis and experimental allergic encephalomyelitis by oral administration of acetylcholine receptor and myelin basic protein: double tolerance. J Neuroimmunol 1995; 63: 79 86. 20 Hafler DA et al. Oral administration of myelin induces antigenspecific TGF-beta 1 secreting T cells in patients with multiple sclerosis. Ann NY Acad Sci 1997; 835: 120 131. 21 Trentham DE et al. Effects of oral administration of type II collagen on rheumatoid arthritis (see comments). Science 1993; 261: 1727 1730. 22 McKown KM et al. Lack of efficacy of oral bovine type II collagen added to existing therapy in rheumatoid arthritis. Arth Rheum 1999; 42: 1204 1208. 23 Bonini C et al. HSV-TK gene transfer into donor lymphocytes for control of allogeneic graft-versus-leukemia (see comments). Science 1997; 276: 1719 1724. 24 Anderson WF, McGarrity GJ, Moen RC. Report to the NIH Recombinant DNA Advisory Committee on murine replicationcompetent retrovirus (RCR) assays (February 17, 1993). Hum Gene Ther 1993; 4: 311 321. 25 Powell SK et al. In vitro analysis of transformation potential associated with retroviral vector insertions. Hum Gene Ther 1999; 10: 2123 2132. 26 Mathisen PM et al. Treatment of experimental autoimmune encephalomyelitis with genetically modified memory T cells. J Exp Med 1997; 186: 159 164. 27 Shaw MK et al. Local delivery of interleukin 4 by retrovirustransduced T lymphocytes ameliorates experimental autoimmune encephalomyelitis. J Exp Med 1997; 185: 1711 1714. 28 Kuchroo VK et al. B7-1 and B7-2 costimulatory molecules activate differentially the Th1/Th2 developmental pathways: application to autoimmune disease therapy. Cell 1995; 80: 707 718. 29 Chen LZ et al. Gene therapy in allergic encephalomyelitis using myelin basic protein-specific T cells engineered to express latent transforming growth factor-beta1. Proc Natl Acad Sci USA 1998; 95: 12516 12521. 30 Moritani M et al. Prevention of adoptively transferred diabetes in nonobese diabetes with IL-10-transduced islet-specific Th1 lymphocytes. J Clin Invest 1996; 98: 1851 1859. 31 Moritani M et al. Transgenic expression of IL-10 in pancreatic islet A cells accelerates autoimmune insulitis and diabetes in non-obese diabetic mice. Int Immunol 1994; 6: 1927 1936. 32 Wogensen L, Lee MS, Sarvetnick N. Production of interleukin 10 by islet cells accelerates immune-mediated destruction of beta cells in nonobese diabetic mice. J Exp Med 1994; 179: 1379 1384.

33 Costa G et al. Targeting rare populations of murine antigenspecific T lymphocytes by retroviral transduction for potential application in gene therapy for autoimmune disease. J Immunol (submitted). 34 Fleugel A et al. Gene transfer into CD4 + T lymphocytes: green fluorescent protein-engineered, encephalitogenic T cells illuminate brain autoimmune responses. Nature Med 1999; 5: 843 847. 35 Kashara N, Dozy AM, Kan YW. Tissue-specific targeting of retroviral vectors through ligand-receptor interactions (see comments). Science 1994; 266: 1373 1376. 13