Magy Immunol/Hun Immunol 2004;3(4):4-20.
Therapeutic treatment of rheumatoid arthritis by gene therapy-induced apoptosis
James M. Woods, PhD 1 (correspondent),
Michael V. Volin2
1Department of Microbiology and Immunology, Chicago College of Osteopathic Medicine, Midwestern University, Downers Grove, IL 60515. Phone: (1) (630) 515-6173, fax: (1) (630) 515-7245, e-mail: firstname.lastname@example.org;
2Department of Microbiology and Immunology, Chicago College of Osteopathic Medicine, Midwestern University, Downers Grove
Gene therapy was initially conceptualized as a treatment for individuals with genetic disorders, where defective genes would be replaced with functional ones. This concept was eventually broadened to include the use of gene therapy as a delivery mechanism for gene products effective in the treatment of diseases. The latter use of gene therapy, essentially as a drug delivery mechanism, was recognized to be particularly useful in the treatment of rheumatoid arthritis because it may have many advantages over traditional therapies. Two groups of target genes that are potentially useful for gene transfer include soluble inflammatory mediators that in theory could suppress the inflammatory process, and apoptotic mediators that may induce cell death, thereby suppressing the accumulation of inflammatory cells in the joint. To date the former group of target genes has received most of the attention, but it is the latter group of apoptosis-inducing targets that will be discussed in this review. We will focus our discussion on target genes that have shown success at inducing apoptosis in animal models of arthritis and will also include discussion of the apoptotic pathways that are altered in the attempts to reduce inflamed synovial tissue.
apoptosis, rheumatoid arthritis, gene therapy, animal models
Érkezett: 2004. április 7. Elfogadva: 2004. április 15.
In rheumatoid arthritis (RA), the synovial tissue is primarily composed of hyperplastic fibroblasts and activated leukocytes including macrophages and lymphocytes. Recently, it has been suggested that increased proliferation or the lack of apoptosis of synovial cells may contribute to the accumulation of cells in joints of patients with RA. Surgical removal of diseased synovial tissue, or synovectomy, has been performed for four decades and has been shown to produce significant and sustained relief from RA symptoms while increasing joint function1–3. An alternative to surgical removal of synovial tissue is chemical or radiation treatment in which radioactive nucleotides are injected into the inflamed joint resulting in radiological destruction of synovial tissue4. Each of these methods has potential adverse side effects and is not always effective in relieving RA symptoms. As a result, alternative treatments such as genetic synovectomy are being pursued in order to achieve better outcomes. Genetic synovectomy involves the transfer of a gene or genes that will induce apoptosis of synoviocytes resulting in the loss of the pannus tissue in the RA joint. Utilization of gene therapy for this application avoids potential problems seen in other gene therapy methods in that it does not require long-term expression of the transgene. As long as the transgene is efficiently taken up by the synovial cells or it induces a substantial bystander effect in which neighboring cells are also killed, new techniques to improve gene delivery systems may not be required. Recently, advancements have been made in animal models that have improved our ability to specifically induce synoviocyte apoptosis in arthritis, demonstrating gene therapy’s promising potential for the reduction of inflamed synovial tissue in RA patients.
Apoptosis is an important mechanism for deleting self-reactive lymphocytes in the tolerance process. Dysregulated apoptosis may contribute to the pathogenesis of autoimmune diseases. Specifically, insufficient apoptosis may perpetuate the survival of self-reactive lymphocytes in the RA joint leading to the heightened inflammation and synoviocyte hyperplasia. In order to study this hypothesis, several groups studied the prevalence of apoptotic cells in RA synovial tissues. Studying the quantity of DNA strand breaks, an indicator of apoptosis, in RA synovial tissue led to conflicting results with some studies showing few breaks while others reported abundant breaks5–8. However, studies of the morphology of the synovial cells including leukocytes consistently show few apoptotic cells in RA tissues5–7, 9. Thus, it is speculated that the DNA strand breaks seen in studies showing abundant DNA strand breaks may be a result of inflammation and not an indicator of apoptosis9–11. Apoptosis can be initiated through two central pathways, one involving the aggregation of death receptor ligands and the other involving mitochondrial dysfunction. Apoptosis can be suppressed in synoviocytes by inhibition of these pathways. As future studies better define the mediators of apoptosis, potential new therapeutic targets will be revealed. In this review, we will concentrate on gene therapy studies that have induced apoptosis of synovial cells in attempting to decrease the cellularity of the RA joint. Specifically, we will study: 1. Fas/Fas ligand (FasL)-induced apoptosis; 2. tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL); 3. the nuclear factor k B (NF-kB) signaling pathway; 4. the p53 tumor suppressor gene; 5. endothelial cell (EC) apoptosis within the synovial vasculature; 6. galectin-1 (Gal-1)-induced apoptosis; and 7. the use of the herpes simplex virus thymidine kinase (HSV-tk) and ganciclovir gene therapy to achieve synovial ablation.
Fas-mediated apoptosis in rheumatoid arthritis
The differentiation antigen CD95 or Fas is a ubiquitously expressed type I membrane receptor protein in the tumor necrosis factor (TNF) -a receptor family. FasL, a type II membrane protein that is also a member of the TNF-a receptor family, can bind to Fas resulting in the initiation of apoptosis. Fas and FasL can either be expressed as transmembrane proteins or as soluble proteins, sFas and sFasL. Synovial fluids from patients with RA contain both sFas and sFasL12. In arthritic synovial fluids, sFasL may compete with membrane-bound FasL for Fas binding thereby preventing the induction of Fas-mediated apoptosis12–14. In RA synovial tissue, Fas is expressed on the surface of fibroblasts while the expression of FasL is a subject of debate6, 7, 13, 15–17. Initially, it was thought that synovial tissue fibroblasts did not express FasL, but recently Harada and colleagues identified FasL protein by Western blot analysis in protein extracts from RA synovial fibroblasts17. In culture, RA synovial tissue fibroblasts undergo Fas-mediated apoptosis, however in vivo only a small percentage of RA synovial tissue fibroblasts are observed undergoing apoptosis7. Additionally, synovial tissue fibroblasts from osteoarthritic (OA) patients do not undergo Fas-mediated apoptosis in vitro18. This difference in susceptibility to Fas-mediated apoptosis was corroborated by Okamoto and colleagues in co-culture experiments that showed that FasL-transfected cells were cytotoxic to synovial tissue fibroblasts from RA patients but not to synovial fibroblasts from OA patients15. Several groups have shown that the ligation of Fas with agonistic anti-Fas antibodies can induce RA synovial tissue fibroblasts to undergo apoptosis6, 16, 19.
Synovial tissue cells including fibroblasts, macrophages, and T cells can undergo various apoptotic pathways. For thorough reviews of these pathways refer to recent articles by Perlman and Pope10, 20. Flow cytometric analysis revealed that peripheral blood monocytes from healthy individuals express Fas and FasL on their surface21. Additionally, spontaneous monocyte apoptosis could be inhibited by antibodies that blocked Fas-FasL interactions indicating that Fas-FasL interactions were responsible for the induction of apoptosis. Fas is a transmembrane protein with a cytoplasmic domain which upon activation interacts with FADD (Fas-associated death domain protein). This in turn activates caspase-8 leading to caspase-3 activation and apoptosis (Figure 1)22. Alternatively, activated caspase-8 functions by activating BID (BH3-interacting death domain agonist) resulting in reduction of mitochondrial membrane potential23. This change in mitochondrial membrane potential causes BAK (Bcl-2-antagonist/killer) and BAX (Bcl-2-associated X protein) to release cytochrome c from the mitochondria. Cytochrome c can then bind APAF1 (apoptotic protease-activating factor 1) and ATP forming the apoptosome resulting in activation of caspase-9 who similarly to caspase-8 can induce apoptosis via caspase-3 activation24–26. Several molecules have been identified that are inhibitory to the apoptotic process. Flip (FADD-like IL-1b converting enzyme-inhibitory protein) can bind to FADD and inhibit the activation of caspase-827. Additionally, Bcl-2 and Bcl-XL can inhibit mitochondrial dysfunction involved in apoptosis10. In RA, peripheral blood monocytes egress from the circulation and enter the synovial tissue which is rich in inflammatory cytokines including TNF-a and IL-1b. These cytokines, in addition to activating the transmigrating monocytes, can also induce the expression of the anti-apoptotic protein Flip21. Perlman and colleagues showed that RA synovial fluid macrophages highly express Flip. They went on to show that decreasing Flip levels using a chemical inhibitor increased the amount of Fas-mediated apoptosis suggesting that Flip is necessary for RA synovial fluid macrophage survival.
Similar to RA ST monocyte/macrophages, RA synovial tissue fibroblast sensitivity to Fas-mediated apoptosis can be altered by antirheumatic drug treatment or cytokines present in the RA synovial joint. An example of drug treatment altering Fas-mediated apoptosis is seen in cells treated with penecillamine in combination with copper sulfate. These cells have increased cell surface Fas expression and experience increased Fas-mediated apoptosis in response to exposure to an agonistic anti-Fas antibody17. For examples of cytokines present in the RA synovial milieu altering Fas-mediated apoptosis we start with a study by Kobayashi and colleagues which showed that RA synovial tissue fibroblasts that were pretreated with TNF-a were more susceptible to Fas-mediated apoptosis28. This increased susceptibility of TNF-a-treated RA synovial tissue fibroblasts to Fas-mediated apoptosis is possibly due to TNF-a treatment causing a reduction of Flip expression. This was confirmed when overexpression of Flip by gene transfection resulted in a partial reduction of TNF-a-induced Fas-mediated apoptosis. It is known that Flip is normally absent from synovial tissues, however in RA it is expressed by synovial tissue fibroblasts29. Contrasting these findings, Wakisaka and colleagues found cytokine-stimulated RA synovial tissue fibroblasts to be less susceptible to Fas-mediated apoptosis. Specifically, they found TNF-a or IL-1b treatment of RA synovial tissue fibroblasts inhibited agonistic anti-Fas antibody induced apoptosis, while upregulating Bcl-2 and downregulating caspase-2 and caspase-319. In support of this finding, Kawakami and colleagues also found Fas-induced apoptosis of RA synovial tissue fibroblasts to be inhibited by TGF-b16. Thus, cytokines normally found in the RA joint milieu such as TNF-a have been shown to either enhance RA synovial tissue fibroblast Fas-mediated apoptosis by down regulating the expression of Flip or to inhibit RA synovial tissue fibroblast Fas-mediated apoptosis by upregulating Bcl-2 expression.
In animal models of arthritis, Fas activation has been shown to mediate the progression of arthritis. Specifically, intraperitoneal (i.p.) injection of an agonistic anti-Fas antibody into a severe combined immunodeficient (SCID) mouse model of arthritis resulted in apoptosis of synovial cells and decreased cartilage destruction30. In another animal study, arthritic HTLV-1 tax transgenic mice injected intra-articularly (i.a.) with an anti-Fas antibody had increased synovial apoptosis and reduced pain and swelling31. Fas activation has also been shown to help suppress the development of arthritis in animal models. For example, collagen-induced arthritic (CIA) mice injected subcutaneously (s.c.) before the peak of arthritis with Chinese hamster ovary (CHO) cells that overexpress FasL and IL-4 had decreased development of arthritis with reduced clinical severity32. Additionally, rat adjuvant-induced arthritis (AIA) could be prevented by pretreatment with bisinolylmaleimide VIII, a facilitator of Fas-mediated apoptosis suggesting that overcoming blockage of Fas-mediated apoptotic signaling may be useful in the treatment of arthritis33. Thus, studies of arthritis in animal models indicate that activation of Fas-mediated apoptosis by synovial cells may be a powerful target for future treatments of RA.
FasL or FADD gene transfer-induced apoptosis
In 1997, Zhang and colleagues used a gene transfer approach to activate the Fas/FasL pathway in order to alleviate synovial hyperplasia in mouse CIA34. The synovium of CIA mice is similar to human RA synovial tissue in that both express high levels of Fas yet very little FasL. Specifically, Fas is highly expressed by both infiltrating leukocytes and activated synovial cells. In contrast the expression of FasL in arthritic synovial tissue is minimal suggesting that its low level is in part responsible for synovial cell survival in arthritis. In order to determine whether an increase in the expression of FasL by arthritic synovial tissue would alter cell survival and disease progression, an adenovirus containing the gene for FasL was injected both i.a. and peri-articularly into inflamed CIA mouse joints. Mouse ankles that were injected with the FasL-containing adenovirus produced significantly more FasL mRNA and FasL protein relative to control adenovirus injected mouse ankles. More importantly, mouse ankles that were injected with the FasL-containing adenovirus had more apoptotic cells in their synovial lining and sublining than control adenovirus injected mouse ankles. This increase in apoptosis peaked two to four days following injection with the FasL-containing adenovirus and resulted in a dramatic reduction in the number of activated synovial lining and sublining cells and infiltrating leukocytes ten to twelve days later34. Treatment with FasL-containing adenovirus prevented disease progression and significantly reduced the severity of inflammation34. Additionally, unlike intravenous (i.v.) injections of adenovirus that can cause liver and spleen damage, this study utilized local injections of adenovirus directly into the joint and avoided systemic toxicity34. While not determining that the paucity of FasL is the cause of CIA, this study clearly demonstrated that transfer of the gene for FasL directly into arthritic synovial tissue cells can induce their apoptosis resulting in a decrease in synovial membrane thickness and more importantly in a cessation of disease progression.
The work of Zhang and colleagues was followed-up by studies of Okamoto and colleagues where they introduced FasL into human arthritic tissue in a SCID mouse model ex vivo15. Okamoto and colleagues, searching for a method of inducing Fas-mediated apoptosis without the toxic systemic side effects of anti-Fas mAb, created a cell line that over-expressed FasL. Specifically, they transfected human FasL into a murine T lymphoma cell line that does not express Fas resulting in FasL transfectants that can induce apoptosis via cell-to-cell interactions15. These FasL transfectants, when cultured with RA synoviocytes, induced apoptosis while having no effect when cultured with OA synoviocytes. The induction of RA synoviocyte apoptosis by FasL transfectants was shown to be a result of Fas binding to surface FasL in a cell-to-cell interaction. They next studied FasL transfectants in SCID mice engrafted with human RA synovial tissue. Specifically, irradiated FasL transfectants or irradiated mock transfectants were injected into human RA tissue engrafted on the back of SCID mice. Histological examination of SCID mice injected with irradiated mock transfectants showed morphological features typical of RA including infiltration of inflammatory cells and synoviocyte proliferation. In contrast, SCID mice injected with irradiated FasL transfectants displayed almost complete elimination of inflammatory cells including macrophages and lymphocytes by seven days after injection. This loss of inflammatory cells correlated with increased apoptosis present in engrafted tissue of SCID mice receiving FasL transfectants, while no apoptotic cells were present in engrafted tissue of SCID mice receiving mock transfectants15.
Expression of surface FasL has also been utilized to prolong the survival of transfected cells designed to secrete anti-inflammatory cytokines for the treatment of CIA32. Specifically, CHO cells transfected with the genes for anti-inflammatory cytokines interleukin (IL)-4 and IL-10 were also transfected with the gene for FasL in an attempt to evade the immune system’s removal of the transfected cells. This concept of evading the immune system was based on the premise that transplanted cells that express FasL could induce apoptosis of Fas expressing allo-activated T cells resulting in prolonged transplant survival35. Specifically, if you could delay the elimination of the transplanted cells whose function is to express the anti-inflammatory cytokines, then there would be a longer time for expression of the beneficial proteins, resulting in improved treatment. Inflammation in CIA mice could be inhibited by s.c. injection of a small number of FasL and IL-4-transfected CHO cells. This effect was greater than that of cells only transfected with IL-4, though interestingly FasL and IL-4-transfected CHO cells had an unexpected reduced survival time. Upon a closer look it was determined that the significant decrease in inflammation was most likely due to a combination of FasL-mediated apoptosis of activated neutrophils and anti-inflammatory properties of IL-432.
A novel method of introducing FasL into an animal model of arthritis was introduced by Hsu and colleagues. They first showed that wild type C57BL/6 (B6) mice upon injection with the Gram-negative bacteria, Mycoplasma pulmonis, were resistant to chronic arthritis while Fas apoptosis-deficient strains of these mice (B6-lpr/lpr mice or B6-gld/gld mice) initially develop acute synovitis that later develops into chronic erosive arthritis36. The difference in susceptibility to chronic arthritis was not a result of defective clearance of M. pulmonis from the joints of B6-lpr/lpr or B6-gld/gld mice as the number of viable mycoplasma organisms in these joints did not differ from wild type B6 mice. Alternatively, the pattern of development of chronic arthritis present in B6-lpr/lpr and B6-gld/gld mice suggested that it is the result of defective functional Fas-mediated apoptosis subsequent to M. pulmonis infection. In order to confirm this hypothesis, they utilized a novel ex vivo approach to introduce FasL into these mice in order to restore the defective Fas-mediated apoptosis. Specifically, they isolated antigen-presenting cells (APCs) derived from Fas-deficient lpr mice and transfected them ex vivo with adenovirus containing the FasL gene (lpr-APC-AdFasL). These Fas-deficient APC would express FasL on their surface and would be capable of inducing Fas-mediated apoptosis upon encountering Fas expressing cells once introduced i.p. into the chronically arthritic animal36. Control lpr-APC-AdLacZ cells or lpr-APC-AdFasL cells were introduced into B6-gld/gld mice once a week for five weeks following mycoplasma infection. After one week control cell-treated B6 gld/gld mice had more mononuclear cell infiltration of synovial tissue than mice treated with FasL expressing lpr-APC-AdFasL cells. By week eight, chronic arthritis, including synovial hyperplasia and bone erosion, had developed in control cell-treated B6-gld/gld mice, whereas mice injected with lpr-APC-AdFasL cells displayed only mild arthritic features. To determine the function of the FasL expressing lpr-APC-AdFasL cells it was first necessary to determine where they migrated once they were injected into the animal. To do this, APCs were transfected with green fluorescent protein and injected into the animal. The green fluorescent protein allowed for visualization of the APCs in the animal days later. The results confirmed that the APCs did not migrate to the synovium but rather migrated to the cervical and inguinal lymph nodes. As a result, there was no detected increase in synovial tissue apoptosis, instead FasL expressing lpr-APC-AdFasL cells induced apoptosis of T lymphocytes in peripheral lymph nodes36.
Fas-mediated apoptosis can be regulated by altering the expression of proteins throughout the Fas/FasL pathway. One prime candidate gene for gene therapy in RA is FADD. FADD plays an important role in mediating apoptosis of RA synovial cells. Adenovirus expressing FADD can induce time- and dose-dependent apoptosis of RA synoviocytes in culture as determined by DNA fragmentation and nuclear condensation experiments37. Additionally, inhibitors of caspase-8 and caspase-3, both downstream of FADD, could reduce the number of apoptotic cells following FADD adenoviral treatment. Next, the FADD gene containing adenovirus was introduced into a SCID mouse model of RA. Specifically, control adenovirus or the FADD gene containing adenovirus was injected locally into RA synovial tissues engrafted upon the backs of SCID mice. After seven days, histopathological analysis of the engrafted RA tissues revealed that while control adenovirus-injected RA synovial tissues exhibited synoviocyte proliferation, leukocyte extravasation, bone erosion, and cartilage destruction, FADD adenovirus-injected RA synovial tissue lacked synoviocyte hyperplasia and inflammatory cells37. In order to determine the fate of FADD adenoviral-treated RA synovial tissue cells the TUNEL method was performed. The TUNEL method showed that while control-treated RA synovial tissue cells engrafted to SCID mice did not undergo apoptosis, extensive cell death was found in RA synovial tissues treated with the FADD containing adenovirus. Taken together, these studies indicate that manipulation of the Fas/FasL pathway via gene therapy has the potential to ameliorate arthritis.
TRAIL and rheumatoid arthritis
TRAIL is a membrane bound member of the TNF family that may be best known for its ability to specifically induce apoptosis of tumor cells38. TRAIL accomplishes this by cross-linking and oligomerizing its receptors into a signaling complex that ultimately can utilize caspase-8 and caspase-3. TRAIL interacts with at least five known receptors, including death receptor (DR) 4 (TRAIL-R1), DR5 (TRAIL-R2), decoy receptor (DcR) 1 (TRAIL-R3), DcR2 (TRAIL-R4), and the soluble receptor osteoprotegerin39. These receptors mediate diverse activities such as orchestrating the induction of apoptosis, inhibition of apoptosis, and mediating bone resorption39. TRAIL and its receptors DR4, DR5, and DcR2, all exhibit a wide tissue distribution, which does not aid prediction of a function in vivo39. In addition to its characteristic type II transmembrane form, a small proportion of TRAIL is expressed as a soluble form (sTRAIL), following cleavage by cysteine proteases40. The membrane bound form of this molecule may be more active than its soluble counterpart, since crosslinking of sTRAIL enhances its apoptotic activity41. A recent study suggests that TRAIL-deficient mice have a severe defect in thymocyte apoptosis and are hypersensitive to CIA. These results suggest that TRAIL may play a key role in thymocyte selection and autoimmune diseases42.
Several in vitro and in vivo studies have been performed investigating whether interfering with TRAIL or its receptors could be an efficacious treatment in RA. The first of these studies reported that i.a. addition of a recombinant adenovirus producing TRAIL, introduced six days after mouse arthritis onset, significantly ameliorated disease43. In contrast, chronic blockade of TRAIL was also accomplished through the use of a recombinant soluble DR5 protein, which consisted of the extracellular domain of human TRAIL receptor DR5. Daily injection of soluble DR5 to mice with CIA significantly enhanced mouse arthritis when assessed clinically and histologically43. Interestingly in this study, TRAIL did not appear to induce apoptosis of inflammatory cells, but rather, inhibited synthesis of DNA thereby preventing lymphocyte cell cycle progression. Similarly, TRAIL did not appear to regulate inflammatory cell apoptosis in a study examining experimental autoimmune encephalomyelitis. Rather, the data suggested that it inhibited the disease by preventing activation of autoreactive T cells44.
In related studies which also suggest that TRAIL up-regulation may ameliorate arthritis, the data does suggest that TRAIL can induce apoptosis in synovial fibroblasts. Primary synoviocyte infection with an adenovirus producing human TRAIL induced significant apoptosis in rabbit synovial fibroblasts as well as synovial cells derived from three out of five RA patients45. Moreover, adenovirus-mediated TRAIL gene transfer in a rabbit model of arthritis induced synovial lining apoptosis and reduced white blood cell invasion45. In an independent study, six of six primary synovial fibroblasts from RA patients were found to stain positively for DR5 protein by immunohistochemistry, and another eight of eight were positive by flow cytometry46. Further, all of these cells were susceptible to apoptosis mediated by TRAIL or an agonistic monoclonal antibody specific for DR546. An additional study using flow cytometry, however, examined RA synovial fibroblasts from four patients and did not find significant expression of TRAIL receptors, and adenovirus-produced TRAIL failed to induce apoptosis of these cells47. Further, they did not detect TRAIL or its receptors on RA synovial fluid lymphocytes, while TRAIL R3 was detected on CD14+ monocyte/macrophages derived from RA synovial fluid47. Taken together these studies suggest that DR5 protein expression on the cell surface of RA synovial fibroblasts is rather heterogeneous. RA patients that do express TRAIL receptors on synovial cells may benefit from therapies that act through these receptors, similar to the mouse and rabbit studies described here.
Finally, a recent gene-modified cell therapy study used an inventive approach based on collagen II (CII)-pulsed dendritic cells transfected with an adenovirus system containing the TRAIL gene under the regulation of the doxycycline-inducible tetracycline response element48. This tactic is rooted in the knowledge that antigen-specific lymphocytes play a role in synovial proliferation. The novelty of priming the antigen presenting dendritic cells with collagen, and expressing TRAIL when the system is turned on by doxycycline, allows for collagen-specific T cell deletion when induced48, 49. Intraperitoneal administration of collagen-pulsed dendritic cells led mainly to cell migration to the spleen, where specific T cell elimination prevented subsequent migration to the joint. In mice receiving the gene-modified cell therapy described with doxycycline, the incidence of arthritis and T cell infiltration to the joint was significantly reduced. In line with this, in vitro splenic T cell proliferation and induction to produce interferon-g were significantly lowered in this same group of mice compared to controls48. While activated splenic T cells underwent apoptosis, the inducible adenoviral construct did not appear to have toxic effects on the mice nor the dendritic cells that carried them. There was, in addition, a delayed appearance of a mildly reduced form of mouse arthritis when the same TRAIL gene-modified cell therapy was used without CII pulsing48.
NF-kB as an inhibitor of rheumatoid arthritis synovial cell apoptosis
NF-kB, a ubiquitous and inducible nuclear transcription factor, consists of heterodimers or homodimers of the NF-kB family members: p50, p52, p65 (RelA), RelB, and c-Rel. In its active form NF-kB usually consists of a heterodimer of p65 and either p50 or p52. P65, RelB, and c-Rel are transactivating subunits, while p50 and p52 are non-transactivating. Activated NF-kB binds to its specific enhancer elements resulting in the transcription of genes involved in the inflammatory process and cell survival regulation. NF-kB’s ability to stimulate transcription can be inhibited by IkB. Specifically, IkB sequesters NF-kB in the cytoplasm inhibiting it from migrating to the nucleus and stimulating transcription. Stimulation of a cell by the inflammatory cytokine, TNF-a, can cause IkB to degrade resulting in the translocation of NF-kB to the nucleus and subsequent transcriptional activation. Mice deficient in p50 are immunodeficient, while mice deficient in p65 are not viable and die before birth50. Mouse embryonic macrophages and fibroblasts derived from NF-kB-deficient mice undergo apoptosis upon stimulation with TNF-a. These same murine cells can be rescued from apoptosis by expression of NF-kB, indicating that NF-kB can prevent apoptosis of cytokine stimulated macrophages and fibroblasts51. Additionally, inhibition of NF-kB DNA-binding activity by either overexpression of IkB or by chemical inhibition using pyrrolidine dithiocarbamate can suppress the expression of inhibitors of apoptosis (iap) genes resulting in the conversion of apoptosis-resistant cells to cells susceptible to TNF-a-induced apoptosis52–54. NF-kB has also been shown to mediate apoptosis induced by radiation and chemotherapeutic agents. Specifically, suppression of NF-kB DNA-binding activity resulted in increased cell apoptosis as a result of treatment with chemotherapeutic agents or radiation55–57.
In RA synovial macrophages, fibroblasts, and endothelial cells NF-kB is highly activated in contrast to non-inflammatory joints58–60. Conti and colleagues showed that normal peripheral blood monocytes only constitutively express homodimers of NF-kB p50. Conversely, monocyte-derived macrophages, that are usually resistant to apoptosis, express both homodimers of NF-kB p50 and heterodimers of NF-kB p65-p50, indicating a role for NF-kB p65-p50 heterodimers in the prevention of apoptosis61, 62. Pagliari and colleagues showed that inhibition of NF-kB can induce macrophages to undergo apoptosis. Specifically, primary human macrophages treated with pyrrolidine dithiocarbamate became apoptotic independent of additional apoptotic-inducing stimuli63. These studies indicate a role for NF-kB in the paucity of macrophage apoptosis in the RA joint.
Inappropriate survival of T cells and neutrophils contribute to the chronic inflammation in RA. Activation of NF-kB is thought to be a regulatory process in the induction of apoptosis by RA synovial tissue and synovial fluid T cells and neutrophils. Inhibition of NF-kB transcription in normal peripheral blood T cells has been shown to result in apoptosis. Specifically, Kolenko and colleagues showed that NF-kB transcription could be inhibited by using a peptide that masks the nuclear localization sequence of NF-kB inhibiting its migration into the nucleus64. Similarly, Wang and colleagues have shown IFN-b-induced inhibition of neutrophil apoptosis involves NF-kB activation65. Whether RA synovial T cells and neutrophils can be induced to undergo apoptosis by inhibition of NF-kB remains to be determined.
To assess the timing of NF-kB activation in arthritis, Miagkov and colleagues used the streptococcal cell wall (SCW)-induced arthritis rat model. They showed that activated NF-kB was absent in normal rat synovium, however in the SCW model, reactivation of arthritis coincided with strong activation of NF-kB66. They next explored the role of NF-kB in the regulation of apoptosis in their in vivo arthritis model, by treating the animals with a proteasome inhibitor peptide which inhibits NF-kB activation. There was a dramatic increase in apoptotic cells in the rat arthritic joint following proteasome inhibitor peptide treatment. These experiments linked NF-kB activation and the resultant apoptosis suppression with arthritis reactivation in the SCW model.
As mentioned, the cytoplasm of resting cells has NF-kB sequestered in an inactive complex with its inhibitor IkB. Phosphorylation of IkB by IkB kinase (IKK) results in its degradation, which allows NF-kB to freely form dimers and translocate to the nucleus where it initiates transcription. Synovial fibroblasts from arthritic tissue constitutively express two IKK genes IKK1 and IKK267. However upon IL-1 and TNF-a stimulation, it is IKK2 that is responsible for IkB degradation and NF-kB-dependent expression of chemokines, cytokines, and adhesion molecule by these cells68. McIntyre and colleagues continued this work by specifically inhibiting IKK2 in mice with CIA. They utilized a small molecule inhibitor of IKK2, BMS-345541. When used prophylactically, BMS-345541 reduced cytokine expression and blocked both inflammation and joint destruction69. These works show that regulating the expression of IKK2 in arthritic synovial cells can alter the inflammatory environment, suggesting IKK as a potential therapeutic target to regulate the inflammatory effects of NF-kB in arthritis.
Induction of apoptosis using gene therapy to inhibit NF-kB activation
Several studies have been conducted which show that a gene therapy approach can be used to induce apoptosis in synovial fibroblasts by inhibiting the activation of NF-kB. Cytokines such as TNF-a can induce human RA fibroblast cell lines to proliferate and have increased survival. Zhang and colleagues determined that this survival advantage provided by TNF-a involved NF-kB. Specifically they showed that transfection of TNF-a treated primary RA fibroblasts with an adenovirus containing a dominant negative (DN) from of IkB resulted in a majority of those cells undergoing apoptosis70. This effect was specific to NF-kB and TNF-a, as control experiments in which DN IkB adenovirus was transfected into non-stimulated RA synovial fibroblasts, or transfection of control adenovirus into TNF-a-treated RA synovial fibroblasts, failed to induce apoptosis. Andreakos and colleagues were also able to inhibit NF-kB activation in synovial fibroblasts by targeting an upstream molecule IKK, which is essential for cytokine-induced IkB degradation and NF-kB activation. Specifically, they inhibited TNF-a-, IL-1-, and lipopolysaccharide-induced NF-kB activation of RA synovial fibroblasts using an adenovirus containing a dominant-negative form of IKK271. Additionally, they showed that inhibition of IKK2 in RA synovial membrane cells using this adenovirus could reduce the expression of cytokines IL-1b and IL-6, chemokine IL-8, and metalloproteinases 1, 2, 3, and 13. These studies show that inhibition of NF-kB activation in synovial fibroblasts by adenovirus containing upstream inhibiters can both induce apoptosis and reduce expression of inflammatory mediators.
Pagliari and colleagues showed that NFkB is constitutively active in primary human macrophages. Macrophages are long-lived cells that are resistant to many apoptotic stimuli. Using an adenovirus containing a nondegradable super-repressor form IkBa they were able to inhibit the constitutive activation of NF-kB in primary human macrophages and induce their apoptosis63. This same group later found that macrophages isolated from RA synovial fluid can also be induced to undergo apoptosis following their expression of a super-repressor form of IkBa that inhibits NF-kB activation10. This work suggests that synovial macrophages, like fibroblasts, may be induced into apoptosis by adenovirus transfection of upstream NF-kB inhibitors.
The downstream manifestations of the inhibition of NF-kB activation resulting in apoptosis are beginning to be identified. Zhang and colleagues determined that the inhibition of NF-kB activation that they induced using a DN IkB adenovirus in TNF-a-activated RA fibroblasts resulted in increased activation of caspase-3. This finding suggests that NF-kB signaling is responsible for the inhibition of TNF-a-induced caspase 3 activation70. Additionally, this group identified another molecule involved in synovial fibroblast apoptosis which acts downstream of NF-kB activation. X-linked inhibitor of apoptosis (XIAP) is expressed by synovial fibroblasts in response to TNF-a stimulation. Transfection of a DN IkB adenovirus in TNF-a treated primary RA fibroblasts resulted in the loss of expression of XIAP, indicating that NF-kB activation is necessary for its expression. XIAP also protects RA synovial cells from TNF-a induced apoptosis, as inhibition of its expression using an adenovirus containing its antisense gene allows these cells to undergo apoptosis70. Thus, NF-kB activation is central in the increased survival of synovial fibroblasts in part due to its activation of XIAP expression and inactivation of caspase-3 activation.
In addition to in vitro studies, gene therapy approaches to inhibit NF-kB activation have been used in animal models of arthritis. In a SCID mouse model of human RA, primary RA synovial fibroblasts were injected into SCID mouse knees. Four weeks later the SCID mice were infected with an adenovirus expressing DN IkB. Two days later the mice were injected with TNF-a and 24 hours later there was extensive RA synovial fibroblast apoptosis70. Miagkov and colleagues, using an adenovirus containing a supper-repressor IkBa induced apoptosis in joints of pristine-induced arthritic rats and SCW-induced arthritic rats66. In another study, they inhibited NF-kB activation suppressing the recurrence of inflammation in the joints of rats with SCW-induced arthritis66. Specifically, NF-kB decoy molecules, which are double stranded oligodeoxynucleotide sequences that contain NF-kB binding sites, were packaged into liposomes and injected into a SCW rat joint. Injection of NF-kB decoy molecules 24 hours before reactivation of SCW arthritis resulted in the inhibition of arthritis development in that joint and contralateral untreated joints, suggesting a systemic preventative effect. Tomita and colleagues also utilized NF-kB decoy oligodeoxynucleotides in the CIA rat model of arthritis72. In this study, rats treated with NF-kB decoys had decreased arthritis as assessed histologically by decreased paw inflammation and radiologically by decreased joint destruction. Additionally, the NF-kB decoy treatment resulted in a reduction of IL-1 and TNF-a mRNA and protein in arthritic joints. Studying the potential therapeutic role of IKK2 in adjuvant-induced arthritis, Tak and colleagues showed that introducing an adenovirus containing a dominant negative form of the IKK2 gene greatly ameliorated joint inflammation as assessed by paw volume73. These studies indicate that gene therapy approaches targeting NF-kB signaling pathways may be beneficial in treating RA.
The p53 gene and apoptosis
The p53 gene, commonly referred to as a tumor suppressor gene, produces a decisive regulatory nuclear phosphoprotein, called P53. This protein binds specific DNA sequences in its action as a key regulator of cell proliferation and survival74. Production of the P53 protein is initiated by DNA damage, halting the cell cycle and thus allowing time for DNA repair or apoptosis when the damage is severe75. P53 can support apoptosis directly or indirectly through the up-regulation of BAX, a pro-apoptotic Bcl-2 family member. The p53 protein can transcriptionally activate genes that inhibit G1 to S phase progression74. Cell arrest at the G1 stage by p53 protein is mediated via p21WALF/cip1, which inhibits the G1 cyclin-dependent kinase complex, which in turn precludes transcription factor activation and retinoblastoma family member phosphorylation76. In normal cells, p53 protein expression is very low, with a half-life of less than 20 minutes77–79. p53 gene transcription can be controlled by proto-oncogenes such as c-myc. Also, p53 loss is often related to cell transformation in vitro and neoplasia in vivo80–82.
In comparison with synovial tissue lining of OA and non-arthritic patients, RA lining has higher expression of both mutant and wild type p5383, 84. Likewise in vitro, fibroblasts from RA synovial tissue exhibit higher p53 protein expression than dermal fibroblasts or those derived from OA synovial tissue6, 8, 9, 85, 86. It is possible that DNA strand breaks induced by chronic inflammation (i.e. TNF-a and oxygen free radicals) and c-myc expression upregulate p53 in RA synovial tissue9, 84, 87. Despite a DNA fragmentation rate of nearly 50% of cells in RA synovial lining, it is rare to find a cell which completes apoptosis, suggesting that apoptosis may be ineffective due to inactive p536, 8, 9, 75, 86.
The p53 gene has been determined to have somatic mutations, similar to those identified previously in human tumors, in 40% of cultured RA synovial tissue and the fibroblasts which constitute its lining83, 88. The majority of the mutations include transition base changes where some are dominant negative and capable of suppressing wild-type p53 function89. Microdissection of RA synovial tissue followed by PCR demonstrated ample p53 transition mutations, principally in synovial tissue lining cells as opposed to sublining cells, which may be indicative of oxidative stress-induced DNA damage90. Further, mutant p53 subclone clusters were identified, hinting that these synovial cells may be the result of clonal expansion.
As opposed to studies demonstrating substantial p53 expression by synovial cells, a report that utilized anti-p53 antibodies towards wild-type and mutant p53 demonstrated only a low amount of wild-type p53 immunopositivity in RA synovium91. Further, a study performed on RA synovial fibroblasts from German patients determined that no p53 mutations were present, including examination of the known “hot spots” within the p53 genome92. In the same study, however, p53 mutations were identified in clones of three RA synovial fibroblast populations that were derived from the United States92. Therefore, the exact percentage of RA synovial fibroblasts with mutant p53 protein has been difficult to determine. However, the identification of specific DNA mutations and the presence of some mutant p53 is suggestive that p53 may be a crucial regulator of RA synovial tissue fibroblast invasiveness, apoptosis, and proliferation in some patients. Taken together, research evidence to date focusing on RA synovial fibroblasts is consistent with the hypothesis that a reduction in the presence or activation of p53 in synoviocytes may be an underlying cause of their aggressive phenotype and dysfunctional cell cycle. Further, this theory implies that upregulation of wild-type p53 protein and its activity may be of therapeutic benefit.
Studies performed in animal models also shed some light on the role of p53 in arthritis. In rat AIA, Tak and coworkers demonstrated that p53 increases gradually in synovial tissues with a peak that coincides with inflammation93. Late in this model, noteworthy apoptosis can be detected in the synovial tissues93. DBA/1 mice that develop CIA exhibit p53 expression and apoptosis in synovial cells90. In contrast, p53 (–/–) DBA/1 mice exhibit increased severity of arthritis by clinical and histological scoring, with nearly nonexistent apoptosis90. Consistently, expression of collagenase-3, IL-1, and IL-6 are significantly higher in p53 (–/–) joints with CIA than those of wild-type controls90. Therefore, studies in rat AIA and mouse CIA further support the theory that increasing wild-type p53 may be of therapeutic benefit for RA patients.
Aupperle, Firestein, and colleagues inactivated RA synovial tissue fibroblast endogenous p53 protein by transferring the viral E6 protein derived from human papilloma virus type 18 in vitro75. Viral E6 binds p53 protein and activates the ubiquitin-proteinase system. Viral E6-treated RA synovial fibroblasts were more impervious to oxidant-induced apoptosis in vitro and exhibited increased growth and invasiveness into cartilage extracts75. Further, the viral E6 construct could transform normal synovial fibroblasts into a more aggressive phenotype, similar to those found in the RA lining. Subsequently, viral E6 transduction of synovial tissue fibroblasts co-implanted with non-arthritic human cartilage into SCID mice suggests that inhibiting endogenous p53 enhances invasiveness and cellularity94.
Overexpression of wild-type p53 by adenoviral vectors induced apoptosis in human synovial cells in vitro95. Similarly in tumors, p53 gene overexpression induces cancer cell apoptosis in vivo and in vitro96–99. When examined in endothelial cells, adenovirus-mediated wild-type p53 inhibited differentiation in vitro as well as angiogenesis in vivo100. Similar to human synoviocytes, rabbit synovial cells in culture that overproduce wild-type p53 undergo significant apoptosis95. Further, adenovirus-mediated wild-type p53 quickly and extensively induces synovial apoptosis in arthritic rabbit knees, without altering the metabolism of cartilage95. The magnitude and breadth of synovial apoptosis in most animals appeared to surpass the infectious capacity of the adenovirus, suggesting that a bystander killing effect was operative. Unanticipated in this study was a significant reduction in leukocytic infiltrate, which is not likely attributed to apoptosis of the IL-1-producing cells which initiate the arthritis in the model examined. This was ruled out because a similar reduction in inflammation was demonstrated using antigen-induced arthritis95. Moreover, since less than 30% of infiltrating leukocytes are typically infected after intra-articular injection of adenovirus, a direct effect of wild-type p53 on the leukocytes also does not seem likely. In any case, overexpression of wild-type p53 via intra-articular gene transfer effectively mediated synovial apoptosis and reduced inflammation95. Taken together, these studies suggest that up-regulation of wild-type p53 could have therapeutic implications in the treatment of RA.
Endothelial apoptosis within the rheumatoid arthritis joint vasculature
In the RA joint, the highly aggressive and hyperplastic synovial lining receives most of its nutrients from blood vessels, a feature noted early in the disease course101. Angiogenesis is the establishment of new blood vessels from preexisting ones, and this neovasculature is critical to the maturity of the pannus in RA. Consequently, disruption of the endothelial nutrient lifeline is another strategy which may hold therapeutic value. Factors that influence whether angiogenesis will be initiated include the proangiogenic and angiostatic mediators found directly adjacent to an existing blood vessel102–104. Another factor includes inhibition of EC apoptosis, which promotes EC survival105. Similar to initiation of angiogenesis, which is dependent on the balance of angiogenic versus angiostatic factors, there also needs to be a predominance of factors promoting EC survival as opposed to those encouraging apoptosis. These factors also include proangiogenic growth factors, endogenous angiogenic inhibitors, and other signals coming from direct cell-cell contact or from components of the extracellular matrix.
Many proangiogenic growth factors inhibit programmed cell death of ECs105. As a result, removal of growth factors from ECs in vitro results in apoptosis. Growth factors, including basic fibroblast growth factor, angiopoietin-1, and vascular endothelial growth factor (VEGF) promote EC survival by regulating gene expression and post-transcriptional regulation of protein kinases105. One method to reduce neovascularization in the RA synovium is to induce programmed cell death of ECs by targeting either the proangiogenic growth factors or their receptors by gene therapy. An example from the study of pancreatic cancer, with potential implications for RA, used an adenovirus to express the soluble VEGF receptor flt-1, which inhibits VEGF activity in a dominant-negative manner106. In vitro, there was no difference between the proliferation of cancer cells infected with control adenoviruses or those producing flt-1. When these pancreatic tumor cells were introduced into SCID mice and injected with experimental or control adenoviruses, significant differences were noted. Tumor growth in the presence of flt-1 was suppressed in comparison to controls. There was also a lower microvessel density, and an elevated tumor apoptosis index106.
Aside from growth factors, EC interaction with extracellular matrix is necessary for cell survival and angiogenesis. Without extracellular matrix interactions, human ECs quickly undergo programmed cell death, as demonstrated by cell shape, protein cross-linking, nuclei fragmentation, DNA degradation, and expression of apoptosis-specific genes107. Vascular cell integrins interacting with the extracellular matrix molecules, such as vitronectin and fibronectin, have been recognized as key regulators of survival as well as proliferation and invasion during the process of angiogenesis108, 109. Induction of angiogenesis promotes EC entry into the cell cycle along with expression of integrin avb3. Once the process is initiated, avb3 antagonists induce apoptosis of the ECs actively participating in proliferative angiogenesis, without affecting quiescent vessels110. Intercellular adhesion may also play a critical role in angiogenesis and EC survival. For example, serum removal induces EC apoptosis, however, cells can be rescued by homophilic adhesion through platelet EC adhesion molecule-1 (PECAM-1, CD-31) interactions111. While inducing cell survival, PECAM-1 interactions do not assist with proliferation, migration, or spreading of ECs111.
EC apoptosis is accountable for partial actions by endogenous inhibitors of angiogenesis, including angiostatin, endostatin, and thrombospondin-1 (TSP-1). Angiostatin is an internal fragment of plasminogen which partially mediates its angiostatic effects by initiating programmed cell death of ECs112–114. It is unclear whether this apoptotic ability is related to its actions as an inhibitor of an a/b ATP-synthase found on the EC surface115, 116. Adenovirally-produced angiostatin in vitro significantly reduces EC viability, and in an in vivo bFGF-induced model of angiogenesis, gene therapy to produce angiostatin inhibits EC migration and capillary formation117. Assessment of the Matrigel plugs used in these studies displayed ECs with a rounded phenotype and nuclei with characteristics indicative of apoptosis117. Arthritis studies using gene therapy to express angiostatin locally have also been accomplished using the mouse CIA model, and the results suggest that this may be an effective method for decreasing neovascularization118. In this study, NIH3T3 fibroblasts were transduced with a control retrovirus or a retrovirus designed to express angiostatin, and then introduced into the knee. By comparison, angiostatin dramatically lowered arthritis-associated neovascularization, pannus formation, cartilage erosion, and onset of CIA in the ipsilateral paw below the knee118.
Endostatin has also been shown capable of inducing EC apoptosis, however, its main action on EC’s has been recognized as an inhibitor of their migration119–121. Endostatin is a 20 kD C-terminal cleavage fragment of collagen XVIII, which acts in part by binding integrins and inhibiting their normal function122. Endostatin treatment of cow pulmonary artery ECs induces apoptosis, as demonstrated using three different methods120. Moreover, endostatin markedly reduced Bcl-2 and Bcl-XL anti-apoptotic protein without altering levels of Bax protein120. With regards to gene therapy studies, endostatin has been evaluated using models of arthritis and cancer123, 124. In a TNF-a-induced inflammatory arthritis model, prophylactic treatment with a lentivirus producing endostatin significantly lowered mean arthritis index scores as well as blood vessel density of the synovium124. In this study, apoptosis of ECs was not examined, and part of endostatin’s effects could be through interfering with TNF-a-induced JNK activation. A noteworthy observation from this study was that endostatin increased expression of another angiostatic protein, TSP-1, in arthritic mice124.
The antiangiogenic properties of TSP-1 are partly mediated by binding to EC CD36, which initiates an apoptotic pathway which activates caspase-3125. An important observation has been that the apoptotic effects of TSP-1 are limited to those EC’s participating in the angiogenic process, and does not effect ECs which are part of quiescent vessels125. While one study examining tumors suggests the feasibility of TSP-1 as an effective anti-angiogenic treatment in vivo126, a study using an animal model of arthritis did not reduce angiogenesis127. In the first study, using a gene therapy approach, the TSP-1 gene was expressed in a prostate cell line of human origin. Xenografts in mice verified reduced growth in vivo, which may be related to a lower density of microvessels in tumors expressing TSP-1126. In the second study performed in rat AIA, TSP-1 release from Hydron pellets did not appear effective at reducing angiogenesis or inflammation127. However, overexpression of this angiostatic protein via gene therapy may achieve higher levels with better success.
A distinct strategy for inducing neovasculature apoptosis branches off from the finding that av integrins are expressed differently in ECs from established vessels versus those that are newly formed110, 128–130. Similarly, this approach is rooted in the findings that diverse tissues have a unique vascular “address”131. The latter study isolated and sequenced those phage peptide library members that homed to vascular tissue of breast carcinoma xenografts, when the library was introduced to the circulation of nude mice. One peptide, with the sequence Arg-Gly-Asp (RGD), specifically bound to the tumors131. Applying this information to CIA in mice, it was next demonstrated that this phage accumulated in inflamed synovium, as opposed to normal synovium, when introduced i.v.132. Moreover, linking the RGD peptide with a peptide that is pro-apoptotic could target it to neovasculature. Further, systemic administration of this chimeric peptide to mice with CIA demonstrated significantly reduced clinical arthritis scores and apoptosis in small vessels of the inflamed synovium132. It has also been demonstrated that a mutant form of the Raf-1 gene can act down-stream in the Ras signaling pathway, blocking EC Raf activity, thereby promoting apoptosis133–135. In a gene therapy approach, the vector producing this mutant was targeted specifically to angiogenic blood vessels in tumor-bearing mice, again exploiting the knowledge that the avb3 integrin is preferentially expressed on new vessels and contributes to viral internalization133, 136–138. Specificity for avb3-expressing cells was shown in vitro and in vivo133. When the mutant Raf-1 gene was targeted to the blood vessels of newly vascularized tumors in mice, 67% of mice showed no signs of a tumor while the other 33% had tumors with masses that were >95% smaller than controls133. In contrast, mice that had the mutant Raf-1 gene introduced without being targeted to avb3-expressing cells of the established melanoma tumors were euthanized because of the tumor size by day 25133. Also important is that the tumorigenic melanoma cells used did not express avb3 and therefore were not acted upon directly, but rather indirectly via apoptosis of the angiogenic endothelium133. Strategies related to those specifically targeting neovascularization in tumors could have important parallels in RA and may be worth examining within the synovium.
Some factors which promote EC survival, such as VEGF interaction with its Flk-1/KDR receptor, activate the PI3 kinase-Akt and Bcl-2 signaling pathways, thereby protecting ECs from apoptosis139–141. Inhibiting the VEGF pro-survival pathways for ECs initiates caspase activation and induces microvascular regression. However, the efficacy of this strategy can be undermined in vivo by the engagement of redundant pro-survival signals142. Therefore, it was speculated that activating caspase-9 may inhibit vascularization by having direct effects down stream. With this in mind, a retrovirus was designed to produce an inducible caspase-9 (iCaspase9) molecule that is activated by a cell permeable dimerizer drug142. Despite the powerful pro-survival signals from bFGF or VEGF, drug-induced dimerization of iCaspase9 in infected ECs could activate endogenous caspase-3 and activate apoptosis142. Moreover, when functional human microvessels were engineered in immunodeficient mice in vivo, a single i.p. injection of the drug can upregulate iCaspase9, and induce apoptosis of infected ECs142. In short, several of these strategies strongly suggest that the expression and controlled activation of inducible death genes in newly formed blood vessels represent a novel strategy worthy of application to animal models of arthritis.
Gal-1 and arthritis
Another protein of interest which can induce apoptosis in arthritis and has been verified in an animal model, termed Gal-1, is an endogenous mammalian lectin143. It is a constituent of the b-galactoside binding protein family and it is endowed with immunomodulatory and growth regulatory activities144, 145. Gal-1 can induce apoptosis of activated T cells and T leukemia cell lines in humans, and it is expressed by stromal cells of the thymus and lymph nodes145. Resting T cells also bind Gal-1, but it does not induce apoptosis in these cells145, 146. Reasoning that the majority of thymocytes go through programmed cell death in the thymus, Gal-1 apoptotic activity was tested on various thymocyte subsets. The phenotype of cells that Gal-1 regulated included negatively selected and nonselected CD4(lo) CD8(lo) thymocytes147. Further, the data suggested that susceptibility of thymocytes to Gal-1 is partly regulated by cell cycle status147. Rabinovich and coworkers supplied additional evidence suggesting that T cell apoptosis induced by Gal-1 is regulated via activation of the AP-1 transcription factor, the c-Jun proto-oncogene, and reduced levels of Bcl-2148. Moreover, the same group demonstrated that rat macrophages also express a Gal-1-like protein that is significantly increased in inflammatory and activated cells. Further, these studies suggest that the Gal-1-like protein in rats, similar to the human protein, has implications in T cell suicide149, 150.
Nearly each and every one of the characteristics described for Gal-1 thus far suggests that it can be used as a therapeutic modality for RA. Thus, Rabinovich, Chernajovsky, and colleagues performed a gene therapy study to examine the effects of Gal-1 in CIA151. Syngeneic fibroblasts constitutively expressing recombinant Gal-1 at high levels were injected one time i.p. into mice with CIA on the day of onset. The Gal-1 protein ameliorated clinical signs of arthritis when assessed side-by-side with mice receiving control transfectants. Clinical manifestations of arthritis in mice were assessed by clinical scores, the number of affected paws, and paw edema. Similar results were obtained by administering recombinant Gal-1 protein each day for 11 days after onset of CIA151. The Gal-1 gene therapy clinical improvements were closely paralleled when pathological measures of CIA were examined in arthritic joints 12 days after disease onset. Ankles of mice receiving the Gal-1 gene displayed lesser cartilage erosion, mononuclear infiltration, and synovitis. Moreover, the arthritogenic process was changed by Gal-1 to a T helper 2 subset at the level of the draining lymph node. This was demonstrated by lower production of interferon-g and higher production of IL-5 when cells were accompanied by type II collagen in culture. As anticipated, susceptibility was augmented to antigen-induced apoptosis in lymph node cells from mice that received the Gal-1 gene151. Thus, gene therapy with Gal-1, similar to addition of recombinant protein, considerably halted the sequence of events in arthritis as examined by clinical, immunological, and histopathological determinants.
HSV-tk and ganciclovir gene therapy to achieve synovial ablation
Tk is an enzyme in HSV that plays a vital role in the synthesis of viral nucleotides and replication of DNA. In mammalian cells, the HSV-tk transgene product can phosphorylate ganciclovir, an antiviral drug, resulting in a monophosphorylated product. Endogenous kinases can further act on the drug to produce a triphosphorylated nucleotide analog that can be incorporated into the DNA during replication152, 153. When encountered by DNA polymerase, the analog induces a premature termination of DNA synthesis and apoptosis of the cell. Consequently, HSV-tk gene therapy requires a local injection of the transgene, whose product must encounter ganciclovir that has been administered systemically153, 154. In a rabbit AIA model, plasmid DNA encoding the HSV-tk gene transfected sites of vigorous synovial inflammation154. Ganciclovir was given two times per day for three consecutive days. Three weeks later, joints receiving HSV-tk/ganciclovir were examined and found to exhibit reduced inflammation as well as histological signs of synovial cytolysis. Also of key importance, there was no cytolytic damage noted in the articular cartilage and bone which are directly adjacent to the synovial lining154.
This strategy of suicide gene therapy also shows promise in primates in a CIA study153. First, an adenovirus encoding the HSV-tk gene was shown to destroy the vast majority of RA synoviocytes when ganciclovir was added to the medium in vitro. Next, analysis of the “bystander effect” suggested that infection of 10%, a relatively minor portion of the cells, could mediate cell death of 85% of the cellular population using this method. Subsequently, at different doses of adenoviral vectors delivered locally, it was demonstrated that the joints of rhesus monkeys with CIA exhibited transgene expression in synovial tissue surrounding articular cartilage, tendons, bone, and subsynovial adipose tissue153. Equally important, there were no signs of infection of muscle tissue, fat, bone, or cartilage by the marker gene. Additionally, there was no histological evidence that the transgene product was expressed outside of the synovial tissue. This demonstration was crucial, since the combination of ganciclovir with HSV-tk is anticipated to efficiently induce cell suicide of all cells. Finally, arthritic monkeys injected locally with adenovirus expressing HSV-tk, followed by 14 days of ganciclovir administration, exhibited amplified levels of synovial apoptosis and ablation of the synovial lining without side effects153.
Insufficient programmed cell death of inflammatory cells may be an underlying cause of RA pathogenesis, signifying that enhanced apoptosis via mechanistic interference may be advantageous. Moreover, several therapies currently in use for RA, such as methotrexate, inflixamab, and sulfasalazine, may already be acting in part through an apoptotic mechanism10. Here, we reviewed various approaches with potential to achieve genetic synovectomy, through the use of gene therapy with pro-apoptotic transgenes. The existing generation of gene therapy viral vectors falls short of expectations when trying to deliver genes that are required to remain expressed in the transfected cells for long periods of time. However, these same vectors are efficient at infecting synovial cells and accomplishing short term expression, which are required to increase apoptosis and ablate the synovium. The use of gene therapy to induce genetic synovectomy may, through clinical trials in humans, establish some of the strategies reviewed here as viable substitutes for surgery.
Dr. James M. Woods is supported by NIH grant AR050985 and by an Arthritis Investigator Award from the Arthritis Foundation, dr. Michael V. Volin is supported by the Illinois Society for the Prevention of Blindness.
The authors would like to thank Brian Colander for his graphics assistance and Loren Guzik for her editorial assistance.
A RHEUMATOID ARTHRITIS LEHETSÉGES KEZELÉSE: GÉNTRANSZFERINDUKÁLT APOPTÓZIS
A génterápiát eredetileg elsődlegesen olyan betegek kezelésére fejlesztették ki, akiknek valamilyen géndefektusuk van, és a hiányzó gént pótolják ezzel az eljárással. Később az indikációs kört géntranszfer formájában kiterjesztették, ami bizonyos betegségek kezelésére alkalmas gének bevitelét jelenti. Ez utóbbi eljárás bevezetésre került rheumatoid arthritisben is, ahol a génterápiának számos előnye lehet más eljárásokkal szemben. A gének két nagy csoportja mindenképpen alkalmas a terápiás felhasználásra: a szolúbilis gyulladásos mediátorok (például citokinek) genetikai szuppressziója gátolja a synovitist, az apoptózis génjeinek indukciója pedig a programozott sejthalál beindítása révén gátolja a gyulladásos sejtek ízületekben való felhalmozódását. Korábban a citokingéntranszferről már beszámoltunk, ezen öszszefoglalóban az apoptózissal kapcsolatos adatokat tekintjük át. Főleg azon célgénekre koncentrálunk, amelyekről kiderült, hogy arthritis-állatmodellekben apoptózist indukálnak. Emellett azon apoptózissal kapcsolatos mechanizmusokat is áttekintjük, amelyek szerepet játszanak a synovialis gyulladásban.
apoptosis, rheumatoid arthritis, génterápia, állatmodellek