类风湿性关节炎是一种以滑膜炎、关节周围骨质侵蚀和关节破坏为特点的自身免疫性疾病。其病因目前尚不明确，但已知其病程进展与关节局部炎性细胞因子增高密切相关，如白细胞介素1和肿瘤坏死因子等。在炎症或自身免疫因素的刺激下，关节局部炎性细胞因子增高，进而激活一系列下游途径，包括NF-κB和AP-1传导通路，及其调控的下游炎症反应蛋白，如调节炎症介质PGE2水平的环氧化酶2 (Cyclooxygenase-2， COX-2)，引起细胞外基质破坏的金属蛋白酶，其他炎性细胞因子及趋化因子等。基于对RA病理机制的研究，目前对RA的治疗主要包括使用解热镇痛消炎药或激素类药物控制炎症和相关症状，早期或中、晚期合并使用疾病调节性抗风湿药物(disease-modifying antirheumatic drugs，DMARDs)，主要包括针对细胞因子的药物或金制剂等。选择性COX-2抑制剂的发现和使用曾被认为是对包括RA在内的关节炎的治疗的重大突破，因为通过选择性的抑制COX-2而不影响COX-1的活性，该类药物有效的避免了传统非甾体消炎药的消化道副作用，为其长期使用提供了可能性。但是，随着该药的广泛应用，其心血管副作用逐渐引起了医疗界的关注。 糖皮质激素作为一种有效的抗炎药物被广泛应用与包括类风湿性关节炎和哮喘等在内的多种炎症和免疫性疾病的治疗，但长期使用该药可引起多种严重附作用，对其抗炎作用的分子机制的理解有助于对炎症疾病的控制。目前多认为激素的抗炎作用主要是通过激素激活的糖皮质激素受体对转录因子NF-κB和AP-1的抑制作用，并被总结为3种模型：1 )IκB-α上调模型，通过上调NF-κB抑制物IκB-α抑制NF-κB的功能；2)蛋白-蛋白相互作用模型，活化的激素受体直接与c-Jun/AP-1或NF-κB p65亚单位结合，抑制AP-1和NF-κB功能；3)竞争模型，认为活化的激素受体与NF-κB和/或AP-1竞争转录共激活因子如CBP/p300和SRC-1等，进而调节目标基因的转录。然而，这三种模型均有不足之处，例如IκB-α 上调模型现已被证实是高度细胞特异性的，而且不上调IκB-α的激素受体突变体依然能抑制NF-κB活性。而最近的研究发现，另一种激素诱导的蛋白糖皮质激素诱导的亮氨酸拉链蛋白(Glucocorticoid induced leucine zipper，GILZ) 同样可以与 NF 和 AP-1相互作用，并在免疫细胞中介导糖皮质激素的免疫抑制和抗凋亡作用。于2003年，Shi等发现在基质干细胞中糖皮质激素也能诱导GILZ的早期显著增高，他们还证明了 GILZ可通过与转录因子c/EBP的作用抑制PPARγ的表达，并进而抑制基质干细胞的脂肪分化能力。 因此，我们希望通过研究骨髓基质干细胞(Bone marrow derived mesenchymal stem cells， BM-MSCs)中 GILZ 对细胞因子刺激的环氧化酶 2 (Cyclooxygenase-2)表达的调控作用，来进一步阐明糖皮质激素的抗炎机制，并为RA的细胞治疗提供一些思路。选择研究COX-2是因为：1)通过调节炎症介质PGE2的合成，COX-2在类风湿性关节炎(Rheumatoid Arthritis，RA)的发病和病程中起着重要作用，抑制COX-2可作为控制RA症状和疾病进程的有效治疗手段；2) COX-2的启动子区域已研究较为清楚，已知该区域有NF-κB和AP-1结合位点，且在RA发病过程中显著升高的细胞因子可通过激活NF-κB和AP-1通路上调COX-2的转录。选择BM-MSCs进行研究则是因为：1)最近的研究表明BM-MSC参与RA的发病和病程进展，在RA发病过程中，由于局部细胞因子浓度升高，受炎症刺激聚集到关节局部参与组织修复的BM-MSCs被阻滞在不同的分化前阶段，成为参与骨质破坏的成纤维细胞样滑膜细胞(Fibroblast-like Synoviocyte， FLS)。2) BM-MSCs 是一种具有多潜能的干细胞，能在特定条件下分化为成骨细胞，软骨细胞，肌细胞等，同时这类细胞免疫原性低，在体外实验中能诱导T细胞的免疫耐受，动物实验也提示BM-MSCs移植不引起排斥反应，因而这类细胞被认为是用于RA或其它关节疾病细胞治疗的极好的选择。 考虑到年龄因素对骨髓基质干细胞生物特性的影响以及类风湿性关节炎的发病年龄。在第一部分实验中我们选用18月大的C57BL/6J小鼠股骨和胫骨骨髓分离出贴壁细胞，然后经免疫磁珠分选纯化得到CD-11b， CD-11c， CD45R/B220， PDCA-1阴性，sca-1阳性的BM-MSCs；随后使用流式细胞仪技术和免疫荧光染色检测纯化细胞的表面抗原；最后使用不同的诱导分化培养条件诱导BM-MSCs分化为脂肪细胞，成骨细胞和肌细胞，确定了所得细胞的基质干细胞特性。 在第二部分实验中我们构建了过表达小鼠GILZ蛋白的逆转录病毒Ret-GILZ，也使用腺病毒载体pLL3.7构建了小发夹结构RNA以敲除(knock-down)内源性小鼠GILZ表达，通过将构建成功的腺病毒载体和包装质粒pCMV-VSV-G (衣壳蛋白)、pHRΔ8.9 VPR (核心蛋白)共转染293T细胞得到相应的腺病毒，并使用这些病毒分别建立了持续过表达GILZ或GILZ knock-down的BM-MSCs细胞；使用Western-blot和逆转录实时PCR技术检测GILZ表达对COX-2蛋白和基因的表达的影响；使用报告基因检测技术检测荧光素酶活性，了解GILZ表达对NF-κB和AP-1依赖的COX-2转录活性的影响；使用免疫荧光染色技术检测GILZ对NF-κB p65亚单位核易位的作用；使用电泳迁移率检测技术检测GILZ对AP-1与COX-2启动子区结合能力的影响。通过这些实验我们发现 1.细胞因子IL-1□和TNF-α上调BM-MSCs中COX-2蛋白表达水平，且处理时间及浓度均影响刺激作用； 2.过表达GILZ抑制BM-MSCs和基质干细胞系MC3T3-E1受细胞因子刺激上调的COX-2蛋白和mRNA表达； 3.GILZ knock-down抑制BM-MSCs中糖皮质激素刺激的GILZ表达，同时影响激素对细胞因子上调的COX-2表达的抑制作用； 4.过表达GILZ可通过抑制转录因子NF-κB和AP-1的转录活性来抑制COX-2的转录，且表达量越高抑制作用越强； 5.GILZ抑制细胞因子刺激的NF-κB p65亚单位核转运； 6.GILZ抑制细胞因子刺激的AP-1与COX-2启动子区的结合能力。 据此，我们得出结论，在BM-MSCs中GILZ通过抑制NF-κB和AP-1的转录活性抑制细胞因子刺激的COX-2表达，介导糖皮质激素的抗炎作用。几乎与我们实验同期，另一研究组在呼吸道上皮细胞也同样证实了 GILZ通过抑制细胞因子激活的NF-κB通路介导激素的抗炎作用。因此，GILZ极有可能是另一种新型的激素抗炎介质。虽然近年来BM-MSCs因其多潜能分化和低免疫原性特点被视为关节疾病细胞治疗的极佳选择，但在RA动物模型所做的研究发现，移植BM-MSCs未能有效改善RA症状，进一步研究发现RA关节局部炎症细胞因子尤其是TNF-α增高，影响了 BM-MSCs的抗炎作用和分化能力。而在我们的实验中，高表达GILZ能有效抑制细胞因子激活的NF-κB和AP-1信号通路，因而我们认为对BM-MSCs加以修饰，如使其过表达GILZ，仍然有可能成为RA的治疗新手段。更重要的是， GILZ能通过抑制PPAR□的表达调控脂肪细胞的分化，最近我们又发现GILZ还可能通过对PPAR□的调控增强BM-MSCs的成骨细胞分化能力，这也就意味着， GILZ可能介导激素的抗炎作用但不影响骨形成，这样就可以有效避免长期使用激素导致的骨质疏松的并发症，进一步增加了 GILZ的可能临床使用价值。 关键词：GILZ 环氧化酶-2(COX-2) 关节炎 基质干细胞

 Rheumatoid Arthritis is an autoimmune disease characterized by synovitis, articular bone erosion and joint destruction. The etiology of RA remains unclear at present, but it is known that the increase of pro-inflammatory cytokines, including IL-1 and TNF, in local joint is highly related with the progress of disease. Under stimulation of inflammation or auto immune factors, the concentration of cytokines in joints increases. As a result, a series of downstream pathways are activated, including NF-B and AP-1 signaling pathway. Accordingly, several inflammatory proteins regulated by these two transcription factors are upregulated, such as cyclooxygenase-2 (COX-2), which is a key enzyme that regulates PGE2 secretion; matrix metalloproteinase (MMPs), which is responsible for extracellular matrix degradation; other cytokines and chemokines. Current therapies for RA include use of nonsteroidal anti-inflammatory drugs (NSAIDs) or glucocorticoids to control inflammation and symptoms, combination of disease-modifying anti-rheumatic drugs (DMARDs) at early or later stage, including drug therapy targeting cytokines or gold compounds. Usage of selective COX-2 inhibitors was once considered a milestone of RA therapy, since by selectively inhibit COX-2 without affect COX-1, this class of drugs has much less severe gastrointestinal side effects. However, concerns increased about their cardiovascular risk with the increasing prescriptions of these drugs. Glucocorticoids(GCs) are regarded as potent anti-inflammatory drugs and have been widely used in many inflammatory and autoimmune diseases, including rheumatoid arthritis (RA), asthma et al. However, their long-term use can lead to a series of severe side effects. Thus, understanding the molecular mechanism underlying how GCs exert their anti-inflammatory effect will help researchers to identify the mechanism of inflammation and develop new anti-inflammatory drugs with fewer side effects. Till now, it's still widely accepted that GCs exert their anti-inflammatory effect activation via interactions between activated Glucocorticoid receptors and transcription factors NF-κB or AP-1. three main models were established to explain the interactions between GCs and NF or AP-1: 1); the IκB-α up-regulatory model, which proposes that GCs induce the expression of IKB-awhich is an inhibitor of NF-κB. thereby inhibits NF-κB nuclear functions; 2) the protein-protein interaction model, which proposes that activated GRs physically interact with c-Jun/AP-1 and with the NF-κB p65 subunit. resulting in inhibition of genes activated by AP-1 or NF-κB; 3) the competition model, which proposes that the activated GRs compete with NF-κB and/or AP-1 for transcription coactivators such as CBP/p300 and SRC-1, and thereby modulate the transcription of target genes. There is controversy, however, regarding these models. For example, the effect of GCs on IκB-α synthesis and subsequently on NF-κB nuclear translocation is cell-type specific. In addition, a GR mutant that does not enhance IκB-α expression was still able to repress NF-κB activity. It is now known that another glucocorticoid induced protein, GILZ (glucocorticoid induced leucine zipper) can also interact with both NF-κB and AP-1, and mediate GC's immunosuppressive and anti-apoptosis effect in immune cells. Shi et al. discovered GCs can induce GILZ rapidly in MSCs, they also confirmed that GILZ can inhibit PPARγ expression by interacting with c/EBP and block adipocyte differentiation of MSCs. Thus, to further understand how GCs exert their anti-inflammartory effect, we decided to determine the regulation effect of GILZ on cytokine induced COX-2 expression in BM-MSCs. The reasons for choosing COX-2 include 1) COX-2 plays an important role in RA by regulating PGE2 synthesis, inhibition of COX-2 is an effective therapy to alleviate RA symptoms and progression. 2) Cox-2 promoter has been well characterized; it contains functional NF-κB and AP-1 binding sites and is known to be activated by both transcription factors through these sites in response to cytokines. The reasons for choosing BM-MSC include 1) BM-MSCs have been recently implicated in the pathogenesis of rheumatoid arthritis, during the early stage of RA, BM-MSCs are recruited to local joints and arrested at various stages of differentiation by inflammory cytokines, therefore, they turn into rheumatoid arthritis fibroblast like synoviocytes which are involved in joint erosion; 2) MSCs are multi-potential stem cells which can differentiate into different cell lineages including osteoblasts, chondrocytes and muscles, and MSCs have immunosuppressive effects, these cells can induce T-cell tolerance in vitro, and transplantation of these cells do not induce rejection in animal models. Thus, these cells are believed to be good candidates for cell therapy of RA or other joint diseases. Cosidering age has certain effects on the characteristic of BM-MSCs and rheumatoid arthritis occurs usually at middle-age. we used 18 months C57BL/6J mice to isolate adherent cells of bone marrow from femurs and tibias in our first part of experiment; then using negative-immuno-depletion and positive-immuno-selection approaches, we got purified CD-11b, CD-11c, CD45R/B220, PDCA-1 negative and sca-1 positive BM-MSCs, and then detected cell surface antigens using FACS assay and immunostaining procedure; finally, we identified their characteristic of MSCs by doing differentiation experiments. In our second part of experiment, we constructed retrovirus Ret-GILZ, which can induce stable GILZ overexpression, we also constructed small hairpin RNA targeting endogenous murine GILZ based on lentivirus plasmid pLL3.7. By cotransfecting these constructs and packaging plasmid pCMV-VSV-G (envelope protein) and pHRA8.9 VPR (core protein) into 293T cells, we harvested lentiviruses which can effectively knock down GILZ expression; then, by using these viruses, we established MSCs stably overexpress GILZ and MSCs in which GILZ expression is knocked down; then we performed Westem-blot and real-time RT-PCR to determine GILZ's effect on COX-2 expression, both at protein level and at mRNA level; we did reporter gene assay to decide GILZ's effect on NF-κB and AP-1 dependent COX-2 promoter transactivity; we employed immunostaining to detect the effect of GILZ on NF-κB p65 subunit nuclear translocation; we also did electophorsis mobility shift assay to verify the effect of GILZ on AP-1's binding activity to COX-2 promoter. From all these experiments, we discovered: 1.Cytokines IL-1β and TNF-α upregulate COX-2 expression in BM-MSCs in a time and dose-dependent manner; 2.Overexpression of GILZ can inhibit cytokines induced COX-2 expression both in mRNA and in protein level in BM-MSCs and mesenchymal stem cell line MC3T3-E1; 3.GILZ knock down reduce GILZ expression in BM-MSCs under stimulation of Dexamethasone, in the meantime, weaken the inhibition effect of Dexamethasone on cytokines induced COX-2 expression; 4.GILZ inhibit Cox-2 gene transcription by bloehing the transactivity of NF-κB and AP-1 in a dose dependent manner: 5.GILZ interact with NF-κB p65 and block its nuclear translocation; 6.GILZ inhibit cytokines induced AP-1 c-Fos binding activity to Cox-2 promoter. Taken together, we came to our conclusion that in BM-MSCs, GILZ mediates GCs' anti-inflammatory effect by inhibiting cytokines induced COX-2 expression via blocking the transactivities of NF-κB and AP-1. While we are preparing our manuscript, another group published there data showing that GILZ mediates GCs' anti-inflammatory effect by inhibiting the activation of NF-κB signaling pathway in response to cytokines in pulmonary epithelia cells. So, GILZ may be a potent mediator of GCs' anti-inflammatory effect. Characterized by their multipotential and immunosuppressive effect, BM-MSCs represent promising candidates for cell therapy of joint diseases. However, transplantation of BM-MSCs failed to confer any benefit in Collagen induced arthritis model of RA. Further research indicated local high concentration of cytokines, especially TNF-α, may influence the immunosuppressive effect of MSCs. In our experiments, overexpress GILZ can effectively inhibit cytokines induced NF-κB and AP-1 signaling pathway, so we believe modification of MSCs, such as overexpression of GILZ, still may be a novel therapy for RA. Most importantly, GILZ can regulate adipogenic differentiation by inhibiting PPARγ expression, and recently, we also discovered that GILZ may also enhance osteoblastic differentiation through modulation of PPARγ, which indicates GILZ can mediated GCs' anti-inflammatory effect without affect bone formation, so it may avoid the side effect of bone lose due to long-term GCs administration, and work as a novel anti-inflammatory therapy. Key words: GILZ, cyclooxygenase-2, arthritis, mesenchymal stem cell