【作者】 赵鹏；

【导师】 李新钢； 曲迅； Rolf Bjerkvig；

【作者基本信息】 山东大学 ， 外科学， 2008， 博士

【摘要】 研究背景恶性胶质瘤是最常见的成人脑肿瘤，约占所有成人脑肿瘤的40％。高分级的胶质瘤患者其五年生存率低于3％，中期生存期小于一年。目前通常采用的治疗手段是手术及术后辅以放疗化疗等综治疗方法为主，但其生存期并未得到明显的改善。近来，随着肿瘤免疫及肿瘤免疫逃逸机制研究的进展，继肿瘤手术、放疗及化疗之后诞生的肿瘤第四治疗模式——肿瘤生物治疗已显示了良好的前景，基于抗肿瘤免疫效应细胞及细胞因子过继疗法的出现及发展，特别是临床前期及临床研究的结果，为提高脑肿瘤患者的治疗效果延长其生存期及生存质量提供了新的希望。在众多抗肿瘤免疫治疗方案中，基于树突状细胞（dendritic cells，DCs）的抗肿瘤疫苗的特异性细胞免疫疗法尤为令人关注。DCs作为专职性抗原呈递细胞在启动机体特异性抗肿瘤免疫应答中起着关键性作用，随着DCs体外制备及培养技术的发展使基于DCs的抗肿瘤疫苗的基础及临床研究取得了较大的突破，在许多临床前期实验和临床试验中，基于DCs的免疫治疗在启动机体抗肿瘤特异性细胞免疫应答中取得了令人鼓舞的结果。目前基于DCs的免疫治疗策略主要有两种：其一是将体外诱导分化的患者DCs直接注射到肿瘤部位或手术切除区；其二是将患者的DCs在体外负载肿瘤抗原，然后再回输入患者体内。研究已证实脑瘤体内注射DCs，能够使其在体内原位获得肿瘤抗原并迁移至局部淋巴结，启动特异性抗肿瘤免疫应答。越来越多的研究表明瘤内注射DCs是一种更有希望的肿瘤治疗策略。在体内肿瘤局部应用DCs的免疫疗法较体外用肿瘤抗原负载的DCs疫苗回输体内的方法主要优点在于在体内DCs捕获抗原更符合DC呈递肿瘤抗原的自然环境。DC在体内环境中处理呈递肿瘤抗原所涉及的可能的共刺激因子是体外负载抗原所不能具备的，体外负载肿瘤抗原的DCs回输入体内所引起的抗肿瘤免疫反应效价低。另外，体外负载的DCs疫苗需要手术后在体外处理数周，而局部应用DCs可以在手术切除同时将DCs注入瘤腔，从而能够在手术切除肿瘤后尽快地调动机体免疫系统清除残留的肿瘤细胞。虽然有令人鼓舞的发现，但是目前此种方法在治疗肿瘤的临床试验中并没有得出切实有效的结论，主要问题是肿瘤微环境下DCs功能缺失。近来有研究表明，在瘤灶内注射的体外分离的DCs仅有极少数迁移到局部淋巴结。这表明，在肿瘤环境中，DCs的迁移功能受到抑制。DCs由抗原呈递区迁移到局部淋巴结是其发挥呈递抗原功能的关键因素，DCs的迁移功能对周围环境的变化极其敏感。肿瘤上清中大量的抑制性因子和肿瘤周围的乏氧微环境是肿瘤微环境的主要特征。然而目前仍不清楚肿瘤微环境中那些因素对DCs的迁移起关键性的抑制作用。因此，研究肿瘤微环境中何种因素对DCs迁移产生影响，以及影响其迁移的机制在DCs肿瘤免疫治疗中是非常有意义的。在实验中，我们在体外比较了在肿瘤上清和／或乏氧环境中人单核细胞来源的DCs迁移的变化，并评价了这两种因素对DCs迁移功能的影响。进而我们研究了肿瘤微环境是通过何种机制影响DCs迁移的，并发现了一条肿瘤微环境中抑制DCs迁移的信号转导通路。在分子机制上对肿瘤逃避宿主免疫监视提出了新的观点，为基于DCs的肿瘤免疫治疗提供了改进方法。目的1、体外分别在乏氧和常氧环境中由单个核细胞诱导分化出DCs，并检测其形态和表型；2、研究肿瘤微环境中肿瘤上清和／或乏氧对DCs迁移的影响；3、研究肿瘤上清和／或乏氧影响DCs迁移功能的机制。方法1、体外单核细胞源性DCs的培养Ficoll-Paque密度梯度离心法分离健康供者外周血中的单核细胞。含2％胎牛血清RPMI 1640培养基中加入PBMCs，调整细胞浓度为1×10⁷／ml，37℃孵育2小时，去除未贴壁细胞。贴壁细胞加入含GM-CSF和IL-4的RPMI 1640培养液中，在乏氧和常氧的条件下培养7天。第三天更换含GM-CSF和IL-4的培养液。第五天加TNF-α（10 ng/ml）促进细胞成熟。七天后收集培养的细胞，洗涤两次，于实验前重新混悬细胞置于无血清的RPMI 1640培养基内。2、肿瘤上清的收集收集处于对数生长期的人恶性胶质瘤U251细胞，计数，以1×10⁵／L的浓度悬浮于完全培养基，培养24h后离心收集上清，过滤冻存于-80℃备用。将DCs细胞置于含有50％肿瘤细胞上清的DC培养基中培养，设不含肿瘤细胞培养上清的DC培养基对照。3、常氧和乏氧的细胞培养条件常氧环境，细胞在常规细胞培养箱内孵育（21％O₂，5％CO₂ and 74％N₂）。乏氧环境，细胞于专用密封的乏氧培养箱中培养（1％O₂，5％CO₂ and 94％N₂）。温度（37℃）和湿度（90％）均保持在理想的状态下。4、乏氧和／或肿瘤上清对DCs迁移影响的检测①应用膜上覆有Matrigel胶的迁移小室来研究DCs对胶质瘤上清和／或乏氧环境在体外研究胶质瘤微环境中这两种因素对DCs迁移功能的影响。体外诱导生成的DCs与两种因素（乏氧和／或胶质瘤上清）作用12小时后，通过膜上DCs细胞数观察各因素对DCs迁移功能的影响。我们将实验分为4组：正常氧条件下DCs培养组（DCs-N），常氧+肿瘤上清组（DCs-T），乏氧组（DCs-H），乏氧+肿瘤上清组（DCs-H-T）。②在基因水平上通过琼脂糖电泳和Real-time PCR检测不同实验组CCR7基因转录的水平。蛋白水平应用流式细胞仪检测细胞膜上的CCR7蛋白的表达水平。③应用Real-time PCR和Western blot技术分别在基因及蛋白水平检测不同实验组MMP-9的表达变化。5、乏氧抑制人DCs产生MMP-9的机制研究①应用乏氧功能基因组芯片分析DCs在常氧和乏氧环境中相关基因转录变化。②应用real-time PCR技术验证常氧和乏氧中DCs腺苷受体基因表达的谱变化。③通过药理学方法应用腺苷受体激动剂和阻断剂研究腺苷受体的功能。④应用特异性腺苷酸环化酶激活剂forskolin和抑制剂SQ22536以及特异性的PKA抑制剂H89研究cAMP／PKA信号转导通路。结果1、乏氧和／或肿瘤上清对DCs迁移的影响①覆有Matrigel胶的迁移小室显示DC-H组的迁移活性（94±16．1 cells／HPF）仅是DC-N组（185±23．6 cells／HPF，P<0．05）的一半。而DC-T组（163±18．1 cells／HPF）和DC-H-T（92±15 cells／HPF）的迁移活性与DC-N组未见明显区别。DC-H组的细胞在DC-N的上清中培养24小时后能够部分恢复其迁移功能。②乏氧能够诱导MMP-9 mRNA的表达上调（P<0．01），而肿瘤上清对MMP-9 mRNA的表达没有作用。在Western blot中DC-N和DC-T所对应的印记明显强于DC-H和DC-H-T。③乏氧在基因和蛋白水平上均可增加DCs中CCR7的表达。肿瘤上清对CCR7的表达未见有明显影响。2、乏氧诱导DCs表达腺苷受体A_2b乏氧功能基因组芯片显示，乏氧下DCs表面腺苷A₁，A_2a，A₃受体的表达水平较常氧没有变化（P>0．05），而mDCs表面腺苷A_2b受体的表达是常氧中的16±0．81倍（P<0．05），且A_2b受体成为mDCs腺苷受体的优势表达受体。imDCs的表达水平没有变化。Real-time PCR检测DCs腺苷表达谱也验证了基因芯片的结果。这些结果表明乏氧能够诱导mDCs表面的A_2b受体上调。3、乏氧抑制DCs的MMP-9产生由腺苷A_2b受体介导为了评估腺苷受体的功能，我们用腺苷受体的激动剂和拮抗剂对其进行干预。在常氧或乏氧状态下mDC-N给予腺苷受体激动剂（NECA，1μM）或A_2b受体特异性拮抗剂（MRS1754，1μM）作用8小时，或不作处理。收集细胞培养上清液酶谱检测显示：mDC-N组的MMP9可见一清楚的条带，而mDC-H的MMP-9则呈现一条弱的条带。MRS1754（1μM）可以消减缺氧所造成的mDC-H产生MMP9减少的效应。然而在常氧状态下，NECA对mDC-N没有促分泌MMP9的效应。酶谱和ELLSA实验均验证这一结果。4、乏氧抑制DCs分泌MMP-9的效应与cAMP／PKA相关A_2b受体是G蛋白耦联受体，它可以通过活化腺苷酸环化酶（AC）激活许多信号传导通路。继发信号转导的研究主要集中在cAMP。首先，AC激活剂forskolin可以在DC-N组中显著降低MMP9的分泌量；其次，在乏氧状态下DCs中cAMP的含量增加，揭示forskolin可以模拟缺氧状态下对MMP-9生成的效应；最后，我们研究了AC抑制剂SQ22536对在乏氧下DCs产生的MMP9效应的影响。乏氧状态下MMP-9被抑制了51±2％，在有SQ22536（1 mM）的存在时，乏氧状态下的MMP9仅被抑制了20±3％（P<0．05，与仅在乏氧状态下相比，n=4）。这一研究结果显示细胞内cAMP增加的水平可以调控乏氧状态下MMP9的分泌量。cAMP可以使蛋白激酶A（PKA）磷酸化，PKA在乏氧时抑制DCs分泌MMP-9的效应可以被PKA特异性抑制剂H89所阻断。在乏氧状态下的DCs，加入H89后可以抵消乏氧对MMP-9的分泌抑制。因此，cAMP/PKA传导通路可以通过作用于A_2b受体消减乏氧对MMP-9分泌的抑制效应。5、乏氧时对MMP-9的抑制效应取决于A_2b受体活性的增高而非腺苷酸环化酶常氧下刺激A_2b受体不能复制乏氧对MMP-9的抑制，然而在乏氧状态下抑制A_2b受却能消减MMP-9的分泌量。为了解释这一现象，我们的研究主要集中于分子cAMP信号间的转导。为了排除乏氧下DCs本身产生的腺苷对cAMP的影响，在常氧中对DC-N和DC-H进行干预。与cAMP显著增加的DC-H相比加入了NECA的DC-N其cAMP的分泌量仅略微增加。仅在DC-H中A_2b选择性阻断剂MRS1754可以消除NECA的刺激效应，表明A_2b的活性在DC-H中高，在DC-N中低。Forskolin的应用表明AC激活可以增加DC-N和DC-H中cAMP的水平，这说明AC的活性不受乏氧影响。A_2b受体阻断剂MRS1754不能抑制forskolin的刺激效应，然而特异性的AC抑制剂SQ22536能够抑制NECA的刺激效应，这表明在乏氧下出现的A_2b介导的MMP-9的抑制效应是经由cAMP实现的。结论1、体外常氧与乏氧状态下均能产生DCs2、乏氧通过抑制MMP-9的产生抑制DCs的迁移，而肿瘤上清对DCs的迁移功能没有抑制效应3、乏氧对于DCs产生MMP-9的抑制效应需要激活腺苷A_2b受体并经由cAMP／PKA途径更多还原

【Abstract】 BackgroundMalignant glioma represent about one-third of all adult primary brain tumors. The prognosis of human malignant glioma remains poor with an overall 5-year survival rate of less than 3% and a median survival time of about 1 year for higher grade tumors such as glioblastoma. Alternative therapeutic approaches to surgery, radiotherapy and chemotherapy are required to improve patient survival. Recent insights into neuroimmunology provide evidence of the immunocompetence of the brain even in the case of tumor development, thus opening new hopes for immunotherapy.Dendritic cells （DCs） play a crucial role during the priming of antitumor-specific immune response. The DCs immunotherapy for inducing specific immunity against tumor has shown promising activity in both preclinical models and clinical observations. Two main DC-based immunotherapies are intensively used for tumor bearing patients. One is injecting DCs directly into tumor or surgical resection region of tumor; the other is Loading DCs with tumor antigens ex vivo and transfusing back to body. Intratumorally injected DCs can acquire and process tumor antigens in situ, migrate via lymphoid vessels to regional lymphoid organs, and initiate significant tumor-specific immune responses, even in the CNS. A cumulating data indicate that intratumoral injection of DCs may be a more promising strategy for the therapy of tumor patients. A major advantage of stimulating DCs with tumor antigen in vivo instead of ex vivo is that DCs encounter the tumor antigen in their natural environment. Other co-factors involved in antigen processing and presentation are already available, simplifying the manipulation required to maximize anti-tumor immune responses. In addition, this therapy can be administered to patients at the same time as tumor resection by surgery, while ex vivo impulse antigen of DCs require a week or more before the patient can receive the immunotherapy. This means that the tumor is rapidly hit with the anti-tumor immune response, is not given any chance to recover from surgical resection and the therapy can take advantage of any inflammation naturally associated with the tumor resection.However, clinical trials have produced contrasted results and no definitive conclusion can be drawn about the real efficacy of this approach. Basic questions, concerning functional deficiency of DCs in tumor microenvironment, are currently unresolved. A recent study asserted that the migration of in vitro generated DCs to the regional lymph nodes was very few observed when injected into tumor tissues. This suggested that DCs lose their migrating activity in tumor tissues. Migration of DCs from the site of antigen recognition to regional lymphoid tissues is crucial for DCs to function as potent antigen presenting cells （APCs）. DCs migration is known to be highly sensitive to microenvironmental changes. A number of immunosuppressive factors in tumor supernatant （TSN） and regional low oxygen （Hypoxia） are particular phenotypes in tumor microenvironment. But to date, it remained unclear which factor in tumor microenvironment play a crucial role in hindering the migration of DCs. So it is very interesting to find which factor in tumor microenvironment can impair the migration of DCs and how to effect on it.In our study, we compared the migrating activities of human monocyte-derived DCs under hypoxic in the present or absent of TSN to their normoxia counterparts in vitro and assessed the effect of these two factors on the migration of human monocyte-derived DCs. Whereafter, we further investigate the mechanisms leading to the suppression of DCs migration under tumor microenvironment, and find a signaling pathway to inhibit DCs migration. So as to provide a new insight into the understanding of molecular mechanisms on how tumor cells escape host immune surveillance in tumor local-tissue microenvironments.Objective1. In vitro generation of monocyte-derived DCs from peripheral blood monocytes （PBMCs） under normoxic and hypoxic conditions and identification of their morphology and phenotype.2. To study the effect of TSN and hypoxia in tumor microenvironment on the migration of DCs.3. To study how to affect the migration of DCs by TSN and/or hypoxia.Methods1. In vitro generation and culture of monocyte-derived DCsPBMCs were isolated by density gradient centrifugation with Ficoll-Paque Plus of buffy coats. PBMC were resuspended in 2% FCS-RPMI-1640 medium at 10⁶/mL and distributed in 6-well plates （2 mL/well）. The plates were incubated in a 37℃incubator for 2h before nonadherent cells were removed. The remaining adherent cells were cultured in the presence of GM-CSF （1000 U/mL） and IL-4 （500 U/mL） in 10% FCS-RPMI-1640 medium. Fresh medium containing GM-CSF and IL-4 was added on Day 3 （2 mL/well）. To induce maturation, LPS （1μg/mL） was added to cells on day 5 of culture for 48h.2. Preparation of TSN and TSN-exposed DCsTSN were prepared by seeding 25cm² flasks with 1×10⁶ U251 glioma cells in 10 ml of complete medium. The culture supernatants were collected after 24 h under hypoxia, centrifuged to remove cells, and stored at -80℃. The effects of TSN on DCs were examined by replacing 50% （vol/vol） of the culture medium with TSN.3. Normoxic and hypoxic culture conditionsFor normoxic conditions, cells were incubated in a regular incubator （21% O₂, 5% CO₂, and 74% N2）. Hypoxic incubation was performed in a sealed, anaerobic work station, where the hypoxic environment （1% O₂, 94% N₂ and 5% CO₂）, the temperature （37℃）, and humidity （90%） were kept constant. 4. To investigate the effect of hypoxia and/or TSN on the DCs migration.①DCs cultured under normoxic or hypoxic conditions were treated with or without50% TSN in medium for 24h and were evaluated for their migratory activity througha Matrigel-coated filter in a Transwell system.②In gene level, using Real-time quantitative RT-PCR to detect the changes of CCR7gene expression in every one of experimental groups, In protein level, using Flowcytometric analysis to detect CCR7 on membrane of DCs.③Using Real-time quantitative RT-PCR and Western blot to detect the MMP-9 ofDCs separate from gene and protein level.5. To investigate the mechanisms of hypoxia suppressing the produce of MMP-9 by human DCs.①To profile the relative expression of hypoxia signal pathway proteins related genes in normoxic and hypoxic DCs by gene microarray analysis.②To verify the expression pattern of adenosine receptor subtypes on DCs in response to hypoxia by real-time PCR.③To evaluate the function of adenosine receptors by pharmacological approach using adenosine receptors agonists and antagonists.④To study the cAMP/PKA signaling pathway by forskolin （a direct activator ofadenylate cyclase）, SQ22536 （a specific adenylate cyclase inhibitor） and the specificPKA inhibitor H89.Result1. Effect of hypoxia and/or TSN on the DCs migration.（1） After 8 h of incubation in the presence of MIP-3β, it was evident that DC-H（94±16.1 cells/HPF） displayed a lower migratory activity by twofold than that ofDC-N （185±23.6 cells/HPF, P < 0.05）. The migratory activity of DC-T （163±18.1cells/HPF） had no significant change when compared with DC-N. DC-H-T （92±15cells/HPF） was the similar with DC-H on migratory activity. Interestingly, themigratory capacity of DC-H was partially recovered when cultured with thesupernatant from DC-N. ②Stimulation with hypoxia induced a twofold increase in the steady-state levels of MMP-9 mRNA （P < 0.01）, but no significant differences were found in the induced levels stimulated with TSN. Both DC-N and DC-T exhibited a distinct band corresponding to MMP-9 in Western blot. In contrast, DC-H and DC-H-T presented a faint band corresponding to MMP-9.③Stimulation both with hypoxia and TSN had no significant differences in the levels of CCR7 mRNA were found compared with DC-N. CCR7 protein was slightly increased by hypoxia.2. Hypoxia induces adenosine receptor A_2b in expression pattern of adenosine receptor subtypes in DCsThe experiment data demonstrated that adenosine receptor A_2b was selectively induced by hypoxia （3-fold in imDCs and 21-fold in mDCs） and other subtypes were not changed. According to this clue, we verified the expression pattern of adenosine receptor subtypes on DCs in response to hypoxia by real-time PCR. The results consistently revealed that the A₁, A_2a and A₃ receptors expression levels in hypoxia remained unchanged compared with them in normoxia. In contrast, A_2b receptor on mDCs was increased by as much as 6±0.11-fold compared with it in normoxia （P <.01）, whereas on imDCs it was not changed. What’s more, the ratio of expression of A_2b receptor to others changed to predominance under hypoxic condition. Taken together, our results suggest that adenosine receptor A_2b on mDCs can be induced by hypoxia, which consistent with previous studies that hypoxia prominently induces A_2b receptor.3. Inhibition of MMP-9 produce by DCs in hypoxia is mediated by A_2b receptormDC-N was treated with or without adenosine receptor agonist （NECA, 1μM） under normoxia for 8 hours. mDC-H was treated with or without A_2b receptor specific antagonist （MRS 1754, 1μM） under hypoxia for 8 hours. Cultured supernatant Gelatin zymography showed that mDC-N exhibited a distinct band corresponding to MMP-9. In contrast, mDC-H presented a faint band corresponding to MMP-9, in agreement with our previous publication. MRS 1754 （1μM） was able to counteract the inhibition of hypoxia on MMP-9 in mDC-H. Under normoxic conditions, however, NECA had no effect on the secretion of MMP-9 in mDC-N incubated for 8 hours. The results obtained by zymography were confirmed by ELISA. To detect whether inhibition of MMP-9 is mediated by other receptors subtypes of DCs, we use each of specific receptor antagonists in different concentration （0, 10, 100, 1000μM） to incubate with mDC-H in hypoxia. Cultured supernatant ELISA showed that the A_2b specific antagonist MRS 1754 at the concentration of 100 nM was able to abrogate the inhibition of hypoxia on MMP-9. The A₁ specific antagonist DPCPX, A_2a specific antagonist SCH58261 and A₃ specific antagonist MRS 1220 were essentially without effect. To rule out the possibility that this suppression on hypoxia could be a result of an increase in MMP-9 secretion by MRS 1754 itself, The A_2b antagonist MRS 1754 （1000 nM） used in DCs under normoxia did not modify the MMP-9 secretion. Among the four adenosine receptors agonists tested (The A₁ specific agonist CPA, A_2a specific agonist CGS21680, A_2b agonist NECA and A₃ specific antagonist IB-MECA); none of them was able to mimic the inhibitory effect of hypoxia on MMP-9. Taken together, these results suggest the suppression of MMP-9 secreted by DCs in hypoxia is involved in activation of adenosine receptor A_2b.4. Inhibition of MMP-9 produce by DCs in hypoxia is cAMP/PKA relatedFirstly, the effect of forskolin, a direct activator of adenylate cyclase, was determined. Forskolin （10μM） significantly decreased MMP-9 secreted by DC-N. Secondly, direct measurements of cAMP levels in DCs showed that hypoxia can increase cAMP production. These results suggested that forskolin can mimicked the action of hypoxia on MMP-9. Finally, the effect of SQ22536, a specific adenylate cyclase inhibitor, on the action of hypoxia was studied. In control cultures, hypoxia decreased MMP-9 production （as measured by ELISA in 8-hour supernatants） by 51±2% （mean±SD）. In the presence of SQ22536 （1 mM）, hypoxia decreased MMP-9 production by only 20±3% （P<.01 compared with hypoxia alone, n=4;）. These findings suggest that the effects of hypoxia on MMP-9 are mediated through increasing of cAMP levels. When added to DCs under hypoxia, H89 abrogated the inhibition of MMP-9. Taken together, the cAMP/PKA pathway could be responsible for the observed effects of hypoxia on MMP-9 through binding to the A_2b receptor. 5. The inhibition of MMP-9 by hypoxia due to the elevation of A_2b receptor activity not adenylate cyclaseWe have showed that stimulation of A_2b receptor was not able to replicate the inhibitory effect of hypoxia on MMP-9 under normoxia, whereas the inhibition of A_2b receptor can counteract the MMP-9 under hypoxia. To explain the phenomena, our studies focused on subsequent signal transduction molecular cAMP. To rule out the effect of adenosine produce by DCs under hypoxia on cAMP, DC-N and DC-H were cultured with interventional agents under normoxia. Our results showed that NECA only mildly increased cAMP levels in DC-N compared with the significant increase in cAMP production in DC-H. The A_2b selective antagonist MRS 1754 eliminated the stimulatory effect of NECA only in DC-H, indicating that A_2b activity is high in DC-H and low in DC-N. Administration of forskolin demonstrated that adenylate cyclase activation augmented cAMP levels in both DC-N and DC-H, suggesting that adenylate cyclase activity was not affected by hypoxia. We noted that A_2b receptor antagonist MRS 1754 can not inhibit the stimulatory effect of forskolin, whereas specific adenylate cyclase inhibitor SQ22536 can inhibit the stimulatory effect of NECA. The observation demonstrated that the inhibition of MMP-9 mediated by A_2b receptor in hypoxia is via cAMP.Conclusion1. We can successfully generate DCs under hypoxic and normoxic conditions in vitro.2. Hypoxia can hinder the migratory of DCs through suppression of DCs-induced MMP-9, whereas TSN has no effect on their migration activity.3. The inhibitory effect of hypoxia on MMP-9 by DCs requires the activation of A2b in a cAMP/PKA-dependent signaling pathway.更多还原