【作者】 王闯；

【导师】 徐江平； 张汉霆；

【作者基本信息】 南方医科大学 ， 药理学， 2011， 博士

【摘要】 目的：阿尔茨海默病(Alzheimer’s disease, AD)是一种以β-淀粉样蛋白斑块和神经纤维缠结为重要标志物并伴随慢性神经炎症及认知功能障碍为特征的神经退行性疾病。通过抑制β-淀粉样蛋白诱导的神经毒性可能成为开发新型缓解神经损伤和减缓阿尔茨海默病进程药物的理论依据。磷酸二酯酶-4(Phosphodiesterse-4, PDE4)作为重要第二信使环磷酸腺苷(Cyclic adenosine monophosphate, cAMP)特异性水解酶可能对脑内多种信号转导通路发挥作用。研究发现,通过抑制PDE4从而激活cAMP介导的cAMP/PKA/CREB信号转导通路,可以改善多种因素引起的认知障碍及炎性疾病。因此,本研究的目的首先通过使用β-淀粉样蛋白(Amyloid beta peptide, Aβ)的不同片段：Aβ1-42或Aβ25-35诱导的AD动物模型,明确Ap的病理作用中是否诱发神经炎症及其pCREB信号通路的下调；并使用PDE4的特异性抑制剂咯利普兰进行干预,明确PDE4抑制剂在改善动物认知功能障碍的同时是否伴随动物脑内神经炎症的改善以明确PDE4对学习记忆功能的改善是否与改善神经炎症有关；并且为进一步阐明这一问题,我们通过使用TNF-α抗体动物脑内注射,单纯通过阻断Aβ诱发的神经炎能看是否能够改善其所诱发的认知功能障碍,以期阐明神经炎症与认知功能障碍间的关系；同时我们使用脑内给予PKA的阻断剂H89以阻断PDE4抑制剂对脑内PKA/CREB信号转导通路的激活效应,看是否通过阻断pCREB介导的信号转导通路能否阻断PDE4抑制剂对认知功能与神经炎症的改善作用,明确PDE4抑制剂引起的cAMP/PKA/CREB信号转导通路与认知功能及神经炎症间的关系。其次,为进一步阐明PDE4亚型在学习记忆中的调节作用,我们首先使用PDE4A基因敲除动物对PDE4A在学习记忆功能及其该基因对神经发生的影响进行了首次研究,并进一步对PDE4A、B及D亚型基因敲除后在Ap诱导的认知功能障碍及神经炎症的调节作用,进一步阐明每种亚型在中枢神经系统的调节作用。因此本研究进一步丰富了PDE4及其亚型在AD中的调节作用,为开发高选择性及高效亚型抑制剂提供了理论依据,并为AD治疗中的抗炎理论提供了实验依据及新的观点。方法：(1)小鼠海马CA1或DG区分别注射老化状态的Aβ1-42 (0.4、1.6μg/side),10天后分别依次进行自主活动测试记录其穿线次数及直立次数。24小时后在同一装置内进行5分钟的新物体识别实验的学习阶段,记录动物对两对称放置的相同物体的识别时间,24小时后进行新物体的识别测试,并分别记录动物对新,旧物体在5分钟内的识别时间,并计算动物对新物体的识别指数。物体识别实验后24小时,进行跳台实验的学习阶段,24小时后对动物的潜伏期进行测试。全部行为学结束后24小时,处死动物并取海马组织,采用Western bloting的方法对pCREB、CREB、BDNF、NF-κBp65、IL-1β进行检测。并且,为验证本实验手术定位的准确性,在整个实验进行前对药物能否准确微量注射进海马CA1或DG区进行了验证,我们采用GFP-标记的慢病毒,按照不同的坐标,打入CA1或DG区,14天后,切片观察病毒能否在CA1或DG区准确表达。(2)首先,我们以空白溶剂和逆序列的Aβ1-42为对照,对ICR小鼠海马CA1区分别注射老化状态的Aβ1-42 (0.1、0.4、1.6pμg/side)。选择中间剂量行PDE4抑制剂咯利普兰进行干预。10天后动物分别进行自主活动测试、物体识别实验、跳台实验及Morris水迷宫测试。全部行为学实验结束后24小时,处死动物并取海马及前额皮层,采用免疫印迹的方法对海马及皮层内的pCREB、CREB、IL-1β、TNF-α、NF-κB p65、COX-2、IL-10、SOD以及海马内Aβ1-42的含量进行测试。并对海马内PDE4A、B、D亚型和亚亚型的表达进行了检测。其次,为验证咯利普兰的量效关系,采用老化状态的Aβ25-35(10μg/side)诱导SD大鼠学习记忆障碍模型,分别采用0.1,0.25,0.5mg/kg的咯利普兰进行干预。12天后进行Morris水迷宫测试及避暗实验测试。行为学实验结束24小时后动物处死并取海马组织,免疫印迹方法测试海马内pCREB、NF-κB p65、Bcl-2及Bax的表达。最后,为验证Ap诱导的动物学习记忆障碍是否与其诱导的神经炎症相关及pCREB在炎症调节中的作用。我们使用老化状态的Aβ1-42分别微量注射进ICR小鼠海马CA1区(0.4μg/side)诱导AD动物模型,验证TNF-α抗体海马CA1区给药能否阻断动物的神经炎症,并采用PKA的阻断剂H89,通过对PKA的阻断验证能否阻断PDE4抑制剂咯利普兰的抗炎及认知增强作用。本实验所有动物手术后10天均进行自主活动测试、物体识别实验、跳台实验及Morris水迷宫测试。行为学测试结束24小时后,采用免疫印迹方法对动物海马内pCREB、CREB、NF-κB p65及TNF-α进行测试。(3)首先采用PDE4A-/-小鼠探讨PDE4A亚型敲除后与野生型相比在学习记忆能力方面的变化及其机制。分别采用自主活动实验、物体识别实验及跳台实验比较两组动物的学习能力。采用免疫印迹方法对动物海马、皮层及纹状体内的pCREB、CREB及BDNF进行检测。并采用免疫组化的方法对新生神经元的增殖与存活能力进行比较。其次,我们采用PDE4D-/-动物,海马CA1区给予0.41μg/侧的Aβ1-42,对PDE4D在Aβ1-42诱导的学习记忆障碍与神经炎症的调节作用进行研究。手术10天后,采用自主活动实验、新物体识别实验、跳台实验及Morris水迷宫实验对动物的学习记忆能力进行测试。采用免疫印迹方法对海马内pCREB、CREB、NF-κB p65、TNF-α、IL-1β、COX-2、IL-10进行测试。最后,我们采用PDE4A、B、D基因敲除动物,海马DG区分别给予老化状态的Aβ1-42 (0.4μg/side),对三种亚型在Aβ1-42诱导的学习记忆障碍与神经炎症中的调节作用进行研究。手术10天后,采用自主活动实验、物体识别实验、跳台实验对动物的学习记忆能力进行检测,采用免疫印迹方法对海马内pCREB、BDNF、NF-κB p65及COX-2进行检测。结果：(1)预实验研究发现,根据文献报道的定位坐标,均能准确定位海马CA1与DG区。研究发现海马CA1或DG区分别给予“老化”状态的Aβ1-42后,动物在自主活动实验中的穿线次数与直立次数均无显著变化。新物体识别实验研究发现,海马CA1或DG区两组剂量的Aβ1-42,均显著诱导了动物的认知指数的显著下降。跳台实验发现,CA1区给予两组剂量Aβ1-42均显著降低了其逃避潜伏期,而DG区仅仅1.6μg/侧的剂量显著降低了动物的逃避潜伏期。与行为学结果相一致,Aβ1-42两组剂量显著降低了CA1区的pCREB及BDNF的表达。而只有1.6μg/侧的剂量显著降低了DG区pCREB的表达,0.4,1.6μg/side均显著降低了DG区的BDNF表达。所有处理对动物的CREB的表达均没有显著影响。两剂量Aβ1-42均显著诱导了海马CA1及DG区的炎症反应,NF-κB p65及IL-1β均显著增高。(2)首先研究发现,Aβ1-42(0.1,0.4和1.6μg/侧)剂量依赖性诱导动物在新物体识别实验、跳台实验及Morris水迷宫实验中认知功能的障碍,而咯利普兰(0.5mg/kg)诱导了上述作用,而各种处理效应对动物的自主活动能力无影响。并且对炎症因子的检测发现,Aβ1-42诱导了动物海马内神经炎症及氧化应激的出现,剂量依赖性地增加海马内IL-1β、TNF-α、NF-κB p65、COX-2水平,并且咯利普兰显著性降低了上述炎症相关因子的释放；对pCREB及CREB的检测发现,Aβ1-42剂量依赖性抑制了CREB的磷酸化,而咯利普兰逆转了上述效应,Aβ1-42及咯利普兰对CREB均无显著性影响；并且研究发现炎症反应轻微诱导了IL-10的生成,而咯利普兰也对IL-10的升高起部分调节作用；研究发现咯利普兰对Aβ1-42本身无降低作用,而Aβ1-42可以显著上调PDE4A5、PDE4A?、PDE4B1、PDE4B2、PDE4B3、PDE4B4、PDE4D4、PDE4D5水平。大鼠的结果同样发现,Ap的核心毒性片段Aβ25-35海马CA1区微量注射(10μg/侧)诱导了动物在Morris水迷宫中空间学习记忆的障碍及避暗实验中长时程学习记忆的障碍。并且Aβ25-35诱导海马内NF-κB p65和Bax的显著性增高及pCREB和Bcl-2的显著性降低。上述指标均被咯利普兰(0.1、0.25和0.5mg/kg)剂量依赖性地逆转。其次,研究发现TNF-α抗体尽管其阻断了炎症反应,使NF-κB p65及TNF-α显著性降低,但仅轻度缓解了动物的学习记忆能力,而TNF-α轻微加重了Ap诱导的学习记忆障碍,轻度削弱了咯利普兰的认知增强作用；而PKA阻断剂H89不仅仅阻滞咯利普兰的认知增强作用而且阻断了咯利普兰对炎症的缓解作用。(3)研究首先发现PDE4A基因敲除动物与野生型动物相比学习记忆能力显著增强,且PDE4A基因敲除对动物的自主活动能力无显著性影响。并且该基因敲除后不仅显著上调海马内pCREB/BDNF信号转导通路,而对皮层及纹状体内该信号转导通路无显著性影响,而且使海马内神经增值与存活能力显著增强；其次研究发现PDE4D基因敲除动物能够逆转海马CA1区微量注射Aβ1-42所诱导的学习记忆障碍,并且显著性逆转了Aβ1-42对CREB磷酸化功能的抑制。尽管对NF-κB p65、TNF-α、IL-1β无显著性影响,PDE4D基因敲除却显著性抑制了Aβ1-42对COX-2的诱导作用,并且轻度增加了IL-10水平。进一步研究发现,而PDE4A与D对虽未能改善海马DG区微量注射Aβ1-42所诱导的诸如NF-κBp65的增加,其却显著性降低了Ap1-42诱导的COX-2的水平,并且显著上调pCREB/BDNF信号转导通路,起到了认知功能增强的作用；而PDE4B基因敲除虽未能显著改善Aβ1-42诱导的认知功能障碍及pCREB/BDNF信号转导通路的下调,但其显著性降低了Aβ1-42诱导的NF-κB p65的过表达,并且奇怪的是PDE4B基因敲除却未能显著降低Aβ1-42所诱导的COX-2的过表达。结论：本研究发现,Aβ1-42微量注射入海马CA1和DG区均能够诱导动物学习记忆功能障碍,并且伴随神经炎症的出现与pCREB/BDNF信号转导通路的下调,为寻找具有抗炎作用并伴有pCREB介导的信号转导通路增强的新的抗老年痴呆药物靶点提供了理论依据；并且研究发现PDE4抑制剂咯利普兰不仅可以逆转Aβ1-42及Aβ25-35分别诱导的动物认知功能障碍,而且可以改善Aβ注射脑区的炎症反应与凋亡反应,并上调pCREB介导的信号转导通路,为选择以PDE4为药物作用靶点奠定了理论基础,不仅对PDE4抑制剂在Aβ诱导的学习记忆障碍中的作用及机制进行了阐明,而且对Ap具有诱导PDE4不同亚亚型的高表达进行了研究,丰富了Ap对pCREB信号转导通路下调的机制并为进行PDE4A、B及D亚型在Ap诱导的学习记忆障碍中的调节作用奠定了基础。神经炎症在AD学习记忆障碍中发挥一定作用,而PDE4抑制剂通过pCREB信号转导通路在学习记忆及神经炎症中发挥重要作用。我们的研究首次明确PDE4A基因敲除具有上调pCREB信号转导通路并具有增强海马神经发生的功能,明确了PDE4A亚型具有调节认知功能的作用；进一步证明了PDE4A及D基因敲除逆转Ap诱导的学习记忆障碍的作用,并且阐明PDE4A及D对Ap诱导的神经炎症无显著改善作用,而PDE4B虽然未能显著上调pCREB信号转导通路,却显著改善了海马内神经炎症,这可能是由于PDE4A及D基因敲除后下调COX-2而PDE4B敲除后未能下调COX-2所致。说明不同部位COX-2下调可能对炎症改善并不起积极作用。更多还原

【Abstract】 Objective:Alzheimer’s disease (AD) is a kind neurodegenerative diseases characterized by the P-amyloid plaques and neurofibrillary tangles as the important marker and associated with chronic neuroinflammation and cognitive dysfunction. By inhibiting neurotoxicity induced by P-amyloid may be theoretical basis to develop the new drugs which to slow down the nerve damage and the process of Alzheimer’s disease. Phosphodiesterase-4 (PDE4), an enzyme that specifically catalyzes the hydrolysis of cyclic AMP (cAMP), plays a crucial role in a variety of signal transduction pathway in brain function. More and more study has revealed that the inhibition of PDE4 mediated cAMP/PKA/CREB signal transduction pathway upregulation reverses the cognitive impairment caused by many factors and inflammatory diseases. Therefore, the main purpose of this study was to clarify the pathological role of Aβand also demonstrated whether different fragments:Aβ1-42 or Aβ25-35 induced the memory deficits with role of neuroinflammation and the reduction of pCREB signaling. We also used the PDE4 specific inhibitor rolipram to demonstrated whether the reversed effect on memory deficits induced by Aβassociated with anti-inflammatory response. In order to further clarify whether neuroinflammation plays important role on memory deficits, we used TNF-a antibody to block the neuroinflammation induced by Aβand to test if the cognitive dysfunction associated with the reduction of the neuroinflammation. In order to clarify the relationship of the memory deficits between the neuroinflammation and the inhibition of pCREB signaling induced by Aβ, we used the PKA blocker (H89) to block the activation effect of rolipram on the PKA/CREB signaling to see the blocking pCREB-mediated signaling pathway can block the PDE4 inhibitor on cognitive function and anti-inflammation effects. Secondly, to further clarify the regulation of the PDE4 subtypes on the learning and memory, we used the PDE4A knockout mice to demonstrated the role of PDE4A on the memory and the neurogenesis. And we also used the PDE4A、B and D subtype gene knockout mice to further clarify the different regulation on the cognitive dysfunction and neuroinflammation induced by Aβ. Therefore, this study further enriches the regulation of PDE4 and its subtypes in the AD and also provides a theoretical basis for developing the highly selective and efficient PDE4 or subtypes inhibitors with the role of anti-inflammaion effect.Methods:(1) CA1 or DG subregions of hippocampus were injected with "aging" Aβ1-42 (0.4,1.6μg/side),10 days later the locomotor activity were tested in order to record the line crossing and rears.24 hours later, mice were individually placed in the center of the box containing two identical objects located in the two diagonal corners. The cumulative time spent in exploring each object was recorded during a 5-min period. 24 hours after training, mice were tested for memory using same procedure except that one of the familiar objects was replaced with a novel object. The time of exploration of each object (Tf and Tn for familiar and novel objects, respectively) was recorded for determination of the recognition index (RI=Tn/Tn+Tf).24 h after the novel object recognition test, the training of the step-down passive avoidance were conducted.24 h after the training using the same procedure except the no shocks were delivered, the step-down latency was recorded, with an upper cut-off time of 300 s.24 h after all the behavioral test, the animals were sacrificed and the pCREB> CREB、BDNF、NF-κB p65、IL-1βin the hippocampus were detected. Moreover, to verify the accuracy of the injected position in the CA1 and DG subregions, we used the GFP-labeled lentiviruses to inject into the CA1 and DG subregions according to the different coordinates,14 days later the virus expression in the CA1 and DG subregions was observed by immunohistochemistry.(2) First, we used the reverse sequence of Aβ1-42 and blank solvent as the control, the aged Aβ1-42 (0.1,0.4,1.6μg/side) were injected into the CA1 subregion of hippocampus. The Rolipram (0.5mg/kg) or Vehicle were administrated 24 h after the surhery.10 days later, the behavioral test of the locomotor activity, the novel object recognition test, the step down passive avoidance and the Morris water maze test were conducted in order.24 h after all the behavioral test, the animals were sacrificed and the pCREB、CREB、IL-1β、TNF-α、NF-κB p65、COX-2、IL-10、SOD and the Aβ1-42 levels in the hippocampus and cortex were detected by western bloting. And PDE4A, B, D subtype expression in the hippocampus also were examined in order to clarify the different effect of Aβ1-42 on the different PDE4 subtypes. Secondly, in order to verify the dose-effect of the rolipram (0.1、0.25、0.5 mg/kg) on the memory deficits induced by Aβ25-35 (10μg/side) in the SD rats, the behavioral test of the Morris water maze test, the passive avoidance test were conducted in order from the day 12.24 h after the last behavioral test, the animals were sacrificed and the pCREB、NF-κB p65、Bcl-2 and the Bax in the hippocampus were detected by western bloting. Finally, in order to verify whether the memory deficits induced by Aβis associated with neuroinflammation and also to clarify the important regulation of pCREB signaling on the memory and neuroinflammation, we microinjected the aged Aβ1-42 into the CA1 subregion of the hippocampus (0.4μg/side) and to tested whether the TNF-αantibody could block the neuroinflammation and memory deficits. And we also used the PKA blocker of the H89 to test whether this blocker could block the effects of the anti-inflammatory and cognitive enhancement when the pCREB signaling was blocked. In this study,10 days after the surgery all the animals were tested for the locomotor activity、the object recognition test、the step down passive avoidance task and the Morris water maze test.24 h after the last behavioral test, the animals were sacrificed and the pCREB、CREB、NF-κB p65 and the TNF-αin the hippocampus were detected by western bloting.(3) We first used the PDE4A subtype knockout mice comparing with wild-type on the learning and memory changes and its mechanism. The locomotor activity, the object recognition test and step down passive avoidance test were conducted in order. The pCREB, CREB and BDNF were detected respectively in the hippocampus, cortex and striatum by the method of Western bloting and the neurogenesis of hippocampus was tested by immunohistochemistry. Second, in order to demonstrated the regulation of PDE4D on the memory deficits and neuroinflammation induced by the Aβ1-42, we microinjected the aged Aβ1-42 (0.4μg/side) into the CA1 subregion of hippocampus in the PDE4D-/-mice.10 days after the surgery, the locomotor activity, the object recognition test and step down passive avoidance test and the Morris water maze were tested in order. And then the pCREB, CREB, NF-κB p65, TNF-α, IL-1β, COX-2, IL-10 in the hippocampus were detected by Western bloting. Finally, in order to clarify the regulation of PDE4A, B and D subtypes on the mempry deficits and neuroinflammation induced by Aβ1-42, the PDE4A, B and D knockout mice were used and the DG subregion of hippocampus was microinjected the aged Aβ1-42 (0.4μg/side).10 days after the surgery, the locomotor activity, the object recognition test and step down passive avoidance test were tested in order. And then the pCREB, BDNF, NF-κB p65 and COX-2 in the hippocampus were detected by Western bloting.Results:(1) Our preliminary data shown that the microinjected position accurately into the CA1 and DG subregions of the hippocampus. The treatment of Aβ1-42 did not changed the locomotor activity between all groups. The recognition index were significantly decreased by Aβ1-42 microinjected into the CA1 or DG subregions of hippocampus. For the step-down passive avoidance test, two doses of Aβ1-42 injected into the CA1 subregion significantly reduced the escape latency (retention), and only the dose 1.6μg/side injected into the DG subregion of hippocampus significantly reduced escape latency (retention). Consistent with the behavior test, Aβ1-42 injected into the CA1 subregion significantly decreased the expression of pCREB and BDNF in the hippocampus. Only the dose of 1.6μg/side microinjected into the DG subregion significantly reduced the expression of pCREB in the hippocampus expect for the expression of BDNF was significantly decreased by the both dose. All treatments did not effect the expression of CREB between all groups. Two doses of Aβ1-42 significantly increased NF-κB p65 and IL-1βexpression in the CA1 and DG subregions of hippocampus.(2) We found that the Aβ1-42 (0.1,0.4 and 1.6μg/side) dose-dependently dexreased the cognitive function in the object recognition test, the step down passive avoidance test and the Morris water maze test, and the rolipram (0.5 mg/kg) reversed these effects. All the treatments had no effect on the locomotor activity. The Aβ1-42 also dose-dependently increased the expression of the IL-1β、TNF-α、NF-κB p65、 COX-2 levels and the rolipram significantly reduced the release of these inflammation-related factors. And the Aβ1-42 also dose-dependently inhibited the phosphorylation of CREB, this was reversed by the rolipram. And we also found that Aβ1-42 dose-dependently slightly induced the formation of IL-10, and rolipram also played regulatory role on the IL-10. Our study also found that rolipram had no effect on the the Aβ1-42 itself. In order to clarify which PDE4 subtypes were impacted by Aβ1-42, we found that Aβ1-42 significantly increased PDE4A5、PDE4A?、PDE4B1、PDE4B2、PDE4B3、PDE4B4、PDE4D4、PDE4D5 levels. We also found that microinfusions of Aβ25-35 into bilateral CA1 subregions (10μg/side) impaired learning ability and long-term spatial memory in the Morris water maze test, it also decreased long-term memory in the passive avoidance test. Levels of NF-κB p65 and Bax were increased while levels of pCREB and Bcl-2 were decreased in the hippocampus in rats treated with Aβ25-35. These effect of Aβ25-35 were reversed by repeated treatment with rolipram in a dose dependent manner. The further study also demonstrated that TNF-αantibody blocked the inflammatory response induced by Aβ1-42 and only slightly alleviated the learning and memory deficits. While the TNF-αslightly increased the Aβ1-42-induced learning and memory impairments. We also found that the PKA inhibitor of H89 not only blocked the cognitive enhancing effect of rolipram and also the anti-inflammatory effects.(3) Mice deficient in PDE4A display significant memory enhancing effect comparied with WT mice. But the locomotor activity was not changed by PDE4A subtype knock out. The PDE4A knockout also increased the expression of pCREB and BDNF in the hippocampus expect for the cortex and striatum. The proliferation and survival of newborn cells were increased significantly in the hippocampus of PDE4A knockout mice compared with WT mice. The PDE4D knockout mice also reversed the memory deficits downregulation of pCREB signaling induced by Aβ1-42 microinjected into the CA1 subregion of the hippocampus. Although the mice deficit in PDE4D did not significantly decreased the expression of the NF-κB p65, TNF-α, IL-1κinduced by Aβ1-42, the PDE4D knockout displayed inhibited effect on COX-2 overexpression and slightly increased effect on IL-10. Further studies also showed that PDE4A and D subtypes knockout reversed the memory deficits induced by microinjected Aβ1-42 into the DG subregion of hippocampus except for PDE4B knockout mice. Consistent with behavioral test, the PDE4A and PDE4D knockout mice reversed the pCREB/BDNF signaling except for PDE4B knockout mice. The PDE4A and PDE4D knockout mice decreased the expression of COX-2 induced by Aβ1-42 except for the NF-κB p65. However, PDE4B knockout mice displayed the downregulation effect on the expression of NF-κB p65 except for the COX-2.Conclusion:Microinjected Aκ1-42 or 25-35 into the CA1 and (or) DG subregions of the hippocampus produced the inflammatory responses, apoptosis, memory deficits and downregulation of pCREB/BDNF signaling. These provide a new theoretical basis which help to find the new drugs targets focus on the anti-inflammatory and accompanied upregulaition of pCREB-mediated signaling in Alzheimer’s disease. PDE4 inhibitor (rolipram) reversed the memory deficits and neuroinflammation induced by Aβand different PDE4 subtypes were increased by Aβ. PDE4 may be play important role on the regulation on the memory deficits and neuroinflammation in AD. We also extended these findings to the demonstration that the protective effect of PDE4 against Aβinsult may be partly attributed to the blockade of inflammatory responses and apoptosis in the hippocampus. And we also demonstrated that PDE4 inhibitor may be through pCREB signaling to reversed the memory deficits and inflammation induced by Aβ. Our study also firstly demonstrated that PDE4A subtype plays important role on the memory and neurogenesis. Further study also clarified that the PDE4A and D subtypes kncokout reversed the memory deficits and downregulation of pCREB signaling except for the neuroinflammation induced by Aβ. Although PDE4B subtype knockout did not reversed the memory deficits and downregulation of pCREB signaling, PDE4B knockout mice displayed anti-inflammatory effect. The regulate effect of PDE4A、B and D subtypes on COX-2 may be contribuate to the different regulation on neuroinflammation. These demonstrated that COX-2 may be play an important role on the progress of anti-inflammatory effect.更多还原