【作者】 周伟；

【导师】 于金明； 侯明；

【作者基本信息】 山东大学 ， 肿瘤学， 2009， 博士

【摘要】 研究背景CXCL12/CXCR4生物学轴是指由趋化因子CXCL12与其特异性受体CXCR4相互作用而构成的一个与细胞间信息传递、细胞迁移有密切关系的偶联分子对,其实质在于CXCR4对其配体CXCL12的高度亲和力和绝对特异性,即CXCR4作为CXCL12的专属受体而存在。CXCL12/CXCR4生物学轴不仅参与白细胞浸润、细胞迁移和器官发育等一系列重要的生理过程,在肿瘤的局部侵袭和器官特异性转移中具有重要作用,包括乳腺癌、肝癌、基底细胞癌等。表遗传机制可以导致基因失活,通过抑癌基因表达缺失,参与肿瘤的发生和发展。在哺乳动物细胞中,最重要的表遗传机制是DNA甲基化。在CXCL12基因5′的转录起始区有一个典型的CpG岛区,该区域缺乏一个真正的TATA调控区,提示CXCL12基因启动子区甲基化可能参与基因转录水平的调节。在哺乳动物中,DNA的甲基化主要通过DNA甲基转移酶DNMT1,DNMT3A和DNMT3B来建立和维持,这三种酶之间存在着协同作用。研究目的研究趋化因子CXCL12及其受体CXCR4在星形细胞瘤和乳腺癌组织中的表达与肿瘤恶性程度之间的关系;初步探讨CXCL12基因的启动子区DNA甲基化与mRNA表达变化的内在因果联系;探讨DNA甲基转移酶的表达与DNA异常甲基化之间的关系;分析DNA甲基化在CXCL12/CXCR4生物轴参与星形细胞瘤和乳腺癌恶性进展中的调控机制。研究方法（一）星形细胞瘤中DNA甲基化对CXCL12/CXCR4生物学轴的基因调控1.临床收集76例脑星形细胞瘤（WHO分级:Ⅱ级20例,Ⅲ级26例,Ⅳ级30例）和10例正常脑组织标本,收集患者的临床和病理资料。2.传统的逆转录聚合酶链反应（RT-PCR）和实时定量PCR方法（SYBR GreenReal-time PCR）,检测CXCL12、CXCR4 mRNA在星形细胞瘤和正常脑组织中的表达情况。采用Kruskal-Wallis检验、Mann-Whitney检验等统计学方法分析CXCL12、CXCR4 mRNA表达与肿瘤病理组织分级、患者性别和年龄等临床参数间的关系。3.DNA经亚硫酸氢钠修饰后,甲基化特异性PCR（methylation specific PCR,MSP）方法,检测CXCL12基因在星形细胞瘤和正常脑组织中的DNA甲基化发生情况。采用卡方检验分析CXCL12基因甲基化与肿瘤病理组织分级、患者性别和年龄等临床参数间的关系;另外分析CXCL12基因甲基化与CXCL12 mRNA表达之间的关系。4.传统的逆转录聚合酶链反应（RT-PCR）和实时定量PCR方法（SYBR GreenReal-time PCR）,检测DNA甲基转移酶DNMT1、DNMT3A和DNMT3B基因在星形细胞瘤和正常脑组织中的mRNA表达情况。采用Spearman秩相关系数检验分析DNMT1、DNMT3A和DNMT3B mRNA表达相互之间的相关性;另外分析DNA甲基转移酶（DNMTs）表达与CXCL12基因甲基化的关系。（二）乳腺癌中DNA甲基化对CXCL12/CXCR4生物学轴的基因调控1.临床收集63例乳腺癌标本（TNM分期:Ⅰ期9例,ⅡA期25例,ⅡB期13例,ⅢA期16例）和20例正常乳腺组织标本,收集患者的临床和病理资料。2.免疫组织化学染色-SP法,检测ER、PR、Her-2、p53和Ki-67蛋白在乳腺癌和正常乳腺组织中的表达情况。3.传统的逆转录聚合酶链反应（RT-PCR）和实时定量PCR方法（SYBR GreenReal-time PCR）,检测CXCL12、CXCR4 mRNA在乳腺癌和正常乳腺组织中的表达情况。采用Kruskal-Wallis检验、Mann-Whitney检验等统计学方法分析CXCL12、CXCR4 mRNA表达在不同年龄、绝经情况、淋巴结转移、组织学分级、病理类型、肿瘤大小以及不同免疫组化指标（ER、PR、Her-2、p53和Ki-67）等临床参数间的关系。4.DNA经亚硫酸氢钠修饰后,甲基化特异性PCR（methylation specific PCR,MSP）方法,检测CXCL12基因在乳腺癌和正常乳腺组织中的DNA甲基化发生情况。采用卡方检验分析CXCL12基因甲基化在不同年龄、绝经情况、淋巴结转移、组织学分级、病理类型、肿瘤大小以及不同免疫组化指标（ER、PR、Her-2、p53和Ki-67）等临床参数间的关系;另外分析CXCL12基因甲基化与CXCL12mRNA表达之间的关系。5.传统的逆转录聚合酶链反应（RT-PCR）和实时定量PCR方法（SYBR GreenReal-time PCR）,检测DNA甲基转移酶DNMT1、DNMT3A和DNMT3B基因在乳腺癌和正常乳腺组织中的mRNA表达情况。采用Spearman秩相关系数检验分析DNMT1、DNMT3A和DNMT3B mRNA表达相互之间的相关性;另外分析DNA甲基转移酶（DNMTs）表达与CXCL12基因甲基化的关系。6.免疫组织化学染色-SP法,检测CXCL12、CXCR4蛋白在乳腺癌和正常乳腺组织中的表达情况。采用卡方检验分析CXCL12、CXCR4蛋白在不同年龄、绝经情况、淋巴结转移、组织学分级、病理类型、肿瘤大小以及不同免疫组化指标（ER、PR、Her-2、p53和Ki-67）等临床参数间的关系;另外分析CXCL12蛋白表达与CXCL12基因甲基化之间的关系。研究结果（一）星形细胞瘤中DNA甲基化对CXCL12/CXCR4生物学轴的基因调控1.实时定量RT-PCR结果显示,星形细胞瘤标本CXCR4 mRNA表达量（6.00±4.77）高于正常脑组织(0.53±0.24,P＜10^-7)。CXCR4 mRNA表达量与组织分级呈正相关(P=1.2×10^-5)。CXCR4 mRNA在不同性别、不同年龄组间的差异均无统计学意义（P＞0.05）。2.实时定量RT-PCR检测CXCL12 mRNA在16例（21.1%）星形细胞瘤表达下调或缺失（均为Ⅱ级）,在47例（61.8%）星形细胞瘤表达上调（Ⅲ级20例,Ⅳ级27例）。CXCL12 mRNA在Ⅳ级星形细胞瘤的表达（7.88±6.12）高于正常脑组织(1.86±0.62,P=4.74×10^-5)。Ⅱ级和Ⅲ级星形细胞瘤CXCL12 mRNA表达,与正常脑组织之间的差异无统计学意义（P值分别为0.534、0.113）。另外,CXCL12 mRNA在不同性别、不同年龄组间的差异均无统计学意义（P＞0.05）。3.CXCL12基因在星形细胞瘤中的DNA甲基化率为34.2%。在26例发生甲基化星形细胞瘤中,21例是不完全甲基化,5例是完全甲基化。所有10例正常脑组织中均未检测到CXCL12基因甲基化。CXCL12基因在Ⅱ级星形细胞瘤的甲基化率为55.0%（11/20）,Ⅲ级为34.6%（9/26）,Ⅳ级为20.0%（6/30）。CXCL12基因甲基化率与星形细胞瘤WHO分级呈负相关(r=-1,P=10^-6)。CXCL12基因甲基化在不同性别、不同年龄组间的差异均无统计学意义（P＞0.05）。在Ⅱ级和Ⅲ级星形细胞瘤中,CXCL12基因发生甲基化的星形细胞瘤样本中的mRNA的表达水平低于未发生甲基化的星形细胞瘤样本（P=0.001）。但在Ⅳ级星形细胞瘤中,CXCL12 mRNA表达量在CXCL12基因发生甲基化与未发生甲基化的星形细胞瘤样本间的差异无统计学意义（P=0.550）。4.实时定量RT-PCR结果显示,在星形细胞瘤中DNMT1、DNMT3A和DNMT3B基因表达相互之间呈正相关,DNMT1和DNMT3A（r=+0.344,P=0.002）、DNMT1和DNMT3B（r=+0.389,P=0.001）、DNMT3A和DNMT3B（r=+0.305,P=0.007）。DNMT1、DNMT3A和DNMT3B基因在CXCL12发生甲基化的星形细胞瘤表达均高于CXCL12未发生甲基化的星形细胞瘤（P值分别为0.025、0.003、0.043）。（二）乳腺癌中DNA甲基化对CXCL12/CXCR4生物学轴的基因调控1.实时定量RT-PCR结果显示,在乳腺癌中CXCR4 mRNA表达水平（1.22±0.80）高于正常乳腺组织（0.65±0.59,P=0.001）。CXCR4 mRNA在乳腺癌中表达上调与淋巴结转移数目（＞3个腋窝淋巴结转移）、Her-2表达情况（3+）密切相关（P值分别为0.013、0.031）。2.实时定量RT-PCR结果显示,在乳腺癌中CXCL12 mRNA表达水平（1.41±1.18）低于正常乳腺组织（2.02±0.71,P=0.031）。CXCL12 mRNA在乳腺癌中表达下调与淋巴结转移数目（＞3个腋窝淋巴结转移）、ER表达情况（0,1+,2+）密切相关（P值分别为0.017、0.047）。但CXCL12 mRNA在不同年龄、绝经情况、组织学分级、病理类型、肿瘤大小以及其他免疫组化指标（PR、Her-2、p53和Ki-67）组间的差异均无统计学意义（P＞0.05）。3.乳腺癌CXCR4蛋白阳性表达率为79.4%。所有20例正常乳腺组织CXCR4蛋白均阴性表达。CXCR4蛋白高表达与淋巴结转移数目（＞3个腋窝淋巴结转移）密切相关（P=0.035）。乳腺癌CXCL12蛋白阳性表达率为38.1%。所有20例正常乳腺组织CXCL12蛋白均阳性表达。乳腺癌CXCL12蛋白表达在不同患者年龄、绝经情况、淋巴结转移数目、组织学分级、病理类型、肿瘤大小以及不同免疫组化指标（ER、PR、Her-2、p53和Ki-67）组间的差异均无统计学意义（P＞0.05）。4.CXCL12基因在乳腺癌中的DNA甲基化率为52.4%。所有20例正常乳腺组织中均未检测到CXCL12基因甲基化的发生。在33例发生甲基化的乳腺癌中,20例为不完全甲基化,13例为完全甲基化。乳腺癌CXCL12基因甲基化状态与淋巴结转移数目（＞3个腋窝淋巴结转移）、ER表达情况（0,1+,2+）密切相关（P值分别为0.032、0.011）。乳腺癌CXCL12基因甲基化状态与CXCL12 mRNA表达呈负相关(r=-0.568,P=1.2×10^-6),与CXCL12蛋白表达水平亦呈负相关（P=0.001）。5.实时定量PCR结果显示,在乳腺癌中DNMT1、DNMT3A、DNMT3B mRNA表达相互之间呈正相关,DNMT1和DNMT3A（r=+0.272,P=0.031）、DNMT1和DNMT3B（r=+0.303,P=0.016）、DNMT3A和DNMT3B（r=+0.389,P=0.002）。DNMT1、DNMT3B基因在CXCL12发生甲基化乳腺癌的mRNA表达高于CXCL12未发生甲基化乳腺癌（P值分别为0.012、0.026）。但DNMT3A mRNA表达在CXCL12发生甲基化组和CXCL12未发生甲基化组间的差异无统计学意义（P=0.185）。研究结论1.在部分星形细胞瘤中（主要是低度恶性星形细胞瘤）,CXCL12基因通过DNMTs参与的DNA甲基化导致其表达下调。CXCR4受体在星形细胞瘤中过表达与星形细胞瘤恶性程度呈正相关。2.趋化因子CXCL12及其受体CXCR4在Ⅳ级星形细胞瘤（胶质母细胞瘤）中高表达,可能通过自分泌作用参与胶质母细胞瘤的局部侵袭。3.乳腺癌细胞通过DNMTs参与的DNA甲基化导致CXCL12表达下调,但保持CXCR4受体的高表达,与淋巴结转移情况密切相关。4.星形细胞瘤和乳腺癌DNA甲基化是导致CXCL12表达失活的主要原因之一,但并不是唯一的原因,可能还存在着其它机制。更多还原

【Abstract】 BackgroundThe CXCL12/CXCR4 signaling axis is a coupled molecular pair. Binding of the chemokine CXCL12 to its receptor CXCR4 activates a variety of intracellular signal transduction pathways and cell migration. CXCR4 exhibites high affinity and absolute specificity for its ligand CXCL12. So CXCL12 appears to be the only ligand for CXCR4. The CXCL12/CXCR4 signaling axis regulates a variety of immune responses, cell migration and embryogenesis in many normal and pathophysiological processes. Recently, it has been established that it can serve as tissue-specific attractant molecules for tumor cells, promoting tumor cell invasion to adjacent sites, including breast, hepatocellular, and basal cell carcinomas.It has emerged that epigenetic events can lead to gene inactivation that contributes to neoplasia. The most important mechanism in mammalian cells is DNA methylation. A typical CpG island on the start site of the transcripton is revealed in the 5’ region of the CXCL12 gene. Moreover, this region of CXCL12 lacks a true TATA-box. It indicated that hypermethylation of the promoter region of CXCL12 likely plays an important role in the regulation of their mRNA levels in malignant tumors. In mammals, DNA methylation patterns are established and maintained by mainly three known functional DNMT genes, namely DNMT1, 3A, and 3B. There is a considerable level of cooperation and functional overlap among them. ObjectiveTo study the expressions of the CXCL12 and CXCR4 genes in primary astrocytoma and breast carcinoma, as well as their correlation with various clinical characteristics. To identify the methylation status of CXCL12, as well as their relationship with their mRNA levels. To study the relationship of DNA methyltransferases with DNA hypermethylation. To analyze the mechanism of DNA hypermethylation involved in tumor progression in primary astrocytortia and breast carcinoma.Methodsi DNA methylation-mediated regulation of CXCL12/CXCR4 axis in primary astrocytoma1. Seventy-six of primary astrocytoma samples （20 of WHO grade II, 26 of WHO grade III, and 30 of WHO grade IV） and 10 normal brain tissues were collected during surgery. All the patients were followed up after surgery.2. The mRNA expression levels of CXCL12 and CXCR4 genes were detected by conventional RT-PCR and SYBGreen Real-time PCR assays in these primary astrocytomas and normal brain tissues. The CXCL12 and CXCR4 mRNA expression levels among various histologic grades and age subgroups were evaluated by using the Kruskal-Wallis test. The comparisons of CXCL12 and CXCR4 mRNA levels between different sex teams were calculated using the Mann-Whitney test.3. After sodium bisulfite modification of genomic DNA, methylation specific PCR （MSP） was carried out to analyze the methylation status of promoter regions of the CXCL12 gene in these primary astrocytomas and normal brain tissues. Correlations of methylation alterations of the CXCL12 gene with clinicopathologic parameters, including histologic grade, sex and age were analyzed by x² test. Comparisons of CXCL12 mRNA levels between different methylation status teams were also calculated.4. The mRNA expression levels of DNMT1, DNMT3A and DNMT3B genes were detected by conventional RT-PCR and SYBGreen Real-time PCR assays in these primary astrocytomas and normal brain tissues. The Spearman rank correlation test was used to determine the links between continuous mRNA values of each DNMT. Relationships of the mRNA expression levels of DNMT1, DNMT3A and DNMT3B genes with the CXCL12 methylation status were also analyzed.ii DNA methylation-mediated regulation of CXCL12/CXCR4 axis in primary breast cancer1. Sixty-three of primary breast carcinoma samples and 20 normal breast tissues were obtained from patients （9 of stage I, 25 of stage IIA, 13 of stage IIB, and 16 of stage IIIA） , when treated with curative resectional surgery. All the patients were followed up after surgery.2. The protein expression levels of ER, PR, Her-2, p53, and Ki-67 in primary breast carcinomas and normal breast tissues were determined by immunohistochemical detection.3. The mRNA expression levels of CXCL12 and CXCR4 genes were detected by conventional RT-PCR and SYBGreen Real-time PCR assays in these primary breast carcinomas and normal breast tissues. Correlations of the CXCL12 and CXCR4 mRNA expression levels with clinicopathologic parameters were analyzed using the Kruskal-Wallis test or the Mann-Whitney test, including age, menopausal status, lymph node status, histologic grade, tumor type, tumor diameter and immunohistochemical staining status （ER, PR, Her-2, p53, and Ki-67）.4. After sodium bisulfite modification of genomic DNA, methylation specific PCR （MSP） was carried out to analyze the methylation status of promoter regions of the CXCL12 gene in these primary breast carcinomas and normal breast tissues. Correlations of methylation alterations of the CXCL12 gene with clinicopathologic parameters were analyzed by x² test, including age, menopausal status, lymph node status, histologic grade, tumor type, tumor diameter and immunohistochemical staining status （ER, PR, Her-2, p53, and Ki-67）. Relationships of the CXCL12 mRNA expression levels with CXCL12 methylation status were also analyzed.5. The mRNA expression levels of DNMT1, DNMT3A and DNMT3B genes were detected by conventional RT-PCR and SYBGreen Real-time PCR assays in these primary breast carcinomas and normal breast tissues. The Spearman rank correlation test was used to determine the links between continuous mRNA values of each DNMT. Relationships of the mRNA expression levels of DNMT1, DNMT3A and DNMT3B genes with the CXCL12 methylation status were also analyzed.6. The protein expression levels of CXCL12 and CXCR4 in primary breast carcinoma and normal breast tissues were determined by immunohistochemical detection. Correlations of the CXCL12 and CXCR4 protein expression levels with clinicopathologic parameters were analyzed by x² test, including age, menopausal status, lymph node status, histologic grade, tumor type, tumor diameter and immunohistochemical staining status （ER, PR, Her-2, p53, and Ki-67）. Relationships of the CXCL12 protein expression levels with CXCL12 methylation status were also analyzed.Resultsi DNA methylation-mediated regulation of CXCL12/CXCR4 axis in primary astrocytoma1. Quantitative analysis of the CXCR4 transcript revealed significantly (P < 10^-7) higher levels in astrocytomas （6.00±4.77, Mean±SD） than in normal brain tissues （0.53±0.24）. The expression of CXCR4 increased with increasing tumor grades (P = 1.2×10^-5). Other parameters such as sex and age of the patients did not show significant correlations （P > 0.05） with CXCR4 mRNA status.2. Quantitative real-time PCR showed that 16 of 76 （21.1%） astrocytoma samples exhibited no or low expression of CXCL12 mRNA. These 16 samples were all in WHO grade II. Meanwhile, 47 of 76 （61.8%） astrocytomas displayed elevated transcription of CXCL12, with 20 of WHO grade III, and 27 of WHO grade IV. The expression levels of CXCL12 mRNA in WHO grade IV astrocytomas （7.88±6.12） were significantly (P = 4.74×10^-5) higher than in normal brain tissues （1.86±0.62）. There were no significantly difference between WHO grade II / III and normal brain tissues （P =0.534, 0.113, respectively）. Other factors such as sex and age showed no links （P > 0.05） with the CXCL12 mRNA levels.3. Hypermethylation of the CXCL12 gene was detected in 34.2% of astrocytomas. Both methylated and unmethylated products were present in 21 of CXCL12-methylated tumors. The methylation frequency of CXCL12 in astrocytomas of WHO grade II was 55.0% （11/20）, in WHO grade III was 34.6% （9/26） and in WHO grade IV was 20.0% （6/30）. No methylation of CXCL12 was observed in the ten normal brain tissues. A significant inverse relationship (r = - 1, P = 10^-6) between CXCL12 methylation and WHO grades were found, but no statistical （P > 0.05） differences were found among different sexes and ages.In WHO grades II - III astrocytomas, the Mann-Whitney test showed a significant difference （P = 0.001） of CXCL12 mRNA levels between CXCL12 methylated and unmethylated groups. However, such statistical difference was not observed in WHO grades IV astrocytomas samples （P = 0.550）.4. Quantitative real-time PCR showed that the mRNA levels of DNMT1, DNMT3A and DNMT3B correlated strongly with each other, as follows: DNMT1 with DNMT3A （r = +0.344, P = 0.002）, DNMT1 with DNMT3B （+r = 0.389, P = 0.001）, and DNMT3A with DNMT3B （r = +0.305, P = 0.007）. Furthermore, the expression levels of DNMT1, DNMT3A and DNMT3B were significantly （P = 0.025, 0.003, 0.043, respectively） higher in the CXCL12-methylated astrocytomas than those in the CXCL12-unmethylated ones .ii DNA methylation-mediated regulation of CXCL12/CXCR4 axis in primary breast cancer1. Quantitative analysis of the CXCR4 transcript revealed significantly （P = 0.001） higher levels in breast carcinoma tissues （1.22±0.80） than in normal mammary tissues （0.65±0.59）. CXCR4 overexpression in primary breast carcinomas was significantly related to lymph node metastasis （> 3 positive lymph nodes metastasis, P = 0.013）, and strong Her-2 expression （P = 0.031）.2. Quantitative analysis of the CXCL12 transcript revealed significantly （P = 0.031） lower levels in breast carcinoma tissues （1.41±1.18） than in normal mammary tissues （2.02±0.71）. Furthermore, significantly （P = 0.017） lower expression levels of CXCL12 were found in tumors with > 3 positive lymph nodes metastasis. A significant difference （P = 0.047） in CXCL12 mRNA levels was also found between estrogen receptor negativity and positive group. However, no links （P > 0.05） were found between CXCL12 mRNA expression levels and age, menopausal status, histological grade, tumor type, macroscopic tumor size, or other standard immunohistochemical parameters （ER, PR, Her-2, p53, and Ki-67）.3. The rate of breast carcinoma specimens stained positive for CXCR4 was 79.4%. All normal breast specimens resulted in negative staining for CXCR4. There was significant correlation （P = 0.035） between aberrant CXCR4 expression and lymph node metastasis （> 3 positive lymph nodes metastasis）.Only 38.1% of breast cancers showed CXCL12 positive staining. All normal breast specimens exhibited positive staining for CXCL12. None of the clinicopathological factors revealed a correlation （P > 0.05） with CXCL12 protein, including age, menopausal status, lymph node metastasis status, histological grade, tumor type, macroscopic tumor size, and immunohistochemical parameters （ER, PR, Her-2, p53, and Ki-67）.4. There were 52.4% of primary breast tumors hypermethylated in the CXCL12 promoter region, whereas hypermethylation of CXCL12 was not observed in any of the 20 normal breast tissues. Methylated and unmethylated products were present in 20 of 33 CXCL12-methylated tumors. CXCL12 methylation was significantly associated with > 3 positive lymph nodes metastasis （P = 0.032）, and estrogen receptor negativity （P = 0.011）. A significant inverse relationship (r = -0.568, P =1.2×10^-6) between CXCL12 methylation status and their mRNA levels was observed. A strong inverse relationship （P = 0.001） between CXCL12 methylation and protein expression was also observed.5. Quantitative real-time PCR showed that the mRNA levels of DNMT1, DNMT3A and DNMT3B correlated strongly with each other, as follows: DNMT1 with DNMT3A （r = +0.272, P = 0.031）, DNMT1 with DNMT3B （r = +0.303, P = 0.016）, and DNMT3A with DNMT3B （r = +0.389, P = 0.002）. The expression levels of DNMT1 （P = 0.012） and DNMT3B （P = 0.026） were significantly higher in the CXCL12-methylated primary breast carcinomas than in the CXCL12-unmethylated ones. But no significant difference （P = 0.185） of DNMT3A was observed between the CXCL12-methylated and CXCL12-unmethylated breast carcinomas. Conclusions1. DNA hypermethylation of CXCL12 is implicated in part of astrocytomas, mainly in low-grade ones, via DNA hypermethylation by DNMTs. The chemokine receptor CXCR4 is overexpressed in astrocytoma, increasing with increasing WHO grade.2. Both CXCL12 and its receptor CXCR4 are up-regulated in astrocytomas of WHO grade IV （glioblastoma）. The presence of CXCL12 and CXCR4 in tumor cells, which constitutes autocrine paracrine CXCL12/CXCR4 signaling, promotes invasion ofglioblastomas.3. DNMTs-mediated DNA hypermethylation of CXCL12 plays an important role in the down-regulation of CXCL12 expression in breast carcinomas. Both up-regulation of CXCR4 and down-regulation of CXCL12 are observed in primary breast carcinomas, which are significantly related to lymph node metastasis status.4. DNA hypermethylation plays an important role in the down-regulation of the CXCL12 gene of astrocytoma and breast carcinoma. Other mechanisms might be involved in its regulation.更多还原