【作者】 许晨璞；

【导师】 王林波；

【作者基本信息】 浙江大学 ， 乳腺癌（专业学位）， 2013， 博士

【摘要】 背景：乳腺癌是世界范围内女性最常见的恶性肿瘤,且近十年来,其发病率呈持续上升趋势,严重威胁着女性健康。乳腺癌作为高度异质性的肿瘤,根据雌激素受体(Estrogen Receptor, ER)表达可分为ER阳性及ER阴性,根据基因分析又可分为luminal A型、luminalB型、基底细胞样型、HER-2过表达型及Normal-like等不同分子分型。尽管乳腺癌治疗手段较为多样,化疗的重要地位仍不可撼动,尤其近年来术前新辅助化疗的广泛开展使部分T2、T3期患者达到降期及保乳的目的,进而提高了患者的肿块切除率及生活质量。尽管尚不明确具体机制,目前陆续有研究显示ER阳性乳腺癌患者中虽有一部分对化疗敏感,但总体而言化疗反应较差,因此循证医学依据对在该类患者中使用新辅助化疗推荐力度较弱,转而较为推荐对ER阳性患者有一定疗效的新辅助内分泌治疗。然而尽管ER阳性乳腺癌患者辅助内分泌治疗疗效较好且总体预后佳,仍有部分患者在术前术后的内分泌治疗中出现耐药甚至复发转移,考虑到该类内分泌治疗无效患者中可能存在化疗敏感者,目前已有循证依据及诊疗常规均建议此时给予化疗,而由于迄今为止尚无分子标记物或临床病理指标可有效预测ER阳性乳腺癌患者化疗疗效,因此在这部分患者的术后辅助及复发后化疗方案选择方面存在一定的盲目性。鉴于ER阳性表达者占所有乳腺癌的75%-80%,故该类患者总数可能较为可观,因此若能找到有效的分子标记物对ER阳性患者进行化疗疗效预测以及预后判断,将有助于筛选该类患者以使其接受有效治疗并进一步预测其转归。FOXA1又称HNF3α(肝细胞核因子3α),为FOX翼状螺旋转录因子家族成员之一,经研究证实在ER阳性乳腺癌细胞表达较正常组织为高且能作为ER的先锋因子,影响其与DNA结合。本文拟探索FOXA1是否可能成为与ER阳性乳腺癌化疗疗效相关的有效分子标记物。鉴于新辅助化疗作为有效的体内药敏实验模型,可用以筛选化疗敏感者及其适合的化疗方案,故本研究首先以ER阳性乳腺癌新辅助化疗患者为研究对象,探索FOXA1与化疗疗效的相关性,若两者确实相关,拟通过体外研究初步探索该相关性的可能机制；最后通过临床回顾性分析探索FOXA1与ER阳性乳腺癌新辅助化疗患者预后的相关性以明确FOXA1在此类患者中是否同时具有预后判断价值。方法和结果：1. FOXAl表达与ER阳性乳腺癌新辅助化疗后病理反应的相关性分析2002年至2012年我科收治的123例ER阳性乳腺癌新辅助化疗患者,行化疗前肿块穿刺标本FOXA1免疫组化染色及Histoscore评分,对术后大体标本HE染色切片进行阅片,根据Miller&Payne病理学评估标准[9]判断新辅助化疗病理反应并分为两组：病理无反应(G1-3,大于10%肿瘤细胞残留)及病理有反应(G4-5,0-10%肿瘤细胞残留)。结果显示病理有反应率25.2%(31例),病理无反应率74.8%(92例)。卡方检验显示ER阳性乳腺癌新辅助化疗患者化疗前穿刺标本FOXA1表达水平与化疗后病理反应(0～10%肿瘤细胞残留vs.大于10%肿瘤细胞残留)相关(P=0.002),FOXA1表达较高者病理反应率较低,肿瘤细胞残留较多。为明确该相关性是否与ER达有关,将ER进行不同分组(评分1,2,3三组；评分1,2与3两组；评分1与2,3两组)并与上述FOXA1表达分别进行卡方检验,发现均不存在相关性(p值分别为0.702,0.345及0.564)。结果提示,FOXA1可能可以独立于ER作为ER阳性乳腺癌患者的化疗疗效预测指标,并具有体外实验进一步研究的价值。2. FOXAl在ER阳性乳腺癌细胞中的耐药特性及其机制研究该部分拟在体外实验水平进一步验证FOXA1表达与ER阳性乳腺癌化疗敏感性相关,并在此基础上通过检测细胞周期、凋亡、增殖、克隆形成、迁移及侵袭等生物学行为及耐药相关分子变化以探索该相关性的可能机制。首先MTT法检测ER阳性乳腺癌细胞株MCF-7、T47D及MCF-7经阿霉素诱导形成的耐药株MCF-7/ADR(以下简称M/A)对临床乳腺癌化疗最常用蒽环类代表药阿霉素及紫杉类代表药紫杉醇的敏感性差异,Western blot检测三者FOXA1蛋白表达差异,结果显示化疗敏感性由高至低为MCF-7、T47D及M/A,而FOXA1表达水平由低至高亦为MCF-7、T47D及M/A。初步提示FOXA1表达水平与ER阳性乳腺癌细胞株化疗敏感性相关。继而使用慢病毒表达系统FOXA1质粒(pMD2G、pSPAX2、pSI-FOXA1)以及对照空载质粒(pMD2G、pSPAX2、pSI-NEG)稳定转染FOXA1弱表达的MCF-7细胞株,嘌呤霉素筛选,Western blot验证转染成功。选取MCF-7/FOXA1-15(以下简称MF15)及MCF-7/FOXA1-20(以下简称MF20)两株FOXA1过表达克隆株、以MCF-7/vector空载对照(以下简称Mneg)连同MCF-7母代细胞作为阴性对照、以FOXA1较高表达的T47D及M/A作为阳性对照,用MTT法检测对阿霉素及紫杉醇化疗敏感性变化；流式技术比较细胞周期、凋亡；MTT检测生长曲线；平板克隆形成实验检测克隆形成能力；Transwell小室测迁移/侵袭能力等功能实验探索FOXA1过表达对ER阳性乳腺癌细胞株生物学行为的影响。结果显示FOXA1过表达的MF15、MF20及较高表达的T47D、M/A细胞株对阿霉素及紫杉醇的化疗敏感性均明显较FOXA1低表达的MCF-7及Mneg差,FOXA1过表达或高表达可促进细胞周期从G1进入到S/G2期、凋亡抑制、增殖加快、克隆形成及迁移侵袭能力增加。随后通过siRNA干扰技术敲低ER阳性乳腺癌细胞株MCF-7、M/A及T47D的FOXA1表达,RT-qPCR及Western blot验证敲低成功。同上述过表达实验方法进行化疗敏感性及生物学行为检测。结果显示FOXA1表达敲低后MCF-7、M/A及T47D化疗敏感性均出现不同程度增加,并可使细胞出现G1期阻滞、凋亡增加、增殖减缓、克隆形成及迁移侵袭能力减弱。基于以上改变使用RT-qPCR及Western blot技术,检测上述FOXA1过表达及敲低实验中耐药相关基因及蛋白表达变化。结果显示在mRNA及蛋白水平,FOXA1过表达或高表达时耐药相关基因GST-pi显著增加；细胞周期相关基因cyclinE1增加,cyclinD1、p53、p27、p21表达下降；抗凋亡基因Bcl-xl明显增加；FOXA1敲低后上述基因出现相反变化。此外,FOXA1表达变化还可影响TGF-β信号通路中关键蛋白TGF-β、经典信号通路中AKT及MAPK(p38)磷酸化蛋白的表达。3.ER阳性乳腺癌新辅助化疗前后FOXA1表达差异及预后相关性分析因前两部分实验结果显示,FOXA1表达可引起ER阳性乳腺癌化疗敏感性及细胞增殖、迁移、侵袭能力改变,故该部分拟探索此类患者中FOXA1在化疗后残留肿瘤组织中的表达情况及其是否具有预后判断价值。首先选取实验第一部分123例患者中具有新辅助化疗后手术大体标本的110例进行术后标本FOXA1免疫组化染色及Histoscore评分,Wilcoxon带符号秩分析显示上述患者新辅助化疗前后FOXA1阳性率及表达量均无明显改变(p值分别为0.078及0.129),但FOXA1染色定位改变显著(p<0.0001),表现为胞核胞浆共染者由化疗前2.7%增至化疗后31.8%。进一步卡方检验及Kaplan-Meier生存分析分别显示化疗后FOXA1染色定位与ER阳性乳腺癌新辅助化疗患者化疗疗效及预后均无相关性(p值分别为0.208及0.828)。随后对第一部分中123例资料完整的ER阳性乳腺癌新辅助化疗患者先后行卡方检验及Kaplan-Meier生存分析,发现以Histoscore大于3作为阳性判断标准时,FOXA1表达不与包括年龄、月经、组织学分类、脉管内癌栓、神经侵犯、癌结节形成、病理分期T、N、M,免疫组化染色ER、PR、HER-2、Ki67、E-cadherin、P53以及luminal分型在内的任何临床病理指标相关,也不与ER阳性、luminalA及luminalB型新辅助化疗患者预后相关。通过尝试,我们发现当以FOXA1评分大于等于20及小于20进行分组时,可较好区分ER阳性及luminalA型而非luminalB型患者的预后(p分别为0.026,0.012及0.691)。卡方检验显示FOXA1评分大于等于20与年龄大于50岁及绝经后状态相关(p值分别为0.012及0.005),但不与ER阳性乳腺癌新辅助化疗疗效相关(p=0.110)。COX回归模型单因素及多因素分析发现FOXA1评分大于等于20可做为ER阳性(单因素p=0.026,多因素p=0.004)及luminalA型(单因素p=0.012,多因素p=0.003)乳腺癌新辅助化疗患者的独立预后预测因子。结论：本实验通过临床回顾性分析提示FOXA1与ER阳性乳腺癌患者化疗疗效相关,可能可以作为该类患者的化疗疗效预测指标。体外实验显示FOXA1表达可能通过影响TGF-β、AKT、MAPK(p38)等信号通路,p53、p27、p21、cyclinE1、 Bcl-xl、GST-pi等分子,及其下游细胞周期、凋亡、增殖等生物学行为进而影响ER阳性乳腺癌细胞对包括蒽环类及紫杉类在内的化疗药物敏感性。ER阳性乳腺癌患者新辅助化疗前肿块穿刺标本与新辅助化疗后手术大体标本中FOXA1染色定位发生变化,但尚不具有临床意义。FOXA1、的Histoscore评分大于等于20分可作为ER阳性及luminalA型乳腺癌新辅助化疗患者预后不佳的独立预测因子。更多还原

【Abstract】 BACKGROUNDAs the most common malignancy in females worldwide, breast cancer with its ever increasing incidence in recent decade, severely threatens the health of women. It is a highly heterogeneous disease and can be categorized into ER positive and ER negative subgroups according to the estrogen receptor status or into luminal A, luminal B, basal-like, HER-2enriched and Normal-like subgroups based on genomic test. Although there is a variety of therapeutic choices for these patients, the fundamental importance of chemotherapy cannot be surpassed, especially with the advent of neoadjuvant chemotherapy which can downgrade the patients in T2or T3to increase their chance of complete resection of tumor or receiving breast conserving surgery and in turn having better life quality. Recent researches revealed that the overall chemotherapy response of ER positive breast cancer patients is not ideal except a proportion of chemosensitive populations in this subgroup despite the lack of understanding of its underlying mechanisms. Therefore evidence based medicine(EBM) weakly recommends its use in these ER positive patients and support the application of neoadjuvant endocrine therapy which is proven to have better efficacy in this subgroup. However, part of the ER positive patients have drug resisrence, relapse or metastasis during treatment with pre-or postoperative endocrine therapy anyway. Both the EBM and clinical routines highly recommend the use of chemotherapy in this situation since these endocrine resistence patients could possibly respond to chemotherapy. Due to the fact that there is no molecular biomarker or clincal pathological surrogate to predict the efficacy of chemotherapy in ER positive breast cancer patients effectively, it is usually difficult to choose the appropriate regimen either postoperative or after recurrence for them. Considering the large amount of ER positive patients, the number of the patients who could benefit from certain chemotherapy regimen is conceivable. Therefore, it is important to find an effective biological marker to predict the therapeutic efficacy of chemotherapy and prognosis of ER positive breast cancer patients, and screen out the sensitive ones who could benefit from effective chemotherapy. FOXA1, also named HNF3a(hepatic nuclear factor3a), is a member of Forkhead box winged-helix transcription factor family. It is found to be expressed higher in ER positive breast cancer cells than in normal breast tissues and can serve as a pioneer factor for ER to influence its combination with DNA. In this thesis, we plan to explore the possible predictive value of FOXA1for therapeutic efficacy of chemotherapy in ER positive breast cancer patients. Firstly, our study wants to explore the correlation between expression of FOXA1and chemotherapeutic efficacy in neoadjuvant chemotherapy settings since it is considered as an effective drug sensitive model to screen out the sensitive patients and their effective regimens. Secondly, we plan to explore the preliminary underlying mechanisms through in vitro tests for this association if it truly exists. Thirdly, we plan to clarify the prognostic relevance of FOXA1in ER positve patients who had received neoadjuvant chemotherapy in past.METHODS AND RESULTS1. Correlation analysis between FOXA1and the pathological response of ER positive breast cancer following neoadjuvant chemotherapy In total, there are123cases of ER positive breast cancer patients who had complete data record and core needle biopsy before neoadjuvant chemotherapy in our hospital from August2002to December2012. At a total of123cases, pathological response(no more than10%tumor residual) was observed in25.2%(31/123) patient while pathological nonresponse(more than10%tumor residual) in74.8%(92/123) of patients. The expression of FOXA1in ER positive breast cancer patients was correlated with pathological response by chi-square statistical analysis (P=0.002), and high FOXA1expression levels were more frequently found in nonresponding patients. Besides that, FOXA1expression was not associated with ER by the same statistical approach(p>0.05), suggesting the possible predictive value of FOXA1in chemotherapeutic efficacy of ER positive breast cancer patients independent of ER expression. However, the mechanism of whether and how FOXA1can lead to chemoresistance is still unknown.2. The characteristic of chemoresistance and its underlying mechanism of FOXA1in ER positive breast cancer cell linesMTT was performed in ER positive breast cancer cell lines MCF-7, T47D and Adriamycin induced chemoresistent cell line MCF-7/ADR using the most widely applied regimen in breast cancer treatment, adriamycin and paclitaxel. The Western blot assay were performed to evaluate FOXA1expression in the above cell lines at the protein level. The results showed that chemosensitivity from high to low is MCF-7, T47D, MCF-7/ADR respectively whereas the FOXA1expression level from low to high is also MCF-7, T47D, MCF-7/ADR respectively, indicating a correlation between FOXA1expression and chemosensitivity in ER positive breast cancer cell lines.Then we used lentiviral expression system to transfect MCF-7cells with FOXA1plasmid (pMD2G、pSPAX2、pSI-FOXA1) and vector plasmid (pMD2G、pSPAX2、 pSI-NEG) with subsequent Puromycin selection and confirmation by Western blot Choose MCF-7/FOXA1-15(MF15) and MCF-7/FOXA1-20(MF20) as overexpression clones, MCF-7and MCF-7/vector as negative control, and T47D, MCF-7/ADR as positive control. In the above six cell lines, MTT assay were performed to evaluate the drug sensitivity and growth curve, flow cytometry to detect any changes in cell cycle and apoptosis, colony formation assay to test the clonal formation, Transwell to test migration and invasion ability. The results showed that the chemosensitivity in negative control group was better than positive control and overexpression clones. Overexpression or high level of FOXA1can promote the cell cycle progression from G1phase to S/G2phase, inhibit apoptosis, promote proliferation, increase clonal formation and ability of migration and invasion.We then used siRNA interference technology to knockdown FOXA1in MCF-7, T47D and MCF-7/ADR followed by confirmation through RT-qPCR and Western blot assay. Then we tested the chemorensitivity and functional changes in the above cell lines by using the same approaches as mentioned above. The results showed that after FOXA1knockdown, the chemosensitivity of MCF-7, T47D and MCF-7/ADR were increased, cell cycle was blocked at G1phase, apoptosis was increased, proliferation was slowed down, the abilities of colony formation, migration and invasion were impaired.We next examined the preliminary mechanism of FOXA1in mediating the above changes related to chemosensitivity using Western blot and RT-qPCR assay. The results showed that overexpression or high levels of FOXA1could lead to upregulation of GST-pi, cyclinE1and Bcl-xl whereas downregulation of p53, p27, p21and cyclinD1both at mRNA and protein levels. Knockdown of FOXA1caused the opposite changes in these genes. Additionally, the change in expression of FOXA1can influence the protein level of TGF-β and phosphorylated AKT and MAPK(p38).Therefore we assumed that FOXA1may influence the possible signaling pathways including TGF-β, AKT and MAPK(p38), the expression level of p53、p27、p21、 cyclinE1、Bcl-x1、GST-pi, and their downstream biological functions such as cell cycle control, apoptosis and proliferation. And the above changes may all contribute to the chemosensitivity of ER positive breast cancer cell lines to chemotherapy including but not limited to adriamycin and paclitaxel.3. The expression variation of FOXA1around neoadjuvant chemotherapy in ER positive patients and relevant prognostic analysisSelect110cases who had operative tissue sections out of123patients being studied in the first part of this thesis. Use the same approach as the first part to stain and score FOXA1in these operative samples. Statistical analysis revealed no significant changes neither in positive rate nor in expression levels before and after neoadjuvant chemotherapy whereas the staining position changed obviously from mainly nuclear to both cytoplama and nuclear. However, this change did not correlate with chemotherapeutic efficacy or prognosis of these ER positive patients.The Chi-square test found that FOXA1expression defined by Histoscore more than3is not correlated with any clinicopathological variables including age, menstrual status, histological categories, lymphovascular embolus, nerve invasion, cancer nodular formation, pathological stages, IHC staining of ER, PR, HER-2, Ki67, E-cadherin, p53and luminal subtypes. Kaplan-Meier survival analysis also revealed no prognostic relevance of FOXA1expression in ER positive, luminal A or luminal B breast cancer patients.By increasing the criteria of FOXA1expression, we found that FOXA1histoscore no less than20and less than20can separate ER positive and luminal A breast cancer patients into groups with different prognosis, and serve as an independent prognostic marker. Although this grouping did not correlate with chemoresponse, it was associated with age and menstrual status.CONCLUSIONOur study revealed that FOXA1expression correlated with chemosensitivity of ER positive breast cancer patients, and could possibly serve as a predicative marker for these patients. In vitro test indicated that FOXA1may influence the possible signaling pathways including TGF-β,AKT and MAPK(p38), the expression level of p53、p27、 p21、cyclinE1、Bcl-xl、GST-pi, and their downstream biological functions such as cell cycle control, apoptosis and proliferation. And the above changes may all contribute to the chemosensitivity of ER positive breast cancer cell lines to chemotherapy including but not limited to adriamycin and paclitaxel. Retrospective study showed that the staining position of FOXA1changed after neoadjuvant chemotherapy without other clinical relevance. FOXA1histoscore no less than20and less than20can separate ER positive and luminal A breast cancer patients into groups with different prognosis, and serve as an independent prognostic marker.更多还原