NOVEL BIOMARKERS FOR AROMATASE INHIBITOR THERAPY

描述（由申请人提供）：雌激素受体阳性（ER+）乳腺癌对内分泌治疗表现出高度可变的反应性。我们无法准确预测疾病的临床病程，导致与高发病率和医疗成本相关的低效临床算法。对ER+乳腺癌基因组的分析揭示了广泛的体细胞突变和基因拷贝数变化，这些突变驱动基因表达和功能异常，这些异常是临床观察到的多种表型的基础。为了加快分子分类的创建，从而转化为临床有用的预后和预测测试，我们进行了一系列新辅助内分泌治疗试验。十多年前设计的初步研究数据表明，在治疗的最初几周到几个月内分析细胞对内分泌治疗的反应可用于确定ER+乳腺癌在辅助内分泌治疗下的长期预后。特别是，“治疗”样本的增殖（Ki67）和ER状态的测量提供了独立于病理分期和分级的预后信息，并形成了术前内分泌预后指数（PEPI）的基础。在本R01的初始阶段，我们通过对临床2期和3期ER+乳腺癌（POL研究）术前来曲唑2期试验的肿瘤样本进行分析，扩展了这一开创性工作，并确定了治疗前和治疗后预测对芳香酶抑制剂（AI）治疗反应的基因表达特征。为了验证POL试验中发现的特征，并深化我们的生物标志物发现工作，我们提出了一项基于ACOSOG Z1031研究的新研究计划，这是一项在新辅助治疗环境中对三种已批准的ai进行随机比较的研究，该研究将于2009年中期完成375例患者的累积。在Aim 1中，我们建议验证两种“治疗中”预后特征，基于ER和Ki67的简单PEPI方法和50基因qRT-PCR内在亚型测定（PAM50），该方法已经被证明在POL研究中应用于来曲唑暴露肿瘤时也能预测不良反应和复发。为了解决从实际角度来看基线表达特征仍然更可取的批评，在Aim 2中，我们将使用mRNA基因表达谱来进一步发现和验证预测人工智能治疗反应性差和复发风险升高的预处理生物标志物。在目标3和目标4中，我们将通过分析与AI耐药生物学标记相关的基因拷贝畸变和突变来确定内分泌治疗耐药的基本分子基础。在资助的最后一年，当无复发生存数据可用时，我们将重新分析整个数据集，并创建一个最终模型，该模型可能使用所有三种数据类型（表达，基因复制和突变）来最佳地预测结果。该应用程序将提供一套福尔马林固定组织耐受性生物标志物方法，可以在随机辅助治疗试验的背景下进一步验证。我们的长期目标是生成一个用于临床结果注释的腔内乳腺癌图谱，以便对ER+乳腺癌的治疗可以有信心地进行个性化定制，并加速新的靶向治疗的出现。公共卫生相关性：雌激素受体阳性（ER+）乳腺癌的主要治疗方法是内分泌治疗，但肿瘤反应变化很大。我们无法可靠地预测肿瘤对药物的敏感性，往往导致额外的治疗，特别是化疗，以试图确保最好的结果，即使这些昂贵且耐受性差的药物对许多有反应性疾病的患者来说是不必要的。该应用程序将提供一套临床试验，不仅可以预测对内分泌治疗的反应，还可以确定内分泌治疗耐药肿瘤的全部分子缺陷，从而设计出新的更有效的靶向治疗方法。

DESCRIPTION (provided by applicant): Estrogen receptor positive (ER+) breast cancers exhibit highly variable responsiveness to endocrine therapy. Our inability to predict accurately the clinical course of the disease leads to inefficient clinical algorithms associated with high morbidity and health care costs. An analysis of the ER+ breast cancer genome reveals a wide range of somatic mutations and gene copy number changes that drive abnormalities in gene expression and function that underlie the diverse phenotypes observed clinically. To speed the creation of molecular classifications that can be translated into clinically useful prognostic and predictive tests we have executed a series of neoadjuvant endocrine therapy trials. The data from the initial studies designed over a decade ago demonstrated that an analysis of the cellular response to endocrine therapy within the first few weeks to months of treatment can be used to determine the long term prognosis of ER+ breast cancer under treatment with adjuvant endocrine therapy. In particular, measurements of proliferation (Ki67) and ER status of "on treatment" samples provides prognostic information that is independent of pathological stage and grade and forms the basis for a preoperative endocrine prognostic index (PEPI). In the initial phase of this R01, we extended this pioneering work through an analysis of tumor samples from a Phase 2 trial of preoperative letrozole for clinical stage 2 and 3 ER+ breast cancer (POL study) and identified both pre-treatment and on- treatment gene expression signatures that predict response to aromatase inhibitor (AI) treatment. To validate the signatures discovered in the POL trial and to deepen our biomarker discovery efforts we present a new research plan based on the ACOSOG Z1031 study, a randomized comparison between the three approved AIs in the neoadjuvant setting that is due to complete accrual of 375 patients by mid-2009. In Aim 1 we propose to validate two "on-treatment" prognostic signatures, the simple PEPI approach based on ER and Ki67 and a 50 gene qRT-PCR intrinsic subtype assay (PAM50) that has already been shown to also predict poor response and relapse when applied to letrozole exposed tumors in the POL study. To address the criticism that a baseline expression signature would still be preferable from a practical standpoint, in Aim 2 we will use mRNA gene expression profiling to further discover and validate pretreatment biomarkers that predict poor responsiveness to AI therapy and elevated relapse risk. In Aims 3 and 4 we will determine the fundamental molecular basis for endocrine therapy resistance by analyzing gene copy aberrations and mutations associated with biological markers of AI resistance. In the last year of the grant we will reanalyze the entire data set when relapse-free survival data becomes available and create a final model that potentially uses all three data types (expression, gene copy and mutation) to optimally predict outcomes. This application will deliver a suite of formalin-fixed tissue-tolerant biomarker approaches that can be further validated in the context of randomized adjuvant therapy trials. Our long term goal is to generate a luminal breast cancer atlas that is annotated for clinical outcomes so that the treatment of ER+ breast cancer can be individually tailored with confidence and the emergence of new targeted treatments can be accelerated. PUBLIC HEALTH RELEVANCE: The principle treatment for estrogen receptor positive (ER+) breast cancer is endocrine therapy but tumor response is highly variable. Our inability to reliably predict tumor sensitivity to often leads to additional treatments, particularly chemotherapy to try and ensure the best outcome, even though these costly and poorly tolerated drugs are not necessary for many patients with responsive disease. This application will deliver a suite of clinical tests that will not only predict response to endocrine therapy but will define the full repertoire of molecular defects in endocrine therapy resistant tumors so new and more effective targeted therapies can be designed.