Improved Diagnosis and Treatment of Cushing's Disease

散发性垂体腺瘤是最常见的人类病理之一，发现在超过10%的一般人群。垂体腺瘤是突变温和的，绝大多数腺瘤的致病机制仍然未知。 我的研究小组专注于库欣病（CD）作为一个模型，以了解所有散发性垂体腺瘤的致病机制。由于ACTH对肾上腺的对数放大效应，在诊断时，CD腺瘤通常很小（<1cm，微腺瘤），最小程度地扭曲垂体结构，并对激素轴产生可预测的影响（生化测试证实CD的存在和缓解）。基因组测序发现USP 8（30%）、USP 48和BRAF基因（共40%）中存在复发性体细胞突变。外遗传失调被假设为其余的基础-大多数垂体腺瘤。 目前的研究试图通过比较腺瘤之间或与同种异体尸检来源的垂体腺的转录组来研究散发性腺瘤的表观基因组机制。活体人类脑垂体的转录组尚不清楚，因此无法确定脑垂体病理学中的差异转录组通路。目前还没有动物模型、患者来源的异种移植模型或细胞系来研究散发性垂体腺瘤的发病机制。我们假设正常人脑垂体转录组将帮助我们发现以前未知的发病机制，散发性腺瘤的基础。 我们发现，虽然促肾上腺皮质激素瘤是良性肿瘤，但它们与恶性肿瘤一样经历代谢重编程。我们发现代谢重编程是通过同工酶转换HK-1到HK-2，乳酸脱氢酶A（LDH-A）到LDH-B和6-磷酸果糖-2-激酶/果糖-2，6-二磷酸酶-3（PFKFB 3）的核靶向介导的。我们发现人促肾上腺皮质激素瘤过度表达葡萄糖转运蛋白1（GLUT 1），从而增加葡萄糖的摄取。然后，我们证明了GLUT 1的表达可以通过促肾上腺皮质激素释放激素（CRH）等促分泌素的刺激进行转录调节。我们已经将这些发现转化为改善促肾上腺皮质激素瘤的FDG-PET检测（葡萄糖摄取的标志物）（NIH方案12-N-0007）。5我们发现CRH刺激导致腺瘤中平均FDG摄取增加。通常，这些肿瘤在MRI成像上是不可见的。通过CRH刺激，盲态神经放射科医生能够在PET成像上检测到40%的MRI不可见垂体腺瘤。基于这些数据，我们现在开始了一项临床试验，以研究DDAVP刺激的PET成像在检测MRI阴性垂体腺瘤引起CD中的作用。 我们发现促肾上腺皮质激素瘤的生存依赖于葡萄糖和糖酵解途径。然后，我们发现，通过使用特异性靶向肿瘤中GLUT 1表达的药物（如组蛋白脱乙酰酶抑制剂-伏立诺他），我们能够降低离体人促肾上腺皮质激素瘤的存活率和激素分泌。正常情况下，促肾上腺皮质激素的分泌是通过视黄酸受体（RXR）和肝X受体（LXR）的异源二聚体激活POMC基因来调节的。我们发现伏立诺他在体外小鼠肿瘤细胞中选择性下调LXR转录，但在正常促肾上腺皮质激素细胞中不下调。基于这些发现，我们现在已经启动了一项FDA批准的CD患者口服伏立诺他的临床试验，以测试其在手术前使激素水平正常化的功效。 使用NIH独有的CD患者的大型临床数据集，我们开发了一个关键的见解，即术后状态代表内源性压力测试。这种见解现在允许临床医生仅从一个术后血清激素数据点预测CD手术后的激素缓解。我们现在已经完成了一项试验，在该试验中，我们测量了患者从手术中醒来（从麻醉中苏醒）时的应激激素水平。我们发现，这些数据点在检测我们的患者是否需要压力激素替代方面非常敏感。 毫米级垂体腺瘤的成功术前成像可改善CD的手术结局。我们正在推进成像，以帮助检测这些腺瘤。我们已经开始并正在继续进行一项首次人体临床试验，以利用垂体腺瘤手术期间使用的腔内MRI线圈。本试验的初步数据分析表明，该线圈可以检测到先前MRI阴性的垂体腺瘤。

Sporadic pituitary adenomas are one of the most common human pathologies, found in more than 10% of the general population. Pituitary adenomas are mutationally bland and the pathogenic mechanisms underlying a vast majority of adenomas remains unknown. My group focuses on Cushings disease (CD) as a model to understand the pathogenic mechanisms underlying all sporadic pituitary adenomas. Due to the log amplification effect of ACTH on adrenal glands, when diagnosed, CD adenomas are often small (<1cm, microadenomas), minimally distort the pituitary architecture and have a predictable effect of the hormonal axes (biochemical testing confirms the presence of and remission from CD). Genomic sequencing discovered recurrent somatic mutations in the USP8 (30%), USP48 and BRAF genes (total 40%). Epigenetic dysregulation is hypothesized to underlie the rest - a majority of pituitary adenomas. Current studies attempt to study epigenomic mechanisms underlying sporadic adenomas by comparing the transcriptome of adenomas with each other or with allogenic autopsy derived pituitary glands. The transcriptome of intra-vital human pituitary gland is unknown, precluding identification of differential transcriptomic pathways in pituitary pathology. There exist no animal models, patient derived xenograft models or cell lines to investigate the pathogenesis of sporadic pituitary adenomas. We hypothesized that the normal human pituitary gland transcriptome will help us discover previously unknown pathogenic mechanisms that underlie sporadic adenomas. We discovered that although corticotropinomas are benign tumors, they undergo metabolic reprogramming much like malignant cancers. We found that metabolic reprogramming is mediated via isozyme switching of Hexokinase-1 (HK-1) to HK-2, lactate dehydrogenase A (LDH-A) to LDH-B and by nuclear targeting of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase-3 (PFKFB3). We found that human corticotropinomas overexpress glucose transporter 1 (GLUT1) that allows increased uptake of glucose. We then demonstrated that GLUT1 expression can be transcriptionally modulated by stimulation with secretagogues such as corticotropin releasing hormone (CRH). We have translated these findings to improve FDG-PET detection (a marker of glucose uptake) of corticotropinomas (NIH Protocol 12-N-0007).5 We found that CRH stimulation led to increased mean FDG uptake in adenomas. Frequently, these tumors are invisible on MRI imaging. With CRH stimulation, blinded neuroradiologists were able to detect 40% of MRI invisible pituitary adenomas on PET imaging. Based on these data, we are now initiating a clinical trial to investigate the role of DDAVP stimulated PET imaging on detecting MRI negative pituitary adenomas that cause CD. We showed that corticotropinomas rely on glucose and the glycolytic pathway for survival. We then found that by using drugs that specifically target GLUT1 expression in tumors (such as a histone deacetylase inhibitor- vorinostat), we were able to decrease survival and hormone secretion in human corticotropinomas ex-vivo. Normally, ACTH secretion is modulated by promoter activation of the POMC gene by a heterodimer of retinoic receptor (RXR) and liver X receptor (LXR). We found vorinostat transcriptionally downregulated LXR selectively in murine tumor cells in-vitro but not in normal corticotrophs. Based on these findings, we have now initiated an FDA approved clinical trial of oral vorinostat in patients with CD to test its efficacy in normalizing hormone levels pre-operatively. Using the large clinical dataset of CD patients that is uniquely available at NIH, we developed a critical insight that the post-operative state represents an endogenous stress test. This insight now allows clinicians to predict hormonal remission after surgery for CD from just one post-operative serum hormone data point. We have now completed a trial where we measured stress hormone levels as the patients were waking up from surgery (emergence from anesthesia). We found that these data points were exquisitely sensitive in detecting the need for stress hormone replacement in our patients. Successful pre-operative imaging of millimeter sized pituitary adenomas can lead to improved surgical outcomes in CD. We are advancing imaging to help detect these adenomas. We have initiated and are currently continuing a first-in-human clinical trial to utilize an intracavitary MRI coil that is used during surgery for pituitary adenomas. Preliminary data analysis from this trial suggests that this coil can detect previously MRI negative pituitary adenomas.