第一部分：梗阻性黄疸患者减黄前后CA19-9变化趋势研究

研究背景

      糖类抗原19-9（CA19-9）是胰腺疾病诊疗过程中应用最广泛的肿瘤标志物之一，准确的CA19-9基线水平对指导胰腺疾病良恶性鉴别、胰腺癌可切除性评估、治疗反应评价及预后预测意义重大。然而，临床实践中观察到，合并梗阻性黄疸的胰腺疾病患者CA19-9水平大多显著升高，使CA19-9测定值不能真实地反映肿瘤分泌的CA19-9真实水平，进而影响疾病诊断及对胰腺癌患者的临床决策。因此，迫切需要制定良恶性梗阻性黄疸患者CA19-9的校正公式，以指导临床实践、避免临床决策偏差。

研究目的

       探究影响CA19-9测定值的因素，推导出在减黄前预测CA19-9真实值的校正公式，为梗阻性黄疸患者治疗方案的选择提供帮助。

研究方法

      纳入2014年1月至2019年1月于北京协和医院行减黄操作且具有明确临床或病理诊断、合并梗阻性黄疸且在减黄操作前后1个月内均测定过CA19-9水平的患者，根据电子病历系统收集患者人口学资料、影像学资料、血清学检验结果、病理结果、手术操作日期及类型、诊断等资料。利用Mann-Whitney U检验比较组间差异，利用Spearman秩和检验评价CA19-9与其他指标相关性，利用二元线性回归探究CA19-9校正公式。

研究结果

    本研究共纳入121例患者（恶性102例、良性19例），恶性组患者CA19-9（p=0.048）、TBil（p=0.048）及中性粒细胞百分比（p=0.013）水平显著高于良性组，红细胞计数（p=0.047）显著低于良性组。恶性患者中，CA19-9水平与TBil无明确相关性，与AST（p＜0.0001）、GGT（p=0.029）及中性粒细胞百分比（p=0.013）具有较强相关性。24.5%的恶性患者在减黄后CA19-9水平无明显下降趋势甚至持续升高，89.5%的良性患者在减黄后CA19-9水平呈明显下降趋势。良恶性患者减黄前后TBil平均下降率无明显差异，恶性患者CA19-9水平平均下降率显著低于良性患者（p=0.014）。通过筛选拟合效果最好的自变量，恶性患者在减黄前预测CA19-9真实值的校正公式为：CA19-9_true = 0.63 × CA19-9_measured - 20.3，良性患者在减黄前预测CA19-9真实值的校正公式为：CA19-9_true = 0.085 × CA19-9_measured + 60.4。在良恶性疾病鉴别中，减黄后CA19-9水平（AUC=0.699）、校正公式所得CA19-9预测真实值（AUC=0.695）诊断效能显著高于减黄前CA19-9水平（AUC=0.598）。

研究结论

    良恶性梗阻性黄疸患者在CA19-9等指标基线水平上存在显著差异，二者减黄前后TBil平均下降率相近，但恶性患者CA19-9水平平均下降率显著低于良性患者。我们成功得到了以减黄前CA19-9水平为自变量预测CA19-9真实值的校正公式。根据我们的CA19-9校正公式，可以更好的对良恶性梗阻性黄疸患者进行鉴别，且该公式在胰腺癌可切除性评估、治疗反应评价及预后预测等方面具有良好的应用前景。

第二部分：RNA结合蛋白PUM2调控胰腺癌化疗耐药的功能及机制研究

研究背景

      胰腺癌起病隐匿、恶性程度高、预后极差。化疗是胰腺癌治疗的重要组成部分，以吉西他滨（GEM）为主的联合化疗是胰腺癌的一线化疗方案之一，但GEM耐药一直是导致胰腺癌复发及患者死亡的瓶颈问题，且目前与GEM的联合化疗或联合靶向治疗方案并不能解决GEM耐药问题。因此，从新的分子生物学角度探究胰腺癌GEM耐药机制以增敏GEM疗效、逆转耐药，对于提高胰腺癌患者总体生存率意义重大。

      RNA结合蛋白（RBP）是一类可以调节RNA转运、剪接、稳定性及翻译的重要蛋白质，RBP的异常表达常导致下游一系列RNA的异常累积或降解、进而导致疾病发生。例如，RNA结合蛋白HuR在胰腺癌中高表达，其可以通过调控下游WEE1、PIM1、IDH1等基因mRNA代谢过程，进而促进胰腺癌细胞增殖能力、侵袭转移能力及对缺氧低糖等应激的耐受性。然而，RBP在胰腺癌GEM耐药中是否发挥功能、何种RBP在胰腺癌GEM耐药中发挥功能以及RBP在胰腺癌GEM耐药中发挥何种功能均缺乏系统性研究。因此，探索在胰腺癌GEM耐药中发挥重要功能的RBP及其具体分子机制，对于进一步深入理解和解决胰腺癌GEM耐药意义重大。

研究目的

      筛选与胰腺癌GEM耐药相关的RBP及其与临床病理特征的相关性，明确RNA结合蛋白PUM2对胰腺癌增殖、迁移及GEM耐药等恶性生物学行为的影响，探究PUM2影响胰腺癌GEM耐药的具体机制。

研究方法

      综合胰腺癌GEM耐药细胞株转录组测序结果、siRNA文库增殖及GEM耐药实验结果，筛选与胰腺癌GEM耐药密切相关的RBP，利用胰腺癌组织芯片PUM2免疫组化染色评价PUM2表达情况与临床病理特征关系。通过敲低及过表达PUM2，利用SRB增殖实验、GEM细胞毒实验及Transwell细胞迁移实验检测体外状态下PUM2对胰腺癌细胞恶性生物学行为的影响，利用小鼠皮下移植瘤模型检测体内状态下PUM2对胰腺癌细胞恶性生物学行为的影响。进而，联合RIP-seq和RNA-seq寻找PUM2在胰腺癌细胞中调控的下游mRNA，利用qRT-PCR、Western Blot、RIP-qPCR、放线菌素D RNA稳定性实验、双荧光素酶基因报告实验验证PUM2对下游mRNA的调控作用，并利用下游基因对PUM2进行功能拯救实验进一步确证PUM2对下游基因的调控作用。最后，结合RIP-seq、RNA-seq和JSAPAR数据库，筛选和PUM2具有相互调控作用的转录因子，利用qRT-PCR、Western Blot、RIP-qPCR及功能拯救实验对二者调控关系进行验证。

研究结果

      基于胰腺癌GEM耐药株及亲本株转录组测序结果，我们发现多种RBP在胰腺癌耐药株中高表达，我们通过siRNA文库及GEM耐药实验筛选出与胰腺癌GEM耐药最相关的RNA结合蛋白——PUM2，胰腺癌组织芯片免疫组化染色提示PUM2高表达是胰腺癌患者不良预后的独立危险因素。体外实验表明PUM2可促进胰腺癌细胞的增殖、迁移及对GEM的抵抗，体内实验表明，敲低PUM2，小鼠皮下移植瘤生长受到明显抑制且对GEM敏感性增加。进一步地，联合RNA-seq和RIP-seq探究PUM2发挥促进胰腺癌GEM耐药过程中所调控的下游RNA，我们发现PUM2可以上调黏着斑相关基因集中的数个基因（ITGA3、ADAM17、ASAP1等）mRNA的稳定性。ITGA3是上述结果中受PUM2调控最显著的下游基因，PUM2可以通过与ITGA3 mRNA 3'UTR区中的PUM结合元件（PBE）结合而稳定ITGA3 mRNA，体外拯救实验证实了ITGA3是PUM2发挥调控胰腺癌GEM药功能的下游分子。最后，我们发现转录因子EGR1与PUM2存在相互调控作用，PUM2可以结合EGR1 mRNA 3'UTR区、EGR1可以结合PUM2基因的启动子区，从而造成级联效应以放大PUM2在胰腺癌耐药中发挥的功能。

研究结论

      RNA结合蛋白PUM2与胰腺癌患者预后密切相关，其通过调控黏着斑相关基因集中ITGA3等基因mRNA稳定性发挥促进胰腺癌GEM耐药功能，且其与转录因子EGR1存在正反馈调控作用。EGR1/PUM2/ITGA3轴的发现对未来胰腺癌患者化疗方案的筛选、探索逆转GEM耐药联合用药方案提供了扎实的实验基础。

Part I. Dynamic change of serum CA19–9 levels in benign and malignant patients with obstructive jaundice after biliary drainage

Background

    CA19-9 is one of the most widely used tumor markers in the diagnosis and treatment of pancreatic diseases. Accurate CA19-9 baseline value is of great significance for guiding differentiation of benign and malignant pancreatic diseases, tevaluation of resectability of pancreatic cancer, evaluation of treatment response and prognosis prediction. However, it has been observed in clinical practice that the level of CA19-9 in patients with pancreatic disease complicated with obstructive jaundice is significantly increased, so that the measured value of CA19-9 cannot truly reflect the true level of CA19-9 secreted by tumor, thus affecting the diagnosis of disease and clinical decision-making in patients with pancreatic cancer. Therefore, it is urgent to develop the correction formulas of CA19-9 in patients with benign and malignant obstructive jaundice to guide clinical practice and avoid improper clinical decision-making.

Objective

    Explore influencing factors of CA19-9 measured value. Derive correction formulas of CA19-9 real value before the biliary decompression in order to help surgeons make proper clinical decision for obstructive jaundice patients.

Methods

    Obstructive jaundice patients who underwent biliary decompression between January 2014 and January 2019 in Peking Union Medical College Hospital, with specific clinical or pathological diagnosis as well as CA19-9 value measured within one month before and after biliary decompression, were brought into our study. Demographic data, imaging data, serological test results, pathological results, operation data, diagnosis and other data were collected according to the electronic medical record system. Mann-whitney U test was used to compare the differences between groups. Spearman rank sum test was used to evaluate the correlation between CA19-9 and other indicators. Binary linear regression was used to explore the correction formulas of CA19-9.

Results

    Our study brought into 121 patients (102 malignant patients and 19 benign patients). CA19-9 value (p = 0.048), TBil value (p = 0.048) and neutrophil percentage (p = 0.013) in malignant group were significantly higher than in benign group, and in contrary, RBC count (p =0.047) in malignant group was significantly lower than that in benign group. In malignant patients, CA19-9 value has no definite correction with TBil value, but was strongly correlated with AST (p < 0.0001), GGT (p =0.029) and neutrophil percentage (p =0.013). CA19-9 value of 24.5% malignant patients did not decline significantly or even increased after biliary decompression, while CA19-9 value of 89.5% benign patients decline significantly. There was no significant difference in the average decline rate of TBil in benign and malignant patients, and the average decline rate of CA19-9 value in malignant patients was significantly lower than that in benign patients (p =0.014). By screening independent variables with the best imitative effect, the correction formula for predicting the true CA19-9 value in malignant patients before biliary decompression was: CA19-9_true = 0.63 × CA19-9_measured - 20.3; the correction formula for predicting the true CA19-9 value in benign patients before biliary decompression was: CA19-9_true = 0.085 × CA19-9_measured + 60.4. In regard to differential diagnosis of benign and malignant diseases, the diagnostic efficacy of CA19-9 value after biliary decompression (AUC=0.699) and CA19-9 predicted true value obtained by correction formula (AUC=0.695) was significantly higher than that of CA19-9 value before biliary decompression (AUC=0.598).

Conclusions

    Benign and malignant patients with obstructive jaundice have significant differences in indicators such as CA19-9 in baseline. Average TBil decline rates after biliary decompression in two groups were similar, but average CA19-9 decline rate in malignant patients was significantly lower than that in begine patients. We successfully obtained the correction formulas for predicting CA19-9 true value using CA19-9 measured value before biliary decompression as independent variable. With help of our correction formulas, patients with benign and malignant obstructive jaundice can be better differentiated. Our correction formulas also have a good application prospect in evaluation for resectablility, treatment response and prognosis prediction of pancreatic cancer.

Part II. Function and mechanism study of RNA-binding protein PUM2 in regulating chemotherapy resistance of pancreatic cancer

Background

    Pancreatic cancer has insidious onset, high malignancy and poor prognosis. Chemotherapy is an important part of the treatment for pancreatic cancer, in which gemcitabine (GEM) based combined chemotherapy is one of the first-line chemotherapy regimen for pancreatic cancer. However, resistance for GEM has always been a bottleneck problem leading to recurrence and death of pancreatic cancer patients, and the current combined chemotherapy regimen or combined targeted therapy regimen with GEM cannot solve the problem of GEM resistance. Therefore, it is of great significance to explore the mechanism of GEM resistance from a new molecular biological perspective to enhance the efficacy of GEM sensitization and reverse drug resistance, so as to improve the overall survival of pancreatic cancer patients. 

    RNA binding protein (RBP) is a kind of important proteins that regulate transportation, splicing, stability and translation of RNA. Abnormal expression of RBP often leads to a series of abnormal accumulation or degradation of downstream RNA resulting in various diseases. For example, the RNA-binding protein HuR is highly expressed in pancreatic cancer, which can promote proliferation, invasion and metastasis of pancreatic cancer cells and the tolerance to hypoxia and hypoglycemia by regulating the metabolism of downstream RNAs, like WEE1, PIM1, IDH1 and other mRNAs. However, there is a lack of systematic study on whether RBP plays a role in GEM resistance of pancreatic cancer, which RBP plays a significant role in GEM resistance of pancreatic cancer, and what specific role RBP plays in GEM resistance of pancreatic cancer. Therefore, it is of great significance to explore RBPs and their specific molecular mechanisms that play an important role in GEM resistance of pancreatic cancer for further understanding and solving GEM resistance of pancreatic cancer.

Objective

    Screen RBP related to GEM resistance of pancreatic cancer and its correlation with clinicopathological indicators. Clarify the effect of RNA-binding protein PUM2 on the malignant biological behaviors of pancreatic cancer, such as proliferation, migration and GEM resistance. Explore the specific mechanism of PUM2 regulating GEM resistance of pancreatic cancer.

Methods

    RBPs closely related to GEM resistance of pancreatic cancer were screened based on transcriptome sequencing, siRNA library proliferation and GEM resistance test results. Relationship between expression level of PUM2 and clinicopathological indicators was evaluated by immunohistochemical staining of tissue chip. SRB proliferation assay, GEM drug resistance assay and transwell cell migration assay were used to detect the effects of PUM2 on the malignant biological behaviors of pancreatic cancer cells in vitro by knocking down or overexpressing PUM2.The effects in vivowere explored by mice subcutaneous xenograft model. Furthermore, RIP-seq and RNA-seq were combined to explore the downstream mRNAs regulated by PUM2 in pancreatic cancer cells, and the regulation of PUM2 on downstream mRNAs was verified by qRT-PCR, Western Blot, RIP-qPCR, actinomycin D RNA stability assay, dual luciferase gene reporter assay and rescue experiments. Finally, combined with RIP-seq, RNA-seq and JSAPAR database, transcription factors with mutual regulation relationship with PUM2 were screened, and the regulatory relationship between the transcriptor factor and PUM2 was verified by qRT-PCR, Western Blot, RIP-qPCR, ChIPqPCR and f rescue experiments.

Results

    Based on transcriptome sequencing results of GEM resistant pancreatic cancer cell line and the parent cell line, several RBPS were highly expressed in GEM-resistant pancreatic cancer cell line. We screened out the RNA-binding protein PUM2, which is most related to gemcitabine resistance of pancreatic cancer, through siRNA library and GEM resistance assay. Immunohistochemical staining of pancreatic cancer tissue chip suggested that high expression of PUM2 was an independent risk factor for poor prognosis of pancreatic cancer patients. In vitro functional experiments showed that PUM2 could promote proliferation, migration and resistance to gemcitabine of pancreatic cancer cells. In vivo experiments showed that knockdown of PUM2 inhibited the growth of subcutaneous transplanted tumor in mice and increased sensitivity to gemcitabine. Further, RNA-seq and RIP-seq were combined to explore the regulation role of PUM2 on downstream RNAs that promote GEM resistance in pancreatic cancer. We found that PUM2 up-regulated mRNA stability of several genes (ITGA3, ADAM17, ASAP1, etc.) in the focal adhesion pathway. ITGA3 was confirmed to be the most significant downstream mRNA of PUM2 regulating gemcitabine resistance in pancreatic cancer by rescue experiments in vitro, and PUM2 could stabilize ITGA3 mRNA by binding to PUM binding element (PBE) in the 3'UTR region of ITGA3 mRNA. Finally, we found the mutual regulation relationship between transcription factor EGR1 and PUM2, that is PUM2 binding to 3'UTR region of EGR1 mRNA, and EGR1 binding to promoter region of PUM2 gene, resulting in a cascade effect that amplifying the role of PUM2 in pancreatic cancer drug resistance.

Conclusions

    RNA-binding protein PUM2 is closely related to the prognosis of pancreatic cancer patients. It promotes GEM resistance of pancreatic cancer by regulating mRNA stability of ITGA3 and other genes in focal adhesion pathway, and it has positive feedback regulation with transcription factor EGR1. The discovery of EGR1/PUM2/ITGA3 axis provides a solid experimental basis for the screening of chemotherapy regiments for pancreatic cancer patients in the future and the exploration of combination regimens reversing GEM resistance.