背景

卵巢癌（OC）是女性生殖系统中死亡率最高的恶性肿瘤^[1]，在理想肿瘤细胞减灭术的基础上辅助铂类为基础的化疗是目前规范化的治疗策略。但是，晚期卵巢癌的生存率仅有29%^[2]。肿瘤的反复复发导致的其对于铂类药物产生耐药性是卵巢癌生存率得不到进一步改善的主要原因之一^[3]。卵巢上皮性癌（EOC）是一组具有多种组织学类型的高度异质性肿瘤^[4]，占卵巢恶性肿瘤的95%，在组织学类型上主要分为浆液性癌、粘液性癌、透明细胞癌、子宫内膜样癌等^[4]。其中高级别浆液性癌（HGSOC）约占60%-70%，侵袭性强，疾病进展迅速，是一类最能体现卵巢癌的生物学特性和临床特征的组织学类型，从而成为有关卵巢癌基础与临床研究的核心类型^[5]。

自1951年人宫颈癌细胞系Hela^[6]建立的60余年来，细胞培养及建系已成为肿瘤研究的重要手段，其使用极大地促进了人类对肿瘤生物学的理解^[7]。细胞系的优势是可以提供一个相对同质性的细胞群，且能无限制复制及传代。体外细胞系和细胞系的异种移植模型也成为研究卵巢癌的重要方法^[8]。2011年，采用二代测序的方法，The Cancer Genome Atlas (TCGA)中描述了各种肿瘤的基因组和表达图谱^[9]。2012年，The Broad-Novartis Cancer Cell Line Encyclopedia (CCLE)详细的描述了约1000个肿瘤细胞系的起源及基因图谱^[7]。在此基础上，Domcke等分析了47种卵巢癌细胞系及卵巢癌肿瘤组织的DNA拷贝数改变（CNAs）、突变类型和mRNA表达水平，并将重点放在HGSOC上，寻找最适合作为体外模型的卵巢癌细胞系^[10]。目前在体外实验中最常采用的HGSOC细胞系主要包括SKOV3、A2780、IGROV1、OVCAR3、CAOV3等。其中，SKOV3和A2780为p53野生型，均具有KRAS基因突变，表明这两株可能为低级别的卵巢上皮癌。IGROV1细胞系被发现为MLH1、MSH3和MSH6基因突变，是Lynch综合征中常见的突变基因，其起源及基因组均具有子宫内膜样癌的特征。OVCAR3和CAOV3都有p53突变和较高的CNAs值，但其CNAs和TCGA中HGSOC组织的相关性较低。Domcke等推荐了一些能更好的模拟HGSOC基因组特征、但在既往研究中较少用到的细胞系，包括KURAMOCHI和OVCAR4等。其中OVCAR4具有p53突变，CNAs与HGSOC的组织相关性强，在关键癌基因和肿瘤抑制因子中具有低的突变频率，被认为能较好的代表HGSOC细胞系。该研究改变了人们对HGSOC细胞系的认知，并且得到了之后多项研究的验证和支持^[11-14]。

随着肿瘤研究的进展，肿瘤干细胞（CSCs）理论于上世纪70年代左右被提出^[15]。CSCs被认为是存在于肿瘤细胞中的一小群具有干细胞自我更新及分化能力的细胞，在肿瘤中长期处于静息状态，可逃避化疗药物的作用，介导肿瘤的耐药及复发^[16]]。近年来针对CSCs表面特异性标志蛋白的单克隆抗体成为肿瘤治疗领域的热点^[17]。研究发现，通过单克隆抗体靶向杀伤CSCs，再配合手术、放化疗等常规抗肿瘤治疗，可达到治愈肿瘤的目的。卵巢上皮性癌最突出的特点为复发及耐药，这高度提示肿瘤中可能存在CSCs细胞^[[18]，因此，如何通过细胞表面特异性标志蛋白识别其中的干细胞进而靶向杀灭，成为卵巢上皮性癌治疗的关键。

本课题组长期致力于卵巢癌干细胞表面标志蛋白的研究，先后两次受国家自然科学基金资助，通过定量蛋白组学技术对卵巢上皮癌的侧群细胞（SP）和非SP细胞膜蛋白进行比较蛋白组学分析，从而筛选出卵巢癌干细胞表面标志物候选蛋白—CDC50A，进一步通过体内体外实验充分证实了CDC50A可作为卵巢癌干细胞表面标志蛋白。本课题组在前期体外研究中以文献中最常用的SKOV3细胞系为主，现为进一步研究CDC50A在HGSOC中的表达及干性特征，本论文选取了具有HGSOC基因组特征的细胞系OVCAR4为研究对象，比较其与SKOV3细胞系在生物学特性上的差异；通过活体成像技术监测OVCAR4细胞系NSG小鼠腹腔及卵巢原位成瘤过程，建立理想的模拟HGSOC腹腔内播散种植转移的动物模型；进一步验证OVCAR4细胞系中CDC50A+细胞的干细胞特性，并在OVCAR4和SKOV3两种细胞系中初步探讨CDC50A作为卵巢上皮性癌干细胞治疗靶点的可行性。

方法

OVCAR4细胞系和SKOV3细胞系的生物学特性比较。采用瑞氏吉姆萨染色比较OVCAR4和SKOV3细胞系形态；七天法细胞计数绘制两种细胞系的生长曲线；流式细胞检测法检测两种细胞系的细胞生长周期分布；CCK-8法比较两种细胞系的顺铂半数抑制率（IC50）；免疫组化方式检测两种细胞系p53表达情况；建立OVCAR4及SKOV3细胞系裸鼠皮下移植瘤模型，比较两种细胞系皮下致瘤能力，免疫组化法观察移植瘤石蜡切片p53抗体染色情况。

卵巢癌细胞系OVCAR4小鼠腹腔及卵巢原位成瘤模型的建立及动态观察。慢病毒转染技术建立OVCAR4-Luciferase细胞系，嘌呤霉素筛选稳定表达Luciferase的OVCAR4细胞，监测其细胞形态与群体倍增时间等均与OVCAR4细胞系无差异；取4～6周龄雌性NSG小鼠，OVCAR4-Luciferase细胞1×10⁷/100μl/只腹腔注射法注入小鼠腹腔内，共3只；于小鼠背部解剖右卵巢，OVCAR4-Luciferase细胞5×10⁶/40μl/只显微注射法注入右卵巢，共5只；定期麻醉小鼠，腹腔注射底物D-Luciferin，通过活体成像仪检测荧光信号，在活体状态下连续观察成瘤情况，适时解剖小鼠，监测成瘤情况。

CDC50A作为上皮性卵巢癌干细胞标记蛋白的体外验证。流式细胞仪分析OVCAR4细胞系贴壁生长细胞中CDC50A+细胞的比例；流式细胞仪分选OVCAR4细胞系中CDC50A+及CDC50A-细胞，qPCR法比较两种细胞CDC50A基因的表达水平；流式细胞仪分选OVCAR4细胞系中CDC50A+及CDC50A-细胞，无血清悬浮培养法比较两种细胞sphere形成能力；OVCAR4细胞系采用无血清悬浮培养的方法，形成sphere，流式细胞学方法检测OVCAR4 sphere对CDC50A的富集作用；流式细胞仪分别分选SKOV3及OVCAR4细胞系CDC50A+细胞，于悬浮培养基中悬浮培养，定期于细胞培养基中加入CDC50A抗体、同型对照Ig-G、PBS，同时设空白组，比较各组细胞sphere形成情况；流式细胞仪分选OVCAR4细胞系CDC50A+及CDC50A-细胞，不同浓度的顺铂处理后，CCK-8法比较不同细胞在OD450nm处的吸光度，比较CDC50A+及CDC50A-的顺铂耐药性。

结果

OVCAR4细胞在光镜下呈多边形状，核仁明显，细胞异型性明显，小团细胞聚集呈簇状生长，并与SKOV3细胞形态形成鲜明对比，SKOV3细胞呈梭形，树突状凸起，细胞大多散在分布，细胞核大、深染，胞质较均匀；SKOV3细胞系的群体倍增时间（PDT）比OVCAR4细胞系明显缩短，细胞生长快，同样时间下细胞生长数量多于OVCAR4；OVCAR4细胞系中G0/G1期所占比例较SKOV3细胞系稍低，但OVCAR4细胞系S期细胞所占比例明显高于SKOV3细胞系；OVCAR4细胞系的顺铂IC50明显低于SKOV3细胞系，OVCAR4对顺铂的反应更敏感；p53抗体在OVCAR4细胞系细胞核染色强阳性，而SKOV3细胞系染色阴性；两组细胞系裸鼠皮下移植瘤模型均可见移植瘤形成，解剖裸鼠，肿瘤均限于接种部位，SKOV3组移植瘤重量明显高于OVCAR4组，差异有统计学意义（P<0.05），OVCAR4皮下移植瘤组织p53染色强阳性，而SKOV3细胞系皮下移植瘤组织p53染色阴性。

建立了稳定的OVCAR4-Luciferase细胞系，顺利的构建了OVCAR4细胞系NSG小鼠腹腔及卵巢原位移植瘤模型，10天后均可于活体成像仪下检测到荧光信号，定期监测肿瘤信号增长良好，适时牺牲小鼠解剖后并未见到腹腔内有大块样肿瘤形成，腹腔组3只小鼠及卵巢原位组5只小鼠腹腔内均可见少许腹水，3/3只小鼠腹腔网膜及肠系膜根部可以见到粟粒样的肿瘤组织形成，3/5只小鼠右侧卵巢形成移植瘤，均为单侧，5/5只左侧卵巢外观均正常，网膜及肠系膜根部肉眼未见到肿瘤组织形成。

流式细胞仪检测发现OVCAR4细胞系中CDC50A+细胞群所占比例为1.03%，符合干细胞比例要求；qPCR结果显示流式细胞仪分选的OVCAR-CDC50A+细胞中CDC50A基因的表达水平明显高于分选的OVCAR-CDC50A-细胞，进一步验证了CDC50A抗体的特异性和分选样本的正确性；OVCAR-CDC50A+细胞sphere形成能力明显高于OVCAR-CDC50A-，差异有统计学意义（P<0.01）；无血清悬浮培养出OVCAR4 sphere，流式分析验证CDC50A+细胞在OVCAR4 sphere中所占的比例显著高于贴壁OVCAR4细胞；悬浮培养基中培养的流式分选出的OVCAR4-CDC50A+细胞及SKOV3-CDC50A+细胞中，CDC50A抗体组中sphere形成数量明显低于Ig-G组、PBS组及空白组，差异有统计学意义（P<0.01），后三组间sphere形成数量无明显差异；经3、6、9ug/ml不同浓度的顺铂处理后，6、9ug/ml 顺铂作用下的OVCAR4-CDC50A-细胞活性显著低于OVCAR4-CDC50A+细胞（P<0.05）。

结论

OVCAR4细胞系具有卵巢高级别浆液癌生物学特征。从细胞的生物学特征上看，OVCAR4细胞系的细胞异型性明显，S期细胞所占比例明显高于SKOV3细胞系，对顺铂治疗的反应更敏感；p53抗体在细胞核染色强阳性，具备HGSOC的生物学特征，能很好地进行体外细胞实验。OVCAR4细胞系可以在裸鼠皮下形成移植成瘤，移植瘤石蜡切片p53抗体免疫组化染色强阳性。SKOV3细胞系确实不表达p53蛋白，其可能并不属于HGSOC细胞系。

成功构建了OVCAR4细胞系NSG小鼠腹腔及卵巢原位成瘤模型，在活体成像系统的动态监测下观察到肿瘤细胞的生长。OVCAR4细胞系腹腔内成瘤形态具有卵巢癌腹腔内播散的生物学特征。

OVCAR4细胞系中CDC50A+细胞具有肿瘤干细胞特征。CDC50A+细胞在OVCAR4细胞系中所占比例符合干细胞比例，OVCAR-CDC50A+细胞中CDC50A基因的表达水平明显高于分选的OVCAR-CDC50A-细胞。OVCAR4-CDC50A+细胞sphere形成能力明显高于OVCAR-CDC50A-细胞；OVCAR4-sphere中CDC50A+细胞比例明显高于贴壁细胞，sphere具有富集CDC50A+细胞的作用；抗体封闭CDC50A+细胞后，sphere形成能力明显下降；不同浓度的顺铂处理后，OVCAR4-CDC50A-细胞抗顺铂治疗能力显著低于OVCAR4-CDC50A+细胞。深入探讨CDC50A作为卵巢癌对铂类耐药的潜在治疗靶点具有积极意义。

Background

Ovarian cancer (OC) is the most lethal gynecological genital tract cancer, the standard treatment for advanced OC is primary debulking surgery followed by the platinum-based chemotherapy. However, the survival rate of advanced ovarian cancer is only 29%. The drug resistance to platinum drugs caused by tumor recurrence is one of the main reasons that the survival rate of ovarian cancer cannot be further improved. Epithelial ovarian cancer (EOC) is a group of highly heterogeneous tumors with a variety of histological types, accounting for 95% of ovarian malignant tumors. In terms of histological types, EOC is mainly divided into serous cancer, mucinous cancer, clear cell cancer, endometrioid cancer and so on . Among them, high-grade serous ovarian cancer (HGSOC) accounts for about 60% - 70%. It is a kind of histologic type that can best reflect the biological and clinical characteristics of ovarian cancer with strong invasion and rapid progress of disease, and thus becomes the core type of basic and clinical research on ovarian cancer.

Since the establishment of HeLa cell line in 1951, cell culture and establishment of tumor cell line has become an important means of tumor research, and its use has greatly promoted human understanding of tumor biology. The advantage of cell line is that it can provide a relatively homogeneous cell group, and can reproduce and propagate without restriction. In vitro cell line and cell line xenotransplantation model become an important method to study ovarian cancer. In 2011, the Cancer Genome Atlas (TCGA) used a second-generation sequencing method to describe the genome and expression map of various tumors. In 2012, the broad Novartis cancer cell line Encyclopedia (CCLE) detailed the origin and gene map of about 1000 tumor cell lines. On this basis, Domke et al. analyzed DNA copy number changes (CNAs), mutation types and mRNA expression levels of 47 ovarian cancer cell lines and ovarian cancer tissues, and focused on HGSOC to find the most suitable ovarian cancer cells as in vitro models. At present, the most commonly used HGSOC cell lines in vitro include SKOV3, A2780, igrov1, OVCAR3, and CAOV3. SKOV3 and A2780 are wild-type p53 with KRAS gene mutations, indicating that these two strains may be low-grade epithelial ovarian cancer. IGROV1 cell line was found to be MLH1, MSH3 and MSH6 gene mutations. It is a common mutation gene in Lynch syndrome. Its origin and genome have the characteristics of endometrioid carcinoma. Both OVCAR3 and CAOV3 have p53 mutation and high CNAs value, but the correlation between CNAs and HGSOC in TCGA is low. Domke et al. recommended some cell lines that can better simulate HGSOC genomic characteristics, but are rarely used in previous studies, including KURAMOCHI and OVCAR4. Among them, OVCAR4 has p53 mutation, CNAs has strong correlation with HGSOC, and has low mutation frequency in key oncogenes and tumor suppressors, which is considered to be a good representative of HGSOC cell line. This study has changed people's cognition of HGSOC cell line, and has been verified and supported by many later studies.

With the development of tumor research, the theory of cancer stem cells (CSCs) was put forward around 1970s. CSCs are considered to be a small group of stem cells with the ability of self-renewal and differentiation in tumor cells, which can in a resting state for a long time in the tumor, can escape the role of chemotherapy drugs, and mediate the drug resistance and recurrence of the tumor. In recent years, monoclonal antibodies against CSCs surface specific marker proteins have become a hot topic in the field of tumor therapy. It was found that monoclonal antibody targeting to kill CSCs, combined with conventional anti-tumor therapy such as surgery, radiotherapy and chemotherapy, can achieve the purpose of curing tumor. The most prominent feature of epithelial ovarian cancer is recurrence and drug resistance, which highly suggests that there may be CSCs cells in the tumor. Therefore, how to identify the stem cells through cell surface specific marker protein and then target to kill them becomes the key to the treatment of epithelial ovarian cancer.

Our research group has been devoted to the study of the surface marker proteins of ovarian cancer stem cells for a long time, and has been supported by the National Natural Science Foundation twice. Through the quantitative proteomics technology, comparative proteomics analysis was carried out on the membrane proteins of side population (SP) and non SP cells of ovarian epithelial cancer, and the candidate protein CDC50A for the surface marker of ovarian cancer stem cells was screened out, further through in vivo and in vitro research we confirmed that CDC50A could be used as the surface marker protein of ovarian cancer stem cells. In the previous in vitro study, our research group mainly focused on SKOV3 cell line, which is the most commonly used in the literature. Now, in order to further study the expression and characteristics of CDC50A in HGSOC, we selected OVCAR4 cell line with HGSOC genomic characteristics as the research object, and compared its biological characteristics with SKOV3 cell line. The intraperitoneal and ovary in situ tumorigenesis model of OVCAR4 cell line in NSG mice was constructed and the growth of tumor cells was observed under the dynamic monitoring of the in vivo imaging system, to further verify the stem cell characteristics of CDC50A positive cells in ovcar4 cell line, and to preliminarily explore the feasibility of CDC50A as the treatment target of ovarian epithelial cancer stem cells in OVCAR4 and SKOV3 cell lines.

Methods

1. Comparison of biological characteristics between ovcar4 cell line and SKOV3 cell line. The comparison of cell morphology of OVCAR4 andSKOV3 was performed by Giemsa staining; The cell growth curve of those two cell lines was sketching by cell counting in 7 days; Cell cycles of those cell lines were analyzed by flow cytometry; The half maximal inhibitory concentration of the two cell lines was compared with cck-8 method; The expression of p53 in the two cell lines was detected by immunohistochemistry; The subcutaneous tumor models of OVCAR4 and SKOV3 cell lines were established to compare the subcutaneous tumorigenic ability of the two cell lines.

2. The establishment and dynamic observation of ovarian cancer cell line OVCAR4 mouse model of intraperitoneal and ovarian tumorigenesis in situ. OVCAR4-luciferase cell line was established by lentivirus transfection technology; Puromycin was used to screen OVCAR4 cell line stably expressing luciferase, and its cell morphology and population doubling time were no different from OVCAR4 cell line; OVCAR4-luciferase cells(1×107 /100 μ L) were injected into the abdominal cavity of 4-6 week old female NSG mice by intraperitoneal injection, 5 mice in total; The right ovary was dissected in the back of mice, and OVCAR4-luciferase cells (5×106 / 40 μ L / mouse) were injected into the right ovary by microinjection, 5 mice in total; Mice were anesthetized regularly, and D-luciferin was injected intraperitoneally. The fluorescence signal was detected by the living imaging instrument, and the tumorigenesis was continuously observed in the living state, and the mice were dissected timely to monitor the tumorigenesis.

3. In vitro validation of CDC50A as a surface marker protein of epithelial ovarian cancer stem cells. The proportion of CDC50A positive cells in the ovarian cancer cell lines OVCAR4 were analyzed by flow cytometry; OVCAR4 CDC50A positive and CDC50A negative cells were sorted by Moflo fluorescent activated cell stream sorter and the expression of CDC50A gene was tested by qPCR; The sphere formation ability of the OVCAR4-CDC50A positive and CDC50A negative cells were tested by serum-free suspension culture technology; The enrichment of CDC50A positive cells in OVCAR4 spheres were tested by immunofluorescence and flow cytometry; OVCAR4 CDC50A positive and  SKOV3 CDC50A positive cells were sorted by Moflo fluorescent activated cell stream sorter and cultured in serum-free suspension culture technology，both kinds of cells were divided into antibody group, Ig-G control group, PBS group and blank group. CDC50A antibody, Ig-G and PBS were added to cell culture medium regularly to compare the formation of cell sphere; The Cisplatin resistance ability of the OVCAR4-CDC50A positive and CDC50A negative cells were tested by CCK-8 method.

Results

1.Under light microscope OVCAR4 cells were in multilateral shape, with obvious nucleoli, obvious cell atypia, small clusters of cells growing in clusters, which were in sharp contrast with the morphology of SKOV3 cells. SKOV3 cells were spindle shaped, dendritic raised, mostly scattered and distributed, with large and deeply stained nuclei and relatively uniform cytoplasm; The population doubling time (PDT) of cell line SKOV3 was shortest for 32.54 hours, while OVCAR4 was longest for 49.2 hours; The proportion of G0 / G1 phase cells in OVCAR4 cell line was slightly lower than that in SKOV3 cell line, but the proportion of S phase cells in OVCAR4 cell line was significantly higher than that in SKOV3 cell line; The IC50 of cisplatin in OVCAR4 cell line was significantly lower than that in SKOV3 cell line, and OVCAR4 was more sensitive to cisplatin; The p53 antibody was strongly positive in OVCAR4 cell line, but negative in SKOV3 cell line; The formation of transplanted tumor was observed in both cell lines of nude mice, the weight of transplanted tumor in SKOV3 group was significantly higher than that in OVCAR4 group (P < 0.05); p53 staining was strongly positive in OVCAR4 and negative in SKOV3 cell line.

2.A stable OVCAR4-luciferase cell line was established, the intraperitoneal and ovary in situ tumorigenesis model of OVCAR4 cell line in NSG mice was successfully constructed. After 10 days, fluorescence signals could be detected under the in vivo imaging system and the growth of tumor signals was good after regular monitoring. No massive tumor formation was found in the abdominal cavity after sacrifice of mouse anatomy. A little ascites was found in the abdominal cavity of 3 mice in the a intraperitoneal model and 5 mice in the ovary in situ model, military tumor tissue was found in the omentum and mesentery root of 3 / 3 mice, transplanted tumor was found in the right ovary of 3 / 5 mice, which were unilateral, normal appearance was found in the left ovary of 5 / 5 mice, no tumor tissue was found in the omentum and mesentery.

3. CDC50A positive population were exist in OVCAR4 and the proportion of CDC50A positive cells were 1.03%; The results of qPCR showed that the expression of CDC50A gene in OVCAR4-CDC50A positive cells was significantly higher than that in OVCAR4-CDC50A negative cells, which further verified the specificity of CDC50A antibody and the correctness of sample selection; The sphere formation ability of OVCAR4-CDC50A positive cells was significantly higher than that of OVCAR4-CDC50A negative cells, the difference was statistically significant (P < 0.01); The proportion of OVCAR4-CDC50A positive cells in OVCAR4 sphere was significantly higher than that of adherent OVCAR4 cells; In SKOV3- CDC50A positive cells and OVCAR4-CDC50A positive cells cultured in suspension medium, the number of sphere formation in CDC50A antibody group was significantly lower than that in ig-g group, PBS group and blank group, the difference was statistically significant (P < 0.01); The activity of OVCAR4-CDC50A negative cells treated with cisplatin was significantly lower than that of OVCAR4-CDC50A positive cells (P < 0.05).

Conclusions

1. OVCAR4 cell line has the biological characteristics of HGSOC. From the perspective of cell biological characteristics, OVCAR4 cell line has obvious cell atypia, the proportion of S phase cells is significantly higher than that of SKOV3 cell line, and is more sensitive to cisplatin treatment; p53 antibody is strongly positive in nuclear staining, with the biological characteristics of HGSOC, which can be used for cell experiments in vitro. OVCAR4 cell line can be transplanted subcutaneously in nude mice, and p53 antibody in paraffin section of transplanted tumor is strongly positive. SKOV3 cell line does not express p53 protein, which may not belong to HGSOC cell line.

2.The intraperitoneal and ovary in situ tumorigenesis model of OVCAR4 cell line in NSG mice was successfully constructed and the growth of tumor cells was observed under the dynamic monitoring of the in vivo imaging system. OVCAR4 grew well after intraperitoneal injection in NSG mice and recapitulated military disease.

3.CDC50A could be considered as a surface marker of ovarian cancer stem cells. The proportion of CDC50A positive cells in OVCAR4 cell line is in line with that of stem cells. The expression level of CDC50A gene in OVCAR4-CDC50A positive cells was significantly higher than that in OVCAR4-CDC50A negative cells. The sphere forming ability of OVCAR4-CDC50A positive cells is significantly higher than that of OVCAR4-CDC50A negative cells; The proportion of CDC50A positive cells in OVCAR4 sphere is significantly higher than that of adherent cells, and sphere has the function of enriching CDC50A positive cells; The sphere forming ability decreases significantly after the antibody blocked CDC50A positive cells; the anti-cisplatin treatment ability of OVCAR4-CDC50A negative cells is significantly lower than that of OVCAR4-CDC50A positive cells after different concentrations of cisplatin treatment.