【作者】 张巨波；

【导师】 汤钊猷；

【作者基本信息】 复旦大学 ， 外科学， 2007， 博士

【摘要】 肝癌（主要是肝细胞癌，hepatocellular carcinoma，HCC）是全球第三位癌症杀手，在我国更高居第二位。手术切除仍为肝癌最有效的疗法，但即使根治性切除，5年内仍有60％～70％的转移复发率，成为进一步提高疗效的瓶颈。对部分不能切除肝癌，局部治疗是一种选择，但同样遇到高转移复发率。肝癌是多血管的实体瘤，为此，抗血管生成是一条有希望的途径。我所王鲁等发现干扰素α（IFNα）通过抑制血管生成而减少肝癌的转移复发，其机理是下调血管内皮细胞生长因子（VEGF）。临床随机对照研究亦证实IFN-α可推迟肝癌术后的复发，成为预防肝癌术后复发的常规之一。我们的临床随机分组研究发现肝癌切除术后给予IFNα治疗可延长患者的生存，推迟复发；但与对照组相比复发率并未降低（用药期间生存率提高了，停药一个阶段后生存率才降低到对照组水平）。因此我们推测IFNα在治疗早期可抑制复发病灶生长，但在后期，肝癌可能对IFNα治疗产生耐药性，因此肿瘤重获生长。如能弄清肝癌对IFNα治疗耐药的机制，将有助于设计新的治疗方案，克服肝癌对IFNα的耐药，以提高IFNα治疗肝癌的疗效。本课题是在此基础上，通过我所建立的高转移人肝癌原位移植裸鼠模型LCI-D20，模拟临床治疗模式，探讨IFNα治疗一段时间停药后是否能够继续使用，再次应用是否能够达到和第一次同样的治疗效果（目的在于回答临床工作中产生的问题，即是否可以对肝癌术后应用IFNα治疗18个月停药后复发病人再次使用IFNα治疗），通过对IFNα第一、二次治疗前、后血管生成相关基因表达谱异同的分析和验证，分析复发肝癌对再次应用IFNα治疗所产生逃逸的分子机制，试图找到克服肝癌对治疗逃逸的方法。1．IFNα停药后再治疗不能有效地抑制肝癌生长和降低微血管密度MVD根据我所前期的实验结果，IFNα在剂量为1.5×10.⁷ .U／kg／d时能有效地抑制高转移人肝癌原位移植裸鼠模型LCI-D20的生长。本次实验中这一剂量在最初治疗的20天中，IFNα治疗组和对照组在肿瘤大小（0.27±0.19 g versus0.68±0.24 g）、微血管密度[22／HP（95％CI=15-29）versus 46／HP（95％CI=32-60）]和血清VEGF浓度（24.5±8.7 pg／ml versus 41.6±12.0 pg／ml）等方面均有显著差异，P＜0.05。停药20天后，肿瘤大小、微血管密度和血清VEGF浓度在两组间的差别已没有统计学意义（P＞0.05）。按照实验设计再次给予IFNα治疗（1.5×10.⁷ .U／kg／d）20天，与对照组相比，再次治疗组的血清VEGF浓度明显减少（58.0±13.7 pg／ml versus 92.8±24.3 pg／ml，P=0.012），但肿瘤大小（3.22±0.62g，3.81±0.92 g）和微血管密度[61／HP（95％CI=48-74），70／HP（95％CI=59-81）]与第一次治疗不同，没有明显的减少（P＞0.05）。2．肝癌前负荷在一定范围内对IFNα的治疗没有明显影响考虑到IFNα第一次和第二次治疗时肝癌的重量不同，为了证实给药前肝癌的重量是否对IFNα治疗产生影响，一组高转移人肝癌原位移植裸鼠模型LCI-D20在无任何干预的情况下生长30天，此时肿瘤大小与IFNα第二次治疗时相近（约1.5g），给予IFNα治疗（1.5×10.⁷ .U／kg／d）20天，与对照组相比，肿瘤重量（2.28±0.63 g versus 3.90±0.80 g）、微血管密度[39／HP（95％CI=30-48）versus 67／HP（95％CI=45-89）]和血清VEGF水平（44.9±10.6 pg／ml versus76.1±15.6 pg／ml）均显著降低（P＜0.05）。3．IFNα治疗不同时间点血管生成相关基因表达差异的研究利用Super Array血管生成相关基因芯片，我们对IFNα第一次治疗前、后和第二次治疗前、后血管生成相关基因的表达谱进行了分析。与治疗前相比，IFNα第一次治疗后有多个促血管生成因子表达下调，如VEGF、angiogenin、bFGF、PDGF-A、TGF、TNF、G-CSF、EGF、IGF、IL-8。IFNα第二次治疗后，除PDGF-A基因表达上调外，其他几个促血管生成因子表达再次下调。4．Real-time PER和Western blot分析VEGF165和PDGF-A在IFNα第一、二次治疗前、后的表达。考虑到血管周细胞对血管内皮细胞具有重要的支持和信号传导作用，而PDGF／PDGF受体在周细胞的信号传导中具有重要意义，因此将下一步实验研究的重点放到了VEGF165和PDGF-A基因和蛋白水平的检测上。通过探针法real-timePCR和Western blot检测发现，VEGF165在第一次和第二次IFNα治疗过程中表达均下调，并且伴有VEGF受体2（Flk-1）磷酸化水平的降低。这些结果表明VEGF／VEGF受体信号传导通路在前后两次治疗中均被有效地抑制。PDGF-A在第一次治疗中表达下调，在第二次治疗中表达反而上调。5．IFNα第二次治疗联合PDGF受体抑制剂imatinib显著降低肿瘤的大小和微血管密度。为了验证PDGF-A上调是IFNα第二次治疗效果差的原因，在IFNα第二次治疗过程中联合应用了PDGF／PDGF受体抑制剂imatinib（100mg／kg／d）治疗20天。结果表明联合用药组的肿瘤大小（2.23±0.43 g）、微血管密度[42／HP （95％CI=32-52）]，与IFNα单纯再次治疗组[3.22±0.62 g，61／HP （95％CI=48-77）]相比均显著降低，更显著低于对照组[3.81±0.92 g，70／HP （95％CI=59-81）]。6．缺氧诱导因子（HIF）1α、2α在各组中和不同细胞系缺氧状态中的表达肿瘤的非限制性增生导致瘤体内缺氧，促使HIF-1α、HIF-2α表达上调。其中HIF-1α是VEGF高表达的主要转录因子。Western blot检测结果显示，IFN-α第一次治疗后HIF-1α、HIF-2α的表达水平均显著降低，停药20天后HIF-1α、HIF-2α的表达水平则明显升高。与对照组相比，HIF-1α表达水平在IFNα再次治疗时仍可以被有效地下调；与第一次IFNα治疗不同，HIF-2α的表达水平在IFNα再次治疗时不再被下调。在正常氧浓度培养条件下，Hela、MHCC-97L、MHCC-97H细胞均只表达较低的HIF1α、HIF2α。当在低氧浓度（1％O₂）下培养5h后，HIF1α、HIF2α表达在3种细胞系中均显著升高。7．超声造影能够显示抗血管治疗对肝癌血流灌注的影响。治疗组裸鼠在给药前、给药的第3天、8天、18天、32天分别进行微泡超声造影，利用软件给出造影强度随时间变化的造影时间强度曲线（TIC曲线），通过公式求出主要参数：产生显影的时间（AT）、达到峰值强度的时间（TTP）及峰值强度（PI）、曲线下面积（AUC）等数据。在拟合回归分析曲线中给出基础强度BI，缩放因子α1，下降曲率α2，阶段曲率α3。通过统计分析发现，在对照组随着肿瘤的增大，AUC有由低到高、再逐渐降低的趋势，而VEGF受体抑制剂Pazopanib治疗组则逐渐降低，给药8天后两组相比有统计学差异。AT在对照组随着时间的延长急剧降低，在治疗组降低的速度明显缓慢。统计分析显示，AUC与MVD正相关，AT与MVD负相关，PI与MVD没有相关性。结论1．IFNα第二次应用治疗肝癌时不能有效地抑制肿瘤的生长和降低微血管密度，肝癌对IFNα再次治疗产生逃逸。2．IFNα治疗高转移裸鼠人肝癌的疗效不受初始治疗时肿瘤负荷影响。3．促血管生长因子PDGF-A在IFNα再次治疗高转移裸鼠人肝癌时高表达，与IFNα再次治疗抑瘤效果差有显著的相关性，此时肝癌血管生成由VEGF依赖转化为PDGF-A依赖。IFNα再次治疗高转移裸鼠人肝癌时联合应用PDGF受体抑制剂imatinib可显著提高治疗效果。4．超声造影技术能够较好地反映肿瘤生长和治疗过程中血流灌注的变化，时间强度曲线下面积（AUC）和造影剂开始显示的时间（AT）分别与微血管密度正相关和负相关，两者有可能作为抗血管生成治疗疗效检测的指标。应用价值1．IFNα先后两次治疗高转移裸鼠人肝癌前后不同时间点血管生成相关基因表达谱发生改变，有利于对抗肝癌血管生成治疗耐药机制的深入研究。2．IFNα再次治疗高转移裸鼠人肝癌时联合PDGF受体抑制剂imatinib可显著提高治疗效果，为临床合理选择抗血管生成药物的使用提供了线索。3．超声造影具有较强的可操作性和较好的依从性，有助于动态观察抗血管治疗的效果，使及时发现耐药性和选择新的抗血管生成药物成为可能。创新点1．首次发现并提出促血管生长因子PDGF-A在IFNα再次治疗高转移裸鼠人肝癌时高表达，血管生成由VEGF依赖转化为PDGF-A依赖，与肝癌对IFNα再次治疗产生逃逸密切相关。2．首次提出IFNα再次治疗肝癌时联合PDGF受体抑制剂imatinib可显著提高治疗效果。3．首次应用超声造影动态检测抗肝癌血管治疗的效果，并提出AUC与MVD正相关，AT与MVD负相关。更多还原

【Abstract】 Hepatocellular carcinoma (HCC), one of the most common malignanciesworldwide and the second cancer killer in China, is a hypervascular tumor, andabundant and tortuous blood vessels distinguish HCC from benign lesions inangiography. The first clues to the endogenous angiogenesis inhibitors came with theobservations that interferon. (IFN-α) may inhibit endothelial cell chemotaxis andproliferation, and IFN-. has been proven to effectively inhibit HCC growth andrecurrence, especially when combined with cytotoxic drugs. Although severalsignaling pathways are involved in the angiogenesis, attention has been focused onthe vascular endothelial growth factor (VEGF) family, especially VEGF-A. Thecritical involvement of VEGF-A in vascular development and angiogenesis is to bindand activate of VEGF receptor-2, which stimulates endothelial migration andproliferation. IFN-. inhibits tumor cells production of VEGF and endothelial cellmotility, and demostrates direct and indirect anti-angiogenic activity. Nowadays,IFN-. is applied in the treatment of HCC, renal cancer, neuroendocrine tumors,prostate cancer, malignant melanoma, chronic myelogenous leukemia, and superficialbladder carcinoma,We previously reported that IFN-. inhibited tumor growth and recurrence innude mice bearing human HCC xenografts, the same is true of the nude mice withhuman HCC after curative resection, mediated by anti-angiogenic effect throughdown-regulation of VEGF but not basic fibroblast growth factor (bFGF). In theLCI-D20 model, anti-angiogenic activity was a more important factor than theanti-proliferative effect of p48-ISGF, protein deficiency, which plays a pivotal role inthe IFN-αassociated signal transduction pathway. Despite the conspicuous decreasein tumor burden and vasculature induced by the treatment with IFN-., little is knownabout the reversibility of these changes after IFN-. therapy is discontinued.A randomized control trial showed that 18 months’ post-operative IFN-.treatment delayed recurrence and prolonged patients’ survival. Unfortunately, tumorsrecurred in a number of patients shortly after IFN-αtreatment was discontinued. Thespeed of vascular regrowth in HCC after IFN-. therapy is discontinued is clinicallyrelevant and may influence the use of IFN-.. Based on the results of our clinical trial and pilot study, this experimental design investigates whether relapse is inevitable,and whether restarted IFN-. treatment is effective. The results demonstrate that withvascular regrowth after IFN-. therapy is discontinued, PDGF-A expression increasesdramatically when IFN-αtreatment is restarted. There is no obvious decrease intumor burden and microvessel density (MVD). Pericytes are another vascular celltype that provides endothelial cells with crucial survival signals, and PDGF/PDGFRpathway regulated pericyte homeostasis, so we postulate that up-regulation ofPDGF-A is responsible for the failure of restarted IFN-αtreatment for HCC. To testthis hypothesis, we added imatinib (Gleevec, Novartis, Basel, Switzerland), thePDGF-receptor inhibitor, to the restarted IFN-αtreatment course, and resulted insignificant inhibition of tumor growth and angiogenesis. Overall, the results indicatethat the up-regulation of PDGF-A was responsible for the failure in the restartedIFN-αtreatment when tumors recurred after the first IFN-αtherapy was discontinued,and long-term disease control can be best achieved by combination therapy withimatinib in an IFN-αrestarted treatment course.1. Restarted IFN-αtreatment was less effective in suppressing tumor growth than the first treatment courseWhen IFN-αtreatment group was compared with the control group, a significantdecrease of tumor weight was observed (0.27±0.19 g versus 0.68±0.24g), MVD wasalso markedly decreased [22/HP(95% CI=15-29) versus 46/HP (95% CI=32-60)].Furthermore, ELISA results showed that the serum VEGF levels were lower in theIFN-αtreatment group than in control group A (24.5±8.7 pg/ml versus 41.6±12.0pg/ml), P＜0.05. When IFN-αtreatment was discontinued for 20 days, there was nosignificant difference in the tumor weight between the follow -up of control group andfollow-up of IFN-αtreatment group. Furthermore, the restarted IFN-αtreatment didnot result in significant inhibitory effect on tumor growth and MVD when comparewith control group [restrated group F: 3.22±0.62 g, 61/HP (95% CI=48-74), controlgroup E: 3.81±0.92 g, 70/HP (95% CI=59-81)], P＞0.05.2. Response of IFN-αtreatment was not influenced by pretreatment tumor weight.The poorer tumor response to the restarted treatment of IFN-αmay have resultedfrom a larger tumor load. To test this possibility, tumors were allowed to grow for 35days when the tumor weight was about 1.5 g (the same size as the tumor before the restarted IFN-αtreatment). IFN-αtreatment for 20 days resulted in a significantreduction of tumor weight compared with control group G, being 2.28±0.63 g versus3.90±0.80 g (P=0.003).3. Changes in the angiogenesis-related gene expression profile of human HCC in nude mice treated by IFN-α.In two independent cDNA microarray analyses, the results showed that IFN-αtreatment down-regulated several pro-angiogenic factors, including VEGF,angiogenin, bFGF, PDGF-A, TGF, TNF, G-CSF, EGF, IGF, and IL-8 in the firsttreatment course. All the pro-angiogenic factors except PDGF-A were againdown-regulated in restarted IFN-αtreatment group compared with control group.4. Effects of IFN-αon the expression of VEGF165, PDGF-A and VEGF receptor 2.RT-PCR with primers showed two transcripts whose base pair lengthcorresponded to VEGF165 and VEGF121, and a minor transcript corresponding toVEGF189. However, Western blotting results showed no immunoreactive proteinsother than VEGF165. These data indicated that IFN-αdown-regulated the expressionof VEGF165.To verify the results from the cDNA microarray analyses, we used RT-PCR andWestern blot assays to study the expression of VEGF165 and PDGF-A. The results ofRT-PCR showed that IFN-αdecreased the VEGF165 expression in the first (IFN-αtreatment group) and restarted IFN-αtreatment course (restarted IFN-αtreatmentgroup) compared with their co ntrol groups, respectively, which corresponds with thechange in serum VEGF concentration.Western blot assays for VEGF165 of HCC tissues under reducing conditionsshowed two bands of 23 and 21 kD, respectively, which is consistent with twoglycosylation variants of VEGF165, as reported previously. No other isoforms ofVEGF were detected. Tumors from IFN-αtreatment group expressed a lower level ofVEGF165 than those of control group in the first treatment course. Similarly, in thesecond treatment course, the VEGF165 concentration of restarted IFN-αtreatmentgroup was lower than that of restarted control group. Thus, VEGF165 expression inthe first and second treatment courses was down-regulated by IFN-αtreatment.The results of RT-PCR show that PDGF-A expression diminished in tumorstreated with IFN-αcompared with the control group in the first treatment course, andincreased in the restarted IFN-αtreatment group compared with its control group. These results are consistent with the cDNA microarray analysis. Western blot assaysshowed that PDGF-A in IFN-αtreatment group was down-regulated significantlycompared with its control group. In contrast, restarted IFN-αtreatment group had ahigher PDGF-A expression level than its control group in the second treatment courseTo assess the inhibiting effect of IFN-αon the VEGFR2 signaling, thephosphorylation status of VEGFR2 was determined using antibodies that recognize aphosphorylation site in the VEGFR2 kinase domain of tyrosine 951 (Y951), whichplays a critical role in pathological angiogenesis. Western blot assays results showedthat VEGFR2 phosphorylation of Y951 was detected at a high level in control groups,but at a very low level in IFN-αtreatment group and restarted IFN-αtreatment group.Thus, both the VEGF165 level and VEGFR2 phosphorylation demonstrated that theIFN-αretains its ability to inhibit VEGF signaling in the first and second treatmentcourses.5. Blockage of PDGF signaling by imatinib significantly improved the efficacy of restarted IFN-αtreatment.Combination treatment with imatinib (PDGF receptor inhibitor) and IFN-αresulted in a significant decrease in tumor burden in the second treatment coursecompared to IFN-αtreatment alone (P=0.01) and imatinib treatment alone (P=0.018), which was accompanied by a decrease in angiogenesis estimated by MVD(P=0.007 for IFN-αtreatment alone; P=0.004 for imatinib treatment alone).6. The high expression of transcription factor HIF-2αin the restarted IFN-αtreatment may trigger the high expression of PDGF-A.The HeLa and MHCC07-H、MHCC07-L cells exposed to hypoxic conditionsexpressed high level of hypoxia inducible factor-1αand 2αcompared with cells in thenormoxic conditions. HIF-1αand HIF-2αare both undergo rapid hypoxia-inducedprotein stabilization and bind identical target DNA sequences. In the first IFN-αtreatment course, the expression of HIF-1αand HIF-2αin the IFN-αtreatment groupB is lower than in the control group A; in the restarted IFN-αtreatment group F, theexpression of HIF-1αwas down regulated again compared with control group E,while HIF-2αmaintained high level in the restarted IFN-αtreatment group F.7. Correlation of contrast agent flow and immunohistochemical results.Contrast-enhanced ultrasonography were performed at day 9, 12, 17, 27, and 42.For the entire cohort (treated and control tumors), AUC and AT were positively andnegatively correlated with MVD and the tumor volume and weight. Conclusions1. Restarted IFN-αtreatment was less effective in suppressing tumor growththan the first treatment course by evasion of antiangiogenic targeting of VEGFsignaling.2. Response of IFN-αtreatment was not influenced by pretreatment tumorweight.3. Up regulation of platelet -derived growth factor-Ais responsible for the failureof restarted interferonαtreatment in hepatocellular carcinoma, and blockage ofPDGF signaling by imatinib significantly improved the efficacy of restarted IFN-αtreatment.4. Quantification of intratumor flow of ultrasonographic contrast agent atgray-scele imaging shows promise for monitoring tumor vascular response toanti-angiogenic therapy.The potential application of this work1. The differential expressed genes, in the first and restrated IFN-αtreatment,identified by angiogenesis associated cDNA microarray analysis provide a valuableresource for basic and clinical studies of angiogenesis in the study of HCC metastasis.2. Restarted IFN-αtreatment was not effective because of up-regulation ofPDGF-A and the VEGF signal pathway remained inhibited. The combination ofPDGF-A inhibitor improved the restarted IFN-αtreatment, and may be a tractableclinical strategy for treating recurrent HCC.3. Quantification of intratumoral flow of ultrasonographic contrast agent atgray-scale imaging shows promise for monitoring tumor vascular response toanti-angiogenic therapy.Originalities of this work1. Analysed and confirmed, for the first time, that restarted IFN-αtreatment isnot effective in inhibiting HCC growth because of up-regulation of PDGF-A, and thecombination of PDGF receptor inhibitor imatinib improved the restarted IFN-αtreatment.2. Established the platform of using contrast-enhanced ultrasonography toassess the response of anti-angiogenic agents on the HCC in the LCI-D20 model, andfound the AUC and AT were positively and negatively correlated with current histologic indices for quantifying angiogenesis MVD.更多还原