【作者】 张晓东；

【导师】 许乙凯；

【作者基本信息】 南方医科大学 ， 影像医学与核医学， 2007， 博士

【摘要】 研究目的1．制备由ASG受体介导的MR对比剂半乳糖基化白蛋白-SPIO（Galactose-Bovine serum albumin-SPIO，Gal--BSA-SPIO），通过兔正常肝脏体内外饱和实验测定其与ASG受体的结合活性及其分布；2．建立兔VX2肝癌模型，初步探讨ASG受体介导的Gal--BSA-SPIO在肝脏肿瘤检出与诊断方面的价值；3．测定Gal--BSA-SPIO与人肝脏ASG受体的结合活性及分布，初步探讨Gal--BSA-SPIO在人肝脏MR成像的潜在应用价值；4．对人肝细胞性结节的CT／MR强化特征与其组织病理及免疫组化特征进行对照分析，以更深刻认识肝细胞性结节的影像学特征，并初步探讨Gal--BSA-SPIO在肝细胞性结节的鉴别诊断中是否具有潜在的优越性?材料和方法1．Gal--BSA-SPIO的制备及体内外饱和实验（1）乳糖基白蛋白的制备及测定采用还原胺法，称量牛血白蛋白240mg，乳糖1.2mg，氰基硼氢化钠816mg，溶于30ml 0.2M PH值8.0的磷酸盐缓冲液，37℃水浴搅拌反应24h～120h：取反应液对蒸馏水低温透析三天，间换透析外液；取透析过的溶液低温离心，取上清液，过葡聚糖凝胶柱层析分离，纯化后溶液冰冻干燥即得产品，为白色固体粉末。苯酚—硫酸比色法测定产品中半乳糖的浓度，考马斯亮蓝G250法测定产品中蛋白的浓度，计算糖基化比率。（2） Gal--BSA-SPIO的制备取适量SPIO过Sepharose 4B层析柱，分离出小粒径SPIO，此溶液对PH值7.4标准缓冲液低温透析24h，浓缩至含Fe量8mg／ml，调PH值为6.5左右。取此溶液适量，加入等体积含1％W／V乳糖基白蛋白溶液，冰浴超声波振荡。未结合的糖蛋白用1M Nacl除去，调溶液PH值为7.4，测定含铁量。Malvern-3000HS激光粒度分析仪测定粒径及其分布，透射电镜测定其核心粒径及观察其形态。（3） Gal--BSA-SPIO体内外饱和实验及其分布体外饱和实验取10g新鲜兔肝组织，加入预冷的全细胞生长液，将组织切成1～2mm³的小块，过30目金属筛，用含0.3mg／ml胶原酶的全细胞生长液稀释，37℃孵育30min，冰浴超声波振荡降解，低温离心，取上清（富含细胞膜碎片）。（1）阻断组：将富含细胞膜碎片的溶液2ml用含0.2 mg或2mg的D（+）—半乳糖37℃孵育30min，然后用含10μmol Fe的SPIO或Gal--BSA-SPIO 37℃孵育30min，PBS、超高速离心洗涤2次，以除去未结合的Gal--BSA-SPIO和SPIO，最后用6ml全细胞生长液悬浮。（2）非阻断组：将富含细胞膜碎片的溶液2ml直接用含10μmol Fe的Gal--BSA-SPIO或SPIO 37℃孵育30min，洗涤2次，6ml全细胞生长液悬浮。各溶液用1.5T MR扫描、测量T2驰豫时间，统计学比较分析。体内饱和实验健康新西兰白兔6只，随机分为两组，阻断组和非阻断组，每组3只，均分别行MR平扫及增强扫描。平扫及增强所用序列：SE T2WI．TR600ms，TE 20ms、80ms，均作轴状位扫描。增强扫描：阻断组预先注入D（+）—半乳糖（2mg／kg）预饱和肝细胞膜上的ASG受体，30min后注入10μmolFe／kg Gal--BSA-SPIO，30min后行MR扫描。非阻断组直接注入10μmolFe／kgGal--BSA-SPIO，30min后行同上扫描。分别测量增强前及增强后实验动物肝脏T2值。分布实验取新西兰白兔2只，分别缓慢静注20μmolFe／kg SPIO及Gal--BSA-SPIO，30min后处死动物，快速取其肝脏及脾脏组织，同时做电镜标本（肝）和石蜡标本（肝、脾）处理。分别于光镜和透射电镜下观察其分布。2．Gal--BSA-SPIO在兔VX2肝癌模型的应用（1）兔VX2肝癌模型的建立：从VX₂荷瘤种兔肿瘤边缘切取生长活跃的鱼肉样组织放于少量生理盐水中，用眼科剪将其剪成约0.5～1mm³瘤块备用。取健康新西兰白兔麻醉，暴露腹腔、将肝脏一叶轻轻牵至体外，以眼科镊在较厚的位置轻轻刺破肝组织，将1粒VX2瘤小块埋入其中。每只兔种植1～3处肿瘤。（2）动物分组：肝VX2肿瘤新西兰白兔20只随机分为2组：SPIO组和Gal—BSA-SPIO组，各10只，每组再随机分为两个剂量组（Ⅰ和Ⅱ），组Ⅰ为5μmol Fe／kg；组Ⅱ为10μmol Fe／kg。（3） MR扫描：1.5T MR头部线圈进行平扫及增强扫描，层厚3-5mm，FOV：87.5×100；扫描序列及参数为：SE T₁ WI：TR300ms，TE20ms、3000ms；SE T₂WI：TR600ms，TE 20ms、80ms；TSE T₂Wh TR3500ms，TE15ms、105ms；GRE T₂^* WI：TR600ms，TE 15ms，18°翻转角度；轴状位、冠状位或矢状位扫描。增强：按实验分组分别经兔耳缘静脉缓慢注入对比剂，30min后行MR扫描，扫描条件同平扫一致。（4）图像观察与测量：观察分析实验动物肝脏VX2肿瘤成瘤情况及MR特征；测量增强前及增强后各组实验动物肿瘤、肝脏的T₁及T₂值；测量增强前后各组实验动物肝脏及肿瘤的信号强度，计算肝脏的信噪比（SNR）、肝脏信号强度下降百分比（PSIL）及肿瘤—肝脏对比噪声比（CNR）。（5）统计学比较分析：采用配对t检验比较各组增强前后肝脏T₁值、T₂值、SNR及CNR是否有显著性差异；采用单因素方差分析比较不同组间T₁值、T₂值、SNR和CNR增强前后差值及肝脏PSIL是否有显著性差异；采用单因素方差分析比较增强前及增强后不同扫描序列间肝脏SNR及CNR是否有显著性差异。多重比较用LSD或Dunnett’s T3，P＜0.05认为有显著性差异。3．ASG受体介导的Gal--BSA-SPIO人肝脏MR受体成像初步研究（1） Gal--BSA-SPIO与人肝脏ASG受体亲和力的初步测定收集外科手术新鲜肝癌标本6例，肝硬化标本4例，正常肝组织3例，按前述（兔肝体外饱和实验）方法制备细胞膜悬液，然后行MR扫描及T2值测定。采用单因素方差分析和独立样本t检验进行统计学比较分析。多重比较用LSD法或Dunnett’s T3。P＜0.05认为有显著性差异。（2）人正常肝脏及病变（新鲜标本）受体分布实验取出低温贮藏的人肝脏不同新鲜组织，冰冻切片（2～4μm），室温风干，用SPIO、Gal--BSA-SPIO及PBS 37℃孵育20s，洗涤3次，除去未结合的SPIO或Gal--BSA-SPIO，室温风干24h，Perl’s染色。显微镜下观察。（3）人肝不同组织（石蜡标本）受体分布实验收集肝脏石蜡标本27例，包括小肝癌12例，转移瘤3例、胆管细胞癌2例、肝硬化再生结节5例，肝细胞结节状增生3例、血管瘤3例、囊腺瘤1例。常规切片（2～4μm），65℃烤箱烤片60min，二甲苯脱蜡至水，室温风干，用SPIO及Gal--BSA-SPIO 37℃孵育10min，PBS洗涤3次，除去未结合的SPIO或ALC-BSA-SPIO，室温风干24h，Perl’s染色。显微镜下观察。（4） CD34免疫组织化学染色对18例肝细胞性结节（小肝癌10例，肝硬化结节5例及肝细胞结节状增生3例）进行CD34标记，染色方法采用链霉亲合素-生物素-过氧化物酶连结法（SABC法）；显微镜下观察，并按Weidner等方法计数10例肝细胞癌肿瘤微血管密度（MVD）。（5）观察18例肝细胞性结节的CT／MR强化特征，并与其组织病理、免疫组织化学进行对照分析。结果1．Gal--BSA-SPIO的制备及体内外饱和实验结果（1） Gal--BSA-SPIO的测定结果：24h、48h、72h、96h、120h，乳糖基白蛋白的糖基化比率分别为8、15、27、32、38。Gal--BSA-SPIO含铁量为2.8mg／ml；平均体积粒径为34.4nm，多分散性为0.35；核心粒径为14.8nm，呈类圆形，大小均匀，表面平滑完整，部分有重叠。（2） Gal--BSA-SPIO体外饱和实验结果：用造影剂孵育前，不同浓度阻断组和非阻断组T2值差异无显著性意义（P＞0.05）。细胞膜悬液用SPIO孵育后，不同浓度阻断组与非阻断组之间差异无显著性意义（P＞0.05）。细胞膜悬液用Gal--BSA-SPIO孵育，非阻断组孵育后T2值减低，与孵育前有显著性差异（P＜0.05），与SPIO孵育后亦有显著性差异（P＜0.05）；不同浓度的D（+）-半乳糖预先饱和ASG受体后，细胞膜悬液用Gal--BSA-SPIO孵育后T2值升高，与直接用Gal--BSA-SPIO孵育有显著性差异（P＜0.05），不同浓度的D（+）-半乳糖之间亦有显著性差异（P＜0.05）。（3） Gal--BSA-SPIO体内饱和实验结果：注射Gal--BSA-SPIO后肝脏T2驰豫时间明显缩短，MR信号明显减低，与周围噪声相当或略低，呈现“黑肝”效应；预先用D（+）-半乳糖预饱和肝ASG受体后，肝脏T2驰豫时间升高，但肝脏MR信号依然很低。（4）受体分布结果：注射SPIO后，兔肝Kupffer细胞内可见较多蓝染铁颗粒沉着，肝细胞内无蓝染铁分布，脾脏内也可见到较多蓝染铁颗粒；注射Gal--BSA-SPIO后，兔肝细胞内可见较多蓝染铁颗粒沉着，而仅有少数Kupffer细胞内可见蓝染铁颗粒沉着，脾组织内罕见蓝染铁颗粒。电镜下观察，注射Gal--BSA-SPIO后肝细胞溶酶体内可见较多高电子致密颗粒，而仅有极少数Kupffer细胞溶酶体内亦可见高电子致密颗粒；注射SPIO后，肝实质细胞内无高电子致密颗粒，而Kupffer细胞溶酶体内可见较多高电子致密颗粒。2．兔VX2肝癌模型MR成像结果（1） 20只兔肝VX2肿瘤均建模成功，MR共检出肿瘤28个，其中SPIO组检出13个，病理解剖15个；Gal-BSA-SPIO组检出15个，病理解剖16个。肿瘤呈类圆形，灰白色，部分肿瘤有纤维包膜，最小直径3mm，最大直径1.2cm。（2） T值测量结果注射SPIO后，正常肝脏T1与平扫均无显著性差异（P＞0.05），注射Gal-BSA-SPIO后，正常肝脏T1与平扫均有显著性差异（P＜0.05），T1增强前后差值不同组间均无显著性差异（F=1.059，P=0.385）。各组增强后正常肝脏T2与平扫均有显著性差异（P＜0.05）：T2增强前后差值不同组间差异均有显著性（F=19.295，P=0.000），组间多重比较示T2增强前后差值101.μmol Fe／kg Gal-BSA-SPIO与其它各组均有显著性差异（P＜0.05），5μmol Fe／kg Gal-BSA-SPIO与同剂量SPIO有显著性差异（P=0.001），与10μmol Fe／kg SPIO无显著性差异（P=0.080）。各组肝VX2肿瘤T1及T2增强前后均无显著性差异（P＞0.05），不同组间T1及T2增强前后差值亦无显著性差异（P＞0.05）。（3）肝脏SNR测量结果SE 300／20 5μmol Fe／kg SPIO增强后肝脏SNR与平扫前无显著性差异（F=2.098，P=0.090），其余各序列不同剂量的SPIO或Gal-BSA-SPIO增强后肝脏SNR与增强前均有显著性差异（P＜0.05）；SE 300／20不同组间SNR增强前后差值无显著性差异（F=2.624，P=0.074）；其余各序列不同组间增强前后差值均有显著性差异（P＜0.05），组间多重比较：SE 600／80、TSE 3500／105及GRE 600／15 10μmol Fe／kg Gal-BSA-SPIO与其它各组均有显著性差异（P＜0.05），5μmol Fe／kg Gal-BSA-SPIO与5μmol Fe／kg SPIO均有显著性差异（P＜0.05）。各组不同序列间肝脏SNR均有显著性差异（P＜0.05）。（4）肝脏PSIL测量结果各序列不同组间肝脏PSIL均有显著性差异（P＜0.05），组间多重比较：SE 300／20 10μmol Fe／kg Gal-BSA-SPIO与5μmol Fe／kgGal-BSA-SPIO及10μmol Fe／kg SPIO有显著性差异（P＜0.05）：SE 600／80除5μmol Fe／kg Gal-BSA-SPIO与10μmol Fe／kg SPIO无显著性差异外（P=0.159），其余各组间均有显著性差异（P＜0.05）；TSE3500／105及GRE600／15各组间均有显著性差异（P＜0.05）。各组不同序列间肝脏PSIL均有显著性差异（P＜0.05），不同序列间多重比较：5μmol Fe／kg SPIO GRE600／15与SE600／80及TSE3500／105有显著性差异（P＜0.05），10μmol Fe／kg SPIO及不同剂量的Gal-BSA-SPIO除SE600／80与TSE3500／105无显著性差异外（P＞0.05），其它不同序列间均有显著性差异（P＜0.05）。（5）肿瘤—肝脏CNR测量结果各序列不同剂量的SPIO或Gal-BSA-SPIO增强后CNR与平扫均有显著性差异（P＜0.05），CNR增强前后差值各序列不同组间亦均有显著性差异（P＜0.05）。组间多重比较：SE300／20 10μmol Fe／kgGal-BSA-SPIO与不同剂量SPIO均有显著性差异（P＜0.05），SE600／80、TSE3500／105、GRE600／15不同组间两两比较均有显著性差异（P＜0.05）。平扫各组不同序列间CNR均有显著性差异（P＜0.05），组间多重比较：10μmol Fe／kg SPIO组SE 300／20与SE600／80 CNR无显著性差异（P=1.0），其余各组不同序列间CNR均有显著性差异（P＜0.05）。增强后各组不同序列间CNR均有显著性差异（P＜0.05），组间多重比较：不同剂量的SPIO及5μmol Fe／kg Gal-BSA-SPIO增强后，SE 600／80与TSE 3500／105无显著性差异（P＞0.05），其它不同序列间均有显著性差异（P＜0.05）；10μmol Fe／kg Gal-BSA-SPIO增强后，不同序列间CNR均有显著性差异（P＜0.05）。（6）组织病理学HE染色，VX2肿瘤组织以鳞癌为主，可见少量腺癌成分，部分肿瘤组织周围可见纤维组织增生、形成纤维性假包膜。瘤周肝组织基本正常。铁染色，注射SPIO组，兔正常肝组织Kupffer细胞内可见蓝染铁颗粒，而肝实质细胞内未见蓝染铁颗粒，肿瘤组织内亦未见蓝染铁颗粒；注射Gal-BSA-SPIO组，兔正常肝组织肝细胞内可见较多蓝染铁颗粒，仅极少数Kupffer细胞内可见蓝染铁颗粒，肿瘤组织内未见蓝染铁颗粒。3．ASG受体介导的Gal--BSA-SPIO人肝脏MR受体成像初步研究（1）受体亲和力测定结果正常肝组织用不同对比剂孵育后，T2值Gal--BSA-SPIO和SPIO有显著性差异（P=0.003）。肝硬化组织用不同对比剂孵育后T2值Gal--BSA-SPIO和SPIO有显著性差异（P=0.001）。肝癌组织用不同对比剂孵育后T2值SPIO和GaI--BSA-SPIO无显著性差异（P=0.829）。孵育前及不同对比剂孵育后，正常肝组织和肝硬化组织T2值均有显著性差异（P＜0.05）。（2）受体分布正常肝组织SPIO孵育后肝细胞膜及胞内均未见蓝染铁，GaI-BSA-SPIO孵育后肝细胞膜及胞内可见大量蓝染铁颗粒；轻中度肝硬化未孵育或SPIO孵育后，肝细胞内可见少量蓝染铁颗粒沉积，细胞膜无蓝染，Gal-BSA-SPIO孵育后肝细胞膜及细胞内可见较多蓝染颗粒沉积；重度肝硬化及癌旁肝硬化组织未孵育及SPIO或Gal-BSA-SPIO孵育后细胞内均可见大量蓝染铁颗粒沉积，Gal-BSA-SPIO孵育后肝细胞膜未见蓝染铁沉积。未孵育及SPIO或Gal-BSA-SPIO孵育后，肝细胞癌罕见蓝染铁颗粒分布，胆管细胞癌及转移瘤均无蓝染铁分布，肝细胞结节状增生及腺瘤组织可见蓝染铁颗粒散在分布，而血管瘤组织无蓝染铁颗粒分布。（3） CD34表达及MVD结果肝细胞结节状增生及肝硬化结节CD34阳性主要分布于汇管区和纤维间隔中的小血管，本组2例结节状增生和1例肝硬化结节在肝血窦壁也见较多的CD34淡黄色表达；10例肝细胞癌组织内呈棕黄色至棕褐色窦隙状弥漫分布，透明变细胞癌CD34无表达或散在弱阳性表达。5例富血供小肝细胞癌MVD值为121.5±40.4：5例乏血供小肝细胞癌MVD值为133.5±27.8。（4）肝细胞结节CT／MR强化特征3例肝细胞结节状增生，平扫均为低密度，其中2例密度欠均匀，增强扫描动脉期明显强化，静脉期及延迟扫描为略高或等密度，病灶周围可见强化血管影，其中1例可见中心瘢痕（延迟强化）；此外1例动脉期及静脉期或延迟扫描均未见明显强化。5例肝硬化结节，2例平扫为等或稍低密度，增强扫描动脉期呈等密度，静脉期呈等或稍低密度；MR检查1例，未见明确结节，此例同时发生原发性肝癌：另2例CT平扫为等密度，动脉期病灶均可见中等偏高强化，静脉期为等密度，延迟扫描为等或稍低密度。小肝癌富血供（动脉期强化大于20个单位）、乏血供（动脉期不强化或强化小于20个单位）各5例。结论1．Gal--BSA-SPIO与肝细胞膜ASG受体具有良好的结合活性，其主要作用于肝实质细胞，可明显缩短肝脏T2驰豫时间，具有良好的肝脏负向强化效果；2．Gal-BSA-SPIO可显著提高肿瘤—肝脏对比噪声比；含铁量相同的Gal-BSA-SPIO对肝脏的负向强化效果优于SPIO；Gal-BSA-SPIO增强效果最佳的序列是GRE *T2WI，其次是SE T2WI和TSE T2WI，SE T1WI最差。3．正常肝组织、轻中度肝硬化组织、肝细胞结节状增生及腺瘤有丰富或较多的Gal-BSA-SPIO聚集；重度或癌旁肝硬化组织摄取Gal-BSA-SPIO明显减少；肝细胞癌罕见Gal-BSA-SPIO分布；肝血管瘤及胆管细胞癌及转移瘤无Gal-BSA-SPIO分布；Gal-BSA-SPIO在人肝脏MR成像如肿瘤检出及良、恶性肿瘤的鉴别方面具有良好的潜在应用价值。4．小肝癌主要由肝动脉供血，多呈富血供表现，CD34呈弥漫强阳性表达；动脉期不强化或伴点片状不强化最主要的原因是坏死和癌细胞透明变；透明变细胞癌不强化的原因在于其微血管分布极少（CD34无表达或弱阳性）、血供差，但其内有少数ASG受体分布，是肿瘤分化好的一种标志。5．Gal-BSA-SPIO在肝细胞性结节的良恶性鉴别方面具有潜在的优势，但对于部分不典型增生结节（DN）和高分化肝细胞癌仍具有一定的局限性，尤其对于重度肝硬化基础上的肝细胞性结节的诊断与鉴别诊断价值可能有限。更多还原

【Abstract】 Objective1. To prepare Galactose-Bovine-Serum-Albumin （Gal-BSA-SPIO） nanoparticles as Asialoglycoprotein Rceptor -directed MR contrast agent, and to test specificity of the agents in rabbit liver tissues in vitro and in vivo MR imaging experiments.2. To investigate and debate the effect and potential values of Gal-BSA-SPIO as MR specific contrast agents targeted imaging for liver VX2 tumor in rabbits.3. To study and evaluate primarily the potential values of Gal-BSA-SPIO as receptor -directed contrast agents for MR imaging with affinity and histologic receptor assay.4. Compare and analyze the enhancement features on CT or MR images with histological pathology and immunohistochemsitry of human hepatocyte nodules to know more profoundly the imaging features of its and to debate primarily whether the Gal-BSA-SPIO as MR contrast agents is more superior.Materials and Methods1. Preparation of Gal-BSA-SPIO and the affinity assays（1） Preparation of Gal-BSA-SPIO and assays By method of reductive amination, bovine serum albumin （240mg）, lactose （1.2mg） and sodium cyanoborohydride （100mg） were dissolved in 30 ml of 0.2 M natrium phosphate（PH 8） and incubated at 37℃for 24 to 120 hours. The reaction mixture were dialyzed extensively （3 days） against distilled water at lower temperature. Aliquots of the dialyzed mixture were centrifuged and the supernatant was withdrawn applied to a sephadex gel chromatography column. The purified conjugates were lyophilized. Total carbohydrate and albumin were determined separately by the phenol-sulfuric acid method and coomassie brilliant blue G-250 to calculate the ratio of galactose to albumin.（2） Preparation of Gal-BSA-SPIO Smaller iron oxide particles were obtained through size fractionation of SPIO with use of a Sepharose 4B column. The separated smaller SPIO was dialyzed against buffer solution for 24 hours and then concentrated to obtain an iron concentration of 8mg/ml （PH6.5）. Fractions were diluted 1:1 into a neutral buffer containing 1% W/V galactose-BSA. Thereafter, sonication was performed in an ice bath. The loosely associated BSA was removed by high salt（1M NaC1） and added by dropwise of ammonium hydroxide to PH 7.4. The iron concentration was tested and the particle size and distribution was measured by a dynamic light scattering （a Malvern Zetasizer 3000 HS）. The core particle size and morphology of Gal-BSA-SPIO were examined by transmission electron microscopy.（3） Affinity assays in vivo and vitroReceptor affinity experiments in vivo Fresh liver tissues in rabbits （10g） were obtained and added with complete growth medium. And then the tissues were cut into 1-2mm³ pieces, minced through a 30 wire mesh, and incubated with 0.3 mg/ml collagenase in complete growth medium for 30 minutes at 37℃. Thereafter, the dispersion was sonicated in an ice bath to disrupt tissue and cell fragments. Samples were centrifuged, and the supematant that contained cell membranes was used for incubation assays. （1）Blocker group Samples （2ml） of the membrane suspensions were incubated for 30 minutes at 37℃with complete medium contained different doses of D（+）-galactose（0.2mg or 2mg） to block existing ASG receptor on the membrane. After this incubation, either SPIO or Gal-BSA-SPIO （contained of 101amol iron） was added to the suspensions. 30 minutes later, samples were washed twice to remove free unbound iron oxide particles. Cell and membrane fragments were then resuspended in 6ml of complete growth medium. （2） Un-blocker group Samples （2ml） of the membrane suspensions were directly incubated with either SPIO or Gal-BSA-SPIO （contained of 10μmol iron）. 30 minutes later, samples were also washed twice and resuspended in 6ml of complete growth medium. Relaxation times of all samples were measured at 1.5T, and statistics were performed.Relaxation studies in vitro Six rabbits were randomly divided into two groups of blocker and un-blocker. MR imaging was separately performed before and after administrated with agents. Spin-echo（SE） images were obtained with following pulse sequences: SE TR/TE 600/20, 80, and transverse section images were obtained. The rabbits of blocker group were administrated firstly with D（+）-galactose （2mg/kg） to block ASG receptor on the hepatocyte membrane, and 30 minutes later, all were administrated with Gal-BSA-SPIO（10μmol Fe/kg）. The animals of unbloked group were directly administrated with GaI-BSA-SPIO（10μmol Fe/kg）. MR imaging of all animals were performed after administration of Gal-BSA-SPIO for 30 minutes, with the same sequence as before. Thereafter, T2 relaxation times pre- and post-enhancement were measured.Histological assays Liver and spleen specimens were examined 30min after injection of 20μmol Fe/kg to study the cellular distribution of Gal-BSA-SPIO or SPIO in two rabbits. After removal, the liver tissue was performed with histological specimens and electron microscopy, and the spleen tissue with histological specimens.2. Application of Gal-BSA-SPIO in VX2 tumor of rabbits（1） Model of VX2 tumor in rabbits Tumor tissue with active growth characters looked like fish was obtained from the border of VX2 tumor in breeder-rabbit. Tissue was added some physiological saline, cut into 0.5-1mm³ with ophthalmic scissor. The rabbits were anaesthetized and paunched exposed the liver. One lobe of the liver was pulled lightly outside, and pierced by ophthalmic forceps. Thereafter, a small tumor piece was put into the pierced hole. One to three pieces of tumor were put in.（2） Animal group 20 rabbits were randomly divided into two groups of SPIO and Gal-BSA-SPIO, and 10 were of each group. Each group was then separately divided into two groups of 5μmol Fe/kg and 10μmol Fe/kg.（3） MR imaging MR imaging was performed at 1.5T with a section thickness of 3-5mm, and a field of view of 87.5^*100. Pulse sequences were used as below: SE T₁WI: TR300ms, TE20ms、3000ms; SET₂Wh TR600ms, TE20ms、80ms; TSE T₂WI-TR3500ms, TE15ms、105ms; GRE T₂^*WI: TR600ms, TE 15ms, 18°Transverse, coronal and sagittal multisections images were obtained. Contrast scan was performed with administration of different agents by the groups.（4） Observation and measure of images Relaxation times of liver and tumor were measured before and after administration of agents. And signal intensity of liver and tumor were measured to calculate SNR of liver, the percent of the decrease of signal intensity of liver and CNR of liver to tumor.（5） Statistical analysis Paired t tests were used to evaluate data from pre- and post -contrast images. One-way ANOVA was used to evaluate PSIL and the difference of T1, T2, SNR, and CNR pre-contrast and pos-contrast. One-way ANOVA was used to evaluate pulse sequences. Multi-comparison was performed with LSD or Dunnett’s T3. The significance of reported P values were defined as P＜0.05.3. Pilot study of Gal-BSA-SPIO as ASG receptor-directed agents for MR imaging in human liver（1） Affinity assays A total of 13 specimens were collected from surgery. Specimens included normal liver tissue（n=3）, cirrhosis（n=4）, hepatocyte carcinoma（n=6）. All specimens were performed to cell and membrane suspensions by the above methods （relaxation times measure of rabbits liver in vivo）. Relaxation times of T2 were measured. One-way ANOVA and Independent-samples were used to evaluate the datas. Multi-comparison was performed with LSD or Dunnett’s T3. The significance of reported P values were defined as P＜0.05.（2） Distribution of ASG receptor in different fresh human tissue All specimens were cut into 2-4μm sections（-20℃）. Sections were then thawed, air dried, and incubated 20 seconds（37℃） with Gal-BSA-SPIO, SPIO, or buffer solution. After incubation, slides were rinsed with saline solution to remove unbound iron. Slides were then dried for 24 hours at room temperature. Subsequent counterstaining for iron was performed with Perls Prussian blue stain.（3） Distribution of ASG receptor in different human tissue （paraffin masses） A total of 27 specimens were collected, consisted of small hepatocyte carcinoma（n=12）, metastases（n=3）, cholangiocarcinoma（n=2）, cirrhosis nodules（n=5）, focal nodular hyperplasia（n=3）, hemangiomas（n=3）,adenomas（n= 1）. Specimens were cut into 2-41am sections（ at room temperature）, and then baked for 60 minutes（65℃）. Sections were dewaxed extensively with xylene. Thereafter, sections were air dried, and incubated with Gal-BSA-SPIO, SPIO, or buffer solution for 10 minutes （37℃）. After incubation, slides were rinsed with saline solution to remove unbound iron. Slides were then dried for 24 hours at room temperature. Subsequent counterstaining for iron was performed with Perls Prussian blue stain.（4） CD34 expression A total of 18 specimens, consisted of small hepatocyte carcinoma（n=10）, cirrhosis（n=5） and focal nodular hyperplasia（n=3）, were performed by CD34 stain with SABC. CD34 expression were observed and MVD were calculated by Weidner method.（5） Compare and analyze the enhancement features on CT or MR images with histological pathology and immunohistochemsitry of human hepatocyte nodules.Results1. Results of preparation of Gal-BSA-SPIO and the affinity assays（1） The ratio of galactose to albumin for 24h, 48h, 72h, 96h, and 120h were separately 8, 15, 27, 32 and 38. The iron concentration of GaI-BSA-SPIO was 2.8mg/ml. The mean volume size was 34.4nm and the index of size distribution was 0.35 measured by dynamic light scattering. The core particle size was 14.8nm and the morphology was spherical shape with slight adherence between some particles observed by transmission electron microscopy.（2） There was no significant difference between the blocker and un-bloeked group before incubation with agents（P＞0.05）.After incubation with SPIO, the different blocker group showed no significant difference with the un-blocker group（P＞0.05）. T2 relaxation times of membrane suspension post-incubation with GaI-BSA-SPIO were lower statistically than pre-incubation and post-incubation with SPIO（P＜0.05）. After incubation with Gal-BSA-SPIO, T2 relaxation times of membrane suspension incubation of D（＋）-Galactose to block the ASG receptor on the membrane were greater statistically than without blocker （P＜0.05）. And there was significant difference between the different doses of D（＋）-Galactose （P＜0.05）.（3） After administration of GaI-BSA-SPIO, T2 relaxation times and signal intensity of liver decreased. The liver looked like "dark liver" since the signal intensity decreased so marked as to near the noise around. T2 relaxation times and signal intensity of liver increased if administrated firstly of D（＋）-Galactose before injected agents.（4） After administration of SPIO, many blue iron deposits were seen in kupffer cells while no in hepatocytes. And blue iron deposits were also seen in spleen tissue. After administration of Gal-BSA-SPIO, many blue iron deposits were seen in hepatocytes while only in few kupffer cells, and no in spleen tissue. Electron-dense particles were seen in lysosomes in hepatocyte after administration of Gal-BSA-SPIO by transmission electron micrograph. Electron-dense particles were seen in lysosomes in kupffer cell while no in hepatocyte after administration of SPIO.2. MR imaging of VX2 tumor in rabbits（1） All of 20 models of VX2 tumor in rabbits were successfully. A total of 28 tumors were detected by MR Imaging, 13 of them with SPIO（15 with anatomy） and 15 of them with Gal-BSA-SPIO（16 with anatomy）. Tumors were hoar and spherical shape. There were seen fibrous tissues proliferation around some tumor, looked like some capsular. The minimal and maximal diameter was separately 3mm and 1.2cm.（2） T1 relaxation time of liver post-contrast with SPIO showed no significant difference with pre-contrast（P＞0.05）, While T1 relaxation time post-contrast with Gal-BSA-SPIO showed significant difference with pre-contrast（P＜0.05）. The difference pre-contrast and post-contrast showed no significant between different groups （F=1.059, P=0.385）. T2 relaxation time of liver post-contrast showed significant differences with pre-contrast（P＜0.05）, and the differences pre-contast and post-contrast showed also significant between different groups （F=19.295, P=0.000）. By multi-comparison, the differences of T2 relaxation times pre-contrast and post-contrast, 10μmol Fe/kg Gal-BSA-SPIO showed significant differences with the other gourps（P＜0.05）, 5μmol Fe/kg Gal-BSA-SPIO showed significant difference with 5μmol Fe/kg SPIO （P=0.001） while no significant difference with 10μmol Fe/kg SPIO（P=0.080）. T1 and T2 relaxation times of tumor post-contrast with different agents showed no significant differences with pre-contrast（P＞0.05）, neither the differences pre-contrast and post-contrast between different groups（P＞0.05）.（3） By pulse sequence of SE 300/20, SNR of liver post-enhanced with 5μmol Fe/kg Gal-BSA-SPIO showed no significant difference with pre-enhanced （F=2.098,P=0.090）. SNR of liver post-enhanced in different groups by other pulse sequences showed significant difference with pre-enhanced（P＜0.05）. The difference pre-contrast and post-enhanced of liver SNR showed no significant difference （F=2.624,P=0.074） by pulse sequence of SE 300/20,while showed significant difference by other sequences. By multi-comparison, 10μmol Fe/kg GaI-BSA-SPIO showed significant difference with the other gourps by pulse sequences of SE600/80, TSE3500/105 and GRE 600/15 （P＜0.05）,and 5μmol Fe/kg Gal-BSA-SPIO showed significant difference with 5μmol Fe/kg SPIO（P＜0.05）. Liver SNR on different pulse sequences showed significant differences（P＜0.05）.（4） There were significant differences of PSIL within different groups on all pulse sequences（P＜0.05）. By multi-comparison, on pulse sequence of SE 300/20, 10μmol Fe/kg GaI-BSA-SPIO showed significant differences with 5μmol Fe/kg GaI-BSA-SPIO and 10μmol Fe/kg SPIO（P＜0.05）. On pulse sequence of SE 600/80, the different groups showed significant differences （P＜0.05）except 5μmol Fe/kg GaI-BSA-SPIO with 10μmol Fe/kg SPIO（P=0.159）. On pulse sequence of TSE 3500/105 and GRE 600/15, the different groups showed significant differences （P＜0.05）. Different pulse sequences showed significant differences（P＜0.05）. By multi-comparison, enhanced with agent of 5μmol Fe/kg SPIO, pulse sequence of GRE600/15 showed significant differences with SE 600/80 and TSE 3500/105（P＜0.05）, and with agent of 10μmol Fe/kg SPIO and different dose of GaI-BSA-SPIO, the different pulse sequences showed significant differences（P＜0.05） except pulse sequence of SE600/80 with TSE3500/105（P＞0.05）.（5）Tumor-liver CNR post-contrast with different agents showed significant differences with pre-contrast （P＜0.05）, and the differences of pre-contrast and post-contrast showed aslo significant differences bwtween different groups（P＜.05）. By multi-comparison, on pulse sequence of SE 300/20, 10μmol Fe/kg GaI-BSA-SPIO showed significant differences with different doses of SPIO（P＜0.05）. On pulse sequences of SE600/80, TSE 3500/105 and GRE 600/15, there were significant differences between different groups（P＜0.05）. Before contrast, there were significant differences within different pulse sequences （P＜0.05）. By multi-comparison, enhanced with agent of 10pmol Fe/kg SPIO, there were significant differences within different pulse sequences （P＜0.05） except between SE300/20 and SE600/80（P=1.0）. After contrast, there were also significant differences within different pulse sequences （P＜0.05）. By multi-comparison, enhanced with agent of SPIO and 5μmol Fe/kg GaI-BSA-SPIO, there were significant differences within different pulse sequences （P＜0.05） except between SE600/80 and TSE 3500/105（P＞0.05）. Enhanced with agent of 5μtmol Fe/kg GaI-BSA-SPIO, there were significant differences within different pulse sequences （P＜0.05）.（6）Stained by Hematoxylin-Eosin, VX2 tumor was composed mainly of squamous cell carcinoma and few of adenocarcinoma. There were seen fibrous tissues proliferation around some tumor, looked like some capsular. After administration of SPIO, many blue iron deposits were seen in kupffer cells while no in hepatocyte and tumor tissue. After administration of Gal-BSA-SPIO, many blue iron deposits were seen in hepatocyte while only in few kupffer cells, and no in tumor tissue.3. Pilot study of GaI-BSA-SPIO as ASG receptor-directed agents for MR imaging in human liver（1） T2 relaxation time of the normal liver tissue incubated with Gal-BSA-SPIO showed a significant difference with SPIO （P=0.003）. And T2 relaxation time of cirrhosis tissue incubated with Gal-BSA-SPIO showed a significant difference with SPIO （P=0.001）. T2 relaxation time of tumor tissue incubated with Gal-BSA-SPIO showed no significant difference with SPIO （P=0.829）. In addition, T2 relaxation time of the normal tissue showed a significant difference with cirrhosis pre-contrast and post-incubation with different agents （P＜0.05）.（2） After administration of SPIO, many blue iron deposits were seen in kupffer cells while no in hepatocyte tissue. After administration of Gal-BSA-SPIO, many blue iron deposits were seen in hepatocyte of normal liver tissue while only in few kupffer cells. Some blue iron deposits were seen in hepatocyte cytoplast while no on membrane in low and moderate cirrhosis tissue pre-incubation or post-incubation with SPIO. Many blue iron deposits were seen both in hepatocyte cytoplast and on membrane in low-grade and moderate-grade cirrhosis tissue pre-incubation or post-incubation with GaI-BSA-SPIO. Lots of blue iron deposits were seen in hepatocyte cytoplast in high-grade cirrhosis pre-incubation or post-incubation with SPIO or Gal-BSA-SPIO. And no blue iron deposits were seen on membrane in high-grade cirrhosis post-incubation with GaI-BSA-SPIO. Few blue iron deposits were seen in carcinoma tissue post-incubation with Gal-BSA-SPIO. No blue iron deposits were seen in metastases and cholangiocarcinoma tissue pre-incubation and post-incubation. Many blue iron deposits were seen in focal nodular hyperplasia and adenomas pre-incubation and post-incubation, while no in hemangiomas（3） CD34 expression was seen mainly in small vessels inside the fibrous tissues in focal nodular hyperplasia and cirrhosis regenerating nodules. The hyperplatic hepatocyte nodules also showed diffuse positive along the sinusoid in two specimen of focal nodular hyperplasia and one specimen of cirrhosis regenerating nodules. The hepatocyte carcinoma tissue almost showed diffuse CD34 expression, except the hyaline carcinoma area showed few dispersible positive. Mean MVD of 5 specimens of full-blooded sHCC was 121.5±40.4, and that of 5 specimens of poor-blooded sHCC was 133.5±27.8.（4） 3 specimens of focal nodular hyperplasia were all hypodense on pre-contrast scan. The nodules were markedly on arterial-phase scan, turned to isodense or slightly hypodense on portal and delayed phase scans. Arteries were seen at the peripheral area, and the central scar with delayed enhanced was seen in one specimen. Another specimen of focal nodular hyperplasia were all hypodense on different phase of post-contrast. 2 specimens of cirrhosis regenerating nodules were isodense or slightly hypodense on pre-contrast scan, and turned to isodense on arterial-phase scan, isodense or slightly hypodense on portal phase scans. 1 specimen was not found on MR imaging, and it was with hepatocyte carcinoma. In addition, 2 specimens of cirrhosis regenerating nodules were isodense on pre-contrast scan, and turned to hyperdense on arterial-phase scan, isodense on portal phase scans and isodense or slightly hypodense on delayed phase scan. Full-blooded and poor-blooded sHcc were separately 5 specimens.Conclusion1. GaI-BSA-SPIO bind well with ASG receptor on hepatocyte membrane. It distributed mainly in hepatocyte and decreased markedly T2 relaxation time of normal liver that brought a good negative enhancement effect in liver.2. Gal-BSA-SPIO improved obviously tumor-liver CNR, and brought a better enhancement effect in in liver than SPIO contained same dose of iron. The pulse sequence on which there was a best enhancement effect with GaI-BSA-SPIO was GRE T2-weighted imaging, SE T2-weighted imaging and TSE T2-weighted imaging were second, and SE T1-weighted imaging was inferior.3. GaI-BSA-SPIO accumulated in normal liver tissue, low-grade and moderate-grade cirrhosis, focal nodular hyperplasia and adenomas, while decrease obviously in great or adjacent to carcinoma cirrhosis tissues. Few were seen in hepatocyte carcinoma and no in hemangiomas, metastases and cholangiocarcinoma. Gal-BSA-SPIO was with a potential value applied to detect tumor of liver and discriminate the benign tumor from malignant.4. Blood of small hepatocyte carcinoma was mainly supplied by liver artery, and many took on full-blooded features on CT or MR imagings. This hepatocyte carcinoma tissue almost showed diffuse CD34 expression. Some small hepatocyte carcinoma showed no or slight enhancement on arterial phase scan, which was because mainly of necrosis and hyaline carcinoma. Hyaline carcinoma was with few microvessel distribution and poor-blooded supplied while with some ASG receptor distribution, which was an index of gooddifferentiation.5. Although Gal-BSA-SPIO showed a potential value in diagnosis the benign tumor from malignant, it had some difficulty in discrimination some dysplastic nodules from good-differentiation hepatocyte carcinoma, especially in diagnosis and differentiation some nodules founded on cirrhosis.更多还原