肝細胞癌及肝細胞移植之細胞增殖研究

Study of Cell Proliferation in Hepatocellular Carcinoma and Hepatocyte Transplantation

袁瑞晃(Ray-Hwang Yuan)

指導教授 ： 李伯皇 ； 陳惠玲

國立臺灣大學/醫學院/臨床醫學研究所/博士(2005年)

https://doi.org/10.6342/NTU.2005.00435

摘要

肝細胞癌是世界上人類常見惡性腫瘤之一，約佔所有惡性腫瘤的百分之四，據統計西元兩千年當年，全世界約有五十六萬四千名新病例發生，同時也有大約相同的病患死於肝細胞癌。肝細胞癌好發於南中國、東南亞、撒哈拉地區之下的非洲(sub-Saharan Africa)以及台灣。雖然造成肝細胞癌的真正機制仍然未能全盤了解，但造成肝細胞癌的主要原因為B型肝炎病毒、C型肝炎病毒、黃麴毒素、酒精、阿爾發-1抗胰蛋白酶缺乏症、酪氨酸血症等。在我國，癌症位居國人十大死因之首。根據行政院衛生署統計，2003年在台灣約有四萬六千位病患因肝細胞癌而住院或接受門診治療，而因肝癌去世的病人佔所有因癌症死亡人數的第一位，2004年台灣因肝癌而死亡的人數也居所有因癌症死亡人數的第二位。雖然超過百分之六十的肝細胞癌合併有肝硬化，外科治療仍然是目前治療肝細胞癌最有效的方法，但能藉由外科手術切除腫瘤的比例仍然很低。雖然在台大醫院的病人當中可以藉由手術切除腫瘤的比例已由1987年的百分之十八提高到1997年的百分之三十三，但是1995年在台灣所有新增的4812位病例中還是只有百分之十的病人可以接受手術切除腫瘤。隨著分子生物以及細胞生物學研究的發展，包括p53基因和其相關分子、血管生成因子、細胞週期調控因子、致癌基因及其接受器、細胞凋亡相關因子、細胞增殖標記、細胞外基質分解相關的蛋白酶以及細胞附著分子等被陸續發現，且和肝細胞癌侵襲性、轉移性、再發性以及病人存活率有關的研究也積極的展開，這些研究成果雖然能夠應用於推測疾病的預後情況，但整體而言對肝癌病人的病情的改善並沒有特別的幫助，因此在這方面的發展還是有極大的空間。對肝細胞癌而言，腫瘤的侵襲性是影響預後最重要的因子，可以造成臨床上腫瘤分級、腫瘤分期、血管侵襲、肝臟內早期再發以及預後的差別，因此本研究希望能夠找到與腫瘤細胞增殖，造成腫瘤侵襲性有關的基因，一方面可以作為病人預後的指標，另一方面可以作為將來治療的標的。
根據美國肝臟協會1996年的統計，肝衰竭是美國第七大死亡原因，每年有25萬人因肝臟疾病入院，且有52,000人因肝臟疾病而死亡，平均一年花費在治療肝臟疾病的醫療費用高達90億美元。國內由於B型肝炎、C型肝炎普遍，因此慢性肝臟疾病病人極多。2003年台灣的慢性肝臟疾病(慢性肝炎及肝硬化)排名十大死亡原因的第六位，而2004年則排名十大死亡原因的第七位。此外急性肝衰竭也是造成病人死亡重要的疾病，由於致病原因的不同急性肝衰竭死亡率可達40~90%，而肝臟移植也是治療急性肝衰竭的方法之一，因此肝臟移植的需求至今有增無減。國內的肝臟移植開始於1984年3月，臺大醫院的肝臟移植開始於1989年，兒童的活體肝臟移植自1996年開始，成人的活體肝臟移植自2000年開始。由於免疫抑制藥物的發展，國內肝臟移植四年平均存活率大於70%，手術成

功率雖然高，但是術後的合併症仍多(如急性感染、排斥及膽道狹窄)。但肝臟來源有限，雖然捐贈的肝臟可以切開後分給不同的人進行移植(Split liver transplantation)，甚至可以近親活體捐肝(living related donor)，但器官缺乏的問題仍然存在，因此除了預防與治療肝臟疾病的研究必須持續進行外，尋找新的移植方法以彌補肝臟器官移植的不足就顯得極為迫切且重要。以上所述顯示肝細胞癌、慢性肝病的研究以及治療方法的改進對國人健康的重要性，也因此本研究主要分成兩大部分，分別針對肝細胞癌分子病理以及肝細胞移植中影響細胞增殖的因素進行探討。

肝細胞癌之細胞增殖研究
根據台大醫院經驗，許輝吉教授曾經報告過許多基因上的變化與肝細胞癌的腫瘤分級、腫瘤分期、血管侵襲、肝臟內早期再發以及預後的差別有關。本論文利用抑制性減差雜交法及微陣晶片技術(suppression subtractive hybridization and microarray)探討肝細胞癌相關基因時發現差異最顯著的就是胰臟再生蛋白1A (Regenerating gene 1A - REG1A)，此蛋白與胰臟再生關係非常密切，而在人類肝細胞癌中的過量表現是一個全新的發現。REG1A位於人類第二對染色體上，而同一區域上有另一與人類肝細胞癌有關係的胰臟炎相關蛋白(Pancreatitis Associated Protein - PAP) 基因存在，兩者前後串聯排列，可能是由相同的祖源基因經過複製而來。雖有報告指出PAP在部分肝細胞癌中有過量的表現，但卻缺乏任何臨床上相關性的研究，因此在肝細胞癌之細胞增殖研究的第一部分研究就決定同時探討REG1A及PAP這兩種和細胞增殖有關的蛋白在人類肝細胞癌中所扮演的角色。
胰臟再生基因(Regenerating gene)是分析接受了百分之九十胰臟切除的大鼠於胰臟再生時所發現的基因，其蛋白具有抗細胞凋亡及增加細胞生長的作用。胰臟再生基因目前被分成四型，REG1A是屬於第一型，而PAP屬於第三型，兩者前後串聯排列於第二對染色體長度95 kb的2p12區域上。REG1A是胰島細胞中β-細胞所合成，可在過度增生的胰島細胞內表現，也和神經細胞以及消化道上皮細胞的增殖有關，但在正常胰島、胰島細胞瘤及再生的肝臟中並不表現。此外在胃癌、大腸直腸癌、膽管細胞癌也可以見到REG1A的過量表現，且隨著細胞分化的變差而增加。PAP和REG1A有百分之四十九的相似性，在正常胰臟中只有極少量的表現，但在急性胰臟炎或慢性胰臟炎時會明顯的增加表現。PAP可表現在正常的小腸、胰臟以及腦下腺，但在其他組織中無法偵測到。此外胃癌、大腸直腸癌、胰臟癌以及四分之一的肝細胞癌中也有PAP的過量表現，這些皆顯示REG1A以及PAP的過量表現與細胞的增殖有密切的關係，但是他們在臨床病理上所扮演的角色仍需進一步的研究。
為了研究腫瘤的病理與臨床之間的關係，我們將腫瘤細胞分化程度分三級：分化良好(第一級)，分化中等(第二級)，分化差(第三級和第四級)。同時也將腫瘤分為四期，第一期及第二期沒有血管侵犯，第三期及第四期則有不同程度的血管侵犯。第一期包括纖維被膜完全包覆而≦2 cm的腫瘤且沒有血管侵犯；第二期包括≦2 cm的腫瘤有肝臟的侵犯，而不論有否主要腫瘤旁邊顯微鏡下可見的衛星腫瘤，或>2cm腫瘤而不論有否肝臟的侵犯以及小的衛星腫瘤在主要腫瘤的旁邊，重要的是沒有血管的侵犯。第三A期可見到纖維被膜內薄壁的血管有腫瘤侵犯現象，但沒有肝門靜脈侵犯或肝臟深層實質內的衛星腫瘤。第三B期在主要腫瘤旁邊的肝門靜脈區可以見到小的肝門靜脈腫瘤侵犯，但主要的肝門靜脈分枝沒有腫瘤侵犯，也沒有肝臟深層實質衛星腫瘤。第四期在主要的肝門靜脈分枝有腫瘤侵犯，肝臟深層實質有衛星腫瘤，腫瘤破裂或是腫瘤侵犯到周圍的器官。此外利用腹部超音波、電腦斷層的影像診斷，輔以臨床上血中胎兒蛋白的增加來診斷腫瘤在肝臟內的再發，並定義一年內的腫瘤再發為早期再發。
在隨機取樣的265位病例當中，包括209位男性以及56位女性，年齡介於14~88歲，平均55.6歲，其中有236位病人(89.1%)可以評估腫瘤早期再發。腫瘤分化程度依次為分化良好(第一級，61個病例)，分化中等(第二級，102個病例)，分化差(第三級和第四級，102個病例)。腫瘤分期為第一期(6個病例)，第二期(115個病例)，第三A期(43個病例)，第三B期(33個病例)以及第四期(68個病例)。結果發現55例(20.8%)有REG1A的過量表現，97例(36.6%)有PAP的過量表現，但在鄰近腫瘤的正常肝臟組織中兩者都沒有過量表現。
PAP過量表現常見於低血清胎兒蛋白值(≦200ng/ml, P=0.039)，以及分期低且沒有血管侵襲的腫瘤(stage I~II，P=0.013)，但REG1A則沒此現象。兩者的過度表現和年齡、性別、慢性B型或C型肝炎、腫瘤大小、腫瘤分級或是腫瘤的早期再發都沒有關係。此外44.8%病例有p53基因突變，14.9%病例有β-catenin基因突變，REG1A及PAP的過量表現都和β-catenin基因突變有關(P<0.00001及P=0.00005)，但和p53基因突變則無關。
進一步分析發現所有病例中，有46例(17.4%)同時有REG1A及PAP的過量表現，而兩者均沒有過量表現者共有159例(60.0%)，因此兩者的一致性非常高(77.4%, P<0.00001)。因此我們將所有的病例依照REG1A及PAP有過量表現(+)與沒有過量表現(-)分成四組：PAP(+)/REG1A(+)，PAP(+) /REG1A(-)，PAP(-)/REG1A(+)以及PAP(-)/REG1A(-)。結果發現四組在年齡、性別、血清胎兒蛋白值以及慢性B型或C型肝炎上都沒有差異。PAP(+) /REG1A(-)病例，多為分級低(grade I，P<0.007)、分期低且沒有血管侵襲的腫瘤(stage I~II，P<0.001)，而且腫瘤在肝臟內早期再發頻率也是四組中最低(P=0.051)。比較PAP(+)/REG1A(-)以及PAP(+)/REG1A(+)病例時發現，在同樣有PAP過量表現的肝癌中，若合併有REG1A的過量表現，則病例多為大於2 cm (P=0.009)，分級高 (grade II~IV，84.8% versus 58.8%，P=0.005)，分期高有血管侵襲 (stage III~IV，60.9% versus 29.4%，P=0.002)的腫瘤，且肝臟內早期再發的頻率也比較高(61.8% versus 31.3%，P=0.006)，顯示REG1A的負面影響甚巨。
由於p53及β-catenin基因突變是肝癌的重要基因突變，病例在分成四組後發現，在沒有PAP過量表現[PAP(-)/REG1A(+)及PAP(-)/REG1A(-)]的肝癌中，大部分的腫瘤都沒有β-catenin基因突變，而有PAP過量表現[PAP(+)/REG1A(+)及PAP(+)/REG1A(-)]時，則不論有無REG1A的過量表現，有β-catenin基因突變腫瘤的比例增加很多(P<0.0001)。分析p53基因突變時發現四組有p53突變的比例並沒有顯著的差異，但在有PAP過量表現[PAP(+)/REG1A(-)及PAP(+)/REG1A(+)]的肝細胞癌中，只有PAP過量表現[PAP(+)/REG1A(-)]的肝癌p53基因突變的比例最低(25.6%)，但是若合併有REG1A的過量表現[PAP(+)/REG1A(+)]時，腫瘤出現p53基因突變的比例就明顯提高(46.2%)，而與其他兩組腫瘤[PAP(-)/REG1A(+)及PAP(-)/REG1A(-)]中p53基因突變的比例相當(48.9%及50.0%)。由以上的結果顯示PAP與REG1A的存在與β-catenin及p53基因突變有微妙的關係存在，因此進一步分析具有PAP過量表現的肝癌[PAP(+)/REG1A(+)及PAP(+)/REG1A(-)]時發現，在有β-catenin基因突變的肝癌中不論是否有REG1A的過量表現，大部分的腫瘤是低分期而沒有血管侵襲(82.1%)且較少有腫瘤早期再發(17.4%)，顯示β-catenin具有比REG1A更強的腫瘤侵襲壓制效應。但在有p53基因突變的肝癌中若同時有REG1A的過量表現，則腫瘤為高分期有血管侵犯(66.7%)、高分級分化差(88.9%)的比例較多，且較多有腫瘤的早期再發(62.5%)，顯示在p53基因突變存在下，REG1A可以抵消PAP使腫瘤傾向侵襲性低的作用。
雖然p53及β-catenin基因突變是肝癌的重要基因突變，且影響肝細胞癌生物行為極深，但仍約有一半的病例並沒有p53或β-catenin基因的突變，其基因變化及影響仍待研究。在沒有p53或β-catenin基因的突變的肝細胞癌病例中，相較於PAP(+)/REG1A(-)病例言，PAP(+)/REG1A(+)病例較多出現分級高(grade II~IV)的腫瘤(P<0.002)，且伴隨分期高有血管侵襲腫瘤的機率是PAP(+)/REG1A(-)病例的三倍(P<0.005)。因此不論是否有p53基因突變的存在，REG1A的過量表現還是會造成侵襲性較強的腫瘤。以上結果顯示，腫瘤只有PAP過量表現者與β-catenin基因突變的腫瘤類似，是傾向於低分級、低分期沒有血管侵犯且較少早期再發；而合併REG1A過量表現者與p53基因突變的腫瘤類似，傾向於高分級、高分期有血管侵犯且較容易發生早期再發。雖然PAP過量表現後腫瘤傾向較好發展的作用會被REG1A過量表現所拮抗，但是β-catenin基因突變對於腫瘤傾向較好進展的效力並不會被REG1A的過量表現所拮抗。最後分析四組病例的五年存活率發現，PAP(+)/REG1A(-)者存活率最高(P=0.044)，但是合併REG1A過量表現者存活率顯然比只有PAP過量表現者低(P<0.0002)。綜合以上結果顯示，PAP和REG1A雖然是同源基因，但在人類肝細胞癌中所扮演的角色卻是相反的。本文可能是最早發現在沒有p53或β-catenin基因突變的肝癌之PAP與REG1A變異表現者，而此兩基因之過量表現呈現相反的效益。
接著在肝細胞癌之細胞增殖研究的第二部分研究是探討另一種同樣與細胞增殖有關的基因-血癌伴隨蛋白18 (leukemia associated protein- LAP18)與人類肝細胞癌的相關性。LAP18最早是在人類血癌細胞HL60內所發現，是一種普遍存於細胞質(cytosol)內的磷酸複合蛋白質(phosphoprotein)，可由細胞外控制細胞增殖、分化的訊息所調節，因而被認為是細胞訊息傳導中(signal transduction)的綜合中繼站(general relay)，此外LAP18也會在細胞分裂過程中調控微細管絲(microtubule filament)的形成與分解，因此與細胞的增殖也有密切關係。LAP18通常在細胞增殖時大量表現，但當細胞開始分化時LAP18的表現會降低。在急性白血病中也可以見到LAP18的大量表現，但當血癌細胞碰到分化促進因子而停止增殖開始分化時，LAP18的表現量也會減少，而同樣的在分化很差的實質腫瘤(solid tumor)且具有高增殖潛能者也可見到LAP18的大量表現。目前文獻上關於LAP18與人類肝細胞癌的報告只有一篇，其中提及B型肝炎相關肝細胞癌會有LAP18的過量表現，但對於LAP18的角色並沒有任何陳述。為了辨認肝細胞癌的預後指標，提供治療的對策，我們研究並討論LAP18和肝細胞癌細胞增殖以及腫瘤侵襲性的關係。
隨機取樣的184個病例包括145位男性以及39位女性，年齡介於14-88歲，平均為55.5歲，其中的124位病例在血漿中可以偵測到B型肝炎表面抗原。腫瘤細胞分化程度依次為分化良好(第一級，39個病例)，分化中等(第二級，79個病例)，分化差(第三級和第四級，66個病例)，腫瘤分期依次為第一期(1個病例)，第二期(76個病例)，第三A期(30個病例)，第三B期(28個病例)以及第四期(49個病例)。結果發現95例(52%)有LAP18的過量表現，且有慢性B型肝炎感染病人有LAP18過量表現的傾向(P=0.06)，但和年齡、性別以及血液中AFP level無關。在組織病理方面，LAP18的過量表現常見於大於5cm腫瘤(P=0.0053)、分級高而分化不良(grade II to IV，P=0.027)以及分期高有血管侵襲的腫瘤(stage IIIA to IV，P=0.0001)。此外，有LAP18過量表現者較常發生腫瘤的早期再發(P=0.022)，且病人七年的存活率顯然比沒有LAP18過量表現者差(P=0.001)。
人類肝細胞癌中p53基因的突變常伴隨高侵襲性的腫瘤以及不好的預後，本研究中雖然有46%病例有p53基因突變，但LAP18的過量表現和p53的突變並沒有直接的相關性。但由於這兩種蛋白與細胞週期的調控都有關，為了解這兩種都會增加腫瘤侵襲性並造成不佳預後的蛋白間是否有某種程度的相互作用，因此我們也針對LAP18有過量表現(+)或沒有過量表現(-)以及p53基因有突變(+)或沒有突變(-)做了組合分析並將病例分成LAP18(+)/p53(+)(共41例)，LAP18(+)/p53(-)(共41例)，LAP18(-)/p53(+)(共32例)，以及 LAP18(-)/p53(-)(共44例)四組。結果發現，LAP18(+)/p53(+)最常見於有高分期有血管侵襲(Stage IIIA to IV，85%，P=0.00003.)、高分級分化不良(II~IV，90%，P=0.073)的腫瘤，而且腫瘤早期再發的機率也是最高(68%, P=0.0035)，甚至比只有p53突變者[LAP18(-)/p53(+)]差。
利用免疫組織染色法來偵測有LAP18 mRNA過量表現的病例中LAP18蛋白的表現時發覺，腫瘤細胞的細胞質中也有廣泛且濃厚的LAP18蛋白表現。這些LAP18蛋白過量表現的地方以位於肝癌組織小樑邊緣(borders of trabeculae)、腫瘤邊緣以及血管內腫瘤的血栓表現特別明顯；而在沒有LAP18 mRNA過量表現的腫瘤中最多只有散在性少量的細胞有LAP18蛋白的表現；鄰近非腫瘤組織中則最多也只有散在性少量細胞有LAP18蛋白的表現。本研究在追蹤的180個月中，178位病人(96.7%)追蹤超過七年或到其死亡，研究結束時仍有41位病人(22.3%)存活，因此用七年當追蹤的終點，結果發現有LAP18 mRNA過量表現者七年的存活率比沒有過量表現者差(P=0.001)。而在配對的研究中觀察LAP18與p53的相對影響發現，同時有LAP18 mRNA過量表現又有p53基因突變[LAP18(+)/p53(+)]的病例七年的存活率最差(P<0.016)。針對有p53突變的腫瘤中，若合併LAP18的過量表現[LAP18(+)/p53(+) versus LAP18(-) /p53(+)]，則多有血管侵襲的腫瘤(P=0.0025)以及預後較差(P<0.035)。至於沒有p53突變的腫瘤中，不論是LAP18(+)/p53(-)或是LAP18(-)/p53(-)，在血管侵犯以及預後上都沒有差異。但和LAP18(-)/p53(-)的病例相比，LAP18(+)/p53(+)常伴隨有血管侵襲 (Stage IIIA to IV，85% versus 34%，P<0.000002)、分化差(Grade II~IV，90% versus 68%，P=0.013)的腫瘤以及高的肝臟內早期腫瘤再發機率(68% versus 32%，P=0.0016)。以上研究顯示，於細胞增殖而表現的LAP18在p53正常的肝癌中與肝癌的進行關係不大，但是在p53突變的肝癌中則似與肝癌的血管侵襲關係密切，似乎為p53下游的標的物。

肝細胞移植之細胞增殖研究
肝臟機能衰竭的治療在早期即已分成三個方向，其一為肝臟的器官移植，其二為人工肝臟，其三為肝細胞移植。肝細胞移植的優點在：(1)若捐贈的肝臟不適合移植時，仍可分離出部分健康的肝細胞；(2)一個肝臟分離出的細胞可以捐贈給數名病人；(3)移植細胞經由靜脈導管輸入，其過程比肝臟移植簡單而安全；(4)可當作病人等待器官移植時的輔助治療；(5)移植的肝細胞可執行部分肝臟功能，延長病人生存的時間，讓原來的肝臟有再生的機會；(6)用於遺傳疾病或其它疾病的基因治療。此外細胞移植所需經費只要器官移植的十分之一，且由於肝細胞基因植入技術進步使得自體移植也不在是個問題，可降低免疫排斥上的問題，實用性不容忽視。
肝細胞移植是一種較不具侵入性的手術，但是肝細胞移植常常會碰到植入細胞在肝臟內無法持續繁殖生長。現階段的動物實驗多半是在細胞移植後利用部分肝臟切除或連續性的肝臟損傷，提供移植細胞繁殖的刺激，但是接受者的肝細胞也接受相同的訊號同樣繼續增殖，無法選擇性只讓移植的細胞繼續繁殖生長。目前應用在細胞移植研究而且結果非常令人滿意的動物模式有三，尿激素纖維蛋白溶酶原活化體基因植入小鼠實驗模式(uPA - urokinase-type plasminogen activator transgenic mouse )、延胡索醯乙醯乙酸基因剔除小鼠實驗模式(FAH - fumarylacetoacetate hydrolase null mouse)以及倒千里光鹼(retrorsine)/百分之七十部分肝臟切除(partial hepatectomy)的動物模式。前兩者是利用代謝所產生的毒物使得原有肝細胞不斷被破壞，移植細胞因代謝正常不會被破壞而佔有生長優勢。第三種實驗模式則是利用藥物抑制原有的接受者肝細胞再生的能力，使移植細胞在肝臟切除後有生長優勢取代不會分裂生長的原有肝細胞。這

些出生時即會致命以及利用毒物抑制原有肝細胞分裂的動物模式在人類並不適用，因此我們希望可以發展更接近人類疾病的動物模式，不但可以選擇性的增加移植細胞的增殖，將來又可能應用到人類疾病的治療。因此在肝細胞移植之細胞增殖研究的第一部分我們將移植具有增殖優勢的基因改造肝細胞，並與移植的正常的肝細胞做相互的對照，以比較在小鼠接受急性肝臟損傷、慢性肝臟損傷以及不做任何損傷等不同的動物模式下，兩種不同的移植細胞在細胞增殖能力以及肝臟重建(repopulation)上的差別。
我們在此部分所使用的移植細胞是不具增殖優勢的正常肝細胞(wild type hepatocytes)以及具有增殖生長優勢的基因改造細胞(p27-/- gene knockout hepatocytes)，而接受細胞移植的動物主要是DPPIV基因剔除鼠(DPPIV-/- gene knockout muce，又叫做CD26 knockout mice)以及DPPIV-/-/Rag2-/- 雙重基因剔除鼠(DPPIV-/-/Rag2-/- double gene knockout mice)。由於細胞移植研究需要辨別所植入的細胞，而文獻上使用的方法都很複雜，本研究則利用簡單的組織化學染色法來區別移植細胞。Dipeptidyl peptidase IV (DPPIV)，是一種外多胜酶(ectopeptidase)，可表現在許多不同的上皮組織包括腎臟、肝臟以及小腸，而DPPIV也是一種T淋巴球訊息傳導有關的細胞表面分化標記。DPPIV gene knockout mice是一種健康的老鼠，由於DPPIV也是肝細胞表面膽管小管(Canaliculi)側的酶，利用組織化學染色法可在正常小鼠肝細胞表面染出紅色膽管小管形態，而DPPIV突變小鼠則無此型態，因此移植正常肝細胞到DPPIV-/- gene knockout mice時，可以很容易分辨出染有紅色膽管小管形態的移植細胞，是追蹤移植細胞很好的標記。將C57BL/6小鼠(DPPIV+/+)的肝細胞移植到接受百分之七十肝臟切除手術的DPPIV-/-小鼠，在移植後一、二、三以及六個月後犧牲小鼠並採其肝臟做切片分析發現，移植細胞不但可以在DPPIV-/- gene knockout mice中存活，且可在細胞周圍膽管小管處清楚的表現DPPIV活性，並與原有肝細胞互相結合。植入後六個月仍可以清楚見到移植細胞，但移植細胞的擴充有限且在肝臟重建比例不到百分之一。由於移植細胞群大部分只含有一或二個細胞，極少數含有三個以上的細胞，這表示移植細胞嵌入接受百分之七十肝臟切除手術小鼠後，由於缺乏進一步的增殖刺激而停止生長，不過這部分的結果證明DPPIV小鼠實驗模式是可行的。
由於缺乏移植後的刺激使得移植細胞增殖受限，因此我們想要探討移植後的肝臟損傷對於移植細胞的增殖以及肝臟重建是否有幫助。本實驗將小鼠分成三組：第一組(對照組)只移植正常肝細胞；第二組(急性傷害組)除了移植正常細胞外並接受急性四氯化碳(CCl4)傷害；第三組(急性傷害合併慢性傷害組)除了移植正常細胞並接受急性CCl4傷害外並於每週接受一次的慢性CCl4傷害，並分別在四週以及八週犧牲。結果發現在移植細胞後四週，不論是第二組急性傷害組或是第三組急性傷害合併慢性傷害組，移植正常細胞造成肝臟重建都不到百分比之一，但是比第一組對照組有效率(P值分別是0.03及0.01)。但是CCl4慢性傷害的強化作用到了八週卻消失了(P值分別是0.09及0.18)。因此對於移植的正常細胞

而言，急性傷害以及慢性損傷是給予移植細胞初期增殖上的優勢，但繼續給予慢性傷害並無法有效增加移植細胞肝臟重建的比例，原因是它們與受贈者肝細胞的增殖能力相同，兩者有相同的分裂週期因此無法造成明顯的差別。
既然移植的正常細胞無法有效的增加肝臟重建，接下來要探討的是移植具有增殖優勢的基因改造細胞是否可以得到較好的肝臟重建。p27-/- gene knockout mice的肝細胞除了對於H-thymidine以及BrdU的鍵結能力增加，Cdk2-kinase的活性也增加，且利用p27-/- gene knockout mice肝細胞可以比正常細胞更有效的挽救FAH gene knockout mcie的肝臟傷害。p27kip1是週期素倚賴性激酶的抑制劑，可以和 Cyclin A以及Cyclin E/Cdk複合體穩定結合而抑制細胞的過度繁殖，是調控細胞週期的重要蛋白。由於p27-/- gene knockout mice與DPPIV-/- gene knockout mice的遺傳背景不相同，移植細胞會被排斥，因此要移植細胞必須培養出具免疫耐受性(immunotolerant)的DPPIV-/- gene knockout mice。活化重組基因第二型(Recombination- Activating Genes 2 -- Rag2) 是製造B淋巴球抗體以及T淋巴球接受器(T-cell receptor, TCR)成熟時在重組過程(VDJ rearrangement process)中所需的特殊重組酶(recombinase)。缺乏此酶的老鼠由於無法進行基因片段重組，因此沒有成熟的B淋巴球及T淋巴球，反而可以接受同種異體(allogeneic)以及異種異體(xenogeneic)的移植細胞，因此我們選定Rag2-/- gene knockout mice)與DPPIV-/- gene knockout mice交配出DPPIV-/-/Rag2-/- double gene knockout mice，此特殊培育的小鼠不但具有免疫耐受性，也和DPPIV-/- gene knockout mice相同可以很容易的利用DPPIV組織化學染色法來區分移植細胞。
移植p27-/- gene knockout肝細胞到只接受百分之七十肝臟切除手術的DPPIV-/-/Rag2-/- double gene knockout mice時可以見到移植細胞在小鼠肝臟中存活，植入後六個月也可以清楚見到移植細胞仍然存在。雖然在移植p27-/- gene knockout肝細胞的小鼠肝臟中偶而可以見到比移植正常細胞時較大的細胞群體積，但兩者造成肝臟重建的比例在統計上沒有顯著的差異。由於移植細胞群大部分也只含有一或二個細胞，結果與移植正常細胞類似，顯示即使移植具有增殖優勢的細胞，由於收贈者肝臟缺乏進一步的增殖刺激也同樣會停止增殖，這表示具有增殖優勢的細胞仍然受到正常肝臟中增殖的調控，不會自行無限制的增殖。
接著我們探討移植後的肝臟損傷對於增殖優勢細胞移植後的增殖以及肝臟重建是否有幫助，本部分將實驗小鼠也分成三組而延續前項實驗的編號：第四組(對照組)只移植p27-/- gene knockout肝細胞；第五組(急性傷害組)除了移植p27-/- gene knockout肝細胞外並給予急性CCl4傷害；第六組(急性傷害合併慢性傷害組)除了移植p27-/- gene knockout肝細胞給予急性CCl4傷害外並合併每週一次的慢性CCl4傷害，之後分別在四週以及八週犧牲。結果發現在移植p27-/- gene knockout 肝細胞四週後，不論是只有給予急性四氯化碳傷害(第五組)或是再加上慢性四氯化碳傷害(第六組)，移植細胞重建的比例都比只移植細胞不做任何傷害者來的有效率(第四組，P值分別是0.08及0.03)，甚至在加上慢性四氯化碳傷
害後肝臟重建比率可以多於百分比之一。與移植正常細胞相異而更有趣的結果是，合併慢性四氯化碳傷害者肝臟重建比例到了八週反而加強了(第六組versus 第四組，P值是0.01)，甚至顯著的比僅有急性傷害者好(第六組versus 第五組，P值是0.02)，且此重建能力的加強在八週時比四週時更明顯 (第六組，第八週versus第六組，第四週，P值為0.049)。
比較移植p27-/- gene knockout 肝細胞與正常細胞時發現，慢性肝臟傷害明顯的增加p27-/- gene knockout 肝細胞的增殖能力。對合併慢性四氯化碳傷害的小鼠來說，在移植p27-/- gene knockout 肝細胞(第六組)四週後，每100x視野中的細胞群數量雖然在統計上和移植正常小鼠肝細胞者(第三組)差異不顯著，但是數量卻增加了兩倍(第六組versus 第三組，P值為0.075)。到了八週時此種差異更加明顯(第六組versus第三組，P值為0.015)，幾乎達到移植正常小鼠肝細胞者的五倍。此外我們比較合併慢性四氯化碳傷害者細胞群數量在四週到八週的變化時發現，移植正常小鼠肝細胞(第三組)者每100x視野中的細胞群數量有減少的趨勢，但是移植p27-/- gene knockout肝細胞(第六組)者每100x視野中的細胞群數量反而有增加的趨勢。
在每個細胞群中平均所含細胞數量方面，移植正常肝細胞者在四週時大部分細胞群所含細胞數量是一個或兩個，此現象在八週時並沒有差異。但相對的移植p27-/- gene knockout肝細胞後，細胞數目介於三到五個的細胞群不論在四週或八週時，都比移植正常肝細胞者多得多。移植正常肝細胞者在八週時，含有六到十個細胞的最大細胞群只有四個，沒有一個細胞群所含的細胞數量大於十個細胞，但移植p27-/- gene knockout肝細胞者含有六到十個細胞的細胞群共有兩百六十八個，而細胞群所含的細胞數量大於十個細胞者甚至有一百一十個，因此移植p27-/- gene knockout肝細胞後，細胞數目大於十個的細胞群甚至佔了全部細胞群的8.16 %。
只針對合併慢性四氯化碳傷害的小鼠而言，若在所有的細胞群中剔除細胞數目只有一個或是兩個者，再加以計算每個細胞群平均所含的細胞數量發現，不論在四週或八週，移植p27-/- gene knockout肝細胞與移植正常肝細胞者在細胞群平均所含細胞數量上都有顯著的差異(P值分別為0.0459及0.0462)。尤其重要的是，移植p27-/- gene knockout肝細胞者最大的細胞群所含細胞數可達五十個細胞，而移植正常肝細胞者只有十個細胞。假設每一個細胞群是立體球型，且是由一個細胞所衍生而來時，我們可以估計移植p27-/- gene knockout肝細胞要得到最大的細胞群平均需分裂八次，而移植正常肝細胞得到最大的細胞群平均只需分裂五到六次。
我們也比較慢性四氯化碳傷害所造成的增殖刺激對於移植p27-/- gene knockout肝細胞與移植正常肝細胞在肝臟重建上的影響，結果發現不論是移植何種細胞，若不做任何肝臟傷害，不論四週或八週後兩者肝臟重建的比例並沒有差異。若加上急性肝臟傷害，肝臟重建的比例只有小幅度的增加，兩者的差異在統計上也不顯著。但當每週給予慢性低劑量四氯化碳刺激時，移植p27-/- gene knockout肝細胞後肝臟被取代的比例有強化的現象，但移植正常肝細胞則沒有變化。雖然在統計上差異不顯著(P = 0.14)，經過四週的低劑量四氯化碳肝臟損傷後移植p27-/- gene knockout肝細胞的肝臟被取代的比例是移植正常肝細胞的兩倍，而到了八週後移植p27-/- gene knockout肝細胞的肝臟被取代的比例是移植正常肝細胞的七倍，差異非常顯著(P = 0.036)。甚至在某些移植p27-/- gene knockout肝細胞的小鼠，經過八週後p27-/-細胞除了位於肝門周圍外也擴展到了肝小葉中間帶，且肝臟重建的比例更可以達到12-15%，而移植正常肝細胞者肝臟重建的比例最多也只有1-1.5%。此外移植正常肝細胞者且不論在有沒有急性或慢性四氯化碳傷害，移植細胞所佔全部細胞的比例都有減少的趨勢，移植p27-/- gene knockout肝細胞後若不給任何傷害，或只給急性傷害移植細胞所佔比例也有減少的趨勢，但若合併有慢性傷害者，移植細胞肝臟重建的比例反而有意義的增高。以上種種結果顯示p27-/- gene knockout肝細胞的增殖還是受到正常肝臟生長調控，但p27-/- gene knockout肝細胞比正常肝細胞存著較為強化的增殖能力，且這種增殖的差異在肝臟受到不斷的慢性傷害產生生長刺激時更加顯著。
由本論文先前的研究顯示，慢性肝臟傷害可以增加移植細胞對於肝臟的重建，而臨床上需要進行肝臟移植的人類肝臟疾病大部分在移植前都有慢性肝臟疾病的存在，因此細胞移植前的慢性肝臟損傷對於移植細胞的影響也是值得研究的。移植前的百分之七十肝臟切除以及急性四氯化碳傷害也可以增加移植細胞肝臟重建的比例，但是這些實驗模式並無法直接應用到人類，因此發展更符合人類肝臟疾病的實驗模式還是需要的。由於在肝細胞移植的動物模式中，當移植細胞進入肝臟時約有70~80%細胞會被肝臟巨噬細胞(Kupffer cell)所破壞，而也有報告減低Kupffer cell的功能而增加移植細胞肝臟重建的比例，因此我們在此也希望探討進行細胞移植時肝臟巨噬細胞所扮演的角色。在肝細胞移植之細胞增殖研究的第二部份(也是本論文的最後部分)是利用慢性四氯化碳傷害來模擬細胞移植前的慢性肝臟損傷，並利用人類經常使用且容易造成急性肝衰竭的藥物- acetaminophen當作移植前的急性肝臟損傷的模式，探討移植前的慢性肝臟損傷如何造成肝臟Kupffer cell功能上的改變而影響移植細胞增殖以及肝臟重建。
Acetaminophen 中毒是臨床上常見的問題，在英國因藥物中毒而住院的病人約有50%和acetaminophen有關；在美國約有10%，而且每年造成約八百例的急性肝衰竭，其中有三分之一因而死亡。Acetaminophen在肝臟中的代謝途徑主要是經由glucuronidation或sulphation，所得到含有劇毒性的結合物(non-toxic conjugates)可經由腎臟排泄。此外acetaminophen也可以經由cytochrome P-450系統代謝，造成高活性的代謝物N-acetyl-p- benzoquinoneimine與肝臟的蛋白結合而造成細胞壞死。在正常情況下N-acetylp-benzoquinoneimine的毒性會被自然界的解毒劑glutathione所中和。但是若服用過量的acetaminophen或因飲酒過量造成glutathione不足，便會導致肝臟centrilobular necrosis甚至致死。由acetaminophen所引起的急性衰竭包括肝衰竭、腎衰竭、低血壓、敗血症、凝固因子病變、腦病變以及腦水腫。通常在服藥過量的24小時內大部分的病人都可以利用N-acetylcysteine拮抗劑以內科療法治療，但對於出現急性肝衰竭的病人則必須立即評估，並決定是否需要肝臟移植來挽救。
這一部分實驗將小鼠分為兩組，第一組只接受Acetaminophen傷害之後移植細胞，第二組合併每週兩次的慢性CCl4肝臟損傷四週後加上Acetaminophen急性傷害再移植細胞，移植後第七天與第十天犧牲小鼠取其肝臟分析。結果發現移植細胞七天後，在移植前受到慢性肝臟損傷的小鼠體內移植細胞的細胞數量及細胞群數量都比只給acetaminophen者明顯的增加(P值分別是<0.001及<0.001)。而在十四天時兩者的細胞數量及細胞群數量也是有顯著的差異(P值分別是0.001及0.004)。但是觀察七天及十四天的結果發現，幾乎所有的細胞群都只含有一個到兩個細胞，而且移植前有慢性肝臟損傷者的細胞數量以及細胞群數量都有減少的趨勢，此證實移植前所給予的肝臟的損傷只能增加初期細胞的增殖，並無法持續增加進入肝臟後移植細胞的量。
我們的實驗證實肝臟細胞生長的調控不論在肝臟受到移植前或移植後的慢性傷害時仍然存在，移植具生長優勢的基因改造細胞確實可以得到較好的結果，但這些細胞還是受到肝臟的生長調控，但是有趣的是何以移植前受到慢性肝臟損傷的小鼠可以在移植後有較多的移植細胞。根據文獻的報告，當細胞進入肝臟後會有七八成細胞受到破壞。雖然破壞植入細胞的因素很多，但是肝臟內的巨噬細胞(kupffer cells)在破壞移植細胞中扮演了極為重要的角色，因此我們也希望探討慢性肝臟損傷是否對於肝臟巨噬細胞有某種程度的影響，直接或間接導致移植細胞的破壞減少使小鼠肝臟中有較多的移植細胞。
CD68 是一種透膜醣蛋白(transmembrane glycoprotein)，在人類的單核球以及組織巨噬細胞中有強烈的表現，利用CD68免疫螢光染色在未經過任何刺激的肝臟中可以見到少量的Kupffer cells散佈在肝臟中，經過慢性四氯化碳傷害後，Kupffer cells不但大量的增加而且大部分的細胞都集中在肝門區域 (poral area)。至於只有acetaminophen的傷害者，Kupffer cells的分佈與未經過任何刺激者相似但數量有增加，細胞的分佈較廣，並不侷限在肝門區域。至於先有慢性四氯化碳傷害再加上acetaminophen的傷害時，Kupffer cells細胞量雖然比只有acetaminophen傷害者多，但是和只有慢性四氯化碳傷害者相比相差並不多且分佈相似，似乎kupffer cells的活化並不時很明顯。為了探討Kupffer cell功能上的變化，我們在體外培養模式下利用脂多醣類(lipopolysaccharide，LPS)刺激kupffer cells，偵測TNF-αmRNA產生量的變化。結果發現合併慢性傷害者在LPS刺激前TNF-αmRNA的表現比只有急性acetaminophen傷害者高(P=0.012)，但是受到LPS刺激後合併慢性傷害者TNF-α mRNA的表現量反而比只有急性acetaminophen傷害者低(P=0.004)，也就是說移植前CCl4的慢性傷害顯著減緩了Kupffer cells受到LPS刺激時TNF-αmRNA的表現能力，至於LPS所產生的刺激是否相似於移植細胞所產生的刺激則需要進一步研究來證實。
我們的研究顯示REG1A、PAP以及LAP18的過量表現與肝細胞癌的進展、

侵襲以及病人預後有相關性。REG1A及PAP雖然與β-catenin基因的突變有關聯性但是REG1A和PAP對於肝細胞癌的進展以及轉移卻有相反作用，PAP減少腫瘤侵襲性而REG1A增加腫瘤侵襲性。LAP18的過量表現雖然與p53基因的突變無直接關係但兩者有協同加強作用，使得腫瘤具有較強的侵襲性。此研究不僅讓我們了解到不同基因可以在肝細胞癌中扮演不同的角色，也認知了組合分析的運用對於發覺不同族群的肝細胞癌是非常重要的，且認知了不同基因導致腫瘤不同的發展可以造成不同的預後。此篇論文提出許多肝細胞癌分子生物層面上新的發現，我們的觀察也提供了不同基因在肝細胞癌腫瘤發展中伴隨不同侵襲性的生物體內證據，此發現可以當作術後病人是否具有高度危險性的指標，也可作為將來發展抗癌藥物的參考，但是真正造成腫瘤發展與轉移的分子機制仍然需要進一步的探討，且不同基因與β-catenin和p53基因突變的關係以及作用機制也需要更進一步的研究。此外本論文對於肝細胞移植所面臨的困難有所探討，並應用p27-/-基因改造肝細胞進行動物活體移植研究。 p27-/- gene knockout肝細胞具有細胞增殖的優勢，因此得到比移植正常肝細胞更好的結果，這是直接在活體中觀察到應用基因改造小鼠肝細胞可以得到強化移植效果的成功報告，但是在考慮應用於人類疾病治療前，除了近程移植的治療效果外，遠程的影響如基因改造細胞是否有產生惡性腫瘤的可能也是必須因此需要長時間的追蹤才能了解的。另一方面利用藥物短暫抑制p27基因而給予移植細胞增殖優勢以取代原有的基因改造細胞，或設計攜帶自殺基因的基因改造肝細胞來作為移植的細胞，而當病人脫離險境甚至康復後再活化自殺基因破壞移植細胞以避免移植細胞癌化的疑慮也是一個可能性。最後我們也探討接受者生物體內影響移植細胞的重要因素，希望可以藉由降低受贈者破壞移植細胞的比例來提高肝臟重建的比例，達到應用於治療的目的。
本論文藉由細胞增殖能力的探討進行肝細胞癌及肝細胞移植兩方面的研究，提供基因治療及細胞治療新的方向，希望對於解決人類肝臟疾病的研究有所助益。

關鍵字

肝細胞癌 ； 肝細胞移植 ； 細胞增殖

並列摘要

Hepatocellular carcinoma (HCC) is one of the common malignancies worldwide, and the incidence was relatively higher in south China, Taiwan, southeastern Asia, and sub-Saharan Africa. According to the annual statistics by Department of Health, Executive Yuan, Taiwan, malignant disease is the number one cause of death in 2003 and 2004. Among them, HCC is the number one cause of death in 2003, and the secondary cause of death in 2004. In addition, chronic liver disease including chronic hepatitis, liver cirrhosis that finally resulting in hepatic failure is the number sixth cause of death in 2003, and the number seventh in 2004. Consequently, nearly ten percent of the people in Taiwan died of liver-related disease in the recently two years. At present, liver transplantation is the only available and effective method for treatment of end-staged liver diseases, but the organ donation is still limited. Therefore, researches on both HCC and liver transplantation are imperative, especially in our country.

Part One: Study of Cell Proliferation in Hepatocellular Carcinoma
Section One: Overexpression of REG1A and PAP in Human Hepatocellular Carcinoma

Even though many clinical studies have been made for prognostic prediction in HCC, the overall outcome of patients with HCC has not been completely changed, and specific prognostic indicators are still lacking. HCCs are genetically heterogeneous neoplasm and the genetic heterogeneity correlates with the variety of etiologic factors. Recent studies have shown that many genes that expressed aberrantly in the neoplastic transformation of liver cells had made the HCC differently in tumor invasiveness, metastatic potential, tumor recurrence and patient survival. These factors include p53 gene, angiogenic factors, cell cycle regulators, oncogenes and their receptors, apoptosis related factors, cellular proliferation markers, growth factors, proteinases that involved in the degradation of extracellular matrix, and adhesion molecules.
To identify the candidate genes that expressed differently in human HCC, suppressive subtraction hybridization and microarray techniques were used in this study, and several genes including regenerating islet-derived 1 alpha (REG1A), secreted phosphoprotein 1 (osteopontin), pancreatic lipase-related protein 2 (PNLIPRP2), defender against cell death 1 (DAD1), ATP-binding cassette protein C11 (ABCC11), glycine receptor alpha 1, apolipoprotein B (APOB), dihydropyrimidinase (DPYS), RB1-inducible coiled-coil 1 (RB1CC1), angiotensinogen proteinase inhibitor and so on were discovered. After literatures reviewing, the REG1A was selected to be studied in this study because it’s never been reported in human HCC before. In addition, another gene called pancreatitis-associated protein (PAP) that belongs to the same Reg gene family and located tandemly in the same chromosome region as REG1A was selected in this study because the PAP had been found overexpressed in human HCC but the clinical significance was till unknown. Therefore, in the first portion of this thesis, we studied the expression pattern of these two cell proliferation-related genes in human HCC, and correlated their expression pattern with the clinicopathological features of the patients.
Regenerating gene (Reg), which is expressed prominently in regenerating pancreatic islet, was first identified in the screening of regenerating islet-derived cDNA library taken from 90% depancreatized rat. The Reg gene family consists of a group of antiapoptotic factors or growth factors for pancreatic islet-cells, neural cells and epithelial cells, and was classified into four classes (I~IV) and contained 17 different members till now. In humans, there are two members in Reg I family, REG1A and REG1B. The REGIA gene encodes a 166-amino-acid protein with a 22 amino acids signal sequence, and is highly represented in human pancreatic secretion. The REG1A protein is identical to the pancreatic thread protein, pancreatic stone protein. The REG1B gene codes for a transcript with 87% homology to the REG1A transcript, but the Reg1B protein has never been characterized and its expression in the pancreas remains questionable. Pancreatitis-associated protein (PAP), a member belongs to Reg III family, is also called islet neogenesisassociated protein (INGAP), pancreatic β cell growth factor, na d REG-III. The human PAP cDNA encodes a 175-amino acid protein with 49% identity with the human Reg protein. PAP protein is merely detectable in normal pancreas, but remarkably increased, representing up to 5% of secreted protein in acute pancreatitis and in some chronic pancreatitis, and is also called hepatocarcinoma-intestine-pancreas (HIP) because it can also be detected in intestine and HCC.
In human, the REG1A, REG1B, RS (REG-related sequence), and PAP genes are clustered tandemly in a 95 kb region on chromosome 2p12, and this gene cluster may have arisen from the same ancestral gene by gene duplication. REG I mRNA was expressed in regenerating or hyperplastic pancreatic islets, and expressed high in high-grade biliary dysplasia in hepatolithiasis. REG I was expressed in colorectal cancer, cholangiocarcinoma, and was associated with the infiltrating growth of gastric carcinoma. PAP mRNA was overexpressed during the acute phase of pancreatitis, and can be detected in pituitary adenoma, gastric cancer, colorectal cancer, and was associated with nodal involvement, distant metastasis, and short survival in pancreatic cancer. These indicated that overexpression of REG1A and PAP correlated with invasiveness of some human malignancy. Even though PAP was detected in about a quarter of primary human HCC, the clinicopathological role of REG1A and PAP expression and their interaction in HCC is not clear. In the present study, we demonstrate that PAP expression is associated with a subset of HCCs that is often low-grade, low-stage tumor and shows high frequency of β-catenin mutation, whereas a coexpression of REG1A leads to more advanced disease and poor prognosis.
From January 1983 to December 1997, 1033 surgically resected primary unifocal HCCs were pathologically assessed at the National Taiwan University Hospital. Of these, 265 patients (209 males, 56 females, from 14 to 88 years old with mean age of 55.6 years) who already had mRNA samples taken from resected primary HCC were selected for this study randomly. The tumor grade was divided into three groups, well differentiated (grade I, 61 cases), moderately differentiated (grade II, 102 cases), and poorly differentiated (grade III and IV, 102 cases). HCC tends to spread in the liver via the portal vein invasion even in advanced stage. At the time of operation, no evidence of regional lymph node or distant metastasis was noted, and minute HCC (≦2 cm) has excellent prognosis. HCC with complete fibrous encapsulation has a favorable four-survival, and vascular invasion, the most crucial step of intrahepatic tumor metastasis, is a crucial prognostic factor for HCC. Therefore a modified tumor staging with special emphasis on the extent of vascular invasion, tumor size (≦2 cm or >2 cm), and encapsulation was adopted. Stage I to II HCCs had no vascular invasion, whereas stage III to IV HCCs had various extent of vascular invasion. Stage I HCC (6 case) included completely encapsulated minute HCC ≦2 cm with no liver invasion. Stage II HCC (115 cases) included minute HCC with liver invasion and/or microscopic satellite close to the main tumor; or larger HCC without or with liver invasion and/or minute satellite close to the main tumor. Stage IIIA HCC (43 cases) had invasion of thin-walled vessels in the tumor capsule, but no portal vain invasion or satellite deep in the liver parenchyma. Stage (33 cases) IIIB HCC had invasion of small portal vein in portal tract near the main tumor, but no invasion of major portal vein branch and satellite deep in the liver parenchyma. Stage IV HCC (68 cases) had invasion of major portal vein branches, satellites extending deeply into the surrounding liver, tumor rupture, or invasion of the adjacent organs. The intrahepatic tumor recurrence was based on imaging diagnosis with ultrasonography and/or computed tomography, supplemented with elevated serum α-fetoprotein. Among these patients, 236 (89.1%) were eligible for the evaluation of early intrahepatic tumor recurrence (≦1 year). Twenty-nine patients who died within 1 year after resection and had no information or were negative for intrahepatic tumor recurrence were excluded from the evaluation of early recurrence.
We used RT-PCR for large–scale analysis of PAP and REG1A mRNA levels in 265 unifocal primary HCCs. The primer sequence used was list in Table 1. PAP and REG1A were overexpressed in 97 (36.6%) and 55 (20.8%) tumors, respectively. Both genes were not detectable in 219 nontumorous liver tissues (Fig. 2). PAP overexpression showed a positive correlation with low α-fetoprotein (AFP) level (≦200ng/ml), and low-stage (stage I to II) HCCs that had no vascular invasion, P=0.039 and P=0.013, respectively, whereas REG1A overexpression did not (Table 2). The expression of PAP or REG1A did not correlate with age, gender, chronic hepatitis B infection, chronic hepatitis C infection, tumor size, tumor grade, or early tumor recurrence. In addition, p53 mutation was found in 99 out of 221 cases (44.8%), while β-catenin mutation was detected in 37 out of 248 tumors (14.9%). We found that PAP and REG1A expression showed strong association with β-catenin mutation (P<0.00001 and P=0.00005, respectively), but not with p53 mutation (Table 3).
Because of the frequent coexpression of PAP and REG1A (46 tumors expressed both PAP and REG1A, with a high concordance rate of 77.4%, P<0.00001), we did a combination analysis to further characterize the effects of PAP and REG1A expression in the tumor progression of HCC. We divided these cases into four groups according to presence or absence of PAP and REG1A overexpression: PAP(+)/REG1A(+), PAP(+)/REG1A(-), PAP(-)/REG1A(+), and PAP(-)/REG1A(-). Among the four groups, HCCs expressing PAP alone were associated with the highest frequencies of low-grade (grade I) and low-stage tumors, P<0.007 and P<0.001, respectively, and hence the lowest early tumor recurrence, P=0.051 (Table 4). Consistent with results shown in Table 2, there was also no significant difference in age, gender, α-fetoprotein level, and chronic hepatitis infection between these four groups (data not shown). In the two groups of HCCs with PAP expression, PAP(+)/REG1A(+) HCCs showed more frequent high-grade (84.8% versus 58.8%, P=0.005), high-stage tumors (60.9% versus 29.4%, P=0.002), and hence high early tumor recurrence rate (61.8% versus 31.3%, P=0.006) as compared with PAP(+)/REG1A(-) HCCs (Table 4).
To elucidate the reasons for the more aggressive tumors and early tumor recurrence in PAP(+)/REG1A(+) HCCs than in PAP(+)/REG1A(-) HCCs, we further analyzed the role of p53 and β-catenin mutations. As shown in Table 5, the former group showed significantly higher frequency of p53 mutation (46.2% versus 25.6%, P=0.036), whereas the two groups did not differ significantly inβ-catenin mutation. Then, we further analyzed the role of p53 andβ-catenin mutations in PAP(+)/REG1A(+) and PAP(+)/REG1A(-) HCCs. As shown in Table 6, PAP(+)/REG1A(+) HCCs tended to have more frequent high-stage tumors, but not of statistical significance, and the majority of HCCs withβ-catenin mutation had low-stage tumors, regardless of the presence or absence of REG1A.
Despite the frequent p53 andβ-catenin mutations, approximately half of HCCs are negative for both mutations. We then analyzed potential role of PAP and REG1A expression in this subset of HCC. In the subset of HCCs expressing PAP, the coexpression of REG1A had adverse effect in the groups of HCCs with PAP expression. As shown in Table 6, HCCs with coexpression of REG1A and PAP had more than 3-fold high-stage tumor than those with PAP expression alone (P<0.005). Finally, we analyzed the survival rates of the four different groups of HCCs. According to the expression pattern of PAP and REG1A, HCCs with PAP expression alone had the best 5-year survival, P=0.044, significantly better than HCC with coexpression of PAP and REG1A, P<0.0002 (Fig. 3).

Section Two: Overexpression of LAP18/Stathmin in Human Hepatocellular Carcinoma
In the second portion of the study, we analyzed another gene that also important for proliferation of the cells. Leukemia associated protein 18 (LAP18), also known as stathmin, p18, p19, pp17, prosolin, pp20, 19K, metablastin, and oncoprotein 18, is a ubiquitous cytosolic phosphoprotein, and its expression is regulated during the development in response to extracellular signals regulating cell proliferation, differentiation, and function. LAP18 is proposed to act as a general relay in signal transduction, possibly integrating diverse signals from the cell's environment and also plays an important role in the regulation of microtubule dynamics during the process of cell mitosis. LAP18 was first identified in HL60 leukemic cells, and was overexpressed in acute leukemia, but decreased expression when leukemic cells ceased to proliferate upon exposure to differentiation agents. Expression of LAP18 was most abundant in fetal and adult brain, spinal cord, and cerebellum, followed by thymus, bone marrow, and testis. Expression was intermediate in colon, ovary, placenta, uterus, and trachea, and at substantially lower levels in all other tissues examined. The fetal liver was among the tissue with most abundant LAP18 expression whereas the adult liver had the lowest. In a given tissue, LAP18 was expressed at higher levels in the proliferative compartment than the adjacent more differentiated cells. In addition to the leukemia, higher expression of LAP18 was noted in poorly differentiated solid tumors with high proliferative potential than more differentiated and less proliferative tumors. These observations suggest a strong correlation of LAP18 expression with cellular proliferation in both normal and malignant cells. To identify potential prognostic indicators of human HCC and potential targets for therapeutic strategies, we hypothesized that LAP18 may play a role during liver carcinogenesis. We show here that the LAP18 mRNA is overexpressed in nearly half of human HCC, and the expression correlates with vascular invasion, particularly when cooperation with p53 mutation in HCC.
One hundred and eighty four patients who had adequate and well-preserved RNA samples taken from resected primary HCCs receiving complete pathological assessment were selected for this study including 145 males and 39 females with a mean age of 55.5 years (range, 14-88 years). The tumor grade was divided into three groups, well differentiated (grade I, 39 cases), moderately differentiated (grade II, 79 cases), and poorly differentiated (grade III and IV, 66 cases), and the tumor staging was classified into I (1 case), II (76 cases), IIIA (30 cases), IIIB (28 cases), and IV (49 cases), as previously described. Among them, 165 (88.7%) were eligible for the evaluation of early intrahepatic tumor recurrence (≦1 year). Nineteen patients who died within 1 year after resection and had no information or were negative for intrahepatic tumor recurrence were excluded from the evaluation of early recurrence.
RT-PCR was used for large-scale analysis of LAP18 mRNA expression in HCC and the primer sequence was list in Table 1. Among the 184 unifocal primary HCCs, LAP18 mRNA was overexpressed in 95 tumor specimens (52%, Fig. 4). As shown in Table 7, LAP18 mRNA was overexpressed tended to occur in HCC with HBsAg in sera, P=0.06, but not correlate with age, gender or α-fetoprotein level. Histopathologically, LAP18 overexpression was found more frequently in HCCs with bigger tumor (> 5cm), less differentiation (grade II to IV), and high-stage (stage IIIA to IV) HCCs that had evidence of vascular invasion, P=0.0053, P=0.027 and P=0.0001, respectively. In addition, HCCs with LAP18 overexpression had more frequent early tumor recurrence (P=0.0022), and worse prognosis, with a lower 7-year survival rate than those without LAP18 overexpression, P=0.001 (Fig. 6).
By immunohistochemistry using specific antibody to detect LAP18 protein expression revealed that HCCs with LAP18 mRNA overexpression showed diffuse and intense predominantly cytoplasmic expression of LAP18 protein in the tumor cells. The LAP18-positive tumor cells were often more intensely immunostained at the outer layers of trabeculae, the tumor borders, and intravascular tumor thrombi (Fig. 5A-C). In contrast, LAP18 protein was absent or positive in scattered tumor cells in HCCs without the LAP18 mRNA overexpression (Fig. 5D). The adjacent nontumor livers also showed scattered immunoreactive liver cells with obvious less intense staining.
p53 mutation is the most commonly genetic alternation in HCC, and is known to correlate with tumor aggressiveness and unfavorable prognosis. In this series, p53 mutation was found in 73 out of 158 cases (46%), and p53 mutation showed a positive correlation with high-stage (stage IIIA to IV) HCCs than HCC without p53 mutation, 71% (52/73) versus 42% (36/85), P=0.00028. As shown in Table 7, however, LAP18 overexpression in HCC did not correlate with p53 mutation.
Both LAP18 and p53 were associated with tumor aggressive, and were also important in cell cycle regulation, we then did a combinational analysis to evaluate the potential interplay between these two important unfavorable prognostic factors. We divided these cases into four groups according to presence or absence of Lap18 overexpression and p53 mutation: LAP18(+)/p53(+), 41cases; LAP18(+)/p53(-), 41cases; LAP18(-)/p53(+), 32cases, and LAP18(-)/p53(-), 44cases. As shown in Table 8, LAP18(+)/p53(+) HCCs had the highest frequencies of vascular invasion (stage IIIA to IV) (85%) and early tumor recurrence (68%) than other groups, P=0.00003 and P=0.0035, and hence, the lowest 7-year survival, P<0.016 (Fig. 7). In HCC with p53 mutation, LAP18(+)/p53(+) HCCs had more frequent vascular invasion (stage IIIA~IV) and lower 7-year survival than LAP18(-)/p53(+) HCCs, P=0.0025 and P<0.035, respectively (Table 8; Fig. 7). In HCCs without p53 mutation, however, the frequencies of vascular invasion and prognosis did not differ significantly between HCC with and without LAP18 overexpression.

Part Two: Study of Cell Proliferation in Hepatocyte Transplantation
Section One: Transplantation of Hepatocyte with Augmented Proliferation Capacity

Currently, orthotopic liver transplantation using liver donated by cadaver or living people is the only available method for effective treatment of acquired and genetic-based liver diseases, especially when these diseases reach their end stages. With the limited availability of donor livers, alternative methods have been actively sought through liver cell transplantation. Studies conducted during the 1990s demonstrated that hepatocytes transplanted to the liver, either by injection into the spleen or by infusion through the portal vein, traverse the liver sinusoids and become incorporated into the liver parenchyma. Hepatocyte transplantation is beneficial because of its less invasiveness, easier to be manipulated, viable hepatocytes can be harvested from liver unsuitable for orthotopic liver transplantation, cells from single liver can be used for multiple recipients, and isolated cells can be frozen and distributed shortly when needed. However, the number of cells that can be introduced into the liver by this method is quite limited and engraftment levels are generally below 1.0%. Attempts to increase the proportion of transplanted hepatocytes in the host liver by stimulating liver regeneration are also quite limited because both host and transplanted hepatocytes respond equally to the proliferative stimulus. At present, three experimental models of extensive liver repopulation have been described: the urokinase-type plasminogen activator (uPA) transgenic mouse, the fumarylacetoacetate hydrolase (FAH) null mouse and the retrorsine / partial hepatectomy treated rat. However, in first two models, replacement of liver mass by transplanted hepatocytes is based on continuous and essentially lethal liver injury produced by the genetic disorder, and strong selection of transplanted wild hepatocytes that escape toxic substances damage can proliferate significantly. In the third model, the proliferation of resident hepatocytes were inhibited by toxic substance after partial hepatectomy that offered transplanted hepatocytes time and space for engraftment and proliferation until nearly total liver was replaced. But these animal models were not practical for routine use in human. Therefore, we design a new model to mimic human chronic liver injury with aggravation by acute insult and transplanted hepatocytes with augmented proliferation capacity to evaluate these particular cells in this disease process.
In studies of hepatocyte transplantation, it’s important to distinguish transplanted cells from resident hepatocytes. Many exogenous markers (fluorescent dyes, radioisotopes, et al.) and endogenous markers (transgenes, highly specific molecular probes, reporter genes, retroviral vectors, sex chromosome, et al.) had been used for this purpose, but it’s more convenience using Dipeptidyl peptidase IV (DPPIV) because it can be detected easily by histochemical staining. DPPIV is an ectopeptidase that can specifically cleave X-proline or X-alanine dipeptides from the NH2-terminus of several biological peptides such as the growth hormone-releasing hormone or substance P. Furthermore, DPPIV has been identified as CD26, a cell surface differentiation marker involved in signal transduction in the T-cell lineage. Targeted inactivation of the CD26 gene yielded healthy mice with normal blood level of glucose when fasted and wild hepatocytes containing normal DPPIV activity can be stained red in a membranous bile canalicular distribution pattern. Therefore, it provides as a marker in tracing the transplanted cells in DPPIV knockout mice (Fig. 8).
DPPIV expression was determined on 5-μm-thick cryostat sections from frozen tissue. Fixation was for 5 min in 95% ethanol/5% glacial acetic acid (99:1, v/v) at 0-10°C, followed by a 5-min wash in 95% ethanol at 4°C. Air-dried slides were incubated for 30-40 min at 37°C in the substrate reagent: 100 mg Gly-Pro-4-methoxy-B-naphthylamide (Sigma) dissolved in 6 ml of dimethyl-formamide and mixed with a 100 mg of Fast blue BB salt (Sigma) in 100 ml TMS (0.1 M Tris maleate, 0.1 M NaCI, pH 6.5). The slides were rinsed two times in TMS, incubated for 5 min in 0.1 M CuS04, and rinsed again in TMS. The slides were fixed for 10 min in cold 4% parafonnaldehyde in 0.15 M NaCI, washed two times in 0.15 M NaCI, rinsed in water, counterstained with hematoxylin, and mounted in glycerol.
Multiple sections from both the middle and left liver lobes were taken from each mouse and DPPIV histochemistry staining was performed. Donor hepatocytes were readily identified by DPPIV enzyme expression in a membranous bile canalicular distribution. Ten l00x fields were chosen randomly in sections from each lobe (20 fields were examined for each section) for determination of the number of DPPIV+/+ cell clusters per l00x field and the number of cells present in each cluster. The number of DPPIV4+/+ hepatocytes in each l00x field was then compared to the average total number of hepatocytes per l00x field to determine percent hepatocyte repopulation by transplanted cells. The slides were examined independently by two observers and scored in a blinded fashion by one of the authors.
In the beginning, we have to determine whether the DPPIV knockout mouse can be used as a model to facilitate identification of transplanted DPPIV4+/+ wild hepatocytes, similar to the naturally occurring Fischer 344 DPPIV-/- mutant rat, we transplanted hepatocytes from wild C57B1/6 mice into DPPIV knockout mice in conjunction with two thirds partial hepatectomy, then sacrificed 1,2,3,6 months later to evaluate the transplanted efficiency. As shown in Fig. 11, transplanted hepatocytes, expressing DPPIV in a bile canalicular pattern, engrafted into the periportal regions and became fully integrated into the parenchymal cord structure, and the transplanted cells can remain stable in the host liver for up to 6 months. The small size of DPPIV4 clusters (one, two. or rarely three cells in two-dimensional liver sections) suggested that after their initial engraftment, the transplanted cells underwent one or two divisions during recovery of the animals from partial hepatectomy, but there was no subsequent expansion of transplanted hepatocytes. Thus, the widely used DPPIV transplantation/detection system, originally developed in the rat, can also be used in mice.
In the following study, we attempts to increase the proportion of transplanted hepatocytes by stimulating liver regeneration through repeated CCl4-induced hepatic necrosis after hepatocyte transplantation. Wild type hepatocytes were isolated and transplanted through the spleen into the liver of recipients. The recipients were divided into three groups. Group 1 received wild hepatocyte transplantation only; group 2 received acute CCl4, (1.2 ml/kg mouse body weight) 24 h after wild hepatocyte transplantation; group 3 received acute CCl4 (1.2 ml/kg mouse body weight) 24 h after p27 null hepatocyte transplantation, followed by weekly injections of CCl4 (0.5 ml/kg mouse body weight) for 4 or 8 weeks. Recipients were sacrificed 1 week after the final treatment with CCl4 or vegetable oil only. Liver segments were flash-frozen for histology and DPPIV staining to quantitate donor cell engraftment and proliferation. Even though the repopulation was less than 1%, after 4 weeks, mice receiving either acute insult only (group 2, 0.34 ± 0.10%) or acute insult combined with chronic liver injury (group 3, 0.53 ± 0.23%) had significant liver repopulation than mice absence of a liver regenerative stimulus (group 1, 0.15 ± 0.08%，P = 0.03, and P =0.01, respectively). However, after 8 weeks, the differences became insignificant (group 2, 0.20 ± 0.16% and group 3, 0.46 ± 0.36% versus group 1, 0.05 ± 0.03%; P = 0.09, and P =0.18, respectively). This indicated that liver injury increased the initial engraftment of transplanted hepatocytes but did not augment the proliferation of transplanted hepatocytes.
Then, we transplanted cells capable of higher proliferation activity to evaluate the proliferation of this particular cells in the diseased liver. p27Kipl is a cyclin-dependent kinase inhibitor that stably interacts with cyclin A and cyclin E/Cdk complexes to negatively regulate cell proliferation. p27 null mice are about 25% larger than wild-type littermates and show hyperplasia and organomegaly of multiple organs, including the liver. Hepatocytes in the p27 null liver show accelerated entry into S phase after partial hepatectomy, and p27 inactivation also synergizes with the inactivation of another kinase inhibitor, pi 8, to stimulate hepatocyte proliferation after partial hepatectomy. These results are consistent with previous studies by Karnezis et al., who demonstrated increased [3H]thymidine and BrdU incorporation into DNA in primary hepatocytes from p27 null mice compared with eild hepatocytes.
Because p27 null mice and DPPIV null mice are on different genetic backgrounds, these studies also required development of an immunotolerant DPPIV null mouse capable of accepting allogeneic hepatocytes. This was accomplished by breeding DPPIV null mice with Rag2 null mice, which are deficient in T and B cells due to inability in VDJ rearrangement, and are capable of accepting both allogeneic and xenogeneic cells. DPPIV-/- mice were crossed with Rag2-/- mice to produce heterozygous mice for both DPPIV and Rag2 gene deletions. The double heterozygous mice were mated together to produce either DPPIV-/-Rag2+/- or DPPIV+/- Rag2-/- mice, which were then crossed with each other to generate DPPIV-/- /Rag2-/- double knockout mice. Fig. 12 shows that DPPIV-/- /Rag2-/- mice accepted hepatocytes from various rodent species, including 129sv x C57B1/6 mice, 129sv x CD-1 mice, Balb c mice, and Sprague-Dawley rats, and the transplanted cells can also be identified by DPPIV histochemical staining.
To compare the study that transplanted wild hepatocyte in conjunction with partial hepatectomy, we transplanted P27-/- null hepatocytes into DPPIV-/-/Rag2-/- mice in conjunction with two thirds partial hepatectomy at the time of cell transplantation and followed proliferation of transplanted cells for 6 months (Fig. 13). In contrast to cessation of proliferation by wild hepatocytes in DPPIV-/- mice after 1 month (Fig. 13A), there was a modest increase in the size of DPPIV+/+ cell clusters at 2-3 months after P27+/+ hepatocyte transplantation (Fig. 13B, C), and although infrequent, larger clusters were clearly evident at 6 months after cell transplantation (Fig. 13D).
Thereafter, we attempts to increase the proportion of transplanted hepatocytes by stimulating liver regeneration through repeated CCl4-induced hepatic necrosis after hepatocyte transplantation. For convenience to compare the groups transplanted with p27 null hepatocytes and wild hepatocytes, we divided the mice into three groups and defined as: group 4, received p27 null hepatocyte transplantation only; group 5, received acute CCl4, (1.2 ml/kg mouse body weight) 24 h after p27 null hepatocyte transplantation; and group 6, received acute CCl4 (1.2 ml/kg mouse body weight) 24 h after p27 null hepatocyte transplantation, followed by weekly injections of CCl4 (0.5 ml/kg mouse body weight) for 4 or 8 weeks. Liver segments were also flash-frozen for histology and DPPIV staining to quantitate donor cell engraftment and proliferation.
We found that when an acute liver regenerative stimulus was included at the time of cell transplantation, the P27-/- null hepatocytes repopulated and formed larger clusters than hepatocytes without any regeneration stimulus after 4 weeks (group 5, 0.48 ± 0.23% versus group 4, 0.18 ± 0.17%) and 8 weeks (group 5, 0.40 ± 0.25% versus group 4, 0.10 ± 0.13%), although the differences were small and not statistically significant (P = 0.08, and P =0.15, respectively). However, transplanted cells proliferated and repopulated significantly better in liver with acute liver insult and chronic liver injury than in a normal quiescent liver after 4 weeks (group 6, 1.27 ± 0.76% versus group 4, 0.18 ± 0.17%; P = 0.03), and even better after 8 weeks (group 6, 3.27 ± 2.06% versus group 4, 0.10 ± 0.13%，P = 0.01). In addition, in groups with acute liver insult, P27-/- null hepatocytes repopulated better in those with chronic liver injury after 8 weeks (group 6, 3.27 ± 2.06% versus group 5, 0.40 ± 0.25%，P=0.02) than those with acute insult only, and in mice received acute insult with chronic liver injury, the repopulation of liver increased significantly from 4 weeks to 8 weeks(1.27 ± 0.76%versus 3.27 ± 2.06%, P=0.049).
Further comparison was made between transplantation of p27-/- null and wild hepatocytes. After four weekly doses of CCl4 (Fig. 14), there was no significant increase in the number or size of DPPIV+/+ clusters in livers of mice transplanted with wild hepatocytes (Table 9, Fig. 14A). In contrast, however, after four weekly doses of CCl4, the livers of mice transplanted with p27-/- null hepatocytes showed an increase in both the number and size of DPPIV4 clusters (Table 9. Fig. 14C), and the total number of cells in 100x field increased by two-fold, although the data did not reach statistical significance because of variations between individual animals (22.13 ± 9.00 versus10.27 ± 2.31, P = 0.075). Four additional treatments with CCl4 (a total of eight weekly doses) markedly increased the number and size of DPPIV4 clusters produced by p27 null hepatocytes (Table 9, Fig. 14D), but no such enhancement was observed with p27 wt hepatocytes (Table 9, Fig. 14B). These differences were clearly statistically significant (35.58 ± 16.10 versus 7.53 ± 5.56 in 100x field, P = 0.015).
To quantitate the proliferation of transplanted cells after repeated liver injury, 20 random fields at l00x magnification were examined from multiple sections of each lobe of the liver to determine the average number of clusters per field and the average number of cells per cluster. As showed in Fig. 15, even though the majority of clusters contained one or two cells with both p27 null and wt hepatocytes, the percentage of clusters containing three or more cells was greater with p27 null hepatocytes after four weekly treatments with CCl4 (Fig. 15A). The difference in cluster size between p27 null and wild hepatocytes became much more dramatic after eight weekly treatments with CCl4. Under these conditions, the largest clusters (four in total) derived from wild hepatocytes contained 6-10 cells and none contained more than 10 cells (Fig. 15B). Under comparable circumstances with p27 null hepatocytes, 268 clusters contained 6-10 cells and 110 clusters contained more than 10 cells (Fig. 15B). In addition, the largest clusters derived from p27 null hepatocytes contained 40-50 cells compared with 6-10 cells with p27 wt hepatocytes (cf. Fig. 14D and B). Assuming that each cluster is spherical and derived from a single transplanted cell, the largest clusters with p27 null hepatocytes represent an average of eight cell divisions, whereas those from p27 wt hepatocytes represent an average of five to six cell divisions. These results demonstrate that transplanted p27 null hepatocytes exhibit augmented proliferative activity compared with transplanted wild hepatocytes in the normal host liver and this difference is cumulative during the regenerative response to repeated liver injury.
Liver repopulation by transplanted hepatocytes was measured too. With transplanted hepatocytes alone (quiescent liver), there was no difference in the percent repopulation between p27 null and wt hepatocytes, either at 4 or 8 weeks after cell transplantation. Addition of a single acute liver regenerative stimulus slightly increased liver repopulation by transplanted cells, but again there was no statistically significant difference between p27 null and wild hepatocytes. However, weekly repeated small doses of CCl4 augmented liver repopulation by p27 null hepatocytes versus wild hepatocytes (Fig. 16). After four weekly cycles of CCl4 administration, the extent of liver repopulation by p27 null hepatocytes was more than double that observed with wt hepatocytes, but the difference was still not statistically significant (P=0.14). After eight weekly cycles of CC4 administration, there was a sevenfold greater liver repopulation by p27 null hepatocytes (3.27%, range 1.2-6.1%) compared with wild hepatocytes (0.46%, range 0.04-0.63%) that was highly statistically significant (p= 0.036). In selected areas of high liver repopulation after 8 weeks of CC l4 treatment (predominantly in the periportal and midzonal regions), p27 null hepatocytes replaced up to 12-15% of total hepatic mass compared with 1-1.5% with wild hepatocytes. These results all indicated that p27 null hepatocyte still under normal control in regenerated liver and can not proliferate spontaneously and indefinitely after liver regeneration finished. On the other hand, chronic liver injury enhanced the proliferation capacity of the p27 null hepatocytes, resulted in augmentation of liver repopulation.

Section Two: Chronic Liver Injury before Cell Transplantation Attenuate Function of Hepatic Kupffer cell and Enhanced Liver Repopulation

Last but not least, we want to find out the effect of preceding chronic liver disease on the proliferation capacity of transplanted cells. In the mean time, we developed an animal model using acetaminophen, a very common analgesic, to induce acute liver injury because acetaminophen poisoning is very common in human. In the UK 50% of poisoning admissions involve acetaminophen. In the United State, acetaminophen intoxication accounts for nearer 10% poisoning admissions and that resulted in acute liver failure in as many as 800 people, and one-third of whom died every year. The major pathways of acetaminophen in liver include glucuronidation or sulphation, and the resulting non-toxic conjugates excreted by the kidney. In addition, another pathway involves the cytochrome P-450 (CYP) system, especially CYP2E1, by which acetaminophen is metabolized to the highly reactive metabolite N-acetyl-p- benzoquinoneimine, that may bind covalently with hepatic proteins causing cellular necrosis. The toxic effect of N-acetylp-benzoquinoneimine can be eliminated by the natural antidote glutathione. When taken overdose, acetaminophen will cause a potentially fatal, hepatic centrilobular necrosis. Acute liver failure induced by acetaminophen is a multisystem disorder, with acute renal failure, hypotension, sepsis, coagulopathy, encephalopathy and cerebral edema. Patients who present within 24 h of an acetaminophen overdose can be managed medically. But for those who developed acute liver failure, meticulous supportive care and transplantation is considered for those patients who will die without liver transplantation.
In the fourth portion, we transplanted wild hepatocytes into acetaminophen intoxicated mice without (group 1) or with (group 2) repeated CCl4 injury twice a week for 4 weeks before cell transplantation, and then to sacrifice the mice 7 and 14 days after the transplantation to evaluate the repopulation capacity of transplanted cells.
We found that after chronic repeated liver regenerative stimulus before cell transplantation, number and cluster number of the transplanted cell in 100x field were larger than those without repeated regenerative stimulus at 7 days (21.53 ± 6.29 versus 8.25 ± 3.98, and 14.86 ± 4.62 versus 6.59 ± 3.39; P<0.001 and P<0.001, respectively; Fig. 17, Table 10). The cell number and cluster number remained larger significantly in group with repeated CCl4 injury at 14 days too (17.29 ± 4.40 versus 10.16 ± 3.31, 12.59 ± 3.19 versus 7.90 ± 1.58; P=0.001, and P=0.004, respectively, Table 10), and therefore better hepatocyte repopulation (0.44% versus 0.16% at 7 days, and 0.35% versus 0.20% at 14 days). However, the cell number, the cluster number and repopulation decreased from 7 days to 14 days in liver with repeated CCl4 injury.
Even though the percentage of clusters containing three or more cells was greater in liver with repeated chronic injury at 7 days (P=0.007), the majority of clusters still contained only one or two cells, and the percentage of cluster containing three or more cells were similar 14 days after transplantation. This indicated that repeated chronic liver injury increased engraftment and stimulated initial proliferation of transplanted cells but the proliferation ability is limited after hepatocyte transplantation without further stimulation.
After cell transplantation, the transplanted cell deposited in the hepatic sinusoid, and the Kupffer cells are activated shortly. Therefore, a significant fraction of transplanted cells were destroyed especially when the transplanted cells are entrapped in portal spaces. It’s also reported that perturbation of Kupffer cells activity might increase engraftment of transplanted cell. To elucidate the role of hepatic macrophage on the engraftment of transplanted cell, we use fluorescence immunohistochemical staining to assess the distribution, and use lipopolysaccharide (LPS) stimulation to evaluate the activity of the hepatic macrophage. The hepatic macrophage distributed scatter throughout the plate in quiescent liver. After acute acetaminophen intoxication, the number of macrophage increased but still exhibited generalized distribution. After repeated chronic liver injury, the hepatic macrophages increased and were distributed prominently in the portal area, and even high up after additional acetaminophen insult (Fig. 18). Before LPS stimulation, the TNF-α mRNA producing ability of the Kupffer cells isolated from mice treated with acute on chronic liver injury was prominent than those from acute injury only (TNF-αmRNA/GAPDH level, 117.8 ± 12.3 versus 73.9 ± 12.4, P=0.012). But after LPS stimulation, macrophages isolated from mice that treated with acute on chronic liver injury had the TNF-α mRNA producing ability significantly lower than those from mice receiving acute acetaminophen only (TNF-αmRNA/GAPDH level, 175.7 ± 54.6 versus 465.6 ± 54.2, P=0.004: Fig. 20). It indicated that chronic repeated liver injury before cell transplantation attenuated the function of hepatic Kupffer cells, and might contribute to increased engraftment of transplanted cells. But the molecular relationship between transplanted cell engraftment and Kupffer cell function deserved further evaluation.
In this study, we demonstrated that REG1A, PAP and LAP18 were all correlated with cell proliferation and tumor invasiveness in hepatocellular carcinoma. We showed that both REG1A and PAP were correlated well with β-catenin mutation, but HCCs with PAP expression alone had the highest frequencies of well differentiated and low-stage tumors, the least frequent early tumor recurrence, and the best 5-year survival, while HCC coexpressed with REG1A is associated with more advanced disease and may contribute to tumor progression. In addition, we showed that LAP18 overexpression was associated with bigger tumor, high-grade, high-stage HCC with vascular invasion, more frequent early tumor recurrence, and worse prognosis. These findings suggest that LAP18 confers growth advantage and facilitates tumor cell growth and invasion capacity. These observations provide novel in vivo evidence of PAP, REG1A and LAP18 on cell proliferation capacity in hepatocellular carcinoma, and might offer specific targets for therapeutic strategies concerning cell proliferation in hepatocellular carcinoma.
Furthermore, we demonstrated augmented hepatocyte transplantation with genetic alternated hepatocytes in vivo. Our results revealed that, p27 null hepatocytes did not differ from their wild counterpart in their engraftment or proliferative activity in quiescent livers, and exhibited a slightly proliferative advantage under an acute proliferative stimulus. This suggests that once the liver mass has returned to normal, the p27 null hepatocytes cannot overcome the normal growth control, but the genetic altered hepatocytes can proliferate significantly after repeated liver regeneration stimulus. However, uncontrolled proliferation often impairs cellular differentiation and can pose unacceptable tumor risk. Therefore, a delicate balance should be achieved that would augment cellular proliferation and benefit transplanted hepatocytes, but would not increase tumor risk by the transplanted cells.
Finally, we demonstrate that chronic liver disease increased engraftment of transplanted cells and this phenomenon might result from attenuation of resident Kupffer cells in the recipient liver. This indicated that instead of doing harm for the patient, chronic liver disease might be beneficial for transplanted cells to be engrafted and repopulated in the liver. Additional studies to attenuate the function of hepatic Kupffer cells and increase the repopulation of transplanted cells are mandatory in the near future.

並列關鍵字

hepatocellular carcinoma ； hepatocyte transplantion ； cell proliferation

參考文獻

Adnane J, Jackson RJ, Nicosia SV, Cantor AB, Pledger WJ, Sebti SM. Loss of p21WAF1/CIP1 accelerates Ras oncogenesis in a transgenic/knockout mammary cancer model. Oncogene 2000;19:5338-5347.

Ahmad TA, Eguchi S, Yanaga K, Miyamoto S, Kamohara Y, Fujioka H, Furui J, Kanematsu T. Role of intrasplenic hepatocyte transplantation in improving survival and liver regeneration after hepatic resection in cirrhotic rats. Cell Transplant 2002;11:399-402.

Ahn J, Murphy M, Kratowicz S, Wang A, Levine AJ, George DL. Down-regulation of the stathmin/Op18 and FKBP25 genes following p53 induction. Oncogene 1999;18:5954-5958.

Alaiya AA, Franzen B, Fujioka K, Moberger B, Schedvins K, Silfversvard C, Linder S, Auer G. Phenotypic analysis of ovarian carcinoma: polypeptide expression in benign, borderline and malignant tumors. Int J Cancer 1997;73:678-683.

Alison MR. Regeneration of liver growth. Physiol Rev 1986;66:499-541.