摘要
膽汁滯留是常見的肝臟疾病起因，可能由先天或後天疾病引起。嬰兒膽汁滯留，定義為嬰兒結合性膽紅素上升，約佔各國醫學中心兒童肝膽疾病的百分之五十，在台灣也是常見的兒童肝臟疾病。根據過去的統計，本國嬰兒得到肝外膽道閉鎖的機率比日本、菲律賓以及西方國家的嬰兒都來的高。而肝內膽汁滯留發生的機率又比肝外膽道閉鎖為高。嬰兒膽汁滯留發生的年紀常在新生嬰兒兩個月之內，臨床表現包括黃疸、淺黃色或灰白色大便、體重過輕、餵食問題以及肝脾腫大。 嬰兒膽汁滯留包含肝外膽汁滯留及肝內膽汁滯留兩種。肝外膽汁滯留有膽道閉鎖，總膽管囊腫，及膽汁阻塞症候群。而肝內膽汁滯留包括非常複雜的原因，通稱為新生兒肝炎症候群。其發生原因包括先天性感染(TORCHES)，染色體異常(18及21對染色體異常)，內分泌異常(腦下垂體低下症，甲狀腺低下症)，肝內膽道構造異常(新生兒硬化性膽道炎，肝內膽道稀少症)，基因性疾病(進行性家族性肝內膽汁滯留)，以及代謝性疾病(肝醣貯積症，溶小體異常，先天性膽汁生成異常，α1 antitrypsin deficiency等)。 由於造成嬰兒膽汁滯留的原因眾多且複雜，因此在診斷膽汁滯留的嬰兒時，檢查及診斷是一項非常困難的工作，但是能夠正確且快速的診斷這樣的病人，對於及時給予治療非常重要，不但可根據疾病給予適當藥物及手術治療，也確保嬰兒可以有正常的生長及發育。 在台大醫院的經驗，過去張美惠教授曾經報告新生兒肝炎的病例分析，在五十六名罹患新生兒肝炎的嬰兒當中，有二十二名(49%)是與巨細胞病毒感染有關。而大約有一半的嬰兒膽汁滯留患者，雖然經過詳細的檢查，仍然無法發現確切的病因。在西方國家有許多常見的代謝性或先天性疾病，包括α1-antitrypsin缺乏以及tyrosinemia，在我國病人雖然經過詳細檢查，但都沒有發現病例，可見人種的差異對於新生兒肝炎或肝內膽汁滯留的病因還是有很大的影響。在台大醫院過去的報告中，百分之七十的病例在六個月之內都會痊癒，約有百分之十四的病人其復原速度較慢，大於六個月以上，另外有百分之十六的病人其癒後較差，包括肝硬化，肝衰竭，以及死亡。 進行性家族性肝內膽汁滯留(PFIC)是一群遺傳性、體染色體隱性遺傳的疾病，它通常發病於嬰兒時期，以膽汁滯留為臨床表徵。在短短的幾個月到幾年間，病人就會發展至肝硬化，肝衰竭，或死亡的情形。PFIC最早是在美國的Amish族群所發現，一開始被報告的病人，大部分都有血清中γ-glutamyl transpeptidase (GGT)低下或正常的現象，血清中膽固醇正常，但在膽汁中的膽酸濃度低下。這些疾病的基因目前已經被發現有三種基因可造成PFIC，第一型及第二型分別由Familial intrahepatic cholestasis 1 (FIC1)以及Bile salt export pump (BSEP)基因所造成，以上兩種病人血清GGT是低下或正常，第三種進行性家族性肝內膽汁滯留是由Multi drug resistant gene 3 (MDR3)基因的變異所造成，這類病人血清中GGT是升高的，MDR3基因變異的病人，他們膽汁中的磷脂非常的低，在兒童早期就會發生膽汁性肝硬化的情形。 從1980到1998年，台大醫院總共有295名因肝內膽汁滯留而住院的嬰兒，其中有46名(15.6%)慢性肝內膽汁滯留，所有病人均從嬰兒期發病(發病年齡都在一歲以內)，並且疾病達六個月以上。 在46個病人當中，Alagille syndrome有十八名，先天性膽酸生成異常有四位，新生兒杜賓強森徵候群有兩位，以及tyrosinemia有一位，在其餘的21名病人當中並未發現任何已知的病因，其中有6位病人血清GGT濃度上升 (>94 U/L，本院嬰兒正常值<94 U/L)，以及15位嬰兒有正常的血清GGT濃度。(表1) 在二十一名不明原因的慢性肝內膽汁滯留病人當中，我們分為high GGT 六位及low GGT 二十一位病人。在high GGT 6位病人我們分析MDR3的cDNA sequencing，cDNA是由肝臟細胞抽取經反轉錄PCR之後直接定序，在六名病人當中發現有一名病人有MDR3的mutation，這個病人有一個homozygotes 719bp deletion (nucleotide 287 to 1005)，這個病人的疾病發生年齡是一歲，其他五名病人都是以新生兒膽汁滯留症為起始表現，其中四名病人包括一對姐弟，他們的肝臟病理學檢查都有很明顯的門脈纖維化並進行到肝硬化，這樣的病理變化和MDR3 mutation的病人無法區分，但是這四名病人都沒有發現MDR3的基因變異，而且分析一名病人的膽汁也發現膽汁內的磷脂濃度正常，因此我們發現MDR3基因變異僅佔high GGT的慢性肝內膽汁滯留的一小部份，仍有一大部份病人具有high GGT的PFIC的表現型，並且很早期就發生膽汁滯留疾病，可能是由於未知的基因變異所造成的。 另外在low GGT的病人方面，我們以病理變化及AFP分成兩組，第一組病人(有五名)他們的病理表現是bland cholestasis，而第二組病人(有八位)他們的病理表現是巨細胞變化，在臨床上第二組病人表現出較高的肝內轉氨酵素及α-fetoprotein，第二組病人死亡率較高。其中有七名病人進行FIC1的基因鑑定，在第一組病人當中有四名病人接受基因檢查，四名病人全部都出現FIC1的基因變異，包括一個74bp deletion，一個98bp deletion，一個nonsense及二個missense mutation，在另外三名沒有發現FIC1基因變異的病人當中，只有兩名發現有BSEP的基因變異，包括二個missense mutation以及一個1bp deletion。因此我們發現在low GGT PFIC的病人，FIC1以及BSEP的基因變異約佔了86% (6/7)，而表現型的特徵有助於來區分FIC1或BSEP的基因變異。 第二部份，我們針對表現為新生兒黃疸的杜賓-強森症候群病人進行MRP2基因的分析，杜賓-強森症候群是一種慢性但良性的高膽紅素血症，通常是在青春期或是成人期被診斷，肝臟內有黑色色素的堆積，但是新生兒的病例非常罕見。我們搜集了四位新生兒黃疸病人，後來證實為杜賓-強森症候群的病人，抽取他們的肝臟cDNA進行全段sequencing的定序，我們發現所有四名病人都有MRP2基因的變異，三名是compound heterozygous，一名是heterozygous mutation，因此證實了新生兒杜賓-強森症候群是由MRP2基因變異造成。並且我們追蹤一名病人達二十年之久，發現這樣的病人，在新生兒出現黃疸之後將會慢慢的恢復正常，一直到青春期之後膽紅素再度上升。比較其它報告的MRP2 mutation，我們發現在較早期出現症狀的病人，他的mutation大多發生在二個功能上重要的ATP-binding cassette proteins上面，因此基因型的嚴重度跟表現型似乎有相關。 第三部份，我們分析在發育過程當中膽小管傳送蛋白的表現，在臨床上發現新生兒及早產兒極易發生膽汁滯留，在生理學上他們的膽汁分泌也尚未達成熟的程度，我們取人類胎兒14到20周的肝臟，與成人肝臟做比較，以real-time PCR偵測BSEP、PFIC1、MRP2、NTCP、MDR3、FXR的表現量，並以免疫組織染色來定位這些蛋白在細胞中的位置，發現胎兒所有的基因表現都比成人來的低，平均表現量BSEP、MRP2、FIC1介於成人標準的30到50%左右，FXR佔成人的75%左右，MDR3及NTCP僅在成人的6.4%及1.8%，另外我們也比較了膽道閉鎖病人的基因表現量，一到二個月大接受手術的膽道閉鎖病人的肝臟，基因表現除了NTCP之外，都較成人標準為高，顯示在膽汁滯留的情況之下這些基因可能有向上調控的情形。另外在免疫組織染色裡面發現胎兒肝臟的免疫組織染色特徵和成人肝臟差異很大，在成人肝臟組織可以發現明顯的BSEP、MRP2、及MDR3的膽小管染色，這是在肝細胞膽小管細胞膜上的線狀表現，而在胎兒肝臟發現這些蛋白的染色和成人不同，MRP2最接近成人，BSEP部分在細胞膜上，部分在細胞內，而MDR3的染色最淡。這些染色在胎兒肝臟中的血液生成細胞並沒有被染色，因此發現在胎兒期間大部份的膽小管傳送蛋白都已經有所表現，但表現量較成人為低，且在肝細胞當中他們並沒有完全被定位到正確的膽小管細胞膜的位置，也因此可能影響了這些傳送蛋白的功能。 第四部份，我們研究肝細胞培養中膽小管傳送蛋白的調控機制，我們以半定量RT-PCR方法，偵測膽小管傳送蛋白在primary hepatocyte culture裡面的表現，發現在肝細胞培養初期，這些傳送蛋白都有表現，約在培養五到七天之後漸漸的降低，到了兩個禮拜之後完全沒有表現，這時如果加上不同的膽酸刺激，BSEP、FIC1、MDR3、MRP2、NTCP都會被表現出來，因此我們認為膽酸的濃度可以調控這些膽小管傳送蛋白的表現。 第五部份，PFIC2動物模式(sPgp knockout mice)的細胞治療，sPgp knockout mice是PFIC2(BSEP基因變異)的動物模式，這樣的老鼠會出現肝腫大以及慢性肝內膽汁滯留的表現型，我們分別以肝臟細胞以及骨髓移植兩種方法進行細胞治療，並以EGFP轉殖小鼠(螢光鼠)來做捐贈者，結果在五隻knockout mice當中有兩隻出現PCR的陽性反應，每隻小鼠的四個肝臟組織當中各有一個出現PCR的陽性反應，在冷凍切片可以發現有少數散在性的陽性螢光細胞，但並未形成聚集，另外在骨髓移植的細胞當中可以看到大多數的螢光細胞是肝內的巨噬細胞，但是仍有散在性的肝細胞出現，高倍下可看見這些細胞呈多角形，是成熟肝臟細胞的形狀，我們以real-time PCR來偵測捐贈者的DNA在受贈的knockout mice肝臟裡面，其sPgp DNA的比例約佔細胞的百分之十左右，我們計算出捐贈者肝細胞在所有肝臟細胞的比例，發現sPgp knockout mice平均3211的細胞有一個是由捐贈者來，而在野生型平均15662個細胞裡有一個肝臟細胞是捐贈者細胞。因此我們發現在細胞移植方面，sPgp knockout mice的肝臟損傷可能並未到達致死程度，且受贈者肝細胞本身再生能力很強，可能是移植後捐贈者細胞不易再生的原因，在骨髓移植細胞方面，我們發現在受贈者肝臟裡可以發現少數的肝細胞，可能是肝細胞和骨髓細胞的融合，將來我們的努力方向希望增加受贈者正常細胞的數目，並證實這些正常細胞可以有功能性，並了解這些細胞是否可以改善肝臟的功能。 本研究提供了膽汁滯留的基因診斷，也對過去無法診斷的疾病做了分析，另外膽小管傳送蛋白的調控研究，有助於將來了解膽汁滯留機制及肝臟傷害；而細胞治療，雖然仍在實驗階段，但對遺傳性疾病及代謝性疾病來說，仍然是未來很有希望成功的治療方式值得繼續努力。
並列摘要
Introduction and Background Cholestasis, or impaired bile flow, is one of the most common and devastating manifestations of hereditary and acquired liver diseases. Cholestasis of infancy, manifesting as conjugated hyperbilirubinemia, accounts 50% of the work load in tertiary referral centers for pediatric hepatology and is a common disorder for children in Taiwan. Chinese infants have a higher incidence of biliary atresia than Japanese, Philippino, and white infants. Intrahepatic cholestasis is even more common than biliary atresia in Chinese infants. Patients usually manifest by conjugated hyperbilirubinemia within the first 2 months of age. Clinical manifestations include jaundice, light-yellowish or clay-color stool, small for age, poor feeding, and hepatosplenomegaly. Infantile cholestasis consisted of extrahepatic cholestasis and intrahepatic cholesatsis. Extrahepatic cholestasis includes biliary atresia, choledochal cyst, and inspissated bile syndrome. Intrahepatic cholestasis consists of a diversity of etiologies, termed as “neonatal hepatitis syndrome”. The etiologies includes congenital infections (TORCHES), chromosomal anomalies (Trisomy 18, 21, etc), endocrine disorders (hypopituitarism, hypothyroidism), structural anomalies (neonatal sclerosing cholangigis, Caroli disease, ductal paucity), genetic diseases (Alagille syndrome, progressive familial intrahepatic cholestasis), and metabolic diseases (tyrosinemia, glycogen storage disease, Niemann-Pick disease, lysosomal disorders, inborn errors of bile acid synthesis). Evaluation of the cholestatic infants remains a difficult task, owing to the diversity of cholestatic syndromes and to the obscure pathogenesis of many of these disorders. Prompt identification and diagnostic assessment of the infant with cholestasis is imperative to recognize disorders amenable either to institute specific medical therapy or effective nutritional and medical support to allow optimal growth and development. PFIC is a group of inherited, autosomal recessive disorders with infantile onset cholestasis. Patients progress to liver cirrhosis and hepatic failure in the early childhood. Three genes causing PFIC have been cloned. Defects in FIC1 (formal genetic name ATP8B1) that encodes a P-type ATPase, and BSEP (formal genetic name ABCB11) that encodes an ABC transporter, are found in patients with low serum GGT levels. Mutations in class III-multidrug-resistance P-glycoprotein (its formal genetic name ABCB4) are responsible for a distinct type of PFIC characterized by high serum GGT levels. Patients with phenotype of PFIC have been described in Japan but no reports about Chinese patients. There has been no report about children of PFIC in Taiwan. About half of our cases with neonatal hepatitis were attributed to CMV infection and the other half were idiopathic. Following the clinical course of patients with neonatal hepatitis revealed that most cases recovered within 6 months. However, some patients progressed to liver cirrhosis and died in the early childhood. Only a few of them had a family history of infantile cholestatic jaundice. It is not known whether hereditary cholestatic syndrome with the above mentioned genetic abnormalities existed in our patients. Comprehensive studies are impeded by the laborious genetic analysis, which is available in only a few research centers in the world. Our study started with identifying the patients with genetic disorders in Taiwan. These patients were previously classified as idiopathic. The diagnosis would directly improve the treatment and management to patients. Secondly, we determined the expression patterns of human fetal and postnatal stage, to correlate with physiological cholestasis in premature and normal infants. Thirdly we apply the primary hepatocyte culture systems to identify the effect of bile acids on the expression of canalicular transporters. The system may be applied to all forms of cholestatic diseases causing by insults such as sepsis and parenteral nutrition. Because current treatment of genetic cholestatic diseases has been unsatisfactory, most patients eventually would need liver transplantation. We have tried cell therapy in a mouse model of PFIC type 2, the sPgp knockout mice, for possible new therapeutic approaches to cholesatic liver diseases. Genetic Studies in patients with PFIC PFIC From 1980 to 1998, a total of 295 patients had been admitted to the Department of Pediatrics in National Taiwan University Hospital for infantile-onset intrahepatic cholestasis. Among them, 46 (15.6%) developed to chronic cholestasis. Among the 46 patients, Alagille syndrome was diagnosed in 18, inborn errors of bile acid synthesis in four, neonatal Dubin-Johnson disease in two, and tyrosinemia in one patient. No identifiable etiology was found in the remaining 21 patients, including 6 patients with high serum GGT levels (above 94U/L according to the normal range of infant in our hospital) and 15 with normal serum GGT levels. High GGT patients To investigate the etiology of the 6 patients with high GGT levels, either endoscopic retrograde cholangiography, magnetic retrograde cholangiography, or operative cholangiography was performed and revealed no structural anomalies of biliary trees. Needle biopsy of the liver was performed in all patients under informed consent from the parents. These patients were followed for 17 months to 16 years. The study has been approved by the institutional review board. Neonatal onset of cholestasis was found in five patients except case 5 who presented pruritus at one year of age. The GGT levels ranged from 1.2 x to 9.8 x of normal value. Liver histology showed significant widening and fibrosis of the portal areas, with progression to liver cirrhosis. In the remaining one patients, liver histology showed frequent spotty necrosis and occasional giant cell transformation. There were no significant changes in portal areas and bile ducts in this patient. Analyses for MDR3 gene. MDR3 gene was analyzed in all the 6 patients. Only one (case 5) had MDR3 mutation. A homozygous 719-bp deletion (nucleotide 287 to 1005 of coding sequence) of liver cDNA was found which encompassed exon 5 to 9 and led to a frameshift. We found that MDR3 mutations accounted for about 2% (1/47) of patients with infantile-onset chronic intrahepatic cholestasis in Taiwan. The only case with novel mutation in MDR3 manifested pruritus at 1 year instead of neonatal jaundice. This implies that MDR3 defect is not a common cause of high-GGT intrahepatic cholestasis in infancy, especially in neonatal cholestasis. Those patients presenting high GGT-PFIC with early onset cholestasis but without MDR3 mutation probably had inheritable disorders remaining to be clarified. Low GGT patients Thirteen patients whose serum GGT levels were normal or low (below 94 U/L according to the normal range of infant in our hospital were included. Phenotype characterization Two distinct histopathological phenotypes were identified. Group 1 included five patients (cases 1-5) and had liver histology characterized by bland cholestasis with minimal lobular disarray. Portal fibrosis was minimal to mild. Bile ducts were normal or slightly reduced in number. Group 2 included eight patients (case 6-13) and showed evident lobular disarray, marked giant cell transformation, or confluent cell necrosis accompanied by pericellular fibrosis. The grouping results from the two pathologists were identical. Group 1 patients had chronic or intermittent cholestasis and pruritus, with mild elevation of transaminase and bilirubin levels. The associated extrahepatic manifestations included diarrhea, rickets, and failure to thrive. Group 2 patients had marked elevation of transaminase and bilirubin levels. Peak serum bilirubin, aspartate aminotransferase, and alanine aminotransferase levels were higher than those in group 1 patients, p=0.02, p=0.001, and p=0.001, respectively. All the 8 patients in group 2 had high AFP levels (range, 2.4 to 16.2 SD score). Whereas all the 5 patients in group 1 had normal AFP levels as compared with age-specific normal range. (p<0.01). Pruritus and failure to thrive were frequent. Genotype characterization Analyses for FIC1 gene. Seven patients had frozen liver tissues available for analyses of FIC1 gene alterations. These included 4 patients in group 1 (bland cholestasis) and 3 patients in group 2 (giant cell transformation). All the 4 group 1 patients had mutations in FIC1 gene. Analyses for BSEP gene. cDNA sequencing was performed in 3 group 2 patients who had no mutations in FIC1 gene. A homozygous 850G→C(V284L), and a heterozygous 1-bp deletion at position 1145 was found in patient 7. The V284L mutation occurs in the intracellular loop between transmembrane spans 4 and 5. The 1145 deletion resulted in frameshift after codon 382 followed by 15 novel amino acids and protein truncation. A homozygous missense mutation 3137 G→A (G1004D) was found in case 8, which located in extracellular loop between transmembrane spans 11 and 12. In this study, genetic cholestasis disease was found in 86% (6/7) of the patients with infantile onset chronic intrahepatic cholestasis and low serum GGT levels. All the mutations were different from that reported from the Amish, European, or Arabic descents. It is a significant finding since the frequency of genetic diseases can vary greatly in people of different ethnic background. For example, alpha-1-antitrypsin deficiency, tyrosinemia, and cystic fibrosis were rarely seen in Asian children whereas Alagille syndrome is frequent in our children. 19 This finding prompt a specific direction of clinical awareness and research effort toward Asian patients of infantile cholestasis in the future. We found that the phenotypic diagnosis, based on clinical, serologic, and histologic features, correlated very well with genetic changes in chronic cholestatic patients with serum low-GGT levels. Two groups of patients were identified. Group 1 was characterized by normal AFP level and bland cholestasis in histology, and high frequency of FIC1 mutation (4 of 4 patients examined). In contrast, group 2 was characterized by high AFP level, giant cell transformation with lobular disarray, and high frequency of BSEP mutations (2 of 3 patients examined). We suggest that phenotypic characterization is a simple method to identify non-familial patients suspected to have FIC1 and BSEP mutations. Subsequent cDNA sequencing is useful to confirm the genetic diagnosis, and serve as a basis for genetic counseling or future gene therapy. We also have found that AFP may serve as an additional laboratory marker to differentiate patient with BSEP mutations from FIC1 mutation patients. Dubin-Johnson Syndrome (DJS) Four patients were included according to the clinical characteristics and hepatic pathology compatible with the diagnosis of Dubin-Johnson syndrome in the National Taiwan University Hospital. Two of the patients had an onset of cholestasis since the neonatal period, while the other two had a history of neonatal hyperbilirubinemia, which was documented since early adolescence. Mutational analysis (DJS) Genomic DNA analysis Biphasic pattern of jaundice attack was observed in one patient followed for 20 years. Mutations of ABCC2 gene, including 3 compound heterozygous, and 1 heterozygous mutations were found, including deletions (2748del136, 3615del229, and Del3399-3400) and missense mutations (L441M, E1352Q, and R393W). We found that all patients diagnosed as DJS before 10 years old have mutations involve one of the two ATP-binding cassettes (ABCs) of the MRP2 (multidrug-resistance protein 2) protein. To our knowledge, this is the first article reporting genetic analysis and long-term follow-up of neonatal DJS. DJS is rarely diagnosed in the neonatal period and only a few cases have been reported. The clinical characteristics and diagnostic investigations are similar with adolescent- or adult-onset DJS, except that more severe cholestasis and hepatomegaly may be found. All of the mutations found in our patients are novel. This is the first report of MRP2 mutations in patients of Chinese descent. The mutations are different from those reported from Japanese patients. Most reported mutations of the ABCC2 gene in DJS involve one or both ABCs. These ABCs are important for ATP binding and are highly conserved. Together with genotype-phenotype correlations in reported childhood cases of DJS, we suggested that disruptions of functionally important ABCs domains in the MRP2/ABCC2 protein maybe related to the earlier onset of the disease. Long-term follow-up of a neonatal-onset case is mandated because DJS may have a biphasic pattern of jaundice attack, with a second attack occurring after adolescence. Careful evaluation of all cases of neonatal jaundice to rule out DJS is suggested, especially in patients with conjugated hyperbilirubinemia with mild or minimally elevated transaminase levels. Developmental Expression of Canalicular transporters BSEP, MRP2, and MDR3 are major hepatic canalicular transporters mediating bile salts, conjugated bilirubin, and phospholipid secretion. Before birth, BSEP and MRP2 expressions have been reported to be either undetectable or at a much lower level than the adults in rats. This study aimed to understand the expression of canalicular transporters during human development, We investigated the expression of genes participating in bile formation and regulation. Human liver samples from fetus at gestational age 14 to 20 weeks, adult livers and liver samples of infants with biliary atresia patients were tested for mRNA expression of BSEP, MDR3, MRP2, NTCP, FIC1, and FXR genes by using real-time RT-PCR. Immunohistochemical staining of BSEP, MDR3, and MRP2 were performed on fetal and adult livers. Fetal expression levels of all the genes tested were lower than adult levels. The mean expression levels of BSEP, MRP2 and FIC1 were 30 to 50% of adult levels. The mean expression levels of FXR were 75% of adult levels, and the mean expression levels of MDR3 and NTCP were only 6.4% and 1.8 % of adult levels, respectively. The immunohistochemical staining of the fetal liver had distinct patterns from the adult liver. In the adult liver the characteristic canalicular staining of BSEP, MRP2, and MDR3 could be seen. In the fetal liver, different staining patterns in the three proteins were noted. MRP2 showed mainly canalicular pattern, slightly fuzzier than adult staining; BSEP showed partially intracellular and partially canalicular pattern; and MDR3 showed faint intracellular pattern, only occasionally canalicular stain could be seen. Our study presents that canalicular transporters are expressed early in the second trimester in human, which is different from the studies in rats showing that BSEP was not expressed until after birth. Many of the bile physiology studies have been performed in rodents. However, physiological differences between human and rodents should be considered. Our results showed that the time of BSEP expression correlates with the time when bile starts to flow to the intestine before 14 weeks of gestation. Although BSEP, MRP2, as well as MDR3 genes were readily expressed before birth, the localization pattern was yet distinct from mature hepatocytes. The reasons may be immature polarization of heaptocytes during development and inadequate targeting of specific proteins to the canalicular membrane. Regulation of canalicular transporters by bile acids in primary hepatocyte cultures In primary hepatocyte culture systems, we found that expression of canalicular transporters gradually decreased with time in primary hepatocyte cultures after 7~14 days. After adding bile satls into culture medium, expression of BSEP (bile salt export), FIC1, and NTCP was induced, implicating that these canalicular transporters may be regulated by intrahepatic bile acid levels. Cell therapy of sPgp knockout mice, a model of chronic cholestasis and progressive familial intrahepatic cholestasis type II Current therapy for PFIC includes palliative therapy: medical therapy (Urso), surgical therapy (partial biliary diversion); and curative therapy: liver transplantation. Major problems of liver transplantation are lack of donors and surgical risks. Cell therapy using hepatocytes or bone marrow cells in treating liver diseases had been successful in a few animal models including metabolic liver disease and partial hepatectomy model. There has been no research on cell therapy on chronic cholestasis animal model. We used green fluorescent protein (GFP) transgenic mice as donors for cell therapy. Isolated hepatocytes were infused into the spleen of sPgp knockout mice; in another study , and bone marrow cells were infused into tail vein in sPgp knockout mice with lethal dose irradiation (700~900 rad). We found that 4 to 8 weeks after hepatocyte transplantation, scattered GFP-positive cells were visible in the liver, especially peri-portal areas. sPgp immunofluorescent staining revealed no positive-stained cells. In mice with bone marrow transplantation, donor cells comprosed about 10 % of recipient liver DNA. Most cells were spindle shaped and were Kupffer cells. Some scattered hepatocytes could be identified. In sPgp knockout mice approximately 1 in 3520 hepatocytes were donor origin, but only 1 in 13863 hepatocytes were donor origin in wild type mice, indicating a higher rate of donor cell incorporation or cell fusion in sPgp knockout mice. The efficiency of bone marrow cells and hepatocytes repopulated in sPgp knockout mice liver was low, probably due to high endogenous regeneration activity. Further efforts in increasing donor cell repopulation and investigate the function restored by transplanted cell is mandatory.