台灣先天性巨大結腸症之基因研究

Abstract

先天性巨大結腸症是腸道的神經在發育時發生異常，導致腸道的末端沒有神經節的分布，進而形成遠端腸道功能性的阻塞。在臨床上的發生率約每五千個新生兒生會有一個人患有此疾病，且發生率在男性，亞洲人種最高。 腸道神經發育過程已知和多個基因相關，而任何發育過程的干擾都有可能形成先天性巨大結腸症。研究的初期，我們針對台灣的巨腸症患者做臨床個案的基因體研究。方法上，我們使用受試者的周邊血抽取DNA，並利用Captured based的次世代基因定序，選取31個報導過和巨腸症相關的基因做分析。在定序了41個家族，共45位患者後有幾個發現。第一，這樣的方法可以在87.5%的家族性患者以及46.67%的偶發性患者中找到可能致病的變異位點，遠高於過往只能在50%的家族性患者以及10~20%的偶發性患者中找到變異位點。第二，我們發現台灣地區的病患，雖然同樣以RET的變異為主，但SEMA3C可能和RET有著類似的重要性，這和國外報導過的有差別。第三，長型的患者中有20%的患者可以確認其致病變異，在未來或可經由人工生殖技術避免致病基因變異遺傳給下一代，降低巨腸症家族的復發率。 除了臨床個案的基因體研究之外，針對L1CAM基因，由於患者帶有的是會導致基因功能缺失，且新發生的突變，同時也符合L1症候群裡的水腦以及發展遲緩等症狀，而被進一步分析。我們使用CRISPR endonuclease的方式讓FVB小鼠帶有一個和我們在患者身上發現的同樣變異，L1CAM exon 18,g.2155 delG，使小鼠的L1cam基因產生一個frameshift的變異，同時也利用ssODN在變異點前方製造了一個限制酶的切點以加速後續分辨小鼠的基因型，並對此變異的致病性做進一步的探討。由於此基因位在X染色體上，故我們只針對L1camemfs1 hemizygous的公鼠做觀察並發現了兩件事。第一，在觀察了三代，共62隻的hemizygous公鼠後，並沒有任何一隻有巨腸症的表現，也沒有過去常提及的水腦，但有明確的公鼠不孕的現象。而公鼠的精液分析正常，體外授精也能順利產生可復育的胚胎，推測其不孕或許和其行為模式異常相關。第二，L1cam在小鼠一共有6個不同的transcripts，而利用不同primer做個別分析時發現，在L1camemfs1小鼠的腦部，所有的transcripts表現量都非常的低，然而在腸道中，短片段的transcripts表現量較wild type來的更高。在使用不同的L1cam抗體做免疫螢光染色時發現小鼠的腸道神經節的確有表現這些非全長的L1cam蛋白質。比較過去報導過L1CAM變異導致L1症候群伴隨巨腸症的患者中，會影響到這些短片段transcripts的變異佔了將近七成。這些短片段是否具有保護腸道神經節正常發育的功能值得進一步探討。 在小鼠之後，我們還有利用斑馬魚做基因的功能研究。斑馬魚有著繁殖數量多，魚體初期透明便於活體觀察等特性而被大量用於發育的相關的研究。我們在初期的基因體實驗中發現一對父女帶有NTRK1 exon7上的點突變，但NTRK1對巨腸症的影響並不明確。為此，我們引進了一隻經由ENU突變導致其ntrk1 exon7帶有終止碼的斑馬魚品系，sa14955，並利用其和神經節帶有GFP螢光的Tg(phox2b:EGFP)w37 的基因轉殖魚配種後得到可以觀察ntrk1缺失後神經發育表現的子代。經過初步的觀察發現在ntrk1缺失的斑馬魚的後半腸道，若有的確有明顯神經結數量減少的現象。未來將增加觀察的斑馬魚數量，以及利用攝影產生Spatiotemporal maps來確認減少的神經節是否有造成蠕動的異常。 在未來，除了繼續斑馬魚的實驗之外，針對易用目前定序方式仍找不到變異位點的長型患者，可進一步利用父母以及患童的Trios全外顯子定序方式試著找到未知的可能致病的基因變異。

Hirschsprung''s disease (HSCR) is a congenital disorder with the absence of myenteric and submucosal ganglion cells within the distal gut. Due to multigenic inheritance and interactions, we employed next‐generation sequencing (NGS) to investigate the genetic backgrounds of Hirschsprung’s disease in Taiwan. We started with a clinical genetic study in patients with HSCR. Genomic DNA extracted from the peripheral blood of HSCR patients was subjected to capture‐based NGS, based on a 31‐gene panel. This study enrolled 45 HSCR patients, including 37 (82.22%) sporadic cases and 8 (17.78%) familial patients in 4 different families. Among these patients, we found possible harmful variants 87.5% of familial patients and 46.67% in sporadic patients. This is much higher than what was reported previously that only 50% of the familial cases and 10~20% of the sporadic cases can found possible harmful variants. We also noticed that SMEA3C may play an important role in Taiwan’s patients which is different from other studies. Lastly, in our study, 20% of our long-segment HSCR patients could find a pathogenic variant. This might be used to prevent the transmission of variants down to their offspring, especially when in vitro fertilization is needed for infertility. From our clinical genetic study, we found an L1CAM de novo frameshift mutation in a female with mild hydrocephalus and skip-type HSCR. A nearly identical L1cam variant was introduced into FVB/NJ mice via the CRISPR-EZ method. A silent mutation was created via ssODN to gain an artificial Ncol restriction enzyme site for easier genotyping. Six L1cam protein-coding alternative transcripts were quantitatively measured. Immunofluorescence staining with polyclonal and monoclonal L1cam antibodies was used to characterize L1cam isoform proteins in enteric ganglia. Fifteen mice, seven males and eight females, generated via CRISPR-EZ, were confirmed to carry the L1cam frameshift variant, resulting in a premature stop codon. There was no prominent hydrocephalus nor HSCR-like presentation in these mice, but male infertility was noticed after observation for three generations in a total of 62 hemizygous male mice. Full-length L1cam transcripts were detected at a very low level in the intestinal tissues and almost none in the brains of these mice. Alternative shorter transcripts encoding the extracellular domains were over-expressed in the intestine of L1cam knockdown mice. Immunofluorescence confirmed no full-length L1cam protein in enteric ganglia. These shorter L1cam isoform proteins might play a role in protecting L1cam knock-down mice from HSCR. In addition to the mice study, we also include zebrafish as our animal model. The advantages of using zebrafish include that hundreds of embryos can be produced in a single clutch and the optical clarity of the developing embryo, which allows live imaging at the organism level. From our clinical genetic study, we found a family with a single nucleotide variant in the exon 7 of NTRK1. We imported sa14955, which was generated via ENU and carried a stop codon at ntrk1 exon 7. By crossing sa14955 with Tg(phox2b:EGFP)w37, we could observe the distribution of ganglion cells in nrtk1 knock-down or knock-out embryos as the neurons carry GFP. The initial results showed that significantly decreasing neurons were noted in the distal intestine of these zebrafish when compared to their wild-type counterparts. We will carry on with spatiotemporal maps to see if the motility of the intestine has also been affected. In the future, a trio whole exome sequence can be applied to the long-segment HSCR patients whose variant can not be identified by the current method.