- 學位論文
以斑馬魚為模式探討nkx2.7和rtn4在人類致病的分子機制
Using Zebrafish as a Model to Study the Molecular Mechanism of nkx2.7 and rtn4 in Human Diseases
摘要
Part 1 因早期心臟發育的錯誤或異常而形成缺陷的心臟被定義為人類先天性心臟缺陷Human congenital heart defects (CHDs)。在早期胚胎發育時，許多心臟專一性的轉錄因子在心臟發育扮演了重要的角色。相較於已有許多報導的突變Nkx2.5,其同源基因（paralogue），Nkx2.6的突變則較少被報導。根據家族遺傳分析，我們在臨床上CHDs的病人中，找到兩個新穎的突變型Nkx2.6，分別為Nkx2.6541mut and Nkx2.6895mut.然而，這兩個突變型Nkx2.6是否會影響心臟發育而造成心臟缺陷還不清楚。因為斑馬魚的在心血管發育研究的優勢，包括了在早期發育不完全仰賴心血管系統提供氧氣的特性，我們利用斑馬魚去研究這個議題。為了分析Nkx2.6541mut and Nkx2.6895mut的特性，我們構築了斑馬魚同源基因（orthologue）的突變型nkx2.7。實驗結果顯示，利用顯微注射nkx2.7WTmRNAs過量表現的胚胎會造成心臟looping的缺失，但是在過量表現Nkx2.7541mut and Nkx2.7895mutmRNAs的胚胎，造成心臟looping的缺失的發生率和控制組沒有顯著差異，顯示突變型nkx2.7在功能上可能有所缺失。在rescue實驗當中，注射wobble nkx2.7WT mRNAs的胚胎能夠部分地rescue由nkx2.7-morpholino (MO) 造成心臟looping的缺失。但是，wobble nkx2.7541mut mRNAs rescue的效果則較低而 nkx2.7895mutmRNAs幾乎沒有rescue的效果。同時，我們也分析兩個突變型nkx2.7在同時缺乏nkx2.7和nkx2.5情況下對於心臟looping缺失rescue的能力。過量表現nkx2.7WTmRNAs能夠部分地rescue由nkx2.7-MO和nkx2.5-MO 造成心臟looping的缺失,但是，兩個突變型nkx2.7並沒有rescue的能力。這些結果，強烈地顯示在對於調控心臟looping的功能上兩個突變型nkx2.7有所缺失。接下來，我們想要了解nkx2.7對於下游基因的啟動的能力。In vivo luciferase分析顯示，過量表現nkx2.7WT mRNAs的胚胎能夠開啟斑馬魚cmlc2啟動子的活性，而過量表現兩個變型nkx2.7斑馬魚胚胎中cmlc2啟動子的活性與控制組沒有顯著差異。而利用WISH （Whole mount in situ hybridization）更進一步確定過量表現nkx2.7WT會異位性開啟斑馬魚內生性cmlc2的轉錄活性，但是過量表現nkx2.7541mut or nkx2.7895mut，則有較少比例的胚胎開啟cmlc2的轉錄活性。這些結果顯示，在開啟下游基因的能力上，兩個突變型nkx2.7都有缺失。為了進一步模擬人類CHD病人的情形，我們利用TALEN，建立了帶有nkx2.7895mut突變的斑馬魚。同時，我們也利用TALEN結合同源互換（homologous recombination）建立了由內生性啟動子驅動的nkx2.7895mut-P2A-GFP. 這些帶有nkx2.7895mut突變的斑馬魚可以幫助我們進一步了解在成人中CHD病人裡心臟細部的分子機制。綜合以上所述，我們發現 （1）兩個Nkx2.6/nkx2.7的突變位點位在有功能的domain之中；（2）過量表現兩型突變型Nkx2.6/nkx2.7並不能造成心臟looping的缺失; (3) 兩型突變型Nkx2.6/nkx2.7啟動斑馬魚心臟專一性cmlc2啟動子的能力較弱；和（4）建立帶有nkx2.7895mut突變的轉殖基因斑馬魚。 Part 2 miR-206是肌肉專一表現的微型核糖核酸（microRNA， miRNA）, 透過調節其下游基因影響肌肉與鄰近組織的發育。先前實驗室利用labeled microRNA pull-down assay 在16 hours post-fertilization (hpf) 的斑馬魚胚胎發現nogo/rtn4為miR-206直接調控的基因。在哺乳類中，非神經細胞的nogo/rtn4, 主要是透過其Nogo-66 domain與神經細胞上的receptors（包含NgR和PirB）的結合，誘發neurite崩解的效果，進而抑制神經的再生。相反地，也有斑馬魚的Nogo-66會促使神經neurite生長的報告。顯然，斑馬魚rtn4 isoforms是否會擁有不同的功能還不清楚；各rtn4 isoform和miR-206在肌肉細胞中對於運動神經生長的影響也尚未釐清。所以，我們首先在肌肉中過量表現不同全長的rtn4 isoforms，以研究他們對運動神經元發育的影響。結果顯示（1）rtn4m和rtn4n兩種isoforms與已知的斑馬魚Nogo-66 domain的功能相同；但是過量表現rtn4l卻會抑制運動神經細胞neurite的生長；和（2）抑制內生性miR-206所呈現運動神經細胞neurite生長受到抑制的phenotype與在肌肉中過量表現rtn4l的phenotype一樣。我們進一步以連續性剔除（serial deletion）與氨基酸突變分析，我們證實rtn4l會抑制neurite outgrowth的關鍵在第104個氨基酸。 另一方面，在哺乳類當中，我們想要了解肌肉中Nogo三型isoforms中的Nogo-A是透過何種機制來達到抑制運動神經的效果。雖然Nogo-A 基因剔除的小鼠會擁有較好的神經再生能力，但是NgR基因之剔除小鼠或是PirB基因剔除之小鼠同時阻擾NgR訊息傳遞路徑並沒有相對改善其神經再生能力。所以，我們合理地假設肌肉中的Nogo-A可能不僅會透過典型Nogo-66/NgR或是Nogo-66/PirB的訊息傳遞路徑，還有可能影響分泌的myokine來抑制神經的neurite的生長。為了證實這個可能性，我們用Sol8肌肉細胞，過量表現Nogo-A，然後將所分泌在細胞外的conditioned media （CM）與運動神經細胞NSC-34共同培養。我們發現CM存在會抑制neurite的生長。進一步偵測神經崩解分子（磷酸化的cofilin）的表現量也發現它會上升，表示神經neurite生長被抑制。接著分析浸泡過量表現Nogo-A 肌肉細胞之CM的運動神經細胞，發現其磷酸化AKT和磷酸化EGFR並沒有顯著變化。這些結果都和已知Nogo-66/NgR路徑造成抑制神經neurite生長時呈現磷酸化AKT表現量降低和磷酸化EGFR表現量上升的現象並不一致。因此，我們認為當肌肉細胞過量表現Nogo-A時，細胞外的myokine可能改變，而影響運動神經細胞neurite 生長，這也表示存在著另一種與Nogo-66/NgR不同的調控路徑。為了要找到由Nogo-A影響分泌到細胞外的分子，我們將會利用二維電泳去分析Nogo-A過量表現肌肉細胞與控制組分泌myokine的差異。這些結果，有助於了解人類脊髓側索硬化症（amyotrophic lateral sclerosis, ALS）這種神經退化性疾病的致病機轉，我們希望為ALS提供另一個可能治療的新對策。
並列摘要
Part 1 Human congenital heart defects (CHDs) is defined as cardiac malformation resulting from error cardiac development. During early embryogenesis, many heart-specific transcription factors play important role in cardiogenesis. Compared to Nkx2.5, which mutations had been enormously reported, mutation of Nkx2.6, a paralogue of Nkx2.5, is less known. We have identified two novel Nkx2.6 mutants, designed as Nkx2.6541mut and Nkx2.6895mut, from CHDs patients based on genetic study. However, it is unclear the effect of these mutants on cardiogenesis. Since zebrafish offers several advantages including less need for cardiovascular system to provide oxygen at embryo stage, we used zebrafish to study this issue. Compared to control group, embryos injected nkx2.7WT exhibited heart looping defect. However, the occurrence rates of heart looping defect were not different between control group and both mutant-injected group. Rescue experiment demonstrated that injection of wobble nkx2.7WT mRNAs could partially rescue the heart looping defect induced by nkx2.7-MO. However, the injection of either wobble nkx2.7541mut mRNAs or wobble nkx2.7895mut mRNAs did not rescue the heart looping defect. We further examined the rescue capabilities of mutated nkx2.7 in the absence of both nkx2.7 and nkx2.5. Results showed that overexpression of wobble nkx2.7WT mRNAs could partially rescue the heart looping defect in both nkx2.7 and nkx2.5 morphants. However, injection of either wobble nkx2.7541mut mRNA or wobble nkx2.7895mut mRNA did not rescue the heart looping defect. Interestingly, in vivo luciferase activity demonstrated that the zebrafish cmlc2 promoter activity of nkx2.7WT mRNAs-injected embryos was greater than that of control embryos, while the cmlc2 promoter activities were not different among nkx2.7WT control group and two nkx2.7mut-injected groups. WISH demonstrated that ectopic expression of heart-specific cmlc2 was observed in the nkx2.7WT–overexpressed embryos. However, embryos injected with either nkx2.7541mut or nkx2.7895mut displayed less ectopic expression of cmlc2. Moreover, the G0 of zebrafish mutant carrying nkx2.7895mut was generated by TALEN, and its F1 offspring have been identified. Also, zebrafish mutant carrying nkx2.7895mut fused with GFP reporter and driven by endogenous nkx2.7 promoter was established by TALEN-mediated homologous recombination. This mutant line with GFP reporter should help us to understand in more detail the molecular mechanism involved in CHD patients during adulthood. We concluded that (1) Two mutation sites are located at Nkx2.6/nkx2.7 functional domain; (2) Mutated Nkx2.6/nkx2.7 exhibits loss of function of heart looping in vivo; (3) Two mutated Nkx2.6/nkx2.7 impair trans-activation ability of heart-specific cmlc2 promoter; and (4) Transgenic line of zebrafish which harbors mutated nkx2.7, nkx2.7895mut , has been generated. Part 2 Muscle-specific microRNA (miRNA), miR-206, plays important role in muscle and its adjacent tissue development by regulating other gene expression at the post-transcriptional level. Previously, we identified a target of miR-206, which is nogo/rtn4, by labeled microRNA pull-down assay we developed from the 16 hours post-fertilization (hpf) zebrafish embryos. In mammals, non-neuronal nogo/rtn4 induces neuron collapse through direct binding of its Nogo-66 domain to neuronal receptor including Nogo-receptor (NgR) and PirB, thereby prevent neuron regeneration. However, there is paper reported that zebrafish Nogo-66 domain promotes rather than inhibits neurite outgrowth. Apparently, it is unknown whether zebrafish rtn4 isoforms have different functions. Additionally, the effects of different rtn4 isoforms and miR-206 within muscle on motoneuron haven’t been clarified yet. Therefore, we first overexpress different rtn4 isoforms within muscle in order to study the effect of them on motoneuron. Our results showed that (1) the function of rtn4m and rtn4n is same with the known function of zebrafish Nogo-66 domain, however, overexpression of rtn4l within muscle inhibits neurite outgrowth; (2) inhibition of endogenous miR-206 results in neurite outgrowth inhibition, resembling rtn4l overexpression in muscle. After subsequent serial deletion and mutagenesis analysis, we proved that the important amino acid residue that exerts inhibitory effect reside at the 104th position of rtn4l protein. Apart from this, we want to clarify that how one of the nogo/rtn4 isoforms, Nogo-A within muscle inhibit neurite outgrowth in mammals. Though Nogo-A knockout mice exhibit enhanced neuron regeneration ability, NgR knockout mice or PirB knockout mice with blocking NgR pathway do not exhibit comparable neuron regeneration ability. Therefore, we reasonably hypothesized that the muscle-specific Nogo-A would not only induce a canonical Nogo-66/NgR or Nogo-66/PirB signaling pathway, but also might influence the secreted myokine, to inhibit neurite outgrowth. To verify this possibility, we overexpress Nogo-A, in Sol8 myoblasts, and collect the conditioned media (CM) to coculture with motoneuron cell line NSC-34. We found that the neurite outgrowth was inhibited. Western blot showed that collapsing marker (phosphorylated cofilin) of the motoneuron cell line after treated with CM derived from myoblast expressing Nogo-A is up-regulated, indicating the neurite outgrowth is inhibited. Subsequent analysis showed that the phosphorylated AKT and phosphorylated EGFR of the motoneuron cell line after treated with CM derived from myoblast expressing Nogo-A remained unchanged. These evidences are inconsistent with the canonical Nogo-66/NgR pathway, which inhibits AKT phosphorylation and induces EGFR phosphorylation. Collectively, we postulate that Nogo-A overexpression in muscle might result in change in secreted myokine, subsequently influence neurite outgrowth of neuron, indicating a presence of non-canonical Nogo-66/NgR pathway. In order to identify the secreted factor influenced by Nogo-A overexpression in myoblasts, we will utilize two-dimensional gel electrophoresis to analyze the difference between myokine secreted from myoblasts expressing Nogo-A and control. These results would shed a light on the pathological mechanism of amyotrophic lateral screlosis (ALS), a progressive neurodegenerative disease and we anticipate providing another possible theraputic alternative to ALS.