使用人類類寡樹突膠細胞模型探討與鋰鹽治療反應相關之遺傳位點功能

Abstract

引言： 躁鬱症是一種常見的精神疾患，其病人會經歷反覆發作的躁期與鬱期。近年來，躁鬱症的病理學研究發現患者大腦白質中寡突膠細胞顯著少於一般人。此外，研究也發現和正常人相比之下，躁鬱症病人大腦皮質中寡突膠細胞相關的基因表現是低於健康組的。在1949年後鋰鹽就是拿來緩解躁期的第一線藥物。雖然目前鋰鹽的治療機轉仍未非常確定，目前有些研究認為鋰鹽可以透過影響Wnt/β catenin路徑來強化寡突膠細胞生成髓鞘的功能使神經訊號傳導更加穩固。然而，儘管鋰鹽被廣泛的使用，並非所有的人都對它有良好的治療反應，而這背後的機轉目前仍不明朗。 因此，如果能找到一些生物指標來預測病人對鋰鹽的反應對臨床上是非常有幫助的。在2016年，一篇全基因組關聯分析研究發現GFRA2基因中的一個位點rs7833426在東亞群體中和正向治療反應有相關。而GFRA2是一個由寡突膠細胞分泌的生長因子(GDNF)配體家族的受體。在這篇研究中，我們透過研究寡突膠細胞相關基因期能找出其他有潛力預測治療反應的基因並使用人工誘導的寡突膠細胞來探討這些基因的功能。最後，我們將在細胞模型中加入鋰鹽，調查是否可以改變因該基因所導致的細胞表現型。 方法： 我們收集了101位台灣躁鬱症患者並執行了全基因組關聯分析並用其結果進行了gene-based分析。由於基因表現在不同組織會不同，我們將分析的結果與寡突膠細胞相關的基因列表結合。在找出最顯著的7個基因後我們在化學分子誘導的寡突膠細胞模型上進行了即時聚合酶連鎖反應以精準定出候選基因。 本論文中我們使用了兩個細胞模型。第一種是化學分子誘導的寡突膠細胞模型。我們將人類纖維母細胞培養在加入4種小分子藥物的培養液中，以使其在4天後轉換成類突膠細胞。第二種為將神經母細胞誘導成寡突膠母細胞。我們使用誘導分化的培養液讓不朽化神經母細胞在40天內分化成為寡突膠母細胞。 我們透過在細胞模型上感染慢病毒來執行RNA干擾，使我們的候選基因表現下降。我們觀測感染後的細胞型態，並使用即時聚合酶連鎖反應以及免疫螢光染色，來測量與寡突膠細胞相關的基因表現。最後，我們會加入氯化鋰來觀察上述寡突膠細胞基因表現是否受到影響。 結果： 全基因組關聯分析結果顯示，只有19個SNP達到顯著水準。在gene-based中，與寡突膠細胞有關的最重要的基因是DOCK2，SSPN，PVALB， RFFL， RHOC， INPP5D和GNG11，p值的範圍為0.00016至0.0027。由於GFRA2（32倍）和INPP5D（43倍）的高表現，我們選擇了這兩個基因作為候選基因。 抑制候選基因表現不會改變化學分子誘導的寡突膠細胞的形態，也不會改變O4和MBP的表現。另一方面，在神經母細胞誘導的寡突膠母細胞中，兩個候選基因的表現都會使O4的表現顯著上升，儘管Olig2的表現保持不變。最後，將10mM LiCl添加至經慢病毒感染後的細胞之培養液後，O4和Olig2的表現沒有改變。 結論： 在這篇研究中，我們使用寡突膠細胞模型來研究鋰治療反應相關基因的功能。我們發現，抑制GFRA2和INPP5D的表現不會影響化學誘導的寡突膠細胞的重編程過程，但會影響神經母細胞誘導之寡突膠母細胞相關基因的表現。這可能表明這兩個基因不影響細胞重編程，而是負責寡突膠細胞的分化。鋰鹽不會改變候選基因遭抑制之細胞中的指標基因表現。因此，需要做進一步的實驗來揭示為何鋰鹽不會影響寡突膠母細胞分化。

Introduction Bipolar disorder (BD) is characterized by episodes of mania and depression that often repeat periodically. Previous studies found that reduction in the number of cortical oligodendrocytes and myelin gene expression deficits were observed in BD. Since1949, lithium has been considered the first line drug for bipolar disorder. It is suggested that lithium could potentially improve the myelination of the oligodendrocyte by affecting Wnt/β catenin pathway. The efficacy of lithium treatment varies between patients. The mechanism behind this phenomenon is still unknown. Hence, it is important to find out some biomarkers to predict the patients’ response to lithium. In 2016, Hou., et al identified in a genome-wide association study that a SNP (rs7833426) within GFRA2, receptor of oligodendrocyte-secreting neuronal growth factor, is associated with positive treatment response from Asian cohorts. In this study, I aim to identify more genes associated with treatment response and measure their functional readout on induced human oligodendrocyte lineage cells. Materials and Methods I performed GWAS and a subsequent gene-based test from data set of 101 BD patients from Taiwan, from which the results were merged with a list of genes associated with oligodendrocytes. Next, I pinpointed candidate genes through quantitative real time polymerase chain reaction (qRT-PCR) on our cell model. Two cell models were applied. In the first model, chemically induced oligodendrocytes, I converted fibroblasts into oligodendrocytes with four drugs. In the second model, neural stem cell derived oligodendrocyte progenitor cell (OPC), I derived immortalized human neural stem cells into oligodendrocytes in 40 days. Candidate genes were knockdown via infection of lentivirus encoding short hairpin RNA. I assayed the result of the knockdown by observing the morphological change and measuring the expression of oligodendrocyte related markers by immunofluorescence staining and qRT-PCR. Lastly, lithium chloride was added to the knockdown cells to check if the expression of those markers was altered. Result Genome-wide association analysis results in single marker levels showed that only 19 SNPs reach the suggestive significant level. In the gene-based testing, the most significant genes and are relevant to the oligodendrocyte cells are DOCK2, SSPN, PVALB RFFL RHOC INPP5D, and GNG11 with the p-values ranged from 0.00016 to 0.0027. Because of the high expression of GFRA2 (32 folds) and INPP5D (43 folds), I picked these two genes as our candidates. Knockdown of the candidate genes did not change the reprogramming procedure and morphology of the chemically induced oligodendrocytes, neither the expression of the O4 and MBP. On the other hand, expression of O4 was significantly upregulated after the knockdown GFRA2 and INPP5D in neural progenitor derived OPCs, despite the expression of the Olig2 remained unchanged. 10mM LiCl was added to the knockdown cell line. The expression levels of the O4 and Olig2 did not altered. Conclusion In this study, I used oligodendrocyte cell models to investigate the functionality of the lithium responsiveness related genes. I found that knockdown of the associated genes, GFRA2 and INPP5D, did not affect the reprogramming procedure of the chemically induced oligodendrocytes. Of note, the expression level of O4 was upregulated in neural stem cell derived OPCs after the knockdown of GFRA2 and INPP5D. This may indicate that these two genes may be in charge of the differentiation of the oligodendrocyte lineage cells. Lithium did not alter the expression of the markers in knockdown cells. Thus, further experiments are needed to unveil the mechanism of why lithium did not affect the differentiation of OPCs.