摘要:關鍵字:葛瑞夫茲氏病,自體免疫疾病,全基因體規模關連研究,人類白血球抗原,疾病危險性預估葛瑞夫茲氏病是全世界造成甲狀腺機能亢進以及甲狀腺眼病變的最重要原因。它是一個影響數個器官的自體免疫疾病,而在臨床以及基礎研究上都佔有重要地位。本病的病因尚未清楚,一般相信是多種原因共同造成,而遺傳因素扮演重要角色。 為了找出葛瑞夫茲氏病的病因,過去十年來已經做了非常多聯鎖分析,其中有些是針對特定區域,有些則是全基因體掃瞄。然而,成果非常有限。此外,也進行了無數的關連研究,在所有的報告結果中,HLA,CTLA4 以及 PTPN22 等算是比較可信的。到目前為止,全世界只有一個全基因體規模關連研究,但該研究的檢力只算中等(檢體數:900病人,1,468對照組),而對全基因體的涵蓋性也不足(只做了大約 15,000 個在基因上頭的 SNP)。 整體說來,葛瑞夫茲氏病的遺傳研究領域有太多可以改進的地方。首先,檢體數量普遍不足。其次,少有以家族檢體來做的研究。第三,大部分的研究是利用白人的檢體做的,目前已有很明確的證據顯示說本病的致病因子可能會有種族差異。第四,到目前為止的唯一全基因體規模關連研究並不夠好。第五:套數變異對葛瑞夫茲氏病的影響從未被檢驗過。第六,亞洲人在典型 HLA 位點上的致病對偶基因型尚未被發現。第七,在找尋其他易罹病基因時,未能一併考慮 HLA 的影響。第八,從未建立利用基因資訊來預測疾病危險性的模型。 本團隊在過去十年來在葛瑞夫茲氏病的遺傳研究上已建立了穩固的基礎,簡單闡述如下:首先,我們目前的葛瑞夫茲氏病檢體數量(1000個病人,1000個對照組以及807個家族檢體)是全球最大的。第二,我們先前已成功地利用家族檢體發現了一個先前從沒有人報告過的危險 SNP。第三,我們是全球唯一利用連鎖分析而能明顯偵測到 HLA 重要性的團隊。第四,我們目前已經就典型的所有六個 HLA 位點進行了基因定型,也發現了在我們族群內特有的致病對偶基因型。第五,我們發展了一個可以推估套數變異的軟體,也將此軟體運用在我們的檢體。第六,我們已經完成了一個先期全基因體規模關連研究(246病人,468對照組),根據目前的分析結果,本研究的品質相當可靠,也的確如預期地發現了在 HLA 的信號,除此之外,還在另外幾個區域發現信號。第七,我們有很好的基礎在找尋其他易罹病基因的過程中,將 HLA 的資料一併納入考慮。第八,我們將利用基因資訊建立葛瑞夫茲氏病疾病危險性預估模型。 在未來的四年計畫期間,我們有信心要完成全世界葛瑞夫茲氏病研究領域內的第一個具有高檢力(2,000病人,2,000對照組)又有完整涵蓋性的全基因體規模關連研究。兩階段性全基因體規模關連研究已被明顯證明可以節省大量支出卻仍能保有幾乎所有的檢力,而因為我們已經有了246病人以及1000對照組的所有基因型,我們可以在第一階段全基因體規模關連分析再額外節省超過百分之六十以上的支出。我們也將整合HLA以及套數變異的資料進入所有的分析,而且將測試 HLA 和其他所有基因的可能交互作用。一旦發現了可能的致病基因以及致病對偶基因型,我們將進行最合適的實驗來測試它們可能造成的功能影響,進而找出真正的致病基因。而最直接在臨床上馬上可以獲益的研究成果則會是我們建立的疾病危險性預估模型。 發現致病基因將能夠大大幫助科學研究。更進一步來說,能發現我們族群本身特有的致病基因或是致病對偶基因型的話那更是非常重要,因為其他族群的研究永遠沒有辦法提供給我們這樣的寶貴資料。而一個好的疾病危險性預估模型將有助於個人化醫療。最後,發現葛瑞夫茲氏病的致病基因不僅對本疾病的研究與治療有幫助,甚至可能一併嘉惠其他自體免疫性疾病的研究。
Abstract: Keywords: Graves’ disease (GD), autoimmune disease, genome-wide association study (GWAS), Human Leukocyte Antigen (HLA), risk prediction modelGraves’ disease [GD (MIM 27500)], the worldwide leading cause of hyperthyroidism and thyroid eye disease, is a common organ-specific autoimmune disorder with both clinical and scientific importance. The etiology of GD is generally accepted to be multifactorial, with several lines of evidence showing significant genetic influence. Many linkage analyses for GD were performed in the past 10 years, with unsatisfactory results. Association studies have been conducted on numerous candidate genes/regions. Among the regions with positive results, only the HLA region, CTLA4 and PTPN22 seem to be promising. The only genome-scale association study for GD had modest power (900 cases and 1,468 controls) but inadequate genome coverage (~15,000 nonsynonymous SNP). There is plenty of room for GD genetic researches to improve. First, the sample sizes have been unsatisfactory. Second, there were very few family-based association studies. Third, the majority of published results were from Caucasians, and there is strong evidence in several examples that the susceptibility alleles for GD are different across populations. Fourth, the only genome-wide association study (GWAS) so far is not good enough. Fifth, copy number variations (CNVs) have not been tested. Sixth, the classical HLA susceptibility loci/alleles in Asians have not been identified. Seventh, the effect of HLA allele(s) has not been incorporated while testing other susceptibility genes. Eighth, there is no model that can predict GD disease risk based on genetic information.Our group has built a solid foundation on genetic study of GD during the past 10 years. First, we have set up the largest GD collection in the world (1,000 GD cases, 1,000 controls and 807 family samples). Second, we conducted a family-based association study on CTLA4 and identified a novel susceptibility SNP (rs733618). Third, we are the only group worldwide to demonstrate significant linkage signal for GD at the HLA region. Fourth, we have finished a large-scale high-quality HLA genotyping (4-digit resolution for all 6 classical HLA loci) and successfully identify susceptibility/protective alleles of GD in our own population. Fifth, we developed a novel method for CNV inference, and can apply it to our upcoming GWAS. Sixth, we conducted a pilot GWAS (246 GD cases/ 468 controls) which demonstrated reliable quality and positive association signals at HLA and a couple of other regions. Seventh, we are in a good position to incorporate HLA genotypes to our GWAS project. Eighth, we can build a GD risk prediction model based on our population-specific genetic information.In the upcoming new 4-year project period, we have the confidence to perform the first statistically well-powered two-stage GWAS (2,000 GD cases and 2,000 controls) for GD, with unbiased genome coverage. A two-stage GWAS will substantially save the cost while maintaining the power. Because of the availability of genome-wide SNP genotypes for 246 GD cases and 1,000 controls, we can further save greater than 60% cost for stage-one GWAS. HLA genotypes and CNVs will also be incorporate in GWAS, and interaction of HLA with other SNPs will be tested. We will then perform functional assays to identify and verify the genuine causative variants. The immediate clinically relevant aim is to build a GD risk prediction model based on genetic information.Identification of susceptibility genes will pave the road for better scientific research. Discovery of population-specific risk alleles/genes will be invaluable because this kind of knowledge can never be generated from other populations. A good disease risk prediction model has the potential for personalized medicine. Furthermore, breakthrough in GD study might even shed light on research of other autoimmune diseases.