選擇性血清素再攝取抑制劑對於不同憂鬱症狀分群的治療反應之藥物遺傳學研究

Abstract

引言: 重度憂鬱症是一個複雜且富含多因素的高異質性之精神疾病。病人對於最常被當作處方藥的選擇性血清素再攝取抑制劑 (SSRIs) 的治療效果根據個體或其症狀的不同而有所不同。大多數的重度憂鬱症患者在服用SSRIs治療後會產生令他們感到麻煩、厭煩的副作用。性功能障礙亦屬於常見的副作用之一且其盛行率在服用SSRIs的病人中高達30%。而另一個嚴重的副作用為藥物治療後出現自殺意念的情況。因此，對於了解治療反應和副作用的相關預測因子是一個十分重要的議題。 而遺傳因素被認為是影響藥物治療反應的重要原因之一。藥物遺傳學研究為探討遺傳變異與重度憂鬱症患者服用抗憂鬱症藥物後的治療效果之間的關係。 方法: 本研究在台灣北部招募了455位重度憂鬱症門診病患。憂鬱的嚴重程度依據21題的漢氏憂鬱量表 (HRSD) 來衡量，而受試者的問卷總分須至少14分才可納入研究，亦在第二、四、八週對受試者再進行一次問卷訪談。此外，所有受試者的基因型藉由Illumina HumanOmniExpressExome BeadChips定序。對單核苷酸多態性 (SNPs) 與個體執行一系列的篩選後，保留了421位病患與4,241,701個SNPs，其平均年齡為43.65歲，71%為女性，服用escitalopram和paroxetine皆為38.48%，18.29%服用fluoxetine，4.75%服用citalopram。 藉由探索性因素分析，並根據病患在研究開始時的HRSD分數分成數個憂鬱症狀因素，並藉由這些症狀因素進行後續的全基因組關聯分析。治療反應定義為兩個變項，分別為「%change」定義為研究初始到每個後續的收案時間點之間的改善幅度，以及「response」定義為其%change的改變量是否有超過50%。另外，我們亦會探討兩個副作用，分別為藥物治療後出現自殺意念與性副作用，皆定義為在所有的訪談時間點間HRSD的分數是否有增加。無論是線性迴歸、邏輯斯迴歸或是重複測量的關聯性分析皆會校正年齡和性別。 結果: 藉由探索性因素分析發現5個症狀因素，分別為核心、失眠、身體焦慮、心理運動的病識感與厭食症狀。症狀的改善幅度在第四週時，最低的為身體焦慮(29.75%)，最高的為失眠(43.47%)。雖然沒有一個位點達到全基因組的顯著水準，但仍有數個位點達到建議的顯著水準(P<5×10-6)，「%change」與「response」分別發現184個與33個SNPs。其中最顯著的位點位於%change為與失眠相關的rs146653208 (P=1.12×10-7)，在response為與失眠相關的rs11759993 (P=1.85×10-6)。 與不同症狀因素相關的基因分別為在%change的LOC101927653 (核心)，COL22A1、ENSG00000227101、ENSG00000236990、ENSG00000258214、ENSG00000212599、ENSG00000251931和CGNL1 (失眠)，IQCJ-SCHIP1、SCHIP1、NRXN3、SNRPN和TIAM1 (身體焦慮)，CCKAR、MGAM、TAS2R38、SLC5A8和UTP20 (心理運動的病識感)，ENSG00000232197、ENSG00000273100、ENSG00000222974、CYTH1、USP36、ENSG00000252818、GNAZ和RTDR1 (厭食)，以及在response的CSE1L和STAU1 (核心)，C8A、C8B和POC1B (心理運動的病識感)。 在副作用的部分，大約有24%的病患在八週藥物治療期間出現自殺意念，17%出現性副作用。與副作用相關的基因分別為與治療後出現自殺意念相關的LINC00478和ZNF267，以及與性副作用相關的CTBP1、CTBP1-AS、LOC100130872和SPON2。 結論: 我們在服用SSRIs的憂鬱患者中，發現數個基因可能會對不同症狀的藥物治療反應或副作用造成一定程度的影響。在未來的研究中，需要在其他的重鬱症患者身上對這些基因進行檢測，並研究與釐清這些遺傳變異在生物學上的功用。

Introduction: Major depressive disorder (MDD) is a complex and multifactorial disorder with heterogeneous syndromal presentations. Patients’ response to commonly prescribed selective serotonin reuptake inhibitors (SSRIs) varies across individuals and symptoms. Moreover, a high percentage of individuals develop bothersome side effects after treated with SSRIs for depression. Sexual dysfunction is among the most common side effect and the prevalence can be as high as 30% in patients treated with SSRIs. In addition, a serious side effect is treatment emergent suicidal ideation (TESI). It is our goal to search for relevant predictors for treatment response as well as side effects. Inherited genetic factors are considered one of the important factors to influence therapeutic drug response. Conducting pharmacogenetics study is the key to identify genetic variants that modify the effects of antidepressant treatment among depression patients. Material and Method: We recruited 455 MDD outpatients in northern Taiwan. Depression severity was rated using the 21-item Hamilton Rating Scale for Depression (HRSD) in all participants, with a minimum score of 14 at baseline for inclusion, and repeated assessment at weeks 2, 4, and 8. All participants were genotyped using Illumina HumanOmniExpressExome BeadChips. After quality controls, 421 patients (mean age 43.65 years, 71% females) with 4,241,701 genotyped and well-imputed SNPs were retained for analysis. The patients were treated with escitalopram (38.48%), paroxetine (38.48%), fluoxetine (18.29%), or citalopram (4.75%). We first performed exploratory factor analysis to identify syndromal factors for baseline HRSD symptoms. In the subsequently analysis, response of each syndromal factor was used for genome-wide association studies. Treatment response was defined by two variables: “% change” (score difference between baseline and follow-up weeks divided by the baseline score), and “response” (≥50% score improvement from baseline to follow-up weeks). We also analyzed two side effects, TESI, and sexual side effect (SSE), which were defined by score increases in HRSD items during follow-up time-points. We performed genetic association testing using logistic regression for binary response, and mixed model for repeated measurement of treatment score % change, with adjustment for age and gender in all models. Results: There were five empirically derived factors for HRSD, namely core, insomnia, somatic anxiety, psychomotor-insight, and anorexia in the MDD patients. The degree of improvement in syndromal severity at week-4 was ranged from 29.75% (somatic anxiety) to 43.47% (insomnia). Several loci showed suggestive signals with p-values<5×10-6, 184 for %change and 33 for response. The most significant markers were rs146653208 (P=1.12×10-7) for %change in continuous insomnia syndromal factor and rs11759993 (P=1.85×10-6) for insomnia binary response outcome. A number of known genes were mapped for the associations with different syndrome improvement, such as LOC101927653 (core), COL22A1, ENSG00000227101, ENSG00000236990, ENSG00000258214, ENSG00000212599, ENSG00000251931 and CGNL1 (insomnia), IQCJ-SCHIP1, SCHIP1, NRXN3, SNRPN, and TIAM1 (somatic anxiety), CCKAR, MGAM, TAS2R38, SLC5A8, and UTP20 (psychomotor-insight), ENSG00000232197, ENSG00000273100, ENSG00000222974, CYTH1, USP36, ENSG00000252818, GNAZ, and RTDR1 (anorexia) for %change, and CSE1L and STAU1 (core), C8A, C8B, POC1B (psychomotor-insight) for response. In addition, there were approximately 24% patients with TESI and 17% patients with SSE during 8-weeks treatment. There were a number of coding genes were associated with TESI or SSE, such as LINC00478, and ZNF267 for TESI, and CTBP1, CTBP1-AS, LOC100130872, and SPON2 for SSE. Conclusion: We found several genes that may affect treatment response for different empirically defined syndromal factors and side effects among SSRIs treated depression patients. Future studies are needed to replicate these findings in MDD patients and to investigate the biological functions of identified genetic variants.