試管嬰兒療程中粒線體含量與胚胎潛能之相關性

Abstract

背景說明: 在早期胚胎發育階段，粒線體為細胞重要的能量工廠且呈現明顯的動態變化。而基於胚胎粒線體相關檢測於試管嬰兒療程(IVF)執行的困難度，過去臨床研究主要討論利用胚胎著床前檢測所計算而得的粒線體相對含量於植入胚胎選擇的可行性。然而，早期囊胚滋養層單點切片所得的粒線體含量之於該顆胚胎的意義、與其他胚胎參數的關聯、以及和早期懷孕之間的貢獻度均仍待商榷。因此將粒線體含量應用於選胚實務前，應先針對胚胎粒線體含量和可能之互動因子進行全面分析以確認其角色。

研究設計、規模、持續時間: 本研究共以三階段評估粒線體完整性及其含量應用:初始階段篩選20個不同不孕背景的胚胎用以檢測粒線體DNA(mtDNA)、相關氧化還原基因、凋亡因子細胞色素c (Cytochrome c, Cyt c) 的潛在缺陷或產量變化，因其變動均可能代償性影響粒線體含量。第二階段於NGS程序建立mtDNA比率演算法以計算胚胎切片內相對粒線體含量，並通過分析2,500餘個人類囊胚的滋養層(TE)切片檢體，以評估粒線體含量和各種IVF特徵之間的相互作用，包括和先進的形態動力學參數之關聯。最終階段回溯分析2,283個染色體套數正常的單一囊胚解凍植入(Single euploid embryo cryotransfer, eSET)療程結果，旨在評估粒線體含量對早期懷孕的影響和權重。

參與者/材料、設置、方法: 研究用胚胎檢體依其特定目的分為四群: 粒線體品質測試(n=20)、mtDNA比率演算法的建立(n=150)、粒線體含量與IVF特徵的相關性(n=2,135)、以及粒線體含量對生殖結果的貢獻度(n=2,283)。最初會先對具有各種個體特徵的胚胎進行mtDNA、相關氧化還原基因、粒線體調控細胞凋亡物質Cyt C篩檢，評估以上參數是否和胚胎本質相關聯、進而影響粒線體含量變化；接以建構mtDNA含量演算法並驗證，以確保單點切片所得之粒線體含量的再現性、可靠性和代表性；續用於演算大量檢體的粒線體含量以評估其與胚胎特徵之間的相互作用；最終使用邏輯回歸和機器學習(ML)模型總結粒線體含量在選胚植入中的效益。

主要結果: 粒線體品質相關分析顯示在卵子捐贈者(借卵:年輕健康女性)和IVF族群(自卵:高齡或有其他難孕指標婦女)內、外觀好或不好的胚胎中，mtDNA (4577bp deletion)、氧化還原基因(Nrf2、Keap1、Nqo1)或Cyt c產量均無發現顯著的突變或差異。然而，大規模的粒線體含量與IVF特徵回歸分析顯示，控制多項母體或胚胎參數後粒線體含量與囊胚的時間軸參數具有高度相關性，包含擴張性囊胚的形成天數 (day of expanded blastocyst formation)，分群之95％信賴區間(CI): [自卵] -0.55~-0.23、[借卵] -0.49~-0.14，p<0.01; 擴張性囊胚的形成時間 (timing of expanded blastocyst formation) ([95％CI]: -0.24~-0.01，p=0.04); 和胚胎性別 ([95％CI]: 0.24~1.71，p=0.01)。經控制有相互作用的囊胚時間軸因子由邏輯回歸或ML模型評估，粒線體含量對早期懷孕終點影響均不顯著(勝算比，OR: 0.98-1.00; 曲線下面積，AUC差異<0.01，p=0.54)。由此得知影響懷孕的貢獻度主要來自於與粒線體含量有互動的囊胚時間軸因子，而非其本身。

限制: 本研究主要限制於胚胎樣本的不可重複利用與回溯分析本質，致使直接與粒線體品質或功能相關的實驗無法於胚胎本身執行，代以利用胚胎衍生物或胚胎切片放大產物間接替代，因此無法完全排除胚胎切片所經全基因體放大(Whole genome amplification, WGA)衍生的Allele drop-out (ADO)、或來源切片的母源DNA汙染(maternal contamination)之可能性。此外回溯分析主要受資料來源實驗室的內定規範(如切片胚胎等級要求)、電子病歷所載之格式化紀錄等限制，於數據中所產生的偏態分布影響。

Study question: To explore the correlation between mitochondrial quantity and the blastocyst developmental potential as well as its contribution to early pregnancy.

What is known already: During the early embryo development, mitochondria, the cellular powerhouse, demonstrates a distinctively kinetic change pattern. Past research primarily focused on the viability of mitochondrial DNA (mtDNA) content for embryo selection, considering its accessibility in preimplantation genetic testing (PGT) within IVF programs. Yet, its quantity implications, associations with other biological aspects, and true contribution to early pregnancy remain ambiguous. Therefore, a comprehensive investigation into mitochondrial content of a single trophectoderm biopsy in early-stage blastocyst is imperative before clinical implementation.

Study design, size, duration: Three stages were established to evaluate mitochondrial quantity and its applications. The initial stage involved screening 20 selected embryos to detect potential defects in mtDNA, associated redox genes, and the proapoptotic factor, cytochrome c (Cyt c), encompassing individuals with diverse infertility issues. In the second stage, an mtDNA ratio algorithm to calculate relative mitochondrial content through NGS procedure was established, analyzing a pool of trophectoderm (TE) biopsies over 2,500 embryos to assess interactions between mtDNA quantity and various IVF features, including advanced morphokinetic parameters. The last stage applied 2,283 single euploid blastocyst cryotransfers aimed at demonstrating the impact and significance of mitochondrial quantity on reproductive outcomes.

Participants/Materials, Setting, Methods: The examined embryo cohort was allocated for specific purposes: mitochondrial quality assays (n=20), establishment of relative mtDNA content algorithm (n=150), correlation between mtDNA content and IVF features (n=2,135), and assessment of mitochondrial parameter contribution to reproductive outcomes (n=2,283). Initially, screening for mitochondrial DNA quality and associated redox genetic anomalies was conducted across embryos with various individual characteristics. Subsequent validation of the mitochondrial quantity calculation algorithm ensured the reliability, repeatability, and representativeness of a single TE biopsy. Then the interactions between mtDNA ratios and numerous IVF features were assessed. Finally, the utility of mitochondrial parameters in embryo selection was evaluated using both logistic regression and machine-learning (ML) models.

Main Results: The investigation into mitochondrial quality unveiled no variants or significant alterations in mtDNA (4577 bp deletion), redox genes (Nrf2, Keap1, and Nqo1), or Cyt c across both good and poor embryos derived from either the oocyte donors (a young and healthy population) or IVF cases (women of advanced maternal age and those with other adverse indications). However, mitochondrial quantity (mtDNA ratio) exhibited strong dependence on specific blastocyst timeline factors: the day of expanded blastocyst formation ([95% Confidence intervals, CI] self oocyte: from -0.55 to -0.23; oocyte donors: from -0.49 to -0.14, p<0.01), timing of expanded blastocyst formation ([95% CI] from -0.24 to -0.01, p=0.04), and gender of embryo ([95% CI] from 0.24, to 1.71, p=0.01). Despite these interactions, the mtDNA ratio did not notably impact reproductive outcomes, as assessed by logistic regression (Odds ratio, OR: 0.98~1.00) or ML models (Area under curve, AUC: improvement <0.01, p=0.54). Therefore, factors contributing to embryo competence and successful pregnancy appear to be primarily linked to time-dependent blastocyst variables rather than mitochondrial quantity.

Limitation, reasons for caution: The primary limitations of present study lie in the sample availability and the retrospective nature. Assessing mitochondrial quality faces constraints in direct function evaluation due to embryo feasibility. Therefore, fundamental bias of allele drop-out (ADO) or possible maternal contamination during the whole genome amplification (WGA) in the NGS procedure cannot be completely excluded. Additional influence to the data analysis could be derived from the precondition for biopsied embryos at the operation laboratory and the formatted electronic medical records.