從分子驅動到細胞群集：MCM4 與 CTM 作為癌症轉移與復發的互補性生物標記

Abstract

本論文題為「From Molecular Drivers to Cellular Clusters：MCM4 and Circulating Tumor Microemboli as Complementary Biomarkers for Cancer Metastasis and Recurrence」，旨在從「分子驅動因子」與「細胞群集結構」兩個層級，建構一套探討癌症轉移與復發風險的互補性生物標記架構。 研究背景在於，肺腺癌(LUAD)與乳癌皆以遠端轉移與治療後復發為主要致死原因。現行多數生物標記著重單一尺度，例如組織切片的分子表現或血液中循環腫瘤細胞(CTCs)計數，對於轉移風險與疾病動態的反映仍有限。本研究提出「分子—細胞軸(molecular–cellular axis)」概念：上游以複製起始因子MCM4代表基因體不穩定與侵襲表型的分子驅動力，下游以循環腫瘤微栓子(circulating tumor microemboli, CTM)代表具功能意義的轉移單位，嘗試從雙尺度整合解讀轉移與復發風險。 本論文分為兩個互補研究。研究一(Study I)聚焦乳癌追蹤門診之CTM。共納入台灣恩主公醫院及新光吳火獅紀念醫院11名乳癌病人，於術後及治療追蹤期間收集 80 筆周邊血液檢體。利用自組裝細胞陣列(self-assembled cell array, SACA)晶片與自動化多光譜影像平台，結合PanCK/Hoechst/CD45/CD3免疫螢光標記與細胞形態判讀，自血液中偵測及計數CTCs與CTM，分析其在疾病痊癒(disease-free, DF)與疾病復發(disease recurrence, DR)族群間的分布差異，並建立ROC曲線比較預測效能。進一步在CTM陽性的21筆檢體(7名病人)中使用PanCK/Hoechst/CD45/ER/CD8標記，將CTM依周邊免疫細胞組成與腫瘤細胞比例分為四種亞型，並於個案層級整合CA15-3、CEA、CTC/CTM計數與影像型態，探討 ER 與 CD8 表現隨治療與時間的變化。結果顯示，在80筆檢體中共偵測到342顆CTCs與31個CTM。CTCs在DF與DR之間分布高度重疊，ROC 曲線之AUC約0.57，預測力有限；反之，CTM計數在DF與DR之間分布差異較明顯，AUC接近0.90，顯示CTM對於區分復發狀態具有更佳判別度。影像分析顯示，CTM可細分為僅含 PanCK⁺腫瘤細胞的同質性團塊，以及合併CD45⁺白血球、CD3⁺/CD8⁺淋巴球之異質性團塊，且腫瘤細胞與免疫細胞的數量比例在不同病例與時間點呈現明顯變化。於個案追蹤曲線中，部分DF病人之CTM保持ER⁺/CD8⁺微環境，而一名DR病人在荷爾蒙治療與疾病進展過程中，CTM出現ER與CD8同步轉為陰性的現象，提示CTM不僅能作為復發量化指標，亦可能反映局部腫瘤—免疫微環境的質變。 研究二(Study II)則以MCM4為核心，探討肺腺癌轉移相關分子驅動因子。透過四個 GEO 微陣列資料集(GSE32863、GSE27262、GSE40275、GSE33356)及TCGA-LUAD RNA-seq資料，篩選出共同差異表達基因(differentially expressed genes, DEGs)，利用癌症Hallmarks、Gene Ontology與KEGG進行功能富集，再透過STRING/Cytoscape/cytoHubba建構蛋白交互作用網絡並選出前十個節點度最高的 hub genes。接著結合TCGA、Lung Cancer Explorer、TNMplot及Human Protein Atlas平台，評估這些hub genes在LUAD之mRNA與蛋白質表現、臨床分期、淋巴結轉移(特別是AJCC N2)及整體存活與疾病進展後存活之關聯，並分析MCM4與多種基質金屬蛋白酶(MMP1、MMP9、MMP12、MMP13)之表達相關性。分析結果顯示，十個hub genes多與細胞分裂與DNA複製相關，在多數資料集中於LUAD顯著上調。當中MCM4在LUAD及LUSC中均明顯高表現，且在TCGA中隨臨床分期(I–IV)上升；蛋白層級與組織免疫染色亦呈現腫瘤組織強於正常肺組織的趨勢。高MCM4表現與較差整體存活及疾病進展後存活顯著相關，尤其在AJCC N2亞群中更為明顯。TNMplot顯示MCM4自正常肺、原發LUAD至轉移LUAD呈現漸進式升高，且與MMP9、MMP12等侵襲相關基因表現高度正相關，支持MCM4可能透過促進基質降解與細胞浸潤參與轉移過程，為潛在的轉移相關標記與治療標的。 綜合兩個研究，本論文提出一套由上游分子驅動因子MCM4至下游CTM細胞團塊的「分子—細胞」整合框架，說明基因體複製壓力與微環境塑形如何在不同癌別與不同尺度上共同影響轉移與復發風險。臨床上，結合分子標記(例如：LUAD中的MCM4)與功能性細胞團塊指標(乳癌追蹤中的CTM數量與亞型)，有助於更精準的風險分層、疾病監測與治療反應評估。惟本研究亦受限於乳癌樣本數較少、CTM極度稀有以及MCM4研究主要仰賴公開轉錄體資料庫等因素，未來需透過更大規模前瞻性臨床研究與功能性實驗驗證，以進一步確認MCM4與CTM在多癌別中的可轉譯性與實際應用價值。

Cancer metastasis and recurrence represent the leading causes of cancer-related mortality, yet current diagnostic biomarkers for monitoring disease progression remain inadequate in terms of timeliness and accuracy. This study proposes an innovative multi-level liquid biopsy framework that integrates molecular-level and cellular cluster-level biomarkers to enhance the predictive capacity for cancer metastasis and recurrence. Through bioinformatics analysis, minichromosome maintenance protein 4 (MCM4) was identified as a critical molecular driver of lung adenocarcinoma (LUAD) metastasis, representing molecular aberrations in cancer progression. Concurrently, in breast cancer patients, circulating tumor microemboli (CTM) were validated via microfluidic platforms as a cellular manifestation with superior predictive capability. The research background demonstrates that despite significant advances in surgical intervention, radiotherapy, chemotherapy, targeted therapy, and immunotherapy, the five-year survival rate for lung adenocarcinoma patients remains disappointing, while the recurrence rate for intermediate to advanced breast cancer patients within three years post-treatment remains as high as 50%. Traditional serum markers such as carcinoembryonic antigen (CEA) and cancer antigen 15-3 (CA15-3) exhibit insufficient sensitivity for early-stage or localized lesions, severely limiting their clinical utility. These inherent limitations create an "unmet medical need," compelling researchers to explore more advanced, multimodal approaches. This study demonstrates that the synergistic combination of MCM4 and CTM biomarkers provides a comprehensive dual perspective on cancer progression. MCM4, as a molecular driver, reveals tumor aberrations at the genomic and transcriptomic levels, providing early signals regarding disease aggressiveness and metastatic propensity. CTM, as a cellular functional marker, directly captures tumor cell clusters with metastatic potential in the circulatory system, providing tangible evidence of imminent metastatic risk. This multi-level integration from molecular to cellular dimensions represents a paradigm shift in the field of cancer biomarkers, demonstrating exceptional value in predicting disease recurrence and establishing a novel paradigm for personalized long-term patient management. Study I Abstract–CTM in Breast Cancer Follow-Up (2024). Distant metastasis and treatment resistance account for the majority of mortality in breast cancer patients. Identifying suitable biomarkers that enable accurate and timely monitoring of disease status and treatment outcomes remains a critical need. Emerging evidence suggests that circulating tumor cell clusters demonstrate stronger associations with metastatic progression and adverse survival outcomes compared to isolated tumor cells, positioning them as valuable prognostic indicators. Methods: This investigation analyzed 80 specimens obtained from 11 breast cancer patients undergoing post-treatment surveillance. A microfluidic platform coupled with advanced imaging technology was employed to quantify circulating tumor cells and microemboli (CTCs/CTM), evaluate their distribution across patient risk groups, and determine their capacity to forecast disease status following therapeutic intervention. The study further characterized the compositional features and classification of CTM subtypes. Results: CTM demonstrated superior discriminatory power relative to isolated CTCs in stratifying patients and forecasting disease progression during post-treatment monitoring, as evidenced by enhanced area under the curve performance. Based on compositional analysis, CTM were classified into four distinct subtypes. When integrated with conventional markers CEA and CA15-3, CTCs and CTM enumeration provided effective monitoring of disease dynamics and therapeutic response in individual patients throughout the follow-up period. Conclusions: CTM and their subtypes represent valuable prognostic markers with significant potential for investigating metastatic mechanisms and enabling continuous surveillance of cancer patients during long-term follow-up. Study II Abstract – MCM4 in LUAD Lung adenocarcinoma (LUAD) is the most common type of non-small cell lung cancer, often diagnosed at advanced, metastatic stages, leading to poor clinical outcomes. Discovery of pivotal biomarkers associated with metastatic progression is essential for advancing early detection strategies and developing targeted therapeutic approaches. Methods: Our investigation examined four microarray datasets from the Gene Expression Omnibus (GSE32863, GSE27262, GSE40275, and GSE33356) alongside The Cancer Genome Atlas data to discover genes exhibiting differential expression patterns in LUAD. We performed functional enrichment analysis of these differentially expressed genes employing Gene Ontology, Kyoto Encyclopedia of Genes and Genomes pathway analysis, and Cancer Hallmark Enrichment Plot methodologies. Network-based hub gene identification was accomplished using Cystoscope software. These hub genes underwent comprehensive evaluation for their expression patterns, survival implications (utilizing the Kaplan-Meier plotter platform), and clinical associations through multiple databases (TCGA, Lung Cancer Explorer, TNMplot, and the Human Protein Atlas). Results: Our analysis revealed 333 genes with consistent dysregulation across datasets, demonstrating enrichment in biological pathways implicated in metastatic processes, such as neovascularization, immune evasion mechanisms, and extracellular matrix remodeling. Network analysis identified ten central hub genes (AURKA, TOP2A, CCNB2, CENPF, MCM4, TPX2, KIF20A, ASPM, MELK, and NEK2). MCM4 demonstrated pronounced elevation in LUAD specimens and exhibited significant correlation with diminished overall survival outcomes. Furthermore, MCM4 expression levels showed association with post-progression survival duration and indicators of invasive behavior. Both immunohistochemical staining and transcriptomic profiling validated MCM4 upregulation at the messenger RNA and protein expression levels. MCM4 expression demonstrated positive associations with multiple matrix metalloproteinase family members, corroborating its participation in facilitating tumor invasive properties. Conclusions: MCM4 emerges as a promising novel biomarker for assessing LUAD metastatic potential and prognostic stratification. Its reproducible overexpression, correlation with metastatic indicators, and clinical relevance position it as a viable candidate for diagnostic applications or therapeutic targeting in advanced LUAD management.