Triple‐negative breast cancer (TNBC) exhibits high rates of early relapse after standard anthracycline–cyclophosphamide (AC) chemotherapy. While chemoresistance and metastasis are classically studied in isolation, mounting evidence suggests they are mechanistically intertwined. This thesis interrogates how AC shapes clonal evolution from primary tumors (PT) to metastatic sites, and establishes a platform to recover and characterize metastasis‐derived circulating tumor cells (CTCs), with the goal of revealing actionable vulnerabilities. I adapted a human TNBC (MDA‐MB‐231) xenograft to a clinically relevant AC regimen that produced stable disease, mirroring chemoresistant residual disease in patients. High‐complexity lineage tracing was combined with single‐cell RNA‐seq to track clonal and transcriptional dynamics across PT, lung metastases, and CTCs. I further developed a high‐throughput workflow to isolate large numbers of viable CTCs from metastasis‐bearing mice, enabling matched multi‐compartment analyses. Targeted metabolomics and Seahorse assays profiled metabolic states. Functional tests included PAEP knockdown and transient pharmacologic modulation of mitochondrial activity in freshly isolated CTCs. AC reduced clonal diversity at the primary site but selectively enriched aggressive clones with high metastatic outgrowth potential; upon re‐challenge, these chemoresistant/pro‐metastatic clones retained and reinforced fitness, revealing durable adaptations. Transcriptionally, chemotherapy imposed a layered, hyper‐adaptive state: treated pro‐metastatic clones coupled glycolysis with mitochondrial resilience, proteostatic and redox programs, and maximized proliferation. Expression‐based discovery identified PAEP (glycodelin) as the top marker of chemoresistant pro‐metastatic clones. PAEP was dispensable in untreated settings but selectively required for chemoresistant metastasis formation, linked to therapy‐induced senescence and secretory program (SASP); its expression predicted poor response and adverse outcomes in TNBC cohorts. Methodologically, I established the first single‐cell, longitudinal dataset incorporating matched PT, lung metastases, and large numbers of viable CTCs recovered at late time points ("metastatic CTCs"). CTCs displayed a unique metabolic phenotype: single‐cell RNA‐seq showed strong OXPHOS enrichment, yet targeted metabolomics revealed global nutrient depletion, high NADPH utilization, and glutathione turnover, consistent with severe oxidative stress. Seahorse assays demonstrated suppressed respiration in CTCs relative to lung metastases, resolving an apparent transcriptome–function paradox via a model in which intravasation triggers an acute adaptive response: respiration is transiently restrained to minimize ROS (short‐term survival), while OXPHOS programs remain transcriptionally primed as a “metabolic scar” for rapid re‐engagement after extravasation. Consistently, brief inhibition of ATP synthase with oligomycin increased ex vivo CTC viability, prolonged intravascular persistence, and enhanced lung seeding, functionally validating the survival advantage of respiratory restraint. Ongoing studies test the reciprocal prediction that oxidative stressors (e.g., arsenic trioxide) preferentially eliminate CTCs. AC reshapes TNBC evolution by pruning diversity yet selecting a restricted set of clones with integrated chemoresistance and metastatic competence. PAEP emerges as a secreted, senescence‐linked effector and therapy‐specific biomarker/target for chemoresistant metastasis. Metastatic CTCs operate under acute metabolic strain and survive by transiently suppressing mitochondrial respiration while maintaining a transcriptionally primed OXPHOS blueprint. These insights open translational avenues: (i) targeting PAEP or downstream pathways, and (ii) exploiting CTC‐specific metabolic vulnerabilities by tipping redox balance or forcing premature respiratory activation to prevent relapse.

LINEAGE TRACING REVEALS PAEP AS A CHEMOTHERAPY-INDUCED METASTATIC DRIVER AND EXPOSES METABOLIC VULNERABILITIES OF CIRCULATING TUMOR CELLS IN BREAST CANCER

DALMASSO, ALBERTO
2025

Abstract

Triple‐negative breast cancer (TNBC) exhibits high rates of early relapse after standard anthracycline–cyclophosphamide (AC) chemotherapy. While chemoresistance and metastasis are classically studied in isolation, mounting evidence suggests they are mechanistically intertwined. This thesis interrogates how AC shapes clonal evolution from primary tumors (PT) to metastatic sites, and establishes a platform to recover and characterize metastasis‐derived circulating tumor cells (CTCs), with the goal of revealing actionable vulnerabilities. I adapted a human TNBC (MDA‐MB‐231) xenograft to a clinically relevant AC regimen that produced stable disease, mirroring chemoresistant residual disease in patients. High‐complexity lineage tracing was combined with single‐cell RNA‐seq to track clonal and transcriptional dynamics across PT, lung metastases, and CTCs. I further developed a high‐throughput workflow to isolate large numbers of viable CTCs from metastasis‐bearing mice, enabling matched multi‐compartment analyses. Targeted metabolomics and Seahorse assays profiled metabolic states. Functional tests included PAEP knockdown and transient pharmacologic modulation of mitochondrial activity in freshly isolated CTCs. AC reduced clonal diversity at the primary site but selectively enriched aggressive clones with high metastatic outgrowth potential; upon re‐challenge, these chemoresistant/pro‐metastatic clones retained and reinforced fitness, revealing durable adaptations. Transcriptionally, chemotherapy imposed a layered, hyper‐adaptive state: treated pro‐metastatic clones coupled glycolysis with mitochondrial resilience, proteostatic and redox programs, and maximized proliferation. Expression‐based discovery identified PAEP (glycodelin) as the top marker of chemoresistant pro‐metastatic clones. PAEP was dispensable in untreated settings but selectively required for chemoresistant metastasis formation, linked to therapy‐induced senescence and secretory program (SASP); its expression predicted poor response and adverse outcomes in TNBC cohorts. Methodologically, I established the first single‐cell, longitudinal dataset incorporating matched PT, lung metastases, and large numbers of viable CTCs recovered at late time points ("metastatic CTCs"). CTCs displayed a unique metabolic phenotype: single‐cell RNA‐seq showed strong OXPHOS enrichment, yet targeted metabolomics revealed global nutrient depletion, high NADPH utilization, and glutathione turnover, consistent with severe oxidative stress. Seahorse assays demonstrated suppressed respiration in CTCs relative to lung metastases, resolving an apparent transcriptome–function paradox via a model in which intravasation triggers an acute adaptive response: respiration is transiently restrained to minimize ROS (short‐term survival), while OXPHOS programs remain transcriptionally primed as a “metabolic scar” for rapid re‐engagement after extravasation. Consistently, brief inhibition of ATP synthase with oligomycin increased ex vivo CTC viability, prolonged intravascular persistence, and enhanced lung seeding, functionally validating the survival advantage of respiratory restraint. Ongoing studies test the reciprocal prediction that oxidative stressors (e.g., arsenic trioxide) preferentially eliminate CTCs. AC reshapes TNBC evolution by pruning diversity yet selecting a restricted set of clones with integrated chemoresistance and metastatic competence. PAEP emerges as a secreted, senescence‐linked effector and therapy‐specific biomarker/target for chemoresistant metastasis. Metastatic CTCs operate under acute metabolic strain and survive by transiently suppressing mitochondrial respiration while maintaining a transcriptionally primed OXPHOS blueprint. These insights open translational avenues: (i) targeting PAEP or downstream pathways, and (ii) exploiting CTC‐specific metabolic vulnerabilities by tipping redox balance or forcing premature respiratory activation to prevent relapse.
17-dic-2025
Inglese
PELICCI, PIER GIUSEPPE
PASINI, DIEGO
Università degli Studi di Milano
227
File in questo prodotto:
File Dimensione Formato  
phd_unimi_R13479.pdf

embargo fino al 20/05/2027

Licenza: Creative Commons
Dimensione 29.29 MB
Formato Adobe PDF
29.29 MB Adobe PDF

I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14242/353908
Il codice NBN di questa tesi è URN:NBN:IT:UNIMI-353908