Lymphoma is currently diagnosed using tissue biopsies and monitored during treatment with PET/CT scans. Although these methods are highly valuable, they may miss very small amounts of remaining disease or detect abnormalities that are not caused by cancer. In this thesis, we investigated whether a bloodplasma test could provide additional information about the presence and activity of lymphoma.

Tumor cells release small fragments of DNA into the bloodstream, known as circulating tumor DNA (ctDNA). By analysing ctDNA, lymphoma can be detected and monitored in a minimally invasive way. To improve the sensitivity of this approach, we developed a new analytical method called MNVista, which identifies highly specific DNA mutation patterns and is able to detect extremely small amounts of tumor DNA.

Using this technique, we studied several types of B-cell lymphoma. We demonstrated that ctDNA not only accurately reflects the genetic characteristics of the tumor but can also be more sensitive than current clinical methods for monitoring treatment response and detecting residual disease.

One of the most remarkable findings was that ctDNA could already be detected in blood samples collected years before lymphoma was diagnosed. In two-thirds of the patients studied, ctDNA was present before diagnosis, and in one case around 9 years earlier. Using our method, we showed that lymphoma can leave detectable molecular traces long before symptoms develop.

These findings contribute to the development of more accurate, less invasive, and more personalised approaches for the diagnosis and monitoring of lymphoma patients.

Classic Hodgkin lymphoma is a blood cancer that mainly affects young adults. Although cancer cells make up less than 5% of the tumor, they can manipulate the abundant immune cells in the surroundings to support tumor growth and enable immune escape. PD-1 checkpoint blockade is a treatment that can restore anti-tumor immunity, but not all patients respond and durable cures remain challenging.

This thesis investigated how the tumor microenvironment influences disease behavior and response to PD-1 blockade, focusing on tumor-associated macrophages (TAMs). First, three candidate tissue biomarkers, c-MET, HLA class II and CD68, were evaluated in samples from an international clinical trial. None reliably predicted patient outcomes, underscoring the need for rigorous biomarker validation.

Next, imaging mass cytometry was used to generate a high-quality single-cell dataset comprising over one million cells from samples of 69 patients. As a proof of concept, PD-L1 expression was profiled on tumor cells and TAMs. PD-L1 was more frequently expressed by TAMs than by tumor cells, and a reproducible analytical framework was established for future spatial studies of PD-1 and other checkpoint pathways.

Finally, laboratory models demonstrated that PD-1 blockade primarily enhances CD4 T-cell activation. Moreover, cancer cells were shown to drive monocyte differentiation into immunosuppressive, PD-L1-positive macrophages. These macrophages suppressed immune-cell activity but also modulated responses to PD-1 blockade. Collectively, these findings demonstrate that both malignant cells and the surrounding immune microenvironment shape treatment response, providing a foundation for developing more personalized and effective immunotherapies for Hodgkin lymphoma.

B-cell lymphomas represent a heterogeneous group of malignancies, with diffuse large B-cell lymphoma (DLBCL) being the most common and aggressive subtype. Aggressive B-cell lymphomas may arise de novo, through transformation from indolent lymphomas such as marginal zone lymphoma (MZL), or in specific clinical contexts such as post-transplant lymphoproliferative disorders (PTLD) and relapsed/refractory DLBCL (R/R DLBCL). This thesis investigates the biological mechanisms underlying lymphoma progression and transformation, and evaluates the utility of circulating tumor DNA (ctDNA) as a biomarker across various aggressive B-cell lymphoma subtypes.
Clinical and molecular analyses identified risk factors for MZL transformation and revealed that transformed MZL frequently acquires features of germinal center B cells. Multi-omics approaches showed only subtle genomic and transcriptomic changes during transformation. Across subtypes, ctDNA emerged as a promising non-invasive biomarker for diagnosis, prognosis, and disease monitoring. In PTLD, ctDNA profiling revealed recurrent genetic alterations and closely reflected tumor characteristics and guide treatment decisions.
In R/R DLBCL, a high ctDNA tumor fraction was associated with poor prognosis. Persistent ctDNA mutations or copy number alterations provided complementary diagnostic value when combined with PET-CT, supporting their integrated use in assessing disease progression and guiding treatment decisions.
Together, these findings underscore ctDNA as a clinically informative biomarker across aggressive B-cell lymphoma subtypes, with potential to improve risk stratification and therapeutic guidance.

Protein synthesis is an energy-intensive process, with translation initiation being a key regulatory step. This thesis focuses on the roles of three translation initiation factors in cancer with the aim to identify novel therapeutic targets.
Amplification of the chromosomal region encompassing the initiation factor 4EBP1 created a synthetic dependency on FGFR1 signaling in cancer. This indicates a therapeutic potential for FGFR1 inhibitors by effectively disrupting phosphorylation of 4EBP1. We showed that 4EBP1 plays a role in regulating genes involved in insulin signaling, glucose metabolism, and the inositol pathway, three pathways contributing to cancer progression.
Treatment of sarcomas with the 4EBP1 inhibitor CR-1-31B induced apoptosis and suppressed growth in vitro and in vivo. Ribosome profiling of cells treated with CR-1-31B identified YAP and TAZ as the critical eIF4A-dependent genes.
Inhibitors of the initiation factor eIF4A like silvestrol showed promising results in treating aggressive lymphomas. However, drug resistance arose through MDR1-mediated efflux of sivestrol. A genome-wide CRISPR/Cas9 screen revealed that activation of NRF2 reduced the efficacy of the eIF4A inhibitor. This resistance could be reverted by blocking FN3K which decreased the stability of NRF2.
Using pancreatic cancer models we showed that genes with long, structured 5’UTRs rely on eIF4A for translation. CR-1-31B treatment effectively suppressed tumor growth by disrupting KRAS-dependent translation, reducing KRAS, MYC, and ERK signaling.
In conclusion, our study demonstrates that disrupting translation initiation, particularly through 4EBP1 and eIF4A inhibition, can suppress tumor progression and this offers new therapeutic strategies for cancer.

This double-degree PhD research investigates two important human viruses, Epstein-Barr virus (EBV) and Hepatitis B virus (HBV), both of which establish lifelong persistent infections and are associated with cancer development. Using an interdisciplinary approach that integrates molecular profiling, functional assays, bioinformatics, and epidemiology, the research aims to understand how these viruses persist, evade immune surveillance, and contribute to chronic disease and cancer, providing insight into mechanisms of viral persistence and pathogenesis.

It examines EBV-encoded miRNAs in B cells and their role in EBV-driven B-cell lymphomas. These miRNAs regulate viral replication, latency programs, and infected B-cell survival by targeting viral and host transcripts. The miRNA profile of the largest EBV-positive classical Hodgkin lymphoma cohort was characterized and compared with EBV-positive Burkitt lymphoma, PTLD, and primary EBV infection controls. BHRF1 miRNAs were specific to PTLD, while BART miRNAs showed stable expression across groups. Highly expressed BART miRNAs were functionally validated, showing reduced cell growth or proliferation upon inhibition. GO analysis linked these miRNAs to apoptosis, proliferation, and cell cycle regulation.
The thesis also reviews occult HBV infection (OBI) in Latin America, highlighting limited and uneven prevalence data and underestimation due to reliance on serology. A study in northeast Colombia identified OBI in 1.1% of blood donors, underscoring limitations of serological screening and the need for DNA-based testing.

Overall, the dissertation highlights host–virus interactions driving EBV and HBV persistence and cancer risk, and supports improved diagnostic and preventive strategies.