CanCell Associate Members
Manuela Zucknick
Institute of Basic Medical Sciences, University of Oslo
Statistical modelling of cancer drug response outcomes in ex vivo experiments
In 2025 we have finalized several projects where we adapted penalized regression models to address typical challenges we encounter when predicting drug response outcomes in large-scale pharmacogenomic screens (D’Alessandro et al. 2025, https://doi.org/10.1002/sim.70148, and Asenso & Zucknick 2025, https://doi.org/10.1080/02664763.2025.2567981): joint prediction of many drugs in parallel, efficient integration of multiple omics data sources in one prediction model, and the need to allow for different effect sizes in different subgroups. An example for the latter case is synthetic lethality, i.e. when a genetic marker is only predictive for drug response in the presence of a mutation in a second gene, because both genes need to be inactivated to result in cell death. We have also continued our statistical model advancements for ex vivo cancer drug screens, extending our bayesynergy (https://doi.org/10.1093/bib/bbab251) model for combination experiments to various multi-output Gaussian process models, which can be applied with or without selection of potential biomarker candidates. The following sub-projects serve as illustrations for the broader modelling framework: Biomarker identification for drug response outcomes (SQUAB): SQUAB can simultaneously perform dose-response estimation and biomarker discovery through a multi-output Gaussian process formulation that allows estimation of highly non-linear dose-response relationships, while incorporating high-dimensional omics data for biomarker discovery. Selection of potential biomarkers is achieved by a regularized horseshoe prior within the kernel construction. This new model leads to better statistical power for biomarker discovery in typical drug combination screens with small numbers of cell lines. Collaboration with Enserink lab. Publication in preparation. Multi-output model for spheroid cultures: We have utilized the multi-output Gaussian process framework to estimate drug response metrics from single-spheroid assays. The model allows us to leverage the correlation between multiple response metrics, such as cell death, growth, and autophagy. Additionally, we can factor in the correlation between repeated measurements of these metrics over time, which allow a more precise estimation. Collaboration with Knævelsrud lab. Further extensions planned as part of a new PhD project which will start in 2026 and in submitted joint funding applications.
Terje Johansen
Autophagy Research Group, UiT The Arctic University of Norway
The archetypal autophagy receptor NBR1: Functional anatomy and mechanisms employed in selective autophagy
Neighbor of BRCA1 (NBR1) is the archetypal selective autophagy receptor found in all major eukaryotic phyla with instances of secondary loss in some organisms. NBR1 is much less studied than its paralog, the first discovered autophagy receptor p62/Sequestosome 1 (SQSTM1) appearing in the evolution at the advent of the Metazoans (Rasmussen et al. 2022). NBR1 and p62 both recognize ubiquitinated substrates, engage the autophagy core machinery and connect them to forming autophagosomes via LIR-ATG8 protein interactions and co-localize in p62 bodies formed by LLPS (liquid-liquid phase separation). NBR1 is twice as large as p62 and contains several domains not found in p62. NBR1 can directly interact with membranes via its amphipathic helix (AH) domain, which p62 cannot. We aim with this 4-year project to obtain a more complete mechanistic understanding of structure function relationships and the roles played by NBR1 and its various domains in selective autophagy processes.
In the first year we have studied the role of the AH domain and its interaction with membranes and with the UBA domain of NBR1. We will complete this study early 2025 in close collaboration with Andreas Brech, Sebastain Schultz and Athanasios Kournoutis at CanCell. A manuscript is being written up on the regulatory role of the AH domain.
We have also worked on a “bottoms up” approach where we so far have used affinity purification mass spectrometry (AP-MS) to identify protein interaction partners of NBR1. We are presently working on validating interactors from a long list of candidates. We will go on using proximity interactomics with UltraID and/or APEX2 and have performed TMT labeling for quantitative proteomics experiments.
We have identified proteins interacting with two different parts of NBR1 (FW- and the ZZ -CC domains) and tested two buffer systems for the affinity purification. Validation experiments are initiated using imaging and in vitro GST pull down assays.
Rasmussen, N.L., Kournoutis, A., Lamark, T. and Johansen, T. 2022. NBR1: The archetypal selective autophagy receptor. J. Cell Biol. 221:e202208092. doi: 10.1083/jcb.202208092.
Tuula Nyman
National Network for Advanced Proteomics Infrastructure Department of Immunology, Oslo University Hospital
Advanced Mass Spectrometry-Based Proteomics
CanCell associate member Tuula Nyman is heading the Proteomics core facility (PCF) in the Department of Immunology, UiO & OUS and is the Project Manager for the National Network for Advanced Proteomics Infrastructure (NAPI, https://www.napi.uio.no/) funded by the Research Council of Norway (RCN).
We provide advanced mass spectrometry-based proteomics and associated data analysis to CanCell members. In 2024 we have contributed to several projects including a collaborative project with another CanCell associate member, Professor Emmet McCormack to characterise a novel three-dimensional model of high-grade serous ovarian carcinoma peritoneal metastases using a porcine decellularized peritoneal-derived ECM scaffold (peritoneal matrix; PerMa). For this, PCF characterized the proteome composition of the native tissues and decellularized scaffolds; and the results showed that porcine PerMa represents a biologically relevant tissue scaffold with similarities to human PerMa in structure, physical properties, and biomolecular composition. There has been growing interest within multiple CanCell groups to study protein secretion using proteomics; both global protein secretion and extracellular vesicle (EV)- mediated protein secretion. PCF/Nyman has a lot of experience in this type of studies from different sample types including cell culture and plasma samples, and we have participated in optimising the sample preparation strategies required, as well as provided advanced data analysis and bioinformatic characterisation of the data.
With proteomics it has become increasingly clear that we should also get spatial information on protein abundance in the samples. PCF has set up mass spectrometry imaging (MSI)-technique which is a novel, innovative MS methodology that generates unique in situ information on abundance and spatial distribution of molecules such as peptides, lipids and metabolites in tissue sections or cell cultures. Importantly, the instrument at PCF allows for accurate MS/MS-based molecule identification. MSI can be widely used as part of cancer research in future CanCell projects.
Emmet Mc Cormack
Department of Clinical Science, University of Bergen
The group’s core activities focus on the development, characterization, and application of clinically relevant models (patient-derived xenografts (PDX), Hu-PDX, Comparative animal models, in vitro decellularized scaffold platforms) and multimodal molecular imaging strategies and how they believe judicious application of these technologies can expedite clinical development of novel therapeutics. More recently, the group has focused on the development of novel immunotherapies (CAR-T and BiTes).
Within the context of CanCell, our group has active collaborations with both the Rusten and Nyman groups. In collaboration with Tor Erik Rusten, we have preclinically examined the impact of mammalian cystine transporters SLC7A9 and SLC7A11 in the malignant normal cell proliferation settings. Experiments have been performed in xenograft murine models, and the manuscript is currently being finalized.
In collaboration with the Tuula Nyman, we published a novel 3D ovarian cancer metastasis model using a decellularised peritoneal matrix to study therapy response. High-grade serous ovarian carcinoma (HGSOC) presents a significant therapeutic challenge. Late-stage disease is frequently associated with peritoneal carcinomatosis. The peritoneal metastases exhibit a unique tumour microenvironment (TME) distinct from the primary tumours and other metastatic sites. Understanding the critical influence of the extracellular matrix (ECM) in shaping the tumour phenotype is essential for the development of effective new therapies. This study introduced a three-dimensional (3D) model of HGSOC peritoneal metastases using a porcine decellularised peritoneal-derived ECM scaffold, referred to as peritoneal matrix (PerMa). We showed that the decellularisation maintains the structural integrity and composition of ECM molecules. Comparative analysis revealed structural, compositional, and mechanical similarities between porcine and human peritoneal matrices, underscoring the porcine model's translational relevance for modelling human peritoneum physiology. The PerMa supported the 3D growth of HGSOC cell lines. The model enabled the assessment of sensitivity to traditional chemotherapy and novel cell-based immunotherapy through confocal imaging and quantification of cell volume. Our model offers a valuable platform for investigating peritoneal carcinomatosis in HGSOC, with the potential to contribute significantly to developing novel therapeutic approaches.
The paper was published in EBiomedicine; A 3D ovarian cancer metastasis model using a decellularised peritoneal matrix to study therapy response. Gjerde CH, EBioMedicine. 2026 Feb;124:106135.
Nikolina Sekulic
Institute of Basic Medical Sciences, University of Oslo
Sekulic Lab – Structural Mechanisms of Genome Stability and Disease
Our research focuses on understanding how structural and allosteric mechanisms regulate protein complexes that safeguard genome stability. We are particularly interested in centromeres and kinetochores, which are essential for accurate chromosome segregation during cell division and whose dysfunction can lead to cancer and other diseases.
We combine biochemistry, biophysics, and advanced structural biology approaches—including X-ray crystallography, cryo-electron microscopy (cryoEM), hydrogen-deuterium exchange mass spectrometry (HDX‑MS), and computational modeling—to visualize dynamic protein assemblies and map interaction surfaces at high resolution. Recently, our work has expanded into cellular context through cryo-tomography, in close collaboration with Andreas Brech from CanCell, enabling us to capture the architecture of protein complexes within intact cells. Our research on chromosome segregation has also been strengthened by new funding to explore these mechanisms in diverse organisms, including fish, with potential applications in biotechnology and medicine.
As an associate member of CanCell, we have collaborated with CanCell researchers to deliver high-resolution insights into relevant protein-protein interactions using experimental or computational approaches (AlphaFold-based structural predictions), which support mechanistic insights and guide experimental design. The CanCell network offers shared expertise and infrastructure, allowing us to connect fundamental structural knowledge to disease-relevant pathways. Such collaborations foster innovation in cancer and genome stability research.
Our ultimate goal is to translate structural knowledge into actionable strategies—whether by guiding vaccine design, engineering therapeutic proteins, or identifying novel drug targets—to improve human health.
