Despite its aggressiveness, glioblastoma often lacks a single dominant, easily druggable oncogenic driver; instead, many tumours converge on aberrant activation of the PI3K/Akt/mTOR signalling network, which is active in almost 90% of cases.

 However, glioblastoma is not simply “PI3K-addicted”: pharmacological inhibition is frequently cytostatic and—if it reduces proliferation at the wrong time—can counteract chemotherapy, meaning that efficacy depends strongly on cell state and treatment schedule.

 In patient-derived glioblastoma models, we found that PI3K signalling can have distinct, cell-specific functions within the same tumour—contributing to the enhanced therapy resistance of stem-like cells, while primarily regulating motility/invasion of more differentiated populations—highlighting why pathway inhibition must be integrated thoughtfully into rational, timed combination therapies, rather than applied as uniform maximal blockade.

Medulloblastoma (MB) is the most common malignant solid brain tumour in children. It typically arises in the cerebellum and is treated with an intensive multimodal regimen. While cure rates have improved, therapy can cause substantial long-term sequelae, and relapse remains difficult to treat. Molecularly, MB is an umbrella diagnosis comprising distinct subgroups (classically WNT, SHH, Group 3, Group 4) with different biology and prognosis.

A key translational concept we pursue in MB is to break intrinsic resistance to therapy-induced cell death so that effective treatment might be achievable with less toxic intensity, drawing from our decades-long experience in the field of apoptosis research.

In parallel, our broader brain-tumour program emphasizes that therapeutic success is also shaped by tumour–microenvironment interactions. Work on extracellular matrix cues such as reelin illustrates how modulating cell–matrix behaviour can alter malignant phenotypes (e.g., adhesion and motility) and provides a conceptual framework for adding microenvironment-focused interventions to cell-death–based strategies in brain tumours.

Our therapeutic work is guided by the idea that malignant brain tumours evolve under treatment pressure—and that therapy can be designed not only to “hit harder”, but to shape how tumour populations move through an adaptive (fitness) landscape. Instead of only creating a single deep “valley” by maximal cell kill, the goal is to steer evolutionary trajectories: anticipating which resistant peaks become accessible under a given intervention and then using sequence, timing, and combinations to constrain escape routes (“fitness landscape steering”).

One concrete translation of this evolutionary thinking is the RIST strategy—a metronomic combination regimen integrating targeted pathway inhibition with cytotoxic therapy (Rapamycin, Sunitinib, Irinotecan, Temozolomide). In glioblastoma models, this combination increased apoptosis and reduced proliferation compared with single agents, and in compassionate-use settings it was able to stabilize tumour burden for prolonged periods in selected, heavily pretreated patients, consistent with a “chronification” concept rather than expecting complete eradication from a single line of therapy.

In parallel, our recent translational focus addresses a second, complementary resistance principle in brain tumours: multicellular connectivity. Building on the concept that glioma cells form therapy-resistant networks, we pursue the Alcatraz Strategy—a multimodal roadmap to physically and functionally isolate tumour cells, analogous to preventing coordinated escape and repair. This framework combines (i) advanced neurosurgical approaches such as supramarginal resection targeting infiltrative network zones beyond the enhancing mass, with (ii) morphological network disruption (e.g., Meclofenamate to impair tumour microtube formation/outgrowth), and (iii) functional disconnection, including inhibition of gap-junction–mediated coupling and reduction of excitatory neuronal input into the malignant network.

These approaches are tightly embedded in our long-term collaborations with the Department of Neurosurgery, where the surgical and translational neuro-oncology programs provide the clinical and experimental backbone for implementing and testing such concepts—from refined resection strategies to perioperative pharmacological add-ons and trial-driven biomarker readouts of network disruption and tumour adaptation.

Our group runs an active science-outreach program that aims to bring biomedical research closer to the public—especially to children and adolescents—by making complex topics tangible, interactive, and relevant to everyday life. Building on our work on misinformation and the observation that young people and parents are specifically targeted by anti-science narratives, we translate core scientific skills (e.g., source evaluation, separating facts from opinions, understanding experimental design, and updating beliefs based on evidence) into age-appropriate, engaging formats.

We go out to schools, organise exhibitions and give talks to the interested public