Novel therapeutic targets and treatment strategies in paediatric brain tumours

The primary goal of Study I was to understand the effects of lithium on the tumour cells in Group 3 medulloblastoma, and whether it extends its radioprotective properties we see in the context of neural progenitors onto the tumour cells. We found that not only did lithium not protect the tumour cells from radiation, but it also possesses anti-tumour properties shown both in vitro and in vivo.

Study II focused on the identification of novel markers of tumour evolution in medulloblastoma. In this study, we performed single-cell RNA sequencing on 14 medulloblastoma (MB) patients, comprised of all 4 molecular subgroups, as well as 6 additional publicly available datasets. By using different computational tools, we found that MB tumour cells cluster according to their copy number variation (CNV) load, with RNA velocity showing that they follow a temporal trajectory from those with low CNV levels to those with high. We then used this information and designated the cluster with a low CNV load as "early" and the clusters with high CNV levels as "later" tumour cells, allowing us to perform a differential gene expression analysis, identifying earlier and later tumour markers. Finally, we used Numbat to differentiate between the early and the late clones in individual samples, which allowed us to identify SOX4 as the only common marker upregulated in the later clones in all Group 3 and Group 4 samples.

The main focus of Study III was to understand the role of microglia in the radiation- induced neuroinflammation proposed to be an important contributor to the loss of cognitive function seen in brain tumour patients that have undergone radiotherapy. We found a delayed microglial response 2 weeks after irradiation exhibiting an interferon-related inflammatory signature, as well as two waves of neuronal asynchrony associated with inflammation and aberrant synaptic regulation, respectively.

The goal of Study IV was to classify microglial cells using deep learning algorithms and to develop spatial statistics pipelines to understand interactions between classified cells. Additionally, it was important for this tool to be adjustable to any cell type and to be applicable to brightfield and fluorescent microscopy.

List of scientific papers

I. Lithium decreases tumour growth in Group 3 medulloblastoma. Yana Ruchiy, Daniela Nunes, Casandra Salinas, Shigeaki Kanatani, Miao Zhao, Gabriela Rosen, Fredrik J Swartling, Carlos FD Rodrigues, Lena-Maria Carlson*, Klas Blomgren *. [Manuscript]

II. Genomic tumor evolution dictates human medulloblastoma progression. Yana Ruchiy#, Ioanna Tsea#, Efthalia Preka, Bronte Manouk Verhoeven, Thale Kristin Olsen, Shenglin Mei, Indranil Sinha, Klas Blomgren, Lena-Maria Carlson*, Cecilia Dyberg*, John Inge Johnsen*, Ninib Baryawno *. Neuro-Oncology Advances. Volume 6, Issue 1, January-December 2024. https://doi.org/10.1093/noajnl/vdae172

III. Microglia adopt temporally specific states after irradiation, correlating with neuronal asynchrony. Alejandro Lastra Romero#, Efthalia Preka#, Giusy Pizzirusso, Luis Enrique Arroyo-García, Georgios Alkis Zisiadis, Nuria Oliva-Vilarnau, Yana Ruchiy, Thea Seitz, Kai Zhou, Arturo Gonzalez Isla, Lara Friess, Ying Sun, Maria Querol Canut, Alia Shamik, Yiran Xu, Changlian Zhu, Carlos F. D. Rodrigues, André Fisahn, Bertrand Joseph, Lena-Maria Carlson, Adamantia Fragkopoulou, Volker M Lauschke, Christer Betsholtz, Ahmed M Osman*, Klas Blomgren *. [Manuscript]

IV. Unravelling microglial spatial organization in the developing human brain with DeepCellMap, a deep learning approach coupled with spatial statistics. Theo Perochon, Zeljka Krsnik, Marco Massimo, Yana Ruchiy, Alejandro Lastra Romero, Elyas Mohammadi, Xiaofei Li, Katherine R. Long, Laura Parkkinen, Klas Blomgren, Thibault Lagache*, David A. Menassa*, David Holcman *. Nature Communications. 16, 1577 (2025). https://doi.org/10.1038/s41467-025-56560-z

# Shared first authorship; * Shared senior authorship