Periodic Reporting for period 1 - STIC-GBM (Spatio-temporal dynamics of immune circuitry in glioblastoma: from single cells to comprehensive models of tumor niches)
Période du rapport: 2024-04-01 au 2026-04-30
Glioblastoma is among the most aggressive malignant brain tumours in adults and remains associated with limited treatment options and poor survival. While immunotherapies have transformed treatment in several cancers, their success in glioblastoma has been limited. A major reason is the highly complex and immunosuppressive tumour microenvironment, in which immune cells are progressively reprogrammed and prevented from mounting effective anti-tumour responses.
STIC-GBM addressed this challenge by combining single-cell genomics, spatial transcriptomics, antibody-based single-cell technologies, computational modelling and functional perturbation approaches. The overall objective was to understand how immune escape develops in glioblastoma across space and time, and to identify cellular programmes or molecular regulators that may be relevant for future therapeutic strategies.
The project had four main scientific objectives: to develop a spatio-temporal single-cell approach for studying immune cell infiltration into glioma; to generate computational methods for analysing spatial tissue niches; to establish workflows for applying these methods to glioma patient material; and to functionally validate candidate regulators of tumour-associated immune programmes.
The pathway to impact combines biological discovery with technological innovation. By generating new methods for analysing complex single-cell and spatial data, the project supports data-driven biomedical research beyond glioblastoma. Its results are relevant for cancer immunology, spatial biology and computational biology, and contribute to the broader European objective of strengthening digital and AI-supported life-science research.
STIC-GBM addressed this challenge by combining single-cell genomics, spatial transcriptomics, antibody-based single-cell technologies, computational modelling and functional perturbation approaches. The overall objective was to understand how immune escape develops in glioblastoma across space and time, and to identify cellular programmes or molecular regulators that may be relevant for future therapeutic strategies.
The project had four main scientific objectives: to develop a spatio-temporal single-cell approach for studying immune cell infiltration into glioma; to generate computational methods for analysing spatial tissue niches; to establish workflows for applying these methods to glioma patient material; and to functionally validate candidate regulators of tumour-associated immune programmes.
The pathway to impact combines biological discovery with technological innovation. By generating new methods for analysing complex single-cell and spatial data, the project supports data-driven biomedical research beyond glioblastoma. Its results are relevant for cancer immunology, spatial biology and computational biology, and contribute to the broader European objective of strengthening digital and AI-supported life-science research.
The project developed and applied a spatio-temporal single-cell methodology to trace immune cell infiltration dynamics in glioma models. This enabled the analysis of immune cell states across both time and tissue context and provided a new experimental framework for studying tumour-immune interactions.
A major achievement was the development of computational methods for high-dimensional single-cell and spatial omics data. The project contributed to scVIVA, a probabilistic framework for modelling niche-specific gene expression variation in spatial transcriptomics, and to CytoVI, a deep probabilistic model for antibody-based single-cell technologies. CytoVI has been accepted for publication in Nature Methods, and scVIVA is under revised consideration at Nature Methods. Both methods have been integrated into the open-source software library scvi-tools.
Workflows were established for the collection and processing of primary glioma patient samples, including protocols for fixed single-cell analysis and spatial transcriptomics-compatible tissue preparation. A prospective cohort of 30 patient samples was collected. The downstream analysis of this cohort could not be completed because stored samples were destroyed during the missile attack on the Weizmann Institute in June 2025.
The project also performed functional perturbation work to identify regulators of tumour-associated macrophage programmes. This led to the identification of Zeb2 as a master regulator of tumour macrophage identity and a potential targetable mechanism in preclinical tumour models. Related work further showed that analogous macrophage programmes contribute to progressive neuroinflammation, extending the relevance of the findings beyond glioma.
A major achievement was the development of computational methods for high-dimensional single-cell and spatial omics data. The project contributed to scVIVA, a probabilistic framework for modelling niche-specific gene expression variation in spatial transcriptomics, and to CytoVI, a deep probabilistic model for antibody-based single-cell technologies. CytoVI has been accepted for publication in Nature Methods, and scVIVA is under revised consideration at Nature Methods. Both methods have been integrated into the open-source software library scvi-tools.
Workflows were established for the collection and processing of primary glioma patient samples, including protocols for fixed single-cell analysis and spatial transcriptomics-compatible tissue preparation. A prospective cohort of 30 patient samples was collected. The downstream analysis of this cohort could not be completed because stored samples were destroyed during the missile attack on the Weizmann Institute in June 2025.
The project also performed functional perturbation work to identify regulators of tumour-associated macrophage programmes. This led to the identification of Zeb2 as a master regulator of tumour macrophage identity and a potential targetable mechanism in preclinical tumour models. Related work further showed that analogous macrophage programmes contribute to progressive neuroinflammation, extending the relevance of the findings beyond glioma.
STIC-GBM generated results beyond the state of the art in both experimental and computational single-cell biology. The spatio-temporal single-cell approach enables immune cell infiltration dynamics to be studied with temporal information and transcriptome-wide resolution in tissue. This provides information that could not be obtained from conventional single-cell or spatial methods alone.
The computational outputs of the project also represent substantial advances. CytoVI addresses a major limitation of antibody-based single-cell technologies: the difficulty of integrating datasets generated with different antibody panels or platforms. scVIVA addresses the need to model how spatial cellular environments shape gene expression variation. By integrating both methods into scvi-tools, the project ensured that these advances are available to the wider research community through public code, documentation and tutorials.
The identification of Zeb2 as a regulator of tumour-associated macrophage programmes provides a mechanistic result with potential translational relevance. Further uptake will require additional validation, larger-scale benchmarking, clinical and preclinical follow-up studies, and continued support for open-source software maintenance. For potential diagnostic or therapeutic applications, further development may require intellectual property support, multicentre validation, regulatory assessment and collaboration with clinical or industrial partners.
The computational outputs of the project also represent substantial advances. CytoVI addresses a major limitation of antibody-based single-cell technologies: the difficulty of integrating datasets generated with different antibody panels or platforms. scVIVA addresses the need to model how spatial cellular environments shape gene expression variation. By integrating both methods into scvi-tools, the project ensured that these advances are available to the wider research community through public code, documentation and tutorials.
The identification of Zeb2 as a regulator of tumour-associated macrophage programmes provides a mechanistic result with potential translational relevance. Further uptake will require additional validation, larger-scale benchmarking, clinical and preclinical follow-up studies, and continued support for open-source software maintenance. For potential diagnostic or therapeutic applications, further development may require intellectual property support, multicentre validation, regulatory assessment and collaboration with clinical or industrial partners.