Biology subjects

Cosma, A.

Publications and source records attributed to Cosma, A..

5 recordsLinked to original sources

bioRxiv2026

A Multidimensional Arithmetical Ontology for High-Parameter Cytometry: Gaussian Primes and Cryptographic Hashing for Cell Population Standardization

Background. The rapid expansion of high-dimensional single-cell technologies has outpaced the development of standardized metadata infrastructure. Current text-based nomenclature for cell populations and gating strategies lacks computational rigor, hindering cross-study reproducibility and automated machine learning. Our previous mathematical frameworks, the Prime Population System (PPS) and the Unique Gating Identifier (UGI), proposed arithmetical solutions but were limited by biological fragility (no inherent link between a protein and its structural isoforms) and by computational bottlenecks caused by exponential integer growth. Objective. To develop a scalable, multidimensional arithmetical ontology that unambiguously identifies cell populations and hierarchical gating strategies without biological ambiguity or computational overflow. Methods. We transition the base marker dictionary from the one-dimensional real axis to the complex plane of Gaussian integers (a+bi). The reviewed human proteome of the UniProt Knowledgebase is paired with Gaussian primes satisfying a > b > 0, each list being totally ordered by a key intrinsic to it, so that the pairing is determined by position alone. Each protein receives a contiguous reserved block of 64 primes: slot 0 addresses the protein with isoform unresolved, and slot k addresses UniProt isoform k, the slot index being read directly off the accession. A Bipartite Complex Ontology encodes positive (a+bi) and negative (b+ai) expression states through Conjugate Pair Symmetry. To bypass the 64-bit integer limits of Goedel exponentiation, hierarchical gating sequences are held as vectorized arrays of complex numbers and compressed by SHA-256 deterministic hashing. Results. Pairing against the reference release assigns 53,296 identifiers, 20,431 protein-level and 32,865 isoform-level, with no collisions and no manual curation at any step. Every isoform record in the release resolves under a single rule, including the 329 records whose accession prefix differs from the entry that cites them. The geometric symmetry conserves biological state, distinguishing positive, negative and isoform-resolved expression upon factorization over [Z][i]. Localized vectorization with SHA-256 hashing yields immutable 64-character hexadecimal fingerprints for entire gating strategies, resolving the integer explosion problem. Conclusion. By combining complex plane topology with cryptographic hashing, this framework provides a universally scalable, computationally lightweight and biologically accurate registry system. The ontology eliminates nomenclature ambiguity and opens the way to machine learning interoperability and centralized global cytometry databases.

bioinformatics↗

bioRxiv2026

Circulating immune profiling reveals impaired monocyte states and trajectories driving immunosuppression in glioblastoma

Glioblastoma (GBM) is an aggressive and lethal brain tumor marked by profound local and systemic immune dysfunction. Yet, the diagnostic and therapeutic relevance of peripheral impairments remains undefined. To clinically dissect their underlying mechanisms and pathological implications, we combined mass and flow cytometry with single-cell RNA-sequencing of peripheral blood mononuclear cells from GBM patients and healthy donors. GBM blood profiles were characterized by heterogeneous changes in classical monocytes, encompassing expanded, reduced and unchanged subsets, presenting distinct functional states, including antigen-presenting, interferon and metabolic subsets. Additional adaptations included myeloid-derived suppressor cell (MDSC) expansion and loss of non-classical monocytes. Trajectory analyses positioned MDSCs as an intermediate state, in continuum with the metabolic subset. Single-cell RNA-sequencing further showed antigen-presenting monocyte propensity to differentiate into tumor-associated macrophages. Circulating monocytes shared a "GBM-classical monocytic signature" exhibiting low MHC class II expression, altered cell-cell communication and increased anti-inflammatory mediators, such as IL1R2 and CD163. Lastly, lymphocyte alterations included decreased proportions of CD4+ T, natural killer (NK) and CD56+ T cells, retaining relatively conserved activation profiles, exemplified by up-regulation of alarmins S100A8/S100A9. These findings map systemic immune reprogramming in GBM, suggesting new avenues for non-invasive biomarker discovery and therapeutic strategies to restore anti-tumor immunity.

cancer biology↗

bioRxiv2024

Adenoviral Delivery of the CIITA Transgene Induces T-Cell-Mediated Killing in Glioblastoma Organoids

The immunosuppressive nature of the tumor microenvironment poses a significant challenge to effective immunotherapies against glioblastoma (GB). Boosting the immune response is critical for a successful therapy. Here, we adopted a cancer gene therapy approach to induce T-cell mediated killing of the tumor through increased activation of the immune system. Patient-based 3D GB models were infected with a replication-deficient adenovirus (AdV) armed with the Class II Major Histocompatibility Complex (MHC-II) Transactivator CIITA gene (Ad-CIITA). Successful induction of surface MHC-II was achieved in infected GB cell lines and primary human GB organoids. Infection with an AdV carrying a mutant form of CIITA with a single amino acid substitution resulted in cytoplasmic accumulation of CIITA without subsequent MHC-II expression. Co-culture of infected tumor cells with either PBMCs or isolated T-cells led to dramatic breakdown of GB organoids. Intriguingly, both wild-type and mutant Ad-CIITA but not unarmed AdV, triggered immune-mediated tumor cell death in the co-culture system, suggesting an at least partially MHC-II-independent process. We further show that the observed cancer cell killing requires the presence of either CD8+ or CD4+ T-cells and the direct contact between GB and immune cells. We did not however detect evidence of activation of canonical T-cell mediated cell death pathways. While the precise mechanism remains to be determined, these findings highlight the potential of AdV-mediated CIITA delivery to enhance T-cell-mediated immunity against GB.

cancer biology↗

bioRxiv2022

T cell landscape definition by multi-omics identifies Galectin-9 as novel immunotherapy target in chronic lymphocytic leukemia

Failure of cancer immunotherapy is linked to T cell exhaustion. To decipher the underlying mechanisms, we explored the T cell landscape in blood, bone marrow and lymph node samples of patients with chronic lymphocytic leukemia (CLL), and spleen samples of a CLL mouse model. By single-cell RNA-sequencing, mass cytometry (CyTOF), and multiplex image analysis of tissue microarrays, we identified a disease-specific accumulation of distinct regulatory T cell subsets and T cell exhaustion stages and their trajectories in CLL lymph nodes. Integration of T cell receptor sequencing data revealed a clonal expansion of CD8+ precursor exhausted T cells (TPEX), suggesting their CLL reactivity. Interactome analyses identified the TIM3 ligand Galectin-9 as a novel immunoregulatory molecule in CLL. Blocking of Galectin-9 in CLL-bearing mice slowed down disease development and reduced the number of TIM3-expressing T cells. Galectin-9 expression correlated with shorter survival of patients with CLL, renal cell carcinoma or glioma. Statement of significanceOur findings for the first time define the T cell landscape in CLL lymph nodes and reshape the current understanding of T cell exhaustion in this malignancy. They further introduce Galectin-9 as novel immune checkpoint with a high potential to overcome resistance to PD1 targeting drugs in CLL and beyond.

cancer biology↗

bioRxiv2021

Standard PBMC cryopreservation selectively decreases detection of nine clinically-relevant T-cell markers

Biobanking is an operational component of various epidemiological studies and clinical trials. Although peripheral blood is routinely acquired and stored in biobanks, the effects of specimen processing on cell composition and clinically-relevant functional markers of T cells still require a systematic evaluation. Here we assessed 25 relevant T-cell markers and showed that the detection of nine membrane markers, e.g., PD-1, CTLA4, KLRG1, CD25, CD122, CD127 and others reflecting exhaustion, senescence and other functions was reduced among at least one T-cell subset following standard processing, although the frequency of CD4, CD8 and regulatory T cells was unaffected. Nevertheless, a six-month-long cryopreservation did not impair the expression levels of many other membrane and all the eight tested intracellular lineage or functional T-cell markers. Our findings uncover that several clinically-relevant markers are particularly affected by processing and the interpretation of those results in clinical trials and translational research should be done with caution.

immunology↗