Biology subjects

Boxberg, M.

Publications and source records attributed to Boxberg, M..

4 recordsLinked to original sources

bioRxiv2026

Pan-cancer virtual spatial transcriptomics from routine histology with Phoenix

Spatial transcriptomics links gene expression to tissue architecture, providing a mechanistic view of cellular organization. Yet existing datasets cover few donors and miss the complexity of human disease. Experimental costs remain prohibitive, and large-scale profiling is impractically slow for population-level studies. Accurate computational methods are urgently needed. Predicting gene expression from standard histology, however, remains an open problem, as current approaches transfer poorly to unseen cohorts and diseases. Here, we present Phoenix, a (latent) flow matching generative model that infers pan-cancer spatially resolved single-cell gene expression with high accuracy. Phoenix analyzes treatment response in silico: Applied to 763 head and neck cancer patients, it identified three new spatial biomarkers that we validated across two cancers (breast cancer, n = 84; ovarian cancer, n = 157) and treatment regimens (platinum, trastuzumab). Phoenix generalizes beyond carcinomas: In a large sarcoma cohort (802 tissue microarray cores), it accurately predicted cell-type-specific signatures in held-out samples and captured chemotherapy-induced immune remodeling. Phoenix also extends across species: In a mouse model, it accurately predicted the expression of pancreatic cancer lineage markers and the mutant mKras^G12D allele in silico. In total, we evaluated Phoenix on over 10,000 patients. Our results establish virtual spatial transcriptomics as a scalable framework for studying tissue organization, therapeutic response, and disease mechanisms.

bioinformatics↗

bioRxiv2022

Proteogenomic analysis reveals RNA as an important source for tumor-agnostic neoantigen identification correlating with T-cell infiltration

Systemic pan-tumor analyses may reveal the significance of common features implicated in cancer immunogenicity and patient survival. Here, we provide a comprehensive multi-omics data set for 32 patients across 25 tumor types by combining proteogenomics with phenotypic and functional analyses. By using an optimized computational approach, we discovered a large number of novel tumor-specific and tumor-associated antigens including shared common target candidates. To create a pipeline for the identification of neoantigens in our cohort, we combined deep DNA and RNA sequencing with MS- based immunopeptidomics of tumor specimens, followed by the assessment of their immunogenicity. In fact, we could detect a broad variety of non-wild type HLA-binding peptides in the majority of patients and confirmed the immunogenicity of 24 neoantigens. Most interestingly, the majority of total and immunogenic neoantigens originated from variants identified in the RNA dataset, illustrating the importance of RNA as a still understudied source of cancer antigens. Moreover, the amount of these mainly RNA-based immunogenic neoantigens correlated positively with overall CD8+ tumor-infiltrating T cells. This study therefore underlines the importance of RNA-centered variant detection for the identification of shared biomarkers and potentially relevant neoantigen candidates. Statement of significanceThe significance of this study lies not only in the potential of our optimized proteogenomic workflow for the discovery of neoantigens (in particular RNA-derived neoantigens) for clinical application, but sheds light on the entity-agnostic prevalence of HLA class I peptide presentation of RNA processing events to be used for tumor targeting.

cancer biology↗

bioRxiv2021

Spatial and temporal plasticity of neoantigen-specific T-cell responses bases on characteristics associated to antigen and TCR

Neoantigens derived from somatic mutations have been demonstrated to correlate with therapeutic responses mediated by treatment with immune checkpoint inhibitors. Neoantigens are therefore highly attractive targets for the development of personalized medicine approaches although their quality and associated immune responses is not yet well understood. In a case study of metastatic malignant melanoma, we performed an in-depth characterization of neoantigens and respective T-cell responses in the context of immunotherapy with Ipilimumab. Three neoantigens identified either by immunopeptidomics or in silico prediction were investigated using binding affinity analyses and structural simulations. We isolated seven T-cell receptors (TCRs) from the patient immune repertoire recognizing these antigens. TCRs were compared in-vitro and in-vivo with multi-parametric analyses. Identified immunogenic peptides showed similar binding affinities to the human leukocyte antigen (HLA) complex and comparable differences to their wildtype counterparts in molecular dynamic simulations. Nevertheless, isolated TCRs differed substantially in functionality and frequency. In fact, TCRs with comparably lower functional avidity and higher potential for cross-reactivity provided at least equal anti-tumor immune responses in vivo. Of note, these TCRs showed a reduced activation pattern upon primary in vitro stimulation. Exploration of the TCR-{beta} repertoire in blood and in different tumor-related tissues over three years, offered insights on the high frequency and particular long-term persistence of low-avidity TCRs. These data indicate that qualitative differences of neoantigen-specific TCRs and their impact on function and longevity need to be considered for neoantigen targeting by adoptive T-cell therapy using TCR-transgenic T cells. Statement of translational relevanceImmunotherapy has demonstrated high efficacy in diverse malignancies. Neoantigens derived from mutations provide promising targets for safe and highly tumor-specific therapeutic approaches. Yet, single determinants of an effective and enduring T-cell mediated tumor rejection are still not well understood. We analyzed in detail seven neoantigen-specific T-cell receptors (TCRs) derived from a melanoma patient targeting three different altered peptide ligands identified by mass spectrometry and prediction analyses. Functional characterization of these TCRs revealed potent anti-tumor reactivity of all TCRs. Of special interest, TCRs with comparably lower affinity demonstrated effective in vivo activity as well as dominant spatial and temporal distribution in blood and tissue. Functional differences of TCR may require further T-cell and/or TCR engineering and should be considered for future clinical trial designs.

immunology↗

bioRxiv2019

Multi-task learning of a deep k-nearest neighbour network for histopathological image classification and retrieval

Deep neural networks have achieved tremendous success in image recognition, classification and object detection. However, deep learning is often criticised for its lack of transparency and general inability to rationalize its predictions. The issue of poor model interpretability becomes critical in medical applications, as a model that is not understood and trusted by physicians is unlikely to be used in daily clinical practice. In this work, we develop a novel multi-task deep learning framework for simultaneous histopathology image classification and retrieval, leveraging on the classic concept of k-nearest neighbors to improve model interpretability. For a test image, we retrieve the most similar images from our training databases. These retrieved nearest neighbours can be used to classify the test image with a confidence score, and provide a human-interpretable explanation of our classification. Our original framework can be built on top of any existing classification network (and therefore benefit from pretrained models), by (i) adding a triplet loss function with a novel triplet sampling strategy to compare distances between samples and (ii) a Cauchy hashing loss function to accelerate neighbour searching. We evaluate our method on colorectal cancer histology slides, and show that the confidence estimates are strongly correlated with model performance. The explanations provided by nearest neighbors are intuitive and useful for expert evaluation by giving insights into understanding possible model failures, and can support clinical decision making by comparing archived images and patient records with the actual case.

pathology↗