Biology subjects

Dislich, B.

Publications and source records attributed to Dislich, B..

4 recordsLinked to original sources

bioRxiv2026

Aurora kinase A is a synthetic lethal target in FANCA-deficient cancers

Loss-of-function genomic alterations in FANCA occur across multiple cancer types, yet no molecularly tailored therapies have successfully exploited this potential vulnerability. Using complementary unbiased approaches, including a genome-wide CRISPR/Cas9 loss-of-function screen and a high-throughput drug screen in isogenic cancer cell-based models, we identified Aurora kinase A (AURKA) as a reproducible synthetic lethal target of FANCA-deficient cancers. Inhibition of AURKA induced chromosomal instability, micronucleation, and early G2/M arrest selectively in FANCA-deficient cells, consistent with an increased reliance on mitotic checkpoint control. Mechanistically, FANCA deficiency is associated with an elevated AURKA expression at both the transcriptomic and protein levels, and with an upregulation of mitotic spindle and G2/M checkpoint gene signatures. Analysis of large-scale cancer genomics datasets, including over 650,000 clinically sequenced tumors, confirms that FANCA is the most frequently altered Fanconi anemia pathway gene across cancers, and that Fanconi anemia-defective tumors exhibit an increased tumor mutational burden and genomic instability. Collectively, our findings point to AURKA inhibition as a promising precision treatment strategy in FANCA-deficient cancers and provide a rationale to further explore this strategy in the clinic.

cancer biology↗

bioRxiv2022

Adversarial attacks and adversarial robustness in computational pathology

Artificial Intelligence (AI) can support diagnostic workflows in oncology by aiding diagnosis and providing biomarkers. AI applications are therefore expected to evolve from academic prototypes to commercial products in the coming years. However, AI applications are vulnerable to adversarial attacks, such as malicious interference with test data aiming to cause misclassifications. Therefore, it is essential for the use of AI-based diagnostic devices to secure them against such attacks before widespread use. Unfortunately, no resistant systems exist in computational pathology so far. To address this problem, we investigate the susceptibility of convolutional neural networks (CNNs) to multiple types of white- and black-box attacks. We demonstrate that both attacks can easily confuse CNNs in clinically relevant pathology tasks and impair classification performance. Classical adversarially robust training and dual batch normalization (DBN) are possible mitigation strategies but require precise knowledge of the type of attack used in the inference. We demonstrate that vision transformers (ViTs) perform equally well compared to CNNs at baseline and are orders of magnitude more robust to different types of white-box and black-box attacks. At a mechanistic level, we show that this is associated with a more robust latent representation of clinically relevant categories in ViTs compared to CNNs. Our results are in line with previous theoretical studies. We show that ViTs are robust learners in computational pathology. This implies that large-scale rollout of AI models in computational pathology should rely on ViTs rather than CNN-based classifiers to provide inherent protection against adversaries.

bioinformatics↗

bioRxiv2021

Benchmarking artificial intelligence methods for end-to-end computational pathology

Artificial intelligence (AI) can extract subtle visual information from digitized histopathology slides and yield scientific insight on genotype-phenotype interactions as well as clinically actionable recommendations. Classical weakly supervised pipelines use an end-to-end approach with residual neural networks (ResNets), modern convolutional neural networks such as EfficientNet, or non-convolutional architectures such as vision transformers (ViT). In addition, multiple-instance learning (MIL) and clustering-constrained attention MIL (CLAM) are being used for pathology image analysis. However, it is unclear how these different approaches perform relative to each other. Here, we implement and systematically compare all five methods in six clinically relevant end-to-end prediction tasks using data from N=4848 patients with rigorous external validation. We show that histological tumor subtyping of renal cell carcinoma is an easy task which approaches successfully solved with an area under the receiver operating curve (AUROC) of above 0.9 without any significant differences between approaches. In contrast, we report significant performance differences for mutation prediction in colorectal, gastric and bladder cancer. Weakly supervised ResNet-and ViT-based workflows significantly outperformed other methods, in particular MIL and CLAM for mutation prediction. As a reason for this higher performance we identify the ability of ResNet and ViT to assign high prediction scores to highly informative image regions with plausible histopathological image features. We make all source codes publicly available at https://github.com/KatherLab/HIA, allowing easy application of all methods on any end-to-end problem in computational pathology.

bioinformatics↗

bioRxiv2020

Furin controls β cell function via mTORC1 signaling

Furin is a proprotein convertase (PC) responsible for proteolytic activation of a wide array of precursor proteins within the secretory pathway. It maps to the PRC1 locus, a type 2 diabetes susceptibility locus, yet its specific role in pancreatic {beta} cells is largely unknown. The aim of this study was to determine the role of furin in glucose homeostasis. We show that furin is highly expressed in human islets, while PCs that potentially could provide redundancy are expressed at considerably lower levels. {beta} cell-specific furin knockout ({beta}furKO) mice are glucose intolerant, due to smaller islets with lower insulin content and abnormal dense core secretory granule morphology. RNA expression analysis and differential proteomics on {beta}furKO islets revealed activation of Activating Transcription Factor 4 (ATF4), which was mediated by mammalian target of rapamycin C1 (mTORC1). {beta}furKO cells show impaired cleavage of the essential V-ATPase subunit Ac45, and by blocking this pump in {beta} cells the mTORC1 pathway is activated. Furthermore, {beta}furKO cells show lack of insulin receptor cleavage and impaired response to insulin. Taken together, these results suggest a model of mTORC1-ATF4 hyperactivation in {beta} cells lacking furin, which causes {beta} cell dysfunction.

physiology↗