Biology subjects

Wendt, M.

Publications and source records attributed to Wendt, M..

10 recordsLinked to original sources

bioRxiv2026

SNAC-DB: An ML-Ready Database for Antibody and NANOBODY(R) VHH-Antigen Complexes with Expanded Structural Diversity and Real-World Benchmarking

Predicting antibody and NANOBODY(R) VHH-antigen complexes remains a critical challenge for state-of-the-art structure prediction models, limiting their impact in therapeutic discovery pipelines. We introduce SNAC-DB, an ML-ready database and curation pipeline enriched with structural biology expertise, designed to accelerate model accuracy and generalization by providing 31-37% expanded structural diversity over existing resources like SAbDab through comprehensive re-curation that extracts maximum value from available experimental structures. SNAC-DB expands coverage by capturing often-overlooked complexes and accurately identifying complete multi-chain epitopes through improved biological-assembly-based logic. Built for ML practitioners, SNAC-DB provides standardized formats with multi-threshold structure-based clustering to enable principled sample weighting during training. Using a rigorous benchmark of public PDB entries deposited post-May 2024 plus confidential therapeutic structures, we evaluate seven leading models (Protenix-v1, OpenFold-3p2, RosettaFold-3, Boltz-2, Boltz-1x, Chai-1, and AlphaFold2.3-multimer) with evaluation methodology tailored to antibody/NAN-OBODY(R) VHH-antigen complexes to ensure correct handling of multi-chain epitopes, revealing systematic performance gaps: success rates rarely exceed 25%, confidence-based ranking fails to identify best predictions even when accurate structures exist in ensembles, and all models consistently struggle with therapeutically relevant NANOBODY(R) VHHs. Systematic evaluation of sampling strategies demonstrates that while generating 1000 samples per target substantially increases the likelihood of producing accurate structures (oracle selection improves from 11.9% to 50.5%), confidence-based ranking remains nearly flat (between 10.9% and 14.9%), revealing that improved ranking mechanisms represent a more tractable path to performance gains. Finally, fine-tuning GeoDock on SNAC-DB yields higher success rates than training on SAbDab (11.0% vs. 7.1% for antibodies; 7.0% vs. 4.0% for NANOBODY(R) VHHs), suggesting that SNAC-DBs expanded structural diversity translates to improved model generalization. Significance StatementComputational antibody/NANOBODY(R) VHH design shows promise but remains unreliable for therapeutic development. SNAC-DB provides 31-37% expanded structural diversity through comprehensive data curation, immediately accelerating model development. Benchmarking seven leading AI models reveals accuracy rarely exceeds 25% on therapeutic targets, with confidence-based ranking failing to identify correct structures even when they exist in model outputs. Training on SNAC-DB increases prediction accuracy, validating that high-quality, diverse training data is critical for advancing computational methods toward clinical impact.

bioinformatics↗

bioRxiv2026

Loss of Mast cells and histaminergic signaling link diet to platelet-mediated NETosis and mammary cancer recurrence

Breast cancer recurrence remains a clinical challenge. The period after the treatment of the primary tumor while cancer cells that evaded initial treatment lay dormant, provides a unique window of opportunity for interventions to prevent recurrence. Specific modifiable factors such as consumption of high fat diets or elevated circulating cholesterol are associated with decreased time to recurrence. Mechanistically, oxidized cholesterol and lipid species have been implicated in the regulation of the tumor microenvironment. This suggests that consumption of food prepared under oxidizing conditions such as pan-frying, may be an underappreciated risk. Using murine models of mammary cancer dormancy, we found that a diet enriched with fat from fried, cured bacon (cfBF) decreased dormancy latency times. Resulting lesions had fewer mast cells (MCs). Loss of MCs alone resulted in reemergence from dormancy. Elevated expression of a MC gene signature in breast tumors was associated with improved progression free and overall survival, highlighting the human relevance of these findings. MCs are a major source of tissue histamine, and lesions from mice fed cfBF had decreased concentrations. Importantly, antagonists of the histamine receptor 2 (H2R) sparked reemergence from dormancy. H2R antagonists are over-the-counter drugs are taken to alleviate gastroesophageal reflux disease. Chronic treatment of mice with H2R-antagonists sensitized platelets towards activation and crosstalk with neutrophils, and subsequent formation of neutrophil extracellular traps (NETs). The loss of platelet or NETosis activity mitigated the H2R-antagonist stimulated reemergence from dormancy. Therefore, we establish a novel metastatic axis which links diet to recurrence via MCs, histaminergic signaling and NETosis: Diet -- MC -- H2R -- (decreased) Platelet Activity -- (decreased) Neutrophil-NETosis -- (decreased) Reemergence from Dormancy. Our data reveal several potential intervention strategies: lifestyle, MC stabilization, histaminergic signaling, and neutrophil and platelet activity.

cancer biology↗

bioRxiv2024

Unveiling inverted D genes and D-D fusions in human antibody repertoires unlocks novel antibody diversity

Antibodies, fundamental to immune defense, derive their diversity primarily from the intricate rearrangement of variable (V), diversity (D), and joining (J) gene segments. Traditionally, D genes in the forward (5-3) direction contribute to this diversity by rearranging with V and J segments. However, the existence and significance of inverted D genes (InvDs), which are D genes oriented in the inverted (3-5) direction, were previously obscured by limitations in data and detection methods. Here, we carried out a comprehensive analysis of a large-scale public next-generation sequencing (NGS) dataset encompassing antibody repertoires from 13 healthy donors using a novel immunoinformatics workflow. Our analysis, for the first time, uncovers the existence of all 25 unique InvDs across all three reading frames within human antibody repertoires, including both naive and memory B cells. This finding challenges previous assumptions, revealing the extensive presence of InvDs and identifying a broad spectrum of D-D fusions, especially those involving InvDs. Notably, InvDs enrich for unique amino acids such as histidine, proline, and lysine, not commonly found in forward D genes, and exhibit reduced use of certain negatively charged and bulky amino acids, including aspartate, tryptophan, and methionine. The unique amino acid profile of InvDs discloses new diversity and functionality in the human antibody repertoire, evidenced by over two dozen documented antibodies featuring InvDs, targeting a wide array of antigens. By opening exciting avenues for immunogenetics research, including new chromatin compaction models, innovative antibody libraries, and advancements in antibody engineering, these findings hold promise for the development of novel therapeutics and vaccines.

immunology↗

bioRxiv2023

Learning the Language of Antibody Hypervariability

Protein language models (PLMs) based on machine learning have demon-strated impressive success in predicting protein structure and function. However, general-purpose ("foundational") PLMs have limited performance in predicting antibodies due to the latters hypervariable regions, which do not conform to the evolutionary conservation principles that such models rely on. In this study, we propose a new transfer learning framework called AbMAP, which fine-tunes foundational models for antibody-sequence inputs by supervising on antibody structure and binding specificity examples. Our feature representations accurately predict an antibodys 3D structure, mutational effects on antigen binding, and paratope identification. AbMAPs scalability paves the way for large-scale analyses of human antibody repertoires. AbMAP representations of immune repertoires reveal a remarkable overlap across individuals, overcoming the limitations of sequence analyses. Our findings provide compelling evidence for the hypothesis that antibody repertoires of individuals tend to converge towards comparable structural and functional coverage. We validate AbMAP for antibody optimization, applying it to optimize a set of antibodies that bind to a SARS-CoV-2 peptide and obtaining 82% hit-rate and upto 22-fold increase in binding affinity. We anticipate AbMAP will accelerate the efficient design and modeling of antibodies and expedite the discovery of antibody-based therapeutics. Availability:https://github.com/rs239/ablm

bioinformatics↗

bioRxiv2021

Development of an integrated structural biology platform specialized for sub-100 kDa protein complexes to support biologics discovery and rational engineering

Developing a biologic medicine requires successful decision making at each step of selection, optimization, and/or combination of the right candidates at early research stages. Knowing the structural information and binding pattern between drug target and discovery candidates greatly increases the possibility of success. With the cryo-EM resolution revolution and rapid development of computational software, we have evaluated and integrated different tools in structural biology and the computation field and established a highly cost-effective platform, which allows us to obtain fast and accurate structural information for biologics projects with a close to 100% success rate and as fast as weeks turn-around time. Here we report four case studies selected from over 40 different protein structures and share how we integrate cryo-EM structure determination, computational structure modeling, and molecular dynamics simulation. With proper decision making and strategic planning, the platform allows us to obtain quality results within days to weeks, including sub-100 kDa complexes which are usually considered as a challenge due to their small size. Our utilization of this differential approach and use of multiple software packages, allows to manage priorities and resources to achieve goals quickly and efficiently. We demonstrate how to effectively overcome particle orientation bias by altering complex composition. In several of our examples, we use glycan density to facilitate interpretation of low-resolution 3D reconstruction and epitope mapping. Protein information plays an important role in our cryo-EM projects, especially in cases where we see significant challenges in obtaining high-resolution 3D maps.

molecular biology↗

bioRxiv2021

Application of an integrated computational antibody engineering platform to design SARS-CoV-2 neutralizers

As the COVID-19 pandemic continues to spread, hundreds of new initiatives including studies on existing medicines are running to fight the disease. To deliver a potentially immediate and lasting treatment to current and emerging SARS-CoV-2 variants, new collaborations and ways of sharing are required to create as many paths forward as possible. Here we leverage our expertise in computational antibody engineering to rationally design/optimize three previously reported SARS-CoV neutralizing antibodies and share our proposal towards anti-SARS-CoV-2 biologics therapeutics. SARS-CoV neutralizing antibodies, m396, 80R, and CR-3022 were chosen as templates due to their diversified epitopes and confirmed neutralization potency against SARS. Structures of variable fragment (Fv) in complex with receptor binding domain (RBD) from SARS-CoV or SARS-CoV2 were subjected to our established in silico antibody engineering platform to improve their binding affinity to SARS-CoV2 and developability profiles. The selected top mutations were ensembled into a focused library for each antibody for further screening. In addition, we convert the selected binders with different epitopes into the trispecific format, aiming to increase potency and to prevent mutational escape. Lastly, to avoid antibody induced virus activation or enhancement, we applied NNAS and DQ mutations to the Fc region to eliminate effector functions and extend half-life.

bioengineering↗

bioRxiv2019

Transglutaminase-2 facilitates extracellular vesicle-mediated establishment of the metastatic niche

AbstractThe ability of breast cancer cells to interconvert between epithelial and mesenchymal states contributes to their metastatic potential. As opposed to cell autonomous effects, the impact of epithelial-mesenchymal plasticity (EMP) on primary and metastatic tumor microenvironments remains poorly characterized. Herein we utilize global gene expression analyses to characterize a metastatic model of EMP as compared to their non-metastatic counterparts. Using this approach we demonstrate that upregulation of the extracellular matrix crosslinking enzyme tissue transglutaminase-2 (TGM2) is part of novel gene signature that only emerges in metastatic cells that have undergone induction and reversion of epithelial-mesenchymal transition (EMT). Consistent with our model system patient survival is diminished when primary tumors demonstrate enhanced levels of TGM2 in conjunction with its substrate, fibronectin. Targeted depletion of TGM2 inhibits metastasis, while overexpression of TGM2 is sufficient to enhance this process. In addition to being present within cells, we demonstrate a robust increase in the amount of TGM2 and crosslinked fibronectin present within extracellular vesicle (EV) fractions derived from metastatic breast cancer cells. Confocal microscopy of these EVs suggests that FN becomes fibrillated on their surface via a TGM2 and Tesin1-dependent process. Upon in vivo administration, the ability of tumor-derived EVs to induce metastatic niche formation and enhance subsequent pulmonary tumor growth requires the presence and activity of TGM2. Finally, we develop a novel 3D model of the metastatic niche to demonstrate that education of pulmonary fibroblasts via pretreatment with tumor-derived EVs promotes subsequent growth of breast cancer cells in a TGM2-dependent fashion. Overall, our studies illustrate a novel mechanism through which EMP contributes to metastatic niche development and distant metastasis via tumor-derived EVs containing abberent levels of TGM2 and fibular FN.

cancer biology↗

bioRxiv2019

Fibroblast growth factor receptor facilitates recurrence of minimal residual disease following trastuzumab emtansine therapy

Trastuzumab-emtansine (T-DM1) is an antibody-drug conjugate (ADC) that efficiently delivers a potent microtubule inhibitor into HER2 overexpressing tumor cells. However, resistance to T-DM1 is emerging as a significant clinical problem. Continuous in vitro treatment of HER2-transformed mammary epithelial cells with T-DM1 did not elicit spontaneously resistant cells. However, induction of epithelial-mesenchymal transition (EMT) via pretreatment with TGF-{beta}1 facilitated acquisition of T-DM1 resistance. Flow cytometric analyses indicated that induction of EMT decreased trastuzumab binding, prior to overt loss of HER2 expression. Kinome analyses of T-DM1 resistant cells indicated increased phosphorylation of ErbB1, ErbB4, and fibroblast blast growth factor receptor 1 (FGFR1). T-DM1 resistant cells failed to respond to the ErbB kinase inhibitors lapatinib and afatinib, but they acquired sensitivity to FIIN4, a covalent FGFR kinase inhibitor. In vivo, T-DM1 treatment led to robust regression of HER2-transformed tumors, but minimal residual disease (MRD) was still detectable via bioluminescent imaging. Upon cessation of the ADC relapse occurred and secondary tumors were resistant to additional rounds of T-DM1, but this recurrent tumor growth could be inhibited by FIIN4. Expression of FGFR1 was upregulated in T-DM1 resistant tumors, and ectopic overexpression of FGFR1 was sufficient to enhance tumor growth, diminish trastuzumab binding, and promote recurrence following T-DM1-induced MRD. Finally, patient-derived xenografts from a HER2+ breast cancer patient who had progressed on trastuzumab failed to respond to T-DM1, but tumor growth was significantly inhibited by FIIN4. Overall, our studies strongly support therapeutic combination of TDM1 and FGFR targeted agents in HER2+ breast cancer.

cancer biology↗

bioRxiv2018

Biased signaling downstream of epidermal growth factor receptor regulates proliferative versus apoptotic response to ligand

Inhibition of EGFR signaling by small molecule kinase inhibitors and monocloncal antibodies has proven effective in the treatment of multiple cancers. In contrast, metastatic breast cancers (BC) derived from EGFR-expressing mammary tumors are inherently resistant to EGFR-targeted therapies. Mechanisms that contribute to this inherent resistance remain poorly defined. Here we show that in contrast to primary tumors, ligand-mediated activation of EGFR in metastatic BC is dominated by STAT1 signaling. This change in downstream signaling leads to apoptosis and growth inhibition in response to EGF in metastatic BC cells. Mechanistically, these changes in downstream signaling result from an increase in the internalized pool of EGFR in metastatic cells, increasing physical access to the nuclear pool of STAT1. Along these lines, an EGFR mutant that is defective in endocytosis is unable to elicit STAT1 phosphorylation and apoptosis. Additionally, inhibition of endosomal signaling using an EGFR inhibitor linked to a nuclear localization signal specifically prevents EGF-induced STAT1 phosphorylation and cell death, without affecting EGFR:ERK1/2 signaling. Pharmacologic blockade of ERK1/2 signaling through the use of the allosteric MEK1/2 inhibitor, trametinib, dramatically biases downstream EGFR signaling toward a STAT1 dominated event, resulting in enhanced EGF-induced apoptosis in metastatic BC cells. Importantly, combined administration of trametinib and EGF also facilitated an apoptotic switch in EGFR-transformed primary tumor cells, but not normal mammary epithelial cells. These studies reveal a fundamental distinction for EGFR function in metastatic BC. Furthermore, the data demonstrate that pharmacological biasing of EGFR signaling toward STAT1 activation is capable of revealing the apoptotic function of this critical pathway.

cancer biology↗

bioRxiv2018

Pyruvate carboxylase supports the pulmonary tropism of metastatic breast cancer

Background: Overcoming systemic dormancy and initiating secondary tumor grow under unique microenvironmental conditions is a major rate-limiting step in metastatic progression. Disseminated tumor cells encounter major changes in nutrient supplies and oxidative stresses compared to the primary tumor and must demonstrate significant metabolic plasticity to adapt to specific metastatic sites. Recent studies suggest that differential utilization of pyruvate sits as a critical node in determining the organotropism of metastatic breast cancer. Pyruvate carboxylase (PC) is key enzyme that converts pyruvate into oxaloacetate for utilization in gluconeogenesis and replenishment of the TCA cycle.\n\nMethods: Patient survival was analyzed with respect to gene copy number alterations and differential mRNA expression levels of PC. Expression of PC was analyzed in the MCF-10A, D2-HAN and the 4T1 breast cancer progression series under in vitro and in vivo growth conditions. PC expression was depleted via shRNAs and the impact on in vitro cell growth, mammary fat pad tumor growth, and pulmonary and non-pulmonary metastasis was assessed by bioluminescent imaging. Changes in glycolytic capacity, oxygen consumption and response to oxidative stress were quantified upon PC depletion.\n\nResults: Genomic copy number increases in PC were observed in 16-30% of metastatic breast cancer patients. High expression of PC mRNA was associated with decreased patient survival in the MCTI and METABRIC patient datasets. Enhanced expression of PC was not recapitulated in breast cancer progression models when analyzed under glucose-rich in vitro culture conditions. In contrast, PC expression was dramatically enhanced upon glucose deprivation and in vivo in pulmonary metastases. Depletion of PC led to a dramatic decrease in 4T1 pulmonary metastasis but did not affect orthotopic primary tumor growth. Tail vein inoculations confirmed the role of PC in facilitating pulmonary, but not extra-pulmonary tumor initiation. PC-depleted cells demonstrated a decrease in glycolytic capacity and oxygen consumption rates and an enhanced sensitivity to oxidative stress.\n\nConclusions: Our studies indicate that PC is specifically required for the growth of breast cancer that has disseminated to the lungs. Overall, these findings point to the potential of targeting PC for the treatment of pulmonary metastatic breast cancer.

cancer biology↗