About

David Weber is the Director of the Center for Biomolecular Therapeutics (CBT) within the Institute for Bioscience and Biotechnology Research (IBBR). His research involves managing scientific studies that investigate mechanisms involved in disease states and developing drugs to treat them. His laboratory focuses on developing small-molecule inhibitors using structure-based drug design methods, particularly targeting cancer, diabetes, and infectious diseases. A significant area of his work includes studying the structure, function, and inhibition of the S100 family of calcium-binding proteins, especially S100B, which is an important marker for malignant melanoma prognosis and contributes to the disease by eliminating the tumor suppressor p53. His team has developed small molecules to restore p53 activity, evaluated in melanoma mouse models, with the goal of creating safe, effective therapeutic options. Weber's work leverages advanced biomedical tools such as medicinal chemistry, structural biology, protein engineering, and biophysics, aiming to translate discovery into therapeutic development to improve human health.

Research topics

• Chemistry
• Medicine
• Biology
• History
• Cancer research

Selected publications

• Enterovirus-induced cleavage of Mitofusin 2 generates mitophagosomes for enveloped virion release

  Science Advances · 2026-04-22

  articleOpen access

  Enterovirus D68 (EV-D68) is a plus-strand RNA virus that primarily causes respiratory infections in infants but, in rare cases, has been associated with the pediatric paralytic disease acute flaccid myelitis. We previously demonstrated that EV-D68 induces nonselective autophagy for its benefit. Here, we demonstrate that the 3C protease of EV-D68 cleaves the mitochondrial fusion protein Mitofusin 2 near its C-terminal HR2 domain, inducing fragmentation of the mitochondrial network. This, in turn, triggers the formation of mitophagosomes, a hallmark of mitophagy, a selective form of autophagy that recycles mitochondria. Multiple hallmarks of mitophagy are observed during infection, including loss of mitochondrial membrane potential and Parkin translocation to the mitochondria, but mitochondrial degradation is blocked during infection. While autophagy plays multiple roles in enterovirus infection, depleting Mitofusin 2 or transiently overexpressing Mitofusin 2, particularly the cleavage-resistant mutant, specifically reduces EV-D68 release from cells without affecting intracellular titers. Our results show that enteroviruses induce mitophagosomes as vectors for nonlytic release of virions from cells.

  PublisherDOI

• Galeterone monotherapy in advanced pancreatic ductal adenocarcinoma: Results from a phase two trial.

  Journal of Clinical Oncology · 2025-01-27

  article

  744 Background: Advanced pancreatic ductal adenocarcinoma (PDAC) is a devastating disease with limited treatment options. Dysregulated MAPK and PI3K signaling in PDAC activates the eukaryotic translational initiation complex, promoting cell growth, chemoresistance and metastatic spread (PMID 25593033). Galeterone, a novel steroidal anti-androgen, downregulates critical mediators of this complex including Mnk1/2 and phosphorylated eIF4E, thereby blocking epithelial-to-mesenchymal transition and NF-kB activity. In-vitro, galeterone exhibits anti-tumor activity both alone and in combination with gemcitabine and slows tumor growth in a MiaPaCa-2 xenograft murine model (PMID 28881737). We present results of the monotherapy arm from an ongoing phase two trial evaluating galeterone +/- gemcitabine for patients with advanced PDAC. Methods: Patients with locally advanced or metastatic PDAC, an ECOG performance status of 0-2 and adequate organ function, who progressed on two prior lines of systemic therapy were eligible. Galeterone was administered orally at 2550mg daily and gemcitabine intravenously at 1000mg/m 2 weekly for three weeks on 28-day cycle. The primary endpoint was radiographic response rate per RECIST v1.1 and secondary endpoints included progression-free survival (PFS), overall survival (OS) and safety. We used a Simon’s optimal two-stage design with a null hypothesis of p 0 =0.05 and one-sided alternative of p 1 =0.20 with an early stopping rule for futility in each arm. Due to poor accrual and no signal of activity, the galeterone alone arm closed after enrollment of four patients. Results: Four patients were evaluable for the primary endpoint. Ages ranged from 43 to 68. Two patients were male and two were female. Two patients were white and two were black; one patient was Hispanic/Latino. All patients had ECOG PS of 1. No patients had a radiographic response; the best response was progressive disease (PD). PFS from cycle 1 day 1 ranged from 0.9 – 1.7 months and overall survival between 1.8 – 3.6 months. The only treatment-related adverse events (TRAEs) were grade 1 dizziness and grade 3 fatigue in one patient each. There were no treatment-related serious adverse events. No patients required dose modification or discontinued due to TRAEs. Conclusions: Galeterone showed no clinical activity, but with a favorable safety profile, as a single-agent in heavily pre-treated patients with advanced PDAC. Activation of the Mnk-eIF4E axis mediates chemoresistance in PDAC and galeterone may prove more efficacious with gemcitabine. This combination arm is currently open to enrollment. Clinical trial information: NCT04098081 . Clinical outcomes with galeterone monotherapy in PDAC. Pt # Age Prior lines of Therapy Time on Treatment (days) Best Response PFS (months) OS(months) 1 68 4 46 PD 1.5 3.6 2 59 2 54 PD 1.7 3.8 3 55 2 47 PD 1.4 2 4 43 2 35 PD 0.9 1.8

  PublisherDOI

• Explainable and Class-Revealing Signal Feature Extraction via Scattering Transform and Constrained Zeroth-Order Optimization

  ArXiv.org · 2025-02-08

  preprintOpen accessSenior author

  We propose a new method to extract discriminant and explainable features from a particular machine learning model, i.e., a combination of the scattering transform and the multiclass logistic regression. Although this model is well-known for its ability to learn various signal classes with high classification rate, it remains elusive to understand why it can generate such successful classification, mainly due to the nonlinearity of the scattering transform. In order to uncover the meaning of the scattering transform coefficients selected by the multiclass logistic regression (with the Lasso penalty), we adopt zeroth-order optimization algorithms to search an input pattern that maximizes the class probability of a class of interest given the learned model. In order to do so, it turns out that imposing sparsity and smoothness of input patterns is important. We demonstrate the effectiveness of our proposed method using a couple of synthetic time-series classification problems.

  PublisherOA PDFDOI

• 1HN, 13C, and 15N resonance assignments of Ca2+-calmodulin bound to a BP2 peptide derived from the sheep STRA6 vitamin A transporter (residues 625–656)

  Biomolecular NMR Assignments · 2025-12-19

  articleOpen accessSenior author

  Vitamin A is essential for vision and many other biological processes required for human health and survival. Extracellular retinol binding protein (RBP) delivers vitamin A into the cell upon binding to the vitamin A transporter, STRA6. However, when retinol free RBP binds to STRA6, it induces vitamin A transport out of the cell. The bi-directionality of vitamin A transport is thought to be regulated further by an intracellular protein-protein interaction (PPI) between STRA6 and the EF-hand Ca2+-binding protein, calmodulin (CaM). Insights regarding how CaM regulates vitamin A transport were originally provided at atomic resolution by a cryoEM structure of the zebrafish STRA6-CaM complex. This cryoEM structure, together with NMR studies, confirmed that three STRA6 helices (i.e., BP0, BP1, and BP2) comprised the CaM-STRA6 binding interface, with BP2 providing the major set of interactions. NMR and other biophysical methods demonstrated that zebrafish BP2 peptide (zfBP2) binding to CaM involved a Ca2+-dependent type 2 binding and functional folding mechanism of action, which could influence structural, dynamic, and allosteric functions of STRA6. To expand our understanding of vitamin A transport to a mammalian STRA6 transporter, the backbone and sidechain 1HN, 13C, and 15N resonances were assigned here for CaCaM (148 residues) when bound to a sheep BP2 peptide (32 residues) (shBP2). Interestingly, the NMR data showed CaCaM resonances were affected differently upon binding shBP2 versus zfBP2. Such differences may be useful for distinguishing important features regarding CaCaM complexes with mammalian versus zebrafish STRA6.

  PublisherOA PDFDOI

• Hierarchical AF2RAVE for Multiconformation Virtual Screening Targeting S100 Ca²⁺-Binding Proteins

  Journal of Chemical Theory and Computation · 2025-09-04 · 2 citations

  article

  Protein function is driven by transitions between metastable conformations, many of which are not conserved across homologues, offering opportunities for selective drug design. Accurately modeling both backbone and side chain metastability, and generating structures suitable for rigid docking in high-throughput virtual screening, is thus desirable yet challenging. Here, we present a hierarchical AF2RAVE pipeline that integrates AlphaFold2 with machine learning-based enhanced sampling to systematically explore the free energy landscape and metastability of protein systems, particularly at both backbone and side chain levels. Applied to the calcium-binding S100 protein family, this approach enables the generation of diverse holo-like conformations, starting from sequence. Retrospective docking and enrichment testing with a new Ca2+-S100B inhibitor data set demonstrates that AF2RAVE-generated structures outperform standard AlphaFold2 and even outperform experimentally resolved X-ray structures in enrichment testing. Our results highlight the potential of AF2RAVE for high-throughput virtual screening and selective inhibitor discovery, particularly for challenging targets such as the Ca2+-S100 family.

  PublisherDOI

• High Throughput Ligand Dissociation Kinetics Predictions using Site-Identification by Ligand Competitive Saturation

  ChemRxiv · 2025-04-02

  preprintOpen access

  The dissociation or off rate, koff, of a drug molecule has been shown to be more relevant to efficacy than affinity for selected systems motivating the development of predictive computational methodologies. These are largely based on enhanced-sampling molecular dynamics (MD) simulations that come at a high computational cost limiting their utility for drug design where a large number of ligands needs to be evaluated. To overcome this, presented is a combined physics- and machine learning (ML)-based approach that uses the physics-based site-identification by ligand competitive saturation (SILCS) method to enumerate potential ligand dissociation pathways and calculate ligand dissociation free energy profiles along those pathways. The calculated free energy profiles along with molecular properties are used as features to train ML models, including tree- and neural network approaches, to predict koff values. The protocol is developed and validated using 329 ligands for thirteen proteins showing robustness of the ML workflows built upon the SILCS physics-based free energy profiles. The resulting SILCS-Kinetics workflow offers a highly efficient method to study ligand dissociation kinetics, providing a powerful tool to facilitate drug design including the ability to generate quantitative estimates of atomic and functional groups contributions to ligand dissociation.

  PublisherDOI

• Characterizing presenteeism among healthcare personnel at an academic medical center across eras of the COVID-19 pandemic

  Infection Control and Hospital Epidemiology · 2025-05-16 · 1 citations

  articleSenior author

  OBJECTIVE: To assess the frequency of and motivations for acute respiratory illness (ARI) presenteeism in healthcare personnel (HCP) during two waves of COVID-19. DESIGN: Survey. SETTING: Large academic medical center, both ambulatory and acute care settings. PARTICIPANTS: All HCPs (n = 11,429) at the University of North Carolina Medical Center were eligible for two voluntary, electronic surveys: pre-Omicron (n = 591, recall period March 2020 - December 2021) and Omicron BA.1 (n = 385, recall period January - April 2022). METHODS: We compared self-reported ARI presenteeism (working despite feeling feverish plus cough and/or sore throat) and motivators across time and demographics. We also estimated effects of workplace perceptions and culture on ARI presenteeism with log-binomial regression, adjusting for age, gender, HCP role, and patient interaction. RESULTS: In the pre-Omicron and Omicron BA.1 eras, 24% and 34% of respondents respectively reported at least one instance of ARI presenteeism. In both eras, clinical frontline HCP were more likely to report ARI presenteeism than other roles, as were HCP primarily providing direct patient care vs not. Pre-Omicron motivators included disciplinary action and sick leave concerns, whereas workplace culture predominated during Omicron. Feeling professional obligation to attend work and observing colleague presenteeism increased ARI presenteeism in both eras. During Omicron, COVID-19 burnout, fatigue, and unclear call-out procedures increased ARI presenteeism. CONCLUSIONS: ARI presenteeism was common and had diverse motivations, including workplace culture, disciplinary action, and sick leave. Efforts to reduce presenteeism should address these factors and prioritize frontline clinical personnel with direct patient interaction.

  PublisherDOI

• Less Is More: The Influence of Pruning on the Explainability of CNNs

  IEEE Access · 2025-01-01

  articleOpen access

  Over the last century, deep learning models have become the state-of-the-art for solving complex computer vision problems. These modern computer vision models have millions of parameters, which presents two major challenges: (1) the increased computational requirements hamper the deployment in resource-constrained environments, such as mobile or IoT devices, and (2) explaining the complex decisions of such networks to humans is challenging. Network pruning is a technical approach to reduce the complexity of models, where less important parameters are removed. The work presented in this paper investigates whether this reduction in technical complexity also helps with perceived explainability. To do so, we conducted a pre-study and two human-grounded experiments, assessing the effects of different pruning ratios on explainability. Overall, we evaluate four different compression rates (i.e., 2, 4, 8, and 32) with 37 500 tasks on Mechanical Turk. Results indicate that lower compression rates have a positive influence on explainability, while higher compression rates show negative effects. Furthermore, we were able to identify sweet spots that increase both the perceived explainability and the model’s performance.

  PublisherDOI

• Hierarchical AF2RAVE for multi-conformation virtual screening targeting S100 Ca$^{2+}$-binding proteins

  ChemRxiv · 2025-06-02

  preprintOpen access

  Protein function is driven by transitions between metastable conformations, many of which are not conserved across homologs, offering opportunities for selective drug design. Accurately modeling both backbone and sidechain metastability and generating structures suitable for rigid docking in high-throughput virtual screening, is thus desirable yet challenging. Here, we present a hierarchical AF2RAVE pipeline that integrates AlphaFold2 with machine learning-based enhanced sampling to systematically explore the free energy landscape and metastability of protein systems, particularly at both backbone and sidechain levels. Applied to the calcium-binding S100 protein family, this approach enables the generation of diverse holo-like conformations, starting from sequence. Retrospective docking and enrichment testing with a new $Ca^{2+}$-S100B inhibitor dataset demonstrates that AF2RAVE-generated structures outperform standard AlphaFold2 and even outperform experimentally resolved X-ray structures in enrichment testing. Our results highlight the potential of AF2RAVE for high-throughput virtual screening and selective inhibitor discovery, particularly for challenging targets such as the $Ca^{2+}$-S100 family.

  PublisherOA PDFDOI

• Fomite workshop recommendations addressing the role of surfaces in virus transmission in the built environment

  mSphere · 2025-08-20

  reviewOpen access

  The emergence of SARS-CoV-2 has led to a need to assess the role of fomites in viral transmission within the built environment. Assessing the role of fomites is necessary for developing intervention strategies for controlling emerging pathogens. A fomite workshop with experts was convened in November 2024 by academia, several government agencies, and public health officials to evaluate existing data and discuss how to mitigate risks. Fomite transmission is influenced by the nature of the built environment, population density and proximity, environmental factors (humidity, heat, etc.), virus survival, surface type, engineering controls (ventilation, physical barriers, etc.), and human behaviors. Based on our current data, direct contact with a contaminated surface/fomite, even for respiratory viruses, presents a risk of viral exposure and transmission by both contact with the fomite and resuspension in the air. Even respiratory viruses can be resuspended from fomites following human and pet movement, activities (e.g., vacuuming, toilet flushing, etc.), or changes in ventilation/indoor airflow. After resuspension from surfaces, microbes can be potentially inhaled (contributing to droplet and/or aerosol exposure) and/or re-deposited from primary to secondary fomites. Development of standard methods (molecular, chemical/physical, and infectivity assays) for detecting the presence of viruses on fomites and human behavior modeling would help to determine the most effective infection prevention strategies.

  PublisherDOI

Recent grants

• NIH Grant S10RR023447

  NIH · $362k · 2008

• NIH Grant S10RR015741

  NIH · $500k · 2002

• NIH Grant S10RR029601

  NIH · $8.0M · 2012

• 800 MHz NMR Spectrometer

  NSF · $400k · 2001–2002

• NIH Grant S10RR031729

  NIH · $164k · 2012

Frequent coauthors

• Kristen M. Varney

  University of Maryland, Baltimore

  134 shared

• William A. Rutala

  University of North Carolina at Chapel Hill

  117 shared

• Paul T. Wilder

  University of Maryland, Baltimore

  116 shared

• Alexander D. MacKerell

  University of Maryland, Baltimore

  104 shared

• Raquel Godoy‐Ruiz

  University of Maryland, Baltimore

  90 shared

• Wenbo Yu

  Hangzhou Vocational and Technical College

  77 shared

• Edwin Pozharski

  67 shared

• Michele Vítolo

  University of Maryland, Baltimore

  54 shared

Education

• Postdoctoral Fellow, Biological Chemistry

  Johns Hopkins Medicine

  1992

• PhD, Chemistry

  University of North Carolina at Chapel Hill

  1988

• B.S., Chemistry

  Muhlenberg College

  1984