Browse Articles

QnAs with Daniel Herschlag

Citrulline drives age-related lipid deposition for healthspan

Mapping the dialogue: Decoding alveolar stem–niche interactions

While cellular atlases have revealed remarkable phenotypic diversity, how cells navigate this landscape to influence tissue behavior remains poorly understood. We present an alveolosphere screening platform for investigating interactions between lung stem cells and their fibroblast niche. We assessed the role of 201 candidate genes in stem cells via imaging, then used chimeric RNAseq analysis for a transcriptome-wide understanding of cell-autonomous effects on stem cells and non-cell-autonomous effects on the niche. This phenome-transcriptome map uncovered cellular states and pathways regulating proliferation, metabolism, and immune signaling. Notably, stem cells influenced scar-forming and immune programs in fibroblasts. This injury response was dependent on stem cell identity; loss of Nkx2.1 , encoding the transcription factor conferring lung epithelial identity, rewired stem cell–niche interactions and had a greater non-cell-autonomous effect than eliminating the cancer genes Trp53, Egfr, or Cdkn2b . Our study highlights how functional atlases complement the cellular diversity revealed by descriptive methods.

High ambient temperature activates a neural circuit for gut glucose uptake in male mice

Physically intelligent capsule robots with embodied memory and logic in the gastrointestinal tract

Miniaturized medical robots offer a promising solution for minimally invasive measurements and interventions in the gastrointestinal (GI) tract. Clinical assessment of GI disorders is commonly guided by threshold-based physiological indicators, including pressure, temperature, and pH, which motivate event-triggered strategies for personalized medicine. However, identifying homeostatic dysregulation and enabling in-situ therapy remains challenging, because ingestible robotic systems must tightly integrate sensing, decision-making, and actuation under severe constraints of size, power, and biosafety. Inspired by the autonomy of microorganisms that operate without neural processing, this work introduces physically intelligent capsule robots (PI Capbots) that enable homeostatic monitoring and targeted delivery within the GI tract, without relying on centralized electronic control. Through embodied stimuli-responsive memory and logic, PI Capbots effectively distill rich, detailed, and redundant physiological information into a small set of decoupled and event-triggered outputs suitable for operations in in vivo environments. In each PI Capbot, multistable metamaterials encode intraluminal pressure as mechanical memory, programmable hydrogels implement orthogonal sensing and logic operations, and helical fibers enable multimodal locomotion. Ex vivo and in vivo studies in large animal models demonstrate the efficacy, robustness, and reproducibility of PI Capbots, highlighting its potential for their translational medical applications.

Preclinical characterization and phase 1 clinical testing of targeting mitochondrial peroxiredoxin 3 in cancer

Dopamine-driven mitochondrial reverse electron transport in immune cells mediates gut–brain ROS signaling during sleep deprivation

Sleep deprivation (SD), together with inevitable stress inherent to conventional SD protocols, can induce oxidative stress and inflammation, thereby increasing the risk of premature death. However, the source and signaling pathways underlying reactive oxygen species (ROS) generation remain unclear. Here, we demonstrate that both mechanical and thermogenetic SD, along with possible stress induced by both protocols, lead to initial ROS accumulation in Drosophila gut subregions, including the proventriculus (PV) and PV-resident hemocytes, via upregulation of dopamine (DA) biosynthesis. Intriguingly, DA acts unconventionally by activating mitochondrial reverse electron transfer (RET), presumably through modifying interactions between the respiratory complex I proteins NDUFV1 and NDUFS3. RET-ROS elicits hemolymphatic IMD/Relish-mediated antibacterial defense. However, during chronic SD, downregulation of the Drosophila APOE/D ortholog Neural Lazarillo promotes the recruitment of hemocytes to the central brain and, together with this process, leads to widespread neuronal ROS accumulation in an Alzheimer’s disease (AD) fly model. Inhibiting RET or hemocytic DA levels extends the survival of animals under chronic SD. Our work reveals DA-driven RET-ROS in innate immune cells during SD, highlights the pivotal role of a gut-innate immune-brain crosstalk in mediating the effect of SD manipulation on aging and AD pathogenesis, and suggests ways to lessen the consequence of SD, a profound health issue in modern society.

Single-cell transcriptome reveals keratinocyte subclusters contributing to altered differentiation and inflammatory responses in atopic dermatitis

High-resolution structure of monomorphic Aβ _1-40 fibrils

Amyloid-β (Aβ) fibrils primarily composed of Aβ 1-40 and Aβ 1-42 form the core of senile plaques in Alzheimer’s disease. Aβ 1-40 fibrils may exhibit significant polymorphism influenced by sample preparation conditions, complicating atomic resolution structural characterization. To establish a reliable structural baseline, we developed a protocol for expressing and purifying recombinant Aβ 1-40 that forms monomorphic fibrils under physiological conditions (pH 7.4). We present a high-resolution structure of these unseeded, monomorphic Aβ 1-40 fibrils obtained using magic-angle spinning NMR spectroscopy (PDB ID 12GB). We obtained unambiguous chemical shift assignments for approximately 90% of the residues and measured over 500 distance and torsion angle restraints. The resolved structure, with a backbone RMSD of 0.63 ± 0.06 Å, shows two monomers per filament plane, with two distinct β-sheets (residues E11-E22 and K28-V39, respectively) running along the fibril axis with H-bonding between each plane, and the two strands linked by a flexible loop region. This structure reveals three continuous hydrophobic cores inside each filament which bury 24 hydrophobic side chains per filament plane: those of L17, F19, A21, V24, A30, I32, M35, V40 between the two β-strands within each monomer and I31, L34, V36, V39 between the two monomers. Small angle X-ray scattering reveals the size and geometry of the fibril cross-section, which is compatible with a two-filament arrangement with a total of 4 monomers per fibril plane.

Context-aware sequence-to-function model of human gene regulation

Abstract Sequence-to-function models have been very successful in predicting gene expression, chromatin accessibility, and epigenetic marks from DNA sequences alone. However, current state-of-the-art models have a fundamental limitation: they cannot extrapolate beyond the cell types and conditions included in their training dataset. Here, we introduce Corgi, a context-aware sequence-to-function model that overcomes this limitation by integrating DNA sequence and trans -regulator expression to predict chromatin accessibility, histone modifications, and gene expression coverage, even in held-out cell types. Trained on a diverse set of bulk and single-cell sequencing datasets, Corgi achieves top performance in joint cross-sequence and cross-cell-type epigenetic track prediction. Additionally, we present an advanced model version, Corgi+, which is state-of-the-art in imputation of epigenetic tracks using only RNA-seq data. We further show that Corgi learns key cell type-specific trans -regulators in a zero-shot manner, and it can predict genomic variant effects in held-out cell types.

Structural basis of complement anaphylatoxin receptor activation by an immunostimulant lead candidate

Activation of the complement cascade is a primary innate immune response mechanism to combat pathogenic infections. Complement anaphylatoxins (i.e., C3a and C5a) exert a robust inflammatory response via prototypical GPCRs (i.e., C3aR and C5aR1). Several peptides derived from anaphylatoxins have shown promise as immunostimulants from therapeutic standpoint by eliciting immune response without excessive inflammation. EP67, a C5a-derived decapeptide, is the most advanced candidate with preclinical indications in antiviral and antibacterial context. Still, the molecular mechanism and the precise receptor target of EP67 remain unclear. Here, we perform a comprehensive pharmacological profiling of EP67 on the human and mouse C3aR and C5aR1 and find that it preferentially activates human C3aR in transducer-coupling assays. Subsequently, we determined four cryo-EM structures of C3aR and C5aR1 in complex with EP67, which elucidate the molecular details of its interaction with, and activation of, these receptors. Interestingly, we observe that EP67 adopts a hook-like structure and binds in the orthosteric pocket of the receptors, analogous to that of the carboxyl terminus of C3a and C5a. We employ site-directed mutagenesis studies to validate the key interactions of EP67 with these receptors and corroborate the structural observations including the engagement of a critical activation switch. Finally, we observe that EP67 induces distinct conformations of the TM7–Helix8 interface for C3aR and C5aR1, which provides a plausible explanation for its ability to preferentially activate C3aR. In summary, our study elucidates molecular insights into the interaction of EP67 with the complement anaphylatoxin receptors, and it should facilitate further optimization for therapeutic applications.

Adversarial dynamical systems characterize when data-driven learning succeeds or fails

Abstract Many systems resist analytical modeling, making data-driven inference of dynamics important. Yet data-driven methods can fail to converge or generalize, leaving open a central question: When can system behavior be learned reliably from data, and when is such learning impossible? We answer this question using adversarial dynamical systems to identify the boundary between accessible and inaccessible regimes. In Koopman operator learning, a leading framework for representing nonlinear dynamics through linear spectral objects, we design optimal data-driven spectral algorithms with convergence and certification guarantees under conditions arising broadly in physical systems. This yields a convergence theory for Koopman-operator approximations and resolves a longstanding open problem in Koopman spectral analysis. Conversely, by constructing adversarial systems, we prove matching impossibility results: without these conditions, no single-sequence limiting procedure can guarantee learning, regardless of data quality. These results sharply characterize when data-driven spectral learning can succeed and when it must fail. We validate the framework on oscillators, chaotic fluid flows and Arctic sea ice concentration forecasting. In the latter, we uncover hidden modes of Arctic sea ice decline, deliver long-range forecasts with geographic error bounds, and outperform state-of-the-art dynamical and deep learning models at substantially lower computational cost, enabling real-time deployment on standard CPUs.

Diffusion-induced instabilities promote cooperation in eco-evolutionary networks

Understanding how cooperation persists despite the advantage of selfish behavior remains a central challenge in evolutionary dynamics. Classical models of public goods dilemmas predict dominance of defectors, yet natural and social systems often sustain cooperation. We study an eco-evolutionary public goods game on complex networks where cooperators and defectors diffuse at different rates. When the isolated system is in a defector-dominated coexistence regime, faster dispersal of defectors than cooperators leads to a symmetry-breaking transition that produces localized clusters of cooperators. In heterogeneous networks, nodes with higher connectivity become significantly more likely to exhibit cooperative dominance. A degree-based mean-field reduction supports this result by showing that network connectivity controls an effective coupling strength proportional to node degree, thereby producing a bifurcation that separates defector-dominated and cooperative states. We also address why not all hubs become cooperative by means of a multistability analysis. These results reveal how asymmetric mobility and heterogeneous connectivity jointly promote cooperation in structured populations.

Reversal of protein chemical aging by enzymatic deglycation

Circadian screening of neutrophils identifies therapeutic targets in multiple sclerosis

Circadian rhythms are intrinsic time-keeping mechanisms that play a critical role in tuning immunity. Here, we investigated the impact of circadian rhythms on the pathogenesis of experimental autoimmune encephalomyelitis (EAE), a mouse model for multiple sclerosis (MS). We demonstrate that circulating neutrophils in blood significantly increase early in EAE, prior to symptoms onset. Importantly, we found that these cells infiltrate the central nervous system (CNS) in a time-of-day (ToD)-dependent manner, with increased infiltration at the onset of the behavioral active phase of the mice (evening). Transcriptomic analysis of CNS-infiltrating neutrophils revealed distinct ToD-dependent gene expression profiles, which identified Formyl peptide receptor 2 (FPR2) as a potential therapeutic candidate, since pharmacological inhibition of FPR2 led to reduced EAE disease severity. Furthermore, combinatorial treatment with a drug that targets VLA-4 (used in clinical practice under the trade name Natalizumab to treat MS) led to additive effects, substantially reducing EAE symptoms. Together, these findings highlight the importance of circadian immune cell dynamics during EAE development and provide a characterization of the circadian immune landscape in an animal model of MS, identifying potential targets for MS therapies.

A clinically deployed dual-compartment biochemical monitoring platform for human liver perfusion

Why human societies adopt rigid moral rules: The efficiency–robustness trade-off

Humans are capable of remarkably flexible moral judgment. Yet societies rely on rigid rules—obligations and prohibitions that apply categorically, even when case-by-case reasoning could yield better outcomes. Why would a species capable of flexibility bind itself to rigid rules? We propose that rigid rules arise as social technologies for managing ambiguity about noncooperation. People often have legitimate reasons for failing to cooperate, yet those reasons are typically opaque to observers, allowing opportunists to disguise selfishness as justified hardship. We formalize this idea with a game-theoretic model. Two cooperative equilibria emerge: a flexible norm that accommodates legitimate excuses but is vulnerable to exploitation, and a rigid norm that closes this loophole by mandating cooperation even when inefficient. Comparing these equilibria reveals an efficiency–robustness trade-off: Flexibility maximizes welfare when trust is secure, whereas rigidity preserves cooperation when trust is fragile. This explains why rigid rules prevail in interactions with strangers, formal institutions, or in tight societies, while flexibility is more common when trust is secure.

Nanowire-like C2H2 assembly in a flow-channel crystal boosts C2H2/CO2 separation at 348 K

Divergent population trajectories despite similar response to temperature in a widespread aerial insectivore

Climate change is shifting when animals breed [C. Parmesan, G. Yohe, Nature 421 , 37–42 (2003) and S. J. Thackeray et al. , Nature 535 , 241–245 (2016)], but it is not clear why some populations keep pace with warming while others fall behind [L. D. Bailey et al. , Nat. Commun. 13 , 2112 (2022) and J. M. Samplonius et al. , Glob. Change Biol. 24 , 3780–3790 (2018)]. Differences could arise from variation in sensitivity to temperature [L. D. Bailey et al. , Nat. Commun. 13 , 2112 (2022)] or constraints on the ability to respond to temperature. Without knowing whether populations differ in sensitivity—or in their ability to act on that sensitivity—we cannot identify which are most at risk. Using 1,555 population-years from 123 populations of tree swallows ( Tachycineta bicolor ), we show that populations have similar sensitivity to local temperature, advancing breeding by about one day per degree of warming. However, northern populations face tighter time constraints and greater exposure to recent warming. Northern populations have advanced laying dates the most, but still experience stronger selection for earlier breeding, especially in warm years; they have also declined most in breeding abundance. These findings suggest that vulnerability to climate change can arise not just from different sensitivity to warming, but from when and where populations can respond effectively. By disentangling sensitivity from timing constraints, our results are consistent with a general mechanism by which even uniformly responsive species can show uneven impacts of climate change across their ranges.

Two-coordinate Cu(I) sites in ZIF-derived porous carbon enable high CO working capacity for CO/N₂ separation