Biology subjects

Levine, H.

Publications and source records attributed to Levine, H..

At least 19 recordsLinked to original sources

bioRxiv2026

Regulatory stochasticity drives opposing phenotypic outcomes in cell-fate decision networks

Gene regulatory network models treat interaction parameters as fixed, although regulatory efficacy fluctuates. We asked how temporal fluctuations in interaction strength reshape phenotype occupancy in cell-fate decision GRN motifs. Across large parameter ensembles, anchored fluctuations largely preserved deterministic occupancies. Additive fluctuations increased occupancy of all-high co-expression states, particularly where high expression was accessible. In contrast, multiplicative fluctuations biased inhibitory interactions toward stronger repression and favored single-high states in a topology-dependent manner. Deterministic controls sampled from noise-induced parameter distributions did not fully reproduce these effects. A Boolean-limit analysis revealed an intrinsic upward bias: loss of repression increased expression regardless of regulator state, whereas stronger repression acted only when the regulator was present. Analyses of epithelial-mesenchymal plasticity and gonadal-fate networks showed increased occupancy of hybrid team-expression states under additive fluctuations. Thus, regulatory noise can reshape the developmental landscape in opposing directions, pushing cell-fate systems toward either progenitor-like or terminally differentiated states.

systems biology↗

bioRxiv2026

The Warburg effect as a metabolic niche construction strategy in the tumor microenvironment

The advantage conferred by the Warburg effect to cancer cells remains unresolved. While extensive research has examined aerobic glycolysis at the intracellular level, emphasizing its possible biosynthetic or ATP-production rate benefits, the fate of the energy-rich lactate exported into the tumor microenvironment (TME) has received comparatively little attention. Yet lactate is far from metabolically inert: it actively shapes the TME, functioning as a toxin to adaptive immune cells and simultaneously promoting growth of several non-cancerous cells in the tumor microenvironment which in turn promote tumor growth. Here, we develop a minimal consumer-resource model incorporating tumor, anti-tumor immune, and pro-tumor stromal populations interacting through shared nutrients, lactate, and metabolic byproducts. We find that lactate secretion imposes a growth-exclusion tradeoff: higher glycolytic allocation suppresses anti-tumor immune cells but simultaneously reduces tumor abundance due to less efficient metabolism. This tradeoff is resolved in the presence of pro-tumor stromal populations capable of metabolic cross-feeding, namely stromal cells that convert lactate into metabolites which feed back to support tumor growth, thereby offsetting the glycolytic cost. The cross-feeding benefit is second-order in population size, requiring a critical tumor mass to operate. Taken together, these results suggest that there may be an ecological advantage of the Warburg effect, contingent on microenvironmental context. Specifically, lactate secretion is beneficial primarily when appropriate stromal partners are present, functioning as a niche-construction strategy that reorganizes the tumor ecosystem in favor of tumor expansion.

ecology↗

bioRxiv2026

A biophysical framework for accurately identifying antigen single-amino acid escape variants and corresponding variant-specific compensatory TCR sequences

The impact of single amino acid substitution on T-cell receptor (TCR) recognition is central to understanding the molecular determinants of TCR specificity and degeneracy during viral mutational escape, cancer recognition, and autoimmunity. In this study, we developed a biophysics-informed computational approach integrating experimental alanine-scan mutagenesis data from the autoimmune-associated ALWGPDPAAA peptide bound to HLA-A*02:01 together with coarse-grained structural modeling. Our approach reconstructs the energetics and structural determinants underpinning the observed loss of recognition by the diabetogenic 1E6 TCR upon single-point mutations, specifically at the critical Pro5 and Asp6 residues. Leveraging the computational models ability to incorporate multiple structural templates into binding predictions, this approach quantitatively reproduces experimentally measured affinity disruptions. Additionally, we apply our approach to identify potential compensatory interactions capable of restoring binding affinity through alternative residue interactions. This integrative computational framework contributes a strategy for inferring TCR-peptide binding energetics at the single amino acid level, guiding the rational design of peptide-based immunotherapeutics, and predicting the functional impacts of clinically relevant peptide variants.

biophysics↗

bioRxiv2026

Early detection of sudden transitions in Notch signalling

Identifying sudden transitions during phenotypic decision-making of complex biological systems can be crucial for our ability to control a cellular state. Yet, prior determination of these sudden transitions or tipping points remains challenging, as biological systems often exhibit only subtle early changes, which are often masked by inherent noise or rapid transition dynamics. Using Notch signaling as a model, we systematically analyze dynamical transitions in Notch-Delta (ND), Notch-Delta-Jagged (NDJ), and Fringe-mediated NDJ systems for both one and two-cell contexts. In the one-cell ND system, critical slowing down (CSD)-based early warning signals (EWSs) reliably capture transitions between sender (S) and receiver (R) states and remain robust to variation in forcing rate. We further find that flickering is a precursor to transitions in one-cell NDJ system. In contrast, flickering does not occur in the two-cell Notch model due to the presence of a supercritical bifurcation. Our analysis also offers insight into how NICD (Notch Intracellular Domain)-driven and Fringe-mediated asymmetries, along with the strength of external signals, control the emergence of flickering. Overall, this study identifies sudden transitions in Notch signaling under demographic noise and can be extended to other noisy biological systems, with potential applications in drug development and targeted therapeutic interventions.

systems biology↗

bioRxiv2026

Ecological Dynamics of Pro-tumor and Anti-tumor Teams in the Tumor Microenvironment

Tumor growth occurs within a complex tumor microenvironment (TME) composed of many interacting cell types. The immune cell types in TME tend to organize into two functional communities: a pro-tumor team and an anti-tumor team, each internally cooperative but mutually antagonistic forming a two-team ecosystem. Quantitatively predicting the ecological outcomes of such interactions remains challenging due to cellular diversity and interaction variability, and the exact dynamical regimes accessible to such a two-team ecosystem remain unknown. Here, we model tumor-immune interactions as a structured ecosystem with two competing teams using a generalized Lotka-Volterra framework and analyze it using the cavity method. We derive phase diagrams that delineate when these two communities coexist, when one dominates, and how these outcomes depend on intra-team cooperation, cross-team inhibition, and ecological heterogeneity. Our work provides a foundation for understanding tumor-immune dynamics from a community ecology perspective.

ecology↗

bioRxiv2025

PROFET Predicts Continuous Gene Expression Dynamics from scRNA-seq Data to Elucidate Heterogeneity of Cancer Treatment Responses

Single-cell RNA sequencing profiles cellular heterogeneity but captures only static snapshots, limiting inference of gene expression dynamics. We developed PROFET (Particle-based Reconstruction Of generative Force-matched Expression Trajectories), a framework that reconstructs continuous, nonlinear single-cell trajectories from sparsely sampled scRNA-seq time series. PROFET combines a particle-based gradient-flow algorithm with simulation-free force matching to accurately infer cellular dynamics. Across mouse and human in vitro datasets and an in vivo axolotl regeneration dataset, PROFET achieved 2.6-12.5X lower prediction error than ten state-of-the-art trajectory inference methods. Applying PROFET to newly generated scRNA-seq data from a palbociclib-treated MCF7 cell line and three published breast cancer patient datasets, we reconstructed treatment-response trajectories and identified a resistant cell subpopulation exhibiting large phenotypic shifts and enrichment of the surface markers UNC5B, TLR3, PCDH19, PROCR, SLITRK6, and SEMA6B. PROFET provides a biologically grounded framework for reconstructing cell-state dynamics from static single-cell data across development, regeneration, and therapeutic response.

bioinformatics↗

bioRxiv2025

The interaction between dynamic ligand signaling andepigenetics in Notch-induced cancer metastasis

Metastatic melanoma presents a formidable challenge in oncology due to its high invasiveness and resistance to current treatments. Central to its ability to metastasize is the Notch signaling pathway, which, when activated through direct cell-cell interactions, propels cells into a metastatic state through mechanisms akin to the epithelial-mesenchymal transition (EMT). While the upregulation of miR-222 has been identified as a critical step in this metastatic progression, the mechanism through which this upregulation persists in the absence of active Notch signaling remains unclear. Here we introduce a dynamical system model that integrates miR-222 gene regulation with histone feedback mechanisms. Through computational analysis, we delineate the non-linear decision boundaries that govern melanoma cell fate transitions, taking into account the dynamics of Notch signaling and the role of epigenetic modifications. Our approach highlights the critical interplay between Notch signaling pathways and epigenetic regulation in dictating the fate of melanoma cells.

systems biology↗

bioRxiv2025

Phase-field Approach to Cellular Blebbing

Bulges in the plasma membrane of cells known as blebs can form spontaneously in a wide range of biological processes but what controls their shape and stability remains incompletely understood. To address this, we introduce a dual phase-field model with coupled order parameters representing the cell cortex and plasma membrane that can quantitatively model blebbing in three dimensions. Simulations and analysis of the model reveal that, depending on whether blebbing occurs by detachment of the plasma membrane or rupture of the actin cortex, blebs can form discontinuously through a saddle node bifurcation or continuously with increasing cortical tension. The model predictions are in good quantitative agreement with existing experimental data for laser-induced cortex rupture.

biophysics↗

bioRxiv2025

Emergent dynamics of cellular decision making in multi-node mutually repressive regulatory networks

Stem cell differentiation during development is governed by the dynamics of the underlying gene regulatory networks (GRNs). Mutually inhibiting nodes/collection of nodes encompass the GRNs that govern differentiation to two distinct fates. But the properties of GRNs that can allow differentiation into n-terminal phenotypes are not understood. In this study, we examine toggle-n networks, encompassing mutual inhibitions among multiple transcription factors, to derive generalized conclusions regarding the dynamics underlying differentiation into n-terminal phenotypes. We show that in steady-state distributions of gene expression, multiple cell state-specific transcription factors are co-expressed, indicating an obligatory multi-step process of multi-lineage differentiation. Furthermore, we show that cytokine signaling and specific asymmetry of regulatory links can lead to directed differentiation towards a particular cell state. Our findings provide valuable insights into the mechanistic aspects of directed differentiation of stem cells.

systems biology↗

bioRxiv2025

Mathematical Characterization of Drug-Induced Persistence in Cancer

We introduce a general phenomenological framework for understanding how phenotypic plasticity gives rise to drug persisters. These persisters, often quiescent but sometimes which again return to cycling, survive in the presence of targeted therapy and eventually can lead to mutants with true resistance. Our framework builds on recent experimental observations regarding variations between and among single-cell clones and the possible role of the drug itself in enhancing the survival strategy. Predictions of our approach include the existence of an optimum drug concentration as well as an optimum drug holiday schedule to minimize the persistence-based threat.

cancer biology↗

bioRxiv2024

A multi-level formalism to capture the characteristics of hybrid phenotypes in Epithelial-Mesenchymal Plasticity

Epithelial-mesenchymal plasticity (EMP) is a cell-fate switching program that enables cells to adopt a spectrum of phenotypes ranging from epithelial (E) to mesenchymal (M), including intermediate hybrid E/M states. Hybrid E/M phenotypes are conducive to cancer metastasis, as they are associated with metastatic initiation, cancer stemness, drug resistance, and collective migration. Boolean models of the gene regulatory networks (GRNs) underlying EMP have yielded valuable insights into the dynamics of E and M phenotypes. However, these models are limited in their ability to capture hybrid phenotypes effectively, as they restrict gene expression to binary states. In contrast, hybrid E/M cells often exhibit partial expression of epithelial and mesenchymal markers. To overcome this limitation, we modified a threshold-based Boolean formalism to incorporate intermediate gene expression levels. The resulting multilevel model reveals novel hybrid steady-states characterized by partial expression of both E and M genes, thereby expanding the phenotypic landscape beyond that represented by traditional Boolean approaches. Notably, these hybrid states exhibit lower frustration compared to their counterparts in classical Boolean models. Furthermore, by resolving dynamical degeneracy, we demonstrate that the hybrid states identified by the multilevel model are more stable. These findings suggest that introducing minimal additional complexity into Boolean models can uncover previously hidden qualitative features of phenotypic landscapes governed by GRNs. Significance StatementHybrid epithelial-mesenchymal phenotypes drive metastasis through collective migration, stemness, and drug resistance. However, traditional Boolean models of epithelial-mesenchymal plasticity (EMP) networks cannot capture the partial gene expression characteristic of these biologically crucial hybrid states. We developed a multilevel Boolean formalism that extends classical binary models to include intermediate expression levels. This approach successfully captures hybrid states with partial expression of both epithelial and mesenchymal genes, exhibiting lower frustration and greater stability than traditional Boolean hybrid phenotypes. While conventional Boolean hybrid states depend on modeling artifacts, multilevel hybrid states emerge independently of such uncertainties. Critically, plasticity experiments revealed that multilevel hybrid states show significantly higher transition probabilities to other hybrid states compared to traditional models, creating a stable "hybrid cloud" that maintains metastatic properties while allowing adaptive responses to environmental stress. This multilevel framework provides a biologically meaningful characterization of hybrid phenotypes essential for understanding metastasis.

systems biology↗

bioRxiv2024

Computational modeling of cancer cell metabolism along the catabolic-anabolic axes

Abnormal metabolism is a hallmark of cancer. Initially recognized through the observation of aerobic glycolysis in cancer nearly a century ago. Also, we now know that mitochondrial respiration is also used by cancer for progression and metastasis. However, it remains largely unclear the mechanisms by which cancer cells mix and match different metabolic modalities (oxidative/reductive) and leverage various metabolic ingredients (glucose, fatty acids, glutamine) to meet their bioenergetic and biosynthetic needs. Here, we formulate a phenotypic model for cancer metabolism by coupling master gene regulators (AMPK, HIF-1, Myc) with key metabolic substrates (glucose, fatty acid, and glutamine). The model predicts that cancer cells can acquire four metabolic phenotypes: a catabolic phenotype characterized by vigorous oxidative processes - O, an anabolic phenotype characterized by pronounced reductive activities - W, and two complementary hybrid metabolic states - one exhibiting both high catabolic and high anabolic activity - W/O, and the other relying mainly on glutamine oxidation - Q. Using this framework, we quantified gene and metabolic pathway activity respectively by developing scoring metrics based on gene expression. We validated the model-predicted gene-metabolic pathway association and the characterization of the four metabolic phenotypes by analyzing RNA-seq data of tumor samples from TCGA. Strikingly, carcinoma samples exhibiting hybrid metabolic phenotypes are often associated with the worst survival outcomes relative to other metabolic phenotypes. Our mathematical model and scoring metrics serve as a platform to quantify cancer metabolism and study how cancer cells adapt their metabolism upon perturbations, which ultimately could facilitate an effective treatment targeting cancer metabolic plasticity. Statement of significanceWe present a theoretical framework that integrates both catabolic and anabolic modes of cancer metabolism, considering the complex interplay between glucose, fatty acids, and glutamine, by coupling genetic regulation with metabolic pathways. Our work characterizes four main metabolic phenotypes in cancer - OXPHOS, glycolysis, hybrid, glutamine-dominant and demonstrates the critical role of Myc on glutamine metabolism in all four phenotypes. The characterization of metabolism can guide our evaluation of patient survival across cancer types.

systems biology↗

bioRxiv2024

Pulse Instabilities Can Shape Virus-Immune Co-evolution

Adaptive immune systems engage in an arms race with evolving viruses, trying to generate new responses to viral strains that continually move away from the set of genetically-varying strains that have already elicited a functional immune response. It has been argued that this dynamical process can lead to a propagating pulse of an ever-changing viral population and concomitant immune response. Here, we introduce a new stochastic model of viral-host co-evolution, taking into account finite-sized host populations and varying processes of immune "forgetting". Using both stochastic and determinstic calculations, we show that there is indeed a possible pulse solution, but for a large host population size and for finite memory capacity, the pulse becomes unstable to the generation of new infections in its wake. This instability leads to an extended endemic infection pattern, demonstrating that the population-level behavior of virus infections can exhibit a wider range of behavior than had been previously realized.

biophysics↗

bioRxiv2023

Regulation of chromatin transcription dynamics by DNA supercoiling

Transcription has a mechanical component, as the translocation of the transcription machinery or RNA polymerase (RNAP) on DNA or chromatin is dynamically coupled to the chromatin torsion. This posits chromatin mechanics as a possible regulator of eukaryotic transcription, however, the modes and mechanisms of this regulation are elusive. Here, we first take a statistical mechanics approach to model the torsional response of topology-constrained chromatin. Our model recapitulates the experimentally observed weaker torsional rigidity of chromatin compared to bare DNA, and proposes structural transitions of nucleosomes into chirally distinct states as the driver of the contrasting torsional mechanics. Coupling chromatin mechanics with RNAP translocation in stochastic simulations, we reveal a complex interplay of DNA supercoiling and nucleosome dynamics in governing RNAP velocity. Nucleosomes play a dual role in controlling the transcription dynamics. The steric barrier aspect of nucleosomes in the gene body counteracts transcription via hindering RNAP motion, whereas the chiral transitions facilitate RNAP motion via driving a low restoring torque upon twisting the DNA. While nucleosomes with low dissociation rates are typically transcriptionally repressive, highly dynamic nucleosomes offer less of a steric barrier and enhance the transcription elongation dynamics of weakly transcribed genes via buffering DNA twist. We use the model to predict transcription-dependent levels of DNA supercoiling in segments of the budding yeast genome that are in accord with available experimental data. The model unveils a paradigm of DNA supercoiling-mediated interaction between genes and makes testable predictions that will guide experimental design.

biophysics↗

bioRxiv2023

Reconstruction of single cell lineage trajectories and identification of diversity in fates during the epithelial-to-mesenchymal transition

Exploring the complexity of the epithelial-to-mesenchymal transition (EMT) unveils a diversity of potential cell fates; however, the exact timing and intricate mechanisms by which early cell states diverge into distinct EMT trajectories remain unclear. Studying these EMT trajectories through single cell RNA sequencing is challenging due to the necessity of sacrificing cells for each measurement. In this study, we employed optimal-transport (OT) analysis to reconstruct the past trajectories of different cell fates during TGF-beta-induced EMT in the MCF10A cell line. Our analysis revealed three distinct trajectories leading to low EMT, partial EMT, and high EMT states. Cells along partial EMT trajectory showed substantial variations in the EMT signature and exhibited pronounced stemness. Throughout this EMT trajectory, we observed a consistent downregulation of the EED and EZH2 genes. This finding was validated by recent inhibitor screens of EMT regulators and CRISPR screen studies. Moreover, we applied our analysis of early-phase differential gene expression to gene sets associated with stemness and proliferation, pinpointing ITGB4, LAMA3, and LAMB3 as genes differentially expressed in the initial stages of the partial versus high EMT trajectories. We also found that CENPF, CKS1B, and MKI67 showed significant upregulation in the high EMT trajectory. While the first group of genes aligns with findings from previous studies, our work uniquely pinpoints the precise timing of these upregulations. Finally, the latter group of genes represents newly identified regulators, shedding light on potential targets for modulating EMT trajectories. Significance StatementIn our study, we investigated cellular trajectories during EMT using a time-series scRNAseq dataset. OT analysis was used to infer cell-to-cell connections from scRNAseq data, allowing us to predict cell linkages and overcome limitations of sequencing such as the need to sacrifice cells for each measurement. This approach allowed us to identify diverse EMT responses under uniform treatment, a significant advancement over previous studies limited by the static nature of scRNAseq data. Our analysis identified a broad set of genes involved in the EMT process, uncovering novel insights such as the upregulation of cell cycle genes in cells predisposed to a high EMT state and the enhancement of cell adhesion marker genes in cells veering towards a partial EMT state. This work enriches our understanding of the dynamic processes of EMT, showcasing the varied cellular fates within the same experimental setup.

systems biology↗

bioRxiv2023

RACER-m Leverages Structural Features for Sparse T Cell Specificity Prediction

Reliable prediction of T cell specificity against antigenic signatures is a formidable task, complicated primarily by the immense diversity of T cell receptor and antigen sequence space and the resulting limited availability of training sets for inferential models. Recent modeling efforts have demonstrated the advantage of incorporating structural information to overcome the need for extensive training sequence data, yet disentangling the heterogeneous TCR-antigen structural interface to accurately predict the MHC-allele-restricted TCR-peptide binding interactions remained challenging. Here, we present RACER-m, a coarse-grained structural template model leveraging key biophysical information from the diversity of publicly available TCR-antigen crystal structures. We find explicit inclusion of structural content substantially reduces the required number of training examples for reliable prediction of TCR-recognition specificity and sensitivity across diverse biological contexts. We demonstrate that our structural model capably identifies biophysically meaningful point-mutants that affect overall binding affinity, distinguishing its ability in predicting TCR specificity of point mutants peptides from alternative sequence-based methods. Collectively, our approach combines biophysical and inferential learning-based methods to predict TCR-peptide binding events using sparse training data. Its application is broadly applicable to studies involving both closely-related and structurally diverse TCR-peptide pairs.

biophysics↗

bioRxiv2023

Modeling the role of immune cell conversion in the tumor-immune microenviroment

Tumors develop in a complex physical, biochemical, and cellular milieu, referred to as the tumor microenvironment. Of special interest is the set of immune cells that reciprocally interact with the tumor, the tumor-immune microenvironment (TIME). The diversity of cell types and cell-cell interactions in the TIME has led researchers to apply concepts from ecology to describe the dynamics. However, while tumor cells are known to induce immune cells to switch from anti-tumor to pro-tumor phenotypes, this type of ecological interaction has been largely overlooked. To address this gap in cancer modeling, we develop a minimal, ecological model of the TIME with immune cell conversion, to highlight this important interaction and explore its consequences. A key finding is that immune conversion increases the range of parameters supporting a co-existence phase in which the immune system and the tumor reach a stalemate. Our results suggest that further investigation of the consequences of immune cell conversion, using detailed, data-driven models, will be critical for greater understanding of TIME dynamics.

cancer biology↗

bioRxiv2023

Dual role of CASP8AP2/FLASH in regulating epithelial-to-mesenchymal (EMT) plasticity

Metastasis consists of sequential steps initiated by cancer cells invading from the primary tumor site into neighboring tissues, followed by entry into the circulatory system and completed by extravasation and growth in distal organs where secondary tumors are formed. Circulating tumor cells, thus, encounter and adapt to multiple environmental changes during their transition from the primary to the secondary tumor sites. Epithelial-to-mesenchymal transition (EMT) is a developmental program that consists of loss of epithelial features concomitant with acquisition of mesenchymal features. Activation of EMT in cancer facilitates acquisition of aggressive traits and cancer invasion. EMT plasticity (EMP), the dynamic transition between multiple hybrid states in which cancer cells display both epithelial and mesenchymal phenotypes, confers survival advantages for cancer cells in the constantly changing environment. Therefore, understanding the molecular mechanisms regulating intermediate phenotypic states along the E-M spectrum is critical. Core EMT transcription factors (EMT-TFs), ZEB, SNAI and TWIST families, play an important role in EMT and its plasticity. In the present study we characterize FLASH as a regulator of EMP and multiple EMT-TFs. We demonstrate that loss of FLASH gives rise to a hybrid E/M phenotype with high epithelial scores even in the presence of TGF{beta}, as determined by computational methods using expression of predetermined sets of epithelial and mesenchymal genes. We demonstrate that FLASH is regulating expression of multiple cell junction proteins with an established role in cancer progression and that its role in EMT is independent of its histone biogenesis role. Further, we show that FLASH expression in cancer lines is inversely correlated with the epithelial score, consistent with its function as a repressor of the epithelial phenotype. Nonetheless, activation of a distinct set of mesenchymal markers concomitant with epithelial markers reveals the complex role of FLASH in EMT and indicates that intermediate E/M states could arise from opposing control by FLASH on different families of EMT-TFs.

cancer biology↗