IFAC-PapersOnLine

Parameter estimation in nonlinear biological dynamical systems is a difficult inverse problem because the governing equations are often stiff or oscillatory, the data are sparse and noisy, and the objective landscape is non-convex. Physics-informed neural networks (PINNs) offer an alternative to purely simulation-based calibration by representing state trajectories with neural networks while penalizing violations of the governing equations. This paper studies the empirical reliability of PINNs for recovering the parameters of the repressilator, a synthetic genetic oscillator formed by three cyclically repressive genes. We use synthetic time-series generated from the standard ordinary differential equation model and train inverse PINNs to estimate the production parameter {beta} and the Hill coefficient n. The study varies observation noise, partial observation of repressors, sampling density, sensitivity to initial parameter guesses, and the difference between stable and oscillatory regimes. The results show that PINNs can reconstruct trajectories accurately when the model structure is correct and the three repressors are observed, but parameter recovery is more fragile than trajectory fitting. Noise, sparse sampling, unobserved variables, and unfavorable initial guesses increase the risk of biased estimates. The stable regime is easier to reconstruct, whereas the oscillatory regime provides richer information but also exposes optimization sensitivity. These findings support PINNs as a useful reverse-engineering tool for small gene-regulatory ODE models, while highlighting the need for repeated runs, uncertainty reporting, and experimental designs that improve identifiability.

Accurate modeling of laryngeal motor control is key to understanding typical and disordered voice production. However, traditional biomechanical plant models based on ordinary differential equations (ODEs) often involve high computational costs and numerical instabilities, limiting their use in real-time closed-loop control frameworks. This study evaluates feature-driven machine learning (ML) regressors, specifically Random Forest (RF), Multilayer Perceptron Neural Networks (NN), and Polynomial Regression (PR), as surrogate forward models mapping laryngeal motor inputs to fundamental frequency and sound pressure level. Training data were generated with two biomechanical vocal fold models: the extended body-cover and the triangular body-cover. Results demonstrate that ML surrogates reduce execution times from seconds to milliseconds (e.g., 2 ms for PR), enabling stable real-time tracking via inverse Jacobian control. While RF provides the highest accuracy, NN and PR offer smoother control signals and smaller memory footprints. A practical performance threshold was identified near N = 1,000 training samples, below which accuracy degraded substantially when models were trained from scratch. These findings support ML surrogates as efficient and adaptable alternatives to direct numerical simulation, providing a foundation for future subject-specific modeling through transfer learning in data-limited clinical scenarios.

In biotherapeutic manufacturing, product quality such as glycosylation profile is typically assessed only after harvest, limiting opportunities for corrective action during cell culture operation. Intracellular nucleotide sugar donors (NSD) directly determine glycosylation outcomes but are rarely measured, even offline, due to analytical complexity and process disruption. As a result, quality-related decisions remain constrained to fixed operating strategies. This work introduces a model-based soft sensing framework to infer NSD concentrations from readily available extracellular measurements. A Bayesian state estimation approach based on the Ensemble Kalman Filter (EnKF) is developed to reconstruct unmeasured intracellular states during CHO cell culture. An imperfect kinetic process model is combined with noisy extracellular measurements, explicitly accounting for process variability and measurement uncertainty through ensemble-based propagation and updates. The framework is validated using four independent experiments with distinct feeding perturbations that are not used for model calibration. Although the open-loop model exhibited substantial mismatch for both extracellular metabolites and intracellular NSDs, EnKF assimilation of extracellular measurements corrected key metabolic profiles. Building on these corrected extracellular dynamics, the EnKF demonstrated robust estimation of a growth-determining amino acid, asparagine, from correlated extracellular states. Based on the improved extracellular and amino acid estimates, the framework further enabled reliable inference of intracellular NSDs across all experiments.

The suspension culture of pluripotent stem (PS) cells in stirred bioreactors poses a delicate balance between maintaining homogeneous cell dispersion and avoiding excessive shear stress that can compromise cell viability and pluripotency. In this study, we used computational fluid dynamics (CFD) coupled with a discrete particle method (DPM) to simulate iPS cell behavior in a 5 mL delta-impeller stirred tank. Our analysis revealed that upward flow at the tank bottom and downward flow at the top are critical for maintaining a stable suspension. To optimize the stirring protocol, we applied Bayesian optimization to identify a time-dependent stirring schedule that begins with a high-speed phase for resuspension, followed by a low-speed phase for sustained suspension with minimal hydrodynamic stress. The optimized schedule demonstrated improved suspension ratio and reduced slip velocity, indicating lower mechanical stress on cells. These findings provide engineering insights into scalable bioreactor operation, contributing to the design of robust iPS cell manufacturing systems.

Hemodynamic forces play a key role in early cardiac morphogenesis, yet many computational studies assume Newtonian blood behavior. Here, we evaluate the impact of nonNewtonian shearthinning rheology on flow patterns, pressure distributions, and wall shear stress (WSS) during cardiac looping using idealized threedimensional models of the embryonic heart tube. Five geometries representing progressive looping stages, from a linear tube to an Sshaped configuration with ventricular ballooning, were analyzed under pulsatile flow using both Newtonian and powerlaw viscosity models. Across all stages, Reynolds numbers (Re {approx} 1-7) and Womersley numbers (Wo {approx} 0.3) indicated laminar, quasisteady flow consistent with embryonic conditions. Incorporating shearthinning rheology produced substantial deviations from Newtonian predictions, with peak systolic WSS differing by up to [~]40% and pressure drops by up to [~]20%. These effects were most pronounced in regions of increased curvature and geometric complexity. These findings demonstrate that nonNewtonian rheology significantly influences predicted hemodynamic environments during cardiac looping and should be incorporated into computational models aimed at understanding mechanobiological regulation of early heart development.

The development of large-scale, three-dimensional human tissues is crucial for various applications in therapeutic tissue engineering, disease modeling, and drug testing. However, due to the diffusion limit of oxygen, the lack of functional vascular networks is a significant limitation in maintaining these engineered tissues in the laboratory. To address this challenge, we present a systematic, model-based design process for artificial supply networks that can ensure a sufficient supply of oxygen and nutrients to engineered human tissue. Our approach combines mathematical models of fluid dynamics, cell metabolism, and network properties to identify key parameters influencing the supply performance. We demonstrate the applicability and possibilities of this design process by simulating different network structures, including cuboid and rhombic do-decahedral honeycombs, under various conditions. Our results show that the structure of the artificial supply network, oxygen concentration, and solute flow within the network strongly influence cellular metabolic activity and viability. We also examine the effects of non-uniform cell density, channel blockage, and long channel length on the oxygen distribution inside the cell-containing tissue compartment. Our findings highlight the importance of considering these factors in the design of artificial supply networks for large-scale engineered human tissues. This study provides a promising approach for quickly exploring the vast design space of possible network structures under different conditions for desired cell and tissue states, ultimately contributing to the development of more efficient and effective tissue engineering strategies.

Mechanistic ordinary differential equation models are widely used in systems biology to represent biochemical networks, population dynamics, cell-state transitions, and other biological processes; however, their predictive value depends critically on accurate parameter estimation from noisy and often sparse experimental data. In this tutorial, we present the Weak-form Estimation of Nonlinear Dynamics (WENDy) method as a forward-solver-free approach that reformulates parameter estimation as a covariance-corrected weak-form regression problem by integrating the model equations against compactly supported test functions. We present the background on the methodology through the lens of the familiar logistic equation, and we demonstrate applications of the method on real experimental data through two systems biology examples: a glycolytic oscillator with relatively dense time-course data and a sparse epithelial-mesenchymal cellstate transition model with multiple experimental replicates. Ultimately, using WENDy, we estimate interpretable biological parameters with uncertainty for systems with noisy and sometimes sparse available experimental data.

Standard target-mediated drug disposition (TMDD) models are widely used to describe nonlinear pharmacokinetics driven by high-affinity drug-target interactions. However, their reliance on instantaneous binding limits their ability to capture delayed and history-dependent dynamics observed in vivo. Here, we introduce a fractional TMDD model that incorporates memory effects through a fractional derivative, thereby generalizing the standard TMDD (sTMDD) framework. Although this fractional TMDD (fTMDD) formulation increases modeling flexibility, it also exacerbates parameter identifiability challenges under typical experimental conditions where only drug concentration data are available. To address this limitation, we derive a fractional quasi-steady-state approximation (fQSSA) that reduces model dimensionality while preserving essential nonlinear and memory-dependent pharmacokinetic dynamics. We further establish an explicit validity condition that quantifies the approximation error of both fTMDD and fQSSA without requiring numerical simulation. This condition reveals that the initial drug-to-target ratio is the primary determinant of QSSA validity, whereas the fractional order has a comparatively minor influence. Application of the proposed framework to recombinant human erythropoietin (rhEPO) data demonstrates that fractional dynamics play a population-dependent role, improving model performance in adults but not in infants. Together, this work provides the first systematic derivation of a QSSA framework for fractional TMDD models, along with rigorous and computable applicability conditions. Our results establish a principled foundation for incorporating memory effects into pharmacokinetic modeling and offer a generalizable framework for nonlinear PK-PD systems involving binding-mediated dynamics. Author summaryMany drugs interact strongly with their biological targets, leading to complex and nonlinear pharmacokinetics that are commonly described using target-mediated drug disposition (TMDD) models. However, these models assume that drug-target interactions occur instantaneously, which limits their ability to capture delayed and history-dependent behaviors observed in real biological systems. In this study, we develop a new modeling framework that incorporates such memory effects by extending TMDD models using fractional calculus. To make the model more practical and computationally efficient, we derive a simplified version based on a quasi-steady-state approximation (QSSA) and provide a clear mathematical condition that determines when this simplification is valid. Our analysis shows that the accuracy of the simplified model is primarily controlled by the initial ratio of drug to target, while the influence of memory effects is comparatively smaller. When applied to experimental data for erythropoietin, our model reveals that memory effects are important in adults but negligible in infants, suggesting that these effects may reflect underlying physiological differences. Overall, this work provides a systematic and interpretable framework for incorporating memory effects into pharmacokinetic modeling, with potential applications to a wide range of drug systems involving complex binding dynamics.

Control of fluids is a hallmark of microfluidic systems and fundamental for the successful application of microfluidic devices. Trigger valves use geometric features to autonomously control the release of fluids in microfluidic devices. Our previous work has adapted geometries used in closed trigger valve systems to enable use in open systems, allowing for open microfluidic devices with up to three trigger valves. Here, we focus on the parallel co-flows produced by sequential release of trigger valves and present a model that predicts their layer widths as a function of the geometric characteristics of the different side channels of each trigger valve. We show layered co-flows with widths as low as 50 microns. Additionally, we expand the use of trigger valves in open microfluidic devices by incorporating 1) varied step heights, 2) devices with up to seven trigger valves, and 3) use of varied fluids and plastics. To validate the implementation and use of these trigger valves in open systems, we have developed a theoretical framework to compare predicted outcomes (i.e., fluid travel distance, velocity, and layering width) with our experimental values. This theoretical work offers applications in various fields, including hydrogel patterning for 3D cell culture, organ-on-a-chip models, at-home sample preparation, and autonomous microfluidic systems for biosensing.

The optimization and control of bioprocesses require robust in silico models that can accurately capture the complex and dynamic behavior of living cells. While hybrid models that combine machine learning with mechanistic equations have emerged as a powerful tools, they often require relatively large datasets and might yield inconsistent predictions that violate the stoichiometry of metabolism. In this study, we introduce FBA-Hyb, a multi-scale hybrid modeling framework that tightly integrates genome-scale metabolic networks via flux balance analysis (FBA) into its architecture. In our FBA-Hyb framework, artificial neural networks predict key FBA inputs (substrate uptake rates and cellular objectives) while a surrogate FBA module translates them into the metabolic fluxes that govern the bioprocess. A key novelty is that the FBA optimization step is replaced by a surrogate generated with symbolic regression, which encapsulates the FBA model into a compact analytical expression. This allows easy backpropagation through the integration of the neural controlled differential equationbased FBA-Hyb bioprocess model. We validated FBA-Hyb against a standard hybrid model (Std-Hyb) using two Escherichia coli fedbatch case studies. In the first study, FBA-Hyb achieved a 42 % average improvement in predictive accuracy (R2) during a leave-one-process-out cross validation. Crucially, FBA-Hyb maintains strict stoichiometric feasibility even during extrapolation. Meanwhile, an alternative approach based on standard architectures leads to stoichiometrically inconsistent solutions in 22 % of the cases analyzed. In the second case study, we demonstrate how FBA-Hyb effectively simulates unmeasured chemical species and discovers a metabolic shift in sulfate-limited regimes during bioprocessing. By providing a modular, biologically consistent, and computationally efficient architecture, FBA-Hyb offers a robust foundation for the next generation of bioprocess models and sustainable process optimization. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=81 SRC="FIGDIR/small/720062v1_ufig1.gif" ALT="Figure 1"> View larger version (28K): org.highwire.dtl.DTLVardef@16f011eorg.highwire.dtl.DTLVardef@b25b5borg.highwire.dtl.DTLVardef@18bd178org.highwire.dtl.DTLVardef@65274e_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LIFBA-Hyb integrates flux balance analysis (FBA) into hybrid bioprocess models. C_LIO_LISymbolic regression discovers a simple closed-form FBA surrogate model. C_LIO_LIThe FBA surrogate ensures accurate reaction stoichiometry. C_LIO_LIA neural network predicting the FBA objective keeps the model flexible. C_LIO_LIFBA-Hyb has superior capabilities and accuracy compared to the current standard. C_LI

Ordinary differential equation models of biochemical reactions are often formulated as stoichiometric systems in which the dynamics arise from a collection of interacting processes. A central challenge is that the functional form of each process is rarely known a priori and may be difficult to infer from data. We propose biochemically informed neural ordinary differential equations (BINODEs), a neural-ODE framework that retains the stoichiometric structure of mechanistic models while representing individual processes by neural networks. In BINODEs, the outputs of neural network processes (NNPs) are mapped to state derivatives through a linear layer analogous to a stoichiometric matrix. This architecture allows biological side information, such as process-specific inputs, sign constraints, and monotonicity assumptions, to be built directly into the model. We characterize the approximation properties of NNPs for several standard biochemical rate laws and show that the proposed framework recovers both trajectories and process-level structure in Monod, Lotka-Volterra, pharmacokinetic, and ultradian endocrine models. These results suggest that BINODEs offer a useful compromise between mechanistic interpretability and data-driven flexibility for modeling partially known biochemical or biological dynamical systems.

Digital twins are increasingly recognized as a transformative technology for precision livestock farming; however, a major bottleneck in their development remains the scarcity of high-quality, high-granularity physiological data. This study introduces the expert-informed conditional diffusion (EICD) framework, a novel approach to synthesizing high-fidelity metabolic time-series trajectories by embedding mechanistic biological principles directly into the generative process. While traditional generative models often prioritize statistical pattern-matching over biological reality, frequently resulting in physiological hallucinations, the EICD framework utilizes a physiology loss function (PhLF) to act as a form of mechanistic regularization. This guardrail penalizes samples that contradict expert-defined constraints, such as the laws of porcine bioenergetics, effectively steering the model toward a realistic physiological manifold. The framework was validated using an empirical dataset of growing pigs under varying thermal conditions. Quantitative results demonstrate near-perfect statistical distributional fidelity, with the model achieving an average Jensen-Shannon divergence (JSD) of 0.062 and a Kullback-Leibler divergence (KLD) of 0.19. The full EICD model produced a mean energy expenditure (EE) of 284.94 {+/-} 38.70 kJ/kg/day, mirroring the empirical average of 281.33 {+/-} 41.58 kJ/kg/day. In contrast, the standard generative diffusion model (i.e., with no physiology guardrail) exhibited significant distributional drift, yielding a mean EE of 334.41 kJ/kg/day. The biological integrity of the model was further assessed using the biological violation rate (BVR), a novel metric defined as the percentage of generated samples that fall outside the physically possible metabolic boundaries established by species-specific laws. While the standard diffusion model produced frequent biological artifacts, the EICD framework successfully suppressed these hallucinations, ensuring that synthetic trajectories remain strictly grounded in mechanistic laws. Despite these advancements, limitations remain at physiological extremes where individual stochasticity is high. By providing a reliable method for generating physiology-consistent synthetic data, this framework provides a robust foundation for the next generation of animal digital twins. HighlightsO_LIA novel expert-informed conditional diffusion (EICD) framework is proposed for physiology-consistent synthetic data generation in precision livestock farming. C_LIO_LIA physiology loss function (PhLF) embeds species-specific bioenergetic laws directly into the generative process as a mechanistic guardrail. C_LIO_LIThe framework achieves near-perfect distributional fidelity (JSD = 0.062) while suppressing physiological hallucinations (BVR = 0.93%). C_LIO_LIAn ablation study confirms that biological consistency is not an emergent property of standard diffusion models but requires explicit mechanistic constraints. C_LIO_LIThe framework provides a scalable solution for synthetic data augmentation in precision livestock farming, supporting the 3Rs and enabling high-throughput in silico experimentation. C_LI

Thin structures such as heart valves and aortic dissection flaps interact dynamically with blood flow in human vessels. Their flexibility and capacity for large deformations generate complex, highly transient hemodynamic patterns over the cardiac cycle. Accurately resolving these interactions remains challenging for conventional boundary-fitted fluid-structure interaction approaches. We present an immersed boundary method for simulating thin structures in incompressible flow on unstructured grids. The method couples a stabilized finite element fluid solver with a nonlinear, rotation-free shell formulation through a direct forcing immersed boundary approach. The framework supports both weak (explicit) and strong (implicit) time-coupling strategies, enabling stable simulations over a wide range of solid-to-fluid density ratios. Hydrodynamic forces acting on thin structures are computed from fluid solutions sampled on both sides of the structure, allowing accurate force reconstruction for zero-thickness shells. To our knowledge, this is the first immersed boundary formulation that couples an unstructured finite element fluid solver with a two-dimensional, rotation-free shell model to simulate interactions between thin structures and incompressible flow. Fluid-structure coupling is achieved using predefined finite element shape functions, which provide consistent projection between Eulerian and Lagrangian fields without additional interpolation procedures. The framework is validated using three-dimensional benchmark problems involving thin structures. Then, valve-like model is used to compare strong and weak coupling strategies. Finally, the method is applied to an idealized type-B aortic dissection model. The proposed approach is implemented within the open-source software CRIMSON, a finite element platform for cardiovascular simulation.

The challenge of stratifying patients for combination therapy is both technically demanding and clinically crucial. In previous work, we introduced a multi-objective optimization framework for identifying optimally synergistic combination protocols that are robust to competing definitions of additivity. This manuscript extends this methodology to quantify how inter-individual variability in drug sensitivity influences the combination doses that optimally balance the competing objectives of synergy of efficacy and synergy of potency (a proxy measure of toxicity). For this methodology, we introduce a voxel-based stratification approach to characterize individuals (model parameterizations) into subgroups based on sensitivity to each drug as a monotherapy and in combination. As a case study, we apply the method to a preclinical dataset of murine response to the combination of an immune checkpoint inhibitor and an antiangiogenic agent. We demonstrate that the algorithm can quantify how the robustly optimal combination therapies vary across different treatment response subgroups and how the algorithm can identify subpopulations for which no meaningfully efficacious combination exists. As applying the methodology requires knowledge of specific parameter values for which measurable biomarkers may be unavailable, we also propose an initiation protocol that permits identification of the parameters necessary to place an individual in a subgroup. This methodology is a step in the direction of determining the right combination therapy for a subgroup and finding the right subgroup for an existing therapy.

Advanced in vitro systems such as organoids and microfluidic organ-on-a-chip platforms enable physiologically richer experimentation, but their complexity creates large parameter spaces and makes it difficult to disentangle the mechanistic roles of transport, mechanics, and extracellular microstructure. Agent-based modelling provides a natural computational counterpart to these systems by representing heterogeneous cells as discrete entities coupled through local rules and environmental fields. However, realistic microenvironment models often remain limited by scalability, simplified extracellular matrix representations, and the practical difficulty of calibrating large numbers of parameters. Here we present Cellfoundry, a computational framework built on a FLAMEGPU2-based modelling template for simulating complex cellular microenvironments. The framework integrates multiple interacting agent populations, including cells, fibrous networks, and focal adhesions mediating attachment dynamics and traction-force transmission. It combines mechanically resolved cell-cell and cell-matrix interactions with multi-species diffusion fields that propagate biochemical signals through the extracellular environment and regulate processes such as metabolism, migration, and cell-cycle progression. Cellfoundry also supports customizable behaviours across multiple cell types, enabling the study of heterogeneous multicellular systems within a unified computational setting. To support reproducible model development and calibration, the framework includes a fibre-network generation module, automated performance benchmarking workflows, post-processing and reporting utilities, and an Optuna-based Bayesian optimization pipeline with configurable single- and multi-objective targets. Two showcase examples illustrate these capabilities: a migration assay calibrated against fibroblast motility descriptors and a multi-objective organoid growth scenario reproducing target population composition and expansion dynamics and over time. Together, these examples demonstrate how Cellfoundry can be used to build, calibrate, and extend mechanistically interpretable models of coupled biochemical and mechanical dynamics in advanced in vitro systems. HighlightsO_LIHighly versatile, GPU-accelerated agent-based framework for cellular microenvironments C_LIO_LIExplicit fibrous ECM networks with dynamic remodelling and focal adhesion agents C_LIO_LICoupled mechanics and multi-species diffusion regulate cell behaviour in a highly customizable environment C_LIO_LIModular architecture with automated benchmarking and Bayesian parameter optimization C_LI

AnnotationMitochondrial heteroplasmy --the coexistence of both wild-type and mutant copies of mitochondrial DNA (mtDNA) within a cell--is a key factor in the pathogenesis of mitochondrial diseases. Classical approaches, which rely solely on the scalar fraction of mutant DNA, fail to fully account for threshold effects, the stochastic nature of heteroplasmy dynamics, and tissue specificity. The aim of the work is to construct a complex stochastic model of heteroplasmy dynamics, which for the first time combines the effects of selection, genetic drift, migration of mitochondrial genomes between tissues and threshold mechanisms of pathology development, for a quantitative assessment of the risk of mitochondrial diseases. In this paper, we propose a complex-phase formalism in which the state of a cells mitochondrial genome is described by a complex number Z = a + ib, where a and b are the absolute numbers of normal and mutant mtDNA copies, respectively. This approach naturally combines information on copy number and heteroplasmy level, and the argument{phi} = arctan (b / a) is interpreted as a phase characterizing the mutant load. Based on this formalism, we developed a stochastic model of tissue dynamics that includes the processes of selection, genetic drift, and intertissue migration of mitochondrial genomes. Using Monte Carlo methods (1000 simulations), we demonstrated that neuronal tissues are characterized by high heteroplasmy variability and a significant probability of reaching a pathological threshold even with a relatively low systemic mutant load. Kaplan-Meier survival analysis demonstrates that the development of pathology is probabilistic and can be described as a time -to-event process . The proposed approach enables quantitative assessment of the individual risk of developing mitochondrial diseases and opens the door to personalized prognosis.

Human pluripotent stem cells (hPSCs) hold significant promise for regenerative medicine, yet optimizing their expansion in three-dimensional bioreactor systems remains challenging due to complex interactions between mechanical forces, metabolic constraints, and aggregate formation dynamics. This study developed and validated a mechanistic mathematical model to predict hPSC proliferation dynamics in vertical-wheel bioreactor (VWBR) systems, incorporating the effects of shear stress and energy dissipation rate (EDR) on cell growth and aggregate dynamics. Seven model variants employing different kinetic formulations for shear stress and energy dissipation rate effects were systematically evaluated through model selection, identifiability analyses, and experimental validation. Experimental data from six bioreactor conditions varying in initial cell density (2 x 104-15 x 104 cells/mL), agitation rate (30-60 RPM), and working volume (100-500 mL) were used for model calibration and selection. Bayesian Information Criterion analysis identified a model combining Michaelis-Menten kinetics for shear stress inhibition with a EDR-mediated aggregate detachment formulation as the best-performing variant, achieving a Mean Relative Prediction Error of 13.97%, comparable to the experimental variability of 16.29%. Independent validation experiments using leave-out data gathered under different media exchange schedules confirmed model accuracy with prediction errors below 14%, consistent with observed experimental variability around 12%. The validated model was used to optimize the media exchange protocol, leading to a 37.5% reduction in media consumption with only a 13.5% reduction in final cell yield, demonstrating its utility for prospective, quantitative bioprocess design in VWBR systems.

Immune cell engagers (ICE) such as bispecific antibodies (bsAbs), within an immunological synapse, bind and link CD3 on a T cell to a target antigen (TAA) on a cancer cell, forming a trimer (CD3:bsAb:TAA complex). With sufficient trimer numbers within the synapse, the T cell can become activated and promote cancer cell killing. Elranatamab, a CD3-bispecific antibody for multiple myeloma, has received FDA and EMA filing acceptance (August 2023 and December 2023, respectively) adding to a growing list of bsAbs that are treating patients. In the drug development stages of ICE bsAbs, mechanistic modeling approaches are often used to attain a greater quantitative understanding of the modality, preclinically, and provide human pharmacokinetic and efficacious dose predictions to aide in Phase 1 trial design. To date, the majority of ordinary differential equation (ODE) trimer models treat the tumor compartment as well-mixed and trimer formation is governed by a bulk population reaction not accounting for individual synapses. This lack of discrimination can lead to imprecise analysis when analyzing results across E:T ratios using metrics like trimers per T cell or trimers per target cell. To this end we developed an ODE trimer model based on single-synapse complexes (one target cell/one immune cell) with 2D cross-linking trimer formation. We show computationally that the number of trimers per synapse is invariant to the value of the E:T ratio for a given free bsAb concentration, a property that cannot be captured by non-synapse models. A simple demonstration of this discrepancy using the well-known Betts trimer model is presented. We then apply the Betts trimer model coupled to a tumor growth inhibition (TGI) module to show that our synapse-based trimer model is easy to substitute in to model TGI, including the addition of a trimer-per-synapse activation threshold function for cell killing. Overall, our model attempts to balance mechanistic fidelity while limiting the complexity of the model.

Computational analysis of physiological and biomedical systems necessitate efficient geometry representations for high fidelity model predictions, including patient or device specificity. Particle-based Lagrangian computational approaches comprise a valuable approach to gain insights from quantitative velocity and pressure data from computational models. Examples include particle dynamics and transport in human vasculature for diseases such as stroke, thrombosis, and embolisms; and modern targeted drug delivery systems in the vascular network and respiratory airways. However, current particle simulation approaches can bear significant computational expense that scales with both number of particles and background fluid mesh resolution. A significant determinant of this computational expense is the contact resolution between particles and anatomically realistic vessel wall. Here, we develop an efficient particle dynamics model that leverages an implicit representation of real anatomical features using a signed distance field to efficiently resolve particle-wall contact. We outline the underlying algorithmic details, followed by a systematic illustration of performance and accuracy using simplified and analytically defined geometries and flow fields. Subsequently, we present a representative simulation of embolic particles along a human vascular segment where we compare our distance field-based approach against classical wall-contact checks based on assessing particle boundary intersection with triangulated surface mesh. Our approach transforms the underlying Lagrangian contact detection operation into an equivalent Eulerian operation, significantly speeding up bulk particle dynamics computations without significantly impacting accuracy or geometric fidelity.

Vertebrate gastrulating mesoderm is a prototypic example of a mesenchymal-like tissue undergoing extensive remodelling. While the tissue may be globally represented as a viscoelastic material, the actual biological material is intrinsically complex. To get to a real understanding of its properties, one needs to move to the mesoscale, linking cellular properties to collective phenomena. Vertebrate embryos also display a remarkable variability in mechanical properties, despite which they robustly complete gastrulation. This study attempts to explore these aspects by dissecting Xenopus mesoderm cell behaviour in a minimal system, using aspiration through a microfluidic system to impose controlled stress to a mesoderm aggregate. We show that beyond estimating global rheology at the tissue scale, it is possible to infer a wealth of information based on cell morphology and dynamics. Our data are consistent with collective behaviour being mostly dictated by the balance between the capacity of cells to stretch and the resistance to cell-cell contacts, which limits cell-cell intercalation and thus tissue remodelling. Importantly, tissues are not only able to transmit stress over a distance, they also clearly react to it through actively reinforcing cell-cell mechanical coupling. This adaptative property is found through a broad range of tissue stiffness, and adhesion strength appears to scale with the elastic modulus, suggesting that cell stiffness may ultimately be the key parameter setting mesoderm rheology and accounting for the large differences observed between embryo batches.