Mechanical cues presented by biomaterial surfaces are powerful regulators of cellular behavior, yet how topographical feature size differentially regulates neuronal and glial spatial organization remains poorly understood. Although anisotropic surface architectures have been shown to promote neural alignment and neurite outgrowth, systematic comparisons across multiple topographical regimes within a common material platform have been limited. To address this gap, we developed a scalable wrinkle-based nanofabrication platform capable of generating anisotropic MXene topographies spanning characteristic nano-, submicron-, and micron-scale regimes. By varying multilayer thickness, the platform enabled systematic shifts in characteristic wrinkle dimensions while maintaining a common material composition and surface chemistry. Using this platform, we found that distinct regenerative processes exhibit different responses across the topographical regimes. While all wrinkled topographies effectively aligned Schwann cells, the nanoscale topographies produced the greatest cellular elongation. In contrast, submicron and microscale topographies showed greater early neuronal alignment, whereas the submicron topography showed the greatest neurite extension while maintaining strong directional guidance following differentiation. These findings reveal distinct structure–response relationships between topographical regimes and neural guidance behavior. More broadly, this scalable wrinkle-based platform provides a versatile strategy for investigating topography-mediated cell–material interactions and informing the design of next-generation regenerative neural interfaces.
Engineering microbiome therapies requires an organizing principle for the mechanisms by which microbiome factors impact the host. We previously developed the “pharmacobiome”, a framework that uses mechanisms of drugs to quantify the host impact of microbiome factors. Here, we generalized this framework into a Microbiome Pharmacology Foundation Model (MPFM) for mechanism-informed design of microbiome therapeutics. We used correlation analysis of vaginal multi-omics data to generate host gene signatures for vaginal microbiome factors and compared the signatures to drug treatment host transcriptomics to construct microbiome factor-drug similarity scores. This Vaginal Microbiome Pharmacology database of 6,719,800 training instances enabled us to train the MPFM, a deep autoencoder model, and to derive experimentally testable predictions of individual microbiome factors and designed combination therapeutic properties. The MPFM training converged with strong performance (r = 0.57, MSE = 0.0058) and the latent space formed a continuous manifold. We analyzed cisplatin in the MPFM to identify anti-cancer metabolites and experimentally confirmed that taurine was Lactobacillus-derived and cytotoxic against cancer cell lines (p < 0.0005). We also analyzed metronidazole to identify antibiotic metabolites and experimentally showed hydroxyisocaproate was Lactobacillus-derived and selectively killed vaginal pathogens (p < 0.05). Finally, we used the model to explore patient-specific microbiome pharmacology and design probiotic combinations to support in vitro fertilization. We demonstrated consistent performance of the MPFM and experimentally supported the validity of predictions. Our continued work simulates probiotic, prebiotic, and postbiotic combinations for patient-specific in vivo sample analysis to develop tailored microbiome therapies for gynecologic conditions.
Angiogenesis plays a critical role in ovarian cancer progression, yet physiologically relevant in vitro models that recapitulate ovarian cell-specific tumor-vascular interactions remain limited. Standardized workflows for ovarian cancer spheroid generation, vascularization, and functional angiogenesis assessment within microphysiological systems are also lacking. Here, we developed a vascularized ovarian tumor microenvironment (TME)-on-chip platform and established a standardized modular framework for ovarian cancer modeling and angiogenesis studies. We systematically compared ovarian cancer cell lines and spheroid formation methods to define reproducible tumor modules with distinct morphological stability and pro-angiogenic phenotypes. Following on-chip integration, tumor viability, structural stability, and invasive behavior were evaluated. Multiple endothelial cell sources were compared to establish an ovarian cell-specific vascular component. Disease-relevant gain- and loss-of-function studies were subsequently performed using lysophosphatidic acid (LPA) and Bevacizumab, respectively. On-chip tumoroids maintained viability, structural stability, and invasive behavior throughout culture. Human ovarian microvascular endothelial cells (HOMECs) supported robust lumen formation, angiogenic remodeling, and tumor-associated vascular interactions. LPA treatment enhanced VEGF secretion and angiogenic network formation, whereas anti-VEGF treatment with Bevacizumab suppressed tumor-driven angiogenesis. This study establishes a standardized, biologically responsive, and ovarian cell-specific vascularized TME-on-chip model for investigating tumor-microvascular crosstalk, angiogenic signaling, and anti-angiogenic therapeutic responses, with future potential for incorporating stromal and immune components.
Glioblastoma (GBM) is a highly aggressive brain tumor with limited treatment options and poor survival rates. In this study, we investigate the potential of peanut-shaped and sphere-shaped microparticles as microrobots for targeted GBM therapy, specifically using the U87 GBM cell line as a model. Our findings demonstrate that both peanut and sphere-shaped microparticles effectively decrease U87 cell viability in a dose-dependent manner (61–77
In the absence of a reliable cure, Osteoarthritis (OA), treatment targets symptomatic pain and inflammation. However, chronically administered drugs often result in adverse side effects. In order to develop an improved treatment, our studies focused on controlled dual drug delivery. Bupivacaine (BUP), a local anesthetic, is used to manage pain directly and dexamethasone (DEX), a glucocorticoid, reduces inflammation. Key obstacles of both BUP and DEX administration, is that at high concentrations both drugs may be toxic, while at low concentrations they can provide pain relief and control inflammation for only a short time. We developed nanoparticles composed of Poly(lactic- co-glycolic acid) (PLGA), a biodegradable polymer, to prolong the release of lower doses of BUP and DEX for at least 30 days. Using the nanoparticle (NP) diffusion profiles, loading capacity, and optimal dose for our target cell types, we tested our dual NP therapy on chondrocytes, macrophages, and in baseline and activated co-cultures in order to compare bolus and NP drug effects either alone or in combination. We found that the combined nanoparticle co-therapy maintained cell viability and reduced inflammation more than either DEX nanoparticle (DexNP) or BUP nanoparticle (BupNP) therapy alone, especially in co-culture, and can therefore ultimately be incorporated into longer term therapeutic protocols. Pathway analysis enabled an approach to understanding the potential synergy between the two drugs. Overall, the macrophage, chondrocyte, and co-culture studies indicate that DualNP therapy maintains viability while reducing IL-8 secretion more than either DexNP, BupNP, and bolus DUAL in all conditions. While further analysis is needed, due to its versatility, our approach can be further developed to potentially improve an array of clinical outcomes.
