Li-Hsin Han

Adolescent idiopathic scoliosis (AIS) is a three-dimensional spinal deformity that can progress during skeletal growth due to asymmetric mechanical loading of the spine. Non-surgical bracing is commonly used to limit curve progression in skeletally immature patients, but treatment success depends on both the magnitude of immediate in-brace correction and patient compliance over the prescribed wear period. Rigid braces can generate corrective forces but may reduce comfort and compliance, while soft-bracing concepts may improve tolerability but require further biomechanical evaluation. Patient-specific finite element (FE) modeling provides a computational approach for quantifying how external forces are transferred through the osteoligamentous thoracic and lumbar spine, rib cage, and pelvis with the torso's external soft tissue (TEST) and abdomen. However, many existing AIS patient-specific FE models simplify the external torso using shell elements or do not include a representative torso external soft tissue and abdominal components. Therefore, the objective of this thesis was to develop an AIS patient-specific FE modeling framework capable of quantifying effects on spinal curve correction and apical intervertebral disc (IVD) stress from generalized brace-like corrective forces in a parametric simulation study. In Specific Aim 1, an AIS patient-specific FE template model was developed by morphing a normative pediatric osteoligamentous thoracic and lumbar spine-rib cage-pelvis FE model to an AIS patient-specific spinal and rib cage geometry and integrating a solid hexahedral torso external soft tissue (TEST) FE model and a simplified solid hexahedral abdominal FE component using connector beams to transmit corrective forces from the external surface of the torso to the osteoligamentous thoracic and lumbar spine-rib cage-pelvis FE model. A mesh convergence study as well as various quality assessments supported the selected element sizes, and frontal-plane geometric comparison showed less than 5% error for the evaluated anatomical dimensions. In Specific Aim 2, the AIS patient-specific FE template model was used in a parametric simulation study to evaluate generalized three-point brace-like corrective forces applied to the axillary, apical, and lumbar regions on the external surface of the TEST FE model. Corrective force magnitude, force-vector angle, and apical force location were varied using apical forces of 40 N, 60 N, and 80 N, force-vector angles of 10°, 25°, and 40° with respect to the transverse plane (XY-plane), and lateral, posterolateral, and anterolateral apical force locations. The simulations demonstrated that increasing force magnitude generally reduced Cobb angle, increased apical vertebral (T9) axial orientation, decreased average von Mises stress on the concave half of the T9-T10 IVD, and increased stress on the convex half of the T9-T10 IVD. Increasing the force-vector angle reduced Cobb angle and concave-side T9-T10 IVD stress, but was also associated with decreased T9 axial orientation. Among the apical force locations, posterolateral and lateral loading were more effective than anterolateral loading for Cobb angle correction and for producing lower concave-side T9-T10 IVD stress relative to the convex side, while anterolateral loading produced the greatest change in T9 axial orientation. The clinical significance of this work is that it provides a patient-specific computational framework for evaluating how corrective force magnitude, direction, apical force location, and interface stress influence immediate spinal correction, internal disc loading, and patient compliance. By quantifying changes from the baseline model state, this approach may help identify loading combinations that maximize initial in-brace correction while limiting undesirable stress distributions on the vertebrae and IVD. Although long-term growth modulation was not directly simulated, the T9-T10 IVD stress response provides a mechanical indicator of how brace-like loading may redistribute apical stresses in a manner relevant to growth modulation principles, such as the Hueter-Volkman law. More broadly, this framework can be adapted to investigate additional brace concepts, loading configurations, curve classifications, and spinal deformities beyond AIS. With appropriate patient-specific geometry, material properties, and loading definitions, the model provides a foundation for future computational studies of non-surgical spinal correction, load transfer, and patient-specific brace optimization.

Early onset scoliosis (EOS) is a heterogeneous spinal deformity affecting children under the age of 10, with curve progression driven by asymmetric vertebral growth dictated by the Hueter-Volkmann Law (HVL). Left untreated, EOS can lead to Thoracic Insufficiency Syndrome (TIS), a condition that impairs breathing and thoracic development. While EOS encompasses multiple etiologies, idiopathic EOS accounts for a substantial proportion of cases and is characterized by unpredictable curve progression. Computational finite element (FE) models have been used to simulate growth and predict deformity progression in adolescent idiopathic scoliosis, but limited models exist for idiopathic EOS. Existing models also lack comprehensive anatomy such as the pelvis, ribcage, and sternum. The purpose of this study was to develop a high-quality hexahedral idiopathic EOS osteoligamentous template FE model of the thoracic-lumbar spine with intervertebral discs (IVDs), spinal ligaments, pelvis, and ribcage and simulate ranges of motion (ROMs) through flexion-extension, lateral bending, and axial rotation, and evaluate deformity progression through 12 months of stress-modulated longitudinal growth. Chest CT scans of a 1-year-old patient with idiopathic EOS were retrospectively obtained from the Children's Hospital of Philadelphia (CHOP). Thoracic and lumbar vertebrae (T1-L5) and ribs were digitally segmented and reconstructed, then meshed with intervertebral discs and spinal ligaments. The pelvis was adapted from a previously developed AIS FE model, and the sternum and costal cartilage were simplified as beams. Material properties were derived from age-scaled adult values reported in literature. The complete osteoligamentous FE model consisted of 251689 elements and met established hexahedral mesh quality criteria: Jacobian [greater than or equal to] 0.5 (94%), warpage [less than or equal to] 40 (98%), skewness [less than or equal to] 60 (98%), minimum angle [greater than or equal to] 30° (97%), maximum angle [less than or equal to] 150° (95%), and aspect ratio [less than or equal to] 5 (86%). A ±0.50 N-m moment was applied to T1 and angular displacement of the thoracic (T1-T12) and lumbar (L1-L5) spines were compared against adult cadaveric in vitro data. The FE model matched the regional trends described in the in vitro data, with generally greater angular displacements across all loading modes. Notably, the FE model predicted high thoracolumbar (T12-L1) involvement during lateral bending towards the concavity of the curve (18°) in contrast with lateral bending towards the convexity of the curve (3°). A sensitivity analysis evaluating the effects of IVD stiffness on spinal flexibility suggested that these results likely reflect the structural effects of the idiopathic EOS deformity. Longitudinal growth was performed using the thermal expansion method in the vertebral bodies of the T1-L5 osteoligamentous FE spine. After antigravitational unloading and gravitational loading, 12 months of growth was simulated using literature-derived normative growth rates modulated by the von Mises stress across vertebral body regions. After 12 months of simulated growth, the T7-L2 Cobb angle increased by 10°, T1-L5 kyphosis decreased by 6°, apical vertebral (T11 vertebra) wedging increased by 1.24° in the frontal plane and decreased by 0.8° in the sagittal plane, and growth strain accumulated asymmetrically consistent with the HVL. Overall, the idiopathic EOS template FE model predicted deformity progression in idiopathic EOS through stress-modulated longitudinal growth. The clinical significance of the idiopathic EOS template FE model developed in this study is that it serves as a computational tool for understanding and predicting curve deformity progression in idiopathic EOS, which is a condition with limited clinical data. Furthermore, this template FE model contributes to the shift towards computational tools in pediatric spine orthopedics and could aid clinicians in patient-specific treatment planning for idiopathic EOS patients. Future studies could integrate algorithms to morph the existing spine geometry using patient-specific anatomical landmarks, develop FE models representative of other EOS etiologies, and evaluate the progression of TIS in EOS patients.

Hydrogels are widely used in tissue engineering, but conventional homogeneous polymerization often creates dense matrices that hinder cell migration and restrict extracellular matrix production. The motivation of this project was to overcome these limitations by developing a heterogeneously crosslinkable hydrogel platform that enables both cell migration and matrix deposition. We present a two-step heterogeneous polymerization approach that introduces spatial variations in matrix density, producing tunable, cell-sized pores that promote migration, proliferation, and matrix synthesis. As an implementation, gelatin was pre-assembled into microribbon-like building blocks using a dynamic molding process, methacrylated to introduce crosslinkable groups, chemically modified, washed, and freeze-dried. Upon rehydration, the ribbons formed a moldable paste that could be mixed with cells and photo-crosslinked into scaffolds with in situ –formed, cell-sized pores. The main novelty of this method is the introduction of chemical modifications with methacrylic anhydride (MAA), acetic anhydride (AceA), and succinic anhydride (SucA), which enable a controlled two-step heterogeneous polymerization and allow independent tuning of scaffold microstructure, mechanics, and degradation. AceA reduced crosslink density and accelerated degradation, whereas SucA promoted swelling, enhanced mechanical strength, and slowed degradation. Cell studies revealed that SucA-modified scaffolds supported superior adhesion and proliferation compared to AceA-modified and unmodified controls. Such work may significantly impact the design of next-generation scaffolds by providing a versatile platform that integrates structural, mechanical, and biochemical control for regenerative medicine applications.

Tissue engineering of load-bearing tissues such as cartilage and meniscus remains a major challenge due to their complex mechanical and structural requirements. This dissertation presents the development and application of a new scaffold--RibbonGel--to address this problem. RibbonGel consists of microribbon-shaped, crosslinkable elastomers fabricated using Dynamic Molding and Two-Step Crosslinking. These scaffolds are designed to mimic key features of the native extracellular matrix (ECM), including regional stiffness, porosity, and alignment, which impact cell behavior. By tuning these parameters, RibbonGel supports cell survival, proliferation, and zone-specific phenotype development. The ability to vary scaffold properties across regions enables the reconstruction of complex tissues with spatially distinct ECM composition, cellular phenotypes, and mechanical functions. The first part of this dissertation examined how side-chain functionalization affects RibbonGel microribbons. Functional groups were introduced to increase hydrophilicity, promote hydrophobicity, or enable photo-crosslinking. Varying the degree of these modifications altered the microribbons' microstructure, mechanical properties, swelling behavior, and degradation rate--parameters that can be optimized for different cell types and tissue applications. The second part developed a microribbon scaffold designed to circumferentially align vascular smooth muscle cells, replicating the organization of native blood vessels. This construction was applied as a lab-on-a-chip platform for drug testing in cardiovascular disease models. The third part used RibbonGel to regenerate zonally distinct meniscus tissues by varying cell source and tuning scaffold alignment, porosity, and stiffness. YAP/TAZ nuclear translocation was used to assess mechanotransducive signaling in response to these matrix cues. The fourth part explored the translational use of RibbonGel in a large animal model, demonstrating its effectiveness in supporting cartilage repair in horses via surgical implantation. The fifth and final part evaluated RibbonGels commercial and clinical potential through over 100 stakeholder interviews conducted as part of the NSF I-Corps National program. This analysis identified key opportunities and challenges related to regulatory approval, reimbursement, and clinical adoption. Together, this dissertation presents RibbonGel as a new scaffold system capable of guiding structurally and functionally distinct tissue regeneration. The work provides a foundation for future clinical translation in meniscus and cartilage repair, as well as broader applications in tissue engineering.

The gold standard to measure arterial health is vasodilation in response to nitric oxide. Vasodilation is generally measured via pressure myography of arteries isolated from animal models. However, animal arteries can be difficult to obtain and may have limited relevance to human physiology. It is, therefore, critical to engineer human cell-based arterial models capable of contraction. Vascular smooth muscle cells (SMCs) must be circumferentially aligned around the vessel lumen to contract the vessel, which is challenging to achieve in a soft blood vessel model. In this study, we used gelatin microribbons to circumferentially align SMCs inside a hydrogel channel. To accomplish this, we created tunable gelatin microribbons of varying stiffnesses and thicknesses and assessed how SMCs aligned along them. We then wrapped soft, thick microribbons around a needle and encapsulated them in a gelatin methacryloyl hydrogel, forming a microribbon-lined channel. Finally, we seeded SMCs inside the channel and showed that they adhered best to fibronectin and circumferentially aligned in response to the microribbons. Together, these data show that tunable gelatin microribbons can be used to circumferentially align SMCs inside a channel. This technique can be used to create a human artery-on-a-chip to assess vasodilation via pressure myography, as well as to align other cell types for 3D in vitro models.

Long-term drug delivery systems are an attractive option to drug manufacturers and patients alike due to simplified dosing schedules and reduced negative side effects which increase patient therapy compliance and quality of life. Being studied for decades, multiple controlled release systems have been developed to sustain the release of target therapeutic agents; however, these methods have various drawbacks such as difficulty in formulation, high cost, etc. that limit widespread commercial usage. One pervasive challenge in these solutions is the burst release phenomena-where an uncontrolled bolus of drug is released immediately upon product usage which lowers the effective lifetime of the drug or medical device. While methods have been developed to address burst release in sustained release systems, these methods often require extra processing steps that waste a portion of the initially loaded drug or may leave the device prone to failure if not processed correctly. To address this issue, a microfluidic release system was developed using a novel method of rapidly, tunably, and cost effectively creating microchannels bundles without the use of non-biocompatible materials. In this study, we investigated the sustained release of both a model small molecule and model large molecule from these novel microchannel bundles of varying channel diameters. From this, we discovered that rate of release was tunably controlled by modifying microchannel diameters in these bundles, significantly extending the product lifetime at increasingly thin channel diameters. Additionally, burst release was decreased by tunably lowering microchannel diameters within the bundle. As a result, this study provides a novel method of creating microfluidic drug delivery systems that could address major issues in current sustained release drug delivery systems.

An estimated 10% of the global population over the age of 60 have significant clinical problems that may be attributed to osteoarthritis. Current clinical solutions alleviate common symptoms; nothing on the market today addresses treatment of osteoarthritis itself. Regeneration of cartilage tissue lost in osteoarthritis is extremely challenging due to an innate lack of blood vessels, lymphatics and nerves. Tissue engineering may provide alternative solutions by developing biomimetic tissue substitutes. The goal of this project is to investigate how combination of biomechanical and biochemical properties can be used to create a pro-chondrogenic microenvironment, by means of controlled mechanotransduction and drug-delivery. To achieve such a goal, a novel technique is used to fabricate a fibrous, three-dimensional scaffold from natural polymer. Microfibers are generated by repeatedly stretching and folding gelatin sheathed in a sacrificial layer of PCL. Increasing the number of stretch-fold cycles reduces the fiber diameter while increasing the number of gelatin cores. After the PCL is dissolved, microfibers of three diameters are collected and UV-crosslinked into scaffolds. This provides scaffolds with microscopic pores that are either below or above than the typical size of cells, controlling of how cells sense and respond to mechanical environment, and subsequent signaling pathways that influence cartilage regeneration. IL-4 encapsulated PLGA microparticles were also synthesized, and embedded within the fibrous scaffold. Drug delivery from the final crosslinked product was measured over time to show continuous drug release for ~1 week. Modulation of phenotype after IL-4 stimulation was also explored. By combining mechanotransduction with drug delivery, this project aimed to investigate immunomodulatory properties of this scaffold to encourage macrophage polarization to the M2 phenotype. While M1 macrophages are known to promote inflammation and recruitment, M2 macrophages are known to release pro-chondrogenic, anti-inflammatory cytokines. Modulation of these properties may influence chondrogenic potential of cells at the injury site. Mechanotransduction of macrophages, in terms of attachment and activation, has been shown to be influenced by scaffold fiber diameter as well as pore size. Administration of IL-4 is known to promote the M2 phenotype both in-vitro and in-vivo. Both properties were combined here to approach the most pro-chondrogenic scaffold. Keywords: 3D Scaffold, Drug Delivery, Macrophage, Mechanotransduction, Microfiber Scaffold

Tissue engineering, which aims to regenerate damaged human tissues or organs with biomaterials and living cells, developed rapidly in the past few decades and has motivated remarkable advancement in the studies on cell engineering, biomaterials, and biofabrication. In spite of such progress, the regeneration of thick tissues, which is needed for repairing a sizable tissue defect, remains difficult. One of the major obstacles is the challenge of creating cell-laden, cell-scale hollow structures that resemble the microvascular networks of our tissues. In an implant for our body, these microvascular networks enable the exchange of nutrients, oxygen, and metabolic wastes between cells and the larger blood vessels, which sustaining the cells' and the tissues' survival. In this study, I developed a novel 4D printing platform that creates microscale, cell-encapsulating, hollow 3D scaffolds for tissue vascularization. Human umbilical vein endothelial cells (HUVECs) were seeded on an inkjet-printed bilayer gelatin micropattern, the cell-lined micropattern was made capable of self-folding into 3D microtubes, at a moment of choice. This work contains four major topics including (a) developing a 4D platform to implement 4D printing of microscale cell-encapsulating hollow 3D scaffold, (b) developing a mathematical model to decipher the dynamics of shape transformation of the self-folding micropatterns, (c) developing biomimetic microvessel/microvascular network using the 4D printing method and investigating how cells behave in response to the microtubular, physical environment given by the 4D printing, and (d) testing the feasibility of vascularizing a 3D tissue-engineering scaffold using the 4D-printed microvessels. This study provides a novel way to create 3D cell-encapsulating tissue engineering scaffold and sheds the light on creating vascularized 3D tissue engineering implants as regenerative medicine.

Tissue engineering is a promising solution to articular cartilage repair. The scaffolds, an essential component for tissue-engineering, are required for simulating cartilage regeneration by mimicking the 3D extracellular matrix (ECM) of natural cartilage. However, existing scaffolds often fails to effectively incorporate cells in 3D and lack the tunability of microstructures, such porosity, to regulate the cells' mechanosensing and subsequent cell bioactivities. To overcome this technological hurdle, FiberGel, a new type of scaffold made of a crosslinkable paste of microfibers, was created. Fabricated by a unique stretch-and-fold method, this new scaffold was capable of replicating the natural 3D framework of ECM and distributing cells evenly in 3D. Our previous study demonstrated that FiberGel was able to promote cartilage tissue formation, while varying the fiber size of FiberGel shown to have significant effects on chondrogenesis. In this study, I explored the potential of FiberGel for medical application by using a clinically relevant, in vitro model. The effect of implanting FiberGel scaffold into a cartilage defect was studied using plugs of living cartilages with subchondral bone, which were harvested from a horse. Horse mesenchymal stem cells were blended with FiberGels of two types of microfibers (subcellular-sized fiber and super cellular-sized fiber), implanted into defects created on the horse cartilage plugs, and incubated for long term cultured. After 42 days, both groups of different fiber size demonstrated dramatically increased integration to the host cartilage, with integration strength increasing from 0 up to 200 kPa. FiberGel made of subcellular-sized fibers were found to promote better integration. Both groups also demonstrated significant increased production of ECM components including collagens and glycosaminoglycan. Overall, the results indicated that FiberGel is highly promising for cartilage repair. This study also suggested that the performance of FiberGel scaffolds, particularly the scaffolds consisting of subcellular size fibers, can further enhance in vivo, in which body motions may further promote chondrogenesis.

The success of the stretch-and-fold method illustrates a more efficient and productive method to manufacture micro/nano-fibers. In the manufacturing process, the theory of Rayleigh-Plateau Instability predicts that surface tension between core and clad interface will overcome the viscosity of the core/clad liquids. This leads to the break of the core into segments. By using a thermoplastic clad (Polycaprolactone, PCL), the dissociation motion of the core's fluid was delayed due to the high viscosity of the melted clad and was finally arrested by to the solidification of the clad at the room temperature. Meanwhile, operating temperature and stretch-and-fold cycle speed are also important factors in the fabrication. To explore the relationship among all these operational factors, experiments were conducted by producing 15/20/25 folds fibers. The manual stretch-and-fold process, weight ratio calculation, scanning electron microscope imaging, and mathematical model approach were the main tasks performed during the entire experiment. Finally, an analysis based on the collected data and mathematical approximation. A prediction is made between stability and related experimental factors.