Biomimetic scaffolds for regeneration of volumetric muscle loss in skeletal muscle injuries

Skeletal muscle injuries typically result from traumatic incidents such as combat injuries where soft-tissue extremity injuries are present in one of four cases. Further, about 4.5 million reconstructive surgical procedures are performed annually as a result of car accidents, cancer ablation, or cosmetic procedures. These combat- and trauma-induced skeletal muscle injuries are characterized by volumetric muscle loss (VML), which significantly reduces the functionality of the injured muscle. While skeletal muscle has an innate repair mechanism, it is unable to compensate for VML injuries because large amounts of tissue including connective tissue and basement membrane are removed or destroyed. This results in a significant need to develop off-the-shelf biomimetic scaffolds to direct skeletal muscle regeneration. Here, the structure and organization of native skeletal muscle tissue is described in order to reveal clear design parameters that are necessary for scaffolds to mimic in order to successfully regenerate muscular tissue. We review the literature with respect to the materials and methodologies used to develop scaffolds for skeletal muscle tissue regeneration as well as the limitations of these materials. We further discuss the variety of cell sources and different injury models to provide some context for the multiple approaches used to evaluate these scaffold materials. Recent findings are highlighted to address the state of the field and directions are outlined for future strategies, both in scaffold design and in the use of different injury models to evaluate these materials, for regenerating functional skeletal muscle.

STATEMENT OF SIGNIFICANCE

Volumetric muscle loss (VML) injuries result from traumatic incidents such as those presented from combat missions, where soft-tissue extremity injuries are represented in one of four cases. These injuries remove or destroy large amounts of skeletal muscle including the basement membrane and connective tissue, removing the structural, mechanical, and biochemical cues that usually direct its repair. This results in a significant need to develop off-the-shelf biomimetic scaffolds to direct skeletal muscle regeneration. In this review, we examine current strategies for the development of scaffold materials designed for skeletal muscle regeneration, highlighting advances and limitations associated with these methodologies. Finally, we identify future approaches to enhance skeletal muscle regeneration.

Programmable Hydrogel Ionic Circuits for Biologically Matched Electronic Interfaces

The increased need for wearable and implantable medical devices has driven the demand for electronics that interface with living systems. Current bioelectronic systems have not fully resolved mismatches between engineered circuits and biological systems, including the resulting pain and damage to biological tissues. Here, salt/poly(ethylene glycol) (PEG) aqueous two-phase systems are utilized to generate programmable hydrogel ionic circuits. High-conductivity salt-solution patterns are stably encapsulated within PEG hydrogel matrices using salt/PEG phase separation, which route ionic current with high resolution and enable localized delivery of electrical stimulation. This strategy allows designer electronics that match biological systems, including transparency, stretchability, complete aqueous-based connective interface, distribution of ionic electrical signals between engineered and biological systems, and avoidance of tissue damage from electrical stimulation. The potential of such systems is demonstrated by generating light-emitting diode (LED)-based displays, skin-mounted electronics, and stimulators that deliver localized current to in vitro neuron cultures and muscles in vivo with reduced adverse effects. Such electronic platforms may form the basis of future biointegrated electronic systems.

A two-component model of host-parasitoid interactions: determination of the size of inundative releases of parasitoids in biological pest control

A two-component differential equation model is formulated for a host-parasitoid interaction. Transient dynamics and population crashes of this system are analysed using differential inequalities. Two different cases can be distinguished: either the intrinsic growth rate of the host population is smaller than the maximum growth rate of the parasitoid or vice versa. In the latter case, the initial ratio of parasitoids to hosts should exceed a given threshold, in order to (temporarily) halt the growth of the host population. When not only oviposition but also host-feeding occurs the dynamics do not change qualitatively. In the case that the maximum growth rate of the parasitoid population is smaller than the intrinsic growth rate of the host, a threshold still exists for the number of parasitoids in an inundative release in order to limit the growth of the host population. The size of an inundative release of parasitoids, which is necessary to keep the host population below a certain level, can be determined from the two-component model. When parameter values for hosts and parasitoids are known, an effective control of pests can be found. First it is determined whether the parasitoids are able to suppress their hosts fully. Moreover, using our simple rule of thumb it can be assessed whether suppression is also possible when the relative growth rate of the host population exceeds that of the parasitoid population. With a numerical investigation of our simple system the design of parasitoid release strategies for specific situations can be computed.

Stochastic epidemics: major outbreaks and the duration of the endemic period

A study is made of a two-dimensional stochastic system that models the spread of an infectious disease in a population. An asymptotic expression is derived for the probability that a major outbreak of the disease will occur in case the number of infectives is small. For the case that a major outbreak has occurred, an asymptotic approximation is derived for the expected time that the disease is in the population. The analytical expressions are obtained by asymptotically solving Dirichlet problems based on the Fokker-Planck equation for the stochastic system. Results of numerical calculations for the analytical expressions are compared with simulation results.

Oxygen diffusivity of biologic and synthetic scaffold materials for tissue engineering

Scaffolds for tissue engineering and regenerative medicine applications are commonly manufactured from synthetic materials, intact or isolated components of extracellular matrix (ECM), or a combination of such materials. After surgical implantation, the metabolic requirements of cells that populate the scaffold depend upon adequate gas and nutrient exchange with the surrounding microenvironment. The present study measured the oxygen transfer through three biologic scaffold materials composed of ECM including small intestinal submucosa (SIS), urinary bladder submucosa (UBS), and urinary bladder matrix (UBM), and one synthetic biomaterial, Dacron. The oxygen diffusivity was calculated from Fick's first law of diffusion. Each material permitted measurable oxygen diffusion. The diffusivity of SIS was found to be dependent on the direction of oxygen transfer; the oxygen transfer in the abluminal-to-luminal direction was significantly greater than the luminal-to-abluminal direction. The oxygen diffusivity of UBM and UBS were similar despite the presence of an intact basement membrane on the luminal surface of UBM. Dacron showed oxygen diffusivity values seven times greater than the ECM biomaterials. The current study showed that each material has unique oxygen diffusivity values, and these values may be dependent on the scaffold's ultrastructure.

Recent Trends in Injury Models to Study Skeletal Muscle Regeneration and Repair

Skeletal muscle injuries that occur from traumatic incidents, such as those caused by car accidents or surgical resections, or from injuries sustained on the battlefield, result in the loss of functionality of the injured muscle. To understand skeletal muscle regeneration and to better treat these large scale injuries, termed volumetric muscle loss (VML), in vivo injury models exploring the innate mechanisms of muscle injury and repair are essential for the creation of clinically applicable treatments. While the end result of a muscle injury is often the destruction of muscle tissue, the manner in which these injuries are induced as well as the response from the innate repair mechanisms found in muscle in each animal models can vary. This targeted review describes injury models that assess both skeletal muscle regeneration (i.e., the response of muscle to myotoxin or ischemic injury) and skeletal muscle repair (i.e., VML injury). We aimed to summarize the injury models used in the field of skeletal muscle tissue engineering, paying particular attention to strategies to induce muscle damage and how to standardize injury conditions for future experiments.

Stochastic epidemics: the expected duration of the endemic period in higher dimensional models

A method is presented to approximate the long-term stochastic dynamics of an epidemic modelled by state variables denoting the various classes of the population such as in SIR and SEIR model. The modelling includes epidemics in populations at different locations with migration between these populations. A logistic stochastic process for the total infectious population is formulated; it fits the long-term stochastic behaviour of the total infectious population in the full model. A good approximation is obtained if only the dynamics near the equilibria is fit.

Self-assembling Peptide Hydrogels Facilitate Vascularization in Two-Component Scaffolds

One of the major constraints against using polymeric scaffolds as tissue-regenerative matrices is a lack of adequate implant vascularization. Self-assembling peptide hydrogels can sequester small molecules and biological macromolecules, and they can support infiltrating cells in vivo. Here we demonstrate the ability of self-assembling peptide hydrogels to facilitate angiogenic sprouting into polymeric scaffolds after subcutaneous implantation. We constructed two-component scaffolds that incorporated microporous polymeric scaffolds and viscoelastic nanoporous peptide hydrogels. Nanofibrous hydrogels modified the biocompatibility and vascular integration of polymeric scaffolds with microscopic pores (pore diameters: 100-250 μm). In spite of similar amphiphilic sequences, charges, secondary structures, and supramolecular nanostructures, two soft hydrogels studied herein had different abilities to aid implant vascularization, but had similar levels of cellular infiltration. The functional difference of the peptide hydrogels was predicted by the difference in the bioactive moieties inserted into the primary sequences of the peptide monomers. Our study highlights the utility of soft supramolecular hydrogels to facilitate host-implant integration and control implant vascularization in biodegradable polyester scaffolds in vivo. Our study provides useful tools in designing multi-component regenerative scaffolds that recapitulate vascularized architectures of native tissues.

A deterministic model of the cell cycle

The variability of the duration of the cell cycle is explained by the phenomenon of sensitive dependence upon initial conditions; as may occur in deterministic non-linear systems. Chaotic dynamics of a system is the result of this sensitive dependence. First a deterministic system is formulated that is equivalent to the Smith-Martin transition probability model of the cell cycle. Next the model is extended to a dynamic process that ranges over the cell generations. A deterministic non-linear relationship between the cycle time of the mother and daughter cell is established. It clarifies the variability of mother-daughter correlation for the different cell types. The model is fitted to two different cell cultures; it shows that the graph of the non-linear relation has the same shape for different cell types.

Crosslinking strategies facilitate tunable structural properties of fibrin microthreads

A significant challenge in the design of biomimetic scaffolds is combining morphologic, mechanical, and biochemical cues into a single construct to promote tissue regeneration. In this study, we analyzed the effects of different crosslinking conditions on fibrin biopolymer microthreads to create morphologic scaffolds with tunable mechanical properties that are designed for directional cell guidance. Fibrin microthreads were crosslinked using carbodiimides in either acidic or neutral buffer, and the mechanical, structural, and biochemical responses of the microthreads were investigated. Crosslinking in the presence of acidic buffer (EDCa) created microthreads that had significantly higher tensile strengths and moduli than all other microthreads, and failed at lower strains than all other microthreads. Microthreads crosslinked in neutral buffer (EDCn) were also significantly stronger and stiffer than uncrosslinked threads and were comparable to contracting muscle in stiffness. Swelling ratios of crosslinked microthreads were significantly different from each other and uncrosslinked controls, suggesting a difference in the internal organization and compaction of the microthreads. Using an in vitro degradation assay, we observed that EDCn microthreads degraded within 24h, six times slower than uncrosslinked control threads, but EDCa microthreads did not show any significant indication of degradation within the 7-day assay period. Microthreads with higher stiffnesses supported significantly increased attachment of C2C12 cells, as well as increases in cell proliferation without a decrease in cell viability. Taken together, these data demonstrate the ability to create microthreads with tunable mechanical and structural properties that differentially direct cellular functions. Ultimately, we anticipate that we can strategically exploit these properties to promote site-specific tissue regeneration.

Reconstruction of the seasonally varying contact rate for measles

With the extended Kalman filter the time dependent contact rate for measles in the SEIR-model is reconstructed from data of the incidence of this infectious disease in the city of New York. It is concluded that although these data show through the years an irregular change in the number of infected children the contact rate is definitely periodic and follows the season. The analysis gives improved values of the parameters in the SEIR-model for this special problem.

A new class of non-linear stochastic population models with mass conservation

We study the effects of random feeding, growing and dying in a closed nutrient-limited producer/consumer system, in which nutrient is fully conserved, not only in the mean, but, most importantly, also across random events. More specifically, we relate these random effects to the closest deterministic models, and evaluate the importance of the various times scales that are involved. These stochastic models differ from deterministic ones not only in stochasticity, but they also have more details that involve shorter times scales. We tried to separate the effects of more detail from that of stochasticity. The producers have (nutrient) reserve and (body) structure, and so a variable chemical composition. The consumers have only structure, so a constant chemical composition. The conversion efficiency from producer to consumer, therefore, varies. The consumers use reserve and structure of the producers as complementary compounds, following the rules of Dynamic Energy Budget theory. Consumers die at constant specific rate and decompose instantaneously. Stochasticity is incorporated in the behaviour of the consumers, where the switches to handling and searching, as well as dying are Poissonian point events. We show that the stochastic model has one parameter more than the deterministic formulation without time scale separation for conversions between searching and handling consumers, which itself has one parameter more than the deterministic formulation with time scale separation for these conversions. These extra parameters are the contributions of a single individual producer and consumer to their densities, and the ratio of the two, respectively. The tendency to oscillate increases with the number of parameters. The focus bifurcation point has more relevance for the asymptotic behaviour of the stochastic model than the Hopf bifurcation point, since a randomly perturbed damped oscillation exhibits a behaviour similar to that of the stochastic limit cycle particularly near this bifurcation point. For total nutrient values below the focus bifurcation point, the system gradually becomes more confined to the direct neighbourhood of the isocline for which the producers do not change.

Static axial stretching enhances the mechanical properties and cellular responses of fibrin microthreads

Fibrin microthreads are a platform technology that can be used for a variety of applications, and therefore the mechanical requirements of these microthreads differ for each tissue or device application. To develop biopolymer microthreads with tunable mechanical properties, we analyzed fibrin microthread processing conditions to strengthen the scaffold materials without the use of exogenous crosslinking agents. Fibrin microthreads were extruded, dried, rehydrated and static axially stretched 0-200% of their original lengths; then the mechanical and structural properties of the microthreads were assessed. Stretching significantly increased the tensile strength of microthreads 3-fold, yielding scaffolds with tensile strengths and stiffnesses that equaled or exceeded values reported previously for carbodiimide crosslinked threads without affecting intrinsic material properties such as strain hardening or Poisson's ratio. Interestingly, these stretching conditions did not affect the rate of proteolytic degradation of the threads. The swelling ratios of stretched microthreads decreased, and scanning electron micrographs showed increases in grooved topography with increased stretch, suggesting that stretching may increase the fibrillar alignment of fibrin fibrils. The average cell alignment with respect to the longitudinal axis of the microthreads increased 2-fold with increased stretch, further supporting the hypothesis that stretching microthreads increases the alignment of fibrin fibrils on the surfaces of the scaffolds. Together, these data suggest that stretching fibrin microthreads generates stronger materials without affecting their proteolytic stability, making stretched microthreads ideal for implantable scaffolds that require short degradation times and large initial loading properties. Further modifications to stretched microthreads, such as carbodiimide crosslinking, could generate microthreads to direct cell orientation and align tissue deposition, with additional resistance to degradation for use as a long-term scaffold for tissue regeneration.

The multifractal structure of arterial trees

Fractal properties of arterial trees are analysed using the cascade model of turbulence theory. It is shown that the branching process leads to a non-uniform structure at the micro-level meaning that blood supply to the tissue varies in space. From the model it is concluded that, depending on the branching parameter, vessels of a specific size contribute dominantly to the blood supply of tissue. The corresponding tissue elements form a dense set in the tissue. Furthermore, if blood flow in vessels can get obstructed with some probability, the above set of tissue elements may not be dense anymore. Then there is the risk that, spread out over the tissue, nutrient and gas exchange fall short.

Tissue Models for Neurogenesis and Repair in 3D

Development and maturation of vascular and neuronal tissues occurs simultaneously in utero, and are regulated by significant crosstalk. We report on the development of a 3D tissue system to model neurogenesis and recapitulate developmental signaling conditions. Human umbilical vein endothelial cells (HUVECs) were seeded inside channels within collagen gels to represent nascent vascular networks. Axons extending from chicken dorsal root ganglia (DRGs) grew significantly longer and preferentially towards the HUVEC seeded channels with respect to unloaded channels. To replicate these findings without the vascular component, channels were loaded with brain-derived neurotrophic factor (BDNF), the principle signaling molecule in HUVEC-stimulated axonal growth, and axons likewise were significantly longer and grew preferentially towards the BDNF-loaded channels with respect to controls. This 3D tissue system was then used as an in vitro replicate for peripheral nerve injury, with neural repair observed within 2 weeks. These results demonstrate that our 3D tissue system can model neural network formation, repair after laceration injuries, and can be utilized to further study how these networks form and interact with other tissues, such as skin or skeletal muscle.

The Effect of Sterilization Methods on the Structural and Chemical Properties of Fibrin Microthread Scaffolds

A challenge for the design of scaffolds in tissue engineering is to determine a terminal sterilization method that will retain the structural and biochemical properties of the materials. Since commonly used heat and ionizing energy-based sterilization methods have been shown to alter the material properties of protein-based scaffolds, the effects of ethanol and ethylene oxide (EtO) sterilization on the cellular compatibility and the structural, chemical, and mechanical properties of uncrosslinked, UV crosslinked, or 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) crosslinked fibrin microthreads in neutral (EDCn) or acidic (EDCa) buffers are evaluated. EtO sterilization significantly reduces the tensile strength of uncrosslinked microthreads. Surface chemistry analyses show that EtO sterilization induces alkylation of EDCa microthreads leading to a significant reduction in myoblast attachment. The material properties of EDCn microthreads do not appear to be affected by the sterilization method. These results significantly enhance the understanding of how sterilization or crosslinking techniques affect the material properties of protein scaffolds.

Porous biomaterial scaffolds for skeletal muscle tissue engineering

Volumetric muscle loss is a traumatic injury which overwhelms the innate repair mechanisms of skeletal muscle and results in significant loss of muscle functionality. Tissue engineering seeks to regenerate these injuries through implantation of biomaterial scaffolds to encourage endogenous tissue formation and to restore mechanical function. Many types of scaffolds are currently being researched for this purpose. Scaffolds are typically made from either natural, synthetic, or conductive polymers, or any combination therein. A major criterion for the use of scaffolds for skeletal muscle is their porosity, which is essential for myoblast infiltration and myofiber ingrowth. In this review, we summarize the various methods of fabricating porous biomaterial scaffolds for skeletal muscle regeneration, as well as the various types of materials used to make these scaffolds. We provide guidelines for the fabrication of scaffolds based on functional requirements of skeletal muscle tissue, and discuss the general state of the field for skeletal muscle tissue engineering.

Hyperosmolar potassium inhibits myofibroblast conversion and reduces scar tissue formation

Scar formation is a natural result of almost all wound healing in adult mammals. Unfortunately, scarring disrupts normal tissue function and can cause significant physical and psychological distress. In addition to improving surgical techniques to limit scar formation, several therapies are under development towards the same goal. Many of these treatments aim to disrupt transforming growth factor β1 (TGFβ1) signaling, as this is a critical control point for fibroblast differentiation into myofibroblasts; a contractile cell that organizes synthesized collagen fibrils into scar tissue. The present study aimed to examine the role of hyperosmolar potassium gluconate (KGluc) on fibroblast function in skin repair. KGluc was first determined to negatively regulate fibroblast proliferation, metabolism, and migration in a dose-dependent manner in vitro. Increasing concentrations of KGluc also inhibited differentiation into myofibroblasts, suggesting that local KGluc treatment might reduce fibrosis. KGluc delivery was confirmed via loading into collagen hydrogels and used to treat a full thickness skin wound in mice. KGluc qualitatively slowed initial closure of the wounds and resulted in tissue that more closely resembled mature, healthy skin (epidermal thickness and dermal-epidermal morphology) when compared to unloaded collagen hydrogels. KGluc treatment significantly reduced the number of myofibroblasts within the dermis while upregulated blood vessel density with respect to unloaded hydrogels, likely a result of disruption of TGFβ1 signaling. Taken together, these data demonstrate the effectiveness of KGluc treatment on skin wound healing and suggest that this may be an efficient treatment to limit scar formation.

Mathematical conservation ecology: a one-predator-two-prey system as case study

A method is presented to analyse the long-term stochastic dynamics of a biological population that is at risk of extinction. From the full ecosystem the method extracts the minimal information to describe the long-term dynamics of that population by a stochastic logistic system. The method is applied to a one-predator-two-prey model. The choice of this example is motivated by a study on the near-extinction of a porcupine population by mountain lions whose presence is facilitated by mule deer taking advantage of a change in land use. The risk of extinction is quantified by the expected time of extinction of the population.

Angiogenic Hydrogels to Accelerate Early Wound Healing

The metabolic disorder diabetes mellitus affects an increasing proportion of the population, a number projected to double by 2060. Non-life-threatening comorbidities contribute to an interrupted healing process which is first delayed, then prolonged, and associated with increased susceptibility to infection and sustained and unresolved inflammation. This leads to chronic non-healing wounds and eventually potential amputation of extremities. Here we examine the use of a bioactive angiogenic peptide-based hydrogel, SLan, to improve early wound healing in diabetic rats, and compare its performance to clinically utilized biosynthetic peptide-based materials such as Puramatrix. Streptozotocin-treated diabetic rats underwent 8 mm biopsy wounding in their dorsum to remove the epithelium, adipose tissues and muscle layer of the skin, and served as a model for diabetic wound healing. Wounds were treated with either Low (1w%) SLan, High (4w%) SLan, PBS, Puramatrix or K2 (an unfunctionalized non-bioactive control sequentially similar to SLan), covered with Tegaderm and monitored on days 0, 3, 7, 10, 14, 17, 21, 28; animals were sacrificed for histomorphic analyses and immunostaining. An LC/MS method developed to detect SLan in plasma allows pharmacokinetic analysis showing no trafficking of peptides from the wound site into the circulation. Low and High SLan groups show similar final outcomes of wound contraction as control groups (Puramatrix, PBS and K2). SLan-treated rats, however, show marked improvement in healing in earlier time points, including increased deposition of new mature blood vessels. Additionally, rats in the Low SLan treatment groups showed significantly improved wound contraction over other groups and significantly improved healing in early time points. Altogether our results suggest this material can be used to "jumpstart" the diabetic wound healing process. This article is protected by copyright. All rights reserved.

* Design of an In Vitro Model of Cell Recruitment for Skeletal Muscle Regeneration Using Hepatocyte Growth Factor-Loaded Fibrin Microthreads

Large skeletal muscle defects that result in volumetric muscle loss (VML) result in the destruction of the basal lamina, which removes key signaling molecules such as hepatocyte growth factor (HGF) from the wound site, eliminating the endogenous capacity of these injuries to regenerate. We recently showed that HGF-loaded fibrin microthreads increased the force production in muscle tissues after 60 days in a mouse VML model. In this study, we created an in vitro, three-dimensional (3D) microscale outgrowth assay system designed to mimic cell recruitment in vivo, and investigated the effect of HGF-loaded, cross-linked fibrin microthreads on myoblast recruitment to predict the results observed in vivo. This outgrowth assay discretely separated the cellular and molecular functions (migration, proliferation, and chemotaxis) that direct outgrowth from the wound margin, creating a powerful platform to model cell recruitment in axially aligned tissues, such as skeletal muscle. The degree of cross-linking was controlled by pH and microthreads cross-linked using physiologically neutral pH (EDCn) facilitated the release of active HGF; increasing the two-dimensional migration and 3D outgrowth of myoblasts twofold. While HGF adsorbed to uncross-linked microthreads, it did not enhance myoblast migration, possibly due to the low concentrations that were adsorbed. Regardless of the amount of HGF adsorbed on the microthreads, myoblast proliferation increased significantly on stiffer, cross-linked microthreads. Together, the results of these studies show that HGF loaded onto EDCn microthreads supported enhanced myoblast migration and recruitment and suggest that our novel outgrowth assay system is a robust in vitro screening tool that predicts the performance of fibrin microthreads in vivo.

On the robustness of the geometrical model for cell wall deposition

All plant cells are provided with the necessary rigidity to withstand the turgor by an exterior cell wall. This wall is composed of long crystalline cellulose microfibrils embedded in a matrix of other polysaccharides. The cellulose microfibrils are deposited by mobile membrane bound protein complexes in remarkably ordered lamellar textures. The mechanism by which these ordered textures arise, however, is still under debate. The geometrical model for cell wall deposition proposed by Emons and Mulder (Proc. Natl. Acad. Sci. 95, 7215-7219, 1998) provides a detailed approach to the case of cell wall deposition in non-growing cells, where there is no evidence for the direct influence of other cellular components such as microtubules. The model successfully reproduces even the so-called helicoidal wall; the most intricate texture observed. However, a number of simplifying assumptions were made in the original calculations. The present work addresses the issue of the robustness of the model to relaxation of these assumptions, by considering whether the helicoidal solutions survive when three aspects of the model are varied. These are: (i) the shape of the insertion domain, (ii) the distribution of lifetimes of individual CSCs, and (iii) fluctuations and overcrowding. Although details of the solutions do change, we find that in all cases the overall character of the helicoidal solutions is preserved.

Jointly Optimized Spatial Histogram UNET Architecture (JOSHUA) for Adipose Tissue Segmentation

Objective. We aim to develop a machine learning algorithm to quantify adipose tissue deposition at surgical sites as a function of biomaterial implantation. Impact Statement. To our knowledge, this study is the first investigation to apply convolutional neural network (CNN) models to identify and segment adipose tissue in histological images from silk fibroin biomaterial implants. Introduction. When designing biomaterials for the treatment of various soft tissue injuries and diseases, one must consider the extent of adipose tissue deposition. In this work, we analyzed adipose tissue accumulation in histological images of sectioned silk fibroin-based biomaterials excised from rodents following subcutaneous implantation for 1, 2, 4, or 8 weeks. Current strategies for quantifying adipose tissue after biomaterial implantation are often tedious and prone to human bias during analysis. Methods. We used CNN models with novel spatial histogram layer(s) that can more accurately identify and segment regions of adipose tissue in hematoxylin and eosin (H&E) and Masson's trichrome stained images, allowing for determination of the optimal biomaterial formulation. We compared the method, Jointly Optimized Spatial Histogram UNET Architecture (JOSHUA), to the baseline UNET model and an extension of the baseline model, attention UNET, as well as to versions of the models with a supplemental attention-inspired mechanism (JOSHUA+ and UNET+). Results. The inclusion of histogram layer(s) in our models shows improved performance through qualitative and quantitative evaluation. Conclusion. Our results demonstrate that the proposed methods, JOSHUA and JOSHUA+, are highly beneficial for adipose tissue identification and localization. The new histological dataset and code used in our experiments are publicly available.