Biodegradable Silk Catheters for the Delivery of Therapeutics across Anatomical Repair Sites

Abstract

Biodegradable silk catheters for the delivery of therapeutics are designed with a focus on creating porous gradients that can direct the release of molecules away from the implantation site. Though suitable for a range of applications, these catheters are designed for drug delivery to transplanted adipose tissue in patients having undergone a fat grafting procedure. A common complication for fat grafts is the rapid reabsorption of large volume adipose transplants. In order to prolong volume retention, biodegradable catheters can be embedded into transplanted tissue to deliver nutrients, growth factors or therapeutics to improve adipocyte viability, proliferation and ultimately extend volume retention.

Two fabrication methods are developed: a silk gel spinning technique, which uses a novel flash-freezing step to induce high porosity throughout the bulk of the tube, and a dip coating process using silk protein solutions doped with a water soluble porogen. Increased porosity aids in the diffusion of drug through the silk tube in a controllable way. Additionally, we interface the porous tubes with ALZET osmotic pumps for implantation into a subcutaneous nude mouse model. The work described herein will discuss the processing parameters as well as the interfacing between pump and cargo therapeutic and the resulting release profiles.

Introduction

Medical catheters are frequently employed for a wide range of clinical applications, such as drainage of bodily fluids (urine drainage from kidney, liquids from cavities), intravenous fluid and drug administration (chemotherapeutics, insulin), cardiological procedures (angioplasty, balloon septostomy), and dialysis. These catheters must interface with patients’ tissues, typically via injection or implantation, for extended periods of time, such as for extended infusion or long-term release of drug or removal of fluids. In these cases, catheter tubes must possess sufficient physical strength to prevent collapse or blockage of the canal, biocompatibility, and low shear for delivery of biologicals or viscous fluids, which may undergo phase changes at high flow (e.g. blood clotting, protein aggregation, etc.). Recently, there has been increasing interest in developing medicated catheters that allow local drug delivery to target tissues. One example of this, are the drug-coated balloon catheters for the treatment of coronary disease.¹

A common issue with catheter implantation is the possibility of invasive surgical removal at the end of treatment. This removal process can disrupt local tissue regeneration and healing, leading to patient discomfort, tissue damage, and potentially necessitating follow-up procedures. Many catheters are designed from flexible, yet non-resorbable polymeric materials, such as natural rubber latex, polyvinyl chloride, polyurethane, polytetrafluoroethylene, and silicon rubber. Therefore, a clinical need exists to design catheters that are mechanically robust, biocompatible, flexible and porous, while also possessing sensitivity to endogenous factors within tissues (e.g. proteases, macrophages, etc.) that allow biodegradation at a predictable and tunable rate.

Bombyx mori silk fibroin is an outstanding material for biomedical application, broadly used in commercially available cosmetic and medical devices. Fibroin is extracted from the silkworm’s cocoons through a water-based process, avoiding the use of harsh chemicals or conditions. Reconstituted silk can be processed in many different formats, such as sponges, films, fibers and micro-particles,² with controllable mechanical properties and degradability.^3–5

The goal of this study was to explore different approaches for fabricating biodegradable, mechanically robust and flexible catheters with tunable porosity. Two different fabrication protocols were tested in order to obtain silk fibroin catheters with controllable architecture and permeability, for controlled release of drugs to target anatomical regions. Functional characterization of these devices was performed both in vitro and in vivo. In particular, we demonstrated the feasibility of long-term delivery of a model drug, using a previously described fat graft retention model^6,7.

Materials and Methods

Preparation of Silk Solutions

Silk fibroin solution was prepared as previously reported.² Briefly, silk fibroin protein was extracted from B. mori cocoons by boiling in a 0.02 M sodium carbonate solution for 10 or 30 minutes (hereafter referred to as 10 or 30 “minutes-extracted” respectively) to remove sericin. The extracted silk fibroin was washed and dried for 24 hours in a chemical hood before being dissolved in 9.3 M LiBr solution at 60°C for 4 hours, at a 20% w/v ratio. This solution was dialyzed against distilled water using Pierce Slide-a-Lyzer cassettes, MWCO 3500 Da (Rockford, IL) for 3 days to remove LiBr. The solution was then centrifuged to remove aggregates that formed during purification. The final concentration of aqueous silk fibroin (hereafter referred to as silk) was ~6–8% w/v. Solutions were stored at 2–5°C until use.

Preparation of Silk Tubes

Porous silk catheters were prepared via two different methods: gel-spinning and dip-coating.^8,9 For the gel-spinning process, 10 minutes-extracted silk fibroin was concentrated to about 15–18% by centrifugal evaporation (CentriVap, Labconco, Kansas City, MO), according to previously published protocols^8,10. Tubular scaffolds were produced by spinning the concentrated silk solutions onto a rotating (200 rpm) and axially reciprocating Teflon-coated wire (outer diameter: 400 μm) with an axial slew rate of 1 mm/sec as previously described.⁸ In some experiments, a stainless steel flow modulator tube (outer diameter: 800 μm) was sheathed around the Teflon wire to create a larger inner lumen fitting on one end of the tube. Concentrated silk solution was then spun around the interface between the Teflon rod and stainless steel fitting. After spinning, the silk tubes were either incubated at room temperature for 10 minutes, or flash-frozen on a dry ice bed, before being transferred to the freeze-drier chamber with the temperature pre-set to −45°C and incubated for 12 hours before putting a vacuum to 100 mTorr for lyophilization. After completion, tubes were treated with methanol for 2 hours and removed from the wires.

For the dip-coating process, the 30 minutes-extracted silk solution was further concentrated to ~18–20 w/v% by centrifugal evaporation, according to previously described methods^9,11. Half of the silk solution was blended with polyethylene oxide (PEO, MW = 900,000 Da; Sigma-Aldrich, St. Louis, MO) at a 80:20 silk:PEO dry weight ratio. Teflon-coated rods (outer diameter: 400 μml length: ~5 cm) were individually dipped into silk solutions with or without PEO to create catheters with gradient porosity. To create non-porous walls, one end of a Teflon rod was dipped into the concentrated silk solution slowly, being careful to create even coatings. The tube was immersed in 90% methanol for 5 minutes to crystallize the layer and allowed to dry at 60°C for 10 minutes, then another layer was coated onto the same end of the Teflon rod. After five complete layers, rods with silk coating were stored in 90% methanol for 12 hours over night, then dried at 60°C for one hour before beginning application of the porous layers. Rods half coated with non-porous silk layers were then dipped into silk-PEO solutions. The non-coated end was slowly dipped into blended solution to ensure even coating and deep enough to allow overlapping of silk-PEO solution over the previously cast silk only layers (overlap approximately 2–3 mm in length). After deposition of one silk-PEO layer, rods were immersed in 90% methanol for 5 minutes and allowed to dry at 60°C for 10 minutes, then another layer was coated onto the same end of the Teflon rod. After five complete layers, rods were stored in 90% methanol for 12 hours over night, and then transferred to deionized water for 48 hours to allow complete leaching of the PEO.

Silk catheter porosity measurements

Silk catheter porosity and pore sizes were measured on 10 μm thin cryosections. Briefly, the silk catheters were embedded in OCT and frozen at −80°C. Several 10 μm thin sections were collected at different heights along the catheters, for at least 5 catheters per group. Brightfield images were acquired and converted in binary on ImageJ. Particle analysis was performed to calculate the pore area. At least 15 images were analyzed for each group.

Scanning Electron Microscopy (SEM)

SEM was used to evaluate the tube wall morphology of silk catheters. Silk protein tubes were imaged using a Zeiss EVO MA10 electron microscope (Carl Zeiss AG, Germany). Samples were cut perpendicular to the longest axis, mounted onto copper tape, and sputtered with a 10 nm layer of gold before imaging (15mA for 90 seconds sputtering).

Two-compartment Test Chamber and Dye Release Assay

A two-compartment testing chamber was 3D printed to assess the diffusional properties of the gradient porosity silk catheters. The chambers were printed with acrylonitrile butadiene styrene (ABS) plastic. Chamber dimensions were 56 mm length × 20 mm wide × 20 mm depth, with two chambers each 26 mm length × 16 mm wide × 18 mm depth, separated by a 2 mm thick wall with a 1.5 mm diameter channel extending through the bifurcation. The non-porous ends of the dip-coated tubes were inserted into the opening of an ALZET (Cupertino, CA) osmotic pump, prefilled with 100 μL of Reactive Red 120 (Sigma-Aldrich, St. Louis, MO) solution. Tubes were inserted through the aperture on the impermeable septum separating the two compartments, such that the overlapping porous/non-porous region was centered in the middle of the aperture. Silicone caulking agent was used to seal the edges of the bifurcation channel to prevent flow between the chambers. Each chamber was filled with 5 mL of 1xPBS solution. Chambers and tubes were stored at room temperature on a shaker plate set at the lowest setting to induce agitation. Dye release was measured at daily time points over two weeks using a spectrophotometric analysis. Absorbance measurements at 540 nm wavelength light were recorded as a means of quantifying the release of dye over time.

Implantation of Catheter Delivery System In Vivo

Mouse experiments were conducted under protocols approved by the Pittsburgh University Institutional Animal Care and Use Committee (protocol ID:16047853). ALZET pumps were loaded with either 100 μl dexamethasone (Sigma-Aldrich, St. Louis, MO) solubilized in 1xPBS at a concentration of 100 ng/cc (N = 10), or with 100 μl 1xPBS alone (N = 13). Silk catheters were affixed to the pumps and the delivery systems were then submerged in 1xPBS and incubated for 48 hours at 37°C, per manufacturer instructions. After informed consent, human lipoaspirate was collected under an IRB-approved exemption (IRB No. 0511186, University of Pittsburgh) from three healthy, de-identified adult female patient donors. Subcutaneous adipose tissue was harvested during abdominoplasty at the University of Pittsburgh Medical Center, where all samples were waste materials collected as a byproduct of surgery. Residual oils and tumescent fluids were removed upon centrifugation-induced separation at 3000 RPM for 3 minutes. Lipoaspirate was grafted into athymic nude mice, a common animal strain used to model soft tissue reconstruction as it is capable of simulating subcutaneous engraftment to assess volume retention supported by the drug delivery system.^12–14 Mice were anesthetized with 2.5% isoflurane (Sigma-Aldrich, St. Louis, MO) prior to surgeries. The delivery systems were implanted subcutaneously in athymic nude-Foxn1^nu mice (Envigo, East Millstone, NJ), accessing the space via incision at the nape of the neck, catheter facing distally once inserted. Subsequently, 1cc lipoaspirate was injected using 16G blunt-tip cannulae surrounding the silk catheter. A third cohort of mice received 1ml injections of lipoaspirate only (N = 13). Animals were allowed food and water ad libitum until sacrifice using CO₂ asphyxiation at Day 28. Adherence to the regulations and standards set forth by University of Pittsburgh and National Institutes of Health Office for Animal Care and Use were strictly followed throughout. Grafts and delivery systems were recovered, followed by volumetric analysis and histology performed on the explanted adipose tissue grafts. Skin was removed from each of the grafts recovered and the resulting volume was measured via gas pycnometer (AccuPyc II, Micromeritics, Norcross, GA) as previously described ¹⁵. H&E staining (Sigma-Aldrich, St. Louis, MO) was performed on paraffin-embedded sections, which were fixed in formalin for 8 days prior to processing and sectioning.

In vivo durability assessment

In vivo durability experiments were conducted under protocols approved by the Tufts Institutional Animal Care and Use Committee (M2016-157). Porous silk catheters were fabricated as described above and cut into 5 mm sections. The silk catheters were implanted on the back of 4–6 weeks old, female Balb/C mice (Charles River Laboratories, Wilmington, MA) subcutaneously, after anesthetizing the animals with 2.5% isoflurane. Briefly, a 1 mm incision was practiced in the skin to create a small subcutaneous pocket; the silk catheters were inserted through the incision and stitches were applied to seal the wound and maintain the constructs in position. Three constructs were implanted for each animal, for a total of 12 animals. Mice were euthanized by CO₂ asphyxiation after 2, 4, 8, 12 weeks (3 animals per time-point). The recovered grafts were fixed in formalin for 24 hours and embedded in OCT media (VWR, Radnor, PA) for cryosectioning. H&E or Masson’s Trichrome staining (Sigma-Aldrich, St. Louis, MO) was performed on the samples to assess cell infiltration. In some experiments, the sectioned samples were stained with DAPI, to visualize cell nuclei. Quantification of the average capsule thickness was performed by ImageJ on at least 5 representative images for each time point, by measuring the total capsule area for each section and reporting the average radius.

Results

Gel-Spun Silk Catheters

Gel-spun tubes were designed to have tunable porosity, in order to control drug diffusion. Figure 1 shows how the porosity may be tuned based on the freezing procedure. Panels A and B, show silk tubes obtained by incubating the spun solution at room temperature for 10 minutes before the freezing step. The silk tubes obtained through this process showed reduced porosity and a continuous structure on the outer surface. We assume that prolonged exposure to room temperatures allowed water exclusion from the silk solution due to the gelation process, and subsequent evaporation. Tube layers are relatively thin (~ 200–300 μm thickness) with high surface area; therefore, water evaporation may be occurring rapidly after deposition. Removal of water may limit ice crystal formation and thus reduce porosity.

Figure 1.

Highly porous silk tubes fabricated through a novel flash-freezing spinning technique. Tubes made by conventional gel-spinning have small pores on the inside of the material, but the outer surface generally lacks porosity (A and B). Tubes made from gel-spun and flash-frozen silk maintain their outer porous morphology (C and D).

Alternatively, gel-spun tubes were flash-frozen after spinning over a dry-ice bed. This method is assumed to limit the protein-water separation and evaporation effects observed in the previous gel-spinning method. In Figure 1, panels C and D show a cross section and outer wall morphology of tubes with very high porosity. Quantification of the pore dimensions showed a broad distribution (range 3.88–141 μm), with mean pore diameter of 12.82±0.13 μm.

The gel-spinning technique allowed us to fabricate silk catheters with modular diameter. As shown in Figure 2(A), a Teflon rod (green) was fit inside the stainless steel flow modulator of the pump. Silk solution was spun around both the flow modulator and Teflon rod, creating a silk tube with a larger inner lumen adapter at one end, which can accommodate quick attachment of the osmotic pump after surgical insertion of the silk tube. This would be ideal for interfacing the catheter with a pump or an injection port external to the augmented tissue.

Figure 2.

Interfacing silk gel spun tubes with ALZET osmotic pumps. A) Schematic showing the assembly of the flow modulator with a Teflon rod. Silk gel was then spun over the interface of both pieces, creating a single, intact silk tube with different inner diameters at each end. Teflon rod OD = 400 μm; flow modulator OD = 800 μm. B and C) Images of the finished assembly. After tubes are gel-spun and processed, the silk tube can quickly disconnect from the osmotic pump. The flow modulator can be reinserted into the silk tube at a later time and attached via suture or adhesive for delivery of therapeutics from the pump.

Dip-Coated, Gradient Porosity Silk Catheters

Silk dip-coating allowed rapid fabrication of flexible, mechanically robust tubes with multiple material phases to accommodate the design of gradient porosity across the length of the tube. Figure 3 depicts a schematic for the design of gradient porosity tubes. These catheters consist of two phases: a proximal, non-porous dense silk phase, and a distal, highly porous silk phase. A shown in the schematics, the proximal end of the catheter could be directly interfaced with the inlet of the ALZET osmotic pumps, due to their higher mechanical strength.

Figure 3.

Schematic of gradient porosity silk catheters. Gradient porosity tubes consist of two phases: a proximal, non-porous dense silk phase, and a distal, highly porous silk phase. As shown, the non-porous end of the catheter may directly interface with the osmotic pump without requiring a flow modulator. The porous end permits the release of cargo therapeutic. As shown, release is directed away from the pump, limiting exposure of drug to tissues immediately surround the pump and non-porous tube end. In practice, the porous end of the tube would be implanted into the bulk of the regenerating tissue, while the pump and non-porous end would exist external to the healing tissue.

SEM analysis of the surface morphology for the two phases and the interface of dip-coated silk catheters shows unique surface topography for each format (Figure 4). The non-porous phase (red) has a very smooth profile, with minimal porosity and roughness. This topography is ideal for interfacing with the osmotic pump since it should permit low friction and ease of insertion to the pump inlet. Furthermore, low surface porosity should reduce diffusion through the material. Alternatively, the porous phase (blue) exhibits pores distributed uniformly along the length of the tube. Compared to the gel-spun catheters, the dip-coated method yielded smaller pores (diameter range 0.87–34.4 μm), with a mean diameter of 4.64±0.46 μm. It is expected that this phase should rapidly permit diffusion of material through the walls and into the surrounding environment. The interface (green) shows the transition between the phases, suggesting that there is no obvious weak point at the phase junction. As depicted in Figure 5 (top panel), dip-coated tubes can be fabricated as a single, porous phase, or as a two-phase construct, where the length ratio of porous to non-porous morphology can be modulated. In the latter case, the length of the non-porous phase may be adjusted to allow diffusion away from the pump, to different depths within the augmented tissue. This is demonstrated in Figure 5 (bottom panel), where the non-porous phase is optically clear, and the flow of dye of can be observed macroscopically. It appears that the walls of the non-porous tube are devoid of dye, indicating there is no immediate dye diffusion through the non-porous material. However, in the porous region, there appears to be immediate diffusion and absorption of dye through the catheter wall. After two weeks of incubation in PBS, we observed that the walls of the non-porous tube have been stained red, suggesting that the dye can diffuse through the non-porous layer, at a much slower diffusion rate compared to the porous phase. Dip-coated silk tubes where chosen for all the further characterization, due to their ease of fabrication and robust mechanical properties.

Figure 4.

Scanning electron microscopy of the surface morphology for the two phases and the interface of dip-coated silk catheters. The non-porous phase (red) has a very smooth profile, while the porous phase (blue) exhibits a rough morphology with pores distributed uniformly along the length of the tube. The interface (green) shows a very smooth transition between the phases, suggesting that there is no obvious weak point at the phase junction.

Figure 5.

Phase ratios of dip-coated tubes can be controlled to tune diffusion distance from the osmotic pump. Top panel: dip coated tubes may consist of only a single, porous phase, or have a two-phase design where the length ratio of porous to non-porous morphology can be changed. Bottom panel: the non-porous phase is optically clear, and the flow of dye of can be observed macroscopically. It was observed that the walls of the non-porous tube are devoid of dye, indicating that there is no immediate diffusion into this region. The porous phase does show rapid uptake of dye into the bulk of the material. After two weeks of incubation in PBS, the walls of the non-porous tube were stained red, suggesting that while diffusion may still occur in this phase, the rate of diffusion is much slower compared to the porous phase.

Two-compartment chamber dye-release assay

To further quantify the release properties of the two-phase system, we connected the catheter to an ALZET osmotic pump loaded with a red dye. To quantify the release of dye from the silk catheters, a two-compartment test chamber was designed (Figure 6A). These two compartments housed either the pump and non-porous tube end, or the porous tube end alone. The connecting aperture was sealed with a silicon caulking agent to prevent leakage between the compartments. We observed that dye was predominantly released from the porous end alone accumulating in the distal chamber, with minimal release near the pump (Figure 6B). Dye release was quantified via spectrophotometry by measuring the absorbance at 520 nm within each well at designated time intervals.

Figure 6.

Bifurcated chamber to measure release kinetics from gradient porosity catheters (A). The chambers containing the non-porous tube regions showed limited dye presence, while the porous tube section had continuous dye release for up to two weeks (B). Dye release was analyzed by measuring absorbance of the fluid within each chamber at 520nm wavelength (C). *p<0.05, **p<0.01.

By day 10, we observed only a slight increase in the absorbance measurements in the chamber containing the non-porous catheter end (Figure 6C), confirming that silk tubes with gradient porosity can direct release of solution away from the pump, targeting the tissue immediately surrounding the porous tube region while limiting exposure to the tissues in proximity to the osmotic pump.

In Vivo Implantation of Drug Delivery System

The potential of this delivery system was further explored in vivo in a fat graft retention model. Previous studies using the athymic mouse as a model for engraftment of human lipoaspirate have identified dexamethasone as a successful therapeutic for increasing volume retention.^6,7 Dexamethasone has been shown to be a potent regulator of adipogenesis in adipose-derived stem cells (ASCs),^16–18 and given that adipose tissue processed into lipoaspirate preserves the native population of adipose-resident ASCs,^19,20 the influence toward adipogenic differentiation may act as a mechanism for volume retention. The delivery system was implanted subcutaneously in athymic mice, at the nape of the neck such that the catheter was oriented toward the tail. Human lipoaspirate was grafted into the subcutaneous space surrounding the catheter. In two cohorts, the ALZET pump was loaded with either dexamethasone (100 ng/ml, solubilized in 1xPBS, N=10) or 1xPBS alone (N=13); the dexamethasone concentration of 100 ng/ml was based on the concentration of the steroid in adipogenic media.²¹ A third cohort received lipoaspirate only as a control condition (N=13). The in-life duration of this study was limited by the pump elution capability, where the maximum loading capacity of the pump is 100 μl, eluting continuously over a 28-day period. The customization of porous and non-porous segments allowed elution of aqueous medium from the pump through the catheter, and directly into the grafted tissue of interest, limiting drug loss at the catheter-pump interface. Figure 7A displays the position of the delivery system and lipoaspirate in the athymic nude mouse immediately after sacrifice at 28 days. It was noted that minimal graft migration occurred in fat grafts containing the delivery systems as shown Figure 7A, where the injected adipose remained atop the catheter, adjacent to the ALZET pump, as the location of initial injection. As expected upon explant, a small fibrous capsule surrounded the human lipoaspirate as well as the ALZET pump.

Figure 7.

Athymic nude mice representing the adipose retention over the 28-day study, with the delivery system show implant subcutaneously, with the silk catheter embedded into the grafted lipoaspirate (A). Volume retention amongst groups is compared. N.S. denotes no statistical significance amongst conditions from a Fisher’s Least Significant Difference post-hoc ANOVA (p = 0.371, α = 0.05) (B). H&E staining revealed morphology at 4x and 20x magnifications typical of grafted adipose with few oil cysts shown in all conditions. Top panels, scale bar: 500 μm; bottom panels, scale bar 100 μm (C). SEM micrographs showing the structure of explanted catheters after 28 days in vivo infusion, at 200 X magnification. White boxes show the magnified area (D).

Figure 7B displays the volume retention after the 28-day period, showing a trend of increased volume retention favoring the dexamethasone condition. Upon statistical analysis, it was shown however, that statistical distinction between groups couldn’t be made (Fisher’s Least Significant Difference post-hoc ANOVA, p = 0.371). Histological analysis revealed highly consistent morphology across conditions suggesting that the drug delivery system was of no detriment to the grafted lipoaspirate. Some oil cysts were present in all conditions, which is typical of grafted adipose tissue (Figure 7C).²² A Possible limitation of this investigation could be inadequate dexamethasone concentration or study duration to appreciate a significant effect. Future studies will seek to optimize the dexamethasone concentration and length of therapy such that maximal and consistent volume retention can be determined.

The explanted catheters were fixed and inspected by scanning electron microscopy for signs of degradation. The SEM micrographs in Figure 7D showed that the catheters maintained structural integrity and a smooth surface, with limited erosion observed in both the dexamethasone and vehicle (PBS) eluting catheters.

Silk Catheter in-vivo durability

Tissue macrophages and immune cells play a fundamental role in the degradation of implanted constructs. For this reason, we tested the durability of our silk catheters in an immunocompetent mouse model over a 12-week period. The silk catheters were implanted in the back of Balb/C mice in subcutaneous pockets. Mice were euthanized after 2, 4, 8 and 12 weeks, and catheters were explanted to assess cell infiltration and degradation. As shown in Figure 8, the implanted tubes engrafted in the host tissue, without creating evident tissue inflammation, even though histological examination revealed formation of a fibrotic, collagen rich capsule around the constructs (Figure 8B). Quantification of the capsule thickness at different time points showed morphological evolution as expected. In particular, after 2 weeks of implantation, the catheters were surrounded by a collagen layer, with granulation tissue still present (capsule thickness: 88±4.5μm). At 4 weeks, the collagen was reorganized in a thin, dense layer with a thickness of 51±8.5μm. After 8 and 12 weeks the capsule became thicker, possibly due to the recruitment of fibroblasts that deposited more extracellular matrix, with an average thickness of 107±4.2μm and 87±2.9μm, respectively. The catheters did not display any obvious sign of degradation both macroscopically and microscopically. Indeed, histological analysis showed very limited cell infiltration even after 12 weeks, as further demonstrated by staining the construct with the nuclear stain DAPI (Figure 8C). These data demonstrate that these catheters can maintain shape and mechanical integrity for extended period of time, which will allow prolonged release of therapeutic molecules.

Figure 8.

Macroscopic appearance of subcutaneously implanted silk catheters after 2, 4, 8 weeks from implantation (top panels, scale bar 5 mm) and correspondent H&E staining (bottom panels, scale bar: 500 μm) (A). Macroscopic appearance of implanted silk catheters after 12 weeks (top panel, scale bar 5 mm) and correspondent Masson’s trichrome staining (bottom panel, scale bar: 500 μm) (B). Brightfield and fluorescent images of DAPI stained catheters after 12 weeks from implantation, scale bar: 500 μm (C).

Discussion

This work describes the fabrication and application of a biodegradable, silk-based catheter with gradient porosity for drug diffusion into tissues. Silk tubes can be manufactured via two different techniques, both yielding materials with tunable porosity, aiding in target drug delivery. In this preliminary study, the dip-coated silk catheters were chosen for in vivo validation because of their ease of fabrication. Nevertheless, gel-spinning represents a valuable alternative for silk catheter fabrication, which may allow a more precise control of the silk tube structure and better reproducibility. We showed that the catheters can be interfaced with a pump, and the drug release profile can be controlled by tuning the porosity gradient, thus allowing targeted release. Local drug delivery through medicated catheters has been successfully used to treat peripheral and coronary artery disease conditions, by using drug-coated balloon catheters to prevent restenosis.¹ We propose a novel delivery method that can be exploited in several different applications where targeted drug delivery is necessary. The advantage of using silk biomaterial is its well-established biocompatibility, mechanical robustness, and controlled degradability into non-toxic byproducts. The latter characteristic is pivotal for this application, allowing the catheter to be left in place and slowly degraded by immune cells and tissue enzymes, without requiring invasive surgical removal. Under the time frames of the present in vivo study, we did not pursue specific degradation assessments, as the goal was device design. Based on our extensive prior studies of in vivo silk degradation, we anticipate that at least 6 months would be required to demonstrate significant loss of volume in the current catheter design. This assumption is based on studies of porous silk sponges in vivo with lipoaspirate, which required 1–2 years for full degradation in rats.^23–26 Silk degradation can be finely tuned in many ways: controlling the formation of β-sheet secondary structures by water annealing, for instance, has been proven to impact degradation rate of various formats of silk scaffolds like films and sponges.^3,4 Another valuable alternative could be to dope the silk protein with rapidly degrading matrix components, such as hyaluronic acid or fibrin, which are approved by FDA for medical use and tissue augmentation. Indeed, we previously showed that the addition of hyaluronic acid and fibrin increased the degradation rate of silk hydrogels in-vitro. ^27,28 This was probably due to the initial degradation of the hyaluronic acid and fibrin components and disruption of the silk polymer network, which, in turn, became more exposed to proteolytic degradation.

We showed that the proposed system could be applied to improve the outcome of soft tissue replacement strategies after traumatic injuries. Autologous fat grafting is a common procedure to restore aesthetic and structural functions after traumas. This technique does not pose risks in terms of immune response and graft rejection, but long-term graft volume retention is extremely variable among subjects, due to graft necrosis, degradation and reabsorption. Delivery of adipogenic factors such as dexamethasone has improved volume retention of grafted adipose tissue and allowed long-term functional defect restoration.^6,7

In this manuscript, we used an athymic mouse model for human lipoaspirate engraftment, to demonstrate the possible application of our catheter system for localized delivery of dexamethasone. Despite our results not showing statistically significant differences between the dexamethasone treated mice and the control groups, we demonstrated that the catheters were able to sustain drug delivery for the duration of the experiment without losing structural integrity, making them suitable for long-term delivery. This study constitutes a proof-of-concept, where further investigations involving higher dosage, higher animal numbers, and extended elution time with refillable pumps will be required to prove the effectiveness of our strategy in vivo.