Replenishable prevascularized cell encapsulation devices increase graft survival and function in the subcutaneous space

Abstract Beta cell replacement therapy (BCRT) for patients with type 1 diabetes (T1D) improves blood glucose regulation by replenishing the endogenous beta cells destroyed by autoimmune attack. Several limitations, including immune isolation, prevent this therapy from reaching its full potential. Cell encapsulation devices used for BCRT provide a protective physical barrier for insulin‐producing beta cells, thereby protecting transplanted cells from immune attack. However, poor device engraftment posttransplantation leads to nutrient deprivation and hypoxia, causing metabolic strain on transplanted beta cells. Prevascularization of encapsulation devices at the transplantation site can help establish a host vascular network around the implant, increasing solute transport to the encapsulated cells. Here, we present a replenishable prevascularized implantation methodology (RPVIM) that allows for the vascular integration of replenishable encapsulation devices in the subcutaneous space. Empty encapsulation devices were vascularized for 14 days, after which insulin‐producing cells were inserted without disrupting the surrounding vasculature. The RPVIM devices were compared with nonprevascularized devices (Standard Implantation Methodology [SIM]) and previously established prevascularized devices (Standard Prevascularization Implantation Methodology [SPVIM]). Results show that over 75% of RPVIM devices containing stem cell‐derived insulin‐producing beta cell clusters showed a signal after 28 days of implantation in subcutaneous space. Notably, not only was the percent of RPVIM devices showing signal significantly greater than SIM and SPVIM devices, but the intraperitoneal glucose tolerance tests and histological analyses showed that encapsulated stem‐cell derived insulin‐producing beta cell clusters retained their function in the RPVIM devices, which is crucial for the successful management of T1D.

showed a signal after 28 days of implantation in subcutaneous space. Notably, not only was the percent of RPVIM devices showing signal significantly greater than SIM and SPVIM devices, but the intraperitoneal glucose tolerance tests and histological analyses showed that encapsulated stem-cell derived insulin-producing beta cell clusters retained their function in the RPVIM devices, which is crucial for the successful management of T1D.
K E Y W O R D S beta cell replacement therapy, cell encapsulation device, human stem cells, prevascularization, transplantation in subcutaneous space, type 1 diabetes 1 | INTRODUCTION Type 1 diabetes (T1D) is an autoimmune disease in which the immune system attacks insulin-producing beta cells in the pancreas. Loss of beta cells results in an inability to produce and secrete adequate levels of insulin, especially in response to changes in peripheral glucose levels, which can lead to hyperglycemia and several long-term complications. [1][2][3] People with T1D must frequently monitor blood glucose levels and require exogenous insulin delivered through pumps or injections. While this strategy helps manage T1D, under and overdosing on exogenous insulin may lead to hyperglycemia and hypoglycemia, respectively, leading to long-term morbidity. A longer-term solution for treating T1D is islet transplantation, which replaces the destroyed cells with functional beta cells. 1,2,[4][5][6] This therapy can recapitulate endogenous beta cells' intricate glucose sensing and insulinreleasing capabilities while eliminating the burdens of patient compliance and dependence on exogenous insulin. A drawback to this therapy is that patients are prescribed long-term immunosuppressive drugs to prevent immune rejection of transplanted cells, which can be fatal. Cell encapsulation using macroencapsulation devices have shown to be a promising approach to address this challenge. 1,2,[7][8][9][10][11][12][13] Encapsulation devices offer a physical barrier to protect transplanted beta cells from immune attack; however, this barrier can limit the diffusion of oxygen and nutrients, causing ischemic stress detrimental to graft survival. This ischemic stress varies based on the transplantation site of the device. 1,2,4,[7][8][9][10]14,15 An ideal transplantation site for cell encapsulation devices includes (1) a dense vascular network that allows for insulin and glucose exchange, along with high oxygen and nutrient supply to the graft; (2) a hospitable microenvironment that prevents initial loss of cells posttransplant; (3) a minimally invasive procedure for implanting, monitoring, and retrieving the graft. Transplantation in the subcutaneous space allows for minimally invasive implantation and retrieval. 5,11,12,[16][17][18][19][20][21][22][23] However, a significant challenge that remains unaddressed in subcutaneous cell transplantation is the loss of beta cell viability that occurs shortly after transplantation. During surgical implantation, the inherently low vasculature in the subcutaneous space is further destroyed, leading to an even lower supply of oxygen and nutrients at the transplantation site. This is particularly detrimental to highly metabolic beta cells. 1,2,7,16,17,22,24,25 Several studies have shown that subcutaneous transplantation sites can be modified to promote neovascularization posttransplantation. These methods involve the use of biologics such as growth factors (beta-fibroblast growth factor [FGF], vascular endothelial growth factor), [26][27][28] chemical modifications of the encapsulation material, 29 anti-inflammatory drugs, 14,30,31 co-delivery of mesenchymal stem cells, [32][33][34] and the use of oxygen generators, 10,35-37 among others.
While these strategies may improve blood vessel formation, these methods require at least 10 days to create a dense vascular network that is well integrated with the host vasculature after cells have already been implanted in the subcutaneous space. Due to this limitation, these methods cannot rescue most of the graft loss that occurs within the initial days of implantation. 7,11,12,15,21,38 Another promising strategy to improve islet vascularization is the prevascularization of encapsulation devices at the transplantation site.
In this approach, a non-vascularized device is implanted before transplanting beta cells in the vascularized site. The advantage of this approach is that the host vasculature is well incorporated at the transplantation site, improving access to oxygen and nutrient supplies. 7,11,20,21,32,39,40 However, a drawback of this strategy is that the vascularized device is typically removed prior to beta cell transplantation. Device removal can rupture some newly formed vascular networks and create a suboptimal microenvironment for subsequent islet transplantation. Several groups have sought to overcome this drawback; however, these approaches involve in vitro prevascularized devices or more complex encapsulation devices that require different membranes, drugs, and surface modification to promote in vivo vascular growth. 7,41-43 To our knowledge, there are currently no strategies for encapsulation devices that create a prevascularized transplantation site that allows for the direct insertion of cells without disrupting the surrounding vasculature.
Here we report a prevascularization strategy using a replenishable encapsulation device that prevents the initial loss of cell viability and function of stem cell-derived insulin-producing beta cell clusters in the subcutaneous space. Our method was designed to prepare the transplantation site such that a functional vascular network surrounds the encapsulation device prior to the transplantation of cells. This approach was developed using thin-film polycaprolactone (PCL) cell encapsulation devices that have been previously shown to maintain the viability and function of insulin-producing cells in the liver lobe for at least 6 months. 44 We have also demonstrated that such devices support the viability of insulin-producing cells in the subcutaneous space by incorporating nutrient depots. 24 Devices were implanted subcutaneously for 2 weeks, after which, they were loaded with stem cell-derived insulin-producing beta cell clusters without disrupting the integrity of the surrounding vascular network. After 28 days, more than 80% of prevascularized replenishable devices showed signal, leading to a measurable C-peptide secretion in response to a glucose challenge.

| Design of subcutaneous thin film PCL device and implantation technique
To create a refillable device, we modified the fabrication of thin film devices described in previous work. 24,45,46 We fabricated a small device that was 2 cm in length and width to permit the insertion of the device in the subcutaneous space with minimal stress to the animal. In brief, the new design included a 1 cm long and 0.6 cm wide neck with an enclosed catheter that allows for a facile opening of the device and the insertion of cells (Figure 1). The long neck was crucial in opening the device and inserting cells without removing the device from the transplant site and disrupting the surrounding vasculature. After inserting the cells, a cauterizer was used to seal the device opening through resistive heating. The circular region, where the cells reside, has a diameter of 1.6 cm and maximum volume capacity of 160 μL. To provide mechanical support, we also incorporated a 100 μm thick PCL border around the device, which keeps it sturdy and prevents it from folding over in vivo. Additionally, transplantation in the subcutaneous space allowed for multiple surgeries where no adverse side effects were observed (based on whole animal and gross site observation). This may not be possible with other implantation sites, where repeated administration may lead to greater adverse events. 16,17,19 Once encapsulation devices were assembled, we next determined the ability of devices to maintain cellular viability and function, with and without prevascularization and/or device removal. Devices were either loaded with cells or immediately transplanted into an unmodified subcutaneous space (Standard Implantation Method [SIM]), or were first implanted as an empty device to establish vasculature before loading with cells (Refillable Prevascularized Implantation Method, RPVIM). For the prevascularized devices, an empty device was allowed to vascularize for 14 days. After that, stem cell-derived insulin-producing beta cell clusters were inserted, and devices were sealed without disrupting the newly formed vasculature around the device. To demonstrate that maintaining an intact vascular network in RPVIM is critical for graft survival, we also compared with a previously reported approach where prevascularized empty devices were explanted and replaced with a SIM device (Standard Prevascularization Method [SPVIM]). We hypothesized that disruption of the prevascularized zone around the device would limit its ability to preserve graft survival in the subcutaneous space.

| Survival of stem cell-derived insulinproducing beta cell clusters in prevascularized replenishable encapsulation devices
We next asked whether prevascularization, with or without removal of the device, affects the survival of stem cell-derived insulinproducing beta cell clusters. These luciferase-labeled cell clusters were generated using a previously established protocol. 47 Moreover, this same trend was observed with the d28 cells, which are believed to be more susceptive to ischemia due to higher metabolic activity. 46,47 The change in bioluminescence compared with the baseline for cells in RPVIM was $60% after

| Function of mature insulin-producing cells within prevascularized replenishable encapsulation devices
Given the viability data, we hypothesized that RPVIM devices would show greater glucose-stimulated insulin secretion than SIM or SPVIM devices. Since the d20 cells are immature and do not respond significantly to glucose challenge, we conducted functional tests on d28 cells that more closely resemble mature betacell-like clusters and demonstrate glucose-stimulate insulin secretion. 47,48 To assess the function of the encapsulated cells, after F I G U R E 2 Vasculature formation around empty implanted devices. (a) Lectin-perfusion assay (stained in purple) was performed to visualize functional vasculature after 7 (n = 6), 14 (n = 6), and 28 days (n = 6) of implantation. (b) Representative images of vascular networks (detected using lectin-perfusion assay) used to quantify changes in (c) total vascular area, (d) number of nodes, and (e) number of branches around the implant. The significance across all experimental groups was performed using One-way ANOVA, followed by Tukey's post hoc test.
F I G U R E 3 Legend on next page.

| DISCUSSION
Our results confirm other studies reporting that implantation within the subcutaneous space does not allow for the survival of beta cells, which are typically highly vascularized in their native environment. 5,19,40,48,49 Low oxygen tension and lack of nutrient supply caused by insufficient vasculature often lead to ischemia and necrosis of the highly metabolic beta cells. 2,4,24,25,39,50,51 It has also been shown that beta cell function and insulin secretion are severely impacted by the dense vascular network through blood flowdependent and independent pathways. 5,38,49 Therefore, it is crucial for encapsulated beta cells to have functional and robust vasculature around the implant.
Here, we demonstrate an easy-to-implement prevascularization approach that aims to relieve the ischemic stress experienced by encapsulated cells upon implantation in immunodeficient NSG mice.
Our results indicate that in the subcutaneous space, prevascularization of replenishable devices increases the survival of encapsulated stem cell-derived insulin-producing beta cell clusters posttransplantation. The nanoporous thin-film encapsulation device was designed such that after prevascularization, cells could be loaded easily through the port without the need to remove the device from the transplant site. After the cells were loaded, the device was resealed using resistive heating. This technique maintains the device's shape and structure while also preserving the surrounding vasculature ( Figure 1).
Additionally, previously published research suggested that a functional, planar encapsulation device should be no more than 550 μm in thickness with a volume fraction of $2.5%. 52 In our encapsulation devices, the total volume of transplanted beta cell cluster mass was 4.423 μL, which results in a volume fraction of $2.7%. Also, we estimated that these loaded encapsulation devices had a thickness of $100 μm, demonstrating that our device design satisfies the optimized parameters. Analysis of the vasculature around the encapsulation device showed that after 14 days of prevascularization in the subcutaneous space, the device is surrounded with twice as many new, functional blood vessels compared with 7 days of implantation ( Figure 2). The presence of these blood vessels plays an important role in cell survival and function inside the encapsulation device, as we observed greater than 80% graft survival in RPVIM devices in the subcutaneous space (Figure 3). The RPVIM devices also show greater performance than SPVIM devices, a technique used previously by several groups to enhance cell engraftment. 7,20,32,39,42 This was also observed in more mature stem cell-derived insulin-producing beta cell clusters, which are highly metabolic and require robust sources of nutrients and oxygen. 46,47 Moreover, the mature cells in prevascularized grafts also demonstrate significantly higher levels of glucose-stimulated insulin secretion compared with controls ( Figure 4). The histological analysis demonstrates in vivo biocompatibility and the presence of Quantification of bioluminescence signal from cells transplanted in devices compared with baseline for (B) d20 and (e) d28 cells. The significance of changes in bioluminescent signal at Day 28 versus baseline was determined using multiple unpaired t-tests, corrected for multiple comparisons using Holm-Sidak method. Quantification of the percent of (c) d20 and (f) d28 grafts showing bioluminescence over a period of 28 days. The significance between survival curves was determined using the Kaplan-Meier test, and comparisons were made using a Log-rank (Mantel-Cox) method.

F I G U R E 4 Glucose response and insulin secretion from RPVIM devices containing d28 stem cell-derived insulinproducing cells. (a) Levels of secreted Cpeptide from cells in Refillable
Prevascularized Implantation Method (RPVIM) devices significantly increase 45 min postintraperitoneal glucose injection. The significance between fasting and glucose groups was determined using a one-tailed unpaired ttest. (b) Systemic C-peptide levels in RPVIM devices are greater post-IPGTT compared with Standard Implantation Method (SIM) and Standard Prevascularization Method (SPVIM) devices. Statistical significance across the groups was determined using a two-way ANOVA fitting a mixed-effects model followed by Tukey's post hoc test.
functional stem cell-derived insulin-producing beta cell clusters inside RPVIM devices ( Figure 5). While functional beta cell clusters were observed in the device, some necrotic areas were also present, likely due to insufficient oxygen supply. This is also indicated by the 50% reduction in cluster mass observed in RPVIM devices ( Figure 3). However, overall, the RPVIM approach resulted in a significant increase in vascular coverage and oxygen availability, leading to better preservation of beta cell cluster mass and function compared with controls. Results from this study confirm that the RPVIM devices effectively provide a hospitable microenvironment for encapsulated stem cell-derived insulin-producing beta cell clusters in the subcutaneous space for 28 days. Future studies include longer-term survival and function studies using the RPVIM devices and testing the ability of these devices to reverse diabetes in a diabetic mouse model. These studies will provide key insights into the clinical benefit of this approach. Our approach involves modification of the subcutaneous site to create a highly vascularized implantation site without introducing multiple devices, materials, or biologics whose long-term safety and biocompatibility need to be accounted for. This strategy utilized only FDA-approved biomaterials and took advantage of the naturally occurring vascularization process in the body. 24,44,45 The lectin perfusion studies demonstrated functional vasculature surrounding the implant at 14 days, which we deemed sufficient time to create a favorable transplantation site. We observed robust vascular response with vessel branching, ingrowth, and outgrowth on and around the encapsulation device in the subcutaneous site. Additionally, since neovascularization is a dynamic process, we expect the vasculature to remodel extensively over time, especially with the addition of encapsulated cells. 5,49 Furthermore, our approach showed that not only is a prevascularized site necessary for subcutaneous implants, but it is also crucial to ensure that we preserve the vascular network while transplanting the cells in the prevascularized site.

| Differentiation into pancreatic cells
To initiate differentiation, we dissociated confluent cultures into singlecell suspensions using TrypLE Select, counted them, and seeded them in six-well suspension plates at a density of 5.5 Â 10 6 cells per 5.5 mL of hESC maintenance media supplemented with 10 ng/mL activin A (R&D Systems) and 10 ng/mL heregulinB (PeproTech). The plates were incubated at 37 C and 5% CO 2 on an orbital shaker set at 100 rpm to induce 3D sphere formation. After 24 h, the spheres were differentiated as previously described. 47

| Transplantation
PCL encapsulation devices were implanted in the subcutaneous space via a small incision in the mouse's lower back. The space between the skin and the muscle layer was dissected using blunt forceps, and saline was injected to create an easily accessible pocket. The encapsulation device was implanted in this newly formed space, and the skin wound was closed using surgical staples. For devices containing cells, and nodes was manually calculated for each image. A node was defined as a point at which two vessels intersect, and a branch was defined as vessels that were extending from the main blood vessel.

| Bioluminescent imaging
Graft-bearing animals were injected IP with D-luciferin solution (Goldbio) at 150 mg/kg 30 min before imaging to capture the peak in bioluminescent intensity. Mice were anesthetized with an isoflurane mixture (2% in 98% O 2 ), and the bioluminescent signal was quantified using a Xenogen IVIS 200 imaging system (PerkinElmer). Images were acquired for 1 min and then analyzed using the Living Image analysis software (Xenogen).
Regions of interest (ROIs) were centered over the location of the devices, and background signal was obtained by capturing the ROI of a nonbioluminescent signal. Photons collected over the acquisition time were counted within the ROI. The same imaging protocol was applied each time to ensure consistency across longitudinal studies.

| Intraperitoneal glucose tolerance test
Mice were subjected to an IPGTT at 28 days posttransplantation to assess function of the grafted cells. The test was split into two parts to prevent undue stress on mice. In the first round, mice were fasted overnight, after which the blood glucose levels and blood samples were obtained via tail vein and cheek bleeds, respectively. After 4 days of rest, mice were fasted overnight, and blood glucose levels were measured from the tail vein. Three milligrams per kilogram of glucose was injected into the intraperitoneal space of mice, and after 45 min, blood glucose and blood samples were obtained from the tail vein, and cheek bleeds, respectively. Cpeptide levels were measured using an ultrasensitive insulin ELISA kit (Alpco 80-CPTHU-CH05).

| Statistical analysis
Data were analyzed using GraphPad Prism software version 9.4. All differences in vasculature between experimental groups were evaluated using one-way analysis of variance (ANOVA) or Two-way ANOVA, followed by Tukey's post hoc test or Student's t-test. Graft survival was compared using Kaplan-Meier survival curves. A p < 0.05 was considered statistically significant.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.