Fibroblast growth factor 7 releasing particles enhance islet engraftment and improve metabolic control following islet transplantation in mice with diabetes

Transplantation of islets in type 1 diabetes (T1D) is limited by poor islet engraftment into the liver, with two to three donor pancreases required per recipient. We aimed to condition the liver to enhance islet engraftment to improve long‐term graft function. Diabetic mice received a non‐curative islet transplant (n = 400 islets) via the hepatic portal vein (HPV) with fibroblast growth factor 7‐loaded galactosylated poly(DL‐lactide‐co‐glycolic acid) (FGF7‐GAL‐PLGA) particles; 26‐µm diameter particles specifically targeted the liver, promoting hepatocyte proliferation in short‐term experiments: in mice receiving 0.1‐mg FGF7‐GAL‐PLGA particles (60‐ng FGF7) vs vehicle, cell proliferation was induced specifically in the liver with greater efficacy and specificity than subcutaneous FGF7 (1.25 mg/kg ×2 doses; ~75‐µg FGF7). Numbers of engrafted islets and vascularization were greater in liver sections of mice receiving islets and FGF7‐GAL‐PLGA particles vs mice receiving islets alone, 72 h posttransplant. More mice (six of eight) that received islets and FGF7‐GAL‐PLGA particles normalized blood glucose concentrations by 30‐days posttransplant, versus zero of eight mice receiving islets alone with no evidence of increased proliferation of cells within the liver at this stage and normal liver function tests. This work shows that liver‐targeted FGF7‐GAL‐PLGA particles achieve selective FGF7 delivery to the liver‐promoting islet engraftment to help normalize blood glucose levels with a good safety profile.

between islets and the liver starts to form by day 3. The majority of islet loss occurs within the first 3 days posttransplant. 12 Hypoxia due to a lack of a blood supply and inflammation contributes to islet loss. 13 We hypothesized that a hepatic microenvironment favoring islet retention and vascularization in the early stages posttransplant would ameliorate early islet loss and aid long-term function.
Preconditioning the host liver with growth factors (GFs) creates a receptive "niche" involving the re-modeling and proliferation of liver cells. 14,15 Systemic GFs, such as triiodothyronine (T3), 16,17 hepatocyte growth factor (HGF), 14 and fibroblast growth factor 7 (FGF7), 18 have been used to increase rat liver cell proliferation and enhance the efficiency of retroviral gene delivery. FGF7 is a small polypeptide member of the FGF family that binds to the FGF7 receptor and has proliferative and antiapoptotic effects on epithelial cells including hepatocytes. 18,19 GFs may promote islet engraftment through: (1) liver cell proliferation and immediate islet "trapping"; (2) upregulation of VEGF, promoting vasculogenesis and early islet vascularization; 20,21 (3) anti-inflammatory activity, thereby aiding islet survival; 22,23 and (4) inhibition of T cell-mediated immune effects 24 reducing islet rejection.
Administering GFs systemically is limited by their short half-life, low tissue penetration, and effects on multiple organs. Low GF concentrations in the targeted organ necessitate dose escalation with off-target effects. 25 We tested targeted GF delivery to the liver to promote short-term liver cell proliferation, enhance islet engraftment, and improved metabolic control in a mouse model of T1D.
PLGA polymer is a biodegradable material used in medical devices. The polymer matrix achieves desirable release kinetics based on the polymer hydration profile. 26 For targeted delivery to the liver, the PLGA polymer may have galactose added to it, 27,28 exploiting asialoglycoprotein receptor (ASGPR)-mediated endocytosis. There are ~25 000 ASGPR in the hepatocyte plasma membrane with a specific binding affinity toward the galactose moiety attached on the PLGA particles. 28 Based upon its ability to induce proliferation in the liver following systemic use, a GF was selected for use in an engineered polymer for targeted delivery to the liver.
Our aim was then to create a microenvironment in the liver suitable for early islet engraftment, using the GF-PLGA polymerassociated complex. In order to do this we: (1) characterized the biodistribution and release kinetics of several formulations and particle sizes in vivo; (2) identified an optimal particle size and dose; (3) co-transplanted GF-loaded galactosylated PLGA (GAL-PLGA) particles concurrently with a non-curative mass of islets via the clinically relevant hepatic portal vein (HPV) into diabetic mice and monitored glycemic control over a 6-week period with histological assessments of islet engraftment in the liver.

| Animals
Male C57Bl/6 mice (8-10 weeks old, Harlan Laboratories) were housed under standard conditions in a 14-h light to 10-h dark cycle and given standard chow and water ad libitum.

| Injection of growth factors and proliferation of cells within liver
In short-term experiments, 12-week-old C57Bl/6 mice (n = 8/group) received the following GFs or vehicle: Group (1)  (R&D Systems ™ ), Group (4) all three GFs, and Group (5) 100μl saline (vehicle), at day −2 and day 0. These doses were based on previous studies 18 with FGF7, which demonstrated that at 24 h following one injection of FGF7 at high dose (5 mg/kg), three of eight mice were anorectic and hypoglycemic (blood glucose levels between 3.0 and 3.9 mmol/L) with signs of distress (Table S1). Therefore, lower doses of FGF7 were used with no adverse effects. Mice were pulsed with BrdU (1 mg dissolved in PBS) via intraperitoneal (i.p.) injection 48 h later and culled 1 h afterwards. Liver lobes were processed for immunohistochemistry (see below) with the antibodies listed in Table S2. The GF associated with the greatest liver cell proliferation FGF7-GAL-PLGA particles achieve selective FGF7 delivery to the liver-promoting islet engraftment to help normalize blood glucose levels with a good safety profile.

K E Y W O R D S
animal models: murine, diabetes: type 1, islet transplantation, regenerative medicine, translational research/science AJT was selected and given s.c. prior to administering islets via the HPV, to determine if glycemic control was improved. This GF was subsequently incorporated into galactosylated PLGA particles.

| FGF7 and effects on insulin secretion and oxygen consumption rates of islets in vitro
Islets (n = 20 islets per well in triplicate) were incubated in 0 (control), 5 ng/ml, and 30 ng/ml FGF7 for 24 h. FGF7 concentrations were based on concentrations released from 1-mg FGF7-GAL-PLGA particles.

| Biodistribution of particles via the HPV or tail vein
PLGA particles were rhodamine-labeled. Mice (n = 3/group) received an injection of 1-mg PLGA particles in 100-µl 30% fetal calf serum (FCS). Non-galactosylated PLGA particles (2, 10, and 22 µm mean diameter) and GAL-PLGA particles (2, 10, and 26 µm mean diameter) were injected with a 30G needle via the HPV and via the tail vein (i.v.); vehicle injections were also run as controls. Mice were culled 24 h postinjection, blood samples collected by cardiac puncture, and the liver, lung, kidney, heart, and spleen were harvested for analysis.

| Biochemical analysis
ALT, albumin, and bilirubin (Alpha Laboratories Ltd.) were analyzed on the Cobas Fara centrifugal analyzer (Roche), human FGF7 and insulin concentrations by ELISA (Thermo Scientific; Mercodia, respectively); insulin content of the pancreas was measured following weighing, homogenization, and sonication. 29

| Histological analysis
We quantified liver cell proliferation, islet engraftment, liver fibrosis, necrosis, and vascularization. Tissue was either fixed, embedded in paraffin, and cut serially (5 µm) or processed using cryosections (8-30 µm). In brief to analyze: (i) proliferation-liver sections were immunostained with BrdU and hepatocyte nuclear factor 4α (HNF4α) to hepatocyte proliferation and total hepatocyte population, respectively, and (ii) PLGA particle detection-rhodamine-labeled particles non-overlapping fields per section were evaluated, using ×20 magnification (Nikon Eclipse E600 fluorescent microscope). The average number of β-cells detected per FFPE section was standardized to the total analyzed fields. 30 (iv) Necrosis and fibrosis-H&E stains were produced using a Shandon Varistain Automated Slide Stainer. Collagen fibers in the liver tissues were detected with picrosirius red (PSR) staining. 31 (v) Vascularization was determined in paraffin-embedded liver sections using immunofluorescence for CD31 32 and the erythroblast transformation-specific-related gene (ERG): a transcription factor specific for endothelial cells; 33 VEGF-A quantification was attempted in liver sections using an immunofluorescence method. 10,33 For each immunostain, control procedures included isotype-matched rabbit monoclonal antibodies. DAPI staining was performed to label nuclei. Slides were mounted using an aqueous medium and imaged using an Operetta High-Content System (PerkinElmer).

| Statistical analysis
Results are expressed as mean ± SEM unless otherwise stated.
Significance was determined by unpaired t tests or one-way ANOVA with Tukey's post hoc testing using Prism 6.0 software (GraphPad Software). A p < .05 was considered significant. pronounced cell proliferation in all organs including the lungs, pancreas, heart, and spleen, as demonstrated by BrdU immunofluorescence staining versus controls ( Figure 1B,C). Therefore, FGF7 was selected for further studies.

| FGF7 has no effect on insulin secretion or OCR in short-term in vitro studies with islets
FGF7 at a dose of 5 or 30 ng/ml had no effect on insulin secretion or OCR ( Figure S1).
The release kinetics showed an initial burst release phase, releasing ~one-third of the FGF7 payload on day 1, declining to 8% release on day 2, and 3% on day 3. Release was maintained at 1% between days 4 and 6 increasing to ~8% release from days 9 to 21 ( Figure 2C). Cumulative in vitro release profiles of the particles and FGF7 delivery dose (ng per mg particles) are shown over 3 weeks ( Figure 2D; Table 1). For 0.1-mg PLGA particles, the FGF7 content was 60 ng and a 70% release of FGF7 over 21 days was ~40 ng FGF7.     No proliferating ß-cells were detected in the pancreases of mice treated with FGF7-GAL-PLGA particles ( Figure S6A). There was no significant difference in pancreatic insulin content between groups ( Figure S6B).

| DISCUSS ION
Transplantation of islets into patients with T1D stabilizes glycemic control, reducing SH 1,5,34 but due to poor engraftment of islets into the liver 35,36 islets from two to three donor pancreases, a scarce resource, are required.
In diabetic rodents, partial hepatectomy preceding intra-portal islet transplantation is associated with improved glycemic outcomes versus islet transplant alone, likely due to GF release, remodeling of the liver niche and liver cell proliferation, improving islet engraftment, and revascularization. 20,30,37 However, partial hepatectomy is not a clinically applicable adjuvant therapy for intraportal islet transplantation in man. FGF7 RNA is expressed in most organs throughout the human body, with moderate expression in the pancreas and no expression in the islet in adulthood. 38 Furthermore, the majority of mice administered these particles were cured following islet transplantation. Importantly, there was no β-cell regeneration in the native pancreas and FGF7 did not augment insulin secretion in our in vitro experiments, consistent with the beneficial effects being mediated by improved hepatic islet engraftment. Of note, there was no evidence of diminished OCRs from islets exposed to FGF7 and therefore no evidence of an adverse effect of FGF7 on islet function in the short term. Subcutaneous FGF7 increased hepatic non-parenchymal cell proliferation; however, glycemic control was not improved following islet transplantation. AJT form between the islets and the liver, occurs largely between days 3 and 28. 43 Subcutaneous FGF7 also causes cell proliferation in other organs including the lungs, pancreas, kidney, heart, and spleen, which limits clinical applicability.
When GAL-PLGA particles were studied over a 21-day period, the route of administration, particle size, and galactosylation influenced its sequestration within organs.

D I SCLOS U R E
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

AUTH O R CO NTR I B UTI O N S
SF, KS, and SJF conceptualized study and obtained funding for the study. OQ formulated FGF7 particles and drafted FGF7 particle methods; SA, PSL, AB, and JN performed animal experiments and laboratory assays. PSL performed immunohistochemistry for vascularization markers and performed additional statistical analyses. SA helped draft the manuscript and performed statistical analyses; JM, PB, and SFG gave technical laboratory assistance. SA, NM, and RC performed and analyzed oxygen consumption rate assays. SJF was PI for the liver regeneration studies, KMS was PI for the FGF7 particle formulation studies, and SF was PI for the FGF7 and metabolic studies; SF drafted and revised the manuscript and figures and performed statistical analyses for the in vivo transplant studies. All authors critically reviewed the manuscript. SF is the guarantor of this work and as such had full access to all the data in this study and takes responsibility for the integrity of the data and the accuracy of the data analysis.