A Novel ex vivo Method for Investigating Vascularization of Transplanted Islets Free

Pancreatic islet function is interwoven with the local microenvironment and dependent on vascularization. When isolated islets are removed from their highly vascularized in vivo niche, they rapidly lose their intra-islet endothelium. Thus, the study of islet function in vitro is intrinsically avascular – an issue that underlies a push for novel models of islet biology that incorporate blood vessels such as the pancreas slice model [1]. Furthermore, rapid revascularization is important for islet graft survival and treatment of type 1 diabetes [2]. Transplanted islets are thought to be revascularized by blood microvessels that grow into islets from the host via angiogenesis [3, 4]. Despite the recognized need for islet vascularization, critical questions remain unanswered. Do vessels invade the islet? Do islet-derived vessels grow out and connect to the nearby microcirculation? How long does it take for islets to become vascularized? Does islet revascularization depend on multiple cell types?

These questions motivate the need to develop biomimetic models to investigate the spatial and temporal dynamics of islet engraftment. A challenge in model development or model selection is balancing the temporal readouts of an in vitro system versus the complexity of a real tissue in vivo environment. Recently, microfluidic [5] and anterior eye chamber models [2] have emerged to define the edges of the gap between in vitro and in vivo approaches. Islet transplantation into microfluidic systems incorporate endothelial lined host channels, endothelial cell sprouting, and vessel perfusion. Transplanting islets into the anterior chamber of the eye allows for longitudinal imaging of perfused vessels. Tradeoffs between the approaches include throughput, complexity, environmental control, and ease of use.

In an attempt to bridge the gap between in vivo and in vitro microvascular models, our laboratory introduced the ex vivo rat mesentery tissue culture model. The mesentery thinness (20 μm–40 μm) makes it ideal for culturing and imaging [6]. Recent validation studies have shown that intact microvascular networks remain viable in mesenteric explants with endothelial cells, pericytes, smooth muscle cells, lymphatic endothelial cells, immune cells, and interstitial cells [7, 8]. Validation studies further support smooth muscle cell functionality and the maintenance of endothelial cell phenotype [9, 10]. We have also demonstrated the usefulness for time-lapse imaging of microvascular network growth including capillary sprouting, the evaluation of angiogenic therapies, and the tracking of exogenously transplanted cells [11, 12]. In the context of islet transplantation, we previously grafted islets to the mesentery in vivo to reverse chemically induced diabetes in mice and rats [13, 14]. The in vivo transplantation of islets on mesenteric tissues and the advantages of evaluating cell dynamics ex vivo motivate the objective of this study to demonstrate the utility of the rat mesentery tissue culture model as a novel tool to investigate islet microvascular integration (Fig. 1).

The University of Florida Animal Care and Use Committee approved all experimental protocols using rat islets. Rat islets were isolated from the pancreases of postnatal day 5 Sprague-Dawley rats (Charles River, Willington, MA, USA) consistent with previously published methods [13‒16]. To briefly summarize, postnatal day 5 rat pups were euthanized by decapitation. The whole pancreas was removed and digested in 0.15-mg mL⁻¹ Liberase TL (Roche, Basel, Switzerland) in Hank’s buffered salt solution (HBSS, Gibco-Thermofisher, Waltham, MA, USA) with 20-mM HEPES for 7 min with strong manual agitation. HBSS, 20-mM HEPES, and 0.5% fetal bovine serum (FBS) were added to stop Liberase enzyme digestion. Digested pancreas tissues were washed four times with HBSS, 20-mM HEPES, and 0.5% FBS at 4°C, resuspended in Histopaque-1119 (Sigma-Aldrich, St. Louis, MO, USA), and carefully overlain with HBSS, 20-mM HEPES, and 0.5% FBS at room temperature. The solution was centrifuged with no breaks for 20 min at 300 g at 20°C. Islets were retrieved from the interface between Histopaque and HBSS phases. Next, the islets were washed three times with a solution of HBSS, 20-mM HEPES, and 0.5% newborn calf serum. Islets were hand-cleaned with a 200-µL pipette, and 100–200 islets were taken from the 10-cm petri dish and placed in a 15-mL conical tube with 4 mL of medium for seeding onto our mesenteric tissues. For islet labeling with DiI, 20-μL DiI was added to 3.5–4 mL of the islet containing medium at 37°C for 5 min and then subsequently incubated at 4°C for 15 min. The islet/DiI solution was centrifuged 3 times at 180 g for 15 s and washed using MEM medium between each centrifugation. Islets were suspended in MEM medium and kept at 37°C before being seeded onto mesenteric tissues.

The rat mesentery culture model was used as described previously [7, 8]. The method described in this paper details the transplantation of islets onto cultured mesentery tissues (Fig. 2). Animal protocols and tissue harvesting were approved by the University of Florida Animal Care and Use Committees. First, adult Wistar rats (Harlan, Indianapolis, IN, USA) were anesthetized with an intramuscular injection of ketamine (80 mg/kg body weight) and xylazine (8 mg/kg body weight). Mesenteric tissues were surgically removed under aseptic conditions and spread the mesenteric tissue onto a plastic stage using the ileum as a point of reference. Rats were euthanized with an intracardiac injection of 0.2-mL Beuthanasia before harvesting mesenteric windows. Tissues were rinsed in DPBS after being excised and immediately placed in serum-free MEM supplemented with 1% penicillin/streptomycin (all from Life Technologies, Carls-bad, CA, USA). Next, tissues were spread onto a polycarbonate filter fitted to a cell-crown insert (CellCrown, Sigma-Aldrich, St. Louis, MO, USA). Islets were seeded onto the tissues at a concentration of 20 islets/20 μL and then incubated for 20 min at 37°C. The tissue and cell-crown insert were carefully inverted into a well of a 6-well culture plate and 4 mL of MEM supplemented with 10% FBS was carefully added into the sides of each well to ensure islets were not knocked off of the tissue. Tissues were incubated at 37°C and medium was replaced daily for 3–5 days. Alternatively, to demonstrate the feasibility for tracking host tissue cell dynamics, additional tissues were harvested from adult Sprague-Dawley EGFP rats (Rat Resource and Research Center, Columbia, MO, USA) and cultured following the same steps detailed above.

Vascular networks were imaged longitudinally by labeling with BSI-lectin conjugated to FITC (FITC-lectin; Sigma-Aldrich; St. Louis, MO, USA). BSI-lectin labeled both blood and lymphatic vessel networks. Each well was supplemented with FITC-lectin (1:40) and incubated for 30 min under standard conditions followed by two washes with lectin-free media. Islets were identified by positive DiI staining for imaging.

Tissues were fixed on day 3 with 100% methanol. We then washed and stained the tissues for insulin, CD31 (PECAM), and neuron glial antigen 2 (NG2). All antibodies were diluted in PBS +0.1% saponin +2% BSA, and tissues were washed three times with PBS +0.1% saponin for 10 min between each antibody incubation. Insulin/PECAM/NG2 panels used 1:1,000 dilution of guinea pig anti-insulin antibody (Dako, Santa Clara, CA, USA), 1:50 dilution of biotinylated mouse monoclonal anti-CD31 primary antibody (BD Biosciences, San Jose, CA, USA), and 1:100 dilution of rabbit anti-neuron glial antigen 2 (NG2; MilliporeSigma, Burlington, MA, USA). Secondary antibodies were used at a concentration of 1:200 for goat anti-rabbit (Jackson ImmunoResearch, West Grove, PA) and goat anti-guinea pig (Abcam, Cambridge, UK) antibodies and 1:500 dilution of streptavidin antibody conjugated to CY2 (Jackson ImmunoResearch, West Grove, PA, USA). For Figures 1, 3, and 4, images were acquired using 4× and 10× objectives on an inverted microscope (Nikon eclipse Ti2) coupled with a Photometrics CoolSNAP EZ camera or an Andor Zyla sCMOS camera. Confocal images in Figure 5 were acquired using the 10× and 20× objectives on a Zeiss LSM 710 coupled with an Axio Observer microscope.

The major contribution of this study is the demonstration of a method for culturing transplanted islets with intact rat mesentery tissues. We show that this method can be used to investigate rapid islet revascularization, which is important for islet survival and function. After 3 days in culture, transplanted islets were associated with multicellular capillaries (Fig. 1). Imaging of lectin or PECAM-positive labeling enabled quantification of angiogenesis from islets (Fig. 3). Angiogenic outgrowth can be characterized by radial sprouting from the islet and the formation of vessel connections. For example, Figure 3b shows representative sprouting and branch point data for 12 islets on 4 tissues. Capillary sprouts were defined as blind-ended lectin-positive vessel segments and branch points were defined as the intersection of two or more segments. Observation of multiple islets near host networks indicates that islets can connect with nearby vessels and other islets (Fig. 3c). Lectin labeling at different time points enabled time-lapse observation of the vessel growth and confirmed capillary sprouting both from the transplanted islet and toward the islet from the host vasculature (Fig. 4). Optical sections (less than 2.5 μm) were taken throughout the islet using confocal microscopy and confirmed PECAM-positive labeled vessels in the same focal plane as the islet core (Fig. 5).

The ability to observe vascular outgrowth from the islet and connections with nearby microvascular networks motivates numerous fundamental questions that highlight the unknowns regarding the cell dynamics involved in islet revascularization. What is the direction of vessel growth? What role do cell types other than endothelial cells play? What cell types come from the islets? What cell types come from the host environment? Are vascular pericytes involved? Is interstitial cell recruitment involved? Does islet ensheathment influence revascularization? Are the newly formed vessels functional? Is islet revascularization regulated by hypoxia? Is islet revascularization dependent on inflammation? Answering these questions requires a method that allows for characterization of islet and host microvascular network interactions over time in an intact network environment. Our representative results establish the culturing of islets with rat mesentery tissue as such a method. Future lineage, mechanistic, and characterization studies will be needed to systematically probe the associated cellular dynamics.

Major advantages of the islet revascularization model described herein are that islets remain accessible, and the tissue is easily imaged. The mesentery model can be combined with techniques such as lineage tracing, genetic reporters, and calcium imaging as a platform for characterizing the islet functionality and glucose responsiveness as a function of the revascularization process. Of particular interest are the endothelial sprouts that emerge radially from islets placed on the mesenteric tissue. The fact that these sprouts form easily and grow to connect with host vasculature is a unique feature of this model. These results suggest that the microenvironment of the mesentery explant encourages islet vascular sprouting. It is notable that such clear sprouting behavior does not readily occur when islets are placed on other surfaces such as tissue culture polystyrene or extracellular matrix hydrogels. In these models, fibroblastic cells tend to overtake those cultures and endothelial sprouts do not form predictably. Thus, the mesentery explant model provides a rich niche for expanding the frontiers of islet physiology in areas where current models are limited.

Since an emergent paradigm of islet revascularization is the coordination of multiple cell types [17], the maintained multi-cell/system complexity in the rat mesentery culture model serves to provide a valuable view. It is thought that transplanted islets are revascularized by blood vessels that grow into islets from the host via angiogenesis [3, 4]. Our representative results suggest that vessels can also sprout from the donor islet and motivate the investigation of cell interactions involved in the vessel connection process. One particular cell type that has been relatively unexplored is the vascular pericyte [18‒20]. Pericytes both elongate over and wrap around endothelial cells. Moreover, they physically interact with endothelial cells by influencing endothelial cell migration via direct integrin signaling and growth factor presentation [21, 22]. The multifaceted role of pericytes is highlighted as they are also considered to be vessel stabilizers and even regulators of capillary blood flow [22‒24]. Almaca et al. demonstrated pericyte control over capillary diameter in islets [25]. In the context of islet revascularization, Landsman [18] demonstrated a role for NG2, which is a common pericyte marker. PECAM and NG2 labeling along vessels suggests the presence of both endothelial cells and vascular pericytes (Fig. 1, 5; control labeling is shown in online suppl. Fig. S1; for all online suppl. material, see www.karger.com/doi/10.1159/000523925). Our observations of NG2-positive pericytes along capillaries connecting transplanted islets to nearby host vessels supports the application of our method for probing pericyte cell lineage to determine, for example, whether pericytes along vessels sprouting from the islets originate from the islet or are recruited from the host environment. Another application enabled by our model is the investigation whether islet revascularization involves lymphatic vessels. Interestingly, we did not observe LYVE-1-positive lymphatic vessel sprouting near the angiogenic islets (data not shown).

The ease of imaging using a standard epi-fluorescent microscope and the ability to manipulate the environment by harvesting tissues from altered rat strains (diseased, aged, etc.) supports the method for use in future applications. For example, trial experiments using tissues harvested from GFP transgenic rats support the recruitment of interstitial cells to the islet perimeter during culture (Fig. 6). The observation of interstitial cell recruitment implicates a role for inflammatory cells and motivates future gain and loss of function studies to determine their importance of these cells. It is important to note that the current method description demonstrates feasibility of a novel approach and highlights the potential for future work focused on enhancing islet outgrowth, enhancing microvascular integration, and probing mechanistic players. Manipulating the culture system can allow for testing the effects of culture conditions, single growth factors, inhibition of specific cell types, and/or culture duration. For the current study, all tissues were cultured with serum. While additional trials support the revascularization of islets in serum-free conditions, future studies will be needed to evaluate how the revascularization process is influenced by changes in the local environment. Future studies will also be needed to evaluate vessel function, including vessel permeability for the intra-islet vessels, the angiogenic vessels outgrowing from the islet, and the nearby angiogenic host vessels.

Another important consideration in islet transplantation is the success rate. A common assay for evaluating islet function is insulin secretion, a method which generally entails pooling multiple islets. Consequently, data related to functional viability do not provide information about islet-specific success rates. Sparse example from the literature detailing in vivo transplantation and isolation studies suggests that there could be a high nonresponder rate [26‒29]. Approximately 60% of murine islets were reported to be lost during the peritransplant period, with the loss being attributed to both beta-cell apoptosis and necrosis [27]. Following human islet isolation, up to 65% remained viable suggesting a margin for improvement [28, 29] prior to transplantation. Interestingly, the potential for high nonresponder rates is similarly seen in our model. The nonresponder rate based on an analysis of 89 islets from 9 tissues and 2 rats was 60 +/− 26% (Avg +/− SD) and the tissue-specific rates ranged from 20% to 94%. For the analysis, tissues were selected for analysis if they contained at least one sprouting islet. Nonresponders were defined as an islet with no observable radial vessel sprouting. Islets embedded within a dense microvascular network without clear sprouting were also categorized as nonresponders. The islet nonresponders could be due to a lack of viability or a lack of responsiveness. Regardless, future studies will be required to determine what factors might enhance islet revascularization and integration with the nearby microvascular networks. Our quantification of sprouting and branch points also supports the possibility that some islets might be more responsive than others (Fig. 3). The variability across individual islets suggests that the amount of microvascular outgrowth might be dependent on the islet microenvironment and starting intra-islet vasculature. For these data, only angiogenic islets were analyzed. Some islets displayed no vascular outgrowth further supporting the importance of initial islet health.

Limitations of this method include the lack of vessel perfusion and the use of rat versus mouse tissue. Recent work from our laboratory suggests these limitations can be addressed with complimentary methods [10, 30, 31]. For example, we introduced a bioreactor design that enables microvascular network perfusion during culture [10]. We also demonstrated that analogous mouse tissues contain multicellular networks and can undergo angiogenesis ex vivo. Our results for this methods article and our previous work demonstrating the potential for introducing perfusion and using murine tissue support the future integration of approaches for specific cell lineage, mechanistic, and functional studies. Another limitation is the use of DiI. While this membrane marker is commonly used for lineage studies, it only is detectable across 3–4 cell divisions. Hence, the DiI labeling might not be observable on cells or vessels originating from the islet. The DiI labeling might also not label all intra-islet cells. Both scenarios offer explanations for why the sprouts shown in representative images are DiI negative. Also, for the current study, islets isolated from Sprague-Dawley rats were transplanted onto tissues harvested from Wistar rats. The strain mismatch was due to logistical issues related to the collaboration across labs. With that being noted, the transgenic GFP rat background strain is Sprague-Dawley, which is the same strain for the islets. So, the apparent vascular outgrowth in the trial studies using tissues harvested from GFP rats (Fig. 6) supports replication of our results. Finally, future studies will be needed to evaluate islet function and integration with diabetic tissues – an environment which would more closely mimic the pathophysiology of diabetes. Regarding function, whether transplanted islets maintain or have compromised islet insulin secretion responses are yet to be determined.

In conclusion, this study documents a method of adding exogenous islets to cultured rat mesenteric tissues. The relevance of using the mesentery as a host tissue is supported by in vivo chronic experiments that investigate islet engraftment [13, 14]. Similar to in vivo results, transplanted islets are able to engraft with the tissue during culture and connect to the vasculature. Our ex vivo approach offers an alternative time-lapse view using the rat mesentery culture model for characterizing the temporal-spatial angiogenesis dynamics involved in islet revascularization. Compared to bottom-up engineered in vitro approaches, including three-dimensional cell scaffold models and microfluidic devices, our method incorporates intact microvessels and the local environment by maintaining the physiological complexity of the tissue in culture.

We would like to thank the members of the Murfee and Phelps laboratories for their help with this collaboration. We specifically thank Walker Hagan and Matthew Becker in the Phelps laboratory for assistance with rat islet isolation.

This study protocol was reviewed and approved by the University of Florida Animal Care and Use Committee (IACUC Protocol # 201710060).

The authors have no conflicts of interest to declare.

This project was supported by funding from the National Institute of Aging (Grant No. R01AG049821) and by a JDRF Agreement Award (Grant No. 2-SRA-2019-781-S-B).

Robert Dolan and Nicholas A. Hodges were responsible for the islet transplantation and tissue culture experiments. Robert Dolan was also responsible for imaging. Jorge Santini-González and Edward A. Phelps were responsible for islet isolation. Robert Dolan, Edward A. Phelps, and Arinola Lampejo contributed to the writing interpretation of the results. Edward A. Phelps also provided input on the motivation for the method development and the interpretation of observations. Walter L. Murfee oversaw the project and contributed to the method design, interpretation of the results, and writing.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.