Transplantation of pancreatic islets necessitates an engraftment process, including revascularization of the graft. Studies of graft vasculature have demonstrated that islets become revascularized during the first post-transplant week through an angiogenic process. If this also involves lymphatic vessels is unknown. The aim of the present study was to functionally evaluate if lymphatic vessels, which are absent in endogenous islets, form after islet transplantation. To achieve this, inbred Wistar-Furth rats were transplanted with 250 syngeneic islets under the renal capsule. Intra-vital microscopy of the graft in combination with interstitial injection of Evans Blue was performed 1 week, 1 month or 9–12 months later. In all animals studied, there was drainage through intra-graft lymphatic capillaries emptying into larger lymphatic vessels associated with the renal capsule. The number was slightly lower 1 week post-transplantation. Most of the lymphatic capillaries were present in the graft stroma, rather than interspersed among the endocrine cells. In some animals, we were able to demonstrate dye in regional lymph nodes. We conclude that unlike endogenous islets, islet grafts develop a lymphatic drainage. Its functional importance and characteristics remain to be established. However, it can be speculated that immune reactions may be facilitated by the presence of lymphatic vessels.
Transplantation of isolated pancreatic islets is performed as a heterotopic implantation, and therefore necessitates an adaptation of the endocrine cells in their new surroundings, i.e. an engraftment. This process includes revascularization, reinnervation and reorganization of the tissue, which enables the surviving endocrine cells to acquire an adequate function (1–3). Despite an extensive cell death in the immediate post-transplantation period (4), enough endocrine cells survive in order to make islet transplantations a realistic therapeutic choice in selected patients (5).
In order to minimize the losses of grafted islets, and thereby reduce the number of required donors for each recipient, interest has focused on this engraftment process, in particular the angiogenic process (2,6). It has been assumed that most of the revascularization of transplanted islets occurs through stimulation of angiogenic processes in the implantation organ microvessels. However, it seems clear that some endothelial cells survive in the islet isolation procedure and participate in the angiogenic process (7), and also that bone marrow-derived angioblasts may contribute (8). Despite this, there is evidence of an impaired revascularization of transplanted islets, since most of the graft capillaries are present in the connective tissue adjacent to the islets (9,10). The reasons for this are at present unknown.
In spite of the interest in revascularization with blood vessels, lymphangiogenesis in transplanted islets has so far received no attention. Islets normally lack functional lymphatic capillaries, even though such vessels can be found associated with adjacent exocrine acini (11–13). In nonobese diabetic, mice lymphatic vessels associated with islets have been suggested to play a key role in the infiltration of dendritic cells into islets (14). Furthermore, there is evidence of intra-tumoral lymphangiogenesis in pancreatic endocrine tumors (15). In view of this, and in the light of the fact that the presence of lymphatic vessels would also enable the exit of antigen-presenting cells and antigenic material to regional lymph nodes, we deemed it of interest to investigate to what extent lymphatic microvessels are present in grafted islets.
Materials and Methods
Male, inbred Wistar-Furth rats weighing 300–350 g, were purchased from B&K Universal (Sollentuna, Sweden). All animals had free access to tap water and pelleted rat food throughout the experiments. Principles of laboratory animal care (NIH publication No. 23, revised 1985) were followed. All experiments were approved by the local animal ethics committee at Uppsala University.
Islet isolation and transplantation
Pancreatic islets were isolated by a collagenase digestion method, as previously described (16). The islets were cultured free-floating for 3–5 days, with 150 islets in each culture dish, in 5 mL culture medium, RPMI 1640 (Sigma-Aldrich, Irvine, UK) supplemented with l-glutamine (Sigma-Aldrich), benzylpenicillin (100 U/mL; Roche Diagnostics Scandinavia, Bromma, Sweden), streptomycin (0.1 mg/mL; Sigma-Aldrich) and 10% (v/v) fetal calf serum (Sigma-Aldrich). The medium was changed every 2nd day.
A total of 25 syngeneic male, Wistar-Furth rats were used as recipients. These animals were anesthetized with an intra-peritoneal injection of ekviticine (a mixture of pentobarbital and chloral hydrate). The left kidney was exposed through a flank incision and 250 islets were implanted under the renal capsule. The animals were then allowed to recover, and were subsequently studied 1 or 9–12 months after transplantation.
In vivo microscopy and detection of lymphatic vessels
The transplanted rats were anesthetized with an intra-peritoneal injection of thiobutabarbital (Inactin™, Research Biochemicals, Natick, MA, USA; 120 mg/kg body weight) 1 week, 1 month or 9–12 months after transplantation. The animals were placed on a heated operating table to maintain body temperature and breathed spontaneously through a tracheostomy. Polyethylene catheters were inserted into the femoral artery and vein. The arterial catheter was connected to a pressure transducer (PDCR 75/1, Druck Ltd., Groby, UK) to allow constant monitoring of the mean arterial blood pressure, whereas the venous catheter was used to continuously infuse Ringer solution (6 mL/kg body weight/h; Fresenius Kabi, Uppsala, Sweden). A subcostal left flank incision was made to visualize the graft-bearing left kidney. The kidney was gently dissected free from surrounding tissues and immobilized in a plastic cup. The kidney was embedded in pieces of cotton wool soaked in Ringer solution and covered with mineral oil (Apoteksbolaget, Uppsala, Sweden) to prevent evaporation and thereby keep the kidney moist and at body temperature.
Microvessels in the graft and surrounding kidney were observed under a Leitz microscope using a ×10 saltwater immersion objective (Leitz, numerical aperture 0.13). Continuous recordings were obtained from a digital camera (TK-C1481EG, JVC Company of America, Wayne, NJ, USA) and later analyzed with Wincoder® (Intervideo, Fremont, CA, USA).
When basal recordings had been obtained a 0.1% solution of Evans Blue in saline (Sigma-Aldrich) was administered into the interstitium of the grafts through sharpened glass micropipettes (tip diameter 8–12 μm) with the aid of a micromanipulator, and the distribution and drainage of the dye were recorded (17,18). In some instances, capillaries and veins in the grafts were also injected with Evans Blue, and the distribution of the dye was observed and recorded.
After the study, the part of the kidney containing the graft was removed and fixed in 10% formaldehyde overnight. The samples were dehydrated and embedded in paraffin.
Paraffin sections, 4-μm thick, were placed onto Superfrost/plus® slides (Mentzel, Germany), deparaffinized in xylene and rehydrated in graded alcohols. Incubation with Peroxidase blocking reagent® (DAKO, Glostrup, Denmark) for 10 min to block endogenous peroxidase was followed by washing in Wash buffer® (DAKO). The VEGFR-3/FLT-4 antibody (RDI, Flanders, NJ, USA) or LYVE-1 antibody (Santa Cruz Biotechnology, USA) diluted 1:100 in a buffer (Antibody dilutent with background reducing components®; DAKO) was applied to the sections for 30 min at room temperature followed by washing in Wash buffer®. The ChemMate EnVision detection kit® (DAKO) was used and developed with diaminobenzidine before counterstained with hematoxylin. Sections were dehydrated and mounted in Pertex mounting medium (HistoLab, Gothenburg, Sweden). Negative control slides were processed identically except that the primary antibody was omitted. The immunohistochemical staining was performed at room temperature using an automated immunostaining instrument (Autostainer plus®, DAKO).
Cannulation of renal lymphatic vessels
In three animals not used for in vivo microscopy, the graft-bearing kidney was visualized as described above. A thin polyethylene catheter (tip diameter approximately 200–300 μm) was inserted into a hilar lymphatic vessel and lymph was collected by free flow for 30 min. Approximately 10 μL was obtained from each rat. At the end of the sampling period, a blood sample was taken from the femoral vein. Insulin concentrations in the lymphatic and serum samples were then analyzed with ELISA (Supersensitive Insulin ELISA, Mercodia AB, Uppsala, Sweden).
Two animals were excluded from the study due to infection in the immediate post-transplantation period, and three due to technical problems in the surgical preparations. Thus, a total of 20 transplanted rats were examined by in vivo microscopy 1 week (n = 4), 1 month after transplantation (n = 7) or after 9–12 months (n = 9). A distinct, whitish well-demarcated graft, which could easily be distinguished from the surrounding renal parenchyma, was seen in all animals (Figure 1). There were large intra-capsular venous blood vessels (Figure 2A), which radiated centripetally from the graft out into the renal parenchyma, irrespective of the length of the observation period.
When Evans Blue was administered interstitially into the grafts, some of the dye was spread diffusely between groups of endocrine cells. However, most of the dye dispersed into surrounding, previously invisible vessels (Figures 2B,C and 3A,B). There were some regional differences within the grafts, but lymphatic capillaries (Figure 4) were generally seen wherever the dye was injected. The time frame of the appearance of the dye within these vessels was in the order of several minutes. These smaller vessels emptied into larger vessels. Almost all of these lymphatic vessels were located in graft stroma, and not between the endocrine cells per se (Figures 3B and 4). There were larger conduit vessels containing Evans Blue in the capsule adjacent to, but distinct from, larger veins which emptied downward into the renal parenchyma (Figure 5A,B). Structures similar to valves were occasionally seen in some of the larger vessels. In some animals we could discern the dye in regional lymph nodes at the origin of the left renal artery.
The number of lymphatic vessels within the grafts was similar at 1 and 9–12 months post-transplantation. In animals studied after 1 week, the number was lower, and in 3 out of the 4 animals studied a network could be seen within the implant with drainage as described above. In the fourth animal, we could discern a lower number of lymphatic vessels, but otherwise the organization was similar to that in the other rats.
When Evans Blue was injected into blood vessels, i.e. vascular structures containing red blood cells, a much faster rate of emptying occurred, and the dye was cleared within a matter of seconds. In all these cases, the dye entered the large venular structures, which emptied downwards into the renal parenchyma. Also with regard to blood microvessels, most were located in the graft stroma, rather than in the endocrine parenchyma.
There were alleged lymphatic capillaries and larger lymph vessels in 9/9 animals transplanted 9–12 months before the study, 6/7 after 1 month and 4/4 after 1 week. It is, however, difficult to make a quantification of these numbers, since identification of the lymph vessels depends on dye injection. Nevertheless, they seemed to be substantially fewer than the number of graft blood vessels at all times after implantation.
Stained vascular structures were seen in all examined grafts when VEGFR-3 and LYVE-1 stained sections of islet grafts were studied (Figures 6 and 7). There were no differences in the number of lymph vessels between grafts obtained 1 or 9–12 months after implantation, but the number was slightly less in rats studied 1 week post-transplantation.
Insulin concentrations in lymph derived from the graft-bearing kidney was approximately 50% of that in serum (0.43 ± 0.05 vs. 0.88 ± 0.10 ng/mL; n = 3).
In the present study, we were able to demonstrate the presence of a functional lymphatic network in islet grafts from 1 week up to 9–12 months post-transplantation. Both lymphatic capillaries and larger confluent vessels were seen throughout the grafts. However, most of the lymphatic capillaries were present in the connective tissue stroma between the transplanted islets, which is similar to the distribution of intra-graft blood vessels (10). The number of lymph vessels was clearly lower than blood vessels. The reason for not directly quantifying this reduction was the difficulties in identifying lymphatic vessels when they were not filled with Evans Blue as lymph vessels do not contain erythrocytes (19).
Interstitially injected Evans Blue was drained through progressively larger lymph vessels, which finally emptied into the kidney parenchyma, presumably into renal intra-lobular lymph vessels, which then exit from the kidney through hilar lymphatics (20,21). In confirmation of this we could observe dye in regional lymph nodes close to the celiac plexus in some animals. The time span for the drainage was in the order of several minutes.
It is difficult to ensure with certainty the exact location of the glass micropipette tip, i.e. whether it is positioned in the stroma between individual islets or directly within an islet. The former is more likely, since the interstitium in islets is very sparse. Thus, if the micropipette is within the islets it is likely to empty the Evans Blue into blood capillaries rather than between blood vessels. This was quite frequently seen, but occasionally lymphatic capillaries were also interspersed between endocrine cells. The larger lymphatic tributaries were, as expected, located in the stroma. Some of them were as large as surrounding veins and occasional valves could be seen, thereby confirming previous reports of presence of lymphatic valves in the rat pancreas (22).
There is a similar distribution of lymphatics in tumours as that in the transplanted islets, viz. large lymph vessels around the tumors and only few interspersed between the cells, in islet endocrine tumors (15) and malignant tumors in general (23,24). One possible reason for these similarities is that lymphatic vessels collapse due to the stress exerted during growth and/or reorganization of cells within tumors and grafts (25). However, since the pattern of lymphatic vessels was the same up to 1 year after implantation, i.e. at a time point when growth or reorganization is unlikely to occur, this seems to be less likely.
Lymphatic ingrowth into islet transplants has, to our knowledge, not been previously studied. However, in whole-organ transplantation this has received some attention. Thus, there is a restoration of lymph channels within 2–3 weeks in renal grafts (26,27). In the present study, this occurred even earlier and lymph channels were seen already after 1 week. Thus, there is clearly a potential for lymphangiogenesis in transplanted organs. The mechanisms behind this are for the moment unknown.
Previous studies on lymphangiogenesis suggest that VEGF-C and its associated receptor VEGFR-3 are necessary for lymphangiogenesis to occur, whereas mainly VEGF-A and VEGFR-2 are involved in blood vessel angiogenesis, even though some overlap occurs (28). It should be noted that the α-cells of normal islets express VEGF-C (29), but the physiological role of this is unknown. Since α-cells are preferentially located to the periphery of the islets, it may be that VEGF-C released from these cells helps to form the stromal network of lymph vessels. In the present study, we found VEGFR-3 positive endothelial cells in the grafts, which speaks in favor of this notion. However, VEGFR-3 is not completely specific for lymphatic endothelium (29,30). In previous studies of the human pancreas, the lymphatic capillaries, but not blood capillaries, have been found to be stained for VEGFR-3 (29). It, therefore, seems likely that both VEGF-C and VEGFR-3 are involved in the process of graft lymphangiogenesis, even though this awaits further confirmation. However, lymphatic endothelial cells can also possess VEGFR-2, and through this receptor VEGF-A may promote their survival in vitro (31). Indeed, e.g. in wound healing, including ingrowth of blood vessels in corneal transplants (32), lymphangiogenesis and blood vessel angiogenesis are stimulated in concert with one another (28). Thus, the present findings can be reconciled with the hypothesis that lymphangiogenesis is, at least partially, secondary to the ongoing angiogenesis.
Another concept, which is of interest when interpreting the present findings, is that of a communicating interstitium, such as seen in the kidney (21). Thus, a loose interstitium may enable a flow of interstitial fluid toward peripherally located larger lymphatic vessels or blood vessels at other locations in the organ. The resolution of the microscope and camera used in the present study did not enable us to evaluate this. However, when we have used antibodies against VEGFR-3, vide supra and another putative antibody against lymphatic endothelium, viz. LYVE-1, we could observe, endothelium-lined vascular structures in the grafts. LYVE-1 is a homolog of the CD44 hyaluronan receptor (33,34). In view of these findings with immunohistochemical stainings, and the demonstration of drainage of interstitially injected Evans Blue into vascular structures we deem it certain that transplanted pancreatic islets contain a lymphatic system. It should be noted that the normal protein concentrations in renal lymph and blood plasma are approximately 2% and 6%, respectively, i.e. a lymph/plasma ratio of 1:3 (35). However, the ratio between insulin in lymph and plasma was 1:2 in the present experiments, which suggests that some drainage of insulin occurs through the lymphatics. How important from a functional point of view this is remains to be established.
One final word of caution refers to the fact that we have so far only studied islets implanted under the renal capsule, not within the clinically used intra-hepatic site. For technical reasons we are at the moment unable to perform such studies.
Thus, the major finding in the present study is that islets syngeneically transplanted under the renal capsule become enmeshed in a network of lymphatic capillaries and larger vessels, which drain the graft interstitium. This alters the microcirculatory physiology of grafted islets when compared to that of endogenous islets, since the latter do not possess a comparable lymphatic drainage (11–13). Furthermore, lymph drainage is likely to facilitate the exit of antigen-presenting cells and antigenic material to regional lymph nodes, which may enhance immune reactions.
The skilled technical assistance of Astrid Nordin is gratefully acknowledged. Financial support was received from the Swedish Research Council (72X-109, 72XD-15043), the Swedish Diabetes Association, the Juvenile Diabetes Research Foundation, the EFSD/Novo Nordisk for Type 2 Diabetes Research Grant, the NOVO Nordic Research Fund, Åke Wibergs Stiftelse, Magnus Bergvalls Stiftelse, Barndiabetesfonden, Anérs Stiftelse, Groschinskys Minnesfond, Svenska Läkaresällskapet and the Family Ernfors Fund.