Intraislet Endothelial Cells Contribute to ... · Intraislet Endothelial Cells Contribute to...

Intraislet Endothelial Cells Contribute toRevascularization of Transplanted Pancreatic IsletsMarcela Brissova,

1Michael Fowler,

1Peter Wiebe,

2Alena Shostak,

1Masakazu Shiota,

2

Aramandla Radhika,1

P. Charles Lin,3

Maureen Gannon,1,2

and Alvin C. Powers1,2,4

Pancreatic islet transplantation is an emerging therapyfor type 1 diabetes. To survive and function, trans-planted islets must revascularize because islet isolationsevers arterial and venous connections; the currentparadigm is that islet revascularization originates fromthe transplant recipient. Because isolated islets retainintraislet endothelial cells, we determined whetherthese endothelial cells contribute to the revasculariza-tion using a murine model with tagged endothelial cells(lacZ knock-in to Flk-1/VEGFR2 gene) and using trans-planted human islets. At 3–5 weeks after transplanta-tion beneath the renal capsule, we found that isletswere revascularized and that the transplant recipientvasculature indeed contributed to the revascularizationprocess. Using the lacZ-tagged endothelial cell model,we found that intraislet endothelial cells not only sur-vived after transplantation but became a functionalpart of revascularized islet graft. A similar contributionof intraislet endothelial cells was also seen with humanislets transplanted into an immunodeficient mousemodel. In the murine model, individual blood vesselswithin the islet graft consisted of donor or recipientendothelial cells or were a chimera of donor and recip-ient endothelial cells, indicating that both sources ofendothelial cells contribute to the new vasculature.These observations suggest that interventions to acti-vate, amplify, or sustain intraislet endothelial cellsbefore and after transplantation may facilitate isletrevascularization, enhance islet survival, and improveislet transplantation. Diabetes 53:1318–1325, 2004

Pancreatic islet transplantation holds great prom-ise for the treatment of type 1 diabetes sincerecent advances in islet isolation and immuno-suppression have led to greatly improved results

(1–3). However, several major challenges currently pre-vent islet transplantation from being widely adapted as atreatment for type 1 diabetes. For example, less-toxic

immunologic interventions are needed to prevent allograftrejection and the recurrence of the autoimmune processthat originally caused type 1 diabetes. Another majorchallenge is that most patients must receive islets isolatedfrom at least two pancreata to become insulin independentand often insulin independence is not permanent (4). Whyislets from at least two pancreata are required to reversediabetes is perplexing as the majority of the pancreas canbe surgically removed without a normal individual becom-ing diabetic. One possible explanation for the requirementof islets from at least two pancreata is that many islets diein the first days after transplantation, before adequatevascular supply is reestablished. Davalli and colleagues(5–7) found that islet cell survival, islet insulin content,and �-cell mass declined 1–3 days after transplantation.This is the period when the islet graft is avascular, sinceislet isolation severs arterial and venous connections; untilrevascularized, transplanted islets are dependent on diffu-sion of nutrients and oxygen from the surrounding tissue.These data suggest that the reduced islet vascularity in theimmediate posttransplant period may contribute to theearly loss of islet cells and their function.

While angiogenesis and revascularization are essentialparts of islet cell engraftment, the molecular events in-volved in the revascularization of transplanted islets areincompletely understood. Using a variety of approaches(dorsal skinfold chamber of Syrian golden hamsters ormice, intravital fluorescence microscopy, histologic anal-ysis, RT-PCR, immunocytochemistry, corrosion casting,and electron microscopy), investigators (8–16) have foundthat capillary sprouting, angiogenesis, and revasculariza-tion begin 2–4 days after islet transplantation and aremostly complete by 10–14 days. The current paradigm isthat islet grafts become revascularized from recipientvessels and endothelial cells (10,13,16–18). Because iso-lated islets contain endothelial cells and capillaries andthese vascular elements are included in the islet trans-plant, we examined whether intraislet endothelial cellscontribute to revascularization of transplanted islets in amurine model in which endothelial cells are tagged withlacZ and in human islets transplanted into immunodefi-cient mice. Our results indicate that both donor andrecipient endothelial cells participate in the revasculariza-tion of transplanted islets. By injecting an endothelium-binding lectin, we were able to demonstrate for the firsttime that intraislet endothelial cells have a capacity tointegrate into the functional vasculature of revascularizedislet graft.

From the 1Department of Medicine, Division of Diabetes, Endocrinology, andMetabolism, Vanderbilt University Medical Center, Nashville, Tennessee; the2Department of Molecular Physiology and Biophysics, Vanderbilt UniversityMedical Center, Nashville, Tennessee; the 3Department of Radiation Oncol-ogy, Vanderbilt University Medical Center, Nashville, Tennessee; and the 4VATennessee Valley Healthcare System, Nashville, Tennessee.

Address correspondence and reprint requests to Alvin C. Powers, Divisionof Diabetes, Endocrinology, and Metabolism, 715 PRB, Vanderbilt University,Nashville, TN 37232. E-mail: [email protected].

Received for publication 17 October 2003 and accepted in revised form 23December 2003.

Additional information for this article can be found in an online appendix athttp://diabetes.diabetesjournals.org.

PECAM, platelet endothelial cell adhesion molecule.© 2004 by the American Diabetes Association.

1318 DIABETES, VOL. 53, MAY 2004

RESEARCH DESIGN AND METHODS

Flk-1 (kinase insert domain-containing receptor, vascular endothelial growthfactor receptor 2) heterozygote mice with lacZ-tagged endothelial cells(Flk-1wt/lacZ) (19), C57BL/6 mice, and NOD-SCID mice were obtained from TheJackson Laboratory (Bar Harbor, ME). To identify the lacZ insert in the Flk-1gene, the offspring of Flk-1wt/lacZ mice (stock no. 002938; background strainC57BL/6) were genotyped by PCR using the following primers to detect theLacZ insertion (5� primer: ATC CTC TGC ATG GTC AGG TC; 3� primer: CGTGGC CGT ATT CAT TTC) and the wild-type locus (5� primer: CAA ATG TTGCTT GTC TGG TG; 3� primer: GTC AGT CGA GTG CAC ATG TT). Mouse isletsfrom Flk-1wt/lacZ donors and human islets were transplanted into an immuno-deficient NOD-SCID mouse model from The Jackson Laboratory (for addi-tional information see http://jaxmice.jax.org/jaxmice-cgi/jaxmicedb.cgi?objtype�pricedetail&stock�001303).Mouse islet isolation. Islets were isolated from Flk-1wt/wt mice by dissectionof the splenic portion of the pancreas followed by collagenase P digestion(Roche Molecular Biochemicals, Indianapolis, IN) as previously described(20). To increase the yield of islets isolated from Flk-1wt/lacZ mice, 3 mlcollagenase P in Hank’s buffered saline (0.6 mg/ml) was first directly infusedinto the pancreas through the bile duct. Groups of two pancreata were thendigested in 6.7 ml collagenase P (0.6 mg/ml) for 4–5 min at 37°C using awrist-action shaker. Islets were handpicked under microscopic guidance andwashed three times with 10 mmol/l PBS containing 1% mouse serum. Finally,200–240 islets were suspended in 30 �l of the same solution and transplantedinto mouse recipients immediately following the isolation procedure.Human islets. Human islets were obtained through the Juvenile DiabetesFoundation Human Islet Distribution Program and from Dr. David M. Harlanat the Transplantation and Autoimmunity Branch of the National Institute ofDiabetes and Digestive and Kidney Diseases of the National Institutes ofHealth. After isolation, human islets were shipped in CMRL media byovernight courier to Vanderbilt and cultured for additional 24 h in CMRLmedia, 95% CO2/5% O2 at 37°C. After culture, 500–2,000 islets were trans-planted into NOD-SCID mice (�48 h after islet isolation).Islet transplantation model. The mice were anesthetized by an intraperi-toneal injection of sodium pentobarbital (50 mg/kg body wt; Abbott Labora-tories, North Chicago, IL). After adequate anesthesia, the left flank wasshaved, prepped, and draped in sterile fashion. With a left flank incision, theleft kidney was identified, exposed, and irrigated with saline. The isletsuspension (30 �l) was injected between the capsule and renal parenchyma ofthe left kidney using a 23-gauge butterfly needle. After withdrawal of theneedle, the insertion point was cauterized and the wound was closed withsubcutaneous sutures (Prolene, size 7-0 with cutting needle; Ethicon, Somer-ville, NJ) and skin staples (Autoclips, 9-mm size; Clay Adams, Parsippany, NJ).Tissue collection. Adult pancreata, as well as kidneys bearing islet trans-plants (3–5 weeks after transplantation), were dissected in ice-cold 10 mmol/lPBS and fixed in freshly prepared 4% paraformaldehyde (Electron MicroscopySciences, Hatfield, PA)/100 mmol/l PBS for 1.5 h on ice. Following fixation, thetissues were washed 4–6 times with 100 mmol/l PBS over a period of 2 h andthen equilibrated in 30% sucrose/10 mmol/l PBS overnight at 4°C. The tissueswere cryopreserved in optimum cutting temperature compound (VWR Scien-tific Products, Willard, OH) at �80°C, and 10- or 60-�m sections weremounted on charged slides.Detection of �-galactosidase activity. In the case of the whole pancreas,the fixed tissue was permeabilized twice for 30 min at room temperature inpermeabilization solution (2 mmol/l MgCl2, 0.01% sodium deoxycholate, and0.02% Nonidet P-40 in 10 mmol/l PBS). �-Galactosidase activity was detectedby incubating the tissue in staining solution (2 mmol/l MgCl2, 5 mmol/l Kferricyanide, 5 mmol/l K ferrocyanide, 100 mmol/l Tris, pH 7.3, and 1 mg/mlX-gal [5-bromo-4-chloro-3-indolyl-�-D-galactopyranoside; Research ProductsInternational, Mt. Prospect, IL]) overnight in the dark at room temperature.Tissue was then rinsed with 10 mmol/l PBS, postfixed in fresh ice-cold 4%paraformaldehyde/10 mmol/l PBS for 1 h at 4°C, and rinsed three times with10 mmol/l PBS. Whole-mount images of the pancreas were collected on anOlympus SZX9 microscope with an Olympus pm-C35 camera using KodakElite Chrome 160T film.

�-Galactosidase activity was also detected on 10-�m cryosections preparedas described above. The cryosections were postfixed with 0.2% glutaralde-hyde/1% paraformaldehyde (Electron Microscopy Sciences) for 15 min atroom temperature, washed three times for 5 min with 10 mmol/l PBS, andpermeabilized with permeabilization solution for 10 min at room temperature.Sections were incubated with X-gal staining solution in a humidified chamberovernight at 37°C, rinsed three times with 10 mmol/l PBS, and mounted withAquaPoly/Mount (Polysciences, Warrington, PA).Immunocytochemistry. Ten-micron cryosections were permeabilized in0.2% Triton X-100 for 10 min at room temperature, blocked with 5% normaldonkey serum (Jackson ImmunoResearch Laboratories, West Grove, PA) for

1.5 h, and then incubated with primary antibodies overnight at 4°C. Secondaryantibodies were applied to the tissue sections for 1 h at room temperature.Both primary and secondary antibodies were diluted in 10 mmol/l PBScontaining 1% BSA and 0.1% Triton X-100. Digital images of the 10-�mcryosections mounted with AquaPoly/Mount were acquired with a MagnaFiredigital camera (Optronics, Goleta, CA) connected to an Olympus BX-41fluorescence microscope.

Freshly isolated islets were attached to MatTek dishes (cat. No. P35G-0-14-C; MatTek, Ashland, MA) precoated with Cell-Tak adhesive (BectonDickinson Labware, Bedford, MA) and fixed in 4% paraformaldehyde/10mmol/l PBS for 25 min at room temperature. The fixation was followed bythree 30-min washes in 10 mmol/l PBS and 3-h permeabilization with 0.3%Triton X-100/10 mmol/l PBS. The islets were blocked with 5% normal donkeyserum/0.15% Triton X-100/10 mmol/l PBS overnight at 4°C and then equili-brated in antibody dilution buffer twice for 20 min at room temperature. Theprimary and secondary antibodies were diluted in 1% BSA/0.2% TritonX-100/10 mmol/l PBS, and the incubations were carried out for 24 h at 4°C.Sixty-micron cryosections of the islet transplants under the kidney capsulewere permeabilized, blocked, and stained under the same conditions asisolated islets. The islets and tissue sections were mounted with AquaPoly/Mount (Polysciences). Samples were subjected to optical sectioning using aZeiss LSM410 or LSM510 META confocal laser scanning microscope. Digitalimages were analyzed and three-dimensionally reconstructed using Meta-Morph 5.0 software (Universal Imaging, Downington, PA).Antibodies. Rat anti-mouse CD31 (platelet endothelial cell adhesion mole-cule [PECAM]-1) (1:200) and mouse anti-human CD31 (1:50) monoclonalantibodies were from BD Biosciences Pharmigen (San Diego, CA). Guinea piganti-human insulin IgG (1:1,000) was from Linco Research (St. Charles, MO),sheep anti-somatostatin IgG (1:1,000) was from American Research Products(Belmont, MA), rabbit anti–Flk-1 IgG (1:500) was a gift from Rolf Brekken atThe Hope Heart Institute, and rabbit anti–�-galactosidase IgG (1:5,000) wasfrom ICN Pharmaceuticals (Costa Mesa, CA). The antigens were visualizedusing appropriate secondary antibodies conjugated with Cy2, Cy3, and Cy5fluorophores from Jackson ImmunoResearch Laboratories. Secondary anti-bodies were used at concentrations recommended by the manufacturer.Lectin infusion and graft assessment. At 3 to 5 weeks after transplantation,mice transplanted with Flk-1wt/lacZ islets were anesthetized with sodiumpentobarbital (80 mg/kg body wt; Abbott Laboratories). Fluorescein isothio-cyanate–conjugated tomato lectin (Lycopersicon Esculentum, 1 mg/ml; Vec-tor Laboratories) was injected into the jugular vein (0.1 ml/mouse) andallowed to circulate for 3 min, after which animals were killed. Kidneysbearing islet transplants were dissected and preserved for cryosectioning asdescribed above. The permeabilization step was omitted because detergentdestabilizes lectin binding. Ten-micron cryosections were blocked with 5%normal donkey serum/10 mmol/l PBS/1 mmol/l Ca2� for 15 min. Incubationswith primary and secondary antibodies were carried out at room temperaturefor 1 h and 30 min, respectively. Both primary and secondary antibodies werediluted in 10 mmol/l PBS containing 1% normal donkey serum and 1 mmol/lCa2�. Washes were performed using the same buffer. Sections were mountedwith Vectashield mounting medium (Vector Laboratories) and subjected tooptical sectioning using a LSM510 META confocal laser scanning microscope.

RESULTS

Expression of endothelial cell markers in isolated

islets and pancreas of Flk-1wt/lacZ mice. Even thoughislet isolation severs arterial and venous connections,isolated islets retain their capillary network (Fig. 1A).Therefore, we asked whether these intraislet endothelialcells contribute to the revascularization of transplantedislets. To follow the fate of the intraislet endothelial cells,we used a model in which endothelial cells are tagged withlacZ (knock-in of lacZ to the Flk-1 locus termed Flk-1wt/lacZ).LacZ encodes the �-galactosidase enzyme. Figure 1B showsprominent X-gal staining (reflecting �-galactosidase activity)of islets in the whole-mount Flk-1wt/lacZ pancreas. Pancreaticsections in Fig. 1C and D demonstrate that lacZ expressionrecapitulates expression of Flk-1. Similar to Flk-1 expression,there was a greater density of lacZ� capillary structures inthe islets compared with exocrine tissue, reflecting thehigher vascularity of islets (Fig. 1C and D).Role of intraislet endothelial cells in the revascular-

ization process of transplanted mouse islets. Immedi-

M. BRISSOVA AND ASSOCIATES

DIABETES, VOL. 53, MAY 2004 1319

ately after islet isolation, wild-type islets (Flk-1wt/wt) weretransplanted into wild-type animals (Flk-1wt/wt, n � 3)(Fig. 1E–G) and mice that contain the endothelial lacZ tag(Flk-1wt/lacZ, n � 4) (Fig. 1H–J). At the same time, murineislets containing tagged endothelial cells (Flk-1wt/lacZ)were transplanted into mice that do not carry the endo-thelial cell marker (n � 10) (Fig. 1K–M). We used NOD-SCID mice as the recipient of Flk-1wt/lacZ islets to avoid apossible immune reaction to endothelial cells expressinglacZ.

Kidneys bearing transplanted islets were retrieved 3–5weeks after transplantation, a time in which the revascu-larization is completed, as shown by previous studies(9–16). In all three types of transplants, the islets werevascularized as detected by Flk-1 expression in vascularstructures within and surrounding the islet graft (Fig. 1E,H, and K). As shown by detection of �-galactosidaseactivity (Fig. 1J and M), both donor and recipient endo-thelial cells were found within the islet graft area positive

for insulin (Fig. 1I and L). Occasionally, there were a fewlacZ� intraislet endothelial cells migrating further awayfrom the transplant into kidney cortex (Fig. 1M). Thesedata indicate that intraislet endothelial cells survive andpossibly contribute to the revascularization process.

To determine whether intraislet endothelial cells partic-ipate in revascularization and to access the structure andcomposition of blood vessels in the revascularized grafts,60-�m sections of the islet grafts were labeled for themouse endothelial marker PECAM-1, which is ubiqui-tously expressed on the surface of all (both donor andrecipient) endothelial cells. The sections were colabeledfor lacZ-encoded �-galactosidase, which is only expressedby the endothelium of donor Flk-1wt/lacZ islets (Fig. 1K, L,and M) and unlike PECAM-1 has a more cytoplasmiclocalization. In both mouse and human native islets andislet grafts, mouse PECAM-1 or human CD31 and Flk-1 arecoexpressed in islet microvasculature (data not shown).Mounted sections were then subjected to optical section-

FIG. 1. Expression of endothelial cell mark-ers in isolated islets, pancreas, and revas-cularized mouse islet grafts. A: Freshlyisolated islet from a Flk-1wt/wt mouse immu-nostained [Flk-1 (red), somatostatin (blue)]and optically sectioned. Three-dimensionalreconstruction was applied to a series of 90optical sections (40� magnification). B:X-gal staining in the whole mount Flk-1wt/lacZ

pancreas as viewed by light microscopy (ar-rowheads point to islets, 40� magnification).C and D: Detection of �-galactosidase activ-ity and endogenous Flk-1 expression in con-secutive 10-�m sections of Flk-1wt/lacZ

pancreas (20� magnification). E, F, and G:Flk-1wt/wt islets transplanted into Flk-1wt/wt

recipients. H, I, and J: Flk-1wt/wt islets trans-planted into Flk-1wt/lacZ recipients. K, L, andM: Flk-1wt/lacZ islets transplanted into NOD-SCID recipients. The kidneys bearing theislet grafts were retrieved 3–5 weeks aftertransplantation. Insulin (Ins, green), Flk-1(red). Dashed lines in panels G, J, and M

show a boundary between graft and kidneycortex (kidney cortex below dashed line).Panels E, H, K at 40� magnification, and allother panels at 20�.

INTRAISLET ENDOTHELIAL CELLS AND VASCULATURE


ing using a laser scanning confocal microscope. Thethree-dimensional reconstruction of optical sectionsthrough the islet grafts (Fig. 2) indicated the existence oftwo types of blood vessels in the revascularized islet graft:1) capillaries formed predominantly of either donor orrecipient endothelial cells directly connected to each otherand 2) chimeric blood vessels formed from a mixture ofdonor and recipient endothelial cells (online data supple-ment 1CandD [available at http://diabetes.diabetesjournals.org]). By examining optical sections of the islet grafts inthree dimensions, both donor and recipient endothelialcells were found to be components of tubular structuresconsistent with vessels that traversed throughout the isletgraft (Fig. 2E and F and online data supplement 1C and D).To estimate the contribution of donor and recipient endo-thelial cells to the graft revascularization, we used Meta-Morph software and calculated the volume of PECAM-1�

and �-galactosidase� endothelial cells in the insulin� graftarea. This calculation was applied to four different grafts,and three to six fields were examined per each graft. Thesedata suggested that as much as 40 � 3% (n � 18, range18–68%) of endothelial cells in the revascularized graftoriginated from the donor islets.Role of intraislet endothelial cells in revasculariza-

tion process of transplanted human islets. Humanislets, transplanted under the kidney capsule of NOD-SCIDmice (n � 4), were analyzed for the presence of the donorand recipient endothelial cells using species-specific anti-bodies to the endothelium-specific marker CD31. Sixty-micron sections of the human islet grafts were stained formouse CD31 (PECAM-1) and human CD31. Mounted sec-tions were then subjected to optical sectioning using alaser scanning confocal microscope. The three-dimen-sional reconstruction of optical sections through the islet

FIG. 2. Three-dimensional reconstruction ofvasculature within the revascularized graft ofFlk-1wt/lacZ islets transplanted into NOD-SCIDrecipient. Sixty-micron sections of the kidneysbearing islet grafts were optically sectionedand three-dimensionally reconstructed. Bothdonor and recipient endothelial cells were la-beled for PECAM-1 (green), and �-galactosi-dase antibody was used to visualize donor-specific endothelial cells (red, [�-gal]). Isletgraft was identified by staining for insulin(blue, Ins). A–F (F is an enlargement of thehighlighted area in E): Representation of twodifferent fields of view in the same islet graft,40� magnification. D: A merge of the images inA–C. E and F: Only endothelial cell markers areshown. The dashed line in A–E shows a bound-ary between graft and kidney cortex (kidneycortex is to the right of dashed line). Whitearrows point to blood vessels derived fromeither donor or recipient endothelial cells, andmagenta arrows point to chimeric blood ves-sels. (See also online data supplement 1A–D.)



grafts (Fig. 3) indicated that human intraislet endothelialcells, similar to the murine islets, survive after transplan-tation. In contrast to mouse intraislet endothelial cells, thehuman intraislet endothelial cells did not form chimericblood vessels. Moreover, examination of human isletgrafts in three dimensions revealed fewer possible connec-tions of human endothelial cells with the host vasculature(Fig. 3D and E) than was seen in mouse islet grafts. Thehuman intraislet endothelial cells also appeared to morereadily migrate into the kidney cortex (Fig. 3F) than wasobserved in transplants of murine Flk-1wt/lacZ islets (Fig.1M).Do intraislet endothelial cells become a part of func-

tional vasculature in revascularized graft? Even

though intraislet endothelial cells survived transplantationand were a part of vascular-like structures within therevascularized islet grafts, it was not known whether theyintegrated into a functional blood vessel network of therevascularized graft. To address this question, mice thatreceived transplants of Flk-1wt/lacZ islets were infused withfluorescein isothiocyanate–conjugated tomato lectin (n �4), which binds with high affinity to the surface of mouseendothelial cells (21). Lectin infusion, a well-establishedtechnique, has been widely used in the area of angiogen-esis (22). For example, many investigators (22) feel that itis the technique of choice for identifying functional bloodvessels since the only way lectin can reach the endothelialcells is via the circulation. Furthermore, it has been shown

FIG. 3. Three-dimensional reconstructionof vasculature within the revascularizedgraft of human islets transplanted intoa NOD-SCID recipient. Sixty-micronsections of the kidneys bearing isletgrafts were optically sectioned andthree-dimensionally reconstructed. Do-nor and recipient endothelial cells werelabeled with species-specific antibodiesto CD31; mouse CD31 (green, mPE-CAM-1), human CD31 (red, hCD31),and insulin (blue), 60� magnification.D: A merge of the images in A–C. E: Anenlargement of the area highlighted inD without insulin staining. The dashedline in A–D shows a boundary betweengraft and kidney cortex (kidney cortexis above dashed line). The arrows in D

and E point to possible vascular connec-tions of human endothelial cells withthe host vasculature revealed by exam-ination of human islet grafts in threedimensions.



previously that L. Esculentum lectin, which is utilized inour studies, binds uniformly to the luminal surface ofmurine endothelial cells in arteries, veins, and capillariesand is colocalized with CD31 in functional blood vessels(22). We confirmed the uniformity of L. Esculentum lectinbinding in both pancreas and kidney. Optical sections ofthe lectin-perfused mouse tissue specimens acquired witha laser scanning confocal microscope show that the lectinsignal is localized in the luminal lining of blood vessels(online data supplement 2A and B).

The lectin-perfused islet grafts were additionally ana-lyzed for expression of insulin and �-galactosidase byimmunocytochemistry and laser scanning confocal mi-croscopy (Fig. 4) (online data supplement 3A–F). Theseries of optical sections was acquired at a 0.5-�m intervalin the axial (z) dimension and an appropriate pinholesetting to alleviate any concerns about overlapping signalsfrom the layers of the specimen above and below the givenfocal plane. Individual optical sections of the islet graftsdemonstrate that in the same focal plane, �-galactosidase,which marks intraislet endothelial cells, is colocalizedwith lectin, thus proving unequivocally the functionality ofthese �-galactosidase–positive blood vessels. The pres-ence of cells double positive for lectin and �-galactosidaseindeed proves that blood flows through donor-derivedblood vessels in revascularized mouse islet grafts (Fig. 4B

and D). Immunocytochemistry of the lectin-perfused spec-imens required several modifications to avoid lectin leach-ing during the staining procedure (see RESEARCH DESIGN AND

METHODS) and underestimated the number of �-galactosi-

dase� endothelial cells (especially in cells with lower�-galactosidase expression). This was based on a compar-ison of �-galactosidase� cells in consecutive sectionsstained with either �-galactosidase antibody (modifiedimmunocytochemistry procedure) or using an enzymaticreaction of �-galactosidase with X-gal substrate. Becausethere were fewer �-galactosidase� cells detected by im-munocytochemistry compared with enzymatic reaction,we did not feel it was appropriate to estimate the ratio of�-galactosidase� cells to cells double positive for �-galac-tosidase� and lectin�. We did find a few cells positive for�-galactosidase and negative for lectin, and, for example,one such cell appeared in Fig. 4B (arrow) (online datasupplement 3C and F). These �-galactosidase�/lectin�

cells could be either intraislet endothelial cells that sur-vived the transplantation but did not establish a lumen orproliferating donor-derived endothelial cells.

In the human islet grafts perfused with L. Esculentum

lectin, we did not find human endothelial cells doublepositive for human CD31 and lectin. While binding of L.

Esculentum lectin to mouse endothelium is well docu-mented (22), it is unknown if this lectin binds to humanendothelial cells. The human CD31� cells remained nega-tive for the lectin even when the sections of human isletgrafts were subsequently stained with L. Esculentum

lectin (a standard method to assess if a lectin binds toendothelial cells). These observations indicate that humanintraislet endothelial cells may lack the N-acetyl-D-glu-cosamine olygomer moieties that are recognized by L.

Esculentum lectin. The formation of tubular structures by

FIG. 4. Intraislet endothelial cells be-come a functional part of revascularizedgraft. NOD-SCID mice bearing Flk-1wt/lacZ

islet grafts were infused with endotheli-um-binding fluorescein isothiocyanate–conjugated tomato lectin (red, lectin) 5weeks after transplantation. Subse-quently, the sections were labeled withantibodies to insulin (green, Ins) and�-galactosidase (blue, �-gal). Ten-mi-cron sections of the kidneys bearingislet grafts were optically sectioned andthree-dimensionally reconstructed. A

and B: 40� magnification. C and D: 60�magnification. Endothelial cells doublepositive for lectin and �-galactosidase(purple) in B and D indicate blood flowthrough donor-derived capillaries. (Seealso online data supplement 3A–F.)



human endothelial cells is evident from the three-dimen-sional projections in Fig. 3; however, it is still possible thatthey were incompletely connected to the recipient’s vas-culature.

DISCUSSION

Transplanted pancreatic islets, unlike transplanted solidorgans, must revascularize to survive and function, andthis process of angiogenesis is likely an important factorthat ultimately determines whether an islet transplantsecretes sufficient insulin to reverse diabetes. The reasonsfor the current requirement of islets from two pancreatafor islet transplantation to reverse diabetes are uncertain,but this limitation is a major obstacle to advancing islettransplantation. While continued focus on improving im-munosuppression after islet transplantation is crucial, abetter understanding of the molecular events of isletengraftment or revascularization could lead to a dramaticincrease in islet survival and improve islet transplantationas a therapy for type 1 diabetes. In contrast to the currentthinking that revascularization is solely a function of thetransplant recipient (10,13,16–18), our results show thatintraislet endothelial cells from the donor survive aftertransplantation and play an important role in the revascu-larization process. The survival of intraislet endothelialcells was recently reported by investigators (19) using thetransgenic Tie2-lacZ mouse model, but this report did notexamine whether these cells can contribute to functionalvasculature.

Three-dimensional reconstructions of optical sectionsthrough islet grafts allowed a more detailed examinationof blood vessels within the revascularized mouse andhuman islet grafts. In mouse Flk-1wt/lacZ islet transplants,donor and recipient endothelial cells cooperated effec-tively during the process of angiogenesis in that somevessels were lined with donor or recipient endothelialcells, while other vessels were a chimera of donor andrecipient endothelial cells. In contrast, there was lessinteraction between human intraislet endothelial cells andrecipient mouse endothelial cells. Examination of opticalsections in three dimensions did not demonstrate chimericblood vessels in the human islet grafts and revealed fewerpossible connections of mouse and human endothelialcells. The limited interaction of mouse and human endo-thelial cells may reflect species differences in endothelialcells. This may also explain the relatively extensive migra-tion of human endothelial cells into the kidney cortex,which was only occasionally found in mouse Flk-1wt/lacZ

islet transplants. In addition, human islets (unlike those ofmice) were procured from brain-dead donors and culturedlonger, which may have influenced their intraislet endo-thelial cell biology.

Even though the presence of vascular connections be-tween capillaries derived from donor and recipient endo-thelial cells seemed possible based on three-dimensionalprojections of islet grafts, it does not prove that thesevessel-like structures are functional. Infusion of endothe-lium-binding lectin in the mouse islet–transplant modeldemonstrated that blood flows through donor-derivedcapillaries and, thus for the first time, showed that theintraislet endothelial cells have a capacity to integrate intothe functional vasculature within the revascularized graft.

These results have several implications for the handling ofpancreatic islets destined for transplantation. Pancreasprocurement and islet isolation procedures must not onlystrive to maximize the health of islet endocrine cells, butmust now also consider the effects on endothelial cellhealth and survival. In contrast to the original Edmontonprocedure, many human islet transplantation centers cur-rently culture pancreatic islets for several days beforetransplantation; the effect of such culture on endothelialcell health and survival after transplantation is unknownand warrants further study. The model system in whichintraislet endothelial cells are tagged with lacZ shouldallow one to address such questions. In addition, ourtransplantation studies with both mouse and human isletsdid not utilize immunosuppression; the effect of immuno-suppressive agents on endothelial cell health and isletrevascularization is not known and warrants further study.

Since islet revascularization may be a limiting factor inislet survival and since transplanted islets have reducedvessel density compared with islets in the pancreas(10,13,23), the capacity of intraislet endothelial cells tointegrate into functional graft vasculature suggests a num-ber of avenues of investigation to enhance the revascular-ization process. For example, can intraislet endothelialcells be activated or primed ex vivo (pretransplantation)to accelerate angiogenesis? Will inclusion of exogenousendothelial cells with transplanted islets promote isletengraftment and revascularization? Emerging evidencesupports a direct interaction between pancreatic islet cellsand endothelial cells. Melton et al. (24) recently suggestedthat the developing aorta was essential for the initiation ofendocrine cell differentiation during pancreas develop-ment. These investigators also found that transgenic over-expression of the angiogenic factor, vascular endothelialgrowth factor (VEGF), under the control of a pancreas-specific promoter from the pancreatic duodenal ho-meobox-1 (PDX-1) gene increased islet size. Whencombined with the current results, this raises the possibil-ity that intraislet endothelial cells and the newly formedislet vasculature may nurture transplanted islet cells inaddition to the reestablishment of islet blood flow. Futurestudies are needed to address this hypothesis.

Because of the shortage of organ donors and the largenumber of patients with type 1 diabetes, investigators aredeveloping alternative sources of insulin-producing cellssuch as stem cells and genetically engineered cells. Liketransplanted pancreatic islets, such cell transplants willlikely need to become vascularized. Our studies wouldsuggest that inclusion of exogenous endothelial cells withsuch insulin-producing cells may facilitate the survival andfunction of these surrogate islet cells. A better understand-ing of how transplanted cells survive and become vascu-larized will also be relevant to the development of celltransplantation therapy for other diseases.

Our studies also give rise to a number of questions aboutthe molecular events of islet engraftment. For example, wetransplanted islets beneath the renal capsule, whereas inhuman transplantation, islets are infused into the portalvein and embolize in the liver. Whether the angiogenic andrevascularization processes are identical in the liver andbeneath the renal capsule is unknown. What signals initi-ate angiogenesis after islet transplantation? Are islet cells



the source of such signals? What are the steps of isletengraftment and capillary sprouting, from breakdown ofthe basement membrane of preexisting vessels within andsurrounding the islet graft to endothelial cell migration?The current findings and future research to better under-stand the molecular events of islet engraftment shouldenhance islet revascularization, improve islet survival af-ter transplantation, reduce the number of pancreatic isletsrequired to reverse diabetes, and improve the outcome ofislet transplantation in type 1 diabetes.

ACKNOWLEDGMENTS

This study was supported by a Merit Review Award fromthe VA Research Service, research grants from the Na-tional Institutes of Health and the Juvenile Diabetes Re-search Foundation International, and the VanderbiltDiabetes Research and Training Center (National Insti-tutes of Health DK20593).

The authors thank Drs. David M. Harlan and BoazHirshberg at the National Institutes of Health for providingsome of the human pancreatic islets and for helpfuldiscussions. We also thank Ninche Alston for technicalassistance.

REFERENCES

1. Biancone L, Ricordi C: Pancreatic islet transplantation: an update. Cell

Transplant 11:309–311, 20022. Robertson RP: Successful islet transplantation for patients with diabetes:

fact or fantasy? N Engl J Med 343:289–290, 20003. Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E, Warnock GL,

Kneteman NM, Rajotte RV: Islet transplantation in seven patients with type1 diabetes mellitus using a glucocorticoid-free immunosuppressive regi-men. N Engl J Med 343:230–238, 2000

4. Ryan EA, Lakey JR, Rajotte RV, Korbutt GS, Kin T, Imes S, Rabinovitch A,Elliott JF, Bigam D, Kneteman NM, Warnock GL, Larsen I, Shapiro AM:Clinical outcomes and insulin secretion after islet transplantation with theEdmonton protocol. Diabetes 50:710–719, 2001

5. Davalli AM, Ogawa Y, Scaglia L, Wu YJ, Hollister J, Bonner-Weir S, WeirGC: Function, mass, and replication of porcine and rat islets transplantedinto diabetic nude mice. Diabetes 44:104–111, 1995

6. Davalli AM, Scaglia L, Zangen DH, Hollister J, Bonner-Weir S, Weir GC:Vulnerability of islets in the immediate posttransplantation period:dynamic changes in structure and function. Diabetes 45:1161–1167,1996

7. Davalli AM, Scaglia L, Zangen DH, Hollister J, Bonner-Weir S, Weir GC:Early changes in syngeneic islet grafts: effect of recipient’s metaboliccontrol on graft outcome. Transplant Proc 27:3238–3239, 1995

8. Carlsson PO, Andersson A, Carlsson C, Hellerstrom C, Hoglund E, King A,Kallskog O, Liss P, Mattsson G, Olsson R, Palm F, Sandler S, Tyrberg B,Jansson L: Engraftment and growth of transplanted pancreatic islets. Ups

J Med Sci 105:107–123, 20009. Griffith RC, Scharp DW, Hartman BK, Ballinger WF, Lacy PE: A morpho-

logic study of intrahepatic portal-vein islet isografts. Diabetes 26:201–214,1977

10. Jansson L, Carlsson PO: Graft vascular function after transplantation ofpancreatic islets. Diabetologia 45:749–763, 2002

11. Lukinius A, Jansson L, Korsgren O: Ultrastructural evidence for bloodmicrovessels devoid of an endothelial cell lining in transplanted pancreaticislets. Am J Pathol 146:429–435, 1995

12. Menger MD, Jaeger S, Walter P, Feifel G, Hammersen F, Messmer K:Angiogenesis and hemodynamics of microvasculature of transplantedislets of Langerhans. Diabetes 38:199–201, 1989

13. Menger MD, Yamauchi J, Vollmar B: Revascularization and microcircu-lation of freely grafted islets of Langerhans. World J Surg 25:509 –515,2001

14. Rooth P, Dawidson I, Lafferty K, Diller K, Armstrong J, Pratt P, SimonsenR, Taljedal IB: Prevention of detrimental effect of cyclosporin A onvascular ingrowth of transplanted pancreatic islets with verapamil. Dia-

betes 38:202–205, 198915. Vajkoczy P, Menger MD, Simpson E, Messmer K: Angiogenesis and

vascularization of murine pancreatic islet isografts. Transplantation 60:123–127, 1995

16. Vajkoczy P, Olofsson AM, Lehr HA, Leiderer R, Hammersen F, Arfors KE,Menger MD: Histogenesis and ultrastructure of pancreatic islet graftmicrovasculature: evidence for graft revascularization by endothelial cellsof host origin. Am J Pathol 146:1397–1405, 1995

17. Carlsson P-O, Palm F, Mattsson G: Low revascularization of experimen-tally transplanted human pancreatic islets. J Clin Endocrinol Metab

87:5418–5423, 200218. Hirshberg B, Mog S, Patterson N, Leconte J, Harlan DM: Histopathological

study of intrahepatic islets transplanted in the nonhuman primate modelusing Edmonton protocol immunosuppression. J Clin Endocrinol Metab

87:5424–5429, 200219. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breitman

ML, Schuh AC: Failure of blood-island formation and vasculogenesis inFlk-1-deficient mice. Nature 376:62–66, 1995

20. Brissova M, Shiota M, Nicholson W, Gannon M, Knobel S, Piston D, WrightC, Powers AC: Reduction in transcription factor pdx-1 impairs normalglucose sensing and insulin secretion by pancreatic islets. J Biol Chem

277:11225–11232, 200221. Inoue M, Hager JH, Ferrara N, Gerber HP, Hanahan D: VEGF-A has a

critical, nonredundant role in angiogenic switching and pancreatic betacell carcinogenesis. Cancer Cell 1:193–202, 2002

22. McDonald DM, Choyke PL: Imaging of angiogenesis: from microscope toclinic. Nat Med 9:713–725, 2003

23. Carlsson PO, Mattsson G: Oxygen tension and blood flow in relation torevascularization in transplanted adult and fetal rat pancreatic islets. Cell

Transplant 11:813–820, 200224. Lammert E, Cleaver O, Melton D: Induction of pancreatic differentiation by

signals from blood vessels. Science 294:564–567, 2001



Intraislet Endothelial Cells Contribute to ... · Intraislet Endothelial Cells Contribute to...

Documents

Transcript of Intraislet Endothelial Cells Contribute to ... · Intraislet Endothelial Cells Contribute to...