Role of Donor-Derived Monocyte Chemoattractant Protein-1 in Murine Islet Transplantation

Bernd Schro¨ppel,*† Nan Zhang,†‡ Peng Chen,* Dongmei Chen,‡ Jonathan S. Bromberg,†‡ and Barbara Murphy*† *Division of Nephrology, †Recanati/Miller Transplantation Institute, and ‡Department of Gene and Cell Medicine, Mount Sinai School of Medicine, New York, New York

Monocyte chemoattractant protein-1 (CCL2/MCP-1) is a proinflammatory chemokine produced by several cell types, including pancreatic islets. High levels of donor-derived CCL2 have been associated with poor islet allograft outcome in patients with type 1 diabetes; however, the causal relationship is unknown. The constitutive and inducible expression of chemokines and their receptors by pancreatic islets in vitro were investigated, specifically the role of donor-derived CCL2 in marginal mass murine islet transplantation. The results showed that inflammatory cytokine stimulation of islets induced de novo expression of CCL2, CCL5/RANTES, CXCL9/MIG, and CXCL10/IP-10 and increased expression of CXCL2/macrophage-inflammatory protein-2. CCL2 mRNA and protein were highly expressed within2dincultures. Transplantation of islets with high levels /of CCL2 into syngeneic recipients led to a significantly greater influx of CCR2؉ cells and higher expression of monocyte macrophage-associated inflammatory cytokines compared with low CCL2-donor islets. The level of pretransplantation CCL2 .inversely correlated (P < 0.0001) with isograft function. In contrast, in CCR2؊/؊ recipients, this correlation was not present A direct toxic effect of CCL2 on islets was excluded by assessing cell viability and insulin release in vitro. In conclusion, CCL2 secreted by islets plays an important role in the immediate islet graft function. Strategies to decrease islet-derived CCL2 release may increase the success of islet transplantation. J Am Soc Nephrol 16: 444-451, 2005. doi: 10.1681/ASN.2004090743

n the immediate period after islet transplantation, the signal monocytes, natural killer, and activated and memory T marginal islet mass is further compromised by ␤ cell dam- cells to migrate to sites of injury as part of the inflammatory I age caused by ischemia-reperfusion and the inflammatory response (13,14). response. In addition to adhesion molecules, such as selectins We hypothesized that the local generation of chemokines by and integrins, the complex process of extravasation and influx islet cells may be important in the initiation and regulation of of leukocyte subsets into the site of tissue injury is mediated, to inflammatory processes during insulitis and islet graft destruc- a significant extent, by the expression of specific chemokines tion. This hypothesis is supported by recent data that demon- and chemokine receptors (1). Monocyte chemoattractant pro- strated that high levels of CCL2 released by human islet cells tein-1 (CCL2/MCP-1) is a proinflammatory chemokine pro- may play an important role in the clinical outcome of islet duced by monocytes/macrophages and lymphocytes in addi- allografts in patients with type 1 diabetes, because high levels tion to many other cell types, including endothelial cells, of donor-derived CCL2 secretion were associated with poor mesangial cells, renal tubular cells, and fibroblasts, in response islet graft outcome (6). Despite recent publications on CCL2 to cytokines, oxidized LDL, and LPS (2–4). Several studies have secretion by pancreatic islets (5–11), no mechanistic data out- now shown that CCL2 is also strongly expressed by human and lining the role of donor-derived CCL2 are available. In the rodent pancreatic islets under various conditions (5–11). The present study, we specifically investigated the role of donor- biologic actions of CCL2 are mainly mediated through the derived CCL2 in murine islet transplantation. interaction with the CC chemokine receptor 2 (CCR2), the other ligands for which include CCL8/MCP-2, CCL7/MCP-3, and CCL12/MCP-5. CCR2 receptors are found on the surface of Materials and Methods inflammatory mononuclear cells, endothelial cells, and fibro- Mice and Diabetic Model blasts (12). Through binding to the CCR2 receptor, its ligands Animals were treated in strict compliance with regulations established by the Institutional Animal Care and Use Committee. Mice were born and housed under specific pathogen-free conditions. The recipients were rendered diabetic by a single intraperitoneal injection of 180 Received September 8, 2004. Accepted November 10, 2004. mg/kg streptozotocin (Sigma, St. Louis, MO) and considered diabetic Published online ahead of print. Publication date available at www.jasn.org. when the tail-vein blood glucose concentration as determined by One- Touch glucometer (Lifescan, Milpitas, CA) was Ͼ300 mg/dl for 2 Address correspondence to: Dr. Bernd Schro¨ppel, Division of Nephrology, consecutive days. Inbred C57BL/6 (H-2b) mice (Jackson Laboratories, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1243, New York, NY 10029-6574. Phone: 212-241-8004; Fax: 212-987-0389; E-mail: bernd.schroppel@ Bar Harbor, ME) were used at 10 to 12 wk of age (25 to 30 g). The CCR2 mssm.edu gene knockout mice (background C57BL/6; F10) were provided by

W.A. Kuziel (Department of Microbiology, Institute for Cellular and sense 5Ј-GATGAATTGGCGTGGAATCT-3Ј. All runs included a control Molecular Biology, University of Texas, Austin, TX), and deletion was for genomic DNA contamination, where the reverse transcriptase was confirmed by genotyping (15). omitted in the sample. The PCR reactions were cycled 40 times after initial denaturation (95°C, 15 min) with the following parameters: denaturation Islet Isolation, Culture, and Transplantation at 94°C for 15s, annealing at 56°C for 20 s, and extension at 72°C for 15 s Islet isolation and their culture and transplantation were previously with temperature transition rates of 20°C/s. Control reactions for product described in detail (5). Briefly, 3 ml of cold Hank’s buffer/collagenase identification consisted of analyzing the melting peaks (°C) and determin- P solution (1.5 mg/ml; Roche Diagnostics, Mannheim, Germany) was ing the length of the PCR products on agarose gel. infused into the pancreatic duct in situ, and the removed pancreas was digested at 37°C for 20 min. Islets were purified on a discontinuous cDNA Array Ficoll gradient (Sigma). For marginal mass syngeneic transplantation, The relative mRNA expression of 67 chemokines and chemokine 200 handpicked islets were transplanted immediately after isolation or receptors and four different housekeeping genes was analyzed with a after 48 h of culture in serum-free medium beneath the renal capsule, nonradioactive GEArray Q series (SuperArray, Bethesda, MD), accord- and tail-vein glucose was measured daily. This represents half the ing to the manufacturer’s protocol. Two micrograms of total RNA was number of islets usually transplanted. Mice were anesthetized with reverse-transcribed into cDNA with Moloney murine leukemia virus ketamine. Posttransplantation glucose reduction was calculated as per- reverse transcriptase (Life Technologies BRL). Biotin-labeled cDNA centage of the pretransplantation value. As an indirect measure of cell probes were hybridized to chemokine and chemokine receptor gene– viability in vitro, proliferation was assessed by a modified methyl- specific cDNA fragments that were spotted on the GEArray mem- tetrazolium (MTT) dye procedure (16). The MTT staining (absorbance branes. The image was analyzed by SuperArray software. The relative at 600 nm) for ␤-TC3 cells was found to increase linearly with the cell expression level of each gene is determined by comparing the signal number from 5 ϫ 103 to 1.2 ϫ 105 cells loaded per well in a 96-well plate intensity of each gene in the array after normalization to the signal of (data not shown). All experiments were done in quintuplicate. Reagents the housekeeping gene cyclophilin A. Arbitrary units were calculated Ϫ were routinely tested for the presence of endotoxin using the Limulus with the following formula: (chemokine signal background signal)/ Ϫ Ͻ Amebocyte Lysate kit (Bio-Whittaker, Walkersville, MD) and contained (cyclophilin A signal background signal). Results of 5% were Ͻ0.1 endotoxin unit/ml. considered negative. For the cDNA array expression analysis, a twofold increase in gene expression was considered to be significant for genes ␤ with basal expression values significantly over background values, Islet and -TC3 Stimulation which was defined as Ͼ5% of the housekeeping gene. The mouse pancreatic insulinoma ␤-TC3 line (gift from Dr. S. Efrat, Albert Einstein College of Medicine, Bronx, NY), which produces both In Situ Hybridization proinsulin I and II and efficiently processes each into mature insulin in In situ hybridization was performed as described previously (5). a manner comparable to normal ␤ cells in isolated islets, were cultured Negative controls included hybridization performed on replicate tissue as described previously (17). Cells (500 islets/ml and 5 ϫ 106 ␤-TC3 sections using the sense riboprobe. cells/ml) were stimulated at 37°C for 24 h in 1 ml of fresh serum-free medium in the presence or absence of a combination of recombinant Glucose-Stimulated Insulin Secretion murine IL-1␤ (R&D Systems, Minneapolis, MN; 2 ng/ml; specific ac- Groups of 30 isolated islets or 3 ϫ 104 ␤ cells were cultured with tivity 1.1 ϫ 106 units/mg), TNF-␣ (BD PharMingen, San Diego, CA; 20 complete CMRL 1066 medium (10% FBS, 100 U/ml penicillin, 0.1 ng/ml; specific activity 1.0 ϫ 109 units/mg), and IFN-␥ (BD PharMin- mg/ml streptomycin, and 2 mM l-glutamine) that contained 5.5 or 16.7 gen; 20 ng/ml; specific activity 1.0 ϫ 108 units/mg). mM glucose in 24-well tissue culture plates (Corning, Corning, NY). Supernatants were analyzed for insulin content using a rodent-specific RNA Isolation and cDNA Synthesis insulin ELISA kit (Crystal Chem, Chicago, IL). Total RNA was extracted from 5 ϫ 106 ␤-TC3 cells, 500 islets, or islet grafts with 1 ml of phenol/guanidine isothiocyanate that contained CCL2 Protein Assay Trizol solution (Life Technologies, BRL, Grand Island, NY). The integ- Aliquots of various pancreatic islet preparations (50 islets/ml; n ϭ rity of total RNA was determined by denaturing agarose gel electro- 13) were cultured in 24-well plates, and supernatants were harvested at phoresis and ethidium bromide staining. For cDNA synthesis, 1 ␮gof different time points and stored at Ϫ80°C until assayed. CCL2 was total RNA was primed with oligo(dT) Superscript Moloney murine measured with a sandwich ELISA kit from R&D Systems according to leukemia virus reverse transcriptase (Life Technologies BRL) as de- the manufacturer’s instructions. Absorbance was measured at 450 nm. scribed previously (5). Statistical Analyses Quantitative Real-Time PCR All experiments were repeated at least three times. Groups were Kinetic PCR was performed on a LightCycler (Roche Diagnostics) by compared using two-sided t test for continuous variables. Values are Ϯ using FastStart QuantiTect SYBR Green PCR kit (Qiagen, Valencia, CA) as expressed as the mean SEM. No corrections were made for multiple a double-strand DNA-specific binding dye using the following primer comparisons. Linear regression and correlations were calculated using pairs or as described previously (5,18). CD64 sense 5Ј-GGGAAGACAC- InStat 2.01 software. CGCTACACAT-3Ј, antisense 5Ј-GGAGATGACA CGGATGCTCT-3Ј; CD68 sense 5Ј-CCAATTCAGGGTGGAAGAAA-3Ј, antisense 5Ј-ATGGG- Results TACCGTCACAACCTC-3Ј; IL-1␤ sense 5Ј-CAACCAACAAGTGATAT- Differential Expression of Chemokines and Their Receptors TCTCCATG-3Ј, antisense 5Ј-GATCCACACTCTCCAGCTGCA; TNF-␣ in ␤ Cells and Isolated Islets under Basal and Stimulated sense 5Ј-CATCTTCTCAAAATTCGAGTGACAA-3Ј, antisense: 5Ј- Conditions TGGGAGTAGACA AGGTACAACCC Ϫ3Ј; macrophage-inflammatory We first characterized the constitutive and stimulated expres- protein 1␣ (MIP-1␣) sense 5Ј-ATGAAGGTCTCCACCACTGC-3Ј, anti- sion of chemokines and chemokine receptors of islets in vitro. 446 Journal of the American Society of Nephrology J Am Soc Nephrol 16: 444-451, 2005

Table 1. Chemokine and chemokine receptor gene As islets are composed of different residential cell populations, expression in murine ␤ -TC3 cells and islets an established ␤ cell line (␤-TC3) was also studied to delineate immediately after isolation (0 h) and after 24 h in the contribution of this cell type. Immediately after the isolation culture, without the addition of cytokines, detected by and purification procedure, islets expressed the transcripts of cDNA arraya specific chemokine genes, namely CCL6, CCL7, CCL17, CCL19, CCL21, CCL28, CXCL2, CXCL5, CXCL11, and CXCL15. The Islets expression of CCL2 was induced after 24 h in culture. In addi- Gene Other Name ␤-TC3 0 h 24 h tion, most of the chemokines identified immediately after isolation were further upregulated after 24 h in culture. The same CC chemokine mRNA expression pattern was observed in ␤-TC3 CCL1 TCA3 – – – in long-term culture. Transcripts for the chemokine receptors CCL2 MCP-1 ––17CCR7 and CXCR5 were found in both ␤-TC3 and isolated islets CCL3 MIP-1␣ –––(Table 1). CCL4 MIP-1␤ –––One mechanism by which invading mononuclear cells de- CCL5 RANTES – – – stroy ␤ cells is by local production of cytokines such as TNF-␣, CCL6 C10 65 35 136 IL-1␤, and IFN-␥. To study whether proinflammatory condi- CCL7 MCP-3 54 19 151 tions result in the induction of chemokines in islets, we exposed ␣ ␤ ␥ CCL8–12 – – – cells in culture to a combination of TNF- , IL-1 , and IFN- for CCL17 TARC 44 23 69 24 h at a concentration used previously (10,11). Cytokine stim- CCL19 MIP-3␤ 19 9 28 ulation of islets led to an induction of CCL2, CCL5/RANTES CCL20 MIP-3␣ –––(regulated on activation normal T cell expressed and secreted), CCL21 SLC 7614CXCL2/MIP-2, CXCL9/MIG, and CXCL10/IP-10. With the ex- ception of CXCL2, all genes were also induced in the ␤ cell line CCL22 MDC – – – (␤-TC3), suggesting a source other than ␤ cells within the islets CCL24 Eotaxin-2 – – – for CXCL2 (Table 2). These findings were confirmed by QPCR CCL25 TECK – – – (data not shown). CCL27 CTAK – – – To determine whether CCL2 mRNA expression was associ- CCL28 MEC 12 9 13 ated with protein synthesis and release, we determined CCL2 CXC levels in culture medium by ELISA and found them to correlate CXCL1 KC – – – well with the pattern of mRNA expression. Islets that were CXCL2 MIP-2 61 25 182 cultured up to 2 d had a progressive increase in CCL2 produc- CXCL4 PF4 – – – tion (Figure 1). These in vitro data suggest that islets and ␤ cells CXCL5 LIX 14 13 62 constitutively expressed the mRNA for several chemokines and CXCL9 MIG – – – chemokine receptors, and exposure to proinflammatory cyto- CXCL10 IP-10 – – – kines upregulated additional potent chemokines, with the po- CXCL11 I-TAC 36 16 17 tential to attract cells expressing CCR2, CCR5, and CXCR3, CXCL12–14 – – – such as activated T cells and monocytes/macrophages. CXCL15 Lungkine 92 47 206 CXCL16 SR-PSOX – – – Other SCM-1 – – – Table 2. IL-1␤, IFN-␥, and TNF-␣ induced chemokine CX3CL1 Fractalkine – – – gene expression in islets and cultured ␤-TC3 cells SDF2 – – – detected by cDNA arraya CCR ␤ CCR1–6 – – – Chemokine -TC3 Islets CCR7 56 8 39 CC CCL2 MCP-1 Ͼϩ10 ϩ2 CCR8–10 – – – CCL5 RANTES Ͼϩ2 Ͼϩ2 CXCR CXC CXCL2 MIP-2 – ϩ2 CXCR1–4 – – – CXCL9 MIG Ͼϩ10 Ͼϩ10 CXCR5 788 CXCL10 IP-10 Ͼϩ10 Ͼϩ5 Other aData are presented as fold variation in gene expression CX3CR1 – – – compared with their respective controls (untreated cells). A Ն a twofold change of gene expression comparing stimulated to Data are presented as percentage of the housekeeping unstimulated cells was considered significant. When genes gene cyclophilin A. Genes in bold signify expression level Ͼ were not detectable in unstimulated controls, the fold change 5% from background. was expressed as Ͼϩ2, ϩ5, or ϩ 10 depending on expression level. J Am Soc Nephrol 16: 444-451, 2005 Chemokines and Islet Transplantation 447

levels of pretransplantation CCL2 by islets resulted in greater intragraft CCR2 mRNA expression (Figure 2). In situ hybridization localized a significant CCR2ϩ mononuclear cellular infiltrate surrounding the islet cells (Figure 3). Molecular mark- ers for leukocytes suggested increased recruitment of monocytes/macrophages (CD64, CD68) and T cells (CD3) into the CCL2-expressing donor grafts (Figure 2), with early recruitment of monocytes/macrophages on day 1 followed by late recruitment of T cells on day 4.

Intragraft Cytokine mRNA Expression Pattern Expression of seven cytokines was analyzed on days 1 and 4 posttransplantation of cultured and uncultured islets to assess their local production. There was no detectable expression of IL-2 and IFN-␥ (data not shown). Both cytokines are mainly produced by T cells and natural killer cells after their activation, Figure 1. Effect of islet cell culture on CCL2 secretion. CCL2 secretion of 13 different islet preparations immediately after suggesting that significant T cell activation does not occur isolation (0 h), 1 and2dofculture. CCL2 concentration is given within 4 d posttransplantation. In addition, there was no dif- as pg/ml supernatant and pg/islet. ference of IL-10 expression in the two groups and time point’s analyzed (data not shown). In contrast, there was evidence for production of monocyte/ High Levels of Pretransplantation CCL2 Effectively Recruit macrophage-derived cytokines. Cultured islets showed higher CCR2ϩ Monocytes/Macrophages and T Cells expression of IL-1␤, TNF-␣, TGF-␤1, and MIP-1␣ when com- To assess whether donor-derived CCL2 has the potential to pared with freshly isolated islets after transplantation (Figure attract CCR2ϩ cells in vivo, we analyzed intragraft mRNA 4). MIP-1␣ is a product of monocytes and macrophages and expression in isografts by QPCR and localized it with in situ primary stimulator of other proinflammatory cytokines such as hybridization. By using the isografts, we were able to deter- IL-1, IL-6, and TNF-␣. This is consistent with the pattern of mine recruitment independent of allogeneic stimulus. High cytokine production seen in cultured transplanted islets. These

Figure 2. Isograft CCL2, CCR2, monocyte/macrophage marker CD64 and CD68, and T cell marker CD3 mRNA expression. Grafts were analyzed by QPCR 1 and 4 d after transplantation of cultured or freshly isolated islets (*P Ͻ 0.05 freshly isolated versus cultured islets). 448 Journal of the American Society of Nephrology J Am Soc Nephrol 16: 444-451, 2005

islets grafts (Ϫ40 Ϯ 15.9 versus Ϫ55 Ϯ 7.7%; P ϭ 0.038). We then examined whether pretransplantation CCL2 levels ranging from 0.46 pg/islet (22.9 pg/ml) to 4.12 pg/islet (206 pg/ml) correlated with islet graft function. We found that the glucose reduction 1 d posttransplantation inversely correlated with the level of CCL2 secretion of islets (P Ͻ 0.0001; Figure 5). This effect was independent of the time interval analyzed (mean posttransplantation glucose reduction day 1 to 2 r ϭϪ0.923 (95% confidence interval [CI], Ϫ0.98 to Ϫ0.74), day 1 to 3 r ϭ Ϫ0.921 (95% CI, Ϫ0.98 to Ϫ0.74), and day 1 to 4 r ϭϪ0.915 (95% CI, Ϫ0.98 to Ϫ0.72). Isografts that were transplanted to Ϫ Ϫ Figure 3. CCR2 was localized by in situ hybridization using an CCR2 / recipients demonstrated reduction of glucose inde- antisense riboprobe, demonstrating mRNA expression (in situ pendent of the level of CCL2 production by islets (Figure 5). signal indicated by the deposition of silver grains) by infiltrat- Although the CCR2Ϫ/Ϫ animals were significantly protected ing mononuclear cells. As negative controls, serial sections from this effect, damage to the islets could not be completely were hybridized with the CCR2 sense probe, demonstrating a accounted for by MCP-1/CCR2 interaction because a minimal very small number of nonspecifically deposited silver grains. decrease in percentage of glucose reduction was seen in the ϫ Original magnification 400. absence of the receptor, suggesting a role for other cytokines or chemokines.

CCL2 Is not ␤ Cell Toxic and Does not Impair Insulin Release In Vitro We and others have shown that islet cell expression of CCR2 mRNA can be induced under various conditions (5,19). Using the more sensitive QPCR assay (data not shown), we demonstrated that islet cells constitutively expressed small amounts of CCR2 mRNA. Therefore, to determine whether the detrimental effect of high levels of CCL2 secretion was due to direct toxic- ity, we measured in vitro glucose-stimulated insulin release and ␤ cell viability, in the presence of recombinant CCL2 at concentrations similar to those produced by cultured islets (Figure 1). Islet cell insulin release at basal (5.5 mM) or high glucose (16.7 mM) conditions was unaffected by rCCL2 at a range of concentrations (100 to 1000 pg/ml; Figure 6). ␤-TC3 cells that were incubated with rCCL2 concentrations ranging from 100 to 105 pg/ml were also evaluated for cell viability using the MTT method. We found that rCCL2 had no direct affect on ␤-TC3 cell proliferation or viability (Figure 7). These results strongly Figure 4. Isograft cytokine mRNA expression. Grafts were an- suggest that CCL2 itself had no deleterious effect on ␤ cell alyzed by QPCR 1 and 4 d after transplantation of cultured or viability or endocrine function. freshly isolated islets (*P Ͻ 0.05 freshly isolated versus cultured islets). Discussion Highly differentiated pancreatic ␤ cells have prominent che- data suggest that mainly monocyte/macrophage-derived cyto- mokine-producing capabilities and, therefore, the potential to kines and not T cell–derived cytokines are likely mediators for induce or alter inflammatory as well as noninflammatory pro- the impaired islet graft function. cesses (5–11). Murine islet cells are not passive bystanders of their own destruction but have the potential to secrete CCL2, Effect of Islet-Derived CCL2 on Islet Isograft Function and this study is the first to show that this is detrimental for We next examined whether the islet cell culture or the infil- immediate posttransplantation graft function. The important trating mononuclear cells are detrimental to islet grafts using a role of the CCR2 pathway in conjunction with low-dose immu- marginal mass transplantation model in streptozotocin-in- nosuppression has been demonstrated in a murine allogeneic duced diabetic syngeneic C57BL/6 mice. In this model, the islet transplantation model (18,20). However, few data are degree of glucose reduction, expressed as percentage reduction available to support the role of donor-derived chemokines in of the pretransplantation glucose concentration, reflects islet transplantation (21). graft function in vivo. The mean glucose reduction over 4 d Our data suggest that islets have the capacity to behave in a posttransplantation was significantly lower when islets were similar manner to cells of the immune system, recruiting leu- transplanted after2dofculture compared with freshly isolated kocytes from the blood stream, resulting in the amplification of J Am Soc Nephrol 16: 444-451, 2005 Chemokines and Islet Transplantation 449

Figure 5. Effect of islet-derived CCL2 on islet isograft function after marginal mass transplantation of streptozotocin-induced diabetic mice. (A) Linear regression of pretransplant CCL2 concentration with glucose reduction 1 d posttransplantation. Linear coefficient (r) and its 95% confidence interval and P-value are shown. High pretransplant CCL2 significantly correlated with poor islet graft function in wild type recipients. This correlation was absent when the CCR2 pathway was disrupted by transplanting CCR2Ϫ/Ϫ mice. (B) Recipients divided into low and high donor-derived CCL2 and the effect on glucose reduction. CCL2 cutoff value was the 50th percentile (82 pg/ml).

Figure 7. Effect of CCL2 on ␤-TC3 cell viability was measured by MTT metabolism. Data are % of control mean Ϯ SEM of three experiments performed in quintuplicate. ␤-TC3 were ␤ Figure 6. Effect of CCL2 on islet and -TC3 insulin secretion. plated in 96-well plates and cultured for 48 hours in the pres- Cells were cultured for 12 h with 0 (vehicle control), 100, and ence of rCCL2 at various concentrations. 1000 pg/ml rCCL2 and then exposed to 5.5 mM or 16.7mM glucose for 1 h. Supernatants were analyzed for insulin content with ELISA. Data are shown as mean insulin secretion Ϯ SEM. are consistent with an inflammatory function for CCL2 in this setting, and migration studies with isolated monocytes have the mononuclear cell infiltration within the islet itself. This shown that CCL2 secreted by islets is biologically active (6,7). supports the data of a recent study in islet transplant patients Our data that in vitro culture of islets augmented CCL2 secre- with type 1 diabetes that found a correlation with high basal tion are in agreement with previous data on human islets donor islet CCL2 production before transplantation with a poor cultured up to 6 (6) or 11 days after isolation (22). Our in vivo clinical outcome (6). Data in the literature are conflicting as to experiments showed that the inflammatory mediator CCL2 whether islets’ expressing CCL2 constitutively may be due to exerts a biologic effect by providing a gradient to attract the differences in species and technique sensitivities. The data CCR2ϩ cells to the site of inflammation. Moreover, the amount that islets express and secrete CCL2 as soon as they are cultured of CCL2 produced negatively correlated with islet graft func- 450 Journal of the American Society of Nephrology J Am Soc Nephrol 16: 444-451, 2005 tion, a correlation absent when the CCR2 pathway was dis- Expression of both CCL19 and CCR7 was also found in glo- rupted. merular mesangial cells, where this pathway was important for Our data suggest that the primary cause of transplanted islet cell survival, suggesting a homeostatic or regulatory role for dysfunction is antigen-nonspecific inflammation at the graft chemokines expressed by nonlymphoid tissue (30). site mediated by the monocyte/macrophage-associated inflam- In conclusion, high pretransplantation CCL2 concentration matory products IL-1␤, TNF-␣, and MIP-1␣. Recent data also has a negative impact on immediate islet graft function as a documented significant expression of macrophage-associated result of increased inflammatory cell recruitment. Therefore, inflammatory genes, including IL-1␤, IL-6, TNF-␣, and CCL2, strategies to decrease CCL2 release by islet preparations or in islets from a brain-dead donor (23), and these islet grafts blocking the CCL2/CCR2 pathway in the recipient might in- were found to have a decreased functionality in vivo and in vitro crease the success of islet engraftment and long-term insulin (24). Together, the data suggest that donor-derived CCL2 leads independence in human islet transplantation. to monocyte/macrophage infiltration, further cytokine secretion, and activation of the innate and adaptive immune re- Acknowledgment sponse to cause graft dysfunction. B.M. was supported by National Institutes of Health research grant Of note, transgenic mice with targeted CCL2 expression for ␤ RO1 AI 49289-01. cells develop intense insulitis, with a predominance of macrophages but no diabetes (25). This suggests that the local pro- References duction of CCL2 is not sufficient to cause tissue destruction but 1. Furie MB, Randolph GJ: Chemokines and tissue injury. might contribute to the recruitment of a first wave of mononu- Am J Pathol 146: 1287–1301, 1995 clear cells, and the final result depends on the presence of other 2. Rovin BH, Yoshiumura T, Tan L: Cytokine-induced pro- chemokines. Our data with early monocyte/macrophage infil- duction of monocyte chemoattractant protein-1 by cul- tration followed by T cells would be consistent with that hy- tured human mesangial cells. J Immunol 148: 2148–2153, pothesis. In addition, in our marginal mass model, which re- 1992 sembles clinical islet transplantation, mild insulitis might affect 3. Prodjosudjadi W, Gerritsma JS, Klar-Mohamad N, Gerrit- islet graft function. Our results exclude, however, that CCL2 is sen AF, Bruijn JA, Daha MR, van Es LA: Production and directly ␤ cell toxic, as it did not affect ␤ cell viability or the cytokine-mediated regulation of monocyte chemoattrac- release of insulin, confirming previous studies with human tant protein-1 by human proximal tubular epithelial cells. islets (6). Kidney Int 48: 1477–1486, 1995 4. Sica A, Wang JM, Colotta F, Dejana E, Mantovani A, Op- Our data have several potential clinical implications. First, penheim JJ, Larsen CG, Zachariae CO, Matsushima K: they provide evidence that islet cell culture affects islet graft Monocyte chemotactic and activating factor gene expres- function in vivo. Transplantation of islets immediately after sion induced in endothelial cells by IL-1 and tumor necro- isolation was done in the original description of the Edmonton sis factor. J Immunol 144: 3034–3038, 1990 protocol (26) and may explain the high success rate not reached 5. Zhang N, Schroppel B, Chen D, Fu S, Hudkins KL, Zhang in other sites despite the same immunosuppressive protocol H, Murphy BM, Sung RS, Bromberg JS: Adenovirus trans- (27). At this point, however, it is unknown whether islet cell duction induces expression of multiple chemokines and culture has a negative impact on graft function in immunosup- chemokine receptors in murine beta cells and pancreatic pressed recipients. Second, measurement of pretransplantation islets. Am J Transplant 3: 1230–1241, 2003 CCL2 level may be a useful tool to preselect islets to increase 6. Piemonti L, Leone BE, Nano R, Saccani A, Monti P, Maffi P, engraftment success rates. Third, strategies to prevent in vitro Bianchi G, Sica A, Peri G, Melzi R, Aldrighetti L, Secchi A, Di Carlo V, Allavena P, Bertuzzi F: Human pancreatic islets CCL2 secretion might offer new approaches in clinical islet produce and secrete MCP-1/CCL2: Relevance in human transplantation. Of note, NF-␬B was found to play an important ␤ islet transplantation. Diabetes 51: 55–65, 2002 role for CCL2 expression in cells and might be a potential 7. Ehrnfelt C, Kumagai-Braesch M, Uzunel M, Holgersson J: target for pretransplantation, ex vivo gene therapy (9). In addi- Adult porcine islets produce MCP-1 and recruit human tion, it was previously shown that the in vitro CCL2 expression monocytes in vitro. Xenotransplantation 11: 184–194, 2004 in human islets can be decreased by cyclosporine and the 8. de Groot M, Schuurs TA, Keizer PP, Fekken S, Leuvenink vitamin B derivative nicotinamide (28). HG, van Schilfgaarde R: Response of encapsulated rat Using a cDNA array approach, we also identified the induc- pancreatic islets to hypoxia. Cell Transplant 12: 867–875, tion of potent chemokines CCL5, CXCL9, CXCL10, and CXCL2 2003 after treatment with cytokines. It is interesting that CXCL9 and 9. Kutlu B, Darville MI, Cardozo AK, Eizirik DL: Molecular CXCL10 are not only chemoattractant but also angiostatic and, regulation of monocyte chemoattractant protein-1 expression in pancreatic beta-cells. Diabetes 52: 348–355, 2003 therefore, of possible importance during islet transplantation, 10. Cardozo AK PP, Gysemans C, Chen MC, Mathieu C, in which revascularization of the islets is required for successful Eizirik DL: IL-1␤ and IFN-␥ induce the expression of di- engraftment (29). However, the role of these chemokines in the verse chemokines and IL-15 in human and rat pancreatic recruitment of inflammatory cells and angiogenesis is unclear islet cells, and in islets from pre-diabetic NOD mice. Dia- and needs to be studied further. We have also shown that islets betologia 46: 255–266, 2003 as well as ␤-TC3 cells expressed the chemokine receptor CCR7, 11. Chen MC, Proost P, Gysemans C, Mathieu C, Eizirik DL: along with both its ligands CCL19/SLC and CCL21/MIP-3␤. Monocyte chemoattractant protein-1 is expressed in pan- J Am Soc Nephrol 16: 444-451, 2005 Chemokines and Islet Transplantation 451

creatic islets from prediabetic NOD mice and in interleu- Luster AD: Donor-derived IP-10 initiates development of kin-1 beta-exposed human and rat islet cells. Diabetologia acute allograft rejection. J Exp Med 193: 975–980, 2001 44: 325–332, 2001 22. Johansson U, Olsson A, Gabrielsson S, Nilsson B, Korsgren 12. Murphy PM, Baggiolini M, Charo IF, Hebert CA, Horuk R, O: Inflammatory mediators expressed in human islets of Matsushima K, Miller LH, Oppenheim JJ, Power CA: In- Langerhans: Implications for islet transplantation. Biochem ternational union of pharmacology. XXII. Nomenclature Biophys Res Commun 308: 474–479, 2003 for chemokine receptors. Pharmacol Rev 52: 145–176, 2000 23. Toyama H, Takada M, Suzuki Y, Kuroda Y: Activation of 13. Maghazachi AA, al-Aoukaty A, Schall TJ: C-C chemokines macrophage-associated molecules after brain death in is- induce the chemotaxis of NK and IL-2-activated NK cells. lets. Cell Transplant 12: 27–32, 2003 Role for G proteins. J Immunol 153: 4969–4977, 1994 24. Contreras JL, Eckstein C, Smyth CA, Sellers MT, Vilatoba 14. Sallusto F, Lanzavecchia A, Mackay CR: Chemokines and M, Bilbao G, Rahemtulla FG, Young CJ, Thompson JA, chemokine receptors in T-cell priming and Th1/Th2-me- Chaudry IH, Eckhoff DE: Brain death significantly reduces diated responses. Immunol Today 19: 568–574, 1998 isolated pancreatic islet yields and functionality in vitro 15. Kuziel WA, Morgan SJ, Dawson TC, Griffin S, Smithies O, and in vivo after transplantation in rats. Diabetes 52: 2935– Ley K, Maeda N: Severe reduction in leukocyte adhesion 2942, 2003 and monocyte extravasation in mice deficient in CC che- 25. Grewal IS, Rutledge BJ, Fiorillo JA, Gu L, Gladue RP, mokine receptor 2. Proc Natl Acad Sci U S A 94: 12053– Flavell RA, Rollins BJ: Transgenic monocyte chemoattrac- 12058, 1997 tant protein-1 (MCP-1) in pancreatic islets produces mono- 16. Bell E, Cao X, Moibi JA, Greene SR, Young R, Trucco M, cyte-rich insulitis without diabetes: Abrogation by a sec- Gao Z, Matschinsky FM, Deng S, Markman JF, Naji A, ond transgene expressing systemic MCP-1. J Immunol 159: Wolf BA: Rapamycin has a deleterious effect on MIN-6 401–408, 1997 cells and rat and human islets. Diabetes 52: 2731–2739, 2003 26. Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E, 17. Efrat S, Fusco-DeMane D, Lemberg H, al Emran O, Wang Warnock GL, Kneteman NM, Rajotte RV: Islet transplan- X: Conditional transformation of a pancreatic beta-cell line tation in seven patients with type 1 diabetes mellitus using derived from transgenic mice expressing a tetracycline- regulated oncogene. Proc Natl Acad Sci U S A 92: 3576– a glucocorticoid-free immunosuppressive regimen. N Engl 3580, 1995 J Med 343: 230–238, 2000 18. Schroppel B, Zhang N, Chen P, Zang W, Chen D, Hudkins 27. Ault A: Edmonton’s islet success tough to duplicate else- KL, Kuziel WA, Bromberg JS, Murphy B: Differential ex- where. Lancet 361: 2054, 2003 pression of chemokines and chemokine receptors in mu- 28. Moberg L, Olsson A, Berne C, Felldin M, Foss A, Kallen R, rine islets allografts: The role of the CCR2 and CCR5 sig- Salmela K, Tibell A, Tufveson G, Nilsson B, Korsgren O: naling pathways. J Am Soc Nephrol 15: 1853–1861, 2004 Nicotinamide inhibits tissue factor expression in isolated 19. Frigerio S, Junt T, Lu B, Gerard C, Zumsteg U, Hollander human pancreatic islets: Implications for clinical islet GA, Piali L: Beta cells are responsible for CXCR3-mediated transplantation. Transplantation 76: 1285–1288, 2003 T-cell infiltration in insulitis. Nat Med 8: 1414–1420, 2002 29. Keane MP, Arenberg DA, Moore BB, Addison CL, Strieter 20. Lee I, Wang L, Wells AD, Ye Q, Han R, Dorf ME, Kuziel RM: CXC chemokines and angiogenesis/angiostasis. Proc WA, Rollins BJ, Chen L, Hancock WW: Blocking the mono- Assoc Am Physicians 110: 288–296, 1998 cyte chemoattractant protein-1/CCR2 chemokine pathway 30. Banas B, Wornle M, Berger T, Nelson PJ, Cohen CD, Kret- induces permanent survival of islet allografts through a zler M, Pfirstinger J, Mack M, Lipp M, Grone HJ, Schlon- programmed death-1 ligand-1-dependent mechanism. dorff D: Roles of SLC/CCL21 and CCR7 in human kidney J Immunol 171: 6929–6935, 2003 for mesangial proliferation, migration, apoptosis, and tis- 21. Hancock WW, Gao W, Csizmadia V, Faia KL, Shemmeri N, sue homeostasis. J Immunol 168: 4301–4307, 2002