In this study we analyzed the role of CCL2, a member of the chemokine family, in early graft damage. Using simultaneous kidney-pancreas transplantation (SPK) as a model, we showed that brain death significantly increases circulating CCL2 levels in humans. We found that in such situations, high donor CCL2 levels (measured before organ recovery and at the onset of cold preservation) correlate with increased postreperfusion release of CCL2 by both the graft and recipient throughout the week following transplantation (n = 28). In a retrospective study of 77 SPK recipients, we found a significant negative association between high donor levels of CCL2 and graft survival. Decreased survival in these patients is related to early posttransplant complications, including a higher incidence of pancreas thrombosis and delayed kidney function. Taken together our data indicate that high CCL2 levels in the donor serum predict both an increase in graft/recipient CCL2 production and poor graft survival. This suggests that the severity of the inflammatory response induced by brain death influences the posttransplant inflammatory response, independent of subsequent ischemia and reperfusion.
Chemokines are a superfamily of small proteins that play a crucial role in establishing immune and inflammatory responses (1–4). Extensive studies of organs recovered from brain-dead rats show that endothelial and parenchymal cells produce chemokines, which in turn, recruit and activate neutrophils (CXCL3/KC, CXCL1/GRO-α and CXCL2/GRO-β), macrophages (CCL2/MCP-1, CCL3/MIP-1α, and CCL4/MIP-1β) and NK and T cells (CXCL10/IP-10). Chemokine production is associated with increased rates of ischemia-reperfusion injury and acute and chronic transplant rejection (5–10). Chemokine damage to the graft can be prevented by treating transplant recipients with steroids (11,12) or agents that block neutrophil adhesion (e.g. recombinant P-selectin glycoprotein ligand (9)), or by inducing protective gene expression (e.g. hemoxygenase-1 (13,14)) in the graft.
In a model of human tissue transplantation, we recently demonstrated that release of the inflammatory chemokine CCL2/MCP-1 by a graft negatively affects clinical outcome (15). In fact, decreased production of CCL2 by pancreatic islets infused in patients with type 1 diabetes is associated with longer lasting insulin independence (15). Presently, there are no available data on how donor CCL2 affects the clinical outcome of human vascularized organ transplantation. In animal models, an intense nonspecific inflammatory response, including CCL2 upregulation, was observed in donor organs after brain death (6,7,16–18). A similar response was reported in studies of human donors (18,19). However, all these studies focused on the inflammatory response in graft biopsy specimens (19–22), in most cases obtained at the time of reperfusion after cold ischemia.
Here we investigate the association between CCL2 levels in donor blood before organ recovery and the clinical outcome of simultaneous kidney-pancreas transplantation (SPK) in recipients with type 1 diabetes.
Patients and Methods
CCL2 levels after SPK transplant: Twenty-eight consecutive patients with type 1 diabetes and end-stage kidney disease (ESRD) who received an SPK transplant from heart-beating donors at the Department of Surgery in the San Raffaele Hospital (University VitaSalute, Milan, Italy) between October 2001 and October 2003 were included in this study. Recipient CCL2 was measured in venous blood samples obtained at day 0 (before transplantation) and 1, 3, 5 and 7 days posttransplantation. Donor CCL2 was measured on serum samples collected at the time of the crossmatch. In a subgroup of patients, CCL2 was also measured during surgery from serum samples collected immediately or 10 minutes after kidney (n = 21/28) or pancreas (n = 8/21) reperfusion (i.e. declamping).
Donor CCL2 levels and transplant outcome: The study group consisted of patients with type 1 diabetes and ESRD who received an SPK transplant at the Department of Surgery San Raffaele Hospital, University Vita Salute (Milan, Italy) between January 1, 1992, and December 31, 2002, and for whom a donor serum sample was available to measure CCL2 (n = 77). Pancreas and kidney grafts were obtained from heart-beating donors through the North Italian Transplant network. Induction therapy consisted of antithymocyte (n = 61) or antilymphocyte (n = 16) globulin. Immunosuppressive therapy consisted of either tacrolimus (n = 16) or cyclosporin-A (n = 61), in association with mycophenolate mofetil (n = 36), azathioprine (n = 41), or prednisone (n = 77). The pancreatic exocrine diversion was enteric in 42 recipients and vesicle in 31 recipients. Four patients received a segmental pancreas transplant. The clinical characteristics of the SPK transplant recipients included in this study are summarized in Table 1.
Table 1. Clinical characteristics and causes of graft loss in 77 kidney/pancreas transplants performed between 1992 and 2002
*Available for 16 of 77 donors.
#Available for 31 of 77 donors.
37.5 ± 7.2
28 ± 9
Diabetes duration (years)
25.6 ± 7.0
Dialysis duration (years)
4.0 ± 2.4
Body weight (kg)
62 ± 10
69 ± 11
22.4 ± 2.9
23.6 ± 2.7
163 ± 142*
7.9 ± 3.4
0.93 ± 0.29#
Cause of death
Cold ischemia time (min)
776 ± 293
Kidney warm ischemia time (min)
41 ± 8
Pancreas warm ischemia time (min)
41.6 ± 8.5
Death with functioning grafts
Causes of graft loss
At the end of follow-up (April 30, 2007), 60 transplant recipients (77.9%) were alive. Of the 17 patients who died during follow-up, both grafts were functioning in 10 at the time of death. The overall median follow-up was 76.4 months (25th–75th percentile: 53.8–123.9) and the median follow-up of the survivors was 87.4 months (65.4–132.3). The cause of graft loss, defined as a return to exogenous insulin therapy (pancreas graft) or development of ESRD requiring dialysis (kidney graft), was abstracted from medical records reviewed by a physician (Table 1). Technical failure (TF) of a pancreas graft was defined as graft loss occurring within 3 months posttransplantation secondary to thrombosis, leaks, bleeding, pancreatitis or infection. Delayed kidney graft function was defined as a <30% creatinine reduction ratio on day 2 posttransplant (i.e. serum creatinine on day 1 minus serum creatinine on day 2, multiplied by 100, and divided by serum creatinine on day 1) (23,24). CCL2 levels were also measured in 21 living kidney donors and 70 healthy individuals matched in age and sex to the SPK recipients.
ELISA assay of CCL2
Serum CCL2 was measured using a sandwich ELISA, as described previously (15). This assay did not cross-react with closely related chemokines, such as CCL7 and CCL8.
Nonnormally distributed variables were compared using the Mann–Whitney U or Kruskal–Wallis H test. Student's t-test was used to compare normally distributed variables. Categorical variables were compared using the Pearson chi-square test. Regression analysis was performed after log transformation of nonnormally distributed variables. Kaplan–Meier survival estimates were compared using the log-rank test. In the analysis of graft survival, patients who died with a functioning graft were considered lost to follow-up. The Cox proportional hazard model estimated the association between selected risk factors and pancreas or kidney loss. Hazard ratio (HR) and 95% CI are presented. Statistical analyses were performed using SPSS 10.0 for Windows (SPSS Inc, Chicago, IL).
Graft release of CCL2 after reperfusion is proportional to the donor level
CCL2 levels in SPK recipients (Figure 1) significantly increased above baseline values after transplantation, reaching a peak between the first and third day, and then gradually decreased to pretransplant levels after 1 week (upper panel). In the subgroup of patients where CCL2 levels were measured immediately and 10 min after kidney and pancreas revascularization, a dramatic increase in CCL2 was observed after graft reperfusion. This early CCL2 peak would have otherwise been missed by the routine daily sampling schedule during the postsurgery period.
Immediately after graft reperfusion, median CCL2 concentrations were significantly higher in the graft vein [218 ng/mL (118–434)] than in the recipient peripheral vein [133 ng/mL (100–302)] (p < 0.001) (Figure 2A). Similar differences in median CCL2 levels were observed 10 min after graft reperfusion [graft: 239 ng/mL (143–509); peripheral vein: 165 ng/mL (119–483), p < 0.001]. Moreover, a positive correlation was observed between graft vein CCL2 levels after graft reperfusion and the area under the curve generated by peripheral vein CCL2 during the first week posttransplantation (data not shown).
To evaluate whether the amount of CCL2 released by a graft is related to the level of donor CCL2, we assessed the relationship between renal vein CCL2 after revascularization and systemic donor CCL2 at the time of organ recovery. We found a positive correlation between these factors, independent of whether CCL2 measurements were taken from samples drawn immediately after revascularization or 10 min later (Figure 2B). Additionally, we evaluated whether donor CCL2 levels affect recipient CCL2 levels during the first week after transplantation. In this case, a significant relationship was also detected between high donor CCL2 and increased recipient CCL2 (Figure 2C).
Taken together these results suggest that donor CCL2 level predicts the degree of CCL2 released by the graft and detected in the recipient during the first week after transplantation.
CCL2 in heart-beating donors: Retrospective evaluation in a cohort of 77 kidney-pancreas donors
To investigate the role of donor CCL2 in the natural history of SPK transplantation, serum concentrations of CCL2 were measured in: (a) 77 heart-beating kidney-pancreas donors; (b) 21 healthy living kidney donors; and (c) 70 matched (age and sex) healthy individuals (Figure 3). The median CCL2 level in the healthy individuals was 64 pg/mL (25th–75th percentile: 50–82 pg/mL). The median level of serum CCL2 in living donors was not significantly different from that in healthy individuals (73 pg/mL, 51–100 pg/mL; p = 0.46), whereas the median CCL2 level in heart-beating kidney-pancreas donors was significantly higher (301.8 pg/mL; 167–484 pg/mL; p < 0.0001).
To investigate if donor CCL2 predicts graft survival, we compared CCL2 levels of the 77 donors with graft survival time in the recipients. We divided heart-beating donors into three groups according to CCL2 serum levels: low CCL2 group (L-group, n = 26, CCL2 level <33th percentile); medium CCL2 group (M-group, n = 26, CCL2 level between the 33th and 65th percentile); and high CCL2 group (H-group, n = 25, CCL2 level ≥66th percentile). The clinical characteristics of these three groups of patients are summarized in Figure 3. Groups were comparable for age, sex, BMI, cause of death and time in critical care before organ recovery. For the purpose of this analysis, L- and M-groups were merged into one group (L/M-group, n = 52). End points for the analysis were either kidney loss (development of ESRD requiring dialysis) or pancreas loss (return to exogenous insulin use). Patients who died but had a functional graft until the day of death were included.
As measured by univariate analysis, recipients of organs from donors in the H-group (CCL2 levels above the 66th percentile) had an increased risk of pancreas loss [HR = 2.31 (95% CI: 1–5.33); p < 0.05], but not kidney loss [HR 2.06 (95% CI:0.68–6.19); p = 0.2] (Figure 4). A multivariate analysis, which included variables significant at p < 0.05 in the univariate analysis, confirmed high donor CCL2 levels as an independent prognostic factor both for pancreas and kidney survival (Table 2). A second multivariate analysis was performed including all available donor variables, regardless of their statistical significance in the univariate analysis. This analysis also indicated that high donor CCL2 concentration is an independent prognostic factor for pancreas survival. Kaplan–Meier analysis was used to confirm that pancreas, kidney and pancreas-kidney survival in the L/M-group was higher than in the H-group (Figure 5A).
Table 2. Cox proportional hazard models of the predictors of graft loss by multivariate analysis
HR (95% CI)
HR (95% CI)
Parsimonious adjusted for recipient-related variables
Recipent sex (M)
Recipient age (year)
Adjusted for variables significant at p < 0.05 in the univariate analysis
Years of dialysis pre-Tx
>1 episodes rejection
MMF vs. AZA
ATG vs. ALG
Adjusted for donor related variables
Donor age (year)
Donor sex (M)
Donor weight (kg)
Cause of death (non traumatic vs. traumatic)
>3 HLA mismatches
Cold ischemia (h)
Warm ischemia (min)
Time in critical care (h)
Cause of graft loss in the L/M-group and H-group is shown in Figure 5B. Overall, TF was the most common cause of pancreas graft loss, with thrombosis (13/15) being the largest component. In the H-group, 32% of the patients lost their pancreas to TF (8/25), whereas only 13.5% did so in the L/M-group (7/52; p = 0.054, chi square test). The univariate relative risk of pancreas loss for thrombosis among H-group patients was 3.02 (95% CI: 0.95–9.6; p = 0.06, logistic regression analysis). CCL2 was not associated with pancreas graft loss to rejection (relative risk 1.6; p = 0.5). Immunologic rejection (mainly chronic) was the most common cause of kidney graft loss. In the H-group, 24% of the patients lost their kidney to rejection (2/25 acute, 4/25 chronic) compared to 11.5% in the L/M-group (2/52 acute, 4/52 chronic; p = 0.15, chi square test). The univariate relative risk of kidney loss to rejection among H-group patients was 2.42 (95% CI: 0.49–8.4; p = 0.16, logistic regression analysis).
The incidence of delayed graft function, expressed as the creatinine reduction ratio, was 48% (12/25) in the H-group and 25% (13/52) in the L/M-group (p = 0.044, chi square test). The univariate relative risk of delayed graft function among H-group patients was 2.76 (1.01–7.5; p = 0.047, logistic regression analysis). Moreover, serum creatinine levels during the first 2 weeks posttransplantation were significantly higher in the H-group than in the L/M-group (Figure 5C). Creatinine clearance remained lower in the H-group than in the L/M-group for up to 5 years posttransplantation (Figure 5C).
In this study we evaluated whether CCL2, a member of the CC chemokine family (25), is a predictor of early graft damage. Using kidney-pancreas transplantation in humans as a model, we showed that high levels of CCL2 in donor serum predict increased graft/recipient CCL2 levels after reperfusion and decreased graft survival.
CCL2 is produced by multiple cell types (i.e. endothelial cells, vascular smooth muscle cells, keratinocytes, fibroblasts, mesangial cells, tubular epithelial cells, lymphocytes and monocyte/macrophages) in response to proinflammatory stimuli, including tumor necrosis factor-α, γ-interferon, lipopolysaccharide, interleukin-1β, platelet-derived growth factor and oxidized LDL. In vitro, subnanomolar concentrations of CCL2 are able to induce chemotaxis of monocytes (26,27) and recruit a subset of T cells (28) and IL-2-activated natural killer cells (29). In monocytes, CCL2 induces not only chemotaxis, but also respiratory burst, rapid induction of arachidonic acid release and changes in Ca2 + concentration (30). Because of its target cell specificity, CCL2 is postulated to play a role in a variety of processes related to organ transplantation, such as acute and chronic rejection (31–35), ischemia/reperfusion injury (36–39) and graft vasculopathy (32,40–42). Supporting this hypothesis, we recently demonstrated that CCL2 released from a graft negatively influences the clinical outcome in a model of human tissue transplantation (15). In fact, CCL2 secreted by pancreatic islets plays a role in the clinical outcome of islet transplant in patients with type 1 diabetes, as decreased CCL2 is associated with longer lasting insulin independence. Until now, no data have been available on how the donor level of circulating CCL2 influences the clinical outcome of human vascularized transplantation.
Here, we demonstrate that brain death significantly increases the circulating level of CCL2 in humans and that the donor CCL2 level predicts the amount of CCL2 released by the graft and recipient after revascularization. The positive correlation between donor CCL2 levels (measured before organ recovery and at the beginning of cold preservation) and postreperfusion recipient CCL2 levels demonstrates that the intensity of the inflammatory response induced by brain death influences the posttransplant inflammatory response, independent of subsequent ischemia and reperfusion. Thus, our results may explain, in part, the inferior outcome of matched organ transplantation from heart-beating donors compared to those from living (un)-related donors (43). The mechanisms by which brain death leads to a systemic CCL2 increase remain unclear. The central catastrophic injury evokes an upsurge of catecholamines along with peripheral tissue vasoconstriction and ischemia. An increase in IL-6 secretion in response to hypoxia has been demonstrated in a variety of cell types, ranging from vascular smooth muscle cells to mononuclear cells (44–51). Additionally, the existence of a loop of CCL2 secretion, involving the IL-6/IL-6Ralpha/gp130 complex (52), suggests that ischemia may be a major trigger of CCL2 secretion (53). However, since brain death also promotes the release of hormones and other inflammatory mediators, what triggers increased CCL2 production is still unknown (19,54–59).
Although a retrospective study has its limits, we found a significant association between high donor CCL2 levels and decreased graft survival. Graft loss was related to early posttransplant complications, including a higher incidence of pancreas thrombosis and delayed renal allograft function. Given the inflammatory basis of these two events, these results were not unexpected (60–62). Although simultaneous kidney-pancreas transplantation is a lifesaving procedure in patients with type 1 diabetes and ESRD, pancreas transplantation is still associated with a high failure rate (63,64). Posttransplantation morbidity is mainly related to pancreas graft complications, with thrombosis being the main nonimmunologic cause of graft loss. At the moment, there are no markers to identify patients at risk for developing graft-related complications (64–67). We suggest that donor CCL2 can be used as a marker to assess the risk of early pancreas graft damage and graft-related pancreas morbidity. Of note, this marker could be measured in the peripheral blood before organ recovery using a quick and cheap reproducible method, such as ELISA, potentially enabling all donors to be screened.
In conclusion, our results further emphasize the influence of the donor proinflammatory state on posttransplantation morbidity and graft survival, and suggest a relevant role for CCL2 as a mediator. The association between donor CCL2 and outcome in SPK transplantation recipients suggests that either CCL2 reduces graft survival by promoting a pro-inflammatory environment, or is itself a marker of damaged grafts. Although it is likely that CCL2 has causal effect, i.e. its presence is detrimental, we cannot exclude the possibility that CCL2 is simply a marker of ‘stress’ in the donor. Strategies aimed at selectively inhibiting CCL2 in heart-beating donors and/or graft recipients could have the potential to improve graft outcome of simultaneous kidney-pancreas transplantation.
This work was supported by Telethon Italy and the Juvenile Diabetes Research Foundation (JT01Y01), Ministero della Salute (Ricerca Finalizzata RF041234). We thank Massimo Cardillo and Mario Scalamogna (Transplant Immunology and Blood Bank, Maggiore Policlinico Hospital, Milano, Italy) for stimulating discussion and critical reading of the article.