Journal Logo



Berney, Thierry2 3 5; Molano, R. Damaris2 5; Cattan, Pierre4 5; Pileggi, Antonello5; Vizzardelli, Caterina5; Oliver, Robert5; Ricordi, Camillo5; Inverardi, Luca5 6

Author Information


Transplantation of islets of Langerhans represents a potential curative treatment for diabetes mellitus (1). Unfortunately, in clinical trials, most transplant recipients do not experience long-term function of the transplanted islets, and insulin-independence is still an uncommon achievement. A majority of islet grafts are lost within the first weeks after transplantation. It is interesting that it has been demonstrated that long-term islet graft survival was not different than that of whole organ pancreatic transplants when only recipients of “state-of-the-art isolation and transplantation” grafts with persisting function at 3 months posttransplantation were considered (2). The fact that primary nonfunction (PNF) or early graft loss also occurs in syngeneic islet transplant (3) or in models of T-cell inactivation (4, 5) suggests that mechanisms other than the concurrence of rejection, recurrence of autoimmunity and toxicity of conventionally used immunosuppressive agents are involved in this phenomenon. Inadequate mass of transplanted islets and lack of vascularization of the graft in the first days after transplantation might partly account for the nonspecific mechanisms of early graft loss. Furthermore, islets are notorious for their high sensitivity to noxious stimuli exerted by the release of macrophage proinflammatory byproducts, such as free radicals and cytokines (6–8). Implanted islets are thus exposed to an early, nonspecific inflammatory injury triggered by activation of the transplant microenvironment, with the participation of humoral and cellular mediators. This early inflammation induces functional stunning or even destruction of the islets and amplifies the subsequent autoimmune and alloimmune reactions. The role of macrophages in the graft microenvironment in PNF was confirmed by reports of prolonged graft survival by treatment with pharmacological agents aiming at macrophage depletion or inactivation, such as 15-deoxyspergualine, clodronate, or gadolinium (9–11), in murine or canine models of islet transplantation. Furthermore, it was first hypothesized and then demonstrated that proinflammatory cytokines were released endogenously within the islets by resident islet macrophages and could inhibit β-cell function, which would account for β-cell damage in the early stages of diabetes (12–14). These findings should be confronted with the recent observation of induction of interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α in islets during the isolation procedure, which was likely to be related to endotoxin contamination of the collagenase blends used (15). Endotoxin contamination of the enzyme blend was also shown to impair islet engraftment in a xenogeneic rat-to-mouse model (16). Islet purification by Ficoll density gradients could also be detrimental to islet viability and is a step associated with significant insulin loss from the islets (17). Indeed, collagenase and Ficoll have been shown to have important endotoxin contamination, with the capacity to induce IL-1β release from human peripheral blood monocytes (18, 19). Apoptotic cell death could be a significant cause of islet PNF, and has been shown to occur after islet isolation (20) and as a response to cytokine stimulation (21). Thus, the availability of the new enzyme blend Liberase™, characterized by almost absent endotoxin contamination and a gentler enzymatic digestion activity associated with increased islet yields (18, 22), allowed us to test the effects of endotoxin on occurrence of PNF.

Analysis of the early inflammatory events is difficult, because it can be confounded by concurrent phenomena of allorecognition and autoimmunity. In this study we have used the model of PNF described by Kaufman, in which a marginal mass of syngeneic islets is transplanted (5, 9). Implantation of a marginal mass of islets is characterized by achievement of normoglycemia in <100% of recipients and/or a delay in reversal of hyperglycemia. The extent of this delay reflects many parameters, including the quality of the islet preparation and the degree of reactivity of the graft microenvironment. Targeted manipulations that lead to a higher percentage of successful grafts and/or to a reduction of the lag to normoglycemia can be interpreted as beneficial in preventing nonspecific destruction. The term PNF, often associated with the concept of poor functional quality of the grafted tissue, might be revisited to include the participation of inflammatory phenomena. We report on the better early function observed with grafts obtained by Liberase™ isolation, compared with the more conventional enzyme mixture Type V collagenase, on the role of endotoxin contamination of the enzyme blends on the development of PNF, and on the cellular mechanisms associated therewith.



Male C57BL/6 (B6) mice were purchased from Hilltop Lab Animals, Inc. (Scottdale, PA). All animals were kept in the University of Miami animal facilities and used in compliance with the United States Department of Agriculture and National Institute of Health regulations. All animal manipulations were conducted and monitored under protocols reviewed and approved by the Institution Animal Care and Use Committee. Islet donors were sacrificed at 12 weeks of age. Islet recipients (7 weeks old) received a single i.v. injection of 200 mg/kg streptozotocin (STZ; Sigma, St-Louis, MO) to induce diabetes. STZ-injected mice were used as recipients after 7 days, only if 2 consecutive nonfasting blood glucose >250 mg/ml were measured in samples collected from the tail vein using a strip glucometer (Elite; Bayer, Tarrytown, NY).

Enzyme blends used for islet isolation.

Two different enzyme blends were used for purpose of comparison. In one set of isolation procedures, we used Collagenase type V (Sigma) at a final concentration of 1.5 mg/ml in Hank’s balanced saline solution (HBSS; Gibco, Long Island, NY). In a second set of experiments, we used Liberase RI (Roche-Boehringer-Mannheim, Indianapolis, IN) at a final concentration of 0.17 mg/ml in HBSS supplemented with 25 mM HEPES buffer (Mediatech, Herndon, VA). In a third set of experiments endotoxin (purified lipopolysaccharide from Eschericia coli 055:B5; Sigma) was added to the 0.17 mg/ml Liberase™ solution at described concentrations.

Endotoxin contents determination.

The endotoxin contents of the collagenase and Liberase™ working solutions were measured by the gel-clot method (Associates of Cape Cod, Falmouth, MA). Samples were run in duplicates. Endotoxin concentrations were obtained by comparing the experimental values to a standard curve generated by serial 1:2 dilutions of a known concentration of reference endotoxin (E. coli O113). The error of the test is±one 2-fold dilution.

Islet of Langerhans isolation procedure.

Murine islets were isolated by means of a previously described technique (11). Briefly, animals were sacrificed under general anesthesia, the abdomen was opened, and the pancreas was exposed and injected with the enzyme solution through the main bile duct until full distension was achieved. The pancreatic tissue was then surgically removed and immersed in the enzyme solution. For collagenase isolations, digestion was performed in a 17-minute incubation at 37°C, with gentle shaking, after which enzyme kinetics were sharply slowed by addition of cold HBSS supplemented with 10% fetal calf serum (Hyclone, Logan, UT). For Liberase™ isolations, digestion was performed in a 30-min static incubation at 37°C, followed by a brief, vigorous, manual shaking. Mechanical disruption of the digested pancreatic tissue was achieved by repeated passages through needles of decreasing gauge until complete release of free islets was observed under the microscope, and the tissue was then filtered through a 450-μm screen. Islet purification was obtained by centrifugation at 900 g for 11 min on discontinuous Euro-Ficoll gradients, and routinely provided islets of purity >90%. Islet purity was assessed by dithizone (Sigma) staining, and the islets were counted and scored for size. An algorithm was used for the calculation of 150 μm diameter islet equivalent number (IEQ). For each isolation, average size of islets was calculated using the following formulaMATH : Before transplant, islets were cultured overnight at 37°C, 5% CO2, in CMRL medium (Gibco) supplemented with 10% fetal calf serum, 2 mmol/l l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 25 mmol/l HEPES buffer (CMRL-10).

Marginal mass islet transplantation.

Immediately before transplantation, islets were handpicked under the microscope and divided in aliquots of 50–300 per recipient. Under general anesthesia induced by methoxyflurane inhalation (Metofane, Schering-Plough Animal Health, Atlanta, GA), a left lombotomy was performed and the left kidney was exteriorized and exposed. A polyethylene catheter was introduced through a breach made in the kidney capsule at the upper pole and brought to the lower pole. Islets were slowly infused and spread at the lower kidney pole. The catheter was then gently removed, and the breach was cauterized. The kidney was repositioned into the abdominal cavity, and the muscular layer and skin were sutured.

Graft function.

Blood glucose levels were measured daily after transplantation on whole blood samples collected from the tail vein to determine the number of days needed to achieve normoglycemia. The first day of graft function was defined as the first of 5 consecutive days of nonfasting blood glucose <200 mg/ml. To rule out residual function of the native pancreas, survival nephrectomy was performed after diabetes reversal to verify that a quick return to hyperglycemia was obtained. For mice remaining hyperglycemic >100 days after transplantation failure was considered to have occurred, and the mice were killed.

Determination of cytokine levels in islet culture supernatants.

Aliquots of 100 handpicked islets were cultured in triplicates at 37°C, 5% CO2, in 2 ml CMRL-10 in 24-well untreated microtiter plates (Corning Inc., Corning, NY). Supernatant was harvested at various time points and concentration levels of IL-1β, TNF-α, and IL-6 were measured by ELISA, using Quantikine M kits (R&D Systems, Minneapolis, MN). A sample of known concentration was run in each test as a quality control. Cytokine levels in the supernatants of murine resident peritoneal macrophages, obtained by peritoneal lavage, resting or stimulated with lipopolysaccharide (LPS) (50 endotoxin units [EU]/ml) and interferon-γ (IFN-γ; 10 U/ml), were measured to serve as positive controls. The detection threshold was 7.5 pg/ml for IL-1β, 16 pg/ml for TNF-α, and 15.5 pg/ml for IL-6. Results are expressed in pg/100 islets.

Determination of cytokine expression in isolated islets.

IL-1β, TNF-α, and IL-6 expression was assessed by quantitative reverse transcription-polymerase chain reaction (RT-PCR) using the Light Cycler™ machine (Roche-Boehringer-Mannheim). After a 3 hr culture at 37°C, 5% CO2 in CMRL-10, at least 700 islets were washed 3 times in phosphate-buffered saline (Gibco) without Ca2+ and Mg2+ and frozen and kept at −80°C until use. RNA extraction from the frozen islets was performed using the RNA NOW-LM kit (Biogentex, Seabrook, TX). Synthesis of first strand cDNA was performed from DNase-treated RNA (DNase I, Gibco), using the Superscript II RT kit (Gibco), with 25 ng/μl oligo(dT)12–18 as primer. The RT product was amplified by PCR in the Light-Cycler™ machine that allows real-time quantification of the PCR product, on the basis of the incorporation of a fluorescent dye in the neosynthesized DNA and its measurement at the end of each PCR cycle. Relative level of initial transcript copy number in each cDNA sample was calculated as previously described (23). PCR conditions were the same for all primer pairs: denaturation at 94°C for 1 min, followed by 35 cycles of 94°C for 1 sec, 55°C for 9 sec, and 72°C for 20 sec, with fluorescence read at 85°C after each cycle. Initial starting concentrations of cDNA were determined in arbitrary units, and the ratio of cytokine/actin levels was used for comparison purposes. Oligonucleotide primers for human β-actin (forward, GATGACCCAGATCATGTTTG; reverse, AGGCTGGAAGAGTGCCTCA), murine IL-1β (forward, CCCATTAGACAGCTGCACTAC; reverse, GTGCTCTGCTTGTGAGGTGCT), murine TNF-α (forward, GGCAGGTCTACTTTGGAGTCATTG; reverse, CATTCGAGACTCCAGTGAATT), and murine IL-6 (forward, GCCAGAGTCCTTCAGAGAGATACAG; reverse, CCCAACATTCATATTGTCAG) were used. The human β-actin forward primer has 100% homology with the murine sequence, and the human reverse primer has a 2-base difference with the murine sequence on its 3′ end. Recovered PCR products were 440 bp (β-actin), 384 bp (IL-1β), 286 bp (TNF-α), and 343 bp (IL-6).

Apoptosis assay.

After an 18 hr culture at 37°C, 5% CO2 in CMRL-10, islets were incubated for an additional 6 hr in CMRL-10 medium, with or without a cocktail of pro-apoptotic cytokines (21, 24). Cytokine concentrations were 50 U/ml for IL-1β, 103 U/ml for IFN-γ, and 103 U/ml for TNF-α. Recombinant cytokines were purchased from R&D Systems. Aliquots of 1000 islets were washed twice in phosphate-buffered saline without Ca2+ and Mg2+ and dispersed into single cell suspensions by gentle continuous pipetting in trypsine-EDTA (0.05% trypsine, 0.53 mM EDTA, Gibco) for 4 min. Enzymatic activity was stopped by the addition of heat inactivated bovine serum and cells were washed twice. Trypan blue exclusion test consistently showed cell viability >95%. Cells were resuspended and incubated for 15 min at room temperature in 50 μl HEPES buffer containing 1 μg/ml propidium iodide (PI) and 2 μg/ml fluorescent Annexin V (Annexin V-FLUOS Staining Kit, Roche-Boehringer-Mannheim). Controls included cells incubated with Annexin-V only, PI only, or no marker. Samples were then diluted in 300 μl HEPES buffer and analyzed within 2 hr on a Coulter flow cytometer (Beckman Coulter, Fullerton, CA) using 488 nm excitation and a 530/30 nm bandpass filter for fluorescein detection and a 650 nm longpass filter for PI detection. For each sample, 5000 cells were analyzed. Results are expressed as percentage of Annexin single positive and Annexin-PI double positive cells.

Statistical analysis.

The Statistica software package (Statsoft, Tulsa, OK) was used for statistical analysis. Data were expressed as mean±standard error of the mean (SEM). Variables were compared using the one-tailed Student’s t test. Kaplan-Meier analysis was performed for diabetes-free survival determination, and differences were assessed with the Mantel-Cox logrank test. Values of P <0.05 were considered significant.


Endotoxin contents of the enzyme blends.

The endotoxin contents of the enzyme blends resuspended according to manufacturer’s instructions were 400 EU/ml for Collagenase type V, and 3.1 EU/ml for Liberase RI. Endotoxin contents of the buffers used for enzyme suspension (HBSS and HBSS supplemented with HEPES, respectively) were <0.03 EU/ml. A control sample of HBSS contaminated with a known quantity (1000 EU/ml) of bacterial LPS gave a result of 800 EU/ml, consistent with the error of the test.

Time to normoglycemia in mice receiving a marginal islet mass.

To assess the effect of the enzyme blend used in the isolation procedure on graft function in the absence of specific immune mechanisms, 50–300 syngeneic islets were transplanted under the kidney capsule of chemically-diabetic B6 mice. Time to normoglycemia and percentage of animals achieving normoglycemia were compared in the experimental groups receiving different numbers of islets. Five different isolation procedures were performed with Collagenase type V and with Liberase RI. The average islet size was 130 μm in the collagenase groups and 133 μm in the Liberase™ groups (P =0.26, n.s.). The number of islets/kg transplanted in each group of mice was comparable for collagenase- or Liberase™-isolated islets. A higher proportion of mice achieved normoglycemia, and time required to reverse diabetes was consistently lower in the Liberase™ groups (Table 1). All animals returned to hyperglycemia within 48 hr of nephrectomy.

Table 1
Table 1:
Dose-response time necessary to return to normoglycemia in chemically-diabetic C57BL/6 mice receiving a marginal mass of syngeneic pancreatic islets, according to the enzyme blend used for islet isolation

To assess the role of endotoxin contamination of the collagenase blend in the observed delayed function, two additional isolation procedures were performed with a Liberase™ solution to which 400 or 800 EU/ml of E. coli LPS were added. Two groups of chemically-diabetic B6 mice received 75 islets obtained from these isolations. The addition of 400 EU/ml to the Liberase™ solution increased the time to diabetes reversal towards that of collagenase-isolated islets (P =0.1 vs, LPS-free Liberase™). Addition of 800 EU/ml further shifted the diabetes reversal time curve (P =0.005 vs. LPS-free Liberase™;Figure 1).

Figure 1
Figure 1:
Time to normoglycemia in streptozotocin-induced diabetic C57BL/6 mice receiving a marginal islet mass. Eight-week-old mice received transplants under the kidney capsule with 75 syngeneic islets isolated with Liberase™ (circles; n=5) or collagenase (squares; n=5). Endotoxin contents of the collagenase solution were 400 EU/ml, and 3.1 EU/ml for the Liberase™. To assess the role of endotoxin, additional mice received the same number of islets, isolated with a Liberase™ solution in which 400 EU/ml (triangles; n=4) or 800 EU/ml endotoxin (diamonds; n=3) had been added.

Cytokine release and expression by cultured islets.

To determine whether the differential endotoxin contents of the enzyme blends used for isolation induced different levels of intra-islet cytokine production and release in the absence of any other stimulus, proinflammatory cytokines (TNF-α, IL-1β, and IL-6) released by the islets were measured in the supernatants after 3 hr and 18 hr in culture and compared. Higher levels of all 3 cytokines studied were consistently found in the supernatants of collagenase-isolated islets, compared with Liberase™. Substantial isolation-to-isolation variability was seen in cytokine levels, as indicated by high SEMs. IL-1β or TNF-α never reached measurable levels in the culture supernatants of Liberase™-isolated islets. Results are shown in Figure 2.

Figure 2
Figure 2:
Cytokine release in the supernatant of cultured islets. Islets isolated with Liberase™ (squares) or collagenase (circles) were cultured in aliquots of 100 IEQ in 2 ml of CMRL-10 for 3 or 18 hr. At the end of the culture time, supernatants were harvested and cytokine levels measured by ELISA.A, TNF-α;B, IL-1β;C, IL-6. Bars indicate ±SEM Results are representative of 6 Liberase™ and 5 collagenase isolation procedures. Cytokine levels measured in the supernatants of resting or stimulated peritoneal macrophages (positive controls) were 16.5 pg/ml (resting) and 1387 pg/ml (stimulated) for TNF-α, 8.9 pg/ml (resting) and 29.2 pg/ml (stimulated) for IL-1β, and 26.4 pg/ml (resting) and 15842 pg/ml (stimulated) for IL-6.

To verify that the enhanced cytokine release observed was related to increased expression, transcripts levels of IL-1β, IL-6, and TNF-α were measured in the isolated islets after 3 hr culture by quantitative RT-PCR, using a Light-Cycler™ machine. There was a 3.4- to 5.3-fold increase in relative levels of cytokine mRNA expression for IL-6 and TNF-α in the collagenase-isolated islets, compared with Liberase™. This increase, consistent with cytokine release data, was not observed for IL-1β, whereas mRNA levels were almost identical in both groups (Table 2).

Table 2
Table 2:
Relative levels of cytokine expression in islets according to enzyme blend used for isolationa

Islet apoptosis.

To determine if the noxious stimuli mediated by the increased production of proinflammatory cytokines observed in collagenase-isolated islets led to increased islet cell death by necrosis or apoptosis, islets were dispersed with trypsin-EDTA into single cell suspensions after an 18 hr incubation and were stained with annexin V-FLUOS (annexin) and propidium iodide (PI). Flow cytometric analysis allowed the identification and quantification of subsets of necrotic (annexin-PI double positive) and apoptotic (annexin single positive) islet cells (Fig. 3). Basal levels of isolation-induced apoptosis were approximately twice as elevated in the collagenase as in the Liberase™ groups, whereas the percentage of necrotic cells was only slightly more elevated after collagenase isolation. Islets isolated with either enzyme blend approximately doubled their percentage of apoptotic cells in response to incubation with a cytokine cocktail described as a potent inducer of islet cell apoptosis (21, 24). However, the level of apoptosis in Liberase™-isolated islets after cytokine induction was lower than the basal level of apoptosis in collagenase-isolated cells. Incubation with the cytokine cocktail did not raise the percentage of necrosis, indicating a selective effect on apoptosis (Table 3). In islets isolated with Liberase™ in which 800 EU/ml LPS had been added, the percentages of apoptotic and necrotic cells were 44.6% and 15%, respectively.

Figure 3
Figure 3:
Flow cytometry analysis of C57BL/6 islets stained with Annexin-V-FLUOS and propidium iodide (PI).A, Islets isolated with collagenase were cultured for 24 hr at 37°C, in a 5% CO2 atmosphere. B, Collagenase-isolated islets were incubated with a pro-apoptotic cytokine cocktail (50 U/ml IL-1β, 103U/ml IFN-γ and 103 U/ml TNF-α) during the last 6 hr of the 24 hr incubation. C, Liberase™-isolated islets were cultured for 24 hr. D, Liberase™-isolated islets were incubated with 50 U/ml IL-1β, 103U/ml IFN-γ and 103 U/ml TNF-α during the last 6 hr of the 24 hr incubation. Islet cells were dispersed with trypsine-EDTA at the end of the incubation period, stained with Annexin-V-FLUOS and propidium iodide (PI) and analyzed by flow cytometry method. Annexin-V single positive cells represent the apoptotic population, and Annexin-V/PI double positive cells are necrotic. The upper half of each panel shows Annexin-V/PI double staining, and the lower half of each panel shows Annexin-V single staining. This figure is representative of 3 Liberase™ and 4 collagenase isolation procedures.
Table 3
Table 3:
Spontaneous and induced apoptosis and necrosis of islet preparations according to enzyme blend used for isolation


This study links several elements that have a detrimental effect on islets isolated for transplantation and demonstrates their biological significance in an in vivo system of PNF.

The marginal islet mass model we have used allows the study of the fate of transplanted islets as a result of nonspecific stimuli occurring during the isolation procedure or in the graft microenvironment, in the absence of immune rejection mechanisms. In this model, the time required to revert diabetes is inversely proportional to the number of syngeneic islets transplanted (5, 9). Thus, an increase over baseline of the time to normoglycemia could result either from the permanent loss of a portion of the islets because of necrotic or apoptotic cell death, or from a transient loss of function.

The high endotoxin contents of crude collagenase blends and their implication on inflammatory insult to the islets because of cytokine production were suggested in the past (15, 18). The newly available, highly purified enzyme blend Liberase™, characterized by minimal levels of endotoxin contamination, has been associated with better reproducibility and higher islet yield and integrity in animal models and in humans compared with collagenase, whereas reports on viability and in vitro function have been contradictory (22, 25–27). However, the implications of the differences in enzyme composition and endotoxin contamination on islet PNF have not yet been studied in an in vivo islet transplantation model. The endotoxin contents of the two enzyme blends used are likely to be the primum movens of the consecutive chain of events reported here, even if causality and sequence are not formally demonstrated.

The dose-response analysis of time to normoglycemia as a function of number of islets transplanted consistently revealed earlier graft function for Liberase™-isolated islets. The critical role of endotoxin contamination in this observation was confirmed by the results obtained by adding endotoxin to the Liberase™ preparation. Transplantation of 75 islets (the amount at which differences between both enzyme blends were the most conspicuous and significant) isolated with a Liberase™ solution in which endotoxin was added at the same concentration present in collagenase, produced a time curve similar to that produced by collagenase-isolated islets. Doubling the endotoxin concentration in the Liberase™ solution further shifted the time curve. These data are in agreement with a recent report, in which endotoxin contamination of the enzyme blend impaired islet engraftment, as suggested by lower insulin recovery from islet grafts removed 3 days after transplantation (16).

Production of pro-inflammatory cytokines by cells of the monocyte-macrophage lineage in response to endotoxin-contaminated collagenase has been reported in an in vitro model (18, 19). More recently, direct expression and production of TNF-α, IL-1β, and IL-6 by the islets after isolation was reported (15). Our study brings new insight into the role of endotoxin stimulation on cytokine production by islets of Langerhans, by reporting consistently higher levels of mRNA transcripts and cytokine release by islets isolated with the endotoxin-containing enzyme blend. The role of other as yet unidentified factors is also outlined, because 1) although endotoxin contents of the collagenase solution were constant, great isolation-to-isolation variability in cytokine levels was observed, and 2) significant IL-6 release was sometimes observed in Liberase™-isolated islets, even though the enzyme solution was endotoxin-free. Response to the question of the cellular origin of these pro-inflammatory cytokines remains conjectural. It has previously been demonstrated that resident intra-islet macrophages produced increased amounts of IL-1β in response to stimulation by endotoxin and that this endogenous cytokine release could in turn inhibit β-cell function (14). Another study points out that cytokine production by human islets, both in terms of mRNA expression and protein release, is markedly increased when the purity of the islet preparation decreases (40% vs. 80% purity), and concludes that the exocrine tissue contained in the preparation could be a major source of cytokines (15). It is interesting that the preparation purities in this murine study were routinely >90%, and IL-6 and TNF-α levels were consistently measurable in the supernatants of collagenase-isolated islets, which would contradict this conclusion. From another standpoint, expression of CD14, the ligand for LPS-LPS-binding protein complex mostly expressed on cells of the monocyte lineage, has been demonstrated on islet α- and β-cells (28), which makes these cell types possible cytokine producers in response to endotoxin. We report very high levels of IL-6, compared with IL-1β or TNF-α, the latter 2 cytokines having always been below detection level in the supernatants of Liberase™-isolated islets. This might suggest that different arrays of cell types express and release IL-6 or the other two cytokines. Two likely candidates for IL-6 production are beta-cells (29), and the recently characterized islet microvascular endothelial cells (30), because endothelial cells are known to produce copious amounts of this cytokine in response to endotoxin stimulation (31, 32). A role for IL-6 in inhibition of glucose-stimulated insulin release and in islet cell destruction has been reported (33, 34).

It is established that apoptosis participates in the death of freshly isolated islets cultured in standard conditions (20, 35). There is evidence that the lack of native extra-cellular matrix components and of selected growth factors is involved in triggering islet programmed cell death in experimental models (20, 36, 37). DNA fragmentation, the morphological hallmark of apoptosis, occurs early in β-cells after stimulation with proinflammatory cytokines (24). Thus the increased apoptotic islet cell death observed after collagenase isolation could occur simply as a result of increased cytokine release. However, the baseline apoptosis observed after endotoxin-free Liberase™ isolation indicates that LPS contents of the enzyme blend is not the sole factor of apoptosis during islet isolation. No data linking endotoxin stimulation with islet apoptotic cell death is as yet available, but evidence that endotoxin could induce programmed cell death is emerging. It is notable that in a model of cultured rat cardiomyocytes, TNF-α-mediated apoptosis was shown to occur as a result of engagement of the CD14 receptor expressed on the cell surface by bacterial LPS (38).

There is one caveat for the comparison of results obtained with either enzyme blends insofar that the methods used in the digestion phase of the isolation were not strictly identical. The mechanical component used for collagenase isolation was gentle and continuous throughout digestion, as opposed to brief and vigorous at the end of a static incubation for Liberase™. These two techniques were adapted from the manufacturers’ instructions and were chosen because they were likely to optimize the isolation yields. It is therefore more accurate to view this study as a comparison between two isolation procedures, rather than between two enzyme blends, although the differences in the digestion phase are likely to add only minor variables, if any, as suggested by the results of the experiments in which LPS was added to Liberase™.

Taken together, these data suggest that endotoxin could play its detrimental role on islet function by stimulating a variety of resident cell types, that is, macrophages, endothelial cells, and endocrine cells, during the isolation process. Direct binding of endotoxin to β-cells expressing CD14 can be hypothesized. Cytokine production elicited by endotoxin stimulation may lead to apoptosis or destruction of the endocrine cells through the release of cytotoxic mediators such as free radicals (6, 14). By the time islets are transplanted, endotoxin has induced overexpression of adhesion and MHC molecules (L. Inverardi, unpublished data) and cytokines along with some residual endotoxin are likely to be carried to the graft site, all of which may lead to additional activation of host effector cells.

This study contributes to the unraveling of the detrimental actions of endotoxin on islet transplant function and demonstrates its role in primary nonfunction in vivo. Early islet cell loss by apoptosis is shown to be a significant component of endotoxin-related islet primary nonfunction. The use of endotoxin-free enzyme blends or other reagents for islet isolation should be a major step ahead in the quest for the eradication of islet primary nonfunction.


The authors thank James Phillips, at the Department of Microbiology and Immunology, University of Miami, who performed the flow cytometry analysis.


1. Hering BJ, Ricordi C. Results, research priorities, and reasons for optimism: islet transplantation for patients with type I diabetes. Graft 1999; 2: 12.
2. Ricordi C. Human islet cell transplantation: new perspectives for an old challenge. Diabetes Rev 1996; 4: 356.
3. Nagata M, Mullen Y, Matsuo S, Herrera M, Clare Salzer M. Destruction of islet isograft by severe non-specific inflammation. Transplant Proc 1990; 22: 855.
4. Chahine AA, Stoeckert C, Lau HT. Local immunomodulation to promote co-stimulatory blockade. Clin Transplant 1995; 9: 215.
5. Kaufman DB, Platt JL, Rabe FL, Dunn DL, Bach FH, Sutherland DER. Differential roles of Mac-1+ cells, and CD4+ and CD8+ T lymphocytes in primary nonfunction and classic rejection of islet allografts. J Exp Med 1990; 172: 291.
6. Rabinovitch A, Suarez-Pinzon WL, Strynadka K, Lakey JR, Rajotte RV. Human pancreatic islet beta-cell destruction by cytokines involves oxygen free radicals and aldehyde production. J Clin Endocrinol Metab 1996; 81: 3197.
7. Corbett JA, Sweetland MA, Wang JL, Lancaster JR Jr, McDaniel ML. Nitric oxide mediates cytokine-induced inhibition of insulin secretion by human islets of Langerhans. Proc Natl Acad Sci USA 1993; 90: 1731.
8. Nussler AK, Carroll PB, Di Silvio M, et al. Hepatic nitric oxide generation as a putative mechanism for failure of intrahepatic cell grafts. Transplant Proc 1992; 24: 2997.
9. Kaufman DB, Gores PF, Field MJ, et al. Effect of 15-deoxyspergualin on immediate function and long-term survival of transplanted islets in murine recipients of a marginal islet mass. Diabetes 1994; 43: 778.
10. Stephanian E, Lloveras JJ, Sutherland DE, et al. Prolongation of canine islet allograft survival by 15-deoxyspergualin. J Surg Res 1992; 52: 621.
11. Bottino R, Fernandez LA, Ricordi C, et al. Transplantation of allogeneic islets of Langerhans in the rat liver: effects of macrophage depletion on graft survival and microenvironment activation. Diabetes 1998; 47: 316.
12. Mandrup-Poulsen T, Bendtzen K, Dinarello CA, Nerup J. Human tumor necrosis factor potentiates human interleukin-1-mediated rat pancreatic β-cell cytotoxicity. J Immunol 1987; 139: 4077.
13. Sandler S, Andersson A, Hellerstrom C. Inhibitory effects of interleukin 1 on insulin secretion, insulin biosynthesis, and oxidative metabolism of isolated rat pancreatic islets. Endocrinology 1987; 121: 1424.
14. Arnush M, Heitmeier MR, Scarim AL, Marino MH, Manning PT, Corbett JA. IL-1 produced and released endogenously within human islets inhibits β cell function. J Clin Invest 1998; 102: 516.
15. Vargas F, Vives-Pi M, Somoza N, et al. Endotoxin contamination may be responsible for the unexplained failure of human pancreatic islet transplantation. Transplantation 1998; 65: 722.
16. Eckhardt T, Jahr H, Federlin K, Bretzel RG. Endotoxin impairs the engraftment of rat islets transplanted beneath the kidney capsule of C57BL/6-mice. J Mol Med 1999; 123.
17. Brandhorst H, Brandhorst D, Brendel MD, Hering BJ, Bretzel RG. Assessment of intracellular insulin content during all steps of human islet isolation procedure. Cell Transplant 1998; 7: 489.
18. Jahr H, Pfeiffer G, Hering BJ, Federlin K, Bretzel RG. Endotoxin-mediated activation of cytokine production in human PBMCs by collagenase and Ficoll. J Mol Med 1999; 77: 118.
19. Vargas F, Vives-Pi M, Somoza N, Fernandez-Llamazares J, Pujol-Borrell R. Endotoxin activity of collagenase in human islet transplantation. Lancet 1997; 350: 641.
20. Rosenberg L, Wang R, Paraskevas S, Maysinger D. Structural and functional changes resulting from islet isolation lead to islet cell death. Surgery 1999; 126: 393.
21. Delaney CA, Pavlovic D, Hoorens A, Pipeleers DG, Eizirik DL. Cytokines induce deoxyribonucleic acid strand breaks and apoptosis in human pancreatic islet cells. Endocrinology 1997; 138: 2610.
22. Linetsky E, Bottino R, Lehmann R, Alejandro R, Inverardi L, Ricordi C. Improved human islet isolation using a new enzyme blend, liberase. Diabetes 1997; 46: 1120.
23. Morrison TB, Weis JJ, Wittwer CT. Quantification of low copy transcripts by continuous SYBR Green I monitoring during amplification. Biotechniques 1998; 24: 954.
24. Cavanagh TJ, Lakey JR, Dwulet F, et al. Improved pig islet yield and post-culture recovery using Liberase PI purified enzyme blend. Transplant Proc 1998; 30: 367.
25. Lakey JR, Cavanagh TJ, Zieger MA, Wright M. Evaluation of a purified enzyme blend for the recovery and function of canine pancreatic islets. Cell Transplant 1998; 7: 365.
26. Brandhorst H, Brandhorst D, Hering BJ, Bretzel RG. Significant progress in porcine islet mass isolation utilizing liberase HI for enzymatic low-temperature pancreas digestion. Transplantation 1999; 68: 355.
27. Somoza N, Vargas F, Armengol MP, Alcalde L, Vives-Pi M, Pujol-Borrell R. Expression of CD14 in human β cells and human insular cell lines. Diabetologia 1994; 37 (suppl 1): 374.
28. Campbell IL, Cutri A, Wilson A, Harrison LC. Evidence for IL-6 production by and effects on the pancreatic beta-cell. J Immunol 1989; 143: 1188.
29. Lou J, Triponez F, Oberholzer J, et al. Expression of alpha-1 proteinase inhibitor in human islet microvascular endothelial cells. Diabetes 1999; 48: 1773.
30. Jirik FR, Podor TJ, Hirano T, et al. Bacterial lipopolysaccharide and inflammatory mediators augment IL-6 secretion by human endothelial cells. J Immunol 1989; 142: 144.
31. Podor TJ, Jirik FR, Loskutoff DJ, Carson DA, Lotz M. Human endothelial cells produce IL-6: lack of responses to exogenous IL-6. Ann N Y Acad Sci 1989; 557: 374.
32. Sandler S, Bendtzen K, Eizirik DL, Welsh M. Interleukin-6 affects insulin secretion and glucose metabolism of rat pancreatic islets in vitro. Endocrinology 1990; 126: 1288.
33. Campbell IL, Kay TW, Oxbrow L, Harrison LC. Essential role for interferon-gamma and interleukin-6 in autoimmune insulin-dependent diabetes in NOD/Wehi mice. J Clin Invest 1991; 87: 739.
34. Paraskevas S, Duguid WP, Maysinger D, Feldman L, Agapitos D, Rosenberg L. Apoptosis occurs in freshly isolated human islets under standard culture conditions. Transplant Proc 1997; 29: 750.
35. Petrik J, Arany E, McDonald TJ, Hill DJ. Apoptosis in the pancreatic islet cells of the neonatal rat is associated with a reduced expression of insulin-like growth factor II that may act as a survival factor. Endocrinology 1998; 139: 2994.
36. Thomas FT, Contreras JL, Bilbao G, Ricordi C, Curiel D, Thomas JM. Anoikis, extracellular matrix, and apoptosis factors in isolated cell transplantation. Surgery 1999; 126: 299.
37. Rabinovitch A, Suarez-Pinzon WL, Shi Y, Morgan AR, Bleackley RC. DNA fragmentation is an early event in cytokine-induced islet beta-cell destruction. Diabetologia 1994; 37: 733.
38. Comstock KL, Krown KA, Page MT, et al. LPS-induced TNF-alpha release from and apoptosis in rat cardiomyocytes: obligatory role for CD14 in mediating the LPS response. J Mol Cell Cardiol 1998; 30: 2761.
© 2001 Lippincott Williams & Wilkins, Inc.