Abstract
For more than a decade Liberase HI was commonly used as the standard enzyme blend for clinical human islet isolation until enforced replacement by collagenase NB1 (NB1). This change resulted initially in a reduction in islet isolation outcome and transplant activities worldwide. This retrospective study was initiated to compare the efficiency of NB1 premium grade with Liberase in 197 human islet isolations. All pancreata were processed between January 2006 and June 2008 utilizing the same procedures for isolation and quality assessment except the administration of preselected lots of either Liberase (n = 101) or NB1 (n = 96). Utilizing Liberase, significantly more digested tissue and purified islet yield was produced compared to NB1. In contrast, the use of NB1 was associated with significantly higher purity and glucose stimulation index during dynamic perifusion. The expression of proinflammatory markers was almost identical except tissue factor expression, which was higher after utilization of Liberase. No difference was found in the percentage of pancreata fulfilling the criteria for clinical islet transplantation. The results suggest that Liberase is more efficient for pancreas dissociation than collagenase NB1 but seems to be more harmful to exocrine cells and islet tissue.
Introduction
Recent studies demonstrated the feasibility of human islet allotransplantation to ameliorate long-term complications in type 1 diabetic patients (12, 25, 28). However, to establish islet transplantation as a regular treatment for type 1 diabetic patients it is essential to isolate frequently a sufficient islet mass from a single donor pancreas (1, 23). To achieve this goal it is of utmost importance to improve the low efficiency of the enzymatic dissociation process, which seems to depend mainly on the quality of the administered enzyme blend (4, 30).
For more than a decade Liberase HI had been the commonly used standard enzyme blend for human islet isolation (11, 17). In 2007 Liberase was replaced in clinical islet transplantation by Serva collagenase NB1 in order to fulfill guidelines for quality assurance. This major change in the isolation procedure has contributed to the reduction in islet isolation outcome and transplant activities worldwide (1). This is in contrast to a previous initial study in 18 human donors, suggesting that collagenase NB1 is superior compared to Liberase HI with regard to islet yield and quality (9). Because no comparison between Liberase HI and collagenase NB1 has been performed yet in a larger population of human donor pancreata, the present study was initiated to analyze the efficiency of both enzymes in nearly 200 human islet isolations.
Materials and Methods
This retrospective analysis includes the data of 197 consecutive human islet isolations performed at the islet isolation center of Uppsala between January 2006 and June 2008. All pancreata were processed utilizing essentially the same procedures except the application of different enzyme blends. During the first 14 months of observation three lots of Liberase HI (Roche, Indianapolis, IN, USA) were used for pancreas digestion. Afterwards, this enzyme blend was replaced by Serva collagenase NB1 (Serva, Heidelberg, Germany) due to regulatory and safety reasons. In the subsequent 14 months five batches of collagenase NB1 were utilized for islet isolation. Prior to final selection each Liberase and NB1 lot was prescreened in three to four donor pancreata according to digestion time, islet cleavage, islet yield, and in vitro function.
Pancreas Procurement
This study was approved by the local ethics committee. Once legal consent had been given pancreata were procured from multiorgan donors with total brain infarction utilizing cold perfusion with University of Wisconsin solution (UWS; ViaSpan©, DuPont Pharmaceuticals Ltd., Herts, UK) or histidine-tryptophan-ketoglutarate (HTK; Köhler Chemie GmbH, Alsbach, Germany). Exclusion and inclusion criteria were applied as previously described (13). Donor pancreata were provided by the Nordic Network for Clinical Islet Transplantation consisting of six donor centers (Helsinki, Finland; Oslo, Norway; Gothenburg, Malmö, Stockholm, Uppsala, Sweden) and processed in a central isolation unit located in Uppsala (24).
Islet Isolation
Isolation, purification, and culture of human islets were performed as previously described (6, 13). Briefly, all enzyme blends were reconstituted with cold Ringer's acetate solution (RAS, Braun, Melsungen, Germany) supplemented with 3 mM CaCl2 (BDH Chemicals, Poole, UK). Collagenase NB1 was additionally supplemented with neutral protease (NP, Serva). Enzyme amount and perfusion volume were adjusted according to the trimmed pancreas weight: for pancreata smaller than or equal to 100 g one vial of either Liberase or collagenase NB1 plus 130 DMC-U NP was dissolved in 60 ml RAS. In the case that the pancreas weight was larger than 100 g, enzyme amount was increased to 1.3 vials of either Liberase or collagenase NB1 supplemented with 150 DMC-U NP and dissolved in a perfusion volume of 100 ml. Enzymes were administered by means of continuous ductal perfusion performed at <40 mmHg for 15 min. Digestion was performed in a continuous digestion-filtration device filled with RAS supplemented with 5.5 mmol glucose, 5 mmol sodium pyruvate, 10 mM nicotinamide, 0.5% penicillin-streptomycin (Gibco-Invitrogen AB, Stockholm, Sweden), and 0.2 μl/ml Pulmozyme (Roche Diagnostics Scandinavia, Bromma, Sweden). After recirculation at 37°C, released tissue fragments were washed for 1 min at 224 × g in cold RAS supplemented as described above, additionally adding 2.5% of ABO-compatible human serum (Blood bank of the University Hospital, Uppsala, Sweden), collected in 200 ml of 1.2-fold concentrated UWS (Apoteket AB, Stockholm, Sweden) supplemented with 5 μl/ml Pulmozyme and stored for 60 min until purification.
Subsequently, the digested tissue was centrifuged for 7 min at 2300 rpm in a refrigerated Cobe 2991 cell processor (Gambro, Lakewood, CO, USA) utilizing a continuous hyperosmolaric Ficoll gradient for islet separation from exocrine tissue (6). After duplicate washing in serum-supplemented RAS, purified islets were resuspended in CMRL 1066 supplemented with 10% ABO-compatible human serum, 1 mM sodium pyruvate, 10 mM nicotinamide, 2 mM L-glutamine, 10 mM HEPES, 50 μg/ml gentamicin, 0.25 μg/ml fungizone (Gibco), and 20 μg/ml ciprofloxacin (Bayer, Leverkusen, Germany). Islet culture was performed for 1–3 days in a semiclosed culture bag system (Baxter Medical AB, Stockholm, Sweden) incubated in a humidified atmosphere (5% CO2) for 1 day at 37°C and subsequently at 24°C (13).
Islet Characterization
Subsequent to purification and after culture islet yield and purity were determined in a standardized procedure (22). All other tests were performed after 24–36 h of culture at 37°C. Duplicate aliquots of 20 equally distributed islets were subjected to dynamic glucose perifusion to assess the stimulation index (16.7/1.67 mmol/L) (13). The intracellular insulin content was normalized to the DNA content determined by means of a fluorometric assay (Quant-it, Invitrogen-Molecular Probes, Stockholm, Sweden). In addition, homogenized islets were also assessed for tissue factor (TF; Imubind, American Diagnostica, Stamford, CT, USA), monocyte chemoattractant protein-1 (MCP-1; Quantikine, R&D Systems, Abingdon, UK), interleukin-6 (IL-6), and IL-8 (R&D Systems).
The criteria for clinical islet transplantation were ≥300,000 islet equivalent (IE), >30% of islet purity, ≤5 ml of purified packed tissue volume, and a biphasic insulin response to dynamic glucose perifusion corresponding to a stimulation index of ≥ 2.0. Islet preparations tested by means of static glucose incubation were excluded from this comparison.
Data Analysis
All statistical analysis was performed utilizing Prism 5.0 for MacIntosh (GraphPad, La Jolla, CA, USA). All results are expressed as mean ± SEM. Comparisons of data were carried out by the unpaired t-test, Fisher's exact test, and chi-square test. Differences are considered significant at p < 0.05; values of p > 0.05 are termed nonsignificant (NS).
Results
Human Pancreas Variables
Pancreas characteristics are shown in Table 1. Donor age, body mass index (BMI), and pancreas weight ranged from 22 to 73 years, 15.4 to 38.1 kg/m2, and 42 to 200 g, and from 16 to 78 years, 17.6 to 40.7 kg/m2, and 36 to 200 g for Liberase and collagenase NB1, respectively. Cold ischemia time varied from 3.3 to 21.8 and 3.1 to 18.0 h for Liberase and collagenase NB1, respectively. None of these variables differed significantly between experimental groups. The percentage of HTK solution or UWS used for pancreas perfusion did not vary between experimental groups. No significant difference was observed either in the proportion of female and male donors procured for Liberase or collagenase NB1 utilization (Table 1) or in the proportion of donor centers allocating organs for the different experimental groups (data not shown).
Human Pancreas Characteristics
BMI, body mass index; UWS, University of Wisconsin solution; HTK, histidine-tryptophan-ketoglutarate; CIT, cold ischemia time; PZ-U, 4-phenylazobenzyloxycarbonyl-L-prolyl-L-leucylglycyl-L-prolyl-D-arginine units, equivalent to Wünsch units.
Islet Isolation Outcome
The amount of collagenase activity infused into the pancreas was in the same range in both experimental groups (Table 1). As shown in Table 2, the average digestion time needed for release of cleaved islets was almost similar for utilization of Liberase or collagenase NB1. The average percentage of undigested tissue remaining in the chamber was lower after utilization of Liberase compared to collagenase NB1 (15.3 ± 1.3% vs. 18.0 ± 1.0%, NS). In correspondence, the amount of packed tissue volume was significantly higher in the Liberase group in comparison to collagenase NB1 (57.3 ± 2.2 vs. 46.9 ± 2.0 ml, p < 0.001). Also, when digested tissue volume was calculated per gram trimmed pancreas weight, Liberase produced significantly more digest compared to collagenase NB1 (0.60 ± 0.02 vs. 0.47 ± 0.02 ml/g, p < 0.001).
Islet Isolation Outcome Utilizing Liberase HI or Serva NB1
IE, islet equivalent.
p < 0.001 by unpaired T-test.
p < 0.01 by unpaired t-test.
In agreement with these observations it was found that Liberase yielded significantly more purified IE per pancreas compared to collagenase NB1 (374,711 ± 20,274 vs. 295,048 ± 13,986 IE, p < 0.01). This difference became even more obvious relating islet yield to the trimmed pancreas weight. As shown in Table 2, purified IE per gram pancreas was significantly lower in the collagenase NB1 group than after Liberase use (4010 ± 232 vs. 2979 ± 149 IE/g, p < 0.001). Comparing final purity after density gradient centrifugation it was found that the utilization of Liberase was associated with a small but highly significant reduction of islet purity (43.0 ± 1.8% vs. 53.8 ± 1.5%, p < 0.001), which did not change during 1–3 days of culture of Liberase- or collagenase NB1-isolated islets (42.1 ± 2.1% vs. 52.2 ± 1.8%, p < 0.001). However, the difference in purity did not affect islet recovery after culture of pancreata processed by either Liberase or collagenase NB1 (Table 2).
Islet Characterization
The details of islet quality assessment are presented in Table 3. It was found that the stimulation index, determined by dynamic glucose perifusion, was significantly higher after pancreas digestion by means of collagenase NB1 than after use of Liberase (7.3 ±1.3 vs. 13.4 ± 1.4, p < 0.05). In agreement with this finding it was observed that the insulin content in Liberase-isolated islets was significantly lower compared to islets released by means of collagenase NB1 (2.4 ± 0.2 vs. 5.0 ± 0.3 ng/ng DNA, p < 0.001).
Human Islet Characterization
TF, tissue factor; MCP-1, monocyte chemoattractant protein-1; IL, interleukin; Tx, transplantation.
p < 0.05 by unpaired t-test.
p < 0.001 by unpaired t-test.
Islet characterization revealed also a highly significant increase in the expression of tissue factor after pancreas dissociation utilizing Liberase in comparison to collagenase NB1 (87.7 ± 18.5 vs. 38.0 ± 3.0 fmol/μg DNA, p < 0.001). In contrast, expression of MCP-1, IL-6, and IL-8 was almost identical regardless of whether Liberase or collagenase NB1 was administered into the pancreas (Table 3).
Considering the final aim of successful islet isolation, it was found that the percentage of islet preparations fulfilling criteria for clinical islet transplantation was slightly but not significantly enhanced in Liberase-isolated islets (58.1%) when compared to islets isolated by means of collagenase NB1 (41.5%, NS).
Discussion
An analysis of more than 400 human islet isolations demonstrated recently that the efficacy of an enzyme blend is extremely important for a successful isolation process and posttransplant function after human islet allotransplantation (21). As a consequence, the selection of a suitable enzyme blend is the first crucial step for the entire process of islet separation including islet transplantation into patients (4, 30). It is therefore not surprising that the enforced change from Liberase to collagenase NB1 has had an enormous impact on clinical islet transplantation observed worldwide (1). Although Liberase HI has not been produced for nearly 2 years, different aspects related to this enzyme still may be of relevance because follow-up products of this enzyme blend will be available in the near future.
The present analysis is the first one that compares the efficiency of Liberase HI and collagenase NB1 in a population of nearly 200 human pancreata processed in the same isolation center. It may reflect exemplarily the situation other isolation centers have experienced as well. It was found that the replacement of Liberase HI by collagenase NB1 resulted in a significant reduction of islet yield. This is in contrast to a previous study in 18 human donors indicating that collagenase NB1 is superior compared to Liberase HI in terms of islet yield and quality (9). Nevertheless, both studies correspond in the observation that the utilization of Liberase is associated with significantly reduced islet purity.
Previous analysis in cultured human islets demonstrated that lower islet purity is a significant risk factor for losing islets during culture (16). However, we did not observe a significant difference between Liberase and collagenase NB1 with regard to islet survival postculture. It cannot be excluded that a reduction of islet purity by approximately 20% may decrease the glucose stimulation index to the extent that was observed after use of Liberase compared to collagenase NB1. Studies in encapsulated rat islets suggest that exocrine tissue significantly reduces the glucose-stimulated insulin response of cultured islets (15). In agreement with this observation it was found in human islets that purity correlates inversely with islet production of TNF-α, IL-1α, and IL-6 (27), cytokines that are associated with decreased stimulated insulin release in human islets (2). In contrast, the present analysis did not detect a significant difference between Liberase- and collagenase NB1-isolated islets with regard to the expression of proinflammatory mediators such as MCP-1, IL-6, and IL-8. An increased expression of tissue factor was measured after use of Liberase despite the continuous addition of nicotinamide to the isolation media (20). Islet-released tissue factor is the initial trigger of prothrombotic reactions after intraportal islet transplantation and correlates with posttransplant outcome in diabetic patients (14, 19). It remains to be clarified whether this finding is of relevance for posttransplant islet function.
Apart from exocrine contaminations, it may be that highly purified enzyme blends contain components that have the capacity to induce proinflammatory pathways in islets. As reported previously, isolated human islets that were exposed to Liberase for 1 h were characterized by the expression of proinflammatory and proapoptotic markers (3). In addition, a significantly decreased insulin release after dynamic glucose challenge and impaired posttransplant function in diabetic immunodeficient mice was observed. The study of Balamurugan et al. (3) suggests that impaired human islet function was induced by internalization of Liberase-derived components. A deterioration of islet in vitro and in vivo function was also observed in rat islets isolated utilizing Liberase HI compared to collagenase P (26). On the other hand, it seems that enzyme internalization is not a characteristic feature of Liberase, because a penetration of islet cells by collagenase was also observed after utilization of FITC-marked collagenase NB1 (10). So far, it is unclear which enzymes are internalized by islet cells but it cannot be excluded that degradation products of collagenase are candidates for this process (4, 5). In a previous study we observed impaired posttransplant function in rat islets isolated by means of degraded collagenase class I and class II (8).
Neutral protease is an essential component of collagenase NB1 that can induce damage in isolated islets as well. Although this enzyme is essential for the dissociation of the peri-islet collagen matrix, neutral protease can significantly reduce islet functional and morphological integrity as demonstrated in rat and human islets (7, 29). The present findings indicate that we found a suitable and effective neutral protease dosage at least for our system. Very little is known about the effect of thermolysin, the complementary protease of Liberase, on islet purity and viability. Considering the enormous importance of neutral proteases for isolation success, it is amazing how little attention has been paid to this factor so far. Therefore, we can only speculate whether a potentially higher potency of thermolysin, as reflected by the higher digest volume and islet yield, is associated with characteristics that are harmful for exocrine and endocrine pancreas cells. Because different assays have been used by Roche and Serva for determination of specific thermolysin and clostridial neutral protease activity, respectively, a proper comparison between these complementary proteases is difficult. This indicates the urgent need for a new assay that utilizes a substrate that can be digested by different proteases (18).
In conclusion, the present comparison demonstrated that the utilization of collagenase NB1 is associated with significantly reduced islet yield compared to the use of Liberase HI. In contrast, islet isolation by means of collagenase NB1 resulted in a significant improvement of islet purity and in vitro function. Our findings may be utilized to optimize the composition of enzyme blends in the future.
Footnotes
Acknowledgments
The authors are grateful to Margareta Engkvist, Nina Mikkola, Magnus Ståhle, Anna Svensson, Anna Andersson, and Elisabeth Wijkström for their excellent work in the isolation facilities and quality assessment laboratories. This study was supported by grants from the National Institutes of Health (USA).