Share this article on:

Refinement of the Automated Method for Human Islet Isolation and Presentation of a Closed System for In Vitro Islet Culture

Goto, Masafumi1,8; Eich, Torsten M.1; Felldin, Marie2; Foss, Aksel3; Källen, Ragnar4; Salmela, Kaija5; Tibell, Annika6; Tufveson, Gunnar7; Fujimori, Keisei1; Engkvist, Margareta1; Korsgren, Olle1

doi: 10.1097/01.TP.0000140882.53773.DC
Clinical Transplantation

Background. The procedure of human islet isolation needs further optimization and standardization. Here, we describe techniques to enhance enzymatic digestion and minimize mechanical forces during the digestion process. The isolation protocol has also been modified to meet current GMP (cGMP) standards. Moreover, the impact of donor- and process-related factors was correlated to the use of islets for clinical transplantation.

Methods. One hundred twelve standardized consecutive islet isolations were evaluated. Metyltioninklorid and indermil (topical tissue adhesive) were applied to detect leakage of collagenase injected and to repair the damaged pancreatic glands. The effects of dye and glue were evaluated in terms of islet yield, islet function using the perifusion assay, and success rate of the isolation. To analyze key factors for successful isolations, both univariate and multivariate regression analysis were performed.

Results. Both Metyltioninklorid and Indermil were effective to prevent leakage of enzyme solutions from the pancreatic glands. Both islet yield and success rate were higher when these tools were applied (4,516.1±543.0 vs. 3,447.7±323.5, P=0.02; 50.0% vs. 21.3%, P=0.02, respectively). No adverse effects on islet function or collagenase activity were observed. Multivariate regression analysis identified the maximal recorded amylase >100 U/L (P=0.026), BMI (P=0.03), and the use of catecholamine (P=0.04) as crucial donor-related factors. In addition, cold ischemia time (P=0.005), the dissection procedure using whole glands with duodenum (P=0.02), and the local procurement team (P=0.03) were identified as crucial isolation-related variables.

Conclusions. A standardized technique of islet isolation is presented applying novel means to improve enzymatic digestion and to meet cGMP standards.

1 Department of Radiology, Oncology and Clinical Immunology, Division of clinical Immunology, University Hospital, Uppsala, Sweden.

2 Department of Surgical Sciences, Division of Transplantation Surgery, University Hospital, Gothenburg, Sweden.

3 Department of Transplantation Surgery, University Hospital, Oslo, Norway.

4 Department of Nephrology and Transplantation, University Hospital, Malmö, Sweden.

5 Division of Transplantation, Surgical Hospital, Helsinki University, Helsinki, Finland.

6 Department of Transplantation Surgery, Karolinska Institute, Stockholm, Sweden.

7 Department of Surgical Sciences, Division of Transplantation Surgery, University Hospital, Uppsala, Sweden.

This study was supported by grants from the Swedish Medical Research Council (16P-13568 and 16X-12219), the Åke Wiberg Foundation, the Nordic Insulin Fund, the Torsten and Ragnar Soderbergs Foundation, the Ernfors Family Fund, Barn Diabetes Fonden, the Swedish Diabetes Association, the Juvenile Diabetes Foundation International, the Knut and Alice Wallenberg Foundation, and the Novo Nordic Foundation.

8 Address correspondence to: Masafumi Goto, M.D., Ph.D., Department of Radiology, Oncology and Clinical Immunology, Division of Clinical Immunology, The Rudbeck Laboratory C11, SE-751 85 Uppsala, Sweden. E-mail:

Received 31 March 2004. Accepted 16 June 2004.

The Nordic Network for clinical islet transplantation was initiated 3 years ago with an aim to establish islet transplantation as a therapeutic option for patients with type 1 diabetes. An important step in this process was to optimize the islet isolation for pancreata retrieved from ordinary organ donors without applying any exclusion criteria other than those generally accepted for organ donation, e.g., kidneys. Also, techniques had to be developed to ascertain the use of organs partially damaged during the harvest procedure. The process of islet isolation from the human pancreas is frequently described in the literature (1–4). However, most reports use reagents not normally in clinical practice or not approved according to current GMP (cGMP) standards. Here, we describe the use of substances and solutions that meet these criteria and in most instances also are to be preferred due to low costs.

A key component in islet isolation is the source of collagenase (5, 6). However, one so far overlooked aspect is the delivery of collagenase to the pancreas. Most isolation teams dissolve the enzyme in 250–300 mL of fluid (1, 6). Of this volume only about 1/3 can be contained in the pancreas. The efficiency of collagenase present in the solution is almost negligible compared with that within the pancreas (7–9). A novel distension procedure was introduced together with techniques to detect and seal damaged areas of the pancreas. By applying these techniques, it was possible to ascertain optimized digestion of most pancreata.

A key component of a Network is transportation of isolated islets between the islet isolation unit and participating transplantation centers. The logistics are by obvious reasons different than those applied within a single center. The so far described procedures for culturing isolated human islets are not optimized to meet the requirements of the newly isolated islets and the procedures are rarely adjusted to cGMP standards. A newly designed islet culture system utilizing a closed bag system that readily meets both the functional and regulatory demands is presented.

To further increase the success rate of islet isolation, analysis of the influence of donor backgrounds and islet isolation-related variables were correlated to the outcome of standardized isolation of 112 consecutive human pancreata.

Back to Top | Article Outline


Variables Related to the Donor, Organ Procurement, and Isolation Procedure

Between March 2000 and October 2003, 350 pancreata were obtained from brain-dead cadaveric multiorgan donors with informed consent either from the organ donor registry or from relatives. The study was approved by the local ethics committee. Between August 2002 and October 2003, the outcome of 112 standardized consecutive islet isolations were correlated to donor- and isolation-related factors. The following factors were retrospectively analyzed from donor charts: age, gender, body mass index (BMI), weight of pancreas, hospitalization length, the history of cardiac arrest, the history of alcohol abuse, and treatment with catecholamines, furosemide, and steroids. Maximal creatinine, γ-GT, CRP, ALAT, ASAT, ALP, and glucose levels, as well as minimal glucose levels, were also analyzed. Cause of death, recorded maximal amylase, and minimal systolic blood pressure (BP) were categorized into two groups (traumatic or nontraumatic, ≥ or < 100 U/L, and ≤ or >90 mm Hg, respectively). Organ procurement- and isolation-related factors included cold ischemia time (CIT), dissection procedures, procurement team, collagenase lot, the use of additional collagenase, the use of dye, collagenase amount per gram of pancreas, harvest starting time (so called “digestion stop time”), dissection time, and digestion rate.

Back to Top | Article Outline

Islet Isolation

Four distant centers (Oslo, Helsinki, Gothenburg, and Malmo) and two local centers (Uppsala and Stockholm) participated. Cold ischemia time was defined as the time from cross-clamp of the aorta to the time of the initiation of dissection procedures at the laboratory. The isolation procedure used is a refinement of the automated method originally described by Ricordi et al. (3). One group of pancreata (n=26) were obtained en bloc together with duodenum (Whole pancreas with duodenum: WPD group [Fig. 1C]), whereas the pancreata were dissected from the duodenum at the operation room in 86 isolations. In 28 of these 86 isolations, the whole glands including the head and uncus area of the pancreas were used (Whole pancreas: WP group [Fig. 1B]), while only the body and tail area were applied in remaining 58 isolations (Partial pancreas: PP group [Fig. 1A]). The main pancreatic duct was cannulated with a Venflon IV cannula (Becton Dickinson Infusion therapy AB, Helsingborg, Sweden) and 10 mL cold Hanks’ balanced salt solution (HBSS) containing Liberase (Roche, Roche Diagnostica, Indianapolis, IN), 20 mM HEPES (GIBCO BRL, Paisley, Scotland), 2 mM CaCl2 (BDH Chemicals Ltd., Poole, United Kingdom), and 4 mM Pefabloc SC PLUS (Roche, Roche Molecular Biochemicals, Indianapolis, IN) (10) with (n=23) or without (n=89) Metyltioninklorid (Apoteket AB, Stockholm, Sweden) was injected. Damaged areas of the pancreata were readily identified after injection of only a few microliters of the dye (Fig. 1D) and repaired by clamping, suturing, electrocautery, or Indermil (Loctite Ireland Ltd., Dublin, Ireland; Fig. 1E). This was followed by continuous perfusion with an additional 50 mL of cold collagenase solution under a constant pressure of about 80 mm Hg using Pharmacia LKB Pump (Pharmacia LKB Biotechnology, Uppsala, Sweden) at 5.3 mL/min. Subsequently, HBSS (10–40 mL) was gradually added to saturate the distention of the gland according to the need of each pancreas. In contrast to previously published studies (1), pancreas dissection was performed during ongoing intraductal perfusion to save time and to facilitate dissection using the space formed between the pancreatic parenchyma and surrounding tissues. After about 30 min, the pancreas was cut into 3–4 pieces and transferred to the Ricordi chamber. To retain undigested tissue and avoid outflow obstruction, a cassette of three stainless steel filters of increasing pore size (0.4, 1.0, and 1.9 mm) were positioned in the upper portion of the digestion chamber. Ringer Acetate solution (1,000 mL, Braun, Melsungen, Germany) supplemented with 5 mL of penicillin-streptomycin (GIBCO BRL), 10 mM nicotinamide (NIC; Sigma-Aldrich, St. Louis, MO), 0.2 mL Pulmozyme (Roche AB, Stockholm, Sweden), 5 mL NaHCO3 (7,5%; GIBCO BRL), and 2.5 mM glucose (Fresenius Kabi, Uppsala, Sweden) was circulated continuously (200 mL/min) through the system. The temperature was kept at 37±1°C throughout both the digestion and harvest procedures using a blood heater device (Biegler Protherm II;E. Biegler GmbH, Mauerbach, Germany). During the first 15 min, the chamber was kept in upright position without shaking, after which it was shaken vertically at a frequency of 220 cycles/min at an amplitude of 5.5 cm. During harvesting, 6–8 l of Ringer Acetate was pumped through the system to gradually empty the tissue contents of the chamber. Throughout the harvesting procedure, the digestion chamber was tilted back and forth about 30° to enhance release of digested materials from the chamber. The chamber was only occasionally shaken by hand. The effluent was collected in conical 250 mL centrifuge tubes containing 3–5 mL of ABO compatible human serum and centrifuged at 224 g at 4°C for 1 min. Pellets were pooled (≈25 mL), resuspended in cold UW solution with a density 1.2 times higher than that of normal UW, and incubated on ice for 60 min. To allow convenient and sterile transfer of the various solutions, transfusion kits (Transfusionsaggregat; JMS Singapore pte Ltd.) and connectors (REF EMC 1401; Baxter Medical AB, Eskilstuna, Sweden) were used. Separation of endocrine and exocrine tissues were performed applying a continuous density gradient (created by mixing Biocoll [Biochrom Seromed KG, Berlin, Germany] of two densities 1.077 and 1.100 g/mL) in a refrigerated COBE 2991 centrifuge (COBE Blood Component Technology, Lakewood, CO) (1, 11, 12). After centrifugation, 20 mL fractions were harvested in 25 cm2 Corning flasks (Corning Incorporated, NY) containing 30 mL of cold Ringer Acetate with 2.5% human serum. Selected fractions were immediately pelleted and washed twice by centrifugation at 224 g. Thereafter, the islets were suspended in CMRL 1066 (ICN Biomedicals, Costa Mesa, CA) supplemented with 10 mM HEPES, 2 mM L-glutamin, 50 μg/mL Gentamycin, 0.25 μg/mL Fungizone (GIBCO BRL), 20 μg/mL Ciproxfloxacin (Bayer healthcare AG, Leverkusen, Germany), and 10 mM NIC and cultured in either untreated culture dishes or flasks Sterilin (Bibby Sterilin Ltd., Stone, United Kingdom) or in a culture bag system Fenwal (Baxter Medical AB) at 37°C in 5% CO2 and humidified air for 1 to 4 days. The culture medium was changed on day 1 and then every other day. The volume and purity of the islets were determined by microscopic sizing on a grid after staining with dithizone and then converted to the standard number of islet equivalents (IEQs). Islet function was assessed as insulin secretion in response to a glucose challenge in a dynamic perifusion system Suprafusion 1000 (BRANDEL, Gaithersburg, MD; 1.67, 16.7, and 1.67 mM glucose). The stimulation index (SI) was defined as the ratio of the total amount of insulin secreted during the high glucose stimulation and that released during low glucose. The isolation was defined as successful if the islets were used for clinical transplantation. In the Nordic Network, the islets are usually transplanted if (1) there is a sufficient amount of islets at the time of transplantation (about 5,000 IEQs/kg of body weight in recipients), (2) there is less than a 20% loss of islets during the 1–4 days in culture, (3) the integrity of the islets remains intact, and (4) specific islet function is demonstrated using the dynamic perifusion glucose challenge assay.



Back to Top | Article Outline

Endotoxin and Collagenase Activity

The amount of endotoxin in Metyltioninklorid was determined using Endosafe KTA2 (Charles River Laboratories International, Inc., Wilmington, DE). Collagenase activity was evaluated by measuring PZ activity (U/mL) according to Wuensch et al. (13).

Back to Top | Article Outline

Statistical Analysis

All statistical analyses were performed using the Statistical Package for the Social Science (SPSS 10.0 for Macintosh; SPSS Inc., Chicago, IL). In univariate analysis, all values are shown as means ± SEM. Continuous variables were compared using two-sided t tests for unpaired observations or one-way factorial ANOVA. Categorical variables were compared using chi-square tests. Correlation was analyzed using Pearson’s correlation tests. According to the results of univariate analysis, stepwise multivariate logistic regression analysis was performed to identify significant affecting factors on the outcome of isolations. Stepwise multivariate regression was performed to evaluate the influence of each variable on the yield and function of islets. The data of islet yield was log-transformed to achieve a normal distribution.

Back to Top | Article Outline


Results of Consecutive 112 Human Islet Isolations

The mean age and BMI were 52.0±1.0 and 25.7±0.4, respectively. The mean pancreas weight and CIT were 85.9±2.6 g and 488.5±19.4 min, respectively. After digestion and harvest, 9.91±0.84 g of pancreatic tissue (including membranes and connective tissue) remained in the chamber. The degree of digestion was 88.5%±0.9%. The mean pellet volume before and after purification procedures was 35.4±1.4 and 1.6±0.2 mL, respectively. The mean islet yield from all isolations was 305,817±19,144 IEQs or 3,631±286 IEQs/g. The mean islet yield from successful isolations (n=29) was 502,133±38,388 IEQs or 6,097±678 IEQs/g, and 237,225±16,493 IEQs or 2,819±249 IEQs/g from failed isolations (n=83). The mean purity of the entire islet preparation and of the pure fraction was 48.9%±1.8% and 78.6%±1.2%, respectively. The mean SI in the dynamic perifusion test was 3.90±0.34.

Back to Top | Article Outline

Effect of Dye Solutions on the Outcome of Islet Isolation

No endotoxin was detectable in the dye solutions (<0.15 EU/mL, triplicate) and the dye had no adverse effect on collagenase activity (Dye-: 33.3±1.3 U/mL, Dye+: 32.1±1.6 U/mL; P=0.59) (Table 1). Islet yield was higher in the isolations where dye solutions were used than those without dye (376,820±36,246 vs. 287,468±21,868; P=0.009). Although the weight of pancreata was higher in the dye-applied group (94.6±4.7 vs. 83.6±3.1; P=0.03), the IEQs/g was also significantly higher (4,207±451 vs. 3,469±342; P=0.02). Successful isolation rate also increased, however, this difference was not statistically significant (34.8% [8/23] vs. 23.6% [21/89]; P=0.27). Notably, if we only analyze the isolations after introducing dye solutions, dye was applied in 75.0% (9/12) of successful isolations, whereas it was used in 48.3% (14/29) of failed isolations (Table 2). No adverse effects on islet function were observed in the case of using dye solutions during isolation procedures (SI: 3.72±0.69 vs. 3.96±0.39; P=0.76).





Back to Top | Article Outline

Effect of Glue on the Outcome of Islet Isolation

Both islet yield and success rate were higher in isolations where glue was used comparing with those without glue (Table 1). However, since the application of glue was introduced late in the series, these differences did not reach statistical significance (IEQs: 404,348±65,524 vs. 300,240±19,805; P=0.08, IEQs/g: 4,479±809 vs. 3,579±299; P=0.14, success rate; 50.0% (3/6) vs. 22.3% (26/106); P=0.18, respectively). No adverse effects on the islet function were observed when glue was applied in the isolations (SI: 6.85±1.55 vs. 3.71±0.34).

Back to Top | Article Outline

Influence of Donor-Related Variables

Univariate analysis of donor-related variables revealed that maximal recorded creatinine and amylase and medical use of furosemide and steroid significantly affected the success rate of the isolations (P=0.02, 0.03, 0.01, and 0.046, respectively) (Table 2). Although both the age and BMI of donors failed to demonstrate significant differences between successful and failed isolations (P=0.28 and 0.25), donors more than 55 years old were associated with increased success rate (35.6% [16/45] vs. 19.4% [13/67]; P=0.06, respectively). Also, BMI more than 26.3 was significantly associated with high success rate compared with the others (40.6% [13/32] vs. 20.3% [16/79]; P=0.03). Furthermore, high BMI was related to increased islet yield and improved islet function (IEQs/g: 4,134±493 vs. 3,448±346; P=0.12, SI: 4.60±0.68 vs. 3.63±0.38; P=0.18).

Back to Top | Article Outline

Influence of Organ Procurement- and Isolation-Related Variables

Univariate analysis of organ procurement- and isolation-related variables revealed that CIT, the procedures of pancreas dissection, procurement team, and collagenase lot significantly affected the success rate of the isolations (P=0.0002, 0.001, 0.004, and 0.02, respectively) (Table 2). CIT was significantly shorter in the successful isolation group compared with the failed group (396.0±22.4 vs. 520.9±24.1; P=0.0002) (Table 2). In addition to the correlation with the success rate of isolations, CIT was also correlated to the islet function (r=−0.26, P=0.005; r=−0.24, P=0.03, respectively). Furthermore, long CIT defined as more than 7 hr caused significant deterioration not only in the success rate but also in islet yield (IEQs/g) and islet function (SI) (Table 3). The success rate of isolations and IEQs were significantly higher in the WPD group (WPD 46.2% [12/26] vs. WP 14.3% [4/28] vs. PP 22.4% [13/58]; P=0.02, WPD 421,760±35,660 vs. WP 249,239±24,801 vs. PP 282,542±29,532; P=0.003), whereas the pancreas weight was significantly higher in WP group than in WPD group (116.3±4.4 g vs. 95.8±3.6 g; P<0.0001) (Table 3). This data indicate that the beneficial effect of WPD group cannot be explained by only the size of pancreata. Interestingly, the number of IEQs/g was associated with that of the digestion rate (Fig. 1A–C, Table 3), suggesting that the favorable effect of WPD group could be attributed to improved dispersion of collagenase and more efficient digestion. There was no significant difference in islet function amongst all groups (P=0.65) (Table 3). The success rate was 50% (11/22) when pancreata were procured by local team, while it was only 20% (18/90) when distant team procured (P=0.004, Table 3). Vast variation in effectiveness was seen amongst different collagenase lots. 93088620 and 93126020 were associated with high success rate (38.7% and 40.0%; defined as efficient), while other lots were linked with low success rate (23.8%, 17.6%, and 14.3%; defined nonefficient). Islet function was also influenced by collagenase lot (Table 3). Collagenase amount more than 5.5 mg/g was significantly associated with the increase of islet yield (IEQs/g) and digestion rate (%) (4,072±350 vs. 2,218±294, P=0.0001; 90.1±0.9 vs. 83.1±1.9, P=0.0007). The success rate was also higher when the total amount of collagenase was more than 5.5 mg/g of pancreas (29.1% vs. 15.4%, P=0.16).



Back to Top | Article Outline

Logistic Regression Analysis of Affecting Factors for Successful Isolations

Based upon the univariate analysis, 12 donor-related variables and 9 organ procurement- and isolation-related variables were included in the stepwise multivariate logistic regression analysis. Analysis of donor-related variables identified high BMI (P=0.03, odds ratio [OR] 1.29) as positively correlated factors, and the maximal recorded amylase >100 U/L (P=0.026, OR 0.03) and the use of catecholamine (P=0.04, OR 0.05) as negatively correlated factors (Table 4). The regression analysis had an overall prediction accuracy of 86.8%, with 88.9% and 86.2% accuracy for predicting successful and failed isolations, respectively. Analysis of organ procurement- and isolation-related variables identified the local procurement team (P=0.03, OR 4.05) and WPD (P=0.02, OR 4.00) as positively correlated factors and long CIT (P=0.005, OR 0.995) as negatively correlated factors (Table 4). The regression model had an overall prediction accuracy of 79.6%, with 68.8% accuracy for predicting successful isolations and 81.7% accuracy for predicting failed isolations.



Back to Top | Article Outline

Multivariate Regression Analysis of Affecting Factors for the Yield and the Function of Islets

The same variables studied in the logistic regression analysis were included in multivariate regression analysis (Table 5). Islet equivalents per gram of pancreas (IEQs/g) were used as outcome variables for the islet yield and log-transformed to approximate a normal distribution. Forward stepwise multivariate regression analysis of donor-related variables identified the minimal recorded systolic blood pressure ≤90 mm Hg (P=0.009, β=- 0.38), maximal recorded amylase >100 U/L (P=0.024, β=-0.19), and the medical history of steroids (P=0.046, β=-0.22) negatively related to islet yield. Regression analysis of organ procurement- and isolation-related variables identified high amount of enzyme per gram of pancreas (P<0.001, β=0.054), WPD (P=0.004, β=0.21), and efficient collagenase lot (P=0.01, β=0.159) positively related to islet yield. Harvest starting time (P=0.001, β=−0.015) was negatively associated with islet yield. Multivariate regression analysis of organ procurement- and isolation-related variables identified harvest starting time (P=0.008, β=−0.128) as negatively related to the islet function.



Back to Top | Article Outline


The overall strategy governing the optimization of the automated method for islet isolation was the notion that the islets must be liberated from surrounding exocrine tissues by enzymatic and not mechanical forces. To limit adverse effects on the islets during the isolation process the processing time must be minimized. The described technique readily allows three persons to complete the procedure from the arrival of the pancreas at the islet laboratory to the time when the islets are placed in culture in only 4 hr.

To retain collagenase activity in the pancreas, we introduced the use of clinical available dye solution and thrombin-based tissue glue to detect and repair damaged areas without creating any harm to the pancreatic gland. The use of Metyltioninklorid revealed that in almost all pancreata leakage was found in one or several locations. Electrocautery, clamping, and suturing of the affected area was not suitable due to the inherent characteristics of the pancreatic parenchyma. In contrast, tissue glue readily closes even large cuts in the parenchyma in only a few minutes. Importantly, these reagents have no adverse effects on collagenase activity or on islet function. The pancreas should be retrieved using the same technique as applied for whole pancreas transplantation. Careful detachment of the pancreas from the duodenum and ligation of accessory duct(s) can readily be made at the islet laboratory. When these procedures were combined with the use of dye and tissue glue, the amount of islets retrieved was significantly increased as was also the success rate (50.0% vs. 21.3%, respectively, P=0.02). Notably, the success rate of our recent 50 isolations has dramatically increased comparing with that of our initial 50 isolations where these techniques were not applied (32.0% vs. 14.0%, respectively, P=0.03).

Several details of the isolation process were changed to allow a controlled enzymatic digestion of the entire pancreas. HBSS or culture media was replaced by Ringer Acetate during the phases of digestion and harvest. Ringer Acetate is supplied at low costs in a standardized sterile system that can easily be applied to create a cGMP system. Importantly, the buffer system in Ringer Acetate is based on acetic acid, whereas the buffer system present in culture media and HBSS requires the presence of carbon dioxide. The temperature was maintained at 37°C during both the digestion and harvesting procedures. This enables an ongoing digestion of tissue fragments remaining in the Ricordi chamber, utilizing the highest enzymatic activity possible combined with minimal mechanical forces, during both digestion and harvesting phases. This is in contrast to most other centers that reduce the temperature between the digestion and harvesting phases from about 37°C to 20°C. The cassette of filters ascertains a continuous release of tissue-fragments without obstruction from the chamber. Collagenase activity is promptly stopped by addition of human serum. The originally described use of FCS for this purpose was recently replaced by albumin to avoid the use of products of animal origin (14). However, albumin available for clinical use does not contain protease inhibitors and consequently has no or only marginal inhibitory effect on the collagenase activity.

The culture bag system applied was originally developed for use at blood banks. It fulfills the requirements for cGMP standards and is designed to store platelets, one of the stickiest cell types. The advantage of using this bag system is the high sterility and safety of the system, no adherence of isolated islets, and a rapid and convenient handling of large tissue volumes at the time of explantation, media exchange, as well as harvest of the islets. Theoretically, the bag system could also provide better oxygenation of islets in culture compared with ordinary flasks or dishes. This, since oxygen readily passes through the plastic wall, i.e., also from underneath the bag. Thus far, no adverse effects on the islet function were observed when bag system was applied (SI of dishes; 3.50±0.38 vs. SI of bag system; 8.79±1.37, P=0.12).

The most significant donor-related variable affecting the outcome of islet isolation was the maximal recorded amylase level. Our result is consistent with the previous study by Lakey et al. (15). However, Zeng et al. (16) did not find any adverse effects of high amylase. This may be due to an insufficient number of donors analyzed (n=8). BMI was also correlated with successful isolation. This is also in accord with previous reports (15, 17). Furthermore, stepwise logistic regression analysis found that the use of catecholamines was negatively correlated with successful isolation. Also, this observation is in agreement with that of Lakey et al. (15). Interestingly, Hering et al. reported failed graft function in two recipients of six patients having received islets from donors treated with catecholamines (18).

The most significant organ procurement- and isolation-related variable affecting the outcome of the isolation procedure was CIT. This observation is consistent with several previous reports (9, 15–17, 19). In the present study, 7 hr of CIT was the time discriminating successful from failed isolations. The influence of CIT may change when the two-layer method for pancreas procurement is adopted (18, 20, 21). The amount of collagenase per gram of pancreas was correlated with high islet yield by multivariate regression analysis. Collagenase more than 7.0 mg/g increased the yield of islets (IEQs/g; 4,515±397 vs. 2,993±382, P=0.008, with a success rate of 30.6% vs. 22.2%, respectively) indicating that additional collagenase should be added if the pancreas weight is more than 70 g. This finding was more pronounced (Table 3) when the amount of collagenase was set to 5.5 mg/g, suggesting that additional enzyme should be added to a pancreas with a weight over 90 g. In contrast to previous reports (5) an obvious difference in isolation outcome, dependent on the batch of Liberase used, was observed.

Retrospectively, if a pancreas retrieved from a donor with a combination of variables related to adverse outcome, i.e., maximal amylase ≥ 100 U/L, BMI <26.3, and use of catecholamine, no islet transplantation could be performed (0/13). Notably, the same outcome was observed (0/16) when considering only maximal amylase activity ≥100 U/L and BMI <26.3.

In conclusion, the present study demonstrates that the novel techniques introduced, i.e., the use of dye and tissue glue to optimize collagenase delivery to the gland, and the careful detachment of the pancreas from the duodenum at the isolation laboratory markedly improve the outcome of human islet isolation. Furthermore, the multivariate statistical analysis suggests that the selection of donors, especially based upon the maximal amylase, BMI, the use of catecholamines, and CIT could further increase the success rate of islet isolation.

Back to Top | Article Outline


We thank Magnus Ståhle, Anette von Malmborg, Ulrika Johansson, Ingrid Skarp, Ingela Carlsson, and Nordmark Arzneimittel GmbH for their excellent technical assistance, and Peter Schmidt for editing the text.

Back to Top | Article Outline


1. Lakey JR, Warnock GL, Shapiro AM, et al. Intraductal collagenase delivery into the human pancreas using syringe loading or controlled perfusion. Cell Transplant 1999; 8: 285.
2. Goss JA, Goodpastor SE, Brunicardi FC, et al. Development of a human pancreatic islet-transplant program through a collaborative relationship with a remote islet-isolation center. Transplantation 2004; 77: 462.
3. Ricordi C, Lacy PE, Finke EH, et al. Automated method for isolation of human pancreatic islets. Diabetes 1988; 37: 413.
4. Gray DW, McShane P, Grant A, et al. A method for isolation of islets of Langerhans from the human pancreas. Diabetes 1984; 33: 1055.
5. Linetsky E, Bottino R, Lehmann R, et al. Improved human islet isolation using a new enzyme blend, liberase. Diabetes 1997; 46: 1120.
6. Brandhorst H, Brandhorst D, Hesse F, et al. Successful human islet isolation utilizing recombinant collagenase. Diabetes 2003; 52: 1143.
7. Gotoh M, Maki T, Satomi S, et al. Reproducible high yield of rat islets by stationary in vitro digestion following pancreatic ductal or portal venous collagenase injection. Transplantation 1987; 43: 725.
8. Horaguchi A, Merrell RC. Preparation of viable islet cells from dogs by a new method. Diabetes 1981; 30: 455.
9. Warnock GL, Ellis D, Rajotte RV, et al. Studies of the isolation and viability of human islets of Langerhans. Transplantation 1988; 45: 957.
10. Lakey JR, Helms LM, Kin T, et al. Serine-protease inhibition during islet isolation increases islet yield from human pancreases with prolonged ischemia. Transplantation 2001; 72: 565.
11. Robertson GS, Chadwick DR, Contractor H, et al. The optimization of large-scale density gradient isolation of human islets. Acta Diabetol 1993; 30: 93.
12. Goss JA, Schock AP, Brunicardi FC, et al. Achievement of insulin independence in three consecutive type-1 diabetic patients via pancreatic islet transplantation using islets isolated at a remote islet isolation center. Transplantation 2002; 74: 1761.
13. Wuensch E, Heidrich HG. [on the Quantitative Determination of Collagenase]. Hoppe Seylers Z Physiol Chem 1963; 333: 149.
14. Shapiro AM, Lakey JR, Ryan EA, et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 2000; 343: 230.
15. Lakey JR, Warnock GL, Rajotte RV, et al. Variables in organ donors that affect the recovery of human islets of Langerhans. Transplantation 1996; 61: 1047.
16. Zeng Y, Torre MA, Karrison T, et al. The correlation between donor characteristics and the success of human islet isolation. Transplantation 1994; 57: 954.
17. Toso C, Oberholzer J, Ris F, et al. Factors affecting human islet of Langerhans isolation yields. Transplant Proc 2002; 34: 826.
18. Hering BJ, Kandaswamy R, Harmon JV, et al. Transplantation of cultured islets from two-layer preserved pancreases in type 1 diabetes with anti-CD3 antibody. Am J Transplant 2004; 4: 390.
19. Benhamou PY, Watt PC, Mullen Y, et al. Human islet isolation in 104 consecutive cases. Factors affecting isolation success. Transplantation 1994; 57: 1804.
20. Kuroda Y, Kawamura T, Suzuki Y, et al. A new, simple method for cold storage of the pancreas using perfluorochemical. Transplantation 1988; 46: 457.
21. Matsumoto S, Qualley SA, Goel S, et al. Effect of the two-layer (University of Wisconsin solution-perfluorochemical plus O2) method of pancreas preservation on human islet isolation, as assessed by the Edmonton Isolation Protocol. Transplantation 2002; 74: 1414.

Human islet isolation; Dye; Tissue glue; Dissection procedure; Multivariate statistical analysis; Culture bag system

© 2004 Lippincott Williams & Wilkins, Inc.