|Home | About | Journals | Submit | Contact Us | Français|
We evaluated 3 commercially available enzymes for pancreatic digestion by comparing key parameters during the islet isolation process, as well as islet quality after isolation.
Retrospectively compared and analyzed islet isolations from pancreata using 3 different enzyme groups: liberase HI (n = 63), collagenase NB1/neutral protease (NP) (n = 43), and liberase mammalian tissue-free collagenase/thermolysin (MTF C/T) (n = 115). A standardized islet isolation and purification method was used. Islet quality assessment was carried out using islet count, viability, in vitro glucose-stimulated insulin secretion (GSIS), glucose-stimulated oxygen consumption rate, and in vivo transplantation model in mice.
Donor characteristics were not significantly different among the 3 enzyme groups used in terms of age, sex, hospital stay duration, cause of death, body mass index, hemoglobin A1c, cold ischemia time, and pancreas weight. Digestion efficacy (percentage of digested tissue by weight) was significantly higher in the liberase MTF C/T group (73.5 ± 1.5 %) when compared to the liberase HI group (63.6 ± 2.3 %) (P < 0.001) and the collagenase NB1/NP group (61.7 ± 2.9%) (P < 0.001). The stimulation index for GSIS was significantly higher in the liberase MTF C/T group (5.3 ± 0.5) as compared to the liberase HI (2.9 ± 0.2) (P < 0.0001) and the collagenase NB1/NP (3.6 ± 2.9) (P = 0.012) groups. Furthermore, the liberase MTF C/T enzymes showed the highest success rate of transplantation in diabetic non-obese diabetic severe combined immunodeficiency mice (65%), which was significantly higher than the liberase HI (42%, P = 0.001) and the collagenase NB1/NP enzymes (41%, P < 0.001).
Liberase MTF C/T is superior to liberase HI and collagenase NB1/NP in terms of digestion efficacy and GSIS in vitro. Moreover, liberase MTF C/T had a significantly higher success rate of transplantation in diabetic NOD Scid mice compared to liberase HI and collagenase NB1/NP enzymes.
Successful human islet isolation is essential for clinical and research applications,1-4 and is greatly influenced by selecting appropriate enzyme(s) for pancreas digestion.5-10 Hence, developing and using consistently high-quality enzymes is a key element for islet isolations. Advancements in obtaining non-good manufacturing practice liberase HI indeed made major progress to achieve successful islet isolation11,12 and allowed for an increase in the number of patients receiving islet transplants.13 However, in 2007, the use of liberase HI for human islet isolations was ceased due to safety concerns of possible prion contamination from bovine tissue-derived raw materials.14 Liberase HI was a mixture of collagenase and neutral protease (NP) and not of highly purified enzymes and hence, lot-to-lot variability was a major concern due to failure to consistently achieve successful islet isolation and long term storage was sporadic.9,12,15 Therefore, there is a great demand for manufacturing GMP-grade, highly pure, and low endotoxin digestion enzyme to replace liberase HI among islet isolation specialists and commercial manufacturers to consistently obtain high quality islets.6,16-18 Consequently, collagenase NB1 supplemented with NP was adopted by many centers globally as an alternative enzyme for pancreas digestion for islet isolation.6,19,20 Indeed, several patients have been transplanted with islets isolated using this enzyme, suggesting that this may be a promising product for islet isolation replacing liberase HI. However, the limitation of collagenase NB1 for its efficacy to isolate islets from younger donors influenced the ultimate islet isolation outcome from different donor populations. Furthermore, it has been reported that the collagenase NB1 enzyme contains degraded collagenase; therefore, higher doses are required to achieve successful isolations.8,21 Collagenase, and NP, and clostripain are produced from Clostridium histolyticum,15 whereas thermolysin is purified from Bacillus thermoproteolyticus rokko.12 Collagenase (class I and II isoforms) supplemented with either Thermolysin or NP is currently used for pancreas digestion. Hence, the combination of CIzyme collagenase HA (containing nondegradable class I [60%] and II [40%]) and NP was found to be effective in isolating islets for transplantation, but are not GMP products.21 Indeed, liberase mammalian tissue-free collagenase/thermolysin (MTF C/T) has been shown to be a promising enzyme for human islet isolation for clinical islet transplantation compared to collagenase NB1/NP.6,22
All types of aforementioned enzymes have been used in islet isolations for research and transplantation purposes. However, the advantage of using a particular enzyme from a specific supplier over another is debatable and often subjectively determined by an isolation team's experiences.21,23 It was reported that the overall isolation outcome of collagenase NB1/NP was comparable with that of liberase HI.5,19 In this retrospective analysis, islet isolations from 221 donor pancreata using 3 commercially available enzymes, liberase HI, collagenase NB1/NP, and liberase MTF C/T enzymes, were compared.
Retrospective data analysis was performed on 221 human islet isolations from cadaveric pancreata. Pancreas donation consent was obtained from donor families for islet transplantation and research. This study was approved by the Institutional Review Board of the Beckman Research Institute of the City of Hope. Donors with elevated hemoglobin A1c (≥6.5%) and donation after cardiac death criteria were excluded from the group for analysis.
Human islet isolations were conducted using the standard islet processing procedure established at City of Hope and described recently.24-26 Briefly, on receiving the pancreas, cleaning and cannulating of pancreas was carried out in a cGMP facility at City of Hope. Then automated perfusion was performed using a perfusion apparatus (BioRep Technologies, Miami, FL) and infusion of one of the following enzyme combinations: liberase HI (Roche Diagnostics, Roche Applied Science, Indianapolis, IN; n = 63), collagenase NB1 with NP (SERVA Electrophoresis GmbH, Heidelberg, Germany; n = 43), or liberase MTF C/T (Roche Diagnostics; n = 115). Once the enzyme perfusion was completed, the distended pancreas was cut into 7 to 10 pieces and loaded into a Ricordi's digestion chamber for digestion at 37°C. During the digestion process, samples were taken every minute and stained with dithizone to visualize the islets and the acinar tissue appearance.27 Switch time, the endpoint of the pancreas digestion, was determined when 50% or more of the islets appeared to be free from acinar tissues. Subsequently, the enzyme digestion was terminated by adding surplus media for enzyme dilution. The tissue was then collected, centrifuged, and pooled in the presence of human serum albumin. Circulating dilution media, along with mechanical dissociation of the pancreatic tissue, was continued until no more free islets were seen. Pooled tissue containing islets was purified using a COBE 2991 and continuous density gradients as described previously.25,28 The islet count and purity was determined microscopically using dithizone staining and total islet count was expressed as islet equivalents (IEQ)29 and islet particle number.25
After the isolation, islets were cultured in Connaught Medical Research Laboratories 1066 medium (pH 7.4) supplemented with human serum albumin (0.5%) and insulin-like growth factor-1 (0.1 μg/mL) at 37°C/5% CO2 for 24 to 72 hours. Postculture islet samples were taken for quality assessments including: islet count, viability, glucose-stimulated insulin secretion (GSIS), glucose-stimulated oxygen consumption rate (ΔOCR),30 and islet transplantation into NOD Scid mice.31
Viability of islets postculture was determined using fluorescent microscopy and fluorescent dyes (fluorescein diacetate/propidium iodide/propidium iodide) following the method used at City of Hope.32 For the GSIS assay, islet aliquots containing approximately 150 IEQ were used from each isolation. Dynamic insulin release, in response to high glucose stimulation, was tested using an in vitro perifusion system following the standard operation procedure used by the City of Hope quality control team.25 Insulin secretion of each effluent was measured using an ELISA kit for human insulin (ALPCO, Salem, NH). Stimulation index (SI) was calculated by dividing the total insulin secreted in high glucose over the total insulin released in basal low glucose for the same period. The SI was expressed as an average of the 2 test results.
Islet assessment for ΔOCR was carried out using 750 IEQ from each isolation, and ΔOCR of islets was measured to evaluate the amount of oxygen consumed during metabolism as previously described in detail.30 The ΔOCR was defined as the difference in measured OCR of islets exposed to 3 mmol/L glucose (Hospira, Inc., Lake Forest, IL) for 15 minutes, followed by exposure to 20 mmol/L glucose for 30 to 45 minutes. Results were expressed in nM O2/100 islets per minute.
Male NOD Scid mice (Jackson Laboratory, Bar Harbor, ME), 10 to 12 weeks of age, were obtained from the Animal Resources Center of Beckman Research Institute of the City of Hope and used as islet recipients. Diabetes was induced in mice by intraperitoneally injecting 50 mg/kg of streptozotocin (Sigma-Aldrich, St Louis, MO) for 3 consecutive days. The mice with hyperglycemia (>350 mg/dL) for at least 2 consecutive days were used as recipients. Three NOD Scid mice were transplanted with an identical number (1200 IEQ) of human islets from same donor under the left kidney capsule. Diabetes reversal was monitored 2 to 3 times per week for 30 days to measure blood glucose levels using a glucometer (LifeScan, Inc., Milpitas, CA). For each isolation, the in vivo graft function was considered successful when two thirds of transplanted mice were normoglycemic (<200 mg/dL) within 2 weeks and maintained normoglycemia levels for more than 3 weeks. The percentage of successful graft function in each enzyme group was calculated by: number of isolations with successful graft function/total number of isolations that transplanted in mice × 100.
GraphPad Prism (GraphPad Software 6.0, La Jolla, CA) was used to analyze the data and generate the figures. One-way analysis of variance was used to compare the 3 enzyme groups, followed by Bonferroni multiple comparisons test to compare the mean values between any 2 groups. Values were expressed as mean ± SEM. For category and percentage variables, χ2 test was used. For all the tests used, P less than 0.05 was considered significant.
Donor characteristics for the 3 different enzyme groups are listed in Table Table1.1. There were no significant differences among the 3 enzyme groups in terms of age, sex, hospital stay duration, body mass index, hemoglobin A1c, cold ischemia time, and pancreas weight. The causes of death were also similar among the 3 groups.
Calculated enzyme concentrations per gram of pancreatic tissue for 3 enzymes are listed in Table Table2.2. Interestingly, the data showed that the ratio of collagenase Wünsch unit to protein content was similar between collagenase NB1/NP and liberase MTF C/T; however, the ratio of NP unit to protein content was different.
The digestion switch time in the liberase HI group was the lowest, but the value was statistically significant only between the liberase HI and the collagenase NB1/NP groups (10.7 ± 0.4 vs 12.3 ± 0.5 min; P = 0.009) (Figure (Figure1A).1A). Digestion efficacy (determined by digested pancreatic tissue weight/initial pancreatic tissue weight before digestion × 100) was significantly higher in the liberase MTF C/T group (73.5 ± 1.5 %) when compared to the liberase HI group (63.6 ± 2.3 %) (P < 0.001) and the collagenase NB1/NP group (61.7 ± 2.9%) (P < 0.001) (Figure (Figure11B).
The IEQ per gram of pancreatic tissue of 3 types of enzymes were not significantly different prepurification (P = 0.374) (Figure (Figure2A),2A), postpurification (P = 0.239) (Figure (Figure2B),2B), and postculture (P = 0.071) (Figure (Figure2C).2C). Furthermore, islet particle number per gram of pancreatic tissue was also not significantly different among the 3 groups (Figure (Figure22).
The percentages of islet preparations yielding more than 250,000 IEQ for liberase HI, collagenase NB1/NP, and liberase MTF C/T were 57% (36 out of 63), 41% (18 out of 43), and 44% (50 out of 113), respectively. Significant difference was observed only between liberase HI and collagenase NB1/NP groups (P = 0.024).
The liberase HI had significantly higher recovery rate postpurification than the collagenase NB1/NP (P = 0.037) (Figure (Figure3A).3A). There was no significant difference between any groups postculture (P = 0.584) (Figure (Figure3B).3B). Tissue volume prepurification in the liberase MTF C/T group was significantly higher than that in the liberase HI group (52.8 ± 1.7 vs 42.2 ± 1.7 mL, P < 0.001) (Figure (Figure4A).4A). After purification, the liberase HI group had the highest mean tissue volume and was significantly higher than the collagenase NB1/NP group (1.4 ± 0.2 vs 0.9 ± 0.1 mL, P = 0.019) (Figure (Figure4B).4B). For each enzyme group, the prepurification IEQ per gram of pancreatic tissue was inversely correlated with donor age (liberase HI: r = −0.294, P = 0.021; collagenase NB1/NP: r = −0.460, P = 0.002; liberase MTF C/T: r = −0.227, P = 0.015) (Figure (Figure5A,5A, D, G). There was no significant correlation between IEQ/g postpurification (and culture) and donor age in all 3 groups (Figure (Figure55).
The islet viability postculture was similar within the 3 groups (Figure (Figure6A).6A). However, the SI for the GSIS was significantly higher in the liberase MTF C/T group (5.3 ± 0.5) when compared to the liberase HI group (2.9 ± 0.2, P < 0.0001) and the collagenase NB1/NP group (3.6 ± 2.9, P = 0.012) (Figure (Figure6B).6B). The ΔOCR of isolated islets from the liberase MTF C/T group was significantly higher than that from the liberase HI group (0.243 ± 0.037 vs 0.116 ± 0.011 nM O2/100 islets per minute, P = 0.014) (Figure (Figure66C).
The islets isolated from the liberase MTF C/T group presented with the highest successful graft function rate (65%) of reversing diabetes after transplanting into diabetic NOD Scid mice (Figure (Figure7).7). This rate was significantly higher than that from the liberase HI (42%, P = 0.001) and the collagenase NB1/NP (41%, P < 0.001) groups (Figure (Figure7).7). The islets from the liberase HI and the collagenase NB1/NP showed a similar success rate of transplantation in mice (Figure (Figure77).
Furthermore, the differences among different lots of same enzymes were also analyzed showing that both liberase HI and collagenase NB1/NP presented with lot-to-lot variability. However, there was no significant lot-to-lot variability observed among the different lots used in liberase MTF C/T enzyme group (Table (Table33).
Human pancreas digestion to free the islets from acinar tissues depends on the utilization of optimum concentrations of collagenase and Thermolysin/NP. However, the lot-to-lot variability, differences in manufacturing methods, and lack of common assays to evaluate enzymatic activity may represent significant problems for standardizing the digestion process across different islet transplant centers.9,12,15 Therefore, head-to-head systematic comparisons of currently used enzymes manufactured from Roche and Serva are of importance in selecting the ideal enzyme. In this retrospective study, in vitro and in vivo islet assessments data were analyzed to evaluate the efficacy of the 3 enzymes used.
Liberase HI revolutionized human islet isolation from 1994 to 2007.5,12,13,33-35 This enzyme blend was used during Edmonton protocol development to isolate islets successfully for clinical transplantation,13 and in fact has been reproduced by other islet transplant centers globally.1,20,26,36-38 The cessation of liberase HI production in 2007 led to a drop in the number of patients transplanted with islets,39 which resulted in production and use of other enzymes for human islet isolation.6,19,20,40 A recent Collaborative Islet Transplant Registry–based study reported that collagenase NB1/NP was predominantly used in replacement of liberase HI for clinical-grade islet production.41 Brandhorst et al18 reported that collagenase NB1/NP had significantly less digested tissue volume and lower islet yields when compared to liberase HI. Our results confirmed the significant difference regarding digested tissue volume between these 2 enzyme types; however, the purified islet yields showed no difference. Islet isolations using liberase HI had the advantage of having shorter digestion switch time as compared to the collagenase NB1/NP but had high endotoxin levels and variability between enzyme lots. In fact, collagenase NB1/NP made progress to advance the field; however, it has been reported that such enzyme may not be highly purified because trace proteolytic enzymes were present, influencing the isolation outcomes and thus lot-to-lot variability.42
In addition to liberase HI and collagenase NB1/NP enzymes, the use of CIzyme collagenase HA and NP was also reported by the University of Minnesota with successful outcomes in autologous and allogeneic human islet transplantation.21,43 This combination of CIzyme collagenase HA and NP is interesting, thus requires further investigation to substantiate this finding. Moreover, a recent study has shown that supplementation of CIzyme collagenase, Thermolysin with Clostripain improved human islet isolation outcomes.44 The new liberase MTF C/T is a highly purified GMP-grade and animal tissue free product with lower endotoxin levels. Therefore, lot-to-lot variability has been minimum providing very stable enzyme mixtures. Liberase MTF C/T enzyme is supplied as a kit containing 2 separate enzymes: collagenase (C) and thermolysin (T). The collagenase vial is a mixture of collagenase class I (60%) and collagenase class II (40%).17,24,45 The use of liberase MTF C/T was first reported by Caballero-Corbalan et al22 in a multicenter trial of islet isolation using liberase MTF C/T. Twelve successful isolations were reported in this study and 5 of them (42%) reached the transplant criteria of more than or equal to 250,000 IEQ islets.22 In this retrospective analysis, 44% of islet preparations using liberase MTF C/T achieved the same transplant criteria (>250,000 IEQ), which is comparable to aforementioned study.22 Liberase MTF C/T showed best digestion efficacy among the enzymes tested. Furthermore, the lot-to-lot variability was minimal compared to the other 2 enzymes analyzed. There were no significant differences among these enzymes in terms of islet yields prepurification, postpurification, and postculture. It is not surprising that the liberase MTF C/T had the highest mean tissue volume because it showed highest digestion efficacy prepurification. Donor age is one of the major influential factors for islet isolation and transplantation outcome.46,47 A study demonstrated significantly superior outcome with human islets from donors 45 years or younger.48 In this study, the donor age inversely correlated with islet yield prepurification for all 3 types of enzyme blends. Interestingly, the negative association was more significant in collagenase NB1/NP, which may be attributed to the fact that such enzymes seem more effective for younger donors.19 Pancreata from younger donors (<15 years old) would still remain a challenge to isolate free islets despite attempts that were made to obtain highly purified islets from donors with certain age range (6 months to 17 years).49,50 In fact, the pancreata from younger donors with lower body mass index yield less tissue volume and therefore, digested tissues could be transplanted into type 1 diabetic patients without purification. Transplanting tissues from younger donors without purification might have a great advantage because of the potential for islet regeneration, particularly if transplanted into the omentum.51-55
It has been suggested that degradation of collagenase class I may have influenced the enzyme properties,21 albeit this is still debatable. This may be attributed to the fact that the current methods used to evaluate the enzymatic activities such as Wünsch56 and thermolysin/NP units are not consistent throughout the industry.57 Thus, influencing intervariability of the enzyme activity that was used to calculate the final enzyme concentration.56,58,59 The pancreas has been used as a “surrogate substrate” for assessing enzyme activity for determination of suitable lot for pancreas digestion for islet isolation. However, this is expensive and not reproducible due to the variability among pancreata, donor characteristics, and personal experience of isolation team. Therefore, developing a well-standardized method to assess enzyme activity is paramount to further advance the field of islet isolation. In particular, such an assay could be used among industries and/or for evaluation of recombinant enzyme product should they become commercially available.60,61 Thus, standardization of the islet isolation protocol could be achieved so that a Biologics License Application62,63 could be approved by the FDA.64
Recently, an FDA-approved clinical-grade collagenase, Xiaflex (Auxilium Pharmaceuticals, Inc., Chesterbrook, PA), is being clinically used to treat Dupuytren's Contracture and Peyronie disease.65 Xiaflex is composed of class I (50%) and II (50%) collagenase without protease and may be a potential source of FDA approved enzyme for pancreas digestion for islet isolation. Xiaflex has been used clinically based on the protein concentration levels, and the enzyme activity has not been delineated, which may further overemphasize the importance of optimizing enzyme concentration for islet isolation. Interestingly, in this study, calculated enzyme concentration per gram of pancreatic tissue revealed that enzyme protein levels could also be used in addition to or as an alternative to enzyme activities (Table (Table22).
Successful pancreas distention is one of the paramount procedures for successful islet isolation. Tremendous progress has been made in pancreas perfusion with collagenase using well-standardized perfusion apparatus. Hyaluronidase has recently been used for rapid infusion of drugs and solutions in clinical practice.66,67 Therefore, it is tempting to speculate that the use of collagenase/NP supplemented with hyaluronidase for rapid diffusion into the pancreatic tissue, especially into fibrotic tissues and/or organs from donor with chronic pancreatitis, may improve islet isolation outcomes.
Islets isolated with liberase MTF C/T presented significantly higher insulin secretion on glucose stimulation, as compared to the liberase HI and the collagenase NB1/NP. Indeed, the islets from the liberase MTF C/T groups showed superior islet graft function when transplanted into diabetic NOD Scid mice. These tests for islet products indicate that liberase MTF C/T is an efficient enzyme for human islet isolation compared with liberase HI and collagenase NB1/NP.
In conclusion, the data presented in this study indicate that liberase MTF C/T is superior to liberase HI and collagenase NB1/NP in terms of digestion efficacy and GSIS in vitro. Moreover, islets from liberase MTF C/T had a significantly higher success rate of transplantation in diabetic mice.
The authors thank Randall Heyn-Lamb and Karen Ramos for providing and interpreting donor information. The authors also thank Chris Orr, Henry Lin, and Leonard Chen for providing comments to the manuscript. Lastly, the authors extend their thanks and appreciation to all islet isolation team members at City of Hope.
This work was supported by a grant from the National Institute of Health (NIH), the Southern California Islet Cell Resources Center (SC-ICR) 5U42RR016607 and the City of Hope (COH). Human pancreatic islets were provided by the NIDDK-funded Integrated Islet Distribution Program (IIDP) at City of Hope.
The authors declare no conflicts of interest.
M.Q., L.V., and B.M. contributed to study design, data collection and analysis, and article writing and revision. K.O., S.B., J.J., J.R., S.S., K.F., Y.M., M.E., and D.D. contributed to data collection and article revision. F.K. and I.H.A. contributed to study conception and design, data interpretation, and to article writing and revision. All authors approved the final version.