ENDOTOXIN CONTAMINATION MAY BE RESPONSIBLE FOR THE... : Transplantation (original) (raw)
Transplantation of organs or cells has become a highly successful therapeutic procedure. However, transplantation of human pancreatic islets has not yet met with similar success in spite of the improvement of isolation techniques and of repeated attempts. The cause(s) of the frequent failure of islet transplantation remain elusive, but it is puzzling that 59% of the grafts never regain function (1). This situation, known as “primary nonfunction,” is thought to be caused by an early inflammatory response to the graft as a result of strong activation of tissue macrophages (2). The better outcome of whole gland transplantation: 46-76% of insulin independence at 1 year(3, 4), suggests that primary nonfunction of islet grafts results from alterations induced in the tissue during the islet isolation procedure. All the methods currently used to prepare islets, and most other dispersed cell grafts, for transplantation are based on the digestion of the tissue with collagenase (5). In the case of islets, digestion is followed by purification on density gradients. There are indications that a sizeable proportion of the islet cells transplanted do not survive the initial few days, even when the graft is successful(6).
Our group has been isolating human islets with the aim of clinical transplantation since 1989, when we started to use the technique of semiautomatic digestion of the pancreas developed by Ricordi et al.(7). In 1992, we introduced islet purification using the COBE cell separator (8, 9). The disappointing results obtained by others have discouraged us from starting to transplant islets to human recipients.
While we, as most investigators in the field, were trying to elucidate why human islet transplantation has such a low rate of success, we made one observation that gave us a lead towards a possible explanation for the occurrence of primary nonfunction in islets grafts.
This observation is that human islet cells express CD14, the receptor for lipopolysaccharide (LPS*) (10) (N. Somoza, J. Fernandez, F. Vargas, M. Vives-Pi, R. Gomis, M.O. Labeta, R. Pujol-Borrell, manuscript in preparation) and related bacterial endotoxins(11, 12).
This observation led us to consider those situations in which islets could be exposed to endotoxin. One situation was during collagenase digestion of the pancreas if the collagenase was contaminated. Commercial collagenases are, in fact, a complex mixture of proteolytic enzymes obtained from_Clostridium histolyticum_ cultures, and it seemed plausible that they could have endotoxin activity (13). In fact, measurement of endotoxin activity in collagenase gave such remarkably high levels of activity that we undertook the present work aiming at: (1) determining the extent of endotoxin contamination of collagenase preparations currently in use to prepare human islets for transplantation; (2) assessing the proinflammatory capacity of these collagenase solutions; and (3) elucidating whether proinflammatory cytokines were induced in the islet tissue to be transplanted during the islet preparation procedure.
MATERIALS AND METHODS
Isolation and purification of human islets. Thirteen human pancreases obtained from cadaveric organ donors were used for this study. The mean age of the donors was 39 years (range, 15-63 years), and the male/female ratio was 9/4. The amount of tissue available for islet separation ranged from 46 to 109 g (mean, 76.5 g). All protocols described in this report were approved by the ethical committee of the “Germans Trias i Pujol” University Hospital.
Human pancreas were digested according to a slight modification of Ricordi's semi-automated method (7, 14), which is similar to the method used in most laboratories active in human islet transplantation. Briefly, the pancreatic duct was cannulated and injected with the collagenase solution (Type P, Boehringer-Mannheim Biochemica, Mannheim, Germany), in Hanks' balanced salt solution (Gibco BRL, Paisley, Scotland) containing Hepes (15 mM; Sigma Chemical, St. Louis, MO) and supplemented with 1% newborn calf serum (Gibco) and 0.04 mg/ml deoxyribonuclease I (DNase I, Sigma) at a ratio of 2 ml/g of tissue. The gland was then loaded into the digestion chamber. After digestion, the dispersed tissue was washed three times in a total volume of 6 L of Hanks' solution. Islets were purified by centrifugation through a continuous density gradient in a COBE cell separator using a Pielograf/Lymphoprep mixture (metrizamide/Ficoll, Pielograf from Juste SAQF, Madrid, Spain; Lymphoprep from Nycomed Pharma AS, Oslo, Norway) as described (7, 9). Free islets were washed at least three times in Hanks' solution and maintained in culture for up to 3 hr; at this point, they were considered ready for transplantation(15). Microbiological controls always gave negative results. Culture medium was RPMI 1640 (Gibco) supplemented with 10% endotoxin-free fetal calf serum (Myoclone, Gibco), 40 mg/ml gentamycin(Normon, Madrid, Spain), and 40 mg/ml penicillin (Roger, Barcelona, Spain). Cultures were kept at 37°C, in a 5% CO2/air humidified incubator.
Samples of pancreatic tissue, islet preparations at different stages of purification, transportation fluids, and media and solutions used for islet separation were collected and kept at -70°C until assayed.
Endotoxin detection assay. To assess endotoxin levels in the different samples, a commercial quantitative chromogenic method was used(Limulus amebocyte lysate Coatest, Endosafe Inc., Charleston, SC). The Limulus amebocyte lysate contains an enzymatic system that is activated in the presence of endotoxin. The lower limit of detection of this assay is 0.06 EU/ml; the C.V. is 1.55.
Separation and culture of purified monocyte-derived macrophages. Human macrophages were obtained from buffy coat preparations (n=2) provided by the hospital blood bank and one from freshly drawn heparinized blood from healthy laboratory staff by separation in a density gradient (Lymphoprep, Nycomed Pharma, Oslo, Norway) and purification by 1 hr of adherence to plastic coated with 2% gelatine (Sigma). After 3 days, 106 cells were stimulated for 1 to 8 hr with samples from transportation and islet isolation fluids, collagenase digestion solutions, and LPS (Escherichia coli, serotype 0111:B4, Sigma).
Measurement of cytokines. Tumor necrosis factor-α(TNF-α) was measured by enzyme-linked immunosorbent assay (ELISA) using a commercial kit (Biotrak, Amersham International, Buckinghamshire, UK), which has a lower limit of sensitivity of 25 pg/ml.
To assess the presence of cytokine transcripts, the reverse transcription-polymerase chain reaction (RT-PCR) was used essentially as described (16). Positive controls were peripheral blood mononuclear cells (2×106) stimulated with LPS (100 ng/ml). RNA was extracted according to the technique of Chomczynski and Sacchi(17). cDNA was prepared by standard methods. All reagents were obtained from Promega (Madison, WI). Polymerase chain reaction(PCR) was performed according to manufacturers instructions using the following primers: for interleukin (IL)-1α (sense, GAGAGCATGGTGGTAGTAGCAACC; antisense, CCCTGCCAAGCACACCCAGTAGTC), for IL-1β(sense, CTTCATCTTTGAAGAAGAACCTATCTTCTT; antisense, AATTTTTGGGATCTACACTCTCCAGCTCTA), for IL-6 (sense, ATGAACTCCTTCTCCACAAGCGC; antisense, GAAGAGCCCTCAGGCTGGACTG), for TNF-α (sense, GACGTGGAGCTGGCCGAG; antisense, CACCAGCTGGTTATCTCTCAGCTC), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (sense, TCTTCTTTTGCGTCGCCAG; antisense, GGGGGCAGAGATGATGACC). Primers were designed in our laboratory to span at least one intron and supplied by Oligo Etc. (Wilsonville, OR) or Genset (Paris, France). For Southern blotting, the following oligoprobes were used: IL-1α, TACGAATTCATCCTGAATGAC; IL-1β, AAAAAAGCTTGGTGATGTCTGG; IL-6, CTCATTGAATCCAGATTGGAA; and TNF-α, AGCCTCTTCTCCTTCCTGATCGTG.
Normalized RT-PCR was performed by adjusting the amount of each cDNA according to that of a constitutive gene (GAPDH) and by working within the exponential phase of the amplification reaction as established in preliminary experiments. One sample containing all reagents but cDNA was included in every run. For Southern blotting, gels were transferred to Hybond™ -N nylon membranes (Amersham International), hybridized, and autoradiographed.
RESULTS
Collagenase used to prepare islets for transplantation has high endotoxin activity. Samples of collagenase of the type used for islet isolation and from the manufacturers that normally supply this enzyme to groups currently transplanting islets, were assayed. Six batches of collagenase from manufacturer 1 (which currently supplies our group) were tested and gave an endotoxin activity that ranged between 28,500 and 360 EU/mg. The two batches of collagenase from manufacturer 2 (less widely used at the moment) had a lower activity: both in the range of 100 EU/mg. One batch from manufacturer 3 had 500 EU/mg. This wide variability of endotoxin activity did not bear relation to the intrinsic activity of the collagenase. To exclude other sources of contamination by endotoxin during the process of islet separation, we measured endotoxin activity in all the fluids and reagents. The results, summarized in Table 1, demonstrate that the only important source of endotoxin is collagenase, although both Ficoll and serum had some endotoxin activity. A new, purified preparation of collagenase from manufacturer 1 was also tested and found to have a low level of endotoxin (2.3 EU/ml).
To determine the degree of exposure of the tissue to endotoxin during the islet isolation, endotoxin activity was monitored along the whole procedure. Endotoxin was measured in the fluid in which the organs were transported to the laboratory (EuroCollins) and in the solution in which the islets were maintained during the procedure (Hanks' balanced salt solution/1% newborn calf serum, v/v), both immediately after digestion and at the moment the islets were considered ready for transplantation. Samples from islet isolations from 13 different glands, carried out using three lots of collagenase, which contained different activities of endotoxin, were assayed. Results are summarized in Table 2. Digestion resulted in a heavy contamination of the fluids and presumably of the islets by endotoxin. The repeated washes by centrifugation reduced markedly endotoxin activity, but levels remained well above the initial ones; in 10 out of 13 separations, they remained above those permitted for human administration(18).
Collagenase and fluids used in the preparation of the islets induce inflammatory cytokines in macrophages. Previous time course experiments exposing macrophages to either collagenase or LPS (positive control) had established that IL-1α, IL-1β, TNF, and IL-6 messages are all strongly induced by collagenase at 4 hr as assessed by RT-PCR. TNF-α measured as protein by ELISA reached high values at 81 hr.
To determine whether the endotoxin activity detected in collagenase of different batches was actually capable of inducing the activation of macrophages, standard preparations of monocyte-derived macrophages were exposed to collagenase for 4 hr and TNF-α production was measured in the supernatants. Table 3 summarizes the results of three experiments using two different batches of collagenase and two types of macrophage preparations. Collagenase solution at the working dilution used to isolate islets induced the release of TNF-α in the order of 1000 pg/106 cells/4 hr, which is similar to the magnitude induced by the standard stimulatory doses of LPS (100 ng/ml). Interestingly, when macrophages were derived from buffy coat preparations, collagenase was a much more potent TNF inducer that LPS, whereas, when derived from freshly drawn blood, the levels of induction were similar.
To determine whether the solutions actually used and generated during the separation of the islets could induce inflammatory cytokines, monocyte-derived macrophages were exposed for 4 hr to these media at 1/2 to 1/5 dilution. At these dilutions, the digestion and the initial washing solutions both induced strongly the appearance of TNF-α, IL-1α, IL-1β, and IL-6 transcripts, whereas the pretransplantation fluid still produced a moderate induction of TNF-α (Fig. 1a), thus indicating that even if the multiple washing steps remove most of proinflammatory cytokine inducing activity, some activity remains detectable (at 1/5 dilution). TNF-α was measured by ELISA on the supernatant of the same cells. Maximal release (950 pg of TNF-α/106 cells) was observed when the macrophages were incubated with the collagenase, but TNF-α (15 pg of TNF-α/106 cells) was also produced after exposure to the fluid from washing the islets before transplantation (Fig. 1b).
Induction of inflammatory cytokines in the islets during the isolation procedure. To determine whether, during the digestion of the pancreatic gland, there was induction of inflammatory cytokines in the islet preparations, transcripts of TNF-α, IL-1α, IL-1β, and IL-6 were measured by RT-PCR in samples of the original pancreatic tissue and in islet preparations, both as they came out from the gradient and after culturing for 90 min (two experiments) or 180 min (two experiment). Islet preparations of different purities were cultured in parallel. The results, depicted in Figure 2a, demonstrate a strong induction of cytokine message during the isolation. Specificity of the amplicons was confirmed by Southern blotting with the corresponding oligo probes (data not shown). The results of RT-PCR were confirmed by measuring by ELISA TNF-α in the supernatants of the islets (see Fig. 2b). Interestingly, cytokine induction correlated inversely with the degree of islet purity; as an example, see Figure 2 and compare PCR bands and ELISA results in lane 90′ left (40% pure islets) with those in lane 90′ right (80% pure islets). In another experiment, the amount of TNF secreted at 3 hr was much higher and again correlated inversely with purity.
DISCUSSION
The results here presented confirmed high levels of endotoxin activity in collagenase preparations used to prepare islets for transplantation into human recipients. This activity was in part incorporated into the tissue to be grafted and, more disturbingly, induced the expression of proinflammatory cytokine in the tissue itself. We think that this observation may help to understand and overcome the difficulties met by the transplantation of pancreatic islets, and it is also relevant to attempts of transplanting other collagenase dispersed cell tissue preparations. In addition, and because collagenase is a reagent of general use in many experimental fields of biomedicine, awareness of its endotoxin activity may help to re-interpret several experiments and to improve protocols, e.g., preparation of thymic stromal cells (19).
The endotoxin activity present in collagenase formulations is high and equivalent to highly toxic concentrations of LPS (20). Many components of gram-negative and -positive bacteria have endotoxin activity, and it is not known at the moment which component(s) of_Clostridium_ cell culture supernatant is responsible for the endotoxin activity of collagenase. We had already reported that it is very efficient in eliciting an inflammatory cytokine response in monocyte-derived macrophages (21), and our present results demonstrate that pancreatic tissue also produces cytokines after exposure to collagenase. We have not tried to establish the precise type of endotoxin through neutralization experiments because manufactures have already produced low endotoxin collagenase, so they probably have this knowledge.
In our view, the obvious implication of the above results for islet cell transplantation is that primary nonfunction, the most common cause of islet graft failure, may be a consequence of the use of collagenase containing endotoxin to obtain the islets. Two synergistic mechanisms may be operating.(i) As some endotoxin is probably carried over with the graft into the host tissue, this would induce proinflammatory cytokine production by host macrophages and would also activate the endothelium (22) enhancing the migration of monocytes, polymorphonuclear cells, and lymphocytes into the graft. This would contribute not only to expanding the local inflammatory response but would also favor the subsequent development of rejection. (ii) Because proinflammatory cytokines are also induced in the tissue to be grafted while islets are cultured before being transplanted, it is highly likely that by the time they are implanted, their levels of HLA and adhesion molecules have increased significantly. On the other hand, these cytokines present in the graft would also act on the host macrophages and endothelial cells activating them and contributing to the induction of local inflammation with the same consequences predicted in (i). We should not forget that the graft, in addition to the above, inevitably contains some dead cells, exocrine tissue loaded with proteolytic enzymes, and that the process of islet separation itself may induce the expression of some stress proteins(23). In fact, considering all of these factors, it would not be totally inappropriate to describe the islet preparations currently being transplanted as a “graft with adjuvant.”
One question is the cellular origin of the transcripts of cytokines detected in the islet preparations. It should be remembered that these preparations contain at least 10% of exocrine tissue, capillary endothelial cells, tissue macrophages, dendritic-like cells, fibroblasts, and other minor cell populations. Macrophages and, as we have reported, islet cells express the endotoxin receptor CD14 (10). Furthermore, endotoxin can activate endothelial cells directly through an alternative LPS receptor(24), and it seems that most cells can respond to high concentrations of endotoxin (25). It is therefore likely that multiple cellular sources contribute to the detected cytokine transcripts. The lower production of TNF-α by the purest islet preparations suggests that the exocrine tissue may produce more cytokines than islets themselves.
It may seem puzzling that the cytokine-inducing effects of collagenase had not being scrutinized before in detail. Several factors have probably contributed. On the one hand, collagenase has been in use in experimental transplantation in rats (a specie relatively resistant to endotoxin) for so long as to become an “accepted” reagent. On the other hand, it is true that the transplantologists have indeed always been highly critical of collagenase as a reagent. In addition, collagenase itself was never been administered to the patients and the protocols used for washing it away from the graft are, as we have shown, reasonably efficient. It was not predicted that the exposure of the future graft to collagenase was sufficient to trigger the production of proinflammatory cytokines.
Most importantly, endotoxin can be removed from collagenase. We believe that the use of low endotoxin collagenase may improve the outcome of islet grafts reducing both the incidence of primary nonfunction and of rejection. It is fortunate that a new generation of collagenases has been developed(26) and that they seem to have a much lower endotoxin content. Work is in progress in our laboratory to demonstrate that the endotoxin in collagenase contributes to the induction of rejection in experimental islet transplantation and also that the use of the new collagenases may substantially improve the success rate of islet transplantation.
Acknowledgments. The authors thank the Departments of Surgery of the Hospital Vall d'Hebró, Hospital Clinic, and Hospital Germans Trias i Pujol for the co-operation in the procurement of the tissue. The authors particularly thank the rest of the islet isolation team (Dr. R. Gomis, Dra. J. Fernández, E. Corominola, Dra. R. Gasa, C. Benito, C. Franco, M. Español, and M. Gudayol) for their help. The authors also thank Dr. E. Espel for help in macrophage obtention and culture, the Microbiology Unit of Germans Trias i Pujol University hospital for performing control cultures, and Dr. Ian Todd (Nottingham, UK) for kindly reviewing the manuscript.
Collagenase and collagenase-contaminated solutions but no other fluids used in islet isolation are capable of inducing inflammatory cytokines in macrophages. (a) a RT-PCR representative fourth experiment of IL-1α, IL-1β, IL-6, and TNF-α induction by different solutions. The amount of mRNA was normalized according to the levels of GAPDH. Lane φ, phageφ 174 (molecular weight marker); lane B, macrophages stimulated with control tissue culture medium; lane 1, macrophages stimulated with collagenase solution (2.5 mg/ml batch -37); lane 2, marcophages stimulated with LPS (100 ng/ml); lane 3, macrophages stimulated with pancreatic transportation medium(0.14 EU/ml) diluted 1/5; lane 4, macrophages stimulated with pretransplantation medium (0.18 EU/ml) diluted 1/5; lane 5, macrophages stimulated with pancreatic pretransplantation medium (9.5 EU/ml) diluted 1/5; lane C, PCR-negative control (reagents without cDNA). (b) TNF measured by ELISA correlates approximately with message levels. Mφ, macrophages.
Cytokines are induced in the pancreatic tissue during the islet isolation process. (a) RT-PCR experiments on the sustrates of two islet isolation procedures: pancreas 212 (PB 212) and pancreas 214 (PB 214). The induction of cytokine message for IL-1α, IL-6, and TNF-α was assessed. Lane φ, phage φ 174 (molecular weight marker); lane T, unprocessed pancreatic tissue; lane F, freshly isolated islets; lane 90′, islets cultured for 90 min (left, 40% pure; right, 80% pure); lane 180′, islets cultured for 180 min (left, 30% pure; right, 50% pure); lane (-), PCR-negative control (reagents without cDNA); lane (+), PCR-positive control (macrophages stimulated with 100 ng/ml LPS). (b) TNF-α was detected on the supernatant of islets cultured for 90 min (PB 212) and 180 min(PB 214), which confirms RT-PCR results. EIN, islets equivalent number.
*Abbreviations: ELISA, enzyme-linked immunosorbent assay; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IL, interleukin; LPS, lipopolysaccharide; PCR, polymerase chain reaction; RT-PCR, reverse transcription-polymerase chain reaction; TNF-α, tumor necrosis factor-α.
REFERENCES
1. Anonymous. Results on all adult islet allografts through Dec 31, 1994. Int Islet Transplant Registry Newslett 1995; 5: 10.
2. 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.
3. Sutherland DER. State of the art in pancreas transplantation. Transplant Proc 1994; 26: 316.
4. Remuzzi G, Ruggenenti P, Mauer SM. Pancreas and kidney/pancreas transplants: experimental medicine or real improvement? Lancet 1994; 343: 27.
5. Gray DW, McShane P, Grant A, Morris PJ. A method for isolation of islets of Langerhans from the human pancreas. Diabetes 1984; 33: 1055.
6. Scopsi L, Andreola S, Socci C, et al. Immunocytochemical detection and characterization of intrahepatic human pancreatic islets after combined liver-islet allotransplantation. Cell Transplant 1994; 3: 449.
7. Vives M, Sarri Y, Conget I, et al. Human islet function after automatic isolation and bovine serum albumin gradient purification. Transplantation 1992; 53: 243.
8. Lake SP, Bassett D, Larkins A, et al. Large-scale purification of human islets utilizing discontinuous albumin gradient on IBM 2991 cell separator. Diabetes 1989; 38(suppl 1): 143.
9. Vargas F, Vives-Pi M, Somoza N, et al. Advantages in using the COBE cell separator and metrizamide gradients for human islet purification. Transplantation 1996; 61: 1562.
10. Somoza N, Vargas F, Pujol-Borrell R. Expression of CD14 in human beta cells and human insular cell lines. Diabetologia 1994; 37 (suppl 1): 374.
11. Espevik T, Otterlei M, Skjak-Braek G, Ryan L, Wright SD, Sundan A. The involvement of CD14 in stimulation of cytokine production by uronic acid polymers. Eur J Immunol 1993; 23: 255.
12. Kusunoki T, Hailman E, Juan TS, Lichenstein HS, Wright SD. Molecules from Staphylococcus aureus that bind CD14 and stimulate innate immune responses. J Exp Med 1995; 182: 1673.
13. Baysto K, Tomlinson M, Cohen J. In-vitro stimulation of TNF-a from human whole blood by cell-free supernatants of gram-positive bacteria. Cytokine 1992; 4: 397.
14. Ricordi C, Lacy PE, Finke EH, Olack BJ, Scharp DW. Automated method for isolation of human pancreatic islets. Diabetes 1988; 37: 413.
15. Ricordi C, Tzakis A, Carroll P, et al. Human islet isolation and allotransplantation in 22 consecutive cases. Transplantation 1992; 53: 407.
16. Somoza N, Vargas F, Roura-Mir C, et al. Pancreas in recent onset insulin-dependent diabetes mellitus. J Immunol 1994; 153: 1360.
17. Chomczynski P, Sacchi N. Single step of RNA isolation by guanidinium thiocianate-phenol-chloroform extraction. Anal Biochem 1987; 162: 156.
18. Center for Drug Evaluation and Research. Guideline on validation of the Limulus amebocyte lysate test as an end-product endotoxin test for human and animal parenteral drugs, biological products, and medical devices. Bethesda, MD: U.S. Department of Health and Human Services Public Health Service, Food and Drug Administration, 1987: 2.
19. Eshel I, Savion N, Shoham J. Analysis of thymic stromal cell subpopulations grown in vitro on extracellular matrix in defined medium. J Immunol 1990; 144: 1554.
20. Kuhns DB, Alvork WG, Gallin JI. Increased circulating cytokines, cytokine antagonists, and E-selectin after intravenous administration of endotoxin in humans. J Infect Dis 1995; 171: 145.
21. Vargas F, Vives-Pi M, Somoza N, Fernández-Llamazares J, Pujol-Borrell R. Endotoxin activity of collagenase in human islet transplantation. Lancet 1997; 350: 641.
22. Pugin J, Schurer-Maly C-C. Lipopolysaccharide activation of human endothelial and epithelial cells is mediated by lipopolysaccharide-binding protein and soluble CD14. Proc Natl Acad Sci USA 1993; 90: 2744.
23. Welch WJ. Mammalian stress response: cell physiology, structure/function of stress proteins, and implications for medicine and disease. Physiol Rev 1992; 72: 1063.
24. Bevilacqua MP, Pober JS, Mendrick DL, Cot RS, Gimbrone MA. Identification of an inductor of endothelial leukocyte adhesion molecule. Proc Natl Acad Sci USA 1987; 84: 9238.
25. Lynn W, Raetz C, Qureshi N, Golenbock D. Lipopolysaccharide-induced stimulation of CD11b/CD18 expression on neutrophils. J Immunol 1991; 147: 3072.
26. 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.
© Williams & Wilkins 1998. All Rights Reserved.