Abstract
The mechanism of the two-layer method (TLM) of pancreas preservation is unclear. Facilitating oxygen diffusion into preserved pancreas has been suggested, but direct measurements of tissue pO2 have yielded conflicting results. The degree of penetration of perfluorocarbon (PFC) into the pancreas during TLM storage is unknown. Segments of porcine pancreas (7.5 cm in length) were preserved either in University of Wisconsin solution (UW) alone (n = 6) or in TLM for 24 h (n = 6). Pancreatic samples were analyzed using Varian INOVA 9.4T MR scanner. External PFC standard was introduced for quantification. Four consecutive transverse images of 4 mm thickness were obtained using a spin-echo sequence. 19Fluorine magnetic resonance spectroscopy (19F MRS) was performed with the same parameters except with more averages. MR data were confirmed by headspace chromatography. PFC standard was readily detected in 19F MR images. There was no signal from pancreas in 19F MR images following either UW or TLM storage. 19F MR spectra typical of PFC were not obtained from either UW- or TLM-preserved pancreas with nonlocalized 19F MRS. Mean concentration of PFC in TLM pancreas measured by head space chromatography was 0.011 nl/g (SD ±0.006), not significantly different from background concentration (0.012 nl/g, SD ±0.006) in UW pancreas (p = 0.42). There was no evidence of penetration of PFC into pancreas tissues investigated either by MR or chromatography in organs preserved at hypothermia by TLM, and mechanisms of TLM remain speculative.
Keywords
Introduction
Perfluorocarbons (PFCs) are chemically and biologically inert liquids that have the exceptional ability to physically dissolve and release significant quantities of respiratory gases (28). They have been used as oxygen carriers to improve oxygen transport and oxygen delivery to tissues (17,28). PFCs were first used for organ preservation as a component of the two-layer method (TLM) of pancreas preservation, developed by Kuroda et al. in 1988 (15). The first clinical application of the TLM of preservation for islet transplantation was reported in 2002 (37). Since then several centers have reported improved islet isolation outcomes with this preservation technique (22,36). Reports of success with single donor islet transplants and marginal donors have also been attributed, either partially or wholly, to the employment of the TLM (9,10,29). However, despite two decades of experimental and clinical application the mechanism of the TLM remains unclear. Improved oxygenation of the pancreas is considered to be the most plausible explanation for the beneficial effects of TLM (18). Oxygen delivery could continuously generate adenosine triphosphate (ATP) within the graft, thereby maintaining the homeostatic ion balances essential for cell membrane integrity (6). In this hypothesis the presumption is that PFC infiltrates the pancreas to facilitate oxygen delivery to the graft because simple gas diffusion from an aqueous solution into a large tissue mass (such as a human pancreas) will be severely restricted. However, direct evidence of improved graft oxygenation is not available and a recent meta-analysis by our group raises questions regarding the clinical efficacy of the TLM (1).
To our knowledge, there are only two reports of direct measurements of oxygen partial pressure in solid pancreases preserved by TLM. Matsumoto et al. used Clark oxygen electrodes to measure tissue oxygenation in the core of canine pancreas preserved in TLM (21). They observed that even under warm ischemic conditions, when preserved by TLM at 20°C oxygenation at the core of the pancreas improved rapidly, reached a level of 15–17 mmHg within 30 min, and was maintained at this level. In the control arm where pancreas was preserved in oxygenated University of Wisconsin (UW) solution at 20°C, tissue pO2 and ATP levels continued to decrease progressively with time. They concluded that TLM could restore the viability of pancreas damaged by warm ischemia. This observation was challenged in a more recent study by Papas et al. (26). They used a homogeneous diffusion reaction model for theoretical calculations of pO2 profiles of a cylindrical pancreas and found that the predicted depth of oxygen penetration was just 1 mm under cold storage conditions and that the oxygenated volume fraction dropped further as the temperature was raised. They confirmed these results by experimental measurements of pO2 with fiber optic sensors placed in the core of 1-cm-thick piece of porcine pancreas. Within 12 min of insertion of the probe, the pO2 in the pancreatic core dropped to zero.
The findings of these two studies therefore conflict. Whereas in the first study the authors speculate that PFC might permeate the organ to facilitate core oxygenation, the second study concludes that primate pancreas remains largely anoxic during TLM preservation, suggesting that the PFC does not penetrate the organ. This is further confounded by the fact that there may well be different outcomes when studying pancreases from small animal species and those from large organs such as human and porcine pancreases. In human islet isolations, Kin et al. reported no significant increase in ATP content of the pancreases that were preserved by TLM, calling into question the validity of the hypothesis that the TLM efficiently delivers oxygen to the pancreas (16). More recently another retrospective analysis of 200 human islet isolations arrived at the same conclusion (3).
Direct diffusion of oxygen in a large pancreas is an extremely slow process. Improved oxygen transfer with PFC would suggest that the PFC itself penetrates the organ (21). The hypothesis that PFC penetrates the pancreas in order to facilitate oxygenation is based on previous observations in pediatric patients with adult respiratory distress syndrome. Ventilation with PFC is associated with absorption across the alveolus and subsequent elimination by evaporation (7,21). To date measurement of PFC uptake in pancreas preserved by TLM has not been attempted. The purpose of this study was to quantify and examine the pattern of distribution of PFC within adult porcine pancreas preserved by TLM.
Materials and Methods
The procedures in this protocol were carried out in accordance with the Animal (Scientific Procedures) Act 1986 UK. Adult porcine pancreases were harvested from breeder pigs and the procurement in general followed the standard techniques for human multiorgan retrieval. Briefly, following terminal anesthesia a midline laparotomy was made and the aorta was cannulated immediately. The abdominal aorta was cannulated and retrograde flushing commenced with 3 L of chilled Marshall's solution (Soltran, Baxter Healthcare, Norfolk, UK) after clamping the supracoeliac aorta, with effluent solution vented from the adjacent vena cava. The gastrocolic ligament was divided to expose the left pancreatic segment. The left segment was then mobilized and removed by dividing the pancreas along a line in front of the aorta downwards from the coeliac axis to the splenic–superior mesenteric vein confluence. The pancreases were preserved in ice by one of two methods described below.
Group 1 pancreases were preserved by the two-layer method for 24 h in a 1.5-L cylindrical jar and maintained at the interface of the bottom layer of PFC (300 ml) and layer of UW solution (300 ml). Immediately following harvest the pancreases were placed at the interface of PFC (precharged with oxygen immediately prior to organ harvesting at a flow rate of 1.5 L/min for 30 min at room temperature) and an equal volume of UW solution (300 ml) in a 1.5l-Nalgene jar. Each pancreas was kept partially immersed in PFC (about 50% of full thickness) with the help of an adjustable fenestrated disc screwed to the lid of the jar. This assembly was in an ice box for 24 h. Perfluorodecalin (C10F18; F2 Chemicals, Preston, UK), the typical compound employed in organ preservation, was the PFC used for all the studies and was precharged with oxygen immediately prior to organ harvesting at a flow rate of 1.5 L/min for 30 min at room temperature. Group 2 control arm pancreases were preserved in a similar fashion but using 300 ml chilled UW solution (Du Pont Pharma, Bad Homburg, Germany) for 24 h. The jars were kept in an ice box during the storage period. Following storage, the surface of the pancreas was thoroughly irrigated with phosphate-buffered saline immediately before the assays were made.
Proton and Fluorine Magnetic Resonance Imaging (1H and 19F MRI)
Preliminary assessment of PFC penetration was made by noninvasive fluorine magnetic resonance imaging with a 9.4 T Varian INOVA MR scanner (Varian Inc., CA, USA). For this a segment of pancreas (n = 4 each in groups 1 and 2), 7.5 cm in length, was divided from the tail end of each pancreas, washed, and placed in a 50 ml universal tube containing UW solution. This was then centrally located in a quadrature volume coil capable of being tuned to both proton (1H) and fluorine (19F) resonance frequencies. A phantom (1 ml syringe filled with perfluorodecalin) was taped to the universal tube containing the pancreatic segment as an external reference standard. Initially, proton imaging was performed to establish appropriate sample location using a spin-echo sequence with a repetition time (TR) of 1 s, echo time (TE) of 20 ms, field of view (FOV) of 45 × 45 mm, 256 × 128 matrix size, and 4 averages. Four consecutive transverse images 4 mm in thickness were obtained. 19F MRI images were subsequently acquired with the same parameters except 512 averages were used.
Magnetic Resonance Spectroscopy (19F MRS)
Nonlocalized 19F MRS was then performed to quantify PFC uptake in the pancreas using the same scanner (9.4 T). For this, 16 K time data points were collected from a sweep width (SW) of 50 kHz using a 90° pulse, repetition time (TR) of 1 s, and 512 averages were collected. Syringes (1 ml) of perfluorodecalin and sodium fluoride were employed as external reference for 19F MRS quantification purposes. Four pancreas harvests were carried out in each group: in one group (TLM) the organs was preserved by the two-layer method for 24 h, and in the other (UW alone) standard preservation in UW was employed for the same duration (n = 4 in each group).
Headspace Gas Chromatography
Headspace gas chromatography was used as an alternative method to quantify PFC uptake in the pancreas. Six porcine pancreases preserved in TLM for 24 h were compared with an equal number of porcine pancreases preserved for the same duration in UW solution (n = 6 in each group). In addition, fresh pancreatic tissue was spiked with perfluorodecalin to serve as a positive control for this experiment: 0.1 μl of perfluorodecalin was added to 500 mg of pancreatic tissue.
Pancreatic tissue (500 mg) was transferred to 50-ml pyrex head space vials containing 2 g of anhydrous magnesium sulphate. The bottles were placed in a microwave oven and heated for 90 s at 200°C to volatilize the PFC. The sample effluent from the head space was analyzed in a gas chromatograph (model 6890; Agilent Technologies, Inc. Wilmington, DE, USA) configured with a flame ionization detector (GC-FID) and a stationary phase 20% w/w SE-30 on solid support Gas Chrom R Mesh 100–120 packed glass column (12 ft x 6 mm OD x 3 mm ID). The carrier gas was helium flowing at a constant rate of 24 ml/min. Beadspace effluent (1 ml) was introduced into the injector port for analysis using a gas-tight syringe. The initial temperature was 100°C for 12 min followed by a ramp rate of 4°C/min for 18 min to a final temperature of 172°C held for 5 min.
Statistical Analysis
Statistical significance for headspace gas chromatography was assessed using Student's t-test to compare the means between the two groups. A value of p < 0.05 was considered as significant.
Results
Proton and Fluorine Magnetic Resonance Imaging
The results of MR imaging are shown for a typical pancreas preserved by TLM in Figures 1 and 2. Proton and fluorine MR imaging was carried out in all pancreatic segments (n = 4 in each group) as a preliminary to spectroscopic evaluation. With proton imaging, the pancreatic segment was clearly imaged but the PFC syringe was not visualized. The PFC syringe was readily detected with fluorine imaging but the pancreatic segment was not visible. The observations in both UW- and TLM- preserved pancreas were the same.
Corresponding 1H (a) and 19F (b) transverse MR images of a typical porcine pancreas stored in TLM for 24 h. Although the experimental phantom is clearly visualized, there is no fluorine signal from the adjacent pancreas, indicating that PFC does not infiltrate into the parenchyma (b). This was consistent in all pancreases from both UW and TLM groups.
Illustration of Figure 1 images diagrammatically. The universal tube has a diameter of 30 mm.
19F Magnetic Resonance Spectroscopy
Due to the lack of fluorine signal from imaging of the pancreas, nonlocalized 19F MRS was performed to estimate the PFC level in the whole pancreatic sample. Nonlocalized spectroscopy has greater sensitivity than 19F MRI to measure PFC because the former allows the detection of any PFC throughout whole sample. Conversely, 19F MRI will only detect PFC if above the limit of detection in a defined volume, which in this instance is 25 mm3. Syringes containing 1 ml of sodium fluoride and perfluorodecalin were used as positive controls. Figure 2 illustrates the typical 19F -MRS spectra obtained from these experiments. Qualitatively there was no difference between the spectra obtained from TLM-preserved pancreases (n = 4) and those in UW solution (n = 4), and the characteristic 19F MRS spectrum of PFC was not seen in either (Fig. 3). Nonlocalized 19F MRS was unable to detect any PFC in the pancreas samples and so 19F MRI investigations were not pursued further.
19F MR spectra obtained from pancreas preserved by simple cold storage (pancreas 3) and by two-layer method (pancreas 1 and 2) compared with representative spectra of PFC and sodium fluoride. Both TLM- and UW-preserved pancreases have identical MR spectrum and no PFC signal is evident.
Method Validation for Gas Chromatography: Linearity and Precision
The method linearity and precision for head space gas chromatography were studied. Linearity data were obtained from serial diluted standards of perfluorodecalin. Linear regression analysis was performed over the range of 0.02% to 0.38% with five concentration levels (0.02%, 0.04%, 0.10%, 0.19%, and 0.38% w/w equivalent) and the linear equation and correlation coefficient obtained. A correlation coefficient (R) of more than 0.99 indicates that the proposed method has a good linearity. The correlation coefficient of the curve for our method was 0.9981 and the factor of curvature 0.9846 (Fig. 4). In addition to this, six separate analyses of the 0.10% w/w standard produced a residual standard deviation (RSD) of 4.5%, indicating good precision.
Standard curve (linear regression graph) of the response obtained from analysis of five different concentration levels of diluted standard for method validation, indicating good method linearity and precision (see text).
Headspace Gas Chromatography
Figure 5 shows the typical GC-FID chromatograms of pancreas spiked with perfluorodecalin (first panel), pancreas preserved for 24 h in the TLM (second panel), and pancreas preserved in UW solution for 24 h (third panel). The mean concentration of perfluorodecalin, as identified by the retention times of the PFD isomers, in the pancreatic tissue samples preserved in TLM was 0.011 ± 0.006 nl/g. This was not significantly different from the concentration of PFC in the control samples preserved in UW solution (0.012 ± 0.42 nl/g, p = 0.42). The concentration of the spiked standard (0.1 μl perfluorodecalin added to 500 mg of fresh pancreas) was 166.07 nl/g.
Headspace chromatograms of PFC from fresh pancreas spiked with perfluorodecalin (positive control, panel 1), from pancreas preserved by TLM for 24 h and washed with phosphate-buffered saline (test sample, panel 2), and from pancreas preserved in UW solution alone (control, panel 3). PFC levels are minimal in both TLM- and UW-preserved pancreas, suggesting lack of PFC penetration in pancreatic tissue.
Discussion
In this study pancreatic tissue concentration of PFC was measured both by magnetic resonance spectroscopy and headspace gas chromatography. No evidence for the uptake of PFC into pancreas during cold preservation by TLM was seen.
Compared to the widely used rat model, the pig pancreas is a more suitable model for preclinical preservation studies because of anatomical and physiological similarities with humans. Also the parenchymal tissue density is similar to that of the human gland (32,35). Moreover, the procurement and preservation procedures in pigs can be fully adapted from the clinical situation. The pig pancreas has therefore been used as a large-animal model for pancreas (35) and islet transplantation (13,25) in previous studies.
19F NMR (nuclear magnetic resonance) is a specific and accurate method for measurement of tissue PFC concentration. It has been used to quantitate the biodistribution of PFC blood substitutes in porcine tissues following intraperitoneal administration of a PFC emulsion (27). Recently, PFC content of targeted nanoparticles has been exploited for the purposes of 19F imaging and spectroscopy to accurately quantify fibrin content in atherosclerotic plaques (14,24). Head space gas chromatography is a well-established method for detection of PFC in biologic specimens (2,5,11). It has been employed for detection of trace amounts of PFC in blood and tissues of premature lambs during liquid ventilation experiments (30). The authors concluded that following uptake of trace amounts (of the order of nl/g of tissue), PFC was eliminated through the lungs upon return of gas ventilation. Headspace chromatography has been recently employed to study the elimination and exhalation kinetics of a PFC ultrasound contrast agent administered intravenously to healthy volunteers (19).
Our current study was designed to focus on the potential for PFC penetration into pancreases under conditions close to those found in the clinic. For this reason, the TLM was set up with preoxygenation of the PFC, although changes in oxygen tension are unlikely to make any difference to the penetration of the chemical itself. Of course, oxygen tension in the PFC would become important if metabolic effects of the TLM are to be investigated, but 19F MR spectroscopy cannot provide information about this. We are currently investigating other MR experiments using different nuclei (31P) for future work.
In the present study, with proton signal MR imaging the pancreas was clearly seen but the reference syringe containing PFC was not visualized. This was expected because there are no hydrogen atoms in the perfluorodecalin molecule. However, the PFC syringe was readily seen with the fluorine imaging MR sequence used, whereas no signal was detected from the pancreatic segment. This indicates that any PFC present in the tissue was below the detection limit of the MR experiment as no fluorine signal could be obtained from the pancreas from a series of volumes of 0.25 mm3. In addition, the characteristic spectrum of PFC was not obtained from the TLM-preserved pancreas by 19F MRS. Further confirmation of this was obtained when 19F MR spectra of UW- and TLM- preserved pancreas were compared with typical spectra obtained from the experimental phantoms (syringes filled with perfluorodecalin and sodium fluoride). There was no difference between the 19F MR spectra between UW- and TLM-preserved pancreas: no 19F resonance was observed in either spectra and they both lacked “PFC” regions. It is unlikely that leakage of PFC from the pancreas tissue during sample could have contributed to the absence of signal. Because the scans were performed immediately after the end of the preservation period and because the samples interrogated were whole blocks of pancreas with an intact capsule, there was very little possibility of PFC leakage prior to the measurements. In order to avoid PFC signal from the preservation solution, these measurements were not carried out in situ and the blocks were suspended in UW solution instead. Initially, nonlocalized spectroscopy was performed on the whole sample volume because nonlocalized spectroscopy has greater sensitivity than 19F MRI to measure PFC, because the former allows the detection of any PFC throughout whole sample. Here again, PFC was not detected. Conversely 19F MRI will only detect PFC if above the limit of detection in a defined volume, which in this instance is 25 mm3.
Headspace gas chromatography confirmed the MR data. There was no significant difference in GC measurements between samples of tissues from pancreases stored either with UW alone or with TLM.
No significant difference in values were measurable in the PFC content of pancreases exposed to PFC and the background concentration of PFC (which would result from environmental contamination) in pancreas stored in UW solution. Taken together, our results demonstrate that PFC does not permeate into a large-animal pancreas when preserved by TLM even for 24 h. These findings are consistent with the observation by Papas et al. that pO2 level in the core of solid pancreas preserved in TLM is zero and that there is no difference in tissue ATP levels between TLM- and UW-preserved human pancreases (26). These results also explain the limited benefit with the TLM of preservation in human islet transplantation, as observed in a recent meta-analysis (1). Any beneficial effect of PFC in the TLM clearly does not depend on PFC penetration and hence is most unlikely to relate to oxygen transport. Alternative mechanisms that require further investigation include the possible ability of PFC to act as a sump for CO2 or other organic molecules.
PFCs, in other circumstances, have been shown not to cross biological membranes. When used for liquid ventilation, there is evidence that PFC is partially absorbed across the alveolus. However, even though the blood–gas barrier is only 0.2–0.3 μm thick (8), only trace amounts of PFC (0.001 mg/g tissue) were detected in peripheral tissues by gas chromatography (4,23). There is no histologic evidence of presence of residual PFC liquid in tissues except for local accumulation of vacuolated macrophages in the lungs and associated lymph nodes. It has been estimated that only approximately 0.5 ml of PFC escapes into the circulation during standard experimental liquid ventilation in dogs (the endotracheal tube being connected to a reservoir containing 1–1.2 L of liquid fluorocarbon for 1 h), either directly or via macrophages (4,12).
A number of investigators have observed that when used for liquid ventilation, perfluorochemicals have a beneficial anti-inflammatory action in the alveolar space. In vitro studies have demonstrated that PFC decreases cytokine and reactive oxygen species production from and attenuates chemiluminescence of lipopolysaccharide-stimulated rabbit and human alveolar macrophages (31,33). Although it was speculated in the past that this immunosuppressive property of PFCs might translate into reduced rejection episodes and improved graft survival following islet transplantation (20) on the presumed basis that PFC might permeate the pancreas, none of the trials comparing TLM with conventional preservation has demonstrated such an advantage and recent studies have revealed that there is no significant difference in immunogenicity of islets isolated from fresh pancreas and those preserved in TLM (34).
Conclusion
There is no evidence of penetration of PFC into large-animal pancreases preserved by the TLM. The hypothesis of a physiological effect mediated by improved gas transfer in a solid pancreas is not plausible. The mechanism of action of the TLM remains unclear and its precise role in pancreas preservation for islet transplantation needs to be revisited.
Footnotes
Acknowledgments
This work was partly funded by the Henry Smith Charity (No. 230102) and Transplants to End Diabetes (No. 1103596). This work was presented at Cryo 2007: 44th annual meeting of the Society for Cryobiology (July 28 to August 1, 2007) and British Transplant Society 10th Annual Congress (March 28–30, 2007).