Search

The object of this study was to evaluate differences in islet diameters, their distribution and both cross sectional surface areas and densities of insulin containing islets between adult and juvenile porcine pancreata using a computerised image analysis system (Improvision). Five adult (A) (2-3 yrs) and 5 juvenile (J) (< 12 mths) Large White pancreata were assessed. Biopsies were taken from 5 different regions (posterior lobe, duodenal lobe, along with the head, body and tail regions of the splenic lobe) of the pancreas and tissue sections stained for insulin. In both A and J pancreata islet numbers increased with decreasing islet diameter, showing a skewed distribution. There was no statistical significance between the cross sectional surface area within A (mean 5.04 x 10(3) microm2) or J (mean 5.99 x 10(3) microm2) pancreata. Assuming islets are spherical, extrapolation from pir2 showed that the mean diameter for A was 80 microm and 87 microm in J. These compared with A 77 microm and J 86 microm diameters using conventional microscopic techniques. The percentage islet volume density relative to exocrine tissue, derived from the principle of Delesse (Area density = volume density), did not significantly differ between each of the 5 areas studied, either in A or J. The percentage islet volume densities did show a significant difference between A (mean 1.83%) (P = 0.001) but not between J pancreata (mean 2.13 %). In conclusion poor islet yields can be attributed to differences in islet volume density of islets within porcine pancreata. These results also suggest that the posterior and duodenal lobes should be used along with the splenic lobe in order to improve porcine islet yields. Furthermore, the current practise of reporting porcine islet yields and the isolation index relative to 150 microm (IEQs) needs to be redefined, based on the assumption that the average size of an adult porcine islet is 80 microm.

The activation of endogenous pancreatic enzymes during automated pancreas digestion may be detrimental to islet isolation. In this report we assessed the activation of trypsin, chymotrypsin, elastase, carboxypeptidases A and B, phospholipase A2, and lipase using a porcine model. Four islet isolations were examined. Duplicate aliquots were taken from the automated circuit at 5-min time intervals up to the completion of pancreas digestion (approx 60 min). One aliquot was activated in vitro with exogenous trypsin in order to convert the enzymes into their active non-"proform," with the exception of trypsinogen, which was activated with exogenous enterokinase. This was done to assess the percentage activation of each individual enzyme (total potentially activatable enzyme release). The extent of activation between isolations was extremely variable. During the closed (recirculating) circuit phase of pancreas digestion there were both gradual and rapid increases in the levels of enzymes released. Peak activity of enzyme activation varied from 13 to 30 min; similarly, total potentially activatable peaks occurred between 13 and 38 min. Lipase and carboxypeptidase B showed greater than 70% activation, chymotrypsin, carboxypeptidase A, and phospholipase A2 between 50% and 70% activation, and trypsin and elastase less than 20%. There were up to 30-fold differences between the four islet preparations. In summary, it is unlikely that poor islet yields are soley explained by variations between collagenases; the variable activation of endogenous pancreatic exocrine enzymes is also likely to be influential to porcine islet yields.

The collagenase digestion phase of islet isolation is variable and unpredictable. This results partly from the vagaries of collagenase itself but also from the complex effects of organ retrieval and storage on collagenase digestion. Improvements in the purification of islets by density gradient centrifugation will not result from the production of new density gradient media, but rather from the continued modification of the biochemical composition of the solvents in which established gradient media are dissolved. The challenge is to produce solutions that will minimise acinar tissue swelling without compromising islet yield and viability. It is sobering to note the low recovery of islets after 48-h tissue culture (56%), and it has to be concluded that present tissue culture techniques for human and porcine islets are inadequate. Part of the problem is that the media and techniques used for islet tissue culture were designed for single cell culture and were not formulated specifically for the culture of insulin-producing mini-organs with a high metabolic rate. Cell viability is notoriously difficult to quantify, and this is no less true for isolated islets than for other tissues. The main problem is that although it is possible to determine the viability of isolated islets by measuring specific cellular functions such as glucose-stimulated insulin release, it is impossible to compare this with the same function performed by islets within the native pancreas. For this reason, it is not possible in absolute terms to determine the effect of islet isolation on islet viability.

Background: Different treatment regimens for minimizing microbial contamination of purified porcine islets were analyzed., Methods: Purified islets, prepared from abattoir-excised pancreata (n=26), were cultured in Hams F-12 with or without antibiotics after islet isolation. Aliquots of pancreas transport media were all contaminated with microbes (100%). Of those islets cultured with no antibiotics, 70% were contaminated, compared with 42% of the islets cultured in penicillin and streptomycin (P/S) (P>0.05). A further 20 consecutive pancreata were then randomized to saline washing (N=10) before islet isolation. Islets were then cultured in either Hams F-12 with P/S (groups C and E) or Hams F-12 with gentamicin, penicillin, and amphotericin B (G/P/A) (groups D and F)., Results: None of those purified islets prepared from washed pancreata and then undergoing culture in Hams F-12 with G/P/A (group F) were contaminated. This compared with 30% contamination when islets were cultured in P/S (group E). Neither antibiotic regimen compromised stimulated insulin release (P=0.13)., Conclusions: In conclusion pancreas washing in saline and culture in G/P/A was shown to eradicate detectable microbial contamination of purified porcine islets isolated from abattoir-excised pancreata.

Methods to minimize the effect of cold ischaemia on porcine islet isolation were investigated. Forty pancreata were randomized to intraductal collagenase delivery in University of Wisconsin solution (UW) or Hanks balanced salt solution (HBSS) (control) both before and after 65 min of cold pancreas storage. Collagenase was also administered in a specially designed cold storage solution (University of Leicester solution, ULEIC), before cold storage. Median islet yield was significantly greater if the pancreas was distended with collagenase in either UW (21524 islet equivalents (IEQ) per g) or ULEIC (19814 IEQ/g) before cold storage, compared with that after distension with HBSS (6924 IEQ/g) following cold storage (P < 0.05). Islet fragmentation, islet purification and glucose-stimulated insulin release were not significantly different after collagenase delivery in either UW or ULEIC compared with those after administration in HBSS. It is concluded that porcine islet yields can be improved significantly by intraductal collagenase administration in either UW or ULEIC immediately after excision of the pancreas.

Neuropeptide Y (NPY) has been shown to decrease insulin secretion from rodent islets. NPY messenger ribonucleic acid (mRNA) has been demonstrated in rat and mouse pancreatic islets. We, therefore, examined human islets for the presence of NPY-encoding mRNA and NPY-like immunoreactivity. Human pancreatic islets were obtained from cadaveric organ donors, using collagenase digestion and purification on BSA density gradients. Northern blot analysis, employing a human NPY riboprobe, revealed specific NPY-encoding mRNA in the islet. Compared to the islet, NPY message abundance was 9-fold higher in the caudate nucleus and 2.4-fold higher in the temporal lobe, but it was 75% lower in the adrenal gland. NPY-like immunoreactivity was present at 2.4 +/- 0.3 fmol/microgram protein in acid-ethanol extracts from the islets. On fast protein liquid chromatography with a reverse phase column, the majority of NPY-like immunoreactivity eluted as a peak with a retention time identical to that of porcine NPY standard. Added NPY (100 nmol/L) decreased (P = 0.001) glucose-stimulated (8 mmol/L) insulin release from the human islets by 45% in a perfusion system. Therefore, human islets synthesize substantial amounts of NPY, which could act as an intra-islet paracrine regulator.

Although human islet allotransplantation has been shown to be effective in restoring normoglycaemia to diabetic recipients, the long-term success rate is relatively low. The poor results of human islet transplantation are surprising given the relative ease of establishing functioning islet allografts in rodent models with minimal immunosuppression and have been attributed to a number of factors of which rejection may play a major part. The impracticality of studying the immunogenicity of human islets in vivo has led us to develop a model specifically to study the allogeneic immune response to the major components of isolated human islet preparations. To try and overcome the variability associated with this methodology, this study was designed to develop and optimize the mixed lymphocyte coculture as an in vitro model for the initial alloresponse to both freshly isolated intact human islets (MLIC) and acinar tissue pieces (MLAC). For this, titration of the islets and acinar tissue, as well as kinetic studies of the response, were used to select optimum conditions for the MLIC which were shown to be ten islets and ten acinar tissue pieces with a coculture duration of 7-9 days. The optimized studies included the use of non-tissue culture microwells to deter fibroblast growth, RPMI + 10% human AB serum to reduce the background lymphocyte response and a low concentration of dithizone to stain the islets prior to handpicking. Freshly isolated human islets were found to stimulate allogeneic lymphocytic proliferation; however, differences in the relative response to islets and acinar tissue led us to test the effect of soluble products from the acinar cells on lymphocyte proliferation in the mixed lymphocyte reaction. An inhibitory effect of the soluble products of acinar cells on the allogeneic lymphocyte proliferative response was found to contribute to the reduced response of the MLAC compared to the MLIC. Human islets were shown to stimulate an allogeneic immune response in vitro, and this MLIC model provides a method for assessing the specific immunogenicity of human isolated islet preparations and associated acinar tissue in transplantation.

Cell and tissue culture techniques have improved considerably since the first attempts to maintain explants of animal tissue in vitro. The two major developments that have allowed these improvements are the ability to produce continuous cell lines, thus allowing reproducible results to be obtained, and the definition of media for different cell types, thereby reducing the need for supplements of serum and other extraneous extracts. The requirements of islets in culture have been more difficult to define, largely because islets do not proliferate in culture and proliferation rate cannot therefore be used to measure the suitability of the medium. Further difficulties arise because islets are highly metabolically active "mini-organelles." Although many studies have been undertaken to try and optimize media for the culture islets of Langerhans, the media most commonly used are commercially available media developed for other cell types. There remains ample scope for further refinement of the composition of islet culture media, with the possibility of different media for islets from different species.

There have been no extensive studies investigating the effect of tissue culture media on the in vitro functional characteristics of rat, porcine and human Islets of Langerhans. We therefore aimed to compare ten commercially available tissue culture media on the basis of their ability to maintain islet viability. Following isolation, islets were cultured free-floating in the ten media (RPMI 1640-11mM glucose (control), RPMI 1640-2.2mM glucose, Dulbecco's MEM, TCM 199, CMRL 1066, Iscove's MEM, Waymouth's MEM, Serum-Free medium, Ex-cell 300, Ham's F-12) and viability was assessed after 24 hr, 3 days, and 7 days on the basis of macroscopic appearance, cell membrane integrity, and insulin secretion in response to glucose stimulation both by dynamic incubation and by perifusion. Each islet species demonstrated physiological insulin release characteristics in all media--however, it was possible to distinguish between the media by comparing the stimulation indices calculated from the insulin release studies. Significantly higher stimulation indices were produced in Iscove's MEM for rat islets, in Ham's F-12 for porcine islets and in CMRL 1066 for human islets. Over the entire culture period a significant deterioration in function was observed in all species cultured in the control media, although this was reversed when islets were cultured in the optimal media. Furthermore, in the case of porcine and human islets a significant improvement in function over the seven-day period was noted in the optimal media. In conclusion, of the commercially available media, the optimal tissue culture medium for rat islets is Iscove's MEM, for porcine islets is Ham's F-12, and for human islets is CMRL 1066.

Purification of islets based on the physical differences in density between exocrine and islet tissue reduces islet yields and remains one of the factors limiting islet transplantation. Immunomagnetic cell separation methods provide an attractive, highly specific alternative capable of rapid, gentle, high volume cell separation, but they require modification to be applied effectively to separation of the much larger tissue fragments involved in islet purification. In this study, mAb to rat exocrine tissue were coupled to 4.5-microns magnetic beads (M450 Dynabeads), before incubation with standard aliquots of rat pancreatic digest. The effect on immunomagnetic islet purification of modifications in the magnetic field and the method of digest release into the field were investigated. The results showed that using vibration to maintain the immunomagnetically labeled digest in suspension in tissue culture medium whose density had been increased by the addition of BSA, significantly improved the purification process. When the digest suspension was slowly released and allowed to drift under gravity through a magnetic field applied across a narrow tube, the use of a quadripole of permanent magnets improved results compared with bipolar or unipolar magnetic fields. By modifying immunomagnetic cell separation techniques in this way, a median islet yield of 77% could be reliably achieved while removing 88% of the contaminating exocrine tissue. The use of such methods in human islet purification would significantly increase the yield of islets from each donor pancreas and increase the success rate of transplantation from single donors.

The density-dependent purification of islets from several species of mammalian pancreata is improved by prior storage of the dispersed, collagenase-digested pancreas in suitable storage solutions, such as University of Wisconsin (UW) solution. The optimal composition of such solutions, however, is not fully established, although previous investigations have suggested separately that cellular impermeants and colloids are important components. To investigate this issue further, dispersed tissues from 7 porcine and 7 human pancreata were stored in UW or in solutions containing the impermeants lactobionate and raffinose, with either no added colloid or in the presence of the colloids hydroxyethyl starch, dextran 40, dextran 250, or Ficoll 400; hydroxyethyl starch-containing solutions in which the principal cation was sodium, rather than potassium, were also studied. Subsequent purification of islets on continuous linear density gradients of BSA was then assessed by insulin/amylase assay of gradient fractions. Islet purity was slightly reduced using solutions containing impermeants but lacking a colloid, compared with using UW. In the combined presence of impermeants and a colloid, however, islet purity was similar to that obtained with UW, and for porcine pancreata, solutions containing Ficoll 400 or dextran 40 were slightly superior to UW. Purity was not, however, influenced by the sodium to potassium ratio of storage media. In conclusion, impermeants and colloids are both essential components of solutions used to preserve pancreatic tissue before islet purification, findings which may be relevant when designing media for use during other phases of islet isolation, e.g., during collagenase digestion/density gradient purification.

Using isolated beta-cells from human islets of Langerhans we have demonstrated that purinergic receptor agonists are functionally coupled to rises in the intracellular calcium ion concentration ([Ca2+]i). The effects of ATP, ADP and AMP have been examined over a range of concentrations, 0.5 to 500 microM. The actions of ATP were more potent than those of either ADP or AMP suggesting that a P2-type of purinergic receptor operates in these cells. Responses to ATP were concentration-related, but exhibited marked desensitisation at high concentrations (> 100 microM). Purinergic receptor agonists elevate [Ca2+]i by mechanisms that involve both Ca2+ influx and Ca2+ mobilisation from intracellular stores. The physiological significance of our data has been discussed, and related to previous studies carried out upon rodent and clonal insulin-secreting cells.

The purification of human pancreatic islets before transplantation relies on the density-dependent separation of islets from exocrine fragments after collagenase digestion of the donor pancreas. The results vary among pancreases despite increasing automation of the digestion and purification processes, reflecting variations in the overlapping densities of islets and contaminating exocrine tissue. Hypothermic storage of both the pancreas and the pancreatic digest alters cell volumes and tissue densities, thereby affecting islet purification. By biochemical analysis of the isopycnic distribution of islets and exocrine tissue fragments from 23 human pancreases on linear continuous density gradients, the effect of various solutions for cold storage of pancreatic digest was studied. The use of the University of Wisconsin cold storage solution, which resulted in a significant decrease in digest volume (P = 0.006) and increase in the densities of both exocrine tissue (P = 0.001) and islets (P = 0.005), produced a significant improvement in islet purity compared with tissue culture medium (P = 0.035), predominantly due to the inclusion of a colloid, which increased the difference in density between exocrine tissue and islets. The addition of large molecular weight cellular impermeants without alteration in the concentration of permeable anions produced no effect. The results of this study support the concept that the use of solutions that minimize cell swelling throughout the process of islet purification would result in significant improvements in density-dependent islet separation, and that such solutions should contain a colloid.

The effects of the mixed alpha/beta-agonist adrenaline on insulin secretion from isolated human islets of Langerhans were studied. In static incubation experiments, adrenaline (0.1 nmol/l to 10 mumol/l) caused a concentration-dependent inhibition of glucose-induced insulin secretion from isolated human islets. However, perifusion experiments revealed that the time-course of the secretory changes induced by adrenaline was complex. When employed at a high concentration (1 mumol/l), adrenaline caused a sustained inhibition of glucose-induced insulin secretion, which could be relieved by the addition of the alpha 2-antagonist yohimbine (10 mumol/l). By contrast, infusion of adrenaline at a lower concentration (10 nmol/l), produced a large initial potentiation of glucose-induced insulin secretion. This response was, however, short-lived and followed by sustained inhibition of secretion, which could be relieved by yohimbine (10 mumol/l). The initial stimulation of insulin secretion provoked by 10 nmol adrenaline/l was abolished when islets were incubated in the presence of the beta-antagonist, propranolol (1 mumol/l), consistent with activation of beta-adrenoceptors. In support of this, treatment of human islets with the selective beta 2-agonist clenbuterol, was also associated with marked stimulation of insulin secretion. By contrast, each of two selective beta 3-agonists tested failed to alter insulin secretion from human islets. The results indicate that human pancreatic B-cells are equipped with both alpha 2- and beta 2-adrenoceptors which can affect insulin secretion. Adrenaline interacts with both of these but the alpha 2-response is predominant and can overcome the tendency of beta 2-adrenoceptors to potentiate insulin release.

University of Wisconsin solution has become the most commonly used vascular perfusate during multiorgan donation world-wide. In the UK however, hyperosmolar citrate remains in common use. The purpose of this prospective randomized study was to compare the effect of systemic perfusion with UW or HOC on subsequent islet yield and purification for pancreata with short cold ischemic times. Seven pancreata were randomized to each group, with the donor age, pancreas weight, and period of cold ischemia being similar in both. Perfusion with UW was shown to inhibit collagenase digestion, and a higher concentration of this enzyme was needed to achieve comparable numbers of islets with good separation of exocrine and islet tissue after a similar period of digestion. There were no differences in the number, size, purity, or viability of islets between the two groups. In conclusion, UW solution offers no benefits over HOC for pancreata with short cold ischemic times, and because of its expense and need to use greater amounts of collagenase enzyme, we continue to use HOC.

The aim of this study was to establish the beneficial effect of storage of pancreatic digest in University of Wisconsin solution on porcine islet purification, the mechanism of this effect, and the components of UW responsible. Ten porcine pancreata were collagenase-digested, and samples of digest were washed and stored for 1 hr in either UW or minimum essential medium at 4 degrees C, prior to separation on continuous linear density gradients of bovine serum albumin. Samples of digest from a further ten pancreata were similarly treated, comparing storage in MEM, UW, and five solutions varying in lactobionate:chloride ratio and raffinose content. The purity of the islet preparations and the densities of islets and exocrine tissue were determined from insulin and amylase assay of aliquots aspirated from these gradients. Washing and storage of digest in UW markedly improved islet purity, compared with MEM, due to an increase in the density of exocrine tissue. Exocrine tissue density following storage was dependent upon the control of acinar cell volume, rather than exocrine enzyme discharge, and was determined primarily by the chloride:lactobionate ratio of the storage solution. Raffinose was of little additional benefit, while the beneficial effect of UW was greater than that due to its lactobionate and raffinose content alone. In conclusion, inadequate purification of islets results from exocrine tissue swelling. This swelling is reduced by storage of the pancreatic digest in UW solution, due primarily to the replacement of chloride by lactobionate in UW.

Although a number of different protocols have been described for the cryopreservation of pancreatic islets of Langerhans, most involve equilibration with 2 M Me2SO followed by slow cooling and fast warming. This paper compares two methods with identical equilibration protocols: one with cooling at 0.25 degree C/min to -40 degrees C prior to transferring to liquid nitrogen and subsequent warming at 200 degrees C/min and sucrose dilution of Me2SO (Method A) and the other with cooling at 0.3 degree C/min to -70 degrees C prior to transferring to liquid nitrogen and warming at 50 degrees C/min followed by stepwise dilution of Me2SO (Method B). Islet viability was assessed by syngeneic transplantation to the renal subcapsular space of streptozotocin-induced diabetic rats. Seven hundred fifty fresh islets in the size range 100-200 microns consistently reversed diabetes (blood glucose < 10 mmol/liter) whereas 1000 Method A cryopreserved islets and 2000 Method B cryopreserved islets were required. Cryopreservation by either method resulted in impaired islet function compared to that of fresh islets, although this functional loss could be partially compensated by the transplantation of greater numbers of cryopreserved islets.

Euro-Ficoll and bovine serum albumin (BSA) are two of the most commonly used density gradient media for the purification of pancreatic islets. Euro-Ficoll is based upon Euro-Collins, a cold storage medium, and must, therefore, be used at 4 degrees C. The ionic composition of BSA, however, is likely to contribute to hypothermic cellular swelling, and this may influence the efficiency of islet purification using this medium at 4 degrees C. Experience in this laboratory also suggested that batch-to-batch variation in islet purity using BSA was related to differences in BSA osmolality. The aim of this study was to assess the effect of gradient medium temperature and osmolality on the purification of human and porcine islets using BSA. Pancreata were collagenase-digested, and islets were purified on continuous linear density gradients of BSA. The distribution of insulin and amylase in each gradient was assayed, and used to calculate the median density of islets and exocrine tissue, and the efficiency of islet purification (% amylase contamination at a fixed insulin yield), using: 1) gradient osmolalities of 300, 400, and 500 mOsm/kg H2O (seven porcine pancreata), and 2) gradients at 4 degrees C and at 22 degrees C (eight human and seven porcine pancreata). Increase in density gradient osmolality produced increases in porcine exocrine tissue density which exceeded changes in islet density, resulting in improved islet purity, maximal at a BSA osmolality of 400 mOsm/kg H2O. For human pancreata there was no significant change in pancreatic tissue densities nor islet purity with temperature.(ABSTRACT TRUNCATED AT 400 WORDS)

A prospective study was undertaken to compare an automated method of porcine pancreatic digestion with a simpler manual procedure. These techniques have not previously been compared directly. After intraductal distension with collagenase, seven porcine pancreata were divided longitudinally; half of each was digested by the automated method and half by the manual technique. Islet yield and purity were measured. Compared with the manual technique, the automated method isolated a significantly greater total volume of islet tissue (median (range) 3.56 (1.39-5.30) versus 1.07 (0.46-1.92) mm3/g, P = 0.022), increased the median (range) number of 105-microns islet equivalents isolated (5875 (2294-8746) versus 1766 (759-3168) per g, P = 0.022) and improved the islet cleavage index (median (range) 92 (89-99) versus 82 (78-92) per cent, P = 0.035). It is concluded that, although the automated method is more complicated to set up, it greatly improves the yield of intact islets from the porcine pancreas.

Euro-Ficoll (EF) and bovine serum albumin (BSA) are the two most commonly used media for the density gradient purification of human pancreatic islets. The aim of this study was to compare these two media with respect to the efficiency of human islet isolation. Ten human pancreata were collagenase-digested, and samples of digest were separated on either a continuous linear density gradient of BSA or a discontinuous gradient of EF (1.108/1.096/1.037/Euro-Collins). Efficiency of islet purification was assessed by insulin and amylase assay of aliquots aspirated from the BSA gradients, and from the interfaces of the EF gradients. Islets were obtained from two interfaces in the EF gradients. Islet yield from the upper interface was generally poor (median 28% of total insulin; range 2-71%), but purity was better than for an equivalent yield using BSA [1% (0-3%) amylase contamination for EF versus 6% (0-37%) for BSA; P = 0.013]. Pooling both EF interfaces increased yield to 66% (17-81%) but markedly reduced purity [46% (0-50%) amylase for EF versus 31% (0-52%) for BSA]. In conclusion, the efficiency of human islet purification is similar, though disappointingly low, with BSA and with EF. Considerable scope exists, therefore, for improvement in the density gradient purification of human islets.

The use of the COBE 2991 cell processor (COBE Laboratories, Colorado) for large-scale islet purification using discontinuous density gradients has been widely adopted. It minimizes many of the problems such as wall effects, normally encountered during centrifugation, and avoids the vortexing at interfaces that occurs during acceleration and deceleration by allowing the gradient to be formed and the islet-containing interface to be collected while continuing to spin. We have produced cross-sectional profiles of the 2991 bag during spinning which allow the area of interfaces in such step gradients to be calculated. This allows the volumes of the gradient media layers loaded on the machine to be adjusted in order to maximize the area of the gradient interfaces. However, even using the maximal areas possible (144.5 cm2), clogging of tissue at such interfaces limits the volume of digest which can be separated on one gradient to 15 ml. We have shown that a linear continuous density gradient can be produced within the 2991 bag, that allows as much as 40 ml of digest to be successfully purified. Such a system combines the intrinsic advantages of the 2991 with those of continuous density gradients and provides the optimal method for density-dependent islet purification.

The purification of large numbers of human pancreatic islets remains one of the limiting factors in islet transplantation. This paper describes and validates a method for accurately and reproducibly determining the density of islets and exocrine tissue in pancreatic digest on the basis of their isopycnic distribution on linear continuous density gradients. The use of this data to analyse and compare the purity of a standard 60% islet yield is described. The results obtained using such gradients will enable factors responsible for the variation in yield between pancreases to be determined and optimized, improving the results and reliability of islet purification.

Density-gradient purification of human pancreatic islets from the collagenase-digested pancreas relies on the exocrine tissue being denser than the islets. Cold storage of the pancreas before and after digestion causes cell swelling, which can decrease the density of pancreatic exocrine tissue and adversely affect subsequent purification. Using 14 human pancreata (seven perfused in situ with hyperosmolar citrate (HOC) and seven with University of Wisconsin solution (UW)), it is shown that storage of the pancreatic digest in UW significantly increases the density of pancreatic exocrine tissue compared with storage in minimal essential medium (MEM) (P = 0.009). This results in an improvement in islet purity (P = 0.036) for HOC- but not UW-perfused pancreata. Storage in UW for 1 h not only prevented the deterioration that occurred in MEM, but resulted in an improvement in islet purity for five of the seven HOC-perfused pancreata. Most pancreata in the UK are perfused with HOC, but storage of the digest in UW results in significantly better islet purity and, when islets cannot be purified immediately, a period of storage will often improve separation and allow islets to be purified.

The selective beta 2-adrenergic agonist clenbuterol was ineffective as a stimulus for insulin secretion when isolated rat pancreatic islets were incubated with glucose at concentrations between 4 and 20 mM. Inclusion of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine led to potentiation of glucose-induced insulin secretion, but did not facilitate stimulation by clenbuterol. Furthermore, maintenance of isolated rat islets for up to 3 days in tissue culture also failed to result in the appearance of a secretory response to beta-agonists. By contrast, clenbuterol induced a dose-dependent increase in insulin release from isolated human islets incubated with 20 mM glucose. Clenbuterol did not increase the basal rate of insulin secretion (4 mM glucose) in human islets. Under perifusion conditions, the secretory response of human islets to clenbuterol was rapid, of similar magnitude to that seen under static incubation conditions and could be sustained for at least 30 min. The increase in insulin secretion induced by clenbuterol was inhibited by propranolol, indicating that the response was mediated by activation of beta-receptors. In support of this, a similar enhancement of glucose-induced insulin secretion was elicited by a different beta 2-agonist, salbutamol, in human islets. The results indicate that the B cells of isolated rat islets are unresponsive to beta-agonists, whereas those of human islets are equipped with functional beta-receptors which can directly influence the rate of insulin secretion.

Islets of Langerhans were isolated from WAGola rats and encapsulated in high mannuronic acid sodium alginate droplets. The capsules were completed in the following ways: (a) poly-l-lysine alone, (b) poly-l-lysine plus high guluronic acid alginate, (c) poly-l-lysine plus high mannuronic acid alginate. Islet viability was assessed every 3-4 days over a period of 4 weeks, by a microfluorometric assay using fluorescein diacetate and propidium iodide, and every 7 days by perifusion. The perifusion results were corrected for DNA content to take islet size differences into account. The results from the microfluorometric assay showed a generally high viability of the islets in all groups at each time interval. The perifusion experiments showed that the response time and stimulation indices of the islets was similar in all groups, but the absolute amount of insulin released was slightly lower from encapsulated islets relative to unencapsulated controls. It was concluded that the presence of a capsule, regardless of its composition, did not adversely affect the membrane integrity of the islets, their response time, nor their proportionate increase in insulin release following glucose stimulation. However, the absolute amount of insulin released from encapsulated islets was reduced during perifusion. This reduction was apparent after 1 week in culture, but remained stable thereafter.

A simple technique for the controlled collagenase digestion of the human pancreas is described. The pancreas is distended with collagenase, and a biopsy taken and divided into 5 pieces that are placed in Universals containing minimal essential medium and dithizone at 39 degrees C. The pancreas itself is incubated in MEM at 39 degrees C. Starting at 5 min and at intervals thereafter, a Universal is removed from the water bath, shaken for 30 sec, and the contents examined by microscopy. As soon as free cleaved islets are seen, the pancreas is placed into one compartment of a kidney-bowl divided in half by a 1-mm mesh. The pancreas is gently teased apart and fluid digest in the empty half of the bowl aspirated and passed through a 500-micron mesh into ice-cold MEM containing 20% newborn calf serum. This process is repeated until the digestion process has ceased. Using this technique on 20 consecutive pancreata, median wt. (range) 53.9 (45.2-72.9) g, we have counted 131,672 (43,516-400,000) islets in the digest, equivalent to 2394 (715-8000) islets/g pancreas. The volume of islet tissue in the digest was 299 (26-1341) mm3 equivalent to 5.81 (0.36-26.81) mm3/g pancreas. In conclusion, we have found this simple technique to be an effective method for the controlled collagenase digestion of the human pancreas.