Frances L. Shienvold

Natural history of intrahepatic canine islet cell autografts.

We have serially followed the function of intrahepatic canine islet autografts in 15 beagle dogs for up to 24 mo. Of these, only 20% sustained normal levels of fasting blood glucose for greater than 15 mo posttransplant. Failure of autograft function was accompanied by a preferential loss of well-granulated beta cells in the engrafted islets. The chronic stimulation of an initially marginal intrahepatic beta-cell mass ultimately resulted in metabolic deterioration and loss of beta cells below the minimal threshold required to maintain normal fasting blood glucose levels. It is possible that transplantation of a larger mass of islets would result in indefinite graft function in dogs. However, it remains to be demonstrated in larger mammals, including humans, whether an islet cell mass that is initially adequate in a heterotropic site such as the liver can remain functionally competent over a prolonged period.

A ganglioside antigen on the rat pancreatic B cell surface identified by monoclonal antibody R2D6.

In an attempt to identify B cell specific antigens, we have generated a mouse monoclonal antibody, R2D6, which is directed against plasma membranes of rat pancreatic B cells but against no other pancreatic cells. R2D6 crossreacted with mouse and guinea pig B cells, but not with human or dog. The B cell specificity of R2D6 was utilized in fluorescence-activated cell sorting to prepare highly enriched separate populations of viable pancreatic islet B cells and A cells. R2D6 also recognized adrenal chromaffin cells, secretory cells in the anterior pituitary, and the myenteric plexus of the gastrointestinal tract. Trypsin, chymotrypsin, papain, ficin, and pronase had no effect on R2D6-binding to dissociated rat islet cells. However, neuraminidase treatment of intact cells reduced R2D6-binding by 75%. The antigen recognized by R2D6, Ag(R2D6), could be quantitatively extracted from rat islets by dichloromethane/methanol (2:1) and, after drying, was soluble in methanol alone as well as in phosphate-buffered saline. When the dichloromethane/methanol extract (DME) was bound to polyvinylchloride microtiter plates, antigenic activity was retained and remained insensitive to pronase. In this solvent-extracted form, antigenic activity was totally destroyed by neuraminidase. Therefore, sialic acid is either an integral part of, or is related sterically to the binding site (epitope) for R2D6. In high performance thin-layer chromatographs of the DME, developed in 60:40:9 chloroform/methanol/2.5 N ammonia, Ag(R2D6) migrated with a relative mobility (Rf) of 0.54 +/- 0.07 (n = 3), which was a position nearly coincident with the purified brain ganglioside, GD1a. The antigen bound to DEAE-Sephacel, was not inactivated by mild treatment with base (which hydrolyzes phospholipids) and eluted in ganglioside fractions upon C18 Sep-Pak and upon silicic acid chromatography. Hence, the solubility characteristics, enzyme sensitivities, and behavior of Ag(R2D6) in four chromatography systems are consistent with its identification as a ganglioside.

Successful Long-Term Survival of Pancreatic Islet Allografts in Spontaneous or Pancreatectomy-induced Diabetes in Dogs: Cyclosporine-induced Immune Unresponsiveness

Nineteen pancreatectomized beagles and three spontaneously diabetic dogs were recipients of canine islet allografts from one or more unrelated donors. The islets, enriched 30–45-fold for endocrine cells and contained in a packed cell volume of <1.5 ml, were engrafted in the livers of recipient animals. Treatment of diabetic recipients with cyclosporine (CsA) was begun 3–5 days before islet transplantation and the initial dosage was adjusted to attain and maintain CsA serum trough levels between 400 and 600 ng/ml. Five dogs with CsA levels less than this (155 ± 35 SEM ng/ml) at the time of transplantation promptly rejected their grafts, whereas rejection was encountered in only 1 of 17 diabetic animals in which the initial level exceeded 400 ng/ml. CsA was discontinued 30, 60, or 90 days after continuous therapy in 10 animals. Graft failure was observed 2 mo after stopping CsA in 1 animal and 5 mo in the other. Eight other islet allograft recipients have sustained fasting euglycemia for 7 and 8 mo in 2 and for at least 2 mo in the remainder. These results demonstrate that short-term CsA therapy prolongs survival of islet allografts and induces a state of immune unresponsiveness to islet alloantigens in dogs with experimental and spontaneous diabetes. The findings are unique for a nonrodent mammal and thus hold promise that similar results may be achieved for islet allografts of other mammalian species, including humans.

The hemidesmosome: New fine structural features revealed by freeze-fracture techniques

SummaryHemidesmosomes along the dermal-epidermal junction of larval and post-metamorphic newt skin have been examined in freeze-fracture replica images correlated with electron micrographs of sectioned material. Larval hemidesmosomal sites are characterized by large (200–300 Å) intramembranous granules arranged into clusters, each of which is aligned with a cytoplasmic hemidesmosomal plaque. In unfixed epidermis the granules remain attached to the A-face, while after glutaraldehyde fixation they are found on both A- and B-faces. Following metamorphosis the clusters are less distinct and localized. Replicas of unfixed B-faces and nearby cytoplasm display elongate, filamentous profiles which traverse the cytoplasmic leaflet and extend onto the B-face. The possibility that these components constitute a filamentous network serving to link tonofilaments, hemidesmosomal plaque, and basal plasmalemma is considered in view of the evidence to date. Hemidesmosomal fine structure as revealed by these studies is compared to features of desmosomes as detailed in the following report.

Preparation of Rat Islet B-Cell-Enriched Fractions by Light-Scatter Flow Cytometry

Flow cytometry has been examined as a method to separate islet cells into homogeneous subpopulations. Collagenase-isolated rat islets were dissociated into single cells and these were analyzed and sorted according to their low forward angle light scattering properties by using automated flow cytometry. Light scatter histograms showed two peaks of viable cells. Radioimmunoassay of hormone content in cell fractions collected across the two peaks showed that glucagon- containing cells were concentrated towards the left side of the left peak and somatostatin-containing cells were concentrated towards the right side of the left peak, whereas insulin-containing cells were clearly enriched in the right peak. The B-cell-enriched fraction (90% B cells, 3% A cells, 2% D cells) exhibited significant insulin secretory responses to glucose (16.7 mM), and 3-isobutyl-1-methylxanthine (0.1 mM), during a 24-h culture period, and these responses were slightly greater than those observed in the original mixed islet cell preparation (66% B cells, 14% A cells, and 4% D cells). These results indicate that flow cytometry can be applied to sort pancreatic islet cells into populations enriched in specific endocrine cell types for further study of the functions of individual cell types.

Immunocytochemical localization of HLA-DR in human islets of Langerhans.

Previous reports of allogeneic transplantation studies in rodents have postulated that the primary carriers of la antigen in islets of Langerhans are passenger leukocytes. We sought to demonstrate directly the localization of the analogous human antigen, HLA-DR, in islet-enriched fractions (IEFs), utilizing a nonpolymorphic monoclonal anti-DR (αDR) antibody. The presence of DR in the IEFs was first demonstrated by radioimmunobinding assay. Light microscopic immunocytochemistry, in frozen sections of intact (unfixed) human pancreas, revealed a staining pattern suggestive of a vascular distribution of DR in islets. Ultrastructural localization of DR was then carried out by indirect immunoperoxidase labeling in the presence of NaN3 (to prevent internalization of bound αDR). The major site of DR expression in the islet was the endothelial cell surface. Endocrine cells were entirely devoid of αDR binding. Nonislet endothelium was also heavily labeled, but acinar and ductal cells were completely negative. Leukocytes bound αDR but were relatively rare in the IEFs. Human islets, therefore, clearly express HLA-DR, but predominantly on insular endothelial cells. Isolation of pure endocrine cell populations, specifically free of endothelium, would appear to be a rational approach to reducing immunogenicity in allogeneic transplantation.

Natural history of multiple intrahepatic canine islet allografts during and following administration of cyclosporine

Cyclosporine (CsA) prevented acute rejection of intra-hepatic canine islet allografts (IHIA) in 5/5 pancreatectomized dogs. Normal fasting blood glucose levels were sustained in these dogs for 210±78 days (mean ± SEM) following withdrawal of CsA. We tested whether combining islets from more than one pancreas would improve function and prolong islet allograft survival during and following administration of CsA. Areas under the glucose disappearance (GDA) and C-peptide response (CPA) curves following i.v. glucose (0.5 g/kg), 1 month posttransplant were not significantly different using islets from 1 or 2 pancreases, whereas GDA and CPA approached normal if islet yields from 3 or more pancreases were combined. Mean islet allograft survival following interruption of CsA decreased with an increase in the number of donor pancreases (one: 210±78 days, vs.two: 113±23 days, vs. >2: 57±5 days). These studies demonstrate that: (1) IHIA uniformly resulted in fasting euglycemia in 36 of 38 diabetic dogs treated with CsA; (2) normal i.v. glucose metabolism required the combined islet yield of 3 or more donor pancreases, suggesting that a substantial number of intrahepatic islet cells are functionally lost despite effective CsA-induced immunosuppression; (3) use of multiple donors to accumulate an increased mass of islets may immunologically compromise allograft survival following discontinuation of CsA (these experiments, however, do not exclude a direct relationship between the duration of CsA therapy and the duration of immune unresponsiveness following interruption of CsA in multidonor islet allografts as an independent variable); and (4) unmodified islets obtained from multiple donors seem to require continuous immunosuppression to prevent rejection.

Identification of Ia-bearing cells in rat, dog, pig, and human islets of Langerhans

Cells expressing Ia-like antigens (Ia) are believed to provide the primary recognition signals for allograft rejection. In order to clarify the localization of Ia within islets of Langerhans in animals most commonly used as transplantation models and in humans, we performed both light microscopic (LM) and electron microscopic (EM) immunocytochemical studies with the following panel of monoclonal antibodies directed against Ia: B1F6 (antidog, antirat, and antihuman Ia) and B2E8 (antidog and antihuman Ia), generated in our laboratory; OX6 (antirat Ia); and L243 (antihuman Ia). By LM, B1F6-labeled or B2E8-labeled dog islets and B1F6-labeled or OX6-labeled rat islets all exhibited Ia-specific localization on 0–15 spheroidal or stellate cells per islet. Ultrastructural identification of these cells by indirect immunoperoxidase labeling of unfixed isolated islets revealed Ia-specific cell surface binding mostly on sparsely distributed cells that exhibited ultrastructural characteristics typical of monocytes or macrophages; no specific correlation between cell shape and cell type was apparent by EM. The presence of a resident population of monocytes and macrophages in these islets was confirmed by cytochemical localization of nonspecific esterase activity. Islet endothelium was Ia-negative, as were endocrine cells. Studies of human islets with B1F6, B2E8, and L243 confirmed our preliminary findings that, although there exists a sparse population of Ia-positive leukocytes similar to those present in dog and rat islets, Ia is also expressed extensively and constitutively on the human islet vascular endothelium. This differential localization of Ia in rats and dogs compared with humans could have important implications for the development of successful strategies for the immuno-modulation of human islet allografts. We have now shown by immunopurification of 125I-labeled splenocyte plasma membrane antigens, that our MoAbs B1F6 and B2E8 also recognize porcine Ia, and immunocytochemical studies with these MoAbs demonstrated that Ia is also localized on the vascular endothelium in pig islets. Hence, this species may prove to be uniquely suitable as a model in which the immune response functions of Ia bearing islet endothelium can be explored.

Effect of anti-Ia antibodies, culture, and cyclosporin on prolongation of canine islet allograft survival.

Evidence in rodents suggests that islet pretreatment to reduce islet immunogenicity will also require some form of immunosuppression of the recipient for islet allograft acceptance in highly reactive donor-recipient pairs. We attempted to ascertain whether outbred dogs would also require treatment of both donor islets and the recipient to prolong islet allograft survival. Untreated canine islets are uniformly rejected in 6-10 days in beagles. Tissue culture alone, at 37°C for 7 days, or treatment of freshly prepared islets with antila monoclonal antibodies (MoAbs) (B1F6 + 7.2) did not prolong canine islet allograft survival. Treatment of culture-maintained canine islets with anti-la MoAbs plus complement resulted in prolongation of islet allograft survival for 188 and 368 days in two of seven pancreatectomized nonimmunosuppressed beagles. The administration of low doses of cyclosporin A (CsA) intramuscularly, to recipients of untreated canine islet allografts had no effect on graft survival. By contrast, six of nine CsA-treated recipients of islets that were also treated with anti-la MoAbs (B1F6 + 7.2) plus complement showed prolongation of graft survival. Euglycemia was sustained for 19, 34, 89, and 300 days after the CsA was discontinued (day 30) in four of these animals. Two animals had unstable grafts from the beginning that failed 23 and 29 days after transplantation. Our results indicate that simple maneuvers like short-term tissue culture at 37°C and treatment of freshly isolated islets with anti-la MoAbs and complement are inadequate to prevent rejection in outbred pancreatectomized beagles. In contrast, low-dosage CsA acts synergistically with the in vitro treatment of islets with anti-la MoAbs to prolong islet allograft survival in outbred dogs with induced diabetes mellitus.

Human pancreatic cell surface antigens with homologous expression in liver, sweat glands, and other exocrine tissues.

We have generated two panels of monoclonal antibodies (MoAbs) that represent a unique array of immunolabeling reagents specific for diverse cell surface and intracellular antigens of ductal epithelial cells (DEC) and of acinar cells (AC) within the human pancreas. Eight of the MoAbs are specific within the pancreas for the DEC plasma membrane (PM), three for DEC intracellular antigens, seven for AC-PM, and one for AC secretory granules. The MoAbs with PM reactivities have permitted us to substantially enhance our assessment of exocrine contamination in our human islet preparations. However, we have been unable to demonstrate consistently the effectiveness of these MoAbs in complement-mediated cytotoxicity protocols designed to improve the purity of our islet preparations by exocrine cytolysis. It is possible that primary cell isolates of pancreas or epithelial organs in general may lack inherent susceptibility to complement-mediated cytolysis or, instead, that this particular panel of MoAbs lacks the appropriate characteristics needed to achieve consistent complement-mediated cytolysis. We have also begun to explore the potentially broader applicability of the anti-DEC MoAbs as immunocytochemical tools and have observed intriguing patterns of cross-reactivities in other exocrine tissues, particularly bile duct epithelial cells in the liver and functionally discrete subsets of DEC in the eccrine sweat gland and the parotid gland. These homologous antigenic patterns, by virtue of their specificities for ductal tissue, most likely reflect the molecular bases for functions common to these tissues, e.g., active ion transport and much secretion. In addition, the anti-DEC and the anti-AC MoAbs exhibit patterns of binding specificity for human cell lines derived from pancreatic adenocarcinoma that suggest the potential value of these MoAbs in tumor diagnosis and therapy. We believe that these panels of MoAbs will find broad utility in immunopathology and in experimental approaches to questions regarding the development, normal function, and pathogenetic mechanisms in the human pancreas and other related organs as well.