WO1990001498A1

WO1990001498A1 - Process for purification of islet cells; monoclonal antibodies directed to islet and acinar cells; and process for making and using same

Info

Publication number: WO1990001498A1
Application number: PCT/US1989/003424
Authority: WO
Inventors: Patrick Soon-Shiong; Roswitha Heintz
Original assignee: Cell Biotech, Inc.
Priority date: 1988-08-12
Filing date: 1989-08-10
Publication date: 1990-02-22
Also published as: AU4071789A

Abstract

The present application relates to the identification and purification of islet cells which are the cells responsible for the production of insulin in mammals, and preferably, human islet cells, using monoclonal antibodies specifically directed to these cells. Further, this application is directed to the monoclonal antibodies useful in the isolation and purification of islet cells. This application is also directed to a method of making the subject monoclonal antibodies and use of the same in the detection of cells for diagnostic purposes.

Description

PROCESS FOR PURIFICATION OF ISLET CELLS; MONOCLONAL ANTIBODIES DIRECTED TO ISLET AND ACINAR CELLS; AND PROCESS FOR MAKING AND USING SAME

BACKGROUND OF THE INVENTION

1 . FIELD OF THE INVENTION :

This invention relates to the fields of biotechnology, including monoclonal antibodies, diabetes treatment and cell separation, and more specifically, to the production and use of monoclonal antibodies in isolating specific cells useful in the treatment of diabetes.

2. ART B ACKGROUND :

Diabetes mellitus is a disease which affects millions of people throughout the world. This disorder of carbohydrate metabolism is characterized by high blood sugar levels resulting from inadequate production or utilization of insulin. Standard treatment today for many of the diabetic patients comprises daily injections of insulin which prevent the patients from lapsing into a diabetic coma. Other treatments for milder forms of diabetes comprise strict dietary control in combination with various oral medications.

Attempts have recently been made by various groups to transplant pancreatic tissue from a donor to a diabetic recipient. Initial efforts in such transplantation have not been particularly successful. Pancreatic tissue is subject to transplantation rejection as a result of the body recognizing the transplanted pancreas as a foreign agent. Thus, it has been found necessary to suppress the transplant recipient's immunological system to limit the rejection. However, the suppression of the immune response in such individuals may cause patient death from infection, kidney disfunction or cancer. Variations on the concept of whole pancreatic transplantation have recently been investigated. A key to these attempts is a long-known discovery that the insulin-producing cells are disposed in the Islet of Langerhans, which is comprised of alpha, beta and delta cells, sometimes referred to herein as islet cells. The major cellular component of the pancreas consists of exocrine tissue including acinar cells, and it has been shown to be a formidable task to separate the islet cells from the acinar cells.

Once the islet cells have been separated from the remainder of the pancreatic tissue, including primarily the acinar cells, they may then be injected into the body, in various forms, for the purpose of producing normal amounts of insulin in diabetic patent's body in an interactive way to provide for the patient a normal physiological response to glucose ingestion.

While transplantation of unpurified dispersed pancreatic tissue has been performed with relative success in dogs, embolization of pancreas microfragments into the portal vein in man has resulted in portal hypertension and necrosis of the liver. Currently, no reproducible method exists for obtaining viable islets compatible with humans, in sufficient quantity and of sufficient and consistent purity.

Current techniques of islet cell separation generally rely on very sensitive and tedious methodologies with varying degrees of success. The standard technique, which is primarily practiced today, is the use of various means of density gradient separation. In fact, density gradient separation techniques standardly requires three to five technicians working five or more hours to isolate islet cells from a single pancreas. In the review article by Sharp, D.W., "The Elusive Human Islet: Transplantation of the Endocrine Pancreas" (In press), Dr. David Sharp comprehensively details the large number of variables involved in human islet isolation. There are two general steps in the isolation of islet cells, namely, pancreatic tissue digestion and islet purification. Islet digestion has been performe for many years on the rat and canine pancreases. However, the fibrous, compact nature of the human pancreatic gland rendered many of the earlier techniques, which were successfully applied to the other mammals, unsuccessful on the human organ.

Only recently have reports emerged on the isolation of the islets without the surrounding exocrine tissue. Extensive studies in pancreatic tissue digestion have been performed and various techniques, including the use of digestion and filtration chambers, pancreatic distension using venous and ductal routes, mechanical intervention including chopping devices and kitchen meat grinders, have been employed. Recently an automated islet isolator has been developed and appears to provide reasonably good results.

The second step of islet purification, has not been so successful. Mechanical purification using filtration screens or elutriator chambers has resulted in preparations which are only approximately 20% pure. Currently, most islet purification techniques utilize density gradient separation. However, density gradient separation has been found to be very inefficient because the differences in densities between the acinar cells and islet cells are inconsistent after the pancreatic islet digestion steps have been utilized.

Two methods of islet cell purification may be proposed. One method is to selectively separate the islet cells from the remainder of the pancreatic tissue, sometimes referred to as positive selection. An alternative approach is to selectively separate the acinar cells from the pancreatic tissue, again resulting in a substantially pure selection of islet cells, in a process sometimes referred to as negative selection. Of course, these two processes of positive and negative selection may be combined for the purpose of improving the islet isolation techniques.

Based upon the above discussion, it is clear that improved methods of islet purification are critical to overcoming the obstacles preventing successful clinical islet transplantation.

SUMMARY OF THE INVENTION

The present application relates to the identification and purification of islet cells which are the cells responsible for the production of insulin in mammals, and preferably, human islet cells. Further, this application is directed to the monoclonal antibodies useful in the isolation and purification of islet cells, both for positive and negative selection. This application is also directed to a method of making the subject monoclonal antibodies and using the same in the detection of islet and acinar cells for diagnostic purposes.

In the method of identifying and purifying islet cells, the first step is to obtain pancreatic tissue from a live or recently deceased donor. The pancreatic tissue is removed from the donor and treated to digest the pancreas into its component cellular material. Any of a number of different digestion procedures may be employed and are well known in the art. However, the presently preferred process is the use of an automated islet isolator described by Ricordi, et al., Diabetes. 37, 413-420 (1988). It has been found that the automated method results in minimum trauma to the ceils and maximization of islet yield.

After digestion of the pancreatic tissue, the next, and most critical step for purposes of the present invention is the purification of islet cells from the digested material.

There are two approaches to the purification of islet cells which may be used separately, or together, to achieve the cleanest sample of viable islet cells. In one approach, a positive selection separation, human islet cells are identified and removed from the digestate of the pancreatic tissue by a monoclonal antibody directed specifically to such human islet cells. For this positive approach, two monoclonal antibodies have been produced, isolated, characterized and tested for recognition of antigen determinants on human islet cells, with the concomitant inability of these monoclonal antibodies to recognize antigen determinants on the human acinar or ductal cells. These monoclonal antibodies can be used in a number of different procedures to isolate the desired islet cells from a pool of digested pancreatic tissue. These monoclonal antibodies are termed herein CIC-1 and ClC-2. CIC-2 is commercially available from Lambda-one, Los Angeles, California, and is cometimes referred to as CBL3. They are preferrably fusions of mouse spleen ceils following antiblast immunication with mouse myeloma. These monoclonal antibodies may be bound to an immobilized support to adsorb the islet cells from solution, immunolabeied and separated from the pool using a cell sorter, such as a fluorescent activated cell sorter, sometimes referred to herein as a FACS, separated based upon the magnetic characteristics of the support to which the monoclonal antibody is attached, or separated by any of a number of other methods known in the art.

The binding of the monoclonal antibodies to the islets is reversible and therefore purified islet cells can be obtained using any of a variety of variations to the present invention.

The present application also relates to a negative selection purification approach, comprising the identification of acinar cells for the purification of islet cells, using monoclonal antibodies specifically directed to these cells. Further, this application is directed to the monoclonal antibodies useful in the isolation and purification of acinar cells. This application is also directed to a method of making the subject monoclonal antibodies and use of the same in the detection of acinar cells for diagnostic purposes.

Several monoclonal antibodies have been isolated, characterized and tested for recognition of antigen determinants on acinar cells of various mammalian species, with the concomitant inability of these monoclonal antibodies to recognize antigen determinants on islet cells. The monoclonal antibodies are the fusion product of mouse spleen cells from mice immunized with human adenocarcinoma and mouse myeloma cells. These monoclonal antibodies can be used in a number of different procedures to isolate the acinar cells in order to purify the islet cells from a pool of digested pancreatic tissue. Common to all of the anti-acinar monoclonal antibodies is the apparent reactivity to blood group H antigen. Numerous such monoclonal antibodies exist and are available in the public domain.

These monoclonal antibodies, termed herein CAC-1 , CAC-2 and CAC-3, may be, for example, bound to an immobilized magnetic or non-magnetic support to draw the acinar cells from solution, immunolabeied and separated from the pool of pancreatic digestate using a fluoresence activated cell sorter, incubated with the pool of pancreatic digestate and mixed with complement to kill the acinar cells thereby leaving a population of islet cells intact and easily separable from the islet cells, or any of a number of other methods known in the art. For example, in one embodiment the monoclonal antibody directed to acinar cells may be bound to magnetic or non-magnetic immobilized supports so that simple retrieval is possible.

Additionally, since it is known that in pancreatic transplants, acinar cells are found in the patients urine if the transplant is rejected, the subject anti- acinar monoclonal antibodies may be used as a diagnostic test as an early marker for transplant rejection.

It is an object of the present invention to provide a monoclonal antibody having a specificity for human islet cells.

It is yet another object of the invention to provide a method for making monoclonal antibodies having specificity for human islet cells.

It is yet another object of the invention to provide a method for the identification of human islet cells. It is yet another object of the present invention to provide a system for the isolation and purification of human islet cells.

It is yet another object of the present invention to provide a system for isolating human islet cells for the purpose of preparing the cells for transplantation.

It is another object of the present invention to provide a diagnostic system for the detection of human islet cells for the determination of pancreatic tissue rejection.

It is an object of the present invention to provide a monoclonal antibody having a specificity for acinar cells.

It is yet another object of the invention to provide a method for making monoclonal antibodies having specificity for acinar cells.

It is yet another object of the invention to provide a method for the identification of acinar cells.

It is yet another object of the present invention to provide a system for the isolation and purification of acinar cells.

It is yet another object of the present invention to provide a system for isolating acinar cells for the purpose of obtaining a purified culture of islet cells for transplantation.

It is another object of the present invention to provide a diagnostic system for the detection of acinar cells for the determination of pancreatic tissue rejection.

Additionally, the combination of the two purification protocols may be used to obtain an islet cell population with increased purity. The foregoing and other objects in the present invention may be understood with reference to the drawings which are briefly described below and the detailed description following.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGURE 1 is a frozen section of a human pancreas showing immunoperoxidase staining of islet cells reacted with anti-islet monoclonal antibody ClC-2. Islet cell staining is clearly seen with no acinar or ductal activity.

FIGURE 2 is a controlled frozen section of human pancreas showing immunoperoxidase anti-acinar staining with anti-acinar monoclonal antibody (McAB) CIC-1 , with no islet cell activity.

FIGURE 3 is a light microscopy view of isolated human islet and acinar cells. The arrow demarcates an islet amongst contaminant acinar tissue.

FIGURE 4 is an immunofiuorescent image of the isolated islet cell with no immunofiuorescent activity of the surrounding acinar tissue, demonstrating the specificity of the fluoroscene isothiocyanate-labelled CIC-2 monoclonal antibody to islet cells and the retention of the islet antigen even after collagenase digestion of the pancreas.

FIGURE 5 is an Immunoperoxidase stain of frozen section of human pancreas, X200, demonstrating that human acinar cells express blood group antigens (positive stain), but islets do not.

FIGURE 6 shows the antigenic determinants recognized by the McAb on the acinar cell surface or cytoplasm is either ⁽ -²⁾ fucose or Gal-GlcNAc.

FIGURE 7 shows a frozen section of human pancreas (X300) incubated with CAC1 and CAC2, demonstrating a strongly positive immunoperoxidase staining of acinar cell membrane and cytoplasm.with no reactivity seen with islet cells. 1_.1

FIGURE 8a shows dispersed human pancreas cells demonstrating an isolated human islet (arrow).

FIGURE 8b shows FITC-labeled cic-2 which clearly identifies the islet without fluorescent activity within the concomitant acinar cells epifluorescence.

Figures 9A and 9B are of the same specimen without moving the stage.

FIGURE 9A is a phase contrast view of an isolated islet enveloped by dispersed acinar cells.

FIGURE 9B shows that when viewed under fluorescent light, these dispersed acinar cells bound, with FITC-labeled McAb clearly identified.

FIGURE 9C shows that the addition of complement to the slide of acinar and cells treated with anti-acinar McAB resulted in specific acinar cell lysis, as demonstrated by trypan blue uptake in the dead acinar cells.

FIGURE 10A shows a dispersed pancreatic cell population after collagenase digestion, demonstrating contaminant acinar cells associated with free islets.

FIGURE 10B demonstrates the obvious acinar cell lysis after McAb treatment with intact viable islets.

FIGURE 10C shows a purified canine islet population. DETAILED DESCRIPTION OF THE INVENTION

ANTI-ISLET MONOCLONAL ANTIBODIES

Monoclonal antibody purification. Four to six-week old Balb/c mice were immunized with a cerivcal squamous cell carcinoma cell line (CIC-2) and the membrane protein of a lung adenocarcinoma cell line (ClC-1) by intraperitoneal injection. Three days after the last booster injection the immunized spleen cells were fused with myeloma P3-X63 AG 8.653 by a modified procedure of Kohler and Milstein (1975) Nature 256:495-497. Two weeks after the fusion the hybrid cultures were selected by microcytotoxicity, enzyme-linked immunosorbent assay testing and immunocytochemical assays. Positive hybridomas were cloned by limited dilution followed by passage into Balb/c mice for ascites production.

Immunocytochemical analysis. To characterize specificity of islet cell binding by CIC-2 and CIC-1 , immunocytochemical studies were performed using frozen sections of human adult pancreas. Monoclonal antibody CAC-2, which we determined has a specificity for human acinar but not islet cells, was used as a control. Indirect immunoperoxidase staining was performed on 4-6 μm sections of frozen tissue by incubation with 50 μl of the primary antibody solution at ambient temperature for 30 min, in a humidity chamber. After washings in three changes of phosphate-buffered saline the sections were incubated with peroxidase-conjugated goat anti-mouse IgG (KPL, Gaithersburg, MD). Negative controls include γG3κ mouse myeloma immunoglobulin, nonreactive ascites harvested from Balb/c mice innoculated with 653.1 mouse myeloma cells, and primary antibody diluent.

Monoclonal antibody isotype. Isotyping of the monoclonal antibodies was performed by immunoperoxidase staining using immunoglobulin subclass- reactive antibodies. Both monoclonal antibodies were of K light chain specificity, with CIC-1 having γG2_a and CIC-2 γG3_D heavy chains. Isolation of viable islet cells and indirect fluorescent staining with fluorochrome-labeled CIC-1 and CIC-2. To determine if these islet-specific monoclonal antibodies would continue to bind isolated viable islet cells following the collagenase digestion process, the binding ability of fluoroscein isothiocyante- or rhodamine- labeled CIC-1 or CIC-2 to freshly isolated adult islets was examined. Four human pancreata were subjected to collagenase digestion using a modification of the method described by Gray et al. Briefly, the pancreatic duct was cannulated with a 22-guage cannula and the gland distended with 4^° C Hank's solution containing 10 mM Hepes, 1.3 mM CaCl2- The pancreas was then immersed in a beaker with 250 ml Hank's solution containing 0.2 mg/ml Type XI collagenase, pH 7.4, at 37°C and continuously perfused with this solution at 0.2 ml/min/g. The endpoint of a digestion was considered reached when the tissue appeared to dissolve to a fine texture. The gland was then cooled in 4°C Hank's solution and gently torn apart with forceps. The fluid with the dispersed pancreas tissue was filtered through a 500-μm sieve and the resultant sediment centrifuged. The tissue pellet thus obtained was examined for islets under an inverted microscope. Islet cell viability was confirmed by trypan blue staining.

Islets of about 30 - 50% purity were incubated with monoclonal antibodies CIC-1 and CIC-2 ascites at their appropriate dilution of 1 :1000 to 1 :2000 for 30 min at room temperature, under mild agitation. After three washings in Hank's solution the cells were reacted with secondary fluoroscein isothiocyante goat anti- mouse IgG for 30 min, followed by three more washings and resuspension in culture medium. The cell preparation was then examined under a Zeiss universal microscope with epifluorescence.

Effects of chemical or enzyme treatment on antiαenic determinants. Nayak et al, (1985) Diabetes 34: 617-619. have shown that positive autoislet cell antibody from human sera recognize an antigenic determinant on human islets that is a sialic acid containing glycolipid (ganglioside). The binding of these monoclonal antibodies was studied after the following biochemical treatment of human pancreas sections: (a) periodic acid; (b) neuraminidase; (c) pronase + trypsin digestion; and (d) extraction with ehoiroform:methanol.

Immunocytochemical testing. Frozen sections of nine adult human pancreata, of known blood group, demonstrated specific islet cell binding by CIC-1 and CIC-2, with no acinar or ductal cell activity (Fig. 1). A control monoclonal antibody (CAC-1 ), known to be acinar cell but not islet cell-reactive, was studied in these same specimens and is shown in Fig. 2.

Monoclonal antibody binding to isolated viable adult human islet cells. Having determined that CIC-1 and CIC-2 bound specifically to human islets in frozen sections of pancreas tissue the next step was to demonstrate the continued ability of these monoclonal antibodies to bind to isolated viable islets following the collagenase digestion process. Figure 3 demonstrates an isolated human islet amongst separated acinar cells or acinar-cell clumps, as viewed under a Zeiss universal microscope without epifluorescent illumination. Fluoroscein isothiocyante-conjugated CIC-2 clearly labels the islet, as seen in Figure 4, indicating that the antigenic binding sites are not denatured by the collagenase digestion process. The preservation of the CIC-1 and CIC-2 antigenic determinants on the islet cell membrane following the collagenase digestion process was established by immunofiuorescent staining of freshly-obtained islets.

Antigenic determinants. Carbohydrate residues of glycosphinogolipids (glycolipids) and glycoproteins are very sensitive to periodate oxidation. Neuraminidase treatment abolishes binding of the antibodies to the antigen, should a sialic acid be a component of the antigen determinant site. Binding of antibody to islet cells was not abolished by a periodate treatment in the case of CIC-2 nor by neuraminidase treatment in both CIC-1 and CIC-2 (Table 1), demonstrating that, unlike the autoantigen described by Nayak et al., supra, the human islet antigenic determinant recognized by CIC-1 and CiC-2 does not contain sialic acid. Treatment of frozen sections with chloroform:methanol and proriase-trypsin digestion allowed characterization as to the anti-lipid or anti-protein nature of the antigenic determinant. Binding was abolished by chloroform :methanol extraction, but was not affected by pronase or trypsin digestion, suggesting that CIC-1 and CIC-2 recognize an antigen that is glycolipid in nature.

Thus, biochemical perturbations of pancreas frozen sections suggest that the antigenic determinants of CIC-1 and CIC-2 do not contain sialic acid. The sensitivity to chloroform:methanol extraction and inability to affect antibody binding by pronase or trypsin enzyme digestion suggests that antigenic determinants of CIC- 1 and CIC-2 are not protein, but lipid, in nature. The epitopes recognized by CIC-1 and CIC-2 are different as, observed by their response to periodate treatment, thus potentially allowing both fluorchrome-labeled monoclonal antibodies to be used simultaneously as islet identification reagents.

FACS Sorting of Antibodv-labelled Cells: CIC-1 and CIC-2 are ideal reagents for utilization in an automated islet purification apparatus, the processes for which are well-known in the art. Existing flow cytometery hardware known in the art is designed for the study of lymphocytes and smaller cells and requires modifications to accommodate the larger islets (100 - 300 μm). Thus, CIC-1 and CIC-2 are important reagents in the invented islet cell purification by an automated cell sorting process.

In summary, two monoclonal antibodies (CIC-1 and CIC-2) have been identified and characterized , which are specific for human islet but not acinar cells. The antigen these antibodies recognize is lipid in nature and, thus, unaffected by collagenase digestion of the pancreas. These properties allow isolated human islet cells to be identified by fluoroscein isothiocyante-labeled CIC-1 and CIC-2, making these antibodies useful reagents in islet cell identification.

ANTI-ACINAR MONOCLONAL ANTIBODIES

Identification of Anti-Acinar Cell Monoclonal Antibodies

Immunoperioxidase staining of frozen human pancreatic tissue was performed using a large battery of monoclonal antibodies (obtained after fusion of spleen cells from Balb/c mice previously immunized with purified membrane proteins of lung, gastric, and colon adenocarcinoma, with myeloma cell line P3-X63.AG 8- 653). Several monoclonal antibodies (CAC-1 , CAC-2 and CAC-3) were identified that bound the human acinar and ductal but not islet cells. In addition, these monoclonal antibodies were found to be ABH and Lewis blood group reactive.

Interspecies crossre activity between these monoclonal antibodies and canine acinar cells were confirmed by immunoperoxidase staining of frozen canine pancreas sections. Negative controls consisted of nonreactive ascites of Baib/c mice innoculated with 653.1 mouse myeloma cells and primary antibody diluent. The cytotoxic nature of these monoclonal antibodies to isolated human and acinar cells was confirmed by trypan blue staining of acinar cells following the addition of rabbit complement at a 1 :20 dilution.

Canine Islet Cell Isolation. Twenty adult mongrel dogs, weighing 15 to 20 kg, were used. The pancreas was processed by the ductai perfusion method as described by Haraguchi et al., with minor modifications. Briefly, the gland was placed into a tempering beaker maintained at 37^°C and perfused at 10 mlJmin via the ductai cannula, in a recirculating pump chamber. The perfusate consisted of HBSS, 10 mmol/L HEPES and type XI collagenase (0.8 mg/mL) (Sigma Chemical Co., St Louis) at 37^°C. The digestion was considered complete when the pancreas developed a sandy, mucoid texture. The tissue was then stripped form the ductai tissue into ice-cold HBSS, gently separated with forceps, filtered through a 500-μ pore stainless steel sieve and washed three times in HBSS by centrifugation. The digestate thus obtained was divided into two equal pellets: (1 ) untreated control and (2) anti-acinar cell monoclonal antibody plus complement treatment and then submitted to Percoll density gradient separation.

Anti-Acinar Monoclonal Antibody Treatment. The digestate was separated into two equal 1 -mL pellets in 100 mL Hank's solution containing 2% fetal calf serum and aprotonin, 100 KIU/mL. One pellet was treated with anti-acinar monoclonal antibodies obtained from mouse ascites ammonium sulfate precipitates at a protein concentration of 5 to 10 μg/mL. Following a 30-min monoclonal antibody incubation, rabbit complement at a 1 :20 dilution was added to the washed and resuspended digestate. During an incubation period of one hour, the cell suspension was washed every 15 minutes by centrifugation and addition of fresh complement-containing Hank's solution. The incubating digestate was maintained in a spinner flask under mild agitation at room temperature. The control digestate pellet was exposed to the same number of washings as the treated digestate above, except without the addition of monoclonal antibodies and complement.

Both control and treated digestate were then subjected to Percoll gradient separation and thereafter examined under a Nikon inverted Diaphot microscope at 100 x and 200 x for determination of islet cell yield and purity.

Static Glucose Stimulation, To determine whether islet cell function was affected following monoclonal antibody treatment, islets were handpicked from the treated digestate for static glucose stimulation. Insulin release in response to 100 and 300 mg% glucose, was determined by RIA assay and the insulin concentration computed as μU/mL/isiet.

Cytotoxic Anti-Aciπar Cell Monoclonal Antibodies.

Immunoperoxidase staining of nine human pancreatic sections confirmed that acinar but not islet cells express blood group antigens. Cross- reactivities of several of these anti-acinar cell monoclonal antibodies were found to occur in canine pancreas tissue. Following collagenase digestion of the canine pancreas, continued binding of these anti-acinar cell monoclonal antibodies to isolated acinar cells was demonstrated by immunofluorescence studies using secondary flourescein isothiocyanate-labeled goat antimouse antibodies.

Lysis of acinar cell membrane occurred after addition of complement to the monoclonal antibody, resulting in 85% to 90% acinar ceil death. Islet ceil viability, on the other hand, was not affected, as demonstrated by trypan blue staining and positive glucose stimulated insulin release from a basal of 3.1 ± 1.5 to 9.4 + 2.1 μU/mL islet following 300 mg% glucose stimulation.

Islet Cell Purification. As shown in Table 2, density gradient separation of the digestate alone resulted in poor purification, obtaining preparations in the range of 9% to 12% purity. In contrast, monoclonal antibody treatment resulted in a significant improvement, achieving islet cell preparations of 65% to 87% purity. Islet cell yield, however, waε significantly impaired by monoclonal antibody treatment, reducing the average yield from the control digestate of 40,000 to 60,000 islets/pancreas by 40% to 45%.

SU MMARY

Human acinar cells but not islets express ABH and Lewis blood group antigens. This difference in antigenic expression continued after collagenase digestion, allowing the development of an immunologic method of islet cell purification using cytotoxic anti-acinar cell monoclonic antibodies. In the canine model, islet cell preparations of 65% to 87% purity have been achieved using this technique. Thus, the potential exists that anti-acinar cell monoclonal antibodies may be applied to large-scale islet cell purification and may play a role in overcoming the technical obstacles preventing successful clinical islet cell transplantation.

Demonstration That Human Acinar Cells, but Not Islets. Express Blood Group Antigens.

Evidence exists that the acinar ceils of the human pancreas express blood group-reactive antigens. In 1981 , Ronger et al., Tissue Antigens. (1981) 18:51 demonstrated, by immunofiuorescent studies using blood group anti-sera, the expression of ABH and Lewis antigens in centra acinar cells. Using lectin- horseradish peroxidase conjugates, Itoh et al. , J. Histochem Cvto 25:81 (1987) confirmed the localization of blood group antigens in the human pancreas.

Immunoperoxidase study. Based on this evidence, frozen sections from nine human pancreata were studied (obtained from donors of known blood type) for blood group expression, using anti-ABH McAb. The indirect immunoperoxidase staining of frozen sections was performed. Briefly, the methods employed were as follows: Frozen sections of human pancreas of 6 μm thickness were fixed in 4% buffered formalin and incubated at room temperature with 50 λ of primary McAb after three washes, diluted with 2.5% BSA-containing PBS at pH 7.2. Following three washes, the sections were then incubated for 30 min with peroxidase-conjugated goat anti-mouse IgG or IgM (KPL, Gaithersburg, MD) at a dilution of 1 :150 with 3% normal goat serum-containing PBS. Negative controls consisted of nonreactive mouse ascites of Balb/c mice innoculated with 653.1 mouse myeloma cells, IgM mouse myeloma immunoglobulin (Litton Bionetics, Charleston, SC) and primary antibody diluent. Peroxidase staining was developed by incubating the sections in 3-amino-9-ethyl carbazole (0.02% W/V) with 2% DMSO in 0.02 M acetate buffer, pH 5.1 , in the presence of 0.03% H2O₂, for 4 min. The sections were then counterstained with Harris hemotoxylin and mounted in 90% glycerol-PBS. Evaluation of the acinar cell and islet reactivity was then performed using a Zeiss universal microscope. The results are shown in Figure 5 and Table 3, below, as follows.

Analysis of data. These studies confirm that human acinar cells, but not islets, express blood group antigens. No exception (other than anti-H McAb) was encountered to the correspondence between blood type of the donor pancreas and blood group McAb. This confirms the report by Itoh, who demonstrated similar correspondence, using lectin-horseradish peroxidase conjugates.

Of interest was the finding that anti-H McAb reacted with acinar cells form donors with blood groups A, B, AB and O. This finding can be explained when one examines the genetics of ABH antigens. The blood group antigens are elaborated in a stepwise fashion by a series of glycosyl transferase enzymes, whose formation is determined by A, B and H genes, respectively. The H antigen is a precursor for A and B antigens, and the immediate precursor for H antigen is Gal- GlcNAc, as shown in Figure 6. Thus, based on the finding that anti-H McAb binds to pancreas tissue from donors of all blood groups, and that α⁽¹'²) fucose and Gal- GlcNAc are terminal oligosaccharide residues common to each blood group (Figure 6), the antigenic determinants recognized by the McAb on the acinar cell surface or cytoplasm is either ^{(1 -2)} fucose or Gal-GlcNAc.

Identification of Blood Group-reactive McAb that Bind Human Acinar Cells but Not Islets

Exploiting the differences in antigenic expression in acinar cells described above, a number of mouse McAb derived from immunization with human tissue of pulmonary, gastric and colon adenocarcϊnomas were examined at various stages of differentiation. The basis for examining McAb derived from tumor sources was the literature reporting the expression of blood group antigens in human adenocarcinomas. Furthermore, blocked synthesis of ABH antigens in a large variety of human cancers is associated with accumulation of the precursor carbohydrate, Gal-GIcNAc. Feizi , et al., Lancet 2:391 (1975), described blood group precursor accumulation in two cases of gastric cancer and Fukishi , et al. J,. Biol. Chem.. 259:4681 (1984), reported on novel glycolipids accumulating in human adenocarcinoma. As discussed above, it is possible that the precursor substance, Gal-GIcNAc, is one of the epitopes recognized by anti-acinar McAb. With such alterations of blood group antigens and accumulation of complex oligosaccharide end-structures, immunization of mice with human tumor tissue expressing these antigens, may result in reagents that recognize a large number of epitopes on the acinar cell membrane; indeed, this may account for the wide cross-reactivity of such McAb (see cross-reactivity and specificity discussed below).

Monoclonal antibody production Briefly, four to six-week-old

Balb/c mice were immunized by s.c. injection with 0.5 mg of purified membrane protein form pulmonary, gastric or colon human adenocarcinoma cell lines or tissues at various stages of differentiation. Three days after the last booster injection, spleen cells were fused with mouse myeloma P3.X63.Ag8-653 using a modified protocol of Kohler and Milstein, supra. Two weeks after fusion the hybridoma supematants were tested for antibody production by enzyme-linked immunosorbent assay (ELISA) and by complement-mediated cytotoxicity, using the microdroplet cytotoxicity test described by Terasaki., et al., Am J. Clin Pathol 69:103 (1978) Clones of the hybridomas with the desired reactivity were obtained by limiting dilution, followed by passage into Balb/c mice for ascites production.

Immunohistochemical identification of McAb with acinar cell binding. Indirect immunoperoxidase staining was performed to determine antigenic reactivity of the mouse ascites thus produced. Details of the staining method have been described above. Staining intensity was graded as: - (negative), + (weakly positive), + (positive) and ^•++ (strongly positive). A reaction, as observed by immunoperoxidase staining, was regarded positive when McAb binding occurred in at least 70% of the acinar cell population, with staining intensity of at least + or ++. Two McAb (derived from lung and stomach human adenocarcinoma) were thus identified with ++ reactivity to human acinar and ductai cells, but not islets, and were designated: CAC1 and CAC2 (Figure 7). Binding of these McAb was observed to occur both at the cell membrane, as well as within the cytoplasm of the acinar cell.

3. Preliminary Characterization of Anti-Acinar McAb .CAC1 and CAC2.

Studies were performed in our laboratory to determine: (a) immunoglobulin class of CAC1 and CAC2 and (b) specificity and cross-reactivity.

Antibody Class. Knowledge of the antibody class and subclass is of importance in determining the strategy of purification. To determine the class of CAC1 and CAC2, an Ouchterlony test (formation of pecipitin lines in agar) was performed. The source of CAC1 and CAC2 was from culture surpematant of hybridoma cells obtained after cloning, to avoid the ambiguous results that can occur from serum or ascites, due to the concurrent presence of normal mouse immunoglobulin. The results indicate that both CAC1 and CAC2 are of the IgM class.

Specificity and Cross Reactivity

"Multiple tissue-reactive" McAb. CAC1 and CAC2 were found to be multiple tissue-reactive on immunocytochemical analysis of frozen sections of various human tissues (Table 4, below). It is well-established that blood group antigens are expressed by many different tissues in the body other than red blood cells. Thus, given the fact that CAC1 and CAC2 recognize blood group antigens, it is not surprising to find that these reagents are so-called "multiple tissue-reactive" McAb. Within the pancreas itself, however, CAC1 and CAC2 are specific in terms of their recognition of acinar cells, with no binding to islets.

Interspecies cross-reactivitv. Immunocytochemical analysis of frozen sections of canine and rat pancreas tissue was performed to assess for interspecies cross-reactivity. CAC1 was found to bind to canine acinar cells, but not islets, while CAC2 reacted with both rat and canine acinar tissue. Again, an explanation for this broad cross-reactivity is found by examining the genetics of ABH antigens. A review on this subject by Oriol, Vox Sano 51 :161 (1986) reported that ABH antigens are expressed in tissue of rat, rabbit, marmoset, baboon and man. Furthermore, Reibel et al., Cell Tiss. Res 237:11 (1984) demonstrated cross-reactivity of rodent and mice antigens with antibodies to human blood group antigens A, B and precursor substance. Thus, these findings strengthen the hypothesis that CAC1 and CAC2 are blood group-reactive McAb that recognize a blood group precursor (possibly αC^1-2) fucose or Gal-GIcNAc) antigen expressed on acinar cells, which is present in several species, including rat, dog and man. While broadly-reacting McAb are not desirable if they are to be applied as specific diagnostic tools, for the purposes of cell separation, this broader specificity of CAC1 and CAC2 may indeed be advantageous.

4. Preliminary Analysis of the Nature of the Acinar Cell Antigen

Lectins are sugar-binding protein (glycoproteins) of nonimmune origin. Thus, using lectins with specific binding properties of known monosaccharides as blocking agents, one could deduce the carbohydrate components of the cell membrane molecule recognized by the McAb. Lectins, therefore, provide a range of readily available and well-defined reagents for the identification of cell surface antigens.

Experiments. Ulex-europaeus type 1 agglutinin is a lectin with specific binding to the saccharide, -L-fucose. To determine whether human acinar cell membrane expresses this saccharide on it cell surface and whether CAC1 and CAC2 binds to the acinar cell by recognition of α-L-fucose, two-stage immunoperoxidase binding and blocking studies of frozen sections of Type O human pancreas were performed as follows: Frozen sections of blood group O human pancreas as fixed in 4% buffered formalin for 2 min and after three washes in PBS were treated at room temperature for 20 min with the following reagents: (a) unlabeled UEA-1 alone (200 μg/ml - negative control); (b) CAC1 and CAC2 alone (1/200, 1/400, 1/800 dilution); (c) UEA-1 treatment followed by CAC1 and CAC2; and (d) horseradish peroxidase labeled-UEA-1 was used to assess direct staining method to demonstrate UEA-1 binding to acinar cells. Immunoperoxidase labelling with secondary antibody was then accomplished as previously described.

Results, (a) UEA-1 treatment alone showed a positive reaction, indicating the expression of α-L-fucose.antigens on human acinar cells and (b) blocking of CAC1 and CAC2 binding by UEA-=1 did not occur, as shown in Table 5, below.

Analysis of data. The inability of UEA-1 to block the binding of CAC1 and CAC2 indicates that the oligosaccharide recognized by these reagents are of a much broader specificity than that of α-L-fucose. alone. In the context of application in cell separation, this is, perhaps, advantageous.

5. Demonstration of Complement-dependent Cvtotoxicitv of CAC1 and CAC2 to Human Acinar Cells, but not Islets

To examine the ability of these McAb to bind complement and selectively lyse human acinar cells, we studied the effects of CAC1 and CAC2 on collagenase-digested human pancreatic tissue WERE STUDIED.

Islet isolation from human pancreatic tissue. Human pancreas tissue was obtained with full consent from kidney cadaver donors. Pilot studies to learn and reproduce the published methods of human islet isolation were performed in approximately 14 human pancreata. The results described here are from the successful isolation of dispersed islets form four pancreata. A method to obtain isolated human islets in large quantities is part of the subject invention and is discussed more fully below.

The islets from four human pancreata were prepared as follows: The gland was dissected free from adjacent tissue and the pancreatic duct cannulated with 22-gauge catheter. The gland was distended with 4°C Hank's solution containing 10 mM Hepes, 1.3 mM CaCl2, and then immediately immersed in a beaker with 150 ml prepared (37°C) Hank's solution-containing 0.6 mg/ml Sigma Type 1S collagenase (Sigma Chemical Co., St. Louis, MA). The pancreas was then continuously perfused with this solution at pH 7.4 at 37°C at 0.2 ml/min/g. The exact endpoint of the digestion process was judged to have been reached by teasing apart a small area of the gland using fine forceps, or when the pancreatic tissue appeared to dissolve to a fine, smooth sand-like texture. The gland was then rapidly cooled by immersion in 4°C Hank's solution cut into several pieces, torn apart using fine forceps, shaving constantly in cold Hank's solution. The digestate thus obtained was passed through a mesh filter (pore size 500 μm) and the filtrate centrifuged in 250-ml centrifuge tubes at 250 xg for 30 sec. The pellet was washed twice more in cold Hank's solution and placed in a spinner flask until McAb treatment. The residual tissue that did not pass through the filter was mechanically dissociated by shaking in a cold-water bath for 10 min and re-filtered. The filtrate was then treated in a manner similar to that described above.

Treatment of digestate with McAb and complement. A 1-ml pellet was suspended in 100 ml Hank's solution contianing 2% FCS and 100 KlU/ml aprotinin. CAC1 and CAC2 (as ascites ammonium sulfate precipitate) was added to the pellet for a final concentration of t to 10 μm/ml. After 30 min incubation in a spinner flask at mild agitation, rabbit complement was added to the washed and resuspended pellet at a 1 :20 dilution. During the incubation period of 1 hr, the cell suspension was washed every 15 min by pelleting and addition of fresh complement-containing Hank's solution. The entire incubation was maintained in a spinner flask at room temperature. As control, a second pellet suspension was treated in exactly the same manner as above, except for the omission of McAb or complement. Both treated and untreated suspensions were then evaluated for: (a) islet identification; (b) binding of CAC1 and CAC2 to acinar cells even after collagenase digestion; (c) complement-mediated acinar cell lysis; and (d) islet cell viability of the McAb- treated pellet as follows:

Islet identification. Islets were identified by visual inspection. In the early phases of the study, to confirm that the human islets were recognized, the use of a vital dye, dithizone, which colored human islets, but not acinar cells, red, was employed. Furthermore, in pilot studies, a FITC-labeled McAb, CIC-2 (which bound only to islets) was used to assist in the identification of dispersed islets. Using a microscope with fluorescent capabilities, FITC-labeled CIC-2 clearly identified isolated human islets (Figs. 8A and SB), but not acinar cells.

Demonstration that collagenase digestion does not denature the acinar cell antigen recognized bv CAC1 and CAC2. To confirm that the acinar antigenic determinants were not denatured by the collagenase digestion process of the pancreas, the dispersed acinar cells were examined by immunofiuorescent labelling. Viable acinar cells were incubated with CAC1 and CAC2 after collagenase digestion. Following washing, the McAb were labeled with FiTC- conjugated goat anti-mouse IgM. Fluorescent staining was evaluated under epifluorescence using a Zeiss universal microscope. Figure 9A demonstrates an isolated islet enveloped by attached acinar cells under phase microscopy. When viewed under epifluorescence, the FITC-labeled McAb binding to the dispersed acinar cells become readily visible (Figure 9B), thus demonstrating that the antigenic determinants recognized by CAC1 and CAC2 are not denatured by the collagenase digestion process. Complement-mediated cell vsis. Using a combination of the immunofiuorescent technique described above and trypan blue uptake to assess cell viability, after the addition of complement, FITC-labeled acinar cells were lysed, as evidenced by uptake of trypan blue (Fig. 9c)

Islet viability. Islet viability was assessed by in vitro perifusion studies. Following McAb treatment, a 200 μl pellet of cell suspension was subjected to glucose perifusion studies. Islet viability was confirmed, as evidenced by significant insulin secretion in response to 400 mg% glucose.

6. Application of CAC1 and CAC2 in Canine Islet Purification

Twenty adult mongrel dogs were used in this study. The pancreas was placed into a tempering beaker maintained at 37°C and perfused at 10 ml/min via the ductai cannula in a recirculating pump chamber. The perfusate consisted of Hank's balanced salt solution, 10 mM hepes and 0.8 mg/ml Sigma Type XI collagenase. The digested tissue was transferred into a container with 4°C HBSS by centrifugation. The digestate thus obtained was divided into two equal pellets: (a) untreated control and (b) anti-acinar cell McAb plus complement-treated, then submitted to density gradient separation.

The following results were obtained:

(a) lysis of acinar cell membrane occurred after addition of complement to McAb resulting in 85-90% acinar cell death (Figures 10A and 10B).

(b) islet viability, on the other hand, was not affected, as demonstrated by trypan blue staining and positive glucose stimulation release from 3.1 ±5.1 to 9.4+2.1 μU/ml/islet following 300 mg% glucose. (c) density gradient separation resulted in poor purification, obtaining preparations in the range of 9-12% purity. In contrast, the McAb-treated digestate resulted in significant improvement in islet preparation, achieving 65-87% purity (Fig 10c).

(d) islet yield however, was significantly impaired by McAb- treatment, reducing the average yield from the control digestate of 40-60,000 islets/pancreas by 40-45%.

The present invention also includes a purification system for islet cells comprising the use of both anti-islet and anti-acinar monoclonal antibodies, as set forth above, to obtain a very pure islet population. It will be obvious to a person of ordinary skill in the art that many variations of the above-described invention may be utilized and still come within the spirit and scope of the present invention.

Claims

C LAI MS

1. A monoclonal antibody having a specificity for human islet cells.

2. The monoclonal antibody of Claim 1 wherein said monoclonal antibody is produced by a fused murine hybrid cell derived from a spleen cell from a mouse immunized with a cervical squamous cell carcinoma.

3. The monoclonal antibody of Claim 1 wherein said monoclonal antibody is produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with a lung adinocarcinoma.

4. The monoclonal antibody of Claim 1 , wherein said monoclonal antibodies are produced by a murine-derived hybrid cell line, said antibodies being capable of specifically binding to an antigenic determinant of human islet cells.

5. The monoclonal antibody of Claim 4 wherein said monoclonal antibodies are capable of specifically binding to an antigenic determinant on human islet cells, but not human acinar cells.

6. A method of making monoclonal antibodies specific for human islet cells, but not acinar ceils, comprising the steps of:

fusing an mouse spleen cells immunized with a human adenocarcinoma to a mouse myeloma cell line to form fused cells;

selecting said fused cells for anti-islet antibody production;

cloning said selected anti-acinar antibody producing hybridoma ceils.

7. A method of selectively isolating and identifying islet ceils from a digested mixture of pancreatic tissue, comprising the steps of:

providing a digested mixture of pancreatic tissue;

binding monoclonal antibodies specific for human islet cells to an immobilized support;

contacting the digested pancreatic tissue with the monoclonal antibody bound to a support under conditions which permit binding of said monoclonal antibody to said islet cells;

removing the remaining digested pancreatic material from the immobilized antibody - islet cell composite; and

removing the islet cells from the monoclonal antibody.

8. The method of Claim 7, wherein said monoclonal antibody is derived produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with a cervical squamous cell carcinoma.

9. The method of Claim 7, wherein said monoclonal antibody is derived produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with a lung adinocarcinoma.

10. The method of Claim 7, wherein said monoclonal antibody is capable of specifically binding to an antigenic determinant of human islet cells.

11. The method of Claim 10 wherein said monoclonal antibodies are capable of specifically binding to an antigenic determinant on human islet cells, but not human acinar ceils.

12. The method of Claim 7 wherein said antibody-islet composite is removed by means of a cell sorter capable of selecting cells labelled with the flourescent -labelled antibody.

13. The method of Claim 7 wherein said antibody-islet composite is removed by the steps of:

fusing said monoclonal antibodies to a support;.

contacting said pancreatic digestate with said support bound monoclonal antibodies to allow said islet cells to adhere to said support bound monoclonal antibodies;

separating the islet cell/monoclonal antibody/support from the remainder of said pancreatic digestate.

14. The method of Claim 13 wherein said support is selected from a magnetic and non-magnetic support.

15. The method of Claim 13 wherein said support is a magnetic material and said support bound islet cells are separated from said digestate material by application of a magnetic field to said digestate to hold said islet cells in place and allowing said islet cells to be freely removed.

16. The method of Claim 13 wherein said support is an immobilized support.

17. The method of isolating human islet cells from a digested human pancreatic mixture comprising the steps of:

providing monoclonal antibodies specific for human islet cells;]

conjugating with said monoclonal antibody an immunofluroescent molecule to form an immunofiuorescent monoclonal antibody;

contacting said immunofiuorescent monoclonal antibody with a digested pancreatic mixture containing said islet cells to form an immunofiuorescent islet cell; and

flowing said pancreatic mixture containing said immunofiuorescent labelled islet cells through a fluorescent activated cell sorter to separate the fluorescing islet cells from the remainder of the digested pancreatic material.

18. A method of detecting human islet cells comprising the steps of:

providing monoclonal antibodies specific for human islet cells;]

conjugating with said monoclonal antibody an immunofluroescent molecule to label said antibody;

contacting said immunofiuorescent labelled monoclonal antibody with a biological fluid containing said islet cells; and

examining said fluid for immunofiuorescent labelled islet cells.

19. A monoclonal antibody having a specificity for acinar cells.

20. The monoclonal antibody of Claim 19 wherein said monoclonal antibody is produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with a pulmonary adenocarcinoma.

21. The monoclonal antibody of Claim 19 wherein said monoclonal antibody is produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with gastric adenocarcinoma.

22. A monoclonal antibody produced by a murine-derived hybrid cell line wherein the antibodies are capable of specifically bonding to an antigenic determinant of acinar cells.

23. The monoclonal antibody of Claims 19 and 22 wherein said antibody is capable of specifically binding to an antigenic determinant on acinar cells selected from human, rate, canine and monkey acinar cells, but not human islet cells from the corresponding animals.

24. A method of making monoclonal antibodies specific for acinar cells, but t cells, comprising the steps of:

providing human carcinoma cells expressive of A, B, H and Lewis blood group antigen;

innoculating a receptor animal with said carcinoma cells under conditions which permit the spleen cells of said receptor animal to produce anti-ABH and Lerwis blood group antibodies, and extracting said spleen;

providing murine myeloma cells;

fusing cells from said spleen with murine myeloma cells;

selecting and isolating said fused cells for anti-acinar antibody production; and

collecting said anti-acinar antibodies produced by said selected fused cells.

25. A method of selectively isolating islet cells from a digested mixture of pancreatic tissue, comprising the steps of:

providing a digestive mixture of pancreatic tissue;

binding monoclonal antibodies specific for acinar cells to a support;

contacting the digested pancreatic tissue with the monoclonal antibody bound to the support under conditions which permit binding of said monoclonal antibody to said acinar cells;

removing the remaining digested pancreatic material from the immobilized antibody - acinar cell composite; and

removing the islet cells from the remaining digested pancreatic material.

26. The method of Claim 25, wherein said monoclonal antibody is derived produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with a gastric adenocarcinoma.

27. The method of Claim 25, wherein said monoclonal antibody is derived produced by a fused hybrid cell derived from a spleen cell from a mouse immunized with a pulmonary adenocarcinoma.

28. The method of Claim 25 wherein said monoclonal antibody is produced by a murine-derived hybrid cell line wherein the antibodies are capable of specifically bonding to an antigenic determinant of acinar cells.

29. The method of Claims 25 through 28 wherein said monoclonal antibody is capable of specifically binding to an antigenic determinant on acinar cells selected from human, rate, canine and monkey acinar cells, but not islet cells from the corresponding animals.

30. . The method of Claim 25 wherein said antibody-acinar composite is removed by means of a cell sorter capable of selecting cells labelled with the flourescent -labelled antibody.

31. The method of Claim 25 wherein said antibody-acinar composite is removed by the steps of:

fusing said monoclonal antibodies to a support;.

contacting said pancreatic digestate with said support bound monoclonal antibodies to allow the acinar cells contained therein to adhere to said support bound monoclonal antibodies;

separating the acinar cell/monoclonal antibody/support from the r emainder of said pancreatic digestate to obtian the remainder including islet cells.

32. The method of Claim 31 wherein said support is selected from a magnetic and non-magnetic support.

33. The method of Claim 31 wherein said support is a magnetic material and said support bound acinar cells are separated from said digestate material by application of a magnetic field to said digestate to hold said acinar cells in place and allowing said acinar cells to be freely removed.

34. The method of Claim 33 wherein said support is an immobilized support.

35. A method of diagnosing pancreatic transplant rejection comprising:

obtaining a sample of biological fluid suspected of containing acinar cells;

contacting said sample of biological fluid with antibodies selected from anti-acinar antibodies and anti-islet antibodies, said antibodies being labelled with a marker for detection;

examining said urine or biological fluid sample for the presence of cells having bound thereto antibodies comprising said detection marker.

36. A method of selectively isolating islet cells from a digested mixture of pancreatic tissue, comprising the steps of:

providing a digestive mixture of pancreatic tissue;

binding lectins which recognize A, B, and H antigens specific for acinar cells to a support;

contacting the digested pancreatic tissue with the lectins bound to the support under conditions which permit binding of said monoclonal antibody t said acinar ceils;

removing the remaining digested pancreatic material from the immobilized lectin - acinar cell composite; and

removing the islet cells from the remaining digested pancreatic material.

37. The method of Claim 36 wherein said lectin is UEA-1.

38. Islet cells obtained by the methods of any of Claims 7 through 17, 25 through 28 or 30 through 34.