We have reproduced a previously described method for the in vitro generation of endocrine cells in adult human pancreatic tissue culture. The aim of this study was to characterize the nature of pancreatic progenitor cells and to identify the factors necessary for their differentiation in this model. During monolayer expansion, two types of cells proliferated sequentially; first cytokeratin 19 (CK19)-positive ductal epithelial cells and then nestin-positive fibroblastoid cells. After the bromodeoxyuridine-labeled cells were traced in differentiated islet buds, some of the proliferating ductal cells had differentiated into endocrine cells, whereas nestin-positive cells could not give rise to endocrine tissue. Serum-free culture was found to be an absolute requirement for the endocrine differentiation to occur. Also, overlay of the cells with Matrigel was essential, whereas nicotinamide had a potentiating effect. The in vitro–generated islet buds released insulin in response to glucose nearly as efficiently as native islets. When transplanted under the kidney capsule of nude mice, only one of five grafts demonstrated further growth with foci of both endocrine and exocrine differentiation. Our results support the previous notion that pancreatic progenitor cells represent a subpopulation of ductal epithelial cells. No evidence was found for the development of endocrine cells from nestin-positive stem cells.

Transplantation of isolated islets from cadaver pancreas is a promising possibility for the optimal treatment of type 1 diabetes (1). However, such an approach is severely limited by the shortage of donor organs. This problem could be overcome if it became possible to generate transplantable islets from stem cells (2). In addition to embryonic sources, stem cells have been identified in many adult tissues and even pluripotent stem cells may be found in adult bone marrow (3). It is at this point not clear that it will become possible to generate fully functional β-cells from primitive stem cells. One attractive approach for the generation of β-cells involves expansion and differentiation of adult human pancreatic progenitor cells, which are closely related with the β-cell lineage and would avoid the controversy and technical problems associated with pluripotent stem cells (4).

Evidence that stem cells reside in the pancreatic ducts is provided by rodent models of pancreas regeneration (57). In vitro–cultured mouse ductal cells seem to provide a source of pancreatic progenitors (8,9). Endocrine differentiation has also been reported in human pancreatic duct cell–enriched cultures (10). However, the exact nature of pancreatic progenitor/stem cells is still controversial. In addition to the duct epithelial cell, another candidate islet progenitor cell has been described recently, characterized by its expression of the neural stem cell marker nestin and lack of established islet and duct cell markers (11,12). Nestin-positive cells were identified within islets and also in centrolobular ducts. These cells were reported to differentiate in vitro into pancreatic endocrine, exocrine, and hepatic phenotypes (12). Differentiation of insulin-producing cells from mouse embryonic stem cells also involved an intermediate cell type expressing nestin (13). However, descriptive analyses of mouse (14) and human (15) development argue against a role of nestin-positive stem cells in islet differentiation.

Several studies have demonstrated that pancreatic ductal cells express markers associated with islet differentiation, particularly Ipf-1/Pdx-1, when placed in tissue culture (1618). These cells, presumed to represent latent endocrine progenitors, were shown to differentiate given the appropriate external stimuli (10). The factors used by Bonner-Weir et al. (10) included fibroblast growth factor-7 (FGF-7) to stimulate ductal cell proliferation and Matrigel, a commercial basement membrane matrix, plus nicotinamide (NIC) to initiate and stimulate endocrine differentiation (1921). In this study, we first reproduced this model and then systematically evaluated the contribution of the various factors involved in the proliferation versus differentiation of precursor cells. Finally, by pulse-chase analysis of proliferating cells, the phenotype of human islet precursor cells was determined.

Cell expansion and differentiation.

Studies by Bonner-Weir et al. (10) provided us a model to investigate endocrine differentiation and the factors involved in human adult pancreatic tissue culture. Human islets were isolated according to previously described methods (22,23) in Uppsala, Sweden. After Ficoll gradient purification, the mixed fractions rich in ductal fragments and poor in islets were collected and shipped on ice to Biomedicum Helsinki for subsequent culture. All procedures were approved by institutional ethical committees in Sweden and Finland.

The fresh tissue was first maintained in serum-free Ham’s F-10 (10 mmol/l glucose; Life Technologies) with 0.5% BSA (fraction V; Sigma Chemical Co., St. Louis, MO) in nonadherent Petri dishes (Sterilin, Stone, Staffordshire, U.K.) at 37°C in a humidified atmosphere of 95% air and 5% CO2 for 2–4 days (24), followed by monolayer expansion in nontreated six-well plates or T-75 flasks (Becton Dickinson) with CMRL 1066 medium (Life Technologies) supplemented with 10% FCS (PromoCell). Within 7–10 days, a monolayer consisting mainly of cuboid epithelial cells had formed substantial plaques, and the media were changed into serum-free Dulbecco’s modified Eagle medium/F12 (National Public Health Institute, Helsinki, Finland) supplemented with ITS (5 mg/l insulin + 5 mg/l transferrin + 5 μg/l sodium selenite; Sigma), 2 g/l BSA, 8 mmol/l glucose, 10 mmol/l NIC (Sigma), and 10 ng/ml FGF-7 (R&D Systems). These conditions were termed as the “standard conditions.” Omission of NIC or FGF-7 was chosen in parallel culture wells. In some wells, 10% FCS was added to observe the serum effect on differentiation. After changing into serum-free medium, Matrigel, a basement membrane preparation from Engelbreth-Holm-Swarm mouse tumor cells (Becton Dickinson) was applied on the top of the cells according to the manufacturer’s instructions with the exception of dilution (1:10) and overnight gelling time at 37°C. Inverted light microscopy was used to monitor morphologic changes. The same cell culture fields were serially photographed at 3, 7, 10, 14, 18, and 21 days after Matrigel overlay. Final samples were collected after 4–5 weeks in culture. Those three-dimensional structures, cysts, and cultivated human islet buds (CHIBs) that protruded from monolayers were hand-picked in six-well plates and mechanically sheared with the stream of culture medium in T-75 flasks. Remaining monolayer cells were collected by a solution of 0.05% trypsin and 0.02% EDTA (Life Technologies). Dithizone staining was used to determine the purity of fresh tissue and to assess quickly the insulin-containing cells in differentiation experiments (25).

Bromodeoxyuridine labeling.

For determining cell proliferation during monolayer expansion, the culture cells were labeled with 10 μg/ml bromodeoxyuridine (BrdU; Zymed) for 24 h at various time points before fixation in 4% paraformaldehyde (PFA). A pulse-chase protocol, in which cells were first labeled with BrdU for 24 h and then incubated in the absence of BrdU until final processing, was used to trace the fates of proliferating cells during later differentiation. The two time points of labeling were performed because pilot experiments indicated that the rapid growth of cytokeratin 19 (CK19)-positive and nestin-positive cells appeared separately at approximately day 3 and 7.

Cell processing and immunocytochemical staining.

For preparing cytocentrifuge slides from the cultured cells, the aggregates were dissociated with 0.05% trypsin and 0.02% EDTA, washed in PBS, and spun to microscope slides (SuperFrost Plus) by centrifugation at 700 rpm for 8 min. Intact monolayer cultures were prepared for analysis in eight-chamber slides (Nunc Lab-Tek). Monolayers and cytospin slides were fixed in 4% PFA for 15 min and rinsed in PBS. Alternatively, harvested cysts/CHIBs were fixed for 1 h in 4% PFA, rinsed with PBS, suspended in 2% agarose-PBS solution, and centrifuged to form compact pellets, which were further embedded in paraffin for sectioning.

Immunostaining was performed in the above preparations to identify various cell types using the following primary antibodies: guinea pig anti-porcine insulin 1:100, rabbit anti-human glucagon 1:500, rabbit anti-human chromogranin A 1:500, rabbit anti-human α-amylase 1:200, mouse anti-human cytokeratin 19 1:50, mouse anti-human vimentin 1:200, and mouse anti-BrdU 1:100 (all from Dako); rabbit anti-Nkx6.1 1:500 (provided by Dr. Ole Madsen, Hagedorn Research Institute, Gentofte, Denmark); rabbit anti–Ipf-1/Pdx-1 1:500 (provided by Dr. Chris Wright, Vanderbilt University, Nashville, TN); and rabbit anti-human nestin 1:100 (provided by Dr. Urban Lendahl, Karolinska Institute, Stockholm, Sweden). Nonspecific binding was blocked by preincubation in 3% normal serum (Zymed) from the species in which the secondary antibody was raised, followed by incubation of primary antibodies for 1 h at room temperature. Biotinylated goat anti-rabbit and biotinylated rabbit anti-mouse IgGs (1:200; Zymed) were used as secondary antibodies. Peroxidase-conjugated streptavidin (1:200; Zymed) was used by developing the substrate of 3-amino-9-ethylcarbazole. Light counterstaining was performed with hematoxylin. Microwave treatment in citrate buffer was necessary to retrieve the antigenicity of CK19, nestin, and BrdU, whereas 0.1% pepsin-0.1 mol/l HCl was optimal for hormone retrieving. For double staining of CK19, nestin, or chromogranin A together with BrdU, the Vectastain ABC-kit (Vector, Burlingame, CA) was used. Double immunofluorescent staining was performed to check for colocalization of nestin/vimentin using conjugated secondary antibodies as follows: FITC-conjugated donkey anti-mouse and TRITC-conjugated donkey anti-rabbit IgGs (1:50; Dako).

Insulin and DNA content.

For the analysis of insulin and DNA content, the cells were washed twice in PBS, resuspended in 300 μl of distilled cold water, and homogenized by sonication on ice. An aliquot of the homogenates in duplicate was analyzed fluorometerically for the DNA content (26), and another was extracted with acid ethanol overnight and measured for insulin content using a solid-phase radioimmunoassay kit (DPC, Los Angeles, CA).

Insulin release.

Dynamic insulin release from differentiated CHIBs was studied with a perifusion system (Superfusion 600; Brandel, Gaithersburg, MD) as described previously (27). Briefly, batches of selected pure human islets or harvested cysts/CHIBs were loaded in separate perifusion chambers and exposed to Krebs-Ringer bicarbonate buffer supplemented with 20 mmol/l HEPES (Sigma) and 0.2% BSA at a flow rate of 0.25 ml/min. After a 60-min stabilizing period in 1.67 mmol/l glucose, the cells were stimulated with 16.7 mmol/l glucose and 10 mmol/l theophylline. Fractions were collected every 4 min and analyzed for their insulin content.

RNA extraction and analysis by Northern blotting.

Total RNA from pure human islets and harvested CHIBs was extracted using the Gen Elute Mammalian Total RNA Kit (Sigma). Total RNA (∼10 μg/lane for CHIBs and 2.5 μg/lane for islets) was fractionated on a 1.2% formalin-agarose gel and transferred to a nylon membrane (Hybond-N; Amersham) by capillary blotting. The cDNA probe was 32P-labeled by a random priming method (Prime-A-Gene Labeling System; Promega). Hybridizations were done in buffer containing 1% SDS, 1 mol/l NaCl, and 8% dextran sulfate overnight at 65°C. The blots were washed at 65°C in 1× SSC and finally in 0.5× SSC. Hybridization signals were visualized using a Bio-imaging analyzer (Fuji Photo Film). The hybridization signals were normalized by the housekeeping gene cyclophilin (28).

Transplantation of CHIBs.

Six- to 8-week-old male athymic nude Balb/c (nu/nu) mice were purchased from Harlan (Horst, the Netherlands). The mice were housed in isolators (Scantainer) and had water and food ad libitum. Animals were anesthetized with fentanyl-fluanisone 3 ml/kg (Hypnorm; Jansen Gilac, Espoo, Finland) and diatsepam 5 mg/kg (Diapam; Orion, Espoo, Finland) intraperitoneally. Approximately 5 μl of packed CHIBs harvested from 4-week-old cultures were injected with a microinjector under left kidney capsule. The animals (n = 5) were killed 3 months after transplantation, the left kidney was removed, and the graft site was dissected and fixed for 6 h in Bouin’s fixative.

Statistics.

The differences at various time points and experimental conditions were analyzed with Student’s t test within two groups and one-way ANOVA and the Fischer’s PLSD test at 95% significance level for multiple comparisons (Statview 4.1; Abacus, Berkeley, CA).

Cell attachment and monolayer expansion phase.

The material in the continuous Ficoll gradients was collected in 15 fractions. A pool of duct-rich and islet-poor fractions was used as our starting material. The endocrine β-cell proportion in these fractions ranged from 3 to 20% (9 ± 1%, based on insulin immunostaining; n = 15) and ductal cell proportion between 30 and 60% (50 ± 3%, based on CK19 immunoreactivity). Nonadherent vessels were reported to favor the attachment of ductal cells rather than islets (10). We also found out that the attached cells represented 27.2 ± 4.7% of original DNA and only 6.7 ± 1.1% of the original insulin content (n = 6). The attached cells started to grow and form a monolayer from day 2. Within 7–10 days of monolayer expansion, two major cellular phenotypes were observed: cuboid epithelial cells and serpiginous spindle-like cells.

Cultures were prepared from cells originating from 35 different donors. Successful monolayer expansion and subsequent differentiation were recorded in 31 of these (89%).

By comparing the insulin and DNA contents in the same aliquots after initial attachment and monolayer expansion, we found a twofold decrease in insulin content accompanied by a similar increase in the DNA content in 5 days (Fig. 1). Consistent with this, a decrease in the percentage of insulin-positive cells as well as glucagon-positive cells was seen by immunostaining, whereas the proportion of CK19-positive cells increased during the same period (Table 1).

Cell type–specific proliferation was analyzed by BrdU incorporation. Two peaks of proliferation appeared separately at days 3–4 and 7–8 with mean BrdU labeling indexes of 7 ± 2 and 12 ± 3% (n = 3), respectively. At day 3, the monolayers mainly consisted of large patches of ductal epithelial cells (77 ± 5%) with limited numbers of nestin-positive cells (4 ± 2%). Almost all BrdU labeling was detected within the CK19-positive population with very few BrdU/nestin double-positive cells (Fig. 2A and B). As shown in Fig. 2L, Ipf-1/Pdx-1 was expressed in the cytoplasm of proliferating ductal cells at this time. From day 5, nestin-positive cells started to grow and reached one-fourth (26 ± 3%) of the total cell population by day 8. The proportion of BrdU/nestin double-positive cells increased to 11 ± 2%. At the same time, the CK19-positive population decreased to 54 ± 7% and only a few BrdU/CK19 double-positive cells were seen (Fig. 2C and D). At this point, double immunofluorescent staining of nestin and vimentin showed that nestin-positive cells formed a subpopulation of vimentin-positive cells (Fig. 2E and F). A small number of preexisting endocrine cells remained during monolayer expansion, but none were BrdU positive (Table 2). Thus, the 80% confluent monolayer formed during the first phase of serum-containing culture was mainly due to the initial proliferation of CK19-positive cells, followed by the growth of the nestin-positive cell population.

Differentiation phase.

During the 14–21 days after Matrigel overlay and serum-free culture, many cell colonies migrated into the gel and formed three-dimensional cystic structures. With time, numerous small dense buds (CHIBs) grew out of the cyst walls and sometimes became separated from the cysts (Fig. 3). These buds turned red after dithizone staining (data not shown). The cysts/CHIBs mainly consisted of CK19-positive ductal cells and hormone-positive endocrine cells, whereas nestin-positive cells were rare (Fig. 2I–K). Ipf-1/Pdx-1 was also expressed in CHIBs. However, different from the cytoplasmic expression pattern in the monolayer, it appeared inside the nuclei of endocrine β-cells identified by sequential Nkx6.1 staining (Fig. 2M and N). Compared with the same aliquot of cells at monolayer culture, final harvested cellular insulin content and insulin-to-DNA ratio increased by eight- and fivefold, respectively (Fig. 1). Measurement of insulin content and immunocytochemistry analysis demonstrated that insulin-producing cells were enriched in the CHIBs (Fig. 1, Table 1).

We used the expression of chromogranin A as a marker of endocrine cells in the CHIBs. Using the pulse-chase protocol, we found that BrdU-labeled endocrine cells appeared only in CHIBs from cultures that had been labeled early (day 3) at the time of active CK19-positive cell proliferation (Fig. 2G). BrdU labeling at the time of nestin-positive cell growth (day 7) did not result in chromogranin A/BrdU double-positive cells (Fig. 2H). Although the overall BrdU labeling remained low in the CHIBs, up to 50% of the labeled cells were endocrine (Table 2).

Critical factors for endocrine differentiation.

It is obvious that the differentiation process involved combined functions of extracellular matrix and soluble factors. Omission of serum was found to be essential for the development of CHIBs. There was no cyst development at all and very little insulin content when the cells were maintained in serum-containing medium. Also, the Matrigel overlay procedure was found to be absolutely necessary for three-dimensional structure formation, and without this manipulation, insulin/DNA ratio was 50% lower than the standard condition. Furthermore, when the ductal cells were allowed to grow on another matrix (804G) found to support their growth (29), CHIB development was prevented (data not shown). Omission of NIC from the medium decreased both the numbers of CHIBs and their insulin content per DNA by ∼40%. Addition of FGF-7 to the serum-free medium had no effect on insulin content, endocrine cell proportion, or CHIB number (Fig. 4, Table 3).

Insulin release.

Perifusion studies were performed to study the CHIBs’ responsiveness to the stimulation of glucose and theophylline. For each perifusion experiment, 100 CHIBs were selected after 5 weeks of culture. Thirty pure islets were hand-picked from freshly isolated human islet fractions to be used as a positive control. The cells were challenged sequentially with 16.7 mmol/l glucose and 16.7 mmol/l glucose plus 10 mmol/l theophylline. As shown in Fig. 5, both islets and CHIBs responded in a biphasic manner to glucose, and theophylline potentiated the response equally in both types of cells. Compared with basal insulin secretion at 1.67 mmol/l glucose, the first-phase insulin response was 4-fold in the CHIBs and 12-fold in the islets. The basal rate of insulin release, as related to the DNA content, was similar in both cell types.

Insulin and glucagon gene expression in CHIBs.

Insulin and glucagon mRNA levels from two separate CHIB preparations were compared with those of freshly isolated islets. As shown in Fig. 6, the normalized insulin mRNA levels in CHIBs were only 4–5% of that found in islets, whereas glucagon mRNA levels were similar in CHIBs and islets. These results are consistent with an immature stage of islet differentiation, and they also confirm that the insulin immunostaining in CHIBs is due to insulin synthesis and not just uptake from the insulin-containing medium, as has recently been shown to be the case in embryonic stem cell differentiation experiments (30).

Transplantation into nude mice.

CHIBs were transplanted under the kidney capsule in five nude mice. Three months after transplantation, only one graft had clearly grown and showed expression of pancreatic endocrine and exocrine markers. It is interesting that in this graft, separate areas contained endocrine (insulin or glucagon positive) cells, ductal cells (CK19 positive), and exocrine acinar (amylase positive) cells (Fig. 2O–R). No amylase-containing cells were ever detected in the CHIBs before transplantation. In the other experiments, only scar tissue was found at the site of the graft.

The only previous study to demonstrate the development of islet buds from adult human pancreatic tissue was published by Bonner-Weir et al. (10). In the present study, we reproduced this culture method and confirmed the capacity of endocrine differentiation from progenitors present in the adult human pancreas. The results also clearly demonstrate that serum-contained factors effectively inhibit this process. Furthermore, by labeling of proliferating cells, we obtained direct evidence that differentiated endocrine cells are derived from CK19- but not nestin-positive cells.

Our experimental strategy began with the identification of proper Ficoll fractions containing the putative precursor cells for in vitro expansion and differentiation. The pellet fractions were cultured several times but without any evidence of epithelial cell proliferation. This could be caused by the low proportion of ductal cells in the pellet and by the damage of digestive enzymes released by large numbers of acinar cells. The cell populations present in upper Ficoll layers were systematically analyzed. We could not, however, identify any pure ductal fractions without contaminating endocrine and acinar cells. Although we used fractions that contained as few islets as possible, the inherent problem of preexisting endocrine cells remained in the differentiation experiments. However, as a whole, our results clearly indicate that the endocrine cells present in the CHIBs have differentiated in vitro and do not result from the migration of preexisting islet cells. The strongest argument for this is the demonstration of endocrine expression in up to 50% of BrdU-labeled cells when the cultures had been labeled during the time of ductal cell proliferation.

Previous studies proposed two possibilities for the islet progenitor cell type in the adult pancreas. Ductal epithelial cells have been demonstrated to give rise to islet tissue in various experimental models of pancreas regeneration (57). Ductal origin of neogenic islets is also suggested by morphologic observations in human tissue (31,32) and in clinical pathologies (33). Another type of tissue stem cell was reported in the rat pancreas to consist of a distinct population of cells expressing the neural stem cell marker nestin (11,12). In our experiments, after the initial proliferation of the ductal cells, we also observed a nestin-expressing population that proliferated actively. Nestin, a type VI intermediate filament protein, requires coexpression of type III proteins such as vimentin for normal assembly (34). Our data confirm the colocalization of nestin and vimentin and demonstrate that nestin-positive cells form a subpopulation of vimentin cells, which have typically been considered as contaminating fibroblasts in primary cultures.

To determine which cell type represents the endocrine precursor, we labeled the monolayer cells with BrdU separately at the waves of CK-positive cell proliferation or nestin-positive cell proliferation and chased their differentiation fates in the CHIBs. The identification of BrdU-positive endocrine cells establishes that the endocrine cells have truly developed in vitro because no BrdU uptake was observed in preexisting endocrine cells. The BrdU pulse, even if lasting for 24 h, is expected to label only part of the cells ever undergoing a proliferative cycle during the monolayer culture. Dilution of the incorporated BrdU as a result of further division of the labeled cells during subsequent culture may also decrease the number of BrdU-positive cells in CHIBs, particularly after early-phase labeling. Moreover, BrdU may have a negative effect on pancreatic cell differentiation (35). Thus, the numbers of BrdU-labeled endocrine cells may underestimate the differentiation of initially proliferating ductal cells.

On the basis of the experimental conditions used in the previously described successful protocol (10), we investigated in more detail the critical role of its individual components. Our results demonstrate that serum-free medium is absolutely necessary for the development of islet cells, and serum completely inhibits the formation of CHIBs. This is consistent with previous observations showing that serum suppresses islet differentiation (27). The exact growth factors responsible for this potent effect remain unknown.

Extracellular matrix has been shown to play a crucial role in cell differentiation through rearrangement of the cytoskeletal network (36). This is consistent with our finding that application of Matrigel on the progenitor cells is essential for cell migration, the organization of three-dimensional cystic structures, and thereafter the protrusion of islet buds. Laminin is reported to be involved in duct lineage selection in the development of mouse embryonic pancreas (20) and to promote β-cell differentiation in cell culture (19). Because the laminin-rich matrix produced by the 804G cell line is reported to support epithelial ductal cell growth (29), we planned to use it for extensive expansion of the ductal monolayer. However, when the ductal cells were sandwiched between the 804G matrix and Matrigel, formation of three-dimensional structures in the differentiation phase was arrested. These two basement membrane matrices contain different types of laminins and various other extracellular matrix components (37,38), which may interact with each other and thus influence islet differentiation. Consistent with previous reports (21,27,39), the addition of NIC promoted the development of endocrine cells. However, NIC was not an absolutely required factor because its omission decreased the formation of CHIBs by only 40%.

FGF-7 is known to be a potent mitogen for a variety of epithelial cell types and to support the growth of embryonic pancreatic epithelium while repressing endocrine development (40,41). As FGF-7 was included in the published protocol, we also included it in the serum-free medium during the differentiation phase. However, we could not identify any major effects on the growth and/or differentiation of the pancreatic cells in these conditions, and omission of FGF-7 did not affect the results. Thus, in further optimization of the method, it seems more logical to apply this type of growth factors for the initial expansion of the progenitor cells rather than in the differentiation phase.

Ductal cells start to reexpress Ipf-1/Pdx-1 when they proliferate, which has been taken as an indication of their precursor cell capacity (16,18,42). We detected Ipf-1/Pdx-1 expression mainly in the cytoplasm of the ductal cells during the expansion phase, whereas nuclear expression was detected in the endocrine cells within the CHIBs. It has previously been shown that an inactive 31-kDa form of Ipf-1/Pdx-1 locates in the cytoplasm of β-cells from where it translocates into the nucleus as larger molecular-weight forms when the cells are activated (43). The 31-kDa form of Ipf-1/Pdx-1 has also been detected in the cytoplasm of human ductal cells (44). Our observations thus suggest that an inactive cytoplasmic form of Ipf-1/Pdx-1 is expressed in the proliferating ductal cells and this then translocates into the nucleus in the fully differentiated endocrine cells of the CHIBs.

After transplantation under the kidney capsule, the CHIBs in most animals failed to engraft. The reason for this is unknown but is most likely not due to immunologic rejection of the graft. Previous studies using fetal human islet tissue or adult human islets have demonstrated that athymic mice tolerate human islet tissue (45,46). Importantly, the microenvironment in the fetal human islet grafts seems to contain all factors needed to promote engraftment and subsequent differentiation of the fetal tissue to mature insulin-producing cells capable of achieving normoglycemia in diabetic recipients. In contrast, the majority of the CHIB grafts failed to engraft, and only scar tissue remained at the site of implantation. These unexpected findings indicate that some factor(s) needed for a successful engraftment is lacking in the CHIBs. The nature of this factor is unknown, but it may be speculated that the CHIBs failed to induce proper revascularization, because the β-cells within the CHIBs were differentiated in vitro to the level that they were capable of responding to glucose and displayed near normal insulin content. Even so, the pancreatic stem cells present in human islet preparations used for clinical transplantation may survive if transplanted together with mature islets. If so, then these pancreatic stem cells may play an important role after islet transplantation in humans by providing a cell source capable of islet neogenesis and thereby enhancing the long-term functional capacity of the graft.

In summary, the results presented here confirm that islet neogenesis can be induced in vitro from progenitors within adult human pancreatic ductal cells. It seems likely that neogenesis of islet cells from precursors transplanted along with the isolated islets contributes to the final endocrine cell content of the islet graft. However, for developing a significant additional source of human islets, improved methods are needed for the expansion and differentiation of the ductal cells without losing their capacity to survive after transplantation.

FIG. 1.

Insulin content and DNA content and their ratio at three different time points during the whole 4- to 5-week culture. Data are the mean ± SE of five separate experiments with cells from different donors. Each data point represents the same number of cells at the onset (one culture well in a six-well plate). Significant differences are indicated as *P < 0.05, **P < 0.01.

FIG. 1.

Insulin content and DNA content and their ratio at three different time points during the whole 4- to 5-week culture. Data are the mean ± SE of five separate experiments with cells from different donors. Each data point represents the same number of cells at the onset (one culture well in a six-well plate). Significant differences are indicated as *P < 0.05, **P < 0.01.

Close modal
FIG. 2.

Immunostaining of monolayers (AF and L), harvested CHIBs (GK, M, and N), and grafted tissue (OR). At day 3, the proliferating cells (blue BrdU-positive nuclei) are CK19-positive (A), and the rare nestin-positive cells do not proliferate (B). At day 7, the CK19-positive cells have stopped growing (C), whereas spindle-shaped nestin-positive cells proliferate actively (D). Double-immunofluorescence staining for nestin (E) and vimentin (F) demonstrates that some of the vimentin-expressing cells also express nestin. Double staining for chromogranin A (red) and BrdU (blue) can be demonstrated only in sections of CHIBs labeled with BrdU at day 3 (G) but not at day 7 (H). Consecutive sections of cyst/CHIB structures demonstrate CK19-positive ductal epithelial structures (I) with peripheral insulin (J) and glucagon (K) positive cells. Ipf-1/Pdx-1 immunoreactivity is detected in the cytoplasm of proliferating ductal cells during monolayer expansion (L) and in the nuclei of endocrine cells in the CHIBs (M), some of which are also positive for Nkx6.1 (N). Immunostaining of consecutive sections of a CHIB graft harvested 3 months after transplantation reveals acinar structures strongly positive for amylase (O, arrow) and separate areas of endocrine differentiation with cells positive for insulin (P, arrow) or glucagon (Q, arrow). Only very few cells in the graft express the ductal marker CK19 (R, arrow).

FIG. 2.

Immunostaining of monolayers (AF and L), harvested CHIBs (GK, M, and N), and grafted tissue (OR). At day 3, the proliferating cells (blue BrdU-positive nuclei) are CK19-positive (A), and the rare nestin-positive cells do not proliferate (B). At day 7, the CK19-positive cells have stopped growing (C), whereas spindle-shaped nestin-positive cells proliferate actively (D). Double-immunofluorescence staining for nestin (E) and vimentin (F) demonstrates that some of the vimentin-expressing cells also express nestin. Double staining for chromogranin A (red) and BrdU (blue) can be demonstrated only in sections of CHIBs labeled with BrdU at day 3 (G) but not at day 7 (H). Consecutive sections of cyst/CHIB structures demonstrate CK19-positive ductal epithelial structures (I) with peripheral insulin (J) and glucagon (K) positive cells. Ipf-1/Pdx-1 immunoreactivity is detected in the cytoplasm of proliferating ductal cells during monolayer expansion (L) and in the nuclei of endocrine cells in the CHIBs (M), some of which are also positive for Nkx6.1 (N). Immunostaining of consecutive sections of a CHIB graft harvested 3 months after transplantation reveals acinar structures strongly positive for amylase (O, arrow) and separate areas of endocrine differentiation with cells positive for insulin (P, arrow) or glucagon (Q, arrow). Only very few cells in the graft express the ductal marker CK19 (R, arrow).

Close modal
FIG. 3.

Formation of three-dimensional structures (cysts with protruding islet buds) with serum-free medium and overlaid Matrigel. Serial events of monolayer cell migration, cyst formation, and islet-bud protrusion within the same microscopic field at days 3 (A), 7 (B), 10 (C), 14 (D), 18 (E), and 21 (F) after Matrigel overlay. Magnification 20×.

FIG. 3.

Formation of three-dimensional structures (cysts with protruding islet buds) with serum-free medium and overlaid Matrigel. Serial events of monolayer cell migration, cyst formation, and islet-bud protrusion within the same microscopic field at days 3 (A), 7 (B), 10 (C), 14 (D), 18 (E), and 21 (F) after Matrigel overlay. Magnification 20×.

Close modal
FIG. 4.

Insulin content and DNA content and their ratio after a 4-week culture in various differentiation conditions. Standard condition, as a control, is serum-free Dulbecco’s modified Eagle medium/F12 medium with 10 ng/ml FGF-7 and 10 mmol/l NIC and Matrigel overlay. The effects of omission of each of these factors or the addition of 10% FCS are shown as percentage of control. Data are the mean ± SE of four separate experiments with cells from different donors. Significant differences between groups are indicated as *P < 0.001 vs. the standard condition.

FIG. 4.

Insulin content and DNA content and their ratio after a 4-week culture in various differentiation conditions. Standard condition, as a control, is serum-free Dulbecco’s modified Eagle medium/F12 medium with 10 ng/ml FGF-7 and 10 mmol/l NIC and Matrigel overlay. The effects of omission of each of these factors or the addition of 10% FCS are shown as percentage of control. Data are the mean ± SE of four separate experiments with cells from different donors. Significant differences between groups are indicated as *P < 0.001 vs. the standard condition.

Close modal
FIG. 5.

Insulin release in perifusion of pure human fresh islets and harvested CHIBs harvested after 4–5 weeks in culture. Experiments were started after a 60-min stabilizing period in low glucose (1.67 mmol/l). The cells were stimulated sequentially with 16.7 mmol/l glucose at 16–52 min and theophylline (10 mmol/l) combined with high glucose at 56–68 min. The effluent insulin concentrations have been normalized against the DNA content of the perifused cells. Data are the mean ± SE of three separate experiments for islets and four separate experiments for CHIBs.

FIG. 5.

Insulin release in perifusion of pure human fresh islets and harvested CHIBs harvested after 4–5 weeks in culture. Experiments were started after a 60-min stabilizing period in low glucose (1.67 mmol/l). The cells were stimulated sequentially with 16.7 mmol/l glucose at 16–52 min and theophylline (10 mmol/l) combined with high glucose at 56–68 min. The effluent insulin concentrations have been normalized against the DNA content of the perifused cells. Data are the mean ± SE of three separate experiments for islets and four separate experiments for CHIBs.

Close modal
FIG. 6.

Insulin and glucagon mRNA expression in two different CHIB preparations and freshly isolated islets analyzed by Northern blotting. The mRNA levels were densitometrically normalized to cyclophilin mRNA expression. Expression in pure islets is equal to 100.

FIG. 6.

Insulin and glucagon mRNA expression in two different CHIB preparations and freshly isolated islets analyzed by Northern blotting. The mRNA levels were densitometrically normalized to cyclophilin mRNA expression. Expression in pure islets is equal to 100.

Close modal
TABLE 1

Endocrine and ductal cell proportions at four culture time points

Total cells (%)
InsulinGlucagonCK19
Fresh tissue (day 0) 9.4 ± 1.2 4.2 ± 0.3 49.9 ± 3.5 
Initial attachment (day 2) 3.4 ± 0.9 Not done 54.0 ± 10.7 
Monolayer (day 5) 1.9 ± 0.3 1.1 ± 0.3 70.0 ± 4.6 
CHIBs (day 28) 22.7 ± 2.6 12.3 ± 1.1 25.1 ± 2.4 
Remaining monolayer 3.1 ± 1.5 0.1 ± 0.04 1.8 ± 0.6 
Total cells (%)
InsulinGlucagonCK19
Fresh tissue (day 0) 9.4 ± 1.2 4.2 ± 0.3 49.9 ± 3.5 
Initial attachment (day 2) 3.4 ± 0.9 Not done 54.0 ± 10.7 
Monolayer (day 5) 1.9 ± 0.3 1.1 ± 0.3 70.0 ± 4.6 
CHIBs (day 28) 22.7 ± 2.6 12.3 ± 1.1 25.1 ± 2.4 
Remaining monolayer 3.1 ± 1.5 0.1 ± 0.04 1.8 ± 0.6 

Data are the means ± SE of four separate experiments with cells from different donors.

TABLE 2

Endocrine cell content and their BrdU labeling in monolayer culture and tracking of the labeled cells in CHIBs

Expanded monolayer
CHIBs
Chro A (%)BrdU (%)BrdU/Chro A (%)Chro A (%)BrdU (%)BrdU/Chro A (%)
Hu35       
 Day 3 1.7 6.1 15.0 0.8 0.05 
 Day 7 0.7 11.6 13.8 2.8 
Hu37       
 Day 3 0.9 6.8 25.6 0.8 0.5 
 Day 7 1.0 7.4 17.5 2.6 
Hu38       
 Day 3 1.5 4.1 25.1 0.7 0.3 
 Day 7 1.2 10.4 19.1 1.9 
Expanded monolayer
CHIBs
Chro A (%)BrdU (%)BrdU/Chro A (%)Chro A (%)BrdU (%)BrdU/Chro A (%)
Hu35       
 Day 3 1.7 6.1 15.0 0.8 0.05 
 Day 7 0.7 11.6 13.8 2.8 
Hu37       
 Day 3 0.9 6.8 25.6 0.8 0.5 
 Day 7 1.0 7.4 17.5 2.6 
Hu38       
 Day 3 1.5 4.1 25.1 0.7 0.3 
 Day 7 1.2 10.4 19.1 1.9 

Individual results are presented from experiments with cells derived from three donors. Chromogranin A was used as a marker of endocrine cells.

TABLE 3

The importance of nicotinamide and FGF-7 on the formation and differentiation of CHIBs

Cell proportion (% of total cells)
Number of CHIBs
InsulinGlucagon
Standard condition 22.7 ± 2.6 12.3 ± 1.1 50 ± 4 
NIC omission 13.1 ± 2.1 1.0 ± 0.8 30 ± 3 
FGF-7 omission 19.7 ± 5.1 13.0 ± 0.9 49 ± 3 
Cell proportion (% of total cells)
Number of CHIBs
InsulinGlucagon
Standard condition 22.7 ± 2.6 12.3 ± 1.1 50 ± 4 
NIC omission 13.1 ± 2.1 1.0 ± 0.8 30 ± 3 
FGF-7 omission 19.7 ± 5.1 13.0 ± 0.9 49 ± 3 

Data are the means ± SE of four separate experiments with cells obtained from different donors. The data represent the development of the same quantity (∼2 μg DNA) of originally seeded tissue in one well of a six-well plate.

This work was supported by grants from the Juvenile Diabetes Research Foundation, the Sigrid Juselius Foundation, the Academy of Finland, and the Helsinki University Hospital Research Funds (EVO).

We thank Dr. Kaija Salmela and the Nordic network for clinical islet transplantation for procurement of the human pancreatic cells. Ms. Päivi Kinnunen is thanked for skilled technical assistance.

1.
Ryan EA, Lakey JR, Paty BW, Imes S, Korbutt GS, Kneteman NM, Bigam D, Rajotte RV, Shapiro AM: Successful islet transplantation: continued insulin reserve provides long-term glycemic control.
Diabetes
51
:
2148
–2157,
2002
2.
Berna G, Leon-Quinto T, Ensenat-Waser R, Montanya E, Martin F, Soria B: Stem cells and diabetes.
Biomed Pharmacother
55
:
206
–212,
2001
3.
Jiang Y, Jahagirdar BN, Reinhardt RL, Schwartz RE, Keene CD, Ortiz-Gonzalez XR, Reyes M, Lenvik T, Lund T, Blackstad M, Du J, Aldrich S, Lisberg A, Low WC, Largaespada DA, Verfaillie CM: Pluripotency of mesenchymal stem cells derived from adult marrow.
Nature
418
:
41
–49,
2002
4.
Serup P, Madsen OD, Mandrup-Poulsen T: Science, medicine, and the future: islet and stem cell transplantation for treating diabetes.
BMJ
322
:
29
–32,
2001
5.
Vinik A, Pittenger G, Rafaeloff R, Rosenberg L, Duguid WP: Determinants of pancreatic islet cell mass: a balance between neogenesis and senescence/apoptosis.
Diabet Rev
4
:
235
–263,
1996
6.
Bonner-Weir S, Baxter LA, Schuppin GT, Smith FE: A second pathway for regeneration of adult exocrine and endocrine pancreas: a possible recapitulation of embryonic development.
Diabetes
42
:
1715
–1720,
1993
7.
Gu D, Sarvetnick N: Epithelial cell proliferation and islet neogenesis in IFN-g transgenic mice.
Development
118
:
33
–46,
1993
8.
Cornelius JG, Tchernev V, Kao KJ, Peck AB: In vitro generation of islets in long-term cultures of pluripotent stem cells from adult mouse pancreas.
Horm Metab Res
29
:
271
–277,
1997
9.
Ramiya VK, Maraist M, Arfors KE, Schatz DA, Peck AB, Cornelius JG: Reversal of insulin-dependent diabetes using islets generated in vitro from pancreatic stem cells.
Nat Med
6
:
278
–282,
2000
10.
Bonner-Weir S, Taneja M, Weir GC, Tatarkiewicz K, Song KH, Sharma A, O’Neil JJ: In vitro cultivation of human islets from expanded ductal tissue.
Proc Natl Acad Sci U S A
97
:
7999
–8004,
2000
11.
Hunziker E, Stein M: Nestin-expressing cells in the pancreatic islets of Langerhans.
Biochem Biophys Res Commun
271
:
116
–119,
2000
12.
Zulewski H, Abraham EJ, Gerlach MJ, Daniel PB, Moritz W, Muller B, Vallejo M, Thomas MK, Habener JF: Multipotential nestin-positive stem cells isolated from adult pancreatic islets differentiate ex vivo into pancreatic endocrine, exocrine, and hepatic phenotypes.
Diabetes
50
:
521
–533,
2001
13.
Lumelsky N, Blondel O, Laeng P, Velasco I, Ravin R, McKay R: Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets.
Science
292
:
1389
–1394,
2001
14.
Selander L, Edlund H: Nestin is expressed in mesenchymal and not epithelial cells of the developing mouse pancreas.
Mech Dev
113
:
189
–192,
2002
15.
Piper K, Ball SG, Turnpenny LW, Brickwood S, Wilson DI, Hanley NA: Beta-cell differentiation during human development does not rely on nestin-positive precursors: implications for stem cell-derived replacement therapy.
Diabetologia
45
:
1045
–1047,
2002
16.
Gmyr V, Kerr-Conte J, Belaich S, Vandewalle B, Leteurtre E, Vantyghem MC, Lecomte-Houcke M, Proye C, Lefebvre J, Pattou F: Adult human cytokeratin 19-positive cells reexpress insulin promoter factor 1 in vitro: further evidence for pluripotent pancreatic stem cells in humans.
Diabetes
49
:
1671
–1680,
2000
17.
Beattie GM, Itkin-Ansari P, Cirulli V, Leibowitz G, Lopez AD, Bossie S, Mally MI, Levine F, Hayek A: Sustained proliferation of Pdx-1(+) cells derived from human islets.
Diabetes
48
:
1013
–1019,
1999
18.
Rooman I, Heremans Y, Heimberg H, Bouwens L: Modulation of rat pancreatic acinoductal transdifferentiation and expression of PDX-1 in vitro.
Diabetologia
43
:
907
–914,
2000
19.
Jiang FX, Cram DS, DeAizpurua HJ, Harrison LC: Laminin-1 promotes differentiation of fetal mouse pancreatic β-cells.
Diabetes
48
:
722
–730,
1999
20.
Crisera CA, Kadison AS, Breslow GD, Maldonado TS, Longaker MT, Gittes GK: Expression and role of laminin-1 in mouse pancreatic organogenesis.
Diabetes
49
:
936
–944,
2000
21.
Otonkoski T, Beattie GM, Mally MI, Ricordi C, Hayek A: Nicotinamide is a potent inducer of endocrine differentiation in cultured human fetal pancreatic cells.
J Clin Invest
92
:
1459
–1466,
1993
22.
Ricordi C, Lacy PE, Finke EH, Olack BJ, Scharp DW: Automated method for isolation of human pancreatic islets.
Diabetes
37
:
413
–420,
1988
23.
Ozmen L, Ekdahl KN, Elgue G, Larsson R, Korsgren O, Nilsson B: Inhibition of thrombin abrogates the instant blood-mediated inflammatory reaction triggered by isolated human islets: possible application of the thrombin inhibitor melagatran in clinical islet transplantation.
Diabetes
51
:
1779
–1784,
2002
24.
Bouwens L: Cytokeratins and cell differentiation in the pancreas.
J Pathol
184
:
234
–239,
1998
25.
Latif ZA, Noel J, Alejandro R: A simple method of staining fresh and cultured islets.
Transplantation
45
:
827
–830,
1988
26.
Hinegardner RT: An improved fluorometric assay for DNA.
Anal Biochem
39
:
197
–201,
1971
27.
Otonkoski T, Ustinov J, Rasilainen S, Kallio E, Korsgren O, Hayry P: Differentiation and maturation of porcine fetal islet cells in vitro and after transplantation.
Transplantation
68
:
1674
–1683,
1999
28.
Pulkkinen MA, Spencer-Dene B, Dickson C, Otonkoski T: The IIIb isoform of fibroblast growth factor receptor 2 is required for proper growth and branching of pancreatic ductal epithelium but not for differentiation of exocrine or endocrine cells.
Mech Dev
120
:
167
–175,
2003
29.
Beattie GM, Rubin JS, Mally MI, Otonkoski T, Hayek A: Regulation of proliferation and differentiation of human fetal pancreatic islet cells by extracellular matrix, hepatocyte growth factor and cell-cell contact.
Diabetes
45
:
1223
–1228,
1996
30.
Rajagopal J, Anderson WJ, Kume S, Martinez OI, Melton DA: Insulin staining of ES cell progeny from insulin uptake.
Science
299
:
363
,
2003
31.
Bouwens L, Lu WG, Dekrijger RR: Proliferation and differentiation in the human fetal endocrine pancreas.
Diabetologia
40
:
398
–404,
1997
32.
Bouwens L, Pipeleers DG: Extra-insular beta cells associated with ductules are frequent in adult human pancreas.
Diabetologia
41
:
629
–633,
1998
33.
Goudswaard WB, Houthoff HJ, Koudstaal J, Zwierstra RP: Nesidioblastosis and endocrine hyperplasia of the pancreas: a secondary phenomenon.
Hum Pathol
17
:
46
–54,
1986
34.
Steinert PM, Chou YH, Prahlad V, Parry DA, Marekov LN, Wu KC, Jang SI, Goldman RD: A high molecular weight intermediate filament-associated protein in BHK-21 cells is nestin, a type VI intermediate filament protein: limited co-assembly in vitro to form heteropolymers with type III vimentin and type IV alpha-internexin.
J Biol Chem
274
:
9881
–9890,
1999
35.
Githens S, Pictet R, Phelps P, Rutter WJ: 5-bromodeoxyuridine may alter the differentiative program of the embryonic pancreas.
J Cell Biol
71
:
341
–356,
1976
36.
Levenberg S, Golub JS, Amit M, Itskovitz-Eldor J, Langer R: Endothelial cells derived from human embryonic stem cells.
Proc Natl Acad Sci U S A
99
:
4391
–4396,
2002
37.
Lefebvre VH, Otonkoski T, Ustinov J, Huotari MA, Pipeleers DG, Bouwens L: Culture of adult human islet preparations with hepatocyte growth factor and 804G matrix is mitogenic for duct cells but not for β-cells.
Diabetes
47
:
134
–137,
1998
38.
Langhofer M, Hopkinson SM, Jones JCR: The matrix secreted by 804G cells contains laminin-related component that participate in hemidesmosome assembly in vitro.
J Cell Sci
105
:
753
–764,
1993
39.
Yonemura Y, Takashima T, Miwa K, Miyazaki I, Yamamoto H, Okamoto H: Amelioration of diabetes mellitus in partially depancreatized rats by poly (adp-ribose) synthetase inhibitors.
Diabetes
33
:
401
–404,
1984
40.
Bottaro DP, Rubin JS, Ron D, Finch PW, Florio C, Aaronson SA: Characterization of the receptor for keratinocyte growth factor: evidence for multiple fibroblast growth factor receptors.
J Biol Chem
265
:
12767
–12770,
1990
41.
Elghazi L, Cras-Meneur C, Czernichow P, Scharfmann R: Role for FGFR2IIIb-mediated signals in controlling pancreatic endocrine progenitor cell proliferation.
Proc Natl Acad Sci U S A
99
:
3884
–3889,
2002
42.
Sharma A, Zangen DH, Reitz P, Taneja M, Lissauer ME, Miller CP, Weir GC, Habener JF, Bonner-Weir S: The homeodomain protein Idx-1 increases after an early burst of proliferation during pancreatic regeneration.
Diabetes
48
:
507
–513,
1999
43.
Macfarlane WM, McKinnon CM, Felton-Edkins ZA, Cragg H, James RFL, Docherty K: Glucose stimulates translocation of the homeodomain transcription factor Pdx1 from the cytoplasm to the nucleus in pancreatic beta-cells.
J Biol Chem
274
:
1011
–1016,
1999
44.
Heimberg H, Bouwens L, Heremans Y, Van De Casteele M, Lefebvre V, Pipeleers D: Adult human pancreatic duct and islet cells exhibit similarities in expression and differences in phosphorylation and complex formation of the homeodomain protein Ipf-1.
Diabetes
49
:
571
–579,
2000
45.
Beattie GM, Otonkoski T, Lopez AD, Hayek A: Functional β-cell mass after transplantation of human fetal pancreatic cells: differentiation or proliferation?
Diabetes
46
:
244
–248,
1997
46.
Tyrberg B, Ustinov J, Otonkoski T, Andersson A: Stimulated endocrine cell proliferation and differentiation in transplanted human pancreatic islets: effects of the ob gene and compensatory growth of the implantation organ.
Diabetes
50
:
301
–307,
2001