p8 protein expression is known to be upregulated in the exocrine pancreas during acute pancreatitis. Own previous work revealed glucose-dependent p8 expression also in endocrine pancreatic β-cells. Here we demonstrate that glucose-induced INS-1 β-cell expansion is preceded by p8 protein expression. Moreover, isopropylthiogalactoside (IPTG)-induced p8 overexpression in INS-1 β-cells (p8-INS-1) enhances cell proliferation and expansion in the presence of glucose only. Although β-cell-related gene expression (PDX-1, proinsulin I, GLUT2, glucokinase, amylin) and function (insulin content and secretion) are slightly reduced during p8 overexpression, removal of IPTG reverses β-cell function within 24 h to normal levels. In addition, insulin secretion of p8-INS-1 β-cells in response to 0–25 mM glucose is not altered by preceding p8-induced β-cell expansion. Adenovirally transduced p8 overexpression in primary human pancreatic islets increases proliferation, expansion, and cumulative insulin secretion in vitro. Transplantation of mock-transduced control islets under the kidney capsule of immunosuppressed streptozotocin-diabetic mice reduces blood glucose and increases human C-peptide serum concentrations to stable levels after 3 days. In contrast, transplantation of equal numbers of p8-transduced islets results in a continuous decrease of blood glucose and increase of human C-peptide beyond 3 days, indicating p8-induced expansion of transplanted human β-cells in vivo. This is underlined by a doubling of insulin content in kidneys containing p8-transduced islet grafts explanted on day 9. These results establish p8 as a novel molecular mediator of glucose-induced pancreatic β-cell expansion in vitro and in vivo and support the notion of existing β-cell replication in the adult organism.
- islet transplantation
- insulin secretion
- β-cell proliferation
to date, it is controversial whether neogenesis from pancreatic precursor cells present in ducts (2) and islets of Langerhans (45) or self duplication of existing β-cells (12) contributes to the formation of new pancreatic β-cells in the adult organism. The debate highlights that mechanisms of β-cell homeostasis and regenerative repair are not well understood. In particular, knowledge about molecular regulation of pancreatic β-cell mass expansion is limited. Important examples of factors associated with β-cell expansion are the incretin hormone glucagon-like peptide (GLP)-1, the GLP-1 analogon exendin (Ex)-4 (28), and the cellular molecular mediator cyclin D2 (17).
Here we report on β-cell-expanding properties of the protein p8, which is considered to be a member of the high-mobility group (HMG)-I/Y transcription factor family despite low sequence homology (14). p8 is expressed in a broad range of tissues (25, 41) and is degraded via the ubiquitin/proteasome pathway (19). In rats, high fetal pancreatic p8 mRNA levels progressively decline during the postnatal period to reach low levels in the adult, indicating a role for p8 during development and organogenesis (25). In the adult rat, p8 expression is upregulated by pancreatic injury and sepsis, and p8−/− mice display increased pancreatitis-induced tissue damage and lipopolysaccharide (LPS)-induced mortality (21, 25, 38, 39). Consequently, p8 is believed to be part of a general cellular stress defense mechanism. In addition, several studies provide evidence that p8 is associated with cell expansion. For example, p8 amplifies Smad transactivation during transforming growth factor (TGF)-β1-induced proliferation of mouse embryonic fibroblasts (MEF) (16) and is required for endothelin-induced mesangial cell hypertrophy in the diabetic kidney (18). Remarkably, monkey COS-7 kidney, rat AR42J pancreatic acinar, and human HeLa cervix epithelial cells with p8 overexpression display enhanced growth compared with mock-transfected controls (25, 41).
Within the pancreas, p8 expression had been reported only in exocrine cells (25, 27) until our previous work (31) demonstrated that p8 is also expressed in primary human islets of Langerhans and several pancreatic cell lines including ductal cells and differentiated β-cells. We further confirmed an association between glucose-stimulated p8 expression and expansion of rat INS-1 β-cells. This observation supports the hypothesis that p8 may be involved in pancreatic β-cell mass expansion not only during development and regenerative tissue repair but also in response to nutrients.
To examine the potential pancreatic β-cell mass-expanding properties of p8 in more detail, we correlated glucose-induced p8 protein expression to cell numbers in expanding, nonconfluent, rat pancreatic INS-1 β-cell cultures. To assess the capability of p8 to increase β-cell mass, we generated transgenic INS-1 β-cells with isopropylthiogalactoside (IPTG)-inducible p8 overexpression and examined p8-induced proliferation and expansion as well as β-cell-related gene expression and insulin secretory function. Furthermore, primary human islets of Langerhans were adenovirally transduced with a p8 expression vector, and transduced islets were tested for proliferation, expansion, and cumulative insulin secretion in vitro. p8-transduced islets were subsequently xenotransplanted under the kidney capsule of immunosuppressed streptozotocin (STZ)-diabetic C57BL/6 mice. Posttransplantation, blood glucose, human C-peptide levels, and body weight were monitored for 8 days, and, finally, insulin content of kidneys explanted on day 9 was measured as a surrogate parameter of β-cell expansion.
MATERIALS AND METHODS
INS-1 β-cell culture.
Basal medium for the rat INS-1 pancreatic β-cell line (kindly provided by C. Wollheim, Geneva, Switzerland) was as follows: RPMI 1640 containing 10 mM HEPES, 1 mM sodium pyruvate, 2 mM l-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, and 50 μM β-mercaptoethanol. The medium was supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 11.1 mM glucose unless otherwise stated.
Cell count and Western blotting.
Cells (1.5 million/60-mm dish) were seeded for correlation of cell count and p8 protein expression. Cell numbers were assessed in a Neubauer chamber. p8 protein expression was assessed by Western blotting of whole cell extracts as previously described (22). Equal amounts of protein were loaded per lane, size fractionated on 10–20% Tris-Tricine Ready Gels (Bio-Rad Laboratories, Munich, Germany), and semi-dry blotted. For detection of p8, a polyclonal p8-antiserum was generated in rabbits using a bacterially expressed full-length p8-glutathione sulfonyl transferase (GST) fusion protein as antigen. Visualization was achieved by a horseradish peroxidase-linked donkey anti-rabbit antibody and ECL Plus Western Blotting Detection Reagent (both from Amersham Biosciences Europe, Freiburg, Germany). After visualization, blotted p8 protein was quantified by densitometric scanning as previously described (34).
Generation, culture, and expansion of IPTG-inducible p8-INS1 β-cells.
Inducible p8 overexpression was achieved by the original LacSwitch system (Stratagene Europe, Amsterdam, The Netherlands) composed of a p3'SS lac repressor vector and the pOPRSV1-CAT lac operator vector. A full-length rat p8 cDNA was cloned as a NotI fragment into the chloramphenicol acetyltransferase (CAT) region flanked by NotI sites of the pOPRSV1-CAT vector. The CAT sequence of the resulting pOPRSV1-p8 vector was thereby removed. INS-1 β-cells were transfected with the NaeI linearized p3'SS vector using Lipofectamine (Invitrogen, Karlsruhe, Germany) and selected for 4 wk with 71.6 μg/ml hygromycin. The resulting stably transfected cells were transfected again with the SacI linearized pOPRSV1-p8 vector and further selected with hygromycin and 0.18 mg/ml neomycin (G418) for another 4 wk. Double-selected p8-INS1 single cell clones were established and expanded in INS-1 basal medium supplemented with 35.8 μg/ml hygromycin, 0.09 mg/ml G418. Clones were tested by stimulation with 5.5 mM IPTG. IPTG-induced p8 expression was confirmed by RT-PCR, and a p8-INS1 clone with sixfold overexpression was chosen for further experiments.
Proliferation of 100,000 p8-INS-1 β-cells/well seeded in 12-well plates was assessed by 24-h incubation in serum-free medium in the presence or absence of 5.6 mM glucose (physiological concentration) and 5.5 mM IPTG. Serum-free conditions were chosen to exclude induction of β-cell expansion by serum factors like insulin-like growth factor (IGF)-I and growth hormone (8, 20). [3H]thymidine was added to the medium during the last 4 h of the 24-h incubation period. [3H]thymidine incorporation was measured by liquid scintillation counting. For determination of cell numbers in the presence or absence of 5.5 mM IPTG, 120,000 p8-INS1 β-cells/well were seeded in 12-well plates and cultured for up to 4 days in a reduced medium containing 2.8% FBS and 5.6 mM glucose. In contrast to proliferation experiments, the medium was supplemented with 2.8% FBS, since serum-free maintenance for longer than 48 h resulted in progressive cell death.
Gene expression in p8-INS-1 β-cells.
p8-INS1 β-cells (100,000/well) were seeded in six-well plates and cultured for 4 days in the presence, and subsequently another 2 days in the absence, of 5.5 mM IPTG. Stimulation medium was changed every 12 h. Total RNA was isolated at different time points using TRIzol (Invitrogen) and tested by rat β-actin PCR (for primer oligonucleotide details, see below) for genomic DNA contamination. Samples free of genomic DNA were reverse transcribed into cDNA with Superscript II using oligo(dT) primers (both Invitrogen). General PCR conditions were as follows: initial denaturing for 2 min at 94°C followed by template-specific numbers of cycles [10 s at 94°C, primer-specific ambient temperature (Ta) for 10 s, 30 s at 72°C] and 2-min final elongation at 72°C. Primer sequences and specific PCR conditions were as follows: rat p8, forward 5′-ccaggtatgatggccaccttgccaccaac-3′ and reverse 5′-ggtgtggtgtccgtggtctggcctcatctcc-3′ (Ta, 60°C; 24 cycles; product length, 292 bp); rat protein encoded by the pancreatic and duodenal homeobox gene (PDX)-1, forward 5′-cctccgccgacaccccagtttg-3′ and reverse 5′-caccgcccccgctcgttgtc-3′ (Ta, 62°C; 22 cycles; product length, 526 bp); rat proinsulin I, forward 5′-ctgcccaggcttttgtca-3′ and reverse 5′-cagaggggtgggcggggagtggt-3′ (Ta, 56°C; 20 cycles; product length, 299 bp); rat glucose transporter-2 (GLUT2), forward 5′-gctggggtggtcctggtctt-3′ and reverse 5′-gcttataacgtgctttccctgtag-3′ (Ta, 54°C; 24 cycles; product length, 433 bp); rat glucokinase, forward 5′-tatgggcgagctggtacgacttgt-3′ and reverse 5′-ctgccgctgccctcctctgatt-3′ (Ta, 61°C; 24 cycles; product length, 450 bp); rat amylin (islet amyloid precursor polypeptide; IAPP), forward 5′-ccagtcctcccaccaaccaa-3′ and reverse 5′-catcccacgccctcctttatt-3′ (Ta, 54°C; 20 cycles; product length, 409 bp); rat β-actin, forward 5′-tacaacctgcttgcagctcc-3′ and reverse 5′-ggatcttcatgaggtagtctgtc-3′ (Ta, 60°C; 18 cycles; product length, 630 bp). Resulting PCR products were size fractionated in agarose gels, stained with ethidium bromide, and visualized under ultraviolet light.
Insulin content and secretion of p8-INS1 β-cells.
p8-INS1 β-cells (100,000/well) were seeded in six-well plates and cultured for 4 days in the presence, and subsequently another 2 days in the absence, of 5.5 mM IPTG. Stimulation medium was changed every 12 h. Insulin secretion and content were measured in individual wells at the time points indicated. Before analysis of insulin secretion (supernatants) and content [insulin extraction from cells by ethanol-H2O-HCl (7.5:2.35:1.5) at 4°C overnight], cells were incubated for 1 h in INS-1 basal medium without FBS but with 0.1% bovine serum albumin (BSA). Insulin concentrations of samples were analyzed using an ultrasensitive rat insulin ELISA (Mercodia, Uppsala, Sweden). For determination of insulin secretion in response to glucose, p8-INS1 β-cells were cultured in the presence of 2.8 mM glucose for 7 days. During the last 4 days, medium was additionally supplemented with 5.5 mM IPTG or vehicle. Cells were then harvested and seeded in medium with 2.8 mM glucose at a density of 10,000 cells/well in 12-well plates. The next day, cells were incubated for 1 h in serum-free INS-1 basal medium supplemented with 0.1% BSA and glucose in different concentrations (0–25 mM). Supernatants were analyzed for insulin concentrations using an ultrasensitive rat insulin ELISA (Mercodia).
Recombinant p8-expressing adenovirus was generated using Adeno-X adenoviral expression system and Adeno-X Rapid Titer Kit (both BD Biosciences-Clontech, Heidelberg, Germany) according to the manufacturer's protocols. Recombinant p8-adenoviruses express the human full-length cDNA of p8 under the control of a cytomegalovirus promoter. The control adenovirus (mock) does not contain p8 cDNA. Adenoviral titers were determined by plaque assay and expressed as plaque-forming units (PFU).
Human pancreatic islet isolation and culture.
Islets were isolated from human pancreases obtained from brain-dead multiorgan donors after obtaining legal consent. Approval from the local ethics committee was obtained. The islets were isolated at the Giessen Islet Isolation and Transplantation Center (Giessen, Germany) according to a modified semiautomated digestion-filtration method and cultured as previously described (3, 33). Human islets of Langerhans were routinely cultured in medium as described above for INS-1 β-cells but supplemented with 5.6 mM glucose using noncoated 10-cm dishes in suspension culture.
Human pancreatic islet transduction and in vitro characterization.
Purified human islets (100/well) were seeded in noncoated 12-well plates in suspension culture and transduced overnight by 106 PFU of p8 or mock adenovirus, respectively. For analysis of p8 overexpression, total RNA was isolated and tested for genomic DNA contamination as described for gene expression in p8-INS1 β-cells. PCR was performed using specific primers for human p8 (forward 5′-ccaggcacgatggccaccttcccaccag-3′ and reverse 5′-gcctcatctccagctctgtctcagcgcc-3′; Ta, 60°C; 24 cycles; product length, 278 bp) and general PCR conditions, as described for gene expression in p8-INS1 β-cells. Resulting PCR products were size fractionated in agarose gels, stained with ethidium bromide, and visualized under ultraviolet light. On days 0, 1, 3, 5, and 7, cumulative insulin levels in the supernatants were analyzed using a human insulin ELISA kit (Linco, St. Charles, MO), and islets were dispersed with 0.5 g/l trypsin and 0.2 g/l EDTA (Invitrogen) for determination of cell numbers and percentage of bromodeoxyuridine (BrdU)-positive cells by fluorescence-activated cell sorting using a BrdU Kit (Roche Diagnostics, Mannheim, Germany).
C57BL/6 mice (8- to 10-wk old) were injected (ip) with a single dose of 200 mg/kg STZ in citrate buffer (pH 4.5). Animals with serum glucose concentrations >300 mg/dl after overnight fasting (day 3, post-STZ injection) were considered to be diabetic. The day before transplantation, diabetic mice were injected with anti-lymphocyte globulin (Accurate Chemical Scientific, Westbury, NY). After exposure to methoxyflurane (Metofane; Schering Plough Animal Health, Union, NJ) vapors for rapid anesthesia, 500 islets/animal were transplanted under the left kidney capsule as previously described (44). Six mice received p8, and another six mice were transplanted with mock-transduced islets. The mice were housed in a pathogen-free environment on a 12:12-h light-dark cycle and fed a standard rodent chow. The Institutional Animal Care and Use Committee approved all the studies.
In vivo characterization of transplanted mice and ex vivo characterization of islet grafts.
During the experimental period, body weight, serum glucose, and human C-peptide levels were monitored daily. Glucose and human C-peptide levels were determined from tail blood using a 2300 STAT glucose analyzer (Yellow Springs Instrument, Yellow Springs, OH) and a human-specific C-peptide ELISA kit (Linco), respectively. On day 9, mice were killed, and kidneys were explanted. Insulin was extracted from minced kidneys with ethanol-H2O-HCl (7.5:2.35:1.5) at 4°C overnight for determination by ELISA (Linco).
Glucose-dependent expansion of native INS-1 β-cells is preceded by enhanced p8 protein expression.
To explore the association between glucose-induced p8 expression and cell proliferation in INS-1 β-cells (31) in more detail, p8 protein expression was directly correlated to cell numbers during 6 days of culture in the presence of standard (11.1 mM) and high (25 mM) glucose concentrations in the medium (Fig. 1). Standard culture conditions with 11.1 mM glucose continuously induced p8 protein expression with a peak on day 4 (Fig. 1A). As expected, high glucose concentrations in the medium increased p8 protein expression more rapidly, reaching a maximum on day 2 and subsequently decreasing to a stable lower level (Fig. 1B). In both settings, cell numbers peaked 2 days postmaximal p8 expression, when cultures reached confluency. Subsequently, cell numbers declined to a lower stable level as observed for p8 protein expression, which indicates regulation of cell growth by contact inhibition. These results suggest that glucose-induced p8 protein expression is counterregulated by the density of the monolayer, since its peak precedes maximum β-cell expansion.
IPTG-induced p8 overexpression in p8-INS1 β-cells enhances proliferation and expansion in a glucose-dependent manner.
To further test the assumption that p8 is involved in glucose-induced β-cell expansion, we generated p8-INS1 β-cells with IPTG-inducible p8 overexpression (see Fig. 3 for demonstration of p8 mRNA overexpression). Maintenance of p8-INS1 β-cells in serum- and glucose-free medium resulted in complete growth stop as reflected by [3H]thymidine incorporation values that were close to background levels (Fig. 2A). Under these conditions, induction of p8 overexpression by IPTG did not significantly increase [3H]thymidine incorporation. In contrast, cell proliferation at a physiological glucose concentration of 5.6 mM was substantially increased by IPTG-induced p8 overexpression (P < 0.05) (Fig. 2A). These results strongly indicate that p8-mediated INS-1 β-cell proliferation depends on the presence of glucose.
Because serum-free culture of p8-INS1 β-cells for >48 h resulted in progressive cell death, the medium containing 5.6 mM glucose was additionally supplemented with 2.8% FBS for assessment of cell numbers during 4 days of expansion ± 5.5 mM IPTG. After 4 days, cell numbers increased threefold in the presence of physiological glucose concentrations without p8 overexpression (absence of IPTG). In contrast, in the presence of IPTG (p8 overexpression), cell numbers on day 4 were elevated by about fourfold. Cell numbers on day 4 were significantly higher (P < 0.01) in IPTG-treated cultures compared with control cultures without IPTG (Fig. 2B).
Transient IPTG-induced p8 overexpression affects pancreatic β-cell gene expression without permanent loss of phenotype in p8-INS1 β-cells.
To determine possible effects of p8 overexpression on pancreatic β-cell phenotype, mRNA expression of important β-cell-specific genes was monitored during and post-5.5 mM IPTG-induced p8 overexpression (Fig. 3). It is noteworthy that PDX-1 is a transactivator of other important β-cell genes such as proinsulin (29), GLUT2 (42), glucokinase (43), and amylin (also known as IAPP) (7). IPTG supplementation stimulated stable p8 mRNA expression (elevated sixfold) in p8-INS1 β-cells within 24 h, whereas mRNA levels of PDX-1, proinsulin, GLUT2, glucokinase, and amylin slightly declined in parallel to constant levels. These results suggest that p8-induced β-cell expansion is associated with reduced β-cell-specific gene expression during the proliferative phase. However, recurrence of original mRNA expression levels for all analyzed genes within 24 h after removal of IPTG strongly indicates no loss of β-cell phenotype due to transient p8 overexpression. Thus reduction of pancreatic β-cell gene expression likely may be the result of enhanced mitosis caused by p8 overexpression as demonstrated in Fig. 2. As a control, β-actin mRNA expression did not change during the investigated culture period.
Cell expansion mediated by transient p8 overexpression does not affect glucose-induced insulin secretion in p8-INS1 β-cells.
To investigate effects of p8 overexpression on β-cell insulin secretory functions during cell expansion, insulin content and secretion were monitored during and after 4 days of exposure of p8-INS1 β-cells to IPTG (Fig. 4). Induction of p8 overexpression by IPTG supplementation resulted in slightly reduced insulin content and secretion, by ∼17%, on days 3 and 4 at the single cell level (Fig. 4A). Similar to the expression of pancreatic β-cell-specific genes (Fig. 3), reduced insulin content and secretion per cell recovered to preexpansion levels within 24 h after removal of IPTG. More importantly, 4 days of exposure to IPTG-induced p8 overexpression and cell expansion did not alter glucose-induced insulin secretion (Fig. 4B). No differences in 0–25 mM glucose-induced insulin responses were detectable between IPTG/p8-expanded and untreated cells. Fast recovery of moderately reduced insulin content and secretion per cell after removal of IPTG strongly suggests that transient p8 overexpression does not result in loss of pancreatic β-cell secretory function.
Adenoviral p8 overexpression enhances cell proliferation and cumulative insulin secretion in human islets of Langerhans in vitro.
Adenoviral transduction was used, because no similarly effective transfection protocols for primary human islets are available currently. Before transplantation of the transduced islets into STZ-diabetic C57BL/6 mice, a proportion of the islets was characterized in vitro (Fig. 5). Compared with mock controls, p8-transduced islets consistently displayed elevated p8 mRNA levels on days 1, 3, 5, and 7 after transduction (Fig. 5A). Mock control levels of p8 transcripts on day 7 were similar to those on day 1 (not shown). Upregulation of p8 expression resulted in a continuous elevation of both BrdU-positive cells (Fig. 5B) and cell numbers (Fig. 5C), with final magnitudes being 2.8- and 6.2-fold higher, respectively, on day 7 relative to mock control levels (both P < 0.001 on days 1, 3, 5, and 7). Moreover, higher cell numbers of p8-transduced islets were paralleled by elevated cumulative insulin concentrations in the supernatants (day 1, P < 0.05; days 3, 5, and 7, P < 0.001) (Fig. 5D). This increased cumulative net insulin output of p8-transduced primary human pancreatic islets indicates enhanced replication within the islet β-cell fraction.
Transplantation of p8-transduced human islets of Langerhans into STZ-diabetic C57BL/6 mice enhances glycemic control in vivo.
To evaluate the ability of p8-transduced human pancreatic islets to reduce hyperglycemia in diabetic animals, mock- and p8-transduced primary human islets of Langerhans were transplanted under the left kidney capsule of STZ-diabetic C57BL/6 mice (Fig. 6). On days −1 and 0, serum glucose levels of all investigated mice were >300 mg/dl, demonstrating overt diabetes (Fig. 6A). Amelioration of hyperglycemia was observed posttransplantation in both the mock and the p8 group. Blood glucose concentrations of animals transplanted with mock-transduced islets and p8-transduced islets decreased similarly during days 1 and 2, demonstrating functional insulin secretion by transplanted β-cells within the transduced grafts. On day 3, serum glucose concentrations in mice transplanted with mock-transduced human pancreatic islets reached a stable level of ∼200 mg/dl (mean ± SD on day 8, 178.3 ± 23.1 mg/dl). In contrast, mice transplanted with p8-transduced islets displayed a continuous amelioration of hyperglycemia (day 3, P < 0.05; day 4, P < 0.01; days 5–8, P < 0.001) to blood glucose levels of ∼70 mg/dl on day 8 (mean ± SD, 70.0 ± 21.5 ml/dl), indicating enhanced secretory function of p8-transduced islet grafts in vivo, possibly due to expansion of transplanted insulin-producing cells within the grafts. As an additional control for the in vivo function of the islet transplants, human C-peptide was measured in the same blood samples from which glucose values were determined (Fig. 6B). Inversely to glucose clearance posttransplantation, human C-peptide serum concentrations rose steadily to a stable level on days 3–8 (mean ± SD on day 8, 0,89 ± 0,076 nmol/l) in the mock group but increased continuously in the p8 group on days 5–8 (P < 0.001; mean ± SD on day 8, 1,46 ± 0,083 nmol/l). This continuous rise of human C-peptide serum concentrations in the animals transplanted with p8-transduced human pancreatic islets further supports the notion of enhanced insulin secretory function of p8-transduced islet grafts in vivo. On day 9, mice were killed, and kidneys were explanted for analysis of insulin content. In line with serum glucose and C-peptide levels, mean normalized insulin content of the transplanted kidney [left (containing islet graft) minus right (without graft) kidney] was twofold above mock control levels in the p8 group (P < 0.001; mean ± SD; mock, 92.5 ± 14 μU/ml, vs. p8, 183 ± 16.2 μU/ml) (Fig. 6C). Transplantation procedures did not affect general nutritional status differentially in the animal groups, because dynamics of mean body weight regulation were similar after transplantation in both experimental cohorts (Fig. 6D).
An appropriate amount of pancreatic β-cells is essential for functional insulin responses to adjust blood glucose to physiological levels. Along this line, glucose and insulin themselves are well known mediators regulating the adaptation of β-cell mass to its functional demands (1, 30). Beyond the major function as a proliferative agent for β-cells, glucose is also an essential prerequisite for IGF-I- and growth hormone-induced expansion of these cells (8, 20). Our study now suggests that glucose-stimulated β-cell mass expansion may be, at least in part, mediated by the protein p8.
The presented results demonstrate that p8 expression is induced in a glucose-dependent manner in expanding pancreatic INS-1 β-cells and downregulated as soon as the density of the INS-1 cell monolayer reaches a status of dense cell contacts with subsequent contact inhibition of cell growth. IPTG-induced p8 overexpression in p8-INS1 β-cells resulted in enhanced proliferation and expansion rates. It is important to note that this induction of cell proliferation in p8-INS1 cells was strongly dependent on the presence of glucose in the medium. At 0 mM glucose, p8 overexpression on its own was not able to increase the proliferation rate of p8-INS1 cells, whereas, in the presence of a physiological glucose concentration of 5.6 mM, p8 overexpression substantially enhanced [3H]thymidine incorporation. These results provide strong evidence that p8 is a molecular mediator of glucose-dependent pancreatic β-cell expansion. In addition, it is important to note that p8 expression by itself is regulated by glucose, as we have demonstrated earlier (31).
These results, together with existing studies demonstrating augmented growth of several p8-overexpressing cell lines (25, 41), indicate that p8 expression is not merely associated with increasing pancreatic β-cell mass; rather, the results suggest that this molecule is a mediator of β-cell expansion on its own. Detailed characterization of IPTG-induced p8 overexpression in INS-1 β-cells revealed a decrease in β-cell-specific gene expression as well as insulin biosynthesis and secretion during p8-mediated cell expansion. It is noteworthy that these alterations are restored to preexpansion levels after termination of p8 overexpression by removal of IPTG. Therefore, the observed reversible reduction of gene expression and function in β-cells during p8 overexpression may be the consequence of a transiently enhanced mitosis rate rather than a permanent dedifferentiation.
However, on the basis of current literature, p8 seems to exert some apparently contradictory growth-related actions. In contrast to augmentation of cell proliferation by p8 overexpression in other cellular systems, MEF from p8−/− mice and human pancreatic Panc-1 and BxPc-3 cells with anti-sense-silenced p8 expression grow more rapidly than p8-expressing controls (24, 40). These divergent observations may be attributed to different effects of p8 in distinctive segments of the cell cycle. Flow cytometry studies revealed that p8 overexpression increases cell numbers of fast-growing HeLa cells in the S and G2/M phases (41). In contrast, 10−7 M 1,25-dihydroxyvitamin D3-induced transient p8 expression in human MCF7/LCC2 breast cancer cells was associated with reduced cell growth and cell cycle arrest in the G1 phase (4). Interestingly, growth of MCF7/LCC2 cells was inhibited by high concentrations (≥10−8 M) but stimulated by lower concentrations of 1,25-dihydroxyvitamin D3. This indicates that distinct modes of p8 function on cell growth may be related to the status of the cell cycle in a given cell, the intensity of external stimuli, and the cellular microenvironment. Thus the use of varying conditions and cell types may be a plausible explanation for the heterogeneity of p8-associated cellular responses in certain experimental settings.
In concordance with the observed growth-enhancing properties of p8 in the INS-1 pancreatic β-cell line, we also found a substantial increase of cell proliferation and expansion due to adenovirally transduced p8 overexpression in cultured primary human islets of Langerhans. Moreover, net cumulative basal insulin secretion was augmented in p8-transduced compared with mock-transduced human islets. However, the relative magnitude of enhanced cumulative insulin secretion in p8-transduced islets was significantly smaller than that observed for rates of proliferation and expansion. Two explanations for this observation may be envisioned. First, as demonstrated for p8-INS1 β-cells during IPTG/p8 overexpression, transient reduction of β-cell-specific gene expression and insulin secretory function in response to an enhanced mitosis rate may also occur in p8-overexpressing primary human pancreatic islets. Second, results of some studies using adenoviral transduction of intact primary pancreatic islets suggest that this method displays a preference for gene transfer into islet cells that lie in the periphery of the pancreatic islet, where a substantial fraction of islet non-β-cells is located (11). Thus the p8 adenovirus has targeted a major population of the islet non-β-cells in the outer rim and a smaller proportion of β-cells in the core of the islets. Increases in cell numbers and BrdU-positive cells in p8-transduced human islets may be, therefore, due in part to p8-induced proliferation of non-β-cells. However, it has also been demonstrated in several studies that adenoviral transduction of human pancreatic islets is able to provide gene transfer into β-cells within the islet core, especially in smaller islets (13). With regard to our study, we strongly believe that a significantly higher cumulative insulin secretion in adenovirally p8-transduced human pancreatic islets can only be explained by targeting of a substantial proportion of islet β-cells by the p8 adenovirus. Because p8 overexpression in p8-INS1 β-cells did not functionally induce, but rather moderately reduced, insulin secretory responses per cell, it is strongly suggested that higher cumulative insulin secretion in p8-overexpressing human islets was due to p8-induced expansion of the islet β-cell population.
The transplantation experiments of p8-transduced human primary islets into STZ-diabetic mice imply that the p8-mediated enhanced β-cell expansion observed in vitro also persisted in vivo. As a consequence, the hyperglycemic state of transplanted animals was continuously ameliorated to almost hypoglycemic levels within 8 days. In contrast, blood glucose concentrations of mice receiving equal amounts of mock-transduced islets were only moderately reduced to a mild diabetic level, indicating proper function of the initially transplanted islet cells but no further enhancement of β-cell secretory function in vivo.
One might argue that amelioration of elevated blood glucose levels measured in mice that were transplanted with p8-transduced islets is caused by functional alterations and/or an increase of the remaining endogenous β-cell mass rather than expansion of transplanted β-cells by p8 overexpression. However, human C-peptide levels in transplanted mice that were not detectable on days −1 and 0 of the experiment steadily increased with time posttransplantation in mice receiving p8-transduced islets, whereas mice receiving mock-transduced islets displayed stable human C-peptide concentrations between days 3 and 8 posttransplantation. This observation indicates that p8 mediates an enhanced insulin secretory response that seems to be due to p8-induced β-cell expansion in vivo. This notion is underlined by a doubling of insulin content of explanted kidneys bearing the p8-transduced islet grafts compared with kidneys harboring mock-transduced islets. Because transient IPTG/p8-induced expansion of p8-INS-1 β-cells did not alter single cell insulin response to glucose in vitro, it seems unlikely that adenovirally transduced p8 overexpression resulted in increased insulin secretion per single human islet β-cell.
In several animal models, p8 is strongly upregulated during experimentally induced pancreatitis, sepsis, and liver damage (21, 25, 37–39). Remarkably, p8−/− mice display increased pancreatitis-induced cell destruction, LPS-induced mortality, and sensitivity of the liver to the hepatotoxic effects of carbon tetrachloride (CCL4). As an underlying mechanism, at least within the exocrine pancreas, a p8-dependent upregulation of the pancreatitis-associated protein (PAP) I and subsequent PAP I-induced inhibition of TNFα-stimulated NF-κB activation was suggested (38). In addition, p8 is involved in downregulation of staurosporine-induced caspase-3/7 and -9 activities as part of a complex with prothymosin-α (23). Thus it could be argued that the p8-transduced islets used in our murine transplantation model benefit from these protective properties of p8 rather than from its proliferation-inducing actions. If this were true, one would expect a continuously fading function of the mock-transduced islet transplants. However, stable blood glucose and C-peptide levels from day 3 to day 8 in mice receiving mock-transduced islets strongly suggest persistent cell survival and proper function within the investigated experimental period. Therefore, we believe that, in this experimental setting, p8-induced expansion of graft β-cells may represent the primary mechanism for improved secretory responses of p8-overexpressing islet grafts as opposed to possible p8-mediated effects on cell survival or protection from apoptosis.
Not unexpectedly, for a proliferative factor, p8 is also upregulated in expanding pancreas and breast cancer cells (32, 35, 36). An association of p8 expression with tumor growth has further been demonstrated in several studies in vitro and in animal models. For example, p8 expression levels in clonal sublines of rat GH3 somatolactotrope pituitary cells positively correlate with tumor size in a nude mice model (26). Consistently, MEF from p8−/− mice transformed with mutated rasV12 protein and EA1 oncogene, as well as p8-negative human MCF7 breast cancer cells and p8 anti-sense-treated mouse LβT2 gonadotrope pituitary cells, failed to form colonies in vitro or tumors in vivo (4, 26, 41). Because of these observations, p8 was suggested to play a role in tumor development.
However, the results from our studies provide no evidence for an oncogenic character of the p8 protein, because endogenous p8 expression levels in expanding native INS-1 β-cells were counterregulated by contact inhibition, which would not be typical for a tumorigenic protein. Moreover, the observation that p8-INS1 β-cells that were expanded by transient p8 overexpression display no alterations in glucose-induced insulin secretion suggests that high cellular p8 levels do not result in dedifferentiation and malignant transformation of β-cells. The maintenance of a differentiated cellular phenotype despite p8 overexpression is further suggested by the result that p8-transduced human islets of Langerhans reduce hyperglycemia more efficiently with time in STZ-diabetic mice compared with mice transplanted with the same amount of mock-transduced islets, at least during the experimental period of 8 days. Of note, because the data presented here were derived from short-term experiments designed to demonstrate proof of principle of the potential of p8 to generally mediate pancreatic islet mass expansion, further long-term experiments are required to analyze tumorigenic and hypoglycemic propensity of transplanted p8-transduced pancreatic β-cells or islets.
With regard to a possible underlying molecular mechanism of p8-induced β-cell mass adaptation, it is known that the amino acid sequence of p8 contains two predicted PKC phosphorylation sites (6). β-Cell proliferation in response to glucose is mediated through activation of phosphatidylinositol 3-kinase (PI 3-kinase) as well as PKB and also PKC isoforms (9, 10, 20). Glucose concentrations can also modulate the proliferative response of β-cells to other hormonal ligands (5) by activation of insulin receptor substrate (IRS)-4 (15) and PKC (20). In recent experiments, we have obtained evidence that the protein p8 is indeed constitutively associated with PKC and binds to PI 3-kinase in a glucose-dependent manner in pancreatic β-cells (unpublished observations). Therefore, at the molecular level, glucose-dependent pancreatic β-cell expansion may be mediated through PKC- and/or PI 3-kinase-dependent phosphorylation of p8.
Taken together, we conclude that p8 represents a glucose-regulated mediator of pancreatic β-cell expansion and thereby may play a role in the adaptation of pancreatic β-cell mass to glycemic demands. If future long-term experiments confirm that p8 overexpression does not alter β-cell function and phenotype, this protein would represent a suitable molecular target for the expansion of these cells from limited human donor material as an alternative source for islet transplantation. This is of importance, since, to date, no effective protocols for the in vitro generation of insulin-producing cells from adult progenitors exist. In summary, our results characterize p8 as a novel glucose-regulated molecular mediator that, when overexpressed, is able to stimulate β-cell expansion. Finally, these results support the notion of existing β-cell replication in the adult organism.
This study was supported by the Juvenile Diabetes Research Foundation (Grant 5-2000-12 to J. Seufert) and the Bundesministerium für Bildung und Forschung of Germany (Grant FKZ-01GN0114 to M. D. Brendel and R. G. Bretzel, and Grant FKZ-01GN0115 to J. Seufert). G. Päth was supported by the Interdisciplinary Center of Clinical Research at the University of Würzburg, Germany (Grant Z-4/57 and z-4/66).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by American Physiological Society