Pax-6 Activates Endogenous Proglucagon Gene Expression in the Rodent Gastrointestinal Epithelium

The mouse small bowel-derived intestinal cell line, STC-1 cells (22), were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% horse serum and 2.5% FBS. GLUTag cells, derived from a murine large intestinal glucagonoma, were cultured in DMEM containing 10% FBS as described previously (23). Caco-2 intestinal cells and HEK 293 and baby hamster kidney fibroblast (BHK) cell lines were obtained from American Type Culture Collection (ATCC, Gaithersburg, MD). Cells were trypsinized, distributed into 6-cm plates, and grown to ∼90% confluence before transfection or adenoviral transduction. Intestinal epithelial cells were isolated from embryonic day 19–20 fetal rats and cultured in DMEM containing 5% FBS as described (24). Primary intestinal cell cultures were incubated at 37°C under 5% CO₂ and 95% O₂ atmosphere for 24 h before viral transduction studies. Briefly, cell lines or primary cells were incubated with adenoviruses for 1 h and then washed and incubated for an additional 24 h in a culture medium containing 5% FBS before harvesting for RNA or protein analysis.

cDNAs encoding for wild-type hamster Cdx-2/3, the dominant negative Cdx-2/3 (Cdx-2/3-HD) encoding the homeodomain and COOH-terminal peptide (180–313 amino acids) (30), and mouse Pax-6 and Sey^neu genes (kind gifts from Dr. M.S. German, University of California, San Francisco, CA) were subcloned into pACCMV-pLpA vector and cotransformed with pJM17 plasmid into 293 cells (25). Successful homologous recombination between these two plasmids resulted in recombinant viruses encoding the specific transcription factors (Ad-Pax-6, Ad- Sey^neu, and Ad-Cdx-2/3, Ad-Cdx-2/3-HD) and the control Ad-βgal. After plaque purification, viruses were amplified in 293 cells and purified by CsCl gradient centrifugation to enrich the viral stocks as previously described (25). Plaque-forming units (PFUs) were assayed using the Adeno-X rapid titer kit (Clontech, Palo Alto, CA). Adenoviral titrations were 2.2 × 10¹⁰ 1 PFU/ml and 2.5 × 10¹⁰ PFU/ml for Ad-Pax-6 and Ad-Cdx-2/3, respectively. Similar viral titers were obtained for Ad-βgal (2.0 × 10¹⁰ PFU/ml) and Ad-Sey^neu (2.1 × 10¹⁰ PFU/ml). In vitro transductions of viruses into STC, GLUTag, 293, or Caco-2 cell lines were completed after 1-h incubation at 37°C, whereas in BHK cells the transduction time was extended to 2 h.

Total RNA was isolated by the Trizol method (Life Technologies, Gaithersburg, MD) from STC-1, GLUTag, primary intestinal cell cultures, or rat intestine. RNA was electrophoresed on a 1.2% formaldehyde agarose gel, transferred to Hybond-N plus, and hybridized to ³²P-labeled glucagon cDNA probes as described (26,27).

Semiquantitative RT-PCR was used to determine expression levels of Cdx-2/3, Pax-6, and proglucagon according to the protocol from Clontech (Palo Alto, CA). The following primers were used for PCR amplification: 5′-AAGAGTGGCGACTCCAGAAGTTG-3′ and 5′-ACCACACCTGTATCCTTGCTTCAG-3′ for Pax-6; 5′-AGCTTCAGTCCCACAAGGCAGAAT-3′ and 5′-GGCCGCAGGAGATGTTGTGAAGAT-3′ for proglucagon; 5′-ATGTACGTGAGCTACCTTCTGGAC-3′ and 5′-GGAGCCACCATTGAGGCCGTGGGC-3′ for Cdx-2/3; 5′-CATGACCACAGTCCATGCCATC-3′ and 5′-CACCCTGTTGCTGTAGCCATATTC-3′ for the control transcript GAPDH. A forward primer 5′-GTGTGTACTTGTGGAGCACTGCCC-3′ priming within intron-B sequences of mouse proglucagon gene and a reverse primer 5′-CGGTTCCTCTTGGTGTTCATCAACC-3′ locating in the exon-3 were used in RT-PCR to quantitate newly synthesized proglucagon primary RNA transcripts in GLUTag cells transduced with Ad-βgal or Ad-Cdx-2/3-HD. For RT-PCR experiments shown in Fig. 5, specific primers were designed to amplify cDNA fragments of 297 and 545 bp for Cdx-2/3 and Pax-6, respectively. Semiquantitative measurements of newly synthesized proglucagon primary RNA were obtained by RT-PCR studies. Total RNA (2.5 μg) was digested with 10 units of RNase-free DNase at 37°C for 1 h. After heat inactivation of DNase at 95°C for 10 min, RNA was reverse transcribed with random hexamers in the presence of reverse transcriptase (RT+) or absence of reverse transcriptase (RT−). The resultant cDNA was amplified using Tag polymerase (44).

After adenoviral transduction, STC-1 or GLUTag cells were grown for an additional 24 h and nuclear proteins were isolated as described (17). Synthetic oligonucleotides corresponding to rat proglucagon promoter G1 or G3 sequences (10) were end-labeled with ³²P-ATP by T4 kinase, purified by G25 spin column (Pharmacia, Montreal, PQ), and incubated with 2–4 μg nuclear proteins. The reaction mixes were loaded onto a 5% nondenaturing polyacrylamide gel, and, after electrophoresis, the gel was exposed to X-ray film for 16–24 h.

A solution of ∼1 × 10¹² virions of either Ad-Pax-6 or Ad-βgal was infused into the large colon of Wistar rats via a modified rectal approach developed in our laboratory. Briefly, animals were fasted for 48 and 16 h before viral infusion, and colonic fecal material was completely cleansed by feeding animals with GoLytely solution (Baxter, Mississauga, ON), followed by three additional bowel irrigations with PBS buffer. To disrupt intestinal mucous, a mucolytic agent, acetylcysteine, was introduced into the rectum and maintained for 30 min before viral infusion. After being mildly sedated with halothane, animals were infused with a viral solution via the rectum by using a 10-cm modified gavage and were kept in the Trendelenburg position for at least 30 min to enhance transduction of adenoviruses into the large intestine. Because the rapid turnover of intestinal epithelium may result in diminished Pax-6 expression in newly formed epithelial cells, animals were killed on day 3 after the viral infusion. The entire colon was excised, cleaned, quickly frozen in liquid nitrogen, ground into fine powder, and used for RNA isolation.

To characterize adenoviral-derived Pax-6 and Cdx-2/3, we analyzed expression of these transcription factors in BHK and 293 cells transduced with an increasing multiplicity of infection (MOI of 2.5, 10, and 25). Western blot analyses revealed dose-dependent increases in levels of Cdx-2/3 and Pax-6 in both cell lines (Fig. 1A).

To study the transcriptional activation of the rat proglucagon promoter by adenoviral-derived Cdx-2/3 and Pax-6, 293 and BHK fibroblasts were transfected with a reporter plasmid containing 300 bp of the rat proglucagon promoter linked to a luciferase reporter gene [−300]GLU-Luc and transduced with increasing MOI of Ad-Cdx-2/3 and Ad-Pax-6 alone or in combination. Both Cdx-2/3 and Pax-6 increased proglucagon promoter activity (Fig. 1B) and, consistent with previous reports (14,28), coexpression of Cdx-2/3 and Pax-6 produced a synergistic activation of proglucagon promoter activity (Fig. 1B). Nuclear extracts isolated from 293 or BHK cells transduced with various viruses at different MOI were used in electrophoretic mobility shift assay (EMSA) experiments with proglucagon promoter G1 and G3 probes (Fig. 1C and D). As shown in Fig. 1C, no detectable DNA-protein G3 complex was observed in cells transduced with increasing MOI of Ad-βgal or Ad-Cdx-2/3. By contrast, a dose-dependent increase in the formation of a Pax-6–G3 complex was observed in cells transduced with increasing MOI of Ad-Pax-6. Cells cotransduced with both Ad-Pax-6 and Ad-Cdx-2/3 appeared to contain a somewhat less abundant G3 complex compared with cells expressing Ad-Pax-6 alone. Similar findings were observed using the G1 oligonucleotide probe, which preferentially binds Cdx-2/3 (Fig. 1D). Moreover, co-transduction of both Ad-Pax-6 and Ad-Cdx-2/3 into 293 or BHK cells at either low or high MOI did not result in detectable formation of a higher molecular weight complex that might be suggestive of heterodimer formation (Fig. 1C and D).

To assess the effect of adenoviral-derived Pax-6 and Cdx-2/3 expression in enteroendocrine cells, we studied mouse GLUTag cells. Basal expression of endogenous Pax-6 and Cdx-2/3 was detectable in GLUTag cells infected with the control Ad-βgal virus (Fig. 2A). An approximate three- to fivefold increase in levels of Pax-6 or Cdx-2/3 was observed following viral transduction, with adenovirus-derived Pax-6 and Cdx-2/3 migrating at the same electrophoretic positions as the endogenous GLUTag (Fig. 2A) or STC-1 proteins (data not shown).

GLUTag cells transduced with Ad-Cdx-2/3 or Ad-Pax-6 exhibited increased proglucagon promoter activity (Fig. 2B), whereas cells infected with a truncated Pax-6 cDNA, Ad-Sey^neu did not exhibit significant differences in the transcriptional activity of the identical proglucagon promoter plasmid (Fig. 2B). Hence, the adenoviral-derived Cdx-2/3 and Pax-6 function as transcriptional regulators of the proximal rat proglucagon promoter in gut endocrine cells (14,17,28–31).

To ascertain whether enhanced expression of Pax-6 or Cdx-2/3 leads to increased levels of intestinal proglucagon mRNA transcripts, and not just proglucagon promoter activity, we transduced GLUTag cells with Ad-Cdx-2/3, Ad-Pax-6, or Ad-βgal for 24 h. The levels of proglucagon mRNA transcripts were increased in GLUTag cells transduced with Ad-Pax-6, in a dose-dependent manner (Fig. 3A). In contrast, GLUTag cells transduced with Ad-Cdx-2/3 or the control Ad-βgal virus did not exhibit changes in levels of endogenous proglucagon gene expression (Fig. 3B). Furthermore, GLUTag cells transduced with Ad-Pax-6 also exhibit increased biosynthesis and secretion of the PGDPs, as assessed by radioimmunoassays for GLP-1 of cell and media contents following viral transduction (data not shown).

To assess whether the Pax-6–dependent activation of enteroendocrine proglucagon gene expression is cell-type specific, we transduced Ad-Pax-6 into HEK 293 cells and BHK fibroblasts, nonendocrine cells that do not normally express the proglucagon gene. In contrast to the activation of endogenous proglucagon gene expression observed in GLUTag endocrine cells, no induction of proglucagon mRNA transcripts was observed in these nonendocrine cell lines following transduction with Ad-Pax-6 (Fig. 3B).

To determine whether Pax-6 or Cdx-2/3 is capable of activating proglucagon gene expression in enteroendocrine cells derived from both the large and small intestine, we transduced GLUTag cells (derived from endocrine cells of the large bowel) and STC-1 cells (derived from endocrine tumors arising from the small bowel) with Ad-Pax-6, Ad-Cdx-2/3, and Ad-Sey^neu viruses. As shown in Fig. 4A, both GLUTag and STC-1 cells transduced with Ad-Pax-6 exhibited similar increases in proglucagon mRNA transcripts relative to levels observed in cells infected with Ad-βgal. In contrast, proglucagon mRNA transcripts were not increased in GLUTag or STC-1 cells transduced with Ad-Cdx-2/3 (Fig. 4). Furthermore, we did not observe additive increases in the levels of proglucagon mRNA transcripts in enteroendocrine cells transduced with both Ad-Cdx-2/3 and Ad-Pax-6 (Fig. 4). Indeed, cells transduced with both Ad-Cdx-2/3 and Ad-Pax-6 did not exhibit the increase in proglucagon mRNA transcripts detected in cells transduced with Ad-Pax-6 alone (Fig. 4), suggesting that Cdx-2/3 may actually inhibit the stimulatory effects of Pax-6 on the intestinal proglucagon promoter.

To determine whether basal levels of endogenous Cdx-2/3 contributed to control of enteroendocrine proglucagon gene expression, we transduced GLUTag cells with the dominant negative Cdx-2/3-HD virus (Fig. 5). A marked reduction of proglucagon promoter activity was observed following transduction of Ad-Cdx-2/3-HD (Fig. 5A). To assess the effect of blocking endogenous Cdx-2/3 activity on newly synthesized proglucagon mRNA transcripts, we performed semiquantitative RT-PCR using specific primers amplifying sequences between intron-B and exon-3 of the mouse proglucagon mRNA transcript. In contrast to the potent inhibition of proglucagon promoter activity (observed in Fig. 5A), we did not observe a reduction in levels of endogenous proglucagon RNA transcripts (Fig. 5B) in GLUTag cells.

To ascertain whether Pax-6 activates endogenous proglucagon gene expression in normal intestinal endocrine cells, we isolated small-bowel epithelial cells from fetal rats and cultured cells for 24 h in vitro before adenoviral transduction. Proglucagon mRNA transcripts were clearly increased in a dose-dependent manner following Ad-Pax-6 transduction. In contrast, proglucagon mRNA transcripts were not increased following infection of primary intestinal cells with Ad-βgal, Ad-Cdx-2/3, or Ad-Sey^neu (Fig. 6A). Similar to findings in GLUTag and STC-1 cells (Fig. 4), coinfection of primary intestinal cell cultures with both Ad-Pax-6 and Ad-Cdx-2/3 abrogated the induction detected with Ad-Pax-6 alone (Fig. 6A). These results demonstrate that Pax-6, but not Cdx-2/3, is sufficient for induction of proglucagon gene expression in both transformed and nontransformed intestinal cells.

To determine whether Pax-6 is also capable of increasing endogenous expression of the intestinal proglucagon gene in the intact gastrointestinal tract of a living animal, we performed an in vivo infusion of Ad-Pax-6 virus directly into the rectum and distal portion of the large bowel of Wistar rats. The large bowel of rats transduced with Ad-Pax-6 exhibited an approximate twofold increase in levels of Pax-6 RNA transcripts, with no difference detected in the levels of endogenous Cdx-2/3 after Ad-Pax6 transduction (data not shown). The results of this in vivo experiment showed that animals given Ad-Pax-6 exhibited an approximate three- to fourfold increase in proglucagon mRNA transcripts in the large bowel relative to control animals infused with Ad-βgal (Fig. 6B).

The failure to observe Ad-Cdx-2/3–dependent activation of proglucagon gene expression in islet and enteroendocrine cells prompted us to assess the transcriptional activity of the Ad-Cdx-2/3 virus in Caco-2 cells, a Cdx-2/3–responsive intestinal epithelial cell line (32). Caco-2 cells transduced with Ad-βgal or Ad-Pax-6 contained comparable levels of endogenous sucrase isomaltase (SI) mRNA transcripts compared with nontreated Caco-2 cells (Fig. 7A). In contrast, the levels of SI mRNA transcripts were induced following transduction of Caco-2 cells with Ad-Cdx-2/3 (Fig. 7A). These findings demonstrate that the failure of Cdx-2/3 to activate endogenous enteroendocrine proglucagon gene expression was not attributable to the presence of a transcriptionally defective Ad-Cdx-2/3 virus.

To exclude the possibility that the effect of Ad-Pax-6 on intestinal proglucagon gene expression requires coexpression of adenoviral proteins or cofactors that directly or indirectly enhance the levels of proglucagon mRNA transcripts, we performed lipofectamine-mediated DNA transfection of a plasmid vector (pBAT7) containing the Pax-6 cDNA alone or with cotransduced Ad-βgal (expressing adenoviral proteins/cofactors). The results of this experiment (Fig. 7B) showed that STC-1 cells transfected with Pax-6 also exhibited increased levels of proglucagon mRNA transcripts compared with cells transfected with empty vector pBAT7 alone (Fig. 7B). No additive increases in proglucagon gene expression were observed in cells transfected with Pax-6 in the presence of Ad-βgal. Taken together, these studies demonstrate that Pax-6 alone, whether delivered via a plasmid or adenovirus, is sufficient for activation of proglucagon gene expression in enteroendocrine cells.

Transcription factors regulating expression of the proglucagon gene in islet cells include Brn4, Pax-6, Pax-2, Cdx-2/3, Isl-1, CREB, Maf factors, HNF-3β, and HNF-3α (12,16,18,19,31,34–39). The majority of evidence linking these transcription factors to proglucagon gene expression is derived from transfection experiments studying the proglucagon promoter in immortalized islet cell lines. However, there is little genetic evidence from human studies linking mutations in these candidate transcription factor genes to abnormalities in proglucagon gene transcription.

Pax-6 was first identified as a proglucagon gene transcription factor in studies of Pax-6 knockout and Pax-6 Sey^neu mutant mice (21,31). Both lines of mice exhibit marked defects in formation of normal islets, with reduced numbers or absent islet α-cells and with concurrent reductions in islet proglucagon mRNA transcripts. Pax-6 binds to rat proglucagon gene promoter sequences and transactivates the promoter in transfection experiments. Intriguingly, our data demonstrate that although abrogation of Pax-6 transcriptional activity markedly reduced proglucagon promoter activity, no effect on proglucagon mRNA transcripts was detected in the same experiment using the mutant Pax6 cDNA. These data illustrate that analysis of relative promoter activity represents only one component of the complexity of factors regulating steady-state levels of RNA transcripts. Furthermore, the essential importance of Pax-6 for the developmental formation of islet cells precludes assessment of a potentially independent role for Pax-6 in the control of proglucagon gene expression in the adult pancreas in vivo.

Pax-6 is also essential for the formation of subsets of gut enteroendocrine cells. Pax-6^−/− mice exhibit reduced numbers of duodenal gastrointenstinal polypeptide (GIP) cells and decreased numbers of gastric cells immunopositive for gastrin and somatostatin (40). Analysis of endocrine cell populations in the gut of Sey^neu mice expressing a dominant negative Pax-6 protein revealed almost complete absence of enteroendocrine cells containing GLP-1 or GLP-2 immunopositivity and a marked decrease in the levels of proglucagon mRNA transcripts in the small and large intestine (19). Nevertheless, it is not possible to infer from these studies whether the reduction in gut proglucagon gene expression is attributable to the importance of Pax-6 as an essential factor for enteroendocrine cell development, or whether the primary role of Pax-6 resides at the level of controlling proglucagon gene expression once enteroendocrine cell development has proceeded normally.

Our results clearly show that Pax-6 alone is sufficient for activation of endogenous proglucagon expression both in vitro and in vivo. This is the first demonstration that enhanced expression of a single transcription factor is capable of activating proglucagon gene expression in isolated cell lines, primary cell cultures, and in the context of the intact intestinal epithelium in vivo. Taken together, these studies provide additional evidence linking the level of Pax-6 expression to the control of intestinal proglucagon gene expression both in vitro and in vivo. In contrast, our data demonstrate that enhanced expression of Cdx-2/3 in primary intestinal cell cultures, or in the intact intestinal epithelium, is not sufficient for activation of enteroendocrine proglucagon gene expression. Hence, although Cdx-2/3 may activate the transfected proglucagon promoter, our data do not support using Cdx-2/3 alone for the enhancement of intestinal proglucagon gene expression.

The finding that GLP-1 agonists ameliorate diabetes in rodents has sparked considerable interest in the potential use of GLP-1 for the treatment of human subjects with type 2 diabetes. Similarly, the intestinotrophic and cytoprotective effects of GLP-2 have prompted studies of the potential use of GLP-2 analogs for the treatment of human patients with intestinal failure (2). An alternative approach to the therapeutic administration of GLPs involves strategies designed to increase the synthesis and secretion of the intestinal GLPs from normal gut endocrine cells in vivo. The gut represents a potentially attractive target for peroral delivery of viral vectors, and several studies have demonstrated the utility of targeting the gut with gene therapy approaches in vivo (41,42,43). Our studies identify Pax-6 as the first transcription factor capable of enhancing proglucagon gene expression within the context of the normal rodent gut epithelium in vivo. Future studies directed at examining the utility of adenoviral-delivered Pax-6 targeted to the gut epithelium for the enhanced production of GLP-1 and GLP-2 in the setting of experimental diabetes or intestinal disease in vivo appear warranted.

This work was supported in part by grants from the Banting and Best Diabetes Center (to D.K.Y.T.) and the Canadian Institutes of Health Research (CIHR) (to D.K.Y.T., P.L.B., and D.J.D.). D.J.D. is supported by a Senior Scientist Award from the CIHR, and P.L.B. is a CIHR Canada Research Chair in Diabetes.

We thank Ruijuan He for expert technical assistance.