Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Gastroenterology. Author manuscript; available in PMC 2010 December 1.
Published in final edited form as:
PMCID: PMC2789913

Foxa1 and Foxa2 Control the Differentiation of Goblet and Enteroendocrine L- and D-Cells in Mice


Background & Aims

The winged helix transcription factors Foxa1 and Foxa2 are expressed in all epithelia of the gastrointestinal tract from its embryonic origin into adulthood. In vitro studies have shown that Foxa1/a2 can transactivate the promoters of Mucin 2 (Muc2), which is expressed in goblet cells, and of preproglucagon, which is expressed in enteroendocrine cells. These findings suggest Foxa1/a2 as critical factors in the differentiation of gut epithelial cells.


Mice with intestine-specific simultaneous deletion of Foxa1 and Foxa2 were derived using the Cre-loxP system and analyzed using histological and molecular means.


Both Foxa1 and Foxa2 were successfully deleted in the epithelia of the small intestine and colon using Villin-Cre mice. Immunohistochemical staining showed that Foxa1/a2 mutants lack glucagon-like peptide-1 and 2 expressing cells (L-cells), and have reduced numbers of somatostatin (D-cells) and peptide YY (PYY) expressing cells (L-cells). Preproglucagon, somatostatin and PYY mRNA levels were also significantly reduced in Foxa1/a2 mutants. Thus, Foxa1 and Foxa2 are essential regulators of these enteroendocrine lineages in vivo. The mRNA levels of transcription factors Islet-1 and Pax6 were significantly reduced in the small intestine, demonstrating that Foxa1 and Foxa2 impact on a transcription factor network in the enteroendocrine lineage. In addition, deletion of Foxa1/a2 caused a reduction in goblet cell number with altered expression of the secretory mucins Muc2, Mucin5b, Mucin5ac and Mucin 6.


The winged helix factors Foxa1 and Foxa2 are essential members of the transcription factor network that governs secretory cell differentiation in the mammalian gastrointestinal tract.


During embryogenesis, the gastro-intestinal epithelia are derived from the definitive endoderm through a series of complex developmental steps. Several transcription factors, including the bHLH transcription factors Math1 and Beta2, neurogenin-3 (Ngn3), the paired box transcription factors Pax4 and Pax6, the zinc-finger transcription factor Krüppel-like factor 4 (Klf4) and insulinoma associated-1 (Insm-1 or IA-1), and the homeodomain transcription factor Nkx2.2 have been shown to play critical roles in the differentiation of different types of epithelial cells of the gastrointestinal tract [19]. Secretory cell lineages, which include goblet, Paneth and enteroendocrine cells, are derived from a common Math1-expressing progenitor, whereas enterocytes are Math1 independent [6]. Ngn3 controls enteroendocrine cell fate commitment of Math1+ secretory progenitors [1, 7]. Beta2 acts downstream of Ngn3 and is required specifically for the differentiation of cholecystokinin and secretin-producing cells [1, 2]. The differentiation of goblet cells, however, is Ngn3 independent. In fact, Ngn3−/− mice have an increased number of goblet cells in the small intestine, possibly due to the failure of stem cells to differentiate along the enteroendocrine lineage [1]. Klf4 is another transcription factor that is required for the terminal differentiation of goblet cells in the colon [10]. Nkx2.2 has been shown to be important for the differentiation of several enteroendocrine cells such as CCK, GIP, gastrin, glucogan and somatostatin [8]. Insm-1 is essential for the differentiation of serotonin, CCK and PYY [9].

The winged-helix transcription factors Foxa1 and Foxa2 are expressed in the definitive endoderm during embryogenesis [1113], and in many adult tissues derived from the endoderm such as pancreas, liver, stomach and intestine [12]. Foxa1 null mice die within the first two weeks of life due to hypoglycemia and mild nephrogenic diabetes insipidus [1417]. Foxa2 null embryos do not elaborate an organized node and notochord, and they die after gastrulation with defects in dorsal-ventral patterning of the neural tube [18, 19]. Due to the early lethality of both Foxa1 and Foxa2 null mice, it has been impossible thus far to determine their role in intestinal epithelial cell differentiation in genetically altered mice. The present study was designed to determine the function of Foxa1 and Foxa2 in intestinal epithelial cell differentiation in adult mice using cell-type specific gene ablation.



The derivation of Foxa1loxP/l+ and Foxa2loxP/+ mice has been described previously [2023]. After mating Foxa1loxP/+ heterozygote mice with Foxa2loxP/loxP homozygous mice, Foxa1loxP/+,Foxa2loxP/loxP mice were obtained. The Villin-Cre mouse [24] was mated with Foxa1loxP/+, Foxa2loxP/loxP mice to produce mice with intestine-specific deletion of Foxa1 and Foxa2. Foxa1loxP/+,Foxa2loxP/loxP or Foxa1loxP/loxP, Foxa2loxP/loxP mice were used as controls in all studies. Genotyping was performed by PCR using genomic DNA isolated from the toe tips of newborn mice. Eight to nine-week old adult male or female mice were used in all experiments except for the growth rate determination.

RNA preparation and Quantitative Real-time PCR (qRT-PCR)

Intestines were dissected and divided into duodenum, jejunum, ileum, colon, and rectum. Total RNA was extracted in TRIzol (Invitrogen) according to the manufacturer’s protocol. RNA was reverse transcribed using 1 µg oligo(dT) primer, Super-Script II reverse transcriptase, and accompanying reagents (Invitrogen), and qRT-PCR was performed as described [25].


Intestines were dissected and divided into duodenum, jejunum, ileum, colon, and rectum, and fixed in 4% paraformaldehyde overnight at 4°C, embedded in paraffin, cut to 6-µm sections and processed for immunohistochemistry as described previously [25, 26]. The primary antibodies for Foxa1/2 (Santa Cruz, 1:1000), chromagranin A (Immunostar, 20085, 1:2000), GLP-1 (Santa Cruz, sc-7780, 1:1000), GLP-2 (Santa Cruz, sc-7781, 1:100), somatostatin (Santa Cruz, sc-7819, 1:1000), PYY (Abcam, ab22663, 1:8000), cholecystokinin (Abcam, ab27441, 1:200), secretin (Abcam, ab8495, 1:100), GIP (Santa Cruz, sc-23554, 1:1000), vasoactive intestinal peptide (Immunostar, 20077), gastrin (Nova Castra, NCL-GASp, 1:5000), and serotonin (Immunostar, 20080, 1:10000), BrdU (US Biologicals, B2850-01) were diluted in PBT (1xPBS with 0.1% Triton x-100 and 2% BSA). Slides were incubated in primary antibody overnight at 4°C, washed in 1xPBS, and then incubated with biotinylated anti-goat antibodies (Vector Laboratories, 1:200) or anti-rabbit antibodies (Vector Laboratories, 1:200) for 30 min at 37°C. Slides were rinsed with 1xPBS and incubated with HRP-conjugated ABC reagent (Vector Elite kit) for 30 min at 37°C. After washing, slides were developed using a DAB substrate kit (Vector Laboratories), and counter-stained with hematoxylin.

Assessment of Body Composition

Eight-week old male mice were injected intraperitoneally with sodium pentobarbital (40 mg/kg) and dual energy x-ray absorptiometry was performed (DEXA; Lunar PIXImus2; General Electric Medical systems, Madison, WI) to determine their body composition [27, 28].

Electron Microscopy (EM)

Intestines were dissected and divided into duodenum, jejunum, ileum, colon, and rectum, washed with 1xPBS, and suspended in a fixative solution containing 2.5% cacodylate-buffered glutaraldehyde and 4% paraformaldehyde (pH 7.4) overnight at 4°C [10, 20].

Chromatin immunoprecipitation (ChIP)

Foxa1 and Foxa2 ChIP were performed as described [29], using antibodies against Foxa1 (a gift from Dr. Günther Schütz) and Foxa2 (a gift from Dr. Jeffrey Whitsett). The ChIP PCR primers spanned the putative Foxa1 or Foxa2 binding site covering −200/−87 bp of the Muc2 promoter (forward primer: 5’-ccaagtttacccagggagtcat-3’; reverse primer: 5’-gcatttgccaagttatcaggaa-3’).

Detection of apoptosis in goblet and enteroendocrine cells

Apoptosis of goblet and enteroendocrine cells was detected by using the ApopTag® Peroxidase In Situ Apoptosis Dectection Kit (Chemicon, S7100), and ApopTag® Red In Situ Apoptosis Dectection Kit (Chemicon, S7165), following manufacturer’s instructions. Briefly, intestinal sections were deparafinized by xylene followed by 95% and 70% ethanol. The sections were then pretreated with 20 µg/mL proteinase K. After quenching endogenous peroxidase with 3% hydrogen peroxide, the sections were incubated with equilibration buffer. Terminal deoxynucleotidyl transferase (TdT) was then applied to the sections and incubated for 1 hr at 37°C, after which the stop/wash buffer was applied to the sections for 10 min at room temperature. After incubation of the sections with anti-digoxigenin conjugate, sections were incubated in Acian blue solution for 20 min. For detection of enterodendocrine cells, sections were incubated with chromagranin A antibody overnight at 4°C before the application of anti-digoxigenin conjugate (rhodamine).


Derivation of mice with intestine-specific deletion of Foxa1 and Foxa2

Foxa1 and Foxa2 play important roles in the formation of the definitive endoderm during gastrulation and in the differentiation of endoderm-derived tissues [18, 19]. In adult mice, Foxa1 and Foxa2 are expressed in all intestinal epithelia [12]. Since both Foxa1 and Foxa2 are present in the crypts where the stem cells reside, we hypothesized that Foxa1 and Foxa2 control epithelial cell differentiation. We derived mice deficient for Foxa1 and Foxa2 in the intestine using the Cre-loxP system as illustrated in Figure 1A. Immunohistochemical staining demonstrated that both Foxa1 and Foxa2 proteins were absent from the small intestine (Fig. 1E), colon and rectum (see supplemental Fig. 1) of Foxa1loxP/loxP,Foxa2loxP/loxP;Villin-Cre mice (termed “Foxa1/a2 mutants” in the following). Because the third member of the FoxA gene family, Foxa3, is also expressed in the intestine [13, 30, 31], we considered the possibility that this gene might be upregulated in Foxa1/a2 mutants to compensate for the loss of Foxa1 and Foxa2. Instead, we found Foxa3 mRNA levels were significantly reduced in the Foxa1/a2 mutants’ jejunum but not in the duodenum (Fig. 1B). Thus, there is no compensatory upregulation of Foxa3 expression in Foxa1/a2 mutants.

Figure 1
Reduced body weight and slower growth rate in mice with intestine specific deletion of Foxa1 and Foxa2

Foxa1/a2 mutants were born at the expected Mendelian ratio and their body weights were similar to control mice at birth (data not shown). However, the growth rate of both male and female Foxa1/a2 mutants was reduced compared to that of controls from postnatal day 3 onward (Fig. 1C, supplemental Fig. 2). At 8 weeks of age, Foxa1/a2 mutants were significantly smaller (26.8%) and shorter (7.3%) than littermate controls (Fig. 1D). DEXA Scan analysis determined that bone mineral content (BMC), lean muscle and fat weight were reduced by 16.8%, 22.3% and 42.4%, respectively (Fig. 1D). The percentage of whole body fat was not significantly different from controls (Fig. 1D).

Defective goblet cell differentiation in Foxa1 and Foxa2-deficient mice

Examination of gross morphology showed that the overall anterior/posterior patterning of the intestine was normal in 8-week old Foxa1/a2 mutants. To determine whether deletion of Foxa1 and Foxa2 affects epithelial cell proliferation, BrdU was given to 8-wk old mice at 1 hr or 48 hr before organ harvest by intraperitoneal injection. The intestinal sections were stained with anti-BrdU antibody. At 1 hr, the BrdU positive cells were present at the bottom of the crypts for both controls and mutants (supplemental Fig. 3). After 48 hr, the BrdU positive cells had migrated comparable distances up the crypt-villus axis in both controls and mutants (supplemental Fig. 3). These data demonstrate that epithelial cell proliferation is not affected by intestine-specific deletion of Foxa1 and Foxa2.

Examination of the morphology of enterocytes or colonocytes showed that these cells displayed normal columnar morphology (supplemental Fig. 3). Paneth cells were also differentiated properly (data not shown). Next, we studied the effects of Foxa1 and Foxa2 deletion on goblet cell differentiation. We stained intestinal sections with Alcian blue, which labels the glycoproteins in the secretory granules of goblet cells. The number of Alcian blue positive cells was reduced throughout the intestinal tract (Fig. 2A), and quantification of goblet cells showed a significant decrease of goblet cell numbers in the duodenum, jejunum and ileum in the mutants compared to controls (Fig. 2B). In both colon and rectum, partially or completely empty vacuoles were observed in goblet cells of Foxa1/a2 mutants by ultrastructural imaging (Fig. 3). Electron microscopy also showed reduced overall granule density, indicating fewer mucins contained within each granule in Foxa1/a2 mutants compared to controls (Fig. 3).

Figure 2
Defective goblet cell differentiation in Foxa1/a2 mutants
Figure 3
Reduced mucin granules in goblet cells of Foxa1/a2 mutants

The major secretory products of goblet cells are the mucin glycoproteins, among which Muc2 is the main component. In vitro studies by van der Sluis and colleagus had shown that Foxa1 and Foxa2 can bind to and transactivate the Muc2 promoter [32]. We found Muc2 mRNA levels decreased in the colon by 87%, and in the rectum by 86% in Foxa1/a2 mutants but not significantly changed in the small intestine (Fig. 4A). Besides Muc2, Mucin 5b (Muc5b), Mucin 5ac (Muc5ac), and Mucin 6 (Muc6) are also present in the intestine. Both Muc5b and Muc5ac mRNA levels were significantly increased in the colon and rectum of Foxa1/a2 mutants (Fig. 4B and 4C). In contrast, Muc6 mRNA levels were decreased in the colon by 99% and in the rectum by 69% in Foxa1/a2 mutants compared to controls (Fig. 4D). Thus, the mucin expression profile is dramatically altered in the absence of Foxa1 and Foxa2.

Figure 4
Defective goblet cell differentiation in Foxa1/a2 mutants

Trefoil factor 3 (Tff3), a protein secreted by goblet cells, plays an important role in the maintenance and repair of the intestinal mucosa [33]. The expression of Tff3 was dramatically reduced in the colon by 75% and rectum by 96% in Foxa1/a2 mutants compared to controls (supplemental Fig. 5A). Because Klf4 had been shown previously to affect goblet cell differentiation in the colon, we determined its expression levels by qRT-PCR. Klf4 expression was not altered in the colon but significantly reduced in the rectum (supplemental Fig. 5B).

To determine whether Foxa1 or Foxa2 can bind to the Muc2 promoter in vivo, we performed ChIP assay with chromatin isolated from the small intestinal epithelium. Shown in Fig. 4E is the putative Foxa binding site as predicted by the JASPAR database, and the chromatin immunoprecipitation experiment demonstrating occupancy of this site by Foxa1 in vivo. Quantification of the ChIP experiment revealed that the Muc2 promoter sequence was enriched in chromatin precipitated with the Foxa1-, but not the Foxa2- specific antibody (Fig. 4E and F). When we performed ChIP on Foxa1/a2 mutant tissue, no enrichment to the Muc2 promoter remained, demonstrating that our antibodies and ChIP conditions were specific for Foxa1 (Fig. 4E and F). Our data indicate that Foxa1 is the preferred transcription factor in binding and activating of the Muc2 promoter. Consistent with our findings, previous in vitro studies also showed that Foxa1 is more efficient than Foxa2 in activating the Muc2 promoter in co-transfection assays [32].

To determine whether the decrease in goblet cells was caused by increased apoptosis, intestinal sections were co-stained with Alcian blue and Tunel. Very few goblet cells were Tunel-positive in either controls or mutant tissue, indicating similar apoptosis rates of these two groups (supplemental Fig. 4A). These data suggest that decreased goblet cells number is more likely due to decreased goblet cell differentiation rather than increased apoptosis.

Enteroendocrine cell differentiation depends on Foxa1 and Foxa2

Foxa2 is an essential transcriptional regulator for the differentiation of pancreatic endocrine cells [21, 26]. Moreover, both Foxa1 and Foxa2 can bind to and activate the preproglucagon promoter in vitro [17, 34]. Two protein products from the preproglucagon gene, Glp-1 and Glp-2, are expressed in intestinal enteroendocrine L-cells. In adult mice, Foxa1 and Foxa2 localize to the intestinal crypts which contain stem cells and transit amplifying cells that produce all epithelial cells (Supplemental Fig 1, [12]). Therefore, we hypothesized that Foxa1 and Foxa2 contribute to enteroendocrine cell differentiation. In order to determine whether enteroendocrine cells were present in the mutant intestinal epithelium, we performed immunohistochemical analysis with antibodies against chromagranin A, a major component of the secretory granules of all enteroendocrine cells. Chromagranin A positive cells were scattered throughout the epithelium of the small intestine of control mice, while only a few chromagranin A positive cells were present in the Foxa1/a2 mutants (Fig. 5). To determine whether the decrease in enterendocrine cell number was due to increased apoptosis, intestinal sections were co-stained with chromagranin A and Tunel. We observed very few chromagranin A/Tunel double positive cells in either controls or mutants (supplemental Fig. 4B), suggesting that decreased enteroendocrine cell differentiation rather than increased apoptosis was the cause of reduced enteroendocrine cell number in Foxa1/a2 mutant mice.

Figure 5
Decreased enteroendocrine cell differentiation in Foxa1 and Foxa2 mutants

Next, we performed immunohistochemical staining with antibodies specific to the dominant hormone produced by each subtype. Glp-1 and Glp-2 positive cells were absent from the small intestine of Foxa1/a2 mutants (Fig. 6). Foxa1/a2 mutants also had fewer somatostatin and PYY positive cells (Fig. 6). In contrast, staining for other principal intestinal hormones including cholecystokinin (CCK), secretin (SEC), gastrin (GAS), serotonin (SER), GIP, and vasoactive intestinal polypeptide (VIP) showed that expression of these enteroendocrine hormones is Foxa1/a2 independent (data not shown). In addition, preproglucagon mRNA levels were undetectable in the intestine of Foxa1/a2 mutants by qRT-PCR (Fig. 7A). Compared with control mice, somatostatin mRNA levels in Foxa1/a2 mutant mice were reduced by 71% and 83% in the duodenum and jejunum, respectively (Fig. 7B). PYY expression is restricted to the ileum and was significantly decreased (80%) in Foxa1/a2 mutants compared to controls (Fig. 7C). Thus, Foxa1 and Foxa2 are required for L-cell (Glp-1, Glp-2, PYY) and D-cell (somatostatin) differentiation in the mouse small intestine.

Figure 6
Decreased differentiation of L- and D-cells in Foxa1/a2 mutants
Figure 7
Isl-1 and Pax6 act downstream of Foxa1 and Foxa2 in the enteroendocrine cell lineage

Foxa1 and Foxa2 act upstream of Isl-1 and Pax6 to control L- and D-cell differentiation

Ngn3, Isl-1, Beta2, Pax4, Pax6, Nkx2.2 and Insm-1 are transcription factors required for normal enteroendocrine cell differentiation [14, 8, 9, 35]. Ngn3 is necessary for the elaboration of all intestinal enteroendocrine cell types and controls enteroendocrine cell fate commitment of Math1+ secretory lineage progenitors [1]. Beta2 acts downstream of Ngn3 to regulate cholecystokinin and secretin-producing cells [1]. To determine at which stage of enteroendocrine cell differentiation Foxa1 and Foxa2 are required, we examined the expression of Ngn3, Isl-1, Beta2, Pax4, Pax6, Nkx2.2 and Insm-1 by qRT-PCR in Foxa1/a2 mutants and controls. Ngn3 and Beta2 expression was unchanged in the Foxa1/a2 mutants jejunum compared to controls (Fig. 7D), suggesting that Foxa1 and Foxa2 act downstream of Ngn3 in the enteroendocrine cell lineage (Fig. 8). In contrast, Isl-1 and Pax6 expression levels were reduced by 59% and 69%, respectively (Fig. 7D and E), while Pax4 expression was unchanged in the ileum of Foxa1/a2 mutants (Fig. 7E). The expression of Nkx2.2 and Insm-1 was unchanged as well (supplemental Fig. 5). Isl-1 and Pax6 are transcription factors that bind to and transactivate the preproglucagon promoter [36, 37]. We propose that Isl-1 and Pax6 act downstream of and coordinately with Foxa1/Foxa2 to regulate L-cell differentiation (Fig. 8). Moreover, Isl-1 can also bind to and transactivate somatostatin [38] promoter. Thus, Foxa1/Foxa2 and Isl-1 cooperate to control D-cell differentiation in a feed-forward loop. Although Isl-1, Pax6 and PYY expression were reduced in Foxa1/a2 mutants, ChIP experiments using Foxa1 or Foxa2 antibodies did not reveal direct binding of Foxa1 or Foxa2 to the Isl-1, Pax6 or PYY promoters (supplemental Fig. 6).

Figure 8
Proposed model for the role of Foxa1 and Foxa2 during differentiation of the secretory cell lineage in the mammalian intestine


The intestinal epithelium renews every three to four days in mice through the differentiation of transit amplifying cells emerging from the crypts. Multiple transcription factors are involved in this process, including Math1, Ngn3, Isl-1, Pax4 and Pax6, NeuroD, Nkx2.2 and Insm-1. Here, we propose that Foxa1 and Foxa2 are additional transcription factors important for proper differentiation of goblet and enteroendocrine cells. Goblet cells function to produce mucus to protect the gut, and alteration in mucin expression has been implicated in intestinal inflammation and carcinogenesis [3941]. We observed decreased numbers of goblet cells throughout the intestinal tract of Foxa1/a2 mutants. Foxa1/a2 mutants also had significantly reduced Muc2 expression in the colon and rectum at age of 8 weeks, while the entire mucin pattern was perturbed (Fig. 4). Taken together, Foxa1 and Foxa2 are important factors for goblet cell differentiation, and possibly the maintenance of intestinal health.

The endocrine cells of the gastrointestinal tract form the largest endocrine system of the body. Hormones produced by enteroendocrine cells control various functions such as glucose metabolism, exocrine pancreatic secretion, the growth and repair of the intestinal epithelium, and the motility of the intestinal wall [35]. Several transcription factors such as Isl-1, Pax6, Pax4, Beta2, Nkx2.2 and Insm-1 are not only important for the expression of gut hormones but also involved in the differentiation of these enteroendocrine cells. The present study shows that Foxa1 and Foxa2 control the differentiation of enteroendocrine L- and D-cells via regulating expression of several important transcription factors including Isl-1 and Pax6 (Fig. 7 D and E). The current model of the differentiation of the secretory cell lineages in the mammalian intestine is summarized in Figure 8.

The principal hormones expressed in L-cells, Glp-1 and Glp-2, were absent from the intestine of Foxa1/a2 mutants (Fig. 6). Transcription of the preproglucagon gene is controlled by numerous transcription factors, including Isl-1, Pax6, and Foxa1 or Foxa2 [17, 26, 34, 36, 42]. Preproglucagon promoter activity is regulated through G1, G2, G3, and G4 elements. Both Foxa1 and Foxa2 can interact with similar affinities to either the G1 or G2 element of preproglucagon promoter in vitro [17, 34]. Recently, an additional conserved Foxa binding site located close to the transcription start site of the preproglucagon promoter has been identified and shown to bind preferentially to Foxa1 in a glucagon-producing pancreatic islet α-cell line [43]. Moreover, Foxa1 null mice have markedly reduced levels of plasma glucagon [17]. Consistent with prior work in the pancreas, our study shows that deletion of both Foxa1 and Foxa2 ablates preproglucagon transcripts in the small intestine (Fig. 7A).

Pax6 is a critical regulator of pancreatic α-cell differentiation during development [42]. It is a transcription factor of preproglucagon gene and inhibition of Pax6 led to decreased endogenous glucagon expression in pancreatic α-cell lines [44]. Mice expressing a dominant-negative Pax6 mutant have reduced preproglucagon mRNA levels and lack Glp-1 and Glp-2 positive cells in the small and large intestine [37]. Pax6 activates the preproglucagon promoter through G1 and G3 elements [45]. In vitro, Pax6 binds to an overlapping site on G1 element with a greater affinity as Foxa1 or Foxa2 [34]. Binding of Foxa1 and Foxa2 impairs Pax6-mediated transactivation of the preproglucagon promoter via G1 and G3 elements in co-transfection assays, suggesting an antagonism between Pax6 and the Foxa factors in the regulation of this gene [34]. We show here that intestine-specific deletion of Foxa1 and Foxa2 causes decreased Pax6, Glp-1 and Glp-2 expression (Fig. 6 and Fig 7). Thus, our data indicate that in small intestine Foxa1 and Foxa2 control Pax6 transcription, and Pax6 acts downstream of and in concert with Foxa1 and Foxa2 to regulate preproglucagon gene transcription in a feed-forward loop.

Isl-1 is another important transcription factor of preproglucagon gene transcription, and Isl-1 null mice exhibit defective islet development and α-cell formation [36, 46]. In early embryogenesis, Isl-1 can be detected in the foregut, implicating its possible role in the gut development [47]. Our data showed that Isl-1 expression is significantly reduced in mice with intestine specific deletion of Foxa1 and Foxa2 (Fig. 7D), suggesting that Foxa1 and Foxa2 may regulate Isl-1 gene transcription and Isl-1 lie downstream of Foxa1/a2-mediated enteroendocrine lineage. Reduction of Isl-1 expression could contribute to the significant decrease of preproglucagon gene expression in Foxa1/a2 mutants.

Foxa1/a2 mutants are smaller and have lower bone mineral content, less lean muscle and body fat than age-matched controls. These phenotypes can be explained, in part, by reduced action of Glp-2, PYY and somatostatin. Glp-2 functions as a key regulator of mucosal integrity, permeability, and nutrient absorption (reviewed in [48]). While we did not observe decreased epithelial cell proliferation in the small intestine of Foxa1/a2 mutants, the lack of Glp-2 may affect nutrient absorption and thus body weight. PYY is important for water and electrolyte absorption [49], and indeed we observed increased water loss via the feces of Foxa1/a2 mutants (data not shown). Somatostatin controls gastric emptying, reduces smooth muscle contraction and blood flow within the intestine, and suppresses the release of several gastrointestinal hormones including gastrin, cholecystokinin, secretin, VIP, GIP, Glp-1 and Glp-2 ([50], reviewed in [51, 52]). Foxa1/a2 mutants have relatively normal intestinal gastrin, cholecystokinin, secretin, VIP, and GIP expression as shown by immunohistochemical analysis and qRT-PCR (data not shown), but we cannot exclude that altered somatostatin levels might affect plasma levels for these hormones.

Since Glp-1 is important for glucose metabolism, intestine-specific deletion of Foxa1 and Foxa2 may have an impact on glucose homeostasis. Glp-1 stimulates insulin and inhibits glucagon secretion from the endocrine pancreas, thus contributing to postprandial glucose uptake (reviewed in [53, 54]). It also inhibits gastrointestinal motility, regulates appetite and food intake (reviewed in [53, 54]). In the pancreas, Glp-1 stimulates β-cell proliferation and neogenesis, and prevents β-cell apoptosis [5558]. Surprisingly, the resting glucose level was normal in Foxa1/a2 mutants (data not shown). One possibility for this finding is that deletion by the Villin-Cre transgene is not complete in the proximal duodenum, and that the remaining L-cells present there produce enough Glp-1 to control glucose homeostasis. Another possibility for the normal glucose homeostasis in our model is the fact that GIP expression is normal in Foxa1/a2 mutants, and both GLP-1 and GIP are insulinotropic peptides. GIP-receptor deficient mice have higher blood glucose levels with impaired initial insulin response in the oral glucose test [59]. GLP-1 receptor null mice had mild glucose intolerance and impaired insulin secretion after intraperitoneal glucose injection [60]. Interestingly, GLP-1 receptor null mice had increased plasma GIP levels, suggesting a compensatory relationship between the two signaling mechanisms [61]. It might be interesting to test the metabolic response of Foxa1/a2 mutants to a high fat diet, which is often used to uncover metabolic defects, in the future.

In conclusion, we demonstrate that intestine-specific deletion of Foxa1 and Foxa2 alters the differentiation of goblet and enteroendocrine L- and D-cells. In goblet cells, Foxa1 binds to the Muc2 promoter preferentially over Foxa2. In L- and D-cells, Foxa1 and Foxa2 control expression of Glp-1/2, somatostatin and PYY both directly and through feed-forward regulation of Isl-1 and Pax6.

Supplementary Material



We thank Dr. Gary Wu for critical reading of the manuscript, and Ms. Karrie Brondell for care of the mouse colony. We acknowledge the Penn Diabetes and Endocrinology Research Center Mouse Metabolic Phenotyping Core (P30-DK-19525) for performing mouse body composition and CLAMS studies, and we want to thank Ravindra Dhir and Xiaoyan Yin for excellent technical support. This study was supported by the NIH (DK R01-053839).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Author contribution

DZY: study concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript

KHK: study concept and design; critical revision of the manuscript for important intellectual content; obtained funding; study supervision

The authors declare that no conflict of interest exists.


1. Jenny M, et al. Neurogenin3 is differentially required for endocrine cell fate specification in the intestinal and gastric epithelium. Embo J. 2002;21(23):6338–6347. [PubMed]
2. Naya FJ, et al. Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice. Genes Dev. 1997;11(18):2323–2334. [PubMed]
3. Larsson LI, et al. Pax 4 and 6 regulate gastrointestinal endocrine cell development. Mech Dev. 1998;79(1–2):153–159. [PubMed]
4. Larsson L, et al. The LIM-homeodomain protein Isl-1 segregates with somatostatin but not with gastrin expression during differentiation of somatostatin/gastrin precursor cells. Endocrine. 1995;3:519–524. [PubMed]
5. Lee CS, Kaestner KH. Clinical endocrinology and metabolism. Development of gut endocrine cells. Best Pract Res Clin Endocrinol Metab. 2004;18(4):453–462. [PubMed]
6. Yang Q, et al. Requirement of Math1 for secretory cell lineage commitment in the mouse intestine. Science. 2001;294(5549):2155–2158. [PubMed]
7. Lee CS, P N, Brestelli JE, Kaestner KH. Neurogenin 3 is essential for the proper specification of gastric enteroendocrine cells and the maintenance of gastric epithelial cell identity. Genes Dev. 2002;16(12):1488–1497. [PubMed]
8. Desai S, et al. Nkx2.2 regulates cell fate choice in the enteroendocrine cell lineages of the intestine. Dev Biol. 2008;313(1):58–66. [PMC free article] [PubMed]
9. Gierl MS, et al. The zinc-finger factor Insm1 (IA-1) is essential for the development of pancreatic beta cells and intestinal endocrine cells. Genes Dev. 2006;20(17):2465–2478. [PubMed]
10. Katz JP, et al. The zinc-finger transcription factor Klf4 is required for terminal differentiation of goblet cells in the colon. Development. 2002;129(11):2619–2628. [PMC free article] [PubMed]
11. Monaghan AP, et al. Postimplantation expression patterns indicate a role for the mouse forkhead/HNF-3 alpha, beta and gamma genes in determination of the definitive endoderm, chordamesoderm and neuroectoderm. Development. 1993;119(3):567–578. [PubMed]
12. Besnard V, et al. Immunohistochemical localization of Foxa1 and Foxa2 in mouse embryos and adult tissues. Gene Expr Patterns. 2004;5(2):193–208. [PubMed]
13. Kaestner KH, et al. The HNF-3 gene family of transcription factors in mice: gene structure, cDNA sequence, and mRNA distribution. Genomics. 1994;20(3):377–385. [PubMed]
14. Vatamaniuk MZ, et al. Foxa1-deficient mice exhibit impaired insulin secretion due to uncoupled oxidative phosphorylation. Diabetes. 2006;55(10):2730–2736. [PubMed]
15. Behr R, et al. Mild nephrogenic diabetes insipidus caused by Foxa1 deficiency. J Biol Chem. 2004;279(40):41936–41941. [PubMed]
16. Shih DQ, et al. Impaired glucose homeostasis and neonatal mortality in hepatocyte nuclear factor 3alpha-deficient mice. Proc Natl Acad Sci U S A. 1999;96(18):10152–10157. [PubMed]
17. Kaestner KH, et al. Inactivation of the winged helix transcription factor HNF3alpha affects glucose homeostasis and islet glucagon gene expression in vivo. Genes Dev. 1999;13(4):495–504. [PubMed]
18. Ang SL, Rossant J. HNF-3 beta is essential for node and notochord formation in mouse development. Cell. 1994;78(4):561–574. [PubMed]
19. Sasaki H, Hogan BL. HNF-3 beta as a regulator of floor plate development. Cell. 1994;76(1):103–115. [PubMed]
20. Gao N, et al. Foxa2 controls vesicle docking and insulin secretion in mature Beta cells. Cell Metab. 2007;6(4):267–279. [PubMed]
21. Sund NJ, Vatamaniuk MZ, Casey M, Ang SL, Magnuson MA, Stoffers DA, Matschinsky FM, Kastner KH. Tissue-specific deletion of Foxa2 in pancreatic beta cells results in hyperinsulinemic hypoglycemia. Genes Development. 2001;15:1706–1715. [PubMed]
22. Sund NJ, et al. Hepatocyte nuclear factor 3beta (Foxa2) is dispensable for maintaining the differentiated state of the adult hepatocyte. Mol Cell Biol. 2000;20(14):5175–5183. [PMC free article] [PubMed]
23. Gao N, LeLay J, Vatamaniuk MZ, Rieck S, Friedman JR, Kaestner KH. Dynamic regulation of Pdx1 enhancers by Foxa1 and Foxa2 is essential for pancreas development. 2008 under review. [PubMed]
24. el Marjou F, et al. Tissue-specific and inducible Cre-mediated recombination in the gut epithelium. Genesis. 2004;39(3):186–193. [PubMed]
25. Gupta R, Gao N, Gorski RK, White P, Hardy OT, Rafiq K, Brestelli JE, Chen G, Stoeckert CJ, Jr, Kaestner KH. Expansion of adult beta-cell mass in response to increased metabolic demand is dependent on HNF-4alpha. Genes Dev. 2007;21(7):756–769. [PubMed]
26. Lee CS, Sund NJ, Bhr R, Herrera PL, Kaestner KH. Foxa2 is required for the differentiation of pancreatic alpha-cells. Development Biology. 2005;278:484–495. [PubMed]
27. Qi Y, et al. Loss of resistin improves glucose homeostasis in leptin deficiency. Diabetes. 2006;55(11):3083–3090. [PubMed]
28. Nagy TR, Clair AL. Precision and accuracy of dual-energy X-ray absorptiometry for determining in vivo body composition of mice. Obes Res. 2000;8(5):392–398. [PubMed]
29. Friedman JR, L B, Le PP, Peiris TH, Arsenlis A, Schug J, Tobias JW, Kaestner KH, Greenbaum LE. Orthogonal analysis of C/EBPbeta targets in vivo during liver proliferation. Proc Natl Acad Sci U S A. 2004;101(35):12986–12991. Epub 2004 Aug 18. [PubMed]
30. Lai E, et al. Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head. Genes Dev. 1991;5(3):416–427. [PubMed]
31. Hiemisch H, Schutz G, Kaestner KH. Transcriptional regulation in endoderm development: characterization of an enhancer controlling Hnf3g expression by transgenesis and targeted mutagenesis. Embo J. 1997;16(13):3995–4006. [PubMed]
32. van der Sluis M, et al. Forkhead box transcription factors Foxa1 and Foxa2 are important regulators of Muc2 mucin expression in intestinal epithelial cells. Biochem Biophys Res Commun. 2008;369(4):1108–1113. [PubMed]
33. Mashimo H, et al. Impaired defense of intestinal mucosa in mice lacking intestinal trefoil factor. Science. 1996;274(5285):262–265. [PubMed]
34. Gauthier BR, et al. Hepatic nuclear factor-3 (HNF-3 or Foxa2) regulates glucagon gene transcription by binding to the G1 and G2 promoter elements. Mol Endocrinol. 2002;16(1):170–183. [PubMed]
35. Skipper M, Lewis J. Getting to the guts of enteroendocrine differentiation. Nat Genet. 2000;24(1):3–4. [PubMed]
36. Wang M, Drucker DJ. The LIM domain homeobox gene isl-1 is a positive regulator of islet cell-specific proglucagon gene transcription. J Biol Chem. 1995;270(21):12646–12652. [PubMed]
37. Hill ME, Asa SL, Drucker DJ. Essential requirement for Pax6 in control of enteroendocrine proglucagon gene transcription. Mol Endocrinol. 1999;13(9):1474–1486. [PubMed]
38. Leonard J, et al. The LIM family transcription factor Isl-1 requires cAMP response element binding protein to promote somatostatin expression in pancreatic islet cells. Proc Natl Acad Sci U S A. 1992;89(14):6247–6251. [PubMed]
39. Van der Sluis M, et al. Muc2-deficient mice spontaneously develop colitis, indicating that MUC2 is critical for colonic protection. Gastroenterology. 2006;131(1):117–129. [PubMed]
40. Velcich A, et al. Colorectal cancer in mice genetically deficient in the mucin Muc2. Science. 2002;295(5560):1726–1729. [PubMed]
41. Bartman AE, et al. Aberrant expression of MUC5AC and MUC6 gastric mucin genes in colorectal polyps. Int J Cancer. 1999;80(2):210–218. [PubMed]
42. Sander M, et al. Genetic analysis reveals that PAX6 is required for normal transcription of pancreatic hormone genes and islet development. Genes Dev. 1997;11(13):1662–1673. [PubMed]
43. Sharma SK, et al. Characterization of a novel Foxa (hepatocyte nuclear factor-3) site in the glucagon promoter that is conserved between rodents and humans. Biochem J. 2005;389(Pt 3):831–841. [PubMed]
44. Gosmain Y, et al. Pax-6 and c-Maf functionally interact with the alpha-cell-specific DNA element G1 in vivo to promote glucagon gene expression. J Biol Chem. 2007;282(48):35024–35034. [PubMed]
45. Philippe J, Morel C, Cordier-Bussat M. Islet-specific proteins interact with the insulin-response element of the glucagon gene. J Biol Chem. 1995;270(7):3039–3045. [PubMed]
46. Ahlgren U, et al. Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells. Nature. 1997;385(6613):257–260. [PubMed]
47. Yuan S, Schoenwolf GC. Islet-1 marks the early heart rudiments and is asymmetrically expressed during early rotation of the foregut in the chick embryo. Anat Rec. 2000;260(2):204–207. [PubMed]
48. Dube PE, Brubaker PL. Frontiers in glucagon-like peptide-2: multiple actions, multiple mediators. Am J Physiol Endocrinol Metab. 2007;293(2):E460–E465. [PubMed]
49. Liu CD, et al. Peptide YY: a potential proabsorptive hormone for the treatment of malabsorptive disorders. Am Surg. 1996;62(3):232–236. [PubMed]
50. Ho LT, et al. Inhibition of somatostatin on glucose-induced release of gastric inhibitory polypeptide in rats. Chin J Physiol. 1987;30(2):45–53. [PubMed]
51. Weckbecker G, et al. Opportunities in somatostatin research: biological, chemical and therapeutic aspects. Nat Rev Drug Discov. 2003;2(12):999–1017. [PubMed]
52. Lucey MR. Endogenous somatostatin and the gut. Gut. 1986;27(4):457–467. [PMC free article] [PubMed]
53. Aaboe K, et al. GLP-1: physiological effects and potential therapeutic applications. Diabetes Obes Metab. 2008 [PubMed]
54. Holst JJ. The physiology of glucagon-like peptide 1. Physiol Rev. 2007;87(4):1409–1439. [PubMed]
55. Xu G, et al. Exendin-4 stimulates both beta-cell replication and neogenesis, resulting in increased beta-cell mass and improved glucose tolerance in diabetic rats. Diabetes. 1999;48(12):2270–2276. [PubMed]
56. Zhou J, et al. Glucagon-like peptide 1 and exendin-4 convert pancreatic AR42J cells into glucagon- and insulin-producing cells. Diabetes. 1999;48(12):2358–2366. [PubMed]
57. Farilla L, et al. Glucagon-like peptide 1 inhibits cell apoptosis and improves glucose responsiveness of freshly isolated human islets. Endocrinology. 2003;144(12):5149–5158. [PubMed]
58. Buteau J, et al. Glucagon-like peptide-1 prevents beta cell glucolipotoxicity. Diabetologia. 2004;47(5):806–815. [PubMed]
59. Miyawaki K, et al. Glucose intolerance caused by a defect in the entero-insular axis: a study in gastric inhibitory polypeptide receptor knockout mice. Proc Natl Acad Sci U S A. 1999;96(26):14843–14847. [PubMed]
60. Scrocchi LA, et al. Glucose intolerance but normal satiety in mice with a null mutation in the glucagon-like peptide 1 receptor gene. Nat Med. 1996;2(11):1254–1258. [PubMed]
61. Pederson RA, et al. Enhanced glucose-dependent insulinotropic polypeptide secretion and insulinotropic action in glucagon-like peptide 1 receptor −/− mice. Diabetes. 1998;47(7):1046–1052. [PubMed]