Intracellular calcium-release and protein kinase C-activation stimulate sonic hedgehog gene expression during gastric acid secretion

Abstract

Introduction

Hypochlorhydria during Helicobacter pylori infection inhibits gastric Shh expression. We investigated whether acid-secretory mechanisms regulate Shh gene expression through Ca2⁺_i-dependent protein kinase C (PKC) or cAMP-dependent protein kinase A (PKA)-activation.

Method

We blocked Hedgehog signaling by transgenically overexpressing a secreted form of the Hedgehog interacting protein-1 (sHip-1), a natural inhibitor of hedgehog ligands, which induced hypochlorhydria. Gadolinium, EGTA+BAPTA, PKC-overexpressing adenoviruses, and PKC-inhibitors were used to modulate Ca²⁺_i-release, PKC-activity and Shh gene expression in primary gastric cell, organ, and AGS cell line cultures. PKA hyperactivity was induced in the H⁺/K⁺-β-cholera-toxin overexpressing mice (Ctox).

Results

Mice that expressed sHip-1 had lower levels of gastric acid (hypochlorhydria), reduced production of somatostatin, and increased gastrin gene expression. Hypochlorhydria in these mice repressed Shh gene expression, similar to the levels obtained with omeprazole treatment of wild-type mice. However, Shh expression was also repressed in the hyperchlorhydric Ctox model with elevated cAMP, suggesting that the regulation of Shh was not solely acid-dependent, but pertained to specific acid-stimulatory signaling pathways. Based on previous reports that Ca²⁺_i-release also stimulates acid secretion in parietal cells, we showed that gadolinium-, thapsigargin- and carbachol-mediated release of Ca²⁺_i induced Shh expression. Ca²⁺-chelation with BAPTA+EGTA reduced Shh expression. Overexpression of PKC-α, -β and -δ (but not PKC-ε) induced Shh gene expression. In addition, phorbol esters induced a Shh-regulated reporter gene.

Conclusion

Secretagogues that stimulate gastric acid secretion induce Shh gene expression through increased Ca²⁺_i-release and PKC activation. Shh might be the ligand transducing changes in gastric acidity to the regulation of G-cell secretion of gastrin.

INTRODUCTION

Sonic Hedgehog (Shh) is expressed by gastric parietal cells^1–3, and correlates with several mechanisms related to parietal cell acid secretion. First, Shh stimulates H⁺/K⁺-ATPase gene expression and subsequently enhances histamine-stimulated acid secretion by parietal cells⁴. Second, Shh protein co-localizes with the proton/potassium pump (H⁺/K⁺-ATPase)³. Third, gastric acid induces processing of the Shh 45kDa precursor form to the 19kDa biologically active form⁵. Fourth, blocking gastric acid secretion with a proton pump inhibitor inhibits Shh gene expression^{2, 6}. The latter observation is relevant to the observation that Helicobactor pylori-infection downregulates Shh gene expression apparently through its ability to induce pro-inflammatory cytokines, which inhibit acid secretion⁶. Parietal cell Shh production is important since tissue-specific deletion of Shh using the Cre-Lox system induces foveolar hyperplasia⁷, which is reminiscent of the pre-malignant changes induced by H. pylori. Nevertheless, the mechanism by which gastric acid regulates Shh is not well understood.

In gastric cells, the process of acid secretion is associated with protein kinase A (PKA) and protein kinase C (PKC) signaling⁸. While the elevation in cAMP levels and activation of PKA pertain to histaminic stimulation⁸, both cholinergic (acetylcholine) and hormonal (gastrin) stimulation of acid secretion increase intracellular calcium (Ca²⁺_i) in parietal cells^{9, 10}. The release of Ca²⁺_i activates several signal transduction pathways including Ca²⁺-dependent PKC isoforms, α and β¹¹. While the effects of Ca²⁺_i on Shh expression have not been previously investigated, Ca²⁺-chelation inhibits Indian hedgehog (Ihh) gene expression in chick chondrocytes¹². Furthermore, in the chick wing bud, PKC sustains Shh gene expression¹³. Collectively, these studies suggest that calcium-dependent PKCs, when activated, transduce elevated Ca²⁺_i levels to downstream targets.

Previously, we reported that Shh enhances acid secretion in gastric parietal cells by increasing H⁺/K⁺-ATPase gene expression thereby increasing enzyme content⁴. In this report, regulation of Shh gene expression was examined in vivo and in vitro. Transgenic overexpression of an inhibitor of Hedgehog (Hh) ligands called hedgehog-interacting protein-1 (Hip-1) from parietal cells resulted in hypochlorhydric mice. This permitted us to examine the effect of low gastric acidity on the Shh gene locus (Shh gene expression). Since Shh gene expression was reduced, we hypothesized that the change occurred in response to the hypochlorhydria. Knowing that PKC regulates Shh gene expression during development, we tested the hypothesis that hypochlorhydria reduces Ca²⁺_i and PKC activation, ultimately inhibiting Shh gene expression.

MATERIALS AND METHODS

Generation of transgenic mice

The Hip-1 cDNA lacking the transmembrane domain (sHip-1)¹⁴ was subcloned downstream of the H⁺/K⁺-ATPase beta-subunit promoter¹⁵. The H⁺/K⁺-ATPase-Hip transgene was injected into fertilized eggs obtained by mating (C57BL/6 X SJL)F1 (UM Transgenic Core). Transgenic founders were bred to C57BL/6 mice. Ctox mice (line 7) were generated as described previously¹⁵.

Omeprazole treatment

Omeprazole (Sigma-Aldrich) stock (80 μmol/ml) was prepared in a 9:1 solution of DMSO:Polyethylene glycol (PEG) 400 (Fluka, Sigma-Aldrich). Intraperitoneal injections of omeprazole (400 μmol/kg) were administered once daily for 3-consecutive days.

Gastric acidity

After opening the stomach along the greater curvature, the gastric contents were collected in 1.5 ml of 0.9% NaCl. The samples were centrifuged to collect a clear supernatant. The hydrogen ion concentration was determined by titrating with 0.005 N sodium hydroxide using the PHM290 pH-Stat Controller titration system (Radiometer, OH).

Western blot analysis

Western blotting was performed according to previously published conditions⁵ using a 1:200 dilution of goat polyclonal anti-Shh (sc-1194, Santa Cruz, CA), 1:500 GAPDH (Molecular Probes, Invitrogen) antibody, and 1:1000 phospho-PKCα/β_II (Thr638/641) antibody (#9375, Cell Signaling, Boston, MA).

RESULTS

Hypochlorhydria in sHip-1 expressing mice

Prior studies of primary parietal cell cultures demonstrated a role for Shh in H⁺/K⁺-ATPase gene expression, implicating an indirect effect of Shh on gastric acid production⁴. Indeed, we recently demonstrated that a proinflammatory cytokine inhibits both gastric acid and Shh gene expression⁶. Therefore to understand the relationship between gastric acidity and Shh expression in vivo, we generated a transgenic mouse expressing the natural inhibitor of Shh called Hip-1. The secreted form of Hip-1 (sHip-1), which lacks the transmembrane domain, inhibits Hh signaling in the intestine by binding the Hh ligands¹⁴. We therefore expressed sHip-1 from the H⁺/K⁺-ATPase-β subunit promoter to block Hh signaling in the stomach (Fig. 1A). To confirm expression of the sHip-1 transgene in the stomachs of 2-month-old mice, Hip-1 mRNA expression in non-transgenic mice was compared to transgenic littermates by RT-qPCR. Hip-1 protein was verified by western blotting. sHip-1 mice showed a 40-fold increase in Hip-1 mRNA (Fig. 1B), and a marked increase in Hip-1 protein (Fig. 1C). There was only a slight reduction in size (2.4 kDa) of the sHip-1 transgene compared to endogenous Hip-1, since the deleted transmembrane domain is only 22 amino acids¹⁴. The increase in Hip-1 corresponded to a significant reduction in Hh-target gene expression (Gli-1 and Ptch-1) demonstrating efficient suppression of hedgehog signaling (Fig. 1D), and was consistent among 3 different founder lines (Supplementary Fig. 1A–C).

Figure 1. Hip-1 mRNA levels and Shh signaling in the gastric fundus of transgenic mice.

A) The secreted hedgehog interacting protein 1 (sHip-1) transgene, lacking the C-terminal transmembrane domain, was expressed downstream of the H⁺/K⁺-ATPase-β subunit promoter, and in frame with the 3′UTR of the human growth hormone (HGH) gene and polyA tail to enhance expression in mammalian cells. The mice expressing the transgene were designated as “sHip-1 mice”. B) Quantitative PCR analysis of Hip-1 mRNA from the sHip-1 mice compared to non-transgenic littermate is shown as the mean +/− SEM for 5 mice per group. C) A representative western blot of Hip-1 protein (85kDa) from a sHip-1 mouse (Founder line 450) compared to non-transgenic littermate is shown. D) Target gene expression of glioma-associated oncogene-1 (Gli-1) and Ptch-1 genes in the sHip-1 fundus versus controls to assay the signaling activity of the Shh pathway (shown is the mean for n = 5 mice per group +/− SEM). P-values are indicated such that * P < 0.05 and ** P < 0.01.

Previous reports have shown that Hh signaling induces H⁺/K⁺-ATPase gene expression and enhances histamine-stimulated acid secretion in primary canine parietal cell cultures⁴. Therefore we examined whether H⁺/K⁺-ATPase gene expression was affected by the overproduction of sHip-1. Indeed, H⁺/K⁺-ATPase-β subunit mRNA and protein were significantly reduced in 3 founder lines (Fig. 2A, C, and Supplementary Fig. 2), despite the normal morphological distribution of parietal cells observed with H&E staining (Fig. 2B, insert). This correlated with reduced levels of gastric acid in 3 founder lines (Fig. 2B, and supplementary Fig. 1D). In contrast, H⁺/K⁺-ATPase-α subunit expression at the protein or mRNA level was not affected (Supplementary Fig. 3).

Figure 2. H⁺/K⁺-ATPase-β expression and acid production in sHip-1 mice.

A) H⁺/K⁺-ATPase-β mRNA expression in the sHip-1 fundus versus non-transgenic littermate measured by RT-qPCR (shown is the mean for n = 5 mice per group +/− SEM). B) Gastric acidity measured by base titration in sHip-1 versus non-transgenic mice. B) (insert) Histological examination of the sHip-1 gastric mucosa. C) H⁺/K⁺-ATPase-β subunit staining (red) in the sHip-1 mice and non-transgenic controls. Low and high-power confocal images are shown in the left and right panels respectively. The DAPI nuclear stain is pseudo-colored in grey. P-values are indicated such that ** P < 0.01.

Shh peptide and H⁺/K⁺-ATPase-β reside in the same complex

Since the hedgehog pathway is associated with acid secretion^{2, 5, 6}, and Shh co-migrates with H⁺/K⁺-ATPase to the parietal cell apical membrane during acid secretion^{3, 16}, we determined whether Shh and H⁺/K⁺-ATPase reside in the same subcellular complex. Indeed, Shh protein immunoprecipitated with the H⁺/K⁺-ATPase-β subunit protein (Supplementary Fig. 4A). Shh exists inside the cell as a 45 kDa precursor¹⁷, then a 39 kDa protein after removal of the signal peptide. Further processing generates a 27kDa C-terminal form, and a 19kDa biologically active form (ShhN)¹⁷. The addition of cholesterol and palmitate to the 19kDa form (ShhNp) modifies the mobility of this form such that it migrates faster in an SDS-gel¹⁸. Surprisingly, the H⁺/K⁺-ATPase-β protein immunoprecipitated only with the faster-migrating processed Shh form, suggesting that only ShhNp formed a complex with the H⁺/K⁺-ATPase-β subunit (Supplementary Fig. 4B).

Increased gastrin expression in sHip-1 mice

Since the sHip-1 overexpressing mice were hypochlorhydric, we examined whether gastrin gene expression also increased. Indeed, gastrin mRNA levels were significantly elevated (Fig. 3A). Moreover, immunohistochemistry and morphometric analyses revealed a 3-fold increase in the number of gastrin-expressing cells (Fig. 3B). By contrast, somatostatin gene expression was depressed in the sHip-1 mice (Fig. 3C) consistent with elevated gastrin levels. Although suppression of Hh signaling correlated with hypochlorhydria and hypergastrinemia^{2, 7}, we considered the possibility that Hh signaling might regulate gastrin by modulating somatostatin gene expression directly. To test this possibility, enriched cultures of primary D cells were prepared and treated with recombinant Shh. D-cells did not express Shh (Supplementary Fig. 5A), but expressed the Hh receptor Ptch-1 (Supplementary Fig. 5B) demonstrating that they are responsive to Hh ligands. We found an increase in somatostatin gene expression with Shh ligand treatment (Fig. 3D). Therefore, reduced somatostatin levels in sHip-1 mice leading to hypergastrinemia might result from the loss of Shh-mediated D-cell stimulation.

Figure 3. Gastrin and somatostatin mRNA expression levels in sHip-1 mice.

A) Gastrin gene expression relative to hypoxanthine–guanine phosphoribosyltransferase (HPRT) in the sHip-1 antrum versus control mice measured by RT-qPCR. Shown is the mean for n = 3 mice +/− SEM. B) Morphometric quantitation of gastrin protein immunostaining in the sHip-1 antrum versus control mice. C) Somatostatin mRNA expression measured by RT-qPCR in the sHip-1 antrum versus non-transgenic littermate. Shown is the mean for n = 3 mice +/− SEM. D) Northern blot of somatostatin expression in an enriched culture of primary canine D-cells treated with or without 0.5μg/ml recombinant 19kDa Shh peptide. D-cells constitute 70% of the cells in the enriched culture. ** P < 0.01.

Downregulation of Shh mRNA expression in sHip-1 mice

Since expression of sHip-1 was sufficient to inhibit acid secretion and expression of the H⁺/K⁺-ATPase enzyme, we examined the effect of acid secretion on endogenous Shh expression. Previously, Suzuki et al. showed in Mongolian gerbils that gastric hypochlorhydria reduces Shh gene expression¹⁹. Moreover, we previously reported that acid is required for processing of the nascent Shh peptide to its biologically active form⁵. However, neither study addresses how acid secretion regulates Shh gene expression. Since gastric acid secretion was reduced in these mice and the Shh locus was still intact in the transgenic mice, we examined the levels of Shh mRNA and protein. Indeed, we found that both Shh mRNA and protein expression were reduced in the sHip-1 mice (Fig. 4A, B). Concluding that the effect on Shh was in response to reduced acid levels⁶, we treated wild-type mice with omeprazole for 3 days. Consistent with reduced gastric acid in the sHip-1 mice, omeprazole blocked acid secretion (Fig. 4C), Shh gene expression (Fig. 4D), and Hh target genes Ptch-1 and Gli-1 (Fig. 4D). Thus acid secretion is required to sustain Shh mRNA and protein expression⁵. This observation prompted us to further investigate how acid secretion regulates Shh gene expression.

Figure 4. Shh mRNA expression in sHip-1, omeprazole-treated, and Ctox mice.

A) RT-qPCR demonstrating Shh mRNA expression in the gastric fundus of sHip-1 mice compared to non-transgenic controls. B) Representative western blot of full-length and processed Shh protein in the gastric fundus of sHip-1 mice versus non-transgenic controls. C) Gastric acidity measured by base titration in omeprazole versus vehicle-treated mice. D) RT-qPCR of Shh, Gli-1, and Ptch-1 mRNA expression from the gastric fundus of omeprazole- versus vehicle-treated mice. E) RT-qPCR of Shh and H⁺/K⁺-ATPase-α mRNA expression from the gastric fundus of Ctox versus non-transgenic mice. Shown are the means for n = 5 mice per group +/− SEM (for Ctox mice n = 3). P-values are indicated such that * P < 0.05 and ** P < 0.01.

cAMP elevation does not induce Shh gene expression during acid secretion

Previous studies correlated histaminic stimulation of acid-secretion with elevated cAMP levels in the parietal cell²⁰. We therefore investigated the effect of elevated cAMP levels on Shh gene expression in parietal cells. We utilized mice overexpressing cholera toxin, an irreversible stimulator of adenylate cyclase, from the H⁺/K⁺-ATPase-β subunit promoter (Ctox mice, line 7), which leads to increased cAMP signaling in parietal cells¹⁵. Despite hyperchlorhydria in these mice at 6 weeks of age¹⁵, Shh expression was inhibited significantly (Fig. 4E), possibly due to hypogastrinemia. We excluded the possibility that parietal cell atrophy was responsible for the loss of Shh expression by showing that H⁺/K⁺-ATPase-α expression was not affected in Ctox mice (Fig. 4E), corroborating previous observations that parietal atrophy does not occur in these mice before they reach 3 months of age¹⁵. The downregulation of Shh expression during hyperchlorhydria suggested that the regulation of Shh expression is not related to acid secretion per se, but to signaling processes regulating this process, and that the effect of omeprazole on Shh gene expression might be indirect. We therefore investigated Ca²⁺_i-release and PKC-activation as a potential mechanism for the modulation of Shh gene expression during acid secretion.

Intracellular calcium release modulates Shh expression during acid secretion

Prior studies have correlated the process of acid-secretion with elevated Ca²⁺_i in the parietal cell. First, cholinergic and hormonal stimulation of acid secretion increases Ca²⁺_i in parietal cells^{9, 10}. Second, thapsigargin- or calcium-sensing receptor (CaSR)-mediated increase in Ca²⁺_i stimulates gastric acid secretion^{20, 21}. Third, gastrin stimulation of acid secretion induces⁵, while omeprazole-mediated acid inhibition suppresses Shh gene expression (Fig. 4D). We therefore hypothesized that Ca²⁺_i stimulates Shh gene expression. To study the effect of Ca²⁺_i on Shh expression, we treated mouse primary gastric cultures with compounds that increase Ca²⁺_i by three different mechanisms: 1) gadolinium which increases Ca²⁺_i by activating the CaSR²¹, 2) thapsigargin which inhibits Ca²⁺-reuptake into the endoplasmic and sarcoplasmic reticulae, and 3) carbachol which elevates Ca²⁺_i via activation of the muscarinic 3 (M3) receptor (Fig. 7). All three compounds increased Shh expression transiently at 3 h, but not at 6 h (Fig. 5A) without affecting cell viability (Supplementary Fig. 6). The induction by Gd³⁺ was stronger than thapsigargin or carbachol, consistent with the ability of Gd³⁺ to perform additional functions such as activation of the MAPK pathway and the induction of diacylglycerol (DAG) production^{22, 23}.

Figure 7. Hypothetical model of Shh in the gastric mucosa.

A) Shown are three mechanisms capable of stimulating acid secretion through an increase in Ca²⁺_i released from the endoplasmic and sarcoplasmic reticulae: 1) hormonal-stimulation with gastrin, 2) cholinergic stimulation with acetylecholine (ACh) through the M3 muscarinic receptor, or 3) alkalinization of the basolateral surface by extruded HCO³⁻ ions leading to the activation of the calcium-sensing receptor (CaSR). Stimulation of acid secretion by histamine release from enterochromaffin-like cells is not shown. Ca²⁺_i-release stimulates acid secretion and Shh gene expression via PKC α and β. PKC-δ might also mediate diacylglycerol (DAG)-induction of Shh gene expression. DAG also stimulates Ca²⁺_i-release and acid secretion. The compounds used in this study, gadolinium (Gd³⁺), carbachol, thapsigargin and TPA, all stimulated Shh gene expression and are known to increase Ca²⁺_i. Shh protein undergoes processing (intracellular location not defined), migrates to the basolateral membrane, or co-migrates with the H⁺/K⁺-ATPase-β subunit to the apical membrane where it is likely to be secreted luminally. B) Luminal Shh potentially targets epithelial cells expressing Ptch-1 such as D-cells to induce non-canonical signaling. In addition, Shh protein reaching the basolateral membrane would target Gli-1-positive mesenchymal cells inducing a secondary signal from the mesenchyme. The outcome of Shh signaling is to induce somatostatin (SST) that in turn inhibits the G-cell and gastrin production, and acid secretion from the parietal cell through paracrine mechanisms.

Figure 5. Effect of Ca²⁺_i release on Shh mRNA expression.

A) Shh mRNA expression in mouse primary fundic culture following gadolinium (Gd³⁺, 0.5 mM), thapsigargin (1 μM) and carbachol (100 μM) versus vehicle treatment for 3 or 6 hours. Shown is the mean of 6 experiments +/− SEM. B) Left Panel, Fura-2 imaging of canine parietal cells with high levels of Ca²⁺_i after 48-hour culture before and after perfusion with EGTA (4mM) plus BAPTA-AM (10μM). Arrows indicate cultured parietal cells in which Ca²⁺_i were depleted after the perfusion. Right Panel, RT-qPCR of Shh mRNA from mouse fundic organ cultures treated with EGTA (4mM) plus BAPTA-AM (10μM) versus vehicle for 12 hours. Each data point is one mouse. * P < 0.05 and ** P < 0.01.

Prior reports have shown that EGTA and BAPTA-AM treatment of cells depletes extracellular and intracellular Ca²⁺ respectively²⁴. Therefore we examined whether chelating parietal cell Ca²⁺_i reduced Shh gene expression. We observed that when canine parietal cells were cultured for 48 h, in the absence of any stimulation, they showed a dramatic increase in Ca²⁺_i levels (Supplementary Fig. 7). Hence, we treated these cells with a cocktail of the chelators 48 h post culture. Calcium chelation depleted parietal cell calcium stores as measured by Fura-2 in canine primary cultures (Fig. 5B, left panel) and reduced Shh gene expression in organ cultures from mouse stomachs (Fig. 5B, right panel). We found that EGTA and BAPTA-AM induced intrinsic factor mRNA levels in organ cultures demonstrating that their effect on Shh expression was specific and was not due to general cellular toxicity of the chelators (Supplementary Fig. 8A).

PKC mediates calcium induction of Shh

An increase in Ca²⁺_i stimulates several signaling pathways including protein kinase C (PKC)¹¹. Calcium and diacylglycerol (DAG) activate the conventional PKCs (cPKC), PKCα-, βI-, βII-, and γ, whereas only DAG activates the novel isozymes (δ, ε, η, θ and μ). Calcium and DAG do not activate the atypical isozymes (ζ and λ). A recent study showed that PKC stimulates Shh gene expression¹³. We therefore hypothesized that cPKCs might mediate the effect of calcium on Shh expression. Indeed, treatment of mouse primary gastric cultures with the PKC activator, TPA, transiently induced Shh expression at 1 h, but not after 3, 6, or 24 h of treatment (Fig. 6A).

Figure 6. Effect of PKC on Shh expression in mouse primary cells and gastric cell lines.

A) RT-qPCR for Shh mRNA expression before and after TPA treatment for 1, 3, 6, and 24 hours (shown is the mean of 3 experiments +/− SEM). B) RT-qPCR of Shh mRNA from mouse primary fundic cell cultures transduced with adenoviral vectors overexpressing PKC-α, β, δ, ε, and empty vector (shown is the mean of 6 experiments +/− SEM). The cells were treated with 50 MOI of each of the individual adenoviruses. C) RT-qPCR measuring Shh in mouse fundic organ cultures treated with Bisindolylmaleimide I (Bis I, inhibits α, β, δ, and ε), PKC-β inhibitor, Ro-32-0432 (PKC-α inhibitor), and rottlerin (PKC-δ inhibitor). Each data point indicates one mouse. D) Shh-luciferase reporter activity normalized to Renilla luciferase in AGS cells treated with TPA versus vehicle. The figure demonstrates the reporter activity of the 0.2kb, 1.5kb, and 4.2kb promoters after 23 h incubation in media followed by 1 h TPA, 24 h TPA treatment, and 1 h TPA followed by 23 h incubation in media (shown is the mean of 3 experiments +/− SEM). * P < 0.05 and ** P < 0.01.

In order to determine which PKC isoform regulated Shh expression, we overexpressed two cPKC (α and β) and two novel PKC (δ and ε) isozymes using adenoviruses in mouse primary gastric cultures. Due to the reported low expression levels of PKC-α and β isoforms in gastric parietal cells²⁵, we first confirmed that the expression of the activated forms of these isoforms was detectable in canine parietal cells (Supplementary Fig. 9, antibody recognizes both α and β). Ectopic expression of PKC-α, β, or δ induced Shh expression, whereas PKC-ε did not have an effect (Fig. 6B). Thus even among two novel PKC isoforms (δ and ε) normally activated by DAG, only PKC-δ induced Shh.

Since PKC-α, β, and δ induced Shh expression, we next determined whether inhibition of these isozymes in mouse organ cultures inhibited Shh. We used bisindolylmaleimide I (PKC-α, βI, βII, δ, and ε inhibitor), Ro-32-0432 (PKC-α inhibitor), Rottlerin (PKC-δ inhibitor), and PKC-β inhibitor (#539654 from Calbiochem). These inhibitors significantly depressed Shh expression (Fig. 6C), supporting the results observed with the adenoviruses. The observed effect was specific for Shh, since bisindolylmaleimide I treatment did not affect intrinsic factor gene expression (Supplementary Fig. 8B).

Phorbol ester activates Shh promoter activity

Since phorbol esters directly stimulate PKC, we examined whether TPA stimulated Shh gene expression. To map the DNA regions conferring phorbol ester induction of the Shh promoter, we generated Shh-luciferase reporter constructs that contained 4.2, 1.5, and 0.2 kilobases upstream of the transcriptional start site (Supplementary Fig. 10). After transfecting these constructs into the human gastric AGS cell line, and treating the transformants with TPA, we observed significant induction of the 4.2kb, 1.5kb and 0.2kb Shh reporters at 24 h (Fig. 6D), but not at 1 h (Figure 6D). The delay could be attributed to either the time required for the luciferase protein to be produced as previously reported²⁶, or to the possibility that long-term treatment with TPA inhibited PKC activity. To exclude the second possibility, we treated AGS cells for 1 h with TPA followed by 23 h incubation in media, which induced luciferase activity similar to the 24 h treatment (Figure 6D). These results demonstrated that a TPA-responsive enhancer resided within the 0.2kb proximal promoter.

DISCUSSION

In the current study, we set out to understand the relationship between gastric acidity and Shh gene expression. Prior studies reported that Helicobacter-induced chronic gastritis represses Shh gene expression^{6, 19, 27}. Recently, Zavros and coworkers showed that a conditional deletion of the Shh gene locus in parietal cells induces foveolar hyperplasia, an epithelial lesion that recapitulates the pre-malignant changes observed with chronic gastritis⁷. We have recently reported that the repression of Shh gene expression during chronic gastritis is linked to the loss of gastric acidity mediated by pro-inflammatory cytokines, e.g., IL-1β⁶. Both Helicobacter infection in vivo and treatment of parietal cell cultures with either IL-1β or TNFα inhibit acid secretion^{28, 29}. Moreover, prior studies have linked parietal cell acid secretion to an increase in intracellular calcium²⁰. None of the prior studies directly link changes in intracellular calcium to Shh gene expression. Therefore we tested the hypothesis that modulating intracellular calcium was the common mechanism by which inflammation-induced hypochlorhydria regulates Shh gene expression.

We increased intracellular calcium by treating primary gastric cell cultures with agents that work through a ligand-receptor mediated mechanism, e.g. carbachol, or that increase or decrease intracellular calcium levels by activating the calcium-sensing receptor, blocking calcium re-uptake mechanisms, or directly chelating calcium (Fig. 7). We previously showed that gastrin stimulates Shh gene expression⁵. Gastrin has also been shown to elevate Ca²⁺_i in target cells⁹. Therefore, our results are consistent with the fact that mechanisms which increase Ca²⁺_i will stimulate Shh gene expression. By contrast, chelating intracellular calcium with BAPTA and EGTA inhibited Shh expression. Since changes in intracellular calcium alone were not sufficient to explain the nuclear events, we considered that Ca²⁺-dependent signal transduction pathways such as protein kinase C were possibly involved. Indeed, we found that overexpression of the Ca²⁺ or DAG- dependent PKC isoforms were sufficient to stimulate Shh gene expression and that the induction mapped to specific DNA elements. Although not determined here, PKC regulatory signals typically map to AP1 elements that bind Fos-Jun heterodimers³⁰. Helicobacter infection increases AP1 binding to consensus elements^{31, 32}. Yet there are few studies that have addressed the role of AP1-mediated transcription during Helicobacter infection. Indeed, in silico analysis demonstrated several putative AP1 sites within the responsive elements of the Shh promoter.

We observed in this study that both Ca²⁺-dependent PKCs α and β stimulated Shh gene expression. However, we also observed that Ca²⁺-independent PKC-δ stimulates Shh gene expression suggesting that other pathways must contribute significantly to the regulation of Shh mRNA expression. Indeed, the activation of PKC-δ is dependent on diacylglycerol (DAG), a second messenger signaling lipid that also stimulates Ca²⁺_i-release⁹. Importantly, Gd³⁺ stimulates the activity of the phospholipase C (PLC) enzyme required for the production of DAG³³, suggesting that Gd³⁺ targets both Ca²⁺_i and DAG production. Also, Gd³⁺ induced Shh gene expression more potently than thapsigargin or carbachol, which could be mediated by Ca²⁺_i-release and DAG production acting synergistically to stimulate Shh expression. It is also important to note that PKC-α and β stimulated Shh expression even though their expression is barely detectable in parietal cells²⁵. We detected low levels of the activated forms of PKCα/β_II in isolated canine parietal cells in this study, suggesting that small amounts of these proteins in parietal cells might be functionally sufficient.

This study suggested that the regulation of Shh expression does not specifically pertain to acid secretion per se, but to signalling pathways associated with this process such as calcium-signaling and PKC. Indeed, we observed using the Ctox mice that, despite hyperchlorhydria (induced through a non-calcium mediated mechanism), Shh gene expression was inhibited possibly due to hypogastrinemia leading to lower Ca²⁺_i levels. This suggested that Shh expression is not directly regulated by the efflux of H⁺ ions from parietal cells, and that the effect of omeprazole on Shh gene expression is possibly indirect. We speculate that omeprazole might be inhibiting Ca²⁺_i-release in parietal cells in vivo, even though this mechanism has not been explored. However, some studies suggest that omeprazole inhibits the alkaline tide between the parietal cell’s basolateral membrane and the blood^{34, 35}, which is necessary for the activation of the CaSR³⁶, and Ca²⁺_i-release²¹. Thus the inhibition of Ca²⁺_i-release by omeprazole might be a plausible mechanism affecting Shh gene expression requiring further investigation.

The studies here revealed additional information supporting the notion that Shh protein expression is acid-dependent. In the hypochlorhydric stomach of the sHIP mice, the 19 kDa processed ligand was absent. Recent studies using confocal analysis of Shh during parietal cell acid secretion indicate that secretagogues stimulate movement of the ligand to the apical surface^{3, 16}. Since acid secretion only occurs at the canalicular (apical) membrane, we previously concluded that the processing must also occur at the apical membrane⁵. We show here that the processed form of Shh co-precipitates with the H⁺/K⁺-ATPase-β subunit, consistent with the notion that Shh moves, at least in part, to the apical membrane of the parietal cell^{3, 16}.

Despite these observations, the mechanism by which Shh travels from the parietal cell to its target cells in the stomach remains unknown. Recent reports have shown that Shh-responsive cells in the gastrointestinal tract reside in the mesenchyme³⁷. However, some epithelial cells also express the Ptch-1 receptor³⁸ (unpublished observation), suggesting that gastric epithelial cells might respond to Hh ligands in a non-canonical manner by inducing Erk1/2³⁹. In this study, we found that D-cells express the Ptch-1 receptor suggesting that they might directly respond to Shh secreted luminally in a non-canonical manner. In contrast, Shh has also been localized to the basolateral membrane of parietal cells suggesting that it is also secreted basolaterally³ to target the mesenchyme. In the latter example, a secondary signal from the mesenchyme might be responsible for regulating gastrin and somatostatin production in G- and D-cells (Fig. 7). Canonical versus non-canonical Hh signaling might explain the disparate temporal yet distinct functions of the ligand.

Despite the evidence that gastric acid stimulated Shh gene expression, this study also showed that Shh is important for maintaining acid secretion, which suggests a feedback mechanism operating between Shh and gastric acid. A recent report shows that conditional deletion of the mouse Shh gene in gastric parietal cells leads to hypochlorhydria and reduced somatostatin expression⁷. Collectively, these changes induce hypergastrinemia⁷. Similarly by blocking Hh ligands, the sHip-1 transgenic mice were hypochlorhydric, exhibited reduced somatostatin expression, and developed hypergastrinemia (Fig. 7).

In conclusion, we show here that gastric acid secretion induced Shh gene expression via the release of Ca²⁺_i, which in turn activates calcium-specific PKC-α and PKC-β. These results provide a more detailed molecular mechanism that links Helicobacter infection to regulation of Shh gene expression and gastrin production.

Abbreviations

Shh

sonic hedgehog

Ptch-1

Patched-1

sHip-1

secreted hedgehog-interacting protein 1

TPA

phorbol-12-myristate-13-acetate

Ca²⁺_(i)

intracellular calcium

PKC

protein kinase C

gadolinium

Gd³⁺

DAG

diacylglycerol