Calcium-sensing receptor deletion in the mouse esophagus alters barrier function

Abstract

Calcium-sensing receptor (CaSR) is the molecular sensor by which cells respond to small changes in extracellular Ca²⁺ concentrations. CaSR has been reported to play a role in glandular and fluid secretion in the gastrointestinal tract and to regulate differentiation and proliferation of skin keratinocytes. CaSR is present in the esophageal epithelium, but its role in this tissue has not been defined. We deleted CaSR in the mouse esophagus by generating keratin 5 CreER;CaSR^Flox+/+compound mutants, in which loxP sites flank exon 7 of CaSR gene. Recombination was initiated with multiple tamoxifen injections, and we demonstrated exon 7 deletion by PCR analysis of genomic DNA. Quantitative real-time PCR and Western blot analyses showed a significant reduction in CaSR mRNA and protein expression in the knockout mice (^EsoCaSR^−/−) as compared with control mice. Microscopic examination of ^EsoCaSR^−/− esophageal tissues showed morphological changes including elongation of the rete pegs, abnormal keratinization and stratification, and bacterial buildup on the luminal epithelial surface. Western analysis revealed a significant reduction in levels of adherens junction proteins E-cadherin and β catenin and tight junction protein claudin-1, 4, and 5. Levels of small GTPase proteins Rac/Cdc42, involved in actin remodeling, were also reduced. Ussing chamber experiments showed a significantly lower transepithelial resistance in knockout (KO) tissues. In addition, luminal-to-serosal-fluorescein dextran (4 kDa) flux was higher in KO tissues. Our data indicate that CaSR plays a role in regulating keratinization and cell-cell junctional complexes and is therefore important for the maintenance of the barrier function of the esophagus.

NEW & NOTEWORTHY The esophageal stratified squamous epithelium maintains its integrity by continuous proliferation and differentiation of the basal cells. Here, we demonstrate that deletion of the calcium-sensing receptor, a G protein-coupled receptor, from the basal cells disrupts the structure and barrier properties of the epithelium.

INTRODUCTION

Stratified squamous epithelia are the body’s first line of defense against the outside environment. They consist of multiple cell layers and play an important role in isolating and protecting underlying structures from adverse conditions, including water loss and mechanical and chemical injuries. The esophageal stratified squamous epithelium consists of one or two layers of basal cells, a few layers of spinous cells or stratum spinosum, and a few layers of granular cells or stratum granulosum; in rodents, the uppermost layer is keratinized. The basal cells can divide and regenerate the whole epithelium in approximately 7 days (56, 61). Although the ability of basal cells to regenerate the esophageal epithelium is widely accepted, the presence of a distinct population of stem cells in the basal layer is a topic of debate (4, 22, 24, 35). Understanding the mechanisms of homeostasis and repair in the esophageal epithelium is of major importance because of the unexplained increased incidence of esophageal disease in the last 40 years and of the morbidity and mortality associated with esophageal cancers (73).

In stratified squamous epithelia like the skin, calcium plays a major role in maintaining the structure and the barrier function of the organ (for a review, see Refs. 29 and 53). A calcium gradient has been described in the mammalian epidermis and is thought to play an important role in regulating proliferation and differentiation of keratinocytes (27, 62). A mathematical model has been developed that attributes this gradient to the impermeability of stratum corneum to calcium, the accumulation of calcium in stratum spinosum and granulosum, and the presence of tight junction proteins impermeable to calcium (2, 3).

Calcium-sensing receptor (CaSR) is a G protein-coupled receptor that was first identified in the parathyroid gland (34, 75). It consists of an extracellular domain, seven transmembrane helices, and an intracellular carboxy-terminal tail (87). One of the main functions of CaSR is to regulate plasma Ca²⁺ concentrations. It also modulates a wide variety of functions in different tissues including secretion, channel activity, gene expression, proliferation, wound healing, and cancer (5, 44, 57, 114). The binding of extracellular Ca²⁺ to the receptor’s extracellular domain activates one or more signaling pathways through the heterotrimeric G proteins (G_q/11, G_i, and G_12/13) and subsequent activation of phospholipase C, production of inositol (1,4,5)-trisphosphate and diacylglycerol, resulting in intracellular Ca²⁺ mobilization and activation of mitogen-activated protein kinase cascade (16, 21, 43, 44, 113).

In skin keratinocytes, CaSR plays an important role in epidermal differentiation and in maintaining barrier function. This is supported by several studies in both cell cultures and animal models (50a, 66, 67, 93, 96, 97).

CaSR is present in the esophageal epithelium, and in cultured esophageal cells, it plays a role in Ca²⁺ mobilization (49) and epithelial remodeling (1). The role of CaSR in the esophageal tissue in vivo has not been investigated yet.

The aim of this study is to examine the role of CaSR in the esophagus. For this purpose, we generated a keratinocyte-specific CaSR knockout (KO) model by breeding (CaSR^Fl+/+) mice whose CaSR gene (exon 7) is flanked by two LoxP sequences (18) with transgenic keratin 5 (KRT5)-CreER mice. The recombination in the KRT5-CreER;CaSR^Fl+/+ mice was induced by tamoxifen injections, causing translocation of CreER to the nucleus and deletion of CaSR in esophageal keratinocytes.

In this study, we demonstrate that deletion of exon 7 by recombination decreased the expression of CaSR in the cell membrane of esophageal tissues. Tissues from ^EsoCaSR^−/− mice showed morphological changes that included rete peg elongation, abnormal keratinization, and bacterial buildup on the luminal surface of the esophagus. Expression of adherens junction proteins E-cadherin and β catenin and tight junction proteins claudin-1, claudin-4, claudin-5, and zonula occludens (ZO1) were reduced. Transepithelial resistance was significantly reduced in KO tissues. We propose that CaSR expression is an important factor in maintaining the structure and barrier properties of the esophageal epithelium.

METHODS

Generation and Genotyping of Conditional CaSR Knockout

We bred CaSR^Flox+/+ (CaSR^loxP/loxP) mice [generated by Chang et al. (18)] with transgenic mice expressing Cre-recombinase under the control of the keratin 5 promoter KRT5-CreER-Rosa26 (generously donated by Jianwen Que, Columbia University) to generate KRT5-CreER; Rosa26;CaSR^Flox+/+ compound mutants. Mice were allowed free access to food and water. Mouse genotypes were determined by PCR analyses of genomic DNAs from tail snips using Terra PCR Direct Genotyping kit (Clontech). PCR primers for Rosa were 3′-AAA GTC GCT CTG AGT TGT TAT-5′ (forward), 3′-GCG AAG AGT TTG TCC TCA ACC-5′ (reverse 1), and 3′-GGA GCG GGA GAA ATG GAT ATG-5′ (reverse 2) with an expected PCR product of 340 bp for homozygote mutants, 340 and 650 bp for heterozygotes, and 650 bp for wild type (84). Primers for detection of CreER were 3′-TGC TGT TTC ACT GGT TAT GCG G-5′ (forward) and 3′-TTG CCC CTG TTT CAC TAT CCA G-5′ (reverse), with an expected PCR product of 671 bp. Primer sets for CaSR^Flox+/+ genotyping were 5′-GTG ACG GAA AAC ATA CTG C-3′ (Lox 7UP) and 5′-CGA GTA CAG GCT TTG ATG C-3′primer (Lox-7LOW). Wild-type alleles amplified a 133-bp DNA PCR product, whereas floxed CaSR alleles amplified a 167-bp DNA product.

Tamoxifen Treatment

Adult male and female (6–8 wk old) KRT5-CreER; Rosa26;CaSR^Flox+/+ compound mutants were injected intraperitoneally with tamoxifen (160 mg/kg body wt) dissolved in corn oil once every 48 h for a total of 3 injections. Simultaneously, adult KRT5-CreER; Rosa26 or CaSR^Flox+/+ of the same age group were also treated with tamoxifen and used as controls. At specific time frames ranging from 10 days to 4 wk after the last tamoxifen injection, the mice were euthanized and tissues were collected for the study. Esophageal tissues were stripped of their muscularis externa, and genomic DNA was isolated using a Gene Elute Mammalian Genomic DNA Miniprep Kit (Sigma-Aldrich). To verify the excision of exon 7, esophageal genomic DNA from tamoxifen-treated mice and control littermates was analyzed by PCR using the following forward and reverse primers, respectively: (CCTCGAACATGAACAACTTAATTCGG) and (CGAGTACAGGCTTTGATGC). Real-time (RT)-PCR products were resolved by electrophoresis using a 2.5% agarose gel containing 10 μg/mL ethidium bromide. The approximate size of each product was determined by comparison to a DNA ladder (Invitrogen). A PCR product of 294 bp from the CaSR gene allele indicates the excision of exon 7 by recombination (18) and was confirmed for every experiment.

After the tamoxifen injections, we tested the optimal time for recombination and maximal reduction of CaSR expression in esophageal tissue and found that it ranged from 7 days to 4 wk after the last tamoxifen injection. These data are consistent with other reports in the literature (54, 78). Mice weights remained unchanged over the course of tamoxifen treatment. The animal studies were approved by Tulane University’s Institutional Animal Care and Use Committee.

Ussing Chamber Measurements

After euthanasia, the esophagus was excised and placed in ice-cold HEPES Ringer. Under a stereoscope, the tissue was stripped of its mucularis externa and cut in four pieces (an additional small piece was used to confirm excision of exon 7 by PCR in KO animals). Each piece of tissue, consisting of the mucosa, the submucosa, and the muscularis mucosa, was cut open and placed between two sides of a leucite slider with an opening of 0.031 cm². Each tissue was mounted between two sides of chambers: one luminal and one basolateral. Both sides of the tissues were initially bathed in HEPES Ringer and bubbled with oxygen. The composition of Ringer solutions is 133 mM Na⁺, 119.8 mM Cl⁻, 5.2 mM K⁺, 25 mM HEPES, 1.2 mM Ca²⁺, 1.2 mM Mg²⁺, 2.4 mM ${\text{HPO}}_4^{2-}$, 0.4 mM ${\text{H}}_{\text{2}}{\text{PO}}_{\text{4}}^{-}$, and 10 mM glucose (osmolality 290 mosmol/k H₂O, pH 7.4, gassed with 100% O₂). HEPES was omitted from the acidified solution and the pH was adjusted to 1.6 with HCl acid. The temperature in the chambers was maintained at 37°C using a circulating water bath system. The mini Ussing chambers were fitted with two sets of KCl agar bridges: one for the luminal side and the other for the basolateral sides. The bridges were connected via Ag-AgCl electrodes to an electrometer (Voltage Current Clamp MC6, Physiologic Instruments, San Diego, CA) equipped with a built-in pulse generator for measurement of tissue resistance. The electrometer allows monitoring of transepithelial current or voltage and compensation for electrode asymmetry and fluid resistance. A data acquisition program (Acquire and Analyze 2.3) was used for continuous recording of data. The tissues were left to equilibrate for 30 min.

To evaluate the permeability of the tissues, fluorescein isothiocyanate dextran (FITC-dextran; molecular mass of 4 kDa) was dissolved in HEPES Ringer and added to the luminal solution at a final concentration of 0.5 mg/mL. Samples of 100 μL were taken from the basolateral side of the tissues at determined time intervals and placed in a 96-well plate for determination of fluorescence using a plate reader. A standard dilution curve was plotted for every experiment to determine concentrations. After 2 h of sampling, the recording was stopped and the luminal solution was switched to a Ringer solution titrated to pH 1.6, containing fluorescein at the same concentration as mentioned above. Data collection was resumed for an additional 90 min.

Total RNA Isolation, Reverse Transcription, and Amplification of mRNA

Total RNA was isolated from homogenized tissues using an “Absolutely RNA” kit (Stratagene) according to manufacturer’s instructions. RNA was converted to cDNA using AffinityScript QPCR cDNA Synthesis Kit (Stratagene). Samples were analyzed by RT-PCR (Mx-Pro 3005P, Stratagene) using Brilliant III Ultra-Fast SYBR Green (Agilent Technologies). Primers were obtained from Integrated DNA Technologies (Coralville, IA). Forward and reverse primers for CaSR Exon 7 quantitation were (CGG AAG CTG CCA GAG AAC TT) and (TCG ATG GTG TTC CGT GAA GG), respectively. Relative quantification of gene expression was carried out using standard curves for the target genes and for three reference endogenous genes (β₂-microglobulin, β-actin, and GAPDH). The 2^−ΔΔCT method was used to determine the relative expression of exon 7 in knockout mice with respect to control. The samples were run in triplicates. Statistical analysis using t test (nonpaired) was performed.

Western Blotting

For whole cell protein extraction, tissues were lysed in Cell-Lytic (Sigma) in the presence of protease inhibitors. For separate cytoplasmic and cell membrane protein extraction, we used Thermo Fisher Scientific Mem-PER Plus Membrane Protein Extraction Kit. According to the manufacturer’s instructions, tissue homogenates were permeabilized with a mild detergent to allow the release of soluble cytosolic proteins. A second detergent was used to solubilize membrane proteins. Protein content was quantified using the Pierce BCA Protein Assay (Thermo-Fisher Scientific) and normalized to 1 μg/μL. Proteins were separated using 10% SDS PAGE under reducing conditions (51). Prestained molecular weight markers were run in parallel lanes (Li-Cor, Lincoln, NE). Samples were run in duplicates or triplicates. After electrophoresis, proteins were transferred to a nitrocellulose membrane. The membranes were blocked for nonspecific binding, incubated overnight at 4°C with primary antibodies, washed, and incubated with secondary antibodies conjugated to a fluorescent dye (Li-Cor). The immunoreactive complex was visualized using Li-Cor Odyssey Infrared System and analyzed with system-specific software. β-actin or GAPDH was used as a loading control and for normalization. For each sample, duplicates or triplicates from every experiment were pooled and their mean was calculated. For every gel, the calculated densitometry values from control animals were averaged, and the individual values for control and knockout mice from each gel were calculated as a percentage of that average. Statistical significance was determined using two-tailed unpaired Student’s t test, unless mentioned otherwise. P < 0.05 was considered significant. The membranes were stripped up to three times using Li-Cor stripping buffer as directed by the manufacturer, blocked, and reprobed as described above. The experiments were repeated at least three times on three different tissues. The antibodies used for Western blot (WB) analysis and immunohistochemistry are listed in Table 1 and were validated in previous publications, as shown.

Table 1.

Primary antibodies used for immunohistochemistry and Western analysis

Host Antibody	Immunogen Clone, Source	Protein MW, kD	Concentrations Used
β-catenin antibody Rb monoclonal	Synthetic peptide corresponding to residues near NH2-terminus of human β-catenin (Abcam E247, 32572) (58)	~97	0.03 μg/mL (WB) 0.06 μg/mL (IHC)
CaSR, mouse monoclonal	Amino acid peptide sequence 15–29 at the extracellular NH2-terminus of human CaSR (Sigma-Millipore) (7)	~130	1–2 μg/mL (WB)
CaSR Rb polyclonal	Synthetic peptide derived from the NH2-terminal of rat calcium sensing receptor (Abcam) (42)	130	2–4 μg/mL (WB)
CaSR Rb polyclonal	Amino-acid peptide sequence 111–210 at the extracellular NH2-terminus of human CaSR (Santa-Cruz H-100)	130	2–4 μg/mL (WB)
CaSR Rb polyclonal	Peptide DDYG RPGIE KFREE, corresponding to amino acid residues 216–229 of human CaSR (ACR-004, Alomone) (38)	130	1–2 μg/mL (WB)
E-cadherin Rb monoclonal	Synthetic peptide surrounding residue 780 of human E-cadherin cytoplasmic region (Cell Signaling) (106, 108)	~130	~0.1 μg/mL (WB) 0.2–0.4 μg/mL (IHC)
Claudin-1 Rb polyclonal	Synthetic peptide within human claudin 1 aa 150 to the COOH-terminus (COOH terminal) (Abcam) (81)	23–18	2–4 μg/mL
Claudin-4 Rb polyclonal	Synthetic peptide (COOH-terminus of human claudin-4) (GeneID: 1364), conjugated to KLH (Sigma Millipore) (45)	23	4 μg/mL (IHC)
Claudin-5 Rb polyclonal	KLH-conjugated linear peptide corresponding to human claudin-5 at the cytoplasmic domain (Sigma Millipore) (86)	17	2 μg/mL (IHC)
ZO1 Rb polyclonal	KLH-conjugated linear peptide corresponding to human ZO1 (Sigma-Millipore)	260	2 μg/mL (IHC)
Ki67 Rb polyclonal	Synthetic peptide conjugated to KLH derived from within residues 1200–1300 of human Ki67 (Thermo Fisher Scientific) (109)	>250	1 μg/mL (IHC)
Filaggrin mouse monoclonal	Human foreskin (Santa Cruz) (111)	26–50	1–2 μg/mL (WB)
GAPDH (14C10) Rb monoclonal	Synthetic peptide near the COOH-terminus (Cell Signaling) (104)	32	0.1 μg/mL (WB)
β-actin Rb monoclonal	Synthetic peptide corresponding to the NH2-terminus of human β-actin (Li-Cor) (Cell Signaling) (25)	37	0.2 μg/mL (WB)

CaSR, calcium-sensing receptor; IHC, immunohistochemistry; MW, molecular weight; Rb, rabbit; WB, Western blot; ZO1, zonula occludens-1; NH2 amino terminus, KLH, keyhole Limpet Hemocyanin.

Immunohistochemistry

Sections of control and ^EsoCaSR^−/− esophageal tissues were fixed overnight in 10% formalin, dehydrated, and embedded in paraffin blocks. Thin sections (5 μm) were cut and mounted on gelatin-coated slides. Sections were dewaxed in xylene and rehydrated sequentially in 100%, 95%, and 70% alcohol. Tissue sections were treated with 0.3% H₂O₂-methanol or Bloxall (Vector Laboratories, Burlingame, CA) to block endogenous peroxidase activity and immunostained using standard procedures. Depending on specific recommendations for each antibody, some tissue sections were treated with citrate buffer at high temperatures to unmask antigens. Sections were blocked and incubated with the primary antibody, washed, and incubated with biotinylated secondary antibody against the IgG of the species providing the primary antibody. Sections were then incubated with avidin:biotinylated enzyme complex (Vector Laboratories). Peroxidase activity was detected using diaminobenzidine (Sigma, St. Louis, MO) as a substrate. Specimens were counterstained with hematoxylin, cleared, and mounted in Permount. In some experiments, following the application of the primary antibody, ImmPRESS-alkaline phosphatase (Vector Laboratories) polymerized reporter enzyme staining system was used, and 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (BCIP/NBT) was used as substrate. The antibodies, their concentrations, and their supply sources are listed in Table 1.

Gram Staining

Paraffin sections were dewaxed in xylene, rehydrated in graded alcohols and PBS, and sequentially stained with crystal violet, Gram’s iodine, and Safranin O, then were cleared, dehydrated, and mounted in Permount.

Lectin Binding

Paraffin sections were dewaxed in xylene, rehydrated in graded alcohol, and treated with 0.3% H₂O₂-methanol or Bloxall (Vector Laboratories) to block endogenous peroxidase activity. The slides were incubated with 3% bovine serum albumin to prevent nonspecific binding, then incubated with the biotinylated lectin at a concentration of 5–10 μg/mL with avidin:biotinylated enzyme complex (Vectastain Elite ABC Kit, Vector Laboratories). Peroxidase activity was detected as described above. The sections were subsequently counterstained with hematoxylin, blued in lithium carbonate, washed and dehydrated through alcohol, cleared, and mounted in Permount for study by light microscopy. For negative controls, the tissues were incubated in the absence of lectin or incubated in the presence of lectin that was reacted with the specific eluting (inhibiting) sugar, as indicated in Table 2.

Table 2.

Distribution of lectin staining in control and ^EsoCaSR^−/− knockout esophageal mouse tissues

Lectins	Sugar Specificity	Stratum Basalis		Stratum Spinosum		Stratum Corneum		Bacteria
		Ctrl	KO	Ctrl	KO	Ctrl	KO	Ctrl	KO
Wheat germ agglutinin	(GlcNAc): Neu NAc	−	++	+++ strata	+++ diffuse	+/−	−	+/−	+++
Ulex europus (gorse)	α-L-fucose	−	+/−	+++ strata	+++ diffuse	+/−	−	+	+++
Concanavalin A	mannose	+++	+++	+++	+	+/−	+	+/−	+++
Soybean agglutinin	α- or β-linked GalNAc>galactose	−	−	+++ strata	+ diffuse	-	+/−	+/−	++

Ctrl, control;; Gal, galactose; GalNAc, N-acetyl galactosamine; GlcNAc, N-acetyl glucosamine; KO, knockout; NeuNAc, N-acetyl neuraminic acid, sialic acid. − indicates no staining; +/− indicates mixed weak staining; + indicates weak staining; ++ indicates moderate staining; +++ indicates strong positive staining.

Chemicals

All chemicals were obtained from Sigma-Aldrich, unless otherwise indicated.

Statistical Analysis

Data are presented as means ± SE. Data were analyzed using unpaired Student’s t test; n is the number of observations unless otherwise indicated.

RESULTS

Ablation of Exon 7 of CaSR Gene Following Tamoxifen Injections

Exon 7 encodes the transmembrane and cytoplasmic tail of CaSR, and its ablation results in a truncated protein that does not translocate to the cell membrane and does not respond to changes in extracellular calcium (18, 97). As described in methods, KRT5-CreER; Rosa26;CaSR^Flox+/+ compound mutants were injected intraperitoneally with tamoxifen to induce recombination. Figure 1A shows a diagram of the CaSR gene in CaSR^Flox+/+ mice, in which lox P sequences flank exon 7, and the location of three primers used to identify the floxed sequence (167 bp), the wild-type sequence (133 bp), and the expected DNA fragment from exon 7 deletion (294 bp). Using primers 2 and 3, genotyping of CaSR^Flox+/+ mice (Fig. 1B) shows PCR products from homozygote (lane 4) and heterozygotes (lanes 2 and 3). We next extracted DNA from esophageal tissue, epidermis, and esophageal muscularis externa of KRT5-CreER- CaSR^Flox+/+ mice treated with tamoxifen. PCR results show that only esophageal and epidermal tissues show the truncated 294 bp fragment from exon 7 (Fig. 1C, lanes 2 and 4), indicating recombination in these tissues and ablation of exon 7, as expected. Lane 5 shows esophageal DNA extracted from a control littermate that was not treated with tamoxifen. These results indicate that recombination can only occur in tissues expressing KRT5 and the CaSR^Flox+/+ sequences. This is because Cre-recombinase is under the control of the KRT5 promoter, which is present in esophageal squamous tissue and in epidermis but not in muscle (46, 77). KRT5-CreER mouse model was used successfully to knock out E-cadherin in the mouse esophagus (48). Using RNA extracted from esophageal mucosae of KRT5-Cre-CaSR^Fl+/+ mice treated with tamoxifen (^EsoCaSR^−/−), we analyzed relative expression of exon 7 compared with control littermates by quantitative real-time PCR (Fig. 1D). Exon 7 was reduced by ~55% ± 4% (n = 5, control mice and n = 4, KO mice, P < 0.01), indicating proper recombination and ablation of exon 7. The lack of complete receptor KO is likely due to an uneven distribution of KRT5-Cre in the tissue, which was described as variegated in a number of studies in which the Cre-Lox technique was used to knock out genes (59, 68, 76).

Fig. 1.

Deletion of exon 7 from calcium-sensing receptor (CaSR) gene in mouse esophagus. A: diagram of the CaSR gene in CaSR^Flox+/+ mouse in which lox P sequences flank exon 7 and the location of three primers used to identify the floxed sequence (167 bp), wild-type sequence (133 bp), and the DNA fragment from exon 7 deletion (297 bp). B: genotyping of CaSR^Flox+/+ mice showing PCR products from homozygotes (167 bp, lane 4) and heterozygotes (167 and 133 bp, lanes 2 and 3). Primers 2 and 3 were used, and lane 1 shows the ladder standards. C: PCR analysis of DNA extracted from tissues of a keratin-5 (Krt5) -CreER; CaSR^Flox+/+ mouse treated with tamoxifen, using primers 1 and 3. Only lanes 2 and 4 show the 297 bp fragment from exon 7. Lane 5 is from a littermate mouse that was not treated with tamoxifen. D: quantitative real time PCR results showing relative expression of exon 7 in RNA extracted from Krt5-CreER; CaSR^Flox+/+ mice treated with tamoxifen (^EsoCaSR^−/−) as compared with tamoxifen-treated controls. Exon 7 was reduced by ~55% ± 4% (n = 5 control and 4 knockout, *P < 0.05). UTR, untranslated region.

Expression of CaSR protein in esophageal tissues and effect of ablation of exon 7.

To confirm expression of CaSR protein in the untreated mouse esophagus, we first performed Western blot experiments in control tissues. As shown in Fig. 2, expression at the expected molecular mass band of 130 kDa was confirmed using an antibody against amino acid peptide sequence 15–29 at the extracellular NH₂-terminus of human CaSR (Sigma-Millipore C0493). The immunoblot, using whole tissue lysate (Fig. 2A), showed the expected CaSR band at 130 kDa but also showed other bands at ~90, 65, 35, and 27 kDa, probably indicating the presence of immature or degraded receptors (11, 39, 107). We observed a similar pattern of staining (multiple bands) in a previous study using protein extracts from pig esophagus and human esophageal biopsies, confirming that the presence of multiple bands was not an artifact (1). Other antibodies to CaSR that were raised in rabbit against the NH₂ or the COOH terminus from Alomone (ACR-004), Millipore (AB5630P), Abcam (79038), and Santa Cruz (H-100) showed closely similar patterns of multiband staining. However, in WB experiments, CaSR antibody Sigma C0493 gave us the best results in both the intensity of staining at the expected molecular weight of the protein (130 kDa) and in the sharpness of the band. (See Ref. 1 and Supplemental Fig. S1 at https://doi.org/10.6084/m9.figshare.9925088.v1).

Fig. 2.

Calcium-sensing receptor (CaSR) antibody specificity to CaSR protein in mouse esophageal tissue. Tissue lysates from two control mice were run on the same gel in lanes 1, 2, 3, and 4. The blots from the gel were then separated and simultaneously incubated with CaSR antibody (Sigma C0493). A: the membrane was incubated with the antibody (Ab) overnight. B: the antibody was preincubated for 3 h with the binding peptide before its incubation overnight with the blot. After incubation with the peptide, all the reactive bands were eliminated, except for the faint bands at 25 and 50 kDa corresponding with the light and heavy chains of endogenous immunoglobulins in the sample (mouse tissue and anti-mouse secondary antibody). MW, molecular weight.

To confirm the specificity of this antibody, tissue lysates from control mice were run on the same gel in lanes 1, 2, 3, and 4 (Fig. 2). The blots from the same gel were separated and simultaneously incubated as follows: The first membrane was incubated overnight with the CaSR antibody (Fig. 2A). The second was incubated overnight with the same CaSR antibody (C0493) that had been preincubated for 3 h with the binding peptide (WHSSAYGPDQRAQ) (Fig. 2B). All incubations were performed at 4°C. All blots were stained with the secondary antibody as described in methods. In the blot of Fig. 2B, all the reactive bands were eliminated except for the faint bands at 25 and 50 kDa, corresponding to the light and heavy chains of endogenous immunoglobulins in the sample (mouse tissue and anti-mouse secondary antibody). On the other hand, the blot in Fig. 2A clearly shows the expected bands at 130 kDa and the smaller fragments.

Western blot experiments on whole tissue lysates showed reduced expression of CaSR by 28% ± 5% (P < 0.05) in KO tissues (n = 8) compared with control (n = 4). We then ran Western blot experiments on separate cytoplasmic and membrane extracts (as described in methods) and compared CaSR expression in control and KO extracts. As shown in Fig. 3A, CaSR protein at 130 kDa was absent in cytoplasmic extracts (lane 2) and only expressed in cell membrane extracts of control tissues (lane 5). CaSR protein was also absent in cytoplasmic extracts from KO tissues (lanes 1 and 3). More importantly, CaSR expression in membrane extracts was reduced by ~52% ± 11% in KO tissues (lanes 4 and 6) as compared with control (lane 5). This experiment strongly indicates that the mature (130 kDa) receptor is present in the cell membrane and not in the cytoplasm, and that deletion of Exon 7 reduces its expression significantly (Fig. 3B, n = 5 control and n = 9 KO, P < 0.01).

Fig. 3.

Calcium-sensing receptor (CaSR) expression in cytoplasmic and membrane fractions of esophageal tissues. A: Western blot analysis of protein extracts from cytoplasmic (lanes 1, 2, and 3) and their corresponding membrane fractions (lanes 4, 5, and 6). Lanes 1 and 4 are, respectively, cytoplasmic and membrane extracts from one knockout (KO) mouse; similarly, lanes 3 and 6 are from another KO mouse and lanes 2 and 5 are from one control (C) mouse. The blot was stained with CaSR (Sigma), β-actin, and GAPDH antibodies. The tissues were extracted 10 days after the last tamoxifen injection. B: bar graphs showing relative expression of CaSR to GAPDH in protein extracts from cell membranes of control (n = 5) and ^EsoCaSR^−/− mice (n = 9) *(P < 0.05). MW, molecular weight.

β-actin expression in control and KO tissues.

We used β-actin and GAPDH as loading control proteins in our WB experiments to semiquantitate CaSR protein expression. In whole tissue lysates, there was no apparent difference in expression of both proteins in tissue from KO or control animals. However, when we ran the experiments using cytoplasmic and membrane extracts separately, we observed that β-actin was greatly reduced in the membrane fraction of KO animals but not in the cytoplasmic fraction. This indicates that CaSR deletion had an effect on β-actin expression, evident only in the membrane fraction lysates (Fig. 3A, lanes 4 and 6). On the other hand, GAPDH expression in the membrane fraction did not change significantly upon CaSR deletion, indicating that the changes in β-actin were not due to experimental artifacts.

Histological Features of ^EsoCaSR^−/− Esophageal Tissues

Hematoxylin and eosin.

^EsoCaSR^−/− mice showed no significant differences in appearance, activity, or food and water intake compared with control littermates. Histological examination of formalin-fixed, paraffin-embedded sections of ^EsoCaSR/tissues stained with hematoxylin and eosin (Fig. 4, B and D, ×400 and ×1,000, respectively) showed subtle but significant differences when compared with control mice also treated with tamoxifen (Fig. 4, A and C, ×400 and ×1,000, respectively). The most obvious of these differences in KO mice were the elongation of rete pegs (Fig. 4B, open red arrow), impaired enucleation of cells in the stratum granulosum (Fig. 4D, red chevron), increased number of lamellar bodies or keratohyalin granules (white arrows, compare Fig. 4, C and D), increased waviness of the tissue, aberrant keratinization, and the presence of a sizeable bacterial colony on the luminal surface of the squamous tissue (Fig. 4B, closed green arrow).

Fig. 4.

Hematoxylin and eosin (H&E) staining of mouse tissue sections from control (A and C) and ^EsoCaSR^−/− tissues (B and D). A and B are at ×400 magnification and C and D are at ×1,000; both groups were treated with tamoxifen. Rete peg elongation is clearly seen in B (open red arrow) and bacterial colonies are present on the luminal surface of the squamous tissue (green-filled red arrow). Impaired enucleation is observed in D (red chevron arrow) alongside an increase in number of lamellar bodies or keratohyalin granules (white arrows). Cornified layer is seen in C, whereas in D, a filamentous layer replaces the stratum corneum (SC). The bar represents 100 μm in A and B and 50 μm in C and D. CaSR, calcium-sensing receptor; SB, stratum basalis; SG, straum granulosum; SS, stratum spinosum.

Gram staining.

We performed Gram staining in control and ^EsoCaSR^−/− tissues to examine the presence of surface bacteria. Figure 5, A and C, shows Gram staining in control tissues at magnifications of ×400 and ×1,000, respectively. In control tissues, the presence of bacteria is limited to very few Gram-negative rods on the luminal surface (Fig. 5C). In ^EsoCaSR^−/− tissues (Fig. 5, B and D, ×400 and ×1,000, respectively), a large number of mixed Gram-negative and Gram-positive colonies covered the entire luminal surface of the tissue (red arrows). Bacteria were also observed in the deeper layers of the tissue, indicating that deletion of CaSR in the tissue not only promotes the adhesion and growth of bacteria on the epithelial surface but also allows bacteria to invade the deep layers of the tissue. This experiment was repeated in three different controls and three KO tissues.

Fig. 5.

Gram staining in control (A and C) and ^EsoCaSR^−/− tissues (B and D). A and B are at ×400 magnification and C and D are at ×1,000. Bacterial colonies are much more widespread and numerous in knockout (KO) tissues. Red arrows indicate bacterial colonies. The experiment was repeated on tissues from three control and three KO mice. Bar represents 100 μm in A and B and 50 μm in C and D. CaSR, calcium-sensing receptor; SB, stratum basalis; SC, stratum corneum; SG, stratum stratum granulosum; SP, stratum spinosum.

Lectin staining.

Lectin binding is a useful tool to identify glycosylation patterns in differentiating cells and to examine flora binding to epithelial surfaces (30, 101). For further morphological examination of KO tissues and simultaneous detection of associated bacteria, we performed staining with the following biotinylated lectins: wheat germ agglutinin (WGA), Dolichos biflorus, Ulex europaeus (gorse), peanut agglutinin, concanavalin A, Vicia villosa, and soybean agglutinin (SBA). The results are shown in Figs. 6–8 and are summarized in Table 2.

Fig. 6.

Histochemical staining with wheat germ agglutinin (WGA; lectin) in control (A and C) and ^EsoCaSR^−/− tissues (B and D). A and B are at ×400 magnification, and C and D are at ×1,000. The regular stratified pattern of the cell membranes in stratum spinosum is seen in A and C with a considerable amorphous cornified layer [stratum corneum (SC)], whereas in the knockout (KO) tissues, the staining pattern is irregular and the uppermost amorphous cornified layer is replaced by a filamentous irregular layer on the epithelial surface. Large pyknotic cells are observed (open black arrows). Bacterial colonies are positive to WGA and much more widespread and numerous in KO tissue. Peroxidase activity was detected using diaminobenzidine as a substrate (brown staining indicates positive staining). Specimens were counterstained with hematoxylin. Stratum basalis (SB) is indicated by the red arrow. Bar indicates 100 μm in A and B and 50 μm in C and D. CaSR, calcium-sensing receptor; SG, stratum granulosum; SP, stratum spinosum.

Fig. 8.

Histochemical staining with concanavalin A (ConA; lectin) in control (A and C) and ^EsoCaSR^−/− (B and D) tissues. A and B are at ×400 magnification and C and D are at ×1,000. In knockout (KO) tissues, the staining pattern indicates projections of the stratum corneum into the lumen and irregular stratification and nuclear distribution. Bacterial colonies are positive to ConA (green filled arrow) and more widespread and numerous in KO animal. Brown staining indicates positive staining to the lectin. Specimens were counterstained with hematoxylin. Stratum basalis (SB) is indicated by the red arrow. Open red arrow indicates a pyknotic cell. Bar indicates 100 μm in A and B and 50 μm in C and D. CaSR, calcium-sensing receptor; SC, stratum corneum; SG, stratum granulosum; SP, stratum spinosum.

Staining with WGA showed normal stratification in control tissues (Fig. 6, A and C at ×400 and ×1,000, respectively), whereas in the KO tissue the stratification was irregular (Fig. 6, B and D at ×400 and ×1,000, respectively). The uppermost amorphous cornified layer was replaced by a filamentous irregular layer. In the stratum granulosum of the KO tissue, the granules (lamellar bodies, keratohyalin) were more abundant than in control tissue and their distribution was irregular. A number of pyknotic cells were also present in the KO tissue (Fig. 6D, open arrows). A considerable layer of bacteria, positive to lectins, is present in the ^EsoCaSR^−/− tissues.

Lectin staining with Ulex and SBA in control tissues (Fig. 7, A and C) also showed a regular stratified pattern of the cell membranes in stratum spinosum, whereas in the ^EsoCaSR^−/− tissues, the staining of the cell membranes was irregular (Fig. 7B) with enhanced cytoplasmic staining (Fig. 7D). There was also an increase in the number and size of the granules (lamellar bodies, keratohyalin) in the granular layers (Fig. 7, B and D, red arrows) and several pyknotic cells (Fig. 7D, open arrow).

Fig. 7.

Histochemical staining with the lectins Ulex and soybean agglutinin (SBA) in control (A and C) and knockout (B and D) tissues, all at ×1,000 magnification. The regular stratified pattern of the cell membranes in stratum spinosum (SP) is seen in A and C with a considerable amorphous cornified layer [stratum corneum (SC)], whereas in the ^EsoCaSR^−/− tissues, the staining pattern was irregular and the uppermost amorphous cornified layer is replaced by a filamentous irregular layer on the epithelial surface. Large pyknotic cells are indicated by the open arrow. Specimens were counterstained with hematoxylin. Brown staining indicates positive staining to the lectin. Bar indicates 50 μm. Stratum basalis (SB) is indicated by the solid red arrow. CaSR, calcium-sensing receptor; SG, stratum granulosum.

The distribution of concanavalin A staining is also notably modified in KO tissues, revealing projections of the stratum corneum into the lumen and irregular stratification (Fig. 8B). Nuclear membrane staining, which is evident in control tissues (Fig. 8C), is reduced in KO tissues, indicating possible modified glycosylation patterns and decreased expression of α-D-mannose (Fig. 8, B and D).

E-Cadherin Expression in ^EsoCaSR^−/− Mice

To investigate the effect of CaSR deletion on cell-cell junction proteins, we performed Western blot experiments on whole tissue lysates from control and ^EsoCaSR^−/− esophageal tissue. Figure 9A shows an immunoblot stained with an antibody to E-cadherin; lanes 2 and 3 are the extracts from control tissues, whereas lanes 4 and 5 are extracts from KO tissues. Figure 9B shows the summary of relative expression of E-cadherin at the expected molecular mass of 130 kDa to β-actin. E-cadherin was significantly reduced by 23% ± 4% in the KO tissues (n = 12 control and n = 23 KO, P < 0.01). It is important to note here that expression of β-actin in the KO tissues was reduced only in the membrane extracts and not in whole tissue lysates.

Fig. 9.

E-cadherin expression in control and ^EsoCaSR^−/− mice. A: immunoblot analysis of E-cadherin in lysates from control mouse esophageal tissue (lanes 2 and 3) and ^EsoCaSR^−/− (lanes 4 and 5). The blot was stained with antibodies to E-cadherin and to β-actin. Expression of E-cadherin was decreased in knockout (KO) mice. B: bar graph showing relative expression of E-cadherin (expected molecular mass ~130 kDa) in control (n = 10) and KO tissues (n = 23) *P < 0.01. C and D: tissues from control and KO mice stained with an antibody to E-cadherin (indigo indicates positive staining; tissue counterstained with Fast Nuclear Red). Staining intensity is reduced in KO. C and D are at ×400 magnification; bar indicates 100 μm. CaSR, calcium-sensing receptor.

To further confirm the observed changes in E-cadherin, we performed immunohistochemistry experiments using an antibody to E-cadherin, as described in methods. Figure 9C shows E-cadherin staining in tissue sections from control mice. E-cadherin staining intensity is most prominent in the basal and suprabasal layers and decreases progressively going toward the lumen. Figure 9D shows tissue from a KO mouse, in which the intensity of E-cadherin staining is remarkably reduced in all layers.

β-Catenin Expression in ^EsoCaSR^−/− Mice

We performed experiments similar to those described above to compare the expression of β-catenin in ^EsoCaSR^−/− and control mice. Figure 10A shows an immunoblot stained with an antibody to β-catenin (Table 1); lanes 1 and 2 are the extracts from KO tissues, whereas lanes 3 and 4 are extracts from control tissues. Figure 10B shows the summary of relative expression of β-catenin at the expected molecular mass of 97 kDa to β-actin. β-catenin was significantly reduced by 36% ± 9% in the KO tissues (n = 8 control and 12 KO, P < 0.05).

Fig. 10.

β-catenin expression in ^EsoCaSR^−/− mice. A: immunoblot analysis of lysates from ^EsoCaSR^−/− (lanes 1 and 2) and control mice (lanes 3 and 4). The blot was stained with antibodies to β-catenin and to β-actin. Expression of β-catenin (expected molecular mass 97 kDa) was decreased in knockout (KO) mice. B: bar graph showing relative expression of β-catenin in control (n = 8) and KO mice (n = 12) (*P < 0.01). C and D: tissues from control and KO mice stained with an antibody to β-catenin (indigo indicates positive staining; tissue counterstained with Fast Nuclear Red). Staining intensity is reduced in KO. C and D are at ×400 magnification; bar indicates 100 μm. CaSR, calcium-sensing receptor; MW, molecular weight.

We also performed immunohistochemistry experiments in control and KO tissues using the antibody to β-catenin. Figure 10C shows β-catenin staining in tissue sections from control mice. Similar to E-cadherin, β-catenin staining intensity is most prominent in the basal and suprabasal layers and decreases progressively going toward the lumen. Figure 10D shows tissue from a KO mouse and indicates that the intensity of β-catenin staining is remarkably reduced in the basal layers.

Claudin-1, Claudin-4, Claudin-5, and ZO1 Expression in ^EsoCaSR^−/− Mice

Claudin-1 is a tight junction protein that plays an important role in maintaining the barrier function in stratified epithelia (33). Using whole tissue lysates, we performed WB experiments on tissues from control and KO mice and stained the blot with claudin-1 (Table 1) and β-actin antibodies. Our data from 4 control and 5 KO tissues indicate that the expression of claudin-1 (molecular mass 28 kDa) was reduced by 31% ± 7% in KO tissues (P < 0.05, Fig. 11, A and B). Immunohistochemistry experiments showed delineation of the cell membranes in control tissues (Fig. 11C) and confirmed decreased expression of claudin-1 in KO tissues (Fig. 11D).

Fig. 11.

Claudin-1 expression in mouse esophageal tissue lysates from control and ^EsoCaSR^−/− mice. A: immunoblot analysis using whole tissue lysates from knockout (KO; lanes 1 and 2) and control (lanes 3 and 4) mouse tissue. The membrane was stained for claudin-1 and β-actin. B: bar graph showing the expression of claudin-1 at 28 kDa in control and KO tissues. claudin-1 is significantly decreased in knockout (KO) tissues (n = 6 control and 10 KO, *P < 0.05). C and D: tissues from control and ^EsoCaSR^−/− mice stained with claudin-1 antibody. Intensity of claudin-1 staining (indigo) is much reduced in KO tissue compared with control. There is clear delineation of cell membranes in control tissues that is absent in KO tissues. CaSR, calcium-sensing receptor.

Claudin-4 and claudin-5 have been identified in human and mouse keratinocytes (19, 41, 70). Reduced expression of these proteins is known to cause a decrease in the barrier properties of epithelia (20, 50). Immunohistochemistry using specific antibodies to claudin-4 and to claudin-5 showed precise delineation of the cell membranes with claudin-4 and 5 in control tissues (Fig. 12, A and C) but a diffuse and marked reduction in the intensity of staining in KO tissues (Fig. 12, B and D).

Fig. 12.

Immunolabeling of claudin-4 (A and B) and claudin-5 (C and D) in esophageal tissues from control and ^EsoCaSR^−/− mice. Using specific antibodies, labeling of claudin-4 and claudin-5 showed marked reduction in the intensity of staining in knockout (KO) tissues (B and D) compared with control (A and C). Cell membranes are clearly delineated in control tissues. There is also stratification and clear delineation of cell membranes in control tissues that become absent and diffuse in KO tissues. CaSR, calcium-sensing receptor.

ZO1 proteins are scaffolding proteins that bind to the cell membrane, claudins, and actin filaments (99). They have also been reported to play a role as signaling molecules (36). In our experiments, immunohistochemistry staining in tissue sections from control mice (Fig. 13A) showed intense staining in the nuclei of the cells in the stratum spinosum and a more homogeneous staining in the cell membranes of the upper layers of the tissue. This nuclear localization is consistent with a role of ZO1 in the maturation of the esophageal cells as they progress from the stratum basalis to the stratum corneum. The distribution of ZO1 in the KO tissues was uneven and irregular (Fig. 13B).

Fig. 13.

Histochemical staining of zonula occludens (ZO1) in esophageal tissues from control and ^EsoCaSR^−/− mice. In control tissues (A), ZO1 staining is intense in nuclei of stratum spinosum cells and there is a homogeneous staining in the cell membranes of the upper layers of the tissue. In knockout (KO) tissues (B), the expression of ZO1 is irregular and less defined. CaSR, calcium-sensing receptor.

Rac/Cdc42 in ^EsoCaSR^−/− Mice

The small GTPases Rac and Cdc42 have been reported to play a key role in regulating the cytoskeleton. In a previous study, we found that cultured esophageal epithelial cells responded to cinacalcet, a CaSR agonist, with major changes in cell shape and formation of filopodia (1). To investigate the effect of CaSR deletion in vivo on expression of Rac/Cdc42 in mouse tissues, we performed Western blot experiments on esophageal tissue lysates from 6 controls and 9 ^EsoCaSR^−/− mice. As shown in Fig. 14, A and B, expression of Rac/Cdc42 decreased by 26% ± 5% in ^EsoCaSR^−/− mice compared with control (P < 0.01).

Fig. 14.

Expression of Rac/Cdc42 and filaggrin in control and ^EsoCaSR^−/− mice. A: Western blot of esophageal tissue lysates stained with an antibody to Rac1 and Cdc42. Lanes 1 and 2 are lysates from control mouse, lanes 3 and 4 are from tamoxifen-treated heterozygote CaSR^{Fl +/−} [partial knockout (KO)], and lanes 5 and 6 are lysates from tamoxifen-treated CaSR^Fl+/+ (^EsoCaSR^−/−). Expression of Rac/Cdc42 was decreased in partial KO and in KO mice. B: bar graphs showing relative expression of Rac1 and Cdc42 in control (n = 6) and KO (n = 9). C: Western blot showing filaggrin expression in control tissues (n = 10) (lanes 1 and 2) compared with KO (n = 10) (lanes 3 and 4). D: there was a significant reduction in expression in KO tissues; *P < 0.05. CaSR, calcium-sensing receptor; MW, molecular weight.

Expression of Filaggrin in ^EsoCaSR^−/− Mice

Profilaggrin is a large protein present in the differentiating stratum spinosum in stratified squamous epithelia. It has an NH₂-terminal domain with two S-100 homologous calcium binding sites and multiple filaggrin repeats in addition to a COOH-terminal domain. Proteolytic enzymes process profilaggrin into multiple filaggrin units, which play an important role in binding keratin intermediate filaments to form the cornified layer, an insoluble impermeable barrier that protects the epithelium (72). Tissue sections from ^EsoCaSR^−/− mice stained with hematoxylin and eosin showed an abnormal cornified layer (Fig. 4). Western blot experiments using tissue lysates from KO and control mice indeed showed a reduction in filaggrin expression of 27% ± 4% (P < 0.01, n = 8 control and n = 10 KO, Fig. 14, C and D), consistent with a dysregulated cornification process.

Ki67 in ^EsoCaSR^−/− Mice

Ki67 protein is a marker of proliferation that is expressed during all active phases of the cell cycle (G1, S, G2, and M) but is absent in resting cells (G0). While the distribution of Ki67 is rather homogeneous (Fig. 12A) in the stratum basalis of control tissues, Ki67-positive cells in the ^EsoCaSR^−/− KO tissues show clusters along the basal lamina (Fig. 15B). Moreover, when we calculated the percentage number of Ki67-positive cells in the basal layer of tissues from four ^EsoCaSR^−/− and four control animals, we found that the percentage of Ki67-positive cells to basal cells was significantly decreased (Fig. 15C).

Fig. 15.

Immunostaining for cell proliferation marker Ki67. A and B: tissues from control and knockout (KO) mice stained with a rabbit antibody to Ki67 and ImmPRESS-AP (Vector) Anti-Rabbit IgG (alkaline phosphatase) Polymer Detection Kit (indigo indicates positive staining). Slides were counterstained with Nuclear Fast Red. A and B are at ×400 magnification; bar indicates 100 μm. C: graph represents the percentage of Ki67-positive cells relative to Nuclear Fast Red-positive basal cells from a set of four KO tissues and four control tissues. *P < 0.05. CaSR, calcium-sensing receptor.

Ussing Chamber Measurements

Our data, based on Western analysis and immunohistochemistry, clearly indicate a significant change in expression and localization of junction proteins in CaSR KO animals. This suggests that the barrier properties of this tissue may also be affected. To functionally test this possibility, we conducted experiments in Ussing chambers on KO and control tissues to measure transepithelial resistance and permeability. In those experiments, the esophagi were stripped and mounted in mini Ussing chambers as described in methods. After equilibration, fluorescein dextran was added to the luminal solution and the tissue voltages were clamped to zero for continuous measurement of tissue resistances.

Transepithelial resistance.

The average transepithelial electrical resistance in control tissues was 2,990 ± 587 Ω·cm² (32 tissues, 8 mice), a value similar to previously reported values in rabbit esophageal tissues (89). On the other hand, the average resistance in KO tissues was 1,223 ± 586 Ω·cm² (16 tissues, 4 mice), a value significantly lower than control (Fig. 16A, unpaired t test, P < 0.05). At 120 min, the luminal solution was replaced with an acidic Ringer solution at pH 1.6 and measurements were resumed. However, the decrease in resistance in control and KO tissues did not reach significance following acid treatment.

Fig. 16.

Transepithelial resistance (TEER) and apparent FITC-dextran flux in tissues from control and ^EsoCaSR^−/− mice. A: transepithelial resistance of tissues mounted in Ussing chambers was 2,990 ± 587 Ω·cm² in control tissue (n = 32) but only 1,223 ± 586 Ω·cm² in knockout (KO) tissues (n = 16, P < 0.05). B: increased paracellular transport of FITC-dextran (4 kDa) in KO tissues (triangle markers) compared with control (circles). At each time point, the amount of FITC-dextran in the bath (nmoles) was calculated from the measured fluorescein concentration. All readings were calculated relative to initial reading at 60 min of each experiment. Acid exposure of the tissue was done by replacing the luminal fluid with acidified Ringer (pH 1.6) at 120 min. As indicated, permeability to FITC-dextran in KO tissues was significantly higher than in control. *P < 0.01, analyzed by two-factor ANOVA with replication. CaSR, calcium-sensing receptor.

Fluorescein flux.

FITC-dextrans of different molecular weights are often used to measure paracellular permeability over time. The barrier properties of the tissue are determined by the tight junctions and by the paracellular components of the epithelial cells forming the tissue. We chose 4 kDa (14 Å) fluorescein dextran because it is a good marker for the paracellular permeability in esophageal tissue (31, 90). Fluorescein dextran accumulation was measured in nmol/cm² and was normalized to the average amount of fluorescein accumulated at 60 min (initial value) in basolateral baths of each experiment. This amount of accumulated fluorescein in the bath was not significantly different over the first 120 min of the experiment between control and KO tissues. However, when the luminal solution was replaced with an acidified Ringer solution (pH 1.6), the relative amount of fluorescein accumulated was significantly higher in KO tissues as compared with controls (Fig. 16B).

DISCUSSION

This is the first report about the role of CaSR in the esophagus in an in vivo mouse model. Knocking out CaSR from the esophageal basal cells using a Cre/Flox mouse resulted in a decrease in expression of adherens junction proteins, impaired keratinization, and bacterial colonization of the epithelium. These findings indicate that in the esophageal epithelium, CaSR plays an important role in the regulation of cell-cell junctional proteins and in the maintenance of the epithelial barrier.

In skin keratinocytes, calcium and CaSR have been reported to play an important role in proliferation, differentiation, and barrier function (94, 95). Targeted deletion of CaSR from the skin caused the loss of the extracellular epidermal calcium gradient and an increase in skin permeability that was most prominent in the neonatal stage (97). On the single cell level, transient elevation of cytoplasmic calcium ion concentration has been demonstrated to precede cornification (64).

Our data indicate an important role of CaSR in the esophagus. We have previously demonstrated in a primary culture of esophageal cells that sustained CaSR activation by the calcimimetic cinacalcet caused the cells to lose their epithelial phenotype and their adherens junction complexes and to acquire extensive filopodia. These modifications of the actin cytoskeleton involved cleavage of E-cadherin and a decrease in β-catenin expression (1). We attributed those changes to the disruption of Ca²⁺ signaling and to altered distribution and internalization of the CaSR in the cells upon chronic activation of the receptor by calcimimetic cinacalcet.

Our present findings confirm that CaSR is present in the native mouse esophageal cell, predominantly in the cell membrane, as demonstrated by WB experiments on membrane protein extracts compared with cytoplasmic fractions (Fig. 3A). Knocking out the receptor caused a decrease in mRNA and in protein levels that was more pronounced in the membrane fraction of the cell. To our knowledge, this is the first report demonstrating conditional knockout of CaSR in the esophagus.

Cell-Cell Junctions

The barrier function of the esophagus is reported to be at least partially dependent on the integrity of the cell-cell junctional proteins. In fact, the hallmark of reflux disease is the presence of “dilated intercellular spaces” in the esophageal epithelium (69). CaSR ablation in our mouse model resulted in a significant decrease in β-catenin detected both by immunohistochemistry and WB (Fig. 10). The decrease in β-catenin was accompanied by a decrease in E-cadherin (Fig. 9) and a decrease in β-actin expression only observed in the cell membrane fraction of tissue lysates (Fig. 3A).

Claudin is a large family of integral membrane proteins considered to be the major players in the formation and maintenance of the permeability barrier in epithelia (33). Claudin-1, 4, and 5 are barrier-forming isoforms reported to limit cation transport and to increase transepithelial resistance (TEER) across tissues (40). Our experiments demonstrated a decrease in expression in the above-mentioned claudins in the KO tissues (Figs. 11 and 12).

The decrease in expression of the above-mentioned adherens and tight junction proteins was accompanied by a decrease in expression of ZO1 in KO tissues. ZO1 is usually expressed as a continuous band around polarized confluent epithelial cells. However, in nonconfluent cultured cells, ZO1 can also localize to the nucleus (36). It is of interest that in our experiments, ZO1 was expressed in the nucleus in the control and KO tissues (to a lesser extent), indicating a possible role of ZO1 in the progression and maturation of esophageal keratinocytes.

These findings support the hypothesis that CaSR deletion in the esophageal epithelium affects the formation of junctional complexes in the cell membrane and subsequently the adhesive properties of the membranes. It is well established that deletion of exon 7 from CaSR results in a truncated protein that does not translocate to the cell membrane, does not respond to changes in extracellular calcium CaSR, and does not activate phospholipase C (18, 97). It remains to be determined whether the observed morphological changes in KO tissues are caused by changes in Ca²⁺ levels (extracellular or intracellular) or by other CaSR-related signaling proteins like filamin, Rho A, and RhoGEF (74). The decrease in β-catenin expression indicates that transcription signaling by β-catenin in the cells is also modified (37). The role of the CaSR in esophageal tissue is consistent with its role in skin keratinocytes, in which it mediates adherens junction formation and cell differentiation (96, 97).

The decrease in β-actin expression in the membrane fraction of ^EsoCaSR^−/− mucosa confirms the close correlation between calcium signaling and the actin cytoskeleton (98, 100). β-actin plays a major role in cell motility, migration, and growth (14, 23), and small fluctuations of calcium are known to induce major changes in the distribution of cytoskeletal actin (103). In our WB experiments, deletion of CaSR caused redistribution of actin in the cells; actin decreased in the cell membrane extracts only, as shown in Fig. 3A, whereas total β-actin remained constant in whole cell lysates. It is also known that the cadherin-catenin complex is linked to and regulates the actin cytoskeleton and that decreased expression of these proteins can alter the distribution of actin (13, 105). The redistribution of β-actin in our experiments was likely a contributing factor to the observed morphological changes in the tissues, as shown in Fig. 4.

The effect of CaSR signaling on actin is further demonstrated by a significant decrease in Rac/Cdc42 expression in CaSR KO mice (Fig. 14, A and B). Rac/Cdc42 are a family of small guanosine triphosphatases that play an important role in actin remodeling in a variety of cells (8, 65, 79). Rac1-effector proteins include a protein complex involved in lamellipodia formation. These data are consistent with the role of CaSR in cell remodeling (1) and motility (10).

Transepithelial Resistance and Fluorescein Flux

Collectively, the downregulation of tight and adherence junction proteins in ^EsoCaSR^−/− suggest that the barrier properties of the esophageal tissue may be compromised in KO animals. To test this possibility, we measured TEER and FITC-dextran flux in control and KO tissues mounted in Ussing chambers. We also measured the relative changes in TEER and FITC-dextran transport after the tissues were exposed to acid (pH 1.6). Acid exposure was aimed at testing whether acid treatment would cause damage that might affect transepithelial transport differently in KO and control tissues.

A decrease of transepithelial resistance measurements is often correlated with an increase in transepithelial ion permeability, whereas a change in FITC-dextran flux is often attributed to paracellular changes in permeability. In our experiments, initial TEER was significantly lower in KO tissues compared with control (Fig. 16A). This finding is consistent with our hypothesis that CaSR-mediated changes of junctional complexes altered transepithelial permeability. After acid treatment, TEER was slightly, but not significantly, reduced in control or KO tissues. Several possibilities may explain this, including the relatively low initial TEER of KO tissues. It is well established that acid at pH 1.6 alone (without pepsin or bile acids) does not cause significant changes in TEER (82, 91).

Luminal-to-serosal FITC-dextran flux also showed differences between control and KO tissues (Fig. 16B). Initial flux of FITC-dextran up to 120 min (measured as the rate of change in concentration per unit time) were not different between KO and control tissues. However, after acid exposure at 120 min, relative FITC-dextran flux in KO tissues continued to increase, whereas that of control tissues remained relatively low. This is consistent with the increased paracellular permeability in KO tissues and may also indicate the lack of significant damage to control tissue with acid exposure of pH 1.6.

Filaggrin, Lectins, and Bacterial Invasion of ^EsoCaSR^−/− Mucosae

Filaggrin is an epidermal protein that has been reported to play a major role in skin and esophageal disease (60, 110). It is synthesized in the stratum granulosum and then dephosphorylated and proteolytically processed in the upper layers of the epithelium, where it binds keratin filaments to form the stratum corneum (88). Our experiments demonstrated a marked decrease in filaggrin expression in the ^EsoCaSR^−/− mucosa (Fig. 14, C and D). The mechanism by which filaggrin is reduced in our KO model could be the result of a defective Ca²⁺-dependent proteolytic process and requires further investigation.

Lectins are plant-derived proteins that bind specific carbohydrate residues in cells. They have been used to study glycosylation and carbohydrate composition and as markers for differentiation in the gut and other tissues (12, 85). In our experiments, we used lectin staining for two purposes: to show differences in glycosylation patterns between control and KO tissues and to stain sugar residues in the walls of the bacterial colonies in KO tissues (83). WGA has binding affinity to N-acetylglucosamine, SBA to α-L-fucose, and Ulex to N-acetylgalactosamine (6, 15, 26). In control tissues, these lectins all quite accurately delineated the cell membranes in the stratum spinosum and granulosum. Although lectin staining was still positive in ^EsoCaSR^−/− tissues, the stain distribution was mostly cytoplasmic and diffuse and underscored impaired stratification. Lectin staining also highlighted the presence of bacterial colonies and the intense rete peg elongation and spiking projections of the epithelial surface.

A novel and unexpected finding in our experiments was the invasion of the esophageal tissue by bacteria in ^EsoCaSR^−/− mice. Several factors could play a role in facilitating bacterial invasion in this setting. First, expression of adherens and tight junction proteins is decreased, diminishing the barrier properties of the epithelium. Second, the cell’s ability to resist bacterial invasion through the regulation of the cytoskeleton and the Rho and Rac small GTPases is likely decreased in KO tissues, facilitating bacterial adhesion (9, 63, 92). Third, calcium-regulated proteins, such as calprotectin, that are crucial for keratinocyte protection against bacterial invasion (17) may be nonfunctional or decreased because of changes in intracellular Ca²⁺. Further studies are needed to investigate these possibilities.

Mechanisms of Epithelial Remodeling in ^EsoCaSR^−/−

In a normal healthy esophagus, basal cells proliferate and migrate toward the lumen, where they become terminally differentiated and shed. This proliferation-differentiation process is responsible for the ability of the basal cells to regenerate a healthy epithelium. We have shown that deletion of CaSR in the mouse esophagus caused epithelial remodeling that can be best explained by Ca²⁺-dependent modification of the proliferation-differentiation axis. This is underscored by our experiments indicating altered distribution and reduced relative expression of the proliferation marker Ki67. Two possible explanations exist for the observed decreased ratio of Ki67-positive cells to basal cells. First, the number of actively proliferating cells may be decreased; second, the number of basal, nonproliferating, nondifferentiated cells may be increased. The latter can be observed in multiple sections of esophageal tissues (Fig. 4B) and could be explained by the impaired progression of the cells from the basal layers to the stratum spinosum. In control tissues, Ki67-positive cells are homogeneously distributed along the basal lamina, whereas in KO tissues, Ki67-positive cells form clusters, so that some proliferating cells are spatially disturbed and are not sitting against the basal lamina (Fig. 15, A and B). This possibility could explain fingering of the tissue, the elongation of rete pegs, and the waviness of the epithelium in ^EsoCaSR^−/− tissues (71).

Impaired Keratinization

A striking microscopic feature of the ^EsoCaSR^−/− is the absence of a defined stratum corneum and the presence of an abundant pool of lamellar bodies. The lamellar bodies observed in the KO tissues are larger in size than in control tissues (Fig. 4, C and D) and seem to be located intracellularly in the KO tissues. The abnormal secretion of lamellar bodies in the skin has been tightly linked to the disruption of the calcium gradient in the tissue (62). The formation of stratum corneum in skin is dependent on the proteolytic degradation of corneodesmosomes in the uppermost layers of stratum granulosum and on the proper formation and extrusion to the extracellular space of lamellar bodies (32, 47, 55). The abnormal size and distribution of lamellar bodies in ^EsoCaSR^−/− tissues, combined with the absence of a well-defined stratum corneum, is a strong indicator that the calcium gradient is disrupted in the tissue. In human skin, defects in secretion of lamellar bodies and abnormal lipid composition lead to a defective barrier and microbial invasion, as observed in atopic dermatitis and hyperkeratosis (28, 80). Esophageal hyperkeratosis is often observed in association with Barrett’s esophagus and in adenocarcinoma, and bacterial colonization is seen in 31% of cases (86a).

Several possible mechanisms working in tandem could account for the observed effects of CaSR deletion on esophageal squamous tissue: 1) intracellular calcium signaling is interrupted, thus affecting the downstream cell signaling mediators in the cell; 2) cell-to-cell interaction is modified via changes in adherens and tight junction proteins, causing abnormal cell shedding and desquamation; and 3) cell to extracellular matrix signaling is modified. The latter is supported by the observations that the basement membrane shape is modified (53a), there is an expansion of the submucosa, and the epithelium is vulnerable to bacterial invasion. Further experiments are needed to examine each of these possibilities.

In conclusion, we provide the first evidence that deletion of CaSR from esophageal tissue weakens cell-cell junctions and disrupts stratification and cornification, causing barrier disruption and bacterial colonization of the tissue. Thus, CaSR seems to play an important role in maintaining the structure and function of the esophageal epithelium. The findings of this study may have significant translational relevance in humans with gastroesophageal reflux disease and other esophageal diseases in which alterations in esophageal cell-cell junctions lead to a compromised mucosal barrier.

GRANTS

This work was supported in part by U54-GM-104940 from the National Institute of General Medical Sciences of the NIH, which funds the Louisiana Clinical and Translational Science Center. Other sources of funding include a VA Merit Award (Bx 001513), a Carol Lavin Bernick Faculty Grant, and a Tulane Bridge Fund Grant.

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Veterans Administration.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.L.N., and S.M.A.-N. conceived and designed research; K.L.B., M.T.I., A.G.H., and S.M.A.-N. performed experiments; N.L.N, S.M.A.-N. and A.G.H. analyzed data; N.L.N and S.M.A.-N. interpreted results of experiments; C.-L.T provided mouse model; S.M.A.-N. prepared figures; N.L.N. and S.M.A.-N. drafted manuscript; N.L.N., and S.M.A.-N. edited and revised manuscript; N.L.N., C.-L.T., K.L.B.; M.T.I. A.G.H., and S.M.A.-N. approved final version of manuscript.