Na+/Ca2+ exchangers regulate the migration and proliferation of human gastric myofibroblasts

Lajos V. Kemény, Andrea Schnúr, Mátyás Czepán, Zoltán Rakonczay Jr., Eleonóra Gál, János Lonovics, György Lázár, Zsolt Simonka, Viktória Venglovecz, József Maléth, Linda Judák, István B. Németh, Kornélia Szabó, János Almássy, László Virág, Andrea Geisz, László Tiszlavicz, David I. Yule, Tibor Wittmann, Andrea Varró, Péter Hegyi


Gastrointestinal myofibroblasts are contractile, electrically nonexcitable, transitional cells that play a role in extracellular matrix production, in ulcer healing, and in pathophysiological conditions they contribute to chronic inflammation and tumor development. Na+/Ca2+ exchangers (NCX) are known to have a crucial role in Ca2+ homeostasis of contractile cells, however, no information is available concerning the role of NCX in the proliferation and migration of gastrointestinal myofibroblasts. In this study, our aim was to investigate the role of NCX in the Ca2+ homeostasis, migration, and proliferation of human gastrointestinal myofibroblasts, focusing on human gastric myofibroblasts (HGMs). We used microfluorometric measurements to investigate the intracellular Ca2+ and Na+ concentrations, PCR analysis and immunostaining to show the presence of the NCX, patch clamp for measuring NCX activity, and proliferation and migration assays to investigate the functional role of the exchanger. We showed that 53.0 ± 8.1% of the HGMs present Ca2+ oscillations, which depend on extracellular Ca2+ and Na+, and can be inhibited by NCX inhibitors. NCX1, NCX2, and NCX3 were expressed at both mRNA and protein levels in HGMs, and they contribute to the intracellular Ca2+ and Na+ homeostasis as well, regardless of the oscillatory activity. NCX inhibitors significantly blocked the basal and insulin-like growth factor II-stimulated migration and proliferation rates of HGMs. In conclusion, we showed that NCX plays a pivotal role in regulating the Ca2+ homeostasis, migration, and proliferation of HGMs. The inhibition of NCX activity may be a potential therapeutic target in hyperproliferative gastric diseases.

  • human myofibroblast
  • sodium-calcium exchanger
  • calcium oscillation

myofibroblasts are contractile, nonexcitable, transitional cells between fibrocytes and smooth muscle cells (17). Myofibroblasts transform from tissue fibroblasts by transforming growth factor-β1 (TGF-β1) and are localized to the subepithelium throughout the whole gastrointestinal tract (7). They are crucial for the production of extracellular matrix and morphogenesis, and they also take part in the inflammatory process related to tissue repair and secrete a broad spectrum of bioactive molecules, altering the microenvironment around epithelial cells in response to different noxa (18, 28). It has been shown in mice and rats that fibroblasts contribute to gastric and esophageal ulcer healing (4, 24, 30). Their involvement in the regeneration of the gastric mucosa after ethanol-induced gastric damage has also been shown in rats (29).

Besides their physiological roles, in pathophysiological conditions, myofibroblasts contribute to chronic gastritis and to tumor development of gastric cancer too (13, 27). Moreover, cancer-associated myofibroblasts reside in the tumor mass after the malignant transformation and promote the angioneogenesis and metastasis formation of the malignant cells (7). There is evidence now that these cells may originate at least partly from bone marrow from mesenchymal stem cells, and they contribute to the formation of mesenchymal stem cell niche and promote tumor growth in mice (33). A further relevance of human gastric myofibroblasts (HGMs) is that they might serve as a predictive factor of the outcome of gastric cancer: high tumor TGF-β1 activity levels in tumor-derived myofibroblasts significantly correlated with worse survival in gastric cancer patients (14).

Intracellular free Ca2+ concentration ([Ca2+]i) is a signal important for cellular processes, like motility and cell division, which determine the behavior of a cell in a particular environment (34). There are a number of pathways by which [Ca2+]i is modulated according to the needs of cell function (32). Mainly, Ca2+ can enter the cell from the extracellular space into the cytosol via nonspecific cation channels, voltage-sensitive Ca2+ channels, and Na+/Ca2+ exchangers (NCXs). It is known that NCX plays an important role in regulating the [Ca2+]i in several cell types (3). NCX alters the migration of rat tendon fibroblasts and migration and proliferation in human pulmonary artery smooth muscle cells (36, 43). Moreover, it has been reported that Ca2+ entry-mode operation of NCX is required for des-Arg10-kallidin- and TGF-β1-stimulated fibrogenesis and participates in the maintenance of the myofibroblast phenotype (35).

The plasma membrane protein NCX is a ubiquitously expressed protein with three isoforms that all can function in forward (also known as Ca2+ exit) mode, or in reverse Ca2+ entry mode, depending on the membrane potential and the chemical gradient of Na+ and Ca2+ (25, 26).

Because there is not enough information on how human myofibroblasts regulate [Ca2+]i, we set out to characterize the Ca2+ homeostasis of human gastrointestinal myofibroblasts, focusing on HGMs and on NCX, and to investigate the relationship between NCX and cell function, such as migration and proliferation.


Chemicals and solutions.

Chemicals and solutions used for cell culture were purchased from Sigma-Aldrich (Budapest, Hungary). All reagents for immunocytochemistry were purchased from Jackson Immunoresearch (Stanford, CA). The mouse anti-NCX3 antibody was a kind gift from Dr. Michela Ottolia (Cedars-Sinai, Los Angeles, CA).

All secondary antibodies, the Target Retrieval Solution, and the Fluoromount medium were purchased from Dako (Glostrup, Denmark). Vectashield mounting medium was from Vector Laboratories (Peterborough, UK).

Polyclonal CD117 (Labvision), DOG1 (clone SP31; Labvision), platelet-derived growth factor receptor (PDGFR) α (epitope specific clone; Labvision), and ki67 (clone B56; Histopathology, Pécs, Hungary) were used in the immuncytochemistry and immunohistochemistry experiments. Chemicals and reagents for PCR techniques were purchased from Promega (Southampton, UK), unless indicated otherwise. All primers were purchased from Bio Basic (Markham, Canada).

The solvent of acetoxymethyl 2-{5-{bis[(acetoxymethoxy-oxo-methyl)methyl]amino}-4-[2-(2-{bis[(acetoxymethoxy-oxo-methyl)methyl]amino}-5-methyl-phenoxy)ethoxy]benzofuran-2-yl}oxazole-5-carboxylate (fura 2-AM; Invitrogen, Paisley, UK) and for Na+-binding benzofuran isophthalate-AM (SBFI-AM; Invitrogen) dyes and for the NCX inhibitor 3-amino-6-chloro-5-[(4-chloro-benzyl)amino]-N-{[(2,4-dimethylbenzyl)amino]iminomethyl}-pyrazine-carboxamide (CB-DMB; Sigma-Aldrich) was dimethyl sulfoxide (DMSO) with pluronic acid (Biotinum) in a ratio of 1:1. The final concentration of DMSO was always below 1%.


The study was approved by the Ethics Committee of the University of Szeged (Szeged, Hungary). All patients gave informed consent.

Patients, isolation, and culture of myofibroblasts.

Three specimens were obtained from patients undergoing gastric tumor resection surgery in the Department of Surgery, University of Szeged, Hungary. Specimens were dissected postoperatively at least 4–10 cm away from the visible tumor border and were transported immediately to the laboratory on ice-cold media. Two other gastric specimens from multiple organ cadaver donors were also obtained following donor surgery and cultured similarly. Additionally, we isolated pancreatic myofibroblasts from a patient with chronic pancreatitis and esophageal, duodenal, and colonic myofibroblasts from two organ cadaver donors. The details of patients can be found in Table 1. The isolation of myofibroblasts was performed using a previously described method (6). Briefly, the specimens were washed and chopped into ∼1- to 2-mm3 pieces and were then bathed in a shaking water bath at 37°C for 15 min with 1 mM dithiothreitol. After being washed, the chopped pieces were incubated for 30 min at 37°C with 1 mM ethylenediamine tetra-acetic acid (EDTA) four times. Samples were cultured for 1–2 wk in Roswell Park Memorial Institute medium supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin, and 2% antibiotic-antimycotic solution. After the cells became confluent, they were trypsinized with 0.25% trypsin-EDTA and were transferred into Dulbecco's modified Eagle's medium (DMEM) complemented with 4 mM l-glutamine containing 10% fetal bovine serum, 1% vol/vol amino acid solution, 1% vol/vol penicillin-streptomycin, and 2% vol/vol antibiotic-antimycotic solution. Medium was replaced every 48 h, and cells were passaged at full confluency up to passage 10. Cultures were continually incubated at 37°C in an incubator gassed with a mixture of 5% CO2 and 95% air. All cultures were handled separately. Experiments were performed on different preparations from different patients, and the samples were never combined. N numbers are given in the text and in the legends as follows: number of independent experiments/number of cells per experiment, unless indicated otherwise.

View this table:
Table 1.

Patients' details

Immunocytochemistry and immunohistochemistry.

HGMs (14,000/chamber) were seeded on chamber slides and allowed to recover overnight. To fix cells, 3.6% paraformaldehyde was used for 30 min and washed three times with phosphate-buffered saline (PBS; Invitrogen). Permeabilization was performed by incubation for 30 min using a filtered, PBS-based solution containing 0.2% Triton X-100 and 0.3% protease-free bovine serum albumin. Cells were then incubated with 10% donkey serum in PBS for 30 min. After being washed three times with PBS, primary antibodies were added to the chambers, and slides were incubated overnight in moist atmosphere at 4°C. The following primary antibodies were used: anti-α-smooth muscle actin (α-SMA) antibody raised in guinea pig (1:100), anti-vimentin antibody raised in mouse (1:400), anti-cytokeratin antibody raised in mouse (Dako), anti-NCX1 and -NCX2 raised in goat (1:50; Santa Cruz Biotechnology, Santa Cruz, CA), and anti-NCX3 and anti-ki67 in mouse. Primary antibodies were removed by sequence washes for 10 min each with 0.14 M NaCl, 0.5 M NaCl, and 0.14 M NaCl dissolved in PBS. HGMs were incubated with secondary antibodies for 60 min in dark and moist conditions. Fluorescein isothiocyanate (FITC)-conjugated anti-guinea pig secondary antibody (1:400), Texas Red-conjugated anti-mouse antibody (1:400), and FITC-conjugated anti-goat and anti-mouse secondary antibodies (1:400) were used as secondary antibodies. After hybridization, slides were washed three times with PBS and covered with Vectashield mounting medium containing 4,6-diaminido-2-phenylindole (DAPI) for nuclear staining and were covered with a cover slip.

Control samples for PDGFRα, DOG1, and CD117 stainings were obtained from gastrointestinal stromal tumor together with peritumoral normal small bowel. Tissues were fixed in 4% buffered formaldehyde and embedded in paraffin, and then 4-μm-thick sections were cut on silanized slides. Deparaffinization and antigen retrieval of sections were made in Target Retrieval Solution pH 6 for PDGFRα and pH 9 for CD117 and DOG1. For nonspecific antigen blocking, 1% BSA-TBS and goat serum were used in a dilution of 1:100. Primary antibodies of CD117, DOG1, and PDGFRα were applied in a dilution of 1:200, 1:100, and 1:80, respectively. The secondary FITC anti-mouse antibody was used in a dilution of 1:400. Nuclear counterstaining was performed by DAPI in a dilution of 1:100. Sections were cover slipped by Fluoromount medium.

Measurement of intracellular Ca2+ and Na+ concentrations.

HGMs (5,000, 30,000, or 100,000 cells) were seeded on a cover slip and were allowed to recover overnight. Next, the cover slip was mounted on an inverted fluorescent light microscope (Olympus, Budapest, Hungary). The cells were bathed in standard HEPES solution at 37°C and were loaded with the Ca2+-sensitive fluorescent dye fura 2-AM (5 μM) for 60 min. After loading, the cells were continuously perfused with solutions at a rate of 6 ml/min. Changes in [Ca2+]i were measured using an imaging system (Cell R; Olympus). HGMs as regions of interests were excited with light at wavelengths of 340 and 380 nm, and the 340-to-380 fluorescence emission ratio (F340/F380) was measured at 510 nm. For intracellular Na+ concentration ([Na+]i) measurements, HGMs (100,000 cells) were handled the same way as for the Ca2+ measurements. The cells were bathed in standard HEPES solution at 37°C and loaded with the Na+-sensitive fluorescent dye SBFI-AM (5 μM) for 50 min. For [Na+]i measurements, the same filters were used as for fura 2-AM, and the F340/F380 ratio was measured at 510 nm. When excited at 340 nm, the Ca2+ indicator fura 2-AM and the Na+ indicator SBFI-AM dyes are sensitive to the [Ca2+]i and [Na+]i, respectively, but insensitive to these ions when excited at 380 nm, thus corresponding to the dye concentration only. Therefore, measuring the F340/F380 ratio is more sensitive than F340 alone, since not only the change in F340 corresponds to the intracellular ion concentrations but the inverse change in F380 as well, since the concentration of the dye changes too. One measurement was obtained per second (40). Compositions of solutions are detailed in Table 2.

View this table:
Table 2.

Composition of solutions


RNA was isolated from HGM cultures using the Qiagen RNeasy Mini Kit (Qiagen House, Crawly, UK) according to the manufacturer's instructions. RNA (3 μg) was reverse transcribed to cDNA. RNA/primer annealing was performed with 0.5 μg oligo(dT) primer at 65°C for 5 min. After cooling, samples were reverse transcribed in a final reaction volume of 30 μl containing the annealed RNA/primer set, 5× avian myeloblastosis virus (AMV) buffer, 1.25 mM deoxyribonucleotide triphosphates (dNTP) mix, 20 units human RNase inhibitor, 15 units AMV-reverse transcriptase, and diethyl dicarbonate (DEPC)-treated water. Reactions were incubated at 42°C for 60 min; enzymes were inactivated at 85°C for 5 min. cDNA (1 μl) was used as template for each PCR in a final volume of 25 μl containing master mix, 10× Taq buffer, 10 nM dNTPs/sample, 2.5 units Taq-polymerase/sample, DEPC-treated water, and 1 μM NCX primer set. The sequences of NCX primers are detailed in Table 3. The expected PCR product sizes were 231, 243, and 154 bp for NCX1, NCX2, and NCX3, respectively. The identities of the PCR products were confirmed by DNA sequencing (performed by Biocenter 2000, Szeged, Hungary). PCR settings were as follows: denaturation at 95°C for 15 s, annealing at 60°C for 15 s, and extension at 72°C for 45 s, 30 cycles. PCR products and DNA HyperLadderPlus (BioLine) were run on a 2% agarose gel containing 0.005% ethidium bromide in Tris-buffered EDTA buffer, and then the gel was illuminated in a UV chamber, and photographs were taken with AlphaImager EC (Alpha Innotech, San Leandro, CA).

View this table:
Table 3.

NCX primer sequences

Voltage-clamp measurements.

Isolated HGMs were cultured as described above. For patch-clamp experiments, the cells were trypsinized, and then the cell suspension was centrifuged at 200 g for 5 min. Thereafter, the cells were suspended in fresh DMEM and put in the setup chamber. HEPES-buffered Tyrode's solution contained (in mM/l) 144 NaCl, 0.4 NaH2PO4, 4.0 KCl, 1.8 CaCl2, 0.53 MgSO4, 5.5 glucose, and 5 HEPES at pH of 7.4. The K+-free solution contained (in mM/l) 135 NaCl, 10 CsCl, 1 CaCl2, 1 MgCl2, 0.2 BaCl2, 0.33 NaH2PO4, 10 tetraethylammonium chloride, 10 HEPES, 10 μM glucose, 20 μM ouabain, 1 μM nisoldipine, and 50 μM lidocaine at pH of 7.4, and the pipette solution for measuring NCX current was 140 CsOH, 75 aspartic acid, 20 TEA-Cl, 5 MgATP, 10 HEPES, 20 NaCl, 20 ethylene glycol tetraacetic acid, and 7 CaCl2 with the pH adjusted to 7.2 with CsOH. One drop of cell suspension was placed within a transparent recording chamber mounted on the stage of an inverted microscope (TMS; Nikon, Tokyo, Japan,) and individual HGMs were allowed to settle and adhere to the chamber bottom for at least 30 min before superfusion was initiated and maintained by gravity. Ventricular myocytes were enzymatically digested from hearts of mongrel dogs as previously described (39). HEPES-buffered Tyrode's solution served as the normal superfusate. Micropipettes were fabricated from borosilicate glass capillaries using a P-97 Flaming/Brown micropipette puller (Sutter, Novato, CA) and had a resistance of 2–5 MΩ when filled with pipette solution. The membrane currents were recorded with an Axopatch-1D amplifier (Axon Instruments, Foster City, CA) using the whole cell configuration of the patch-clamp technique. After establishing high (1–10 GΩ)-resistance seals by gentle suction, the cell membrane beneath the tip of the electrode was disrupted by further suction or by applying 1.5-volt electrical pulses for 1–5 ms. The membrane currents were digitized using a 333-kHz analog-to-digital converter (Digidata 1200; Axon Instruments) under software control (PCLAMP 6.0; Axon Instruments). The results were analyzed using software programs purchased from AXON (PCLAMP 6.0). Experiments were carried out at 37°C.

Migration assay.

HGMs (routinely 120,000 cells) were seeded on six-well plates and were allowed to recover overnight in full media. On the following day, the HGM monolayer was gently scratched by a P2 tip in the middle of the well. Only wells containing even-sided and sharp-edged wounds were used for experiments. After gentle washing for three times with serum-free media, an inverted light microscope was used to measure and photograph wounds. Reagents were then added to the wells in serum-free media, and HGMs were incubated in a CO2 incubator at 37°C for 24 h. Migration was evaluated by counting the cells in the same area of the wound after 24 h as reported earlier (6).

Proliferation assay.

Proliferation assays were performed as previously described (6). HGMs (routinely 50,000 cells) were seeded on cover glasses. After overnight recovery, they were synchronized by 30 h of serum starvation, and then 10 μM Click-iT 5-ethynyl-2-deoxyuridine (EdU) (Alexa Fluor 488 Imaging Kit; Invitrogen) was added to the cells for overnight incubation with or without treatment. After overnight incubation, HGMs were fixed and permeabilized according to the manufacturer's protocol and to detect EdU incorporation. DAPI was used for nuclear staining. The proliferation rate was calculated by normalizing the number of EdU-positive cells to the DAPI-stained cells in at least 10 fields of ×200 magnification per sample.

Statistical analysis.

Values are shown as means ± SE. Statistical analyses were performed using nonparametric Kruskal-Wallis tests with post hoc Wilcoxon tests for pairwise comparisons and Bonferroni correction to test post hoc significance within groups. Simple pairwise comparisons were tested with Student's t-test. P < 0.05 was accepted as significant.


Immunocytochemical identification of cell cultures.

To identify myofibroblasts from different gastrointestinal tissues (pancreas, esophagus, stomach, duodenum, colon), cultured cells were subjected to immunocytochemical analysis using antibodies to vimentin and α-SMA as specific markers of myofibroblasts. Cytokeratin and desmin antibodies were used for detecting epithelial and muscle cells, respectively. α-SMA and vimentin verified the presence of myofibroblasts (Fig. 1), whereas desmin and cytokeratin negativity proved that no epithelial or smooth muscle cells were isolated. Purity of the myofibroblast cell cultures was 100%.

Fig. 1.

Isolated human myofibroblasts from different tissues. Representative immunocytochemistry images of myofibroblasts isolated from the stomach, duodenum, pancreas, esophagus, and colon. α-Smooth muscle actin (α-SMA)-stained sections were also counterstained with antibodies against desmin, vimentin, or cytokeratin. Nuclei were counterstained with 4,6-diaminido-2-phenylindole (DAPI) (blue). Cells showed positive staining for α-SMA (left) and vimentin (right). Cytokeratin and desmin stainings were negative, so the cells were thereby considered myofibroblasts. White scale bar indicates 1 mm. For positive and negative controls, the white scale bar indicates 100 μM. For negative and cytokeratin positive controls, the Capan-1 pancreatic ductal cell line was used. Cultured cancer cells from human high-grade pancreatic neuroendocrine carcinoma were used for positive control to desmin; N = 2 experiments.

Spontaneous Ca2+ oscillation in myofibroblasts.

Interestingly, depending on the source of isolation, about 30–50% of the cells showed spontaneous Ca2+ oscillations (Fig. 2A), but in pancreatic myofibroblasts no oscillation was observed. Figure 2B shows a representative experiment in which oscillations occur in some cells, independently from each other. Note that not all cells show spontaneous oscillatory activity. Oscillations were not synchronized in neighboring cells. We tested different confluency levels (10–50–100%) to see whether oscillation depends on the number of cells seeded on the cover slip but we found no correlation (data not shown). Cells showed stable but greatly variable oscillation patterns from cell to cell with different amplitudes and frequencies. Because the proportion of oscillating cells (53.0 ± 8.1%) was the highest in the gastric myofibroblasts (Fig. 2C) and they showed the highest oscillation frequency (Fig. 2D), we decided to focus our further experiments on HGMs. There were no significant differences between the samples isolated from different patients. Amplitude data on oscillation were calculated by normalizing the amplitude of the wave to the baseline resting Ca2+ signal ratio of F340/F380. HGMs displayed oscillatory activity with amplitudes of 1.67 ± 0.062 (N = 4/10). Average frequency was 0.35 ± 0.05 oscillations/min (N = 4/10).

Fig. 2.

Intracellular Ca2+ oscillations in human myofibroblasts. Human gastric myofibroblasts (HGMs) were seeded on cover glass; after overnight incubation, they were mounted on an inverted microscope and loaded with the Ca2+-sensitive fluorescent dye fura 2-AM to record changes in intracellular Ca2+ concentration. A: representative curves of different Ca2+ signals in standard HEPES solution in HGM. The diagram on the left shows a nonoscillatory cell, whereas the right is a representative curve of an oscillatory cell. Note that spontaneous oscillatory waves are strictly monophasic. B: the color pictures show a representative experiment in standard HEPES buffer-bathed HGMs during fura 2-AM microfluorometry using ×20 magnification. White arrows indicate cells showing no spontaneous Ca2+ oscillations while black arrows indicate cells with spontaneous Ca2+ oscillations. White scale bar indicates 0.5 mm. The 340-to-380 fluorescence emission ratios (F340/F380) are represented as colors as indicated in the color bar (B, bottom). The bars charts show percentage of cells demonstrating oscillatory activity (C) and oscillatory frequencies in myofibroblasts isolated from different gastrointestinal organs (D). While pancreatic myofibroblasts showed no oscillatory activity, esophageal, duodenal, colonic, and gastric myofibroblasts demonstrated oscillation. The highest proportion of oscillatory cells was found in gastric myofibroblasts (N = 6/5).

HGMs are positive for CD117 and PDGF-R, but negative for DOG1.

When focusing on the α-SMA- and vimentin-positive HGMs (Fig. 3A), we performed a ki67 staining to determine if the proportion of oscillatory cells might correspond with the cell cycle, i.e., the S phase. We found that only 34 ± 3.1% of cells stained positive for ki67 (Fig. 3B). The maximum proportion of ki67-positive cells was always below 50%, and in fura 2-AM experiments occasionally 100% of cells showed spontaneous oscillations. It can be concluded that oscillations are unlikely to be due to cell cycle; perhaps they are the result of cell division and migration together, which might explain the heterogeneity of Ca2+ oscillations.

Fig. 3.

HGMs are positive for CD117 and platelet-derived growth factor receptor (PDGF-R) but negative for DOG1. A: representative immunocytochemistry images of HGMs. Cells showed positive staining for vimentin (red) and α-SMA (green). Nuclei were counterstained with DAPI (blue). White scale bar indicates 50 μm; N = 2 experiments. B: further immunotypization of HGMs. DAPI-ki67 costaining (left) revealed that only 34 ± 3.1% of cells were positive for ki67 (N = 6 experiments). PDGF-R and CD-117 (c-kit) expression was also detected in HGMs (center), but the DOG1 (right), specific marker for interstitial cells of Cajal, was negative in all samples. White scale bar indicates 1 mm; N = 3. C: positive control immunohistochemistry for PDGF-R, CD-117, and DOG1 with DAPI costaining. While anti-PDGF-Rα antibody shows characteristic immunostaining in mucosal subepithelial myofibroblasts, CD117 and DOG1 immunoreaction highlights interstitial cells of Cajal in the myenteric plexus. White scale bar indicates 100 μm; N = 1.

Because spontaneous intrinsic electrical activity is well known in the interstitial cells of Cajal (ICC), we checked different ICC markers on HGMs (Fig. 3B). We found that HGMs stained positive for CD117 but negative for DOG1. Because DOG1 is the most specific ICC marker (12) and should correlate with CD117 staining, it can be concluded that CD117 is only a dedifferentiation marker in HGMs (44). PDGF-R, a stromal myofibroblast marker, was also positive for the HGMs.

Intracellular Ca2+ oscillations are dependent on the extracellular Na+ and Ca2+.

We used ion withdrawal techniques on nonoscillating and oscillating HGMs (Fig. 4) to identify the mechanism responsible for the oscillatory activity. Withdrawal of Ca2+ from the extracellular solution caused no effect on nonoscillatory cells, but a cease of Ca2+ oscillation occurred in oscillatory cells (Fig. 4A). Readdition of Ca2+ to the extracellular space made oscillatory cells continue oscillation (Fig. 4A, right). Clearly, oscillation depends on extracellular Ca2+, indicating the role of a plasma membrane Ca2+ channel/transporter. Na+ removal also stopped oscillation and caused a relatively high-amplitude single Ca2+ wave in oscillatory and nonoscillatory cells as well (Fig. 4B), in some cases with repeated waves with fade-out effect (Fig. 4C). These findings suggest a Na+-dependent Ca2+ transport mechanism in HGMs. Removal of both ions had no effect on nonoscillating cells but stopped oscillations in the other type (Fig. 4D). Oscillations returned immediately after readdition of ions, which agrees with the former indication of a plasma membrane Na+-dependent Ca2+-exchange mechanism. Removal of K+ before removing Na+ had no effect on the Ca2+ signaling in nonoscillating cells and did not stop oscillations in oscillatory cells (Fig. 4E), suggesting that the oscillations were K+ independent. Removal of Mg2+ elevated the basal Ca2+ levels in HGMs, but in 16.8 ± 3.6% (N = 6) of the nonoscillating cells it caused Ca2+ oscillations (Fig. 4F). Furthermore, in oscillating HGMs, Mg2+ removal had no effect on Ca2+ oscillations, suggesting that the oscillations are independent of Mg2+ (Fig. 4G).

Fig. 4.

Intracellular Ca2+ oscillations are dependent on the extracellular Ca2+ and Na+. Representative intracellular traces of different Ca2+ signals provoked with ion withdrawal techniques in HGMs during fura 2-AM microfluorometry. Ca2+ (A), Na+ (B and C), Ca2+ and Na+ (D), and potassium then Na+ removal (E) in nonoscillatory (left) and oscillatory (right) cells. Removal of magnesium elevated the intracellular Ca2+ levels of nonoscillatory cells (F, left), but in some cells (16.8 ± 3.6%) it caused Ca2+ oscillations (F, right). Magnesium withdrawal did not affect Ca2+ oscillations (G). Note that spontaneous oscillatory waves were strictly monophasic; N = 8/10.

Intracellular Na+ also depends on extracellular Na+ and Ca2+.

Suspecting the presence of a Na+-dependent Ca2+ exchange mechanism in HGMs, we determined to investigate the effects of ion withdrawal techniques on the [Na+]i levels using SBFI-AM dye. Withdrawal of the extracellular Ca2+ caused an increase in the [Na+]i (Fig. 5A, right), meaning that not only extracellular Na+ regulates the [Ca2+]i, but Na+ and Ca2+ regulate each other's entry to the cells. Removal of extracellular Na+ caused a decrease in [Na+]i (Fig. 5B, left). Removal of both Na+ and Ca2+ also decreased the [Na+]i (Fig. 5B, right), however, the decrease was greater with the presence of extracellular Ca2+ (Fig. 5C). Not only the absolute change of the fluorescence (Fig. 5C, left) but the relative change to the baseline fluorescence (Fig. 5C, right) was also significantly different in the presence of extracellular Ca2+, providing more functional evidence for the presence of a plasma membrane NCX.

Fig. 5.

Intracellular Na+ concentration depends on the extracellular Ca2+ and Na+ concentrations. Representative traces of different intracellular Na+ signals provoked with ion withdrawal technique in HGMs during Na+-binding benzofuran isophthalate-AM microfluorometry (A and B). C: the absolute (left) and relative (normalized to the baseline, right) change in fluorescence intensity after ion withdrawal techniques. Data are shown as means ± SE. *P < 0.05 vs. Na+/Ca2+ free; N = 4/10.

Effect of different Ca2+ transport inhibitors on the intracellular Ca2+ and Na+ levels of HGMs.

Cells were constantly perfused with standard HEPES solution with or without reagents. We tested the L-type Ca2+ channel blocker verapamil and NCX blockers CB-DMB and NiCl2 (Fig. 6). The amiloride derivative pan-NCX inhibitor CB-DMB in 1 μM concentration reversibly lowered [Ca2+]i (Fig. 6A), whereas stopping its administration recovered [Ca2+]i but did not affect oscillation. However, it should be noted that CB-DMB 1 μM concentration inhibited only 31.2% of HGMs, but higher concentrations inhibited all HGMs. The administration of 10 μM CB-DMB resulted in a decrease in [Ca2+]i in all oscillatory and nonoscillatory cells (Fig. 6B). Also, there was no significant difference in [Ca2+]i decrease when nonoscillatory cells were treated with 1 or 10 μM concentrations of CB-DMB (Fig. 6, A vs. B). NiCl2 (1 mM), another potent pan-NCX inhibitor, slightly decreased [Ca2+]i with a small overshoot after stopping its administration (Fig. 6C, left). NiCl2 (1 mM) stopped oscillation in oscillatory cells, but it came back after stopping the administration (Fig. 6C, right). These findings also confirm that spontaneous oscillation is most likely maintained by NCX or at least it is involved in generating/maintaining oscillations.

Fig. 6.

The effect of Na+/Ca2+ exchanger (NCX) inhibitors on the intracellular Ca2+ ([Ca2+]i) and Na+ ([Na+]i) concentration of HGMs. Effects of 1 (A) and 10 (B) μM 3-amino-6-chloro-5-[(4-chloro-benzyl)amino]-N-{[(2,4-dimethylbenzyl)amino]iminomethyl}-pyrazine-carboxamide (CB-DMB), 1 mM NiCl2 (C), and 100 μM verapamil (D) in nonoscillatory (left) and oscillatory (right) HGMs on [Ca2+]i. [Na+]i in SBFI-AM microfluorometric experiments, when inhibiting NCX with 1 and 10 μM CB-DMB (E) and 1 mM NiCl2 (F); N = 4/10.

Verapamil (10 and 100 μM) had no effect on [Ca2+]i (Fig. 6D), suggesting that the oscillation is probably not voltage-mediated. Administering CB-DMB in 1 and 10 μM concentrations resulted in a decreased [Na+]i level, but there was no significant difference between the two concentrations (Fig. 6E). The other NCX inhibitor NiCl2 did not alter significantly the [Na+]i (Fig. 6F). These results suggest the importance of NCX not only in the regulation of [Ca2+]i but also in the maintenance of the [Na+]i in HGM cells.

mRNA and protein expression of NCX in HGMs.

Based on the functional experiments, we investigated the presence of the NCX isoforms at the mRNA and protein levels. RT-PCR confirmed the mRNA expression of all three NCX isoforms (Fig. 7A) on a transplant patient (columns on left) and on a cancer patient (columns on right), which was confirmed by DNA sequencing. Immunocytochemistry revealed that NCX1, NCX2, and NCX3 are also present at the protein level in HGMs (Fig. 7B).

Fig. 7.

NCX mRNA and protein are expressed in HGMs. A: RT-PCR confirmed the mRNA expression of NCX1, NCX2, and NCX3 (n = 2) in HGMs. Left, PCR products from cadaver donors; right, products from gastric cancer patient healthy tissue. Expected PCR product sizes are: NCX1, 231 bp; NCX2, 243 bp; NCX3, 154 bp. In water blind experiments, water was used as template. B: immunocytochemistry revealed that NCX1, NCX2, and NCX3 are present in HGMs. White scale bar indicates 50 μm. No specific staining was detected when the primary antibodies were omitted (data not shown); N = 3.

Measuring NCX current.

After the microspectrofluorometry measurements, the electrogenic features of the cells were examined with whole cell configuration of the patch-clamp technique. For recording the NCX current, special K+-free bath and pipette solutions were used (see above) to block the Na+, Ca2+, and K+ currents and the Na+/K+ pump current. The descending limb of the ramp was used to plot the current-voltage curve (Fig. 8A). The cell capacitance was measured by applying a 10-mV hyperpolarizing pulse from -10 mV. The holding potential was -90 mV. The capacity was calculated by integration of the capacitive transient divided by the amplitude of the voltage step (10 mV).

Fig. 8.

Lack of measurable NCX current (INCX) in HGMs. A: the current-voltage relationship of INCX was measured by ramp pulses. The cells were initially depolarized to a holding potential of −40 to 60 mV and then hyperpolarized to −140 mV and depolarized back to the holding potential. NCX current was recorded as Ni2+-sensitive current using the descending limb of the ramp protocol. B: representative current traces recorded from HGMs (left) and in control dog cardiomyocytes (right) in K+-free bath solution for control (top), after addition of 10 mM NiCl2 (middle). The NCX current is the Ni2+-sensitive current (bottom), i.e., subtracting the trace recorded in the presence of 10 mM NiCl2 from that measured in the absence of NiCl2; N = 5.

The NCX (NiCl2-sensitive) current (Fig. 8B, left) was not measurable in HGMs; however, the same protocol applied to cardiomyocytes resulted in a measurable NCX current (Fig. 8B, right). Probably this is a result due to the low number or density of the NCX expressed on the surface of the HGMs.

Motility is decreased by NCX inhibitors.

In scratch wound migration assays, we used the inhibitors to examine their effects on the motility of HGMs (Fig. 9A). CB-DMB (1 μM) and 1 mM NiCl2 significantly decreased motility (by 28.4 ± 3.9 and 34.3 ± 6.6%) compared with the basal level (Fig. 9B). Stimulating migration with 100 ng/ml insulin-like growth factor II (IGF-II) caused a twofold increase in motility (203.9 ± 9.8% vs. basal). CB-DMB (1 μM) significantly inhibited stimulated migration (by 65.6 ± 8.4%), whereas 1 mM NiCl2 treatment caused the strongest inhibition (98.9 ± 10.9 vs. 203.9 ± 9.8%) (Fig. 9B). We tried to use higher concentrations of CB-DMB to test dose-dependency, but the cells detached above 1 μM after a few hours.

Fig. 9.

Motility is decreased by NCX inhibitors in HGMs. A: raw images taken at 0 and 24 h from scratch wound migration assays. HGMs were treated with different inhibitors with or without 100 ng/ml insulin-like growth factor II (IGF-II) treatment. Black lines mark the original wound edges at 0 h. From these edges HGMs made their way into the cleared area by moving toward the opposite direction. Under basal conditions (without 100 ng/ml IGF-II stimulation), HGMs showed poor motility in serum-free media after 24 h. CB-DMB (1 μM) and 1 mM NiCl2 decreased the number of migrated cells. In response to 100 ng/ml IGF-II stimulation, the basal migration rate elevated, more cells migrated into the scratched wound area, and all the inhibitors seemed to decrease motility. B: the bar chart shows normalized migration rates from the experiment shown in A. The no. of cells migrated into the wound area under basal conditions without any treatment was considered as 100%. Data are shown as means ± SE. *P < 0.05 vs. basal and #P < 0.05 vs. 100 ng/ml IGF-II stimulated; N = 4.

Proliferation is inhibited by blocking NCX.

Next, we tested whether NCX inhibitors have an impact on cell division (Fig. 10). IGF-II (100 ng/ml) was applied to stimulate proliferation of HGMs. We found that CB-DMB (1 μM) inhibited cell division (by 55.8 ± 5.5%), whereas 1 mM NiCl2 caused a 78.3 ± 5.9% decrease in cell proliferation without IGF-II stimulation (Fig. 10B). IGF-II (100 ng/ml) stimulation increased basal proliferation rate to nearly twofold (191.3 ± 10.1%). CB-DMB (1 μM) greatly inhibited cell division, but 1 mM NiCl2 caused the strongest inhibition in the IGF-II-stimulated HGMs.

Fig. 10.

NCX inhibition decreases proliferation rates in HGMs. A: HGMs were seeded on cover glasses and then subjected to 5-ethynyl-2-deoxyuridine (EdU) Click-iT immunocytochemistry. The figure shows representative images from EdU (yellow) immunocytochemistry proliferation assays with DAPI (blue) nuclei staining. HGMs were treated with NCX inhibitors and/or 100 ng/ml IGF-II. The administration of inhibitors decreased the number of EdU-positive cells under basal conditions. IGF-II (100 ng/ml) significantly increased EdU positivity while all inhibitors seemed to greatly inhibit this stimulatory effect. B: the bar chart shows normalized proliferation rates from experiments in A. The number of EdU-positive cells at basal conditions without any treatment was considered as 100%. Data are shown as means ± SE. *P < 0.05 vs. basal and #P < 0.05 vs. 100 ng/ml IGF-II stimulated; N = 3.


The role of the stromal microenvironment is essential in tumor development and angiogenesis. The cells of this microenvironment, like stellate cells in pancreatic carcinomas or myofibroblasts in the gastric carcinomas, contribute to the progression of the diseases by producing inflammatory mediators, cytokines, and collagens (7a, 18, 28). Tumor-derived myofibroblasts have increased migration and proliferation rates compared with adjacent tissue myofibroblasts, and their cellular proteomes and secretomes differ as well (19). Consequently, regulating stromal cells might lead to therapeutic drugs that can modify tumor development and growth. Therefore, in our study, we focused on the Ca2+ homeostasis, migration, and proliferation of HGMs.

In our study, we showed that esophageal, gastric, duodenal, and colonic myofibroblasts display spontaneous Ca2+ oscillations, whereas pancreatic myofibroblasts do not. However, it should be noted that the lack of oscillations in the pancreatic myofibroblasts might not reflect the physiological state in the human pancreas, since these specimens were isolated from a patient suffering from chronic pancreatitis. For that reason, the lack of oscillations might be due to the chronic inflammatory state, the damage from pancreatic enzymes, or the significant amount of epithelial-mesenchymal transitions. However, further studies are needed to determine whether oscillations occur in healthy or tumor-derived pancreatic myofibroblasts.

Among the oscillatory myofibroblasts, the HGMs displayed the highest number of oscillations with the highest frequency. These oscillations depended on the extracellular concentrations of Ca2+ and Na+. Because [Na+]i did not vary in our experiments and the L-type Ca2+ channel blocker verapamil did not affect Ca2+ oscillations, the presence of any voltage-mediated ion influx can be excluded. Oscillations did not depend on cell passage number or confluency on the carrier cover glasses. The function of this phenomenon is unknown; it may be related to muscle cell origin, or it may probably designate a different state in the cell cycle. However, our results suggest that cell cycle could not be the main reason for oscillations. Perhaps they are the result of cell division and migration together, which might explain the heterogeneity of oscillations. Our results suggest that nerve innervations are not required for oscillations or contractions of at least 50% of HGMs. Therefore, these cells should be considered to have nerve-independent spontaneous oscillatory activity. Administration of carbachol (data not shown) caused a single elevation in Ca2+ concentration but did not switch on the oscillatory activity of the cells. However, bioactive molecules (matrix metalloproteinases, TGF-β, insulin-like growth factor binding protein-5, IGF-II progastrin, endothelin-1) were shown to modulate the proliferation and activation of stromal myofibroblasts (10, 15, 28, 30). Therefore, it is more likely that these cells are rather regulated by themselves in an autonomous way and by bioactive molecules by their neighboring cells (such as epithelial cells), thus contributing to inflammation and cancer progression.

In mouse embryonic stem cells, Ca2+ oscillations are mediated by inositol 1,4,5-trisphosphate and store-mediated Ca2+ entry (23). However, in HGMs, Ca2+ clearly enters from the extracellular space, whereas in mouse embryonic stem cells oscillations may occur in the absence of Ca2+ as well, depending on the cell cycle.

With ion-withdrawal techniques, we showed that an extracellular Na+- and Ca2+-dependent transport mechanism plays an important role in regulating the cytosolic Ca2+ and Na+ levels in HGMs, suggesting the presence of NCX. By the K+ withdrawal technique, we showed that the Na+- and Ca2+-dependent transport is independent of K+, thus ruling out the involvement of Na+/Ca2+-K+ exchangers in controlling the Ca2+ homeostasis of HGMs. Our study indicates that extracellular Mg2+ might have a regulatory role in the Ca2+ oscillations, since the removal of Mg2+ caused Ca2+ oscillations in some nonoscillating cells. Probably that is due to the decreased intracellular Mg2+ concentration, which activates the NCX (41). Furthermore, NCX mRNA and protein were expressed in HGMs. The NCX inhibitor CB-DMB (which inhibits the forward and reverse mode of NCX) stopped the oscillations even after washout of the agent, highlighting the noncompetitive irreversible effect of CB-DMB. NiCl2, which is a competitive divalent anion for Ca2+ (20), caused cease of oscillations for the period of administration, but oscillation returned after washout.

In nonexcitable HeLa cells, the mitochondrial NCX controls the pattern of cytosolic Ca2+ oscillations, but in that study cells were pretreated with histamine to display oscillation (16). In the same study in human fibroblasts, the mitochondrial NCX inhibitor CGP-37157 increased the frequency of the baseline oscillations in cells having spontaneous activity and induced the generation of oscillations in cells without spontaneous activity. In HGMs, the possible role for mitochondrial NCX-generated oscillation is unlikely, since CB-DMB does not inhibit mitochondrial NCX (37).

Earlier work from Jacob (21) showed that oscillations mediated by Ca2+-induced Ca2+ release from endoplasmic reticulum (ER) in endothelial cells are synchronized in neighboring cells. However, this was not the case in our experiments. In rabbit urethral ICC, the oscillations were abolished when ryanodine receptors were blocked with tetracaine or ryanodine, also suggesting the role of intracellular Ca2+ stores in originating oscillations (22). However, that may not be the mechanism by which HGMs generate oscillation, since CB-DMB does inhibit nor mitochondrial nor ER Ca2+ channels/transporters in applied concentration (37), and no oscillations could be observed after applying it. In pulmonary artery smooth muscle cells, NCX plays an important role in regulating cytosolic Ca2+ levels; however, in these cells Ca2+ increases via the NCX and store-operated entry (43). Although NCX regulates the cytosolic Ca2+ in the HGMs in a similar way, store-operated entry does not play a role in HGMs, since inhibiting the NCX results in Ca2+ depletion, not store depletion-mediated Ca2+ entry.

Our results indicate that the Ca2+ oscillations may be sustained by plasmalemmal NCX in HGMs. Our results demonstrate that NCX current was not detectable in HGMs, and the administration of 10 mM NiCl2 did not influence the currents recorded in control conditions (in K+-free bath solution in the absence of Ni2+). To validate our experimental method, NCX current was measured in single cells isolated from left ventricles of dog hearts. The NCX current in myofibroblasts may be under the limit of our experimental sensitivity, which could be due to the low number or low density of NCX transporters in the plasma membrane of HGMs.

By controlling the intracellular Na+ and Ca2+ levels in HGMs, NCX may regulate different important physiological processes. Therefore, we investigated the effect of NCX inhibitors on the proliferation and migration rates of HGMs.

Migration was inhibited in HGMs under basal conditions following NCX inhibition by CB-DMB and nickel chloride. It is clear that migrating cells need intracellular free Ca2+ for actin polymerization, which is necessary for the formation of lamellipodia (8), suggesting that decreasing [Ca2+]i might result in decreased motility. Furthermore, it has been shown that NCX1.1 associates with the F-actin cytoskeleton in NCX1.1-expressing Chinese hamster ovarian cells (5).

There are other proteins like calpain and myosins that utilize [Ca2+]i for organizing the construction of focal adhesions also necessary for migration (2). It is believed that Ca2+ entry, mediated by store depletion, through the reverse mode of NCX increases [Ca2+]i, thus stimulating motility in rat tendon fibroblasts (36). In Madin-Darby canine kidney cells, the plasma membrane Ca2+-ATPase and particularly the NCX are necessary for cell migration, and the lack of NCX function almost fully abrogates migration (9).

IGF-II acts through insulin-like growth factor receptor I receptor, increasing motility via various ways, such as Akt, MAP, and PI3, depending on the cell type (31). IGF-II caused a twofold increase in motility of HGMs, which was decreased by both CB-DMB and NiCl2. These data suggest that NCX plays an important role not only in the basal but also in the IGF-II-stimulated migration of HGMs.

Basal proliferation rates were also impaired by either CB-DMB or NiCl2 treatment. That may be a result of the above-mentioned lack of free [Ca2+]i, which is necessary for signaling key events in cytoskeletal organization during S-phase or M-phase (11). Furthermore, without proper Ca2+ signaling, resting cells cannot get over the G0 phase (1). In pulmonary artery smooth muscle cells, Ca2+ enters the cytosol after store depletion via the reverse mode of NCX, and the NCX inhibitor KB-R7943 inhibited cell proliferation via disorganization of intracellular Ca2+ signaling (43).

IGF-II treatment greatly increased cell division in HGMs. However, administering NCX inhibitors abrogated this effect. In addition, NiCl2 decreased both basal and IGF-II-stimulated proliferation to surprisingly low levels, and we believe that, besides blocking NCX, this effect might be due to the inhibition of other enzymes as well (38). Thus we think that these effects of NiCl2 in long-term (24-h) proliferation experiments may contribute to the decreased proliferation rates of HGMs. Therefore, such an enzyme inhibitor effect in migration experiments also cannot be excluded.

In conclusion, we showed for the first time that cultured HGMs display nonsynchronized spontaneous monophasic Ca2+ oscillations, which depend on the extracellular Ca2+ and Na+. In oscillatory and nonoscillatory HGMs, [Ca2+]i and [Na+]i are dependent on extracellular Ca2+ and Na+, suggesting a role for NCX. We showed that NCX1, NCX2, and NCX3 mRNA and protein are present in HGMs and are involved in the migration and proliferation of HGMs. Functional experiments showed that NCX is necessary for proper basal- and IGF-II-stimulated migration and particularly for proliferation of HGMs. Because myofibroblasts are involved in pathophysiological conditions, such as chronic inflammation or tumor development, further investigations are needed to determine if modulating NCX function has a therapeutic effect on hyperproliferative gastric diseases.


This study was supported by the National Development Agency (TÁMOP-4.2.2.A-11/1/KONV-2012-0035, TÁMOP-4.2.2-A-11/1/KONV-2012-0052 TÁMOP-4.2.2.A-11/1/KONV-2012-0073), the North West Cancer Research Funds, the Hungarian Scientific Research Fund (K76844 to J. Lonovics, NF100677 to P.H. and NF105758 to Z. R akonczay, Jr.), the Hungarian Academy of Sciences (BO 00174/10/5 to Z. Rakonczay, Jr.), and the Rosztoczy Foundation (to L. V. Kemény).


The authors hereby declare that there are no conflict of interests to disclose.


Author contributions: L.V.K., A.S., M.C., Z.R.J., E.G., G.L., Z.S., V.V., J.M., L.J., I.B.N., L.V., A.G., and L.T. performed experiments; L.V.K., A.S., M.C., L.J., K.S., J.A., L.V., and D.I.Y. analyzed data; L.V.K., A.S., M.C., Z.R.J., V.V., K.S., J.A., D.I.Y., T.W., and A.V. interpreted results of experiments; L.V.K. and A.S. prepared figures; L.V.K. drafted manuscript; J.L. and P.H. conception and design of research; P.H. edited and revised manuscript; P.H. approved final version of manuscript.


We thank Dr. Michela Ottolia (Cedars-Sinai, Los Angeles) for providing the anti-NCX3 antibody.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 7a.
  9. 8.
  10. 9.
  11. 10.
  12. 11.
  13. 12.
  14. 13.
  15. 14.
  16. 15.
  17. 16.
  18. 17.
  19. 18.
  20. 19.
  21. 20.
  22. 21.
  23. 22.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
  42. 41.
  43. 43.
  44. 44.
View Abstract