Electrophysiological properties of anion exchangers in the luminal membrane of guinea pig pancreatic duct cells

The pancreatic duct epithelium secretes the HCO3−-rich pancreatic juice. The HCO3− transport across the luminal membrane has been proposed to be mediated by SLC26A Cl−–HCO3− exchangers. To examine the electrophysiological properties of Cl−–HCO3− exchangers, we directly measured HCO3− conductance in the luminal membrane of the interlobular pancreatic duct cells from guinea pigs using an inside-out patch-clamp technique. Intracellular HCO3− increased the HCO3− conductance with a half-maximal effective concentration value of approximately 30 mM. The selectivity sequence based on permeability ratios was SCN− (1.4) > Cl− (1.2) = gluconate (1.1) = I− (1.1) = HCO3− (1.0) > methanesulfonate (0.6). The sequence of the relative conductance was HCO3− (1.0) > SCN− (0.7) = I− (0.7) > Cl− (0.5) = gluconate (0.4) > methanesulfonate (0.2). The current dependent on intracellular HCO3− was reduced by replacement of extracellular Cl− with gluconate or by H2DIDS, an inhibitor of Cl−–HCO3− exchangers. RT-PCR analysis revealed that the interlobular and main ducts expressed all SLC26A family members except Slc26a5 and Slc26a8. SLC26A1, SLC26A4, SLC26A6, and SLC26A10 were found to be localized to the luminal membrane of the guinea pig pancreatic duct by immunohistochemistry. These results demonstrate that these SLC26A Cl−–HCO3− exchangers may mediate the electrogenic HCO3− transport through the luminal membrane and may be involved in pancreatic secretion in guinea pig ducts.


Introduction
The pancreas plays a pivotal role in digestion. Pancreatic acini secrete digestive enzyme-rich neutral fluid that is not dependent on the presence of the CO 2 /HCO 3 − -buffer system.
However, ducts secrete a HCO 3 − -rich fluid, which is dependent on the presence of CO 2 /HCO 3 − -buffer, and that neutralizes acid chyme in the duodenum [32]. The generally accepted model for HCO 3 − transport involves Cl − -HCO 3 − exchangers that operate in parallel with cAMP-activated Cl − channels [cystic fibrosis transmembrane conductance regulator (CFTR)] and Ca 2+ -activated Cl − channels, such as TMEM16A/ANO1, on the luminal membranes of duct cells [42,49]. TMEM16A/ANO1 is also found specifically in the apical membranes of the acinar cells and is the critical channel for the control of acinar fluid secretion [33]. In addition, H + -K + pumps and K + channels are expressed on the luminal membrane of pancreatic ducts [11,28,45]. K + channels are important for setting the resting membrane potential and for providing the driving force for anion transport, and may provide the transport partners for H + -K + pumps [10]. Electrophysiological studies have found a luminal Cl − conductance in rat pancreatic ducts [6,27]. Single-channel recordings revealed small-conductance Cl − channels on the luminal membrane of duct cells, which were identified as CFTR Cl − channels [6,7]. The HCO 3 20, 48]. SLC26A6 has been found to be electrogenic with a 1Cl − /2HCO 3 − exchange stoichiometry in Xenopus oocytes and HEK 293 cells [19,38]. Consistently with this, deletion of Slc26a6 altered the overall stoichiometry of apical Cl − -HCO 3 − exchange in native mouse interlobular ducts, suggesting the upregulation of a Cl − -HCO 3 − exchanger with different stoichiometry [41]. Previous studies have demonstrated a functional coupling between CFTR Cl − channels and Cl − -HCO 3 − exchange activity in isolated pancreatic interlobular ducts [15,43]. Furthermore, a computational model suggested that the HCO 3 − /Cl − permeability ratio of apical Cl − channels of 0.4 was able to support HCO 3 − secretion [50]. However, few studies have examined the electrophysiological properties and regulation of HCO 3 − conductance across the luminal membrane of native pancreatic duct cells.
The aim of the present study was to identify HCO 3 − conductance that is important for pancreatic secretion. For this purpose, we directly measured HCO 3 − currents through the luminal membrane of guinea pig pancreatic duct cells using the patch-clamp method in the inside-out configuration. We demonstrated that the inward conductance is dependent on intracellular HCO 3 − and extracellular Cl − , and is blocked by H 2 DIDS, an inhibitor of anion transporters, and thus conclude that such inward conductance is carried out via anion exchangers on the luminal membrane. Furthermore, we report the expression and localization of the SLC26A family in the interlobular and main pancreatic duct using molecular biological and immunohistochemical analyses.

Preparation of pancreatic duct cells from guinea pigs
Female Hartley guinea pigs (240-450 g, n = 35) were sacrificed by carbon dioxide stunning in accordance with protocols approved by the Animal Experimentation Committee, Kansai Medical University. Pancreatic ducts were isolated by enzymatic digestion and microdissection from the pancreas as previously described [12]. S a n t a C r u z B i o t e c h n o l o g y ) a n d 2 -m e t h y l -8-(phenylmethoxy)imidazo[1,2-a]pyridine-3-acetonitrile (Sch28080; Santa Cruz Biotechnology) were dissolved in DMSO at a 1000-fold concentration for application. The current was recorded in the inside-out configuration using the EPC 800 patch-clamp amplifier (HEKA). The amplifier was driven by Clampex 9 (Axon) in order to allow the delivery of a voltage-ramp protocol with concomitant digitization of the current. The membrane potential was generally held at 0 mV, and the command voltage was varied from − 80 to + 80 mV over a duration of 800 ms every 10 s.

RT-PCR analysis
RNA was extracted from the interlobular (outside diameter of 50-150 μm) and main ducts (outside diameter of around 500 μm) from three independent guinea pigs using the RNeasy Plus Micro kit with DNase I (Qiagen). RT-PCR analysis was performed using the OneStep RT-PCR kit (Qiagen) with primers designed to recognize different types of transporters (

Immunolocalization
Immunolocalization was performed on the guinea pig pancreas. The pancreas was obtained from female Hartley guinea pigs (n = 3) in accordance with protocols approved by the Animal Experimentation Committee, Kansai Medical University. The guinea pigs were anesthetized with isoflurane and a mixture of medetomidine (0.5 mg/kg body weight), midazolam (5.0 mg/kg b.w.), and butorphanol (2.5 mg/kg b.w.), and perfused transcardially with 4% paraformaldehyde. The pancreas was fixed with 4% paraformaldehyde in PBS for 24 h, embedded in paraffin, and sectioned. Detailed methods for immunohistochemistry were described previously [12]. Briefly, autofluorescence was blocked by 0.1 M Tris-glycine. Non-specific binding was blocked with 2% normal donkey serum in PBS.

Statistics
Data are shown as means ± SEM. A one-way analysis of variance (ANOVA) or Student's paired t test was applied, and P < 0.05 was considered significant. Data were analyzed in Igor or Microsoft Excel.

Results
Bicarbonate conductance through the luminal membrane of the interlobular pancreatic duct cells We recorded macroscopic currents from excised inside-out patches from the luminal membrane of the interlobular pancreatic duct cells of guinea pigs under unstimulated conditions. Figure 1a shows the macroscopic current-voltage (I-V) relationships in the presence of intracellular HCO 3 − at different concentrations (0, 16, 33, 65, and 130 mM). As we used the standard NMDG-Cl pipette solution and the bathing solution containing KHCO 3 , the inward current was due to HCO 3 − efflux through the luminal membrane. The inward conductance determined from the linear section of the I-V relationships (from − 80 to − 60 mV) increased with intracellular HCO 3 − concentration (Fig. 1a). The linear plot of conductance with the HCO 3 − concentration had a sigmoid relationship (Fig. 1b). The half-maximal effective concentration (K d ) value for the effects of HCO 3 − and Hill coefficient were 31.5 ± 5.1 mM and 3.5 ± 0.4 (n = 5), respectively. We also measured inward HCO 3 − currents in the bathing solution containing 130 mM NaHCO 3 . The inward conductance increased to 2.04 ± 0.95 nS in NaHCO 3 from 0.34 ± 0.08 nS in NaCl (data not shown; n = 13). Thus, there was a minor contribution of K + conductance under unstimulated conditions.   (Fig. 2c; n = 6). Replacement of HCO 3 − with thiocyanate (SCN − ) slightly shifted the reversal potential in a positive direction, but the inward conductance had little change ( Fig. 2d; n = 6). Finally, replacement of HCO 3 − with iodide (I − ) did not cause a marked difference in the reversal potential or the inward conductance (data not shown; n = 6). We calculated the permeability ratio (P X /P HCO3 ) from the shift in the reversal potential (ΔV rev ) when anion X is substituted for internal HCO 3 − [13]; that is, from:

Ion selectivity of the bicarbonate conductance
where R, T, and F have their conventional thermodynamic meanings. The sequence of the permeability ra-

Bicarbonate conductance is dependent on luminal Cl −
To evaluate the activities of Cl − -HCO 3 − exchangers on the apical membrane of interlobular pancreatic ducts of the guinea pig, Ishiguro and colleagues replaced Cl − with gluconate in the lumen [14,16,43]. Similarly, we recorded macroscopic currents with extracellular solution containing 120 mM gluconate and 10 mM Cl − . With the control intracellular solution , E rev was − 46.6 ± 4.5 mV with a standard NMDG-Cl pipette solution ( Fig. 2a; n = 5) and − 63.6 ± 3.9 mV with gluconate-rich pipette solution ( Fig. 3a; n = 5), demonstrating a significant difference (ANOVA). We also compared the inward HCO 3 − conductance with gluconate-rich and standard NMDG-Cl pipette solutions (Fig. 3b). The HCO 3 − conductance was significantly lower with the gluconate-rich pipette solution (0.66 ± 0.20 nS) than with standard NMDG-Cl pipette solutions (1.30 ± 0.09 nS) (n = 5, ANOVA). Additionally, as described in the previous section, the inward conductance significantly decreased when HCO 3 − in the bath was replaced with Cl − , indicating that there was a minor contribution from Cl −dependent current (Fig. 3b, right). However, the inward conductance was not significantly different with gluconate-rich pipette solution (Fig. 3b, left)

Regulation of bicarbonate conductance by intracellular ATP and cAMP
In pancreatic duct cells, cAMP and Ca 2+ signaling pathways play a role in fluid secretion. As CFTR Cl − channels were regulated by intracellular cAMP [6,8,29] and ATP [40], we tested their effects on bicarbonate conductance. Application of intracellular 2 mM ATP-Mg significantly increased the inward conductance from 1.51 ± 0.59 to 5.70 ± 2.18 nS (Fig. 5a, b; n = 13). The addition of 1 mM cAMP further increased the inward conductance to 14.8 ± 5.57 nS (n = 4). cAMP also activated the marked outward conductance, which was attributed to Cl − influx, most likely through CFTR Cl − channels. Therefore, we tested the effects of intracellular ATP-Mg and cAMP with the pipette solution including CFTRinh-172 at 20 μM. In the presence of CFTRinh-172, application of intracellular 2 mM ATP-Mg and 1 mM cAMP had little effect on the conductance in either direction (Fig. 5c): the inward conductance was not significantly increased in the presence of ATP-Mg (1.11 ± 0.39 nS) or cAMP (1.37 ± 0.27 nS) as compared with the control (0.98 ± 0.29 nS) ( Fig. 5d; n = 11

Pancreatic duct epithelia expressed a variety of SLC26A family members
It is known that anion exchangers in pancreatic duct cells are members of the SLC26A family [26]. Two members of the family, SLC26A3 (DRA; downregulated in adenoma) [25] and SLC26A6 (PAT1; putative anion transporter-1) [24,46], were reported to be expressed in the luminal membrane of pancreatic ducts and function as Cl − -HCO 3 − exchangers [9,19,20]. Interlobular ducts from guinea pigs expressed mRNAs encoding Slc26a3 and Slc26a6 [43]. In the present study, we evaluated the expression of all members of the SLC26A family using RT-PCR analysis on isolated interlobular and main ducts. Figure 6a shows the isolated interlobular and main pancreatic ducts expressing CFTR and GAPDH. Then, we screened all 11 members of the SLC26A family from the interlobular ducts ( Fig. 6b; n = 3 animals) and main ducts ( Fig. 6c; n = 3 animals), along with GAPDH and a duct marker of carbonic anhydrase II (CA2). We also screened all primer sets from the total RNA of the kidney as a positive control (Fig. 6d).

Immunolocalization of the SLC26A family in pancreatic duct cells
The immunolocalization of the SLC26A family was examined with paraffin sections of guinea pig pancreas. Immunofluorescence ascribed to the SLC26A exchanger was colocalized with Ezrin, an A-kinase anchoring protein, to the luminal membrane of the pancreatic duct (Fig. 7). In the guinea pig pancreas, immunofluorescence of SLC26A6 was detected on the luminal membranes of duct cells (Fig. 7a), as reported for the rat pancreas previously [20]. SLC26A6 were colocalized with Ezrin to the luminal membranes (Fig. 7b, c). The immunofluorescence on the luminal membranes was diminished with SLC26A6 antibody, which was pre-absorbed with the corresponding antigen for the negative control (Fig. 7d). Additionally, a strong signal ascribed to SLC26A1 was detected and colocalized with Ezrin to the luminal membrane ( Fig. 7e-g). We also detected immunofluorescence of SLC26A4 and SLC26A10 on the luminal membrane of the duct cells (Fig. 7i-k and m-o, respectively). The immunofluorescence was reduced when antibodies were pre-absorbed with the corresponding antigens (Fig. 7h, l, p). We used HPA036055 (Atlas) as the anti-SLC26A3 antibody, but failed to immunostain SLC26A3 in the guinea pig pancreas. We stained with PECAM-1, a blood vessel marker, to distinguish between pancreatic ducts and blood vessels (Fig. 7q, r).

Discussion
In the present study, we applied patch electrodes on the luminal membrane of guinea pig pancreatic duct cells and recorded macroscopic currents in the inside-out configuration. The inward conductance was dependent on the intracellular HCO 3 − concentration (Fig. 1) and was reduced when intracellular HCO 3 − was replaced with Cl − , glc − , or MES − (Fig. 2) or extracellular Cl − was replaced with glc − (Fig. 3). Furthermore, the inward conductance was decreased in the presence of H 2 DIDS, an inhibitor of Cl − -HCO 3 − exchangers (Fig. 4). These electrophysiological findings suggested that Fig. 6 RT-PCR analysis of the SLC26A family. Ethidium bromide-stained agarose gels show RT-PCR products generated from total RNA isolated from the interlobular (I) and main (M) pancreatic ducts. a Control experiment shows the amplification of Cftr (623 bp) and Gapdh (610 bp). No DNA fragment was amplified with the template without reverse transcription (RT). The primers for the RT-PCR analysis from the interlobular (b) and main (c) ducts gave the expected fragment length for Slc26a1-11 ( exchangers but also through CFTR Cl − channels on the luminal membrane [7]. The permeability ratio sequence of the Cl − channel in inside-out patches from the rat pancreatic duct cells was NO 3 − > Cl − > HCO 3 − > gluconate [7], and that in wholecell patches from the guinea pig pancreatic duct cells was Br − > I − = Cl − > HCO 3 − > ClO 4 − > aspartate [29]. These were different from the permeability ratio sequence of the inward conductance obtained in the present study: SCN − > Cl − = gluconate = I − = HCO 3 − > MES − (Fig. 2). Similarly, a previous study reported that the anion selectivity of SLC26A6 in Fig. 7 Immunolocalization of the SLC26A family in the interlobular pancreatic duct.  [30]. Single-channel and whole-cell conductance through Cl − channels was reduced in the presence of HCO 3 − [7,29], whereas the inward conductance was increased with increasing intracellular HCO 3 − in our experiments (Fig. 1). These results suggest that HCO 3 − efflux occurs by pathway independent from the Cl − channels. We followed previous studies that evaluated the activities of Cl − -HCO 3 − exchangers on the apical membrane of pancreatic ducts by replacing extracellular Cl − with gluconate [14,16,43], and observed that the reversal potential shifted to a negative direction and the inward HCO 3 − conductance decreased (Fig. 3). We detected SLC26A1, SLC26A4, SLC26A6, and SLC26A10 on the luminal membrane of the interlobular pancreatic duct (Fig. 7). SLC26A6 was localized to the luminal membrane of interlobular pancreatic ducts of humans [24] and rats [20], as well as to the intestine, kidney, parotid gland, and heart [1,20,21,24,47]. SLC26A6 cloned from guinea pig pancreatic ducts mediated Cl − -HCO 3 − exchange in HEK 293 cells [44]. SLC26A4 (pendrin) was localized to the apical membranes of the submandibular duct, type B and non-A, non-B intercalated cells in the cortical collecting duct of the kidney, and thyroid follicular cells, and was expressed in inner ear [3,36,37,39]. SLC26A4 mediates HCO 3 − secretion across the apical membrane in Calu-3, a human airway epithelia cell line, monolayers [5] and in the cortical collecting ducts [37]. SLC26A1 identified as sulfate/bicarbonate/oxalate exchangers was expressed in the liver and kidney, and to a lesser extent, in the pancreas and testis [2,35], and detected on the basolateral membrane of kidney and liver epithelial cells [18,34]. SLC26A10 was found at the mRNA level in the heart and sarcoma [1,4], but its function is unknown. Although the previous study demonstrated localization of SLC26A3 to the apical membrane of mouse pancreatic duct cells [9], we were unable to immunostain SLC26A3 in guinea pig pancreatic duct cells. The immunostaining signal in the guinea pig pancreas may be underestimated because we were only able to use antibodies against the human SLC26A family. Future studies are needed to establish the functional relevance of SLC26A molecules in pancreatic ducts. We found that intracellular ATP and cAMP activated anion conductance on the luminal membrane in guinea pig pancreatic duct cells (Fig. 5a, b), as observed in rat pancreatic duct cells [6]. It was reported using HEK293 cells that CFTR stimulated by forskolin activated anion exchange of SLC26A3, SLC26A4, and SLC26A6 [19]. Thus, the increased conductance was attributed to activation of CFTR Cl − channels by intracellular ATP and cAMP [6,8,29,40], and activation of Cl − -HCO 3 − exchangers by activated CFTR [19]. As anion conductance was not significantly increased in the presence of CFTRinh-172 in the pipette solution (Fig. 5c, d), we concluded that intracellular ATP and cAMP may not directly regulate Cl − -HCO 3 − exchangers.
We found that H 2 DIDS applied intracellularly inhibited inward HCO 3 − conductance by 50% in excised inside-out patches from the luminal membrane (Fig. 4). A previous study demonstrated that other disulfonic stilbenes, 4,4′d i n i t r o s t i l b e n e -2 , 2 ′ -d i s u l p h o n i c a c i d a n d 4 , 4 ′diisothiocyanostilbene-2,2′-disulfonic acid, blocked CFTR Cl − channels when applied to the cytoplasmic face of membrane patches, with K d values (at 0 mV) of 160 and 80 μM, respectively [23]. It is likely that disulfonic stilbenes are able to act on Cl − -HCO 3 − exchangers from not only the outside but also from the inside of the cell membrane.  Fig. 9 Model of HCO 3 − transport in a pancreatic duct cell. Intracellular HCO 3 − is derived from CO 2 through the action of carbonic anhydrase (CA) and from HCO 3 − uptake via the Na + -HCO 3 − cotransporter. H + is extruded at the basolateral membrane by the Na + -H + exchanger and H + -K + pump. HCO 3 − efflux across the luminal membrane is mediated by Cl − channels (CFTR and TMEM16A/ANO1) and electrogenic Cl − -nHCO 3 − exchangers (SLC26A1, 4, 6, and/or 10; n > 1). K + channels provide an exit pathway for K + and play a vital role in maintaining the membrane potential, which is a crucial component of the driving force for anion secretion. Luminal H + -K + pumps may provide a buffering/protection zone for the alkali-secreting epithelium. Primary active transport is indicated by filled circles In conclusion, we used the patch-clamp technique in the inside-out configuration and demonstrated that the HCO 3 − conductance through the luminal membrane is mediated by Cl − -HCO 3 − exchangers under physiological HCO 3 − concentrations in pancreatic duct cells. Our findings suggest that SLC26A1, SLC26A4, SLC26A6, and SLC26A10 may be involved in the HCO 3 − transport through the luminal membrane. The SLC26A family may also play a role in pH homeostasis in the pancreatic lumen and duct cells. The direct measurement of the HCO 3 − current from the interlobular duct and its functional characterization helps to propose a useful model for HCO 3 − secretion from the pancreatic duct epithelia ( Fig. 9).

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All procedures performed in studies involving animals were in accordance with the ethical standards of the institution or practice at which the studies were conducted.