Effects of acid-base variables and the role of carbonic anhydrase on oxalate secretion by the mouse intestine in vitro

Hyperoxaluria is a major risk factor for calcium oxalate kidney stones and the intestine is recognized as an important extra-renal pathway for eliminating oxalate. The membrane-bound chloride/bicarbonate (Cl−/) exchangers are involved in the transcellular movement of oxalate, but little is understood about how they might be regulated. , CO2, and pH are established modulators of intestinal NaCl cotransport, involving Na+/H+ and Cl−/ exchange, but their influence on oxalate transport is unknown. Measuring 14C-oxalate and 36Cl fluxes across isolated, short-circuited segments of the mouse distal ileum and distal colon we examined the role of these acid-base variables and carbonic anhydrase (CA) in oxalate and Cl− transport. In standard buffer both segments performed net oxalate secretion (and Cl− absorption), but only the colon, and the secretory pathway were responsive to and CO2. Ethoxzolamide abolished net oxalate secretion by the distal colon, and when used in tandem with an impermeant CA inhibitor, signaled an intracellular CA isozyme was required for secretion. There was a clear dependence on as their removal eliminated secretion, while at 42 mmol/L was also decreased and eradicated. Independent of pH, raising Pco2 from 28 to 64 mmHg acutely stimulated net oxalate secretion 41%. In summary, oxalate secretion by the distal colon was dependent on , CA and specifically modulated by CO2, whereas the ileum was remarkably unresponsive. These findings highlight the distinct segmental heterogeneity along the intestine, providing new insights into the oxalate transport mechanism and how it might be regulated.

intestinal oxalate transport and overall homeostasis to be defined.
In the small intestine, PAT1 (Putative Anion Transporter 1; Slc26a6) is a Cl À /HCO À 3 exchanger expressed predominantly in the small intestine, located in the apical membrane of the villus epithelium, with roles in nutrient and nonnutrient-linked Cl À absorption, as well as HCO À 3 secretion and absorption, and intracellular pH regulation (Wang et al. 2002(Wang et al. , 2005Simpson et al. 2007Simpson et al. , 2010Singh et al. 2008Singh et al. , 2010Walker et al. 2011;). Studies of PAT1-KO mice have shown this transporter is also involved in oxalate secretion by the duodenum (Jiang et al. 2006) and ileum (Freel et al. 2006). Another important apical Cl À /HCO À 3 exchanger is DRA (Down-Regulated in Adenoma; Slc26a3), which is prominently involved in Cl À absorption and HCO À 3 secretion along the length of the intestine (Walker et al. 2008;Singh et al. 2010;Freel et al. 2013;Whittamore et al. 2013;Xiao et al. 2014). Most recently, using DRA-KO mice we have shown that DRA also contributes to oxalate absorption by the ileum, cecum, and distal colon . The development of transgenic mice for these transporters has revealed the tremendous value of such animal models for advancing our understanding of intestinal oxalate transport and the roles of individual anion exchangers. Yet, there remains scant information on some of the overall fundamental oxalate-transporting characteristics of the mouse intestine, and furthermore how they might be regulated. Such knowledge will aid a better understanding of the patho-physiology behind disorders of oxalate metabolism, and is vital given the interest in developing the therapeutic potential of the intestine as a means to combat hyperoxaluria (Hatch et al. 2011;Robijn et al. 2011).
One of the most essential functions of the apical Cl À / HCO À 3 exchangers in the intestine occurs when coupled with sodium/proton (Na + /H + ) exchange to perform electroneutral NaCl and fluid absorption. This is a highly regulated process and a broad range of hormonal, immune, and nervous system inputs can modify and coordinate NaCl cotransport and HCO À 3 secretion by the intestine in a segment-specific manner (Kato and Romero 2011). The effects on Na + /H + and Cl À /HCO À 3 exchange activity in the ileum and colon in response to alterations in systemic acid-base homeostasis has previously been established for humans, rats, and rabbits in vivo and in vitro (Charney and Feldman 1984;Charney et al. 1995;Charney and Dagher 1996;Gennari and Weise 2008), and subsequently for mice in vitro (Goldfarb et al. 2000;Charney et al. 2004b). Elegant studies on the acute effects of CO 2 in the rat distal colon demonstrated rapid shifts in the expression of these transporters at the apical membrane signaled through intracellular pH and [HCO À 3 ], mediated by the catalytic enzyme carbonic anhydrase (CA) (Charney et al. 2002b(Charney et al. , 2004a. The components of the bicarbonate buffering system (pH, HCO À 3 and CO 2 ) therefore represent additional, and often overlooked, regulatory factors controlling intestinal electrolyte transport (Goldfarb et al. 1988;Charney and Dagher 1996), but their effects on oxalate handling have not been considered. As oxalate is a substrate of the Cl À /HCO À 3 exchangers, we hypothesized that it too would display characteristics of being modified and regulated by these acid-base variables. The aim of this study was therefore to systematically examine how changes to pH, HCO À 3 and CO 2 impact oxalate transport by the mouse intestine in vitro. Given we know relatively more about intestinal Cl À transport than oxalate, and because Cl À is one of the primary substrates of the anion exchangers, collecting information on the associated Cl À fluxes as part of these investigations offered a useful reference for interpreting the responses of oxalate in the context of what we presently understand about Cl À transport. The following study reports how oxalate and Cl À fluxes across the mouse distal ileum and distal colon were affected by alterations in acid-base variables and the role of CA.

Experimental animals
The following experiments used the wild-type (WT) C57BL/6 mouse strain. All mice were from breeding colonies housed at the AAALAC (Association for Assessment and Accreditation of Laboratory Animal Care)-accredited animal facility within the Biomedical Sciences Building at the University of Florida, where they were given free access to standard mouse chow (diet 2018S; Harlan Teklad) and water. For flux studies, male and female mice aged 2-5 months (20-30 g body mass) were killed by cervical dislocation following prior sedation induced by brief inhalation of isoflurane (≤15 sec). The entire lower portion of the intestinal tract (proximal ileum to distal colon) was then dissected out and placed in ice-cold buffer for immediate preparation for transport experiments. All animal experimentation was approved by the University of Florida Institutional Animal Care and Use Committee (IACUC) and performed in accordance with the National Institutes of Health "Guide for the Care and Use of Laboratory Animals."

Epithelial transport experiments
Unidirectional fluxes of oxalate and Cl À were measured simultaneously under symmetrical, short-circuit conditions across pairs of intact, isolated tissues taken from the distal ileum (4 cm length immediately proximal to the ileo-cecal valve) and distal colon (4 cm length immediately proximal to the peritoneal border and representing the lower 30% of the large intestine). After removing the outer connective tissue each segment was opened longitudinally along the mesenteric border to form a flat sheet. From each segment a pair of tissues were prepared with each one mounted on a slider (P2304; Physiologic Instruments, San Diego, CA), exposing a gross surface area of 0.3 cm 2 , and secured into a modified Ussing chamber (P2300). The mucosal and serosal surfaces were bathed with 4 mL buffered saline and maintained at 37°C while being simultaneously gassed and stirred with the appropriate hydrated gas mixture (Table 1). Each individual preparation was continuously voltage clamped to 0 mV with an automatic voltage clamp (model VCCMC6, Physiologic Instruments, San Diego, CA).
To measure the transepithelial fluxes of oxalate and Cl À , 0.27 lCi of 14 C-oxalate (specific activity 115 mCi/ mmol), and 0.09 lCi of 36 Cl (specific activity 571 lCi/ mmol) were added to either the mucosal or serosal chamber and which was then designated as the "hot side". The addition of 14 C-oxalate to the "hot side" required the respective addition of 0.9 lmol/l "cold" oxalate in the form of Na 2 Ox to achieve a final concentration of 1.5 lmol/l oxalate. This was matched by 1.5 lmol/l Na 2 Ox on the opposing "cold side". At 15 min intervals 1 mL samples were taken from the "cold side" to detect the appearance of these tracers, along with a recording of short-circuit current (lA) and open-circuit potential difference (mV). Each sample taken from the "cold side" was immediately replaced with 1 mL of warmed buffer. At the beginning and end of each experiment a 50 lL sample was taken from the "hot side" and used to calculate the specific activity (dpm/mmol) of each isotope. The activity of 14 C-oxalate and 36 Cl in all samples was determined by liquid scintillation spectrophotometry (Beckman LS6500, Beckman-Coulter, Fullerton, CA) with quench correction following dissolution in 5 mL scintillation cocktail (Ecoscint A, National Diagnostics, Atlanta, GA). Using a series of external standards for each isotope, the validity of counting dual-labeled samples was independently established thus allowing the individual activities of 14 C-oxalate and 36 Cl to be calculated on the basis of their relative counting efficiencies after modifying the detection channels to minimize overlap in their respective energy spectra.
The epithelial responses to the CA inhibitors, ethoxzolamide, and N-3500, as well as to increased CO 2 partial pressure (PCO 2 ) were performed as part of a paired experimental design. This involved commencement of an initial "control" period (0-45 min; Period I), after which the Table 1. The nominal concentration (mmol/L) of salts composing the buffers used in the following study. Each bicarbonate-containing buffer was equilibrated with a 95% O 2 /5% CO 2 gas mixture at 37°C prior to measuring pH, total CO 2 , and osmolality. To elevate PCO 2 in the standard, 21 mmol/L HCO À 3 buffer an 89% O 2 /11% CO 2 mixture was used. designated treatment was applied and the effects recorded for a further 60 min (45-105 min; Period II). In a bid to try and distinguish the involvement of intracellular CA isoforms (CAI and CAII) from the extracellular apical membrane-bound CAIV (Goldfarb et al. 2000), we utilized the impermeant CA inhibitor N-3500 (Delacruz et al. 2010). To target the external CAIV, N-3500 was added to the mucosal bath at the end of the initial control period (Period I). At the conclusion of Period II the membrane-permeant ethoxzolamide was then applied for a final third period (105-165 min; Period III) to inhibit all CA activity. The response to ethoxzolamide was tested in all buffers, whereas the effects of increasing PCO 2 , and the impact of N-3500 (coupled with ethoxzolamide) were examined in standard 21 mmol/L HCO À 3 buffer only. For each buffer shown in Table 1, the data collected during the initial control period of these paired experiments was subsequently used to independently compare the effects of varying [HCO À 3 ] and pH. Table 1 displays the nominal salt composition and characteristics of each buffer. The standard bicarbonate buffer contained 21 mmol/L HCO À 3 and was gassed with 95% O 2 /5% CO 2 to achieve a pH and PCO 2 of 7.4 and 28 mmHg, respectively. To reduce [HCO À 3 ] to 7 mmol/L, two-thirds of the NaHCO 3 was replaced with 14 mmol/L sodium gluconate, and to help limit the increase in osmolality when increasing [HCO À 3 ] to 42 mmol/L, the [NaCl] was reduced by 21 mmol/L. The [HCO À 3 ] and PCO 2 of each bicarbonate buffer was calculated following re-arrangement of the Henderson-Hasselbalch equation using measurements of pH and total CO 2 (tCO 2 ). pH was measured with an accupHast combination microelectrode (Fisher Scientific) connected to a Beckman 690 pH meter (Beckman-Coulter, Fullerton, CA), and tCO 2 by a Corning 965 CO 2 analyzer (Corning Ltd., Halstead, Essex, U.K.). To achieve a HCO À 3 /CO 2 free buffer NaHCO 3 was replaced with equimolar HEPES buffers and gassed with 100% O 2 . To modify pH under these conditions the concentrations of these complementary HEPES buffers were modified accordingly and in some cases necessitated the addition of mannitol to preserve osmolality ( Table 1). As part of this study was examining the effects of CA inhibition, the HCO À 3 /CO 2 -free buffers did not include a CA inhibitor. In all cases, the serosal buffer contained 10 mmol/L glucose, with an equivalent 10 mmol/L mannitol included in each mucosal buffer. To inhibit endogenous prostanoid production all buffers contained 5 lmol/ L indomethacin (Sigma, St. Louis, MO). The CA inhibitor ethoxzolamide was sourced from Sigma, and the 11% CO 2 /89% O 2 gas mixture from Airgas Inc. A concentrated stock solution of ethoxzolamide in DMSO was made fresh on the day of each experiment and added to both mucosal and serosal chambers to a final concentration 100 lmol/L, the resulting amount of DMSO in each half chamber was 0.05%. The impermeant CA inhibitor N-3500 was custom synthesized (Delacruz et al. 2010) and dissolved in standard HCO À 3 buffer prior to addition to the mucosal bath at a final concentration of 100 lmol/L. The isotope 14 C-oxalate was a custom preparation from ViTrax Radiochemicals (Placentia, CA) and 36 Cl was purchased as HCl from Amersham Biosciences (Piscataway, NJ).
The fluxes of oxalate and Cl À in the absorptive, mucosal to serosal (J ion ms ) and secretory, serosal to mucosal (J ion sm ) directions were calculated from the change in activity of 14 C-oxalate or 36 Cl detected on the "cold side" of the chamber at each 15 min sampling point, having corrected for dilution with replacement buffer between samples. These flux rates were expressed per cm 2 of tissue surface area per hour. The recordings of short-circuit current (I sc ; lA) and transepithelial potential difference (mV) were used to calculate transepithelial conductance (G t ; mS/cm 2 ) following Ohm's Law. For each segment net fluxes (J ion net ) were calculated as: J ion net = J ion ms À J ion sm from tissues matched on the basis of conductance (no greater than a AE 15% difference in G t between pairs of tissues from the distal ileum, and not exceeding AE 25% for tissue pairs from the distal colon).
The following data are presented as mean AE SE. For experiments conducted as a paired design, a repeatedmeasures, one-way ANOVA was used to evaluate the epithelial response to treatment with ethoxzolamide, 11% CO 2 , or N-3500 followed by ethoxzolamide, at each subsequent 15 min time point compared to the preceding control period. This control period value was taken as the mean of data points collected between 0 and 45 min.
Significant differences following the experimental treatment were subsequently distinguished by multiple comparisons to the corresponding control value using Holm-Sidak post-hoc tests. Differences in flux rates and electrical characteristics in the presence of buffers with varying [HCO À 3 ] or pH were assessed by one-way ANO-VA followed by Holm-Sidak multiple pairwise comparisons. For data failing to meet the assumptions of approximate normality and equality of variance, the equivalent nonparametric tests were performed. The results of all statistical tests were accepted as significant at P < 0.05. Statistical analysis was performed with SigmaStat v3.5 and the figures drawn using SigmaPlot v11.0 (Systat Software Inc. San Jose, CA).

Effects of varying [HCO À
3 ] Tables 2 and 3 show that in standard buffer (21 mmol/L HCO À 3 ) the distal ileum and distal colon mediated net oxalate secretion. In Table 2, reducing extracellular [HCO À 3 ] to 7 mmol/L or eliminating it altogether, did not significantly impact oxalate fluxes by the distal ileum. The latter condition, did result in net Cl À secretion (À2.49 AE 1.24 lmol/cm 2 /h), through a 30% decrease in J Cl ms . This indicates that even with CA activity intact, endogenous metabolic production of HCO À 3 was insufficient to sustain Cl À absorption and was thus dependent on an extracellular supply of HCO À 3 /CO 2 . In this segment of the lower small intestine, when [HCO À 3 ] was increased to 42 mmol/L both oxalate and Cl À fluxes were significantly affected cutting J Ox sm and J Cl ms by 45 and 24%, respectively, with no accompanying alterations to I sc or G t (Table 2). Conversely, in the distal colon (Table 3), eliminating HCO À 3 /CO 2 exclusively reduced J Ox sm by 28% consequently abolishing net oxalate secretion by this segment. The net secretion of oxalate was also greatly diminished at 42 mmol/L HCO À 3 , again through a decreased J Ox sm flux. Surprisingly, net Cl À absorption was independent of [HCO À 3 ], even in HEPES buffer where J Cl ms was not diminished, although J Cl sm was 25% higher ( Table 3).

Effects of changing pH
The absence of HCO À 3 and CO 2 permitted buffer pH to be manipulated independently of these two variables. Under these circumstances pH did not exert any significant impacts on net oxalate or Cl À transport for either segment examined. For the distal ileum, altering pH between 6.9 and 7.9 did produce a significant increase in I sc from À1.70 to À4.22 leq/cm 2 /h, which was approxi- mately equivalent in magnitude to the resulting net Cl À secretion at pH 7.9 (5.38 AE 1.67 lmol/cm 2 /h). In addition, this higher pH also produced an increase in G t (Table 4). Even though the absence of HCO À 3 and CO 2 abolished net oxalate secretion by the distal colon, changing pH under these circumstances did not produce any other dramatic effects on oxalate fluxes by this segment, although J Cl ms was significantly lower at pH 7.9 this did not translate to a significant change in net Cl À flux ( Table 5).

Effect of increasing PCO 2
There were no significant effects of increasing PCO 2 on oxalate or Cl À fluxes across the distal ileum aside from a reduction in G t (Fig. 1). In contrast, 11% CO 2 produced a clear, rapid increase in the secretory flux of oxalate by the distal colon ( Fig. 2A) which translated to a significant 41% enhancement of net oxalate secretion from À17.50 AE 3.26 to À24.60 AE 3.13 pmol/cm 2 /h. There were, however, no accompanying changes to Cl À fluxes ( Fig. 2B), although I sc gradually became positive (Fig. 2C).

Role of carbonic anhydrase
In Figure 3, application of the CA inhibitor ethoxzolamide did not impact any parameter in the distal ileum, with the exception of G t which showed a very modest reduction (Fig. 3C). For the distal colon, there were dramatic changes to oxalate fluxes in response to CA inhibition, where net secretion was completely abolished (Fig. 4A). This was primarily through a 27% reduction in J Ox sm , with a smaller rise in J Ox ms . There was also an exclusive decrease in J Cl ms , lowering net Cl À absorption by 60% (Fig. 4B), accompanied by a modest, but significant change in direction of I sc (Fig. 4C). Targeting the external CA at the apical membrane of the distal colon with N-3500 did not significantly diminish J Ox sm , only the subsequent addition of ethoxzolamide was able to inhibit net oxalate secretion (Fig. 5A). Similarly, this final maneuver also abolished net Cl À absorption via J Cl ms , where N-3500 had previously no effect (Fig. 5B), and was accompanied by an increasingly positive I sc (Fig. 5C).
When ethoxzolamide was used to inhibit CA in the absence of extracellular HCO À 3 /CO 2 , the distal ileum proved itself refractory to this maneuver also, with no significant effects on oxalate or Cl À transport ( Fig. 6A and  B), although there was a transient increase in I sc (Fig. 6C). The absence of net oxalate secretion by the distal colon under HCO À 3 /CO 2 -free conditions did not change following the application of ethoxzolamide with no subsequent effect on unidirectional fluxes (Fig. 7A).
14.10 AE 0.61 (6) 1.59 AE 1.08 (6) À0.66 AE 0.25 (12) 12  Effects of elevated PCO 2 on oxalate and chloride transport by the distal ileum. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 / h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, shortcircuited segments of distal ileum in standard bicarbonate buffer following an increase in PCO 2 using 11% CO 2 (mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 8 wild-type mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min). Effects of elevated PCO 2 on oxalate and chloride transport by the distal colon. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 / h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, shortcircuited segments of distal colon in standard bicarbonate buffer following an increase in PCO 2 using 11% CO 2 (mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 8 wild-type mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min).  Figure 3. Effects of carbonic anhydrase inhibition on oxalate and chloride transport by the distal ileum. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 /h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, short-circuited segments of distal ileum in standard bicarbonate buffer following application of the carbonic anhydrase inhibitor ethoxzolamide (100 lmol/L, mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 15 wildtype mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min).  Figure 4. Effects of carbonic anhydrase inhibition on oxalate and chloride transport by the distal colon. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 /h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, short-circuited segments of distal colon in standard bicarbonate buffer following application of the carbonic anhydrase inhibitor ethoxzolamide (100 lmol/L, mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 8 wildtype mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min).  Figure 5. Effects of carbonic anhydrase inhibition on oxalate and chloride transport by the distal colon. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 /h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, short-circuited segments of distal colon in standard bicarbonate buffer following application of the membraneimpermeant carbonic anhydrase inhibitor N-3500 (100 lmol/L, mucosal only), followed by ethoxzolamide (100 lmol/L, mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 6 wild-type mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min).  Figure 6. Effects of carbonic anhydrase inhibition on oxalate and chloride fluxes by the distal ileum in HEPES buffer. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 /h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, short-circuited segments of distal ileum in HCO À 3 /CO 2 -free HEPES buffer following application of the carbonic anhydrase inhibitor ethoxzolamide (100 lmol/L, mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 8 wild-type mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min).

Distal ileum
The reduction in J Cl ms seen previously with CA inhibition in standard HCO À 3 buffer was also evident in HEPES buffer and eliminated net Cl À absorption (Fig. 7B).

Discussion
Previous work comprehensively established that states of acute respiratory and metabolic acidosis or alkalosis produce rapid, reversible alterations to NaCl absorption, HCO À 3 secretion and fluid handling by the ileum and colon in vivo. These effects were subsequently shown to be specific responses by the apical Na + /H + and Cl À / HCO À 3 exchangers to extracellular HCO À 3 concentration ([HCO À 3 ] e ), PCO 2 and pH, through corresponding alterations to intracellular pH (pH i ) and HCO À 3 ([HCO À 3 ] i ), mediated by CA. As oxalate is a substrate of the intestinal Cl À /HCO À 3 exchangers, the aim of this study was to systematically examine how these acid-base variables impacted oxalate transport by the mouse intestine in vitro. Under standard buffer conditions we found the distal ileum and distal colon each mediated net oxalate secretion and Cl À absorption, but only the latter segment and notably the J Ox sm pathway, were exclusively responsive to changes in HCO À 3 e , PCO 2 and dependent on CA activity. Oxalate secretion, but not Cl À absorption, by the distal colon was acutely stimulated by increasing PCO 2 . These findings highlight the distinct segmental heterogeneity of oxalate transport in the intestine, provide important new insights into the characteristics of the transport mechanism, and strongly imply that oxalate secretion by the mouse distal colon is specifically regulated by CO 2 .

Effects of acid-base variables on Cl À transport by the mouse intestine
The responses of intestinal Na + and Cl À transport to pH, HCO À 3 and CO 2 , and the role of CA, have been studied extensively in the rat model, and to a lesser degree in the mouse. As this was the first study investigating the relationship between oxalate and these acid-base variables in the mouse, simultaneously measuring Cl À fluxes provided important points of reference to these earlier investigations and assisted subsequent interpretation of the associated oxalate fluxes. Net Cl À absorption by the rat distal ileum in vivo, and specifically J Cl ms in vitro, was indirectly proportional to extracellular pH, whether produced by alterations in PCO 2 , [HCO À 3 ] e or in HCO À 3 /CO 2 -free HE-PES buffer (Charney and Feldman 1984;Kurtin and Charney 1984;Wagner et al. 1986;Vaccarezza and Charney 1988;Charney et al. 1991), and also dependent on CA (Charney et al. , 2002a. Subsequent work on this segment in the mouse showed that increasing PCO 2 from 21 to 70 mmHg stimulated net Cl À absorption in identi-  Figure 7. Effects of carbonic anhydrase inhibition on oxalate and chloride fluxes by the distal colon in HEPES buffer. The unidirectional oxalate fluxes, J Ox (pmol/cm 2 /h), and Cl À fluxes, J Cl (lmol/cm 2 /h), measured across isolated, short-circuited segments of distal colon in HCO À 3 /CO 2 -free HEPES buffer following application of the carbonic anhydrase inhibitor ethoxzolamide (100 lmol/L, mucosal + serosal), are shown in Panels A and B, respectively. Panel C displays the responses of short-circuit current (I sc ) and transepithelial conductance (G t ). Each data point represents mean AE SE of tissue pairs from n = 12 wild-type mice. An asterisk represents a statistically significant change from the preceding control period (0-45 min). cal fashion (Charney et al. 2004b). In contrast, we found ileal Cl À fluxes were unresponsive when subjected to a similar elevation of PCO 2 (Fig. 1B). Reducing pH from 7.61 to 7.09 in HEPES buffer stimulated J Cl ms by the rat ileum 16% (Vaccarezza and Charney 1988), yet we found Cl À fluxes by the mouse ileum were not significantly altered between pH 6.9 and 7.9 (Table 4). The absence of HCO À 3 /CO 2 reversed net Cl À absorption to net secretion exclusively through a reduction in J Cl ms (Table 2), indicating endogenous HCO À 3 production was insufficient to support apical Cl À /HCO À 3 exchange and was thus dependent on HCO À 3 supplied from the serosal bath. This notion was corroborated by the inability of CA inhibitor ethoxzolamide to impact Cl À fluxes (Figs. 3B and 5B), and consistent with the abolition of DIDS-sensitive (mucosal Cl À -dependent) HCO À 3 secretion by the mouse distal ileum following the removal of serosal HCO À 3 /CO 2 (Uchiyama et al. 2007;Zhang et al. 2007). A role for CA in the mouse ileum cannot be dismissed entirely as Uchiyama et al. (2007) observed that Cl À -dependent HCO À 3 secretion was reduced~30% in the presence of 100 lmol/L acetazolamide. In contrast, net Cl À absorption by the rat ileum was sustained in the absence of HCO À 3 /CO 2 (Vaccarezza and Charney 1988), and furthermore was sensitive to methazolamide (Charney et al. 2002a), indicating a prominent contribution of CA to ileal transport.
These acid-base variables also exert their effects on Na + and Cl À absorption by the rat distal colon in vitro in a similar manner (Goldfarb et al. 1988), a segment characterized as sensitive to CO 2 rather than pH (Charney and Dagher 1996;Charney et al. 2004a). Subsequent work on the mouse distal colon in vitro revealed trends largely consistent with the rat, where raising PCO 2 to 69 mmHg exclusively stimulated J Cl ms (Goldfarb et al. 2000;Charney et al. 2004b), while reducing pH from 7.61 to 7.09 (in HEPES buffer) notably enhanced net Cl À absorption through a reduction in J Cl sm (Goldfarb et al. 2000). Unlike these previous studies, we found increasing PCO 2 did not stimulate Cl À fluxes by the distal colon (Fig. 2B), and net Cl À absorption was also independent of pH (Table 4). In the absence of HCO À 3 /CO 2 (pH 7.6), net Cl À absorption in the rat and mouse distal colon was abolished due to a reduction of J Cl ms , relative to standard 21 mmol/L HCO À 3 buffer (Goldfarb et al. 1988(Goldfarb et al. , 2000Charney et al. 2004a), indicating a limited contribution from metabolic HCO À 3 production. Conversely, Table 2 shows J Cl ms and J Cl net , were independent of HCO À 3 /CO 2 , although interestingly the secretory J Cl sm flux was significantly higher in this HEPES buffer (discussed further on). In our study, HCO À 3 i was therefore sufficient to maintain apical Cl À /HCO À 3 exchange (represented by J Cl ms Þ, in the absence of external HCO À 3 /CO 2 , thus implying a substantial contribution from CA. This is supported by the observation that on average, ethoxzolamide reduced J Cl ms by a similar magnitude (~3-4 lmol/cm 2 /h) whether HCO À 3 /CO 2 was present ( Fig. 4B and 5B) or not (Fig. 6B). However, if CA is providing HCO À 3 i from CO 2 to drive apical Cl À /HCO À 3 exchange, then it is curious as to why increasing PCO 2 failed to stimulate J Cl ms (Fig. 2B) as shown reproducibly in previous studies, particularly as [HCO À 3 ] i increases from 11.3 to 18.3 mmol/L when PCO 2 is raised from 21 to 69 mmHg (Dagher et al. 1992).
The reasons for the above discrepancies between our present work on the mouse and previous results are not clear. In relation to studies with rats, species-specific differences in the functional expression of the Cl À transporter(s), particularly the anion exchanger, AE1 (Slc4a1), may be one explanation. Both DRA and AE1 are present in the rat colon , where they are suggested to operate as respective Cl À /OH À and Cl À / HCO À 3 exchangers at the apical membrane, thus endowing AE1 as the principal Cl À transporter Binder 1999, 2000). This is significant considering the CO 2 -stimulated increase in J Cl ms by the rat distal colon correlated with enhanced AE1 expression (Charney et al. 2004a). In contrast to a prominent role for AE1 in the rat large intestine, there is no evidence that this is the case for the mouse. Utilizing immunocytochemistry, Alper et al. (1999) found that antibodies directed against AE1 did not stain enterocytes from the mouse colon, and thus lacks AE1 expression. Although we note that more recent work has detected AE1 mRNA in the mouse large intestine (Gawenis et al. 2010). If AE1 is absent from the mouse intestine this might explain why increasing CO 2 failed to stimulate Cl À absorption in this study. For the mouse (and human) intestine, DRA rather than AE1 has become recognized as the main Cl À /HCO À 3 exchanger (Hoglund et al. 1996;Schweinfest et al. 2006;Kato and Romero 2011;Freel et al. 2013). Qualitatively, the mouse and rat share a near-identical pattern of DRA expression along the large intestine (Talbot and Lytle 2010), yet the relative contributions of DRA and AE1 to Cl À transport for the latter model remain undetermined. In addition to species-specific differences, our data also showed similar departures from other mouse studies (Goldfarb et al. 2000;Charney et al. 2004b). As strain-related differences in intestinal transport do exist between mice this might offer some explanation. For example, duodenal calcium and phosphate absorption by C57BL/6 mice were found to be different from C3H/He mice (Armbrecht et al. 2002), whereas the stimulation of colonic ion secretion in Sv129 mice was distinct from the C57BL strain (Flores et al. 2010). For both of these studies the disparities between strains were, in part, the result of differences in respective transporter expression. A similar scenario, per-haps related to functional AE1 expression, may explain why Cl À absorption by the ileum and colon of the Balb/C mouse used by Charney et al. (2004b) responded to PCO 2 , but the C57BL/6 strain used here did not. We note that AE1 expression was absent from the large intestine of CD1 mice (Alper et al. 1999), whereas AE1 mRNA was detected in the colon of mice on a mixed 129SvEv/Black Swiss background (Gawenis et al. 2010). However, this line of reasoning becomes somewhat uncertain in relation to the findings of Goldfarb et al. (2000), who presented data pooled from C57BL/6J and DBA/2J mice, based on no measurable differences in Na + and Cl À flux rates, CA isozyme expression or CA activity between strains.
Oxalate secretion by the distal colon is stimulated by CO 2 Although elevated PCO 2 unexpectedly failed to stimulate J Cl ms , it dramatically enhanced oxalate secretion ( Fig. 2A). A major portion (70-75%) of CO 2 -stimulated NaCl absorption by the rat distal colon corresponds to the CAdependent trafficking of NHE3 and the anion exchanger AE1 (Slc4a1) to the apical membrane (Charney et al. 2002b(Charney et al. , 2004a. As J Ox sm was also acutely stimulated by CO 2 (Fig. 2A), and dependent on CA (Fig. 4A), independent of any changes in G t (Figs. 2C and 4C), we considered whether alterations to membrane transporter expression might also explain this response. If this increase in J Ox sm was due to changes in an apical Cl À /HCO À 3 exchanger such as AE1 or DRA, then we would anticipate an accompanying increase in J Cl ms , but this was not the case (Fig. 2B). Furthermore, while DRA accounts for 50% of J Cl ms in the mouse distal colon, we have shown that it is involved in transcellular oxalate absorption rather than secretion , and notably PCO 2 was also without effect on J Ox ms ( Fig. 2A). While PAT1 has been identified as the apical anion exchanger responsible for oxalate secretion by the small intestine (Freel et al. 2006;Jiang et al. 2006), the apical transporter(s) involved in the large intestine have not been resolved. Considered to be most prominent in the small intestine, PAT1 expression does extend into the large intestine (Wang et al. 2002;Hatch et al. 2011), yet its function there is uncertain. We have recently shown PAT1 contributes to sulfate (SO 2À 4 ) secretion by the mouse cecum ), but whether this also applies to oxalate and the distal colon has yet to be revealed. Interestingly, PAT1 is considered responsive to systemic acid-base status, as PAT1mediated HCO À 3 secretion by the mouse duodenum in vivo was decreased when the systemic acidosis induced by isoflurane anesthesia was left uncorrected (Singh et al. 2008). Although PAT1 contributes to J Ox sm in the distal ileum (Freel et al. 2006), we have shown here that oxalate secretion by this same segment was unaffected when subjected to acidotic conditions in vitro, that is, where [HCO À 3 ] = 7 mmol/L and pH 6.9 (Table 3), and following an increase in PCO 2 (Fig. 1A). Previous work on the rabbit distal colon showed that J Ox sm and net oxalate secretion could be stimulated by cAMP with characteristics bearing resemblance to electrogenic Cl À secretion (Hatch et al. 1994). Notably, CO 2 can elicit cAMP production by recombinant mammalian transmembrane adenylyl cyclases (tmACs) with an EC 50 of~2 mmol/L (Townsend et al. 2009), which is similar to the [CO 2 ] achieved with 11% CO 2 here (2.1 mmol/L). In contrast, CO 2 -stimulated oxalate secretion by the mouse distal colon appeared to be independent of a cAMP-mediated pathway, as Cl À fluxes (Fig. 4B) and I sc (Fig. 4C) indicated no substantial induction of Cl À or HCO À 3 secretion. Furthermore, the application of 10 lmol/L forskolin (a potent agonist of the tmACs) does not stimulate oxalate secretion by the mouse distal colon (Whittamore, J. M. and Hatch, M., unpublished observations).

Intracellular bicarbonate mediates changes to intestinal ion transport
CO 2 -stimulated Cl À absorption (Charney and Dagher 1996) culminating in the trafficking of AE1 to the apical membrane (Charney et al. 2004a). The underlying signaling mechanism(s) regulating intestinal Cl À absorption and secretion by the rat colon have not been determined, and while we have yet to identify the transporter(s) responsible for oxalate secretion by the mouse large intestine, it is interesting to consider whether this secretion might share the same regulatory pathway(s) involving [HCO À 3 ] i and CA.

Possible transport mechanisms involved in intestinal oxalate transport
While [HCO À induction of a chronic metabolic acidosis significantly reduced urinary oxalate excretion (Bushinsky et al. 2001). However, similar chronic acid loading of normocalcemic rats did not reveal any significant changes to either urinary oxalate excretion or serum oxalate (Green et al. 2005), suggesting a systemic acidosis does not alter renal oxalate handling or oxalate metabolism. The DRA-KO mouse is a model of the disease congenital chloride diarrhea (CCD) and exhibits a chronic metabolic alkalosis with respiratory compensation (Walker et al. 2008;Xiao et al. 2014). In our report on this model , urinary oxalate excretion and serum oxalate were decreased, associated with an induction of net oxalate secretion by the intestine. We emphasize that this change in transport was due to a reduction in J Ox ms from the absence of DRA, rather than enhanced secretion, but we did record a (nonsignificant) 40% increase in J Ox sm by the distal colon (P = 0.10). The intestinal phenotype and urine pH of these mice suggested they harbored the same acid-base disturbance but we did not perform a blood-gas analysis to verify their overall acid-base status. The impact of CCD on the mass-balance of oxalate in humans has not been assessed. A survey of 35 patients diagnosed with, and treated for, this disease found urinary oxalate excretion was within the "normal" range (Wedenoja et al. 2008), however, only very few of these individuals exhibited any systemic acid-base abnormalities (median serum HCO À 3 = 25 mmol/L, blood pH and PCO 2 were not reported). Finally, CA inhibitors, such as acetazolamide, are used clinically for treating various disorders including glaucoma, edema, seizures, and altitude sickness. A potential complication for patients is the development of metabolic acidosis and a propensity for calcium phosphate kidney stone formation, associated with increased urine pH and hypocitraturia (Matlaga et al. 2003;Mirza et al. 2009). The impacts of CA inhibitors on oxalate homeostasis in vivo, however, are limited and inconclusive. A significant rise in urinary oxalate excretion and some mixed calcium phosphate/ oxalate stones have been reported in patients taking acetazolamide (Ahlstrand and Tiselius 1987), whereas other studies with CA-inhibiting drugs have revealed no changes to urinary oxalate handling (Higashihara et al. 1991;Welch et al. 2006;Kaplon et al. 2011).
In summary, oxalate secretion (but not Cl À absorption) by the mouse distal colon was acutely stimulated by increasing PCO 2 in vitro. The secretory pathway, J Ox sm , was found to be exclusively responsive to changes in HCO À 3 , CO 2 , and dependent on CA activity, but not pH. These results strongly suggest oxalate secretion by this segment is specifically regulated by CO 2 . In contrast, net oxalate secretion by the ileum was generally unresponsive to alterations in these same acid-base variables. These find-ings highlight some of the distinct segmental heterogeneity in oxalate transport that exists along the intestine, but also provides important new insights into the characteristics of the underlying transport mechanisms and how they might be regulated, thus helping to direct future work in this area.