INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE KIDNEY RESPONDS TO metabolic acidosis by increasing H⁺ secretion at several sites along the nephron to restore homeostasis. Along the proximal tubule there is increased brush border Na⁺/H⁺ antiporter activity and basolateral Na⁺(HCO₃)²₃ cotransport (1, 24). The cortical collecting duct (CCD) reverses the polarity of its flux from secreting HCO₃ to secreting H⁺ (21, 31), whereas there is increased H⁺ secretion in the outer medullary collecting duct (OMCD) (41) and inner medullary collecting duct (IMCD) (6, 15, 44).

Under baseline conditions, the OMCD has the highest rate of H⁺ secretion (HCO₃ absorption) of the collecting duct segments (28). Net HCO₃ absorption is stimulated by metabolic acidosis in vivo, primarily by increasing the bafilomycin-sensitive H⁺ flux (41). In addition, acute elevation of PCO₂ (22) or 3 h of exposure to a low pH in vitro (41) results in increased net HCO₃ absorption. The mechanisms underlying the adaptive increase in H⁺ secretion have been investigated previously. In acute studies in turtle bladder (14) and rabbit proximal tubules and collecting ducts (29), CO₂ was shown to stimulate exocytic fusion of vesicles containing H⁺ pumps with the luminal membrane. A correlation between H⁺ secretory rate and exocytosis was demonstrated in the turtle bladder (14), and this exocytic process depended upon changes in cell calcium (10, 43). The cytoskeleton has been shown to be important in mediating this process, because disruption of microtubules or microfilaments impairs H⁺ secretion (29, 30, 35).

Recent studies of this adaptation in the OMCD of the inner stripe (OMCD_i) (41) showed that the minimum time of exposure to low pH in vitro that would result in clear cut stimulation of H⁺ secretion was 3 h, suggesting that time was needed to make new H⁺ pumps or move the preformed ones to the luminal membrane. Further, an equivalent stimulation was observed after exposing OMCDs to pH 6.8 for 3 h or to pH 6.8 for 1 h followed by 2 h at pH 7.4 (41). The stimulation of the overall HCO₃ absorptive flux was 32%, corresponding to a 53% increase in the H⁺-ATPase-dependent flux with no change in the flux due to H⁺-K⁺-ATPase. The division of the acid incubation into 1 h at pH 6.8, called the stimulus period, followed by 2 h at pH 7.4, called the recovery period, allowed us to investigate which of the segments of this incubation was crucial in mediating the adaptation. Because comparable changes in H⁺-ATPase-mediated HCO₃ absorption resulted from acidosis in vivo and in vitro, an investigation into the in vitro adaptation would be likely to provide direct insight into how the OMCD responds to this perturbation in vivo.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals. Female New Zealand White rabbits weighing 1.7-2.8 kg and maintained on normal laboratory chow (Purina lab diet 5326; Purina Mills, Richmond, IN) plus free access to tap water were used for these experiments. The animals were weighed and killed by intracardiac injection of 100 mg pentobarbital sodium after premedication with ketamine (44 mg/kg) and xylazine (5 mg/kg) (41).

Microperfusion in vitro. Kidneys were removed, and 1- to 2-mm coronal slices were made and transferred to chilled dissection medium containing (in mM) 145 NaCl, 2.5 K₂HPO₄, 2 CaCl₂, 1.2 MgSO₄, 5.5 D-glucose, 1 trisodium citrate, 4 sodium lactate, and 6 L-alanine; it was gassed with air yielding pH 7.4 and 290 ± 2 mosmol/kg (29, 41). OMCDs were dissected from the corticomedullary rays, and special attention was taken to obtain the ducts from deep within the inner stripe (below the termination of S3 segments and adjacent to the medullary thick ascending limbs of Henle's loop). To maximize the reproducibility of this isolation, relatively short segments (0.5-0.7 mm) were obtained.

In vitro microperfusion was performed according to the method of Burg (8) as previously described (29, 31, 41, 42). An OMCD_i was isolated and transferred to a 1.2-ml temperature-controlled and environmentally controlled chamber mounted on an inverted microscope. It was perfused and bathed at 37°C with Burg's solution containing (in mM) 120 NaCl, 25 NaHCO₃, 2.5 K₂HPO₄, 2 CaCl₂, 1.2 MgSO₄, 5.5 D-glucose, 1 trisodium citrate, 4 sodium lactate, and 6 L-alanine; it was gassed with 94% O₂-6% CO₂ to yield pH 7.4 and 290 ± 2 mosmol/kg (29, 41). The specimen chamber was continuously suffused with 94% O₂-6% CO₂ to maintain bath pH at 7.4 (32). Bathing solution was exchanged by a peristaltic pump at a rate of 14 ml/h to maintain constant solute concentrations.

The collecting end of the segment was sealed into a holding pipette using Sylgard 184 (Dow Corning, Midland, MI). The length of each segment was measured using an eyepiece micrometer. Samples of tubular fluid were collected under water-saturated mineral oil by timed filling of a 14-nl pipette. Collections during each period were made in triplicate.

Bicarbonate transport. The concentration of total CO₂ (assumed to be equal to that of HCO₃) in perfusate (C_O) and collected fluid (C_L) was measured by microcalorimetry (Picapnotherm; Microanalytical Instrumentation, Mountain View, CA). Because there is no net water absorption in the rabbit OMCD (3, 17), the rate of HCO₃ transport (J_HCO3) was calculated as J_HCO3 = (C_O  C_L) × (V_L/L), where V_L is the rate of collection of tubular fluid (1.5-2 nl/min), L is tubular length (in mm), and J_HCO3 is in picomoles per minute per millimeter tubular length. When J_HCO3 > 0, there is net HCO₃ absorption, equivalent to net H⁺ secretion.

The sensitivity of the Picapnotherm was 10-20 counts/pmol total CO₂, so that for samples of 14 nl, there were 140-280 counts/mM total CO₂. The coefficient of variation for a 20 mM standard measured in quadruplicate was <0.5% (<23 counts/sample of 4,500 counts) (42). This level of sensitivity allowed us to reliably detect HCO₃ differences of 1 mM between perfused and collected fluids. In practice, we perfused tubules at 2-3 nl · min¹ · mm¹, which generally resulted in a difference of 4-6 mM between perfused and collected fluids.

Transepithelial voltage. Transepithelial voltage (V_te) was measured using the perfusion pipette as an electrode. The voltage difference between calomel cells connected via 3 M KCl agar bridges to perfusing and bathing solutions was measured with a high-impedance electrometer (World Precision Instruments, Sarasota, FL). Collections of tubular fluid were initiated once the V_te had stabilized (30-45 min), and readings were recorded at the conclusion of each collection. Since the H⁺ pump is electrogenic, V_te tended to increase with increased H⁺ secretion rate and to not change when adaptation was prevented.

Viability. Evidence of damaged cells and gross leak of perfusate was continually assessed by the inclusion of 0.15 mg/ml FD & C green dye to the perfusate during the study (41, 42). The experiment was discarded if tubular damage was detected.

Experimental protocols. J_HCO3 was measured under standard conditions (with Burg's solution in lumen and bath) before and after a 3-h incubation. In a previous study, we showed that 3 h was the minimum time at pH 6.8 required to stimulate net HCO₃ absorption and that an incubation for 1 h at pH 6.8 followed by 2 h at pH 7.4 gave comparable results (41). To investigate the mechanisms involved in this adaptation, we divided the incubation period into two parts: a "stimulus" or early period at pH 6.8, and a "recovery" or late period at pH 7.4, with the rationale that the OMCD cell physiology, metabolism, and protein synthesis would be better maintained after an acid stimulus if the external pH were restored to normal. Such an incubation period was previously shown in the CCD to cause a net reversal of the polarity of net HCO₃ flux from secretion to absorption (40).

The pH 7.4 incubation solution was prepared by mixing three parts DMEM containing 44 mM NaHCO₃ (Life Technologies, Gaithersburg, MD), plus five parts Burg's solution containing 25 mM NaHCO₃ and 1 part HCO₃-free dissection solution, yielding a HCO₃ concentration of 28 mM (40, 41). The acid solution contained three parts DMEM without NaHCO₃ (GIBCO), plus two parts Burg's solution and four parts dissection solution, yielding a HCO₃ concentration of 6 mM (40, 41). Solutions were applied to both apical and basolateral surfaces of the OMCD. The bathing solutions contained 30 U/ml penicillin, 30 µg/ml streptomycin (GIBCO), and 3.3% fetal bovine serum. When the control solution was gassed at 37°C with 94% O₂-6% CO₂, the pH was 7.40 ± 0.02, whereas the acid solution gave a pH of 6.8 ± 0.02.

Agents were added to the incubation in three different formats: 1) for the whole 3 h, 2) for 3-15 min prior to and during the 1 h at pH 6.8, and 3) for the 2-h incubation at pH 7.4. The addition of an agent 3-15 min before the incubation at pH 6.8 allowed us to be certain that the agent had entered the cells in time for the exposure to the acid medium. Concentrations of these agents were chosen to be in the lowest published or tested effective range, to minimize nonspecific effects. All agents were purchased from Sigma Chemical (St. Louis, MO) except for acetoxymethyl ester of 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA-AM), which was from Molecular Probes (Eugene, OR). Anisomycin (10-50 µM) was used to reversibly inhibit protein synthesis, and actinomycin D (4 µM) was used to inhibit DNA transcription. Colchicine (10-100 µM) was used to disrupt microtubules (14, 29, 35), whereas lumicolchicine (10 µM) was used as an inactive analog. Each agent was added to the luminal and bathing solutions. Several of these agents have been previously shown to prevent the CCD from decreasing HCO₃ secretion after in vitro acid incubation (46).

To chelate intracellular calcium activity during the acid incubation, we used the permeant chelator, BAPTA-AM (10-20 µM) (33, 38). We also used an inhibitor of the endoplasmic reticulum (ER) Ca-ATPase, thapsigargin (100 nM) (37), which selectively depletes calcium from this organelle, to investigate the role of ER calcium stores in the adaptation to low-pH incubation. To determine whether the calcium-regulated adaptation to low pH in vitro was mediated by calmodulin, segments were exposed to the specific inhibitor, calmidazolium (R-24571, 30 nM) (9, 27). To determine whether protein kinase C (PKC) mediates the adaptation to low pH in vitro, we exposed OMCD segments to the PKC inhibitor, staurosporine (100 nM) (23), for 10 min prior to and for the whole hour of low-pH incubation.

Inhibitors were not present for the collections of tubular fluid. After the incubation, collections were generally completed within 2 h so that the total time of each experiment was usually 6-7 h.

Analysis and statistics. Data are presented as means ± SE. Paired comparisons (baseline vs. postincubation experimental) for each tubule were analyzed by paired t-test using statistical software (NCSS, Kaysville, UT); significance was asserted when P < 0.05. Multiple comparisons were analyzed using the Bonferroni correction (45).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Protein synthesis inhibition. Anisomycin was applied to the luminal and basolateral solutions 3-15 min before the tubule was exposed to the pH 6.8 incubation. Four of the tubules were treated with 50 µM and three with 10 µM for the 63- to 75-min period, which includes the entire hour at pH 6.8. Table 1 and Fig. 1 show that this early application of anisomycin prevented adaptation of the OMCD, meaning that it prevented the previously observed 30% increase in net HCO₃ absorption and a parallel 0.7-mV increase in V_te (42). There were small significant differences in flow rate and collected HCO₃ concentration, which were not considered important. The experimental net HCO₃ flux was 1 ± 1% higher than baseline [not significant (NS)].

View this table: [in this window] [in a new window]   Table 1.   Data for HCO3 transport before and after 3-h incubation at low pH

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Net bicarbonate absorption (JHCO3, in pmol · min1 · mm1) in individual outer medullary collecting ducts (OMCDs) perfused and bathed in Burg's solution (40 mm PCO2, 25 mM HCO3, pH 7.4): effect of anisomycin in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h, and JHCO3 was measured again (experimental). First, anisomycin was applied 3-15 min before and during the pH 6.8 incubation [early: open circles (50 µM, n = 4) and open squares (10 µM, n = 3) with solid connecting lines]. Second, anisomycin was applied during the pH 7.4 incubation [late: solid circles (50 µM, n = 2) and solid squares (10 µM, n = 3) with dashed connecting lines]. There was no significant change in JHCO3 with early anisomycin application, but there was a significant increase with late application.

On the other hand, late application of anisomycin to five tubules (2 tubules at 10 µM and 3 tubules at 50 µM during pH 7.4 incubation) did not inhibit the adaptation (Fig. 1), which was manifested by a 0.5-mV increase in V_te, an additional decrease of 0.8 mM in collected HCO₃ concentration, and a 4.4-pmol increase in HCO₃ flux (all significant changes, Table 1). The experimental net HCO₃ flux was 32 ± 1% higher than baseline (P < 0.001).

RNA synthesis inhibition. To assess the role of RNA transcription in mediating this adaptation, we added actinomycin D (4 µM) to apical and basolateral solutions at 3-15 min before exposure to pH 6.8 media, which was removed at the end of the low-pH incubation. Actinomycin D completely prevented the adaptation of the OMCD to acid incubation (Fig. 2; Table 1), as there were no significant changes in V_te, collected fluid HCO₃ concentration, or net HCO₃ flux. The experimental net HCO₃ flux was 1 ± 1% higher than baseline (NS).

View larger version (7K): [in this window] [in a new window]   Fig. 2.   JHCO3 in individual OMCDs: effect of actinomycin D in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h and JHCO3 was measured again (experimental). Actinomycin D (4 µM) was applied 2-15 min before and during the pH 6.8 incubation. There was no significant change in JHCO3 with early actinomycin D application.

Microtubule disruption. To document the role of the microtubules in mediating this adaptation, we added colchicine for the entire 3-h period to the apical and basolateral solutions of four tubules (3 tubules at 10 µM and 1 tubule at 100 µM). Colchicine completely prevented the adaptation of the OMCD to acid incubation (Fig. 3; Table 1), as there were no significant changes in any of the measurements. The experimental net HCO₃ flux was 1 ± 1% higher than baseline (NS).

View larger version (9K): [in this window] [in a new window]   Fig. 3.   JHCO3 in individual OMCDs: effect of colchicine in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h and JHCO3 was measured again (experimental). First, colchicine was applied for the entire incubation [open circles (10 µM, n = 3) and open square (100 µM, n = 1) with solid connecting lines]. Second, lumicolchicine, the inactive congener, was applied for the entire incubation at 10 µM (solid circles, dashed connecting lines). There was no significant change with colchicine, but there was a significant increase with lumicolchicine.

In contrast, the inactive congener, lumicolchicine (10 µM), during the 3-h incubation did not prevent the adaptation, as seen by an increase of 0.6 mV in voltage, an additional decrease of 0.9 mM in collected HCO₃ concentration, and an increase of 3.4 pmol net HCO₃ flux (all significant changes; Fig. 3 and Table 1). The experimental net HCO₃ flux was 29 ± 1% higher than baseline (P < 0.01).

Intracellular calcium chelation. To determine the role of intracellular calcium in mediating this adaptation, we added the permeant agent BAPTA-AM (20 µM) to the apical and basolateral solutions of four tubules beginning 15 min before the incubation at pH 6.8. BAPTA-AM is cleaved by cells to the intracellular calcium chelator, BAPTA. BAPTA substantially attenuated the adaptation of the OMCDs, as seen by insignificant changes in net HCO₃ flux, voltage, and collected fluid HCO₃ concentration (Table 1; Fig. 4). The experimental net HCO₃ flux was only 3 ± 1% higher than baseline (P < 0.05). Similar results were obtained when 10 µM BAPTA was added for the entire 3-h incubation (data not shown).

View larger version (10K): [in this window] [in a new window]   Fig. 4.   JHCO3 in individual OMCDs: effect of 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h and JHCO3 was measured again (experimental). First, BAPTA-AM was applied 15 min before and during the pH 6.8 incubation [early: open circles (20 µM, n = 4) with solid connecting lines]. Second, BAPTA-AM was applied during the pH 7.4 incubation [late: solid circles (20 µM, n = 3) with dashed connecting lines]. There was no significant increase in JHCO3 with early BAPTA application, but there was a large significant increase with late BAPTA application.

In contrast, addition of BAPTA-AM (20 µM) to three tubules during the 2-h incubation at pH 7.4 failed to prevent the adaptation, manifested by a significant further decrease of 1.9 mM in collected HCO₃ concentration and a 4.2 pmol increase in net HCO₃ flux (Fig. 4; Table 1). The collected HCO₃ concentration was smaller in part because of the slightly slower perfusion rate after the incubation. The 0.5-mV increase in V_te just failed to reach significance (P = 0.06). The experimental net HCO₃ flux was 32 ± 1% higher than baseline (P < 0.01).

Intracellular calcium transport. To elucidate a role for calmodulin in affecting this adaptation to an acidic environment, we applied the calmodulin kinase inhibitor calmidazolium (30 nM) to the apical and basolateral solutions of four tubules 15 min before and during the incubation at pH 6.8. Adaptation was completely prevented, and there were no significant changes in voltage, collected HCO₃ concentrations, or HCO₃ flux after calmidazolium (Fig. 5; Table 1). The experimental net HCO₃ flux was 2 ± 1% higher than baseline (NS).

View larger version (8K): [in this window] [in a new window]   Fig. 5.   JHCO3 in individual OMCDs: effect of calmidazolium in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h and JHCO3 was measured again (experimental). Calmidazolium (30 nM) was applied 15 min before and during the pH 6.8 incubation. There was no significant change in JHCO3 with early calmidazolium application.

To determine the role of internal calcium stores in the ER, we applied the ER Ca-ATPase inhibitor thapsigargin (100 nM) to the luminal and bathing solutions of four tubules beginning 15 min before the incubation at pH 6.8 and ending when the media were changed to pH 7.4. Adaptation was completely blocked by early application of thapsigargin, as noted by no significant changes in voltage, collected HCO₃ concentration, and net HCO₃ flux (Fig. 6; Table 1). The experimental net HCO₃ flux was 0 ± 1% higher than baseline (NS).

View larger version (9K): [in this window] [in a new window]   Fig. 6.   JHCO3 in individual OMCDs: effect of thapsigargin in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h and JHCO3 was measured again (experimental). First, thapsigargin was applied 15 min before and during the pH 6.8 incubation [early: open circles (100 µM, n = 4) with solid connecting lines]. Second, thapsigargin was applied during the pH 7.4 incubation [late: solid circles (100 µM, n = 3) with dashed connecting lines]. There was no significant change in JHCO3 with early thapsigargin application, but there was a significant increase with late thapsigargin application.

In contrast, when thapsigargin (100 nM) was added late during the pH 7.4 incubation in three tubules, adaptation was not prevented (Fig. 6). There was an increase of 0.5 mV in voltage, a further decrease of 1.7 mM in collected HCO₃ concentration, and a 4.1-pmol increase in net HCO₃ flux (all significant changes, Table 1). The experimental net HCO₃ flux was 35 ± 1% higher than baseline (P < 0.01).

PKC inhibition. To determine a role for PKC in mediating this adaptation, we applied the general inhibitor staurosporine (100 nM) to the apical and basolateral solutions of four tubules beginning 15 min before and continuing through the 1-h pH 6.8 incubation. Adaptation was nearly completely prevented (Fig. 7), as seen by no significant changes in V_te or in collected fluid HCO₃ concentration (Table 1). There was a 0.4-pmol increase in net HCO₃ flux that was significant, but this represented only a 3 ± 1% increase in the experimental over the baseline flux (P < 0.05).

View larger version (7K): [in this window] [in a new window]   Fig. 7.   JHCO3 in individual OMCDs: effect of staurosporine in in vitro acidosis. After baseline values were obtained, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h and JHCO3 was measured again (experimental). Staurosporine (100 nM) was applied 15 min before and during the pH 6.8 incubation. There was a very small but significant increase in JHCO3 with early staurosporine application.

Effect of colchicine on V_te. From Table 1, there can be seen a good correlation between increases in HCO₃ absorption (electrogenic H⁺ secretion) and V_te. Therefore, we used the V_te to indicate changes in luminal H⁺ secretion that would presumably be mediated by exocytosis; such experiments would allow us to determine the critical periods of exocytosis of H⁺ pumps during the 3-h incubation. As seen in Fig. 8 and Table 2, the 3-h incubation (1 h at pH 6.8, 2 h at pH 7.4) in the absence of colchicine resulted in an increase in V_te at each hour of incubation. When colchicine was applied during the 1 h at pH 6.8, there was no increase in V_te, but after its removal there were increases in V_te after each hour of incubation at pH 7.4 (see Fig. 8 and Table 2). When colchicine was applied during the 2-h incubation at pH 7.4, there were no significant increases in V_te during those periods, although there was a significant increase during the 1 h at pH 6.8 in the absence of colchicine. When colchicine was applied during the entire 3-h period, there were no significant increases in V_te after any hour of incubation, consistent with the studies summarized above and in Table 1.

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Transepithelial voltage (Vte, in mV) as a function of time of incubation in colchicine-treated OMCDs. After initial values were obtained at 0 h, each OMCD was incubated at pH 6.8 for 1 h and then at pH 7.4 for 2 h (total incubation 3 h) and Vte was measured each hour. Colchicine (10 µM) was applied during the 1 h at pH 6.8 (solid triangles), during the 2 h at pH 7.4 (open squares), or during the entire 3-h incubation (open circles). Increases in Vte were only seen during the periods without colchicine. Controls, incubated in the complete absence of colchicine (solid circles), showed a consistent increase in Vte. Data represent means of 5 tubules in each group, and SE values (0.2-0.26) are presented only for the control and 3-h colchicine groups. * P < 0.05 compared with previous period. # P < 0.05 compared with initial period.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Metabolic acidosis in vivo results in an increase in urinary acid excretion, and the adaptation occurs in several segments along the nephron. The collecting duct is the final site of regulation. We and others have previously shown in the rabbit CCD that the polarity of net HCO₃ flux converted from secretion to absorption after acid loading in vivo (12, 17, 21, 31). In addition, we have recently shown that metabolic acidosis in vivo results in a 31% increase in net HCO₃ absorption, which corresponded to a 61% increase in bafilomycin-sensitive H⁺ secretion, in the OMCD_i (41).

Using a monoclonal antibody directed against the 31-kDa subunit protein of the vacuolar H⁺-ATPase, Bastani et al. (5) showed striking changes in the immunocytochemical distribution of H⁺-ATPase in the inner stripe of the outer medulla of acidotic rats vs. controls. Control rats showed few rim-labeled cells, with most of the antibody staining in a cytoplasmic vesicular pattern. Most collecting duct cells from acidotic rats showed a loss of vesicular cytoplasmic staining, with all of the label appearing on the apical membrane in a rim pattern within 1 day of acid treatment. This polarization of the H⁺-ATPase, which was not associated with significant increases in H⁺-ATPase mRNA or protein (5), represents an adaptive response that would mediate enhanced apical H⁺ secretion in the OMCD. These findings agree with the morphological findings of Madsen and Tisher (18) in the intercalated cells of the acidotic rat outer medulla, which showed a significant increase in surface density of the apical plasma membrane, along with a decrease in apical tubulovesicular structures. They suggest that shuttling and exocytosis of H⁺ pumps could mediate the adaptation to metabolic acidosis. Indeed, when cultured IMCD cells are acidified, there is a large increase in exocytotic rate, which would be expected to mediate a rise in H⁺ secretion (20, 33). These and other results would support our findings of increased H⁺ secretion in the OMCD_i of acidotic rabbits.

To further examine this tubular adaptation or conditioning, we used the model of metabolic acidosis in vitro, which has been previously shown to induce CCDs to reverse the polarity of net HCO₃ flux from secretion to absorption (26, 46) and, in the perfused OMCD_i, to increase overall net HCO₃ absorption by 32%, corresponding to a 53% increase in bafilomycin-sensitive H⁺ secretion and a ~0.7-mV increase in V_te (41). We have divided the 3-h incubation period into 1 h at pH 6.8 and 2 h at pH 7.4 to try and determine the critically sensitive period for the OMCD to adapt to this perturbation. Previously, we were unable to demonstrate such an adaptation if the total incubation time was reduced to 2 h (41), and a similar observation was observed in the CCD, for which the effect of 2-h acid incubation failed to reduce HCO₃ secretion (46).

Our data show that this adaptation can be prevented if protein or RNA synthesis were to be inhibited by the application of anisomycin or actinomycin D, respectively, during the 1 h at low pH. Interestingly, the late application of anisomycin during the 2 h at pH 7.4 failed to prevent the adaptation, suggesting that the important proteins are made early in the conditioning process. One possibility by which this adaptation might occur is through the synthesis of new H⁺ pumps. However, as noted above, Bastani et al. (5) failed to show increases in H⁺-ATPase (31-kDa subunit) protein or mRNA in rat cortex or medulla during chronic metabolic acidosis. The findings of enhanced apical and reduced cytoplasmic staining of the H⁺ pump in the outer medulla would suggest that exocytosis of H⁺ pumps to the apical membrane is a more likely adaptation. Previous studies have demonstrated in the turtle bladder and rabbit proximal tubule and collecting duct that CO₂ rapidly stimulates exocytotic fusion of vesicles with the luminal membrane (14, 29, 35). These vesicles were shown to contain H⁺ pumps (7), and there was a correlation between H⁺ secretion rate and apical surface area or exocytosis (14, 36). Perhaps, in our studies, there was a synthesis of proteins involved in mediating the process of exocytosis.

There is good evidence from nerve terminal and chromaffin cell studies that vesicle docking, membrane fusion, and exocytosis involve several interactions among trafficking proteins (4, 25, 39). The finding of syntaxin in the apical plasma membrane of rat renal collecting duct cells (19) and its putative receptor synaptobrevin/vesicle-associated membrane protein (VAMP) in aquaporin-2-containing vesicles (16) support the view that these proteins can play a role in targeting of these vesicles to the apical plasma membrane in the rapid mediation of vasopressin-induced water permeability.

More recently, synaptosomal-associated protein (SNAP-25), syntaxin, synaptophysin, and synaptobrevin have been shown to be expressed in cultured IMCD cells and are believed to play a key role in regulated H⁺ secretion (2). Thus the synthesis of one or more of these trafficking proteins may help to mediate exocytotic H⁺ secretion in the OMCD_i in response to metabolic acidosis.

Other investigators have demonstrated upregulation of H⁺-ATPase during metabolic acidosis. Fejes-Toth and Naray-Fejes-Toth (11) showed an upregulation of H⁺-ATPase 31-kDa subunit mRNA expression in H⁺-secreting intercalated cells of acid-loaded compared with alkali-loaded rabbits. Moreover, Garg and Narang (13) showed increased N-ethylmaleimide-sensitive ATPase (presumably H⁺-ATPase) activity during acidosis. Perhaps H⁺ pumps are being synthesized in the OMCD_i under the present experimental conditions. However, it might be expected, if this were true, that the late application of anisomycin would partially prevent the adaptation, but it did not. Therefore, it would seem that much of the protein synthesis mediating or accompanying this regulation is accomplished during the first hour of low-pH incubation. At present, it is not possible to know which proteins are synthesized to mediate the increased rate of H⁺ secretion observed during metabolic acidosis in vivo or in vitro.

We have also observed in the present study that colchicine, which disrupts microtubules in the cytoskeleton and prevents exocytosis (14, 29, 35), when used during the entire 3-h incubation, completely prevented the adaptation, but did not affect the baseline H⁺ transport rate. This finding would support the role of exocytosis and the contributions of the cytoskeleton and neurosecretory-type trafficking proteins toward mediating the increase in H⁺ secretion during metabolic acidosis. The inactive agent, lumicolchicine, failed to prevent the adaptation after the 3-h incubation. Because microtubules are expected to be important in shuttling H⁺ pumps during the entire incubation, we believed that the colchicine should have been applied during the entire 3-h incubation. Indeed, when we examined V_te on an hourly basis during the 3-h incubation, we found that the V_te failed to increase during each hour of colchicine application, indicating that exocytosis of H⁺ pumps is likely to be occurring on a continuous basis in response to metabolic acidosis. Even after the 1 h at pH 6.8, that is, during the pH 7.4 incubation, the V_te continually increased (with the largest increase after 1 h at pH 7.4). These results suggest that the signal to continue the adaptation (exocytosis of H⁺ pumps) was no longer a reduced intracellular pH.

We investigated some of the intracellular signaling that was expected to occur during the adaptation to acidosis. Previous studies have shown that acidification of the cell causes a transient increase in cell calcium in H⁺-secreting epithelia (10, 34, 43). Moreover, buffering or depleting cell calcium changes prevents the acid-induced increase in H⁺ secretion by turtle bladder mitochondria-rich cells (10, 43) and cultured IMCD cells (33). In the present studies, we observed that buffering cell calcium with comparable concentrations of the chelator BAPTA prevented any adaptive increase in H⁺ secretion in response to in vitro acid incubation, but did not affect the baseline HCO₃ transport rate. Furthermore, we suggest that by inhibiting the ER Ca-ATPase, and its potentially regulatable calcium stores (37), with thapsigargin, intracellular calcium is involved in the signaling pathways leading to the adaptive increase in H⁺ secretion. Thus both buffering intracellular calcium with BAPTA and preventing the recruitment of calcium from the ER with thapsigargin have similar effects on the adaptation of the OMCD. Previous studies have shown that thapsigargin does not substantially alter intracellular calcium despite causing the release of calcium from ER stores (37), suggesting that calcium from the ER is in some other way involved in initiating calcium-dependent signaling pathways. Because we incubated the segments for 15 min before starting the 1-h incubation at pH 6.8, we were confident that any transient effects on cell calcium would be eliminated and that the predominant effect on ER calcium stores and transport would be inhibited during the incubation.

Calmodulin also appeared to be involved in the adaptation to acidosis, because an inhibitor, calmidazolium, which antagonizes calmodulin and a secondary calcium/calmodulin-dependent protein kinase (9, 27), also prevented an increase in H⁺ secretion without compromising the baseline transport rate. Since calmodulin modulates the effects of free intracellular calcium, these data suggest that intracellular calcium signaling might be critical for the early adaptation of the OMCD to metabolic acidosis in vitro and perhaps in vivo. The dynamics of calcium interactions involved in this adaptation appear to be far more complicated than originally believed and are worthy of further study.

The cascade of signaling that results in changes in cell calcium is probably preceded by changes in PKC activity. We tested this by using the PKC inhibitor staurosporine applied just before and during the pH 6.8 incubation. Most of the adaptation was prevented by this agent, suggesting that PKC is involved in the adaptive process. We realize that this inhibitor may have other overlapping effects and that additional experiments with more specific inhibitors and perhaps with PKC activators should be anticipated in the future.

To summarize, the OMCD_i adapts to low-pH incubation (metabolic acidosis in vitro) similarly to its adaptation response to in vivo metabolic acidosis (41). By dividing the in vitro incubation into a stimulus at pH 6.8 for 1 h and a recovery at pH 7.4 for 2 h, we have been able to better understand some aspects of this adaptation. Exposure to low pH for 1 h stimulates RNA and protein synthesis and causes increases in intracellular calcium and in PKC activity, probably through the cascade beginning with phospholipase C, which cleaves phospholipids in the cell membrane and generates two intracellular messengers, diacylglycerol and inositol trisphosphate (IP₃). Blockade of the PKC activation or of calmodulin, loss of intracellular stores of calcium, chelation of cell calcium to prevent a transient rise, or disruption of the cytoskeleton all seemed to prevent the adaptation of the OMCD_i to increase H⁺ secretion in response to the in vitro acid incubation, without affecting the baseline transport rate. We can speculate from these data that new H⁺ pumps may be synthesized, or more likely, the preformed H⁺ pumps are shuttled from intracellular vesicles, to be exocytosed onto the apical membrane to increase H⁺ secretion, an adaptive response appropriate to metabolic acidosis. The possibility that the PKC/calcium signaling system is involved to initiate the processes of vesicle docking, membrane fusion, and exocytosis in the adaptive response to metabolic acidosis will require further investigation.