The molecular basis for the renal compensation to respiratory acidosis and specifically the role of pendrin in this condition are unclear. Therefore, we studied the adaptation of the proximal tubule and the collecting duct to respiratory acidosis. Male Wistar-Hannover rats were exposed to either hypercapnia and hypoxia [8% CO2 and 13% O2 (hypercapnic, n = 6) or normal air (controls, n = 6)] in an environmental chamber for 10 days and were killed under the same atmosphere. In hypercapnic rats, arterial pH was lower than controls (7.31 ± 0.01 vs. 7.39 ± 0.01, P = 0.03), blood HCO3− concentration was increased (42 ± 0.9 vs. 32 ± 0.24 mM, P < 0.001), arterial Pco2 was increased (10.76 ± 0.4 vs. 7.20 ± 0.4 kPa, P < 0.001), and plasma chloride concentration was decreased (92.2 ± 0.7 vs. 97.2 ± 0.5 mM, P < 0.001). Plasma aldosterone levels were unchanged. In the proximal tubule, immunoblotting showed an increased expression of sodium/bicarbonate exchanger protein (188 ± 22 vs. 100 ± 11%, P = 0.005), confirmed by immunohistochemistry. Total Na/H exchanger protein expression in the cortex was unchanged by immunoblotting (119 ± 10 vs. 100 ± 11%, P = 0.27) and immunohistochemistry. In the cortex, the abundance of pendrin was decreased (51 ± 9 vs. 100 ± 7%, P = 0.003) by immunoblotting. Immunohistochemistry revealed that this decrease was clear in both cortical collecting ducts (CCDs) and connecting tubules (CNTs). This demonstrates that pendrin expression can be regulated in acidotic animals with no changes in aldosterone levels and no external chloride load. This reduction of pendrin expression may help in redirecting the CNT and CCD toward chloride excretion and bicarbonate reabsorption, contributing to the increased plasma bicarbonate and decreased plasma chloride of chronic respiratory acidosis.
- cortical collecting duct
- chloride regulation
- intercalated cells
hypercapnia and consecutive respiratory acidosis with concomitant hypoxemia are well-known pathological conditions observed in different disease states. Hypercapnia is also encountered in the care of acute respiratory distress syndrome and is therefore not rare in a clinical setting (20). Hypercapnia induces acidosis of the respiratory type, due to a rise in blood Pco2 which will displace the CO2⇋HCO3−+H+ equilibrium, leading to a decrease in blood pH. In the early phase of hypercapnia (the initial rise in Pco2), extrarenal factors are very important to buffer the pH (14). This is in contrast to chronic elevation of Pco2, where the kidney is the major player in stabilizing pH (37). However, the molecular mechanisms by which the kidney compensates a chronic respiratory acidosis are incompletely understood.
In the proximal tubule, 70–80% of the filtered bicarbonate is reabsorbed, mainly via the apical Na/H exchanger (NHE3) and, to a lesser extent, via an apical vacuolar proton pump (2, 3, 15, 28). Most studies suggest that adaptation of proximal tubular transport is important for renal compensation during prolonged respiratory acidosis (5, 32, 39). This occurs via activation of the sodium/proton exchanger NHE3 and the sodium/bicarbonate exchanger NBCe1 (32, 33, 39). However, the importance of the increased proximal tubular bicarbonate reabsorption has been questioned (34, 48).
In addition to the proximal tubule, the collecting duct plays a major role in the final acidification of the urine during acid-base challenges by modulations of intercalated cell transport mechanisms (27, 40, 41). The importance of these cells is clear in metabolic acidosis. In this situation, an increase in the activity of the proton pump occurs in conjunction with a decrease in abundance of the anion exchanger pendrin, resulting in a decreased excretion of bicarbonate by the cortical collecting duct and in a further acidification of the urine (13, 45).
Pendrin is a bicarbonate/chloride exchanger located in the apical domain of the type B and non-A non-B intercalated cells (19). Traditionally, intercalated cells are considered to be primarily acid-base regulators, and therefore pendrin was initially thought to be involved in the maintenance of the acid-base status since its expression is regulated in metabolic acidosis and alkalosis (13, 45). However, the pendrin null mice presented with no acid-base disturbance during basal conditions (31), and evidence has been presented that pendrin function is involved in processes which may mediate mineralocorticoid-induced hypertension (43). Furthermore, as many of the animal models involving metabolic acidosis or alkalosis imply major changes in chloride intake, which by themselves appear to influence pendrin regulation (30, 42, 46), the exact role of pendrin in acid-base physiology is still to be determined. The influence of respiratory acidosis on pendrin regulation is unknown and is interesting, as this condition represents a pure primary acid-base disturbance with no load of electrolytes superimposed. Respiratory acidosis is usually observed with concomitant hypoxemia. Accordingly, we designed the experimental conditions to mimic this situation and studied the renal adaptation to respiratory acidosis.
Experiments were performed using male Wistar-Hannover rats (160–180 g, Dept. of Physiology, Univ. of Aarhus, Aarhus, Denmark). The animal experiments were performed following the Danish legislation on animal experiments and according to license 2005/516–1032 for use of experimental animals issued by the Danish Ministry of Justice.
Chronic normobaric hypercapnia and hypoxia.
Rats were placed in normal cages inside an airtight Plexiglas chamber. The chambers were closed, and the internal Po2, Pco2, temperature, and relative humidity (RH) were controlled. Air conditioning devices were attached to each box and the RH was maintained at 45% in each cage and the temperature between 20 and 22°C. Rats were allowed free access to pellet food (Altromin 1320) and tap water during the entire experiment. Pair feeding was not performed due to the obligation of keeping the chamber closed during the experiment. Food intake, however, was recorded at the end of the experiment and was altogether lower in the hypercapnic group (478 g in total for hypercapnic rats vs. 614 g for controls during the 10 days for all rats).
In the hypercapnic group, the chamber was kept closed, allowing for a progressive rise in Pco2 and decline in Po2 due to the rats' metabolism. Pco2 and Po2 in the chamber were monitored and recorded continuously on a computer. The sensors for O2 (Servomex 570A) and for CO2 (Qubit S 153) were calibrated on a daily basis. When the desired level of CO2 was reached, a Labview computer program turned on an air pump (Reciprotor 506 R), which drove air out of the chambers. By pressure equilibrium, the equivalent amount of fresh air would then enter the box through a 2-m-long (to impair back-diffusion) gastight tube. The pump was set to be turned off when a value <8.3% CO2 was recorded and was automatically turned on when the CO2 reached 8.3% inside the Plexiglas box. Due to dead spaces and various other delays, this setting allowed maintenance of the CO2 percentage within 8.3 ± 0.5%. Due to the metabolism of the rats, the oxygen concentration decreased in the chambers to ∼13%, mimicking respiratory failure. Control rats were placed in the same type of airtight Plexiglas chamber controlled in the same way, but the limit for CO2 was set at 0.2%, resulting in an almost normal atmosphere in the chamber. Dehumidification and air mixing were established in parallel in the control rats. After 10 days of exposure to this environment, the rats were killed.
Anesthesia and sampling of blood and tissue.
To minimize the changes in composition of inspired air during the time of death, air exiting the hypercapnic chamber was collected in airtight “Douglas” bags during the last day. When the hypercapnic rats were euthanized, this air mixed with isoflurane was used in the evaporative anesthesia system (Forane, Abbott Laboratories, Gentofte, Denmark). For control rats, normal room air mixed with isoflurane was used. After anesthesia, a large laparotomy was performed and the abdominal aorta was cannulated. Blood was collected by a three-way valve system. Subsequently, the right renal artery was clamped and the right kidney was removed, dissected into zones [cortex and outer stripe of outer medulla (cortex/OSOM), inner stripe of outer medulla (ISOM) and inner medulla (IM)], and processed for immunoblotting as described below. The left kidney was fixed by retrograde perfusion as described below.
Plasma and blood values.
One aliquot of the arterial blood sample was analyzed immediately for pH, Pco2, Po2, Cl−, Ca2+, and lactate in an ABL system 615 (Radiometer, Copenhagen, Denmark). Blood bicarbonate and base excess were calculated with the ABL 615 software from the measured parameters. The remaining blood was centrifuged for 15 min at 4,000 g to remove blood cells, and the plasma was analyzed for sodium, potassium, and creatinine using a Vitros 950 (Johnson&Johnson). Osmolality was measured with an automatic cryoscopic osmometer (Osmomat 030, Gonotech). Plasma aldosterone concentrations were determined using a commercially available radioimmunoassay kit (Coat-A-Count, Diagnostic Products, Los Angeles, CA).
The following previously characterized antibodies were used: rabbit polyclonal anti-NBCE1 (1159AP) against rat NBCE1 (26); rabbit polyclonal anti-NHE3 L546 against rat NHE3 (11, 16) (kind gift from Dr. M. Knepper, LKEM, National Institutes of Health, Bethesda, MD) for Western blotting and immunohistochemistry; rabbit polyclonal anti-NHE3C00 raised against a peptide similar to the peptide used to generate L546 for immunohistochemistry (kind gift from Alicia McDonough, Univ. of Southern California Keck School of Medicine, LA, CA) (47); rabbit polyclonal anti-V1-ATPase B1 subunit (H7659) against rat V1-ATPase B1 subunit (4); rabbit polyclonal anti-pendrin (RA3466/2671) against mouse pendrin, previously characterized in rats (19); and rabbit polyclonal anti-α-ENaC subunit (6) (kind gift from Dr. B. C. Rossier, Dept. of Pharmacology, Univ. of Lausanne, Lausanne, Switzerland). Rabbit polyclonal anti-β-ENaC (5161) and anti-γ-ENaC (5163) against rat β- and γ-ENaC were developed against peptides with similar amino acid sequences as previously used (24). Immunoblotting and immunohistochemistry revealed identical patterns compared with previously developed antibodies to similar peptides. Peptide preabsorption on immunoblotting and immunohistochemistry revealed specific abatable bands and labeling (data not shown). Moreover, immunohistochemistry and immunoblotting using the antibodies showed an upregulation of the proteins in samples from DDAVP-treated Brattleboro rats as demonstrated previously (7) (data not shown). Polyclonal rabbit anti-aquaporin-2 (AQP2 H7661) against rat AQP2 was developed against the same sequence as previously used (29). Immunoblotting and immunohistochemistry using this antibody revealed identical patterns compared with previously developed antibodies to similar peptides. Peptide preabsorption on immunoblotting and immunohistochemistry revealed specific abatable bands and labeling (data not shown).
A commercially available mouse monoclonal antibody against rat calbindin D-28K (Research Diagnostics, Flanders, NJ) was used.
The dissected renal cortex/OSOM, ISOM, and inner medulla were homogenized (Ultra-Turrax T8 homogenizer, IKA Labortechnik, Staufen, Germany) in ice-cold isolation solution containing 0.3 M sucrose, 25 mM imidazole, 1 mM EDTA, 8.5 μM leupeptin, and 1 mM Pefabloc at pH 7.2. The homogenates were centrifuged at 4,000 g for 15 min at 4°C to remove whole cells and larger organelles, and the supernatant was collected and kept on ice. All samples were heated to 65°C for 15 min after addition of SDS containing sample buffer and then stored at −20°C. Immunoblotting was performed as previously described (13). Results are presented as densities relative to the control group.
Fixation of rat kidneys and immunohistochemistry.
A needle connected by tubing to PBS and fixative reservoirs pressurized to 160 mmHg with air was inserted in the abdominal aorta of isoflurane-anesthetized rats, and the inferior caval vein was cut to establish an outlet. After a 15-s perfusion with cold PBS (pH 7.4), the kidney was perfused with cold 3% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.4) for 3 min. The kidney was removed and sectioned into 2- to 3-mm transverse sections and immersion fixed for additionally 1 h, followed by 3 × 10-min washes with 0.1 M cacodylate buffer, pH 7.4. Immunohistochemistry was performed as previously described (13).
Counting of pendrin-positive cells relative to total number of nuclei in CCDs and CNTs was performed, as described in detail (13), with the modification that nuclei were labeled using a nuclear counterstain, TO-PRO-3 iodide (642/661, Invitrogen, Molecular Probes). The fraction of pendrin-positive cells in hypercapnic rats was normalized to the fraction of pendrin-positive cells in control rats.
For immunoelectron microscopy, small pieces of kidney cortex were cut from slices of fixed kidney (see above), cryoprotected in 2.3 M sucrose, and frozen in liquid nitrogen and prepared as described previously (13).
Evaluation of pendrin subcellular localization by immunoelectron microscopy.
We evaluated the subcellular localization of pendrin in pendrin-positive cells by electron microscopy. Briefly, we identified a minimum of six pendrin-positive cells in the CNT and in the CCD of three animals in each group (total number of cell counted: 106 pendrin-positive cells). The cells were picked randomly as they were encountered during electron microscopy. Cells were selected for quantification if they were well preserved, showing a well-defined nucleus and organelles. Digital pictures were acquired at ×22,000 magnification on a FEI Morgagni electron microscope operating at 80 kV using the Multiple Image Alignment (MIA) function of AnalySIS software. Each cell was imaged by merging 25 single images into one MIA-image. In AnalySIS software, the “touch-count” function was used, and each gold particle was manually identified and classified as either being localized to the apical plasma membrane (within 30 nm) or in the cytoplasm of the pendrin-positive cell. To quantitatively assess the subcellular localization of pendrin, we calculated the ratio of the plasma membrane-associated gold particles to the total number of gold particles counted in the same cell. For each animal, the average of this ratio was calculated separately for cells from CNT and CCD, and the mean in each group (n = 3) was then computed and analyzed statistically
Values are presented as means ± SE. Comparisons between two groups were made by an unpaired t-test. P values <0.05 were considered significant.
Plasma pH, Cl−, and Po2 were reduced, whereas plasma Pco2 was increased in chronic hypoxic hypercapnia.
At the time of death, hypercapnic rats presented with a significant decrease in their blood pH compared with the control group (P = 0.037) (Table 1). This decrease in pH was consecutive to a marked increase in arterial Pco2 levels (P < 0.001) despite a concomitant rise in bicarbonate levels (P < 0.001) (Table 1). The base excess was markedly increased in the treated group compared with the control group (P < 0.001). There was also a decrease in arterial Po2 in the hypercapnic group (P < 0.001). Using the ABL system, the plasma concentration of chloride was significantly decreased in the hypercapnic animals (P < 0.001), whereas the plasma concentration of sodium was slightly increased (P < 0.05) and potassium was unchanged (P = 0.28). Using pure plasma and flame photometry, however, plasma potassium was slightly decreased (P < 0.05). Plasma osmolality and aldosterone were not different between the groups (P = 0.5 and = 0.9) and there was no difference in plasma lactate between the groups (P = 0.4).
In the proximal tubule, NHE3 abundance was unchanged and NBCe1 abundance was increased.
The total abundance of sodium/proton exchanger NHE3 was assessed in the kidney cortex. By semiquantitative immunoblotting, there was no change in the expression of the transporter in the cortex/OSOM (P = 0.27) (Fig. 1). This was confirmed by light microscopy, where an unchanged abundance of NHE3 in proximal tubules was observed using the C00 antibody (Fig. 2, A and B). The L546 antibody showed identical results to C00 (data not shown). As NHE3 is situated in the apical domain of the proximal tubular cells of both control and treated rats, we were not able to distinguish changes in subcellular localization (Fig. 2, A and B) with the resolution of immunoperoxidase microscopy, although this does exclude trafficking over short distances within the labeled area of the brush border. NHE3 labeling in the cortical and medullary thick ascending limb was also not modified, in agreement with an unchanged abundance of NHE3 in the ISOM by Western blotting (data not shown).
As NBCe1 provides the main basolateral exit for bicarbonate in the proximal tubule, the relative change in abundance of this protein was also determined by semiquantitative immunoblotting. The abundance of NBCe1 was increased in the rats exposed to hypercapnia as assessed by immunoblotting (188 ± 22% in hypercapnic rats against 100 ± 11% in control rats, P = 0.005) (Fig. 1, A and B). Immunohistochemistry for this transporter confirmed an increase in NBCe1 expression, apparent as a relative increase in labeling intensity in the basolateral part of the proximal tubule in hypercapnic rats (Fig. 2, C–F). It is also known that NBCe1 is distributed unequally along the proximal tubule, being highly expressed in S1, less in S2, and probably not in S3 of rat proximal tubule (35). Subsegments of the proximal tubule were identified according to their localization and morphological criteria, such as the height of the brush border and the infoldings of the basolateral membrane (25). NBCe1 labeling was present in S1 and S2 in both groups, and was not visible in S3. We could not ascertain the increase in abundance to only one of the segments as both S1 and S2 appeared generally stronger in hypercapnic animals, although a variation in intensities between nephrons was visible. Again, by the resolution of light microscopy we could not assess modifications in subcellular localization of the protein (Fig. 2, E and F).
Pendrin abundance was markedly downregulated in the cortical CNT and CCD of hypercapnic rats.
Pendrin protein abundance was clearly decreased in rats exposed to hypercapnic conditions compared with controls, as assessed by semiquantitative immunoblotting (51 ± 9% in hypercapnic rats vs. 100 ± 7% in control rats, P = 0.003) (Fig. 3). This was confirmed by immunohistochemistry, which demonstrated an overall decrease in the intensity of labeling in hypercapnic rats and an average decrease in the size of cells showing immunoreactivity for pendrin (Fig. 4, A and B). These changes were apparent both in the cortical labyrinth, which encloses both CCDs and CNTs, and in the straight segment of the collecting duct in medullary rays (Fig. 4, C–F). Laser confocal microscopy of sections labeled for both calbindin-28K (marker to distinguish between CNT and CCD) and pendrin confirmed that the decrease in abundance was evident in both the CNTs and the CCDs (Fig. 5). Moreover, we counted the pendrin-positive cells relative to the number of nuclei in five CCDs and CNTs from each animal of the two groups using sections labeled for both pendrin and calbindin. The total number of counted cells in the CNTs was 613 in controls and 647 in hypercapnic rats. The total number of counted cells in CCDs was 846 in control rats and 1,073 in hypercapnic rats. The fraction of pendrin-positive cells was decreased in both CCDs and CNTs of hypercapnic rats compared with controls, although significance was reached (P = 0.002) only in CCDs (Fig. 6). The P value in CNTs was 0.06, which indicates a very likely decrease in pendrin-positive cells also in CNTs, although it did not reach our significance criteria. The lower fraction of pendrin-positive cells in hypercapnic animals could be due to either a decrease in the number of type B and non-A non-B intercalated cells or to a decrease in pendrin expression in the cells, leading to a decrease in the number of cells expressing pendrin at detectable levels.
Pendrin subcellular localization appeared unchanged.
By light microscopy, considerable cell-to-cell heterogeneity of the subcellular repartition of pendrin was evident in control and treated rats (Figs. 4 and 5), and no distinct pattern of trafficking could be ascertained to any of the groups. We therefore performed immunogold electron microscopy of tissue from the cortex of three rats in each group. We addressed here the whole population of cells showing immunoreactivity for pendrin. In the rat, this population is composed of type B intercalated cells and non-A non-B cells, the latter being present mostly in the CNT (18). In both groups, pendrin was located in the apical membrane and in intracellular membranous structures, predominantly in the apical part of the cells (Fig. 7). The ratio of the number of gold particles in the apical membrane divided by the total number of gold particles in the cell was calculated in a minimum of six cells from both the CNT and the CCD per animal. In neither the CNT (0.33 ± 0.02 in control vs. 0.36 ± 0.08 in hypercapnic rats, P = 0.77) nor the CCD (0.28 ± 0.01 in control vs. 0.22 ± 0.03 in hypercapnic rats P = 0.10) was the ratio of gold in the apical membrane to total cell gold modified by chronic hypoxic hypercapnia (Fig. 8). Since only three animals were studied in each group, however, we cannot exclude that a small change in trafficking would have gone undetected by this method.
The abundance of the vacuolar H+-ATPase B1 subunit was unchanged in the cortex and outer medulla of hypercapnic rats.
The vacuolar H+-ATPase in the CCD is situated in the apical membrane of type A intercalated cells and in the basolateral membrane of type B intercalated cells. The B1 subunit is specifically associated with the V-type proton pumps located in the CNT and collecting duct. There was no significant modification of the abundance of the B1 subunit of the proton pump between the groups in either homogenates of tissue from cortex/OSOM or ISOM as assessed by immunoblotting (Fig. 9).
ENaC subunits did not show changes in abundance or enhanced apical targeting.
Since ENaC has been shown to be downregulated in metabolic acidosis induced by NH4Cl loading using a gelled diet (17), we performed immunoblotting and immunohistochemistry for the ENaC subunits. There were no significant changes in abundance of the α-, β-, or γ- subunits of ENaC (Fig. 10, A and B) in cortex/OSOM homogenate. Immunochemistry can assess modification of trafficking of ENaC subunits, which is an important regulatory event for ENaC. Immunohistochemistry for β- and γ-ENaC did not reveal any obvious change in trafficking of these subunits (data not shown).
AQP2 abundance was unchanged in the two groups.
We measured the abundance of AQP2 in the inner medulla and inner stripe of outer medulla of both groups by semiquantitative immunoblotting to further ensure that there was no volume depletion in the treated rats compared with controls. There was no significant difference in the abundance of the water channel in these two zones (125 ± 28 vs. 100 ± 11% in ISOM, P = 0.42 and 119 ± 15% in hypercapnic rats vs. 100 ± 5% in control rats in IM, P = 0.24).
Rats exposed to hypoxic hypercapnia developed a partially compensated respiratory acidosis.
Rats exposed to hypoxic hypercapnia presented a clear increase in their arterial Pco2 and a decrease in Po2. This confirmed the establishment of a mixed hypercapnia-hypoxemia model, mimicking a condition of respiratory failure. The plasma concentration of bicarbonate was increased in hypercapnic rats, allowing for a slight decrease in pH only and confirming a state of partially compensated chronic respiratory acidosis. This is expected, since complete correction of pH is not observed in this condition (37). Given the blood-gas analysis, there was no apparent mixed acid-base disorder in these rats (37). Moreover, the hypoxemia in the treated group was relatively mild, as no lactate elevation was obvious in any of the groups. Therefore, from an acid-base point of view, our model is representative of a pure chronic compensated respiratory acidosis.
In this model, the plasma osmolality, the aldosterone level, and the abundance of AQP2 and ENaC proteins were unchanged, solid arguments against volume depletion and dehydration. We also observe a clear decrease in plasma chloride concentration in the treated animals. The decrease in plasma chloride is part of the response to respiratory acidosis and is of renal origin (23). In the classic view of acid-base physiology, the kidney is expected to compensate for a drop in pH induced by CO2 by retaining bicarbonate and increase proton secretion and the decreased plasma chloride is considered a side effect of the renal changes in proton and bicarbonate transport. According to the Stewart (38) approach, the pH is dependent on Pco2, strong ion difference, and concentrations of weak acids, and bicarbonate is not considered an independent determinant of pH. Accordingly, chronic respiratory acidosis can be interpreted as an acidification due to an increase in Pco2, which the kidney compensates for by increasing the strong ion difference in the plasma, e.g., by increasing the excretion of chloride relative to sodium. Both theoretical approaches thus offer explanations of how the primary change in Pco2 is translated by the kidney into the observed changes in plasma bicarbonate and chloride levels. Since our model includes moderate hypoxemia, as observed in most natural and clinical situations of hypercapnia, it should be considered whether hypoxemia could account for all the adaptations observed. Although we cannot exclude some participation, we find it to be unlikely, since a more global profile of downregulation of membrane transporters would be expected in response to hypoxemia, e.g., as described in ischemic renal failure (12, 21).
The proximal tubule adapts to chronic respiratory acidosis.
Evidence suggests that the proximal tubule plays a role in the increase in bicarbonate reabsorption and proton secretion in chronic hypercapnia (39). In our setting, total NHE3 protein abundance was not upregulated by chronic respiratory acidosis. However, this unchanged NHE3 abundance does not preclude the known increase in NHE3 activity (32, 39), since the specific activity of NHE3 may be regulated, e.g., by phosphorylation (49) and trafficking (47). We also did not address the V-ATPase in the proximal tubule, which also contributes to proton secretion in this segment.
NBCe1 is situated on the basolateral side of proximal tubule cells and mediates bicarbonate exit out of the proximal tubular cell mainly in S1 and S2 (35). The relative abundance of NBCe1 was clearly increased in the cortex/OSOM of hypercapnic rats relative to controls, and immunohistochemistry revealed this increase to take place in both S1 and S2 segments. An increased activity of NBCe1 also occurs in chronic hypercapnic rabbits (32), and an increased synthesis of NBCe1 is described in vitro using an immortalized proximal tubule cell line exposed to high CO2 levels for 12 h (9). This finding, however, is in contrast to the known unchanged abundance of NBCe1 in metabolic acidosis induced by NH4Cl loading (22). This difference might be linked to a specific regulatory pathway in respiratory acidosis.
In summary, in chronic respiratory acidosis, the proximal tubule seems to participate in renal compensation by increasing its absorption of bicarbonate through activation and increased expression of NBCe1, and activation of NHE3 independently of protein expression (5, 9, 32, 39).
Adaptation of the collecting duct to chronic respiratory acidosis.
Compensation with regard to respiratory acidosis is characterized by a decrease in plasma chloride together with an increase in plasma bicarbonate. Increased bicarbonate reabsorption seems to take place partly in the proximal tubule, but it is unclear whether this is accompanied by decreased proximal tubular chloride reabsorption (5). In addition, the collecting duct is also likely to have a role in this compensation via regulation of intercalated cells. In chronic respiratory acidosis, the activity of the proton pump in the CCD and medullary collecting duct is increased (8), suggesting a role for type A intercalated cells in the generation of increased plasma bicarbonate and decreased plasma chloride. In the present study, the abundance of the B1 subunit of the proton pump was not modified. However, the proton pump may undergo cell type-specific expression regulation, which cannot be detected by immunoblotting. Furthermore, unchanged expression does not exclude activation of the pump (8), which, together with modification of subcellular localization (36), may be sufficient to increase acid secretion in the collecting duct as is described in metabolic acidosis (1).
The apical anion exchanger of type B intercalated cell, pendrin, is known to excrete bicarbonate and reabsorb chloride (31). As assessed by immunoblotting and immunohistochemistry, the abundance of pendrin was clearly decreased in the cortex of hypercapnic rats. The decreased labeling intensity was clearly apparent in both CNTs and CCDs, as assessed by laser and light microscopy, and was seen together with a reduction in the fraction of CCD cells showing immunoreactivity for pendrin. In addition to modifications of expression, pendrin has also been described to be regulated by changes in subcellular localization in mice (43, 44, 46). In the present study, however, subcellular redistribution of pendrin was not obvious given the high variability between cells. Quantification of immunogold labeling using electron microscopy confirmed the large cell-to-cell variability and did not reveal significant differences in the fraction of membrane-associated pendrin in CNT nor CCD between the groups. Since only three animals were analyzed in each group by electron microscopy, however, it cannot be excluded that a difference of the magnitude observed could have been detected with statistical significance, notably in the CCD, if the number of animals studied had been higher. Furthermore, we did not differentiate the type B intercalated cells from the non-A non-B intercalated cells due to the lower abundance of this latter cell type in rats than in mice. Thus we cannot formally exclude that a cell type-specific regulation of trafficking occurred.
In any case, the 49% decrease in pendrin expression in the cortex, even without obvious changes in subcellular distribution, implies that the total abundance of pendrin in the apical membrane would also be decreased to ∼50% in hypercapnic rats compared with controls. Interestingly, the observed increased plasma bicarbonate and decreased plasma chloride are exactly the predicted effects of reduced pendrin activity, as could be anticipated to result from this reduced pendrin abundance. Therefore, the downregulation of pendrin may participate in redirecting the CNT and CCD toward excretion of chloride and reabsorption of bicarbonate. Together with an increased activity of A cells and the described adaptations of the proximal tubule, this event is likely to contribute to the observed plasma values.
Pendrin has also been described as playing a role in the distal tubular response to mineralocorticoids (43), but since no difference in aldosterone level was found here between the control and the hypercapnic group, the observed effect on pendrin expression was probably not mediated by this hormone. Moreover, no load of either sodium or chloride was imposed on the animals in this setting. In fact, hypercapnic animals even had a lower NaCl intake than controls due to lower food intake, which therefore cannot explain the observed downregulation of pendrin (46). Our data thus support that the downregulation of pendrin was seen in response to a primary respiratory acid-base equilibrium disturbance induced by hypercapnic hypoxemia.
We describe here the adaptation of the proximal tubule and the collecting duct to hypercapnia and hypoxia, mimicking a condition of chronic respiratory failure. The proximal tubule seems to play a role in the maintenance of high plasma bicarbonate levels as indicated by the increase in the abundance of NBCe1. The CNT and CCD appear also involved in the compensation as a clear reduction in the abundance of the anion exchanger pendrin was observed in response to chronic respiratory acidosis. In this study, pendrin was regulated independently of aldosterone levels and of chloride loading, revealing a potential direct effect of acid-base imbalance on the regulation of this protein. Finally, this downregulation of pendrin may contribute to generate and maintain the observed high plasma bicarbonate and low chloride level in chronic respiratory acidosis.
This study was supported by the NOVO-Nordisk Foundation. The Water and Salt Research Center is established and supported by the National Danish Science Foundation. Sophie de Seigneux was supported by a Medicine and Biology grant (FSMB).
We thank Mogens Koed and Gøsta Dam Jensen for invaluable help in making the environmental chambers and Einar Larsen for construction of the electronic control system. We also thank Lotte Vallentin Holbech, Inger Merete Paulsen, Line V. Nielsen, Ida Jalk, Zhila Nikrozi, and Albert Meier for excellent multifaceted technical expertise and help. We warmly thank Arvid B. Maunsbach, Alicia McDonough, Mark Knepper, and Jeppe Prætorius for very helpful discussions.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society