Acid-base transport by the renal proximal tubule

Abstract

Each day, the kidneys filter 180 L of blood plasma, equating to some 4,300 mmol of the major blood buffer, bicarbonate (HCO₃⁻). The glomerular filtrate enters the lumen of the proximal tubule (PT), and the majority of filtered HCO₃⁻ is reclaimed along the early (S1) and convoluted (S2) portions of the PT in a manner coupled to the secretion of H⁺ into the lumen. The PT also uses the secreted H⁺ to titrate non-HCO₃⁻ buffers in the lumen, in the process creating “new HCO₃⁻” for transport into the blood. Thus, the PT – along with more distal renal segments – is largely responsible for regulating plasma [HCO₃⁻]. In this review we first focus on the milestone discoveries over the past 50+ years that define the mechanism and regulation of acid-base transport by the proximal tubule. Further on in the review, we will summarize research still in progress from our laboratory, work that addresses the problem of how the PT is able to finely adapt to acid–base disturbances by rapidly sensing changes in basolateral levels of HCO₃⁻ and CO₂ (but not pH), and thereby to exert tight control over the acid–base composition of the blood plasma.

Introduction

The maintenance of a physiological ratio of the major blood-plasma buffer parameters, [HCO₃⁻] and P_co2, is at the root of stabilizing blood pH, as described by the Henderson-Has-selbalch equation:

$$\mathrm{p}\mathrm{H}=\mathrm{p}\mathrm{K}+\text{log}\frac{[{{\text{HCO}}^{-}}_3]}{s·{\mathrm{P}}_{\mathrm{C}{\mathrm{O}}_2}}$$ [1]

The lungs control plasma P_co2 by exhaling the CO₂ produced during aerobic respiration (1). The kidneys (2) regulate plasma [HCO₃⁻] by (i) reabsorbing the HCO₃⁻ filtered in the glomeruli (~4,300 mmol/day) and (ii) transporting into the plasma “new HCO₃⁻” that neutralizes the H⁺ arising from sources such as the metabolic production nonvolatile acids (~70 mmol/day).

The renal proximal tubule (PT) is the major site of HCO₃⁻ reabsorption, reclaiming ~80% of the HCO₃⁻ filtered by the glomerulus. Nearly all of the remaining 20% is reclaimed along the distal nephron segments (3). As shown in Figure 1A, the PT cell secretes H⁺ from the cytosol across the microvillus apical membrane into the lumen via the Na-H exchanger (NHE), mainly NHE3 (4, 5), and the vacuolar-type H⁺-ATPase (6). Carbonic anhydrase IV is a membrane-associated carbonic anhydrase (CA) tethered – by a glycosylphosphatidylinositol (GPI) anchor – to the outer leaflet of the apical membrane along the tubule lumen, where it converts secreted H⁺ and luminal HCO₃⁻ to CO₂ and H₂O (7). The CO₂ and H₂O rapidly reenter the cell across the apical membrane, facilitated, as we will describe later, by the aquaporin 1 (AQP1) channel. In the cytosol, CO₂ and ‘H₂O are converted back into HCO₃⁻ and H⁺ by CA II, the HCO₃⁻ exiting with Na⁺ across the basolateral membrane via the renal electrogenic Na⁺/HCO₃⁻ cotransporter (NBCe1-A) at a stoichiometry of 3:1 into the interstitial space and ultimately the blood (8). Other solutes (e.g., glucose, lactate, glutamine) move from the lumen into the PT cell via a variety of Na⁺-coupled transporters and other mechanisms, and subsequently either undergo metabolism in the PT cell or move into the blood (i.e., reabsorption). Cl⁻ and certain other solutes move from lumen to blood across tight junctions by diffusion and solvent drag. Accompanying the net movement of all of the above solutes is the osmotic movement of water, mainly through apical and basolateral AQP1 (9–11).

Fig. 1.

Model of acid–base metabolism by the proximal tubule (PT). A) Mechanism of HCO₃⁻ reabsorption in the early PT. The H⁺ pump (i.e., V-ATPase) and Na-K pump are unlabeled. B) Formation of “new” HCO₃⁻. AQP1 = aquaporin 1; CA II and CA IV = carbonic anhydrases 2 and 4; GDH = glutamate dehydrogenase; Gln = glutamine; Glu = glutamate; α-KG²⁻ = α-ketoglutarate; NBCe1-A = electrogenic Na/HCO₃ cotransporter (1 variant A); NHE3 = Na-H exchanger 3; OAA = oxaloacetate; PEP = phosphoenolpyruvate; PEPCK = phosphoenolpyruvate carboxykinase; SLC6A19 = system B neutral amino acid transporters; SNAT3 (aka SLC38A3) = system N amino acid transporter. *Chronic metabolic acidosis up-regulates NHE3, GA, GDH, PEPCK and SNAT3.

The majority of the H⁺ secreted into the PT lumen titrates filtered HCO₃⁻, the result of which is “HCO₃⁻ reabsorption” (Fig. 1A). The remainder of the secreted H⁺ leads to the creation of “new HCO₃⁻,” introduced above, because the efflux of an H⁺ across the apical membrane is coupled to the efflux of HCO₃⁻ across the basolateral membrane. Viewed from the perspective of the secreted H⁺ to which it is coupled, this new HCO₃⁻ has 3 components (Fig. 1B):

• (a)A tiny fraction remains free in the lumen and is thereby responsible for lowering luminal pH. The remainder titrates filtered buffers other than HCO₃⁻, which renal physiologists have somewhat artificially divided into the following 2 classes.
• (b)When these buffers are anything other than NH₃, we call the result “formation of titratable acid.” Examples of these filtered buffers are dibasic inorganic phosphate (HPO₄²⁻), urate and creatinine. The PT (which can achieve a final luminal pH of ~6.8) is responsible for about half of the titration of HPO₄²⁻ (pK ≅ 6.8) and a very small fraction of the titration of urate (pK ≅ 5.8) and creatinine (pK ≅ 5.0). Distal nephron segments make a far larger contribution for these latter buffers.
• (c)The final component of new HCO₃⁻ is coupled to ammonium secretion. The PT is the principal site of renal ammonium synthesis (12), and the excretion of this NH₄⁺ is the major route for excreting H⁺ equivalents in the urine, in the following sequence of events. First, Na/ amino acid cotransporters mediate the uptake of glutamine (Gln) across both the apical and basolateral membranes. The system B neutral amino acid transporters (SLC6A19 and SLC6A20) mediate the constitutive uptake of Gln across the apical membrane (13), whereas the system N amino acid transporter 3 (SNAT3 or SLC38A3) – up-regulated by acidosis – mediates Gln uptake across the basolateral membrane (13). Once inside the PT cell, the Gln enters the mitochondrion and undergoes hydrolysis via glutaminase to form NH₄⁺ plus glutamate, which then undergoes oxidative deamination via glutamate dehydrogenase to produce NH₄⁺ plus α-ketoglutarate (α-KG²⁻). The newly formed NH₄⁺ dissociates to form intracellular H⁺ and NH₃. The PT cell extrudes the H⁺ across the apical membrane via NHE3 and H⁺ pumps. The NH₃ exits in parallel, probably both through the membrane lipid and via AQP1 (14). Finally, the luminal H⁺ titrates the luminal NH₃ to form NH₄⁺, much of which ultimately appears in the urine. The metabolism of α-KG²⁻ produces CO₂ and glucose (“glu-coneogenesis”). In response to chronic metabolic acidosis, the PT adaptively up-regulates ammonium synthesis to promote H⁺ excretion (12, 13, 15, 16).

In addition to responding to chronic acid-base disturbances, the PT also responds to acute disturbances. We will discuss these below in the section titled “Regulation by acid–base parameters.” To set the stage for this discussion on regulation, we will first introduce the molecular components of acid–base transport at the apical and basolateral sides of the PT cell.

Apical H⁺ extrusion

NHE3

The Na-H exchanger 3 (NHE3) is the most important H⁺-extruder along the PT apical membrane (17, 18). NHE3 utilizes the inward Na⁺ gradient – created by the basolateral Na-K pump – to exchange 1 H⁺ for 1 Na⁺, thereby extruding H⁺ against its electrochemical gradient (Fig. 1A). Adaptive responses to chronic respiratory or metabolic acidosis involve increasing the abundance and activity of NHE3 protein in the apical membrane (19).

The acute regulation of NHE3 involves the trafficking and recycling of NHE3 along the microvilli of the apical membrane via interaction between the C-terminal PDZ motif of NHE3 with the PDZ binding domain of the NHE regulatory factor-1 (NHERF-1). In turn NHERF-1 links NHE3 to the actin cytoskeleton through association with ezrin (20). Ezrin also serves as a low-affinity cAMP kinase (PKA) anchoring protein (AKAP), enabling PKA to phosphorylate NHE3 (21). The actin-based motor myosin VI is involved in trafficking NHE3 from the tip to the base of the microvilli thereby acutely suppressing H⁺ extrusion (20). A complex combination of phospho-serine modifications at the NHE3 carboxy terminus between amino acid residues 455 and 832 – mediated both by PKA and protein kinase C (PKC) – are involved in controlling the majority of these acute regulatory stimuli. Generally, NHE3 activity is inhibited in response to hormonal or other stimuli that, via cAMP formation, enhance PKA. Conversely, NHE3 activity is stimulated in response to PKC activation, with serine modification being necessary (but not sufficient) for increased NHE3 activity (21).

Vacuolar type H⁺ pump or ATPase

Luminal acidification via the vacuolar-type H⁺-ATPase (V-ATPase) is required for a portion of the HCO₃⁻ reabsorption along the PT (Fig. 1A). This apical-membrane H⁺ pump is also the major H⁺-secreting protein in the thick ascending limb (TAL) and the distal nephron, specifically in the α-intercalated cells of the distal convoluted tubule (DCT), connecting tubule and the collecting duct (6). The multi-subunit V-ATPase consists of 2 major structures; an integral membrane V_o ring-like structure that mediates H⁺ movement (subunits a-e) and a peripheral cytoplasmic V₁ ATPase (subunits A-H). Mutations in the kidney-specific subunit B1 encoded by ATP6V1B1 are found in individuals with auto-somal-recessive distal renal tubular acidosis (dRTA) with deafness (type 1b). Mutations occurring in another kidney specific subunit a4, encoded by ATP6V0A4 cause dRTA with preserved hearing (22). The trafficking of the a4 subunit to the apical membrane along distal segments appears to be the major regulatory mechanism for V-ATPase activity in response to acid or NaHCO₃ loading (23).

Apical CO₂ uptake

AQP1 – discovered by Peter Agre and coworkers (24) – is present in high abundance at the apical and basolateral membranes of the PT, where it plays a central role in the near-isotonic reabsorption of H₂O across the PT epithelium (11, 25). Preliminary data suggest that AQP1 also is responsible for a major component of CO₂ movement across the PT apical membrane, thereby playing a major role in HCO₃⁻ reabsorption.

More than a century ago, Overton concluded that NH₃ and other neutral amines – compared with their positively charged conjugate weak acids (e.g., NH₄⁺) – rapidly move through biological membranes (26). This observation and others with neutral weak acids supported Overton’s hypothesis that the cell membrane is predominantly lipid. Others later concluded that all gases move through all membranes simply by dissolving in the membrane lipid (Overton’s rule). However, in the mid-1990s, Waisbren et al demonstrated that the apical membranes of gastric parietal and chief cells have no detectable permeability to either CO₂ or NH₃ (27) – the first evidence that the century-old dogma is not entirely correct. In 1998, Nakhoul et al (28) as well as Cooper and Boron (29), showed that AQP1, overexpressed in Xenopus oocytes, serves a CO₂ channel – the first evidence for a “gas” channel. More evidence has accumulated that AQP1 behaves not only as a water channel but also as a CO₂ channel (30–34).

Some investigators have concluded that CO₂ does not move through channels, based upon experimental or theoretical arguments (35–38). However, those conclusions have themselves been challenged (39, 40). It might be noted that most of the evidence regarding the gas-channel hypothesis is based on work with model systems. Thus, it would be timely to explore the gas-channel model in a physiologically relevant system. If AQP1 is really a gas channel, then – as outlined in Figure 1A – it ought to contribute to the diffusion of CO₂ across the apical and basolateral membranes of PT cells. Moreover, the movement of CO₂ through apical AQP1 ought to contribute to the reabsorption of HCO₃⁻. Preliminary work on isolated, perfused mouse PTs (41) suggests that the knockout of AQP1 reduces maximal HCO₃⁻ reabsorption by ~50% CO₂. Moreover, whereas the knockout of AQP1 has no effect on the transepithelial flux of HCO₃⁻ from the basolateral solution (“bath”) to the lumen, it reduces the transepithelial flux of CO₂ by ~60%. This work would be the first evidence that a channel plays a physiologically important role in a mammalian tissue.

Basolateral HCO₃⁻ efflux

NBCe1-A (SLC4A4)

In 1983, using isolated perfused salamander PTs, Boron and Boulpaep were the first to demonstrate the existence of a Na/HCO₃ cotransporter, and to show that an electrogenic Na/HCO₃ cotransporter is the major route of HCO₃⁻ efflux across the PT basolateral membrane (42). This transporter is electrogenic because it mediates net exit of Na⁺ and more than 1 HCO₃⁻. This cotransport is independent of Cl⁻ but blocked by stilbene derivatives such as 4-acetamido-4’-isothiocyanato-2,2′-stilbene disulfonate (SITS) (42).

In mammalian cells, the Na⁺:HCO₃⁻ stoichiometry must be 1:3 to produce a net efflux of HCO₃⁻ from the PT cell. Indeed, Soleimani et al, working on basolateral membrane vesicles, measured a stoichiometry of 1:3 (8). In 1998, Romero and colleagues expression-cloned the cDNA that encodes the salamander renal electrogenic Na/HCO₃ cotransporter (NBC) – the first Na⁺-coupled HCO₃⁻ transporter to be cloned – and found that NBC is in the same gene family as the anion exchangers AE1-AE3 (43). The original renal electrogenic NBC was eventually renamed NBCe1-A (44), following the discovery of 2 additional splice variants expressed in other tissues (45, 46) plus a second electrogenic NBC (47, 48) and 3 homologous electroneutral transporters (49–51). NBCe1-A is expressed at the basolateral membrane of the PT, at the highest levels in the S1 segment, consistent with its dominant role in mediating renal HCO₃⁻ reabsorption. Preliminary work (52) suggests that NBCe1-A, as expressed in Xenopus oocytes, actually transports carbonate (CO₃⁼). Mutations within the SLC4A4 gene that encodes human NBCe1-A – 12 such mutations are known (53–61) – produce a devastating phenotype that includes type 2 (proximal) RTA (44, 62) and, depending upon the mutation, defects in the eye, teeth and mental development.

AE2 (SLC4A2)

Using microelectrodes on isolated PTs, Kondo and Frömter demonstrated Cl-HCO₃ exchange at the basolateral membrane of the S3 segment. Under physiological conditions, this transporter would couple the exit of 1 HCO₃⁻ into the interstitium to the uptake of 1 Cl⁻ (63). By semiquantitative polymerase chain reaction (PCR), Brosius et al detected AE2 in the convoluted and straight PT as well as more distal segments of rat kidney (64). Thus, AE2 may account for the anion-exchange activity in the S3 segment.

Carbonic anhydrases

The carbonic anhydrase enzymes effectively bypass the slow reaction in the sequence $\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\stackrel{\mathit{\text{Slow}}}{\rightleftharpoons }{\mathrm{H}}_2\mathrm{C}{\mathrm{O}}_3\stackrel{\mathit{\text{Fast}}}{\rightleftharpoons }{{\text{HCO}}_3}^{-}+{\mathrm{H}}^{+}$, and thus are critical for HCO₃⁻ reabsorption and the creation of new HCO₃⁻. Figure 1A summarizes the disposition of CAs in the human PT.

CA II is the archetypal mammalian-class CA. It is a 29-kDa cytosolic protein that is ubiquitously expressed, and is among the fastest of CAs, achieving a turnover rate for CO₂ hydration of 1 × 10⁶/s. Except for the thin ascending limb, CA II is present throughout the nephron, accounting for ~95% of the CA activity in the kidney (7). As illustrated in Figure 1A, cytosolic CA II plays a central role by converting the CO₂ that enters across the apical membrane into H⁺ for secretion into the lumen plus HCO₃⁻ for export across the basolateral membrane. The importance of CA II is illustrated by the effect of inherited CA II deficiency, which causes mixed proximal and distal (type 3) RTA (22) accompanied by osteopetrosis (because of impaired osteoclast function) and cerebral calcification (65).

In humans, the 5% of renal CA activity not due to CA II is accounted for by 2 integral membrane proteins: CA IV and CA XII. The GPI-linked CA IV is localized both api-cally and basolaterally along the PT and TAL, and apically in α-intercalated cells of the cortical collecting duct and cells of the medullary collecting duct (7). As shown in Figure 1 A, apical CA IV catalyses HCO₃⁻ dehydration, thereby consuming secreted H⁺ as well as generating membrane-permeant CO₂ and H₂O. Consistent with this role, CA IV has a higher HCO₃⁻ affinity and a greater HCO₃⁻ dehydratase activity than CA II (66).

The role of basolateral CA IV is unclear. Preliminary work suggests that CA IV minimizes the changes in surface pH caused by CO₃⁼ transport mediated by NBCe1-A, but does not substantially change the activity of the cotransporter (52). Thus, CA IV may protect nearby proteins from extreme pH fluctuations caused by NBCe1-A.

CA XII is a single-span transmembrane protein with its N terminus and catalytic domain in the extracellular space. This protein is exclusively basolateral, present in the human PT but far more abundant in the TAL, DCT and principal cells (which reabsorb NaCl but not NaHCO₃) of the collecting duct (67). In the TAL and DCT – as suggested above for CA IV with CO₃⁼ transport mediated by NBCe1 -A – CA XII might minimize pH fluctuations caused by H⁺ transporters.

Regulation of proximal-tubule acid secretion

Regulation by hormones

Angiotensin II

Angiotensin II (ANG II) is perhaps the most powerful hormonal modulator of Na⁺, fluid, and HCO₃⁻ reabsorption by renal PTs. As we will see below (see the section “Involvement of apical AT₁ receptors,” below), basolateral CO₂ and HCO₃⁻ are equally powerful. Burg and Orloff were the first to examine the effect of ANG II on fluid reabsorption in isolated perfused PTs (68). Since then, many studies involving isolated tubules (69–74) and micropuncture (75– 77) have reported that luminal or basolateral ANG II has biphasic concentration-dependent effects, increasing the fluid reabsorption rate (J_v) and the HCO₃⁻ reabsorption rate (J_HCO3) at low ANG II concentrations, but decreasing J_v and J_HCO3 at higher concentrations. ANG II acts via angiotensin II receptors type 1 (AT₁ (78, 79), which are G protein-coupled receptors (GPCRs), for both its stimulatory and inhibitory effects (80, 81).

Dopamine

Acting via autocrine and paracrine effects, dopamine is a potent natriuretic hormone, acting in part by reducing apical NHE3 protein levels and thereby decreasing Na⁺ and volume reabsorption by the PT (82).

Endothelin

Endothelins are small peptides that induce vasoconstriction but also have other important physiological roles. In the renal PT, chronic metabolic acidosis increases endothelin-1 (ET-1) expression, enhancing its autocrine action on the apical endothelin-B receptor, which in turn stimulates NHE3 (83).

Nitric oxide

Nitric oxide (NO) is produced in renal tubular epithelium by the inducible isoform of nitric oxide synthase (iNOS). In situ microperfusion studies on mouse PTs show that the knockout of iNOS reduces J_v and J_HCO3. Nevertheless, the mouse maintains normal acid–base status via unknown compensatory mechanisms (84).

Regulation by acid–base parameters

The 4 fundamental acid–base disturbances are metabolic alkalosis and acidosis, and respiratory alkalosis and acidosis – defined below. Each of these disturbances can be acute or chronic. As discussed in conjunction with Equation 1, the renal or respiratory systems compensate for each disturbance by appropriately altering blood [HCO₃⁻] or P_CO2, thereby returning arterial pH toward normal. In the process, a mixed-type acid–base disturbance develops (85). Here we will focus on the response of the PT to the 4 acute acid–base disturbances.

Metabolic acidosis (i.e., a decrease in plasma [HCO₃⁻] at a fixed P_co2, resulting in a fall in plasma pH) is compensated largely by an increase in alveolar ventilation, which lowers P_co2. If the source of the defect is extrarenal, the kidney – including the PT (86, 87) – will rapidly adapt by increasing H⁺ secretion. For example, Soleimani et al prepared brush-border or basolateral-membrane vesicles from rabbit PT suspensions preexposed for 2 hours to a pH 6.9 / 5% CO₂ solution. They found that the acidosis increased both apical NHE3 and basolateral NBC activities (88).

Metabolic alkalosis (i.e., an increase in plasma [HCO₃⁻] at a fixed P_co2, causing a rise in pH) is compensated by changes opposite to those for metabolic acidosis. In isolated perfused PTs, increases in basolateral [HCO₃⁻] cause a fall in J_HCO3 that is reversed by also raising P_co2 (89). Creating acute metabolic alkalosis in rats by infusing HCO₃⁻ causes a fall in J_HCO3 as assessed by free-flow micropuncture (90).

Respiratory alkalosis (i.e., a decrease in plasma P_co2, causing a rise in pH) is compensated by a decrease in J_HCO3. For example, Cogan found that hyperventilating a rat to reduce arterial P_co2 by ~20 mm Hg causes a marked fall in J_HCO3, as determined by free-flow micropuncture (91).

Respiratory acidosis (i.e., an increase in plasma P_co2, causing a fall in pH) is compensated by an increase in J_HCO3. More than half a century ago, a series of 3 papers demonstrated that acute respiratory acidosis in dogs rapidly stimulates renal acid secretion (92–94). Based on additional experiments, these authors concluded that it is most likely an increase in P_co2 – rather than a decrease in pH – that controls renal acid secretion in respiratory acidosis. In the more modern era, Chan and Giebisch, working specifically on microperfused PTs, showed that increasing basolateral (BL) pH – either by lowering [CO₂]_BL or by increasing [HCO₃⁻]_BL– causes J_HCO3 to fall dramatically (95). Later work by Cogan, using free-flow micropuncture, showed that acute respiratory alkalosis in rats leads to a decrease in J_HCO3, whereas acute respiratory acidosis has the opposite effect (91).

In addition to the above work on acute acid-base disturbances, others have investigated chronic effects. For example, working with cultured PT cells, Alpern and colleagues have found that chronic (1–7 days) metabolic or respiratory acidosis increases NHE3 activity in cultured cells in a process that involves activation of c-Src and autocrine signaling by endothelin (83, 96–99).

As informative as the above studies have been, all of the maneuvers involved concomitant changes in at least 2 of the 3 key acid-base parameters: pH, [HCO₃⁻] and P_CO2. Thus, it was impossible to determine the extent to which any 1 of the 3 parameters produced the observed effects. Indeed, with classical approaches, it is impossible to change just 1 of the aforementioned 3 parameters as related in Equation 1.

The conundrum of how to isolate the effects of the 3 acid-base parameters was finally addressed in 1995 by Zhao et al (100), who developed a technique for generating out-of-equilibrium (OOE) solutions to produce isolated changes in pH or [HCO₃⁻] or [CO₂], while holding the other 2 parameters constant. The technique was first applied to tubules in a series of experiments by Zhao et al (101), who introduced OOE solutions to the basolateral (BL) surface of rabbit S2 proximal tubules. As described in the next 3 sections, Zhou et al (102) later monitored J_HCO3 while systematically varying, one at a time, pH_BL, [HCO₃⁻]_BL and [CO₂]_BL in isolated, perfused the rabbit proximal tubules (S2 segments).

Basolateral pH

In the first series of experiments (Fig. 2A, B), we increased pH_BL from 6.8 to 8.0 while holding [HCO₃⁻]_BL at 22 mM and [CO₂]_BL at 5%. The middle symbol in Figure 2A (triangle) represents the J_HCO3 under standard, equilibrated conditions: a pH_BL of 7.40, a [HCO₃⁻]_BL of 22 mM and a [CO₂]_BL of 5%. The triangle in Figure 2B represents the corresponding intracellular pH (pH_i). Using OOE technology to lower pH_BL to 6.8 or raise it to 8.0 – always at a [HCO₃⁻]_BL of 22 mM and a [CO₂] _BL of 5% – produces a surprising result: no change in J_HCO3 (Fig. 2A). Nevertheless, the isolated changes in pH_BL produce sizeable changes in pH_i (Fig. 2B). Thus, the PT cell can not acutely regulate J_HCO3 in response to changes in either pH_BLor pH_i per se.

Fig. 2.

Effect of isolated changes in basolateral (BL) pH and [HCO₃⁻] on the rate of HCO₃⁻ reabsorption (J_HCO3) and steady-state intracellular pH. A) Effect of isolated changes in pH_BL, obtained using out-of-equilibrium solutions, on J_HCO3. B) Effect of isolated changes in pH_BL on steady-state pH_i. C) Effect of isolated changes in [HCO₃⁻]_BL on J_HCO3. D) Effect of isolated changes in [HCO₃⁻]_BL on steady-state pH_i. All experiments were performed at 37°C. Data from (102); reproduced with permission in accordance with terms of original publication, Copyright ©2005 by the National Academy of Sciences.

Basolateral [HCO₃⁻]

In the second series of experiments (Fig. 2C, D), we increased basolateral [HCO₃⁻] from 0 to 22 to 44 mM while holding pH_BL at 7.40 and [CO₂]_BL at 5%. The middle symbol in Figure 2C represents the J_HCO3 under the same standard, equilibrated conditions as in Figure 2A. As shown in Figure 2C, we found that graded increases in [HCO₃⁻]_BL cause a fall in J_HCO3. As summarized in Figure 2D, the isolated increase of [HCO₃⁻]_BL from 0 to 22 mM also causes a substantial increase in pH_i. In summary, isolated increases in [HCO₃⁻]_BL produce the appropriate compensatory response for whole-body acid-base balance: a fall in J_HCO3.

Basolateral [CO₂]

In the final series of these experiments (Fig. 3A, B), we increased [CO₂]_BL from 0 to 20% while holding pH_BL at 7.40 and [HCO₃⁻]_BL at either 22 mM (solid symbols) or 0 mM (open symbols). The triangle in Figure 3A represents the same standard, equilibrated conditions as in Figure 2A and C. Figure 3A shows that isolated increases in [CO₂]_BL from 0% to 20% cause graded increases in J_HCO3. As shown in Figure 3B, the graded increases in [CO₂]_BL also cause substantial decreases in pH_i. The data obtained at a fixed [HCO₃⁻]_BL of 0 mM (open symbols) are similar, but translated along the y-axis. The 2 J_HCO3 plots in Figure 3A show that isolated increases in [CO₂]_BL produce the appropriate compensatory response for whole-body acid–base balance: a rise in J_HCO3. A comparison of the 2 J_HCO3 plots in Figure 3A shows that a reduction in [HCO₃⁻]_BL causes an upward shift of the plot, consistent with the data in Figure 2C (compare data at 22 vs. 0 mM).

Fig. 3.

Effect of isolated changes in basolateral (BL) [CO₂] on the rate of HCO₃⁻ reabsorption (J_HCO3) and steady-state intracellular pH. A) Effect of isolated changes in [CO₂]_BL, obtained using out-of-equilibrium solutions on J_HCO3. B) Effect of isolated changes in [CO₂]_BL on steady-state pH_i. All experiments were performed at 37°C. Data represented by solid symbols from ref. (102); reproduced with permission in accordance with terms of original publication, Copyright ©2005 by the National Academy of Sciences. Data represented by open symbols are new.

Although not shown in Figure 2 or 3, none of the isolated changes in pH_BL, [HCO₃⁻]_BL or [CO₂]_BL produced statistically significant effects on J_v Thus, the appropriate adjustments of J_HCO3 to acid-base disturbances do not perturb fluid re-absorption in vitro, and thus presumably would not perturb blood pressure in vivo.

In the 1980s, Al-Awqati and colleagues showed that increases in P_co2 lead to the insertion of H⁺ pumps into apical membranes in turtle bladder (103) as well as PTs and cortical collecting tubules from rabbit (104). It is possible that the same fundamental processes are at work in both the Al-Awqati experiments and the OOE experiments in Figure 3. Al-Awqati and collaborators suggested that CO₂ might act by lowering pH_i and thereby modulating [Ca²⁺]_i Indeed, changes in [Ca²⁺]_i appear to be involved in the OOE experiments. Bouyer et al found that adding equilibrated 5% CO₂ / 22 mM HCO₃⁻ (pH 7.40) has no effect on [Ca²⁺]_i when added to the PT lumen but causes a slowly developing and sustained increase when added to the bath (105). Moreover, switching the bath from a CO₂/HCO₃⁻-free solution to an OOE solution containing 22 mM HCO₃⁻/pH 7.40, but virtually no CO₂ (“pure HCO₃⁻”) causes no change in [Ca²⁺]_i, whereas switching the bath from a CO₂/HCO₃⁻-free solution to an OOE solution containing 5% CO₂/pH 7.40 but virtually no HCO₃⁻ (“pure CO₂”) produces the same rise in [Ca²⁺]_i as does the equilibrated CO₂/HCO₃⁻ solution. Thus, it is the CO₂ that causes [Ca²⁺]_i to rise.

The first portion of Al-Awqati’s hypothesis – that a fall in pH_i is the trigger for the insertion of H⁺ pumps – appears not to be true in the OOE experiments. In Figure 4A, we replot the J_HCO2 data from the previous 2 figures as a function of pH_i. It is clear that pH_i. is not the unique determinant of J_HCO3. In fact, the large changes in pH_i produced by isolated changes in pH_BL (diamonds) elicit no change in J_HCO3. In retrospect, it is perhaps not surprising that pH_i is not the unique determinant of J_HCO3 because, from the perspective of pH_i. regulation, one might have expected the H⁺ extruders at the apical membrane to have the opposite pH_i dependency of the electrogenic Na/HCO₃ cotransporter at the basolateral membrane. A further problem with using pH_i as the unique signal for regulating pH_BL is that different acid–base disturbances can generate the same pH_BL but different pH_i values. In other words, the pH_i signal is degenerate (i.e., “knowledge” of pH cannot inform the cell about the status of pH_BL or the identity of the acid–base disturbance). Furthermore, if pH_i were the critical signal, the dynamics of pH_i regulation (e.g., the recovery of pH_i from an acute intracellular acid load; see (106–108)) would make J_HCO3 quite unstable. The PT cell – faced with the dilemma of selfishly regulating its own pH_i while yet regulating blood pH as part of its raison d’être – seems to have evolved the only way it could have: uncoupling pH_i regulation from J_HCO3.

Fig. 4.

Replots of data from Figures 2 and 3. A) Dependence of J_HCO3 on pH_i for various out-of-equilibrium (OOE) protocols. Here we plot the J_HCO3 data from Figure 2A versus the pH_i, data from Figure 2B, and do the same for Figure 2C versus Figure 2D, and for Figure 3A versus Figure 3B, respecting the symbols and colors in Figures 2 and 3. Data represented by solid symbols from ref. (102); reproduced with permission in accordance with terms of original publication, Copyright ©2005 by the National Academy of Sciences. B) Dependence of J_HCO3 on pH_BL for various OOE protocols. As in panel A, we again replot the data from Figures 2 and 3, but here as a function of pH_BL. C) Dependence of J_HCO3 on the [CO₂]_BL/[HCO₃⁻]_BL ratio for various OOE protocols. As in panels A and B, we replot the data from Figures 2 and 3, but here as a function of the ratio [CO₂]_BL/[HCO₃⁻]_BL. D) Dependence of J_HCO3 on the [HCO₃⁻]_BL/[CO₂]_BL ratio for various OOE protocols. As in panel C, we replot the data from Figures 2 and 3, but here as a function of the ratio [HCO₃⁻]_BL/[CO₂]_BL. None of the 4 parameters on the x-axes can uniquely predict J_HCO3. BL = basolateral; J_HCO3 = HCO₃⁻ reabsorption rate.

Figure 4B is similar to Figure 4A but a comparable plot with pH_BL on the abscissa, shows that pH_BL is also not the unique determinant of J_HCO3 (see the vertical spread of J_HCO3 data at pH_BL = 7.4).

Figure 4C is yet a third replot of the J_HCO3 data in Figures 2 and 3, but this time as a function of the ratio [CO₂]_BL/[HCO₃⁻]_BL. Note the vertical spread of J_HCO3 data at [CO₂]_BL/[HCO₃⁻]_BL = ∞. Figure 4D is similar, but is a plot of J_HCO3 versus the inverse ratio, [HCO₃⁻]_BL/[CO₂]_BL. Here, note the vertical spread of J_HCO3 data at [HCO₃⁻]_BL/[CO₂]_BL = 0. Thus, neither of these ratio parameters is a unique determinant of J_HCO. In fact, the Henderson-Hasselbalch equation tells us that the ratio on the abscissa of Figure 4D is simply 10^(pHBL-pK). In other words, the abscissas in Figure 4B–D are merely transformations of one another. The PT cell clearly responds to isolated changes in [HCO₃⁻]_BL (Fig. 2C) and [CO₂]_BL (Fig. 3A). However, when faced with simultaneous changes in [HCO₃⁻]_BL and [CO₂]_BL, the cell evidently integrates this information in such a way that the [HCO₃⁻]_BL influences the response to [CO₂]_BL– as we saw in Figure 3A – and vice versa. One possibility is that HCO₃⁻ competes with CO₂ for binding to a common receptor.

Involvement of EGFR

In the wake of the above work with OOE solutions (102), a critical question is, how does the PT sense acute changes of [CO₂]_BL and/or [HCO₃⁻]_BL and transduce the signal(s) in the PT cell to regulate bicarbonate reabsorption? In studying the literature on gas-sensing by other organisms, Patrice Bouyer (then in our group) learned that Gilles-Gonzales et al (109) had demonstrated that the bacterium Sinorhizobium meliloti (formerly Rhizobium meliloti) senses low O₂ levels using a 2-component system comprising the regulatory proteins, FixL and FixJ. Low O₂ levels activate the His-kinase activity of the heme protein FixL, which activates FixJ, which in turn activates genes encoding enzymes for nitrogen fxation (see (110)). In 1993, Chang et al found that the ability of the plant Arabidopsis to respond to ethylene, which acts as a hormone (111), depends on the protein ETR1 (112), the C-terminal portion of which is remarkably similar to both FixL and FixJ of the bacterial 2-component system. Because the mammalian cells do not have histidine kinases, Bouyer hypothesized that the CO₂-sensing mechanism of renal PTs requires a receptor tyrosine kinase (RTK) or a receptor-associated (i.e., soluble) tyrosine kinase (sTK) that would interact with a membrane-bound CO₂ sensor.

Zhou and Bouyer began to test a variety of tyrosine-kinase inhibitors for their ability to inhibit the response of the rabbit PT to CO₂. Luckily, the second drug on the list – PD168393, a cell-permeant, highly specific and irreversible inhibitor of the ErbB family of receptor tyrosine kinases (113) – blocked the J_HCO3 response to changes in [CO₂] (114). Another ErbB inhibitor, BPIQ-I, also inhibits the response to [CO₂]_BL. More-over, preliminary data suggest that PD168393 blocks the response to alterations in [HCO₃⁻]_BL (115). The PT expresses both ErbB1 (aka, EGFR or HER1) and ErbB2 (aka, HER2). Preliminary data suggest that exposing PT suspensions to CO₂/HCO₃⁻ causes an increase in the tyrosine-phosphory-lation of ErbB1 and ErbB2 (116). Thus, the CO₂/HCO₃⁻ signal-transduction pathway may pass through ErbB1 and/or ErbB2. Our current signal-transduction model, upon which we will expand in the next several sections, is summarized in Figure 5.

Fig. 5.

Model of CO₂/HCO₃⁻ sensing and signal transduction in the proximal tubule. The transporters are identified in the legend to Figure 1. AnG II = angiotensin II; AT1 = angiotensin receptor type 1; EGFR = epidermal growth factor receptor; PKC = protein kinase C; PLC = phospholipase C; Gq = G protein q; RPTPγ = receptor protein tyrosine phosphatase γ.

Involvement of RPTPγ

We became interested in receptor protein tyrosine phosphatase γ (RPTPγ) because it is a receptor protein tyrosine phosphatase (see (117)) that has an extracellular ligand-binding domain that is homologous to the canonical CAs. Joseph Schlessinger’s group cloned the cDNA encoding RPTPγ (118) and created a RPTPγ-knockout mouse (119). Barnea et al pointed out that the CA-like domain (CALD) of RPTPγ lacks 2 of the 3 His residues needed for coordinating Zn²⁺, and thus suggested that the CALD would be catalytically inactive (118). Preliminary work by Skelton et al (120) indeed suggests that RPTPγ lacks CA activity, but that a combination of 4 mutations (which render the CALD more like CA II) engenders CA activity. Moreover, preliminary work by Zhou suggests that PTs from the RPTPγ-null mouse cannot respond to alterations in either [CO₂]_BL (121) or [HCO₃⁻]_BL (122). RPTPγ mRNA is present in kidney (123), and preliminary work that exploits a newly developed antibody suggests that RPTPγ is expressed at the basal – but not the lateral – membrane of the PT (124). We hypothesize that the CALD of RPTPγ senses CO₂ and/or HCO₃ ⁻ and that the phosphatase domain of RPTPγ then remodels ErbB1, ErbB2 and/or other proteins responsible for transmitting the CO₂/HCO₃⁻ signal.

Involvement of apical AT₁ receptors

We have already discussed the powerful role of ANG II – which acts through apical and basolateral receptors – in controlling J_HCO3 in the PT (see the section “Angiotensin II,” above). An interesting aspect of ANG-II physiology is that the PT secretes an angiotensin-related substance into the lumen (125) (see also (126–128)). Working with rabbit PTs, we found that adding ANG II to the lumen (in addition to the amount secreted) or bath modulated the response to changes in [CO₂]_BL and vice versa (74). A follow-up study (129) showed that the luminal addition of saralasin, a pep-tide ANG II antagonist (130), or candesartan, a non-peptidic antagonist of specifically AT₁ receptors (131), blocks the response of rabbit PTs to changes in [CO₂]_BL. Thus, the response to alterations in [CO₂]_BL requires an active apical AT₁ receptor that is presumably stimulated by secreted ANG II. Because basolateral saralasin had no effect on the [CO₂]_BL dependence of J_HCO3, we can conclude that the PT does not secrete ANG II basolaterally. Finally, PTs from the AT_1A-null mouse (132, 133) exhibit no [CO₂]_BL-dependent changes in J_HCO3. Thus, the J_HCO3 response to altered [CO₂]_BL specifically requires AT_1A receptors at the apical membrane.

To further explore the role of ANG II in the CO₂ signal-transduction pathway, we added lisinopril, an antagonist of angiotensin-converting enzyme (ACE), to the lumen. We were surprised to find that 240 nM luminal lisinopril has no effect on the [CO₂]_BL dependence of J_HCO3. However, this same dose of the ACE inhibitor, when added to the bath, totally eliminates the J_HCO3 response, and even 60 nM basolateral lisinopril produces an inhibition. Thus, it is likely that the PT secretes preformed ANG II and that basolateral lisinopril acts by blocking the conversion of ANG I to ANG II within intracellular vesicles. We have begun using basolateral lisinopril to block the endogenous production of ANG II, allowing us to explore the isolated effect of adding ANG II to the lumen. Preliminary observations suggest that luminal 10^–11 M ANG II has little or no effect on J_HCO3 when [CO₂]_BL is 0%, but that the effect of this dose of ANG II increases in a graded fashion as we raise [CO₂]_BL to 5% and then to 20% (134). Thus, at least part of the explanation for how basolateral CO₂ controls J_HCO3 is that CO₂ may enhance the action of luminal ANG II. Note that basolateral CO₂ has the opposite effect on the response to basolateral 10⁻¹¹ M ANG II: increasing levels of [CO₂]_BL reduce the stimulation by ANG II in a graded fashion (74). Thus, the signal-transduction pathways for [CO₂]_BL and ANG II interact in a complex way.

Involvement of protein kinase C (PKC)

It is generally accepted that the stimulatory effect of low-dose AT_1A in the PT occurs through PKC. Indeed, PKC activation stimulates apical NHE3 (135), enhancing Na⁺ reabsorption (72), as well as HCO₃⁻ and fluid reabsorption (136). Moreover, the stimulatory effect of low-dose ANG II appears to occur via PKC in the case of enhanced H⁺-pump activity (137), enhanced apical Na-H exchange activity (138, 139) and enhanced HCO₃⁻ and water reabsorption (136). Evidence from rat PTs points to PKC-ζ as being the critical isoform (139). In preliminary work on isolated perfused rabbit S2 PTs, a PKC inhibitor eliminates the CO₂-evoked increase in J_HCO3 (140). This result is consistent with the hypothesis that PKC plays an important role in the CO₂ signal-transduction pathway.

Soluble adenylyl cyclase and G protein–coupled receptors

Although it appears to play no role in sensing acid–base disturbances in the PT, the cytoplasmic/soluble adenylyl cyclase (sAC) is an evolutionarily conserved, cytosolic HCO₃^– chemosensor related to cyanobacterial adenylyl cyclases. Upon activation by HCO₃⁻, sACs catalyze the conversion of ATP to cAMP (141). In the kidney, sAC has been localized to the TAL, distal tubule and collecting duct (142). In a collecting-duct cell model, sAC increases Na⁺ reabsorption in response to alkalosis (143). In the dogfish, sAC plays an important role in systemic acid–base homeostasis, specifically in the gills. Here, alkalosis – presumably by raising pH_i and therefore [HCO₃⁻]_i – stimulates V-ATPase insertion into the basolateral membrane, enhancing H⁺ absorption into the body (144). A potential conundrum, assuming that the acid–base sensitivity of sAC uniquely reflects changes in [HCO₃⁻]_i^, is that both systemic metabolic alkalosis and systemic respiratory acidosis would raise [HCO₃⁻]_i.

In 2003, Ludwig et al (145) were the first to report that ovarian cancer G protein–coupled receptor (OGR1) is a proton-sensing receptor that stimulates inositol phosphate formation. Half-maximal activation of OGR1 occurred at pH 7.50, and was increasingly stimulated at more acidic pH, maximizing at pH 6.8. A related receptor, GPR4 is also proton sensitive – in this case stimulating cAMP formation. Conserved extracellular histidine residues in OGR1 and GPR4 are important for H⁺ sensing (145). The potential role for OGR1 or a related GPCR in acid–base sensing along the nephron is an attractive possibility. Finally, in Drosophila, a pair of GPCRs – Gr21a and Gr63a – act as a sensor for a CO₂-related substance (146). OOE technology presumably could identify the true ligand of this insect sensor. It is not clear whether mammals have GPCRs with a similar function.

Concluding Remarks

Every cellular and bodily function depends on pH, everything from control of the cell cycle at one extreme to the muscle contraction that underlies exercise at the other. Thus, the regulation of intracellular pH – and the whole-body acid-base homeostasis on which pH regulation depends – are of major importance. The past century has seen the defining of pH and buffering power, the realization that the lungs excrete CO₂ and that the kidneys (including the proximal tubule) excrete acid into the urine, the discovery that acid-base status regulates these processes, the discovery of pH_i regulation, the identification and cloning of the responsible acid-base transporters and advances in the understanding of regulatory pathways. In the present review, we focus on the mechanism of HCO₃⁻ reabsorption by the PT cell (Fig. 1). Figure 5 summarizes our view of the acute regulation of this process. RPTPγ appears to be central in the response to alterations in [HCO₃^–]_BL (Fig. 2C) and [CO₂]_BL (Fig. 3A), and the figure suggests that HCO₃⁻ and CO₂ may compete for binding to RPTPγ. However, we really do not yet know the identities of the sensors for molecular HCO₃⁻ and CO₂. Although an RTK such as ErbB1/2 also appears to be essential for the response to alterations in [HCO₃⁻]_BL and [CO₂]_BL, we do not know how the signal crosses from the basolateral to the apical membrane. Following the action of luminal ANG II on the AT1A receptor, one can imagine how G protein q (Gq) might activate the apical H⁺-extrusion mechanisms. However, it is not clear if or how the signal crosses back to the basolateral membrane to stimulate NBCe1-A. Although many unknowns remain, it is already clear that the acute regulation of acid–base transport by the PT depends not on blood pH per se but on the 2 parameters that define blood pH: [HCO₃^–]_BL and [CO₂]_BL.