INTRODUCTION

TOP ABSTRACT INTRODUCTION DETERMINANTS OF CELL VOLUME REGULATORY VOLUME INCREASE REGULATORY VOLUME DECREASE CHANGES IN THE SET... SUMMARY REFERENCES

THE NECESSITY TO RAPIDLY REGULATE volume after shrinkage or swelling is obvious in cells from lower organisms, which are continually exposed to osmotic stress. Yet, in mammals, where osmolarity of the internal milieu is carefully regulated, cells also exhibit rapid volume-regulatory responses. Are volume-regulatory pathways in mammalian cells merely a vestige that have no significant physiological role? The answer to this question lies in the realization that mammalian cells are still subject to volume changes, despite a relatively constant osmolarity. This is most apparent in transporting epithelia, where slight changes in the large apical and basolateral fluxes could lead to rapid changes in cell volume. Furthermore, even if cell volume were not threatened, mechanisms are still needed to alter cell volume, such as during hypertrophy and atrophy, apoptosis, and differentiation. Thus mammalian cells clearly require volume-regulatory mechanisms, despite their isosmotic existence. Although the principal volume-regulatory transporters have been identified and characterized, their activation by changes in cell volume is poorly understood and their physiological role is largely unexplored. This review focuses on the salient points of volume-regulatory ion transporters that have direct relevance to mammalian physiology, with particular emphasis on the distinction between anisosmotic and isosmotic volume regulation, the role of intracellular Cl, and the need for cell volume regulation in vivo. Space does not permit a detailed discussion of the intracellular sensing and signaling mechanisms, and readers are referred to a recent, comprehensive review on this subject (94).

	DETERMINANTS OF CELL VOLUME

TOP ABSTRACT INTRODUCTION DETERMINANTS OF CELL VOLUME REGULATORY VOLUME INCREASE REGULATORY VOLUME DECREASE CHANGES IN THE SET... SUMMARY REFERENCES

In discussing cell volume it is important to distinguish between factors that determine steady-state volume and mechanisms that correct acute perturbations in steady-state volume. This has been a source of confusion, and it is best to consider these as entirely separate processes. Transporters responsible for correcting acute changes in cell volume are usually not active at steady-state volume. Thus steady-state cell volume is not a balance between their ongoing, opposing actions. Rather, cells appear to establish a steady-state volume and then activate volume-regulatory transporters only when cell volume deviates from this "set point." This is far more efficient than having opposing volume-regulatory transporters continuously active.

Steady-State Cell Volume

Another simple maneuver that causes cell swelling is replacement of extracellular Na⁺ with K⁺. Because neither the Gibbs-Donnan equilibrium nor other osmotic forces are altered (115), elimination of the outward K⁺ gradient is directly implicated in cell swelling. By virtue of the selective permeability of the plasma membrane to K⁺ over Na⁺, this gradient results in an outward K⁺ current and a negative membrane potential that dictates a low intracellular concentration of permeant anions, primarily Cl. The source of energy is ultimately the Na⁺-K⁺ pump, explaining why its inhibition often leads to cell swelling. This ability to extrude Cl (coupled with extrusion of Na⁺ by the Na⁺-K⁺ pump) appears to be the principal mechanism that maintains steady-state cell volume. In cells with high anion permeability and low membrane potential, such as erythrocytes, it is the net cation extrusion by the Na⁺-K⁺ pump coupled with low membrane permeability to both Na⁺ and K⁺ that limits Cl entry. In erythrocytes from carnivores, which lack Na⁺-K⁺ pumps, net cation extrusion occurs through combined action of Ca²⁺ pumps (Ca²⁺-ATPases) and Na⁺/Ca²⁺ exchangers. The former generates a very large inward Ca²⁺ gradient that is then used to pump out Na⁺ via the latter (127). Cl accompanies Na⁺ efflux by virtue of the electrogenicity of both transporters.

Acute Changes in Steady-State Volume

Cells instead recruit additional mechanisms to correct acute deviations in their volume. These processes, termed regulatory volume increase (RVI) and regulatory volume decrease (RVD), occur through activation of specific transporters in the plasma membrane that mediate net fluxes of osmotically active molecules (and therefore water). Not surprisingly, both [Cl]_i and macromolecule concentration play key roles in governing these volume-regulatory responses. A further adaptation to osmotic stress is the accumulation of organic osmolytes such as sorbitol, sn-glycero-3-phosphorylcholine, betaine, and taurine, which occurs in response to increased ionic strength rather than cell shrinkage (70, 207) and therefore is not truly a volume-regulatory mechanism. Instead, these compounds replace high intracellular salt concentrations that can perturb protein structure and impair enzyme function (18, 70), enabling cells to function in high osmolarity, such as in the renal medulla. Organic osmolytes also appear in the brain during chronic hypernatremia and comprise the so-called idiogenic osmoles (195). Organic osmolytes do have an impact on volume regulation, because they must be rapidly removed when normal osmolarity is restored in order to prevent cell swelling.

The mechanism by which cells sense changes in their volume is unknown but clearly is not through changes in osmolarity, ionic strength, or ion concentration, since isosmotic volume changes also activate RVI and RVD. Cell volume could be sensed chemically as changes in the intracellular concentration of impermeant molecules or mechanically through some form of stretch receptor. The former possibility, proposed by Minton et al. (129) and based on the concept of macromolecular crowding, is supported by direct experimental data. According to this theory, it is not cell volume per se that is regulated but rather the state of intracellular water. In erythrocytes, volume-regulatory transporters can be activated in the absence of changes in cell volume by altering intracellular protein concentration or by adding agents that perturb protein hydration (157). These findings have been extended to nonerythroid cells, specifically renal tubular cells (83) and barnacle muscle (200). The question that remains is the nature of the sensor for macromolecular crowding. Mechanical sensing of cell volume is an attractive theory, since cell structure is directly threatened, but the evidence is indirect. Agents that perturb the actin cytoskeleton can inhibit RVD (24, 25, 42, 49, 184), and channels activated directly by membrane stretch can also be activated by cell swelling (176). Mechanical sensing of changes in cell volume remains speculative since no role for the cytoskeleton has been demonstrated in RVI, and the role of stretch-activated channels in RVD remains unproven. The issue of how cells sense their volume is far from resolved, and it is quite possible that both chemical and mechanical sensors are employed.

Consequences of Cell Shrinkage and Swelling

	REGULATORY VOLUME INCREASE

TOP ABSTRACT INTRODUCTION DETERMINANTS OF CELL VOLUME REGULATORY VOLUME INCREASE REGULATORY VOLUME DECREASE CHANGES IN THE SET... SUMMARY REFERENCES

Volume-Regulatory Transporters

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Transporters involved in regulatory volume increase. NKCC1, Na+-K+-2Cl transporter; NHE1, Na+/H+ antiporter.

The mechanism by which cell shrinkage activates NKCC1 or NHE1 is unknown. Because activation occurs within minutes, it cannot be due to increased synthesis of transporters. Current data suggest that protein phosphorylation is involved. Activation of NKCC1 by shrinkage is associated with phosphorylation of the transporter (on serine and threonine, but not tyrosine) and can be blocked by kinase inhibitors and mimicked by inhibitors of protein phosphatases (90, 163). Although these agents have a similar effect on NHE1 activity (9), cell shrinkage does not increase phosphorylation of the transporter (53). The data imply the existence of a volume-sensitive protein kinase (or more than one) that phosphorylates NKCC1 and a regulatory protein that activates NHE1. Hopefully, the recent demonstration of a volume-sensitive kinase that phosphorylates NKCC1 in vitro will lead to identification of this kinase (145). Although the identity of this kinase is unknown, other volume-sensitive kinases have been tentatively identified. One of these is myosin light chain kinase (MLCK), which is activated by shrinkage in vascular endothelial and smooth muscle cells (89), mesangial cells (203), and glial cells (188). Inhibition of MLCK blocks activation of NKCC1 in shrunken endothelial cells (89) and activation of NHE1 in shrunken astrocytes (188). However, MLCK is not the volume-sensitive kinase that phosphorylates NKCC1, and the mechanism by which myosin light chain phosphorylation influences NKCC1 and NHE1 is unknown. Hypertonicity increases phosphorylation of stress-activated protein kinases and mitogen-activated protein (MAP) kinases (45, 69). It is not known whether this is an effect of hypertonicity or cell shrinkage, and phosphorylation of MAP kinase also occurs after hypotonicity, suggesting a nonspecific stress response rather than a specific response to cell volume. However, phosphorylation and activation of the tyrosine kinases p59^fgr and p56/59^hck occurred in neutrophils shrunken both hypertonically and isosmotically (91), directly implicating cell shrinkage.

Hypertonic vs. Isosmotic Shrinkage

Can thermodynamics explain the lack of hypertonic RVI? When hypertonicity is produced with an impermeant molecule such as sucrose or mannitol, the concentration of extracellular ions is unchanged while intracellular ions increase in concentration, thereby reducing the driving force for influx. This can actually lead to further cell shrinkage (101) and is accentuated by the occasional use of impermeant molecules in place of NaCl in the reference isotonic solution. However, hypertonic RVI is still infrequent when NaCl is the osmotic agent. Given reasonable estimates of intracellular ion activities, activation of NKCC1 should result in net influx of ions, primarily because of the inwardly directed Cl gradient. In the absence of accurate determinations of intracellular ion activities, it can always be argued that NKCC1 is at equilibrium. However, substitution of external Na⁺ with K⁺ to equalize their concentrations still does not result in net influx and RVI, despite the fact that the driving force for influx is increased sevenfold (143, 168). Furthermore, thermodynamics cannot explain a lack of hypertonic RVI by NHE1 and anion exchange, provided some CO₂ and HCO₃ are present. These results indicate that the lack of RVI in hypertonic medium is a regulatory event rather than a thermodynamic effect. This could be because the responses to the two types of shrinkage are inherently different (40) or could be explained by superimposed regulatory effects of ionic strength or intracellular ion concentrations. In either event it is clear that cells can sense how they are being shrunk.

Although the difference between isosmotic and hypertonic shrinkage could be sensed as a change in ionic strength or an agonist-induced signal, the most logical candidate is [Cl]_i. Cl is a minority anion in most cells, yet it is the major permeant anion, with the majority of anions being impermeant proteins and organophosphates. While hypertonic shrinkage increases [Cl]_i in parallel with all other cytoplasmic constituents, [Cl]_i decreases with isosmotic shrinkage, since the fluid moving out of cells has a higher Cl concentration than cytoplasm. As with hypertonic shrinkage, the concentration of all other cytoplasmic constituents increases (except for K⁺, which should change little since it comprises the vast majority of intracellular cations). The decrease in [Cl]_i is actually greater than the decrease in cell volume, making Cl concentration a more sensitive indicator of changes in cell volume than cell volume itself under isosmotic conditions. As shown in Fig. 2, if Cl accounts for one third of cellular anions and a cell loses one sixth of its volume via loss of KCl, half the cell Cl is lost and [Cl]_i theoretically decreases by 40% (1/2 the original Cl content in 5/6 the original volume). It is not surprising then that [Cl]_i has important effects on volume-regulatory transporters.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Theoretical relationship between intracellular Cl concentration ([Cl]i) and cell water during isosmotic shrinkage, where steady-state [Cl]i comprises one third of intracellular anions and volume changes are mediated by isosmotic fluxes of KCl.

[Cl]_i inhibits NKCC1, and this occurs through inhibition of phosphorylation and probably through a direct effect on the transporter as well. Phosphorylation of NKCC1 is stimulated by lowering [Cl]_i and inhibited by raising [Cl]_i (57, 111), probably via effects on the same kinase activated by cell shrinkage (110). However, this cannot be the principal mechanism that blocks hypertonic RVI, since there is still a substantial increase in NKCC1 phosphorylation with hypertonic shrinkage, suggesting that Cl has an additional, direct effect on NKCC1. The inhibition of NKCC1 by Cl in internally perfused squid axons (16) is consistent with this hypothesis. In a sense this represents classic feedback inhibition of NKCC1 and provides an ideal yet simple mechanism for regulating cell volume. The nature of this Cl effect is unclear. In endothelial cells, hypertonic and isosmotic shrinkage activate unidirectional influx through NKCC1 to about the same degree, yet net influx only occurs after the latter (143). This suggests that intracellular Cl blocks net influx while allowing the partial reaction [K⁺(Rb⁺)/K⁺ exchange] to continue. Evidence for inhibition of NKCC1 by internal Cl has also been obtained in Ehrlich ascites cells (101), salivary acinar cells (168), and tracheal epithelial cells (58). In many cells NKCC1 apparently functions in a similar mode at steady-state volume as well (21, 46, 101, 116, 143, 150), essentially "running in neutral" due to inhibition by intracellular Cl, yet poised for rapid influx when [Cl]_i decreases. Because cells already expend energy to counteract swelling due to the Gibbs-Donnan equlibrium, it makes little sense to allow net influx via NKCC1 or NHE1 at steady-state volume.

The mechanism by which hypertonic shrinkage prevents RVI by NHE1 has not been examined. In the absence of CO₂/HCO₃, RVI could be limited by the alkalinization that accompanies Na⁺ influx and rapidly shuts off NHE1. For instance, if shrinkage produces a 0.2 unit alkaline shift in the pH set point for NHE1 and the internal buffering capacity is 30 mM/pH unit, NHE1 shuts off after an influx of only 6 mM Na⁺. The fact that more substantial RVI via NHE1 can occur in the nominal absence of CO₂/HCO₃ (after isosmotic shrinkage) indicates that there is still enough CO₂/HCO₃ to permit some Cl/HCO₃ exchange. Because hypertonic RVI requires more Na⁺ influx, it may be substantially slower under these conditions. However, there appears to be an additional mechanism since hypertonic shrinkage can actually suppress activation of NHE1, presumably due to an effect of internal Cl (168).

Inhibition of RVI by intracellular Cl may explain the variable responses to hypertonic shrinkage noted in Table 1. Epithelial cells, which frequently exhibit hypertonic RVI, may have a significantly lower [Cl]_i that is below the threshold for inhibition of RVI even after hypertonic shrinkage. Some epithelia that do not exhibit RVI under standard hypertonic conditions do so in the presence of agents that raise cAMP levels (43, 54, 202), most likely via activation of Cl channels and reduction of [Cl]_i. Hypertonic RVI by C6 glioma cells appears to require additives as well, including insulin and PGE₁ (133). This could also be due to a decrease in [Cl]_i, but the possibility that these agents might modulate the inhibitory effect of Cl cannot be ruled out. An additional additive that confers hypertonic RVI on renal epithelial cells is butyrate (135, 170). The mechanism is not known, but the RVI is only partly accounted for by monovalent ions (170), suggesting that butyrate stimulates the synthesis of organic osmolytes. The occasional observation of hypertonic RVI in other cells and the discrepancies between studies employing similar cells could be due to subtle differences in culture conditions or the inadvertent presence of agonists or growth factors, perhaps autocrine in nature. Hypertonic RVI may also depend on the rate of shrinkage. In renal proximal tubules, no shrinkage occurred when tonicity was increased 1.5 mosM/min, whereas the cells behaved as perfect osmometers at 3 mosM/min (104). A similar response has been described in C6 glioma cells (134), but these cells do exhibit some RVI after acute hypertonic shock (133). The basis for this so-called isovolumetric regulation is unknown but could relate to the fact that [Cl]_i is not markedly elevated during shrinkage. This phenomenon is probably not widespread, since most tissues do not exhibit RVI during hypertonicity in vivo, which also develops gradually.

Hypertonic Shrinkage In Vivo

Based on both in vitro and in vivo data, it appears that mammalian cells are poised to undergo RVI after isosmotic shrinkage but not hypertonic shrinkage. This might seem appropriate since these cells reside in an environment of constant osmolarity. But is the osmolarity of mammalian plasma and interstitial fluid truly that constant? Despite the best efforts of the hypothalamus, kidneys, colon, and skin, significant water losses still occur during water deprivation. Thus osmolarity fluctuates with the frequency of water intake. In various rodents and ungulates, particularly those that reside in arid areas, water loss that exceeds 20% of body weight is well tolerated (112, 181). Deprivation for as short as 1 day can lead to a 6% or greater weight loss (112, 181). Some of this fluid comes from the intestinal lumen (113) and there is also excretion of electrolytes, so the corresponding rise in osmolarity is not as great. Deprivation of water increases plasma osmolarity 6-7% after 8 days in kangaroo rats (193), 4 days in dogs (224), and 3 days in Moroccan goats (76). In humans, strenuous exercise in hot conditions without water can result in water losses up to 5% of body weight (8% of total body water, 7% increase in osmolarity) without functional impairment (26, 100). Because periods of water deprivation are probably common in mammals, hypertonicity appears to be a routine occurrence that is well tolerated.

The lack of hypertonic RVI may instead be due to potential detrimental effects on the organism. Substantial water losses are often well tolerated, without circulatory collapse, because the loss is distributed across all compartments (intracellular, interstitial, and intravascular). If RVI occurred, the water loss would be concentrated in the one third of body water that is not intracellular. A 10% loss of body water would instead result in a 30% decrease in extracellular volume, including intravascular volume. Thus the absence of RVI helps maintain vascular volume during water deprivation, essentially allowing tissues to serve as a water reservoir (112). Another serious consequence of RVI would be the profound transcellular shift of K⁺. Correction of a mere 1% cell shrinkage via cellular uptake of KCl could decrease plasma K⁺ concentration as much as 3 mM.

Because of the potential detrimental effects of hypertonic RVI and the seemingly minimal detrimental effects of mild hypertonic shrinkage, RVI may be sacrificed for the benefit of the whole organism. Exceptions appear to be intestinal epithelium (138) and brain (3, 28, 29, 137, 192) and cells cultured from these sites (116, 133). Because intestinal epithelium is one of the few tissues routinely exposed to sudden changes in osmolarity, hypertonic RVI would be appropriate, particularly since cell shrinkage could disrupt the mucosal barrier. The brain is a critical organ with limited capacity to shrink without structural damage because of the inelasticity of the skull. Another exception may be chondrocytes. The extracellular fluid in cartilage is hypertonic compared with plasma, owing to the high concentration of fixed, polyanionic macromolecules, and this can change with applied load (64). Hypertonic RVI is occasionally observed in chondrocytes in situ but not in isolated chondrocytes (Ref. 63 and A. C. Hall, personal communication). The specific mechanism that allows hypertonic RVI to proceed in some tissues and cells but not in others is unknown.

Isosmotic Shrinkage In Vivo

	REGULATORY VOLUME DECREASE

TOP ABSTRACT INTRODUCTION DETERMINANTS OF CELL VOLUME REGULATORY VOLUME INCREASE REGULATORY VOLUME DECREASE CHANGES IN THE SET... SUMMARY REFERENCES

Volume-Regulatory Transporters

An additional transporter that mediates RVD is the K⁺-Cl cotransporter (KCC), which transports K⁺ and Cl stoichiometrically in either direction across the plasma membrane (98). Although structurally related to NKCC1 and inhibited by furosemide, this transporter is distinguished by its Na⁺ independence and insensitivity to bumetanide. Among physiological anions, it has a strict requirement for Cl and, not surprisingly, is the principal mechanism for RVD in cells with a high [Cl]_i such as erythrocytes. Studies of KCC have been hampered by the lack of a specific inhibitor, but, based on mRNA abundance, the transporter is probably more prevalent than previously thought (47, 48, 132). For instance, it appears to be responsible for about one-half the RVD in aortic endothelial cells (160).

The mechanisms by which cell swelling activates volume-sensitive channels and KCC are poorly understood. The nature of the K⁺ channels involved suggests Ca²⁺ as a signal, and evidence in support of this has been obtained primarily in epithelial cells (22, 31, 72, 121, 122, 172, 173, 189, 221). The hypothesis based on these data is that cell swelling induces Ca²⁺ influx, possibly via stretch-activated channels, leading to a rise in [Ca²⁺]_i and subsequent activation of BK channels. However, in many other cells only a small rise in [Ca²⁺]_i occurs and is not required for activation of K⁺ efflux (52, 74, 125, 161, 222). One possible explanation for this discrepancy is the distinction between dependence on Ca²⁺ and activation by Ca²⁺. The K⁺ channels may require Ca²⁺ for activity, but swelling could open them through a separate mechanism. Maneuvers that block increases in [Ca²⁺]_i (such as removing extracellular Ca²⁺) often lower basal [Ca²⁺]_i, thereby preventing channel activation by any mechanism. Further evidence against the Ca²⁺ theory is that Ca²⁺ is not involved in the activation of anion channels or K⁺-Cl cotransport in swollen cells. Because the evidence for Ca²⁺ signaling of RVD is quite compelling in epithelial cells, it is reasonable to hypothesize two pathways for activation of K⁺ channels by cell swelling: a mechanosensitive Ca²⁺ influx that is the primary pathway in epithelia and a Ca²⁺-independent pathway operational in nonepithelial cells. The nature of the Ca²⁺-independent signal is unknown, but phospholipase A₂ has been implicated in volume-sensitive K⁺ and anion channels (136, 147), whereas protein kinase inhibition has been implicated in the activation of KCC in swollen erythrocytes (11, 82, 84). Unfortunately, it has not yet been possible to determine whether the cotransporter itself is phosphorylated.

Hypotonic vs. Isosmotic Swelling

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Regulatory volume decrease (RVD) in aortic endothelial cells, measured as changes in cell content of Na+ and K+. Top: cells placed in medium made hypotonic by removal of NaCl. Middle: RVD after isosmotic swelling. Cells were preincubated in isosmotic KCl medium and then placed back in standard isosmotic medium. Bottom: Cells were placed in hypotonic medium with 50 µM bumetanide. Dashed line in each panel represents expected values for complete RVD. [Adapted from O'Neill and Klein (143).]

Hypotonic Swelling In Vivo

Isosmotic Swelling In Vivo

Acidosis can also increase cell volume, a response that is distinct and opposite from the effect of pH on the Gibbs-Donnan equilibrium. By titrating negative charges on impermeant molecules, a reduction in pH decreases the tendency of cells to swell. However, this is more than offset by accumulation of the anions that accompany the H⁺ and replace impermeant anions titrated by the H⁺ (79). If intracellular buffering capacity is 30 mM/pH unit, lowering cell pH by 0.5 with a monovalent acid would lead to accumulation of 15 mM anion. Cell swelling occurs during all three types of acidosis: organic, inorganic, and respiratory. With organic acids such as lactic acid, the anion accumulates not only within cells in which they are synthesized but also in other cells, because the acids are weak and permeant and because the anions can be transported (105). The resulting swelling has been demonstrated in astrocytes (80, 81, 105) and erythrocytes (191). Hydrochloric acidosis, the principal inorganic acidosis in vivo, can be created in vitro by replacing HCO₃ with Cl. The assumption is that HCO₃ leaves cells in exchange for Cl, transmitting the acidosis to the cytoplasm. Because of intracellular buffering, the amount of HCO₃ leaving the cell and replacement anions entering will be greater than the drop in HCO₃ concentration, leading to an increase in osmotically active molecules. In renal proximal tubular cells, a decrease in HCO₃ concentration from 25 to 5 mM increased cell volume by 6% (198), but there was no accumulation of Cl (199), suggesting a different counteranion. In respiratory acidosis CO₂ diffuses into cells and dissociates into H⁺, which are buffered, and HCO₃, which accumulates and is partly exchanged for extracellular Cl. Accordingly, an increase in CO₂ from 5 to 15% produced swelling of renal proximal tubules accompanied by an increase in the cellular content of both HCO₃ and Cl (199). An additional mechanism of swelling in acidosis is activation of counterregulatory pathways. In astrocytes exposed to organic acids, activation of NHE1 by intracellular acidification leads to a second phase of swelling (80, 81). It is not known whether RVD occurs during acidification in vitro, and limited data are available in vivo. Respiratory acidosis (10% CO₂ for 3 h) did not increase cell water in rat heart, brain, spleen, liver, or skeletal muscle (171), implying that RVD had occurred. In contrast, exposure of rat skeletal muscle to 30% CO₂ for 95 min increased cell water by 11% (177), but such severe acidosis may have overwhelmed or inhibited RVD. In muscle biopsies from humans during vigorous exercise, there was a small, insignificant increase in intracellular water despite substantial cellular lactate accumulation (190), again implying some volume correction.

Because the negative membrane potential is largely responsible for maintaining cell volume by limiting Cl accumulation, depolarization can lead to cell swelling. In vivo, this would be most apparent in excitable tissues such as nerve and muscle, particularly where depolarization is due to opening of Na⁺ channels. The expected swelling has in fact been demonstrated in nerve fibers (78) and in cultured myogenic cells (185). Although the Na⁺-K⁺ pump can rapidly export the excess Na⁺, much of this is exchanged for K⁺, with the resulting net efflux of only one osmotically active molecule per cycle. Thus the Na⁺-K⁺ pump is an inefficient volume-regulatory transporter that cannot rapidly relieve cell swelling. Swelling of skeletal muscle during repetitive contraction would be particularly problematic since it could compromise regional blood flow. Such swelling occurs during vigorous exercise, accompanied by an increase in Cl content (123, 190), but the swelling cannot be ascribed solely to repetitive depolarization since acidosis could also contribute to the swelling. The increase in intracellular water is small, implying that volume regulation has occurred. Volume regulation via Ca²⁺-activated K⁺ channels could serve the dual purpose of repolarizing the membrane potential (15). A dramatic example of isosmotic swelling in vivo, possibly related to depolarization and acidosis, occurs after ischemia (39, 215). Although ischemia directly swells cells in vitro (92), in vivo this may require reperfusion (215). The marked swelling of endothelium probably contributes to reduced blood flow and further ischemia, since perfusion with hypertonic solutions maintains blood vessel patency and organ perfusion (39), and prevents ischemic damage (2). The cellular events responsible for cell swelling during reperfusion and the role of volume-regulatory transporters are unknown.

	CHANGES IN THE SET POINT FOR STEADY-STATE CELL VOLUME

TOP ABSTRACT INTRODUCTION DETERMINANTS OF CELL VOLUME REGULATORY VOLUME INCREASE REGULATORY VOLUME DECREASE CHANGES IN THE SET... SUMMARY REFERENCES

Steady-state volume varies over the life span of a cell, particularly during growth and differentiation. This obviously requires suppression of normal volume-regulatory mechanisms, but volume-regulatory transporters may also play a more direct role. Activation of NHE1 or NKCC1 in the absence of cell shrinkage could lead to cell enlargement, whereas volume-independent activation of RVD pathways could lead to cell shrinkage. Although the role of volume-regulatory transporters in adjusting steady-state cell volume has not been established, the rapid activation of NKCC1 (152, 153, 155, 156, and W. C. O'Neill, J. D. Klein, and S. T. Lamitina, unpublished observations) and NHE1 (60, 175) by growth factors lends credence to this concept. Addition of serum produces a rapid increase in cell volume (15% over 10 min) in endothelial cells, which is mediated by NKCC1 (W. C. O'Neill, J. D. Klein, and S. T. Lamitina, unpublished observations), and a similar response occurs in growth-stimulated fibroblasts (126, 152, 153). A more chronic stimulation of NKCC1 is also seen with growth factors (Ref. 19 and W. C. O'Neill, J. D. Klein, and S. T. Lamitina, unpublished observations) and contributes to the increase in cell volume during the growth cycle (19). This switch in NKCC1 from its normal mode of K⁺/K⁺ exchange at steady-state volume to a net influx mode implies that growth stimuli can modulate or overcome the putative inhibitory effect of intracellular Cl. Activation of NHE1 by growth factors also increases cell volume (51), which, like NKCC1, requires disinhibition of the transporter at steady-state volume, in this case by shifting the pH dependence to the right (51). The mechanism by which growth factors activate NHE1 and NKCC1 is not understood. Phosphorylation of NHE1 is increased (178), but a mutated antiporter lacking phosphorylation sites is still activated by mitogens, although to a lesser degree (210). Regulation of NKCC1 by growth factors has not been examined at the molecular level and, whatever the mechanism, it must overcome the normal inhibition of net influx (presumably due to intracellular Cl) to increase cell volume. Volume-regulatory transporters may also be responsible for cell enlargement during transformation. Fibroblasts expressing the ras oncogene were found to be 32% larger than control cells, with upregulation of both NKCC1 and NHE1 (95, 126), whereas cell enlargement accompanying cytomegalovirus infection is associated with enhanced activity of NHE1 (27).

Cell shrinkage is an early and dramatic event during apoptosis (6, 7, 149). The mechanism is unknown but appears to be biphasic. In dexamethasone-treated lymphoid cells, shrinkage begins in 12 h and progresses until chromatin condensation (36 h). Although this is associated with a net loss of K⁺, other cytoplasmic components are lost as well, since buoyant density does not change (7). The second phase of shrinkage after chromatin condensation is due to cellular fragmentation. Apoptotic shrinkage is inhibited by blocking K⁺ channels (6), but it is not known whether these are the normal volume-regulatory channels. Cell shrinkage also occurs during the latter stages of erythrocyte maturation and may be mediated by the K⁺-Cl cotransporter (97, 99), which is expressed predominantly in young erythrocytes (20, 62, 98, 142). Altered regulation of this transporter may induce additional shrinkage that contributes to sickling in hemoglobin SS disease (13).

The mechanism that couples cell volume to cell growth and differentiation is unknown. The macromolecular crowding theory of volume regulation could account for the regulation of volume-regulatory transporters by cell growth. As cell content of protein and other macromolecules changes, cell volume would be perceived as too small or too large, and RVI or RVD would be activated. This is consistent with the constant relative content of cell water during the cell cycle in HeLa cells (131) and the constant buoyant density during early apoptosis in lymphoid cells (7), suggesting a constant ratio of macromolecules to cell water. However, this scenario cannot account for the very early activation of NKCC1 and/or NHE1 and the resulting increase in cell volume produced by growth factors well before any increase in protein content. These data are more consistent with coupling in the other direction, with the increase in cell volume stimulating growth. Such coupling of cell growth to cell volume would ensure, for instance, that cells reach a certain size before mitosis. Consistent with this, inhibition of NKCC1 reduces growth responses in endothelial cells (155) and fibroblasts (19, 126, 139, 140, 154, 156) but not in all studies (1). Inhibition of NHE1 also blunts growth responses in a variety of cells (12, 60, 93, 164, 212), but the effect is variable and it is not clear whether it is mediated through changes in cell volume or intracellular pH (8). Because growth factors often activate both NHE1 and NKCC1, dual inhibition may be required to prevent cell swelling, possibly explaining the failure of amiloride or loop diuretics alone to completely inhibit growth. In fact, studies in fibroblasts showed that both were required to completely block DNA synthesis (154). These results suggest that activation of NHE1 and/or NKCC1 leads to an early increase in cell volume that may signal other aspects of cell growth. The role of cell shrinkage in apoptosis is even less clear. Hypertonic shrinkage is reported to both induce (14, 149) and inhibit (55) apoptosis, and prevention of cell shrinkage with K⁺ channel blockers does not prevent apoptosis (6).

The concept that cell volume controls cell growth is considerably strengthened by evidence that cell volume also regulates cell metabolism. Swelling (either hypotonic or isosmotic) of hepatocytes (109, 167, 194) increases protein synthesis and decreases protein degradation, whereas hypertonic shrinkage has the opposite effect (71) in liver and in other cells (162). Parallel effects are seen on glycogen synthesis and breakdown (71, 159). This process also represents a volume-regulatory mechanism, since osmotically active molecules are consumed during synthesis of protein and glycogen and released during proteolysis and gycogenolysis. In the liver, changes in cell volume serve to couple metabolism to nutrient uptake and hormone action (71). Nutrient uptake and insulin both cause cell swelling (the latter via stimulation of NKCC1 and NHE1), leading to an increase in the content of protein and glycogen. Prevention of cell swelling by inhibiting NKCC1 and NHE1 blocks this anabolic response. On the other hand, glucagon causes cell shrinkage (probably via opening of K⁺ channels), thereby inducing a catabolic response (proteolysis and glycogenolysis). There is in fact a linear relationship between hepatocyte volume and protein synthesis and proteolysis along which many anabolic and catabolic stimuli fall (Fig. 4). The anabolic and catabolic effects of cell volume may exist in skeletal muscle as well (106-108), but the role of cell volume in the action of insulin in target tissues other than liver is unclear (223). The signaling mechanisms by which cell volume controls cell metabolism and growth are unknown but could be related to known growth signaling cascades. For instance, cell swelling has been shown to activate tyrosine kinases, phosphorylate (and activate) MAP kinases (180, 204), and to increase c-jun phosphorylation (180), c-jun mRNA abundance (38), and phosphorylation of ribosomal protein S6 (109). To what extent any of these is responsible for the metabolic effects of cell volume is unknown.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Correlation between cell volume and proteolysis in perfused rat liver. Perturbations included anisosmotic media, hormones, channel and transporter blockers, and nutrients. [Adapted from Haussinger (71).]

	SUMMARY

TOP ABSTRACT INTRODUCTION DETERMINANTS OF CELL VOLUME REGULATORY VOLUME INCREASE REGULATORY VOLUME DECREASE CHANGES IN THE SET... SUMMARY REFERENCES

The purpose of this review is to dispel the notion that cell volume regulation in higher organisms is a curious but physiologically irrelevant phenomenon. Osmolarity of the milieu interieure is not as constant as is often assumed, and significant threats to cell volume still exist in an isosmotic environment under both physiological and pathological conditions. In addition to maintaining cell volume, volume-regulatory mechanisms are also employed to initiate changes in cell volume, as required for cell growth and differentiation. The close link between volume-regulatory transporters, cell volume, and cell growth and metabolism indicates that regulation of cell volume is a fundamental and important property of mammalian cells. The sensing and signaling mechanisms for cell volume regulation remain a mystery, and elucidation of these pathways will provide important new information on other aspects of cell physiology as well.