Optimal cellular function and viability rely on mechanisms that allow the cell to adapt to changes in physical or biochemical conditions found in the extracellular environment (Hochachka and Somero, 2002). Reference to these mechanisms as “cellular stress responses” reflects the fact that macromolecules and biochemical processes function optimally under very specific biochemical conditions (e.g., temperature, concentration, ionic strength, etc.). Deviation from these optimal biochemical conditions results in suboptimal function and thus cellular stress. Cellular stress responses mitigate the detrimental effect that changes in environmental conditions can have on cell function by providing the means by which the cell can homeostatically re-establish optimal function under otherwise sub-optimal conditions. The fundamental importance of these pathways is illustrated by functional genomic studies showing that the expression of a significant percentage of all genes within the yeast genome is altered in response to changes in various extracellular conditions (Gasch et al., 2000; Causton et al., 2001). Moreover, proteins involved in cellular stress responses represent the most highly conserved among all organisms (Kultz, 2003).
The physical and biochemical parameters of greatest relevance to eukaryotic cell function include nutrient availability, oxygenation, temperature, pH, and osmolality. Unicellular eukaryotic organisms such as yeast, as they exist in nature, are very likely exposed to extremes of each of these environmental conditions. However, cells of higher eukaryotes that maintain a homeostatically regulated internal environment (i.e., mammals) are exposed to very limited changes in temperature and osmolality under normal physiologic conditions. In mammals, the regulation of intravascular fluid volume and plasma electrolyte concentrations by the kidney results in a homeostatic balance in which plasma osmolality does not normally vary by more than 2–3% (Guyton and Hall, 1996). However, the critical function of the kidney to reabsorb water and thus concentrate urine is dependent on the physiologic generation of an extremely hyperosmotic environment within the renal medulla (Knepper and Rector, 1996). Osmotic stress in mammals has thus been thought to be of particular and perhaps unique relevance to cells located in the medulla of the kidney, as there has been no direct experimental evidence supporting the biologic significance of osmotic stress and the osmotic stress response to cells outside the kidney.
In this review, a new perspective on osmotic stress in mammals is presented that is based on recent studies of the osmoregulatory transcription factor, NFAT5/TonEBP (hereafter referred to as NFAT5). The unanticipated importance of the mammalian osmotic stress response pathway defined by NFAT5 to the function of lymphocytes in vivo during an immune response not only demonstrates the biologic importance of the mammalian osmotic stress response outside the kidney, but further suggests that dependence on an osmotic stress response is a function of not only extracellular osmotic stress, but also cellular metabolic activity. These insights highlight the critical yet largely unappreciated importance of intracellular water homeostasis to normal cell function in vivo.
Approximately 60% of body weight in man consists of water with two-thirds being present inside cells and most of the remaining water comprising the interstitial fluid outside cells that forms the extracellular environment (Guyton and Hall, 1996). Water has thus appropriately been referred to as the “universal solvent” of life because the myriad cellular biochemical processes that define biologic systems take place in the aqueous or water-soluble phase (Henderson, 1913; Hochachka and Somero, 2002). Consistent with this, thermodynamic and kinetic parameters typically used to describe biochemical reactions assume that the concentration of water, as a solvent, is in vast excess relative to the specific reactants and therefore has no effect on reaction equilibria and kinetics. However, distinct theoretical and experimental perspectives (Parsegian et al., 2000) indicate that the notion of water functioning solely as a solvent does not adequately reflect the importance of the chemical activity of water on biochemical reactions in vivo. Rather, water is more accurately considered to be not only the solvent but also a critical reactant, the concentration of which can influence a potentially wide spectrum of biochemical processes. These perspectives are supported experimentally by an “osmotic stress” approach that assesses the effect of altering the activity of water on a defined biochemical process through the use of solutes that lower the activity of water but do not otherwise interact with the system, thus allowing quantitation of the number of water molecules participating in that process (Rand, 2004). This approach has been applied to a variety of biochemical processes, including membrane channel dynamics, water soluble enzyme activity, and DNA–protein interactions, in each case demonstrating that reducing the activity of water itself can have a significant effect on the biochemical process in question, whether it be through effects on the preferential hydration of protein surfaces or on specific biochemical reactions that involve changes in hydration (Rand et al., 2000).
The importance of the concentration of water within the cell can also be framed in terms of the macromolecular crowding and confinement model, which considers the effects of excluded volume on biochemical processes within an the cell (Ellis, 2001; Minton, 2001). The environment that exists within a living cell is clearly not composed of a dilute aqueous solution of reactants. Rather, the complex intracellular environment is more accurately described as “crowded,” wherein macromolecules, in aggregate, exist at high concentration and occupy a significant fraction (∼20–30%) of total cell volume (Minton, 2001). Based on this analysis, the degree of crowding by macromolecules, representing excluded volume (i.e., the extent to which available volume is reduced), significantly affects biochemical processes. As a result, the functions of macromolecules within the cell are exquisitely sensitive to small alterations in available intracellular volume that might result from changes in the intracellular concentration of water (Ellis, 2001; Ho, 2003). Biologic systems have thus presumably evolved to function within well-defined ranges of concentration of not only reactants but also water, with the homeostatic regulation of the concentration of water within the cell thus representing a fundamentally critical process. The biologic implication of this concept is that small changes in the concentration of water within the cell have the potential to globally alter biochemical processes, and in so doing establish an intracellular environment detrimental to normal cell function and viability.
The movement of water between aqueous solutions that are separated by a semi-permeable membrane is driven by differences in the concentration of solute particles (also referred to as osmolytes) on either side of the membrane, with water flowing down its concentration gradient from the solution with a lower osmolyte concentration into the solution with a higher concentration. This physical biochemical process, referred to as osmosis, is of fundamental and universal significance to all biologic organisms that utilize a semi-permeable membrane to separate the intracellular milieu from the extracellular environment. The osmotic flow of water crosses the plasma membrane occurs by passive diffusion through water-permeable protein channels called aquaporins (King et al., 2004). In the living cell, changes in osmolyte concentration occurring either within or outside the cell can thus lead to the very rapid flow of water across the plasma membrane and marked alterations in cell volume, but only to the extent that the osmolytes are membrane-impermeable. Given that osmosis resulting from differences in the concentration of membrane-impermeable osmolytes affects the tonicity or volume of the cell, osmotic stress mediated by osmolytes that are predominantly membrane-impermeable is also referred to as tonic stress. Exposure of a cell to a hypertonic environment in which the osmolality outside the cell is greater than that present within the cell results in the osmotic efflux of water, a reduction in cell volume (i.e., cell shrinkage) and an increase in the concentration of all intracellular constituents. The resulting biochemical disequilibrium gives rise to a wide spectrum of deleterious effects on cell function and, in the event biochemical homeostasis is not restored, ultimately results in apoptotic cell death (Table 1). Given the universal role of water as both solvent and reactant, such pleiotropic effects of hypertonic stress on cell function, previously referred as “molecular mayhem” (Proft and Struhl, 2004), are not surprising. However, the effects described above have been observed under ex vivo experimental conditions that involve exposing cultured cell lines to hypertonic stress through the addition of osmolytes such as sodium chloride or sorbitol to the culture medium. The relevance of such experimental manipulations to the biology of mammalian cells outside the kidney, as they exist in vivo, remains unclear.
Table 1. Pleiotropic effects of hypertonic stress on cell function
Specific mechanisms are necessary to rapidly and dynamically compensate for the potentially lethal effects of osmotic stress. The cellular osmotic stress response represents the means by which cells adapt to osmotic stresses in order to optimize cell function and ensure survival. All cells respond to osmotic stress by altering the concentration of osmolytes within the cell in order to reduce or eliminate the difference in intracellular versus extracellular osmolyte concentration, thereby eliminating any change in intracellular water concentration and the concomitant change in cell volume that might occur by osmosis. An immediate cellular response to hyperotonic stress takes place within seconds and involves increases in the intracellular concentrations of charged ions such as potassium, sodium, and chloride that are mediated by pre-existing ion transport systems, including the Na+-K+-2Cl− co-transporter, the Na+/H+ exchanger, and the Cl−/HCO3− exchanger (McManus et al., 1995; Lang et al., 1998). The instantaneous reduction in cell volume due to the osmotic efflux of water induced by acute hypertonic stress is thus rapidly corrected by what is referred to as a regulatory volume increase. The use of charged ions as compensatory osmolytes, however, results in an increase in intracellular ionic strength that is not compatible with normal cell function. Therefore, the cellular response to hypertonic stress is also mediated by a delayed or adaptive response involving the transcriptional regulation of genes that allow ionic osmolytes to be replaced with non-ionic osmolytes (Burg et al., 1997; Handler and Kwon, 2001). Non-ionic osmolytes, also referred to as compatible or non-perturbing osmolytes, consist of small, non-charged organic molecules that do not affect cell function even at relatively high intracellular concentrations (Yancey et al., 1982). In mammalian cells, these include polyols such as sorbitol and myo-inositol, neutral amino acids, or their derivatives such as taurine or alanine, and methylamines such as betaine. Accumulation of compatible osmolytes within the cell thus maintains intracellular water homeostasis without otherwise impairing normal biochemical function.
The osmotic stress response pathway, best characterized in the yeast Saccharomyces cerevisiae by functional genetic and biochemical studies (Hohmann, 2002; Westfall et al., 2004), represents a paradigmatic signaling pathway that functions to sense the extracellular environment and translate that information into an appropriate program of gene expression (Fig. 1A). In yeast, this pathway links the putative osmotic sensor proteins Sln1 and Sho1 to the evolutionarily conserved mitogen-activated kinase Hog1, the homolog of the mammalian p38 mitogen-activated protein (MAP) kinase. Hog1 regulates the transcription factors Sko1, Hot1, and Msn2p/Msn4, which control genes involved in glycerol synthesis such as glycerol-3-phosphate dehydrogenase (GPD1), thus providing a means to increase the intracellular concentration of compatible osmolytes (Fig. 1B).
The mammalian adaptive osmotic stress response pathway (Fig. 1B) has only more recently been defined as a result of the identification of the mammalian osmoregulatory transcription factor NFAT5, originally designated tonicity enhancer binding protein or TonEBP (Miyakawa et al., 1999b). TonEBP was first identified in a functional genetic screen for proteins capable of binding to a specific tonicity-responsive DNA sequence present within the enhancer region of genes induced in response to hypertonic stress (Miyakawa et al., 1999b), and was also purified by DNA affinity chromatography and designated osmotic response element binding protein (Ko et al., 2000). TonEBP was also identified independently based on protein sequence similarity within the DNA binding domain to transcription factors of the dorsal/rel/NFκB/NFATc family and was designated NFAT5 (Lopez-Rodriguez et al., 1999), NFATL1 (Trama et al., 2000), and NFATz (Pan et al., 2000). Similar to the transcriptional response to osmotic stress in yeast, NFAT5 also regulates the expression of osmocompensatory genes (Fig. 1B). Although the list of NFAT5 target genes is likely incomplete, those genes identified to date function to either increase the intracellular concentration of compatible osmolytes or, in the case of HSP70, to presumably facilitate normal protein folding in an intracellular environment that has been altered as a result of the osmotic loss of intracellular water or an increase in intracellular ionic strength (Table 2). NFAT5 is itself regulated by hypertonicity at multiple levels, including RNA accumulation, protein stability, post-translational modification, and subcellular localization (Trama et al., 2000; Woo et al., 2000; Dahl et al., 2001; Lee et al., 2002). Phosphorylation of NFAT5 by kinases such as p38, fyn (Ko et al., 2002), protein kinase A (Ferraris et al., 2002), and ATM (Irarrazabal et al., 2004) regulate both cytosol-to-nuclear translocation and the function of transcriptional activation domains present within the protein (Lee et al., 2003). Direct interaction between NFAT5 and either PKA or ATM has been demonstrated (Ferraris et al., 2002; Irarrazabal et al., 2004), suggesting that NFAT5 may be a direct substrate for these kinases.
Unless otherwise noted, identification of a gene as an NFAT5/TonEBP target is based on a combination of genetic and cell-based functional assays as well as direct DNA binding studies.
Although ex vivo assays indicate that NFAT5/TonEBP directly regulates the TauT gene (Ito et al., 2004), NFAT5 null mice exhibit no reduction in TauT expression in the kidney in contrast to the significantly reduced expression of AR, SMIT, and BGT1 (Lopez-Rodriguez et al., 2004).
Loss of NFAT5 function in vivo results in the impaired induction of ATA2 mRNA in response to hypertonicity, thus providing functional genetic evidence that NFAT5 regulates ATA2 gene expression. However, direct regulation of the ATA2 gene by NFAT5 has not as yet been demonstrated. The ATA2 gene is thus most accurately considered a potential NFAT5-target gene.
Lopez-Rodriguez et al. (2004); Kasono et al. (2005)
The Pax2 transcription factor has also recently been shown to be induced by hypertonicity in renal medullary cell lines and in the kidney, where its expression varies with osmolality (Cai et al., 2005). Partial loss of Pax2 function mediated by small interfering RNA resulted in increased apoptotic cell death under hypertonic conditions (Cai et al., 2005). However, in contrast to the ubiquitous expression of NFAT5 mRNA (Trama et al., 2000), Pax2 exhibits a much more restricted pattern of expression (Dressler et al., 1990). Thus, the extent to which Pax2 contributes to the mammalian osmotic stress response remains unclear.
OSMOTIC STRESS IN MAMMALS
The critical importance of NFAT5 in mediating adaptation to hyperosmotic stress within the kidney was demonstrated by targeted disruption of the Nfat5 gene in the mouse, which resulted in a null allele (Lopez-Rodriguez et al., 2004). Although the majority of homozygous Nfat5−/− animals died late in gestation or perinatally, approximately 3% of the homozygous progeny survived past weaning. These animals exhibited marked atrophy of the renal medulla that was associated with impaired activation of several NFAT5-regulated osmoprotective genes (Lopez-Rodriguez et al., 2004). Consistent with this result, transgenic over-expression of a dominant inhibitory form of NFAT5 in the epithelial cells of the renal collecting tubules resulted in an impaired capacity to concentrate urine, leading to progressive hydronephrosis, that was also associated with defects in the expression of osmo-regulatory genes (Lam et al., 2004). Thus, as expected based on the unique physiology of the mammalian kidney, which is characterized by the generation of a hyperosmotic environment within the renal medulla during antidiuresis, the osmotic stress response pathway defined by the NFAT5 transcription factor is clearly essential for normal renal function.
In contrast to the increasingly well-defined and critical role for NFAT5 and the adaptive osmotic stress response pathway within the context of the unique physiology of the kidney, the function of NFAT5 and the significance of the osmotic stress response outside the kidney have only more recently become apparent. Initial studies demonstrated that while NFAT5 mRNA is ubiquitously expressed in all tissues, expression of NFAT5 protein was largely restricted to the thymus and activated lymphocytes (Trama et al., 2000). However, given that the osmolality of blood, and presumably that of interstitial tissue fluids, are maintained within a very narrow window of osmolality as described above, the function of NFAT5 and the relevance of the osmotic stress response in lymphocytes have remained unclear. Initial functional studies of role of NFAT5 in the immune system demonstrated that partial loss of NFAT5 function mediated by transgenic expression of a dominant inhibitory form of the NFAT5 protein resulted in lymphoid hypocellularity and impaired T lymphocyte growth and survival under conditions of hyperosmotic stress (Trama et al., 2002). This result was further substantiated in mice engineered by gene targeting to express a non-functional, dominantly inhibitory NFAT5 deletion mutant that exhibited partial loss of NFAT5 function in the heterozygous state and complete loss of function in the homozygous state (Go et al., 2004). Although homozygous animals die perinatally, consistent with the phenotype of animals that are homozygous for the null allele (Lopez-Rodriguez et al., 2004), heterozygous animals were viable and exhibited an immunologic phenotype very similar to that of animals expressing a dominant negative transgene, including an impaired T cell-dependent antibody response upon immunization (Go et al., 2004). Moreover, direct measurements demonstrated that normal lymphoid tissue osmolality was greater than that of blood (Go et al., 2004), providing evidence that hyperosmotic conditions are not limited to the renal medulla. Interestingly, while heterozygous animals exhibited no defect in renal function under conditions of normal hydration, water deprivation revealed an impaired ability to concentrate urine specifically in male mice (Park et al., 2004). To the extent that NFAT5 function is limited to the regulation of osmocompensatory gene expression (Table 1), functional genetic studies thus indicate that T cell development in the thymus and lymphocyte function during an immune response exhibit greater dependence on an functional osmotic stress response pathway than does renal function under conditions of normal (i.e., ad lib) hydration (Trama et al., 2002; Go et al., 2004).
CONTRIBUTIONS TO OSMOTIC STRESS IN VIVO
Further consideration of the critical role of the osmotic stress response to lymphocyte function provides interesting insights into not only osmotic stress associated with specific tissue microenvironments, but also the potential effects of cell metabolism on intracellular water homeostasis as a source of cellular osmotic stress. Direct measurement of tissue osmolality by vapor pressure osmometry indicates that, in contrast to the highly precise regulation of intravascular osmolality, tissue osmolality appears to be significantly more variable, with lymphoid and hepatic tissues being hyperosmolar as compared to blood, brain, or lung (Go et al., 2004). Osmotic stress under normal physiologic conditions outside the kidney thus appears to represent a previously unappreciated form of extracellular environmental stress. However, relative to the extremes of hyperosmotic stress to which cells of the renal medulla may be exposed during antidiuresis, the osmolality of tissues such as the spleen or thymus, as measured using vapor pressure osmometry, is only modestly elevated above that of blood (Go et al., 2004). One significant limitation in the measurement of tissue osmolality by vapor pressure osmometry is poor spatial resolution. Such measurements are incapable of measuring the osmolality present within a distinct tissue microenvironment such as the germinal center of a lymph node where extensive cell proliferation takes place. In addition, this approach does not permit distinction between intravascular and extravascular (interstitial) tissue fluid. Thus, the values for tissue osmolality obtained by vapor pressure osmometry may represent an underestimate of the actual osmolality to which cells may be exposed in vivo within specific tissue microenvironments.
Assuming that the extracellular osmotic stress to which lymphocytes are exposed in vivo is in fact limited, the expression of NFAT5 protein in thymus and activated lymphocytes (Trama et al., 2000) and the observed dependence of lymphocyte function on the NFAT5 osmotic stress response pathway (Trama et al., 2002; Go et al., 2004) would suggest that there are other mechanisms through which lymphocytes are exposed to osmotic stress in the thymus and during an immune response occurring within peripheral lymphoid tissues. Interestingly, immature thymocytes and activated lymphocytes undergo rapid cell proliferation, which is associated with a high rate of metabolic activity (Shortman et al., 1990; Frauwirth and Thompson, 2004). In addition, although most tissues express low or indetectable levels of the NFAT5 protein, NFAT5 protein is readily detectable in essentially all constitutively proliferating cell lines maintained in tissue culture (Lopez-Rodriguez et al., 1999; Miyakawa et al., 1999b; Ko et al., 2000; Trama et al., 2000). These observations suggest that cell proliferation or the state of cellular metabolic activity may determine the extent to which a cell is dependent on the osmotic stress response. As previously suggested (Ho, 2003), changes that occur within the cell as a result of active metabolism/cell division likely give rise to the same alterations in intracellular water homeostasis that also result from exposure to overt extracellular hypertonicity. Active cell metabolism and cell proliferation lead to a decrease in available intracellular volume as a result of the induction of macromolecular biosynthesis. In addition, biosynthesis involves the consumption of precursors such as amino acids and inositol, which function as intracellular osmolytes. These changes result in a decrease in the concentration of intracellular water and thus represent an osmotic stress that is functionally identical to exposing a cell to extracellular hypertonicity. Consistent with this hypothesis, culture of human fibroblasts under isotonic conditions lacking free amino acids, which leads to a reduction in the concentration of intracellular amino acids and a concomitant reduction in cell volume, results in activation of NFAT5, as measured by increased NFAT5 protein abundance and enhanced nuclear translocation (Franchi-Gazzola et al., 2001). Thus, the osmotic stress response pathway defined by NFAT5 appears to mediate cellular adaptation not only to overt extracellular osmotic stress, but also rather to any process that affects intracellular water homeostasis.
The critical importance of the NFAT5 osmotic stress response to both kidney function (Lam et al., 2004; Lopez-Rodriguez et al., 2004) and adaptive immunity (Trama et al., 2002; Go et al., 2004) highlights a concept in which functional osmotic stress represents the manifestation of alterations in intracellular water homeostasis occurring by distinct mechanisms within uniquely different physiologic contexts (Fig. 2). Altered intracellular water homeostasis in the kidney results predominantly from the marked extracellular hyperosmotic stress generated within the renal medulla that is necessary to permit reabsorption of water. Given that cells present within the renal medulla are typically non-proliferative, nor are they highly active metabolically, the contribution of intracellular osmotic stress is limited. In contrast, while extracellular osmotic stress within lymphoid tissues is limited, particularly in comparison to the renal medulla, altered intracellular water homeostasis and thus osmotic stress likely results from the high level of metabolic activity associated with rapid cell proliferation. Extracellular osmotic stress and cellular metabolic activity in combination may be thus considered to give rise to functional osmotic stress, which reflects the degree to which the concentration or activity of water within the cell no longer supports optimal biochemical functions. Thus, although the mechanisms contributing to functional osmotic stress in the kidney and in lymphoid tissues differ significantly, intracellular water homeostasis is altered to an extent that makes each tissue similarly dependent on the NFAT5 osmotic stress response.
As previously mentioned, while mice with partial loss of NFAT5 function exhibited impaired lymphocyte function (Trama et al., 2002; Go et al., 2004), impaired renal function was observed only under conditions of water deprivation (Park et al., 2004). The greater sensitivity of lymphocytes to partial loss of NFAT5 function thus suggests that a high rate of metabolic activity (e.g., cell division) may render a cell particularly sensitive to osmotic stress as compared to a cell that is in a metabolically more quiescent state. For example, during the course of an adaptive immune response quiescent T and B lymphocytes are activated to undergo massive macromolecular biosynthesis and rapid cell division within a very short period of time in order to effectively counter the very rapid rate of replication of infectious pathogen. Small changes in the intracellular concentration of water may be much more deleterious to a metabolically highly active cell that is critically dependent on the proper integration of a wide spectrum of interdependent metabolic and biosynthetic pathways, as compared to a cell that is metabolically less active. As a result, the significantly lower level of extracellular osmotic stress that may be present in tissues other than the kidney, when combined with potential osmotic stress inherent in active cell metabolism or cell division, may be sufficient to render a cell critically dependent on the induction of an adaptive osmotic stress response. Considerations such as these suggest that the osmotic stress response must be considered within the context of intracellular water homeostasis, which is influenced not only by the extracellular environment present within a given tissue, but also by cellular metabolic activity.
Unexpected observations often reveal important new insights. The critical importance of the NFAT5 osmotic stress response pathway to the adaptive immune response in vivo has given rise to interesting new insights concerning intracellular water homeostasis in mammals and the potential contribution of both the extracellular tissue microenvironment and cell-intrinsic metabolic activity to osmotic stress. The availability of animal models and cell lines deficient in NFAT5 activity, and thus defective in the osmotic stress response defined by NFAT5, should provide useful tools to further explore the importance of intracellular water homeostasis in mammals within various biologic contexts. Moreover, realization of the fundamental importance of intracellular water homeostasis to mammalian cell biology will no doubt provide the impetus to further define the molecular basis of the mammalian osmotic stress response pathway. Finally, given the fundamental importance of intracellular water homeostasis to basic cell biology, further consideration of osmotic stress and the mammalian osmotic stress response within the context of pathophysiologic conditions such as cancer or inflammation may lead to novel therapeutic approaches to treat human diseases.