Challenges and intriguing problems in comparative renal physiology

The August Krogh principle that for every physiological problem there is an animal uniquely suited by nature to study it has been cited many times to emphasize the importance of the comparative approach to understanding specific physiological functions. Perhaps nowhere has this been more apparent than in the area of vertebrate renal physiology. The classical historical example,known to most students of comparative and renal physiology, is the definitive demonstration of renal tubular secretion by Marshall and his colleagues using the aglomerular goosefish Lophius piscatorius(Marshall and Vickers, 1923; Marshall and Grafflin, 1928). Other examples include insight into tubule fluid secretion viastudies with these aglomerular and other fishes(Beyenbach, 2004) and the understanding of the electrophysiology of tubular transport processes through the exploitation of the rather large amphibian renal tubules(Dantzler, 1988). In this brief commentary, I suggest a few other intriguing problems in renal physiology for which the comparative approach can provide important insights.

Knowledge of the exact pressures and the capillary surface area available for ultrafiltration in the renal glomerulus is essential to a quantitative description of this process. As indicated below, such knowledge cannot be obtained for any known mammalian glomerulus but could be obtained for some avian glomeruli.

In a classical anatomical study of the renal glomerulus, William Bowman(1842) examined kidneys from mammals (including humans), birds, reptiles, amphibians and fishes in order to evaluate the structural relationships between them. Differences between mammalian and avian glomerular capillary networks were evident in Bowman's drawings and are shown even more clearly in more recent studies of structure(Spinelli et al., 1972; Casotti and Braun, 1995). In mammals, the glomerular capillaries form a complex network of thoroughfare channels that anastomose freely (Spinelli et al., 1972), whereas in birds a single unbranched glomerular capillary coils around the periphery of the renal corpuscle(Casotti and Braun, 1995). Indeed, in the small, superficial loopless avian nephrons (sometimes called`reptilian-type' nephrons), there may be only a single capillary loop(Fig. 1). This simple structure offers a unique opportunity to examine pressure profiles in a high-pressure glomerular filtration system as described below.

In this equation, l is the length of the cylinder, x is the distance along the cylinder, A is the glomerular capillary area available for filtration and Lp is hydraulic conductivity. Since random micropuncture measurements of the glomerular capillary hydrostatic pressure (P_GC) in mammalian glomeruli show little variation (Brenner et al.,1971), P_GC apparently decreases little along the length of the capillaries. Although the small decrease in P_GC along the length of the capillaries is essential for flow to occur, it can be ignored in the simplest treatments of the above relationship and P_GC can be considered to be constant throughout the length of the cylinder. The transmural hydrostatic pressure difference is then equal to the difference between P_GC and the hydrostatic pressure in the space outside the capillaries (Bowman's space), which is continuous with the lumen of the proximal tubule. This pressure in Bowman's space (P_BS) is also considered to be constant. Because the protein that is filtered and enters Bowman's space is negligible, it can be ignored and the colloid osmotic pressure of the plasma(P_COP) alone can be considered the net colloid osmotic pressure that opposes the hydrostatic pressure driving filtration. Although P_GC apparently decreases only slightly along the length of the capillaries, the P_COP rises markedly as filtration occurs. This results from the fact that the protein concentration in the capillaries increases reciprocally with the fraction of water remaining in them and P_COP increases as an exponential function of the increasing protein concentration (Landis and Pappenheimer, 1963). Filtration will occur only as long as the net outwardly directed hydrostatic pressure(P_GC-P_BS) exceeds the opposing P_COP. The difference between these two opposing pressures is the net ultrafiltration pressure (P_UF).

Models based on mammalian micropuncture studies(Brenner et al., 1972) indicate that if filtration equilibrium is reached along the length of the capillaries,the glomerular filtration rate is proportional to plasma flow along the capillaries. Changes in plasma flow under these circumstances alter the shape of the rising P_COP curve. However, the complexity of the capillary network in the mammalian glomerulus prevents determination of the precise site of each pressure measurement or of the exact profile of the P_COP curves along the length of the capillaries from the afferent to the efferent end. No such pressure measurements have yet been made in avian glomerular capillaries, but the simplicity of the single loop of glomerular capillary in the superficial loopless nephrons(Fig. 1) should permit direct determination of the pressure profiles in glomeruli from any species that are accessible to micropuncture. This will take some exploration and careful,meticulous work, but it should be possible for the interested investigator. Birds, like mammals, are homeotherms with high arterial pressures. Therefore,it appears likely that such measurements, which cannot be made in mammals,would provide very important information concerning glomerular function in both vertebrate classes.

The integrated rate of glomerular ultrafiltration is determined not only by P_UF but also by the glomerular capillary area available for filtration (A in the above equation) and on the water permeability of the capillary wall (hydraulic conductivity, Lp). With the apparent exception of the single loop of glomerular capillary in the avian superficial loopless nephrons, the glomerular capillaries in other avian nephrons are moderately complex and in mammals form a highly complex branching network. The total area available for filtration in these complex networks is a function of the length, diameter, and number of the capillary branches. Moreover, the specific morphology of the capillary network, including the capillary dimensions and branching pattern, as well as the microrheological properties of the blood, determine the distribution of blood flow and, thus,the area used for filtration. Because the area available for filtration is unknown in such complex glomerular networks, A and Lp are usually treated together in the equation above as the ultrafiltration coefficient, K_f. However, in superficial avian glomeruli found accessible to micropuncture, it should be possible not only to make the pressure measurements discussed above but also to determine the area available for filtration in this single loop and, thus, to calculate a specific value for Lp, at least for one type of high pressure glomerulus. This information, together with the pressure profiles in the same glomeruli, would be enormously valuable for obtaining a quantitative understanding of glomerular ultrafiltration.

In vivo and in vitro micropuncture and microperfusion studies of proximal tubules in nephrons of fishes (at least lampreys; Logan et al., 1980),amphibians (Windhager et al.,1959), reptiles (Dantzler and Bentley, 1978a; Stolte et al.,1977), birds (Laverty and Dantzler, 1982), and mammals(Gottschalk and Mylle, 1959; Walker et al., 1941) have indicated that reabsorption of water and solutes occurs at approximately equivalent rates, i.e. fluid reabsorption is approximately isosmotic. Furthermore, these studies have demonstrated that in mammals and amphibians,at least, the overall process is dependent on the transport of sodium. The mechanism by which water and solutes are transported isosmotically across epithelia has always posed an interesting intellectual challenge. However,studies of the above type have now indicated that in mammalian and amphibian proximal tubules the reabsorptive process actually involves the establishment of small transepithelial osmotic gradients, although all the details of the coupling are not fully understood (Braun and Dantzler, 1997). In other words, the reabsorptive process in the proximal tubules of these vertebrates is not truly isosmotic as originally thought.

In reptiles, however, the reabsorptive process for solutes and water in the proximal tubule appears to be quite different from that in mammals and amphibians. In fact, studies of this process in reptiles have raised a number of intriguing, unanswered questions about the coupling of solute and water transport across epithelia, the answers to which might materially advance our general understanding of transepithelial transport. However, the answers to these questions will only be obtained through further studies in this vertebrate class.

The initial micropuncture and microperfusion studies on lizards and snakes,mentioned above (Dantzler and Bentley,1978a; Stolte et al.,1977), indicated that water and sodium are (or at least can be)reabsorbed at osmotically equivalent rates, but they did not demonstrate that water reabsorption is dependent on sodium reabsorption. In fact, studies with isolated, perfused garter snake (Thamnophis sp.) proximal tubules indicate that neither sodium nor chloride is essential for normal fluid reabsorption (Dantzler and Bentley,1978a). These studies involved replacement of sodium or chloride or both with a number of substances commonly used as substitutes for them[e.g. choline, tetramethylammonium (TMA), or lithium for sodium; methyl sulfate for chloride; or sucrose for both sodium and chloride]. The critical results of these studies are summarized in Fig. 2. When sodium in the perfusate is replaced with choline, net fluid reabsorption nearly ceases(Fig. 2). However, when sodium in the bathing medium is also replaced with choline so that the composition of the two solutions is identical, net fluid reabsorption returns to the control level (Fig. 2). The results are the same when sodium is replaced with TMA, when chloride is replaced with methyl sulfate, and when both sodium and chloride are replaced with sucrose(Fig. 2). However, net fluid reabsorption does not change when sodium is replaced with lithium in the perfusate or in both the perfusate and bathing medium(Fig. 2). Fluid reabsorption at the control rates, regardless of the composition of the perfusate and bathing medium, is isosmotic (at least within the margin of error of the cryoscopic measurement system used) and can be at least partially inhibited by cyanide and cold (Dantzler and Bentley,1978a). However, even when sodium is the primary cation in the media, fluid reabsorption is not inhibited by removal of potassium from the bathing medium or by addition of ouabain or other cardiac glycosides (highly specific inhibitors of Na⁺-K⁺-ATPase) to it (Dantzler and Bentley, 1978a,b). This latter observation apparently reflects the lack of sensitivity of the Na⁺-K⁺-ATPase in these reptilian tubules to cardiac glycosides (Dantzler, 1972). Moreover, even with sodium as the primary cation, net fluid reabsorption is independent of the buffer system (bicarbonate, phosphate or Tris) used, but it is reduced about 18-25% by the removal of colloid from the peritubular fluid(Dantzler and Bentley,1978a).

These studies on snake renal tubules are fascinating because, in contrast to studies on all other vertebrate species, they indicate that isosmotic fluid reabsorption in the proximal tubule can occur in the absence of sodium and chloride. However, they do not provide any information on the mechanism involved in such transepithelial transport. One suggestion has been provided by quantitative structural studies on these isolated, perfused tubules(Dantzler et al., 1986). Within a few minutes of the replacement of sodium in both the perfusate and bathing medium with choline, the cells double in size and the intercellular spaces nearly quintuple. During this time, the areas of the lateral and apical cell membranes approximately double, but their surface densities remain constant. This observation means that although the larger cells in the absence of sodium have proportionally larger surface areas, the volume-to-surface area ratio remains constant. The source of the membrane utilized in the rapid increase in membrane area (perhaps intracellular vesicles) is unknown, but certainly merits examination. In any case, these changes in membrane surface area are correlated with the maintenance of the control rate of net fluid reabsorption when sodium is replaced by choline(Dantzler et al., 1986),suggesting that they may be related to the mechanism involved in such fluid reabsorption. For example, they may permit a small, previously insignificant driving force (perhaps even the osmotic pressure gradient across the epithelium generated by the colloid in the bathing medium) to be sufficient to produce a control level of net fluid reabsorption(Dantzler et al., 1986). However, there is no direct evidence for such an occurrence. It is merely speculation. Moreover, it does not provide an intellectually satisfying explanation for the maintenance of such a constant rate of net fluid reabsorption in the presence and absence of sodium. The process remains a mystery that should intrigue the inquisitive investigator. Devising experimental strategies to determine the mechanism(s) and to measure the actual force(s) involved will, of course, test the ingenuity of such an investigator. However, the solution of the problem, which could provide unique insight into the coupled transport of solutes and water across epithelia, is certainly exciting enough to warrant whatever effort it takes.

Urates form the major end product of nitrogen metabolism excreted by the kidneys of birds, all reptiles except some chelonians from a freshwater aqueous environment, and a few primarily terrestrial amphibians(Dantzler, 1988). Uricotelism is one physiological adaptation that permits animals that cannot produce urine more concentrated than the plasma (amphibians and reptiles) or only modestly more concentrated than the plasma (birds) to occupy relatively arid terrestrial habitats (Dantzler,1988). Uricotelism also plays a role in renal excretion of organic cations and regulation of acid-base balance (Dantzler, 1988, 1995). Urates are freely filtered by the nephrons of uricotelic vertebrates, but tubular secretion accounts for most of the urinary excretion(Dantzler, 1988).

Although urea is the major end product of nitrogen metabolism excreted by the kidneys of all mammals, a small amount of uric acid is produced and urates do appear in the urine (Roch-Ramel,1979). In most mammals, filtered urate undergoes net reabsorption by the renal tubules (Roch-Ramel,1979). However, in a few mammalian species (pigs, rabbits), net secretion of urate by the renal tubules commonly occurs(Chonko, 1980; Roch-Ramel et al., 1980). In humans, tubular reabsorption of filtered urate predominates, but tubular secretion also occurs (Roch-Ramel,1979). Understanding the mechanism of urate secretion by the renal tubules of reptiles and birds, where it dominates, may aid in the understanding of the mechanism of urate secretion in mammals such as humans,where it normally plays a minor role in urate excretion. Most of the physiological information on the tubule secretory process has come from studies with isolated perfused snake and chicken renal tubules(Brokl et al., 1994; Dantzler, 1973, 1976; Randle and Dantzler, 1973) and is summarized in the descriptive models shown in Figs 3 and 4.

In snakes and chickens, net transepithelial secretion occurs in the proximal tubule only (Brokl et al.,1994; Dantzler, 1973, 1976). Secretion occurs equally in all segments of snake proximal tubules, but possible variation along the proximal tubules has not yet been determined in chickens(Brokl et al., 1994; Dantzler, 1976). There is no evidence of net transepithelial reabsorption of urate in snake and chicken tubules; however, in contrast to the transepithelial transport of other organic anions in these tubules, there is a substantial passive reabsorptive backflux that appears to move between the cells (Figs 3 and 4; Brokl et al., 1994; Dantzler, 1973, 1974, 1976). In snake tubules and probably also chicken tubules, the magnitude of this backflux varies directly with the luminal perfusion rate, resulting in net secretion that also varies directly with perfusion rate (Dantzler,1973; Brokl et al.,1994). The overall pattern of transepithelial transport in snake renal tubules is remarkably similar to that observed for urate secretion in isolated perfused rabbit proximal renal tubules(Chonko, 1980; Dantzler, 1973). The similarity of these patterns has implications for physiological and possibly pathological function in rabbit tubules and the molecular basis of transport in snake tubules (see paragraphs discussing mammalian pathology and mechanism of basolateral urate transport below).

Net transepithelial secretory transport of urate is a saturable process,but the K_m for this process in snake tubules (∼150μmol l^-1) is well below the normal plasma urate concentration in these animals (∼400-500 μmol l^-1; Dantzler, 1982). This observation indicates that the transepithelial secretory process is normally about 60-70% saturated and suggests that changes in plasma urate concentration have little effect on net secretion. This probability suggests, in turn, that changes in the rate of flow of filtrate (or perfusate) along the proximal tubule may be particularly important in determining net transepithelial secretion and final urinary excretion(Dantzler, 1995). It appears very likely that flow rate plays the same role in urate secretion by chicken tubules, although this has yet to be shown definitively.

In both snake and chicken tubules during the process of transepithelial secretion, urate is transported into the cells at the basolateral membrane against an electrochemical gradient and then moves from the cells to the lumen down an electrochemical gradient (Figs 3 and 4; Brokl et al., 1994; Dantzler, 1973). These are, of course, the same basic steps observed for the transepithelial secretion of other organic anions in these and other vertebrate species(Dantzler, 1988).

However, in snake renal tubules the apparent permeability of the basolateral membrane to urate is much greater than that of the luminal membrane (Fig. 3), just the opposite of the situation for other organic anions (Dantzler, 1973, 1974, 1976). The arrangement of these membrane permeabilities provides an inefficient system for net transport to go in the secretory direction and, considered in isolation, appears to be highly non-physiological. Nevertheless, net secretion does occur in perfused nephrons. Moreover, this high permeability of the basolateral membrane may account, in part, for the much lower intracellular concentration of urate established at steady-state in nonperfused than in perfused snake tubules(Brokl et al., 1994; Dantzler, 1973), a situation just the opposite of what is observed for other organic anions(Brokl et al., 1994; Dantzler, 1973, 1974, 1976). In addition, the apparently higher rate of urate transport into the cells at the basolateral membrane in perfused than in non-perfused tubules may be related to a luminal anion that is absorbed and exchanged for urate at the basolateral membrane(see discussion of the basolateral transport mechanism below).

In reptiles and birds, what appear to be non-physiological observations that (1) urate transport into the cells at the basolateral membrane is dependent on perfusion of the lumen with filtrate (or an appropriate substitute), (2) the basolateral membrane has a higher passive permeability to urate than the luminal membrane, and (3) substantial passive paracellular urate reabsorption can occur, all may actually have substantial physiological significance when considered in the context of the intact kidney. The small loopless (reptilian-type) nephrons in the avian kidney and all nephrons in the reptilian kidney can filter intermittently. At the time when one of these nephrons is not filtering, these three apparently non-physiological characteristics of urate transport may actually perform the important physiological function of preventing accumulation and precipitation of poorly soluble urate in the tubule lumen or cells, thereby also preventing blockage of the tubule lumen or damage to the cell structure. Similarly, if net urate secretion by rabbit proximal tubules also includes a balance between an energy-requiring secretory flux and a substantial passive backflux, as seems likely (Chonko, 1980), this backflux may prevent accumulation and precipitation of urate in the lumens of nephrons that stop filtering during pre-renal acute renal failure. This might occur at the same time that secretion of hippurate by the general organic anion secretory system (see below for brief discussion of organic anion secretory system) is actually producing fluid secretion and helping to maintain some flow in the lumen, as suggested by Grantham and Wallace(2002). A similar pattern might also occur under these circumstances in human tubules, although there is no information on this process.

The mechanism by which urate is transported into the cells against an electrochemical gradient at the basolateral membrane has not been completely determined for either reptilian or avian nephrons and merits intensive study,especially since it may provide insights into the secretory process in mammalian nephrons, including human nephrons. In reptilian nephrons, the process is completely separate from the general physiological process for other organic anions (Dantzler,1988). However, inhibition of the basolateral urate transport step by anion-exchange inhibitors (e.g. the disulfonic stilbene SITS; Mukherjee and Dantzler, 1995) suggests that the process may involve anion-exchange transport(Fig. 3). Although this seems likely, the possible counter anion for which exchange of urate might occur has not been identified, despite attempts to promote exchange with numerous mono-,di- and tricarboxylates, by analogy with the general organic anion transport system (see below; Y. K. Kim and W. H. Dantzler, unpublished observations). Thus, the details of this process in the reptilian nephron remain a mystery.

In avian nephrons, urate transport into the cells against an electrochemical gradient at the basolateral membrane appears to involve both a process independent of that for other organic anions (as in the reptilian nephrons) and the accepted process for other organic anions (generally monovalent organic compounds of molecular mass less than 0.4 kDa, e.g. p-aminohippurate (PAH; Fig. 4). The accepted physiological process for other organic anions is a tertiary active transport system, the terminal step of which involves the transport of the organic anion into the cells against its electrochemical gradient in exchange for α-ketoglutarate moving out of the cells down its electrochemical gradient (Fig. 4). The outwardly directed gradient for α-ketoglutarate is maintained in turn by metabolism and by α-ketoglutarate transport into the cells against its electrochemical gradient via coupling to the movement of sodium into the cells down its electrochemical gradient(sodium-dicarboxylate cotransport; Fig. 4). Finally, the inwardly directed gradient for sodium is maintained by its transport out of the cells via basolateral Na⁺-K⁺-ATPase, the primary, energy requiring step in the tertiary active transport process (Fig. 4). This general process was first proposed by Burckhardt and Pritchard, essentially simultaneously, from data obtained with basolateral membrane vesicles from rat kidney (Pritchard, 1987, 1988; Shimada et al., 1987). It was then demonstrated to function this way in intact rabbit, chicken, snake and flounder renal tubules (Brokl et al.,1994; Chatsudthipong and Dantzler, 1991, 1992; Dantzler et al., 1995; Miller and Pritchard, 1991). It appears quite possible that urate may be transported by an independent system in one portion of a given avian proximal tubule and by the common organic anion transport system in another portion of the same tubule, but the difficulty of teasing avian tubules from fresh tissue has not yet permitted this to be determined. It is also possible that one transport system dominates in one group of nephrons and the other transport system in another group of nephrons. Internephron heterogeneity, depending on the distribution of the transporters, could be important for urate secretion during intermittent nephron function. In any case, the possibilities of intra- and internephron heterogeneity for the urate transport systems in the avian kidney certainly merit examination.

Nothing is known about the molecular basis of either the pure basolateral urate transport process or the general basolateral transport process for other organic anions in either reptiles or birds. However, in all mammals studied,the terminal step in the basolateral tertiary active transport system(α-ketoglutarate/organic anion exchanger) involves two members of the Organic Anion Transporter (OAT) family, OAT1 and OAT3(Wright and Dantzler, 2004). These two transporters have different specificities for organic anion substrates and these specificities vary for the same transporter cloned from different species. Of particular interest for the present discussion, rabbit OAT3, when expressed in a heterologous cell line, can transport urate in exchange for α-ketoglutarate (X. Zhang and A. Bahn, unpublished observation; noted in Wright and Dantzler,2004) but has almost no affinity for PAH(Wright and Dantzler, 2004). OAT3 may be the critical basolateral transporter in the urate secretory system in rabbit proximal tubules, although this has yet to be studied directly. If this turns out to be the case, it is possible that orthologs of OAT3 are the critical basolateral transporters for urate in reptiles and birds, given the physiological similarities between the rabbit urate secretory system and the reptilian (and probably avian) urate secretory systems noted above(Chonko, 1980; Dantzler, 1973, 1976). This is an area in which appropriate molecular and physiological studies could provide significant information about the details of the urate secretory process in reptiles, birds, and mammals.

Much less is known about the urate transport step at the luminal membrane than at the basolateral membrane, although movement from the cells into the lumen is very clearly down an electrochemical gradient (Figs 3 and 4). Absolutely nothing is known about this transport step in avian nephrons. It has yet to be studied. Moreover, those studies that have been performed on reptilian nephrons present a very confusing picture of this process. In isolated, perfused snake proximal renal tubules, transport of radiolabeled urate from the cells to the lumen is not affected in any way by the addition of probenecid (inhibitor of organic anion transport), SITS or unlabeled urate itself to the lumen, thus providing no evidence for mediated transport(Dantzler and Bentley, 1979; Mukherjee and Dantzler, 1985). Studies with brush border membrane vesicles (BBMV) from snake renal tubules also provided no evidence for mediated transport(Benyajati and Dantzler, 1988). These data, in conjunction with the relatively low urate permeability of the luminal membrane (Fig. 3), are compatible with completely passive diffusion of urate from cells to lumen. However, it is very difficult to reconcile the rather large flux of urate across the luminal membrane during net transepithelial secretion with simple passive diffusion. Perhaps urate is sequestered in vesicles during the transcellular portion of the secretory process and then extruded into the lumen by exocytosis, as suggested by Miller and his colleagues for the transport of fluorescein in teleost and some other tubules(Miller and Pritchard, 1994; Miller et al., 1993). But exocytosis does not appear compatible with the observed rate of transport of urate. Determining the nature of this transport step is going to require unusual and creative approaches, but it is critical to the understanding of the transepithelial transport and excretion of this very important anion.

Comparative studies have played and continue to play a critical role in our understanding of renal function and of the way in which diverse species thrive in all environments. The problems and challenges discussed above are ones that I believe are particularly intriguing, which will yield information of fundamental importance, and that cry out for innovative approaches by creative and inquisitive young investigators.

The writing of this commentary and the personal research discussed herein were supported in part by National Science Foundation Grant IBN9814448 and preceding grants.