Portal de Periódicos da CAPES

The Juxtaglomerular Apparatus

2003; American Society of Nephrology; Volume: 14; Issue: 6 Linguagem: Inglês

10.1097/01.asn.0000069221.69551.30

1533-3450

Jürgen Schnermann,

Diet and metabolism studies

Homer Smith died in 1962, the year before I started my career in the Department of Physiology of the University of Goettingen. At that time, his original work was still well known and widely quoted. Several decades later, this has of course changed, and the newer generation of renal physiologists and nephrologists is often unaware of who Homer Smith was and what he did. This is not totally surprising given the time elapsed and the short half-life of scientific fame and the fact that the electronic library, which we more and more rely on exclusively, does not go back beyond 1965 or so. I would therefore like to begin my presentation by honoring the man who is commemorated by this award. Born in 1895 in Denver, he received an A.B. from the University of Denver in 1917. He joined the armed forces and was transferred to the Chemical Warfare Station in Washington D.C., where he studied the biologic effects of gases, specifically of mustard gas, under the supervision of E.K. Marshall, who later provided proof for tubular secretion of organic compounds and its saturable nature. After the end of the war, Homer Smith enrolled in graduate studies at Johns Hopkins University and received the D.Sc. degree in chemistry in 1921. His early scientific work revolved around the chemotherapeutic value of arsenic compounds and the chemistry of secondary valence. After 2 years in the laboratory of Walter Cannon at Harvard University (1923–1925), he turned to studies of marine biology and of the physiology of fish. Around 1930, he began to develop and apply quantitative methods to studying the function of the mammalian kidney. He was appointed chairman of the Department of Physiology of the University of Virginia at the age of 30 yr. In 1928, he became chairman of the Physiological Laboratories at New York University College of Medicine, a position that he held until 1961. In the late 1920s, when Homer Smith started his work on the kidney, renal physiology as a field was not particularly advanced; in fact, it was considerably underdeveloped. This is reflected, for example, in the following statements of Alfred Richards, one of the other greats of kidney physiology in that same time: The literature of investigation of the kidney, since Bowman, contains no great illuminating discoveries comparable to those which are so conspicuous in other fields. Many of the questions of kidney function which are now being actively debated by physiologists are practically identical with those which were subjects of controversy seventy-five years ago…. Is the Malpighian body with its contained glomerulus a filter or is it a “secretory” structure? Is the epithelium which composes the tubule walls capable of secreting substances from blood into lumen of tubule, or is its task that of permitting or enabling restoration to the blood of substances lost to it in passage through the glomerulus? (41). These were very basic questions indeed! Thirty years later, by the end of the “Homer Smith era,” the function of the kidney at the organ level was adequately described in quantitative terms of filtration, perfusion, absorption, and secretion, through his work and through the work of others that he had stimulated. One could argue, as Robert Berliner has done, that his summary and interpretation of the available knowledge may be his most important contribution to the field (5). He wrote three textbooks on renal physiology and pathophysiology of which the second, The Kidney, was enormously influential, because it was the first modern and correct summary of how the kidney works in quantitative terms (58,61,63). His scientific work is summarized in a little over 100 original publications — papers used to be fewer and longer then. His major scientific contribution was to co-invent, simultaneously with Richards (42), the clearance methodology for measuring GFR and RBF, and then apply it in a wide array of conditions, mostly in clinical conditions, since Homer Smith had long-standing and intense collaborations with his clinical colleagues at New York University (14,23,54,64). Among his other contributions are the development of the concept of free water clearance with urinary dilution being a consequence of NaCl absorption across a water impermeable segment, the distal nephron, the discovery that urea back-diffusion accounts for its relatively low clearance and its dependence on urine flow (10), the notion that bulk absorption of the majority of filtered water occurs in the proximal tubule by active absorption of Na followed by passive absorption of water, the concept of “obligatory” reabsorption in the proximal tubule and “facultative” reabsorption in the distal tubule, and the introduction of the concept of glomerulotubular balance. Homer Smith was a much sought-after speaker as judged from the number of talks he was invited to give, and it is from the transcripts of these talks that one can glean a more personal view of who Homer Smith was. The titles of some of these presentations indicate his interests, titles such as “Error in Physiology” (1936), “Plato and Clementine” (1947), “Objectives and Objectivity in Science” (1949), “Agnosticism versus Atheism” (1956), “De Urina” (1958), and “The Biology of Consciousness” (1959) among others. It is apparent that he was broadly interested; he was a humanist, an evolutionary biologist, a paleontologist, a philosopher, and a musician, somewhat of a renaissance man. Throughout his life, he struggled with the place of man in nature and therefore with the interface of science, philosophy, and religious beliefs. He did not believe in the supernatural or in creationism, and was skeptical about all established religions. In his words: All human history reveals that transcendental metaphysics is not only futile but dangerous. Those who have, frequently by dishonest means, foisted upon the nai[Combining Diaeresis]ve and gullible their own unsupported speculations have served to retard man’s self realization more than any other misfortune that has ever befallen him. History also reveals that man does not need any brand of transcendental metaphysics — his lasting contentments and achievements he has found wholly within the frame of reference that takes things as they are in the here and now (62). In addition to his textbooks, he wrote four other books, Kamongo, or The Lungfish and the Padre (Viking Press, 1932), The End of Illusion (Harper, 1935), Man and His Gods (Little Brown, 1952), and From Fish to Philosopher (Little Brown, 1953). Particularly pleasant to read is From Fish to Philosopher. His enthusiasm and understanding for comparative physiology has substantial resonance in this era in which we learn more and more about the extent of evolutionary conservation both at the genomic and functional level. The Juxtaglomerular Apparatus I would like to turn to the topic of this overview, the mechanisms of juxtaglomerular cell communication, with these quotes of Homer Smith that are as true and puzzling now as they were then: Examining the pattern of the human kidney, we must not be surprised to find that it is far from a perfect organ. In fact, it is in many respects grossly inefficient. It begins its task by pouring some 125 cc of water into the tubules each minute, demanding for this extravagant filtration one quarter of all the blood put out by the heart. Out of this stream of water, 99 per cent must be reabsorbed again. This circuitous method of operation is peculiar, to say the least…. In consequence of the circuitous pattern of the filtration and reabsorption of water, nearly half a pound of glucose and over three pounds of sodium chloride per day, not to mention quantities of phosphate, amino acids and other substances, must be saved from being lost in the urine by being reabsorbed from the tubular stream. There is enough waste motion here to bankrupt any economic system — other than a natural one… (59). The astounding conclusion that over three pounds of NaCl must be filtered and absorbed per day was a derivative of the recognition of the nature of glomerular filtrate formation as an ultrafiltration process and the determination of the GFR magnitude by the inulin clearance. Avid absorption takes place along the proximal part of the nephron by an array of transporters, and their activity reduces the amount of Na to an apparently trivial number by the time the collecting duct is reached. There is good evidence that most or all of the regulation of Na absorption that is necessary to match salt intake and excretion over the physiologic range occurs along the collecting duct through variations in Na absorption across the highly regulated epithelial Na channel ENaC. The critical role of ENaC-dependent Na absorption is highlighted by observations in patients and mice, in which ENaC deficiency is associated with severe Na loss and volume depletion, that in the case of the ENaC knockout mice, regardless of ENaC subunit, lead to rapid postnatal death, at least to a substantial extent by volume loss and circulatory collapse (1,20,33). The same outcome is seen in mice that are deficient in the mineralocorticoid receptor, and therefore do not respond to aldosterone, the most important regulator of ENaC (4). Conversely, ENaC overactivity as in Liddle syndrome causes severe volume-dependent arterial hypertension with all its dire consequences (17,55). Because the capacity of the collecting duct is limited, one could envisage unwanted Na loss by collecting duct overloading with excess Na when there is an increase in GFR or when there is a decrease in Na absorption in the proximal nephron. It is now clear that such changes in NaCl delivery are sensed in the juxtaglomerular region of the nephron and are used as a signal that feeds back to the glomerulus and tends to return distal Na delivery to its original state. This feedback system is anatomically represented in the juxtaglomerular apparatus (JGA), the complex of epithelial, mesangial, and vascular cells formed as the result of the return of the thick ascending limb to its origin, a fact well appreciated and illustrated in Homer Smith’s textbook (Figure 1). In essence, specialized epithelial cells called macula densa (MD) cells make contact with a mesangial cell type, and through them with the smooth muscle and renin-producing granular cells of the glomerular arterioles. It is now clear that this structure serves two functions, both related to Na homeostasis. One is the vascular feedback control called tubuloglomerular feedback or TGF for short; the other is NaCl-dependent control of renin secretion and renin synthesis. Although both of these effector systems respond to the same input, they differ in their temporal characteristics. Figure 1. : Schematic diagram of a nephron demonstrating the return of the thick ascending limb to the vascular pole of its own glomerulus. A histological view of the juxtaglomerular apparatus is shown in the upper left hand corner. The diagram is from Smith H, The Kidney (Plate I) (61).Macula densa control of vascular tone has a fast response time with full activation or deactivation occurring within 20 to 60 s, it compensates for random and high frequency variations of the MD signal in the vicinity of its operating point, it acts as a minute-to-minute controller of salt delivery to the distal nephron, and prolonged out-of-range deviations of the macula densa signal cause resetting of the response curve. Macula densa control of renin secretion has a slow response time and a slow frequency response (about 2 to 3 mHz or cycle times between 5 and 8 min), it responds to prolonged deviations of the MD signal, those associated with changes in EC volume for example, it acts as a long-term controller of body salt balance through changes in systemic angiotensin II levels, and it is inexhaustible because the MD signal can also affect renin transcription. Characteristics of Tubuloglomerular Feedback Experimental studies of tubuloglomerular feedback began with the microinjection experiments that Klaus Thurau designed and initiated (69). After outlining the proximal segments that belong to the same nephron, its distal tubule was punctured and fluid was manually injected. When this fluid was an isotonic solution of either NaCl or NaBr, the previously identified proximal tubule would usually completely collapse, and this was interpreted as indicating a severe reduction in GFR. In retrospect, it would appear that the choice of the experimental conditions was fortunate because complete cessation of GFR is not the typical response to distal saturation with NaCl. A subsequent approach permitted an evaluation of the TGF response in a more quantitative way (50). After identifying the segments of a proximal tubule by dye injection, loops of Henle were microperfused from the last superficial proximal segment. Flow in an upstream segment was interrupted by injecting an immobile block, and fluid collections were made in a segment even further upstream for measurement of GFR, and in the distal tubule for measurement of Na or Cl concentration. The results of such experiments showed that as perfusion rate increased there was an increase in Na concentration and a parallel decrease in single nephron GFR (SNGFR). A decrease of SNGFR was not seen when the perfusate was an isotonic mannitol of Na sulfate solution (50). A further refinement (Figure 2, left) consisted of performing nine repeat measurements of SNGFR at nine different flow rates, and this design revealed a sigmoidal relationship between SNGFR and end-proximal flow rate, with SNGFR falling by about 50%, and the flow causing a half-maximal response being precisely in the range of normal end-proximal flows (7). In a variant approach, first performed in collaboration with Erik Persson, stop flow pressure, a close correlate of glomerular capillary pressure, was measured instead of SNGFR; this technique is technically somewhat easier and has proven useful as a screening method (48). Stop flow pressure shows the same sigmoidal relationship to perfusion rate, with the exception that responses are less easily triggered in the low-flow range, and the midpoint is therefore shifted to the right (Figure 2, right). Thomson and Blantz (68) developed a technique in which the tubule is not blocked, and in which loop flow is perturbed by the addition or withdrawal of fluid at low rates. Tubular flow is measured upstream from the perturbation by a videometric quasi-continuous method, permitting an assessment of the compensatory power of the entire feedback loop. A compensation of one would be optimal because it would indicate that an addition of 1 nl/min had caused a fall in upstream flow of 1 nl/min. Maximal error compensation of between 0.8 and 0.9 was observed in response to small perturbations around normal flows. Figure 2. : (Left) Relationship between the rate of loop of Henle perfusion and single nephron GFR (SNGFR) determined as the product of early proximal tubular flow rate (VEP) and the tubular fluid to plasma inulin ratio. Data are from Briggs et al. (7). (Right) Relationship between the rate of loop of Henle perfusion and the associated change of stop flow pressure (PSF). Data are from Schnermann et al. (48).An important characteristic feature of TGF is its tendency to reset by shifting its operational range into a higher or lower flow range and by changing its sensitivity by either decreasing or increasing the slope of the TGF function. This typically occurs when the signal is forced out of operating range for too long. Often, left shifts are situations of increased angiotensin II, states of volume depletion for example, and right shifts are states of reduced angiotensin II and perhaps increased nitric oxide, volume expansion for example. In summary, the TGF control system describes an inverse and sigmoidal relationship between SNGFR and VLP; its operating point is at the function midpoint, its optimal error compensation is around the operating point, it can reset its operational range and its sensitivity to flow perturbations, it has a tendency for stable oscillations at 35 to 50 mHz (28), it is a single nephron mechanism although TGF-induced smooth muscle activation can spread by electrotonic coupling (24), and its vascular effect is long-lasting. Mechanisms of JGA-Mediated Regulation of Vascular Tone and Renin Secretion The information pathway linking the tubular lumen with the vascular effector cells sequentially involves the different cell types of the JGA. It begins with a signaling event in the tubular lumen that affects two end points, contractility and renin secretion, through initiating an epithelial cell response, most likely of macula densa cells, followed by an extracellular transmission mechanism that bridges the JGA interstitium. Luminal Signal Tubuloglomerular Feedback. To identify the nature of the tubular signal that is required to alter the vascular endpoint, a modified technique was used in which test solutions were perfused into the distal tubule, a site that is much closer to the JG region so that one can assume that the test solutions reach the sensing site largely unmodified (49). Pure NaCl solutions elicited the full response at a constant flow rate, and responses were seen in the hypotonic range with saturation reached at about 60 mM (Figure 3). Half-maximum concentration was about 35 mM, which we estimate to be close to the normal NaCl concentration in the macula densa region. Studies in an isolated rabbit preparation show that half-maximum and saturation concentrations may be even lower in that species (40). Using the same technique of retrograde perfusion, normal responses were seen when Na was substituted with small univalent cations, but not when Cl was substituted with a number of anions, the exception being a Br substitution (49). We concluded from this that variations in Cl concentrations are responsible for causing the TGF response, whereas variations in Na concentration or osmolarity are not a mandatory requirement for the vascular reaction. To summarize, the luminal signal for TGF is dependent on graded changes of Cl and possibly K, the Cl concentration causing half-maximum response is about 35 mM, substitution of Na by other monovalent cations has no effect on the response, and signaling is independent of flow, osmolarity, and luminal calcium concentration (47). Figure 3. : Relationship between the NaCl concentration in solutions perfused into the distal tubule at a rate of 15 nl/min and the associated percent change in SNGFR. C1/2 indicates the NaCl concentration causing the half-maximum decrease of SNGFR. Data are from Schnermann et al. (49).Renin Secretory Response. Macula densa-dependent renin secretion has been proposed and studied in vivo by Vander (74), but renin secretion is under the control of important drives in addition to that of the macula densa; it was therefore difficult to decide whether and to what extent these other drives were involved. To study MD-dependent renin secretion in isolation, Skott and Briggs (56) developed a technique in which the thick ascending limb and its adherent glomerulus were dissected from rabbit kidneys and perfused with the Burg technique through the ascending limb. The perfused preparation was retracted into an outer pipette that functions as a superfusion chamber. After immersion of this whole setup in oil, the superfusate carrying the released renin would form droplets that could be collected in defined time intervals for measurement of renin activity (Figure 4, A and B). The change in renin release in response to changing the luminal perfusate composition is macula densa-mediated, because this preparation does not consist of much more than the tubuloglomerular contact area and because it obviously has no baroreceptor or regulated sympathetic input. As shown in Figure 4C, an increase in NaCl concentration caused a reversible inhibition and a decrease in NaCl a reversible stimulation of renin release with maximal deflections being reached after about 20 min (29). Resolution of the concentration dependency of renin secretion showed that the renin response, like TGF, covers the hypotonic concentration range below 60 mM, with a half-maximum Cl concentration of about 25 to 30 mM (19). To summarize, the luminal signal for the renin secretory response is dependent on graded changes of Cl, the Cl concentration causing the half-maximum response is about 25 to 30 mM, substitution of Na by Rb or choline has no effect on the response, substitution of Cl by isethionate or acetate eliminates the response, and the response is largely determined by NaCl concentration, not NaCl load. Figure 4. : (A) Preparation of a glomerulus (Glom) with attached thick ascending limb (TAL) isolated from rabbit kidney. The pipette perfusing the thick ascending limb is visible on the left. In the rabbit, the macula densa cells (MD) form a plaque that bulges into the tubular lumen (56). (B) The specimen attached to the perfusion pipette has been withdrawn into an outer glass pipette that serves as a superfusion chamber. After establishing a superfusion flow, the superfusion pipette is immersed in mineral oil. The superfusate containing the released renin forms droplets that are withdrawn in time intervals for renin analysis (57). (C) Increasing perfusate NaCl concentration from Cl/Na 7/26 mM to 122/141 mM causes a reversible decrease in renin secretion rate (top); decreasing it from Cl/Na 122/141 mM to 7/26 mM causes a reversible increase in renin secretion. Data are from Weihprecht et al. (29) and Lorenz et al. (77).Epithelial Cell Response Elicited by a Luminal NaCl Concentration Change. Loop diuretics such as furosemide, bumetanide, and others have been observed to completely block TGF responses and to stimulate renin secretory responses (30,79). In a study in which we determined the dose dependence of TGF and transport inhibition, we found that bumetanide inhibited TGF at exactly those concentrations at which a change in distal Cl concentration indicated inhibition of transport, indicating that these two events were causally linked. Recent studies with the K channel blocker U37883A, as well as earlier experiments with Ba from our laboratory, showed that K channel blockade largely blocked TGF responses (45,73). Furthermore, Lorenz et al. (32) recently showed that ROMK knockout mice have very small or no TGF responses. The identical effect of inhibitors of two entirely different transporters indicates that their common role in determining NaCl transport activity rather than any other effects of NKCC2 or ROMK is an early epithelial event in the macula densa pathway. Studies by Rainer Greger (16) in cortical TAL segments had shown that the NaCl cotransport activity is modified by Cl with an affinity constant of about 50 mM, whereas the affinities for Na or K are so high that the cotransporter is essentially always saturated. Together with the inhibitory effect of loop diuretics, we believe that this finding may explain why, as alluded to earlier, the TGF response is regulated by changes in Cl concentration and why it is apparently Na independent. Implicit in this argument is the notion that affinity constants of NKCC2 in the thick ascending limb are comparable to those in the MD. This notion has been challenged by the recent findings of Gimenez and Forbush, who have observed that the Cl affinity of the NKCC2 B isoform, the isoform most likely expressed in macula densa cells, had a much higher affinity, only 9 mM, when expressed in oocytes (13). It is still unclear which of the epithelial cell changes that are dependent on increased activity of NKCC2 are part of the juxtaglomerular signaling pathway. NaCl cotransport activation has been shown to cause increases of cytosolic Na and K concentration, an increase of cytosolic Cl concentration, cell depolarization, probably an increase in cytosolic Ca concentration, cell swelling, and cell alkalinization (12,25–27,43,44). All of these changes could be involved in initiating downstream events, perhaps with the exception of cell alkalinization, because loop diuretics were found to increase cell pH without eliciting a TGF response (12). A related problem is to determine which of the transporters and receptors that have been identified by the work of several laboratories, most notably those of Bell/Lapointe (3) and Persson (38), to be present in apical or basolateral membranes of macula densa cells play a role in juxtaglomerular regulatory pathways. Up until now, with the exception of NKCC2 and ROMK, none of the other membrane proteins have been unequivocally implicated in the TGF or renin pathway (Figure 5). Ongoing studies will hopefully shed some light on the roles of the luminal NHE, H/K ATPase, angiotensin II receptors, and others in determining the vascular response. A role for NO generated by macula densa nNOS has been established that on the basis of inhibitor studies consists of an attenuation of the full vasoconstrictor response (39,78). Figure 5. : Summary of membrane transporters and surface receptors found in apical and basolateral membranes of macula densa cells. The presence of NOS 1 and COX-2 in macula densa cells is indicated. For proteins outlined in red, a connection to the specific function of the macula densa cells to affect vascular tone or renin secretion has been established. See text for references.Mechanisms Mediating between the Macula Densa and Effector Cells Tubuloglomerular Feedback. It is widely assumed that an increase in luminal NaCl concentration causes changes in the interstitium of the JGA that alter the function of granular and smooth muscle cells. Because of the cellular discontinuity, the most common notion is that these changes are changes in the appearance and interstitial concentration of a paracrine mediator. A number of such mediators have been invoked on the basis of pharmacologic inhibitor studies, but conclusive evidence and therefore a consensus has been difficult to reach. It is in this area of research that the use of knockout mice has been and will continue to be quite helpful. For example, mice with knockout mutations in the thromboxane receptor have perfectly normal TGF responsiveness, suggesting that thromboxane is not a major player in this response under normal conditions (66). Rather solid pharmacologic evidence in various in vivo and in vitro preparations has implicated adenosine 1 receptors (A1AR) as participating in the TGF response. For example, the A1AR antagonist FK 838 fully prevented the vascular response to an elevation in luminal NaCl concentration in an in vitro perfused rabbit preparation in which the thick ascending limb was perfused and in which diameter changes of the perfused afferent arterioles served as TGF end point (40). Thus, on the assumption that the endogenous A1AR ligand, adenosine, may trigger the TGF response through this receptor subtype, as originally suggested by Osswald (37), Daqing Sun in our lab with the help of Linda Samuelson (65) generated mice with a knockout mutation in this receptor subtype. In fact, TGF responses of stop flow pressure or single nephron GFR were found to be entirely absent in A1AR−/− mice (65). The dependence of TGF responsiveness on the presence of intact A1 adenosine receptors was confirmed in another strain of A1AR knockout mice that was independently generated by Fredholm and studied in the laboratory of Erik Persson (8). Thus, we conclude that an increase in luminal NaCl causes the appearance in the JGA interstitium of adenosine and that adenosine through A1AR activation causes TGF vasoconstriction. The exact source of adenosine still remains to be identified. Adenosine is presumably generated inside macula densa cells by ATP degradation, and it could be exported by one of the so-called equilibrative nucleoside transporters that are found in many membranes. Alternatively, adenosine may be generated extracellularly from released ATP. Bell et al. (2) have provided some evidence that ATP may in fact be released in a NaCl concentration-dependent fashion, perhaps through a nonspecific anion channel, and an attenuation of TGF responses by an inhibitor of ecto-5′nucleotidase has been observed by Thomson et al. (67). Renin Secretory Response. While adenosine is probably responsible for high NaCl-induced vasoconstriction, it is less certain that it plays a major role in the stimulation of renin secretion by low luminal NaCl. For example, the stimulation of renin secretion caused by furosemide, likely mediated to a major extent by the macula densa pathway, was found to be unaltered in isolated perfused kidneys of A1AR knockout mice (52). The conclusion that inhibition of adenosine production is not the primary cause for the stimulation of renin release by low NaCl agrees with earlier studies in the isolated JGA preparation in which the addition of adenosine to the bath did not induce a major inhibition of renin secretion (31). On the other hand, prostaglandins have been invoked in the macula densa pathway of renin secretion for a while. An involvement of prostaglandins in macula densa-mediated renin secretion was supported by the observation in the isolated rabbit JGA preparation that two nonsteroidal cyclooxygenase (COX) inhibitors essentially entirely blocked the renin stimulation caused by low NaCl (15). The interpretation of these findings was not entirely clear because macula densa cells were not believed to express COX-1, the only known COX enzyme at the time. A breakthrough development was the demonstration by Harris et al. (18) that macula densa cells as well as surrounding cTAL cells express COX-2, the second isoform of COX, usually called the inducible isoform. We subsequently