The Glomerular Regional Intrarenal Circulations
The extremely high flow rate which characterizes renal perfusion is confined to the renal cortex. The results of a number of representative studies performed with a variety of techniques in the dog under barbiturate anaesthesia are summarized in Figure 2.1. In these studies cortical blood flow ranged from 4-5 to 5-9 ml/g/min.
Blood flow in the renal papilla was in the neighbourhood of 0-3 ml/g/min, considerably less than that in most metabolically active organs. Outer medullary flow was intermediate, and close to that which characterizes flow in most metabolically active tissues.
A gradient of blood flow from the outer to the inner cortex has been demonstrated with a number of techniques, including microsphere delivery to the kidney (McNay & Abe 1970, Slotkoffet al. 1971, Stein et al. 1973b), and rubidium delivery (Steiner & King 1970). Microsphere delivery, perhaps the most widely used method to date for assessing regional cortical perfusion, has been criticized because of the possibility that the axial distribution of such a large moiety could result in streaming and preferential downstream delivery (Wallin et al. 1971).
Blood flow in the renal papilla was in the neighbourhood of 0-3 ml/g/min, considerably less than that in most metabolically active organs. Outer medullary flow was intermediate, and close to that which characterizes flow in most metabolically active tissues.
A gradient of blood flow from the outer to the inner cortex has been demonstrated with a number of techniques, including microsphere delivery to the kidney (McNay & Abe 1970, Slotkoffet al. 1971, Stein et al. 1973b), and rubidium delivery (Steiner & King 1970). Microsphere delivery, perhaps the most widely used method to date for assessing regional cortical perfusion, has been criticized because of the possibility that the axial distribution of such a large moiety could result in streaming and preferential downstream delivery (Wallin et al. 1971).
However, this criticism is certainly not applicable to ionic rubidium delivery, as employed by Steiner and King (1970). Stein et al. (1973b) has assembled evidence that streaming does not modify microsphere delivery sufficiently to produce a systematic artifact favouring outer cortical perfusion.
Katz et al. (1971), Wallin et al. (1971), Bankir et al. (1973), Kallskog et al. (1972) and Chenitz et al. (1976) have estimated regional glomerular blood How in the rabbit, rat and dog. The results are Summarized in Table 2.1. Systematic differences Were found in the blood flow to individual outer and inner cortical glomeruli in all three species, but the differences were considerably smaller than the differences in blood flow per gram.
The gradient of How per unit tissue mass from the outer to the inner cortical regions can largely be accounted for on the basis of differences in the number of glomeruli per gram in the regions, in accord with the anatomical analyses of Horster et al. (1971) and Chenitz et al. (1976).
The gradient of How per unit tissue mass from the outer to the inner cortical regions can largely be accounted for on the basis of differences in the number of glomeruli per gram in the regions, in accord with the anatomical analyses of Horster et al. (1971) and Chenitz et al. (1976).
Nephron function is initiated in the glomerulus where hydrostatic pressure in the glomerular capillary exceeds the offsetting forces-hydrostatic pressure in Bowman’s space and colloid oncotic pressure in the capillary-«to create an ultrafiltrate. This concept was placed on a firm basis by the classical micropuncture studies of Richards and collaborators, which Richards summarized in 1939.
This was the first, and perhaps still the most important contribution made by micropuncture to our understanding of renal function. Until recently glomerular capillary pressure had been measured directly only in amphibians where glomeruli are accessible to puncture, and various indirect approaches had been required to assess capillary pressure in mammals, as reviewed by Smith (1951) and Winton (1956).
As arterial pressure is reduced, urine formation generally ceases at a pressure of about 75 mmHg in the antidiuretic state, but is measurable with pressures as low as 40-50 mmHg after osmotic loading or when other manoeuvres which prevent tubular reabsorption, such as cooling the kidney or administration of cyanide, are added. Winton (1956) suggested that the difference was attributable to intrarenal tubular reabsorption with reduced filtration rates.
Similarly, progressive increments in ureteral pressure result in a progressive decrement in urine flow and glomerular filtration rate. Both indirect approaches led to estimates of glomerular capillary pressure which range from 60 to 80 per cent of mean arterial pressure. Similar objections, however, can be raised to both manoeuvres.
First, both reduced arterial and raised ureteral pressures dilate afferent arterioles and thus may modify glomerular capillary pressure (Thurau 1964). Equally important, in both cases filtration may continue despite cessation of urine liow because of maximal reabsorption of filtered solute and water. It is likely on both counts that effective filtration pressure was overestimated by the indirect techniques.
A mutant strain of Wistar rat in which superficial glomeruli are occasionally accessible to direct puncture has made a more direct measure of hydrostatic pressure in the glomerular capillary net possible (Brenner et al. 1972). In some small primates the occasional glomerulus is also sufficiently superficial to allow direct micropuncture (Maddox et al, 1974a).
Measurements in both species provided a lower estimate of glomerular capillary pressure, averaging about 50 per cent of arterial pressure, or 44 mmHg in the rat. The forces offsetting capillary hydrostatic pressure are the oncotic pressure of the plasma proteins in the glomerular capillaries (averaging 18 mmHg) and hydrostatic pressure in Bowman’s space (averaging 10 mmHg). Moreover, the ultrafiltration process itself raises plasma protein concentration.
Oncotic pressure, therefore, rises to 35 mml lg in the efferent arteriole so that the not driving force for ultrafiltration averages only 10 mml lg. These observations have provided new insight into the forces acting within the glomerulus. In the setting of these experiments filtration equilibrium is reached during the passage of blood from the afferent to the efferent end of the glomerular capillaries, i.e. hydrostatic pressure in Bowman’s space plus the increased oncotic pressure in the glomerular capillaries resulting from the ultrafiltration of fluid are sufficient (44 mmHg) to offset capillary hydrostatic pressure; glomerular filtration has ceased by the time the blood reaches the efferent end of the glomerular capillaries. The major implication of these observations is that filtration within a glomerulus is limited by plasma flow normally.
Before we extrapolate from observations made in these anatomically unusual glomeruli, there are several reasons for reserving judgement. First, these studies are possible only when the kidney is exposed, and there is considerable evidence to suggest that the anaesthesia and surgery required to expose the kidney result in a striking reduction in renal blood flow, due to arteriolar constriction (see Burger et al. 1976 for references).
Secondly, the unusual subcapsular location of the glomeruli which can be punctured may reflect an unusually long afferent arteriole and, therefore, an unusually large pressure drop proximal to the glomerulus. Moreover, indirect evidence suggests important Species differences; for example, multiple observations suggest that filtration equilibrium is less likely in the dog than in the rat (Baer & Navar 1973, Knox et al. 1975, Chenitz et al. 1976).
Finally, according to this concept the filtration fraction should be constant. A host of studies, many reviewed below in sections dealing with the control of the renal circulation, has demonstrated striking variation in the filtration fraction as renal perfusion rises or falls. It is possible that ultimately each of these problems will be answered, but at present it is necessary to reserve judgement on the universality of filtration equilibrium.
It is possible from the direct estimates of glomerular capillary pressure, oncotic pressure in the glomerular capillaries and net Huid flux across the glomerular membrane to estimate the ultrafiltration coefficient of the glomerular capillary membrane (Renkin & Gilmore 1973, Deen et al. 1973). In addition, on the basis of assumed, but reasonable, estimates of the surface area available for filtration it is possible to compare the permeability to water flux across glomerular capillaries with capillaries in other vascular beds.
Such estimates suggest that glomerular capillary permeability exceeds by a factor of 50 or more that of other mammalian capillaries, such as those in the skeletal muscle (Deen et al. 1973). The effective ultrafiltration coefficient calculated in this way is the product of the effective hydraulic permeability and surface area of the glomerular capillaries.
Under a number of circumstances where all of the measurements (except that of . surface area available for filtration) were made. rather striking variation in the calculated ultrafiltration coefficient appeared. Blantz er al. (l970) reported a marked influence of angiotensin on the ultrafiltration coefficient. Another example was piovaled by Bayliss er al. (l970), who studied surface glomeruli in rats over a wide range of plasma protein concentration.
The filtration rate failed to increase with decreases in arterial protein concentration, despite a rise in net ultrafiltration pressure, because of the offsetting effect of a simultaneous reduction in the calculated filtration coefficient. Considerable changes in the calculated filtration coefficient have been observed in a number of other circumstances.
A sharp drop in the coefficient occurred during mannitol administration (Blantz 1974). Gertz et al. (1969) and Daugharty et al. (1971) concluded that the increase in glomerular filtration rate which followed plasma volume expansion was in large part due to the increase in plasma flow and the recruitment of an increase in the glomerular capillary surface area-which Daugharty et al. assumed to be longitudinal although alternative possibilities were not assessed.
Gassee et al. (1974) evaluated the sieving coefficients for polyvinylpyrrolidone as a function of filtration pressure in dogs and suggested that changes in the area available for filtration played a major role in keeping glomerular filtration rate constant at increased perfusion pressures. It seems much more likely that such a rapid change in the filtration coefficient reflects a variable surface area available for filtration, rather than changes in the intrinsic permeability of the capillary membrane.
Whether this reflects a variable perfusion pathway between capillaries which have a different intrinsic permeability or a modification of the intrinsic permeability of the capillaries is not yet clear, but examination of the architecture of the glomerulus provides some insight.
Major advances made in our understanding of glomerular microvascular anatomy also provide insight into function. Boyer (1956) defined the vascular pattern of the renal glomerulus by painstaking reconstruction from serial sections, which revealed an anastomotic rete of capillary channels. Independent parallel capillary 100ps were not identifiable. The detailed studies on the embryogenesis of the glomerulus performed by Hall and Roth (1956) and by Lewis (1958) also illuminate the final structure. Glomerular capillaries develop in situ; a mass of mesenchymal cells
This was the first, and perhaps still the most important contribution made by micropuncture to our understanding of renal function. Until recently glomerular capillary pressure had been measured directly only in amphibians where glomeruli are accessible to puncture, and various indirect approaches had been required to assess capillary pressure in mammals, as reviewed by Smith (1951) and Winton (1956).
As arterial pressure is reduced, urine formation generally ceases at a pressure of about 75 mmHg in the antidiuretic state, but is measurable with pressures as low as 40-50 mmHg after osmotic loading or when other manoeuvres which prevent tubular reabsorption, such as cooling the kidney or administration of cyanide, are added. Winton (1956) suggested that the difference was attributable to intrarenal tubular reabsorption with reduced filtration rates.
Similarly, progressive increments in ureteral pressure result in a progressive decrement in urine flow and glomerular filtration rate. Both indirect approaches led to estimates of glomerular capillary pressure which range from 60 to 80 per cent of mean arterial pressure. Similar objections, however, can be raised to both manoeuvres.
First, both reduced arterial and raised ureteral pressures dilate afferent arterioles and thus may modify glomerular capillary pressure (Thurau 1964). Equally important, in both cases filtration may continue despite cessation of urine liow because of maximal reabsorption of filtered solute and water. It is likely on both counts that effective filtration pressure was overestimated by the indirect techniques.
 |
| Regional blood flow rates in the dog kidney were measured with different techniques. The intravascular transit studies involved the local measurement of the transit of an intravascular indicator such as tagged proteins (Evans Blue or radio-iodinated serum albumin) or P32 tagged red cells. The measurements made with delivery limited technique involve the use of indicators such as Rb (rubidium) or tagged microspheres (MI) which are trapped in the kidneys, so that the amount in the tissue reflects the amount delivered, or blood flow. Washout studies were performed with the highly diffusible radioactive tracers krypton (Kr) and xenon (Xe) and external counting. In the studies listed under slotkoff and Hollenberg, the measurements were made with two methods in the same dogs. Note the similarity of perfusion in each zone with each method, with the exception of the inner medulla. Rubidium produces a falsely high flow estimate, presumably because of delivery via tubules. In contrast, RIA measures only inflow, whereas the other indicators measure transit and thus show a higher flow because outflow exceeds inflow considerably in the inner medulla. |
A mutant strain of Wistar rat in which superficial glomeruli are occasionally accessible to direct puncture has made a more direct measure of hydrostatic pressure in the glomerular capillary net possible (Brenner et al. 1972). In some small primates the occasional glomerulus is also sufficiently superficial to allow direct micropuncture (Maddox et al, 1974a).
Measurements in both species provided a lower estimate of glomerular capillary pressure, averaging about 50 per cent of arterial pressure, or 44 mmHg in the rat. The forces offsetting capillary hydrostatic pressure are the oncotic pressure of the plasma proteins in the glomerular capillaries (averaging 18 mmHg) and hydrostatic pressure in Bowman’s space (averaging 10 mmHg). Moreover, the ultrafiltration process itself raises plasma protein concentration.
 |
| Table 2.1: Glomerular blood flow in outer, mid and inner cortex. |
Oncotic pressure, therefore, rises to 35 mml lg in the efferent arteriole so that the not driving force for ultrafiltration averages only 10 mml lg. These observations have provided new insight into the forces acting within the glomerulus. In the setting of these experiments filtration equilibrium is reached during the passage of blood from the afferent to the efferent end of the glomerular capillaries, i.e. hydrostatic pressure in Bowman’s space plus the increased oncotic pressure in the glomerular capillaries resulting from the ultrafiltration of fluid are sufficient (44 mmHg) to offset capillary hydrostatic pressure; glomerular filtration has ceased by the time the blood reaches the efferent end of the glomerular capillaries. The major implication of these observations is that filtration within a glomerulus is limited by plasma flow normally.
Before we extrapolate from observations made in these anatomically unusual glomeruli, there are several reasons for reserving judgement. First, these studies are possible only when the kidney is exposed, and there is considerable evidence to suggest that the anaesthesia and surgery required to expose the kidney result in a striking reduction in renal blood flow, due to arteriolar constriction (see Burger et al. 1976 for references).
Secondly, the unusual subcapsular location of the glomeruli which can be punctured may reflect an unusually long afferent arteriole and, therefore, an unusually large pressure drop proximal to the glomerulus. Moreover, indirect evidence suggests important Species differences; for example, multiple observations suggest that filtration equilibrium is less likely in the dog than in the rat (Baer & Navar 1973, Knox et al. 1975, Chenitz et al. 1976).
Finally, according to this concept the filtration fraction should be constant. A host of studies, many reviewed below in sections dealing with the control of the renal circulation, has demonstrated striking variation in the filtration fraction as renal perfusion rises or falls. It is possible that ultimately each of these problems will be answered, but at present it is necessary to reserve judgement on the universality of filtration equilibrium.
It is possible from the direct estimates of glomerular capillary pressure, oncotic pressure in the glomerular capillaries and net Huid flux across the glomerular membrane to estimate the ultrafiltration coefficient of the glomerular capillary membrane (Renkin & Gilmore 1973, Deen et al. 1973). In addition, on the basis of assumed, but reasonable, estimates of the surface area available for filtration it is possible to compare the permeability to water flux across glomerular capillaries with capillaries in other vascular beds.
Such estimates suggest that glomerular capillary permeability exceeds by a factor of 50 or more that of other mammalian capillaries, such as those in the skeletal muscle (Deen et al. 1973). The effective ultrafiltration coefficient calculated in this way is the product of the effective hydraulic permeability and surface area of the glomerular capillaries.
Under a number of circumstances where all of the measurements (except that of . surface area available for filtration) were made. rather striking variation in the calculated ultrafiltration coefficient appeared. Blantz er al. (l970) reported a marked influence of angiotensin on the ultrafiltration coefficient. Another example was piovaled by Bayliss er al. (l970), who studied surface glomeruli in rats over a wide range of plasma protein concentration.
The filtration rate failed to increase with decreases in arterial protein concentration, despite a rise in net ultrafiltration pressure, because of the offsetting effect of a simultaneous reduction in the calculated filtration coefficient. Considerable changes in the calculated filtration coefficient have been observed in a number of other circumstances.
A sharp drop in the coefficient occurred during mannitol administration (Blantz 1974). Gertz et al. (1969) and Daugharty et al. (1971) concluded that the increase in glomerular filtration rate which followed plasma volume expansion was in large part due to the increase in plasma flow and the recruitment of an increase in the glomerular capillary surface area-which Daugharty et al. assumed to be longitudinal although alternative possibilities were not assessed.
Gassee et al. (1974) evaluated the sieving coefficients for polyvinylpyrrolidone as a function of filtration pressure in dogs and suggested that changes in the area available for filtration played a major role in keeping glomerular filtration rate constant at increased perfusion pressures. It seems much more likely that such a rapid change in the filtration coefficient reflects a variable surface area available for filtration, rather than changes in the intrinsic permeability of the capillary membrane.
Whether this reflects a variable perfusion pathway between capillaries which have a different intrinsic permeability or a modification of the intrinsic permeability of the capillaries is not yet clear, but examination of the architecture of the glomerulus provides some insight.
Major advances made in our understanding of glomerular microvascular anatomy also provide insight into function. Boyer (1956) defined the vascular pattern of the renal glomerulus by painstaking reconstruction from serial sections, which revealed an anastomotic rete of capillary channels. Independent parallel capillary 100ps were not identifiable. The detailed studies on the embryogenesis of the glomerulus performed by Hall and Roth (1956) and by Lewis (1958) also illuminate the final structure. Glomerular capillaries develop in situ; a mass of mesenchymal cells
differentiates into capillary endothelial cells and mesangial elements. The anatomic relationship between the mesangial and endothelial cells is therefore much more intimate than if capillary elements represented an invagination, as was thought earlier. The relationships of the vascular structure suggested to Lewis that ‘the large channels . . . of the peripheral plexus may exist as shunts through the plexiform glomerulus from afferent to efferent arterioles’.
Barger and Herd (1971) also suggested that larger ‘thoroughfare’ channels in the glomerular periphery may remain patent, while smaller anastomosing vessels open and close under different physiological conditions, thus increasing and decreasing the available surface area for filtration. There is, moreover, the possibility that different capillaries have different filtration characteristics, providing another means by which glomerular filtration might be controlled.
Berliner (1973) concluded that there is excellent evidence in mammals that variations in glomerular filtration rate depend on changes in the volume of filtrate formed by each glomerulus rather than on the number of glomeruli contributing to the process. However, he pointed out that ‘the physiologic data on which these conclusions are based do not bear on the possibility of variation in the number of capillary loops per glomerulus open at any time’.
These anatomic observations relate directly to the concept of filtration equilibrium: under the conditions of filtration equilibrium the bulk of glomerular filtrate is formed across only a fraction of the available glomerular capillary surface area, and the volume of glomerular filtrate formed is dependent principally on the volume of glomerular plasma flow and not on available surface area.
\
However, variation in the capillary filtration coefficient, which is now well documented, requires an explanation. The possibility that hltration occurs primarily in some specialized glomerular capillaries which have a variable perfusion provides an attractive explanation.
Shunts provide an alternative explanation for variation in calculated glomerular filtration 006 efficient. Zlabek (1957) identified relatively thick capillaries connecting afferent and efferent arterioles only in human juxtamedullary glomeruli, providing a potential bypass of ultrafiltration in this region. Strong support came from Ljungqvist (1964) who identified a similar connection between the afferent and the efferent arterioles.
He suggested, moreover, that the adrenergic innervation of the continuous juxtaglomerular arterioles provides a basis for variations in their circulation, especially in the juxtamedullary region (Ljungqvist & Wagermark 1970). These observations were extended recently by Takazakura et al. (1972) in man, who found that the number of juxtamedullary glomeruli showing such a bypass increased with increasing age, becoming dominant by the fifth decade.
Barger and Herd (1971) also suggested that larger ‘thoroughfare’ channels in the glomerular periphery may remain patent, while smaller anastomosing vessels open and close under different physiological conditions, thus increasing and decreasing the available surface area for filtration. There is, moreover, the possibility that different capillaries have different filtration characteristics, providing another means by which glomerular filtration might be controlled.
Berliner (1973) concluded that there is excellent evidence in mammals that variations in glomerular filtration rate depend on changes in the volume of filtrate formed by each glomerulus rather than on the number of glomeruli contributing to the process. However, he pointed out that ‘the physiologic data on which these conclusions are based do not bear on the possibility of variation in the number of capillary loops per glomerulus open at any time’.
These anatomic observations relate directly to the concept of filtration equilibrium: under the conditions of filtration equilibrium the bulk of glomerular filtrate is formed across only a fraction of the available glomerular capillary surface area, and the volume of glomerular filtrate formed is dependent principally on the volume of glomerular plasma flow and not on available surface area.
\
However, variation in the capillary filtration coefficient, which is now well documented, requires an explanation. The possibility that hltration occurs primarily in some specialized glomerular capillaries which have a variable perfusion provides an attractive explanation.
Shunts provide an alternative explanation for variation in calculated glomerular filtration 006 efficient. Zlabek (1957) identified relatively thick capillaries connecting afferent and efferent arterioles only in human juxtamedullary glomeruli, providing a potential bypass of ultrafiltration in this region. Strong support came from Ljungqvist (1964) who identified a similar connection between the afferent and the efferent arterioles.
He suggested, moreover, that the adrenergic innervation of the continuous juxtaglomerular arterioles provides a basis for variations in their circulation, especially in the juxtamedullary region (Ljungqvist & Wagermark 1970). These observations were extended recently by Takazakura et al. (1972) in man, who found that the number of juxtamedullary glomeruli showing such a bypass increased with increasing age, becoming dominant by the fifth decade.
Spinelli et al. (1972) were unable to identify such a structure despite an elegant technique based on fixation in vivo, ‘Microfil’ perfusion and scanning electron microscopy. The ages of the animals that were studied were not noted, however. More recently Ljungqvist (1976) has performed a painstaking study, providing beautiful documentation of the ultrastructure of the continuous arteriole which traverses the juxtaglomerular apparatus in the juxtamedullary but not the cortical glomeruli. Such an arrangement provides a cogent explanation for the pale cortex and congested medulla characteristic of settings in which filtration ceases, which so influenced Trueta’s thinking (1947).
If such a potential for variation in intraglomerular perfusion exists, is there an intraglomerular control system? Considerable, albeit circumstantial, evidence suggests such a role for the mesangium. Goormaghtigh (1942) first pointed out the fibrillar structure of the mesangial cells three decades ago, clearly indicating that this could provide a contractile function.
Supporting morphologic evidence based mainly on electron microscopy has been reviewed by Hornych et al. (1972), who demonstrated, moreover, that glomerular capillary volume in vivo fell strikingly in response to angiotensin, especially in superficial cortical glomeruli. Several independent observations support such a concept. Bernik (1969) demonstrated spontaneous contractile activity in elements of human glomeruli grown in tissue culture, which was attributed to the mesangial cells.
Becker (1971) developed a highly specific antiserum to human smooth muscle actomyosin which showed great affinity for the mesangial contractile elements, providing an immunochemical link between the contractile elements of smooth muscle and the mesangium. He suggested that contraction of the mesangium may play a significant role in regulating glomerular blood flow.
Hornych (1972) attributed the dose-related constriction of glomerular capillaries which occurred even at subpressor doses to angiotensin-mediated activation of the mesangial contractile elements. Sraer et al. (1974) and Osborne et al. (1975) have demonstrated specific, high affinity receptors for angiotensin in glomeruli, which Osborne at al. localized to mesangial cells, and Sraer at al. demonstrated would reduce the volume of the glomeruli in vitro.
Given these observations it was disappointing that Myers et al. (l975a) could not detect any influence of either angiotensin or norepinephrine on the glomerular capillary ultrafiltration coefficient, suggesting an unaltered capillary surface area available for filtration, as discussed above. More recently Blantz et al. (1976) were able to demonstrate a striking influence of angiotensin on the ultrafiltration coefficient in an elegant study in the Munich-Wistar rat.
Their protocol differed from that of Myers in that the rats were volume expanded, thus creating an enhanced and more consistent sensitivity to angiotensin. Moreover, the increased glomerular plasma flow consequent to plasma volume expansion resulted in initial filtration disequilibrium, making it possible to demonstrate a reduction in the ultrafiltration coefficient with angiotensin.
Perhaps the combined effects of anaesthesia and trauma in the antidiuretic state, the state in which Myers et al. (1975) performed their study, reduced the ultrafiltration coefficient to a physiological minimum.