Medical Physiology A Cellular and Molecular Approach, Updated 2nd Ed.


Gerhard Giebisch and Erich Windhager

The kidneys serve three essential functions. First, they function as filters, removing metabolic products and toxins from the blood and excreting them through the urine. Second, they regulate the body’s fluid status, electrolyte balance, and acid-base balance. Third, the kidneys produce or activate hormones that are involved in erythrogenesis, Ca2+ metabolism, and the regulation of blood pressure and blood flow.


We begin this discussion with a macroscopic view and progress to the microscopic level as we describe the nephron, the functional unit of the kidney that is repeated approximately 1 million times within each kidney.

The kidneys are paired, retroperitoneal organs that comprise a complex mixture of vascular and epithelial elements

The human kidneys are paired, bean-shaped structures that lie behind the peritoneum on each side of the vertebral column (Fig. 33-1A). They extend from the 12th thoracic vertebra to the third lumbar vertebra. The two kidneys together comprise somewhat less than 0.5% of the total body weight; in men, each kidney weighs 125 to 170 g, whereas in women, each kidney weighs 115 to 155 g.


Figure 33-1 Structure of the urinary system. B, Posterior view of the right kidney.

A fibrous, almost nondistensible capsule covers each kidney (Fig. 33-1B). In the middle of the concave surface, a slit in the capsule—the hilus—serves as the port of entry for the renal artery and nerves and as the site of exit for the renal vein, the lymphatics, and the ureter. The hilus opens into a shallow space called the renal sinus, which is completely surrounded by renal parenchyma except where it connects with the upper end of the ureter. The renal sinus includes the urine-filled spaces: the renal pelvis proper and its extensions, the major and the minor calyces. Blood vessels and nerves also pass through the sinus. The renal capsule reflects into the sinus at the hilus so that its inner layers line the sinus and the outer layers anchor to the blood vessels and renal pelvis.

A section of the kidney (Fig. 33-1B) reveals two basic layers, the cortex (granular outer region) and the medulla (darker inner region). The granularity of the cortex results from the presence of glomeruli, microscopic tufts of capillaries, and a large number of highly convoluted epithelial structures in the form of tubules. The medulla lacks glomeruli and consists of a parallel arrangement of tubules and small blood vessels.

The medulla is subdivided into 8 to 18 conical renal pyramids, whose bases face the cortical-medullary border; the tip of each pyramid terminates in the renal pelvis. At the tip of each pyramid are perforations, almost invisible to the naked eye, through which urine flows into the minor calyces of the renal sinus.

The kidneys are unique in having a very high blood flow and glomerular capillaries that are bounded by upstream and downstream arterioles

Although the kidneys comprise less than 0.5% of total body weight, they receive ~20% of the cardiac output. This high blood flow provides the blood plasma necessary for forming an ultrafiltrate in the glomeruli. The renal circulation has a unique sequence of vascular elements: a high-resistance arteriole (the afferent arteriole), followed by a high-pressure glomerular capillary network for filtration, followed by a second high-resistance arteriole (the efferent arteriole), which is followed by a low-pressure capillary network that surrounds the renal tubules (peritubular capillaries) and takes up the fluid absorbed by these tubules.

The main features of the renal vascular system are illustrated in Figure 33-1B and C. A single renal artery enters the hilus and divides into anterior and posterior branches, which give rise to interlobar and then arcuate arteries. The latter arteries skirt the corticomedullary junction, where they branch into ascending interlobular arteries that enter the cortex and give rise to numerous afferent arterioles. These give rise to glomerular capillaries that rejoin to form efferent arterioles. For nephrons in the superficial portion of the cortex, the efferent arterioles are the origin of a dense peritubular capillary network that supplies oxygen and nutrients to the tubules in the cortex. The afferent and efferent arterioles determine the hydrostatic pressure in the interposed glomerular capillaries. The tone of both arterioles is under the control of a rich sympathetic innervation, as well as a wide variety of chemical mediators.

Very small branches of the arcuate artery, or the proximal portion of the interlobular artery, supply a subpopulation of juxtamedullary glomeruli that are located at or near the junction of cortex and medulla. The efferent arterioles of these nephrons descend into the renal papillae to form hairpin-shaped vessels called the vasa recta, which provide capillary networks for tubules in the medulla. Some 90% of the blood entering the kidney perfuses superficial glomeruli and cortex; only ~10% perfuses juxtamedullary glomeruli and medulla.

Lymph vessels, which drain the interstitial fluid of the cortex and may contain high concentrations of renal hormones such as erythropoietin (EPO), leave the kidney by following arteries toward the hilus. Lymphatics are absent from the renal medulla, where they would otherwise tend to drain the high-osmolality interstitial fluid, which is necessary for producing concentrated urine (see Chapter 38).

The functional unit of the kidney is the nephron

Each kidney consists of 800,000 to 1,200,000 nephrons. Each nephron is an independent entity until the point at which its collecting duct merges with the collecting duct or ducts from one or more other nephrons.

A nephron consists of a glomerulus and a tubule. The glomerulus is a cluster of blood vessels from which the plasma filtrate originates. The tubule is an epithelial structure consisting of many subdivisions, designed to convert the filtrate into urine. These two entities—vascular and epithelial—meet at the blind end of the tubule epithelium, which is called Bowman’s capsule or the glomerular capsule. This capsule surrounds the glomerulus and contains Bowman’s space, which is contiguous with the lumen of the tubule. It is here that filtrate passes from the vascular system into the tubule system.

The remainder of the nephron consists of subdivisions of the tubule (Fig. 33-2). The epithelial elements of the nephron include Bowman’s capsule, the proximal tubule, the thin descending and thin ascending limbs of the loop of Henle, the thick ascending limb of the loop of Henle, the distal convoluted tubule, and the connecting tubule. The connecting tubule leads further into the initial collecting tubule, the cortical collecting tubule, and the medullary collecting ducts. (See Note: Definition of a Nephron)


Figure 33-2 Structure of the nephron.

Within the renal cortex, as noted earlier, one can distinguish two populations of nephrons (Fig. 33-2). Superficial nephrons have short loops extending to the boundary between outer and inner medulla. Juxtamedullary nephrons, which play a special role in the production of concentrated urine, have long loops that extend as far as the tip of the medulla.

The renal corpuscle has three components: vascular elements, the mesangium, and Bowman’s capsule and space

The renal corpuscle, the site of formation of the glomerular filtrate, comprises the glomerulus, Bowman’s space, and Bowman’s capsule. During the development of the kidney, the interaction between the ureteric bud—which gives rise to the urinary system from the collecting duct to the ureters—and the surrounding loose mesenchyme (Fig. 33-3A) leads to the branching of the ureteric bud and condensation of the mesenchyme (Fig. 33-3B). These condensed cells differentiate into epithelium that forms a hollow, S-shaped tubular structure (Fig. 33-3C) that gives rise to the nephron’s tubular elements between Bowman’s capsule and the connecting segment. The distal portion of the S-shaped tubular structure elongates and connects with branches of the developing ureteric bud (Fig. 33-3D). At the same time, the blind proximal end of this S-shaped tubule closely attaches to the arterial vascular bundle that develops into the glomerular capillary tuft. Thinning of the epithelium on one circumference of the blind end of the S-shaped tubule leads to the emergence of the future parietal layer of Bowman’s capsule. In contrast, the opposite visceral layer thickens and attaches to the glomerular capillaries (Fig. 33-3D). These visceral epithelial cells later fold and develop into podocytes (Fig. 33-3E).


Figure 33-3 A-G, Development of glomerulus and Bowman’s capsule. H, The capillary lumen. The four major layers of the glomerular filtration barrier. I, Model of podocyte foot processes and slit membrane. The diagram shows nephrin and other proteins of the slit diaphragm. (A-E modified from Ekblom P: In Seldin DW, Giebisch G: The Kidney, 2nd ed, pp 475-501. New York: Raven Press, 1992; H modified from Kriz W, Kaissling B: In Seldin DW, Giebisch G [eds]: The Kidney: Physiology and Pathophysiology, 3rd ed, pp 587-854. New York: Raven Press, 2000.)

In the mature kidney (Fig. 33-3F), foot processes of the podocytes cover the glomerular capillaries. These podocytes, modified epithelial cells, thus represent the visceral layer of Bowman’s capsule. Beginning at the vascular pole, the podocytes are continuous with the parietal layer of Bowman’s capsule. Glomerular filtrate drains into the space between these two layers (Bowman’s space) and flows into the proximal tubule at the urinary pole of the renal corpuscle.

The glomerular filtration barrier between the glomerular capillary lumen and Bowman’s space comprises four elements with different functional properties (Fig. 33-3G): (1) a glycocalyx covering the luminal surface of endothelial cells, (2) the endothelial cells, (3) the glomerular basement membrane, and (4) epithelial podocytes.

The glycocalyx consists of negatively charged glycosaminoglycans (see Chapter 2) that may play a role in preventing leakage of large negatively charged macromolecules. Endothelial cells of the glomerular capillaries are almost completely surrounded by the glomerular basement membrane and a layer of podocyte foot processes (Fig. 33-4). The exception is a small region toward the center of the glomerulus, where the endothelial cells have neither basement membrane nor podocytes and come into direct contact with mesangial cells resembling smooth muscle. Filtration occurs away from the mesangial cells, at the peripheral portion of the capillary wall, which is covered with basement membrane and podocytes. The endothelial cells contain large fenestrations, 70-nm holes that provide no restriction to the movement of water and small solutes—including proteins or other large molecules—out of the lumen of the capillary (Fig. 33-5). Thus, the endothelial cells probably serve only to limit the filtration of cellular elements (e.g., erythrocytes).


Figure 33-4 Glomerular capillaries covered by the foot processes of podocytes. This scanning electron micrograph shows a view of glomerular capillaries from the vantage point of Bowman’s space. The outer surfaces of the capillary endothelial cells are covered by a layer of interdigitating foot processes of the podocytes. The podocyte cell body links to the foot processes by leg-like connections. (Courtesy of Don W. Fawcett.)


Figure 33-5 Inner aspect of glomerular capillaries, showing fenestrations of endothelial cells (arrows). This scanning electron micrograph shows a view of the glomerular capillary wall from the vantage point of the capillary lumen. Multiple fenestrations, each ~70 nm in diameter, perforate the endothelial cells. (From Brenner BM: Brenner and Rector’s The Kidney, 7th ed, vol 1, p 10. Philadelphia: Saunders, 2004.)

The basement membrane, located between endothelial cells and podocyte foot processes (Fig. 33-3G), separates the endothelial layer from the epithelial layer in all parts of the glomerular tuft. The basement membrane itself has three layers (Fig. 33-3H): (1) an inner thin layer (lamina rara interna); (2) a thick layer (lamina densa), and (3) an outer thin layer (lamina rara externa). The basement membrane makes an important contribution to the permeability characteristics of the filtration barrier by restricting intermediate-sized to large solutes (molecular weight >1 kDa). Because the basement membrane contains heparan sulfate proteoglycans, it is especially restricts large, negatively charged solutes (see Fig. 34-4).

Podocytes have foot interdigitating processes that cover the basement membrane (Fig. 33-4). Between the interdigitations are filtration slits (Fig. 33-3H), which are connected by a thin diaphragmatic structure—the slit diaphragm—with pores ranging in size from 4 to 14 nm. Glycoproteins with negative charges cover the podocytes, filtration slits, and slit diaphragms. These negative charges contribute to the restriction of filtration of large anions (Fig. 33-4). Nephrin, neph1, podocin, and other membranes organized on lipid rafts of podocytes form the slit diaphragm (Fig. 33-3I). Phosphotyrosine motifs on the intracellular domains of some of these proteins may recruit other molecules involved in signaling events that control slit permeability. The extracellular domains of nephrin, neph1, and FAT1 from adjacent podocytes may zip together to help form the filtration slit. In Finnish-type nephrosis, the genetic absence of nephrin leads to severe proteinuria. (See Note: Finnish-Type Nephrosis)

Supporting the glomerular capillary loops is a network of contractile mesangial cells, which secrete the extracellular matrix. This network is continuous with the smooth muscle cells of the afferent and efferent arterioles. The matrix extends to the extraglomerular mesangial cells (Fig. 33-3F). The juxtaglomerular apparatus (JGA) includes the extraglomerular mesangial cells, the macula densa, and the granular cells. The macula densa (Latin [dense spot]) is a region of specialized epithelial cells of the thick ascending limb, where it contacts its glomerulus (Fig. 33-3F). These cells have strikingly large nuclei and are closely packed, and thus they have a plaque-like appearance. The granular cells, also called juxtaglomerular or epithelioid cells, in the wall of afferent arterioles are specialized smooth muscle cells that produce, store, and release renin (see Chapter 40). The JGA is part of a complex feedback mechanism that regulates renal blood flow and filtration rate (see Chapter 34), and it also indirectly modulates Na+ balance (see Chapter 40) and systemic blood pressure (see Chapter 23).

The tubule components of the nephron include the proximal tubule, loop of Henle, distal tubule, and collecting duct

Figure 33-6 illustrates the ultrastructure of the cells of the different tubule segments. Table 33-1 lists these segments and their abbreviations. Based on its appearance at low magnification, the proximal tubulecan be divided into the proximal convoluted tubule (Fig. 33-6A), and the proximal straight tubule (Fig. 33-6B). However, based on ultrastructure, the proximal tubule can alternatively be subdivided into three segments: S1, S2, and S3. The S1 segment starts at the glomerulus and includes the first portion of the proximal convoluted tubule. The S2 segment starts in the second half of the proximal convoluted tubule and continues into the first half of the proximal straight tubule. Finally, the S3 segment includes the distal half of the proximal straight tubule that extends into the medulla. (See Note: Tubule Segments of the Nephron)


Figure 33-6 Structure of tubule cells along the nephron. Because of the variability among tubule segments, the cross sections of the tubule are not to scale.

Table 33-1 Tubule Segments of the Nephron

Tubule Segment


Proximal convoluted tubule


Proximal straight tubule


Thin descending limb of loop of Henle


Thin ascending limb of loop of Henle


Thick ascending limb of loop of Henle


Distal convoluted tubule


Connecting tubule


Initial collecting tubule


Cortical collecting tubule


Outer medullary collecting duct


Inner medullary collecting duct


Both the apical (luminal) and basolateral (peritubular) membranes of proximal tubule cells are extensively amplified (Fig. 33-6A, B). The apical membrane has infoldings in the form of a well-developed brush border. This enlargement of the apical surface area correlates with the main function of this nephron segment, namely, to reabsorb the bulk of the filtered fluid back into the circulation. A central cilium, which may play a role in sensing fluid flow, protrudes from the apical pole of proximal tubule cells and nearly all tubule cells.

The basolateral membranes of adjacent proximal tubule cells form numerous interdigitations, bringing abundant mitochondria in close contact with the plasma membrane. The interdigitations of the lateral membranes also form an extensive extracellular compartment bounded by the tight junctions at one end and by the basement membrane of the epithelium at the other end. Proximal tubule cells contain lysosomes, endocytic vacuoles, and a well-developed endoplasmic reticulum. Proximal tubule cells are also characterized by a prominent Golgi apparatus (see Chapter 2), which is important for synthesizing many membrane components, sorting them, and targeting them to specific surface sites. From the S1 to the S3 segments, cell complexity progressively declines, correlating with a gradual decrease of reabsorptive rates along the tubule. Thus, the cells exhibit a progressively less developed brush border, diminished complexity of lateral cell interdigitations, a lower basolateral cell membrane area, and a decrease in the number of mitochondria.

In comparison with the S3 segment of the proximal tubule, the cells lining the descending and ascending thin limbs of the loop of Henle are far less complex (Fig. 33-6C, D), with few mitochondria and little cell membrane amplification. In superficial nephrons, the thin ascending limbs are extremely abbreviated (Fig. 33-2). However, they form a major part of the long loops of the juxtamedullary nephrons.

Epithelial cells lining the thick ascending limb of the loop of Henle, which terminates at the macula densa, are characterized by tall interdigitations and numerous mitochondria within extensively invaginated basolateral membranes (Fig. 33-6E). This complex cell machinery correlates with the key role these cells play in making the medullary interstitium hyperosmotic.

Until the latter part of the 20th century, morphologists defined the classic distal tubule—on the basis of light microscopic studies—as the nephron segment stretching from the macula densa to the first confluence of two nephrons in the collecting duct system. Today, we subdivide the classic distal tubule into three segments, based on ultrastructural studies: the distal convoluted tubule (starting at the macula densa), the connecting tubule, and the initial collecting tubule. What was classically termed the early distal tubule is mainly the distal convoluted tubule, whereas the classically termed late distal tubule is mainly the initial collecting tubule.

The distal convoluted tubule begins at the macula densa and ends at the transition to the connecting tubule (Fig. 33-6F). The cells of the distal convoluted tubule are similar in structure to those of the thick ascending limb. However, significant cell heterogeneity characterizes the tubule segments that follow.

The connecting tubule, which ends at the transition to the initial collecting tubule, consists of two cell types: connecting tubule cells and intercalated cells. Connecting tubule cells (Fig. 33-6G) are unique in that they produce and release renal kallikrein, a local hormone whose precise function is still uncertain. We discuss intercalated cells later.

The two segments following the connecting tubule, the initial collecting tubule (up to the first confluence) and the cortical collecting tubule (after the confluence), are identical. They are composed of intercalated and principal cells, which exhibit striking morphological and functional differences. Intercalated cells, similar in structure to the intercalated cells of the connecting tubule, make up about one third of the lining of these collecting tubule segments (Fig. 33-6H, I). They are unusual among tubule cells in that they lack a central cilium. One subpopulation of these cells (A- or α-intercalated cells) secretes H+ and reabsorbs K+, whereas another (B- or ß-intercalated cells) secretes HCO3Principal cells make up about two thirds of the cells of the initial collecting tubule and cortical collecting tubule (Fig. 33-6H, I). Compared with intercalated cells, principal cells have fewer mitochondria, only modestly developed invaginations of the basolateral membrane, and a central cilium on the apical membrane. Principal cells reabsorb Na+ and Cl and secrete K+.

The medullary collecting duct is lined mostly by one cell type that increases in cell height toward the papilla (Fig. 33-6J, K). The number of intercalated cells diminishes, beginning at the outer medullary collecting duct. Cells in this segment continue the transport of electrolytes and participate in the hormonally regulated transport of water and of urea. At the extreme end of the medullary collecting duct (i.e., the “papillary” collecting duct or duct of Bellini), the cells are extremely tall.

The tightness of tubule epithelia increases from the proximal to the medullary collecting tubule

Epithelia may be either “tight” or “leaky,” depending on the permeability of their tight junctions (see Chapter 5). In general, the tightness of the tubule epithelium increases from the proximal tubule to the collecting duct. In the leaky proximal tubule, junctional complexes are shallow and, in freeze-fracture studies, show only a few strands of membrane proteins (see Chapter 2). In contrast, in the relatively tight collecting tubule, tight junctions extend deep into the intercellular space and consist of multiple strands of membrane proteins. Tubule segments with tight junctions consisting of only one strand have low electrical resistance and high solute permeability, whereas tubules with several strands tend to have high electrical resistance and low permeability.

Gap junctions (see Chapter 6) provide low-resistance pathways between some, but not all, neighboring tubule cells. These gap junctions are located at various sites along the lateral cell membranes. Electrical coupling exists among proximal tubule cells, but not among heterogenous cell types, such as those found in the connecting and collecting tubules.


The nephron forms an ultrafiltrate of the blood plasma and then selectively reabsorbs the tubule fluid or secretes solutes into it

As they do for capillaries elsewhere in the body, Starling forces (see Chapter 20) govern the flow of fluid across the capillary walls in the glomerulus and result in net filtration. However, in the case of the glomerular capillaries, the filtrate flows not into the interstitium, but into Bowman’s space, which is contiguous with the lumen of the proximal tubule.

The main function of renal tubules is to recover most of the fluid and solutes filtered at the glomerulus. If the fluid were not recovered, the kidney would excrete the volume of the entire blood plasma in less than half an hour. The retrieval of the largest fraction of glomerular filtrate occurs in the proximal tubule, which reabsorbs NaCl, NaHCO3, filtered nutrients (e.g., glucose and amino acids), divalent ions (e.g., Ca2+, HPO2−4, and SO2−4), and water. Finally, the proximal tubule secretes NH+4 and a variety of endogenous and exogenous solutes into the lumen.

The main function of the loop of Henle (i.e., thin descending limb of loop of Henle [tDLH], thin ascending limb of loop of Henle [tALH], thick ascending limb of loop of Henle [TAL]) is to participate in forming concentrated or dilute urine. The loop does this by pumping NaCl into the interstitium of the medulla without appreciable water flow, thus making the interstitium hypertonic. Downstream, the medullary collecting duct exploits this hypertonicity by either permitting or not permitting water to flow by osmosis into the hypertonic interstitium. In humans, only ~15% of the nephrons, the juxtamedullary nephrons, have long loops that descend to the tip of the papilla. Nevertheless, this subpopulation of nephrons (Fig. 33-2) is extremely important for creating the osmotic gradients within the papilla that allow water movement out of the lumen of the entire population of medullary collecting ducts. As a result of this water movement, urine osmolality in the collecting ducts can far exceed that in the plasma.

TAL cells secrete the Tamm-Horsfall glycoprotein (THP). Normal subjects excrete 30 to 50 mg/day into the urine, thus accounting—along with albumin (<20 mg/day)—for most of the protein normally present in urine. THP adheres to certain strains of Escherichia coli and may be part of the innate defense against urinary tract infections. THP also constitutes the matrix of all urinary casts, defined as cylindrical debris in the urine that takes the shape of the tubule lumens in which it is formed. (See Note: Tamm-Horsfall Protein)

The classic distal tubule and the collecting duct system perform the fine control of NaCl and water excretion. Although only small fractions of the glomerular filtrate reach these most distally located nephron sites, these tubule segments are where several hormones (e.g., aldosterone, arginine vasopressin) exert their main effects on electrolyte and water excretion.

The juxtaglomerular apparatus is a region where each thick ascending limb contacts its glomerulus

Elements of the JGA play two important regulatory roles. First, if the amount of fluid and NaCl reaching a nephron’s macula densa (Fig. 33-3F) increases, the glomerular filtration rate (GFR) of that nephronfalls. We discuss this phenomenon of tubuloglomerular feedback in Chapter 34.

The second regulatory mechanism comes into play during a decrease in the pressure of the renal artery feeding the various afferent arterioles. When the afferent arteriole senses decreased stretch in its wall, this baroreceptor (see Chapter 23) directs neighboring granular cells to increase their release of renin into the general circulation. We discuss the renin-angiotensin-aldosterone axis, which is important in the long-term control of systemic arterial blood pressure, in Chapter 40.

Sympathetic nerve fibers to the kidney regulate renal blood flow, glomerular filtration, and tubule reabsorption

The autonomic innervation to the kidneys is entirely sympathetic; the kidneys lack parasympathetic nerve fibers. The sympathetic supply to the kidneys originates from the celiac plexus (see Fig. 14-3), and it generally follows the arterial vessels into the kidney. The varicosities of the sympathetic fibers release norepinephrine and dopamine into the loose connective tissue near the smooth muscle cells of the vasculature (i.e., renal artery as well as afferent and efferent arterioles) and near the proximal tubules. Sympathetic stimulation to the kidneys has three major effects. First, the catecholamines cause vasoconstriction. Second, the catecholamines strongly enhance Na+ reabsorption by proximal tubule cells. Third, as a result of the dense accumulation of sympathetic fibers near the granular cells of the JGA, increased sympathetic nerve activity dramatically stimulates renin secretion.

Renal nerves also include afferent (i.e., sensory) fibers. A few myelinated nerve fibers conduct baroreceptor and chemoreceptor impulses that originate in the kidney. Increased perfusion pressure stimulates renal baroreceptors in the interlobular arteries and afferent arterioles. Renal ischemia and abnormal ion composition of the interstitial fluid stimulate chemoreceptors located in the renal pelvis. These pelvic chemoreceptors probably respond to high extracellular levels of K+ and H+ and may elicit changes in capillary blood flow.

The kidneys are also endocrine organs: they produce renin, the biologically active form of vitamin D, erythropoietin, prostaglandins, and bradykinin

Besides renin production by the JGA granular cells (see Chapter 40), the kidneys play several other endocrine roles. Proximal tubule cells convert circulating 25-hydroxyvitamin D to the active metabolite, 1, 25-dihydroxyvitamin D. This hormone controls Ca2+ and phosphorus metabolism by acting on the intestines, kidneys, and bone (see Chapter 52), and is important for developing and maintaining bone structure.

Fibroblast-like cells in the interstitium of the cortex and outer medulla secrete EPO in response to a fall in the local tissue PO2 (see p. 453). EPO stimulates the development of red blood cells by action on hematopoietic stem cells in bone marrow. In chronic renal failure, the deficiency of EPO leads to severe anemia that can be treated with recombinant EPO. (See Note: Erythropoietin)

The kidney releases prostaglandins and several kinins, paracrine agents that control circulation within the kidney. These substances are generally vasodilators and may play a protective role when renal blood flow is compromised. Tubule cells also secrete angiotensin, bradykinin, cAMP, and ATP into the lumen, but the precise function of such local secretion is poorly understood.


Numerous tests can assess renal function. Some are applicable only in animal experiments. Others are useful in clinical settings and fall into two general categories: (See Note: In Vitro Preparations for Studying Renal Function in the Research Laboratory)

1. Modern imaging techniques provide outstanding macroscopic views of renal blood flow, filtration, and excretory function.

2. Measurements of so-called renal clearance of various substances evaluate the ability of the kidneys to handle solutes and water.

This section focuses on clearance measurements, which compare the rate at which the glomeruli filter a substance (water or a solute) with the rate at which the kidneys excrete it into the urine. By measuring the difference between the amounts filtered and excreted for a particular substance, we can estimate the net amount reabsorbed or secreted by the renal tubules and can thus gain insight into the three basic functions of the kidney: glomerular filtration, tubule reabsorption, and tubule secretion. Although widely used, clearance methods have the inherent limitation that they measure overall nephron function. This function is overall in two different senses. First, clearance sums many individual transport operations occurring sequentially along a nephron. Second, clearance sums the output of all 2 million nephrons in parallel. Hence, clearance cannot provide information on precise sites and mechanisms of transport. Such information can, however, come from studies of individual nephrons, tubule cells, or cell membranes. One can also apply the clearance concept to other problems, such as clearance of bile by the liver (see Chapter 46) or clearance of hormones from the blood.

The clearance of a solute is the virtual volume of plasma that would be totally cleared of a solute in a given time

All solutes excreted into the urine ultimately come from the blood plasma perfusing the kidneys. Thus, the rate at which the kidney excretes a solute into the urine equals the rate at which the solute disappears from the plasma, provided the kidney does not produce, consume, or store the solute. Imagine that, in 1 minute, 700 mL of plasma will flow through the kidneys. This plasma contains 0.7 L × 142 mM or ~100 mmol of Na+. Of this Na+, the kidneys remove and excrete into the urine only a tiny amount, ~0.14 mmol. In principle, these 0.14 mmol of Na+ could have come from only 1 mL of plasma, had all Na+ ions been removed (i.e., cleared) from this volume. The clearance of a solute is defined as the virtual volume of blood plasma (per unit time) needed to supply the amount of solute that appears in the urine. Thus, in our example, Na+ clearance was 1 mL/min, even though 700 mL of plasma flowed through the kidneys.

Renal clearance methods are based on the principle of mass balance and the special anatomy of the kidney (Fig. 33-7). For any solute (X) that the kidney does not synthesize, degrade, or accumulate, the only route of entry to the kidney is the renal artery, and the only two routes of exit are the renal vein and the ureter. Because the input of X equals the output of X,



Figure 33-7 Solute mass balance in the kidney. See text for details.

PX,a and PX,v are the plasma concentrations of X in the renal artery and renal vein, respectively. RPFa and RPFv are the rates of renal plasma flow (RPF) in the renal artery and vein, respectively. UX is the concentration of X in urine. image is urine flow (the overdot represents the time derivative of volume). The product UX · image is the urinary excretion rate, the amount of X excreted in urine per unit time.

In developing the concept of renal clearance, we transform Equation 33-1 in two ways, both based on the assumption that the kidneys clear all X from an incoming volume of arterial plasma. First, we replace RPFa with the inflow of the virtual volume—the clearance of X (CX)—that provides just that amount of X that appears in the urine. Second, we assign the virtual venous output a value of zero. Thus, Equation 33-1 becomes the following:


Solving for clearance,


This is the classic clearance equation that describes the virtual volume of plasma that would be totally cleared of a solute in a given time (Table 33-2A). We need to know only three parameters to compute the clearance of a solute X: (See Note: Symbols of the Clearance Equation)

1. the concentration of X in the urine (UX);

2. the volume of urine formed in a given time (image); and

3. the concentration of X in systemic blood plasma (PX), which is the same as PX,a in Equation 33-1.

Table 33-2 Renal Clearance


Together, the three basic functions of the kidney—glomerular filtration, tubule reabsorption, and tubule secretion—determine the renal clearance of a solute. In the special case in which the kidneys completely clear X from plasma during a single passage through the kidneys (PX,v = 0 in Equation 33-1), the renal clearance of X equals RPFa in Equation 33-1. Because p-aminohippurate (PAH) is just such a special solute, its clearance is a good estimate of RPFa, which we simplify to RPF (Table 33-2B). We discuss RPF in Chapter 34.

For all solutes that do not behave like PAH, the renal venous plasma still contains some X. Thus, the virtual volume cleared of X in a given time is less than the total RPF. For most solutes, then, clearance describes a virtual volume of plasma that would be totally cleared of a solute, whereas in reality a much larger volume of plasma is partially cleared of the solute.

We can use a clearance approach to estimate another important renal parameter: GFR, which is the volume of fluid filtered into Bowman’s capsule per unit time. Imagine a solute X that fulfills two criteria. First, X is freely filtered (i.e., concentration of X in Bowman’s space is the same as that in blood plasma). Second, the tubules do not absorb, secrete, synthesize, degrade, or accumulate X. Thus, the amount of X that appears in the urine per unit time (UX · image) is the same as the amount of X that the glomerulus filters per unit time (PX · GFR):


The input to Bowman’s space is also known as the filtered solute load and is generally given in millimoles (or milligrams) per minute. Rearranging Equation 33-4.


Equation 33-5 is in exactly the same form as the classic clearance equation (Equation 33-3). In other words, GFR is CX if X has the required properties. As discussed in Chapter 34inulin is just such a solute (Table 33-2C).

A solute’s urinary excretion is the algebraic sum of its filtered load, reabsorption by tubules, and secretion by tubules

The homeostasis of body fluids critically depends on the ability of the kidneys to determine the amount of a given solute that they excrete into the urine. Renal excretion rate (EX) depends on three factors (Fig. 33-8):

1. the rate of filtration of X (FX): the filtered load (FX = GFR · PX);

2. the rate of reabsorption of X (RX) by the tubules; and

3. the rate of secretion of X (SX) by the tubules.


Figure 33-8 Factors contributing to the net urinary excretion of a substance.

This interrelationship is expressed quantitatively as follows:


For some substances (e.g., inulin), no reabsorption or secretion occurs. For most substances, either reabsorption or secretion determines the amount present in the final urine. However, for some substances, both reabsorption andsecretion determine excretion.

If a solute is only reabsorbed, but not secreted, we can rearrange Equation 33-6 to obtain the rate of reabsorption Equation 33-6:


Conversely, if a solute is only secreted, but not reabsorbed, the rate of secretion is as follows:


In applying Equations 33-7 and 33-8, we must keep in mind two important limitations. First, to estimate the rate at which a substance appears in the filtrate—the filtered load—from the product GFR · PX, we assume that PX is the freely filterable concentration of X. Indeed, many substances, particularly univalent electrolytes, urea, glucose, and amino acids, are freely filterable. However, if the solute binds to protein, for example, then it will not be freely filterable. For these solutes, including Ca2+, phosphate, Mg2+, and PAH, it is necessary to measure plasma binding and correct for the nonfilterable fraction of the solute. Second, for us to apply the mass balance equation (Equation 33-6), the kidney must not synthesize, degrade, or accumulate the solute. An example of a solute that is synthesized by the kidney is ammonium. Examples of solutes degraded by the kidney include glutamine and glutamate (which are deaminated to yield ammonium) as well as several other amino acids and monocarboxylic and dicarboxylic acids.

When the kidney both reabsorbs and secretes a substance, clearance data are inadequate to describe renal handling. For example, if the proximal tubule completely reabsorbed a solute that a later segment then secreted, clearance data alone would imply that only filtration and some reabsorption occurred. We would have no reason to implicate secretion. Complex combinations of reabsorption and secretion occur with K+, uric acid, and urea.

Another useful parameter for gauging how the kidney handles a freely filtered solute is the fractional excretion (FE), which is the ratio of the amount excreted in the urine (UX · image) to the filtered load (PX · GFR):


According to Equation 33-3, however, the term (UX • image/PX) is simply the clearance of X (CX):


As discussed in Chapter 34, one estimates GFR by measuring the clearance of inulin (CIn). Thus, the fractional excretion of a freely filterable solute is the same as the clearance ratio:


Microscopic techniques make it possible to measure single-nephron rates of filtration, absorption, and secretion

Because clearance methods treat the kidney as a “black box,” reflecting the activity of many single nephrons and nephron segments, it is very difficult to determine which nephron segments are responsible for which transport processes. It is also impossible to determine which nephrons are responsible for overall urinary excretion. To learn how single nephrons function, and to understand how individual nephron segments contribute to overall nephron function, physiologists developed a series of invasive techniques for studying renal cells in the research laboratory (Fig. 33-9). (See Note: In Vitro Preparations for Studying Renal Function in the Research Laboratory)


Figure 33-9 Methods for studying renal function in the research laboratory.

To apply the concept of clearance to a single nephron site, one uses the free-flow micropuncture approach (Fig. 33-9A) and measures the concentration of the solute in the tubule fluid at that site (TFX), volume flow at that site (i.e., the collection rate), and plasma concentration (PX). By analogy with the macroscopic clearance equation, we can write a clearance equation for a single nephron:


Compared with Equation 33-3, here TFX replaces UX and “Volume collection rate” replaces image. We can use this basic equation to compute the amount of fluid that a single nephron filters, as well as the amounts of fluid and solutes that a single tubule segment handles.

Single-Nephron Glomerular Filtration Rate If X in Equation 33-12 is a marker for GFR (e.g., inulin or In), one can calculate single-nephron GFR (SNGFR) using an equation that is similar to that in Table 33-2C to calculate total GFR:


Using the numeric values for rat kidney shown in Figure 33-10, we can use Equation 33-13 to compute the SNGFR:



Figure 33-10 Measurement of single-nephron glomerular filtration rate. Data are for the rat.

Handling of Water by Tubule Segments In a Single Nephron We also can use the same information that we used to compute SNGFR to calculate the rate of water reabsorption between the glomerulus and the micropuncture site. The fraction of filtered water remaining at the micropuncture site is as follows:


Substituting the expression for SNGFR (Equation 33-13) into Equation 33-15, we have the following:


Thus, to know the fraction of filtered water remaining at the collection site, we do not need to know the collection rate, only the concentrations of inulin in blood plasma and at the collection site. In the example of Figure 33-10, in which the tubule reabsorbed two thirds of the fluid, the fraction of filtered water remaining at the sampling site is (1 mg/mL)/(3 mg/mL) or ~0.33.

The fraction of filtered water reabsorbed is 1 minus the fraction of filtered water remaining:


In our example, the fraction of filtered water reabsorbed is 1 − 0.33 or ~0.67.

Handling of Solutes by Tubule Segments in a Single Nephron We can use the same concepts of single-nephron clearance to quantitate the reabsorption or secretion of any solute along the tubule. The first step is to estimate the fraction of the filtered solute remaining at the puncture site. This parameter—the fractional solute delivery—is the ratio of the amount of solute appearing at the micropuncture site to the amount of solute filtered at the glomerulus (i.e., single-nephron filtered load):


The numerator is the product of volume collection rate and tubule solute concentration (TFX), and the denominator is the product of SNGFR and the plasma solute concentration (PX):


In Equations 33-15 and 33-16, we saw that the ratio (Volume collection rate/SNGFR) is (PIn/TFIn). Making this substitution in Equation 33-19, we obtain an alternative expression for fractional solute excretion:


The advantage of Equation 33-20 over Equation 33-19 is that no measurements of collection rates are required.

Equation 33-20 is important for understanding the transport of a solute along the nephron. If (TFX/PX)/(TFIn/PIn) is greater than 1, secretion has occurred. Merely observing that the solute concentration in tubule fluid increases along the length of the nephron does not necessarily mean that the tubule secreted the solute; the concentration would also increase if water were reabsorbed. We can conclude that secretion has occurred only if the solute concentration in tubule fluid, relative to its concentration in filtrate, exceeds the concentration ratio of inulin. If (TFX/PX)/(TFIn/PIn) is less than 1, reabsorption has occurred.

Rather than referring to the fraction of the solute excreted (i.e., remaining) up to the point of the micropuncture site, we can also refer to the fractional solute reabsorption at the puncture site. By analogy to Equation 33-17 for water, this parameter for a solute is merely 1 minus the fractional solute excretion:


Thus, applying the principles of clearance and mass balance to a single nephron, we can calculate SNGFR as well as the fractional reabsorption of water and solutes at the micropuncture site.


As discussed in Chapters 44 and 45, the epithelium of the gastrointestinal system continues to modify the contents of the gastrointestinal tract up until the point where the contents finally exit the body. The situation is very different in the mammalian urinary system. By the time the fluid leaves the most distal portion of the collecting duct, the fluid has the composition of final urine. Thus, the renal pelvis, ureters, bladder, and urethra do not substantially modify the urine volume or composition.

The ureters propel urine from the renal pelvis to the bladder by peristaltic waves conducted along a syncytium of smooth muscle cells

The ureters serve as conduits for the passage of urine from the renal pelvis into the urinary bladder (Fig. 33-1A). Located in the retroperitoneum, each ureter loops over the top of the common iliac artery and vein on the same side of the body and courses through the pelvis. The ureters enter the lower, posterior portion of the bladder (ureterovesical junction), pass obliquely through its muscular wall, and open into the bladder lumen 1 to 2 cm above, and lateral to, the orifice of the urethra (Fig. 33-11A). The two ureteral orifices, connected by a ridge of tissue, and the urethral orifice form the corners of a triangle (bladder trigone). A flap-like valve of mucous membrane covers each ureteral orifice. This anatomical valve, in conjunction with the physiological valve-like effect created by the ureter’s oblique pathway through the bladder wall, prevents reflux of urine back into the ureters during contraction of the bladder.


Figure 33-11 A and B, Anatomy of the ureters and bladder. In B, ureteral smooth muscle cells generally have a resting membrane potential of ~–60 mV, mainly determined by a high K+ membrane permeability. Na+ channels speed the upstroke of the action potential, although Ca2+channels are mainly responsible for the action potential.

The lumen of each ureter is lined by transitional epithelium, which is above a submucosal layer of connective tissue, as well as an inner longitudinal and an outer circular layer of smooth muscle. Ureteral smooth musclefunctions as a syncytium and is thus an example of unitary smooth muscle (see Chapter 9). Gap junctions (see Chapter 6) conduct electrical activity from cell to cell at a velocity of 2 to 6 cm/s. Chemical or mechanical stimuli (e.g., stretch) or a suprathreshold membrane depolarization may trigger an action potential (Fig. 33-11B) of the plateau type (see Chapter 9).

Contraction of ureteral smooth muscle is similar to that of other smooth muscle (see Chapter 9), in which Ca2+-calmodulin activates myosin light chain kinase (MLCK). cAMP-dependent protein kinase (PKA) can phosphorylate MLCK, thereby lowering the affinity of MLCK for Ca2+-calmodulin and impairing phosphorylation of myosin light chains. This mechanism may, at least in part, account for the relaxing effect of cAMP on smooth muscle.

Ureteral peristaltic waves originate from electrical pacemakers in the proximal portion of the renal pelvis. These waves propel urine along the ureters and into the bladder in a series of spurts, at frequencies of two to six per minute. The intraureteral hydrostatic pressure is 0 to 5 cm H2O at baseline, but it increases to 20 to 80 cm H2O during peristaltic waves. Blockade of ureteral outflow to the bladder, as by a kidney stone, causes the ureter to dilate and increases the baseline hydrostatic pressure to 70 to 80 cm H2O over a period of 1 to 3 hours. This pressure is transmitted in retrograde fashion to the nephrons and thus creates a stopped-flow condition in which glomerular filtration nearly comes to a halt. Hydronephrosis, dilation of the pelvis and calyces of the kidney, can evolve over hours to days. Patients complain of severe pain (renal colic) resulting from distention of involved structures. If not cleared, the obstruction can cause marked renal dysfunction and even acute renal failure. With persistent obstruction, the pressure inside the ureter declines to a level that is only slightly higher than the normal baseline. Even though the patient produces no urine (anuria), glomerular filtration continues, albeit at a markedly reduced rate, a condition reflecting a balance between filtration and fluid reabsorption by the tubules.

Although ureteral peristalsis can occur without innervation, the autonomic nervous system can modulate peristalsis. As in other syncytial smooth muscle, autonomic control of the ureters occurs by diffuse transmitter release from multiple varicosities formed as the postganglionic axon courses over the smooth muscle cell. Sympathetic input (through aortic, hypogastric, and ovarian or spermatic plexuses) modulates ureteral contractility as norepinephrine acts by excitatory α-adrenergic receptors and inhibitory ß-adrenergic receptors. Parasympathetic input enhances ureteral contractility through acetylcholine, either by directly stimulating muscarinic cholinergic receptors (see Chapter 3) or by causing postganglionic sympathetic fibers to release norepinephrine, which then can stimulate α-adrenoceptors. Some autonomic fibers innervating the ureters are afferent pain fibers. In fact, the pain of renal colic associated with violent peristaltic contractions proximal to an obstruction is one of the most severe encountered in clinical practice.

Sympathetic, parasympathetic, and somatic fibers innervate the urinary bladder and its sphincters

The urinary bladder consists of a main portion (body) that collects urine and a funnel-shaped extension (neck) that connects with the urethra (Fig. 33-11A). Transitional epithelium lines the bladder lumen. Three poorly defined layers of smooth muscle make up the bulk of the bladder wall, the so-called detrusor muscle. At the lower tip of the trigone, the bladder lumen opens into the posterior urethra (i.e., distal part of bladder neck), which extends over 2 to 3 cm. The wall of the posterior urethra contains smooth muscle fibers of the detrusor muscle interspersed with elastic tissue, together forming the internal sphincter (Table 33-3). Immediately adjacent is the external sphincter, made up of voluntary, mainly slow-twitch striated-muscle fibers.

Table 33-3 Overview of the Urethral Sphincters


Internal Sphincter

External Sphincter

Type of muscle



Nerve reaching the structure



Nature of innervation



In humans, bladder smooth muscle appears to lack gap junctions, a finding suggesting the absence of electrotonic coupling between cells. Thus, bladder smooth muscle is probably “multiunit” (see Chapter 9), with a 1:1 ratio between nerve endings and smooth muscle cells. Contraction of bladder smooth muscle is typical of other smooth muscle cells.

The bladder and sphincters receive sympathetic and parasympathetic (autonomic) as well as somatic (voluntary) innervation (Fig. 33-12). The sympathetic innervation to the bladder and internal sphincter arises from neurons in the intermediolateral cell column of the tenth thoracic to the second lumbar spinal cord segment (see Chapter 14). The preganglionic fibers then pass through lumbar splanchnic nerves to the superior hypogastric plexus, where they give rise to the left and right hypogastric nerves. These nerves lead to the inferior hypogastric/pelvic plexus, where preganglionic sympathetic fibers synapse with postganglionic fibers. The postganglionic fibers continue to the bladder wall through the distal portion of the hypogastric nerve. This distal portion also contains the preganglionic parasympathetic axons discussed in the next paragraph.


Figure 33-12 Autonomic and somatic innervation of the bladder.

The parasympathetic innervation of the bladder originates from the intermediolateral cell column in segments S2 through S4 of the sacral spinal cord. The parasympathetic fibers approaching the bladder via the pelvic splanchnic nerve are still preganglionic. They synapse with postganglionic neurons in the body and neck of the urinary bladder.

The somatic innervation originates from motor neurons arising from segments S2 to S4. Through the pudendal nerve, these motor neurons innervate and control the voluntary skeletal muscle of the external sphincter (Fig. 33-12).

Filling the bladder activates stretch receptors that initiate the micturition reflex, a spinal reflex arc also under the control of higher central nervous system centers

Bladder tone is defined by the relationship between bladder volume and internal (intravesical) pressure. One can measure the volume-pressure relationship by first inserting a catheter through the urethra and emptying the bladder and then recording the pressure while filling the bladder with 50-mL increments of water. The record of the relationship between volume and pressure is a cystometrogram (Fig. 33-13, blue curve). Increasing bladder volume from 0 to ~50 mL produces a moderately steep increase in pressure. Additional volume increases up to ~300 mL produce almost no pressure increase; this high compliance reflects relaxation of bladder smooth muscle. At volumes higher than 400 mL, additional increases in volume produce steep rises in “passive” pressure. Bladder tone, up to the point of triggering the micturition reflex, is independent of extrinsic bladder innervation.


Figure 33-13 A cystometrogram.

Cortical and suprapontine centers in the brain normally inhibit the micturition reflex, which the pontine micturition center coordinates. The pontine micturition center controls both the bladder detrusor muscle and the urinary sphincters. During the storage phase, stretch receptors in the bladder send afferent signals to the brain through the pelvic splanchnic nerves. One first senses the urge for voluntary bladder emptying at a volume of ~150 mL and senses fullness at 400 to 500 mL. Nevertheless, until a socially acceptable opportunity to void presents itself, efferent impulses from the brain, in a learned reflex, inhibit presynaptic parasympathetic neurons in the sacral spinal cord that would otherwise stimulate the detrusor muscle. Voluntary contraction of the external urinary sphincter probably also contributes to storage.

Pathophysiology of Micturition

Lesions in the nervous system can lead to bladder dysfunction, the characteristics of which will depend on the site of the neural lesion. Three major classes of lesions can be distinguished.

1. Combined afferent and efferent lesions. Severing both afferent and efferent nerves initially causes the bladder to become distended and flaccid. In the chronic state of the so-called “decentralized bladder,” many small contractions of the progressively hypertrophied bladder muscles replace the coordinated micturition events. Although small amounts of urine can be expelled, a residual volume of urine remains in the bladder after urination.

2. Afferent lesions. When only the sacral dorsal roots (sensory fibers) are interrupted, reflex contractions of the bladder, in response to stimulation of the stretch receptors, are totally abolished. The bladder frequently becomes distended, the wall thins, and bladder tone decreases. However, some residual contractions remain because of the intrinsic contractile response of smooth muscle to stretch. As a rule, a residual urine volume is present after urination.

3. Spinal cord lesions. The effects of spinal cord transection (e.g., in paraplegic patients) include the initial state of spinal shock in which the bladder becomes overfilled and exhibits sporadic voiding (“overflow incontinence”). With time, the voiding reflex is re-established, but with no voluntary control. Bladder capacity is often reduced and reflex hyperactivity may lead to a state called “spastic neurogenic bladder.” Again, the bladder cannot empty completely, resulting in the presence of significant residual urine. Urinary tract infections are frequent because the residual urine volume in the bladder serves as an incubator for bacteria. In addition, during the period of “overflow incontinence,” before the voiding reflex is re-established, these patients have to be catheterized frequently, further predisposing to urinary tract infections.

The voiding phase begins with a voluntary relaxation of the external urinary sphincter, followed by the internal sphincter. When a small amount of urine reaches the proximal (posterior) urethra, afferents signal the cortex that voiding is imminent. The micturition reflex now continues as pontine centers no longer inhibit the parasympathetic preganglionic neurons that innervate the detrusor muscle. As a result, the bladder contracts, expelling urine. Once this micturition reflex has started, the initial bladder contractions lead to further trains of sensory impulses from stretch receptors, thus establishing a self-regenerating process (Fig. 33-13, red spikes moving to the left). At the same time, the cortical centers inhibit the external sphincter muscles. Voluntary urination also involves the voluntary contraction of abdominal muscles, which further raises bladder pressure and thus contributes to voiding and complete bladder emptying.

The basic bladder reflex that we have just discussed, although inherently an autonomic spinal cord reflex, may be either facilitated or inhibited by higher centers in the central nervous system that set the level at which the threshold for voiding occurs. Because of the continuous flow of urine from the kidneys to the bladder, the function of the various sphincters, and the nearly complete emptying of the bladder during micturition, the entire urinary system is normally sterile.


Books and Reviews

Benzing T: Signaling at the slit diaphragm. J Am Soc Nephrol 2004; 15:1382-1391.

Kriz W, Bankir L: A standard nomenclature for structures of the kidney. Kidney Int 1988; 33:1-7.

Seldin DW, Giebisch G (eds): The Kidney: Physiology and Pathophysiology, 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 2000.

Smith H: The Kidney: Structure and Function in Health and Disease. New York: Oxford University Press, 1951.

Weiss RW: Physiology and pharmacology of the renal pelvis and ureter. In Walsh PC, Rettig AB, Vaughan E, Wein AJ (eds): Campbell’s Urology, 7th ed, pp 839-870. Philadelphia: WB Saunders, 1998.

Journal Articles

Maxwell PH, Osmond MK, Pugh CW, et al: Identification of the renal erythropoietin-producing cells using transgenic mice. Kidney Int 1993; 44:1149-1162.

Shannon JA: The excretion of inulin by the dog. Am J Physiol 1935; 112:405-413.

Smith HW, Finklestein N, Aliminosa L, et al: The renal clearances of substituted hippuric acid derivatives and other aromatic acids in dog and man. J Clin Invest 1945; 24:388-404.

Tamm I, Horsfall FL Jr: A mucoprotein derived from human urine which reacts with influenza, mumps, and Newcastle disease viruses. J Exp Med 1952; 95:71-97.

Walker AM, Bott PA, Oliver J, MacDowell MC: The collection and analysis of fluid from single nephrons of the mammalian kidney. Am J Physiol 1941; 134:580-595.