Renal Physiology

Renal physiology of the mouse

Pierre Meneton, Iekuni Ichikawa, Tadashi Inagami, Jürgen Schnermann


As the transgenic and gene-targeting technology has become an invaluable experimental approach to study the function of gene products, the need has been expanded to assess the physiology in the mouse, which is virtually the only animal species to which that new genetic technology can apply. In this regard, renal physiologists have also received fruits of success from modern technology in several key areas, and areas are expanding in both depth and scope.

  • transgenic
  • gene targeting
  • kidney
  • sodium
  • blood pressure
  • micropuncture

over the last half century, a variety of techniques has been employed to study renal physiology and cardiovascular function in laboratory animals. The need to apply similar techniques to the mouse arose as transgenic and targeted gene null mutation (or gene-knockout) technologies became accessible. This is because transgenic rats have been produced in only a very few laboratories (40), and the targeted gene disruption is not yet applicable to rats. Even if a readily usable embryonic stem (ES) cell system for rats becomes available, only a very few investigators will be able to afford the higher cost of maintaining a rat breeding colony. Thus it is clear that investigators interested in genes related to or responsible for the regulation of cardiovascular and renal physiological functions must work with mice for the foreseeable future.

However, mice are much more sensitive to stress than rats and require a great deal of technical refinement for the determination of 10–20% changes in quantitative parameters such as blood pressure (BP) and cardiac output. Surgical manipulation to insert a cannula for a pressure transducer, vascular occlusion, and manipulations of various organs in survival surgery is more difficult in mice than in rats, due not only to their small size (note that body weight of a young adult mouse is 25 g, which is only one-tenth as large as rats of a similar age) but also to greater sensitivity of mice to surgical stress and infection. Gene null mutated mice are generated from two strains bearing different coat color to facilitate recognition of chimeric mice. For example, many, although not all, targeted gene deletions are done with embryonic stem cells derived from the 129 Ola strain and blastocysts from C57BL/6. Homozygous null mice thus have a variably mixed genetic background of the two strains. Fortunately, the basal BP levels of the 129 Ola and C57 BL/6 are very close (49), and F2, F3, or F4 crosses of 129 Ola and C57BL/6 may not have a very wide spread in BP; however, other properties can have spreads in the progenies. Preferably, therefore, gene null mice are backcrossed to either of their parental strains. Genetic background becomes essentially identical in 7–10 generations of backcrosses.

Balance and Clearance Studies

Investigators have documented balance and clearance studies in the mouse as early as the 1970s (14, 58-62). After this period, however, the mouse was essentially abandoned and replaced by the rat for the same experimental purposes. The success of gene-targeting technologies in the mouse has prompted investigators recently to return to this species. The requirement is the availability of metabolic cages small enough to collect the urine and feces excreted by each mouse in 12 or 24 h. For example, some commercial companies (Marty Technologie, Paris, France) produce a cage in which the mouse resides above a tapered glass ball so that the urine is effectively separated from the feces.

By using such cages, the daily excretion of sodium and potassium in the urine and feces was measured in wild-type mice and mice carrying a null mutation for the colonic isoform of the proton/potassium-ATPase, which plays an important role in potassium conservation (36). Potassium and sodium contents in the feces were determined in the supernatant after resuspension overnight in 0.75 N nitric acid. It was noted that the urinary excretion of potassium and sodium is 10 times greater than the fecal excretion in the mice fed a standard diet containing 1% sodium and 1% potassium (Fig. 1). It was also evident that the mutant mice lose more potassium in the feces than wild-type mice, but not in the urine. Under a potassium-free diet, potassium excretion is dramatically reduced in the urine and to a lesser extent in the feces. In this condition, the fecal excretion of potassium approaches the same order of magnitude as the urinary excretion, pointing to the fact that fecal ion excretion should always be taken into account in balance studies performed in animals that are maintained on a potassium- or sodium-poor diet. In the above experiment, it was found that the adult mice (outbred strain NSA) drink 4–7 ml of water and excrete 1–2 ml of urine/day. They eat 4–5 g of chow and excrete ∼0.75 g of feces/day (Table1).

Fig. 1.

Urinary (A) and fecal excretions (B) of potassium and sodium in wild-type (●) and mutant (○) NSA mice. Data are means ± SE of 3 different experiments, corresponding to a total of 14 wild-type and 16 mutant mice. Animals were fed a control diet for 9 days and then a potassium-free diet for 18 days (shaded area). Urinary volume and ion concentrations were determined every day for each animal. Feces were collected each day and pooled over a 3-day period before solubilization and ion concentration measurement. Inset: urinary potassium excretion between days 12 and 27, with a magnified scale on y-axis.

View this table:
Table 1.

Compilation of determinations of some plasma, urine, and feces parameters in mice fed a standard diet

Like rats, mice should be pair-fed in these experiments, because, even though they eat a near normal amount of a potassium- or sodium-free diet, they still have a tendency to lose weight compared with mice fed a normal diet. For example, the NSA mice lose ∼10% of their initial body weight after 3 wk on a potassium-free diet. Depending on the mouse strain, the results may differ, however, because some strains accept a sodium-free diet much more easily than others. Taste-preference experiments have shown that C57BL/6 mice have a strong aversion to sodium added to the drinking water at any concentration, whereas 129Sv mice prefer salty water vs. pure water up to a concentration of 150 mM sodium chloride (4). BALB/c mice seem to be intermediate between C57BL/6 and 129Sv mice (42).

Balance studies have been performed in mice fed a sodium-free or a high-salt (8% sodium) diet during a 1- to 2-wk period to compare wild-type mice to atrial natriuretic peptide transgenic mice (65) or null mutant mice (29). Both the transgenic (C3H strain) and the null (mixed C57BL/6 and 129Sv background) mice and their wild-type siblings ate the sodium-free or the high-salt diet well, as their body weights remained essentially unchanged, and they consumed similar quantities of sodium-free or high-salt food during the period. Results of the balance studies for the C3H wild-type mice are shown in Fig.2.

Fig. 2.

Daily dietary intake (○) and urinary excretion (●) of water, sodium, chloride, and potassium in wild-type C3H mice fed an 8% (A) or sodium-free (B) diet during a 2-wk period. Data are means ± SE; n = 11 mice for each diet. Intake and urinary excretion of water and ions were determined each day. Adapted from Ref. 65.

Strain differences have been described in mice for kidney weight (50), renal pathology (15), and function (24). The renal function study compared the clearances of sodium, potassium, osmolality, and water during the 12-h nocturnal period in male and female mice of 18 inbred strains and 6 F1 hybrids. No clear difference was found between the sexes in any strain, but large differences were found between strains; up to 6- to 10-fold differences for sodium, potassium, osmolar, and water clearances. The reason for these large differences has not been elucidated but could be explained by variations in the dietary habit and/or structure/function of the kidney. Thus the data suggest that the phenotype expression in renal clearances induced by a mutation may be affected substantially by the genetic background of the mouse strain. It is therefore desirable for the analysis of gene targeting to be performed in animals of multiple genetic backgrounds to verify the specific consequence of the mutation of the gene.

Renal function and morphology change with age, and this also depends on the mouse strain (25). The changes in renal clearance that occur from age 2–6 mo can be particularly variable; e.g., the clearance for sodium will increase twofold in the CBA strain but will decrease almost twofold in the B6 strain during this period.

Some investigators have performed balance studies in nonanesthetized mice for a period shorter than 12 h (34, 58-60, 62). Caution should be given to the circadian rhythms of urinary excretion that have been described for water, sodium, potassium, and creatinine. Like rats, mice are nocturnal animals that have higher rates of excretion during the night than during the day. The variations may be as high as 20% for creatinine, 30% for water, and 40–50% for sodium and potassium in Black Swiss mice (10). As most experiments are performed during the day, it should be noted that the physiology of the mouse during the day, i.e., resting, corresponds to the human physiology at night, if any, and not that during the day.

Clearance studies are especially useful in mice for determining glomerular filtration and renal blood flow rates. The determination of glomerular filtration rate in nonanesthetized mice has been performed by measuring the total clearance of a single injection of51Cr-ethylene-diaminetetraacetic acid into the tail vein (22, 23, 25, 48) or the clearance of endogeneous creatinine (12). The51Cr-ethylene-diaminetetraacetic acid-clearance method does not require urine collection and has the advantage of estimating the extracellular fluid volume by extrapolation of the initial volume of distribution of the tracer. However, the sampling of 50 μl of venous blood by puncture of the retrobulbar vein plexus every 10 min during 1 h is problematic due to the small total blood volume of the mouse (∼1.5 ml). The endogeneous creatinine-clearance method has the great advantage of avoiding any manipulation of the animals, but it is necessary to perform urine collection in metabolic cages, ideally over a 24-h period, to take advantage of time-averaged estimates. This method is also problematic in that blood sampling should be repeated many times during the 24-h period to take into account the circadian rhythm and variations in plasma creatinine levels due to feeding. The availability of analyzers (Biolyzer from Kodak, for example) that can measure the creatinine level in 10 μl of plasma may resolve this issue. Another potential problem relates to poor accuracy of creatinine measurements. Measured plasma creatinine values reported in the literature vary by a factor of 10 (Table 1). Rodents, and particularly mice, have high concentrations of noncreatinine chromogens in plasma and urine that can interfere with routine human clinical assays. As a consequence, these assays, which are based on the alkaline picrate Jaffé reaction, greatly overestimate creatinine concentration in the plasma and urine compared with high-performance liquid chromatography determination (38). Among the several modifications designed to improve the assay, Fuller's earth procedure gives a good estimate of creatinine concentration in the urine, but not in the plasma, which cannot be completely cleared of the noncreatinine chromogens (26, 39). Given the difficulty of performing routine high-performance liquid chromatography determinations, an interesting alternative is represented by the enzymatic method on the basis of the use of creatinase and sarcosine oxidase, which gives a good estimate of creatinine concentration in the urine and to a lesser extent in the plasma (39). Furthermore, the enzymatic method is available either as a kit (17) or in a dry-slide assay (56, 63). Several investigators have recently reported that 14C-labeled inulin administered via an osmotic pump and 24-h urine collections can also be used for the chronic measurement of glomerular filtration rate (33a).

As for the clearances of sodium, potassium, osmolality, and water, there are large differences in the glomerular filtration rate between mouse strains (22, 23). Even within one strain, values are highly dependent on ages and weights. The normalization of glomerular filtration rate, and more generally of clearances, is, therefore, an important issue in comparing data from one study to those in another. The investigation of 18 inbred and 6 F1 hybrid mouse strains has shown that glomerular filtration rate correlates better with body weight than with kidney weight (22, 23). This was also shown for the clearances of sodium, potassium, osmolality, and water (24).

Determination of effective renal blood flow rate by measurement of125I- or 3H-para-amino hippurate clearance after an intravenous injection appears feasible in nonanesthetized mice (33) but has the same potential problems that have been described above for the determination of glomerular filtration rate.

All the parameters characterizing renal function have also been measured in anesthetized mice. In this case, glomerular filtration rate is commonly measured by the clearance of 3H-,14C- or fluorescein isothiocyanate-labeled inulin, which is perfused intravenously throughout the experiment (11, 14, 18, 29, 31,36, 41, 60). Effective renal blood flow rate can be measured by the clearance of para-amino hippurate (11, 31), but recently the use of 0.5-mm V-series flow probes (Transonic Systems, Ithaca, NY) placed around the left renal artery has allowed the direct assessment of blood flow rate (1, 21). Extrapolation of data from anesthetized mice into those from animals in the conscious state requires caution as in other species, small or large.

The large variation in plasma potassium concentration observed in mice fed a standard diet (Table 1) may be related, at least in part, to mouse strain differences (14). These differences may be of physiological relevance, but they may also be due to red cell hemolysis during blood sampling, which artifactually raises the potassium level (66).

Measurements of Hormones Closely Linked to Renal Functions

The measurements of vasoactive peptides and enzymes in the mouse are confronted with the same problems that arise in this type of study in other species. Ideally, blood collection is performed in unanesthetized animals without euthanizing or stressing them. Rapid sampling of a small blood volume by puncture of the retrobulbar venous plexus seems to be the most efficient method as it provides the lowest values for renin concentration, which is very dependent on sympathetic tonus and stress. If possible, blood collection should be avoided in anesthetized animals because even if all the animals are tested “similarly” in one experiment, it is unclear whether the results can be extrapolated to physiological situations in vivo. For example, we have tested the effect of different commonly used anesthetics on renin secretion in the mouse. The results show that basal plasma renin concentration [930 ± 87 ng angiotensin I ⋅ mg 1 ⋅ h 1,n = 10] is increased 10-fold by ether [10,786 ± 987, n = 10], pentobarbital [9,868 ± 435,n = 10], inactin [8,403 ± 720, n = 10], but not by a mixture of ketamine-xylazine [1,250 ± 786, n = 10] at 7 and 0.4 mg/100 g body wt, respectively. The artifactual increase in renin secretion induced by the other anesthetics is accompanied by either a decrease, no change, or an increase in the arterial BP, depending on the anesthetic. The BP is only slightly reduced by the ketamine-xylazine mixture (10–15%). However, this anesthetic triggers a marked decrease in heart rate by some 50%.

Renin-Angiotensin-Aldosterone System

Plasma angiotensinogen concentration can be determined indirectly by measuring angiotensin I generation in the presence of an excess of semipurified mouse submaxillary renin. The concentrations of angiotensin I produced are usually determined by RIA (35). In general, plasma angiotensinogen concentration is higher in females than in males and is higher in one-renin-gene compared with two-renin-gene mouse strains. Typical values in the mouse (100–300 nmol/l) are much lower (10 times) than in humans. Plasma renin determination can be performed by measuring the generation of angiotensin I in the absence (renin activity) or in the presence of an excess of rat angiotensinogen (renin concentration). Of note, the conditions of the assay appear to be different between rat and mouse renins with an optimal pH of 6 and 8.5, respectively. At pH 6 or 7, the concentration of mouse renin is underestimated by a factor of two or three (5). Typically, plasma renin concentrations vary between 500 and 1,500 ng angiotensin I ⋅ ml 1 ⋅ h 1in one-renin-gene mouse strains and between 5,000 and 20,000 ng angiotensin I ⋅ ml 1 ⋅ h 1in two-renin-gene mouse strains. The values are uniformly higher in males than in females especially in the two-renin-gene mouse strains. One should be very cautious about the difference existing between one-renin (e.g., C57BL/6, BALB/c, C3H)- and two-renin (e.g., DBA/2,129Sv, Swiss, SWR)-gene mouse strains, especially when work is being done with knockout mice that usually carry a mixed background, 129Sv/C57BL6. In contrast to plasma angiotensinogen concentration, plasma renin concentration in the mouse is very high even in one-renin-gene strains compared with human values (at least 100 times more). However, plasma renin activity in the mouse is very similar to that in humans, i.e., in the range of 5–10 ng angiotensin I ⋅ ml 1 ⋅ h 1depending on the sex and the strain. It is interesting to note that there is no significant difference in plasma renin activity between one-renin- and two-renin-gene mouse strains. The renin-angiotensin system is very different between mouse and humans, even though the output of the system (renin activity) is similar; in the mouse this is achieved by a very high renin concentration and a very low angiotensinogen concentration, whereas the opposite is true in humans. For this reason, the assay of renin concentration is preferred over that of renin activity. Plasma angiotensin-converting enzyme (ACE) activity can be measured in the mouse by the spectrophotometric method of Cushman and Cheung (13), which is based on the hydrolysis of the synthetic substrate hippuryl-histidyl-leucine. Typical values are ∼400 nmol ⋅ ml 1 ⋅ min 1, which are on the same order of magnitude as in humans. As plasma renin activity is similar in the mouse and in humans, similar levels of plasma angiotensin II are expected in the two species (10–50 pg/ml) (55). Measurement of plasma angiotensin II is difficult in the mouse, as it is in other species. Indeed, the difficulty in controlling ACE and aminopeptidase activities during and after blood collection easily introduces artifactual variations in the measured levels. An additional problem in the mouse is the low sensitivity of the available RIAs for angiotensin II that necessitate the use of 0.5 ml of plasma (which implies the death of the animal) in contrast to the determination of angiotensinogen, renin, and ACE that can be performed on 10 μl of plasma.

Measurements of plasma aldosterone were done by using human RIAs in wild-type and angiotensinogen null mutant mice, which were given either a normal (0.46% Na, Purina Mills), low (0.02% Na)-, or high-sodium (3.15% Na) diet for 2 wk (45). In both wild-type and null mutant mice, plasma aldosterone concentration was low during the high-sodium diet, averaging 7 ± 4 and 19 ± 6 ng/dl, respectively, and remained so during the normal-sodium diet, averaging 36 ± 10 and 17 ± 3 ng/dl, respectively. Moreover, during the low-sodium diet, plasma aldosterone concentration was comparably elevated in wild-type and null mutant mice, averaging 757 ± 236 and 774 ± 229 ng/dl, respectively. Clearly, therefore, plasma aldosterone level is highly dependent on dietary sodium intake in mice, as documented in many other species. Of note, in null mice, plasma aldosterone increased just as markedly as in wild-type mice after 2 wk of dietary sodium restriction. Because these mutant mice are completely devoid of angiotensin II, it is clearly not essential for achieving the degree of hyperaldosteronism that is compatible with survival during dietary sodium restriction. To examine the significance of the marked hyperkalemia that developed in null mutants during sodium restriction, the effect of concurrent dietary sodium and potassium restriction (0.03% Na, 0.02% K) on aldosterone was tested. Interestingly, normalization of plasma potassium led to a fall in plasma aldosterone concentration to 47 ± 20 ng/dl. Overall, these studies revealed that the secondary hyperaldosteronism during extracellular fluid volume volume depletion is highly dependent on both angiotensin II and plasma potassium (Fig. 3) (44). The aldosterone values obtained are high, possibly due in part to the plasma glucocorticoid levels, which are much higher in mice than in humans. It is desirable that a specific RIA for measuring aldosterone in the mouse become available.

Fig. 3.

Correlation between plasma potassium concentration and aldosterone concentration simultaneously determined in Agt−/−mice (○) and Agt+/+ mice (●). Data collected from Agt−/− mice placed on 4 different sodium- and potassium-content diets were pooled. Data are from Agt−/− mice on 5-day low-sodium or low-sodium/potassium diet. Each point represents data from a single mouse. A significant correlation was noted between these 2 parameters in both of these animal groups.

Kallikrein-Kinin System

The kallikrein-kinin system is another example of an important vasoactive system that can be assessed in the mouse by using small plasma or urine volumes (<50 μl).

Plasma kininogens can be determined indirectly by measuring the generation of bradykinin after treatment with an excess of trypsin (47). The two kininogen genes that exist in the mouse species code for kininogens that appear to be sensitive to kallikrein and form bradykinin (P. Meneton, unpublished results). The generated bradykinin can be measured by using rat RIAs. Plasma kininogen levels in the mouse are found to be slightly higher than in the rat (∼10 μg bradykinin eq/ml). Interestingly, in contrast to the rat, there is no indication of the presence of acute-phase kininogen during inflammation in the mouse.

Tissue kallikrein is classically measured in urine by quantifying the generation of bradykinin in the presence of an excess of semipurified bovine kininogens (2). By using this method in the mouse, urinary kallikrein activity is found to be at least five times higher than in the rat (∼50 ng bradykinin ⋅ μl 1 ⋅ 30 min 1), reflecting a very active kallikrein-kinin system. A notable difference from other species is the insensitivity of mouse tissue kallikrein to aprotinin.

To our knowledge, bradykinin has not yet been measured in mouse plasma. Given the high kininogen concentration and tissue kallikrein activity, one would expect relatively high levels compared with in the rat. As in other species, the main problem for the determination of endogenous bradykinin is the excessively easy artifactual generation or degradation during blood collection, which makes the assay even more difficult than for angiotensin II. Urinary bradykinin has been recently measured in the mouse with reported values of ∼0.5 pmol/ml, which are indeed 30 times higher than in humans (16).

Atrial Natriuretic Factor (ANF)

Plasma level of ANF has also been measured in mice. In a study by Okubo et al. (43), mice were decapitated and truncal blood was collected promptly into tubes kept on ice in the presence of EDTA (10 mM). The plasma was separated and kept frozen at −70°C until measurement. A commercial RIA (RIK9103, Peninsula Laboratories, Belmont, CA) was used to analyze the concentration of immunoreactive (ir) ANF. Plasma irANF was measured directly from 25 μl plasma. Plasma ANF concentration was determined after 14 days of the three different dietary regimens specified above for aldosterone measurements in wild-type and angiotensinogen null mutant littermates. In both wild-type and mutant mice, plasma ANF concentration was significantly higher during a high- or normal-sodium diet compared with a low-sodium diet, a pattern found in many other species. Of interest, the value was higher in null than in wild-type mice on high (∼3-fold)- and normal (∼1.5-fold)-sodium diets. A further study revealed that the ANF expression is markedly activated in the ventricle of null mutant mice, a phenomenon uniformly seen in failing hearts.

BP Measurements

The method of BP measurement in small animals has been developed mostly for rats, due to availability of relatively uniform populations in large numbers, dependability in generating animal models of hypertension, and successful breeding of genetically hypertensive models of diverse etiology (9, 19, 28, 46). In the absence of absolutely correct BP reference standards, the methodological approach has been to demonstrate the concordance of BP values measured by several different methods. In essence, such agreement was seen between systolic BP values obtained by an indirect method using a tail-cuff sphygmomanometer and a direct method using an intra-arterial catheter connected to a pressure transducer. A generally good agreement was reported between these methods in the rat (9, 19, 28, 46), provided that inordinate stress like heating to 36°C, prolonged restraint, or a combination of the two (19) was eliminated.

The mouse has interesting characteristics in that many strains with varying basal BP have been identified and inbred by Schläger (49). These mice strains, derived from Swiss mice, seem to have been segregated over a long period, and their mean arterial pressure (MAP) seems to be different.

In general, the systolic BP of mice under various manipulations are somewhat lower than those of rats or dogs, seldom exceeding the 180 ± 190 mmHg level. Thus a MAP of 150 ± 155 mmHg seems to be the maximum that can be reached by a clip on the renal artery or deoxycorticosterone-salt administration, as reported by Johns et al. (30).

For the determination of BP in mice, the following methods are presently available: 1) the indirect method using a tail-cuff sphygmomanometer (30, 32); 2) a direct method using an intra-arterial indwelling catheter connected to a pressure transducer outside of the body (27, 30, 32); and 3) an intra-aortic indwelling catheter connected to a small pressure sensor and transmitter placed in the abdominal cavity in a telemetric system as summarized in Table 2.

View this table:
Table 2.

Comparison of 3 methods for blood pressure measurement in the mouse

These methods had been developed for rat and dog studies, validated, and well accepted in both (6, 8, 19, 32, 46, 64). To adapt these methods for use in mice, components of the BP-measurement system had to be miniaturized from those used for rats, which are 10 times as large as mice, although diameters of the aorta and carotid arteries are different by only three- to fourfold. We review below the technical problems and limitations of these methods in mice.

Indirect Tail-Cuff Sphygmomanometer Method

This noninvasive method has been most widely and readily used for rat studies. However, this method measures systolic pressure, and diastolic pressure is estimated with certain assumptions by subtracting pulse pressure from systolic pressure. Some serious problems of this method are the requirements of prewarming the animal in a restrainer, and training the animal for 3–7 days (30, 32). Unless these precautions are taken, the animals feel stress, and the BP reading tends to be higher than that by other methods. Repeated training will reduce this significantly. Also, a considerable technical improvement has been made for this method. Use of a light beam sensing device improved the sensitivity and requirement of prewarming. In an IITC (Woodland Hills, CA) instrument with (model 84) or without (model 179) a computer, the prewarming and measurement can be done at an ambient temperature of 29°C. For mouse BP measurement, a cylindrical restrainer (2.5 × 10 cm, model 84 IITC) is used for an animal of 20–40 g with a small tail cuff (1/4 in. in diameter, model B60 IITC). As conscious animals are used, it is essential to minimize excitation caused by environmental discomfort. While prewarming, restraint of the animal and inflation of a tail cuff are inevitable, but noise should be minimized.

The major problem inherent to the tail-cuff method is the discomfort due to restraining and heating (19). The high temperature of 36°C used in earlier models should be avoided. Blood odor from surgery will excite the animal. Keeping the restrainer in a semidarkened condition with a cover or using a dark-colored restrainer is useful. It is our experience that a cleaned restrainer, free from foreign scent, reduces the noise. We assign one cleaned restrainer for one animal during a given training and measurement course. It is also important that all of the measurements are done by one person and during a fixed time period of the day, when BP is stable (e.g., 11 AM–1 PM or 12–3 PM).

An instrument developed by J. E. Rogers and J. P. Rogers of Visitech Systems (30) uses a steel cover with a cross section in a –⊓– shape, which can be placed over an animal on a platform and affixed to the platform by a magnetic force (32). The tail would also be placed between two pieces of plastic ridges, having a steel cover on the top. This tail restrainer also contains a pulse-detection sensor. Regardless of the type of instrument, training of animals is essential for this method and is the most time- and labor-intensive process.

Various training schedules are used by different investigator groups. Johns et al. (30) train mice in 3 days, with a schedule that consists of a total of six 5-min periods in restrainer on day 1 and a total of six 10-min sessions. The duration is extended to 10 min ondays 2 and 3, during which the tail cuff is inflated in quick succession. Krege et al. (32) noted that a 7-day training period is necessary. Each session is scheduled at a fixed time between 1 and 5 PM. We allow 1 wk, training the mouse every day with six inflations each between 11 AM and 1 PM. It is important to train the mouse so that it will expect the session at a given time of the day and will walk into the restrainer spontaneously without showing resistance.

Johns et al. (30) recommend the use of a conical metallic insert to cover the mouse's head in the cylindrical restrainer (IITC, model 84), which has been described to improve the behavior of mice and reduce training to 3 days. The cylindrical restrainer (IITC), or one with a steel cover (Visitech), attains comparable levels of concordance with the direct method (30, 32). The results given in Fig.4 are the BP of F2 of a cross between 129 Ola and C57BL/6. The two filled circles at the lower edge of the figure are those of mice that had a clogged intra-arterial catheter.

Fig. 4.

Comparison of blood pressure determined by a sphygomomanometer and by a direct method of indwelling intracarotid artery catherter. F2 male mice produced from F1 of C57BL/6 X 129 Ola were trained for 2 wk for tail-cuff instrument. After final measurement of systolic pressure by this method, animals were catheterized and mean arterial pressure was recorded for 48 h after surgery. In both modes, animals were conscious. ●, Partially occluded intracarotid artery catheter.

Direct Method in Conscious, Chronically Cannulated Mice

Several sites of arteries are candidates for the catheterization These include the left common carotid artery, femoral artery, or abdominal aorta. For studies with a large number of animals, and to minimize catheter occlusion, the carotid artery site seems to be the best. A catheter in the femoral artery is prone to occlusion unless the tip reaches well into the abdominal aorta to the level of the iliac artery. Because the anatomy of the mouse femoral artery is such that it is bifurcated with a sharp bend, it is not easy for a cannula to reach the abdominal aorta. Catheterization into the abdominal aorta requires a midline laparotomy and displacement of the entire intestine. The mouse is particularly slow in recovering from such drastic surgery. Usually, it takes 4–7 days before the animal begins to regain a normal growth rate. The failure rate of animals to recover from the surgical stress is high.

Compared with these two sites, the left common carotid artery is the most convenient site even though the left carotid artery is ligated. The right carotid artery and vertebral arteries provide sufficient blood to the brain. Recovery from the peripheral surgery on the carotid or femoral will take 1–2 days.

For the direct methods, a cannula is prepared by pulling a PE-10 tubing in hot water (∼65°C). A long, tapered cannula will be cut blunt (without a bevel) to make the tip ∼0.2 mm in outer diameter (OD). The other end is connected to PE-50 tubing and heated in a flame to seal the connection. It is made straight by placing it in a cooler water bath, then stored immersed in heparin-containing saline (500 IU/ml) (27). Cannulas made of microrenathane (0.04-in. OD and 0.025-in. inner diameter), which is pulled to a thinner tubing, is more resistant to coagulation and gives much improved patency, lasting longer than 4 days (37).

Under pentobarbital sodium anesthesia (50 mg/kg body wt ip), or alternatively, 300 μl of 2:1 mixture of ketamine (100 mg/ml) and xylazine (20 mg/ml) intramuscularly (im), the left common carotid artery is exposed and a cannula is inserted with a minimum of tissue damage and bleeding. At 8 mm the cannula will reach the aortic arch. At this position the cannula will remain patent for at least 3–4 days. If the cannula is 9 mm long and extends into the ascending aorta, it can remain patent for a longer period because of a higher flow rate. In this case it is important that the orifice is not pressed against the wall of the aortic arch. The cannula can be fixed with a suture (the common carotid artery will be occluded). The distal end of the cannula will be threaded under the skin and will exit at the back of the nape, then be heat sealed. It is secured to the skin or to a plastic button, connected through heparin-filled tubing to a disposable pressure transducer (TXD-310, Micro Med, Louisville, KY), then to a BP analyzer (Micro Med), embedded under or on the outer surface of the skin, with the heat-sealed stub (1.0 ± 1.5 cm) sticking out of the skin. This analyzer prints a digital reading of systolic, diastolic, and MAP, as well as heart rate averaged over a designated period with repeated outputs. Analog wave forms can be recorded by using a preamplifier and strip-chart recorder.

Animals will recover from anesthesia in 1 h. However, it takes a minimum of 8–10 h before the animal becomes free from surgical stress, as judged by a peaking of BP reading, locomotion, and eating. It is important to allow a minimum of 24 h for recovery from the initial surgical stress. At 48 h, the BP readings are practically identical to those taken at 24 h in most cases. The patency of the cannula can be judged by pulse pressure. A robust pulse pressure in the range of 25–30 mmHg (depending on heart rate) indicates that the tip of the cannula is free from partial occlusion by clotting or close contact with vessel walls.

In BP determination, by any one of the direct methods, each mouse is placed in a mouse cage with bedding in which the animal has been kept so that it has its own scent. Again, keeping the animals in a quiet environment and covering the cage with a lab towel that has been previously exposed to the animal helps to relieve the mouse from tension due to a new environment.

As to anesthetics used in the surgery, pentobarbital sodium or the ketamine-xylazine mixture mentioned above is reliable and wears off rapidly. However, pentobarbital sodium lowers BP by >15 mmHg; hence, it cannot be used during the determination of BP unless a relative BP response to a drug is determined. On the other hand, we observed only slight depressor effects (<5 mmHg) with ketamine-xylazine, or inactin. Hemodynamic determination with animals under anesthesia is usually much easier than with conscious, freely moving animals; however, most of the anesthetics depress BP and heart rate. Pentobarbital sodium used at 50 mg/kg depresses the BP by as much as 20 ± 40 mmHg; inactin has much less effect on BP. The individual difference in sensitivity to a given amount of anesthetics also limits the use of anesthetized animals. For fully conscious animals, again, we emphasize the need for minimizing environmental stress. Intrinsic obstacles to this requirement are remaining stress of surgery, anesthesia, and possible discomfort or even slight pain or discomfort due to twisting of the tube exiting the back of the neck. Another important source of stress is the scent of other mice. In addition to confirming the identical pressure reading at 24 and 48 h after the surgery, it is imperative to use a very-low-resistance swivel to connect the cannula to the pressure transducer (Stoelting model 53625, Woodale, IL). When the BP measured directly is higher than that of a well-trained mouse, we would consider that the animal is under stress.

It is also important to pay attention to the diurnal cycle of the BP. To obtain a meaningful and reproducible result, the measurement should be done at a stable time period of the day, for example, 11 AM–1 PM or 1–4 PM.

The accuracy of the direct method is considered to be more reliable than the indirect tail-cuff method for various reasons. Foremost is that the tail-cuff method causes more variation due to stress by the restrainer and heating. As discussed above, the direct method, although it can be done with good reproducibility, may not be completely without stress, nor without variation due to ambulation, eating, and fear. BP should be measured after a long rest. Again, one person must do the determination.

Telemetric Method

Because of the satisfactory experience of many investigators working on rats with a telemetry system (for example, see Refs. 8 and 32), a similar system appeared desirable for mice. The telemetric system of Data Science System is now available with a sufficiently small pressure sensor-transmitter unit suitable for a mouse that, when placed in the abdominal cavity, does not cause discomfort to an animal even smaller than 20 g. The cannula is processed by a proprietary method to prevent occlusion for several months. Highly reliable data may be obtained by this system, but it is not completely free from problems. According to the manufacturer's instructions, the pressure-sensing cannula from the transducer-transmitter unit is supposed to be inserted into the abdominal aorta, and the cannula is affixed to the outer wall of the aorta by a cement. Because the stress of such a major type of surgery is severe (which consists of midline laparotomy and displacement of the entire intestine, and installation of a pressure-sensing cannula into the abdominal aorta), it is our experience that recovery from the surgical stress takes longer than 4–5 days.

On the other hand, the telemetric system is exceedingly well suited to long-term studies of continuously administered drug effects and to screening effects of gene mutation, to mention a few uses. However, the marked fragility and sensitivity of the mouse to the midline laparotomy will prevent rapid screening of mice with altered BP, pulse pressure, or short-term BP responses to various challenges or stresses.

Overall, a method best suited to the purpose may be chosen (Table 2). The indirect method is least cumbersome to start, but a long period of training (7–14 days) is required, which presents a major logistic problem. Less than a 5-mmHg difference by using this method may not be reproducible.

The direct methods, which may involve relatively less surgical stress on the carotid artery or abdominal aorta via the femoral artery, have been highly reliable, and the training and experience of the researcher have advanced. The method is now gaining recognition for good reliability. The complete ligation of the left carotid artery does not seem to be a serious factor affecting BP reading. Greater than 85% of heparin-soaked cannulas remain open. A 5-mmHg difference may be accepted as a real individual difference. There is no question that a digital BP-measuring unit is of great help. However, for BP measurement, it is essential to watch the animals and to pick the right period when an animal has been resting. The need for a single dedicated researcher for the experiment is also essential, as in the indirect method.

Telemetry is probably the ideal choice. If one can afford to allow 4–5 days for an animal to stabilize after postsurgery recovery, one can enjoy the luxury of recording diurnal cycles, computer-assisted removal of abnormal peaks by a jerky motion, and so on. However, for a rapid characterization of effects of gene manipulation on BP, pulse pressure, and ventricular contractility with a large number of animals, the direct method is the least expensive and most readily accessible, although one major drawback of the latter method is that an animal usually has to be killed after one cannula is occluded.

Micropuncture Measurements

The use of micropuncture techniques has contributed greatly to the progress in the understanding of nephron function that has occurred in the 1960s and 1970s. These studies were usually performed in rats, although dogs and hamsters have occasionally also been used. Of note, functional and morphological differences between different strains of mice are much greater than between the commonly used inbred rat strains (14). Two papers using this technique in mice have already appeared despite all present shortcomings, and it is expected that there will be a further extension in the use of this approach. In principle, utilization of the micropuncture technique itself is in no way impeded by the fact that mice are only one-tenth the size of a normal laboratory rat. This is already documented by the fact that micropuncture has been successfully used in the exposed papillae of hamsters as well as in young rats (3, 31). Furthermore, de Rouffignac et al. (14) demonstrated the feasibility of micropuncture in mice in the single study published before the recent resurgence of interest in this species. Thus the major challenge is related to the practical aspect of how to generate a preparation that is reproducible and hemodynamically stable, and therefore permits the creation of a reliable single nephron database.

Thiobutabarbital (inactin) has been the preferred anesthetic in rats because it provides stable circulatory conditions for the duration of a several-hour experiment. This anesthetic has also been used in mice (at a dose of 100 mg/kg ip), even though in our experience it does not readily induce anesthesia when used alone. Induction of anesthesia is more manageable in combination with ketamine (100 mg/kg im). Unlike in rats, the anesthetic effect of inactin in mice usually does not last for the duration of a micropuncture experiment, and supplementary im doses of ketamine are often necessary. The preparation of the kidney for micropuncture requires careful dissection, particularly around the lower pole of the kidney where the fat-embedded ureter is relatively tightly adherent to the kidney capsule. The size of the lucite cup used for kidney immobilization in the flank approach used in our laboratory has to be adapted to the smaller size of the abdominal space. Catheterization of the ureter for unilateral urine collection, the most practical way for clearance determinations and for guaranteeing unobstructed urine flow in the kidney under study, has remained a problem, owing to the tendency of catheter obstruction by urinary crystals. The present solution to this problem may be the use of a bladder catheter, where the caliber of the tube can be much wider. We consider continuous recording of arterial BP an absolute requirement in the present stage of data collection. Arterial BPs measured in either the carotid or femoral artery have been found to be between 80 and 100 mmHg in nontransgenic mice. Recurrent measurements in plasma can only be done with sensitive methods because the total blood volume of the mouse is only ∼1.5 ml. By using [125I]iothalamate as a filtration rate marker, three 5-μl blood collections in 45- to 60-min intervals using Drummond microcaps will remove only ∼1% of the total blood volume. Other determinations, particularly using RIA and ELISA methods, can be done on a terminal blood sample, but recurrent measurements would require miniaturization.

Analysis of superficial nephron fluid absorption by micropuncture has been performed in a limited number of studies. Superficial nephron glomerular filtration rate (SNGFR) ranges from 6 to 15 nl/min, and these variations appear to be due for the most part to differences in kidney or body weight (14, 53, 57). Normalized for a kidney weight of 200 mg (body wt 30 g), SNGFR determined in the proximal tubule is ∼11 nl/min (Fig. 5). Figure6 suggests that the relationship between kidney weight and SNGFR can be linearly extrapolated, within the errors of these measurements, to values obtained in rats, suggesting that proximal SNGFR in both adult rats and mice is on the order of 40–50 nl ⋅ min 1 ⋅ g kidney wt 1. Young rats in a weight range comparable to that of adult mice, on the other hand, have a much lower SNGFR than predicted from the relationship shown in Fig. 5. It is also of note that this relationship cannot be extrapolated to larger organisms such as rabbits, dogs, or humans where SNGFR per gram kidney weight is much lower than found in rats and mice. Only a small number of fluid collections have been performed in distal tubules of mice. In a series of 10 paired collections from distal and proximal segments of the same nephrons, we found a mean SNGFR of 12.8 ± 0.8 nl/min in proximal tubules and 10.8 ± 0.67 nl/min in distal tubules with the mean difference of 2.03 ± 0.46 nl/min being significant at P< 0.01. The relatively small increase in SNGFR caused by eliminating the tonic constrictor effect of the tubuloglomerular feedback may be due to some degree of volume expansion resulting from a relatively high rate of volume replacement (350 μl/h or ∼1.2 ml ⋅ h 1 ⋅ 100 g body wt 1).

Fig. 5.

Relationship between body weight and superficial nephron glomerular filtration rate (SNGFR) in mice (●). Data are from Refs. 14, 53, and57 and unpublished observations from our laboratories. For comparison, a value is given for a young rat of comparable size (▴; data from Ref. 3).

Fig. 6.

Relationship between kidney weight and SNGFR in mice (triangles) and rats of varying body size (circles). Data for mice are from Refs. 14and 53 and unpublished results from our laboratories. (Data for rats are from Ref. 7).

Like in the rat, fluid absorption along the accessible part of the proximal convoluted tubule elevates nonabsorbable marker concentrations to about twice that of plasma concentrations, indicating a fractional fluid absorption around or slightly less than 50% (53, 57). Considering that SNGFR is ∼10 nl/min, fluid absorption along the proximal convoluted tubule is on the order of 5–6 nl/min. Because the length of the proximal convolution in the mouse is 3–4 mm, fluid absorption in the mouse is ∼1.5–2 nl ⋅ min 1 ⋅ mm 1, a value slightly lower than typically found in rats (∼3 nl ⋅ min 1 ⋅ mm 1).

Tonic activation of the tubuloglomerular-feedback (TBF) system is suggested by the observation that SNGFR determined by proximal collections is significantly higher than that found in distal tubules. We recently observed that in mice with a null mutation in either the NHE3 gene or the aquaporin-1 gene, two transgenic mouse strains characterized by inhibition of proximal fluid absorption, proximal-distal SNGFR differences were found to be markedly augmented. This indicates that chronic suppression of proximal fluid absorption causes a TGF adaptation that consists of sensitization and resetting of the TGF function curve to the left. It is likely that this adjustment is a consequence of the extracellular fluid volume depletion that results from impairment of proximal tubule function.

Existence of TGF regulation has been corroborated in microperfusion experiments in which glomerular capillary pressure, derived indirectly from stop-flow pressure (Psf), was evaluated during changes in loop of Henle flow rate (Fig.7). Like in the rat, Psf was found to decrease in response to an increase in loop of Henle perfusion rate in a nonlinear fashion. Maximum elevation of flow rates caused a 24% decrease in Psf (from 37.2 ± 1.6 to 28.2 ± 1.9 mmHg) similar in magnitude to what has been reported previously in rats (Fig. 7 A). However, commensurate with the smaller size of the mouse, the sensitive flow range was shifted to lower flows (Fig. 7 B). The flow rate causing half-maximum responses (V 1/2) was ∼8.7 ± 0.4 nl/min, a value that appears to be within the range of physiological flows. In a comparable study in rats, V 1/2 averaged 18.9 ± 0.97 nl/min. Thus, in the mouse, as in the rat, the operating point of the TGF system is positioned in the steep part of the TGF function, providing efficient regulation during both increments and decrements of loop flow rate. The micropuncture approach to assess TGF responses of Psf was subsequently applied to transgenic mouse strains. In mice with null mutations in the angiotensin II type 1A receptor and in the ACE gene, TGF responses were found to be essentially absent (51, 54). In contrast, TGF responses in mice with a knockout mutation in the nitric oxide synthase I gene were indistinguishable from normal (51).

Fig. 7.

A: relationship between loop of Henle perfusion rate and stop-flow pressure (Psf) in mice and in a comparable study in rats. B: relationship between loop of Henle perfusion rate and change of Psf in mice and rats. Data are from Refs. 52and 54.


Recent advancement in producing recombinant mice has drastically increased the interest in establishing methodologies to assess the function of various organs. For the assessment of renal physiology, several techniques that had been initially developed for rats have been sucessfully adapted for use in mice. In mice, a greater caution is often required due to interstrain variability, high sensitivity to anesthesia, and small absolute quantity in blood volume. With these cautions taken into account, data from recombinant mice are expected to shed new light on the renal physiology of mammals.


We thank Drs. Gary Shull and Alhenc-Gelas for the work conducted in their laboratories, Drs. Kotaro Numaguchi and Masakazu Shiota for helpful intellectual input, Edward Price for expert technical assistance, and Mary Lee Jones for editorial assistance during the preparation of this manuscript.


  • Address for reprint requests and other correspondence: I. Ichikawa, Vanderbilt Univ. School of Medicine, Div. of Pediatric Nephrology, C-4204 Medical Center North, 21st and Garland Aves., Nashville, TN 37232-2584 (E-mail:iekuni.ichikawa{at}

  • The work conducted in our laboratories and cited in this article was supported by National Institutes of Health Grants DK-50594, HL-41496, DK-39626, and DK-48816 (P. Meneton); DK-44757 and DK-37868 (I. Ichikawa); HL-35323 and HL-58205 (T. Inagami); and DK-37448 and DK-40042 (J. Schnermann), and by Institut National de la Santé et de la Recherche Médicale.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 33a.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
  42. 41.
  43. 42.
  44. 43.
  45. 44.
  46. 45.
  47. 46.
  48. 47.
  49. 48.
  50. 49.
  51. 50.
  52. 51.
  53. 52.
  54. 53.
  55. 54.
  56. 55.
  57. 56.
  58. 57.
  59. 58.
  60. 59.
  61. 60.
  62. 61.
  63. 62.
  64. 63.
  65. 64.
  66. 65.
  67. 66.
View Abstract