Cell Physiology

Analysis of organ physiology in transgenic mice

Shobha Rao, A. S. Verkman


The increasing availability of transgenic mouse models of gene deletion and human disease has mandated the development of creative approaches to characterize mouse phenotype. The mouse presents unique challenges to phenotype analysis because of its small size, habits, and inability to verbalize clinical symptoms. This review describes strategies to study mouse organ physiology, focusing on the cardiovascular, pulmonary, renal, gastrointestinal, and neurobehavioral systems. General concerns about evaluating mouse phenotype studies are discussed. Monitoring and anesthesia methods are reviewed, with emphasis on the feasibility and limitations of noninvasive and invasive procedures to monitor physiological parameters, do cannulations, and perform surgical procedures. Examples of phenotype studies are cited to demonstrate the practical applications and limitations of the measurement methods. The repertoire of phenotype analysis methods reviewed here should be useful to investigators involved in or contemplating the use of mouse models.

  • rodent
  • surgery
  • anesthesia
  • cardiovascular
  • lung
  • kidney
  • brain

the dramatic recent expansion in genomic database information has motivated the development of new approaches to characterize gene expression and function. One emerging subject is the mass cataloging of gene expression profiles, developmental programs, and regulatory patterns. Evolving technologies including gene microarrays are of great utility in mass data compilation. A potentially more consequential subject is the physiological role of genes, the subject of functional genomics. Mice are, at present, the preferred mammalian species for genetic manipulations because of the availability of pluripotent embryonic stem cells and inbred strains and the relatively low breeding and maintenance costs. The two basic techniques used in the creation of transgenic mice are integration of foreign DNA into a fertilized oocyte by random chromosomal insertion and homologous recombination in embryonic stem cells that are then introduced into zygotes (123, 160). Transgenic mice serve as sophisticated tools to probe protein function, as models of human disease, and as hosts for the testing of gene replacement and other therapies. Embryonic stem cell libraries for mouse gene deletion are being developed, which will make it possible to generate knockout mice rapidly and without the need to analyze gene structure, construct targeting vectors, and screen embryonic stem cell clones. Murine functional genomics is thus of central importance in modern biomedical research. A major challenge is the small size of mice, which requires a repertoire of imaginative experimental procedures to assess mouse organ physiology.

Many animals, including sheep, goats, dogs, cats, rabbits, guinea pigs, and rats have been used as models to study functional physiology. In mice, there is an extensive literature on in vitro studies to elucidate cellular mechanisms such as cell culture, biochemical assays, and electrophysiological recordings. However, there is a relative paucity of in vivo functional studies in mice, in part due to technical difficulties associated with their small size and the lack of an extensive database on anesthetic procedures and normal physiological parameters. For example, a traditional method in organ physiology is the cannulation of ducts/organs to infuse and collect fluids. Although procedures such as pancreatic duct cannulation may be routine in larger animals, their direct application in mice can present formidable challenges. There are unique concerns in performing anesthesia and surgery in mice, including the challenges in routine monitoring of cardiorespiratory parameters and blood gas analysis during anesthesia and the risks during surgery. Blood loss of ∼300 μl in an adult mouse can produce hypovolemic shock.

This review is focused on the feasibility and limitations of procedures to study organ physiology in mice. Although the emphasis is on measurements in living mice, important in vitro organ preparations are cited, such as preparations involving organ perfusions. Methods are described to study physiology in the cardiovascular, pulmonary, renal, gastrointestinal, and neuromuscular systems. To maintain a sharp focus on organ physiology, this review does not address other issues regarding mouse phenotype analysis such as embryonic development and neoplasia, nor does it address specialized areas such as reproductive, visual, and auditory function. Where appropriate, references have been provided to demonstrate practical applications of the described methods to mutant or transgenic mouse models. The references have been selected to provide a useful starting point in the design of mouse phenotype investigations.


There are many considerations in concluding that a gene is physiologically significant in mice and relevant to human biology. Such considerations apply even if the physiological measurements are carried out with meticulous care and statistical rigor. Defects in organ function can result directly from protein deletion or indirectly from secondary factors such as altered organ development/structure, hemodynamics, or serum chemistries. For example, changes in a gross phenotype such as exercise tolerance can result from a myriad of unrelated abnormalities in muscle protein function, mouse behavior, hemodynamics, hematocrit, and serum electrolyte composition. Phenotype results can be influenced by genetic background. C57BL/6J mice have a higher respiratory rate (270 breaths/min) than C3H/HeJ mice (177 breaths/min) (141). 129/SvJ mice absorb cholesterol twice as fast as C57BL/6J mice (61). Outbred mice strains such as Swiss Webster, CD1, and CARL/ChGo have substantially more genetic heterogeneity than inbred strains such as C57BL/6J or 129/SvJ.

Gene deletion can produce complex compensatory changes involving organ development and expression of other proteins. For example, the failure of dystrophin-deficient mdx mice to exhibit myofibrosis and cardiomyopathy, characteristic of Duchenne muscular dystrophy, is related to upregulation of utrophin, a homologous protein that compensates for the loss of dystrophin (38). Phenotype results can be influenced by environmental factors (temperature, light cycle), diet, cage crowding, and housing conditions. Mice may be maintained in conventional or specific pathogen-free housing or under gnotobiotic (germ free) conditions in which the mice have been rederived by aseptic cesarean section and are raised under sterile conditions. The term specific pathogen free refers to mice that have been found to be free from specified pathogenic microorganisms. An example where housing conditions affect phenotype is the spontaneous development of colitis in interleukin-2- and interleukin-10-deficient mice raised under conventional and specific-pathogen-free conditions but not under germ-free conditions (20, 72).

Phenotype can be affected by age and sex, and there may be diurnal variations. Normal mice have decreased heart rate during daylight hours (when they sleep more) and increased heart rate at night. This circadian variation is absent in transgenic mice overexpressing the cardiac specific GTP-binding protein Gsα (146). Because infections and immune status can influence phenotype, it is generally desirable to use litter-matched mice maintained in the same environment. Because of differences in mouse and human physiology, the extrapolation of data in mice to humans must be made with caution. For example, the maximum osmolality of mouse urine (>3,000 mosmol/kgH2O) is much greater than that of human urine (∼1,000 mosmol/kgH2O). Protein expression patterns and thus the interpretation of phenotype studies may also be species dependent. For example, water channel aquaporin-4 (AQP4) is expressed in both proximal tubule and collecting duct in mouse but only in collecting duct in rats and humans (18, 149). Thus the challenges in mouse phenotype analysis include not only technical factors in transgenic mouse generation and physiological recordings but also numerous genetic, environmental, and species/strain-specific factors.


The selection of anesthetic regimens for mouse surgery and physiological monitoring is important. Many anesthetic regimens affect cardiorespiratory and other organ systems. The response to an anesthetic agent depends on mouse weight, sex, age, and metabolic rate (39, 55, 110). Obese and older mice generally require reduced anesthetic dose. Withholding food and water before anesthesia, a common practice in humans and large animals to prevent regurgitation and aspiration pneumonia, is not necessary in mice. Mice do not regurgitate, and their small glycogen reserve and rapid metabolism contraindicate extended fasting. However, food and water are sometimes withheld for 4–6 h before surgery to reduce the chances of puncture of the gastrointestinal tract during intraperitoneal injections (161). An anesthetized recumbent mouse is particularly prone to hypothermia due to its high surface area-to-mass ratio and high metabolic rate (110). Continuous monitoring of core temperature is generally done using a rectal probe. Thermostatically controlled heating pads and heating lamps are used to maintain core temperature. When mechanical ventilation is indicated, a tracheostomy is generally performed for intratracheal insertion of a 20-gauge needle or polyethylene tubing of suitable diameter. Spontaneous respirations are abolished using pancuronium (0.2 mg/kg). Endotracheal intubation has also been done in mice by direct laryngoscopy, in which the ventral neck is transilluminated for visualization of the vocal cords (6). One study on mechanical ventilation recommended an inspiratory time of <0.1 s, with a tidal volume of 0.35–0.45 ml and a respiratory rate of 120–170 min−1 (24). Barotrauma is a potentially serious complication of mechanical ventilation. Avoidance of excessive tidal volumes and appropriate monitoring (see below) are indicated during mechanical ventilation.

Table 1 summarizes the commonly used anesthetic agents in mice. Avertin (2,2,2-tribromoethyl and tertiary amyl alcohol) and barbiturates are often used. However, both agents can depress cardiovascular function, and Avertin can cause peritonitis if not prepared freshly (26). Fentanyl-flunasine used alone is an excellent neuroleptanesthetic that produces immobility and analgesia for minor procedures such as collection of saliva and tears. Urethane produces stable anesthesia but is associated with a prolonged recovery time. Urethane in combination with chloralose produces minimal cardiovascular depression (24, 140). Ketamine is effective when used with the α-adrenergic blocker xylazine. Ketamine-xylazine anesthesia can cause hypotension, especially in pregnant mice, which is reversed by atipemazole (12, 39). Inhalation anesthesia permits excellent control of the depth of anesthesia. The anesthesia circuit consists of an anesthetic vaporizer, oxygen supply, carbon dioxide absorber and scavenging system. After induction, anesthesia is administered through a face mask or an endotracheal tube. Methoxyflurane, halothane, and isoflurane have been used in mice. Isoflurane may be preferable to methoxyflurane because of its rapid induction and recovery with minimal cardiovascular depression (120).

View this table:
Table 1.

Anesthesia regimens in mice

Monitoring anesthetic depth in mice involves evaluation of reflexes, and, when indicated, measurement of heart rate, blood pressure, and respiratory rate. Visual counting of chest excursions to estimate respiratory rate is not practical in mice. Simple devices are available to measure respiratory rate, such as piezoelectric crystals implanted on opposing left and right surfaces of the chest wall (146). Useful reflexes to monitor the onset and depth of anesthesia include the righting response, toe pinch response, pupillary constriction, and tongue retraction. For practical purposes, loss of the righting reflex indicates the onset of anesthesia and loss of the toe pinch reflex is used as an indicator to proceed with surgery. Shallow respirations or deep gasping breaths signals deepening of anesthesia. Cyanosis with bluish discoloration of the tongue indicates hypoxia and thus warrants oxygenation. If mechanical ventilation is used, arterial blood gas analysis is advisable to adjust ventilator settings. However, repeated blood loss from sampling is a problem since mice only tolerate three withdrawals of 70 μl of blood even if replaced by an equal volume of saline (35). Recently developed pulse oximeters for use in mice should be useful for monitoring oxygenation.


The principal parameters describing cardiovascular physiology in mice include heart rate, mean arterial pressure, left ventricular pressure generation (dP/dt), and cardiac output. Heart size and myocardial wall thickness are ancillary parameters. The small size of the murine heart and the rapid heart rate/brief cardiac cycle time make physiological evaluation challenging. Invasive and noninvasive approaches to assess the above parameters are described, as well as approaches to assess cardiovascular function in unanesthetized mice.

Electrocardiography in mice is used to record heart rate and electrophysiological parameters. The average heart rate in conscious mice is ∼500 beats/min (range 470–650 beats/min) and ∼425 beats/min (range 300–550 beats/min) in anesthetized mice (4, 26, 62). Berul et al. (6) recorded a 6-lead electrocardiogram (ECG) in C57BL/6J mice using subcutaneously inserted 27-gauge needles as electrodes and also obtained epicardial recordings using pacing wires placed on the exposed atrial and ventricular surfaces. P-R, QRS, and Q-T intervals in C57BL/6J mice were 54, 30, and 109 ms, respectively. A three-lead ECG in newborn mice was recorded using small adhesive electrodes; P-R and QRS values of 40 and 22 ms, respectively, were reported in BALB/c− neonatal pups (98). Conduction abnormalities with bradycardia (∼233 beats/min) and prolonged P-R interval (∼74 ms) were found in pups from mothers injected with maternal autoantibodies (anti-SSA-Ro and anti-SSB/La) (98). Prolonged repolarization and sinus node recovery times were found in an α-myosin heavy-chain (α-MHC403/+) mutant mouse model of familial hypertrophic cardiomyopathy (7). Prolonged depolarization times were reported in thyroid receptor (TRα1/β)-deficient mice (59).

Arterial blood pressure recording is required for assessment of cardiovascular function and monitoring during surgical procedures and noncardiac physiological measurements. Invasive monitoring of blood pressure is generally done by carotid or femoral arterial catheterization using fluid-filled catheters and pressure transducers. The principal technical difficulties are related to the small arterial size and fragile walls, as well as blood loss during arterial exposure and cannulation. Cannulation of the abdominal aorta has been done in young mice (67). Mean arterial blood pressures are generally measured. The average mean arterial blood pressure in anesthetized mice is 80–100 mmHg, with averaged systolic and diastolic pressures of 112–124 and 48–62 mmHg, respectively (56). Chronic blood pressure monitoring has been done in nonanesthetized mice by exteriorizing a femoral catheter at the back of the neck. Stable mean arterial pressure (116 ± 1 mmHg) and heart rate (627 ± 21 beats/min) were followed for 35 days in ambulatory mice (97). Hypertension (150–159 mmHg) and an altered baroreceptor response were reported in chronically monitored transgenic mice overexpressing the human renin and angiotensin genes (102). Noninvasive tail cuff methods have also been used in which tail blood flow is measured by light transmittance or piezoelectric transduction (71, 146). Telemetric systems with a transmitter implanted in the peritoneal cavity have been used to record blood pressure, ECG, and heart rate in freely moving mice (59).

Measurement of left ventricular contractility (dP/dt) by cardiac catheterization has been done using a suitable catheter with a micromanometer inserted into the left ventricle via the carotid artery (37, 113). A concern in left ventricular cannulation is the effect of increased outflow resistance on cardiac function. Lembo et al. (76) reported enhanced contractility in insulin-like growth factor-I (IGF-I)-deficient mice with increased dP/dt from 4,023 to 6,395 mmHg/s; Koch et al. (69) reported similar increased contractility in transgenic mice expressing the β-adrenergic receptor kinase inhibitor. Additional catheterization measurements are needed to establish baseline values in different mouse strains and to assess effects of preload and afterload on contractility (63,77, 79). To date, more detailed measurements of intracardiac pressures (left atrial pressure, pulmonary capillary wedge pressure) have not been reported in mice. However, right ventricular pressure measurements have been made by catheterization of the jugular vein (136) and by direct insertion of a 26-gauge needle into the right ventricle (68). Right ventricular peak pressure in control mice (15 mmHg) increased to 22 mmHg in mice lacking atrial natriuretic peptide (ANP).

Measurement of cardiac output in mice is challenging. Conventional thermodilution methods requiring right-sided cardiac catheterization are not practical. Reference microsphere techniques have been used in mice in which radiolabeled microspheres are introduced into the left ventricle by catheterization, and peripheral blood samples are withdrawn by femoral or caudal tail arterial cannulation (4, 126). Cardiac output is computed from injected radioactivity, reference blood flow, and measured radioactivity. Using the microsphere method, cardiac outputs of 12–16 ml/min were reported in C3H and C57BL/6J mice (4, 126). Cardiac output has also been estimated using ultrasonic flow probes and transesophageal echocardiography. Although values obtained by ultrasonographic and echocardiographic methods were similar [7.4 and 7.9 ml/min (Refs. 53and 127)], they were substantially lower than that determined by the microsphere method. Additional studies are needed to establish normal cardiac outputs in mice and to standardize measurement procedures.

Ultrasonic echocardiography provides a noninvasive approach to assess left ventricular size and function (53, 140). Specialized instrumentation is required because of the small size of the mouse heart and the rapid heart rate. Good correlation was found in left ventricular mass determined by M-mode echocardiography and direct weight measurement (140). M-mode echocardiography has been used to assess left ventricular hypertrophy in a model of aortic constriction (36). However, images of mouse heart by two-dimensional echocardiography with available instrumentation have been marginal, even with transesophageal and transperitoneal placement of multiarray transducers. Also, a concern in some of the reports is the low heart rates (250–350 beats/min), which may be related to anesthesia and/or increased thoracic pressure from apposition of the transducer to the chest wall. Another developing noninvasive technology to assess murine heart function is magnetic resonance imaging (MRI) (58, 62). MRI permits reasonably accurate measurement of ventricular volume, ejection fraction, and myocardial mass. In addition, direct visualization of the myocardium permits detection of infiltrative processes (8). ECG-gated MRI has been used to document impaired cardiac function with low-ejection fraction in mice expressing tumor necrosis factor-α (TNF-α) (10) and in mice overexpressing cardiac-specific β1-adrenergic receptors (31). With improvements in instrumentation, such as custom-designed coils, MRI is likely to become an important tool in assessing mouse cardiac physiology.

Vascular contractility studies have been done in mice in which the aorta or carotid artery are studied in vitro (41,139). The aorta or the carotid artery is commonly used in mice for vasomotor tone studies. Vascular tone in response to agents such as vasoconstrictors is measured using an isometric force transducer. For example, in response to acetylcholine, mice lacking the sarcoplasmic reticulum Ca2+-ATPase isoform 3 manifest 50% less aortic relaxation than wild-type mice (80). Intravital microscopy has been used to study the microcirculation of organs such as skin and muscle (51,101). The surgically exposed microvascular bed is viewed directly after fluorescent labeling of erythrocytes or leukocytes for determination of flow velocity, vessel diameter, and vascular permeability (51, 119).


Analysis of pulmonary function in mice involves measurement of respiratory rate, lung volumes, static lung compliance, dynamic lung compliance, airway resistance, and gas-diffusing capacity. Table2 provides a comparison of respiratory parameters in mice and humans. Figure1 A defines lung volumes as assessed in humans by spirometry and forced respiratory maneuvers. Tidal volume refers to the volume of inhaled/exhaled air during normal respiration and is usually assessed in mice by plethysmography. Although technically challenging, some of the other lung volumes have been determined in mice. Vital capacity, the air volume inhaled and exhaled with maximal effort, has been measured in mice by forced deflation maneuvers. Total lung capacity has been estimated in mice from the volume of air that inflates lungs to pressures of 25–30 cmH2O (141). Functional residual capacity, the air volume at the end of a normal respiration, has been assessed in mice using a neon dilution method (47). Lung compliance provides static information about lung mechanical elasticity and is estimated by plethysmography and related methods (see below). Airway resistance (or the reciprocal quantity, airway conductance) provides dynamic information relating airflow to transthoracic pressures. As in humans, arterial blood gas (PaO2, PaCO2, pH) analysis is useful in mice. Finally, diffusion of carbon monoxide provides a quantitative measure of the gas transfer properties of the alveolar-capillary barrier.

View this table:
Table 2.

Respiratory parameters in mice and humans

Fig. 1.

Definition of lung parameters. A: lung volumes as assessed by spirometry in humans (see Table 2 for lung volumes in humans and mice). B: airway responsiveness measured in mice using plethysmography. Pressure changes during a single breath following administration of a bronchoconstrictor. PIP, peak inspiratory pressure; PEP, peak expiratory pressure; T i, inspiratory time from the start to the end of inspiration;T e, expiration time from the end of inspiration to the end of expiration; T r early relaxation phase of expiration when pressure decays to 36% of its initial value. See text for definition and application of the parameter Penh, which is derived from these quantities. [Adapted from Hamelmann et al. (45).] C: schematic of a pressure-volume curve showing normal hysteresis.


In addition to the basic respiratory parameters (respiratory rate, tidal volume, minute ventilation), plethysmography alone or in conjunction with other methods has been used to measure lung volumes, lung compliance, and airway resistance (21,28, 144). Plethysmography is conventionally categorized as pressure or volume plethysmography. The pressure plethysmograph chamber is a constant volume chamber that is sealed to the atmosphere. Chest compression and relaxation accompanying respiration creates pressure changes that are related to tidal volume. Flow is deduced from the derivative of the volume-time curve. In contrast, the volume plethysmograph is open to the atmosphere and connected to a spirometer for flow measurement (134).

Pressure plethysmography with a two-compartment chamber, one for restraining the neck and the other for the body, has been used for lung function studies in mice (25, 75); however, neck compression and restraint increases sympathetic activity and respiratory rate. Chambers have been designed for measurements in unrestrained mice (28, 45), although motion and sniffing artifacts are present. Isothermal conditions are required during pressure plethysmography to relate volume to pressure.

Plethysmography has been used in conjunction with other methods to assess lung volumes. For example, vital capacity in mice has been estimated using a forced deflation maneuver (64). An anesthetized, mechanically ventilated mouse is placed in a plethysmography chamber and instrumented to measure tracheal airflow and airway pressure. Lungs are inflated to total lung capacity and then deflated rapidly with a negative pressure to residual volume. Forced vital capacity in control mice was 35 ml/kg and reduced by ∼50% in mice deficient in the cystic fibrosis transmembrane conductance regulator (CFTR) Cl channel (64); functional residual capacity in mice was ∼17% of total lung capacity (75).

Plethysmography has been used to measure airway resistance and lung compliance in mice (13, 91,167). Continuous measurements of airflow and transpulmonary pressure are made using a plethysmograph chamber with appropriate transducers. In older methods (1), lung compliance was estimated from the difference in lung volume divided by the difference in intrapleural pressure at the beginning and end of inspiration. Because airflow is zero at these times, the volume-to-pressure ratio depends only on static lung properties. Airway resistance was estimated from the ratio of airflow to intrapleural pressures at points of equal volume in the respiratory cycle, where static compliance factors are the same. In modern measurements, airway resistance and lung compliance are determined by more sophisticated regression algorithms utilizing the full time courses of pressures, volumes, and flows (13, 27, 91). One issue in the design and evaluation of such measurements is how to deal with the chest wall compliance; strategies including thoracotomy (to equalize external and pleural pressure) and direct intrapleural pressure monitoring have been used. Table 2 provides airway resistance and lung compliance values in normal mice.

Airway resistance following a challenge with a bronchoconstrictor is the principle parameter measured in airway responsiveness studies (50). After an intravenous injection of acetylcholine, airway resistance increased in a dose-dependent manner and was fourfold higher in DBA/2 mice than C57BL/6J mice (16). A useful parameter to assess airway resistance is the enhanced pause ratio (Penh), which provides an index of the slowing of late expiration during bronchoconstriction (45). Penh is defined by the relation: Penh = [PEP/PIP] (T eT r)/T r, where PEP and PIP are peak expiratory and inspiratory pressures,T e is the total expiration time from the start to the end of expiration, and T r is the time in which the area under the pressure-time curve during expiration decays to 36% of its initial value (Fig. 1 B). In C57BL/6J mice sensitized with ovalbumin, the Penh value following an aerosolized methacholine challenge was 4.0, compared with a value of 0.9 with aerosolized saline.

Forced oscillatory mechanics provides another approach to assess airway resistance in mice. Oscillating air pressures at higher than normal breathing frequencies are applied to the body surface. The airflow response to variations in pressure is measured and airway resistance is computed from the pressure-flow relationship (50).

Pressure-Volume Curves

Pressure-volume curves have be used to estimate lung compliance and lung volumes in mice (Fig. 1 C). Pressure-volume curves have been generated in excised lungs or in dead intubated mice either with or without the chest wall intact (141,142). Immediately before death, the lungs are degassed with 100% O2 to produce complete alveolar collapse. The lungs are then inflated and deflated, and the pressures/volumes are recorded. In most studies in mice, lungs are incrementally inflated with air-to-peak pressures of 25–30 cmH2O and then deflated to negative pressures of 5–10 cmH2O. Lung compliance has been estimated from the slope of the linear portion of the deflation curve (142, 144). Pressure-volume curves in normal lungs show hysteresis due in part to surface tension effects related to lung surfactant. Pressure-volume curves in neonatal mice deficient in surfactant protein B lack hysteresis (144). Because volumes are determined at arbitrarily set pressures, it is important to report the pressure limits used for generating pressure-volume curves. The lung volumes given in Table 2 have been determined using pressure limits of −5 to 30 cmH2O (141).

Diffusion Capacity

Lung diffusion capacity is assessed from the rate at which carbon monoxide disappears from the alveolar airspaces (25,47). Diffusion capacity is determined from the ratio of carbon monoxide uptake to the difference in the partial pressures in alveoli and capillaries. In the single-breath method, lungs are inflated for 10 s with a mixture containing equal amounts of carbon monoxide and neon. A true alveolar sample is collected (after discarding the first half of the sample) for analysis of carbon monoxide content. In the rebreathing method, the mouse rebreathes a specified volume of 0.28% CO and 0.14% He for 60 s before sample collections. A CO diffusion capacity of 0.9 μl · min−1 · mmHg−1 · g−1was reported in normal mice (25). Small sample volumes and dead space effects are potential sources of error in measurements of carbon monoxide diffusion capacity in mice.

Alveolar Fluid Transport

An important function of the alveolar epithelium is the transport of electrolytes and water to maintain adequate hydration and, clinically, the formation and resolution of pulmonary edema. Alveolar fluid is strongly absorbed in the neonatal lung in preparation for air breathing. Alveolar fluid clearance has been measured in fluid-filled mouse lungs using radiolabeled albumin as a volume marker. Fluid is sampled at one or more fixed time points to determine the amount of absorbed fluid. Measurements have been done in in situ and ex vivo perfused lungs and in ventilated mice (96). Alveolar fluid clearance in mice is rapid (1–2% per minute), blocked by various transport inhibitors such as amiloride, and accelerated (by up to 1.5-fold) by β agonists. Alveolar fluid clearance was not affected by deletion of water channels AQP1, AQP4, or AQP5 (2,83), despite substantial decreases in osmotically induced water permeability of the airspace-capillary barrier. Osmotic water permeability in perfused mouse lung is very high, as measured by pleural surface (14) and gravimetric (135) methods. Hydrostatically driven lung fluid filtration has also been measured in mouse lung by a gravimetric method in which lung weight is measured continuously in response to changes in perfusate pressures (135). It was proposed that gravimetry may be useful to study capillary filtration in various isolated mouse organ preparations.


The kidney is responsible for salt/water homeostasis and the elimination of nitrogenous wastes. Renal function involves glomerular filtration and tubular secretion and/or reabsorption of water, electrolytes, sugars, and amino acids (Fig.2 A). Hormonal regulation by antidiuretic hormone (vasopressin), ANP, and components of the renin-angiotensin system play an important role in renal salt and water handling. The murine renal system is relatively easy to study. Collected urine is assayed for pH, electrolyte concentrations, urea, glucose, total protein, and other substances. Serum is analyzed for electrolyte concentrations, blood urea nitrogen (BUN), and creatinine. Urine and plasma osmolalities provide information about urinary concentrating and diluting abilities. Additional renal indexes include the glomerular filtration rate and renal blood flow. Measurements of renal function under stress (e.g. water deprivation, water loading, saline infusion) provide information about urinary concentrating/diluting ability, salt clearance, acidification, and other functions.

Fig. 2.

Urinary concentrating ability in mice deficient in aquaporin water channel proteins. A: schematic of a nephron showing sites of aquaporin (AQP) expression. Excretion of metabolic wastes and conservation of water and electrolytes occurs by glomerular filtration and tubular reabsorption and secretion. B: urine osmolalities in mice of indicated genotype before and after a 36-h water deprivation. [Adapted from Chou et al. (17) and Ma et al. (86).]

Mouse urine can be collected by catheterization of the urinary bladder or after spontaneous voiding. Collection under oil or while observing the mouse is important to prevent evaporative losses. Specially designed metabolic cages allow for separation of urine and feces for long-term urine collections and water balance studies (100). The average 24-h water intake for a 30-g mouse is 4–10 ml, and average urine output is 2–3 ml (108, 137). These values can increase remarkably: for example, daily water intake increased fivefold in the Os/+ inbred mouse model of diabetes insipidus (137), fourfold in angiotensinogen-deficient mice (109), threefold in water channel AQP1-deficient mice (17), eightfold in AQP3-deficient mice (85), and three- to fivefold in mice lacking Cl channel (CLC–K1) (95). Urine osmolality is generally 1,000–2,000 mosmol/kgH2O in mice given free access to water, and urine pH is in the range of 6.7–7.4. Low urine osmolalities have been reported in angiotensinogen null mice [644 mosmol/kgH2O (109)], AQP1-deficient mice [600–700 mosmol/kgH2O (86)], and AQP3-deficient mice [<300 mosmol/kgH2O (85)]. Freezing point depression osmometry is accurate for determination of osmolality when at least 50-μl sample volumes are available, and vapor pressure osmometry is useful for smaller volumes. Densitometry measurements of specific gravity using bromobenzene-kerosene density gradients has been used for samples down to 1 μl (133). Specialized assay methods have been developed for determination of electrolyte and urea content in nanoliter fluid samples collected by micropuncture or isolated tubule microperfusion (129).

Plasma and urinary electrolytes are generally measured by flame photometry. Average plasma electrolyte concentrations in mice are (in mM) 140–145 Na+, 115–120 Cl, 5.0–5.8 K+, and 20–23 HCO3 . Normal serum osmolality in mice is 330–345 mosmol/kgH2O, substantially greater than that in humans (130). As in humans, urinary electrolyte composition varies widely depending on volume status, dietary intake, and age, even within the same strain (44).

BUN and serum creatinine provide semiquantitative information about glomerular filtration rate (GFR) and volume status. BUN in mice is in the range of 8–30 mg/dl. As in analysis of renal function in humans, BUN is sensitive to mouse hydration and metabolic state. Serum creatinine is in the range of 0.3–0.6 mg/dl. Creatinine concentrations are sensitive to muscle mass, intestinal degradation, renal excretion, and the presence of urinary chromogens and so provide only a qualitative index of GFR. In transforming growth factor-β1-overexpressing mice, nephrotic syndrome with ascites and progressive azotemia occurs with elevated BUN to 90 mg/dl (70). Male mice lacking adenine phosphoribosyl transferase have elevated BUN resulting from renal interstitial damage (138). GFR is assayed quantitatively by creatinine clearance or better (since some tubular creatinine secretion occurs) by inulin clearance (43, 73). GFR is affected by age, body weight, and mouse strain, with hybrids showing a lesser decline of renal function with age (44). Creatinine clearance studies using a single injection of 51Cr-EDTA provide information about GFR (from clearance rate) and extracellular fluid volume (from initial concentration). Stockelman et al. (138) reported a creatinine clearance of 15–18 μl · g−1 · min−1 in control mice that was reduced to ∼8 μl · g−1 · min−1 in mice lacking adenine phosphoribosyl transferase. Inulin is a small, uncharged fructose polymer that does not bind to plasma proteins and is efficiently filtered by the glomerulus. Measurement of inulin clearance involves placement of intravenous and bladder catheters, establishing a constant serum inulin concentration by continuous infusion, and measuring urine volumes and inulin concentrations. Angiotensinogen-deficient mice had an approximately threefold reduction in inulin clearance compared with wild-type mice (109). Other nonsecreted, nonabsorbed macromolecules that have been used to measure GFR include iothalamate and fluorescently labeled dextrans.

Single nephron GFR is measured by renal micropuncture (81). Under anesthesia, the kidney is exposed via a flank incision, placed in a Lucite cup, and covered in mineral oil. Timed urine collections are made by direct micropuncture of the nephron with glass micropipettes, generally at superficial sites in the end proximal tubule or distal tubule. Single nephron GFR is determined from the collected volume and concentration of an impermeant marker (such as iothalamate) relative to that in serum. Single nephron GFRs of normal mice are in the range of ∼8–12 nl/min (81,129). Mice deficient in AQP1 water channels have impaired fluid absorption in proximal tubule as detected by micropuncture sampling of end proximal tubular fluid (128).

Urinary concentrating ability is assessed from the effects of water deprivation and/or V2 receptor agonists on urine and serum osmolalities and urine volume. Normal mice concentrate their urine to >3,000 mosmol/kgH2O after a 24–48 h water deprivation. Marked urinary concentrating defects have been documented in mice lacking CLC–K1 (95). Our lab has generated a series of knockout mice lacking the renal aquaporin water channels (sites of expression shown in Fig. 2 A). Figure 2 B shows remarkable urine hyposmolality in mice lacking AQP1 and AQP3, with an inability of AQP1 null mice to increase their urine osmolality in response to water deprivation. These studies have provided useful information about the urinary concentrating mechanism (151). The V2 agonist 1-desamino-8-d-arginine vasopressin (DDAVP) increases collecting duct water permeability so that the osmolality of excreted urine is approximately that of the renal medullary interstitium. In response to a 1 μg/kg intraperitoneal bolus injection of DDAVP, angiotensinogen-deficient mice failed to increase urine osmolality, whereas wild-type mice had an ∼11-fold increase in urine osmolality. Angiotensinogen-deficient mice also failed to lower urine osmolality in response to acute water loading, indicating defective urinary diluting ability (109). Urine osmolality in AQP1 null mice is also unresponsive to DDAVP challenge (86) because of defective proximal tubular absorption (128) and countercurrent exchange (17,114).

Water clearance and urinary diluting ability have been measured in anesthetized rats and mice subject to acute volume expansion (60, 150). After placement of intravenous and bladder catheters, urine is collected following infusion of hypotonic solutions. In a volume expansion study, urine output increased 12-fold and Na+ excretion increased 6-fold in wild-type mice but not in guanyl cyclase A-deficient mice, indicating that guanyl cyclase A is required for signal transduction of the natriuretic and diuretic factors produced by the heart (66).

Analysis of renal function in mice maintained on special diets has provided information about salt excretory ability and urinary acidification. Dietary studies with high salt (8% NaCl) or low salt (0.008% NaCl) given for several weeks are used to assess kidney function with or without volume expansion (150). When subjected to acute volume expansion, ANP-deficient mice on a low-salt diet had approximately fourfold diminished salt excretion, establishing a role for endogenous ANP in the natriuretic response (60).

Renal blood flow has been measured in anesthetized mice using an ultrasound flow probe positioned around the renal artery (40, 145). Mean renal blood flow in mice was ∼1.8 ml/min at a mean arterial pressure of 94 mmHg (145). Renal blood flow following volume expansion was increased by 84% due largely to a reduction in renal vascular resistance (145). Renal medullary and cortical blood flow has also been estimated using laser Doppler flowmetry (40). Increased renal perfusion pressure (90–140 mmHg) or saline loading resulted in ∼25% increased medullary blood flow without changes in cortical flow, suggesting that medullary flow is not as tightly regulated as cortical flow (40).


The gastrointestinal tract carries out nutrient intake and digestive, absorptive, secretory, and barrier functions. Ingested food mixes with saliva and passes into the acid environment of the stomach for proteolytic hydrolysis. Pancreatic exocrine secretions containing amylase, proteases, and lipases facilitate digestion of carbohydrates, proteins, and fat. Bile and bile salts contribute further to the digestion of lipids. Digested nutrients are absorbed mainly in the small intestine, and water absorption and stool desiccation occur in the colon. Basic analysis of gastrointestinal function includes measurements of food/water intake and urine/stool excretion using metabolic cages and analysis of stool for fat and water content. As in human physiology, serum concentrations of lipase, amylase, transaminases, triglycerides, albumin, etc. provide information about liver and pancreatic function. Specific analyses of gastrointestinal functions in mice are described below.

Collection and Analysis of Gastrointestinal Secretions

Cannulation of the parotid duct has been done in anesthetized BALB/c mice (93, 107). Basal salivary volume flow was ∼80 μl/30 min, with Na+ and K+concentrations of 61 and 15 mM, respectively. Collection of total secreted saliva in anesthetized mice has been done by different methods, including a mini suction device (84), placing a mouse face down on a sloping acrylic plate covered with filter paper (106), and uptake into microcapillaries positioned at the orifice of the parotid papilla (105). In response to pilocarpine stimulation, ∼225 mg of saliva were collected in wild-type mice in 5 min, with remarkably less (∼80 mg) in mice lacking water channel AQP5 (84). Figure3 A shows the viscous saliva produced by AQP5 null mice, and Fig. 3 B shows that the saliva was hyperosmolar and hypertonic, indicating normal salt transport into the salivary gland acinus without adequate water flow.

Fig. 3.

Saliva production in mice lacking AQP5 water channels.A: saliva collected in 5 min in wild-type (+/+), heterozygous (+/−), and AQP5 null (−/−) mice showing reduced and viscous saliva in AQP5 null mice. B: osmolality and Na+ concentration of saliva from wild-type and AQP5 null mice. [Adapted from Ma et al. (84).]

Gastric secretions have been collected in fasted anesthetized mice. After laparotomy, the esophagus and pylorus are ligated, PE-50 tubing is inserted through an incision in the anterior portion of the stomach to infuse saline, and PE-240 tubing is inserted in the antrum to collect the perfusate and gastric secretions. Gastric acid output is determined by titration of collected fluid (34). Gastric pH was 1.5–2.1 in fasted mice. Basal gastric acid secretion was 0.6 meq/min in control mice and undetectable in gastrin-deficient mice (34). A two- to threefold increase in gastric acid secretion was reported in normal mice following the administration of gastrin, histamine, or carbachol (34) but was not affected by deletion of AQP4, a water channel expressed in gastric parietal cells (156).

Exocrine pancreatic secretions have been collected and analyzed for amylase, lipase, and bicarbonate content. Briefly, the common pancreatic duct is identified, isolated from the liver, ligated, and cannulated with polyethylene tubing (inner diameter 150 μm). Basal pancreatic flow rate was ∼11 μl/h, increasing approximately twofold with cholecystokinin/secretin stimulation (87). Basal amylase output was ∼58 U/30 min in Swiss mice (148). The proximal duodenum has been cannulated to measure pancreatic bicarbonate secretion in CFTR-deficient mice (52). PE-50 was passed via the stomach into the duodenal bulb and a ligature secured around the pylorus. Basal bicarbonate output in wild-type mice was 4.7 μmol · cm−1 · h−1 and reduced approximately twofold in CFTR-deficient mice. Bicarbonate production following luminal acidification increased by ∼50% in wild-type mice but did not increase in CFTR-deficient mice (52).

Cannulation of the bile duct has been done in mice (99,155). Briefly, the lower end of the common bile duct is ligated and cannulated with PE-10 tubing below the entrance of the cystic duct (159). Biliary flow was ∼2 μl/min in wild-type mice (99). Mice deficient in ferrochelatase showed a 4-fold increase in bile salt excretion and an ∼80-fold increase in plasma bile salt concentration (99). Of the >10 bile acids identified in mouse bile, the major component is muricholic acid (131, 158). A threefold decrease in total fecal bile acid content was reported in mice deficient in cholesterol 7α-hydroxylase (131).

Gastrointestinal Motility

Gastric emptying time and intestinal transit time are used as indexes of gastrointestinal motility. Gastric emptying has been studied in mice using radiolabeled meals (11, 94). The mouse is euthanized at 30 min, the cardia and pyloric ends of the stomach are clamped, the stomach along with the small and large intestines are removed, and radioactivity is measured. Gastric emptying has been expressed as the percent of total counts found in the small intestine and colon. Gastric emptying in normal mice was 40–48%, increasing to ∼67% after exposure to cold stress (11).

Measurement of intestinal transit time in fasted mice has been done by oral administration of a radiolabeled charcoal meal (32,111). The intestine is removed, cut into the multiple segments, and radioactivity of each segment determined. Various numerical procedures, such as the slope of a radioactivity vs. intestinal segment plot or geometric center determination, have been used to quantify intestinal transit (104,118). Earlier methods of measuring intestinal transit time included administration of nonradioactive markers such as charcoal or phenol red and expressing the distance traveled by the leading edge of the meal as a percentage of the total intestinal length (94, 121). Morphine, a known inhibitor of gastrointestinal motility, reduced transit time by 22% in wild-type mice but did not affect intestinal transit time in μ-opioid-receptor-deficient mice (121).

Nutrient Absorption

The intestinal absorption of carbohydrates, proteins, and fats has been studied by oral and duodenal nutrient tolerance tests. Test nutrients are administered by gastric gavage or by infusion directly into the lumen of the duodenum (following laparotomy). Blood is collected for assay of the absorbed nutrient at timed intervals. The capacity of the intestine to digest and absorb carbohydrates has been assessed by administration of a mixture of d-glucose and maltose (9) or d-xylose (57,152). Complete absorption of xylose occurs in 45 min in normal adult mice and at 2 h in neonatal mice (57). The absorption of orally administered radiolabeled leucine has been studied in mice as an index of intestine amino acid absorption (9).

Fat malabsorption in various intestinal diseases generally occurs much earlier than carbohydrate or protein malabsorption (103). Exocrine pancreatic insufficiency and biliary disorders cause fat malabsorption. Steatorrhea is a common presenting sign with loose pale-colored stools. The Sudan 3 dye test distinguishes split/digested fats from neutral or undigested fat (65). Digested fats are detected by vigorously mixing a 3-mm-diameter pellet of feces with a drop of 36% acetic acid on a glass slide. A drop of Sudan dye is added, a coverslip is applied, and the mixture is brought to a boil three times. Microscopically, the split fats appear as orange globules that form spicules on cooling. Neutral or undigested fats are detected by Sudan 3 as above, except that the boiling process is omitted. Fecal fat analysis is done on collected stools using a chloroform extraction procedure (82). Normal mice on standard chow have ∼7% fecal fat (compared with stool dry weight). Increased fecal fat (55% of dry weight) has been reported in plasma triglyceride lipase 2-deficient suckling mouse pups (82).

Fat malabsorption also affects absorption of fat-soluble vitamins A and E. The vitamin A tolerance test in mice has been done by administration of an intragastric bolus of retinyl palmitate in corn oil followed by collection of serum samples (for retinyl palmitate assay) at 1, 2, 4, and 10 h after administration (158,159). Mice expressing liver-specific human apolipoprotein B (Hu BTg+/0 apo B−/−) exhibit defective fat malabsorption, with undetectable serum retinyl palmitate at 10 h, compared with wild-type mice that have peak serum concentrations at 2 h (162).

Colonic water absorption has been studied in rats by insertion of a preweighed agarose cylinder. The net amount of water gained or lost is determined by measuring changes in the weight of the agarose plug (29, 164). A similar approach should be useful in mice. Net fluid transport in the small and large intestine has been studied in mice by the ligated-loop assay. Briefly, the intestines are exposed and the contents are flushed out through a small distal incision. Jejunal, ileal, or proximal colonic loops are identified, ligated, and filled with saline containing [3H]polyethylene glycol as a nonabsorbable volume marker (147). In a continuous flow approach where the colonic lumen was perfused by a constant infusion pump, transepithelial water permeability was reduced 50% by AQP4 deletion (157).

Mucosal Barrier Integrity

The structural and functional integrity of the gastrointestinal mucosal barrier is important for protection from various luminal agents such as acids, enzymes, bacteria, viruses, and toxins (117). Barrier functional studies in mice have been carried out by exposure of the stomach to various concentrations of acid, alcohol, and urea (30). The gastrointestinal mucosa is then examined histopathologically for evidence of hemorrhage or mucosal damage/repair. When given 2.5% dextran sodium sulfate in their drinking water, mice lacking the intestinal trefoil factor developed fatal hemorrhage and enterocolitis (92).


Neuromuscular and behavioral data in mice generally should be evaluated in the context of overall systemic abnormalities. For example, hypotension and dehydration can produce marked secondary effects on neuromuscular performance. Initial observations of mouse activity, gait, and general behavior can provide clues about central or peripheral neuromuscular abnormalities and behavioral disorders. Behavior results from a complex interaction between genetics and environment. To minimize environmental variations, behavioral tests are generally carried out at the same time on litter-matched, same-sex pairs (22, 116). Sensory and motor deficits can affect performance on behavioral tests and need to be ruled out. For example, assessment of visual acuity by observing a mouse near the edge of a table cannot be done if the mouse stumbles due to motor deficit. Conversely, primary neuromuscular or behavior abnormalities can produce multiple systemic defects. For example, newborn mice lacking the G protein Golf have defective olfaction, leading to an inability to suckle and failure to thrive (5).


Locomotion is a complex coordinated event involving direct transmission of signals from the motor cortex through the corticospinal tract to the spinal cord and indirect transmission through multiple accessory pathways of the basal ganglia, the cerebellum, and the brain stem. Impaired mobility is a common presenting sign of generalized illness. Detailed gait analysis using video cameras to record stride frequency and forelimb/hindlimb swing have been reported to quantify locomotion defects (19, 153). Exploratory tests of spontaneous locomotion provide a composite index of neuromuscular performance and behavior (56,165). After placement in an unfamiliar environment, mouse locomotion is assessed by video analysis, counting for example the number of times the mouse crosses grid lines. Purkinje cell degeneration gene-deficient (pcd) mice with hereditary degenerative ataxia show an average of 21 grid crossings in 5 min compared with control mice with 227 crossings (165).

Sensorimotor Evaluation

Analysis of reflexes in mice provides information about afferent sensory and efferent motor pathways. Commonly assessed simple reflexes include the righting reflex, the grasp reflex, and the geotaxic response. The righting reflex tests the ability of the mouse to right itself when turned on its back. The grasp reflex tests the ability of the mouse to grasp tightly onto an 18-gauge needle placed in the palm of each limb. The geotaxic response tests the ability of the mouse to turn around and walk up a 45° incline when placed head down on the incline (154). Additional specific tests of sensorimotor function in mice include the rotorod, the hot plate test, and the tail flick test (78, 125). The rotorod test involves placing the mouse on a rod with accelerating rotational speed and measuring the performance time as the time taken for the mouse to fall off the rod (Fig. 4 A). Rotorod performance depends on intact sensorimotor, cerebellar, and muscular systems. On the rotorod test, 92-wk-old prion protein-deficient mice fall from the rod in 10–20 s, whereas wild-type mice remain on the rod for >180 s (125). Enkephalin-deficient and dopaminergic D2 receptor-deficient mice also show significant impairment of rotorod performance (3,88). The hot plate test measures the response to pain caused by a thermal stimulus. The mouse is placed on a hot plate, and the time is determined for the mouse to lick its paws or to jump off the hot plate (168). Porphobilinogen deaminase-deficient mice with peripheral neuropathy remain on the hot plate for 45 s, approximately five times longer than wild-type control mice (78). Physical endurance as an index of muscle strength has been tested by exercising mice on a treadmill (Fig. 4 B). Older mice (34–36 mo) were unable to run on a 5° incline, and their running time was 50% reduced on a 0° incline compared with young mice (112).

Fig. 4.

Assessment of mouse neuromuscular performance.A: rotorod. Motor coordination is tested by placing mice on the rotorod apparatus in which rotational speed is continuously increased. Performance time is defined as the time taken to fall off the rod. B: treadmill. Mice are forced to run on a treadmill at specified speed and incline. A shock plate is positioned just behind the moving belt. C: schematic of Morris water maze. Mice search for an underwater platform using external visual cues.

Aggression, Anxiety, and Maternal Behavior

Specialized behavioral tests have been used to assess aggression, anxiety, and maternal behavior. Adenosine A2receptor null mice show increased biting frequency compared with wild-type mice, indicating increased aggression (74). In the resident intruder test, an intruder mouse from a different cage is placed in the cage of the test mouse. Aggressive mice attack an intruder mouse in <5 min. 5-Hydroxytryptamine 1B-, adenosine A2 receptor-, and monoamine oxidase A-deficient mice all display enhanced aggression by attacking the intruder mouse in <2 min (74, 124). Reluctance to move around in an open field, quantified by the frequency of entries into the open arms of an elevated maze, is taken as a parameter of anxiety (23). Adenosine A2 receptor mice exhibit fewer entries into the open field compared with wild-type mice, reflecting increased anxiety, whereas endothelial nitric oxide synthase null mice show more entries, suggesting decreased anxiety (54). Neophagia (aversion to eating new food) in an unfamiliar environment is also taken as a sign of anxiety and has been used to test the anxiolytic action of benzodiazepines and other agents (23). Various maternal behavior tests have been devised. After separation from newborn pups, a primiparous mother mouse will retrieve her pups and exhibit nesting, crouching, and nursing behaviors. Female mice deficient in the G protein Golf lack this nurturing response, leading to mortality among pups (5). It should be noted that, despite a number of apparently clear-cut reported examples of abnormal mouse behaviors, behavioral evaluation remains relatively subjective, sometimes difficult to reproduce in different labs, and potentially subject to multiple interpretations.

Learning and Memory

Learning in association with memory is a complicated process involving different regions of the brain including the hippocampus, neocortex, and amygdala. Complex learning, simple associative learning, and avoidance learning are examined in mice most often using the Morris water maze test. In one version of the Morris water maze test, a platform with a white flag is submerged in a pool of water. To avoid drowning, normal mice use the white flag as a visual cue to associate the flag with the platform (115, 132). In the hidden version of the Morris water maze test, the submerged platform remains in a fixed location without a white flag (Fig. 4 C). The mouse must analyze multiple spatial cues in the chamber surrounding the pool to locate and swim to the platform. Calcium calmodulin kinase II-deficient mice show impaired performance on the hidden version of Morris water maze test and take twofold longer (20 s) than the wild-type mice to locate the platform, suggesting impaired spatial learning (132). Passive avoidance tests make use of an aversive stimulus to study memory. In one such test the mouse is placed in the lit portion of a two-compartment chamber, and entry into the dark chamber produces a foot shock. When tested at 1 day after the learning period, the duration of time that the mouse avoids the dark chamber is recorded as a measure of memory (116). In an interesting study using the foot shock test, mice deficient in the β-subunit of the neuronal nicotinic acetylcholine receptor had better retention of the foot shock memory than wild-type controls. Injection of nicotine increased retention of the foot shock memory in wild-type but not the null mice (116).

Electrophysiological Studies

Nerve and motor performance in mice are assessed by nerve conduction velocity, electromyography (EMG), compound muscle action potentials, and electroencephalography (EEG) (78,143, 154). Peripheral nerve conduction studies of the facial, sciatic, tibial, and caudal nerves have been reported in anesthetized mice. Stimulating electrodes are placed percutaneously near the nerve being tested, and recording electrodes are placed in the corresponding innervated muscle. After a 0.05-ms square wave stimulus, the latency, amplitude, and area of the resultant action potentials are recorded. Nerve conduction velocity of the sciatic nerve was reduced fourfold in protein tyrosine phosphatase-deficient mice, resulting from developmental abnormalities of large myelinated fibers (154). EMG recordings are obtained using bipolar stainless steel or platinum-iridium wires inserted into the muscle belly. Motor unit action potentials are elicited by gentle pressure on the opposite limb to induce a withdrawal response (78, 154). A twofold increase in motor unit action potential amplitude and increased recruitment of motor units was found in porphobilinogen deaminase-deficient mice that develop peripheral neuropathy (78). Compound muscle action potentials of the gastrocnemius muscle have been recorded using a unipolar electrode placed on the sciatic nerve at a paraspinal site, a ground electrode placed on the back, and a reference electrode placed on the Achilles tendon. The amplitude of the gastrocnemius compound muscle action potential in progressive motor neuropathy (pmn) mutant mice was 35 mV compared with 80 mV in wild-type mice (42). EEG in mice uses surgically implanted gold-plated screws on the cranium over the cerebral hemispheres. EEG has been used in mice to study sleep patterns and seizure activity and threshold (143).

Cerebral Blood Flow, Perfusion, and Edema Studies

Cerebral blood flow has been measured in mice using a 0.5-mm laser Doppler flow probe immobilized on the intact skull with adhesive. Doppler flowmetry provides a relative rather than absolute measure of perfusion (24, 166). Doppler flowmetry has been used to study autoregulation and CO2 sensitivity of cerebral blood flow (24). Regional blood flow in anesthetized 129/SvJ mice was relatively constant at mean arterial pressures of 40–130 mmHg. In response to hypercapnia induced by inhalation of 5% and 10% CO2, regional blood flow increased by 38% and 77%, respectively (24).

Cerebral edema is the common result of various types of injury to brain cells, such as ischemic stroke and tumors, as well as metabolic insults such as hyponatremia. Brain tissue water content has been measured by a densitometric method using bromobenzene/kerosene density gradients (90) and by determination of wet-to-dry weight ratios (46, 48). Figure5 A is a transmission electron micrograph showing reduced swelling of astrocytic foot processes in brains from AQP4 water channel-deficient mice after acute water intoxication (89). Quantitative analysis of multiple micrographs showed a substantial reduction in astrocytic foot process area at 30 min after water intoxication (Fig. 5 B). Cerebral edema, quantified by the densitometric method, was seen after water intoxication but was reduced in the AQP4-deficient mice (Fig.5 C).

Fig. 5.

Reduced brain edema after water intoxication in AQP4-deficient mice. Mice were injected with water (20% body wt) intraperitoneally. A: transmission electron micrograph of mouse brain showing swelling of astrocytic foot processes (arrows), with reduced swelling in AQP4-deficient mice at 30 min following water intoxication. B: quantitation of astrocytic foot process area showing markedly increased area in wild-type mice after water intoxication. C: cerebral edema assessed by densitometric measurement of the specific gradient of brain fragments. Increased specific gravity indicates decreased brain water content. [Adapted from Manley et al. (89).]

MRI has been used in stroke models to measure infarct size, changes in vascular perfusion, and diffusion abnormalities during ischemia and after reperfusion (46, 163). The noninvasive nature of MRI makes it useful to study mouse models of neurodegenerative diseases and monitor drug responses. Limited signal-to-noise ratio in present day MRI is a concern that will likely be addressed using higher magnetic field strengths and small radiofrequency receiver coils (122).


The analysis of mouse phenotype is much like human medicine, except that subjective information is not available and obtaining physical and laboratory information is often challenging. For genetic manipulations that can produce multiorgan abnormalities, a “review-of-systems” approach, as in human medicine, is required to evaluate an ensemble of clinical findings in terms of primary vs. secondary defects. Additional variables are introduced when phenotype depends on mouse age, sex, strain, environment, and other factors. As described in this review, methodological developments for investigation of organ physiology in mice are rapidly advancing in terms of instrumentation and surgical approaches. However, in many areas there is a paucity of information about baseline physiological parameters, and there lacks a consensus about measurement guidelines. Establishing guidelines for performing and interpreting mouse organ physiology will become increasingly important with the expanding database of mouse phenotype information. Last, we anticipate over the next decade the introduction of automated mouse phenotype analysis for high-throughput screening in drug discovery and toxicity studies.


We thank Drs. Tonghui Ma, Geoffrey Manley, and Michael Matthay for critical review of the manuscript and helpful suggestions.


  • This work was supported by National Institute of Health Grants DK-35124, HL-51854, HL-42368, and DK-43840, and Grant R613 from the National Cystic Fibrosis Foundation.

  • Address for reprint requests and other correspondence: A. S. Verkman, 1246 Health Sciences East Tower, Cardiovascular Research Institute, Univ. of California, San Francisco, San Francisco, CA 94143-0521 (E-mail: verkman{at}itsa.ucsf.edu;http://www.ucsf.edu/verklab).


View Abstract