The development and widespread use of genetically altered mice to study the role of various proteins in biological control systems have led to a renewed interest in methodologies and approaches for evaluating physiological phenotypes. As a result, cross-disciplinary approaches have become essential for fully realizing the potential of these new and powerful animal models. The combination of classical physiological approaches and modern innovative technology has given rise to an impressive arsenal for evaluating the functional results of genetic manipulation in the mouse. This review attempts to summarize some of the techniques currently being used for measuring cardiovascular, renal, and pulmonary variables in the intact mouse, with specific attention to practical considerations useful for their successful implementation.
the advent of molecular techniques to produce targeted gene mutations in the mouse has opened nearly unlimited opportunities for studying the physiological and pathophysiological role of almost any functional or regulatory protein in the intact animal. Mouse models with genetic manipulations specifically designed to investigate specific proteins are being developed at a remarkable rate, and the availability of these “designer mice” has opened avenues of investigation that would have once been considered implausible. Along with these genetic advances, considerable progress has been made in our ability to evaluate the resulting phenotypes at the tissue, organ, and whole animal level. These advances are evidenced by an ever increasing number and sophistication of studies that evaluate physiological behavior in intact mice, which have been made possible by a combination of new computer and electronic technologies, investigator ingenuity, and perseverance and not a small amount of mental and manual dexterity. Commensurate with this progress, there have been a plethora of illuminating review articles summarizing techniques and methodologies that have been developed for evaluating function across a wide range of physiological disciplines, including cardiovascular, renal, pulmonary, behavioral, neuro-, and electrophysiology (14, 17, 29, 45, 46,56, 63, 69, 70, 91). It is not the goal of this review, therefore, to provide yet another synopsis of currently available approaches, but rather to provide a practical guide to the use and implementation of these methodologies, primarily for those investigators who may not have extensive experience with in vivo experimentation in the mouse. The ubiquitous nature of new transgenic approaches has, after all, resulted in an inevitable blurring of the once bold line between molecular biology and classical physiology. By necessity, investigators wishing to take advantage of these powerful tools must become at least minimally versed in both disciplines. Because the practical experiences in our own laboratory are primarily centered on cardiovascular, renal, and pulmonary studies, this review shall focus on these areas of investigation and will necessarily reflect our understanding of the implementation of these techniques. It is relevant to point out, as many others have, that the mouse is not simply a small rat, and special care must be taken to resist the urge to directly compare data from mice to that from rats. Oddly enough, in our experience, the mouse seems to more closely resemble a small rabbit in terms of obvious functional characteristics; compared with the rat, it is somewhat fragile, difficult to maintain blood pressure under anesthesia, and has a labile blood pressure that is difficult to elevate experimentally by either acute or chronic intervention.
There are essentially two categories of investigation that are most commonly used in the phenotypic analysis of newly generated mouse models: chronic, often noninvasive (or minimally invasive) screening for functional variations, and acute instrumentation for in-depth evaluation of specific physiological variables. Examples of chronic approaches include echocardiography, tail-cuff blood pressure evaluation, renal balance studies, and whole animal plethysmographic evaluation of pulmonary function. In addition, there has been important recent progress in the area of chronic instrumentation and the use of telemetric techniques for evaluating function in a relatively undisturbed animal. Acute approaches can be categorized into in vivo procedures, such as cardiac catheterization, regional blood flow measurements, renal clearance and micropuncture studies, and airway pressure-flow measurements, as well as ex vivo procedures, such as Langendorff and isolated working heart preparations (37), isolated smooth muscle preparations (79), and isolated lung preparations (67). Because the requirements for adequate phenotypic evaluation of mutant mouse strains is becoming more rigorous in the eyes of many reviewers, this article will describe the various screening methods often used to initially evaluate organ function in mice, but will focus more intently on the more analytic techniques used to confirm various phenotypic traits.
The choice of anesthesia for use in mice is a crucial decision and depends largely on the type of experiment being performed as well as the personal preference and experience of each investigator. Variables to be considered are background strain of mouse, duration of the procedure, whether the procedure is terminal or recovery in nature, and the physical constraints of the planned instrumentation. Anesthetic regimens are of two types, injectable and inhaled, and the various compounds used in mice have been extensively reviewed (69). The most commonly used anesthetics in mice include the injectable agents avertin, pentobarbital, inactin (thiobutabarbital), chloralose-urethane, ketamine (usually combined with other agents such as acepromazine, xylazine, and/or diazepam), and inhaled agents isoflurane, methoxyflurane, and halothane. We have adopted the use of several anesthetics in our laboratory and the choice depends on several factors. For long, nonrecovery procedures, such as cardiac catheterization or renal micropuncture, we typically use a combination of ketamine and inactin given as separate intraperitoneal injections: 2 μl/g body wt of 25 mg/ml ketamine given first, followed by 2 μl/g body wt of 50 mg/ml inactin. This combination allows for quick induction and fairly prolonged action with minimal supplementation required. The cardiodepressor effects are mild as evidenced by mean arterial pressure in the 80–90 mmHg range and left ventricular dP/dt ∼9,000 mmHg/s under basal conditions (58, 59); these values are only slightly lower than those observed in awake mice (66). The advantage of using both agents in a long procedure is that the ketamine-inactin combination will initially produce a deeper anesthesia that permits relatively invasive procedures. Then as the shorter acting ketamine begins to wear off, the level of anesthesia can be carefully titrated via intravenous supplementation and during constant monitoring of heart rate and blood pressure. It should be recognized that inactin does not have the very long duration of action that is seen in the rat and therefore must be supplemented regularly; nonetheless, it is generally accepted that its use for recovery surgeries should be avoided. For nonrecovery procedures, a tracheotomy is usually performed using a short length of PE-90 tubing to ensure airway patency, but the animal is usually allowed to breath spontaneously without the aid of a ventilator. However, it is fairly common to provide anesthetized mice with a stream of 100% O2 to prevent hypoxia. It must also be noted that anesthetized mice are particularly vulnerable to hypothermia and body temperature must be constantly monitored and carefully regulated. We monitor and maintain temperature via a rectal thermistor probe and a feedback-controlled, warmed surgical table (Vestavia Scientific, Birmingham, AL). Supplemental heat is often provided using a heat lamp to ensure even warming.
We recently began to explore the use of inhaled anesthetics (isoflurane in our case) for prolonged analytic procedures. The advantages of isoflurane are that it preserves sympathetic vasomotor activity, has minimal cardiodepressor effects, and allows very careful, minute-to-minute control of the anesthetic plane. By way of comparison, left ventricular dP/dt measured under isoflurane averages ∼13,000 mmHg/s and mean arterial pressure is usually ∼90–100 mmHg (unpublished observations). The disadvantages of isoflurane are that it requires expensive equipment, can be physically cumbersome, especially when the procedure requires extensive instrumentation and multiple changes in animal posture, and the depth of anesthesia can be volatile if its administration is not finely and carefully controlled, often through the use of mechanical ventilation. We attempted to combine isoflurane with other agents, such as ketamine or buprenorphine, to help stabilize the anesthetic plane, but many of the benefits of inhaled anesthesia are diminished by this approach. In our experience, isoflurane anesthesia can be quickly induced by manually restraining the animal and placing its head in a mask fashioned from a 50-ml Falcon tube through which a steady stream of 4% isoflurane is introduced. After induction of deep anesthesia, usually within 30 s, the animal is removed from the mask and quickly orally intubated, and isoflurane is provided at 2.25% via spontaneous breathing or through a ventilator. Oral intubation using a 20-gauge stainless steel or Teflon needle can be accomplished by illuminating the ventral surface of the neck with a bright light source and then visualizing the vocal cords (4) or, more simply, by quickly exposing the trachea via a small neck incision and advancing the cannula under direct visualization of the airway. With minimal practice, intubation can be accomplished in less than 30 s and a steady flow of anesthetic can be reestablished well before the animal begins to awaken. Of course, inhaled anesthetics are also ideal for recovery procedures because they are so well tolerated and have fast recovery times. After cessation of isoflurane, mice usually begin to regain consciousness within 2 min and are fully ambulatory in 5–10 min. These properties make it an ideal choice for the implantation of indwelling catheters or telemetry probes or for surgical manipulations, such as transthoracic aortic banding or coronary artery ligation. Furthermore, because of its minimal cardiodepressant effects and quick induction and quick recovery time, the use of isoflurane is a preferred choice for brief analytic procedures that require sedation, such as echocardiography (80).
We also used ketamine mixtures in our lab for both recovery procedures and nonrecovery analytic procedures. For mice, a mixture of 67 mg/ml ketamine, 3.3 mg/ml xylazine, and 1.7 mg/ml acepromazine given intraperitoneally at a dose of 1.5 μl/g body wt seems to work quite well. This cocktail is made by mixing standard preparations as follows: 8 ml of 100 mg/ml ketamine, 2 ml of 20 mg/ml xylazine, and 2 ml of 10 mg/ml acepromazine. This mixture has moderate cardiodepressant effects, but is well tolerated and produces a surgical plane of anesthesia for ∼1 h when given as a single intraperitoneal injection. We commonly use this preparation for recovery surgeries such as catheter implantation and renal artery clipping. It provides ample time to complete the surgery, permits free manipulation of the animal's posture and position, and the animal generally is fully ambulatory within 2 h. In addition, we used this anesthetic for the analysis of airway reactivity in nonrecovery experiments. Unlike the barbiturates, the ketamine mixtures appear to preserve more of the airway responsiveness to bronchoconstrictor challenges such as acetylcholine administration.
For very short procedures, we often used avertin, which has the advantage of having very fast induction and recovery times. Induction of a surgical plane of anesthesia usually occurs within 2–3 min of injection and can last 15–30 min; animals also awake quite suddenly, becoming ambulatory within minutes of the initial signs of recovery. The preparation, which is no longer available commercially, is made as a 100% stock solution by mixing 10 g of tribromoethanol with 10 ml of tertiary amyl alcohol. This stock is then diluted to 2.5% and given as an intraperitoneal injection at a dose of 15–17 μl/g body wt (44). Both solutions should be stored in the dark at 4°C, and, in our experience, the 100% stock is stable for at least 6 mo, but the 2.5% stock should be prepared fresh at least every 2 wk. We often used smaller doses (10–12 μl/g body wt) as a sedative for performing echocardiographic measurements, but even at these lower doses, avertin is cardiodepressant, often yielding heart rates of <400 beats/min and blood pressures of ∼70–80 mmHg. Avertin, which is most often used for embryo transfer, can be an effective anesthetic for brief recovery surgical procedures, but, because of its short duration of action and rapid recovery period, it should be used with caution. Supplements can be given, but there are reports of acute necrotic and inflammatory changes with higher doses and increased mortality with repeated administrations (95). In any case, investigators should keep in mind that the choice of anesthesia should depend not only on the procedure being performed, but also on the organ system or systems being studied. For instance, urethane is often used in the rat by neuroscientists because it permits maintained neural activity and blood pressure, but it can be a poor choice for renal investigations because of intraperitoneal toxicity and derangements in renal hemodynamics and fluid handling (38, 75).
It has been fortuitous, but not altogether coincidental, that the advent of genetically altered mice as models for physiological investigation has corresponded with significant advancements in electronics, instrumentation, and computer technology. Microchip-based instrumentation has permitted the miniaturization of sensors and transducers to a size and performance level suitable for the mouse. In addition, computer technology and digital recording equipment, along with significant advances in data archival media, have allowed for greater accuracy, ease, and dependability in recording equipment.
In general, there have been two obstacles to overcome in adapting existing methodologies for use in the mouse: the size, which is about one-tenth the size of the traditional rat model, and the speed or frequency-response requirements, which are up to twofold greater than in the rat. Equipment manufacturers, having been convinced of the importance and endurance of mouse models, have for the most part been responsive to the particular challenges presented by an animal that weighs only 30 g. For example, since first being used in the intact mouse, high-fidelity microtip transducers from Millar Instruments (Houston, TX) have been reduced in size by 20% (from 2 Fr. to 1.4 Fr.), making left ventricular pressure measurements routine (59). Transit-time flow probes (Transonics Systems) have been miniaturized to a point where they are capable of accurately measuring flow in vessels as small as 0.25 mm in diameter, such as the renal artery (33), and further miniaturization is on the horizon. Telemetric implants have even been miniaturized to the point where heart rate, temperature, and blood pressure can be monitored continually in unrestrained mice for months at a time (5,24t can be as high as 30,000 mmHg/s at a heart rate of ∼700 beats/min. Signals that occur this quickly cannot be monitored with conventional fluid-filled catheters, because the signal is dampened by the necessarily small lumen diameter. In addition, conventional paper strip-chart recorders are unable to accurately follow signals at this high rate. To achieve sufficient frequency-response characteristics, left ventricular pressure measurements are therefore typically made using Millar Mikro-Tip pressure transducers coupled to a computer-based recording system operating at a sampling speed of 1,000 or even 2,000 Hz (59). A sample tracing of left ventricular pressure and dP/dt is shown in Fig. 1. Figure 1, top, shows the actual data points that describe the ventricular pressure pulse and illustrates the importance of a high frequency-response, both in measuring and recording equipment. At baseline function (Fig. 1 A, dP/dt ≅10,000 mmHg/s), the majority of the pressure increase during systole occurs in ∼10 ms and it is clear that a sampling rate of 1,000 Hz is adequate. At extremely elevated performance levels (Fig. 1 B, dP/dt ≅ 30,000 mmHg/s), the majority of the pressure increase occurs in <5 ms, and a higher sampling rate is necessary to faithfully record this waveform. By contrast, dP/dt values in the rat range from ∼7,000 to perhaps 16,000 mmHg/s. Although current technology allows for fluid-filled catheter systems that are nicely responsive (frequency-response curves that are flat to perhaps 150 Hz), these figures also demonstrate that only high-fidelity microtip transducers, with a flat frequency response up to 10,000 Hz (i.e., Millar), are sufficient for evaluating left ventricular pressure in mice. For other pressure measurements, we found fluid-filled systems to be more than adequate.
Vessel isolation and cannulation.
Many, if not most, of the techniques used to evaluate cardiovascular, renal or pulmonary function involve the surgical isolation and cannulation of an artery and or vein and the techniques used do not vary substantially from those used in larger animals. The right tools are essential for these procedures, and great care must be taken to prevent even slight blood loss: the total blood volume of a mouse is 2–3 ml and the loss of even 100 μl of blood can represent a significant hemorrhage. We found use of a low-power binocular dissecting microscope to be essential for procedures, and adding ×0.5 objectives can increase the field of view as well as the working distance. Dissecting tools include several pairs of very fine forceps (i.e., Dumont #5, straight or angled), small dissecting scissors, and a set of Vannas scissors. For fluid-filled catheters, we found it convenient to pull very fine cannulas over an open flame from large-bore, thick-walled polyethylene tubing. With the use of this process, the tubing is placed just above the hot flame until it becomes molten and then it is removed and quickly stretched to produce a fine capillary. With a little practice, this approach can be used to make catheters of almost any dimension. To minimize the dead space for venous catheters, we start with polyethylene tubing with an outer diameter (OD) of ¼ in. and a wall thickness (WT) of in. (6.4 mm OD × 1.6 mm WT, Fisher Scientific). Conversely, to increase the relative lumen size for arterial catheters, we start with polyethylene tubing with an OD of or ½ in. (9.5 or 12.7 mm) and a WT of in. The resulting catheter has a relatively large ID-to-OD ratio and therefore better frequency-response characteristics than conventional PE tubing (i.e., stretched PE-10) when measuring arterial blood pressure. With this approach, we have been able to fashion catheters as small as 100 μm OD and have managed to insert up to four catheters into a single femoral vein.
For actual cannulation, the vessel is carefully isolated from surrounding tissue and associated vessels using two pairs of fine forceps. We found it useful to polish the tips of Dumont #5 forceps to remove burrs and to slightly blunt them to prevent accidental puncture of the vessel (the veins are especially delicate and can be easily torn). Once isolated, three loops of 7–0 silk suture are placed around the vessel; the distal-most ligature is tied to prevent back flow of blood and enough tension is placed on the proximal ligature so as to temporarily occlude flow. The center ligature is loosely knotted and is to be used to secure the catheter in place and prevent leakage when the proximal tie is released. Once blood flow is occluded, a short longitudinal incision is made in the vessel and the cannula is introduced. Alternatively, some investigators use a 23- or 25-gauge hypodermic needle, bent at an angle at the tip, to puncture the vessel and to use as an introducer. Once the cannula is in place, the central ligature is tightened and the tension is removed from the proximal tie; the catheter can then be advanced further and secured into place. All surgical incisions are closed using cyanoacrylate adhesive, even in nonrecovery procedures, to prevent evaporative fluid loss and tissue desiccation.
It is often useful to monitor various cardiovascular parameters over long periods of time, weeks to months, without surgical interventions. We have used three well-described techniques to accomplish this: tail-cuff measurement of blood pressure, echocardiography, and exercise tolerance. In addition, radiotelemetry devices have recently been developed that are small enough to permit measurement of a variety of variables such as electrocardiogram, temperature, activity, and blood pressure. Finally, for in-depth analysis of cardiovascular function in vivo, high-fidelity measurements of left ventricular function, cardiac output, and regional blood flow can be made in acute and, in some cases, chronic preparations.
The measurement of systolic pressure through tail sphygmomanometry has been a standard technique for the long-term evaluation of blood pressure in the rat and has been applied to the mouse under a variety of experimental paradigms (7, 42, 85). Although similar measurements in the mouse are technically more challenging, updated instrumentation, using computer-controlled pulse detection and data-acquisition technology, was recently developed and validated and provides a practical approach for these measurements in the mouse (53). We are currently using an integrated system (Visitech, Apex, NC) that permits measurement of tail-cuff pressure on four mice simultaneously and requires a minimal training period (4–7 days). Using this system, we can perform measurements on 12–16 mice per hour. As with any tail-cuff approach, measurements must be obtained over an extended period of days (usually ∼5–7 days, in addition to the training period), to obtain a reasonable determination of blood pressure in a group of animals. As an illustration of the effectiveness of the tail-cuff approach in the mouse, we examined the development of hypertension in mice that were fed N G-nitro-l-arginine methyl ester (to inhibit nitric oxide production) over a 30-day period and the results are shown in Fig. 2. We have used this system extensively and have found that it can generally discern differences in blood pressure as low as 10–15 mmHg.
Although the tail-cuff method provides a simple estimate of arterial blood pressure, it has a number of disadvantages that can limit its usefulness. Primary among these is the stress it produces in the subject animals. Although acclimating the animals to the instrument helps to some degree, it must be recognized that these blood pressure measurements are taken under some degree of anxiety and stress. This is most easily evidenced by the high heart rates measured by tail cuff: as high as 650–700 beats/min, compared with ∼500 beats/min in the resting animal. Unlike the rat, which usually becomes quite comfortable in restraint after moderate training, the stress level in mice (as reflected in heart rate) remains elevated even after several weeks of training. Fortunately, the available evidence indicates that the blood pressure measurements are less affected by the imposed stress (61). Measurements of tail-cuff pressure are also notoriously variable from reading to reading, and this problem can be exacerbated by the software detection algorithms designed to automatically find the systolic pressure end point during cuff inflation. To explore this potential problem we recorded tail-cuff measurements while simultaneously monitoring intra-arterial blood pressure using implanted telemetry devices; a sample tracing is shown in Fig. 3. When the systolic end point is clearly discernible and is without movement artifact, there is very good agreement between arterial pressure and tail-cuff pressure and most of the variation is due to actual moment-to-moment variation in the subject's blood pressure. However, investigators should be cautioned that although the automatic detection algorithms are generally quite good when the animal is quiet, movement artifacts, which occur quite often even in trained animals, can result in erroneous measurements that should be discarded. It is therefore important to visually inspect the pulse tracings while making these measurements and to make note of those measurements that do not conform to preestablished criteria. It is also important to take multiple readings to compensate for the overall variability (see Fig. 2 legend). Finally, it is interesting to note that the tail-cuff pressure generally correlates well with mean, rather than systolic, pressure, a finding that is consistent with previous reports (53). We presume this observation to be due to a significant pressure gradient between the carotid artery measurement site and the tail. This would not be surprising given the relatively small caliber (∼0.5–1.0 mm) of the aorta. Thus, as in the rat, the tail-cuff method can be a valuable approach for screening and monitoring changes in blood pressure over extended periods of time if used with proper caution and diligence. However, as is also the case in the rat, the limitations of the tail-cuff approach prescribe that independent verification of blood pressure phenotype by an alternative method be obtained.
In similar fashion to evaluation of tail-cuff pressure over extended periods of time, it is often useful to evaluate cardiac function over a period of weeks or months, and transthoracic echocardiography of the mouse heart has become well established as a method for providing useful indexes of myocardial performance. Use of M-mode and pulsed-Doppler echocardiography to assess ventricular function in the mouse has gained widespread use as a noninvasive method for evaluating cardiac function and has been previously reviewed in detail (12, 45, 46). Although echocardiographic methods in mice do not differ conceptually from those in rat, actual images are more difficult to obtain due to the small size, increased rate of shortening, and increased relative distance of the heart from the chest wall. As newer, more advanced imaging technology has become available, spatial and temporal resolution for imaging the mouse heart has become more and more accurate. Two-dimensionally directed M-mode echocardiography permits a number of parameters to be determined, including left ventricular end-diastolic and end-systolic dimension and posterior and anterior (septal) wall thickness. From these dimensions, one can calculate variables such as left ventricular fractional shortening and left ventricular mass. Furthermore, pulsed Doppler measurements of transvalvular aortic and mitral flow can be used to obtain ejection phase indexes of systolic and diastolic function. When combined with heart rate-matched M-mode images, one can calculate a variety of performance indexes such as aortic acceleration and ejection time, velocity of circumferential shortening, peak aortic flow velocity, and early and late diastolic transmitral velocity. Although Doppler measurements of flow velocity are limited by the ability to direct the ultrasound beam parallel to flow in the mouse, measurements such as ejection time and acceleration time can be reliably determined. Newer generation high-frequency transducers and scanners provide real time analysis of two-dimensional images with high temporal resolution, allowing more accurate serial assessment of left ventricular mass (20).
In our laboratory, mice are sedated with an intraperitoneal injection of 2.5% avertin (10 μl/g body wt) and the chest is shaved. The animal is placed in a supine position and warmed using an isothermal heating pad. Ultrasound studies are performed using an Interspec Apogee X-200 ultrasonograph and a dynamically focused 9-MHz annular array transducer with an axial resolution of 0.2 mm. Acoustic coupling is achieved through a gel-filled offset applied to the shaved chest. Two-dimensional targeted M-mode studies are taken from either the long or short axis at the level of greatest left ventricular diameter. Systolic aortic outflow and diastolic transmitral inflow velocities are determined from angulated parasternal long-axis views using a pulsed wave Doppler transducer. Left ventricular function can be determined under baseline conditions and during β-adrenergic stimulation by intraperitoneal injection of isoproterenol. Tracings are recorded on S-VHS tape, and frozen images are later digitized. Measurements are made from the digital images using image analysis software (NIH Image), and three beats are averaged for each measurement. Sample M-mode and pulsed-Doppler wave forms are shown in Fig.4. In the M-mode image in Fig.4 A, the anterior and posterior left ventricular walls are clearly defined and permit reasonably accurate determination of end-systolic and end-diastolic dimensions and wall thickness. As illustrated in Fig. 4 B, Doppler flow recordings allow time-based measurements such as heart rate (from R-R interval), acceleration time, and ejection time, as well as measurement of flow-dependent variables such as peak aortic velocity.
A pressing issue in echocardiographic analysis is the method of sedation, if any. Numerous studies reported on the effects of various anesthetics on echo-derived heart function, and several labs reported studies performed without sedation. A recent study compared xylazine-ketamine vs. avertin anesthesia on cardiac function in mice using both echocardiography and closed-chest catheterization (40). The authors reported that compared with avertin, xylazine-ketamine anesthesia produced severe bradycardia with secondary effects on loading conditions, and their findings illustrated the need for comprehensive assessment of left ventricular function in the mouse with specific attention demanded regarding the mode of sedation. Another study compared the effects of halothane vs. ketamine-xylazine (10), which also concluded that echo-derived parameters are highly sensitive to anesthetic regimen and that halothane is superior to ketamine-xylazine for obtaining relevant performance parameters. The feasibility and effectiveness of performing echocardiography studies in restrained nonanesthetized mice have also been evaluated (94). In this study, parameters obtained from awake mice were compared with those from pentobarbital- and ketamine-xylazine-anesthetized mice. Predictably, findings showed that both anesthetics result in a marked decrement in heart rate and performance variables compared with awake mice; it should be noted, however, that the echo-derived heart rate in awake mice was substantially higher than heart rates determined in undisturbed animals by telemetry. In any case, it is clear that the interpretation of echocardiographic data must be interpreted with caution, including serious consideration of the anesthetic state of the animal. As with tail-cuff measurement of blood pressure, it is recommended that differences in cardiac performance observed with echocardiography be verified using an alternative methodology (see Cardiac catheterization).
Experimental mouse models with deficits in cardiovascular function can be expected to exhibit an impaired ability to tolerate exercise. It has been our experience, in fact, that many animal models that appear phenotypically normal under resting conditions can exhibit deficits in their ability to tolerate even moderate exercise, and such measurements can provide a useful integrated index of overall cardiovascular integrity. To objectively evaluate exercise tolerance in mice, we combined an exercise treadmill (Omnipacer, Accuscan Instruments, Columbus, OH) with a custom-designed, personal computer-based detection system, permitting automatic quantification of an animal's performance (27). A motor-driven treadmill, with adjustable belt speed and shock grid, was modified by installing an infrared detection system consisting of two infrared detectors and logic circuitry assembled on a printed circuit board. When a mouse running on the treadmill blocks the infrared sensor, this “failure” produces a signal to the digital input/output board inside a personal computer, which is then recorded and displayed in real time by custom software. This program checks and records the status of the infrared switches (on/off) once every second to determine the exercise status of the subject. A similar system was described in detail in a report that evaluated exercise performance, in conjunction with telemetric recording of heart rate, in mice overexpressing ventricular myosin regulatory light chain (24). In our experience, it is important to acclimate the animals to the treadmill before evaluating performance. However, as the goal is usually to evaluate basal exercise performance rather than the response to exercise conditioning, we prefer to keep the acclimation regimen quite mild, usually 15 min at a relatively slow speed (10 m/min) for 2 or 3 days. Our measurement protocol, then, usually consists of a 40–50 min exercise period in which the treadmill is set at a 7° incline and the speed is increased by 5 m/min every 10 min. This protocol is repeated three times over 5 days.
Telemetry and indwelling catheters.
It is accepted that the most reliable of cardiovascular measurements are those obtained using radiotelemetry. Although implants suitable for measuring electrocardiogram, temperature, and activity in mice have been available for a number of years (24, 48, 49), only recently has the miniaturization of these devices permitted blood pressure measurement in the mouse (65). These devices (Data Sciences International, St. Paul, MN), which enable long-term continuous recording of blood pressure without restraint, have become accepted as the gold standard for accurate evaluation of blood pressure in the intact, conscious animal. Although these implants provide significant benefit over other methodologies, their use in the mouse has presented several potential limitations. First, despite being significantly miniaturized, the transmitters are still relatively large: ∼1 × 1.5 × 2.3 cm in size (∼ 2.3 ml) and ∼3.5 g, roughly 10–15% of the normal body weight of a mouse. In earlier applications, the sensing catheters on these devices were implanted in the abdominal aorta in a nonoccluding fashion, with intraperitoneal placement of the transmitter, and their recommended use was limited to 30 g or larger mice. Furthermore, even in these larger mice, the size of the catheter in relation to the abdominal aorta was such that could often occlude flow to the hindquarters, leading to a discouragingly low success rate. To solve these problems, Carlson and Wyss (6) recently published a report describing carotid cannulation and subcutaneous placement of the transmitter on the animal's back, obviating the need for the more invasive abdominal implantation. Using this protocol, these investigators reported a success rate of >90% in mice as small as 19 g. This technique was further modified so that the transmitter was placed along the flank of the animal, greatly simplifying the implantation procedure (5). The advantages of subcutaneous implantation are such that it is recommended under certain circumstances for use in rat as well as in the mouse: it is less stressful surgically and is characterized by a faster return to presurgical weight and circadian patterns than abdominal placement. Nonetheless, it is important to note that telemetry studies have consistently demonstrated that full recovery from anesthesia and surgery does not occur for 5–7 days, as indicated by the return of normal circadian rhythms in activity, blood pressure, and heart rate.
Other disadvantages of radiotelemetry are that it is expensive and does not allow vascular access for the administration of experimental agents. As an alternative, several investigators have adapted the use of indwelling catheters and swivel tethering systems to monitor intra-arterial blood pressure in mice (47, 50, 61, 62). In this procedure, catheters are implanted in the femoral vessels or carotid artery and tunneled subcutaneously to the nape and exteriorized through a spring, which is secured to the animal's back via a Teflon button. The spring can then be connected to a swivel device at the top of the cage to allow free movement of the tethered animal. Although catheter construction varies widely, we found that polyurethane tubing (Micro-Renathane, 0.25 mm OD, Braintree Scientific, Braintree, MA) pulled to a small diameter in hot oil is very effective and usually remains patent for extended periods. With the use of this approach, stable blood pressure and heart rate measurements have been recorded for as long as 5 wk (61). Because this technique allows for continual vascular access and blood pressure monitoring without disturbance to the animal, it can be used to evaluate a variety of cardiovascular parameters, such as the sensitivity of blood pressure to precisely regulated electrolyte intake, baroreflex sensitivity, and blood pressure and heart rate variability (47, 50).
As indicated in an earlier section of this discussion (seeInstrumentationt recordings obtained during the experiment to then be recalibrated using the values obtained at the end of the experiment.
Although measurements of left ventricular pressure have proved to be extremely valuable in assessing cardiac performance in genetically altered mice, the analysis of pressure alone cannot account for potential differences in loading conditions of the heart from one mouse to the next. To reliably evaluate cardiac contractility, several groups of investigators sought to simultaneously evaluate left ventricular pressure and volume to obtain load-independent indexes of contractile function. For example, we reported the use of echocardiography in conjunction with simultaneous recording of left ventricular pressure to derive end-systolic pressure-dimension relationships in intact mice (13). Although this approach proved effective, it did not permit continuous beat-to beat evaluation of pressure-volume relationships and the image-based data analysis was extremely labor intensive. Investigators also successfully used sonomicrometry to evaluate cardiac dimension during left ventricular pressure measurements (18, 54). In this procedure, piezoelectric crystals are implanted on or within the myocardium in an open-chest preparation, and measurements of left ventricular dimension are recorded simultaneously with pressure recordings via a Millar catheter placed in the left ventricular cavity. As demonstrated in one recent study (18), this technique can be quite effective when used with two pairs of crystals and can yield reasonable pressure-volume relationships. However, this technique has the distinct disadvantage of being very technically demanding and highly invasive, requiring significant intrathoracic manipulation as well as open-chest measurements. This is probably reflected in low values of cardiac output and ejection fraction (<5 ml/min and 30%, respectively). The most promising and productive technique for evaluating pressure-volume relationships in mice has been by conductance measurement using a Millar 1.4 pressure catheter with four integrated platinum electrodes. This method, first reported in the mouse by Georgakopoulos et al. (31), permits the generation of an instantaneous signal for left ventricular volume that is based on the time varying conductance measured by the two pairs of electrodes within the left ventricular chamber. Essentially, changes in the electric field generated by one pair of electrodes during chamber filling and ejection is sensed as a change in voltage in the other pair of sensing electrodes. Signal conditioning and processing is accomplished through specialized instrumentation, such as the Aria-1 Conductance System from Millar Instruments. This technique can be used in the closed- or open-chest animal and has the advantages of being no more technically demanding than standard left ventricular catheterization. Disadvantages include difficulty in calibrating the volume signal and the need to account for the component of the conductance signal that is not dependent on chamber dimension, that is, the parallel conductance due to surrounding structures (i.e., myocardium). There have been several reports that explore the use of this technique in the mouse in considerable depth and the reader is directed to these references for more technical consideration of the relevant issues (22, 23, 30,31, 93). Sample pressure-volume traces from an isoflurane-anesthetized animal using the conductance method are shown in Fig. 5.
Because cardiac contractility can be significantly influenced by changes in heart rate, it can often be difficult to compare cardiac performance in animals that have different resting heart rates. Furthermore, in our experience, it is not unusual for genetically manipulated animals to exhibit chronotropic as well as inotropic phenotypes. Thus, in many instances it can be useful, if not essential, to electrically pace the heart in an in vivo preparation. To accomplish this, a bipolar pacing wire with a blunted tip is inserted into the right jugular vein and advanced to the right atrium. This pacing electrode is then connected to a suitable stimulator, and a normal sinus rhythm can be produced at any rate above the normal resting heart rate. In this manner, ventricular performance in animals with differing resting heart rates can be compared by pacing them at a common frequency (15). In addition to providing a means to normalize heart rate between animals, atrial pacing allows for the determination of valuable force-interval relationships in mouse models of cardiac dysfunction. The most straightforward of these is the force-frequency relationship, which has been interpreted in terms of variable amounts of Ca2+ made available to the contractile myofilaments as a function of the interval between beats.
Blood flow and cardiac output.
Evaluation of regional blood flow and cardiac output is also an important component for thorough analysis of any cardiovascular phenotype, and a variety of approaches and techniques has been adopted for use in the mouse. In a comprehensive series of studies in a mouse model overexpressing atrial natriuretic factor, direct intraventricular injection of radioactive microspheres was used to estimate blood volume, cardiac output, and regional blood flows in intact, awake mice (1, 2). Although effective, this method has the disadvantage of yielding measurements at only one time point. In addition, a number of attempts has been made to use indicator dilution methods for analyzing cardiac output in mice, but this methodology is difficult to apply in small animals for a variety of reasons, including loss of diffusible indicator in the pulmonary vascular circuit, lack of vascular access to the pulmonary artery, and limited capacity for blood sampling (36, 86). In one study, however, recordings of blood conductivity during bolus injections of 5% glucose were used to successfully estimate cardiac output in mice as well as rats (86). Another alternative is the use of pulsed-Doppler flowmetry, which has been used both invasively and noninvasively in mice (41, 68). In our studies, a 1.5-Fr. Doppler flow probe with a 20-MHz crystal was used in conjunction with a Doppler flowmeter from Millar instruments. Surgical preparation is nearly identical to that described above for the Millar Mikro-Tip pressure transducer, except that the tip of the Doppler crystal is positioned in the ascending aorta ∼5 mm from the aortic valve and guided by the online recording so that a maximum peak flow velocity and stable wave form are achieved. This technique has the advantage of being able to monitor online changes in pulsatile blood flow but the disadvantage of only measuring flow velocity, which is difficult to convert to absolute measurements of bulk flow.
Perivascular flow probes have also been effectively employed in mice, and although their use was originally limited by the size of the available probes, more recent models have been miniaturized sufficiently to permit faithful recordings of regional blood flow and cardiac output. Transit-time flowmetry from Transonics Systems (Ithaca, NY) has become the gold standard for measuring blood flow in the mouse and has been applied to measurements of cardiac output in the open-chest preparation (26), as well as renal (33), descending aorta (31), and carotid (21) blood flows. The transit time flow probes are generally simple to use and very reliable and require surgical isolation of only a short segment of artery (2–4 mm). After the probe is placed around the desired artery, the acoustical path is filled with a couplant. We easily manage to measure renal, carotid, hindlimb, and mesenteric blood flows in mice as small as 20 g. A new generation of probes is also available that should more readily permit chronic placement of probes for monitoring blood flow in the awake animal. This advancement is especially welcome in terms of measuring cardiac output, because various methods have yielded vastly contrasting values. Open-chest measurements of cardiac output using flow probes generally yield values of 3–6 ml/min, whereas preliminary measurements in closed-chest animals are in the area of 15 ml/min. This latter value would be more consistent with other techniques used to evaluate aortic flow, including conductivity dilution (∼15 ml/min) and radioactive microspheres (∼16 ml/min) (1, 86).
The techniques applied for the evaluation of renal function in mice have been previously reviewed and are similar to those used in the rat, except of course for limitations in the size of plasma and urine samples (56, 63, 69). Analysis of pressure and flow relationships in the kidney present the same challenges as in cardiovascular studies as discussed above. Here we will discuss three major approaches for measuring renal function: balance studies in awake mice, clearance and hemodynamic studies in anesthetized mice, and micropuncture studies.
Analysis of long-term electrolyte balance has been successfully performed in mice, using either rat metabolic cages or specially designed cages for mice, which are now available commercially. In our lab we use a simple cage design consisting of a Plexiglas cylinder, divided into an upper chamber, which houses the mouse, and a lower chamber, which houses a tear drop-shaped glass ball that effectively separates feces from urine (55, 64, 74). Mattson and coworkers (11, 62) combined balance studies with the implantation of indwelling arterial and venous catheters to permit monitoring of blood pressure, blood sampling (with simultaneous replacement), and electrolyte infusion. It is essential to recognize that the blood volume of a typical 30-g mouse is very small, perhaps 2–2.5 ml, and therefore that blood sampling must be kept to a minimum. A blood sample of even 100 μl represents a significant hemorrhage and can be expected to have hemodynamic consequences. When possible, blood samples should therefore be replaced with an equal amount of blood from a donor animal.
We managed in our lab to miniaturize many of the common assays for evaluation of renal function to the point where only a few microliters of sample are required. New generation blood gas instruments are able to analyze several electrolytes on small capillary samples. For example, we use a Chiron model 348 blood gas analyzer (Medfield, MA) that can measure Po 2, Pco 2, pH, Na+, K+, and Cl− (or Ca2+) on 40 μl of whole blood. Using a Corning model 480 flame photometer (Medfield, MA), we are also able to measure Na+ and K+ on only 4 μl of plasma or urine by manually diluting samples rather than using the automatic diluter. Plasma and urine osmolality can be measured using a Fiske One-Ten freezing-point depression osmometer (Norwood, MA), and the 15-μl sample can be largely recovered after the analysis. A Labconco model 442 digital chloridometer (Kansas City, MO) can measure Cl− on 10-μl samples. Finally, we used a standard picric acid-based creatinine assay to measure endogenous creatinine levels in mice (available from Sigma, St. Louis, MO; procedure No. 555). We miniaturized this assay by using half-area 96-well microplates so that the total reaction volume is 120 μl. However, creatinine concentration in mouse plasma is very low (∼0.25 mg/dl), necessitating the use of rather large sample volumes in the assay: we typically use 40 μl of plasma and 80 μl of alkaline picrate to perform the assay. It is also important to note that mouse plasma contains a large amount of non-creatinine cromagens; for this reason it is imperative to measure the absorbance of the sample before and after acidification of the reaction volume, the difference being equivalent to the level of creatinine-picrate complex. Blood samples can be obtained from awake mice by saphenous vein puncture or by tail cut and in sedated mice by retroorbital puncture (43).
Clearance and hemodynamic studies.
In acutely instrumented mice, measurements of electrolyte excretion, inulin clearance, and para-aminohippurate (PAH) clearance are commonly used. To evaluate glomerular filtration rate (GFR), we successfully adapted the use of FITC-inulin for use in both whole kidney and micropuncture samples (57). For whole kidney clearance measurements, animals are infused with a 1% solution of FITC-inulin (Sigma) at 0.15 μl · min−1 · g body wt−1. Plasma samples (20–30 μl) are drawn midway through each urine collection period, with replacement of whole donor blood. Plasma or urine aliquots of 4 μl are then diluted with 196 μl of 10 mM HEPES buffer (pH 7.4) into 96-well microplates. The samples are then analyzed using a microplate fluorometer with an excitation at 485 nm and emission at 538 nm. In our experience, GFRs range between 0.8 and 1.0 ml · min−1 · g kidney wt−1(57, 60), which is consistent with most of the current literature (8, 9, 35, 78) and is also consistent with values obtained from rat. It is interesting to note that on a kidney weight basis, GFR in mice and rats are comparable, but when corrected for body weight, GFR in the mouse is perhaps twice that in the rat, reflecting the increased kidney weight-to-body weight ratio in mice.
The estimation of renal plasma flow using PAH has also been used in mice by several investigators. Spurney and coworkers (78) used [3H]PAH to measure renal plasma flow and reported values ∼2.5 ml · min−1 · g kidney wt−1. More recently, several groups reported the use of a colorimetric assay for evaluating PAH clearance in mice infused with 2–5% PAH (9, 88), but the findings were limited due to sampling restrictions. These investigators reported PAH clearances ranging from 2 to 4.5 ml · min−1 · g kidney wt−1. We miniaturized a colorimetric assay for PAH (90), enabling measurement on as little as 10 μl of plasma. In this assay we dilute plasma and urine sample 1:10 with dichloracetic acid and, after centrifuging, mix 40 μl of the supernatant with 40 μl of the color reagent (p-dimethylaminobenzaldehyde) into 96-well microplates, which are read at a wavelength of 450 nm. By cannulating the renal vein to sample renal venous effluent blood, we found that the extraction ratio for PAH generally ranges between 0.8 and 0.9 at plasma PAH concentrations up to at least 0.08 mg/ml. These values compare favorably to those obtained in rats, which are generally reported to range between 0.6 and 0.9. It is valuable to note that the plasma concentration of PAH remains <0.1 mg/ml even at infusion rates of 6 μg · g body wt−1 · min−1(corresponding to an infusion of 4% PAH at a rate of 3 μl/min). Values for renal blood flow obtained from these experiments averaged ∼3 ml · min−1 · g kidney wt−1.
Direct measurements of renal blood flow using flow probes are generally preferable to estimates using PAH clearance, and current technology using transit time and/or laser-Doppler flowmetry permits such approaches. Gross and coworkers (32-35) evaluated renal blood flow and pressure-natriuresis responses using a Transonics Systems 0.5 mm V-series perivascular flow probe to measure total renal blood flow and two fiber optic strands in conjunction with a Transonics laser-Doppler flowmeter to determine cortical and medullary flow. They reported total renal blood flow values of ∼7 ml · min−1 · g kidney wt−1, which correlates reasonably well with values obtained from PAH measurements. These investigators also found that total, as well as cortical renal, blood flow was well autoregulated between 80 and 140 mmHg in normal mice but that medullary blood flow was not autoregulated. These investigators used long-term changes in renal perfusion pressure (induced by ligating the celiac and mesenteric arteries and lower abdominal aorta) to evaluate autoregulatory behavior. We and others used an aortic clamp to transiently alter renal perfusion pressure to evaluate autoregulation (82). Sample tracings from these experiments are shown in Fig.6, demonstrating efficient autoregulatory behavior in the mouse above ∼90 mmHg. Although these data are qualitatively similar to those obtained from rat, the upper end of the pressure-flow relationship has not been determined in the mouse, because it has proved to be difficult to increase renal perfusion pressure above ∼150 mmHg.
In vivo approaches for renal micropuncture and microperfusion have proved to be quite feasible, requiring only moderate adaptation of the techniques used in rat (see Refs. 71, 72 for review). For example, Wang and coworkers (89) reported in situ microperfusion of proximal tubules to evaluate net rates of fluid and bicarbonate reabsorption in Na/H exchanger isoform 3 (NHE3) knockout mice. Free flow micropuncture measurements were reported in transgenic mice as early as 1994 (77) and since then, several different mutant models have been studied by analyzing samples obtained from both proximal and distal tubules (60, 73,84). Free-flow measurements yielded a profile of nephron function that is not altogether different from that seen in the rat. Although values of single nephron GFR are lower in mice, averaging perhaps 12–15 nl/min compared with 30–35 nl/min in rat, fractional reabsorption from late proximal and early distal puncture site is comparable to the rat: 40–50% and 70–80%, respectively, in normal animals. In practical terms, we found that proximal punctures in the mouse are on the same size and flow scale as distal collections in the rat and mouse distal collections are somewhat smaller. Preparation of the kidney, although similar to that in the rat, requires considerably more care due to the small size. Also, the ureter of the mouse kidney is tightly adherent to the medial margin of the kidney and is therefore very difficult to dissect free without bleeding. For this reason, we modified the Lucite kidney holder by adding a second opening: the first opening is located traditionally, near the center of the holder, so as to accommodate the renal vessels emerging from the hilum, and the second opening is located in one corner so as to accommodate the ureter emerging from the caudal pole of the kidney.
We miniaturized the technique for evaluating FITC-inulin clearance to include measurements of single nephron GFR (57). With the use of this approach, micropuncture samples (5 nl or greater) are deposited between oil columns in a small 1-μl microcapillary tube (microcaps, Drummond Scientific, Broomall, PA). After 0.5–1 nl of 500 mM HEPES (pH = 7.4) is added to each sample to normalize pH, the microcaps are placed on the stage of an inverted microscope fluorometer (of the sort typically used for intracellular calcium measurements). With the use of an excitation wavelength of 480 nm and emission wavelength of 530 nm, the samples are digitally imaged and then analyzed for fluorescence intensity. This technique has the advantages of being simple to use, highly sensitive, inexpensive, and nonradioactive; furthermore, it can be performed on samples as small as 5 nl and it does not consume the sample.
Evaluation of tubuloglomerular feedback (TGF) responses, primarily by measuring changes in stop-flow pressure in the open-loop configuration, has also proved to be feasible in a wide variety of genetic mouse models, and the technique is the same as that used in the rat. We found, for example, that TGF responses in NHE3 knockout mice are intact compared with their wild-type littermates (60). In these studies, the stop-flow pressure measurements were also supported by analysis of proximal-distal differences in single neprhon GFR, which compares the filtration rate under conditions of intact and interrupted flow to the macula densa and can be viewed as a measure of the prevailing strength of the TGF signal at the time of the measurement. NHE3 knockouts demonstrated a significant difference in distal vs. proximal single nephron GFR, reflecting the persistence of a robust TGF response in these animals. Sample traces of stop-flow pressure at varying tubule perfusion rates obtained in our laboratory are shown in Fig. 7. It is interesting to note that the dynamic range of the TGF response is much lower than in the rat; that is, the loop flow rate causing a half-maximum response is between 10 and 15 nl/min (∼14 nl/min in the example shown), consistent with the lower endogenous flow rate observed in the mouse. These low flow rates are difficult to discriminate because of limitations in the perfusion equipment, and whereas differences in maximal responses should be apparent using the stop-flow technique, shifts in the TGF relationship may be more difficult to discern in mice. This possibility is perhaps illustrated in the study noted above (56). Although stop-flow pressure measurements and proximal-distal single nephron GFR differences both confirmed activity of the TGF system in NHE3 knockouts, the stop-flow data suggested that the system was not different between the two groups of mice, whereas the proximal-distal data suggested that the strength of the TGF signal was augmented in the knockouts. Similar findings have been reported in other strains of mice (83).
As in cardiovascular and renal investigations, analysis of pulmonary function can be performed chronically in awake mice by whole body plethysmography or acutely in anesthetized mice by measurement of airway pressure-flow relationships. Parameters that can be readily evaluated include breathing frequency, tidal and minute volumes, airway reactivity (resistance), lung compliance, and diffusion capacity.
Several commercial systems are available for plethysmographic measurements of respiratory function in mice and these can be found in single chamber or two-chamber models. We have had experience with the BioSystem XA from Buxco Electronics (Sharon, CT), which can continuously evaluate a number of derived parameters in unrestrained, awake mice. This unit consists of a whole body chamber and integral reference chamber and pneumotachograph. A bias flow of various gas mixtures (O2 and N2), controlled by needle flow valves, can be delivered to a port on the chamber to maintain O2 concentration within the chamber at a desired level. An aerosol inlet port is also available for delivery of nebulized bronchoactive agents. When this plethysmograph is sealed, a respiring mouse creates pressure fluctuations, which relate to the animal's tidal volume when the animal is breathing quietly and to the effort of breathing when the breathing is labored. Alternatively, when the pneumotach is open, fluctuations relate to the animal's flow rate when the animal is breathing quietly and to the effort of breathing when the breathing is labored. Analog pressure/flow signals from the plethysmograph are analyzed by the software package to evaluate the following parameters: inspiratory and expiratory time, peak inspiratory and expiratory flow, tidal volume, relaxation time, minute ventilation, frequency of breathing rate, end-inspiratory and end-expiratory pause, and enhanced pause (Penh). Use and validation of this system for measuring these variables was previously published (39) and the technique has been used extensively.
To evaluate the effectiveness of this system for monitoring alterations in ventilation in a normal mouse, we measured the above parameters under normal conditions and immediately after exposure to hypoxia (21% and 10% O2, respectively). As the data in Fig.8 demonstrate, a distinct hyperventilation (increases in breathing frequency, tidal volume and minute ventilation) in response to breathing low O2 is evident using this system. Investigators should be aware, however, that these types of measurements must be made in resting mice that have been well acclimated to plethysmograph chamber. We observed that active (nonresting) mice typically display irregular, high-frequency/low-volume breathing patterns (as high as 650 breaths/min), which are consistent with virtually constant sniffing behavior.
Airway responsiveness to bronchoconstrictor challenge is probably the most widely measured parameter using whole body plethysmography. Generally, increasing doses of methacholine, ranging from 1 to 100 mg/ml, are delivered to the animal by aerosol through a port in the top of the chamber. Although airway resistance cannot be directly measured using a single chamber plethysmograph, one of the derived parameters, Penh, has been shown to reflect changes in airway resistance under appropriate conditions. This dimensionless parameter is based on the notion that during bronchoconstriction, the main alteration in the pressure signal from the chamber occurs during early expiration (when the air in the lungs become more greatly compressed) and that the resulting change in the recorded waveform can be quantified by comparing the chamber pressure during early expiration with the pressure during late expiration. Because the calculation of Penh is related to respiratory timing, rather than on pressure-flow relationships, it is perhaps most effectively used as a screening tool for evaluating differences in airway reactivity between groups of mice. It is recommended, therefore, that any observed differences in Penh be confirmed by a more thorough analysis of pressure-flow relationships to evaluate airway resistance.
Pulmonary mechanics and airway reactivity.
Whole body plethysmography provides a convenient noninvasive and effective method for screening pulmonary function over extended periods of time. However, to more precisely evaluate pulmonary mechanics and airway reactivity, acute methods can be employed to generate a variety of functional parameters in the mouse. Again there are several commercial systems available that offer the necessary instrumentation and software to make sophisticated measurements of lung function. Such systems generally involve intubation or surgical tracheotomy and subsequent measurements of airway pressure and flow relationships. In its simplest form, for example, lung compliance can be measured by generating pressure-volume relationships in excised lungs or in dead mice by incrementally inflating and deflating the lungs with a syringe attached to a manometer (51, 81).
To evaluate pulmonary mechanics and airway reactivity in our lab, we have adopted the use of the Mouse Reactivity System (MRS), which was developed by Costa and coworkers at the US Environmental Protection Agency (Research Triangle Park, NC) (28, 52). In this procedure, mice are anesthetized with a ketamine-xylazine-acepromazine mixture, as described earlier, and neuromuscular blockade is achieved with an intravenous injection of decamethonium (10 μg/g body wt). We found that this anesthetic regimen is more effective for preserving airway reactivity than the use of barbiturates. The animal is connected to an endotracheal tube with five ports (see Fig. 9). Two of the ports are connected to the inspiratory and expiratory ports of a computer-controlled ventilator. One side port is connected to a pressure sensor and monitors airway pressure and the final two side-ports are connected to a highly sensitive differential pressure transducer and serve as a pneumotach for monitoring airway flow. Mechanical ventilation is accomplished by positive pressure using a mass flow controller (MFC) that delivers a constant inspiratory flow rate of 50–60 ml/min. The inspiratory/expiratory cycle is controlled electronically by the MRS through a pair of solenoid valves, so that inspiratory and expiratory times are both 250 ms in duration. This respiratory pattern results in a frequency of 120 breaths/min, a tidal volume of 0.225–0.250 ml, and a minute ventilation of 27–30 ml/min, and establishes arterial blood gas values at normal levels. A recent paper by Volgyesi et al. (87) described a similar system that provides constant flow inflation and includes end-inspiratory and end-expiratory pauses for the evaluation of airway resistance and compliance. As these investigators point out, in contrast to larger animals, the mouse can be especially vulnerable to dynamic hyperinflation during periods of bronchoconstriction, which can be avoided by lengthening the expiratory time.
The airway pressure and flow signals are recorded and analyzed by the MRS to determine peak inspiratory and expiratory flow, total lung resistance, and dynamic lung compliance according to the algorithms described by Bates and co-workers (3), and illustrated in Fig. 9. Airway outlet pressure (Pao) and tracheal air flow (differential pressure from the pneumotach) are directly measured and lung tidal volume is determined by integrating the flow signal. Then, on a breath-by-breath basis, total lung resistance is calculated as they19). If desired, animals can also be analyzed during spontaneous breathing to determine spontaneous inspiratory and expiratory time, tidal volume, respiratory rate, and minute volume.
It has been our experience that measurements of lung compliance can be highly variable, depending on the methodology used to derive pressure/volume relationships in the lung. Although the computerized system described above provides an estimate of dynamic lung compliance, we found that these measurements often do not correlate well with values of static lung compliance more commonly measured in mice postmortem. Thus to evaluate static lung compliance in living mice, the computer-controlled ventilation system described above was adapted to allow the operator to carefully control the ventilation pattern during inspiration and expiration. To make measurements of static compliance, tidal ventilation is interrupted for a single cycle of maximal lung inflation followed by an interrupted deflation, as illustrated in Fig.10. The deflation occurs in sequential steps of passive exhalation (2-ms duration), each followed by airway occlusion (300-ms duration). At points of zero airflow (∼250 ms after blocking expiration), airway pressure is recorded and analyzed, and the maximal slope of the volume-pressure curve represents the static lung compliance. In this manner we are able to make repeated determinations of static lung compliance over extended periods of time (up to 4 h). Furthermore, by nonsurgically intubating mice, we have been able to make repeated measurements on separate days. Values of static compliance made in this manner correlate well with postmortem measurements obtained by syringe inflation of the lung (see dashed line in Fig. 10 C). There are a variety of commercial systems available for making similar measurements in mice and they make use of several different approaches for evaluating lung mechanics. For example, the flexiVent system from Scireg (Montreal, Quebec) can evaluate lung resistance and elastance and, using a forced oscillation technique, can measure lung impedance, from which a variety of other parameters can be derived.
Measurements of diffusion capacity are also an important component in evaluating overall pulmonary function. In mice, as in other species, diffusing capacity can be evaluated from the disappearance rate of CO2 from alveolar air spaces, either in a single breath or rebreathing method (16). As an alternative, we recently measured the rate of gas exchange by adopting a method in which anesthetized mice are first ventilated at a fixed rate with 95% O2-5% CO2 for 10 min. During this time, two arterial blood gas sample are obtained to establish a baseline for arterial Pco 2. Mice are then ventilated with 100% O2, and arterial blood gas samples are taken at 30 s, 3 min, 6 min, and 9 min. The rate of Pco 2 decay in the arterial samples is taken as an index of the rate of gas exchange (52, 92).
Another important functional parameter in the lung is the rate of water and electrolyte transport. Alveolar fluid clearance has been measured in the mouse lung using radioactive albumin as a marker. In this procedure, the lungs of ventilated mice are instilled with ∼0.3 ml of Ringer solution containing 131I-labeled albumin. At fixed time points after instillation, fluid samples are withdrawn from the lungs and analyzed for total protein and labeled albumin. The relative rate of change in albumin concentration provides an index of fluid clearance from the lungs. To evaluate the regulation of fluid transport, inhibitors such as amiloride and ouabain can be included in the instillate (25, 76).
The use of genetically altered mice for studying the functional role of specific proteins or protein polymorphisms has become widespread. With this flexibility in molecular technology has come the realization that cross-disciplinary approaches are necessary to fully exploit these powerful animal models. Because many of the investigators generating these new animal models often have limited experience in physiological methodologies, it has been the goal of this review to provide some practical insights regarding the use of various techniques commonly used to evaluate organ function in the intact mouse. Although this review is not intended to be comprehensive, it has hopefully provided a brief description of some of the practicalities involved in performing these experiments and, perhaps more importantly, an awareness of some of the common problems and challenges that can be encountered with mice and how these challenges can be met.
The author thanks Drs. R. Paul, G. Shull, W. Noonan, and W. Cupples for critical review of this manuscript and helpful suggestions.
This work was supported by National Institutes of Health Grants DK-57552–01, DK-50594–05, and HL-58010–01.
Address for reprint requests and other correspondence: J. N. Lorenz, Dept of Mol. And Cell. Physiology, Univ. of Cincinnati, PO Box 670576, Cincinnati, OH 45267-0576 (E-mail:).
- Copyright © 2002 the American Physiological Society