INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE APPLICATION of genetic engineering methods to mice has greatly expanded the ability to study the molecular foundations of physiology and disease. Many of these models, including those with gene manipulations without known direct cardiac specificity, display dramatic changes in cardiac morphology and/or integrated cardiovascular physiology (9, 14, 21, 22). This has greatly increased the need for rigorous quantification of cardiovascular dynamics in intact animals (7). Such analysis is achievable in larger animals and humans on the basis of measurements of simultaneous cardiac chamber pressure and volume. Relations between and based on these variables have been established as among the most specific and precise means of assessing intrinsic passive and active properties of the intact heart, as well as quantifying its energetics and interaction with the arterial system (27). To date, however, application of this approach to mice has been stymied by methodological limitations. In this study, we demonstrate the feasibility and utility of conductance micromanometry for assessment of cardiovascular dynamics and chamber mechanics of in vivo murine hearts. We report the first pressure-volume relation analysis in this species and test the hypothesis that chamber mechanical behavior is similar between mouse and human, further supporting the use of murine models to study human disease.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Volume catheter system design. A miniaturized combined conductance catheter-micromanometer was custom fabricated to our specifications by Millar Instruments (Houston, TX). The catheter has an outer diameter of 0.44 mm, occupying ~20% of the cross-sectional area of the aortic valve and 0.76 µl of left ventricular (LV) cavity volume. It contains a micromanometer (Millar 1.4-Fr, SPR-671) flanked by two platinum electrodes 0.5 mm in length and 5 mm apart (see Fig. 1). In addition to this combined catheter, custom four-electrode conductance catheters (NuMed, 0.32 mm OD) were used in combination with a separate micromanometer (Millar, SPR-671) in these studies.

View larger version (44K): [in this window] [in a new window]   Fig. 1.   Photograph of the new miniaturized conductance-micromanometer catheter used for this study. The catheter is 0.4 mm in cross-sectional diameter. A short distal nylon extension (Ext) facilitates placement and stabilization of the catheter within the ascending aorta. The two electrodes [distal (DE) and proximal (PE)] are used to generate volume signal. DE is positioned to lie at or just above the aortic valve. A 1.4-Fr Mikrotip micromanometer [pressure transducer (PT)] is located between electrodes and falls within the midcavity of left ventricle (LV).

Animal preparation. Twenty four C3H/HeJ mice (Jackson Laboratories) of either sex weighing 20-35 g were housed under diurnal lighting conditions and allowed food and tap water ad libitum. Animal treatment and care were provided in accordance with institutional guidelines, and the protocol was approved by the Animal Care and Use Committee of the Johns Hopkins University. Anesthesia was initiated with methoxyflurane inhalation followed by intraperitoneal injection of urethan (750 mg/kg) and -chloralose (50 mg/kg) dissolved in normal saline. Supplemental intraperitoneal anesthesia (one-fifth dose) was provided if needed so that the animals remained unresponsive to tail pinch by forceps as assessed by changes in heart rate and blood pressure. A heating pad was placed underneath the animal, and the temperature was set to 37.5°C. The animals were intubated with a blunt 19-gauge needle via a tracheotomy and were ventilated with a custom-designed constant-pressure ventilator with 100% oxygen at 120 breaths/min and a tidal volume of 200 µl. The chest was entered through an anterior thoracotomy under dissecting microscope visualization, and a small apical stab was made at the LV apex, leaving the pericardium as intact as possible. The pressure-volume catheter was then advanced retrogradely into the LV along the cardiac longitudinal axis with the distal tip in the aortic root (Fig. 2A) and proximal electrode just within the endocardial wall of the LV apex. Advancing the micromanometer across the valve confirmed the lack of hemodynamically significant pressure gradients due to transvalvular catheter placement (data not shown). Catheter position was verified at autopsy. In studies in which separate pressure and volume catheters were employed, the volume catheter was placed via the LV apex as described above, and the micromanometer was placed in the midchamber through a separate small stab incision.

View larger version (58K): [in this window] [in a new window]   Fig. 2.   A: X-ray photograph of mouse on its side with the catheter positioned in LV and ascending aorta. Arrow indicates catheter. B: time-series tracings of LV volume, LV pressure, and thoracic aortic flow at both steady state and during transient inferior vena caval (IVC) occlusion. Arrow, onset of IVC occlusion. Vp, offset term. See Volume catheter system design for details. C: comparison of beat-to-beat stroke volume (SV) determined by flow probe and microconductance catheter during IVC occlusion. Results of linear regression between the 2 variables are shown. D: corresponding pressure-volume loops from time-series plots in B, with conductance volumes calibrated using regression relation of C. Dashed line, end-systolic pressure-volume relationship (ESPVR).

Statistical analysis. All data are reported as means ± SE. Differences in the intercept of the stroke volume-stroke volume plots used to calibrate the conductance catheter were analyzed using multiple linear regression. Significance was assumed at P < 0.05. Confidence intervals for allometric plots were calculated by linear regression.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Pressure-volume relations and catheter calibration. Table 1 provides baseline hemodynamics measured with our anesthesia protocol. Heart rate and systolic pressure were nearly identical to values reported in conscious mice (6), and dP/dt_max was nearly three times higher than that reported in isolated mouse heart preparations (12) and 50% greater than in other intact mouse models. Thus, despite the open-chest preparation, the cardiovascular data were very physiological. Figure 2B displays time-series data during transient inferior vena caval occlusion. The volume scale in this plot is the uncalibrated conductance catheter signal in arbitrary units. As shown in the tracings, there was a commensurate decline in chamber filling and stroke volume observed in both the conductance catheter signal and in the aortic flow probe signal. There was minimal heart rate change during the few seconds required to obtain these data. To calibrate relative volume changes measured by the catheter conductance signal, stroke volumes were calculated by beat-to-beat integration of the descending thoracic aorta volume-flow signal and were compared with simultaneous volumes derived from the uncalibrated conductance signal. As shown in Fig. 2C, the two stroke volume values were highly linearly correlated over the broad loading range. Similar results were obtained in 10 hearts, with a pooled regression r² of 0.983 and a mean intercept of 1.65 µl (P < 0.0001). The conductance signal was converted with these regressions yielding relative volumes in microliters. The small nonzero intercept implied slight nonlinearity of the underlying catheter signal vs. true volume relation, consistent with electric field theory (30).

Assessment of changes in contractility with isoproterenol and propranolol. An important feature of pressure-volume analysis is its ability to identify cardiac-specific alterations in the intact animal. Figure 3 displays an example of the acute effects of -adrenergic stimulation (isoproterenol at 150 ng · kg¹ · min¹ ip) and -blockade (bolus injection of propranolol at 1 mg/kg iv) in the same heart. Simultaneous flow was not obtained in this example, so volume data are presented in relative units with a constant instrumentation gain and offset between interventions. The ESPVR shifted leftward and became steeper with isoproterenol and shifted rightward and became more shallow after propranolol, similar to results from other species. The volume-axis intercepts of all three relations were little altered acutely.

View larger version (33K): [in this window] [in a new window]   Fig. 3.   A: pressure-volume loops at baseline (Con; solid curves) and after administration of isoproterenol (Iso; dashed curves) and propranolol (Pro; dotted curves). Iso shifted the baseline ESPVR upward to left, consistent with increased contractility, whereas Pro did the opposite. The relations also show the flat diastolic pressure-volume curve typical of intact blood-perfused murine heart. B: relations between beat-to-beat maximum first derivative of ventricular pressure with respect to time (dP/dtmax) and end-diastolic volume (EDV) derived from data in A. These relations also provide a measure of chamber contractility (19) and show similar changes, consistent with enhancement or depression after -receptor agonist or antagonist exposure. Resting value of dP/dtmax (top right point of each relation) actually declined between control and Iso, whereas the relation itself shifted leftward, indicating a clear rise in contractility.

Mouse vs. human time-varying elastance. Pressure-volume data were used to calculate the time course of ventricular activation indexed by time-varying chamber elastance in murine hearts. The instantaneous pressure-volume ratio P(t)/[V(t)  V₀], where V₀ is the intercept of the ESPVR (in relative volume units), was calculated from rest cardiac cycles and normalized to both peak amplitude and time to achieve peak amplitude. This normalized elastance [E_N(t_N)] curve has recently been shown to be highly conserved in human hearts despite a broad range of diseases and reflects properties of muscle force development measured at the chamber level (29). Figure 4A displays comparisons of the E_N(t_N) curves derived from 52 human subjects and from 15 mice and reveals the two curves to be nearly superimposable despite profound disparities in the kinetics of chamber contraction between the species.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   A: time- and amplitude-normalized LV elastance [EN(tN)] curves describing temporal course of chamber stiffening throughout cardiac cycle. Results for mouse are compared with curve previously reported in humans (27) and display striking similarities throughout time course. Thus, although there are clearly differences between normal human and murine hearts with respect to time required to reach maximal chamber stiffness (due to nearly 10-fold difference in heart rate) and peak stiffness achieved (related to chamber size), the temporal course of chamber stiffening [i.e., the EN(tN) curve] appears to be a fundamentally conserved chamber mechanical property. B, left: allometric plot of log Ees vs. log M, where Ees is LV end-systolic elastance and M is body mass, for cow (17), human (18), pig (13), sheep (23), dog (5), rabbit (24), rat (16), and mouse. B, right: plot of log Ees vs. log Ea across species, where Ea is arterial elastance. Relation was linear (r2 = 0.98) with a slope of 1.09 and intercept of 0.15. Solid line, regression line; dashed lines, 95% confidence intervals.

Allometric plots of ventricular and effective E_a. To compare the values of E_es and E_a derived from the present analysis to values reported in other mammalian species, we made allometric plots for both variables. Data for other species were obtained from prior reported literature (5, 13, 16-18, 23, 24). The allometric equation is parameter = a × M^b, where a and b are allometric variables and M is body mass; thus, log(parameter) = log(a) + blog(M). Results of these logarithmic plots are displayed in Fig. 4B and reveal a linear inverse relation between ventricular elastance and body mass. The slope (the allometric coefficient b) for this relation was 1.27. A similar relation and slope were derived for effective E_a (1.18, data not shown). The similarity of the two slopes means there was a correlation between the two properties, E_es and E_a (see Fig. 4B, right), indicating preservation of ventricular-arterial coupling independent of heart size, heart rate, or vascular scaling (32).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study we demonstrate the first analysis of cardiac function by pressure-volume relations in the mouse. Reproducible and physiological values were obtained through this analysis, thereby extending this method, the "gold standard" for the assessment of cardiovascular dynamics in the clinic and in basic research, to the study of murine cardiovascular physiology, where there is great potential in translating the effects of genetic and molecular changes to those at the chamber level. Although use of the conductance catheter to measure intraventricular volume is hardly new, the use of the same electrodes for both stimulation and sensing represents a return to a concept first introduced by Rushmer et al. (26) and Geddes et al. (8) but later abandoned by Baan et al. (3) and others in favor of separate stimulating and sensing electrodes. Two-electrode systems can be associated with electrode polarization that may alter the resistance between them over time. Over the relatively short-term duration of its present use, however, we did not observe any change in intraelectrode resistance. Confirmation was made by comparing results to those measured with a four-electrode catheter with two stimulating and two sensing electrodes. The results were identical to those with the two-electrode system. Similar adequacy of a two-electrode system was recently reported in rabbit hearts (1). Although the resolution of volume changes even in the small murine heart might be improved by the addition of more segments, the single pair provided excellent signals, correlating strongly with flow probe data, and seems adequate. Clearly, use of fewer electrodes enables miniaturization of the catheter and makes combined pressure-volume or pressure-volume-Doppler flow catheters feasible.

The catheter volume signal is noncalibrated, and conversion to absolute volumes requires estimation of both an offset term and the amplitude of a time-varying term. Calibration of the offset has been previously achieved by either the hypertonic saline dilution method (4) or by matching of global ejection fraction and stroke volume to an independent standard. However, even small volumes of hypertonic saline in the mouse could easily result in marked hemodynamic changes, given the small circulating blood volume. Fractional shortening data based on echo imaging might be used to estimate ejection fraction, but this will require validation for the mouse. Lack of offset calibration only limits precise estimation of the volume-axis intercepts of the pressure-volume relations. By setting the gain and offset of our signal generator-voltmeter at constant values, we easily discerned relative shifts in the pressure-volume relations, as shown in Fig. 3 with propranolol.

Because the reported values for murine E_es and E_a in the present study represent the first such measurements in this species, we sought to compare the results relative to those reported in other larger mammals. Chamber volume is known to scale proportionately with body mass, M (15), whereas systolic pressure is generally conserved independently of M (33). Thus one would expect elastance (pressure/volume) to scale by M¹, yielding an allometric coefficient of 1.0. The measured value of 1.2 is close to this expected value. Similarly, the E_a should also vary by M¹, and the slope of this allometric relation was 1.18. Thus E_es and E_a covary, with a slope near 1.0 (measured slope was 1.09), indicating conserved ventricular-arterial interaction and thus optimal cardiac efficiency and power output across species.

Perhaps the most physiologically intriguing aspect of this study involved the comparison of the time course of ventricular activation between human and mouse. Despite a predominance of the V₁ myosin adenosinetriphosphatase in mice, which has been shown to provide a 50% increase in maximum velocity over the V₃ isotype (25) present in humans and larger mammals, the temporal pattern of chamber activation was remarkably conserved. This similarity in mechanical phenotype at the chamber level should greatly facilitate mechanistic studies in normal and diseased hearts and perhaps allow for direct comparisons to be made between mice and humans in relation to the pathophysiology of disease processes.

Although we obtained hemodynamics that were in many ways comparable to those measured in conscious mice, the methodology employed was invasive. In addition, the calibration in the present study was based on thoracic aortic flow and may have underestimated the true cardiac output. Ongoing efforts aim at designing catheters with a Doppler crystal at the distal end, thereby minimizing the time for surgery and the invasiveness required to obtain cardiac output for the purposes of calibrating the conductance catheter. Reversal of the positioning of the pressure and flow sensors relative to the electrodes will facilitate placement in a closed-chest animal in the future. Still, our protocol provides a method to obtain highly physiological cardiovascular hemodynamics at a single study point and should still be of considerable value to many studies.

In summary, we demonstrate that real-time continuous pressure-volume analysis can be applied in live mice, yielding physiological results with many similarities to data measured in larger animals and humans. Even in the absence of full-volume calibration, this approach can provide useful insights into contractile changes that may not be as easily interpretable on the basis of steady-state dP/dt_max measurements alone. Furthermore, combining measures of both ventricular and vascular properties provides truly integrated cardiovascular physiological assessment so that molecular mechanisms of a disease process can be determined in a species with full genetic control. Such evaluations will also be of tremendous value as we develop more specific novel therapies for cardiovascular disease.