Congestive heart failure (CHF) is most prevalent in aged individuals and elicits a spectrum of cardiovascular and muscular perturbations that impairs the ability to deliver (Q̇o2) and utilize (V̇o2) oxygen in skeletal muscle. Whether aging potentiates the CHF-induced alterations in the Q̇o2-to-V̇o2 relationship [which determines microvascular Po2 (PmvO2)] in resting and contracting skeletal muscle is unclear. We tested the hypothesis that old rats with CHF would demonstrate a greater impairment of skeletal muscle PmvO2 than observed in young rats with CHF. Phosphorescence quenching was utilized to measure spinotrapezius PmvO2 at rest and across the rest-to-contractions (1-Hz, 4–6 V) transition in young (Y) and old (O) male Fischer 344 Brown-Norway rats with CHF induced by myocardial infarction (mean left ventricular end-diastolic pressure >20 mmHg for YCHF and OCHF). In CHF muscle, aging significantly reduced resting PmvO2 (32.3 ± 3.4 Torr for YCHF and 21.3 ± 3.3 Torr for OCHF; P < 0.05) and in both YCHF and OCHF compared with their aged-matched counterparts, CHF reduced the rate of the PmvO2 fall at the onset of contractions. Moreover, across the on-transient and in the subsequent steady state, PmvO2 values in OCHF vs. YCHF were substantially lower (for steady-state, 20.4 ± 1.7 Torr for YCHF and 16.4 ± 2.0 Torr for OCHF; P < 0.05). At rest and during contractions in CHF, the pressure driving blood-muscle O2 diffusion (PmvO2) is substantially decreased in old animals. This finding suggests that muscle dysfunction and exercise intolerance in aged CHF patients might be due, in part, to the failure to maintain a sufficiently high PmvO2 to facilitate blood-muscle O2 exchange and support mitochondrial ATP production.
- heart failure
- skeletal muscle
- oxygen exchange
the pathological condition of congestive heart failure (CHF) is characterized by exercise intolerance, reductions in exercising muscle blood flow (Q̇m) (17, 38, 54, 60), and slower pulmonary oxygen uptake (V̇o2) dynamics (23, 33, 49). Aging may also be associated with multiple cardiovascular structural alterations [e.g., resistance artery rarefaction (4), decreased muscle capillary-to-fiber ratio (12, 45), but unchanged or increased capillary density (24, 30)] and functional alterations [e.g., impaired myogenic response and endothelial function in resistance arterioles (12, 16, 34, 36, 52)], which compromise the ability of structures to deliver (Q̇o2) and off-load O2 (V̇o2) within skeletal muscle. By measuring the pressure of O2 within the microvasculature (PmvO2), the dynamic balance between V̇o2 and Q̇o2 at the site of capillary-myocyte O2 exchange can be investigated (5, 7, 44). We have recently demonstrated that aged skeletal muscle exhibits a lower resting PmvO2 as well as PmvO2 dynamics characteristic of compromised convective and diffusive O2 delivery across the rest-to-contractions transition (6). It is presently unknown whether, or to what extent, aging potentiates the CHF-related perturbation in PmvO2 dynamics. However, given the aforementioned age-related decrements in skeletal muscle vascular function, it is likely that CHF concomitant with senescence could dramatically impair the matching of Q̇o2 to V̇o2 in skeletal muscle across the rest-exercise transition.
Measurements of PmvO2 via phosphorescence quenching (50) provides a real-time assessment of the relationship between muscle Q̇o2 and V̇o2 at rest and across exercise transients (5, 7), as well as detecting perturbations of this relationship (i.e., altered PmvO2 indicative of changed fractional O2 extraction) in major disease conditions (e.g., CHF) (15, 20). Specifically, PmvO2 provides an index of extraction at the microvascular level (32), demonstrated mathematically as follows: (1) Therefore, the kinetic profile of PmvO2 is representative of the dynamic Q̇O2-to-V̇o2 ratio.
In this investigation, we utilized the Fischer 344 Brown-Norway (F344xBN) rat model of aging (27) to test the general hypothesis that CHF concurrent with old age will elicit changes in microvascular oxygenation at rest and across the rest-contractions transition in the spinotrapezius muscle. Specifically, because of impaired skeletal muscle arteriolar endothelium-dependent vasodilation manifest in old rats (35), rats with CHF (14), and elderly individuals with CHF (34), we hypothesized that aging will exacerbate the degree of Q̇o2-to-V̇o2 mismatching (evidenced by a lower PmvO2) induced by the CHF condition in skeletal muscle both at rest and across the rest-contractions transition. Knowledge of these microcirculatory O2 exchange impairments in aging and pathological conditions is fundamental to understanding the mechanisms of skeletal muscle dysfunction and exercise intolerance in affected individuals.
All procedures were approved by the Institutional Animal Care and Use Committee at Kansas State University.
Young (Y) (4–6 mo; n = 20) and old (O) (≥26 mo; n = 20) male F344xBN rats were used in this study. These rats were specifically selected for this investigation because they represent young and late middle age-old rats according to the lifespan for the F344xBN strain (27). In addition, the F344xBN rat has the distinct advantage over the F344 rat because, unlike the F344, it does not develop many of the age-related pathologies that proliferate in their highly inbred cousins (9). Rats were housed individually at 23°C and were maintained on a 12:12-h light-dark cycle. All rats were fed rat chow and water ad libitum.
Myocardial infarction procedures.
Rats were assigned randomly to undergo either sham or myocardial infarction (MI) procedures, as described previously (37). Briefly, rats were anesthetized with a 5% isoflurane-O2 mixture, intubated, connected to a rodent respirator (Harvard model 680), and maintained on a 2% isoflurane-O2 mixture. A left thoracotomy was performed between the fifth and sixth ribs (∼1.5 cm in length) to expose the heart. The pericardial sac was opened, and the heart was exteriorized. In rats receiving a MI, a 6-0 suture was used to encircle and ligate the left main coronary artery ∼2–4 mm distal to its origin. Sham operations were completed by using the same surgical procedures with the exception that the coronary artery was not ligated. The lungs were hyperinflated, and the ribs approximated with 3-0 gut. The muscles of the thorax were sewn together with 4-0 gut, and the skin incision was closed with 3-0 silk. Lidocaine (1.5 mg/kg every 2 h for 8 h) and buprenorphine (0.03 mg/kg every 12 h for 24 h) were administered subcutaneously for postoperative pain alleviation, and ampicillin (50 mg/kg every 24 h for 10 days) was injected subcutaneously to minimize the chance for infection. After surgery, anesthesia was withdrawn and the rats were extubated and monitored for 8–12 h postoperation.
Ten weeks after MI or sham procedures, the rats were anesthetized with pentobarbital sodium (30 mg/kg ip, supplemented as needed). The right carotid artery was isolated and, by means of an introducer, cannulated with a 2-Fr catheter-tip pressure manometer (Millar Instruments). The manometer was advanced into the left ventricle (LV) in a retrograde fashion to measure LV end-diastolic pressure (LVEDP) and the rate of pressure change within the chamber (LV dP/dt). Subsequently, the manometer was replaced with a fluid-filled catheter (PE-50) to monitor arterial blood pressure and heart rate for the duration of the experiment (Digi-Med BPA model 200). This fluid-filled catheter was used for the administration of additional anesthesia and for the infusion of phosphorescent probe. Rectal temperature was monitored and maintained at 37°C with a heating pad.
The left spinotrapezius was exposed as described previously (2, 15). Briefly, the skin and fascia were carefully removed from the caudal portion of the dorsal aspect of the muscle. Vascular and neural tissues branch primarily from the scapular origin of the spinotrapezius and were left undisturbed. Stainless steel electrodes were used to stimulate the muscle. The cathode was placed in close proximity to the motor point (0.5–1.0 cm caudal to the scapula), whereas the anode was sutured in place at the caudal edge of the muscle, near the fourth thoracic vertebrae. Stimulation parameters (i.e., voltage and placement of electrodes) were held constant among all animals. The phosphor, palladium meso-tetra-(4-carboxyphenyl)-porphyrin dendrimer (R2; Oxygen Enterprises, Philadelphia, PA), was infused at a dose of 15 mg/kg through the arterial cannula ∼15 min before each experiment.
The muscle was kept moist with a Krebs-Henseleit bicarbonate-buffered solution equilibrated with 5% CO2-95% N2 at 37°C during a 10-min stabilization period after exposure and throughout the subsequent experiment. The muscle was stimulated to contract at 1 Hz (∼4–6 V, 2.0-ms pulse duration, twitch contractions) for 5 min with a Grass S88 stimulator. PmvO2 measurements were recorded every 2 s throughout rest and exercise.
On completion of the experiment, each rat was euthanized with an overdose of anesthesia (50 mg/kg ia pentobarbital sodium). The thorax was opened, and the lungs and heart were excised. The right ventricle (RV) was separated from the LV, and all tissues were weighed and normalized to the body weight of each animal.
PmvO2 measurements and calculations.
The probe of a PMOD 1000 frequency domain phosphorimeter (Oxygen Enterprises, Philadelphia, PA) was positioned ∼2 mm above the spinotrapezius, as described by Bailey et al. (2). A light guide contained within the probe focused excitation light (524 nm) on the medial region of the exposed spinotrapezius (∼2.0 mm diameter, to ∼500 μm deep). The PMOD 1000 uses a sinusoidal modulation of the excitation light (524 nm) at frequencies between 100 Hz and 20 kHz, which allows phosphorescence lifetime measurements from 10 μs to ∼ 2.5 ms. In the single-frequency mode, 10 scans (100 ms) were used to acquire the resultant lifetime of the phosphorescence (700 nm) and repeated every 2 s (for review, see Ref. 56). The phosphorescence lifetime was obtained computationally based on the decomposition of data vectors to a linearly independent set of exponentials (57).
The Stern-Volmer relationship allows the calculation of PmvO2 from a measured phosphorescence lifetime using the following equation (50): (2) where kQ is the quenching constant (Torr/s) and to and t are the phosphorescence lifetimes in the absence of O2 and at the ambient O2 concentration, respectively. For R2, in in vitro conditions similar to those found in the blood, kQ is 409 Torr/s and to is 601 μs (29). Because the R2 is tightly bound to albumin in the plasma and is negatively charged, in combination with the extremely high albumin reflection coefficients in skeletal muscle (for review, see Ref. 47), the Po2 measurements are ensured to result from signals within the microvasculature, rather than the surrounding muscle tissue (43). The phosphorescence lifetime is insensitive to probe concentration, excitation light intensity, and absorbance by other chromophores in the tissue (50). The effects of pH and temperature are negligible within the normal physiological range, which was maintained herein (29, 41).
Citrate synthase activity.
Citrate synthase activity (CSa), a mitochondrial enzyme and marker of muscle oxidative potential, was measured in duplicate from spinotrapezius muscle homogenates according to the method of Srere (53). CSa, expressed as micromoles per minute per gram wet weight, was measured spectrophotometrically with a Spectramax 190 microplate (Molecular Devices, Sunnyvale, CA) in 300-μl aliquots at 30°C.
From anatomic dissection and morphological measurements, MI rats were categorized based on lung congestion [lung weight-to-body weight ratio (LW/BW)] and RV hypertrophy [RV to body weight ratio (RV/BW)]. Rats with both LW/BW and RV/BW greater than 4 standard deviations above the mean for the age-matched sham animals were considered to be in CHF (YCHF and OCHF groups). Rats receiving a sham operation comprised the sham (Ysham and Osham) groups.
KaleidaGraph software (Kaleidagraph 3.5) was used to describe the time course of each PmvO2 response using an exponential function, following a time delay (TD): (3) where τ is the time constant of the response, and ΔPo2 (SS) is the difference between rest and the steady-state value.
When a marked undershoot occurred in the PmvO2 response before the attainment of a steady state (i.e., PmvO2 falling transiently below the steady-state value), a second exponential term was included in the model to reduce the residual sum of squares: (4) where A1 and A2 are the amplitudes of the two components of the response, respectively. For Y responses (both sham and CHF groups), the single exponential with TD provided an excellent fit to the PmvO2 data at the onset of contractions as judged from 1) coefficient of determination (r2), 2) sum of the squared residuals, and 3) visual inspection of the raw data and the fit of the residual error to a linear model (7). The PmvO2 responses from old animals required a more complex model with two exponentials (as described above), each with independent delays, to fit the PmvO2 response (6, 15). A two-way ANOVA among groups was performed on mean arterial pressure (MAP), heart rate (HR), LVEDP, LV dP/dt, LW/BW, RV/BW, CSa, baseline PmvO2, end-contracting PmvO2, PmvO2 deltalow (i.e., baseline PmvO2 − minimum PmvO2), and from the modeling results (TD, τ, and k PmvO2; i.e., PmvO2 deltalow/τ). A Student-Newman-Keuls test was used for post hoc analysis. To determine whether the PmvO2 “undershoot” demonstrated significance, we calculated the z-statistic to test the null hypothesis that the undershoot was not different from zero for sham and the mean of the sham for CHF (13). In all instances, a significance level of P < 0.05 was accepted.
Successful experiments were performed on eight Ysham and nine Osham rats and five rats for each of the CHF groups. The remaining animals were either euthanized for humane reasons during the 10-wk period after receiving a MI, did not fit the criteria for CHF as detailed in methods, or demonstrated unstable blood pressure responses during anesthesia and were not included in data analyses. CSa was reduced significantly in the spinotrapezius of YCHF and OCHF groups vs. values for age-matched sham animals (Table 1). In addition, there was a strong tendency (P = 0.08) for CSa values to be lower in OCHF than in YCHF spinotrapezius muscles (Table 1).
Indexes of heart failure and hemodynamic variables.
As illustrated in Fig. 1, there was clear evidence of lung congestion and RV hypertrophy in the CHF groups (i.e., greatly increased LW/BW and RV/BW; for further clarification, see methods). In addition, LVEDP was increased and LV dP/dt was depressed to a similar degree in YCHF and OCHF animals (Fig. 1). By these criteria, therefore, both YCHF and OCHF animals demonstrated an equivalent severity of heart failure. MAP and HR were lower in CHF animals than in age-matched sham animals (Table 1).
Representative PmvO2 profiles for both sham and CHF groups in response to electrical stimulation are illustrated in Figs. 2 and 3 for Y and O spinotrapezius, respectively. In Osham and OCHF groups, the precontracting baseline PmvO2 was depressed significantly vs. condition-matched Y values (Table 2). Furthermore, the PmvO2 baseline was 20% lower in OCHF than in Osham. In response to contractions, the change in steady-state (precontracting baseline − end-contracting) PmvO2 was lower in Osham than in Ysham group (8.4 ± 0.9 vs. 12.5 ± 1.9 Torr; P < 0.05) and in OCHF than in YCHF group (5.8 ± 0.5 vs. 12.6 ± 2.8 Torr; P < 0.05). In Osham vs. Ysham, there were no significant differences in the initial TD and time constant (Table 2). However, the Osham and OCHF responses required the more complex two-exponential model fit because of the presence of a PmvO2 undershoot (end-contracting PmvO2 minus absolute nadir value at the end of the exponential fall; see Fig. 3 and Table 2), which was not apparent in the Ysham response (Fig. 2). In addition, the PmvO2 undershoot was greater in Osham vs. Ysham (2.5 ± 0.9 and 0.9 ± 0.6 Torr, respectively; P < 0.05) as well as in OCHF vs. YCHF (Table 2). Neither the Ysham nor YCHF PmvO2 undershoots were significantly different from zero, and the two-exponential model did not provide a better fit (r value P > 0.05). For both Osham and OCHF animals, the secondary rise of PmvO2 began ∼2 min after the onset of contractions (Table 2).
The present investigation demonstrates that the driving pressure of oxygen in the microcirculation (i.e., PmvO2), which facilitates transcapillary O2 flux, is significantly lower in skeletal muscle from OCHF than in muscle from YCHF rats at rest and at any given time point across the rest-contractions transition (Fig. 4). These findings demonstrate that CHF induces a blunting of the dynamic PmvO2 response (i.e., altered Q̇o2-to-V̇o2 ratio), which is exacerbated by aging. The lower PmvO2 exhibited in OCHF rats could potentially increase phosphocreatine (PCr) degradation and glycogen utilization in the contracting muscle (58), ultimately contributing to premature fatigue. In addition, because the slowing of PmvO2 kinetics with CHF (Figs. 2 and 3) may reflect a reduced net muscle O2 exchange, the altered PmvO2 profiles demonstrated herein provide a mechanistic link to the slowed pulmonary V̇o2 kinetics that is commonly found in heart failure patients. Specifically, reducing O2 availability (i.e., PmvO2) can exert a modulatory effect on V̇o2 kinetics (18, 25, 28). Therefore, the lower PmvO2 values in OCHF would be predicted to slow muscle V̇o2 kinetics (and accordingly pulmonary V̇o2 kinetics) (3) as a result of the low capillary O2 driving pressure.
Plasticity of Q̇o2 and V̇o2 with CHF: impact of aging.
In the present investigation, we sought to compare the effects of CHF with aging on the potential for skeletal muscle O2 exchange. Indeed, Y and O rats that underwent the MI procedure demonstrated similar degrees of severe LV dysfunction and CHF based on predetermined hemodynamic and morphometric criteria (Fig. 1). In Y rats, CHF elicits several structural (e.g., reduced capillary-to-fiber ratio) (59) and functional (e.g., reduced endothelium-dependent vasodilation) (14) modifications within the microcirculation, which impact the ability to deliver and distribute O2 within skeletal muscle. In spinotrapezius muscle of YCHF rats, for example, there is a reduced percentage of capillaries supporting continuous red blood cell (RBC) flow at rest (65% in CHF vs. 87% in Sham) (26) and during the contracting steady state (48), thus diminishing the functional O2 diffusing capacity of the muscle (DmO2) (19). Even with the reduced DmO2 in YCHF rats, blood flow increases adequately with contractions to maintain fairly high PmvO2 values that are commensurate with their age-matched sham groups. Thus any reduction in blood-muscle O2 transport would not be caused by a reduced O2 driving pressure (PmvO2), as measured herein, but rather might be caused by some combination of the impaired microcirculatory perfusion (26, 48) and reduced muscle oxidative capacity (Table 1).
In contrast, in OCHF, there is expected to be a cumulative reduction in functional DmO2 due to a reduced percentage of RBC-perfused capillaries as seen in YCHF animals (26, 48), which would be compounded by the effects of aging (i.e., reduced lineal density of RBC-perfused capillaries) (51). Both at rest and during the transition to muscular contractions, the ability to deliver O2 (i.e., Q̇o2) is impaired in the old rat (Osham, Fig. 3) (6). Furthermore, in spinotrapezius muscle from OCHF rats, the effects of venous congestion (i.e., elevated LVEDP, Fig. 1) and a lower perfusion pressure (MAP, Table 1) would act to reduce the blood pressure differential across the capillary bed (31). This reduced arteriovenous pressure difference might retard RBC flow in perfused capillaries and consequently prolong the interaction between the RBC and myocyte, allowing the Po2 in the RBC to more closely approximate extravascular Po2 values (i.e., lower PmvO2).
Mathematically, the reduced PmvO2 in Osham and OCHF animals could be the consequence of either a reduced Q̇o2 (as discussed previously) and/or an elevated V̇o2. The latter option does not appear to be valid in this instance because the lower PmvO2, at least in the Osham, occurs in the presence of an unchanged resting V̇o2 in spinotrapezius from O vs. Y adult animals (51). In contrast, in CHF, there is evidence that resting metabolic rate (RMR) is elevated (42, 55) and that the extent of the elevation is correlated with the severity of heart failure (40). Several mechanisms may contribute to an increased RMR in CHF, including 1) a reduced bioavailability of nitric oxide [reduced nitric oxide would relieve the inhibition of cytochrome-c oxidase (10, 61) and possibly increase V̇o2], 2) increased catecholamines [e.g., epinephrine-stimulated increase in RMR (46)], and 3) potential alterations in mitochondrial uncoupling proteins. An increased resting V̇o2 of OCHF spinotrapezius muscle coupled with a reduced Q̇o2 provides one plausible mechanism for the substantially reduced PmvO2 observed. One factor that does mitigate against an increased resting V̇o2 in OCHF is the oxidative potential (CSa), which showed a tendency (P = 0.08) to be reduced in OCHF vs. YCHF (Table 1). Nonetheless, the Q̇o2-to-V̇o2 ratio is clearly attenuated in OCHF, which, as illustrated in Fig. 4, results in a lower PmvO2 at any point from rest to the contracting steady state. This reduced capillary O2 driving pressure may, in part, be responsible for the greater reliance on nonaerobic energy sources (e.g., PCr) after exercise onset manifest in skeletal muscle from CHF patients (1, 11, 39). Therefore, in the OCHF condition, skeletal muscle V̇o2 kinetics, which closely mirror pulmonary V̇o2 kinetics (3, 22), would be slowed significantly due to reduced PmvO2 and an impaired DmO2, which both act to compromise O2 delivery.
CHF and PmvO2 dynamics.
The primary aim of this investigation was to explore the effects of aging on the PmvO2 dynamics at rest and in the steady state of contractions in animals with CHF. However, it is insightful to consider briefly the alterations in PmvO2 dynamics observed in YCHF and OCHF animals. In the present investigation, severe heart failure was present in both Y and O animals, which resulted in a reduced CSa in both groups. Indeed, the blunted PmvO2 dynamics in the YCHF animals in the present study is similar to that found previously for Y animals with severe, but not moderate, indexes of heart failure (15). This would suggest that the reduced oxidative capacity of the YCHF and OCHF muscle may have a greater effect on the PmvO2 profile than altered Q̇o2 kinetics per se. The delay before PmvO2 declined across the first few seconds of contractions, i.e., primary TD was not different between YCHF and OCHF (Table 2), indicating that the initial increase (i.e., what has been termed phase I) of the blood flow response might not be different across age groups in the CHF condition. However, the presence of the PmvO2 “undershoot” in OCHF (Table 2) suggests that the secondary increase in blood flow (i.e., phase II) could be sluggish relative to that of V̇o2 (5, 21). Therefore, it appears that the effects of CHF on muscle hemodynamics (Q̇o2) and metabolic responses (V̇o2) is compounded by the aging process. The mechanistic basis for this observation may well relate to the impaired endothelium-dependent vasodilation present in old muscle (35) and in CHF (14). It is plausible that CHF, concomitant with aging, results in a severe reduction in the ability of resistance vessels to vasodilate in response to endothelium-mediated challenges (e.g., flow-induced) and consequently that Q̇o2 dynamics across the rest-contractions transition might be slowed, forcing PmvO2 to very low values.
To control for the strength of contractions in this preparation, we held the relative intensity of contractions (i.e., voltage range 4–6 V) as well as electrode placement on the motor point constant for all animals. Therefore, it was assumed that a similar recruitment pattern occurred with respect to motor unit activation and the number of fibers stimulated among groups. Although technical considerations precluded measurement of arterial blood gases in the present investigation, we have previously published evidence that neither CHF nor aging in the rodent model alter arterial blood gases (15, 51).
MAP was lower in the CHF groups vs. sham animals (Table 1). In healthy animals, reductions in MAP to ∼70 mmHg do not discernibly affect PmvO2 kinetics (8). In contrast, the lower MAP in CHF animals, coupled with venous congestion (i.e., elevated LVEDP; Fig. 1), may have constrained blood flow dynamics and thus PmvO2 dynamics in response to muscular contractions. However, despite the lower MAP values, there was no significant correlation (P > 0.1) between either the PmvO2 primary TD or time constant and MAP in any of the CHF or sham responses.
There is significant experimental evidence suggesting that CHF alters the relationship between O2 delivery (Q̇o2) and O2 consumption (V̇o2) within skeletal muscle. This study demonstrates for the first time that aging significantly reduces PmvO2 in the spinotrapezius muscle of CHF rats at rest along with producing alterations in PmvO2 dynamics during the transition from rest to muscle contractions, thereby suggesting that important age-related changes in the Q̇o2-to-V̇o2 relationship occur in the CHF condition. The lower PmvO2 values in old animals with and without CHF would act to reduce V̇o2 in accordance with Fick's law, as well as elicit greater perturbations to the intracellular milieu (i.e., increased H+ production and greater PCr degradation) (58). These findings, therefore, provide a number of potential mechanisms that could contribute to the extreme slowing of the pulmonary V̇o2 kinetics found in elderly CHF patients. Whether the slowing of pulmonary V̇o2 kinetics found in elderly CHF patients is due to impairment of Q̇o2 relative to V̇o2 within the muscle will require further investigation.
This work was supported, in part, by National Institutes of Health Grants AG-19228 (T. I. Musch), HL-50306 (D. C. Poole), and AG-25622 and HL-71270 (B. J. Behnke); a Grant-in-Aid from the American Heart Association, Heartland Affiliate (D. C. Poole); and National Aeronautics and Space Administration Grants NAG2-1340 and NCC2-1166 (M. D. Delp).
The authors thank K. Sue Hageman for excellent technical assistance. In addition, this study could not have been completed without the assistance of Drs. Paul McDonough and Danielle Padilla as well as Clay Greeson, Kyle Ross, and John Russell.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society