INTRODUCTION

Top Abstract Introduction Methods Results and discussion References

THE ABILITY TO SUSTAIN muscular exercise is dependent in large part on the ability of the muscles to generate ATP aerobically at rates commensurate with the energy demands of the task. In the steady state, however, the aerobic cost [O₂ uptake (O₂)] of a given work rate is not appreciably different among subjects differing in age, gender, physical fitness, or state of training. The clues to the control, therefore, reside in the non-steady-state response profiles. Although it is generally agreed that O₂ in muscle is controlled by features of the high-energy phosphate bond utilization {e.g., change in creatine phosphate (PCr) concentration, the phosphorylation potential [ADP ×(P_i /ATP)], or the free energy for ATP splitting}, much of the detail remains to be elucidated (4, 11).

A major impediment to resolving these issues in humans has been the inability to establish precisely parameters of the kinetics of O₂ and of its putative intramuscular control mediators simultaneously during exercise. Furthermore, the amplitude of the responses should be appropriate to allow the kinetics to be characterized with sufficient precision for justifiable control inferences to be drawn (9). We have, therefore, developed a technique that allows the dynamic features of O₂ and the concentrations of the relevant intramuscular phosphate metabolites to be established simultaneously, by using remote sensing of respired volume and gas concentrations, during quadriceps exercise performed in a whole body NMR superconducting magnet.

Quadriceps exercise has been used only rarely for spectroscopic investigations (5, 13, 14) because of the technical problems of accommodating such a large muscle mass in conventional magnets. Pulmonary gas-exchange monitoring has been performed in conjunction with this exercise modality by Takahashi et al. (14), who used steady-state bag collections of exhaled gas, and by Evans et al. (5), who used a breath-by-breath approach. In neither case was the goal of the investigation to estimate dynamic profiles of response. Furthermore, in the latter study (5), which is more pertinent to the present study, no information is presented on how the breath-by-breath measurements were accomplished (for example, if the monitoring system was situated in the magnet room, how the interference problems were overcome; if it was in an adjacent shielded room, how were concerns addressed that related to issues such as analyzer delays and response fidelity?). We believe such details to be important; therefore, we describe the details of a procedure for the simultaneous determination of intramuscular phosphate profiles and breath-by-breath O₂ during exercise.

	METHODS

Top Abstract Introduction Methods Results and discussion References

Attempts to measure ventilation and pulmonary gas exchange on a breath-by-breath basis, with a human subject enclosed in a whole body NMR unit, present a series of technical challenges. The common feature is the requirement that there be no ferrous components within the vicinity of the magnet. This was achieved by housing both the ventilation measurement module (VMM; Interface Associates, Laguna Niguel, CA) and the quadrupole mass spectrometer (Airspec QP9000; Clinical and Scientific Equipment, Gillingham, Kent, UK) in the control room adjacent to the magnet. A turbine was custom designed, with high-grade stainless steel mountings, and the intake to the mass spectrometer sampling line was positioned at the mouth, as shown in Fig. 1. The cable to the VMM and the sampling capillary tube to the mass spectrometer were each 45 ft. long. However, to maintain an appropriate response fidelity for the respired-gas profiles (Figs. 2 and 3), it was necessary to increase the bore of the mass spectrometer sampling capillary from 0.5 to 1.0 mm, thus increasing the flow from 50 to 250 ml/min. Although this extended-length sampling configuration led to the transit delay's being increased from 410 to 2,680 ms, the 5-95% rise time was still <80 ms. The delay and 5-95% rise time were determined by means of a small, dead-space solenoid valve that switched between gases of two known concentrations; the switch was triggered by an electrical switch which also served as time zero for the transit-delay and rise-time measurements, as previously described (3).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Schematic representation of procedure for exercising in the whole body NMR unit. Subject exercises by alternately extending rubber bands arranged as stirrups attached to the feet. Subject breathes through a turbine volume sensor, with respired gas concentrations sampled by mass spectrometry. These devices are housed in a room remote from the NMR unit. M. Spec. Capill., mass spectrometer sampling capillary. See METHODS for further details.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   On-line breath-by-breath display of inspiratory and expiratory volume (VI and VE, respectively), and respired PCO2 and PO2 by mass spectrometry. A: display with standard mass spectrometer inlet line. B: inlet line to mass spectrometer was 45 ft. long (see METHODS). Solid vertical line, end of volitional deep inspiration. * Respired gas consequence.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Profiles of respired PCO2 and PO2 during moderate exercise with standard and extended inlet line. Profiles are phase aligned to demonstrate that fidelity of mass spectrometer response is not degraded by the remote sampling technique. See METHODS for further details.

One of the biggest limitations to NMR spectroscopy is the inherent lack of sensitivity consequent to the low signal-to-noise characteristics. Because the VMM turbine system normally operates at a frequency range up to 5 kHz, the addition of a low-frequency filter at the VMM then allowed the ³¹P signals to be sampled without this interference (e.g., Fig 4).

View larger version (64K): [in this window] [in a new window]   Fig. 4.   Sequential display of Pi, phosphocreatine (PCr), and 3 peaks of ATP during a rest-exercise-rest protocol. Display has been configured in this manner so that profile of change of Pi is not obscured by that of PCr. ppm, Parts per million.

Square-wave changes of work rate were obtained by having the subject exercise in the prone position (as shown in Fig. 1) with the transmit-receive coil placed under the right quadriceps muscles. A broad, nonelastic strap over the hips served to stabilize the subject during exercise. Rubber stirrups were attached to brass bars, which were an integral part of a custom-designed plastic insert to the magnet bore. This fitted snugly into the inner bore of the magnet with sufficient outward recoil to provide a firm friction hold. The stirrups were attached to the feet by means of a foot strap that was designed to fit snugly to the foot and, therefore, to minimize any lateral movement. On command, subjects performed alternate knee-extensor exercise over a constant excursion (i.e., the entire extent of the available bore diameter) at a constant frequency of 32 cycles/min. One of the investigators observed the leg excursions to ensure that they were maintained both in the vertical plane and over the entire bore diameter. This proved to be of no concern. Exercise timing was arranged such that the acquisition of the ³¹P NMR signal occurred with the leg stationary in the recovery phase. Thus, neither the radio frequency (RF) coil nor the volume of interest made spatial excursions, as with cycling-type exercise (6), allowing good-quality spectra to be obtained. Different work rates were achieved by substituting rubber stirrups, each with different recoil characteristics. To date, this has allowed a range of oxygen uptakes to be attained which extends to almost 2 l/min.

The experiments were carried out in a 1.5-T superconducting magnet (General Electric, Signa Advantage) with a 0.5-m bore by using a one-pulse ³¹P-NMR acquisition. A surface coil (8-in. transmit, 5-in. receive), tuned to a frequency of 25.85 MHz for phosphorus, was placed under the quadriceps of the dominant leg, midway between the knee and hip joint (Fig. 1). The coil was securely fastened to the table, and displacement of the leg over the coil was prevented by securing the subject with a broad, nonelastic strap over his hips. The magnetic field homogeneity was optimized by shimming to the proton signal of muscle water. The RF excitation pulse was set at a level to give a maximum PCr signal at a 5-s repetition rate. Free induction decays for ³¹P spectra were collected every 1,875 ms, with a spectral width of 2,500 Hz and 512 data points. Data were averaged every eight acquisitions, and the dynamic signals for the three ATP peaks, PCr, and P_i could be determined every 15 s during the rest-exercise-rest condition (Fig. 4). Signal intensities of each resonance were calculated by means of the time-domain variable projection VARPRO-fitting program and by using the appropriate prior knowledge of the ATP multiplets (12).

The T₁-saturation factor was assumed to remain constant for each resonance throughout the experiment, and all phosphate metabolite levels are given relative to their preexercise values. The techniques for the breath-to-breath ventilatory and gas-exchange determinations have previously been described in detail (3).

	RESULTS AND DISCUSSION

Top Abstract Introduction Methods Results and discussion References

The ability to make valid physiological inferences from parameter estimation of the non-steadystate response of a physiological variable to a particular dynamic forcing regime depends to a large extent on the confidence limits for the estimation. For a function that changes exponentially, such as the Variable Projection VARPRO-fitting program pulmonary O₂ response (O_{2 t}, where t is time) to square-wave dynamic exercise (8, 15)

(1)

where O_{2 ss} is the steady-state O₂ increment,  is a short delay-like component attributable to the muscle-to-lung transit delay (15), and  is the time constant of the response which is considered, on the basis of computer modeling (2), to be equivalent (within ~10%) to that of the muscle O₂ kinetics and empirically shown to be the case by Grassi et al. (7). The 95% confidence limit (K_n) for the estimation of  in Eq. 1 can be characterized by the equation

(2)

where is a constant that depends on the value of the underlying time constant and, hence, the amount of data relevant to the estimation, and S₀ is the SD of the breath-to-breath fluctuation in O₂ (9). This breath-to-breath "noise" has previously been demonstrated to be an uncorrelated Gaussian stochastic process, the SD of which is largely independent of metabolic rate (9). In Eq. 2, n is the number of independent-like transitions that have been time aligned and superimposed to minimize the influence of the breath-to-breath noise.

To date, attempts to draw control inferences from the dynamic response of O₂ with respect to that of components of change in the high-energy phosphate pool (e.g., decrease in intramuscular PCr) have been hampered by 1) the PCr and O₂ responses being determined on different days in different locations; and 2) O_2 ss being so small [on the order of 100 ml/min for plantar flexion exercise (10)]. This would require the superimposition of ~100 like transitions to achieve a comparable confidence limit for the estimation of  as for an amplitude of 1 l/min (e.g., Fig. 5, in which the work rate was of heavy intensity and the O₂ attained at the end of the work bout was 75% of this subject's maximum for this exercise modality). Attempts to provide a large value for O_{2 ss} (1) by using cycle ergometer exercise (i.e., ranging from ~350 to 900 ml/min) unfortunately did not compare the dynamic O₂ responses, either simultaneously or with a similar appropriately large-amplitude change of PCr. The kinetics of the PCr change were obtained in response to a different exercise format (plantar flexion), i.e., not only necessitating an appreciably lower work rate but also utilizing different muscles.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Representation of simultaneously determined changes of PCr concentration (top) and pulmonary O2 uptake (O2; bottom) in response to and during recovery from 6-min square-wave exercise in single representative subject.

Our attempts to overcome these technical limitations were based on two considerations: 1) utilizing an exercise format that would allow a large muscle group to be involved, i.e., providing a large O_{2 ss} (Fig. 5); and 2) determining the time-course of both O₂ and changes in the intramuscular high-energy phosphate pool simultaneously. Our strategy for remote sensing of both respired gas volumes and partial pressures of O₂ and CO₂ to be utilized in the breath-by-breath gas-exchange algorithms (3) required that the fidelity of the intrabreath gas partial pressure profile not be distorted, despite the increased delay which resulted from the extended capillary sampling line. As shown in Figs. 2 and 3, this was achieved (as described in METHODS) by increasing the diameter and the flow down the capillary tube. Figure 2 represents a subject exercising on two different occasions: with the standard 6-ft. sampling line (A) and with a 45-ft. sampling line (B). The transit delay is demonstrated by the breath marked with an asterisk; in both cases, this was a consequence of a volitional deep inspiration. Figure 3 illustrates a series of representative breaths which were aligned and overlaid in a subject who spontaneously couples his breathing frequency to the exercise cadence and who was exercising in one case with the short and in the other case with the long inlet line. The dynamic response features are functionally indistinguishable. Consequently, as the prolonged transit delay could be readily incorporated into the breath-by-breath algorithm, the magnitude and the time course of the O₂ response to constant-load cycle ergometry are indistinguishable between tests using the standard and the "extended" sampling line (Fig. 6A). This fidelity of respired gas signal transmission should be seen to be a requirement of remote sensing techniques, such as those employed in this study: appreciable slowing of the dynamic response could lead to major errors in the O₂ computation.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   A: breath-by-breath determination of pulmonary O2 in response to 6-min square-wave transition to 100 W on cycle ergometer from control phase of 20 W, with "standard" mass spectrometer sampling line () and extended (45-ft.) sampling line (). Simultaneously determined changes of pulmonary O2 (B) and PCr concentration (C) in response to and during recovery from 6-min square-wave exercise in NMR magnet on 2 occasions in a single representative subject.

The other essential requirement is that the devices used in the magnet for the breath-to-breath gas-exchange measurements should not interfere with the determination of the phosphate profiles. As shown in Fig. 4, the responses of PCr, P_i, and the three phosphate groups of ATP during a study comprising a control period, 6 min of dynamic exercise, and recovery are notably free of interference. This allows the dynamic features of the PCr response to be discerned and compared with those of O₂, either from a single transition, as shown in Fig. 5, or with respect to two identical transitions, as shown in Fig. 6, B and C.

Several additional factors are likely to limit the interpretation of phosphate and O₂ profiles during this kind of exercise. These relate to issues such as coil position and muscle recruitment patterns. We took care to minimize the influence of these factors as far as is possible. For example, we ensured that the coil was located in the same position for each individual subject and that it was located in the same relative position for different subjects. We cannot readily determine the recruitment pattern of the quadriceps during the spectroscopic measurements; neither can we ensure that precisely the same muscle areas are sampled among different subjects. However, we believe that our technique is valid to relate the dynamic profiles of O₂ to those of the intramuscular phosphates for each individual subject. The issue of contributions of muscles not sampled by the coil, especially at higher work rates (i.e., influencing O₂ but not the sampled PCr pool), is a further concern for this and other similar investigations of these control processes. We utilized the broad stabilizing strap (Fig. 1) in an attempt to minimize the influence of such extraneous contraction. The concern remains, however.

We propose, therefore, that techniques such as those described above will allow investigators to establish simultaneously the dynamic profiles of both O₂ and the intramuscular phosphate groups of interest. Furthermore, the large amplitude of response (e.g., Fig. 5) improves the likelihood that justifiable control inferences may be drawn from the response profiles, especially when several repetitions (9) are utilized to reduce the influence of the "noise."