INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE ROLE OF O₂ as a modulator of intracellular bioenergetics is often overlooked, but its importance has been well documented (10, 11, 31, 32). Recent skeletal muscle studies have demonstrated that intracellular PO₂ was reduced by systemic hypoxia (20), and this may explain the reduced maximal mitochondrial respiratory rate [maximum O₂ uptake (O_{2 max})] observed in these conditions (19). These data may represent the initial steep portion of a hyperbolic relationship between in vivo maximal muscle O_{2 max} and intracellular PO₂ and support the theory that O_{2 max} in exercise-trained muscle is normally limited by O₂ supply. However, measurements under hyperoxic conditions were not performed. Consequently, the relationship between these variables, with increased rather than decreased O₂ availability, has not been resolved.

In combination, magnetic resonance spectroscopy to determine myoglobin (Mb) saturation, an endogenous probe of tissue oxygenation (29), and the functionally isolated human quadriceps muscle model (2) provide the opportunity to study the relationships between intracellular and intravascular events in humans. With the use of these techniques, the aim of this study was to elucidate the relationship between skeletal muscle O_{2 max} and intracellular PO₂ under conditions of hypoxia (12% O₂), normoxia (21% O₂), and, in particular, hyperoxia (100% O₂). We specifically wished to test the hypothesis that increased intracellular PO₂ would allow an increase in O_{2 max} but that diminishing returns would be evident with increasing O₂ availability as cellular metabolism makes the transition from O₂ supply dependence to O₂ supply independence.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Five healthy nonsmoking male competitive bicycle racers regularly riding 200-400 miles/wk volunteered to participate in this study, which had been approved by the University of California San Diego Human Subjects Committee. Health histories and physical examinations were completed, and written informed consent was obtained according to the University of California San Diego Human Subjects Committee requirements. The physical characteristics of the subjects were as follows: age, 25.0 ± 1.9 (SE) yr; height 184.9 ± 2.0 cm; and weight 81.5 ± 3.8 kg. Subjects performed an incremental cycle ergometry test (100-W initial work rate increased by 25 W/min until fatigue), and their O_{2 max} averaged 4.76 ± 1.2 l/min or 58.4 ± 1.1 ml · min¹ · kg¹ in this test.

Exercise model. Subjects lay supine on a padded bed with the knee-extensor (KE) ergometer placed in front of them (illustrated in Ref. 20). The resistance to KE was provided via a fiberglass bar attached to the crank of the ergometer and to a specially designed shin brace worn by the subject. This ergometer was a prototype, and as such work rates could not be measured in conventional units of work, but work rate was prescribed and measured as a percentage of the maximum resistance (weight in g, resisting the fly-wheel turning) at the end of the preliminary graded maximal test. Sixty dynamic contractions of the KE muscles per minute were performed. This exercise model isolates muscular contraction to the quadriceps femoris muscle group, a muscle group with a roughly equal number of slow-twitch and fast-twitch muscle fibers (17) and a Mb concentration of 4-6 mg/g of muscle (15). Contractions of this muscle group caused the lower part of the leg to extend from ~90 to 170° flexion. The momentum of the flywheel returned the relaxed leg to the start position. Subject reports, force tracings, electromyography, and T2-weighted magnetic resonance imaging (2, 18, 22) support the conclusion that active contractions are limited to the quadriceps muscles during this exercise modality.

Preliminary tests. Subjects performed two to three training bouts on the dynamic KE apparatus to ensure familiarity with this exercise modality. The final two practice periods involved two graded maximal KE exercise tests and a simulated final experiment adhering to the planned protocol but without any vascular catheterization. Pulmonary minute ventilation, O₂ uptake (O₂), and CO₂ production were calculated by a commercially available software package (Consentius Technologies, Salt Lake, UT) integrated with a Perkin-Elmer MGA 1100 mass spectrometer, a gas-mixing chamber, and a Fleisch pneumotachograph #3 (Hans-Rudolph) (22). In hyperoxic conditions, the technical limitations of measuring ventilatory O₂ exchange at very high inspired O₂ concentrations ([O₂]) precluded collection of pulmonary gas-exchange data (30).

Experimental protocol. Within 1 wk of the preliminary studies, subjects returned in the morning to the laboratory, where two catheters (radial artery and left femoral vein) and a thermocouple (left femoral vein) were emplaced by using a sterile technique as previously reported (16, 22). During exercise, iced saline was infused through the femoral venous catheter at flow rates sufficient to decrease blood temperature at the thermocouple by >1°C. Infusions were continued for 15-20 s until femoral vein temperature had stabilized at its new lower value. Saline injection rate was measured by weight change in a reservoir bag suspended from a force transducer, which was calibrated before and after each experiment. The calculation of blood flow was performed on thermal balance principles as detailed by Andersen and Saltin (3). Blood samples were taken from the arterial and femoral venous catheters to quantify arteriovenous concentration differences.

Determination of intracellular PO₂. Days after completion of the catheter studies, subjects traveled by airplane from San Diego to Philadelphia, where maximal exercise was reproduced in a 2.0-T Oxford imaging magnet to allow the determination of intracellular PO₂ by proton magnetic resonance spectroscopy of Mb (Fig. 1 in Ref. 20). Spectra were collected from the muscle region below the 7-cm-diameter surface coil that was double-tuned to proton (85.45 MHz) and phosphorus frequencies (34.59 MHz) and placed over the rectus femoris portion of the quadriceps group (25), ~20-25 cm proximal to the knee (Fig. 1 in Ref. 20). For these studies, this "sensitive region" was <100 cm³ of muscle, thus isolating signal detection predominantly to the rectus femoris (1). Details of the theory behind O₂-sensitive Mb signals have been published previously (4, 20). Fractional deoxy-Mb (^fdeoxy-Mb) was determined by normalizing signal areas to the average signal obtained during minutes 9 and 10 of cuff ischemia at suprasystolic pressure (270 mmHg) (Fig. 1B). Intramuscular O₂ depletes within 6-8 min of occlusion (29). Therefore, the plateaued signals obtained during the last 2 min of cuff occlusion represent complete deoxygenation of Mb and are used to estimate total Mb concentration within the muscle (Fig. 1B). In a separate method development study, contractions performed during minute 9 of ischemia did not change the deoxy-Mb signal intensity. As illustrated in Fig. 1, no other peaks were visible within 10 parts/million of the deoxy-Mb resonance at any time. Although deoxy-Hb in solution gives rise to signals in the region of the deoxy-Mb peak (14), it has previously been shown that the visibility of deoxy-Hb in vivo is greatly reduced compared with that of deoxy-Mb (28). Here the Hb signals are lost within the baseline noise (Fig. 1, B and C) and thus do not affect the deoxy-Mb measurements.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Representative individual deoxymyoglobin (deoxy-Mb) spectra collected at rest (A), after 10 min of suprasystolic cuff occlusion (B), and during maximal exercise (C). ppm, Parts/million.

Blood analyses. Samples (3-4 ml) of arterial and femoral venous blood were withdrawn from the catheters anaerobically to measure PO₂, PCO₂, pH, O₂ saturation, and Hb concentration. All measurements were made on an IL 1306 blood-gas analyzer and IL 482 CO-oximeter (Instrumentation Laboratories, Lexington, MA). Between each sample, electrodes were calibrated and demonstrated acceptable reproducibility (SD of repeated determinations: PO₂ and PCO₂, 1.5 Torr; pH 0.003). [O₂] was calculated as 1.39 ml O₂ × Hb concentration (in g/100 ml) × measured O₂ saturation (in %) + 0.003 ml O₂/100 ml blood × measured PO₂ (in Torr). Arteriovenous [O₂] difference was calculated from the difference in radial artery and femoral venous [O₂]. This difference was then divided by arterial concentration to give O₂ extraction. Muscle O₂ was calculated as the product of blood flow and arteriovenous O₂ difference, whereas O₂ delivery was calculated as the product of blood flow and arterial [O₂].

Thigh volume measurement. With the use of thigh length, circumference, and skinfold measurements, thigh volume was calculated to allow an estimate of quadriceps femoris muscle mass as suggested by Andersen and Saltin (3) and others (13). It should be recognized that this calculation of muscle mass and the consequent normalizing of blood flow and O₂ assume that this is the only muscle mass involved in the KE exercise, an assumption recently verified in the KE human exercise model (18).

Muscle O₂ transport conductance and mean capillary PO₂ calculations. With the use of the measured intracellular PO₂ values, muscle O₂ transport conductance (D_O2) and mean capillary PO₂ ) were calculated as described previously (26) but only at 100% of WR_max. Briefly, a numerical integration procedure is used to determine the value of D_O2, assumed constant along the capillary, that produces the measured femoral venous PO₂ (Pv_O2), given the measured arterial PO₂ (Pa_O2). As these calculations utilized the measured Mb-associated PO₂ (P_MbO₂), they were no longer burdened by the assumption of a low intracellular PO₂ at WR_max (26). An additional, and at this time unavoidable, assumption of this calculation is that the only explanation of O₂ remaining in the femoral venous blood is diffusion limitation of O₂ efflux from the muscle microcirculation. Perfusion-O₂ heterogeneity and perfusional or diffusional shunt are considered negligible. To the extent that these phenomena do not contribute O₂ to femoral venous blood, the parameter D_O2 is a conductance coefficient that expresses the diffusing capacity that would be required to achieve the measured O_{2 max}, assuming only diffusion limitation. This assumption cannot be avoided presently because of the lack of specific means for detecting perfusion-O₂ heterogeneity and shunt. is the numerical average of all PO₂ values computed, equally spaced in time, along the capillary from the arterial to the venous end. Whereas is useful for graphical purposes (see RESULTS), our conclusions come from statistical comparisons of D_O2 among the three experimental conditions.

Statistical analyses. At maximal exercise, variables were tested for a significant difference among the three different inspired O₂ conditions by repeated-measures ANOVA and a Tukey post hoc analysis. All statistics were computed by using a commercially available software package (Graph Pad, San Diego, CA). Variables were considered significantly different when the P value was 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Intracellular PO₂. Under resting conditions, the proton resonance for Mb lay under the water peak, and the deoxy-Mb was not evident. After inflation of the thigh cuff, proximal to the surface coil, the deoxy-Mb signal increased for the initial 6 min and plateaued for the final 4 min, as previously described (20, 29). On the basis of this plateau in the deoxy-Mb signal amplitude, this signal averaged over minutes 9 and 10 of vascular occlusion was considered to be 100% of the deoxy-Mb signal. Before the exercise protocol was commenced, there was again no discernible deoxy-Mb signal. In normoxic maximal exercise, the deoxy-Mb signal was 55 ± 5% of the maximum deoxy-Mb signal (P_MbO₂ = 3.0 ± 0.7 Torr). During maximum hypoxic exercise, the deoxy-Mb signal achieved 60 ± 4% of the maximum signal (P_MbO₂ = 2.3 ± 0.5 Torr). In hyperoxia, the deoxy-Mb signal reached an average of 48 ± 5% of the maximum deoxy-Mb signal (P_MbO₂ = 4.1 ± 0.7 Torr). Thus in comparison with normoxia, the intracellular PO₂ in hypoxia was significantly lower, whereas in hyperoxia it was significantly elevated. Again, as in previous studies (20), the cessation of exercise in each condition produced a rapid reduction in the deoxy-Mb signal and, therefore, a large increase in P_MbO₂ within 20 s.

Quadriceps O₂ and blood flow, pulmonary O₂, and work rate. Quadriceps O₂ varied with inspired O₂, being lower in hypoxia than normoxia and greater in hyperoxia than normoxia (Table 1). Because of the experimental design, a square-wave protocol, the classic criteria used to establish the attainment of O_{2 max}, a plateau in the O₂-work rate relationship, was not attainable. However, based on data attained in previous KE studies (19), the competitive nature of the subjects, and the reproducibility of WR_max in the preliminary and experimental study days, it was evident that these subjects were at or extremely close to O_{2 max} in each condition. Maximal muscle blood flow was not different between hyperoxia and normoxia but was significantly lower in hypoxia (Table 1). Pulmonary O_{2 max} at maximal KE rose to only 45% of that previously recorded during maximal conventional cycle ergometry, indicating that whole body maximal effort was not invoked. Although hyperoxic pulmonary O_{2 max} could not be measured because of the high fraction of inspired O₂ (FI_O2), there was no difference in pulmonary O_{2 max} observed between hypoxia and normoxia (Table 1). Work rate reflected the muscle O_{2 max} measurements, being significantly greater in hyperoxia and significantly lower in hypoxia than in normoxia [maximum resistance applied to the ergometer = 920 ± 67, 1,160 ± 101, and 1,300 ± 104 g in hypoxia, normoxia, and hyperoxia, respectively (Table 1)].

Blood O₂ content, O₂ delivery, and O₂ extraction. At maximal exercise, both arterial O₂ content and venous O₂ content were significantly reduced in hypoxia and significantly increased in hyperoxia, whereas Pv_O2 was only statistically reduced in hypoxia (Table 1). O₂ delivery reflected the changes in arterial O₂ content and blood flow and consequently was significantly reduced in hypoxia and significantly elevated in hyperoxia in comparison with normoxia (Table 1). O₂ extraction, at maximal exercise, was not significantly influenced by the inspired [O₂].

Hb-O₂ saturation, blood PO₂, and D_O2. At maximal exercise, Sa_O2 was significantly reduced in hypoxia and significantly elevated in hyperoxia (Table 1). As expected, the magnitude of Sa_O2 reduction in hypoxia was much greater than the elevation in hyperoxia because of the 100% ceiling for Hb-O₂ saturation. Venous O₂ saturation was significantly lower in hypoxia and higher in hyperoxia than in normoxia. Pv_O2 in hyperoxia, compared with normoxia, was elevated but did not achieve statistical significance. With this exception noted, Pa_O2, Pv_O2, and were significantly decreased in hypoxia and significantly elevated in hyperoxia in comparison with normoxia (Table 1). Additionally, in the three FI_O2, Pv_O2 and each varied proportionately with O_{2 max}. D_O2 did not vary with the FI_O2 during maximal exercise (Table 1).

Quadriceps femoris weight. Estimated average quadriceps femoris weight was 2.5 ± 0.16 kg. Although all data are presented in absolute terms in this manuscript (Table 1), this measurement allows the calculation of mass-specific blood flows, muscle O₂, and the mitochondrial O₂ estimates illustrated in Fig. 2.

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Comparison of relationship between mitochondrial respiratory rate and O2 availability in vitro made by Wilson and colleagues (32) (A) with present in vivo measurements of relationship between mitochondrial O2 uptake (O2) and intracellular PO2 (B). Mitochondrial O2 was calculated based on assumed mitochondrial fiber volume of 7.5%, myofibril volume of 80%, and muscle density of 1.06 g/cm. Muscle mass was 2.5 kg, calculated from anthropometric measurements. [O2], O2 concentration; P50, O2 half-saturation pressure; FIO2, fraction of inspired O2. C: theoretical combination of A and B to illustrate how Mb-associated PO2 data fit with O2 supply dependence of maximum O2 in intact, normal humans. Linear expression of O2 transport (equation) is a straight line of similar slope (muscle O2 transport conductance; DO2) in hypoxia, normoxia, and hyperoxia, but with lower and higher intercept in hypoxia and hyperoxia due to lower and higher mean capillary PO2 (), respectively, at maximum O2. PmitoO2, mitochondrial PO2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In a progression from the work of the past two decades investigating intracellular PO₂ and its physiological role (5, 6-8, 31, 32), it has recently been documented that, in exercising skeletal muscle, there is a substantial vascular-to-intracellular PO₂ gradient that may be manipulated by altering the FI_O2 (20). The present data support this observation, with the smallest PO₂ gradient of 26.7 Torr in hypoxia and 32.2 Torr in normoxia and the largest gradient of 36.9 Torr in hyperoxia. Here it is interesting to note the small downstream effect of a large increase in Pa_O2 (from 125 to 618 Torr) in the periphery. This may be explained by the rapid fall in Pa_O2 in the first one-tenth of the capillary due to the large PO₂ gradient from blood to cell. As the blood continues to traverse the capillary with a much reduced PO₂, calculated is elevated above normoxia but not as substantially as the original increase in Pa_O2 (a 17% increase in vs. a 400% increase in Pa_O2). Additionally, the significant increase in O_{2 max} associated with both an increase in O₂ delivery and the O₂ gradient from blood to cell supports the theory that O₂ supply plays an important role in determining O_{2 max} in trained skeletal muscle. However, perhaps the most novel observation here is that, unlike the proportional linear increase in O_{2 max} with an increase in P_MbO₂ from hypoxia to normoxia (including a hypothetical point at the origin), hyperoxia increased O_{2 max} relatively less than P_MbO₂, suggesting that at this point the capacity to utilize O₂ (maximal mitochondrial capacity) is starting to play a role in this condition (Fig. 2B).

Intracellular PO₂ as a determinant of O_{2 max}. The present data reveal a hyperbolic relationship between intracellular PO₂ and O_{2 max} (Fig. 2B). This suggests that, in hyperoxia, there is the expected rise in intracellular PO₂ (due to increased ), but this elevated O₂ availability is now in excess of mitochondrial capacity (Fig. 2B), thus suggesting that intracellular PO₂ is a determinant of O_{2 max} in 12, 21, and 100% O₂ but that, in the latter case, the increased intracellular PO₂ results in diminishing returns with respect to an increase in O_{2 max}. These observations are consistent with cellular metabolism that is moving toward a transition between O₂ supply as a determinant of O_{2 max} and O₂ demand as a determinant of O_{2 max} (Fig. 2A). This is illustrated in Fig. 2C, where further increases in intracellular PO₂, beyond those recorded in hyperoxia, have smaller effects on O_{2 max} until a plateau is reached and O_{2 max} becomes invariant with intracellular PO₂. From this point, intracellular PO₂ is no longer a determinant of skeletal muscle O_{2 max}. This hyperbolic relationship, originating from the origin, between O₂ tension and cellular respiration is in agreement with data previously described by Wilson et al. (32) in kidney cells (Fig. 2A). We again (20), although now with more conclusive data, suggest that these findings represent the hyperbolic relationship between in vivo muscle O₂ and intracellular PO₂, supporting the concept that maximal respiratory rate (O_{2 max}) is normally limited by O₂ supply. In fact, we suggest that our data may describe a relationship similar to those in the study by Wilson et al. (32) by reconciling the hyperbolic expression of O₂ utilization in Fig. 2A with the linear expression of O₂ transport (equation in Fig. 2C), as theoretically illustrated in Fig. 2C. Thus this equation, when O₂ is plotted against intracellular PO₂, is a straight line of similar slope (D_O2) in hypoxia, normoxia, and hyperoxia but with a lower and higher intercept in hypoxia and hyperoxia due to the lower and higher , respectively, at O_{2 max} (Table 1). The intersection of these lines with the mitochondrial O₂-PO₂ hyperbolic relationship shows how the present P_MbO₂ data fit with O₂ supply dependence of O_{2 max} in intact, normal humans. The conclusion that D_O2 constrains O_{2 max} is identical, but the data are essentially independent of the data relating O_{2 max} to , which is supported by the present and previously published data (Table 1; Fig. 2; Refs. 12, 19, 23).

Intracellular PO₂ vs. mitochondrial PO₂. The present data once again raise the issues: 1) why at normoxic or hyperoxic O_{2 max} did the P_MbO₂ not fall to the level reached in hypoxia, and 2) why in all three conditions was the Mb desaturation far less at maximum exercise than under conditions of cuff occlusion? A possible explanation may be found by attempting to reconcile our Mb-associated data with the recent measurements of cytochrome aa₃ oxidation-reduction state in exercising skeletal muscle (9). These data illustrated a progressive decrease in the concentration of oxidized cytochrome aa₃ (which correlated highly with rising muscle lactate efflux), with increasing muscle O₂ extraction. At O_{2 max}, the magnitude of this redox response was equivalent to that observed at death or complete anoxia, suggesting that near depletion of O₂ at the mitochondrial level accompanies maximal exercise intensities (9). These findings are in stark contrast to our P_MbO₂ data, which revealed a constant intracellular PO₂ [which did not correlate with rising muscle lactate efflux (21)] with increasing work intensity. However, a reconciliation of these data is possible by approaching them with similar logic employed to explain the observation that there is a large gradient from blood to cell and Pv_O2 (representative of end-capillary PO₂) does not fall to zero even at O_{2 max} (20, 27). That is, a finite D_O2 exists, and this may limit O₂ transport (26). Thus a mitochondrial D_O2, which limits O₂ conductance from the cytosol to mitochondria, may explain both the difference in O₂ availability outside and within the mitochondria as well as the inability for P_MbO₂ to fall to a greater extent before the cessation of high-intensity exercise. In this scenario, a gradient exists from capillary to cytosol (~30 Torr) and from cytosol to mitochondria (~2 Torr). It should be noted that, although the gradient is vastly different in each case, the physiological significance may well be equal.

Summary. The ability to study both intravascular and intracellular O₂ availability in combination with variations in FI_O2 during maximal exercise in trained human skeletal muscle has revealed 1) that there is a very low cytoplasmic PO₂ at O_{2 max}, 2) that variations in systemic O₂ supply alter intracellular PO₂ and these changes are consistent with the concept that O₂ supply limits O_{2 max} in trained human skeletal muscle, and 3) a disproportionate increase in intracellular PO₂ in hyperoxia, suggesting that, in trained skeletal muscle, mitochondrial metabolic limits to muscle O₂ are being approached under these conditions.