INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THERE HAS LONG BEEN INTEREST in measuring the relative flux of metabolic pathways in skeletal muscle as well as other tissues. Specifically, skeletal muscle can utilize a wide variety of substrates. During rest, ~40% of oxygen uptake (O₂) is used for carbohydrate oxidation, and the rest, ~60%, is used for fat oxidation (9). As the metabolic rate is increased, such as during progressively intense contractile exercise, not only is the overall rate of substrate utilization increased but also the relative fraction of carbohydrate utilization increases. During intense exercise, the respiratory quotient (considering only carbohydrates and fats) can approach a value of 1.0, which would be a condition where 100% of the O₂ was being used for carbohydrate oxidation (9). This is usually accompanied by an increase in both muscle and venous blood lactate concentration ([lactate]), leading to the classic interpretation that lactate accumulation is a result of an energy supply-demand mismatch (50). However, on the basis of both old (3) and recent (6, 8, 7, 36) data, it is probably more correct to conclude that lactate accumulation is not simply the result of a supply-demand mismatch but is more closely related to the balance of production and utilization of pyruvate, the cytosolic redox state (24), and the balance of lactate production in the cytosol and lactate utilization in the mitochondria (7). Additionally, the shift from fat to carbohydrate utilization may involve management of citric acid cycle (CAC) intermediate (CACI) concentrations via ancillary reactions of the CAC that include carbohydrates (12, 28).

Since originally proposed (10, 11), the use of ¹H- and ¹³C-nuclear magnetic resonance (NMR) and ¹³C-isotopomer analysis has become a very powerful tool for the study of yeast (46), liver (48), and cardiac (21, 31) metabolism. This combination of methodologies provides information not obtainable with use of more traditional measurement methodologies (6, 8, 23, 36, 44). However, there are many technical difficulties to be overcome before this methodology can be used as robustly in skeletal muscle. As evidence, a recent literature search found only three papers published (2 of which are addressed below) that combined ¹H- and ¹³C-NMR and ¹³C-isotopomer analysis to study skeletal muscle metabolism.

In the first of these papers, Szczepaniak et al. (49) demonstrated that prolonged [2-¹³C]acetate infusion in the intact, anesthetized rabbit can result in very well-enriched and easily interpretable ¹H and ¹³C spectra and observed (among other things) that anaplerosis was much greater in resting skeletal muscle than seen in analogous settings in, for example, the rodent heart. This is significant because anaplerosis, defined as the flux through citrate synthase not due to the flux of acetyl-CoA units directly oxidized in the first two turns of the CAC (28), can be measured most correctly from a well-enriched ¹³C spectrum from a metabolically active tissue that is known to be in metabolic as well as isotopic steady state (32).

To determine whether such measurements could be made from intact rodent skeletal muscle by using ¹H- and ¹³C-NMR methods, in the second of these papers we have recently demonstrated that the competition between utilization of acetate and lactate can be detected and monitored in the hindquarter muscles of the intact, anesthetized rat (5). This demonstration is important because of the long-term potential for this methodology to provide a wealth of information about metabolic regulation not obtainable from other types of methodologies (20, 31, 49). In our study (5), which used the simplest possible combination of oxidizable substrates to minimize the complications of the analysis, our principal finding was that, contrary to the more classic view of lactate, exogenous lactate was readily oxidized by rat muscle, even during contraction.

Because our original implementation of this methodology was in the intact, anesthetized rat, our results may have been complicated by the possibility of label uptake, metabolism, and/or scrambling in tissues other than skeletal muscle. To address this possible limitation, we have modified our procedure to administer the exact same label, in the same concentration of substrates, for the same period of time, but now into the isolated perfused rat hindquarter. With such a preparation and its isolated circulation, we would continuously sample the perfusate, both inflow and outflow, and interpretation of the data would be simplified. Furthermore, if we could demonstrate that such a model could be used for these types of studies, it would be possible in the future to combine this model with pharmacological interventions that would poison one or more metabolic pathways (15, 39) without concern for the possibility that use of these poisons would alter the function of one of the other organ systems or, worse, would kill the animal. Therefore, using the model of the isolated perfused rat hindquarter, we employed ¹H- and ¹³C-magnetic resonance spectroscopy (MRS) and isotopomer analysis to monitor the patterns of ¹³C-labeled lactate and acetate utilization by the isolated rat hindquarter muscles during rest and contraction.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

All procedures were approved by the Institutional Animal Care and Research Advisory Committee. Female Sprague-Dawley rats (267 ± 22 g, Harlan Sprague Dawley, Indianapolis, IN) were housed in the Animal Resources Center of the University of Texas Southwestern Medical Center with a 12:12-h light-dark cycle and access to water and laboratory chow ad libitum.

Hindquarter Surgical Preparation

Muscle contraction. After 30 min of hindquarter perfusion, the muscles of the left limb were stimulated to contract by application of external electrical stimulation to the sciatic nerve. Custom-built silver electrodes connected to an electrical stimulator and isolation unit (Grass models S88 and S7, respectively, New Astromed, West Warwick, RI) were used to deliver 100-ms trains of pulses (100 Hz, 0.2-ms duration, no delay) at twice the motor threshold voltage (typically 1-2 V) at 8 tetani/min for 30 min. The muscles of the right limb were the internal resting contralateral controls. The left calcaneal tendon was attached to a force transducer (Grass model FT-10, New Astromed), and the output was recorded on a strip-chart recorder (model 8K23, Soltec, San Fernando, CA). Input voltage was increased over the 30 min of stimulation to maintain the initial contractile force output.

Perfusate. The perfusate was a modified Krebs-Henseleit buffer of 115 mM sodium chloride, 5.9 mM potassium chloride, 25 mM sodium bicarbonate, 1.2 mM magnesium sulfate, 1.2 mM sodium phosphate, 2.5 mM calcium chloride, 4% bovine serum albumin, 2 mM [1,2-¹³C]acetate, and 5 mM [3-¹³C]lactate and suspended washed bovine erythrocytes, prepared from blood from a local meat-processing plant. Blood was collected in a plastic vessel containing acid-citrated dextrose (23 mM citric acid, 51 mM sodium citrate, and 82 mM glucose) in volumes of 4-5:1, respectively. This solution was washed and spun three to five times in equal volumes of saline and two to four times in equal volumes of a modified version of the above perfusate (5.5 mM glucose, without labeled substrates). Washed erythrocytes were subsequently stored at ~50% hematocrit in the modified perfusate with added penicillin-streptomycin (Sigma Chemical, St. Louis, MO) and used within 7 days. The day of surgery, erythrocytes were washed two to three times in the modified perfusate (without glucose) and brought up to a physiological hematocrit (35.8 ± 0.7) in the original perfusate.

Tissue Preparation

MRS

¹³C-MRS. Proton-decoupled ¹³C-MR spectra were obtained (32-K blocks, spectral width = 20,000 Hz) with a pulse repetition rate of 1 s and a pulse width of 8 µs (45° flip angle). The ¹H decoupling scheme was WALTZ16 with H₂ of ~2,000 Hz. For each gastrocnemius extract, the number of acquisitions (na) = 10,000 for a total collection time of ~7 h and 20 min. For each soleus extract, na = 20,000 for a collection time of ~14 h and 40 min.

¹H MRS. ¹H-MR spectra were obtained (8-K blocks, spectral width ± 6,000 Hz) with a pulse width of 8 µs (90° flip angle). Solvent signal was suppressed with a continuous low-power 2-s presaturation pulse with a bandwidth of 0.1 parts/million (ppm). For each extract, na = 64 for a total collection time of ~10 min.

Spectral analysis. Raw NMR data files were downloaded to magnetic tapes and analyzed on a dedicated data-analysis workstation (SunSPARC station 10 GX, Sun Microsystems, Mountain View, CA) by using data-analysis software (NMR1, Tripos, St. Louis, MO). The free-induction decays were Fourier transformed, and the areas of each relevant resonance were determined by curve fitting by using a Lorentzian function over a distance of ±2.5 line widths.

Assays of Total Metabolite Concentrations

After fluorometric analysis was completed for a muscle extract, each data point for each assay was expressed relative to the wet weight of the muscle and relative to the total creatine content (25). We found that the reference base had little effect on the relative concentrations of any of the metabolites. Therefore, we expressed these values relative to the wet weight of the muscle so that labeled and total metabolite concentrations could be directly compared and fractional enrichment values could be easily calculated.

Statistics

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Physiological Measurements

1

1

1

1

¹³C-MRS Analysis

For many technical reasons, it is much more difficult to make an absolute, rather than a relative, quantitation of the amount of a metabolite producing a spectral peak. However, there are many important types of assessments of any spectral data that require some knowledge of the overall enrichment of a particular spectral peak (32). To aid in making more useful judgments about the relative sizes of the spectral peaks in the spectra, we have chosen to present the relative sizes of the peaks in two different ways: with reference to taurine C-2 (Table 1) and to each other (Table 4). When the sizes of peaks are expressed relative to an endogenous (unenriched) natural abundance peak such as taurine (Table 1), it is more possible to reference the size of any peak to the possible contributions from natural abundance (endogenous, unenriched) or enrichment via incorporation of exogenous (enriched) [3-¹³C]lactate or [1,2-¹³C]acetate. In particular, this will facilitate comparison between different spectra (where the taurine can be expected to be, but where relative label enrichment may not be, relatively constant) or more importantly, between different muscle types. On the other hand, when the sizes of peaks are expressed relative to each other (Table 4), peak area ratios much more reflect the relative incorporation of label with relatively little ability to assess any differential contribution from naturally abundant (endogenous, unenriched) moieties.

Evidence of [1,2-¹³C]acetate incorporation. After examination of the spectra, and calculation of the spectral peak areas (Table 1), it was clear that [¹³C]acetate readily entered both gastrocnemius and soleus muscle tissue and was utilized for metabolism. In gastrocnemius muscle, there were peaks (Table 1) arising from acetate C-2 [three peaks: a singlet at, and a doublet centered around, 24.2 ppm] and acetate C-1 (a doublet centered around 182.3 ppm). In most soleus extracts, although the overall signal-to-noise ratio was lower, doublets were visible both at acetate C-1 and C-2.

Evidence of [3-¹³C]lactate incorporation. In gastrocnemius muscle, incorporation of [3-¹³C]lactate was always clearly evident from the large central C-3 peak (21.0 ppm) and the doublet arising from the [2,3-¹³C]lactate molecule (Table 2). Multiple-peak (multiplet) patterns were also centered about the lactate C-2 (69.4 ppm) and C-1 (183.3 ppm) (Table 4). That this lactate entered directly into metabolism was evident from the singlet in the pyruvate C-3 (26.7 ppm) region. In soleus, although there was also signal visible in all three lactate carbons, there were peaks from [2,3-¹³C]lactate (Table 4). In some extracts, it was also possible to see a small singlet from pyruvate C-3.

Other evidences of incorporation of [1,2-¹³C]acetate or [3-¹³C]lactate. It was also possible to see evidence of incorporation of combinations of [1,2-¹³C]acetate itself, acetate-derived [1,2-¹³C]acetyl-CoA, [3-¹³C]lactate itself, or lactate-derived [2-¹³C]acetyl-CoA in gastrocnemius. For example, a doublet could be detected in the methyl carbon of acetylcarnitine (-CO-¹³CH₃) (centered around 21.5 ppm). Also, there was also a doublet in the C-2 (54.2 ppm) and multiplets in the regions corresponding to the resonances of C-1 (175.5 ppm) and C-4 (30.5 ppm) of acetoacetate. There were singlets visible from the C-1 carbon of glucose (93.1 and 97.0 ppm) and many peaks in regions of other glucose carbons, the precursors of which could be [3-¹³C]pyruvate. Signal was also visible in regions corresponding to the resonances of other carbohydrate metabolites such as glyceraldehyde, glycerol 3-phosphate, fructose, and glycogen. In contrast, in the soleus, there were few regions of such label incorporation. Most prominent of them was the label incorporation visible in the regions corresponding to the resonances of C-1, C-2, and C-4 of acetoacetate.

Oxidative Incorporation

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Sources of 13C labeling in glutamate. Primary sources of acetyl-CoA pool entering citric acid cycle and resultant glutamate labeling pattern are shown. A: labeled () and unlabeled () carbon atoms of acetyl-CoA from primary sources available to this muscle preparation. -, Acetyl-CoA arises from glycogen (intracellular directly from pyruvate or intra- or extracellular via lactate) or intracellular triglycerides (Fc0); -, acetyl-CoA arises from exogenous [3-13C]lactate (Fc2); -, acetyl CoA arises from exogenous [1,2-13C]acetate (Fc3). n, No. of data points; OAA, oxaloacetate; -KG, -ketoglutarate. B: these different acetyl-CoA isotopomers produce 1 of 5 different labeling patterns in C-4 of glutamate. Unlabeled acetyl-CoA produces a singlet solely from its 1.1% natural abundance. [3-13C]-derived [2-13]acetyl-CoA produces a singlet or doublet after incorporation into citrate and label mixing at succinate or fumarate, and then condensation with another [2-13]acetyl-CoA, [1,2-13C]acetate-derived [1,2-13C]acetyl-CoA, produces a doublet. Doublet of doublets, commonly referred to as a quartet, is produced after incorporation of either substrate into citrate, label mixing at succinate or fumarate, and then condensation with a(nother) [1,2-13C]acetate. Sum is a 9-line multiplet. Because of small degree of carbon-carbon coupling, across 2 carbons, labeling pattern in C-1 and C-2 of glutamate is not detectable by this method, and these carbons are depicted by crosshatched circles.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   13C spectra from rested and contracted gastrocnemius muscle extracts. A: glutamate C-4 region from contracted muscle. B: glutamate C-3 region from contracted muscle. C: glutamate C-4 region from rested muscle. D: glutamate C-3 region from rested muscle. 13C resonances are visable from succinate (Succ) C-23 and glutamate C-4 and C-3. In glutamate C-4 region, singlet (S), doublet 34 (D34), doublet 45 (D45), and quartet (Q) are labeled. In glutamate C-3 region, singlet plus triplet (S+T), triplet wings (T), and doublet (D) are labeled. ppm, Parts/million.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   13C spectra from rested and contracted soleus muscle extracts. A: glutamate C-4 region from contracted muscle. B: glutamate C-3 region from contracted muscle. C: glutamate C-4 region from rested muscle. D: glutamate C-3 region from rested muscle. 13C resonances are visible from succinate C-23 and glutamate C-4 and C-3. In glutamate C-4 region, S, D45, D34, and Q are labeled. In glutamate C-3 region, S+T, T, and D are visible.

Label in CACI. Other CAC metabolites were also enriched (Table 1). Multiplets (predominantly 3 peaks) were observed for malate C-2 (71.2 ppm) and C-3 (43.4 ppm) as well as citrate C-2-C-4 (44.4 and 45.3 ppm) and C-1-C-5 (180.0 ppm). Peaks in areas of the malate C-4 (181.6 ppm) were visible in resting gastrocnemius and contracting soleus muscle extracts. Signal in the regions where fumarate C-2-C-3 (136.2 ppm) and succinate C-1-C-4 (183.1 ppm) resonate was visible in gastrocnemius but not soleus muscle. Distinct multiplets arose from incorporation of label into the C-2-C-3 of succinate (35.0 ppm) in gastrocnemius muscle (Fig. 1) and to a lesser degree in soleus muscle (Fig. 2). Muscle contraction increased signal from succinate C-1-C-4 (Table 1) and the ratio of succinate C-2-C-3 to glutamate C-4 in gastrocnemius muscle (Table 4).

¹H-MRS

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Sources of 1H labeling in methyl carbon of lactate. Sources of 1H-nuclear magnetic resonance spectral peaks arising from a mixture of 12C and13C lactate are shown. Doublet at 1.32 ppm arises from summed resonance of 3 equivalent protons of unlabeled () methyl carbon split by J coupling to single proton on adjacent carbon. Pair of doublets arise when this same doublet is itself split by J coupling to its labeled () carbon. Ratio of area under center doublet to that of summed area under doublets on either side is equal to ratio of [3-12C]lactate to [3-13C]lactate in this extract.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   1H spectra from perfusate. 1H magnetic resonance spectra from methyl carbons of labeled and unlabeled acetate (ac) and lactate (lac) are shown for perfusate samples taken (from bottom to top) before perfusion began, perfusion during rest, and perfusion during contraction.

Although [2-¹³C]acetate was visible in spectra from the perfusate (Fig. 5), it was more difficult to detect in the muscle extracts themselves (Fig. 6). The acetate peaks were larger in contracting soleus than in gastrocnemius muscle, but, in both muscles, the ratio of labeled to total acetate was not different after contraction (Fig. 6). In contrast, peaks corresponding to [3-¹²C] and [3-¹³C]lactate were clearly visible in all muscle extracts and were generally much larger than the analogous acetate peaks (Fig. 6). During contraction, the ratio of labeled to total lactate was visibly decreased in gastrocnemius but not in soleus muscle (Fig. 6).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   1H spectra from rested and contracted gastrocnemius muscle. 1H magnetic resonance spectra are displayed from rested (bottom) and contracted (top) gastrocnemius (A) and soleus (B) muscle extracts. 1H resonances from methyl carbons of labeled and unlabeled ac, alanine (ala), and lac are shown.

Fluorometric Metabolite Assays

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We have previously demonstrated this combination of NMR methodologies to study rat skeletal muscle substrate utilization (5); however, these experiments involved systemic administration of substrate via the jugular vein into an intact anesthetized rat. Although clear and analyzable results were obtained, we could not discount the possibility that the pattern of substrate availability may had been influenced by substrate metabolism in any of several internal organs, particularly the heart and liver. Therefore, our goal in the present study was to modify the methodology to isolate the hindquarter muscles and then reexamine the pattern of substrate utilization in an analogous manner. We believe these data represent the first use of this combination of NMR methodologies to study the effects of metabolic stress in isolated perfused rat hindquarter muscles. We found that 1) our application of the isolated, perfused rat hindquarter preparation produces a similar pattern of utilization with the same exogenous substrates as in an intact animal with systemic venous infusion (5); 2) [3-¹³C]lactate is oxidatively incorporated in contracting, as well as resting, muscle; 3) entry and exit from the CAC via nonoxidative pathways are components of normal muscle metabolism; and 4) nonoxidative CAC entry of carbohydrate-derived pyruvate during contraction may be required for normal muscle function.

The preparation was very stable. We used a hindquarter flow rate adequate to elicit its highest possible O₂. In preliminary experiments, we found that the hindquarter muscle remained viable for over 2 h of perfusion, even with perfusion in which buffer without erythrocytes was used (unpublished observations). Because the bioenergetic state of the muscle would be expected to be reflected by PCr concentration ([PCr]) (34), we used this as an index of tissue viability. [PCr] found in resting muscle of the intact rat is ~16 (43) to 23 µmol/g (35) for gastrocnemius and between 11 (43) and 13 µmol/g (35) for soleus. We found that [PCr] in extracts of gastrocnemius was 18.65 µmol/g in resting and 13.75 µmol/g in contracting muscle; in extracts of soleus it was 11.03 µmol/g in resting and 10.02 µmol/g in contracting muscle (Table 5). Inasmuch as it would be expected that [PCr] would decline during the 1) 10-60 s between the end of perfusion and the freeze clamping, 2) inevitable hydrolysis of creatine phosphate during the extraction process, and 3) time the extract was at room temperature in the magnet, we believe such high [PCr] is abundant evidence of a well-oxygenated, stable intact muscle preparation and that the lower gastrocnemius [PCr] in contraction is evidence of the metabolic load from the contractile activity.

We imposed muscle contraction to increase the metabolic demand of those muscles. Although we employed a low-intensity-stimulation protocol (14, 17), it was still severe enough that the muscle fatigued to end force outputs of 30-50% of the initial force output and [PCr] fell ~35%. Our inability to detect an increase in hindquarter O₂ during contraction was likely due to the combination of the 1) small mass of contracting muscle (14) and 2) constant flow rate during rest and contraction (hindquarter flow was kept constant and in our hands represented a maximally dilated preparation; data not shown). Thus our result was not surprising because, typically, increased O₂ during muscle contraction in the isolated perfused rat hindquarter preparation is brought about by increased pump flow rate (which we kept constant) over that of rest and much less that of contractile rate (16).

As in the whole-animal infusion preparation (5), labeled acetate was readily taken up by this muscle preparation. We chose to provide 2 mM acetate because this was midway between the normally quite low (0.15-0.20 mM) concentration of acetate found in the arterial blood of rats (38) and the 3.5 mM vascular acetate used by Szczepaniak et al. (49), about as high a concentration as is normally found in humans after renal dialysis against acetate-containing buffer (45). The normally occurring 0.2 mM acetate would not be high enough to result in enrichment of the acetate-derived acetyl-CoA pool adequate to produce analyzable ¹³C spectra. On the other hand, although the 3.5 mM concentrations would provide the greatest likelihood of acetyl-CoA pool enrichment, such a concentration might require addressing questions of methodological artifact. Thus, at a 2 mM concentration, we could expect that incorporation of [1,2-¹³C]acetate would be adequate to allow for accurate spectral analysis with minimal complication from an excessive concentration load.

In the gastrocnemius spectra, labeled acetate was easily visible as three peaks at C-2 and a doublet at C-1. The three peaks in C-2 arose from the sum of two separate pools of [2-¹³C]acetate: a pool of [2-¹³C]acetate likely derived from [3-¹³C]lactate via acetyl-CoA hydrolase (27) in combination with the pool of [1,2-¹³C]acetate. The latter would be the sole source of the doublet in C-1 and account for the observed acetate doublets in soleus muscle spectra. Labeled carbons from acetate also enriched the acetylcarnitine and acetoacetate pools. There was a doublet in the methyl resonance from acetylcarnitine (-CO-¹³CH₃) and multiplets in the regions where the C-1 and C-4 of acetoacetate resonate. These could only arise from incorporation of the labeled acetate.

Both muscles also readily incorporated lactate. We chose to provide 5 mM lactate because this was about midway between the 1.0 mM lactate concentration normally found in the arterial blood of resting rats (38) and the blood lactate concentrations normally found after exhaustive exercise in humans (15-20 mM) (9) or in the Thoroughbred horse (up to 30 mM) (40). Thus we chose a [lactate] that was about what could be expected to be found in the muscles of these animals after in situ contractions. At this concentration, we could expect that the incorporation of [3-¹³C]lactate would be adequate to allow for accurate spectral analysis with minimal complication from excessive concentration load.

There was a large singlet at lactate C-3, a smaller peak at C-2, and one still smaller at C-1 (Table 1). The multiplet patterns on either side of the C-3, C-2, and C-1 singlets in gastrocnemius muscle (Table 2) meant that, from the original [3-¹³C]lactate, label scrambling had produced the following: [3-¹³C]lactate, [2,3-¹³C]lactate, [2-¹³C]lactate, [1,2-¹³C]lactate, and [1-¹³C]lactate. The small side peaks on either side of the C-3 singlet in soleus muscle (Table 2) meant that, from the original [3-¹³C]lactate, label scrambling had produced the following: [3-¹³C]lactate, [2,3-¹³C]lactate, [2-¹³C]lactate, [1-¹³C]lactate but no visible amount of [1,2-¹³C]lactate. These lactate isotopomers were produced by some combination of labeled acetyl-CoA entering the CAC at citrate with subsequent mixing through CAC turnover and/or label scrambling at succinate and/or fumarate and then exiting at malate and/or oxaloacetate to generate labeled pyruvate, which produced all the observed combinations of label in lactate. Alternatively, one cycle through gluconeogenesis as far as phosphoenolpyruvate and then back through the glycolytic pathway to pyruvate could produce some of this pattern of label in lactate as well, and this type of pathway utilization is supported by label incorporation in alanine C-2 in all muscles (Table 1). Although it might appear surprising that this pyruvate cycling would occur in contracting muscle, normally considered bioenergetically stressed, this is consistent with a continual and concurrent entry of substrate into glycogen synthesis, which has been demonstrated in both resting and exercising rat muscle (18). Our observation of label in the area where glucose C-1 resonates as well as signal in several other glycolytic intermediate regions is also consistent with this observation, although signal from natural abundance of these glycolytic intermediates cannot be ruled out.

Based on the direct-analysis method of isotopomer calculations (32), which depends on the relative ratio of the peaks in the glutamate C-3 and C-4 multiplets (Table 2), the fraction of acetyl-CoA entering the CAC from [1,2-¹³C]acetate in resting gastrocnemius muscle was 56% (F_c3), whereas 24% came from [3-¹³C]lactate (F_c2), and the remainder, 19%, came from endogenous sources (F_c0) (Table 3). In soleus muscle, these relative fractional enrichments were 73, 16, and 11%, respectively (Table 3). During contraction, these fractions were not different. It is interesting to note that, even during contraction, where one might have expected an increase in the rate of production of endogenous (unenriched) lactate to lower the rate of incorporation of exogenous lactate, this calculated F_c2 did not decrease.

With this experimental design, the calculated F_c2 depends mostly on the relative area of the singlet of the glutamate C-4 region (Table 2), the size of which is the summed contribution of unlabeled (therefore, in a 98.9:1.1 ratio of 2-¹²C to 2-¹³C) acetyl-CoA (from all endogenous sources, at a 1.1% natural abundance in ¹³C) and [2-¹³C]acetyl-CoA, which in this experiment can only come from exogenous [3-¹³C]lactate. During contraction, it is expected that the rate of production of endogenous (i.e., unenriched, or 98.9% ¹²C) lactate in the cell would increase. This would increase the overall ratio of ¹²C to ¹³C lactate in the muscle (a sum of the intracellular, interstitial, and vascular volumes). If this occurred, the rate of oxidation of lactate remained constant, and the proportion of intracellular vs. extracellular lactate oxidation remained constant, the rate of entry of unlabeled lactate into the citric acid cycle would increase in proportion. ¹H spectra (Fig. 5) from perfusates indicate that the relative fraction of unlabeled (the doublet labeled ¹²C lactate) to labeled (the adjacent doublets labeled ¹³C lactate) actually increased during contraction. This means that the perfusate was less enriched in[ ¹³C]lactate, thereby diluting the lactate-derived acetyl-CoA pool. In contrast, ¹H spectra (Figs. 4 and 6) from muscle extracts indicate that there was no difference in the ratio of unlabeled to labeled lactate. Because the muscle lactate is the sum of endogenous (98.9% unenriched) and exogenous (from Fig. 5, ~50% unlabeled) lactate, for F_c2 to remain constant, the fraction of exogenous lactate oxidized must have increased during contraction.

Furthermore, it is important to remember that relative oxidative fractional utilization we calculated may actually underestimate the total fraction of acetyl-CoA that entered the CAC oxidatively from lactate because we cannot determine from these data alone how much of the endogenous unenriched acetyl-CoA comes from endogenous unenriched (or, for that matter, exogenous) lactate. Moreover, the singlet in the glutamate C-4 region, arising from first-time incorporation of [3-¹³C]lactate, visibly increased during contraction in both gastrocnemius and soleus muscle (Figs. 2 and 3). Such an oxidation of exogenous lactate, even during this period of metabolic stress, is inconsistent with the concept of lactate accumulation as an index of anaerobic threshold (50). In contrast, it provides strong support for the view of lactate not as a metabolic end or waste product (13) but as a mobile glycogen (treating glycogen-derived lactate oxidized by an adjacent cell or in another part of the same cell as having the same function and energetic cost as one-half of a glycogen-derived glucosyl unit), which is in a readily oxidizable form (6), perhaps because lactate can be found in a subcellular compartment or in a form in which it can be preferentially oxidized (47). This has previously been observed in the heart (47) where there was evidence that vascular lactate was preferentially transferred to mitochondria. We found a similar pattern in gastrocnemius extracts (Table 6), where the ¹³C enrichments of alanine C-3 and acetyl-CoA C-2 were not different but were different from that in lactate C-3. Such a pattern is consistent not only with an extracellular (6) but also with an intracellular lactate shuttle mechanism (7).

One of our goals was to demonstrate that it would be possible to provide ¹³C-labeled substrates to intact rat muscle and detect the effect of an altered metabolism on their relative incorporation. Ideally, any such ¹³C experiment would result in enough of a ¹³C enrichment of the acetyl-CoA pool that an accurate isotopomer multiplet analysis could be done. The accuracy of an isotopomer multiplet analysis is increased as a function of the fractional enrichment of the acetyl-CoA pool entering the CAC (32). Such enrichment will increase with increasing concentration of ¹³C-labeled acetyl-CoA precursors and infusion or perfusion time. By comparing the area of the taurine C-2 peak (as an internal reference) with the glutamate C-4 multiplets, we can estimate the degree to which the glutamate pool is enriched in ¹³C. With use of this sort of estimation, the total amount of glutamate enriched in ¹³C at the singlet position was ~5%, which was no different from what we found previously with a jugular infusion (over the same 30-min period) of the same mixture and concentration (5 mM [3-¹³C]lactate and 2 mM [1,2-¹³C]acetate) of substrates (5). Thus this isolated hindquarter preparation took up exogenous lactate no differently than muscle in the intact, anesthetized rat (5).

Careful examination of the in vivo ¹³C spectra from the only other comparable experiment, 3.5 mM vascular [2-¹³]acetate in the rabbit (49), using the same method of estimation (comparison with the size of the endogenous taurine C-2 peak), at the 30-min point (half time of enrichment of 17 min), the glutamate singlet was ~13% enriched in ¹³C. At this point, it is impossible to determine how much of the difference between 5% (our data and Ref. 5) and 13% (in the rabbit) (49) may be due to any intrinsic differences between the rabbit and the rat, differences in the endogenous glutamate concentrations, the use of [2-¹³]acetate alone vs. [3-¹³C]lactate + [1,2-¹³C]acetate in combination, or the use of acetate at 3.5 vs. 2 mM concentration. Nonetheless, we have clearly demonstrated that this methodology can be used to monitor substrate utilization in the isolated perfused rat hindquarter.

Anaplerotic entry of carbohydrates to maintain CACIs has been speculated to be required to support normal muscle oxidative capacity during exercise (12). Although we could not calculate anaplerosis from our data, as done in the rabbit by using a 2-h infusion of 3.5 mM [2-¹³C]acetate (49), evidence supporting nonoxidative entry of labeled carbohydrate could be inferred from the labeling pattern in contracting gastrocnemius muscle. The ratio of label incorporation in the succinate C-2-C-3 region vs. that of glutamate C-4 increased in gastrocnemius muscle during contraction (Table 4). Additionally, the pattern of succinate label incorporation during contraction does not reflect the doublet 45 observed in the glutamate C-4 region but instead demonstrates an increased singlet (Fig. 2). In heart, the succinate isotopomer composition has been demonstrated to reflect that of both glutamate and -ketoglutarate (22). Our results in gastrocnemius muscle indicate a difference in labeled substrate incorporation in -ketoglutarate (in exchange with glutamate) vs. succinate. Such a difference in labeling pattern between succinate and the glutamate--ketoglutarate isotopomers, particularly considering the pattern of succinate label incorporation, is consistent with nonoxidative labeled lactate entry into the CAC via pyruvate, most likely through pyruvate carboxylation.

In contrast to contracting gastrocnemius muscle, limited succinate label incorporation was visible in contracting soleus muscle (Fig. 2, Table 1), although the total concentration of succinate was about twofold higher (Table 5). Although the reason for this is unclear, it was not due to an inability to detect labeled succinate. ¹³C incorporation into carbons of other metabolites (e.g., glutamate) was readily observed in soleus muscle; thus, if succinate had been similarly enriched, peaks would have been as visible in the soleus muscle spectra. This observation suggests that there was a difference between gastrocnemius and soleus muscle in how the labeled substrates were metabolized once they entered the cells. However, there was other evidence of nonoxidative entry of endogenous (unenriched) carbohydrate into contracting soleus muscle. For example, there was an increase in total succinate concentration in contracting soleus muscle (Table 5). This was not accompanied by increases in labeled succinate pools; thus this increase must have arisen from unlabeled sources. Pyruvate carboxylase activity has been shown to increase with muscle contraction (29); therefore, it is possible that endogenous carbohydrate-derived pyruvate was responsible for the increased succinate concentration during contraction, although contribution via fumarate from the purine nucleotide cycle cannot be ruled out. Further support for nonoxidative entry of endogenous carbohydrate during contraction in soleus muscle was obtained from the combined results of 1) no change in oxidative fractional enrichment values from rest (Table 3) and 2) decreased fractional enrichment of the CAC pool [as estimated by the sum of the labeled concentrations of malate C-3, succinate C-2-C3 and citrate C-2 (Table 1)/the sum of the total concentrations of the same metabolites (Table 5)] from 0.05 ± 0.02 during rest to 0.03 ± 0.01 during contraction.

In summary, we have demonstrated the application of ¹H- and ¹³C-MRS and ¹³C-isotopomer analysis of the utilization of [3-¹³C]lactate and [1,2-¹³C]acetate by resting and contracting muscles of the isolated perfused rat hindquarter preparation. This combination of methodologies avoids some of the methodological limitations of radioactive tracers (6, 20) and therefore represents a powerful alternative means by which to monitor the regulation of substrate utilization by intact mammalian skeletal muscle during normal and stressed metabolic conditions. Our results indicate that 1) our application of the isolated perfused rat hindquarter preparation produces a similar pattern of utilization with the same exogenous substrates as in an intact animal with systemic venous infusion (5); 2) [3-¹³C]lactate is oxidatively incorporated in contracting, as well as resting muscle; 3) entry and exit from the CAC via nonoxidative pathways are components of normal muscle metabolism; and 4) nonoxidative CAC entry of carbohydrate-derived pyruvate during contraction may be required for normal muscle function.

Perspectives

It has been suggested that carbohydrate depletion during prolonged exercise may impair aerobic energy production by reducing the levels of CACIs (42). If O₂ and oxidative phosphorylation were limited not only by oxygen itself but also by CAC pool size, then this would indicate a requirement for maintenance of CACI concentrations in working muscle. Results from these experiments are consistent with the concept that nonoxidative entry of carbohydrate-derived pyruvate is required for normal muscle function. Thus our data provide a direct mechanistic explanation for the observed carbohydrate shift (9) during incremental exercise and explain the necessity for carbohydrates in fulfilling a metabolic requirement that is not possible via free fatty acid metabolism alone.