INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE BIOPSY TECHNIQUE has been used for many years to measure muscle metabolites, including glycogen, and it is vital that the methods employed give an accurate representation of the muscle glycogen stores. Glycogen is found in granules, and thus there is the potential for a high degree of variability among biopsy samples. An electron-microscopic study of glycogen granules by Friden et al. (5) demonstrated that there are two different particle-sized populations in human skeletal muscle. These two separate populations of glycogen particles were stored in five different locations within the muscle cell, and the authors speculated that these populations could be two forms of the glycogen molecule that were utilized at different times from distinct locations. Despite these potential variables, Essen and Henriksson (4), Harris et al. (6), and Hultman (9) have shown that, in human muscle, the procedure is very reliable and that muscle biopsy samples are representative of the whole muscle.

Traditionally, there have been two methods for measuring glycogen, either by acid hydrolysis (AC) (15) or by enzymatic hydrolysis (EZ) (3). Although it has been stated that the variability of the analytic technique is about ±10-13 mmol glucosyl units/kg dry weight (dw) (6), no detailed comparison of AC vs. EZ has been published determining whether the methods are comparable. Bergmeyer (3) has suggested that EZ analysis of purified glycogen gives values that are 5% higher than those obtained by using AC, whereas Passonneau and Lauderdale (15) claimed that the two methods give equivalent results. None of above-mentioned authors provided data to substantiate his claim. In the only direct comparison of AC and EZ that we are aware of, Jansson (10) performed the AC and EZ methods in human muscle over a wide concentration range and stated that there was no systematic difference between the values obtained by using either method. However, no data were provided to validate this finding.

Before Jansson's study (10) it was common for glycogen measurements to be made in the precipitate that had been pretreated with perchloric acid (PCA) (11, 15, 16). This method was thought to make the most efficient use of a muscle sample by allowing for the precipitation of proteins and extraction of PCA-soluble metabolites while also permitting the determination of glycogen in the same piece of tissue. However, it has been reported that only a portion of the total glycogen was extracted during PCA treatment (7, 9, 11, 17). Jansson (10) compared this method to AC and EZ and documented that ~15-25% of the glycogen was PCA soluble and hence any measurement of glycogen only from the precipitate would underestimate the total. Jansson's findings discouraged the analysis of PCA precipitate to determine total glycogen but did not stimulate further research regarding the possibility that these two fractions represented two physiological pools.

Recently, investigators (1, 12, 14), while identifying and studying glycogenin, the protein core of the glycogen molecule, realized that there were two pools of glycogen in rodent skeletal muscle and other tissues such as liver and heart. This work used only rodent resting muscle and had not investigated exercise metabolism or human tissue.

The above-mentioned studies determined that one of the fractions was of smaller molecular weight (400,000) and was relatively rich in protein, prompting use of the name proglycogen (PG). The other type, a larger molecule, was what is recognized as "classic" glycogen or macroglycogen (MG; mol wt 10,000,000). The two forms differ in the ratio of protein to carbohydrate. PG was found to precipitate in trichloracetic acid because of its 10% protein component (12). MG, with a protein content of only 0.35%, was soluble.

No one has systematically compared the two traditional methods, nor have these methods ever been compared with the MG/PG determination for glycogen. Furthermore, PG and MG concentrations have not been examined in human muscle under any circumstances. However, before studies into the physiological response and regulation of these pools can be undertaken, a thorough evaluation of these methods is required. The purposes of this study were 1) to compare AC and EZ determinations for total glycogen; 2) to establish whether MG and PG exist in human muscle and whether they can be measured in a "biopsy-sized" sample; 3) to evaluate the reproducibility of the AC, EZ, PG, and MG determinations; and, finally, 4) to compare the AC, EZ, and MG+PG values for total muscle glycogen.

	METHODS

Top Abstract Introduction Methods Results Discussion References

To evaluate the precision and variability and compare the various methods (AC, EZ, and MG+PG separation) of measuring muscle glycogen, a variety of approaches was employed. The muscle samples had been collected during a number of different studies and conditions ranging from exhaustive exercise to resting muscle. Every sample in each comparison weighed 1-3 mg dw (i.e., a biopsy-sized sample) and was measured in duplicate.

Comparison I. The first comparison used rodent muscle because most of the initial MG/PG data are derived from this species and because large muscle samples could be obtained. Muscle samples of various known glycogen concentrations from rodent hindlimb were freeze-dried, combined, and mixed thoroughly to create three large pools for repeated measures. A total of 45 samples was analyzed. Independent samples (n = 15) from each pool were measured for glycogen by using AC (n = 5), EZ (n = 5), or MG+PG (n = 5).

Comparison II. Human muscle has a much wider range of muscle glycogen concentration than does rodent muscle. Hence human muscle samples, which from previous determination were known to have a wide range of total glycogen concentration, were combined and mixed thoroughly to give three pools of distinctly different glycogen concentrations. A total of 18 samples was analyzed for this comparison. Independent samples (n = 6) from each pool were measured for glycogen by using AC (n = 3) or MG+PG (n = 3).

Comparison III. Individual human muscle samples were analyzed by using each of the three methods. These were studied to compare the AC, EZ, and MG/PG determinations when total glycogen concentration is measured in small biopsy samples. Because of limited sized, not all samples could be analyzed by using all three methods. MG/PG was performed on every sample (n = 50), EZ on 43 samples, AC on 20 samples, and, of the latter, 15 were analyzed by EZ as well.

Comparison IV. Human muscle samples (n = 28) were obtained from an independent laboratory, in which they had previously been analyzed for total glycogen by EZ. These were subsequently measured for MG and PG by a technician blind to the existing data for EZ and then were compared with the prior independent determinations.

Comparison V. Rat muscle was used to determine the percent recovery of a given amount of exogenous glycogen by using MG+PG separation. A single pool of muscle was used, and a portion was analyzed for MG+PG. A glycogen solution was made from oyster glycogen (Sigma G-8571), and a portion of this solution was analyzed for glucosyl units. Muscle samples had a known quantity of this exogenous oyster glycogen solution added during the PCA-extraction procedure (n = 9). After the samples were assayed for glucosyl units, the previously determined MG+PG concentration was subtracted from the total concentration determined for the exogenous plus muscle sample. To find the percent recovery, this value was divided by the exogenous glycogen concentration and then multiplied by 100.

Analysis. The AC glycogen method used was adopted from Passonneau and Lauderdale (15) and is described as follows. Freeze-dried muscle samples of between 2 and 3 mg were hydrolyzed with 2 M HCl and then were heated for 2 h at 85-90°C, followed by neutralization with 2 M NaOH. The extracts were then analyzed in duplicate fluorometrically by using the Bergmeyer (3) method for determining glucosyl units.

Statistics. The total glycogen for the MG/PG determination was obtained by summing the two fractions. The fraction of MG to PG was determined by dividing the concentration of the MG by the summed total (MG/MG+PG) and reporting it as a percent (×100%). The coefficient of variation (CV) was obtained by the formula SD/mean. Linear regression analysis was performed within comparison sections, and 95% confidence intervals were used to determine agreement among analysis methods. T-tests were used to determine difference between means of total glycogen for each of the three methods. In all comparisons, differences were accepted as significant at the 0.05 probability level.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Comparison I. The important findings of this section were that the three measurements were reproducible in rat muscle and that MG+PG was reliable for biopsy-sized samples. Generally, CVs were within the published range, although the CV increased with lower total glycogen concentration (Table 1). However, SE values were small, and the higher CV was a reflection of the low glycogen concentration. One of the "normal" glycogen pools had a high CV for AC because the value for one of the five samples was unusually high.

Comparison II. In the pooled human muscle, the repeatability was similar to that in comparison I (Table 2). Over a wide range of total glycogen concentrations, the SD and CV of the MG/PG determination in human muscle were within the range reported by Harris et al. (6). The MG/PG determination demonstrated a strong correlation with the AC method (r = 1.0), and the line of identity with a slope equal to 1 fell within the 95% confidence interval, illustrating that there was no difference between the two methods. (Fig. 1) There was no significant difference in the total glycogen values determined by AC and by MG+PG.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   A comparison of total muscle glycogen in 3 pools of human muscle by using acid hydrolysis (AC) and macroglycogen (MG)+proglycogen (PG) methods (comparison II) Each point is a single sample from 1 of the 3 pools of muscle. dw, Dry wt. Solid line, linear regression analysis: y = 1.01x + 8.38, r = 1.0, n = 9; dashed line, line of identity with slope = 1; dotted lines, 95% confidence interval.

Comparison III. The third comparison of muscle glycogen for individual biopsy samples of human muscle presents evidence that the MG/PG determination is comparable to the AC and EZ methods (Figs. 2-4). The correlation coefficients as well as the line of identity and regression line not being significantly different demonstrate how similar the three methods were in giving a precise measurement of total glycogen. The correlation coefficients for EZ vs. MG+PG (n = 43), AC vs. MG+PG (n = 20), and EZ vs. AC (n = 15) are 0.96, 0.99, and 0.97, respectively. The ratio of MG to PG varied greatly in these individual samples, with MG representing from 6 to 38% depending on total glycogen concentration. The percentage of MG relative to PG is not constant. With increases in total glycogen (MG+PG), the percentage of MG increases as well. The relationship is curvilinear (Fig. 5). There appears to be a cluster of MG concentrations ranging from 4 to ~50 mmol glucosyl units/kg dw when the total glycogen concentration is <300. At higher glycogen concentrations, the PG concentration does not increase a great deal, but the MG fraction becomes a greater contributor, although never the equal of the PG fraction.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   A comparison of total muscle glycogen in human muscle by using AC and MG+PG methods (comparison III). Points and lines are defined as in Fig. 1. Linear regression analysis: y = 0.92x + 3.8, r = 0.99, n = 20.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   A comparison of total muscle glycogen in human muscle by using enzymatic (EZ) and MG+PG methods (comparison III). Points and lines are defined as in Fig. 1. Linear regression analysis: y = 0.84x + 30.0, r = 0.96, n = 43.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   A comparison of total muscle glycogen in human muscle by using EZ and AC methods (comparison III). Points and lines are defined as in Fig. 1. Linear regression analysis: y = 0.87x + 14.6, r = 0.97, n = 15.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   A comparison of MG values and MG+PG total values obtained in n = 50 individual muscle samples.

Comparison IV. In this comparison human biopsy samples were analyzed for MG/PG and then compared with EZ determinations made by an independent laboratory. Again, the MG+PG total glycogen correlated well (r = 0.97) with the values previously measured by using the EZ method (Fig. 6). Again, there was no significant difference between the line of identity and regression line.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   A comparison of total glycogen in human muscle by using EZ values previously obtained by an independent laboratory and MG+PG methods (comparison IV). Points and lines are defined as in Fig. 1. Linear regression analysis: y = 0.93x + 19.2, r = 0.97, n = 26.

Comparison V. This part of the study was designed to test how much of a given amount of glycogen is actually recovered by using the MG and PG extraction procedure. A 10-µl sample of a 43 µM solution of exogenous glycogen was added to a measured amount of rat muscle, and the glycogen concentration was compared with the concentration of an independent aliquot from the same pool. The percent recovery averaged 105 ± 8.5% (n = 9), and the exogenous glycogen was entirely recovered in the MG portion.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This study was designed to compare the two well-accepted methods of measuring the total glycogen in a muscle biopsy sample over a wide range of total glycogen concentrations. It was also conducted to determine whether MG and PG can be measured in biopsy-sized samples and whether this method can be applied to human tissue. The precision of this new technique was also evaluated and compared with that of the two accepted methods. This study presents the first determinations of MG+PG for human muscle and strongly suggests that there are two forms, and probably two metabolic pools, of glycogen in human muscle.

In rat muscle (comparison I), there was no statistical significance in total glycogen between the AC and EZ methods. In addition, the AC and EZ methods were found to be equivalent in reproducibility (Table 1). Similarly, in human muscle we found that there was no systematic difference in the two measurements because the line of identity with a slope of 1 was within the limits of the 95% confidence interval (Fig. 1). This is in agreement with a comment made by Jansson (10). Harris and Hultman (6) determined that in the measurement of muscle glycogen, there was a variation of 8.43 mmol glucosyl units/kg dw because of analytic procedure error alone. The SD for routine analysis of muscle glycogen samples taken during exercise experiments was found to be between 10 and 13 mmol glucosyl units/kg dw (6). Our data for rodent and human muscle glycogen are in agreement with these findings. Therefore, each of the three methods gives low and similar CV values that are comparable with findings in the literature.

The reproducibility in the pooled human muscle samples is in agreement with other studies, and the variances (Table 2) in the present study were within the published range (6). The repeated measurement of MG, PG, and MG+PG in human muscle (n = 9) had a CV range from 0.2 to 21%. The majority of the CV values displayed in Table 2 lie between 3 and 6%. Essen and Henriksson (4) found that there was a CV of ~6.3% when they measured glycogen within single muscle fibers. In a comparison of the human muscle total glycogen for the MG/PG determination for the low, medium, or high concentrations with those determined by the AC method, there was no statistical significance among the values. The correlation among these methods was 1.0 with slope of 1, and the line of identity was almost identical to the regression line, demonstrating that the MG+PG technique is as precise as the traditional AC and EZ methods (Fig. 1). The MG/PG determination for individual muscle samples (n = 20) compared with AC also gave extremely high correlations. Thus the three methods are comparable.

MG/PG values calculated from comparison II suggest that as the total glycogen concentration rises so does the percentage of MG. We found that at a MG+PG concentration of 43.8 mmol glucosyl units/kg dw the MG/PG was 13:87. As the total increased to 181.8 mmol glucosyl units/kg dw, the ratio increased to 19:81 and to 25:75 as the total concentration reached 340.2 mmol glucosyl units/kg dw. The percentage of MG in the high-concentration pool was significantly higher than the percentage in the low pool. This is in accordance with Jansson's (10) results, which displayed an increasing percentage of soluble glycogen as the total concentration increased. She found that PCA-soluble muscle glycogen constituted 25% of the total glycogen content, which was 350 mmol/kg dw, and increases in total glycogen above that concentration seemed to be mainly due to the soluble type (MG). In the present study, in the individual muscle sample with the highest total glycogen (529.6 mmol glucosyl units/kg dw), the ratio was 38:62. Alonzo et al. (1) also showed that a greater percentage of PG relative to MG (molar percentage: 85% PG to 15% MG) in resting rabbit muscle. Interestingly, Friden et al. (5) reported that two populations of glycogen particles were clearly distinguished and, of the 144 particles counted in a pool in the intermyofibrillar space, ~76% were of the smaller particle size. This consensus suggests that MG+PG can be detected by electron microscopy.

Glucose analyses in the PCA extracts of human muscle before hydrolysis yielded low values (0.11-3.61 mmol/kg dw) (10). This demonstrated that glycogen was not hydrolyzed into glucose residues during the PCA-extraction procedure. Jansson (10) also demonstrated that the relationship between PCA-soluble and -insoluble glycogen was not influenced by strength of the PCA in the range from 0.5 to 3 M or the type of acid (PCA vs. trichloracetic acid). Nor was it affected by the freeze-dry procedure or the weight of the samples in the range from 0.2 to 2 mg. The benefit of using the MG+PG determination is in the ability to separate the two fractions of muscle glycogen and hence to study the metabolism of each one individually.

There are reports, with regard to different pools of glycogen, that date back to 1900, when Nerking (20), and later, to 1934, when Willstaetter and Rhodewald (21) concluded that tissue glycogen appeared in two forms: an acid-extractable or free form and an acid-nonextractable or protein-fixed form. It was not resolved whether the glycogen-protein complexes were artifacts or whether they constituted a physiological entity. Since then, investigators have shown different sizes, different concentrations, and different solubilities of glycogen molecules. Integrating the results of past studies suggests that data have consistently pointed to the existence of two different pools of muscle glycogen, with the smaller molecular/granular form being most common.

Lomako et al. (14) were able to demonstrate the movement of labeled glucose initially into the PG fraction, and then 30 min later the label appeared in the MG fraction while simultaneously falling in the PG fraction. A similar study conducted by Huang et al. (8) in resting rodent muscle demonstrated the incorporation of [³H]glucose into the PG fraction first and then into the MG fraction under conditions of moderate increases in [³H]glycogen synthesis. These results suggest that PG is the plausible precursor and that PG synthesis precedes that of MG. From the limited human data in the present study, it is speculated that PG is the small and dynamic, intermediate form of muscle glycogen, whereas MG is the larger storage form that appears to increase on a relative basis as the total glycogen increases. It is possible that when the glucose environment is favorable and the PG has reached a critical limit, a portion is synthesized into MG, and it is conceivable that the two pools may eventually equilibrate. With this new separation technique, it is now possible to study each pool of glycogen exclusively to understand how the fractions change under conditions of breakdown and resynthesis.

In summary, the AC and EZ determinations are not systematically different. MG and PG do exist in human skeletal muscle and can be measured accurately in biopsy-sized samples. The AC, EZ, and MG+PG determinations are reproducible and give the same values for total muscle glycogen.