A modified post-fixation procedure increases the contrast of glycogen particles in tissue. This paper provides a step-by-step protocol describing how to handle the tissue, conduct the imaging, and use stereological methods to obtain unbiased and quantitative data on fiber type-specific subcellular glycogen distribution in skeletal muscle.
With the use of transmission electron microscopy, high-resolution images of fixed samples containing individual muscle fibers can be obtained. This enables quantifications of ultrastructural aspects such as volume fractions, surface area to volume ratios, morphometry, and physical contact sites of different subcellular structures. In the 1970s, a protocol for enhanced staining of glycogen in cells was developed and paved the way for a string of studies on the subcellular localization of glycogen and glycogen particle size using transmission electron microscopy. While most analyses interpret glycogen as if it is homogeneously distributed within the muscle fibers, providing only a single value (e.g., an average concentration), transmission electron microscopy has revealed that glycogen is stored as discrete glycogen particles located in distinct subcellular compartments. Here, the step-by-step protocol from tissue collection to the quantitative determination of the volume fraction and particle diameter of glycogen in the distinct subcellular compartments of individual skeletal muscle fibers is described. Considerations on how to 1) collect and stain tissue specimens, 2) perform image analyses and data handling, 3) evaluate the precision of estimates, 4) discriminate between muscle fiber types, and 5) methodological pitfalls and limitations are included.
Glycogen particles are composed of branched polymers of glucose and various associated proteins1 and constitute an important fuel during high metabolic demands2. Although not widely recognized, glycogen particles also constitute a local fuel, where some subcellular processes preferentially utilize glycogen despite the availability of other and more long-lasting fuels as plasma glucose and fatty acids3,4.
The importance of storing glycogen as a subcellular specific localized fuel has been discussed in several reviews5,6 mainly based on some of the earliest documentations of the subcellular distribution of glycogen by transmission electron microscopy (TEM)7,8. The first studies used different protocols to increase the contrast of glycogen from histochemical staining techniques to negative and positive stainings9,10. An important methodological development was the refined post-fixation protocol with the potassium ferrocyanide-reduced osmium11,12,13,14, which significantly improved the contrast of glycogen particles. This refined protocol was not used in some of the pioneering work on exercise-induced glycogen depletion15 but was re-introduced by Graham and colleagues16,17.
Based on the 2-dimensional images, the subcellular distribution of glycogen is most often described as glycogen particles located in three pools: subsarcolemmal (just beneath the surface membrane), intermyofibrillar (between the myofibrils), or intramyofibrillar (within the myofibrils). However, glycogen particles could also be described as associated with, for example, sarcoplasmic reticulum7 or nuclei18. In addition to the subcellular distribution, the advantage of TEM-estimated glycogen content is also that quantification can be conducted at the single fiber level. This allows investigation of fiber-to-fiber variability and correlative analyses with fiber types and cellular components as mitochondria and lipid droplets.
Here, the protocol for the TEM-estimated fiber type-specific volumetric content of the three common subcellular pools of glycogen (subsarcolemmal, intermyofibrillar, and intramyofibrillar) in skeletal muscle fibers is described. The method has been applied to skeletal muscles from humans19, rats20, and mice21; as well as birds and fish22; and cardiomyocytes from rats23.
The present protocol using human biopsied skeletal muscle samples was approved by The Regional Committees on Health Research Ethics for Southern Denmark (S-20170198). Muscle biopsies were obtained through an incision in the skin from the vastus lateralis muscle using a Bergström needle with suction after local anesthesia was given subcutaneously (1-3 mL of Lidocaine 2% per incision). If isolated whole rat muscles were used, the animals were sacrificed by cervical dislocation before the muscle biopsies were obtained, in accordance with the guidelines of the animal ethics committee at Odense University Hospital, Denmark.
1. Primary fixation, post-fixation, embedding, sectioning, and contrasting
2. Imaging
3. Image analyses
4. Calculations
Using this protocol, glycogen particles appear black and distinct (Figures 1 and Figure 2). The normal values of glycogen are depicted in Figure 3. These data are based on a total of 362 fibers from 41 healthy young men as collected in different previous studies19,24,29,30,31. Here, it can be seen that intermyofibrillar glycogen values are distributed close to normal, whereas both intramyofibrillar and subsarcolemmal glycogen show a skewed distribution, where fibers sometimes have an excessive amount of glycogen. It is important to note that in normal-sized muscle fiber (diameter of 60-80 µm), intermyofibrillar glycogen is the largest pool constituting around 80% of total glycogen content. Intramyofibrillar and subsarcolemmal glycogen each constitute around 10% of the total content.
Figure 1: Imaging and fiber typing. (A) Each fiber is imaged in a randomized systematic order. (B) Example of an image from the subsarcolemmal space. (C) Example of an image from the myofibrillar space. (D) In each myofibrillar image, the width of one Z-disc is measured (red lines). The measurements of a total of 12 Z-discs (one per image) give a coefficient of error of approximately 0.03. (E) The typical distribution of the average fiber Z-disc width in 6-10 fibers of each of the 10 biopsies. From each biopsy, 2-3 fibers are defined as types 1 and 2 based on the within-biopsy distribution. The images originate from a biopsy of m. vastus lateralis of a powerlifter included in a previous study29. m: mitochondria and Z: Z-disc. Please click here to view a larger version of this figure.
Figure 2: Glycogen analyses. (A) Subsarcolemmal glycogen volume per surface area is estimated by point counting using a grid size of 180 nm x 180 nm within a region defined by the length of the outermost myofibril and the subsarcolemmal region perpendicular to this length (blue dotted lines). (B) The myofibrillar volume fraction is estimated by point counting using a grid size of 400 nm x 400 nm. (C) The volume fraction of intramyofibrillar glycogen is estimated by point counting using a grid size of 60 nm x 60 nm. (D) The volume fraction of intermyofibrillar glycogen is estimated by point counting using a grid size of 180 nm x 180 nm. In A–D, the red circles indicate hits (a cross that hits a glycogen particle). (E) The estimated coefficient of error for a stereological ratio estimate24 for 2 to 12 analyzed images. The coefficient of error is estimated based on the number of counts and therefore varies between samples based on the glycogen concentration. It is often relatively low when the glycogen content is high and vice versa. (F) The coefficient of variation of glycogen particle diameter after measuring 2-99 particles. Please click here to view a larger version of this figure.
Figure 3: Normal values of the three subcellular pools of glycogen in skeletal muscle. The violin plots are based on 362 fibers from 41 healthy young men (18-39 years old). The fibers originate from previous studies, wherein biopsies from m. vastus lateralis in a resting or control condition were obtained19,24,29,30,31. Values are shown as a box plot with a marker for the median and a box indicating the interquartile range. The lines represent upper and lower adjacent values. The boxes are overlaid by kernel density plots. Please click here to view a larger version of this figure.
The critical step of the method is the use of reduced osmium by potassium ferrocyanide during post-fixation. The selectivity of this modified fixative for glycogen detection cannot be fully explained by chemistry, but also includes experimental findings demonstrating no detection of such particles in tissues known to be free of glycogen or in the extracellular space11.
Critical parameters are the precision of the estimates and the fiber-to-fiber variation. By following the present protocol for imaging, a coefficient of error between 0.1 and 0.2 of the estimates of the different pools of glycogen per fiber is obtained. This level of error is well below the variation between individual fibers (Figure 3). It is encouraged to report such precision estimates when estimating the volumetric content of glycogen. The presented fiber typing method is validated against myosin ATPase isoform29. The Z-disc thickness and mitochondrial volume fraction can also be used in combination to indicate fiber type, but not mitochondrial volume fraction alone32.
The major limitations of the method are the inability to detect the very small glycogen particles and that profiles of glycogen particles may overlap in the projected image28. The first limitation invalidates a true measure of the average particle size. This becomes a severe bias when the glycogen particles are being degraded during high metabolic demands, whereas the bias may be insignificant when the glycogen particles grow from medium to a larger size during glycogen resynthesis or super-compensation. While this may have huge implications for the estimate of average glycogen particle size at low glycogen levels, the estimates of volumetric glycogen concentrations are robust, since small, unobserved glycogen particles contribute very little toward the total glycogen content. The second limitation originates from the condition, where the glycogen particles are much smaller than the thickness of the sections. This bias is mostly present at very high glycogen concentrations and could be investigated by comparing the glycogen volume fractions of sections with different thicknesses. If a thicker section is not paralleled by a high glycogen volume fraction, it must be due to an underestimation due to more overlapping particles in the thickest section. In previous studies, the glycogen volume fraction correlates with the glycogen concentration within the range from 50 to 600 mmol kg dw-1 indicating no pronounced overlapping of particles. However, if the glycogen concentration increases above this level, there is no increase in intermyofibrillar glycogen indicating overlap33. This can be solved by extrapolating the relationship between the glycogen volume fraction and the concentration at the lower glycogen concentrations.
Based on the nm resolution provided by TEM, this protocol is at present the only method to estimate the subcellular distribution of glycogen. In addition, the methodology also permits a large-scale quantitative approach (as described here), where quantitative values can be obtained at the single fiber level. This is of immense importance in skeletal muscles with high heterogeneity in fiber recruitment during various types of exercise2, where glycogen-dependent fatigue mechanisms only occur in some fibers. The method also has potential for other excitable tissues as cardiomyocytes, where glycogen is known to be essential for normal heart function and critical during ischemia23,34.
The authors have nothing to disclose.
This work was supported by the Swedish Olympic Committee.
1,2-Propylene oxide | Merck | 75-56-9 | |
Embedding 812 resin medium kit | Taab | T031 | |
Glutaraldehyde solution 25% | Merck | 1.04239.0250 | |
ITEM | Olympus | Imaging software | |
Leica EM AC20 | Leica | Automatic contrasting system | |
OSIS Veleta digital camera | Olympus | ||
Osmium tetroxide 4% solution | Polysciences | 0972A | |
Philips CM 100 Transmission EM | Philips | ||
Potassium hexacyanoferrate (II) trihydrate | Sigma-Aldrich | 455989-245G | |
Sodium cacodylatbuffer 0,2 M ph 7.4 | Ampliqon.com | AMPQ40989.0500 | |
Ultra-microtome Leica UC7 | Leica | ||
Ultrostain lead citrate 3%, stabilised solution | Leica | 16707235 | |
Uranyl acetate dihydrate | Polysciences | 6159-44-0 |