In order to evaluate 3 staining methods for demonstration of glycogen in thin sections, 2 tissues containing an abundance of this carbohydrate in β-particle form were studied. Tissues were aldehyde-fixed, postfixed in osmium tetroxide, embedded in Araldite and sectioned in the usual manner without special precautions. The first method for staining thin sections employed a sequential combination of periodic acid, thiosemicarbazide and osmium tetroxide vapour, while in the second procedure a silver protein solution was substituted for the osmium tetroxide vapour. The third technique utilized periodic acid, sodium chlorite and uranyl acetate, also in sequential combination. Each method yielded glycogen particles of greater electron density than were seen in sections stained by the usual uranyl acetate-lead citrate procedure. Under high magnification, considerable method-dependent variation in the appearance of the glycogen granules was noted. Particulate substructure, only faintly visible in routinely stained sections, was easily resolved with the periodic acid-thiosemicarbazide-silver protein technique. Conversely, periodic acid-thiosemicarbazide—osmium tetroxide completely obscured this substructure. With periodic acid—sodium chlorite—uranyl acetate, glycogen particles appeared larger, more confluent, and of a less regular outline than with the other methods. Sections were also stained by incubation in periodic acid prior to treatment with lead citrate. The alteration in appearance of the glycogen granules produced by this modification was so great that high-resolution analysis of particle size and substructure could not be undertaken. The usefulness of the procedures investigated here resides in their ability to stain glycogen in thin sections in an intense and selective manner.

The ultrastructural demonstration of glycogen in thin sections of osmium-fixed tissue presented a considerable problem for early workers in electron microscopy. This difficulty resulted from the paucity of chemical groups in glycogen capable of binding or reducing osmium tetroxide (Revel, 1964). Thus, even if glycogen particles were structurally preserved by osmium fixation, their electron density was so low that effective visualization was not possible (Fawcett, 1955; Nilsson, 1962). Double fixation with aldehydes and OsO4 did not obviate this problem, for although aldehydes improved structural preservation, the glycogen particles were still of insufficient electron density to be adequately resolved.

Successful efforts to find an electron-dense stain for glycogen began with the report by Swift & Rasch (1958) of the staining of glycogen particles in osmium-fixed tissues by phosphotungstic and phosphomolybdic acid. Watson (1958 a) confirmed this finding with phosphomolybdic acid, but in a later report (1958b) showed that lead hydroxide was a superior stain for glycogen. Subsequently, the classic study of Revel, Napolitano & Fawcett (1960) conclusively demonstrated that glycogen particles were stained intensely by lead hydroxide after either osmium tetroxide or potassium permanganate fixation. Lead citrate was shown to have similar properties (Reynolds, 1963).

Though quite effective for demonstrating glycogen, lead solutions were nonspecific stains which enhanced the contrast of many cellular components (Revel, 1964) and did not reveal or reflect any of the specific chemical properties of glycogen. Consequently, attempts were made to develop a selective staining reaction for polysaccharides. A number of investigators (Suzuki & Sekiyama, 1961; Movat, 1961; Steiner & Carruthers, 1961; Marinozzi, 1961) adapted the periodic acid-methenamine silver method of Gomori (1952) for staining of polysaccharides in plastic-embedded thin sections, and Movat (1961) was successful in demonstrating glycogen with it. This finding, though disputed by Rambourg & Leblond (1967), has been confirmed independently by Marinozzi (1963) and Thiéry (1967). However, the technique was also non-specific, for the silver reaction product was deposited on many intracellular structures, particularly membranes (Marinozzi, 1961). Furthermore, the coarse granularity of the precipitate precluded high-resolution analysis of the glycogen particles (Movat, 1961; Marinozzi, 1961; Thiéry, 1967).

Subsequently, Hanker et al. (1964) introduced the osmiophilic reagent thiosemicarbazide (TSC) for the ultrastructural demonstration of macromolecules in thin sections. This reagent binds to 1,2-dialdehyde groups generated by the periodic acid oxidation of carbohydrates and subsequently reduces osmium tetroxide to electrondense osmium black. The specificity of this reaction is similar to the PAS stain as used in light microscopy (Seligman et al. 1965; Hanker et al. 1965). Intense, selective staining of glycogen particles in thin sections of fixed tissue was produced by this procedure (Hanker et al. 1965).

The ultrastructural methods for staining glycogen were reviewed by Thiéry (1967), who introduced a modification of the technique of Hanker et al. (1964) in which a solution of silver protein replaced OsO4 vapour. This reagent, by eliminating the need for osmium tetroxide vapour, greatly increased the convenience of the method. In addition, the staining produced by this modification was particularly delicate and revealed a definite substructure in the glycogen particles.

Recently, in a study of glycogen-rich tissues, we have employed 3 methods for staining glycogen particles; 2 involved the use of TSC while the third utilized sodium chlorite and uranyl acetate. Though each procedure clearly demonstrated the glycogen granules, many differences were noted in their ultrastructural appearance depending upon the method used. These findings are the subject of this report.

Preparation of tissue for electron microscopy

Two glycogen-rich tissues were investigated. The first, human skeletal muscle from three patients with Type V glycogenosis, commonly called McArdle’s disease (McArdle, 1951), was obtained by open biopsy under local anaesthesia and maintained at rest length in a muscle tension clamp (Price, Howes, Blumberg & Pearson, 1965). The biopsies were immediately fixed for 2 h in 2·5 % glutaraldehyde or 4% paraformaldehyde in Millonig’s buffer (Millonig, 1961) at pH 7·4. Following initial fixation, blocks measuring 1 mm3 were dissected from the surfaces of the specimens and fixed for an additional 12 h. The blocks were stored in buffer for a period not exceeding 24 h prior to embedding. The second tissue, chick glycogen body, was obtained from the lumbar enlargement of the spinal cord of 20-day White Leghorn chick embryos. After being cut into i-mm’ blocks, they were fixed and stored in the same manner as the skeletal muscle.

Following primary fixation, both tissues were postfixed in 1 % osmium tetroxide in Millonig’s buffer for 1 h, after which they were dehydrated in graded ethanols and embedded in Araldite. One-micron thick sections for light-microscopic survey were cut with glass knives and stained with toluidine blue-O. Thin sections were cut with diamond knives and mounted on uncoated copper or gold grids.

Staining methods for glycogen

The 3 methods used for the demonstration of glycogen required periodic acid oxidation. For these procedures sections were mounted on inert gold grids. All steps of the staining, except rinsing in H.O, were performed by floating the grids on a drop of reagent solution. (Reagents were purchased from the following suppliers: thiosemicarbazide, Eastman Organic Chemicals; silver protein (mild), Merck and Co.; lead citrate, K and K Laboratories, Inc.; uranyl acetate, Fisher Scientific Co.; sodium chlorite, Matheson, Coleman and Bell. Periodic acid and osmium tetroxide were obtained from several sources.)

Periodic acid-thiosemicarbazide-osmium tetroxide (PA-TSC-OT)

The method described here is essentially similar to the technique published by Hanker et al. (1964). Thin sections mounted on gold grids were incubated in 1 % periodic acid for 30 min, following which they were rinsed for 1 min in a jet of distilled water. The sections were then treated for 60 min with a 1 % solution of thiosemicarbazide dissolved in 25 % acetic acid. Following a second rinse in distilled water, the sections were reacted with osmium tetroxide vapour by mounting the grids on coverslips which were placed in small Coplin jars containing OsO4 crystals. The containers were sealed with stopcock grease and incubated for 3 h in a waterbath at 60 °C to volatilize the osmium.

Periodic acid-thiosemicarbazide-silver protein (PA-TSC-SP)

This method, a modification of the previous technique, was reported by Thiéry (1967). The procedure given here differs from his description in that TSC was dissolved in 25% acetic acid rather than 10%, that sections were rinsed with distilled water rather than acetic acid, and that sections were mounted on grids before staining instead of after it. Using this method, sections were reacted with periodic acid, rinsed in distilled water, incubated in thiosemicarbazide and rinsed again in a manner identical to the PA-TSC-OT technique. At this point the sections were treated with i % aqueous solution of silver protein for 30 min followed by a final i-min rinse with a jet of distilled water. The silver protein reagent was prepared fresh and used immediately in the dark.

Periodic acid-sodium chlorite-uranyl acetate (PA-SC-UA)

This technique was developed by M. J. Kamovsky & J. P. Revel (personal communication) but has not been published. Sections were reacted with 1 % periodic acid for 30 min and rinsed with distilled water as in the other methods. Subsequently, they were incubated for 15 min in a 2·5 % solution of sodium chlorite dissolved in 10% acetic acid. The sections were again rinsed, following which they were treated for 10 min with a saturated aqueous solution of uranyl acetate. A final rinse in a jet of distilled water completed the procedure. The sodium chlorite reagent must also be prepared fresh and used in the dark.

Other grids were stained by the conventional method employing lead citrate in sequential combination with saturated uranyl acetate in 50% alcohol (UA-LC). Lead citrate staining was also employed alone following oxidation with 2·5 % periodic acid for 30 min, as recommended by Marinozzi (1963) and modified by Perry (1967).

Two sets of control grids employing incomplete staining reactions were prepared. In the first, water was substituted for periodic acid, while in the second, water or acetic acid was substituted for thiosemicarbazide (first 2 methods) or sodium chlorite (last method).

McArdle’s disease (McArdle, 1951) is an uncommon type of glycogenosis which affects skeletal muscle and is due to a genetically determined deficiency of the glycolytic enzyme phosphorylase (Schmid & Mahler, 1959; Pearson, Rimer & Mommaerts, 1959). The ultrastructure of skeletal muscle in this condition was first described by Schotland, Spiro, Rowland & Carmel (1965). Several subsequent reports (Delwaide et al. 1967; Boudouresques et al. 1967; Brownell, Trevor-Hughes, Goldby & Woods, 1969) have confirmed their findings, and the appearance of the tissues studied here does not differ significantly from that described in these reports. The muscle fibres are distorted by large subsarcolemmal and intermyofibrillar aggregates of glycogen (Fig. 1). These intermyofibrillar pools occur mainly in the I-band, but the number of particles located within the myofibrils is also increased. There is no constant association of the particles with intracellular membranes, nor are the glycogen aggregates membrane-bound. Individual granules resemble the β-particles described by Drochmans (1962). The biochemical structure of the glycogen molecule in this condition has been reported to be normal (Mommaerts et al. 1959).

Fig. 1.

Skeletal muscle from a patient with Type V glycogenosis stained with uranyl acetate and lead citrate. Large aggregates of particulate glycogen are located beneath the sarcolemma and between the myofibrils. The number of particles between myofilaments within the myofibrils is also increased. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Fig. 1.

Skeletal muscle from a patient with Type V glycogenosis stained with uranyl acetate and lead citrate. Large aggregates of particulate glycogen are located beneath the sarcolemma and between the myofibrils. The number of particles between myofilaments within the myofibrils is also increased. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

The ultramicroscopio appearance of the embryonic chick glycogen body has been described by Revel et al. (1960). The cytoplasm of the cells consists of a large pool of glycogen particles. Nuclei are peripherally placed and a few organelles are located in a perinuclear position (Fig. 2).

Fig. 2.

Glycogen body from 20-day chick embryo stained with uranyl acetate and lead citrate. The cytoplasm of the cells consists almost entirely of a large pool of glycogen particles, g, glycogen; m, mitochondrion; n, nucleus, × 15000.

Fig. 2.

Glycogen body from 20-day chick embryo stained with uranyl acetate and lead citrate. The cytoplasm of the cells consists almost entirely of a large pool of glycogen particles, g, glycogen; m, mitochondrion; n, nucleus, × 15000.

In the 3 methods used for the selective staining of glycogen, sections were oxidized with periodic acid but not treated with lead solutions. Such sections, when compared with tissue stained by the usual uranyl acetate-lead citrate (UA-LC) procedure, exhibited glycogen particles which appeared to be of greater electron density than with the routine method (Figs. 3-8). At low magnification (less than x 15000) results with the 3 stains appeared quite similar. Since periodic acid oxidation removes much of the reduced osmium from the tissues (Merriam, 1958; Seligman, 1966), glycogen particles were revealed as densely stained globular bodies against a pale background.

Fig. 3.

Skeletal muscle as in Fig. 1 stained by the periodic acid-thiosemicarbazideosmium tetroxide (PA-TSC-OT) method. The electron-dense glycogen particles stand out in sharp relief against the paler structures of the muscle fibre. Though much reduced in electron density, the structure of the mitochondria and myofibrils is still discernible. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Fig. 3.

Skeletal muscle as in Fig. 1 stained by the periodic acid-thiosemicarbazideosmium tetroxide (PA-TSC-OT) method. The electron-dense glycogen particles stand out in sharp relief against the paler structures of the muscle fibre. Though much reduced in electron density, the structure of the mitochondria and myofibrils is still discernible. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Though each method stained glycogen intensely, differences were noted in the size and configuration of the granules. In Table 1 the size distribution of glycogen particles from both tissues is presented for each method as well as for routine UA-LC staining. Even though the range of particle size was quite great, certain pertinent observations can be made. In all cases, the average particle diameter was greater in the glycogen body than in muscle. With each method, the difference between the 2 tissues was statistically significant (P < 0·01). A further observation was that particle diameter in both the glycogen body and skeletal muscle, when measured after PA-SC-UA staining, was significantly larger than with the other techniques (P < 0·01).

Table 1.

Diameter of glycogen particles with various staining methods

Diameter of glycogen particles with various staining methods
Diameter of glycogen particles with various staining methods

In addition to these quantitative variations, there were qualitative differences in the appearance of the glycogen granules. In sections conventionally stained with uranyl acetate and lead citrate, glycogen appeared as discrete or partially interconnected spheroidal particles which demonstrated a faint but constant substructure (Figs. 9, 10). With both PA-TSC-OT and PA-TSC-SP, the general outline of the particles was maintained, but no substructure could be seen with certainty in the PA-TSC-OT material (Figs. 11, 12). Glycogen stained with PA-TSC-SP, however, revealed a definite substructure within the granules which was similar to that seen in UA-LC stained particles, but more easily resolved (Figs. 13, 14). Each particle appeared to be composed of globular subunits measuring about 3 nm in diameter (inset, Fig. 14). With the PA-SC-UA method glycogen granules appeared as a series of interconnected units which were quite variable in size and shape (Figs. 15, 16), and usually exhibited a light centre which gave them a doughnut-like appearance (inset, Fig. 16). The globular subunits were not well defined by this technique.

Fig. 4.

Glycogen body as in Fig. 2 stained by the PA-TSC-OT procedure. Electrondense glycogen granules dominate the ultrastructural appearance. The shadow of a nucleus is faintly visible, g, glycogen; n, nucleus, × 15000.

Fig. 4.

Glycogen body as in Fig. 2 stained by the PA-TSC-OT procedure. Electrondense glycogen granules dominate the ultrastructural appearance. The shadow of a nucleus is faintly visible, g, glycogen; n, nucleus, × 15000.

Fig. 5.

Skeletal muscle as in Fig. 1 stained by the periodic acid-thiosemicarbazidesilver protein (PA-TSC-SP) method. The tendency of glycogen to aggregate in the I-band is apparent in this section. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Fig. 5.

Skeletal muscle as in Fig. 1 stained by the periodic acid-thiosemicarbazidesilver protein (PA-TSC-SP) method. The tendency of glycogen to aggregate in the I-band is apparent in this section. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Fig. 6.

Glycogen body as in Fig. 2 stained by the PA-TSC-SP technique. The section is composed almost entirely of densely stained glycogen particles, g, glycogen, × 15000.

Fig. 6.

Glycogen body as in Fig. 2 stained by the PA-TSC-SP technique. The section is composed almost entirely of densely stained glycogen particles, g, glycogen, × 15000.

Fig. 7.

Skeletal muscle as in Fig. 1 stained by the periodic acid-sodium chloriteuranyl acetate (PA-SC-UA) procedure. The similarity to Figs. 3 and 5 is apparent. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Fig. 7.

Skeletal muscle as in Fig. 1 stained by the periodic acid-sodium chloriteuranyl acetate (PA-SC-UA) procedure. The similarity to Figs. 3 and 5 is apparent. A, A-band; I, I-band; Z, Z-line; g, glycogen; m, mitochondrion, × 15000.

Fig. 8.

Glycogen body as in Fig. 2 stained by the PA-SC-UA method. The density of the glycogen granules is similar to Figs. 4 and 6. g, glycogen; m, mitochondrion, × 15000.

Fig. 8.

Glycogen body as in Fig. 2 stained by the PA-SC-UA method. The density of the glycogen granules is similar to Figs. 4 and 6. g, glycogen; m, mitochondrion, × 15000.

Fig. 9.

Skeletal muscle stained with uranyl acetate and lead citrate. The glycogen particles appear as a series of interconnected granules in which a faint substructure can be discerned, mf, myofibril, × 100000.

Fig. 9.

Skeletal muscle stained with uranyl acetate and lead citrate. The glycogen particles appear as a series of interconnected granules in which a faint substructure can be discerned, mf, myofibril, × 100000.

Fig. 10.

Glycogen body stained with uranyl acetate and lead citrate. The glycogen particles are also seen as a series of interconnected granules displaying a faint substructure. At higher magnification (inset) the substructure is well resolved and shows the glycogen granules to be composed of subunits measuring about 3 nm in diameter. When compared with Fig. 9, the larger size of the particles in the glycogen body is apparent, x 100000; inset, × 250000.

Fig. 10.

Glycogen body stained with uranyl acetate and lead citrate. The glycogen particles are also seen as a series of interconnected granules displaying a faint substructure. At higher magnification (inset) the substructure is well resolved and shows the glycogen granules to be composed of subunits measuring about 3 nm in diameter. When compared with Fig. 9, the larger size of the particles in the glycogen body is apparent, x 100000; inset, × 250000.

Fig. 11.

Skeletal muscle stained by the PA-TSC-OT technique. Though glycogen particles are well stained, no substructure is noted, mf, myofibril, × 100000.

Fig. 11.

Skeletal muscle stained by the PA-TSC-OT technique. Though glycogen particles are well stained, no substructure is noted, mf, myofibril, × 100000.

Fig. 12.

Glycogen body stained by the PA-TSC-OT procedure. Electron-dense interconnected glycogen particles are clearly delineated. Even at high magnification (inset) a well-defined substructure cannot be resolved. The larger size of the particles in the glycogen body as compared to skeletal muscle (Fig. 11) is apparent, × 100000; inset, × 250000.

Fig. 12.

Glycogen body stained by the PA-TSC-OT procedure. Electron-dense interconnected glycogen particles are clearly delineated. Even at high magnification (inset) a well-defined substructure cannot be resolved. The larger size of the particles in the glycogen body as compared to skeletal muscle (Fig. 11) is apparent, × 100000; inset, × 250000.

Fig. 13.

Skeletal muscle stained by the PA-TSC-SP method. The glycogen particles have an easily resolvable substructure which is composed of globular units measuring about 3 nm in diameter, × 100000.

Fig. 13.

Skeletal muscle stained by the PA-TSC-SP method. The glycogen particles have an easily resolvable substructure which is composed of globular units measuring about 3 nm in diameter, × 100000.

Fig. 14.

Glycogen body stained by the PA-TSC-SP technique. The glycogen particles also have a clearly resolved substructure composed of small globular fragments which measure about 3 nm in diameter; at high magnification (inset) this subunit is quite similar to that seen in the uranyl acetate—lead citrate stained tissue (Fig. 10). The larger size of the glycogen granules in the glycogen body is also apparent, x 100000; inset, × 250000.

Fig. 14.

Glycogen body stained by the PA-TSC-SP technique. The glycogen particles also have a clearly resolved substructure composed of small globular fragments which measure about 3 nm in diameter; at high magnification (inset) this subunit is quite similar to that seen in the uranyl acetate—lead citrate stained tissue (Fig. 10). The larger size of the glycogen granules in the glycogen body is also apparent, x 100000; inset, × 250000.

Fig. 15.

Skeletal muscle stained by the PA-SC-UA method. The glycogen particles assume a less regular, more interconnected appearance. Substructure within the granules is not clearly visualized, m, mitochondrion; mf, myofibril, × 100000.

Fig. 15.

Skeletal muscle stained by the PA-SC-UA method. The glycogen particles assume a less regular, more interconnected appearance. Substructure within the granules is not clearly visualized, m, mitochondrion; mf, myofibril, × 100000.

Fig. 16.

Glycogen body stained by the PA-SC-UA procedure. The glycogen particles have an irregular outline and are more highly interconnected. Many granules have a lucent centre, imparting a doughnut-like appearance. Substructure is not well resolved. Because of the irregular shape, differences in particle size between muscle and glycogen body are less apparent with this procedure, × 100000; inset, × 250000.

Fig. 16.

Glycogen body stained by the PA-SC-UA procedure. The glycogen particles have an irregular outline and are more highly interconnected. Many granules have a lucent centre, imparting a doughnut-like appearance. Substructure is not well resolved. Because of the irregular shape, differences in particle size between muscle and glycogen body are less apparent with this procedure, × 100000; inset, × 250000.

When grids were stained with lead citrate alone following oxidation with 2·5% periodic acid, as recommended by Perry (1967), the structure of glycogen was altered considerably (Figs. 17, 18). Margins of the β-particles were obscured, and in their place a continuum of irregularly sized granules measuring between 3 and 24 nm in diameter was observed. The smallest particles resembled subunits seen with UA-LC and PA-TSC-SP staining; however, it was difficult to relate them to a larger particulate structure. Because of this highly variable appearance, size analysis of the β-particles could not be performed.

Fig. 17.

Skeletal muscle stained with lead citrate preceded by incubation in periodic acid. Though glycogen stains intensely, the particulate structure has been greatly altered. In such sections glycogen appears to be composed of granules which vary in size from 3·0 to 24 nm. The smallest of these resemble the subunits seen in sections stained with uranyl acetate and lead citrate (Fig. 9) or PA-TSC-SP (Fig. 13), though it is difficult to relate these units to any larger structural component, mf, myofibril, × 100000.

Fig. 17.

Skeletal muscle stained with lead citrate preceded by incubation in periodic acid. Though glycogen stains intensely, the particulate structure has been greatly altered. In such sections glycogen appears to be composed of granules which vary in size from 3·0 to 24 nm. The smallest of these resemble the subunits seen in sections stained with uranyl acetate and lead citrate (Fig. 9) or PA-TSC-SP (Fig. 13), though it is difficult to relate these units to any larger structural component, mf, myofibril, × 100000.

Fig. 18.

Glycogen body stained with lead citrate preceded by incubation in periodic acid. The ultrastructural alterations of the glycogen particles in this tissue resemble those seen in muscle treated in a similar manner (Fig. 17). × 100000.

Fig. 18.

Glycogen body stained with lead citrate preceded by incubation in periodic acid. The ultrastructural alterations of the glycogen particles in this tissue resemble those seen in muscle treated in a similar manner (Fig. 17). × 100000.

In control specimens employing an incomplete staining protocol, glycogen did not appear electron-dense. The particles presented instead as electron-lucent bodies against a darker background (Figs. 19, 20). Thus, the pictures produced by the incomplete staining procedures were virtually the negative image of the complete reactions.

Fig. 19.

Glycogen body stained by the PA-TSC-OT technique but with H2O substituted for PA. The glycogen particles appear as electron-lucent globules against a dark background, g, glycogen, × 15000.

Fig. 19.

Glycogen body stained by the PA-TSC-OT technique but with H2O substituted for PA. The glycogen particles appear as electron-lucent globules against a dark background, g, glycogen, × 15000.

Fig. 20.

Glycogen body stained by the PA-TSC-OT procedure but with H2O substituted for TSC. Glycogen again appears as electron-lucent particles against a dark background. Other staining methods using an incomplete reaction protocol yielded glycogen particles of similar electron density, g, glycogen, × 15000.

Fig. 20.

Glycogen body stained by the PA-TSC-OT procedure but with H2O substituted for TSC. Glycogen again appears as electron-lucent particles against a dark background. Other staining methods using an incomplete reaction protocol yielded glycogen particles of similar electron density, g, glycogen, × 15000.

Though lead solutions produce satisfactory delineation of glycogen particles in most situations, the non-specific nature of this staining method has long been recognized (Revel, 1964). In addition, lead staining of glycogen is often unpredictable, considerable variation in density being noted with different preparations of reagent and even with the same preparation at different times. In view of these difficulties,3methods for staining glycogen which do not employ lead have been investigated. Both the PA-TSC-OT and PA-TSC-SP techniques are based on a principle similar to the light-microscopic PAS stain (Hanker et al. 1965; Seligman et al. 1965). Thus, they are capable of staining substances with dialdehyde linkages or chemical groups which can be converted to dialdehydes by periodic acid oxidation. The PA-TSC-OT method has been used to study the extraction of glycogen by diastase digestion from Epon-embedded liver tissue (Rosa & Johnson, 1967), the distribution of glycogen in McArdle’s disease (Vye, 1968), extraction of glycogen by en bloc staining with uranyl acetate (Vye & Fischman, 1970), and the structure of Armanni-Epstein lesions in renal tubular cells of animals with experimental diabetes (Vye, Friederici & Vargas, submitted for publication). The PA-TSC-SP technique has been employed in the investigation of mucopolysaccharide synthesis by the Golgi apparatus of intestinal epithelial cells (Thiéry, 1969), and to localize glycogen in spermatozoa (Anderson & Personne, 1970). The histochemical basis of the PA-SC-UA technique has not been clearly elucidated, nor is it known if its reaction mechanism is similar to the 2 methods discussed above.

When the diameters of glycogen particles in the chick glycogen body and muscle from patients with McArdle’s disease were compared, it was found that with all 4 staining methods the granules were significantly larger in the glycogen body. Revel et al. (1960) reported an average diameter of glycogen particles in permanganatefixed, lead-hydroxide-stained chick glycogen body to be about 30 nm. However, he noted considerable variability of particle size, with a range from 17 to 100 nm in diameter. In two cases of McArdle’s disease, Schotland et al. (1965) reported that glycogen particles ranged from 15 to 50 nm in diameter, but did not report a mean value. Biava (1963), in a study of many human tissues including skeletal muscle, stated that the mean diameter of monoparticulate glycogen varied from 22 to 30 nm following osmium tetroxide fixation and lead hydroxide staining. A specific figure for skeletal muscle, however, was not given. Wanson & Drochmans (1968a), on the other hand, reported a mean diameter of glycogen particles in glutaraldehyde-fixed, lead-hydroxide-stained rabbit muscle to be 27·3 ± 3 nm. Therefore, the particle size noted here for lead-stained skeletal muscle is in good agreement with previous investigations. With chick glycogen body the mean particle diameter of 37-1 nm reported in this study is somewhat larger than the 30 nm given by Revel et al. (1960). However, in view of the large range of particle size in both studies, such a discrepancy is not surprising. Because the difference in particle size between glycogen body and skeletal muscle has been noted with all staining methods, it is probable that these figures reflect a real difference in size of the glycogen granules between the 2 tissues.

Perhaps the most interesting question raised by these results relates to the method-dependent appearance of the glycogen particles in sectioned tissue. The clearest elucidation of glycogen β-particle structure has been obtained from the study of negatively stained preparations of isolated glycogen from both rat liver (Drochmans, 1962) and rabbit skeletal muscle (Wanson & Drochmans, 1968a), in which the granules were found to be composed of small, regular subunits about 3 nm in diameter. Limitations imposed by thin sectioning make analysis of particle structure in fixed tissue considerably more difficult. Section thickness is hard to control precisely and granularity of the image at high magnification may be confused with particle substructure. Despite these problems, a substructure similar to that seen in negatively stained particles was present in lead-stained material. This is in accord with the previous findings of Revel et al. (1960) and Biava (1963). Staining with PA-TSC-SP improved the resolution of the subunits, as previously noted by both Thiéry (1967) and Anderson & Personne (1970). In view of this fact, it is curious that the PA-TSC-OT method, which is based on the same chemical principle, not only fails to amplify the substructure, but actually obscures it.

With PA-SC-UA, the particles were significantly larger and had a somewhat different appearance; only a faint substructure was visible. These results raise the question of what is actually being stained by the various methods. In rat liver, isolated glycogen particles have been shown to have enzymically active protein bound to them (Leloir & Goldemberg, 1960; Luck, 1961). Wanson & Drochmans (1968 a) were unable to reduce the protein content of isolated rabbit muscle glycogen particles below 3% without destroying the structural integrity of the particles, and these protein-containing particles were demonstrated to possess phosphorylase activity (Wanson & Drochmans, 1968b). Thus, it seems likely that glycogen in the cell exists as a protein-polysaccharide complex in which some of the enzymes of glycolysis are present. The PA-TSC-OT and PA-TSC-SP procedures are based on the reaction of dialdehydes produced by periodic acid oxidation. Consequently, with these methods the polysaccharide, which contains many diol groups, should be the major component that is stained. On the other hand, uranyl acetate, at least when used alone, is a relatively non-specific stain capable of reacting with proteins and nucleic acids (Watson, 1958 a). It is possible, therefore, that the unusual configuration of glycogen particles seen with PA-SC-UA staining may result from staining of protein as well as the polysaccharide component of the glycogen particle.

Further evidence for the method-dependent appearance of the glycogen particles was provided by the curious morphological changes that were wrought when lead staining was preceded by periodic acid oxidation. The chemical basis of these alterations is obscure.

The principal advantage of the techniques investigated here is that each provides a selective method for the demonstration of glycogen in thin sections. Furthermore, unlike the methods described by de Bruyn (1968) and Bradbury & Stoward (1967, 1968), special preparation of the tissues during fixation or prior to embedding are not required. The PA-TSC-SP method produces an ultrastructural configuration of the glycogen particles which is in accord with data from negative staining and with the appearance of the granules in lead-stained sections. The PA-TSC-OT method has been quite useful in differentiating glycogen from ribosomes in sections of liver which have been double-fixed with glutaraldehyde and osmium tetroxide (M. V. Vye, unpublished observations). It is not possible, however, to state with certainty which stain, if any, demonstrates glycogen as it truly exists in situ. The major disadvantages of the methods relate to the time required for preparation of reagents and staining, as well as the necessity of using uncoated inert grids. The lack of staining of other tissue components may be a disadvantage in some situations, although this is not the case when the selective effect of a ‘special stain’ is sought.

Because of their ability to stain polysaccharides selectively in tissues fixed and embedded by standard techniques, these cytochemical methods should be of value in many areas of ultrastructural investigation.

The authors wish to express their appreciation to Misses Rita Yambot, Irena Kairys and Mrs Lucy Vedegys for their excellent technical assistance. This work was supported by Grant No. 315 from the General Research Support Grant of the University of Illinois (to M.V.V.) and No. GB-7591 from the National Science Foundation and No. N-69-32 from the Chicago and Illinois Heart Association (to D.A.F.).

Anderson
,
W. A.
&
Personne
,
P.
(
1970
).
The localization of glycogen in the spermatozoa of various invertebrate and vertebrate species
.
J. Cell Biol
.
44
,
29
51
.
Biava
,
C.
(
1963
).
Identification and structural forms of human particulate glycogen
.
Lab. Invest
.
12
,
1179
1197
.
Boudouresques
,
J.
,
Roger
,
J.
,
Toga
,
M.
,
Daniel
,
F.
,
Dubois
,
D.
,
Aquaron
,
R.
&
Poncet
,
M.
(
1967
).
Un cas de myopathie type McArdle (observation clinique, données biochemiques, constataious histopathologiques et ultrastructurales)
.
Revue neurol
.
117
,
650
662
.
Bradbury
,
S.
&
Stoward
,
P. J.
(
1967
).
The specific cytochemical demonstration in the electron microscope of periodate-reactive mucosubstances and polysaccharides containing Uc-glycol groups
.
Histochemie
11
,
71
80
.
Bradbury
,
S.
&
Stoward
,
P. J.
(
1968
).
The use of monofluorophenyl-hydrazine for the localization of periodate-reactive mucosaccharides and polysaccharides in the electron microscope
.
In Electron Microscopy
, vol.
2
(ed.
D. S.
Bocciarelli
), pp.
61
62
.
Rome
:
Tipografía Poliglotta Vaticana
.
Brownell
,
B.
,
Trevor-Hughes
,
J.
,
Goldby
,
F. S.
&
Woods
,
H. F.
(
1969
).
McArdle’s myopathy: A report of a case with observations on the muscle ultrastructure
.
J. neurol. Sci
.
9
,
515
526
.
De Bruyn
,
W.C.
(
1968
).
A modified OsO4 — (double) fixation which selectively contrasts glycogen
.
In Electron Microscopy
, vol.
2
(ed.
D. S.
Bocciarelli
), pp.
65
66
.
Rome
:
Tipografía Poliglotta Vaticana
.
Delwaide
,
P. J.
,
Reznik
,
M.
,
Lemaire
,
R.
,
Lelièvre
,
P.
&
Bonnet
,
F.
(
1967
).
A propos d’une nouvelle observation d’absence de phosphorylase dans le muscle strié
.
Revue neurol
.
116
,
119
140
.
Drochmans
,
P.
(
1962
).
Morphologie du glycogène: Etude au microscope électronique de colorations négatives du glycogène particulaire
.
J. Ultrastruct. Res
.
6
,
141
163
.
Fawcett
,
D. W.
(
1955
).
Observations on the cytology and electron microscopy of hepatic cells
.
J. natn. Cancer Inst
.
15
,
1475
1503
.
Gomori
,
G.
(
1952
).
Microscopic Histochemistry. Principles and Practice
.
Chicago
:
University of Chicago Press
.
Hanker
,
J. S.
,
Seaman
,
A. R.
,
Weiss
,
L. P.
,
Ueno
,
H.
,
Bergman
,
R. A.
&
Seligman
,
A. M.
(
1964
).
Osmiophilic reagents: A new cytochemical principle for light and electron microscopy
.
Science, N.Y
.
146
,
1039
1043
.
Hanker
,
J.
,
Weiss
,
L.
,
Dmochowski
,
H.
,
Katzoff
,
L.
&
Seligman
,
A.
(
1965
).
A new cytochemical principle for light and electron microscopy of functional groups of macromolecules with reagents having an osmiophilic moiety
.
J. Histochem. Cytochem
.
13
,
3
.
Leloir
,
L. F.
&
Goldemberg
,
S. H.
(
1960
).
Synthesis of glycogen from uridine diphosphate glucose in the liver
.
J. biol. Chem
.
235
,
919
923
.
Luck
,
D. J.
(
1961
).
Glycogen synthesis from uridine diphosphate glucose. The distribution of the enzyme in liver cell fractions
.
J. biophys. biochem. Cytol
.
10
,
195
209
.
Marinozzi
,
V.
(
1961
).
Silver impregnation of ultrathin sections for electron microscopy
.
J. biophys. biochem. Cytol
.
9
,
121
133
.
Marinozzi
,
V.
(
1963
).
The role of fixation in electron staining
.
JI R. microsc. Soc
.
81
,
141
154
.
Mcardle
,
B.
(
1951
).
Myopathy due to a defect in muscle glycogen breakdown
.
Clin. Sci
.
10
,
13
33
.
Merriam
,
R. W.
(
1958
).
The contribution of lower oxides of osmium to the density of biological specimens in electron microscopy
.
J. biophys. biochem. Cytol
.
4
,
579
582
.
Millonig
,
G.
(
1961
).
Advantages of a phosphate buffer for OsO4 solutions in fixation
.
J. appl. Phys
.
32
,
1637
.
Mommaerts
,
W. F. H. M.
,
Illingworth
,
B.
,
Pearson
,
C. M.
,
Guillory
,
R. J.
&
Seraydarian
,
K.
(
1959
).
A functional disorder of muscle associated with the absence of phosphorylase
.
Proc. natn. Acad. Sci. U.S.A
.
45
,
791
797
.
Movat
,
H. Z.
(
1961
).
Silver impregnation methods for electron microscopy
.
Am. J. clin. Path
.
35
.
528
537
.
Nilsson
,
O.
(
1962
).
Electron microscopy of glandular epithelium in the human uterus
.
J. Ultrastruct. Res
.
6
,
422
431
.
Pearson
,
C. M.
,
Rimer
,
D. G.
&
Mommaerts
,
W. F. H. M.
(
1959
).
Defect in muscle phosphorylase: A newly defined human disease
.
Clin. Res
.
7
,
298
.
Perry
,
M. M.
(
1967
).
Identification of glycogen in thin sections of amphibian embryos
.
J. Cell Sci
.
2
,
257
264
.
Price
,
H. M.
,
Howes
,
E. J. JR.
,
Blumberg
,
J. M.
&
Pearson
,
C. M.
(
1965
).
An improved biopsy technique for light and electron microscopic studies of human skeletal muscle
.
Lab. Invest
.
14
,
194
199
.
Rambourg
,
A.
&
Leblond
,
C. P.
(
1967
).
Electron microscopic observations on the carbohydraterich cell coat present at the surface of cells in the rat
.
J. Cell Biol
.
32
,
27
53
.
Revel
,
J. P.
(
1964
).
Electron microscopy of glycogen
.
J. Histochem. Cytochem
.
12
,
104
114
.
Revel
,
J. P.
,
Napolitano
,
L.
&
Fawcett
,
D. W.
(
1960
).
Identification of glycogen in electron micrographs of thin tissue sections
.
J. biophys. biochem. Cytol
.
8
,
575
589
.
Reynolds
,
E. S.
(
1963
).
The use of lead citrate at high pH as an electron-opaque stain in electron microscopy
.
J. Cell Biol
.
17
,
208
212
.
Rosa
,
F.
&
Johnson
,
F. B.
(
1967
).
Enzymatic digestion in the cytochemical demonstration of glycogen
.
J. Histochem. Cytochem
.
15
,
14
20
.
Schmid
,
R.
&
Mahler
,
R.
(
1959
).
Chronic progressive myopathy with myoglobinuria: Demonstration of a glycogenolytic defect in the muscle
.
J. clin. Invest
.
38
,
2044
2058
.
Schotland
,
D. L.
,
Spiro
,
D.
,
Rowland
,
L. P.
&
Carmel
,
P.
(
1965
).
Ultrastructural studies of muscle in McArdle’s disease (deficiency of muscle phosphorylase)
.
J. Neuropath, exp. Neurol
.
24
,
629
644
.
Seligman
,
A. M.
(
1966
).
New methods in electron microscopical cytochemistry
.
J. Histochem. Cytochem
.
14
,
745
746
.
Seligman
,
A. M.
,
Hanker
,
J. S.
,
Wasserkrieg
,
H.
,
Dmochowski
,
H.
&
Katzoff
,
L.
(
1965
).
Histochemical demonstration of some oxidized macromolecules with thiocarbohydrazide (TCH) or thiosemicarbazide (TSC) and osmium tetroxide
.
J. Histochem. Cytochem
.
13
,
629
639
.
Steiner
,
J. W.
&
Carruthers
,
J. S.
(
1961
).
Studies on the fine structure of the terminal branches of the biliary tree
.
Am. J. Path
.
38
,
639
661
.
Suzuki
,
T.
&
Sekiyama
,
S.
(
1961
).
Application of methenamine silver stain for electron microscopy
.
J. Electron Microsc., Tokyo
10
,
36
38
.
Swift
,
H.
&
Rasch
,
E.
(
1958
).
Localization of nucleic acids, proteins and polysaccharides with the electron microscope
.
Scient. Instrum. News
3
,
1
12
.
Thiéry
,
J. P.
(
1967
).
Mise en évidence des polysaccharides sur coupes fines en microscopie électronique
.
J. Microscopie
6
,
987
1018
.
Thiéry
,
J. P.
(
1969
).
Role de l’appareil de Golgi dans la synthèse des mucopolysaccharides; étude cytochimique
.
J. Microscopie
8
,
689
708
.
Vye
,
M. V.
(
1968
).
Ultrastructural study of two cases of McArdle’s disease
.
Lab. Invest
.
18
,
336
.
Vye
,
M. V.
&
Fischman
,
D. A.
(
1970
).
The morphologic alteration of particulate glycogen by en bloc staining with uranyl acetate
.
J. Ultrastruct. Res
.
33
,
278
291
.
Wanson
,
J. C.
&
Drochmans
,
P.
(
1968a
).
Rabbit skeletal muscle glycogen: A morphological and biochemical study of glycogen β-particles isolated by the precipitation-centrifugation method
.
J. Cell Biol
.
38
,
130
150
.
Wanson
,
J. C.
&
Drochmans
,
P.
(
1968b
).
Detection of phosphorylase with the electron microscope
.
In Control of Glycogen Metabolism
(ed.
W. J.
Whelan
), pp.
169
177
.
New York and London
:
Academic Press
.
Watson
,
M. L.
(
1958a
).
Staining of tissue sections for electron microscopy with heavy metals
.
J. biophys. biochem. Cytol
.
4
,
475
478
.
Watson
,
M. L.
(
1958b
).
Staining of tissue sections for electron microscopy with heavy metals. II. Application of solutions containing lead and barium
.
J. biophys. biochem. Cytol
.
4
,
727
730
.