A biochemical investigation into the inhibitory effects of glucosamine and N-acetylglucosamine on the aggregation in vitro of embryonic chick muscle cells

Sialic acid-containing glycoproteins occur in a number of animal tissues, characteristically as a membrane component (Cook, 1968; Spiro, 1970). Location of sialic acid moieties at the plasma membrane, in particular, had led to their suggested involvement in a number of surface-mediated biological functions, notably cell adhesion (Ambrose, 1966; Curtis, 1967). That sialic acid has a role in this phenomenon has been implied from investigations which showed a decreased adhesiveness of tissue cells to be a consequence of the specific enzymic removal of surface sialic acid (Weiss, 1961; Kemp, 1970) although an increased level of this compound at the cell periphery has also been held to account for the non-adhesiveness of some neoplastic cell strains (Forrester, Ambrose & Stoker, 1964). These findings led to the suggestion that the adhesive process may be impaired by levels of surface sialic acid outside a certain range (Kemp, 1970).

It was considered possible to examine this suggestion by testing the effect of exo genous glucosamine (2-amino-2-deoxy-D-glucose) on the aggregation of tissue cells. This hexosamine, which is incorporated into cell sialic acid (Molnar, Lutes & Winzler, 1965; Kohn, Winzler & Hoffman, 1962), could uncouple glycoprotein synthesis as an extension of its inhibitory effect on other aspects of metabolism (Bekesi & Winzler, 1969; Bekesi, Bekesi & Winzler, 1969) and thereby modify the chemistry of the cell surface. Indeed, Rizki (1961) found that a 5% (w/v) solution of glucosamine reduced the mutual adhesiveness of larval Drosophila lamellocytes and Garber (1963), using trypsin-dissociated embryonic chick cells, proposed that a derangement of the cell surface could account for the inhibited aggregation of glucosamine-treated cells. The latter study also reported N-acetylglucosamine (2-acetamido-2-deoxy-D-glucose) to be without effect on aggregation and it has therefore been employed in this present investigation as a control for the administration of glucosamine.

Previous investigations into the biochemical effects of exogenous glucosamine have indicated that this glucose analogue may interfere with both glycogen and glucose metabolism (Spiro, 1958; Kono & Quastel, 1962). Similarly, since glucosamine is incorporated into uridine diphosphate-(UDP) N-acetylglucosamine (a precursor of sialic acid) after the step by which this amino-sugar nucleotide controls its own formation by feedback inhibition (Kornfeld, Kornfeld, Neufeld & O’Brien, 1964), addition of exogenous glucosamine initiates an uncontrolled synthesis of UDP-N acetylglucosamine which proceeds with a concomitant depletion of the uridine and adenine nucleotide pools (Bekesi & Winzler, 1969). Since these effects could be expected to depress cellular energy levels, which of itself may influence the extent of aggregation, it was decided to examine this facet of the metabolic effects of glucosamine as a necessary prelude to investigating the glycoprotein metabolism of glucosamine treated cells.

It is the aim of this present study, then, to monitor the effect of glucosamine on some parameters of cellular energy production and to attempt to relate these to the state of cellular aggregation.

Cells were dissociated from the skeletal muscle of 9-day-old chick embryos by treatment for 10 min with a 0·25 % (w/v) solution of crude trypsin (Burroughs Wellcome, 1:300) at 37 °C, according to the standardized method employed in this laboratory (Kemp, Jones, Cunningham & James, 1967). The dissociated cells were washed in Hanks’s balanced salts solution and finally resuspended in Eagle’s minimal essential medium (MEM). The cell density of cells in the sus pension was estimated using a modified Fuchs–Rosenthal haemocytometer. Sterile conditions were maintained throughout the dissociation of the tissue and its subsequent culture.

Aggregation of the dissociated cells was quantitatNely estimated by the gyratory shaker method (Moscona, 1961). The cell suspension was placed into 25-ml Erlenmeyer flasks in aliquots of 3 ml at a concentration of 3·0 × 10⁵ cells/ml. D-Glucosamine hydrochloride or N-acetylglucosamine (Sigma Chemical Company) in Eagle’s MEM was then added to the flask contents to a final concentration of 0·023 M, control flasks receNing an equal volume (0·1 ml) of Eagle’s MEM instead of hexosamine. Resulting from the addition of glucosamine as its hydrochloride, the pH of the medium decreased and was therefore corrected with 0·1 ml of 4·4 % (w/v) sodium bicarbonate. This same volume of bicarbonate solution was dispensed to N-acetylglucosamine and control cell suspensions. Contents of all the flasks were then adjusted to pH 74 by equilibration against a s% CO₂/95 % air mixture (Paul, 1965) before being sealed and incubated at 37 °C on a gyratory shaker rotating at 70 rev/min. The process of cellular reaggregation was followed throughout the culture period by determining the dimensions of a given number of aggregates with a calibrated eyepiece micrometer. The resulting data were analysed by linear regression. Viability was examined by estimating the percentage of cells excluding particles of lissamine green dye (Kemp et al. 1967) in a 2% (w/v) aqueous solution.

Intracellular levels of adenine nucleotides, pyruvic acid and lactic acid were determined in an acid-soluble fraction isolated at given times from the cellular material by a standardized ex-traction procedure. It was essential for all steps to be performed rapidly and in the cold (4 °C) to minimize breakdown of the labile intermediates.

Contents of each culture flask were poured into a ground-glass homogenizer (Griffiths tube) pre-cooled to —15 °C and the cellular material was then centrifuged for 90s at 1000s. The supernatant was removed and retained for measurement of ‘extracellular’ lactate. The cell pellet was homogenized with 5 vol of 0·6 N perchloric acid and centrifuged at 2000g for 10 min. The acid-soluble extract was removed and the insoluble material triturated with 2·5 vol of 0·3 N perchloric acid. After centrifugation the 2 acid-soluble extracts were combined and brought to pH 6 by the cautious addition of 2 N KOH. These solutions were allowed to stand for at least 30 min to ensure complete precipitation of the insoluble potassium perchlorate. The clear extract was removed after centrifugation and adjusted to a known volume with distilled water before being stored frozen overnight.

After the frozen extracts had been allowed to reach the temperature of ice-water, intermediates were estimated enzymically by ultraviolet spectrophotometry, utilizing test-combinations supplied by C. F. Boehringer und Soehne GmbH (Mannheim, W. Germany). Adenosine triphosphate (ATP) was determined with glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.12) and phosphoglycerate kinase (EC 2.7.2.3) by the method of Biicher & Schuart (1964). Pyruvic acid, adenosine diphosphate (ADP) and adenosine monophosphate (AMP) were assayed in the same mixture (Adam, 1963) by the successNe additions of lactate dehydrogenase (EC 1.1.1.27), pyruvate kinase (EC 2.7.1.40) and myokinase (EC 2.7.4.3). Lactic acid in both the deproteinized sample and the suspension medium was assayed after the method of Scholz, Schmitz, Biicher & Lampen (1959) using lactate dehydrogenase. To obtain comparable results, care was taken throughout the experimental series to use a constant cell density since preliminary observations indicated that the levels of adenine nucleotides per unit number of cells decrease with increasing cell concentration. This phenomenon has also been reported for Ehrlich ascites cells (Bickis & Quastel, 1965).

The effect of exogenous hexosamines on the labelled carbon dioxide (¹⁴CO₂) released by cells rotated in Eagle’s MEM was estimated at intervals after the addition of 3 μCi of D-[U-¹⁴C]-glucose (specific actNity 16·7 μCi/mg) (Radiochemical Centre, Amersham) to the cell suspensions at zero time. The ¹⁴CO₂ was trapped in hydroxide of Hyamine (Rohm & Haas, Inc.) using a method of Snyder & Godfrey (1961) as modified by Dunn, Owen & Kemp (1970). The radio activity was measured to 1 % error in a toluene-based scintillation fluid (containing Fluoralloy dry mix TLA, Beckman Instruments Ltd.) with a Beckman LS 100 liquid scintillation spectrometer. The results were expressed as disintegrations per minute (dpm) after correction against an empirically determined quench curve. In view of the suggestion (Martin, 1967) that dilution of labelled glucose may occur on its entrance into a metabolic pool, thus causing a delay in the appearance of label in the carbon dioxide, it was decided to extend the observation period from 8 to 24 h.

Glycogen metabolism of the cells was estimated from the incorporation of labelled glucose into glycogen at given intervals after the addition of 3 μCi of D-[U-¹⁴C]glucose to the preparations. Glycogen was isolated and purified using a method developed by Goodridge (1968). Radio actNity accumulated in the glycogen was measured directly rather than, as suggested by Good ridge (1968), after its reduction to glucose. The glycogen was dissolved in 0·5 ml of distilled water and this solution solubilized in 12 ml of scintillation fluid by adding 1 ml of Beckman Bio-Solv Solubiliser, BBS 3 (Beckman Instruments Ltd.). Isotope actNity in the samples was then estimated by liquid scintillation spectrometry to an error of 1 %, the counts being conver ted to dpm after correcting for the experimentally determined efficiency.

Culture flasks were removed hourly from the gyratory shaker and the rate of oxygen uptake of the cell suspensions measured by a polarographic technique (Lessler, Malloy & Schwab, 1965). The linear decrease in the oxygen saturation of the suspension was recorded for 3 min at 37 °C using the oxygen sensor of a Biological Oxygen Monitor (Yellow Springs Instrument Co. Inc., California, U.S.A.) which was calibrated against air-saturated Eagle’s MEM.

Glucosamine at 0·023 M inhibited the aggregation of embryo chick muscle cells over an 8-h period (Fig. 1) for, when compared with the controls (Fig. 7), aggregates from glucosamine-treated cell suspensions were of a smaller size (Fig. 9). As evinced by the statistical analysis (Fig. 1), control aggregates also gained size more rapidly than those treated with this amino sugar, indicating that the inhibitory effect was continuously exerted throughout the 8-h experimental period. The N-acetyl analogue in equimolar concentration had a similar effect for muscle cell aggregation although the inhibition was not as pronounced as that exerted by glucosamine. At 8 h, glucosamine had effectively inhibited the size of aggregates by 18-9% and N-acetylglucosamine by 16·4%.

The viability of glucosamine-treated single cells at the end of the experimental period was 88 % of the controls, whereas that of cells incubated with N-acetylgluco samine was not significantly different.

Hourly measurements revealed a distinct oscillation in the level of ATP in embryonic muscle cells. Nevertheless, it will be seen (Fig. 2) that there was consistently more ATP in control cells than in cells incubated with either hexosamine. Of the 2 agents, glucosamine induced the greater depression of the level of this nucleotide and limited the range of its variability. Over the 8-h period there was a 36% (mean) loss of the ATP in glucosamine-treated cells as compared to the 27 % loss caused by its analogue. The hexosamines not only reduced cellular ATP but also total adenine nucleotides (Tables 1, 2). An average loss of 32% of the total ATP, ADP and AMP occurred during the 8-h incubation with glucosamine, an effect which is more pronounced than the 19% loss of total nucleotide caused by N-acetylglucosamine. The higher control value for cellular ATP was reflected by an ATP/AMP ratio greater than found in hexosamine-treated cells (Table 2). When considered over the 8-h period, the ATP/AMP ratio of control cells varied about a mean of 13-7. The administration of N-acetylglucosamine and glucosamine reduced this ratio to 8-75 and 8-o respectively, indicative of a greater degradation of ATP.

When cells were rotated for 8 h in the presence of N-acetylglucosamine or glucosamine, there was a resulting increase of the intracellular lactate/pyruvate ratios above the range determined for controls (Table 3). Although comparison of these data is made difficult by the periodic fluctuation of these metabolites between a considerable range of peak and basal levels, the mean values about which the ratios characteristically oscillated were 6:1, 8:1 and 10: 1 for controls,./V-acetylglucosamine- and glucosamine treated cell suspensions respectively. Correspondingly, the amount of extracellular lactic acid produced by glucosamine-treated cells (Fig. 3) was consistently higher than that produced by the controls or from cells incubated with N-acetylglucosamine. Glucosamine-treated cells secreted 229% more lactate into the medium than did the controls at 4 h and 99% more at 8 h. In contrast, although cells rotated in the presence of N-acetylglucosamine lost 85 % more lactic acid than the controls at 4 h, this trend was reversed at 8 h, there being 25% less lactate in the medium of the treated cells than of the controls.

Both hexosamines inhibited the release of¹⁴CO₂ from cells rotated in Eagle’s MEM containing uniformly labelled glucose and the inhibition was manifest throughout the 24-h period studied (Fig. 4). At 6 h the difference between the inhibitory effects of glucosamine and N-acetylglucosamine was slight but by 8 h the inhibition exerted by glucosamine was the greater by 6%. When the amount of radioactive carbon dioxide released by the controls had doubled 4 h later, the disparity between the inhibitory effects of the two amino sugars was more accentuated; glucosamine in hibiting ¹⁴CO₂ output by 18% and N-acetylglucosamine by 9%.

When cells were rotated in vitro with ¹⁴C-labelled glucose, the radioactivity sub-sequently measured in isolated glycogen was maximal for all preparations at 4 h (Fig. 5). After this time, utilization of glycogen resulted in loss of radioactivity from this fraction. The addition of glucosamine to cell suspensions caused 42 % less labelled glucose to be incorporated into glycogen at 4 h; N-acetylglucosamine exerted a 30% inhibition at this time (Fig. 5). From about the 6-h stage, N-acetylglucosamine no longer appeared inhibitory because of the more rapid disappearance of radioactive glycogen from controls. At 12 h this amino sugar showed no apparent inhibitory effect on cellular glycogen metabolism, although the radioactive glycogen in cells treated with glucosamine was still 45 % lower than in the controls.

Glucosamine induced stimulation of oxygen uptake by cells (Fig. 6). This enhanced uptake was observed throughout the test period and by the end of the experiment, glucosamine-treated cells had utilized 59% more oxygen than had the controls. The N-acetyl amino sugar did not have a stimulatory effect; moreover, after the 5-h stage, N-acetylglucosamine imposed a slight inhibition which by 8 h had depressed the total oxygen uptake by 13%.

It may be concluded from this investigation that glucosamine is instrumental in inhibiting the aggregation in vitro of embryonic chick muscle cells rotated on a gyratory shaker. This finding endorses the results of a previous inquiry on trypsin-dissociated chick liver and retinal cells (Garber, 1963) which established that the aggregation of those cells was inhibited by glucosamine. The present results are not, however, consonant with Garber’s contention (1963) that the JV-acetyl analogue in equNalent concentration was without effect on aggregation since it had been found that this amino sugar is inhibitory for muscle cells (Fig. 1). A plausible interpretation of this difference is that muscle cells may utilize N-acetylglucosamine more efficiently than do lNer and retinal cells. Our preliminary observations indicate that this sugar is actively incor porated into macromolecular components of muscle cells whereas it had been shown (McGarrahan & Maley, 1962) that rat liver utilizes N-acetylglucosamine poorly compared with the same amount of glucosamine. Furthermore, mannosamine (2-amino-2-deoxy-D-mannose) is rapidly taken up by Sarcoma 180 ascites cells (Raisys & Winzler, 1970) but the acetylated analogue is almost totally excluded. Such variations in the ability of cells to metabolize glucosamine and N-acetylglucosamine may well be fundamental in explaining the different effects ascribed to the N-acetylhexosamine.

The slight toxicity of glucosamine to cells in the suspension was foreseen as this property of exogenous glucosamine has been noted with a number of tissues (Rubin, Springer & Hogue, 1954; Bekesi, Molnar & Winzler, 1969; Kornfeld & Ginsburg, 1966; Rizki, 1961). Pilot experiments indicated the excessive lethality of 0·035M (≃ 0·75%, w/v) glucosamine on muscle cells. Consequently, a 0·023 M (≃ 0·5, w/v) solution was employed which noticeably inhibited cell aggregation but with a mini-mized toxicity.

The present results indicate that both glucosamine and N-acetylglucosamine in-hibit the conversion of radioactive glucose to glycogen. This corroborates the findings of Spiro (1958), who demonstrated a similar effect in rat lNer slices, with glucosamine the stronger inhibitor. In contrast, Martin (1967) reported that this amino-sugar does not interfere with the metabolic transfer of glucose to glycogen in rat diaphragm in vitro. This contradiction may be resolved by reference to the work of Kono & Quastel (1962), who point towards the ratio of glucose to glucose analogue as having a definite bearing on the degree to which the analogues inhibit several enzymes of glycogen metabolism. This is to be expected since the amount of substrate relative to its inhibitory analogue inevitably circumscribes the extent of enzyme inhibition. In this present study a ratio of hexosamine to glucose of about 4:1 was employed. Spiro (1958) used hexosamines in concentrations up to 8 times greater than glucose but Martin (1967) used the sugars in equimolar concentration. These variations probably modulate the inhibitory effects of the 2 glucose analogues.

The limited formation of labelled glycogen that occurs in the presence of exogenous glucosamine, coupled with the increased production of extracellular lactic acid pre-sumably from stored glycogen, would suggest a 2-fold effect of glucosamine on glycogen metabolism. First, there is an apparent inhibition of the synthesis of glycogen from glucose and, secondly, an enhancement of glycogenolysis. It should be recalled that the glycogenolytic effect of both glucosamine and N-acetylglucosamine has been attributed to their stimulation of glycogen phosphorylase (EC 2.4.1.1) actNity (Kono & Quastel, 1962); the larger decrease in the glycogen content caused by gluco-samine being due to its additional activity of inhibiting the UDP-dependent enzyme, glycogen synthetase(UDP-glucose-glycogen glucosyl transferase, EC 2.4.1.11). The results of this investigation are consistent, then, with such an interpretation of the effects of glucosamine and N-acetylglucosamine on glycogen metabolism.

Since both amino-sugars decreased the ATP/AMP ratios, their stimulatory effect on phosphorylase actNity may be functionally explained on the basis that this enzyme of glycogen breakdown is itself directly stimulated by AMP (see Krebs & Fischer, 1962) as well as by a release of inhibition through loss of ATP (Maddaiah & Madsen, 1966). The adenine nucleotide balance that prevails in hexosamine-treated cells would similarly favour the further degradation of the hexose phosphate derived from glycogen catabolism since an increase of AMP relative to ATP may inhibit fructose 1,6-diphosphatase, EC 3.1.3.11 (Newsholme & Underwood, 1965) and overcome the ATP inhibition of phosphofructokinase, EC 2.7.1.11 (Passonneau & Lowry, 1962), thereby actNating glycolysis. The increased lactic acid found in the presence of gluco-samine and N-acetylglucosamine is in accord with this view, though an enhanced production could also be due in part to released inhibition of the pyruvate kinase (EC 2.7.1.40) reaction owing to a decrease in the level of ATP (Tanaka, Sue & Morimura, 1967).

It is noteworthy that since exogenous hexosamine is incorporated into UDP-N acetylglucosamine after the step by which this nucleotide-sugar regulates its own formation by feedback inhibition (Kornfeld et al. 1964), administration of glucosamine to muscle cells may cause the uncontrolled synthesis of UDP-N-acetylglucosamine with a corresponding depletion of uridine triphosphate (UTP) (Bekesi & Winzler, 1969). It follows that this depletion would moderate glycogen synthesis since UTP is required for transfer of glucosyl moieties to the glycogen chain in the glycogen syn-thetase reaction. In this connexion, Zancan & Hers (1965) showed that glucosamine in equivalent concentration to glucose inhibited the formation of the glycogen precursor, UDP-glucose (UDPG), by 30-40%, which could also release a UDPG inhibition of phosphorylase (Maddaiah & Madsen, 1966). This interpretation would therefore provide a rationale for glucosamine inhibiting glycogen synthetase (Kono & Quastel, (1962), although an inhibition of hexokinase (EC 2.7.1.1) by the amino-sugars (Spiro, 1958) may also be expected to contribute to the decreased synthesis of glycogen from glucose.

The observed inhibition of ¹⁴CO₂ production from glucose in hexosamine-treated cells could result from competition between the hexosamines and labelled glucose for the hexokinase reaction (Spiro, 1958). Support for this view can be drawn from findings (Kono & Quastel, 1962) that both glucosamine and N-acetylglucosamine inhibited the hexokinase of rat liver slices when analogue and glucose were utilized in the 4:1 ratio of this present study. The inhibition of the release of ¹⁴CO₂ is insufficient to explain the more manifest effects that these amino sugars have on other aspects of cellular

metabolism which is therefore in keeping with the proposal that hexokinase is not the site at which glucose analogues inhibit glycogen metabolism (Kono & Quastel, 1962). The present results show that both glucosamine and N-acetylglucosamine depress the cellular ATP level, lending support to earlier reports that cells incubated with glucosamine and low glucose possess a reduced intracellular ATP pool (Creaser & Scholefield, 1960; Bekesi & Winzler, 1969). An explanation for this effect may be found in the postulated excessNe production of UDP-N-acetylglucosamine in cells incubated with hexosamine. Exogenous hexosamine must be phosphorylated prior to its further metabolism and this could then be expected to cause a coincident fall in the level of ATP. The diminished levels of total adenine nucleotides in hexosamine-treated cells probably also arises from this sequestration of ATP in the formation of hexosa mine phosphates - a synthesis which does not contribute to energy metabolism. The decreased ATP/AMP ratio of hexosamine-treated cells may be attributed to the con tinued phosphorylation of these amino sugars. A resulting excess of AMP would consequently lead to its irreversible deamination (Imai, Riley & Berne, 1964) and thereby effect a net loss from the adenine nucleotide pool which is otherwise too great to have derNed from the limited cytotoxicity of glucosamine. The oscillation in the level of cellular ATP may reflect successive activity peaks (see Mitchison, 1969) of certain enzymes in a cell system induced into synchrony by the initial trypsinization procedure (Stubblefield, 1968).

It seems reasonable to assume that the increased glycolytic flux resulting from a glucosamine-stimulated glycogenolysis would, as well as elevating the production of lactic acid, increase the amount of substrate oxidized by way of the Krebs cycle. This view is reinforced by the present finding that glucosamine stimulates oxygen uptake of chick muscle cells. It is significant in this respect that the decreased glycogenolysis of N-acetylglucosamine-treated cells, as judged by the diminishing level of lactic acid in the medium after the 5-h stage, coincides with their lower oxygen uptake.

In brief, the metabolic effects of hexosamines in this present investigation may be attributed to their alteration of the adenine nucleotide balance, which is causal in depressing cellular energy metabolism. These results also indicate that the inhibited aggregation of cells rotated with glucosamine or N-acetylglucosamine may be corre-lated with this reduced energy state of the cells. That the adhesive process depends upon the involvement of an active metabolism has been proposed by a number of investigators studying the effects of metabolic inhibitors and low temperature on aggregation (Giudice, 1965; Humphreys & Uehara, 1964; Kemp et al. 1967; Stein berg, 1962). The inhibitory effect of exogenous hexosamines, whatever its derivation, would seem to have sequelae for several aspects of cellular intermediary metabolism (Kono & Quastel, 1962; Bekesi et al. 1969). On this basis, it is considered that the inhibitory effects of glucosamine and N-acetylglucosamine could encompass glyco protein metabolism at which particular level these hexosamines may impair the adhesive process. The glycoprotein metabolism of hexosamine-treated chick muscle cells is currently under study.

We wish to thank Mr David Williams for processing the photomicrographs. This study was supported by generous grants provided by the Science Research Council.