GLIA 4:417-423 (1991)
Axon-Glia Interactions in the Cra-yfish: Glial Cell Oxygen Consumption Is Tightly Coupled to Axon Metabolism PA'L T. HAFtGITTAI AND EDWARD M. LIEBERMAN
Department of Ph siology, School of Medicine, East Carolina University, dreenuille, North Carolina 27858-4354
KEY WORDS
Carboxyatractyloside,Metabolism
ABSTRACT Oxygen consumption (Q0J of single isolated axons and their associated glial cell sheath was investigated under a variety of conditions to determine the contribution of each cell type to whole tissue Qo,. It was found that the Qo, of the sheath, in the absence of a functional axon, represented approximately 30%of the total tissue Qp,. When the axon was injected with carboxyatractyloside, an inhibitor of mitochondria1 oxidativephosphorylation that is membrane impermeant, electrophysiologicalproperties of the axon were not affected and glial sheath respiratory activity was stimulated by 1.7 to 2.7 times the untreated control level. These results suggest that glial cell metabolic activity is regulated by the metabolic activity of the axon. Depending o p the experimental conditions the glial sheath accounts for 30% to nearly 100% of the Qo, of axon-glial cell tissue. On the basis of these and morphometric measurements we estimate that in a normally functioning axon-glial cell system the glial sheath accounts for 90%of the tissue Qo,.
INTRODUCTION
cells in the axon-glial cell system has a major effect on the metabolic activity of the other and thereby contribIn a previous study from this laboratory (Lieberman utes to the discrepancy in the JNa/-P ratio noted above. et al., 1990)the stoichiometry of Naf transport (JNa) to In this paper, we report on experiments designed to high energy phosphate consumption (Q+) in isolated evaluate the individual contributionsof axons and their axon-gliapreparations from the crayfish was estimated associated glia t o total tissue oxygen consumption (QO,) to be approximately an order of magnitude greater than and factors that may alter that relationship. the generally accepted value of 3/1 (Caldwell, 1960). ratio is not unprecedented Although a high J,$-P (Klahr and Bricker, 1965),the magnitude of the discrepMATERIALS AND METHODS ancy requires that interactions between axons and their Nerve Preparation associated glial cells be explored further. All experiments were performed on medial giant axAxons and glia are known to interact with each other in a variety of ways, including electrophysiologic re- ons of the crayfish, Procambarus clarkii, using a sponses of the glia to action potential generation in the method of isolation modified from Wallin (1967). The axon (Lieberman and Hassan, 1988; Lieberman et al., ventral nerve cord was removed from the animal, placed 1989;Orkand et al., 1966),to the stimulation of ion and in a Lucite chamber,and desheathed of its perineurium. substrate uptake by the glia in response to changes in For electrophysiological experiments, the medial giant extracellular [K'l (Liebermanand Hassan, 1988;Salem axon was partially isolated from surrounding small et al., 19751, and to mutual transfer of macromolecular nerve fibers leaving the axon within the nerve cord for substances (Grossfeld et al., 1988; Lasek et al., 1977; Gainer et al., 1977).It is, therefore, possible that metaReceived August 12,1990; accepted November 7,1990. bolic activity (by itself, or by its products) of one of the Address reprint requests to Edward M. Lieberman at the address given above. 0 1991 Wiley-Liss, Inc.
418
HARGITTAI AND LIEBERMAN
Experimental Treatments mechanical support. For the oxygen consumption experiments, the axon with its surrounding glial sheath was Three types of axon-glialcell preparations were used. completely isolated from the nerve cord for a length of approximately 1-1.5 cm and tied at each end with fine 1.For the control preparation, the isolated axon-glial silk ligatures. Details of the method have been previcell preparation was maintained at physiological condiously published (Hargittai et al., 1987). tions throughout. 2. For the glial sheath preparation, the isolated axon Electrophysiology was perfused internally with the crayfish physiological solutionthat is normally used to superfuse it externally. Microelectrodeswith a tip resistance between 10 and This procedure washes out the axoplasm and destroys 40 M a were used to impale the giant axon for monitor- all metabolic and electrophysiologicalcapability of the ing of resting membrane and action potentials. The axon. This preparation is equivalent to a split axon electrophysiological responses were monitored with a preparation with the exception that the attendant glial WPI model 750 high impedance preamplifier, oscillo- cell destruction that occurs on cutting open the axon is scope, and strip chart recorder. Action potentials were avoided (Coehlo et al., 1960). Under these circumgenerated at the cephalic end of the nerve cord by stances, the glial sheath should be physiologically and externally applied electrical stimuli. Axons were used metabolically intact but isolated from its relationship only if they had resting membrane potentials 3 -80 with a functional axon. The Qo, of this preparation mV and action potentials 2 110mV and were capable of represents the basal metabolic activity of the glia in propagating the entire length of the cephalic and tho- isolation from its axon. racic nerve cord. 3. For the metabolically inhibited but electrically excitable axon, the axon was microinjected with a membrane-impermeant mitochondrial poison that inhibits Solutions oxidative phosphorylation. The glial cells were not directly affected by this treatment since the inhibitor was All axons were isolated and superfused in a modified confined to the cytoplasm of the axon (Stubbs, 1979; Van Harreveld (1936) crayfish physiological solution Verity et al., 1983).This was verified in control experi(PS)containing(in mM): 190 NaC1,5.4 KC1,13.5 CaCl,, ments in which intact isolated axons were superfused 2.6 MgCl,, and 20 Tris buffer, pH 7.4. The artificial with M carboxyatractyloside, which is 10-1,000 intracellular perfusion fluid used for internal injection times greater than usually used for complete inhibition of carboxyatractyloside (Sigma)contained (in mM): 140 of mitochondrial respiration. Under these conditions K2S04,15 NaC1,50 sucrose,and 10 Tris-HC1,pH 7.4. All neither excitability properties of the axon nor respirasolutions used were 420 to 430 m0sm. Experiments tion of the axon-glial cell preparation were affected for were performed at room temperature (21-23°C). periods of several hours (data not shown). The Qo of For experiments in which the [K+lowas greater than this preparation represents the metabolic activity of the the normal 5.4 mM, KC1 was substituted for Tris-HC1or glial sheath cells when functionallycoupled to the axon. conversely Tris-HC1replaced KC1 to make low K+ solutions. Metabolic poisons [cyanide (2 mM), ouabain (2 mM), or carboxyatractyloside M)] were RESULTS added directly to the superfusate or intracellular perfuInternally Perfused Axon-Glial Cell Preparation sion fluid and the pH adjusted to 7.4 with Tris-HC1or Tris-base as required. To differentiate between the contribution of axons and glial cells to the intact tissue O2consumption, the Qo, was measured before and after internally perfusing Oxygen Consumption of Isolated Axons the axon with normal crayfish physiological solution, a and Glial Cells procedure that effectively abolishes all electrophysioThe technique developed in this laboratory for the logical and metabolic activity of the axon. The results of study of oxygen consumption of single isolated crayfish these experiments (Table 1)demonstrate that the glial giant axons and their associated glial cells has been sheath, representing approximately 5% of the. total previously described (Hargittai et al., 1987). The preparation volume, accounted for 29% of the Qo, of method is based on the principle of oxygen quenching of the intact preparation. Both the intact and internally pyrene fluorescence. The 0, probe-pyrene, dissolved perfused axon preparations showed the expected sensiin oil and placed in a 200 pm diameter SpectraPor tivity to cyanide (CN), which inhibited respiration by hollow dialysis tube-was inserted into a quartz capil- approximately 80%. Ouabain inhibited Qo, of the intact lary with the isolated axon and viewed with an epifluo- axon preparation by approximately 30% in one set of rescence microscope. The change in fluorescence, as O2 experiments (Table 1) and 20% in the second set of was removed from the solution by the axon-glial cell experiments (Table2). For all experiments in this series preparation, was monitored with a photomultiplier the average inhibition was 27% and was statistically system. significant at the P > .05 level. The Qo, of both the
419
AXON-GLIA INTERACTIONS TABLE 1. Oxygen consumption (&J of intact and internally perfused axon-glial cell preparations a Qoz(mol02/1 tissue x min) x
Cell preparation Intact Perfused
PS
429 127
** 21** 105 (9) (11)
10-6
Ouabain
CN
301 f 88* (9) 112 f 14** (11)
76 f 23* (9) 26 4*3* (11)
*
"All values given as Qo2i S.E.M.Numbers in parenthesesrepresentthenumber of axons used for the determinations. CN, cyanide; PS, physiological solution. *Statistically different (Student's t test, paired samples) from its intact control at P < .05. *'Statistically different (Student's t test, unpaired samples) from the intact group treated similarly relative to inhibitor treatment at P < .05.
TABLE 2. The effect of carboxyatractyloside {CATR) on oxygen consumption of axon-glial cell preparationsa
Cell preparation Intact Vehicle-injected CATR-injected
Qo, (rno102/1 tissue x min) x 10-6 PS
Ouabain
CN
303 55 (4) 188 f 33** (3) 507 f 80** (6)
251 f 54 (4) 151 f 32* (3) 537 f 225** (6)
31 f 19* (4) 40 f 37* (3) 101 60* (6)
*
*
*Allvalues given as go2* S.E.M.Numbers in parentheses representthe number of axons used for the determinations. CN, cyanide: PS, physiological solution. *Statistically different (Student'st test, paired samples) from its intact control at P < .05. **Statisticallydifferent (Student's t test, unpaired samples) from the intact group treated similarly relative to inhibitor treatment at P < .05.
control and ouabain-treated axons were also sensitive to external [K+l increasing with decreasing external [K+] (Table 3). This finding is similar to the results obtained by measurements of Q+ (Lieberman et al., 1990). These results suggest that Qo, is an accurate, quantitative measure of energy consumption for both transport and non-transport related activities of cells. Effect of Carboxyatractyloside on Axons and Glia
As discussed in Materials and Methods, externally applied carboxyatractyloside does not affect electrophysiological function of intact crayfish axons when the agent is applied externally or injected internally. It is also necessary to show that the substance is an effective inhibitor of crayfish nerve tissue mitochondrial respiration. Cell free homogenates of crayfish nerve tissue and suspensions of crayfish nerve mitochondria were used for these control experiments. The method described by Lemasters and Hackenbrock (1980), modified for use with crayfish nerve tissue, was used for preparation of mitochondria.A Yellow Springs Instrument-Clark electrode O2probe system and temperature-controlledincubation chamber were used. We tested the effectiveness of carboxyatractylosideand CN on both state 4 and state 3 respiration of the mitochondria. State 4 respiration, O2 uptake in the absence of ADP, was achieved with 5 mM succinate as substrate added to a reaction mixture containing (in mM); 220 KCI, 0.5 EGTA, 5 MgCl,, 7.5 phosphate buffer, and 25 Hepes buffer, pH 7.4. State 3 respiration was induced by adding 1 mM ADP to mitochondria operating in state 4.
TABLE 3. The effectof extracellular potassium on Qo,of the crayfish giant axon-glial cell preparationa
Cell preparation
[K'I, (mM) 0.1
5.4
13.5
wQo,
516 f 113 (4) 140 f 20
453 f 102 (5) 100
463 f 135 (3) 85 i 6
Qoz %
344 f 99 (3) 128 5 8
383 f 122 (4) 100 [89 f 91
719 (1)
Co,ntrol Ouabain
100
* S.E.M.Averages and S.E.M.are for all determinations (unpaired samples). % of the &, + S.E.M. of the preparation aQo2 is given in (mol 02/1tissue X min) X
under experimental conditions are referenced to the Qo, in normal (5.4 mM K+) crayfish physiological solution (100%)for both the control and ouabaintreated preparations. Numhers in brackets show the ?6 effect of ouabain on the Qoz of the axon-glial cell preparation. All percentage values determined from paired experiments. Numbers in parenthesesare the numbers of axonsused for the determinations.
In six individual determinations on three different mitochondrial preparations carboxyatractyloside M), which blocks the adenine nucleotide translocase of the mitochondrial membrane, inhibited state 4 and state 3 respiration by 47.1 6.2 and 64.5 +- 5.3% (mean ? S.E.M.),respectively. CN, in all cases, reduced respiration to the baseline level (reactionmedium without substrate and mitochondria) for an effective 100% inhibition of respiration.
*
Electrically Excitable, Respiration-Inhibited Axon Preparation
To define further the extent of interaction between axons and their associated glial cells, a series of experiments were designed to maintain selectively all functional properties of glial cells and the normal electrophysiological properties of the axon but to inhibit axonal respiration. Under these circumstances glial cells should maintain their relationship to an axon performing most of its normal functions, although at the expense of its sizable pool of -P, for up to 2 h (Lieberman et al., 1990). To accomplishthe goal of selectiveinhibition of axonal respiration we microinjected carboxyatractyloside into M was the axoplasm so that a final concentrationof reached in the intracellular space of the axon. Mitochondrial poisons such as cyanide, dinitrophenol, or rotenone are membrane-permeable substances that could not be restricted to one cell even with direct injection of these substances into the cytoplasm of the axon. On the other hand, carboxyatractyloside,a plant-derived mitochondrial poison (Vignais et al., 19731, is a membraneimpermeant substance specific for the adenine nucleotide translocase and has been shown to be an effective inhibitor at concentrations as low as lo-? M. For these experiments two controls were used, the normal, untreated axon preparation and a preparation in which only the vehicle for the carboxyatractyloside (artificial intracellular fluid) was injected into the axon. The results shown in Table 2 demonstrate that cannu-
420
HARGITTAI AND LIEBERMAN
lation of the .axon and injection of artificial axoplasm reduced the Qo, to 62%of the uninjected control preparation. On the other.hand, carboxyatractyloside injection stimulated the Qo to 270%of the injected control (167%of the intact, untreated control). CN caused the expected 80-90% inhibition of respiration in all three of the. preparations described above. Ouabain sensitivity of Qo, of the carboxyatractyloside-treated Ereparation was lost, suggestingthat the increase in the Qo, was due to an increase in ouabain-insensitivehigh energy phosphate use by glia.
Morphometric Analysis of Axon-Glial Cell Membrane Surface Areas and Mitochondria Distributions Fig. 1 shows an electron micrograph of a crayfish axon and its glial sheath. For 36 such electron micrographs analyses of the relative surface areas of the glial sheath cells to axon surface and the distribution of mitochondria have been made. The purpose of this analysis was to correlate the Qo, of the axon-glial cell preparation with the morphology of the cells in order to estimate more closely the individual contribution of glia and axons to the whole tissue Qo,. Morphometric data were collected from six nerve cords and six electron micrographs of each cord at a print magnification of 25,000 X , Parameters were calculated using the methods and formalisms described by Weibel (1979). For the comparison of glia and axon surface areas, surface densities (total plasma membrane area per unit glia volume) were calculated and were 13.8 1.3 and 1.2 -t- 0.1 (mean S.E.M.), respectively. Based on this analysis, membrane surfacearea of glial cells that make up the sheath was nearly 11 times greater than the area of axon membrane with which the glial sheath is associated. These results suggest that energy utilization of the glia (at least for ion transport) may represent up to 90% of the energy consumption of the intact system. The number and/or volume of mitochondria in a cell usually are taken to indicate the ability of the cell to generate or replenish high energy phosphate. To measure this, the volume density (percentage of total volume of the cell represented by the mitochondria)in glia and in the giant axon were calculated to be 3.6%and 0.6%,a ratio of 6. The results suggest, as found for direct measurements of QO2, that glia should be extremely active users of 02.
*
*
DISCUSSION In a recently published investigation of crayfish nerve tissue (Lieberman et al., 1990) the estimate of the -P required to transport Na+ (JNa)against its electrochemical gradient was approximately an order of magnitude higher (JNa/-P = 30) than previously reported for nerve tissue from other species. Since we were unable to
identify errors in measurement that could account for more than 20-30% of the discrepancy it was necessary to explore the possibility that interactions between axons and their associated glial cells modified energy requirements of ion transport. Oxygen Consumption of Axons and Glial Cells Approximately 30%of the oxygen consumption of an isolated intact axon glial cell preparation remained when the axon was internally perfused with normal crayfish physiological solution. A similar Qo, distribution was found for the split squid nerve fiber preparation (Coelho et al., 1969). The high Ca2+and Na+ and the low K+concentrationsof the artificial extracellular solution (perfusion fluid) effectively abolished all metabolic activity of the axon and any possibility of normal selective membrane permeability, membrane potential maintenance, or electrical excitability of the axon. The results of this treatment were compared with those of experimentsin which the axon membrane potential and excitability were maintained but axonal ATP production was abolished by the membrane-impermeant poison, carboxyatractyloside. Following carboxyatractyloside treatment the oxygen consumption of glial cells was stimulated to nearly 270%of the vehicle-injected intact control or 170% of the untreated intact axon-glial cell preparation. The interactive relationship between the axon and its associated glial cells makes it difficult to dissociate the contribution that glial cells make to the total Qo, of the resting (steady-state) intact axon-glial cell preparation. Depending o n the metabolic and functional condition of the axon the Qo, of the glial cells could represent 30%to nearly 100%of the intact tissue-extraordinarily high values when it is recognized that the glial cell layer volume represents only 5%of the total tissue volume. Although an exact value for the contribution of the axon to the intact tissue Qo, cannot be given,we suggest that 10% is a reasonable estimate consistent with the Qo, and morphometric measurements presented here. The estimate is based on the difference in Qo, between the vehicle-injected, intact preparation (Table 2) and Qo, remaining after internal perfusion of the axon (Table l),which are 62 and 29%,respectively, giving a difference of 33% versus their .control values. In the first case the 62%represents the Qo, of the tissue remaining following mechanical manipulation and injection of artificial axoplasm. Since the axon is electrically functional and still sensitive to both ouabain and CN it is assumed that most of the change in respiration is quantitative and not qualitative in nature, affecting both axon and glia similarly.Further change in respiration followinginternal perfusion of the axon must be due to the loss of Qo,of the axon. From this estimate the glia contribute approximately 70% of the tissue Qo,. This value represents a conservative estimate in light of the significant metabolic interaction that exjsts between the axon and its glial sheath. Since glial Qo, increases
AXON-GLIA INTERACTIONS
Fig. 1. Electronmicrogra h ofthe crayfish medialgiant axon (MGA) with its associated adaxona! (Gad)and intermediate (GJ glial sheaths in crosssection.An isolated ‘antaxon- lial cell preparation, as utilized in this study, would inclufe the mefial giant axon, adaxonal, and intermediate glial layer. Note the extensive infoldings of the adaxonal glial plasma membrane (M,) associated with the axolemma or axon
421
lasma membrane (Ma) and several alternating la ers of membrane[mited glial cytoplasm and connective tissue &TI. Mitochondria (MITO) are present in both the axon and the several layers of lial cytoplasm. LGA, lateral giant axon. Bar in the lower left corner ofthe micrograph represents 1 Fm.
422
HARGITTAI AND LIEBERMAN
of the previous investigation (Lieberman.et al., 1990) cannot be explained on the basis of -P or Qo, measurement error. As suggested in the previous investigation, an explanation based on differences in the nature of the membrane steady-state during Qwpmeasurementsversus those of electrophysiological conductance measurements that, by definition, perturb the steady-state seems more likely. It can be concluded that the metabolic activity of axons and that of its associated glia have a close, possibly obligatory, interactive relationship, as has been previously found for their interactions in regulation of the periaxonal ionic environment (Lieberman and Hassan, 1988). Metabolically, glia may account for 90% or more of the oxygen consumption of the intact untreated axon-glial sheath preparation although the glial layer represents only 5% of the tissue volume. In addition, the glia respond to an electrically active but metabolically compromised axon by increasing their own metabolic activity (QO,) to nearly 3 x the control level and conversely decrease their activity when isolated from the axon. It may be speculatedthat metabolic products produced and released into the periaxonal Oxygen Requirement for -P Production space by a metabolically compromised axon is a stimulus to the glial cell to increase in its own activity as a The original stimulus for this investigation was to compensatory homeostatic mechanism. Conversely, determine the amount of oxygen consumed to replenish glial cells isolated from functional axons and therefore the high energy phosphate used by the ouabain-sensi- their metabolic signals, reduce their activity to a just tive transport system of the nerve membrane. It is self-maintenancelevel. generally accepted that for each mole of atomic oxygen utilized by mitochondria1oxidative phosphorylation approximately 3 mol of ATP are generated, the exact ratio ACKNOWLEDGMENTS being dependent on the substrate used and its point of entry into the metabolic cycle (Lemasters, 1984). Thus, We would like to thank Dr. Bob Grossfeldfor his many the -PI0 ratio provides an additional control on the useful editorial suggestions during the writing of this accuracy of the nerve Q-p measurements made in the manuscript and Ms. Paulette Hahn for her technical previous investigation (Lieberman et al., 1990), in par- help with data collection, morphometrics, and electron ticular, whether the Q-p measurements reported ear- microscopy. lier represent a good estimate of the -P consumed in This research was supported, in part, by a grant from cellular processes. the Army Research Office DAAG29-82-K-0182. For 17 paired experiments, Qo,of untreated control axons was 278 +- 33 and four ouabain-treated axons REFERENCES 210 t 35 x mol O2/I tissue x min. The two groups are statistically different (P < .05). Therefore the ouaD.G. and Lieberman, E.M. (1988) Studies of axon-glial cell bain-sensitive Qo, of intact axon-glial cell preparations Brunder, interactions and eriaxonal K' homeostasis: I. The influence ofNa', mol 0 2 A tissue x min, respectively. was 68.1 x K', C1- and chofinergic agents on the membrane potential of the adaxonal glia of the crayfish medial giant axon. Neuroscience, Using 10%of the ouabain-sensitiveportion of the mean 25~951-959. of these values to represent Qo, of the axon (63 x l o p 6 Caldwell, P.C. (1960) The phosphorus metabolism of s uid axons and mol OzA tissue x min) and converting from a volume to its relationship to the active transport of sodium. J. P\ysiol. (Lond.), 152:545-560. a surface area normalization (Lieberman et al., 1990), R.R., Goodman, J.W., and Bowers, M.B. (1960) Chemical the Qo, of the axon is 0.56 x 10-l' mol 02/cm2x s. The Coelho, studies of the satellite cells of the squid giant nerve fiber. Exp. Cell Q-p of the axon under the same conditions is Res., 2O:l-11. H., Tasaki, I., and Lasek, R.J. (1977) Evidence for the glia 2.6 ~ . 1 0 - 'moLkm2 ~ x s (Lieberman et al., 1990). The Gainer, neuron transfer h othesis from intracellular perfusion studies of Q-.P/QO, ratio, at 5.4 mM K+O,is 4.6,or, in more ususquid 'ant axon. Y C e l l Biol., 74:524-530. al terms, the -P/O ratio is 2.3. This value is well within Grossfelg R.M., Hinge, M.A., Lieberman, E.M., and Stewart, L.C. (1988) Axon-glia transfer of a protein and carbohydrate. Glia, the range expected for oxidative phosphorylation and 1:292-300. reaffirms that our previous measurements of -P utili- Hassan, S. and Lieberman, E.M. (1988) Studies of axon-glial cell interactions and eriaxonal K+ homeostasis: 11.The effect of axonal zation of the axon under a variety of conditions are not stimulation, chohergic agents and transport inhibitors on the in serious error. The l o x greater than expected transresistance in series with the axon membrane. Neuroscience, port of Na+ per -P calculated from our measurements 25~961-969.
dramatically when the axon is metabolically inhibited but electrophysiologically functional, it would not .be unreasonable to expect a dramatic decrease in glial Qo, when the glia are dissociated from a normally functioning axon. In this case a glial Qo, contribution of 70% would underestimate the contribution of the glial cell to total tissue Qo, under normal conditions. Finally, morphometric studies of the crayfish medial giant axon-glialcell layer showed that the surface area of the glial cells in the sheath is approximately 11times greater than the surface area of the associated axon. If it can be assumed that the transport ATPase system of the glial cell and axon are similar in their transport rates and -P utilization per unit surface area (Brunder and Lieberman, 1988;Hassan and Lieberman, 1988;Lieberman and Hassan, 1988) then the disparity in surface area reinforces the probability that the glial contribution to tissue Qo, is much greater than that of the axon. This is also consistent with the fact that the volume fraction of mitochondria in the glial layer is also several times greater than that of the axon (6 x 1.
AXON-GLIA INTERACTIONS Hargittai, P.T., Ginty, D.D., and Lieberman, E.M. (1987) A pyrene fluorescence technique for measurement of oxygen consumption of single isolated axons. Anal. Biochem., 163:418-426. Klahr, S. and Bricker, N.S. (1965) Ener etics of anaerobic sodium transport by the fresh water turtle %ladder. J. Gen. Physiol., 48:571-580. Lasek, R.J., Gainer, H., and Barker, J.L. (1977)Cell-to-celltransfer of glial proteins of the s uid 'ant axon. J. Cell Biol., 74501-523. Lemasters, J.J. (1984) e A P-to-oxygen stoichiometriesof oxidative phosphorylation by rat liver mitochondria. J. Biol. Chem., 259:13123-13130. Lemasters, J.J. and Hackenbrock, C.R. (1980)The energized state of rat liver mitochondria. J.Bwl. Chem., 25535674-5680. Lieberman, E.M. and Hassan, S. (1988) Studies of axon-glial cell ' homeostasis: 111. The effect of anisosinteractions and eriaxonal K motic media antpotassium on the relationship between the resistance in series with the axon membrane and glial cell volume. Neuroscience, 25:971-981. Lieberman, E.M., Abbott, N.J., and Hassan, S. (1989) Evidence that glutamate mediates axon-Schwann cell signaling in the squid. Glia, 2:94-102. Liebeman, E.M., Pascarella, J., Brunder, D., and Hargittai, P.T. (1990) Effect of extracellular potassium on ouabain-sensitive con-
'8-1
423
sumption of high energy phos hate by crayfish 'ant axons: A stud of the energy requirement kr transport in t8e steady-state. J! Neurochem., 55:155-164. Orkand, R.K., Nicholls, J.G., and Kuffler, S.W. (1966) Physiological roperties of glial cells in the central nervous system of amphibia. J. geuro hysiol., 29:788-806 Salem, i D . , HammerschlagiR., Bracho, H., and Orkand, R.K. (1975) Influence of potassium ions on accumulation and metabolism of ['4Clglucose by glial cells. Brain Res., 86:499-503. Stubbs, M. (1979) Inhibitors of the adenine nucleotide translocase. Pharmacol. Ther. 7:329-349. Van Harreveld, A. (1936) Physiological saline for freshwater crustaExp. Biol. Med., 34:428-432. ceans. Proc. SOC. Veriiy, M.A., Brown, W.J., and Cheung, M.K. (1983)Failure of atractyloside to inhibit synaptosomalmitochondria1energy transduction. Neurochem. Res., 8:159-166. Vignais, P.V., Vignais, P.M., and Defaye, G. (1973)Adenosine diphosphate translocase in mitochondria. Nature of the receptor site for carboxyatractyloside(Gummiferin).Biochemistry, 12:1508-1519. Wallin, B.G. (1967) Intracellular ion concentration in single crayfish axons.Acta Physiol. Scand., 70:419430. Weibel, E.R. (1979)StereologicalMethods, Vol. 1:PractdcalMethodsfor BiologicalMorphology. Academic Press, London.
Axon-Glia Interactions in the Cra-yfish: Glial Cell Oxygen Consumption Is Tightly Coupled to Axon Metabolism PA'L T. HAFtGITTAI AND EDWARD M. LIEBERMAN
Department of Ph siology, School of Medicine, East Carolina University, dreenuille, North Carolina 27858-4354
KEY WORDS
Carboxyatractyloside,Metabolism
ABSTRACT Oxygen consumption (Q0J of single isolated axons and their associated glial cell sheath was investigated under a variety of conditions to determine the contribution of each cell type to whole tissue Qo,. It was found that the Qo, of the sheath, in the absence of a functional axon, represented approximately 30%of the total tissue Qp,. When the axon was injected with carboxyatractyloside, an inhibitor of mitochondria1 oxidativephosphorylation that is membrane impermeant, electrophysiologicalproperties of the axon were not affected and glial sheath respiratory activity was stimulated by 1.7 to 2.7 times the untreated control level. These results suggest that glial cell metabolic activity is regulated by the metabolic activity of the axon. Depending o p the experimental conditions the glial sheath accounts for 30% to nearly 100% of the Qo, of axon-glial cell tissue. On the basis of these and morphometric measurements we estimate that in a normally functioning axon-glial cell system the glial sheath accounts for 90%of the tissue Qo,.
INTRODUCTION
cells in the axon-glial cell system has a major effect on the metabolic activity of the other and thereby contribIn a previous study from this laboratory (Lieberman utes to the discrepancy in the JNa/-P ratio noted above. et al., 1990)the stoichiometry of Naf transport (JNa) to In this paper, we report on experiments designed to high energy phosphate consumption (Q+) in isolated evaluate the individual contributionsof axons and their axon-gliapreparations from the crayfish was estimated associated glia t o total tissue oxygen consumption (QO,) to be approximately an order of magnitude greater than and factors that may alter that relationship. the generally accepted value of 3/1 (Caldwell, 1960). ratio is not unprecedented Although a high J,$-P (Klahr and Bricker, 1965),the magnitude of the discrepMATERIALS AND METHODS ancy requires that interactions between axons and their Nerve Preparation associated glial cells be explored further. All experiments were performed on medial giant axAxons and glia are known to interact with each other in a variety of ways, including electrophysiologic re- ons of the crayfish, Procambarus clarkii, using a sponses of the glia to action potential generation in the method of isolation modified from Wallin (1967). The axon (Lieberman and Hassan, 1988; Lieberman et al., ventral nerve cord was removed from the animal, placed 1989;Orkand et al., 1966),to the stimulation of ion and in a Lucite chamber,and desheathed of its perineurium. substrate uptake by the glia in response to changes in For electrophysiological experiments, the medial giant extracellular [K'l (Liebermanand Hassan, 1988;Salem axon was partially isolated from surrounding small et al., 19751, and to mutual transfer of macromolecular nerve fibers leaving the axon within the nerve cord for substances (Grossfeld et al., 1988; Lasek et al., 1977; Gainer et al., 1977).It is, therefore, possible that metaReceived August 12,1990; accepted November 7,1990. bolic activity (by itself, or by its products) of one of the Address reprint requests to Edward M. Lieberman at the address given above. 0 1991 Wiley-Liss, Inc.
418
HARGITTAI AND LIEBERMAN
Experimental Treatments mechanical support. For the oxygen consumption experiments, the axon with its surrounding glial sheath was Three types of axon-glialcell preparations were used. completely isolated from the nerve cord for a length of approximately 1-1.5 cm and tied at each end with fine 1.For the control preparation, the isolated axon-glial silk ligatures. Details of the method have been previcell preparation was maintained at physiological condiously published (Hargittai et al., 1987). tions throughout. 2. For the glial sheath preparation, the isolated axon Electrophysiology was perfused internally with the crayfish physiological solutionthat is normally used to superfuse it externally. Microelectrodeswith a tip resistance between 10 and This procedure washes out the axoplasm and destroys 40 M a were used to impale the giant axon for monitor- all metabolic and electrophysiologicalcapability of the ing of resting membrane and action potentials. The axon. This preparation is equivalent to a split axon electrophysiological responses were monitored with a preparation with the exception that the attendant glial WPI model 750 high impedance preamplifier, oscillo- cell destruction that occurs on cutting open the axon is scope, and strip chart recorder. Action potentials were avoided (Coehlo et al., 1960). Under these circumgenerated at the cephalic end of the nerve cord by stances, the glial sheath should be physiologically and externally applied electrical stimuli. Axons were used metabolically intact but isolated from its relationship only if they had resting membrane potentials 3 -80 with a functional axon. The Qo, of this preparation mV and action potentials 2 110mV and were capable of represents the basal metabolic activity of the glia in propagating the entire length of the cephalic and tho- isolation from its axon. racic nerve cord. 3. For the metabolically inhibited but electrically excitable axon, the axon was microinjected with a membrane-impermeant mitochondrial poison that inhibits Solutions oxidative phosphorylation. The glial cells were not directly affected by this treatment since the inhibitor was All axons were isolated and superfused in a modified confined to the cytoplasm of the axon (Stubbs, 1979; Van Harreveld (1936) crayfish physiological solution Verity et al., 1983).This was verified in control experi(PS)containing(in mM): 190 NaC1,5.4 KC1,13.5 CaCl,, ments in which intact isolated axons were superfused 2.6 MgCl,, and 20 Tris buffer, pH 7.4. The artificial with M carboxyatractyloside, which is 10-1,000 intracellular perfusion fluid used for internal injection times greater than usually used for complete inhibition of carboxyatractyloside (Sigma)contained (in mM): 140 of mitochondrial respiration. Under these conditions K2S04,15 NaC1,50 sucrose,and 10 Tris-HC1,pH 7.4. All neither excitability properties of the axon nor respirasolutions used were 420 to 430 m0sm. Experiments tion of the axon-glial cell preparation were affected for were performed at room temperature (21-23°C). periods of several hours (data not shown). The Qo of For experiments in which the [K+lowas greater than this preparation represents the metabolic activity of the the normal 5.4 mM, KC1 was substituted for Tris-HC1or glial sheath cells when functionallycoupled to the axon. conversely Tris-HC1replaced KC1 to make low K+ solutions. Metabolic poisons [cyanide (2 mM), ouabain (2 mM), or carboxyatractyloside M)] were RESULTS added directly to the superfusate or intracellular perfuInternally Perfused Axon-Glial Cell Preparation sion fluid and the pH adjusted to 7.4 with Tris-HC1or Tris-base as required. To differentiate between the contribution of axons and glial cells to the intact tissue O2consumption, the Qo, was measured before and after internally perfusing Oxygen Consumption of Isolated Axons the axon with normal crayfish physiological solution, a and Glial Cells procedure that effectively abolishes all electrophysioThe technique developed in this laboratory for the logical and metabolic activity of the axon. The results of study of oxygen consumption of single isolated crayfish these experiments (Table 1)demonstrate that the glial giant axons and their associated glial cells has been sheath, representing approximately 5% of the. total previously described (Hargittai et al., 1987). The preparation volume, accounted for 29% of the Qo, of method is based on the principle of oxygen quenching of the intact preparation. Both the intact and internally pyrene fluorescence. The 0, probe-pyrene, dissolved perfused axon preparations showed the expected sensiin oil and placed in a 200 pm diameter SpectraPor tivity to cyanide (CN), which inhibited respiration by hollow dialysis tube-was inserted into a quartz capil- approximately 80%. Ouabain inhibited Qo, of the intact lary with the isolated axon and viewed with an epifluo- axon preparation by approximately 30% in one set of rescence microscope. The change in fluorescence, as O2 experiments (Table 1) and 20% in the second set of was removed from the solution by the axon-glial cell experiments (Table2). For all experiments in this series preparation, was monitored with a photomultiplier the average inhibition was 27% and was statistically system. significant at the P > .05 level. The Qo, of both the
419
AXON-GLIA INTERACTIONS TABLE 1. Oxygen consumption (&J of intact and internally perfused axon-glial cell preparations a Qoz(mol02/1 tissue x min) x
Cell preparation Intact Perfused
PS
429 127
** 21** 105 (9) (11)
10-6
Ouabain
CN
301 f 88* (9) 112 f 14** (11)
76 f 23* (9) 26 4*3* (11)
*
"All values given as Qo2i S.E.M.Numbers in parenthesesrepresentthenumber of axons used for the determinations. CN, cyanide; PS, physiological solution. *Statistically different (Student's t test, paired samples) from its intact control at P < .05. *'Statistically different (Student's t test, unpaired samples) from the intact group treated similarly relative to inhibitor treatment at P < .05.
TABLE 2. The effect of carboxyatractyloside {CATR) on oxygen consumption of axon-glial cell preparationsa
Cell preparation Intact Vehicle-injected CATR-injected
Qo, (rno102/1 tissue x min) x 10-6 PS
Ouabain
CN
303 55 (4) 188 f 33** (3) 507 f 80** (6)
251 f 54 (4) 151 f 32* (3) 537 f 225** (6)
31 f 19* (4) 40 f 37* (3) 101 60* (6)
*
*
*Allvalues given as go2* S.E.M.Numbers in parentheses representthe number of axons used for the determinations. CN, cyanide: PS, physiological solution. *Statistically different (Student'st test, paired samples) from its intact control at P < .05. **Statisticallydifferent (Student's t test, unpaired samples) from the intact group treated similarly relative to inhibitor treatment at P < .05.
control and ouabain-treated axons were also sensitive to external [K+l increasing with decreasing external [K+] (Table 3). This finding is similar to the results obtained by measurements of Q+ (Lieberman et al., 1990). These results suggest that Qo, is an accurate, quantitative measure of energy consumption for both transport and non-transport related activities of cells. Effect of Carboxyatractyloside on Axons and Glia
As discussed in Materials and Methods, externally applied carboxyatractyloside does not affect electrophysiological function of intact crayfish axons when the agent is applied externally or injected internally. It is also necessary to show that the substance is an effective inhibitor of crayfish nerve tissue mitochondrial respiration. Cell free homogenates of crayfish nerve tissue and suspensions of crayfish nerve mitochondria were used for these control experiments. The method described by Lemasters and Hackenbrock (1980), modified for use with crayfish nerve tissue, was used for preparation of mitochondria.A Yellow Springs Instrument-Clark electrode O2probe system and temperature-controlledincubation chamber were used. We tested the effectiveness of carboxyatractylosideand CN on both state 4 and state 3 respiration of the mitochondria. State 4 respiration, O2 uptake in the absence of ADP, was achieved with 5 mM succinate as substrate added to a reaction mixture containing (in mM); 220 KCI, 0.5 EGTA, 5 MgCl,, 7.5 phosphate buffer, and 25 Hepes buffer, pH 7.4. State 3 respiration was induced by adding 1 mM ADP to mitochondria operating in state 4.
TABLE 3. The effectof extracellular potassium on Qo,of the crayfish giant axon-glial cell preparationa
Cell preparation
[K'I, (mM) 0.1
5.4
13.5
wQo,
516 f 113 (4) 140 f 20
453 f 102 (5) 100
463 f 135 (3) 85 i 6
Qoz %
344 f 99 (3) 128 5 8
383 f 122 (4) 100 [89 f 91
719 (1)
Co,ntrol Ouabain
100
* S.E.M.Averages and S.E.M.are for all determinations (unpaired samples). % of the &, + S.E.M. of the preparation aQo2 is given in (mol 02/1tissue X min) X
under experimental conditions are referenced to the Qo, in normal (5.4 mM K+) crayfish physiological solution (100%)for both the control and ouabaintreated preparations. Numhers in brackets show the ?6 effect of ouabain on the Qoz of the axon-glial cell preparation. All percentage values determined from paired experiments. Numbers in parenthesesare the numbers of axonsused for the determinations.
In six individual determinations on three different mitochondrial preparations carboxyatractyloside M), which blocks the adenine nucleotide translocase of the mitochondrial membrane, inhibited state 4 and state 3 respiration by 47.1 6.2 and 64.5 +- 5.3% (mean ? S.E.M.),respectively. CN, in all cases, reduced respiration to the baseline level (reactionmedium without substrate and mitochondria) for an effective 100% inhibition of respiration.
*
Electrically Excitable, Respiration-Inhibited Axon Preparation
To define further the extent of interaction between axons and their associated glial cells, a series of experiments were designed to maintain selectively all functional properties of glial cells and the normal electrophysiological properties of the axon but to inhibit axonal respiration. Under these circumstances glial cells should maintain their relationship to an axon performing most of its normal functions, although at the expense of its sizable pool of -P, for up to 2 h (Lieberman et al., 1990). To accomplishthe goal of selectiveinhibition of axonal respiration we microinjected carboxyatractyloside into M was the axoplasm so that a final concentrationof reached in the intracellular space of the axon. Mitochondrial poisons such as cyanide, dinitrophenol, or rotenone are membrane-permeable substances that could not be restricted to one cell even with direct injection of these substances into the cytoplasm of the axon. On the other hand, carboxyatractyloside,a plant-derived mitochondrial poison (Vignais et al., 19731, is a membraneimpermeant substance specific for the adenine nucleotide translocase and has been shown to be an effective inhibitor at concentrations as low as lo-? M. For these experiments two controls were used, the normal, untreated axon preparation and a preparation in which only the vehicle for the carboxyatractyloside (artificial intracellular fluid) was injected into the axon. The results shown in Table 2 demonstrate that cannu-
420
HARGITTAI AND LIEBERMAN
lation of the .axon and injection of artificial axoplasm reduced the Qo, to 62%of the uninjected control preparation. On the other.hand, carboxyatractyloside injection stimulated the Qo to 270%of the injected control (167%of the intact, untreated control). CN caused the expected 80-90% inhibition of respiration in all three of the. preparations described above. Ouabain sensitivity of Qo, of the carboxyatractyloside-treated Ereparation was lost, suggestingthat the increase in the Qo, was due to an increase in ouabain-insensitivehigh energy phosphate use by glia.
Morphometric Analysis of Axon-Glial Cell Membrane Surface Areas and Mitochondria Distributions Fig. 1 shows an electron micrograph of a crayfish axon and its glial sheath. For 36 such electron micrographs analyses of the relative surface areas of the glial sheath cells to axon surface and the distribution of mitochondria have been made. The purpose of this analysis was to correlate the Qo, of the axon-glial cell preparation with the morphology of the cells in order to estimate more closely the individual contribution of glia and axons to the whole tissue Qo,. Morphometric data were collected from six nerve cords and six electron micrographs of each cord at a print magnification of 25,000 X , Parameters were calculated using the methods and formalisms described by Weibel (1979). For the comparison of glia and axon surface areas, surface densities (total plasma membrane area per unit glia volume) were calculated and were 13.8 1.3 and 1.2 -t- 0.1 (mean S.E.M.), respectively. Based on this analysis, membrane surfacearea of glial cells that make up the sheath was nearly 11 times greater than the area of axon membrane with which the glial sheath is associated. These results suggest that energy utilization of the glia (at least for ion transport) may represent up to 90% of the energy consumption of the intact system. The number and/or volume of mitochondria in a cell usually are taken to indicate the ability of the cell to generate or replenish high energy phosphate. To measure this, the volume density (percentage of total volume of the cell represented by the mitochondria)in glia and in the giant axon were calculated to be 3.6%and 0.6%,a ratio of 6. The results suggest, as found for direct measurements of QO2, that glia should be extremely active users of 02.
*
*
DISCUSSION In a recently published investigation of crayfish nerve tissue (Lieberman et al., 1990) the estimate of the -P required to transport Na+ (JNa)against its electrochemical gradient was approximately an order of magnitude higher (JNa/-P = 30) than previously reported for nerve tissue from other species. Since we were unable to
identify errors in measurement that could account for more than 20-30% of the discrepancy it was necessary to explore the possibility that interactions between axons and their associated glial cells modified energy requirements of ion transport. Oxygen Consumption of Axons and Glial Cells Approximately 30%of the oxygen consumption of an isolated intact axon glial cell preparation remained when the axon was internally perfused with normal crayfish physiological solution. A similar Qo, distribution was found for the split squid nerve fiber preparation (Coelho et al., 1969). The high Ca2+and Na+ and the low K+concentrationsof the artificial extracellular solution (perfusion fluid) effectively abolished all metabolic activity of the axon and any possibility of normal selective membrane permeability, membrane potential maintenance, or electrical excitability of the axon. The results of this treatment were compared with those of experimentsin which the axon membrane potential and excitability were maintained but axonal ATP production was abolished by the membrane-impermeant poison, carboxyatractyloside. Following carboxyatractyloside treatment the oxygen consumption of glial cells was stimulated to nearly 270%of the vehicle-injected intact control or 170% of the untreated intact axon-glial cell preparation. The interactive relationship between the axon and its associated glial cells makes it difficult to dissociate the contribution that glial cells make to the total Qo, of the resting (steady-state) intact axon-glial cell preparation. Depending o n the metabolic and functional condition of the axon the Qo, of the glial cells could represent 30%to nearly 100%of the intact tissue-extraordinarily high values when it is recognized that the glial cell layer volume represents only 5%of the total tissue volume. Although an exact value for the contribution of the axon to the intact tissue Qo, cannot be given,we suggest that 10% is a reasonable estimate consistent with the Qo, and morphometric measurements presented here. The estimate is based on the difference in Qo, between the vehicle-injected, intact preparation (Table 2) and Qo, remaining after internal perfusion of the axon (Table l),which are 62 and 29%,respectively, giving a difference of 33% versus their .control values. In the first case the 62%represents the Qo, of the tissue remaining following mechanical manipulation and injection of artificial axoplasm. Since the axon is electrically functional and still sensitive to both ouabain and CN it is assumed that most of the change in respiration is quantitative and not qualitative in nature, affecting both axon and glia similarly.Further change in respiration followinginternal perfusion of the axon must be due to the loss of Qo,of the axon. From this estimate the glia contribute approximately 70% of the tissue Qo,. This value represents a conservative estimate in light of the significant metabolic interaction that exjsts between the axon and its glial sheath. Since glial Qo, increases
AXON-GLIA INTERACTIONS
Fig. 1. Electronmicrogra h ofthe crayfish medialgiant axon (MGA) with its associated adaxona! (Gad)and intermediate (GJ glial sheaths in crosssection.An isolated ‘antaxon- lial cell preparation, as utilized in this study, would inclufe the mefial giant axon, adaxonal, and intermediate glial layer. Note the extensive infoldings of the adaxonal glial plasma membrane (M,) associated with the axolemma or axon
421
lasma membrane (Ma) and several alternating la ers of membrane[mited glial cytoplasm and connective tissue &TI. Mitochondria (MITO) are present in both the axon and the several layers of lial cytoplasm. LGA, lateral giant axon. Bar in the lower left corner ofthe micrograph represents 1 Fm.
422
HARGITTAI AND LIEBERMAN
of the previous investigation (Lieberman.et al., 1990) cannot be explained on the basis of -P or Qo, measurement error. As suggested in the previous investigation, an explanation based on differences in the nature of the membrane steady-state during Qwpmeasurementsversus those of electrophysiological conductance measurements that, by definition, perturb the steady-state seems more likely. It can be concluded that the metabolic activity of axons and that of its associated glia have a close, possibly obligatory, interactive relationship, as has been previously found for their interactions in regulation of the periaxonal ionic environment (Lieberman and Hassan, 1988). Metabolically, glia may account for 90% or more of the oxygen consumption of the intact untreated axon-glial sheath preparation although the glial layer represents only 5% of the tissue volume. In addition, the glia respond to an electrically active but metabolically compromised axon by increasing their own metabolic activity (QO,) to nearly 3 x the control level and conversely decrease their activity when isolated from the axon. It may be speculatedthat metabolic products produced and released into the periaxonal Oxygen Requirement for -P Production space by a metabolically compromised axon is a stimulus to the glial cell to increase in its own activity as a The original stimulus for this investigation was to compensatory homeostatic mechanism. Conversely, determine the amount of oxygen consumed to replenish glial cells isolated from functional axons and therefore the high energy phosphate used by the ouabain-sensi- their metabolic signals, reduce their activity to a just tive transport system of the nerve membrane. It is self-maintenancelevel. generally accepted that for each mole of atomic oxygen utilized by mitochondria1oxidative phosphorylation approximately 3 mol of ATP are generated, the exact ratio ACKNOWLEDGMENTS being dependent on the substrate used and its point of entry into the metabolic cycle (Lemasters, 1984). Thus, We would like to thank Dr. Bob Grossfeldfor his many the -PI0 ratio provides an additional control on the useful editorial suggestions during the writing of this accuracy of the nerve Q-p measurements made in the manuscript and Ms. Paulette Hahn for her technical previous investigation (Lieberman et al., 1990), in par- help with data collection, morphometrics, and electron ticular, whether the Q-p measurements reported ear- microscopy. lier represent a good estimate of the -P consumed in This research was supported, in part, by a grant from cellular processes. the Army Research Office DAAG29-82-K-0182. For 17 paired experiments, Qo,of untreated control axons was 278 +- 33 and four ouabain-treated axons REFERENCES 210 t 35 x mol O2/I tissue x min. The two groups are statistically different (P < .05). Therefore the ouaD.G. and Lieberman, E.M. (1988) Studies of axon-glial cell bain-sensitive Qo, of intact axon-glial cell preparations Brunder, interactions and eriaxonal K' homeostasis: I. The influence ofNa', mol 0 2 A tissue x min, respectively. was 68.1 x K', C1- and chofinergic agents on the membrane potential of the adaxonal glia of the crayfish medial giant axon. Neuroscience, Using 10%of the ouabain-sensitiveportion of the mean 25~951-959. of these values to represent Qo, of the axon (63 x l o p 6 Caldwell, P.C. (1960) The phosphorus metabolism of s uid axons and mol OzA tissue x min) and converting from a volume to its relationship to the active transport of sodium. J. P\ysiol. (Lond.), 152:545-560. a surface area normalization (Lieberman et al., 1990), R.R., Goodman, J.W., and Bowers, M.B. (1960) Chemical the Qo, of the axon is 0.56 x 10-l' mol 02/cm2x s. The Coelho, studies of the satellite cells of the squid giant nerve fiber. Exp. Cell Q-p of the axon under the same conditions is Res., 2O:l-11. H., Tasaki, I., and Lasek, R.J. (1977) Evidence for the glia 2.6 ~ . 1 0 - 'moLkm2 ~ x s (Lieberman et al., 1990). The Gainer, neuron transfer h othesis from intracellular perfusion studies of Q-.P/QO, ratio, at 5.4 mM K+O,is 4.6,or, in more ususquid 'ant axon. Y C e l l Biol., 74:524-530. al terms, the -P/O ratio is 2.3. This value is well within Grossfelg R.M., Hinge, M.A., Lieberman, E.M., and Stewart, L.C. (1988) Axon-glia transfer of a protein and carbohydrate. Glia, the range expected for oxidative phosphorylation and 1:292-300. reaffirms that our previous measurements of -P utili- Hassan, S. and Lieberman, E.M. (1988) Studies of axon-glial cell interactions and eriaxonal K+ homeostasis: 11.The effect of axonal zation of the axon under a variety of conditions are not stimulation, chohergic agents and transport inhibitors on the in serious error. The l o x greater than expected transresistance in series with the axon membrane. Neuroscience, port of Na+ per -P calculated from our measurements 25~961-969.
dramatically when the axon is metabolically inhibited but electrophysiologically functional, it would not .be unreasonable to expect a dramatic decrease in glial Qo, when the glia are dissociated from a normally functioning axon. In this case a glial Qo, contribution of 70% would underestimate the contribution of the glial cell to total tissue Qo, under normal conditions. Finally, morphometric studies of the crayfish medial giant axon-glialcell layer showed that the surface area of the glial cells in the sheath is approximately 11times greater than the surface area of the associated axon. If it can be assumed that the transport ATPase system of the glial cell and axon are similar in their transport rates and -P utilization per unit surface area (Brunder and Lieberman, 1988;Hassan and Lieberman, 1988;Lieberman and Hassan, 1988) then the disparity in surface area reinforces the probability that the glial contribution to tissue Qo, is much greater than that of the axon. This is also consistent with the fact that the volume fraction of mitochondria in the glial layer is also several times greater than that of the axon (6 x 1.
AXON-GLIA INTERACTIONS Hargittai, P.T., Ginty, D.D., and Lieberman, E.M. (1987) A pyrene fluorescence technique for measurement of oxygen consumption of single isolated axons. Anal. Biochem., 163:418-426. Klahr, S. and Bricker, N.S. (1965) Ener etics of anaerobic sodium transport by the fresh water turtle %ladder. J. Gen. Physiol., 48:571-580. Lasek, R.J., Gainer, H., and Barker, J.L. (1977)Cell-to-celltransfer of glial proteins of the s uid 'ant axon. J. Cell Biol., 74501-523. Lemasters, J.J. (1984) e A P-to-oxygen stoichiometriesof oxidative phosphorylation by rat liver mitochondria. J. Biol. Chem., 259:13123-13130. Lemasters, J.J. and Hackenbrock, C.R. (1980)The energized state of rat liver mitochondria. J.Bwl. Chem., 25535674-5680. Lieberman, E.M. and Hassan, S. (1988) Studies of axon-glial cell ' homeostasis: 111. The effect of anisosinteractions and eriaxonal K motic media antpotassium on the relationship between the resistance in series with the axon membrane and glial cell volume. Neuroscience, 25:971-981. Lieberman, E.M., Abbott, N.J., and Hassan, S. (1989) Evidence that glutamate mediates axon-Schwann cell signaling in the squid. Glia, 2:94-102. Liebeman, E.M., Pascarella, J., Brunder, D., and Hargittai, P.T. (1990) Effect of extracellular potassium on ouabain-sensitive con-
'8-1
423
sumption of high energy phos hate by crayfish 'ant axons: A stud of the energy requirement kr transport in t8e steady-state. J! Neurochem., 55:155-164. Orkand, R.K., Nicholls, J.G., and Kuffler, S.W. (1966) Physiological roperties of glial cells in the central nervous system of amphibia. J. geuro hysiol., 29:788-806 Salem, i D . , HammerschlagiR., Bracho, H., and Orkand, R.K. (1975) Influence of potassium ions on accumulation and metabolism of ['4Clglucose by glial cells. Brain Res., 86:499-503. Stubbs, M. (1979) Inhibitors of the adenine nucleotide translocase. Pharmacol. Ther. 7:329-349. Van Harreveld, A. (1936) Physiological saline for freshwater crustaExp. Biol. Med., 34:428-432. ceans. Proc. SOC. Veriiy, M.A., Brown, W.J., and Cheung, M.K. (1983)Failure of atractyloside to inhibit synaptosomalmitochondria1energy transduction. Neurochem. Res., 8:159-166. Vignais, P.V., Vignais, P.M., and Defaye, G. (1973)Adenosine diphosphate translocase in mitochondria. Nature of the receptor site for carboxyatractyloside(Gummiferin).Biochemistry, 12:1508-1519. Wallin, B.G. (1967) Intracellular ion concentration in single crayfish axons.Acta Physiol. Scand., 70:419430. Weibel, E.R. (1979)StereologicalMethods, Vol. 1:PractdcalMethodsfor BiologicalMorphology. Academic Press, London.
