Calcium-dependent regulation of potassium permeability in the glial perineurium (blood-brain barrier) of the crayfish

h’eurosrience Vol. 38. No. 1, pp. 175485, Printed in Great Britain

030&4522/90 S3.00 + 0.00 Pergamon Press plc 0 1990 IBRO

1990

CALCIUM-DEPENDENT REGULATION OF POTASSIUM PERMEABILITY IN THE GLIAL PERINEURIUM (BLOO~BRAIN BARRIER) OF THE CRAYFISH A. M. BUTT,* P. T. HARGITTAIand E. M. LIEBERM& Department

of Physiology,

School of Medicine, East Carolina University, Greenville, NC 27858, U.S.A.

Abstract-The physioio~~al basis of the high selective potassium permeability of the crayfish gliai ~rineu~um was studied. The transient spike-like perineuriai potential generated in high external &+I was used as a measure of barrier K+ permeability. The medial giant axon membrane potential was used to monitor interstitial [K+]. Perineurial current-voltage relations of the perineurium were used to measure electrical resistance and to determine changes in K+ conductance of the perineurial barrier. Of a range of cations studied only Rb+, in addition to K+ generated a large transient sheath potential. In some experiments “regenerative” multiple spikes were observed during the continued exposure of the perineurium to high [Rb+],. This degree of ion selectivity is typical of glial cell membranes and K channels. Barrier conductance increased only very briefly in Rb+; the potential falling rapidly to a steady S-10 mV. The PC,/P,, and the Pn /PK ratios at the peak transient potential were similar suggesting the liability site for these cations was the same. The permeability of Rb+ in the plateau phase was significantly lower than K+ suggesting that high [Rb+], may act to block K+ channels. The K+-selective permeability was reversibly blocked by extracellular Ba *+ at both the peak and the plateau phase, in a concentrationdependent manner. Other K-channel blocking agents, tetraethylammonium ions (10 mM), caesium ions (20 mM), and 3&diaminopyridine (0.5 mM) were ineffective. The effect of Ba*+ on the peak potential was similar to the removal of external Ca*+ or exposure to the Ca2+channel blockers, verapamii (IO-’ M) or La3+ (5 mM). The time- and ~n~entm~on~~ndent reversible biock of the K+ permeability of the perineurium was consistent with the known action of these agents on voltage-gated Ca2+ channels in nerve and glia. La’+ caused an irreversible decrease in perineurial conductance and K+ influx. Lanthanum titration of the negative charges of glial membranes and mucopolysaccharide matrix of the interce.llular space suggest they may be important factors in determining the magnitude of the perineurial leak and paracellular K+ permeability. Electron microscopic examination of La 3+ distribution demonstrated a diffusion barrier at the outer layer of perineurial glia. The binding of La’+ at the basolateral membranes of the glial barrier suggested this was the site at which La 3+ had its physiolo~~l actions. The results suggest that the increase in glial membrane K+ conductance in high (K+], was most likely due to voltage-gated Ca*+ and K+ channels and Ca2+-activated K+ channels of the membranes of

perineurial glia.

The crayfish perineurium is of comparative interest in terms of its role as a “blood-brain barrier” as it provides an opportunity to study extracellular ion regulation by a glial epithelium in vivo. The high degree of ion selectivity and regulation displayed by the crayfish perineurium is more typical of cell membranes or tight epitheha such as the vertebrate blood-brain barrier,’ rather than an apparently leaky epithelium with low electrical resistance and relatively high permeability.‘~” In the first paper in this series” the electrophysiological properties of the perineurium were described as it related to K+ permeability and ion transport regulation and was shown to be primarily exclusively permeable to potassium where the P,,jP, ratio is

*Present address: Biomedical Sciences Division, Kina’s College, University of London, Camden Hih Road, KensinKton. London W8 7AH. U.K. tTo who~co~es~ndence should he addressed. Abbreuiutims: 3,4-DAP, 3.4-diaminopyridine; EDTA, ethylenediaminetetra-acetate; R,, sheath resistance; TEA, tetraethylammonium.

approximately 0.1. Permeability and ion transport of the perineurium in elevated [K+Jo was largely a function of the high K+ permeability of the glial cell membranes, providing a route for rapid cellular transport of K+ away from areas of high concentration, coupled with low paracellular permeabiljty. In this investigation a variety of agents have been used that were expected to help differentiate between various ce!lular and paracellular pathways for ionic substance movement through the perineurium and further define the role of barrier permeability in interstitial ion regulation. The results of this study showed that the increase in glial ~~eability in high external (K+] was dependent on Ca2+ suggesting the presence of Ca*+ channels on the perineurial glial membranes. Calcium-dependent K+ channels may be important in modulating glial K+ permeability and the rapid transport of K+ through the perineurium. These po~ibiliti~ and their physiolo~cal role in interstitiai ion regulation were discussed. Preliminary accounts of portions of this work have been published in abstract forrn.8,9.22 175

A. M.

176 EXPERIMENTAL

BUTT et al.

PROCEDURES

The animal preparation and electrophysiological methods used in this investigation have been fully described in the first article in this series.” The previous investigation in this series demonstrated that the trans-perineurial sheath potential, measured with standard electrophysiological techniques, was highly dependent on K+ and could be used as an indicator of the high degree of selective K+ permeability of the glial perineurial barrier.‘,” Therefore, perineurial potential in 100 mM w+], was used to monitor glial permeability and to test the hypothesis that glial membranes are responsible for the ion selectivity. Current-voltage relations were used to measure perineurial resistance and served as an indicator of K+ permeability and conductance of the perineurial glial cell membranes. The change in [K+] in the adaxonal space adjacent to the medial giant axon was used to estimate rate constants for the bulk transport of K+ through the perineurium. Permeability ratios were estimated using the Moreton modification of the Constant Field equation as described in the previous paper in this series.” Solutions of RbCl were made by substituting for NaCI. Low calcium solutions were obtained by substituting with Tris, and solutions containing LaCI, or BaCl, were made by replacing equal concentrations of CaCl,. Other channel blockers, tetraethylammonium (TEA, Sigma), 3,4-diaminopyridine (3,4-DAP, Sigma) and verapamil (Sigma) were added directly to normal saline. All experiments were carried out at room temperature, 20-25’C. Following electrophysiological measurements of the effects of La)+, La3+ was precipitated, in place, with Na,SO, and the nerve cord prepared for electron microscopic examination. Nerve cords were fixed in 2% glutaraldehyde buffered with veronal acetate overnight at 4°C. and postfixed in 1% osmium tetroxide before dehydration and embedding in epoxy resin. Ultrathin sections were stained with lead citrate before examination under a transmission electron microscope.

RESULTS

In high [K+], a transient spike-like perineurial sheath potential was generated, K+ transport into the nerve cord interstitium was restricted and Na+ was effectively excluded suggesting a high degree of regulation of the selective permeability of sheath to a number of small ionic substances. It was, therefore. of interest to investigate the effects of agents known to influence K+ permeability to further characterize the mechanisms of interstitial ion regulation. Eflects of rubidium As demonstrated in the first study of this series” only Rb+, of all the cations tested for their ability to cause a significant perineurial sheath potential, was comparable with K+. In four paired-control experiments the early transient potential generated in IOOmM Rb,+ was similar in amplitude to the potential in 100 mM [K ‘I0 (34.5 rt 5 and 40.3 f 2.2 mV, respectively) but was of a considerably shorter duration, and fell rapidly to a potential signiBcantly lower (P < 0.05) than that in K+ [approximately IO and 22 mV, respectively (see Ref. 17)]. Multiple regenerative spike potentials were often observed with continued exposure to 1OOmM Rb,+ (Fig. IA) in experimental conditions.

The peak perineurial potential was markedly less sensitive to Rb’ below a concentration of 20mM. indicating a concentration-dependent increase In perineurial selectivity above 20-25mM. similar to that observed in K’ (Fig. IB).” immediately following the peak, Rb’ selectivity decreased rapidly toward the control level giving rise to a low steadystate potential in Rb+ which was only slightly concentration sensitive. In the case where multiple Rb+-induced potential spikes were generated there were also transient increases in the selectivity of the perineurial sheath to Rb’ (Fig. 2, inset). The highly selective mechanism for K’ and Rb’ transport suggests cellular membrane K’ channels are the site of high perineurial permeability. Steady-state Rb’ transport across the perineurium was slower than for K’ (Fig. lC,D), and interstitial rubidium concentration, [Rb+],s, rose only to 7.7 +0.5mM (n =4) after 2min in IOOmM Rb,‘. compared with 22 + 5 mM K& in 100 mM K,,’ (n = 4). indicating a lower steady-state permeability of the perineurial barrier to Rb’ in addition to the low selectivity during this phase of the sheath potential (Fig. 2). The change in sheath resistance (&,,) during the initial spike potential was the same in 100 mM Rb,’ and K,‘. in paired experiments, suggesting that the barrier was equally permeable to both ions during this period. The R,, recovered more rapidly in Rb+ than in K +, returning to the pre-experimental control R,, during the high Rb,’ pulse. At 90 s the R,, in Rbwas 237 + 19 Rem’ compared with the control of 255 + 27 or an 87% recovery (data not shown). In K +. recovery was 35% during the same period. The rapidity of recovery of the R, and permeability accounts for the shorter duration of the spike potential. the low steady-state potential level in high [Rb+l,,. and the limit on the accumulation of interstitial Rb’. Efects of K a-channel blockers A high degree of K+ selectivity is characteristic of membrane K + channels as is suggested to be present in the basolateral membranes of the crayfish perineurial glia. ” We used a variety of agents, namely, Rb’. caesium ions (Cs’). barium ions Ba’+, TEA ions. and 3.4-DAP. Between them, these agents are known to block most K’ channels.” When applied extracellularly, only Baz+ was observed to have any effect on the high K,* -induced perineurial sheath potential change (Fig. 3, top row). The effect of Ba?+ (substituting for Ca2+) was to reversibly reduce both the initial spike and steadystate sheath potential induced by 100 mM [K +I,-,.The effect of Ba*+ was concentrationdependent, and at 2 mM the IOOmM [K+10 potential was slightly reduced at the peak (IO-15% in three experiments). Increasing Ba*+ to 4.5 and 9mM caused a large decrease in the early transient potential whereas lowering [Ca: +],1by the same amount had no effect of

Glial Ca2+dependent

K+ transport

177

60s

Fig. 1. Effect of Rb,+ on sheath and axon potential and trans.-perineurial Rb+ transport. (A) Effect of [Rb+], on trans-perineurial voltage. Traces were from paired experiments in a single animal. [Rb+], was 100, 50 and 25 mM from top to bottom. The period of high Rb+ superfusion is shown by the bar below the traces in panel C. Above 25 mM [Rb+], a transient spike potential was observed which fell rapidly to a low steady-state potential. In 1OOmM [Rb+], two or more spontaneous spike potentials were sometimes observed with no further experimental manipulation. (B) Semilog plot of [Rb+], vs the peak (unfilled symbols) and steady-state (filled symbols) perineurial voltage. Points are means + S.E.M. (bars); where there are no bars, S.E.M. smaller than size of symbol. Results were from paired experiments in four animals. The peak potential was Nemstian above 20 mM but deviated sharply from the theoretical in lower concentrations. The steady-state potential increased only slightly with [Rb+],. (C) Effect of Rbi on the membrane potential of the axon in an intact sheath preparation. The traces shown here are from the same experiment shown in panel A above. The depolarization of the axon during high Rb+ exposure is correlated with the increase in Rb+ in the periaxonal space. (D) Rate constants for perineurial uptake of Rb+ (unfilled triangles) and K+ (unfilled circles) in relation to external ion concentration. Data for this analysis were obtained from experiments shown on the left where the change in axon membrane potential is converted to change in ion concentration at the axon surface. Results from eight animals. Uptake was significantly lower in Rb+ and was not obviously concentration-dependent compared with K+.

its own. In one experiment in which trans-perineurial transport of K+ was measured under several concentrations of Ba*+ there was no evidence that Ba*+ affected steady-state trans-perineurial K+ transport. The K+ transport rate constants were 0.36,0.48,0.52, 0.49 and 0.54 x 10e3/s for 0 Ba’+ (control), 4.5 mM Ba2+, 9mM Ba2+, 13.5mM Ba2+ and recovery, respectively. The results suggest that the initial rapid development of a high selectivity and permeability for K+ was delayed but not blocked in Ba*+, while bulk transport through the perineurium was not greatly affected. Caesium (20mM), TEA (I-1OmM) and 3,4-DAP (0.5 mM) did not affect the perineurial sheath poten-

tial or the rate constants for K+ transport. The membrane permeant 3,4-DAP was observed to prolong the action potential of the medial giant axon in the sheath-intact preparation, as previously observed in the de-sheathed preparation.*’ Caesium and TEA had no obvious effect on the medial giant axon in either a sheath intact’ or de-sheathed preparation, while Ba*+ caused a slow depolarization of the axon. E&cts

of [Ca*+],

Barium, in the experiments presented here, replaced an equal amount of Ca2+, and under these conditions Ba*+ substitution of Ca*+ may affect Ca*+ conductance3 and fluxes.3’ To determine if Ca2+ is

A. M. BUTTef al.

178

Fig. 2. E&t of K$ and Rb,+ on perineurial sheath permeability and selectivity. In high q perineuriai selectivity for K.+ increases rapidly and is anon for the period of high external fK+] exposure. In high Rb+ a similar initial selectivity change occurs but the response is transient with steady-state selectivity for Rb+ recovering to nearly the control level. Results shown are from four paired-control experiments. The untIlled fquan represents the average minimum PEI/Pabratio obtained for all experiments (n = 7). The inset shows the pattern of selectivity change that occurs during repetitive spiking that often occurs during exposure to high Rb,+ . At the peak of each potential spike

the ratio falls to approximately the same value (0.1) even though the peak potential may be smaller (Fig. 1A, top trace).

important for perineurial K’ selectivity, the direct effects of reducing superfusate [Ca*+) were tested. Exposure of the ventral nerve cord to O.l-O.01 the normal [Cat*], for 30min si~ifi~ntiy reduced the initial spike potential in both IOOmM K,f (Fig. 4A) and Rb,+ (not shown). When external Ca” was completely removed and the chelating agent EDTA (1 mM) added, sheath selectivity was irreversibly and completely lost. For this reason the use of Ca*+ chelators was avoided. The effect of Ca2+ depletion on the peak permeability ratio was time-dependent. with the greatest effects seen between 30 and 60 min (Fig. 5, inset). Steady-state permeability was not affected by Ca2+ depletion except in a slowing of the rate at which the minimum P&PI, ratio was reached. The final Pet/P, ratios reached in low Ca2* were the same under all conditions (Fig. 5). There was no difference in the K’-induced peak sheath potential from control when low [Ca2+10+ 1OOmM KJ was given without pretreatment with low [Ca2+],. Similarly. following 30-6Omin Ca2+ depletion, 100 mM G containing normal [Ca’+] (13.5 mM) resulted in a sheath potential not significantly different from a normal Ca*+ control. These results suggest that external Ca2+ must be present for an increase in perineurial K+ permeability to take place during a high K+-pulse. External Ca2+ depletion had no effect on steady-state

Ca2+(.I mM) -_L--LJ--L~

Fig. 3. The Meet of low Co’+ and Cd+ channel blockers on the K+-induced pcriarurial abaatb pote&aI. InrUex~tllJlowainthir8pMtbtcontrolrbatb~wcnypo*ihrcriyOmVfEadw~ offZmV).Thrpariodofi00K+cxporurririDdiaudbythebslindalloftbrrqidpprrpIpOIIllLial ahi~pndthc~of~exponvtpriodbytbcbqliaamgofrfK~~~bythc arrows on the tmce at the top IeR. EWerib@ tbe expt&entsrromtoptobottom:sJk’+mvMbIyndvesd the K+-dcpMknt transient sbeatb potential in a concentrationdqendent mamber (e&u OfOinM ita2’ shown). l%ere vu aIso a sly11 reduxion of the steady-state potattiaI. Ca’+ at 0.1 or kn of ita -1 concentration rwersibIy blocked ttte tmnaient sheath potential without e&t on tIm rsdy-atatc ptWau. Verapamil (IO-’ M). a fairly speci& bIocker of voltqeqat.4 Ca’+ chum& rcvcrsI#y bb&ot! iha ariy transiant peak potential and reducedthe steady-state plrtau. La’+ (SUM) com@a@Iy but nvaribIy blocked the transient peak potential and nearIy abolished the steady-state potantial. The e&et on the steady-state was partially reversed on washout of the La”.

179

Glial Ca2+-dependent K+ transport

-101

r

0

.

120

60 Tim*

160

(WC)

0.04 0

.

20

40

60

80

100

120

lime(HC)

Fig. 5. The effect of low [Ca”], and time of depletion on the development of the K+ selectivity in high w+],. With

0

0

120

60 lima

160

(a-)

Fig. 4. Effect of reducing external (Ca2+] from 13.5 rnM to 1.35 and 0.1 mM on the K+-dependent perineurial voltage (panel A) and resistance (panel B). The period of high K$ exposure is approximately 150 s beginning at 0 time. (A) The effect of 1.35 (filled circles) (n = 4) and 0.1 mM Ca2+ (unfilled triangles) (n = 2) on the K+-induced sheath potential. The control condition (13.5 mM Ca2+) is indicated by the unfilled circles. A direct relationship is demonstrated between [Ca2+], and the early transient sheath potential but not the plateau potential suggesting that two different mechanisms may be involved in potential development over time. (B) Trans-perineurial resistance is also affected by the external [Ca2+] in a manner suggesting a requirement for Cai+ in the development of trans-perineurial ion permeation pathways. The effect of 1.35mM Ca2+ (filled circles) and the control are shown (n = 4).

for K+ transport during a IOOmM K,,+-pulse, which were 0.6 &-0.11 x IO-‘/s in low CaZ+ (n = 8) compared with 0.5 + 0.08 x 10e3/s in controls (n = 4). In normal [K+],, superfusion with 1.35 mM Cd+ for 30 min led to 17 rf: 4% increase in perineurial resistance (n =4). In 1OOmM G (with reduced Ca*+) the fall in resistance during the initial spike potential was reduced by approximately 50% in 1.35 mM Cd+ (Fig. 4B), indicating a Ca*+ dependency of the change in perineurial K+ permeability and conductance, especially in the initial phase of the high K +-response. rate constants

Effects of Ca ‘+-channel blockers

The effects of the organic Ca*+-channel blocker, verapamil and the ionic blocker, La3+ on perineurial K+ permeability were investigated. Exposure to verapamil (1O-5-1O-4 M) or La3+ (5 mM) reversibly blocked the early transient K+-dependent sheath potential (Fig. 3, lower two rows). The perineurial potential was depressed by La’+ and only slowly rose. to a potential comparable with the steady-state voltage observed in Rb+ (see Fig. 1). After La)+

decreasing q+ (1.35 mM, unfilled circles; 0. I mM, unfilled squares) the rate of development of high K+ selectivity is slowed compared with the control (tilled circles) although the final PCI/PKvalue is not statistically different from the control. This property can account for the loss of the peak transient potential without an effect on the plateau level. The inset shows the time dependency of Ca2+ depletion on the slowing of the development of the early high K+ selectivity of the perineurial glial sheath. The filled symbols represent the control followed in time by unfilled diamonds (7 min), triangles (51 mitt) and squares (66min), after exposure to 0.1 mM Ca2+.

washout the amplitude of the K+-induced transient sheath potential recovered completely but the rise to the peak and subsequent decay was slowed. La3+ had similar effects on the high Rb+ potential. These results suggested that La3+ had secondary, less reversible effects on perineurial ion permeability and conductance. The change in perineurial resistance during 100 mM [K+lO, and the effects of La’+ are illustrated in Fig. 6. When the nerve cord was bathed in saline containing 5 mM [La3+], the resistance increased immediately, reaching an almost maximum level within 5 min, indicating a peripheral site of action. In five animals, R,, increased 50% from 233 + 28 R cm* to 360 + 45 R cm*. Lanthanum blocked the increase in conductance in 100 mM [K+],, and transperineurial resistance did not change significantly during a 2-3-min test pulse. Trans-perineurial resistance remained high and did not return to control levels following La3 + washout (resistance was 160 f 20% of controls in paired measurements from three animals). The decrease in resistance during 100 mM G following La3+ washout was approximately 40% compared with 65% for the pre-La3+ control. The absolute change in R,, was the same (approximately 125 Qcm*) for both pre- and postLa’+ conditions. Consistent with the changes in resistance during La3+ treatment are the changes in the Pc,/PK ratio in La’+. As shown in Fig. 7, La3+ reversibly prevented the rapid K+-induced decrease in the PC,/PK ratio similar to the effect of Ca2+ depletion (Fig. 5). The permeability ratio did not decrease to the same level in high K: and La3+ compared with the pre-La)+

180

A. M. BUT et al

&-

E JOO-

0

E 6 200. _-

Jz

OJ

8

0

-..

..

60

120 l-ma

180

(set)

0.0

Fig. 6. The effect of 5 mM La’+ on the K+-induced perineurial R, change. Resistance changes in lOOmM [K+], before (unfilled triangles), during (unfilled circles) and after (filled circles) La’+ treatment. Lanthanum increased resting perineurial resistance by 50% and blocked the [K+mndent increase in conductance corresponding to the loss of the transient perineurial potential. Resistance did not return to control levels following washout of La’+ for 3045min. remaining at 160% of controls in paired experiments. The G-dependent increase in conductance recovered but remained higher for a longer time than controls. Horizontal bars (S.E.M.) indicate the variation in time when measurements were made. The absence of horizontal or vertical error bars indicates errors smaller than the size of the symbol.

control. Following La’+ washout the K,+ -induced permeability ratio fell to a lower level than the control. The inset of Fig. 7 is plotted as a percentage change from the lowest Pa/P, to expand the lower ratios to show the extent of the differences before and after La3+. Two separate effects of La’+ at the barrier were observed. The first was an acute reversible block of the initial transient spike potential due to the slowing of the K+-dependent increase in perineurial K+ permeability and conductance with La3+ in the superfusate. The second effect was a long-term irreversible increase in steady-state resistance due to La)+. The long-term effect was observed when the nerve cord had been bathed in La”+ for 5 min or longer and did not require continued presence of La)+. Figure 8A illustrates the depolarization of the axon as adaxonal [K+] increased during exposure to high [K+Jo in the presence and absence of La’+. In the presence of La3+, the influx of K+ was markedly reduced and [K+li increased only 3.63 + 0.88 mM during 2 min superfusion with 100 mM G as compared with approximately IS mM in the absence of La3+. This corresponded to the decrease in the K+dependent perineurial sheath potential and conductance, indicating that the bulk of influx of K’ into the interstitial space is a function of perineurial glial membrane permeability and transcellular K+ transport under normal circumstances. Following La’ + washout the rapid influx (and voltage and conductance changes) recovered. Rate constants for pccineurial K+ transport were significantly lower in the presence of La3+ (P < 0.05, paired t-test), and recovered completety following La”+ washout (Fig. 8B). The increase in interstitial (K+] during this period (first 60 s) appeared to be approxi-

0

20

40

60 60 lime (MC)

100

120

Fig. 7. Effect of La’+ on the change in perineurial sheath KC permeability. La’+, like CaZ+ depletion delayed the development of the high selectivity to K+ in high K$. At the end of the K+pulse selectivity for K+ in La3+ (filled circles) was not as great as in the control @en circles) (PCI/PK in La ‘+ =b.572 f 0.118 compared \;rith the control. 0.337 + 0.120. n = 6) indicatine a block of both Ca2+- ‘and K+%kctike channels. Folloking washout (open triangles) the selectivity, in high G was significantly greater (Pa/P, =0.106 kO.019) suggesting that a nonselective permeability pathway present in the resting state was restricted by La3+ in an irreversible manner. These results are consistent with the change in R,, illustrated in Fig. 6.

mately half that in controls but considering that transport rates recovered after La-‘+ washout the results suggest that the apparent slowing of K+uptake is due to the effect of La3 + on the rate of initial increase in K+ selectivity not to a long-term decrease in transport. In summary, the reversible block of the rapid transient perineurial K + and Rb+ potential, conductance and selectivity in the absence of external Ca2+ or in the presence of Ba2+, La3+ or verapamil suggests that the high selectivity of the perineurium is dependent on a Ca’+-dependent increase in K +-uptake by the perineurial glia. The site of action of the various agents used in this study are presumed to be Cal+ channels, Ca’+ influxI or directly on Ca’+-activated K + channels. MorphologJl

Because La’+ is an electron-dense substance and restricted to extracellular spaces of tissues (see e.g. Ref. 21) its distribution may be determined by electron microscopic techniques providing an indication of the location of the penneabitity barrier. The results are also of interest because they provide an opportunity to combine physiological studies of the effect of LaS+ on Ca’+-dependent K+ permeability and selectivity with morphological studies of La3’ distribution to indicate the site of action of La’+ as a Ca2+ and K+channel blocker. In the de-sheathed nerve cord, exposure to La” resulted in a prokmgation of the medial giant axon action potential duration from < 1 ms to 2-3 ms’& (Butt and Lie&man, unpuMished observations). In the intact preparation, during i 5 tin exposure to La’+ there was no effect on the action potential,

181

Cilia1 Ca2+-dependent K+ transport

--I”‘5mM

8

Id+7

‘*O: I 0.5.

0\\ 8

1

-0

----0

L1 ----A-------0.010

- - - SO ’ - .

A - . 100 ’ - ’

[K+lo (mu) Fig. 8. The effect of La’+ on the K+-induced axon membrane potential change in an intact sheath preparation (panel A) and perineurial sheath K+ transport rate (panel B). (A) Change in medial giant axon membrane potential in the intact preparation during 2-min exposure to 100 mM K+ bathing solution (filled bar under traces) before, during and after treatment with 5mM La’+. The change in axon membrane potential reflects an increase in interstitial [K+]. In controls, there was a rapid depolarization of the axon membrane. In La ‘+, K+ influx was greatly slowed and interstitial [K+] rose by an average of only 3.63 + 0.88 mM (n = 7) in 2 min of 100 mM [K+],, corresponding to the block of perineurial K+ permeability. The effect of La’+ on transport rate was reversible (panel B) although the development of high K+ selectivity was slowed. (B) Kate constants for perineurial K+-uptake before (unfilled circles), during (triangles) and after (filled circles) La 3+ treatment. Potassium transport rate was inversely related to external [K+], and was reversibly blocked by La’+. Following La’+ washout K+ transport was not statistically different in three paired controls although there was a trend towards a slightly greater

transport rate constant after washout of La’+. suggesting Las+ is prevented from entering the adaxonal space due to a peripherally located diffusion barrier. To verify these observations and for comparison with earlier studies,*’ the distribution of electrondense La3+ within the nerve cord was investigated electron microscopically. The nerve cord was prepared for electron microscopy following electrophysiological measurement of resistance changes, as described above, and after La’+ washout. An ultrathin section through the perineurial sheath and medial giant axon is shown in Fig. 9. The morphology of the crayfish ventral nerve cord has been described previously,*‘.*’ and will only be discussed briefly here. There is an outer neural lamella with distinct collagen fibres and a discontinuous outer layer of cells with large prominent nuclei, thin cytoplasm, and few intracellular inclusions, notably large mitochondria and granule-filled vesicles. Within this there was a thin, two to three cells thick layer, lying on the basal lamina. The cytoplasm of these cells was granular and with large densely packed mitochondria. Between this thin layer of cells and the medial giant axon there were three distinct regions characterized by the size of the extracellular space. There was an outer, thick layer with thin strands of cells supported within large collagen-filled extracellular spaces. In the de-sheathed preparation, these outer layers are removed. Within this there were five or so layers of glial cells interspersed with more dense extracellular material and collagen fibres. Immediately adjacent to the medial giant axon were two to

three concentric layers of periaxonal glia. A barrier to the free movement of La3+ across the perineurium was observed, and there was no evidence of La3+ in the extracellular spaces between adaxonal glia. Lanthanum accumulated at the basolateral membranes of the peripheral cell layer of the perineurium, which appeared to form a continuous barrier to La3+ permeation (at arrows in Fig. 9). It seems likely that the site of accumulation of La3+ is also the site of the crayfish perineurial barrier and the likely site of blocking action of La’+. Although a complete morphological study was not undertaken there were no obvious distinct tight junctions, consistent with previous studies.20*2’ Lanthanum remained bound to collagen fibres in the neural lamella after washout, as observed previously,2’ indicating the presence of anionic sites, probably of glycosaminoglycans, embedded within the collagen matrix. 3o The strong binding of La)+ within the neural lamella, and its accumulation at the basolateral membranes of the inner perineurial glia and within intercellular clefts, imply that these are the likely sites for the long-term effects of La3+ on sheath permeability and resistance, consistent with an effect on membrane surface charges.15 DISCUSSION

In the first paper of this series” it was demonstrated that the crayfish perineurium, in the presence of high external [K+], develops a large and highly selective permeability to K+ which leads to the

182

A. M. Burr er al.

Fig. 9. Electron micrograph of La ‘+-treated following electrophysiological studies. Lanthanum was seen as black electron-dense deposits (at arrows) accumulating in the intercellular clefts and at the basoiateraf membranes of the glial perineurial barrier and bound to collagen fibres of the outer neural lamella (NL). There was no evidence of La’+ in the interstitial spaces between inner perineurial glia (IPG) or periaxonal glia (PAG) surrounding the medial giant axon (MGA). Magnification x4000.

generation

of a large transient

spike potential

across

the glial perineurial blood-brain barrier and an accumulation of KC in the extracellular space of the intact nerve cord.’ The results of these experiments suggested the presence of specific K+ conductance channels in the membranes of the perineurial glia. These characteristics, having, been described in the previous investigation, were used here to further characterize perineurial K’ permeability and transport under a variety of conditions known to act on membrane K+ and Ca*+ channels.

The blocking action of Caj,+ -depletion, verapamil, Ba*+ and La3+ on the high [K+],- and [Rb+],induced early transient sheath potential suggest the presence of both voltage-dependent Ca*+ charm&s and Ca*+-activated K+ conductance channels at the basolateral membranes of perineuriaf glia. -14.19.29.32 Voltage-dependent Ca2 + channels and Ca’ +dependent K+ channels have both been demonstrated in giial cell~.*~~~**~ The lack of effect of TEA and other K+-channel blockers does not, by itself, rule out Ca*+-dependent K+ channels,’ and agents used

Glial Ca*+dependent K+ transport extracellularly in this study do not block all glial K channels in all cases6,io For example, external Ba2+, TEA and 3,4-DAP did not block the high K-conductance of Muller cells in a slice preparation, but were effective when cells were dissociated.24 The fact that 3,4-DAP entered the nerve cord and diffused to the giant axon to produce a prolongation of the action potential suggests that its lack of action on the sheath potential and transport was due to the properties of the K+ channels and not to a diffusion restriction of the agent. Barium may directly block glial K+ permeability:33 however, in these studies the results suggest Ba2+ may also act by a blocking of Ca2+ influx through Ca2+ conductance channels.3,‘9.33 Several properties of the K+-induced increases in trans-perineurial permeability and selectivity for K+ and Rb+ suggests that our results can be explained by proposing the existence of at least two K+-channel types involved in maintaining high K+ selectivity of the perineurial gliai4 a rapidly activated Ca2+activated K+-channel and a voltage-activated K+channel. Alternatively, the results might be also explained by proposing that the Ca2+-activated K+channel is both Ca2+ and voltage-sensitive.” In the absence of Cd+ the channel opens slowly as high K,+ depolarizes both faces of the perineurial glia (Figs 3, 5). In low Cd+, the steady-state transport of K+ through the sheath was not affected although the development of the high selectivity and permeability was delayed. The concentration and voltage dependence of perineurial K+ permeability, described previously, has characteristics associated with inward rectifying K+ channelsI as demonstrated in other glia.” Calciumdependent K+ channels are also known to have a complex sensitivity to voltage and the concentrations of both Ca’+ and K+,19 and Ca2+-activated K+ channels may have characteristics of inward rectifiers.” In short, our experiments do not allow us to distinguish between the one or two K+-channel proposal. To make this determination will require patch clamp analysis of the perineurial glia channel populations. The results of our experiments with Rb+ are particularly instructive since the repetitive voltage spikes generated by Rb+ may be explained by the participation of Ca2+- and voltage-activated K+ channels in the perineurial glial cell membrane. The increase in perineurial permeability in Rb+ was more transient than in K+, and conductance of the barrier increased only briefly before recovering towards resting levels. In contrast, the sheath in high KJ does not immediately recover its initial low selectivity and conductance and R, increases only in direct proportion to the change in intrasheath K+ and the K+ electrochemical gradient across the perineurial sheath. On the other hand, Rb+-induced changes in sheath ion selectivity, conductance and potential are blocked by low Ca2+ and La)+ as are the

183

K+-iniluced changes. The results with high Rw suggest a use- and voltage-dependent blocking action of Rb+.2* At the peak of the sheath potential (the time of maximum inward driving force on Rb+) Rb+ enters the K+-channel and produces a conductance block. As the block develops the potential falls to one third to a quarter of its peak value allowing the Rb+ block to be relieved as K+ (the normal pet-meant species) is better able to compete for the channel. Since intrasheath [Rb+] increases only slightly during the rapid voltage change, the Rb+ electrochemical driving force remains high. With relief of the Rb+ block an increase in perineurial permeability can again occur resulting in a second and third spike potential. The cycle can repeat itself to give several spontaneous spikes until interstitial [Rb+] increased sufficiently to reduce the electrochemical gradient below the critical level. It should be noted in this regard that the change in P,,/P,, ratio that occurs with each spontaneous spikes is the same (Fig. 2, inset) but the amplitude of the third spike (Fig. 1A) is smaller, reflecting an increasing accumulation of interstitial Rb+. A comparison between the permeability of Rb+ and K+ is of interest with regard to the types of channels that may be present on the perineurial glial cell membrane. Using the data of Fig. 2 a plot of the ratio between PRb/PK (data not shown) shows that at the peak of the transient sheath potential the ratio is approximately 0.68 which is consistent with the ratio found for Ca2+-activated K+ channels in Aplysia neurons and the delayed rectifier channels in snail neurons (see Ref. 19). In the steady-state (plateau) the ratio decreases to 0.1 reflecting the inactivation of membrane Ca2+-activated K+ channels or more likely a block of these channels by Rb+. Under these circumstances, Rb + probably traverses the perineurial barrier solely through the highly restrictive paracellular pathway. There was a part of the trans-perineurial potential in elevated [K+lo which was not sensitive to Ba2+, verapamil, La3+ or Cd+ -depletion, corresponding to perineurial K+ permeability other than by the highly selective Ca2+-dependent K+ transport pathway. This potential had a selectivity sequence to monovalent cations most similar to Eisenman’s sequence IV,12.17and indicated a weakly-charged pathway, most likely the paracellular pathway and intercellular junctions. Electron microscopic examination of La3+-treated tissue demonstrated a barrier-to-ion movement at the peripheral basolateral membranes of the glial perineurium with accumulation of La3+ at these membranes and in the intercellular clefts indicating the outer layers of the perineurial glia are the likely site of La3+ action and the site of the permeability and selectivity barrier of the crayfish perineurium (Fig. 9). Impermeability of the perineurium to Na+ I7 but not K+ and other cations suggest cation binding sites in the proteoglycan matrix of the intercellular

184

A. M. BUT er al.

spaces may play an important role in interstitial ion homeostasiP as well as barrier permeability. The results also suggest paracellular permeability may be lower than indicated by trans-perineurial electrical resistance” and the scarcity of tight junctions in the crayfish perineurium.20 The electrical resistance may be more a reflection of transcellular permeability*’ through voltage-sensitive K+ channels activated by applied current flow. When perineurial K+ transport was blocked by La3+ or Rb+, barrier permeability approached levels observed in vertebrate blood-brain barrierI which is determined by low paracellular permeability through intercellular tight junctions modulated by active transport mechanisms.’ The low rate of K+-uptake in the presence of La3+ or steady-state Rb+-uptake, may thus be a measure of paracellular transport. There is evidence to support the presence of active Na-K exchange and Na-K-Cl co-transport in the crayfish perineurium2 but their role in interstitial ion regulation is likely to be over the long-term and in lower extracellular [K+). At present the role of active transport processes in ion regulation as described in these investigations is not known. In summary, K+-selective permeability

of the crayfish perineurium is modulated by a Ca*+-dependent mechanism at the basolateral membranes of the glial barrier and is ion- and voltage-sensitive. Calcium influx through voltage-dependent Ca*+ channels may be the trigger for an increase in glial K+ permeability. Calcium-dependent K+ channels activated by influx of Ca2+ appears to be important in the increase in perineurial permeability. The low resting perineurial resistance and its dependence on the electrochemical gradient would suggest that K + permeability changed in response to a combination of glial membrane potential and extracellular [K+], and the high degree of perineurial K+ selectivity may be due to inward rectifying K’ channels (cf. Ref. 24). Based on the results of this and the previous investigation” a diagrammatic representation of the structure of the perineurial sheath and proposed cellular mechanisms for movement of K+ is shown in Fig. IO. The normally high [Ca”] of crayfish haemolymph may explain the predominance of Ca-dependent transport at the perineurium, and its voltage and ion dependence would be well suited to playing a role in K+ spatial buffering of the CNS.“.r3.” A second route for barrier ion transport was suggested by the results; a restrictive, low selectivity pathway, possibly the paracellular pathway, unaffected by K+-channel blockers, low [Ca*+], and Ca*+-channel blockers with the exception of La’+.

K+(Rb+l

PAGP-

-

-&m-e

Axon Fig. 10. A diagrammatic representation of the perimurial sheath and the pathways for K+ and Ca*+ movement between the interstitial and outaide fluid space. BM, basolateral membrane; NL, neural lamella; PAG, periaxonal glia (Schwann alls); PNG, perineurial giia. The space betwun the perineurial glia and periaxonal glia repraents the region of tbe inner perineurial ghal layers as indicated in Fig. 9. K+. along with Cl-, move across the perineurial glial membrane through two types of K’ channels; Car+activated and voltage-sensitive K+ channels and through tbe interuIhtlar matrix. The driving forces for K+ transport are concentration forces and ekctrical forces The ebt%ical force (“spatial buffering”) is the current generated from the voltage gradient produoed by K’ depolarization of the outer surface of the perineurial alls during a high [K+Jpulse and is indicated in the diagram by “I”. The sequmu~ of events stqPrsted by the results of our invutigations are pqmsed to occur as follows. The depolarization opens Car+ channels. The Ca” that enters the all activates the Car+-activated K+ channels increasing the sekctivity of the membrane for K + and causing a massive depolarization of the outer cell surface and the devehqmtettt of a sbath potential. since verapamil and Baz+ and Ca’+ depletion ail appear to block the Ca’+ influx but only slow the deve&pment of the permeability incmaae to K+, K+ may alsomove through the ail via another set of K+ channels. The efkets of La’+ are more general, not only blocking the Ca2+channel but also the K+ channels and binding to and condensing the mucopolysaccharide layer to decrease diffusion through the interulltdar matrix. Rb+ also appears to move through these channels but because of its tendency to block K+ channels, when it enters under large forces, its influx is self-limiting resulting in a rapid repolarimtion of the sheath. As the tissue repolarizts. the block is rrlicvad and the Rb+ influx may be restored leading to a regoner&e voltage spike. Based on the results of eBlux studies the channel structures and pathways for K + transport appear to be symmetrical. AcknowUgerrrenrs-We would like to thank Dr N. J. Abbott for useful discussions during this study and her review of earlier drafts of this manuscript. We are appreciative of the help of Pauktte Hahn with the electron mkroscopy. This work was supported in part from graots (to E. M. L.) from the Anny Research G&e, DAAG&-K0023. and from the National !Scknce Foundation. INT 861115OH.

REFERENCES

I. Abbott N. J., Moreton R. B. and Pichon Y. (1975) Electrophysiological analysis of potassium and sodium movements in crustacean nervous system. J. exp. Bid. a3, U-115. 2. Abbott N. J. and Pichon Y. (1987) The gliaf Mood-brain barrier of crustand cephalopods: a review. J. fhysiol., Paris 82, 304-3 I 3.

Glial Ca2+dependent

K+ transport

185

3. Almers W. and McCieskey E. W. (1984) The non-selective conductance in calcium channels of frog muscle: calcium selectivity in a single file pore. J. Plryfiol., Land. 353, 585608. 4. Ballanyi K., Grafe P. and Ten Bruggencate G. (1987) Ion activities and potassium uptake mechanisms of glial cells in guinea-pig olfactory cortex slices. J. Physiol., Land. 383, 159-174. 5. Barrett E. F. and Barrett N. J. (1976) Separation of two voltage-sensitive potassium currents, and demonstration of a tetrodotoxin-resistant calcium current in frog motoneurones. J. Physiol., Land. US, 737-774. 6. Bevan S., Chiu S. Y., Gray P. T. A. and Ritchie J. M. (1985) The presence of voltage-gated sodium, potassium and chloride channels in rat cultured astrocytes. Proc. R. Sot. Land. B225, 299-313. 7. Bradbury M. W. B. (1979) The Concept of a Blood-Brain Burrier. Wiley, Chichester. 8. Butt A. M., Hargittai P. and Lieberman E. M. (1986) Ion selectivity of the crayfish blood brain barrier: an electrophysiological study. Neurosci. Abstr. 12, 1259. 9. Butt A. M., Hargittai P. and Lieberman E. M. (1987) Role of mucopolysaccharides in the ion perm-selectivity of the crayfish nerve cord perineurium. Fe&. Proe. 46, 1275. IO. Chiu S. Y., Shrager P. and Ritchie J. M. (1984) Neuronal-type __ Na+ and K+ channels in rabbit cultured Schwann cells. Nature 311, 156157. 11. Coles J. A. and Orkand R. K. (1983) M~i~ation of potassium movement through the retina of the drone (Apis rne~fi~r~) by gIia1 uptake. J. Physiol.. Land. 340, 157-174. 12. Eisenman 6. {1%2)‘Cation selective glass electrodes and their mode of operation. Biophys. J. 2 (Suppf. 2), 259-323. 13. Gardner-Medwin A. R. (1983) Analysis of potassium dynamics in mammalian brain tissue. J. Physiol. x15, 393-426. 14. Hagiwara S. and Byerly L. (1981) Calcium channel. A. Rev. Neurosci. 4, 69-125. 15. Hagiwara S. and Takahashi K. (1974) Surface density of calcium ion and calcium spikes in the barnacle muscle fiber membrane. J. gen. Physiol. 50, 583-601. 16. Hansen A. J., Lund-Andersen H. and Crone C. (1977) K+-permeability of the blood-brain barrier, investigated by aid of a K+-sensitive microelectrode. ACIUphysiol. scund. 101, 438-445. 17. Hargittai P. T., Butt A. M. and Lieberman E. M. (1990) High potassium selective permeability and extracellular ion regulation in the glial perineurium (blood-brain barrier) of the crayfish. Neuroscience 38, 163-173. 18. Hartz T. and Ulbricht W. (1973) Comparison of the effects of calcium and lanthanum on the crayfish giant axon. Pjitigers Arch. 345, 281-294. 19. Hille B. (1984) Ionic Chunnefs of Exciruble Membrunes. Sinauer Association Inc., Massachusetts. 20. Lane N. J. and Abbott N. J. (1975) The organization of the nervous system in the crayfish Frucam&~ clarkii, with emphasis on the blood-brain interface. J. Ce# Tiss. Res. 156, 173-187. 2t. Lane N. J., Swales L. S. and Abbott N. J. (1977) ~nt~nurn penetration in crayfish nervous system: observations on intact and ‘desheathed’ preparations. J. Cell Sci. 23, 315-324. 22. Lieberman E. M., Hargittai P. T., Pascarella J. and Abbott N. J. (1985) Role of epineurium in ionic homeostasis of the crayfish CNS: an eiectrophysiological study. Biophysiol. J. 46, 146a. 23. MacVicar B. A. (1984) Voltage-dependent calcium channels in glial cells. Science 226, 1345-1347. 24. Newman E. A. (1985) Voltage-dependent calcium and potassium channels in retinal glial cells. Nurure 317, 809-811. 25. Orr L. and Lieberman E. M. (1987) Studies of the cardiac-like action potential in crayfish giant axons induced by platinized tungsten metal electrodes. J. exp. Biol. 128, l-17. 26. Quandt F. N. and MacVicar B. A. (1986) Calcium activated potassium channels in glial cells. Neuroscience 19,29-41. 27. Schofield P. K. and Treheme J. E. (1984) Localization of the blood-brain barrier of an insect: electrical model and analysis. J. exp. Biol. 109, 319-331. 28. Standen N. B. and Stanfield P. R. (1980) Rubidium block and rubidium permeability of the inward rectifier of frog skeletal muscle fibres. J. Physic& 304, 415435. 29. Thomas M. V. (1984) Voltage-clamp analysis of a calcium-mediated potassium conductance in cockroach (Peripkmeru umericanu) central neurones. J. Fhysiai. 350, 159-178. 30. Treheme J. E., Schofield P. K. and Lane N. J. (1982) Physiolo~~l and ultras~u~ur~ evidence for an extracellular anion matrix in the central nervous system of an insect (Periplaneta americana). Brain Res. 247, 255-267. 31. Vestergaard B. B., Stampe P. and Christopherson P. (1987) Voltage dependence of the Ca2+-activated K+ conductance of human red cell membranes is strongly dependent on the extracellular K+ concentration. J. Membrane Biol. 95, 121-130. 32. Waltz W., Shargool M. and Hertz L. (1984) Barium-induced inhibition of K+ transport mechanisms in cortical astrocytes-its possible contribution to the large Ba2+evoked extracellular K+ signal in brain. Neuroscience 13, 945-949. 33. Walz W. and Wilson D. C. (1986) Calcium entry into cultured mouse astrocytes. Neurosci. Kerr. 67, 301-396. (Accepted 31 October 1989)