Developmental Brain Research, 65 (1992) 101-114 © Elsevier Science Publishers B.V. All rights reserved. 0165-3806/92/$05.00

101

BRESD51395

Postnatal development of electrogenic sodium pump activity in rat hippocampal pyramidal neurons Atsuo Fukuda and David A. Prince Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA 94305 (U.S.A.) (Accepted 24 September 1991)

Key words: Electrogenic sodium pump; Na+,K+-ATPase; Glutamate; Development; Hippocampus; Intracellular recording; Antibody; Immunocytochemistry

We assessed the development of electrogenic sodium pump (Na+ pump) activity in CA1 pyramidal neurons of rat hippocampal slices by studying the prolonged hyperpolarization which foEows glutamate-induced depolarization (postgiutamate hyperpolarization or PGH) at different postnatal ages. We also examined the development of membrane-bound enzyme in the hippocampal CA1 subfield with light microscopic immunocytochemistry and an antiserum against Na+,K+-ATPase. The PGH, which has previously been shown to be due to activation of an electrogenic Na+ pump in adult hippocampai CA1 neurons, was eliminated by strophanthidin, a Na+,K+-ATPase inhibitor, at all ages. It was unaffected by several potassium channel blockers, an intracellular calcium chelator, intracellular CI- injection or tetrodotoxin (TTX) perfusion. The PGH thus appeared to be independent of K+ and CI- conductances and produced by an electrogenic Na+ pump in adult and immature animals activated in large part by entry of Na+ through the glutamate receptor-channel complex. The size (integrated area) of the PGH was directly proportional to the area of preceding glutamate-induced depolarization (GD) and relatively voltage independent. Similar GDs could be elicited from postnatal day (P) 7 to P ~ 35, however, only very small PGHs were produced in neurons from P7-11 animals. A ratio of PGH area to GD area (PGH ratio) was calculated for each neuron and used to compare Na + pump activity at different ages. There was a significant inc~-.ase in the mean PGH ratio with age when F7-11, P21-25 and P35-39 groups were compared. Na+ pump activity estimated front the PGH ratio is very low in the first postnatal week but develops gradually over the first 5 weeks of life. Immunostaining for Na+,K+-ATPase in adult rat hippocampi revealed a punctate reaction product surroundiag pyramidal cell bodies, whereas the staining was uniform along plasmalemma of dendrites in stratum radiatum and stratum oriens. By contrast, only minimum staining was present surrounding cell bodies and dendrites of P'/hippocampi and staining in stratum pyramidale was not punctate at this age. Na+,K*-ATPase activity estimated grossly from immunocytochemieal staining is very low in the first postnatal week, increases during the first 5 weeks and develops a characteristic focal localization. These results suggest that Na+,K÷-ATPase levels in hippocampal CA1 pyramidal neurons are low and insufficient to allow substantial activity of the Na ÷ pump in the first postnatal week, with gradual attainment of adult functional levels over subsequent weeks, The period of development of the PGH is comparable with that found using immunocytochemical staining for Na+,K÷ATPase. Lower levels of electrogenic Na+ pump activity at early stages of development could be one factor contributing to the increased susceptibility of immature hippocampus to ictal discharges associated with prolonged membrane depolarizations. INTRODUCTION Na+,K+-ATPase, the membrane-bound enzyme which is the basis of the sodium pump (Na + pump), maintains Na + and K + iop concentrations across membranes of wide variety of cell types6's6 and the Na + pump also functions as an electrical current generator 6. Electrogenie Na + pumps exist commonly in mammalian CNS where they have an important role in not only regulating ionic gradients but also generating electrical currents that affect neuronal excitability4. An increase in intracellular Na + concentration ([Na+]i) is the primary stimulus for the electrogenic Na + pump and the rate of pumping is proportional to [Na+]i 62. Increases in Ha + conductance during depolarizing events are likely to increase [Na+]i and result in activation of the Na + pump. Periods of

intense activity that cause increases in [Na+]l may therefore be followed by a slow hyperpolarization generated because the Na + pump extrudes approximately three Na + ions for every two K + ions it takes ia 6. Indeed, the hyperpolarizations following long-lasting tetanic activation 12 and spreading depression t7 in hippocampal pyramidal neurons are thought to be Na+,K+-ATPase dependent. Na +,K+-ATpase activity assayed from whole brain homogenate is low in immature brain and increases with brain maturation 43. Although it is known that the timetable of brain maturation is regionally different, few studies are available that focus on the ontogenesis of Na+,K+-ATPase in different areas 16, because most biochemical assays of Na+,K+-ATPase activity have been done using homogenated brain tissue. It is important to

Correspondence: A. Fukuda, Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA 94305, U.S.A. Fax: (1) (415) 725-7459.

102

understand how Na+ pump activity changes in the hippocampus during maturation, because it has been postulated that developmental or induced decreases in Na+,K+-ATPase can contribute to epileptogenesis in this structure1”33,and it is known that immature hippocampus is xclatively seizure prone15*s8. Histochemical observations indicate a continuous development of Na+,K+-ATPase activity between PS and P30 in rabbit hippocampusr6,and it has also been suggested that lower Na+,K+-pump activity may make region CA1 more seizure prone than CA3 in immature (P&12) rabbits”. However, the chronological and quantitative development of the physiological features of Na+ pump activity in the relatively seizure susceptible CA1 pyramidal neurons has not been examined. Application of the excitatory amino acid glutamate to hippocampal CA1 neurons is known to produce a depolarization (glutamate-induced depolarization or GD) associated with an increase in cationic conductance7113166 that results in an increase in [Na+]i. Previous experiments have shown that the prolonged hyperpolarization which follows GD (postglutamate hyperpolarization or PGH) is due to activation of an electrogenic Na” pump caused by Na+ entry, and that the PGH is a useful physiological assay of electrogenic Na+ transport63. Recently, the activation of the Na+ pump by increases in [Na+]* after glutamate excitation was also confirmed in cultured rat cerebral neurons using radiolabelled K+ uptake*! We therefore elected to use the PGH to quantify functional development of Na* pump activity. We also employed a recently reported immunocytochemical approach for localization of Na+,K+-ATPase in the hippocampus ‘*. Some of these results have been published in abstract form”, MATERIALS AND METHODS Electrophysiological recordings The techniques used here for preparingand maintaininghippoc-

ampal slices in vitro and obtaining intracellular recordings were similar to those described previously from the same laboratorya. Sprague-Dawley rats ages P7 to P 235 (date of birth = PO) were deeply anesthetized with sodium pentobarbital(50 @kg, intraperitoneal) and decapitated. A block of the brain including the hip pocampus was quickly removed and placed in cold (4’C) oxygenated modified Ringer solution, in which sucrose was substituted for NaCl (ref. l), [Cat’] was reduced and [Mg*+]was increased. The solution contained (in mM): sucrose 252, KC12.5, NaH2P04 1.25, MgSO, 2.0, MgCI, 3.0, CaC12OS, NaHCOs 26, glucose 10. Pilot experiments have indicated that use of this suc:ose saline tends to yield *healthier’slices particularly at young ages, as judged by the number of healthy cells encountered. T:ansverse slices with a thickness of 400 pm were cut in the modified Ringer solution with a vibratome (Lancer). Slices were transferred to an interface type recording chamber for incubation and perfused with normal Ringer solution consisting of (in mM): NaCl 124, KC15.0, NaH,PO, 1.25, MgSO.,2.0, CaC12 2.0, NaHCO, 26, glucose 10, which had a pH of 7.4 when saturated with 95%02, 5% CO*. The bathing solution was

miiintained at a temperature of 37 k 1°C.

Intracellular recordings were made from neurons in stratum pyramidale of area CAlb. Microelectrodes pulled with a horizontal puller (Sutter Inst.) and filled with 4 M potassium acetate (pH 7.2) had an impedance of 60-120 MQ. In some experiments, as indicated below, recording electrodes were filled with either 3 M KCl, 2 M CsCl, or 0.3 M 1,2-bis(o-aminophenoxy)ethane-N,N,N’,N’,tetraacetic acid (BAPTA) in 2.5 M potassium acetate. A highimpedance amplifier (Neuro Data IR 283) was used for current clamp recording and bridge balance was continuously monitored. Signals were recorded and stored on video cassette tape using a digitizer unit (Neuro-Corder DR-484). Data were digitized on-line or off-line and analyzed on a computer with data-analysis software (pClamp SO), Throughout this report results are expressed as mean + SD.; analysis of variance (ANOVA) and linear regression analysis were used as appropriate to establish statistical significance (criterion: P < 0.01). Focal applications of L-glutamate (20 mM, dissolved in perfusion medium) were made by applying brief pressure pulses of nitrogen gas (2-3 kg/cm*) to broken micropipettes (tip diameter, 2-10 pm) containing the glutamate solution. The duration of the pulse could be varied from 50 ms to 2.0 s to alter the size of depolarization, and the volume of glutamate solution ejected in this fashion was estimated at 10-106 pl from the diameter of the droplet. The tip of the glutamate micropipette was placed in the slice in stratum radiatum

within 100pm of the recording pipette at the expected position of major apical dendrites. The glutamate pipette was advanced slowly through the tissue until an abrupt and maximum GD (+30-45 mV from resting potential) was obtained. In all cases, the stereotyped response consisted of a GD followed by a PGH (Fig. 1A). An increase in the duration of the glutamate pulse resulted initially in an increase in amplitude of the GD; the GD increased in duration after reaching maximum amplitude (Fig. 1A). Action potentials were evoked during the initial portion of the response but became inactivated within one second. An increase in the duration of the glutamate pulse resulted in an increase in the amplitude and duration of the PGH as well (Fig. lA), and the size of the PGH (measured as the area of response) was proportional to the duration of the glutamate pulse (Fig. lB), The area of the PGH was directly proportional to the area of the preceding GD, for all sires of GD evoked (Fig. lC), The ratio of the PGH area to the preceding GD area (PGH ratio) had a relatively constant value in a single cell for a series of glutamate pulses (Fig. lC), but varied from cell to cell. Therefore PGH ratios for each glutamate application were averaged to represent the PGH ratio of a given cell. In some neurons from animals older than 3 weeks, fast transient hy perpolarizations could be evoked preceding the GDs, unless the glutamate pipette was appropriately positioned to activate maximum responses (not shown). These hyperpolarizations reversed polarity at about -70 to -80 mV, and were most likely mediated by increases in Cl’ conductance as a result of excitation of adjacent GABAergic interneurons 61. Neurons with transient hyperpolarizations were not used for analysis, A second pressure-ejection pipette, filled with stupphanthidin (100 NM, dissolved in perfusion medium) was used in some experiments. When necessary, addition of tetrodotoxin (TTX) (1 PM) to the perfusion solution was used to block voltage depen&nt Na+ conductances. lmmunocytochemical staining

Methods were similar to those described previously’*. A rabbit polyclonal antibody, raised against purified denatured catalytic a subunit (consisting of al, a2 and a3 isoforms) from bovine brain Na+,K+-ATPase and characterized as specific for the catalytic a subunit in rodent brain preparations***‘* was kindly provided by Dr. George J. Siegel (University of Michigan). Sprague-Dawley rats with ages of W, P20-21 and P3S (n = 3 in each age group) were anesthetized with sodium pentobarbital (SOmg/kg, intraperitoneal), perfused through the left ventricle with sodium periodate (0.01 M)-lysine (0.1 M)-paraformaldehyde (4%) fixative3’, and the brains removed and fixed for an additional 8 h. A block of brain

103

similar to that used in electrophysiological experiments was cut and embedded in paraffin. Transverse 6/~m sections were mounted on gelatin-coated slides and incubated in 20% normal goat serum/0.1 M phosphate-buffered saline for I h, 1:500 Na+,K+-ATPase antiserum at 4°(2for 16-18 h, 1:250 goat anti-rabbit IgG for I h, avidin-HRP for I h, and 0.025% diaminobenzidJne in 50 mM Tris buffer-0.01% hydrogen peroxide for 5 min. The immune sera were diluted in 1% normal goat serum/0.1 M phosphate-buffered saline. The Na+,K+ATPase antiserum was replaced by normal rabbit serum for the controls. The sections were cleared, mounted in Permount and photographed under a light microscope (Leitz). The following drugs were used: L-glutamate, monosodium salt; strophanthidin (both from Sigma); BAPTA (Molecular Probes); TrIg (Calbiochem); tetraetbylammonium (TEA) chloride (Eastman Kodak); and goat anti-rabbit IgGlavidin-HRP (Vector Lab, PK

4001).

RESULTS

Developmental changes in membrane properties Intracellular recordings were obtained from neurons in the pyramidal layer of the hippocampal C A l b region at ages P7 to P ~> 35. Neurons were grouped for statistical analysis according to the postnatal age, i.e., P7-11, 1)21-25 and P35-39. Results are summarized in Table I.

A

eomsl

B

100,m~

700 e00

a00mlt

G'

/

•

~e~.l|

In brief, P7-11 and P21-25 cells had significantly higher input resistances than those of the 'adult' (P35-39) group. 1)7-11 cells had lower amplitude and longer durat~o~ spikes than P21-25 or P35-39 neurons. Although A N O V A revealed a significant difference in resting membrane potentials between age groups, the results of linear regression (r = 0.15) indicated no significant correlation between resting membrane potenti~| and age. These data are similar to those previously reported ~'51.

Postglutamate hyperpolarizations during development PGHs of up to 15 mV were always evoked following GDs and lasted as long as 3 rain in adults (Fig. 2A). Although glutamate-induced depolarizations qualitatively similar to those in adults could be elicited in immature cells from P7-11 animals, only a minimal PGH could be obtained even following a maximal GD (Fig. 2B). The higher input resistance of neurons from this age group (Table I) and their capacity to generate trains of spikes (Fig. 2C) indicated that the absence of a significant PGH was not merely due to cell deterioration. In addition, the presence of slow afterhyperpolarizations (AHPs) indicated that a slow Ca2+-activated K + current (/AMp) could be elicited in this period (Fig. 2D). Although neurons from animals ir~ the P21-25 age range had similar membrane properties to those of adult cells, with the exception of a relatively higher input resistance (Table I), PGHs were significantly smaller than those in adult animals (see Fig. 7).

400TABLE I

Developmental changes in passive membrane properties IO0 .

,

.

,

.

|

.

,

..o

0 100 300 500 GLU Pulse OuraUon(ms)

0

4OO

0 GD Area (mY.e)

Fig. 1. Dose-response relations of glutamate-induced depolarization (GD) and postglutamate hyperpolarization (PGH). A.-C from same adult neuron. A: consecutive responses to glutamate (GLU) pressure pulses (arrowheads) of increasing duration. The GD and PGH increased in amplitude and duration as the duratiori of the pressure pulse was increased. Note the large decrease ~a input resistance during the GD due to a conductance increase produced by GLU. The apparent decrease in input resistance during the PGH is due to rectification. Downward voltage deflections in this and subsequent figures are responses to hyperpolarizing constantcurrent pulses (-0.2 nA). Resting membrane potential (Vm): -58 mV; input resistance (RN): 44 Mr2. B: plots of area of the PGH as a function of GLU dose (duration of pulse) indicate a linear relationship (r = 0.98). C: plots of PGH area as a function of preceding GD area show a linear relationship (r --= 0.99). Linear regression lines in B and C were fitted by least-squares method.

Action potential parameters measured from single spikes evoked by brief (30-50 ms) depolarizing current pulses. Amplitude: from threshold to peak; duration: at spike base. Apparent input resistance was calculated from the peak voltage response to a 0.l-0.3 nA hyperpolarizing current pulse (100-200 ms). Values are presented as mean +- S.D.

Postnatal day (number of cells)

Action potential amplitude (mV) duration (ms)

PT-I I

P21-25

P35-39

(n =4~)

(n= ST)

(n=74)

66.4 +_ 10.0' 83.8 --- 7.2 85.4 +-- 6.8 2.91 +- 0.9# '~ 2.04 +-- 0.37 2.00 -4- 0.28

Resting potential (mY) -68.4 - 7.6 -63.9 - 5.1-65.2 - 6.1 Input resistance (Mf~)

44.8 _ 19.1' 38.0 - 9.3* 30.9 - 8.0

*P < 0.005 when compared with values of P35-39 (post-ANOVA Tukey-test).

104

B

A Postnatal Day 37

Postnatal Day 9

I

o.ot

k I rain

D

C

[

I

!

....

20 ms

I'

,I-I

10 mV 1.0 nA 400 ms

Fig. 2. Responses to glutamate pulses in an adult and immature neuron. A: typical GD-PGH responses in an adult, postnatal day (P37), neuron following two consecutive GLU pulses of increasing duration (arrowheads). The PGH ratio of this cell was 1.35. Resting Vm, -66 mV; RN, 28 MCL B: representative responses from immature (139) neuron. Although GDs similar to those of adult neurons were evoked in response to OLU pulses (arrowheads), only small POHs were generated. The PGH ratio of this cell was 0.16. Resting Vm, -62 mV; RN, 48 MQ. C-D: membrane potential responses (V) to intracellular injection of hyperpolarizing and depolarizing current pulses (/) at resting potential in the cell of B. C: neuron generates a train of spikes that show frequency adaptation. D: activation of 9 action potentials by current pulse resulted in the generation of a slow AHP (V). Dotted line, resting membrane potential.

Na ~"pump mediation of postglutamatehyperpolarization during development W e tested the effectsof strophanthidin, a cardiotonic steroid which is a reversible inhibitor of Na+,K +ATPase 67, on the PGH. Strophanthidin (100/~M in perfusion medium) was applied from pressure ejection pipettes positioned on the surface of sliceswithin 200/~m of the recording electrodes. As expected, the PGI-I, regardless of its size or the age of the animal from which the slice was obtained, was completely and reversibly blocked by strophanthidin (n -" 7, Fig. 3A). In contrast to its depressant effect on the PGH, strophanthidin enhanced the size of the A H P (Fig. 3B), indicating that blockade of the P G H was not due to indirect effects of the drug to increase [K+]o, as previously suggested in pyramidal neurons of neocortical slices47. Effects of intracellularapplication of B A P T A , a C a 2+ che, itor, on the A H P and P G H were compared in imniature and mature cells. Intracellularrecordings were obtained by using electrodesfilledwith 300 m M B A P T A

in 2,5 M potassium acetate (n = 8). Within the first few minutes of impalement the AHPs were abolished indicating blockade of the underlying Ca2+-acfivated K + conductances (Fig. 4). In no case was any reduction in amplitude of the PGH observed when AHPs were abolished (Fig. 4). The above data confirm that the PGH recorded in immature neurons, although smaller than in mature cells, is produced by a Na + pump as in adults s3. Application of glutamate produced a GD followed by a PGH at all membrane potentials tested. Hyperpolarization of the membrane potential by constant current to levels more negative than the reversal potential of the AHP which followed directly evoked trains of spikes did not reverse or attenuate the PGH (n = 5, Fig. 5A). The GD was increased in amplitude as membrane potential was hyperpolarized. The PGH duration also increased with hyperpolarization, however no obvious change in amplitude was seen (Fig. 5A), perhaps because of the apparent decrease in input resistance due to rectification at hyperpolarized membrane potentials (note the de-

105 creased amplitude of responses to test pulses at hyperpolarized potentials in Fig. 5A). The area of the PGH increased proportionally to the increase in the GD area as membrane potential was hyperpolarized, therefore the PGH ratio (i.e. the ratio of PGH area to GD area) was maintained at a similar value at various membrane potentials (Fig. 5B). In younger animals, although the PGHs were already small at resting potential, no reversal of the PGHs was observed during hyperpolarization beyond expected EK (-85 mV). Thus the PGH was relatively voltage independent, making a significant contribution of various K + and CI- conductances unlikely. As reported previously c~, perfusion with 0 mM Ca2+/5

Control

Strophanlhidin

Wash

1 "'

ul i[ 1

1ov

1 inln

B

Control

~lmnthldin

. . . . . . . . . . . . . . . . . . .

iJ"l

..........

J~

Fig. 3. Blockade of the PGH by strophanthidin. A: strophanthidin

(100/~M in perfusion medium) applied to the surface of the slice from a pressure ejection pipette reversibly blocked the PGH in a P36 and a PI2 neuron. I)36 (upper traces): identical pulses of GLU (arrowheads, 200 ms) were applied to the neuron before (Control), 7 min (Strophanthidin) and 60 rain (Wash) after application of strophanthidin. Resting Vm, -57 mV; RN, 44 MQ. PI2 (lower traces): focal application of strophantliidin in a P12 neuron completely blocked the PGH within 5 rain (Strophanthidin) with partial recovery after 40 rain of washing (Wash). Resting Vm, -60 mY; RN, 48 MO. Note the smaller amplitude of the PGH in the PI2 vs the P36 neuron. The GDs were prolonged by strophanthidin in both cells, indicating that the PGH normallyshortens the duration of the GD. B: same neuron as in A (PI2). The AI-IP following trains of 7 action potentials (V) elicited by injection of a II0 ms depolarizing current pulse (/) was enhanced rather than being attenuated following the the same s*,rophanthidin application that eliminated the PGH. Action potentials are truncated.

mM Mn 2+ solution, which blocks voltage dependent Ca 2+ conductance and transmitter release, did not reduce either the GD or the PGH in adult (P >~35, n - 4) or immature (P = 14, n - 2) neurons (not shown). This finding suggests that Na + is the principle charge carrier for the GD, and that Ca 2+ does not make a large contribution. A few adult and immature neurons, impaled with 3 M KCl-filled electrodes, had GDs, PGHs and PGH ratios that were similar to those in which potassiumacetate electrodes were used, even though CI- containing electrodes shifted the reversal potential of spontaneous inhibitory postsynaptic potentials (IPSPs) so that these events became positive-going at resting potential (n ffi 5, not shown). Impalement with electrodes containing 2 M CsCi, which reduces most K + conductances 7, caused a gradual depolarization of the cell membrane potential and generation of exceptionally large and prolonged presumed Ca 2+ spikes 5'7, however, the PGH was not blocked either in adult or immature cells (n = 9, Fig. 5C). Another K + channel blocker, TEA (10 mM TEA-CI substituted for equimolar NaCI in the perfusion solution) was tested on 4 mature neurons with or without TFX (1 /~M). In both conditions TEA was effective as judged by prolonged action potential duration and generation of Ca 2+ spikes, but at most only small changes in the size of PGH occurred (Fig. 5D). The addition of TFX did not affect the PGH, suggesting that Na + conductances associated with glutamate receptor activation, rather than with voltage dependent Na + channels, are the principa! contributors to subsequent activation of the PGH. Reduction in [Na+]o, produced by substituting choline chloride for equimolar NaCI, decreased amplitudes of the GD and abolished the PGH, as expected if this treatment reduced Na + entry and [Na+]l (not shown). Although we did not specifically test the effects of T r x on the PGH of immature neurons, we have no reason to believe that voltage dependent Na + conductances would play a more significant role in PGH activation than in mature cells. In fact, the known lower density of Na + conductances in immature neurons ~ makes it highly unlikely that TFX would have such an effect. Thus the PGH (Na + pump) is activated by entry of Na + through glutamate receptor-channels as reported previously in cultured cells2e, and the PGH ratio is relatively voltage independent with little contamination from K + or CI- conductances.

Na + pump activity estimated from the PGH ratio during development Since the area of the GD and the PGH are presumably in large part related to Na + influx through the T r x insensitive, glutamate-activated ionophore and Na + extrusion by the Na + pump, respectively, the PGH ratio

106 A BAPTA

P37

BAPTA

,,,,, ,,

,I--1

B

GLU IllJllllllll

jiiJmi;jlllllimmlmlmmil;,omv t i

i

I or.v

12"0nA

imin 200 ms Fig. 4. Persistence of the PGH in neurons impaled with BAPTA-containing electrodes. A: P37; left trace: the PGH persisted without a change in size 50 rain after impalement with electrode containing 300 mM BAPTA in 2.5 M potassium acetate. Right traces: the slow AHP following a train of spikes (V) elicited by a 120 ms depolarizing pulse (/) was abolished completely shortly after impalement. Resting Vm, -67 mV; RN, 25 MQ. B: P14; the PGH was not attenuated by intracellular injection of BAPTA (left trace, 30 rain after impalement) while the slow ,~-IP was abolished completely (right traces). Resting Vm, -70 mV; RN, 32 MQ. Spikes are truncated in A and B.

(PGH area/GD area) should provide an index of the net Na + pump activity in the cell. As indicated above, the lack of a significant effect of membrane potential on this measure, and the absence of detectable contaminating ion conductances make the PGH ratio suitable for studying developmental changes in Na + pump activity using current clamp techniques. PGH and GD areas were measured by integrating the membrane potential deflection from resting potential (baseline, defined as 0 mV) with time, using a data analysis program. The PGH ratio was calculated for each neuron and was used to compare Na ÷ pump activity at various ages. Since there were substantial changes in membrane properties during development (Table I), the PGH ratio of each cell from each age group was plotted as a function of membrane properties (i.e. action potential, input resistance and resting potential), and linear regression analysts was performed to rule out any membrane property-related differences in the PGH ratio (Fig. 6). There were no significant correlations between the PGH ratio and action potential amplitude or resting potential in any age group. However, the PGH ratio in the P21-25 group had weak but significant correlation with input resistance. This suggested that a leak component through damaged membrane might affect the PGH ratio.

Therefore, we arbitrarily established criteria for cell 'health' in each age group (see Table I). Only n~urons with resting potential more negative than -50 mV, action potential amplitude ~55 mV (P7-11) or 70 mV (P21-25 and P35-39), and input resistance higher than 35 Mg (P7-11), 30 MQ (P21-25), or 25 MQ (P35-39) were used for analysis of developmental changes in PGH ratios. The PGH ratios at ages P7-11, P21-25 and P35-39 were 0.25 --. 0.19 (n = 12), 1.12 --. 0.66 (n = 17) and 1.74 ± 0.72 (n - 28), respectively, and the differences between age groups were statistically significant (Fig. 7). The PGH ratio also had a strong and significant correlation with age when regression analysis was applied (r = 0.69; Fig. 7). Thus Na + pump activity, estimated from the PGH ratio, develops continuously during the period between P7 and P39,

Immunostaining for Na ÷,K ÷.ATPase during development No attempts were made to quantitate the density of staining, however, there were gross changes in staining pattern with age. In the adult hippocampus the stratum pyramidale, stratum radiatum and stratum oriens of CA1, CA2 and CA3 were all immunostained for Na+,K+-ATPase (Fig. 8A). In the CA1 region, pyramidal cell bodies as well as their apical dendrites appeared

107 to be immunostained (Fig. 8B). With higher magnification, the immunostaining pattern of Na +,K+-ATPase surrounding the pyramidal cell body plasmalemma appeared punctate with thick and dense staining, in addition to the diffuse background immunostaining; in contrast, only uniform staining density was observed in the ~tratum oriens and stratum radiatum of the same region (Fig. 8C). Pyramidal cell apical dendrites in the stratum radi-

B

,4 4'

4'

o ~k o

3 o O

~

O 3

O

O

O 2

Oo

~

O O

O

O

OO 0(30 OO

OO 2.

O O O

l

~ A

1

1

1

B Oee

GLU •65mV ~

0 -,.,--,.,.,.~.~ 40 60 80

0 100

Action Potential (mV)

414mY ,~

-108 mV

~-

-g4mV ~ 200 p - 6 5 mV

-lOemV~ ~

40mV

loo

GD Area (mY.s)

I rain

C

100

D Control

el~

eWl

vL__2 II

800m8

~

~

gOmV rain

100 m8

Fig. 5. Effects of membrane potential and potassium channel blockers on the PGH. A: PGH was not reversed or decreased in size as membrane potential was hyperpolarized from resting level of-65 mV by passing steady current through the microelectrode. Dose of GLU was identical at each membrane potential (arrowheads, 150 ms). 1)41; RN, 32 MQ at -65 inV. B: plots of PGH area as a function of preceding GD area at various membrane potentials for the cell of A. PGH area increased in proportion to the increase in GD area as the membrane potential was hyperpolarized. The ratio of PGH area to GD area (PGH ratio) was thus relatively unaffected by membrane potential. Linear regression line: r ffi 0.99. C (upper trace): intracellular Cs + (2 M CsCl.containing microelectrode) did not block the PGH. Note the subthreshold spontaneous voltage deflections due to enhancement of depolarizing IPSPs and ,~PSPs that resulted from a depolarizing shift in the CI- equilibrium potential and blockade of K + conductances by CsCI. There was no apparent increase in membrane conductance during the PGH, probably because of blockade of rectification by Cs +. Lower traces: individual complex and prolonged depolarizing response (V) is elicited by a short depolarizing current pulse (/), indicating that K + conductances were largely blocked. Constant hyperpolarizing current (-0.7 nA) was used to hold membrane potential at -80 mV.

• .,.,.,.~,.,., -80 -70

• . , • , --, • , • , -50 0 20 40 60 80 100 0

-60

Resting PoteNial (mV)

Input Resistance (Mn)

Fig. 6. Relationships between the PGH ratio and passive membrane properties during development. The PGH ratio was calculated for each neuron from all age groups by dividing the PGH area by the preceding GD area for each successful trial, and averaging trials. Each symbol (0, P7-11; &, P21-25; O, I)35-39) represents a~ average value in one neuron. A: plots of the PGH ratio vs action potential amplitude. Linear regression revealed no significant correlation between these variables in any age group (r = 0.37, -0.20 and -0.13 in P7-11, 1)21--25 and 1)35--39, resp.). B: plots of the PGH ratio as a function of resting potential also shows no significant correlations in any age group (r = 0.21, 0.28 and 0.24 in P7-11, P21-25 and 1)35--39, resp.). C: although there was no significant correlation between the PGH ratio and input resistance in the P7-11 (r = 0.33) and 1)35-39 (r ffi 0.38) groups, a significant correlation (r ffi 0.60, P < 0.01) was present in the P21-25 group.

atum appeared to be immunostained for Na+,K +ATPase along their plasmalemmas (Fig. 8C). No specific immunostaining was obtained when a hippocampal section was incubated in normal rabbit serum as a control (Fig. 8D). In the hippocampal CA1 region of the P7 animals, only minimum overall background immunostaining for Na+,K+-ATPase was present compared with control (Fig. 8E). Comparison of P7 and 1)35 CA1 regions showed that there was an increase in overall uniform density in stratum oriens, stratum pyramidale and stratum radiatum, as well as the appearance of focal dense

1)37; RN, 45 MQ. D: PGHs elicited with identical applications of GLU (arrowheads, 400 ms pulses) in a single cell before (Control) and after perfusion with solution containing 1/~M T r x and 10 mM TEA (TrxfFEA). PGH persists with only slight decrement in amplitude and duration (upper traces) in spite of the blockade of Na + and K + conductances produced by TrXfI'EA. Prolonged spikes mask the peak of the GD after TrXtl'EA. There is an apparent increase in RN produced by perfusion of 'ITXfrEA. In the lower traces, a Na + spike (Control, I0 elicited by a 40 ms depolarizing pulse (1, left) was abolished and replaced by a prolonged Ca2+ spike during perfusion with TI"XfI'EA (right). A larger depolarizing current (/) was required to generate the high threshold Ca2+ spike (I0. TIXfrEA also blocked the apparent increase in membrane conductance during the PGH, presumably through effects on rectification (c.f. responses to hyperpolarizing pulses in Control and TrX/TEA). P35; resting Vm, -65 mV; RN, 40 MfL

108 reaction product along the plasmalemma of pyramidal cell bodies, with age (c.f. Fig. 8C,E). The focal densities in stratum pyramidale were observed along with substantial diffuse immunostalning for Na+,K+-ATPase as early as P20 (Fig. 8F). Between P20 and I'35, there was a small increase in overall uniform background density in stratum oriens, stratum pyramidale and stratum radiatum, however, there were no further marked increases in punctate densities along pyramidal cell bodies (c.f. Fig. 8C,F). The punctate immunostalning for Na+,K +ATPase, thus developed between F7 and P20. DISCUSSION

Posmatal development of CAI pyramidal cell membrane properties Developmental changes in membrane properties of CA1 pyramidal neurons recorded in these experiments (Table I) were similar to those previously reported in rabbit40 and rat st hippocampal CA1 neurons. The smaller and broader action potential could be due to reduced Na + channel density and delayed rectifier K + conductance (IK), as in immature neocortical pyramidal neurons ls'~'34. T~.e higher input resistance in the younger neurons can be attributed, at least in part, to their smaller size4s. There were no significant correlations between resting potentials and age of cells, despite lower activity of Na + pump in immature cells. This might indicate that other factors which determine membrane potential, such as membrane conductances active close to resting potential, are counterbalancing the possible effects of reduced Na ÷ pumping, In neocortical neurons, the density of Na + channels increases almost 10 fold from embryonic day 16 to PS0is. A similar developmental sequence in hippocampal neurons could result in a smaller intracellular Na + load during usual depolarizations in immature neurons. Since immature hippocampal slices also have a lower rate of energy and oxygen consumption and more resistance to anoxia than mature ones2s, it is possible that other metabolic factors might influence the resting potentials of immature hippocampal neurons.

of Ca2+-activated K + conductance (IKCCa))such as Ic or IAHP3° was eliminated, since the PGH was not reversed at the expected K + equilibrium potential and not blocked either by perfusion of Mn2+/low Ca2+ solution or intracellular injection of EGTA. In order to study developmental changes in Ha + pump activity, we first confirmed some of the above findings in all age groups, and concluded that PGHs, although small, are also produced in immature neurons by Na + pump activation. The GD appears t o be primarily due to influx of Na + regardless of postnatal age. Although there might be some Ca2+ entry through glutamate receptor-coupled channels and voltage dependent channels, it is not likely to be large enough to account for a significant portion of the GD, since Mn2+/Ca2+-free solution did not alter the GD in mature (see also ref. 63) or immature animals (not shown). The GD area increased as the membrane was hyperpolarized, as expected if the potential were mediated by a Na + conductance (Fig. 5A). Since the PGH area, which we assume reflects net electrogenic Na + pump current activated by Ha + entry during the GD, was directly proportional to the GD area in all age groups (Fig. 1C), the PGH ratio was maintained at the several membrane potentials tested (Fig. 5B). Increases in voltage dependent membrane conductances at hyperpolarized membrane potentials (c.f. conductance pulses at -65 mV and -108 mV in Fig. 5A), which would be expected to decrease the amplitude of a pump potential~, might

rmO,eg

1.Y4:1:0.72 0

•

0

s

i 'I"

0 1.1R'J:O.M

0 0

(n,,17)

@

0 0

2-

I

...... o'6

•

• • . . . . . . "". . . . . . • i ~ ; ~ ........

0.25"0.19

(n-im

i,

•

............ i"" t A

• ,..°""

o .....i " ° i i ;I

Activation of Na + pump by Na + entry following glutamate-induced depolarization A previous study in mature hippocampal neurons~ concluded that the PGH is a Ha + pump potential from following evidence: the PGH is (1) blocked by strophanthidin; (2) reduced and slowed by reduction in temperature; (3) relatively voltage independent and (4) accompanied by no conductance change under manual voltage clamp, except a transient conductance increase in the initial part of the PGH that could be caused by residual glutamate. Possible mediation by one of the components

(n-m)

y,mO.O64x.O.2S p