C.D. Barnes and 0. Pompeiano (Eds.) Progress in Brain Research, Vol. 88 0 1991 Elsevier Science Publishers B.V.

307 CHAPTER 23

Noradrenergic and locus coeruleus modulation of the perforant path-evoked potential in rat dentate gyrus supports a role for the locus coeruleus in attentional and memorial processes C. Harley Department of Psychology, Memorial University of Newfoundland, St. John’s, Newfoundland, Canada

The perforant path-dentate gyrus synapse has provided a model system for functional neural plasticity in adult mammalian brain. NMDA-dependent long-term changes in neural connectivity occur at this synapse in response to highfrequency input. Norepinephrine (NE) applied exogenously or released endogenously can initiate both a short- and a long-term potentiation (LTP) of the dentate gyrus response to perforant path input. Triggering of the potentiated response depends on preceptor activation and does not require a high-frequency stimulus. An increase in locus coeruleus (LC) activity can initiate both short and LTP of the perforant path

response, although a reduction in LC activity does not alter baseline perforant path responses. This chapter considers differences between NE modulation in uitro and in vivo, differences and similarities between NE-LTP and frequencyinduced LTP, and the surprising specificity of NE effects at the perforant path synapse. Studies of NE in the dentate gyrus support a role for the LC in promoting both short- and long-term enhancement of responses to complex sensory inputs and are consistent with a role for the LC in memorial as well as attentional processes.

Key words: dentate gyrus, hippocampus, locus coeruleus, P-receptors, norepinephrine, long-term potentiation

Introduction During the last two decades the model system characterized by perforant path activation of the dentate gyrus-evoked potential has proved seminal for a new neurobiology of learning and memory. The demonstration that the perforant path dentate gyms synapse is capable of long-term potentiation (LTP) by repetitive high-frequency stimulation of the perforant path (Bliss and Mmo,

1973), and the subsequent insight that activation of a new class of voltage and chemically dependent synaptic channels, the NMDA receptor complex, was the trigger event for LTP has provided physiological support for the Hebbian model of associative learning (Wigstrom and Gustafsson, 1985). The present review discusses evidence that another synapse, the noradrenergic synapse originating from the locus coeruleus (LO, participates

308

in initiating long-term changes in the response of the dentate gyrus to perforant path input. After reviewing the effects of norepinephrine (NE) application and LC activation on the dentate gyrusevoked potential, we will consider briefly how these effects might be understood at the cellular level and, finalIy, speculate as to the role of these effects in the functioning of the dentate gyrus. Dentate gyms-evoked potential modulation by NE B.

Long-lasting potentiation In 1983 we reported that NE, iontophoresed for 2-5 min at the dentate gyrus cell body layer in the anesthetized rat, produced a significant potentiation of the perforant path evoked population spike which lasted for many minutes (Neuman and Harley, 1983). A single experiment documenting the evoked potential change induced by NE is shown in Figure 1. The profile of NE-induced long-term change averaged over 16 experiments is shown in Figure 2. Although the experiment in Figure 1 shows an example of an increase in the field EPSP, there were no consistent changes in EPSP amplitude over the 16 experiments of Figure 2. Population spike increases were reliable however, and whether or not the induced changes in population spike were long-lasting, an initial short-lasting potentiation was typical. ,%-Receptormediation Using the brain slice preparation we were able to confirm that NE applied at 10 p M for 10 min could consistently produce potentiation of the perforant path-evoked population spike (Lacaille and Harley, 1985). In 25% of these experiments the changes were long-lasting. Increases in the EPSP slope were seen in 50% of the experiments, but they did not account for the entirety of the population spike increases even in those experiments. In the brain slice, it was possible to identify the &receptor as the adrenergic receptor critical for NE's potentiation effect. Isoproterenol, a P-agonist, produced the population

w

NE, 100 nA

: 50 rn Sec.

C.

D.

25 min after B

E

'

I

\ Chart time 1 min

50 rnin after B

Fig. 1. Top panel (left). Schematic of the recording and iontophoretic site in the dentate gyrus and of the data display system. Top panel (right). Tracing of perforant path-evoked potential indicating measurement of the population spike amplitude. Panel A is a polygraph writeout of the control evoked potentials taken prior to the application of norepinephrine (NE) in panel B. Three to four min following initiation of N E iontophoresis there is a clear increase in population spike amplitude as seen in panel C. Panel E demonstrates the long-lasting nature of the potentiation. An increase in EPSP height is also observable. The 50 msec calibration is the "real-time" calibration for the field potential. (From Neuman and Harley, 1983.)

spike potentiation, and timolol, a &antagonist, blocked the occurrence of NE-induced potentiation. a,-Receptor activation had no effect or a depressant effect on the population spike.

Similarities of NE modulation and high-frequencyinduced L TP Sarvey's group replicated our observation of long-lasting facilitation by NE in the brain slice

309

14~r

-3

NE

T

I

9

I

15

I

24

T -

r

I

30

Time (min)

Fig. 2. Average potentiation of population spike amplitude in 16 observations of NE-induced long-lasting potentiation. The perforant path was stimulated once every 10 sec. Each point represents an average of the 10 preceding population spikes. Bars represent standard errors of the mean. (From Neuman and Harley, 1983.)

preparation and made several important additional observations. In their experiments 5-50 p M NE applied for 30 min invariably produced a long-lasting potentiation of the population spike. They also found, measuring the EPSP in the dendritic layer, that long-lasting EPSP slope increases consistently accompanied long-lasting potentiation of the population spike (Stanton and Sarvey, 1987). P-Receptor activation was again found to be the critical event for initiating both short- and long-lasting potentiation (Stanton and Sarvey, 1985a). Of particular interest is their series of slice experiments demonstrating critical links between high frequency-induced LTP and the long-lasting potentiation induced by NE. First, they provided evidence that high frequency-induced LTP is “virtually eliminated” in the dentate gyrus if NE is depleted by 6-OHDA (Stanton and Sarvey, 1985a). Similarly the P,-blocker, metoprolol, also eliminated high frequency-induced LTP as it had NE-induced LTP (Stanton and Sarvey, 1985a). Second, they showed that NE-induced long-lasting effects and high frequency-induced LTP can both be blocked by applications of protein synthesis inhibitors (Stanton and Sarvey, 1984,1985~) and that both initiate increases in CAMP concentrations (Stanton and Sarvey, 1985b). Third, and most importantly, they showed that application of

an NMDA receptor blocker prevents short and long-lasting NE-induced potentiation just as it prevents high frequency-induced LTP (Burgard et al., 1989; see Fig. 3). Bliss’s group has also provided evidence of a parallel between high frequency-induced and NE-induced potentiation. Both manipulations produce a significant increase in potassium-evoked glutamate release in the dentate gyrus (Lynch et al., 1985; Lynch and Bliss, 1986). The ability to block NE-induced potentiation and high frequency-induced LTP by similar manipulations, and in particular by the use of an NMDA blocker, suggests that these two forms of potentiation may converge on the same plasticity mechanism. Stanton and Heinemann (1986) demonstrated that one effect of NE in the slice is to increase the entry of calcium into cells of the dentate gyrus during a high-frequency stimulus. This effect was blocked by propranolol. Curiously 240

0 LTP or

200

-0

y

a E

NELLP

1 /IM ontogonist

180

a 10#M ontogonist

160

0

-E

140

0

120

a 0

m

.-“

100 80

0-APV HFT

0-APV

50gM

CPP

NE

Fig. 3. The effects on NMDA antagonists on long-term potentiation (LTP) and NELLP (norepinephrine-induced long-lasting potentiation). Open bars show potentiation of population spike amplitude produced by a high frequency stimulus train (HFT) or 50 p M NE. Black bars show the effects of a 1 Km concentration of either D( -)APV or CPP on the potentiation produced by HFT or NE. Hatched bars show t h e effect of 10 p M D(-)APV or CPP on potentiation. Each bar is the mean% baseline amplitude fS.E.M. of the population spike taken at the end of the final wash period. The number of experiments in each group appears within the bar. Asterisk denotes a significant difference ( P < 0.05, ANOVA plus Duncan’s test) in the mean compared to control potentiation. (From Burgard et al., 1989.)

310

the increase in calcium influx was not observed in the molecular layer, only in the cell layer. The entry of calcium is known to be pivotal in the cascade of events initiated by NMDA receptor activation (Sarvey et al., 1989). Gray and Johnston (1987) have shown that NE and p-agonists increase voltage-dependent calcium currents in whole clamped granule cells, while Radhakrishnan and Albuquerque (1989) have recently reported that propranolol reduces NMDA-activated currents. The model of NE action which emerges from the work of Sarvey and Stanton is one in which NE makes more probable the activation of NMDA receptors by single perforant path pulses. Lacaille and Schwartzkroin (1988) reported that one action of NE via @receptor activation is a relatively prolonged, but mild, depolarization of granule cell membranes accompanied by an increase in membrane resistance - probably mediated by a decreased potassium conductance. Stanton, Mody and Heinemann confirmed the observation of a p-mediated depolarization with increased membrane resistance for dentate granule cells (Stanton et al., 1989). In their report such depolarizations were long-lasting and could continue for more than an hour after NE washout. The long-lasting depolarizations were blocked by an NMDA receptor blocker, but a short-lasting depolarization was still seen. Since neither Sarvey nor Stanton see a short-lasting increase in the population spike induced by NE when an NMDA receptor blocker is present, it does not seem that the depolarization effect contributes to the enhanced population spike. It would appear that the actions of NE which are important for the enhanced population spike are those which enhance or activate NMDA currents. The depolarization effect would provide a condition favorable for NMDA receptor activation by glutamate pulses, in addition, as previously mentioned, NMDA currents would themselves be enhanced. One puzzling and possibly contradictory observation is the report that iontophoresed NE increases theta cell activity in the dentate gyrus

(Rose and Pang, 1989). Given the presumed inhibitory role of theta cells, one might expect a reduced probability of LTP-like effects with NE. Since this does not seem to be the case, it may be that cellular, not dendritic, inhibition is enhanced by NE’s effect on theta cells. It has been observed that cellular inhibition has little effect on the probability of LTP, although it would be expected to enhance the specificity of circuit output, while dendritic inhibition can effectively prevent LTP-like changes (Douglas, personal communication). Finally, in the Pang and Rose study, the effect of iontophoresing the P-agonist, isoproterenol, was an increase in granule cell firing, as predicted from the depolarizing effects of preceptor activation, as well as an increase in theta cell firing. If NE in the dentate gyrus acts as a facilitator of NMDA-induced plasticity then it is possible to argue that NE potentiation should be specific to synapses activated during the interval when NE is exerting its effects and that NE potentiation of the evoked potential participates in the Hebbian properties of LTP, ie., the requirement for temporal contiguity of synaptic inputs. Experiments have not yet demonstrated the synaptic specificity of NE potentiation. In an early attempt to look at specificity, using the slice preparation, we assessed the requirement for temporal contiguity by comparing potentiation induced by pairing perforant path activation and NE application with potentiation induced by applying NE during a period of no perforant path stimulation. Similar levels of potentiation were seen in the washout period in both experiments (Lacaille and Harley, 1985). However given the likelihood that NE-induced depolarization was still present at the onset of the washout, the results seemed inconclusive as regards the pairing specificity of NE potentiation. More recently, however, Dahl and Sarvey applied the P-agonist isoproteronol for 30 min and then waited for a 30 min washout period without stimulation; adrenergic potentiation of EPSP slope (population spikes were not evoked) was as great

31 1

as that seen when stimulation occurred at regular intervals throughout the period of NE agonist application (Dahl and Sarvey, 1990). This seems to contradict the hypothesis that NE long-lasting effects are dependent on NMDA receptor activation. Can NE enhance response to synaptic inputs for an extended period, whether or not they are active at the time of NE release? If so, how? The slice work provides support for the hypothesis that LTP and NE-induced potentiation are interdependent and converge on common mechanisms in the dentate gyrus, however a parallel story has not yet been developed in uiuo. Evidence for a role of NE in dentate gyrus LTP processes in uiuo is, at best, equivocal. Depletion of NE in viuo has been reported in one study to attenuate the EPSP slope increase produced by high-frequency stimulation in the dentate gyrus but not to prevent LTP of the population spike (Bliss et al., 1983). In a second study NE depletion affected baseline responding and appeared to increase LTP of the EPSP slope while reducing LTP of the population spike (Robinson and Racine, 1985). When the effects of depletion on baseline responses were taken into account, no overall effect of NE depletion on dentate gyrus LTP was observed. We have also failed to observe attenuation of LTP when NE release is blocked by clonidine (unpublished observations). This may not be a serious discrepancy since it appears more difficult to elicit dentate LTP in the slice than in uiuo. Facilitating factors, e.g., blockade of GABA inhibition (Wigstrom and Gustafsson, 1983) and, possibly, depolarization by NE may be required for reliable induction of dentate LTP in the slice. If N E operates to promote long-lasting synaptic change in the dentate gyrus by facilitating LTP, then it operates in conjunction with a number of potential facilitating events. NE’s role may be more or less important depending on those other events. NE has also been shown to play a facilitatory role in the LTP of EPSP slope in area CA3 of the hippocampus (Hopkins and Johnston, 1984), although NE alone is not sufficient in CA3 to

produce LTP-like changes if only single pulses are used as stimuli. This is a P-receptor-mediated effect and p-blockers prevent CA3 LTP from occurring in the slice. Interestingly, blockade of GABA inhibition overcomes this P-blocker effect and LTP can again be induced. Here LTP is not produced via an NMDA mechanism so that it is clear NE’s role in facilitating long-term change need not be restricted to NMDA synapses. Johnston et al. (1988) hypothesized that the increase in voltage-dependent calcium conductance which they have observed in granule cells and the decrease in calcium-activated potassium conductances which has also been reported to be produced by NE in granule and pyramidal cells (Haas and Rose, 1987; Madison and Nicoll, 1982) may act together in CA3 cells to promote a calcium induction of LTP. Winson and Dahl (1985) replicated our report of NE-induced long-lasting facilitation in uivo using 5 min of iontophoresed NE. The development of the effect in their figures parallels the development seen here in Figures 1 and 2. Winson and Dahl’s pharmacological effects in the dentate gyrus, however, indicate a much more complex pattern of receptor effects than that reported with bath-application of adrenergic blockers. In their study iontophoresis of the Pblocker, sotalol, at the cell body layer produced significant potentiation of the population spike. We had also observed this effect during our own iontophoretic studies (unpublished observations). In general a agonists, in their study, produced population spike increases and P-agonists produced decreases. This is the opposite of what had been reported with NE agonists and antagonists in the slice, and makes it clear that it will be necessary to evaluate the effects of synaptic NE release in order to understand the working system. NE, itself, only produced an effect at the cell body layer if iontophoresed for 5 min. The change seen, as already mentioned, was a long-lasting potentiation of the population spike. In all cases any effects on the population spike of NE ago-

312

nists and antagonists were only seen in a minority of experiments. This is reminiscent of Lacaille and Schwartzkroin’s (1988) experiments with microdrop NE application in the slice, where less than half of the NE applications produced changes. Such variability is generally ascribed to difficulties in appropriate localization of the applied substance. EPSP slope changes were common in the Winson and Dahl(1985) experiments, and these varied as a function of iontophoretic location. EPSP slope and amplitude were decreased if NE or its agonists and antagonists were iontophoresed in the mid-third of the dendritic region. The population spike was usually decreased by applications at that level. EPSP slope was increased if NE was iontophoresed directly above the cell body layer, but the population spike was still decreased. No consistent EPSP changes were seen with NE iontophoresed at the cell body layer (Winson and Dahl, 1985). Again, the question of interest is: which of these “localized” effects embodies the actual effect of NE release?

Selectivity of NE effects in dentate gyms A recent report of NE effects in the slice by Dahl and Sarvey (1989) is of particular interest when considering functional effects of NE release. They provide evidence that the potentiating action of NE is only seen for the evoked potential produced by stimulation of the medial perforant path, the lateral perforant path-evoked potential is significantly depressed by NE application. Both effects are produced by the P-agonist, isoproterenol and both can be blocked by propranolol or by an NMDA blocker. Given the diffuse nature of NE innervation in the hippocampus it is of some interest that NE modulation is so selective as to differentially effect glutamate synapses on the same dendritic tree. There is also other evidence that the lateral and medial pathways can be independently modulated. Naloxone has been reported to block LTP only on the lateral and not on the medial perforant path (Bramham et al., 1988). Most of the work on NE-induced

potentiation of the perforant path has been done on the medial perforant path potential. The observation of selective potentiating and depressant effects of NE on the two perforant path inputs should be investigated in viuo. Evoked potential potentiation by LC activation

We have been particularly concerned with whether or not release of NE by LC terminals in the dentate gyrus would produce the same short and LTP effects as had been seen with direct application of NE in this area.

Electrical stimulation of LC Early short reports of the effects of LC electrical stimulation suggested that, while single pulse LC stimulation had little discernable effect on the perforant path-evoked potential, trains of pulses were effective in producing potentiation of the dentate gyrus potential. In one study both the pre- and postsynaptic components of the evoked potential were enhanced (Bliss and Wendlandt, 1977); in the other study, in which recordings were made at the dendritic and the cell body layer, only the population spike, not the EPSP slope, was increased by LC stimulation (Assaf et al., 1979). Dahl and Winson provided the first full-length report of the effects of electrical stimulation in the LC area on the dentate gyrus-evoked potential (Dahl and Winson, 1985). In an earlier exploration of the brainstem with 1000 Hz triple pulses Winson had not observed any effects on the dentate gyrus potential with 200 p A 100 pS pulses in the LC region (Winson, 1980). In the 1985 study using a train of 6-12 50 Hz pulses which ended 50 msec before the perforant path pulse, Dahl and Winson found a potentiation of the population spike recorded at the cell body layer and a decrease of the EPSP recorded at the dendritic layer. The field EPSP measured at the cell layer was unaffected, but the field EPSP measured in the dendritic layer was consistently decreased. Dahl and Winson suggest that this indicates NE

.-. --7

165

2

313

a

spike amplitude spike onset nepsp slope

N.10

1L5

135-

E 125115-

5

0

2 165

{ 1351 E 125,

% 105:

lOmin

20min

30mln

LOmin

Time from LC stirnulotion onset

,r 7 1

: g 5 1 -6 20 .-"

75.1 Lton

p--...-~.e-

t

f -?--k - - y--j ~

1

LO 60 80 LC-PP IS1 i n milliseconds

Fig. 4. The lower graph illustrates potentiation of the population spike amplitude produced by a 15 msec burst of LC 333 H z stimulation at varying LC-PP interstimulus intervals (ISIS) from 20-80 msec ( n = 7). Note the absence of a potentiating effect on EPSP slope and population spike latency. The upper graph shows the pattern of LTP of the population spike amplitude in ten experiments in which 50 pairs of LC-PP stimulation were given using a 40 msec ISI. The baseline potential was based on the preceding 10 min of 0.1 Hz perforant path-evoked responses. Note the apparent decrease in EPSP slope during LC-PP pairing. Mean and standard errors for both sets of data are shown. (From Harley et al., 1989.)

may radically alter source and sink distribution in the granule cells. More than six pairings of LC and PP stimulation resulted in an attenuation of LC potentiation effects. Identification of LC stimulation effects was related to histological placement within 300 p M of the LC. One thousand (1000) Hz trains were again ineffective in producing these changes. We found that either the 15 msec 333 Hz LC train which had produced population spike potentiation in the study of Assaf et al. (1979) or the 10 Hz train employed by Bliss and Wendlandt (19771, was effective in producing a significant potentiation of the perforant path-evoked potential (Harley et al., 1989). As reported by Assaf et al. (1979) the optimal interval for maximum potentiation with the 5 pulse train required initiating stimulation of the LC 40 msec prior to activation of the perforant path. See the lower panel in Figure 4. This interval was consistent with a reported antidromic conduction time from the hip-

pocampus to the LC of 20-70 msec (Nakamura and Iwama, 1975). It should be mentioned as a caveat that a currently unidentified system running near the median raphe also produces dentate gyrus population spike potentiation with similar raphe-perforant path stimulation intervals (Assaf and Miller, 1978). This system has been shown to suppress granule cell firing during the period of enhancement. Not only did the short-term potentiation of the population spike match what we had seen with direct applications of N E but, of equal interest, it was also possible to produce LTP by repeatedly pairing electrical stimulation of the LC and activation of the perforant path. See the upper panel of Figure 4. In 60% of the experiments there was a significant decrease in EPSP slope, while in the remaining 40% there was either no change or an increase. Surprisingly, contralateral "LC" stimulation also appeared to produce population spike potentiation at a similar interval. Difficulties arose when we attempted to block the effects of electrical stimulation of the LC with the /?-blocker, propranolol. N o attenuation of the potentiation occurred (Fig. 5). Our difficulty was reminiscent of the earlier story of the rewarding effects of electrical stimulation in the vicinity of the LC. While rewarding effects could be shown to be rather well localized to the vicinity of the nucleus, and while electrodes supporting such

6or al

U 3

LC electrical

130

'r: -

a

6

120

2-

110

ln

100

a

90

Before

propmnolol

20 min

40 min

6 0 m i n after

I

Fig. 5 . The lack of effect of propranolol (20-30 mg/kg ip), a 0-receptor blocker, on the immediate potentiation effect produced by LC electrical stimulation 40 msec prior to perforant path activation. (From Harley et a l , 1989 )

314

I4O

tI

selective enough to allow us to study LC activation per se. We decided to change tactics and use micropipette ejections of glutamate to stimulate the nucleus more selectively.

N.20

Glutamate activation of locus coeruleus Glutamate activation of LC produces a clear potentiation of the perforant path-evoked population spike (Fig. 6; Harley and Milway, 1986). In some experiments the potentiation is long-lasting (Fig. 7). The effects of glutamate can be blocked by systemic propranolol (20-30 mg/kg) or by timolol or propranolol delivered by cannulae in the dentate gyrus (Harley and Evans, 1988). A direct comparison of glutamate ejection in the LC alternating with electrical stimulation from an electrode placed within 100 p m of the pipette tip can be seen in Figure 8. Both manipulations produce potentiation of the perforant path population spike, but only the glutamate activation effects are blocked by subsequent propranolol administration. It is, no doubt, possible to induce NE release by electrical stimulation in the vicinity of the LC nucleus. NE release in the forebrain has been demonstrated following periods of electrical stimulation in the vicinity of the LC (Tanaka et al.,

I II

GLU I

90 -5

I

0

I

10

5 Time (rnin)

Fig. 6. The effects of 100-150 nl 0.5 M glutamate ejected in the LC on the perforant path-evoked population spike in the dentate gyrus averaged over 20 animals. (From Harley and Milway, 1986.)

effects had been shown, in pioneering studies by Segal (Segal and Bloom, 1976a,b) to affect cell activity in the hippocampus, it has not been possible in subsequent work to demonstrate their dependence on the functioning of LC (Clavier et al., 1976; Corbett et al., 1977; Corbett and Wise, 1979). We were concerned that electrical stimulation in the fiber-rich LC area was simply not Mean Control Values

e

Population spike EPSP slope

amplitude 6 . 7 m v 3.2 mv/ms

Pop spike onset

latency

T M 18

3.4ms

.. . . . . -.. ... . .. 9

u o"\

120

80-

GLU

t

.

.

-*

.

.*

*

.

.. ..

a

.

5 mins

315 I LC Glutamate 1601

3

(4)-

eZa LC Electrical

140

0 c -

Q

E, 0

Y

130

'20 110

Q

Ln

100

90

Before 20min propranolol

40rnin

60 rnin a f t e r

Fig. 8. Effective propranolol (20-30 mg/kg iv) attenuation of the potentiation of population spike amplitude elicited by glutamate activation of the LC. At the same time potentiation produced by LC electrical stimulation was not attentuated by propranolol. Each animal was tested using a coupled electrode and micropipette assembly in the LC. Numbers in parentheses indicate animals used in each average. (From Harley et al., 1989.)

1976). Furthermore Washburn and Moises (1989) have recently reported, using the stimulation parameters of Assaf et al., (1979) that potentiation of the dentate gyrus population spike was blockable by systemic propranolol and could also be blocked by the a,-agonist, clonidine. These are encouraging results. We would prefer to use electrical stimulation to produce NE release in the dentate gyrus as it is a much easier methodology for the chronic studies which would enable us to investigate the time course of NE-induced long-term changes in the behaving animal. We have had some success in obtaining potentiation in awake animals with glutamate ejection near the LC via cannulae, but the technique is awkward and unsatisfactory for long-term study. In the Washburn and Moises (1989) study, they were careful to use LC currents which only produced half-maximal potentiation of the population spike. It is possible that our stimulation was too effective and that the NE release evoked competed successfully for binding sites with propranolol. Alternatively, our stimulation may have released peptides as well as NE, and the peptide release itself was adequate to produce potentiation. It has been shown that low-frequency stimu-

lation of monoamine pathways preferentially promotes monoamine release while high-frequency stimulation induces the additional release of peptides which often have a similar effect on the target cell (Lundberg et al., 1986). Only the monoamine effects are reduced by postsynaptic monoamine blockers. An important aspect of the Washburn and Moises study was their care in evaluating EPSP slope changes in the dendrites. They found no significant change in EPSP slope either in the dendrites or at the cell body layer. The issue of EPSP increases may be an important one for the hypothesis that NE promotes LTP effects, since increases in EPSP slope are generally considered a hallmark of LTP. While EPSP slope increases may be demonstrated reliably in the slice it is not clear that this is so with LC activation. In the two in vivo studies which have taken care to monitor dendritic EPSP slope changes, the site at which consistent EPSP increases have been reported in the slice, neither has found reliable increases in EPSP slope. It would appear that LC activation most frequently potentiates the coupling of the EPSP and population spike rather than potentiating the EPSP slope itself. Such enhanced population spike effects are also seen with frequency-induced LTP (Bliss and Gardner-Medwin, 1973; Taube and Schwartzkroin, 1988). They do not rule out the hypothesis that NE promotes LTP, but suggest that for some reason LC release of NE in the dentate favors increases in E-S coupling rather than in EPSP slope. This difference between LC effects in v i m and NE effects in the slice makes it particularly important to evaluate NMDA blockade effects on LC-induced potentiation. As another approach to a more selective LC stimulation paradigm we have taken advantage of the work (Ennis and Aston-Jones, 1986) showing excitatory inputs to the LC from the paragigantocellularis (PGi) region of the medulla. We have found that 2-4 0.5 msec pulses of 200 to 800 FA at 333 Hz in the PGi produce a clear potentiation of the dentate gyrus population spike at 30-40

316

PGI-PP spikes

PP spikes

Fig. 9.. The potentiating effect of two 0.5 msec pulses of paragigantocellularis stimulation at 800 F A and 333 Hz 30 msec prior to a perforant path (PP)pulse. This potentiating effect was completely blocked by 4 pg/kg clonidine, an a*blocker. Recovery of the stimulation effect occurred about 30 min after the injection of clonidine.

msec latencies (Fig. 91. Longer intervals are not effective. In the experiment shown, 4 pg/kg iv of clonidine blocked the PGi-induced potentiation (Babstock and Harley, unpublished observations). The effective latency range and the clonidine results are consistent with the hypothesis that the potentiating effects we observed are being mediated via LC activation and NE release in the dentate gyrus. Long-lasting effects of repeated PGI stimulation have also been observed. Locus coeruleus unit activity and dentate gyrus-evoked potential modulation

In other recent experiments on the relationship between LC activity and dentate gyrus-evoked potential modulation, Susan Sara and I have recorded from LC cells, using a variety of methods to increase or decrease LC firing, while concomitantly monitoring the perforant path-evoked potential (Harley and Sara, 1990). We found that silencing the LC with iv injections of clonidine had no observable effect on the dentate gyrusevoked potential. This was also reported by Washburn and Moises (1989) although they did not directly evaluate LC silence. Tail pinch, while briefly increasing LC activity, typically did not

modulate the evoked potential, however intravenous catheterization of the penile vein was associated at times with increases in population spike amplitude (see Fig. 11). This manipulation might be expected to produce more potent changes in LC activity. Injections of 50 nM glutamate in the LC characteristically produced a rush of firing followed by a loss of cell recording due, we have assumed, to depolarization block as has been described previously (Adams and Foote, 1988). In the experiment illustrated in Figure 11, cells appeared to continue to fire at elevated rates after the initial rush of activity; however, the height of the recorded waveforms diminished below the spike window of the ratemeter and the activity was not counted. The waveform shape appeared unchanged. Typically the rate of single unit firing after glutamate went from 1.2 Hz to 6-10 Hz for 10-20 sec and then we were unable to count the cell or cells accurately. Within 5-10 min after each glutamate ejection in Figure 10, we found nearby LC cells that had normal rates of activity. The oscillograph photographs (Fig. 10) show that the cells recorded prior to the time of glutamate ejection all met the criterion of a stereotyped excitation and inhibition pattern following tail pinch. It can also be seen that this particular LC recording displayed a rhythmic character prior to the second glutamate ejection which was also seen in several other experiments. The second glutamate ejection disrupted the rhythmicity but the rhythmicity returned within 5 min of the ejection. The pattern of acute excitement followed in a matter of minutes by a return to basal activity patterns was observed in all the effective glutamate experiments. These data suggest that the initial phasic increases in the dentate gyrusevoked potential depend on an abrupt and substantial release of NE in the hippocampus. The subsequent long-lasting increases in the population spike, once triggered, continue in the presence of basal LC activity, just as such increases can be observed following the washout of NE from the slice preparation. In general, in the

t

A

Perforant path potential each 10 SecS

~-

___

Fig. 10. Oscilloscope tracings of multiunit LC baseline activity recorded during the experiment of Figure 10. The lower trace illustrates the characteristic response to tail pinch seen with LC units. The upward line between the excitatory and inhibitory phases of the pinch response is an artifact. The upper trace is a recording of rhythmic activity observed at the same site. This activity was disrupted by glutamate ejections but returned within 5 min of the ejection.

glutamate experiments we were able to trigger both a short-term potentiation by increasing LC cell firing and a characteristic LTP which did not appear to depend on continued LC activity. The EPSP slope was increased in parallel in some, but not all. of these cases. Implications for hippocampal functioning: NE facilitation of synaptic plasticity

What is the relation of LC modulation of the dentate gyrus-evoked potential to the real world functioning of the hippocampus? Cellular recording evidence strongly supports the hypothesis that

7

miorant pain poieni~aieach 1osecs

IJpop"la"o~" splke=

Lc-cirachvlt;

. .-

Fig. 11. The effects of 50 flM ejections of glutamate in the LC o n the perforant path-evoked population spike and on multiunit cell firing in the LC recorded simultaneously. During the period shown in the top panel two glutamate ejections were made. Each evoked an increase in L C cell firing, as seen in the lower panels, which immediately preceded a significant increase in population spike amplitude. A third phasic increase in population spike amplitude was noted during an iv catherization procedure.

the hippocampus generates a cellular spatial map (O'Keefe and Nadel, 1978). Hippocampal cells have been shown to respond selectively to specific places in the environment. They apparently become coupled to a set of distal environmental cues such that any subset of those cues is effective in activating the map. The environmental cue information reaches the hippocampus by way of

318

the perforant path connections from the entorhinal cortex. McNaughton and Morris (1987) have proposed a simplified model of the operation of the spatial hippocampal network based on the assumption that the synaptic inputs carrying environmental information become coupled to a given set of hippocampal cells through a process of Hebbian association. Hippocampal neurons encode environmental information from the entorhinal cortex in a matrix of weak and strong synaptic connections. Given partial or complete environmental input into such a matrix, the appropriate cell firing pattern is generated. The synaptic weights are determined by the coupling of inputs with a set of cell activations from detonator synapses which provide the strong depolarization condition necessary for NMDAinduced long-term plasticity. The process of Hebbian association then generates increases in the synaptic weights of the concurrently active environmental inputs. The detonator concept seems unlikely since it presumes hardwired templates of cells which fire in a given environment. I would suggest that, in lieu of a set of prewired detonators, the hippocampus receives sets of facilitatory inputs which when active, singly or in concert, can increase the likelihood of strengthening a set of synaptic inputs activated by a particular environmental configuration. The representation of that input set then becomes, as postulated, fixed in the hippocampal matrix. The LC would function as one of those facilitatory systems, but there appear to be multiple facilitating inputs. Any diffuse input which, for example, reduced the probability of inhibition would increase the likelihood of NMDA-mediated changes in synaptic weights. It has been shown that other diffuse systems such as the massive supramammillary input to the dentate gyrus are also capable of potentiating perforant path input (Mizumori et al., 1989). It is not yet certain, at the cellular level, by what single or multiple mechanisms the LC promotes long-lasting modification of the hippocampal network.

Reduction of inhibition appears unlikely given the reports of increased activation of theta units by NE; on the other hand, as discussed earlier, NE does produce an increase in glutamate release, an increase in Ca’+currents, a reduction in Ca++-mediated K+currents and depolarization, all of which are cellular changes consistent with a greater probability of dentate response to input. While I have focused on the evidence for a role of the LC in promoting long-lasting potentiation of the glutamatergic perforant path input to the dentate gyrus, it appears from our work that a short-lasting potentiation, as in other areas of the central nervous system, is more likely with brief LC activation. It seems fitting that a mechanism which can produce transient potentiation of cellular responses to a putative sensory input should also, on an apparent continuum of degree or duration of activation, be linked to an increased future response to that input. The short term effect would presumably increase the intensity of experience of or promote attention to the input while the long-term effect could be viewed as the promotion of memory. The two processes of attention and memory have long been functionally linked in the psychological literature. It now appears that they are linked at the mechanistic level as well. These ideas were discussed two decades ago by Kety (Harley, 1987). The picture of LC function which emerges from the study of the dentate gyrusevoked potential supports the hypothesis that the LC is part of a system which increases the likelihood of the physiological changes which underly the functions “noticing” and “remembering.” It does not, however, argue that these events will not happen in the absence of the LC since even in current models there are diverse ways to enhance the functioning of glutamatergic synapses. Increases in LC activity appear to occur upon introduction into novel environments in particular (see Aston-Jones et al., Foote et al., and Sara and Segal, this volume) and might underly the enduringly enhanced population spike reported in novel, enriched environments by Sharp et al.

319

(1985) or the enduringly enhanced population spike observed during discriminative operant conditioning (Skelton el aL, 1987). Given only a facilitator role, it is probable that relatively subtle and specific behavioral assessment will be necessary to dissect the contribution of LC modulation to normal hippocampal functioning. Nonetheless this postulated role is consistent with the observation that loss of multiple facilitator systems, including the LC, as appears to occur in Alzheimer's disease, is devastating for memory function (McGeer, 1984). Finally, while I have suggested that diffuse systems may act, in general, as facilitators of long-term changes in the hippocampal network, it is increasingly clear that even in that structure the targets of NE action may be quite specific. Thus, for example, in the slice it appears that N E s effects on dentate gyrus circuitry may be restricted to potentiation of only the medial perforant path. Anatomical data suggest that the medial perforant path carries visual and auditory information from the environment while the lateral perforant path carries primarily olfactory input (Swanson et aZ., 1987). Interestingly, after ablation of the fornix which significantly disrupts subcortical inputs to the dorsal hippocampus, including those of the LC, place cell responding is dominated by local olfactory cues rather than distal environmental cues (Shapiro et al., 1989). Spatial behavior appears much less flexible in such animals. It is possible that NE acts to enhance distal environmental cue coupling to the hippocampal network while reduced NE activity, for example, may favor olfactory cue coupling. Interestingly, in that regard, blockers that interfere with noradrenergic and cholinergic inputs in the dentate, and which might now be expected to attenuate LTP processes in the medial perforant pathway, interfere with spatial learning (Gill et al., 1989); while blockers which interfere with the opiate system that promotes LTP in the lateral perforant pathway promote spatial learning (Decker et al., 1989). Do granule neurons function as members of two different and, possibly,

conflicting networks with participation in each being, in part, gated by LC input? The LC system, which has generally been regarded as diffuse and nonspecific, may have a selective and distinct role in promoting certain forms of coding in the hippocampal network while attenuating others. It is unlikely that multiple, diffuse systems will play synonymous roles in all forms of attentional and memorial processes, Acknowledgements

This research was supported by the Natural Sciences and Engineering Research Council of Canada and by the French Ministry of Research and Technology. References Adams, L.M. and Foote, S.L. (1988) Effects of locally infused pharmacological agents on spontaneous and sensoryevoked activity of locus coeruleus neurons. Bruin Res. Bull., 21: 395-400. Assaf, S.Y. and Miller, J.J. (1978) Neuronal transmission in the dentate gyrus: Role of inhibitory mechanisms. Brain Rex, 151: 587-592. Assaf, S.Y., Mason, S.T. and Miller, J.J. (1979) Noradrenergic modulation of neuronal transmission between the entorhinal cortex and the dentate gyrus of the rat. J. Physiol. (London), 292: 52P. Bliss, T.V.P. and Gardner-Medwin, A.R. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the unanaesthetized rabbit following stimulation of the perforant path. J. Physiol. (London), 232: 357-374. Bliss, T.V.P and L m o , T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J. Physiol. (London),232: 331-356. Bliss, T.V.P. and Wendlandt, S. (1977) Effects of stimulation of locus coeruleus on synaptic transmission in the hippocampus. Proc. Znt. Union Physiol. Sci., 13: 81. Bliss, T.V.P., Goddard, G.V. and Riives, M. (1983) Reduction of long-term potentiation in the dentate gyrus of the rat following selective depletion of monoamines. J. Physiol. (London), 334: 475-491. Bramham, C.R., Errington, M.L. and Bliss, T.V. (1988) Naloxone blocks the induction of long-term potentiation in the lateral but not in the medial perforant pathway in the anesthetized rat. Brain Res., 449: 352-356. Burgard, E.C., Decker, G. and Sarvey, J.M. (1989) NMDA receptor antagonists block norepinephrine-induced longlasting potentiation and long-term potentiation in rat dentate gyrus. Bruin Res., 482: 351-355.

320 Clavier, R.M., Fibiger, H.C. and Phillips, A.G. (1976) Evidence that self-stimulation of the region of the locus coeruleus does not depend upon noradrenergic projections to telencephalon. Brain Res., 113: 71-82. Corbett, D. and Wise, R.A. (1979) Intracranial self-stimulation in relation to the ascending noradrenergic fiber system of the pontine tegmentum and caudal midbrain: A moveable electrode mapping study. Brain Res., 177: 423426. Corbett, D., Skelton, R.W. and Wise, R.A. (1977) Dorsal noradrenergic bundle lesions fail to disrupt self-stimulation from the region of the locus coeruleus. Brain Res., 137: 37-44. Dahl, D. and Sarvey, J.M. (1989) Norepinephrine induces pathway-specific long-lasting potentiation and depression in the hippocampal dentate gyrus. Proc. Natl. Acad. Sci. USA, 86: 4776-4780. Dahl, D. and Sarvey, J.M. (1990) P-Adrenergic agonist-induced long-lasting synaptic modifications in hippocampal dentate gyrus require activation of NMDA receptors, but not electrical activation of afferents. Brain Res. (In press). Dahl, D. and Winson, J. (1985) Action of norepinephrine in the dentate gyrus. I. Stimulation of locus coeruleus. Exp. Brain Res., 59: 491-496. Decker, M.W., Introini-Collison, I.B. and McGaugh, J.L. (1989) Effects of naloxone on Morris water maze learning in th? rat: Enhanced acquisition with pretraining but not post-training administration. Psychobiology, 17: 270-275. Ennis, M. and Aston-Jones, G. (1986) A potent excitatory input to the nucleus locus coeruleus from the ventrolateral medulla. Neurosci. Lett., 71: 299-305. Gill, T.M., Decker, M.W. and McGaugh, J.L. (1989) Concurrent muscarinic and beta-adrenergic blockade impairs inhibitory (passive) avoidance and Morris water maze performance. SOC.Neurosci. Abstr., 15: 733. Gray, R. and Johnston, D. (1987) Noradrenaline and betaadrenoceptor agonists increase activity of voltage-dependent calcium channels in hippocampal neurons. Nature (London), 327: 620-622. Haas, H.L. and Rose, G.M. (1987) Noradrenaline blocks potassium conductance in rat dentate granule cells in uitro. Neurosci. Lett., 78: 171-174. Harley, C.W. (1987) A role for norepinephrine in arousal, emotion and learning? Limbic modulation by norepinephrine and the Kety hypothesis. Prog. Neuro-Psychopharmacol. Biol. Psychiatty., 11: 419-458. Harley, C.W. and Evans, S. (1988) Locus-coeruleus-induced enhancement of the perforant-path evoked potential: Effects of intradentate beta blockers. In C.D. Woody, D.L. Alkon and J.L. McGaugh (Eds.), Cellular Mechanisms of Conditioning and Behavioral Plasticity, Plenum, New York, pp. 415-423. Harley, C.W. and Milway, J.S. (1986) Glutamate ejection in the locus coeruleus enhances the perforant path-evoked population spike in the dentate gyrus. Exp. Brain Res., 63: 143-150. Harley, C.W. and Sara, S.J. (1990) Locus coeruleus cell activ-

ity and potentiation of the perforant path evoked dentate gyrus population spike. Sac. Neurosci. Abstr., 16: 264 Harley, C., Milway, J.S. and Lacaille, J.C. (1989) Locus coeruleus potentiation of dentate gyrus responses: Evidence for two systems. Brain Res. Bull., 22: 643-650. Hopkins, W.F. and Johnston, D. (1984) Frequency-dependent noradrenergic modulation of long-term potentiation in the hippocampus. Science, 226: 350-352. Johnston, D., Hopkins, W.F. and Gray, R. (1988) Noradrenergic enhancement of long-term synaptic potentiation. In P.W. Landfield and S.A. Deadwyler (Eds.) Long-Term Potentiation: From Biophysics to Behauior. Alan R. Liss, New York, pp. 355-376. Lacaille, J-C. and Harley, C.W. (1985) The action of norepinephrine in the dentate gyrus: Beta-mediated facilitation of evoked potentials in uitro. Brain Res., 358: 210-220. Lacaille, J-C. and Schwartzkroin, P.A. (1988) Intracellular responses of rat hippocampal granule cells in iitro to discrete application of norepinephrine. Neurosci. Lett., 89: 176-181. Lundberg, J.M., Rudehill, A,, Sollevi, A,, TheodorssonNorheim, E. and Hamberger, B. (1986) Frequency-dependent and reserpine-dependent chemical coding of sympathetic neurotransmission-Differential release of noradrenaline and neuropeptide Y from pig spleen. Neurosci. Lett., 109: 341-348. Lynch, M.A. and Bliss, T.V.P. (1986) Noradrenaline modulates the release of [14C] glutamate from dentate but not from CAl/CA3 slices of rat hippocampus. Neuropharmacology, 25: 171-176. Lynch, M.A., Errington, M.L. and Bliss, T.V.P. (1985) Longterm potentiation of synaptic transmission in t h e dentate gryus: Increased release of [14C]glutamate without increase in receptor binding. Neurosci. Lett., 62: 123-129. Madison, D.V. and Nicoll, R.A. (1982) Noradrenaline blocks accomodation of pyramidal cell discharges in the hippocampus. Nature (London), 299: 636-638. McGeer, P.L. (1984) The 12th J.A.F. Stevenson Memorial Lecture: Aging, Alzheimer’s disease, and the cholinergic system. Can. J. Physiol. Pharmacol., 62: 741-754. McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system. TINS, 10: 408-415. Mizumori, S.J.Y., McNaughton, B.L. and Barnes, C.A. (1989) A comparison of supramammillary and medial septa1 influences on hippocampal field potentials and single unit activity. J. Neurophysiol., 61: 1-17. Nakamura, S. and Iwama, K. (1975) Antidromic activation of the rat locus coeruleus neurons from hippocampus, cerebral and cerebellar cortices. Brain Rex, 99: 372-376. Neuman, R.S. and Harley, C.W. (1983) Long-lasting potentiation of the dentate gyrus population spike by norepinephrine. Brain Res., 273: 162-165. O’Keefe, J. and Nadel, L. (1978) The Hippocampus as a Cognitive Map, Clarendon Press, Oxford, 570 pp. Radhakrishnan, V. and Albuquerque, E.X. (1989) Effect of beta adrenergic blockers on N-methyl-D-aspartate

321 (NMDAhctivated channels of rat hippocampal neurons. Soc. Neurosci. Abstr., 15: 828. Robinson, G.B. and Racine, R.J. (1985) Long-term potentiation in the dentate-gyrus: Effects of noradrenaline depletion in the awake rat. Brain Res., 325: 71-78. Rose, G.M. and Pang, K.C.H. (1989) Differential effect of norepinephrine upon granule cells and interneurons in the dentate gyrus. Brain Res., 488: 353-356. Sarvey, J.M., Burgard, E.C. and Decker, G. (1989) Long-term potentiation: Studies in the hippocampal slice. J. Neurosci. Methods, 28: 109-124. Segal, M. and Bloom, F.E. (1976a) The action of norepinephrine in the rat hippocampus. 111. Hippocampal cellular response to locus coeruleus stimulation in the awake rat. Brain Rex, 107: 499-511. Segal, M. and Bloom, F.E. (1976b) The action of norepinephrine in the rat hippocampus. IV. The effects of locus coeruleus stimulation on evoked hippocampal unit activity. Brain Res., 107: 513-525. Shapiro, M.L., Simon, D.K., Olton, D.S., Gage, F.H., 111, Nilsson, 0. and Bjorklund, A. (1989) Intrahippocampal grafts of fetal basal forebrain tissue alter place fields in the hippocampus of rats with fimbria-fornix lesions. Neuroscience, 32: 1- 18. Sharp, P.E., McNaughton, B.L. and Barnes, C.A. (1985) Enhancement of hippocampal field potentials in rats exposed to a novel, complex environment. Brain Rex, 339: 361-365. Skelton, R.W., Scarth, A.S., Wilkie, D.M., Miller, J.J. and Philips, A.G. (1987) Long-term increases in dentate granule cell responsivity accompany operant conditioning. J. Neurosci., 7: 3081-3087. Stanton, P.K. and Heinemann, U. (1986) Norepinephrine enhances stimulus-evoked calcium and potassium concentration changes in dentate granule cell layer. Neurosci. Lett., 67: 233-238. Stanton. P.K. and Sarvey, J.M. (1984) Blockade of long-term potentiation in rat hippocampal CA1 region by inhibitors of protein synthesis. J. Neurosci., 4: 3080-3088. Stanton, P.K. and Sarvey, J.M. (1985a) Depletion of norepinephrine, but not serotonin, reduces long-term potentiation in the dentate gyrus of rat hippocampal slices. J. Neurosci., 8: 2169-2176. Stanton, P.K. and Sarvey, J.M. (1985b) The effect of highfrequency electrical stimulation and norepinephrine on

cyclic AMP levels in normal versus norepinephrine-depleted rat hippocampal slices. Brain Res., 358: 343-348. Stanton, P.K. and Sarvey, J.M. (1985~)Blockade of norepinephrine-induced long-lasting potentiation in the hippocampal dentate gyrus by an inhibitor of protein synthesis. Brain Res., 361: 276-283. Stanton, P.K. and Sarvey, J.M. (1987) Norepinephrine regulates long-term potentiation of both the population spike and dendritic EPSP in hippocampal dentate gyrus. Brain Res. Bull., 18: 115-119. Stanton, P.K., Mody, I. and Heinemann, U. (1989) A role for N-methyl-D-aspartate receptors in norepinephrine-induced long-lasting potentiation in the dentate gyrus. Exp. Brain Res., 77: 517-530. Swanson, L.W., Kohler, C. and Bjorklund, A. (1987) The limbic region. I. The septohippocampal system. In A. Bjorklund, T. Hokfelt and L.W. Swanson (Eds.), Handbook of Chemical Neuroanatomy, Vol. 5, Integrated systems of the CNS, Part I , Elsevier, Amsterdam, pp. 125-227. Tanaka, C., Inagaki, C. and Fujiwara, H. (1976) Labeled noradrenaline release from rat cortex following electrical stimulation of locus coeruleus. Brain Res., 106: 384-389. Taube, J.S. and Schwartzkroin, P.A. (1988) Mechanisms of long-term potentiation: EPSP/spike dissociation, intradendritic recordings, and glutamate sensitivity. J. Neurosci., 8: 1632-1644. Washburn, M. and Moises, H.C. (1989) Electrophysiological correlates of presynaptic alpha 2-receptor-mediated inhibition of norepinephrine release at locus coeruleus synapses in dentate gyrus. J. Neurosci., 9: 2131-2140. Wigstrom, H. and Gustafsson, B. (1983) Large long-lasting potentiation in the dentate gyrus in uitro during blockade of inhibition. Brain Res., 275: 153-158. Wigstrom, H. and Gustafsson, B. (1985) On a long-lasting potentiation in the hippocampus: A proposed mechanism for its dependence on coincident pre- and postsynaptic activity. Acta Physiol. Scand., 123: 519-522. Winson, J. (1980) Influence of raphe nuclei on neuronal transmission from perforant path through dentate gyrus. J. Neurophysiol., 44: 937-950. Winson, J. and Dahl, D. (1985) Action of norepinephrine in the dentate gyrus. 11. Iontophoretic studies. Exp. Brain R ~ s .59: , 497-506.