Memory and the Brain: Unexpected Chemistries and a New Pharmacology

NEUROBIOLOGY OF LEARNING AND MEMORY 70, 82–100 (1998)ARTICLE NO. NL983840

Memory and the Brain: Unexpected Chemistriesand a New Pharmacology

Gary Lynch

University of California, Irvine, California 92697–3800

Efforts to characterize long-term potentiation (LTP) and to identify its substrateshave led to the discovery of novel synaptic chemistries, computational algorithms, and,most recently, pharmacologies. Progress has also been made in using LTP to developa ‘‘standard model’’ of how unusual, but physiologically plausible, levels of afferentactivity create lasting changes in the operating characteristics of synapses in the corti-cal telencephalon. Hypotheses of this type typically distinguish induction, expression,and consolidation stages in the formation of LTP. Induction involves a sequence con-sisting of theta-type rhythmic activity, suppression of inhibitory currents, intense syn-aptic depolarization, NMDA receptor activation, and calcium influx into dendriticspines. Calcium-dependent lipases, kinases, and proteases have been implicated inLTP induction. Regarding the last group, it has been recently reported that thetapattern stimulation activates calpain and that translational suppression of the proteaseblocks potentiation. It is thus likely that proteolysis is readily driven by synaptic activ-ity and contributes to structural reorganization. LTP does not interact with treatmentsthat affect transmitter release, has a markedly differential effect on the currents medi-ated by colocalized AMPA vs NMDA synaptic receptors, changes the waveform of thesynaptic current, modifies the effects of drugs that modulate AMPA receptors, and issensitive to the subunit composition of those receptors. These results indicate thatLTP is expressed by changes in AMPA receptor operations. LTP is accompanied bymodifications in the anatomy of synapses and spines, something which accounts for itsextreme duration (weeks). As with various types of memory, LTP requires about 30min to consolidate (become resistant to disruption). Consolidation involves adhesionchemistries and, in particular, activation of integrins, a class of transmembrane recep-tors that control morphology in numerous cell types. Platelet activating factor andadenosine may contribute to consolidation by regulating the engagement of latent inte-grins. How consolidation stabilizes LTP expression is a topic of intense investigationbut probably involves modifications to one or more of the following: membrane environ-ment of AMPA receptors; access of regulatory proteins (e.g., kinases, proteases) to thereceptors; receptor clustering; and space available for receptor insertion. Attempts toenhance LTP have focused on the induction phase and resulted in a class of centrallyactive drugs (‘‘ampakines’’) that positively modulate AMPA receptors. These compoundspromote LTP in vivo and improve the encoding of variety of memory types in animals.Positive results have also been obtained in preliminary studies with humans. q 1998

Academic Press

Efforts to understand the cellular substrates of long-term potentiation (LTP)have been motivated by the assumption that explanations for LTP would ac-count for how memories are stored. Less apparent, but also important as a

Address requests for reprints and correspondence to Gary Lynch, CNLM, University of Califor-nia, Irvine, Irvine, CA 92697–3800.

821074-7427/98 $25.00Copyright q 1998 by Academic PressAll rights of reproduction in any form reserved.

AID NLM 3840 / 6v16$$$341 09-14-98 08:47:40 nlmoa AP: NLM

83MEMORY AND THE BRAIN

rationale, was the idea that insight into the mechanisms underlying potentia-tion would eventually lead to new types of brain chemistries and novel classesof drugs. LTP is exotic by almost any physiological criterion and it is notunreasonable to expect that its substrates and pharmacologies are as wellunusual. A third, and later developing, motive for studying LTP was the expec-tation of obtaining empirically derived ‘‘synaptic learning rules.’’ Simulationsincorporating varying degrees of biological realism had demonstrated that therequirements for changing synaptic strength (e.g., the ‘‘Hebb rule’’) dominatethe computational capabilities of networks. From this it was a small step to theidea that detailed descriptions of the spatiotemporal patterns of physiologicalactivity that induce LTP would help explain the computations carried out bylocal circuits in the cortical telencephalon.

Much of the current work on LTP reflects the above points and falls into threecategories. The first is concerned with the analysis of the synaptic changesresponsible for the potentiation effect. Developments have reached a pointwhere an outline of how the LTP is induced, expressed, and maintained iswithin reach. This is not to say that all questions are about to be answered,but rather that it is now possible to construct models that largely satisfy thetheoretical constraints imposed by the diverse and peculiar phenomenology ofLTP. A second and much smaller body of work has evolved around the idea ofusing information about the biochemistry of LTP to guide the development ofmemory-enhancing drugs. At least one new class of centrally active drugshas emerged from this effort. These compounds (‘‘ampakines’’) enhance fast,excitatory transmission in brain, promote the formation of LTP, and selectivelyenhance several varieties of memory. Finally, several rules relating patternsof afferent activity to the induction of LTP have been specified. These impartinteresting, and in some cases unexpected, computational properties to simula-tions of networks found in hippocampus or cortex. Indeed, algorithms extractedfrom the simulations plus LTP-based learning rules have proven to be competi-tive with traditional approaches in dealing with certain classes of problems.

This article reviews a limited portion of the work on the substrates of LTPand attempts to use information about them in designing neuropharmaceuti-cals. A recent review of the computational work that has grown out of LTP,which is not discussed here, can be found in Granger, Myers, Whelpley, andLynch (1995).

INDUCTION AND EXPRESSION OF LONG-TERM POTENTIATION

A great deal of progress has been made in unraveling the mechanisms re-sponsible for the induction of LTP by naturalistic patterns of brain stimulationand in identifying the synaptic changes responsible for its expression. Keydiscoveries include the following:

1. A deep relationship exists between the brain’s theta rhythm—a patternproposed to be key to learning (see Vertes & Kocsis, 1997, for an recent andinteresting review)—and the machinery that induces LTP (Larson, Wong, &Lynch, 1986; Larson & Lynch, 1986); this involves a peculiar refractory periodexhibited by inhibitory connections (Larson & Lynch, 1986, 1988; Mott &Lewis, 1991).

2. Induction of LTP involves intense depolarization followed by activation


84 GARY LYNCH

of NMDA-type glutamate receptors (Collingridge, Kehl, & McLennan, 1983)and then an increase in postsynaptic calcium (Lynch, Larson, Kelso, Barrio-nuevo, & Schottler, 1983).

3. A combination of pharmacological, molecular, and immunocytochemicalstudies indicate that the calcium-activated protease calpain links calcium in-flux, the trigger for LTP, to lasting synaptic modifications.

While once disputed, it is now widely accepted that calpain is present in thebrain and is an agent in several forms of brain pathology. The mounting bodyof evidence on these points has also lent indirect support to the hypothesisthat stimulation of calpain is the threshold step in the cascade leading to LTP.However, two technically challenging questions needed to be resolved beforecontinuing beyond the hypothesis: (a) Does selective suppression of calpainactivity block the induction of LTP? (Drugs that block calpain block LTP butinfluence a range of proteases.) (b) Are the very brief bursts of afferent activityused to induce LTP sufficient to activate calpain? Results relevant to bothissues have recently become available.

Selective depletion of calpain selectively depresses the capacity of synapsesto generate stable LTP. Translational suppression with antisense oligonucleo-tides was used to reduce calpain concentrations in cultured hippocampal slicesto about 50% of normal values. This was shown to substantially lower calpain-mediated proteolysis in the slices. There were no detectable differences in thebaseline physiologies of control and calpain-depleted slices, a not unexpectedresult given that calpain’s activation requirements are such that the enzymemust be inactive except for very unusual circumstances. Percentage LTP wasgreatly reduced in the depleted slices compared with controls despite the ap-parently normal expression of transient forms of synaptic potentiation (Van-derklish, Bednarski, & Lynch, 1996). An earlier series of experiments showedthat depletion of calpain decreases the vulnerability of hippocampal slicesto the pathophysiological effects of excessive stimulation of NMDA receptor(Bednarski, Gall, Lynch, & Vanderklish, 1995). Together these studies linkcalpain to the production of LTP and provide a biochemical basis for the generalhypothesis that plasticity and pathology are points on a continuum.

Theta-pattern stimulation activates calpain. The discovery that a fragmentof the structural protein spectrin is a marker for sites at which calpain hadactivated (Siman, Baudry, & Lynch, 1984; Seubert, Baudry, Dudek, & Lynch,1987) opened the way to tests of where in brain the enzyme is triggered byvarious manipulations. With the advent of antibodies selective to the fragment(Roberts-Lewis, Savage, Marcy, Pinsker, & Siman, 1994; Saido, Yokota, Nagao,Yamaura, Tani, Tsuchiya, Suzuki, & Kawasaki, 1993), mapping calpain activ-ity has become a routine practice in a growing number of laboratories. Thiswas extended to a direct test of whether the minimal synaptic stimulationneeded to induce LTP would be sufficient to stimulate the protease. Studiesappropriate to this question revealed that two theta bursts (four pulses in a30-ms burst repeated over a 200-ms interval) caused the appearance of thespectrin fragment in the dendritic zones containing the pertinent synapses(Vanderklish, Saido, Gall, Arai, & Lynch, 1995).

4. Robust LTP is readily produced without changing (i) the frequency prop-erties of excitatory synapses (Muller and Lynch, 1989), (ii) the effects of treat-



ments that increase the probability of release (Muller, Joly, & Lynch, 1988a),or (iii) the currents passed by postsynaptic NMDA receptors (Muller, Turnbull,Baudry, & Lynch, 1988b; Kauer, Malenka, & Nicoll, 1988). These widely repli-cated results indicate that LTP does not require changes in transmitter release.

5. Tests of the hypothesis that LTP is due to a change in the resistance ofdendritic spines have proven negative (Jung, Larson, & Lynch, 1991; Larson &Lynch, 1991); by exclusion, this and the immediately preceding point arguefor the idea (Lynch & Baudry, 1984) that postsynaptic glutamate receptorsare the agents of LTP expression.

6. Changing the subunit composition of AMPA receptors changes the expres-sion of LTP (Vanderklish, Neve, Bahr, Arai, Hennegriff, Larson, & Lynch,1992), as expected if the receptors are the agents of expression.

7. LTP is accompanied by changes in the waveform of synaptic responses(Abros-Ingerson, Larson, Xiao, & Lynch, 1991; Abros-Ingerson, Xiao, Larson, &Lynch, 1993) and by alterations in the pharmacology of AMPA receptors (Xiao,Bahr, Staubli, Vanderklish, & Lynch, 1991; Staubli, Ambros-Ingerson, &Lynch, 1992; Kolta, Lynch, & Ambros-Ingerson, 1998), findings that were alsoanticipated from the argument that expression is achieved by modifying recep-tor operating characteristics. The first result has been confirmed by two groups(Stricker, Field, & Redman, 1996; Xie, Liau, Baudry, & Berger, 1997); argu-ments continue about the second (Asztely, Hans, Wigstrom, & Gustatsson,1992) but are likely to be resolved with the arrival of powerful new modulatorsof AMPA receptors (see below).

Modeling studies have reproduced the observed changes in waveform, phar-macological sensitivities, and synaptic currents that define LTP by simplyincreasing the open/closing rates of simulated AMPA receptors (Ambros-Inger-son & Lynch, 1993). A more recent effort suggests that clustering of the recep-tors could account for the waveform and size effects (Xie et al., 1997). Testswith a broader set of AMPA receptor modulators affecting different receptorrate constants will likely provide the phenomenological constraints neededfor choosing between receptor biophysical versus receptor distribution/numberhypotheses of expression.

CONSOLIDATION OF LONG-TERM POTENTIATION

As with memory, LTP takes several minutes to consolidate or become resis-tant to disruption (Barrionuevo, Schottler, & Lynch, 1980). As more is learnedabout induction and expression, the question of what goes on in those fewminutes to stabilize potentiation has received increasing attention. Existingclues about the nature of consolidation are as follows:

8. LTP is accompanied by changes in the anatomy of synapses and spines(Lee, Schottler, Oliver, & Lynch, 1981; Lee, Oliver, Schottler, & Lynch, 1981;Chang & Greenough, 1984; Desmond & Levy, 1986; and others). This oftenrepeated result accounts for the extraordinary stability of potentiation: macro-structures can persist indefinitely in the face of continual replacement of theirconstituent proteins.

9. LTP induction is facilitated by stimulation of platelet-activating factorreceptors (Arai & Lynch, 1992; Bazan, Packard, Teather, & Allan, 1997) andreversed by stimulation of adenosine receptors (Arai, Kessler, & Lynch, 1990).


86 GARY LYNCH

10. Hypoxia reverses LTP if applied within minutes of induction but iswithout effect if administered 30 min after induction (Arai, Larson, & Lynch,1990).

11. LTP is erased by 5-Hz afferent stimulation applied immediately afterinduction (Barrionuevo et al., 1980; Staubli & Lynch, 1990; Larson, Xiao, &Lynch, 1993); the efficacy of stimulation in reversing potentiation decreasesto zero by 30 min after induction (Staubli & Chun, 1996a, 1996b).

12. Reversal of LTP is mediated by adenosine receptors (Larson et al., 1993;Staubli & Chun, 1996b; Abraham & Hugget, 1997). This ties in with the preced-ing point because 5-Hz stimulation releases adenosine (Cunha, Vizi, Ribeiro, &Sebastiao, 1996).

13. Agents that interfere with the extracellular interactions of neural celladhesion molecules (NCAMs) block the formation of stable LTP but do notreverse already induced potentiation (Luthi, Laurent, Figurov, Muller, &Schachner, 1994; Rønn, Bock, Linnemann, & Jahnsen, 1995).

14. Antagonists of integrins, adhesion receptors of a type very different thanNCAMs, block the stabilization of LTP (Staubli, Vanderklish, & Lynch, 1990;Xiao, Staubli, Kessler, & Lynch, 1991; Bahr, Staubli, Xiao, Chun, Esteban, &Lynch, 1997).

Integrins are dimeric (a, b) transmembrane adhesion receptors that figureprominently in cell–cell and cell–matrix relationships throughout the body.Surprisingly little is known about the types and functions of integrins in adultbrain. Work from this laboratory demonstrated that receptors with integrinepitopes, and that bind to a consensus sequence in matrix targets for a subclassof integrins, are concentrated in forebrain synapses (Bahr & Lynch, 1992; Bahret al., 1997). Light and electron microscopic immunocytochemical experimentsfrom other groups then showed that b1 subunits are present in hippocampus(Grooms, Terracio, & Jones, 1993; Paulus, Baur, Schuppan, & Roggendorf,1993) and that a8 subunits are highly concentrated in hippocampal synapses(Einheber, Schnapp, Salzer, Cappiello, & Milner, 1996).

A large subclass of integrins recognize their extracellular target proteinsvia the amino acid sequence RGD (arginine–glycine–aspartate) (Piersch-bacher and Ruoslahti, 1984). Binding in these instances can be reduced bysmall, competitive peptides (e.g., GRGDSP). Infusion of diverse peptide antago-nists of RGD binding into hippocampal slices had no evident effects on synapticpotentials or the complex physiological responses to theta bursts (Staubli etal., 1990; Xiao et al., 1991; Bahr et al., 1997). It did, however, prevent theformation of stable LTP; that is, potentiation induced in the presence of thecompounds steadily decreased during the 1 h following its induction. The dosedependency of this effect corresponded to that for peptide suppression of inte-grin-mediated adhesion in various tissues (Cardwell & Rome, 1988; Haskel &Abendschein, 1989). Similar-sized peptides with no relationship to the RGDsite did not interact with LTP (Xiao et al., 1991; Bahr et al., 1997). A recentreport showed that integrin RGD antagonists markedly affect the expressionof kindled seizures in slices while having no effect on baseline physiology(Grooms & Jones, 1997). This provides a second instance of integrins contribut-ing to the long-lasting changes elicited by bursts of high-frequency afferentactivity.

15. Integrin antagonists reverse LTP.



Integrins in many peripheral tissues exist in a quiescent state requiring anactivation event, often triggered by other types of transmembrane receptors,for their adhesive properties to appear (Newton et al., 1997, for a recent re-view). The conversion from a latent to an active state can involve severalminutes (van Willigen, Hers, Gorter, & Akkerman, 1996; Newton, Thiel, &Hogg, 1997 for examples) and it is thus possible that the protracted consolida-tion period for LTP, and by inference possibly memory, reflects the time neededto mobilize adhesion receptors. Assuming that integrins contribute to consoli-dation, it can be expected that the peptide antagonists would (i) reverse LTPin a time-dependent fashion, and (ii) be effective over the same period as arehypoxia and low-frequency synaptic activity (Arai et al., 1990b; Staubli &Chun, 1996a, 1996b). These predictions were recently tested by microperfusingan integrin antagonist to one of two recording sites in hippocampal slices atvarious times before and after LTP was simultaneously induced at both sites.Applications 10 min before or up to 10 min after induction caused LTP at theexperimental site to decay steadily relative to that at the within slice controlsite. However, microperfusion at later points after induction had no detectableeffect on potentiation (Staubli et al., submitted).

The above results constitute strong evidence that integrin activation is acritical step in the consolidation of LTP. The close correspondence betweenthe time courses for LTP reversal by the antagonists versus low-frequencystimulation also strongly suggest that the two effects are linked, i.e., thatsynaptic activity erases LTP by deactivating or blocking integrins.

Integrins and the correlates of consolidation. The hypothesis that activationof latent integrins consolidates LTP is conservative in that it invokes biologicalphenomena known to occur in several types of cells in a variety of circum-stances calling for lasting changes. It is also the case that integrin activationleads to cytoskeletal reorganization (cf. Clarke & Hynes, 1996) and morphologi-cal changes and thus could account for the stable shifts in the anatomy ofsynapses and spines that accompany LTP. These anatomical effects, actingpresumably through their influences on local chemistries, provide by far themost plausible explanation for the extreme persistence of LTP. Note also thatintegrin activation, and the cytoskeletal events following on it, occur withinthe same minutes-long time frame as LTP consolidation. Finally, the integrinhypothesis ties together a variety of seemingly unrelated findings regardingthe stabilization of LTP. Thus, for example, consolidation requires stimulationof platelet-activating factor (PAF) receptors and is reversed by stimulation ofadenosine receptors (see point 9 above), two receptor events that have noevident relationship other than activation (PAF; van Willigen et al., 1996 fora recent study) and deactivation (adenosine; see Thiel, Chambers, Chouker,Fischer, Zourelidis, Bardenhauer, Arfors, & Peter, 1996) of integrins.

Missing links between induction and consolidation of stable LTP. Produc-tion of lasting (consolidated) potentiation appears to involve calcium-activatedenzymes and adhesion receptors. How are these connected? If, as proposed,integrin activation stabilizes new synaptic configurations, then it follows thatsome event must have destabilized the earlier organization. This could be therole played by calpain. Partial digestion of spectrin and other structural pro-teins should have the effect of disassembling the submembrane cytoskeletonand thus the intracellular anchors on one side of the terminal/spine adhesive


88 GARY LYNCH

junction. Calpain may also act directly on adhesion receptors: the intracellulardomain of NCAM 180 (Sheppard, Wu, Rutishauser, & Lynch, 1991) and certainintegrins (cf. Du, Saido, Tsubuki, Indig, Williams, & Ginsberg, 1995) are sub-strates for the protease. It is not unreasonable to assume that breaking downcomponents of the adhesion complex (matrix proteins, transmembrane recep-tors, complexes of proteins organized around cytoplasmic tail of the receptors,crosslinking proteins, submembrane cytoskeleton, etc.) would relax normallypresent constraints on the morphology of the synapse.

It is important to emphasize that the above arguments are concerned withpotentiation lasting for weeks; the discussed mechanisms are probably notneeded for those versions of LTP lasting hours. Phospholipases (Massicotte,Vanderklish, Lynch, & Baudry, 1991) and protein kinases (Finn, Browning, &Lynch, 1980; Barria, Muller, Derkach, Griffith, & Soderling, 1997; Hayashi,Ishida, Katagiri, Mishina, Fujisawa, Manabe, & Takahashi, 1997; Strack,Choi, Lovinger, & Colbran, 1997; Lau & Huganir, 1995) have been implicatedin LTP and by themselves could account for transient potentiation. Even withthis restriction, the above arguments are general in nature and leave unad-dressed a long list of specifics. Important among these is the question of whattypes of NCAM changes are required for stable LTP to develop. Are the cell–cell adhesive interactions mediated by individual CAMs temporally suspendedor are the receptors cleaved and then later replaced? If the latter, what arethe proteases involved? There is also no information on whether the matrixin the synaptic region is left unchanged during LTP formation or instead isreshaped so as to contribute to the maintenance of new shapes. Experimentsaddressing certain of these issues are in progress. Beyond this, there are goodreasons to assume that advances in our understanding of how adhesion junc-tions are maintained and modified will point the way to a more completeexplanation for the remarkable persistence of LTP.

REGIONAL VARIATIONS IN LTP-RELATED ADHESION CHEMISTRIES

Given the evidence that integrins are located in adult synapses, and in partaccount for the peculiarly protracted process of LTP consolidation, it becomesof interest to ask if subunit composition of integrins varies, across brain regionsand cell types. There are at least 11a and 7 b subunits that occur with sufficientfrequency to be candidates for elements of brain integrin dimers. The questionof integrin variations is potentially of importance because the type of ‘‘outside-in’’ signaling performed by the receptors is affected by their subunit composi-tion. Regional differences could, therefore, translate into differences in plastic-ity and neurotropism; indeed, they could be predictive of regional specializa-tions in types of encoding.

It is evident from the first studies to address the above questions that thebrain is highly differentiated with regard to integrin expression. An exampleof this can be seen in Fig. 1, which is a film autoradiogram illustrating thedistribution of a8 at the level of the amygdala. Note the dense concentrationsof mRNA in the deep layers of cortex and the virtual absence of signal in thesuperficial and middle layers. Labeling is dense in basolateral division of theamygdala but almost undetectable in corticomedial and central divisions.Other a8 subunits in cortex have distributions that are complementary to thatfor a8 , i.e., dense superficially and absent in the deep zones (Pinkstaff et al.,



FIG. 1. Distribution of a8 integrin mRNA in the ventrolateral forebrain of the rat. Note thedense expression in the deep layers of the neocortex (arrow), the endopiriform nucleus (EN), andsome, but not all, nuclei in the basolateral division of the amygdala (BLA). Expression in immedi-ately adjacent subdivisions of the amygdala (e.g., the central division: CE) and the superficiallayers of the neocortex is very light. Other integrin a subunits have been found for these areaswith low levels of a8 ; expression is thus regionally selective and complementary (see Pinkstaff,Lynch, & Gall, 1998).

1998). Enough data should shortly be available to allow for identification ofaxby combinations and thus of known integrin subtypes (e.g., fibronectin recep-tors, laminin receptors). But the results already collected suggest that syn-apses, the brain’s adhesive junctions, will prove to be differentiated with regardto both transmitter and adhesion receptors.

LINKING CONSOLIDATION AND EXPRESSIONOF LONG-TERM POTENTIATION

Hypotheses about how consolidation processes result in stable potentiationare constrained by the presumed duration of the potentiation. If, as the avail-able evidence indicates, LTP is expressed by AMPA receptors, then one-timelasting changes in the receptors as with partial proteolysis (Bi, Chang, Molnar,McIlhinney, & Baudry, 1996) would create potentiation with temporal charac-teristics governed by the rate of receptor turnover. Absent special assumptions,the resultant LTP would be expected to begin decreasing shortly after itsinception. Effects resembling this have been described. In other instances,however, LTP appears to be remarkably stable over the longest span (weeks)in which stimulation and recording electrodes can be maintained in position(Staubli & Lynch, 1987). Explanations for this would appear to require changesthat (i) far outlast the turnover of membrane proteins and (ii) secondarilyinfluence the operating characteristics of now present AMPA receptors andtheir replacements. Four overlapping mechanisms whereby adhesion-medi-ated consolidation might satisfy these constraints are proposed below:

a. Modifying the membrane environment of the receptors. The binding prop-


90 GARY LYNCH

erties of AMPA receptors are very sensitive to the protein and lipid environ-ment in which the receptors are located (cf. Terramani, Kessler, Lynch, &Baudry, 1988; Hall, Kessler, & Lynch, 1992). Modifying adhesive relationshipswould alter cytoskeletal control of membrane characteristics and thereby sec-ondarily modify receptor behavior.

b. Greater access by modulatory proteins. Modifying the relationship be-tween the transmembrane adhesion receptors and the subsynaptic cytoskele-ton could remove barriers offered by the latter to the movement of kinasesand other proteins that modulate the operational characteristics of AMPAreceptors (see Strack et al., 1997).

c. Clustering of receptors. Integrins cluster under various conditions and anot unexpected consequence of this would be to aggregate other transmem-brane receptors. As mentioned, modeling studies suggest that this would mod-ify the probabilities of transmitter binding in such a way that a potentiatedresponse would eventuate (Xie et al., 1997).

d. Modifying the space available for receptor insertion. Disassembly–reas-sembly of an adhesive junction might well be expected to alter the size andshape of the junctional surface. If so, then the space available for receptorsmight also be changed. This could lead to the activation of previously latentreceptors or the insertion of additional copies (see Lynch & Baudry, 1984, foran early version of this idea). Either event would enhance the postsynapticcurrents generated by release of a fixed amount of transmitter.

While underspecified, the first two ideas are, in principle at least, testable.Techniques are now available for applying agonists for about the duration ofa transmitter release event to membrane patches containing small numbersof AMPA receptors. The resultant evoked currents closely resemble excitatorypostsynaptic currents. Moreover, recent experiments show that manipulationsthat alter the size or waveform of the patch responses to millisecond transmit-ter pulses produce very similar effects on the size and waveform of synapticresponses (Arai, Kessler, Xiao, Ambros-Ingerson, Rogers, & Lynch, 1994; Arai,Kessler, Ambros-Ingerson, Quan, Yigiter, Rogers, & Lynch, 1996). Accordingly,synaptic elements that modify AMPA receptors so as to produce potentiatedresponses in situ could generate qualitatively similar ‘‘potentiation’’ of patchresponses to very brief applications of glutamate.

The significance of producing ‘‘potentiation’’ in an excised patch is directlyrelated to the stringency of the definition of LTP. There are a number oftreatments that alter AMPA receptor binding and it can be assumed that atleast some of these will affect the current flow. Criteria in addition to enhancedresponses are needed to establish that a given manipulation creates an LTP-like effect. Pertinent in this regard are the above-noted observations that LTP,in addition to increasing amplitudes, also changes the waveform of the synapticresponse and the manner in which it is altered by AMPA receptor modulators.These observations impose a demanding set of criteria to be satisfied before atreatment can be said to have reproduced LTP.

AMPAKINES: A NEW CLASS OF RECEPTOR MODULATORY DRUGSTHAT PROMOTE LTP AND ENHANCE MEMORY

Advances in understanding how LTP is generated provided a variety of cluesfor the development of drugs to enhance memory. As noted, induction of LTP



FIG. 2. Effects of two positive modulators of AMPA-type glutamate receptors on ligand gatedcurrents in membrane patches excised from hippocampal pyramidal neurons. Results obtainedwith the ampakine CX516 are shown on the left; the effects of the benzothiadiazide compoundcyclothiazide are on the right. The top two panels are for 400-ms applications of a saturatingconcentration of glutamate; the left and right bottom panels are for 1-ms pulses of the transmitter.Note that cyclothiazide has potent effects on 400-ms infusions but little influence on the responseto the short pulse. The ampakine is about equally effective in both instances. Results obtainedwith the short pulses are predictive of effects on monosynaptic transmission in hippocampus andcortex (see Arai et al., 1994, 1996; Arai & Lynch, 1998).

involves the unblocking of voltage-sensitive NMDA receptor channels and in-creases in postsynaptic calcium, related events that require intense and pro-longed depolarization of the synaptic region. Two factors conspire against thisoccurring: (i) AMPA receptor-mediated currents, which have to provide thedepolarizing drive, are brief, and (ii) neurons in the cortical telencephalonreceive potent feedforward GABAergic connections that shunt and offset theexcitatory currents. Short, high-frequency bursts of afferent discharges of atype commonly recorded in the cortical telencephalon reduce the first difficulty,while spacing the bursts apart by about 200 ms minimizes the countervailinginfluence of inhibitory synapses. The latter result arises because GABAergicconnections in hippocampus and cortex become refractory after activation, withthe maximum reduction in transmission occurring at about 200 ms. GABA-bautoreceptors are responsible for these effects. That the ideal conditions forinducing LTP—30-ms bursts separated by Ç200 ms (‘‘theta bursting’’)—match well the theta pattern and cell spiking recorded during a variety ofbehaviors (Otto, Eichenbaum, Wiener, & Wible, 1991) strongly suggests thatLTP is a commonplace occurrence during learning.

Positive Modulators of AMPA-Type Glutamate Receptors

Increasing the size and duration of AMPA receptor currents will increasethe magnitude of NMDA receptor-mediated calcium influxes and thus promotethe induction of LTP. Given the likelihood that LTP is the substrate of variousforms of learning, facilitation of AMPA receptors should also enhance memory.This idea led to the invention of a new class (ú100 compounds) of drugs,


92 GARY LYNCH

‘‘ampakines,’’ that increase AMPA receptor currents, freely cross the blood–brain barrier, and promote LTP induction in vivo. Ampakines are positivemodulators in that they have no detectable agonist/antagonist actions butinstead influence the response of the AMPA receptor after transmitter hasbeen bound. Membrane patch studies indicate that the drugs slow (a) the rateat which AMPA receptor channels stop opening in the continuous presence ofagonists (desensitize) and (b) the speed with which receptor currents returnto baseline following a 1-ms agonist pulse (deactivation) (Arai et al., 1994,1996). There is also a possibility of an effect on the opening/closing rates ofthe AMPA receptor channel. One way of visualizing the actions of the drugsis to assume that they reduce all rate constants leading away from the ‘‘trans-mitter bound/receptor sensitized/channel closed’’ state of the receptors. In thisscheme, ampakines slow transmitter dissociation, the shift from sensitized todesensitized states, and the opening of the channel. The effects of the modula-tors on deactivation, rather than on desensitization (see Fig. 2), are predictiveof their actions on synaptic current (Arai & Lynch, 1998).

Infusion of ampakines into hippocampal slices results in a rapid, fully revers-ible increase in the duration of AMPA receptor-mediated synaptic responses.Certain of the drugs also increase the amplitude of the responses while othershave little effect in this regard. These results are obtained at drug thresholds,EC50s, and relative potencies that are close to those found with excised patches.Slice studies have also made the important point that ampakines have sub-stantially greater effects, and act at lower threshold concentrations, in multi-synaptic networks than they do on monosynaptic responses (Sirvio, Larson,Rogers, & Lynch, 1996; Arai et al., 1996). This is not entirely unexpected giventhat glutamatergic neurons are strung together in very long chains, each linkof which provides an additional site of action for a positive modulator of AMPAreceptors. These results raise the possibility that ampakines will more pro-foundly affect behaviors involving complex networks than those mediated bysimpler, reflex-like circuits.

PET studies indicate that ampakines enter the brain within a minute of inter-peritoneal injections (Staubli, Perez, Xu, Rogers, Ingvar, Stone-Elander, & Lynch,1994b). Chronic recording work has shown that they prolong the duration ofexcitatory synaptic responses (Staubli, Rogers, & Lynch, 1994a) and markedlyfacilitate the induction of LTP in freely moving rats (Staubli et al., 1994b). Inaccord with the idea of larger effects on complex circuits, experiments usingimmediate-early gene expression indicate that ampakines shift the balance ofaggregate neuronal activity in cortex versus striatum of behaving rats to favorthe cortex (Palmer, Hess, Larson, Rogers, Gall, & Lynch, 1997).

Ampakines Have Selective Effects on Behavior

Ampakines at high doses cause seizures but have subtle effects on ratbehavior at concentrations expected from in vitro experiments (patches andslices) to be physiologically active. Tested in trained rats at dosages usedin memory studies, the drugs had no obvious effects on response latency,motivation, attention, or fine motor control (cf. Larson, Lieu, Petchpradub,LeDuc, Ngo, Rogers, & Lynch, 1995). The most evident and reproducedeffect not explicitedly involving memory is a modest reduction in open fieldactivity (Granger, Staubli, Davis, Perez, Nilsson, Rogers, & Lynch, 1993).



They also cause a small but reliable increase in the speed with which well-trained rats collect rewards in a spatial maze (Granger, Deadwyler, Davis,Moskovitz, Rogers, & Lynch, 1996). The combination of reduced exploration,faster performance of complex tasks, and no obvious effect on state variablesconstitutes a unique drug profile.

Much of the selectivity of the drugs can be traced to their being modulatorsas opposed to agonists/antagonists: ampakines facilitate networks already en-gaged by the demands of behavior rather than affecting all excitatory synapses.Selectivity also follows from the earlier noted greater influence of ampakineswithin complex networks and, presumably following from this, their shifting ofaggregate neuronal activity in telencephalon to favor cortex. These argumentspoint to the conclusion that ampakines should increase the extent to whichcortical networks are engaged by, and contribute to, behavior. Enhanced cere-bral dominance would explain how the drugs reduce exploratory activity(greater suppression of behavioral excitability by cortex and hippocampus)while accelerating practiced performance in a complex environment (suppres-sion of competing, unlearned behaviors).

It should also be noted that LTP is a thresholded effect and even smallenhancements of excitatory drive can have profound effects on its induction(see Staubli et al., 1994b, for a demonstration); in this sense, ampakinesare likely to have disproportionate (selective) effects on behaviors requiringLTP-based memory encoding. Finally, it is possible that receptor subtypepreferences influence the classes of behavior targeted by ampakines. AMPAreceptors are composed of various groupings of four homologous subunits,each of which has two splice variants. Certain combinations are used repeat-edly throughout the brain and thus constitute legitimate receptor sub-classes (Gold, Hennegriff, Lynch, & Gall, 1997). Because different receptorsubclasses are located in different brain areas (see Gold et al., 1997, for afirst map), preferences for subunits would translate into preferences forregions or cell types.

Ampakines Enhance the Encoding of Memory

Retention scores for a transient form of memory. A sizable literature docu-menting the effects of ampakines on rat memory has grown up over the past4 years. The initial studies used a within-subject design in which rats weregiven drug or vehicle, allowed to collect rewards in a radial maze, and thentested several hours later (well beyond the half-life of the drug) for memory ofthe earlier choices. Retention scores were substantially higher on drug versusvehicle days (Staubli et al., 1994a, 1994b). This result, which has been repli-cated several times, established that ampakines enhance the encoding of aform of memory that decays over an 8- to 12-h period. Confirmation of thiswas obtained in another task using olfactory rather than spatial cues (Staubliet al., 1994b).

Radial maze experiments also established that blood concentrations of am-pakines required for memory enhancement were close to the media concentra-tions of the same drug needed for positive effects on polysynaptic responsesin brain slices. Moreover, the magnitude of the differences in potencies betweendifferent ampakines in excised patches and slices were reproduced in the mem-ory studies (Davis, Moskovitz, Nguyen, Arai, Lynch, & Granger, 1997). These


94 GARY LYNCH

patch-to-slice-to-animal linkages provide good evidence that ampakines influ-ence memory via their intended receptor targets.

Acquisition of long-term memories. Eye-blink conditioning (Shors, Servat-ius, Thompson, Rogers, & Lynch, 1995) and the formation of learned fear(Rogan, Staubli, & LeDoux, 1996) provide two classical examples of long-termmemory that are facilitated by ampakines. The drugs also reduce the numberof trials needed by well-trained rats to acquire a two-odor discrimination (Lar-son et al., 1995). This last study used a within-subject design and establishedthat ampakines are effective in promoting a commonplace form of long-termmemory in animals extremely familiar with the learning problem. It alsoshowed that the positive memorial effects of the drug were not accompaniedby changes in performance.

Novel procedures for dealing with complex problems. There is evidence thatampakines promote the encoding of strategies for dealing with complex prob-lems. The relevant experiments involved rats trained over a 30-day period toasymptotic performance on a delayed match/nonmatch to sample paradigm.The decay of memory (90/% correct to random choices) over a 40-s period inthis task has been well characterized. Ampakines were administered everyother day beginning in the fifth week of the study and effects on performanceassessed. First-day effects were small but scores for the majority of rats im-proved substantially over 2 weeks of treatment. Unexpectedly, the improve-ments persisted for at least 1 week, the longest time tested, after drug injec-tions were stopped. This suggested that the rats may have gradually learnedsomething about the task during the repeated sessions in which they wereadministered the drug. Detailed analyses then revealed that the improvementsin retention scores occurred in association with a reduction of proactive inter-ference; i.e., errors on one trial were no longer predictive of behavior on the nexttrial. Neurophysiological correlates of this gradually developing, ampakine-induced change in behavior were obtained with chronic recording. A patternof firing generated by certain types of error and that continued through theintertrial interval was reduced by the ampakine treatments (Hampson, Rogers,Lynch, & Deadwyler, 1988a, 1988b). These results strongly suggest that theampakine allowed the rats to learn procedures that otherwise would not beacquired, even with extended training.

Age-associated memory losses. Six- and eight-hour retention scores in spa-tial mazes (see above) are substantially poorer in middle-aged rats (Ç14months) than young adults. While this transient form of memory improves inboth groups with extended training on the maze, the young–old differencepersists; accordingly, the deficit is not due to age-related declines in the abilityto adapt to new problems but rather seems to reflect impairments in encodingand/or storage processes. Middle-aged rats showed a sizable improvement onampakine versus vehicle days; moreover, scores for the older rats on drug dayswere not detectably different from the control scores for the younger animals.

Effects on memory in humans. Ampakines have been tested in three clinicaltrials with positive effects on memory. Probably the most obvious change wasobtained in a study of delayed recall of nonsense syllables in 65- to 75-year-old subjects: a nearly twofold improvement was obtained at the highest doseused. The changes appeared to be dose dependent, with the effective concentra-



FIG. 3. Effects of the ampakine CX516 on delayed recall of nonsense syllables by 65- to 75-year-old subjects. As shown, the subjects in this study had poorer recall of 10 syllables (1 or 2items) after a 5-min delay than did young adults in earlier experiments (4 or 5 items). A highlysignificant drug effect (ANOVA) was obtained, with the highest dose group recalling about twiceas many items as the placebos (see Lynch et al., 1997). The effective dosage here corresponds tothat found in earlier studies on recall in animals.

tion corresponding to that found to enhance various types of memory in ani-mals (Lynch, Granger, Ambros-Ingerson, Davis, Schehr, & Kessler, 1997) (Fig.3). Smaller improvements were obtained in a separate study using lower dosesand young adult subjects. A battery of measures were administered, withtesting beginning shortly after drug ingestion in an effort to compensate forthe suboptimal doses by aligning the encoding periods with peak blood levels.Statistically significant improvements were obtained in 24-h retention of asso-ciations between complex visual cues, recognition of odors after a delay of 45min or more, and over-days improvements in a visuospatial maze. Borderlineincreases were obtained in immediate recall of card positions but there wereno differences in cued recall of verbal memory. Visual recognition, motor perfor-mance, and general intellectual functioning were not detectably different ondrug versus placebo days and the subjects were unable to guess whether theyhad been given drug or placebo (Ingvar, Ambros-Ingerson, Davis, Granger,Kessler, Rogers, Schehr, & Lynch, 1997).

It should be emphasized that each of the human studies employed a rela-tively small number of subjects and, excepting the low-dose experiment justcited, were intended to evaluate the safety of an entirely new class of drugs.Much more extensive testing will be needed before strong conclusions can bedrawn. Of particular interest in this regard would be results with ampakinesthat are far more potent than the variant used in the first human studies. But


96 GARY LYNCH

from initial results, the ampakines appear to be a successful application ofinformation about the neurobiology of LTP to the construction of memory-enhancing drugs.

REFERENCES

Abraham, W. C., & Huggett, A. (1997). Induction and reversal of long-term potentiation by re-peated high-frequency stimulation in rat hippocampal slices. Hippocampus, 7, 137–145.

Ambros-Ingerson, J., Larson, J., Xiao, P., & Lynch, G. (1991). LTP changes the waveform ofsynaptic responses. Synapse, 9, 314–316.

Ambros-Ingerson, J., & Lynch, G. (1993). Channel gating kinetics and synaptic efficacy: A hypothe-sis for expression of long-term potentiation. Proceedings of the National Academy of Sciencesof the United States of America, 90, 7903–7907.

Ambros-Ingerson, J., Xiao, P., Larson, J., & Lynch, G. (1993). Waveform analysis suggests thatLTP alters the kinetics of synaptic receptor channels. Brain Research, 620, 237–244.

Arai, A., Kessler, M., Ambros-Ingerson, J., Quan, A., Yigiter, E., Rogers, G., & Lynch, G. (1996).Effects of a centrally active benzoylpiperidine drug on AMPA receptor kinetics. Neuroscience,75, 573–585.

Arai, A., Kessler, M., & Lynch, G. (1990a). The effects of adenosine on the development of long-term potentiation. Neuroscience Letters, 119, 41–44.

Arai, A., Kessler, M., Xiao, P., Ambros-Ingerson, J., Rogers, G., & Lynch, G. (1994). A centrallyactive drug that modulates AMPA receptor gated currents. Brain Research, 638, 343–346.

Arai, A., Larson, J., & Lynch, G. (1990b). Anoxia reveals a vulnerable period in the developmentof long-term potentiation. Brain Research, 511, 353–357.

Arai, A., & Lynch, G. (1992). Antagonists of the platelet activating factor receptor block long termpotentiation in hippocampal slices. European Journal of Neuroscience, 4, 411–419.

Arai, A., & Lynch, G. (1998). The waveform of synaptic transmission at hippocampal synapses isnot determined by AMPA receptor desensitization. Brain Research, in press.

Asztely, F., Hans, E., Wigstrom, H., & Gustafsson, B. (1992). Aniracetam-evoked potentiationdoes not interact with long-term potentiation in the CA1 region of the hippocampus. Synapse,11, 342–345.

Bahr, B. A., & Lynch, G. (1992). Purification of an Arg–Gly–Asp selective matrix receptor frombrain synaptic plasma membranes. Biochemical Journal, 281(Pt 1), 137–142.

Bahr, B. A., Staubli, U., Xiao, P., Chun, D., Ji, Z. X., Esteban, E. T., & Lynch, G. (1997). Arg–Gly–Asp–Ser-selective adhesion and the stabilization of long-term potentiation: Pharmacologicalstudies and the characterization of a candidate matrix receptor. Journal of Neuroscience, 17,1320–1329.

Barria, A., Muller, D., Derkach, V., Griffith, L. C., & Soderling, T. R. (1997). Regulatory phosphory-lation of AMPA-type glutamate receptors by CaM-KII during long-term potentiation. Science,276, 2042–2045.

Barrionuevo, G., Schottler, F., & Lynch, G. (1980). The effects of repetitive low frequency stimula-tion on control and ‘‘potentiated’’ synaptic responses in the hippocampus. Life Science, 27,2385–2391.

Bazan, N. G., Packard, M. G., Teather, L., & Allan, G. (1997). Bioactive lipids in excitatory neuro-transmission and neuronal plasticity. Neurochemistry International, 30, 225–231.

Bednarski, E., Gall, C., Lynch, G., & Vanderklish, P. (1995). Translational suppression of calpain-1 reduces NMDA-induced spectrin proteolysis and pathophysiology of cultured hippocampalslices. Brain Research, 694, 147–157.

Bi, X., Chang, V., Molnar, E., McIlhinney, R. A. J., & Baudry, M. (1996). The C-terminal domainof glutamate receptor subunit 1 is a target for calpain-mediated proteolysis. Neuroscience,73, 903–906.

Cardwell, M. C., & Rome, L. H. (1988). RGD-containing peptides inhibit the synthesis of myelin-like membrane by cultured oligodendrocytes. Journal of Cell Biology, 107, 1551–1559.

Chang, F.-L. F., & Greenough, W. T. (1984). Transient and enduring morphological correlates of



synaptic activity and efficacy change in the rat hippocampal slice. Brain Research, 309, 35–46.

Clark, E. A., & Hynes, R. O. (1996). Ras activation is necessary for integrin-mediated activation ofextracellular signal-regulated kinase 2 and cytosolic phospholipase A2 but not for cytoskeletalorganization. Journal of Biological Chemistry, 271, 14814–14818.

Collingridge, G. L., Kehl, S. J., & McLennan, H. (1983). Excitatory amino acids in synaptic trans-mission in the Schaffer-commissural pathway of the rat hippocampus. Journal of Physiology(London), 334, 33–46.

Cunha, R. A., Vizi, E. S., Ribeiro, J. A., & Sebastiao, A. M. (1996). Preferential release of ATPand its extracellular catabolism as a source of adenosine upon high- but not low-frequencystimulation of rat hippocampal slices. Journal of Neurochemistry, 67, 2180–2187.

Davis, C. M., Moskovitz, B., Nguyen, M. A., Arai, A., Lynch, G., & Granger, R. (1997). A profileof the behavioral changes produced by facilitation of AMPA-type glutamate receptors. Psycho-pharmacology, 133, 161–167.

Desmond, N. L., & Levy, W. B. (1986). Changes in the postsynaptic density with long-term potenti-ation in the dentate gyrus. Journal of Comparative Neurology, 253, 476–482.

Du, X., Saido, T. C., Tsubuki, S., Indig, F. E., Williams, M. J., & Ginsberg, M. H. (1995). Calpaincleavage of the cytoplasmic domain of the integrin beta 3 subunit. Journal of BiologicalChemistry, 270, 26146–26151.

Einheber, S., Schnapp, L. M., Salzer, J. L., Cappiello, Z. B., & Milner, T. A. (1996). Regional andultrastructural distribution of the alpha 8 integrin subunit in developing and adult rat brainsuggests a role in synaptic function. Journal of Comparative Neurology, 370, 105–134.

Finn, R. C., Browning, M., & Lynch, G. (1980). Trifluoperazine inhibits hippocampal long-termpotentiation and the phosphorylation of a 40,000 dalton protein. Neuroscience Letters, 19,103–108.

Gold, S., Hennegriff, M., Lynch, G., & Gall, C. (1996). Relative concentrations and seizure-inducedchanges in mRNAs encoding three AMPA receptor subunits in hippocampus and cortex.Journal of Comparative Neurology, 365, 541–555.

Granger, R., Deadwyler, S., Davis, M., Moskovitz, B., Rogers, G., & Lynch, G. (1996). Facilitationof glutamate receptors reverses an age-associated memory impairment in rats. Synapse, 22,332–337.

Granger, R., Myers, R., Whelpley, E., & Lynch, G. (1995). Intrinsic and emergent corticohippocam-pal computation. In Long-Term Potentiation. (Vol. 2, pp. 371–392). Cambridge, MA: MITPress.

Granger, R., Staubli, U., Davis, M., Perez, Y., Nilsson, L., Rogers, G. A., & Lynch, G. (1993). Adrug that facilitates glutamatergic transmission reduces exploratory activity and improvesperformance in a learning dependent task. Synapse, 15, 326–329.

Grooms, S. Y., & Jones, L. S. (1997). RGDS tetrapeptide and hippocampal in vitro kindling inrats: Evidence for integrin-mediated physiological stability. Neuroscience Letters, 231, 139–142.

Grooms, S. Y., Terracio, & Jones, L. S. (1993). Anatomical localization of beta 1 integrin-likeimmunoreactivity in rat brain. Experimental Neurology, 122, 253–259.

Hall, R. A., Kessler, M., & Lynch, G. (1992). Evidence that high- and low-affinity DL-n-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) binding sites reflect membrane-depen-dent states of a single receptor. Journal of Neurochemistry, 59, 1997–2004.

Hampson, R., Rogers, G., Lynch, G., & Deadwyler, S. (1998a). Facilitative effects of the ampakineCX516 on short ter memory in rats: Correlations with hippocampal unit activity. Journal ofNeuroscience, 18, 2740–2747.

Hampson, R., Rogers, G., Lynch, G., & Deadwyler, S. (1988b). Facilitative effects of the ampakineCX516 on short term memory in rats: Enhancement of delayed non-match to sample perfor-mance. Journal of Neuroscience, 18, 2748–2763.

Haskel, E. J., & Abendschein, D. R. (1989). Deaggregation of human platelets in vitro by an RGDanalog antagonist of platelet glycoprotein IIb/IIIa receptors. Thrombosis Research, 56, 687–695.

Hayashi, Y., Ishida, A., Katagiri, H., Mishina, M., Fujisawa, H., Manabe, T., & Takahashi, T.(1997). Calcium- and calmodulin-dependent phosphorylation of AMPA type glutamate recep-


98 GARY LYNCH

tor subunits by endogenous protein kinases in the post-synaptic density. Brain Research.Molecular Brain Research, 46(1–2), 338–342.

Ingvar, M., Ambros-Ingerson, J., Davis, M., Granger, R., Kessler, M., Rogers, G. A., Schehr, R. S., &Lynch, G. (1997) Enhancement by an ampakine of memory encoding in humans. ExperimentalNeurology, 146, 553–559.

Jung, M. W., Larson, J., & Lynch, G. (1991). Evidence that changes in spine neck resistance arenot responsible for expression of LTP. Synapse, 7, 216–220.

Kauer, J. A., Malenka, R. C., & Nicoll, R. A. (1988). A persistent postsynaptic modification medi-ates long-term potentiation in the hippocampus. Neuron, 1, 911–917.

Kolta, A., Lynch, G., & Ambros-Ingerson, J. (1998). Effects of aniracetam after LTP induction aresuggestive of interactions on the kinetics of the AMPA receptor channel. Brain Research, 788,269–286.

Larson, J., Lieu, T., Petchpradub, V., LeDuc, B., Ngo, H., Rogers, G., & Lynch, G. (1995). Facilita-tion of olfactory learning by a modulator of AMPA receptors. Journal of Neuroscience, 15,8023–8030.

Larson, J., & Lynch, G. (1986). Induction of synaptic potentiation in hippocampus by patternedstimulation involves two events. Science, 232, 985–988.

Larson, J., & Lynch, G. (1988). Priming, NMDA receptor activation, and LTP induction by thetapattern stimulation. In E. A. Cavalhiero, J. Lehmann, & L. Turski (Eds.), Frontiers in excit-atory amino acid research (pp. 411–418). New York: Liss.

Larson, J., & Lynch, G. (1991). A test of the spine resistance hypothesis of LTP expression. BrainResearch, 538, 347–350.

Larson, J., Wong, D., & Lynch, G. (1986). Patterned stimulation at the theta frequency is optimalfor induction of hippocampal long-term potentiation. Brain Research, 368, 347–350.

Larson, J., Xiao, P., & Lynch, G. (1993). Reversal of LTP by theta frequency stimulation. BrainResearch, 600, 97–102.

Lau, L. F., & Huganir, R. L. (1995). Differential tyrosine phosphorylation of N-methyl-D-aspartatereceptor subunits. Journal of Biological Chemistry, 270, 20036–20041.

Lee, K., Oliver, M., Schottler, F., & Lynch, G. (1981). Electron microscopic studies of brain slices:The effects of high frequency stimulation on dendritic ultra structure. In G. Kerkut & H. V.Wheal (Eds.), Electrical activity in isolated mammalian CNS preparations (pp. 189–212).New York: Academic Press.

Lee, K., Schottler, F., Oliver, M., & Lynch, G. (1980). Brief bursts of high-frequency stimulationproduce two types of structural changes in rat hippocampus. Journal of Neurophysiology, 44,247–258.

Luthi, A., Laurent, J-P, Figurov, A., Muller, D., & Schachner, M. (1994). Hippocampal long termpotentiation and neural cell adhesion molecules L1 and NCAM. Nature, 372, 777–779.

Lynch, G., & Baudry, M. (1984). The biochemistry of memory: A new and specific hypothesis.Science, 224, 1057–1063.

Lynch, G., Granger, R., Ambros-Ingerson, J., Davis, C. M., Schehr, R., & Kessler, M. (1997).Evidence that a positive modulator of AMPA-type glutamate receptors affects delayed recallin aged humans. Experimental Neurology, 145, 89–92.

Lynch, G., Larson, J., Kelso, S., Barrionuevo, S., & Schottler, F. (1983). Intracellular injectionsof EGTA block induction of hippocampal long-term potentiation. Nature, 305, 719–721.

Massicotte, G., Vanderklish, P., Lynch, G., & Baudry, M. Modulation of DL-alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic Acid-quisqualate receptors by phospholipase-A2: A necessarystep in long-term potentiation. Proceedings of the National Academy of Sciences of the UnitedStates of America, 88, 1893–1897.

Mott, D. D., & Lewis, D. V. (1991). Facilitation of the induction of long-term potentiation byGABAB receptors. Science, 252, 1718–1720.

Muller, D., Joly, M., & Lynch, G. (1988a). Contributions of quisqualate and NMDA receptors tothe induction and expression of LTP. Science, 242, 1694–1697.

Muller, D., & Lynch, G. (1989). Evidence that changes in presynaptic calcium currents are notresponsible for long-term potentiation in hippocampus. Brain Research, 479, 290–299.

Muller, D., & Lynch, G. (1988). Long-term potentiation differentially affects two components of



synaptic responses in hippocampus. Proceedings of the National Academy of Sciences of theUnited States of America, 85, 9346–9350.

Muller, D., Turnbull, J., Baudry, M., & Lynch, G. (1988b). Phorbol ester-induced synaptic facilita-tion is different than long-term potentiation. Proceedings of the National Academy of Sciencesof the United States of America, 85, 6997–7000.

Newton, R. A., Thiel, M., & Hogg, N. (1997). Signaling mechanisms and the activation of leukocyteintegrins. Journal of Leukocyte Biology, 61, 422–426.

Otto, T., Eichenbaum, H., Wiener, S. I., & Wible, C. G. (1991). Learning-related patterns of CA1spike trains parallel stimulation parameters optimal for inducing hippocampal long-termpotentiation. Hippocampus, 1, 181–192.

Palmer, L. C., Hess, U. S., Larson, J., Rogers, G. A., Gall, C. M., & Lynch, G. (1997). Comparisonof the effects of an ampakine with those of methamphetamine on aggregate neuronal activityin cortex versus striatum. Molecular Brain Research, 46, 127–135.

Paulus, W., Baur, I., Schuppan, D., & Roggendorf, W. (1993). Characterization of integrin receptorsin normal and neoplastic human brain. American Journal of Pathology, 143, 154–163.

Pierschbacher, M. D., & Ruoslahti, E. (1984). Cell attachment activity of fibronectin can be dupli-cated by small synthetic fragments of the molecule. Nature, 309, 30–33.

Pinkstaff, J. K., Lynch, G., & Gall, C. M. (1998). Integrin b-1 mRNA expression in rat brain:Localization and upregulation in neurons and glia by seizure activity. Brain Research, 55,265–276.

Roberts-Lewis, J., Savage, M. J., Marcy, V. R., Pinsker, L. R., & Siman, R. (1994). Immunolocaliza-tion of calpain 1-mediated spectrin degredation to vulnerable neurons in the ischemic gerbilbrain. Journal of Neuroscience, 14, 3934–3944.

Rogan, M. T., Staubli, U. V., & LeDoux, J. E. (1997). AMPA receptor facilitation accelerates fearlearning without altering the level of conditioned fear acquired. Journal of Neuroscience, 17,5928–5935.

Rønn, L. C. B., Bock, E., Linnemann, D., & Jahnsen, H. (1995). NCAM-antibodies modulate induc-tion of long-term potentiation in rat hippocampal CA1. Brain Research, 677, 145–151.

Saido, T. C., Yokota, M., Nagao, S., Yamaura, I., Tani, E., Tsuchiya, T., Suzuki, K., & Kawashima,S. (1993). Spatial resolution of fodrin proteolysis in postischemic brain. Journal of BiologicalChemistry, 268, 25239–25243.

Seubert, P., Baudry, M., Dudek, S., & Lynch, G. (1987). Calmodulin stimulates the degradationof brain spectrin by calpain. Synapse, 1, 20–24.

Sheppard, A., Wu, J., Rutishauser, U., & Lynch, G. (1991). Proteolytic modification of neural celladhesion molecule (N-CAM) by the intracellular proteinase calpain. Biochimica BiophysicaActa, 1076, 156–160.

Shors, T. J., Servatius, R. J., Thompson, R. F., Rogers, G., & Lynch, G. (1995). Enhanced gluta-matergic neurotransmission facilitates classical conditioning in the freely moving rat. Neuro-science Letters, 186, 153–156.

Siman, R., Baudry, M., & Lynch, G. (1984). Brain fodrin: Substrate for calpain I, an endogenouscalcium-activated protease. Proceedings of the National Academy of Sciences of the UnitedStates of America, 81, 3572–3576.

Sirvio, J., Larson, J., Rogers, G. A., & Lynch, G. (1996). Effects of facilitating glutamatergictransmission in the trisynaptic intrahippocampal circuit. Neuroscience, 74, 1025–1035.

Staubli, U., Vanderklish, P., & Lynch, G. (1990). An inhibitor of integrin receptors blocks long-term potentiation. Behavioral and Neural Biology, 53, 1–5.

Staubli, U., Ambros-Ingerson, J., & Lynch, G. (1992). Receptor changes and LTP: An analysisusing aniracetam, a drug that reversibly modifies glutamate (AMPA) receptors. Hippocampus,2, 49–58.

Staubli, U., & Chun, D. (1996a). Factors regulating the reversibility of long term potentiation.Journal of Neuroscience, 16, 853–860.

Staubli, U., & Chun, D. (1996b). Proactive and retrograde effects on LTP produced by theta pulsestimulation: Mechanisms and characteristics of LTP reversal. Learning & Memory, 3, 96–105.


100 GARY LYNCH

Staubli, U., & Lynch, G. (1987). Stable hippocampal long-term potentiation elicited by ‘theta’pattern stimulation. Brain Research, 435, 227–234.

Staubli, U., & Lynch, G. (1990). Stable depression of potentiated synaptic responses in the hippo-campus with 1–5 Hz stimulation. Brain Research, 513, 113–118.

Staubli, U., Perez, Y., Xu, F., Rogers, G., Ingvar, M., Stone-Elander, S., & Lynch, G. (1994b).Centrally active modulators of glutamate (AMPA) receptors facilitate the induction of LTPin vivo. Proceedings of the National Academy of Sciences of the United States of America, 91,11158–11162.

Staubli, U., Rogers, G., & Lynch, G. (1994a). Facilitation of glutamate receptors enhances memory.Proceedings of the National Academy of Sciences of the United States of America, 91, 777–781.

Strack, S., Choi, S., Lovinger, D. M., & Colbran, R. J. (1997). Translocation of autophosphorylatedcalcium/calmodulin-dependent protein kinase II to the postsynaptic density. Journal of Biolog-ical Chemistry, 272, 13467–13470.

Stricker, C., Field, A. C., & Redman, S. J. (1996). Changes in the quantal parameters of EPSCsin rat CA1 neurons in vitro after induction of long term potentiation. Journal of Physiology(London), 490, 443–454.

Terramani, T., Kessler, M., Lynch, G., & Baudry, M. (1988). Effects of thiol-reagents on [3H]-AMPA binding to rat telencephalic membranes. Molecular Pharmacology, 34, 117–123.

Thiel, M., Chambers, J. D., Chouker, A., Fischer, S., Zourelidis, C., Bardenheuer, H. J., Arfors,K. E., & Peter, K. (1996). Effect of adenosine on the expression of beta(2) integrins and L-selectin of human polymorphonuclear leukocytes in vitro. Journal of Leukocyte Biology, 59,671–682.

Van Willigen, G., Hers, I., Gorter, G., & Akkerman, N. (1996). Exposure of ligand-binding siteson platelet integrin AIIb/B3 by phosphorylation of the B3 subunit. Biochemical Journal, 314,769–779.

Vanderklish, P., Bednarski, E., & Lynch, G. (1996). Translational suppression of calpain blockslong-term potentiation. Learning & Memory, 3, 209–217.

Vanderklish, P., Neve, R., Bahr, B. A., Arai, A., Hennegriff, M., Larson, J., & Lynch, G. (1992).Translational suppression of a glutamate receptor subunit impairs long-term potentiation.Synapse, 12, 333–337.

Vanderklish, P., Saido, T. C., Gall, C., Arai, A., & Lynch, G. (1995). Proteolysis of spectrin bycalpain accompanies theta-burst stimulation in cultured hippocampal slices. Molecular BrainResearch, 32, 25–35.

Vertes, R. P., & Kocsis, B. (1997). Brainstem–diencephalo–septohippocampal systems controllingthe theta rhythm in the rat. Neuroscience, 81, 893–926.

Xiao, P., Bahr, B. A., Staubli, U., Vanderklish, P. W., & Lynch, G. (1991). Evidence that matrixrecognition contributes to stabilization but not induction of LTP. Neuroreport, 2, 461–464.

Xiao, P., Staubli, U., Kessler, M., & Lynch, G. (1991). Selective effects of aniracetam across receptortypes and forms of synaptic facilitation in hippocampus. Hippocampus, 1, 373–380.

Xie, X., Liaw, S-J., Baudry, M., & Berger, T. (1997). Novel expression mechanism for synapticpotentiation: Alignment of pre-synaptic release site and post-synaptic receptor. Proceedingsof the National Academy of Sciences of the United States of America, 94, 6983–6988.


Date post:	10-Oct-2016
Category:	Documents
Upload:	gary-lynch
View:	213 times
Download:	0 times

Memory and the Brain: Unexpected Chemistries and a New Pharmacology

Documents