RESEARCH ARTICLE

Reflections on glycogen and β-amyloid: why doesglycogenolytic β2-adrenoceptor stimulation not rescue memoryafter β-amyloid?

Marie Gibbs

Received: 12 March 2014 /Accepted: 30 April 2014# Springer Science+Business Media New York 2014

Abstract Normally noradrenaline release ~30min after train-ing in the day-old chick is essential for memory consolidationby simultaneously increasing both glycogenolysis, by its stim-ulation of β2-adrenergic (AR) receptors, and glycogen syn-thesis, by its stimulation of α2-AR receptors in astrocytes. Atthe same time noradrenaline stimulation of β3-AR receptorsincreases glucose uptake solely in astrocytes. Intracerebralinjection of small oligomeric β-amyloid protein (Aβ1–42)(Aβ) 45 min before one-trial bead discrimination learning inday-old chicks abolishes consolidation of memory 30 minpost-learning. The ensuing memory loss can be rescued byinjection of selective β3- and β2-AR agonists (CL316243 andzinterol), which also have the ability to consolidate weakly-reinforced learning into long-term memory. However, al-though CL316243 rescues Aβ-induced memory loss over asimilar time period to when it consolidates weak learning (upto 25 min post training), zinterol is effective over a morelimited time period and unexpectedly it does not rescue atthe time it promotes glycogenolysis. Injection of Aβ into thehippocampus and the locus coeruleus (LoC) also producessimilar memory deficits and injection of both AR agonists intoa cortical area can rescue memory from LoC Aβ. We havepreviously shown that β3-AR stimulation increases astrocyticglucose uptake and have suggested there may be sensitizationor upregulation of the receptor. Since β2-AR stimulation doesnot rescue memory at the time it promotes glycogenolysis, butthe receptor does not appear to be impaired, it is suggested thatAβ may be causing an impairment in the synthesis of readilyavailable glycogen.

Keywords Glycogen .Memory formation . Beta-amyloid .

β-adrenergic receptors

Alzheimer’s disease in its earliest stages is characterized by aninability to form new memories. It is recognized that thedisease is caused by a build up and aggregation of neurotoxicAβ peptides in the brain, particularly the longer species suchas Aβ1–42. Recent evidence suggests that it is the build up ofsoluble, low oligomeric Aβ, consisting of dimers and trimers,prior to extensive aggregation that causes the neurotoxicity(Klein et al. 2001; Walsh and Selkoe 2004). Injection ofsoluble, small oligomeric Aβ directly into selected brain areascauses an inability to form new memories in day-old chickens(Gibbs et al. 2009; Gibbs et al. 2010b; Gibbs and Gibbs 2013;Mileusnic et al. 2007.

Noradrenergic dysfunction is implicated as an importantcomponent of AD (Guerin et al. 2009; Hertz 1989;Weinshenker 2008; Marien et al. 2004 with locus coeruleus(LoC) and parts of the neocortex being involved early indevelopment of the disease (Braak and Del Tredici 2004.The LoC is the site of most of the noradrenenrgic neuronesin the brain and sends projections to all parts of the brain. Thetiming of the memory loss after Aβ is consistent with impair-ment of noradrenergic transmission (Gibbs and Summers2002). Activation of the β3-ARs by the selective agonistCL316243 completely antagonizes the effect of exogenousAβ onmemory (Gibbs et al. 2010b), and in fact the receptor issensitised to the agonist, suggesting it may be upregulated.The main action of this AR subtype in memory consolidationis to stimulate glucose uptake in astrocytes (Gibbs et al.2008c; Hutchinson et al. 2007), so up-regulation of the astro-cytic glucose transporter (GLUT1) could also be involved incounteracting the effects of Aβ. Unexpectedly, the selectiveβ2-AR agonist zinterol, that normally consolidates memory,was unable to do so after Aβ treatment. Stimulation of

M. Gibbs (*)Monash Institute of Pharmaceutical Sciences, Monash University,Parkville, Melbourne, 3052 VIC, Australiae-mail: marie.gibbs@monash.edu

Metab Brain DisDOI 10.1007/s11011-014-9563-y

astrocytic β2-ARs increases metabolic energy flux by break-ing down glycogen and provides glutamate for neuronaltransmission (Gibbs et al. 2006; Gibbs et al. 2007; Gibbs etal. 2008b).

If Aβ exposure depletes glycogen levels, then zinterolwould not be able to promote memory consolidation by in-creasing glycogen breakdown. We have shown that insulinincreases astrocytic glycogen, but we also have evidenceindicating that insulin is unable to rescue memory after Aβtreatment (Gibbs et al. 2009). The different memory consoli-dation outcomes for activation of β2 and β3-ARs and insulinreceptors after Aβ, together with some of our more recentfindings (Gibbs et al. 2009) suggest that Aβ1–42 exerts itsamnestic effects specifically by impairing glycogen metabo-lism in astrocytes.

The breakdown of glycogen is essential for memoryformation. During one-trial aversive learning in the day-old chick, we have established the exact timing of threeperiods of glycogenolysis (Gibbs et al. 2008b).Inhibition of glycogenolysis leads to memory loss.Various neurotransmitters, including serotonin, ATP andnoradrenaline promote glycogenolysis (Quach et al.1982; Magistretti 1988; Cambray-Deakin et al. 1988;Kong et al. 2002; Darvesh and Gudelsky 2003).Serotonin acting on 5-HT2B receptors triggers the firstglycogenolytic period 5 min after training, whereas nor-adrenaline is associated with the second glycogenolyticperiod (Gibbs and Hertz 2014). Noradrenaline promotesboth glycogenolytic and glycogen synthesizing effectsvia β2 - and α2- ARs respectively. β2-AR stimulationis the signal leading to the second glycogenolytic period20–30 min after training, but β2-ARs are not implicatedin the first period.

Recent experiments with the chick model in our laboratoryhave examined the early effects of β-amyloid (Aβ) on mem-ory processing. We have shown intracerebral injection ofsoluble Aβ1–42 prior to training results in the loss of memory30 min after training which can be rescued by β3-AR stimu-lation 20 min after training, but not by β2-AR stimulation atthis time. This rescue occurs during the second period ofglycogenolysis, at a time critical for memory formation. Inthis paper we review some of the published data and furtherexplore the role of β2-ARs in memory rescue and their inabil-ity to stimulate glycogen breakdown.

Methods

Chick bead discrimination learning

Chicks show robust learning when they are trained to associ-ate bead colour with an aversive taste (Gibbs 2008). Chicksare given one 10 s exposure to a red bead where the bead is

tainted with a bitter tasting chemical aversant, methyl anthra-nilate (Sigma- Aldrich). Tested once, at discrete times afterthis training trial, chicks will avoid the red bead but continueto peck a blue bead that has never tasted bitter. The number ofpecks in the 10 s period are counted on a digital recorder. Priorto the learning trial, the chicks have an equal preference forbeads of both colours. The formation of permanent memorydepends on the strength of the reinforcer on the bead: theconcentrated aversant (100 % anthranilate) results in perma-nent memory whereas after tasting the diluted aversant a labilememory results that lasts only 30 min. Labile memory can beconsolidated into permanent memory by injection ofneuromodulators or hormones within the 30 min of memoryestablishment. These tasks allow very accurate timing ofevents following learning and accurate timing of any memoryinterference caused by drugs. In the discriminative bead task,memory is measured as the choice to peck the coloured beadsor not and not as the withholding of a response as is the case inpassive avoidance. With good memory a discrimination ratio(disc.ratio) of 0.9 to 1.0 occurs, whereas with poor memorythe disc.ratio approaches 0.5 and both beads receive equalnumbers of pecks; up to 10 or so pecks on each bead is notuncommon.

Drug administration

Drugs are administered by bilateral intracerebral injections of1 – 5 μl the volume depending on the brain region targeted.Injections can be targeted to specific brain areas withouthaving to cannulate the brain or to use anaesthesia, and thechicks do not show any discomfort from intracerebral injec-tions and will peck at beads in their usual manner very shortlyafter the injections. We can target different brain regions suchas the avian cortical area (IMM- intermediate medialmesopallium), the hippocampus (Gibbs et al. 2008a), themedial striatum (Gibbs and Summers 2003) and the locuscoeruleus (Gibbs et al. 2010a) to investigate the effect of Aβand the contribution of these areas to memory processing.Stock Aβ1–42 (2.2 mM) is made up in DMSO and diluted in0.9 % NaCl within 30 min of injection to limit aggregation.,controls are injected with the appropriate concentration ofDMSO (see details (Gibbs et al. 2009; Gibbs et al.2010b). A single dose of 10 pmol Aβ1–42 injected intoeach hemisphere of the IMM (sensory cortical region)45 min before training reveals memory impairment after30 min (Fig. 1) (adapted from (Gibbs et al. 2010b).Memory is still impaired 24 h after training. Chicksgiven Aβ1–42 24 h before training, showed that memoryis again unaffected on test 30 min after training but islost by 120 min (large open symbols Fig. 1) demon-strating unimpaired labile memory but an absence ofconsolidation into long-term storage, illustrating the ro-bust nature of the effect of Aβ in our model.

Metab Brain Dis

Challenges to Aβ- induced memory loss are made byintracerebral injection of the β2-AR selective agonistzinterol (zinterol hydrochloride, Bristol-Myer Squibb,Noble Park, Australia) and the β3-AR agonistCL316243 and the selective antagonist SR592230A(Sigma-Aldrich Inc., St Louis, MO, USA).

Statistical analysis

Generally group sizes started around 12–16. Chicks that didnot peck the red bead during training or avoided the blue beadon test are eliminated from the final data analysis, althoughthis was no more than one or two chicks in each group.

Data were analyzed using SPSS v.20 software. All data areexpressed as the mean±SEM. Comparisons between groupswere performed by one-way and two-way independent mea-sures ANOVAwith either Dunnett’s test or simple main effectpost hoc analyses where appropriate. Two-tailed tests of sig-nificance were conducted and a type I error rate of .05 wasadopted.

All experimental procedures were in accordance with theguidelines approved by the Monash University animal EthicsCommittee and comply with the 1997 Australian Code ofPractice for the Care and Use of Animals for ScientificPurposes. All efforts were made to minimize both the suffer-ing and number of animals used. Chicks were killed on thecompletion of each experiment by CO2 inhalation.

Results

Memory retention after Aβ

Investigation of the timecourse of memory after Aβ treatmentrevealed that memory loss was not apparent 10 min aftertraining but developed after 30 min (Fig. 1 adapted from

(Gibbs et al. 2010b). Memory retention after DMSO controlremained high at test. Aβ (10 pmol/hem) was injected intoIMM 45 min before training and memory was tested inseparate groups at various intervals after training: this showedthat memory retention was unimpaired on test between 10 and30 min after training, but it declined rapidly by 35 min afterand remained poor at 120 min [F (7,119)=9.60;P<.001]).Injection 5 min before training showed the same pattern ofretention loss (Gibbs et al. 2010b). In the present experiment,injection of Aβ or DMSO control was made 24 h prior totraining and also yielded the same pattern, memory wasunaffected by Aβ on test until 30 min after training, butimpaired at 120 min (large open symbols Fig. 1; F (2,38)=8.84;P<.001]). The timecourse of memory retention after Aβresembles that seen when the anthranilate concentration usedin the training protocol is diluted to 20 %, i.e. memoryretention is good for 30 min and then declines (Gibbs andSummers 2002).

Challenge to Aβ-induced memory loss by β3-and β2-adrenoceptor agonists

During the period when memory retention is good afterweakly-reinforced training, i.e. the first 30 min after training,we have shown that this memory can be consolidated intopermanent long-term storage by noradrenaline and selectiveadrenoceptor agonists (Gibbs and Summers 2002). The sameeffect occurred, i.e. consolidation of the memory trace, whenthe selective β3-adrenoceptor agonist CL316243 was givenafter Aβ-induced memory loss (Gibbs et al. 2010b). WhenCL316243 (3 pmol/hem) was injected bilaterally into IMMwithin 20 min after training in the presence of Aβ (10pmol/hem) memory was consolidated and found to be stillpresent 120 min after training (Gibbs and Summers 2002).Injection 30 min after training did not rescue the Aβ-inducedmemory loss (Fig.2a [F (6,97)=12.59; P<.001]).

In contrast to CL316243, zinterol which also normallyconsolidates weakly-reinforced learning 20 min after training,was unable to do so when Aβ 45 min was given prior totraining (Gibbs et al. 2010b). In Fig, 2B it is clear that zinterol(30 pmol/hem) is not able to rescue when injected 15, 20 or25 min after training [F (7,108)=4.32; P<.001], whereaswithout Aβ zinterol is normally able to consolidate weakly-reinforced learning injected up until 30 min post training(Gibbs and Summers 2002).

We challenged Aβ-induced memory loss 20 min aftertraining with a higher dose of zinterol – 100 pmol/hem,however even this high dose was unable to counter the mem-ory loss. The disc.ratio for 30 pmol/hem was 0.636±0.5compared to the disc.ratio of 0.60±.091 for the dose of 100pmol/hem. Because of the time period over which zinterolfails to rescue memory after Aβ, i.e. the time over whichzinterol is known to stimulate glycogenolysis, this result raises

Fig. 1 Retention after training with Aβ injected 45 min or 24 h beforetraining. Memory is retained in both conditions when tested 30 min aftertraining but is not present at 120 min. Loss is significant on test 35 minafter training

Metab Brain Dis

the possibility that Aβ interferes with either glycogenolysis orglycogen synthesis. The possibility that soluble Aβ interfereswith theβ2-adrenoreceptors seems unlikely as memory can berescued 2.5 to 10 min after training.

To confirm the proposition that zinterol does promotememory consolidation by initiating glycogenolysis around20 min but not at 2.5 min after training, the ability of zinterolto promote memory consolidation at these two time waschallenge by prior injection of the glycogenolysis inhibitorDAB (Fig. 3).

Saline or DAB (100 pmol/chick) were injected subcutane-ously either 5 min before or 20 min after weakly-reinforcedtraining. Zinterol injected to all groups either at 2.5 or 25 minafter training was only challenged by DAB at +25 but not at +2.5 min [F2,55=3.09; P=.035]. This supports the conclusionthat glycogenolyis is involved in memory processing 20 or25 min after training but not at the earlier times.

Is CL316243 working via stimulation of the β3-AR?

Activation ofβ3-ARs increases glucose uptake into astrocyteswithin 5 min (Hutchinson et al. 2007; Gibbs et al. 2008c) anddoes so via the GLUT1 transporter. Aβ is known to decreaseboth GLUT1 and GLUT3 transport in several brain regions(Hoojimans et al. 2007). To confirm that the β3-AR agonistacts via the β3-AR receptor we challenged the dose responsefunction of CL316243 with the selective β3-AR antagonistSR59230A (Fig. 4). Three groups of Aβ treated chicks(45 min before training) were injected into IMM with 0.03to 0.3 pmol/hem CL316243 twenty min after training. Asubmaximal dose of SR59230A (10 pmol/chick) was injectedsubcutaneously 5 min after training (Gibbs and Summers2002; Gibbs 2008). The CL316243 dose response functionwas shifted to the right in the presence of the antagonist. Therewas no interaction between drug and dose [F=1,53=0.71;P=.402. There were significant drug effects between salineand SR59230A treatment [F1,52=14.24; P=.001].

Aβ in different brain locations

It is known that several different brain regions are involved inmemory processing in the chick (Gibbs 2008). Aβ (10pmol/hem) was injected 45 min prior to training into theMedial Striatum (MSt) (Gibbs and Summers 2003), IMM(Gibbs and Summers 2002), hippocampus (HP; (Gibbs et al.2008a) or the Locus coeruleus (LoC) (Gibbs et al. 2010a). Thevolume of the injection varied according to the site with 5 μlinjections into each IMM; 2 μl injections into each MSt; 1 μlinjections into each hippocampus or one single 5 μl injectioninto the hindbrain between the two LoC, through the 4thventricle (Gibbs et al. 2010a). Aβ injections in all areas excepttheMSt resulted inmemory loss 120 min after training (Fig. 5;[F (4,72)=6.27;P=.001]).

Aβ injected into the LoC 45 min prior to training resultedin memory loss when the chicks were tested 40 min and laterafter training but there was no loss on test at either 10 or30 min after training, the same pattern as seen when Aβ wasinjected into the IMM (Fig. 6; [F (4,73)=11.99;P=.001]).

Fig. 2 Time of challenges to Aβinjected 45 min before training.A. The β3-AR agonist CL316243challenges Aβ when injectedfrom 30 min before to 20 minafter training. B. The β2-ARagonist zinterol only challengedthe effect of Aβ injected between−5 min to +10 min but not later

Fig. 3 Zinterol administered 2.5 min after training facilitates weaktraining, which is not challenged by DAB, however when given 25 minafter training the facilitation is challenged by DAB. Hence, zinterol’saction 25 min after training is likely to be via an action on glycogenolysis

Metab Brain Dis

Challenge of Aβ-induced memory loss in LoC by injectioninto the IMM

The next question asked was whether Aβ-induced memoryloss resulting from injection of Aβ into the LoC could berescued by injection of AR agonists elsewhere in the brain.

Memory loss was induced by injection of 10 pmol Aβ intothe area of the LoC 45 min prior to training. Twenty min aftertraining three groups of chicks received CL316243 (0.1, 0.3 or1.0 pmol/hem) or zinterol (30 pmol/hem) injected into IMM.The dose range for CL316243 was that used following Aβrather than that used for facilitating weak-learning, becausewe have found that CL316243 after Aβ in IMM rescuesmemory at lower doses (Gibbs et al. 2010b).

In contrast to the earlier findings zinterol given 20min aftertraining into IMM effectively rescued memory loss inducedbyAβ in the LoC (Fig. 7). Likewise, CL316243 (1 pmol/hem)also rescued memory loss [F=4,70=11.132;P< .001].Therefore, the memory loss induced in the LoC was effective-ly rescued by injection of zinterol as well as CL316243 in theIMM. The dose of CL316243 that was effective was the sameas that which consolidated weakly-reinforced training.

Discussion

Involvement of glycogen in the inability of β2-AR activationto rescue Aβ-impaired memory

Aβmemory impairment is not rescued a β2-AR agonist at thetime that zinterol normally promotes memory consolidationvia glycogenolysis and this implies that β2-AR stimulationdoes not result in glycogenolysis when it is needed about20 min after training. Although β2-AR stimulation can

Fig. 4 The ability of CL316243 to rescue from the effect of Aβ waschallenged by submaximal dose of the selective antagonist SR592230A,showing that the effect of CL316243 was due to action on β3-AR

Fig. 5 Memory loss 120 min after training with Aβ injected 45 minbefore training into different brain regions

Fig. 6 Retention tested at different time intervals, in separate groups ofchicks, following injection of Aβ into the locus coeruleus 45 min beforetraining

Fig. 7 The effect of Aβ injected into the locus coeruleus 45 min beforetraining was challenged by CL316243 or zinterol injected into IMM20 min after training. DMSO is the control (solid black)

Metab Brain Dis

increase glucose uptake into astrocytes between 30 min to 2 h(Gibbs et al. 2008c; Hutchinson et al. 2007) it does not do so atearlier times and 2 h is too late for it to be promoting memoryconsolidation by stimulating glucose transport or glycolysis.In the absence of Aβ there would have to be uptake of glucosefor the synthesis of glycogen in the longer term. However, theturnover of glycogen for immediate use is very rapid (secondsto minutes) and zinterol can deplete a readily available glyco-gen store within a short time −10 min (see Gibbs andHutchinson 2012). There is a readily accessible pool of gly-cogen in astrocytes with rapid turnover that is rapidly depletedif synthesis is inhibited (Walls et al. 2008)We have shown that10-15 min of β2-AR stimulation is sufficient to deplete thisreadily available astrocytic pool. Therefore we speculate thatzinterol may not be able to rescue memory after Aβ byincreasing glycogenolysis because the astrocytic glycogenstores have already been depleted.

As glycogen selectively supports de novo synthesis ofglutamate by pyruvate dehydrogenation and carboxylation inastrocytes, the outcome of reduced glycogenolysis will be adecrease in glutamate available for neuronal synaptic activa-tion. Neurons are unable to synthesize glutamate and gluta-mate is transferred from astrocytes to neurons as glutamine(Gibbs et al. 2007; Hertz 2013; Schousboe et al. 2013).

The inability of activation of β2-ARs to rescue memoryafter Aβ could either be at the receptor level, e.g. (Wang et al.2010; 2011) or be due to a lack of glycogen. Our finding thatzinterol can rescue Aβ-impaired memory when given close totraining seems to suggest that the receptor function is notimpaired in the short-term.

Glycogen synthesis

As well as controlling the breakdown of glycogen, noradren-aline via α2-ARs is also responsible for its synthesis (Hertzet al. 2007). We have evidence suggesting that in astrocytes itis the postsynaptic α2C-AR subtype that is involved in glyco-gen synthesis, whereas postsynaptic α2A-ARs inhibit nor-adrenaline reuptake – inhibiting the former blocks memory,whereas inhibiting the latter results indirectly in memoryenhancement (Hutchinson et al. 2011). Insulin, as well asincreasing the uptake of glucose, also increases glycogensynthesis in rat brain (Morgenthaler et al. 2008) and in chickastrocytes via Akt and GSK3 (Hutchinson et al. 2008).

In chick astrocytes, the response to insulin is slower thanthat toβ3-AR agonists and glucose uptake only increases after30 min of exposure (Gibbs et al. 2008c); this result suggeststhat enhancement of memory by insulin is not because itincreases glucose uptake. It has been suggested that the CNSinsulin receptor is functionally different from those in classicalinsulin target tissues, e.g. muscle, and memory improvementby insulin is not due to a direct effect on glucose metabolism

(Zhao et al. 2004) which raises the question of whether itcould be due to increased glycogen synthesis.

Glycogen synthesis and GSK3

Both insulin and α2-AR signaling within the cell involve PI3kinase (P13K). Activation of the PI3K pathway activates Akt(protein kinase B) and increases cellular glucose uptake(Johnston et al. 2003; Vannucci et al. 1998). In addition, Aktmodulates the glycogen synthase kinase (GSK)-3 pathway byphosphorylating GSK-3α at its serine 9 residue, therebyinactivating it. GSK3 exists in two closely related formsGSK3α and GSK3β - GSK3α is involved in the regulationof Aβ peptides, and GSK3β is involved in tau protein phos-phorylation. GSK3 negatively regulates glycogen synthase,the rate limiting enzyme for glycogen production. GSK3 is aconstitutively active kinase that phosphorylates a broad rangeof substrates including glycogen synthase (Hooper et al.2008). Overactivation of GSK-3 has been implicated in AD.Thus AKT activation stimulates glycogen synthesis.

Recently the importance of glycogen for long-term mem-ory formation has been demonstrated by Duran et al. (Duranet al. 2013). Mice lacking glycogen synthase in the brain thatshowed a significant deficiency in the acquisition of an asso-ciative learning task and in the concomitant activity-dependent changes in hippocampal synaptic strength.

Our results implicate astrocytes in Aβ-induced memoryloss and suggest that Aβ-impairment of glycogen synthesisvia GSK-3 leads to reduced glycogen levels and may accountfor the inability of β2-ARs and insulin to rescue Aβ-impairedmemory. Thus activation of β2-AR that normally promotesmemory consolidation by stimulating the breakdown of gly-cogen may be blocked after Aβ because there is no readilyavailable glycogen pool. Also because of overactivation ofGSK-3, insulin stimulation of glycogen synthesis does notoccur. Therefore increasing glycogen breakdown will notrescue memory after Aβ. Because glycogen has an importantrole in de novo synthesis of glutamate in astrocytes, long-termpotentiation and synaptic plasticity which are glutamate de-pendent will also be impaired. A recent report links elevatedGSK3 activity and subsequent inhibition of glycogen synthe-sis with mental retardation caused by Fragile X syndrome(Min et al. 2009). These authors suggest that activation ofGSK3 inhibits glycogen synthase and glycogen synthesisresulting in reduced glycogen stores and as a result there isno glycogenolysis, resulting in inhibition of learning andfailure of memory consolidation.

GSK3 is one of several kinases that have been implicated inAβ toxicity, as well as other neurodegenerative conditions(e.g. Hooper et al. 2008; King et al. 2014) and it has beenproposed that GSK3 inhibitors may be able to rescue cogni-tive impairment in these conditions (King et al. 2014).

Metab Brain Dis

β3-AR activation and rescue of Aβ-impaired memory:possible involvement of glucose metabolism

β3-AR activation normally promotes consolidation of memo-ry in chicks given weakly reinforced training is associatedwith an increase in glucose uptake into astrocytes (Gibbs et al.2008c) but not into neurons (Hutchinson, unpublished) within5min of incubation with theβ3-AR agonist CL316243 (Gibbset al. 2008c). β3-ARs are found on astrocytes from chickforebrain (Hutchinson et al. 2007; Catus et al. 2011) and inmouse astrocytes where activation stimulates glucose uptake(Catus et al. 2011). β3-AR mRNA is present in cortical andbrain stem areas of human brain although in relatively lowquantities compared to β1- and β2-AR mRNA (Rodriguezet al. 1995). β3-ARs increase glucose transport by activatingthe astrocytic/endothelial GLUT1 transporter rather than theneuronal GLUT3 or the insulin sensitive GLUT8 (avian) orGLUT 4 (mammalian) transporters.

Activation of β3-ARs by CL316243 rescues and/or pre-vents memory loss caused by Aβ in the chick and it does so ata lower dose of the agonist than that required to promoteconsolidation under normal conditions (Gibbs et al. 2010b)There is no a priori reason to suppose they could not beinvolved in human learning and memory and they do providea target for prevention or alleviation of the consequences ofearly AD, particularly as they appear to be sensitized or evenup-regulated after Aβ. β3-ARs are found in several regions ofrat brain (Summers et al. 1995). It is interesting to note thathuman myometrial β3-ARs are upregulated during pregnancyin humans (Rouget et al. 2005). A recent report showing thatenvironmental novelty prevents Aβ-induced impairment ofLTP in hippocampal slices (Li et al. 2013) found this resultcould be mimicked by the non selective β-AR agonist iso-prenaline, which stimulatesβ3- as well asβ2-ARs. The role ofβ3-ARs should be examined.

In conclusion, it appears that β3-AR stimulation is able torescue memory from the early exposure to Aβ1–42 and does sovia stimulation of glucose uptake. β2-AR stimulation is how-ever not able to rescue memory by increasing breakdown ofglycogen suggesting that glycogen synthesis maybe impairedby exposure to soluble Aβ. The question remains as to hownoradrenaline, stimulating β2-ARs, rescues memory forma-tion from Aβ within 5 min of injection, when it cannot do solater.

Refererences

Braak H, Del Tredici K (2004) Alzheimer's disease: intraneuronal alter-ations precede insoluble amyloid-beta formation. Neurobiol Aging25(6):713–718

Cambray-Deakin M, Pearce B, Morrow C, Murphy S (1988) Effects ofneurotransmitters on astrocyte glycogen stores in vitro. JNeurochem 51(6):1852–1857

Catus SL, Gibbs ME, Sato M, Summers RJ, Hutchinson DS (2011) Roleof β-adrenoceptors in glucose uptake in astrocytes using β-adrenoceptor knockout mice. Br J Pharmacol 162(8):1700–1715

Darvesh AS, Gudelsky GA (2003) Activation of 5-HT2 receptors inducesglycogenolysis in the rat brain. European Journal of Pharmacology464(2):135–140

Duran J, Saez I, As G, Guinovart JJ, Delgado-Garcia JM (2013)Impairment in long-term memory formation and learning-dependentsynaptic plasticity in mice lacking glycogen synthase in the brain.Journal of Cerebral Blood Flow & Metabolism 33(4):550–556

Gibbs M, Gibbs C (2013) Deleterious effects of soluble beta amyloid oncognition, antagonism by saline and noradrenaline, a role for mi-croglia. Neuroscience 230:62–71

Gibbs ME (2008) Memory systems in the chick: Regional and temporalcontrol by noradrenaline. Brain Res Bull 76:170–182

Gibbs ME, Anderson DG, Hertz L (2006) Inhibition of glycogenolysis inastrocytes interrupts memory consolidation in young chicks. Glia54:214–222

Gibbs ME, Bowser DN, Hutchinson DS, Loiacono RE, Summers RJ(2008a) Memory processing in the avian hippocampus involvesinteractions between b-adrenoceptors, glutamate receptors and me-tabolism. Neuropsychopharmacology 33:2831–2846

Gibbs ME, Gibbs Z, Hertz L (2009) Rescue of Aβ1-42-induced memoryimpairment in day-old chick by facilitation of astrocytic oxidativeenergy metabolism: Implications for Alzheimer's disease. JNeurochem 109 (Suppl.1):230–236

Gibbs ME, Hertz L (2014) Serotonin mediation of early memory forma-tion bia 5-HT2B receptor induced glycogenolysis in the day oldchick. Frontiers in Pharmacology. doi:10.3389/fphar.2014.00054

Gibbs ME, Hutchinson DS (2012) Rapid turnover of glycogen in mem-ory formation. Neurochem Res 37(11):2456–2463

Gibbs ME, Hutchinson DS, Hertz L (2008b) Astrocytic involvement inlearning and memory consolidation. Neurosci Biobehav Rev 32:927–944

Gibbs ME, Hutchinson DS, Summers RJ (2008c) Role of β-adrenoceptors in memory consolidation: β3-adrenoceptors act onglucose uptake and β2-adrenoceptors on glycogenolysis.Neuropsychopharmacology 33:2384–2397

Gibbs ME, Hutchinson DS, Summers RJ (2010a) Noradrenaline releasein the locus coeruleus modulates memory formation and consolida-tion; roles for alpha- and beta-adrenergic receptors. Neuroscience170(4):1209–1222

Gibbs ME, Lloyd HGE, Santa T, Hertz L (2007) Glycogen is a preferredglutamate precorsor during learning in 1-day-old chick: biochemicaland behavioral evidence. J Neurosci Res 85(15):3326–3333

Gibbs ME, Maksel D, Gibbs Z, Hou X, Summers RJ, Small DH (2010b)Memory loss caused by β-amyloid protein is rescued by a β3-adrenoceptor agonist. Neurobiol Aging 31:614–624

Gibbs ME, Summers RJ (2002) Role of adrenoceptor subtypes in mem-ory consolidation. Prog Neurobiol 67:345–391

Gibbs ME, Summers RJ (2003) α2-Adrenoceptors in the basal gangliahave a role in memory consolidation and reinforcement.Neuropharmacol 45:355–367

Guerin D, Sacquet J, Mandairon N, Jourdan F, Didier A (2009) Earlylocus coeruleus degeneration and olfactory dysfunctions in Tg2576mice. Neurobiol Aging 30(2):272–283

Hertz L (1989) Is Alzheimer's disease an anterograde degeneration, orig-inating in the brainstem, and disrupting metabolic and functionalinteractions between neurons and glial cells? Brain ResRev 14:335–353

Hertz L (2013) The glutamate-glutamine (GABA) cycle: importance oflate postnatal development and potential reciprocal interactions be-tween biosynthesis and degradation. Frontiers in endocrinology 4

Metab Brain Dis

Hertz L, Peng L, Dienel GA (2007) Energy metabolism in astrocytes:high rate of oxidative metabolism and spatiotemporal dependenceon glycolysis/glycogenolysis. J Cereb Blood Flow Metab 27(2):219–249

Hooper C, Killick R, Lovestone S (2008) The GSK3 hypothesis ofAlzheimer's disease. J Neurochem 104(6):1433–1439

Hoojimans CR, Graven C, Dederen PJ, Tanila H, van Groen T, Kilaan AJ(2007) Amyloid beta depsition is related to decreased glucosetransporter-1 levels and hippocampal atrophy in brains of agedAPP/PS1 mice. Brain Res 1181:93–103

Hutchinson DS, Catus SL, Merlin J, Summers RJ, Gibbs ME(2011) Adrenoceptors activate noradrenaline-mediated glyco-gen turnover in chick astrocytes. J Neurochem 117(5):915–926

Hutchinson DS, Summers RJ, Gibbs ME (2007) β2- and β3-adrenoceptors activate glucose uptake in chick astrocytes by distinctmechanisms: a mechanism for memory enhancement? J Neurochem103:997–1008

Hutchinson DS, Summers RJ, Gibbs ME (2008) Energy metabolism andmemory processing: role of glucose transport and glycogen inresponses to adrenoceptor activation in the chicken. Brain ResBull 76(3):224–234

Johnston AM, Pirola L, Van Obberghen E (2003) Molecular mechanismsof insulin receptor substrate protein-mediated modulation of insulinsignalling. FEBS Lett 546(1):32–36

King MK, Pardo M, Cheng Y, Downey K, Jope RS, Eo B (2014)Glycogen synthase kinase-3 inhibitors: rescuers of cognitive impair-ments. Pharmacology & therapeutics 141(1):1–12

KleinWL, Krafft GA, Finch CE (2001) Targeting small Abeta oligomers:the solution to an Alzheimer's disease conundrum? Trends Neurosci24(4):219–224

Kong EK, Peng L, Chen Y, Yu AC, Hertz L (2002) Up-regulation of 5-HT2B receptor density and receptor-mediated glycogenolysis inmouse astrocytes by long-term fluoxetine administration.Neurochem Res 27(1–2):113–120

Li S, Jin M, Zhang D, Yang T, Koeglsperger T, Fu H, Selkoe DJ (2013)Environmental Novelty Activates β2-Adrenergic Signaling toPrevent the Impairment of Hippocampal LTP by Aβ Oligomers.Neuron 77(5):929–941

Magistretti PJ (1988) Regulation of glycogenolysis by neurotransmittersin the central nervous system. Diabete & Metabolisme (Paris) 14:237–246

Marien MR, Colpaert FC, Rosenquist AC (2004) Noradrenergic mecha-nisms in neurodegenerative diseases: a theory. Brain ResRev 45(1):38–78

Mileusnic R, Lancashire C, Clark J, Rose SP (2007) Protection againstAbeta-induced memory loss by tripeptide D-Arg-L-Glu-L-Arg.Behav Pharmacol 18(3):231–238

Min WW, Yuskaitis CJ, Yan Q, Sikorski C, Chen S, Jope RS, BauchwitzRP (2009) Elevated glycogen synthase kinase-3 activity in Fragile Xmice: Keymetabolic regulator with evidence for treatment potential.Neuropharmacol 56(2):463–472

Morgenthaler FD, van Heeswijk RB, Xin L, Laus S, Frenkel H, Lei H,Gruetter R (2008) Non-invasive quantification of brain glycogenabsolute concentration. J Neurochem 107(5):1414–1423

Quach TT, Rose C, Duchemin AR, Schwartz JC (1982) Glycogenolysisinduced by serotonin in brain: identification of a new class ofreceptor. Nature 298:373–375

Rodriguez M, Carillon C, Coquerel A, Le Fur G, Ferrara P, Caput D,Shire D (1995) Evidence for the presence of beta 3-adrenergicreceptor mRNA in the human brain. Brain Res Mol Brain Res29(2):369–375

Rouget C, Bardou M, Breuiller-Fouche M, Loustalot C, Qi H, Naline E,Croci T, Cabrol D, Advenier C, Leroy MJ (2005) β3-Adrenoceptoris the predominant β-adrenoceptor subtype in human myometriumand its expression is up-regulated in pregnancy. J Clin EndocrinolMetab 90(3):1644–1650

Schousboe A, Bak LK, Waagepetersen HS (2013) Astrocytic control ofbiosynthesis and turnover of the neurotransmitters glutamate andGABA. Frontiers in endocrinology 4

Summers RJ, Papaioannou M, Harris S, Evans BA (1995) Expression ofβ3-adrenoceptors in rat brain. Br J Pharmacol 116:2547–2548

Vannucci SJ, Clark RR, Koehler-Stec E, Li K, Smith CB, Davies P,Maher F, Simpson IA (1998) Glucose transporter expression inbrain: relationship to cerebral glucose utilization. Dev Neurosci20:369–379

Walls AB, SickmannHM, BrownA, Bouman SD, RansomB, SchousboeA, Waagepetersen HS (2008) Characterization of 1,4-dideoxy-1,4-imino-d-arabinitol (DAB) as an inhibitor of brain glycogen shuntactivity. J Neurochem 105(4):1462–1470

Walsh DM, Selkoe DJ (2004) Deciphering the molecular basis of mem-ory failure in Alzheimer's disease. Neuron 44(1):181–193

Wang D, Govindaiah G, Liu R, De Arcangelis V, Cox CL, Xiang YK(2010) Binding of amyloid β peptide to β2 adrenergic receptorinduces PKA-dependent AMPA receptor hyperactivity. TheFASEB Journal 24(9):3511–3521

Wang D, Yuen EY, Zhou Y, Yan Z, Xiang YK (2011) Amyloidβ peptide-(1-42) induces internalization and degradation of β2 adrenergicreceptors in prefrontal cortical neurons. Journal of BiologicalChemistry 286(36):31852–31863

Weinshenker D (2008) Functional consequences of locus coeruleus de-generation in Alzheimer's disease. Current Alzheimer research 5(3):342–345

Zhao W-Q, Chen H, Quon MJ, Alkon DL (2004) Insulin and the insulinreceptor in experimental models of learning and memory. EuropeanJournal of Pharmacology 490(1–3):71–81

Metab Brain Dis