Article history:Received 14 May 2011Received in revised form 27 October 2011Accepted 4 November 2011Available online 12 November 2011
Keywords:Alzheimer's diseaseDiabetesLearning and memoryStem cellNeurogenesis
Type 2 diabetes is a risk factor for Alzheimer's disease. Insulin receptor desensitisation has been found in Alz-heimer brains, which may be the underlying link. Glucose-dependent insulinotropic polypeptide (GIP), anincretin hormone, normalises insulin signalling in diabetes. GIP and the GIP receptors are widely expressedin the brain, and GIP has been shown to have growth factor and neuroprotective properties. Here we inves-tigate the potential therapeutic properties of different doses of the protease resistant long-lasting GIP recep-tor agonist D-Ala2GIP and the antagonist (Pro3)GIP in C57Bl/6 mice. We found that after acute injection, D-Ala2GIP had few effects on general behaviour in the open field at any dose tested (2.5, 25, 100, or 250 nmol/kg i.p.). In memory tests, no change was observed, whilst (Pro3)GIP at 25 nmol/kg i.p. impaired memory for-mation. In a chronic study over 4 weeks, mice injected with D-Ala2GIP (2.5 or 25 nmol/kg i.p.) and (Pro3)GIP(25 nmol/kg i.p.) learned a water maze task and object recognition task without impairment. In LTP record-ing in area CA1, both (Pro3)GIP as well as D-Ala2GIP enhanced LTP formation. In addition, the proliferation ofneuronal progenitor cells in the dentate gyrus was increased both by D-Ala2GIP and (Pro3)GIP. The resultsshow that the antagonist (Pro3)GIP has agonistic effects in chronic use, and both (Pro3)GIP and the agonistD-Ala2GIP are safe to use in wt mice and induces no major behavioural side effects nor impairments in learn-ing whilst enhancing LTP and neuronal progenitor cell proliferation, which may be useful in treating neuro-degenerative diseases.
Incretins such as the Glucose-dependent insulinotropic polypep-tide (GIP) are currently under investigation as a potential treatmentfor diabetes. Glucose-dependent insulinotropic peptide (GIP) is a 42amino acid peptide, which belongs to the secretin-glucagon growthfactor family of gastrointestinal regulatory peptides (Baggio andDrucker, 2007; Drucker and Nauck, 2006). The hormone GIP en-hances the release of insulin under hyperglycemic conditions andhelps to re-sensitise the insulin response (Baggio and Drucker,2007). To this effect, novel long-acting analogues of GIP have beendeveloped that have a much longer biological half life than the nativepeptide (Gault et al., 2008). However, GIP has additional physiologicalactivities and also acts as a growth factor and neurotransmitter. GIPpromotes growth, differentiation, proliferation and survival ofbeta-cells (Irwin et al., 2006; Trumper et al., 2001) and of neuronalprogenitor cells in the dentate gyrus (Nyberg et al., 2005). In addition,GIP and GIP receptor expression have been found in the large pyrami-dal neurons in the cortex and the hippocampus (Nyberg et al., 2005,
ol of Biomedical Sciences, Cro-nited Kingdom. Tel.: +44 28
).
rights reserved.
2007). GIP mRNA and GIP receptor mRNA are expressed in several re-gions including the hippocampus, cerebellum and olfactory system(Nyberg et al., 2007; Usdin et al., 1993). We also recently showedthat novel stable analogues such as n-acetylated-GIP enhance synap-tic plasticity (LTP) in the hippocampus, whilst the antagonist (Pro3)GIP reduced LTP. Importantly, the detrimental effects that beta-amyloid has on synaptic plasticity was reversed by the GIP analogueN-AcGIP (Gault and Holscher, 2008). Co-administration of GIP andAβ1-40 abolishes the severe impairments of spatial learning andmemory induced by intracerebroventricular infusion of Aβ (1–40)during the water maze task (Figueiredo et al., 2010). GIP also has neu-roprotective and neuro-regenerative properties. In one study, GIP hasbeen shown to promote axonal regeneration after sciatic nerve injury.GIP receptor-deficient mice were associated with impaired spontane-ous nerve regeneration compared with controls (Buhren et al., 2009).GIP also activates the proliferation of neuronal progenitor cells andtherefore may contribute to neurogenesis. Chronic intracerebroven-tricular infusion of GIP increased the rate of progenitor cell prolifera-tion in the hippocampus of adult rats. In contrast, GIP receptor KOmice displayed a marked decrease in cell proliferation in the DENTATEGYRUS (Nyberg et al., 2005, 2007). An increase of apoptosis was alsofound in the hippocampus of rats immunised against GIP (Tian etal., 2010). Since other incretins such as Glucagon-like peptide-1shows protective effects in mouse models of Alzheimer's disease
295E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
(McClean et al., 2011), we postulate that GIP also could potentially actas a neuroprotective agent in neurodegenerative diseases (Holscher,2010).
We therefore tested several doses of the GIP receptor agonistD-Ala2GIP (Hinke et al., 2001) and the antagonist (Pro3)GIP (Gaultet al., 2008) in wild type C57Bl/6J mice to estimate the effects on gen-eral behaviour, memory formation, synaptic plasticity in area CA1,and on progenitor cell proliferation in the dentate gyrus of thehippocampus.
2. Materials and methods
2.1. Animals
Eleven-week-old C57Bl/6J male mice from Harlan laboratory wereused for the experiments described. Mice weighed 20 g±1 at thestart of the experiment. Animals were maintained on a 12/12 hlight–dark cycle (lights on at 08 h00, off at 20 h00). Animals were in-dividually caged, and received food and water ad libitum. After twoweeks of acclimatisation, mice were handled for two weeks prior tothe study. All animal experiments were licenced by a UK Home Officeproject licence in agreement with UK and EU laws.
2.2. Drug administration and procedures
2.2.1. Acute studyThe C57Bl/6J mice were injected at a volume of 10 ml/kg with sa-
line solution (0.9% NaCl) (control group), (Pro3)GIP (25 nmol/kgbody weight) or D-Ala2GIP at four different doses: 2.5 nmol/kg,25 nmol/kg, 100 nmol/kg or 250 nmol/kg of body weight. Eachmouse received intraperitoneal injection 30 min before the beha-vioural task. A set of mice was tested in the object recognition task(ORT) after a 3 h delay. Two weeks later, the same set of animalswas assessed in the Object Location Task (OLT) with 5 min delay be-tween acquisition and testing phase. A different set of mice was test-ed in ORT after a 24 h delay and object location memory of mice wasassessed in OLT (3 h delay before testing) 2 weeks later. As the sameenvironment and apparatus were used for the ORT and OLT, a delay oftwo weeks was allowed between both tasks, in order to decrease theimpact that the previous environment could have on the OLT and toleave sufficient time to wash out the peptide, as only acute effectsof the drugs were studied in this part
2.2.2. Chronic studyThe C57Bl/6J mice were injected intraperitoneally at a volume of
10 ml/kg with saline solution (0.9% NaCl) (control group), (Pro3)GIP(25 nmol/kg body weight) or D-Ala2GIP at two different doses(2.5 nmol/kg and 25 nmol/kg body weight) once daily (the evening)for 30 days.
The behavioural task commenced on day 22 by the MWM and wasfollowed by the open-field task on day 29 and object recognition taskon day 30 with delay of 3 h between the acquisition task and the testtask. Half of the animal was tested for their synaptic plasticity byelectrophysiological techniques and the half left was used forimmunohistochemistry.
2.3. Behavioural tasks
2.3.1. Object recognition task (ORT)The apparatus consisted of an open-field arena (58 cm in diame-
ter; 31 cm high walls) constructed in aluminium with painted greywalls and grey floor. The open-field was dimly illuminated by a 60-W lamp placed 2 m directly above the arena. The objects were redcubes (1.8 cm wide) and white balls (2.6 cm). The arena and objectswere cleaned with 70% of ethanol after each mouse trial to preventthe build-up of olfactory cues. Mice received a session of 5 min in
the empty open-field to habituate them to the apparatus and testroom. The mouse was placed in the middle of the open-field andwas free to explore it. During that time, motor activity was recordedby total path, number of lines on the floor crossed and speed. Thenumber of rearing events (forepaws elevated from the floor) wasanalysed as index of exploratory behaviour. The anxiety level wasassessed by the percentage of time spent in the centre versus periph-ery of the arena and the number of grooming sessions. Grooming wasdefined as behaviour when the mouse stopped running and start lick-ing, chewing and scratching at its fur. Twenty four hours after the ha-bituation each mouse was subject to a 10 min acquisition trial, duringwhich they were placed in the open-field in the presence of two iden-tical objects A (cube or ball) situated at 15 cm from the arena wall(Acquisition task). On completion of 10 min exploration, the mousewas returned to its cage for a 3 h or 24 h delay. After retention inter-val, the mice were placed back into the box and exposed to the famil-iar object A and to a novel object B for a further 10 min (Test task).The objects were placed in the same locations as the previous ones.The position of the novel object was fully counterbalanced (half left,half right) in a random manner in order to avoid preferences notbased on novelty (see Fig.). Locomotor activity (number of linescross), speed (cm/s), travel path and the total time spent exploringeach of the two objects (when the animal's snout was directly towardthe object at a distance ≤2 cm), were recorded. A recognition indexwas defined as the amount of time exploring the familiar object orthe novel object over the total time spent exploring both objectstimes 100 was used to measure recognition memory: (TA or TB/(TA+TB))∗100. In the acquisition and retention trial, if the explora-tion time was b30 s and b15 s respectively, the mice were excludedfrom the trial. A video camera was fixed 2 m above the centre pointof the arena and was attached to a video recorder, monitor and acomputer. The movement of the animals in the open-field wastracked using computerised tracking system (Biosignals, New York)to analyse the images.
2.3.2. Object location task (OLT)The apparatus and procedures were the same as in the ORT. The
objects were blue cubes (1.8 cm wide) and black-red bottle tops(.,2 cm diameter; 2.8 cm high), which were randomised betweenmice. After a session of 5 min habituation in the open-field, the ani-mals were submitted to a 10 min acquisition trial with two identicalobjects (pair of cubes or pair of bottle tops). After a delay of 5 minor 3 h, the mice received a second trial (Test trial) identical to thefirst trial except that one of the objects was placed in the new loca-tion. The measurement and Tracking were the same as in the ORT.
2.3.3. Morris water maze taskThe maze was made of white opaque plastic with a diameter of
120 cm and 40 cm high walls, and was filled with water at 25 °C toavoid hypothermia. A small escape platform (10×6.5×21.5 cm) wasplaced at a fixed position in the centre of one quadrant, 25 cm fromthe perimeter, and was hidden 1 cm beneath the water surface. Theroom contained a number of fixed visual cues on the walls. The acqui-sition trial phase consisted of 4 training days (Days 1–4) and four tri-als per day with a 15-min inter-trial interval. Four points equallyspaced along the circumference of the pool (North, South, East,West) served as the starting position, which was randomised acrossthe four trials each day. If an animal did not reach the platform within90 s, it was guided to the platformwhere it had to remain for 30 s, be-fore being returned to its home cage. Mice were kept dry, betweentrials, in a plastic holding cage filled with paper towels. The pathlength and escape latencies were recorded. One day after finishingthe acquisition task (Day 5), a probe trial was performed in order toassess the spatial memory (after a 24 h delay). The platform was re-moved from the maze and animals were allowed to swim freelyfor 60 s.
296 E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
2.4. Surgery and LTP recording in the hippocampus, area CA1
The technique used for testing LTP in the hippocampus was exact-ly as described in (Gengler et al., 2012). Mice were anaesthetised withurethane (ethyl carbamate, 1.8 g/kg, i.p.) for the duration of all exper-iments. Electrodes (tungsten with Teflon coating, Bilaney, UK) wereimplanted at coordinates: 1.5 mm posterior and 1.0 mm lateral forthe recording electrode, and the stimulating electrode 2.0 mm poste-rior to bregma and 1.5 mm lateral to the midline. The electrodes werelowered through the cortex and the upper layers of the hippocampusand into the CA1 region until the appearance of a negative deflectingexcitatory postsynaptic potential (EPSP) that had a latency of ca.10 ms. fEPSPs were recorded on a computerised stimulating and re-cording unit (PowerLab, ADI instruments, USA). The program activat-ed a constant current stimulus isolation unit (Neurolog, UK). Theweak high frequency stimulation (HFS) protocol for inducing LTPconsisted of 3 trains of 100 stimuli at 200 Hz, inter-train interval 1 s,interstimulus interval 5 ms, with the stimulation strength set at 50%of the maximum fEPSP response. This weak protocol was used in
Fig. 1. Measurement of spontaneous behaviour during open-field task of C57Bl/6 mice injec250 nmol/kg 30 min before the task. (A) path length, (B) number of lines crossed, (C) numband (F) the ratio of time spent in the centre of the arena to the periphery. Data represent m
order to assess if peptides could facilitate LTP as the control groupwas not potentiated at a maximal rate (Gault and Holscher, 2008).The stronger high frequency stimulation protocol consisted of 3 trainsof 200 stimuli, inter-train interval 1 s, interstimulus interval 5 ms(200 Hz), with the stimulation strength set at 75% of the max. fEPSPresponse. LTP was measured as % of baseline fEPSP. This strong HFSprotocol was found to potentiate LTP at a higher level under thesestimulation conditions (Gault and Holscher, 2008).
2.5. Immunohistochemistry
Animals were administered BrdU (180 mg/kg bw; i.p.) 18 h priorto being anaesthetised with pentobarbitone (0.3 ml; Euthanal, BayerAG, Leverkusen, Germany) and perfused transcardially with PBS fol-lowed by 4% paraformaldehyde. The brains were removed and putinto 30% sucrose in PBS overnight. Immunohistochemistry for BrdUwas performed on 45 μm free floating sections. Endogenous peroxi-dase activity was quenched by incubation of sections in 3% hydrogenperoxide. Denaturation of DNA involved incubation in 2 N HCl,
ted i.p. with saline, (Pro3)GIP, or D-Ala2GIP at 2.5 nmol/kg, 25 nmol/kg, 100 nmol/kg orer of rearings, (D) speed, (E) anxiety levels as measured in number of grooming eventsean±S.E.M. of 30 mice per group.
297E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
followed by 0.1 M borax for 10 min. Sections were incubated in a pri-mary antibody for BrdU (1:200, mouse monoclonal anti-BrdU, Sigma)overnight at 4 °C. Then secondary antibody (1:200, horse anti-mouse,Vector elite ABC kit, mouse, Vector laboratories) was applied. Sectionswere incubated in an avidin biotin enzyme reagent and incubated inVector SG substrate chromogen (see Gengler et al., 2012 for details).The sections were analysed using an Olympus CX 40 microscope,using stereological techniques. This involves starting the sectioningrandomly and collecting every 5th section throughout the granulecell layer (GCL) of the dentate gyrus was analysed using a ×40 objec-tive and representative images were taken using a 5.1 MPix digitalcamera. For each group, 4 mice brains were analysed. Between
Fig. 2. Object recognition memory tested 3 h after training. Recognition index (RI) for famS.E.M. of saline group (n=15) (A), (Pro3)GIP group (n=15) (B), and D-Ala2GIP groups a250 nmol/kg (n=14) (F); (t-test, *Pb0.05, **Pb0.01).
8 and 12 sections were taken from each brain. All positive cells inthe DENTATE GYRUS were counted using ImageJ software (freewareof the NIH, http://rsbweb.nih.gov/ij/).
2.6. Statistics
Statistical analyses were performed using Prism (Graphpad soft-ware Inc. USA) with the level of probability set at 95%. Standard er-rors of the means are shown in the figures. Data from the open-fieldand immunostaining were analysed by one-way ANOVA followed bya Bonferroni post-hoc when Pb0.05 to measure the difference be-tween groups. Student's paired t-test was used for the ORT and
iliar and novel object during the recall test (after a 3 h delay). Data represent mean±t 2.5 nmol/kg (n=12) (C), 25 nmol/kg (n=16) (D), 100 nmol/kg (n=16) (E), and
298 E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
OLT when the time spent exploring the familiar object/location wascompared to the time spent exploring the novel object/location. Atwo-way repeated measures ANOVA was used to analyse differencesbetween groups, effects over time and interactions for every behav-iour measure (path length, escape latency and swim speed) of ac-quisition task and to determine the difference of time spentbetween quadrants for each group in the probe trial, followed by aBonferroni post-hoc test. To evaluate difference in synaptic activitybetween groups, a two-way repeated measures ANOVA was per-formed for the post-HFS baseline with group as independent vari-able and time as dependent variable. The same analysis wasperformed for the pre-HFS baseline to assess difference betweengroups over time.
Fig. 3. Object recognition memory tested 24 h after training. Recognition index (RI) for famS.E.M. of saline group (n=12) (A), (Pro3)GIP group (n=12) (B) and D-Ala2GIP groups250 nmol/kg (n=13) (F); (t-test, **Pb0.01).
3. Results
3.1. Acute study
3.1.1. Open-field task
3.1.1.1. Motor activity. There was no difference in the distance coveredin the 5 min arena in the open-field arena (one-way ANOVA,F=2.227, P>0.05; Fig. 1A). However, mice injected with D-Ala2GIPat 2.5 nmol/kg and 100 nmol/kg showed a difference in the numberof line crossed compared to the control group (F=2.787, Pb0.05;Fig. 1B). No difference in speed was found between groups(F=2.226, P>0.05; Fig. 1D).
iliar and novel object during the recall test (after a 24 h delay). Data represent mean±at 25 nmol/kg (n=14) (C), 25 nmol/kg (n=14) (D), 100 nmol/kg (n=13) (E), and
299E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
3.1.1.2. Exploratory activity. There was no difference in rearing be-tween groups (F=0.9574, P>0.5; Fig. 1C).
3.1.1.3. Anxiety levels. The number of grooming episodes was not dif-ferent between groups (F=1.289, P>0.05; Fig. 1E). However theD-Ala2GIP group at the concentrations of 100 nmol/kg spent lesstime in the centre than near the walls of the arena, as compared tothe D-Ala2GIP group at 25 nmol/kg (F=3.301, Pb0.05; Fig. 1F).
3.1.2. Object recognition taskAfter a delay of 3 h, there was a difference in the recognition index
(RI) of novel vs. familiar object for the saline group (Student's pairedt-test, t=3.536, Pb0.01; Fig. 2A), D-Ala2GIP groups at 2.5 nmol/kg(t=2.055, Pb0.05; Fig. 2C), at 25 nmol/kg (t=1.910, Pb0.05; Fig. 2D)and at 250 nmol/kg (t=1.995, Pb0.05; Fig. 2F). However, the (Pro3)GIP (t=0.3769, P>0.05; Fig. 2B) and D-Ala2GIP at 100 nmol/kg(t=0.4226, P>0.05; Fig. 2E) did not display a bias for the novel objecttowards the familiar one as shown in the RI close to 50%.
Fig. 4. Object location memory tested 5 min after training. Recognition index (RI) for familiaS.E.M. of saline group (n=12) (A), (Pro3)GIP group (n=12) (B) and D-Ala2GIP groups at 2250 nmol/kg (n=12) (F); (t-test, **Pb0.01, ***Pb0.0001).
After a delay of 24 h, only saline group (t=3.443, Pb0.01; Fig. 3A)and D-Ala2GIP groups at 25 nmol/kg (t=2.397, Pb0.01; Fig. 3D)spend more time exploring the novel object than the familiar one.
3.1.3. Object location taskAll groups spend more time exploring the object place in the novel
location after a delay of 5 min and 3 h (Pb0.01 and Pb0.0001)(Fig. 4). However, a decrease in time of exploration for the noveltycould be observed for the groups injected with D-Ala2GIP at highdoses (100 and 250 nmol/kg) when mice where tested 3 h after theacquisition task (Fig. 5).
3.2. Chronic study
3.2.1. Morris water maze taskAll mice learned to locate the hidden escape platform and showed
a decrease of escape latency across days of training in the acquisitiontrial (two-way ANOVA, groups: F=15.22, Pb0.0001; days: F=28.22,
r and novel location during the recall test (after a 5 min delay). Data represent mean±.5 nmol/kg group (n=14) (C), 25 nmol/kg (n=12) (D), 100 nmol/kg (n=12) (E) and
Fig. 5. Object location memory tested 3 h after training. Recognition index (RI) for familiar and novel location during the test task (after a 3 h delay). Data represent mean±S.E.M. ofsaline group (n=12) (A), (Pro3)GIP group (n=12) (B) and D-Ala2GIP groups at 2.5 nmol/kg (n=14) (C), 25 nmol/kg (n=13) (D), 100 nmol/kg (n=12) (E) and 250 nmol/kg(n=12) (F); (t-test, *Pb0.05, **Pb0.01, ***Pb0.0001).
300 E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
Pb0.0001, Fig. 6A). A Bonferroni post-hoc test compared the escapelatency of different groups per day, and found a significantdifference between the D-Ala2GIP group at 2.5 nmol/kg compared tothe control group on day 4 (Pb0.01), on day 5 (Pb0.05) and on day6 (Pb0.05). A two-way ANOVA analysis was conducted comparingthe escape latency between the (Pro3)GIP group and the salinegroup. A significant difference of escape latency was found betweengroups (F=9.43, Pb0.01), over time (F=16.80, Pb0.0001) with asignificant difference on day 3 (Pb0.05). The same analyses weredone between the D-Ala2GIP group (25 nmol/kg) and the salinegroup and showed significant difference of escape latency betweengroups (F=8.11, Pb0.01), over time (F=15.75, Pb0.0001), with asignificance difference of escape latency on day 3. There was a signif-icant decrease in path length across training sessions (F=15.41,Pb0.0001; Fig. 6B), although no differences were detected betweengroups (F=1.927, p=0.1633, ns). Group injected with 2.5 nmol/kgwere slower than control group with significant difference on Days3 to 6 (two-way ANOVA; Pb0.0001; post-hoc test Pb0.01, Pb0.001)whereas no deficiencies in swimming abilities were found for theother group (F=2.249, P>0.05; Fig. 6C).
In the probe trial, the overall analysis revealed that mice spendhigher proportion of time searching in the quadrant that containedthe escape platform during training (F=22.32, Pb0.0001; Fig. 6D).No difference was found between groups (F=0.6569, P>0.05) indi-cating an intact spatial memory for the previous location of the es-cape platform for all the groups.
3.2.2. Open-field taskOnly the group injected with 2.5 nmol/kg of D-Ala2GIP differed in
the parameters analysed compared to the control group (Fig. 7).
3.2.2.1. Motor activity. There was no difference between groups in thedistance covered in the 5 min arena (one-way ANOVA, F=0.8093,P>0.05; Fig. 7A). Groups were also identical in the number of linecrossed (F=0.9490, P>0.05; Fig. 7B) and in speed (F=0.8110,P>0.05; Fig. 7D).
3.2.2.2. Exploratory activity. There was no difference in number ofrearings (F=0.641, P>0.05) between groups (Fig. 7C).
Fig. 6.Water maze task performance after 21 days of drug injection. (A) escape latency, (B) Path length and (C) swim speed over 6 days during acquisition training (2-way ANOVA;Pb0.0001; post-hoc test *Pb0.05, **Pb0.01, ***Pb0.0001). Data represent mean±S.E.M. of 14 D-Ala2GIP (2.5 nmol/kg) mice and 12 for the other groups (t-test, *Pb0.05 andΔPb0.05 between (Pro3)GIP group and control group and Δ=Pb0.05 between D-Ala2GIP (25 nmol/kg) group and control group). (D) Time spent in each quadrant of the watermaze during the probe trial (2-way ANOVA). Data represent mean±S.E.M. of 14 D-Ala2GIP (2.5 nmol/kg) mice and 12 for the other groups.
301E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
3.2.2.3. Anxiety levels. The number of grooming episodes was not dif-ferent between groups (F=0.8182, P>0.05; Fig. 7E). However, micetreated with D-Ala2GIP at 2.5 nmol/kg display a significant lowerratio of time spent in the centre of the arena indicating a higher anx-iety level for this group compared to the control group (one-wayANOVA, F=3.209; p=0.0316; post-hoc test, Pb0.05; Fig. 7F).
3.2.3. Object recognition taskIn the test trial, there was a difference in the recognition index
(RI) of novel vs. familiar objects for the saline, (Pro3)GIP andD-Ala2GIP (25 nmol/kg) group (paired t-test, t=3.228, t=2.362,t=3.356 respectively, Pb0.05 and Pb0.01; Fig. 8) whilst no
difference was found for the group receiving DAla2GIP at 2.5 nmol/kg reflecting an impairment in recognition memory for this group(t=0.0183, P>0.05; Fig. 8C).
3.2.4. Assessment of LTP in area CA1 of the hippocampusA two-way repeated measure ANOVA showed a difference be-
tween pre-HFS baseline over time (F=2.72, Pb0.0001), but did notfind an effect between groups (F=2.45, P>0.05) nor an interactionof group and time (F=0.99, P>0.05), indicating that pre-HFF base-lines were similar between groups (Fig. 9A). A two-way repeatedmeasure ANOVA did not find a difference between baselines post-HFS over time (F=1.1311, P>0.05) but found a difference between
Fig. 7. Measurement of spontaneous behaviour during open-field task of mice injected i.p. with saline, (Pro3)GIP, D-Ala2GIP at 2.5 nmol/kg or 25 nmol/kg during 27 days. (A) pathlength, (B) number of lines crossed, (C) number of rearings, (D) speed, (E) anxiety levels as measured in number of grooming events and (F) the ratio of time spent in the centre ofthe arena to the periphery Data represent mean±S.E.M. of 14 DAla2GIP (2.5 nmol/kg) mice and 12 for the other groups (t-test, *Pb0.05).
302 E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
groups (F=132.6, Pb0.0001). When we compared the control groupwith the (Pro3)GIP group and the D-Ala2GIP group with the (Pro3)GIPgroup, a two-way repeated measures ANOVA revealed that weak HFSstimulation of the Schaffer collaterals induced a significant robust andlong-lasting LTP in (Pro3)GIP group compared to both groups(F=131.4, Pb0.0001 and F=182.2, Pb0.0001, respectively), whereno LTP was found (F=0.0005, P>0.05). (Fig. 9A).
In a separate study, the pre-HFS baselines of WT mice were ana-lysed by a two-way repeated measures ANOVA (Fig. 9B). No differ-ence was found between the saline group and D-Ala2GIP group(F=1.24, P>0.05), and no effect of time (F=0.74, P>0.05) and nointeractive effect of group and time was determined (F=0.61,P>0.05), indicating that pre-HFF baselines were stable and similarbetween groups over time. A two-way repeated measure ANOVAdid not find a difference between post-HFS baselines over time(F=0.11, P>0.05; Fig. 13) but found a difference between groups(F=162.4, Pb0.0001), indicating that chronic injection of D-Ala2GIPat 25 nmol/kg displayed a stable and long lasting LTP compared tothe control group.
3.2.5. Increase of stem cell proliferation with GIP analogues(Pro3)GIP and D-Ala2GIP (25 nmol/kg) treated mice exhibited
more BrdU-positive cells than control mice (Pb0.05 and Pb0.001 re-spectively). In the Dentate gyrus, D-Ala2GIP treated mice had signifi-cantly higher number of BrdU-positive cells than (Pro3)GIP treatedmice (Pb0.0001; Fig. 10).
4. Discussion
The widespread distribution of GIP receptors in the brain has beenknown for some time (Nyberg et al., 2005); however what roles thesereceptors play in normal brain function has been unclear. Thus, in ourstudy we tested the effects of systemic administration of GIP ana-logues, both acutely and chronically, on cognition, synaptic plasticityand neurogenesis in the hippocampus of normal (wild type) mice. Forthe acute study, mice were injected with peptides, 30 min beforeassessing their spontaneous behaviour in the open-field task as wellas before the acquisition trial of the object recognition task (ORT)and object location task (OLT). The injection protocol was chosensince a study conducted in our group showed that i.p. injection ofD-Ala2GIP at 25 nmol/kg in C57Bl/6J mice resulted in increased levelsof this peptide in the brain, 30 min and 3 h post-injection, confirmingthat D-Ala2GIP could cross the blood brain barrier (Hunter andHolscher, 2011). Different doses of D-Ala2GIP, a very low dose(2.5 nmol/kg) and very high doses (100 nmol/kg and 250 nmol/kg)were tested in order to assess if GIP analogues have side effects onmouse behaviour and to define which concentration of D-Ala2GIP isthe most effective. Our results show that D-Ala2GIP peptide at thedose 25 nmol/kg and 250 nmol/kg as well as (Pro3)GIP at 25 nmol/kg i.p. do not affect motor activity and exploratory behaviour ofmice after acute or chronic application in comparison to the controlgroup. This suggests that these GIP analogues at this concentrationdo not have aversive side effects on mouse behaviour. However,mice injected with 100 nmol/kg showed a decrease in the number
Fig. 8. Object recognition memory of mice injected with saline, (Pro3)GIP, D-Ala2GIP at 2.5 nmol/kg or 25 nmol/kg once-daily for 29 days. Recognition index (RI) for familiar andnovel objects during the recall task (after a 3 h delay). Data represent mean±S.E.M. of saline group (n=11) (A), (Pro3)GIP group (n=12) (B) and D-Ala2GIP groups at2.5 nmol/kg (n=12) (C), 25 nmol/kg (n=11) (D) (t-test, *Pb0.05; **Pb0.01).
303E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
of line crossings and spent less time in the centre of the arena com-pared to the control group, showing a slightly higher anxiety level.
The learning and recognition memory of mice have been assessedby the object recognition task, based on the spontaneous, differentialexploration of familiar and novel objects (Dere et al., 2007). Adminis-tration of peptides prior to the acquisition trial allowed us to assessthe effects of these peptides on the encoding of object characteristicsand early stages of memory consolidation (Ennaceur, 2009). Addi-tionally, different delays between the acquisition trial (learning)and the test trial (memory recall) have been employed to assessshort-term and long-term object location/recognition memory. Ourresults indicate that object recognition memory was intact over a3 h test period in D-Ala2GIP groups, and high doses of this drugwere well tolerated by the mice. The group injected with (Pro3)GIPfailed to investigate the novel object in ORT after acute administra-tion, indicating that this GIP receptor antagonist impairs memory re-tention. Lack of exploration was also noted in mice injected withD-Ala2GIP at 100 nmol/kg. The result could be explained by the factthat these mice were more anxious than the other groups, alteringhow they did learn this task. When the object recognition memorywas tested 24 h after the acquisition trial, only mice injected with25 nmol/kg D-Ala2GIP learned the task. However all groups shownormal performance in the OLT after a short or a long retentiondelay. This suggests that all groups not only encoded and maintainedthe spatial location in which it was encountered but also showed anintact short-term and long-term object location memory.
The injection protocol of 21 days of D-Ala2GIP was chosen, sincethis treatment has been found to improve insulin signalling and glu-cose tolerance in both normal and obese diabetic rodents (Green etal., 2004; Hinke et al., 2002). Additionally, prolonged administration
of 25 nmol/kg D-Ala2GIP did not lead to GIP-receptor desensitisation(Porter et al., 2010) and did not affect blood glucose levels(Figueiredo et al., 2010; Irwin et al., 2006). Spontaneous behaviour,tested over a 5 min period in the open-field task, showed that chronicadministration of D-Ala2GIP and (Pro3)GIP at 25 nmol/kg does nothave aversive effects on general behaviour, as motor activity, explor-atory behaviour and anxiety levels did not differ from the controlgroup. However, mice injected with 2.5 nmol/kg spent less time inthe centre of the arena compared to the control group, indicating aslightly higher anxiety level.
No impairment in spatial water maze long-term memory wasfound in any of the groups in the chronic study. Mice injected with2.5 nmol/kg of D-Ala2GIP were slower to find the platform than con-trol mice, which could be due to their decreased swim speed duringthe acquisition task, as their spatial memory is intact 24 h after thelast training day. In contrast, mice injected chronically withD-Ala2GIP (25 nmol/kg) or (Pro3)GIP (25 nmol/kg) learned the taskslightly quicker. Both drugs therefore seem to improve the learningof spatial memory slightly, though the overall distance swum toreach the target was not different. This information is consistentwith the finding that GIP-overexpressing transgenic mouse showedincreased learning in a Y-maze task (Ding et al., 2006). However, ef-fects found with the GIP receptor antagonist (Pro3)GIP were unex-pected. In fact, (Pro3)GIP seems to have the same beneficial effectthan D-Ala2GIP in learning and memory when injected chronically.The physiological effects of this GIP receptor antagonist in chronic ad-ministration are those of an agonist. (Pro3)GIP is known to block GIPsignalling and inhibit GIP-induced insulin secretion with up to 86%maximal inhibition (Gault et al., 2003). However, (Pro3)GIP wasalso found to be effective in treating diabetes and induced an
Fig. 9. In vivo LTP recording in area CA1 of mice injected for 30 days with saline, 25 nmol/kg of (Pro3)GIP or 25 nmol/kg or D-Ala2GIP. (A) The GIP receptor antagonist (Pro3)GIP facil-itated LTP induced by a weak HFS protocol in the area CA1 of the hippocampus of C57Bl/6 mice (2 way repeated measures ANOVA, Pb0.0001 between the saline and (Pro3)GIP group,and between the DAla2GIP and (Pro3)GIP group). (B) Chronic injection of the GIPR agonist DAla2GIP facilitated LTP induced by a weak HFS protocol compared to the control group(2 way repeated measures ANOVA, Pb0.01). Data represent mean±S.E.M. of 6 mice per group.
304 E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
improvement of insulin resistance and a normalisation of blood glu-cose levels in chronic treatment of diabetic mice (Gault et al., 2005;Irwin et al., 2007). Also, we found a clear enhancement of LTP afterchronic injection of (Pro3)GIP in the present study. This is in contra-diction to the inhibitory action on LTP that (Pro3)GIP showed inacute application (Gault and Holscher, 2008). Furthermore, the en-hancement of neuronal progenitor cell proliferation in the Dentategyrus by (Pro3)GIP also mimics the effects of a GIP receptor agonistin chronic application. There are several interpretations of how thissurprising effect may be caused. It has been shown previously intests of GIP receptor and GLP-1R KO mouse strains that the interrup-tion of incretin signalling shows surprisingly few effects on the ho-meostasis of blood glucose (Pamir et al., 2003; Scrocchi et al., 2000).Even double GIP receptor and GLP-1R KO mice exhibit normal bodyweight, and Plasma glucagon and the hypoglycemic response to exog-enous insulin are normal (Hansotia et al., 2004). It appears that com-pensatory mechanisms are activated which maintain blood glucosecontrol. Although serum total GLP-1 levels in GIP receptor KO micewere unaltered, insulin responses to GLP-1 were significantly greater
in the KO mice compared to wild type mice. Furthermore, GLP-1-induced cAMP production was also elevated twofold in the pancreasof the KO animals (Pamir et al., 2003). In addition, GLP-1R KO micehave elevated plasma GIP levels and increased ß-cell sensitivity toGIP, demonstrating that disruption of one component of the enteroin-sular axis may be compensated for by another signalling pathway(Pederson et al., 1998). These observations show that compensatorychanges in incretin signalling occur in order to maintain physiologicallevels of blood glucose (Preitner et al., 2004). It is feasible that thecontinuous block of GIP receptor by (Pro3)GIP induced an upregula-tion in GLP-1 signalling, which produced the observed agonistic ef-fects in LTP and stem cell proliferation. However, we have shown inthe past that GLP-1R KO mice (Abbas et al., 2009) as well as GIP re-ceptor KO mice (Faivre et al., 2011) show clear impairments in cogni-tion, LTP and stem cell proliferation, suggesting that such acompensation may be possible in blood glucose homeostasis but notin neuronal signalling in the brain.
Another interpretation is that (Pro3)GIP may be a mixed antago-nist/agonist. In acute studies, the compound shows clear antagonistic
Fig. 10. D-Ala2GIP increases the proliferation of neuronal progenitor cells in the subgranular zone of the DENTATE GYRUS. D-Ala2GIP at 25 nmol/kg i.p. for 3 weeks enhanced theproliferation of neuronal progenitor cells compared to both (Pro3)GIP and control group (One-way ANOVA; Pb0.0001). *=Pb0.05 saline vs. (Pro3)GIP group, ***=Pb0.01 salinevs. DAla2GIP group, $$$=Pb0.01 DAla2GIP vs. (Pro3)GIP group. Data represent mean±S.E.M. of 6 mice per group. Sample micrographs are shown with BrdU positive cells in the DG(arrow).
305E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
effects when it competes with the more effective native GIP. In chron-ic application, however, it may act as an agonist during episodeswhen there is no endogenous GIP available. Since native GIP has avery short half life of a few minutes in the blood stream (Hinke etal., 2002), and long-lasting protease resistant GIP analogues have ahalf life of several hours (Deacon, 2004), it is feasible that the weakagonistic effects of (Pro3)GIP outlast native GIP and act as a weak ag-onist during times when the endogenous GIP levels are too low to ac-tivate brain GIP receptors. Further research is required to identify theunderlying biochemical processes of the acute and chronic (Pro3)GIPeffects in the CNS in more detail. We also show that chronic applica-tion of the GIP receptor agonist D-Ala2GIP enhances LTP in the hippo-campus, similar to the enhancement found in our previous acutestudy (Gault and Holscher, 2008). The enhancement of neuronal pro-genitor proliferation in the subgranular zone of the Dentate gyrus byGIP has been demonstrated in the past (Nyberg et al., 2005). Here,we demonstrate that (Pro3)GIP as well as DAla2GIP enhances cell pro-liferation. This is also consistent with our previous study of GIP recep-tor KO mouse brains that show a decrease of neuronal progenitorproliferation (Faivre et al., 2011).
The fact that D-Ala2GIP enhances memory formation, LTP and pro-genitor cell proliferation in wild type mice without inducing adverseside effects is of great interest. Blood glucose levels were not affectedby the drug in non-diabetic animals (Figueiredo et al., 2010; Irwin etal., 2006), showing that the drug is safe to use and does not affectblood glucose levels in normoglycaemic conditions. Furthermore,the observation that DAla2GIP enhanced neuronal progenitor prolif-eration in the brain is also of great importance, as the increase ofsuch cells could be of use for repairing neuronal damage (Blurton-
Jones et al., 2009). Other incretin analogues such as exendin-4 (Li etal., 2009), (Val8)GLP-1 (Gengler et al., 2012) or liraglutide (McCleanet al., 2011) have shown clear improvements in mouse models of Alz-heimer's disease or Parkinson's disease. Long-acting analogues ofincretins of GIP or GLP-1 analogues may be of value in treating neuro-degenerative disorders (Holscher, 2010). However, further researchin mouse models of neurodegenerative diseases will be required totest the neuroprotective properties of GIP analogues.
Acknowledgement
The work was funded by a grant from the Alzheimer Research UK.The authors gratefully acknowledge the support by Prof. Peter Flattand Dr. Victor Gault.
References
Abbas, T., Faivre, E., Holscher, C., 2009. Impairment of synaptic plasticity and memoryformation in GLP-1 receptor KOmice: interaction between type 2 diabetes and Alz-heimer's disease. Behav. Brain Res. 205, 265–271.
Baggio, L.L., Drucker, D.J., 2007. Biology of incretins: GLP-1 and GIP. Gastroenterology132, 2131–2157.
Blurton-Jones, M., Kitazawa, M., Martinez-Coria, H., Castello, N.A., Muller, F.J., Loring,J.F., Yamasaki, T.R., Poon, W.W., Green, K.N., LaFerla, F.M., 2009. Neural stem cellsimprove cognition via BDNF in a transgenic model of Alzheimer disease. Proc.Natl. Acad. Sci. U. S. A. 106, 13594–13599.
Buhren, B.A., Gasis, M., Thorens, B., Muller, H.W., Bosse, F., 2009. Glucose-dependentinsulinotropic polypeptide (GIP) and its receptor (GIPR): cellular localization,lesion-affected expression, and impaired regenerative axonal growth. J. Neurosci.Res. 87, 1858–1870.
Deacon, C.F., 2004. Circulation and degradation of GIP and GLP-1. Horm. Metab. Res. 36,761–765.
306 E. Faivre et al. / European Journal of Pharmacology 674 (2012) 294–306
Dere, E., Huston, J.P., De Souza Silva, M.A., 2007. The pharmacology, neuroanatomy andneurogenetics of one-trial object recognition in rodents. Neurosci. Biobehav. Rev.31, 673–704.
Ding, K.H., Zhong, Q., Xie, D., Chen, H.X., Della-Fera,M.A., Bollag, R.J., Bollag,W.B., Gujral, R.,Kang, B., Sridhar, S., Baile, C., Curl, W., Isales, C.M., 2006. Effects of glucose-dependentinsulinotropic peptide on behavior. Peptides 27, 2750–2755.
Drucker, D.J., Nauck, M.A., 2006. The incretin system: glucagon-like peptide-1 receptoragonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. Lancet 368,1696–1705.
Ennaceur, A., 2009. One-trial object recognition in rats and mice: methodological andtheoretical issues. Behav. Brain Res. 215, 244–254.
Faivre, E., Gault, V.A., Thorens, B., Holscher, C., 2011. Glucose-dependent insulinotropicpolypeptide receptor knockout mice are impaired in learning, synaptic plasticity,and neurogenesis. J. Neurophysiol. 105, 1574–1580.
Figueiredo, C.P., Pamplona, F.A., Mazzuco, T.L., Aguiar Jr., A.S., Walz, R., Prediger, R.D.,2010. Role of the glucose-dependent insulinotropic polypeptide and its receptorin the central nervous system: therapeutic potential in neurological diseases.Behav. Pharmacol. 21, 394–408.
Gault, V.A., Holscher, C., 2008. Protease-resistant glucose-dependent insulinotropicpolypeptide agonists facilitate hippocampal LTP and reverse the impairment ofLTP induced by beta-amyloid. J. Neurophysiol. 99, 1590–1595.
Gault, V.A., O'Harte, F.P., Harriott, P., Mooney, M.H., Green, B.D., Flatt, P.R., 2003. Effectsof the novel (Pro3)GIP antagonist and exendin(9–39)amide on GIP- and GLP-1-induced cyclic AMP generation, insulin secretion and postprandial insulin releasein obese diabetic (ob/ob) mice: evidence that GIP is the major physiological incre-tin. Diabetologia 46, 222–230.
Gault, V.A., Irwin, N., Green, B.D., McCluskey, J.T., Greer, B., Bailey, C.J., Harriott, P.,O'Harte, F.P., Flatt, P.R., 2005. Chemical ablation of gastric inhibitory polypeptidereceptor action by daily (Pro3)GIP administration improves glucose toleranceand ameliorates insulin resistance and abnormalities of islet structure in obesity-related diabetes. Diabetes 54, 2436–2446.
Gault, V.A., Kerr, B.D., Irwin, N., Flatt, P.R., 2008. C-terminal mini-PEGylation of glucose-dependent insulinotropic polypeptide exhibits metabolic stability and improvedglucose homeostasis in dietary-induced diabetes. Biochem. Pharmacol. 75,2325–2333.
Gengler, S., McClean, P.L., McCurtin, R., Gault, V.A., Holscher, C., 2012. Val(8)GLP-1 res-cues synaptic plasticity and reduces dense core plaques in APP/PS1 mice. Neuro-biol. Aging. 33, 265–276.
Green, B.D., Gault, V.A., Flatt, P.R., Harriott, P., Greer, B., O'Harte, F.P., 2004. Comparativeeffects of GLP-1 and GIP on cAMP production, insulin secretion, and in vivo antidia-betic actions following substitution of Ala8/Ala2 with 2-aminobutyric acid. Arch.Biochem. Biophys. 428, 136–143.
Hansotia, T., Baggio, L.L., Delmeire, D., Hinke, S.A., Yamada, Y., Tsukiyama, K., Seino, Y.,Holst, J.J., Schuit, F., Drucker, D.J., 2004. Double incretin receptor knockout (DIRKO)mice reveal an essential role for the enteroinsular axis in transducing the glucore-gulatory actions of DPP-IV inhibitors. Diabetes 53, 1326–1335.
Hinke, S.A., Manhart, S., Pamir, N., Demuth, H., R., W.G., Pederson, R.A., McIntosh, C.H.,2001. Identification of a bioactive domain in the amino-terminus of glucose-dependent insulinotropic polypeptide (GIP). Biochim. Biophys. Acta 1547,143–155.
Holscher, C., 2010. Incretin analogues that have been developed to treat type 2 diabeteshold promise as a novel treatment strategy for Alzheimer's disease. Recent Pat. CNSDrug Discov. 5, 109–117.
Hunter, K., Holscher, C., 2011. Liraglutide, (dAla2)GIP and Ac-GIP Cross the Blood BrainBarrier. poster ARUK Conference, Leeds.
Irwin, N., McClean, P.L., O'Harte, F.P., Gault, V.A., Harriott, P., Flatt, P.R., 2007. Early ad-ministration of the glucose-dependent insulinotropic polypeptide receptor antag-onist (Pro3)GIP prevents the development of diabetes and related metabolicabnormalities associated with genetically inherited obesity in ob/ob mice. Diabeto-logia 50, 1532–1540.
Li, Y., Perry, T., Kindy, M.S., Harvey, B.K., Tweedie, D., Holloway, H.W., Powers, K., Shen,H., Egan, J.M., Sambamurti, K., Brossi, A., Lahiri, D.K., Mattson, M.P., Hoffer, B.J.,Wang, Y., Greig, N.H., 2009. GLP-1 receptor stimulation preserves primary corticaland dopaminergic neurons in cellular and rodent models of stroke and Parkinson-ism. Proc. Natl. Acad. Sci. U. S. A. 106, 1285–1290.
McClean, P., Parthsarathy, V., Faivre, E., Hölscher, C., 2011. The diabetes drug Liraglutideprevents degenerative processes in a mouse model of Alzheimer's disease. J. Neu-rosci. 31, 6587–6594.
Nyberg, J., Anderson, M.F., Meister, B., Alborn, A.M., Strom, A.K., Brederlau, A., Illerskog,A.C., Nilsson, O., Kieffer, T.J., Hietala, M.A., Ricksten, A., Eriksson, P.S., 2005. Glucose-dependent insulinotropic polypeptide is expressed in adult hippocampus and in-duces progenitor cell proliferation. J. Neurosci. 25, 1816–1825.
Nyberg, J., Jacobsson, C., Anderson, M.F., Eriksson, P.S., 2007. Immunohistochemical dis-tribution of glucose-dependent insulinotropic polypeptide in the adult rat brain. J.Neurosci. Res. 85, 2099–2119.
Preitner, F., Ibberson, M., Franklin, I., Binnert, C., Pende, M., Gjinovci, A., Hansotia, T.,Drucker, D.J., Wollheim, C., Burcelin, R., Thorens, B., 2004. Gluco-incretins controlinsulin secretion at multiple levels as revealed in mice lacking GLP-1 and GIP re-ceptors. J. Clin. Invest. 113, 635–645.
Scrocchi, L.A., Hill, M.E., Saleh, J., Perkins, B., Drucker, D.J., 2000. Elimination ofglucagon-like peptide 1R signaling does not modify weight gain and islet adapta-tion in mice with combined disruption of leptin and GLP-1 action. Diabetes 49,1552–1560.
Tian, J.Q., Wang, Y., Lin, N., Guo, Y.J., Sun, S.H., Zou, D.J., 2010. Active immunization withglucose-dependent insulinotropic polypeptide vaccine influences brain functionand behaviour in rats. Scand. J. Immunol. 72, 1–7.
Trumper, A., Trumper, K., Trusheim, H., Arnold, R., Goke, B., Horsch, D., 2001. Glucose-dependent insulinotropic polypeptide is a growth factor for beta (INS-1) cells bypleiotropic signaling. Mol. Endocrinol. 15, 1559–1570.
Usdin, T.B., Mezey, E., Button, D.C., Brownstein, M.J., Bonner, T.I., 1993. Gastric inhibito-ry polypeptide receptor, a member of the secretin-vasoactive intestinal peptide re-ceptor family, is widely distributed in peripheral organs and the brain.Endocrinology 133, 2861–2870.
