Leptin Gene Attenuates Neuronal PDF

ORIGINAL ARTICLE

Leptin gene therapy attenuates neuronal damages evokedby amyloid-b and rescues memory decits in APP/PS1 miceR Perez-Gonzalez1,2, MX Alvira-Botero3, O Robayo4, D Antequera1,2, M Garzon3, AM Martn-Moreno2,5,6, B Brera2,5, ML de Ceballos2,5

and E Carro1,2

There is growing evidence that leptin is able to ameliorate Alzheimers disease (AD)-like pathologies, including brain amyloid-b (Ab)burden. In order to improve the therapeutic potential for AD, we generated a lentivirus vector expressing leptin protein in aself-inactivating HIV-1 vector (HIV-leptin), and delivered this by intra-cerebroventricular administration to APP/PS1 transgenicmodel of AD. Three months after intra-cerebroventricular administration of HIV-leptin, brain Ab accumulation was reduced. Byelectron microscopy, we found that APP/PS1 mice exhibited decits in synaptic density, which were partially rescued by HIV-leptintreatment. Synaptic decits in APP/PS1 mice correlated with an enhancement of caspase-3 expression, and a reduction insynaptophysin levels in synaptosome preparations. Notably, HIV-leptin therapy reverted these dysfunctions. Moreover, leptinmodulated neurite outgrowth in primary neuronal cultures, and rescued them from Ab42-induced toxicity. All the above changessuggest that leptin may affect multiple aspects of the synaptic status, and correlate with behavioral improvements. Our datasuggest that leptin gene delivery has a therapeutic potential for Ab-targeted treatment of mouse model of AD.

Gene Therapy (2014) 21, 298308; doi:10.1038/gt.2013.85; published online 16 January 2014

Keywords: animal model; brain; electron microscopy; leptin; cell culture

INTRODUCTIONAlzheimers disease (AD) is the most common form of dementia inthe elderly.1 Pathologic hallmarks of AD include senile plaques(of which b-amyloid peptide (Ab) is a major component),intracellular neurobrillary tangles containing hyperphosphorylatedtau, synapse dysfunction, progressive memory loss and eventualdeath of brain neurons.2 Loss of connectivity, caused by neuronaldeath, and synapses is thought to underlie cognitive decline in AD.3

There is a growing consensus that Ab is toxic to synapses.46 Indeed,senile plaques are associated with a local loss of dendritic spines.79

Leptin, an adipocyte-derived hormone, exhibits a large range ofcentral and peripheral actions.10 Leptin receptors have beenidentied in neurons in many brain regions, including thehippocampus,11,12 providing evidence for leptins neuroprotectiveproperties on proliferation and survival signaling.13 Recent studieshave demonstrated the potential benecial effects of leptin inseveral transgenic animal AD models, namely modulation of theamount of Ab, and reduction of tau phosphorylation.1416 Inaddition, impaired behavior, memory and synaptic plasticity havebeen observed in leptin-decient rodents.17,18 Importantly, clinicalevidence shows alterations in the circulating levels of leptin in ADas patients with this disorder presents lower levels of leptin in theblood.19,20 In this study, we have used APP/PS1 mice displaying AD-like alterations2123 to investigate whether the cognitive andneuropathological disturbances associated with amyloidosis couldbe reversed by chronic treatment with leptin.

We have shown that overexpression of leptin by gene transferwith viral vectors resulted in a modulation in proliferation of

neuronal progenitors and neuroprotection.24 Recombinantlentiviral vectors are powerful tools for gene transfer to thecentral nervous system and hold great potential as a therapeuticgene therapy strategy for neurological disorders.25 These lentiviralvectors are versatile tools because of both their relatively largecloning capacity and their ability to transduce non-dividing cells.Viral vector gene delivery of leptin via stereotactic injection intothe central nervous system has proven to be a viable approach totreat transgenic mouse models of AD. Leptin treatment wasadministered to a 6-month-old APP/PS1 mice, and after 3 monthsof treatment, we found that leptin treatment reduced Abaccumulation, and alleviated synaptic alterations. In addition,we observed improvements in memory 3 months after vectordelivery. These results suggest a novel and improved approach fordelivery of leptin to the central nervous system for treatment of AD.

RESULTSLeptin rescues behavioral decitsFor optimal gene delivery approach, we generated lentiviralvectors that were able to efciently transfect most cell types.We inserted the cDNA of leptin into a lentiviral vector driven bythe cytomegalovirus promoter. We previously reported that theresulting vector was able to transfect neuronal cells in vivo, andproduced high levels of leptin expression.24

Then, we determined whether HIV-leptin-treated APP/PS1 miceshowed behavioral and cognitive recovery. Working memory wasevaluated by measuring the rate of spontaneous alternations in

1Neuroscience Group, Instituto de Investigacion Hospital 12 de Octubre (i 12), Madrid, Spain; 2Center for Networker Biomedical Research on Neurodegenerative Diseases(CIBERNED), Madrid, Spain; 3Department of Anatomy, Histology and Neuroscience, Universidad Autonoma de Madrid, Madrid, Spain; 4Universidad Pedagogica y Tecnologica deColombia, Boyaca, Colombia and 5Neurodegeneration Group, Cajal Institute-CSIC, Madrid, Spain. Correspondence: Dr E Carro, Neuroscience Group, Instituto de InvestigacionHospital 12 de Octubre (i 12), Av de Cordoba s/n, Madrid 28041, Spain.E-mail: [email protected] address: MD Anderson Cancer Center, Arturo Soria 270, 28033 Madrid, Spain.Received 14 October 2013; revised 5 December 2013; accepted 16 December 2013; published online 16 January 2014

Gene Therapy (2014) 21, 298308& 2014 Macmillan Publishers Limited All rights reserved 0969-7128/14

www.nature.com/gt

the T-maze. This particular model of spatial memory was selectedbased upon its sensitivity of detection of cognitive decits in ADmouse models.26,27 APP/PS1 mice showed a moderate decit inthis behavioral task, the baseline rate of spontaneous alternationsbeing: 73.74.9% in non-transgenic mice versus 59.35.3% inAPP/PS1 mice (P 0.068; Students t-test). HIV-leptin treatmenthad only a partial effect on the performance of APP/PS1 mice; thealternation rate following chronic treatment was 66.35.1%.However, in APP/PS1 mice, the latency (seconds) to perform thecorrect choice was signicantly increased compared with non-transgenic mice, whereas this time was reduced after HIV-leptintreatment (Po0.05; Figure 1a).

Short-term memory was evaluated using a paradigm of non-spatial visual recognition memory, subjecting the animals to anovel object recognition task, as previously described.28

Retention in male APP/PS1 transgenic mice was signicantlyimpaired relative to non-transgenic mice, whereas in HIV-leptin-treated APP/PS1 mice the impaired cognition was completelyrestored (Po0.05; Figure 1b). This result is similar to that of arecent study using different mouse model of AD, the TgCRND8transgenic mice, and a different treatment consisting of acontinuous delivery of leptin using a subcutaneous pump for 8weeks.16

Leptin reduces Ab plaque accumulationPreviously, we did not detect a signicant enhancement inendogenous leptin levels in plasma samples in APP/PS1 mice afterHIV-leptin treatment (data not shown), suggesting a veryrestricted and subtle target, preventing possible peripheral sideeffects. In addition, there was no signicant difference in glucoselevels or body weight observed in leptin- or void vector (vv)-treated APP/PS1 mice, as described previously.

Leptin has been shown to reduce brain Ab levels in 6-month-old CRND8 transgenic mice following 8 weeks of treatment.16

We demonstrated that Ab staining in the brain parenchyma of6-month-old APP/PS1 mice was signicantly reduced after 3months of treatment with HIV-leptin (Figure 2a). Stereologicalanalysis of amyloid immunostaining in APP/PS1 mice treated withleptin showed a 34% decrease in the cerebral cortex, and a 22%decrease in the hippocampus (Figure 2b) compared with emptyvector-treated APP/Ps1 mice. We also used thioavin-S staining todetect amyloid deposits, and conrmed the reduction in amyloiddeposits after HIV-leptin treatment (Figure 2c).

To obtain a direct quantitative measurement of the Ab levelsafter HIV-leptin treatment, we performed an ELISA analysis ofhuman Ab40 and Ab42. The study revealed that APP/PS1 mice

treated with HIV-leptin for 3 months had a signicant decrease inhuman Ab42 levels in the cerebral cortical and hippocampalregions (Table 1), whereas human Ab40 levels was only reduced incerebral cortex (Table 1).

We then investigated whether HIV-leptin treatment altered theprocessing of brain APP into Ab peptides. The amyloidogenicroute was evaluated by testing the expression and the enzymaticactivities of BACE1. We measured BACE1 expression in the corticaland hippocampal samples, using western blot (WB) assay. HIV-leptin-treated mice had reduced the expression of BACE1 incortical (Figure 2d) and hippocampal (Figure 2e) samples, withsimilar levels to those observed in non-transgenic mice.

Because one proposed mechanism of Ab clearance is microglialAb phagocytosis, we also tested the involvement of leptin in theregulation of activated microglia. Microglial cells were activatedin mice transfected by HIV-leptin as an increased areashowing tomato lectin positive cells in both cerebral cortex andhippocampus from HIV-leptin APP/PS1 mice was observed(Supplementary Figures a and b). This microglial activationcould be directly related to the decrease in the load of Ab, giventhe role of the microglia in the degradation of amyloid plaques.

Leptin modulates synaptic decitsBecause leptin has been proposed to induce synaptic plasticity,2931

and in order to assess whether early cognitive decline in APP/PS1mice and restoration after HIV-leptin treatment was associatedwith synaptic alterations, we measured synapse density in CA1pyramidal neurons of the dorsal hippocampus in non-transgenicand APP/PS1 mice treated with HIV-vv or with HIV-leptin. The areastudied in each of the 12 animals was 8820 mm2, and the total areaanalyzed in each experimental group was 35 280mm2. Overall,there was a predominance of asymmetric synapses in all threeexperimental groups (average percentage of asymmetric synapsesin (1) non-transgenic mice: 97.670.73%; (2) APP/PS1-HIV-vv:97.481.16%; (3) APP/PS1-HIV-leptin: 98.781.10%). Therefore,we did not discriminate between synapse type (symmetrical orasymmetrical) in the estimation of the number of synapses perunit area. Also, we observed that most of the synaptic contactsoccurred between axon terminals and dendritic spines, and lessamong axon terminals and dendritic shafts. A gross analysis of theultrastructure showed that, compared with APP/PS1 mice, non-transgenic controls had an abundant number of synapsesbetween axon terminals and dendritic spines. In addition, themorphology of the subcellular elements was well conserved incontrol animals, as expected, whereas the overall organization ofthe subcellular elements was somewhat less conserved in APP/PS1

Figure 1. HIV-leptin treatment of APP/PS1 mice results in the attenuation of memory impairment. (a) Increased latencies in the spontaneousalternation of APP/PS1 mice in the T-maze were partially normalized after HIV-leptin treatment (3 months). (b) In the novel-object recognitiontask, recognition memory was expressed as exploratory preference in the retention test. The recognition index, representing exploratorypreference, was reduced in APP/PS1 mice compared with non-transgenic mice (control), and recovered after HIV-leptin administration. Dataare expressed as means.e.m., *Po0.05, n 8 for non-transgenic control group, n 8 for APP/PS1 control group, n 12 for HIV-leptin-treatedAPP/PS1 mice. vv, void vector.

Leptin gene therapy in APP/PS1 miceR Perez-Gonzalez et al

299

& 2014 Macmillan Publishers Limited Gene Therapy (2014) 298 308

mice. The most prominent characteristics being the presence ofexcessively dilated and vacuolated dendrites, a decreased numberof synaptic proles, enlarged mitochondria, distended andirregular extracellular spaces, and a less compact and structuredaspect of the tissue (Figure 3, upper and middle panels). Finally,the ultrastructural morphology of HIV-leptin-treated APP/PS1 miceappeared to uctuate between control and APP/PS1 mice.Extracellular spaces appear less distended and subcellular prolesappear more compact and less vacuolated. In addition, the

number of synaptic proles also seems to be greater comparedwith APP/PS1 mice (Figure 3, lower panel).

Given the absence of data about synaptic density in APP/PS1transgenic mice, we rst decided to compare the synaptic densitybetween transgenic and non-transgenic mice. Analysis of thesynaptic proles showed a statistically signicant reduction inthe percentage of synapses in dendritic spines compared with thetotal percentage of synapses in APP/PS1 mice (Po0.05) (Table 2).Then, we investigated whether 3 months of gene therapy with

Figure 2. HIV-leptin APP/PS1 mice exhibit reduced levels of Ab deposits compared with HIV-vv APP/PS1 mice. (a) Representative images ofamyloid staining in cerebral cortex (upper panels), and hippocampus (bottom panel) in APP/PS1 mice treated with HIV-vv (vv) or HIV-leptin(leptin). Scale bar, 20 mm. (b) Cerebral cortex and hippocampus Ab burden is decreased in HIV-leptin-treated APP/PS1 mice. Brain Ab burdenrepresents the percentage area covered by Ab immunoreactivity. (c) Representative staining of amyloid deposits (Thioflavin-S staining)showing lower number of aggregates in the brain of APP/PS1 mice after HIV-leptin treatment. Scale bar, 20 mm. (d and e) BACE1 expression isreduced in cerebral cortical (d) and hippocampal (e) lysates of HIV-leptin-treated APP/PS1 mice. b-actin was used as control for proteinloading. Representative WBs and densitometry histograms are shown. Data are expressed as means.e.m., *Po0.05, **Po0.01, n 8 for non-transgenic control group, n 8 for APP/PS1 control group, n 8 for HIV-leptin-treated APP/PS1 mice. c, control; vv, void vector; Lep, leptinvector; Cx, cerebral cortex; Hip, hippocampus.


300

Gene Therapy (2014) 298 308 & 2014 Macmillan Publishers Limited

leptin could alter this observation. However, we did not found anyenhancement in the number of synapses in dendritic spines inHIV-leptin-treated APP/PS1 mice versus empty vector-treatedcontrols (83.941.30 in HIV-vv APP/PS1 mice versus 86.381.61in HIV-leptin APP/PS1 mice, P 0.077; KruskalWallis test).

Synapse loss has been associated with Ab-induced neuro-toxicity in synaptosomes.32 In addition, a recent study hasdemonstrated that caspase-3 activity is increased in Tg2576hippocampal synapses, and correlates with spine degenerationand decit in hippocampal-dependent memory.33 Thus, weexamined caspase-3 expression in cortical and hippocampalsynaptosome preparations from non-transgenic and APP/PS1mice with and without leptin treatment. We observed a markedincrease in procaspase-3 expression in synaptosomes preparedfrom cerebral cortex of HIV-vv mice (Po0.01), which reduced afterHIV-leptin treatment (Po0.05; Figures 4a and b). In these samples,the expression of active forms of caspase-3 (17 and 19 kDa) wasundetectable by WB (Figure 4a). In the hippocampal synaptosomesamples, the enhancement in procaspase-3 expression was notstatistically signicant (P 0.069) in APP/PS1 mice (data notshown). However, active forms of caspase-3 were markedlyincreased in these transgenic mice, showing a tendency towardreduction after HIV-leptin treatment (Figures 4c and d).

These ndings were also associated with an increased expres-sion of procaspase-3 in the total tissue homogenates fromcerebral cortex of APP/PS1 mice, whereas these effects werereverted after HIV-leptin treatment (Po0.05; Figures 4e and f).In the hippocampus, no signicant differences were observedbetween experimental groups (data not shown). Notably, wefound that procaspase-3 was more abundant in the synaptosomalfraction when compared with the total tissue homogenate. Thesendings suggest that caspase-3 is preferentially localized andselectively activated in the synaptic compartment, reinforcing therole of caspase-3 in synaptic dysfunction in AD.33 In view of allthese ndings, we measured whether leptin was able to modulateAb-induced cell death. Using the Cell Death Detection ELISAPLUS

kit, we found an increased tendency in cell death in the cerebralcortex from APP/PS1 mice (P 0.06; Figure 4g), and this effect wasfully reverted after HIV-leptin treatment (Po0.05; Figure 4g).

To provide additional evidence of the synaptic molecularmodications occurring in synaptosomes, we quantied the levelsof synaptophysin, as pre-synaptic marker in the synaptosomepreparations. We found that levels of synaptophysin weresignicantly reduced in synaptosome fractions prepared fromcerebral cortex (Figures 5a and b), and hippocampus of APP/PS1mice as compared with non-transgenic mice (Figures 5c and d).HIV-leptin treatment in APP/PS1 mice was able to increase theselevels of synaptophysin in synaptosomal preparations from bothbrain areas (Figures 5b and d). To determine whether leptin affectssynaptophysin expression in neuronal cell cultures, corticalneurons were grown with the different treatments included inthe media. Leptin treatment for 48 h produced a marked increasein synaptophysin expression, whereas Ab42 tended to reduce this

effect (Figures 5e and f). When both molecules were added to themedia, leptin was able to block the negative effects induced byAb42 on synaptophysin expression (Figures 5e and f). These in vitrondings are consistent with the results from in vivo studiesshowing the effects of leptin on protein expression in synapto-somes of APP/PS1 mice, and support the direct involvement ofAb42 in the process.

Previously, we conrmed that lentiviral administration did notgenerate any non-specic dysfunction in neurons. We comparedthe expression of caspase-3, and synaptophysin in intact (sham)APP/PS1 mice with that of HIV-vv-treated APP/PS1 mice.No differences were observed regarding the expression of theseproteins in the cerebral cortex or hippocampus of APP/PS1 mice(data not shown).

Leptin modulates neurite length in neuronal culturesPrevious studies have shown neurite degeneration caused by Abin cultured neurons as one of the initial events in the process ofneuronal death,34 which ultimately leads to reduced neuritelength.35 As leptin rapidly enhances the motility and density ofdendritic lopodia, neurite length and subsequently increases thedensity of hippocampal synapses,29 we investigated the role ofleptin in neurite degeneration evoked by Ab in cortical andhippocampal neuronal cell cultures. Thus, we studied the effect ofboth stimuli, leptin and Ab42, on neurite length throughimmunocytochemistry of calbindin-labeled cells, which allows tovisualize mature neurons and the different types of neurites.Treatment of cortical neuronal cell cultures with both stimuli for18 h produced modest changes. Only leptin induced an increasein the length of secondary neurites, whereas Ab42 did not producesignicant alterations (Figures 6a and c). After 18 h, resultsobtained from hippocampal neurons were similar (Figures 6eand g). However, when neurite length was assessed 36 h aftertreatment addition, we observed that Ab42 had produced amarked reduction (43%) in the total length of neurites fromcortical neurons (control: 140.130 mm; Ab42: 80.92.1 mm).In addition, the reduction in neurite length induced by Ab42 wascompletely antagonized by the administration of leptin (Figures6b and d). These alterations were mainly caused by changes inprimary and secondary neurites. We also investigated theresponse of hippocampal neurons to Ab42, and found that theeffects were less marked (Figures 6f and h). However, treatment ofneuronal cell cultures with Ab42 also reduced (23.6%) the totallength of neurites from hippocampal neurons (control:199.212.3 mm; Ab42: 152.22.2 mm; Figures 6f and h).In summary, Ab42 had a detrimental impact on both cortical andhippocampal neurons in culture reected by a reduction in neuritelength. This effect was prevented by leptin treatment.

DISCUSSIONDevelopment of new approaches for the delivery of neuroactivepeptides is of fundamental importance for advancing thetherapeutics of AD. Here, we showed that leptin produced froma lentiviral vector regulated levels of Ab, and synaptic alterationsin the brains of APP/PS1 mice. Our current study uses a newapproach to examine this issue based on gene therapy as a viableapproach of leptin delivery to the brain using a lentiviral vector.24

Lentiviral vectors are valuable tools for neurobiology researchbecause of their ability to transduce non-dividing cells, such asneurons.36 This in vivo technique ensured a stable and effectiveoverexpression of leptin, providing a more physiologic leptindelivery than is possible with direct infusion. Lentiviral vectorsexhibit an inherent nontoxic capacity to migrate long distances,and a remarkable ability to modulate host brain cells.19,24,3739

In addition, leptin protein transduced in neurons was expressed inamounts sufcient to exert leptin-specic biological effects,

Table 1. Leptin regulates brain Ab40 and Ab42 levels in APP/PS1 mice

HIV-vv HIV-leptin

Cerebral cortexAb40 (pgml

1) 32.856.82 9.325.34a

Ab42 (pgml 1) 35.236.50 6.312.10a

HippocampusAb40 (pgml

1) 16.044.45 9.492.59Ab42 (pgml

1) 14.631.71 7.531.50a

Data are expressed as means.e.m. aPo0.05, versus HIV-vv mice; n 8 forHIV-vv-treated APP/PS1 mice, n 8 for HIV-leptin-treated APP/PS1 mice.


301


but insufcient for detection of alteration levels brain parenchymaor peripheral circulation, in accordance with previous studiesusing leptin gene therapy.40,41 These results demonstrate theutility of locally elevating the expression of leptin as a potentialtherapeutic approach for AD, although essential features areneeded to ensure its translation to the clinic in the near future.

Lentiviral vector-mediated gene therapy offers many advan-tages compared with other viral systems and, as a therapeuticapproach, is very attractive particularly for neurodegenerativedisorders, including AD, as it can be delivered to the affected area,and one treatment offers long-term expression. After delivery intothe nervous system they induce no signicant immune responses,

there are no unwanted side effects of the vectors per se to date,and manufacturing and safety testing for clinical applications arewell advanced. However, signicant issues remain in some areasof neural gene therapy including dening the optimum ther-apeutic gene(s), increasing the specicity of delivery, regulatingexpression of potentially toxic genes, and designing clinicallyrelevant strategies.

Our ndings indicated that 3 months gene therapy-mediatedleptin administration reduced the accumulation of Ab deposits inAPP/PS1 mice, coinciding with previous studies using anothermouse model of amyloidosis and leptin treatment with asubcutaneous Alzet miniosmotic pump.16 HIV-leptin treatment in

Figure 3. HIV-leptin treatment prevents loss of synapses in the CA1 layer of hippocampus of APP/PS1 mice. Representative photomicrographsby electron microscopy from non-transgenic mice (upper panel), HIV-vv-treated APP/PS1 mice (middle panel) and HIV-leptin-treated APP/PS1mice (bottom panel). (a) A dendritic shaft (D) makes asymetric synapses (red arrows) with different axonic terminals (T) and, occasionally,symetric synapses (yellow arrows). (b) Longitudinally and transversally cut dendritic shafts (D) and asymetric synaptic contacs (red arrows)between dendritic spines (SP) and axonic terminals (T). (c) A dendrite makes different synaptic contacts (red arrows). (d) The curved red arrowindicates how a dendritic spine develops from a dendritic shaft. (e) Asymetric synaptic contacts between spines (SP) and terminals (T). Notedifferent dendritic shafts, one of them is longitudinally cut where a dendritic spine forms (dashed line). (f ) Asymetric synapse between alongitudinally cut dendrite (D) and a terminal (T). In addition, a spine synapse (SP). Scale bar, 0.5 mm, n 4 for non-transgenic control group,n 4 for APP/PS1 control group, n 4 for HIV-leptin-treated APP/PS1 mice.


302


APP/PS1 mice resulted in the improvement in expression ofsynaptophysin in the cerebral cortex and hippocampus, andpartial restoration of learning and memory impairment, incomparison with APP/PS1 mice treated with HIV-vv. Diverseevidences indicate that Ab oligomers detected within plaquesinduce synapse loss,42 and memory decits in the Tg2576 mousemodel,43,44 and that leptin reduces brain amyloid burden.16

More recently, the relationship between leptin and synapticdisruption has been documented as it was shown to prevent theAb-induced deterioration of hippocampal long-term potentiation.45

This study showed that leptin prevents Ab-induced AMPA-receptorremoval from the synapses, thus preventing synapse molecularchanges. In our present study, we show that apart from molecularchanges, leptin has neuroprotective effects on the synapsestructure, thus providing new insight into the coordinatingmechanisms between leptin and Ab-induced neurodegenerationin AD. The characteristic sequence of events associated withAb-induced neurodegeneration starts with the early loss of synapticcontacts, followed by neurite damage, neuronal degeneration, and,nally, neuronal death.46,47 We found that synapse density wassignicantly decreased in the CA1 layer of the hippocampus inAPP/PS1 mice, and HIV-leptin treatment partially attenuated thisimpairment. In support of these results, the APP/PS1 transgenicmouse model showed a signicant decrease in synaptophysin insynaptosomes compared with wild-type mice. Our experimentsconrmed that Ab neurotoxicity is involved in this synaptophysindownregulation, as previously reported,35 and a reversal of thisdecrease was observed in the HIV-leptin-treated APP/PS1 mice.

Table 2. Number of synaptic profiles in spines and shafts in APP/PS1mice compared with control mice

Control APP/PS1-HIV-vv

Total of synapses per mm3 1.52e0.12 1.230.13Spine synapses per mm3 1.340.12 1.040.12Shaft synapses per mm3 0.170.01 0.200.02% spine synapses 88.400.66 83.941.30a

Data are expressed as means.e.m. aPo0.05 (MannWhitney U-test), n 4for non-transgenic control group, n 4 for APP/PS1 control group, n 4 forHIV-leptin-treated APP/PS1 mice.

Figure 4. HIV-leptin treatment prevents caspase 3 activation in synaptosome preparations from APP/PS1 mice. (a) Representative WB showingthe expression of procaspase-3 in cortical synaptosome preparations. Note that cleaved caspase-3 was undetectable. b-actin was used as controlfor protein loading. (b) Densitometric quantification of changes in procaspase-3 expression. (c) In hippocampal synaptosomes cleaved caspase-3was increased in APP/PS1 mice compared with wild-type mice, and was reduced after HIV-leptin treatment. b-actin was used as control forprotein loading. (d) Densitometric quantification of changes in cleaved caspase-3 levels in hippocampal synaptosome preparations.(e) Representative WB, and densitometric quantification (f ) of procaspase-3 expression in lysates from cerebral cortex of wild-type and APP/PS1mice. b-actin was used as control for protein loading. (g) Cell death detected by Cell Death Detection ELISAPLUS kit, is reduced in the cerebralcortex of HIV-leptin-treated APP/PS1 mice. Data are expressed as means.e.m., *Po0.05, **Po0.01, n 8 for non-transgenic control group, n 8for APP/PS1 control group, n 12 for HIV-leptin-treated APP/PS1 mice. c, control; vv, void vector; Lep, leptin vector.


303


We speculate that the neuroprotective effect of leptin againstsynapse loss is due to a decrease in the toxicity of Ab accumulation,probably by decreasing caspase-3 expression, whose upregulationin hippocampal CA1 dendritic spines led to a reduction in spinedensity.33

The inhibitory effect of leptin on Ab accumulation is likelyattributable to its inhibition of APP processing through theamyloidogenic pathway, modulating BACE expression, a majorb-secretase involved in APP processing. In agreement withpublished data,48 we found a modest difference in BACEexpression in APP/PS1 mice compared with non-transgenicmice. This result is not surprising, as there is a progressive age-dependent BACE expression that could be affecting themechanism of APP processing. Although BACE1 levels wereincreased in 1218-month-old APP/PS1 mice,19,49 no differenceswere observed in 6- month-old APP/PS1 mice compared with non-transgenic mice.49 However, after 3 months of treatment with HIV-leptin, we found that BACE expression was reduced in APP/PS1mice compared with control transgenic mice.

In the brain of APP/PS1 mice, a discrete neuronal loss isassociated with Ab plaques.50 Consistent with these ndings, weobserved a slight increase in cell death in our model of transgenicmice, which was reversed in HIV-leptin-treated APP/PS1 mice.The neuroprotective effect of leptin seen in APP/PS1 mice is in line

with previous reports showing protection against Ab-inducedinjury of primary cortical neuronal cultures.45,51,52 Normalcognitive function largely depends on the proper connectionbetween neurons, whereas treatment with Ab, reduces neuriteoutgrowth,35 leptin promotes neurite outgrowth inhypothalamic,53 and hippocampal neurons.29 Therefore, wehypothesized that leptin might induce changes in neuritemorphology, thereby contributing to the changes inhippocampal synaptic efcacy in AD. The nding that leptincompletely antagonized the reduction in neurite length evoked byAb conrms our hypothesis.

It has been demonstrated that leptin administration into theCA1 region of the hippocampus improves memory processing inmice performing T-maze foot-shock avoidance and step-downpassive avoidance tests.54 On the other hand, cognitive decits arethought to be associated with leptin-deciency in rodents,18 andare also prevalent in humans with obesity-related diseases, suchas type II diabetes.55 As the decrease in the number of synapses isless marked in leptin-treated APP/PS1 mice compared withAPP/PS1, it seems that leptin could be protecting from synapseloss, or inducing synaptogenesis, thereby preventing cognitiveimpairment. However, we cannot exclude the possibility thatleptin improves memory through a mechanism other than thedecrease of Ab accumulation, because adult neurogenesis has

Figure 5. Leptin modulates in vivo and in vitro synaptophysin expression. (a and c) Representative WBs of synaptophysin in cortical (a) andhippocampal (c) synaptosome preparations from wild-type and APP/PS1 mice treated with HIV-vv or HIV-leptin. b-actin was used as control forprotein loading. (b and d) Quantitative analysis of synaptic proteins in cortical (b) and hippocampal (d) synaptosome preparations. n 8 for non-transgenic control group, n 8 for APP/PS1 control group, n 12 for HIV-leptin-treated APP/PS1 mice. c, control; vv, void vector; Lep, leptinvector. (e) Representative WB showing an increase in synaptophysin protein levels in neuron cultures with or without exposure to Ab42(5mgml 1). Proteins were extracted 48h after treatments. b-actin was used as control for protein loading. (f ) Quantitative analysis of the effectsof leptin and Ab42 on synaptophysin levels. Data are expressed as means.e.m., *Po0.05, **Po0.001, n 3 independent experiments. LP, leptin.

Figure 6. Leptin prevents the reduction in neurite length evoked by Ab in neuronal cultures. (a and b) Representative immunofluorescenceimages of calbindin cortical neurons 18 h (a), and 36 h (b) after leptin (100 nM) or Ab42 (5mgml

1) treatments. (c and d) Total neurite length,and neurite length of each order were estimated in photomicrographs of calbindin cortical neurons 18 h (c), and 36 h (d) after treatments,and using a stereological procedure. (e and f ) Representative immunofluorescence images of calbindin hippocampal neurons 18 h (e), and36 h (f ) after leptin (100 nM) or Ab42 (5mgml

1) treatments. (g and h) Total neurite length, and neurite length of each order were estimated inphotomicrographs of calbindin hippocampal neurons 18 h (g), and 36 h (h) after treatments, and using a stereological procedure. Data areexpressed as means.e.m., *Po0.05, n 3 independent experiments.


304



305


been proposed to mediate hippocampal-dependent learning andthe therapeutic actions of antidepressants,56,57 and we haverecently reported that leptin stimulates proliferation of newlyformed hippocampal neurons in APP/PS1 mice.24

In summary, our experimental approach using a lentiviral viralvector allows the diffusion of the therapeutic factor, leptin, in thebrain of a mouse model of AD. Our study demonstrates that leptinnot only can ameliorate brain Ab accumulation, but also synaptic,cognitive, behavioral and Ab-induced neuronal network dysfunc-tions. Our ndings indicate that targeting the aberrant networkactivity with leptin or related drugs might be of therapeuticbenet in the prevention or treatment of AD.

MATERIALS AND METHODSLentiviral vectorsLentiviral vectors were produced using a four-plasmid transfection system,as described previously.24 Briey, a cDNA coding for mouse leptin wassubcloned in the XbaI-BamHI site of the pRRLsin18.PPT.CMV.eGFP.Wpretransfer vector, with a small segment of the viral hemagglutinin coatprotein as an epitope tag (HA). Empty vector (no insert) was used as acontrol (void vector, HIV-vv) in all experiments. Packaging, purication andtitter determination of the lentivirus were performed as describedpreviously.37 The packaging construct and the vesicular stomatitis virus Gprotein envelope included the pCMVDR-8.92, pRSV-Rev and pMD.Gplasmids, respectively. The transfer vector (13 mg), the envelope (3.75mg)and the packaging plasmids (3.5mg) were co-transfected with calciumphosphate in 293 T cells (5 106 cells/dish) cultured in Dulbeccos modiedEagles medium (DMEM, Gibco, Carlsbad, CA, USA) with 10% FCS, 1%glutamine and 1% penicillin/streptomycin. Medium was changed 2hprior to transfection, and replaced after 24h. Conditioned medium wascollected 24h later, cleared (1000 rpm/5min) and concentratedE100-fold(19 000 rpm/1.5 h). The pellet was resuspended in phosphate-bufferedsaline with 1% bovine serum albumin, and the virus stored at 80 1C. Viraltitter was determined by HIV-1 p24 ELISA (PerkinElmer, Wellesley, MA, USA).

AnimalsWe used the double-transgenic APP/PS1 mice, B6.Cg-Tg (APPSwe,PSEN1dE9)/J mouse line (Jackson Laboratory, Bar Harbor, ME, USA: stockno. 005864), which expresses human APP (Swedish mutation) andpresenilin 1 with a deletion in exon 9 (APP/PS1). A total of 20 six-month-old double transgenic APP/PS1 male, and 8 age-matched non-transgenicmale littermates were used as controls. Vector suspensions (2ml permouse) were stereotaxically injected in each lateral ventricle (braincoordinates are expressed as mm from bregma: 0.6 posterior, 1.1 lateraland 2 ventral) with a 10ml syringe at a rate of 1 ml min 1. Animals wereweighed at the beginning and at the end of experimental period(3 months). After 3 months of treatment, when mice were 9 months ofage, animals were deeply anesthetized and transcardially perfused eitherwith saline for biochemical analysis, or 4% paraformaldehyde in 0.1Mphosphate buffer (PB), pH 7.4 for immunohistochemical analysis.All animals were handled and cared for according to the Council Directive2010/63/UE of 22 September 2010.

Behavioral testingAfter adaptation to human handling, behavioral tests were conducted inAPP/PS1 and non-transgenic mice over 11-day period, as previouslydescribed.28 On each study day, spontaneous alternation, tested with aT-maze, a behavioral test of spatial memory,27 was the rst parameterevaluated, followed by the object recognition test (days 10 and 11), a testof non-spatial visual recognition memory.58 In the T-maze, the number ofalternations and errors (entries to previously visited arms), and the time tocomplete each session (latency) were recorded. In the object recognitiontest, the recognition index, was dened as the ratio of the time spentexploring the novel object over the time spent exploring both objects inpercentage, and was used to measure non-spatial memory.

Primary neuronal culturesPrimary neuronal cultures from the cerebral cortex and hippocampus,obtained from Wistar rat embryos on prenatal day 17 (E17), wereperformed as previously described.12 Briey, pregnant female patients

were anesthetized with isourane (Forane, Baxter, Deereld, IL, USA), andsubjected to cesarean section in order to remove fetuses and their brains.Cerebral cortex and hippocampus were dissected, incubated for 5min inNeurobasal medium (Gibco) at 37 1C and mechanically dissociated during5min and nally centrifuged at 210 g for 5min at 21 1C. The cells weresuspended in Neurobasal medium (Gibco) supplemented with B27 (Gibco),2mM glutamine, 100U/ml penicillin, 100mgml 1 streptomycin and0.25mgml 1 amphotericin B, and plated at a density of 2.5 105 cellsper square centimeter onto poly-Lysine (1mgml 1; Sigma, St Louis, MO,USA)-coated multi-well plates. The cultures were maintained at 37 1C in ahumidied atmosphere containing a 5% CO2 and cultivated for 7 daysprior to experimentation. Then, cultures were incubated in fresh mediumwith or without 100 nM leptin, alone or in combination with 5 mgml 1

oligomeric Ab42. Ab42 was previously dissolved in acetic acid 0.1M, andthen was dissolved in sterile distilled water as reported previously.19

Determination of neurite outgrowthCortical and hippocampal neuronal cultures were seeded at 5104 cells cm 2 on poly (lysine)-coated 24-well plates. Eighteen and 36 hafter treatment (100 nM leptin, and 5 mgml 1 oligomeric Ab42), cells werexed with 2% paraformaldehyde, and labeled with an anti-calbindinantibody. Images were obtained using a Zeiss LSM 510 Meta scanning laserconfocal microscope with a 40 objective, converted to 8-bit grayscaleimages for analysis. The term neurite refers to any projection from the cellbody of a neuron, which can either be an axon or a dendrite. Processeswith lengths equivalent to one or more diameters of a cell body werecounted as neurites. Morphometric analysis was performed with ImageJ1.43 u software (National Institute of Health, USA) with the NeuronJ plug-in.Morphological analysis of neurite length was performed as describedpreviously,12 based on a method adapted from a study by Arendt et al.59

A centrifugal ordering system was used to establish branch or neuriteorder in a neuron, and branch order was taken into consideration whencalculating total neurite length, as well as the length of neurite of eachorder. The average neurite length was calculated and subsequentlyanalyzed using a two-way ANOVA. At least 30 neurons were examined foreach group.

ImmunohistochemistryFixed brains were cut on a vibratome (Leica Microsystems, Barcelona,Spain) at 50 mm, tissue sections were collected in cold PB 0.1M, andincubated overnight with primary antibodies at 4 1C. To detect Ab deposits,brain sections from APP/PS1 mice were pre-incubated with 88% formicacid and immunostained as previously described.60 The following primaryantibodies were used: mouse anti-Ab (1:500, MBL, Nagoya, Japan), rabbitanti-Ab (1:500, Millipore, Billerica, MA, USA). And the secondary antibodieswere: biotinylated anti-mouse IgG (1: 400, Vector Laboratories, Burlingame,CA, USA), and Texas Red goat anti-rabbit IgG antibody (1:1000, JacksonImmunoresearch, West Grove, PA, USA). Immunostaining of microglial cellswas performed with rodhamine-labeled tomato-lectin (Vector Laboratories).Images were captured using a Zeiss LSM 510 Meta scanning laser confocalmicroscope (Carl Zeiss Microimaging, GmbH, Gottingen, Germany). Ab areawas analyzed stereologically as described previously,60 and plaque burdenwas expressed as percentage of brain area stained with Ab. Pictures weretaken using light microscopy (Zeiss microscope; Carl Zeiss Microimaging,GmbH) at a magnication of 10. Three or four areas were sampled withineach section and semiquantitative analysis of labeled signal was performedusing NIH Image J software. Amyloid deposits in parenchyma were alsoevaluated using Thioavin S (ThS) staining.

Synaptosome preparationHippocampal tissue was homogenized in homogenization buffer contain-ing 320mM sucrose, 4mM HEPES (pH 7.4), 1mM EGTA, 1mM PMSF andprotease inhibitor cocktail (Roche, Madrid, Spain) in a glass Dounce tissuegrinder, as previously described.33 The homogenate was centrifuged at1000g for 10min and the resulting supernatant was centrifuged at12 000 g for 15min. The pellet was resuspended in homogenization bufferand centrifuged at 13 000 g for 15min to obtain the nal pellet containingthe synaptosome-enriched fraction.

Biochemical measurementsFor WB analysis, samples were homogenized in lysis buffer containing150mM NaCl, 20mM Tris-HCl (pH 7.4), 1% Nonidet P-40, 1mM PMSF and


306


protease inhibitor cocktail (Roche), and then 2% SDS-containing buffers,electrophoresed and blotted. Antibodies used included: rabbit anti-BACE1(1:1000; Sigma), rabbit anti-caspase-3 (1:1000, Cell signaling, Danvers, MA,USA) and mouse anti-synaptophysin (1:1000, Millipore). Mouse leptin levelsin plasma samples from APP/PS1 and non-transgenic mice were determinedusing the Quantikine Mouse Leptin Immunoassay (R&D Systems; Minnea-polis, MN, USA). Blood glucose was measured after 12h fasting using aglucometer (Accu-Chek, Roche). Levels of human soluble Ab40 and Ab42 incerebral cortex and hippocampus fractions from APP/PS1 mice weredetermined with human-specic ELISA kits (Biosource, Invitrogen, Camarillo,CA, USA), according to the manufacturers instructions. The nal Ab40 andAb42 values were determined following normalization to total protein levels.For cell death quantication, DNA fragmentation in cortical samples of APP/PS1 mice undergoing apoptosis was detected using a Cell Death DetectionELISAPLUS kit (Roche Diagnostics), according to the manufacturers protocol.

Electron microscopyMice were intracardially perfused with saline solution followed by 2%paraformaldehyde and 3.8% acrolein in 0.1M PB, pH 7.4. The brains wereremoved and post-xed overnight in 2% paraformaldehyde in 0.1M PB.Coronal 500mm thick sections containing the hippocampus were cut on aSeries 3000 vibratome (Technical Products International, St Louis, MO, USA).With the help of a magnifying glass, 1.01.5mm3 trapezoid shaped blocks oftissue of the hippocampal CA1 area were cut with a scalpel, for a total ofthree blocks of tissue per mouse. The 36 tissue blocks were placed inindividual, coded and blind to condition glass vials, post-xed in 1% osmiumtetroxide in 0.1M PB for 2 h, dehydrated in a series of cold graded ethanolsand propylene oxide, incubated for 24h in a 1:1 mixture of propylene oxideand epoxy resin (Embed 812; Electron Microscopy Sciences, Fort Washington,PA, USA) and then in 100% epoxy resin for 1215 h. Each block was thentransferred to at-ended embedding capsules containing epoxy resin. Theposition of each one of the three tissue blocks belonging to each mousewas adjusted using a wooden stick in order to ensure that each had adifferent orientation relative to the other. The tissue was polymerized at60 1C for at least 48 h, after which blocks were trimmed, ultrathin-sectionedon an ultramicrotome and collected on 400 mesh copper grids andanalyzed under a Jeol JEM 1010 electron microscope at 80 kV coupled to adigital camera (GATAN BioScan Gatan Inc., Munchen, Germany).

Thirty micrographs at 25 K magnication were taken of each of the 36copper grids in a diagonal fashion by rows from corner to corner of themesh. To select the micrographs that were going to be analyzed, weconducted a random sampling, blind as to condition, in which one out ofevery three micrographs from every grid was selected. The total numberof synapses, the proles where these synapses occurred (dendritic shaftsor spines) and the relationship between the number of shaft synapses andspine synapses were counted using ImageJ software, version 1.46 using anunbiased counting frame area of 20 um2. The classication of cellularelements was based on the descriptions made by Peters et al.61

Asymmetric synapses were identied by the presence of thickpostsynaptic densities, whereas symmetric synapses were identied bythe presence of thin pre- and postsynaptic specializations. In addition,synapses were classied as shaft synapses, that is, those occurring onthe surface of a dendrite, or spine synapses, that is, those occurring at thebulbous end of a relatively short (o2 mm) protrusion separated from thedendrite by a neck or narrower region.

Statistical analysisResults are expressed as meanstandard error of the mean (s.e.m.).Students t-test was used when comparing two groups, and a two-wayANOVA followed by post hoc Tukey test was used for multiplecomparisons. Non-parametric tests were used when variances were nothomogeneous. All calculations were made using SPSS v15.0 software.Statistical signicance was set at Po0.05.

CONFLICT OF INTERESTThe authors declare no conict of interest.

ACKNOWLEDGEMENTSThis work was supported by grants from Instituto de Salud Carlos III (PI06/0155,FIS2009/01636), Fundacion Investigacion Medica Mutua Madrilena (2008/93, 2010/

0004) and CIBERNED (BESAD-P.2010). We thank Agnieszka Krzyzanowska, PhD, for thecareful revision of this manuscript.

REFERENCES1 Hardy J, Selkoe DJ. The amyloid hypothesis of Alzheimers disease: progress and

problems on the road to therapeutics. Science 2002; 297: 353356.2 Selkoe DJ. Alzheimers disease: genes, proteins, and therapy. Physiol Rev 2001; 81:

741766.3 Coleman PD, Yao PJ. Synaptic slaughter in Alzheimers disease. Neurobiol Aging

2003; 24: 10231027.4 Shankar GM, Li S, Mehta TH, Garcia-Munoz A, Shepardson NE, Smith I et al.

Amyloid-beta protein dimers isolated directly from Alzheimers brains impairsynaptic plasticity and memory. Nat Med 2008; 14: 837842.

5 Hsieh H, Boehm J, Sato C, Iwatsubo T, Tomita T, Sisodia S et al. AMPAR removalunderlies Abeta-induced synaptic depression and dendritic spine loss. Neuron2006; 52: 831843.

6 Snyder EM, Nong Y, Almeida CG, Paul S, Moran T, Choi EY et al. Regulationof NMDA receptor trafcking by amyloid-beta. Nat Neurosci 2005; 8: 10511058.

7 Spires-Jones TL, Meyer-Luehmann M, Osetek JD, Jones PB, Stern EA, Bacskai BJ et al.Impaired spine stability underlies plaque-related spine loss in an Alzheimersdisease mouse model. Am J Pathol 2007; 171: 13041311.

8 Spires TL, Meyer-Luehmann M, Stern EA, McLean PJ, Skoch J, Nguyen PT et al.Dendritic spine abnormalities in amyloid precursor protein transgenic micedemonstrated by gene transfer and intravital multiphoton microscopy. J Neurosci2005; 25: 72787287.

9 Tsai J, Grutzendler J, Duff K, Gan WB. Fibrillar amyloid deposition leads to localsynaptic abnormalities and breakage of neuronal branches. Nat Neurosci 2004; 7:11811183.

10 Carro EM. Therapeutic approaches of leptin in Alzheimers disease. Recent Pat CNSDrug Discov 2009; 4: 200208.

11 Friedman JM. Leptin, leptin receptors, and the control of body weight. Nutr Rev1998; 56: s38s46, discussion s54-75.

12 Alvira-Botero X, Perez-Gonzalez R, Spuch C, Vargas T, Antequera D, Garzon M et al.Megalin interacts with APP and the intracellular adapter protein FE65 in neurons.Mol Cell Neurosci 2010; 45: 306315.

13 Weng Z, Signore AP, Gao Y, Wang S, Zhang F, Hastings T et al. Leptin protectsagainst 6-hydroxydopamine-induced dopaminergic cell death via mitogen-activated protein kinase signaling. J Biol Chem 2007; 282: 3447934491.

14 Marwarha G, Dasari B, Prasanthi JR, Schommer J, Ghribi O et al. Leptin reduces theaccumulation of Abeta and phosphorylated tau induced by 27-hydro-xycholesterol in rabbit organotypic slices. J Alzheimers Dis 2010; 19: 10071019.

15 Fewlass DC, Noboa K, Pi-Sunyer FX, Johnston JM, Yan SD, Tezapsidis N et al.Obesity-related leptin regulates Alzheimers Abeta. FASEB J 2004; 18: 18701878.

16 Greco SJ, Bryan KJ, Sarkar S, Zhu X, Smith MA, Ashford JW et al. Leptin reducespathology and improves memory in a transgenic mouse model of Alzheimersdisease. J Alzheimers Dis 2010; 19: 11551167.

17 Sharma AN, Elased KM, Garrett TL, Lucot JB. Neurobehavioral decits in db/dbdiabetic mice. Physiol Behav 2010; 101: 381388.

18 Li XL, Aou S, Oomura Y, Hori N, Fukunaga K, Hori T et al. Impairment of long-termpotentiation and spatial memory in leptin receptor-decient rodents.Neuroscience 2002; 113: 607615.

19 Dietrich MO, Spuch C, Antequera D, Rodal I, de Yebenes JG, Molina JA et al.Megalin mediates the transport of leptin across the blood-CSF barrier. NeurobiolAging 2008; 29: 902912.

20 Power DA, Noel J, Collins R, ONeill D. Circulating leptin levels and weight loss inAlzheimers disease patients. Dement Geriatr Cogn Disord 2001; 12: 167170.

21 Moolman DL, Vitolo OV, Vonsattel JP, Shelanski ML. Dendrite and dendritic spinealterations in Alzheimer models. J Neurocytol 2004; 33: 377387.

22 Trinchese F, Liu S, Battaglia F, Walter S, Mathews PM, Arancio O et al. Progressiveage-related development of Alzheimer-like pathology in APP/PS1 mice.Ann Neurol 2004; 55: 801814.

23 Zhang H, Gong B, Liu S, Fa M, Ninan I, Staniszewski A et al. Synaptic fatigueis more pronounced in the APP/PS1 transgenic mouse model of Alzheimersdisease. Curr Alzheimer Res 2005; 2: 137140.

24 Perez-Gonzalez R, Antequera D, Vargas T, Spuch C, Bolos M, Carro E et al. Leptininduces proliferation of neuronal progenitors and neuroprotection in a mousemodel of Alzheimers disease. J Alzheimers Dis 2011; 24(Suppl 2): 1725.

25 Lundberg C, Bjorklund T, Carlsson T, Jakobsson J, Hantraye P, Deglon N et al.Applications of lentiviral vectors for biology and gene therapy of neurologicaldisorders. Curr Gene Ther 2008; 8: 461473.

26 Stewart S, Cacucci F, Lever C. Which memory task for my mouse? A systematicreview of spatial memory performance in the Tg2576 Alzheimers mouse model.J Alzheimers Dis 2011; 26: 105126.

27 Deacon RM, Rawlins JN. T-maze alternation in the rodent. Nat Protoc 2006; 1: 712.


307


28 Spuch C, Antequera D, Portero A, Orive G, Hernandez RM, Molina JA et al. Theeffect of encapsulated VEGF-secreting cells on brain amyloid load and behavioralimpairment in a mouse model of Alzheimers disease. Biomaterials 2010; 31:56085618.

29 OMalley D, MacDonald N, Mizielinska S, Connolly CN, Irving AJ, Harvey J et al.Leptin promotes rapid dynamic changes in hippocampal dendritic morphology.Mol Cell Neurosci 2007; 35: 559572.

30 Pinto S, Roseberry AG, Liu H, Diano S, Shanabrough M, Cai X et al. Rapid rewiringof arcuate nucleus feeding circuits by leptin. Science 2004; 304: 110115.

31 Shanley LJ, Irving AJ, Harvey J. Leptin enhances NMDA receptor function andmodulates hippocampal synaptic plasticity. J Neurosci 2001; 21: RC186.

32 Gylys KH, Fein JA, Yang F, Wiley DJ, Miller CA, Cole GM et al. Synaptic changes inAlzheimers disease: increased amyloid-beta and gliosis in surviving terminals isaccompanied by decreased PSD-95 uorescence. Am J Pathol 2004; 165: 18091817.

33 DAmelio M, Cavallucci V, Middei S, Marchetti C, Pacioni S, Ferri A et al. Caspase-3triggers early synaptic dysfunction in a mouse model of Alzheimers disease.Nat Neurosci 2011; 14: 6976.

34 Ivins KJ, Bui ET, Cotman CW. Beta-amyloid induces local neurite degeneration incultured hippocampal neurons: evidence for neuritic apoptosis. Neurobiol Dis1998; 5: 365378.

35 Abad MA, Enguita M, DeGregorio-Rocasolano N, Ferrer I, Trullas R. Neuronalpentraxin 1 contributes to the neuronal damage evoked by amyloid-beta and isoverexpressed in dystrophic neurites in Alzheimers brain. J Neurosci 2006; 26:1273512747.

36 Zhang Y, Wang H, Pan H, Bao X, Li M, Jin J et al. Gene delivery into primarycerebral cortical neurons by lentiviral vector. Cell Biol Int 2006; 30: 777783.

37 Carro E, Spuch C, Trejo JL, Antequera D, Torres-Aleman I. Choroid plexus megalinis involved in neuroprotection by serum insulin-like growth factor I. J Neurosci2005; 25: 1088410893.

38 Carro E, Trejo JL, Spuch C, Bohl D, Heard JM, Torres-Aleman I. Blockade of theinsulin-like growth factor I receptor in the choroid plexus originates Alzheimers-likeneuropathology in rodents: new cues into the human disease? Neurobiol Aging2006; 27: 16181631.

39 Antequera D, Vargas T, Ugalde C, Spuch C, Molina JA, Ferrer I et al. Cytoplasmicgelsolin increases mitochondrial activity and reduces Abeta burden in a mousemodel of Alzheimers disease. Neurobiol Dis 2009; 36: 4250.

40 Boghossian S, Ueno N, Dube MG, Kalra P, Kalra S. Leptin gene transfer in thehypothalamus enhances longevity in adult monogenic mutant mice inthe absence of circulating leptin. Neurobiol Aging 2007; 28: 15941604.

41 Iwaniec UT, Boghossian S, Lapke PD, Turner RT, Kalra SP. Central leptin genetherapy corrects skeletal abnormalities in leptin-decient ob/ob mice. Peptides2007; 28: 10121019.

42 Kofe RM, Meyer-Luehmann M, Hashimoto T, Adams KW, Mielke ML, Garcia-Alloza Met al. Oligomeric amyloid beta associates with postsynaptic densities and corre-lates with excitatory synapse loss near senile plaques. Proc Natl Acad Sci USA 2009;106: 40124017.

43 Lesne S, Koh MT, Kotilinek L, Kayed R, Glabe CG, Yang A et al. A specic amyloid-beta protein assembly in the brain impairs memory. Nature 2006; 440: 352357.

44 Lesne S, Kotilinek L, Ashe KH. Plaque-bearing mice with reduced levels ofoligomeric amyloid-beta assemblies have intact memory function. Neuroscience2008; 151: 745749.

45 Doherty GH, Beccano-Kelly D, Yan SD, Gunn-Moore FJ, Harvey J. Leptin preventshippocampal synaptic disruption and neuronal cell death induced by amyloidbeta. Neurobiol Aging 2013; 34: 226237.

46 Walsh DM, Selkoe DJ. Oligomers on the brain: the emerging role of solubleprotein aggregates in neurodegeneration. Protein Pept Lett 2004; 11: 213228.

47 Klein WL. Abeta toxicity in Alzheimers disease: globular oligomers (ADDLs) asnew vaccine and drug targets. Neurochem Int 2002; 41: 345352.

48 Ding Y, Qiao A, Wang Z, Goodwin JS, Lee ES, Block ML et al. Retinoicacid attenuates beta-amyloid deposition and rescues memory decits inan Alzheimers disease transgenic mouse model. J Neurosci 2008; 28:1162211634.

49 Torres M, Jimenez S, Sanchez-Varo R, Navarro V, Trujillo-Estrada L, Sanchez-Mejias Eet al. Defective lysosomal proteolysis and axonal transport are early pathogenicevents that worsen with age leading to increased APP metabolism and synapticAbeta in transgenic APP/PS1 hippocampus. Mol Neurodegener 2012; 7: 115.

50 Calhoun ME, Wiederhold KH, Abramowski D, Phinney AL, Probst A, Sturchler-Pierrat Cet al. Neuron loss in APP transgenic mice. Nature 1998; 395: 755756.

51 Wan J, Fu AK, Ip FC, Ng HK, Hugon J, Page G et al. Tyk2/STAT3 signaling mediatesbeta-amyloid-induced neuronal cell death: implications in Alzheimers disease.J Neurosci 2010; 30: 68736881.

52 Guo Z, Jiang H, Xu X, Duan W, Mattson MP. Leptin-mediated cell survival signalingin hippocampal neurons mediated by JAK STAT3 and mitochondrial stabilization.J Biol Chem 2008; 283: 17541763.

53 Bouret SG, Draper SJ, Simerly RB. Trophic action of leptin on hypothalamicneurons that regulate feeding. Science 2004; 304: 108110.

54 Farr SA, Banks WA, Morley JE. Effects of leptin on memory processing. Peptides2006; 27: 14201425.

55 Gispen WH, Biessels GJ. Cognition and synaptic plasticity in diabetes mellitus.Trends Neurosci 2000; 23: 542549.

56 Santarelli L, Saxe M, Gross C, Surget A, Battaglia F, Dulawa S et al. Requirementof hippocampal neurogenesis for the behavioral effects of antidepressants.Science 2003; 301: 805809.

57 Shors TJ, Miesegaes G, Beylin A, Zhao M, Rydel T, Gould E. Neurogenesis in theadult is involved in the formation of trace memories. Nature 2001; 410: 372376.

58 Dewachter I, Reverse D, Caluwaerts N, Ris L, Kuiperi C, Van den Haute C et al.Neuronal deciency of presenilin 1 inhibits amyloid plaque formation andcorrects hippocampal long-term potentiation but not a cognitive defect of amyloidprecursor protein [V717I] transgenic mice. J Neurosci 2002; 22: 34453453.

59 Arendt T, Zvegintseva HG, Leontovich TA. Dendritic changes in the basal nucleusof Meynert and in the diagonal band nucleus in Alzheimers disease a quanti-tative Golgi investigation. Neuroscience 1986; 19: 12651278.

60 Carro E, Trejo JL, Gomez-Isla T, LeRoith D, Torres-Aleman I. Serum insulin-likegrowth factor I regulates brain amyloid-beta levels. Nat Med 2002; 8: 13901397.

61 Peters A, Palay SL, de Webster H. The Fine Structure of the Nervous System Neuronsand their supporting cells. Oxford University Press: New York, 1991.

Supplementary Information accompanies this paper on Gene Therapy website (http://www.nature.com/gt)


308


title_linkIntroductionResultsLeptin rescues behavioral deficitsLeptin reduces Abeta plaque accumulationLeptin modulates synaptic deficits

Figure1HIV-leptin treatment of APPsolPS1 mice results in the attenuation of memory impairment. (a) Increased latencies in the spontaneous alternation of APPsolPS1 mice in the T-maze were partially normalized after HIV-leptin treatment (3 months). (b) InFigure2HIV-leptin APPsolPS1 mice exhibit reduced levels of Abeta deposits compared with HIV-vv APPsolPS1 mice. (a) Representative images of amyloid staining in cerebral cortex (upper panels), and hippocampus (bottom panel) in APPsolPS1 mice treated with Leptin modulates neurite length in neuronal cultures

DiscussionTable 1 Figure3HIV-leptin treatment prevents loss of synapses in the CA1 layer of hippocampus of APPsolPS1 mice. Representative photomicrographs by electron microscopy from non-transgenic mice (upper panel), HIV-vv-treated APPsolPS1 mice (middle panel) and HIV-lTable 2 Figure4HIV-leptin treatment prevents caspase 3 activation in synaptosome preparations from APPsolPS1 mice. (a) Representative WB showing the expression of procaspase-3 in cortical synaptosome preparations. Note that cleaved caspase-3 was undetectable. beFigure5Leptin modulates invivo and invitro synaptophysin expression. (a and c) Representative WBs of synaptophysin in cortical (a) and hippocampal (c) synaptosome preparations from wild-type and APPsolPS1 mice treated with HIV-vv or HIV-leptin. beta-acFigure6Leptin prevents the reduction in neurite length evoked by Abeta in neuronal cultures. (a and b) Representative immunofluorescence images of calbindin+ cortical neurons 18thinsph (a), and 36thinsph (b) after leptin (100thinspnM) or Abeta42 (5thinspMaterials and methodsLentiviral vectorsAnimalsBehavioral testingPrimary neuronal culturesDetermination of neurite outgrowthImmunohistochemistrySynaptosome preparationBiochemical measurementsElectron microscopyStatistical analysis

A5A6ACKNOWLEDGEMENTSA7

Documents

Leptin Gene Attenuates Neuronal PDF