RanBP9 overexpression reduces dendritic arbor … OVEREXPRESSION REDUCES DENDRITIC ARBOR AND SPINE DENSITY H. WANG, aM. LEWSADDER, E. DORN, S. XUb AND M. K. LAKSHMANAa* a Section of

Neuroscience 265 (2014) 253–262

RANBP9 OVEREXPRESSION REDUCES DENDRITIC ARBOR ANDSPINE DENSITY

H. WANG, a M. LEWSADDER, a E. DORN, a S. XU b ANDM. K. LAKSHMANA a*

aSection of Neurobiology, Torrey Pines Institute for

Molecular Studies, 11350 SW Village Parkway, Port Saint Lucie,

FL 34987, USA

bFlorida Institute of Technology, 150 West University

Boulevard, Melbourne, FL 32901, USA

Abstract—RanBP9 is a multi-domain scaffolding protein

known to integrate extracellular signaling with intracellular

targets. We previously demonstrated that RanBP9 enhances

Ab generation and amyloid plaque burden which results in

loss of specific pre- and postsynaptic proteins in vivo in a

transgenic mouse model. Additionally, we showed that the

levels of spinophilin, a marker of dendritic spines were inver-

sely proportional to the RanBP9 protein levels within the

synaptosomes isolated from AD brains. In the present study,

we found reduced dendritic intersections within the layer 6

pyramidal neurons of the cortex as well as the hippocampus

of RanBP9 transgenic mice compared to age-matched wild-

type (WT) controls at 12 months of age but not at 6 months.

Similarly, the dendritic spine numbers were reduced in the

cortex at only 12 months of age by 30% (p< 0.01), but not

at 6 months. In the hippocampus also the spine densities

were reduced at 12 months of age (38%, p< 0.01) in the

RanBP9 transgenic mice. Interestingly, the levels of

phosphorylated form of cofilin, an actin binding protein that

plays crucial role in the regulation of spine numbers were

significantly decreased in the cortical synaptosomes at only

12 months of age by 26% (p< 0.01). In the hippocampal

synaptosomes, the decrease in cofilin levels were 36%

(p< 0.01) at 12 months of age. Thus dendritic arbor and

spine density were directly correlated to the levels of

phosphorylated form of cofilin in the RanBP9 transgenic

mice. Similarly, cortical synaptosomes showed a 20%

(p< 0.01) reduction in the levels of spinophilin in the

RanBP9 transgenic mice. These results provided the

http://dx.doi.org/10.1016/j.neuroscience.2014.01.0450306-4522 � 2014 The Authors. Published by Elsevier Ltd.

*Corresponding author. Tel: +1-772-345-4698; fax: +1-772-345-3649.

E-mail address: [email protected] (M. K. Lakshmana).Abbreviations: AD, Alzheimer’s disease; ANOVA, analysis of variance;APP, amyloid precursor protein; BACE1, b-site APP cleaving enzyme1; BCA, Bicinchoninic acid; BDNF, Brain-derived neurotrophic factor;DRG, dorsal root ganglion; F-actin, filamentous-actin; HD, Huntington’sdisease; LC3, Light chain 3; LRP, low-density lipoprotein receptor-related protein; LTD, long-term depression; LTP, long-termpotentiation; MAPK, Mitogen-activated protein kinase; PBS,Phosphate buffered saline; PCR, polymerase chain reaction; PD,Parkinson’s disease; PFA, paraformaldehyde; Rho-GEF, Rho guaninenucleotide exchange factor; SDS–PAGE, Sodium dodecyl sulfate–Polyacrylamide gel electrophoresis; SEM, Standard error of mean;TEM, transmission electron microscopy; TFEB, transcription factor,EB; TG, transgenic mice; TGF-b, Transforming growth factor-b; WT,wild-type.

253

Open access under CC BY-

physical basis for the loss of synaptic proteins by RanBP9

and most importantly it also explains the impaired spatial

learning and memory skills previously observed in the

RanBP9 transgenic mice.

� 2014 The Authors. Published by Elsevier Ltd.

Key words: RanBP9, cofilin, dendritic arbor, spine density,

transgenic mice, Golgi staining.

INTRODUCTION

Neuronal morphology is crucial to our understanding of

information processing and communication in the brain

because neuronal shape is directly related to the

computations performed by the neuron (Spruston,

2008). The two most important morphological

characteristics of neurons are dendritic arbor structure

and dendritic spine density. The shape, size, and

complexity of dendritic trees can modulate action

potential propagation (Vetter et al., 2001) and influence

the firing pattern of a neuron (Mainen and Sejnowski,

1996). Similarly, the shape and the number of dendritic

spines play important roles in synaptic plasticity.

Increasing evidence indicates that deficient structural

neuronal network connectivity is a major, if not primary,

cause of several neurodegenerative disorders including

Alzheimer’s disease (AD) (Knobloch and Mansuy,

2008), Huntington’s disease (HD) (Spires et al., 2004)

and Parkinson’s disease (PD) (Day et al., 2006).

Moreover, changes in the structure and function of

dendritic spines contribute to several physiological

processes including synaptic transmission and learning

and memory (Kennedy et al., 2005; Tada and Sheng,

2006). Therefore identification of molecules that

inadvertently contributes to loss of dendritic arbor and

spine density is crucial in understanding their role in

neurodegenerative diseases.

We previously demonstrated that RanBP9 forms a

multi-protein complex with amyloid precursor protein

(APP), low-density lipoprotein receptor-related protein

(LRP) and b-site APP cleaving enzyme 1 (BACE1),

thereby regulate Ab generation (Lakshmana et al., 2009).

Consistent with our report, RanBP9 was recently found to

be within the clusters of RNA transcript pairs associated

with markers of AD progression (Arefin et al., 2012),

supporting our idea that RanBP9 might play a critical role

in the pathogenesis of AD. Our subsequent investigations

Open access under CC BY-NC-ND license.

NC-ND license.

254 H. Wang et al. / Neuroscience 265 (2014) 253–262

using transgenic mice overexpressing RanBP9 confirmed

that RanBP9 in fact regulates Ab generation and amyloid

plaque burden in vivo in the mouse brain. More

importantly, RanBP9 overexpression led to decreased

levels of specific presynaptic and postsynaptic proteins in

the brain, whereas RanBP9 null mice showed increased

levels of synaptic proteins (Lakshmana et al., 2012; Woo

et al., 2012a). These data taken together with

demonstrations by others that RanBP9 interaction with

plexin-A coordinates semaphorin3A signaling controlling

axonal outgrowth (Togashi et al., 2006) and also the

demonstration that RanBP9 regulates Brain-derived

neurotrophic factor (BDNF)-mediated neuronal

morphology and survival through the Mitogen-activated

protein kinase (MAPK) and Akt pathways (Yin et al.,

2010), all suggest essential role for RanBP9 at the

synapses. Also our finding that RanBP9 is present

throughout the neuron including neurites in the primary

neuronal cultures and within the whole dendritic network

in the adult brain (Lakshmana et al., 2012), followed by

our demonstration that RanBP9 levels are significantly

increased in the brains of patients with AD (Lakshmana

et al., 2010), as well as APP transgenic mice (Woo et al.,

2012a; Wang et al., 2013) strongly implicates RanBP9 to

play pivotal role in the regulation of synaptic density.

More recent work from our laboratory demonstrated an

excellent correlation between RanBP9 protein levels at

the synapses and the loss of spinophilin, a marker of

spines, in a brain region-specific manner due to defects

in the mitochondrial bioenergetics (Palavicini et al.,

2013). Interestingly, in the same study we found

normalized levels of RanBP9 protein relatively more in

the synaptosomes than the whole homogenate or

cytosolic fractions in the human brain (Palavicini et al.,

2013). Thus the enriched presence of RanBP9 at the

synapses is further evidence that RanBP9 has a crucial

role in the regulation of spines and synapses.

Here we further extended our studies and

demonstrated that RanBP9 overexpression in the

transgenic mice leads to significant reductions in the

dendritic arbor and spine density in both the hippocampus

and the cortical brain regions in an age-dependent

manner. Furthermore, phosphorylated form of cofilin, a

filamentous-actin (F-actin) severing protein that increases

the turnover of F-actin was highly reduced in the

synaptosomes which in turn correlated well with spine

density. Thus, we have now provided the physical basis

for loss of pre- and postsynaptic proteins by RanBP9.

EXPERIMENTAL PROCEDURES

Chemicals and antibodies

Ethanol (cat # E7023) and Xylene (cat # 2476) were

purchased from Sigma–Aldrich (St. Louis, MO, USA).

Tissue-Tek O.C.T. compound (cat # 4583) was

purchased from Sakura FineTek USA Inc. (Torrance, CA,

USA). Anti-flag-tag antibody (M2, cat # F3165) was

purchased from Sigma (St. Louis, USA). Polyclonal

phospho-cofilin antibody (cat # 3311), polyclonal

spinophilin antibody (cat # 9061S) and polyclonal Light

chain 3 (LC3) antibody (cat # 4599) were purchased from

Cell Signaling (Danvers, MA, USA). Monoclonal anti-

drebrin antibody (cat # D029-3) was purchased from MBL

international corporation (Woburn, MA, USA). Anti-TFEB

polyclonal antibody (cat # LS-C118813) was purchased

from LifeSpan Biosciences, Inc. (Seattle, WA, USA).

Polyclonal anti-TGF-b1 antibody (cat # NBP1-67698) and

caspase 3 antibody (cat # NB100-56708) were

purchased from Novus biologicals (Littleton, CO, USA).

Mouse monoclonal antibody against beta-actin (cat #

A00702) and polyclonal antibody against lamin-A were

purchased from Genscript USA Inc. (Piscataway, NJ,

USA). Secondary antibodies such as peroxidase-

conjugated AffiniPure goat anti-mouse (Code # 115-035-

146) and ant-rabbit (code # 111-035-144) IgGs were

purchased from Jackson ImmunoResearch Laboratories

(West Grove, PA, USA).

Mice

All animal experiments were carried out based on

ARRIVE guidelines and in strict accordance with the

National Institute of Health’s ‘Guide for the Care and

Use of Animals’ and as approved by the Torrey Pines

Institute’s Animal Care and Use Committee (IACUC).

Generation of RanBP9 transgenic mice have been

described previously (Lakshmana et al., 2012). The

RanBP9 specific primers used in the polymerase chain

reaction (PCR) is as follows. The forward primer is 50-

gcc acg cat cca ata cca g-30, and the reverse primer is

5-tgc ctg gat ttt ggt tct c-3’. Positive mice were then

backcrossed with native C57Bl/6 mice and the colonies

were expanded. RanaBP9 transgenic line 629 which

expressed the transgene in most brain regions was

used in this study. All mice were backcrossed to

maintain in the C57Bl/6 background, expanded and

genotyped and used at the specified ages. To avoid the

influence of gender, only male mice were used for both

WT and RanBP9-Tg genotypes.

The mice were fed with ad libitum food and water all

the time. The food is the irradiated global rodent chow

from Harlan. The mice were maintained in a 12-h light/

dark cycle at a temperature of 21–23 �C and a humidity

of 55 ± 10. After weaning, mice were kept in home

cages comprising single sex, single genotype and

groups of only five mice per cage. All of the mice lived

in an enhanced environment with increased amounts of

bedding and nesting materials.

Golgi staining

We used the FD Rapid Golgi Stain kit (FD

Neurotechnologies) to perform Golgi staining following

manufacturer’s protocol. Briefly, the mice were

euthanized and the brains were removed rapidly and cut

into small blocks of about 10 mm. The tissue blocks

were then rinsed briefly in double distilled water to

remove blood from the surface. The tissues were

immersed in the impregnation solution made by mixing

equal volumes of solutions A and B and changed the

solution after 24 h and stored for two weeks in the dark.

The tissues were then transferred to solution C,

replaced the solution again after 24 h and stored at 4 �C

H. Wang et al. / Neuroscience 265 (2014) 253–262 255

in the dark for a week. The tissues were rapidly frozen in

Tissue-Tek solution to prevent ice crystal formation which

might damage the sections. The brain was oriented such

that the plane of sectioning was perpendicular to the base

of the brain and then serial sections were cut in a rostral to

caudal direction at about 120-lm thickness in a cryostat

at �21 �C. The sections were then mounted on gelatin-

coated microscopic slides with a drop of solution C and

excess solution was removed with a Pasteur pipette and

the slides were dried naturally at room temperature.

Images were acquired in a fluorescent microscope

(Axio Examiner D1) using bright light. For quantification

of dendritic intersections, images of pyramidal neurons

from the layer 6 of cortex and the CA1 region of the

hippocampus were captured by selecting well-stained

neurons randomly at 40� magnification with water

immersion and for the analysis of dendritic spine density

images were acquired randomly at 100� magnification

with oil immersion. The automated quantitation of

dendritic intersections was done by Sholl analysis by

installing Sholl analysis plugin in the Image J application

folder. This plugin automates the task of doing Sholl

analysis on a neuron. Its algorithm is based on how

Sholl analysis is done manually by creating a series of

concentric circles around the soma of the neuron, and

counts how many times the neuron intersects with the

circumference of these circles. The images were first

converted into 8-bit grayscale images. Thresholding was

done to maintain similar background and noise on all

neurons. The pixels were converted into microns from

the scale of approximately 0.16873607 lm/pixel for 40�magnification images and 0.067060678 lm/pixel for

100� magnification images. The number of

intersections of dendrites was calculated with concentric

spheres positioned at radial intervals of 2 lm. The

dendritic morphology and spine quantification were done

by a blinded analyzer. The criteria used for analyzing

neurons are as follows: the pyramidal neurons had to be

fully impregnated and located either in the layer 6 of

cortex or the CA1 region of the hippocampus without

truncated branches and the soma located centrally

within the 120-lm section depth. The criteria for spines

included impregnation intensity allowing visibility of

spines, a low level of background, spines counted only

on dendrites starting at more than 85 lm distal to the

soma and after the first branch point.

Isolation of synaptosomes

To isolate synaptosomes, mice were euthanized under

isoflurane anesthesia and cortical and hippocampal

tissues from RanBP9 transgenic (TG) and age-matched

WT mice were weighed and dounced in a grinder using

Syn-PER synaptic protein extraction reagent (cat #

87793) purchased from Thermo Scientific (Rockford, IL,

USA). Immediately before use protease inhibitor mixture

for mammalian cells from Sigma (cat # P8340) was

added to the Syn-PER reagent. The homogenate was

centrifuged at 2000g for 10 min to remove cell debris.

The resulting supernatant was centrifuged at 15,000g

for 20 min. The supernatant formed the cytosolic

fraction and the synaptosome pellet was gently

resuspended in Syn-PER synaptic protein extraction

reagent. The amounts of total proteins in the

homogenate, cytosolic fraction and synaptosomes were

measured by Bicinchoninic acid (BCA) method and

compared. The quality of synaptosome preparation was

verified by immunoblotting for two cytosolic proteins

(TGFb and LC3), two nuclear proteins (lamin-A and

transcription factor, EB (TFEB)) and two synaptic

proteins (spinophilin and drebrin A).

Immunoblotting

Total protein concentrations of synaptosomes were

measured by BCA method (Pierce Biotechnology Inc.,

Rockford, USA). Equal amounts of proteins were loaded

into each well and subjected to Sodium dodecyl sulfate–

Polyacrylamide gel electrophoresis (SDS–PAGE)

electrophoresis. The proteins were then transferred onto

PVDF membranes, blocked with 5% milk and incubated

overnight with primary antibodies followed by one hour

incubation with HRP-conjugated secondary antibodies.

The protein signals were detected using Super Signal

West Pico Chemiluminescent substrate (Pierce,

Rockford, IL, USA).

Immunohistochemistry for caspase 3

RanBP9 transgenic and age-matched WT control mice

were deeply anesthetized using isoflurane and perfused

with 4% paraformaldehyde (PFA) in Phosphate buffered

saline (PBS). The brains were removed quickly and

immersed again in PFA solution with gentle rocking at

4 �C for 24 h. The rest of the immunohistochemical

staining procedure was exactly as published from our

laboratory (Palavicini et al., 2013). Images were

acquired by a laser-scanning confocal microscope

(Nikon 90i C1 SHS, Melles Griot laser system). The

images were deconvoluted, filtered and analyzed with

Image-Pro Plus 3D Suite software.

Transmission electron microscopy (TEM)

To prepare samples for TEM analysis, 3 ll of

synaptosomes were applied onto a TEM grid and

allowed to incubate for 3–4 min. Excess liquid was

removed with the edge of a kim wipe. The sample was

washed with 30–40 ll of deionized water and stained

with 4–5 ll of 2.5% uranyl acetate. The grid was

washed with 30–40 ll of deionized water and dried for

10 min before analyzing under TEM. Ultrastructure of

synaptosomes was imaged with a Joel 1010 TEM

(Peabody, MA, USA). Images were captured with a

Hamamatsu (Bridgewater, MA, USA) digital camera by

using AMT (Danvers, MA, USA) software.

Statistical analysis

Immunoblot signal for phospho-cofilin and spinophilin

were quantified using Image J software. Cofilin and

spinophilin levels in WT and RanBP9 transgenic mice

were analyzed by Student’s t-test. The differences in the

number of spines in the WT versus TG mice were

analyzed by Student’s t-test using Instat3 software

256 H. Wang et al. / Neuroscience 265 (2014) 253–262

(GraphPad Software, San Diego, CA, USA). We used

two-tailed p value assuming populations may have

different standard errors. The differences in the number

of dendritic intersections versus the radial distance from

soma in the WT versus TG mice were analyzed by a

one-way analysis of variance (ANOVA) followed by the

post hoc test. The data presented are mean ± Standard

error of mean (SEM). The data were considered

significant only if the p< 0.05, ⁄ indicates p< 0.05, and⁄⁄p< 0.01, ⁄⁄⁄p< 0.001.

Fig. 1. RanBP9 overexpression reduces dendritic intersections in the

pyramidal neurons of layer 6 of cortex at only12 months of age but not

at 6 months. A, Representative photomicrographs of Golgi-stained

cortical pyramidal neurons shown for 6- and 12-month- old mice

overexpressing RanBP9 (TG) and age-matched wild-type (WT)

controls. B, Sholl analysis of Golgi-stained neurons by Image J

software. The ordinate represents the distance from soma in lm and

the abscissa represents number of dendritic intersections that cross

along the concentric circles at defined distance from soma. Signif-

icant differences in the dendritic arbor were observed only in those

dendritic branches that originate approximately between 30 and

60 lm from soma as indicated by asterisks. ⁄p<0.01 in RanBP9

transgenic mice compared to WT mice by ANOVA followed by post

hoc test. Scale bar = 25 lm. The data are mean ± SEM, n= 6 for

each of RanBP9 TG and WT mice.

RESULTS

RanBP9 overexpression leads to age-dependentreduction in dendritic arbor in the pyramidal neuronsof cortex and the hippocampus

In order to understand the role of RanBP9 in synaptic

damage, we generated RanBP9 transgenic mice by

cloning 3x-flag-RanBP9 cDNA in the mouse thy-1 gene

cassette in the pTSC21K plasmid as described

previously (Lakshmana et al., 2012). We used thy-1

promoter to restrict RanBP9 expression to the postnatal/

adult brain only so that any adverse effect of RanBP9

during embryonic development may be prevented. It is

well known that the degree of complexity of dendritic

trees can modulate action potential propagation (Vetter

et al., 2001) and influence the intrinsic firing pattern of a

neuron (Mainen and Sejnowski, 1996). Particularly, the

action potential propagation is strongly influenced by the

number of dendritic branching points and the rate of

increase in dendritic membrane area (Vetter et al.,

2001). In order to understand the physical basis for loss

of synaptic proteins as well as learning deficits in the

RanBP9 transgenic mice (Lakshmana et al., 2012; Woo

et al., 2012a; Palavicini et al., 2013), we first directly

quantified the numbers of dendritic intersections in the

layer 6 pyramidal neurons of cortex and the CA1 region

of the hippocampus, the two most vulnerable brain

regions in AD. We analyzed 30 Golgi-stained neurons

per age per genotype of mice. Thus a total of 120

neurons were analyzed in the WT mice and another 120

neurons from the RanBP9 transgenic mice were

analyzed. We performed Sholl analysis of Golgi-stained

neurons by measuring the number of dendrites that

cross circles at different radial distances from the cell

body. The Sholl analysis plugin for Image J automates

the task of doing Sholl analysis on a neuron by creating

a series of concentric circles around the soma of the

neuron and counts how many times the neuron

intersects with the circumferences of these circles. Thus

Sholl analysis provides unbiased and automated

information on the dendritic branching patterns of neurons.

Analysis of dendritic arbor structure in the RanBP9

transgenic mice revealed a visible effect of RanBP9 on

the pyramidal neurons of layer 6 cortex at 12 months of

age but had no effect at 6 months of age (Fig. 1A, B).

Statistical analysis revealed significant reductions in

those dendritic intersections originating roughly between

30 lm and 60 lm from the soma of cortical neurons.

Quantitative data in the hippocampus suggested that

small reductions in the dendritic complexity of pyramidal

neurons in the CA1 region can be observed even at 6

months of age at about 60 lm from soma in the

RanBP9 transgenic mice (TG) compared to age-

matched wild-type (WT) mice (Fig. 2A, B), though it was

not statistically significant. However, more robust and

statistically significant reductions were seen in the

hippocampus in 12-month old mice starting from 15 lmand extending as far as 70 lm from soma. Thus

hippocampus is relatively more vulnerable brain region

in terms of loss of dendritic branches and complexity by

RanBP9 overexpression (Fig. 2A, B).

RanBP9 overexpression reduces number of dendriticspines in the pyramidal neurons of cortex and thehippocampus

It is now widely accepted that dendritic spines are

anatomical specializations on neuronal cells that form

distinct compartments that isolate input from different

synapses and are crucial for excitatory synaptic

transmission. Therefore, like dendritic arbor, the number

of spines can have a great impact on the neuronal

Fig. 3. Reduced spine density in the layer 6 of cortical pyramidal

neurons of brains from 12-month-old mice overexpressing RanBP9.

A, Representative examples of Golgi-stained cortical pyramidal

neurons showing dendritic segments at 100� magnifications to

display spines in the 6- and 12-month-old mice overexpressing

RanBP9 (TG) and age-matched wild-type (WT) controls. B, Semi-

automated quantitation of spine numbers per 10-lm dendritic

segment by image J software was subjected to statistical analysis.⁄⁄p< 0.01 in RanBP9 TG mice versus WT mice by Student’s t-test.The data are mean ± SEM, n= 6 for each of RanBP9 TG and WT

mice.

Fig. 4. Reduced spine density in the CA1 region of the hippocampal

pyramidal neurons of brains from 6- and 12-month-old mice over-

expressing RanBP9. A, Representative examples of Golgi-stained

hippocampal neurons showing dendritic segments at 100� magnifi-

cations to display spines in the 6- and 12-month-old mice over-

expressing RanBP9 (TG) and age-matched wild-type (WT) controls.

B, Semi-automated quantitation of spine numbers per 10-lmdendritic segment by image J software was subjected to statistical

analysis. ⁄⁄⁄p< 0.001 in RanBP9 TG mice versus WT mice by

Student’s t-test. The data are mean ± SEM, n= 6 for each of

RanBP9 TG and WT mice.

Fig. 2. RanBP9 overexpression reduces dendritic intersections in the

pyramidal neurons of CA1 region of the hippocampus at only 12

months of age but not at 6 months. A, Representative photomicro-

graphs of Golgi-stained hippocampal pyramidal neurons shown for

6- and 12-month- old mice overexpressing RanBP9 (TG) and age-

matched wild-type (WT) controls. B, Sholl analysis of Golgi-stained

neurons by Image J software. The ordinate represents the distance

from soma in lm and the abscissa represents number of dendritic

intersections that cross along the concentric circles at defined

distance from soma. Significant differences in the dendritic arbor

were observed only in those dendritic branches that originate

approximately between 15 and 70 lm from soma in the 12-month-

old mice as indicated by asterisks. ⁄p< 0.01 in RanBP9 transgenic

mice compared to WT mice ANOVA followed by post hoc test. Scale

bar = 25 lm. The data are mean ± SEM, n= 6 for each of RanBP9

TG and WT mice.

H. Wang et al. / Neuroscience 265 (2014) 253–262 257

function. Given the role of RanBP9 in reducing synaptic

proteins such as PSD95 and spinophilin (Lakshmana

et al., 2012; Palavicini et al., 2013), we hypothesized

that RanBP9 would also significantly reduce the number

of spines. This is especially true because spinophilin

which is a marker of spines is significantly reduced in

RanBP9 overexpressing APDE9 transgenic mice

(Palavicini et al., 2013). Similar to dendritic arbor, spine

density was not altered at 6 months of age in the

pyramidal neurons of the layer 6 cortex of RanBP9

transgenic mice compared to WT mice (Fig. 3A, B).

However, at 12 months of age spine density was

significantly reduced by 29% (p< 0.01) in the RanBP9

mice compared to WT mice (Fig. 3A, B). In the

hippocampus, similar to cortex, spine density was not

altered in the pyramidal neurons of CA1 region at 6

months of age. However, at 12-months the reduction

was 38% (p< 0.001) in the RanBP9 transgenic mice

versus WT controls (Fig. 4A, B). Although endogenous

versus exogenous expression of RanBP9 was 1:1 in the

Fig. 5. Characterization of synaptosomes by biochemical and mor-

phological methods. A, The purity of the synaptosomes prepared

from mouse brains was verified by immunoblotting the cytosolic (C),

homogenate (H) and synaptosomal (S) fractions for two cytosolic

proteins (TGFb and LC3), two nuclear proteins (lamin-A and

transcription factor, EB (TFEB)) and two synaptic marker proteins

(drebrin A and spinophilin). Please note enrichment of synaptic

proteins and the absence of nuclear proteins or cytosolic proteins in

the S fractions, attesting to the purity of synaptosomes. B, Trans-

mission electron microscopy (TEM) images at 20,000 magnification

showing intact synaptosomes. Arrows indicate the preservation of

postsynaptic densities at the synapses.

258 H. Wang et al. / Neuroscience 265 (2014) 253–262

cortex and only 1:0.8 in the hippocampus (Palavicini et al.,

2013), more robust reduction in the spine density in the

hippocampus (38%) compared to cortex (29%), clearly

suggest that hippocampus is more vulnerable to the

effect of RanBP9. Thus age- and brain region-specific

effect of RanBP9 on the spine density within the

pyramidal neurons was confirmed.

RanBP9 overexpression decreases levels ofphosphorylated form of cofilin in the synaptosomesof cortex and hippocampus

Dendritic spines are the postsynaptic sites of most

excitatory synapses in the brain and are highly enriched

in polymerized F-actin which drives the formation and

maintenance of mature spines. Cofilin is an F-actin-

severing protein that increases the turnover of F-actin

by severing the filaments and creating new barbed ends

for F-actin growth (Moriyama et al., 1990; Yahara et al.,

1996; Carlier et al., 1997; Lappalainen and Drubin,

1997; Rosenblatt et al., 1997). We recently showed that

RanBP9 dephosphorylates cofilin in primary

hippocampal neurons (Woo et al., 2012a). Since cofilin

is a key regulator of actin dynamics and because

dendritic spines are rich in actin molecules which

provide shape and structure to the spines, we wanted to

assess whether reduced spine density in the RanBP9

transgenic mice is due to changes in the levels of

phosphorylated cofilin protein.

Synaptosomes consist of presynaptic terminals

attached to postsynaptic dendritic spines that are pinched

off from the adjoining dendritic shaft, suggesting that they

can also serve as a model to study dendritic spines in

isolation. Therefore we isolated and quantified cofilin

protein levels in synaptosomes instead of whole brain

homogenates which might provide overall changes in the

neuron and is likely to dilute the effects of transgene.

Brain extracts were prepared as cytosolic (C),

homogenate (H) and synaptosomal (S) fractions by

centrifugation. We determined the purity of

synaptosomes by two independent methods. We first

qualitatively looked for two proteins in each of C, H and S

fractions. Cytosolic proteins such as TGF-b and LC3

were almost completely absent in the S fractions, but as

expected were present in both the C and H fractions

(Fig. 5A, left panels). Similarly nuclear proteins such as

lamin-A and transcription factor, EB (TFEB) were

completely absent in the S and C fractions, though

substantial amounts of these protein could be detected in

the H fractions (Fig. 5A, middle panels). Finally, we could

detect enriched amounts of two synaptic proteins, drebrin

A and spinophilin in the synaptosomal fractions relative to

H or C fractions (Fig. 5A, right panels). As such, the

synaptosomes can be used to reflect changes in protein

levels in the spines.

To determine whether synaptosomal architecture is

preserved in our preparation by another independent

method, we used TEM to examine the synaptosomes at

the ultrastructural level. As shown in Fig. 5B, we

observed synaptosomes with intact tightly opposed pre-

and post-synaptic elements held in close proximity with

each other. The postsynaptic density observed as dark

and thick layer are shown (arrows in Fig. 5B), which

represents the pinched-off dendritic spines. The plasma

membrane of most of the synaptosomes appeared

continuous suggesting that the cytoplasmic contents

inside the synaptosomes are not perturbed. Thus we

confirmed the integrity of our synaptosome preparations

by both biochemical and morphological methods.

Next, we quantified phosphorylated form of cofilin in the

synaptosomes isolated from the cortical and hippocampal

brain tissues and compared betweenWT and RanBP9 TG

mice. Similar to changes in the dendritic intersections as

well as spine density, cofilin levels in the cortical

synaptosomes isolated from 6-month old RanBP9

transgenic mice were not significantly altered (only 10%

reduction) when compared to control mice (Fig. 6A). At

12 months, however RanBP9 transgenic mice showed a

reduction of cofilin protein by 26% (p< 0.05) in the

cortical synaptosomes (Fig. 6A, B). Hippocampus also

did not show significant reductions (only 11%) in the

cofilin levels at 6 months of age. By 12 months of age,

the reduction was 36% (p< 0.01) in the RanBP9 TG

mice compared to WT controls (Fig. 6A, B). Thus the

reduction in the levels of phosphorylated form of cofilin

was consistent with changes in the dendritic arbor and

spine density.

RanBP9 overexpression decreases spinophilin levelsin the cortical synaptosomes

We previously demonstrated that RanBP9 overexpression

in the APDE9 mice significantly reduced spinophilin levels

Fig. 6. RanBP9 overexpression decreases phosphorylated form of cofilin protein levels in the synaptosomes of cortex and hippocampus. A,

Cortical and hippocampal synaptosomes from RanBP9 transgenic (TG) and age-matched wild-type (WT) control mice prepared from 6- and

12-month old mice were subjected to SDS–PAGE electrophoresis and probed with anti-phospho-cofilin antibody to detect phosphorylated form of

cofilin. Immunoblotting using flag specific monoclonal antibody detected flag-tagged exogenous RanBP9 in the TG mice but not in WT mice. Actin

was detected as a loading control. B, Image J quantitation did not reveal significant changes in the levels of cofilin in the cortex at 6 months of age

but by 12-months the levels were reduced significantly by 26%. Similarly, in the hippocampus cofilin levels were significantly reduced only at

12-months (36%). ⁄⁄p< 0.01 in RanBP9 TG mice versus WT control mice by Student’s t-test. Data are mean ± SEM, n= 5 for each of TG and WT

mice.

H. Wang et al. / Neuroscience 265 (2014) 253–262 259

in the synaptosomes (Palavicini et al., 2013). However it is

not clear whether RanBP9 transgenic mice also show

decreased spinophilin protein in the synaptosomes.

Decreased spine density in the RanBP9 transgenic mice

observed in the present study also prompted us to

quantify spinophilin levels in the synaptosomes.

Consistent with changes in cofilin levels, synaptosomes

isolated from the cortex did not show any alteration in

spinophilin protein at 6 months of age (Fig. 7A). At 12

months, however RanBP9 transgenic mice showed a

20% reduction (p< 0.01) when compared to

synaptosomes prepared from WT mice (Fig. 7A, B).

Thus although the extent of reduction in spinophilin

levels is lower than that of cofilin levels in the

synaptosomes, a decreased trend for both proteins in

synaptosomes is consistent with reduced spine density.

Fig. 7. RanBP9 overexpression decreases spinophilin protein levels

in the synaptosomes of cortex. A, Cortical synaptosomes from

RanBP9 transgenic (TG) and age-matched wild-type (WT) control

mice prepared from 6- and 12-month-old mice were subjected to

SDS–PAGE electrophoresis and probed with anti-spinophilin anti-

body to detect spinophilin protein. Immunoblotting using flag specific

monoclonal antibody detected flag-tagged exogenous RanBP9 in the

TG mice but not in WT mice. Actin was detected as a loading control.

B, Image J quantitation did not reveal significant changes in the levels

of spinophilin in the cortex at 6 months of age but by 12-months the

levels were reduced significantly by 20%. ⁄⁄p< 0.01 in RanBP9 TG

mice versus WT mice by Student’s t-test. Data are mean ± SEM,

n= 5 for each of TG and WT mice.

RanBP9 overexpression does not alter activatedcaspase 3-positive cells

The presence of activated caspase 3 is an indicator of

neurodegeneration in the brain. We stained for activated

caspase 3 by immunohistochemistry using an antibody

which specifically recognizes activated form of caspase

3. At both 6 and 12 months of age we could see only

few cells stained for activated caspase 3 in the cortex

as well as hippocampus of both the WT and RanBP9

transgenic mice (Fig. 8), suggesting that reduced spine

density as well as spinophilin and cofilin protein levels

are unlikely due to neurodegeneration.

Fig. 8. Cells positively stained for caspase 3 in the cortex and

hippocampus in the WT and RanBP9 TG mice. Representative brain

sections from cortex and hippocampus stained with anti-caspase 3

(red) and counter-stained with DAPI (blue). Only few Caspase 3

positive cells (red) were observed in both the RanBP9 TG mice and

the WT mice at 6 and 12months of age. (For interpretation of the

references to color in this figure legend, the reader is referred to the

web version of this article.)

260 H. Wang et al. / Neuroscience 265 (2014) 253–262

DISCUSSION

Here we report that RanBP9 overexpression in mice

results in age-dependent reductions in the dendritic

arbor and spine density in the pyramidal neurons of

layer 6 of cortex and the CA1 region of hippocampal

brain regions. It is interesting to note that the reductions

in dendritic intersections as well as spine density were

similarly altered in an age- and brain region-specific

manner. In addition, the reduced spine density in the

synaptosomes by RanBP9 was directly correlated with

the reduced protein levels of phosphorylated cofilin as

well as spinophilin. These results are consistent with

several properties of RanBP9 demonstrated previously

by others and from our laboratory.

We recently demonstrated by both immunohisto

chemistry and immunoblots that RanBP9 overexpression

in the APDE9 transgenic mice decreases the levels of

spinophilin, a marker of spines in the cortex and

hippocampus at 12 months of age (Palavicini et al.,

2013). We also showed that reduced spinophilin levels

were accompanied by reduced mitochondrial activity in

the synaptosomes, suggesting that the loss of

spinophilin is due to defects in mitochondrial

bioenergetics. The present finding of reduced spine

density by RanBP9 is consistent with decreased

spinophilin in the synaptosomes in the same brain

regions. Thus our recent demonstration of loss of

spinophilin (Palavicini et al., 2013) and other pre- and

post-synaptic proteins by RanBP9 (Lakshmana et al.,

2012; Woo et al., 2012a) can now be directly attributed

to loss of dendritic intersections and spines. A large

body of accumulating data points to the dendritic spines

as the principal signaling hub responsible for transducing

excitatory synaptic transmission and for the expression

of postsynaptic plasticity such as long-term potentiation

(LTP). LTP in turn is considered the physical basis for

learning and memory. Therefore loss of spines can also

explain impaired spatial learning skills demonstrated

previously from our laboratory using both a T maze

paradigm (Palavicini et al., 2013) and Morris water maze

(Woo et al., 2012a) in mice overexpressing RanBP9. But

it is important to remember that spines can undergo

structural changes within minutes and the spine density

estimated in the present study reports only steady state

levels at the time. This also explains why there was

significant reduction in the spine density at 12 months

but not at 6 months of age. It is also possible that even

in the absence of significant alterations in the spine

numbers, existing spines may not be functionally

efficient in signal processing, compromising neuronal

plasticity. As RanBP9 is enriched in the complex neuritic

processes in the cultured primary neurons and in the

dendritic processes in the neurons of adult brain

(Lakshmana et al., 2012), RanBP9 is expected to play

pivotal role at the synapses. Moreover, RanBP9 is a

ligand for Rho guanine nucleotide exchange factor (Rho-

GEF) (Bowman et al., 2008) and as Rho family

GTPases are essential regulators of actin polymerization

(Luo, 2002) as well as formation of dendritic spines

(Zhang et al., 2005; Tolias et al., 2007; Xie et al., 2007;

Saneyoshi et al., 2008; Wegner et al., 2008), RanBP9 is

also likely to play crucial role in the regulation of spine

density. More recent studies demonstrated that even in

the adult brain, Rho GTPases play pivotal role in the

plasticity of dendritic spines (Martino et al., 2013). The

actin cytoskeleton has long been suspected to be crucial

in controlling the development and stability of dendritic

spines as dendritic spines are highly enriched in actin.

Actin cytoskeleton is mainly regulated by Rho GTPases

that includes Rho, Rac, and cdc42 subfamilies (Etienne-

Manneville and Hall, 2002) which can regulate activity-

dependent structural plasticity by which dendritic spines

are produced and modified. Thus as a Rho GTPase

ligand, RanBP9 might contribute to the loss of dendritic

intersections and spines, which can account for the loss

of hippocampal-dependent learning and memory skills in

mice overexpressing RanBP9. In fact, RanBP9

overexpression results in drastic reductions in the growth

and branching of neurites from dorsal root ganglion

(DRG) neurons (Togashi et al., 2006) and cerebellar

neurons (Cheng et al., 2005).

Another most important observation in the present

study is the direct correlation between loss of spines as

H. Wang et al. / Neuroscience 265 (2014) 253–262 261

well as dendritic intersections and the reduced levels of

phosphorylated form of cofilin in the synaptosomes

derived from the hippocampus and cortical brain regions.

But what is not clear from this study is whether reduced

spine density is the cause or consequence of changes in

cofilin levels. Primary neuronal cultures transfected with

RanBP9 gene also resulted in decreased levels of

phosphorylated form of cofilin as determined by

immunoblots, though non phosphorylated form of cofilin

i.e., total cofilin levels were increased (Woo et al.,

2012a). Our present results of decreased cofilin levels in

the synaptosomes are consistent with these data and

confirm the direct and positive relationship between

phosphorylated cofilin levels and loss of spines. This is

also consistent with the finding that both LTP and long-

term depression (LTD) are impaired in mice lacking

cofilin with associated learning deficits in spatial,

aversive and reward learning (Rust et al., 2010). As

spine dynamics are completely dependent on the

cytoskeletal network, a change in the levels of actin-

remodeling protein such as cofilin is expected to alter

spine numbers. Binding of cofilin with actin filaments

increases the removal of actin monomers from the sharp

end of the filaments leading to filament depolymerization

which can alter the shape and size of spines. Although

RanBP9 decreased the levels of phosphorylated cofilin,

it is not clear how exactly RanBP9 influences the

phosphorylation status of cofilin. An obvious mechanism

may involve protein kinases and phosphatases. We

previously showed that RanBP9 binds both low-density

lipoprotein receptor-related protein (LRP) and b1-integrinand regulate their endocytosis (Woo et al., 2012b). The

Src and LIM kinases are activated downstream of

b1-integrins, and LIM kinase is known to control

phosphorylation of cofilin (Bernard, 2007). Thus RanBP9

can indirectly influence the phosphorylation of cofilin and

therefore actin dynamics, which in turn can alter spine

density. Additionally, synaptosomes derived from

RanBP9 transgenic mice showed reduced spinophilin.

Since spinophilin is considered a marker of dendritic

spines, reduced levels of spinophilin in the

synaptosomes is also consistent with reduced spine

density in the RanBP9 transgenic mice. Overall, these

data suggest that reduced dendritic arbor and spine

density in the RanBP9 transgenic mice may be due to

decreased levels of constitutively active phosphorylated

form of cofilin as well as spinophilin.

Acknowledgements—This work was supported by National Insti-

tute of Aging (NIA)/NIH Grant Nos. (1R03AG032064-01, M.K.

Lakshmana), 1R01AG036859-01, M.K. Lakshmana). We are

grateful to the staff at TPIMS vivarium for their support through-

out this study.

REFERENCES

Arefin AS, Mathieson L, Johnstone D, Berretta R, Moscato P (2012)

Unveiling clusters of RNA transcript pairs associated with markers

of Alzheimer’s disease progression. PLoS One 7:e45535.

Bernard O (2007) Lim kinases: regulators of actin dynamics. Int J

Biochem Cell Biol 39:1071–1076.

Bowman AL, Catino D, Strong JC, Randall WR, Kontrogianni-

Konstantopoulos A, Bloch RJ (2008) The rho-guanine

nucleotide exchange factor domain of obscurin regulates

assembly of titin at the Z-disk through interactions with Ran

binding protein 9. Mol Biol Cell 19:3782–3792.

Carlier MF, Laurent V, Santolini J, Melki R, Didry D, Xia GX, Hong Y,

Chua NH, Pantaloni D (1997) Actin depolymerizing factor (ADF/

cofilin) enhances the rate of filament turnover: implication in actin-

based motility. J Cell Biol 136:1307–1322.

Cheng L, Lemmo S, Lemmon V (2005) RanBPM is an L1-interacting

protein that regulates L1-mediatedmitogen-activated protein

kinase activation. J Neurochem 94:1102–1110.

Day M, Wang Z, Ding J, An X, Ingham CA, Shering AF, Wokosin D,

lliiic E, Sun Z, Sampson AR, Mugnaini E, Deutch AY, Sesack SR,

Arbuthnott GW, Surmeier DJ (2006) Selective elimination of

glutamatergic synapses on striatopallidal neurons in Parkinson

disease models. Nat Neurosci 9:251–259.

Etienne-Manneville S, Hall A (2002) Rho GTPases in cell biology.

Nature 420:629–635.

Kennedy MB, Beale HC, Carlisle HJ, Washburn LR (2005) Integration

of biochemical signalling in spines. Nat Rev Neurosci 6:423–434.

Knobloch M, Mansuy IM (2008) Dendritic spine loss and synaptic

alterations in Alzheimer’s disease. Mol Neurobiol 37:73–82.

Lakshmana MK, Yoon IS, Chen E, Bianchi E, Koo EH, Kang DE

(2009) Novel role of RanBP9 in BACE1 processing of amyloid

precursor protein and amyloid beta peptide generation. J Biol

Chem 284:11863–11872.

Lakshmana MK, Chung JY, Wickramarachchi S, Tak E, Bianchi E,

Koo EH, Kang DE (2010) A fragment of the scaffolding protein

RanBP9 is increased in Alzheimer’s disease brains and strongly

potentiates amyloid-beta peptide generation. FASEB J

24:119–127.

Lakshmana MK, Hayes CD, Bennet SP, Bianchi E, Reddy KM, Koo

EH, Kang DE (2012) Role of RanBP9 on amyloidogenic

processing of APP and synaptic protein levels in the mouse

brain. FASEB J 26:2072–2083.

Lappalainen P, Drubin DG (1997) Cofilin promotes rapid actin

filament turnover in vivo. Nature 388:78–82.

Luo L (2002) Actin cytoskeleton regulation in neuronal

morphogenesis and structural plasticity. Ann Rev Cell Dev Biol

18:601–635.

Mainen ZF, Sejnowski TJ (1996) Influence of dendritic structure

on firing pattern in model neocortical neurons. Nature

382:346–363.

Martino A, Ettorre M, Musilli M, Lorenzetto E, Buffelli M, Diana G

(2013) Rho GTPase-dependent plasticity of dendritic spines in the

adult brain. Front Cell Neurosci 7:1–11.

Moriyama K, Matsumoto S, Nishida E, Sakai H, Yahara I (1990)

Nucleotide sequence of mouse cofilin cDNA. Nucleic Acids Res

18:3053.

Palavicini JP, Wang H, Bianchi E, Xu S, Rao JS, Kang DE,

Lakshmana MK (2013) RanBP9 aggrvates synaptic damage in

the mouse brain and is inversely correlated to spinophilin levels in

Alzheimer’s brain synaptosomes. Cell Death Dis 4:e667.

Rosenblatt J, Agnew BJ, Abe H, Bamburg JR, Mitchison TJ (1997)

Xenopus actin depolymerizing factor/cofilin (XAC) is responsible

for the turnover of actin filaments in Listeria monocytogenes tails.

J Cell Biol 136:1323–1332.

Rust MB, Gurniak GB, Renner M, Vara H, Morando L, Gorlich A,

Sassoe-Pognetto M, Banchaabouchi MA, Giustetto M, Triller A,

Choquet D, Witke W (2010) Learning, AMPA receptor mobility

and synaptic plasticity depend on n-cofilin-mediated actin

dynamics. EMBO J 29:1889–1902.

Saneyoshi T, Wayman G, Fortin D, Davare M, Hoshi N, Nozaki N,

Natsume T, Soderling TR (2008) Activity-dependent

synaptogenesis: regulation by a CaM-kinase kinase/CaM-kinase

I/bPIX signaling complex. Neuron 57:94–107.

Spires TL, Grote HE, Garry S, Cordery PM, Van Dellen A, Blakemore

C, Hannan AJ (2004) Dendritic spine pathology and deficits in

experience-dependent dendritic plasticity in R6/1 Huntington’s

disease transgenic mice. Eur J Neurosci 19:2799–2807.

262 H. Wang et al. / Neuroscience 265 (2014) 253–262

Spruston N (2008) Pyramidal neurons: dendritic structure and

synaptic integration. Nat Rev Neurosci 9:206–221.

Tada T, Sheng M (2006) Molecular mechanisms of dendritic spine

morphogenesis. Curr Opin Neurobiol 16:95–101.

Togashi H, Schmidt EF, Strittmatter SM (2006) RanBPM contributes

to Semaphorin3A signaling through plexin-A receptors. J

Neurosci 26:4961–4969.

Tolias KF, Bikoff JB, Kane CG, Tolias CS, Hu L, Greenberg ME

(2007) The Rac1 guanine nucleotide exchange factor Tiam1

mediates EphB receptor-dependent dendritic spine development.

Proc Natl Acad Sci U S A 104:7265–7270.

Vetter P, Roth A, Hausser M (2001) Propagation of action potentials

in dendrites depends on dendritic morphology. J Neurophysiol

85:926–937.

Wang H, Dey D, Carrera I, Minond D, Bianchi E, Xu S, Lakshmana

MK (2013) COPS5 (Jab1) increases b-site processing of amyloid

precursor protein and Ab generation by stabilizing RanBP9

protein levels. J Biol Chem 288:26668–26677.

Wegner AM, Nebhan CA, Hu L, Majumdar D, Meier KM, Weaver AM,

Webb DJ (2008) N-wasp and the arp2/3 complex are critical

regulators of actin in the development of dendritic spines and

synapses. J Biol Chem 283:15912–15920.

Woo JA, Jung AR, Lakshmana MK, Bedrossian A, Lim Y, Bu JH, Park

SA, Koo EH, Mook-Jung I, Kang DE (2012a) Pivotal role of the

RanBP9-cofilin pathway in Ab-induced apoptosis and

neurodegeneration. Cell Death Differ 19:1413–1423.

Woo JA, Roh SE, Lakshmana MK, Kang DE (2012b) Pivotal role of

RanBP9 in integrin-dependent focal adhesion signaling and

assembly. FASEB J 26:1672–1681.

Xie Z, Srivastava DP, Photowala H, Kai L, Cahill ME, Woolfrey KM,

Shum CY, Surmeier DJ, Penzes P (2007) Kalirin-7 controls

activity-dependent structural and functional plasticity of dendritic

spines. Neuron 56:640–656.

Yahara I, Aizawa H, Moriyama K, Iida K, Yonezawa N, Nishida E,

Hatanaka H, Inagaki F (1996) A role of cofilin/destrin in

reorganization of actin cytoskeleton in response to stresses and

cell stimuli. Cell Struct Funct 21:421–424.

Yin YX, Sun ZP, Huang SH, Zhao L, Geng Z, Chen ZY (2010)

RanBPM contributes to TrkB signaling and regulates brain-

derived neurotrophic factor-induced neuronal morphogenesis

and survival. J Neurochem 114:110–121.

Zhang H, Webb DJ, Asmussen H, Niu S, Horwitz AF (2005) A GIT1/

PIX/Rac/PAK signaling moduleregulates spine morphogenesis

and synapse formation through MLC. J Neurosci 25:3379–3388.

(Accepted 22 January 2014)(Available online 31 January 2014)