The GABAergic septohippocampal connection is impaired in a mouse model of tauopathy

Accepted Manuscript The GABAergic septohippocampal connection is impaired in a mouse model of Tauopathy H. Soler, J. Dorca-Arévalo, M. González, S.E. Rubio, J. Avila, E. Soriano, M. Pascual PII:

S0197-4580(16)30219-6

DOI:

10.1016/j.neurobiolaging.2016.09.006

Reference:

NBA 9719

To appear in:

Neurobiology of Aging

Received Date: 25 May 2016 Revised Date:

5 September 2016

Accepted Date: 8 September 2016

Please cite this article as: Soler, H., Dorca-Arévalo, J., González, M., Rubio, S.E., Avila, J., Soriano, E., Pascual, M., The GABAergic septohippocampal connection is impaired in a mouse model of Tauopathy, Neurobiology of Aging (2016), doi: 10.1016/j.neurobiolaging.2016.09.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

TE D

PARV-positive GABAergic interneuron EP

GABAergic SH neuron

M AN U

SC

RI PT

Hippocampus

AC C

VLW mouse

WT mouse

Medial septum

P-Tau

Principal cells

ACCEPTED MANUSCRIPT 1

The GABAergic septohippocampal connection is impaired in a mouse model of

2

Tauopathy

3 4

H. Soler1,2, J. Dorca-Arévalo1,2, M. González1,2, S. E. Rubio1,2, J. Avila2,3, E.

5

Soriano1,2,4,5 and M. Pascual1,2,4

7

1

8

2

9

(CIBERNED, ISCIII), Spain;

RI PT

6 Department of Cell Biology, University of Barcelona, Barcelona, Spain;

Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas

10

3

11

4

12

5

13

Correspondence should be addressed to Marta Pascual ([email protected])

SC

Centro de Biología Molecular Severo Ochoa (CSIC-UAM), Madrid, Spain; Vall d′Hebron Institute of Research, Barcelona, Spain

M AN U

Institució Catalana de Recerca i Estudis Avançats, Barcelona, Spain

14 15

Key words: Alzheimer’s disease, hippocampal interneurons, parvalbumin-positive

16

cells, VLW mice, hyperphosphorylated Tau.

17

EP AC C

19

TE D

18

1

ACCEPTED MANUSCRIPT ABSTRACT

2

Alzheimer’s disease (AD), the most common cause of dementia nowadays, has been

3

linked to alterations in the septohippocampal pathway (SHP), among other circuits in

4

the brain. In fact, the GABAergic component of the SHP, which controls hippocampal

5

rhythmic activity crucial for learning and memory, is altered in the J20 mouse model of

6

AD—a model that mimics the amyloid pathology of this disease (Rubio et al., 2012).

7

However,

8

hyperphosphorylation and aggregation of the microtubule-associated protein Tau. To

9

evaluate whether tauopathies alter the GABAergic SHP, we analyzed transgenic mice

10

expressing human mutated Tau (VLW transgenic strain). We show that pyramidal

11

neurons,

12

interneurons in 2- and 8-month-old (mo) VLW mice accumulate phosphorylated forms

13

of Tau (P-Tau). By tract-tracing studies of the GABAergic SHP, we describe early onset

14

deterioration of GABAergic septohippocampal (SH) innervation on PARV-positive

15

interneurons in 2-mo VLW mice. In 8-mo animals, this alteration was more severe and

16

affected mainly P-Tau-accumulating PARV-positive interneurons. No major loss of

17

GABAergic SHP neurons or PARV-positive hippocampal interneurons was observed,

18

thereby indicating that this decline is not caused by neuronal loss but by the reduced

19

number and complexity of GABAergic SHP axon terminals. The decrease in

20

GABAergic SHP described in this study, targeted onto the PARV-positive/P-Tau-

21

accumulating inhibitory neurons in the hippocampus, establishes a cellular correlation

22

with the dysfunctions in rhythmic neuronal activity and excitation levels in the

23

hippocampus. These dysfunctions are associated with the VLW transgenic strain in

24

particular and with AD human pathology in general. These data, together with our

25

previous results in the J20 mouse model, indicate that the GABAergic SHP is impaired

26

in response to both amyloid-β and P-Tau accumulation. We propose that alterations in

27

the GABAergic SHP, together with a dysfunction of P-Tau-accumulating PARV-positive

28

neurons, contribute to the cognitive deficits and altered patterns of hippocampal activity

29

present in tauopathies, including AD.

is

cells,

and

by

some

another

pathophysiological

Parvalbumin

(PARV)-positive

hallmark:

the

hippocampal

AC C

EP

TE D

M AN U

mossy

characterized

SC

AD

RI PT

1

30

2

ACCEPTED MANUSCRIPT 1

INTRODUCTION

2 Amyloid-β deposits and Tau neurofibrillary tangles (NFTs) are the most typical

4

neuropathological hallmarks of Alzheimer’s disease (AD)(Bloom, 2014; Castellani et

5

al., 2010; Selkoe, 1991). The cholinergic component of the basal forebrain, including

6

the medial septum/diagonal band of Broca (MS/DB) and the nucleus basalis

7

magnocellularis, innervating the cerebral cortex and the hippocampus, undergoes

8

progressive deterioration in AD and also during normal aging. Moreover, it is believed

9

that reduced function of the cholinergic septohippocampal pathway (SHP) may underlie

10

early impairment of hippocampus-based episodic memory (Whitehouse et al., 1982;

11

Ypsilanti et al., 2008).

12

In addition to the cholinergic pathway, the SHP has a second important component,

13

namely GABAergic neurons (Freund and Antal, 1988; Gulyas et al., 1990). This long-

14

range

15

interneurons, which in turn regulate the activity of pyramidal neurons. It has been

16

proposed that activation of GABAergic SHP neurons leads to the selective inhibition of

17

inhibitory interneurons, thereby allowing the synchronous activation of large numbers of

18

pyramidal neurons. Thus, the GABAergic SHP may provide correct levels of excitation,

19

as well as regulate synchronous neuronal activities, including the theta and gamma

20

rhythms, which are essential for memory and cognition (Bland et al., 2007; Buzsáki,

21

2005; Freund and Gulyas, 1997; Garner et al., 2005; Hangya et al., 2009; Kitchigina et

22

al., 2013; Sotty et al., 2003; Toth et al., 1997). Recent studies have reported alterations

23

also in the GABAergic component of the SHP (Loreth et al., 2012) associated with AD

24

in a triple-transgenic mouse line (TauPS2APP). In this regard, a mouse model with a

25

considerable accumulation of amyloid-β deposits (hAPPSw,Ind; J20 mice, Mucke et al.,

26

2000; Palop et al., 2003) shows a dramatic and early decrease in the GABAergic SHP.

27

This decline is caused by the reduced number and complexity of GABAergic SHP axon

28

terminals, which in turn correlates with changes in network activity and in the

29

electrophysiology of the target hippocampus (Rubio et al., 2012; Vega-Flores et al.,

30

2014).

31

Another feature observed in AD and in other neurodegenerative diseases is the

32

hyperphosphorylation and aggregation of the microtubule-associated protein Tau. In

33

this context P-Tau protein accumulates in the somatodendritic compartment that finally

34

forms NFTs. NTFs are thought to cause synaptic and neuronal dysfunction and

M AN U

pathway

terminates

specifically

on

GABAergic

hippocampal

AC C

EP

TE D

inhibitory

SC

RI PT

3

3

ACCEPTED MANUSCRIPT neuronal death. In addition, recent evidence indicates that soluble P-Tau protein is a

2

key toxic species (Spires et al., 2006). Tauopathy leads to a loss of synapses and

3

neurons, thus leading to cognitive and behavioral impairment (Ballatore et al., 2007;

4

Frost et al., 2014).

5

The VLW mouse line over-expresses human Tau with four tubulin-binding repeats and

6

three mutations (G272V, P301L, and R406W), associated with frontotemporal

7

dementia and Parkinsonism linked to chromosome 17 (FTDP-17). Previous analysis of

8

VLW mutant mice show that cortical and hippocampal neurons accumulate prefibrillary

9

Tau-aggregates

in

somata

and

dendrites—a

RI PT

1

phenomenon

resembling

the

accumulation associated with AD (Lim et al., 2001). VLW mice present severe memory

11

impairment, including retention of episodic memory (Rodríguez-Navarro et al., 2008). In

12

addition, hyperexcitability and epileptic activity have recently been described in this

13

mouse model (García-Cabrero et al., 2013).

14

Here we addressed whether the phosphorylation of pathological Tau leads to

15

alterations in the GABAergic SHP connection that could cause cognitive deficits

16

associated with tauopathies. Our data indicate that VLW mice accumulate P-Tau in

17

pyramidal neurons and in PARV-positive interneurons in the hippocampus. VLW

18

mutant mice show a decrease in GABAergic SHP innervation, and this reduction is not

19

caused by neuronal loss but by a reduced number and complexity of GABAergic SHP

20

synaptic contacts.

23

M AN U

TE D

EP

22

AC C

21

SC

10

4

ACCEPTED MANUSCRIPT 1

MATERIALS AND METHODS

2 3

Animals

4 For the histological procedures, we used wild-type male mice (C57BL/6J strain; 2–3

6

and 8–9 month-old; n = 4-6 per group) and homozygous transgenic male mice with the

7

same genetic background that over-express human Tau with four tubulin-binding

8

repeats and three mutations (G272V, P301L, and R406W) associated with FTDP-17

9

(VLW line; 2–3 and 8–9 months old; n = 5–6 per group). All animals were kept on a 12

10

h light/dark schedule with access to food and water ad libitum. All animal experiments

11

were performed in accordance with the European Community Council Directive and the

12

National Institute of Health guidelines for the care and use of laboratory animals.

13

Experiments were also approved by the local ethical committees.

M AN U

SC

RI PT

5

14 15

Detection of septohippocampal fibers

16

Animals were deeply anesthetized [mixture 10/1 Ketolar (Parke-Davis)/Rompun

18

(Bayer), 0.003 ml/g] and stereotaxically injected with an anterograde tracer (10%

19

biotinylated dextran-amine (BDA) 10000 MW, Molecular Probes) in the MS/DB

20

complex. Each animal received midline injections of the tracer into the MS/DB complex

21

at one anteroposterior (AP) level, and at two dorsoventral (DV) injection points by

22

iontophoresis (7 µA positive direct current, 7 s on–off cycle). Stereotaxic coordinates

23

were (from Bregma): AP + 0.7, and DV -3.0 and -3.7. This protocol results in intense

24

BDA labeling in the MS/DB complex area, which contains the highest proportion of

25

GABAergic SH neurons (Pascual et al., 2004). Five or six days later, the animals were

26

deeply anesthetized and perfused with 4% paraformaldehyde in 0.12M phosphate

27

buffer. The brains were frozen and 30-µm sections were cut. Coronal sections were

28

stored in a cryoprotectant solution (30% glycerin, 30% ethylenglycol, 40% 0.1 M

29

phosphate buffer) at -20ºC until use. To visualize BDA, after blocking, the sections

30

were incubated overnight at 4ºC with the ABC complex (Vectastain ABC Kit; Vector

31

Laboratories, Burlingame, CA, USA) diluted 1/100. Peroxidase activity was developed

32

with diaminobenzidine intensified with nickel ammonium sulfate and cobalt chloride

33

(DAB/Ni-Co), and H2O2. Thereafter, sections were mounted onto gelatinized slides,

34

Nissl-stained, and coverslipped with Eukitt.

AC C

EP

TE D

17

5

ACCEPTED MANUSCRIPT 1 2

Immunodetection

3 Some hippocampal sections from each of the iontophoretically injected animals were

5

processed for the double immunodetection of BDA and several interneuronal markers

6

(Matyas et al., 2004; Pascual et al., 2004; Rocamora et al., 1996). After blocking, free-

7

floating sections of the tissue were incubated overnight at 4ºC with the ABC complex

8

(Vectastain ABC Kit; Vector Laboratories) diluted 1/100 simultaneously with well-

9

characterized rabbit polyclonal antibody (Swant Antibodies, Switzerland) against

10

parvalbumin (PARV, 1/3000) or glutamic acid decarboxylase 65/67 (GAD65/67,

11

Chemicon International, Temecula, CA, USA) diluted 1/1000. BDA was revealed using

12

DAB/Ni-Co, yielding a black end-product in SH fibers. Primary antibodies were then

13

visualized by sequential incubation with biotinylated secondary antibodies and the ABC

14

complex (2 h each, Vector Laboratories). The peroxidase reaction was developed with

15

DAB to produce a brown end-product. The sections were mounted onto gelatinized

16

slides, dehydrated, and coverslipped with Eukitt.

17

In order to detect the accumulation of P-Tau, tissue sections from VLW mice were

18

blocked and incubated overnight with AT-180 mouse anti-phosphothreonine 231 in Tau

19

(1/300) (Goedert et al., 1994). Primary antibody visualization was performed by

20

sequential incubation with anti-mouse biotinylated secondary antibodies and ABC

21

complex (Vector Laboratories). Peroxidase activity was developed with DAB/Ni-Co in

22

order to intensify the signal. Some sections of iontophoretically injected VLW animals

23

were processed for double detection of BDA injected in the MS/DB (DAB/Ni-Co, black

24

signal) and P-Tau immunodetection (DAB alone, brown signal).

25

For double immunofluorescent detection of P-Tau and PARV, calretinin (CALR),

26

calbindin (CALB) or somatostatin (SOM), the sections were incubated overnight

27

simultaneously with rabbit anti-PARV (1/3000), anti-CALR (1/2000), anti-CALB (1/3000,

28

Swant Antibodies, Switzerland) or anti-SOM (1/1000, Chemicon International,

29

Temecula, CA, USA) antibodies and AT-180 mouse anti-PThr231 (1/300). Some

30

sections were stained with rabbit anti-PARV antibody (1/3000) combined with mouse

31

antibodies AT-8(1/500) and 12E8(1/500) in order to recognize Tau protein

32

phosphorylated in Ser202/Thr205 and Ser262/Ser356 respectively. In all double

33

immunodetections, the sections were incubated with Alexa Fluor 568 goat anti-rabbit

34

IgG and Alexa Fluor 488 goat anti-mouse IgG (Invitrogen, Carlsbad, CA, USA) for 2 h.

AC C

EP

TE D

M AN U

SC

RI PT

4

6

ACCEPTED MANUSCRIPT They were then mounted onto slides, coverslipped with Mowiol 4-88 (Merck,

2

Darmstadt, Germany), and viewed under a confocal microscope.

3

Sections corresponding to the MS were used for fluorescent immunodetection of

4

PARV. After blocking, tissue was incubated overnight with anti-PARV antibody and

5

then incubated with Alexa Fluor 488 donkey anti-rabbit IgG (Invitrogen). The sections

6

were mounted and coverslipped with Mowiol 4-88.

RI PT

1

7 8

Analysis of the histological sections

9

Microscopic observations were focused on sections corresponding to the MS and to

11

dorsal (sections between 1.60 and 2.30 mm posterior to Bregma) and ventral (sections

12

2.90 and 3.40 mm posterior to Bregma) hippocampal levels, following the atlas

13

reported by Paxinos and

14

density of GABAergic SH neurons, PARV-immunopositive cells in the MS/DB complex

15

in serial sections of brains (8-mo WT and VLW mice, n=4–6 animals/genotype; 3–5

16

sections/animal) were counted manually under a fluorescent optical microscope (Nikon

17

E600; Nikon Corp., Kawasaki, Japan), and the results were expressed as number of

18

PARV-positive cells in the MS/DB area per section.

19

To estimate the density of hippocampal interneurons and the percentage of these cells

20

contacted by GABAergic SH fibers, the density of interneurons containing PARV or P-

21

Tau, as well as the percentage of these interneurons receiving the BDA-positive

22

pericellular baskets, was calculated in the distinct regions of the hippocampal area

23

(DG, CA3 and CA1) of each section (4–6 animals/genotype and age, 4–8

24

sections/animal). The area comprising the hippocampal regions of each section was

25

quantified using ImageJ program. The density of hippocampal interneurons was

26

defined as the density of interneurons per square millimeter. A similar procedure was

27

performed for GAD65/67-immunopositive cells. However, in this case, due to the large

28

number of GAD65/67-positive cells, we selected several sample areas for each

29

section. The selected samples (125-µm-wide stripes) contained all hippocampal layers

30

(perpendicularly from the ventricle to the pial surface) and each section included 1

31

stripe in the DG, 2 in the CA3, and 3 in the CA1. For GAD65/67 quantification, 4–6

32

animals/genotype and age were used, and 4–8 sections/animal were analyzed. Data

33

were represented as above (density of interneurons per square millimeter, and

34

percentage of GAD67/65-positive cells contacted by GABAergic SH fibers).

SC

10

AC C

EP

TE D

M AN U

Franklin (Paxinos and Franklin, 2001). To estimate the

7

ACCEPTED MANUSCRIPT To assess the complexity of the GABAergic SH contacts, synaptic boutons around the

2

somata of hippocampal interneurons were counted under a conventional microscope

3

for GAD65/67-positive, for P-Tau, and for PARV-positive interneuron subpopulations,

4

using the procedure described above.

5

To quantify the intensity of the AT-180 immunodetection signal, optical density was

6

obtained by scanning (SilverFast) and analyzed by Fiji software. Four VLW mice/age

7

(2- and 8-mo) and 5-7 sections/animal processed simultaneously were included in the

8

analysis.

9 Statistical analysis of histological data

SC

10

RI PT

1

11

For the statistical analysis, we estimated the densities of GABAergic SH (PARV-

13

immunopositive) cells per section for each mouse. For each hippocampal interneuronal

14

population studied (PARV- ,GAD65/67 or P-Tau-immunopositive neurons), the density

15

of interneurons per square millimeter, percentage of contacted interneurons, and

16

number of boutons per single basket were estimated for each hippocampal area or for

17

total hippocampal region. In all statistical analyses, the number of animals was

18

considered as replicates (n= 4–6 animals/genotype and age, 4–8 sections/animal were

19

used).

20

Histological data were processed for statistical analysis with StatGraphics Plus 5.1

21

(Statistical Graphics, Rockville, MD, USA). 2-sample Student’s t test (2-tailed) or Mann-

22

Whitney W test (2-tailed) were used (when the samples fitted a normal distribution or

23

not, respectively) to examine differences between the experimental groups. The

24

significance was set at P< 0.05 for the 2-tailed tests. Statistical information is reported

25

throughout the text in the format: statistic (degrees of freedom) = value, p = value.

27 28

TE D

EP

AC C

26

M AN U

12

8

ACCEPTED MANUSCRIPT 1

RESULTS

2 Distribution of P-Tau in VLW mice.

4

To characterize the distribution of P-Tau, we performed immunodetection using AT-180

5

antibody on sections from 2- and 8-mo VLW mice. The AT-180 antibody specifically

6

recognizes Tau protein phosphorylated in Threonine 231. As previously described (Lim

7

et al., 2001), in adult animals (8-mo), neurons in the pyramidal layer of the CA1

8

accumulated high levels of Thr231P-Tau in the soma and proximal dendrites (Fig. 1A).

9

In addition, these animals also accumulated AT-180-positive P-Tau in mossy cells in

10

the dentate gyrus (DG, Fig. 1G), in some pyramidal neurons in the CA3 region (Fig.

11

1A), and in some sparsely distributed cells in all hippocampal regions (Fig. 1D, arrows).

12

The AT-180 signal in the mossy cells was weaker in 2-mo mice than in older mice (DG,

13

Fig. 1B). In addition, CA1-CA3 pyramidal neurons and some neurons scattered in

14

distinct hippocampal strata and regions presented P-Tau accumulation since 2 mo (Fig.

15

1E, arrows). Quantification of the intensity of the AT-180-positive immunostaining in the

16

CA3 and CA1 pyramidal layer and in the hilus of the DG revealed no differences

17

regarding the level of AT-180-positive P-Tau accumulation in 2-mo VLW mice,

18

compared to that of 8-mo VLW mice (CA3: 156382.05±2827.4 and 157629.2±1805.5

19

arbitrary units, respectively, t(6)=0.38, P=0.70, Student’s t test; CA1: 158063.7±637.3

20

and 158668.4±1942.6 arbitrary units, respectively, t(6)=0.29, P=0.77, Student’s t test;

21

DG: 155344.5±2484.4 and 159484.6±2140.3 arbitrary units, respectively, t(6)=1.2,

22

P=0.24).

23

Our recent data (Rubio et al., 2012), showed an important decrease in SH innervation

24

in a mouse model of AD with amyloid-β accumulation; since we were interested in

25

analyzing SH innervation in VLW mice, we first tested whether SH neurons

26

accumulated P-Tau in this mouse model. We found that no cells expressing AT-180-

27

positive P-Tau were present in the MS/BD region of 2- or 8-mo animals (Fig. 1C). In

28

addition, no accumulation was observed in control mice in any region analyzed (Fig.

29

1F). Thus, our results show that in VLW mice P-Tau accumulate in CA1 and CA3

30

pyramidal neurons, in the mossy cells and in neurons scattered throughout

31

hippocampal layers.

AC C

EP

TE D

M AN U

SC

RI PT

3

32 33

PARV-positive hippocampal interneurons accumulate P-Tau in VLW mice.

9

ACCEPTED MANUSCRIPT A more detailed analysis of scattered AT-180-positive cells in 8-mo mice revealed that

2

they were located mainly in the strata oriens and radiatum close to the pyramidal layer

3

(Fig. 1D,E), and also near the granule cell layer. The distribution and morphology of

4

these AT-180-positive cells seemed to correspond to basket and axo-axonic

5

hippocampal interneurons containing PARV (Freund and Buzsaki, 1996; Matyas et al.,

6

2004). To ascertain whether the PARV-positive neurons of VLW mice present P-Tau,

7

we performed double immunofluorescent detection against PARV and AT-180. Our

8

results showed that some, but not all, PARV-positive cells in the CA1, CA3 and DG

9

accumulated AT-180-positive P-Tau forms in soma and proximal dendrites of 8-mo

10

animals (Fig. 2A-C, D-F). Double-positive cells were also observed in the CA1, CA3

11

and DG of 2-mo animals (Fig. 2J-L). In order to examine whether P-Tau also

12

accumulates in other hippocampal populations, we performed double immunodetection

13

against P-Tau and CALR, SOM and CB, markers for specific populations of

14

hippocampal interneurons (Freund and Buzsaki, 1996; Gulyas et al., 1996; Matyas et

15

al., 2004). No co-localization of P-Tau and CALR was observed in any region analyzed

16

(Fig. 2G-I). Neither was colocalization detected between P-Tau and SOM or CB (data

17

not shown). In order to analyze whether PARV-positive cells accumulate Tau protein

18

phosphorylated exclusively in Thr231 or also Tau phosphorylated in other residues, we

19

performed double immunodetections combining PARV and AT-8 or 12E8, which

20

recognize Ser202/Thr205 and Ser262/Ser356, respectively. Although an intense

21

positive signal was present in the pyramidal cell layer, no AT-8- or 12E8-positive signal

22

was found in PARV-positive cells in VLW mice (Suppl. Fig 1A-D). Thus, our data

23

indicated that PARV-positive hippocampal neurons in this mouse model specifically

24

accumulate Tau phosphorylated in Thr231.

SC

M AN U

TE D

EP

AC C

25

RI PT

1

26

No major loss of GABAergic hippocampal interneurons in VLW mice.

27

Our results showed that hippocampal interneurons accumulate P-Tau forms in VLW

28

mice. Such accumulation has been described to be an initial feature of

29

neurodegeneration (Ballatore et al., 2007; Frost et al., 2014); consequently, we studied

30

whether VLW mice present GABAergic neuronal death and, accordingly, an alteration

31

in the number of hippocampal interneurons. With this aim, we performed GAD65/67

32

immunodetections on 2- and 8-mo VLW and control mice in order to stain all

33

GABAergic cells in the hippocampus (Fig. 3A,B). First, we analyzed the density of

34

hippocampal interneurons during the maturation, comparing the number of GAD- and 10

ACCEPTED MANUSCRIPT PARV-positive cells in 2- and 8-mo VLW and WT mice. Our data showed that the

2

density of GAD-positive (2-mo: 183.0±15.6 and 8-mo: 209.2 ±13.72 GAD-positive

3

cells/mm2) and PARV-positive (2-mo: 51.49±5.17 and 8-mo: 52.15±2.19 PARV-positive

4

cells/mm2) cells remained constant in the former and also in the two stages of WT mice

5

(2-mo: 186.19±13.08 and 8-mo: 184.55±7.22 GAD-positive cells/mm2 and 2-mo:

6

48.50±8.09 and 8-mo: 48.05±1.83 PARV-positive cells/mm2), thereby indicating no

7

increase or decrease in the number of hippocampal interneurons in WT or VLW mice

8

between 2- and 8-mo.

9

In addition, no reduction in the density of GAD65/67-positive neurons in the

10

hippocampus of VLW mice compared to WT animals was detected in 2- or 8-mo mice

11

(t(6)=0.15, P=0.88 and t(9)=1.49, P=0.16 respectively, Student’s t test) (Fig. 3C). Since

12

PARV-positive cells of VLW mice accumulated P-Tau, we also specifically analyzed the

13

population of PARV-positive interneurons (Fig. 3D,E). Our data indicated no alterations

14

in the density of PARV-positive cells in 2- or 8-mo VLW mice (t(6)=0.38, P=0.71 and

15

t(8)=0.25, P=0.80 respectively, Student’s t test) (Fig. 3F), thus excluding the

16

occurrence of hippocampal interneuron degeneration in these animals.

17 Reduction

GABAergic

SH

innervation

on

GAD-positive

interneurons

19

accumulating P-Tau.

20

To trace the GABAergic SHP in the VLW mouse model, we performed iontophoretic

21

injections of BDA in the MS, following a protocol which virtually fills the MS/DB complex

22

with tracer (Pascual et al., 2004). As shown previously, SHP axons innervated all

23

layers of the hippocampus and DG, and two types of fiber were clearly distinguished,

24

namely thin axons with numerous “en passant” boutons, corresponding to cholinergic

25

fibers, and thick GABAergic axons that established complex basket-like nets with large

26

boutons around the perisomatic region of hippocampal interneurons (Freund and Antal,

27

1988; Pascual et al., 2005; Rocamora et al., 1996; Rubio et al., 2012).

28

Sections from 2- and 8-mo VLW mice and the corresponding control WT animals were

29

simultaneously stained for BDA and GAD65/67, the later as a marker for all GABAergic

30

hippocampal interneurons. The pattern of GABAergic SH innervation in VLW mice was

31

indistinguishable from that observed in the WT group (Fig. 4A,B). No significant

32

differences between 2-mo VLW and control mice were observed in the percentages of

33

contacted GAD65/67-positive hippocampal interneurons (Fig. 4C-E), neither in the

34

complexity of GABAergic SH contacts (Fig. 4C,D,F).

TE D

18

AC C

EP

of

M AN U

SC

RI PT

1

11

ACCEPTED MANUSCRIPT However, although no statistical differences were observed at 8-mo, there was a clear

2

downward trend in the percentages of GAD65/67-positive hippocampal interneurons

3

contacted by the GABAergic SH fibers between VLW and control mice (Fig. 4 G,H,J).

4

Neither were differences in the complexity of GABAergic SH contacts on GAD-positive

5

hippocampal interneurons (Fig. 4K,L,N). To address whether the downward trend in the

6

GABAergic SH innervation observed at 8 months specifically affects P-Tau-

7

accumulating interneurons, we co-stained sections from 8-mo VLW mice with P-Tau

8

antibodies and BDA (Fig. 4I,M). While the percentage of P-Tau-positive cells receiving

9

GABAergic SH innervation was similar to the overall population of GAD-positive cells in

10

VLW mice (Fig. 4H,I,J), the complexity of the synaptic contacts (boutons/cell) was

11

reduced by 40% (t(6)=7.36, P-value = 0.00015, Student’s t test) (Fig. 4L,M,N). Our data

12

support the notion that GABAergic SHP innervation is decreased selectively in

13

hippocampal interneurons accumulating P-Tau in mice carrying the VLW mutations.

M AN U

SC

RI PT

1

14

GABAergic SH innervation is specifically diminished on PARV-positive cells in

16

VLW transgenic mice.

17

One of the main targets of GABAergic SH fibers is the PARV-positive population of

18

interneurons. These basket and axo-axonic PARV-containing cells are essential for

19

correct hippocampal function. In this regard, they are responsible for establishing

20

hippocampal synchronous rhythms, which are crucial for learning and memory (Cardin

21

et al., 2009; Freund and Buzsaki, 1996; Pike et al., 2000).

22

Since VLW mice have cognitive deficits (García-Cabrero et al., 2013; Rodríguez-

23

Navarro et al., 2008) and the above results showed a considerable accumulation of P-

24

Tau specifically in basket and axo-axonic PARV-positive interneurons, we next

25

combined the detection of BDA with PARV labeling in order to study whether P-Tau

26

accumulation in PARV-positive cells alters their GABAergic SH innervation specifically

27

in these hippocampal interneurons. As shown previously, our data indicated an early

28

accumulation of P-Tau forms in PARV-positive cells, so we next analyzed 2-mo VLW

29

mice. A reduced percentage of PARV-positive cells contacted by GABAergic SH fibers

30

was visible only in the DG of VLW mice (t(6)=4.28; P-value= 0.032, Student’s t

31

test)(Fig. 5G). The number of SH boutons per PARV-positive neuron was also

32

diminished in the CA3 in these animals (t(6)=2.49, P-value= 0.02, Student’s t test)(Fig.

33

5C,E and H).

AC C

EP

TE D

15

12

ACCEPTED MANUSCRIPT In 8-mo old mice, the pattern of SH innervation and the distribution of PARV-positive

2

interneurons were indistinguishable between WT and VLW mice (Fig. 5A,B). A

3

considerable decrease in the percentage of contacted PARV-positive cells (t(8)=2.53,

4

P-values=0.016, Student’s t test) and in the complexity of GABAergic SH contacts was

5

observed when considering all hippocampal regions of 8-mo VLW mice, compared to

6

control mice and (t(8)=2.31, P-value=0.02, Student’s t test)(Hipp, Fig. 5I and Fig. 5J,

7

respectively). Although there were fewer PARV-positive cells innervated by GABAergic

8

SHP axons in all hippocampal regions, the most severe effect was present in the DG

9

(t(8)= 4.28, P-values=0.0002, Student’s t test) and CA1 (t(8)=3.14, P=0.003

10

respectively, Student’s t test)(Fig. 5I). The reduction in the number of boutons per

11

PARV-positive cell affected mainly the CA1 and CA3 (t(8)= 2.75 and 2.62 respectively,

12

P-value=0.01, Student’s t test)(Fig. 5J).

13

Next, we analyzed the progression of GABAergic SH innervation during maturation,

14

comparing GABAergic connection in 2- and 8-mo WT and VLW mice. As previously

15

described (Rubio et al., 2012), no significant changes were observed in the percentage

16

and complexity of GABAergic SH contacts on PARV-positive cells in any hippocampal

17

region, except in CA1(Suppl. Fig. 2A,B), when comparing 2- and 8-mo WT mice. The

18

percentage of PARV-positive neurons innervated by GABAergic SH fibers was

19

significantly decreased in the CA1 of 8-mo WT mice (t(8)=2.47, P=0.02, Student’s t

20

test). Subsequently, we analyzed VLW mice. Although the complexity of GABAergic

21

SH contacts on PARV-positive cells remained constant in 2- and 8-mo VLW mice

22

(Suppl. Fig. 2D), a considerable decrease in the percentage of innervated PARV-

23

positive neurons was observed (Suppl. Fig. 2C). Our data showed a significant

24

diminution in the percentage of contacted PARV-positive neurons in the DG, CA1, and

25

the total hippocampus (t(8)=2.64, P=0.013; t(8)=0.29, P=0.006; t(8)=2.02, P=0.046

26

respectively, Student’s t test). These data indicate an accelerated decrease in

27

GABAergic SH innervation in VLW mice compared to WT mice.

28

We thus conclude that GABAergic SHP innervation is reduced early and specifically on

29

PARV-positive interneurons in VLW mice accumulating P-Tau.

AC C

EP

TE D

M AN U

SC

RI PT

1

30 31 32

No major loss of GABAergic SH neurons in the MS/DB

33

To explore whether the decrease of GABAergic SH innervation in VLW mice is related

34

to neuronal loss in the MS/DB complex, we counted GABAergic PARV-positive SH 13

ACCEPTED MANUSCRIPT neurons (Freund, 1989) in the septum. Statistical analyses indicated that the number of

2

GABAergic SH neurons remained constant in 8-mo VLW mice (average 124.9±14,21

3

cells/section) compared to WT age-matched controls (121.69±5.6 cells/section;

4

U(8)=0.229, P-value= 0.82, Mann-Whitney test)(Fig. 6A,B). These data indicate that

5

the reduction in GABAergic SH innervation in VLW mice is not related to neuronal loss

6

in the MS/DB, but rather reflects decreased axonal and perisomatic basket complexity.

RI PT

1

7 8 9

SC

10

AC C

EP

TE D

M AN U

11

14

ACCEPTED MANUSCRIPT DISCUSSION

2

The neuropathological hallmarks of AD are amyloid-β deposits and the aggregation of

3

P-Tau protein, accompanied by a progressive deterioration of the cholinergic and

4

GABAergic component of the basal forebrain system (Bloom, 2014; Castellani et al.,

5

2010; Rubio et al., 2012; Ypsilanti et al., 2008). Electrophysiological and biochemical

6

data obtained in transgenic mouse models suggest that AD does not involve only

7

neuronal and synaptic degeneration but rather causes aberrant network activity that

8

interferes with cognitive functions. In addition, AD is associated with an increase in the

9

incidence of epileptic seizures (García-Cabrero et al., 2013; Palop and Mucke, 2010,

10

2009; Verret et al., 2012), which may be responsible for alterations of neural network

11

activity. In this regard, some studies point to GABAergic neuron dysfunction as the main

12

factor responsible for network alterations associated with cognitive deficits in AD and

13

aging (Andrews-Zwilling et al., 2012, 2010; McQuail et al., 2015; Palop and Mucke,

14

2010; Palop et al., 2007). The GABAergic SHP regulates hippocampal rhythmic activity

15

by selective innervation on GABAergic hippocampal interneurons (Buzsaki, 2002;

16

Hangya et al., 2009), which in turn control large number of pyramidal and granular

17

neurons (Freund and Buzsaki, 1996; Freund and Katona, 2007).

18

We recently found that amyloid-β accumulation in J20 mice induces abnormal

19

GABAergic SH innervation that correlates with altered patterns of neuronal activity in

20

the hippocampus and with internal processes related to operant rewards (Rubio et al.,

21

2012; Vega-Flores et al., 2014). In the present study, we sought to examine whether the

22

second AD-associated pathological hallmark, hyperphosphorylation of Tau protein, is

23

also associated with alterations in the GABAergic SHP innervating hippocampal

24

interneurons. First, we demonstrate an accumulation of AT-180-positive P-Tau in

25

pyramidal neurons in the CA1-3 regions, in the mossy cells in the DG, and also in

26

basket and axo-axonic PARV-positive hippocampal interneurons of VLW mice. To the

27

best of our knowledge, this study is the first to describe P-Tau accumulation in PARV-

28

positive interneurons and mossy cells. It has been shown that PARV-positive

29

interneurons display remarkably extensive axonal arborizations, which are believed to

30

synchronize the firing of hundreds of hippocampal pyramidal neurons (Freund and

31

Buzsaki, 1996; Matyas et al., 2004). Interestingly, Thr231P-Tau accumulation was

32

restricted to PARV-positive interneurons (with no expression in CALB-, CALR- or SOM-

33

positive interneurons), thereby suggesting that this process selectively affects the

34

interneurons responsible, to a large extent, for the mechanisms that drive synchronous

AC C

EP

TE D

M AN U

SC

RI PT

1

15

ACCEPTED MANUSCRIPT circuits in the hippocampus (Cardin et al., 2009; Korotkova et al., 2010; Mann and

2

Paulsen, 2005; Sohal et al., 2009). Furthermore, in contrast to hippocampal pyramidal

3

neurons, which present P-Tau recognized by AT-180, AT-8 and 12E8 antibodies,

4

PARV-positive interneurons accumulated P-Tau specifically in Thr231. This observation

5

points to a distinct regulation of Tau phosphorylation in different hippocampal neuronal

6

types, probably linked to differential signaling pathways. The significance of such

7

differential phosphorylation remains to be elucidated.

8

Hippocampal hyperexcitability can lead to the degeneration of GABAergic septal

9

neurons (Garrido Sanabria et al., 2006). In order to analyze a possible toxic effect of P-

10

Tau accumulation in PARV-immunoreactive cells on either septal neurons or

11

hippocampal interneurons, we quantified the density of PARV-positive cells in the

12

hippocampus and in the septum (Figs. 3 and 6). Our data indicate no neuronal loss as a

13

direct effect of P-Tau accumulation. Accordingly, the hyperexcitability described in VLW

14

mice (García-Cabrero et al., 2013) does not appear to be a consequence of

15

hippocampal interneuron death.

16

GABAergic hippocampal neurons are the specific targets of the GABAergic SH

17

connection involved in the establishment of theta and gamma rhythms in the

18

hippocampus

19

Hangya et al., 2009). In addition, the GABAergic SH connection has an

20

electrophysiological function that is indispensable for spatial learning and memory

21

(Düzel et al., 2010; Fries, 2015; Gangadharan et al., 2016). In this regard, previous

22

studies report cognitive deficiencies caused by septal GABAergic neuron degeneration

23

(Dwyer et al., 2007; Köppen et al., 2013; Pang et al., 2011; Roland et al., 2014). Our

24

previous data indicate a dramatic reduction of GABAergic SH innervation in a mouse

25

AD model accumulating amyloid-β (Rubio et al., 2012). In addition, a decrease in

26

GABAergic SH activity and memory deficits associated with theta rhythms have been

27

reported in response to amyloid-β pathology (Rubio et al., 2012; Villette et al., 2011). In

28

order to determine whether P-Tau accumulation affects the pattern of GABAergic SH

29

innervation, we subjected VLW mice to tracer injections. Our data show a slight

30

reduction in GABAergic SH innervation on GAD65/67-positive neurons and a marked

31

decrease in the complexity of GABAergic SH contacts, specifically on P-Tau-positive

32

interneurons. This observation is reinforced by a considerable reduction of GABAergic

33

SH innervation (both percentage of cells and density of innervation) in PARV-positive

34

interneurons, which accumulate P-Tau in their somatodendritic domain. The reduction

M AN U

SC

RI PT

1

AC C

EP

TE D

(Buzsaki, 2002; Freund and Antal, 1988; Gangadharan et al., 2016;

16

ACCEPTED MANUSCRIPT of GABAergic SH innervation on PARV-positive neurons was detected early in the DG

2

and CA3 of 2-mo VLW mice, thereby suggesting a possible effect of P-Tau

3

accumulation during the development of the GABAergic SH connection in these areas.

4

The alterations were more obvious at 8 months, and the differences between 2- and 8-

5

mo VLW mice were more dramatic than in WT mice. These observations therefore also

6

indicate early synaptic degeneration of GABAergic SH contacts on PARV-positive

7

neurons in VLW mice.

8

P-Tau accumulation in neurons affects dendritic spines, synapses and basic processes

9

such as axonal transport (Ballatore et al., 2007; Frost et al., 2014; Hernández and Avila,

10

2007; Ittner et al., 2010). It has been reported that basket and axo-axonic interneurons

11

are crucial for rhythmic hippocampal activity and that alterations in these interneurons in

12

humans correlate with hyperexcitability and epilepsy (DeFelipe, 1999; Freund and

13

Katona, 2007). Together, our data indicate a central role of PARV-positive interneurons

14

in P-Tau pathology and suggest that the hyperexcitability and cognitive alterations

15

described in VLW mice are a consequence of a PARV-positive interneuron dysfunction

16

caused by either P-Tau accumulation, the loss of GABAergic SH input on these cells, or

17

a combination of both. Tau mediates amyloid-β toxicity in dendrites in AD. Moreover,

18

Tau-reduction ameliorates susceptibility to excytotoxic seizures and cognitive deficits

19

induced by amyloid-β (Crimins et al., 2013; Ittner et al., 2010; Roberson et al., 2007). All

20

these results support the notion that Tau makes a critical contribution to the synaptic

21

alterations associated with AD. Accordingly, our present data show impairment in

22

GABAergic SH synapses in a model of tauopathy, similar to the alteration in the same

23

connection associated with amyloid-β accumulation in J20 mice. This observation

24

suggests that P-Tau may be the common alteration responsible for GABAergic synaptic

25

impairment, and, as a consequence, of the altered hippocampal rhythmical activity and

26

cognitive deficiencies associated with AD. Additional physiological animal models

27

combining amyloid-β accumulation with the phosphorylation of Tau protein, or mice

28

expressing human mutated Tau under Tau endogenous promoter, or histological

29

analysis of samples from AD patients will shed light on the role of Tau protein in

30

GABAergic innervation in AD.

AC C

EP

TE D

M AN U

SC

RI PT

1

31 32

Our data, together with our previous results, demonstrate that both amyloid-β and P-

33

Tau accumulation impair GABAergic SH innervation. We propose that the loss of

34

GABAergic SH input contributes to the altered patterns of synchronous activity and the 17

ACCEPTED MANUSCRIPT 1

cognitive deficits associated with early manifestations of AD. In addition, our findings

2

suggest that a PARV-positive interneuron dysfunction, caused by P-Tau accumulation,

3

combined with a decrease in GABAergic SH innervation on these neurons, underlies

4

the cognitive deficits present in tauopathies, including AD.

5

RI PT

6

AC C

EP

TE D

M AN U

SC

7

18

ACCEPTED MANUSCRIPT FIGURE LEGENDS

2

Figure 1. Distribution pattern of P-Tau in VLW mice. (A and B) P-Tau expression is

3

present in the hippocampus of 8-month-old (mo) (A) and 2-mo (B) VLW mice. (C) No P-

4

Tau expression is observed in the medial septum (MS) or Diagonal Band of Broca (DB)

5

nuclei of the septal region in 8-mo VLW mice. (D and E) P-Tau expression in the CA1 of

6

8-mo (D) and 2-mo (E) VLW mice. (F) No P-Tau expression was observed in the CA1 of

7

8-mo WT mice. Arrows indicate P-Tau expression in the stratum oriens (so). (G) Mossy

8

cells showing P-Tau accumulation in the dentate gyrus (DG) of 8-mo VLW mice.

9

Abbreviations: sp: stratum pyramidale, sr: stratum radiatum, h: hilus, sg: stratum

10

granulosum, sm: stratum moleculare. Scale bars: (in C) 500 µm applies to A and B,

11

1000 µm applies to C; (in G) 50 µm applies to D, E and F, and 100 µm applies to G.

SC

RI PT

1

12

Figure 2. PARV-positive interneurons accumulate P-Tau in VLW mice.

14

(A-F and J-L) Colocalization of P-Tau (labeled in green) and PARV-positive (labeled in

15

red) interneurons in the CA1 (A-C) and DG (D-F) of an 8-mo VLW mouse and in the DG

16

of a 2-mo VLW mouse (J-L). No colocalization of P-Tau (green) with CALR (red) is

17

present in the CA1 of 8-mo VLW mice (G-I) Abbreviations: so: stratum oriens, sp:

18

stratum pyramidale, sr: stratum radiatum, h: hilus, sg: stratum granulosum, sm: stratum

19

moleculare. Scale bar: (in C) 100 µm applies to A-L.

TE D

M AN U

13

20

Figure 3. No loss of GABAergic hippocampal interneurons in VLW mice. (A and B)

22

GABAergic hippocampal interneurons in the CA1 of an 8-mo WT (A) and VLW (B)

23

mouse. (D and E) PARV-positive hippocampal interneurons in CA1 region of an 8-mo

24

WT (D) and VLW (E) mouse. (C and F) The number of GAD-positive (C) and PARV-

25

positive interneurons (F) per section was quantified in 2-mo and 8-mo WT mice (gray)

26

and compared to 2-mo and 8-mo VLW mice (white). Abbreviations: so: stratum oriens,

27

sp: stratum pyramidale, sr: stratum radiatum, HIPP: total hippocampus. Scale bar: (in B)

28

200 µm applies to A-D.

AC C

EP

21

29 30

Figure 4. Reduction of GABAergic SH innervation on GAD-positive interneurons

31

accumulating P-Tau. GABAergic innervation on GAD-positive hippocampal

32

interneurons. (A and B) GABAergic SH fibers (in black) contacting GAD-positive

33

interneurons (in brown) are observed in the CA3 of an 8-mo WT (A) and VLW (B)

34

mouse. (C-F) GAD65/67-positive somata of interneurons (in brown) enveloped with 19

ACCEPTED MANUSCRIPT many GABAergic SH synaptic boutons (arrow heads, in black) of a 2-mo WT (C) and

2

VLW (D) mouse. Percentage of GAD-positive interneurons contacted by GABAergic SH

3

fibers (E) and complexity of GABAergic SH contacts (number of boutons per GAD-

4

positive interneuron) (F), comparing 2-mo VLW (in white) and WT mice (in gray). (G-N)

5

GAD65/67-positive somata of interneurons (in brown, G, H,K and L) and P-Tau-positive

6

interneurons (in brown, I and M) enveloped with many GABAergic SH synaptic boutons

7

(in black) of an 8-mo WT (G and K) and VLW mice (H,I,L and M). Percentage of GAD-

8

positive interneurons contacted by GABAergic SH fibers (in gray and white, J) and

9

complexity of GABAergic SH contacts (number of boutons per GAD-positive

RI PT

1

interneuron) (in gray and white, N), comparing WT (in gray), VLW (in white) and also

11

the percentage and complexity of GABAergic SH contacts on P-Tau-positive

12

interneurons (in black, J and N, respectively). **P= 0.0016 *** P=0.00015.

13

Abbreviations: so: stratum oriens/ sp: stratum pyramidale/ sr: stratum radiatum. Scale

14

bars: (in B) 200 µm applies to A and B; (in D) 25 µm applies to C,D,K,L,M; (in I) 25 µm

15

applies to G,H and I.

M AN U

SC

10

16

Figure 5. GABAergic SH innervation on PARV-positive cells is reduced in VLW

18

mice. (A and B) GABAergic SH fibers (in black) contacting PARV-positive interneurons

19

(in brown) are observed in the CA3 of an 8-mo WT (A) and VLW (B) mouse. (C-F)

20

GABAergic SH basket with some synaptic contacts around the soma of PARV-positive

21

interneurons in a 2-mo (C) and an 8-mo WT (D) mouse, and a 2-mo (E) and 8-mo VLW

22

(F) mouse. (G-J) The percentage of PARV-positive hippocampal interneurons contacted

23

by GABAergic afferents (G and I) and the number of boutons/target of these baskets (H

24

and J) were quantified in WT and VLW mice at 2-mo (G,H) and 8-mo (I,J). Data

25

represent mean for each experimental group. All error bars are ± SEM.* P<0.05; **

26

P<0.01, *** P<0.0001. Abbreviations: so: stratum oriens, sp: stratum pyramidale, sr:

27

stratum radiatum, HIPP: total hippocampus. Scale bars: (in B) 200 µm applies to A and

28

B; (in F) 25 µm applies to C-F.

AC C

EP

TE D

17

29 30

Figure 6. No major loss of GABAergic neurons in the MS/DB complex. (A and B)

31

GABAergic SH neurons immunolabeled with PARV in MS/DB region of an 8-mo WT (A)

32

and VLW (B) mouse. No differences in the number of GABAergic septal neurons were

33

detected in VLW compared to WT mouse. Scale bar: (in B) 500 µm applies to A and B.

34 20

ACCEPTED MANUSCRIPT Supplementary Figure 1. No AT-8- or 12E8-positive signal was found in PARV-

2

positive cells in VLW mice. (A and B) 12E8-positive signal (A) and AT-8-labeling (B)

3

are present in the soma and apical dendrites of CA1 pyramidal neurons in 8-mo VLW

4

mice. (C,E,G) No colocalization of 12E8-positive P-Tau (in green, C) and PARV-positive

5

interneurons (in red, E) is present in VLW mice. (D,F,H) No AT-8-positive P-Tau

6

accumulation (in green) in PARV-positive hippocampal neurons is observed in VLW

7

animals. Abbreviations: so: stratum oriens, sp: stratum pyramidale, sr: stratum

8

radiatum. Scale bar: (in B) 100 µm applies also to A; (in H) 25 µm applies to C-H.

RI PT

1

9

Supplementary Figure 2. GABAergic SH input on PARV-positive interneurons

11

during hippocampal maturation. The percentage of PARV-positive hippocampal

12

interneurons contacted by GABAergic afferents (A, C) and the number of

13

boutons/targets of these baskets (B,D) were quantified, comparing 2-mo and 8-mo WT

14

(A,B) and VLW (C,D) mice. Data represent mean for each experimental group. All error

15

bars are ± SEM.* P<0.05; ** P<0.01

16 17

ACKNOWLEDGEMENTS

18

The authors thank the personnel of the Advanced Digital Microscopy Facilities at the

19

Institute for Research in Biomedicine for support. We are grateful to D. E. Parruca da

20

Cruz for his participation in some preliminary experiments of this study. This work was

21

supported by funds from the Ministry of Economy and Competitiveness of Spain

22

(SAF2013-42455-R) awarded to E. Soriano.

TE D

REFERENCES

EP

23 24

M AN U

SC

10

Andrews-Zwilling, Y., Bien-Ly, N., Xu, Q., Li, G., Bernardo, A., Yoon, S.Y., Zwilling,

26

D., Yan, T.X., Chen, L., Huang, Y., 2010. Apolipoprotein E4 Causes Age- and

27 28 29 30

AC C

25

Tau-Dependent Impairment of GABAergic Interneurons, Leading to Learning and Memory Deficits in Mice. Journal of Neuroscience 30, 13707–13717. doi:10.1523/JNEUROSCI.4040-10.2010

Andrews-Zwilling, Y., Gillespie, A.K., Kravitz, A. V., Nelson, A.B., Devidze, N., Lo, I.,

31

Yoon, S.Y., Bien-Ly, N., Ring, K., Zwilling, D., Potter, G.B., Rubenstein, J.L.R.,

32

Kreitzer, A.C., Huang, Y., 2012. Hilar GABAergic Interneuron Activity Controls

33

Spatial Learning and Memory Retrieval. PLoS ONE 7, e40555.

34

doi:10.1371/journal.pone.0040555 21

ACCEPTED MANUSCRIPT 1

Ballatore, C., Lee, V.M.-Y., Trojanowski, J.Q., 2007. Tau-mediated

2

neurodegeneration in Alzheimer’s disease and related disorders. Nature

3

reviews Neuroscience 8, 663–72. doi:10.1038/nrn2194

4

Bland, B.H., Declerck, S., Jackson, J., Glasgow, S., Oddie, S., 2007. Septohippocampal properties of N-methyl-D-aspartate-induced theta-band

6

oscillation and synchrony. Synapse 61, 185–197.

8 9

Bloom, G.S., 2014. Amyloid-β and tau: the trigger and bullet in Alzheimer disease pathogenesis. JAMA neurology 71, 505–8. doi:10.1001/jamaneurol.2013.5847 Buzsaki, G., 2002. Theta oscillations in the hippocampus. Neuron 33, 325–340.

SC

7

RI PT

5

Buzsáki, G., 2005. Theta rhythm of navigation: Link between path integration and

11

landmark navigation, episodic and semantic memory. Hippocampus 15, 827–

12

840.

M AN U

10

13

Cardin, J.A., Carlen, M., Meletis, K., Knoblich, U., Zhang, F., Deisseroth, K., Tsai,

14

L.H., Moore, C.I., 2009. Driving fast-spiking cells induces gamma rhythm and

15

controls sensory responses. Nature 459, 663–667.

17

Castellani, R.J., Rolston, R.K., Smith, M.A., 2010. Alzheimer Disease. Disease-a-

TE D

16

month : DM 56, 484–546. doi:10.1016/j.disamonth.2010.06.001 Crimins, J.L., Pooler, A., Polydoro, M., Luebke, J.I., Spires-Jones, T.L., 2013. The

19

intersection of amyloid beta and tau in glutamatergic synaptic dysfunction and

20

collapse in Alzheimer’s disease. Ageing Research Reviews 12, 757–763.

21

doi:10.1016/j.arr.2013.03.002

23 24 25 26

DeFelipe, J., 1999. Chandelier cells and epilepsy. Brain : a journal of neurology 122

AC C

22

EP

18

( Pt 1, 1807–22.

Düzel, E., Penny, W.D., Burgess, N., 2010. Brain oscillations and memory. Current opinion in neurobiology 20, 143–9. doi:10.1016/j.conb.2010.01.004

Dwyer, T.A., Servatius, R.J., Pang, K.C., 2007. Noncholinergic lesions of the medial

27

septum impair sequential learning of different spatial locations. J Neurosci 27,

28

299–303.

29 30

Freund, T.F., Antal, M., 1988. GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature 336, 170–173. 22

ACCEPTED MANUSCRIPT

3 4 5 6 7 8 9

347–470. Freund, T.F., Gulyas, A.I., 1997. Inhibitory control of GABAergic interneurons in the hippocampus. Can J Physiol Pharmacol 75, 479–487. Freund, T.F., Katona, I., 2007. Perisomatic inhibition. Neuron 56, 33–42. doi:10.1016/j.neuron.2007.09.012

RI PT

2

Freund, T.F., Buzsaki, G., 1996. Interneurons of the hippocampus. Hippocampus 6,

Fries, P., 2015. Rhythms for Cognition: Communication through Coherence. Neuron 88, 220–235. doi:10.1016/j.neuron.2015.09.034

Frost, B., Götz, J., Feany, M.B., 2014. Connecting the dots between tau dysfunction

SC

1

and neurodegeneration. Trends in cell biology 25, 46–53.

11

doi:10.1016/j.tcb.2014.07.005

12

M AN U

10

Gangadharan, G., Shin, J., Kim, S.-W., Kim, A., Paydar, A., Kim, D.-S., Miyazaki, T., Watanabe, M., Yanagawa, Y., Kim, J., Kim, Y.-S., Kim, D., Shin, H.-S.,

14

2016. Medial septal GABAergic projection neurons promote object exploration

15

behavior and type 2 theta rhythm. Proceedings of the National Academy of

16

Sciences of the United States of America 113, 6550–5.

17

doi:10.1073/pnas.1605019113

18

TE D

13

García-Cabrero, A.M., Guerrero-López, R., Giráldez, B.G., Llorens-Martín, M., Ávila, J., Serratosa, J.M., Sánchez, M.P., 2013. Hyperexcitability and epileptic

20

seizures in a model of frontotemporal dementia. Neurobiology of Disease 58,

21

200–208. doi:10.1016/j.nbd.2013.06.005

23 24 25

Garner, H.L., Whittington, M.A., Henderson, Z., 2005. Induction by kainate of theta

AC C

22

EP

19

frequency rhythmic activity in the rat medial septum-diagonal band complex in vitro. J Physiol 564, 83–102.

Garrido Sanabria, E.R., Castaneda, M.T., Banuelos, C., Perez-Cordova, M.G.,

26

Hernandez, S., Colom, L. V, 2006. Septal GABAergic neurons are selectively

27

vulnerable to pilocarpine-induced status epilepticus and chronic spontaneous

28

seizures. Neuroscience 142, 871–883.

29

Goedert, M., Jakes, R., Crowther, R. a, Cohen, P., Vanmechelen, E.,

30

Vandermeeren, M., Cras, P., 1994. Epitope mapping of monoclonal antibodies

31

to the paired helical filaments of Alzheimer’s disease: identification of 23

ACCEPTED MANUSCRIPT 1

phosphorylation sites in tau protein. The Biochemical journal 301 ( Pt 3, 871–

2

877.

3

Gulyas, A.I., Gorcs, T.J., Freund, T.F., 1990. Innervation of different peptide-

4

containing neurons in the hippocampus by GABAergic septal afferents.

5

Neuroscience 37, 31–44. Gulyas, A.I., Hajos, N., Freund, T.F., 1996. Interneurons containing calretinin are

RI PT

6 7

specialized to control other interneurons in the rat hippocampus. J Neurosci 16,

8

3397–3411.

Hangya, B., Borhegyi, Z., Szilagyi, N., Freund, T.F., Varga, V., 2009. GABAergic

SC

9

neurons of the medial septum lead the hippocampal network during theta

11

activity. J Neurosci 29, 8094–8102.

12 13 14

M AN U

10

Hernández, F., Avila, J., 2007. Tauopathies. Cellular and molecular life sciences : CMLS 64, 2219–33. doi:10.1007/s00018-007-7220-x Ittner, L.M., Ke, Y.D., Delerue, F., Bi, M., Gladbach, A., van Eersel, J., Wölfing, H., Chieng, B.C., Christie, M.J., Napier, I.A., Eckert, A., Staufenbiel, M.,

16

Hardeman, E., Götz, J., 2010. Dendritic function of tau mediates amyloid-beta

17

toxicity in Alzheimer’s disease mouse models. Cell 142, 387–97.

18

doi:10.1016/j.cell.2010.06.036

19

TE D

15

Kitchigina, V., Popova, I., Sinelnikova, V., Malkov, A., Astasheva, E., Shubina, L., Aliev, R., 2013. Disturbances of septohippocampal theta oscillations in the

21

epileptic brain: reasons and consequences. Experimental neurology 247, 314–

22

27. doi:10.1016/j.expneurol.2013.01.029

AC C

EP

20

23

Köppen, J.R., Winter, S.S., Stuebing, S.L., Cheatwood, J.L., Wallace, D.G., 2013.

24

Infusion of GAT1-saporin into the medial septum/vertical limb of the diagonal

25 26 27

band disrupts self-movement cue processing and spares mnemonic function. Brain structure & function 218, 1099–114. doi:10.1007/s00429-012-0449-7

Korotkova, T., Fuchs, E.C., Ponomarenko, A., von Engelhardt, J., Monyer, H., 2010.

28

NMDA receptor ablation on parvalbumin-positive interneurons impairs

29

hippocampal synchrony, spatial representations, and working memory. Neuron

30

68, 557–69. doi:10.1016/j.neuron.2010.09.017

31

Lim, F., Hernández, F., Lucas, J.J., Gómez-Ramos, P., Morán, M. a, Avila, J., 2001. 24

ACCEPTED MANUSCRIPT 1

FTDP-17 mutations in tau transgenic mice provoke lysosomal abnormalities

2

and Tau filaments in forebrain. Molecular and cellular neurosciences 18, 702–

3

714. doi:10.1006/mcne.2001.1051

4

Loreth, D., Ozmen, L., Revel, F.G., Knoflach, F., Wetzel, P., Frotscher, M., Metzger, F., Kretz, O., 2012. Selective degeneration of septal and hippocampal

6

GABAergic neurons in a mouse model of amyloidosis and tauopathy.

7

Neurobiology of Disease 47, 1–12. doi:10.1016/j.nbd.2012.03.011

RI PT

5

Mann, E.O., Paulsen, O., 2005. Mechanisms underlying gamma ([`]40 Hz’) network

9

oscillations in the hippocampus--a mini-review. Progress in Biophysics and

10 11

Molecular Biology 87, 67–76.

SC

8

Matyas, F., Freund, T.F., Gulyas, A.I., 2004. Immunocytochemically defined interneuron populations in the hippocampus of mouse strains used in

13

transgenic technology. Hippocampus 14, 460–481.

14

M AN U

12

McQuail, J.A., Frazier, C.J., Bizon, J.L., 2015. Molecular aspects of age-related

15

cognitive decline: the role of GABA signaling. Trends in molecular medicine 21,

16

450–60. doi:10.1016/j.molmed.2015.05.002

Mucke, L., Masliah, E., Yu, G.Q., Mallory, M., Rockenstein, E.M., Tatsuno, G., Hu,

18

K., Kholodenko, D., Johnson-Wood, K., McConlogue, L., 2000. High-level

19

neuronal expression of abeta 1-42 in wild-type human amyloid protein

20

precursor transgenic mice: synaptotoxicity without plaque formation. J Neurosci

21

20, 4050–4058.

23 24 25 26

EP

Palop, J.J., Chin, J., Roberson, E.D., Wang, J., Thwin, M.T., Bien-Ly, N., Yoo, J., Ho, K.O., Yu, G.Q., Kreitzer, A., Finkbeiner, S., Noebels, J.L., Mucke, L., 2007.

AC C

22

TE D

17

Aberrant excitatory neuronal activity and compensatory remodeling of inhibitory hippocampal circuits in mouse models of Alzheimer’s disease. Neuron 55, 697–711.

27

Palop, J.J., Jones, B., Kekonius, L., Chin, J., Yu, G.Q., Raber, J., Masliah, E.,

28

Mucke, L., 2003. Neuronal depletion of calcium-dependent proteins in the

29

dentate gyrus is tightly linked to Alzheimer’s disease-related cognitive deficits.

30

Proc Natl Acad Sci U S A 100, 9572–9577.

31

Palop, J.J., Mucke, L., 2010. Amyloid-[beta]-induced neuronal dysfunction in 25

ACCEPTED MANUSCRIPT 1

Alzheimer’s disease: from synapses toward neural networks. Nat Neurosci 13,

2

812–818.

3 4 5

Palop, J.J., Mucke, L., 2009. Epilepsy and cognitive impairments in Alzheimer disease. Archives of neurology 66, 435–40. doi:10.1001/archneurol.2009.15 Pang, K.C.H., Jiao, X., Sinha, S., Beck, K.D., Servatius, R.J., 2011. Damage of GABAergic neurons in the medial septum impairs spatial working memory and

7

extinction of active avoidance: effects on proactive interference. Hippocampus

8

21, 835–46. doi:10.1002/hipo.20799

RI PT

6

Pascual, M., Perez-Sust, P., Soriano, E., 2004. The GABAergic septohippocampal

10

pathway in control and reeler mice: target specificity and termination onto

11

reelin-expressing interneurons. Molecular and Cellular Neuroscience 25, 679–

12

691. doi:10.1016/j.mcn.2003.12.009

M AN U

13

SC

9

Pascual, M., Pozas, E., Soriano, E., 2005. Role of class 3 semaphorins in the

14

development and maturation of the septohippocampal pathway. Hippocampus

15

15, 184–202. doi:10.1002/hipo.20040

17 18

Paxinos, G., Franklin, K.B.J., 2001. The mouse brain in stereotaxic coordinates, 2nd ed. Academic, San Diego, Calif. London.

TE D

16

Pike, F.G., Goddard, R.S., Suckling, J.M., Ganter, P., Kasthuri, N., Paulsen, O., 2000. Distinct frequency preferences of different types of rat hippocampal

20

neurones in response to oscillatory input currents. J Physiol 529, 205–213.

22 23 24 25

Roberson, E.D., Scearce-Levie, K., Palop, J.J., Yan, F., Cheng, I.H., Wu, T., Gerstein, H., Yu, G.Q., Mucke, L., 2007. Reducing endogenous tau ameliorates

AC C

21

EP

19

amyloid beta-induced deficits in an Alzheimer’s disease mouse model. Science 316, 750–754.

Rocamora, N., Pascual, M., Acsady, L., deLecea, L., Freund, T.F., Soriano, E.,

26

1996. Expression of NGF and NT3 mRNAs in hippocampal interneurons

27

innervated by the GABAergic septohippocampal pathway. Journal of

28

Neuroscience 16, 3991–4004.

29

Rodríguez-Navarro, J. a, Gómez, A., Rodal, I., Perucho, J., Martinez, A., Furió, V.,

30

Ampuero, I., Casarejos, M.J., Solano, R.M., de Yébenes, J.G., Mena, M. a,

31

2008. Parkin deletion causes cerebral and systemic amyloidosis in human 26

ACCEPTED MANUSCRIPT 1

mutated tau over-expressing mice. Human molecular genetics 17, 3128–43.

2

doi:10.1093/hmg/ddn210

3

Roland, J.J., Stewart, A.L., Janke, K.L., Gielow, M.R., Kostek, J.A., Savage, L.M., Servatius, R.J., Pang, K.C.H., 2014. Medial septum-diagonal band of Broca

5

(MSDB) GABAergic regulation of hippocampal acetylcholine efflux is

6

dependent on cognitive demands. The Journal of neuroscience : the official

7

journal of the Society for Neuroscience 34, 506–14.

8

doi:10.1523/JNEUROSCI.2352-13.2014

9

RI PT

4

Rubio, S.E., Vega-Flores, G., Martínez, A., Bosch, C., Pérez-Mediavilla, A., del Río, J., Gruart, A., Delgado-García, J.M., Soriano, E., Pascual, M., 2012.

11

Accelerated aging of the GABAergic septohippocampal pathway and

12

decreased hippocampal rhythms in a mouse model of Alzheimer’s disease.

13

FASEB journal : official publication of the Federation of American Societies for

14

Experimental Biology 26, 4458–67. doi:10.1096/fj.12-208413

17

M AN U

16

Selkoe, D.J., 1991. The Molecular Pathology of Alzheimer’s Disease. Neuron 6, 487–498. doi:10.1016/0896-6273(91)90052-2

Sohal, V.S., Zhang, F., Yizhar, O., Deisseroth, K., 2009. Parvalbumin neurons and

TE D

15

SC

10

18

gamma rhythms enhance cortical circuit performance. Nature 459, 698–702.

19

doi:10.1038/nature07991

20

Sotty, F., Danik, M., Manseau, F., Laplante, F., Quirion, R., Williams, S., 2003. Distinct electrophysiological properties of glutamatergic, cholinergic and

22

GABAergic rat septohippocampal neurons: novel implications for hippocampal

23

rhythmicity. J Physiol 551, 927–943.

AC C

EP

21

24

Spires, T.L., Orne, J.D., SantaCruz, K., Pitstick, R., Carlson, G. a, Ashe, K.H.,

25

Hyman, B.T., 2006. Region-specific dissociation of neuronal loss and

26 27

neurofibrillary pathology in a mouse model of tauopathy. The American journal of pathology 168, 1598–1607. doi:10.2353/ajpath.2006.050840

28

Toth, K., Freund, T.F., Miles, R., 1997. Disinhibition of rat hippocampal pyramidal

29

cells by GABAergic afferents from the septum. J Physiol 500 ( Pt 2, 463–474.

30

Vega-Flores, G., Rubio, S.E., Jurado-Parras, M.T., Gomez-Climent, M.A., Hampe,

31

C.S., Manto, M., Soriano, E., Pascual, M., Gruart, A., Delgado-Garcia, J.M., 27

ACCEPTED MANUSCRIPT 1

2014. The GABAergic septohippocampal pathway is directly involved in internal

2

processes related to operant reward learning. Cerebral cortex (New York, NY :

3

1991) 24, 2093–2107. doi:10.1093/cercor/bht060 Verret, L., Mann, E.O., Hang, G.B., Barth, A.M.I., Cobos, I., Ho, K., Devidze, N.,

5

Masliah, E., Kreitzer, A.C., Mody, I., Mucke, L., Palop, J.J., 2012. Inhibitory

6

interneuron deficit links altered network activity and cognitive dysfunction in

7

Alzheimer model. Cell 149, 708–21. doi:10.1016/j.cell.2012.02.046

8 9

RI PT

4

Villette, V., Poindessous-Jazat, F., Simon, A., Lena, C., Roullot, E., Bellessort, B., Epelbaum, J., Dutar, P., Stephan, A., 2011. Decreased rhythmic GABAergic septal activity and memory-associated theta oscillations after hippocampal

11

amyloid-beta pathology in the rat. J Neurosci 30, 10991–11003. Whitehouse, P.J., Price, D.L., Struble, R.G., Clark, A.W., Coyle, J.T., Delon, M.R.,

M AN U

12

SC

10

13

1982. Alzheimer’s disease and senile dementia: loss of neurons in the basal

14

forebrain. Science 215, 1237–1239.

15

Ypsilanti, A.R., Girao da Cruz, M.T., Burgess, A., Aubert, I., 2008. The length of hippocampal cholinergic fibers is reduced in the aging brain. Neurobiol Aging

17

29, 1666–1679.

EP AC C

18

TE D

16

28

A

8mo

C

8mo

CA1 ACCEPTED MANUSCRIPT

CA3

2mo

B CA1

RI PT

DG

MS

CA3

M AN U

DG

D

E

F

so

TE D

so

sp

sp

sr

EP

sr

2mo

AC C

8mo

SC

DB

8mo

WT

G sm

so

sg sp

h

sr

8mo

ACCEPTED MANUSCRIPT

A

P-TAU C

B

so

RI PT

sp sr

PARV

D

P-TAU F

PARV E

SC

8 mo

so

M AN U

sp sr

G

CALR H

P-TAU I

TE D

so sp

PARV K

AC C

2 mo

J

EP

sr

P-TAU L h sg sm

WT

A

VLW C

B

so sp

so sp

GAD-positive neurons/mm2

sr

sr

200 150 100

50 0

WT E

D

VLW F

AC C

EP

SC

TE D

M AN U

PARV

so sp sr

WT VLW

250

RI PT

GAD65/67

ACCEPTED MANUSCRIPT

so

2 mo

8 mo

PARV-positive neurons/mm 70

WT VLW

60 50

sp

40 30 20

sr

10 0

2 mo

8 mo

2

ACCEPTED MANUSCRIPT

WT B

8 mo

A

VLW

RI PT

sr

sr

sp

sp so

WT

C

D

VLW E

45 40

G

H

K

WT GAD WT L

EP

GAD

M AN U

GAD

GAD GAD

F

8 7

30

6

25

5

20

4

15

3

10

2

5

1

0

0

VLW I

VLW

VLW J

WT-GAD VLW-GAD

9

N. boutons/cell

% contacted cells

TE D

GAD

AC C

8 mo

2 mo

35

SC

so

WT-GAD VLW-GAD VLW-P-TAU

45 40 35 30 25 20 15 10 5 0

P-TAU

M

% contacted cells

VLW N

8 7 6 5 4 3 2 1

P-TAU

0

***

**

WT

VLW

ACCEPTED MANUSCRIPT

8 mo

A

8 mo

B

sr sr

sp so

RI PT

sp so

2 mo D

8 mo E

2 mo F

8 mo

60

% contacted PARV-positive cells

WT

*

VLW

2 mo

50 40 30

10 0

CA3

AC C

DG

I

60

CA1

10 8 6 4

0

HIPP

% contacted PARV-positive cells

***

WT VLW

*

2

WT VLW

50

8 mo

12

EP

20

Number of boutons / PARV-positive cells

H

TE D

G

M AN U

SC

C

DG

J

CA3

CA1

Number of boutons / PARV-positive cells

12

WT VLW

*

*

10

*

*

40

HIPP

8

**

30

6

20

4

10

2

0

0

DG

CA3

CA1

HIPP

DG

CA3

CA1

HIPP

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT . VLW mice show accumulation of P-Tau forms in PARV-positive interneurons. . No major loss of GABAergic hippocampal interneurons is observed in VLW mice. . GABAergic SH innervation is reduced on PARV-positive interneurons in VLW mice.

RI PT

. VLW mice present a decrease in GABAergic SH contacts on P-Tau-positive interneurons.

AC C

EP

TE D

M AN U

SC

. P-Tau accumulation in PARV-positive interneurons impairs the GABAergic SH connection