Enhanced synaptic plasticity and spatial memory in female but not male FLRT2-haplodeficient mice

The Fibronectin Leucine-Rich Transmembrane protein 2 (FLRT2) has been implicated in several hormone -and sex-dependent physiological and pathological processes (including chondrogenesis, menarche and breast cancer); is known to regulate developmental synapses formation, and is expressed in the hippocampus, a brain structure central for learning and memory. However, the role of FLRT2 in the adult hippocampus and its relevance in sex-dependent brain functions remains unknown. We here used adult single-allele FLRT2 knockout (FLRT2+/−) mice and behavioral, electrophysiological, and molecular/biological assays to examine the effects of FLRT2 haplodeficiency on synaptic plasticity and hippocampus-dependent learning and memory. Female and male FLRT2+/− mice presented morphological features (including body masses, brain shapes/weights, and brain macroscopic cytoarchitectonic organization), indistinguishable from their wild type counterparts. However, in vivo examinations unveiled enhanced hippocampus-dependent spatial memory recall in female FLRT2+/− animals, concomitant with augmented hippocampal synaptic plasticity and decreased levels of the glutamate transporter EAAT2 and beta estrogen receptors. In contrast, male FLRT2+/− animals exhibited deficient memory recall and decreased alpha estrogen receptor levels. These observations propose that FLRT2 can regulate memory functions in the adulthood in a sex-specific manner and might thus contribute to further research on the mechanisms linking sexual dimorphism and cognition.

Behavioral analysis of FLRT2 +/− mice shows sex-specific changes in memory recall. Before examining the effects of single-allele FLRT2 deletion on hippocampus-dependent learning and memory functions, we examined the impact of this mutation on locomotor activity by comparing the performance of adult male and female FLRT2 +/− mice to their sex-corresponding wild type littermate controls in the open field test (OFT). No significant main effect of sex or genotype, nor sex x genotype interactions in total distance traveled ( Fig. 2A) were observed indicating that having lower levels of FLRT2 +/− does not interfere with regular spontaneous locomotor activity. When average velocity during the test was measured (Fig. 2B), significant main effect of the sex and main effect of the genotype were observed with no significant interaction between the two factors. However, post hoc analysis showed no significant differences between groups. Additionally, motor coordination was examined by assessing the performances in the accelerating Rota-rod test of adult male and female mice. Compatible with observations from other authors 31 , a difference in performance was observed between male and female wild type animals. This pattern was also observed when male vs female FLRT2 +/− mice were compared, showing a significant main effect of sex on motor coordination (Fig. 2C). No significant differences were found between male wild type vs male FLRT2 +/− mice or between female wild type vs female FLRT2 +/− mice (Fig. 2C). We next examined the impact of the single-allele FLRT2 deletion on learning and memory employing the Morris Water Maze test (MWM), which is a standardized test widely used to examine the properties of hippocampus-dependent spatial learning and memory 32,33 . During all training days, there were no observable differences in the escape latencies of either male or female heterozygous FLRT2 +/− mice when compared to those of their corresponding wild type littermate controls (Fig. 2D). Taken together, these observations indicate that the FLRT2 single-allele deletion does not results in any deleterious impact either on gross motor performance or on the initial phases of learning acquisition. We additionally conducted a probe trial in order to examine spatial memory retention in heterozygous FLRT2 +/− mice. A two-way ANOVA revealed a sex x genotype interaction with male FLRT2 +/− mice spending less and female FLRT2 +/− spending more time in the target quadrant where the platform was placed during the training period compared to their respective wild-type littermates (Fig. 2E). These observations suggest a sex-dependent effect of FLRT2 haplodeficiency on memory recall. Examinations of swimming speeds during the performance in the water maze, in both male and female FLRT2 +/− mice, showed that both male and female heterozygous FLRT2 +/− mice presented with swimming speeds comparable to those of their sex-respective littermate wild type controls (Fig. 2F). These observations propose that FLRT2 might participate in vivo in hippocampus-mediated memory recall in a sex-specific manner.
Loss of a single FLRT2 allele enhances hippocampal synaptic plasticity in female but not male mice. To examine the potential implication of FLRT2 in the hippocampal synaptic function, we examined ex vivo the electrophysiological properties of the monosynaptic circuit established between the CA3 region and the pyramidal inputs of the CA1 region via the Schaffer collateral pathway, a circuit proposed to be implicated in the formation and maintenance of spatial memories 34 . To this aim, we first analyzed the properties of basal synaptic transmission using standard input/output protocols (Materials and Methods) in hippocampal slices obtained from male or female wild type controls and FLRT2 +/− mice. No differences in the voltage-dependence of synaptic transmission were detected between male FLRT2 +/− and their corresponding wild type controls and a right shift was observed in FLRT2 +/− female mice as compared to their related wild type female controls (Fig. 3A). Next, we examined the properties of long-term synaptic potentiation (LTP) in hippocampal slices from wild type controls and heterozygous FLRT2 +/− male and female mice. Recordings of extracellular field potentials and analysis of field-slopes indicated that whereas no significant differences were apparent between male FLRT2 +/− and male wild type mice, a significant enhancement in synaptic potentiation was observed in female FLRT2 +/− mice relative to their wild type female littermates (Fig. 3B,C).
Female but not male FLRT2+/− mice have reduced levels of EAAT2. Glutamatergic signaling plays a key role in the modulation of hippocampal synaptic plasticity and learning and memory 35 . Previous studies have also shown that 17β-estradiol (E2), the primary female sex hormone, can up-regulate both the mRNA and protein levels of the excitatory amino acid transporter 2 (EAAT2, known as GLT-1 in rodent literature) 36,37 . mice do not present with distinctive gross morphological abnormalities in body or brain structure. (B) Microphotographs of Nissl-stained coronal brain sections showed no obvious differences in shape and major cytoarchitectonic brain organization between wild type and FLRT2 +/− male or female mice. (C) Significant main effect of the sex was observed for body weights (p < 0.001, F (1,20) = 44.07) with a consistent pattern between wild type and FLRT2 +/− mice. (D) No significant differences were observed in brain weights (ns, p > 0.05). (E) FLRT2 +/− males and females presented with significantly lower levels of FLRT2 mRNA in the hippocampus as compared to their respective wild type littermate controls, thus exhibiting a main effect of the genotype (p < 0.0001, F (1,19)  Therefore, in an effort to address potential molecular signaling pathways that could underlie the sex-dependent memory phenotype observed in FLRT2 +/− mice, we examined the levels of the vesicular glutamate transporter 1 (VGLUT1) and of EAAT2 in FLRT2 +/− mice by conventional Western blot assay. Whereas no changes were observed for the levels of VGLUT1 in either male or female mice, FLRT2 +/− females exhibited significantly reduced levels of EAAT2 compared to their female counterpart, while no differences were observed between wild type and FLRT2 +/− males (Fig. 4A,B).
Sexually dimorphic levels of ERα and ERβ in FLRT2 haplodeficient mice. Estrogen receptor-mediated signaling plays a central role in the regulation of the adult brain function and has a critical impact on neuronal morphology, synaptic plasticity and learning and memory processes [38][39][40][41][42][43][44][45] . Given the sex-dependent differences in behavior and synaptic plasticity observed upon FLRT2 haplodeficiency, we also examined the levels of the Estrogen Receptor Alpha (ERα) and the Estrogen Receptor Beta (ERβ) in male and Behavioral analysis of FLRT2 +/− mice shows sex-specific changes in memory recall. Open field test did not reveal significant changes in locomotor activity in male or female FLRT2 +/− mice compared to their corresponding wild type littermates, as determined by measuring (A) total distance traveled. When the average velocity during the test was examined, however, a significant main effect of sex (p = 0.004, F (1,46) = 9.23) and a main effect of genotype (p = 0.009, F (1,46) = 7.72) was apparent as indicated in (B) with data showing no significance differences between groups after post hoc analysis. (C) Examinations in the accelerating Rotarod test showed a significant main effect of sex (p < 0.00001, F (1,50) = 20.01) with no main effect of genotype on motor coordination. (D) No significant differences in spatial learning were detected between either male or female FLRT2 +/− mice when compared to the their respective wild type littermate controls as determined by escape latencies during training days in the Morris water maze (p > 0.05, n = 8-14). However, there was a statistically significant interaction between the effects of sex and genotype on long-term memory retention (E) during probe trial (p = 0.0001, F (1,32)  female FLRT2 +/− mice hippocampi. Interestingly, while the levels of ERα were dramatically reduced only in male FLRT2 +/− mice (Fig. 4C), the levels of ERβ were significantly reduced only in female FLRT2 +/− mice compared to those of their corresponding littermate controls (Fig. 4D), suggesting that FLRT2 haplodeficiency has a differential impact on the levels of estrogen receptors depending on the sex.
Female FLRT2 haplodeficient mice exhibit altered hippocampal CA1 pyramidal neurons apical dendritic complexity. The morphological characteristics of neurites and dendritic spines (e.g. neuritic length, spine number and morphology) in hippocampal neurons play key roles in determining synaptic activity and memory-related plasticity functions 46,47 . Moreover, FLRT3, a member of the FLRT family of proteins also expressed in the hippocampus, has been implicated in the regulation of neurite outgrowth and synaptic development 20 . On these grounds, and given that our observations indicate an involvement of FLRT2 in hippocampal synaptic plasticity, we next explored whether the functional effects of FLRT2 haplodeficiency could be accompanied by morphological alterations in dendritic complexity. To this aim, we conducted Golgi-Cox staining of hippocampal CA1 pyramidal neurons obtained from wild type and FLRT2 +/− female mice (materials and Methods). Since our data indicated a possible sex-hormone-related activity of FLRT2, and as the effect of estradiol (perhaps the most important sex-related hormone in mice) on dendritic spines (of hippocampal CA1 pyramidal cells) is more pronounced on the apical tree 46 (see also 48 ), we therefore focused our analyses on this neuronal structure (Fig. 5). Whereas no significant differences in general, macroscopic appearance or in the area of the soma were detected ( Fig. 5A-C), neurons from FLRT2 +/− female mice presented with a significantly lower dendritic length; with a decreased number of ends and nodes ( Fig. 5D-F), as well as with a lower number of dendritic intersections (Fig. 5G). Interestingly, when dendritic spines were examined, neurons from FLRT2 +/− female exhibited a significant decrease in number of thin spines, whereas Stubby, Mushroom and Long Thin spines appear unaltered (Fig. 5H).

Discussion
While the identification of FLRT2 protein and its localization in the brain were described about eighteen years ago 19 , the functional roles of FLRT2 in the adult nervous system has remained virtually unknown. Data presented here derived from studies on FLRT2 haplodeficient (FLRT2 +/− ) mice points towards a potential functional involvement of FLRT2 in adult synaptic plasticity and memory recall and proposes FLRT2 as a molecular candidate implicated in the sex-dependent regulation of memory functions.
The use of FLRT2 +/− mice has precedents in the scientific literature, including studies addressing the functional relevance of FLRT2 in the morphogenesis of the developing heart 18 and studies on the role of FLRT proteins in the developing brain reporting that FLRT2 proteins act as critical signaling molecules regulating cortical neuron migration and brain neuronal layering 22 . In these previous developmental studies 18,22 , the single allele deletion of the FLRT2 gene resulted in a significant reduction of FLRT2 protein levels. Here we extended the studies on FLRT2 to both sexes and to adulthood. Both female and male FLRT2 +/− adult mice examined here reached normal life spans, had no observable external physical defects, and presented with significant reduction in the levels of FLRT2 mRNA. Female and male FLRT2 +/− mice showed also regular brain morphologies, no detectable brain macroscopic cytoarchitectonical defects, and no prominent motoric dysfunctions. However, upon examination of hippocampus-dependent learning and memory using the Morris water maze, a significant augmentation in memory recall in female FLRT2 +/− mice was observed, whereas FLRT2 +/− male mice presented  Different brain neuronal LRR proteins have been implicated in adult neuroplasticity, neurogenesis, and memory storage [49][50][51][52][53] . In addition, the association between LRR-containing proteins and sex hormones activity in the brain 54,55 , as well as in other organs [56][57][58] , has been also previously described. Moreover, FLRT2 has been implicated in processes differentially regulated between males and females such as prostate cancer 59 and chondrogenesis [60][61][62] and in several primarily female and hormone-related physiological and pathophysiological processes in humans, including menarche 63 and breast cancer 64,65 . In humans, different lines of evidence indicate that some of the differences in the cognitive function -and related pathologies-between males and females can be in part attributed to the differential regulation by sex-hormone signaling [66][67][68][69][70][71][72] . Studies in animal models, including mouse and rat, have also showed the importance of sex-related hormones and the sex-dependent differential effect of αand β-estrogen receptors deficiency on neuronal morphology, synaptic plasticity, and memory functions [38][39][40][41][42][43][44][45] .
Interestingly, it has been recently shown that FLRT2 is differentially regulated by β-estradiol, the main female sex hormone 73 . Additionally, it has been also reported that estradiol deficiency demotes synapses and impairs long-term potentiation (LTP) exclusively in females 48,69 . Furthermore, male and female FLRT2 +/− mice studied here not only showed a sex-dependent reduction in αand β-estrogen receptors, but only female FLRT2 +/− mice presented with reduced levels of the excitatory amino acid transporter 2 (EAAT2), thus suggesting a FLRT2-related synergistic crosstalk between sex-hormones and glutamatergic signaling. Further research is therefore required to verify and characterize this new molecular signaling path. EAAT2 (a.k.a. GLT-1), is a high affinity transporter for glutamate vastly expressed throughout the CNS that is accountable for more than 90% of glutamate reuptake in the CNS 74,75 . Additionally, increased levels of GLT-1 have been shown to exert negative effects on LTP and prolonged activation of the GLT-1 results in memory impairments in rats 76 . The regulation of GLT-1 by estradiol is very well established in non-neuronal brain cells 37,77,78 . Interestingly, it has been also shown that stimulation with β-estradiol results in an increased hippocampal GLT-1 expression 79 . These observations are in agreement with the improved MWM performance as well as with the enhancement of LTP and with the observed reduced protein levels of GLT-1 found here specifically in FLRT2 +/− females.
Here we also describe that whereas FLRT2 haplodeficiency results in augmented post-tetanic potentiation and enhanced memory recall in female brains, it also results in altered dendritic morphology, an observation at first glance rather paradoxical. One of the three members of the FLRT family of proteins, FLRT3 has shown similar effects in vivo. It has been additionally shown that a reduction in the levels of FLRT3 (which has a completely different pattern of tissue localization compared to FLRT2) reduces afferent input strength and dendritic spine number in neurons from the dentate gyrus 21 . During development, both FLRT2 and FLRT3 proteins have been shown to have dual -and rather opposing-roles as neuronal adhesion and repulsion molecular elements, although the pathways underlying these actions have only recently begun to be identified and characterized 22,80 . FLRT proteins are also known to modulate Growth Factor signaling 16,17,81,82 and our data here further suggest a potential implication of sex-hormone dependent mechanisms of regulation of the FLRT2 function. Thus, how FLRT2 can concomitantly enhance plasticity-mediated responses and decrease apical dendritic length and thin spine numbers remains to be elucidated. One could however, cautiously speculate that FLRT2 could -for example-mediate in the promotion of dendritic spine clustering, which is known to result in the nonlinear summation of synaptic inputs thus inducing synaptic strengthening 83 . FLRT2 could also dually influence spine turnover or negatively regulate the number of basal spines versus positively boosting the activity-dependent increase in the number of entirely newly formed spines during plastic events (see also 47 ). The growing variety of FLRT2 signal complexity makes also plausible to propose that perhaps the here observed, transitory enhancement of plasticity and memory recall is concomitant with alterations in dendritic complexity in female FLRT2 +/− neurons due to the induction of a gene-dependent compensatory mechanism (see also 84,85 ). As also known for other proteins 86 , one example of this compensation could in fact be the here reported decrease in the levels of the EAAT2 transporter, which could contribute to increase the levels of excitatory neurotransmitter in the synaptic cleft so all dendritic spines would be receiving stronger synaptic inputs. Additional research is therefore required to clarify the diversity of FLRT2 functions in the adult brain and to explore the possible existence of a molecular feedback mechanism engaged in homeostatically regulating the neuronal effects of FLRT2 down-regulation in a sex-dependent manner.
Taken together, our data propose FLRT2 as a novel molecular element that might importantly contribute in vivo to learning-related synaptic plasticity in the adulthood, a hypothesis that also invites for further research in the field of the sexually dysmorphic cognitive function and FLRT proteins in both physiological and pathophysiological contexts.

Materials and Methods
Animals. qRT-PCR. Animals were sacrificed by neck-dislocation and brain dissection performed in ice-cold aCSF solution. Following hippocampi collection, tissue was stored at -20 °C in RNA later ® solution (Ambion, Austin, Texas). qRT-PCR was performed as described elsewhere with minor modifications 87  Behavioral testing. All experiments conducted in this study were carried out during the light phase of the animals day-night cycle by an experimenter blinded to the genotype of the animals. Animal numbers are specified in figure legends. At least one week before the behavioral testing, mice were single caged and handled for 5 min daily in order to reduce stress from manipulations throughout the experiments. Additionally, at least 1 h before each experiment mice were placed in the room where behavioral testing was performed for habituation. Rota-rod. The Rota-Rod test was performed as previously described [89][90][91] . Briefly, each animal had a 60 s accommodation period (pre-training) before starting with the first trial on the Rota-rod Treadmill (USB Rota Rod "SOF-ENV-57X", MedAssociates Inc., St. Albans USA). The speed at the start was 4 rpm and was gradually accelerated until 40 rpm during the period of 5 min. The trial was stopped either at the end of 5 min or when the animal falls of the rod. The time spent on the Rota-rod was automatically recorded and used to evaluate the performance of mice. Each animal's performance was recorded in 3 trials with 15 min breaks in-between and the average was used to check for any deficits in motor coordination in FLRT2 +/− mice as compared to their wild type littermates 92 .
The tank (122 cm diameter, 76 cm height) was filled with tap water at room temperature ((22 ± 1) °C, 48 cm depth). Water was made opaque with non-toxic white paint (Viewpoint, France). Motor adeptness and visual capacity of mice were checked on one pre-training day (two trials). During visual acuity examination, a visible square platform (10 cm) with a red flag was placed in the tank. If the animal was not able to locate the platform within 2 min, it was guided to it. On the first training day, the platform was submerged 1.5 cm beneath the water surface. Once placed in the pool of water, animals had to find the hidden platform. Handmade distal cues were mounted outside of the pool as navigation aids. Following protocols previously used with success by our group 89,91 , each mouse was given 3 trials per day for 3 training days (TD) with inter-trial period of 1 h. Latencies to reach the platform and swimming speeds were recorded/analyzed using an automated system (Monochrome/ Near Infra-red camera coupled to tracking software: Videotrack [02-WATERMAZE-NSHD-LBW], Viewpoint, France). A probe trial (PT) was conducted 24 h after the last training trial. For this, the platform was removed to test memory recall and animals were allowed to swim for 30 s. Times spent in the target quadrant (in which platform used to be) were used as a readout of long-term memory.

Extracellular recordings.
To record evoked field excitatory postsynaptic potentials (fEPSPs) borosilicate glass pipettes (2)(3)(4)(5) were filled with aCSF solution and used as recording electrodes, located ~400 μm from the Teflon-coated tungsten wire bipolar stimulating electrode (~50 µm diameter tip) at the CA3-Schaffer Collateral-CA1 synaptic region, as previously published 91 . Input/Output (I/O) curves were obtained by repeated stimulation with pulses of voltage (0-6 V, 1 V increments, 100 μs duration) with an inter-stimulus interval of 10 s. The decaying slopes of field potentials were normalized to the maximum value and the average was compared between FLRT2 +/− and wild type mice of both sexes. In long-term potentiation (LTP), stimulation inducing 40-50% of maximal amplitudes were used for baseline (15 min) recordings (pulse duration = 100 μs, 0.03 Hz). LTP was induced by 5 trains/100 Hz (10 μs/pulse, 4 s interval between trains, same amplitude/duration as for base line recordings). Synaptic potentiation was determined by analyzing the changes in the slopes of the fEPSP decaying phase using baseline values as reference. AxoClamp-2B amplifier (Bridge mode) and a Digidata-1440 interface (Axon Instruments, MolecularDevices, Berkshire, UK) were used for recordings and pClamp-10 software (Molecular Devices,Berkshire, UK) was used for offline analysis. Dendritic Morphology Analysis. Brains from FLRT +/+ (n = 5) and FLRT +/− (n = 4) mice (8 weeks old) were used for neuronal morphological analysis with the FD Rapid GolgiStainTM kit following the instructions of the manufacturer (Cat Nr. PK401, FD Neurotechnologies, INC). Blind analysis was performed on tertiary branches of the apical trees of pyramidal neurons from the CA1 hippocampal region, an area widely known for its implication in learning and memory 95,96 and where FLRT2 proteins are endogenously expressed 22,23 . Randomly picked, consistently impregnated neurons relatively isolated from adjacent impregnated neurons were selected to ensure that the examined dendrites originated from identified cells. The morphometric properties of neurons were examined using the Neurolucida (version 10) system. At least four neurons per brain (two per brain hemisphere) were used for analysis. For dendritic spine analysis, a segment of 40 µm per neuron was examined. Spine type was determined attending to the following morphometric parameters (length/width < 1 = Stubby; length/ width < 1 and width > 0.6 µm = Mushroom; length/width > 1 and length < 1 µm = Thin; length/width > 1 and 1 µm ≤ length < 2 µm = Long Thin).
Western blot. Total membrane proteins were isolated as previously described 97 . Hippocampi were homogenized in Synaptic Protein Extraction Reagent (Syn-PER, Thermo Fisher Scientific, MA USA) and centrifuged in a bench top centrifuge at 1200 × g for 10 min at 4 °C. In addition, supernatants were collected and centrifuged at 15,000 × g for 20 min at 4 °C. Pellets with the synaptosomal fraction were solubilized in a buffer containing: 1.5% SDS, 100 mM NaCl, 20 mM Tris-HCl pH 7.5, Protease Inhibitor Cocktail (Sigma Aldrich, Vienna, Austria), and gently agitated at room temperature for 1 h. Next, protein concentrations of samples were determined using the BCA protein assay kit (Pierce, Rockford, USA). Following the protein quantification, the samples were mixed with 4 × Laemmli Sample Buffer (Bio-Rad Laboratories, Richmond, CA, USA) and incubated for 30 min at room temperature. Electrophoresis was performed using the Criterion electrophoresis cell (Bio-Rad Laboratories, Richmond, CA, USA) with 10 µg of protein per sample loaded on 1D-SDS-PAGE (10% separating gel), and followed by the transfer of proteins onto the PVDF membrane (Millipore, Darmstadt, Germany) at 350 mA of constant current for 2 h at 6 °C using a semi-dry transfer system (PerfectBlue ™ Electroblotter Sedec ™ M, Peqlab Biotechnologie GmbH, Erlagen, Germany). Next, membranes were blocked in TBS containing 5% non-fat dry milk for 1 h at room temperature, followed by overnight incubation with diluted primary antibody at 4 °C. Antibodies used were VGLUT1 (1: No. 7076, Cell Signaling) and anti-guinea pig IgG (1:10000; Cat. No. 706-035-148, Jackson Immuno Research) secondary antibody for 1 h at room temperature. Membranes were developed with ECL Western blotting detection system (Bio-Rad Laboratories) and proteins visualized by FluorChem HD2 system (Alpha Innotech, San Leandro, CA, USA). Protein levels were quantified by densitometry using ImageJ software (http://rsbweb.nih. gov/ij/).

Statistical analysis.
Statistical analyses were carried out using GraphPad Prism Software, version 7.0 (San Diego, CA, USA). T-test was used for comparisons between two groups and two-way ANOVA followed by Tukey's multiple comparisons tests as posthoc comparisons for multiple groups where appropriate. For latency, swimming speed, Input/output curves and long-term potentiation measurements two-way repeated measures ANOVA was used. P values ≤ 0.05 were considered significant. All data are expressed as means +/− standard error.