LRP4 mediated synaptic plasticity in the piriform cortex

Background Low-density lipoprotein receptor-related protein 4(LRP4) plays a critical role in the central nervous system (CNS), including hippocampal synaptic plasticity, maintenance of excitatory synaptic transmission, fear regulation, as well as long-term enhancement. Results In this study, we found that Lrp4 was highly expressed in the piriform cortex and located in the second layer of the piriform cortex. When the transmembrane domain (TMD) and the intracellular domain (ICD) were missing, the Lrp4 ECD/ECD mice appeared to be smaller, and the brain’s weight decreased, compared with the control mice. Simultaneously, nding food was prolonged for Lrp4 ECD/ECD mice in the buried food-seeking test. In the piriform cortex of Lrp4 ECD/ECD mice, the spine density of layer (cid:0) increased, and the frequency of both miniature excitatory postsynaptic current (mEPSC) and spontaneous excitatory postsynaptic current (sEPSC) enhanced. This study indicated that LRP4 mediated synaptic plasticity in the piriform cortex. Moreover, it also suggested that TMD and ICD of LRP4 are nonnegligible for the LRP4 function in the piriform cortex.


Introduction
LRP4 plays an essential role in forming and maintaining synapses and synaptic plasticity, and excitatory transmission in CNS, and it is expressed in the hippocampus, olfactory bulb, cerebellum, and neocortex [1][2][3]. Recently, Zhang et al. found that genetic deletion of Lrp4 increased Aβ plaques formation in Alzheimer's disease (AD) mice and exacerbated the de cits in neurotransmission, cognition, and synchrony between the hippocampus and prefrontal cortex [4]. Astrocytic LRP4 played a crucial role in AD pathology and cognitive function. Sun et al. found that astrocytic LRP4 regulates ATP release [1]. Glutamate release of the hippocampal neurons was impaired because of ATP release enhancement in Lrp4 knockout astrocytes [1]. These research results show that LRP4 plays a signi cant role, including hippocampal synaptic plasticity, excitatory synaptic transmission, fear regulation, and long-term enhancement [5][6][7].
In our study, Lrp4 LacZ/+ mice were utilized to locate the Lrp4 expressing region. Intriguingly, except the hippocampus in the CNS as previous,Lrp4 was also supremely expressed in the piriform cortex, which played a critical role in transmitting olfactory signals. To explore whether LRP4 involves the sense of smell, Lrp4 ECD/ECD mice were investigated. Finding food was prolonged in the buried food-seeking test, implying the olfactory function was impaired in Lrp4 ECD/ECD mice. Besides, we also found that the body and brain weight of Lrp4 ECD/ECD mice were lighter than the control mice. A further study utilizing Lrp4 ECD/ECD mice indicated that piriform cortical neuronal dendritic spine density increased. Compared with the control mice, Lrp4 ECD/ECD mice exhibited a higher frequency of sEPSC and mEPSC in the piriform cortical neurons. These data indicate that the TMD and ICD of LRP4 are nonnegligible for the LRP4 mediating synaptic plasticity in the piriform cortex.

Animals
Lrp4 LacZ/+ mice were described before; in brief words, β-galactosidase (β-gal) protein cassette, including stop code and a polyadenylation termination signal, was inserted into the downstream of Lrp4 promoter [1].Lrp4 ECD/ECD mice (JAX stock #013157) were described before, which introduced a stop codon just upstream of Lrp4 TMD[6, 7, 18]. Mice were housed in a room 12-h light/dark cycle, 23-25°C, with ad libitum access to rodent chow diet and water. Experiments involving animals were conducted according to the "guidelines for the care and use of experimental animals" issued by Nanchang University, following the directive 2010/63/EU to protect animals used for scienti c purposes. For in vivo experiment, surgery was executed with sodium pentobarbital anesthesia (50 mg/kg, ip injection), and all efforts were made to minimize suffering [19]. Male mice were utilized for the experiments, and after terminal experiments, the mice were euthanized by carbon dioxide inhalation.
Luminata TM Crescendo Western HRP Substrate was added. Immunoreacted bands were captured by an enhanced chemiluminescence system (BIO-RAD, USA).

Open-eld test
In behavioral tests, the activity levels of the mice were evaluated at P50. The open eld (40 × 40 × 20cm) was used to measure the mice's moving distance over 10 min. The data were recorded using a video camera, and the data were analyzed using the behavior analysis software ANY-maze (Stoelting Co., Wood Dale, IL, USA).
Buried food-seeking test Mice were food-deprived for 2 d, trained for 2 d, and tested continuously 3 d, with ad libitum access to enough water all the time. Food was randomly placed in the box and was buried under padding for 0.5cm in testing trials. The mice seizing the food with their front paws and biting with mouth were regarded as nding the food. The time was recorded when the mice were placed in the container and found the food.
Mice brains were xed for 8-10 h in 2%(m/v) paraformaldehyde (PFA) at room temperature and then were transferred into 30% sucrose solution at 4°C. The brain slices were added PBS (phosphate-buffered sodium, pH 7.4) in a wet box, washing the slices at room temperature 3 times with PBS. After rinsing for 10 min, adding dye solution, putting the slices at 37°C for 8 h. After the reaction, brain slices were washed 3 times with PBS.
Brain slices were washed with PBS 3 times, and then the images were captured by a microscope (Olympus FSX100).

Nuclear fast red counterstaining
Put the X-gal-stained or co-stained brain slices into a vitreous tank containing nuclear fast red staining solution for 5 min. Slides with the slices were put into glass tanks containing 50%, 75%, and 90%ethanol in sequence, each for 4 min. The slides were transferred into 100% ethanol 2 times. Then the slides were put into xylene for 5 min. At last, the slides were sealed with mounted in Hydro mount (National Diagnostics). Images were captured by an inverted uorescence microscope (Olympus FSX100).

Immuno uorescent staining
The brain slices were rinsed with PBS at room temperature and were immersed with antibody blocking

Golgi staining
Golgi staining was performed following the FD Rapid Golgi Stain™ Kit (FD Neuro Technologies, PK-401, USA). Staining solution D and solution E were mixed with ultra-pure water in a ratio of 1:1:2. Dying at room temperature for 10 min. Slides with the slices were washed in ultra-pure water twice, then put into the plate hole containing 50%, 75%, 90%, and 100% ethanol in sequence, each for 4 min. Dehydration for 3 times, then the slides were put into xylene for 1 h and mounted in Hydro mount (National Diagnostics).
Images were captured by an Olympus uorescence microscope (FSX100), and dendritic spines were counted with image J.

Electrophysiological recording
The electrophysiological recording was performed as previously described [21,22]

Statistical analysis
Data were statistically analyzed with GraphPad Prism 6.0 software systems, and the values were expressed as means ± standard error (means ± SEM). One-way ANOVA (Fig 1C), two-way ANOVA (Fig 3B), and t-test analyzed the normality distributed data. All tests were two-sided. * p< 0.05, ** p < 0.01.

Results
Lrp4 was highly expressed in the piriform cortex Lrp4 LacZ/+ mice were utilized to locate the Lrp4 expression region by X-gal staining because X-gal is the substrate for β-gal. X-gal staining results showed that Lrp4 was expressed in many brain regions, such as the piriform cortex, hippocampus, and cerebral cortex (Fig. 1A). To verify the expression of Lrp4 in the brain regions, as mentioned above, we conducted western blotting experiments. The results also showed that Lrp4 was highly expressed in the piriform cortex, hippocampus, and cerebral cortex (Fig. 1B). To detect the expression pro le of Lrp4 in the piriform cortex, we also used quantitative uorescence PCR to quantify the expression of Lrp4 in the piriform cortex in postnatal wild-type mice (Fig. 1C). Lrp4 was at a low and stable level in adolescence and began to be highly expressed in adulthood in the piriform cortex, suggesting that Lrp4 expression was related to the development and may involve the olfactory pathway.
GFAP-positive cells in layer of the adult piriform cortex have high expression of Lrp4 Taking advantage of Lrp4 LacZ/+ mice,we used immunohistochemical co-staining to identify the location of LRP4. The results indicated that Lrp4 was mainly expressed in the layer of the piriform cortex ( Fig. 2A). The co-staining assay results suggested that X-gal co-stained with anti-GFAP (astrocyte and neuron stem cell marker) and GFAP-negative cells (Fig. 2B).
Lrp4 ECD/ECD mice have a regular structure of the piriform cortex ECD of LRP4 could maintain critical signaling thresholds for development [17]. Therefore, Lrp4 ECD/ECD mice could develop much better than Lrp4 null mice because the latter mice are dead at birth [9,15,16].
The body and brain weight of Lrp4 ECD/ECD mice were lighter than the control mice ( Fig. 3A-3D).
It was unclear whether the morphology of the piriform cortex in Lrp4 ECD/ECD mice changed. Firstly, immuno uorescent staining was carried out to observe the piriform cortex region of the mice. There was no remarkable difference in the thickness of Lrp4 ECD/ECD mice compared with the control mice (Fig. 2C,  2D), and the piriform cortical neuron density was similar in the two types of mice (Fig. 2E). We speculated that LRP4 ECD maintained the typical structure of the piriform cortex.
Morphology of neurons in Lrp4 ECD/ECD mice brain changed To explore whether the morphology of piriform cortical neurons in Lrp4 ECD/ECD mice change or not, we used Golgi staining to observe the neuronal dendritic spines in the Lrp4 ECD/ECD mice, compared with littermate control mice (Fig. 4). There are two different types of neurons in the second layer of the piriform cortex (Fig. 4B). One type is semilunar (SL) cell lacking basal dendrites, and the other one is super cial pyramidal (SP) cell with both apical dendrites and basal dendrites. In the piriform cortex, Lrp4 ECD/ECD mice showed an increase in the mature spine (mushroom type) and total spine density on SP neurons than the control mice (Fig. 4C), which implies a potential increase in functional synaptic transmission. Except for a bit of rising of the thin type of spine, Lrp4 ECD/ECD mice exhibited similar spine density on SL neurons than the control mice (Fig. 4D). The results suggested that the TMD and ICD of LRP4 were nonnegligible for the spine maturation and maintaining of neurons in the piriform cortex.
Enhanced excitatory postsynaptic transmission of the piriform cortical neurons in Lrp4 ECD/ECD mice In whole-cell patch-clamp con guration, piriform cortex pyramidal neurons were recorded in mice.
Compared with the control mice, Lrp4 ECD/ECD mice exhibit a high frequency of sEPSC (Fig. 5B) and mEPSC (Fig. 5D) of piriform cortical neurons. No change was observed in the amplitude of mEPSC and sEPSC (Fig. 5C, 5E). The results suggested hyperfunction of glutamatergic transmission in Lrp4 ECD/ECD mice, consistent with increased spine density (Fig. 4).

Impaired olfactory function in Lrp4 ECD/ECD mice
Both spine density and electrophysiology of piriform cortical neurons were changed in Lrp4 ECD/ECD mice. In order to explore the function of LRP4 in the olfactory pathway, a buried food-seeking test (Fig. 6) was performed in the two types of mice. Firstly, in the open-eld test, we found that the locomotive ability of Lrp4 ECD/ECD mice was not affected (Fig. 6A), without change in total travel distance (Fig. 6B) and average speed (Fig. 6C). Then, the buried food-seeking test was conducted. The mice were food-deprived for 2 d before training 3 d, and testing was started following training (Fig. 6D). Mice were free to access enough water all the time. Lrp4 ECD/ECD mice spent more time nding the buried pellet chow in the test trials than control mice (Fig. 6E, 6F), suggesting that olfactory function may be impaired.
In conclusion, we reported that Lrp4 was highly expressed in the piriform cortex, and Lrp4 ECD/ECD mice exhibited enhanced spine density in the piriform cortex, hyperfunction of the excitatory postsynaptic current in the piriform cortical neurons, and impaired olfactory function. These results implied that LRP4 might maintain the olfactory signal transmission pathway in the piriform cortex. Although LRP4 ECD seems to play a prominent role in piriform cortical development, both the TMD and the ICD of LRP4 may also be nonnegligible to the piriform cortical development and function.

Discussion
The piriform cortex with synaptic plasticity involves the encoding of olfactory information, associative memory, and sensory processing [23][24][25]. The piriform cortex containing the densely packed cell bodies exhibits highly structural plasticity, such as dendritic remodeling, spine genesis, and synaptic reorganization [26]. As a higher olfactory center receiving direct input from the olfactory bulb, the piriform cortex comprises the largest area of the olfactory cortex, which is vital for olfactory function processing its connection with all of the entorhinal cortical domains [27,28]. Like the hippocampus, the piriform cortex with a three-layered cortical structure belongs to an evolutionally conserved paleocortex, phylogenetically ancient structure [26,29].
Layer II of the piriform cortex is developed prenatally and is devoid of postnatally proliferative capacity. However, in layer II, it is often found that a subpopulation of immature neurons is characterized by the expression of doublecortin (DCX) and polysialylated neural cell adhesion molecule (PSA-NCAM)[26, [30][31][32]. Therefore, it is believed that the piriform cortex has a slight potential for adult neurogenesis [32][33][34]. Therefore, all GFAP-positive cells in the piriform cortex may not be astrocytes because PSA-NCAM positive neurons may also express GFAP [35].
The olfactory system's essential components include sensory neurons located on the olfactory mucosa, olfactory bulb, and olfactory cortex [36]. The most critical part of the olfactory cortex is the piriform cortex [37]. The olfactory cortex mainly integrates the olfactory signals, forms olfactory memories, and integrates speci c olfactory signals with sensory information, such as color, taste, shape, and spatial location [38]. Much complex odor signal analysis and integration rely on higher-level central structures, such as the piriform cortex [39]. The olfactory signal analysis in the piriform cortex may be related to speci c olfactory memory and may involve olfactory sensitization and passivation [40]. It was deeply understood the recognition mechanism of odor molecules in the past ten years, the projection of olfactory mucosa to the olfactory bulb, and olfactory bulbs' spatial orientation to speci c olfactory signals [41]. However, it is still unclear how to complete the integration and modulation of olfactory signals by the high cortex. Here we demonstrated that Lrp4 was highly expressed in the olfactory pathway, especially in the piriform cortex. In this area, we found that Lrp4 was also highly expressed in piriform cortical GFAP-positive cells except for GFAP-negative cells, and LRP4 may regulate the transmission of olfactory signals.
In the buried food-seeking test, results showed that the time to nd food was signi cantly longer in Lrp4 ECD/ECD mice comparing with the control mice. The test results suggested that Lrp4 ECD/ECD mice may have olfactory dysregulation. At the same time, Lrp4 ECD/ECD mice also showed reduced body weight and brain weight. We also found that Lrp4 ECD/ECD mice showed typical tight-knit morphology in the previous report [6]. This kind of limbs clinging to the tail after the tail's suspension also appeared in the neurological disease model mice, suggesting that the brain's neurological function in the Lrp4 ECD/ECD mice may be impaired.
Ariana Kariminejad et al. had reported one case that a patient identi ed the novel homozygous mutation c.289G > T in Lrp4 exon 3. This nucleotide exchange leads to a premature stop codon at amino acid 97 (p.E97X) at the beginning of the large extracellular domain. The patient had mixed-type hearing loss, vertebral anomalies, and renal hypoplasia [42]. Another study reported a novel splice variant in Lrp4 (c.316t1G > A), and the missense variant adds 29 non-native amino acids with premature stopcodon,which causes the Lrp4 encoding to terminate prematurely. The patient had short feet, frontal bossing, and other symptoms [43]. However, these ndings suggest that the ECD of LRP4 plays a vital role in limb development, kidney development, and brain development; they may also imply that both the TMD and ICD of LRP4 may be nonnegligible in the development.
Wnt signaling regulates brain development and synapse maturation [44]. LRP4 has an antagonistic function on LRP6-mediated Wnt/β-catenin activation. Eva Ramos-Fernández et al. reported that Wnt-7a stimulates dendritic spine formation in the hippocampus through glycogen synthase kinase-3 β (GSK-3β) inhibition, triggering β-catenin/T cell factor/lymphoid enhancer factor (TCF/LEF)-dependent gene transcription and promoting postsynaptic density-95 (PSD-95) protein expression, infecting the synapse plasticity [45]. Our results showed that the dendritic spine density in the piriform cortex of Lrp4 ECD/ECD mice signi cantly increased. Furthermore, we hypothesized that the ECD of LRP4 might promote dendritic spine formation. SP dendritic spines in the piriform cortex increased signi cantly, while the density of dendritic spines of SL cells did not change. The neuronal dendritic spines are related to neuronal physiological functions, and they are likely to participate in different neural circuits. On the other hand, neuronal dendritic spine density also affects the same afferent arrival of neurons. Two types of particular neurons in the piriform cortex may be involved in different neural circuits.
This study found enhanced excitatory synaptic transmission of the piriform cortical neurons in Lrp4 ECD/ECD mice. Pohlkamp et al. examined synaptic function in the Lrp4 ECD/ECD mice by recording thetaburst long-term potentiation (LTP) in CA3-CA1 Schaeffer collaterals [7]. CA3-CA1 projections are a classic model for measuring and understanding synaptic plasticity. There was a substantial de cit in late-phase LTP in Lrp4 ECD/ECD mice compared with the control mice [7], and the loss of the ICD and TMD may severely impair the LRP4 function. Sun et al. found that Lrp4 knockout astrocytes suppress glutamatergic transmission in the CNS [1]. The frequency of sEPSC and mEPSC in hippocampal CA1 pyramidal neurons was reduced, and synaptic plasticity was impaired in Lrp4 knockout mice [1,6]. In our research, sEPSC and mEPSC frequency in piriform cortical neurons were enhanced in Lrp4 ECD/ECD mice. Although consistent with the increase of the piriform cortex's dendritic spine density, it was different from other reports [1,6]. There are two possible reasons for the inconsistencies. One is that different brain areas were investigated, and the other reason is that the mechanism involving in the dysfunction may be caused by Lrp4 expression in GFAP-negative cells.

Conclusions
In conclusion, our results showed that Lrp4 was highly expressed in the piriform cortex, especially in layer a, and LRP4 localized in both GFAP-positive and GFAP-negative cells. In Lrp4 ECD/ECD mice, piriform cortical neuronal dendritic spines density increased, and the olfactory function was impaired. The results suggest that the TMD and ICD of LRP4 are nonnegligible for the LRP4 function in the piriform cortex while maintaining the olfactory signal transmission pathway. The molecular regulating mechanism needs further exploration.

Availability of Data and Materials
The datasets used or analyzed in our study are available from the corresponding author on reasonable request.

Ethics Approval and Consent to Participate
All experiments involving animals were conducted according to the "guidelines for the care and use of experimental animals" issued by Nanchang University. The Committee on the Ethics of Animal Experiments of the University of Nanchang approved the protocol.

Consent for Publication
Not applicable (C) Representative images of neuron staining (anti-NeuN) of the two types of mice brain. (D) The thickness of the piriform cortex in layers , , and did not change in Lrp4ECD/ECD mice. (E) The neuronal density was not different in the three layers (mice per type, n = 7; scale bar = 50 μm; values were means ± SEM).

Figure 3
Lower body and brain weight of Lrp4ECD/ECD mice (A) Representative images of one-month-old Lrp4ECD/ECD mice comparing with the control mice. Lrp4ECD/ECD mice were smaller than control mice.