Introduction

Schizophrenia is a highly debilitating psychiatric disorder that affects approximately 1% of the population worldwide [1, 2]. Cognitive deficits, which are often apparent before the onset of psychosis, are a core feature of schizophrenia. However, traditional antipsychotic medications primarily target positive symptoms and have poor efficacy in treating neurocognitive dysfunction symptoms [3,4,5]. Although the dopamine dysregulation hypothesis explains the mechanism of positive symptoms, the N-methyl-D-aspartate (NMDA) receptor hypofunction hypothesis provides the rationale for schizophrenia symptoms in general [6,7,8]. NMDA receptor blockers (ketamine or phencyclidine) induce positive syndrome, negative syndrome, and cognitive deficits in human and rodent models. Dizocilpine (MK-801), one of the NMDA receptor blockers, has been used to mimic broad spectrum of schizophrenia-like behaviors in animal models [9,10,11].

One of the most consistent findings in postmortem brain studies of subjects with schizophrenia is decreased glutamic acid decarboxylase 67 (GAD67) expression in a subset of gamma-aminobutyric-acid (GABA)ergic neurons, including parvalbumin (PV) neurons [12, 13]. A hypothesis expounds that the activation of GABAergic interneurons (for example: PV + neuron) via NMDA receptors exerts a local feedback inhibition on the glutamatergic neurons [14,15,16]. A blockade of NMDA receptors causes an excessive release of glutamate in the cerebral cortex [17]. This disruption of prefrontal cortical inhibitory circuits is thought to contribute to the development of the cognitive deficits observed in schizophrenia because GABAergic transmission is crucial for normal cognitive functions [18, 19]. It has been shown that MK-801 treatment impairs cognitive functions and reduces the density of PV-positive (PV +) cells in the rat medial prefrontal cortex (mPFC) and hippocampus [20]. Furthermore, the calbindin (CB)-containing GABAergic interneurons is also reduced in the mPFC and hippocampus by MK-801 while CB expression is crucial for normal cognitive functions in brain [20, 21]. Relatedly, decreases in the relative densities and morphology disorders of PV + and CB + interneurons have been observed in the hippocampus of patients with schizophrenia [22,23,24]. PV + and CB + interneurons could thus become a potential target for alterations induced by the chronic treatment of NMDA receptor blocker MK-801 [25,26,27]. In addition, the cingulate cortex is a functionally heterogeneous region involved in diverse cognitive and emotional processes [28]. The negative symptoms of schizophrenia, relatedly, are generally associated with altered functioning in several cerebral regions, such as the prefrontal cortex and cingulate cortex [29, 30].

Brain-derived neurotrophic factor (BDNF) is a key member of the neurotrophic family and plays important roles in sustaining neuronal health and synaptic plasticity. Reduction of BDNF level has been reported in the brain of the schizophrenia patients and BDNF is considered as therapy target for schizophrenia [31, 32]. Recent studies have shown that low-intensity pulsed ultrasound (LIPUS) is capable of increasing levels of BDNF and improving cognitive dysfunction in brain disorder rat models [33, 34]. In particular, transcranial ultrasound stimulation (TUS) offers the opportunity to stimulate specific brain regions [35, 36]. However, the ability of LIPUS to improve symptoms of schizophrenia is still unclear. Based on this background, the aim of the current study was to investigate the effects of LIPUS on cognitive deficits induced by the subchronic administration of MK-801. Furthermore, investigating if LIPUS restores PV + and CB + cells and alteration of BDNF level in the brain could provide insights regarding the possible mechanisms underlying the treatment of schizophrenia.

Materials and Methods

Pulsed Ultrasound System

The LIPUS system used in this study was established following the same procedures used in our previous study (Fig. 1A). LIPUS was generated by a 1-MHz focused piezoelectric transducer (A392S; Panametrics, Waltham, MA, USA) with 50 ms burst lengths at a 5% duty cycle and a repetition frequency of 1 Hz. The spatial-peak temporal-average intensity (ISPTA) over the focused transducer head was 528 mW/cm2 and was measured with a radiation force balance (RFB, Precision Acoustics, Dorset, UK) in degassed water. The focused transducer was mounted on a removable cone filled with deionized and degassed water, the tip of which was capped by a polyurethane membrane, with the center of the focal zone placed about 5.7 mm away from the cone tip. The focused transducer was positioned using a stereotaxic apparatus in order to direct the acoustic beam to the desired region (3.0 mm posterior and 2.5 mm lateral to the bregma) of the brain. Each rat hemisphere was treated by LIPUS daily with triple sonications. The duration of each sonication was 5 min, and there was an interval of 5 min between each sonication.

Fig. 1
figure 1

Experimental design and effects of low-intensity pulsed ultrasound (LIPUS) stimulation on body weight and locomotor activity. (A) Schematic diagram of LIPUS setup. (B) Flow chart of the experimental procedure. Rats were treated with MK-801 (0.5 mg/kg) or saline twice daily for 7 days. Two days after the last injection, the LIPUS group and the MK-801 + LIPUS group were stimulated with LIPUS while the Sham group and the MK-801 group just were anesthetized daily for 5 days. (C) The effects of LIPUS stimulation on body weight of rats treated with MK-801. (D) The effects of LIPUS stimulation on the movement distance of rats treated with MK-801 over 10 min in an open field test (24 h after the last LIPUS stimulation). * denotes significantly different between Sham group and MK-801 group. (*, p < 0.05; **, p < 0.01). # denotes significantly different between Sham group and MK-801 + LIPUS group. (#, p < 0.05; ##, p < 0.01). n = 8 for all groups

Full size image

Animals and Treatment Protocols

All procedures involving animals were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals. The study protocol was approved by our Animal Care and Use Committee. Male Sprague–Dawley (SD) rats weighing from 230 to 280 g were used in this study. The animals were habituated in home cage for 7 days. After habituation period, the animals were randomly assigned into four groups, namely, the Sham (sham-operated) group, the LIPUS group, the MK-801 group, and the MK-801 + LIPUS group. The Sham group and the LIPUS group was intraperitoneal injected with saline while the MK-801 group, and the MK-801 + LIPUS group injected with MK-801(0.5 mg/kg) twice daily for 7 days. Two days after last injection, the treatment of LIPUS began. Before the LIPUS stimulation, each animal was anesthetized in the prone position by inhalation of 2% isoflurane in 2 l/min oxygen, and the body temperature was maintained at 37 °C using a heating pad. The head of each rat was mounted on a stereotaxic apparatus (Stoelting, Wood Dale, IL, USA), and the top of the hair around cranium was removed with electric shaver and hair removal cream for LIPUS stimulation. During treatment of LIPUS, four groups were anesthetized continuously. LIPUS would be turned on when the LIPUS group and the MK-801 + LIPUS group accepted treatment while the Sham group and MK-801 group were just anesthetized. The only different step between four groups is that LIPUS was turned on or not. Twenty-four hours after last treatment, the animals were assigned for biochemical analysis, histological assessment, and behavioral evaluation (Fig. 1B).

Western Blot Analysis

Twenty-four hours after the last LIPUS stimulation for each rat, the rat was sacrificed for biochemical analysis. Fresh brain tissues in the hippocampus and mPFC were homogenized by T-Per extraction reagent supplemented with the Halt Protease Inhibitor Cocktail (Pierce Biotechnology, Inc.). Lysates were centrifuged and the supernatants were harvested, and protein concentrations were assayed with Protein Assay Reagent (Bio-Rad, CA, USA). Samples containing 30 μg protein were resolved using 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to Immun-Blot® polyvinyldifluoride (PVDF) membranes (Bio-Rad, CA, USA). After blotting, the membranes were blocked for at least 1 h in milk (anchor), and then the blots were incubated overnight at 4 °C in a solution with antibodies raised in rabbit against GAD67(GTX101881, GeneTex, 1:1000), CB(GTX130856, GeneTex, 1:1000), PV(GTX11427, GeneTex, 1:1000), BDNF(GTX132621, GeneTex, 1:1000), GDNF(GTX02540, GeneTex, 1:1000), or VEGF(GTX102643, GeneTex, 1:1000) or β-actin(GTX109639, GeneTex, 1:2000). After being washed with PBST buffer, each membrane was incubated with the secondary antibodies (Goat Anti-Rabbit IgG HRP, GTX213110-01, GeneTex, 1:5000 or 1:10,000 for β-actin) for 1 h at room temperature. After being washed with PBST buffer, signals were developed using a Western Lightning ECL reagent Pro (Bio-Rad, California, USA). The gel image was captured using an ImageQuant™ LAS 4000 biomolecular imager (GE Healthcare Life Sciences, Pennsylvania, USA) and analyzed using Image J software (Image J, National Institute of Health, Bethesda, MD, USA) to estimate the integral optical density of the protein bands.

Histopathological Evaluation

Five rats from each group were prepared for histological assessment. The rats were perfused with saline and 10% neutral buffered formalin at 24 h after the last LIPUS stimulation. The brains were removed, embedded in paraffin, and then serially sectioned into 10-μm-thick slices. In brief, blocking with Foetal Bovine Serum (SKU 04–001-1A, biological industries USA for 10 min at 37 °C. sections were incubated with primary antibodies (Calbindin: GTX634833, GeneTex, 1:100; Parvalbumin: GTX89558, GeneTex, 1:100) overnight at 4 °C. Then, the samples were incubated DyLight594-combined with the corresponding secondary antibody for 30 min at 37 °C, washed with PBS buffer solution for 3 times, incubated with DAPI buffer solution. Three different sections per rat were analyzed at the same exposure level. At least three randomly distributed fields from the mPFC were captured for each section. Double immunofluorescence was observed under a fluorescence microscope (Leica DM 6000B, Mannheim, Germany) and photographed. The number of cells positive for PV-DAPI-double-label and CB-DAPI-double-label was counted in an area of 534 × 400 μm2 in three non-overlap** fields in the cingulate cortex of the mPFC using a magnification of × 200.

Behavioral Assessment

Spontaneous object recognition (SOR) was performed in a Y-shaped apparatus as described previously [37, 38]. The Y-shaped apparatus had high homogenous white walls constructed from Plexiglas to prevent the rat from seeing out into the surrounding testing room. The apparatus walls were 40 cm high, and each arm was 27 cm long and 10 cm wide. A video camera was mounted on a tripod above the apparatus to record all the trials. All of the rats were habituated to the Y-shaped apparatus over 2 consecutive days. These habituation sessions consisted of two 5-min exposures to the empty Y-shaped apparatus in normal lighting conditions. Each trial consisted of two phases, sample and choice, separated by a retention delay. In the sample phase, two identical objects were placed at the end of the two exploration arms. The sample phase ended when the rat had explored the identical objects for a total of 25 s, or after 3 min has passed, whichever came first. Touching the object with noise or whiskers was defined as exploration. In the choice phase, the Y-shaped apparatus contained the sample object in one arm and a new object in the other arm. The locations of familiar and novel objects were counter-balanced. The time spent in exploring the novel and familiar objects was recorded for the 1-min choice phase. The time spent exploring each object was recorded and analyzed with an EthoVision video tracking system. The discrimination ratio was determined as (novel − familiar)/(novel + familiar).

The performance of reversal learning was assessed by Morris water maze [39]. A circular plastic pool (diameter: 180 cm, height: 60 cm) was filled with nontoxic black ink water (23 ± 2 °C) and was divided arbitrarily into four equally-sized quadrants. An escape platform (diameter: 15 cm, height: 40 cm) was submerged 2 cm below the surface of the water in the center of one of the quadrants. Visual stimuli were placed on the walls around the pool to provide the rats with spatial cues. For each trial, rats were released from one of the two farthest points from the platform. The rat swam until it located the hidden platform or, if it failed to find the platform in 60 s, the rat was guided to the platform. Rats then remained on the platform for 15 s before the next trial began. After 4 trials of session 1, rats spent 15 min under a heat lamp and then 5 min in their home cage before receiving 4 more similar trials in session 2 but with the platform in a new location. The time to reach the platform was used to evaluate the performance of reversal learning.

Statistical Analysis

All data were expressed as mean ± SEM. A two-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was performed to determine the individual and interactive effects of LIPUS treatment and MK-801 administration through the four groups. Only body weight of rats was analyzed with Repeated measure ANOVA. The level of statistical significance was set at p value ≤ 0.05.

Results

Ultrasound Stimulation Ameliorates Behavioral Impairments in Schizophrenia

First, significant reduction was found in the MK-801 group as compared to the Sham group from day 13 to day 17 (6th day of drug injection to 1st day of LIPUS treatment). This reduction did not be reversed in the LIPUS + MK-801 group. Moreover, the LIPUS group did not show significant change as compared to the Sham group (Fig. 1C). To confirm the effects of LIPUS on MK-801-induced schizophrenia-like behaviors in rats, the tests of open field behavior, SOR, and reversal learning were used. In the open field task, the MK-801 group did not exhibit aberrant locomotor activity as compared to the Sham group (1005.5 ± 30.0 versus 957.8 ± 59.5, p = 0.77; Fig. 1D), and no significant effects were observed in the LIPUS group and the MK-801 + LIPUS group (955.4 ± 60.0, p = 0.73; Fig. 1D) as compared to the MK-801 group (Fig. 1D).

In the SOR task, the mean discrimination ratios did not differ among the four groups at the immediate and 5-min retention delays (Fig. 2A, B). At the 24-h retention delay, the MK-801-treated group rats exhibited significant impairment on the SOR task as compared to the saline-treated sham group (0.1 ± 0.1 versus 0.4 ± 0.0, p < 0.05; Fig. 2C). Moreover, the MK-801 + LIPUS group exhibited a significant increase in the mean discrimination ratio as compared to the MK-801-treated group (0.4 ± 0.1 versus 0.1 ± 0.1, p < 0.05; Fig. 2C). However, no significant change was observed in the mean discrimination ratio in the LIPUS group as compared with the sham group.

Fig. 2
figure 2

Effects of LIPUS stimulation on behaviors of rats treated with MK-801. Spontaneous object recognition (SOR) task performance with (A) an immediate, (B) 5 min, and (C) 24 h retention delay by sham rats and those treated with LIPUS or MK-801 or MK-801 and LIPUS. (D) Swim time of rats treated with MK-801 or saline over trials during session 1 and session 2 of reversal learning task. (E) Swim time of four groups at trial 1 of session 2 in reversal learning task. At the 24 h retention delay, rats treated with MK-801 were significantly impaired compared with saline-treated sham rats. LIPUS stimulation significantly increased the discrimination ratio in the MK-801 + LIPUS group compared with the MK-801 group. * and † denote significantly different from Sham group and MK-801 group, respectively. (*,†, p < 0.05). n = 8 for all groups

Full size image

In the reversal learning task, both the sham and MK-801 groups showed reduced swim times during the trials of session 1 and session 2 (Fig. 2D). On trial 1 of session 2, the MK-801-treated group exhibited a significantly longer swim time than the sham group (56.6 ± 2.4 versus 30.8 ± 6.3, p < 0.05; Fig. 2D). Subsequent analyses of LIPUS treatment effects focused on this trial (Fig. 2E). The LIPUS group exhibited no significant effects as compared with the sham group. However, LIPUS treatment significantly reduced the swim times in the MK-801 + LIPUS rats in comparison to the rats with MK-801-induced reversal learning deficits (35.8 ± 5.7 versus 56.6 ± 2.4, p < 0.05; Fig. 2E).

LIPUS Reverses MK-801-Induced Decreases in CB and PV Levels in the Rat Hippocampus

GABAergic interneuron dysfunction is involved in the pathophysiology of schizophrenia. To evaluate the effect of LIPUS on relative GABAergic dysfunction in schizophrenia, we examined whether LIPUS reverses MK-801-induced downregulations of GABAergic neuron markers in rats. Significant reductions in the hippocampal levels of CB (0.47 ± 0.03 versus 0.95 ± 0.10, p < 0.05; Fig. 3C) and PV (0.52 ± 0.08 versus 0.95 ± 0.06, p < 0.05; Fig. 3D), but not GAD67 (0.80 ± 0.06 versus 0.76 ± 0.06, p = 0.64; Fig. 3B) were observed in the MK-801 group as compared to the Sham group. LIPUS stimulation significantly increased the hippocampal levels of CB (0.90 ± 0.04, p < 0.05; Fig. 3C) and PV (0.87 ± 0.03, p < 0.05; Fig. 3D) but not GAD67 (0.83 ± 0.03, p = 0.95; Fig. 3B) in the MK-801 + LIPUS group as compared to the MK-801 group. LIPUS did not affect the hippocampal levels of CB (0.84 ± 0.10, p = 0.60; Fig. 3C), PV (0.96 ± 0.13, p > 1.00; Fig. 3D) and GAD67 (0.83 ± 0.09, p = 0.83; Fig. 3B) in the LIPUS group as compared to the Sham group. However, no significant changes were observed in the levels of GAD67 (0.96 ± 0.11 versus 1.09 ± 0.07, p = 0.71; Fig. 3F), CB (0.82 ± 0.10 versus 0.88 ± 0.10; p = 0.96. Figure 3G), and PV (0.78 ± 0.09 versus 0.95 ± 0.06, p = 0.06; Fig. 3H) in the mPFC of MK-801 group as compared to the sham group. Moreover, the LIPUS did not show obvious effects on GAD67 (1.02 ± 0.06, p = 0.83; Fig. 3F), CB (0.78 ± 0.11, p = 0.60; Fig. 3G) and PV levels (0.95 ± 0.11, p > 0.99; Fig. 3H) in the LIPUS group as compared to the sham groups. These findings indicated the safety and potential of LIPUS to provide improvement in GABAergic function.

Fig. 3
figure 3

LIPUS stimulation enhanced PV and CB expression in the hippocampus and mPFC of rats treated with MK-801. At 24 h after last LIPUS stimulation, all animals were sacrificed for western blot. (A) Immunoblots of GAD67, CB, and PV in the hippocampus of rats in the Sham, LIPUS, MK-801, and MK-801 + LIPUS groups. Quantitative levels of (B) GAD67, (C) CB, and (D) PV in the hippocampus. (E) Immunoblots of GAD67, CB, and PV in the mPFC of rats in the Sham, LIPUS, MK-801, and MK-801 + LIPUS groups. Quantitative levels of (F) GAD67, (G) CB, and (H) PV in the mPFC. * and † denote significantly different from Sham group and MK-801 group, respectively. (*,†, p < 0.05). n = 5 for All groups

Full size image

LIPUS Reverses MK-801-Induced Losses of PV + and CB + Cells in the Rat Cingulate Cortex

The levels of PV and CB were not changed significantly in total mPFC tissue in rats after MK-801 treatment (Fig. 3E, G, and H). To investigate the effects of MK-801 and LIPUS on PV + and CB + cell densities, we labeled PV or CB in the cingulate cortex of the mPFC and counted the numbers of PV + and CB + cells. Consistent with the previous study [20], MK-801 treatment significantly decreased the cell counts of PV + (88.9 ± 10.6 versus 191.9 ± 19.8, p < 0.01; Fig. 4C) and CB + (134.6 ± 7.5 versus 236.9 ± 20.0, p < 0.01; Fig. 4E) cells in the cingulate cortex in the MK-801 group as compared to the Sham group, whereas PV + and CB + cells in other subregions were unaffected. LIPUS significantly reversed the reductions of PV + (196.1 ± 20.8 versus 88.9 ± 10.6, p < 0.001; Fig. 4C) and CB + (228.9 ± 13.9 versus 134.6 ± 7.5, p < 0.01; Fig. 4E) cells in the MK-801 + LIPUS group as compared to the MK-801 group. Moreover, LIPUS did not affect the numbers of PV + and CB + cells in the LIPUS group as compared to the Sham group.

Fig. 4
figure 4

LIPUS attenuates MK-801-induced losses of PV + and CB + cells in the cingulate cortex of the mPFC. At 24 h after the last LIPUS stimulation, all animals were sacrificed for immunofluorescent analysis. (A) At the three subregions, cingulate cortex 1(Cg1), the prelimbic cortex (PrL) and infralimbic cortex (IL) in the medial prefrontal cortex (mPFC) of rat, the PV + and CB + cells were counted. (B) Representative immunofluorescent images of PV (green) and DAPI (blue) staining in the Sham, LIPUS, MK-801, and MK-801 + LIPUS-treated rats in the cingulate cortex of mPFC. The PV-immunoreactive neurons (arrowheads) were double-labelled by PV and DAPI. Image at high magnification showing the methods used for quantification of PV + . (C) Density of PV + cells in the mPFC. (D) Representative immunofluorescent images of CB (green) and DAPI (blue) staining in the Sham, LIPUS, MK-801, and MK-801 + LIPUS-treated rats in the cingulate cortex of mPFC. The CB-immunoreactive neurons (arrowheads) were double-labelled by CB and DAPI. Image at high magnification showing the methods used for quantification of CB + . (E) Density of CB + cells in the mPFC. Nuclei were counterstained with DAPI. Scale bars = 100 μm for all images. * and † denote significantly different from Sham group and MK-801 group, respectively. (**,††, p < 0.01; †††, p < 0.001). n = 5 for All groups

Full size image

LIPUS Reverses MK-801-Induced BDNF Variation in the Rat Brain

Previous studies have shown that impaired BDNF/TrkB signaling is involved the pathology of schizophrenia. In this study, BDNF levels were decreased (0.47 ± 0.02 versus 0.73 ± 0.05, p < 0.05; Fig. 5B) in the hippocampus in the MK-801 group as compares to the Sham group. However, LIPUS stimulation significantly increased hippocampal BDNF levels in the MK-801 + LIPUS group (0.96 ± 0.09 versus 0.47 ± 0.02, p < 0.05; Fig. 5B) as compared to the MK-801 group. The LIPUS group did not exhibit significant change (0.83 ± 0.07 versus 0.73 ± 0.05, p = 0.83; Fig. 5B) as compared to the Sham group. In contrast, MK-801 elevated the BDNF levels in the MK-801 group (0.85 ± 0.08 versus 0.50 ± 0.07, p < 0.05; Fig. 5F) as compared to the Sham group in the mPFC. This trend was reversed by LIPUS stimulation in the MK-801 + LIPUS group (0.49 ± 0.07 versus 0.85 ± 0.08, p < 0.05; Fig. 5F) as compared to the MK-801 group while the LIPUS group did not show obvious change (0.63 ± 0.12 versus 0.50 ± 0.07, p = 0.19; Fig. 5F) as compared to the Sham group. Furthermore, no significant changes of GDNF and VEGF levels were observed in the hippocampus and mPFC after MK-801 treatment, and LIPUS stimulation did not affect the levels of GDNF and VEGF (Fig. 5A, C, D, E, G, and H).

Fig. 5
figure 5

LIPUS stimulation reversed BDNF levels in the brain of rats treated with MK-801. (A) Immunoblots of BDNF, GDNF, and VEGF in the hippocampus of rats in the Sham, LIPUS, MK-801, and MK-801 + LIPUS groups. Quantitative levels of (B) BDNF, (C) GDNF, and (D) VEGF in the hippocampus. (E) Immunoblots of BDNF, GDNF, and VEGF in the mPFC of rats in the Sham, LIPUS, MK-801, and MK-801 + LIPUS groups. Quantitative levels of (F) BDNF, (G) GDNF, and (H) VEGF in the mPFC. * and † denote significantly different from Sham group and MK-801 group, respectively. (*,†, p < 0.05). n = 5 for All groups

Full size image

Discussion

Deficits in cognitive functions such as attention and memory have been reported in schizophrenia patients. Reversal learning impairment, another deficit in cognition in schizophrenia, is involved in the dysfunction of cognitive flexibility which leads to failures in adapting to altered environmental situations [4, 40, 41]. All of the schizophrenia-related symptoms lower the quality of life of patients, and because of the poor efficacy of traditional antipsychotics in treating cognitive symptoms, it is important to develop novel approaches for treating cognitive deficits in schizophrenia [2, 5]. Here, we demonstrated that TUS reversed the cognitive impairments observed in MK-801-treated rats, but did not affect the performance of saline-treated rats.

In this study, our data showed that LIPUS attenuated the schizophrenia-like behaviors and neuropathology in the MK-801 + LIPUS rats. The MK-801-treated rats exhibited significant impairment on the SOR task at 24 h but not at the immediate and 5 min retention delays. These results suggest that the MK-801 treatment induced long-term memory dysfunction (Fig. 2C, retention delay = 24 h) but not the loss of interest for novel objects (novelty-preference) in the rats (Fig. 2A, retention delay = immediate). The observation that the MK-801-treated group and sham group did not differ in swim times during the initial 4 acquisition trials in the water maze suggests that the MK-801 treatment did not significantly affect the general ability to successfully complete the water maze task (Fig. 2D). Furthermore, the MK-801-treated rats did not reveal a deficit in the ability to complete the initial task by reaching the hidden platform in the water maze test, but they did show a perseveration behavior by rounded in the old target quadrant (time spent shown in Suppl. Figure 1) and increased swim times in the first trial after platform relocation. LIPUS improve this perseveration behavior in the MK-801 + LIPUS group (Fig. 2E) and did not affect the behavior in the LIPUS group. Furthermore, the LIPUS group did not show abnormal during training on the total trials in the water maze. Our results revealed that LIPUS significantly reversed this memory dysfunction and the perseveration behavior but did not affect the performance of saline-treated rats (Fig. 2E).

Recently, numerous reports have stated a hypothesis that NMDA receptor dysfunction impairs GABAergic interneurons, resulting in an excitatory and inhibitory (E/I) imbalance that contributes to schizophrenia-like behaviors [42,43,44]. The NMDA receptor blocker MK-801 not only induces neuropathological changes such as GABAergic interneuron, PV, and CB level reductions in the hippocampus and PFC but also causes cognitive symptoms, positive symptoms, and negative symptoms in animals that support this hypothesis of schizophrenia [20, 45, 46]. To investigate how LIPUS attenuates MK-801-induced cognitive impairments and the potential of LIPUS for treating positive symptoms or negative symptoms, it is necessary to examine whether LIPUS ameliorates MK-801-induced reductions of GABAergic interneurons. We chose the hippocampus and PFC for histochemical analysis because these two brain regions are most implicated in schizophrenia [47].

Our data revealed that subchronic MK-801 treatment reduced the levels of PV and CB but not GAD67 in the hippocampus and that LIPUS reversed these reductions in the MK-801 + LIPUS rats (Fig. 3A–D). However, PV, CB, and GAD67 levels in the mPFC were not affected significantly after MK-801 treatment (Fig. 3E–H). Due to both cognitive symptoms and negative symptoms being associated with E/I imbalance in the PFC [48,49,50], we examined whether subchronic MK-801 treatment impaired PV + and CB + cells in the mPFC via immunofluorescence. We found that the cell densities of PV + and CB + cells were significantly decreased in the cingulate cortex of the mPFC after subchronic MK-801 treatment and that LIPUS reversed these reductions (Fig. 4C, E). According to the hypothesis of schizophrenia, GABAergic interneuron dysfunction results in E/I imbalance that causes the symptoms of schizophrenia. From this point of view, the fact that LIPUS restores the cell densities of PV + and CB + cells in the cingulate cortex of the mPFC indicates the potential of LIPUS to treat not only the cognitive symptoms but also the positive symptoms and negative symptoms of schizophrenia. In this study, CB + interneurons were not differentiated from CB + cells because of weakly expression of CB in pyramidal neurons in prefrontal cortex. However, CB + pyramidal neurons also play important role in spatial memory. It is worthwhile to investigate the potential of LIPUS in the treatment of cognitive impairment induced by CB + pyramidal neurons dysfunction [51, 52].

There are accumulating evidences that BDNF may be involved in the pathophysiology of patients with schizophrenia [32, 53]. In the present study, subchronic MK-801 treatment decreased BDNF levels in the hippocampus but increased the BDNF levels in the mPFC (Fig. 5B, F). These inconsistencies may reflect differences in BDNF responses dependent on brain region in the animal model. Although decreased BDNF level may play a role in the generation of schizophrenia-like behaviors, elevated BDNF level also was found in the cingulate cortex of partial schizophrenia patients while overexpression of BDNF also induced behavior deficits in animals [54, 55]. In vivo and in vitro studies have shown that LIPUS stimulation increases BDNF expression in glial cells via TrkB-Akt signaling pathways [56,57,58]. Interestingly, in the present study, LIPUS simultaneously reversed the MK-801-induced decreased level of BDNF in hippocampus and the increased level of BDNF in prefrontal cortex. These results indicate the potential of LIPUS in the treatment of the neural disorders which are associated with alteration (including overexpression and downregulation) of BDNF expression (Fig. 5B, F).

In conclusion, the present study demonstrated that LIPUS stimulation significantly alleviated MK-801-induced behavioral deficits. The beneficial effects of LIPUS may be attributed partially to its enhancement of PV and CB expressions and rebalancing of BDNF levels in the hippocampus and mPFC (Fig. 6). More studies should be conducted to confirm the exact therapeutic mechanisms of LIPUS and to clarify the parameters of LIPUS that could enhance its treatment efficacy. These findings indicate that TUS may represent a novel therapeutic approach for schizophrenia.

Fig. 6
figure 6

The schematic diagram indicates the potential mechanisms of LIPUS stimulation-induced treatment for schizophrenia. LIPUS can (A) rebalance the BDNF levels and (B) enhance the expressions of PV and CB

Full size image