Modulation of Hippocampal Gamma Oscillations by Dopamine in Heterozygous Reeler Mice in vitro

The reelin haploinsufficient heterozygous reeler mice (HRM), an animal model of schizophrenia, have altered mesolimbic dopaminergic pathways and share similar neurochemical and behavioral properties with patients with schizophrenia. Dysfunctional neural circuitry with impaired gamma (γ) oscillation (30–80 Hz) has been implicated in abnormal cognition in patients with schizophrenia. However, the function of neural circuitry in terms of γ oscillation and its modulation by dopamine (DA) has not been reported in HRM. In this study, first, we recorded γ oscillations in CA3 from wild-type mice (WTM) and HRM hippocampal slices, and we studied the effects of DA on γ oscillations. We found that there was no difference in γ power between WTM and HRM and that DA increased γ power of WTM but not HRM, suggesting that DA modulations of network oscillations in HRM are impaired. Second, we found that N-methyl-D-aspartate receptor (NMDAR) antagonist MK-801 itself increased γ power and occluded DA-mediated enhancement of γ power in WTM but partially restored DA modulation of γ oscillations in HRM. Third, inhibition of phosphatidylinositol 3-kinase (PI3K), a downstream molecule of NMDAR, increased γ power and blocked the effects of DA on γ oscillation in WTM and had no significant effect on γ power but largely restored DA modulation of γ oscillations in HRM. Our results reveal that impaired DA function in HRM is associated with dysregulated NMDAR–PI3K signaling, a mechanism that may be relevant in the pathology of schizophrenia.

The reelin haploinsufficient heterozygous reeler mice (HRM), an animal model of schizophrenia, have altered mesolimbic dopaminergic pathways and share similar neurochemical and behavioral properties with patients with schizophrenia. Dysfunctional neural circuitry with impaired gamma (γ) oscillation (30-80 Hz) has been implicated in abnormal cognition in patients with schizophrenia. However, the function of neural circuitry in terms of γ oscillation and its modulation by dopamine (DA) has not been reported in HRM. In this study, first, we recorded γ oscillations in CA3 from wildtype mice (WTM) and HRM hippocampal slices, and we studied the effects of DA on γ oscillations. We found that there was no difference in γ power between WTM and HRM and that DA increased γ power of WTM but not HRM, suggesting that DA modulations of network oscillations in HRM are impaired. Second, we found that N-methyl-D-aspartate receptor (NMDAR) antagonist MK-801 itself increased γ power and occluded DA-mediated enhancement of γ power in WTM but partially restored DA modulation of γ oscillations in HRM. Third, inhibition of phosphatidylinositol 3-kinase (PI3K), a downstream molecule of NMDAR, increased γ power and blocked the effects of DA on γ oscillation in WTM and had no significant effect on γ power but largely restored DA modulation of γ oscillations in HRM. Our results reveal that impaired DA function in HRM is associated with dysregulated NMDAR-PI3K signaling, a mechanism that may be relevant in the pathology of schizophrenia.

INTRODUCTION
Reelin, a glycoprotein of the extracellular matrix, controls cell migration and layering in the developing brain, promotes the formation of synaptic circuits, and regulates synaptic transmission and plasticity in the postnatal and adult brain (Campo et al., 2009;Hwa and Gabriella, 2016). During development, reelin is expressed by the Cajal-Retzius cells in the hippocampus and cortex and granule cells in the cerebellum, whereas in the adult brain, reelin is secreted by GABAergic interneurons in the cortex and hippocampus (Ogawa et al., 1995;Knuesel, 2010;Brosda et al., 2011;Yuki et al., 2015).
The hippocampal CA3 region plays a specific role in memory processes (Cherubini and Miles, 2015) and attention (Vinogradova, 2001;Bygrave et al., 2019) and controls dopamine (DA) release by forming a functional circuit in the ventral tegmental area (Luo et al., 2011). Extensive recurrent axon collaterals of CA3 pyramidal neurons connected with neighboring neurons, including GABAergic interneurons, compose a local circuit, and the interaction of pyramidal neurons and interneurons within the circuit generates synchronized activity, such as gamma (γ) frequency oscillations (30-80 Hz) (Traub and Wong, 1982;Bartos et al., 2007). γ oscillations are able to synchronize local, inter-region, or long-range neuronal activity and to promote information exchanges between neurons (Colgin, 2011;Fries, 2015) and are associated with higher brain function, such as attention, perceptual binding, learning, and memory (Womelsdorf and Fries, 2007;Buzsaki and Wang, 2012).
Schizophrenia has been suggested to be caused by the failure of integrating local and distributed neural circuits (Andreasen, 2000;Lee et al., 2003;Gallinat et al., 2004;Spencer, 2011;Jadi et al., 2016). In fact, studies have found that abnormal γ oscillations are associated with multiple symptoms of schizophrenia, such as hallucinations and delusion (Lee et al., 2003;Spencer et al., 2004). Schizophrenia is known to be associated with altered DA level (Winterer and Weinberger, 2004;Toda and Abi-Dargham, 2007), which influences information processes underlying cognitive process and may contribute to abnormal γ oscillations observed in patients with schizophrenia.
As an animal model of schizophrenia, HRM shows abnormal dopaminergic function, including reduced DA D1 and D2 receptors (D1R and D2R) in the striatum, reduced D1R-and D2R-mediated locomotor response (Matsuzaki et al., 2007), and increased expression of D2R in the striatum (Varela et al., 2015); and it also shows altered dopaminergic fiber densities in different brain areas, such as increase in the densities of tyrosine hydroxylase-immunoreactive (TH-IR) neurons in the hippocampus but decrease TH-IR neurons in the shell of the nucleus accumbens (Nullmeier et al., 2014). DA modulates fast network oscillations in the γ frequency band of rat hippocampus (Andersson et al., 2012) and beta frequency band of the mouse anterior cingulate cortex (Steullet et al., 2014); however, little is known about DA modulation of network oscillations in HRM.
Dopamine modulation of γ oscillations in rat hippocampus is involved in N-methyl-D-aspartate receptor (NMDAR)-dependent mechanism (Andersson et al., 2012). Methamphetamine, a psychostimulant, known to induce a strong DA release, enhances γ oscillations recorded in rat hippocampal slices also involved in NMDAR activation (Li et al., 2019). Studies have demonstrated that NMDAR is dysfunctional in schizophrenia. The cortical hyperexcitability and reduced function of NMDAR in parvalbumin-expressing inhibitory interneurons in schizophrenia are associated with increased γ activity (Spencer, 2011). A single dose of an application of the NMDAR antagonist MK-801 induces psychotic symptoms in humans and schizophrenia-like phenotype in animals, increases peak power, and reduces peak frequency of γ oscillations (Carlen et al., 2012;Lemercier et al., 2017).
Reelin increases NMDAR-dependent synaptic transmission and plasticity in the postnatal hippocampus (Qiu et al., 2006). Reelin deficiency causes increased expression of NR2A and NR2B of NMDAR subunits in the hippocampus from HRM (Isosaka et al., 2006). Blocking reelin secretion rapidly changes the subunit composition of NMDAR to a predominance of NR2B-containing NMDAR in cultured hippocampal neurons (Campo et al., 2009). The altered expression of NMDAR subunits may contribute to the modulation of network oscillations of HRM.
By binding to apolipoprotein E receptor 2 and very-lowdensity lipoprotein receptor (ApoER2/VLDLR), reelin activates different signaling cascades, one of which is phosphatidylinositol 3-kinase (PI3K) signaling pathway, and increases synaptic transmission by enhancing PI3K-dependent postsynaptic AMPAR insertion (Qiu et al., 2006;Ishii et al., 2016). PI3K is one of the downstream molecules in NMDAR activation, in which calcium influx through the NR2B subunit of NMDAR leads to the activation of PI3K (Brennan-Minnella et al., 2013). A previous study shows that nicotinic modulation of hippocampal γ oscillations involves PI3K activation (Wang et al., 2017). These studies indicate that PI3K may be involved in the modulation of γ oscillations in HRM.
In this study, we investigated γ oscillation and its modulation by DA in HRM using extracellular field potential recording to determine whether there are altered dopaminergic modulations of γ oscillation in HRM and the possible mechanisms associated with it.

Electrophysiological Recording, Data Acquisition, and Statistical Analysis
The hippocampal slices were maintained at a temperature of 32 • C at the interface between the ACSF and warm humidified carbogen and allowed to equilibrate in this medium for 1 h prior to recording. Extracellular field potentials were recorded from the stratum pyramidale of Cornu ammonis 3c (CA3c) of the hippocampus, using glass microelectrodes containing ACSF (resistance, 2-5 M ). Field potentials were amplified using NeuroLog NL106 AC/DC amplifiers (Digitimer Ltd., Welwyn Garden City, United Kingdom) and band-pass filtered between 0.5 and 500 Hz using NeuroLog NL125 filters (Digitimer Ltd., Welwyn Garden City, United Kingdom). Electromagnetic interference from the main supply was eliminated from the recordings with the use of HumBug 50-Hz noise eliminators (Digitimer Ltd., Welwyn Garden City, United Kingdom). The recordings were digitized at a sample rate of 2 kHz using a CED 1401-plus ADC board (Cambridge Electronic Design, Cambridge, United Kingdom).
Data were analyzed offline using the Spike2 software (Cambridge Electronic Design). Power spectra were generated to provide a quantitative measure of the frequency components. Power spectra were constructed for 60-s epochs using a fast Fourier transform algorithm.
It has been widely accepted that in vitro γ oscillations ranged from 20 to 80 Hz, because the recorded γ oscillations in brain slices are temperature dependent and the slice recordings performed mostly at 32 • C rather than 37 • C. There is a linear relationship between peak frequency of network oscillations and temperature, in which an increase of 1 • C in temperature of brain slices corresponds to an increase of 2.3 ± 0.4 Hz in the oscillation frequency (Dickinson et al., 2003;Lu et al., 2012).
The area under the curve between 20 and 60 Hz was used to quantify the γ power. Autocorrelograms were calculated in Spike2 using a 500-ms lag from the same local field potential trace used for γ power calculation. The decay time constant (tau) of the autocorrelation peaks is a measure of the regularity of the oscillation and generated by fitting the autocorrelation peaks with an exponential function: Y = exp(−a * X).

Statistical Analysis
All statistical analyses were performed using IBM SPSS Statistics 22 software (IBM, Armonk, NY, United States). The Shapiro-Wilk test was used in testing the normality of the data. Parametric data were expressed as mean ± standard error of the mean. The paired and unpaired Student's t-tests were used to compare two groups of parametric data. One-way analysis of variance (ANOVA) and repeated-measures (RM) ANOVA were used to compare three or more group means. Non-parametric data were expressed as median ± interquartile range. The Wilcoxon ranksum and signed-rank tests were used to compare the two groups of non-parametric data. One-way and RM ANOVAs on ranks were used for three or more group comparisons. The parametric two-way ANOVA was used to analyze experimental data derived from two-factor designs with or without RM. The two-way ANOVA on ranks was used to analyze non-parametric data. A P-value < 0.05 was considered statistically significant.

Dopamine Increased Gamma Power in Wild-Type Mice but Not in Heterozygous Reeler Mice
Defective reelin signaling influences the mesolimbic dopaminergic pathways (Pujadas et al., 2014). Thus, we tested whether DA modulation of γ oscillations was altered in HRM. After stable γ oscillations were induced by carbachol in hippocampal CA3 for at least 30 min, 20 µM of DA was applied. In WTM, DA increased the γ power by 53.8 ± 11.5% of the control [CCh + DA, 1,095.24 (586.52, 3,932.55) vs.

Wortmannin Increased Gamma Power and Largely Blocked Dopamine-Mediated Increase in Gamma Power in Wild-Type Mice
Previous studies indicate that reelin acts on its receptor and activates the PI3K-Akt-mammalian target of rapamycin (mTOR) pathway (Hwa and Gabriella, 2016). Therefore, we examined the effect of wortmannin, a potent and selective inhibitor of PI3K, at a physiological dose (Wang et al., 2017) on γ oscillations of rat hippocampal slices from WTM and HRM. When wortmannin was applied to hippocampal slices, γ power was significantly increased by 39 ± 12% in WTM (CCh + Wort, 973.39 ± 252.78 µV 2 vs. CCh, 715.89 ± 175.34 µV 2 , F (2,9) = 9.908, P = 0.001, RM ANOVA, Figures 4A,B,E), and a further application of DA (20 µM) caused an additional 23 ± 7% increase in γ power, but such an increase did not reach statistical significance compared with that in wortmannin (1,150.03 ± 273.81 µV 2 vs. CCh + Wort, T = 1.801, P = 0.09, RM ANOVA, followed by the Holm-Sidak method, Figures 4A,B,E). These results indicate that wortmannin largely blocked DA-mediated enhancement of γ power in WTM. Neither wortmannin nor wortmannin + DA had any effect on peak frequency of γ oscillations in WTM ( Figure 4F).

DISCUSSION
Our main findings are as follows: (1) DA enhanced γ power in WTM but not in HRM.
(2) MK-801 induced a larger increase in γ power, occluded the effect of DA in WTM, induced a small increase in γ power, and partially restored the effect of DA in HRM. (3) Wortmannin induced a larger increase in γ power, blocked the effect of DA in WTM, and caused no significant increase in γ power but largely restored the effect of DA in HRM.

Altered Dopamine Modulation of Hippocampal Gamma Oscillation in Heterozygous Reeler Mice
Dopamine at a concentration of 20 or 200 µM increased γ power in hippocampal slices in WTM, which differs from the observation that DA at a concentration of 200 µM reduced γ oscillations induced by carbachol in area CA3 of rat hippocampus (Weiss et al., 2003), suggesting that species difference may exist in DA modulation of γ oscillations.
In HRM, we demonstrated that in vitro hippocampal γ oscillation was intact in HRM but that DA modulation of γ oscillations was impaired. Loss of sensitivity to DA for hippocampal γ power in HRM may be related to altered densities of dopaminergic fibers in different brain areas: increased in the hippocampus but reduced in the ventral tegmental area and nucleus accumbens in HRM (Nullmeier et al., 2014). The expression profile of DA receptors in the hippocampus of HRM is relatively sparse but decreased D1 and D2 receptors in the striatum (Matsuzaki et al., 2007) or altered expression pattern of D2R in different brain areas: An increased expression in the striatum but decreased expression in the frontal cortex (Varela et al., 2015) was reported.

Receptor Partially Restores Dopamine Sensitivity in Heterozygous Reeler Mice
Clinical symptoms of schizophrenia are associated with altered cortical neuronal oscillations in γ rhythms. NMDAR antagonists induce psychotic symptoms in humans and a schizophrenia-like phenotype in animals (Spencer, 2011;Jadi et al., 2016). In this study, NMDAR antagonist increased γ power in the hippocampal slices of WTM, which is in agreement with a previous study that a single application of the NMDAR antagonist MK-801 in rats increased the power and reduced the peak frequency of γ oscillations (Lemercier et al., 2017).
Compared with WTM, the same dose of MK-801 caused a relative small increase in γ power in HRM, which may be related to the possible alteration in the composition of NMDAR subunits in the hippocampus. It was reported that blocking reelin secretion increases the NR2B subunit in cultured hippocampal neurons (Campo et al., 2009). HRM also showed increased NR1 but reduced NR2C in the frontal cortex (van den Buuse et al., 2012). Additionally, during neural maturation, a marked decrease in NR1/NR2B receptor participation to NMDAR-mediated synaptic currents concomitant with the accumulation of reelin at active synapse was observed in cultured hippocampal neurons, suggesting that reelin regulates NMDAR surface trafficking and synaptic subunit composition (Sinagra et al., 2005;Groc et al., 2007). Reelin also regulates NMDAR function via increased tyrosine phosphorylation of NR2A and NR2B receptors and increases NMDAR-mediated synaptic plasticity in the adult hippocampus (Qiu et al., 2006). These studies indicate that sufficient reelin is required to control the subunit composition and function of NMDAR in hippocampal neurons and that reelin deficiency causes altered composition and reduced function of NMDAR, which will likely contribute to the altered response of γ oscillation to MK-801.
In WTM, MK-801 occluded DA-mediated increase in γ power, indicating that DA enhancement of γ oscillation is through NMDAR activation. This is in agreement with previous reports that DA-mediated (Andersson et al., 2012), nicotine-mediated (Wang et al., 2017), and methamphetamine-mediated (Li et al., 2019) increase in γ oscillation in rat hippocampus are all involved in NMDAR activation.
In HRM, MK-801 partially restored DA-mediated response of γ oscillation. The explanation for this result could be that blunted DA modulation of γ oscillations by overactivation of NMDAR in HRM may be attenuated by MK-801, as observed in the case that intensive NMDAR activation mediated nicotine (100 µM) inhibition of γ oscillations (Wang et al., 2017). However, reduced NMDAR-dependent synaptic long-term potentiation in HRM (Iafrati et al., 2014) suggests that NMDAR activity may be at a relative low level in HRM. Although detailed mechanisms for the partial restoration of DA enhancement of γ power remain to be further studied, our results are supported by the observation that NMDAR antagonists, ketamine or Ro25-6981 (selective inhibitor of GluN2B), restored synaptic and memory function in HRM (Iafrati et al., 2014). The similar roles between Ro25-6981 and ketamine in HRM imply that correcting NMDAR composition from an immature form (GluN2B) to mature form (GluN2A) is important in recovering normal synaptic transmission in HRM. One study showed that MK-801 altered subunits of NMDAR in the young adult rat prefrontal cortex (Xi et al., 2009), although it is not known whether MK-801 affects the composition of NMDARs in the hippocampus in HRM. Interestingly, MK-801 and ketamine not only alter NMDAR composition but also have a partial agonist effect on D2 receptor (Kapur and Seeman, 2002), which may be critical in DA modulation of γ power, especially in HRM.
Dopamine alone increased the peak frequency of oscillatory activity in HRM. This effect was reversed, whereas DA effect on γ power was partially restored in the presence of MK-801, which suggests that NMDAR activation is required for DA-mediated oscillatory frequency.

Blocking Phosphatidylinositol 3-Kinase Largely Restores the Dopamine Sensitivity in Heterozygous Reeler Mice
Similar to the effects of MK-801 on γ oscillations, wortmannin, a PI3K inhibitor, caused a substantial increase in γ power in WTM and a small, insignificant increase in HRM, indicating that the endogenous PI3K activity is different between WTM and HRM and that sensitivity of γ oscillations to PI3K activity is reduced in HRM.
In WTM, wortmannin was able to occlude DA enhancement of γ power, which indicates that PI3K is also involved in DA modulation of γ oscillation. In HRM, DA-mediated response was largely increased in the presence of wortmannin. Our result is in agreement with the report that blocking NMDAR and its downstream signaling molecule, the mTOR, rescued the deficit of function and behavior in HRM (Iafrati et al., 2014). Studies also demonstrated that reelin, acting through the PI3K, positively modulates the activity of mTOR kinase, which is required in the stimulation of dendrite outgrowth, and activates downstream proteins, such as the p70S6K, which are known to participate in the control of protein translation (Jossin and Goffinet, 2007;Ventruti et al., 2011). Because PI3K is an upstream signaling molecule of mTOR (Lussier et al., 2016) and a downstream molecule of NMDAR (Perkinton et al., 2002;Man et al., 2003;Crossthwaite et al., 2004), it is reasonable to assume that the restoration of DA enhancement of hippocampal γ oscillations in HRM in the presence of wortmannin is likely through inhibition of the NMDAR-PI3K signaling pathway.
As reelin activates PI3K (Beffert et al., 2002) and enhances synaptic transmission via PI3K-dependent synaptic insertion of AMPARs in adult hippocampus (Qiu et al., 2006), HRM with remarkable reelin deficiency may have a low level PI3K activity, which may explain the blunted response of γ oscillations to PI3K inhibitor. However, it is unclear how DA modulation of γ oscillations is largely recovered in the presence of a PI3K inhibitor in HRM. Although the mechanism of this observation remains to be further determined, our results with respect to the large restoration of DA sensitivity in the presence of PI3K inhibitor nevertheless indicate a possibility on how to correct abnormality in DA function in HRM.
The results of this study demonstrated that the altered DA modulation of γ oscillations in HRM is associated with dysregulated NMDAR-PI3K signaling, establishing a link between DA-and NMDAR-mediated signaling, network oscillations, and reelin, which might be relevant to the field of schizophrenia research.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The animal study was reviewed and approved by the Animal Experimentation Ethics Committee of Xinxiang Medical University (protocol number: 11401300017419). All experiments were performed in accordance with the guidelines of the Animal Care and Use Committee of Xinxiang Medical University.

AUTHOR CONTRIBUTIONS
LW, DZ, MW, and YW performed the research. CL and JL designed the research. CL, LW, DZ, and MV analyzed the data. LW, DZ, and CL wrote the manuscript. CL, LW, and MV revised the manuscript. All authors approved the final manuscript for publication.

FUNDING
This study was supported by the National Natural Science Foundation of China (Nos. 81771517 and 81271422) and the Key Science and Technology Project of Henan (No. 182102310209).