All experiments and procedures used in this study were approved by the University at Buffalo Animal Care and Use Committee in accordance with the National Institutes of Health Guideline for the Care and Use of Laboratory Animals. Animal breeding and PE treatment were performed as previously described [27]. The electrophysiological, behavioral and biochemical studies were conducted in distinct cohorts of control and PE male rats. To mitigate the potential litter effects, rats from different litters (2 rats/litter) were used for each experiments.

Male rats (8–10 weeks) underwent two different behavioral tests: Elevated plus maze and zero maze as described previously [27]. The plus maze videos were manually coded and zero maze videos were analyzed with the ANY-maze 6.3 software (Stoelting Co., Wood Dale, IL, USA).

Brain slices containing the DRn were obtained from control and PE rats (8 to 10 weeks) as previously described [28]. Following recovery at room temperature for at least 1 h, slices were transferred to a recording chamber (Warner Instruments, Hamden, CT, USA) mounted on a fixed upright microscope and continuously perfused (2 to 3 ml/min) with artificial cerebrospinal fluid (ACSF) of the following composition (in mM): 119 NaCl; 2.5 KCl; 2.5 CaCl₂; 1.3 MgSO₄; 1 NaH₂PO₄; 26.2 NaHCO₃; 11 glucose and continuously bubbled with a mixture of 95% O₂/5% CO₂ at 30 ± 1°C. Somatic cell-attached and whole-cell recordings were performed from dorsal medial (dmDRn) and ventral medial (vmDRn) subdivisions of the DRn using potassium-gluconate based intracellular solution (in mM): 120 potassium gluconate; 10 KCl, 10 Na₂ -phosphocreatine, 10 HEPES, 1 MgCl₂, 1 EGTA, 2Na₂-ATP, and 0.25 Na-GTP; pH 7.3; osmolality 280–290 mOsm/kg. Serotonin neurons were identified by their unique electrical properties, including slow firing activity, large after-hyperpolarization and 5-HT_1A receptor-induced membrane hyperpolarization [28, 29]. Excitatory postsynaptic currents (EPSCs) were evoked by single square-pulses (duration = 100–200 μs) delivered at 0.1 Hz. To access paired-pulse ratio (PPR), a pair of EPSCs were evoked with an inter-stimulus interval of 50 ms. The intensity of the stimulus was adjusted to evoke 75 % of the maximal amplitude of EPSCs. α-amino-3-hydroxy-5-methyl-4- isoxazolepropionic acid (AMPA) receptor-mediated EPSCs were recorded from neurons voltage clamped at −70 mV in the presence of a GABA_A receptor antagonist picrotoxin (100 μM). The cell input resistance and access resistance (10 – 20 mΩ) were monitored throughout the experiment using 5 mV hyperpolarizing voltage steps (500 ms duration). Recordings were discarded when the input and series resistance changed by more than 10–20%. To isolate miniature action potential independent EPSCs (mEPSCs), tetrodotoxin (TTX, 1 µM) was added to Ringer’s solution in the bath. All data were acquired using the pClamp 10.2 software (Molecular Devices, Union City, CA, USA).

The levels of 2-arachydonyl glycerol (2-AG) and anandamide (AEA) in rat DRn slices were quantified using mass spectrometry and deuterated internal 2-AG and AEA standards as described previously [30, 31].

RNA was extracted from brainstem slices using the RNeasy mini kit (Qiagen) and cDNA synthesis was carried out using the SuperScript III First Strand synthesis system (ThermoFisher). qPCR was performed using PowerUp SYBR green (ThermoFisher) on a StepOnePlus real-time PCR system (Applied Biosystems). The following primers were used: CB1R, Forward: AAGTCGATCCTAGATGGCCTT and Reverse: TCCTAATTTGGATGCCATGTCTC; Actin, Forward: GAGGCTCTCTTCCAGCCTTC and Reverse: CGGATGTCAACGTCACACTT. Quantification was performed using the 2-ΔΔCt method with actin serving as the housekeeping gene. Control reactions lacking template cDNA were included in each run. qPCR data were analyzed by comparing ΔCt values using an unpaired T-test.

Evoked excitatory postsynaptic currents (eEPSCs) were analyzed using Clampfit 10.2 software (Molecular Devices, Union City, CA, USA). The amplitude of eEPSCs is determined by measuring the average current during a 2 ms at the peak of each eEPSC, subtracted from the baseline current and normalized to the mean baseline amplitude recorded for at least 5 min prior drug application. The PPRs (eEPSC₂/eEPSC₁) were averaged for at least 60 trials. Miniature excitatory postsynaptic currents (mEPSCs) were analyzed with Mini Analysis Software (Synaptosoft, Decatur, GA). The synaptic events were selected using amplitude threshold (5 pA), rise time (1 ms), and area threshold (30 fC). All selected events were further visually inspected to prevent noise from compromising the analysis. The decay kinetic of mEPSCs was fitted with mono-exponential decay of the following equation: y = A₀ + Ae^kt, where A₀ is the baseline current, A is the peak current and k is the decay time constant. The data distribution was tested for normality using Shapiro–Wilk test. Parametric (independent and paired t-tests), and non-parametric tests (Mann-Whitney test) were used for normally and not normally distributed data, respectively. For comparison between groups, the independent t-test and analysis of variance (ANOVA) using post hoc Tukey tests were used. For mEPSC comparison, the Kolmogorov–Smirnov tests (K–S tests) were used. Statistical analysis was performed with Origin 9.0 software (MicroCal Software Inc., Northampton, MA, United States) and significance was set at α = 0.05. The results in the text and figures are expressed as mean ± SEM.

Chemicals and salts were obtained from Fisher Scientific (Pittsburgh, PA, USA). N-(Piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide (AM 251), Picrotoxin purchased from Tocris Biosciences (Minneapolis, MN, USA), S-Nitroso-N-acetyl-DL-penicillamine (SNAP), Nω-Nitro-L-arginine methyl ester hydrochloride (L-NAME), 2-(4-Carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (carboxy-PTIO) and 8-(4-Chloorophenylthio)-guanosine 3’, 5’-cyclic monophosphate (8 pCPT-cGMP) were purchased from Sigma-Aldrich (Saint Louis, MO, USA).

The impact of PE on anxiety-like behaviors was tested in the elevated plus maze (EPM) and zero maze tests (ZM). In the EPM test, compared to control, PE rats spent less time in open arms (Control: 122.4 ± 17 s; PE: 63.63 ± 14.55 s; Fig. 1A₁), and more time in closed arms (control: 180.5 ± 17.9 s; PE: 245.6 ± 18.5 s; Fig. 1A₂), with no significant difference in open arms (control: 19.0 ± 3.08; PE: 17.00 ± 1.87; Fig. 1A₃), and closed arms entries (control: 10.75 ± 2.87; PE: 10.75 ± 0.75; Fig. 1A₃), indicating that PE increases anxiety-like behaviors without altering locomotor activity. In ZM test, no differences in open quadrant time were observed (control: 81.74 ± 6.17 s, n = 20; PE = 78.22 ± 7.45 s; Fig. 1B₁) between cohort. However, PE rats made fewer entries into open quadrants than control (control: 16.00 ± 1.46; PE: 11.40 ± 1.2; Fig. 1B₂) and spent more time on quadrants they were initially on, reluctant to venture toward farther quadrants, as shown by the representative tracking plots (Fig. 1B₃). Collectively, these results indicate that PE increases anxiety-like behavior in adult rats.

Serotonin neuron of the DRn are clustered in various groups, including the dmDRn and vmDRn subdivisions, which project extensively to medial prefrontal cortex (mPFC) and central amygdala (CeA), and play a key role in regulating anxiety-like behaviors [32, 33]. Therefore, we examined whether PE alters the function of dmDRn and vmDRn 5-HT neurons. To that end, we first conducted cell-attached recordings, which revealed that in control rats, dmDRn and vmDRN 5-HT neurons were not spontaneously active (0.01 ± 0.005 Hz; Fig. 2A₁-A₂), which was consistent with previous reports [28, 34]. In contrast, in PE rats, most dmDRn and vmDRN 5-HT neurons exhibited spontaneous firing activity (0.31 ± 0.05 Hz; Fig. 2A₁-A₂), demonstrating that PE increases the excitability of adult DRn 5-HT neurons. We next assessed the impact of PE on the intrinsic electrical properties DRn 5-HT neurons and found that PE had no effects on cell input resistance (PC: 1.30 ± 0.12 GΩ; PE: 1.05 ± 0.11 GΩ; n = 8; t-test, t ₍₇₎ = −1.52, p > 0.05), membrane capacitance and evoked firing activity. These results indicate that the PE-induced increase in the excitability of DRn 5-HT neurons is not mediated by alterations of their intrinsic electric properties.

The firing activity of DRn 5-HT neurons is tightly regulated by the strength of their excitatory inputs [35, 36]. Consequently, the increased activity of DRn 5-HT neurons in PE rats could be mediated by a potentiation of DRn glutamate synapses. We first tested this possibility by examining whether blockade of glutamatergic transmission can reverse the increased activity of 5-HT neurons in PE rats. We found that administration of DNQX (20 µM) and AP-5 (20 µM), antagonists of α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor (AMPAR) and N-methyl-D-aspartate receptor (NMDAR), respectively, significantly reduced the PE-induced activation of 5-HT neurons (PE: 0.37 ± 0.04 Hz; PE + DNQX & AP5: 0.16 ± 0.02 Hz; Fig. 2B₁-B₂). Such a finding indicates that a persistent potentiation of DRn glutamate synapses contributes to the PE-induced activation of 5-HT neurons.

Next, to directly examine the impact of PE on glutamate synapses, we assessed the probability of glutamate release using the paired-pulse ratio (PPR) and coefficient of variation (CV) of eEPSCs. Compared to control, PE significantly reduced the PPR (Control: 1.30 ± 0.03; PE: 1.13 ± 0.02; Fig. 3A₁-A₂) and CV of eEPSCs (Control: 0.47 ± 0.02; PE: 0.33 ± 0.01; Fig. 3A₃), thereby indicating persistent potentiation of glutamate release in the DRn. Consistent with this conclusion, we also found that compared to control rats, PE enhanced the average mEPSC frequency (Control: 4.56 ± 0.44 Hz; PE: 9.77 ± 0.67 Hz; Fig. 3B₁-B₂) as indicated by the leftward shift of the cumulative distribution of mEPSC frequency (K-S test: p < 0.05, Fig. 3B₃). In contrast, PE did not affect the average amplitude (Control: 13.53 ± 1.09 pA; PE: 14.38 ± 1.63 pA; Fig. 3B₄), the cumulative distribution (K-S test: p > 0.05, Fig. 3B₅) and decay kinetic of mEPSCs (Tau control: 5.12 ± 0.36 ms; Tau PE: 5.59 ± 0.31 ms; Fig. 3C₁-C₂), indicating the PE did not alter the function of postsynaptic AMPARs. Taken together, these results indicate that the PE-induced increased electrical activity of 5-HT neurons, is in large part, mediated by a persistent potentiation of glutamate release in the DRn.

The strength of DRn glutamate synapses is regulated by various signaling molecules, including nitric oxide (NO) and endocannabinoids (eCBs) signaling. We have previously shown that NO and eCBs retrogradely control the strength of DRn glutamate synapses [37, 38]. Thus an increase in NO and eCBs signaling induce long-term potentiation and depression of glutamate synapses in the DRn, respectively [37–39]. Because PE alters the function of neuronal nitric oxide synthase (nNOS) function [40] and enhances the formation of reactive nitrogen species (RNS), including peroxynitrite (ONOO^-) [41, 42], we hypothesized that an enhanced NO signaling could mediate the PE-induced potentiation of DRn glutamate synapses. We first tested this hypothesis by examining whether reducing tonic NO signaling using NOS inhibitors and NO scavengers can reverse the potentiation of glutamate synapses in PE rats. Remarkably, inhibition of NOS with L-NAME (100 µM) significantly reduced the amplitude of eEPSCs only in DRn slices of PE rats (Control: 108.2 ± 1.98 % of baseline; PE: 58.21 ± 5.5 % of baseline; p < 0.05, ɳ²_P = 0.73, PE vs PC; Fig. 4A₁). Similarly, bath application of the NO scavenger PTIO (100 µM) significantly reduced the amplitude of eEPSCs in PE, but not control rats (Control: 102.36 ± 7.67% of baseline; PE: 70.85 ± 5.92 % of baseline; p < 0.05, ɳ²_P = 0.42, PE vs PC; Fig. 4A₂). These observations suggest that an increase in tonic NO signaling contributes to the PE-induced potentiation of DRn glutamate synapses. Consistent with this notion, reducing NO function in the DRn with PTIO (100 µM) inhibited the frequency and the amplitude of mEPSCs in PE (PE: 13.53 ± 0.55 Hz; PE + PTIO: 5.85 ± 0.65; Fig. 4B₁-B₂), but not in control rats (Control: 5.68 ± 0.41 Hz; Control + PTIO: 5.88 ± 0.78 Hz; Fig. 4C₁-C₂). Collectively, these results indicate that an enhanced tonic NO signaling mediates the PE-induced potentiation of DRn glutamate synapses.

In principle, if the PE-induced potentiation of DRn glutamate synapses were to be mediated by an enhanced tonic NO signaling and persistent activation of soluble guanylate cyclase and protein kinase G (sGC/PKG) signaling cascade, then due to an occlusion effect, direct activation of this signaling pathway should not potentiate eEPSC amplitude in PE rats. We tested this possibility by conducting three independent sets of experiments. First, we examined the effect of NOS activation on the amplitude of eEPSCs in control and PE rats. As expected, we found in control rats, administration of NO donor SNAP (200 µM) potentiated the eEPSC amplitude (143.96 ± 10.90 % of baseline; Fig. 5A₁), and increased glutamate release, as indicated by the decrease in PPR (Baseline: 1.21 ± 0.08; SNAP: 1.11 ± 0.08; Fig. 5A₂) and CV (Baseline: 0.45 ± 0.04; SNAP: 0.30 ± 0.02; Fig. 5A₃). In contrast, administration of SNAP (200 µM) did not alter the eEPSC amplitude in PE rats (100.64 ± 3.91 % of baseline; p < 0.05, ɳ²_P = 0.50, PE vs PC; Fig. 5A₁). Next, we examined the effect of NO precursor L-Arg on eEPSC amplitude in control and PE rats and found that L-Arg increased the eEPSC amplitude in control, but not in PE rats (Control: 129.06 ± 4.29 % of baseline; PE: 108.57 ± 7.77 % of baseline; p < 0.05, ɳ²_P = 0.78, PE vs PC; Fig. 5B₁). The L-Arg induced potentiation of eEPSC amplitude was mediated by an increase in glutamate release as indicated by the decrease in PPR (Baseline: 1.24 ± 0.09; L-Arg: 1.11 ± 0.06; Fig. 5B₂) and CV of eEPSCs (Baseline: 0.36 ± 0.07; L-Arg: 0.28 ± 0.06; Fig. 5B₃). Finally, because activation of sGC/PKG pathway mediate NO-induced potentiation of DRn glutamate synapses, we assessed the impact of PE on the sGC/PKG pathway-mediated potentiation of eEPSC amplitude. We found that only in control rats, activation of sGC with A350619 (100 µM) enhanced the eEPSC amplitude (Control: 162.68 ± 11.86 % of baseline; PE: 90.40 ± 5.41 % of baseline; p < 0.05, ɳ²_P = 0.68, PE vs PC; Fig. 5C). Similarly, activation of PKG with 8pCPT-cGMP (100 µM), potentiated the amplitude of eEPSCs only in control rats (Control: 156.23 ± 2.87 % of baseline; PE: 98.13 ± 5.86 % of baseline; p < 0.05, ɳ²_P = 0.88, PE vs PC; Fig. 5D). Together these results support the hypothesis that a persistent activation of NO/sGC/PKG signaling pathway contribute to the PE-induced potentiation of DRn glutamate synapses.

Glutamate synapses in the DRn are also controlled by tonic eCB signaling [38]. Manipulations that increase and decrease tonic eCB signaling depress and potentiate the strength of these synapses, respectively [38]. Thus, it is possible that an impaired eCB signaling in PE rats could also contribute to the potentiation of DRn glutamate synapses. We examined this possibility by assessing the effects of PE on tonic eCB signaling measured by the blockade of CB1R-induced potentiation of EPSC amplitude [38]. Consistent with a previous report [38], in control rats, blockade of CB1Rs with AM 251 (3 µM) potentiated the eEPSC amplitude and increased glutamate release (163.81 ± 9. 2 % of baseline; Fig. 6A), indicating the presence of a tonic eCB signaling in the DRn. In contrast, in PE rats, the AM 251-induced potentiation of eEPSC amplitude was blunted (103. 9 ± 7.25 % of baseline; p < 0.05, ɳ²_P = 0.54, PE vs PC; Fig. 6A), indicating that PE impaired tonic eCB signaling. Mechanistically, PE could impair tonic eCB signaling by inhibiting eCB synthesis and release, reducing CB1R function, or a combination of both. To test these possibilities, we first measure the levels of anandamide (AEA) and 2-arachydonyl glycerol (2-AG) two major eCBs (i.e. AEA and 2-AG) in the DRn and found that PE did not significantly alter AEA (Control: 24.03 ± 5. 73 pmol/g; PE: 19.34 ± 5.25 pmol/g; Fig. 6B) and 2-AG levels (Control: 33.85 ± 3.93 nmol/g; PE: 23.699 ± 3.17 nmol/g; Fig. 6B), indicating that the impaired tonic eCB signaling in PE rats was not mediated by a deficit in AEA and 2-AG synthesis.

Next, we tested the effect of PE on the function and expression of CB1R. To that end, we examined the impact of PE on the depression of EPSC amplitude induced by the CB1R agonist WIN 55, 212-2 (WIN, 10 µM). We found that WIN55, 212-2 depressed the eEPSC amplitude in control (76.26 ± 7.5 % of baseline; Fig. 6C), but had no effect in PE rats (106.05 ± 10.12 % of baseline; p < 0.05, ɳ²_P = 0.53, PE vs PC; Fig. 6C), indicating an impaired CB1R function. We also assessed the expression of CB1Rs in the DRn using qPCR and found that compared to control, PE significantly reduced the levels of CB1R mRNA (53.65 ± 5.68 % of control; Fig. 6D). Collectively, these results indicate that PE impairs tonic eCB signaling by reducing CB1R function.

The findings that PE persistently enhances NO function while impairing tonic eCB signaling raised the possibility of a functional cross talk between these two systems. Specifically, because nNOS activity is regulated by baseline network activity, we hypothesized that the PE-induced blockade of tonic eCB function, which increases neuronal activity, may contribute to enhanced NO signaling in the DRn of PE rats. We tested this possibility by examining whether in control rats, blocking tonic eCB with the CB1R antagonist/reverse agonist AM 251, could mimic the increased NO signaling observed in PE rats. Tonic NO signaling was assessed by the depression of eEPSC amplitude induced by NO scavenger. We found that incubating DRn slices from control rats with AM 251 (3 µM) significantly increased the depression of eEPSCs induced by the NO scavenger PTIO (100 µM) (In control: 103.25 ± 1.46; In AM 251: 72.22 ± 2.07 % of baseline; p < 0.05, ɳ²_P = 0.95, Control vs Control+AM251; Fig. 6E), thereby suggesting that blocking eCB function enhances tonic NO signaling. In contrast, in PE rats, which already exhibit a strong PTIO-induced depression of eEPSCs, blockade of CB1R with AM 251 did not further enhanced the PTIO- induced depression of eEPSC amplitude (PE control: 62.52 ± 1.4 % of baseline; PE AM 251: 58.38 ± 1.43; p > 0.05, ɳ²_P = 0.19, PE vs PE + AM251; Fig. 6F). Taken together, these results strongly support the hypothesis that an impaired CB1R function contributes, at least in part, to the increased tonic NO signaling in the DRn of PE rats.