Projections from infralimbic medial prefrontal cortex glutamatergic outputs to amygdala mediates opioid induced hyperalgesia in male rats

Abstract

Repeated use of opioid analgesics may cause a paradoxically exacerbated pain known as opioid-induced hyperalgesia (OIH), which hinders effective clinical intervention for severe pain. Currently, little is known about the neural circuits underlying OIH modulation. Previous studies suggest that laterocapsular division of the central nucleus of amygdala (CeLC) is critically involved in the regulation of OIH. Our purpose is to clarify the role of the projections from infralimbic medial prefrontal cortex (IL) to CeLC in OIH. We first produced an OIH model by repeated fentanyl subcutaneous injection in male rats. Immunofluorescence staining revealed that c-Fos-positive neurons were significantly increased in the right CeLC in OIH rats than the saline controls. Then, we used calcium/calmodulin-dependent protein kinase IIα (CaMKIIα) labeling and the patch-clamp recordings with ex vivo optogenetics to detect the functional projections from glutamate pyramidal neurons in IL to the CeLC. The synaptic transmission from IL to CeLC, shown in the excitatory postsynaptic currents (eEPSCs), inhibitory postsynaptic currents (eIPSCs) and paired-pulse ratio (PPR), was observably enhanced after fentanyl administration. Moreover, optogenetic activation of this IL-CeLC pathway decreased c-Fos expression in CeLC and ameliorated mechanical and thermal pain in OIH. On the contrary, silencing this pathway by chemogenetics exacerbated OIH by activating the CeLC. Combined with the electrophysiology results, the enhanced synaptic transmission from IL to CeLC might be a cortical gain of IL to relieve OIH rather than a reason for OIH generation. Scaling up IL outputs to CeLC may be an effective neuromodulation strategy to treat OIH.

Introduction

Opioids are still milestone analgesics for modern anesthesia and pain management.^1,2 Despite its well-accepted positive effects, a growing number of studies from animals to humans have shown that high-dose or long-term exposure to opioid analgesics might induce a paradoxical increase in pain called opioid-induced hyperalgesia (OIH).^3–6 Unlike opioid tolerance, further uptake of opioid is almost futile for pain control in OIH patients. Indeed, the mechanisms of OIH are complicated and inexplicit. There are famous hypotheses including neuroinflammation in the spinal cord, sensitization of the amygdala, as well as other cellular and molecular mechanisms in the pain matrix,^7–10 only few from which directly identified probable neural circuits modulating OIH. Interestingly, compared to females and healthy males, it’s more likely for male patients to experience hyperalgesia after a bolus of fentanyl, which is a synthetic opioid widely used in clinical practice nowadays. ¹¹ Therefore, it’s worth exploring which neural circuits may be responsible for OIH in patients of the specific gender.

The medial prefrontal cortex (mPFC) continuously undergoes morphological and functional changes that may account for behavioral adaptations in response to pain and opioid exposure.^12–14 As a crucial part of the mPFC, the rat infralimbic cortex (IL, the homolog of human ventromedial prefrontal cortex)^13,15 regulates multiple behaviors, such as drug seeking, pain associated emotion and fear extinction.^16–18 Importantly, the IL has direct connections with the amygdala and activation of the IL- Amygdala circuit could alleviate anxiety and fear. ¹⁹ In this regard, the IL sends a dense projection to the amygdala, especially the central part of the amygdala (CeA).^20,21 As the main output nucleus of the amygdala, the CeA integrates nociceptive information with multimodal information about the internal and external environment of the body. ²⁰ However, it remains ambiguous whether and how the connections between the IL and the amygdala may be involved in the development and progression of OIH.

Traditionally, the CeA is divided into three areas including the central medial amygdala (CeM), the lateral central amygdala (CeL) and the capsular central amygdala (CeC, also referred to as the CeLC) ²² . Our previous studies have suggested that the high expression of ERK (Extracellular Signal-Regulated Kinase), CaMKIIα and mGluR1 (Metabotropic Glutamate Receptor 1) in the CeLC induced fentanyl-induced hyperalgesia in male rats.^7,10,23 Additionally, there is evidence showing that the IL regulates CeLC neurons via feedforward inhibition indirectly, which is impaired in the arthritis pain model.^17,24,25 In this study, we verified the activation of CeLC in fentanyl-induced hyperalgesia of male rats with c-Fos immunostaining. We hypothesized that the IL modulates OIH through direct monosynaptic connections with the CeLC. To test this, we used in vitro patch-clamp electrophysiology with optogenetics to investigate the synaptic transmission from IL to CeLC in OIH male rats. Then, we opto-stimulated or chemo-inhibited IL-CeLC pathway respectively to test its effects on sensory hypersensitivity after fentanyl injection. With this study we are the first to analyze the role of IL-CeLC circuit in OIH male rats, thus providing new references for exploring the mechanism and treatment of OIH in the future.

Materials and methods

Animals

Male Sprague-Dawley rats weighing 60–80 g were purchased from the Laboratory Animal and Biomedical of South-Central University for Nationalities. These rats were prepared for virus injection, marking the beginning of the experiment. All rats were housed in plastic cages (4 rats per cage) under a 12 h light/dark cycle in a temperature- and humidity-controlled animal care room with access to food and water ad libitum. All animal experiments were conducted following the National Institutes of Health guide for the care and use of laboratory animals and Ethical Issue of the International Association for the Study of Pain.

Fentanyl-induced hyperalgesia model

The well-established Fentanyl-Induced Hyperalgesia model that mimics the high-dose opioid treatment used in human surgeries was induced as described in detail previously. ²⁶ Fentanyl was purchased from Yichang Humanwell Pharmaceutical Co. Ltd (China). To establish this OIH model, fentanyl was injected subcutaneously at a dose of 60 μg/kg for each rat, who required 4 times repeated injections in total, with an interval of 15 min, resulting in a total dose of 240 μg/kg. The control animals were injected with the same amount of normal saline. Additionally, all rats for OIH model were raised to 350–400 g before fentanyl administration.

Anterograde tracking

To verify if CeLC neurons receive monosynaptic glutamate input from IL neurons, a cre-dependent anterograde transsynaptic AAV (packaged and purchased from Brain VTA, Wuhan, China) was injected into the IL. Rats were anesthetized with ketamine (50 mg/kg) and xylazine (7.5 mg/kg) and then positioned in a stereotaxic frame (RWD Life Science, Shenzhen, China). After drilling a hole in the skull, 0.2 μL AAV1/2-CaMKIIα-Cre-EGFP(PT0198, 1.05*10¹³ genome copies/ml) was injected into the right IL (AP: +3.0 mm, ML: −0.5 mm, DV: −4.9 mm)^17,27 at a speed of 0.1 μL/min by using the 10-μL Hamilton syringe. The needles were kept in the position for another 10 min to allow the virus to diffuse. Four weeks after the virus injection, these rats were sacrificed and the brain slices were then obtained as described below in “Histology and imaging”.

In vivo optogenetics

After rats were deeply anesthetized, 0.2 μL AAV2/9-CaMKIIα-hChR2-EYFP (PT0296, 5.41*10¹² genome copies/ml) or AAV2/9-CaMKIIα-EYFP (PT0107, 5.33*10¹² genome copies/ml) was injected into the right IL (AP: +3.0 mm, ML: −0.5 mm, DV: −4.9 mm) at a speed of 0.1 μL/min. ¹⁷ The rats were raised for at least four more weeks to recover for the virus infection and then anesthetized again for optic fibers (0.37 NA, Ø2.5 mm, Inper Ltd, China) implantation. Optic fibers were implanted into the right CeLC (AP: −2.3 mm, ML: −4.2 mm, DV: −7.5 mm) and behavior tests were performed at least 1 week later. To activate (IL-CeLC pathway) glutamate terminals in the CeLC from IL, a 470 nm blue light stimulator (Inper B1-470, Inper Ltd, China) were connected to the optic fiber by an optic cable. For mechanical sensitivity assessment, the light was delivered at 20 Hz, 15 ms pulse, 15 mW/mm² mode for 15 s until the reflex happened. While for thermal sensitivity assessment, the light lasted for 30 s at most for each test to prevent neuronal damage. The interval between two stimulations was at least 5 min.^27,28

Chemogenetics

To specifically inhibit IL- CeLC pathway, the chemogenetic technology DREADD (designer receptors exclusively activated by designer drugs) was applied. Totally 0.2 μL AAV2/9-CaMKIIα-hM4D (Gi)-EGFP (PT0296, 5.41*10¹² genome copies/ml) was injected into the IL of each rat. The mutated human muscarinic receptors hM4Di can be exclusively activated by clozapine-n-oxide (CNO), which can also inhibit the glutamate terminals from IL in the CeLC (IL-CeLC pathway). ²⁹ After 4 weeks recovery, a 33-gauge stainless steel cannula (RWD Life Science, Shenzhen, China) was implanted into the right CeLC (AP: −2.3 mm, ML: - 4.2 mm, DV: −7.5 mm) and behavior tests were performed at least 1 week later. The CNO (0.5 mM, 0.5 μL), dissolved in sterile saline was microinjected into the CeLC at the speed of 0.1 μL/min 30 min before behavior tests. After infusion, the injector was kept in the same place for another 5 minutes for drug diffusion. For the control group, the same amount of normal saline was injected into the CeLC.

Behavioral assessment

Prior to behavioral testing, all rats (350–400 g) were placed in the special chamber to acclimate to the experimental environment for 1 h each day at the same time as the experiment happened for 3 days prior to the day of the experiment. At the experiment day the nociceptive thresholds of the left hind paw of the rats were evaluated by mechanical and thermal hypersensitivity 6.5 h after the last fentanyl injection according to previous reports.^23,30,31 The rats were then randomly divided into different experimental groups. Some rats were tested before, during, and after light stimulation, while some rats were tested before and after drug administration.

Mechanical nociceptive threshold assessment

Paw-withdrawal mechanical thresholds (PWMT) were measured with von Frey filaments (North Coast, San Jose, CA, USA). The rats were placed in a transparent plastic partition with wire mesh floors to habituate until immobility before testing. A series of calibrated filaments from 0.16 g to 26 g was applied perpendicularly to the plantar surface of the left hind paw in an appropriate force for about 13 s, and withdrawal or licking of the paw was defined as the positive response. The interval between two filament stimulation was at least 5 min. PWMT was determined by the “up and down” method using a Dixon nonparametric test as described previously. ³²

Thermal nociceptive latency assessment

Thermal withdrawal latency (PWTL) was determined by Hargreaves test previously.^7,8 On the day of the measurement, the rats were lying down in a clear Plexiglas chamber on the top of a glass pane quietly. The radiant thermal stimulator (BME-410C, Biomedical Engineering, Boerni Science and Technology Co., Ltd, Guangzhou, China) was adjusted to the appropriate intensity to stimulate the left hind paw plantar until the rats exhibited withdrawing or licking the paw and the time was recorded. The cutoff time was set at 30 s to prevent tissue damage. All rats were measured at least 3 times with an interval of 5 min and the average of these latencies was the PWTL.

Histology and imaging

Rats were deeply anesthetized, and then transcardially perfused with PBS (37°C, pH 7.4), followed by ice-cold paraformaldehyde (PFA). Brains were dissected and postfixed in 4% PFA overnight at 4°C. After successive equilibration in 20% and 30% sucrose in PBS, coronal slices (30 μm) were cut with a cryostat (Leica CM1950) and stored in PBS at 4°C until immunostaining. Brain slices with IL or CeLC regions were selected according to the anatomical landmarks. For c-Fos immunostaining, brain sections were washed three times with PBS for 5 minutes every time. Then, the brain sections were blocked in 5% normal donkey serum solution containing 0.3% Triton X-100 for 1.5 h. After washing with PBS, the slices were incubated in the rabbit anti-c-Fos(1:500, #2250, Cell Signaling Technology) overnight at 4°C. Sections were then washed in PBS and incubated in goat anti-rabbit Cy3 (1:200, Abbkine, A22220) or Dylight 488 goat anti-rabbit (1:200, Abbkine, A23220). Sections were rinsed in PBS and mounted on microscope slides with Antifade Mounting Medium with DAPI (P0131, Beyotime). To verify the location of AAV transfection, the image of brain slices with EYFP and EGFP were obtained under 488 nm (light) excitation and the image of Cy3 was obtained under 543 nm (light) excitation. All the images were captured with a Nikon N2 confocal microscope.

A series of slices containing the CeLC (bregma −2.4 to −2.64) were imaged by confocal microscopy with a 20 × immersion lens and collected at a resolution of 2048 × 2048 pixels. The same laser and scanning settings were used for all confocal images within an experiment to allow comparison across groups. In general, coronal sections from three to five mice were used for quantitative analysis. Three to five images were taken for each mouse, and the analyzed values of these images were averaged to determine the value for each individual. Series of images were captured from the confocal microscope and converted to 8-bit gray scale images, and then the area and mean gray values of white color clusters were measured using the Image-J software. Quantification of c-Fos labeling neurons was estimated in the form of optical density with the same threshold. The positive neurons were defined with large nuclei stained diffusely and staining above basal background. ³³

Electrophysiological recordings

Slice preparation

Rats were anesthetized and decapitated. The brains were removed rapidly and put in the ice-cold cutting solution containing (in mM) 213 sucrose, 3 KCl, one NaH₂PO₄, 0.5 CaCl₂, five MgCl₂, 26 NaHCO₃, and 10 glucose. Coronal brain slices (300 μm) containing CeLC were obtained from the right hemisphere as described previously ²³ with oxygenated (95% O₂ and 5% CO₂) cutting solution at 4°C by a Vibratome (Leica VT 1000S). The slices were allowed to recover in artificial cerebrospinal fluid (ACSF) at 37°C for at least 1 h before recording. The ACSF contained (in mM) 125 NaCl, 5 KCl, 1.2 NaH₂PO₄, 2.6 CaCl₂, 1.3 MgCl₂, 26 NaHCO₃, and 10 glucose. All chemicals were obtained from Sigma-Aldrich.

Whole cell patch clamp recording

To verify the AAV transfection, only slices from rats that ChR2 were strictly expressed in IL are allowed to be used. Slices containing CeLC was transferred to a recording chamber and continuously perfused (2 mL/min) with oxygenated ACSF. Recording pipettes (WPI, USA, 3–5 MΩ resistance) were filled with the following internal solution (in mM):145 KCl, five NaCl, 10 HEPES, five EGTA, 4 Mg-ATP, and 0.3 Na₃- GTP, pH adjusted to 7.3 with KOH and osmolarity to 280 mOsm/kg with sucrose. All chemicals were obtained from Sigma-Aldrich. Data were obtained using EPC10 amplifier and Patchmaster software (HEKA, Germany), filtered and sampled at 10 kHz with a dual 4-pole Bessel filter (Warner Instruments, Hamden, CT). All electrophysiology data analysis was done with Clampfit software (Molecular Devices, USA).

For light activation, ChR2 terminals, 2 ms blue light (pE-300white; CoolLED Ltd) was delivered through a ×40 water-immersion objective lens (Eclipse FN1; Nikon). To examine the monosynaptic function between IL and CeLC, eEPSCs were recorded at −70 mV in voltage clamp mode. Slices were perfused with sodium channel blocker TTX (1 µM) and followed by the addition of 4-aminopyridine (4-AP, 100 µM), a potassium channel blocker, to facilitate glutamate release from synaptic terminals and then CNQX (20 µM) were applied to block glutamate receptors. ¹⁹ Light evoked postsynaptic currents were recorded at −70 mV (eEPSCs) and 0 mV (eIPSCs) in voltage clamp mode respectively. The non-NMDA receptor antagonist CNQX (20 µM) was added to ACSF to block glutamate receptors and GABA_AR antagonist picrotoxin (PTX) (100 μM) was applied to block GABA receptors. Synaptic latencies of eEPSCs and eIPSCs were determined as the time interval between the onset of light stimulation and the onset of current at holding potentials of −70 and 0 mV, respectively. To record the PPR, two light pulses (2 ms duration) with an interval of 100 ms were delivered to the CeLC. The PPR was calculated as the ratio of the amplitude of the second EPSC to that of the first.

Statistical analysis

Most data were expressed as mean ± SEM., and statistical analysis was performed by GraphPad Prism 7.0 software (GraphPad Software, San Diego, CA). Two-way RM ANOVA with Sidak’s multiple comparisons test was used for behavior assessment. We use One-way ANOVA followed by the Tukey’s multiple-comparison tests to analyze differences between three or more groups. The student’s t-test was used between two groups for electrophysiology and immunofluorescence experiments. Statistical significance was set at p < .05.

Results

The activity of the right central nucleus of amygdala is increased in an opioid-induced hyperalgesia model of male rats

Fentanyl-Induced Hyperalgesia model was induced in male SD rats to mimic OIH. The timeline was shown in Figure 1(a), the pain thresholds were tested before and after fentanyl or saline injection. There was no statistical difference in the basal mechanical and thermal pain threshold between Saline and Fentanyl group. As shown in Figures 1(b) and (c), the detectable behavioral hypersensitivity peaked at 6.5 h after fentanyl administration and lasted for 3 days. These results were consistent with previous reports that rats exposed to opioids experience hyperreflexia symptoms in a time-dependent manner.^7,23,30

Figure 1.

Repeated fentanyl injection induced mechanical and thermal hyperalgesia in rats. (a) Rats were randomly divided into fentanyl and saline groups and behavior testing was time dependent: mechanical and thermal pain thresholds were tested before (0 days) and after (3 h, 6.5 h, 12 h, 1 day, 2 days, 3 days, 4 days, 5 days, 6 days) fentanyl or saline administration (60 µg/kg, repeated 4 times with 15-min intervals subcutaneously, n = 5 in each group). (b), (c) Time course of mechanical pain thresholds (b) and thermal pain thresholds (c) in the fentanyl and saline groups. [(b), F _interaction (9, 80) = 3.908, p = .0004; F _time (1, 80) = 39.69, p < .0001, n = 5; (c), F _interaction (9, 80) = 6.967, p < .0001; F _time (1, 80) = 61.04, p < .0001, n = 5, two-way RM ANOVA with Sidak’s multiple comparisons test] *p < .05, **p < .01***p < .001, ****p < .0001, Data are presented as the mean ± SEM.

As one of the immediate early genes (IEG), c-Fos would respond to extrinsic cellular stimuli and be induced by action potential firing, which was therefore chosen as an excitatory marker.^34,35 Given that the peaked behavioral hypersensitivity was 6.5 h after fentanyl injection in our OIH model, we assessed the neuronal activity of both the right and left CeLC by c-Fos immunofluorescence staining in the fentanyl and saline groups. The timeline was shown in Figure 2(a). We found that the level of c-Fos was significantly higher in the right CeLC region after fentanyl administration compared with the saline group (Figure 2(b) and (c), p = .0004, n = 4), while on the left side, there was no significant difference between the two groups (Figure 2(d) and (e), p = .979, n = 4).

Figure 2.

Fentanyl administration increases the c-Fos expression in the right CeLC. (a) Illustration of the experimental timeline. (b) Representative c-Fos immunofluorescence images of the right CeLC in fentanyl and saline groups. (c) The number of c-Fos positive neurons in the right CeLC (n = 4, p = .0004, unpaired t-test). (d) Representative c-Fos immunofluorescence images of the left CeLC in fentanyl and saline groups. (e) Number of c-Fos positive neurons in the Left CeLC (n = 4, p = .979, unpaired t-test) compared with saline-injected rats. ***p < .001, data are presented as the mean ± SEM.

Monosynaptic and functional glutamate projection from infralimbic medial prefrontal cortex-excitatory neurons to central nucleus of amygdala

Among the five cellular layers in IL, the primary output neurons of the IL are excitatory pyramidal neurons in the deep layer, which is a major source of excitatory input to the amygdala.^36,37 It has been shown that calcium/calmodulin-dependent protein kinase IIα (CaMKIIα) is expressed selectively in glutamate pyramidal neurons in the cortex, so we use CaMKIIα to label the output-projecting pyramidal cells in the IL. ³⁸ Previous studies have suggested projections from IL to CeLC ¹⁷ but there is no direct evidence to show that IL glutamate neurons could project to CeLC directly. To verify if CeLC neurons receive inputs from the IL, we applied two strategies. First, we stereotaxically injected anterograde trans-synaptic AAV1/2 tagged with an enhanced green fluorescent protein (EGFP) under the control of CaMKIIα promoter (AAV1/2-CaMKIIα-Cre-EGFP) into the IL, which can transduce axon terminals to its projecting targets and thus express there (Figure 3(a)).^39,40 4 weeks later, we found that EGFP were expressed in all layers of the IL and certain neurons in CeLC, which were CeLC neurons that receive monosynaptic IL input (Figure 3(b)). Next, we used electrophysiology with ex vivo optogenetics to determine whether there were functional connections between IL and CeLC. We injected AAV2/9-CaMKIIα-hChR2-EYFP into IL and recorded the evoked excitatory postsynaptic currents (eEPSCs) stimulated by blue light (470 nm; 2 ms) in CeLC slices (Figure 3(c)). As revealed in Figures 3(d) and (e), the eEPSCs were abolished by the application of tetrodotoxin (TTX, 1 μM) and recovered by 4-AP (100 μM), and then completely blocked by adding non-NMDA receptor antagonist CNQX (20 μM). This further confirmed the monosynaptic connection between IL glutamate neurons and the CeLC.

Figure 3.

IL Glutamate projection to CeLC neurons. (a) Experimental scheme: injection of AAV- Cre-EGFP into IL and the projection to CeLC. (b) Fluorescence images representing the location of EGFP expression in the IL and CeLC neurons. (c) Scheme for recording light-evoked postsynaptic currents in the CeLC by optogenetic activation of the IL projections. (d) Light-evoked excitatory postsynaptic currents were completely blocked by TTX (1 μM) and recovered by TTX plus 4-AP (100 μM), and then were blocked by CNQX (20 μM). (e) Summarized plots of the eEPSCs amplitudes in (d). [(e), n = 7 neurons from three mice, F (3, 20) = 70.31, p < .0001; one-way ANOVA with Tukey’s multiple-comparison test]. ****p < .0001, Data are presented as the mean ± SEM.

The synaptic transmission from infralimbic medial prefrontal cortex to laterocapsular division of the central nucleus of amygdala is enhanced in opioid-induced hyperalgesia rats

To further investigate the synaptic mechanism between IL and CeLC in OIH and control rats, we used optogenetics with patch clamp methods as mentioned before (Figure 3(c)). We noticed that IL fibers may project to the centrolateral (CeL), centromedial (CeM) and basolateral amygdala (BLA). It’s possible that the activation/inhibition of IL- CeLC pathway by the light (or CNO) could affect these neighboring areas. However, the light (or CNO) had little effect on these neighboring areas due to the specific location of the optic fiber and cannula implantation (Figures 5(b) and 6(b), Sup. Figure 1). To minimize the tissue damage as suggested in many studies,^27,33,41 we held the optic fiber/cannula above the objective brain region 0.2–0.3 mm (Sup. Figure 1(c), (d), (h)). Although the dual viral strategy is more precise, the projection from IL to CeLC was not very rich (Figure 3(b)). Considering the relatively lower efficiency of anterograde transsynaptic AAV compared to the virus designed to express in specific area,^40,42 we selected the strategy descriped in the Materials and methods for further exploration. Moreover, since we have found dense projections between IL and BLA, we also postulated that OIH may affect the inhibitory/excitatory (I/E) balance from IL to BLA. However, OIH had little influence on the amplitudes of IL-evoked EPSCs and IPSCs as well as their ratios, suggesting that output to BLA are more resilient against OIH than those to CeLC (Sup. Figure 2). That’s why we didn’t continue to study the influence of IL-BLA pathway on pain sensitivity in our OIH rats. These findings were different from that of the arthritis model, ¹⁷ indicating that the mechanism of OIH was different from the inflammatory pain.

The IL was transfected with the light-sensitive channelrhodopsin-2, which allowed us to observe the fibers transduced to the CeLC (Figure 4(a)). As revealed in Figure 4(b), opto-stimulation of IL projecting fibers in the CeLC induced both excitatory postsynaptic currents (eEPSCs) and inhibitory postsynaptic currents (eIPSCs). The onset latency of the eIPSCs was significantly larger than that of eEPSCs (Figure 4(c)), suggesting that the inhibitory transmission between IL and CeLC were disynaptic. Light activation of IL terminals in the amygdala slices evoked glutamate receptor-driven IPSCs that were abolished by the application of CNQX (20 μM) or GABA_AR antagonist picrotoxin (PTX) (100 μM) in the ACSF (Figure 4(d)).

Figure 4.

The potentiated synaptic transmission from IL to CeLC in OIH rats. (a) Representative images showing the ChR2 expressed in IL and IL terminals in CeLC. (b) Representative traces of IL-evoked EPSCs and IPSCs after opto-stimulation (blue bars, 470 nm, 2 ms) in CeLC at −70 and 0 mV, respectively. (c)The latency of eIPSCs compared with that of eEPSCs. (eEPSCs: n = 12 neurons/5 rats; eIPSCs: n = 9 neurons/4 rats. p < .0001) (d) Representative traces showing effects of 6-cyano-7-nitroquinoxaline-2,3-dioneis (CNQX, 20 μM) or picrotoxin (PTX, 100 μM) on IL-evoked EPSCs/IPSCs. Cells were held at −70 and 0 mV, respectively. (e) Representative traces of the evoked EPSCs/IPSCs in CeLC neurons of control and OIH rats following light activation of IL fibers. (f), (g) The amplitudes of evoked EPSCs (f) and IPSCs (g) in OIH rats and the saline controls (unpaired t test, (f): p = .06, n = 17 neurons/5 rats in OIH rats and control rats respectively; (g): n = 18 neurons/5 rats in OIH rats and control rats respectively). (h) The evoked IPSC/EPSC ratios in OIH rats and the saline controls. (unpaired t test, F: p = .024, n = 10 neurons/4 rats in OIH rats and n = 11 neurons/4 rats in control rats respectively). (i) Representative traces of the paired-pulse ratio (PPR) of eEPSCs in the CeLC neurons from control and OIH rats. (j) The PPR in the CeLC neurons in OIH rats and the saline controls (unpaired t test, p = .0098, n = 10 neurons/5 rats in OIH rats and n = 9 neurons/5 rats in control rats respectively) *p < .05; **p < .01, ***p < .001, data are presented as the mean ± SEM.

Our previous study showed that enhanced synaptic plasticity of CeLC plays a crucial role in the pathophysiology of OIH.^7,23Our group also found that IL layer V pyramidal neurons showed hyperexcitability in OIH rats. ³¹ Since we have demonstrated that there was monosynaptic functional transmission between IL and CeLC, we postulated that OIH may affect the inhibitory/excitatory (I/E) balance from IL to CeLC. By recording eEPSCs and eIPSCs between OIH and control rats (Figure 4(e)), we found that both excitatory and inhibitory transmission of CeLC were enhanced in OIH rats (Figure 4(f) and (g)), with the inhibitory-excitatory (I/E) balance in CeLC neurons shifted to inhibition (Figure 4(h)). We further postulated that the glutamatergic transmission between IL and CeLC may also be potentiated under OIH. To testify this, we evaluated the probability of presynaptic glutamate release in IL-to- CeLC connections via paired-pulse ratio (PPR) of eEPSCs, which is known to be inversely correlated with the presynaptic glutamate release. ⁴³ We delivered two consecutive light pulses with 100 ms intervals to excite the IL afferents to get PPR (Figure 4(i)). The PPR was lower in OIH rats than the saline controls, reflecting the enhanced probability of glutamate release at presynaptic terminal from IL inputs to CeLC after fentanyl administration (Figure 4(j)). Taken together, the above results suggest that in OIH male rats, the synaptic transmission from IL to CeLC is potentiated, probably through presynaptic mechanisms that lead to the disruption of I/E balance.

Activation of the infralimbic medial prefrontal cortex-central nucleus of amygdala pathway ameliorates opioid-induced hyperalgesia by inhibiting central nucleus of amygdala

To detect whether the potentiated synaptic transmission from IL to CeLC was the cause of OIH, we activate the IL-CeLC pathway in normal rats by in vivo optogenetics for behavioral assessment. The timeline was shown in Figure 5(a). AAV under the control of CaMKIIα promoter, carrying ChR2 or not (AAV2/9-CaMKIIα-ChR2-EYFP or AAV2/9- CaMKIIα-EYFP), was injected into the IL to label the output-projecting pyramidal cells and the optic fiber was implanted in the CeLC (Figure 5(b)). As shown in Figure 5(c), for naive rats, activation of the IL- CeLC pathway (20 Hz, 15 ms pulse, 15 mW/mm²) by blue light did not affect the mechanical pain threshold (n = 7, p = .9417). After fentanyl injection, both groups developed hyperalgesia (Figure 5(d) and (e)). The activation of IL inputs in rats expressing AAV-ChR2-EYFP terminals in the CeLC notably increased both mechanical and thermal pain thresholds but did not alter the response of rats expressing only AAV-EYFP terminals. (Figure 5(d), p < .0001, Figure 5(e) p < .0001, n = 7). As shown in Figures 5(f) and (g), compared with rats merely expressing EYFP, ChR2-mediated activation of IL-CeLC inputs decreased c-Fos expression in CeLC after fentanyl administration in ChR2-EYFP group (p < .0001, n = 4). Taken together, optogenetically activated IL terminals in CeLC drastically ameliorated OIH and downregulated the activity of the CeLC.

Figure 5.

Optical stimulation of IL terminals in CeLC is sufficient to ameliorate OIH and downregulate the activity of CeLC. (a) Experimental paradigm for optogenetic activation: Neurons in the IL were transduced with ChR2 -EYFP or EYFP. The blue light was delivered via optical fibers implanted into the CeLC after 3 weeks of viral incubation. (b) Left: Schematic diagram of the location of virus injection and optical fiber implantation. Right: Fluorescence images showing AAV-EYFP expression in IL (scale bars 1000 μm) and CeLC optical fiber implantation location (scale bars 100 μm). (c) Mechanical paw withdrawal threshold when optogenetic activation of IL-CeLC circuit in naïve rats. (n = 7, F _interaction (1, 24) = 0.0054; p = .9417;) (d and e) Mechanical paw withdrawal threshold and thermal paw withdrawal latency of the ChR2 group and EYFP group. [(d), F _interaction (3, 36) = 15.02, p < .0001; F _time (3, 36) = 76.11, p < .0001 n = 7; (e), F _interaction (3,36) = 8.455, p = .0002;F _time (3, 36) = 20.57, p < .0001, n = 7, two-way RM ANOVA with Sidak’s multiple comparisons test] (f) Representative c-Fos immunofluorescence images in the right CeLC after optogenetic activation of IL- CeLC pathway (scale bars: left and middle:100 μm; right: 50 μm). (g) Number of c-Fos positive neurons in the right CeLC after optogenetic activation of IL- CeLC pathway (n = 4, unpaired t-test, p < .0001) ****p < .0001, data are presented as the mean ± SEM.

Inhibition of the infralimbic medial prefrontal cortex-central nucleus of amygdala pathway aggravates opioid-induced hyperalgesia by activating central nucleus of amygdala

To further explore the role of the IL-CeLC pathway in OIH, we decided to inhibit IL input to CeLC using Chemogenetics in vivo. We transfected IL with AAV encoding chemogenetic inhibitory DREADD (designer receptors exclusively activated by designer drugs) - hM4Di-EGFP and clozapine-n-oxide (CNO), which can inhibit the projecting area in the CeLC exclusively . ⁴⁴ As the timeline of the experiment shown in Figure 6(a), at 28 days after virus incubation, the rats were randomly divided into two groups: the saline group that would be microinjected with saline into the CeLC 30 min before the behavior tests after fentanyl administration, and the CNO group that would be injected with CNO at the same time with the saline group. Before fentanyl injection, the basal pain thresholds were measured to identify the hyperalgesia status. After behavior tests, all rats were sacrificed for immunostaining to confirm the location of AAV injection and cannula implantation as shown in Figure 6(b). First, we have identified that chemogenetic inhibition of IL-CeLC pathway had no effect on rats without fentanyl (Figure 6(c), n = 5, p = .8518). Then as shown in Figures 6(d) and (e), CNO could enhance both mechanical and thermal hyperalgesia after fentanyl injection (Figure 6(d), p = .0461, 6E, p = .0268, n = 7). After the behavior test, immunostaining showed that inhibition of IL-CeLC pathway increased the level of c-Fos in the CeLC (Figure 6(f) and (g), p = .028, n = 4). These results suggested that the inhibition of IL-CeLC pathway might aggravate OIH by increasing CeLC activity. Combined with the IL-CeLC projection activation effect on OIH, the increased glutamate release probability from IL to CeLC might be a protective response to relieve OIH, while inhibiting this probability appears to have the opposite effect.

Figure 6.

Chemogenetic inhibition of IL terminals in CeLC exacerbates OIH and increases the activity of CeLC (a) Experimental diagram for Chemogenetic inhibition: neurons in the IL were transduced with hM4Di-EGFP, CNO or saline was delivered via a cannula implanted in the CeLC after 4–6 weeks of virus incubation. (b) Left: Schematic diagram of the location of virus injection and cannula implantation. Right: Fluorescence images showing AAV-EGFP expression in IL and CeLC cannula location (scale bars 1000 μm). (c) Mechanical paw withdrawal threshold when chemogenetically inhibited the IL-CeLC pathway in naïve rats [n = 5, F _interaction (1, 16) = 0.03605, p = .8518; two-way RM ANOVA with Bonferroni’s multiple comparisons test]. (d), (e) Mechanical paw withdrawal threshold and thermal paw withdrawal latency of the CNO group and NS group. [(d): n = 7, F _interaction (2, 24) = 2.394, p = .1127; F _time (2, 24) = 63.83, p < .0001 (e): F _interaction (2, 24) = 2.721, p = .0861; F _time (2, 24) = 37.96, p < .0001; two-way RM ANOVA with Sidak’s multiple comparisons test] (f) Representative c-Fos immunofluorescence images of the right CeLC after Chemogenetic inhibition of IL- CeLC pathway (scale bars: left and middle:100 μm; right: 50 μm). (g) Number of c-Fos positive neurons in the right CeLC after Chemogenetic inhibition of IL- CeLC pathway (n = 4, unpaired t-test, p = .028). *p < .05, data are presented as the mean ± SEM.

Discussion

As far as we know, this is the first study to focus on OIH in terms of upregulated neural circuits and validate the existence of monosynaptic glutamatergic connections between IL and CeLC in morphology using anterograde transsynaptic virus and in function by electrophysiology with optogenetics. OIH was successfully induced by fentanyl injection and an obvious activation of the right CeLC was observed. Specifically, we found that the synaptic transmission between IL and CeLC was strengthened in OIH male rats. In addition, optogenetic activation of IL-CeLC pathway reversed OIH by inhibiting the CeLC, while chemogenetic inactivation of IL-CeLC pathway exacerbate OIH by activating the CeLC, indicating that the enhanced glutamate release from IL to CeLC could inhibit the activity of CeLC. Taken together, these results suggested a rather surprising effect in OIH rats that the enhanced synaptic transmission from IL to CeLC might be a protective response to relieve pain, which could be utilized as a neuromodulation strategy for the treatment of OIH.

Over the past 20 years, the area of pain-related hemispheric lateralization in the amygdala has raised a lot of concern. There is a general trend that when the left and right amygdala have been assessed separately, the right CeA is pronociceptive while the left has little or positive effect on pain. ²² Accordingly, ERK activation occurs in the right CeA plays a dominant role in inflammation-induced peripheral hypersensitivity. ⁴⁵ Engagement of opioid receptor signaling in the right CeA contributes to pain responses produced by morphine. ⁴⁶ Previous studies also identified the activation of ERK and mGluR1 in the right but not the left CeLC after fentanyl induced hyperalgesia.^10,23 This agrees with our finding that the modulation in OIH rats is functionally lateralized to the right hemisphere. Thus, the right IL-CeLC pathway was selected for behavioral tests. Also, eEPSCs and eIPSCs were recorded by whole-cell patch-clamp in the right CeLC. Clinically, most opioids are administered intravenously or orally, thus entering the central nervous system to produce analgesia. Our present discoveries in hemispheric lateralization may provide new ideas for a better way of opioid dosage to treat OIH.

An enduring issue in the management of pain is gender difference. It’s well known that chronic pain has a higher prevalence among female around the world. ⁴⁷ Studies also suggest that female tends to suffer from more severe postoperative and procedural pain than men. ⁴⁸ However, a greater number of male patients may experience hyperalgesia after fentanyl treatment. ¹¹ Gender differences in the effects of opioids on hyperalgesia have been described in humans and in animal models for reasons that remain unclear.^49,50 Our preliminary data have revealed that several male-specific genes were involved in the regulation of peripheral and spinal nociceptive processing during OIH. In the present study, we focused on the sensitivity to opioid antinociception solely in male. Not only for the purpose of our research continuity, but also in order to unveil the distinct mechanisms of hyperalgesia and therapeutic consequences in this specific gender.

As one of the primary targets for pain treatment, mPFC is widely investigated in multiple pain models, including neuropathic and inflammatory pain.^14,51,52 It has been demonstrated that projections from IL to CeA played a critical role in fear extinction, anxiety, and drug seeking.^19,53 Previous studies have identified both IL and CeLC as key centers for OIH modulation.^8,10,23,31 In our previous study, we found that IL exhibited hyperexcitability in OIH rat, ³¹ which is in agreement with our present finding that the glutamate release probability from IL to CeLC was potentiated in OIH. However, the activation of IL input to CeLC could inhibit the activity of CeLC and thus significantly mitigate mechanical and thermal hyperalgesia after fentanyl injection, indicating that the enhanced projection from IL to CeLC could surprisingly be a protective response to ameliorate pain sensitivity in OIH. Similarly, we previously identified that the CeLC-PAG-RVM-spinal cord descending facilitative pain pathway potentiated in OIH rats ⁸ (PAG, Periaqueductal gray; RVM, Rostral ventromedial medulla). As other study demonstrated that activating IL-CeA projection could exerted anxiolytic and fear-releasing effects. ¹⁹ We have very much confidence to believe that activation of the IL-CeLC pathway in pathological conditions may play a positive therapeutic role. Since reports demonstrated that the CeA received feedforward inhibition from mPFC in nociceptive information transmission, ²⁵ we speculated that this typical pain pathway was inhibited when IL-CeLC projections were stimulated. But as a matter of fact, we need more evidences to verify such proposal. On the one hand, it appears to be a different case in other forms of pain, such as arthritic pain. ¹⁷ On the other hand, the activity of CeLC neurons that receive monosynaptic IL input needs further exploration to grasp a more comprehensive understanding.

Prelimbic cortex (PL), another subregion of the medial prefrontal cortex, is next to IL anatomically, displays close functional connections with IL and has been identified to play a crucial role in chronic pain.^14,54 Recently, we found that the oligodendrocyte decreased in the PL in OIH rats, which might affect the synaptic transmission between the PL and its projections. ³⁰ A few projections from PL to CeLC also exist. ¹⁷ Unlike PL which is mainly responsible for pain related behaviors and emotions, reports have shown that IL was more associated with cue-associated behavior, such as drug seeking compared to the other parts of the mPFC.^13,55 Traditionally, pain is a salient sensory stimulus that elicits neural responses to protect us from injury. ⁵⁶ As a kind of paradoxical pain, OIH cannot be relieved by increasing opioids. We posited that the state of hypersensitivity could be a special cue for the brain to make some changes to alleviate hyperalgesia when opioids overdosed. Therefore, this hyperalgesia signal might activate those pain-responsive neurons to potentiate synaptic transmission in the IL as a cortical gain projecting to the CeLC to prevent OIH.^51,57 It probably acts as a compensatory mechanism to the attenuative effects on amygdala of the decreased PL excitability under similar circumstances. However, this cortical gain of IL may be too weak to compensate for the hyperactivity of the CeLC,^7,8,10 so that the hyperalgesic state persists.

The specific function of PFC at a given time may thus depend on the precise behavioral context and neuron function. ⁵¹ For example, inflammatory pain can be improved by the addition of BDNF (Brain-derived neurotrophic factor) into the IL, ⁵⁸ while mGluR5 in the IL might intensify nociception in arthritic rats. ⁵⁹ One study also reported that pharmacogenetic inhibition of IL pyramidal neurons had no significant impact on pain-related wiping ⁶⁰ which was consistent with our study. Activating or silencing IL-CeLC projections did not significantly affect mechanical pain thresholds in naive rats, indicating that such cortical gain might only be triggered in OIH rats. In addition, we also found that the state of hyperalgesia could last for about 3 days after fentanyl administration. However, we only select the time point for electrophysiology when our rats appeared to be the most sensitive to the stimuli. This cortical gain may have been disrupted when acute pain turned into chronic pain.^51,57 Whether this cortical gain still exists in other timepoints needs to be explored in the future study.

Neurons in CeLC expressing CaMKIIα (CeLC ^CAM) may play a critical role in OIH development. Although CeLC neurons are mostly GABAergic, CaMKIIα is widely expressed in the CeA. ²⁴ We speculate that some IL-CeLC neurons expressing CaMKIIα are under excitatory control. ⁶¹ Given the vital role of CaMKIIα in synaptic plasticity, ⁶² we can further infer that long-term potentiation of IL- CeLC ^CAM synapses could induce long-lasting analgesic effects through increased inhibition of the typical descending facilitative pain pathway. ⁸ Indeed, one study has shown that selectively inhibiting the activation of CaMKII neurons in the amygdala can alleviate paclitaxel-induced pain. ⁶³ Our previous study also revealed that inhibition of CaMKIIα in the CeA could attenuates OIH in rats. ⁸ Although in naive rats, one study demonstrated that optogenetic activation of CeA neurons expressing CaMKIIα-EYFP induced a transient mechanical hypersensitivity, ⁶⁴ which is contradictory to our present discovery that activation of the IL-CeLC pathway had no effect on their pain threshold. However, It has been reported that GABAergic CeA neurons expressing CaMKIIα (CeA ^CAM neurons) also project to the lateral parabrachial nucleus (LPBN) and optogenetic stimulation these cells attenuated nociceptive response. ⁶⁵ Therefore, it is also possible that scaling up IL input to CeLC may alleviate OIH by activating CeA ^CAM -LPBN pathway in some extent. These findings suggested that the same cells exert in a dynamical form according to different pathophysiological conditions. How IL modulate OIH through CeLC ^CAM neurons needs further exploration.

Much attention should also be paid to the heterogeneity of the CeLC. CeLC contains two distinct subpopulations of GABAergic neurons that express somatostatin (SOM+) or protein kinase Cδ (PKCδ+)^66–68 and they have distinct functions in pain modulation. Activating SOM + neurons would alleviate neuropathic and inflammatory pain, while activating PKCδ+ neurons might enhance pain sensitivity after formalin administration and nerve injury.^64,69,70 It is not accurate to identify the cell type only by morphology and thus the cell type projections from IL to CeLC remain unclear.^71,72 Based on our electrophysiological results, most of the responding cells in the CeLC had regular spiking. There are also reports showing that the distribution of these two cell types had the same electrophysiology properties,^69,70 which hinders us from identifying them through electrophysiological approaches as well. Nevertheless, in physiological conditions, the PKCδ+ and SOM + CeLC neurons are interconnected and repress each other to restrict the output of the amygdala.^73,74One study found that activation of ERK in CeA mostly occurred in PKCδ+ neurons after formalin injection. ⁶⁴ Our previous study also found that ERK in CeLC was upregulated in OIH rats. ²³ So, we speculate that the PKCδ+ neurons may induce hyperalgesia by activating ERK after fentanyl administration. Considering that optical activation of SOM + neurons can increase the inhibitory control of PKCδ+ neurons, ⁶⁴ the IL fibers may have more projections to SOM + neurons in the CeLC. When activating IL-CeLC pathway, the potentiated synaptic transmission from IL may activate SOM + cells and inhibit ERK upregulation by inhibiting the activity of PKCδ + neurons and thus recover the balance of excitation and inhibition in CeLC.

There are, however, several limitations in our study. First, although we have identified the potentiated synaptic transmission from IL to CeLC could be a cortical gain to alleviate OIH, the molecular mechanism of the synaptic change remains unclear. Then, the projections from IL to CeLC are not very abundant compared with the projections from IL to basolateral amygdala (BLA), even though researches have demonstrated that the BLA was associated with fear, anxiety and pain related emotions rather than pain sensation.^43,75 It is interesting to explore the activity of IL-BLA projections in OIH in the future. Moreover, earlier studies demonstrated that the CeA inhibitory microcircuits participate in fear extinction. ⁷⁶ Whether the IL projection would affect the microcircuits in the CeLC after fentanyl administration has not been examined.

Conclusion

Briefly, this study has demonstrated that the glutamate pyramidal neurons in IL send direct projections to CeLC and these projections are critical for modulating OIH. The enhanced synaptic transmission from IL to CeLC can be a cortical gain to relieve OIH. Scaling up this cortical gain may be an effective neuromodulation strategy to treat OIH, potentially through precise targeting of neurocircuit pathways, similar to precise targeting of receptor pathways in pharmacology.

Supplemental Material

Supplemental Material - Projections from infralimbic medial prefrontal cortex glutamatergic outputs to amygdala mediates opioid induced hyperalgesia in male rats

Supplemental Material for Projections from infralimbic medial prefrontal cortex glutamatergic outputs to amygdala mediates opioid induced hyperalgesia in male rats by Ling-Ling Cui, Xi-Xi Wang, Han Liu, Fang Luo and Chen-Hong Li in Molecular Pain.