Background
It has been gradually known that pain is a complex experience consisting of sensory-discriminative, affective-motivational, and cognitive-evaluative dimensions [
1,
2]. Furthermore, there is now a consensus of idea that noxious information is processed by a distributed and interconnected neural network, referred to as neuromatrix, in the brain [
3‐
6]. Unlike physiological state, pathological pain, when becomes persistent or chronic, can affect various higher brain functions (such as perception, emotion, cognition, and memory) through ascending pain pathways, leading to consequences of cognitive decline and mental disability. In the past three decades, the most advanced understanding about pain is that pathogenesis or chronicity of pain is attributable to sensitization of primary sensory neurons and synaptic plasticity in dorsal horn of the spinal cord [
7‐
9]. To date, the mechanisms by which inflammatory or neuropathic pain is processed at the lower level of the pain pathway have been well characterized [
7‐
11]. However, in clinic, chronic pain often results in not only sensory dysfunction (spontaneous pain, hyperalgesia and allodynia, etc.) but also emotional and cognitive disorders such as anxiety, amnesia and depression [
12‐
14]. Unfortunately, so far, the influences of pathological pain on the higher brain functions are not clear and this may hinder the advances in clinical pain therapy. Therefore, unraveling how pain affects the emotion- or cognition-controlling regions at a higher level of the "pain matrix" would definitely improve our understanding of the process of pain chronicity and provide novel strategies for treating negative emotional symptoms of chronic pain in the clinical setting [
4,
6].
There is substantial evidence indicating that the hippocampal formation (HF), an integral component of the limbic system [
15,
16], is involved in pain processing besides its well documented roles in learning and memory formation [
17‐
19]. Melzack and Casey (1968) proposed that the limbic forebrain structures, including the HF, play important roles in the 'aversive drive and affect that comprise the motivational dimension of pain' [
20]. Anatomically, the HF is positioned as a key interconnecting structure in Papez's circuit of the limbic system, mediating a variety of biological functions, including learning and memory, anxiety, emotion and sensorimotor integration [
15,
21,
22]. More recent evidence using the atlas registration-based event-related (ARBER) analysis technique and whole-brain functional magnetic resonance (fMRI) imaging or 18F-fluorodeoxyglucose positron emission tomography (PET) clearly shows that dorsal, but not ventral, part of the rat HF was activated by subcutaneous formalin injection [
23,
24]. Previous studies using electrophysiological [
25‐
31] and neurochemical/biochemical [
32‐
37] assays have demonstrated that the neuronal activities (pyramidal or interneuronal) and protein expression/activation within the HF could be altered by pain and stress. Moreover, intra-hippocampal microinjection of lidocaine [
38], or antagonists acting at N-methyl-D-aspartic acid (NMDA) receptor [
39,
40], 5-HT
2A/2C receptor [
41] and platelet-activating factor receptor [
42] could result in an analgesic effect in the formalin test. Clinical observations show that electrical stimulation of the HF evokes painful sensations in humans [
43,
44] and hippocampal lesion can partially alleviate chronic pain [
45,
46]. Taken together, the above previous reports provide convergent evidence for the critical involvement of HF in pain processing and support the possibility that there might be some kinds of synaptic plasticity occurred in the HF characterized by functional changes in synaptic transmission and modulation as well as structural changes in synaptic connection under the condition of peripheral persistent nociception.
With regard to the impact of pain upon the brain, it has been revealed that chronic pain states can change the structure and morphology of the brain, namely central structural plasticity, which probably results in long-term dysfunction of synaptic transmission and modulation at different levels of the central nervous system (CNS) [
47]. In addition, long-term potentiation (LTP), a form of functional neuroplasticity, was also found to be associated with pain processing in recent years, except for its wide use as a unique synaptic model for learning and memory [
48‐
50]. Actually, there have been a number of previous studies investigating LTP phenomenon in multiple pain-related CNS regions, including the spinal cord dorsal horn [
51‐
53], primary somatosensory cortex (S1 area) [
54,
55], amygdala [
55,
56], anterior cingulate cortex [
55,
57‐
61] and so on. As regards the HF, an enhanced LTP by pain was also reported in one previous study [
62]. In that study, it was found that LTP in the CA1 area could be facilitated by tail tip amputation-induced injury in mice and the increased synaptic efficacy was accompanied by a strong up-regulation of the immediate early gene product Egr1 [
62]. Because CA1 receives inputs from Schaffer collaterals of CA3 pyramidal cells which are innervated by entorhinal-dentate gyrus (DG) output [
15,
16,
22], it is of particular importance to see whether there are parallel changes in synaptic connection and function in the DG area in response to persistent nociception. Furthermore, based upon the studies from Khanna's group, the neuronal activities in CA1 differ from each other in response to formalin-induced nociception, namely at the time when a discrete population of putative pyramidal cells are selectively activated, a large number of CA1 cells are suppressed in a widespread and prolonged manner, implicating a 'signal-to-noise' processing of pain in the CA1 area [
27‐
31]. However, the interrelationship between different populations of CA1 neurons or between DG and CA1 regions are still not clear and requires to be further studied by multisite recording approaches.
The planar multi-electrode array (pMEA) is a unique and well-established tool for investigating, at a macroscopic level, the electrophysiological properties of living brain slices containing intact networks of neurons, providing a bridge between single cell testing and behavioral studies [
63]. Compared to traditional electrophysiology, the pMEA technique allows one to detect the activity of neuronal networks in both space and time [
63‐
65], to record multiple sites simultaneously [
66], and to make stimulating the recorded cells possible [
67]. To visualize the spatial and temporal information of pMEA recording, two-dimensional current source density (2D-CSD) imaging can also be used [
68,
69]. Therefore, in the present study, using pMEA (i.e., Panasonic's MED64 system, see [
68‐
70]) recordings combined with 2D-CSD imaging on acute hippocampal slices, we examined potential effects of peripheral persistent nociception on spatial and temporal plasticity of synaptic connection and function in the HF. The animal pain models we used are the bee venom (BV) test and the formalin test, both of them being well-developed animal models of persistent, inflammatory pain [
71‐
76]. The results showed robust changes in both spatial and temporal plasticity of synaptic connection and function following peripheral persistent nociception.
Methods
Animals
Experiments were carried out on male albino Sprague-Dawley rats provided by Laboratory Animal Facilities of both Capital Medical University (CCMU) and the Fourth Military Medical University (FMMU). All animals were with ages of 4 weeks old (weighing 120-160 g) and were housed in groups of five per cage under controlled laboratory conditions (12 h light/12 h dark, temperature 22-26°C, air humidity 55-60%). They had free access to commercial rat pellets and tap water. The experimental procedures were approved by the Institutional Animal Care and Use Committee at both CCMU and FMMU. All animals were maintained and cared for in compliance with the guidelines set forth by the International Association for the Study of Pain [
95]. The number of animals used and their suffering were greatly minimized. The experiments were blinded; all experimental rats were randomly divided into five groups: (1) naïve rats without treatment; (2) rats with subcutaneous injection of 0.9% sterile, isotonic saline solution; (3) rats with subcutaneous injection of whole BV solution; (4) rats with subcutaneous injection of formalin; and (5) rats with subcutaneous injection of 0.6 ml bupivacaine (0.25%) into the hind paw 10 min prior to ipsilateral BV treatment.
Induction of persistent pain
Persistent pain was induced with the BV test as described previously [
72]. The BV used in this study was lyophilized whole venom from
Apis mellifera (Sigma, St. Louis, MO) dissolved in 0.9% sterile saline. A volume of 50 μl saline containing 0.2 mg BV was used during the whole experiment, because previous studies have shown that 4 μg/μl was the optimal dose to produce a prolonged pain-related behavioral response [
71]. With respect to the formalin text, for each injection, 0.05 ml of 5% formalin (37.5-40% formaldehyde solution diluted in 0.9% sterile saline) was used in the present study [
74,
92,
94]. The whole BV or diluted formalin solution was administered by subcutaneous injection into the posterior plantar surface of the left hind paw of rats [
72]. The animals were carefully handled during the process to reduce the possible interruption of results caused by handling-induced stress. Intraplantar injection of the same volume of physiological saline served as the control group.
Preparation of multi-electrode array
Procedures for the preparation of the Multi-Electrode Dish (Panasonic, MED probe) were almost the same as described by [
70]. The device had an array of 64 planar microelectrodes, each 50 × 50 μm in size, arranged in an 8 × 8 pattern (inter-electrode distance, 300 μm). The microelectrode's large size resulted in lower impedance, enabling both reliable stimulation and a higher signal to noise ratio when recording. Before use, the surface of the MED64 probe was treated with 0.1% polyethyleneimine (Sigma, St. Louis, MO; P-3143) in 25 mM borate buffer (pH 8.4) overnight at room temperature. This coating helped establish sufficient adhesion of the slice to the probe surface, resulting in enough perfusion by the recording buffer (2-3 ml/min) to keep the slice healthy for more than 6 h of fEPSP recording [
70]. The probe surface was rinsed three to five times with sterile distilled water before immediate use. In general, the MED64 probes could be re-used for approximately 30-40 recording sessions with a mean duration of 4-6 h. Electrode properties could be maintained constant by carefully cleaning the probe with deionized water following each recording session.
Preparation of acute hippocampal slices
The general procedures for preparing acute hippocampal slices were similar to those described previously [
68‐
70]. Male Sprague-Dawley rats aged 25-30 days were sacrificed by decapitation after anesthesia with 4% sodium pentobarbital (0.1 ml/100 g, i.p.) 2 h after subcutaneous saline or BV or formalin injection. Subsequently, the whole brain was rapidly removed and immediately soaked in ice-cold, oxygenated preparation buffer of artificial cerebrospinal fluid (ACSF) for approximately 1-2 min. The ACSF contained 124 mM NaCl, 3.3 mM KCl, 1.2 mM KH
2PO
4, 2.4 mM MgSO
4, 10 mM glucose, 26 mM NaHCO
3, 2.5 mM CaCl
2, and had a pH of 7.4 adjusted by gassing with 5% CO
2/95% O
2. Appropriate portions of the brain were then trimmed and the remaining brain block was placed on the ice-cold stage of a vibrating tissue slicer (Dosaka, DTK-1000). Here, it deserves mentioning that all the present experiments were performed on the right (i.e. contralateral to the BV injection side) anterior hippocampus of rats in three groups, ranging from Bregma -2.52 mm to Bregma -4.08 mm according to the Atlas of the Rat Brain [
96]. The stage was immediately filled with oxygenated and frozen ACSF. The thickness of each tissue slice was set at 350-400 μm. Each slice was gently taken off the blade by a writing brush, trimmed, and immediately soaked in an incubation chamber containing the oxygenated ACSF for 2 h at room temperature.
Electrophysiological recordings
After incubation, one slice was selected and positioned on the MED64 probe in such a way that the whole HF was entirely covered by the 8 × 8 array. Once the slice settled, a netting ballast (U-shaped platinum wire with regularly spaced hair pieces) was carefully disposed on the slice to immobilize it. For the electrophysiological recordings, the probes with immobilized slices were connected to the stimulation/recording component of MED64. The slice was continuously perfused with oxygenated, fresh ACSF at the rate of 2-3 ml/min with the aid of a peristaltic pump (PERI-STAR™, WPI, USA). After a 20 min recovery of the slice, one of the 64 available planar microelectrodes was selected from the 64-switch box for stimulation following visual observation through a charge-coupled device camera connected to an inverted microscope. When not specified, monopolar, biphasic constant current pulses (30-199 μA, 0.1 ms duration) generated by the data acquisition software were applied to the PP at 0.1 Hz. Field potentials evoked at the remaining sites were amplified by the 64-channel main amplifier and then digitized at a 20 kHz sampling rate. The digitized data were displayed on the monitor screen and stored on the hard disk of a microcomputer. Five successive responses were averaged automatically in real time by the recording system. The viability of the slices was kept constant across different sets of recording sessions by measuring the threshold for evoking fEPSP of adequate amplitude.
Experimental procedures
After selecting the best stimulation site and stabilizing the synaptic responses for about 30 min, an I-O curve was first determined for each group using the measurements of fEPSP amplitude or slope in response to a series of stimulation intensities from 30 μA to 199 μA (30 μA, 60 μA, 90 μA, 120 μA, 150 μA, 180 μA, 199 μA). Because of the technical limits of the stimulus generator, higher intensities (>199 μA) could not be applied and were not tested in the present study. The intensity of the test stimulus was then adjusted to elicit 40-60% of the maximum based on the I/O curves. Next, the stability of the whole recording system was checked by recording baseline responses for another 30 min (3 × 10 min). For LTP induction, the TBS protocol was used, which consisted of 10 bursts, each containing 4 pulses at 100 Hz with an inter-burst interval of 200 ms. It is widely accepted that such a protocol resembles in vivo conditions and has been suggested as a method to establish a link between artificial and natural synaptic activity [
97]. In addition, LTP induced by such stimulation appears to be more robust and stable than that induced by other means [
98]. To standardize tetanization strength in different experiments, the TBS strength was set at an intensity evoking almost half of the maximal magnitude of fEPSP. After TBS, the test stimulus was repeatedly delivered (at the identical intensity as baseline) once every 10 min for more than 2 h to allow for the observation of any changes in LTP magnitude and duration.
In experiments regarding pharmacological characterization of fEPSP, the stability was first determined by recording the baseline responses for about 30 min as described above. Then, TTX (0.5 μM and 1 μM), AP5 (50 μM and 100 μM) and CNQX (10 μM) were bath applied to separate slices at a rate of 2 ml/min. For the high magnesium-low calcium solution, the concentration of CaCl2 was lowered to 0.25 mM and the concentration of MgSO4 was raised to 4.0 mM in the ACSF. In another set of experiments, either AP5 (100 μM) or CNQX (10 μM) was bath applied at 2 h after LTP induction (a sustained peak level reached at this time) to observe their actions on spatial and temporal plasticity in the HF. In any of the above cases, complete solution exchange was achieved within 10 min of drug infusion or ionic substitution. Subsequently, fresh ACSF washed in until the drug effects vanished and the normal synaptic responses essentially recovered.
Current source density analysis
In this study, the 2D-CSD was computed in an attempt to identify current sources and sinks in any direction within the plane of each hippocampal slice. In general, the 2-dimensional current density
I
m
in the presence of a field potential Φ was given as
Since the measured field potential (Φ
i, j
) was recorded on a planar array of 64 electrodes, the second partial derivatives at the center of particular electrodes could be computed from the measured field potential on that electrode and its neighbors as
By considering the medium as ohmic with homogeneous conductance (
σ
X
=
σ
Y
=
σ), and under the assumption of equidistant electrodes (Δ
X
= Δ
Y
= Δ), the normalized CSD (I*
i, j
) could be defined and computed as
With the normalized CSD values at the center of the electrodes, it became possible to compute the density at any point (x, y) within the 8 × 8 array using bilinear interpolation. After all of the above calculations, we used the color yellow to represent positive currents (sources), the color blue to represent negative currents (sinks), and the color black to map zero current. Finally, CSD images at selected time points were plotted across all 64 recording sites for each group of slices.
Drugs
All drugs were purchased from Sigma-Aldrich and, except for CNQX, were dissolved in deionized water as stock solutions for frozen aliquots. They were diluted to the desired concentration in ACSF before immediate use. CNQX was dissolved in dimethylsulfoxide (DMSO, final concentration 0.1%) and prepared as described above.
Offline analysis
For quantification of the I-O relationship, the amplitude and slope of fEPSP were analyzed off line by the MED64 Conductor. For LTP data, the amplitude and slope of evoked fEPSP were normalized and expressed as a percentage of the averaged value measured during the last 10 min baseline period. Evaluations of drug effects were carried out on the basis of the difference between the pre-drug recording and the 10 min after drug infusion (when the drug effect was most potent). The total number of effective fEPSP (> 20% baseline) reliably recorded over the HF was counted by an experimenter unaware of the experimental design and averaged across slices for each group. Data sets included results from only one slice per rat (n = number of slices). All data were presented as mean ± S.E.M. When necessary, the statistical significance was determined using either the Student's t test (paired and two-independent sample) or one-way ANOVA (post-hoc Fisher's PLSD). The level of P < 0.05 was assumed as statistically significant.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
X-YZ, M-GL, D-LY and Y-W mainly, YH and D-DW partially collected the primary data; X-FC established the 2D-CSD analysis method and conducted the imaging transformation work; F-KZ and HL helped in MED64 system setup and slice preparation; JC and X-SH designed the experiments; JC and M-GL wrote the manuscript. All authors have read and approved the manuscript finally.