If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Guangdong Key Laboratory of Clinical Molecular Medicine and Diagnostics, Guangzhou First People's Hospital, School of Medicine, South China University of Technology, Guangzhou, Guangdong, China
Guangdong Key Laboratory of Clinical Molecular Medicine and Diagnostics, Guangzhou First People's Hospital, School of Medicine, South China University of Technology, Guangzhou, Guangdong, China
Address reprint requests to Xiangcai Ruan, Department of Anesthesia and Pain Medicine, The Sixth Affiliated Hospital, Sun Yat-Sen University, Guangzhou, Guangdong 510655, China.
Projections from the periaqueductal gray (PAG) to the rostral ventromedial medulla (RVM) are known to engage in descending pain modulation, but how the neural substrates of the PAG-RVM projections contribute to neuropathic pain remains largely unknown. In this study, we showed somatostatin-expressing glutamatergic neurons in the lateral/ventrolateral PAG that facilitate mechanical and thermal hypersensitivity in a mouse model of chemotherapy-induced neuropathic pain. We found that these neurons form direct excitatory connections with neurons in the RVM region. Inhibition of this PAG-RVM projection alleviates mechanical and thermal hypersensitivity associated with neuropathy, whereas its activation enhances hypersensitivity in the mice. Thus, our findings revealed that somatostatin neurons within the PAG-RVM axial are crucial for descending pain facilitation and can potentially be exploited as a useful therapeutic target for neuropathic pain.
Perspective
We report the profound contribution of somatostatin neurons within the PAG-RVM projections to descending pain facilitation underlying neuropathic pain. These results may help to develop central therapeutic strategies for neuropathic pain.
The midbrain periaqueductal gray (PAG) and its descending projections to the rostral ventromedial medulla (RVM) is an essential component of descending pain modulation system.
Recently, the ventrolateral PAG has attracted increasing attention for its function in descending pain facilitation. The magnitude of ventrolateral PAG functional connectivity is correlated to the spontaneous and allodynic pain in patients with painful diabetic polyneuropathy.
Increased firing of ventrolateral PAG neurons seems to be a key reason for allodynia and hyperalgesia to occur in another form of painful polyneuropathy, ie, chemotherapy-induced neuropathic pain (CIPN).
have been identified to strengthen synaptic input to ventrolateral PAG after peripheral nerve injury. However, how the neural substrates of the PAG-RVM projections contribute to neuropathic pain remains understudied.
Neuronal activity of glutamatergic and GABAergic neurons within the l/vlPAG have been demonstrated to play critical roles in regulating pain and itch sensation.
Furthermore, l/vlPAG glutamatergic neurons contain 2 non-overlapping subtypes, of which tachykinin 1-expressing neurons play a key role in facilitating itch but not pain processing, while somatostatin-expressing (l/vlPAGSST) neurons are not associated with itch-induced scratching behavior.
However, knowledge on the role of these somatostatin neurons in pain processing remains elusive. Accordingly, we hypothesize that l/vlPAGSST neurons participate in the PAG-RVM descending pain-modulating system and may have an important role in the central processing of neuropathic pain.
This study set to dissect the functional connections of l/vlPAGSST neurons between PAG and RVM in a mouse model of CIPN that develops profound mechanical and thermal hypersensitivity.
Male C57BL/6N and SST-Cre mice were used in this study. C57BL/6N mice were brought from either Guangdong Provincial Medical Laboratory Animal Center (Guangzhou, China). SST-Cre mice (JAX: 010708) were originally purchased from the Jackson Laboratory. All experiments and procedures in this study included appropriate measures to relief suffering and pain, and also use as few animals as possible, were approved by and done in accordance with the Animal Care and Use Committee of Sun Yat-Sen University, Guangzhou, China. The Experimental Animals were housed separately on a 12:12 h light-dark cycle (lights on at 7:00 am) with access to food and water ad libitum.
Oxaliplatin Injection
Dissolved oxaliplatin (Shanghai Aladdin Reagent, China) in 5% glucose solution (1 mg/mL) were i.p injected with 4 mg/kg to the CIPN mice. The control animals received an equivalent amount of 5% glucose vehicle prepared in sterile distilled water. The chemotherapy oxaliplatin or its vehicle control was i.p injected once a day for 5 consecutive days. Behavior Tests were used to assess the emergence and maintenance of neuropathic pain.
von Frey Test
Mice were adapted to the test room for a minimum of 2 days before the behavior test. The following behavioral tests were conducted in a double-blind manner in respect to the treatment or mouse.
To test mechanical sensitivity, von Frey test used a series of von Frey filaments (0.16–2.56 g, Stoelting, Wood Dale, IL) with logarithmically increasing stiffness to vertically stimulate the hind paws of mice. The withdrawal threshold was determined using the up-down testing paradigm.
In some experiments, we also measured the frequency of withdrawal reflex to mechanical stimuli. We made 3–5 replicate measurements from each hind paw of each mouse, and the values of the 2 paws were averaged.
Hargreaves Test
In order to study radiative heat pain, a Hargreaves apparatus was used to determine paw withdrawal latency to a noxious heat stimulus, which is the average value of at least 4 measurements of each paw in the 5-minute test period with a 20 s cut-off time to avoid tissue injury. We made 3–5 replicate measurements from each hind paw of each mouse, and the values of the 2 paws were averaged.
Hot Plate Test
To test the thermal nociception, a transparent plexiglass cylinder on the plate (50°C) was utilized. the onset of rapid rear paw lifting and/or rear paw retraction/licking was recorded.
Tail Flick Test
For tail flick test, the mice were gently confined in cotton gloves to avoid stressful responses. Then the distal one-third of the tail was immersed into water at 52°C to record the tail flick latency. A cut-off time of 10 s was set to avoid tissue injury.
Open-Field Test
As previously reported, a 40 × 40 × 40 cm polystyrene cube was used for the test to assess locomotor activity.
The mice were initially placed in the rear left square of the cube and allowed to explore freely during a 10 min period. Locomotor traces videotyped were analyzed with LabState software (AniLab, Ningbo, China). The cube was cleaned with a 70% ethanol and dried for each individual mouse test.
Stereotaxic Injections
Mice were anesthetized (pentobarbital sodium 100 mg/kg, i.p.; MYM Technologies.) in a stereotactic device (RWD Life Science, 68513). After exposing the skull surface through the midline incision, the bilateral skull windows were exposed. A hole (about 1.5 mm in diameter) was drilled to introduce a calibrated glass microelectrode into the target nuclei. The microelectrode was connected to an infusion pump (Drummond, Nanoject III) with an injection rate of 20–30 nL/min into the target nuclei. The target nuclei were located as anterior–posterior (AP) from bregma, medio-lateral (ML) from the midline, and dorso-ventral (DV) from the brain surface in mm.
To manipulate neuronal activity of l/vlPAGSST neurons, SST-Cre mice of 6 to 8 weeks old were utilized. The l/vlPAG area (AP, −4.6 mm; ML, ±0.5 mm; DV, −2.7 mm) was located and injected with AAV-hSyn-DIO-hM3Dq-mCherry (AAV2/5, 5.0 × 1012 v.g./mL, S0192-5) or AAV-hSyn-DIO-hM4Di-mCherry (AAV2/5, 4.7 × 1012 v.g./mL, S0193-5). The l/vlPAG area was injected bilaterally in a volume of 300 nL at a rate 20–30 nL/min for each side 6IU. AAV-hSyn-DIO-mCherry (AAV2/5, 5.1 × 1012 v.g./mL, S0240-5) in equivalent volume and infusion speed was used as control.
To determine the projection neurons, 2-m-old wild type (WT) mice were used and AAV-hSyn-ChR2-EYFP (AAV2/9, 5.3 × 1012 v.g./mL, S0318-9) was unilaterally injected into the l/vlPAG location. Details of the injection procedure were same as the viral injections described above. A month later, tetrodotoxin (TTX) and 4-Aminopyridine (4-AP) were used to conduct slice electrophysiology experiments described as below.
To manipulate the l/vlPAGSST projection to the RVM, a retrograde AAV tracer strategy was utilized. SST-Cre mice of 6 to 8 weeks old were bilaterally injected with AAV2/Retro-FLEX-Flpo (AAVrg, 1.0 × 1013 v.g./mL, S0273-2R) into the RVM. Location of the RVM was as follows: AP, −6.00 mm; ML, 0.00 mm; DV, −5.85 mm. Simultaneously, AAV-fDIO-hM3Dq-mCherry (AAV2/9, 4.7 × 1012 v.g./mL, S0337-9) or AAV-fDIO-hM4Di-mCherry (AAV2/9, 4.4 × 1012 v.g./mL, S0336-9) was bilaterally injected into the l/vlPAG. Control was a same volume of AAV-fDIO-mGFP (AAV2/5, 5.2 × 1012 v.g./mL, S0289-5). Details of the injection procedure were same as the viral injections described above. All the AAV viruses in this study were purchased from Shanghai Taitool Bioscience (Shanghai, China)
Pharmacogenetic Manipulations
The binding of Clozapine-N-oxide (CNO) with hM4Di− or hM3Dq− expressing neurons can hyperpolarize or depolarize the cell membrane potential, and thus inhibit or enhance the activity of targeted neurons. For in vivo chemogenetic manipulation, the mice was i.p. injected with CNO (Sigma, C0832; 1 mg/kg,), and pain-like behaviors was tested 30–40 min later.
Immunofluorescent Staining
After behavioral tests, the rats were overdosed with sodium pentobarbital (i.p., 100 mg/kg) and fixed transcardially. The brain was removed and fixed in PFA (4%) overnight at 4°C, and then frozen in sucrose solution (30%) in PBS at 4°C for dehydration. The brain was then sectioned (40 μm thick) using a vibratome (Leica CM 1950, Germany) and immunohistochemical staining was performed. The tissue sections were blocked in PBST (3% BSA, PBS with 0.3% Triton X-100) containing 5% normal donkey serum (NDS, Abcam) at room temperature for 30 minutes. The brain sections were then incubated with primary antibody at 4°C overnight, and incubated with secondary antibody at room temperature for 2–3 hours. Our primary antibodies used for immunohistochemistry (IHC) were rabbit anti-DsRed (1:1000, Clontech) and rabbit anti-GFP (1:1000, Thermo Fisher Scientific), and our secondary antibodies included Cy3-conjugated donkey anti-rabbit IgG and donkey Alexa 488-conjugated anti-rabbit IgG (both 1:500, Jackson ImmunoResearch Laboratories). Images were visualized using a Nikon Eclipse Ni-E fluorescence microscope, and a Fiji image processing package was use to count cell number automatically or manually.
SE mice were overdosed with sodium pentobarbital and coronal midbrain slices (220 µm) were generated with a vibrating blade microtome (VT1200S, Leica Biosystems) in an ice-cold solution (in mM): sucrose, 50; NaCl, 95; KCl, 1.8; MgSO4, 7; KH2PO4, 1.2; NaHCO3, 26; CaCl2, 0.5; and glucose, 15 (300–305 mOsm) and then transferred to 34°C oxygenated ACSF containing (in mM) NaCl, 127; KCl, 1.8; KH2PO4, 1.2; CaCl2, 2.4; MgSO4, 1.3; NaHCO3, 26; and 15 (300–305 mOsm) glucos. The recording chamber was perfused with oxygenated ACSF.
Patch pipettes with a resistance of 3–5 MΩ were used for whole-cell patch-clamp recordings. High power blue light (∼475 nm, UHP-Mic-LED-475, Prizmatix) was used to activate Channelrhodopsin-2 and promote neuronal depolarization. In current clamp recordings, the evoked postsynaptic responses were recorded in voltage-clamp mode after a 7 to 10 min stabilization in cesium (Cs2+)-based internal solution. The internal solution contained (in mM) CsMeSO3, 130; Na-GTP, 0.3; HEPES, 10; CaCl2, 1; QX-314, 2; EGTA, 11; MgCl2, 1; and Mg-ATP, 2 with pH 7.3 and ∼295 mOsm. To record the laser-evoked excitatory postsynaptic currents (EPSCs), holding voltage was set at -70 mV, and 0 mV to record the laser-evoked inhibitory postsynaptic currents (IPSCs) recording. 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(f)quinoxaline (NBQX, Tocris, 1044) 10 μM was used to inhibit AMPA receptors, and GABAA receptors were blocked with Picrotoxin (PTX, 1128) 50 μM. TTX (Abcam, ab120054) 1 μM and 4-AP (Sigma-Aldrich, 275875) 100 μM were used to block action potential based synaptic transmission and re-establish monosynaptic transmission under blue light evocation. The paired pulse ratio (PPR) was determined by paired pulse stimulation (1ms; 50ms pulse interval) as the ratio of the second response amplitude to the first response amplitude.
The amplitude and latency of EPSCs and IPSCs were manually assessed from the averaged traces low-pass filtered at 4 kHz and digitized at 20 kHz (Digidata 1550; Molecular Devices). When the EPSCs and the IPSCs have multiple peaks, the latency of the first emerged 1 was noted. RVM neurons of mice injected with oxaliplatin were recorded in a week after injection.
Statistical Analysis
All data are expressed as the mean ± SEM. Prism (GraphPad software) was used for statistical analysis. All statistical analyses were two tailed comparisons. Mann Whitney test, Wilcoxon signed rank test, two-way ANOVA with Bonferroni's post hoc test, and paired t-test were used to analyze the data when appropriate. Statistical significance was set at P < .05.
Results
l/vlPAGSST Neurons Facilitate Hypersensitivity
To examine the role of l/vlPAGSST neurons in central nociceptive perception, chemogenetic method of “designer receptors exclusively activated by designer drugs” (DREADDs)
were utilized to manipulate the activity of l/vlPAGSST neurons in SST-Cre mice. Adeno-associated virus (AAV) encoding the inhibitory DREADD receptor hM4Di was injected into the l/vlPAG bilaterally (Fig 1A and 1B). After sufficient expression of hM4Di, we tested pain-like behaviors through the von Frey test, Hargreaves test, hot plate test, and tail flick test 30 min after i.p. injection of CNO. Open-field test was utilized to exclude possible motor deficits associated with the intervention. Chemogenetic silencing of l/vlPAGSST neurons significantly increased withdrawal thresholds and withdrawal latencies in the von Frey, Hargreaves, hot plate, and tail flick tests (Fig 1C–1F, respectively). The total traveling distance in the OFT remained unaffected (Fig 1G), indicating that the changes in the pain-like behavior tests are unlikely to arise due to a motor deficit associated with the intervention. These findings indicate that l/vlPAGSST neurons have a proalgesic role in the mediation of pain perception.
Figure 1Chemogenetic silencing of l/vlPAGSST neurons attenuates hypersensitivity. (A) Schematic injection of AAV-DIO-hM4Di-mCherry or AAV-DIO-mCherry into bilateral l/vlPAG. (B) hM4Di-mCherry-labelled neurons in the l/vlPAG. Scale bar, 300 μm. (C–F) Withdrawal thresholds and withdrawal latencies after administration of CNO (1 mg/kg, i.p.) or vehicle on pain-like behaviors in von Frey test (C), Hargreaves test (D), hot plate test (E), and tail flick test (F). Two-way ANOVA (n = 10 SST-Cre mice). (G) Locomotive activity after chemogenetic inhibition of l/vlPAGSST neurons. Two-way ANOVA (n = 10 SST-Cre mice). (H) Top, Time course of vehicle or oxaliplatin injection-induced pain-like behaviors. Two-way ANOVA with Bonferroni's post hoc test (n = 9 SST-Cre mice). Bottom, Timeline of the behavioral experiment. (I–L) Withdrawal thresholds and withdrawal latencies after administration of CNO or vehicle on pain-like behaviors of the neuropathic mice in von Frey test (I), Hargreaves test (J), hot plate test (K), or tail flick test (L), Two-way ANOVA (n = 10 SST-Cre mice). All error bars represent SEMs. *P < .05, **P < .01, and ***P < .001. Oxal, oxaliplatin.
The SST-Cre mice received 5 consecutive oxaliplatin administration 21 days after the AAV injection. Ten days later, the mice were given pain-like behavior tests after CNO administration. We found that the mechanical withdrawal threshold was gradually decreased during the 10-day period after 5 consecutive oxaliplatin administrations (Fig 1H). Silencing of l/vlPAGSST neurons with CNO significantly alleviated mechanical and thermal hypersensitivity in these mice (Fig 1I–1L).
Next, we used excitatory DREADD to assess whether the activation of l/vlPAGSST neurons enhanced pain hypersensitivity. After injection of AAV-DIO-hM3Dq-mCherry or its control into the l/vlPAG bilaterally (Fig 2A and 2B), the SST-Cre mice received 5 consecutive oxaliplatin administrations 21 days after the virus injection. We consistently found that chemogenetic activation of l/vlPAGSST neurons with CNO further decreased withdrawal thresholds and withdrawal latencies in the von Frey, Hargreaves, and hot plate tests (Fig 2C–2E, respectively). In addition, the activation of l/vlPAGSST neurons did not significantly affect the total traveling distance in the open-field test (Fig 2F). Together, our results indicate that l/vlPAGSST neurons play a facilitatory role in neuropathic pain sensitization.
Figure 2Chemogenetic activation of l/vlPAGSST neurons facilitates pain processing. (A) Schematic injection of AAV-DIO-hM3Dq-mCherry or AAV-DIO-mCherry into the bilateral l/vlPAG. (B) hM3Dq-mCherry-labelled neurons in l/vlPAG. Scale bar, 300 μm. (C–E) Withdrawal thresholds and withdrawal latencies after administration of CNO (1 mg/kg, i.p.) or vehicle on pain-like behaviors in von Frey test (C), Hargreaves test (D), hot plate test (E), and tail flick test (F). Two-way ANOVA with Bonferroni's post hoc test (n = 9 SST-Cre mice). *P < .05, **P < .01, and ***P < .001. Error bars represent SEMs.
We then assessed whether l/vlPAGSST neurons also participate in the PAG-RVM descending pain-modulating system. As shown in Fig 3A, we obtained whole-cell patch-clamp recordings from neurons in RVM slices after injection of AAV-DIO-ChR2-EYFP into the l/vlPAG of SST-Cre mice (Fig 3A). With bath application of TTX and 4-AP, photostimulation to the l/vlPAGSST neurons evoked EPSC but not IPSC in the RVM neurons (Fig 3B), consistent with previous finding that l/vlPAGSST neurons are glutamatergic.
The evoked EPSCs were blocked by NBQX (Fig 3C), suggesting a direct excitatory connection between l/vlPAGSST neurons and the RVM neurons.
Figure 3Increased l/vlPAGSST glutamatergic inputs to RVM neurons. (A) Schematic injection of AAV-DIO-ChR2-EYFP into l/vlPAG and patch-clamp site in RVM. (B) Example traces of blue-laser-evoked EPSCs in an RVM neuron in ACSF (black) and in the absence and presence of NBQX, TTX, and 4-AP (red). Blue bar, blue-light stimulation (475 nm, 1 ms). A total 10 of 12 recorded RVM neurons from six mice showed light-evoked EPSCs. (C) Analysis of blue-laser-evoked EPSCs in RVM neurons in the absence or presence of NBQX. Wilcoxon signed-rank test (n = 6 neurons from 6 mice). (D) Example traces of blue-laser-evoked EPSCs in RVM neurons from mice treated with vehicle or oxaliplatin. Blue bar, LED stimulation (475 nm, 1 ms). (E) Oxaliplatin induced larger EPSCs in RVM neurons. Mann-Whitney test (n = 10 neurons from 4 mice treated with vehicle, n = 10 neurons from 4 mice treated with oxaliplatin). (F) PPRs (P2/P1) for l/vlPAGSST glutamatergic neurons in RVM from vehicle and oxaliplatin mice. Mann-Whitney test (n = 10 neurons from 4 mice treated with vehicle, n = 10 neurons from 4 mice treated with oxaliplatin). Error bars represent SEMs. *P < .05, **P < .01, and ***P < .001. EPSCs, evoked excitatory postsynaptic currents; Oxal, oxaliplatin.
CINP Increases the Excitability of l/vlPAGSST-RVM Projection
We then performed slice recording experiments combined with optogenetics to investigate the potential role of the l/vlPAGSST-RVM circuit in the development of CINP. SST-Cre mice received 5 consecutive injections of oxaliplatin or vehicle after injection of AAV-DIO-ChR2-EYFP into the l/vlPAG bilaterally. We were able to detect that photostimulation to the l/vlPAGSST neurons evoked much larger EPSCs in the RVM neurons from the oxaliplatin mice than the vehicle mice (Fig 3D and 3E). The PPR of individual neuron also showed paired-pulse depression in the RVM of the neuropathic mice (Fig 3D and 3F), suggesting that neuropathic pain may increase the l/vlPAGSST-RVM synapse strength and the excitatory inputs of RVM neurons.
Activation of l/vlPAGSST-RVM Pathway Alleviates Hypersensitivity
Next, we examined the functions of l/vlPAGSST-RVM projections using a retrograde dual-recombinase system.
Three weeks after injection of AAV-fDIO-hM3Dq-mCherry or control into the l/vlPAG and AAVrg-FLEX-Flpo into the RVM (Fig 4A and 4B), the SST-Cre mice received 5 consecutive oxaliplatin administrations and activation with CNO. As expected, activation of this projection with CNO enhanced the decreased withdrawal thresholds and withdrawal latencies in the neuropathic mice (Fig 4C–4F). The effect of locomotor activity on pain-like behaviors was excluded by showing that locomotor activity was not significantly affected (Fig 4G).
Figure 4Chemogenetic manipulation of l/vlPAGSST–RVM projection modulates pain-like behaviors. (A) Schematic injection of AAV-fDIO-hM3Dq-mCherry or AAV-fDIO-mGFP into l/vlPAG and AAVrg-FLEX-Flpo into RVM. (B) hM3Dq-mCherry-labelled neurons in l/vlPAG. Scale, 300 μm. (C–F) Withdrawal thresholds and withdrawal latencies after administration of CNO (1 mg/kg, i.p.) or vehicle on pain-like behaviors in von Frey test (C), Hargreaves test (D), hot plate test (E) and acetone test (F), with two-way ANOVA (n = 12 SST-Cre mice). (G) Locomotive activity after chemogenetic activation of l/vlPAGSST projection. Two-way ANOVA (n = 12 SST-Cre mice). (H) Schematic injection of AAV-fDIO-hM4Di-mCherry or AAV-fDIO-mGFP into l/vlPAG and AAVrg-FLEX-Flpo into RVM. (I) hM4Di-mCherry-labelled neurons in the l/vlPAG. Scale bar, 300 μm. (J–M) Withdrawal thresholds and withdrawal latencies after administration of CNO (1 mg/kg, i.p.) or vehicle on pain-like behaviors in von Frey test (J), Hargreaves test (K), hot plate test (L) and acetone test (M). Two-way ANOVA (n = 11 SST-Cre mice). (N) Locomotive activity after pharmacogenetic inhibition of l/vlPAGSST projection. Two-way ANOVA (n = 11 SST-Cre mice). All error bars represent SEMs. *P < .05, **P < .01, and ***P < .001.
Inhibition of l/vlPAGSST-RVM Pathway Enhances Hypersensitivity
We then injected AAV-fDIO-hM4Di-mCherry or control into the l/vlPAG and AAVrg-FLEX-Flpo into the RVM, and then treated the SST-Cre mice with oxaliplatin or vehicle 3 weeks after viruses’ injection (Fig 4H and 4I). Consistently, inhibition of the projection with CNO alleviated the decreased withdrawal thresholds and withdrawal latencies in the pain-like behavior tests (Fig 4J–4M, respectively), without a significant effect on the locomotor activity in these mice (Fig 4N). In addition, we also determined the spontaneous and affective components of pain utilizing conditioned place preference (CPP) testing with gabapentin (GBP; Fig. S1A and S1B). We found that inhibition of l/vlPAG neurons projecting to the RVM could also induce CPP in mice treated with oxaliplatin as GBP, whereas activation of these neurons alleviated the CPP effect of GBP (Fig. S1C and S1D). Altogether, these data indicate that l/vlPAGSST-RVM projection facilitates mechanical and thermal hypersensitivity associated with neuropathic pain.
Dissection of l/vlPAGSST Neurons Innervates Brain Areas
Next, we determined how l/vlPAGSST neurons are regulated by other brain regions using a Cre-dependent rabies-virus-based trans-synaptic retrograde tracing strategy. Two weeks after injection of AAV5-CAG-FLEX TVA-mCherry and AAV5-CAG-FLEX-RG into the l/vlPAG, EnvA-coated Rabies-GFP virus was injected into the same location as well (Fig 5A). We found that retrogradely labeled (GFP+) neurons were mainly located in the bed nucleus of stria terminalis (BNST), ventral medial preoptic area (VMPO), central amygdala (CeA), primary somatosensory cortex, trunk region (S1Tr), anterior hypothalamic area, anterior part (AHA), lateroanterior hypothalamic nucleus (LA), zona incerta (ZI), magnocellular nucleus of the lateral hypothalamus (MCLH), dorsomedial hypothalamus (DMH), ventromedial hypothalamus (VMH), cuneiform nucleus, (CnF), parabrachial nucleus (PBN), Barrington's nucleus (Bar), and nucleus raphe magnus (NRM). These regions may transmit sensory, motor, or mental information to l/vlPAGSST neurons and then modulate pain processing.
Figure 5Retrograde tracing results show brain regions projected with l/vlPAGSST neurons. (A) Coronal sections of injection site in l/vlPAG and monosynaptic rabies tracing results in dorsal raphe nucleus (DRN). (B–J) Example monosynaptic rabies tracing coronal sections of anatomical localization of the BNST, VMPO, CeA, BLA, S1Tr, AHA, LA, ZI, MCLH, DMH, VMH, CnF, PBN, Bar, LC, and NRM. Scale bars represent 300 μm (n = 3). BNST, bed nucleus of stria terminalis; VMPO, ventral medial preoptic area; CeA, central amygdala; BLA, basolateral amygdala; S1Tr, primary somatosensory cortex, trunk region; AHA, anterior hypothalamic area, anterior part; LA, lateroanterior hypothalamic nucleus; ZI, zona incerta; MCLH, magnocellular nucleus of the lateral hypothalamus; DMH, dorsomedial hypothalamus; VMH, ventromedial hypothalamus; CnF, cuneiform nucleus; PBN, parabrachial nucleus; Bar, Barrington's nucleus; LC, locus coeruleus; and NRM, nucleus raphe magnus.
Activation of medullary dorsal horn γ isoform of protein kinase C interneurons is essential to the development of both static and dynamic facial mechanical allodynia.
However, how the neural substrates of the most known circuit in the descending pain-modulating systems (ie, the PAG-RVM pathway) contribute to neuropathic pain remains unclear.
In this study, we identified l/vlPAGSST neurons facilitating mechanical and thermal hypersensitivity associated with neuropathy. We showed that these neurons form direct excitatory synapses with neurons in the RVM region. Importantly, inhibition of this l/vlPAGSST-RVM projection alleviates mechanical and thermal hypersensitivity in the mice, whereas its activation enhances mechanical and thermal hypersensitivity. Thus, our data suggest that this l/vlPAGSST-RVM projection modulates central mechanisms of descending pain facilitation underlying neuropathic pain.
The l/vlPAG neurons have critical roles in central modulation of somatosensory and interoceptive processing.
In the present study, we showed that direct chemogenetic inhibition of SST+ neurons in the l/vlPAG relieves hypersensitivity activation associated with peripheral neuropathy in mice, while chemogenetic activation of this subtype enhances the hypersensitivity activation, suggesting a proalgesic role of SST+ neurons in the descending pain-modulating system. This outcome seems to be opposite to the gating effect of PAG glutamatergic neurons in pain processing.
However, there might be subpopulations of PAG glutamatergic neurons other than the l/vlPAGSST neurons that execute descending pain inhibition, thereby suppressing pain signal processing. Although the global effects of l/vlPAG glutamatergic neurons lead to inhibitory control of pain processing, some specific subpopulations (eg, somatostatin neurons in this study) may have distinct pain-facilitatory functions. Additional experiments are needed to further dissect the functional roles of the discrete subpopulation of PAG neurons.
An ultrastructural study of the projections from the midbrain periaqueductal gray to spinally projecting, serotonin-immunoreactive neurons of the medullary nucleus raphe magnus in the rat.
Our study demonstrated that l/vlPAGSST neurons form direct excitatory synapses with RVM neurons. Mechanical and thermal hypersensitivity is associated with increased excitatory connection between l/vlPAGSST neurons and the RVM in neuropathic mice. Selective activation of RVM-projecting l/vlPAGSST neurons enhances pain hypersensitivity, whereas selective inhibition of the projection alleviates pain hypersensitivity in neuropathic mice. It is important to note as well that our findings did not include the intrinsic excitability and spontaneous synaptic activity of RVM neurons receiving projections from l/vlPAGSST neurons, which should be determined in future studies. Altogether, these findings suggest that chemotherapy-induced hyperexcitation of the l/vlPAGSST-RVM projection plays a facilitatory role in the central processing of neuropathic pain.
The descending pathways could be more complicated, as l/vlPAGSST afferents could form synapses with a mixture of RVM subpopulations projecting to the dorsal horn of the spinal cord.
In this study, we focused the functional role of l/vlPAGSST neuron in promotion of pain perception; however, we could not rule out the possibility that other subtypes of neurons in the PAG also facilitate neuropathic pain. l/vlPAGSST neurons could also influence pain sensation through other brain regions aside from the RVM. For example, projection from l/vlPAG to the locus coeruleus has been demonstrated to suppress descending analgesia.
Taking advantage of an RV-based retrograde tracing method, we found that l/vlPAGSST neurons also project to various brain regions that transmit sensory, motor, and mental information. These findings further support the notion that the PAG can serve as a hub for multisensory integration.
In summary, our findings reveal that l/vlPAGSST neurons anatomically and functionally project to the RVM and exert a descending facilitatory effect on neuropathic pain responses in male mice. This central mechanism may offer a potential central therapeutic target for the treatment of neuropathic pain.
Author contributions
YZ, XH, WX, and XR designed experiments. Animal breeding, behavioral testing, data analysis, and the construction of the manuscript was performed by YZ, XH, WX, HL, JD, and XR performed experiments. All authors contributed to editing and revisions of the manuscript.
An ultrastructural study of the projections from the midbrain periaqueductal gray to spinally projecting, serotonin-immunoreactive neurons of the medullary nucleus raphe magnus in the rat.
Activation of medullary dorsal horn γ isoform of protein kinase C interneurons is essential to the development of both static and dynamic facial mechanical allodynia.
This project was supported by Guangzhou Science, Technology and Innovation Commission (grant 202206060004 to XW; grant 201607010119 to XR), Natural Science Foundation of Guangdong Province (grants 2020A151501012 to XR), and National Natural Science Foundation of China (grants 82171209 and 81271196 to XR).
The authors have no conflicts of interest to declare.
00010-X&id=gr1.jpg){kind=link}
00010-X&id=gr2.jpg){kind=link}
00010-X&id=gr3.jpg){kind=link}
00010-X&id=gr4.jpg){kind=link}
00010-X&id=gr5.jpg){kind=link}