Identification and characterization of rostral ventromedial medulla (RVM) neurons synaptically connected to the urinary bladder afferents in female rats with or without neonatal cystitis

Abstract

The neurons in the rostral ventromedial medulla (RVM) play a major role in pain modulation. We have previously shown that early-life noxious bladder stimuli in rats resulted in an overall spinal GABAergic disinhibition and a long-lasting bladder/colon sensitization when tested in adulthood. However, the neuromolecular alterations within RVM neurons in the pathophysiology of early life bladder inflammation have not been elucidated. In this study, we have identified and characterized RVM neurons that are synaptically linked to the bladder and colon and examined the effect of neonatal bladder inflammation on molecular expressions of these neurons. A transient bladder inflammation was induced by intravesicular instillation of protamine sulfate and zymosan during postnatal days 14 through 16 (P14–16) followed by pseudorabies virus PRV-152 and PRV-614 injections into the bladder and colon, respectively, on postnatal day P60. Tissues were examined 96 hours post-inoculation for serotonergic, GABAergic, and enkephalinergic expressions using In situ Hybridization and/or Immunohistochemistry techniques. The results revealed that >50% of RVM neurons that are synaptically connected to the bladder (i.e., PRV-152+) were GABAergic, 40% enkephalinergic, and about 14% expressing serotonergic marker TpH2. Neonatal cystitis resulted in a significant increase in converging neurons in RVM receiving dual synaptic inputs from the bladder and colon. In addition, neonatal cystitis significantly downregulated GABA transporter VGAT with a concomitant increase in TpH2 expression in bladder-linked RVM neurons suggesting an alteration in supraspinal signaling. These alterations of synaptic connectivity and GABAergic/serotonergic expressions in RVM neurons may contribute to bladder pain modulation and cross-organ visceral sensitivity.

Graphical Abstract

In Brief

Talluri et al identified and characterized RVM neurons that are synaptically linked to the bladder and the colon and studied the effect of early life bladder inflammation on molecular expressions of theses neurons. They demonstrated alterations of synaptic connectivity and GABAergic/serotonergic expressions in bladder-related RVM neurons in rats with neonatal cystitis.

1. Introduction

The intrinsic mechanism of altered nociceptive processing following early-life bladder inflammation is not fully understood. The postnatal day 1 through 21 is a critical developmental period for descending pain modulatory pathways in rodents (Hathway, Koch, Low, & Fitzgerald, 2009). Studies have documented the lack of efficacy of opioidergic pathway and long-term visceral hyperalgesia in rodents with early-life bladder inflammation (DeBerry, Ness, Robbins, & Randich, 2007; DeBerry, Randich, Shaffer, Robbins, & Ness, 2010; A. Randich, Uzzell, DeBerry, & Ness, 2006; Skagerberg, Bjorklund, & Lindvall, 1985). We have established the role of spinal GABAergic suppression in the development of visceral hypersensitivity and overlapping chronic pelvic pain (CPP) following neonatal cystitis in rats (Kannampalli et al., 2017; J. N. Sengupta et al., 2013; J. Zhang et al., 2017). The descending pathway from RVM to the spinal cord plays a major role in the modulation of pain processing and altered functions of RVM neurons contribute critically to the development of hyperalgesia and chronic pain (Chen & Heinricher, 2019; Heinricher, 2016). Transsynaptic pseudorabies virus (PRV) tracings from the bladder base and bladder body identify a multi-synaptic circuit of central neurons including brainstem nuclei that influence urinary bladder functions (Marson, 1997; Nadelhaft & Vera, 1995; Vizzard, Erickson, Card, Roppolo, & de Groat, 1995). Studies have also indicated that the physiological and behavioral responses to painful stimuli are manifested either by facilitating or inhibiting centrally projecting spinal neurons through intensity-dependent activation of separate subsets of neurons in the RVM (Carlson, Maire, Martenson, & Heinricher, 2007; Fields, Bry, Hentall, & Zorman, 1983; Fields, Vanegas, Hentall, & Zorman, 1983; H. S. Li, Monhemius, Simpson, & Roberts, 1998; Mason, 2001; Neubert, Kincaid, & Heinricher, 2004; M. Zhuo & Gebhart, 1990; M Zhuo, Sengupta, & Gebhart, 2002). Although the net balance of RVM descending modulations together with primary afferent input ultimately determine the excitability of spinal neurons, recent reports document that the ascending sensory inputs after peripheral noxious stimulation or injury are not observed until the third postnatal week in rats (Sadler & Kolber, 2016). The pain modulating neurons in the RVM undergo structural, molecular, and pharmacological plasticity in persistent pain models (Heinricher, 2016). Therefore, in-depth knowledge of how early-life bladder inflammation influences the development of supraspinal pain modulatory circuits could reveal the underlying mechanisms of neuroplasticity including neuromolecular and functional changes that cause painful bladder syndrome (PBS) and/or chronic pelvic pain (CPP).

Studies have documented that CPP is not restricted to one pelvic organ, but often overlaps with other pelvic organs through a common neural pathway (Brumovsky & Gebhart, 2010; Pezzone, Liang, & Fraser, 2005; Ustinova, Fraser, & Pezzone, 2006; Ustinova, Gutkin, & Pezzone, 2007; Winnard, Dmitrieva, & Berkley, 2006). We have previously documented that following neonatal (postnatal day14–16) cystitis, rats exhibit long-lasting colonic hyperalgesia to colon distension when tested in adulthood (Kannampalli et al., 2017; Miranda et al., 2011). However, it is unknown whether synaptic plasticity of distinct neuronal populations and/or the convergent neurons within RVM contribute towards the underlying mechanism of the cross-organ sensitization and visceral hypersensitivity as observed clinically in patients.

The present study tested the hypothesis that the alteration of synaptic inputs and neuromolecular expression of RVM neurons that are synaptically linked to the bladder may contribute to the development of neonatal cystitis-induced CPP. Our objectives were to identify and characterize bladder-linked RVM neurons and convergent neurons from the bladder and colon using PRV tracings. Furthermore, we intended to examine the long-term effect of early-life bladder inflammation on GABAergic and serotonergic (5HTergic) RVM neurons in a rat model of neonatal cystitis. This neuromolecular characterization of RVM neurons following neonatal cystitis may delineate the dysregulation of RVM descending influence on pain transmission during the development of pelvic pain.

2. Materials and methods

2.1. Animals

Since clinically PBS is prevalent in females, all experiments were undertaken in female Sprague Dawley rats (Leppilahti et al., 2005). Adult female rats weighing 200 to 250g (Charles River Laboratory, Wilmington, MA) and newborn female pups of postnatal day 14 (P14) were used in this study. Rats were housed under controlled conditions with a 12-hour light-dark cycle and had free access to food and water. Based on our previous findings that the long-term visceral pain following neonatal cystitis is independent of phases of the estrous cycle of the rats, we have included female rats in any phase of the estrous cycle for this study (Miranda et al., 2011). All experiments were performed according to approved guidelines by the Institutional Animal Care and Use Committee (IACUC) at the Medical College of Wisconsin (AUA 0355), and the International Association for the Study of Pain (IASP). We also followed the humane pain management recommendation of the Office of Laboratory Animal Welfare (OLAW).

2.2. Treatment Protocol

The schematic diagram of PRV-based approaches to trace projections from the bladder and colon to RVM is shown in fig. 1a. In the first treatment protocol as shown in fig.1b, adult naïve rats were injected with PRV-152 and PRV-614 into the bladder and distal colon, respectively, on postnatal day 60 (P60). In the second protocol as shown in fig. 1c, the experimental group female pups from P14 to P16 were anesthetized with isoflurane (1.5% induction and 1.0% maintenance with flow rate 1L/min) and received transurethral injection of protamine sulfate (1% in saline, 0.3ml) and left inside the bladder for 20 minutes followed by zymosan (1% in saline, 0.3 ml) for another 20 minutes. The control group pups received two doses of saline (0.3 ml) at similar intervals as described for the experimental group. Neonatal experimental and saline control groups received PRV-152 and PRV-614 on P60 as described for adult naïve rats. Brainstem tissues were collected 96 hours post-inoculation on P64 from naïve and neonatal saline and protamine+zymosan-treated rats to evaluate the PRV transport to RVM neurons.

Figure 1:

Schematic representation of anatomical pathways of PRV transport from bladder and colon to the medulla and the treatment protocols. Illustration of possible neural connecting pathways of the primary sensory afferent fibers from the bladder and colon to the RVM (a). Treatment protocol for naïve rats with no intravesical treatment received PRV injections into the bladder and colon on postnatal day P60 (b). Treatment protocol for female rat pups that received either intravesical protamine sulfate + zymosan or saline from P14 through P16 followed by PRV injections on P60 as described for naïve group (c). Tissues were collected from the naïve and treated groups 96 hours post viral inoculation on P64 for further analysis.

2.3. Fluorescence tagged PRV virus injections into the bladder and the colon

Rats were anesthetized with isoflurane (2% induction and 1.5% maintenance with flow rate 1 L/min) and a laparotomy was performed to expose the urinary bladder and the descending colon (Rouzade-Dominguez, Miselis, & Valentino, 2003). The pseudorabies viruses PRV-152 (PRV-GFP, 9.16X10⁸ pfu/ml) and PRV-614 (PRV-RFP, 1.52x10⁹ pfu/ml) were injected at 4–5 sites (0.5 µl per site) into the muscular layer of the bladder and colon, respectively (Rouzade-Dominguez et al., 2003; Valentino, Kosboth, Colflesh, & Miselis, 2000). The PRV viruses were a generous gift from Dr. Lynn Enquist, Department of Molecular Biology, Princeton University. The tip of the injection needle was held at the site of injection for a minute to allow the solution to get distributed into the muscle layers. The abdominal muscle was closed in layers by suturing with 3–0 absorbable sutures. The skin was closed using 3–0 silk sutures. Rats were injected with antibiotic (Enrofloxacin 20 mg/Kg i.m.) and analgesic (Rimadyl 5mg/Kg i.m.) postoperatively for 3 days. After 96 hours following viral inoculation, rats were deeply anesthetized with sodium pentobarbital (50 mg/Kg, i.p.) and the chest was opened by mid-sternal incision and perfused transcardially with ice-cold 0.1M Sorensen’s phosphate buffer followed by PFA-lysine-periodate fixative (McLean & Nakane, 1974).

2.4. Application of Fluorogold (FG) on the L6-S1 spinal cord in naïve rats followed by PRV virus injection into the bladder

In a separate set of experiments, to identify descending RVM neurons linked to the bladder afferents and secondary spinal neurons, fluorogold (FG) was applied on the dorsal surface of the lumbosacral (L6-S1) segments of the spinal cord in naïve rats followed by PRV-152 injections into the bladder three days after FG application. Briefly, rats were anesthetized with sodium pentobarbital (50 mg/Kg, i.p.) and the skin was opened at the lumbosacral level of the spinal cord to perform a laminectomy as described before (Wang & Wessendorf, 1999). The dura membrane was opened and a gel foam (~8 mm³) containing 2µl of 2.5% solution of FG in saline (Fluoro-Gold, Fluorochrome LLC., Denver, CO) was placed on the dorsal surface of the spinal cord. The wound was closed in layers (muscles followed by skin) with 3–0 sutures and rats were returned to the cage. Seventy-two hours following spinal FG application and PRV-152 injection, transcardiac perfusion was carried out as described above in section 2.3. The brainstem and lumbosacral (L6-S1) spinal cord (SC) tissues were collected 96 hours post-viral inoculation and evaluated for FG, PRV, and FG+PRV positive neurons in the RVM (supplementary data).

2.5. Antibody Characterization

The target amino acid sequences and Genbank Accession numbers for the GFP, RFP, TpH2, and VGAT antibodies used in this study were described in Table 1. The specificity of the antibodies was evaluated by IHC. To examine the cross-reactivity between GFP and RFP antibodies, RVM tissue sections from rats that only received PRV-152 (GFP+) in the bladder were immunostained with a mixture of chicken anti-GFP and rabbit anti-RFP as primary antibodies (Table 1). The specificity of the immunostaining was evident from a bright green neuronal staining for GFP without any red fluorescence staining for RFP. This indicates the specificity of anti-GFP and anti-RFP antibody staining without noticeable cross-reactivity. The specificity of the TpH2 antibody was analyzed by double immunostaining of RVM tissues with anti-TpH2 and anti-5HT (serotonin marker) antibodies (Table 1). The colocalization of TpH2 and 5HT expression in RVM neurons indicates the specificity of the TpH2 antibody for targeting serotonergic neurons in this study. The specificity of the VGAT antibody was examined by preabsorbing the antibody with a specific blocking peptide against which the VGAT antibody was raised (1 µg peptide per 1 µg antibody) and incubated for 60 min at room temperature. The peptide preabsorbed antibody failed to show any neuronal staining as compared to the unabsorbed antibody (supplementary data, fig. S4).

Table 1:

List of primary and secondary antibodies used in this study

Primary Antibody	Vendor/Source	Antibody information	Dilution
Chicken pAb to GFP	Abcam, Cambridge, UK	Catalog# ab13970 Accession: P42212.1 (1-238aa)	1:1000
Rabbit pAb to RFP	Abcam, Cambridge, UK	Catalog# ab62341 Accession: Q9U6Y8.1 (1-225aa)	1:1000
Mouse mAb anti-TpH2	Millipore, Billerica, MA	Catalog# AMAB91108 Accession: EDM16695.1 (C-terminal 269-320aa)	1:1000
Rabbit pAb VGAT	Alomone, Jerusalem, Israel	Catalog# AGT-005 Accession: O35458 (N terminal 106-120aa)	1:1000
Rabbit pAb 5HT (Serotonin)	ImmunoStar, Hudson, WI	Catalog# 20080 RRID: AB_572263	1: 2000
Secondary Antibody	Vendor/Source	Antibody information	Dilution
Alexa Fluor 488 goat anti-chicken	Invitrogen, Carlsbad, CA	Catalog# A11039 RRID: AB_2534096	1:1000
Alexa Fluor 568 goat anti-rabbit	Invitrogen, Carlsbad, CA	Catalog# A11036 RRID: AB_10563566	1:1000
Alexa Fluor 568 goat anti-mouse	Invitrogen, Carlsbad, CA	Catalog# A11019 RRID: AB_143162	1:1500

2.6. Stereological quantification

PRV labeled GFP+ (i.e., from the bladder) and RFP+ (i.e., from the colon) neurons were quantified to determine their anatomical distribution in the RVM and the percentage of converging neurons receiving inputs from the bladder and colon. The rostro-caudal axis of RVM ranged from the rostral edge of inferior olives to the caudal margins of the trapezoid body from bregma −10.8 mm to −12.8 mm according to rat atlas of Paxinos (Paxinos & Watson, 2006). For counting neurons with PRV expression in RVM subnuclei, the regions within the tissue section were selected according to rat atlas (Paxinos & Watson, 2006) to include raphe magnus (RMg) along with nuclei gigantocellularis (Gi), gigantocellularis pars alpha (GiA), and lateral paragigantocellularis (LpGi) as previously described (Hurley, Banfor, & Hammond, 2003; Millan, 2002; Parra, Nguyen, Hurley, & Hammond, 2002). Since no difference was observed in PRV tracing patterns between the left and right sides of the RVM, the labeled neurons on both sides were summed up for quantification. The immunostaining and stereological counting were carried out for sections that were chosen rostro-caudally at an interval of five serial sections. All counting/image analysis was carried out in a blinded manner for statistical analysis.

2.7. Immunohistochemical (IHC) identification of RVM neurons receiving synaptic inputs from the bladder and colon in naïve rats

The brainstem and the spinal cord tissues were post-fixed in PFA-lysine-periodate fixative overnight and the following day they were incubated in 10% sucrose overnight at 4°C. The next day, tissues were transferred to 20% sucrose for 8 hours, and then to 30% sucrose overnight on a shaker at 4°C. The pontomedullary region of the brainstem was sectioned from caudal to rostral at a thickness of 30 µm using a cryostat (Microm cryostat, Thermo Fisher Scientific, Waltham, MA) and stored in cryoprotectant solution at −20°C. For immunostaining of rostro-caudal distribution of PRV+ neurons, floating sections (n=3, 3 sections each for rostral, mid, and caudal levels/rat) were washed three times in PBS wash buffer (1X phosphate-buffered saline (PBS), 0.1% Triton X-100, 0.01% Sodium azide, 0.01% NGS) and then incubated in blocking buffer containing 10% NGS (Product ID: 005–000-121, Jackson Immuno Research, West Grove, PA) in wash buffer for 2 hours at room temperature on an orbital shaker. Subsequently, tissue sections were incubated in primary antibodies against GFP and RFP (Table 1) in 5% NGS overnight at 4°C with gentle shaking. The following day, tissue sections were washed four times and incubated with secondary antibodies Alexa Fluor 488 goat anti-chicken and Alexa Fluor 568 goat anti-rabbit antibodies (Table 1) for 2 hours at room temperature on a shaker. Finally, the tissues were washed 3 times and mounted on glass slides with an anti-fade mounting medium (Product ID: H-1000, Vectasheild, Burlingame, CA). The images of RVM serial sections were captured rostro-caudally under 2X, 4X, and 10X objectives (Nikon eclipse 50i microscope, Nikon, Tokyo, Japan) using red and green fluorescence filters. Transsynaptically labeled PRV infected neurons from the bladder and colon were identified by capturing images under individual fluorescence channels. Similarly, RVM and SC (L6-S1) sections were examined for PRV-152 and FG labeling under green and UV filters, respectively. Colocalized neurons were identified upon merging images taken with independent filters. All cells/neurons exhibiting green, red, and/or merged fluorescence were counted using the ImageJ cell counter tool (NIH, Bethesda, MD). The percentage of convergent neurons from the bladder and colon was identified by counting neurons that were dually labeled with GFP and RFP.

2.8. In situ hybridization (ISH) assay for characterizing GABAergic and enkephalinergic markers on GFP+ RVM neurons in naïve rats

The RNAscope® Multiplex Fluorescent v2 assay was used to study the expression of various genes simultaneously at a single molecule level as per the manufacturer’s protocol (Advanced Cell Diagnostics, Newark, CA). The target sequences and GenBank accession numbers for the probes were described in Table 2. Briefly, RVM tissue sections (n=3, 3 sections/rat) were mounted on SuperFrost Plus slides (Fisher Scientific, Hampton, NH) and baked for 30 mins at 60°C. Tissues were then fixed with 4% PFA for 90 mins at room temperature and dehydrated sequentially with 50%, 70%, and 100% ethyl alcohol for 5 minutes each at room temperature. The tissue sections were pre-treated with hydrogen peroxidase for 10 mins at room temperature and washed with distilled water. The target retrieval was performed at 99°C for 10 mins followed by a brief wash and drying the slides in 100% ethyl alcohol. Tissues were then treated with protease III for 15 mins at 40°C in the HybEZ™ Oven. After three washes, specific probe sets/mixture was added to each tissue section (Table 2) and incubated for 2 hours at 40°C. The slides were then washed with wash buffer and the RNAscope Multiplex Fluorescent Detection Reagents v2 kit was used to amplify the signal sequentially using AMP1, AMP2, and AMP3 amplifiers and further developed with specific TSA plus fluorophores. Finally, the tissues were mounted with mounting medium containing DAPI (Product ID: H-1200, Vectasheild) and coverslipped. The RVM tissue sections were imaged on a confocal microscope (Nikon A1-R, Japan). The images were captured using blue/DAPI, green, red, and far-red fluorescence channels under similar settings of exposure, gain, and gamma adjustment at 60X magnification. The PRV+ GABAergic and/or enkephalinergic double labeled neurons were counted using the ImageJ cell counter tool (NIH, Bethesda, MD).

Table 2-.

Probes used for RNAscope ISH protocol

RNAscope® Probe	Probes (ACDBio)	Accession No:	TSA fluorophores (Perkin Elmer) (1:500-1:1000)
GFP-C1	409011	AF2759553.1 (probe: 12-686bp)	TSA plus fluorescein NEL741E001KT
GFP-C3	409011-C3	AF2759553.1 (probe: 12-686bp)
RFP-O2-C2	562541-C2	AF168419.2 9 (probe: 2-856bp)	TSA plus Cyanine 3 NEL744E001KT
Rn-Gad1-C2	316401-C2	NM_017007.1 (probe: 950-1872bp
Rn-Penk-C3	417431-C3	NM_017139.1 (probe: 242-1369bp)	TSA plus Cyanine 5 NEL745E001KT
Rn-Slc32a1-C1 (VGAT)	424541	NM_031782.1 (probe: 388-1666bp
Rn-Slc32a1-C3 (VGAT)	424541-C3	NM_031782.1 (probe: 388-1666bp

2.9. IHC analysis of RVM neurons receiving synaptic inputs from the bladder and colon following neonatal cystitis

RVM tissues taken from neonatal saline and zymosan-treated rats were post-fixed and processed for cryosectioning as described in section 2.7. Tissue sections (n=3/group, 3 sections/rat) were immunostained with GFP and RFP antibodies as described in section 2.7. The images of RVM sections from experimental and control groups were captured under 2X, 4X, and 10X objectives (Nikon eclipse 50i microscope, Nikon, Tokyo, Japan) using red and green fluorescence filters. Colocalized neurons were identified upon merging images taken with independent filters. All cells/neurons exhibiting green, red, and/or merged fluorescence were counted using the ImageJ cell counter tool (NIH, Bethesda, MD). The percentage of convergent neurons from the bladder and colon was identified by counting neurons that were dually labeled with GFP and RFP.

2.10. IHC analysis of TpH2 expression in RVM neurons receiving synaptic inputs from the bladder following neonatal cystitis

To visualize TpH2 expression, RVM tissue sections (n=3/group, 9 sections/rat) were treated with target retrieval solution (Product ID: S1699, Dako, Santa Clara, CA) for 30 mins at 70°C followed by incubation with chicken anti-GFP and mouse anti-TpH2 antibodies (Table 1). The sections were rinsed and treated with secondary antibodies Alexa Fluor 488 goat anti-chicken and Alexa Fluor 568 goat anti-mouse antibodies (Table 1). The TpH2 expression pattern and the colocalization of GFP and TpH2 expression in RVM neurons of control and experimental groups were visualized with a fluorescence microscope (Nikon Eclipse 50i, Japan) at 2X, 4X, and 10X objectives under similar settings of exposure, gain, and gamma adjustment. For each section, GFP+, TpH2+, or GFP+/TpH2 double labeled were counted using the ImageJ cell counter tool.

2.11. In situ hybridization (ISH) assay for characterizing VGAT mRNA expression in GFP+ RVM neurons following neonatal cystitis

The RNAscope® Multiplex Fluorescent v2 assay was used to study the expression of VGAT mRNA copies as described in section 2.8. The RVM tissue sections from neonatal saline and zymosan-treated rats (n=3/group, 3 sections/rat) were incubated with GFP/VGAT probe mixture (Table 2). The images were captured on a confocal microscope (Nikon A1-R, Japan) using blue/DAPI, green and far-red channels under similar settings of exposure, gain, and gamma adjustment at 60X magnification. To analyze VGAT mRNA copies in individual RVM neurons, we selected 18 large diameter and 27 medium diameter neurons per group based on their ferret’s diameter. The ferret’s diameter was measured using the ImageJ analysis tool by averaging the largest and smallest length across the neuron and converting pixels to µm. The VGAT mRNA copies within each neuron were expressed as individual fluorescence-labeled dots. The images were converted to grayscale and the number of dots was counted for each neuron after adjusting the threshold using the “analyze particles” option of the ImageJ program.

2.12. IHC analysis of VGAT expression in GFP+ RVM neurons following neonatal cystitis

To visualize VGAT expression in PRV labeled neurons, RVM sections (n=3/group, 3 sections/rat) were treated with target retrieval solution (Dako) followed by incubation with chicken anti-GFP and rabbit anti-VGAT antibodies (Table 1). This was followed by incubation with Alexa Fluor 488 goat anti-chicken and Alexa Fluor 568 goat anti-rabbit antibodies (Table 1). After washing, the tissues were mounted on a glass slide along with a mounting medium containing DAPI (Vectasheild). The VGAT immunostaining of GFP+ RVM neurons in control and experimental groups was captured on a confocal microscope (Nikon A1-R, Japan) using blue/DAPI, green and red channels under similar settings of exposure, gain, and gamma adjustment at 60X magnification. The intensity of VGAT expression in GFP+ neurons (45 neurons/group) for both the experimental and control groups was evaluated using the ImageJ program. The fluorescent images were converted to grayscale and a fixed region of interest (ROI) was selected around individual neurons to measure the intensity of VGAT staining. The actual staining intensity for individual neurons was quantified by subtracting the background staining and represented as the mean integrated density of staining.

2.13. Data Analysis

Images were edited using Adobe Photoshop and CorelDraw X8. Statistical analysis was performed using GraphPad Prism (version 9.2). All the fluorescence data were analyzed in a blinded manner and to accommodate color-blind readers we used magenta pseudo-color in all illustrations except for fig. 4. Results are expressed as the mean ± S.E.M. and p<0.05 was considered statistically significant. One-way ANOVA with Bonferroni’s multiple comparison test was performed for analyzing rostro-caudal tracing of PRV+ RVM neurons and distribution of dual-labeled PRV+ neurons in different RVM nuclei. An unpaired t-test with Welch’s correction was used to compare the data between neonatal cystitis and control groups.

Figure 4:

Molecular characterizations of PRV- labeled RVM neurons in adult naïve rats. Representative images of 2-color immunohistochemistry (GFP-green and TpH2-magenta) analyses of bladder specific RVM neurons are shown in (a1-a3). White arrows represent GFP+ neurons expressing TpH2, and the scale bar is 100µm. Representative images of 3-color in situ hybridization (GFP-green, Gad1-red, and PENK-white) analyses of bladder specific RVM neurons are shown in (b1-b4). The white arrows represent GFP+ neurons having both GAD1 and PENK expression while the yellow arrow represents GFP+ cell with either GAD1 or PENK expression. Similarly, representative images of 3-color in situ hybridization (GFP-green, RFP-red, and VGAT-white) analyses are shown in (c1-c4). The white and yellows arrows represent converging neurons with and without VGAT expression respectively. Scale bar is 25µm.

3. Results

3.1. Transsynaptic tracing of RVM neurons following PRV injections into the bladder and colon in adult naïve rats

In non-treated adult naïve rats, the rostro-caudal distribution of PRV+ RVM neurons synaptically connected to the bladder and colon are shown in fig. 2 (stereotaxic representation from Paxinos rat atlas (fig. 2: a1–c1) (Paxinos & Watson, 2006), GFP+/bladder (fig. 2: a2–c2) and RFP+/colon (fig.2: a3–c3). One-way ANOVA analysis reveals that the bladder-linked PRV-152 GFP+ RVM neurons are significantly higher at the mid-level of the RVM compared to the caudal or rostral levels (F=6.66, df=26, p<0.05, fig. 2d). A similar pattern of distribution is observed for colon-linked PRV-614 RFP+ RVM neurons (F=5.586, df=26, p<0.05, fig. 2d). Overall, PRV-infected neurons in the RVM demonstrate bilateral distribution with more labeling in the mid-level compared to caudal and rostral levels of the RVM. Since no differences in the distribution pattern of PRV labeled neurons were observed between the left and right sides of the RVM, the labeled neurons from both sides were totaled up for quantification. Based on this finding, mid-level sections were used for all further analyses.

Figure 2:

Rostro-caudal distribution of bladder- and colon-linked RVM neurons. Stereotaxic representations of RVM sections (bregma −10.8 mm to 12.8 mm) taken at rostral, middle and caudal levels for analyzing PRV expression (a1,b1,c1). Representative images from the 2-color IHC show the distribution patterns of GFP+ bladder and RFP+ colon neurons in the rostral (a2-a3), middle (b2-b3) and caudal (c2- c3) RVM levels. The scale bar is 250 µm. Quantification of PRV-labeled neurons in RVM tissue sections in different RVM levels is shown in (d). Data is represented as mean ± SEM, n=3 (3 sections/level/rat). For GFP+ bladder neurons *p<0.05 middle vs rostral and caudal, for RFP+ colon neurons #p<0.05 middle vs rostral and caudal. Comparison among levels were analyzed using one-way ANOVA followed by post-hoc Bonferroni test.

3.2. Identification of RVM neurons involved in ascending and descending spinal projections in adult naïve rats

In agreement with previous results (Wang & Wessendorf, 1999), the FG application on the dorsal surface of the lumbosacral (L6-S1) spinal cord was found mostly in the dorsal portion of the dorsolateral funiculus, with the highest intensity of labeling near the region of substantia gelatinosa (supplementary data, fig. S2: a1–a2). The PRV-GFP staining is observed primarily around the central canal, the dorsal commissure, and along the lateral edge of the dorsal horn extending from Lissauer’s tract to the region of the sacral parasympathetic nucleus (SPN). Interestingly, FG labeling and PRV staining are in distinct regions in the spinal cord without overlapping (fig. S2: b1–b3), whereas PRV+ neurons are identified within the RVM nuclei receiving FG tracing from the spinal cord and about 59%±2.3% of GFP positive neurons exhibiting FG labeling (fig. S3: a–c). This finding indicates that the descending RVM neurons are synaptically linked to bladder afferents via the spinal-RVM axis and may play a critical role in the overall spinal sensory modulation of CPP.

3.3. Distribution pattern of converging RVM neurons in different RVM nuclei with dual projections from bladder and colon in adult naïve rats

The RVM nuclei were identified around the stereotaxic position of bregma −11.88 mm as referred to in Paxinos rat brain atlas in fig. 3a (Paxinos & Watson, 2006). The representative images of the distribution pattern of converging bladder connected GFP+ and colon connected RFP+ neurons in different RVM subnuclei are shown in fig. 3 (b–d). One-way ANOVA analysis reveals that the percentage of bladder-specific RVM neurons receiving dual projections from the bladder and colon is higher in the Gi/GiA nuclei compared to RMg and LpGi (F=5.6, df=26.p<0.05). Similarly, the percentage of colon-specific RVM neurons receiving dual projections from the colon and bladder is higher in the Gi/GiA nuclei compared to RMg and LpGi (F=11.52, df=26, p<0.05, fig. 3e). The current findings are the more detailed report of our preliminary study which demonstrates that the transsynaptic RVM labeling from the bladder and colon were present in regions implicated in descending pain control, such as RMg, Gi, GiA, and LpGi (Hoelzel et al., 2019).

Figure 3:

Immunohistochemical analysis of converging neurons from bladder and colon in different RVM nuclei of adult naïve rats. Stereotaxic representation of mid-level section (bregma −11.88 mm) showing RVM nuclei (a1, magenta dots) selected for analysis of PRV expression (Paxinos & Watson, 2006). Lower magnification images of bladder specific (GFP+) and colon specific (RFP+) neurons within the identified nuclei (a2-a3). The scale bar is 200µm. Higher magnification images showing distribution patterns of single labeled and double labeled neurons in RMg (b1-b3), Gi/GiA (c1-c3) and LpGi (d1-d3). The scale bar is 100µm. The white arrows represent convergent neurons in different RVM nuclei. Quantification of percentage of bladder and colon specific convergent neurons in different RVM nuclei (e). Data is represented as mean ± SEM, n=3 (3 sections/rat), For % bladder convergent neurons, *p<0.05 Gi/GiA vs RMg and LpGi, for % colon convergent neurons, #p<0.05 Gi/GiA vs RMg and LpGi. Comparison of PRV distribution among RVM nuclei were analyzed using one-way ANOVA followed by post-hoc Bonferroni test.

3.4. RVM neurons synaptically connected to the bladder express serotonergic, GABAergic, and enkephalinergic markers in adult naïve rats

In the next set of experiments, we characterized the constitutive expression of TpH2, GAD1, VGAT, and PENK in PRV-positive neurons as shown in fig. 4 and Table 3. IHC staining reveals that about 14% of PRV labeled (GFP+) neurons are TpH2 positive (fig. 4a). RNAscope ISH analysis of GABA synthesizing enzyme glutamic acid decarboxylase 1 (GAD1), GABA vesicular transporter (VGAT), and proenkephalin (PENK) marker expressions demonstrated that about 61% GFP labeled RVM neurons are GAD1+ and 47% are VGAT+ suggesting that most inputs received by bladder afferents are possibly inhibitory neurons (fig. 4: b–c). Approximately 41% of GFP-labeled RVM neurons exhibit PENK expression (fig. 4b). Interestingly, 40% of GFP+/GAD1 labeled RVM neurons also express PENK, emphasizing that the dual GABAergic and enkephalinergic RVM inputs may play a dominant role in bladder pain modulation. Additionally, we also evaluated VGAT expression in converging RVM neurons (GFP+ and RFP+) using RNAscope ISH probes (fig. 4c). About 50% of these converging neurons are VGAT+ indicating a possible involvement of RVM GABAergic neurons in cross-organ sensitization/overlapping pain following neonatal cystitis (Table 3).

Table 3:

Molecular characterization of PRV labeled RVM neurons

Type of analysis	Neuronal Marker (GFP/Bladder; RFP/Colon)	% of labelled neurons in RVM
IHC	GFP+/TpH2	14.3 (108/751)
ISH	GFP+/GAD1	60.9 (39/64)
	GFP+/PENK	40.6 (26/64)
	GFP+/VGAT	47.4 (28/59)
	GFP+/GAD1/PENK	40.6 (26/64)
	GFP+/RFP+/VGAT	50 (6/12)

3.5. Increase in bladder-specific RVM neurons in neonatal (P14–16) cystitis rats and the effect of neonatal cystitis on the converging RVM neurons from bladder and colon

The effect of neonatal cystitis on bladder- (GFP+) and colon- (RFP+) linked RVM neurons was examined in the experimental group and compared to saline-treated group (fig. 5: a–b). The average numbers of GFP+ and RFP+ neurons are significantly higher in the neonatal zymosan-treated group as compared to the saline-treated group (GFP+ RVM neurons: 112.8±8.5 vs 79±5.9, t=−3.24, df=14.3, p<0.05 and RFP+ RVM neurons: 92.6±8.8 vs 67.1±6.9, t=−2.26, df=15.14, p<0.05 vs saline controls, fig. 5c). Additionally, the dual-labeled cells from the bladder (GFP+) and the colon (RFP+) were analyzed for zymosan and saline-treated groups. Statistical analysis reveals that the percentage of dual-labeled (GFP+/RFP+) neurons from bladder and colon in the RVM are significantly higher in neonatal zymosan-treated group compared to saline controls (%bladder colocalization: 27.9±0.9 vs 23.1±0.7, t=−3.97, df=15.26, p<0.05 and %colon colocalization: 32.2±1.1 vs 28.2±1.7, t=−2.19, df=14.32, p<0.05 vs controls, fig. 5d).

Figure 5:

Immunohistochemical analysis of PRV positive RVM neurons in neonatal zymosan and saline-treated rats. Representative images of PRV expression in RVM neurons for zymosan- (a1-a3) and saline- (b1-b3) treated groups. Scale bar is 250µm. Quantitative analysis of bladder and colon specific RVM neurons in experimental and control groups, *p<0.05 vs saline (c). Quantitative analysis of the effect of neonatal treatment on converging neurons in zymosan and saline groups, *p<0.05 vs saline controls (d). Data represented as mean ± SEM, n=3/group (3 sections/rat). Unpaired t-test with Welch’s correction was used to compare the data between the groups.

3.6. Effect of neonatal cystitis on TpH2 expression in bladder innervating RVM neurons

We examined the effect of neonatal cystitis on TpH2 expression in RVM neurons and also in bladder-specific GFP+ RVM neurons using IHC (fig. 6: a–b). TpH2 positive neurons are significantly higher in the cystitis group compared to the saline group (101.8±4.5 vs 44.8±1.9, t=−11.48, df=35.1, p<0.005 vs controls, fig. 6c). Similarly, the percentage of GFP+ neurons expressing TpH2 is significantly higher in the cystitis group compared to controls (22.7±1.8 vs 10.9±0.7, t=−6.06, df=22, p<0.005 vs controls, fig. 6d).

Figure 6:

Immunohistochemical analysis of TpH2 expression in GFP+ RVM neurons in neonatal zymosan and saline-treated rats. Representative images show GFP-positive and TpH2-positive neurons from zymosan treated (a1-a3) and saline control (b1-b3) groups. White arrows indicate bladder specific RVM neurons co-expressing TpH2. The scale bar is 100 µm. Quantitative analysis of TpH2 positive RVM neurons in experimental and control groups, **p<0.005 vs saline controls (C). Data represented as mean ± SEM, n=3/group, 9 sections/rat. Quantitative analysis of percentage GFP+ RVM neurons expressing TpH2 in experimental and control groups, **p<0.005 vs saline controls (d). Data represented as mean ± SEM, n=3/group, 6 sections/rat. Unpaired t-test with Welch’s correction was used to compare the data between the groups.

3.7. Downregulation of GABA Transporter (VGAT) both at mRNA and protein levels in bladder innervating RVM neurons following neonatal cystitis.

To further evaluate the effect of early-life bladder insults on the RVM GABAergic pathway, we analyzed VGAT mRNA expression in RVM neurons that are synaptically linked to the bladder in neonatal zymosan and control groups using ISH. The expression of mRNA copies for GFP and VGAT within RVM neurons were monitored as individual dots within individual RVM neurons for the experimental (fig. 7: a1–a3) and saline (fig. 7: b1–b3) groups. In the control group, there were clusters of mRNA expression pooled together in each cell, whereas, in the neonatal zymosan group, the mRNA copies were well separated as distinct dots and fewer in number. The quantification of VGAT mRNA expression within selected GFP+ RVM neurons for both the groups is shown in fig. 7c. The RVM neurons were grouped as medium and large neurons with average diameters of 23.2±4.2 µm and 40.8±7.7 µm, respectively. Neonatal zymosan group exhibited significant downregulation in VGAT mRNA expression compared to saline-treated controls in both medium diameter (26.5±0.8 vs 62±4.1, t=−8.4, df=28.34, p<0.005 vs controls, fig. 7c) and large diameter neurons (65.6±4.7 vs 95.3±5.2, t=−4.2, df=33.7, p<0.005 vs controls, fig. 7c). We further examined the effect of treatment on VGAT protein expression in RVM neurons that are GFP positive. Similar to VGAT mRNA expression, we observed higher VGAT protein expression in saline controls compared to neonatal zymosan-treated group (fig. 8: a–b). The analysis revealed that the integrated density of VGAT staining in the experimental group is significantly lower compared to saline-treated controls (integrated density: 1.8±0.1x10⁵ vs 4.5±0.3x10⁵, t=−8.56, df=42.24, p<0.005 vs controls, fig. 8c).

Figure 7:

In situ hybridization analysis of GABA transporter VGAT mRNA expression in GFP+ RVM neurons in neonatal zymosan and saline-treated rats. RNAscope analyses of GFP and VGAT mRNA expression in zymosan (a1-a3) and saline-treated (b1-b3) groups. Bold white arrows represent the large diameter and hollow white arrows represent medium diameter GFP+ neurons expressing VGAT. The scale bar is 25µm. Quantitative data analysis for VGAT mRNA copies in large and medium diameter neurons in experimental and control groups, **p<0.005 vs saline in large diameter neurons, and **p<0.005 vs saline in medium diameter neurons (c), Data represented as mean ± SEM, n=3/group (3 sections/rat). The RVM neurons were grouped as large, 18 neurons/group and medium 27 neurons/group with average diameters of 40.8±7.7µm and 23.2±4.2µm respectively. Unpaired t-test with Welch’s correction was used to compare the data between the groups.

Figure 8:

Immunohistochemical analysis of VGAT protein expression in GFP+ RVM neurons following neonatal cystitis. Representative images of GFP and VGAT protein expression in experimental (a1-a3) and saline-treated (b1-b3) groups. The bold white arrows represent GFP+ neurons expressing VGAT, and hollow white arrows represent GFP+ neurons without VGAT expression. The scale bar is 25µm. Quantitative data of VGAT protein expression measured as integrated density of VGAT staining within GFP+ neurons for experimental and control groups, **p<0.005 vs saline controls (c). Data represented as mean ± SEM, n=3/group, 3 sections/rat). Unpaired t-test with Welch’s correction was used to compare the data between the groups.

4. Discussion

In this study, we used the pseudorabies virus (PRV)-based tracing approach to identify RVM neurons that are synaptically connected to the bladder and colon of naïve and neonatally developed cystitis rats. This viral vector constitutes a unique class of retrograde tracers for investigating the neuronal circuit organization in the central nervous system (CNS). The limitation of passive dye tracers like Fast Blue, Fluorogold, 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate (DiI), and Cholera toxin conjugated peroxidases (Ct-HRP) is the migration of tracer only within the tracer-uptake neurons via the axonal transport. Therefore, it does not show the network of connectivity from the peripheral organ to CNS. The ability of PRV to infect, replicate and transsynaptically migrate across the connecting neurons can reveal the neural network, pathways, and projections in CNS. Our results have shown a bilateral distribution of single and dual virus-labeled cell somas with distinct differences in the expression pattern in different RVM nuclei. The molecular characterization of the identified neurons indicates that most of these neurons are GABAergic and enkephalinergic with a small percentage expressing serotonergic marker TpH2. Importantly, neonatal cystitis resulted in a significant increase in bladder- and colon-specific RVM neurons and converging neurons (GFP+/RFP+) suggesting increase synaptic connections from the bladder and colon. This finding suggests that during the neural developmental phase painful bladder inflammation can influence neuromolecular changes and synaptic connectivity of RVM neurons involved in altered bladder functions like overactive bladder and overlapping pelvic pain. In addition, a significant downregulation of GABA transporter VGAT expression in bladder-linked RVM neurons suggests the alteration in supraspinal GABA signaling in neonatally induced cystitis rats. In a recent study, we have documented a transcriptional dysregulation of spinal GABA receptors and transporters and an overall suppression of spinal GABAergic transmission in neonatal cystitis rats (J. N. Sengupta et al., 2013). Our behavioral and functional studies further established the role of GABA disinhibition in spinal sensitization and long-lasting overlapping pain in this model of neonatal cystitis (Kannampalli et al., 2017; J. N. Sengupta et al., 2013; J. Zhang et al., 2017). Therefore, the current results and our previous findings emphasize that the alteration of GABAergic neurotransmission along the RVM-spinal cord axis plays a critical role in the development of cross-organ sensitization and overlapping pelvic pain following neonatal cystitis. Earlier studies have documented that at the spinal and supraspinal levels, the distribution patterns of virus-labeled cells following the injection of PRV into the lower urinary tract overlap with the cells exhibiting increased c-fos gene expression due to noxious stimulation and injuries within the same region. Thus, it indicates a direct involvement of spinal and supraspinal axis in bladder pain processing mechanism (Birder & de Groat, 1992, 1998; Birder, Roppolo, Erickson, & de Groat, 1999; Alan Randich, Mebane, DeBerry, & Ness, 2008; Sadler & Kolber, 2016; Vizzard, 2000).

RVM is comprised of functionally and neurochemically distinct classes of neurons and the pain sensation and affective behavior can be either facilitatory or inhibitory by the activation of separate subsets of RVM neurons (Terayama, Dubner, & Ren, 2002; M. Zhuo & Gebhart, 1997; M Zhuo et al., 2002). The raphe magnus (RMg) and gigantocellularis (Gi) nuclei within the RVM project to the spinal cord and enhance or diminish nociceptive trafficking through the descending pain modulatory pathways (Millan, 2002; Wei, Dubner, & Ren, 1999). PRV tracings from the bladder and colon in our study are mainly confined within the RVM nuclei that are considered to be involved in descending pain modulation of pelvic viscera suggesting that the ascending transmission from the bladder and the descending modulatory pathways in the RVM may be functionally interconnected through a positive and negative feedback loop (Chen & Heinricher, 2019; D’Mello & Dickenson, 2008).

The adverse effects of noxious visceral stimuli during the neonatal period on the neural development of ultimate pain processing pathways have not been well defined. A clinical correlation and long-term follow up within interstitial cystitis (IC) patients demonstrate that 10–30% of patients reported having experienced childhood bladder related problems including infection (Jones & Nyberg, 1997). Recent studies have documented that in experimental models, early-life somatic and visceral intervention induces a long-lasting spinal hypersensitivity and disruption of GABAergic and opioidergic neurotransmission (DeBerry et al., 2007; Kannampalli et al., 2017; Miranda et al., 2011; Miranda, Peles, McLean, & Sengupta, 2006; Miranda, Peles, Shaker, Rudolph, & Sengupta, 2006; A. Randich et al., 2006). Several recent studies documented the role of dual GABAergic and enkephalinergic (opioidergic) RVM neurons in mechanical pain sensitivity (Francois et al., 2017; Y. Zhang et al., 2015). These studies further indicate that the majority of RVM-derived descending inputs were dual GABAergic and enkephalinergic, and, in adult rats, silencing or activation of these RVM neurons specifically altered the behavioral sensitivity to heat and mechanical stimuli. However, the involvement of this set of RVM neurons in the visceral pain pathway is poorly understood. Although we didn’t undertake a detailed characterization of the opioidergic pathway in the present study, we were interested in evaluating whether bladder-specific RVM neurons co-express both opioidergic and GABAergic markers. In the current study, we observed about 40% of GFP+ /GAD1 labeled RVM neurons also expressed PENK (proenkephalin), a marker for opioidergic neurons, emphasizing a possible involvement of these RVM neurons in bladder pain modulation. explanation for clinical observations ofBesides molecular studies, our lab is actively engaged in developing novel analgesic strategies for treating the pain associated with interstitial cystitis/painful bladder syndrome (IC/PBS) which can circumvent the opioid-induced severe CNS effects. In a recent communication, we have reported a bifunctional opioid-based compound with less adverse CNS effect that demonstrates potent antinociception by attenuating pelvic afferent inputs from the urinary bladder of zymosan-treated rats (Terashvili et al., 2021).

At the spinal level, neonatal tissue injury also induces a persistent disruption in the balance between primary afferent-evoked excitation and inhibition of adult spinal projection neurons by altering sensory input onto GABAergic interneurons (J. Li & Baccei, 2019). At the same time, C-fiber inputs to spinal dorsal horn neurons significantly strengthen during early postnatal days suggesting an acceleration of the ongoing C-fiber synapses development (Baccei, Bardoni, & Fitzgerald, 2003; Fitzgerald, 2005; J. Li & Baccei, 2019). However, the expansion of nociceptive primary afferent projections to the superficial dorsal horn, which occurs following neonatal tissue damage is not observed after injuries at later postnatal ages (Ruda, Ling, Hohmann, Peng, & Tachibana, 2000; Walker, Meredith-Middleton, Cooke-Yarborough, & Fitzgerald, 2003). On the other hand, a recent PRV tracing of bladder related RVM neurons establishes that the supraspinal neural circuitry which underlies descending pain modulation pathway in the adult animal is already organized in the neonatal rat during postnatal day P2 even though the supraspinal pain modulatory mechanism does not become functional until the third postnatal week (i.e., day 21), and independent of noxious sensory input to the RVM early-in-life (Géranton, Tochiki, Chiu, Stuart, & Hunt, 2010; Hathway et al., 2009; Schwaller, Kwok, & Fitzgerald, 2016; Silva, Costa-Pereira, Martins, & Tavares, 2016; Sugaya, Roppolo, Yoshimura, Card, & de Groat, 1997). In our current study, increased bladder-linked RVM neurons in neonatal cystitis rats indicate an alteration in neural circuitry following early-life bladder intervention during postnatal development. Moreover, colocalization of spinally applied FG and bladder injected PRV-GFP positive RVM neurons in this study also emphasizes the major role of a fast acting spino-bulbo-spinal loop which may be activated by ascending nociceptive inputs from the bladder and in turn drives descending feedback modulation of spinal nociception.

Recent clinical studies documented that patients with irritable bowel syndrome and colitis often experience sensory and motor dysfunction of the urinary bladder and conversely, patients diagnosed with interstitial cystitis experience irritable bowel syndrome (IBS) like symptoms (Alagiri, Chottiner, Ratner, Slade, & Hanno, 1997; Dellis, Kostakopoulos, & Papatsoris, 2019; Doiron, Kogan, Tolls, Irvine-Bird, & Nickel, 2017). Similarly, cross-organ sensitization and overlapping pelvic pain have also been reported in experimental models of cystitis and colitis (Bielefeldt, Lamb, & Gebhart, 2006; Brumovsky & Gebhart, 2010; Laird, Souslova, Wood, & Cervero, 2002; Malykhina, 2007; Malykhina et al., 2006; Pezzone et al., 2005; Qin, Malykhina, Akbarali, & Foreman, 2005; J. Sengupta, 2009; Yoshikawa et al., 2015). We have recently demonstrated that the intravesical zymosan administration in female rats during the neonatal period results in a long-term colonic hypersensitivity when tested in adulthood (Kannampalli et al., 2017; Miranda et al., 2011). Several retrograde tracing studies attempted to decipher the neuronal circuitry involved in the convergence of sensory information from discrete pelvic organs at different levels of the nervous system (Christianson et al., 2007; Nadelhaft & Vera, 1995; Rouzade-Dominguez et al., 2003; Vizzard et al., 1995). At the spinal level, the transsynaptic viral tracing from bladder and colon demonstrated dual-labeled neurons mainly in the lumbosacral paraganglionic column and the dorsal commissure area (Rouzade-Dominguez et al., 2003). The converging RVM neurons innervating from both organs indicate the existence of a common neuronal pathway that encompasses supraspinal input along with the peripheral and spinal components. Therefore, the increase in converging RVM neurons in neonatal cystitis rats could provide an explanation for clinical observations of viscero-visceral (i.e., overlapping pain) and viscero-somatic (i.e., expanded dermatomal hyperalgesia) referred pain in cystitis patients, where these neurons may have been involved in neurogenic sensitization via antidromic pathways to produce functional changes of neighboring organs and corresponding somatic areas.

Several morphological and functional studies deciphering brainstem descending pain modulatory pathways mainly focused on RVM serotonergic and non-serotonergic inputs to spinal and primary afferent neurons (Dogrul, Ossipov, & Porreca, 2009; Kwiat & Basbaum, 1992; Lu & Perl, 2007; Potrebic, Fields, & Mason, 1994; Zhao et al., 2014). The recent imaging study indicates a descending RVM serotonergic projection to the TRPV1 positive central terminals, and blockage of serotonergic inputs attenuates spinal sensitization and excitatory afferents innervating the spinal dorsal horn regions (Kim et al., 2014). Several reports also suggest GABAergic projections from RVM to spinal dorsal horn as direct input to postsynaptic neurons such as spinal interneurons and ascending projection neurons (Aicher, Hermes, Whittier, & Hegarty, 2012; Hossaini, Goos, Kohli, & Holstege, 2012; Kato et al., 2006; Reichling & Basbaum, 1990). PRV tracing from the bladder demonstrated that bladder-specific serotonergic RVM neurons are mostly expressed in the caudal raphe with the highest number in the raphe magnus (Ahn, Saltos, Tom, & Hou, 2018). Similarly, our PRV tracing experiments have demonstrated that about 14% of RVM neurons connected to the bladder are TpH2 positive with major expression in RMg, Gi, and LpGi nuclei. Molecular approaches of selective inhibition or activation of RVM TpH2-expressing neurons demonstrated that serotonergic projections from the RVM are important for the facilitation of pain in inflammatory/neuropathic pain states and in long-lasting pain sensitization (Cai, Wang, Hou, & Pan, 2014; Wei et al., 2010). In experimental models of diabetic neuropathy and chemotherapy-induced neuropathy, alteration in RVM 5HT-mediated signaling and increased recruitment of descending serotoninergic projections to the spinal cord are implicated in pain hypersensitivity (Costa-Pereira, Serrao, Martins, & Tavares, 2020; Silva et al., 2016). A significant increase in TpH2 expression in bladder-specific RVM neurons in neonatal cystitis rats in the current study emphasizes the involvement of descending serotonergic pathway in the development of bladder pain syndrome.

We further examined the effect of neonatal cystitis on GABAergic RVM neurons that are synaptically linked to bladder afferents and observed a significant downregulation of GABA transporter VGAT both at mRNA and protein levels in this set of neurons in the RVM. These findings and our previous studies together indicate that GABA disinhibition both at the spinal and supraspinal levels may be involved in long-term spinal sensitization and medullary descending pain modulation in patients with IC\PBS (Kannampalli et al., 2017; Miranda et al., 2011; J. N. Sengupta et al., 2013; J. Zhang et al., 2017). To our knowledge, this is the first study reporting descending GABA projections to be a part of the RVM-bladder pathway. Furthermore, VGAT expression in 50% of the converging RVM neurons also emphasizes a possible involvement of RVM GABAergic neurons in cross-organ sensitization/overlapping pelvic pain in cystitis rats. Using optogenetic/chemogenetic manipulations of genetically identified dorsal horn and RVM neurons, Francois et al (2017) identified a set of RVM GABAergic neurons that facilitate mechanical pain by inhibiting dorsal horn enkephalinergic/GABAergic interneurons (Francois et al., 2017). RVM GABAergic neurons are also reported to project onto parvalbumin-positive spinal interneurons in inner lamina II (Antal, Petko, Polgar, Heizmann, & Storm-Mathisen, 1996; Petitjean et al., 2015). These findings together suggest that multiple parallel GABAergic systems are involved in spinal pain processing of distinct sensory modalities.

In summary, we report that the RVM neurons innervating from the bladder and colon exhibit altered molecular expression and synaptic connectivity when newly born pups (P14–16) develop cystitis. Compared to saline-treated control rats, neonatal zymosan-treated cystitis rats exhibit an increase in TpH2 expression and downregulation of GABA transporter VGAT in RVM neurons that are synaptically linked to the bladder. These findings provide clear evidence that the alterations in synaptic connectivity as well as GABAergic/serotonergic neurotransmission in bladder related RVM neurons may play a pivotal role in the underlying mechanism of cross-organ sensitization and long-lasting pelvic pain as observed in patients with IC\PBS.

Supplementary Material

fS1

Supplement Figure 1: Schematic diagram showing fluorogold (FG) and PRV treatment protocol in naïve rats. Illustration showing possible neural connecting pathways of the primary sensory afferents from the bladder and spinal dorsal surface to the RVM (a). The treatment protocol for naïve rats (b). FG was applied onto the dorsal side of the spinal cord in naïve rats on P57 followed by PRV injections into the bladder on P60 and tissues were collected 96 hours post viral inoculation.

fS2

Supplement Figure 2: Spinal cord sections showing FG application sites on the dorsal horn and PRV tracing from the bladder in naïve rats. Fluorogold (a1) and PRV-152 (a2) labeling are shown under lower magnification, scale bar is 250 µm. Higher magnification images of the box area identify distinct distribution patterns of FG and PRV-152 in the left dorsal horn area of the spinal cord section with less than 10% colocalization of FG and PRV-152 tracing in spinal neurons (b1-b3). Arrows represent PRV+ neurons without FG labeling. The scale bar is 100 µm.

fS3

Supplement Figure 3: Distribution pattern of descending spinal projections and ascending inputs from the bladder in RVM neurons in naïve rats. Lower magnification images of RVM sections with PRV-152 and FG labeling within different RVM regions (a1-a3). The scale bar is 200 µm. Higher magnification of box1 (b1-b3) and box2 (c1-c3) showing PRV 152 and FG in left and right sides of RVM section. The scale bar is 100 µm. The block white arrows represent GFP+ neurons that are merging with FG and the hollow white arrows represent GFP+ neurons without FG labeling. Data analysis has shown that about 59%±2.3% of GFP+ neurons exhibit FG labeling (n=4, 3 tissue sections/rat, data is represented as mean ± SEM).

fS4

Supplement Figure 4: The analysis for specificity of different antibodies used in this study. Representative images of double immunostaining of RVM tissue sections from rats which only received PRV-152 (GFP+) in the bladder for GFP (green) and RFP (magenta) expression (a-c). The tissue sections were immunostained with a mixture of chicken anti-GFP and rabbit anti-RFP as primary antibodies, scale bar is 200µm. Representative images of double immunostaining of RVM sections with anti-TpH2 and anti-5HT antibodies (d-f). scale bar is 100µm. Representative images of VGAT staining using peptide preabsorbed VGAT antibody (g-i) and unabsorbed VGAT antibodies (j-l) for separate sets of RVM sections. scale bar is 50µm. White arrows in (j-l) represent VGAT specific staining with peptide unabsorbed antibody.

Highlights.

• RVM neurons receiving synaptic inputs from bladder and colon are present in regions involved in descending pain modulation.

• Bladder related RVM neurons express GABAergic, enkephalinergic and serotonergic markers.

• An alteration in converging neurons in RVM receiving dual innervation from bladder and colon in neonatal cystitis rats.

• Downregulation of VGAT expression in the bladder related RVM neurons in neonatal cystitis rats.

Abbreviations:

RVM

Rostral Ventromedial Medulla

PRV

Pseudorabies Virus

TpH2

Tryptophan hydroxylase 2

5-HT

5-Hydroxytryptamine

GABA

Gamma-aminobutyric acid

FG

Fluorogold

RMg

Raphe Magnus

Gi

Gigantocellular reticular nucleus

GiA

Gigantocellularis Pars Alpha

LpGi

Lateral Paragigantocellular Nucleus

VGAT

Vesicular GABA Transporter

PFA

Paraformaldehyde

GFP

Green Fluorescent Protein

RFP

Red Fluorescent Protein

NGS

Normal Goat Serum

pAb

Polyclonal Antibody

mAb

Monoclonal Antibody

TSA

Tyramide Signal Amplification

GAD1

Glutamic Acid Decarboxylase 1

PENK

Proenkephalin

ISH

In situ Hybridization

IHC

Immunohistochemistry

IC

Interstitial Cystitis

IBS

Irritable Bowel Syndrome

CNS

Central Nervous System

SC

Spinal Cord

PBS

Painful Bladder Syndrome

CPP

Chronic Pelvic Pain

Data availability statement:

Data will be made available upon request.