Acyloxyacyl hydrolase modulates pelvic pain severity

Abstract

Chronic pelvic pain causes significant patient morbidity and is a challenge to clinicians. Using a murine neurogenic cystitis model that recapitulates key aspects of interstitial cystitis/bladder pain syndrome (IC), we recently showed that pseudorabies virus (PRV) induces severe pelvic allodynia in BALB/c mice relative to C57BL/6 mice. Here, we report that a quantitative trait locus (QTL) analysis of PRV-induced allodynia in F2_CxB progeny identified a polymorphism on chromosome 13, rs6314295, significantly associated with allodynia (logarithm of odds = 3.11). The nearby gene encoding acyloxyacyl hydrolase (Aoah) was induced in the sacral spinal cord of PRV-infected mice. AOAH-deficient mice exhibited increased vesicomotor reflex in response to bladder distension, consistent with spontaneous bladder hypersensitivity, and increased pelvic allodynia in neurogenic cystitis and postbacterial chronic pain models. AOAH deficiency resulted in greater bladder pathology and tumor necrosis factor production in PRV neurogenic cystitis, markers of increased bladder mast cell activation. AOAH immunoreactivity was detectable along the bladder-brain axis, including in brain sites previously correlated with human chronic pelvic pain. Finally, AOAH-deficient mice had significantly higher levels of bladder vascular endothelial growth factor, an emerging marker of chronic pelvic pain in humans. These findings indicate that AOAH modulates pelvic pain severity, suggesting that allelic variation in Aoah influences pelvic pain in IC.

INTRODUCTION

Interstitial cystitis/bladder pain syndrome (IC/BPS or IC) is a condition of debilitating pelvic pain and voiding dysfunction, afflicting as many as eight million patients in the United States (5). Although IC etiology remains unknown, hypothesized pathogenic mechanisms include bladder inflammation, barrier dysfunction, infection, nerve injury, pelvic floor dysfunction, and mast cell activation (32). Because of the profound impact on patients and many unanswered clinical and pathogenesis questions, the National Institute of Diabetes and Digestive and Kidney Diseases established the Multi-Disciplinary Approaches to Chronic Pelvic Pain (MAPP) Research Network, its flagship effort to understand mechanisms of urological chronic pelvic pain syndromes (UCPPS). The MAPP Network is a multicenter network dedicated to clinical, epidemiological, and mechanistic studies of UCPPS (15, 38).

One proposed model of IC pathogenesis involves a positive feedback loop, whereby substance P-containing peripheral nerves activate mast cells, in turn releasing inflammatory mediators that induce urothelial inflammation (23). In this model, histamine released by mast cells feeds back onto peripheral nerves to maintain sustained substance P release and mast cell activation. Consistent with this model, some patients have elevated mast cell counts in the bladder lamina propria and high levels of urinary histamine metabolites, and mast cell activation in bladder lamina propria is correlated with IC symptoms, supporting a role for mast cells in at least a subset of patients (1, 6, 22, 47, 80).

Recent studies in rodent neurogenic cystitis recapitulate IC clinical findings and roles for mast cells. Infection of rats with attenuated Bartha’s pseudorabies virus (PRV) via the tailbase muscle induces a centrally mediated neurogenic cystitis accompanied by mast cell activation that was attenuated by either resection of bladder innervation or ablation of Barrington’s nucleus, a brain center of bladder control (34, 35). In mice, PRV promoted mast cell trafficking to the bladder lamina propria and increased bladder permeability driven by mast cell production of tumor necrosis factor (TNF) (9, 11, 13). PRV also induced pelvic allodynia that was subject to colonic modulation, reminiscent of the exquisite sensitivity of many IC patients to specific dietary constituents (41, 42, 65, 67, 70, 75). Although pelvic allodynia was dependent on mast cells, allodynia was strictly dependent on histamine through sensory activation of Ca²⁺/calmodulin-dependent kinase II (CaMKII) in the dorsal horn of the sacral spinal cord (65, 88). Together, these studies identify peripheral and central mechanisms contributing to pelvic pain in a model mimicking IC. Thus, PRV neurogenic cystitis is a pelvic pain model that is recognized by the MAPP Network for its clinical relevance (44).

We recently reported that PRV differentially induces pelvic allodynia in mouse strains (69). Here, we employ a genetic strategy to identify novel loci modulating pelvic pain. Different Mus musculus strains offer distinct advantages that informed our genetic strategy. First, we have historically employed C57BL/6J mice (B6) within our laboratory for the PRV model of neurogenic cystitis, and knockout mice are most commonly generated on the B6 background. Depending on the model system, B6 mice are typically biased toward Th1 immune responses, whereas BALB/c mice (C) are often biased toward Th2 responses (30, 48); thus, a B6xC strategy offers the potential to identify inflammatory loci potentially modulating pelvic pain responses. Finally, our previous studies demonstrating quantitative differences in PRV-induced pelvic allodynia in B6 and BALB/c mice indicate that a B6xC strategy would likely identify loci mediating differential pain responses. In this study, we identified a SNP linked to Aoah (acyloxyacyl hydrolase) that is associated with pelvic allodynia. AOAH-deficient mice recapitulate key aspects of IC and have elevated expression of vascular endothelial growth factor (VEGF), an emerging IC biomarker. AOAH is expressed along the bladder-brain axis for a potential role in modulating pelvic pain via the bladder sensory system. These findings suggest that Aoah is a novel genetic modulator of pelvic pain that may play a role in the pathogenesis of IC.

MATERIALS AND METHODS

Animals.

Only female mice were studied except bacterial infection experiments. C57BL/6 (B6), BALB/c (C), CB6F1 (BALB/c X C57BL/6 F1 hybrids), and CXB-6 (the recombinant inbred strain) were purchased from Jackson Laboratory. F2 mice were bred in the Center for Comparative Medicine of Northwestern University. AOAH knockout mice (Aoah^−/−) on the C57BL/6 background, a gift from Dr. Robert Munford (National Institute of Allergy and Infectious Diseases, Bethesda, MD), were generated as described previously (49–51). Adult mice aged at 10–14 wk were used for all experiments. All mice were housed under specific pathogen-free conditions. All experiments were performed using protocols approved by Northwestern University Institutional Animal Care and Use Committee.

Murine neurogenic cystitis.

Murine neurogenic cystitis was performed as described previously (65). Briefly, 10- to 14-wk-old female mice were injected 2.29 × 10⁶ pfu of Bartha’s strain of PRV via the abductor caudalis dorsalis muscle. Ultraviolet irradiated/heat inactivated PRV was used as negative control inoculum in sham-treated mice.

Murine urinary tract infection.

Murine urinary tract infection (UTI) was performed as described previously (68). To induce murine post-UTI chronic pain, 6-mo-old male mice were instilled via transurethral catheter with 10⁸ CFU of Sϕ874 (a K-12 Escherichia. coli strain).

Behavioral testing.

Mice were tested before PRV inoculation or Sϕ874 infection (baseline) and at 1–4 days after PRV administration, or PID 5, 7, 14, and 21 for Sϕ874 infection. Referred hyperalgesia and tactile allodynia were quantified using von Frey filaments applied to the abdomen and plantar regions of the hind paw, as previously described (69). Where possible, allodynia experiments were blinded by providing the tester with only coded mouse cages. Coat color differences precluded blinding of parental and F1_CxB mice (Fig. 1A), and randomness of recombinant genotypes in F2_CxB mice rendered blinding unnecessary during allodynia determination for quantitative trait locus (QTL) (Fig. 1B).

Fig. 1.

Quantitative trait locus (QTL) analysis for pelvic allodynia. A: genetic background affects pelvic allodynia. Pseudorabies virus (PRV)-infected BALB/c (C) mice exhibit severe allodynia, whereas F1_CxB mice exhibit C57BL/6 (B6)-like allodynia. B: allodynia was quantified for individual F2_CxB mice at postinfection day 4 (PID4). C: log-of-odds scores (LOD) for association with individual SNPs was determined using R/qtl; dashed line indicates threshold for significance. Only single nucleotide polymorphism (SNP) rs6314295 exceeded threshold. D: allodynia of F2 mice relative to genotype. E: allodynia induced by PRV was significantly greater in C mice and CXB-6 mice, relative to B6, at PID4 (***P < 0.001). F: map of home genome around SNP rs6314295 with nearby genes, as indicated by arrow, which were identified in databases (NCBI map coordinates).

Genotyping.

Genomic DNA was extracted from tails of 96 F2 mice using the Qiagen DNeasy Blood and Tissue Kit. The Illumina Mouse Medium-Density (MD) Linkage Panel was used for genotyping, performed by the Genome Center at Washington University. Using Illumina the GoldenGate Assay and Illumina Beadstation, the MD panel genotyped each mouse for 1,449 SNPs, of which 878 SNPs were informative that distinguishes between C57BL/6J and BALB/cJ genome. Genotype data were analyzed using R/qtl software (7). Genes in the region of interest were determined using Mouse Genome Informatics from Jackson Laboratory. Reported map positions were retrieved from the SNP (single nucleotide polymorphisms) database of the National Center for Biotechnology Information (NCBI).

Measurement of AOAH expression.

The sacral spinal cord was dissected from mice immediately following euthanasia by cervical dislocation and stored at −80°C. All tissues were homogenized using ice-cold TRIzol (Invitrogen), according to the manufacturer’s protocol. RNA was reverse transcribed using an RT2 First Strand Kit (SABiosciences). Quantitative real-time PCR analysis was performed using RT2 qPCR Mastermix (SABiosciences) in a MJ Research Chromo 4 thermocycler. Primers were designed for Aoah (forward, 5′-GGGTGTGTGGTACTGGTATCT-3′ and reverse, 5′-TGAACCAAGAAATAGCAGGCG-3′); L19 (forward, 5′-CCATGAGTATGCT CAGGCTTCAGA-3′; reverse, 5′-TACAGGCTGTGATACATGTGGCGA-3′). Changes in relative gene expression were normalized to ribosomal L19 mRNA levels, using the relative cycle threshold method.

Histology and immunofluorescence.

All mice were anesthetized with isoflurane, and the bladders were collected for histology using standard protocols. For pathological assessment, bladder sections were stained with hematoxylin and eosin (H&E) by the Mouse Histology & Phenotyping Laboratory, Northwestern University. To visualize, mast cell bladder sections were deparaffinized, rehydrated, and stained with 0.1% toluidine blue (Sigma) for 10 min. Mast cells were quantified from two nonadjacent sections. For immunofluoresence, mice were perfused with PBS followed by 4% paraformaldehyde. The bladder, spinal cord, and brain are removed and processed for paraffin embedding by the Mouse Phenotyping and Histology Laboratory of Feinberg School of Medicine. Four-micrometer sections were deparaffinized and rehydrated with standard methods. Immunofluorescence staining was optimized for each tissue by comparing wild-type and AOAH-deficient tissues. For bladder, sections were blocked with both Fc Receptor Blocker and Background Buster (Innovex) for 30 min each. Sections were incubated at 4°C overnight with rabbit anti-AOAH antibodies (1:100, Santa Cruz sc-135110) followed by detection with Alexa fluor 488 donkey anti-rabbit IgG (1:500, Invitrogen A21206) and counterstaining with DAPI. For spinal cord, sections were blocked as described above, and tissue sections were incubated at 4°C overnight with goat anti-AOAH antibodies (1:100, Santa Cruz sc-163692) followed by detection with Alexa fluor 488 donkey anti-goat IgG (1:500, Invitrogen A11055). For brain, antigens were retrieved using a decloaking chamber and Target retrieval solution (Dako S1699) in two steps. First, the sections were incubated at 120°C for 20 s and then 90°C for 20 s. Sections were rinsed and blocked as above. Sections were incubated at 4°C overnight with rabbit anti-AOAH antibodies (1:100, Santa Cruz sc-135110) followed by detection with Alexa fluor 488 donkey anti-rabbit IgG as above. In all cases, omission of the primary antibodies was used as a control. Brain structures were identified by comparison with a mouse brain atlas by a blinded reviewer (85). VEGF immunoreactivity was detected similarly with the following modifications: anti-VEGF mouse monoclonal antibody C-1 (Santa Cruz sc-7269) was used at a dilution of 1:100 and detected with Alex fluor 488 goat anti-mouse IgG (Invitrogen A11029) at a dilution of 1:500.

ELISA.

Mouse urine was collected, and TNF concentration was determined by mouse TNF-α DuoSet ELISA kit (R&D Systems) according to the manufacturer’s instructions. For VEGF determination, bladders were harvested and homogenized in 0.1 ml of PBS using a Tissue Tearor homogenizer (Biospec Products). Bladder VEGF was then quantified using a mouse VEGF DuoSet ELISA kit (R&D Systems).

Sacral spinal cord recording.

The sacral spinal cord was prepared as previously described (68). Briefly, the spinal column was exposed at upper lumbar under anesthesia, perfused with artificial cerebrospinal fluid (ACSF), and then transected at the middle of the lumbar enlargement immediately after euthanasia. The distal part of the spinal cord with dorsal and ventral roots was transferred to a dish containing ACSF. After separation of sacral ventral roots 1–3 (S1–S3) and dorsal roots on each side of the cord and removing all other roots, six recording electrodes were mounted on the ventral roots (S1–S3 on each side) and were then connected to six DAM 50 amplifiers (WPI) in differential mode with 1,000× gain, high-pass filtering at 300 Hz, and low-pass filtering at 20 kHz. Amplifier outputs were transferred to an A-to-D interface (DT1322A, Molecular Devices), and signals were digitized at 50 kHz and acquired using pCLAMP v.9.1 software (Molecular Devices). Spontaneous action potentials were recorded for 30 min. For each animal, the firing units per root were the average from all recorded roots.

To record evoked motor outputs in response to dorsal stimulation, the stimulation threshold was first determined by adjusting the intensity of a 0.2-ms current pulse. Then, five pulses at 25 Hz with intensities of 1, 1.5, 2, and 5 times the threshold stimulus were used. The evoked peak-to-peak compound action potential (coAP) was averaged for each pulse, P2–P5, relative to the coAP at P1.

Statistical analysis.

Results are expressed as means ± SE. When the data were compared between two groups, they were analyzed with Student's t-test, while data compared from more than two groups were analyzed by one-way ANOVA followed by a posttest comparison using either Bonferroni’s or Tukey’s multiple comparison test. All analyses were done using Prism software (GraphPad). A value of P < 0.05 was considered statistically significant.

RESULTS

Pelvic allodynia of neurogenic cystitis induced by PRV is a quantitative trait.

Visceral nociception is manifested as tactile allodynia and corresponding tactile hypersensitity referred to a skin surface, and quantifying allodynia is an accepted measure in MAPP Network mechanistic studies (44). Previously, we (65) described that infecting female C57BL/6 mice with PRV via tail base resulted in progressive tactile allodynia specific to the pelvic region, whereas sham mice infected with inactivated PRV did not elicit such a response. More recently, we (69) detected significant, differential pelvic nociception in C57BL/6 and BALB/c mice, where BALB/c mice developed greater allodynia. These data suggest that genetic factors may determine pelvic pain susceptibility associated with neurogenic cystitis.

Here, we examined the potential genetic contribution to pelvic allodynia by comparing BALB/c and C57BL/6 with F1 hybrid mice, F1_CxB (JAX CB6F1). After PRV infection, tactile allodynia was quantified as increased responses, relative to baseline, to pelvic stimulation with graded von Frey filaments (Fig. 1A). PRV induced progressive pelvic allodynia in C57BL/6 mice, whereas BALB/c mice developed more severe pelvic allodynia. Pelvic stimulation of female F1_CxB mice also resulted in progressive allodynia that was significantly greater than baseline responsiveness by PID 3 and 4 (Fig. 1A; P < 0.05). However, F1 hybrids exhibited only moderate pelvic allodynia, similar to that observed in C57BL/6 mice. Thus, the severe pelvic pain phenotype of BALB/c mice behaved as a recessive trait in the F1 generation, suggesting that dissecting mechanisms of neurogenic cystitis pelvic pain modulation is amenable to genetic approaches.

F1_CxB mice were intercrossed to generate a cohort of F2 female progeny. Quantifying PRV-induced allodynia, we found that the extent of pelvic pain response in the F2 generation was distributed across a range from mice with no allodynia to mice with a 3,400% increase response (Fig. 1B). The broad phenotypic distribution in the F2 generation suggests that the pelvic pain severity phenotype is inherited as a quantitative trait and is modulated by allelic variants at multiple genetic loci.

To identify the loci modulating pelvic pain severity, we undertook QTL analysis in F2 mice, using the magnitude of PRV-induced pelvic allodynia increase as the phenotypic variable. To facilitate subsequent QTL analysis, we scored the pelvic pain as the logarithm of percent increase in pelvic allydonia at PID 4. A total of 96 F2 female mice were analyzed, and pelvic pain scores showed a bimodal distribution (log₂ range 0–12), with 49% of mice having a score <6 and 51% a score >6. Of the F2 mice, 26% had no pelvic pain response (Fig. 1B).

To genotype the F2 mice, genomic DNA was isolated from 96 mice and genotyped using the Illumina Mouse MD Linkage Panel (878 of 1,449 SNPs are informative between BALB/c and C57BL/6). Linkage analysis was performed using R/qtl software to identify potential QTLs associated with pelvic pain severity (7). A one-dimensional genome scan strategy was used. Statistical significance for such studies is traditionally calculated by a logarithm of odds (LOD) score, where LOD >3 is considered significant (45). Only one locus, SNP rs6314295 on chromosome 13, exceeded the threshold for significance (LOD score of 3.11), suggesting a strong association between the locus and the phenotype of severe pelvic pain (Fig. 1C). A second locus, defined by SNP rs13477272, did not achieve significance (LOD score of 2.54), possibly due to limitations in sample size or assay variability. SNP rs13477272 resides on mouse chromosome 3 within Adamtsl4, encoding ADAMTS-like protein-4, a member of a thromobospondin repeat family of extracellular matrix-associated proteases involved in neural development, with ADMATS-LP4 especially expressed in the eye and associated with ocular disorders (14). Rs13477272 lies within a region syntenic with human chromosome 1q21.2, and nearby loci with potential relevance to pain include Arnt (transcriptional regulation), Pip5k1a (phosphoinositide signaling), Cers2 (ceramide biosynthesis), and Sv2a (neurotransmitter release).

To evaluate the impact of SNP rs6314295 genotype on pelvic pain severity, we stratified the F2 mice on the basis of their genotype. Animals homozygous for C57BL/6 alleles had an average pelvic pain score of 7.8, whereas those heterozygous or homozygous for BALB/c alleles were at 4.4 or 6.2, respectively (Fig. 1D). This finding of greater pain associated with the B6/B6 genotype at rs6314295 among F2 mice was unexpected, because BALB/c mice possess the severe pain phenotype (Fig. 1A). This suggests that the pelvic pain phenotype is dependent on interactions between the rs6314295 locus and other loci, also consistent with the relatively large variability in pelvic pain scores of F2 (Fig. 1B).

To confirm the QTL association with pelvic pain, we employed CXB recombinant inbred (RI) mice, a panel of strains derived from the parental strains BALB/c and C57BL/6 that have unique mosaic genotypes resulting from intercrossing of progeny from an initial cross between BALB/c and C57BL/6 (89). The CXB RI strains have been used in the genetic analysis of numerous complex or potentially complex physiological phenotypes including differences in thyroid function (Graves' disease) and pulmonary inflammation as well as behavioral phenotypes including avoidance, exploration, and locomotor activity (53, 56, 59, 81). Within the CXB panel, we identified the CXB-6 RI strain as a strain harboring the B6 alleles at rs6314295 and rs13477272. PRV-infected CXB-6 mice exhibited significantly greater allodynia than B6 mice at PID 4, supporting the possibility that the region near rs6314295 modulates pelvic pain severity, perhaps in acting in concert with rs13477272 (Fig. 1E).

Aoah as a candidate gene for pelvic pain severity.

We hypothesized that pelvic pain QTL studies would identify factors associated with neural transmission, such as ion channels or neurotransmitter receptors, but the 1.5 LOD confidence interval of rs6314295, spanning a 10 MB region, contained no such genes (Fig. 1F). To identify candidate genes near rs6314295 modulating pelvic pain, we compared whole sacral spinal cord genome transcription profiles of C57BL/6 mice in sham, PID 2 and PID 4 groups following PRV infection from our previous study (88). Reviewing our sacral spinal cord microarray data, we observed that Aoah was the only induced gene within this region during PRV infection, and we confirmed Aoah induction by quantitative (q)RT-PCR (Fig. 2A; P < 0.05). AOAH is a lipase expressed in neutrophils and other leukocytes that detoxifies lipopolysaccharide (LPS) by selectively removing the secondary fatty acyl chains from the lipid A moiety (24, 31, 49, 58, 76). However, recently AOAH polymorphisms have been correlated with asthma, and “trans-eQTL” studies showed that altered AOAH expression is associated with numerous polymorphisms previously correlated with chronic inflammatory diseases including ulcerative colitis (3, 26). Together, this suggests AOAH has additional host functions beyond detoxifying LPS.

Fig. 2.

Identification of the gene encoding acyloxyacyl hydrolase (Aoah) as the candidate gene for pelvic pain. A: Aoah expression, relative to L19, was undetectable by qRT-PCR in sham mice but was induced by postinfection day (PID) and day 4. Bartha’s pseudorabies virus (PRV)-induced allodynia in WT (B) and Aoah^−/− mice (C). D: PRV induced no significant change in tactile sensitivity (50% threshold) of the plantar region of paw at PID4. E: pelvic allodynia was significantly higher in PRV-infected Aoah^−/− than in WT mice in response to the finest von Frey filament (*P < 0.05).

Given the unexpected finding of a neutrophil gene as a modulator of pelvic pain, we obtained AOAH-deficient (Aoah^−/−) mice with a targeted disruption of the first Aoah exon (50, 51). To evaluate the impact of AOAH on pelvic pain severity, we quantified pelvic pain behavior in Aoah^−/− female mice during neurogenic cystitis. Aoah^−/− mice showed baseline responses that were significantly greater than those of wild-type B6 mice, and allodynia was induced in both PRV-infected B6 and Aoah^−/− mice by PID 2 to 4 (Fig. 2, B and C; P < 0.05). In contrast, PRV induced no change in tactile sensitivity of the plantar region of the hindpaw of wild-type and Aoah^−/− mice (Fig. 2D), consistent with our previous reports and suggesting that AOAH is not a global modulator of pain but instead is a regional pain modulator exhibiting specificity for the pelvic region (67). Focusing on only the smallest stimulus evoked by the finest von Frey filament, Aoah^−/− mice behaved with progressive responsiveness that was significantly greater than baseline by PID 2 (Fig. 2E; P < 0.05). In contrast, wild-type mice failed to exhibit pelvic allodynia in response to the finest filament. These data suggest that Aoah modulates pelvic pain at baseline and in response to a nociceptive stimulus. However, these data also identify a paradox regarding the role of Aoah expression in pain: Aoah is induced during PRV-induced pain, but AOAH-deficient mice have elevated pain.

AOAH expression along the bladder-brain axis.

We observed minimal expression of Aoah mRNA in the sacral spinal cord of normal mice, but this was induced strongly in response to PRV (Fig. 2A). To further examine potential roles for AOAH as a pelvic pain modulator, we assessed expression along the bladder-brain axis by immunofluorescence (Fig. 3). In the bladder, AOAH immunoreactivity was detectable in the suburothelial lamina propria of B6 mice as punctate staining resembling a meshwork of fine fibers, consistent with the lamina propria as the site of bladder sensory fibers; yet similar staining was not detectable in bladders of AOAH-deficient mice (compare Fig. 3, A and B). Diffuse and punctate staining specific to B6 bladders was also visible within the urothelium, and in the detrusor smooth muscle layer, albeit less intense. As in the bladder, AOAH immunoreactivity was detectable in the sacral spinal cord of B6 mice that was not observed in the sacral cord of AOAH-deficient mice (Figs. 3, C and D). AOAH staining was localized to cell bodies in the dorsal horn in lamina I–II, the site of sensory interneurons (4). In the brain, AOAH immunoreactivity was observed in multiple regions (Table 1). Most notably, in B6 mice, AOAH staining was observed in cell bodies of the Barrington’s nucleus, a center of bladder voiding control (74, 82), but similar staining was largely absent in the Barrington’s nucleus of AOAH-deficient mice (Fig. 3, E and F). Unexpectedly in AOAH-deficient Barrington’s nucleus, we observed occasional cell bodies that exhibited AOAH immunoreactivity. Given that AOAH-deficient mice were generated by insertion of a neomycin resistance cassette into the first exon, we speculate that these residual AOAH-positive cells indicate leakiness of the knockout in some cells, possibly by expression of an alternatively spliced AOAH variant (76). Nonetheless, AOAH staining colocalized with the neuron-specific marker NeuN and was also observed in Purkinje cell bodies and fibers in the cerebellum (Fig. 3, G and H). Thus, AOAH is expressed in neurons at multiple sites along the bladder-brain axis, so AOAH is potentially poised to modulate bladder sensory signals.

Fig. 3.

Acyloxyacyl hydrolase (AOAH) expression along the bladder-brain axis. Paraffin sections were stained with anti-AOAH antibodies or neuronal marker NeuN and visualized by immunofluorescence. A and B: AOAH expression in the bladder of C57BL/6 (B6) or Aoah^−/− mice, respectively. Lamina propria staining along fibers (arrows) occasionally projecting to luminal surface (arrowhead). C and D: AOAH expression in the sacral spinal cord of B6 or Aoah^−/− mice, respectively. AOAH-positive cell bodies (C and inset) were evident in dorsal horn lamina I and II. E and F: AOAH expression in the Barrington’s nucleus of B6 or Aoah^−/− mice, respectively. AOAH-positive cell bodies were evident throughout the Barrington’s nucleus; in E, the boundary with the cerebellum is shown for orientation. G: AOAH staining (green) colocalized with the pan-neuronal marker NeuN (red) in double-labeled cells (arrow, yellow). H: AOAH immunoreactivity was observed in the cerebellum with Purkinje cell bodies (arrow) and Purkinje cell fibers (arrowhead). Scale bars, 100 µm.

Table 1.

Brain regions with AOAH immunoreactivity

Brain Region
Glomerular layer of olfactory bulb
Paraventricular nucleus of the hypothalamus
Medial longitudinal fasciculus
Periaqueductal gray
Bed nucleus of stria terminalis
Pontine reticular nucleus
Dorsal motor nucleus of the vagus
Medial vestibular nucleus parvicellular
Barrington’s nucleus
Locus coeruleus
Cerebellar Purkinje cells
Corpus callosum
Subcoeruleus dorsal and ventral

AOAH, acyloxyacyl hydrolase.

Sacral spinal cord responses to AOAH deficiency.

PRV induces a neurogenic cystitis with features of neuropathic pain. Given that the sacral spinal cord is a critical nexus of bladder sensation and control (86), we examined the sacral spinal cord for evidence of altered responsiveness. We recently demonstrated hyperexcitability within the sacral spinal cord of mice exhibiting chronic allodynia following transient UTI (68). By use of an isolated preparation of sacral spinal cord, spontaneous firing and short-term depression (STD) were quantified in ventral roots as a measurement of coupled sensory and motor functions (36, 37). This approach quantifies motor output of flexion (withdrawal) reflex, a reflex that has long been used as a pain index, and mice with post-UTI chronic pelvic allodynia had sacral spinal hyperexcitability manifested as deficits in STD, consistent with neuropathic pain (68). Here, we quantified STD of motor output to assess normal plasticity by stimulating dorsal inputs and recording ventral outputs. Singular dorsal roots were stimulated with five consecutive current pulses (P1–P5) at 40-ms intervals at multiple stimulus intensities. Responses to pulses P2–P5 were then normalized to first pulse (P1) to facilitate comparisons between these two groups. Relative to sham-infected mice, the sacral spinal cords of PRV-infected wild-type mice exhibited defects in STD at multiple pulses (P2–P4) and multiple stimulus intensities (Fig. 4A, top and bottom; P < 0.05). In contrast, Aoah^−/− mice with spontaneous allodynia exhibited a small STD deficit manifested only at P2 and only at the lowest stimulus intensity (Fig. 4B and data not shown; P < 0.01).

Fig. 4.

Acyloxyacyl hydrolase (AOAH) modulates sacral spinal cord excitability. Spontaneous action potentials and evoked potentials were quantified in sacral spinal cords ex vivo at ventral roots S1–S3. A: short-term depression (STD) was quantified in sham and pseudorabies virus (PRV)-treated mice as compound action potentials and expressed as percent initial pulse (%P1). At 2× current intensity (top) and 5× current intensity (bottom), and PRV-treated mice showed reduced STD at P2–P4. B: STD was quantified as compound action potentials and expressed as %P1. At 1× current intensity, Aoah^−/−mice showed reduced STD at P2. C: no difference of spontaneous firing activity in sacral spinal cords between Aoah^−/− and WT mice. *P < 0.05, **P < 0.01, ***P < 0.001.

Since AOAH-deficient mice exhibited only modest STD deficits, spontaneous sacral spinal cord activity was quantified using pCLAMP software to recognize discrete action potential characteristics and then determine the number of spontaneously firing neurons within the flexion reflex. Spinal cords of Aoah^−/− mice with spontaneous allodynia exhibited spontaneous firing similar to that of wild-type mice, suggesting no difference in spontaneous activity of sacral spinal circuits between these two groups (Fig. 4C). Based on the absence of large effects on intrinsic spinal excitability in this isolated preparation, these data suggest Aoah^−/− mice exhibit spontaneous pelvic pain due to alterations of the bladder-brain axis that lie mostly outside the spinal cord or altered sensory processing of spinal circuits that is largely manifested only in situ. Alternatively, given the localization of AOAH to the dorsal horn, it is possible that ascending sensory information that is shaped by AOAH is only modestly reflected in ventral output.

AOAH modulates spontaneous and peripherally induced pain.

Since Aoah was identified in QTL mapping using PRV neurogenic cystitis in female mice, we sought to determine whether modulation of pelvic pain by AOAH extends to other models and/or male mice. IC patients often have a history of UTI before the onset of IC (84); therefore, we recently developed a post-UTI chronic pain model induced by E. coli strain SΦ874 that mimics a potential infectious etiology of IC. We found that post-UTI chronic pelvic pain in B6 female mice persisted for at least 35 days following a single, transient infection, and the pain was associated with sacral spinal cord hyperexcitability (68). In addition, whereas PRV induces pelvic allodynia that is subject to modulation by other visceral inputs (66), SΦ874 infection of the bladder presumably induces pain that is strictly mediated along the bladder-brain axis. Here, Aoah^−/− male mice previously infected with SΦ874 showed significantly greater allodynia on PID 21 relative to wild-type males (Fig. 5A; P < 0.05). To further evaluate visceral hypersensitivity associated with AOAH deficiency in female mice (Fig. 2), we quantified visceromotor reflex (VMR) as electromyographic activity of abdominal musculature in response to bladder distension with saline (17, 60). In wild-type mice, only modest activity was evoked in response to bladder distension, whereas AOAH-deficient mice had significantly increased VMR at multiple intravesical pressures (Fig. 5B; P < 0.05). Together, these data demonstrate that AOAH modulation of pelvic pain is not restricted to centrally or peripherally mediated models of infectious etiology of IC (PRV or E. coli, respectively); nor is pain modulation restricted by sex. Moreover, spontaneously increased VMR in AOAH-deficient mice is consistent with elevated sensory responses emanating from the bladder.

Fig. 5.

Acyloxyacyl hydrolase (AOAH) modulates visceral pain. A: male mice were infected with E. coli strain Sϕ874 and assessed for allodynia at postinfection day (PID)21. Pelvic allodynia was significantly higher in infected Aoah^−/− mice than in WT B6 mice. *P < 0.02. B: visceromotor reflex (VMR) was quantified in female mice as Electromyography (EMG) activity in response to bladder distension under anesthesia. EMG was significantly elevated in Aoah^−/− mice relative to WT B6 mice. *P < 0.05.

AOAH modulates bladder pathology.

Because AOAH modulates pelvic pain, we compared bladder pathology from wild-type and Aoah^−/− mice. H&E-stained sections of bladders from Aoah^−/− mice were grossly normal, compared with wild-type mice (Fig. 6A). Clinical studies suggest mast cell involvement in the pathophysiology of IC, and mast cells mediate bladder inflammation in various rodent cystitis models, including PRV neurogenic cystitis (28, 65, 67, 72, 73). We hypothesized that bladder mast cells mediate pelvic hypersensitivity of Aoah^−/− mice. Staining with toluidine blue to reveal granulated mast cells, we observed significantly more mast cells in the bladder of Aoah^−/− than of wild-type mice (Fig. 6B; P < 0.05), suggesting that AOAH modulates mast cell recruitment to the mouse bladder.

Fig. 6.

Acyloxyacyl hydrolase (AOAH) modulates mast cell and urothelial lesions. A: hematoxylin-eosin (H&E)-stained sections of urinary bladders of WT and Aoah^−/− mice. Representative histology images of H&E-stained section from WT (left) and Aoah^−/− mice (right). Scale bar, 200 µm (top), 100 µm (bottom). B: mast cells were quantified in toluidine blue-stained sections. Significantly more mast cells were identified in bladders of Aoah^−/− mice relative to WT mice (*P < 0.05). C: bladder section from an Aoah^−/− mouse infected with pseudorabies virus (PRV), where lesions (yellow arrows) were identified as areas of urothelium with multiple, contiguous TUNEL-positive nuclei (green) lacking overlying uroplakin III (red). D: lesions were quantified in nonserial bladder sections. Significantly more lesions were detected in bladders of Aoah^−/− mice infected with Bartha’s PRV relative to WT mice (*P < 0.05).

Pathological assessment of patient biopsies from the IC Database Study revealed an association between patient symptoms and urothelial lesions (80). Similarly, PRV neurogenic cystitis results in mast cell-dependent urothelial lesions due to TNF acting upon TNFR1 (9, 11, 13). We examined the bladders of PRV-infected Aoah^−/− mice for evidence of apoptotic lesions by TUNEL staining of nonserial sections of PID 4 bladders and staining for uroplakin III, a marker of intact urothelium (Fig. 6C). Apoptosis was seldom visualized in bladders of either wild-type or Aoah^−/− sham mice (Fig. 6D). In contrast, although both wild-type and Aoah^−/− PRV-infected mice developed significantly more apoptotic lesions than sham mice, apoptotic lesions were more prevalent in Aoah^−/− mice than in wild-type mice (Fig. 6, C and D). Finally, we observed that Aoah^−/− mice had urinary TNF of 33.9 ± 0.8 pg/ml, whereas none was detected in urine of wild-type mice (P < 0.05), supporting a role of mast cells and TNF as mediators of increased apoptotic lesions in bladders of Aoah^−/− mice.

Elevated VEGF in AOAH deficiency.

Altered bladder vascularization has long been associated with at least a subset of IC patients (78). Similarly, elevated VEGF in IC patient biopsies has been associated with pain, and VEGF receptors exhibit altered expression (39, 71). We quantified VEGF in bladder homogenates of mice by ELISA (Fig. 7A). Bladder homogenates of B6 mice with PRV-induced neurogenic cystitis had modestly elevated VEGF relative to B6 mice (28%, P < 0.05), consistent with previous reports of increased vascular permeability in PRV neurogenic cystitis (8, 34). In contrast, AOAH-deficient mice had bladder VEGF that was elevated 5.4-fold (P < 0.05). To localize the source of elevated bladder VEGF, we stained bladder sections of wild-type and AOAH-deficient mice. VEGF immunoreactivity was detected in the urothelium of B6 mice; yet VEGF staining was more intense in the urothelium of AOAH-deficient mice (Fig. 7B, compare panels), suggesting that AOAH deficiency results in increased bladder VEGF production specifically in the urothelium. Together with pelvic pain and data supporting mast cell-mediated bladder pathology, these findings demonstrate that AOAH deficiency recapitulates key aspects of IC pathogenesis.

Fig. 7.

Elevated vascular endothelial growth factor (VEGF) in acyloxyacyl hydrolase (AOAH)-deficient bladders. A: VEGF was quantified in bladder homogenates by ELISA. Bladders were harvested from B6 mice, uninfected or infected with pseudorabies virus (PRV) to induce neurogenic cystitis or from AOAH-deficient mice. Bladder VEGF was significantly elevated in PRV-treated and AOAH-deficient mice relative to untreated B6 mice (*P < 0.05). B: immunofluorescent imaging of wild-type C57BL/6 (B6; left) and AOAH-deficient bladder sections (right) stained with anti-VEGF antibodies. VEGF immunoreactivity appears more intense in urothelium of AOAH-deficient bladder sections (compare arrows in left and right). Images were captured at ×20 magnification.

DISCUSSION

UCPPS (urological chronic pelvic pain syndromes) cause untold suffering in as many as 10% of adults and are a challenge to clinicians who lack diagnostic tools and effective therapies, stemming largely from a lack of mechanistic understanding (33). Using an unbiased genetic screen in a clinically relevant IC model (44) (42, 66, 69), we identified Aoah as a novel locus that modulates pelvic pain (Figs. 1, 2, and 6). Prior epidemiologic work suggests a genetic component in IC, chronic prostatitis/chronic pelvic pain syndrome and irritable bowel syndrome (reviewed in Ref. 19). Indeed, polymorphisms in IL-4 and β₂-adrenergic receptor (ADBR2) have been correlated with IC, consistent with roles for dysregulated inflammation and the partial efficacy of tricyclic antidepressants in some patients, respectively. AOAH may function in human disease beyond its known roles in acute infection and immune response (16, 27, 31, 49, 61). Supporting a link between Aoah and chronic disease, a recent trans-eQTL study found that HLA polymorphisms associated with diverse chronic inflammatory conditions were also associated with altered Aoah expression (26). Aoah expression is also elevated in IC biopsies (29). Together with findings reported here, these studies support a role for Aoah in chronic disease, including pelvic pain, and Aoah polymorphisms will be assessed for linkage to pelvic pain in MAPP Network patients. However, these findings also illuminate a paradox for the role of AOAH in pelvic pain. We identified Aoah by virtue of PRV-induced expression in the sacral spinal cord (Fig. 2), consistent with elevated expression in IC biopsies. We therefore hypothesized that elevated AOAH induces pain, but this must be reconciled with the observations that AOAH deficiency elevates pain responses (Fig. 2). PRV induction of spinal cord Aoah is likely a fortuitous accident of using a herpesvirus in our studies, as multiple herpesviruses modify lipid metabolism during the viral life cycle (46, 52, 77, 87), suggesting that PRV may employ AOAH to modulate membrane lipids. But because AOAH deficiency is associated with increased pain and inflammation, we speculate that AOAH expression is regulated by feedback and thus is elevated in IC bladders and other chronic disease as a defensive mechanism.

AOAH may impact pelvic pain severity at multiple compartments along the bladder-brain axis, as suggested by detectable immunoreactivity in the bladder lamina propria, the sacral dorsal horn, and several brain regions (Fig. 3). Altered dorsal root reflex in AOAH-deficient mice is consistent with central nervous system (CNS) hyperexcitability (Fig. 4) and suggests that AOAH modulates neuronal function, possibly at the level of diminished inhibitory control (68). In contrast to central functions, we also observed elevated mast cell accumulation in Aoah^−/− bladders and increased urothelial lesions that are a hallmark of PRV neurogenic cystitis and associated with IC symptoms (Fig. 6 and Refs. 12, 40, 80). PRV-induced urothelial lesions are a consequence of TNF released during mast cell activation (12). We observed increased urinary TNF levels, suggesting tonic mast cell activation, perhaps leading to enhanced mast cell accumulation due to stimulation of urothelial RANTES secretion (10). As mast cell histamine drives pelvic pain responses through H1R and H2R (65, 70), AOAH appears to modulate peripheral induction of pelvic pain at the level of mast cells. Thus, AOAH may affect both central and peripheral mechanisms in pelvic pain, and conditional Aoah alleles may facilitate dissection of specific contributions to pelvic pain modulation by various tissues.

We have demonstrated that AOAH modulates spontaneous pelvic pain and pelvic pain induced by central and peripheral triggers (Figs. 2 and 5). However, the full extent and specificity of AOAH sensory modulation remains to be established. We did not observe significant paw allodynia in AOAH-deficient mice, but we detected AOAH immunoreactivity in the dorsal horn of the thoracic and cervical spinal cord (Fig. 3 and data not shown), raising the possibility that AOAH modulates other sensory inputs and therefore may play a role in other pain models or clinical conditions. Consistent with this, AOAH is expressed in diverse brain structures, for example the olfactory bulb (Table 1). Nonetheless, our focus in these studies was pelvic pain, and we note that several structures implicated in bladder control and/or pelvic pain also express AOAH (Table 1 and Fig. 3). Among the UCPPS, men suffer chronic prostatitis/chronic pelvic pain syndrome (CPPS), a condition analogous to IC. In CPPS patients, fMRI studies identified three regions correlated with chronic pain symptoms: the anterior cingulate cortex, the cerebellum, and the pontine micturition center, the human correlate to Barrington’s nucleus in rodents (25). We detected AOAH expression in each of these structures implicated in UCPPS pain, but Barrington’s nucleus is particularly striking for known roles in modulating bladder function (83). Indeed, circuits have been mapped that provide inputs to Barrington’s nucleus, and we note that several of these also express AOAH, including the periaquaductal gray, the paraventricular nucleus of the hypothalamus, and the bed nucleus of the stria terminalis (Table 1 and Ref. 82). Finally, AOAH is expressed in the dorsal motor nucleus of the vagus, possibly consistent with the known association of IC with bowel comorbidities (32, 55). Therefore, within the brain, these findings suggest that AOAH can influence pelvic pain via circuits known to modulate visceral sensation and function that are consistent with clinical features of IC.

AOAH is best understood as a lipase that detoxifies LPS through its hydrolase activity that removes secondary acyl chains from lipid A (24, 31, 57, 58), but its function in the CNS is unclear. Presumably, AOAH acts on host lipids in the CNS. Previous studies found that AOAH possesses phospholipase activity in vitro as well as lysophospholipase, diacylglycerollipase, and acyltransferase activities (58). We found elevated levels of long-chain ceramides in Aoah^−/− spinal cords and reduced pain in PRV-infected mice following treatment with the ceramide synthase inhibitor fumonisin B1 (data not shown). Because ceramides have been linked to pain, this suggests that AOAH modulates central pathways of pelvic pain through sphingolipid metabolism (20, 21, 43, 62). Alternatively, a recent high-throughput screen for serine hydrolase inhibitors demonstrated that AOAH was subject to inhibition by JZL 195 (2). JZL 195 is an inhibitor of fatty acid amide hydrolase and monoacylglyceral lipase, raising the possibility that AOAH has a role in metabolism of endocannabinoids, lipids known to modulate pain (63). Therefore, through its potential to modify multiple classes of important bioactive lipids, AOAH is poised to affect pain signaling.

Because AOAH may act at diverse points along a sensory pathway from the bladder to the CNS, AOAH may also contribute to the etiology of pelvic pain by distinct mechanisms. IC symptoms are exacerbated by stress (32), so any tendency toward neuronal hyperexcitability would lower the threshold for inducing bladder mast cell activation, thereby triggering pain and potentially establishing a “vicious cycle” and chronicity (reviewed in Refs. 23, 79). Thus, AOAH may set a threshold for responses that trigger pelvic pain sensation and chronicity. Similarly, IC patients also often have a history of UTI preceding chronic pelvic pain (84), and we find that AOAH deficiency results in significantly greater development of post-UTI chronic pain (Fig. 5). Although complete clinical AOAH deficiency is unlikely among most IC patients, our genotyping data suggest that the pain phenoytpe of Aoah alleles is dependent on interactions with other loci (Fig. 1). It has been postulated that, analogous to Toll-like receptor 4 (TLR4)-dependent post-UTI chronic pain observed in mice, IC might be initiated in patients with genetic susceptibility to pelvic pain who then suffer a UTI with a bacterial strain possessing a chronic pain phenotype (64). Therefore, in addition to potential roles in bladder sensation, Aoah polymorphisms may also foster the development of IC via post-UTI chronic pain in susceptible patients through reduced LPS detoxification and consequently increased TLR4 signals. Finally, AOAH deficiency in mice leads to robustly elevated bladder VEGF, appearing predominantly as increased VEGF expression in the urothelium, mimicking clinical findings of aberrant VEGF signaling in IC bladders, including recent findings from UCPPS patients in the MAPP Network (18). This aberrant VEGF signaling in IC, combined with altered expression of neuropilin VEGF coreceptors, is postulated as an alternative sensory pathway in the bladder, thus suggesting another potential mechanism of AOAH modulation of pelvic pain (54, 71).

In summary, AOAH affects putative clinical biomarkers and central and peripheral mechanisms of pelvic pain that are consistent with IC. Thus, Aoah provides a novel biomarker candidate for pelvic pain, and AOAH represents a new therapeutic target for treating IC.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases awards R01 DK-066112S1 (D. J. Klumpp) and MAPP U01 DK-82342 (D. J. Klumpp and A.J. Schaeffer).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

W.Y., C.N.R., C.J.H., J.M.R., and D.J.K. conceived and designed research; W.Y., R.E.Y., M.C.J., C.N.R., J.D.D., and J.M.R. performed experiments; W.Y., R.E.Y., M.C.J., C.N.R., J.D.D., and J.M.R. analyzed data; W.Y., M.C.J., C.N.R., C.J.H., J.M.R., A.J.S., and D.J.K. interpreted results of experiments; W.Y., R.E.Y., M.C.J., J.D.D., J.M.R., and D.J.K. prepared figures; W.Y. and D.J.K. drafted manuscript; W.Y., C.J.H., A.J.S., and D.J.K. edited and revised manuscript; A.J.S. and D.J.K. approved final version of manuscript.