1. Introduction

Interstitial cystitis/bladder pain syndrome (IC/BPS) is a debilitating urologic pain disorder that affects roughly 6% of women in the United States.4 Care and treatment for IC/BPS represents a substantial burden on the healthcare system, but these expenditures have not resulted in high patient or provider satisfaction with available treatments.53 Difficulty treating IC/BPS stems in part from a lack of consensus amongst researchers and clinicians regarding the underlying pathophysiology of the disorder.9 However, most agree that multiple overlapping mechanisms likely contribute to the disease state, and efforts are underway to improve patient phenotyping.

One such effort, driven by the NIDDK-funded Multidisciplinary Approach to the Study of Chronic Pelvic Pain (MAPP) Research Network, has demonstrated that female IC/BPS patients with chronic overlapping pain conditions (COPCs), such as irritable bowel syndrome (IBS) and temporomandibular disorder (TMD), show heightened ex vivo cytokine release to stimulation with lipopolysaccharide (LPS), a classic agonist of one highly conserved component of the innate immune system, toll-like receptor-4 (TLR4).47,49 In a study conducted at the University of Iowa as part of the MAPP Epidemiology Phenotyping Study (EPS),10,36 in 66 female IC/BPS patients (40 with comorbid COPCs and 26 with IC/BPS only), a composite score of ex vivo TLR4-stimulated cytokines (interleukins [IL] 1β and 6) from peripheral blood mononuclear cells (PBMCs) was found to be significantly elevated in those patients with comorbid COPCs. Furthermore, this composite score was associated with the spatial extent of comorbid pain measured on a comprehensive body map. In a subsample of patients who underwent experimental pain testing (n = 32), greater pain sensitivity to pressure at the thumbnail was marginally associated with higher TLR4 composite scores as well.47 Together, these results were interpreted to suggest that central nervous system amplification and maintenance of pain (ie, nociplastic pain) is associated with ex vivo TLR4-stimulated cytokine/chemokine release in IC/BPS. This possibility is supported by a number of animal models indicating a role for TLR4 in pain augmentation in the central nervous system and studies demonstrating that the ex vivo peripheral and central immune responses to TLR4 stimulation are linked.15,20,33 Confirming these relationships would provide a foundation for better patient phenotyping and point to mechanistic targets for further investigation.

Following the MAPP EPS (2009-2015), a new cohort of urologic pelvic pain patients was recruited for the MAPP Research Network Symptom Phenotyping Study (SPS, 2015-2021).8 In the MAPP SPS, a modified protocol for measuring the LPS ex vivo cytokine/chemokine response was adopted across 6 recruiting sites, allowing for the opportunity to conduct a critical confirmatory analysis of the MAPP EPS findings.8 Clinically, MAPP studies have shown that COPCs/widespread pain in IC/BPS patients is associated with greater psychosocial difficulties and worse quality of life, indicating an urgent need to establish underlying pain mechanisms in this subset of patients.34 In this study, we examined 135 female IC/BPS patients from the MAPP SPS to determine whether ex vivo TLR4-stimulated cytokine/chemokine release, this time measured across a larger number of cytokines/chemokines, distinguished patients with comorbid COPCs from those with IC/BPS only. We also conducted analyses to determine whether the extent of widespread pain and experimental pain sensitivity were associated with this response. Our primary purpose was to determine whether we could confirm the relationship between the TLR4 ex vivo cytokine/chemokine response and characteristics of nociplastic pain.

2. Methods 2.1. Sample

The MAPP Research Network SPS enrolled 620 Urologic chronic Pelvic Pain Syndrome patients for longitudinal follow-up of symptoms and phenotypic characteristics (ClinicalTrials.gov Identifier: NCT02514265).8 Because of funding limitations, only a subset of the collected biomarker samples could be analyzed. A total of 155 female IC/BPS participants were selected for biomarker analysis; a power calculation was conducted before selecting the number of female participants based on previous data.47 The Cohen d effect size for the difference between IC/BPS only and IC/BPS + comorbid COPCs on the LPS-stimulated composite score (IL-1b + IL-6) was d = 0.67. Assuming 2-tailed hypothesis testing, alpha = 0.05, and an allocation ratio of 1:2 (pelvic pain only: pelvic pain comorbid, the rough distribution in the original article), 82 subjects would be required to detect the effect of interest with 0.80 power. We chose a larger number of samples for analysis, enough to detect an effect size of d = 0.50, because of the changes to the protocol (eg, whole blood rather than PBMCs and the larger number of cytokines/chemokines tested).

The 155 individuals were selected because they provided biomarker samples at baseline, 6 months, and 18 months with concurrent neuroimaging data. These longitudinal data will be analyzed as part of an ancillary R01 to the MAPP network (R01DK123164). As the primary purpose of this article is to attempt a conceptual validation of MAPP EPS findings, only female patients are analyzed who also completed the full battery of COPC self-report criteria, so the final sample consisted of 135 IC/BPS participants.

2.2. Demographic information

Patient demographics were collected by self-report, and body mass index was calculated from height and weight.

2.3. Clinical pain, chronic overlapping pain conditions, and extent of widespread pain

Overall pelvic pain severity (PPS) was calculated using a composite measure composed of questions from the genitourinary pain index (GUPI) and interstitial cystitis symptom index (ICSI), as described previously.21

Chronic overlapping pain conditions were assessed using the Complex Multi-Symptom Inventory and standardized diagnostic criteria.59 Rather than administering all diagnostic criteria to all patients, patients first complete the screener, which contains items that “trigger” full diagnostic criteria the administration of the full diagnostic criteria for COPCs that are relevant for that individual, which limits response burden by only administering relevant questionnaires. Possible diagnostic modules include chronic fatigue syndrome, irritable bowel syndrome, fibromyalgia, temporomandibular joint disorder, and migraine.16,19,40,60,61

Patients were asked to indicate whether they had pain and the severity of any pain on a 0 to 10 scale, using a 76-site body map adapted from the Collaborative Health Outcomes Information Registry project and further reduced to 12 nonpelvic regions for the assessment of widespread pain.45 Sites with severity of pain rated at least 4 were counted.

2.4. Cytokine/chemokines under ex vivo stimulated and unstimulated conditions 2.4.1. Differences between the multidisciplinary approach to the study of chronic pelvic pain Epidemiology Phenotyping Study and Symptom Phenotyping Study protocols

Several important differences are noted between the MAPP EPS and MAPP SPS TLR stimulation protocols. First, in the MAPP EPS, a single site (University of Iowa) collected samples for ex vivo stimulation, whereas in the MAPP SPS, all 6 recruiting sites took part. Second, in the MAPP EPS, PBMCs were isolated for stimulation, whereas in the MAPP SPS, a commercially available whole blood ex vivo stimulation assay was used (TruCulture, Myriad RBM) to allow all sites to collect samples in a consistent and efficient manner. In the MAPP EPS, LPS concentrations of 50 ng/mL were used to stimulate samples for 72 hours, whereas in the MAPP SPS, the concentration was 100 ng/mL (the standard concentration available) for an incubation period of 24 hours. For these reasons, this study represents a conceptual validation of the MAPP EPS findings, rather than a direct replication.

2.5. Symptom Phenotyping Study protocol

The TruCulture system uses vacutainers preloaded with TLR4 agonist (LPS) or control media (unstimulated condition), which are kept frozen at −20°C until being thawed for one hour at room temperature, or overnight in a standard refrigerator before use. Approximately 1 mL of whole blood is drawn directly into each tube and then kept in a tabletop incubator at 37°C for 24 hours. After incubation, the supernatant is isolated using a valve separator included in the kit and stored in −80°C freezer for batch analysis. All samples were sent to a central biorepository at the University of Denver Anscultz Medical Campus Biorepository Core Facility under the supervision of the Director and Codirector of the MAPP Tissue Analysis and Technology Core. Then, 50 µL of thawed supernatant was analyzed for 7 cytokine/chemokines using Luminex Xmap technology (Austin, TX) with R&D systems high-performance assays. These were monocyte chemoattractant protein-1 (MCP-1, range of assay: 3.0-1890 pg/mL), macrophage inflammatory protein 1-α (MIP-1α, 18-9840 pg/mL), IL-1β (0.34-1500 pg/mL), IL- 6 (0.95-3400, pg/mL), IL-8 (0.78-2800 pg/mL), IL-10 (0.46-2000 pg/mL), and tumor necrosis factor-α (TNF-α; 0.78-2000 pg/mL). Unstimulated samples for IL-1β, 6, 8, 10 and TNF-α were diluted 2x, whereas the LPS condition was diluted 20x. Unstimulated samples for MCP1 and MIP-1α were diluted 4x, and the LPS condition was diluted 10x. Values below the limit of quantification (LOQ) were set to one half of the LOQ value. This family of cytokines and chemokines was selected because they represent different aspects of the inflammatory response and are all promoted by the transcriptional factor NF-ΚB, whose upregulation is a well-established consequence of TLR4 stimulation.38

2.6. Pressure pain sensitivity

As in the MAPP EPS, pressure pain sensitivity was measured using the Multimodal Automated Sensory Testing (MAST) system (Arbor Medical Innovations, Ann Arbor, MI).24 The MAST system includes a electromechanical stimulator to deliver pressure stimuli and a touchscreen-based rating scale to capture participant responses.22 Following a familiarization procedure to reduce testing anxiety, an ascending sequence of incremental pressure stimuli were delivered to the participants' dominant thumbnail by a 1-cm2 rubber probe attached to the MAST stimulator. Pressure intensity started at 0.5 kgf/cm2 and increased in 0.5 kgf/cm2 steps at a ramp rate of 4.0 kgf/cm2/s. Each pressure was held constant for 5 second and was separated by a 20-second interstimulus rest interval. Participants rated perceived pain intensity after each stimulus using a digital 0 to 100 NRS displayed on the touchscreen (0 = no pain; 100 = pain as bad as you could imagine). The test was completed when the participant reached his or her pain tolerance and asked that the test be stopped, the participant reported a pain intensity of ≥80/100, or a maximum possible pressure intensity of 10 kgf/cm2 was delivered. Data were automatically uploaded by the system to the MAPP Network Data Coordinating Center through a secure file transfer protocol for analysis. A 3 parameter logistic model was used to estimate the within-person inflection point on the stimulus–response curve between PPT and tolerance, referred to as Pain50.23 To ensure standardization across sites, scripted participant instructions were used and research staff completed annual in-person training.

2.7. Statistical analyses

All analyses were performed in the R programming language, version 3.6.1.

2.7.1. Comparison of biomarker sample to the rest of the multidisciplinary approach to the study of chronic pelvic pain Symptom Phenotyping Study baseline sample

To determine whether the 135 participants differed from the rest of the female IC/BPS participants in the MAPP SPS baseline sample (n = 172) on variables that could plausibly influence immune parameters, we compared patient age, body mass index, depression scores, anxiety scores (Hamilton anxiety and depression scale), perceived stress (perceived stress scale), Pain50 scores, number of painful sites selected on the body map, and proportion of each sample with a COPC by t test and Χ2 tests for continuous and categorical variables, respectively.

2.7.2. Comparison of interstitial cystitis/bladder pain syndrome only and interstitial cystitis/bladder pain syndrome + chronic overlapping pain condition groups 2.7.2.1. Transformation of cytokine/chemokine values for analysis

To conduct parametric analyses with appropriate covariates, we proceeded to transform cytokine/chemokine values with values in the detectable range using Box–Cox transformations (all TLR4-stimulated cytokines/chemokines; unstimulated TNF-α, IL-1β, and IL-8).5 The Shapiro–Wilks test statistic, where values of >0.97 indicate acceptable normality, was used to evaluate these transformations and all exceeded this threshold.

2.7.2.2. Principal components analysis

We subsequently used a principal components analysis based on the correlation matrix between the 7 transformed cytokine/chemokine values under the TLR4-stimulated condition retaining components with an eigenvalue greater than one. Factor scores for the resulting components were then extracted by the regression method. Principal components analysis has previously been used as a dimension reduction technique for inflammatory variables.29,31

2.7.2.3. General linear models

The primary form of analysis was a mixed-effects linear model with the TLR4 component scores as dependent variables with patient age and body mass index included as covariates. A random intercept term was included for site of collection. The independent predictor of interest was COPC group status. We also conducted analyses with the degree of widespread pain, and Pain50 ratings as independent predictors of the TLR4 component scores.

We recreated an inflammatory composite variable analogous to that used in the original study, which was a simple mean of the z-scores for LPS-stimulated IL-1β and IL-6 from this study. These values were used rather than those from the MAPP EPS data set due to the substantial differences in the protocol.

Although not the focus of the current article, we also repeated these analyses with the unstimulated values of TNF-α, IL-8, MCP-1, and IL-1β (transformed) as dependent variables, to determine if unstimulated values were associated with COPC status. Because a large percentage of unstimulated IL-10 (84%), IL-6 (78%), and MIP-1α (56%) values fell below the detectable limit of the assay, we did not compare these cytokines between the groups.

3. Results

Demographic and clinical information is shown in Table 1. Values of stimulated and unstimulated cytokines/chemokines are shown in Table 2.

Table 1 - Urologic chronic pelvic pain syndrome symptoms and distribution of chronic overlapping pain condition types by chronic overlapping pain condition status. IC/BPS only (n = 36) IC/BPS + COPC (n = 99) All (n = 135) Mean SD Mean SD Mean SD Age, y 47.21 16.32 40.97 14.17 42.63 14.97 Body mass index 25.64 5.31 26.96 5.46 26.61 5.43 Age of symptom onset,y 32.03 15.86 27.64 14.66 28.82 15.06 Genitourinary pain severity 9.26 5.16 13.48 6.15 12.35 6.17 Urinary symptom severity 11.68 3.87 16.45 5.18 15.18 5.29 N % n % Chronic overlapping pain conditions  Fibromyalgia 0 0 10 10.1  Myalgic encephalomyelitis/chronic fatigue syndrome 0 0 25 25.3  Irritable bowel syndrome 0 0 57 57.6  Temporomandibular disorder 0 0 46 46.5  Migraine 0 0 53 53.5

COPC, chronic overlapping pain condition; IC/BPS, interstitial cystitis/bladder pain syndrome.

Table 2 - Unstimulated and lipopolysaccharide-stimulated cytokine values by chronic overlapping pain condition status. IC/BPS only (n = 36) IC/BPS + COPC (n = 99) All (n = 135) Median 25th-75th percentile Median 25th-75th percentile Median 25th-75th percentile Unstimulated (pg/mL)  monocyte chemoattractant protein-1 118 78-156 103 78-148 109 78-151  macrophage inflammatory protein 1-alpha* 58 36-136 36 36-108 36 36-112  Interleukin-1β 1 1-2 1 <1-3 1 <1-3  Interleukin-6* 1 1-1 1 1-1 1 1-1  Interleukin-8 37 23-68 37 37-79 37 37-77  Interleukin-10* <1 <1-<1 <1 <1-<1 <1 <1-<1  Tumor necrosis factor-α 4 3-5 5 5-6 4 4-6 Stimulated (pg/mL)  Monocyte chemoattractant protein-1 1182 896-1658 1715 1033-2542 1526 1000-2310  Macrophage inflammatory protein 1-alpha 38,497 23,818-58,503 48,184 37,793-75,389 45,923 33,951-72,730  Interleukin-1β 5427 4195-9035 8982 4686-14,479 8309 4316-8309  Interleukin-6 14,605 10,209-21,923 21,218 15,739-25,878 20,244 14,374-25,815  Interleukin-8 11,793 7197-15,349 14,812 9316-20,675 12,550 9004-19,368  Interleukin-10 53 24-83 68 41-107 65 37-98  Tumor necrosis factor-α 3807 2294-5648 4448 2899-5591 4123 2715-5591

*Greater than 25% of values below the limit of quantification.

COPC, chronic overlapping pain condition; IC/BPS, interstitial cystitis/bladder pain syndrome.

3.1. Comparison of biomarker sample to the rest of the baseline multidisciplinary approach to the study of chronic pelvic pain Symptom Phenotyping Study cohort

There were no significant differences between the MAPP biomarker sample and the rest of the MAPP SPS baseline cohort on patient age, BMI, depression scores, anxiety scores, perceived stress scores, Pain50 scores, number of painful sites selected on the body map, and proportion of each sample with a COPC (all P < 0.05; data not shown).

3.2. Principal components analysis

Two components were extracted, one with high positive loadings on all 7 cytokine/chemokines that explained 63.1% of the variance and a second component with high positive loadings for MCP1, IL-8, and IL-10, a near zero loading for MIP-1α, and negative loadings for IL-6, IL-1, and TNF-α, which explained an additional 16.1% of the variance. These findings are consistent with the first component representing a global response to LPS-stimulation, and a second component more specific to anti-inflammatory activity, the regulatory function of IL-8, and chemotactic activity (MCP1). See Figure 1 for principal components analysis loading plot.

Figure 1.:

Principal components analysis loading plots for the cytokines/chemokines (transformed scales) under the TLR4-stimulated condition. IL, interleukins; MCP-1, monocyte chemoattractant protein-1; MIP-1, macrophage inflammatory protein-1; TLR4, toll-like receptor 4; TNF-α, tumor necrosis factor-α.

3.3. Comparison of interstitial cystitis/bladder pain syndrome and interstitial cystitis/bladder pain syndrome + chronic overlapping pain condition groups on toll-like receptor 4 composite scores

In models controlling for patient age, body mass index, and site of collection, IC/BPS + COPC patients were found to have significantly elevated TLR4 global composite scores, (P < 0.01) but not the TLR4 anti-inflammatory/regulatory/chemotactic composite scores (P > 0.05). See Figure 2A and B for differences in the TLR4 global composite score and Cohen d effect sizes for each of the 7 stimulated cytokine/chemokines by COPC status. The basic difference in the TLR4 global composite score was apparent when stratified by site of collection, strengthening the generalizability of the results (Supplemental Fig. 1, available at https://links.lww.com/PAIN/B739). The z-score composite of ex vivo TLR4-stimulated IL-6 and IL-1β was also significantly higher in the IC/BPS + COPC group (P < 0.05). See Table 3 for model estimates.

Figure 2.:

(A) Ex vivo TLR4 global composite score for female IC/BPS patients stratified by presence (n = 99) or absence (n = 36) of COPCs. (B) Cohen d effect sizes for individual cytokines/chemokines on the transformed scale associated with IC/BPS only status. COPC, chronic overlapping pain condition; IC/BPS, interstitial cystitis/bladder pain syndrome; IL, interleukins; MCP-1, monocyte chemoattractant protein-1; MIP-1, macrophage inflammatory protein-1; TLR4, toll-like receptor 4; TNF-α, tumor necrosis factor-α.

Table 3 - General linear model estimates for relationship between pain variables and toll-like receptor 4–stimulated composite scores. Estimate SE df t value P TLR4 global composite score  Intercept −0.670 0.481 110.343 −1.393 0.167  Age (y) 0.003 0.006 130.289 0.451 0.653  BMI 0.003 0.015 129.671 0.177 0.860  COPC (yes) 0.608 0.188 128.939 3.232 0.002  Intercept −0.319 0.479 97.096 −0.667 0.506  Age (y) −0.001 0.006 129.277 −0.160 0.873  BMI 0.005 0.016 128.691 0.341 0.734  Body map sites 0.071 0.032 125.837 2.258 0.026  Intercept −0.070 0.553 113.500 −0.127 0.899  Age (y) −0.001 0.006 130.400 −0.112 0.911  BMI 0.008 0.016 129.700 0.516 0.607  Pain50 −0.048 0.069 129.600 −0.695 0.489  Intercept −0.574 0.518 111.100 −1.107 0.270  Age (y) 0.000 0.006 130.400 0.054 0.957  BMI 0.004 0.016 129.600 0.268 0.789  Pelvic pain severity 0.026 0.016 127.900 1.613 0.109 TLR4 anti-inflammatory, regulatory, chemotactic composite score  Intercept −1.089 0.494 114.403 −2.205 0.030  Age (y) 0.006 0.006 130.623 1.024 0.308  BMI 0.025 0.016 129.622 1.599 0.112  COPC (yes) 0.263 0.195 128.789 1.348 0.180  Intercept 0.962 0.474 114.361 −2.031 0.045  Age (y) 0.004 0.006 129.988 0.734 0.464  BMI 0.026 0.016 129.194 1.637 0.104  Body map sites 0.054 0.032 125.527 1.686 0.094  Intercept −1.845 0.520 124.280 −3.549 <0.001  Age (y) 0.004 0.006 128.286 0.792 0.430  BMI 0.030 0.015 130.722 2.007 0.047  Pain50 0.233 0.066 130.970 3.507 <0.001  Intercept −0.897 0.517 121.466 −1.734 0.085  Age (y) 0.005 0.006 130.977 0.784 0.434  BMI 0.028 0.016 129.973