|Home | About | Journals | Submit | Contact Us | Français|
Background.Racial disparities exist in gynecological diseases. Variations in Toll-like receptor (TLR) genes may alter signaling following microbial recognition.
Methods.We explored genotypic differences in 6 functional variants in 4 TLR genes (TLR1, TLR2, TLR4, TLR6) and the adaptor molecule TIRAP between 205 African American women and 51 white women with clinically suspected pelvic inflammatory disease (PID). A permutated P < .007 was used to assess significance. Associations between race and endometritis and/or upper genital tract infection (UGTI) were explored. Logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs).
Results.The TT genotype for TLR1 rs5743618, the GG genotype for TLR1 rs4833095, the CC genotype for TLR2 rs3804099, the TLR6 rs5743810 T allele, and the CC genotype for TIRAP rs8177374 significantly differed between races (P < .007). African American race was associated with endometritis and/or UGTI (OR, 4.2 [95% CI, 2.0–8.7]; P = .01). Among African Americans, the TLR6 rs5743810 T allele significantly decreased endometritis and/or UGTI (OR, 0.4 [95% CI, .2–.9]; P = .04). Additionally, rs5743618, rs4833095, and rs8177374 increased endometritis and/or UGTI, albeit not significantly.
Conclusions.Among women with PID, TLR variants that increase inflammation are associated with African American race and may mediate the relationship between race and endometritis and/or UGTI.
Pelvic inflammatory disease (PID), the infection and inflammation of the female upper genital tract, is a condition that can lead to infertility, chronic pelvic pain, and ectopic pregnancy . Microbes such as Neisseria gonorrhoeae, Chlamydia trachomatis, Mycoplasma genitalium, and aerobes and anaerobes commonly associated with bacterial vaginosis (BV) have been implicated in the etiology of PID . Race is a risk factor for PID, as well as chlamydia, gonorrhea, and BV [3–6]. Racial disparities exist in other inflammatory diseases including preterm birth, systemic lupus erythematosus, and cardiovascular disease [7–9]. The reasons for these racial differences are not completely understood. Socioeconomic, behavioral, and psychosocial factors only explain some of the racial disparities [10, 11]. Studies have found that African Americans are more likely to carry alleles that upregulate proinflammatory cytokines including interleukin 6, interleukin 1, and tumor necrosis factor (TNF) [12–14]. This may result in an imbalanced immune response, leading to excessive and harmful inflammation. As African Americans are at increased risk for inflammatory diseases, genetic variations that lead to increased inflammatory responses in the presence of infectious agents may also be more frequent in this group.
The innate immune system relies on pattern recognition receptors (PRRs) for microbial recognition. Toll-like receptors (TLRs) are a family of PRRs that are essential for microbial elimination through the induction of inflammatory cytokine and chemokine genes . The adaptor molecules Toll/interleukin-1 receptor domain–containing adaptor protein (TIRAP) and myeloid differentiation primary response protein 88 (MyD88) help to mediate TLR signaling. Examination of human tissues for expression of TLRs has revealed messenger RNA for TLRs 1–9 in uterine epithelium . TLR2 is highly expressed in fallopian tubes and the cervix , whereas TLR4 is weakly expressed in fallopian tubes , and transcripts of TLR4 and its adaptor MD-2 are not detected in human cervical epithelial cells . TLR2 forms homodimers with itself or heterodimers with TLR1 or TLR6 to bind pathogen ligands, whereas TLR4 requires the adaptor protein MD-2. TLR2 and TLR4 can recognize several bacteria associated with PID, including C. trachomatis, N. gonorrhoeae, and M. genitalium [18–20]. Although TLRs play a critical role in cytokine expression, few studies have examined racial differences in functional TLR polymorphisms. Barreiro et al suggests that variants in TLR genes may result in differential contributions to disease outcomes following infection in different populations . For example, after examining the evolution of TLRs and their contribution to host defense, the authors found that Europeans had a diminished TLR2:TLR1-mediated immune response reducing inflammation. Thus, the authors concluded that those who carry certain alleles may avoid excessive inflammation, possibly inferring a selective advantage .
We have previously found that TLR1 variants are associated with C. trachomatis, upper genital tract infection, and reduced pregnancy rates among African American women with PID . As racial disparities in reproductive and infectious diseases exist and TLR genes are involved in immunopathology, our objective was to determine the racial differences in the genotypic distribution of functional TLR variants among women with clinically suspected PID. We hypothesized that African American women would be more likely to carry TLR variants that are reported to increase inflammatory responses compared to white women. In addition, we examined the associations between race, TLR single-nucleotide polymorphisms (SNPs), and upper genital tract infection and inflammation.
This study utilized data from the PID Evaluation and Clinical Health (PEACH) study, a randomized clinical trial to compare inpatient and outpatient treatment in preventing long-term complications among 831 women with mild to moderate PID. The methods of subject recruitment, data collection, and follow-up have been reported elsewhere . In brief, between March 1996 and February 1999 women aged 14–37 years who had clinically suspected PID—defined as a history of pelvic discomfort for <30 days, findings of pelvic organ tenderness (uterine or adnexal) on bimanual examination, and leukorrhea and/or mucopurulent cervicitis and/or untreated but documented gonococcal or chlamydial cervicitis—and gave informed consent were eligible for the PEACH study. The University of Pittsburgh Institutional Review Board approved the study.
Our substudy population consists of 205 women of self-reported non-Hispanic African ancestry, 51 of non-Hispanic white ancestry, and 34 women of “other” ancestry. These women were included in a previous PEACH substudy that tested the association between TLR variants and chlamydial PID among African American women . All women who had data on C. trachomatis infection and who had previously stored buffy coats (n = 237) or serum samples (n = 50) available were genotyped for TLR and adaptor molecule SNPs. There were no significant differences in the genotype frequencies between women genotyped with buffy coats or serum samples. This population represents the original PEACH study in which 74.7% of participants are African American, 16.0% are white, and 9.2% are of other races. We excluded women of “other” (n = 34) ancestry from the statistical analyses because the group was composed of several racial groups and was small in size.
As part of the PEACH protocol, gynecological examinations were performed at baseline and 5 and 30 days after treatment. During these examinations, vaginal smears were Gram stained for BV in a central laboratory by standardized methods described by Nugent et al . Endometrial biopsy specimens were obtained for histological examination. Additionally, chlamydial polymerase chain reaction (PCR) and gonococcal culture were performed. Histological endometritis was based on a modification of the criteria proposed by Kiviat et al  and defined as the presence of at least 5 neutrophils in the endometrial surface epithelium in the absence of menstrual endometrium and/or at least 2 plasma cells in the endometrial stroma. Upper genital tract infection (UGTI) was defined as a positive endometrial C. trachomatis PCR or endometrial N. gonorrhoeae culture. Cervical and endometrial specimens were also stored and later used to test for M. genitalium, using a microwell plate–based PCR assay (MgPa-IMW) targeting the MgPa gene.
For this study, functional SNPs (those which introduce an amino acid change), including 2 from TLR1 (rs5743618, rs4833095), 1 from TLR2 (rs3804099), 1 from TLR6 (rs5743810), 1 from TLR4 (rs4986790), and 1 from TIRAP (rs8177374), were analyzed. All SNPs were chosen from HapMap (http://hapmap.ncbi.nlm.nih.gov/) and were genotyped by fluorescence polarization . PCR reactions included 2.5 µL of 1× PCR buffer (Invitrogen), 1 µL of magnesium chloride, 4 µL of dNTPs, 1.5 µL of each primer, 0.1 µL of Taq polymerase (Invitrogen), and 13.4 µL of distilled water, for a total volume of 25 µL. Amplification was performed using a Peltier Thermal Cycler (MJ Research). Thermal cycling conditions were 95°C for 3 minutes, then 35 cycles of 95°C for 30 seconds for denaturing, 55°C for 30 seconds for annealing, and 72°C for 30 seconds for extension, and a final extension step of 72°C for 1 minute. PCR products were resolved by electrophoresis in a 3% agarose gel and visualized under ultraviolet light after ethidium bromide staining. Genotypes were assigned by direct comparison to controls of sequence-confirmed genotypes, and a 5% random resample was included for consistency of the genotyping. All SNPs were tested for deviations from Hardy–Weinberg equilibrium and allele frequencies in our cohort were compared to those reported on HapMap.org from a population of African ancestry in the southwestern United States, a white group (Utah residents with ancestry from Northern and Western Europe), and a healthy African American population from Pittsburgh, Pennsylvania. No genotyping errors were detected.
We compared demographic, clinical, and behavioral characteristics, obtained from the baseline interview and examination, between African American and white women using χ2 test, Fisher exact test, or analysis of variance. These variables included age, marital status, education, employment, insurance, temperature, white blood cell count, C-reactive protein, bilateral adnexal tenderness, cervicitis, time to treatment, C. trachomatis, N. gonorrhoeae, BV, M. genitalium, sexual history, smoking, drug use, and douching. Our primary analyses compared genotype frequencies for each SNP between racial groups. We aimed to use an additive model unless the cell size was too small for logistic regression (<5), and then genotypes were combined to run a dominant model. To correct for multiple testing, we performed 1000 permutation tests across all SNPs to generate an empirical null distribution and determined that an observed P value < .007 would be significant (at empirical α ≤ .05). Clinically suspected PID has low specificity and may include women without “true” upper genital tract infection and inflammation. Endometritis is an acceptable alternative to salpingitis for the diagnosis of PID. However, it is a spotty disease (endometrium may look normal in some areas even though disease is present) and biopsy may miss some cases of upper genital tract infection and inflammation. Therefore, to examine upper genital infection and inflammation, we examined the association between racial groups and histologically confirmed endometritis and/or chlamydial and/or gonococcal UGTI. Last, we stratified by race and examined the association between the TLR SNPs and endometritis and/or UGTI. Logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs). Models examining race and endometritis and/or UGTI were adjusted for chlamydia, gonorrhea, and BV. These variables were chosen a priori. In addition, we considered baseline variables that differed between our groups as possible confounders. However, these variables were not associated with our outcome and adding them to our model did not change our estimate by more than 10%. Therefore, they were not included in the final model. All analyses were performed using SAS/Genetics version 9.1.3.
Important demographic characteristics including age, education, insurance, and unemployment did not differ between white and African American women (Table (Table1).1). White participants were more likely than African American participants to be married (20.4% vs 6.9%; P = .004), but also more likely to report a recent new sexual partner (17.7% vs 8.3%; P = .048). White participants were also less likely to report condom use in the past 4 weeks (43.5% vs 60.6%; P = .037). There was no difference in the number of lifetime sexual partners between groups. African Americans were more likely than whites to have an elevated temperature (9.9% vs 0.0%; P = .012) and were also slightly more likely to have elevated C-reactive protein levels (50.0% vs 25.0%; P = .063). African Americans were more likely to have endometrial and/or cervical C. trachomatis (46.0% vs 31.3%; P = .042), endometrial and/or cervical N. gonorrhoeae (34.9% vs 4.4%; P < .0001), and BV (66.2% vs 29.8%; P < .0001), but there was no difference in M. genitalium infection. There was a striking difference in the number of African Americans with endometrial C. trachomatis (19.5% vs 4.1%) and endometrial N. gonorrhoeae (21.3% vs 2.2%) compared to whites. Behavioral characteristics were similar between groups, although African Americans were less likely than whites to smoke (43.4% vs 66.7%; P = .003).
Significant differences were found in the genotype distribution for variants in the TLR1, TLR2, TLR6, and TIRAP genes (Table (Table2).2). African American women were significantly more likely than white women to carry the TLR1 rs5743618 TT genotype (OR, 9.8 [95% CI, 4.4–22.1]; P < .0001) and the rs4833095 GG genotype (OR, 8.6 [95% CI, 3.7–20.3]; P < .0001). Results were similar for the TLR2 rs3804099 CC genotype (OR, 3.9 [95% CI, 1.5–9.8]; P = .0033) and the TIRAP rs8177374 CC genotype (OR, 7.1 [95% CI, 2.9–17.5]; P < .0001). In contrast, African American women were significantly less likely to carry 1 or 2 of the TLR6 rs5743810 T alleles (OR, 0.1 [95% CI, .1–.3]; P < .0001). No significant differences were found for the TLR4 rs4986790 TT genotype (P = .2402).
Logistic regression was used to examine the association between racial groups and confirmed endometritis and/or UGTI. African American women were significantly more likely to have endometritis and/or UGTI (OR, 5.9 [95% CI, 2.8–12.5]; P < .0001). Results were significant when UGTI (OR, 8.9 [95% CI, 2.6–30.0]; P = .0004) and endometritis (OR, 3.9 [95% CI, 3.9–8.2]; P = .0003) were considered separately. After adjustments for C. trachomatis, N. gonorrhoeae, and BV, results were still significant (adjusted OR, 3.3 [95% CI, 1.3–8.2]; P = .0114]. Although our power was limited, we conducted an exploratory analysis to examine SNP associations with endometritis and/or UGTI, stratified by race (Table (Table3).3). Among African Americans, women who carried 1 or 2 of the TLR6 rs5743810 T alleles had decreased odds of endometritis and/or UGTI (OR, 0.4 [95% CI, .2–.9]; P = .0447). There was a similar trend among white participants, although the model was not statistically significant (OR, 0.5 [95% CI, .1–2.1]; P = .4162). Although results were nonsignificant, African Americans who carried the TLR1 rs5743618 TT genotype (OR, 1.9 [95% CI, .9–4.1]; P = .1536), the rs4833095 GG genotype (OR, 1.8 [95% CI, .9–3.5]; P = .1776) and the TIRAP rs8177374 CC genotype had increased odds of endometritis and/or UGTI (OR, 3.4 [95% CI, .8–14.3]; P = .2468). Results were similar among white women and similar when endometritis and UGTI are considered separately. However, white women who carried the TLR2 variant rs3804099 had increased odds of endometritis and/or UGTI (OR, 3.9 [95% CI, .8–17.7]; P = .0669) compared to women who carried the CT or TT genotypes.
We found that African American women had increased odds of upper genital tract infection and inflammation compared to whites. These results are similar to those of Hillier et al, who found that nonwhite race was associated with endometritis after adjustments for BV, gonorrhea, and chlamydia in 178 women with clinical PID (adjusted OR, 2.3 [95% CI, 1.1–4.8]) . Trends were observed for TLR variants as well. Women who carried the T allele for TLR6 SNP rs5743810 had decreased odds of endometritis and/or UGTI. In addition, trends toward increased odds of endometritis and/or UGTI were found for women carrying TLR1 and TIRAP SNPs. Our sample size did limit our power. In addition, endometritis is a spotty disease and may not indicate all cases of upper genital tract inflammation . Furthermore, PID is polymicrobial and women with endometritis could have a variety of microorganisms that could induce a different set of cytokines when recognized by TLRs. Studies have shown that cytokine profiles of genital tract microbes are heterogeneous [30, 31]. We have previously found that functional TLR1 SNP rs5743618 was associated with chlamydial infection, UGTI, and reduced pregnancy in African American women . This may suggest that TLR1 rs5743618 mediates the relationship between African American race and upper genital tract infection.
Our results show that compared to white women, African American women were more likely to carry TLR variants that may increase signaling and less likely to carry a TLR6 variant that may decrease signaling. Hawn et al reported that the T allele for TLR1 variant rs5743618 expressed significantly greater NF-κB signaling in transfected HEK293 cells compared to the G allele . The TLR1 SNP rs4833095 A allele was found to impair the TLR response to Pam3CSK4 (P < .05) . In a population-based case-control study among 1312 tuberculosis patients and controls, the G allele significantly increased the risk of tuberculosis in African Americans (P = .009) . It should be mentioned that rs5743618 and rs4833095 are reported to be in strong linkage disequilibrium with each other as well as with TLR1 SNP rs5743551. These SNPs may be measuring the same association. In a study of 410 Chinese patients with major trauma, the TLR2 rs3804099 C allele was associated with increased production of interleukin 10 (P = .002), interleukin 8 (P = .004), and TNF-α (P = .005) compared to the wild T allele . TIRAP SNP rs8177374 is thought to attenuate TLR2 signal transduction and it is suggested that homozygotes have an overactive response, resulting in more severe outcomes . These studies suggest that TLR SNPs that result in an amino acid change may alter the function of TLR signaling and downstream cytokine responses. However, little is known about the function of these SNPs in the genital tract and in different populations.
Our data were obtained from a large, multicenter, prospective, randomized clinical trial, with comprehensive demographic, clinical, and obstetric measurements. Although not all women in the PEACH study had blood samples available for analyses, important demographic and clinical characteristics between women with and without blood samples did not differ. This is the first study to our knowledge to examine the racial differences in the distribution of functional TLR SNPs among women with PID. However, our sample size limited our power and only a few selected functional SNPs were included in our analyses. Thus, other TLR SNPs should be further explored. As we relied on self-reported race, admixture is possible even after stratification. Although the use of nucleic acid amplification testing for the identification of C. trachomatis was available at the time of the PEACH study, this method was not available for the detection of N. gonorrhoeae. We relied on culture, which is a less sensitive method and may have missed some women who had upper genital tract gonococcal infection and misclassified them as UGTI negative. This could bias our results toward the null when we examined associations with endometritis and/or UGTI. However, we expect that some women misclassified as UGTI negative would have been diagnosed with endometritis and ultimately included in the correct category for our analysis.
Compared to white women, African American women with clinically suspected PID were more likely to carry variants in the TLR1, TLR2, and TIRAP genes, which may increase signaling and were more likely to have upper genital tract infection and inflammation. In contrast, African Americans were less likely to carry a TLR6 variant, which may decrease signaling. Women who carried the TLR6 variant were less likely to have upper genital tract infection and inflammation. Our results need to be replicated in a larger cohort that includes a greater number of white women. However, this study suggests that variations in TLR genes that lead to increased cytokine expression may mediate the association between African American race and upper genital tract infection and inflammation.
Financial support.This work was supported by the Agency for Healthcare Research and Quality (grant number HS08358-05 to R. N.) and the National Institute of Allergy and Infectious Diseases (grant number AI084024 and grant number U19 A1084024 to T. D.).
Potential conflicts of interest.All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.