Chlamydia trachomatis is an important cause of female infertility worldwide.1,2 The sequelae of chlamydial infection are likely due to immunopathologically mediated events in which both the chlamydial 60-kd heat shock protein (Hsp-60) and genetic predisposition play roles. For instance, in prior studies of pelvic inflammatory disease (PID), infertility, and trachoma due to C trachomatis infection, risk factors included antibody to chlamydial Hsp-60 and specific human leukocyte antigens (HLAs).3–6 Recently, polymorphisms in the promoter regions for tumor necrosis factor α and interleukin 10 were associated with trachoma, a scarring disease due to ocular infection with C trachomatis.7,8 Genetic polymorphisms for other immunoregulatory genes (eg, interferon γ, transforming growth factor β1, and interleukin 6) have been shown to influence progression of certain infectious and noninfectious diseases9–11 and, therefore, could potentially effect susceptibility to C trachomatis infection and disease.
Human leukocyte antigen class II molecules (ie, DR, DQ, and DP) present peptides to CD4 T cells and restrict the range of cellular and humoral responses to antigens.12 These alleles may present different chlamydial or host-derived peptides that evoke damaging, protective, or regulatory immune responses by CD4 T lymphocytes.13 In a previous study we found three HLA DQ alleles (DQA*0101, DQA*0102, and DQB*0501) to be associated with an altered risk for C trachomatis–associated tubal factor infertility.3 Microimmunofluorescence antibody to C trachomatis served as evidence of C trachomatis infection. However, because of the large number of HLA alleles tested in this population, significant findings may have arisen through chance, or because a closely linked allele of another immunoregulatory gene such as DRB1, DRB5, or cytokine genes was responsible for the findings. As an example, DRB1*1503 is closely linked with DQA*0602 and in our prior study was detected in none of 37 C trachomatis seropositive women, in comparison with five of 56 (9%) C trachomatis seronegative women (P = .15) (unpublished data, C. R. Cohen).
We designed the current study to confirm the association between C trachomatis microimmunofluorescence antibody and HLA DQA*0101, DQA*0102, and DQB*0501 among women with tubal infertility; to test for associations with DRB genotypes that are known to be closely linked with these DQ alleles; and to examine for cytokine gene polymorphisms.
MATERIALS AND METHODS
Written informed consent was obtained from each participant in the study. The Institutional Review Board for Human Subjects at the University of Washington and the Human Subjects Review Board at Kenyatta National Hospital, Nairobi, Kenya approved the protocol. In addition, human experimentation guidelines of the United States Department of Health and Human Services were followed.
Between October 1998 and November 1999, women presenting to the gynecology clinic at Kenyatta National Hospital in Nairobi who had a history of at least 12 months of infertility and regular menses, distal tubal occlusion on hysterosalpingogram, and a sexual partner with normal semen analysis were recruited. After informed consent was obtained, women were administered a questionnaire that included sexual, infertility, and sexually transmitted disease information. Blood was obtained for C trachomatis microimmunofluorescence serology and collection of peripheral blood mononuclear cells for molecular HLA class II genotyping and for immunogenetic analysis of cytokine loci.
Antibody against C trachomatis was measured using the microimmunofluorescence assay of Wang and Grayston.14 Women with a serovar-specified immunoglobulin (Ig) M or IgG titer of at least 1:16 were considered positive. Deoxyribonucleic acid (DNA) was isolated from peripheral blood mononuclear cells by an automatic extractor (model 340 A, Applied Biosystems, Foster City, CA) for HLA class II typing as performed by a reverse dot blot sequence–specific oligonucleotide method. The second exons of DQA1, DQB1, DRB1, DRB3, DRB4, and DRB5 were amplified using polymerase chain reaction (PCR)12 and were labeled by incorporation of digoxigenin-labeled deoxyuridine triphosphate during the PCR. Labeled PCR products were hybridized to allele-specific probes selected from the second exon of DQA1, DQB1, DRB1, DRB3, and DRB4 that were immobilized on nylon membranes. Positive reactions were visualized by a color precipitation reaction.
Cytokine polymorphism genotyping for transforming growth factor-β1, tumor necrosis factor α, interleukin 6, interleukin 10, and interferon γ was performed on extracted DNA by PCR using the Cytokine Genotyping Tray (One Lambda Inc., Canoga Park, CA). Briefly, preoptimized primers, approximately 19 μL of purified DNA, 1 μL of Taq polymerase, and specially formulated deoxynucleoside triphosphate buffer mix underwent PCR along with negative and positive controls following the manufacturer's instructions. Ten microliters of each PCR reaction was transferred to a 2.5% agarose gel and electrophoresed until the red tracking dye had migrated 0.5 cm into the gel (about 3–5 minutes). Results were interpreted using the worksheets provided with each tray.
We used our published results as assumptions to calculate the sample size for the current investigation.3 With an α of .05, 90% power, and a 68% prevalence for DQA*0102 in the chlamydia seronegative women, we needed to recruit 70 women to demonstrate an odds ratio (OR) of 0.18 in C trachomatis seropositive women compared with C trachomatis seronegative women; a similar number was required to demonstrate an OR of 6.65 assuming a 14% prevalence for DQB*0501 in the chlamydia seronegative women, whereas 88 women were required to show an OR of 5.0 assuming an 18% prevalence for DQA*0101 in the chlamydia seronegative women. Data were analyzed using SPSS-for-Windows 9.0 (SPSS Inc., Chicago, IL). Univariate analyses used χ2 and Fisher exact tests for comparison of categoric data and Mann–Whitney and Student t tests for continuous variables. In the presence of zero values, Haldane's modification of Woolf's formula was used to calculate the OR and 95% confidence interval (CI).15 Because we limited our analysis to the previously described DQ alleles and their linked DR alleles, correction for multiple comparisons was not performed.
Of the 70 women with tubal factor infertility recruited, 35 (50%) were seropositive for C trachomatis: 33 (47%) had IgG and two (3%) had only IgM. Table 1 presents selected demographic, sexual, and sexually transmitted infection data among 31 chlamydia seropositive and 32 seronegative cases; data were missing for four C trachomatis seropositive and three C trachomatis seronegative women. Chlamydia trachomatis seropositive and seronegative women with tubal factor infertility had similar mean ages (28.4 years versus 27.7 years, P = .50), marital status (P = .08), and median lifetime number of sexual partners (three versus three, P = .18). Chlamydia seropositive women reported an earlier age of first sexual intercourse (16.6 years versus 18.8 years, P < .001) and an increased median number of prior pregnancies (one versus zero, P < .04) relative to chlamydia seronegative women. A history of PID was reported by 19 (30%) women but was not statistically more common in C trachomatis seropositive women (Table 1). Because of statistical considerations, only HLA class II alleles found in 10% or more of participants were analyzed (Table 2). Our current investigation failed to demonstrate an association between DQA*0102 (13 [37%] versus 19 [54%]; OR 0.5; 95% CI 0.2, 1.4), DQB*0501 (14 [40%] versus eight [23%]; OR 2.3; 95% CI 0.7, 7.3), and DQA*0101 (15 [43%] versus 13 [37%]; OR 1.3; 95% CI 0.4, 3.7) in C trachomatis seropositive and seronegative women with tubal factor infertility (Table 2). However, in the 35 C trachomatis seronegative women with tubal factor infertility, HLA-DR1*1503 was detected in seven (20%) women, whereas it was absent from the 35 C trachomatis seropositive women (OR 0.05; 95% CI 0, 0.7). In addition, DRB5*0101, an allele in linkage disequilibrium with DRB1*1503, was also less commonly detected in C trachomatis seropositive women (6%, versus 26% in C trachomatis seronegative women; OR 0.2; 95% CI 0.02, 1.0).
Polymorphisms for tumor necrosis factor α-308 promoter; transforming growth factor β1-10 and -25 codons; interleukin 10-1082, -819, and -592 promoters; interleukin 6-174 promoter; and interferon γ+872 intron are described in Table 3. The three tumor necrosis factor α-308 polymorphisms were observed in the Kenyan population, with the majority (85%) exhibiting the -308G/G genotype associated with lower production of tumor necrosis factor α. Five of the eight tested transforming growth factor β1 alleles were found, with 60% and 30% of genotypes associated with high and intermediate production of transforming growth factor β1, respectively. All six interleukin 10-1082, -819, and -592 alleles were detected. Most subjects (60%) had an interleukin 6 genotype consistent with intermediate production, whereas 40% had a genotype associated with low production of interleukin 6. The three genotypes in codon 1 for interferon γ were detected, with the majority (61%) exhibiting a genotype associated with low production of this cytokine. Overall, we did not find any statistically significant differences between C trachomatis seropositive and seronegative subjects in genotype or allele frequencies for the cytokine polymorphisms studied (Table 3). Furthermore, in univariate analysis DRB1*1503 was not associated with any cytokine allelic variation (data not shown).
Our former study demonstrated that DQA*0102 was found less commonly among women with C trachomatis–associated tubal infertility.3 Because statistical significance of this finding may have arisen by chance because of multiple comparisons, we set forth to confirm this result and test for other alleles linked to DQA*0102 at the DQB locus (ie, DQB*0602, DQB*0604) and DRB1 and B5 loci (DRB1*1503, DRB5*0101). In particular, we were interested in DRB1*1503 because in the original study this allele was undetected in C trachomatis seropositive women (unpublished data, C. R. Cohen et al), is known to be closely linked with DQA*0602, and has been commonly identified in the Kenyan population (DRB1*1503 was detected in 37 of 165 [22%] sex workers in Nairobi16).
Although we failed to confirm our earlier observation of an association between DQA*0102 and decreased odds of C trachomatis microimmunofluorescence antibody in women with tubal infertility, we did observe that alleles linked to DQA*0102 at the DRB1 and B5 loci (DRB1*1503 and DRB5*0101) were found less commonly among C trachomatis–associated infertility cases than among infertile women without microimmunofluorescence antibody to C trachomatis. In fact, if we combine the HLA class II genotype results observed in this study with those previously reported,3 the statistical significance with DRB1*1503 is even further enhanced (chlamydia microimmunofluorescence seropositive, 0%, versus chlamydia microimmunofluorescence seronegative, 16%; P < .001). Therefore, the association of DQA*0102 we reported earlier is most likely explained through linkage disequilibrium with DRB1*1503. DRB1*1503 is known to be in linkage with DRB5*0101, and thus the statistical association of DRB5*0101 with microimmunofluorescence seronegative tubal infertility is also likely due to these two alleles existing as a haplotype.
The negative correlation of DRB1*1503 and DRB5*0101 with C trachomatis microimmunofluorescence antibodies in women with tubal infertility might be explained by one of several mechanisms. It is possible that this HLA class II allele is associated with an innate or acquired immune response that 1) reduces the risk of C trachomatis infection, 2) reduces the risk of C trachomatis–associated tubal scarring, 3) increases the risk of subjects with prior chlamydial infection being misclassified as microimmunofluorescence seronegative, or 4) increases the risk of tubal infertility caused by another factor that causes tubal infertility, such as Neisseria gonorrhoeae. However, this last point appears unlikely because DRB1*1503 was not enriched in the control group in comparison with an earlier study of women in Nairobi.16 Although this relationship could not be explained through linkage disequilibrium with the cytokine gene polymorphisms that we studied, we cannot exclude the possibility that another closely associated immunoregulatory gene might be responsible for our finding. As well, because DRB1*1503 and DRB5*0101 are themselves so closely linked, it will be difficult to separately evaluate their independent effects on susceptibility to C trachomatis infection and tubal damage.
Researchers in The Gambia found specific genotypes of the interleukin 10 and tumor necrosis factor α promoter sequences that may be correlated with an increased risk of trachomatous conjunctival scarring.7 We were unable to confirm a similar association with C trachomatis–associated tubal infertility in this population of women. Their reported association of a specific genotype of the interleukin 10 promoter was restricted to a single ethnic group. Thus our combined observations could suggest linkage to unmeasured polymorphorisms in other genes near the interleukin 10 locus. Similarly, we did not find any statistical relationship between C trachomatis serostatus and genes encoding for polymorphisms in transforming growth factor β1, interleukin 6, or interferon γ. For most of the candidate polymorphisms a specific genotype was predominant, thereby limiting our power to detect significant differences. In addition, because our control group also had evidence of tubal scarring most probably caused by an infectious cause other than C trachomatis infection, our control group might not have been the most suitable. It is plausible that certain polymorphisms might select for women with tubal infertility due to causes other than C trachomatis infection, thereby reducing our ability to find differences in cytokine polymorphisms between women with and without evidence of C trachomatis infection. It will be important in future studies to include other control groups such as fertile women without antibody to C trachomatis. Our decision to only evaluate women with tubal infertility was predicated on the association between specific DQ alleles and C trachomatis antibodies found in women with tubal infertility but not found in a control group of women undergoing tubal ligation. We felt that this dichotomy may have occurred as a consequence of an effect of HLA class II phenotype on the immunopathogenesis of chlamydia-associated tubal scarring.
The current study suggests that our previously published associations between DQA*0101, DQA*0102, and DQB*0501 and microimmunofluorescence antibody to C trachomatis in women with tubal infertility likely represented linkage disequilibrium with other HLA genes such as DRB1*1503/DRB5*0101. A future investigation should be conducted to test whether the DRB1*1503/DRB5*0101 haplotype is associated with protection against C trachomatis infection in a general Kenyan population. We hypothesize that individuals with DRB1*1503/DRB5*0101 may resist C trachomatis disease and/or clear cervical infection without inflammatory pathology. In addition, further genetic analyses of other loci involved in innate and adaptive immunity are necessary to fully elucidate the immunogenetic determinants of C trachomatis immunopathology.
1. De Muylder X, Laga M, Tennstedt C, Van Dyck E, Aelbers GN, Piot P. The role of Neisseria gonorrhoeae
and Chlamydia trachomatis
in pelvic inflammatory disease and its sequelae in Zimbabwe. J Infect Dis 1990;162:501–5.
2. Ville Y, Leruez M, Glowaczower E, Robertson JN, Ward ME. The role of Chlamydia trachomatis
and Neisseria gonorrhoeae
in the etiology of ectopic pregnancy in Gabon. Br J Obstet Gynaecol 1991;98:1260–6.
3. Cohen CR, Sinei S, Bukusi E, Bwayo J, Holmes K, Brunham R. Human leukocyte antigen class II DQ alleles associated with Chlamydia trachomatis
tubal infertility. Obstet Gynecol 2000;95:72–7.
4. Kimani J, Maclean IW, Bwayo JJ, MacDonald K, Oyugi J, Maitha GM, et al. Risk factors for Chlamydia trachomatis
pelvic inflammatory disease among sex workers in Nairobi, Kenya. J Infect Dis 1996;173:1437–44.
5. Lichtenwalner AB, Patton DL, Cosgrove Sweeney YT, Gaur LK, Stamm WE. Evidence of genetic susceptibility to Chlamydia trachomatis
-induced pelvic inflammatory disease in the pig-tailed macaque. Infect Immun 1997;65:2250–3.
6. Toye B, Laferriere C, Claman P, Jessamine P, Peeling R. Association between antibody to the chlamydial heat shock protein and tubal infertility. J Infect Dis 1993;168:1236–40.
7. Conway DJ, Holland MJ, Campbell AE, Bailey RL, Krausa P, Peeling RW, et al. HLA class I and II polymorphisms and trachomatous scarring in a Chlamydia trachomatis
–endemic population. J Infect Dis 1996;174:643–6.
8. Mozzato-Chamay N, Mahdi OS, Jallow O, Mabey DC, Bailey RL, Conway DJ. Polymorphisms in candidate genes and risk of scarring trachoma in a Chlamydia trachomatis
-endemic population. J Infect Dis 2000;182:1545–8.
9. Khani-Hanjani A, Lacaille D, Hoar D, Chalmers A, Horsman D, Anderson M, et al. Association between dinucleotide repeat in non-coding region of interferon-gamma gene and susceptibility to, and severity of, rheumatoid arthritis. Lancet 2000;356:820–5.
10. Powell EE, Edwards-Smith CJ, Hay JL, Clouston AD, Crawford DH, Shorthouse C, et al. Host genetic factors influence disease progression in chronic hepatitis C. Hepatology 2000;31:828–33.
11. Pantelidis P, Fanning GC, Wells AU, Welsh KI, Du Bois RM. Analysis of tumor necrosis factor-alpha, lymphotoxin-alpha, tumor necrosis factor receptor II, and interleukin-6 polymorphisms in patients with idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2001;163:1432–6.
12. Kwok WW, Nepom GT, Raymond FC. HLA-DQ polymorphisms are highly selective for peptide binding interactions. J Immunol 1995;155:2468–76.
13. Cohen CR, Brunham RC. Pathogenesis of chlamydia-induced pelvic inflammatory disease [review]. Sex Transm Infect 1999;75:21–4.
14. Wang S, Grayston JT. Human serology in Chlamydia trachomatis
infection with microimmunofluorescense. J Infect Dis 1974;130:388–97.
15. Woolf B. On estimating the relation between blood group and disease. Ann Hum Genet 1955;19:241–3.
© 2003 The American College of Obstetricians and Gynecologists
16. Dunand VA, Ng CM, Wade JA, Bwayo J, Plummer FA, MacDonald KS. HLA-DR 52- and 51-associated DRB1 alleles in Kenya, east Africa. Tissue Antigens 1997;49:397–402.