Serological Investigation of Chlamydia trachomatis Heat Shock Protein 10

Abstract

The humoral immune response to Chlamydia trachomatis 10-kDa heat shock protein (Chsp10) in populations of Russian and French origin was studied by using a recombinant Chsp10 enzyme-linked immunosorbent assay. A physiological but not a serological correlation of Chsp10 exposure with Chsp60 exposure was observed in the Russian population. In the French population studied, there was a significant association between detection of anti-r-Chsp10 immunoglobulin G (IgG) antibodies and chronic genital tract infections. Chsp10 residues 50 to 67 were found to contain an immunodominant although not universal B epitope. Cross-reactions with Chlamydia pneumoniae or Escherichia coli GroES protein are limited but may occur. Our study suggests that detection of anti-Chsp10 IgG antibodies is associated with chronicity of C. trachomatis genital tract infection and does not parallel that of anti-Chsp60 IgG antibodies.

Heat shock proteins, referred to as chaperonin 60 and chaperonin 10, belong to families of widely conserved proteins found in procaryotes and eucaryotes and facilitate the proper folding of numerous other proteins (5). Bacterial heat shock proteins are commonly immunodominant antigens recognized following infection and have been shown to elicit antibody and T-cell responses. The 60-kDa heat shock proteins (Hsp60s) of bacterial pathogens are often implicated in autoimmune inflammatory damage, resulting from molecular mimicry of their human homologs (9). Chlamydia trachomatis 60-kDa heat shock protein (Chsp60) has been associated with the pathogenesis of C. trachomatis-associated ectopic pregnancy and tubal infertility (3, 15, 16). The stress response in Chlamydia reticulate bodies is characterized by Chsp60 induction and by reduction in major outer membrane protein and lipopolysaccharide levels, as shown in an in vitro model of persistent infection (1). This stress response is believed to interrupt the normal progression of reticulate bodies to infectious elementary bodies, resulting in a longer-term persistent infection. Such persistent infections may serve as antigenic reservoirs for potentially immunopathogenic anti-Hsp immune system responses (2).

The C. trachomatis serovar A hyp operon has been cloned and found to be homologous to the groE stress response operon of Escherichia coli, with the hypA and hypB genes (encoding Chsp10 and Chsp60, respectively) being cotranscribed (10). More recently, the hypA gene of C. trachomatis serovar E has been cloned and found to be closely homologous to Hsp10s of other chlamydiae (7). We used purified recombinant Chsp10 to study the association between the immune response to Chsp10 and Chsp60 and the contribution of Chsp10 to the humoral immune response in different population groups.

MATERIALS AND METHODS

Study populations.

A total of 173 women attending the Department of Gynecology and Obstetrics, Leningrad Regional Hospital, St. Petersburg, Russia, including 49 with normal pregnancies (NP group), 52 with a history of more than three consecutive spontaneous abortions (induced abortions not included) (SA group), and 72 with ectopic pregnancies (EP group), were studied. Patient sera were examined by the microimmunofluorescence (MIF) serological assay and were evaluated for the presence of anti-Chsp10 and anti-Chsp60 antibodies.

A total of 187 women with normal pregnancies, attending the Center of Obstetrical Gynecology, Amiens, France, during the first trimester of pregnancy were enrolled for screening of C. trachomatis infections. Patient sera were examined by the MIF assay and evaluated for the presence of anti-Chsp10 immunoglobulin G (IgG) antibodies. Urine samples, taken on the same day as the serum samples, were tested by the transcription-mediated amplification (TMA) direct-detection assay for the presence of C. trachomatis rRNA. Finally, sera from 33 patients (19 women and 14 men) with documented chlamydial infections (a positive direct-detection result or MIF IgM or IgG seroconversion, or elevated MIF titers and relevant clinical data), who had had two to four follow-up serological examinations over 6 months, and sera from 36 patients (24 women and 12 men) with a positive direct fluorescence assay (DFA) result for C. trachomatis, were tested by the MIF assay and evaluated for the presence of anti-Chsp10 IgG antibodies. These patients were seen at the Center for Sexually Transmitted Diseases, Amiens, France.

DFA.

Urethral or cervical swab specimens were screened for C. trachomatis by the DFA (Syva Microtrack) as specified by the manufacturer.

TMA assay.

Urine samples were centrifuged and processed for C. trachomatis-specific rRNA detection as specified by the manufacturer (BioMérieux–Gen-Probe). The TMA assay is based on amplification of a rRNA target sequence via DNA intermediates. In this assay, the specific amplified rRNA sequence is detected by using a chemiluminescent single-stranded DNA probe, complementary to the amplicon (6). Labeled RNA-DNA hybrids are measured in a luminometer.

MIF assay.

The indirect MIF assay of Wang and Grayston (16), modified by Orfila and Eb (12), was performed with different bacterial strains prepared from egg yolk sacs (C. trachomatis LB1 (L2 serovar), Chlamydia psittaci Loth, and Chlamydia pneumoniae IOL 207). For each patient, various serum dilutions (1:16 to 1:2,048 for IgG and IgA detection or 1:12 to 1:96 for IgM detection) were tested on slides with acetone-fixed preparations of infected or noninfected (as negative control) eggs. The presence of species-specific antibodies was assessed by adding fluorescein isothiocyanate-conjugated anti-human IgG antiserum and fluorescein isothiocyanate-conjugated anti-human IgA and IgM antiserum (1/100 dilution) (TAGO Inc., Burlingame, Calif.). For the detection of species-specific IgM antibodies, tested sera were pretreated with rheumatoid factor (RF) absorbant (Behring Diagnostics Inc.) to avoid false-positive results. A titer of ≥16 was considered positive for IgG and IgA, and a titer of ≥12 was considered positive for IgM.

Detection of antibodies to Chsp60.

Recombinant Chsp60 was expressed as an N-terminally His₆-tagged molecule and purified by nickel-chelate affinity (International Microbio). The serum antibody response to chlamydial Hsp60 was determined by a prototype enzyme-linked immunosorbent assay (ELISA). Microtiter plates were coated with recombinant Chsp60 protein (International Microbio). Aliquots of sera diluted 1:500 in phosphate-buffered saline (PBS)-0.5% bovine serum albumin–0.2% Tween 20 were added to the wells, and the plate was incubated at 37°C for 60 min. The wells were then washed three times with PBS-Tween 20 and incubated with a 1:10,000 dilution of peroxidase-conjugated goat antibody to human IgG. After an additional 60-min incubation at 37°C, the wells were washed three times and the peroxidase substrate, tetramethylene benzidine was added. After incubation at room temperature in the dark for 30 min, the reaction was stopped with 0.5 M H₂SO₄ and the plates were read at 450 nm. Each sample was assayed in duplicate in recombinant Chsp60-coated wells, as well as in duplicate in noncoated wells, as a control. The final optical density at 450 nm (OD₄₅₀) value attributed to each serum sample was the difference between the mean value of the duplicate determination in coated wells and the mean value of the corresponding duplicate determination in noncoated wells.

Detection of antibodies to Chsp10.

Recombinant Chsp10 was expressed as a His₆-tagged molecule and purified by nickel chelate affinity chromatography (7). Microtiter plates were coated with 3 μg of recombinant Chsp10 per ml and saturated with 5% skim milk. Aliquots of sera diluted 1:80 in PBS–0.2% Tween–0.5% skim milk were added to the wells, and the plate was incubated at 37°C for 60 min. The wells were then washed three times with PBS-Tween 20 and incubated with a 1:1,000 dilution of alkaline phosphatase-conjugated goat antibody to human IgG. After an additional 60-min incubation at 37°C, the wells were washed three times and the alkaline phosphatase substrate, paranitrophenylphosphate, was added. After incubation at room temperature for 60 min, the reaction was stopped with 3 M NaOH and the plates were read at 405 nm. A positive sample was defined as one yielding an OD₄₀₅ value that was at least 2 standard deviations above the mean value obtained with a panel of 30 MIF-negative samples, that is, an O.D.₄₀₅ > 0.12. Index values were calculated as the OD₄₀₅/0.12.

Inhibition ELISA.

Patient sera with detectable anti-Chsp10 IgG antibodies were incubated overnight in test tubes at 4°C with different concentrations of a synthetic peptide, L18Q (LGTGKKDDKGQQLPFEVQ), corresponding to amino acids (aa) 50 to 67 of Chsp10 (Néosystem) and with E. coli recombinant (rGroES) (Sigma). rGroES was used at 10 μg/ml, while the 18-mer peptide was used at 0.6, 6, and 60μg/ml. From each of these tubes, 50 μl was dispensed in duplicate on the precoated Chsp10 plates. A sample without inhibitor was used to estimate 100% binding. A sample with bovine serum albumin at 60 μg/ml was used to estimate nonspecific inhibition. The inhibition coefficients (IC) were calculated from the final absorptions of the test samples by using the formula IC (%) = 100 × [A(0) − A(x)]/A(0), where A(0) is the absorption of serum solution without inhibitor and A(x) is the absorption of serum solution with peptide inhibitor at a concentration x. Inhibition curves were constructed for several serum samples (data not shown).

Statistical analysis.

Any difference between groups of data was calculated by the χ² test and Student’s t test. The 95% confidence intervals were calculated.

RESULTS

Performance and significance of anti-Chsp10 antibody detection.

The sensitivity and specificity of anti-Chsp10 IgG antibody detection, in comparison with direct detection (either by DFA or by TMA), were estimated to be 41.7 and 86.1%, respectively, in the population group of French origin (Table 1).

TABLE 1.

Percentages of positive anti-Chsp10 IgG responses and mean anti-Chsp10 index values^a

Population	MIF assay status	No. of patients	% Anti-Chsp10 positive (95% confidence interval)
Russian
NP	−	15	40 (16–69)
	+	34	17.6 (7–37)
SA	−	17	11.8 (1–34)
	+	35	22.9 (11–39)
EP	−	8	12.5 (0–47)
	+	64	39.1 (27–51)
Total			27.7 (21–35)
French
Directb +	+	36	41.7 (26–58)
Follow-up	+	33	45.4 (28–62)
Direct −	−	39	7.7 (3–21)
	+	148	15.5 (9–21)
Total			21.9 (17–27)

^aThe mean index (± standard errors) for the Russian and French populations, respectively, are as follows: MIF negative, 4 ± 2.4 and 1.9 ± 0.8; MIF positive, 5.5 ± 1.8 and 3.9 ± 1.2; total, 5 ± 1.4 and 3.5 ± 0.9. 

^bDirect detection by DFA or TMA. 

In the 33 patients followed up in Amiens for 6 months, 67% of the anti-Chsp10-positive patients had symptomatic or asymptomatic C. trachomatis upper or chronic genital tract infections whereas 67% of the anti-Chsp10 negative patients had current C. trachomatis lower genital tract infections. Therefore, there was a significant association between anti-Chsp10 IgG antibody production and upper or chronic genital tract infections (P < 0.02). Among the follow-up sera, four were from patients with anti-Chsp10 IgG seroconversion and two were from patients in whom anti-Chsp10 IgG titers declined over time. In all other patients, anti-Chsp10 IgG titers remained stable over time (data not shown).

Correlation between humoral immune responses to Chsp10 and Chsp60 antigens.

We have previously shown that in the same female population of Russian origin, high anti-Chsp60 titers have no particular association with illness in women seronegative for Chlamydia by MIF. However, in that study we showed that in MIF-positive women, a high anti-Chsp60 titer was significantly associated with an increased risk of ectopic pregnancy (P < 0.001) (14). We found similar associations for immune responses to Chsp10. Among Chlamydia MIF assay-positive patients, a significantly higher percentage of positive anti-Chsp10 serologic test results were seen in the EP group (39.1%) than in the NP (17.6%) and the SA (22.9%) groups. No such correlation was observed in the MIF assay-negative patients. No correlation between anti-Chsp10 mean index values and ectopic pregnancy or spontaneous abortion was observed. Mean anti-Chsp10 index values were higher in the Chlamydia MIF assay-positive than in the Chlamydia MIF assay-negative patients. The results are shown in Table 1.

Surprisingly, in the MIF assay-negative N.P. group of Russian origin, 40% of the patients had positive anti-Chsp10 ELISA serologic test results. No correlation between anti-Chsp10 index values and MIF titers was observed (data not shown). In spite of the link between Chsp10 and Chsp60 exposure, no serologic correlation between anti-Chsp10 and anti-Chsp60 titers was observed (Fig. 1). Immune responses to Chsp10 and Chsp60 are apparently independent of one another.

FIG. 1.

Humoral immune responses to Chsp10 and Chsp60 in a group of 173 women (normal pregnancies, spontaneous abortions, and ectopic pregnancies included).

Specificity of anti-Chsp10 humoral immune response.

Sera from patients with a positive anti-Chsp10 titer by ELISA, of both Russian and French origin, were studied by inhibition ELISA (i) with a synthetic 18-mer peptide L18Q (aa 50 to 67), corresponding to the C. trachomatis Chsp10 species-specific epitope of monoclonal antibody M1.2 reported by LaVerda and Byrne (7), and (ii) with E. coli rGroES protein. The results are shown in Table 2. Limited competition between E. coli rGroES and Chsp10 was observed in the population of Russian origin (6.2%), as well as in the French population with documented C. trachomatis infection (13.3%), suggesting Chlamydia specificity of the anti-Chsp10 serologic test result. In contrast, 46.1% of the French female population with a negative C. trachomatis result by direct detection and positive anti-Chsp10 serologic test result showed evidence of cross-reaction with rGroES protein, suggesting that extrachlamydial factors account for anti-Chsp10 IgG antibodies (Table 2).

TABLE 2.

Percentage of positive ELISA results in the anti-Chsp10-positive population

Population	% Positive (95% confidence interval)
	L18Q inhibition	GroES inhibition	MIF for C. pneumoniaea
Russian (n = 173)	33.3 (20–46)	6.2 (1–18)	17.9 (8–34)
French
Directb +/follow-up (n = 99)	33.3 (16–50)	13.3 (4–30)
Direct − (n = 187)	19.2 (8–41)	46.1 (40–52)
Total	26.8 (15–39)	28.6 (17–41)	39.1 (20–63)

^aCalculated in the Chlamydia MIF assay-positive population. 

^bDirect detection by DFA or TMA. 

Variable inhibition coefficients (IC) with the L18Q synthetic peptide, ranging from 20 to 90%, were observed for 33.3 and 26.8% of the Russian and French anti-Chsp10-positive serum samples, respectively, suggesting a C. trachomatis specificity of the observed anti-Chsp10 response and indicating that the Chsp10 region spanning aa 50 to 67 contains at least one immunodominant epitope in a subset of C. trachomatis-infected individuals. Interestingly, IC ≥ 50% was observed in 14.6% of the Russian population and in only 5.4% of the French population. It has been shown that C. trachomatis serotype E Hsp10 and C. pneumoniae AR-39 Hsp10 have 78.4% identity. Cross-reactivity between C. trachomatis Hsp10 and C. pneumoniae Hsp10 exists, since 17.9 and 39.1% of Chlamydia MIF-positive Russian and French women, respectively, were positive by MIF for C. pneumoniae only but not for C. trachomatis (Table 2). This corresponds to 14.5 and 16% cross-reactions in the anti-Chsp10 positive women of Russian and French origin, respectively. Sera from approximately one-third of these women exhibit an inhibition effect by the L18Q peptide.

DISCUSSION

The sequelae of chlamydial infections, such as infertility or ectopic pregnancies, are associated with chronic inflammation and tubal changes, probably due to immunopathologically mediated events following long-term upper genital tract infection. Studies involving serologic tests to diagnose C. trachomatis upper genital tract injury are hindered by the fact that more than one organism may invade the upper genital tract and by an inability to establish an appropriate “gold standard” for an infected upper genital tract. A predictive serologic test for upper genital tract infection would be a desirable diagnostic tool. In one study, involving PCR of biopsy tissue as a standard of comparison, Chsp60 ELISA demonstrated 43% sensitivity and 100% specificity (4). Much correlative evidence exists between the antibody response to Chsp60, which displays 48% identity to the human Hsp60 (11), and the development of chronic inflammatory sequelae (3, 15), although no direct evidence exists to support an autoimmune mechanism linked to Chsp60. Very little is known about the biological and immunological properties of Chsp10.

The C. trachomatis serovar A and E Hsp10 is 36% identical to the E. coli GroES protein (8) and 35.6% identical to its human mitochondrial homolog (8). Its serological investigation should be interesting because of its coexpression with Chsp60. Our results suggest that an anti-Chsp10 IgG response in Chlamydia MIF assay-positive individuals indeed correlates with upper or chronic genital tract infections and ectopic pregnancies. Therefore, anti-Chsp10 antibodies might be considered a positive predictive factor for chronic infection, independent of anti-Chsp60 antibodies, since no serologic correlation was found between the two types of antibodies. Recombinant Hsp10 has been cloned from C. trachomatis serotype E (7), and recombinant Hsp60 has been cloned from serotype L2 (International Microbio); both serotypes belong to the same serogroup. Thus, serotype variations are not likely to be at the origin of the lack of serologic correlation. The absence of such serologic correlation, in spite of coexpression and probable association between the two antigens, might be due to genetic factors. In mice, the antibody response to Chsp60 is genetically determined and is partly linked to H-2 genes (18). Furthermore, population-type dependence of anti-L18Q antibody levels might also be due either to genetic restriction or to multiple infections with repeated B-cell stimulation, occurring in the Russian population.

The 42% sensitivity in comparison to direct detection might be due either to genetic predisposition of specific patients or to different immunological responses associated with primary infection, reinfection, or ascending infection. La Verda and Byrne (7) have reported a difference between the calculated and the observed gel mobility for the native and recombinant Chsp10, which may be due to posttranslational modification occurring in C. trachomatis but not in E. coli. Absent posttranslational modification may account for the low sensitivity of the ELISA technique if the recombinant protein lacks certain epitopes.

The 86% specificity may be partially explained by a total of 10.5% possible cross-reactions with GroES protein or other bacterial homologs, as suggested by the GroES inhibition ELISA experiments.

The lack of correlation between anti-Hsp10 and MIF titers is not surprising, since MIF detects primarily antibodies against the two major exposed elementary-body antigens: major outer membrane protein and lipopolysaccharide. Overproduction and release of Chsp10 during the infectious process may be induced by the host primary immune response to C. trachomatis-infected cells.

A surprisingly high percentage of positive anti-Chsp10 sera was found in the MIF assay-negative, NP group of Russian origin. Because of the high degree of relatedness between Hsp of different bacteria, it is possible that in at least some of those reactive patients, the positive anti-Chsp10 sera were the result of infection by microorganisms other than C. trachomatis. Indeed, clinical data on the Russian population suggest a high percentage of undiagnosed or untreated infection in this group. This renders the Russian NP control group a rather atypical one and explains the significant difference between this group and the corresponding French control group (40 and 8% positive anti-Chsp10 serologic test results, respectively), as well as our inability to establish a cutoff value in the recombinant Chsp60 ELISA, on the basis of the Russian control group.

One can hypothesize that among other pathogens, C. pneumoniae infection might play a role in the observed level of anti-Chsp10 antibody response. The region spanning aa 50 to 67, containing an epitope, previously reported by LaVerda and Byrne (7) to be C. trachomatis species specific, is probably shared by C. pneumoniae, since L18Q inhibition is observed for some of the serum samples that were exclusively C. pneumoniae positive in the MIF assay.

Our study suggests that in spite of possible cross-reactive antibody binding to C. pneumoniae or other pathogens, detection of anti-Chsp10 antibodies might represent a useful marker for C. trachomatis upper or chronic genital tract infection in patients positive for C. trachomatis by direct detection or with Chlamydia antibodies detected by the MIF assay. Interestingly, Chsp10 seems to represent a marker independent of Chsp60. Whether prolonged expression of Chsp10, leading to stimulation of humoral and/or T-cell-mediated immune responses, is a cause or a consequence of chronicity of infection remains to be determined. More prospective studies on self-reactive B-cell immunity to Chsp10, on T-cell-mediated immune responses possibly elicited by Chsp10, and on elucidation of specific subsets of C. trachomatis-infected individuals who develop immune responses to Chsp10 are needed.

Screening programs have been demonstrated to reduce the overall prevalence of chlamydial infection in the tested population and to reduce the incidence of subsequent pelvic inflammatory disease in previously screened women (13). The inclusion of anti-Chsp10 antibody detection in such screening programs should be considered.