N. Y. Academy of Science, New York, N.Y.. 18. Nowicki, S., M. ... Parsons, N. J., P. V. Patel, E. L. Tan, J. R. C. Andrade, C. A. Nairn, M. Goldner, J. A. Cole, and H.
INFECTION AND IMMUNITY, June 1997, p. 2094–2099 0019-9567/97/$04.0010 Copyright © 1997, American Society for Microbiology
Vol. 65, No. 6
Pelvic Inflammatory Disease Isolates of Neisseria gonorrhoeae Are Distinguished by C1q-Dependent Virulence for Newborn Rats and by the sac-4 Region STELLA NOWICKI,1,2* PRASHANTH RAM,1 TUAN PHAM,1,2 PAWEL GOLUSZKO,1 STEPHEN MORSE,3 GARLAND D. ANDERSON,1 AND BOGDAN NOWICKI1,2 Department of Obstetrics & Gynecology1 and Department of Microbiology & Immunology,2 The University of Texas Medical Branch at Galveston, Galveston, Texas 77555-1062, and Centers for Disease Control and Prevention, Atlanta, Georgia3 Received 18 October 1996/Returned for modification 12 December 1996/Accepted 12 March 1997
The virulence mechanism of Neisseria gonorrhoeae in pelvic inflammatory disease (PID) is not well understood, and an objective diagnostic method to identify patients with PID is lacking. We investigated the hypothesis that development of PID was associated with a C1q-dependent virulence property of gonococcal strains. Recent development of a C1q-dependent experimental model of gonococcal infection (S. Nowicki, M. Martens, and B. Nowicki, Infect. Immun. 63:4790–4794, 1995) created an opportunity to evaluate this hypothesis in vivo. Therefore, the virulence of 32 clinical isolates (18 PID isolates and 14 local infection [LI] isolates) was evaluated in experimental rat pups. A serum bactericidal assay was used to characterize a gonococcal serum-resistant (serr) phenotype. PCR primers designed to amplify a suitable-size gonococcal sac-4 DNA fragment (unique for serum-resistant donor JC1) were used to evaluate the association of serum-resistant genotype sac-4 with two phenotypes: C1q-dependent virulence expressed in vivo and resistance to bactericidal activity of human serum expressed in vitro. Strains were also characterized by auxotyping and serotyping. Of 32 gonococcal strains, 15 (46.7%) caused C1q-dependent bacteremia in rat pups and were sac-4 positive and serr. However, of the 15 isolates, 13 (87%) represented strains associated with human PID and 2 (13%) were associated with LI. None of the strains that were completely serum-sensitive (sers) and sac-4 negative produced C1q-dependent bacteremia in rat pups, suggesting that both serr and sac-4 were required for infection. The serum-resistant recombinant recipient of sac-4 produced C1q-dependent bacteremia in the rat model similarly to the serum-resistant donor of sac-4; the serum-sensitive parent strain did not produce bacteremia. These data suggest that sac-4-mediated serum resistance conferred C1q-dependent virulence and is a unique characteristic associated with PID. These newly identified features may contribute to the understanding of the pathogenic mechanism of PID-associated strains and open perspectives for establishing novel diagnostic methods. A nonprimate model of gonococcal infection in rat pups was recently developed (19). In this model, intraperitoneal (i.p.) inoculation of serum-resistant (serr) N. gonorrhoeae JC1 together with human complement C1q resulted in bacteremia that lasted 6 to 7 days. Preincubation of strain JC1 with C1q protected N. gonorrhoeae from the bactericidal effect of rat pup serum both in vitro and in vivo instead of initiating killing via activation of the classical complement pathway (19). To the contrary, direct binding of C1q by Klebsiella pneumoniae results in killing of the target cell (1). Thus, paradoxical C1qmediated resistance to complement killing of N. gonorrhoeae JC1 may represent an unrecognized mechanism of gonococcal virulence. Resistance to the complement-dependent bactericidal effect of normal human serum (NHS) appears to be a crucial virulence factor for N. gonorrhoeae (21). The known mechanisms whereby gonococci are resistant to killing by NHS and complement have been separated into two types, stable (not affected by subculture) and unstable (lost on subculture) (21). C1q-dependent virulence of stable serr N. gonorrhoeae may be analogous to the mechanism employed by sers (unstable serr) gonococci in which host CMP-N-acetylneuraminic acid is used to sialylate lipooligosaccharide (LOS) to escape the bactericidal effect of serum (10, 20). Changes in the carbohydrate composition of LOS, changes in the outer membrane proteins, or changes in the binding-blocking antibody are among proposed mechanisms (4, 11). Loci called sac-1, sac-3, and sac-4
Pelvic inflammatory disease (PID) is a serious and costly condition among women of reproductive age (5, 7). The cost of treating PID in the United States is estimated to reach $10 billion by the year 2000 (7). PID frequently results in severe irreversible sequelae, such as infertility, ectopic pregnancy, and chronic pelvic pain (2, 6, 7, 13, 28). The pathogenesis of PID is not well understood, and an objective diagnostic method for identifying patients who are at risk for PID is lacking. Traditionally, PID has been characterized as a major complication of genital tract infection caused by sexually transmitted pathogens Neisseria gonorrhoeae and Chlamydia trachomatis and anaerobes and facultative bacteria from the lower genital tract (7). Whether the development of gonococcal PID is determined by host susceptibility or by the virulence of the gonococcus remains to be determined. Elucidating the contribution of these factors would enhance our understanding of the mechanism of gonococcal infection and perhaps lead to the improvement of diagnostic, therapeutic, and prophylactic approaches. In the past, the lack of a simple experimental model of gonococcal infection has significantly limited studies of host and microbial properties contributing to the pathogenesis of PID. * Corresponding author. Mailing address: Division of Infectious Disease, Department of Obstetrics & Gynecology, The University of Texas Medical Branch at Galveston, 301 University Blvd., Galveston, TX 77555-1062. Phone: (409) 772-7598. Fax: (409) 747-0475. 2094
PID ISOLATES OF N. GONORRHOEAE
VOL. 65, 1997
have been implicated in the molecular mechanism of serum resistance (3, 29). A 2.2-kb region of sac-4 cloned from N. gonorrhoeae JC1 DNA was able to transform N. gonorrhoeae F62 from a sers phenotype to a stable serr phenotype (12); however, the contribution of sac-4 to virulence is not well understood. There are no clear conclusions to investigations of the association of PID with stable serum resistance of N. gonorrhoeae; however, this has been determined with disseminated gonococcal infection (25). According to a previous study, PID isolates appear to represent a heterogeneous group with respect to serum sensitivity (23). Therefore, the role of serum resistance as a virulence factor in gonococcal PID remains to be investigated. The onset of gonococcal PID is often clinically associated with menstruation (27). We hypothesize that contact of gonococci with menstrum containing serum rich in C1q could be an unknown mechanism that increases the virulence of gonococcal strains and contributes to the onset of PID or septic spread (17, 18). The clinical features of gonococcal PID that are similar to those of disseminated gonococcal infection include capacity to cause bacteremia, peritonitis, Fitz-Hugh-Curtis syndrome, and septic spread in the host or the fetus from mothers with untreated PID during pregnancy. Furthermore, we hypothesize that not all strains of N. gonorrhoeae are able to establish PID. By testing strains isolated from patients with PID and local infections (LI) for C1q-dependent virulence, we attempted to address these two important issues. We provide novel evidence that PID, but not LI, isolates of N. gonorrhoeae display C1q-dependent virulence in the rat pup model. The molecular mechanism of the C1q-mediated virulence was found to be associated with the sac-4 region and stable serum resistance.
2095
was removed and stored at 220°C for subsequent use in PCRs. The sequences of the primers were as follows: primer A, 59 TATCTGCAGCATCTCCTTTCCA ACC 39; and primer B, 59 TAGGAATTCCTCTGAAGGTTACGG 39. PCR was performed with a Gene Amp PCR reagent kit (Perkin-Elmer Cetus, Norwalk, Conn.). The reaction mixture contained 2 ml of DNA template, 200 mM each deoxynucleotide phosphate, 2.5 U of Amplitaq polymerase, and 1.0 mM each primer in a total volume of 100 ml. The reaction mixture was overlaid with 50 ml of mineral oil. The amplification reaction consisted of 30 cycles of 1-min denaturation at 92°C, 1-min annealing at 55°C, and 1-min extension at 72°C. An aliquot of 20 ml from each tube was removed for electrophoretic analysis on a 2% agarose gel. Serotyping and auxotyping. Typing of gonococcal isolates was performed at the CDC according to previously published procedures (9). In brief, serotyping has been determined by evaluating the major membrane protein I. Protein I exists in two distinct species, IA and IB, identified by sets of monoclonal antibodies. Auxotypes have been determined on the basis of gonococcal requirements for specific nutrients: Arg requires arginine, Pro requires proline, and Proto requires cystine-cysteine for growth. Serum bactericidal assay. The serum bactericidal assay was a modification of the procedure used by McCutchan et al. (11). Gonococci (90% P1 Opa1) were removed from an overnight GC base agar (GCBA) plate culture with a sterile Dacron swab and suspended in GCBB to an A600 of 0.2; a 1:1,000 dilution was made, and 0.1 ml of the diluted cell suspension was mixed with 0.1 ml of 10, 25, and 50% NHS (Sigma, St. Louis, Mo.) in GCBB. Another 0.1-ml aliquot of the cell suspension was mixed with 0.1 ml of 10, 25, and 50% heated human serum at 56°C for 30 min. Both tubes were incubated at 37°C for 45 min. At the end of the incubation period, 10-fold and 100-fold dilutions were made from the assay mixtures; 0.05-ml volumes were spread on GCBA plates in triplicate and incubated for 24 h at 37°C in a 5% CO2 incubator, and then the numbers of CFU of each strain per milliliter were determined. Total serr N. gonorrhoeae JC1 and completely sers N. gonorrhoeae F62 were pretested to evaluate the bactericidal activity of NHS used for experiments. Some human sera, such as lot 124 H9412 obtained from Sigma, were not used for evaluation due to killing properties on JC1. Construction of recombinant molecules and DNA sequencing. As previously described, purification of plasmid DNA digestion with restriction endonucleases and ligation to the pUC19 vector were performed (26). DNA sequencing was performed by the method of Sanger et al. (24). Statistical methods. For statistical evaluation of the associations between two factors, the Fisher exact test was utilized to determine the exact significance levels (P value). The observed significance levels (P values) imply the existence of a statistically significant association between the tested factors.
MATERIALS AND METHODS Bacterial strains and culture conditions. Gonococcal PID isolates (15 strains) and LI isolates (11 strains) were obtained from the Centers for Disease Control and Prevention (CDC) in Atlanta, Ga. This collection represents strains isolated from patients with either PID or lower genitourinary tract infection, from different regions in the United States. Additionally, three PID strains and three LI strains were obtained from the University of Texas Medical Branch (UTMB) clinics. N. gonorrhoeae JC1 (serr) and two isogenic strains, WM3 (F62 recombinant selected following transformation with pWM3 containing the sac-4 region) (serr) and F62 (sers), were provided by R. Hull (12). In brief, sers F62 was transformed by plasmid pWM3, which is pHC79 containing 2.2 kb of the sac-4 region cloned from JC1. Selection of serum-resistant recombinants was further performed by submitting transformants to a serum assay. The selected serr recombinant resulting from a double-crossover event did not maintain plasmid pHC79, as tested by growth on appropriate antibiotic media and DNA hybridization with pHC79 (12). The control F62 was subjected to mock transformation (12). All strains were grown on chocolate agar plates or modified Thayer-Martin agar plates (Remel Laboratories, Richardson, Tex.) and incubated for 18 to 24 h at 37°C in 5% CO2. Liquid cultures were grown in GC base broth (GCBB) supplemented with 1% supplement VX (Difco) (0.04% sodium bicarbonate) and 5 g of glucose per liter. Experimental model. Sprague-Dawley rats (3- to 5-day-old newborns) were divided into two groups and inoculated i.p. with 100 ml of a suspension of gonococci (5 3 107 CFU/ml) pretreated with either 100 mg of C1q per ml or bovine serum albumin (BSA) (control). Alternatively, control heat-inactivated C1q (100°C, 30 min) was also used (19). Five animals from each group were sacrificed 4 h postinoculation. Twenty-microliter blood samples were collected without anticoagulant from heart punctures, modified Thayer-Martin agar plates were inoculated with 10-fold dilutions, and the number of CFU per milliliter was determined after 48 h of incubation. PCR. Deletion mutants of pWM3 constructed by McShan et al. (12) showed that only the 39 region of sac-4 encoding a 29-kDa peptide was necessary for conferring partial serum resistance. In order to determine the prevalence of this region in gonococcal isolates, PCR primers were designed to amplify a suitablesize fragment (344 bp) of DNA from the 39 region of sac-4 necessary for transforming the serum-sensitive laboratory strain F62 to serum-resistant WM3. A 1-ml volume from an overnight GCBB culture was centrifuged, and the pellet was resuspended in 200 ml of distilled H2O. The suspended cells were heated at 100°C for 20 min, cooled, and centrifuged for 5 min; the resulting supernatant
RESULTS C1q-dependent virulence for rat pups differentiates PID from LI isolates. To investigate our hypothesis that unique strains of N. gonorrhoeae are associated with PID, 18 PID and 14 LI isolates of N. gonorrhoeae obtained from CDC and UTMB were tested for C1q-dependent virulence in newborn rats. Two groups of 3- to 5-day-old rats were inoculated i.p. with either 100 ml of 5 3 106 gonococci and human complement C1q (100 mg/ml) or, as a control, 100 ml of 5 3 106 gonococci and BSA (100 mg/ml). Previously published results showed that bacteremia with N. gonorrhoeae JC1 and four PID strains occurred in the presence of C1q but showed a significantly lower rate in the presence of either heat-inactivated C1q or BSA (19). Accordingly, to reduce the significant cost of C1q, control experiments were performed in the presence of BSA, excluding pilot experiments. Previous kinetics studies had indicated an exponential gonococcal growth at 4 h for virulent JC1 and a lack of growth for commensal Neisseria cinerea (19). At 4 and 18 h, the positive-control strain JC1 was present at levels of 5 3 104 and 9 3 105 CFU/ml; the negative-control N. cinerea strain was not detected in the blood at any tested time point from 30 min to 24 h. To evaluate the optimal time for blood culture that would differentiate virulent from nonvirulent isolates, blood was drawn at 0.5, 2, 4, 6, 8, and 24 h postinoculation and cultured on agar plates. Four hours postinoculation was considered to be the earliest convenient and optimal time to differentiate between virulent and nonvirulent isolates. At 4 h postinoculation, blood was obtained from each rat pup (n 5 340), diluted, and used to inoculate agar medium.
2096
NOWICKI ET AL.
INFECT. IMMUN.
TABLE 1. Phenotypic and genotypic characterization of isolates of N. gonorrhoeae associated with different clinical forms of gonococcal infection Strain
Auxotype
Serovar
C1q virulencea
Presence of sac-4b
Serum sensitivityc
Source
2004 2005 2016 2019 2081 1653 1657 1665 1671 2014 2075 2086 1655 1659 2009 2010 2015 2025 2037 2059 2065 1658 1660 1663 1667 1673 4221 4222 4223 3111 3112 3113
Pro Pro Pro Pro Pro Pro Pro Pro Pro Arg Arg Arg Arg Arg Proto Proto Proto Proto Proto Proto Proto Proto Proto Proto Proto Proto NT NT NT NT NT NT
IA6 IB3 IB14 IB1 IB7 IB4 IB1 IB3 IB2 IB3 IB6 IB6 IA4 IA4 IB6 IB2 IB3 IB3 IB3 IB7 IB3 IB4 IB2 IB3 NTd NT NT NT NT NT NT NT
N Y N N Y N N N N Y Y Y N N Y N Y Y Y N Y N N N Y Y Y Y Y N N N
2 1 2 2 1 2 2 2 2 1 1 1 1 1 1 2 1 1 1 2 1 2 2 2 1 1 1 1 1 2 2 2
S R S S R S S R S R R R S S R S R R R S R S S S R R R R R S S S
PID PID PID PID PID LI LI LI LI PID PID PID LI LI PID PID PID PID PID PID PID LI LI LI LI LI PID PID PID LI LI LI
a
Capacity of tested strain to cause infection of blood in the presence of C1q. Y, yes (positive blood culture); N, no (negative blood culture). b 1, detection of 344-bp sac-4 region DNA by PCR; 2, negative PCR. c S, serum sensitive; R, serum resistant. d NT, not tested.
Table 1 shows a comprehensive listing of detailed characteristics of each tested isolate. Table 2 shows a comparative analysis of all tested properties based on the clinical origin of strains. Of 32 strains preincubated with C1q prior to inoculation of rat pups, 15 were recovered from the blood at 4 h. Of the 15 gonococcal strains recovered from rat pup blood, 13 (87%) were PID strains and 2 (13%) were LI strains (Table 2). Cultures for all control animals inoculated with 32 strains excluding C1q (n 5 160) were negative. Fisher’s exact test (PID, P # 0.001) evaluated the differences as highly significant. The number of CFU observed in positive cultures ranged from 2 3 103 to 7 3 105/ml. Detection of sac-4 region among PID and LI isolates by PCR. The sac-4 region was necessary for transformation of sers F62 to serr WM3 (12, 14a). Primers A and B were used to amplify by PCR a 344-bp DNA segment in the 39 end of the sac-4 region. The amplified product was detected by agarose gel electrophoresis (Fig. 1 and 2). All PCR-positive strains showed an amplification product of identical size which migrates close to the 350-bp marker. Two PCR amplification products were subcloned to plasmid pUC19, and DNA inserts were sequenced. A 344-bp sequence from the 39 end of the sac-4 region was identified, thereby confirming that the analyzed PCR products represented the sac-4 region of the tested
TABLE 2. Virulence markers evaluated for differentiation of PID from LI isolates Marker
No. of strains
% PID
PID
LI
PID 1 LI
C1q virulenceb Yes No
13 5
2 12
15 17
87 29
sac-4 Detected Not detected
13 5
4 10
17 15
77 33
IB3c Yes No
4 11
2 7
6 18
67 61
Protod Yes No
7 8
5 6
12 14
58 57
Serum resistance Resistant Sensitive
13 5
3 11
16 16
82 31
C1q vir, sac-4, serre Yes No
13 5
2 12
15 17
87 29
Pa
#0.001*
#0.03*
$0.75
$0.75
#0.01*
#0.001*
a
Statistically significant difference. Capacity of tested strain to cause C1q-dependent bacteremia. c Most frequent serotype among PID strains. d Most frequent auxotype associated with PID strains. e Presence of all three characteristics together: expression of C1q-dependent bacteremia, detection of sac-4, and serum resistance. b
strains. Of the 32 strains tested, 17 were sac-4 positive and 15 were sac-4 negative (Table 2). Of the 17 sac-4-positive strains, 13 (77%) are PID isolates and 4 (23%) are LI isolates. In contrast, of the 15 sac-4-negative strains, 5 (33%) and 10 (67%) represent PID and LI, respectively. The observed differences between PID and LI strains in detection of sac-4 were significant (P # 0.03) and specific for the majority of PID isolates. Auxotypes and serotypes predominating among PID and LI isolates. Typing of PID and LI isolates was performed to determine if there was any association between auxotype and serotype and the ability to produce bacteremia in rat pups. Among the identified PID-associated auxotypes were Proto
FIG. 1. The 344-bp region of sac-4 targeted for PCR amplification. The physical map of sac-4 region is as described by McShan et al. (12); pHC79 (bars) and the insert sac-4 DNA of JC1 (line) are shown. The PCR target (open bar and dotted lines) is indicated. Abbreviations representing endonuclease restriction sites are as follows: A, AvaI; B, BamHI; H, HindIII; N, NarI; P, PstI; S, Sau3A; and V, PvuII.
PID ISOLATES OF N. GONORRHOEAE
VOL. 65, 1997
2097
FIG. 2. PCR characterization of selected PID and LI isolates. An agarose gel of the amplified 344-bp DNA segment of sac-4 indicates the presence of sac-4 in strains causing PID and its absence in strains from LI. Lane ST, standard; lane 1, positive control (N. gonorrhoeae JC1); lanes 2 to 5, PID strains; lanes 6 to 9, LI strains. Fragment sizes are indicated in base pairs.
(seven strains), Pro (five strains), and Arg (three strains). The PID strains belonged to several serovars, including IB3 (six strains), IB6 (three strains), and IB7 (two strains) (Table 1). The remaining strains represented single serotypes IB1, IB2, IA6, and IB14. The most frequent association of auxotype/ serovar (Proto/IB3) was represented by four strains (37%) (Table 2). Among the identified LI auxotypes were Proto (five strains), Pro (four strains), and Arg (two strains). The LI strains belonged to several serovars, including IA4, IB2, IB3, and IB4 (two strains of each) and IB1 (one strain). The most frequent association of auxotype/serovar of tested LI strains, Arg/IA4, was represented by two isolates. There was no association observed between auxotype/serovar and the C1q-dependent virulence of these strains (P $ 0.95). There was also no observable correlation between the auxotype/serotype of all tested strains (PID and LI) and sac-4 positivity. Also, the serum resistance was not associated with any particular auxotype or serotype. Does the serum-resistant phenotype differentiate PID from LI isolates? Standard criteria used by other investigators for evaluation of stable serum resistance of N. gonorrhoeae include survival rate of 50% or greater in 25 or 50% NHS. The sac-4 locus encodes partial stable serum resistance (12). Recombinant strain WM3 (sac-4 positive) showed 30% survival in 25% NHS (12). To evaluate potential association between sac-4 and serr, the sensitivity of 32 strains of N. gonorrhoeae to NHS was tested in 10 and 25% pooled NHS. The majority of the tested strains survived at a rate of greater than 50% in 10 and 25% NHS, and serr did not correlate with the clinical source and serum-resistant sac-4 genotype. The majority of PID strains from the CDC collection showed a survival rate of 10 to 30% in 50% of NHS. The criteria selected for evaluation of resistance to NHS among sac-4-positive isolates were as follows: strains showing 10% survival or greater in 50% NHS were classified as serum resistant, while strains showing less than 10% survival in 50% NHS were classified as serum sensitive. These criteria of resistance to the bactericidal activity of NHS allowed differentiation of PID strains from LI strains of N. gonorrhoeae. The results (Table 2) indicate that of 32 strains, 16 were serum resistant, 13 (82%) of the PID strains and 3 (18%) of the LI strains. Association of C1q-dependent virulence of N. gonorrhoeae isolates with serum resistance and sac-4 among clinical isolates. The results (Table 3) showed that 100% of strains expressing the serum-resistant phenotype and genotype were vir-
ulent in the C1q-dependent animal model. Interestingly, one serum-resistant but sac-4-negative LI strain was avirulent for the rat pups (Table 1). Two additional sac-4-positive and sers LI strains were also not virulent in the rat pup model. A set of three characteristics, C1q-dependent bacteremia, serr, and sac-4, was associated predominantly with PID strains (13 of 15 [87%]). Strains expressing only one feature, the serum-resistant phenotype or the serum-resistant genotype, did not develop bacteremia in the experimental model (P $ 0.001). Serum-resistant recombinant WM3 expresses C1q-dependent virulence. The results (Table 2 and 3) show that bacteremia in the rat pup model was strictly correlated with the serumresistant phenotype and genotype. To define whether C1qdependent bacteremia of PID isolates could be conferred by the sac-4 region, experimental bacteremia was evaluated by using two isogenic gonococcal strains, F62 (sers) and recombinant WM3 (serr) (F62 recipient of sac-4). Donor of sac-4 gonococcal strain JC1 (serr) was used as a positive control (19). In the presence of C1q, strain JC1 and the recombinant strain WM3 were isolated from the bloodstream at 2, 4, 8, 12, and 24 h postinoculation (Fig. 3). N. gonorrhoeae F62 was not capable of disseminating in the presence of C1q at any tested time point, including 2, 4, 8, 12, and 24 h. The control blood cultures of rat pups demonstrated that in the absence of C1q, none of the three strains caused bacteremia. These experiments demonstrated that recipient of sac-4 WM3 (serr) acquired C1q-dependent virulence. TABLE 3. Association of C1q-dependent virulence of N. gonorrhoeae isolates with serum resistance and sac-4 among clinical isolates Characteristic
No. of strains
% Virulent
Virulent
Avirulent
Total
sac-4a Detected Not detected
15 0
2 15
17
88
Serum resistance Resistant Sensitive
15 0
1 16
16 16
94
sac-4, serr Yes No
15 0
0 17
15 17
100
a
P
#0.001
#0.001
#0.001
The presence of the 344-bp DNA of the sac-4 region was detected by PCR.
2098
NOWICKI ET AL.
FIG. 3. Association of C1q-dependent virulence with sac-4 in isogenic strains WM3 and F62. The experiments were performed in the presence of C1q. In vivo virulence was represented by the number of gonococci (CFU per milliliter) recovered from pup’s blood drawn at 2, 4, 8, 18, and 24 h postinoculation. Hatched bars, JC1; open bars, WM3; and closed bars, F62. Each bar represents the mean value of three independent experiments and standard errors for each group of data. The average number of pups tested per experiment was six.
DISCUSSION Here, we present the first experimental evidence that gonococcal PID strains can be differentiated from LI strains by their virulence in an experimental animal model. The development of bacteremia in rat pups was associated with gonococcal strains that possess the capacity to display C1q-dependent virulence and the presence of the sac-4 DNA region. All strains that were capable of surviving in rat pups expressed the serumresistant phenotype. Virulence in rat pups was not associated with a particular auxotype and serovar; conversely, all sac-4negative and sers strains were avirulent for rat pups. This finding suggested that the serum resistance, sac-4 region, and C1q-dependent virulence were common characteristics for 72% of PID isolates. Five (28%) strains associated with PID were avirulent for rat pups. There are two potential explanations for clinical association of these strains with PID. (i) The strains may express a different virulence mechanism(s) or their virulence was enhanced by mixed infection. (ii) The patients had compromising host conditions, such as immunodeficiency. Of 14 LI strains, 9 (64%) show a common pattern but one which is different from that of the majority of PID isolates. These common features include a lack of the 344-bp DNA fragment of sac-4, serum sensitivity, and a lack of C1q-dependent virulence for pups. Of the five remaining LI strains, two (14%) showed characteristics of PID isolates and were virulent for pups. The other three LI isolates that carried a single feature, sac-4 or serr, were avirulent for pups. A hypothetical explanation for the lack of virulence in the serr sac-4-negative strain may include a different, sac-4-independent mechanism of serr (for example, determined by the LOS structure associ-
INFECT. IMMUN.
ated with sac-1) (3). Two strains that were sac-4 positive and sers did not express the serum-resistant phenotype necessary for virulence of N. gonorrhoeae in the C1q-dependent experimental model. These unique strains are currently under investigation. Overall, our study indicates that three specific traits may be associated with gonococcal strains isolated from PID patients: (i) C1q-dependent bacteremia in the rat pup model, (ii) stable resistance to killing by NHS, and (iii) the presence of the 344-bp fragment sac-4. It is tempting to speculate that women with LI who are infected with serr sac-4-positive strains are at increased risk of developing PID. Fourteen percent of the LI strains were serr and sac-4 positive and classified as virulent in the rat pup model. This value is similar to the proportion of women with local gonococcal infections who develop PID (15 to 25%) (13). The present data suggested that sac-4 may confer C1q-dependent virulence. Direct evidence that the sac-4 locus conferred C1q-dependent virulence was obtained by using recombinant serr strain N. gonorrhoeae WM3 (12). The experiment demonstrated that the recipient of sac-4, WM3, but not its parent strain, F62, acquired C1q-dependent virulence. A 40% difference in the levels of JC1 and WM3 CFU may be expected, since sac-4 confers a stable but partial serum resistance (12). N. gonorrhoeae JC1 possesses other properties that may increase gonococcal virulence for pups. For example, the carbohydrate composition of the principal LOS species of strain JC1 does not allow bactericidal antibodies to bind (14). Nevertheless, strain WM3, containing the serum-sensitive LOS epitopes of strain F62 (14), was capable of causing bacteremia, suggesting a key role of sac-4 in bacteremic spread. Similarly, the association between C1q-dependent virulence and sac-4 among isolates from patients with PID suggests a significant contribution of sac-4 to the virulence of these isolates. Further investigation into the possible contribution of C1qdependent virulence to development of gonococcal infection in the upper genital tract in some reproductive-age women is needed. To our knowledge, this seems to be the first phenotype characteristic for the majority of PID, but not LI, strains. Human female volunteers are excluded from PID experimentation. The different anatomy and physiology of the male genital tract allowing study of male urethritis using PID isolates may not be directly related to female infection. The rat pup model, even if not ideal, allows analysis of PID strains in vivo. In 1991, a report of the National Institutes of Health prioritized studies of female health and PID (7). Despite attempts to identify risk factors for PID development, it was not clear whether human and/or gonococcal factors are important in the development of PID (22). Rice and Schachter showed that strains from acute PID patients were more sensitive to human serum than strains from chronic PID patients (23). The killing of more-sensitive strains was postulated to generate a more aggressive, nonspecific immune response by activation of the complement cascade and chemotactic response and to stimulate rapid damage of infected tissue, resulting in the clinical manifestation of the acute stage. Our results suggest that both human and gonococcal factors are required for virulence of N. gonorrhoeae. The subcomponents of C1q involved in interactions with gonococci and tissue receptors to promote bacteremic spread of gonococci in rat pups have not been elucidated. However, we hypothesize that C1q binds to PID isolates via an antibody-independent mechanism (15). C1q binding to gonococci may result in C1q-mediated attachment of N. gonorrhoeae to C1q receptors of rat pup tissues (16) and an assembly of membrane attack complex in a nonlethal configuration to PID (serr) isolates (8). In turn,
PID ISOLATES OF N. GONORRHOEAE
VOL. 65, 1997
attachment would allow efficient colonization, while serum resistance would allow survival and systemic spread. Auxotypes and serovars are stable properties that are useful for typing N. gonorrhoeae strains and permit analysis of epidemiology and patterns of geographical spread of gonococcal strains (9). Some studies have suggested that particular auxotypes and serotypes may be associated more frequently with PID strains; nevertheless, serotyping of gonococcal isolates has not helped in the diagnosis of PID (7). Pili, Opa, and ironbinding proteins are important for gonococcal virulence, but these features have not been useful in identifying strains capable of causing PID (7). The present finding offers a novel and attractive option for further basic and clinical investigations of PID. The use of the rat pup model, serr assays, and sac-4 PCR may allow efficient detection of high-risk PID strains and patients and should result in more-aggressive treatment. Identification and treatment of patients at high risk for the development of PID (based on characteristics of gonococcal strains in the early stage of infection) could reduce the number of patients with serious complications (5, 7) and may significantly decrease the economic burden of PID. Further investigations of mechanisms of gonococcal PID are in progress and may in turn contribute to the development of a vaccine against gonococcal PID. ACKNOWLEDGMENTS This study was supported by grants from the John Sealy Foundation and The James W. McLaughlin Fellowship Fund and in part by a grant from the NIH (NIDDKR01 DK42029). We thank Rosemary Martinez, Linda Morrow Poindexter, and Mac McConnell for their editorial work. Statistical evaluation was performed by Elbert Whorton, Department of Preventive Medicine and Community Health, Division of Epidemiology & Biostatistics, UTMB, Galveston, Tex. REFERENCES 1. Alberti, S., G. Marques, S. Camprubi, S. Merino, J. M. Tomas, F. Vivanco, and V. J. Benedi. 1993. C1q binding and activation of the complement classical pathway by Klebsiella pneumoniae outer membrane proteins. Infect. Immun. 61:852–860. 2. Buchan, H., M. Vessey, M. Goldacre, and J. Fairweather. 1993. Morbidity following pelvic inflammatory disease. Br. J. Obstet. Gynecol. 100:558–562. 3. Cannon, J. G., T. J. Lee, L. F. Guymon, and P. F. Sparling. 1981. Genetics of serum resistance in Neisseria gonorrhoeae: the sac-1 genetic locus. Infect. Immun. 32:547–552. 4. Cannon, J. G., T. M. Buchanan, and P. F. Sparling. 1983. Confirmation of association of protein I serotype of Neisseria gonorrhoeae with ability to cause disseminated infection. Infect. Immun. 40:816–819. 5. Department of Health and Human Services and Public Health Services. 1991. Healthy people 2000, p. 496–508. U.S. Department of Health and Human Services and Public Health Services, Washington, D.C. 6. Eisenstein, B. I., and A. T. Masi. 1981. Disseminated gonococcal infection (DGI) and gonococcal arthritis (GCA). I. Bacteriology, epidemiology, host factors, pathogen factors, and pathology. Semin. Arthritis Rheum. 10:155– 172. 7. Expert Committee on Pelvic Inflammatory Disease. 1991. Pelvic inflammatory disease research directions in the 1990s. Sex. Transm. Dis. 1:46–64. 8. Harriman, G. R., E. R. Podack, A. I. Braude, L. C. Corbeil, A. F. Esser, and J. G. Curd. 1982. Activation of complement by serum-resistant Neisseria gonorrhoeae: assembly of the membrane attack complex without subsequent cell death. J. Exp. Med. 156:1235–1249. 9. Knapp, J. S., E. G. Sandstrom, and K. K. Holmes. 1985. Overview of epidemiological and clinical applications of auxotype/serovar classification of
Editor: V. A. Fischetti
2099
Neisseria gonorrhoeae, p. 6–12. In G. K. Schoolnik, G. F. Brooks, S. Falkow, C. E. Frasch, J. S. Knapp, J. A. McCutchan, and S. A. Morse (ed.), The pathogenic Neisseriae: Proceedings of the Fourth International Symposium, Asilomar, Calif. American Society for Microbiology, Washington, D.C. 10. Mandrell, R. E., A. J. Lesse, J. V. Sugai, M. Shero, J. M. Griffiss, J. A. Cole, N. J. Parsons, H. Smith, S. A. Morse, and M. A. Apicella. 1990. In vitro and in vivo modification of Neisseria gonorrhoeae lipooligosaccharide epitope structure by sialylation. J. Exp. Med. 17:1649–1664. 11. McCutchan, J. A., D. N. A. Katzanstein, D. N. A. Norquist, G. Chikami, A. Wunderlich, and A. I. Braude. 1978. Role of blocking antibody in disseminated gonococcal infection. J. Immunol. 121:1884–1888. 12. McShan, W. M., R. P. Williams, and R. A. Hull. 1987. A recombinant molecule from a disseminating strain of Neisseria gonorrhoeae that confers serum bactericidal resistance. Infect. Immun. 55:3017–3022. 13. Morse, S. A., and K. K. Holmes. 1994. Gonococcal infections, p. 670–685. In P. D. Hoeprich, M. C. Jordan, and A. R. Ronald (ed.), Infectious diseases. J. B. Lippincott Co., Philadelphia, Pa. 14. Nowicki, S. 1991. Possible mechanisms of serum resistance of Neisseria gonorrhoeae from disseminating infections, p. 621–626. In M. Achtman, P. Kohn, C. Marchal, G. Morelli, A. Seiler, and B. Thiesen (ed.), Neisseriae 1990: pathogen-host interactions. Walter de Gruyter, New York, N.Y. 14a.Nowicki, S., et al. Unpublished data. 14b.Nowicki, S., et al. U.S. patent O8/532,674 pending. 15. Nowicki, S., B. Nowicki, and M. Martens. 1993. Complement C1q on the surface of gonococcal cells from pelvic inflammatory disease, abstr. 147, p. 125. In Program and Abstracts of the 3rd World Congress for Infectious Diseases, Acapulco, Mexico. International Society for Infections Diseases in Obstetrics and Gynecology, Washington, D.C. 16. Nowicki, S., B. Nowicki, and M. Martens. 1993. Species-dependent attachment of Neisseria gonorrhoeae to the uterus, abstr. 19. Infectious Diseases Society for Obstetrics and Gynecology Annual Meeting, Stowe, Vt. 17. Nowicki, S., B. Nowicki, and M. Martens. 1994. Host factors in the attachment of gonococcal cells to pelvic tissue, p. 292–294. In E. W. Ades, R. F. Rest, and S. A. Morse (ed.), Microbial pathogenesis and immune response. N. Y. Academy of Science, New York, N.Y. 18. Nowicki, S., M. Martens, A. Kaul, and B. Nowicki. 1994. Gonococcal attachment to human ovarian tissue, p. 730–734. In C. J. Conde-Glez, S. Morse, P. Rice, F. Sparling, and E. Calderon (ed.), Pathobiology and immunobiology of Neisseriaceae. National Institute of Public Health, Cuernavaca, Morelos, Mexico. 19. Nowicki, S., M. Martens, and B. Nowicki. 1995. Gonococcal infection in a nonhuman host is determined by human complement C1q. Infect. Immun. 63:4790–4794. 20. Parsons, N. J., P. V. Patel, E. L. Tan, J. R. C. Andrade, C. A. Nairn, M. Goldner, J. A. Cole, and H. Smith. 1988. Cytidine 59-monophospho-N-acetyl neuraminic acid and a low molecular weight factor from human blood cells induce lipopolysaccharide alteration in gonococci when conferring resistance to killing by human serum. Microb. Pathog. 5:303–309. 21. Rice, P. A. 1989. Molecular basis for serum resistance in Neisseria gonorrhoeae. Clin. Microbiol. Rev. 2:S112–S117. 22. Rice, P. A., and J. Schachter. 1991. Pathogenesis of pelvic inflammatory disease. What are the questions? JAMA 266:2587–2593. 23. Rice, P. A., W. M. McCormack, and D. L. Kasper. 1980. Natural serum bactericidal activity against Neisseria gonorrhoeae isolates from disseminated, locally invasive, and uncomplicated disease. J. Immunol. 124:2105–2109. 24. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463–5467. 25. Schoolnik, G. K., T. M. Buchanan, and K. K. Holmes. 1976. Gonococci causing disseminated gonococcal infection are resistant to the bactericidal action of normal human serum. J. Clin. Invest. 58:1163–1173. 26. So, M., W. S. Dallas, and S. Falkow. 1978. Characterization of an Escherichia coli plasmid encoding for synthesis of heat-labile toxin: molecular cloning of the toxin determinant. Infect. Immun. 21:405–411. 27. Sweet, R. L., M. Blankfort-Doyle, M. O. Robbie, and J. Schachter. 1986. The occurrence of chlamydial and gonococcal salpingitis during the menstrual cycle. JAMA 255:2062–2064. 28. Washington, A. E., W. Cates, and A. A. Zaidi. 1984. Hospitalizations for pelvic inflammatory disease. Epidemiology and trends in the United States, 1975 to 1981. JAMA 251:2529–2533. 29. West, S. E. H., and V. L. Clark. 1989. Genetic loci and linkage associations in Neisseria gonorrhoeae and Neisseria meningitidis. Clin. Microbiol. Rev. 2:S92–S103.
