Microbial Pathogenesis 1991 ; 10 : 373-384

Effect of human chorionic gonadotropin (hCG) on Neisseria gonorrhoeae invasion of and IgA secretion by human fallopian tube mucosa Gary L . Gorby,' ,2 * Christopher M . Clemens, 2 Lee R . Barley2 and Zell A . McGee2 ' Omaha VA Medical Center and Department of Medical Microbiology Creighton University School of Medicine, Omaha, NE 68131, U .S.A . and 2 Center for Infectious Diseases, Diagnostic Microbiology and Immunology, University of Utah School of Medicine, Salt Lake City, Utah 84132, U .S .A . (Received October 17, 1990 ; accepted in revised form January 18, 1991)

Gorby, G . L . (Omaha VA Medical Center and Dept of Medical Microbiology Creighton University School of Medicine, Omaha, NE 68131, U .S .A .), C . M . Clemens, L . R . Barley and Z . A . McGee . Effect of human chorionic gonadotropin (hCG) on Neisseria gonorrhoeae invasion of and IgA secretion by human fallopian tube mucosa . Microbial Pathogenesis 1991 ; 10 : 373-384 . The possible effect of human chorionic gonadotropin (hCG) on the mucosal immune response and susceptibility of the fallopian tube mucosa to invasion by Neisseria gonorrhoeae (gonococci) was investigated in the fallopian tube organ culture (FTOC) model . Immunohistochemical and radioreceptor assay techniques showed specific high affinity binding of hCG in vitro to the -g apices of non-ciliated fallopian tube cells (K d approximately 10 M) . Continuous exposure of the FTOC mucosa to hCG during infection with gonococci resulted in a marked increase (6- to 1 5-fold) in IgA secretion and significantly reduced gonococcal invasion (invasion score range 0 .7 to 1 .75) compared to infected control tissue which was not exposed to hCG (invasion score range 2 .9 to 4 .95, P S 0 .01) . By contrast, exposure of the mucosa to hCG during the 24 h preceding gonococcal infection followed by the removal of hCG from the system at the time of infection resulted in enhanced gonococcal invasion (invasion score range 7 .95 to 9 .7, P < 0 .001) . We conclude that hCG can modulate the mucosal immune response and susceptibility of fallopian tube epithelium to gonococcal invasion . Key words : gonococci ; human chorionic gonadotropin ; human fallopian tube ; IgA .

Introduction The human fallopian tube mucosa consists of ciliated and non-ciliated columnar cells .' In the fallopian tube organ culture (FTOC) model of gonococcal salpingitis, Neisseria gonorrhoeae (gonococci) first attach to the non-ciliated cells . Next, the surfaceassociated organisms are endocytosed and transported to the base of the cells where they are exocytosed into the subepithelial space .' From this location the invading microbes may elicit a local inflammatory response ultimately resulting in serious sequelae such as pelvic inflammatory disease (PID), infertility, and ectopic pregnancy ;

*Author to whom reprint requests should be addressed at : Section of Infectious Diseases, Omaha VA Medical Center, 4101 Woolworth Ave., Omaha, NE 68105, U .S .A . 0882-4010/91/050373+12 503 .00/0

© 1991 Academic Press Limited



374

G . L . Gorby et al.

or gain access to the bloodstream through which they spread to cause disseminated gonococcal infection (DGI) . 3 Factors that alter gonococcal attachment to the epithelium or influence the apparently misguided transport functions of the mucosal cells could have a profound effect on the vulnerability of the fallopian tube to gonococcal infection . Cases of PID and DGI in women are most likely to occur around the time of menses,' but non-gonococcal, non-chlamydial PID does not display this temporal clustering . 4 These considerations have led us and others' to hypothesize a temporal variation in the susceptibility of the fallopian tube mucosa to gonococcal invasion that is related in some way to the menstrual cycle via the hormonal milieu . Variations in the steroid hormonal milieu are known to result in morphological changes in the epithelial cells of the fallopian tube mucosa,' and in some animal models the hormonal milieu alters immunoglobulin secretion by endometrial cells .' Little is known about non-steroid endocrine effects on fallopian tube epithelial function s (e .g . peptide hormones such as the gonadotropins) . Because of the apparent active participation of the fallopian tube epithelium in gonococcal PID, the menstrual variation in the incidence of gonococcal PID and DGI, and the possible effect of the hormonal milieu on mucosal IgA production we performed a series of experiments to : (1) quantitate gonococcal invasion of the fallopian tube mucosa, (2) investigate the effect of human chorionic gonadotropin (hCG) on gonococcal invasion, and (3) measure the effect of hCG on the secretion of IgA by the fallopian tube mucosa . These experiments were done to determine if the hormonal milieu plays a part in the susceptibility of the fallopian tube to gonococcal infection .

Results Immunohistochemical detection of hCG binding Since it was not known whether hCG could bind specifically to fallopian tube epithelium, we performed immunohistochemical studies on fallopian tube tissue that had been incubated with hCG in vitro . Fallopian tube mucosa that was not exposed to hCG showed no evidence of immunostaining whether or not pre-immune or antihCG rabbit serum served as the primary antibody [Fig . 1(C)] . When pre-immune rabbit serum served as primary antibody, hCG-exposed tissue similarly revealed no immunostaining [Fig . . 1(B)] . However, when anti-hCG rabbit serum was used as the primary antibody, hCG-exposed tissue showed dense immunostaining of the apices of non-ciliated cells only [Fig . 1(A)] . None of the ciliated cells displayed evidence of hCG binding . hCG radioreceptor assay of fallopian tube tissue To determine whether the hCG binding observed by immunohistochemistry was 1251 receptor mediated, a radioreceptor assay was performed . Specific binding of labeled hCG to fallopian tube homogenates was 0 .65% while non-specific binding was 3 .6% . Despite the low amount of specific binding, a dose-dependent inhibition of radiolabeled hormone binding by unlabeled 'cold' hCG was observed [Fig . 2(A)] . This dose-response curve is characteristic of receptor-mediated binding . The Woolf -' plot [Fig . 2(B)] revealed a dissociation constant (Kd ) of 1 .58xi0 M when outliers .73x10-'M were excluded or 6 when all points were included . Effects of the hormonal milieu on gonococcal invasion After establishing the presence of hCG receptors in the fallopian tube mucosa, we investigated the effects of hCG on gonococcal invasion of FTOC . Four experimental



hCG effects on gonococcal invasion

375

Fig . 1 . Immunohistochemical localization of hCG binding to FTOC. (A) Mucosa was incubated with hCG for 40 h and stained with rabbit anti-hCG (primary antibody) followed by goat anti-rabbit IgG (bridging antibody) followed by peroxidase rabbit anti-peroxidase (PAP complex) . Note the dense immunostaining of the apices of non-ciliated fallopian tube cells and lack of staining of the ciliated cells . (B) Mucosa was incubated as above with hCG but rabbit pre-immune sera was substituted as the primary antibody in the immunostaining procedure . Neither the ciliated nor non-ciliated cells displayed immunostaining which indicates that the staining in (A) is due to a specific interaction of the anti-hCG antisera with the hCGexposed tissue . (C) Mucosa was incubated for 40 h but was not exposed to any hCG . The immunostaining procedure was identical to (A) . There was no evidence of immunostaining in this treatment group which indicates that the staining in (A) is due to the detection of hCG bound to non-ciliated cells . Bar = 10 µm .



G, L . Gorby et al.

376

(A)

100 90 80 70

0 60 C

VC m

50 40 30 20 10 0

I

10

1000

100

IE4

IE5

1E6

IE7

Cold hCG (pg) 400 (B )

300

m

4

200 --------------------------100

I I I I I

100

200

300

400

500

600

PM/l free hCG

Fig . 2 . (A) Percentage of the maximum radiolabel bound versus the amount of competing unlabeled hCG . (B) Woolf plot of hCG radioreceptor assay data . Dotted lines = 95% confidence interval .

groups were subjected to quantitative invasion scoring : non-infected control tissue (C), infected control tissue (IC : no hCG exposure), infected tissue exposed to hCG continuously (HC), and infected tissue that had been exposed to hCG only for the 24 h that preceded infection with gonococci (HP) . Mean invasion scores among the various experimental groups were significantly different (P < 0 .05) from each other (Table 1) except for the non-infected control and continuous hCG groups in experiment 1 (P=0 .11) . The non-infected control group [Fig . 3(A)] had a low-level rather than

Table 1 Effect of hCG on gonococcal endocytosis in a FTOC system Group' mean invasion score+ SE' Trial 1 2 3

Infection duration

C

HC

HP

IC

32 h 32 h 40 h

0 .3+0.21 0 .6±0.13 0 .4+0.11

0 .7+0.15 1 .6±0 .26 1 .7+0.25

9 .7+1 .8 8 .5±0.73 8 .0+0.65

3 .1+0 .54 2 .9±0.37 4 .3+0.33

'C : the non-infected control tissue; HC : infected tissue exposed to hCG continuously ; HP: tissue that was exposed to hCG prior to infection only ; IC : infected control tissue that was not exposed to hCG . b SE : standard error .

hCG effects on gonococcal invasion

377

Fig . 3 . Appearance of FTOC depending on infection status and hCG exposure . (A) Control tissue (not infected) demonstrating normal appearance of FTOC after a total of 56 h in culture . (B) Infected tissue not exposed to hCG . Note the presence of intracellular gonococci in vacuoles (arrows) at the bases of several non-ciliated cells . (C) Infected tissue exposed to hCG continuously . Note the conspicuous absence of intracellular gonococci in comparison to the tissue depicted in (B) . (D) Infected tissue exposed to hCG only for the 24 h preceding infection . Note the distortion of the cells, which are engorged with intracellular gonococci (arrows), and a mass of gonococci in the subepithelial region . Bar = 10 pm .

zero invasion score because of rare misinterpretation of intracellular structures that appeared similar to an intracellular gonococcus even though no bacteria were present . The infected control groups [Fig . 3(B)] possessed mean invasion scores significantly higher than the non-infected control group (P < 0 .0008 in all three experiments) . Exposing the FTOC to hCG continuously [Fig . 3(C)], decreased the invasion scores compared to untreated infected controls resulting in a two- to four-fold lower mean score (experiment 1 : P = 0 .0005, experiment 2 : P = 0 .010, and experiment 3 : P < 0 .0001 ) . By pre-treating the tissue with hCG for 24 h and then removing it just prior to infection [Fig . 3(D)], the gonococcal invasion of non-ciliated fallopian tube cells was enhanced two- to three-fold compared to infected control tissue (P < 0 .001 in all three experiments) .

Effect of hCG on FTOC IgA secretion The IgA levels in the FTOC media were measured for each of the groups in the invasion experiments relative to the non-infected control (Table 2) . IgA levels were very similar among the infected control and the pre-infection hCG-exposed group (P > 0 .71 at 32 and 40 h), but the infected tissue exposed to hCG continuously displayed a marked increase in IgA content : roughly six-fold higher at 32 h and 1 5-fold higher at 40 h for comparisons to the other groups (P < 0 .0001 at 32 and 40 h) . 95% confidence intervals (C .I .) for the difference between the medians of the pre-infection hCGexposed group and the infected control group overlap 0, but the 95 .5% C .I . for the



G . L . Gorby et al.

378

Table 2

Relationship of IgA levels in FTOC fluid to the hormonal

milieu IgA level ng/ml (95 .8% C . I .)' Trial

Infection duration

C

1 2

32 h 40 h

-

HC

HP

IC

482 (390-576) 79 .3 (44-119) 78.6 (41-108) 811 (610-950) 47 .0 (30 .5-84) 57.8 (24-81)

'See key to abbreviations under Table 1 . Values represent IgA levels in excess of the non-infected control group . Thus, there are no values reported for the non-infected control group.

difference between the medians to FTOC exposed to hCG continuously and the infected control was 320 to 500 ng/ml at 32 h and 638 to 905 ng/ml at 40 h . The 95 .5% C .I . for the difference between the medians of infected tissue exposed to hCG continuously and tissue exposed to hCG prior to infection only was 310 to 490 ng/ml at 32 h and 607 to 895 ng/ml at 40 h . When IgA levels were determined based on a standard curve constructed from secretory IgA serially diluted in fresh medium rather than 32 h FTOC medium, the median IgA levels ranged from 39 .4 ng/ml (95 .8% C .I . : 18 .9 to 65 .6) for non-infected control tissue to 1180 ng/ml (95 .8% C .I . : 955 to 1408) for infected tissue exposed to hCG continuously . In comparison to the first curve, this second standard curve was linear over a narrower range making estimates of IgA levels from either extreme of the curve inherently less accurate than the values reported above . IgA levels in noninfected control tissue were not significantly different from infected tissue not exposed to hCG or infected tissue exposed to hCG only prior to infection (P > 0 .24 for pairwise comparisons of all three groups at both 32 and 40 h) . Only the infected tissue exposed to hCG continuously had significantly higher IgA levels in comparison to all the other groups at both 32 and 40 h (P < 0 .0001) . Thus regardless of the standard curve used to calculate the IgA concentrations, the combination of gonococcal infection and hCG exposure causes a statistically significant increase in the IgA content of the FTOC medium as shown by hypothesis testing compared to infection alone . This difference is relatively large according to confidence interval analysis .

Discussion It is well-known that hCG and luteinizing hormone (hLH) share the same cell-surface receptors on theca, granulosa, and luteal ovarian cells .' Luteinizing hormone, through these specific receptors, controls the development of the Graafian follicle, ovulation, and the transformation of the follicle into the corpus luteum .' Receptor-bound hormone (hLH or hCG) is internalized (receptor-mediated endocytosis) by ovarian cells ."' Our data suggest that hCG binds in a similar specific manner to fallopian tube epithelium . The low amount of specific binding in our radioreceptor assay can be accounted for by the length of time tissue was stored and a freeze-thaw cycle effect . In additional experiments with rat ovarian tissue, one freeze-thaw cycle caused a six-fold reduction in specific binding under identical assay conditions (data not shown) . This suggests that degradation of the hCG receptor prior to assay was significant and might have affected Kd determinations . Evidence of hCG binding to non-ovarian tissue has been provided previously by Ziecik et a/L who detected low capacity, high affinity receptors for hLH/CG in the porcine uterus ." Recently, hCG receptors have been identified in the rabbit uterus," the non-pregnant human uterus, 12 and in the human fallopian



hCG effects on gonococcal invasion

379

12 tube . Others have detected hCG or a related molecule bound to human endometrial tissue during pregnancy . 13 These molecules were bound in an orientation-specific manner that differed from the orientation seen in typical receptor-mediated hCG binding . It is unclear which of these mechanisms of binding is represented by our data . The K d value derived from our receptor assay is higher than the Kd of 10 -10 M reported for typical ovarian hCG receptor binding,' which suggests an interaction of relatively lower affinity in the fallopian tube . It is possible also that the dose-response curve that we observed represents lower affinity hCG binding to a receptor for one of 1014 the other glycoprotein hormones with which hCG shares its alpha subunit . Our invasion studies have demonstrated that hCG in high concentrations is capable of blocking gonococcal invasion of fallopian tube non-ciliated cells . The invasion scoring system has been a reliable means of quantifying this effect . In contrast, exposing the FTOCs to hCG for 24 h prior to infection and then removing the hormone just before inoculation with gonococci resulted in marked enhancement of invasion scores . There are a number of potential explanations for these observations . (1) hCG down-regulates an as yet uncharacterized gonococcal receptor, and removal of hCG subsequently results in a rebound up-regulation of the receptor (Receptor Modulation Hypothesis) . (2) hCG competes with a gonococcal hCG-like molecule for an hCG receptor which leads to receptor-mediated endocytosis (Biological Mimicry Hypothesis) ." (3) hCG modifies intracellular transport mechanisms, including the endocytosis of gonococci and secretion of IgA (Intracellular Traffic Hypothesis) . (4) hCG alters gonococcal attachment which has a resultant effect on subsequent invasion . Although the Receptor Modulation Hypothesis is a viable explanation for our observations, hard data are lacking, and substantiation will require further study of epithelial cell membrane components under a variety of hormonal conditions . Similarly, these studies did not evaluate the possible effect of hCG on attachment . In contrast, the Biological Mimicry Hypothesis is given some support by the work of a number of investigators who described the production of an hCG-like molecule by a number of bacterial species . 16-24 Recent work by Robinson and McGee using gold immunoprobelabeling and western blot techniques has suggested the presence of an hCG-like molecule in some gonococcal strains ." Because these techniques are susceptible to false-positive results via cross-reactive antibody/antigen interactions, one must reserve judgement on the presence of a gonococcal hCG-like molecule until a candidate molecule is isolated and tested for its ability to bind to hCG receptors . Biological Mimicry of hCG is an attractive hypothesis for the pathogenesis of gonococcal salpingitis since the hCG receptor undergoes receptor-mediated endocytosis, $ and hCG/LH receptor numbers in the ovary increase dramatically following the midcycle hLH surge .' During the luteal phase of the menstrual cycle hLH serum levels fall, resulting in a large pool of unbound hCG/LH receptors prior to menses .' This physiology could explain the menstrual clustering of gonococcal PID and our observations of hCG effects on the FTOC model . In contrast to previous studies of FTOC IgA secretion which demonstrated a marked rise in IgA secretion by infected FTOCs compared with controls, 25 we observed a milder rise in IgA content in the culture fluid of infected tissue compared to noninfected control tissue unless hCG was present continuously . In the latter case, IgA content increased markedly . It is not clear whether the hormonal milieu plays merely a permissive role in the IgA response to gonococcal infection, or if hCG alone is capable of stimulating IgA secretion . It is possible that our results differ from the previous report 25 because of differences in patient hormonal status at the time the fallopian tube was obtained or because a different gonococcal strain was used . Further study with highly purified hormones, gonadotropins and steroids, in the FTOC system with



380

G . L . Gorby et al .

tissue obtained during different phases of the menstrual cycle will be required to better define these effects . Understanding the effect of the hormonal milieu on IgA secretion is important especially in the context of gonococcal production of an IgA1 protease ." The role of IgA in the observed effects of hCG on gonococcal invasion in our FTOC model is uncertain . Theoretically, secretory IgA that is immunoreactive with gonococcal attachment or invasion factors could reduce invasion . It is well-established that human sera from individuals not previously infected with gonococci contain antibodies against N . gonorrhoeae, presumably because of previous exposure to other Neisseria spp . such as N. meningitidis ." Perhaps the same is true of genital secretions . 28 Secretory IgA from human colostrum does react with gonococcal components on Western blots of whole organism lysates (Gorby, unpublished observations), and because of the common mucosal immune system secretory IgA that is immunoreactive is likely to be present in genital secretions as well . The significance of this immunoreactivity in vivo or in vitro is unknown, and unfortunately the quantity of FTOC fluid remaining after IgA assay was insufficient to allow us to examine the immunoreactivity of the IgA produced in these particular experiments . However, the IgA levels observed in the FTOC media at the end of the experimental infections are significantly lower than physiological secretory IgA concentrations at other mucosal surfaces . 29 It is likely that fallopian tube secretory IgA levels are much higher in vivo, because the entire fallopian tube secretion is not diluted into 60 ml of culture medium in vivo . Moreover, during the early phases of the FTOC infection (attachment), the IgA levels are very low because fresh medium is added to the system at the time of gonococcal inoculation into the FTOC . Thus the effects of the hormonal milieu on IgA secretion may be important in vivo, but IgA is not likely to contribute much to the reduction of gonococcal invasion in our system . Certainly, alterations of IgA levels cannot explain the enhancement of invasion when fallopian tube tissue is exposed to hCG prior to infection . Elucidating the roles of secretory IgA and the gonococcal IgA1 protease in host susceptibility to gonococcal PID will require testing of the invasion capacity of IgA1 protease-deficient and protease-producing isogenic variants in the FTOC system in the presence of physiological concentrations of immunoreactive secretory IgA . The endocytosis of gonococci, rather than being a traitorous act by the fallopian tube mucosa, may represent an attempt by the non-ciliated epithelial cells to 'process' foreign antigens so that the body may mount an immune response . 3 Taken together, the observed effects of hCG on IgA content and endocytosis of gonococci may reflect hormonal modulation of the genital tract immune system . Theoretically, it is advantageous for the reproduction of the species to minimize the presentation of foreign antigens to the genital mucosal immune system during times when elevated levels of hCG or hLH are present (e .g . early pregnancy or midcycle fertilization) by augmenting IgA secretion and reducing endocytosis of antigens . This modulation would minimize the generation of a cell-mediated immune response which can lead to pre-eclampsia or rejection of the developing fetus . 3o,3 ' These considerations lend credence to the Intracellular Traffic Hypothesis of hCG action on the fallopian tube . Further definition of the effects of the hormonal milieu on the antigen-processing and IgA-secreting functions of the genital tract epithelium has important implications for understanding the control of intracellular molecular traffic ; the pathogenesis of infectious diseases ; and the function of mucosal immunity in fertility, sterility, and pathologic conditions of pregnancy . Materials and methods

Media . Solid media for cultivating Neisseria gonorrhoeae, 'GClso agar', consisted of GC agar base (Difco Laboratories, Detroit, Michigan) supplemented with 2% IsoVitaleX (BBL Micro-



hCG effects on gonococcal invasion

381

biology Systems, Cockeysville, Maryland) . Media for FTOC consisted of Eagle Minimum Essential Medium containing Earle's salts and L-glutamine (GIBCO Laboratories, Grand Island, New York) that was buffered with 0 .05 M N-2-hydroxyethyl-piperazine-N'-2-ethanesulfonic acid (Sigma, St Louis, Missouri) to pH 7 .45 (HEPES-MEM) .' For pans of the FTOC experiments the HEPES-MEM was supplemented with vanocomycin HCI 5µg/ml (Eli Lilly, Indianapolis, Indiana) and colistimethate sodium 3 pg/ml (Parke-Davis, Morris Plains, New Jersey) . Microorganisms and inoculum preparation . Neisseria gonorrhoeae strain UU1008pilus', 0 was used in these studies . Piliation status was confirmed by negative-staining transmission electron microscopy ." P .II - status was suggested by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) analysis of outer membrane preparations, and failure of three monoclonals directed against different P .Ils (P .Ila, P .Ilc, and P .Ile) to react with outer membrane preparations of the organism in Western blots 32 (antibodies kindly provided by Janne Cannon, University of North Carolina, Chapel Hill) . Strain UU1008 was isolated originally from the fallopian tube of a patient with PID . The organisms were grown and prepared for use in FTOC experiments as previously described .' Fallopian tube sources . Tissue from premenopausal non-pregnant women who were having surgery for dysfunctional uterine bleeding or leiomyomata was used, but tissue from postmenopausal women, women with malignancies, women with a past history of PI D or gonococcal infection was excluded . The tissue was transported on ice in HEPES-MEM containing antibiotics (see below) . Preparation and maintenance of FTOC . The procedures followed were essentially those of McGee et al.' with a few modifications . Fallopian tube pieces were selected from the midportion (ampullary region) of the specimen to control for variable degrees of mucosal redundancy that are seen in the isthmic and fimbriated portions of the fallopian tube . A 2 .5 mm diameter hair transplant punch biopsy tool (George Tiemann, Plainview, New York) was used to cut fallopian tube pieces of equal size . The selectivity of mucosal region and method of tube sectioning minimizes variations in surface area and tissue volume between individual pieces, which could affect the relative growth of gonococci in the organ culture fluid and the amount of IgA secreted into the media . Three fallopian tube pieces were placed in each 35x10 mm tissue culture dish (Falcon Plastics, Oxnard, California) which contained 1 ml of HEPES-MEM plus antibiotics . For some experimental groups the medium was supplemented with partiallypurified hCG at a concentration of 500 IU/m1 (Sigma) . This hCG preparation was devoid of estrogen activity according to the manufacturer. This preparation contained no detectable IgG, IgM, or IgA according to ELISA testing (data not shown) . Immunohistochemical detection of hCG binding . FTOCs prepared as above were incubated with HEPES-MEM alone or HEPES-MEM supplemented with hCG for 40 h . The tissue was fixed, 33 embedded in paraffin, and cut into 3-5-µm thick sections . Immunocytochemical detection of tissue-bound hCG was achieved with a peroxidase-anti-peroxidase method (Accurate Chemical and Scientific, Westbury, New York) ." Both the hCG-exposed and nonexposed tissues were processed according to the manufacturer's protocol using anti-hCG antibody or pre-immune serum as the primary antibody . The slides were coded following processing and examined without knowledge of the treatment group being viewed via light microscopy. Fifty sections were examined . Positive immunostaining was identified as a dark brown precipitate . hCG radioreceptor assay of fallopian tube tissue . Thirteen fallopian tubes were placed in a storage buffer consisting of 0 .1 M phosphate buffered saline (PBS) containing 0 .05 M EDTA, 0 .05 M N-ethylmaleimide, and 0 .002 M phenylmethylsulfonylfluoride (Sigma) . 35 The tissue in buffer was quick-frozen in acetone-dry ice and stored up to 3 months at -70°C . Highly purified 1251 labelled hCG was a generous gift of Dr William O'Dell (University of Utah School of Medicine, Salt Lake City, Utah) . The assay was performed according to the method of Keinanen et al. 15 Each assay tube contained the following : (1) 250 µl of tissue homogenate ; (2) 125 µl of assay buffer containing various amounts of unlabeled hCG (1 ug, 1 x 10 5 pg, 1 x 10 4 pg, 3 x 10 3 pg, 1 x 10 3 pg, 300 pg, 100 pg, 30 pg, 0 pg) to generate a dose-response curve ; (3) 125 µl of assay buffer containing 10000 cpm 125 1-labeled hCG (35 pg) . The assay was run in triplicate . Results were expressed as a percentage of the maximum radiolabel bound versus the



382

G . L . Gorby et al

amount of unlabeled hCG . A dissociation constant (K d ), which is an index of the receptor affinity for hCG, was estimated using the Woolf plot . 38

Effect of hCG on gonococcal invasion of FTOC . The organ cultures were split into four treatment groups . All four groups were incubated in HEPES-MEM plus antibiotics for the first 24 h . Two of the groups were supplemented also with hCG . After the first 24 h the cultures were washed twice for 30 min each, and then fresh medium identical to the wash medium was added . This 24 h incubation step followed by washing allows any contaminating serum immunoglobulins to diffuse out of the tissues and be washed away ."" Three of the groups, including one exposed to hCG for 24 h, were washed with HEPES-MEM alone . The fourth group was washed with HEPES-MEM supplemented with hCG . One of the groups that had not been exposed to hCG was designated the uninfected control . The other groups were infected with a standard gonococcal inoculum .' The infection was allowed to proceed for 3240 h, when the FTOC medium was sampled, and serial 10-fold dilutions were plated on chocolate agar . Organ cultures that were contaminated or whose media colony counts deviated more than three-fold from the mean of all four groups were excluded from invasion scoring . The remaining FTOC medium was filter-sterilized and frozen at -70°C for future IgA quantitation . Tissue from FTOCs was fixed, embedded, and sectioned as above . The sections were stained according to the method of Sowter and McGee 33 which stains gonococci red, epithelial cell cytoplasm pale green, and nuclei/nuclear debris blue so that intracellular microbes can be differentiated from cellular elements . In one experiment the tissue was fixed in modified Karnovsky's fixative (2% paraformaldehyde, 1 % glutaraldehyde in 0 .1 M phosphate buffer pH 7 .4) on ice for 0 .5 h . The tissue was then submitted for Epon embedding, thick sectioning (1 lim), and toluidine blue staining . Tissue stained by either method was mapped, and the degree of invasion in each group was assessed . Initially the slides were coded so that the observer did not know the identity of the experimental group being evaluated . Each tissue section on a given slide was mapped into suitable segments delineated with an ocular micrometer reticle divided into 100 parts (10xeyepiece and 100xoil immersion objective) . At least 60 segments were mapped per treatment group . Twenty mapped segments per treatment group were selected randomly and identified on the map . The degree of gonococcal invasion was determined for each of the 20 segments . A score of 1 was given for any 5 unit linear distance along the segment that contained intracellular gonococci . Thus, a given 100 unit mucosal segment received a score between 0 and 20 . A mean segment score was calculated (n = 20) for each treatment group . The Kruskal-Wallis analysis of variance and Wilcoxon rank sum test were employed to determine if the invasion scores for each treatment group differed significantly from the others ." IgA ELISA . A 'sandwich' type ELISA for IgA was devised which utilized rabbit anti-human IgA (Sigma 1-6008 ; IgG fraction) as an antigen capture antibody and affinity-purified goat anti-human IgA-horseradish peroxidase conjugate (Sigma, alpha chain specific ; A-7032) diluted in 0 .1 M PBS 1 % bovine serum albumin (BSA) as the detection antibody . Immunodiffusion testing of these antibodies revealed no cross-reactivity with IgG or IgM (data not shown) . IgA from human colostrum (Sigma ; I-1010) served as a standard . Coating of the plate with the antigen capture antibody took place overnight at 4°C in a 0 .1 M carbonate buffer pH 9 .6 . Wash buffer consisted of 0 .1 M PBS and 0 .05% (v/v) Tween 20 (Sigma) pH 7 .2, and the substrate solution consisted of 0 .04% o-phenylenediamine, 0 .012% H2O2 in 0 .2 M phosphate-0 .1 M citrate buffer pH 5 .0 . To control for possible interfering substances secreted by or present in the fallopian tube tissue, secretory IgA was diluted to 1 fig/ml in HEPES-MEM 1% BSA and added to an equal volume of non-infected control organ culture fluid that had been incubating for 32 h . Two-fold serial dilutions of this mixture in HEPES-MEM 1% BSA served as the concentrations for construction of a standard curve (range 500 ng/ml to 3 .75 ng/ml) . Experimental organ culture fluid (32 or 40 h) was diluted similarly in serial two-fold fashion (range 1 : 2 to 1 :16) with HEPES-M EM 1 % BSA for measurement of IgA levels in triplicate . Thus an IgA level relative to the 30 h control was calculated for each dilution, and a mean and median were determined . A second standard curve was constructed by diluting secretory IgA in fresh HEPES-MEM 1% BSA instead of 32 h organ culture fluid . IgA levels were calculated using both curves . At each timepoint post-infection, 12 samples per group were measured by ELISA . Because one could not validly assume that the data had a normal distribution, non-parametric statistical methods were used for analysis . Ninety-five per cent confidence intervals (C .I .) were calculated for the medians of each group ." The Kruskall-Wallis analysis of variance and Wilcoxon rank sum test were used to determine if IgA levels of each treatment group differed significantly from the



hCG effects on gonococcal invasion

383

others ." Ninety-five per cent confidence intervals for the difference between medians of individual treatment groups were calculated utilizing a Wilcoxon-based method .' The assay was performed as follows : (1) coating as above ; (2) wash (three times, 3 min/wash) ; (3) antigen capture (2 h at 37°C) ; (4) wash (three times) ; (5) antibody-enzyme conjugate binding (2 h at 24°C) ; (6) wash (three times) ; (7) substrate development at 24°C in the dark ; (8) stop reaction by addition of 50 pl 1 .5 M H 2 SO 4 ; (9) absorbance reading at 492 nm .

This work was supported by NIH grant A1271 01 . The authors wish to thank E . N . Robinson, Jr, for helpful suggestions during the initial phases of this work . We thank W . D . O'Dell and J . Griffin for providing reagents, sharing their equipment, and offering suggestions during the radioreceptor assay portion of the study . We are grateful to our surgical colleagues Dr Phil Clark, Dr Kent Rasmussen, Dr Dan Chichester, Dr Kent Farnsworth, Dr Carl Woolsey for providing us with the tissue required for these experiments . We wish to thank Ellen Gordon and Camilla Christensen for secretarial assistance .

References

1 . McGee ZA, Johnson AP, Taylor-Robinson D . Human fallopian tubes in organ culture : preparation, maintenance, and quantitation of damage by pathogenic microorganisms . Infect Immun 1976, 13 : 608-18 . 2 . McGee ZA, Johnson AP, Taylor-Robinson D . Pathogenic mechanisms of Neisseria gonorrhoeae : observations on damage to human fallopian tubes in organ culture by gonococci of colony type 1 or type 4 . J Infect Dis 1981 ; 143 : 413-22 . 3 . McGee ZA, Gorby GL, Wyrick PB, Hodinka R, Hoffman LH . Parasite-directed endocytosis . Rev Infect Dis 1988; 10(S2) : 311-16 . 4 . Sweet RL, Blankfort-Doyle M, Robie MO, Schacter J . The occurrence of chlamydial and gonococcal salpingitis during the menstrual cycle . JAMA 1986 ; 255 : 2062-4 . 5 . Jansen RPS . Endocrine response in the fallopian tube . Endocr Rev 1984 ; 5 :525-51 . 6 . Wira CR, Sullivan DA, Sandoe CP . Estradiol-mediated control of the secretory immune system in the uterus of the rat . Ann NY Acad Sci 1983; 409 : 534-51 . 7 . Yamato M, Nakano R, Iwasaki M, Ikoma H, Furukawa K . Luteinizing hormone receptors on human ovarian follicles and corpora lutea during the menstrual cycle . Obstet Gynecol 1986 ; 68 : 200-3 . 8 . Conn PM, Conti M, Harwood JP, Dufau ML, Catt KJ . Internalization of gonadotropin-receptor complex in ovarian luteal cells . Nature 1978; 274 : 598-600. 9 . Ahmed CE, Sawyer HR, Niswender GD . Internalization and degradation of human chorionic gonadotropin in ovine luteal cells : kinetic studies . Endocrinology 1981 ; 109 : 1380-7 . 10 . Ziecik AJ, Stanchev PD, Tilton JE . Evidence for the presence of luteinizing hormone/human chorionic gonadotropin-binding sites in the porcine uterus . Endocrinology 1986 ; 119 : 1159-63 . 11 . Jensen JD, Odell WD . Identification of LH/hCG receptors in rabbit uterus . Proc Soc Exp Biol Med 1988 ;189 :28-30. 12 . Reshef E, Lei ZM, Rao ChV, Pridham DD, Chegini N, Luborsky JL . The presence of gonadotropin receptors in nonpregnant human uterus, human placenta, fetal membranes and decidua . J Clin Endocrinol Metab 1990; 70 : 421-30 . 13 . Cruz RI, Anderson DM, Armstrong EG, Moyle WR . Nonreceptor binding of human chorionic gonadotropin (hCG) : detection of hCG or a related molecule bound to endometrial tissue during pregnancy using labeled monoclonal antibodies that bind to exposed epitopes on the hormone . J Clin Endocrinol Met 1987 ; 64: 433-40 . 14 . Puett D . Human choriogonadotropin . Bioessays 1986 ; 4 . 70-5 . 15 . Gorby GL, Barley LR, Clemens CM, McGee ZA . Microbial invasion : a covert activity? Can J Microbiol 1988 ;84:507-12 . 16 . Livingston WC, Livingston AM . Some cultural, immunological, and biochemical properties of Progenitor cryptocides . Trans NY Acad Sci 1974 ; 36 : 569-82 . 17 . Cohen H, Stramp A . Bacterial synthesis of substance similar to human chorionic gonadotropin . Proc Soc Exp Biol Med 1976 ; 152 : 408-10 . 18 . Acevedo HF, Slifkin M, Pouchet GR, Pardo M . Immunohistochemical localization of a choriogonadotropin-like protein in bacteria isolated from cancer patients . Cancer 1978 ; 41 : 1217-29 . 19 . Maruo T, Cohen H, Segal SJ, Koide SS . Production of choriogonadotropin-like factor by a microorganism . Proc Nat] Acad Sci USA 1979 ; 76 : 6622-6 . 20 . Acevedo HF, Koide SS, Slifkin M, Maruo T, Campbell-Acevedo EA . Choriogonadotropin-like antigen in a strain of Streptococcus faecalis and a strain of Staphylococcus simulans : detection, identification, and characterization . Infect Immun 1981, 31 : 487-94 . 21 . Backus BT, Affronti LF . Tumor-associated bacteria capable of producing a human choriogonadotropinlike substance . Infect Immun 1981 ; 32 : 1211-15 .



384

U . L . Ciorby et al.

22 . Acevedo HF, Campbell-Acevedo E, Kloos WE . Expression of human choriogonadotropin-like material in coagulase-negative Staphylococcus species . Infect Immun 1985, 32 : 1211-15 . 23 . Domingue GJ, Acevedo HA, Powell JE, Stevens VC . Antibodies to bacterial vaccines demonstrating specificity for human choriogonadotropin (hCG) and immunochemical detection of hCG-like factor in subcellular bacterial fractions . Infect Immun 1986 ; 53 : 95-8 . 24 . Robinson EN Jr, McGee ZA . Detection of human hormone-like molecules elaborated in vitro by Neisseria gonorrhoeae . J Cell Biochem 1988; suppl 11213 : 30 . 25 . Cooper MD, McGee ZA, Mulks H, Koomey JM, Hindman TL . Attachment to and invasion of human fallopian tube mucosa by an IgAl protease-deficient mutant of Neisseria gonorrhoeae and its wild-type parent . J Infect Dis 1984 ; 150 : 737-44 . 26 . Plaut AG, Gilbert JV, Artenstein MS, Capra JD . Neisseria gonorrhoeae and Neisseria meningitidis : extracellular enzyme cleaves human immunoglobulin A . Science 1975 ; 190 : 1103-5. 27 . Hicks CB, Boslego JW, Brandt B . Evidence of serum antibodies to Neisseria gonorrhoeae before gonococcal infection . J Infect Dis 1987 ; 155 : 1276-81 . 28 . Ison CA, Hadfield SG, Bellinger CM, Dawson SG, Glynn AA . The specificity of serum and local antibodies in female gonorrhoeae . Clin Exp Immun 1986 ; 65 : 198-205. 29 . Vincent C, Revillard JP . Sandwich-type ELISA for free and bound secretory component in human biological fluids. J Immunol Methods 1988 ; 106 : 153-60 . 30 . Klonoff-Cohen HS, Savitz DA, Cefalo RC, McCann MF . An epidemiologic study of contraception and preeclampsia . JAMA 1989 ; 262 : 3143-7 . 31 . Beer AE . Immunology, contraception, and preeclampsia . JAMA 1989 ; 262 : 3143-7 . 32 . Black WJ, Schwalbe RS, Nachamkin I, Cannon JG . Characterization of Neisseria gonorrhoeae protein II phase variation by use of monoclonal antibodies . Infect Immun 1984 ; 453-7 . 33 . Sowter C, McGee ZA . Evaluation of a new technique for the demonstration of gonococci and other micro-organisms in host cells . J Clin Pathol 1976 ; 29 : 433-7 . 34 . Sternberger LA, Hardy PH, Cuculis JJ, Meyer HG . The unlabeled antibody enzyme method of immunohistochemistry-preparation and properties of soluble antigen-antibody complex (Horseradish peroxidase-anti horseradish peroxidase) and its use in identification of spirochetes . J Histochem Cytochem 1970; 18 : 315 . 35 . Keinanen KP, Kellokumpu S, Metsikko MK, Rajaniemi HJ . Purification and partial characterization of rat ovarian lutropin receptor . J Biol Chem 1987 ; 262 : 7920-6. 36 . Reightley DD, Cressie NAC . The Woolf plot is more reliable than the Scatchard plot in analyzing data from hormone receptor assays . J Steroid Biochem 1980 ; 13 : 1317-23 . 37 . Brandtzaeg P . Mucosal and glandular distribution of immunoglobulin components . Immunology 1974 ; 26 : 1101-14 . 38 . Kutteh WH, Hatch KD, Blackwell RE, Mestecky J . Secretory immune system of the female reproductive tract . I . Immunoglobulin and secretory component-containing cells . Obstet Gynecol 1988 ; 71 : 56-60 . 39 . Rimm AA, Hartz AJ, Kalbfleisch JH, Anderson AJ, Hoffman RG . Basic biostatistics in medicine and epidemiology . Norwalk: Appleton-Century-Crofts, 1980 : 267-81 . 40 . Campbell MJ, Gardner MJ . Calculating confidence intervals for some non-parametric analyses . In : Gardner MJ, Altman DG, eds . Statistics with confidence . London : British Medical Journal, 1989 : 71-79 .

Effect of human chorionic gonadotropin (hCG) on Neisseria gonorrhoeae invasion of and IgA secretion by human fallopian tube mucosa.

The possible effect of human chorionic gonadotropin (hCG) on the mucosal immune response and susceptibility of the fallopian tube mucosa to invasion b...
2MB Sizes 0 Downloads 0 Views