Cervical Antibodies to Herpes Simplex Virus Proteins in Pregnancy and Puerperium: A Pilot Study

Objective: This study was undertaken to evaluate the changes in total and anti-herpes simplex virus (HSV)-specific cervical IgA and IgG antibody profiles during and after pregnancy. Methods: Serum and cervical secretions were obtained from pregnant patients before 20 weeks gestation, at 34–36 weeks gestation, and at 6 weeks postpartum and tested for total IgA and IgG antibody and for IgA and IgG to HSV proteins by Western blot. Results: Seven women were HSV seronegative, 14 HSV-1 seropositive, and 14 HSV-2 ± HSV-1 seropositive. Minimal changes in the serum anti-HSV profiles were seen over the 3 visits. The total cervical IgA, IgG, and protein levels did not change between the 2 pregnancy visits but tended to increase at the postpartum visit. No consistent change in cervical HSV-specific IgA and IgG was seen during pregnancy, but the levels increased markedly at the postpartum visit. Conclusions: Lower cervical anti-HSV antibody levels may be related to the previously reported increased frequency of a reactivation of HSV during late pregnancy. Further evaluation is necessary to confirm and quantify the changes in genital immunity during pregnancy and to evaluate whether the increased levels at the postpartum visit are sustained.

change during pregnancy, z the effect of pregnancy on mucosal antibodies has not been assessed. To evaluate the effects of pregnancy on circulating and local cervical total and anti-HSV-specific immunoglobulins, we used quantitative assays to assess cervical total IgA and IgG levels and semiquantitative methods to analyze the HSV-specific IgA and IgG in serum and cervical secretions collected during the first half of pregnancy, at 34-36 weeks gestation, and at 6 weeks postpartum.

Study Population and Protocol
Pregnant women were enrolled in the study from the prenatal clinics of the University of Washington Medical Center and Harborview Medical Center between January 1991 and March 1992. All subjects gave informed consents as approved by the University of Washington Institutional Review Board. The patients were enrolled before 20 weeks gestation and were seen again at 34-36 weeks gestation and 6 weeks postpartum. Of the 55 patients who were enrolled, 37 completed at least 2 visits. Of the 18 (33%) who did not complete the study, 13 were lost to follow-up, moved from the area, and 4 were removed for obstetric reasons (spontaneous abortion in 3 and placenta previa in 1). All women had blood drawn at entry (8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) weeks) to determine their HSV serologic statuses. Herpes viral cultures of the labia and cervix were performed, and cervical secretions were sampled at all visits. A brief questionnaire regarding current symptoms of a genital or oral HSV infection was completed at each visit. Viral cultures were performed as previously described. The serum specimens were assayed for IgA and IgG antibodies to HSV-1 and HSV-2 by Western blot analyses, zl'zz Cervical Secretions The cervical secretions were obtained as previously described. TM Briefly, a vaginal speculum was inserted and 2 ophthalmic tear-flow test strips ("Snostrip," Akorn, Inc., Abita Springs, LA) were placed sequentially on the cervical os until they were saturated to the shoulder, approximately 10 sec each. This technique provided a constant volume of secretions with each sample (mean volume 357 +__ 77 Il/sample). TM The combined papers from a single sampling were trimmed at the shoulder, and the saturated strips were then sealed in sterile INFECTIOUS DISEASES 1N OBSTETRICS AND GYNECOLOGY vials containing 500 !1 of a 0.01 sodium azide solution in phosphate-buffered saline (PBS) and frozen at 20C.

Specimen Preparation
The samples were thawed. After the Sno-strips were removed, the vials were vortexed, then centrifuged at 13,960 g for 15 min. The supernatant volumes were recorded, and a 50-11 portion was removed from each sample for protein analysis (Bio-Rad, Richmond, CA) as previously described. TM  Tween, the blots were incubated for min with a commercial Western blotting detection system based on chemiluminescence as directed by the manufacturer (Amersham, Arlington Heights, IL). The blots were then covered with plastic wrap and exposed to Hyperfilm-ECL (Amersham) for a range of times: 10 sec to 3 min. The film was then developed in a Kodak X-Omat processor. All specimens from each patient were tested against the same lot of ECL-Western blot strips and in the same immunoblot run. For an accurate comparison of antibody profiles across time points, each set of blots for HSV-specific IgA or IgG was developed for the same length of time. The changes in relative levels of IgA or IgG to HSV were scored if at least 3 major bands showed a marked difference in intensity or if the bands were present in specimen and absent in the other specimens being compared. The reactive bands were identified by migration characteristics as described previously, zz,z3 Statistical Analysis The median levels of protein and total IgG and IgA from the cervix at each visit were compared using the 2-tailed Wilcoxon's signed-rank test. z4 The changes scored as increased or decreased levels of HSV-specific cervical IgA and IgG by ECL-Western blot between visits were evaluated using a 2tailed sign test. z4

HSV Serostatus
Of the 37 patients who completed the study, 35 subjects had at least 2 cervical samples taken for local antibody measurements that were Hemoccult negative. Of these, 7 (20%)were HSV seronegative, Nineteen of 35 (54%) women had evaluable cervical specimen pairs from third-trimester and postpartum visits (Table 1). Overall, the median values of total protein, IgA, and IgG rose 3-fold to 5-fold in the postpartum samples. This rise was especially apparent in local IgA which rose from a median of 1.32 to 6.3 meg/ml (P < 0.005). Of interest, this rise was seen mainly in the HSV seronegative and HSV-1 seropositive subsets. The median IgA values rose approximately 10-fold in seronegative women (P 0.08) and HSV-1 seropositive women (P 0.03) and by 2-fold in HSV-2 seropositive women (P 0.35) ( Table 1).
The specimen volume, our measure of adequacy and uniformity of the sampling technique, showed no significant change between the pre-and postpartum samples (P 0.50, P 0.51, P 0.35, and P 0.63 for HSV seronegative, HSV-1 seropositive, HSV-2 seropositive, and all patients, respectively) or between the HSV seronegative and HSV-1 vs.

HSV-I IgA Profiles
Fourteen women who were seropositive for HSVhad samples taken both at the enrollment visit and during the third trimester for comparison of HSV cervical antibody profiles. Of these sample pairs, 3 had no detectable IgA to HSV-1 in either sample. Of the 11 with detectable antibodies, 5 had unchanging profiles of HSV-1 cervical IgA and 3 had increases and 3 had decreases in the number and intensity of reactive bands on ECL-Wesetern blot ( Table 2). An example of the low prenatal levels of cervical IgA to HSV is shown in Figure   1A (lanes and 2). Note that detectable cervical IgG to HSV was present at all sampling points (Fig.  1B). The serum IgA (Fig. 1C)    (6%) 12 (75%) 3 (I 9%) 16"* aECL-Western blot profiles of cervical IgA and cervical IgG were compared between enrollment and 3rd-trimester blots and between 3rd-trimester and postpartum blots, as illustrated in Figures and 2. Numbers in parentheses are the percentages of subjects with a given result within the population of pairs scored ("Total"). Increase at least major bands on ECL-Western blot were markedly increased in intensity or were present in the second but not the first blot of the pair. Decrease at least major bands on ECL-Western blot were markedly decreased in intensity or were present on the first but not the second blot of the pair. All blots for the same subject were run at the same time under the same conditions for each antibody type (IgA or IgG).
bDoes not include HSV-I seropositive and 4 HSV-2 seropositive women who either had no detectable IgA antibody to HSV in their samples or <10 ng of protein in one or both samples. *P 0.004 by 2-tailed sign test. **P 0.08 by 2-tailed sign test.
profiles to HSV-1 (Fig. 1D)  the 6 pairs with changes were either similar between the pairs or showed discordant change, suggesting that sampling differences alone were not sufficient to explain the changes. The sixth pair had a lower volume in the third-trimester sample along with decreasing apparent levels of IgA to HSV-1; however, her IgG levels increased between the 2 specimens.  (Table 2). Figure 1A Table 2). An example of a decrease in cervical IgA band number and intensity is shown in Figure 2A (lanes and 2). This patient also had a decrease in cervical IgG in the third-trimester sample (Fig. 2B, lanes and 2).

Postpartum Cervical HSV-2 IgA
Of the 14 HSV-2 seropositive women, 9 had specimens taken at 6 weeks postpartum. Three specimens were not used for comparison of antibody profiles because they contained blood. Five of the 6 (83%) available pairs showed clear increases in IgA complexity and intensity in the postpartum samples and the sixth showed a decrease ( Table  2). The patient in Figure 2A  ECL-Western blots were performed to detect cervical antibodies to HSV-2 (.,1) and serum antibodies to HSV-2 (C,D). The left blot in each panel is from specimens taken at enrollment, the middle blot from specimens taken during the third trimester, and the right blot from specimens taken 6 weeks postpartum. Note the dramatic change in intensity and number of bands in the cervical IgA profile of the postpartum specimen compared with the cervical IgA profiles from specimens taken during pregnancy. In contrast, the cervical IgG and serum IgA and IgG profiles changed little over time.
body, 7 had no change in IgG profiles, 3 had increased intensity of their profiles, and 4 had decreased intensity of their profiles at the third trimester (Table 2). Of the 14 HSV-2 seropositive patients, creased postpartum. The increasing intensity and complexity of cervical HSV antibody profiles, both IgA and IgG, at the postpartum visit are striking.
It is tempting to assume that these relatively low levels of cervical antibodies late in pregnancy reflect broader immune suppression, which may account for the previously reported increased rate of positive genital herpes cultures during the third trimester. 1 However, a study of more women earlier in pregnancy and for a longer period postpartum is necessary to evaluate whether the increasing levels seen postpartum are the normal levels or whether they are transient evaluations. In previous studies of nonpregnant women with either first-episode or recurrent genital HSV infections, the cultures were not positive when the cervical IgA antibody level was ->1:2.17 In addition, the mean duration of genital shedding of HSV with both first-episode and recurrent outbreaks was 3 days shorter among women who developed secretory IgA compared with those without IgA. 17 These and more recent studies 19 are suggestive that IgA may play a role in limiting the duration of HSV shedding. However, since only patient had a positive culture at any visit in our study, only concurrent, more frequent sampling for both virus and antibody in the cervix could address this question.
The lack of change in serum IgG and IgA antibody levels over the course of pregnancy and the postpartum visit is consistent with previous studies of antibody levels in pregnant women, z The serum antibody patterns do not account for the observed changes in cervical antibody levels. Although no changes were seen in the cervical levels of total protein, IgG, or IgA between the enrollment and third-trimester sampling points, the enrollment visit occurred anytime before 20 weeks gestation. The changes related to pregnancy may occur early in pregnancy, but our sampling protocol did not include first-trimester visits in all cases. One important finding of our study was the clear increase in total cervical IgA levels between the third-trimester and postpartum sampling. However, the total cervical IgG levels did not change significantly from the third-trimester to the postpartum sampling.

It is not clear why the increase in cervical total
IgA and HSV-specific IgA was only 2-fold in HSV-2 positive patients compared with 6-fold in HSV-1 only or HSV antibody negative women. It did not appear to be related to the sampling technique.
The volume as measured in each sample was similar at all 3 time points and in all 3 serologic groups, but only a small number of women were studied. Differences in HSV reactivation in the genital tract may be a factor influencing this observation. Our test method could not detect antibodies that were complexed with antigen. Any apparent changes in HSV-2 cervical antibodies may be affected by viral shedding. More detailed study of changes in cervical antibodies during and after pregnancy among women with genital HSV-1 infections would help to elucidate whether the specific type of HSV infecting the genital area has an effect on local antibody levels during and after pregnancy. Such studies will require more patients, more frequent sampling, and more quantitative testing methods. The differences in local antibodies seen during pregnancy compared with postpartum values may be related to the hormone changes of pregnancy. Of interest, the serum antibody levels in pregnant women do not show significant changes during pregnancy compared with postpartum. In a study of secretory component production by uterine tissues in ovariectomized rats, progesterone was found to cause a marked decrease in the production of secretory component, z5 This effect, which was dose dependent, occurred whether or not the animals were pretreated with estradiol. The high levels of progesterone present during pregnancy may inhibit the production of secretory component which would then limit the local IgA antibody levels.
These factors could be further evaluated by more frequent sampling of cervical IgA levels throughout pregnancy and pairing them with progesterone levels since the concentration of progesterone increases markedly over the course of pregnancy. If progesterone specifically affects secretory component and not overall immunoglobulin production, it would account for differences in local antibodies without concomitant changes in serum antibody levels.
For the evaluation of the changes in local immunity suggested by this pilot study, a more comprehensive and detailed study is required. Comparing a control group of nonpregnant women sampled at similar intervals with a group of women before, during, and after pregnancy would allow an evaluation of the changes over time not related specifically to pregnancy. More frequent sampling for both local antibody levels and HSV reactivation detected by culture and polymerase chain reaction should be included to assess the relationship between local immunity and HSV shedding. In addition, a measurement of antibody levels to an antigen not present in the genital tract such as tetanus would allow a delineation of the changes in levels related to antigenic stimulation compared with nonspecific immune-system activation or changes related to pregnancy. The development of more precise methods for collection and quantitation of HSVspecific antibody levels in cervical secr6tions rather than just a comparison of changes in intensity and complexity of Western blot patterns is necessary to evaluate potential pregnancy-related changes in local antibody. Furthermore, since Western blots measure antibody levels to denatured proteins rather than the intact virion, a correlation of antibody levels with neutralizing activity, viral shedding, and cellular immune responses in the genital tract is necessary to evaluate the clinical significance of changes in ECL-Western blot antibody responses over time.
The methods used in our study could be adapted for a study of the mucosal immune responses to other antigens as well. An evaluation of genital immunity, including antibody response, will be crucial in evaluating the response to human immunodeficiency virus, especially to vaccines designed to protect against human immunodeficiency virus infection.