1785 CONCISE COMMUNICATION Detection of Kaposi's Sarcoma±Associated Herpesvirus in Oral and Genital Secretions of Zimbabwean Women Thomas M. Lampinen,1,3 Shalini Kulasingam,1 1Departments of Epidemiology and 2Pathobiology and 3Center for AIDS Juno Min,4 Margaret Borok,6 Lovemore Gwanzura,7 and STD, University of Washington, Seattle; 4Department of Medicine 5 Downloaded from https://academic.oup.com/jid/article/181/5/1785/2191825 by guest on 29 September 2021 4 8 9 and Center for AIDS Research, Stanford University, Stanford, Julie Lamb, Kassam Mahomed, Godfrey B. Woelk, 6 7 2 2 California; Departments of Medicine, Medical Laboratory Sciences, Kurt B. Strand, Marnix L. Bosch, 8Obstetrics and Gynecology, and 9Community Medicine, 10 10,11 Daniel C. Edelman, Niel T. Constantine, University of Zimbabwe, Harare; 10Department of Pathology, David Katzenstein,4,5 and Michelle A. Williams1 University of Maryland and 11Institute of Human Virology, Baltimore, Maryland Kaposi's sarcoma±associated herpesvirus (KSHV) in oral and genital secretions of women may be involved in horizontal and vertical transmission in endemic regions. Nested polymerase chain reaction assays were used to detect KSHV DNA sequences in one-third of oral, vaginal, and cervical specimens and in 42% of peripheral blood mononuclear cell (PBMC) specimens collected from 41 women infected with human immunode®ciency virus type 1 who had Ka- posi's sarcoma (KS). KSHV DNA was not detected in specimens from 100 women without KS, 9 of whom were seropositive for KSHV. A positive association was observed between KSHV DNA detection in oral and genital mucosa, neither of which was associated with KSHV DNA detection in PBMC. These data suggest that KSHV replicates in preferred anatomic sites at levels independent of PBMC viremia. Detection of genital-tract KSHV only among relatively immunosuppressed women may provide an explanation for infrequent peri- natal transmission of KSHV. Kaposi's sarcoma±associated herpesvirus (KSHV), also In Harare, Zimbabwe, the incidence of KS has increased 20- known as human herpesvirus 8 (HHV-8), appears to be causally fold during the past 10±15 years, making it the most common associated with Kaposi's sarcoma (KS) [1, 2]. The modes of malignancy in children [3], men [4, 5], and women [5]. The high KSHV transmission are incompletely understood. However, incidence of KS and the possible role of women in KSHV like other pathogenic herpesviruses, KSHV might be trans- transmission led us to screen for KSHV DNA present in genital mitted via mucosal secretions. In northern Europe and the and oral specimens collected from Zimbabwean women with United States, where KS is not endemic and childhood KSHV and without clinically evident KS. infections seem rare, epidemiological studies have found evi- dence for transmission of KSHV between male homosexual Methods partners. In contrast, in many regions where KS is endemic, KSHV infection is commonly acquired during childhood [1, 2]. Recruitment, interview, and evaluation of study participants. All patients with clinical evidence of KS referred to central hos- pitals (Parirenyatwa and Harare Hospitals) in Harare, Zim- Received 23 July 1999; revised 14 January 2000; electronically published babwe were assessed at a designated clinic by 1 physician (M.B.). 15 May 2000. Nonmenstruating patients with intact uteri who were >18 years of Informed consent was obtained from all participants, and human exper- age and able to provide informed consent were eligible to partic- imentation guidelines of the US Department of Health and Human Services ipate in this study, regardless of their KS treatment history.Between and the Medical Research Council of Zimbabwe were followed in the con- duct of this investigation. July and October 1996, we enrolled 41 women with clinical diag- Grant support: National Institutes of Health training grants T37- noses of KS; 33 (80%) had histologic con®rmation of the di- TW00049 and NIH AI 07140 and the University of Zimbabwe Research agnosis recorded in their medical record. Clinically de®ned stages Board. of KS and human immunode®ciency virus type 1 (HIV-1) serostatus Reprints or correspondence: Thomas M. Lampinen, University of Wash- ington, Dept. of Epidemiology, Box 357236, Seattle, WA 98195 (tlamp@u were abstracted from the medical record. All but 2 eligible women .washington.edu). participated. For comparison, 100 women without clinical KS were recruited The Journal of Infectious Diseases 2000;181:1785±90 q 2000 by the Infectious Diseases Society of America. All rights reserved. between September 1996 and January 1997 from the outpatient 0022-1899/2000/18105-0039$02.00 gynecology department of Harare Maternity Hospital. Over 90% 1786 Lampinen et al. JID 2000;181 (May) of invited women agreed to participate. The high prevalence of TCC-30). Each 20-mL PCR reaction contained 20±400 ng of DNA, HIV-1 among women seeking routine gynecologic care in Harare 5.0 pmol of each primer, 1.0 U of Taq polymerase, 50 mM each allowed comparison of the frequency of KSHV among women with dNTP, 67 mM Tris (pH 8.8), 2 mM MgCl2,11mM (NH4)2SO4,10 (n=47 ) and without (n=53 ) HIV-1 infection. Womenparticipating mM 2-mercaptoethanol, and bovine serum albumin (100 mg/mL). in the study were generally unaware of their HIV-1 or KSHV se- PCR conditions were as follows: 947C for 1 min, 607C for 1 min, rologic status at the time of enrollment and interview. Posttest 727C for 1 min (40 cycles); and 727C for 30 min (1 cycle). PCR counseling was provided with HIV-1 serologic results. products were puri®ed over gel extraction columns (QIAquick Gel Demographic, medical, reproductive, and sexual histories were Extraction Kit; QIAGEN), were then either sequenced directly or collected during in-person interviews using a structured question- cloned using a T-A cloning system (pCR2.1/INV/F'; Invitrogen, naire. Two experienced nurse-midwives performed physical and Carlsbad, CA), and were then sequenced using M-13 universal Downloaded from https://academic.oup.com/jid/article/181/5/1785/2191825 by guest on 29 September 2021 pelvic examinations with sample collection. primers. Blood samples. Lymphocyte subset analyses were performed Data analysis. CD41 T cell counts !50 per cubic mL (the lower on 1-mL aliquots of EDTA-anticoagulated blood specimens using limit of detection,n=3 ), were assigned a midpoint value of 25. a ¯uorescence-activated cell sorter counter (Becton Dickinson, KSHV-seropositive women with unmeasurable antibody levels Cockeysville, MD). Anticoagulant acid-citrate-dextrose±blood (n=2 ) were assigned a titer of 2. KSHV antibody titers were log2 specimens were collected for peripheral blood mononuclear cell transformed, and geometric mean titers (GMTs) were calculated (PBMC) DNA extraction after Ficoll-Hypaque cell separation. after exclusion of nonreactive specimens. Kruskal-Wallis tests were Sera were collected for syphilis (Treponema pallidum hemagglutin- used to compare continuous variables across multiple groups; pair- ation assay), herpes simplex virus (HSV)±1, HSV-2, and HIV-1 wise (Mann-Whitney) comparisons were performed only when testing (the latter by EIA with Western blot con®rmation). Anti- overall tests were signi®cant. Pearson's x2, Fisher's exact test, and bodies to KSHV were identi®ed in 1 : 50 serum dilutions using an odds ratios (ORs) with 95% con®dence intervals (CIs) were cal- ELISA based on a recombinant antigen derived from the open- culated in analyses involving dichotomous and categorical reading frame (ORF) 65 capsid region [6]. variables. Vaginal, cervical, and oral samples. After insertion of a spec- ulum, Dacron swabs were used to collect specimens for microbio- logical and polymerase chain reaction (PCR) assays. Three swabs Results were separately collected from the anterolateral vaginal wall (with avoidance of pooled cervical secretions), 2 were collected from the Description of the study population. The 41 women with ectocervix, and 2 were collected from the cervical os. Genital swabs KS were HIV-1 seropositive. Most (39 [95%] of 41) had gen- were placed in 2 mL of RPMI media (Life Technologies, Gaith- eralized KS lesions, with visceral KS noted in 10 (24%). The ersburg, MD) and vortexed to remove cellular material. Vaginal average self-reported duration of KS lesions was 18 months secretions adhering to the speculum were used for pH measure- (range, 1±62). Among those with available information, half ments and for potassium hydroxide (amine ªwhiffº) tests. Oral- (17/34) had received some form of treatment for their KS before rinse samples were collected using 10 mL of sterile saline. Genital study enrollment (9, chemotherapy; 7, radiation therapy; 1, 2 7 and oral samples were frozen at 70 C for later DNA extraction. chemotherapy and radiation therapy). DNA testing. Vaginal, cervical, oral rinse, and PBMC samples 1 Women with KS had signi®cantly fewer peripheral CD4 were centrifuged at 2000 g for 1 min. Pellets were resuspended in lymphocytes per milliliter (median, 140; range, 25±391; n= 200 mL of PBS, and DNA was extracted using QIAGEN Blood Kits (QIAGEN, Chatsworth, CA), then quantitated by spectro- 32) than either HIV-1±seropositive women without KS (median, ! photometry. PCR ampli®cation assays were performed as described 274; range, 63±949;n=37 ;P .001 ) or HIV-1±seronegative elsewhere, using 0.5 mg of DNA [7]. Brie¯y, 2 distinct, nested PCR women without KS (median, 719; range, 33±1888;n=45 ; primer pairs that amplify nonoverlapping regions of the KSHV P ! .001). Women with KS were more likely than women with- major capsid gene were used to detect KSHV DNA. Only samples out KS to have antibodies to KSHV ORF 65 (table 1), and that consistently ampli®ed with both sets of primers were consid- the mean antibody titer was 5- to 6-fold higher in women with ered to be positive for KSHV DNA. DNA integrity and the absence KS (GMT, 761; range, 2±32,768), than in women without KS of inhibitors were con®rmed in each sample by PCR ampli®cation (GMT, 138, range, 2±1024;P=.05 ).