| Sign In to gain access to subscriptions and/or personal tools. |
© 2006 SAGE Publications (C2) Saliva, Breast Milk, and Mucosal Fluids in HIV TransmissionPresented at the Fifth World Workshop on Oral Health and Disease in AIDS, Phuket, Thailand, July 6–9, 2004, sponsored by Prince of Songkla University, Thailand, the International Association for Dental Research, the World Health Organization, the NIDCR/National Institutes of Health, USA, and the University of California-San Francisco Oral AIDS Center.
1 Center for AIDS Prevention Studies, University of California San Francisco, San Francisco, CA, USA Correspondence: * corresponding author, shafer{at}psg.ucsf.edu
The oral environment has received various amounts of attention in association with HIV infection and pathogenesis. Since HIV infection occurs through mucosal tissue, oral factors—including tissue, fluids, and compartments—are of interest in furthering our understanding of the diagnosis, infectivity, transmission, and pathogenesis of disease. This report reviews: (1) HIV testing and diagnoses with oral fluids; (2) post-natal acquisition of HIV in association with breast-feeding from HIV-positive mothers; and (3) oral sex and HIV transmission. In the first, we examine how oral fluids are used to detect HIV infection and review current consensus on the role of salivary molecules as markers for immunosuppression. Second, lactation-associated HIV acquisition is reviewed, with special consideration of emerging issues associated with the impact of anti-retroviral therapies. Last, we consider current data on the risk of HIV infection in association with oral sex. Investigation of these diverse topics has a common goal: understanding how HIV presents in the oral environment, with an aim to rapid and accessible HIV diagnosis, and improved prevention and treatment of infection.
Key Words: HIV • saliva • breast-feeding • oral mucosa • oral fluid • post-natal transmission
The oral environment has received various amounts of attention in association with HIV infection and pathogenesis. Since HIV infection occurs through mucosal tissue, oral factors—including tissue, fluids, and compartments—are of interest in the diagnosis, infectivity, transmission, and pathogenesis of disease. This report reviews:
Understanding how HIV presents in the oral environment may lead to rapid and accessible HIV diagnosis, and improved prevention and treatment.
Question 1: Are there oral fluid markers—for example, neopterin—for predicting disease progression? Question 2: Can oral fluids be used for diagnostics in the HIV-infected individual? Question 3: What are the routes of HIV transmission via breast milk (BM)? Are tonsillar and gastrointestinal tissues (GIT) the primary sites of initial infection? Question 4: How do other conditions in infants and mothers affect post-natal transmission? Question 5: What are the benefits and safety of ART, especially nevirapine, given to mothers and infants to prevent the transmission of HIV associated with lactation? Question 6: What is the rate of infectivity of HIV associated with oral sex?
Question 1: Are there oral fluid markers—for example, neopterin—for predicting disease progression? There are many advantages to the use of oral fluids for diagnostic purposes compared with blood, including: ease of collection (important in developing countries), non-invasive technique, and less risk of cross-infection. We review the potential use of salivary molecules (immunoglobulins, cytokines, acute reactant proteins, neopterin) as markers of the mucosal immune system function. Currently, the top two markers for assessing HIV disease progression include plasma viral load and CD4+ lymphocyte count. Plasma levels of HIV viremia discriminate risk at all levels of CD4+ lymphocyte counts and predict their subsequent rate of decline (Mellors et al., 1997); thus, this measure is considered the gold standard. Mellors et al.(1997) showed that other markers to consider in assessing HIV disease status and risk of death (in order of predictive strength) were serum neopterin levels, serum beta 2-microglobulin levels, and candidiasis or fever.
Here, we review some other potential candidates for consideration as oral markers of disease progression that have been considered, including: IgA, calprotectin, secretory leukocyte protease inhibitor (SLPI), Th1/Th2 cytokines, neopterin, and beta-2-microglobulin. Since oral diseases such as candidiasis are also correlated with decreased HIV viral load (Baqui et al., 1999a), the assessment of these markers in the presence or absence of oral disease has been postulated to be useful for assessing disease progression. HIV infection may be associated with a serum IgA hypergammaglobulinaemia; however, conversely, secretory IgA levels appear to be lowered in HIV infection. Further, the correlation between serum total IgA levels and CD4 count is poor, leading researchers to conclude that salivary IgA would be a poor marker of HIV disease progression (Sweet et al., 1995). Salivary calprotectin levels are increased in the presence of candidiasis, presumably as a consequence of local inflammation, leading to suggestions that this marker may be useful. However, by controlling for confounding variables, researchers found that, in the presence of HIV infection, there is an overall decrease in salivary calprotectin concentration levels, but these correlated poorly with CD4 counts (Sweet et al., 2001). SLPI is a well-known inhibitor of HIV in saliva and other mucosal secretions (Shugars et al., 2002). SLPI levels have been shown to increase by 17% in plasma and 10% in saliva in HIV patients, and this increase was more marked in patients with high viral loads (> 10,000 copies/mL) compared with low loads (< 400 copies/mL). However, alterations in SLPI levels could be used to distinguish only broad subject groups and were not sufficient to be used to predict the progression of individuals (Baqui et al., 1999b). The measurement of local mucosal Th1/Th2 cytokines (IL-4, IL-10, IL-2, IFN-.
Question 2: Can oral fluids be used for diagnostics in the HIV-infected individual? Some of the problems associated with HIV antibody testing in saliva—such as the low levels of IgG, degradation of IgG by bacterial proteases, and interference of assay techniques by salivary mucins—have been overcome by the collection of oral mucosal transudates. This involves placing an absorbent pad in the buccal sulcus for a few minutes to facilitate collection of IgG as it passes through the oral mucosa. Storage in a preservative diluent greatly facilitates the subsequent testing with assays reportedly achieving sensitivities and specificities of 99.9 to 100% and 99 to 99.9%, respectively (Gallo et al., 1997; Malamud, 1997; Chohan et al., 2001). An oral fluid-based test for antibodies to human immunodeficiency virus (HIV), equivalent to serum in its accuracy but safer and easier to use, is now available in the United States and elsewhere. The development of the oral test involved overcoming technical obstacles to the use of oral fluid as a testing medium, including low immunoglobulin G (IgG) titers, suboptimal assay performance, protease degradation of IgG, high viscosity, and lack of a standardized method of specimen collection, all of which contribute to suboptimal assay performance. The currently available oral HIV test utilizes a collection device to isolate a mucosal transudate component of oral fluid rich in IgG. A vial containing a preservative solution facilitates the transport of stable, low-viscosity specimens to the laboratory for testing with an ELISA and confirmatory Western blot assay, specifically designed for use with oral fluid. Non-HIV medical conditions and oral pathologies do not appear to affect oral test results. It is hoped that the availability of this patient-friendly, portable diagnostic test for antibodies to HIV will facilitate identification of greater numbers of infected individuals, with the ultimate goals of early identification, early treatment, and prevention of disease transmission.
UNAIDS estimates that more than 600,000 infants were infected by human immunodeficiency virus (HIV-1) through mother-to-child transmission (MTCT) worldwide, with the majority (> 90%) occurring in sub-Saharan Africa in 2003 (UNAIDS, 2004). The majority of MTCT occurs during pregnancy and birth; however, up to a third may be accounted for during breast-feeding (Datta et al., 1994; Bulterys et al., 1995; Bertolli et al., 1996; Bobat et al., 1997; Ekpini et al., 1997; Miotti et al., 1999; John-Stewart et al., 2004). It has been estimated that breast-feeding-related risk ranges from 30–45% at 24 months post-delivery (Gaillard et al., 2004). Transmission risk can be reduced to 1% by treating HIV-infected women in the third trimester, implementing obstetric interventions, and avoiding breast-feeding. Short-course prophylaxis with anti-retroviral drugs administered to mothers pre-partum and during labor, and for 1 week post-partum to new-born infants, can reduce MTCT up to 50% (Dabis et al., 1999; Guay et al., 1999; Shaffer et al., 1999; Wiktor et al., 1999; Jackson et al., 2003; Moodley et al., 2003). Unfortunately, several studies and analyses have shown that breast-feeding abrogated some or all of the benefit of anti-retroviral prophylaxis (Leroy et al., 2003; Coutsoudis et al., 2004).
Question 3: What are the routes of HIV transmission via breast milk? Are tonsillar and gastrointestinal tissues (GIT) the primary sites of initial infection?
Question 4: How do other conditions in infants and mothers affect post-natal transmission?
Question 5: What are the benefits and safety of ART, especially nevirapine, given to mothers and infants to prevent the transmission of HIV associated with lactation? However, this success is complicated by the risk of development of resistance virus in the mother and in infected infants (Jackson et al., 2000; Cunningham et al., 2002; Kantor et al., 2002; Eshleman et al., 2004). In an important assessment of risk for breast-feeding infants, plasma and breast milk from 32 women from Chitungwiza, Zimbabwe, who had participated in HTPN 023 and received SD NVP, were assessed for HIV-1 RNA levels and NVP-associated reverse-transcriptase mutations (Lee et al., 2005). At 0, 2, 8, and 20–32 (reported as an average of 24 weeks) weeks after the intervention, plasma and breast-milk HIV-1 RNA levels were measured and NVP-associated mutations identified. Breast-milk sodium levels were measured to examine the correlation between sodium levels (with elevated sodium levels > 3 S.D. from normal thought to be diagnostic of subclinical mastitis) and breast-milk HIV-1 RNA levels. Plasma HIV-1 RNA levels did not differ significantly at 0, 2, 8, and 24 weeks after SD NVP exposure, with an 8-week median HIV-1 RNA level of 4.56 log10 copies/mL (25–75% range from 3.78–4.91 log10 copies/mL). NNRTI-associated mutations were found in 21/28 (75%) of women at 2 weeks, 11/32 (34%) at 8 weeks, and 5/27 (19%) at 24 weeks following SD NVP. At the 2-week time-point, 10/28 (36%) women had a single mutation and 11/28 (39%) had 2 or more mutations; at 8 weeks, 5/32 (15%) had a single mutation and 6/32 (18%) had 2 or more mutations. Having a plasma NNRTI-associated resistance mutation at 8 weeks was significantly associated with having a lower baseline median CD4+ lymphocyte cell count of 325.5 cells per mm3 compared with a baseline median CD4+ lymphocyte cell count of 408 cells per mm3 in women who did not develop resistance mutations (p = 0.03). The 8-week HIV-1 RNA levels were higher in women who developed mutations, but were not statistically different in women who did not develop resistance (4.72 vs. 4.01 log10 copies/mL, p = 0.06). The mutations that were seen in plasma varied at different time points, with mutations at codon 181 predominating at 2 weeks (57%), but with mutations at codons 103, 106, 188, and 190 also detected, while K103N was the predominant NNRTI-associated resistance mutation at 8 and 24 weeks. Breast-milk HIV-1 RNA levels were significantly lower at 2 weeks (1.62 and 1.54 log10 copies/mL in the left and right breast, respectively) than at 8 weeks (2.20 and 2.25 log10 copies/mL in the left and right breast, respectively), following SD NVP (p < 0.01). At 8 weeks, paired right and left breast-milk HIV-1 RNA levels were significantly correlated (r = 0.79, p = 0.003), but with higher levels in milk from the right breast (p = 0.002). Elevated sodium levels, indicative of mastitis (> 12 mmol/L), at 8 weeks were seen in 44% of the paired samples tested, but 17/29 (59%) of women had at least one breast-milk sample with an elevated sodium level. The breast-milk HIV-1 RNA level was significantly higher in those with indications of mastitis in the left breast-milk samples (1.69 vs 2.74 log10 copies/mL, p = 0.024), and trending toward significance in right breast-milk samples (1.94 vs. 2.86 log10 copies/mL, p = 0.099). Overall, a breast-milk sodium level > 12 mmol/L was associated with a 6.23 increased odds of viral RNA level higher than the study median of 2.21 log10 copies/mL at 8 weeks (95% CI 2.03–19.17). There was no significant correlation between plasma and left breast-milk HIV-1 RNA levels (r = 0.04, p = 0.87), but there was some correlation between plasma and right breast HIV-1 RNA levels (r = 0.42, P = 0.066). Overall, plasma HIV-1 RNA levels were higher than paired breast-milk HIV-1 RNA levels (p < 0.0001). Twenty paired breast-milk and plasma samples were available for sequence analysis at the 8-week time-point. Ten pairs (50%) had concordant findings between breast-milk and plasma samples (6 pairs, wild-type (WT); 3 pairs, K103N; 1 pair, K103N and G190A); the other 10 (50%) pairs were discordant (4 pairs with WT in plasma and K103N in breast milk, 1 with WT in breast milk and K103N in plasma, and the remaining 5 pairs with dissimilar resistance mutations in both breast milk and plasma). The presence of an NNRTI-associated mutation in plasma was associated with a 13.5-increased odds of having at least one breast-milk sample with an NNRTI-associated mutation (95% CI = 0.95–687.88). In summary, the study showed a high frequency of NNRTI-associated resistance mutations in maternal plasma, with a change in the predominant mutation over time. Different levels of HIV-1 RNA, as well as variation in the mutations detected in paired breast-milk and plasma samples, suggest the possibility that these sites represent different compartments. It is unknown where or how these compartmental differences or the NNRTI-associated mutations found in breast milk may be associated with increased transmission risk to infants. Anti-retroviral drugs, nevertheless, will continue to pose promising effects, and investigations are under way to assess the benefits and risks of this approach toward reducing post-natal transmission (Gaillard et al., 2004; Safrit et al., 2004). Tenofovir diproxil fumarate (tenofovir), a nucleotide reverse-transcriptase inhibitor, has been demonstrated to prevent transmission of simian immunodeficiency virus (SIV) when administered near the time of virus challenge. A low dose of tenofovir (4 mg/kg/d) partially protected newborn macaques inoculated orally with virulent stock of SIVmac251 at a dose of 100,000 TCID50 (Van Rompay et al., 2001). The 4 placebo-treated animals all became persistently infected after a single inoculation. In contrast, 5 of 8 (62%) animals treated with 2 doses of tenofovir at a dose of 4 mg/kg/d remained seronegative. Two different peri-exposure dosing regimens were used that provided protection in 3 of 4 animals (75%) and 2 of 4 animals (50%), which were not statistically different. This study and others in non-human primates are important in the continued assessment of chemoprophylaxis, immunotherapeutic, and vaccine approaches to post-partum HIV prevention in infants (Van Rompay et al., 1998, 2003). Current studies planned or under way, assessing anti-retroviral interventions to reduce PTMCT in resource-constrained settings, include providing anti-retroviral therapy to infected mothers to reduce viral load in plasma and breast milk, and trials assessing the provision of anti-retroviral therapy to HIV-negative infants during periods of breast-feeding (Gaillard et al., 2004). At sites throughout Africa, studies are planned or under way assessing single or dual anti-retroviral therapies, ultrashort regimens, or one-week trials of SD NVP compared with NVP and zidovudine, intermediate duration (6 weeks) interventions of NVP regimens, and prolonged duration (6 months) interventions (reviewed in Gaillard et al., 2004). Trials of maternal treatment as prophylaxis for infants at risk of HIV from lactation using highly active anti-retroviral treatment (HAART) are also in preparation or under way. Since maternal viral load is significantly associated with infection risk in infants in both breast-feeding and non-breast-feeding women (Cooper et al., 2002; Dorenbaum et al., 2002), approaches to reducing maternal viral load during late pregnancy and after delivery are especially promising. As well as reducing transmission risk, there are other benefits, such as improved immune and health status in mothers and resistance profiles lower than those associated with SD NVP. These studies have important challenges as well, assessing safety and toxicity in both mother and child in the short and long term ("Nucleoside exposure in the children of HIV-infected women receiving antiretroviral drugs: absence of clear evidence for mitochondrial disease in children who died before 5 years of age in five United States cohorts," 2000). Indeed, the many potential benefits from anti-retrovirals for PTMCT could be contradicted by potentially significant and not well-quantified adverse events, including premature delivery, perinatal and neonatal mortality, hematological effects, resistance, impact on future therapy options, rare but possibly fatal mitochondrial abnormalities (Barret et al., 2003), and effects of drug exposures on growth and development.
Question 6: What is the rate of infectivity of HIV associated with oral sex? Since HIV was identified as a sexually transmitted infection, there has been considerable interest in the transmission risk associated with oral sex, especially fellatio, with an infected individual, and specifically, what the infectivity rate is of HIV in association with oral sex. Early studies found no independent risk for fellatio in multivariate analyses (Schechter et al., 1986a; Darrow et al., 1987; Kingsley et al., 1987; Winkelstein et al., 1987; Detels et al., 1989); however, the high correlation among multiple sexual practices in most sexually active persons left open the possibility that risk existed but could not be detected in the available data. Subsequently, case reports of HIV transmission in association with oral sex accumulated among men who have sex with men (MSM), and heterosexual couples who denied practicing other risk behaviors, including anogenital, vaginal intercourse, or intravenous drug use (DeGruttola and Mayer, 1987; Lifson, 1988; Rozenbaum et al., 1988; Detels et al., 1989; Lane et al., 1991; Murray et al., 1991). Despite uncertainty about this potential infection route, many researchers accepted that fellatio, while not an efficient route of transmission, potentially posed a small risk of HIV infection (Lyman et al., 1986; Schechter et al., 1986b; DeGruttola and Mayer, 1987; Goldberg et al., 1988; Lifson, 1988; Spitzer and Weiner, 1989; Lane et al., 1991; Chen and Samarasinghe, 1992; Keet et al., 1992; CDC, 1997; Rothenberg et al., 1998). Some quantitative estimates of risk of HIV seroconversion have been published, principally from cohort studies of MSM (Detels et al., 1989; Samuel et al., 1993; Page-Shafer et al., 1997; Vittinghoff et al., 1999). A particularly well-documented study (Detels et al., 1989) identified and attributed only one seroconversion to oral sex (fellatio) among 232 men followed for 9330 person-years of observation. A study pooling data on seroconverters from multiple sites—Amsterdam, San Francisco, and Sydney (Page-Shafer et al., 1997)—found small but significant associations for incident infection in association with oral sex. An important study by Vittinghoff et al.(1999) estimated infectivity in association with various sexual behaviors and condom use among MSM. They found that the per-contact risk among MSM of unprotected receptive fellatio with ejaculation with an HIV-positive or ’unknown HIV status’ partner, estimated at 4/10,000 (95% CI, 0.01%, 0.17%), was slightly lower than the estimated risk of acquiring HIV per act of protected receptive anal sex (0.18%; 95% CI, 0.10%, 0.28%) and much lower than in unprotected insertive anal sex (0.06%; 95% CI, 0.025%, 0.19%). A limitation of these as well as other studies is the possible overestimation of risk due to the inclusion of persons in the analyses who engaged in higher-risk practices, such as anal sex (Page-Shafer et al., 1997; Vittinghoff et al., 1999; Celum et al., 2001). Keet et al.(1992) found that more than half of incident HIV infections attributed to receptive oral sex (fellatio) were misattributed due to response bias, wherein a high proportion of study participants did not report anogenital sex in written questionnaires, but later did report this practice in face-to-face interviews, leading researchers to conclude that oral acquisition of HIV occurs, but its frequency may be overestimated because of reluctance to report more stigmatized practices, including anal sex. Two more recent studies underscore the very low infectivity of HIV in association with oral sex in heterosexual and MSM populations. In a longitudinal study of serodiscordant heterosexual couples, del Romero et al.(2002) found no incident HIV infections in over 19,000 unprotected orogenital contacts with an HIV-infected partner. In a study of MSM HIV testers in San Francisco who practiced only oral sex (Page-Shafer et al., 2002), no prevalent or incident HIV infections were detected in an estimated 1519 person-years of risk exposure (Balls et al., 2004). Animal models have demonstrated the biologic plausibility of HIV infection from oral exposure to infected semen: Macaques exposed orally and non-traumatically to SIV easily acquire infection (Baba et al., 1996). These studies use high concentrations of SIV and invasive methods of infection which do not approximate human exposure, where infectivity is demonstrably lower. Maher et al.(2004) reported that although HIV binds to oral mucosal, there is limited progression to infection. These authors developed a mucosal model, using human palatine tonsil with intact external epithelium, to study events after oral exposure to HIV. When applied to the external epithelium, semen from an HIV-seropositive patient and cell-free virus both established HIV infection in individual tonsillar cells. However, clusters of infected tonsillar cells were detected where the epithelial surface was damaged. Investigation of the initial events in HIV transmission revealed extensive and stable binding of HIV virions and seminal cells to tonsil epithelium. In experiments modeling physiologically relevant events, the addition of seminal plasma resulted in enhanced virion binding to epithelial cells; however, there was limited progression from binding to primary infection (Maher et al., 2004). In conclusion, the low probability of infection in oral sex, combined with the high prevalence of this sexual practice, has suggested to many researchers the presence of potential inhibitory or defensive factors in the oral cavity (Shugars et al., 2002). Investigations into these factors continue to be needed to develop protective agents or pathways for HIV prevention.
Understanding the pathogenesis of HIV in the oral cavity has led to the successful development of diagnostic tools that allow for the detection of HIV infection without the need for venipuncture and its associated occupational risks. CD4+ lymphocyte counts and HIV RNA levels measured in blood provide a measure of disease progression; attempts to find a biologic agent that is present in oral fluid/secretions, and providing prognostic value in the same manner, have not been hugely successful. However, insight gained into the pathogenesis of HIV in the oral cavity is germaine to our understanding of HIV disease transmission through the oral cavity, as occurs in the setting of post-natal breast-feeding-associated transmission of the disease, and possibly during intra-partum transmission of HIV. Viral characteristics appear to differ in breast milk compared with plasma; it is not yet clear how the presence of drug-resistant mutations in breast milk affects transmission of HIV, or disease progression when transmission occurs. Understanding the protective elements that result in the minimal risk of HIV acquisition associated with oral sexual exposure can instruct further research into HIV prevention methods.
The authors acknowledge the helpful assistance of Joyce Balls, MPE, and Esther Lee, MS, in the preparation of data presented at the workshop. We also acknowledge the support of Dr. David Katzenstein, MD, and the HPTN 023 Study Team for allowing the presentation of data from that study at the 5th World Oral AIDS Conference. Authors received funding from the following sources: K. Page-Shafer from the NIDCR (1 R01 DE12911); Simon Sweet from the Department of Oral Medicine, Pathology and Microbiology, Guys, Kings and St Thomas’ Dental Institute, London; Seble Kassaye from the Doris Duke Charitable Foundation (20000648); and Charles Ssali from the NIH/NIMH (5 P50 MH42459). All authors received support from the organizers of the 5th World Workshop on Oral Health and Disease in AIDS.
Advances in Dental Research, Vol. 19, No. 1,
152-157 (2006)
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
Abstract
Introduction
, IL-12) in saliva may be indicative of oral-associated cell-mediated immunity, which might reflect the susceptibility of HIV patients to oropharyngeal candidiasis (