Can HIV infection occur through the lining of the mouth?

This article is more than 19 years old. Click here for more recent articles on this topic

HIV from semen or breast milk can bind to and infect the cells lining the mouth and the tonsils, according to two test tube studies published on 1st December in The Journal of Infectious Diseases and The Journal of Acquired Immune Deficiency Syndromes. One study also shows that low concentrations of alcohol can enhance the ability of the virus to infect mouth cells.

Infection of cells in the lining of the mouth is not regarded as a major route for HIV transmission. However, despite the low risk of HIV infection through the mouth, these two studies were designed to analyse whether the cells lining the mouth and tonsils can be infected with HIV and provide a potential route to HIV transmission.

Although both studies show that a small number of these cells can become infected with HIV, they do not prove that this process is linked to HIV transmission, or that these cells would become infected following exposure to HIV in semen or milk. Of note, the studies were carried out using tissue and cell cultures that were exposed to HIV for longer periods than would occur under normal circumstances within the mouth. In addition, they were also done without the presence of saliva, which has been reported to have HIV inactivating properties in previous studies.

Infection of mouth cells

In the first study, investigators from University of California Los Angeles examined whether HIV can infect cells from the lining of the mouth or ‘oral epithelium’. After growing cells in culture, they exposed them to high concentrations of HIV particles for 16 hours.

Glossary

tonsils

Two oval lymph node-like structures situated where the mouth joins the throat.

 

oral

Refers to the mouth, for example a medicine taken by mouth.

receptor

In cell biology, a structure on the surface of a cell (or inside a cell) that selectively receives and binds to a specific substance. There are many receptors. CD4 T cells are called that way because they have a protein called CD4 on their surface. Before entering (infecting) a CD4 T cell (that will become a “host” cell), HIV binds to the CD4 receptor and its coreceptor. 

strain

A variant characterised by a specific genotype.

 

gp120

A glycoprotein on the HIV envelope. gp120 binds to a CD4 receptor on a host cell, such as a CD4 T lymphocyte (CD4 cell). This starts the process by which HIV fuses its viral membrane with the host cell membrane and enters the host cell.

Although these cells do not have CD4 receptors on their surface, the researchers discovered that they could be infected with HIV, since they could detect the virus’s genetic material and HIV particles within the cells. However, as the final concentration of HIV used in the experiments was equivalent to semen or breast milk viral loads of between 200,000 and 1,000,000 copies/ml and exposure times of 16 hours were used, it is likely that the infection rate would be many times lower under normal circumstances.

The investigators went on to examine the effect of alcohol exposure on this process by adding concentrations of up to 4% ethanol to the cell cultures. Treatment with 4% ethanol for less than 10 minutes increased the HIV infection of the mouth cells three- to sixfold. This was observed with strains of HIV that normally use either the CCR5 co-receptor to enter T-cells, as well as those that use the alternative co-receptor CXCR4, which usually appear late in HIV disease and are associated with rapid CD4 T-cell loss.

“In this study, we found that 3% to 4% ethanol treatment significantly stimulated HIV infection of primary oral epithelial cells, suggesting that ethanol may play a direct role in HIV oral transmission,” conclude the investigators. “The concentration of 3% to 4% ethanol is similar to that contained in various beers.”

The effect of alcohol was also seen when the cells were treated with 4% ethanol five minutes or two hours before the HIV particles were added. As the effects of ethanol on HIV infectivity were the same under these conditions, the authors conclude that “it is unlikely that ethanol stimulation is caused by the interaction between ethanol and HIV viral particles.” Rather, they suggest that ethanol causes a persistent change in the membrane properties of the oral epithelial cells that increases the chance of HIV infection. However, this effect was lost four hours after ethanol exposure.

The investigators also assessed the ability of a strain of HIV that does not have the gp120 molecule in its outer coat. As this molecule binds to CD4 receptors to allow entry to T-cells, this strain of virus cannot infect CD4 T-cells. However, it could infect the mouth cells and its infectivity was enhanced by alcohol, suggesting that HIV can enter mouth cells by a gp120-independent mechanism. The authors speculate that other proteins in the viral envelope may be responsible for binding to oral cells and subsequent infection.

HIV infection of tonsil cells

In the second study, investigators from the universities of Minnesota and Alabama applied semen from an HIV-positive patient to small pieces of human tonsil embedded in agar jelly. The fresh tonsil tissue was arranged so that only its outer surface, or epithelium, was exposed to the semen.

The tonsils are regarded as a site where HIV infection is particularly likely to occur within the mouth, since they form part of the lymphatic system and contain leukocytes (white blood cells) that are susceptible to HIV infection.

After staining the tonsil tissue with an antibody-based technique, occasional HIV-infected cells were found when they were examined under the microscope. However, clusters of more than one HIV-infected cell were seen only where the tonsil tissue was damaged.

Similar results were seen when the researchers applied cell-free HIV to the tonsil tissue.

In both of these experiments, infected tonsil cells were only found after four to seven days of contact with the HIV-infected semen or HIV particles. This exposure is many times longer than would be expected after receptive oral sex, suggesting that infection of these cells would be less likely under normal circumstances, even if the tonsil surface was damaged by infection or trauma.

Using HIV particles genetically engineered to glow green, the researchers went on to examine the binding of HIV particles to the tonsil surface. They found that HIV binds strongly to the surface of the tonsil epithelium cells in damaged and undamaged areas after three to six hours’ exposure. Cells from semen, including sperm cells and round cells also bound to these cells, and were found deep within the tonsil tissue.

The authors acknowledge that the high degree of HIV binding to these cells is at odds with the low rates of infection observed in epidemiological studies. “Our results establish a distinct separation between extensive and rapid binding events and the relatively rare establishment of primary infection in recipient tissue and they highlight the protective capacities of an intact epithelium,” they conclude.

Only 50% of the bound HIV particles were removed after washing the tonsil tissue once and culturing it for another three to five hours. The investigarors point out that the HIV bound persistently to the tonsil surface “may remain infectious, thereby increasing the time frame for establishing primary infection in the recipient.”

However, they did not attempt to replicate the continuous washing process that would occur in the mouth due to the secretion of saliva and repetitive swallowing. Therefore, under normal conditions in the mouth, the length of time for which HIV would remain bound to the tonsil epithelium is likely to be many times shorter than that observed in this study.

References

Maher D et al. A model system of oral HIV exposure, using human palatine tonsil, reveals extensive binding of HIV infectivity, with limited progression to primary infection. J Infect Dis 190: 1989-1987, 2004.

Zheng J et al. Ethanol stimulation of HIV infection of oral epithelial cells. J Acquir Immune Defic Syndr 37: 1445-1453, 2004.