Defining the interaction of HIV-1 with the mucosal barriers of the female reproductive tract

Abstract

Worldwide, HIV-1 infects millions of people annually, the majority of whom are women. To establish infection in the female reproductive tract (FRT), HIV-1 in male ejaculate must overcome numerous innate and adaptive immune factors, traverse the genital epithelium, and establish infection in underlying CD4(+) target cells. How the virus achieves this remains poorly defined. By utilizing a new technique, we define how HIV-1 interacts with different tissues of the FRT using human cervical explants and in vivo exposure in the rhesus macaque vaginal transmission model. Despite previous claims of the squamous epithelium being an efficient barrier to virus entry, we reveal that HIV-1 can penetrate both intact columnar and squamous epithelial barriers to depths where the virus can encounter potential target cells. In the squamous epithelium, we identify virus entry occurring through diffusive percolation, penetrating areas where cell junctions are absent. In the columnar epithelium, we illustrate that virus does not transverse barriers as well as previously thought due to mucus impediment. We also show a statistically significant correlation between the viral load of inocula and the ability of HIV-1 to pervade the squamous barrier. Overall, our results suggest a diffusive percolation mechanism for the initial events of HIV-1 entry. With these data, we also mathematically extrapolate the number of HIV-1 particles that penetrate the mucosa per coital act, providing a biological description of the mechanism for HIV-1 transmission during the acute and chronic stages of infection.

Fig 1

Macroscopic (left) and microscopic (right) illustrations of the female reproductive tract. (A) Upper reproductive tract. Included are the endocervix and uterus. Tissue is primarily comprised of simple columnar epithelium. Target cells in the lamina propria consist of dendritic cells (green), T cells (blue), and macrophages (purple). (B). Lower reproductive tract. Included are the ectocervix and vagina. Intraepithelial target cells include Langerhans cells (red). Target cells in the lamina propria consist of T cells (blue) and macrophages (purple). The tissue is stratified squamous epithelium composed of the following epithelial layers (distal to proximal): stratum corneum (C), stratum granulosum (D), stratum spinosum (E), and stratum basale (F).

Fig 2

(Top) Entry of HIV-1 into the columnar epithelium of the endocervix and the stratified squamous epithelium of the ectocervix (100×). (A) Identified virions are shown in red, tissue background is shown in green, and DAPI is shown in blue. (B) Identified virions are shown in red, cytokeratin 7 is shown in green, and DAPI is shown in blue. (C) Magnified view of the region within the white box illustrated in panel B. (D) Identified virions are shown in red, tissue background is shown in green, and DAPI is shown in blue. (E) Identified virions are shown in red, WGA is shown in green, and DAPI is shown in blue. (F) Magnified view of the region within the white box illustrated in panel E. (G) Histogram illustrating the percentage of virus penetration in the endocervix (Endo; light gray) and ectocervix (Ecto; dark gray) of ex vivo human cervical tissue. (Middle) Analysis of PA-GFP HIV-1 in individual human cervical explant donors. Interquartile box plots of virus penetration depths of each donor sample are shown. (H) Endocervical donors. (I) Ectocervical donors.

Fig 3

Entry of PA-GFP HIV-1 into the columnar epithelium of NA-treated and untreated human endocervical samples. NA enzymes (sialidases) are glycoside hydrolase enzymes that remove sialic acid from mucins and disrupt the gel-forming interactions in mucus. Untreated samples incubated with PA-GFP HIV-1 are indicated by arrowheads. (A) Identified virions are shown in red, tissue background is shown in green, and DAPI is shown in blue. (B) Identified virions are shown in red, Muc5ac is shown in green, and DAPI is shown in blue. NA-treated samples incubated with PA-GFP HIV-1 (arrowheads). (C) Identified virions are shown in red, tissue background is shown in green, and DAPI is shown in blue. (D) Identified virions are shown in red, Muc5ac is shown in green, and DAPI is shown in blue. (E) Scatter dot plots of virus penetration depth in NA-treated and untreated samples. n = 22 per treatment. Error bars represent standard errors of the mean (SEM). (F) Box plots of the number of penetrating virions in NA-treated and untreated samples (P = 0.018).

Fig 4

(Top) Analysis of PA-GFP HIV-1 using the in vivo macaque FRT (100×). (A) Identified virions are shown in red, tissue background is shown in green, and DAPI is shown in blue. (B) Identified virions are shown in red, WGA is shown in green, and DAPI is shown in blue. (C) Magnified view of the region within the white box illustrated in panel B. (Bottom) Scatter dot plots of PA-GFP HIV-1 virus penetration depths in in vivo macaques 4 and 12 h postinoculation and 4 h postexposure in rhesus macaque explants. n, number of animals. (D) Four hours postinoculation. (E) Twelve hours postinoculation. (F) Macaque explants 4 h postexposure. Error bars represent SEM.

Fig 5

(Top) BSA and PA-GFP HIV-1 penetration depths in the in vivo macaque model. (A) Deconvolved panel image of human ectocervical explant incubated with fluorescently tagged BSA (green) for 4 h. Adherens junctions are shown in red, and DAPI is shown in blue (40×). (B) Deconvolved panel image of in vivo rhesus macaque vagina 4 h postexposure with BSA (red) and DAPI (20×). (C) Scatter dot plots of BSA and virus penetration depth in the in vivo macaque vagina. n, number of animals. ***, P < 0.001. (Bottom) The role of adherens junctions in PA-GFP HIV-1 penetration depths in human cervical explants. (D) Scatter dot plot representing depths of virion penetration in relation to intact cellular junctions. The distance between virus and robust cellular junctions was calculated by subtracting the depth of virus penetration from the depth of cellular junctions in the same area of tissue. A value of 0 (gray dotted line) represents virions found at the same depth as robust adherens junctions. All values above this line are representative of those virions penetrating to areas where cellular junctions are no longer intact, while all values below the gray line are individual virions that were located below or within robust cellular junctions. (E) Scatter dot plot of virus penetration depths for EDTA-treated and untreated ex vivo human ectocervical tissue. n, number of donors. ***, P < 0.001. Error bars represent SEM.

Fig 6

Mechanistic model of how HIV-1 interacts with the stratified squamous epithelium of the lower FRT. (A) The majority of the time, when virus comes in contact with a healthy stratified squamous epithelia, the virus does not interact with or penetrate the intact tissue. (B) Among those areas where epithelium integrity is compromised and cellular junctions are either absent or degraded, virions can readily interact with or penetrate the vaginal or ectocervical epithelium. Additionally, inflammation is known to increase the density of activated T cells in close proximity to possible infected or infectious substrates within the lumen, increasing the chances of HIV-1 infection in the lower FRT. In these instances, virions are more apt to enter 1 to 3 μm, with a few virions entering past 10 μm in a gradient-like manner via simple diffusion.

Fig 7

Extrapolation of the number of penetrating virions to in vivo conditions based on cervical explant data. At an inoculant concentration of 500 ng/ml p24, HIV-1 penetration of the squamous epithelium was observed at an average of 1.21 penetrating virions/z-scan. All analyzed regions were 60 μm wide and 12 μm thick, accounting for a surface area of 720 μm² (left). By simple data extrapolation and assumption of approximately equal exposure throughout the vaginal vault, we can predict 1.68 × 10⁵ HIV-1 penetrating virions/cm². With the surface area of the vaginal vault being 87.5 cm², we estimate 1.47 × 10⁷ virions penetrating per exposure at a concentration of 500 ng/ml p24. However, these data are denotative of the high virus titers used in our experiments and is not reflective of natural infection. In an attempt to compare our data to in vivo conditions, we found the largest reported amount of HIV-1 in semen, during acute infection, is 4.0 × 10⁶ virions/ml; for chronic infection, the largest reported amount is 5.0 × 10³ virions/ml. Since 1 virion contains approximately 3,000 gag proteins, 1 pg of p24 represents 7,900 virions and 500 ng p24/ml is 3.95 × 10⁹ virions/ml. Therefore, if the acute viral load in semen contains 4 million virions/ml, we predict there will be up to 1.49 × 10⁴ penetrating viruses in the vaginal vault. During chronic infection and by the same data extrapolation method, a viral inoculum that consists of 5.0 × 10³ virions/ml will contain approximately 18 penetrators (right). Likewise, when performing similar extrapolation in endocervical explants, very few penetrating virions were calculated to enter the cervical canal to interact with the 20-cm² simple columnar epithelial surface area.