Rapid Infection of Oral Mucosal-Associated Lymphoid Tissue with Simian Immunodeficiency Virus

See allHide authors and affiliations

Science  20 Aug 1999:
Vol. 285, Issue 5431, pp. 1261-1265
DOI: 10.1126/science.285.5431.1261


The early events during infection with an immunodeficiency virus were followed by application of pathogenic simian immunodeficiency virus atraumatically to the tonsils of macaques. Analyses by virologic assays and in situ hybridization revealed that the infection started locally in the tonsils, a mucosal-associated lymphoid organ, and quickly spread to other lymphoid tissues. At day 3, there were few infected cells, but then the number increased rapidly, reaching a high plateau between days 4 and 7. The infection was not detected in the dendritic cell–rich squamous epithelium to which the virus was applied; instead, it was primarily in CD4+ tonsillar T cells, close to the specialized antigen-transporting epithelium of the tonsillar crypts. Transport of the virus and immune-activating stimuli across this epithelium would allow mucosal lymphoid tissue to function in the atraumatic transmission of immunodeficiency viruses.

Virologic and immunologic events during the initial period of human immunodeficiency virus–type 1 (HIV-1) infection have important consequences for vaccine design and the eventual clinical course (1, 2). The level of virus or set-point that develops after acute infection correlates with the rapidity with which the disease develops and acquired immunodeficiency syndrome (AIDS) appears (3). To observe early events during the transmission of an immunodeficiency virus, and to establish the role of mucosal-associated lymphoid tissue (MALT) at body surfaces, we applied simian immunodeficiency virus (SIV) directly to the surface of the tonsils in macaques. We then followed the kinetics of virus multiplication and spread within and from a single lymphoid organ.

This experimental design also provided information on the relative roles of two types of tissue in the early replication of virus. One tissue is the stratified squamous epithelium that overlies the tonsils. It is rich in dendritic cells (DCs) and comparable to the lining of the vagina and anus, tissues that are implicated in genital transmission of HIV-1. The other tissue is the lymphoid component of the tonsil and is comparable to MALT found in the rectum. A critical feature of MALT is a specialized epithelial covering that contacts the underlying lymphoid tissue and contains membranous or microfold “M” cells. Antigens, including virions (4, 5), are transported through M cells without the need for trauma or inflammation (6–9). Beneath the epithelium lie numerous DCs (10,11), which are important for capturing antigens and initiating T cell–mediated immunity (12). There are many observations in tissue culture indicating that DCs contribute to the capture of HIV-1 and SIV and subsequent transmission to T cells (13–17).

We first administered pathogenic SIVmac251 to juvenile or adult rhesus monkeys of either sex, by intravenous (iv) and tonsillar routes (18). For the latter, 200 to 3000 infectious units were applied directly to the palatine and lingual tonsils. To avoid trauma, we touched the tonsils lightly with a cotton-wool swab soaked with culture medium containing cell-free virus (19). Thereafter, the infection was monitored in the blood (Fig. 1). Using limiting dilution assays (20), we detected infectious virus in peripheral blood mononuclear cells (PBMCs) of all animals within 1 week, and titres peaked at 3 to 4 weeks (Fig. 1A). SIV levels then fell but persisted for more than a year (Fig. 1A). SIV antigen was detected in the blood of all animals with varying persistence over time (Fig. 1B). SIV-specific antibodies appeared by week 3 after infection in all monkeys (21), and the number of CD4+ T cells fell shortly after the acute stage of infection (Fig. 1C).

Figure 1

Markers of infection over time in rhesus macaques after intravenous (iv, upper panels) or tonsillar inoculation (lower panels) with SIVmac251. (A) Cell-associated viral load in PBMCs was determined by a limiting dilution coculture technique and the endpoint calculated. Viral loads are expressed as the number of infectious units per 106 PBMCs. (B) Levels of plasma antigenemia, measured as picograms of p27 capsid antigen per millilter. (C) Absolute CD4+ T cells in blood (but not determined in animal 7068) after FACS analysis for CD4 in a lymphocyte gate. The four-digit numbers are macaque designations; d indicates an AIDS-related death.

In the next experiment, two animals each were monitored at days 2, 3, and 4, and one animal at days 7 and 23. At each time point, the animals were euthanized and lymphoid tissues were taken for virology and histology examination. Infectious virus at a very low level was already detectable in the palatine tonsil at day 2 in one of two animals. By day 3, virus was detectable in the tonsil in both animals and reached plateau levels on day 4 and 7 (Fig. 2A). Virus was detected in PBMCs in one of two animals at day 3 and in both at day 4, but the titres of SIV were much less than those seen in the tonsil at day 7 (Fig. 2A). Among the lymphoid tissues we examined, the submandibular and retropharyngeal nodes that drain the tonsil had the most rapid infection (Fig. 2A). By day 3-4, virus had spread to the spleen and distal lymph nodes (mesenteric, axillary), but replication was slow at these time points. By day 23, virus titres in distal lymphoid tissues reached the levels found in tonsils at day 4-7 (Fig. 2A). Therefore, SIV infection was beginning locally in the tonsils and then spreading rapidly to other lymphoid tissues.

Figure 2

Cell-associated viral load in various lymphoid organs and blood in rhesus macaques (A) at different time points after tonsillar SIVmac251 infection or (B) at one time point after iv infection. Mononuclear cells were prepared by forcing the organs through nylon sieves, and their viral load was determined as described in Fig. 1. The absence of a bar means that no virus was detected; nd means the organ was not obtained; ne means the culture could not be evaluated because of contamination; dpi, days post infection; Ln, lymph node; tons, tonsil; cervic, cervical; occip, occipital; retrophar, retropharyngeal; submand, submandibularly; axill, axillary; mesent, mesenteric; Peyer's, Peyer's patches; PBMC, peripheral blood mononuclear cells. The asterisk in (B) indicates an organ in which the end point was not reached because of poor cell yield.

We examined the intestine to rule out uptake of swallowed virus into gut MALT. Abundant virus was only found by day 23 by in situ hybridization in the MALT of rectum, duodenum, and small intestine, so it is unlikely that swallowed virus was infecting the monkeys. To verify that the rapid tonsil infection was not due to trauma and entry of virus into the blood, we infected an animal with 2000 TCID50 intravenously and examined tissues at day 4. Infection was detected in the tonsil, but its extent was similar to that in spleen and only slightly greater than that in PBMCs and other lymphoid tissues (Fig. 2B). In contrast, application of SIV to the tonsil yielded an infection that was much more intense in tonsil than in spleen or blood (Fig. 2A), indicating that atraumatic application of SIV to the tonsil infects the MALT directly and not via the blood stream.

We measured the rate of infection in the MALT by enumerating the infected cells in sections using in situ hybridization with radiolabeled antisense RNA probes (22). Some infected cells were seen at day 3 in palatine and lingual tonsils; however, the number of infected cells then expanded rapidly, peaking between days 4 and 7 (Fig. 3). Small numbers of productively infected cells were observed at day 7 in distal nodes, spleen, and MALT, but only at day 23 were such cells numerous (Fig. 3).

Figure 3

Summary of the number of infected cells per square millimeter of section, according to in situ hybridization of the indicated lymphoid tissues, for all animals infected by the tonsillar route. The absence of a bar means that no infected cells were seen, for example, at day 2 in tonsils. NA, not available.

Infected cells were noted in all compartments of the tonsil (except for the stratified surface epithelium), including the lymphocyte-rich lymphoepithelium (LE) of the crypts (containing CD1a+immature DCs), the germinal centers (GCs) of the B cell follicles, and the outermost extrafollicular lymphoid tissue (ELT) (Fig. 4A). When counted, 86 and 89% of the infected cells were in the ELT at days 4 and 7, respectively, 8% in the LE at both times, and 6 and 3% in GCs. ELT is rich in CD3+ T cells (Fig. 4B), but most infected cells were not in classical deep T cell areas of the ELT, as defined by the presence of mature DCs with the p55 marker (Fig. 4C).

Figure 4

Localization and identification of infected cells in different tissues by in situ hybridization and immunocytochemistry. (A) Rapid productive infection in tonsil, 4 days after atraumatic application of SIVmac. The section is stained in red for CD1a, which marks immature DCs in the crypt epithelium [CE; higher magnification in (F)], and then counterstained with hematoxylin to outline the GC and ELT. Even at this early time point, there are numerous productively infected cells, hybridizing strongly with a radiolabeled antisense SIV probe (black silver grains). Infected cells are found along the CE, in ELT, and less frequently in the GC. (B) Same as (A), but the tonsil was examined at day 7 and stained brown for CD3 T cells. Black, productively infected cells are located in regions that are rich in T cells (arrowheads), that is, in the ELT rather than the GC. There are no infected cells in or against the overlying stratified squamous epithelium (arrow). (C) Same as (B), but the section was stained in red for p55, a marker of mature DCs that are found in the deeper regions of the tonsil. Most of the black, productively infected cells are located near the CE and ELT, which lack the red network of mature DCs expressing p55 (arrow). (D) The sublingual tonsil at day 7 after infection. The squamous epithelium of the tonsil, except for one cell close to the arrow, is not infected, but many underlying tonsillar lymphoid nodules are full of heavily infected cells (black profiles). (E) Double labeling of the CE for cytokeratin (red) and   SIV RNA (black silver grains) after proteinase K pretreatment. Productively infected cells (for example, at arrows) abut the keratin-positive CE. (F) Productively infected cells (black) in the CE are CD1a negative (red). (G) Productively infected cells (black) in the extrafollicular T cell areas are CD4+ (red, arrows). There is some loss in intensity of the radiolabel and CD4 images because these lie in different planes of optical focus. (H) Productively infected cells (black) lack the CD68 antigen (red) that is abundant in macrophages and found in smaller amounts on DCs. (I) High levels of infected, strongly CD68+(red) macrophages in an animal who died of SIV infection. Because CD68+ infected macrophages were difficult to identify between days 2 and 23 after atraumatic infection of the tonsil, we autopsied an animal that had died from SIV infection. In this animal, CD68+ infected cells were numerous as shown (arrow). (J) Infected cells (arrow) in the efferent lymphatic of the tonsil at day 4 after atraumatic application of SIV to the tonsil. Infected cells in a lymphatic could transmit the virus to the adjacent lymph nodes, for example, in the cervical nodes. (K) Relatively low levels of infection in distal lymphoid tissue at day 7 after atraumatic application of SIV to the tonsil. In this section of inguinal node, the deep T cell–rich cortex (T) is marked by the network of mature p55+DCs in red and occupies the entire center of the section. It is surrounded by peripheral cortex with B cell follicles but no red DCs. A single GC has several actively infected cells (arrows). (L) Same as (J), but the gut-associated lymphoid tissue is illustrated in a section of the colon. Only a single positive cell (arrow) is found. (M) Widespread infection of distal lymphoid tissue in the duodenum 3 weeks after atraumatic application of SIV to the tonsil. Original magnifications: (A and B) ×25, (C and D) ×6.25, (E) ×12.5, (F) ×100, (G and H) ×157.5, (I) ×100, (J) ×50, (K) ×6.25, (L and M) ×25.

A similar situation was noted in lingual tonsil, which consists of many lymphoid nodules, each connected to the pharynx by an invaginating crypt. The number of infected cells varied from one tonsil nodule to another, but most nodules were infected, and infected cells were numerous especially in p55-negative regions of the ELT (Fig. 4D). In the lingual tonsil at day 7, 74% of infected cells were in ELT, 6% in LE, and 20% in GCs.

We expected that productive infection with SIV would begin in stratified squamous epithelium that constitutes the external covering of the tonsil (Fig. 4B), tongue (Fig. 4D), and buccal cavity (21). This epithelium is comparable to the surface of the vagina and anus, and is rich in DCs that can capture and transport immunodeficiency viruses (13–17). When we infected the monkeys, we applied SIV directly to the tonsillar squamous epithelium. Breaks in this epithelium also could have provided a conduit for SIV to access susceptible lymphocytes. However, infected cells were not seen beneath the squamous epithelium that directly covered the lymphoid tissue at days 2 to 7 (for example, Fig. 4B), except for one cell in the squamous epithelium covering the tongue on day 7 (arrow, Fig. 4D), but they were abundant along many of the crypts in lingual and palatine tonsils (Figs. 4, A, C, and E). This observation suggests that SIV accessed MALT directly through antigen-transporting crypt epithelium (for example, by means of antigen-transporting M cells) after gentle application to noninflamed tonsils.

To identify the infected cells that hybridized with antisense SIV RNA probes, we double labeled the sections for RNA and different cellular antigens (23). The productively infected cells were labeled for CD4 (Fig. 4G) and CD3 (21), but not for the immature DC marker CD1a (Fig. 4F) and the macrophage marker CD68 (Fig. 4H). As a positive control for macrophage infection, we studied a monkey who died from SIV-AIDS. This CD4+ T cell–depleted animal had many infected CD68+ tonsillar macrophages (Fig. 4I).

At day 3 and day 4, we also noted productively infected cells in the efferent lymphatics of the palatine tonsil (Fig. 4J). The lymphatics were engorged with lymphocytes (“sinus lymphocytosis”) at these early time points, suggesting that productively infected cells can spread to adjacent lymphoid organs via the efferent lymph (24).

The systemic spread of infection was also followed. A few productively infected cells were detected in both the follicular and ELT of spleen (21), lymph nodes (Fig. 4K), and MALT (Fig. 4L). By day 23, the infection was active in all lymphoid tissues (Fig. 4M), when SIV was trapped on follicular dendritic cells of the GCs as described (24).

Ruprecht and colleagues have shown that SIV can be transmitted by the oral route (24, 25), a route that is also implicated in HIV-1 transmission (1, 27). Our finding that MALT is an explosive site for oral transmission of SIV suggests that oral MALT or Waldeyer's ring is a potential site for HIV-1 transmission during oral sex, parturition, and breast feeding, and by extension, rectal MALT can mediate anogenital transmission. Two other findings of our study were unexpected: (i) infection expands rapidly within MALT, peaking in the space of a few days, and (ii) there is a lack of infection in the stratified squamous epithelium that lines the pharynx and surfaces of the palatine and lingual tonsil (Fig. 4, B and D), even though it is this kind of DC-rich epithelium that is thought to be critical in sexual transmission across the vagina and anus.

The distinguishing feature of MALT is the presence of a specialized overlying epithelium through which antigens can be transported in the absence of any trauma [reviewed in (6–9)]. In the gut-associated MALT, the dome of the Peyer's patches and solitary lymphoid nodules are covered by a specialized columnar epithelium in which M cells transport antigens including proteins, small particles, and virions (4, 5, 7, 9, 28). In the pharyngeal MALT or tonsils (nasal, palatine, sublingual), there are deep invaginations or crypts lined by a ramifying, keratinized, lymphocyte-rich epithelium with M cells (29, 30). M cells could have a second role beyond virion transport, that is, the transport of antigens and inflammatory stimuli that activate DCs and T cells locally. This lymphoepithelium was shown to be an active site for chronic HIV-1 infection (10, 31), but it is now evident that this is an efficient area for transmission and acute infection as well (Figs. 2,3, and 4).

When studied at the single-cell level, the infected cells were almost entirely CD4+ T cells. The infected cells could be double labeled for viral RNA and for CD3 and CD4, but not for the CD1a, p55, and CD68 antigens that are abundant on immature DCs, mature DCs, and macrophages, respectively. However, recent tissue culture studies indicate that DCs can initiate infection, but once begun, activated T cells become the major site for infection (32), presumably because T cells have much higher levels of CD4 and CCR5 than do DCs (33). Possibly our in situ method would not detect DCs with low levels of viral transcripts.

Our findings, by documenting the rapid development of an infection with immunodeficiency virus within MALT, emphasize new demands for the development of an HIV-1 vaccine. Because viral replication is so robust, reaching a plateau locally in 4 to 7 days and systemically in less than 3 weeks, vaccine-induced immune mechanisms have to be activated very quickly after mucosal transmission.


View Abstract

Navigate This Article