Gastrointestinal Tract as a Major Site of CD4+ T Cell Depletion and Viral Replication in SIV Infection

See allHide authors and affiliations

Science  17 Apr 1998:
Vol. 280, Issue 5362, pp. 427-431
DOI: 10.1126/science.280.5362.427


Human and simian immunodeficiency virus (HIV and SIV) replicate optimally in activated memory CD4+ T cells, a cell type that is abundant in the intestine. SIV infection of rhesus monkeys resulted in profound and selective depletion of CD4+ T cells in the intestine within days of infection, before any such changes in peripheral lymphoid tissues. The loss of CD4+ T cells in the intestine occurred coincident with productive infection of large numbers of mononuclear cells at this site. The intestine appears to be a major target for SIV replication and the major site of CD4+ T cell loss in early SIV infection.

It is now thought that ongoing HIV replication results in a continual loss of CD4+ T lymphocytes that is nearly balanced by the production of new CD4+ T lymphocytes (1). This model explains some of the puzzles of HIV infection, but the events that occur in the initial stage of infection remain largely unexplored. Although it is clear that HIV targets lymphoid tissue, nearly all studies in this area have focused on peripheral blood and lymph nodes. These studies overlook the fact that the gastrointestinal tract contains most of the lymphoid tissue in the body (2, 3). Furthermore, it is likely that the behavior of HIV in the unique immunologic environment of the intestinal mucosa differs from that observed in the periphery.

The gut-associated lymphoid tissue (GALT) consists of organized lymphoid tissue (Peyer's patches and solitary lymphoid follicles) as well as large numbers of activated memory T lymphocytes diffusely distributed throughout both the intestinal lamina propria and epithelium. The proportion of activated memory CD4+ T cells is much greater in the intestinal lamina propria than in peripheral blood or lymph nodes (2, 4-6). Because HIV replicates most efficiently in activated memory CD4+ T cells (7), the intestinal tract may be the preferred target for initial infection and replication of HIV. The effects of primary HIV infection on GALT are difficult to assess in humans, largely because of the difficulties involved in examining the intestinal tract within days of infection. The SIV macaque model provides a useful tool for examining the pathogenesis of acquired immunodeficiency syndrome and the early events in infection (8).

Groups of macaques were inoculated intravenously with the following molecular clones of SIV (9): pathogenic SIVmac239 (10); a macrophage-competent derivative of SIVmac239, designated SIVmac239/316 (11); and an attenuated,nef-deletion derivative of SIVmac239, designated SIVmac239Δnef (12). Two animals in each group were killed at 7, 14, 21, and 50 days after inoculation. Lymphocytes were separately harvested from the intestinal epithelium and lamina propria as well as from the blood, spleen, and both mesenteric and axillary lymph nodes of each animal (13). Although intestinal intraepithelial lymphocytes (IELs) and lamina propria lymphocytes (LPLs) were collected and analyzed separately, only minor changes were detected in the IEL population (14); thus, only the LPL data are presented here. Lymphocytes from all tissues were examined for expression of CD3 (a pan–T cell marker), CD4 (T helper cells), CD8 (cytotoxic/suppressor cells), CD25 (interleukin-2 receptor, activated T cells), and CD45RA (naı̈ve T cells) by multiparameter flow cytometry (13). To localize virus and examine the numbers and phenotypes of infected cells, we performed in situ hybridization for SIV and immunohistochemistry on intestine and peripheral lymphoid tissues (15).

Infection with pathogenic SIV (SIV- mac239 or SIVmac239/316) consistently resulted in rapid and profound depletion of CD4+ cells exclusively in the lamina propria of the intestinal tract (jejunum, ileum, and colon) (Fig.1A). These declines were evident by 7 days after infection, as compared to controls or animals infected with SIVmac239Δnef (P < 0.01) (16). The nadir of intestinal CD4 depletion was reached by 14 to 21 days after infection in SIVmac239-infected animals but was delayed to 50 days in SIVmac239/316-infected animals.

Figure 1

Sequential changes in CD4+ and CD8+ T cell populations in lymphoid tissues of macaques 7, 14, 21, and 50 days after infection with molecular clones of SIV. (A) Intestinal lamina propria; (B) lymph nodes, spleen, and blood. Each bar represents the mean of two (infected) or four (uninfected control) animals.

Immunohistochemistry on intestinal segments confirmed the CD4+ T cell depletion in the lamina propria and to a lesser extent in the Peyer's patches and solitary lymphoid nodules (14). Immunohistochemistry and quantitative analysis of CD3 did not reveal any significant change in the total number of T cells in the intestine (14). These results, in conjunction with the observed increase in the percentage of CD8+ T cells (Fig.1), indicate that the CD4+ T cell depletion was accompanied by an increase in absolute numbers of CD8+ T cells.

In marked contrast, there were minimal changes in the percentages of CD4+ lymphocytes in the blood, spleen, and lymph nodes from these same animals at the same time points. Slight decreases in CD4+ cells were observed in the axillary and mesenteric lymph nodes of animals infected with pathogenic clones of SIV, but these differences were not significant (Fig. 1). Percentages of CD4+ T cells in the blood were increased at days 7 and 14 after infection, but the absolute numbers were not significantly altered at these time points.

To rule out the possibility of CD4 down-modulation or gp120-mediated CD4 masking in the intestinal CD4 depletion, we performed three-color analysis with antibodies to CD3, CD4, and CD8. No significant increase in the proportion of CD3+CD4CD8(double-negative) T cells was observed, which would have been expected if either CD4 masking or down-modulation were involved (Fig. 2). This strongly suggests that the CD4+ T cells were eliminated from the intestinal tract by lytic or apoptotic mechanisms or, alternatively, by altered trafficking of mucosal lymphocytes.

Figure 2

Three-color flow cytometry dot plots comparing lymphocytes isolated from the intestinal lamina propria (jejunum LPL) with those from the axillary lymph node (LN) from a normal uninfected macaque and from macaques infected with SIVmac239 at 7 and 21 days after infection (pi). Each set of top and bottom panels corresponds to the same animal. Plots were generated by first gating through lymphocytes and then through CD3+ T cells.

To determine whether these results were unique to molecular clones of SIVmac239, we inoculated two additional animals with uncloned SIVmac251. When these animals were killed at 14 days after infection, they also had marked CD4+ T cell depletion in the intestinal lamina propria (CD4+ T cell percentages decreased to 8% or less in the jejunum and colon and 15% or less in the ileum) without significant changes in peripheral lymphoid tissues (14). The speed and extent of the CD4+ T cell depletion in these animals was indistinguishable from what was observed in animals infected with SIVmac239 (Fig. 1A). An additional three animals inoculated with SIVmac251 were killed 5 months after infection to determine whether the CD4+ cells returned after the acute phase of infection. CD4+ T cells were markedly decreased in the intestine of these animals as well (to less than 10% of LPLs) (14). Thus, intestinal CD4+ T cell depletion is consistent for different pathogenic strains of SIV and appears to persist throughout the course of infection. Selective depletion of CD4+ T cells from the intestine also appears to occur in HIV-infected humans. Several studies have suggested that CD4+ T cell depletion is more pronounced or occurs sooner in the intestine than in peripheral blood, even in the relatively early stages (first several months) of HIV infection (5, 17, 18).

In contrast to pathogenic SIV, infection with attenuated virus (SIVmac239Δnef) did not result in a significant loss of intestinal CD4+ cells (Fig. 1) but was associated with marked T cell activation, as determined by up-regulation of CD25 (interleukin-2 receptor). Marked increases in CD25 coexpression were detected in both CD4+ and CD8+ T cell subsets in SIVmac239Δnef-infected animals (Fig.3).

Figure 3

Flow cytometry dot plots comparing CD25 expression on intestinal LPLs from an uninfected macaque and from macaques infected with SIVmac239 or SIVmac239Δnef. Plots were generated by gating on lymphocytes.

The virulence of the molecular clones used here correlated with the rapidity and degree of intestinal CD4+ T cell depletion, and correlated inversely with CD25 expression on the remaining CD4+ cells. Infection with SIVmac239 resulted in rapid and profound CD4+ depletion in the intestine, and it decreased the relative expression of CD25 on the remaining CD4+ intestinal lymphocytes. Infection with SIVmac239Δnef was associated with marked up-regulation of CD25 on intestinal CD4+ T cells and to a lesser extent on CD8+ cells (Figs. 3 and 4). The effects of infection with SIVmac239/316 on CD25 expression were intermediate between the others, with the highest expression 21 days after infection. Infection with SIVmac239/316 resulted in more CD25 expression on CD8+ cells and on the remaining CD4+ T cells than did infection with SIVmac239 (Fig. 4). Expression of CD25 was consistently higher in the jejunum of both normal and SIV-infected macaques than in the ileum or colon. Furthermore, expression of CD25 was consistently higher on CD4+ cells than on CD8+ lymphocytes (Fig. 4). Although more organized lymphoid tissue is present in the ileum, these data suggest that jejunal lymphocytes (which are mostly LPLs) are more activated than ileal lymphocytes, which contain many organized lymphoid nodules containing fewer activated cells. This is consistent with the recent finding that CD4 depletion in the intestine of HIV-infected patients selectively occurs in the lamina propria rather than in the organized lymphoid tissue of the intestine (18).

Figure 4

CD25 expression by CD4+ and CD8+ lymphocytes in the intestinal lamina propria of macaques 7, 14, 21, and 50 days after infection, as measured by flow cytometry. Data were generated by first gating on CD4+ or CD8+ lymphocytes. Each bar represents the mean of two (infected) or four (control) animals.

Our findings also confirm that the normal intestine contains larger numbers of activated memory CD4+ T cells than do peripheral lymphoid tissues. The expression of CD25 is consistently higher, and the expression of CD45RA (naı̈ve cells) is consistently lower, on intestinal lymphocytes than on lymphocytes obtained from the peripheral lymphoid tissues or mesenteric lymph nodes (2, 4-6). Although a suitable memory cell marker has not been found that cross-reacts with the rhesus macaque, these data confirm that the intestinal mucosal lymphoid tissue of rhesus macaques is similar in composition to that of humans. Moreover, these data indicate that the major target cells of HIV and SIV are far more abundant in the intestinal tract than in peripheral tissues.

Further insights into the early pathogenesis of SIV were gained by examining the distribution and phenotype of SIV-infected cells in the intestine by in situ hybridization, alone and combined with immunohistochemistry, at sequential time points after SIVmac239 infection. Coinciding with the onset of CD4 depletion (7 days after infection), many SIV-infected lymphocytes were present in the intestinal tract of animals infected with pathogenic clones of SIV (Fig. 5A). Also, there were more virus-infected cells in the intestine than in the peripheral lymphoid tissues, as previously described (19). Initially, many infected cells were present throughout the intestinal lamina propria as well as in the T-dependent areas of organized lymphoid nodules (Peyer's patches and solitary lymphoid follicles). At later time points (21 and 50 days after infection) when CD4+ T cell depletion was marked, the number of virus-infected cells was greatly diminished and the remaining infected cells were mainly limited to the organized lymphoid nodules. Conversely, the percentage of infected macrophages increased at later time points. Although infected macrophages were rare in animals at day 7 after infection, combined immunohistochemistry for macrophages (HAM-56) and in situ hybridization for SIV demonstrated an increase in the percentage of infected macrophages at later time points, which were mainly located within organized lymphoid nodules (Fig. 5B). This is consistent with the results of previous experiments using uncloned SIVmac251 (20).

Figure 5

Localization of SIV-infected cells in the intestine of macaques. (A) In situ hybridization for SIV in the jejunum showing large numbers of infected lymphocytes (black) in the lamina propria 7 days after infection with SIVmac239. Scale bar, 100 μm. (B) Intestinal lymphoid nodule double-labeled by in situ hybridization for SIV (black) and immunohistochemistry (brown) for HAM-56 (macrophages) in a macaque 21 days after infection with SIVmac239. Note the presence of several SIV-infected macrophages having large black nuclei and brown cytoplasm (arrowheads). A few SIV-infected lymphocytes are also visible in this field (arrows). Scale bar, 50 μm.

Combined, the in situ hybridization and flow cytometry data suggest the following course of events: Initially, large numbers of activated memory CD4+ T cells are constitutively present throughout the intestinal lamina propria. These initial target cells are rapidly infected and serve as sources for viral amplification and dissemination. However, this large pool of activated memory CD4+ T cells in the intestine is rapidly depleted, leaving only naı̈ve lymphocytes and macrophages to serve as viral host cells. The decrease in the pool of susceptible cells results in a decline in viral load in the tissue. However, new lymphocytes are continually recruited to or produced in the organized lymphoid tissues (GALT and lymph nodes), and these may become activated by antigenic stimulation to serve as fresh host cells for viral replication, thus perpetuating the infection. This would explain why the numbers of virus-infected lymphocytes are reduced in GALT at the later time points of infection.

These observations strongly suggest that intestinal lymphoid tissue is a crucial target organ in the initial pathogenesis of SIV infection. On the basis of this evidence, we hypothesize that the gastrointestinal tract, and not the peripheral lymphoid tissue, is the major site of early SIV and HIV replication and amplification, resulting in profound and rapid CD4+ T cell loss. If so, then the acute phase of infection with SIV or HIV should be viewed primarily as a disease of the mucosal immune system. There are important ramifications of considering SIV/HIV as a mucosal disease, including the design of therapies that target the intestinal tract and vaccines that stimulate an effective mucosal immune response. These results suggest that an important advantage of a modified live virus vaccine that targets the intestinal mucosa would be its ability to stimulate an appropriate immune response at the site involved in the earliest stages of viral infection.

  • * To whom correspondence should be addressed. E-mail: alackner{at}


View Abstract

Navigate This Article