HIV infection and the gastrointestinal tract

Introduction Gastrointestinal mucosa and other mucous membranes occupy a unique anatomical niche: the interface between a sterile, internal environment and a contaminated, external environment. There is a polarity in epithelia that is different from all other tissues, in which one side of the epithelial cell faces ‘self’, whereas the other side faces ‘non-self’. Epithelial cells require intimate contact with the external environment in order to carry out their primary functions, e.g. nutrient absorption in the intestine, gas exchange in the lungs, etc. The need for such contact with the external environment makes mucous membranes inherently vulnerable to infectious organisms, because of the lack of a strong physical barrier. The problem is compounded by a large surface area that characterizes most mucous membranes. For example, in the intestine, adaptations such as mucosal folds, villi and microvilli lead to a 600-fold increase in surface area beyond that present in a simple tube. An elaborate system has evolved to protect the gastrointestinal tract from pathogens, which is part of a common mucosal immune system. The afferent and efferent limbs of the mucosal immune system are anatomically separate, but intermingled. The afferent limb represents the cell populations and structures involved in the production of an immune response, e.g. antigen presentation and lymphoproliferation, whereas the efferent limb involves the cell populations and structures directly involved in the immune response, e.g. antibody producing and cytotoxic mononuclear cells. The afferent system includes discrete lymphoid follicles (Fig. 1), which are overlaid by a follicle-associated epithelium, including microfold, or M cells. M cells are specially adapted epithelial cells, without long microvilli or overlying mucus, which promote the transcellular transit of particulate antigens to antigen-presenting macrophages that lie immediately beneath the cells, and then to mucosal lymphoid follicles (Fig. 2). Further maturation of cells occurs during migration through mesenteric lymphoid follicles and the systemic circulation, via the thoracic duct. After further maturation in the spleen and elsewhere, lymphoid cells return to the intestinal mucosa. The lymphoid cells in the efferent compartments, the epithelial layer, and the lamina propria are scattered diffusely throughout the intestine, in keeping with the defensive function along the length of the intestine. Differences in subpopulations of lymphoid cells, and in immune functions, are found in the epithelial layer and in the lamina propria.Fig. 1: Whole mount preparation demonstrating a lymphoid follicle in the lamina propria and submucosa of the colon. (Hematoxylin and eosin, 125×).Fig. 2: Diagram of an M cell. The basal surface of the cell contains a pocket in which lymphocytes (L) and macrophages (MAC) can be found. The M cell (M) efficiently transports particulate material to the underlying lymphoid cells. Reprinted with permission from Amerongen et al. [1].Mucosal immunity has many homologies to systemic immunity, and several distinctions. Although many consider the intestinal mucosa to be physiologically inflamed, it is quite uninflamed, despite microbial provocation. Secretory immunoglobulins and the process of immune exclusion from the internal environment, as well as the lack of complement activation by IgA, reflect the non-inflammatory tone of the intestines. Importantly, inflammation impairs mucosal function by several mechanisms. Intestinal mucosa has long been known to be a target for HIV and related viruses, based upon the route of exposure and other aspects of mucosal immunology (Fig. 3). The aim of this review is to document the progress in the field of HIV infection in the gastrointestinal tract over the past few years. Whereas new information is partly recapitulation of earlier work, discussion will concentrate on recently published studies.Fig. 3: RNA in-situ hybridization for HIV in rectal mucosa. (a) Diffuse labelling in the germinal center of a lymphoid follicle, corresponding to trapping of immune complexes containing virions by follicular dendritic cells. Original magnification 100×. (b) Labelling of a single cell in the lamina propria. Original magnification 440×.Viral penetration Early conceptions of HIV transmission were that it occurred through mucosal trauma or other breaches in the physical barriers of the genital epithelium or perineal skin. However, infection also occurs after atraumatic inoculation of mucous membranes. The route could be through epithelial cells, between epithelial cells, or by the normal route for particulate sampling, the M cells. Initial studies, using rabbit ileum containing Peyer's patches including M cells, showed the penetration of HIV from lumen to lamina propria (Fig. 4) [1]. Fotopoulous and colleagues [2] constructed polarized monolayers using an intestinal cell line (Caco-2), which resembles M cells when grown in the presence of B lymphocytes. A lymphocyte tropic, or X4 HIV strain, crossed the M-cell monolayers and infected underlying CD4 target cells. Transport required both lactosyl cerebroside and CXCR4 receptors, which were expressed on the apical surface of the Caco-2 cells. Antibodies specific for each receptor blocked transport. In contrast, a monotropic, or R5, strain was unable to cross the M-cell monolayers and infect underlying monocytes. Caco-2 cells do not express CCR5 under normal circumstances, but transfection of these cells with CCR5 complementary DNA enabled transport of R5 virus, demonstrating that HIV-1 transport across M cells is receptor mediated. However, the investigators also studied human tissues and found that follicle-associated epithelium expressed galactosyl ceramide and CCR5, but not CXCR4.Fig. 4: Transmission electron photomicrograph demonstrating penetration of HIV through rabbit M cells, in vitro. Adherence (a) and (b), followed by transport in endosomal vacuoles (c), and (d) delivery of virions to the basal epithelial pocket and close association with a lymphocyte (L). Reprinted with permission from Amerongen et al. [1].Van de Perre [3] and Meng and colleagues [4] also investigated viral penetration in the small intestine. They noted that the upper gastrointestinal tract is a principal route of HIV-1 entry in cases of mother-to-child transmission [3]. The phenotype of the newly acquired virus is predominantly R5 and not X4, although both viruses are frequently inoculated onto the mucosa. Primary cultures of jejunal epithelial cells expressed galactosyl ceramide and CCR5, but not CXCR4, and the cells transferred R5, but not X4, viruses to target cells [4]. Transfer was not inhibited by the fusion inhibitor T-20 (Enfuvirtide; Roche Pharmaceuticals, Nutley, NJ, USA), but was substantially reduced by colchicine and by low temperature, consistent with endocytotic uptake and microtubule-dependent transcytosis by the epithelial cells. Bouhlal and colleagues [5] demonstrated that infection of HT29 cells was enhanced twofold when semen was added to HIV before incubating with epithelial cell cultures. The enhancing effect of semen was complement dependent, as evidenced by blockage of the generation of C3a/C3adesArg in semen by heat or ethylenediamine tetraacetic acid treatment, and by the suppression of semen-dependent enhancement with monoclonal antibodies directed to complement receptor 3. The investigators concluded that the activation of complement in semen may facilitate the infection of epithelial or lamina propria cells. Hocini and Bomsel [6] tested the specificity of transcytosis across the epithelium by incubating with secretory IgA or IgG, purified from colostrum, and showed that they inhibited both transcytosis and the infection of target cells. Devito and colleagues [7] studied highly HIV-exposed but uninfected individuals, and also found that IgA from plasma, saliva, and mucosal secretions inhibited HIV-1 transcytosis across an epithelial cell membrane system, whereas IgA isolated from low-risk, healthy control subjects did not. Compartmentalization of viral infection The results of several studies have suggested that there is compartmentalization of viral infection between blood and mucosa. Selective quasispecies transmission after systemic or mucosal exposure of macaques to SIV was studied by Nieldez and colleagues [8]. Macaques were inoculated intravenously, intrarectally, or intravaginally with SIVmac251. The patterns of virological and immunological events differed substantially between vaginally inoculated animals, who had transient viremia and late seroconversion, and intravenously or intrarectally inoculated monkeys, who had persistent viremia and early seroconversion. Analysis of the envelope gene nucleotide sequences revealed specific viral variants that were associated with vaginal transmission. Couedel-Courteille and colleagues [9] also studied rectal infection in rhesus macaques inoculated with SIVmac251. Infection was established initially in certain paracolic lymph node chains draining the rectum. Specific sequences of infection within lymphoid follicles and among different lymphoid chains were noted and contrasted with those in intravenously infected animals. The compartmentalization of infection is partly caused by viral factors. Harouse and colleagues [10] demonstrated different pathogenic sequelae from infection by X4 and R5 SHIV-derived viruses. The R5 virus clone caused a dramatic loss of intestinal CD4 T cells followed by a gradual depletion in peripheral CD4 T cells, whereas infection with the X4 clone caused a profound loss in peripheral T cells that was not paralleled in the intestine. Other evidence of viral compartmentalization was provided by Poles and colleagues [11], who compared genotypic and phenotypic resistance patterns of HIV-1 RNA isolated from colonic mucosa, plasma and peripheral blood mononuclear cells. There was high concordance in detecting mutations in the reverse transcriptase and protease genes as well as in phenotypic resistance patterns. However, different genotypic features in isolates from plasma and the other tissue compartments were observed in some cases. Cell targets for HIV infection Several groups have examined cell targets in the intestine. Initial studies documented the ability of HIV to replicate in intestinal epithelial cell lines [12,13], as well as the presence of HIV DNA, RNA, and protein antigens in intestinal mucosa from clinical specimens [14–16], although convincing evidence of the high-level replication of HIV in epithelial cells in clinical samples is lacking. Smith et al. [17] and Meng et al. [18] examined purified lamina propria lymphocytes and macrophages from the normal human small intestine. Lamina propria lymphocytes expressed CD4, CCR5, and CXCR4. In contrast, lamina propria macrophages expressed CD4 but neither CCR5 nor CXCR4 (Fig. 5). Intestinal lymphocytes supported replication by R5 and X4 isolates of HIV-1, whereas lamina propria macrophages were permissive to neither. Regulated upon activation: normal T cell expressed/secreted (RANTES), macrophage inflammatory protein (MIP) 1α, and MIP-1β inhibited the infection of intestinal lymphocytes by an R5 viral isolate, suggesting that R5 infection was mediated by CCR5. The investigators concluded that resident lamina propria lymphocytes, not macrophages, are the target mononuclear cell for HIV-1 infection in intestinal mucosa. Lapenta and colleagues [19] also showed that human lamina propria lymphocytes, in contrast to autologous peripheral blood lymphocytes, are permissive to both X4 and R5 strains. Anton and colleagues [20] demonstrated enhanced levels of functional CCR5 and CXCR4 on human mucosal T lymphocytes obtained from intestinal biopsies, compared with peripheral blood lymphocytes. In another study, Poles and colleagues [21] showed that a greater susceptibility of mucosal lymphocytes to HIV-1 infection compared with peripheral blood lymphocytes was associated with a greater expression of chemokine receptors.Fig. 5: Flow cytometric analysis of intestinal lymphocytes (left) and macrophages (right). Both CD4-positive and CD4-negative intestinal lymphocytes expressed CCR5 and CXCR4, whereas intestinal macrophages expressed HLA-DR antigen but neither CCR5 nor CXCR4. Of note is the fact that blood monocytes (insets) did express CCR5 and CXCR4. Reprinted with permission from Meng et al. [18].These data, in combination with studies of viral penetration, suggest that the primary infection of intestinal mucosa involves R5 virus, which crosses the epithelial layer by several mechanisms, including transcytosis across M cells, and infects lamina propria lymphocytes. Infection of intestinal mononuclear cells with X4 virus is also possible, but by a different, internal route, the systemic circulation, as opposed to the external environment. HIV pathogenesis The topic of immunopathogenesis involves two separate but overlapping phenomena, the loss of lymphoid cells and the loss of immune function. Several studies of clinical specimens, using immunohistochemistry or flow cytometry, showed a disproportionate early loss of CD4 lymphocytes in lamina propria, compared with peripheral blood [22], with greater relative losses of CD4 lymphocytes from lamina propria than from lymphoid follicles, at least early in the disease course [23]. Veazey and colleagues [24] showed that primary SIV infection of rhesus monkeys with SIVmac239 resulted in profound and selective depletion of CD4 T cells in the intestine within days of infection, before any changes occurred in peripheral lymphoid tissues. Similar findings were reported by Kewenig and colleagues [25]. Fackler et al. [26] and Schmidt and colleagues [27] found a homogenous distribution of HIV proviral DNA in peripheral blood and intestinal mucosa but more p24 production in lamina propria. Their results suggested that mucosal HIV production is upregulated at the transcriptional or translational The effect of HIV infection upon cell by has also been There to be many more cells than cells, suggesting that uninfected cells may as as suggested in peripheral lymph and colleagues showed that the of normal lamina propria lymphocytes, isolated from specimens, with HIV-1 the of cells observed during The process was mediated by and was related to the of RNA by HIV-1 could to the depletion of lamina propria T cells, of viral In other studies, and colleagues studied lymphocytes in primary SIV infection, and documented an early of T cells, but depletion of resident T cells, the cell that includes CD4 cells. The showed an increase in the expression of but between expression and viral suppression and colleagues studied the effect of HIV-1 infection on mucosal lymphocyte and activation in infected and control The of cells was in the intestine, whereas the of activation as well as of mucosal were different from for this is that the subjects were with of infected in the of the The of mucosal inflammation in the pathogenesis of immune is a In past studies, showed among clinical changes on rectal and mucosal HIV protein but not with the presence of HIV protein and RNA expression during disease and were in the of the studies demonstrated a lymphoid in the of the disease by the expression of a of T and The of cells in the lamina propria during disease the of these data, that clinical and intestinal are directly related to the presence of HIV in the mucosa. and colleagues found of mucosal inflammation in from those with inflammatory and from healthy including the expression of CCR5 and CXCR4, and HIV HIV to a clinical the was not for specificity but not to the for the of in a is to an with The pathogenesis of this is not the as the pathogenesis of immune However, the that it is related directly to HIV infection has been from which examined the of days of combination documented in of gastrointestinal of an association However, those studies did not the presence of HIV in mucosa to There are two lines of the pathogenesis of related to mucosal HIV of intestinal have been at the of who examined the of epithelial by a of intestine in an or by an epithelial cell that cell and normal and in of intestinal a that can be and to resistance can be by changes in transcellular or the may be caused by the of the between cells. changes in of showed normal resistance in subjects without but resistance in with An increase in was suggesting as opposed to There was evidence of a in absorption or The effect was by incubating a with of mononuclear cells The effect was also by with and inhibited by with The investigators concluded that the effect of HIV upon mucosal was mediated. was not the effect was caused by the of or epithelial cell both of which could increase more some have also been found to epithelial cell function and colleagues an for Epithelial cell is partly by the of such as the increase intestinal and promote was studied by immunohistochemistry for in and control subjects A of was in small and colonic epithelial cells from and The results were in subjects with or without with combination for as to an increase in the of although to levels than in The concluded that some of HIV infection to changes in epithelial cells. The that HIV directly epithelial cells was investigated Other studies by and colleagues and by and colleagues showed that of the intestinal cell with to the to an increase in which was associated with both and epithelial resistance (Fig. The effect was with both as well as The receptor through which these changes was investigated further by and colleagues CCR5 or CXCR4 were not to be because they are expressed on the surface of the epithelial whereas is more to from mononuclear cells in the lamina propria, which faces the basal and The protein receptor is a for HIV and SIV and viral fusion and infection of although it is an for protein was found in of small and colonic mucosa, lymph and but not in many other immunohistochemistry using antibodies to was noted in the basal membrane (Fig. was examined in cells with and were obtained with of was inhibited by which is a selective protein and by with antibodies to or to galactosyl but not by antibodies to CXCR4. the changes in were found to be HIV strain The concluded that HIV is a of and is of epithelial cell HIV and that of et al. from the of are not HT29 cells were not or to for then with antibody and antibody and examined by with to in part with permission from et al. of a jejunal from an using an antibody to can be on both the and basal membranes of the epithelial cells. Reprinted with permission from et al. immune in HIV infection with systemic immunity, the clinical of HIV infection with immune Intestinal disease caused by or is in with peripheral blood CD4 lymphocyte than The of highly has been associated with mucosal immune and a in the of Several studies examined mucosal immune function in HIV and colleagues studied the of cells in the lamina propria of and control subjects by The of cells was in both Whereas the of cells were in both the and from compared with the of cells was in the colon. changes paralleled the results of studies of in The investigators also the plasma cell and noted that it was suggesting that and and of mucosal B cells. and colleagues studied secretory immunity during the course of SIV infection at and They the in IgA and the increase in They were unable to IgA, although there was they also by in a in mucosal function. In other studies, and colleagues phenotypic and functional of cells isolated from and showed normal or and despite CD4 and colleagues studied the by flow in SIV infection and found the However, they noted that the cells in intestinal mucosa were predominantly cells, and that the loss of CD4 cells was by an increase in cells. T lymphocytes may be in and colleagues isolated from and rectal specimens of infected They were and in their function. and colleagues using a infected SIV showed that the of SIV in gastrointestinal mucosa is to that in rhesus In a of study, and colleagues an HIV followed by rectal with an HIV-1 virus, to that resistance to the transmission of HIV-1 virus can be by present at the mucosal of The resistance was by cells in and required mucosal as provided The was by mucosal delivery of with the results that the of mucosal may be for the of a viruses through a mucosal route, such as In other studies, and colleagues SIV in the jejunal mucosa of vaginally with the were from colonic vaginal after systemic infection, but with plasma viral and disease and colleagues studied but macaques who and were to mucosal SIV monkeys more and than animals. In SIV infection was by cell from these mucosal IgA antibodies have been reported to with in highly but uninfected individuals, but are and colleagues rectal from and control subjects in a IgA levels did not between and uninfected IgA antibodies were in most whereas was found in most rectal of and colleagues also found low levels of IgA antibody production after the of in vitro. suggest that immunity may be more than secretory immunity in immunity to not be that secretory immune function is in the of a infection of the lamina propria. However, could be an of in the intestine. of few of the of in the gastrointestinal tract have been compared the of on intestinal mucosa and peripheral blood by studies immediately and days after combination of the were most had gastrointestinal and had was associated with in gastrointestinal Similar relative in and in CD4 lymphocyte were found in blood and mucosa (Fig. was also associated with a in the of cells as by in-situ a that with the in mucosal viral and colleagues the of of in mucosal and peripheral blood mononuclear cells, and found that levels of HIV-1 RNA in in peripheral blood and mucosa, a in virus in both and colleagues examined the effect of on the of HIV RNA and DNA from rectal in who have with In that to was associated with the suppression of RNA, but not DNA, suggesting viral infection in cell studies in from obtained immediately before combination and days (a) The tissue RNA by from et al. (b) The plasma RNA by in intestinal mucosa after the of can mucosal immunity with clinical including the of such as and Although the of immune is it be For example, and colleagues studied the expression of CCR5 and the intestinal receptor on subpopulations of lymphocytes, and found a profound of lymphocytes and lymphocytes with despite the control of viral replication with immunity in on lymphocytes to the suggest that immune may despite and colleagues examined lymphocyte and mucosal cell expression in with with had reduced CD4 which with treatment, in with mucosal cell expression was in all with at to normal by in without but in with In intestinal mucosa has long been known as a target for HIV and related viruses. penetration through the epithelium to be receptor mediated and to CCR5 viruses. that CXCR4 may the intestinal via the In intestinal lymphocytes to be the target for There is compartmentalization of HIV infection, and different sequelae based on both viral and factors. There is an early and disproportionate loss of CD4 lymphocytes from the mucosal compared with peripheral The of cell to be and may both infected and cells. immunity includes both and immune with the more of have that mucosal HIV is as to suppression as is plasma and there is a for immune in intestinal mucosa. HIV may a in producing intestinal in mucosal which are associated with may from from mononuclear cells in the lamina propria, or from with certain expressed on the membranes of intestinal epithelial cells, to changes and the of