The ‘immunologic advantage’ of HIV-exposed seronegative individuals

Introduction In 1989, a curious phenomenon was described: HIV-specific T-cell responses to the viral envelope and core proteins could be detected in antibody-positive and antigen-negative sexual partners of known HIV-positive men [1]. Two other reports confirmed that initial observation on a total of six exposed seronegative (ESN) individuals, and the author raised the possibility that exposure to HIV that did not result in seroconversion and infection could be associated with the exclusive priming of T lymphocytes [2,3]. Analyses performed in different cohorts of individuals at high risk of HIV infection, including healthcare workers parenterally exposed to HIV and healthy newborns of HIV-infected mothers, revealed that HIV-specific CD4+ T helper cells, but not antibodies, were present in these persons [4,5]. These observations led to the hypothesis that viral exposure resulting in the exclusive priming of HIV-specific T cells could be associated with protection against the establishment of HIV infection [6]. This hypothesis was greatly strengthened by the independent observations that although the majority of commercial sex workers in Nairobi (the Pumwani cohort) became HIV-infected within a year, a sizable minority, subsequently estimated to be around 15% of the individuals tested, showed resistance to infection [7]; and that HIV-specific cytotoxic T lymphocytes (CTLs) could be isolated from healthy uninfected newborns of HIV-infected mothers [8]. The novel concept of ‘resistance’ to HIV infection in HIV-exposed individuals was proposed, and the search for immune correlates of such protection against HIV infection was initiated at that point. Subsequent pivotal reports showed that in HIV-exposed but uninfected individuals a particular genetic background, epitomized by the Δ32 deletion in the CCR5 receptor gene, can be detected [9], the production of soluble factors, including the CD8+ cell antiviral factor (CAF) and beta-chemokines, is increased [10–12], secretory HIV-specific IgA as well as T helper cells and CTLs can be observed in cervico-vaginal fluids and ejaculates [13,14], and natural killer (NK) cell activity is particularly potent [15]. Thus, 15 years after the first description of the detection of HIV-specific T helper cells in seronegative individuals, the ‘immunologic advantage’ possibly conferring resistance to HIV infection can be summarized as being correlated with the elicitation of systemic and mucosal cell-mediated immunity, presumably within favorable genetic and innate immunity settings. The suggested multiple components of this ‘immunologic advantage’ are summarized in Fig. 1 and will be discussed in detail within this review.Fig. 1: The ‘immunologic advantage’ tree. Multiple genetic characteristics are the roots of the peculiar exposed seronegative (ESN) phenotype. These properties result in the activation of multiple immune effector mechanisms that allow the ESN immune system to handle HIV as if HIV was a ‘normal’ virus and prevent the establishment of infection. This response prevents the initiation of a chronic infection, CD4+ cell depletion, and the development of AIDS. In this figure, CD4+ T cells and cytotoxic T lymphocytes are HIV-specific lymphocytes.Terminology Some methodological notes are needed: a consensus on how to define individuals with reduced susceptibility to HIV infection has not been reached. None of the definitions so far proposed is fully satisfactory. Thus, the definitions of ESN individuals and that of highly exposed persistently seronegatives overlook the possible presence of mucosal IgA in these individuals. The exposed uninfected and the multiple (or highly) exposed but uninfected definitions are based on the possibly erroneous assumption that these individuals have never undergone a subclinical and time-limited infection [6]. Noticeably, there is not the final proof that the immune and genetic correlates described herein confer an absolute resistance to HIV infection; rather, we believe that these correlates are associated with a robust down-modulation of the susceptibility to such infection. Lacking better definitions, the historic ESN acronym will be used in this review; we will define the immune and genetic correlates of this clinical state as associated with ‘reduced susceptibility’ to HIV infection. It should be noted, however, that the ESN individuals must not be confused with either long-term non-progressive (LTNP) patients, who have not developed AIDS for an extended period after HIV seroconversion, or with elite suppressors (or controllers), who show very low set-point viral load after acute infection. The latter groups are productively infected with HIV, whereas the HIV genome or proviruses are rarely, if ever, detectable in ESN individuals [2–5,7–13]. We will also concentrate on observations and studies conducted in humans. Notably, some of these observations are still controversial; the points that are still not fully clarified will be highlighted in the manuscript. Additionally, these studies are often based on small numbers of individuals, and reported results are sometimes almost anecdotic; we will summarize how many ESN individuals have been investigated to draw the main conclusions summarized herein. Finally, attempts to compare results obtained in different cohorts of ESN individuals have often limited success for a number of reasons, and in particular because a clear definition of who should be classified as lacking ESN. A clear definition of who should be classified as ESN is still required. The small number of ESN individuals analyzed notwithstanding, the overall observations done on ESN individuals have allowed the creation of a quilt whose design is getting more and more complex, but from which a recognizable pattern is slowly emerging. Immune correlates of reduced susceptibility to HIV infection: cell-mediated immunity HIV-specific CD4+ T cell responses were initially described in heavily HIV-exposed but seronegative (ESN) men enrolled from the MACS cohort (n = 5/5) [3]. These data were confirmed in other ESN groups including healthcare workers (n = 8/12) and healthy neonates born of HIV-infected mothers who did not receive antivirals (n = 8/23) [4,5]. HIV-specific CTLs in ESN individuals were first described some years later [8,16–19] in a dozen individuals. The lag was probably due to the fact that in the early 1990s, it was technically easier to measure CD4+ T cell than CTL responses. CD4+ T cells HIV-specific CD4+ T lymphocytes isolated from ESN individuals were found to produce interleukin (IL)-2 and to proliferate after stimulation with HIV peptides [20–22]. The HIV-specific T cells of ESN individuals were subsequently shown to produce low quantities of IL-10 in comparison with those from HIV-infected individuals [13]. HIV envelope (env)-specific CD4+ T cells of ESN individuals were also shown to generate high levels of CC chemokines, in particular RANTES and MIP-1β [11,23] (RANTES observed in 9/12 ESN individuals enrolled; MIP-1β observed in 24/25 ESN individuals studied), and to be capable of suppressing in vitro the replication of macrophage-tropic HIV strains [11]. Given the pivotal role of CC chemokines in modulating receptor binding and replication of HIV [24,25], these results are particularly important in connecting the ESN status and CD4+ T cell responses. Finally, higher levels of tumor necrosis factor (TNF)-α and TNF-β mRNA were detected in ESN individuals than in HIV-infected patients and healthy controls, both at the systemic level in peripheral blood lymphocytes and at the mucosal level in biopsies of the genital tract [26] (n = 9). No clear trends were identified regarding interferon (IFN)-γ production, as some groups reported that this cytokine is reduced [27–30] and other research indicated that IFN-γ is increased [12,31] in ESN individuals compared with HIV-infected individuals (Table 1). Notably, recent data suggested that HIV-specific T cell responses in ESN individuals could be dampened by an excess of regulatory T (Treg) cells. Consequently, the removal of CD4+CD25+ Treg revealed the presence of previously undetected strong HIV-specific T-cell responses in ESN newborns and neonates [35].Table 1: CD4+ T lymphocytes in exposed seronegative individuals: functional and phenotypic characterization.CD8+ T cells HIV-specific CTLs have been described in several different ESN cohorts and many authors claim that these cells make a fundamental contribution to modulating resistance to HIV infection. HIV-specific CTLs have been observed and characterized in the Pumwani Kenyan cohort of sex workers both at systemic [19,36] and mucosal levels [37]; in injecting drug users [38,39]; and in sexual partners of HIV-infected patients [14,22,26]. These studies have analyzed a total of approximately 100 ESN individuals; HIV-specific CTLs were observed in the majority (>70%) of such ESN (Table 2).Table 2: CD8+ T lymphocytes in exposed seronegative individuals: functional and phenotypic characterization.The detection of HIV-specific CTLs in ESN individuals raises important scientific questions: only the successful infection of host cells, that is, infection resulting in at least one complete cycle of viral replication, allows the effective presentation of viral peptides within a binary complex with a human leukocyte antigen (HLA) class I molecule [41]. The detection of HIV-specific CTLs in ESN individuals thus seems to indicate that HIV has managed to infect the host, but that its further propagation has been contained by immune mechanisms. The recent description of an alternative mechanism of processing and presentation by HLA class I molecules of exogenous antigens known as cross-priming could, nevertheless, explain the presence of CTLs in ESN individuals in the absence of actual infection. According to this mechanism, dendritic cells can process the virus and present it to CTL precursors in the absence of viral replication [42]. If this is the case, the presence of HIV-specific CTLs in ESN individuals would be the consequence, not necessarily of an infection, but of a different and presumably more efficient processing pathway of HIV antigens within dendritic cells. T-cell responses Comparison of HIV-specific T lymphocyte responses in ESN and HIV-infected individuals also suggests that the differences in susceptibility to HIV infection between these two groups resides in the quality rather than in the quantity of their immune responses. Thus, HIV-specific CD8+ CTLs of the ESN individuals recognize HIV epitopes that are different from those recognized by cells of HIV-infected patients [43]; rare conserved CD4+ T-cell epitopes within the HIV Env protein are immunodominant in ESN individuals, but are rarely recognized by HIV-infected patients [32]; and Gag-stimulated CD8+ T lymphocytes isolated from ESN individuals are characterized by higher levels of intracellular perforin and granzymes than those isolated from the HIV-infected partners [40] (n = 30 ESN individuals studied). Role of continual exposure The persistence of both HIV-specific CD4+ T cell and CTLs in ESN individuals seems to be strictly dependent on continuous virus exposure. In fact, various studies using different cohorts of ESN individuals that were followed longitudinally confirm that repeated exposure to HIV is necessary to maintain protective immunity. To summarize, HIV-specific CD4+ T cell and CTLs responses disappeared within 6–9 months after cessation of exposure to the virus in uninfected newborns of HIV-infected women [4] (n = 8/23) and in healthcare workers having reported a single professional exposure to HIV-infected body fluids [5] (n = 8/12); the concentration of HIV-specific IgA was significantly diminished in ESN women who underwent counseling and reported the adoption of safe-sex procedures [44] (n = 14/15); late seroconversion concomitant with the waning of HIV-specific CD8+ T cell responses occurred in two Kenyan HIV-resistant sex workers who interrupted commercial sex work for a period of time, probably due to reduced antigenic exposure [45]; and CD8+ cell non-cytotoxic responses (CNARs), observed in nearly half of 35 ESN individuals examined, were shown to decline in time after the last exposure to HIV [10]. Subsequent reports also showed that both HIV-specific CD4+ T cell and CTL responses are more frequent in ESN women with more recent sexual exposure [26,40,46]; the magnitude of CD4+ T cell responses correlates with the frequency rather than with the duration of virus exposure in sex workers [32]; and an inverse correlation is detected between exposure to virus and in-vitro susceptibility of peripheral blood mononuclear cells (PBMCs) to HIV infection [39]. These studies involved more than 100 ESN individuals recruited in different parts of the world. The observation that maintenance of possibly protective HIV-specific immunity in both the systemic and mucosal compartments might be contingent upon repeated antigen-specific immune stimulation suggests that exposure to HIV does not result in the generation of long-lasting memory cells. It will be important to analyze this issue in depth. Available data on naive and memory T lymphocyte subpopulations in ESN individuals show the presence of a low naive/memory cell ratio: an observation that resembles what has been seen in HIV infection [33,34]. Additionally, Gag-specific central memory CD4+ and CD8+ T cells, as well as terminally differentiated CD8+ T cells were augmented, whereas CD8+ effector memory cells were found to be reduced in ESN individuals compared with HIV-infected individuals (studies performed in 15 ESN individuals). The increase in terminally differentiated lymphocytes was suggested to play a role in determining the resistant phenotype [33]. Immune activation With few notable exceptions [47,48] (n = 45 and n = 20 ESN individuals analyzed, respectively), a consensus seems to emerge that indicates that ESN individuals are characterized by a generalized immune activation. This observation was made when CD4+/CD25+, CD8+/CD38+/CD45RO, and both HLADR-expressing CD4+ and CD8+ activated T lymphocytes, were analyzed in the peripheral blood of ESN individuals [26,34,49] and mucosally in commercial sex workers and their heterosexual partners [32] (a total of 93 individuals was investigated). It is also important to underline that recent results by Suy et al. [49] indicate that peripheral lymphocytes expressing CCR5 and CXCR4, the major HIV coreceptors, are upregulated on peripheral blood CD4+ T lymphocytes of sexually exposed ESN individuals (21 heterosexual couples were analyzed). Finally, some authors have also observed an increase of CD8+/CD28+ cells in ESN women [26]. This finding is very intriguing, given that these cells produce CAF, a soluble factor considered to be responsible for the non-cytotoxic antiviral response exerted by CD8+ cells via the inhibition of HIV RNA transcription [50,51]. The expression of CAF has been previously described to be present in nearly half of 35 studied ESN individuals [10]. The observation that generalized immune activation, including an increase of lymphocytes bearing the major HIV coreceptors, might be associated with reduced susceptibility to HIV infection is curious given that T-cell activation facilitates spreading of HIV infection [52–54]. A plausible explanation could be that immune activation is a favorable factor in inducing an immune resistance to primary HIV infection in ESN individuals, whereas, during the course of HIV propagation, it facilitates the replication of the virus and the consequent progression of the disease. Immune correlates of reduced susceptibility to HIV: humoral immunity ESN individuals also display some unconventional humoral immune responses that may play a role in HIV neutralization: a very effective protective mechanism against viral infections. Two kinds of HIV-related humoral immune responses have been described in ESN individuals so far: antibodies to cellular proteins involved in the HIV infection/entry process, and HIV-specific mucosal antibodies. The induction of anticell immune responses is rather common, and antilymphocyte antibodies have been observed early in sera from HIV-infected patients [55–57]. As the majority of these antibodies recognize HLA or CD4 molecules, their appearance had first been attributed to the cytopathic effect of HIV infection or to the long-lasting exposure to blood derivatives, as in the case of haemophiliacs [58,59]. In this regard, HIV-blocking IgG directed against HLA class I and CD4 molecules have been found in sera of ESN individuals [60–63] (n = 14/31). Anti-CD4 antibodies of ESN individuals recognize epitopes exposed by gp120 binding, suggesting that such antibody response is generated after repeated, long-lasting exposure to HIV in either horizontal or vertical transmission [61,63]. Although anti-CD4 antibodies were also found in some HIV-seropositive individuals and in some healthy blood donors, such anti-CD4 antibodies recognized epitopes different from those seen by antibodies found in sera of ESN individuals [61,62,64,65]. Another cellular protein, CCR5, is also targeted in ESN individuals by IgA and IgG antibodies at the mucosal and systemic levels. These antibodies are directed toward a conformational epitope corresponding to the second extracellular loop of CCR5 (YAAAQWDFGNTMCQ), which is not involved in HIV binding (6/48 ESN individuals analyzed). Thus, anti-CCR5 antibodies are likely to exert their possibly protective role through the downregulation of the CCR5 protein [66]. The effect of anti-CCR5 antibodies could be due to the recognition and the interaction with specific epitopes or, alternatively, to steric hindrance. Recent data showing that the mechanism of action of these antibodies is mediated by the internalization of the receptor through a clathrin-dependent pathway [67] seem to suggest that CCR5-specific antibodies mediate their effect secondarily to epitope-specific recognition. Anti-CCR5 antibodies do not affect physiologic immune functions, due to the redundancy in chemokine receptor family, but could possibly play an initial role in protection [68]. Notably, both anti-CD4 and anti-CCR5 antibodies have been described as specific markers of HIV-exposure in Asian and Caucasian but not in African ESN individuals [69] (anti-CCR5 and anti-CD4 antibodies were detected in 10.7 and 5.5% of the 149 enrolled ESN individuals, respectively). This discrepancy could be due to differences in the genetic background, in the route of exposure, or in the different environmental conditions, which can modulate immune responses to microbes [52,70–72]. The generation of anti-CCR5 antibodies has been attributed to several mechanisms: Ditzel et al.[73] showed that CCR5 can act as an alloantigen in CCR5Δ32 homozygous individuals. Anti-CCR5 antibodies directed toward epitopes different from those seen by antibodies of ESN individuals have been also observed in healthy individuals not previously exposed to HIV [74–76]. This finding could be explained by autoimmune phenomena triggered by membrane perturbations unrelated to HIV stimuli, such as exogenous or endogenous viruses or local inflammation. Alternatively, ESN individuals could have undergone priming with other (possibly cross-reactive) viruses or proteins, and once exposed to HIV, they could possibly mount a secondary response, directed toward allo-antigen and self-antigens associated with viral particles [77]. Allo-immune and auto-immune responses have been found in HIV-infected patients [58,61,78,79] and such potentially autoimmunity-associated immune responses could play a protective role in preventing HIV infection [75,80]. For instance, the HIV-neutralizing human monoclonal antibodies 2F5 and 4E10 are produced in natural infection, recognize conserved gp41 epitopes [80–82], and can bind membrane phospholipids with kinetics comparable with those of anticardiolipin immunoglobulins generated in autoimmune syndromes [81]. The ability of these antibodies to bind cardiolipin has nevertheless recently been questioned by two independent groups [83,84]. These authors did not confirm the cardiolipin-binding properties of 2F5 and 4E10 and showed that, even if 4E10 can bind phospholipids, such binding has a much lower affinity compared with the one for gp41. These discrepancies seem to reflect technical difficulties that could be explained by a low affinity of the 2F5/4E10–phospholipid interaction or, alternatively, by the fact that the anionic lipids could be only a small portion of the antigen-binding paratope. IgA antibodies are the most abundant isotypes found in mucosal secretions and epithelia and take part in several effector pathways that may protect the host from mucosal infection and clear the virus [85]. Soluble antibodies can compete with HIV for attachment to epithelial cells [86], participate in opsonization, activate complement-mediated cell lysis, induce antibody-dependent cell-mediated cytotoxicity and As HIV is by sexual and the genital is the main initial it is that HIV-specific antibodies could be detected in the of African ESN individuals were detected in of the ESN individuals enrolled in the In this regard, IgA to the conformational epitope of CCR5 in ESN individuals were to of HIV a epithelial cell but monoclonal antibodies against other of CCR5 had effect on This finding likely a different of CCR5 at the mucosal level immune protection (or in ESN individuals has also correlated with HIV-specific mucosal IgA antibodies [13]. These IgA have been observed in secretions of ESN individuals from cohorts with different genetic such as heterosexual women from (n = IgA detected in the of and in the secretions of of the enrolled ESN and (n = IgA detected in of as well as in sex workers from studies enrolled a total of ESN individuals; IgA were observed in and of the two respectively), (n = IgA detected in of (n = IgA detected in of and (n = IgA detected in and of individuals using two different These antibodies were also detected in the of ESN partners of HIV-infected women (n = IgA detected in of individuals). In low levels of HIV-specific IgG have also been found in sexual partners of HIV-seropositive individuals with Finally, recent results of a of HIV-infected mothers indicate that HIV-specific IgA can be observed in such in HIV-exposed in IgA were detected uninfected after 1 of The detection of IgA in mucosal secretions of ESN individuals in the absence of detectable HIV-specific IgG in their sera might seem from the of antibody strong antiviral IgA responses with have been observed in the absence of and IgG in Thus, et a CD4+ T antiviral IgA response that is to cells and is generated in the absence of cell via class or This phenomenon was observed in and in which a of IgG and in the presence of strong IgA could be Recent performed have the different for the induction of IgA and IgG antibody responses. In this latter et showed that to functional mucosal IgA responses against T-cell dependent antigens does not through and can be independent of and in IgA to IgG are dependent on and Thus, they suggested a different pathway response for the different for mucosal IgA responses are different from those for systemic antibody responses and In ESN individuals (n = resistance to HIV infection has been associated with HIV-neutralizing antibodies directed to a small within the of the portion of gp41 This with the in gp120 a conserved It is that the of gp41 is not recognized by IgA in HIV-infected individuals. In other ESN HIV-neutralizing IgA recognize a immunodominant of gp41 which is recognized by HIV-seropositive individuals. more than one HIV can modulate the susceptibility to HIV-specific IgA of ESN individuals have been shown to prevent HIV CD4+ T cells in in-vitro infection of and to HIV in an in-vitro using cell IgA from ESN individuals were also shown to primary HIV from different and ESN individuals were analyzed Finally, recent results showed that the and concentration of a chemokine that cells in the epithelial is increased in the majority of of ESN individuals. A correlation between and of HIV-infected and neonates was detected in a cohort Notably, of with a in a increase of cells in the (Table immunity in exposed seronegative individuals: functional and phenotypic important regarding these results is that mucosal IgA antibodies have been observed in some but not groups of ESN individuals. Thus, studies performed in some cohorts of African and ESN individuals did not show mucosal HIV-specific IgA et (n = whereas results in ESN were reported in cohort discrepancy could be due to technical difficulties associated with detection of IgA in mucosal which can have very low of such antibodies In although secretory IgA are only to mucosal levels can due to different and to the of when in multiple ESN cohorts with different were analyzed in the using the and the of HIV-specific were detected in some but not such cohorts et in These results suggest that may be involved in induction detection of HIV-specific mucosal immune responses. Finally, although mucosal HIV-specific antibodies might play a role in preventing horizontal HIV transmission at the mucosal it seems that these antibodies do not play role in the of infection through immunity and the of susceptibility to HIV infection The role of innate immunity in the of susceptibility to HIV infection is and few reports a possible role for this of the immune The activation of cells in the early of and the possible role of these cells in the of acute viral replication has been in a of infection A inverse correlation between the CC of host cells and levels of in HIV infection has also been reported an increase in cell activity has been observed in HIV-exposed but uninfected injecting drug suggesting a possible protective role of cells [15]. It has also been shown that low numbers of cells are associated with progression to AIDS in HIV-infected individuals These results were confirmed by et who also that, although killer cell as a are on cells of ESN individuals, HLA molecules that bind are in these individuals (n = a high associated with levels and an increased were detected in ESN individuals (n = The increased expression of the receptor was recently confirmed (n = these observations seem to the that cell activity is in ESN individuals Finally, results of a very recent that analyzed