From the aChair of Immunology DISP LITA Vialba, University of Milano, Milano, Italy bInfectious Diseases Unit, Ospedale Santa Maria Annunziata, Firenze, Italy cLaboratory of Biology, Don C. Gnocchi Foundation, IRCCS, 20148, Milano, Italy.

AUTHOR DISCUSSION: These study data suggest that a high frequency of perforin-enriched cells could be another correlate of possible protection against HIV infection in a list that includes: (i) CD4 and type 1 cytokine- and beta chemokine-secreting T helper cells [21-25,49]; (ii) CD8 and MCH class I-restricted cytotoxic T lymphocytes [26-31,50]; (iii) HIV-1 specific mucosal IgA antibodies [46,51-53]; (iv) cell secreted antiviral factors including CAF [54,55]; and (v) increased natural killer cell activity [56]. It seems reasonable to suggest that these correlates should be reproduced in the design of preventative vaccines.

Repeated exposure to HIV does not necessarily result in seroconversion and overt infection, and multiple cohorts of HIV exposed but seronegative individuals have been described and characterized over time [1,2]. A remarkably elevated frequency of HIV-specific T helper and cytotoxic T lymphocytes has been observed in ESN [22-33], thus exposure to HIV occurs in these individuals and this exposure is massive enough to induce the presence of antigen-specific memory lymphocytes in the peripheral blood. An alteration of the CD4: CD8 ratio, mostly secondary to an augmented number of CD8 T lymphocytes, as well as a decrease in CD4 naive cells and an increase in CD4 memory lymphocytes, resulting in modifications of the naive/memory ratio, has been described as well in ESN [34,35]. These findings suggest that the same immunologic stimuli/signals that lead to prolonged immune activation in HIV-seropositive patients also occur in ESN.

In the light of recent data obtained in HIV-infected patients, and showing that chronic antigenic exposure can alter the memory/naive subsets of antigen-specific T lymphocytes [9-13], we investigated the possibility that, because ESN are repeatedly exposed to HIV antigens, T-lymphocyte subsets could be altered in these individuals. Results showed that HIV-specific memory and naive subpopulations are indeed skewed in ESN, in whom an increase of central memory lymphocytes, and of TD CD8 T lymphocytes, is observed. These data confirm that exposure to HIV occurs in high-risk seronegative individuals; the observation that naive and CM are decreased in ESN indicate that this exposure is robust enough to skew the EM/CM ratio. Finally, the significant increase in late effector CD8 T lymphocytes seen in ESN suggests a role for these cells in preventing actual infection in the face of continuous exposure to HIV

It is of interest to notice that the prevalent gag-specific CD4 lymphocyte subpopulation was different in ESN compared to HIV patients. Thus, whereas gag-specific IL-2 producing cells were predominant in ESN, gag-specific IFNγsecreting lymphocytes were mostly observed in HIV patients. Control over viraemia in long-term non progressors (LTNP), in HIV-infected, antiviral therapy-naive individuals with low viraemia, and in aviraemic individuals treated with HAART during primary infection, was shown to be associated with the detection of gag-specific CD4 lymphocytes that produce IL-2 [9-13]. These cells are long-lived and have the capacity to persist as long-term memory cells. On the contrary, HIV-specific CD4-specific, IFN-γ-producing cells are characterized by a limited proliferative ability and are prevalent in HAART-untreated patients with high viral load, as well as in patients in whom therapy is unable to control HIV replication [36-43]. The observation that HIV-exposure in ESN might favor the generation of IL-2 rather than IFNγ-producing, gag-specific cells suggests that in this setting viral exposure elicits the same lymphocyte population that is associated with lack of progression in LTNP.

Thymus-generated naive T lymphocytes traffic through lymphoid organs until they encounter the specific antigen and become memory cells [44-47]. Memory cells can be divided into major subsets called CM and EM lymphocytes: these cells are characterized by different clusters of molecules, by distinct effector functions, and by peculiar migratory capacities. Thus, CM cells express CCR7 and CD62L [3-5], and predominantly produce IL-2. The physiological lineage relationship proceeds, in the presence of antigen, from naive to EM through the CM intermediate [3-6]. EM had initially been considered to be the most potent effector cells [3,4], nevertheless recent data suggested that EM revert into CM in the absence of antigen [7,8]. If this is the case, it makes sense that CM lymphocytes are more abundant in ESN compared to HIV-infected individuals. Thus, in ESN fewer CM would differentiate into EM and/or more EM would revert to CM because antigen stimulation is not continuous but rather discrete, and a lower antigenic load is expected in ESN as compared to what is observed in HIV patients.

Repeated observations have demonstrated that the magnitude and the breadth of HIV-specific immune response wane in ESN within a year of the last exposure. Thus: (i) HIV-specific T helper and CTL responses disappear within months after cessation of exposure to the virus in uninfected newborns of HIV-seropositive women [23] and in health care workers reporting a single professional exposure to HIV-seropositive body fluids [24]; and (ii) the concentration of HIV-specific IgA is significantly diminished in ESN women who adopt safe sex procedures [48]. All the ESN enrolled herewith reported having been exposed to HIV within 4 months prior to the study period; analyses in which we attempted to correlate memory and naive T subsets to the last at-risk sexual episode did not yield definitive results. It will be of interest to repeat these studies in bigger groups of ESN.

Analysis of both gag-specific lymphocytes showed that TD CD8 T lymphocytes were augmented in ESN. In particular, the CM/TD and the EM/TD ratios for gag-specific CD8 T cells were reduced in ESN compared to HIV patients; in both cases this was driven by the higher percentages of TD cells observed in ESN. CTL kill virus-infected targets with an array of mechanisms; it was recently demonstrated that the lythic, perforin-dependent pathway is the predominant defensive modality exerted by cytotoxic T lymphocytes against HIV [18,19]. HIV-infection is nevertheless characterized by the accumulation of pre-TD CTL that contain and secrete low amounts of perforin [14-16]; this phenomenon has been linked to the lack of control of the immune response on HIV replication in infected patients [16,17]. The observation presented herewith that TD CD8 T lymphocytes are expanded in ESN is interesting because these cells are known to be perforin enriched. On one hand this result could simply be a marker of immunologic memory to repeated viral exposures. On the other hand, it is tempting to speculate that the high number of TD, perforin-enriched CD8 T lymphocytes seen in ESN could be at least partially responsible for the prevention of HIV infection that may be present in these individuals. The observation that none of these individuals has seroconverted over a 7 years (1997-2004) observation period might favour this possibility.

Despite multiple and repeated exposures to HIV-1 some individuals never become infected (reviewed in [1,2]). It has been observed that viral exposure is associated in exposed seronegative (ESN) individuals with the activation of cell mediated immunity, thus, a high frequency of HIV-specific memory T lymphocytes is detected in these subjects (reviewed in [1,2]). Memory and naive T cells can be differentiated based on the expression of a number of surface markers, including the chemokine receptor CCR7 and the RO and RA isoforms of CD45, into different subsets that include central memory (CM), effector memory (EM), and terminally differentiated (TD) lymphocytes: CM cells express molecules, such as CCR7 and CD62L, that are lost upon differentiation into EM cells [3-5]. Recent data suggest that the physiological lineage relationship between these subsets proceeds from naive to EM through the CM intermediate [3-6]; CD8 TD lymphocytes are effector cells highly enriched in perforin that are specialized in the lysis of viral infected targets. The relative proportion of these lymphocyte susbets was convincingly shown to be driven by antigenic exposure in the presence of cytokines such as interleukin (IL)-7 and IL-15 [5-8].

The proportion of memory and naive T lymphocyte subsets is skewed in individuals who are chronically exposed to high concentrations of antigen. In particular, low IL-2/high interferon (IFN)γ-producing CD4 T lymphocytes with a minimal proliferative ability, as well as pre-TD CD8 EM T lymphocytes, tend to accumulate in HIV infected individuals [9-17]. Thus, HIV-specific 45RA-/CCR7- and 27+/28- CD8+ T cells are abundant in chronic HIV infection. Because: (i) these cells have a lower content of perforin compared to TD lymphocytes; and (ii) recent studies showed that the perforin-mediated pathway is the predominant anti-HIV mechanism used by CD8 T lymphocytes [18,19], it was suggested that the accumulation of such cells could be associated with the lack of control of the immune system on viral replication that characterizes HIV infection.

Fifteen HIV ESN individuals were enrolled in the study. In each case the ESN was the sexual partner of an HIV infected patient; in each couple a prolonged history of penetrative sexual intercourse without condom (and no other known risk factors) was reported. Inclusion criteria for the ESN was a history of multiple unprotected sexual episodes for at least 3 years with at least an episode of at-risk intercourse within 4 months prior to the study period. Self-administered questionnaires show that the couples reported an average of 25 unprotected sexual episodes/year (range 12 to > 50) in the 3 years previous to the study; vaginal intercourse was the rule and anal sex was not reported by any of the participants in the study. The serostatus of the ESN, analysed by ELISA and Western blot techniques at regular intervals, has always been negative. The demographic characterization of the ESN and of the HIV infected partners is shown in Table 1. HIV plasma viraemia was undetectable in 8/15 HIV-infected partners; all but one of the HIV partners were receiving highly active antiretroviral therapy (HAART) at the time of the study (see Table 1). It is of note that because in all cases the diagnosis of HIV infection was done during the chronic phase of disease, unprotected penetrative sexual intercourse was initiated long before HAART. Fourteen HIV-infected individuals and 15 healthy controls (HC) were also enrolled in the study. HIV patients and HC were age- and-sex-matched with the ESN. Nine of the 14 HIV patients had undergone antiretroviral therapy for at least 1 year at the time of the study; CD4 cell counts were in the 250 ×106-750 ×106/l range and HIV plasma viremia was < 50 copies/ml in 9/14 cases. All ESN, HIV-seropositive, and HC had been longitudinally followed for at least 3 years (prior to the study period) by the Department of Infectious Diseases, Santa Maria Annunziata Hospital in Florence. This allowed us to exclude from the study ESN and HC in whom sexually transmitted diseases or any other pathology had been reported in that time period. The ESN were characterized on the basis of the presence of CCR5-Δ32 alleles; a heterozygous deletion was detected in one individual. All ESN, HIV patients and low-risk uninfected individuals agreed to donate peripheral blood mononuclear cells (PBMC). The Research Ethics Committee of the Santa Maria Annunziata Hospital, Florence, Italy, approved this protocol. Written informed consent was obtained from all patients before enrolment.

Memory and naive T lymphocyte subsets have not yet been analysed in ESN: individuals in whom exposure to HIV is repeated over time, but seroconversion does not occur. Preliminary results have shown that the CD4/CD8 and the naive/memory ratios (45RA/RO) ratio are altered in these individuals [20]. We verified whether repeated antigenic exposure in the absence of disease would influence the maturative patterns of HIV-specific subsets of T lymphocytes; results showed that this is indeed the case.

Background: Repeated exposure to HIV is not always associated with infection and multiple cohorts of HIV-exposed but seronegative individuals (ESN) have been described. HIV-specific CD4 and CD8 T lymphocytes are detected both in HIV patients and in ESN; we verified whether different patterns of HIV-specific memory T lymphocytes would be detected in individuals in whom exposure to HIV results or does not result in infection.

Methods: Gag-specific T cells were analysed in 15 ESN, 14 HIV patients, and 15 healthy controls using extensive flow cytometry analysis.

Results: Data confirmed that gag-specific T lymphocytes are present in ESN. Gag-specific T cells mainly secrete interleukin-2 in ESN and interferon-γin HIV patients. In addition the CD4/CD8 and the memory/naive ratios are altered, central memory (45RA-/CCR7+) CD4 and CD8 T lymphocytes are more abundant, and terminally differentiated (45RA+/CCR7- and 27-/28-) CD8 T lymphocytes are augmented in ESN individuals.

Conclusions: Exposure to HIV occurs in high risk seronegative individuals; the observation that naive cells and CM are skewed in ESN indicate that this exposure is robust enough to modulate the CM/EM ratio. The increase in late effectors and in natural killer cells seen in ESN suggests a role for these cells in preventing actual infection.

The frequency of gag (p17+p24)-specific, IL-2 and IFNγ-secreting, CD4 and CD8 T lymphocytes was evaluated in the peripheral blood of 15 ESN, 14 HIV patients, and 15 HC who referred never having been exposed to HIV. Results showed that these populations of lymphocytes are present both in ESN and HIV patients but, as expected, are practically undetectable in HC. CD4, gag-specific, IL-2-secreting lymphocytes were more present in ESN whereas CD4, gag-specific IFNγ-secreting cells seemed to be more expressed in HIV patients; finally, CD8, gag-specific, IFNγ-secreting lymphocytes were the least frequently detected cells both in ESN and HIV patients. These differences were confirmed when antigen-specific cytokine-producing T cells were normalized based on the percentage of memory CD4 or CD8 T cells (data not shown). Representative dot plots are shown in the upper part of Fig. 1; total results are presented in the lower part of the same figure.

Absolute CD4 counts were similar in HC and ESN whereas the percentage of CD4 T lymphocytes was reduced in ESN compared to HC, with the lowest values observed in HIV patients (ESN versus HC, P = 0.045; HIV versus HC, P = 0.01; ESN versus HIV, P = 0.038). Absolute CD8 counts were increased both in ESN and HIV patients compared to HC, the same trend was observed when the percentage of CD8 was measured. As a consequence, the CD4/CD8 ratio was reduced in ESN (median = 1.3) compared to HC (median = 1.9), with the lowest value observed, as expected, in HIV patients (median = 0.59).

We further investigated memory and naive lymphocyte subsets in ESN and HIV patients by stimulating in vitro lymphocytes with gag proteins (p17 + p24) and then analyzing cells based on the expression of the chemokine receptor CCR7. Thymic-derived naive CD4/RA/CCR7 cells lose the expression of RA once they encounter antigens in the periphery. The maintenance or lack thereof of the CCR7 marker divides memory cells into central memory (CCR7+) and effector memory (CCR7-) lymphocytes. Results presented in Fig. 3 showed that central memory CD4 T cells (CD4+RA-CCR7+) were augmented (P = 0. 038) whereas effector memory CD4 T cells (4+ RA-CCR7-) were diminished in ESN compared to HIV patients (P = 0.016).

Similarly to what was observed in antigen-specific populations of CD4 T cells, central memory (CD45RA-/CCR7+) CD8 T lymphocytes were augmented whereas effector memory (CD45RA-/CCR7-) cells were diminished in ESN compared to HIV patients (P = 0.004). Finally, TD, late effector lymphocytes (CD45RA+/CCR7-) were increased in ESN compared to HIV patients (P = 0. 009). This result was confirmed by the analysis of CD8+/CD27+/CD28- (pre-TD) and of CD8+/27-/28- (effector cells) lymphocyte subpopulations. Thus, CD8+/CD27+/CD28- were higher in HIV patients whereas CD8+/CD27-/CD28- lymphocytes were increased (P = 0.012) in ESN. These data are shown in Fig. 4 and in Table 2.

In ESN both gag-specific IL2-secreting and IFNγ-secreting CD4 T cells mostly displayed a central memory (CCR7+/CD45RA-) maturation phenotype. In contrast, the majority of gag-specific, cytokines-producing CD4 T lymphocytes cells were effector memory (CCR7-/CD45RA-) in HIV-infected patients. Similar results were obtained analysing gag-specific, cytokine-secreting memory CD8 T lymphocyte subsets (data not shown).

Ratios between the various subsets of memory and naive CD4 and CD8 T lymphocytes were evaluated for all ESN and HIV patients. The CM/EM ratios for gag-specific CD4 and CD8 T lymphocytes were higher in ESN compared to HIV patients (CD4, P = 0.041); these results likely reflect the lower antigenic load present in ESN. In contrast, both the EM/TD and the CM/TD ratios for gag-specific CD8 T lymphocytes were lower in ESN compared to HIV patients (EM/TD; P = 0.029). These results are summarized in Fig. 5.