Evaluation of the interlaboratory concordance in quantification of human immunodeficiency virus-specific T cells with a gamma interferon enzyme-linked immunospot assay.

The gamma interferon (IFN-gamma) enzyme-linked immunospot (ELISPOT) assay is a reference method for the ex vivo monitoring of antigen-specific T cells and a primary tool for assessing clinical trials of human immunodeficiency virus (HIV) or cancer vaccines. Four experienced laboratories in Paris compared their results with this method by exchanging frozen blood samples from eight HIV-seronegative and eight HIV-seropositive subjects. Each laboratory measured the IFN-gamma-producing cells specific for HIV, Epstein-Barr virus, cytomegalovirus, and influenza using the same set of peptides and the same ELISPOT reader but its own ELISPOT technique. The cutoff values for positive responses (50 or 100 spot-forming cells/10(6) peripheral blood mononuclear cells over background) were consistent with the binomial statistic criterion. The global qualitative concordance, as assessed by the kappa index, ranged from 0.38 to 0.92, that is, moderate to excellent, and was better for non-HIV 9-mer peptide pools than for HIV 15-mer peptide pools. The interlaboratory coefficient of variation for the frequency of virus-specific T cells was 18.7% (data are expressed on a log scale). Clustering analysis of HIV-positive subjects showed qualitative agreement for ELISPOT results from all four laboratories. Overall, the good interlaboratory qualitative concordance of IFN-gamma ELISPOT assays with only the peptide source and ELISPOT reader in common suggests that a qualitative comparison of interlaboratory findings is feasible. Nonetheless, a single set of standard operating procedures should be used in multicenter trials to improve standardization.

Weak anti-HIV CD8(+) T-cell effector activity in HIV primary infection.

HIV-specific CD8(+) T cells play a major role in the control of virus during HIV primary infection (PI) but do not completely prevent viral replication. We used IFN-gamma enzyme-linked immunospot assay and intracellular staining to characterize the ex vivo CD8(+) T-cell responses to a large variety of HIV epitopic peptides in 24 subjects with early HIV PI. We observed HIV-specific responses in 71% of subjects. Gag and Nef peptides were more frequently recognized than Env and Pol peptides. The number of peptides recognized was low (median 2, range 0-6). In contrast, a much broader response was observed in 30 asymptomatic subjects with chronic infection: all were responders with a median of 5 peptides recognized (range 1-13). The frequency of HIV-specific CD8(+) T cells among PBMC for a given peptide was of the same order of magnitude in both groups. The proportion of HIV-specific CD8(+)CD28(-) terminally differentiated T cells was much lower in PI than at the chronic stage of infection. The weakness of the immune response during HIV PI could partially account for the failure to control HIV. These findings have potential importance for defining immunotherapeutic strategies and establishing the goals for effective vaccination.

Broad, intense anti-human immunodeficiency virus (HIV) ex vivo CD8(+) responses in HIV type 1-infected patients: comparison with anti-Epstein-Barr virus responses and changes during antiretroviral therapy.

The ex vivo antiviral CD8(+) repertoires of 34 human immunodeficiency virus (HIV)-seropositive patients with various CD4(+) T-cell counts and virus loads were analyzed by gamma interferon enzyme-linked immunospot assay, using peptides derived from HIV type 1 and Epstein-Barr virus (EBV). Most patients recognized many HIV peptides, with markedly high frequencies, in association with all the HLA class I molecules tested. We found no correlation between the intensity of anti-HIV CD8(+) responses and the CD4(+) counts or virus load. In contrast, the polyclonality of anti-HIV CD8(+) responses was positively correlated with the CD4(+) counts. The anti-EBV responses were significantly less intense than the anti-HIV responses and were positively correlated with the CD4(+) counts. Longitudinal follow-up of several patients revealed the remarkable stability of the anti-HIV and anti-EBV CD8(+) responses in two patients with stable CD4(+) counts, while both antiviral responses decreased in two patients with obvious progression toward disease. Last, highly active antiretroviral therapy induced marked decreases in the number of anti-HIV CD8(+) T cells, while the anti-EBV responses increased. These findings emphasize the magnitude of the ex vivo HIV-specific CD8(+) responses at all stages of HIV infection and suggest that the CD8(+) hyperlymphocytosis commonly observed in HIV infection is driven mainly by virus replication, through intense, continuous activation of HIV-specific CD8(+) T cells until ultimate progression toward disease. Nevertheless, highly polyclonal anti-HIV CD8(+) responses may be associated with a better clinical status. Our data also suggest that a decrease of anti-EBV CD8(+) responses may occur with depletion of CD4(+) T cells, but this could be restored by highly active antiretroviral treatment.

Type 1 CD4(+) T-cell help is required for induction of antipeptide multispecific cytotoxic T lymphocytes by a lipopeptidic vaccine in rhesus macaques.

We have optimized the induction of antiviral cytotoxic T lymphocytes (CTL) in rhesus macaques by a lipopeptide vaccine containing seven peptides from simian immunodeficiency virus (SIV) Nef and Gag proteins and a strong T-helper peptide from tetanus toxoid (TT) that is promiscuous in humans (peptide TT 830-846). Two of the eight immunized macaques showed T-helper (Th) cell proliferation and a specific synthesis of gamma interferon in response to TT 830-846 peptide. They also showed multispecific cytotoxic activity against three to five of the immunizing SIV peptides. These results show the importance of a strong specific type 1 Th response for inducing a multispecific CTL response in vivo, which is essential for the development of an anti-human immunodeficiency virus vaccine.

Altered ex vivo balance between CD28+ and CD28- cells within HIV-specific CD8+ T cells of HIV-seropositive patients.

The CD8+CD28- cell population in the blood of HIV-infected individuals is considerably expanded. Yet the cause of this expansion is not clear. The recent demonstration of identical TCR-rearranged genes in CD8+CD28+ and CD8+CD28- expanded T cells of HIV-seropositive patients supports the hypothesis that these two subpopulations are phenotypic variants of the same lineage. To further elucidate the precise relationship between them, we measured the fraction of CD28+ and CD28- T cell subsets in IFN-gamma-producing CD8+ T cells by intracellular staining and cytofluorometry as a functional test for ex vivo recognition of epitopes derived from HIV-1, Epstein-Barr virus (EBV) and influenza virus. HIV-specific CD8+ T cells were predominantly CD28 in all the eight HIV-seropositive subjects tested. In contrast, the anti-EBV and anti-influenza CD8+ T cells were mainly CD28+ in these patients as well as in HIV-seronegative individuals. This supports the notion that the CD8 CD28 hyperlymphocytosis observed in HIV infection is due mainly to chronic activation and differentiation of HIV-specific memory CD8+CD28+ T cells into terminally differentiated CD8+CD28-lymphocytes, because of intense HIV-1 replication and without any important bystander activation. This clarification of the mechanisms underlying the CD8+CD28- expansion in HIV-induced pathogenesis may have important therapeutic implications.

Administration of interleukin 13 to simian immunodeficiency virus-infected macaques: induction of intestinal epithelial atrophy.

Increase Th2 cytokine production may contribute to some clinical manifestations of HIV infection, and studies have suggested that IL-13 rather than IL-4 is involved in these conditions. We directly tested this hypothesis by administrating IL-13 to SIV-infected macaques. SIV-infected rhesus macaques received a daily subcutaneous injection for 21 days of either IL-13 (10 microg/kg/day) or a placebo. The four macaques treated with IL-13 experienced body weight loss (9.95 +/- 0.71%) related to intestinal tract damage: they all suffered from a complete atrophy of duodenal villi. This was presumably due to premature epithelial cell death: proliferating Ki67+ cells in glandular crypts were as numerous as in control animals, but many epithelial cells developed apoptosis. The duodenal mucosa was infiltrated with cells expressing CD56 and PEN5, two markers of NK cells, and there was a deregulation of local cytokine and chemokine production characterized by a decrease in IL-10 gene expression (25% of controls) and an increase in gene expression for IFN-gamma (4-fold control), MIP-1alpha (8-fold control), and MIP-1beta (13-fold control). Thus, IL-13 can induce digestive epithelial cell injury in vivo in primates infected with a retrovirus. Therefore, its role should be considered in digestive manifestations of HIV infection as well as in other disorders associated with intestinal epithelial atrophy.

Study of the immunogenicity of different recombinant Mengo viruses expressing HIV1 and SIV epitopes.

Recombinant Mengo viruses expressing heterologous genes have proven to be safe and immunogenic in both mice and primates, and to be able to induce both humoral and cellular immune responses (Altmeyer et al., 1995, 1996). Several recombinant Mengo viruses expressing either a large region (aa 65-206) of the HIV1 nef gene product, or cytotoxic T lymphocyte (CTL) epitopic regions from the SIV Gag (aa 182-190), Nef (aa 155-178) and Pol (aa 587-601) gene products were engineered. The heterologous antigens were expressed either as fusion proteins with the Mengo virus leader (L) protein, or in cleaved form through autocatalytic cleavage by the foot-and-mouth disease virus 2A protein. Rhesus macaques and BALB/c mice inoculated with the Mengo virus SIV recombinants failed to develop CTL responses against the SIV gene products, while one of the HIV-Nef recombinants induced a weak CTL response in mice directed to an HIV1 Nef peptide spanning positions 182-198. In contrast, BALB/c mice immunized with vaccinia virus recombinants expressing HIV1 Nef developed a strong CTL response to the 182-198 peptide and also responded to a second peptide spanning positions 73-81. These results indicate that Mengo virus recombinants expressing HIV1 Nef and SIV CTL epitopes are weak immunogens. One of the fusion recombinants expressing SIV CTL epitopes failed to infect macaques even when used at high doses, while the recombinant expressing HIV1 Nef as a fusion protein failed to infect BALB/c mice. These results demonstrate that the expression of certain heterologous sequences as fusion proteins with L can result in the loss of the ability of the recombinant to infect normally susceptible animals.

Phase I trial of recombinant adenovirus gene transfer in lung cancer. Longitudinal study of the immune responses to transgene and viral products.

Animal studies indicate that the use of replication-deficient adenovirus for human gene therapy is limited by host antivector immune responses that result in transient recombinant protein expression and blocking of gene transfer when rechallenged. Therefore, we have examined immune responses to an adenoviral vector and to the beta-galactosidase protein in four patients with lung cancer given a single intratumor injection of 10(9) plaque-forming units of recombinant adenovirus. The beta-galactosidase protein was expressed in day-8 tumor biopsies from all patients at variable levels. Recombinant virus DNA was detected by PCR in day-30 and day-60 tumor biopsies from all patients except patient 1. A high level of neutralizing antiadenovirus antibodies was detected in patient 1 before Ad-beta-gal injection whereas it was low (patient 3) or undetectable in the other two patients. All patients developed potent CD4 type 1 helper T cell (Th1) responses to adenoviral particles which increased gradually over time after injection. Antiadenovirus cytotoxic T lymphocyte responses were consistently boosted in the two patients examined (patients 3 and 4). Sustained production of anti-beta-galactosidase IgG was observed in all patients except patient 1. Consistent with anti-beta-gal antibody production, all patients except patient 1 developed intense, dose-dependent Th1 responses to soluble beta-galactosidase which increased over time. Strong beta-galactosidase-specific cytotoxic T lymphocyte responses were detected in patients 2, 3, and 4. Our results clearly show that despite the intensity of antiadenovirus responses, transgene protein expression was sufficient to induce strong and prolonged immunity in three patients. Recombinant adenovirus injected directly into the tumor is a highly efficient vector for immunizing patients against the transgene protein.

Direct ex vivo simian immunodeficiency virus (SIV)-specific cytotoxic activity detected from small intestine intraepithelial lymphocytes of SIV-infected macaques at an advanced stage of infection.

Human immunodeficiency virus (HIV) induces a profound disorganization of the lymphoid tissues with marked abnormalities of the immune system at the terminal stage of infection. Since the digestive mucosal immune system is by far the largest lymphoid organ of the body, we attempted to evaluate its functional activity in advanced stages of simian immunodeficiency virus (SIV) infection in the SIV-macaque model of HIV infection. Two chronically intravenously SIV-infected macaques, including one at the AIDS stage, were studied. Intestinal intraepithelial lymphocytes (IEL) were isolated, analyzed, and compared to lymphocytes obtained from blood, spleen, and different lymph nodes: IEL were predominantly CD8+ T lymphocytes expressing the alphaE beta7 integrin and lacking the CD28 coactivatory molecule. A direct ex vivo SIV-specific cytotoxic activity was prominently found in the IEL of both macaques and was weaker or absent in the other sites. To our knowledge, this is the first report of SIV-specific cytotoxic activity from small intestine IEL in SIV-infected macaques. Considering the high similitude of the SIV-macaque model with the HIV infection in humans, these results may be highly important for the pathogenesis of HIV infection and more generally important for the characterization and function of digestive CD8+ IEL population.

Differential distribution of galactosylceramide, H antigen, and carcinoembryonic antigen in rhesus macaque digestive mucosa.

The most common route of transmission of HIV is via the mucosa. We compared human and macaque intestinal epithelia to determine whether the SIV macaque system can be used as a model to study HIV transmission by the rectal route. The overall morphology of the macaque gut mucosa is very similar to that of humans. Differentiation markers follow the same pattern as in the human system. The carcinoembryonic antigen (CEA) is apical in epithelial cells of the rectum and is absent from the small intestine. Blood group H antigen is expressed by enterocytes but not by colonocytes or rectocytes. Galactosylceramide, a potential alternative receptor for HIV in epithelial cells, is expressed in all intestinal segments as in humans. In absorptive cells it is apical and intracellular in the rectum, colon, and cecum, whereas it is only intracellular in small intestine. In goblet cells the galactosylceramide is present in intracellular vacuoles in all segments. It is also present on the membrane of mucous granules in colon and small intestine but not in rectum. We therefore believe that the macaque digestive tract may constitute a good model for the human digestive tract in the transmission of lentiviruses.

Long-term humoral and cellular immunity induced by a single immunization with replication-defective adenovirus recombinant vector.

This study examines the suitability of replication-defective adenovirus vectors for engineering recombinant vaccines. The immunological abilities and limitations of E1-deleted adenoviruses containing the lacZ gene (Ad-beta-gal) were investigated by examining the humoral and cellular immune responses to the beta-galactosidase protein. BALB/c mice (H-2d) were given in a single injection of recombinant adenovirus. The cytotoxic T lymphocyte (CTL) response of spleen cells was evaluated. Recognized target cells were H-2d-derived tumor cells transfected by the lac Z gene, or incubated with the 876-884 beta-galactosidase peptide known to be restricted by the Ld molecule of the major histocompatibility complex. A long-lasting beta-galactosidase-specific cytotoxic T cell response was obtained. By contrast, CTL from mice immunized with the Ld-restricted peptide were less specific for the endogenous epitope presented by the transfectants expressing beta-galactosidase. Ad-beta-gal-immunized mice were also protected against an intra-cerebral challenge with a recombinant vaccinia virus expressing the lac-Z gene. These results suggest that Ad-beta-gal-induced CTL have protective abilities in vivo. The induction of beta-galactosidase-specific T helper lymphocytes and humoral IgG responses were also examined. A proliferative response occurred only late after immunization and the primed T lymphocytes produced interleukin-2, but no interleukin-4. A humoral IgG response to the beta-galactosidase protein was detected 15-30 days after a single immunization and remained stable for 6 months without boosting. Lastly, we followed the evolution of the immune response over the course of successive immunizations. The magnitude and kinetics of the cellular and humoral responses were similar to those obtained after a single immunization. Consistent with these observations, an adenovirus-specific neutralizing antibody response was detected as early as the second immunization. Thus, a single immunization with a replication-defective adenovirus recombinant vector induces long-lasting humoral and cellular immune responses specific to the transgene product.

B-cell continuous epitopes of the SIVmac-251 envelope protein in experimentally infected macaques.

The humoral immune response of 34 macaques experimentally infected with SIVmac-251 was studied using a combination of an epitope library and synthetic peptides. The course of the immune response was checked for up to 9 months postinfection with a panel of clones expressing SIV fragments. A systematic study was performed with synthetic peptides covering the whole transmembrane (TM) and external (SU) envelope proteins. Seven major immunodominant epitopes were characterized. Four are localized in the SU protein: one in the V1 region (111-130), one in the Cys loop of the V3 region (311-330) and two in the C-terminal end (501-520 and 511-530). Three are localized in the TM protein: one in the extracellular domain (601-619), one in the anchor domain (731-750) and one in the intracytoplasmic domain (861-881). Among these epitopes, only one, 601-619, was found to be reactive with all sera and can be defined as the principal immunodominant epitope.

Presence of virion protein x (Vpx) of simian immunodeficiency virus SIVmac 251 in target cells in vivo.

Localization of virion-associated protein x (Vpx) of SIVmac 251 was studied in lymph nodes and liver of six SIVmac-infected monkeys. Vpx was found associated with the network of follicular dendritic cells and macrophages in lymph nodes and/or livers from five out of six animals by immunohistochemistry. Although the humoral response to Vpx occurs in only 50% of the animals, the presence of Vpx in target cell or antibodies to Vpx in all the monkeys studied, suggests that Vpx may be necessary for viral replication in vivo.

Simian immunodeficiency virus as a model for vaccination against HIV. Induction in rhesus macaques of GAG- or NEF-specific cytotoxic T lymphocytes by lipopeptides.

The protection against infection by HIV probably requires the induction of both neutralizing Abs and CTL responses. Vaccination by attenuated HIV is hardly acceptable and the use of viral genes inserted in recombinant living vectors needs further development, especially with respect to safety. The peptidic vaccination is a promising approach but free peptides are usually poorly immunogenic. Because potent immune responses have been obtained in mice with modified peptides such as lipopeptides, we have designed a study to assess the immunogenicity of lipopeptides in nonhuman primates. Seven lipopeptides were synthesized, derived from known immunogenic regions of the simian immunodeficiency virus (SIV) NEF and GAG proteins. Twelve rhesus macaques, randomly chosen and not selected on their MHC basis, were immunized subcutaneously with the seven lipopeptides in IFA. An MHC class I-restricted and CD(8+)-mediated CTL response has been observed in seven macaques directed against one or two of the synthetic immunizing peptides in each case. These CTLs were able to lyse autologous target cells infected with a recombinant vaccinia virus expressing the SIV nef or gag genes, suggesting that they recognized the naturally processed peptides. These activities are detectable in peripheral blood cells for at least 10 mo after the last immunization. Abs against the immunizing peptides have also been observed in all cases. This study demonstrates that lipopeptides can generate cytotoxic and humoral immune responses in a large number of unselected animals and this approach may thus be worth considering in the vaccination against HIV.

Characterization of B-cell epitopes in the envelope glycoproteins of simian immunodeficiency virus.

We identified previously a neutralizing epitope in the V2 domain of the simian immunodeficiency virus (SIVmac) external envelope protein. The present study reports identification of five additional linear epitopes of SIVmac (isolate 251) by immunological screening of a peptide library expressed in yeast, using SIVmac-infected macaque sera. Three epitopes were localized in the envelope glycoproteins and the two others in the reverse transcriptase and in the Rev regulatory protein. Antibody response against the four envelope epitopes was monitored for 2 years in 12 macaques experimentally infected by SIVmac251. These four envelope regions represent major immunodominant epitopes of the SIVmac. Two epitopes are located in the V3 domain (a.a. 311-330) of the external gp130 and near the amino terminal part (a.a. 601-619) of the transmembrane gp36, in regions similar to those identified in HIVs, demonstrating immunological similarities between the envelopes of SIVs and HIVs. SIV-specific immunodominant epitopes were also identified in the V1 (a.a. 111-130) and V2 (a.a. 171-190) domains of the external gp130. In particular, antibody response against the V2 neutralizing region seems to play some role in the control of disease progression in SIVmac-infected macaques.

Cytotoxic T lymphocyte response against multiple simian immunodeficiency virusA (SIV) proteins in SIV-infected macaques.

To identify the target proteins of CD8+ T lymphocytes we have explored the cytolytic immune responses of 12 rhesus macaques experimentally infected with the simian immunodeficiency virus (SIVmac). Target cells were autologous B cell lines presenting SIVmac proteins after infection with recombinant vaccinia viruses. The eight following proteins were studied: ENV, POL, GAG, NEF, VIF, REV, TAT, and VPX. Macaque PBMC stimulated with Con A and expanded in T cell growth factor-containing medium produced cell lines with cytolytic activity in the majority of infected animals (9/12). The structural proteins ENV, POL, and GAG were recognized by cell lines derived from nine, eight, and six macaques, respectively. The small regulatory proteins also represented efficient CTL targets, a specific activity being detected against NEF (8/12), REV (7/12), VPX (7/12), TAT (6/12), and VIF (5/12). Most cytotoxic responses (except those directed against ENV) were mediated by CD8 cells and were MHC class I restricted. Limiting dilution analysis allowed us to quantify the frequency of CTL precursors and confirmed the high immunogenicity of multiple SIV proteins. Three different patterns of response could be defined: six animals were able to recognize at least six of the eight tested target proteins, two of them reacting with all eight target proteins. The other three responder macaques reacted only against a few SIV proteins, whereas no cytotoxic activity was detected in the three remaining infected macaques and in the nine negative controls. The six animals responding against multiple proteins were still healthy 12 to 22 mo after infection with two of them presenting a decrease in circulating CD4 cells concurrently to the disappearance of the CTL response. Conversely, three nonresponder or low responder macaques developed an overt disease after 4 to 12 mo, and two other presented a very low level of CD4 cells, suggesting that the pattern of response may be of prognostic value.

Three epitopic peptides of the simian immunodeficiency virus Nef protein recognized by macaque cytolytic T lymphocytes.

In 8 of 12 experimentally infected macaques, the Nef SIVmac 251 protein was recognized by cytolytic T lymphocytes (CTL) and appeared strongly immunogenic. Here, we report experiments which have been performed by using synthetic peptides to precisely determine the epitopes recognized by macaque CTL. Three epitopes of the Nef protein have been defined as CTL targets in three macaques. The epitopic peptides are located in the central region of the protein, and all of them show high homology with peptides of the human immunodeficiency virus type 1 Nef protein recognized by human CTL in association with several human leukocyte antigen molecules. These results suggest that (i) the Nef protein is a good candidate for vaccination not only because of its early expression but also because of its high immunogenicity for CTL, (ii) long peptides covering the central region of this protein could be used as vaccines and could cross the major histocompatibility complex barrier in a large variety of individuals, and (iii) the rhesus macaque is a good animal model in which to test for protection by CTL.

HIV isolation from pulmonary cells derived from bronchoalveolar lavage.

In a study of 75 alveolar cell co-cultures from 55 HIV-seropositive subjects, p24 antigen production was identified by ELISA in 11 out of 26 (42%) of unseparated cell cultures, 15 out of 39 (38%) purified alveolar macrophage cultures and eight out of 10 purified alveolar lymphocyte samples. Positivity in unseparated cell cultures was associated with an alveolar lymphocytosis greater than 30%. Negative macrophage cultures were significantly more likely in the early stages of HIV infection with none positive when the subject had a peripheral CD4+ cell count greater than 300/microliters. Alveolar macrophage infection thus appears to increase with HIV disease progression.

Correlation between CD4 cell counts and cellular and plasma viral load in HIV-1-seropositive individuals.

We conducted a study of 152 HIV-1-seropositive individuals in order to evaluate the possible correlations between the isolation of HIV from peripheral blood mononuclear cells or from plasma and CD4 cell counts. HIV was isolated from only 36% of plasma samples, and the isolation rate was closely related to CD4 cell counts, increasing gradually from 0% in subjects with greater than 800 x 10(6)/l CD4 cells to 88% in those with less than 100 x 10(6)/l CD4 cells. In contrast, HIV was isolated from 92% of cell samples (99% in subjects with less than 900 x 10(6)/l CD4 cells, 46% in those with CD4 counts greater than or equal to 900 x 10(6)/l). Since most cell samples were positive, a scoring method was designed to quantify the cellular viral load. The results obtained demonstrated that the cellular viral load was closely related to CD4 counts. We also found that the cellular viral load was higher in subjects with either positive plasma isolation or positive p24 antigenaemia. The measurement of the cellular viral load by this scoring method appears to be useful for the management of HIV-seropositive individuals and for the evaluation of therapeutic trials.

Autoreactive cytotoxicity in HIV-infected individuals.

A possible role for autoimmunity in the pathogenesis of HIV infection has been suggested, based upon the certain degree of homology shared by HIV gp41 and MHC class II molecules. A number of humoral markers of autoimmunity have since been found in seropositive subjects. We have evaluated the cellular autoreactive response in HIV-infected individuals. Our study demonstrates the existence of a cytolytic activity, present in seropositive but not in seronegative subjects. This activity is mediated by CD3+ T cells, which only occasionally express the CD8 or the CD4 surface markers. Effector cells do not appear to exert their activity in a MHC-restricted fashion, since allogeneic target cells could also be killed, recovered from allogeneic seropositive as well as from seronegative subjects. Several types of target cells were lysed: T cell blasts and Epstein-Barr virus (EBV) transformed B cells, suggesting that the target antigen is common to at least these two cell types. The fact that cells from seronegative individuals were lysed argues against the recognition of an HIV-specific antigen. The nature of the target determinants and the identity of the effector cells are discussed.