Chimeric human/murine monoclonal IgM antibodies to HIV-1 Nef antigen expressed on chronically infected cells

Human IgM antibody (Ab) to gangliosides induced cytolysis of HIV-1-infected cells by homologous human complement. We expected that any human IgM Ab reactive with HIV-1 infected cells could cause complement-mediated cytolysis. The trans-chromosome mouse (TC mouse) contains human chromosomes harboring genes responsible for immunoglobulin production. Spleen cells from TC mice immunized with recombinant Nef were fused with mouse myeloma cells to generate hybridomas, and we selected those that produced human mu-chain-positive Abs reactive with Nef fixed on an ELISA plate. However, the L-chain of the monoclonal Abs (mAbs) were murine lambda in type and were chimeric, and we could not succeed in obtaining mAb with human mu- and human kappa-chains. The chimeric mAbs reacted with the HIV-1 infected cells as seen with flow cytometric analysis, and the surface expression of Nef was also detectable on chronically infected OM10.1 cells which had no detectable gp120. However, although the reaction of the chimeric IgM mAb with HIV-1-infected MOLT4 cells induced C3 deposition on cell surfaces on incubation with fresh human serum, the cells remained unlysed, as determined by 51Cr release assay. The amount of Nef antigen on the cells might not have been high enough to overcome the function of HRF20 (CD59) that restricts formation of membrane attack complexes of homologous complement. However, combination of anti-Nef IgM mAb with other IgM mAbs reactive with the surface of HIV-1-infected cells may induce a synergistic effect in complement mediated cytolysis.     

Diagnostic surface expression of SWAP-70 on HIV-1 infected T cells

Following immunization with HIV-1 infected cells, a hybridoma cell line termed 9F11 was established from the P3U-1 myeloma line fused with lymphocytes from a trans-chromosome (TC) mouse, that harbors human chromosomes containing immunoglobulin genes. The 9F11 human IgM monoclonal antibody (9F11 Ab) reacts with HIV-1 infected MOLT4 cells but not with uninfected MOLT4 cells, and causes immune cytolysis with homologous human complement at a concentration as low as 0.4 microg/ml. This Ab was used to perform immunoscreening of a cDNA expression library derived from HIV-1 infected cells. All positive cDNA clones contained SWAP-70 cDNA. SWAP-70 RNA and protein expression are much stronger in HIV-1 infected cells. SWAP-70 was also detected on the surface of HIV-1 infected cells by flow cytometric analysis. The monocyte cell line U937 cells expresses SWAP-70 on its cell surface regardless of whether it was infected with HIV-1. Furthermore, among PBMCs surface expression of SWAP-70 was detected on CD21+, CD56+ and CD14+ cells. Although CD3+ cells scarcely express SWAP-70 on their surface, once activated, they become positive. SWAP-70 may therefore serve as a marker for T cell differentiation as well as for HIV-1 infection.     

The IgM antibody level against ganglioside GM2 correlates to the disease status of HIV-1-infected patients

HIV-1 infection induces the expression of high level of GM2 ganglioside on infected cells and IgM antibody (Ab) against GM2 can cause complement (C)-mediated cytolysis of HIV-1-infected cells. Since GM2 is immunogenic in human, we proposed that an anti-GM2 IgM Ab may be produced by some HIV-1-infected patients and the titer of this Ab might provide some insight into the progress of the disease. On this premise, the amount of IgM Ab against GM2 was determined in 124 HIV-1-infected patients and 111 seronegative donors. As expected, the anti-GM2 IgM Ab titers of the patients was significantly higher than that of the seronegative donors while the total IgM levels remained unchanged. In addition, we determined the CD4+ cell count and the HIV-RNA load in the HIV-1-infected patients. The results showed a positive correlation between the anti-GM2 IgM Ab titer and CD4+ cell count but a negative correlation between the anti-GM2 IgM Ab titer and HIV-RNA load. These suggest that anti-GM2 IgM Ab induced and/or enhanced by HIV-1 infection causes C-mediated cytolysis of HIV-1-infected cells in vivo to a certain extent, and may help lower the plateau level of the HIV-RNA load. Therefore, the amount of IgM Ab against GM2 may be related to the prognosis of HIV-1 infected patients.     

NK cell lysis of HIV-1-infected autologous CD4 primary T cells: requirement for IFN-mediated NK activation by plasmacytoid dendritic cells

In vivo, several mechanisms have been postulated to protect HIV-1-infected cells from NK surveillance. In vitro, previous research indicates HIV-1-infected autologous CD4(+) primary T cells are resistant to NK lysis. We hypothesized that NK lysis of HIV-1-infected target cells would be augmented by the presence of accessory cells and/or accessory cell factors. In this study, we show that stimulation of plasmacytoid dendritic cells (PDC) with the TLR9 agonist, CpG ODN 2216, triggered NK lysis of HIV-1-infected autologous CD4(+) primary T cells. PDC-stimulated NK lysis was dependent upon MHC class I (MHC-I) down-regulation on infected cells, and primary HIV-1 isolates that exhibited enhanced MHC-I down-regulation were more susceptible to NK-mediated lysis. PDC-stimulated NK lysis of HIV-1-infected autologous CD4(+) primary T cells was blocked by neutralizing Abs to type 1 IFN and was perforin/granzyme dependent. Overall, our data suggest that HIV-infected cells are not innately resistant to NK lysis, and that exogenous NK stimulation derived from PDC can trigger NK cytotoxicity against HIV-1-infected autologous CD4(+) primary T cells.     

Presence of IgM Antibodies Which Sensitize HIV-1-Infected Cells to Cytolysis by Homologous Complement in Long-Term Survivors of HIV Infection

Although human cells are resistant to homologous human complement due to the presence of species-specific membrane inhibitors, a naturally occurring IgM antibody which recognizes an asialo-oligosaccharide can sensitize HIV-1-infected cells for complement-mediated cytolysis. Therefore, we investigated whether long-term survivors of HIV-1 infection harbor such antibodies in their sera. Thirty of 31 sera from HIV-1 seropositive hemophilia patients who have survived HIV-1 infection 10 years or more showed appreciable cytolytic activity, while only 2 sera of 10 seropositive patients presumed to have been infected with HIV-1 (due to sexual contact) more recently showed cytolytic activity. On the other hand, only 7 out of 43 sera from seronegative hemophilia patients showed cytolytic activity. Immunofluorescence staining for IgM on HIV-L -infected cells essentially correlated with the cytolytic capacity of the sera. Therefore, naturally occurring IgM antibodies and/or generated IgM antibodies reactive with the HIV-L -infected cells in patients might have been responsible for long-term survival due to complement-mediated immune cytolysis which may, in conjunction with cytotoxic T lymphocytes, synergistically suppress the infected cells in vivo. Therefore, the transfusion of such IgM antibodies could be effective for the treatment of HIV-L -infected individuals.     

An HIV-1 gp120 envelope human monoclonal antibody that recognizes a C1 conformational epitope mediates potent antibody-dependent cellular cytotoxicity (ADCC) activity and defines a common ADCC epitope in human HIV-1 serum

Among nonneutralizing HIV-1 envelope antibodies (Abs), those capable of mediating antibody-dependent cellular cytotoxicity (ADCC) activity have been postulated to be important for control of HIV-1 infection. ADCC-mediating Ab must recognize HIV-1 antigens expressed on the membrane of infected cells and bind the Fcγ receptor (FcR) of the effector cell population. However, the precise targets of serum ADCC antibody are poorly characterized. The human monoclonal antibody (MAb) A32 is a nonneutralizing antibody isolated from an HIV-1 chronically infected person. We investigated the ability of MAb A32 to recognize HIV-1 envelope expressed on the surface of CD4(+) T cells infected with primary and laboratory-adapted strains of HIV-1, as well as its ability to mediate ADCC activity. The MAb A32 epitope was expressed on the surface of HIV-1-infected CD4(+) T cells earlier than the CD4-inducible (CD4i) epitope bound by MAb 17b and the gp120 carbohydrate epitope bound by MAb 2G12. Importantly, MAb A32 was a potent mediator of ADCC activity. Finally, an A32 Fab fragment blocked the majority of ADCC-mediating Ab activity in plasma of subjects chronically infected with HIV-1. These data demonstrate that the epitope defined by MAb A32 is a major target on gp120 for plasma ADCC activity.     

Antibody-dependent cellular cytotoxicity against primary HIV-infected CD4+ T cells is directly associated with the magnitude of surface IgG binding

Antibody (Ab)-dependent cellular cytotoxicity (ADCC) is thought to potentially play a role in vaccine-induced protection from HIV-1. The characteristics of such antibodies remain incompletely understood. Furthermore, correlates between ADCC and HIV-1 immune status are not clearly defined. We screened the sera of 20 HIV-1-positive (HIV-1(+)) patients for ADCC. Normal human peripheral blood mononuclear cells were used to derive HIV-infected CD4(+) T cell targets and autologous, freshly isolated, natural killer (NK) cells in a novel assay that measures granzyme B (GrB) and HIV-1-infected CD4(+) T cell elimination (ICE) by flow cytometry. We observed that complex sera mediated greater levels of ADCC than anti-HIV-1 envelope glycoprotein (Env)-specific monoclonal antibodies and serum-mediated ADCC correlated with the amount of IgG and IgG1 bound to HIV-1-infected CD4(+) T cells. No correlation between ADCC and viral load, CD4(+) T cell count, or neutralization of HIV-1(SF162) or other primary viral isolates was detected. Sera pooled from clade B HIV-1(+) individuals exhibited breadth in killing targets infected with HIV-1 from clades A/E, B, and C. Taken together, these data suggest that the total amount of IgG bound to an HIV-1-infected cell is an important determinant of ADCC and that polyvalent antigen-specific Abs are required for a robust ADCC response. In addition, Abs elicited by a vaccine formulated with immunogens from a single clade may generate a protective ADCC response in vivo against a variety of HIV-1 species. Increased understanding of the parameters that dictate ADCC against HIV-1-infected cells will inform efforts to stimulate ADCC activity and improve its potency in vaccinees.     

Inhibition of HIV-1 infectivity through an innate mechanism involving naturally occurring IgM anti-leukocyte autoantibodies

In prior studies, we show that naturally occurring IgM anti-leukocyte autoantibodies (IgM-ALA) bind to CD3, CD4, CCR5, and CXCR4 receptors. These observations prompted us to determine whether IgM-ALA have a role in inhibiting HIV-1 infectivity by inhibiting viral entry into cells. We show that purified IgM, but not IgG, from individual sera of both normal and HIV-1 infected individuals is highly inhibitory (>95%) to HIV-1 viral infectivity both in vitro using PHA plus IL-2 activated PBL and in vivo using the human PBL-SCID mouse. Inhibition was observed with physiological doses of purified serum IgM and even after IgM was added 3 days postinfection in the in vitro assays. Absorbing purified serum IgM either with leukocytes or immobilized recombinant CD4 significantly decreased (>80%) the inhibitory effect on HIV-1 infectivity. IgM inhibited by >90% syncytia formation with the X4-IIIB infected SupT-1 cells indicating therefore that IgM inhibits viral attachment to core-receptors. IgM mediated anti-HIV-1 activity was highly specific as only certain IgM-ALA, obtained from human B cell clones inhibited HIV-1. IgM from certain HIV-1 infected individuals were not inhibitory to some R5-HIV-1 viral strains indicating that certain HIV-IgM may lack Abs reactive to strain specific coreceptor epitopes. These data indicate that an innate immune mechanism which is present from birth i.e., IgM-ALA, has a role in inhibiting HIV-1 viral entry into cells. Validation of this data with other in vivo models will be needed to determine whether in vivo administration or enhancement of IgM-ALA, e.g., through a vaccine, could prolong the asymptomatic state in HIV-1 infected individuals.     

CD4 immunoadhesin, but not recombinant soluble CD4, blocks syncytium formation by human immunodeficiency virus type 2-infected lymphoid cells.

Recombinant soluble CD4 (rCD4) has been shown to be an effective inhibitor of human immunodeficiency virus type 1 (HIV-1) and HIV-2 infection of lymphoid cells in vitro. In this report, we characterized the effects of rCD4, the V1V2 fragment of CD4, and the immunoadhesin CD4-immunoglobulin G on syncytium formation between lymphoid cells infected by HIV-1 or HIV-2 and uninfected cells. All three molecules blocked HIV-1-mediated syncytium formation, but only CD4-immunoglobulin G blocked HIV-2-mediated syncytium formation. rCD4 and the V1V2 fragment of CD4 enhanced HIV-2-mediated syncytium formation. These results suggest that the process of cell fusion is significantly different between HIV-1- and HIV-2-infected cells.     

The HIV-1 Vif protein mediates degradation of Vpr and reduces Vpr-induced cell cycle arrest

Prior work has implicated viral protein R (Vpr) in the arrest of human immunodeficiency virus type 1 (HIV-1)-infected cells in the G2 phase of the cell cycle, associated with increased viral replication and host cell apoptosis. We and others have recently shown that virion infectivity factor (Vif ) also plays a role in the G2 arrest of HIV-1-infected cells. Here, we demonstrate that, paradoxically, at early time points postinfection, Vif expression blocks Vpr-mediated G2 arrest, while deletion of Vif from the HIV-1 genome leads to a marked increase in G2 arrest of infected CD4 T-cells. Consistent with this increased G2 arrest, T-cells infected with Vif-deleted HIV-1 express higher levels of Vpr protein than cells infected with wild-type virus. Further, expression of exogenous Vif inhibits the expression of Vpr, associated with a decrease in G2 arrest of both infected and transfected cells. Treatment with the proteasome inhibitor MG132 increases Vpr protein expression and G2 arrest in wild-type, but not Vif-deleted, NL4-3-infected cells, and in cells cotransfected with Vif and Vpr. In addition, Vpr coimmunoprecipitates with Vif in cotransfected cells in the presence of MG132. This suggests that inhibition of Vpr by Vif is mediated at least in part by proteasomal degradation, similar to Vif-induced degradation of APOBEC3G. Together, these data show that Vif mediates the degradation of Vpr and modulates Vpr-induced G2 arrest in HIV-1-infected T-cells.     

Flow cytometric method to monitor the destruction of CD4+ cells following their fusion with HIV-infected cells.

Syncytium formation between HUT-78 cells persistently infected with human immunodeficiency virus type 1 (HIV-1) and uninfected CD4-bearing MOLT-4 or CEM cells results in a rapid destruction of the MOLT-4 or CEM cells. This syncytium formation is due to the interaction between the gp120 glycoprotein expressed by the persistently HIV-1-infected HUT-78 cells and the CD4 receptor present on MOLT-4 or CEM cells. A flow cytometric method has been applied to separate the infected (HUT-78) from the uninfected (MOLT-4, CEM) cell populations. This method is based on a modified DNA staining protocol which clearly shows the differences in DNA content between HUT-78 cells, on the one hand, and MOLT-4 or CEM cells, on the other hand. Using this flow cytometric method we have demonstrated that those compounds (i.e., sulfated polysaccharides, aurintricarboxylic acid) that interact with gp120 (of the HIV-infected cells) or CD4 (of the uninfected cells) suppress syncytium formation and concomitant destruction of the CD4+ cells.     

Maintenance of HIV-specific CD4+ T cell help distinguishes HIV-2 from HIV-1 infection

Unlike HIV-1-infected people, most HIV-2-infected subjects maintain a healthy CD4+ T cell count and a strong HIV-specific CD4+ T cell response. To define the cellular immunological correlates of good prognosis in HIV-2 infection, we conducted a cross-sectional study of HIV Gag-specific T cell function in HIV-1- and HIV-2-infected Gambians. Using cytokine flow cytometry and lymphoproliferation assays, we show that HIV-specific CD4+ T cells from HIV-2-infected individuals maintained proliferative capacity, were not terminally differentiated (CD57-), and more frequently produced IFN-gamma or IL-2 than CD4+ T cells from HIV-1-infected donors. Polyfunctional (IFN-gamma+/IL-2+) HIV-specific CD4+ T cells were found exclusively in HIV-2+ donors. The disparity in CD4+ T cell responses between asymptomatic HIV-1- and HIV-2-infected subjects was not associated with differences in the proliferative capacity of HIV-specific CD8+ T cells. This study demonstrates that HIV-2-infected donors have a well-preserved and functionally heterogeneous HIV-specific memory CD4+ T cell response that is associated with delayed disease progression in the majority of infected people.     

Human IgM monoclonal antibody to ganglioside GM2 and complement suppress virus propagation in ex vivo cultures of lymphocytes from HIV-1 infected patients.

HIV-1 infection induces aberrant ganglioside GM2 expression on infected cell lines, and human IgM anti-GM2 monoclonal antibody (L55 Ab) together with normal fresh human serum (FHS) as a source of complement causes complement mediated cytolysis of HIV-1 infected cells as well as HIV-1 particles. We report here that high expression of GM2 was also detected on HIV-1 infected lymphocytes from HIV-1 seropositive patients. L55 Ab effectively suppressed the generation of HIV in the presence of FHS in primarily cultured lymphocytes from HIV-1 infected patients in ex vivo experiments, and the suppression was enhanced additively by AZT. These data suggest that L55 Ab may increase the therapeutic effect of chemotherapy.     

Induction of apoptosis by HIV-1-infected monocytic cells

We have previously described a soluble 6000-Da peptide produced by an HIV-1-infected human macrophage cell line, clone 43(HIV), which induces apoptosis in T and B cells. We have identified this factor as the novel cDNA clone FL14676485 that encodes for the human hypothetical protein, FLJ21908. The FL14676485 cDNA clone was isolated from a 43(HIV) lambda ZAP Escherichia coli expression library and screened with a panel of rabbit and mouse anti-apoptotic Abs. We transfected the FL14676485 clone into Bosc cells and non-HIV-1-infected 43 cells. Western blot analysis of lysates from the FL14676485-transfected 43 cells and Bosc cells using anti-proapoptotic factor Abs revealed a protein with a molecular mass of 66 kDa corresponding to the size of the full-length gene product of the FL14676485 clone, while Western blot of the supernatant demonstrated a doublet of 46-kDa and 6000-Da peptide that corresponds to our previously described proapoptotic factor. Primary HIV-1(BaL)-infected monocytes also produce the FLJ21908 protein. Supernatants from these transfected cells induced apoptosis in PBMC, CD4(+), and CD8(+) T and B cells similar to the activity of our previously described proapoptotic factor. PCR analysis of 43 cells and 43(HIV) cells revealed a base pair fragment of 420 bp corresponding to the FL14676485 gene product in 43(HIV) cells, but not in 43 cells. The FLJ21908 protein induces apoptosis through activation of caspase-9 and caspase-3. We have further demonstrated that the FLJ21908 protein has apoptotic activity in the SH-SY5Y neuronal cell line and can be detected in brain and lymph tissue from HIV-1-infected patients who have AIDS dementia. The FLJ21908 protein may contribute to the apoptosis and dementia observed in AIDS patients.     

Vpr cytopathicity independent of G2/M cell cycle arrest in human immunodeficiency virus type 1-infected CD4+ T cells

The mechanism of CD4(+) T-cell depletion in human immunodeficiency virus type 1 (HIV-1)-infected individuals remains unknown, although mounting evidence suggests that direct viral cytopathicity contributes to this loss. The HIV-1 Vpr accessory protein causes cell death and arrests cells in the G(2)/M phase; however, the molecular mechanism underlying these properties is not clear. Mutation of hydrophobic residues on the surface of its third alpha-helix disrupted Vpr toxicity, G(2)/M arrest induction, nuclear localization, and self-association, implicating this region in multiple Vpr functions. Cytopathicity by virion-delivered mutant Vpr protein correlated with G(2)/M arrest induction but not nuclear localization or self-association. However, infection with whole virus encoding these Vpr mutants did not abrogate HIV-1-induced cell killing. Rather, mutant Vpr proteins that are impaired for G(2)/M block still prevented infected cell proliferation, and this property correlated with the death of infected cells. Chemical agents that inhibit infected cells from entering G(2)/M also did not reduce HIV-1 cytopathicity. Combined, these data implicate Vpr in HIV-1 killing through a mechanism involving inhibiting cell division but not necessarily in G(2)/M. Thus, the hydrophobic region of the third alpha-helix of Vpr is crucial for mediating G(2)/M arrest, nuclear localization, and self-association but dispensable for HIV-1 cytopathicity due to residual cell proliferation blockade mediated by a separate region of the protein.     

Different effects of Nef-mediated HLA class I down-regulation on human immunodeficiency virus type 1-specific CD8(+) T-cell cytolytic activity and cytokine production

A previous study using a Nef-defective human immunodeficiency virus type 1 (HIV-1) mutant suggested that Nef-mediated down-regulation of HLA class I on the infected cell surface affects the cytolytic activity of HIV-1-specific cytotoxic T-lymphocyte (CTL) clones for HIV-1-infected primary CD4(+) T cells. We confirmed this effect by using a nef-mutant HIV-1 strain (NL-M20A) that expresses a Nef protein which does not induce down-regulation of HLA class I molecules but is otherwise functional. HIV-1-specific CTL clones were not able to kill primary CD4(+) T cells infected with a Nef-positive HIV-1 strain (NL-432) but efficiently lysed CD4(+) T cells infected with NL-M20A. Interestingly, CTL clones stimulated with NL-432-infected CD4(+) T cells were able to produce cytokines, albeit at a lower level than when stimulated with NL-M20A-infected CD4(+) T cells. This indicates that Nef-mediated HLA class I down-regulation affects CTL cytokine production to a lesser extent than cytolytic activity. Replication of NL-432 was partially suppressed in a coculture of HIV-1-infected CD4(+) T cells and HIV-1-specific CTL clones, while replication of NL-M20A was completely suppressed. These results suggest that HIV-1-specific CD8(+) T cells are able to partially suppress the replication of HIV-1 through production of soluble HIV-1-suppressive factors such as chemokines and gamma interferon. These findings may account for the mechanism whereby HIV-1-specific CD8(+) T cells are able to partially but not completely control HIV-1 replication in vivo.     

R5 human immunodeficiency virus type 1 (HIV-1) replicates more efficiently in primary CD4+ T-cell cultures than X4 HIV-1

In this report, we present evidence that R5 human immunodeficiency virus type 1 (HIV-1) replicates more efficiently in primary CD4+ T cells than X4 HIV-1. By comparing CD3/CD28-costimulated CD4+ T-cell cultures infected by several X4 and R5 HIV-1 strains, we determined that R5-infected CD4+ T cells produce more virus over time than X4-infected CD4+ T cells. In the first comparison, we found that more cells were infected by the X4-tropic strain LAI than by the R5-tropic strain JR-CSF and yet that higher levels of viral production were detected in the R5-infected cultures. The differential viral production was partially due to the severe cytopathic effects of the X4 virus. We also compared cultures infected with the isogenic HIV-1 strains NL4-3 (X4) and 49.5 (R5). We found that fewer cells were infected by the R5 strain, and yet similar levels of viral production were detected in both infected cultures. Cell death played less of a role in the differential viral production of these strains, as the cell viability remained comparable in both X4- and R5-infected cultures over time. The final comparison involved the primary R5-tropic isolate KP1 and the primary dual-tropic isolate KP2. Although both strains infected similar numbers of cells and induced comparable levels of cytopathicity, viral production was considerably higher in the R5-infected culture. In summary, these data demonstrate that R5 HIV-1 has an increased capacity to replicate in costimulated CD4+ T cells compared to X4 HIV-1.     

Natural killer cells in perinatally HIV-1-infected children exhibit less degranulation compared to HIV-1-exposed uninfected children and their expression of KIR2DL3, NKG2C, and NKp46 correlates with disease severity

NK cells play an integral role in the innate immune response by targeting virally infected and transformed cells with direct killing and providing help to adaptive responses through cytokine secretion. Whereas recent studies have focused on NK cells in HIV-1-infected adults, the role of NK cells in perinatally HIV-1-infected children is less studied. Using multiparametric flow cytometric analysis, we assessed the number, phenotype, and function of NK cell subsets in the peripheral blood of perinatally HIV-1-infected children on highly active antiretroviral therapy and compared them to perinatally exposed but uninfected children. We observed an increased frequency of NK cells expressing inhibitory killer Ig-like receptors in infected children. This difference existed despite comparable levels of total NK cells and NK cell subpopulations between the two groups. Additionally, NK cell subsets from infected children expressed, with and without stimulation, significantly lower levels of the degranulation marker CD107, which correlates with NK cell cytotoxicity. Lastly, increased expression of KIR2DL3, NKG2C, and NKp46 on NK cells correlated with decreased CD4+ T-lymphocyte percentage, an indicator of disease severity in HIV-1- infected children. Taken together, these results show that HIV-1-infected children retain a large population of cytotoxically dysfunctional NK cells relative to perinatally exposed uninfected children. This reduced function appears concurrently with distinct NK cell surface receptor expression and is associated with a loss of CD4+ T cells. This finding suggests that NK cells may have an important role in HIV-1 disease pathogenesis in HIV-1-infected children.