INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The envelope of human immunodeficiency virus type 1 (HIV-1) is comprised of host cell membrane-derived proteins and lipids incorporated into the envelope when the virion buds from an infected cell (reviewed in references 34 and 48). More than 20 different host cell-derived proteins have been identified in the HIV-1 envelope, including major histocompatibility complex class I (MHC-I) and MHC-II; the adhesion molecules CD44; LFA-1, -2, and -3; and ICAM-1 and ICAM-3 (2, 4, 21, 33). These virion-associated, host cell-derived proteins can serve as markers by which to identify the type of cell from which a virion budded (4, 6, 15). The molecular phenotype of the HIV virion envelope has been used to determine whether HIV virions produced in vivo budded from a macrophage (M) or an activated T cell (27, 38). Incorporation of host cell-derived proteins into virions is not random or simply a function of expression level or density on the cell surface, since proteins that are highly expressed on infected cells, such as CD4, CD45, and the coreceptors CXCR4, CCR3, and CCR5, are not incorporated into virions (7, 15, 21, 25, 29).

Many cellular proteins incorporated into HIV-1 virions retain their biological function. For example, CD44 on the virion has been shown to bind hyaluronic acid (20) and CD55 (decay-accelerating factor) or CD59 present in the virion envelope can provide resistance to complement-mediated lysis (42, 43). The HIV virion envelope is enriched for HLA-DR but not DP or DQ (2, 6, 18, 45), and virion-associated MHC-II can bind and present the superantigen Staphylococcus enterotoxin B to resting T cells, resulting in T-cell activation (39). These observations demonstrate that virion-associated host cell proteins are functional and may play a role in HIV pathogenesis.

Normally, T cells require two signals to become fully activated. Signal one is antigen (Ag) specific and is generated by binding of the T-cell receptor (TCR) to Ag-MHC complexes on the Ag-presenting cell (APC). The second signal, a costimulatory signal, is generated by CD28 on the T cell interacting with CD80 (B7-1) or CD86 (B7-2) on an APC (reviewed in reference 19). We have previously reported that microvesicles and HIV-1 virions incorporate high levels of MHC-I and MHC-II upon budding (2, 5) and have hypothesized that virion- or microvesicle-associated MHC-I or MHC-II, with or without bound antigenic peptides, could bind to and signal through the TCR on responding T cells. It has not been previously determined whether CD80 and CD86 are incorporated into budding HIV-1 virions or microvesicles. Because TCR signaling in the absence of costimulation can lead to anergy or apoptosis, we examined whether microvesicles and/or HIV-1 virions incorporate CD80 or CD86 into their membranes. Here we report that HIV infection of cell lines, M, and primary peripheral blood mononuclear cells (PBMC) upregulates cell surface expression of MHC-II and that virions derived from all of these cells incorporated MHC-II. CD86 was detected on virions produced from 17 of 21 sets of different virus isolates propagated on different cells, whereas CD80 was detected on virions from only 3 of the same 21 viruses produced from CD80- and CD86-expressing cells. Microvesicles were also enriched for CD86, whereas CD80 was excluded. CD45 was identified as a protein that was highly expressed on microvesicles but not on HIV-1 virions.

These data suggest that HIV has evolved to preferentially incorporate some immunoregulatory proteins, such as MHC-II and CD86, but to exclude other proteins like CD45 and CD80. The host cell molecules incorporated into virions influenced the biological effects of the virus. Noninfectious, MHC-containing HIV virions derived from the CEMX174/T1 cell line triggered cell death in resting PBMC, whereas noninfectious, MHC-negative virions derived from the matched CEMX174/T2 cell line did not. These findings suggest that HIV has evolved to preferentially incorporate certain immunoregulatory proteins into virions, potentially contributing to the ability of the virus to evade the immune system and contribute to pathogenesis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell lines. Uninfected cell lines H9 (13), CEMX174/T1, CEMX174/T2 (44), and TBLCL-CD4 (30) were cultured in RPMI 1640 medium with 5% heat-inactivated fetal bovine serum, 2 mM L-glutamine, penicillin G at 100 U/ml, and streptomycin sulfate at 100 µg/ml (complete medium). Chronically HIV-1-infected cell lines MN/H9, NL4-3/H9, NL4-3/CEMX174/T1, NL4-3/CEMX174/T2, and NL4-3/TBLCL-CD4 were also cultured in complete medium. All cell lines were split twice weekly at 3 × 10⁵ cells/ml, were mycoplasma negative (PCR Mycoplasma Detection Kit; American Type Culture Collection, Manassas, Va.), and were cultured in complete medium.

Virus stocks and preparation of virions inactivated by Aldrithiol-2 (2,2'-dithiodipyridine). H9, CEMX174/T1, CEMX174/T2, and TBLCL-CD4 cells chronically infected with HIV-1_NL4-3 were cultured as described previously (35). For experiments involving the induction of cell death, conformationally authentic noninfectious HIV-1 virions were prepared as previously described (1, 40). Concentrated (1,000×) virus preparations were produced by sucrose density gradient banding in a continuous-flow centrifuge (1, 5, 40). For the virion precipitation experiments, different HIV-1 isolates were examined, including patient isolates PI08-436, P2-285, P419, and P115 derived from ex vivo expansion of primary PBMC (a generous gift from Antonio Valentin, National Cancer Institute [NCI] at Frederick). Clade B, R5 patient isolates HIV-1_91US054, HIV-1_92US727, and HIV-1_92US657 (49), grown in PBMC activated with phytohemagglutinin (PHA) plus interleukin-2 (IL 2; 10 U/ml; Hoffman-La Roche, Nutley, N.J.), were acquired from the National Institute of Allergy and Infectious Diseases AIDS Research and Reference Reagent Program. HIV-1_SF162 (R5) (10), HIV-1_89.6 (X4 and R5 dual tropic) (12), and HIV-1_NL4-3 (X4) were also grown in PBMC activated with PHA-plus-IL-2 (47). HIV-1_Adn-M (17) and HIV-1_Ba-L (16) were produced from primary monocyte-derived M (MDM) cultures (see below). Microvesicles, used as a control reagent, were isolated from supernatants of uninfected cell cultures in a manner identical to that used for virus preparation from infected cells (5). All virus and microvesicle stocks were stored at 70°C or in vapor phase liquid nitrogen until use.

Isolation and culture of PBMC and M. PBMC were isolated by density centrifugation (Ficoll-Hypaque; Pharmacia, Uppsala, Sweden) from citrate-anticoagulated peripheral blood obtained from healthy, HIV-1-seronegative donors at the NCI at Frederick. PBMC were cultured in AIM-V medium (Gibco, Gaithersburg, Md.) with 2% human AB serum (Sigma, St. Louis, Mo.). Elutriated monocytes from HIV-negative donor leukopacs were grown at 2 × 10⁶ cells per well on ultralow-attachment six-well Costar plates in RPMI 1640 medium (Biosource International, Camarillo, Calif.) supplemented with penicillin, streptomycin, gentamicin, amphotericin B, L-glutamine (Quality Biological, Gaithersburg, Md.), HEPES buffer (Sigma), and 10% fetal bovine serum (Biosource International). The monocytes were incubated at 37°C under 7% CO₂ and 90% humidity for 7 days to generate MDMs. MDMs were infected with 10 50% tissue culture-infective doses of either HIV-1_Ba-L or HIV-1_ADA for 2 h, washed with phosphate-buffered saline (PBS; Biosource International) to remove free virus, and refed with culture medium. The infected MDMs were incubated for an additional 18 days with medium changes every 5 days. MDMs were stained on day 18 for intracellular HIV-1 core antigen using the KC57 monoclonal antibody (MAb; Beckman-Coulter, Miami, Fla.) and determined to be greater than 80% infected (data not shown). Culture supernatants were found to be positive for HIV-1 p24 by enzyme-linked immunosorbent assay (Beckman-Coulter) at 14 days postinfection. At day 18 postinfection, culture supernatants were harvested and the MDMs were recovered by centrifugation for flow cytometric analysis.

Cell counts and viability. Total cell numbers and viability were determined by trypan blue analysis. Cells were counted on a hemocytometer in triplicate, and the percentage of dead cells was determined by the formula [dead/(live + dead)] × 100. Error bars represent 1 standard deviation of the mean. P values were calculated by using a one-tailed, equal-variance Student t test of experimental measurements versus a PBS control. Statistical analysis was performed with Microsoft Excel (Microsoft, Redmond, Wash.).

Flow cytometry. Immunofluorescent staining of PBMC and MDMs (3 × 10⁵ per condition) was performed at 4°C for 30 min by using isotype immunoglobulin G1 (IgG1) (X40), IgG2a (X39), and V4 (non-gp120-interacting domain on CD4) and HLA-DR (L243) MAbs from Becton Dickinson Immunocytometry Systems (San Jose, Calif.). MAbs reactive with CXCR4 (12G5), CCR5 (3A9), CD45 (HI30), CD55 (IA10), CD80 (L307.4), CD86 (IT2.2), and MHC-I (G46-2.6) were all purchased from Pharmingen (San Diego, Calif.). All antibodies were phycoerythrin coupled. Following antibody staining, cells were washed three times with 250 µl of staining buffer and fixed with 2% paraformaldehyde overnight at 4°C prior to data acquisition on a FACS Calibur flow cytometer using CellQuest software (Becton Dickinson Immunocytometry Systems). Samples were gated on viable cells by forward and 90° light scatter, and at least 15,000 live-cell events were acquired for each sample. Acquired data were analyzed by using FlowJo software (Tree Star, Inc., San Carlos, Calif.).

Western blot analysis. Cells, virions, and microvesicles were solubilized in lysate buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 0.05 M Tris hydrochloride buffer [pH 7.5], 0.15 M NaCl, 1 mM EDTA, 1% aprotinin, 1 mM phenylmethylsulfonyl fluoride). Cell lysates were cleared by microcentrifugation at 12,000 × g for 5 min at 4°C. HIV-1 virions and microvesicles (50 µg of total protein equivalents per lane) for electrophoresis were run separately on discontinuous SDS-polyacrylamide (4 to 20% gradient) gels under nonreducing or reducing conditions. Proteins were transferred onto Immobilon-P membranes by a semidry blotting technique (Millipore, Bedford, Mass.), and specific proteins were detected by immunoblot analysis with a MAb against CD45 (HI30; Pharmingen), a rabbit polyclonal Ab to CD55 (H-319; Santa Cruz Biotechnology, Santa Cruz, Calif.), a goat polyclonal IgG against CD80 (N-20; Santa Cruz Biotechnology), a mouse MAb to CD86 (IT2.2; Pharmingen), a mouse MAb to MHC-I (a generous gift from Hidde Ploegh, Massachusetts Institute of Technology, Cambridge), or a mouse MAb to MHC-II (L243; American Type Culture Collection). Primary antibodies were detected with horseradish peroxidase-conjugated, species-specific goat secondary antibodies (Bio-Rad, Hercules, Calif.) and enhanced-chemiluminescence reagents (Amersham, Arlington Heights, Ill.).

VPA. A whole-virion immunoprecipitation assay (VPA) was performed essentially as previously described (2, 40), except that it was performed with a 96-deep-well (2.2 ml) plate (Marsh Biomedical Products, Inc., Rochester, N.Y.) or microcentrifuge tubes. Comparable input amounts of infectious or Aldrithiol-2-inactivated virus preparations (p24^CA at 10,000 pg/ml or reverse transcriptase equivalents at 2,500 pg/ml) were incubated overnight at 4°C on a rocker with each MAb at 10 µg/ml in PBS plus 3% bovine serum albumin (BSA) in a total volume of 500 µl in deep-well plates sealed with aluminum plate sealers (Beckman, Fullerton, Calif.). Pansorbin cells (formalin-fixed Staphylococcus aureus strain Cowan; 25 µl; Calbiochem, La Jolla, Calif.) were incubated with PBS-3% BSA or with rabbit anti-mouse IgG (Sigma) under saturating conditions and washed three times in PBS plus 3% BSA. Pansorbin-Ab complexes were added directly to virus complexed with the mouse MAbs, and after incubation at 20°C for 30 min with rocking, virion Ab-Pansorbin complexes were precipitated by centrifugation (2,000 × g, 30 min). The residual virus content of the supernatant after immunoprecipitation was determined by p24 capture immunoassay (AIDS Vaccine Program, NCI at Frederick) or reverse transcriptase assay (Cavidi). The MAbs used in the VPA were the same MAbs used in the flow cytometry experiments. Clearance by a particular Ab in this assay is indicative of the presence of immunoreactive antigens on the virion surface (2). It is likely that a threshold density of a host cell-derived protein in the virus envelope is required to precipitate the virus and that the amount of virus precipitated depends in part on the density of a given protein in the envelope of the virus. However, because the VPA readout involves quantitation of a viral protein, this assay measures how many virions have been precipitated by the Ab-Pansorbin complex and not the number of host cell-derived proteins on a virion. Adding a rabbit anti-mouse secondary Ab to the Pansorbin cells allowed us to detect CD80 on virions that appeared to be CD80 negative when precipitated with the anti-CD80 MAb alone (D.G., unpublished observation). Error bars represent 1 standard deviation of the mean of triplicate measurements. P values were calculated by using a one-tailed, equal-variance Student t test of experimental measurements versus isotype control measurements. Statistical analysis was performed by using Microsoft Excel. Proteins for which immunoprecipitation with a specific MAb yielded a value statistically significantly greater than the value for the isotype control were considered to be incorporated into the virions at significant levels.

HLA-DR genotyping. DRB1 genotyping was performed by using a combination of PCR sequence-specific priming (31) and single strand-strand conformation polymorphism (8) analyses.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Differential incorporation of CD80, CD86, MHC-I, and MHC-II into HIV-1 virions. HIV preferentially incorporates or excludes different host cell proteins when budding from an infected cell. We have hypothesized that the presence of MHC molecules or the costimulatory protein CD80 or CD86 in the HIV-1 virion envelope could contribute to HIV pathogenesis (14). HIV incorporates MHC-I and MHC-II upon budding from infected T cells or macrophages in vitro (2, 6, 9) and in vivo (26, 27, 41), but it had previously not been determined whether the costimulatory proteins CD80 and CD86 are also incorporated into the HIV-1 virion envelope. By using a sensitive, Ab-based VPA, we performed an initial survey of seven primary HIV isolates derived from PBMC, two M-derived isolates, and three laboratory isolates to determine whether CD80, CD86, MHC-I, and MHC-II were incorporated into the virions. All of the virions incorporated MHC-II, except the virions derived from the MHC-II-negative CEMX174/T2 cell line (Fig. 1). None of the virions incorporated significant levels of CD80, and 9 of the 12 viruses incorporated CD86 (Fig. 1). There was variable incorporation of MHC-I into the virions, depending on the virus and the cells from which the virus was produced (Fig. 1). These data suggested that, depending on the virus and the cell from which it was derived, there could be differential incorporation of CD80, CD86, MHC-I, and MHC-II into the virion envelope and that CD86 was more readily incorporated into virions than was CD80.

View larger version (32K): [in this window] [in a new window]   FIG. 1.   Survey analysis of differential incorporation of CD80, CD86, MHC-I, and MHC-II into virions of a panel of HIV-1 isolates propagated in PBMC, M, or cell lines. Immunoprecipitation of primary HIV-1 isolates, M-tropic isolates, and cell line-adapted virions was performed by using a MAb-based VPA as described in Materials and Methods. MAbs to immunoregulatory proteins CD80, CD86, MHC-I, and MHC-II were used to characterize primary isolates (108-436, 2-285, P419, 115, 92US657, 92US727, and 91US054 [blue shades]), M isolates (Ada-M 98-4 and Ba-L 98-6 [red shades]) and cell line-adapted isolates (MN/H9, NL4-3/T2, and NL4-3/TBLCL-CD4 [green shades]). A polyclonal antiserum raised against microvesicles derived from the TBLCL-CD4 cell line served as a positive control for maximal virion precipitation, and an isotype-matched mouse anti-CD4 MAb served as a negative control. The data shown are representative of two independent experiments performed in triplicate with <10% variability in the magnitude of virion clearance. Error bars represent 1 standard deviation of the mean of triplicate measurements. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student's t-test significance of differences between experimental measurements and isotype control measurements).

Profile of immunoregulatory molecules incorporated into M-derived HIV-1 virions. Monocyte-derived M expressed low to moderate levels of CD4 and both coreceptors CXCR4 and CCR5 (Table 1). M infected with Ada-M, Ba-L 98-4, and Ba-L 98-7 showed increased expression of CD45, CD55, CD86, MHC-I, and MHC-II (Table 1). The uninfected and infected M expressed low levels of CD86 and low to undetectable levels of CD80 (Table 1). Characterization of the proteins incorporated into the M-derived virions revealed that CD80 was detectable on the Ba-L 98-7 and Ada-M 98-3 virions and CD86 was detectable on the Ba-L 98-4, Ba-L 98-7, and Ada-M 98-3 virions. Interestingly, the M-derived virions did not incorporate detectable amounts of CD45 or CD55 (Fig. 2), despite moderate levels of CD45 and CD55 expression on the M (Table 1). Lastly, all three M tropic viruses incorporated significant levels of MHC-I and MHC-II (Fig. 2). These data support the premises that MHC-II is preferentially incorporated into M-derived virions and that CD55 and CD45 are preferentially excluded.

View this table: [in this window] [in a new window]   TABLE 1.   Cell surface expression of cellular proteins on uninfected and HIV-1-infected Ma

View larger version (36K): [in this window] [in a new window]   FIG. 2.   Profile of immunoregulatory molecules incorporated into HIV-1 virions produced from infected M. M were isolated and cultured as described in Materials and Methods. M were mock infected or infected with Ada-M 98-3, Ba-L 98-4, or Ba-L 98-7. M- derived virions were characterized for the presence of CD4, CD45, CD55, CD80, CD86, MHC-I, and MHC-II in the virion envelope. The data shown are representative of two separate experiments, each performed in triplicate. Error bars represent 1 standard deviation of the mean of triplicate measurements. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t-test significance of differences between experimental measurements and isotype control measurements). Microves., microvesicles.

Profile of immunoregulatory molecules incorporated into PBMC-derived HIV-1 virions. We next characterized the cell surface expression and incorporation of cell surface proteins with immunoregulatory function into representative X4, R5, and dual-tropic HIV-1 virions produced from primary PBMC. The levels of CD4, CXCR4, CCR5, CD45, CD55, CD80, CD86, MHC-I, and MHC-II on activated PBMC revealed that the majority of the cells were CD4 and CXCR4 positive and CCR5 negative (Table 2). The majority of the cells expressed low levels of CD80 and moderate levels of CD55, CD86, and MHC-I. As observed with the M, CD45 was the most highly expressed molecule and cell surface MHC-II expression was increased by HIV-1 infection (Table 2).

View larger version (34K): [in this window] [in a new window]   FIG. 3.   Profile of immunoregulatory molecules incorporated into HIV-1 virions produced from infected PBMC. PBMC were isolated and cultured as described in Materials and Methods. PHA- and IL-2-activated PBMC were mock infected or infected with CCR5-tropic SF162, dual-tropic 89.6, or CXCR4-tropic NL4-3. PBMC-derived virions were characterized for the presence of CD4, CD45, CD55, CD80, CD86, MHC-I, and MHC-II in the virion envelope. The data shown are representative of two separate experiments, each performed in triplicate. Error bars represent 1 standard deviation of the mean of triplicate measurements. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t-test significance of differences between experimental measurements and isotype control measurements). Microves., microvesicles.

Comparison of the profiles of immunoregulatory molecules incorporated into cell line-derived HIV-1 virions and microvesicles. We next sought to characterize and compare the immunoregulatory proteins incorporated into HIV-1 virions and microvesicles. Microvesicles are nonviral membrane vesicle particles of unknown biological and immunological significance that bud from the surface of cells (5, 18). Identification and quantitation of cellular proteins associated with HIV-1 virions have been complicated by the presence of these microvesicles that inevitably copurify with HIV virions (5, 18). We have previously shown that microvesicles contain high levels of 2-microglobulin, MHC-I, and MHC-II (5), but it had not been previously determined whether microvesicles contain CD45, CD55, CD80, or CD86.

View larger version (72K): [in this window] [in a new window]   FIG. 4.   Virion and microvesicle preparations contain high levels of CD45, CD55, CD86, MHC-I, and MHC-II but not CD80. Virions and microvesicles derived from the T1, T2, TBLCL-CD4, and H9 cell lines were purified by sucrose density gradient ultracentrifugation. TBLCL-CD4 cell lysates served as a positive control and a way to determine the sensitivities of the different antibodies in the Western blot assays. Virion and microvesicle preparations (50 µg of total protein per lane) and the TBLCL-CD4 lysates were analyzed on an SDS-5 to 20% nondenaturing polyacrylamide gel under reducing or nonreducing conditions. Immunoblots were probed with a MAb to CD45 (HI30), a polyclonal serum to CD55 (H-319), a goat polyclonal serum to CD80 (N-20), a MAb to CD86 (IT2.2), a MAb to MHC-I, and a MAb to MHC-II (L243). The results shown are representative of at least three independent Western blot assays for each protein.

View larger version (43K): [in this window] [in a new window]   FIG. 5.   Profile of immunoregulatory molecules incorporated into HIV-1 virions produced from chronically infected continuous cell lines. Chronically infected cell lines were maintained in culture as described in Materials and Methods. Cell-free HIV-1NL4-3 derived from the T1, T2, TBLCL-CD4, and H9 cell lines was purified by sucrose density gradient ultracentrifugation. The purified virion preparations were characterized for the presence of CD4, CD45, CD55, CD80, CD86, MHC-I, and MHC-II in the virion envelope. The data shown are representative of three separate experiments, each performed in triplicate. Error bars represent 1 standard deviation of the mean of triplicate measurements. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t-test significance of differences between experimental measurements and isotype control measurements). Microves., microvesicles.

Host cell-derived HIV-1 virion-associated proteins affect virion-triggered cell death. As described in this report and elsewhere, HIV incorporates MHC molecules when it buds from infected cells. We have postulated that virion-associated, host cell-derived proteins might play a role in HIV pathogenesis (2), but previously it has been difficult to distinguish between cell death due to the direct effects of viral replication and lysis from indirect effects due to noninfectious virions. Specifically, we have proposed that MHC molecules incorporated into the HIV virion can interact with the TCR and other receptors on the surface of a T lymphocyte to induce anergy or apoptosis (2, 39). We have recently developed a procedure by which to inactivate HIV infectivity without affecting the conformational integrity of the virion surface proteins (1, 40). These conformationally and functionally intact but noninfectious virions interact authentically with target cells and provide a powerful tool with which to evaluate the role host cell-derived proteins present on the HIV-1 virion play in pathogenesis, independently of productive infection.

View larger version (23K): [in this window] [in a new window]   FIG. 6.   Virion-associated cellular molecules play a role in virion-triggered cell death. The effect of virion-associated, host cell-derived molecules in HIV pathogenesis was examined by using conformationally authentic, noninfectious HIV-1 virions. Noninfectious, Aldrithiol-2-inactivated virions (p24CA equivalents at 50 ng/ml) or microvesicles (total protein at 10 µg/ml) were used to pulse resting PBMC from a healthy, HIV-seronegative donor. After 10 days, the PBMC were enumerated for total cell numbers and percent viability by trypan blue analysis. The data shown are representative of three separate experiments. Error bars represent 1 standard deviation of the mean of triplicate measurements. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t-test significance of differences between experimental measurements and PBS [control] measurements).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A hallmark of HIV infection is the functional impairment of CD4⁺ T lymphocytes that precedes an eventual decline in circulating CD4⁺ T cells. The mechanism(s) behind this HIV-induced unresponsiveness or "anergy" and eventual apoptosis of CD4⁺ T cells remains unclear. Here we propose that host cell-derived immunoregulatory proteins present in the envelope of noninfectious virions could impact HIV pathogenesis. Specifically, binding of gp120 to CD4, virion-associated MHC molecules to TCRs, and virion-associated CD86 to CD28 on T lymphocytes could lead to T-cell activation, differentiation, anergy, or apoptosis. T cells normally require two signals to become fully activated. Signal one is Ag specific and is initiated by TCR binding to Ag-MHC complexes on the APC. The second, or costimulatory, signal is generated by CD28 on the T cell binding to CD80 or CD86 on the APC (reviewed in reference 19). We therefore determined whether HIV-1 virions or microvesicles incorporate CD80 or CD86. When a sensitive immunoprecipitation procedure was used, CD80 was detected on only 3 of 21 viruses derived from CD80-expressing cells whereas CD86 was detected on 17 of 21 viruses derived from CD86-expressing cells (Fig. 1, 2, 3, and 5). Additionally, virions and microvesicles derived from the T1, T2, and TBLCL-CD4 cells preferentially incorporated CD86 compared to CD80 (Fig. 4 and 5), despite approximately equivalent levels of CD80 and CD86 expression on the three cell lines (Table 3). These results suggest that CD86 is generally incorporated into budding virions and microvesicles, whereas CD80 is generally excluded. The molecular mechanisms behind the preferential incorporation of CD86 and exclusion of CD80 remain to be elucidated, but this phenomenon could be mediated by the cytoplasmic domains, which bear no similarity to one another (3).

The immunological significance of microvesicles enriched for MHC-I, MHC-II, and CD86 but not CD80 is also unclear. However, in some experiments, microvesicles have suppressed virus-specific T-cell responses (M. T. Esser, unpublished observation). CD80 and CD86 do not simply play redundant roles in the immune system (19). Antibodies that bind CD86 block the development of Th2 T cells and can exacerbate inflammation, whereas Abs that bind CD80 can reduce the severity of inflammation in certain models of autoimmunity (24, 37). These and other studies raise the possibility that interactions with CD86 present on virions and microvesicles may help differentiate naive T cells into Th2-like effectors (11, 46). Interestingly, there is an increase in the percentage of CD86-expressing CD4⁺ T lymphocytes in HIV-infected individuals (A. Valentin, personal communication). Microvesicles may be a mechanism the immune system uses to down-regulate ongoing inflammatory responses. HIV and other viruses may have exploited this microvesicle secretion pathway as a way to enhance virion assembly and as a mechanism to suppress a T-cell-mediated immune response.

We also undertook these studies in the hope of identifying a protein present on microvesicles that was not present on HIV-1 virions. As mentioned previously, microvesicles can be roughly the same size as HIV virions and band at the same density (1.13 to 1.16 g/ml) as HIV-1 virions in a sucrose gradient and are an inevitable contaminant of all HIV-1 preparations (5, 18). Toward this end, we identified CD45 as a molecule that was present at high concentrations on microvesicles but was not detected on virions (Fig. 1, 2, 3, and 5), despite being the most highly expressed protein on all of the cells examined (Tables 1, 2, and 3). These results extend the findings of Nguyen and Hildreth that CD45 was not incorporated into HIV-1_RF derived from the Jurkat T-cell line (29). Importantly, the presence of CD45 on microvesicles but not on virions may provide a way in which to purify HIV-1 virions of contaminating microvesicles. Microvesicle-free HIV-1 preparations would have practical applications for biochemical analyses. The ability to remove microvesicles from purified virus preparations may also be advantageous for the production of inactivated HIV-1 vaccines.

The mechanism(s) that determines which proteins are incorporated into the budding HIV-1 virion is not well understood. The incorporation of immunoregulatory proteins into virions was not random, since some highly expressed proteins, like CD45, were excluded from virions while others, like MHC-II, appeared to be specifically incorporated (Fig. 1, 2, 3, and 5). Nguyen and Hildreth have proposed that HIV-1 buds selectively from glycolipid-enriched membrane domains called lipid rafts (29). Supporting this hypothesis, we found the lipid raft marker CD55 on T-cell-derived virions, suggesting that the T-cell-tropic virions budded from these rafts whereas the CD55-negative, M-tropic virions may have budded via a different mechanism. In this regard, it is worth noting that in HIV-1-infected M virions accumulate in intracellular vacuoles and are rarely seen budding from the plasma membrane (32) whereas T-cell-derived viruses bud predominantly from the plasma membrane (22, 23). CD55 may be a useful marker with which to dissect the different pathways that M-tropic and T-cell-tropic virions use to egress from an infected cell. It is also possible that HIV-1-encoded proteins directly bind to host cell proteins to facilitate their incorporation into the mature virion. We have previously reported that a 43-amino-acid region in the cytoplasmic tail of gp41 is required for the efficient incorporation of MHC-II, but not MHC-I, into HIV-1 virions derived from T-cell lines and PBMC (36). These results suggest that HIV-1 may have evolved to specifically incorporate MHC-II into the virion as a mechanism of immune evasion.

Regardless of the specific mechanism that HIV uses to incorporate host cell-derived proteins, it is clear from the experiments with HIV-1_NL4-3-AT2 virions that host cell-derived proteins can dramatically affect viral pathogenicity (Fig. 6). The HLA-DR genotype of the T1, T2, TBLCL-CD4, and H9 cell lines; the phenotype of the virion envelope; and whether the HIV-1_NL4-3-AT2 virions triggered cell death are summarized in Table 4. MHC-containing, T1-derived virions triggered cell death, whereas MHC-negative, T2-derived virions did not (Fig. 6 and Table 4), strongly suggesting that virion-associated MHC molecules contribute to HIV pathogenesis. Importantly, the T1- and T2-derived virions differed only in MHC expression (Fig. 5 and Table 4) due to the fact that the T2 cell line has a deletion in chromosome 6 (44). Interestingly, AT2-inactivated NL4-3/H9 virions did not trigger cell death despite containing MHC-II (Fig. 6). This may have been due to the fact that H9-derived virions did not contain CD86, whereas the T1- and TBLCL-CD4-derived virions did (Fig. 5 and Table 4), or it may have been due to the fact that the H9-derived virions contained HLA-DR 0400, whereas the T1-derived virions contained HLA-DR 0701 and the TBLCL-CD4-derived virions contained HLA-DR 1501, 1104 (Table 4). Future studies will determine whether the HLA-DR phenotype of the virus or the responder PBMC affects HIV-1-triggered apoptosis.

Virion-associated MHC molecules could play several roles in HIV pathogenesis. The natural ligands for MHC-II are the TCR and CD4, and virion-associated MHC-II could enhance the avidity of the virion to increase infectivity as reported by Cantin et al. (6). Alternatively, virion-associated MHC-I and MHC-II could bind to TCRs on CD8⁺ and CD4⁺ T lymphocytes, respectively, to trigger apoptosis. Because the AT2-inactivated virions were not infectious, our data favor the second interpretation. Noninfectious-virion-triggered cell killing is especially relevant in the light of recent data from Lawn and Butera demonstrating that virions isolated from patient plasma during primary viremia did not contain MHC-II molecules whereas virions isolated late in infection or from patients with opportunistic infections contained high levels of MHC-II (26). Our current evidence supports the hypothesis that HIV has evolved specific strategies by which to acquire MHC-II as a way to thwart the host immune response.

In summary, this study demonstrated that the incorporation of host cell proteins into virions and microvesicles was not random. HIV-1 infection of M, PBMC, and cell lines increased cell surface expression of MHC-II, and all of the viruses examined incorporated MHC-II. CD86, but not CD80, was preferentially incorporated into both microvesicles and virions. CD45 was identified as a molecule that was highly expressed on microvesicles but excluded from virions. Studies with noninfectious HIV-1_NL4-3-AT2 virions revealed that host cell-derived proteins can dramatically affect the pathogenicity of HIV-1 virions. Dissection of the mechanisms by which HIV acquires host cell immunoregulatory proteins and the role virion-associated host cell proteins play in triggering cell death will advance our understanding of HIV pathogenesis.